Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Physiological mediators of psychosocial stressor effects on the growth of a hormone-responsive mouse.. Rowse, Gerald J. 1993

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_1993_fall_phd_rowse_gerald.pdf [ 8.9MB ]
Metadata
JSON: 1.0086495.json
JSON-LD: 1.0086495+ld.json
RDF/XML (Pretty): 1.0086495.xml
RDF/JSON: 1.0086495+rdf.json
Turtle: 1.0086495+rdf-turtle.txt
N-Triples: 1.0086495+rdf-ntriples.txt
Original Record: 1.0086495 +original-record.json
Full Text
1.0086495.txt
Citation
1.0086495.ris

Full Text

PHYSIOLOGICAL MEDIATORS OF PSYCHOSOCIAL STRESSOR EFFECTSON THE GROWTH OF A HORMONE-RESPONSIVE MOUSE MAMMARYCARCINOMAbyGERALD JAMES ROWSEB.Sc., The University of British Columbia, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTSFOR THE DEGREE OF DOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of AnatomyWe accept this thesis a conforming to the required standardUNIVERSITY OF BRITISH COLUMBIAMay 1993© Gerald James RowseIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignaDepartment of ^A notkom 1The University of British ColumbiaVancouver, CanadaDate^-:"1 tA It! 1, J 13DE-6 (2/88)ABSTRACTThis thesis investigated the role of specific immune and endocrine variables in mediatingthe effects of social housing condition on the growth of the transplantable androgen-responsiveShionogi mouse mammary carcinoma (SC 115). Mice were reared either individually (I) or group(G) housed, and then either remained in their rearing groups (II, GG) or were rehoused (IG, GI).Splenic NK cell activity of tumor cell- and vehicle-injected mice from the 4 experimentalhousing groups was investigated at 1 d post tumor cell- or vehicle-injection. Splenic NK cellactivity was suppressed in tumor cell-injected mice compared with vehicle-injected controls.Overall, mice of the GI group had significantly greater splenic NK cell activity than mice of the IGgroup. NK cell activity of tumor infiltrating lymphocytes in mice of the GI and IG groups wasthen investigated using a modification of the sponge allograft model. Overall, NK cell activitywas greater in tumor cell-injected than in vehicle-injected sponges, and was greater in tumor cell-injected sponges of mice in the GI group than in those of mice in the IG group. Finally, theeffects of modulating in vivo levels of NK cell activity on tumor growth rate were assessed.Injection ip of polyinosinic:polycytidylic acid (poly I:C, 100 Al lmg/m1) or of anti-asialo GM1(ASGMi, 100 IA of a 1:5 dilution) every 5 d for 2 wk maintained stimulation or suppression(respectively) of splenic NK cell activity relative to that in saline-injected control mice, and had noeffect on tumor growth rate in mice of the IG group. In mice of the GI group, stimulation of NKcell activity by poly I:C was accompanied by a significant stimulation of tumor growth ratecompared with that of ASGMi-injected or control mice. These studies suggest that NK cells mayplay an important role in mediating the stimulation of SC 115 tumor growth rate in mice of the GIgroup.In addition, the possibility that 1) selection for a slow growing, hormone-independentphenotype, or 2) alterations in plasma hormone levels may mediate the differential tumor growthrates was examined. Slow growing tumors from mice of the IG group had a morphologicalappearance similar to that of mice from the other experimental groups and dissimilar to that ofslow growing androgen-independent tumors grown in females. Further, tumor cells from mice ofthe IG group showed greater proliferation in response to in vitro stimulation withdihydrotestosterone or hydrocortisone than tumor cells from mice of the GI group. In addition,mice of the GG and II groups had elevated basal levels of plasma testosterone at 1 d whichdeclined significantly by 3d. Mice of the IG group had low basal plasma testosterone levels,whereas mice of the GG group had elevated basal plasma testosterone levels at all time points. Incontrast, basal plasma corticosterone levels were significantly greater in mice of the IG group thanin mice of all other groups at 1, 3 and 7 d. These data suggest that altered plasma levels ofsteroid hormones may also, in part, mediate the effects of psychosocial stressors on thedifferential tumor growth rate observed in this model, whereas selection for a subpopulation ofSC 115 cells with altered hormone responsiveness is likely not involved.Table of ContentsABSTRACT ^Table of Contents ivList of Figures ^ viiList of Tables ixAcknowledgement ^Forward ^ xiCHAPTER 1 INTRODUCTION ^ 1A) Human Studies of Stress and Cancer ^  1A.1) Definition of Stress and Psychosocial stressors ^ 1A.2) Affective Disorders and Cancer ^ 4A.3) Personality / Coping Style and Cancer 7A.4) Childhood Stress/Loss and Cancer ^ 8A.5) Life Stress/Psychosocial Stressors and Cancer. ^ 9A.6) Social Contacts and Cancer ^  10B) Animal Models Of Stressor Effects on Cancer. ^  11C) Psychoneuroendocrinology and Cancer  16C.1) Hormones and Stress ^  16C.2) Hormones and Cancer  19D) Tumors, Stress and the Immune System. ^ 29D.1) Tumor Immunology ^ 29D.2) Psychoneuroinununology 38E) Psychosocial Stressors ^ 41F) Chronic vs Acute Stressors 45G) Animal-Tumor Model for Studying Effects of Psychosocial Stressors ^ 46ivH) Thesis Objectives ^  58CHAPTER 2 GENERAL METHODS  ^ 61A. Tumor Model ^ 61A.1) Dissociation. ^  61A.2) Freezing of Tumor Cells. ^ 62A.3) Thawing of Tumor Cells 63A.4) Transplantation of Tumor Cells ^ 63A.5) Monitoring Tumor Growth 63B. Animal Model. ^ 64CHAPTER 3 IMMUNE STUDIES  ^ 66A. Completion of Splenic NK Cell Activity Time Course Assay. ^ 66Introduction ^ 66Methods And Materials ^  73Results ^  76Discussion 86B. Natural Killer Cell Activity At The Tumor Site. ^  91Introduction ^ 91Materials and Methods ^ 92Results ^ 95Discussion  102C. In Vivo Modulation of NK Cell Activity and Its Effects on the DifferentialGrowth of the SC 115 Tumor Observed in the Model. ^  110Introduction ^  110Materials and Methods ^  112Results ^  114Discussion  135viCHAPTER 4 ENDOCRINE STUDIES ^ 139A) Morphological Studies of the SC 115 Tumor Grown in Male and Female Mice. 139Introduction ^  139Methods And Materials ^  141Results: ^  143Discussion-  151B) Selected Studies of Endocrine Functioning in Mice from the 4 ExperimentalHousing Groups^  157Introduction  157Methods And Materials ^  159Results ^  161Discussion  169CHAPTER 5 DISCUSSION ^ 177REFERENCES ^ 186viiList of Figures Figure #^ Page #Figure 1. Experimental Design of Model I.^ 50Figure 2. Tumor Growth in Male Mice in the Four Housing Groups.^52Figure 3. Experimental Design of Model II.^ 55Figure 4. Tumor Growth in Male Mice in the 4 Experimental Housing Groups. ^57Figure 5. Lytic Activity of Splenic NK Cells 3 d Post Injection.^ 69Figure 6. Lytic Activity of Splenic NK Cells 1 Wk Post Injection. 72Figure 7. Relative Spleen Weights of Tumor Cell-Injected Mice From the 4Experimental Groups in the First Wk Post Injection.^ 83Figure 8. Lytic Activity of Splenic Natural Killer (NK) Cells 1 d Post Injection.^85Figure 9. Time Course Study of White Blood Cells Infiltrating Polyurethane Sponges. ^98Figure 10. Pilot Study of the Time Course of NK Cell Activity in Sponge-InfiltratingLymphocytes.^ 101Figure 11. NK Cell Activity in Sponge-Infiltrating Lymphocytes at 3 d PostInjection in Mice from the Experimental Groups.^ 104Figure 12. NK Cell Activity in Sponge-Infiltrating Lymphocytes at 7 d PostInjection in Mice from the Experimental Groups.^ 106Figure 13. Effect of a Single Injection of Anti-AsialoGM1 on In Vivo NK CellActivity in Mice Previously Treated With Poly I:C^ 116Figure 14. Effect of Dosage of Anti-AsialoGM1 and Time Post Injection onIn Vivo NK Cell Activity in Mice^ 119Figure 15. Effect of Repeated Injections of Anti-AsialoGM1 on In Vivo NK CellActivity in Mice Over Time^ 122Figure 16. Effect of Repeated Injections of Poly I:C on In Vivo NK Cell Activityin Mice Over Time.^ 124viiiFigure 17. Effect of Anesthesia on the Ability of Repeated Injections of Poly I:Cto Stimulate In Vivo NK Cell Activity in Mice Over Time.^126Figure 18. Effect of Repeated Injections of Poly I:C on In Vivo NK Cell Activity inMice Over Time.^ 129Figure 19. Effect of In Vivo Modulation of NK Cell Activity on The DifferentialTumor Growth Rates Observed in GI and IG. ^ 132Figure 20. Effect of In Vivo Modulation of NK Cell Activity on The DifferentialTumor Growth Rates Observed in GI and IG Mice at 18 Days Post Injection. 134Figure 21. Morphology of SC 115 Tumors Grown in Male and Female Mice. ^145Figure 22. Development of Osteoid-Like Regions in Tumors of Female Mice.^148Figure 23. Comparison of MSA and S-100 Staining of Serial Sections FromOsteoid-Like Regions of a Tumor Grown in a Female Mouse.^153Figure 24. Plasma testosterone levels.^ 163Figure 25. Plasma corticosterone levels. 165Figure 26. Tumor Growth in Mice of The GI and IG Groups. ^ 170Figure 27. In vitro hormone response of tumor cells from IG mice and GI mice. ^172ixList of Tables Table #^ Page #Table 1. Body Weights of Tumor Cell- and Vehicle-Injected Mice From the 4Experimental Housing Groups 1, 3 and 7 d Post Injection. ^78Table 2. Spleen Weights of Tumor Cell- and Vehicle-Injected Mice From the 4Experimental Housing Groups 1, 3 and 7 d Post Injection. ^81Table 3. Histochemical Staining of the Osteoid-like Pools in the Female Mice. ^150Table 4. Effect of Social Status (in Group Housed Mice) on Plasma Hormone Levels. ^167AcknowledementsI would especially like to thank my supervisors, Drs. Joanne Weinberg and JoanneEmerman for all the training, advice, guidance, support and friendship that they have shown meover the years. You've really helped to make this my "second home" for me. I would like tothank Drs. Reid, Waterfeild, Bellward and Worth for the expert advice and guidance that theyprovided for me on technical matters. I would like to thank Rosemary Rowan, Darcy Wilkenson,Wayne Yu and Shannon Wilson for their unfailing help and support. I would also like to thankShannon Wilson for her friendship, commiseration and many "road trips" (as graduate student lifeis about more than just research). I would like to thank the graduate students of Anatomy and myother friends in Anatomy for the friendship, encouragement and support that they provided. Iwould like to thank Jay (and the other "Blue Jays") and the Anatomy hockey pool for finallyletting me win.I would like to thank my parents for all the love, support and encouragement that theyhave always given me. Finally, I would like to thank my wife, Diane. Thank you for always beingthere to share the highs and the lows of graduate student life with me. Thank you for yourunfailing support and belief in me.xiForward Portions of this thesis have been previously published as follows:Rowse, G.J., Weinberg, J., Bellward, G.D. and Emerman, J.T. (1992). Endocrine Mediation ofPsychosocial Stressor Effects on Mouse Mammary Tumor Growth. Cancer Lett. 65:85-93. Rowse, G.J., Rowan, R.E., Worth, A.J., Reid, P.E., Weinberg, J. and Emerman, J.T. (1990). AHistological Study of the Shionogi Adenocarcinoma Grown in Male and Female Mice. Histol. Histopathol. 5:485-491.Rowse, G.J., Rowan, R.E., Weinberg, J. and Emerman, J.T. (1990). Alterations in SplenicNatural Killer Cell Activity Induced by the Shionogi Mouse Mammary Tumor. Cancer Lett. 54:81-87.For all portions of these papers that are reported in this thesis, Gerry Rowse was the majorcontributor involved in conducting the research, analyzing the data and writing the papers.CHAPTER 1: INTRODUCTION:A) Human Studies of Stress and CancerThe concept that psychological variables may alter the susceptibility of humans to diseaseprocesses, including cancer, is not new. For example, Galen observed that "melancholy women"were more prone to breast cancer than were "sanguine women" (LeShan, 1959). In theeighteenth century a number of noted British physicians recorded empirical observations of anassociation between breast cancer and personality amongst their patients (Le Shan, 1959).However, observations such as those presented above are anecdotal in nature and are subject toconsiderable observer bias. It was not until the middle of the 20th century that this questionbegan to be examined by controlled scientific studies (Levenson and Bemis, 1991). Since the1950's epidemiological studies have examined the association of such psychological attributes asaffective state, coping style and stressful life events with the occurrence, progression and/orprognosis of cancer.A.1) Definition of Stress and Psychosocial stressorsTo investigate scientifically the possible role of stress in the initiation, promotion orprogression of cancer, it is important to understand the current concepts of stress and to utilize astandardized terminology. The term stress has been much used and abused in both popular andscientific literature. The word has been used to refer interchangeably to a mental state, anexternal cue in the environment which is perceived as threatening or as the physiological changeswhich occur in response to external cues. For the purposes of this dissertation, the conventionsestablished by Selye will be used. Stress will be defined as "an alteration in the body's hormonaland neuronal secretions caused by the central nervous system in response to a perceived threat"1Further, a stressor will be defined as "a change in an organism's internal or external environmentwhich is perceived by the organism as threatening".The modern concepts of stress arose from the pioneering works of Cannon and of Selye inthe early part of the century (Mason, 1975a, 1975b, Selye, 1975). Cannon viewed stress as anyevent which disturbed the state of homeostasis within the organism. He considered stressors toinclude such events as exposure to cold, lack of oxygen, low blood sugar, loss of blood andemotional stimuli. Cannon emphasized the role of the autonomic nervous system (bothsympathetic and parasympathetic) in responding to a stressor so as to restore homeostasis withinthe organism. In contrast, Selye emphasized the role of the pituitary-adrenal system in respondingto noxious stimuli. He proposed, as a part of his theory of the general adaptation syndrome, thatglucocorticoid release from the adrenal gland was a general and non-specific response by theorganism to all forms of noxious stimuli, and equated "biologic stress" with the release ofglucocorticoids (Mason, 1975a, 1975b, Selye, 1975). He suggested that unchecked or chronicstress resulted in exhaustion of the coping mechanism and then disease. Cannon and Selye eachfocused primarily on a specific hormone system as being the critical component of a generalresponse to noxious stimuli. However, subsequent research has demonstrated that during thestress response, the brain is capable of modulating most if not all endocrine functions and thatchanges in the function of many endocrine systems may be important components of the stressresponse.Although both Cannon and Selye primarily focused on the effects of noxious physicalstimuli on the organism, more recently, it has been demonstrated that psychological stimuli, suchas exposure to a novel environment, may be as effective as noxious physical stimuli in eliciting astress response (Mason, 1975b, Weinberg and Levine, 1980). In fact, the response to bothphysical and psychological stimuli appears to depend on the cognitive processing of the stimuli bythe organism. Following exposure to a stimulus, a number of factors, such as past experience,23emotional state or other environmental factors, will influence whether or not the organismperceives the stimulus as threatening. Mason suggests that the common mechanism underlyingthe nonspecific response of the organism to diverse noxious stimuli, as reported by Selye, is infact the organism's perception of the stimulus as threatening (Mason, 1975b). Thus, it has beensuggested that both noxious physical stimuli and psychological stimuli act through a commonmechanism of increasing the organisms level of emotional arousal (i.e. level of fear or anxiety)(Mason, 1975b). The level of activity of the hypothalamic-pituitary-adrenal axis is a goodindicator of an animal's level of emotional arousal and is accepted as a central feature of the stressresponse (Hennessy and Levine, 1979).Moreover, it has been demonstrated that following exposure to an aversive stimulus,cognitive or psychological processes may act to reduce the impact (i.e. level of emotional arousal)of a stressor on the animal (Levine et al, 1978, Weinberg and Levine, 1980). Examples of suchcognitive processes include habituation and coping responses. Habituation has been defined asthe process whereby a novel stimulus that initially increases emotional arousal, but which is notintrinsically aversive, loses it's ability to induce a stress response as the animal learns to associatethe stimulus with a nonthreatening outcome (Levine et al, 1978). In contrast, coping may bedefined as the process that enables an animal to decrease its physiological response to a noxiousor aversive stimulus (Levine et al., 1978). Coping may involve the performance of a behavioralresponse (such as fighting) or it may rely on purely psychological mechanisms (such aspredictability of the stressor), but ultimately, coping occurs via cognitive processes that reducethe impact of the stressor on the emotional response of the animal (Levine et al., 1979).An example of a purely psychological coping process is the effect that predictability of astressor has on the response of the organism to that stressor. It is known that the physiologicaland hormonal responses to predictable or signaled shock are significantly less than those elicitedby unsignaled shock, and that animals will choose signaled over unsignaled shock, even though4the signaled shock may be of greater intensity and of longer duration (Weinberg and Levine,1980). An example of a behavioral coping response is the fighting observed when male ratsreceive electric shocks in pairs (Weinberg et al., 1980). Such animals exhibit a stereotypedfighting response to electric shock, that is, rearing, boxing and biting at each other. This responsesignificantly decreases the physiological consequences (e.g. ACTH level and degree of gastriculceration) of the electric shock compared with that seen in animals shocked individually. Theprotective effect of this coping response does not rely on the physical effects of fighting per se, asthe reduced physiological responses are still observed if the rats are separated by a clear Plexiglasbarrier and thus are only able to rear and posture but not to make physical contact (Weiss et al.,1976). Thus, this coping response appears to rely on the opportunity for the animal to perform astereotyped or organized behavioral response which, through cognitive processes, somehowreduces emotional arousal. It has been suggested that this behavior may be equivalent todisplacement behavior in humans (Weinberg and Wong, 1983). The concept that fightingrepresents a type of coping response is discussed further in section B of the introduction.In summary, many types of aversive stimuli, both physical and psychological, affect theorganism through the common mechanism of altering emotional arousal. Further, the ability ofanimals to cope with a stressor is known to decrease the physiological response of the animal tothe stressor (Levine et al., 1978). As stated previously, coping, whether involving purelypsychological processes or involving a behavioral response, acts to decrease the animal's level ofemotional arousal in response to the stressor (Levine et al., 1978).A.2) Affective Disorders and CancerAffective or emotional state has been strongly implicated as playing an important role inthe etiology of cancer. A number of studies have demonstrated that depression and traitsassociated with depression may be associated with increased cancer incidence/and or poorerprognosis. For example, one study looked prospectively at the relationship between personality5and the incidence of cancer in 2,020 male workers at Western Electric using the MinnesotaMultiphasic Personality Inventory (MMPI) to assess personality (Shekelle et al., 1981). In thisstudy, men working at the Chicago Western Electric plant in 1957-58, who were between theages of 40 and 55, were given a medical exam to insure that they were healthy and were asked tofill out the MMPI forms. Seventeen years later the medical status of the subjects was examined.The overall cancer incidence was 10.5% (212 men) and it was found that men who scored high ona depression scale (380 men) had a significantly greater cancer mortality rate, exhibiting twice theincidence observed in non-depressed subjects. The authors note that the stronger association ofdepression with cancer mortality than with cancer incidence suggests that depression may affecttumor progression rather than tumor initiation. Similarly, Whitlock & Siskind (1979) found astatistically significant increase in the cancer mortality rate during a 4 year follow-up of 126patients (initially cancer free) hospitalized for depression when compared with the normalpopulation.Studies also suggest the existence of a link between depression and length of survival inpatients with advanced cancer. Blumberg eta!. (1954) noted that patients with rapidlyprogressing tumors had a characteristic pattern of scores on the MMPI test, including a highrating on the depression subscale. This was in contrast to patients with slowly progressing orarrested tumors. In another study, 23 patients with various forms of malignant disease wereasked to fill out a self report symptom inventory of psychological status (SCL-90) (Derogatis etal., 1976). It was found that depression was significantly linked with a poor prognosis, but notwith various demographic characteristics including sex, religion, social class and marital status.Finally, a study by Levy eta!. (1988) indicated an association between depression in 34 womenbeing treated for a first recurrence of breast cancer and poor survival (less than 2 years,p<0.08).Further, a patient's level of joy (as measured by SCL-90) was a significant predictor of length ofsurvival in these women (p<0.01), being better than number of metastatic sites or the physiciansrating of prognosis. Thus, this body of experimental evidence supports the early observation ofGalen (Le Shan, 1959) that depression is linked with cancer incidence and prognosis.The role of depression in the development or prognosis of cancer remains controversialfor several reasons. First, several large prospective studies have found no association betweendepression and cancer incidence or mortality. A prospective study of 6,848 adult residents ofAlameda County assessed the link between depressive tendencies and cancer using the HumanPopulation Laboratory questionnaire in 1965 (Kaplan and Reynolds, 1988). After a 17 yearfollow-up, no significant association between depression and either cancer incidence or cancermortality was found. In another study, 8,932 females undergoing routine breast exams between1968 and 1972 were asked to complete the MMPI form. This study failed to demonstrate a linkbetween depression at the start of the study and the 117 cases of breast cancer observed during a10 year follow up (Hahn and Petitti, 1988). Second, many studies in this area have been criticizedfor methodological flaws (Bieliauskas and Garron, 1982; Levenson and Bemis, 1991). Forexample the studies of Blumberg et al. (1954) and Derogatis eta!. (1976) have been criticized forfailing to control for stage of disease. In addition studies have been criticized for small, biasedsamples, inclusion of patients with different types of cancer, failure to control for stage of diseaseand treatment of cancer, retrospective subject bias and failure to standardize measures ofpsychosocial factors (Levenson & Bemis, 1991). Further, it has been suggested thatdifferentiating among various depressive disorders and examining the chronicity of the depressionis important but that many studies failed to consider these issues (Levenson & Bemis, 1991).Another affective or emotional state suggested to be associated with cancer incidence isthe feeling of hopelessness and/or helplessness. Interestingly, reported feelings of hopelessnessor helplessness have also been considered as an indicator of depression (Bieliauskas and Garron,1982). Here again, studies of the association of this trait with cancer have produced conflicting6results, demonstrating positive associations (Schmale and Iker, 1966; Wirsching, 1982) or noassociation (Casselith, 1988).A.3) Personality / Coping Style and CancerA second psychosocial factor that has been suggested to be important in the incidence orthe progression of cancer is personality and/or coping style. It has been convincinglydemonstrated that personality is a factor in influencing the outcome of certain disease states. Thebest example of this is the association of heart disease with the type A personality (Rosenman eta/., 1964). Similarly, based on her observations of melanoma patients, Temoshok has proposedthat a new personality type, the type C personality, may be associated with an increased risk ofcancer (Temoshok and Fox, 1984). The type C personality is described as being intermediatebetween the type A and B personalities: an individual who strives to maintain a sense of pleasantinterpersonal atmosphere with control over angry expressions, and a sense of themselves as wellliked. According to this view, the type C personality differs from the type B personality in thatthe former represses negative emotions while the latter does not. Temoshok investigated 106patients with malignant melanoma and found that significantly more of the patients who died orrelapsed during an 18 month follow-up were of the type C personality (Temoshock and Fox,1984). An important component of the type C personality is the repression of negative emotions.This characteristic has also been linked to cancer incidence in a number of studies (Grossarth-Maticek, 1985, Kissen and Eysenck, 1962, van der Ploeg, 1989, Wirsching eta!, 1982). Forexample, a 10 year (1966 to 1976) prospective study of 1353 residents of a Yugoslavian villagedemonstrated that suppression of emotions was significantly correlated with increased risk ofcancer (Grossarth-Maticek eta!., 1985). Also, using suppression of emotions or poor expressionof emotions as the criteria, Wirsching et a/. (1982) correctly predicted the diagnosis of 75% of58 patients undergoing breast biopsy. Similarly, Greer and Morris (1975) found that patientsexhibiting a "fighting spirit" had significantly increased 5 year survival rates.7It should be recognized that affect is an important component of personality and thatsuppression of emotion, a key factor in Temoshok's cancer prone personality, is considered bysome to be a reflection of depression. Thus it is possible that all of the studies mentioned aremeasuring various facets of the same personality variable or personality type.A.4) Childhood Stress/Loss and CancerCancer is a disease which is believed in general to have a long development period. Forexample, colon cancer is thought to involve mutations in up to 4 discrete genes (oncogenes andsuppressor genes, Vogelstein, 1989). Thus, some researchers feel that if psychosocial stressorsreally do play a role in the initiation or progression of cancer, it is important to look at early timesin the patient's history for evidence of psychosocial perturbations. This approach is exemplifiedby a prospective study conducted by Thomas eta!. (1979). In this study, psychological andmedical data were collected from 1,337 Caucasian male medical students attending John HopkinsMedical School in 1946. The researcher used this information to investigate prospectively theassociation of quality of early family life with subsequent cancer incidence. Of 913 students whofilled out a family attitudes questionnaire (FAQ), 48 subsequently developed cancer during the 30year follow-up. It was found that the cancer patients had previously reported significantly lesscloseness with their parents (p<0.01) than had the control subjects. This finding has recently beenreplicated in a retrospective study in which it was found that women with breast andgynecological cancers reported significantly less closeness with their parents using Thomas's FAQscore, than did matched controls (Gehde and Baltrusch, 1990).89A.5) Life Stress/Psychosocial Stressors and Cancer.There have long been anecdotal reports of stressful life events (or psychosocial stressors)preceding the onset or recurrence of cancer. Recently, a number of studies have demonstratedlinks between life stressors and cancer. For example, retrospective studies have demonstratedthat severely stressful life events, such as the death of a close friend or a family member, are morecommon in a 1 to 6 year period preceding the diagnosis of breast cancer (Cooper et al., 1989,Forsen, 1991, Geyer, 1992) or other types of cancer (Taylor et al., 1988) than in controlpopulations. Also, severely stressful life events were reported more frequently in womenexperiencing relapse of breast cancer than in women who did not experience a relapse (Forsen,1991, Ramirez et al., 1989). Retrospective studies of the link between life stressors and cancerhave been criticized because cancer patients, knowing their diagnosis, may exhibit bias to negativeinterpretations of life events (Levenson and Bemis, 1991). However, it has recently beendemonstrated that recall bias is not a significant factor in these studies (Geyer, 1992).In contrast, a number of studies have failed to demonstrate an association betweennegative life events and cancer incidence. For example, cancer incidence in prisoners of warfollowing World War II and the Korean war did not differ significantly from that in the normalpopulation (Keehn, R.J. eta!., 1974,1980). Ewertz (1986) found no association between theassumed bereavement caused by widowhood or divorce and the diagnosis of breast cancer in1782 women diagnosed with breast cancer and 1738 randomly selected control subjects.However, there is no indication of how well matched the two populations were with respect toage, socioeconomic status or ethnic origin. Further, studies of loss such as this have beencriticized for failing to take into account the individual's perspective of a supposedly stressfulevent. For example, to some, divorce may be viewed as a release and as such a joyous occasion(Eysenck, 1988).10A.6) Social Contacts and CancerThere is evidence that social relationships have a protective effect on general health andwell being (House et al., 1988). This effect is especially observed after traumatic events such asautomobile accidents and heart attacks (Frasure and Prince, 1985, Porritt, 1979). There is alsoevidence that social relationships may have beneficial effects on cancer patients. One largeprospective study found that low levels of social contact and social isolation increased the risk ofcancer mortality in both men and women (Kaplan and Reynolds, 1990). The study also found thatdecreased social contact in women increased the risk of developing cancer (Kaplan and Reynolds,1990). Further, several prospective studies have suggested that a correlation exists betweenperceived levels of social support or social contact and survival in women newly diagnosed withbreast cancer (Ell et al., 1992, Waxier-Morrison et al., 1991). Perhaps the most convincingevidence for the effects of social support on cancer survival come from two studies whichindependently examined the effects of psychological counseling on survival in patients withadvanced breast cancer. These studies demonstrated that psychotherapy can significantly prolongthe survival of women with advanced breast cancer (Grossarth-Maticek et al, 1989, Spiegel etal., 1989). These findings are particularly impressive because Spiegel set out to disprove thetheory that psychological counseling could affect the course of breast cancer. Further, one ofthese studies demonstrated that psychotherapy was as effective as chemotherapy in prolonging thelife of women with advanced stage metastatic breast cancer and can exhibit synergistic effectswith chemotherapy in prolonging life (Grossarth-Maticek et al., 1989).Thus, although there are many scientific studies that strongly suggest anassociation between cancer and psychological or emotional variables, the topic remainscontroversial. This controversy arises from two sources. First, it is probable that our attempts tostudy the effects of discrete psychological variables such as personality type or stressful life eventsare too simplistic. It is likely that affective state, personality, coping styles and social interactions1 1combine to affect how people view life events and ultimately how any of these factors will impacton cancer. Thus it is probably somewhat simplistic to look at one characteristic in isolationwithout examining all other variables. Further it is possible that subtle variations in severalvariables could have an impact on cancer initiation or progression and yet be beyond our currentability to measure or accurately control. Second, the lack of precise knowledge of the events andtiming involved in the genesis and progression of cancer likely hamper studies of the effects ofstressors on cancer.B) Animal Models Of Stressor Effects on Cancer.As noted above, research on the role of stressors in modulating tumor growth is complexand some studies have yielded ambiguous or even negative results. It is likely that the confusionin this field of research arises, at least in part, from an inability to control the frequency or severityof stressors experienced by the subjects and difficulty defining the temporal association of thestressors with the induction of the tumor. Thus, a number of researchers have turned to animalmodels to study the possible link between stressors and cancer. Animal models provide the abilityto control all study variables including the type of stressor, the severity of the stressor and thetiming of application of the stressor relative to tumor injection or induction. Studies using animalmodels have demonstrated that stressors can affect tumor growth. However, despite theincreased control over study variables, the results of animal studies are also often inconsistent andcontradictory. Data have demonstrated that stressors may cause an increase, a decrease, or haveno effect on tumor growth. A number of factors have been suggested to be important indetermining the effect that a stressor will have on tumor growth. These factors include the typeof tumor used, the type of stressor used, the timing of application of the stressor (i.e. before orafter tumor induction), whether the stressor is chronic or acute and whether or not the animal hasa coping response available to help it deal with the stressor (Justice, 1985, Sklar and Anisman,1981, Solomon and Amkraut, 1981).1 2Type of Tumor and Timing of Application of the Stressor. It is well known that stressorsaffect different types of tumors differently. For example, stressors are known to decrease thegrowth of most tumors induced by the chemical carcinogen 7,12 dimethylbenz[a]anthracene(DMBA) (Newberry, 1978, Ray and Pradhan, 1974), whereas stressors generally increase thegrowth of virally-induced tumors (Amkraut and Solomon, 1972; Riley, 1981). Justice (1985)suggests that stressor effects on viral tumors (immunoresponsive due to the expression of viralantigens) are due to the ability of stressors to affect immune functioning. Stressors applied duringthe growth of virally induced tumors suppress the immune system, allowing increased growth ofthe tumor. However, if the stressor is applied before or in the early phases of tumor induction,rebound enhancement of immune functioning occurs following cessation of the stressor andresults in decreased growth of the virally induced tumor (Amkraut and Solomon, 1972; Sklar andAnisman, 1981). Justice (1985) further predicts that, with non virally-induced tumors, theimmune system does not play a role in mediating the effects of stressors on tumor growth.Instead, stressors applied after tumor induction or injection appear to inhibit the growth of thesetumors by other, possibly hormonal, mechanisms. In contrast, stressors applied prior to tumorinduction or growth stimulate tumor growth by a rebound enhancement mechanism which wouldoccur following the cessation of the stressor. Justice claims that these two factors, the type oftumor and the timing of application of the stressor relative to the tumor induction or injection, aresufficient to explain the variable results observed in this field.Although Justice's theory explains many apparent discrepancies in this field of study, hismodel is not completely satisfactory for several reasons. First, although Justice's theory explainsthe growth of virally-induced tumors, it does not completely explain tumor growth in non virally-induced tumors such as carcinogen-induced tumors. Many studies support Justice's claim thatstressors applied during tumor induction (using the chemical carcinogen DMBA) suppress ordelay the growth of tumors. However, a number of studies have been published which have the1 3opposite finding: that chronic stressor administration during tumor induction by DMBA actuallydecreases tumor latency and increases tumor growth (Steplewski, 1985, Steplewski et al., 1987,Tejwani eta!., 1991). Second, Justice does not consider studies which use the psychosocialstressor of differential housing. Interestingly, these studies show that individual housing as well asa change in housing condition typically increase the growth of various non viral tumors(Dechambre and Gosse, 1973, Sklar and Anisman, 1980, Steplewski eta!., 1987). Finally,studies have shown that stressors increase both the number and the size of metastases. Thus thetype of tumor and the timing of stressor application are likely not the only factors responsible foraltering the effect that stressors have on the growth of tumors.Environmental Variables. Environmental variables are known to have an importantimpact on tumor growth. For example, mice switched from group to individual housing conditionshow increased tumor growth rates compared with that of mice remaining in their group (Sklarand Anisman, 1980, Weinberg and Emerman, 1989). Further, it has been demonstrated thathousing condition can affect the impact of physiological stressors on tumor growth. Sklar andAnisman (1980) demonstrated that group housed mice exhibited increased growth of atransplanted syngeneic tumor following exposure to an acute electric shock, whereas mice movedfrom group to individual housing did not have an increase in tumor growth following shockexposure.In addition to housing condition, the general environment in which the animals live (i.e.the intensity and timing of lighting in the animal room, the levels of noise and disturbance in theanimal room, etc.) can affect their tumor growth. Riley et al. (1981) showed that in virgin femaleC3H mice, mammary tumor incidence was much greater in mice housed in "conventional" animalhousing facilities than if mice were housed in a "low stress" environment. In "conventional"animal housing facilities, mice were housed in stainless steel cages on open racks in a communalanimal room, exposing the mice to the daily activities of cage cleaning, bleeding procedures and1 4other stress-inducing experimental manipulations of other animals in the room. In the "low stressenvironment", mice were housed in plastic cages with bedding in specially designed racks whichminimized the spread of pheromones between cages. Further, the "low stress" environmentcontrolled the temperature variation, the light/dark cycle and the degree of disturbance induced byexperimental and maintenance procedures (Riley et al., 1981). Data demonstrated that basalplasma corticosterone levels, one indicator of arousal or stress, were 10 to 20 times lower in micehoused in the "low stress" environment compared with those in "conventional" housing (Riley etal., 1981). Similarly, studies have demonstrated that, in animals housed in "low stress"conditions, some tumors which had previously been considered to be stress-nonresponsive couldbe shown to be modulated by stressors. Thus, some of the variability in the literature could be aresult of failure to control adequately for such variables as housing condition, noise, disturbancesand other stressors in communal animal housing facilities and pheromones released by stressedanimals.Acute Vs Chronic Stressors. It has also been demonstrated that the chronic vs acutenature of the stressor can alter the effect of that stressor on tumor growth. For example,exposure to a single (acute) session of electric shock at the time of injection of a syngeneicmastocytoma in mice reliably increased tumor growth. If, however, the animals were given 4 or 9daily shock sessions before the tumor injection and acute shock session, tumor growth was notenhanced (Sklar and Anisman, 1981). In other experimental models, chronic stress has beenshown to decrease tumor growth (Gershben et al, 1974, Marsh et al., 1959, Molomut et al,1963, Ray and Pradhan, 1974).CopingAbility. The ability of animals to cope with a stressor is known to decrease thephysiological response of the animal to the stressor (Levine et al., 1978). As stated previously,coping, whether involving purely psychological processes or involving a behavioral response, actsto decrease the animal's level of emotional arousal in response to the stressor (Levine et al.,1 51978). For example, as described previously, it has been demonstrated that rats subjected toelectric shock in pairs exhibit a stereotyped fighting behavior during the electric shock whichdecreases the number and severity of gastric lesions, the elevation in blood pressure and theplasma levels of ACTH compared with that in rats receiving the same amount of shock asindividuals (Conner et al., 1971, Weiss et al., 1976, Williams and Eichelman, 1971). Copingprocesses may also effect stressor-induced changes in tumor growth. For instance, it has beendemonstrated that the growth of a syngeneic mastocytoma is slower in mice exposed to escapableshock than in mice receiving an identical amount of inescapable shock (yoked paradigm) (Sklarand Anisman, 1981). Similarly, male mice that have experienced a change in housing conditionfrom group to individual housing exhibit increased tumor growth, whereas mice transferred fromindividual to group housing do not show an increase in tumor growth (Dechambre and Gosse,1973, Sklar and Anisman, 1980, Weinberg and Emerman, 1989). It has been suggested that theprotective effect of being moved from individual to group housing may be due to the fighting thatoccurs in the group (Weinberg and Emerman, 1989). That is, fighting may serve as a kind ofbehavioural coping response that helps the mice deal with the stressor of changing housingcondition. It is not intuitively obvious why fighting, which would appear to be a stressor in itself,may act as a coping response. However, as mentioned previously, it has been suggested thatfighting acts as a coping response in rats receiving electric shock. Fighting is suggested to belongto a class of behaviors referred to as consummatory behaviors (Levine et al., 1979).Consummatory behaviors are stereotyped sequences of goal oriented activities such as eating,drinking and mating (Hennessy and Levine, 1979, Weinberg and Wong, 1983). Consummatoryresponses are known to decrease both the emotional arousal and the activation of thehypothalamic-pituitary-adrenal axis induced by stressors (Levine and Coover, 1976, Levine et al.,1979). Thus, fighting, like other consummatory behaviors, may act to decrease an organisms'level of emotional arousal.16C) Psychoneuroendocrinology and CancerC.1) Hormones and StressHypothalamic-Pituitary Axis: It has long been known that stressors may influenceendocrine function. As previously mentioned, pioneering work by Cannon and Selye in the earlypart of the twentieth century demonstrated that stressors could alter the secretions of adrenaline,noradrenaline and glucocorticoids. More recently, it has been demonstrated that the centralnervous system is involved in the control of almost all endocrine functions, either directly orindirectly. One method whereby the central nervous system controls endocrine function is via thehypothalamic-pituitary axis. The pituitary, a small neuroendocrine gland located at the base of thebrain, is divided into three lobes- anterior, middle and posterior.The posterior lobe or neurohypophysis is derived from neural ectoderm. Cells located inthe hypothalamus synthesize and secrete the peptide hormones oxytocin and antidiuretic hormone.Axons of these cells project to the posterior pituitary. Oxytocin and antidiuretic hormone arestored in axon terminals and secreted from the posterior pituitary following the appropriatestimulation. The middle lobe is functionally unimportant in man. The anterior lobe or parsdistailis is derived from oral ectoderm and consists of a number of distinct cell types which areresponsible for the production of various peptide hormones. Some of these peptides have directeffects on target tissues (i.e. prolactin, growth hormone, somatostatin and 0-endorphin), whereasothers function as trophic hormones to stimulate the release of hormones by other endocrineglands (i.e. leutinizing hormone/follicle stimulating hormone, adenocorticotrophic hormone[ACTH], thyroid stimulating hormone). The activity of the anterior pituitary is modulated by thehypothalamus which produces releasing hormones that induce secretion of pituitary peptides. Thereleasing hormones are delivered to the pituitary by the hypothalamo-hypophyseal portal system,a specialized system of vessels. As an example of how this system functions, one can examine the1 7control of glucocorticoid secretion. Glucocorticoids are secreted by steroid producing cells of theadrenal cortex which are stimulated to synthesize and release glucocorticoids by the peptideACTH (Axelrod and Reisine, 1984). ACTH is produced by corticotroph cells of the anteriorpituitary and travels in the bloodstream to the adrenal glands. The primary mechanism ofstimulating ACTH synthesis and secretion is via the peptide corticotrophin releasing factor (CRF)(Axelrod and Reisine, 1984, Jacobson and Sapolsky, 1991), which is produced by cells of thehypothalamus and released into the hypothalamic-hypophyseal portal system (Axelrod andReisine, 1984, Reisine et al., 1986). Studies have revealed that vasopressin and cholecystokininmay also be produced by CRF-containing hypothalamic neurons under certain conditions and thatthese substances also stimulate ACTH production by pituitary corticotrophs, possibly actingsynergistically with CRF (Jacobson and Sapolsky, 1991, Reisine et al., 1986). Finally, thecatecholamines adrenaline and noradrenaline are known to stimulate pituitary corticotrophs toproduce ACTH via (32 adrenoreceptors (Jacobson and Sapolsky, 1991, Reisine eta!., 1986). Thestimulation by catecholamines is not as great as that induced by CRF (Axelrod and Reisine, 1984).Plasma glucocorticoids in unstressed organisms follow a circadian rhythm, being lowest duringsleep and highest just after waking. In response to stressors, CRF release from the hypothalamusis increased, resulting in increased ACTH and glucocorticoid production and secretion. Therelease of glucocorticoids is tightly controlled at several levels of the hypothalamic-pituitary-adrenal axis. First, high levels of plasma glucocorticoids act to inhibit the production ofglucocorticoids by cells of the adrenal cortex. Elevated plasma glucocorticoid levels also act todecrease the sensitivity of pituitary corticotrophs to CRF. Further, glucocorticoids inhibit therelease of CRF from the hypothalamus. The inhibition of CRF release caused by elevated plasmaglucocorticoid levels has been demonstrated to be due, in part, to direct actions on thehypothalamus and in part to action on other regions of the brain such as the hippocampus.Activation of the hypothalamic-pituitary-adrenal axis is a central feature of the stressresponse. In addition, many stress-induced changes in hormone secretion are modulated, at least1 8in part, by the action of the hypothalamic-pituitary-adrenal axis. For example, hormonalsecretion by the gonads is affected at several levels by glucocorticoids. The hypothalamic releaseof gonadotropin-releasing hormone (GnRH) is decreased by increased CRF levels and bychronically elevated plasma glucocorticoid levels. Evidence suggests that this may be due to adirect effect on a corticosteroid regulatory element in the GnRH gene (Brann and Mahesh, 1991).However, corticosterone- and CRF-stimulated increases in opioid and catecholamine levels in thehypothalamus also appear important in blocking GnRH release and both naloxone and a-adrenergic antagonists block the effects of chronically elevated glucocorticoids on LH secretion(Brann and Mahesh, 1991, Rivier and Rivest, 1991). Chronically elevated corticosteronedecreases the response of anterior pituitary gonadotroph cells to GnRH. Further, the gonadscontain functional glucocorticoid receptors and elevated glucocorticoid or ACTH levels havebeen demonstrated to inhibit the response of Leydig cells to LH (Brann and Mahesh, 1991, Rivierand Rivest, 1991). Thus, the stress-induced alterations in the secretion of sex hormones involvemultiple mechanisms acting at multiple levels. Similarly, stress-induced elevations of CRF,ACTH and glucocorticoids have been shown to be involved in the altered secretion of manyhormones other than those of the hypothalamic-pituitary-gonadal axis, including insulin andglucagon.Autonomic Nervous System: A second mechanism by which the central nervous systemmay control hormonal secretion is through modulation of autonomic nervous function. Forexample, stimulation of sympathetic nerves has been demonstrated to decrease the release ofinsulin and stimulate the release of glucagon from islet of Langerhans cells in the pancreas. It hasfurther been demonstrated that this effect is reproducible by the direct administration ofadrenergic agonists in the isolated pancreas (Yamaguchi, 1992). During stress responses,increased activity of sympathetic nerves to target organs such as the anterior pituitary, the adrenalmedulla and the endocrine pancreas increases the local concentrations of noradrenaline (Axelrodand Reisine, 1984, Yamaguchi, 1992). Further, both increased activity of sympathetic nerves1 9and increased concentrations of plasma glucocorticoids act to increase the synthesis and secretionof adrenaline from chromaffin cells of the adrenal medulla (Axelrod and Reisine, 1984,Yamaguchi, 1992). Thus, stressor-induced alterations in the secretion of hormones such asglucocorticoids and catecholamines can interact both with each other and with other hormone-producing cells to cause wide spread changes in endocrine function following stressors.C.2) Hormones and CancerHormones play an important role in the initiation and/or progression of some tumors. Forexample, it has been reported that 40 - 60% of human cancers are etiologically associated withsex hormone exposure, either endogenously or exogenously (Li eta!., 1991). Cancers are oftendivided into two broad classes, hormone-responsive and hormone-nonresponsive, based on theirhormone responsiveness. Hormone-responsive tumors typically arise from the endometrium,breast, prostate, immune system (lymphomas and leukemia) and endocrine system (Lippman etal., 1985). In these tissues, hormones play a crucial role in modulating cell proliferation.Although the precise role of hormones in the carcinogenesis of these hormone-responsive tissuesis not known, it is clear that hormones are required for the transformations to occur (Welsh,1985). Further, most of the tumors that develop are initially dependent on the presence of thehormone for proliferation to occur (Dickson et al., 1992, Miller et al., 1990). Thus, modulationof hormone levels could drastically affect the growth of hormone-responsive tumors.Steroid Hormones and Growth Factors: It is known that both during development and inmany adult tissues, hormones and growth factors are involved in regulating the proliferation anddifferentiation of cells (Miller and O'Neill, 1990). This is exemplified by the role of steroids,especially estrogens and androgens, in regulating the growth, differentiation and function of avariety of normal tissues including the gonads, pituitary and secondary sex organs (Miller andO'Neill, 1990).20Steroids readily enter the cell, diffusing through the plasma membrane due to theirlipophilic properties. In the cell, steroids bind to intracellular receptors specific for the individualclass of steroid. Steroid receptors have been demonstrated to be a family of related proteins withmolecular weights ranging from approximately 50 to 100 kDa in humans. (Jensen, 1992). Thereceptors contain a steroid binding domain, a hinge region and a DNA binding domain ofapproximately 66 - 68 amino acids (Jensen, 1992, Liao, 1992). The receptors bind to the DNA bya pair of zinc finger domains each of which consists of a peptide loop stabilized by zinc ions(Jensen, 1992). Binding of a steroid by the receptor is thought to cause a conformational changein the receptor, releasing a 90 kDa heat shock protein dimer from the DNA binding region of thereceptor, thus activating the hormone-receptor complex and allowing it to bind DNA (Jensen,1992, Liao, 1992). The steroid receptors bind to certain DNA sequences termed hormone-responsive elements (HREs) and there is apparently overlap in the ability of different steroidreceptors to bind to different HREs (Jensen, 1992, Liao, 1992). The HREs are located upstreamof various cellular genes and affect the transcription of the gene, causing either promotion orinhibition of transcription depending on the cell type and the genes involved (Liao, 1992).Administration of estrogen (17 0-estradiol) to cells of human breast cancer cell lines in cultureresults in increased transcription of a number of cellular genes and ultimately increasesproliferation of the cells (Dickson et al., 1990). However, the genes which are critical toestrogen's ability to stimulate cell proliferation have not yet been identified (Dickson et al., 1990).Both estrogens and androgens have been shown to stimulate the transcription of a number ofproto-oncogenes including peptide growth factors, growth factor receptors and transcriptionfactors. Estrogen treatment of human breast cancer cells in culture induces increased expressionof the growth factors transforming growth factor-a (TGF-a), insulin like growth factor I and II(IGF-I and -II), as well as the epidermal growth factor receptor (EGFR) and the nucleartranscription factors c-fos, c-jun and c-myc (Dickson eta!., 1992). In the prostate of castratedrats, testosterone administration rapidly up regulates the expressions of the proto-oncogenes c-21fos, c-myc and c-Id-ras (Thompson, 1992). The relative importance of hormone-inducedstimulation of growth factors and their receptors versus nuclear proto-oncogenes incarcinogenesis is controversial but it is likely that both play a role (Lippman and Dickson, 1987,Vander Burg et aL, 1992). Further, it has been suggested that the effects of estrogens oncarcinogenesis may be due to the ability of the hormone both to promote tumor development bystimulating cell growth and to initiate transformation in dividing cells by interfering withmicrotubule assembly, thereby inducing aneuploidy (Barrett, 1992, Metzler et al., 1992).Hormonal Control of Normal Breast Cells: In most hormone-responsive normal tissues,complex interactions of several hormones and growth factors are involved in mediating thebalance of proliferation and differentiation of cells. Further it has been suggested that imbalancesin the hormones could contribute to carcinogenesis or promotion of tumor growth (Ho et al.,1992, Kuttenn eta!., 1992). As breast cancer growth in an animal model is a focus of this thesis,the hormonal control of breast cells (both normal and malignant) will be considered in greaterdetail. The breast is a good example of a hormone-responsive tissue.Both the growth and the functional activity of the breast epithelium are tightly controlledby hormones. Unlike most other organs in the body, the breast is not fully formed at birth. Thebreast of female rodents consists of an epithelial rudiment and an underlying mammary fat pad(Imagawa et al., 1990, Sakukura, 1991). At this stage, the epithelial rudiment consists of shortducts that extend into the underlying fat pad. The gland remains in this form until puberty, atwhich time there is a significant outgrowth of the ducts to form branched ductal trees that occupymost of the mammary fat pad. The mammary gland then undergoes cyclic expansion andinvolution of the ductal epithelium under the influence of the female sex steroids. The increasedgrowth of the mammary gland involves further branching of the ductal structure and thedevelopment of lobuloalveolar sacs at the termination of the branches. During pregnancy, thedevelopment of lobuloalveolar sacs is greatly stimulated and secretory alveoli fill the gland,displacing the fat cells. Following the cessation of lactation, the mammary gland undergoesinvolution and returns to the state seen in the cycling gland.Studies have revealed that the hormonal control of the breast involves a network ofinteracting hormones and growth factors. In vivo experiments in rodents have revealed that anumber of hormones are involved in the different stages of the development of the breast. It hasbeen shown that either gonadectomy or hypophysectomy of female mice prevents thedevelopment of the breast epithelium at all stages of development (Haslam, 1987, Imagawa et al.,1990). In gonadectomized, hypophysectomized and adrenalectomized prepubertal female mice,the addition of exogenous estrogen, growth hormone and either progesterone or corticosterone isrequired for the induction of normal development of the ductal tree (Haslam, 1987, Imagawa etal., 1990), whereas in post pubertal virgin female mice, the addition of estrogen, progesterone andprolactin is required to induce the formation of the lobuloalveolar structures (Haslam, 1987,Imagawa eta!., 1990). It has been shown that the elevated serum levels of progesterone in thefemale during gestation block the premature expression of differentiated functions by themammary epithelial cells (Haslam, 1987).In an attempt to define more accurately the relative contributions of different hormonesand growth factors to the development of the mammary epithelium, a number of researchers haveinvestigated the growth of breast epithelium in vitro. In vitro studies have primarily utilized 1 of2 different methods; early studies used whole gland or explant cultures whereas more recentstudies have examined the growth of relatively pure populations of breast epithelial cells in cellculture (Imagawa et al., 1990). These studies have revealed several important findings regardingthe stimulation of breast epithelium. Organ cultures generally have demonstrated that theproliferation of breast epithelium at different stages of mammary gland development have similarhormonal requirements to those observed in in vivo studies (Imagawa eta!., 1990). A notableexception is the finding that unlike the in vivo situation, explants of breast tissue do not proliferate2223in response to estrogen in defined medium (Haslam, 1987, Topper and Freeman, 1980). It hasbeen speculated that estrogen may exert it's mammogenic effect in an indirect manner (Haslam,1987), that is, by affecting the ability of cells to respond to other hormones and by stimulating theproduction of other hormones (see below).Cell culture experiments have revealed that the stroma of the mammary gland is importantin modulating the ability of the breast epithelium to respond to hormones and growth factors. Forexample, the growth of breast epithelial cells from rodents is dramatically affected by thesubstratum on which they are grown (Emerman and Pitelka, 1977). In contrast to cells grown onplastic substrata, cells grown on collagen gels exhibit normal morphology and produce breast milkcomponents, considered to be an expression of the differentiated mammary phenotype (Emermanand Pitelka, 1977, Emerman eta!., 1977). In addition growing breast cells within a collagen gelmatrix produced significantly greater proliferative responses to lactogenic hormones than growingcells on plastic (Imagawa et al., 1990). Cell culture experiments using defined media haveallowed the investigation of the minimal hormonal requirements for the proliferation of breastepithelial cells. The minimal medium which is capable of maintaining breast epithelial cells inculture contains insulin and either bovine serum albumin or phospholipids (Imagawa eta!., 1990).Using this medium, it has been demonstrated that progesterone and/or prolactin are capable ofstimulating the growth of cultured mammary cells (Imagawa et al., 1990). Interestingly, althoughestrogen does not increase cell proliferation in this system, it was demonstrated that the cells dopossess estrogen receptors and that estrogen stimulates upregulation of the progesteronereceptors in these cells (Haslam, 1987). It has been suggested that the in vivo effects of estrogenresult from it's effect on progesterone receptors and it's ability to stimulate prolactin secretion(Imagawa et al., 1985). As well, a number of growth factors have been demonstrated to affectthe growth of mammary epithelial cells in vitro (Dembinski and Shiu, 1987). Epithelial growthfactor (EGF) and transforming growth factor-a (TGF-a) have been shown to stimulate theproliferation of breast cells and to inhibit the differentiated functions of these cells (Dembinski and2 4Shiu, 1987, Imagawa et al., 1990). Further, the combination of EGF and bFGF can substitute forprolactin in the induction of alveolar growth in vitro (Imagawa eta!., 1990). Transforminggrowth factor-I3 (TGF-0) has been demonstrated to have an inhibitory effect on the growth ofmammary cells (Dembinski and Shiu, 1987, Imagawa eta!., 1990).Implanting small pellets containing hormones or growth factors into the mammary glandsof mice has demonstrated the functional activity of these compounds in the organism. It wasdemonstrated that in prepubertal mice (3-4 wk old), estrogen has a local or direct stimulatoryeffect on the development of the ductal epithelium but does not exert a systemic effect onmammary growth as the nonimplanted contralateral mammary glands are not affected (Haslam,1987). Interestingly, in postpubertal mice (10 wk old), the estrogen pellet exerts a systemic effect(Haslam, 1987). Implants of pellets containing EGF result in an initial increase in the growth ofthe breast epithelium in virgin mice, followed by a subsequent decline in growth by 3 d postimplantation (Imagawa eta!., 1990). It has been suggested that this bimodal effect of EGF maybe due to the down-regulation of the EGF receptor in mice chronically exposed to EGF.Importantly estrogen, in addition to affecting serum prolactin levels and the expression ofmammary epithelial cell progesterone receptors as discussed previously, acts to increase theproduction of EGF, up regulate the expression of the EGF receptor and decrease the secretion ofTGF-13 (discussed below). In contrast, it was demonstrated that implants of TGF-0 appear toinhibit the growth of the ductal cells (Daniel and Robinson, 1992). TGF-13 has been found to belocalized in the stroma surrounding nondividing ductal epithelial cells, but not in the stromasurrounding the cells in the growing ends of the ducts (Daniel and Robinson, 1992).In vitro studies of normal and malignant human breast epithelial cells, as well as cell linesderived from human breast cancers, have also greatly added to our understanding of the hormonalcontrol of breast epithelial cells. Data indicate that estrogen in the presence of serum stimulatesthe growth of normal human mammary epithelial cells (Kutten eta!., 1986), human breast2 5carcinomas in primary culture (Emerman et al., 1990), and human breast cancer cell lines(Dickson et al, 1986). In the absence of serum, however, estrogen generally does not stimulatethe growth of breast cells (Gableman and Emerman, 1992). Similar results have been observed instudies of rodent mammary epithelial cells in culture (Haslam, 1987). The mechanism wherebyserum allows estrogen to stimulate breast epithelial cell growth is not currently known. However,studies have demonstrated that estrogen causes an up regulation of EGF receptors on normalhuman mammary epithelial cells (Colomb eta!., 1991) and in human breast-cancer derived celllines (Berthois et al., 1989), as well as increasing the synthesis of EGF- or TGF-a-like molecules(Dickson, 1986). Further, it has been demonstrated that estrogen down-regulates the productionof TGF-13 mRNA in human breast cancer cell lines (Jeng and Jordan 1991, Nutt eta!., 1991). Ifestrogen is able to up regulate EGF receptors and decrease the production of the inhibitorygrowth factor TGF-13, then the permissive effect of serum may be to supply an exogenous sourceof EGF for cell growth. This possibility is supported by the finding that in serum-free medium,estrogen has a synergistic effect on the ability of EGF to stimulate the growth of primary culturesof human breast epithelial cells (Gableman and Emerman, 1992) This summary clearlydemonstrates that the control of growth and differentiated functions in mammary epithelial cellsinvolves an interacting network of hormones and growth factors.Hormonal Control of Breast Cancer: Breast cancers also appear to be regulated bycomplex interactions among a number of hormones and growth factors. Although thetransformation of breast cells is dependent upon the presence of hormones and growth factors(Boot et al, 1981), breast tumors exhibit considerable variability in the hormones and growthfactors that are required for proliferation (Ethier and Cundiff, 1987, Ethier and Moorethy, 1991,Platica et al., 1991). Furthermore, it has been demonstrated that many experimental breasttumors progressively lose their hormonal responsiveness (Sluyser, 1987). It has also been shownthat preneoplastic mammary cells produced by in vitro transformation with chemical carcinogenshave similar growth factor requirements for in vitro growth as normal mammary epithelial cells.2 6The most basic requirements are insulin and EGF (Kittrell et al., 1992) or insulin and prolactin(Ganguly eta!., 1982). It has been demonstrated that approximately half of rat primary mammarytumors, generated in vivo by carcinogen treatment, exhibit similar in vitro growth factorrequirements as normal mammary epithelial cells (Ethier and Cundiff, 1987). Interestingly, whenthese cultured tumor cells are injected into mammary fat pads, they are no longer tumorigenic andproduce only normal appearing ductal structures. In contrast, tumor cells which were found toexhibit growth factor independence in vitro produced tumors when injected into mammary fatpads (Ethier and Cundiff, 1987, Ethier and Moorethy, 1991). The transition from hormone-dependence to hormone-independence has been studied in a pregnancy-dependent mammarytumor (Imagawa et al., 1992, Matsuzawa, 1986). This tumor forms hyperplastic alveolar nodule-type preneoplastic lesions in virgin female mice. The development of the tumor is stimulated bythe elevated levels of estrogen, progesterone and prolactin that occur during pregnancy, but thetumor subsequently regresses during lactation. Chronic stimulation with estrogen andprogesterone or repeated pregnancies can induce the outgrowth of ovarian hormone-responsivetumors which do not regress during lactation and finally, ovary-independent tumors (Imagawa etal., 1992, Matsuzawa, 1986). The transition from a pregnancy-dependent state to an ovary-independent state is observed to be accompanied by increased anaplasia and increased growth rate(Imagawa et al., 1992). Further, data indicate that while pregnancy-dependent tumors are similarto normal mammary epithelial cells in their requirement for hormones (estrogen, progesterone andprolactin) or growth factors (EGF and bFGF) to proliferate in culture, ovary-independent tumorsare not stimulated to proliferate by these factors (Imagawa et al., 1992). A similar phenomenahas been demonstrated in carcinogen-induced mammary tumors; cells that are tumorigenic in vivohave been shown to grow independently of EGF, insulin or cholera toxin in vitro, whereas cellsthat form preneoplastic lesions in vivo are dependent on these factors for growth in vitro (Ethierand Cundiff, 1987). Using this tumor model, the growth factor-independent cells have beendemonstrated to achieve EGF autonomy by the autocrine production of EGF (Ethier andMoorethy, 1991). Thus, although hormones and growth factors are important in thetransformation of mammary tumors, with time the tumors seem to achieve autonomy from theeffects of hormones and growth factors.Breast Epithelial Cells and Androgens: Of particular interest to this thesis is the findingthat breast epithelial cells respond to androgens. Treatment with androgenic compounds has beendemonstrated to inhibit the growth of carcinogen-induced tumors in rats (Teller et al., 1966) and25 % of advanced human breast cancers regress in response to androgen treatment (TheCooperative Breast Cancer Group, 1961). The antiproliferative actions of these androgeniccompounds were postulated to occur by decreasing the release of gonadotrophins by the pituitary(Blackburn and Albert, 1959). More recently, it has been demonstrated that some human breastcancer cell lines possess androgen receptors (Labrie et al., 1992, Ormandy et al., 1992). Normalhuman breast epithelial cells also possess androgen receptors (De Winter et al., 1991, Wagner andJungblut, 1976). Further, it has been estimated that as many as 78 % of human tumors possessandrogen receptors and that up to 25 % of breast cancer patients whose tumors are progesterone-receptor negative still possess androgen receptors (Lea et al., 1989). These authors suggest thatthe ability of high dose progesterone treatment to cause tumor remission in some breast cancerpatients whose tumors are progesterone-receptor negative is due to the presence of androgenreceptors, as progesterone is suggested to cross-react with the androgen-receptor (Bullock et al.,1978).Studies of androgen's effects on breast cancer cells have produced varied results. Anumber of studies using human breast cancer cell lines demonstrated that physiologicalconcentrations of androgens have an antiproliferative effect on tumor cell growth (Labrie et al.,1992). The specificity of androgen for the androgen receptor was insured by the ability ofandrogen antagonists to block the actions of androgenic compounds (Labrie et al., 1992).Studies indicate that androgenic compounds completely block the growth stimulatory effects ofestrogen on human breast cancer cell lines (Labrie et al., 1992). However, several studies suggest272 8a possible role for androgens in breast epithelial cell carcinogenesis and the stimulation of thegrowth of breast cancer cells. First, alterations in testosterone metabolism in breast cancerpatients is a consistent finding (Moore eta!., 1986, Secreto eta!., 1991). Further, studies haveshown that lymph from vessels draining breast carcinomas have significantly higher levels ofandrogens than those observed in the patient's blood (Hamed eta!., 1991). It has also beenobserved that fluid in apocrine breast cysts contains high levels of both EGF and androgenmetabolites (Lai et al., 1989). The authors speculate that the androgens may stimulate theproduction of EGF in these abnormal breast cells (Lai et al., 1989). Interestingly, apocrine breastcysts are associated with an increased risk of developing breast cancer (Secreto eta!., 1989).Thus, it has been suggested that altered androgen metabolism in the patient may be associatedwith the initiation or the progression of breast cancer (Secreto eta!., 1989). Finally, it has beendemonstrated that treatment of some human breast cancer cells lines with androgen results in upregulation of progesterone-receptor expression (Ormandy et al., 1992). Thus, it is clear that bothnormal and malignant breast epithelial cells possess functional androgen receptors which can playa role in altering the growth of these cells.Hormone Nonresponsive Tumors: Altered hormone levels can also indirectly affecttumors that are considered hormone-nonresponsive. This effect can be mediated by severalmechanisms. First, some hormones are involved in the general control of cellular metabolism. Ithas been demonstrated that all mammalian cells in culture require hormones and growth factors toproliferate (Carney eta!., 1981, Hayashi and Sato, 1976, Rizzino and Sato,1978). For instance,the basic requirements for growth of small cell lung carcinoma cells in culture are hydrocortisone,insulin, transferrin, estradiol and selenium (Carney et al., 1981). However, in vivo otherhormones or growth factors may substitute for some of these requirements. Second, it has beendemonstrated experimentally that carcinogen-induced tumorigenesis of the skin, liver and lungare influenced by the hormonal environment of the host organism. For example, female micenormally exhibit higher basal levels of growth hormone than their male counterparts and this2 9increased level of growth hormone in female mice has been demonstrated to delay carcinogen-induced tumorigenesis in the liver (Blanck et al. , 1992). Third, altered hormone levels may affecthost processes which have a critical impact on the growth of the tumor. For example, it has beendemonstrated that elevated levels of glucocorticoids and heparin inhibit angiogenesis (Follananand Haudenschild, 1980). Thus alterations in hormones can have very important effects on bothhormone-responsive and hormone-nonresponsive tumors.D) Tumors, Stress and the Immune System.D.1) Tumor ImmunologyThe question of whether or not the immune system is capable of recognizing andresponding to tumors has been the source of considerable controversy in the last 90 years.Studies of tumor transplantation conducted in the early 1900's by Erlich and others seemed todemonstrate that the immune system has the ability to reject tumors. However, subsequentstudies demonstrated that the tumor rejection observed was in fact due to histoincompatabilitiesbetween the tumor cells and the recipient mice. This finding lead to the discovery of the majorhistocompatability complex (MHC), and an understanding of it's role in immune recognition offoreign cells and of the diversity of MI-IC molecules in allogeneic strains of animals of the samespecies. This discovery caused the concept of immune system-mediated rejection of tumors to fallinto disfavor.In the 1950's tumor rejection was demonstrated using tumors from syngeneic donor mice,here the donor and recipient were demonstrated to share the same MHC phenotype. Transplantsof normal skin grafts from the donor mouse to the recipient mouse were not rejected and failed toimmunize the recipient against a subsequent tumor graft from the donor mouse (Prehn and Main,1957). Since that time, numerous studies have demonstrated immune mediated-rejection of30tumors in experimental animals due to recognition of tumor-specific antigens. The discovery ofantibodies and humoral immunity caused further interest in the concept of antitumor immunity.However, there has been a general failure to demonstrate tumor-specific antigens that arerecognizable by antibodies (Schreiber eta!., 1988). Most antitumor antibodies were found toreact either with other normal tissues or with antigens expressed on fetal tissues. The overallfailure to demonstrate the existence of tumor-specific antigens caused the concept of antitumorimmunity once again to fall into disfavor by the late 1960's. The discovery of cytotoxic Tlymphocyte (CTL)-mediated immunity to virally infected cells and bacteria lead to yet anotherresurgence of interest in antitumor immunity in the early 1970's (Burnet, 1970). Current researchindicates that cell-mediated immunity and, to a lesser extent, humoral-immunity can play a role incontrolling the growth of tumors.Cytotoxic T lymphocytes (CTL): In contrast to the relative difficulty in demonstratingtumor-specific antigens with antibodies, CTL clones have been generated which specificallyrecognize tumor cells. CTL were the first cells which were found to be capable of cell-mediatedcytotoxicity against cells of the organism with an altered surface antigen expression (i.e. virallyinfected cells and tumor cells); they recognize and directly act on the tumor cell to lyse it. CTLrecognize antigen associated with class I molecules of the major histocompatability complex(MHC), which are expressed on the surface of all cells (Unanue and Cerrottini, 1989).Studies of the mechanism by which T lymphocytes recognize target cells have given usinsight into why T cells may be better suited to recognize transformed cells than are antibodies. Itis well known that T lymphocytes recognize foreign antigen in conjunction with class I (CTL) orclass II (Thelper cells) MHC through the T cell receptor (TcR) (Hedrick, 1988). It has beendemonstrated that the association of antigen with MI-IC molecules involves the partial proteolyticdigestion of the native peptide and the association of peptide fragments with the MEC molecule(Unanue and Cerottini, 1989). Class I MHC molecules associate with peptide fragments that are31produced by the digestion of the cell's own peptides, while class II MHC molecules associate withpeptide fragments of molecules from outside the cell (Unanue and Cerottini, 1989). Recently thethree dimensional structure of a class I MHC molecule has been elucidated using X raycrystallography (Hedrick, 1988). The upper surface of the molecule is in the form of a beta-pleated sheet topped by 2 alpha helixes. The alpha helixes form the sides of an elongated cleft andthe beta-pleated sheet forms the base of the cleft. Comparison of the three dimensional structureof the MHC molecule with its amino acid composition reveals that the variable regions of thepeptide would be those sequences involved in forming the top, sides and base of the cleft(Hedrick, 1988). Studies have shown that the smallest peptide fragment recognized by a Tlymphocyte is 8 amino acids long, but that the length varies, depending on the peptide and the Tlymphocyte clone involved (Hedrick, 1988, Unanue and Cerottini, 1989). It has been shown thatthe substitution of a single amino acid in a protein can dramatically increase the affinity of a CTLclone for that molecule (Hedrick, 1988). For example, studies looking at the reactivity of amouse T lymphocyte clone specific for pigeon cytochrome c showed that it reacts with higheraffinity to antigen-presenting cells which are presenting an insect cytochrome c peptide fragmentthan to the pigeon cytochrome c fragment. Analysis of the two peptide fragments showed thatthey differ only in that the insect peptide had an alanine molecule deleted (Hedrick, 1988). Thus itcan be seen that T lymphocytes respond to small peptide fragments of proteins (which do nothave to be membrane proteins) associated with MHC molecules. Small changes in the structureof a given peptide fragment are sufficient to alter the binding affinity of the TcR for the MHCantigen complex. This suggests that a point mutational change of an amino acid in the center of aglobular protein, which may not significantly alter the three dimensional structure of the protein,could render the molecule antigenic for T lymphocytes while not changing its antibody affinity atall.Although the previously mentioned studies suggest that tumor-specific antigens couldtheoretically be generated by a single point mutation in a cellular gene, experimental proof of such32an event occurring was lacking until several years ago. Several labs have used tumor-specificCTL clones to identify different tumor-specific antigens in both human and experimental animaltumors. One recent study generated antigenic variants of the non-immunogenic murine tumorP815 (a mastocytoma). The investigators subsequently generated CTL specific for one antigenicP815 subline (De Plain et al., 1988). Following this, they constructed a genomic DNA library forthe antigenic subline of P815. They used the CTL clone to screen non-antigenic P815 parentalline cells transfected with the DNA library for the presence of the antigen. Followingidentification of the portion of the DNA library containing the antigenic gene, they sequenced thegene and compared it with the mRNA of the non-antigenic parental line. The results demonstratethat the gene, which is expressed in both antigenic and non-antigenic P815 cells, is made antigenicby a single point mutation. This strongly suggests that tumor-specific antigens do exist and canbe generated by mutation of existing cellular genes. Thus, in addition to expressing normalcellular genes and fetal genes in an abnormal manner, tumor cells can also express antigenic,mutated cellular genes. The immune system probably can respond to both types of antigen.The stimulation of CTL is thought to involve both CTL and helper T lymphocytes. TheT helper (Th) cells recognize tumor antigen fragments in association with class II MHC moleculeson the surface of antigen-presenting cells, which include both macrophages and B lymphocytes(Unanue and Cerottini, 1989). Recognition of the antigen-MHC complex and simultaneousstimulation by cytokines produced by the antigen-presenting cell causes the Th cell to secrete thelymphokines interleukin 2 (IL2) and interferon (IFN) and to proliferate (Street and Mosman,1991). CTL precursor cells, referred to as virgin CTL, are not cytolytic. When the TcR of avirgin CTL recognizes a tumor-specific antigen fragment in association with class I MHC on thesurface of the tumor cell, this causes an increase in the expression of IL2 receptors on the CTL'scell surface. The combination of binding the antigen-MT-IC complex and binding the IL2produced by the Th cell causes the CTL to proliferate and become cytotoxic. The generation of33effective CTL-mediated immunity in the organism following virus infection requires about 5 to 8days to reach peak activity levels (Herberman and Ortaldo, 1981).Natural Killer Cells: A second cell type found to be important in mediating antitumorimmunity is the natural killer (NK) cell. NK cells were discovered in the early 1970's followingthe observation that nonimmunized human peripheral blood lymphocytes and mouse splenocytesconsistently produce low levels of lysis of certain tumor cell lines (Greenberg et al., 1974,Kiessling et al., 1975). This activity was subsequently found to be due to a unique population ofeffector cells which were termed NK cells. NK cells have been defined as large granularlymphocytes which have the spontaneous ability to recognize and lyse a variety of normal, virally-infected and malignant cell types in an MHC-nonrestricted fashion (Herberman, 1985). Thesecells have been shown to lack some of the characteristic markers of T lymphocytes, Blymphocytes and macrophages (Herberman, 1985). Importantly, it has been shown that NK cellsdo not rearrange or express the genes for either the T cell receptor or immunoglobins, keymarkers for T and B lymphocytes respectively (Robertson et al., 1990, Trinchieri, 1989). Unlikethe unique specificity of individual T lymphocytes for different antigens, NK cells are thought topossess a broad range of target cell specificities. However, there are indications that there isheterogeneity among NK cells with respect to target cell specificity (Dawson et al., 1992,Herberman, 1985). Despite more than 20 years of research into the phenomenon of NK cellactivity, the NK cell receptor has yet to be characterized. Recently, however, several putative NKcell receptor molecules have been isolated for various species including mice, rats and humans(Anderson, 1992, Giorda et al., 1992, Harris et al., 1992). It is possible that NK cells may utilizemore than one receptor molecule to increase the diversity of cell types that can be recognized(Anderson, 1992). As mentioned previously, NK cells are not MHC-restricted and thus do notrequire that their target cells express MHC class I or class II molecules for recognition and lysis.In fact, a number of studies have demonstrated that there may be an inverse relationship betweenthe expression of class I MHC molecules and the ability of NK cells to lyse the target cell34(Dawson et al., 1992). Interestingly, studies using B lymphocytes as target structures havedemonstrated that the expression of some class I MEC alleles protects the cells against NK celllysis, whereas the expression of other class I MHC alleles does not (Dawson et al., 1992).Molecular analysis of the MHC alleles has revealed that the ability of class I MHC molecules toprotect B lymphocytes from lysis by NK cells involves certain amino acid molecules which form asubpocket on the floor of the antigen-binding groove (Dawson et al., 1992). Substitution of asingle amino acid in this region was shown to abrogate the ability of the class I MHC molecule toblock NK cell lysis (Dawson et al., 1992). The exact mechanism by which MHC moleculesregulate NK cell activity is controversial (Dawson et al., 1992, Ljunggren and Kane, 1990).A variety of different molecules are capable of stimulating the activity and/or proliferationof NK cells. Treatment of NK cells with IL-2 or members of the interferon family (IFN-a, IFN-13,IFN-y) has been demonstrated to increase dramatically the lytic ability of NK cells (Ellis et al.,1989, Ortaldo et al., 1984, Trinchieri et al., 1984), as well as to extend the spectrum of cellslysed, allowing NK cells to lyse target cells which are not normally sensitive to NK cell lysis.These activated NK cells are referred to as lymphokine-activated killer cells (LAIC Ellis et al.,1989; Trinchieri et al., 1984). IL-2 has been found to have the most potent and the most diverseeffects on NK cells, stimulating increased NK cell activity, increased cell proliferation and thesecretion of a number of different cytokines including 1FN-y and TNF-a (Robertson and Ritz,1992). Research has demonstrated that there are many different subtypes of interferons a/I3 andthat they differ in their ability to stimulate the cytotoxic activity of NK cells (Herberman, et al.,1983, Li et al., 1990). In contrast to IL-2, IFN only weakly stimulates the proliferation of NKcells. Studies indicate, however, that some interferons may synergize with 1L-2 in stimulating theproliferation of NK cells (Robertson and Ritz, 1992). More recently, several other cytokines havebeen discovered that are capable of modulating NK cell activity, including IL-4 and IL-12. IL-4alone has a weak stimulatory effect on NK cell cytolytic activity, but in combination with 1L-2 ithas been shown to inhibit the ability of IL-2 to induce LAK activity and proliferation in NK cells3 5(Nagler et al., 1988). IL-12 was originally discovered as a product of some malignant Blymphocyte cell lines which stimulated the cytolytic activity of both T lymphocytes and of NKcells (Naume et al., 1992, Robertson et al., 1992). Studies indicate that IL-12 is similar to IL-2 inits potent ability to stimulate the lytic activity of NK cells, but that IL-12 does not stimulate theproliferation of NK cells (Naume et al., 1992, Robertson et al., 1992). It has been demonstratedthat, in the presence of the appropriate T helper lymphocyte population, IL-12 is a potentstimulator of the secretion of IFN-y by NK cells (Naume et al., 1992). In vitro studies utilizingantibodies to IL-12 indicate that the production of this molecule by normal peripheral bloodlymphocytes may be an important physiological mechanism for controlling NK cell activity in vivo(D'Andrea et al., 1992).NK cells have been shown to constitutively express the intermediate affinity LL-2 receptor(p'75) and approximately 10 % of human peripheral blood lymphocytes also express the p55 IL-2receptor to form the high affinity IL-2 receptor heterodimer (Robertson and Ritz, 1992,Robertson et al., 1992). Stimulation of NK cells with IL-2, IL-12 or IFN has been demonstratedto cause a dramatic up regulation of the expression of both IL-2 receptor molecules on the surfaceof NK cells (Naume et al., 1992, Robertson et al., 1992). Thus, it is likely that all of thesemolecules would increase the response of NK cells to suboptimal doses of IL-2.It has also been demonstrated that in vitro incubation of NK cells with target cells,including virally-infected cells, bacterial cells and tumor cells, stimulates NK cells (Biron andWelsh, 1982, Djeu et al., 1982, Rabinowich et al., 1992), inducing increased lytic activity, upregulation of IL-2 receptors and the secretion of the cytolcines IFN-y and TNF-a (Chong et al.,1989, Rabinowich et al., 1992). One study demonstrated that antibodies to IFN-a could blockthe stimulation of NK cells by target cells (Djeu et al., 1982). Thus NK cells are thought to bepotent mediators of antitumor immunity against blood-borne metastases and primary tumors atearly stages of tumor growth.3 6The ability of NK cells to recognize a broad range of transformed cells, and to respondimmediately has prompted the suggestion that NK cells are the body's first line of defense againsttumor cells (Herberman and Ortaldo, 1981). In experimental tumor models, NK cell have beenshown to play an important role in the control of blood borne metastases (Talmadge et al., 1980).Studies of the kinetics of clearance of radioactively labeled tumor cells from the body followingtheir injection into the bloodstream indicate that NK cells likely attack tumor emboli which havebecome arrested in the vascular beds of target organs (Aslakson et al., 1991, Johnson et al., 1990,Greenberg et al., 1987). Further, experimental models have demonstrated that tumor cell sublineswith decreased sensitivity to NK cell lysis in vitro have increased metastatic ability in vivo(Talmadge et al., 1980). In humans, studies have indicated that the patient's level of NK cellactivity is negatively correlated with the degree of lymph node involvement in breast cancer (Levyand Herberman, 1985). Studies in animals have also demonstrated that in some tumors, NK cellactivity can affect growth of the primary tumor as well as the metastases (Talmadge et al., 1980).These studies suggest that NK cell activity may be important in controlling the early growth of theprimary tumor when the tumor burden is small and that NK cells subsequently areimmunosuppressed as the tumor grows in size (Talmadge et al., 1980).Macrophages: A third effector cell which is active in cell-meditated antitumor immunity isthe macrophage. Macrophages, previously mentioned as playing a role in the presentation ofantigens to Th lymphocytes, are also known to be capable of exerting a tumoricidal effect ontumors (Adams et al. , 1985, Adams and Johnson, 1982). Tumoricidal macrophages exert theireffect either by phagocytosing and destroying the tumor cells or by the release of variouscytotoxic and cytostatic cytokines (Adams and Johnson, 1982, Schwamberger eta!., 1991). Therecognition of antigens by macrophages is poorly understood, but it has recently been suggestedthat they recognize negatively charged phospholipids such as phosphatidylserine (Fidler andSchroit, 1988). The abnormal distribution of phospholipids is postulated to allow macrophages to37recognize old cells, dead cells and tumor cells. Macrophages must be activated to exhibittumoricidal activities (Fidler and Schroit, 1988), and this activation may be induced bylymphokines (of which y-interferon is the most active) or certain bacterial cell wall products suchas lipopolysaccaride (Fidler and Schroit, 1988).Antibody Directed Cellular Cytotoxicity: A fourth mechanism whereby the immunesystem can recognize and lyse tumor cells is through antibody-directed cellular cytotoxicity(ADCC) (Kushner and Cheung, 1992). In ADCC, the recognition of the foreign cell occursthrough the binding of an antibody to the tumor cell. The antibody (typically of the immunoglobinclass gamma or IgG) is specific for and binds to a foreign antigen on the surface of the tumor cell.A number of cell types are then capable of recognizing the antibody-antigen complex and lysingthe tumor cells. Cells involved in mediating ADCC include a special class of cytotoxic cellstermed natural cytotoxic cells (NC), NK cells, macrophages and neutrophils (Kushner andCheung, 1992, Liesveld et al., 1991). NC cells and macrophages are the most potent mediatorsof ADCC and they possess receptors with the greatest affinity for the activated antibody-antigencomplex (Liesveld et al., 1991).Thus, the antitumor activity of the immune system is a diverse and potent defensemechanism which allows the body to guard itself against the growth of abnormal or transformedcells. The broad specificity and spontaneous activity of NK cells, NC cells and macrophagesallow them to recognize and destroy small foci of tumors. The specificity and potent cytotoxicityof CU, although taking longer to develop, allows them to destroy larger tumors.38D.2) PsychoneuroimmunologyClassical immunologists have long considered the immune system as being largely separatefrom other systems in the body. The immune system has mechanisms which were thought toprovide autonomous control of both cell division and cytolytic functions. However, in the past 15to 20 years it has been demonstrated that other systems do indeed interact with the immunesystem and that these interactions are bi-directional. This field of study is termedpsychoneuroimmunology. As the name implies, psychoneuroimmunology is the study of theinterconnections of the immune system, the endocrine system and the central nervous system. Ihave previously discussed the interconnections of the endocrine system and the central nervoussystem; in this section I will discuss the interactions of these two systems with the immune system.It has long been known that hormones can affect immune functioning. As early as 1936,Selye discovered that chronically elevated glucocorticoids can have a profound effect on theimmune system, causing marked thymic involution (Selye, 1975). It was subsequently discoveredthat thymic involution occurs because chronically elevated glucocorticoids cause a markedsuppression of immune function and lysis of developing T lymphocytes. A second key step in thedevelopment of the concept of psychoneuroimmunology was the finding that immune responsescould be conditioned just like other physiological functions. In classical conditioning, anunconditioned stimulus (UCS), which elicits a desired physiological response, is administered inconjunction with a conditioned stimulus (CS) which is initially neutral and does not elicit theresponse. After many such pairings, the presentation of the conditioned stimulus alone will elicitthe physiological response. In 1975, Ader and Cohen (Ader and Cohen, 1975) demonstrated thatimmune responses could be conditioned. They paired the injection of the immunosuppressivedrug cyclophosphamide (UCS) with the taste of saccharin (CS), and demonstrated thatsubsequent presentation of saccharin alone suppressed the antibody response to the administrationof sheep red blood cells compared to that seen in control animals. This indicated that the central39nervous system could act by some mechanism to suppress immune responses, since theassociation of the CS and the UCS is a central nervous system-mediated event. This finding hasbeen repeated (Gorczynski, 1987, Rogers et al., 1975; Wayner et al., 1978) and extended toinclude conditioning of cell-mediated immune responses (Bovbjerg eta!., 1982) and of NK. cellactivity (Greenberg et cd.,1986, Solvason eta!. 1988).These findings sparked interest in the ability of the central nervous system to control or atleast modulate the actions of the immune system. It was discovered that various neuropeptidesand hormones could affect immune responses in vitro. These various peptide molecules mayaffect lymphocyte functioning both in vivo and in vitro. However, the results have beenconflicting due to methodological differences in the source of the hormones, the doses ofhormones used, the presence of serum and the source and purity of white blood cells used (Dunn,1989). Experiments have generally demonstrated that both 13-endorphin and ACTH decreaseantibody production, while thyrotropin, growth hormone and prolactin stimulate antibodyproduction (Bernton, 1991, Blalock, 1992; Dunn, 1989; Johnson 1992; Kelley, 1991). T cellproliferation and functional activity has been reported to be stimulated by endorphins, growthhormone and prolactin, and inhibited by chorionic gonadotrophin (Bernton, 1991, Blalock, 1992;Dunn, 1989; Johnson 1992; Kelley, 1991). Researchers have subsequently discovered that notonly do lymphocytes respond to neuropeptides but that lymphocytes can produce neuropeptide-like molecules during an immune response. It was first demonstrated that lymphocytes producemolecules which are similar to the pituitary hormones ACTH and 0-endorphin during an immuneresponse to viral infections or transformed cells (Blalock and Smith, 1980). The amino acid andmRNA nucleotide sequences of mouse splenic- and pituitary-derived ACTH have beendemonstrated to be identical (Galin eta!. 1990, Smith eta!., 1990). Other studies havedemonstrated that lymphocytes may produce a number of different peptide hormone moleculesincluding prolactin, thyrotropin, growth hormone, chorionic gonadotrophin, luteinizing hormoneand follicle stimulating hormone (Blalock, 1992; Carr and Blalock 1991). It has been suggested4 0that different stimulating agents (e.g. concanavlin A vs allogeneic lymphocytes) result in therelease of different peptide hormones (Blalock, 1992). Further, lymphocytes have been shown topossess specific receptors for a number of different neuropeptide and hormone molecules. Thebest characterized of these receptors is the ACTH receptor. Lymphocyte receptors for ACTHhave been demonstrated which exhibit similar binding affinity to that reported for adrenal corticalcells (Clarke and Bost, 1989; Smith eta!., 1987). Interestingly, isolated rat B lymphocytes havebeen demonstrated to possess approximately 3 times the number of ACTH receptors that Tlymphocytes possess (Clarke and Bost, 1989). Stimulation of the lymphocytes to proliferatecauses an approximate doubling of ACTH receptor number (Clarke and Bost, 1989). The ACTHreceptors are functional in that ACTH causes a dose-dependent increase in the cytoplasmic levelsof the second messenger cyclic adenosine monophosphate (Clarke and Bost, 1989).The central nervous system can also affect the functioning of the immune system directlythrough the actions of peripheral nerves. It is known that the lymphoid organs are innervated bythe autonomic nervous system and recently it has been demonstrated by ultrastructural studiesthat these nerve processes make direct synaptic contacts with the lymphocytes (Felton, 1992).Cells of the immune system possess functional receptors for adrenaline (Madden, 1991),vasoactive intestinal peptide (Ottaway, 1991), and substance P (McGillis et al., 1991) andneurons containing these neuropeptides have been demonstrated in most lymphoid organs (Feltonand Felton, 1991). These neuropeptides have also been demonstrated to be capable of alteringimmune functioning in vitro.Thus, the central nervous system is clearly able to influence the functioning of the immunesystem through the release of various hormones and neuropeptide molecules. The immune systemis also capable of altering endocrine and central nervous system functioning. As mentionedpreviously, immune cells undergoing specific immune responses produce various peptide hormonemolecules and it has been suggested that these molecules are capable of interacting with endocrine41tissues to influence endocrine functions (Blalock, 1992). This concept is controversial, however,and it has been questioned whether or not the peptides secreted by the immune system couldreach sufficient concentrations in the blood to affect the endocrine system (Dunn, 1989). Finally,certain cytokines released by white blood cells during an immune response are potent stimulatorsof endocrine functions. For example, it has been demonstrated that interleukin 1 released byactivated macrophages is a very potent stimulator of ACTH release by pituitary corticotrophs, butit is not clear if this is a direct effect or mediated by stimulation of CRF release from thehypothalamus (Dunn, 1989). Thus, there are bi-directional links between the central nervoussystem, the endocrine system and the immune system which allow coordination of the functionalactivities of these 3 systems.E) Psychosocial StressorsAnimal models used to investigate the effects of stressors on the physiological functions ofthe animal have demonstrated that many different kinds of stressors may affect the organism.Stressors may be divided into two broad categories, physical and psychological. Physicalstressors involve an insult to the tissues of the organism such as heat, cold or laparotomy, whereaspsychological stressors induce the anticipation of threat or harm (Lazarus, 1971). Studies havedemonstrated that a simple psychological stressor such as moving a rodent to a new cage cancause significant elevations in plasma glucocorticoid levels which may be as great as thoseobserved with physical stressors such as electric shock. Psychosocial stressors, a type ofpsychological stressor, result from responses to intense social interactions, the lack of suchinteractions or even perhaps to other subtle social interrelationships (Asterita, 1985). Socialinteractions, or the lack of them, may have profound effects on both humans and animals. Inrodents, psychosocial stressors may include individual housing, crowded housing conditions andchanges in housing condition. In the research of this dissertation, we focused on psychologicalstressors as we feel that they are more relevant to the human condition than are physical stressors.4 2Individual Housing. Individual housing is known to cause marked changes in thebehavior, endocrine function and immune competence of rodents. The most profound change isan increase in aggression of males seen in both mice and rats following even relatively briefperiods of individual housing (Brain, 1975). Increases in inter-male aggression induced byindividual housing occur more often in mice than in rats and are best documented in mice (Brain,1975). Studies suggest that increases in the activity of neuronal circuits containing serotonin andnoradrenaline in limbic regions of the brain may be involved in mediating this phenomenon(Frances eta!., 1990, Gentsch eta!., 1990, Olivier B. eta!., 1989). Individual housing has beendemonstrated to decrease plasma glucocorticoid levels, increase plasma testosterone levels, anddecrease adrenaline turnover in the adrenal glands of male mice (Brain, 1975). These hormonalchanges are indicative of decreased stress levels and it has been suggested that individual housingis less stressful than group housing for mice (Brain, 1975). One basis for this argument is theobservation that in the wild, male mice are territorial and will subordinate or kill all adult malemice in their territory (Crowcroft et al., 1963, Mackintosh, 1970). Thus individual housing issuggested to induce territoriality in male mice (Brain, 1975). However, it has also beendemonstrated that individually housed animals are hyperresponsive to stressors (Giralt andArmario, 1987, Hatch et al., 1965). Thus, it is currently controversial whether or not individualhousing is a stressor in rodents. Finally, individual housing is known to alter the immunecompetence of rodents. Individually housed mice have been demonstrated to have greaterantibody responses to immunization with foreign proteins (Rabin et al., 1987 a, Salvin et aL, 1990,Vessey, , 1964). Also, in vitro T lymphocyte responses to mitogenic stimulation (concanvalin A,phytohemagglutinin, pokeweed mitogen) (Raab et al., 1986, Rabin et al., 1987b) and macrophagestimulation by bacterial antigens (Salvin et al., 1990) are greater in individually housed mice thanin group-housed mice. However, differential housing does not appear to affect immunecompetence equally in all strains of mice (Rabin et aL, 1987b). Interestingly, individual housinghas been shown to have variable effects on resistance of mice to different diseases. Individually43housed mice were shown to be more resistant to malaria (Plaut et al., 1969), but less resistant toencephalomyocarditis virus (Freidman et al., 1969) and to West Nile virus (Ben-Nathan andFeuerstein, 1990) than group housed mice.Crowding. Crowding in mice can be defined by the number of mice that are able tointeract with each other rather than the number of animals per square metre (Christian et al.,1965). Some of the earliest studies on the effects of crowding on mice were conducted in the1950's by Christian eta!. (1965). These studies demonstrated that as the density of mice in aconfined population increased, there was a significant increase in the weight of the adrenal glandand a corresponding decline in the weights of the thymus, testes and other sex organs (Christian etal., 1965). These physiological changes had been previously described by Selye (1975) ashallmarks of a stress response. Subsequent studies have demonstrated that housing rodents incrowded conditions caused marked changes in basal plasma levels of several hormones includingglucocorticoids (Brain and Nowell, 1970, Bronson, 1972, Peng et al., 1989), testosterone (Koikeand Noumura, 1989, Sayegh et al., 1990), TSH (Restrepo and Armario, 1989), GH (Restrepoand Armario, 1989), insulin (Restrepo and Armario, 1989), LH (Bronson, 1972) and FSH(Bronson, 1972) in the first 2 wks post grouping. However, by 4 wks post grouping, there wereno effects of crowded housing conditions on basal plasma hormone levels (Ortiz eta!.,1984,1985). Crowded housing conditions have also been demonstrated to affect immunefunction. Crowding has been shown to decrease antibody production in response to foreignantigens (Brayton and Brain, 1974, Edwards and Dean, 1977) and to decrease T lymphocyteproliferation in response to mitogens (Rabin eta!., 1987a). Generally, crowding has been foundto have a depressive effect on resistance to disease, including Salmonella infection (Edwards andDean, 1977), malarial infection (Plaut eta!., 1969) and viral leukemia (Ebbesen eta!., 1991).Thus crowding is probably a stressor although some adaptation may occur over time.4 4Dominance. It has been demonstrated that, in male animals, most effects of crowdedhousing conditions can be attributed to the establishment of a dominance hierarchy amongst theanimals (Bronson, 1972, Christian, 1970). Mice are reported to form a non-linear dominancehierarchy, with one dominant male and several equal subordinate males (Christian et al., 1965).Fighting is critical to the establishment and maintenance of the hierarchy, with the dominant maleinitiating and winning all encounters (Brain, 1972, Lyte eta!., 1990). Such fighting behavior isobserved in most newly formed groups of adult male rodents, even when the group consists ofonly 2 animals (Brain and Nowell, 1970). As mentioned previously, prior individual housingincreases the aggressiveness of rodents compared with that of continuously group housed animals.Subordinate rank in rodents and the defeat associated with such a position has been demonstratedto produce more dramatic alterations in endocrine and immune activity than are observed in thedominant male. Subordinate or defeated mice are known to have higher plasma levels ofglucocorticoids and lower levels of androgens than their dominant cage mate (Brain and Nowell,1970, 1971, Raab et al., 1986). As well, subordinate mice exhibit higher activity of tyrosinehydroxylase (a key enzyme in the synthesis of catecholamines) in the adrenal medulla (Raab et al.,1986) and increased release of opioid peptides (Miczek et al., 1982). Subordinance or defeat hasalso been shown to affect the immune system. Defeated rodents exhibit decreased antibodyproduction in response to bovine sera (Vessey, 1964), SRBC (Beden and Brain, 1982) andbacterial antigens (Ito et al., 1983). In addition, defeated mice have increased susceptibility toviral leukemia (Ebbensen et al., 1991), decreased responses to T lymphocyte mitogens (Raab etal., 1986) and, in aged mice, decreases in NK cell activity (Ghoneun et al., 1987). However, ithas also been shown that defeated mice have increased levels of macrophage phagocytic activitywhich may be blocked by opioid antagonists (Lyte et al., 1990). This study further demonstratedthat defeat-induced enhancement of immune function is not opioid mediated in all strains of mice(Lyte et al., 1990).In summary, it is apparent that psychosocial stressors are important in modulating theactivity of the central nervous system, the endocrine system and the immune system in animals.Thus, it is important to control for the effects of these potent stressors in all experiments.F) Chronic vs Acute StressorsAcute application of a stressor to an organism is known to cause alterations in the plasmalevels of a number of hormones and neuropeptides including ACTH (Armario, 1985, Armario etal., 1988), 0-endorphin (Tejwani et al., 1991), prolactin (Kant eta!., 1983,1985), LH (Briski andSylvester, 1988, 1987), growth hormone (Kant et al., 1983, 1985), adrenaline and noradrenaline(Cox eta!., 1985, DeTurck and Vogel, 1980, Kvetnansky eta!., 1977, 1979), testosterone(Demura eta!., Frankel and Ryan, 1981, 1989, Sapolsky, 1986), glucocorticoids (Armario, 1985,Armario eta!., 1988, Kant eta!., 1983,1985) and many other hormones. However, in contrast toacute stressors, chronic or repeated administration of a stressor has been shown to result in adecrease in the responsiveness of the organism to the stressor. This effect, termed habituation oradaptation, has been demonstrated to occur for both physical manifestations of the stress response(for example, stress induced analgesia and body core hypothermia) and endocrine responses(Levine et al., 1978). It is proposed that the phenomenon of habituation is a result of cognitivechanges rather than physical alterations of the hormone secreting cells, since habituation to onestressor does not provide protection from a second stressor and therefore could not be acting atthe level of hormone secretion (Kant et al., 1985). Interestingly, not all hormones showhabituation during exposure to chronic stressors and there are contradictory findings about theability of stressors to cause habituation of other hormones. ACTH, prolactin and catecholaminesecretion in response to chronic stressors have generally been demonstrated to consistently exhibithabituation (Armario, 1985, Armario eta!., 1988, Cox eta!., 1985, DeTurck and Vogel, 1980,Kant eta!., 1983, 1985, Kvetnansky eta!., 1977, 1979). Surprisingly, glucocorticoid levels haveoften been demonstrated not to decline with repeated presentation of the same stressor (Armario454 6eta!., 1984a, 1988, 1990, Hennessey and Levine, 1977, Irwin eta!., 1986, Kant, 1983). This islikely due to the ability of chronic stressors to induce hyper-responsiveness of the adrenalcorticotroph cells for ACTH (Armario eta!., 1988). It has been shown that plasma hormonelevels may return to basal levels but if a subsequent stressor is applied, there is hyper-responsiveness to this stimulus. Other hormones such as growth hormone and TSH are reportednot to alter their secretion patterns in response to repeated presentations of a stressor (Armario etal., 1984b, Kant et al., 1983). Different strains of inbred laboratory rodents have beendemonstrated to exhibit differences in the magnitude of their hormone response to a given stressorand in the ability of chronic application of that stressor to induce adaptation of the hormoneresponse (Kvetnansky eta!., 1979). Thus, when considering the effects of a stressor on anorganism, one must take into account the chronicity of the stressor and the potential of thehormones being considered to undergo adaptation to chronic stress.G) Animal-Tumor Model for Studying Effects of Psychosocial StressorsAn animal-tumor model which examines the effects of psychosocial stressors on thegrowth of a hormone-responsive mouse mammary tumor has been developed in our laboratory.In this model, the psychosocial stressors of differential housing has been demonstrated to induceboth increases and decreases in tumor growth rate.The Tumor. The tumor used in this model is the Shionogi carcinoma 115 (SC115). This tumorarose spontaneously in the breast epithelium of a female mouse of the DD/S strain. Following 19generations of transplantation, a subline arose which grew more rapidly in male mice than infemale mice. When 2 x 106 SC 115 cells are injected subcutaneously in the interscapular region ofa male mouse, a palpable tumor arises in approximately 7 d and grows to a mass of 2 to 3 g by 21d. When a similar tumor inoculum is injected into a female mouse, tumor growth rate isconsiderably slower. In females, a tumor may require 40 d to reach a mass of approximately 1 g.4 7The SC 115 tumor, similar to many human breast tumors, is hormone-responsive and anumber of different hormones have been demonstrated to modulate its growth rate. Cells of theSC 115 tumor have been demonstrated to possess functional androgen receptors and physiologicalconcentrations of androgens stimulate the growth rate of the tumor (Bruchovsky and Rennie,1978, Emerman and Siemiatkowski, 1984, Kitamura et at, 1979). However, the SC 115 tumor isheterogeneous and contains both androgen-responsive and androgen-nonresponsive cells.Removal of androgens from the environment results in the outgrowth of a slower growingandrogen-independent cell population (Bruchovsky and Rennie, 1978, Emerman and Worth,1985, Kitamura eta!., 1979). Pharmacological doses of glucocorticoids have been shown tostimulate the growth of the SC115 tumor, both in vivo (Watenabe eta!., 1982) and in vitro(Darbe and King, 1987). The SC 115 tumor has been shown to possess estrogen receptors and torespond to supraphysiological doses of estrogen (Nohno eta!., 1982). However, the stimulatoryeffects noted with pharmacological doses of estrogen are apparently mediated by cross-reactivitywith the androgen receptor and thus are only apparent at suboptimal androgen levels (Luthy etal., 1988, Noguchi eta!., 1987, Nono eta!., 1982). More recently, it has been demonstrated thatSC 115 cells are stimulated to grow by basic fibroblast growth factor (bFGF) and that anti-bFGFantibodies partially inhibit the growth stimulatory actions of androgens and glucocorticoids invitro (Furuya et al., 1990, Tanaka et al., 1990). It has also been demonstrated that TGF-13 maysuppress the ability of testosterone to stimulate the growth of SC 115 cells (Yamanishi et al.,1990). Thus, a number of hormones and growth factors are known to affect the growth of the SC115 tumor, but androgens are the primary stimulator of cell growth. Further it has beendemonstrated that morphological differences exist between tumors grown in male and femaleDD/S mice (Emerman and Worth, 1985, Kitamura eta!., 1979). Tumors grown in male miceexhibit a typical epithelial morphology, with cells appearing round to cuboidal and arranged inclumps or sheets with very little connective tissue separating the cells (Emerman and Worth,1985). In contrast, tumors grown in female mice display a fibroblast-like morphology with48spindle-shaped cells arranged in loose sheets or irregular cords, separated by large amounts ofconnective tissue.The Model. The model is based on a study by Sklar and Anisman (1980) which demonstratedthat individual housing as well as a change in housing group could markedly affect the growth of asyngeneic mastocytoma tumor. Thus, Drs. Emerman and Weinberg designed a series of studies toinvestigate the effects of different social housing groups on the growth of the SC 115 tumor. Thefirst study (Weinberg and Emerman, 1989) investigated the effects of housing group on thegrowth of the SC 115 tumor. The design of the study was as follows (figure 1): Male mice of theDD/S strain were housed either individually [I] or in groups of 3 [0], at time of weaning (3 wk ofage). When the mice were 2 to 4 months of age, they were injected with tumor cells and housedas follows. Mice raised as individuals were either rehoused as a male/female pair [IP] or wererehoused in groups of 5 males [IG]. Mice raised in groups either remained in their groups [GO]or were rehoused as individuals [GI]. Half of the mice in each housing group were subjected toan acute daily stressor consisting of exposure to 1 of 5 novel environments, 15 min/d, 5 d/wk.This daily stressor was designed to produce an acute rise in levels of plasma glucocorticoids. By23 d post tumor injection there were significant differences in tumor growth among the differentgroups (Figure 2). Overall, mice experiencing acute daily novelty stress had significantly largertumors than mice which did not experience this stressor (p<0.05). This effect was mostprominent in mice of the GI and IP groups, and in fact, tumor growth was not significantly alteredby novelty stress in mice of the GO group. In addition, collapsed across the novelty stresscondition, there were significant differences in tumor growth among mice in the 4 housing groups(p<0.05). Mice of the GG group had a tumor growth rate which was similar to that previouslyreported for this tumor (approximately 2 g at 3 wk). In contrast, tumor growth rate wassignificantly greater in mice of the GI group than that of mice in the GO group. Whereas, mice inthe IG group exhibited a significant retardation of tumor growth rate compared with tumorsgrown in mice of all the other housing groups (p<0.01). Plasma corticosterone levels were alsoFigure 1. Experimental Design of Model I.49Male Mice at Time of Weaning(3 Weeks)Housed in Groups^Housed as^of 3^ Individuals(G) (I)Remain as Rehoused as Rehoused as Rehoused asGroups of Individuals a Male/Female Groups of3 (GI) Pair 5(GG) (IP) (IG)5051Figure 2. Tumor Growth in Male Mice in the Four Housing Groups.^Points represent mean± SEM. GU, raised and maintained in sibling groups of three; GI, raised in sibling groups ofthree, then separated and housed singly; IG, raised singly housed, then rehoused in groups of 5males; IP, raised singly housed, then rehoused with a female. All mice were injected with 2 x 106SC115 tumor cells. At 23 d post tumor cell-injection, tumor growth collapsed across group, wasgreater in the presence of acute daily novelty stress than in its absense, p<0.05. In addition,collapsed across novelty stress, all groups differed significantly from each other in tumor size by23 d, GI>GG>1P>IG, p's<0.05. (Weinberg and Emerman, 1989).54_V^ IG152No Acute Novelty Stress0 2 4 6 8 10 12 14 16 18 20 22 24Days Post Injection544,..c^3co758 2E3i-1Acute Novelty Stress0 2 4 6 8 10 12 14 16 18 20 22 24—*-- GGGI—A— ipV^ IGDays Post Injection53measured in these mice at 3 wks post tumor injection and group formation. In mice not exposedto the acute daily novelty stressor, all mice exhibited low basal plasma corticosterone levels andthere were no significant difference among mice from the 4 housing groups. Mice which hadpreviously been exposed to the novelty stress were given a final exposure to novelty stressimmediately before termination. It was found that there was still a significant corticosteroneresponse to the acute daily novelty stress and that there were no significant differences amongmice of the 4 housing groups in their corticosterone response to this novelty stressor.A second experiment was conducted to study the effects of group vs individual housing aswell as a change in housing group on the growth of the SC 115 tumor. The design of the studywas as follows (figure 3): Male mice of the DD/S strain were housed either individually [I] or ingroups of 3 males [G], at time of weaning (3 wk of age). When the mice were 2 to 4 months ofage, they were injected with tumor cells and housed as follows. Mice raised as individuals eitherremained as individuals [II] or were rehoused in groups of 5 [IG]. Mice raised in groups eitherremained in their groups [GG] or were rehoused as individuals [GI]. All mice were subjected toan acute daily stressor consisting of exposure to 1 of 5 novel environments, 15 min/d, 5 d/wk.This study replicated the tumor growth results of the previous study; in mice of the GU group(the standard colony housing group), tumors grew to a mass of approximately 2 g by 3 wk,whereas mice of the GI group had significantly increased tumor growth rates and mice of the IGgroup had significantly slower tumor growth rates compared with those of mice in the GG group(Figure 4). Interestingly, mice of the II group had a tumor growth rate that was intermediatebetween that of GO housed mice and that of GI housed mice. This study also examined the basalplasma levels of testosterone and dihydrotestosterone in mice from the 4 housing groups at 3 wkpost tumor injection, as androgens are known to have an important stimulatory effect on SC 115growth. There were no significant differences in the plasma levels of testosterone ordihydrotestosterone among mice of the 4 housing groups. The mice were also studied fordifferences in humoral or cellular immunity as there are suggestions that the immune system mayFigure 3. Experimental Design of Model II.54Male Mice at Time of Weaning(3 Weeks)Housed in Groups^Housed as^of 3^ Individuals(G) (I)Remain as Rehoused as Remain as Rehoused asGroups of Individuals Individuals Groups of3 (GI) (II) 5(GG) (IG)All Mice Exposed to Acute Daily Novelty Stress(15 min/d, 5 d/wk)5556Figure 4. Tumor Growth in Male Mice in the 4 Experimental Housing Groups. Pointsrepresent mean ± SEM. GG, raised and maintained in sibling groups of three; GI, raised in siblinggroups of three, then separated and housed singly; IG, raised singly housed, then rehoused ingroups of 5 males; II, raised and maintained singly housed. All mice were injected with 2 x 10 6SC 115 tumor cells. At 18 d post tumor cell-injection„ GI=II>GG>IG, p's<0.05. (Weinberg andEmerman, 1989).573.503.00ID'^2.50I–I(.9^2.0011-Cc^1.5002m1-^to o0.500.000^ 10^ 20DAYS POST INJECTION58play a role in modulating the growth of the SC115 tumor (Kitamura eta!., 1979, Nono et al.,1986, Watenabe et al., 1982). This study examined general measures of immune competence ofB and T lymphocytes, looking at secondary antibody production in response to a foreign protein,bovine serum albumin (BSA) and the ability of T lymphocytes to proliferate in response to themitogen concanavalin A (Con A) (Weinberg and Emerman, 1989). The presence of a tumorresulted in a general suppression of both B and T lymphocyte functions in all tumor-injected micerelative to their vehicle-injected counterparts at 3 wks post injection and group formation(p<0.05). Further, it was shown that, collapsed across the tumor/vehicle-injected condition, IGmice, who develop the smallest tumors, demonstrated a greater antibody response to BSA thandid mice of the GI group, those mice who develop the largest tumors (p<0.057). T lymphocyteproliferation in response to Con A did not differ significantly among tumor-injected mice from the4 housing groups. However, in vehicle-injected mice, mice of the GG group exhibitedsignificantly greater T lymphocyte proliferation than mice in the IG group (p<0.05).H) Thesis ObjectivesIn light of the ability of stressors to affect both the immune and the endocrine system, thisthesis investigated the role of selected immune and endocrine variables in mediating the effects ofpsychosocial stressors on tumor growth rate as observed in our model.Several studies suggest that the immune system may play a role in mediating the effects ofpsychosocial stressors on SC 115 tumor growth rate. First, it has been demonstrated that, inDD/S mice, injection of heat killed Staphylococcus aureus prior to tumor cell-injection results indecreased growth of the SC 115 tumor (Nohno eta!., 1986). As the injection of Staphylococcusaureus is known to stimulate the immune system, the authors conclude that the immune systemmay affect growth of the SC 115 tumor. Thus, our laboratory is interested in the role of theimmune system in mediating the differential tumor growth rates observed in our model. There are59several effector cells which could be involved in mediating the differential tumor growth includingCTL and NK cells. Previous studies in our laboratory demonstrated that the presence of theSC 115 tumor significantly stimulates splenic NK cell activity compared with that of vehicleinjected mice at 3 and 7d post injection (Rowan, 1992). Furthermore, at 3 d post tumor cell-injection, when the greatest stimulation of NK cell activity was observed, mice of the GI group(largest tumors) had significantly greater levels of splenic NK cell activity than did mice in allother groups (Rowan, 1992). Thus, the role of NK cells in mediating the effects of psychosocialstressors on growth of the SC115 tumor in our model was examined further. The time coursestudy of splenic NK cell activity was completed by examining NK cell activity of tumor cell- andvehicle-injected mice from the 4 experimental housing groups at 1 d post injection. As well, NKcell activity of tumor-infiltrating lymphocytes was studied in mice of the GI (largest tumors) andIG (smallest tumors) at 3 and 7 d post injection when NK cell activity of tumor-infiltratinglymphocytes was demonstrated to be maximal. Finally, the effect of direct in vivo modulation ofNK cell on the differential tumor growth rates observed in our model was investigated in mice ofthe GI and IG groups.The endocrine system is also likely to be involved in mediating the effects of psychosocialstressors on tumor growth in our model, as the SC 115 tumor is hormone responsive and stressorsare known to alter plasma hormone levels. Thus, to determine if alterations in plasma hormonelevels may be involved in mediating the effects of psychosocial stressors on tumor growth rate inour model, plasma levels of testosterone and corticosterone were measured in mice of the 4experimental housing groups in the first wk post tumor cell-/vehicle-injection. Furthermore, theSC 115 tumor contains subpopulations of cells with different degrees of hormone responsivenessand selection for hormone-nonresponsive cells (i.e. tumors grown in female mice) results in atumor with a slow growth rate. Hormone-nonresponsive tumors grown in female mice have beendemonstrated to exhibit altered morphological characteristics compared with hormone-responsivetumors grown in male mice housed in our standard colony conditions. Thus, the morphological60characteristics of tumors grown in mice from the 4 experimental housing conditions wereexamined and compared with tumors grown in male and female mice housed in our standardcolony conditions. As well, the hormone responsiveness of tumor cells was directly measured inan in vitro assay. The ability of tumor cells from mice of the GI (largest tumors) and IG (smallesttumors) groups to proliferate in response to in vitro stimulation with either dihydrotestosterone orhydrocortisone was examined.CHAPTER 2: GENERAL METHODS:A. Tumor Model.The androgen-responsive SC115 mouse mammary carcinoma, subline class I (Bruchovsky& Rennie, 1978), was maintained by serial transplantation in male mice of the DD/S strain. Forpropagation, tumors were dissociated to a single cell suspension (as described below) and malemice (2-6 months old), housed in the standard colony condition, were injected with 2 x 106cells/mouse. The standard colony condition consisted of male mice raised and housed in siblinggroups of 3.A.1) Dissociation.Tumors weighing approximately 2 g were dissected free of extraneous tissue with steriletechnique and finely minced with opposing scalpel blades. The tissue was transferred to adissociation flask and approximately 15 ml Saline-Trypsin-Versine (STV) added. STY consistedof 0.05% trypsin (1:250) and 0.025% EDTA (Sigman Chemical Co., St. Louis, MO) in Ca2+ -Mg2+ - free Saline A, pH. 7.3. The flask was then gently swirled for 2 min; the contentstransferred to a 50 ml conical centrifuge tube and spun at 80 x g for 1 min in a bench top clinicalcentrifuge (ICU). The supernatant was then transferred to a second 50 ml centrifuge tube with anequal volume of Dulbecco's Modified Eagle's Medium (DMEM; Terry Fox Laboratory,Vancouver, BC) and 5% Calf Serum (CS, to inactivate the trypsin). The tube was spun at 200 xg for 4 min to enrich for epithelial cells. The pellet was then resuspended in 5 ml DMEM andplaced in a 370 C waterbath.The tissue in the original centrifuge tube was transferred to the dissociation flask and 15ml STY was added. The flask was left shaking at 100 rpm on a gyrator shaker (Junior Orbit6162Shaker, Lab-Line Instruments, Inc., Ill) in a 370 C incubator for 7 min. The contents of the flaskwere transferred to a 50 ml centrifuge tube and centrifuged at 80 x g for 1 min. The supernatantwas collected and combined with an equal volume of DMEM and 5% CS. This was spun for 5min at 400 x g, resuspended in 5 ml DMEM and placed in a 370 C waterbath. The remainingtissue was placed back in the flask for a third and final dissociation with STV for 7 min. Thesupernatant was collected as described above, combined with an equal volume of DMEM and 5%CS, and centrifuged. The resulting pellet was resuspended in 5 ml DMEM. All 3 cell suspensionswere then pooled and passed through a 50 pm Nitex filter (Tetko, Inc., Elmsford, NY) to removecell aggregates and debris. The resulting single cell suspension was centrifuged at 340 x g for 5min and the pellet resuspeneded in 20 ml DMEM. An aliquot of this cell suspension was diluted1:10 with DMEM and counted on a haemocytometer using trypan blue exclusion to determine thenumber of viable cells. The plasma membranes of dead cells are not able to prevent trypan blue(pH 7.2) from entering the cytoplasm, and therefore the dead cells stain blue. The suspension wasthen adjusted to the concentration desired for either freezing (see below) or injecting into mice(see below).A.2) Freezing of Tumor Cells.Although the SC 115 carcinoma was maintained in vivo, SC 115 cells were also stored in aliquid-nitrogen (LN2) storage tank. To freeze SC 115 cells, dissociated SC 115 cells were adjustedto a concentration of 1 to 1.5 x 107 cells/ml in freezing media (50% DMEM + 44% CS + 6%dimethylsulfoxide (DMS0)). Freezing vials were labeled with cell type and date of dissociationand 1 ml of cell suspension in freezing media was aliquoted to each vial. Vials were then slowlyfrozen according to a schedule provided by the manufacturer of the freezing tank (Handi-Freezefreezing tray, Union Carbide), before being transferred to a liquid nitrogen storage tank (MVECryogenics, TA).63A.3) Thawing of Tumor Cells.Frozen vials of SC 115 tumor cells were removed from the LN2 storage tank and rapidlythawed in a 370 C water bath. The tumor cell suspension was transferred to a 15 ml conicalcentrifuge tube, diluted with an equal volume of warm (370 C) DMEM and spun for 5 mm at 400x g. The supernatant was discarded, the pellet resuspended in 10 ml DMEM and spun for 5 minat 400 x g (wash step). This wash step was repeated. The pellet was resuspended in 10 ml ofDMEM and an aliquot counted as previously described. Cells were resuspended at 2 x 107cells/ml for tumor injection.A.4) Transplantation of Tumor Cells.For tumor cell-injection, the cell suspension was adjusted to 2 x 107 cells/nil in DMEM.The total volume of tumor cell suspension required was 100 il x the number of mice injected +100 ill to allow for retention of fluid in the head of the syringe. The mice were lightlyanesthetized with ether and then injected s.c. into the interscapular region with 100 Ill of tumorcell suspension. Care was taken to lift the skin high up so as not to penetrate underlying tissueand to plunge the needle in deep enough to minimize the amount of cell suspension that leakedback out.A.5) Monitoring Tumor Growth.For tumor propagation, mice were palpated twice weekly after tumor injection, beginningon d 5 - 8 when a palpable tumor generally first appears and terminating at 18 - 21 days, whentumor size was between 2 and 3 g. When tumors were measurable, caliper measurements weretaken and tumor weights calculated according to the formula (Simpson - Herron & Lloyd, 1970)length (cm) x [width (cm)]2 — g2For experiments, mice were palpated twice a week, starting on day 6 or 7. Tumor weightswere calculated when caliper measurements were able to be taken and the last measurement wastaken on the day of termination.B. Animal Model.Mice were housed in polycarbonate cages (18 x 29 x 13 cm) with stainless steel lids, corncobb bedding (Sanicel) and received food (Purina mouse chow pellets) and water ad libitum.Cages were placed on stainless steel racks in a room with a 12 h dark/light cycle (0700 h - 1900h). The room was relatively free from extraneous building noise and remained at a constanttemperature of 22° C.Following weaning (3 wk of age), male mice were housed individually [I] or in siblinggroups of 3 [G]. When the mice were 2 to 4 months of age, they were injected s.c. in theinterscapular region with a single cell suspension of 2 x 106 cells in 100 ill of DMEM (tumourgroups) or with DMEM alone (Vehicle groups) and housed as follows:1) II - males raised individually housed remained individually housed for the experiment.2) IG - males raised individually housed were placed in groups of 5 animals for theexperiment.3) GG - males raised in sibling groups of 3 remained in their sibling groups for the experiment.4) GI - males raised in sibling groups of 3 were rehoused as individuals for the experiment.6465Animals in all groups were subjected to an acute daily stressor consisting of exposure, 15min/d, 5 d/wk (prior to 1200 h), to 1 of 5 novel environments. The exposure to novelenvironments followed a set order of rotation as follows:1) a round clear plastic jar 9 cm in diameter and 7 cm high with a white plastic screw top lid,2) a polypropylene box 12 x 10 x 4 cm with a lid,3) a covered cardboard box with cardboard divisions forming compartments 7 x 7 x 14 cmhigh,4) a plastic cup (220 ml - 10 cm in height and top diameter of 6.5 cm, base diameter of 4.75cm) with lid,5) a clean cage (empty of bedding, food or water bottle) with a standard cage top.All lids had hole punched in them for adequate ventilation.For termination, the room was closed off the night before the experiment at 1800 h and noone entered the room prior to the start of the experiment. Previous data demonstrate that, underthese conditions, animals are in a non-aroused state, as measured by plasma corticosterone levels(Weinberg & Bezio, 1987). For termination, each cage was quietly removed from the colonyroom and carried to an adjacent laboratory, the animals removed from their home cages, weighedand decapitated immediately. Where appropriate, trunk blood was collected and the requiredtissues removed from the animal.CHAPTER 3 IMMUNE STUDIES:A. Completion of Splenic NK Cell Activity Time Course Assay.IntroductionTumor growth may be influenced by an animal's social housing condition (Riley, 1981,Riley eta!., 1981, Steplewski eta!., 1987). Group housed animals typically have smaller tumorsand show increased rates of tumor regression compared with individually housed animals. Wehave demonstrated that the transplantable androgen-responsive Shionogi mouse mammarycarcinoma (SC 115), a typical adenocarcinoma with a sheet-like epithelial morphology (Emermanand Worth, 1985), also has differential growth rates when male mice are housed under differentconditions (Weinberg and Emerman, 1989). Possible physiological mediators of this differentialtumor growth were investigated in this thesis. The first set of studies focused on the immunesystem.It has been demonstrated that the immune system is important in regulating the growth ofsome tumors (Cerottini and Brunner, 1974, Herberman and Ortaldo, 1981). Previous studiesprovide indirect evidence that the growth of the SC 115 tumor is modulated by the immunesystem. First, it has been shown that the original androgen-dependent SC 115 tumor, which doesnot grow in females or castrated males, can be induced to grow in DD/S mice by theadministration of pharmacological doses of glucocorticoids (Watanabe et al., 1982). The authorsconclude that the growth-promoting effect of glucocorticoids is due in part to suppression of hostimmunity and in part to stimulation of a glucocorticoid receptor in the SC 115 cells. Second, itwas found that the development of the SC 115 tumor is retarded by the injection of666 7Staphylococcus aureus on the day of tumor cell-injection (Nohno eta!, 1986). This again isstrongly suggestive of a role for the immune system in modulating the growth of the SC 115tumor. The immune cells which are responsible for this modulatory effect on SC 115 tumorgrowth have not yet been determined. However, one possible candidate is the NK cell. NK cellsare currently thought to be one of the body's first lines of defense against tumor cells (Haller eta/., 1977, Herberman and Ortaldo, 1981, Wei and Heppner, 1987). Data suggest that these cellsare active very early in the antitumor immune response when the tumor burden is low and laterare subject to immunosuppression as the tumor grows in size (Gerson et al., 1981, Wei andHeppner, 1987). Studies have indicated that the degree of axillary lymph node involvement inbreast cancer patients, a measure of the degree of metastasis of the tumor and thus an importantprognostic indicator, exhibits a significant negative correlation with the patient's level of NK cellactivity (p<0.05, Levy et a/, 1985). It has also been shown that the patient's level of NK cellactivity 15 months after removal of the primary tumor is negatively correlated with the chance ofrecurrence at 5 years (Levy eta!, 1991).In light of the important antitumor role played by NK cells, a study was undertaken toinvestigate the role of NK cell activity in modulating the differential tumor growth rates observedin mice of the 4 housing groups of our experimental model. The initial part of this study, in whichI played a significant role, investigated splenic NK cell activity in both tumor cell- and vehicle-injected mice from the 4 housing groups at 3 d and 7 d post injection and group formation(Rowan, 1992). Overall, data indicated a significant effect of both the presence of the SC 115tumor and differential housing group on splenic NK cell activity. At 3 d post injection and groupformation, there was a significant (p<0.001) stimulation of NK cell activity in the spleens of tumorcell-injected mice compared with their vehicle-injected counterparts (Figure 5). In addition, therewere significant differences among tumor cell-injected mice of the 4 housing groups in their levelof splenic NK cell activity (p<0.001). Interestingly, post-hoc analysis revealed that mice of the GI68Figure 5. Lytic Activity of Splenic NK Cells 3 d Post Injection. Points represent mean ± SEM.Mice from the 4 experimental groups (as described in figure 2) were injected either with 2 x 106SC 115 cells (open symbols) or vehicle (solid symbols). Three days later, the lytic activity ofsplenic NK cells was measured. NK cell activity was significantly greater in tumor cell-injectedthan in vehicle-injected mice, p<0.05. Within the tumor injected condition, GI > GO = II,p's<0.05. (n's: GG-tumor (GG-T)(9), GO-vehicle (GO-V)(7); GI-tumor (GI-T)(6), GI-vehicle(GI-V)(5); II-tumor (II-T)(6), II-vehicle (II-V)(6); IG-tumor (IG-T)(11), IG-vehicle (IG-V)(15).Spontaneous release was usually less than 15% of the total release and always less than 20%.10080200695:1^9:1^18:1^37:1^75:1 150:1Effector:Target ratio70group (who develop the largest tumors) had significantly greater NK cell activity, at the lowereffector to target cell ratios, than did mice of all other housing groups. No differences were foundin NK cell activity among vehicle-injected mice from the 4 housing groups. At 7 d post injectionand group formation (Figure 6), the spleen cell populations of tumor cell-injected mice once againhad significantly greater levels of splenic NK cell activity than did those of vehicle-injected mice(p<0.001). However, at 7 d, there were no significant differences in splenic NK cell activityamong mice from the 4 housing groups in either the tumor cell- or the vehicle-injected conditions.Furthermore, splenic NK cell activity of tumor cell-injected mice was significantly lower at 7 dthan at 3 d post injection (p<0.001). Thus, this study demonstrated that the presence of theSC 115 tumor significantly stimulated NK cell activity in mice from the 4 housing groups.However, at 3 d post injection, the differential levels of NK cell activity observed among tumorcell-injected mice of the 4 housing groups were not in the direction that we would hypothesize.That is, mice of the GI group, in which the largest tumors will develop (Weinberg and Emerman,1989), also had the highest levels of splenic NK cell activity. Further, it was observed that thetumor-induced stimulation of splenic NK cell activity was clearly declining by 7 d post tumor cell-injection. Therefore it is possible that maximal stimulation of NK cell activity occurred earlierthan 3 d post tumor cell-injection. This possibility is supported by the finding that injections ofbiological response-modifiers such as IL-2 or y IFN causes stimulation of NK cell activity by 18to 24 h post injection (Ortaldo eta!., 1989, Talmdge eta!., 1985). Thus, as part of thisdissertation, a study was undertaken to examine splenic NK cell activity in tumor cell- andvehicle-injected mice from the 4 housing groups at 1 d (24 h) post injection and group formation.In addition, data on spleen weights and body weights of tumor cell- and vehicle-injected micefrom the 4 housing groups at 1 d, 3 d and 7 d post injection were examined. Measurements ofsplenic NK cell activity at 3 and 7 d post group formation were reported in a previous thesis(Rowan, 1992) and at the same timepoints, spleen weight and body weight were measured. Asthese data were not reported in the previous thesis, they are reported and analyzed here. As the 1d data described in this dissertation were part of the complete time course study (Rowse eta!.,71Figure 6. Lytic Activity of Splenic NK Cells 7 d Post Injection. Points represent mean ± SEM.Mice from the 4 experimental groups (as described in figure 2) were either injected with 2 x 106SC 115 cells (open symbols) or vehicle (solid symbols). At 7 d, the lytic activity of splenic NKcells was measured. (n's: GG-T(9), GG-V(8); GI-T(5), GI-V(6); II-T(6), II-V(6); IG-T(15), IG-V(15)). Spontaneous release was usually less than 20% of the total release and always less than30%.10080200725:1^9:1^18:1^37:1^75:1 150:1Effector:Target ratio1990), the discussion of this study will include data on spleen weight, body weight and NK cellactivity from all 3 test days (1, 3 and 7 d).METHODS AND MATERIALSAnimals: Seventy eight male DD/S mice were used in this experiment. Animals werereared and housed as described previously in the General Methods section.Tumor Cells: SC 115 tumor cells were prepared and injected as described previously inthe General Methods section.NK cell assay: NK cell activity was assayed as described by Greenberg (Kraut andGreenberg, 1986) with spleens from mice from the 4 housing groups injected with either tumorcells or vehicle. Briefly, animals were weighed, exposed to ether (10-15 sec) and rapidlyterminated by decapitation. Spleens were then aseptically removed.Dissociation of Spleens. In a laminar flow hood (Canadian Cabinets Co., Ottawa, Ont.)each spleen was placed in a sterile 60 mm petri dish (spleens were not pooled) and washed withapproximately 10 ml warm (37°C) RPMI 1640 media (Terry Fox Laboratory). The spleens werethen aseptically transferred to a second set of 60 mm petri dishes containing sterile stainless steelwire mesh screens and approximately 15 ml RPMI. The spleens were cut into several sectionsand gently pressed through the wire mesh using the flat end of a plunger from a plastic 10 mlsyringe. The spleen cell suspensions were transferred to 50 ml polypropylene conical centrifugetubes using 10 ml disposable pipettes. The screens were each washed with approximately 15 mlRPMI which was then added to the centrifuge tubes. The spleen cells were centrifuged for 2 minat 600 x g to pellet the cells. The supernatants were decanted and the cells resuspended in 3 mltrisma base-NH4C1 solution and incubated at 25°C for 3 min to lyse the red blood cells. The cellswere pelleted again by spinning at 600 x g for 2 min and the trisma base solution was decanted.7374The cells were washed twice in approximately 10 ml RPM" containing 10 mM Hepes buffer(Sigma Chemical Co., St. Louis, MO).Removal of B Lymphocytes. Nylon wool columns were prepared by placing 0.7 g nylonwool (Fenwal, Deerfield, IL) into 10 ml pipettes and packing the wool to the 7 ml mark. Thecolumns were then wrapped in aluminum foil and autoclaved. On the day of the experiment, thecolumns were conditioned by first wetting the column with 10 ml warm RPM! (tapping the side ofthe column with a 10 ml pipette until all of the nylon wool was wet with medium) and thenpouring 20 mls of RPM! containing 10 mM Hepes and 10 % fetal bovine serum (Grand IslandBiology Co., Burlington, Ontario) (RPM! + H + FBS) through the column and incubating it for 1h in a humidified incubator at 95 % air, 5% CO2. The spleen cells were resuspended in 1 ml ofRPM + H + FBS and placed on the column with a pastuer pipette one drop at a time. RPM' + H+ FBS (0.5 ml) was then added to each column to ensure that the spleen cells entered the nylonwool. The columns were incubated for 1 h (as above). The T lymphocytes were eluted from thecolumns with 10 ml RPM! + H + FBS (added to the columns drop-wise) and collected in a 15 mlpolystyrene centrifuge tube (Falcon). This procedure has been shown to eliminate B lymphocytesand macrophages, thus enriching the spleen cell preparation for T lymphocytes and NK cells(Julius et al., 1973).Target Cells. Yac-1 lymphoma cells (a generous gift of Dr. D. Chow, Manitoba Instituteof Cell Biology), a Maloney lymphoma virus infected lymphocyte cell line which is highlysusceptible to NK cell lysis, were stored in liquid nitrogen in 1 ml aliquots (107 cells/till). At least48 h before use in an assay, the cells were rapidly thawed as described for SC 115 cells in GeneralMethods, resuspended in 50 ml RPM! + H + FBS at a concentration of 5 x 104 cells/ml, placed ina 75 cm2 tissue culture flask (Corning Glass Works, Corning, NY) and incubated at 37°C in 95 %air, 5% CO2. The cells were resuspended in fresh medium at 5 x 104 cells/ml every 24 h toensure that cells remained in an exponential growth phase. For the experiment, 1 x 107 Yac cells75were placed in a 15 ml round bottomed centrifuge tube (Falcon), spun for 2 mm at 600 x g andthe supernatant decanted. The cells were then resuspended in 10 ml fresh RPMI + H + FBS,centrifuged for 2 min at 600 x g and the supernatant decanted, leaving approximately 100 gilsupernatant behind with the cell pellet. Sodium [51Cr] chromate (500 RCi/ml, Amersham,Oakville, Ont.) (100 ul) was added to the tube and the cells resuspended by hand vortexing. Thecells were then incubated in a water bath at 37°C for 80 min with frequent hand vortexing tomaintain the cells in suspension. Following this, the cells were washed twice with 10 ml RPMI +10mM hepes and resuspended at 105cells/m1 of medium (RPMI + H + FBS).Final Preparation and Plating of Cells. Viable effector cells were counted on ahemocytometer and resuspended in RPMI + H + FBS at a concentration of 1.5 x 107 cells/mlmedium. For each spleen, cells were plated by adding 100 IA spleen cell suspension to 3 wells ofrows 1 and 2 of a 96 well v-bottomed plate (Flow Laboratories Inc., McLean, VA) and adding100 IA media to wells of rows 2 through 5; spleen cells were serially diluted by resuspending thecells in row 2 and placing 100 ill of this suspension into the wells of row 3. The cells in row 3were then resuspended and 100 ill of this suspension placed in the wells of row 4. This processwas repeated for rows 4, 5 and 6. This serial dilution produces effector to target cell ratios of150:1, 75:1, 37.5:1, 17:1, 9:1, 4.5:1 (3 wells per ratio per spleen). Aliquots of 100 t.t1 target cellsuspension (105 cells/m1) were added to each well of effector cells. In addition, 6 wells wereplated with 100 ill of target cells and 100 IA of media alone to measure the spontaneous lysis ofthe target cells and 6 wells were plated with 100 pi of target cells and 100 ill of 1 N HC1 tomeasure the total radioactivity available to be released by the target cells. Plates were incubatedfor 6 h at 37°C in 95% air, 5% CO2. The plates were centrifuged for 10 min at approximately150 x g, 100 lil of supernatant was removed and the [51Cr] release from lysed target cells wasdetermined by gamma counting. The percent of specific lysis (chromium released) at eacheffector to target cell ratio was computed using the formula:Test CPM -Spontaneous CPM% specific lysis=^ x 100Total CPM - Spontaneous CPMwhere Test CPM = counts in experimental wells containing target cells and effector cells.Spontaneous CPM = counts in wells containing only target cells.Total CPM = counts obtained by adding 100 gl 1N HCI to target cells to lyse all cells.To ensure that NK cells were in fact the primary cells responsible for the target cell lysisobserved in these assays, 2 additional groups (GG, n=3, of and GI, n=3) of mice were injectedwith tumor cells. Three days later spleens were removed, dissociated and treated with anti-AsialoGM1 (Wako Chemical USA Inc., Dallas, Tx) and complement (Cedar Lane Laboratories Ltd.,Hornby, Ontario), a procedure shown to suppress NK cell activity (Kasai et al., 1980). Thetreated spleen cells were then assayed for NK cell activity.Data Analysis: Statistical analyses were performed with appropriate analyses of variancefor the factors of Group, Tumor and Ratio, with repeated measures on the last factor. Significantmain effects were analyzed by Tukey's post-hoc tests.ResultsBody Weights: Body weights were analysed separately for tumor cell- and vehicle-injected conditions at 1 d, 3 d and 7 d post injection and group formation. The body weights ofvehicle-injected animals from the 4 groups did not differ at 1 d or 3 d. At 7 d, vehicle-injectedanimals in the GI group were heavier than mice in the other 3 groups (F(3,31) = 7.3, P<0.01).There were no significant differences in body weights among the tumor cell-injected animals fromthe four experimental groups on any of the termination days (Table 1). Since body weights of thetumor cell-injected animals do not differ, we conclude that neither body weights nor nutritionalvariables affect tumor growth in this model.767 7Table 1. Body Weights of Tumor Cell- and Vehicle-Injected Mice From the 4 ExperimentalHousing Groups 1, 3 and 7 d Post Injection. Male mice were injected with either 2 x 106 SC 115cells or vehicle. Mice were weighed and immediately terminated at 1, 3 and 7 d post injection.The body weights of vehicle-injected animals from the 4 groups did not differ at 1 d or 3 d. At 7d, vehicle-injected animals in the GI group were heavier than mice in the other 3 groups (p<0.05).There were no significant differences in body weights among the tumor cell-injected animals fromthe four experimental groups on any of the termination days.Table 1.Body Weight (g)GG GI^II IG1 Day Tumor 36.1 1 37.2 37.6 35.9± 0.98 ± 0.98 ± 0.95 ± 0.701 Day Vehicle 36.6 36.5 38.9 37.4± 1.35 ± 0.83 ± 1.63 ± 0.993 Day Tumor 33.2 32.6 31.8 34.5± 0.68 ± 1.96 ± 2.32 + 0.963 Day Vehicle 34.5 36.4 34.8 32.8± 0.72 ± 0.88 ± 1.55 ± 0.727 Day Tumor 36.6 34.1 35.8 34.2±0.88 ±1.15 ±1.62 ±1.007 Day Vehicle 37.3 42.8 36.9 35.0± 1.24 + 1.44 ± 0.77 ± 0.851. Mean ± sem.787 9Spleen Weights: Spleen weights were adjusted for the body weight of the animal (relativespleen weight). Overall, there were significant effects of Tumor, F(1,195) = 19.86, P<0.001,Group, F(3,195) = 5.75, P<0.001 and Day, F(2,195) = 22.09, P<0.001. Tumor cell-injected micehad significantly larger relative spleen weights than those of vehicle-injected mice (p<0.001)(Table 2). Separate analyses of relative spleen weights in tumor cell- and vehicle-injected animalsrevealed that, in tumor cell-injected mice, relative spleen weights increased significantly over days(1 d < 3 d < 7 d, p<0.001). In addition, there was a significant Day x Group interaction (F(6,102)= 3.326, p<0.01). At 1 d post injection, mice of the GI group had significantly larger relativespleen weights than mice of the IG group (p<0.05, Figure 7). At 3 d, there were no significantdifferences in relative spleen weights among groups, whereas by 7 d, mice of the IG group hadsignificantly larger relative spleen weights than did mice of the II group (p<0.05) (Figure 7). Incontrast, in vehicle-injected mice, there were significant effects of Group and no Group x Daysinteraction. Overall, relative spleen weights were greater at 3 d than at 1 d (p<0.02), but relativespleen weights at 7 d post injection were not significantly different from spleens of mice at either1 d or 3 d post injection (Table 2).Splenic NK Cell Activity at I d Post Injection: At 1 d post injection and group formation,the ANOVA revealed significant main effects of Group (F(3,64) = 15.16, P<0.001), Tumor(F(1,64) = 12.05, P<0.001), and Ratio (F(5,320) = 821.98, P<0.001), as well as a significantGroup x Tumor x Ratio interaction (F(15,320) = 3.94, P<0.001) (Figure 8). Overall, tumor cell-injected animals showed decreased splenic NK cell activity compared with vehicle-injectedcontrols. Importantly, post-hoc analysis of the Group x Tumor x Ratio interaction revealed thatfor both tumor- and vehicle-injected conditions, GI animals generally had the greatest and IGanimals generally had the least NK cell activity (p<0.05). In addition, it is interesting that in thevehicle-injected condition, II animals were similar to GI animals in their high levels of NK cellactivity, whereas in the tumor-injected condition II animals had significantly lower NK cell activity80Table 2. Spleen Weights of Tumor Cell- and Vehicle-Injected Mice From the 4 ExperimentalHousing Groups 1, 3 and 7 d Post Injection. Methods were as described in table 1. Spleenswere removed immediately following termination of the mice. In tumor cell-injected mice,relative spleen weights increased significantly over days (1 d < 3 d < 7 d, p<0.001). In vehicle-injected mice, relative spleen weights were greater at 3 d than at 1 d (p<0.02), but relative spleenweights at 7 d post injection were not significantly different from spleens of mice at either 1 d or 3d post injection.Table 2Tumour Injected Vehicle InjectedSpleen Wt.(g)Spleen Wt. Spleen Wt.(g)Spleen Wt.Body Wtx100(g)Body Wtx100(g)1 DAY GO 0.133 1 0.371 0.149 0.392± 0.008 ± 0.029 ± 0.014 ± 0.045GI .0148 0.396 0.150 0.364±0.009 ±0.021 ±0.018 ±0.030II 0.117 0.313 0.125 0.322± 0.003 ± 0.010 ± 0.006 ± 0.005IG 0.108 0.302 0.122 0.327± 0.006 ± 0.017 ± 0.005 ± 0.0133 DAY GO 0.168 0.506 0.181 0.526± 0.010 ± 0.027 ± 0.021 ± 0.060GI 0.172 0.536 0.170 0.476±0.017 ±0.058 ±0..041 ±0.128II 0.140 0.443 0.123 0.353±0.008 0.020 ±0.012 ±0.031IG 0.152 0.442 0.142 0.432± 0.006 ± 0.014 ± 0.007 ± 0.0217 Day GO 0.260 0.711 0.142 0.385±0.019 ±0.50 ±0.014 ±0.044GI 0.186 0.540 0.146 0.344±0.023 ±0.054 ±0.020 ±0.043II 0.170 0.483 0.108 0.292±0.009 ±0.041 ±0.006 ±0.014IG 0.238 0.713 0.183 0.514± 0.015 ± 0.062 ± 0.020 ± 0.047811 Mean ± sem82Figure 7 . Relative Spleen Weights of Tumor Cell-Injected Mice From the 4 ExperimentalGroups in the First Wk Post Injection. Points represent mean ± SEM. Mice from the 4experimental groups (as described in figure 2) were injected with 2 x 106 SC1 15 cells, terminated1, 3 or 7 d later. At 1 d post injection, mice of the GI group had significantly larger relativespleen weights (spleen weight/body weight) than mice of the IG group (p<0.05). At 3 d, therewere no significant differences in relative spleen weights among groups, whereas by 7 d, mice ofthe IG group had significantly larger relative spleen weights than did mice of the II group(p<0.05). Points represent the mean of "n" mice ± sem. (n's: 1 d: GG(9); GI(6); II(6); IG(1 1);3 d: GG(9); GI(5); II(6); IG(1 5); 7 d: GG(7); GI(7); II(6); IG(13)).0.81.00.00 1^2^3^4 5^6^7^883Days Post Injection8 4Figure 8. Lytic Activity of Splenic Natural Killer (NK) Cells I d Post Injection. Points representmean ± SEM. Mice from the 4 experimental groups (as described in figure 2) were either injectedwith 2x106 SC 115 cells (open symbols) or vehicle (solid symbols). Twenty-four hours later, thelytic activity of splenic NK cells was measured. Within the tumor injected condition, GI and GG> IG and II, p's<0.05. Within the vehicle injected condition, GI = II > GO = IG, p's<0.05. (n's:GG-T(7), GG-V(11); GI-T(7), GI-V(9); II-1(6), II-V(7); IG-T(13), IG-V(12)). Spontaneousrelease was usually less than 20% of the total release and always less than 30%.5:1^9:1^18:1^37:1^75:1 150:11008020085Effector:Target ratio8 6than GI animals (p<0.05). Further, data on NK cell activity at 1 d were compared with the datapreviously obtained at 3 d and 7 d post injection. Collapsed across the group condition, vehicle-injected mice at 1 d post injection and group formation did not have significantly different levelsof splenic NK cell activity than their vehicle-injected counterparts at 3 d and 7 d (F(2,96) = 0.211,P>0.8). In contrast, tumor cell-injected mice have significantly lower levels of splenic NK cellactivity than did tumor cell-injected mice at 3 d and at 7 d (F(2,95) = 322.558, P<0.001).Anti-asialo GM1 Treatment: When spleen cells from mice injected with tumor cells 3 dprior to termination were treated with anti-Asialo GM1 and complement prior to the NK cellassay, there was a greater than 80 % decrease in NK cell activity in all cases compared to spleencells from the same animals that were not treated with anti-Asialo GM1 and complement (data notshown).DISCUSSIONThis is the first study to demonstrate that the SC 115 tumor is capable of stimulatingsplenic NK cell activity. Stimulation of NK cell activity followed a definite time course. Activitywas suppressed 1 d post tumor cell-injection, was maximally stimulated at 3 d and had begun todecline by 7 d. Anti-Asialo GM1 treatment of stimulated spleen cells resulted in a greater than 80% decrease in activity. This indicated that the main effector cell in this assay was the NK cell,although it is possible that allospecific cytotoxic T lymphocytes were involved to a small degree.Inducers of NK cell activity such as interferon maximally stimulate NK cell activity 18 -24h following their injection (Ortaldo eta!., 1989, Talmadge eta!., 1985). In the present study, wefound that maximal stimulation of NK cell activity occurred at 3 d post tumor cell-injection. Thisdelay in NK cell stimulation would be expected since the immune system must first recognize thetumor and respond with the production of lympholcines before NK cell stimulation can occur.87This phenomenon has been demonstrated in other forms of immune stimulation. For example, ithas been shown that augmentation of NK cell activity is maximal 3 d following an injection ofbacterial cells (Klein and Kearns, 1989, Savary and Lotzova, 1987). The decline in NK cellactivity seen in tumor cell-injected animals at one week may be due to tumor mediated-suppression of NK cell activity. The suppression of NK cells by tumors is a well documentedphenomenon (Fulton, 1987, Savary and Lotzova, 1987, Wei and Heppner, 1987).We have previously shown (Emerman and Siemiatkowski, 1984a) that SC 115 tumor cellsinjected into male mice raised under our standard laboratory conditions (groups of 3, GG)produce a palpable tumor within 6-8 d, which grows to a mass of 2-3 g in approximately 3 wk.Moreover, growth rate of the SC 115 tumor may be increased or decreased by housing mice underdifferent conditions (Weinberg and Emerman, 1989). Tumor growth rate is increased in GI and IImice and decreased in IG mice compared to that seen in mice in the standard GG group. In thepresent study, we demonstrate an effect of differential housing on NK cell activity in mice in the 4experimental groups. At 1 d post-injection, GI animals generally had the greatest NK cell activitywhereas IG animals had the lowest NK cell activity in both tumor and vehicle-injected conditions.At 3 d, when NK cell activity was markedly stimulated in tumor cell-injected mice, GI animals inthe tumor injected condition again had the greatest NK cell activity, whereas in the vehicle-injected condition, the groups did not differ from each other. Finally, at 7 d, although NK cellactivity was still stimulated to some extent in tumor cell-injected mice, mice in the 4 experimentalgroups did not differ from each other.We hypothesized that GI and II animals, which have the largest tumors, would havedecreased NK cell activity, and that IG animals, which have the smallest tumors, would haveincreased NK cell activity. Surprisingly, we found that at 3 d post injection, when NK cellstimulation was the greatest, GI animals had greater NK cell activity than IG animals. Further GIand II animals did not have similar NK cell activity at this time.8 8One explanation for these data is that NK cell activity in the spleen may not accuratelyreflect the activity of NK cells at the tumor site. For example, experimental findings demonstratethat the recruitment of lymphocytes, including NK cells, from the circulation (Greenberg et al.,1986, Migliori eta!., 1987, Zangemeister-Wittke eta!., 1989), and the ability of NK cells toinfiltrate solid tumor targets (Jaaskelainen et al., 1989, Wei and Heppner, 1987) may be criticalfactors in regulating tumor growth. Thus, it is possible that NK cells of mice in the IG group arebetter able to remain localized at the tumor site than are NK cells from mice of the GI group.This could result in mice of the IG group having higher levels of tumor-infiltrating NK cells andlower levels of splenic NK cells than mice in the GI group.A second possibility is that NK cell activity may not be a significant regulator of SC 115tumor growth. This view is supported by the findings of Kitamura (1980), who demonstrated thatthe original SC 115 tumor could grow in female nude mice (who lack T lymphocytes but have highlevels of NK cell activity) but not in female DD/S mice. The authors concluded that Tlymphocytes are an important mediator of SC 115 cell growth. However, the class 1 subline of theSC 115 tumor used in this experiment differs from the original SC 115 cell line in that tumors dogrow in female DD/S mice. It is possible that this change could be the result of a modification ofthe immunogenicity of the SC115 tumor, such as down regulation of class I WIC expression.Decreased expression of the class I Mit-IC molecules by tumor cells has been demonstrated todecrease the ability of CTL to lyse some tumors while increasing the susceptibility of the cells toNK cell-mediated lysis (Ljunggren and Kane, 1985).Finally, a third possibility is that NK cells may positively modulate the growth of theSC 115 tumor. It is possible that the growth of the SC 115 tumor is stimulated by hormones orcytokines released by the activated NK cells at the site of the tumor. This possibility is supportedby a study which demonstrated that increased NK cell activity in preneoplastic alveolar nodules in89the breast of female mice stimulated the development of cancerous outgrowths from the nodules(Wei and Heppner, 1989). The difference in NK cell activity between the II and GI groups, bothof which develop large tumors, may be due to the factors discussed above or could indicate thatdifferent mechanisms may be regulating tumor growth in these two groups of mice.Previously, it was demonstrated that at 3 wk post injection, tumor cell-injected mice hadsignificantly enlarged spleens compared with their vehicle-injected counterparts (Weinberg andEmerman, 1989). However, no significant differences in spleen weights among mice in the 4experimental groups were observed at that time. The present study measured spleen weights at 1,3 and 7 d following injection. Overall, spleen weight was lowest in both tumor cell- and vehicle-injected animals at 1 d post injection. At 3 d post injection, there was a significant increase inspleen weight in both tumor cell- and vehicle-injected animals. The spleen weights of tumor cell-injected animals continued to increase and were significantly greater at 7 d than at both 1 and 3 d.In the vehicle-injected condition on the other hand, spleen weights at 7 d were intermediate tothose observed at 1 and 3 d. These results suggest that several different processes may beaffecting spleen weight. First, the finding that both tumor cell- and vehicle-injected mice exhibitenlarged spleens at 3 d post injection and group formation suggests that a non-specific process isoperating at this time. It is possible that exposure to the acute daily novelty stress regimen or thestress/trauma associated with the injection process may be responsible for the increase in spleenweights at 3 d. At 7 d post injection and group formation, spleen weights continue to increase intumor cell-injected mice, but not in vehicle-injected mice. This suggests that the presence of atumor may induce an increase in spleen weights which becomes apparent at this time. These datasupport the previous finding that the presence of a tumor increases spleen weights in mice at 3 wkpost injection (Weinberg and Emerman, 1984).The observed increase in spleen weights of tumor cell-injected mice at 7 d does notcorrelate with the changes in NK cell activity in these mice as splenic NK cell activity is declining90while spleen weights are increasing at this time. It is possible that the increased spleen weights at7 d in tumor cell-injected mice reflects an increase in the number of cytotoxic T lymphocytes.Alternatively, the increase in spleen weights could reflect a tumor-induced increase in the numberof erythoblasts in the spleen as, in mice, the spleen is capable of erythyropoiesis (Milas and Scott,1978). An increase in erythropoiesis could correlate with the gradual suppression of splenic NKcell activity as erythroblasts have been shown to exert a suppressive effect on NK cell activity(Savary and Lotzova, 1987). Histological studies of spleen cell composition and perhaps a FACS(flourescence activated cell sorter) analysis of the relative composition of cell types in spleens oftumor cell- and vehicle-injected animals would distinguish between these possibilities.In summary, this study demonstrates that the SC 115 tumor is capable of stimulating NKcell activity. Further, differences in social housing condition alter the degree of NK cell activationproduced by tumor cell-injection. However, it is not yet clear if modulation of NK cell activityhas an effect on the growth rate of this tumor. An increase in the spleen weight in tumor cell-injected mice, but not in vehicle-injected mice was observed at 7 d post injection. It is not clear atthis time if this increase in spleen weights reflects an increase in lymphoid cell numbers or anincrease in erythropoeisis in these spleens. We conclude that either NK cell activity in the spleendoes not reflect the activity at the tumor site or NK cells are not an important mediator of thedifferential tumor growth observed in this model, or that NK cells may positively modulate thegrowth rate of SC 115 cells.91B. Natural Killer Cell Activity At The Tumor Site.IntroductionIn the previous study, it was demonstrated that the presence of the SC 115 tumor caused asignificant stimulation of splenic NK cell activity. However, the relevance of this finding to themodulation of the differential tumor growth rates observed in mice in the experimental housinggroups of our model is unclear. Thus, the present study was undertaken to investigate thehypothesis that NK cell activity of tumor-infiltrating lymphocytes differs among animals of theexperimental housing groups and that these differences are involved in mediating the differentialtumor growth rates observed in this model.The previous study demonstrated that tumor-induced stimulation of NK cell activity in thespleen was greatest at 3 d post tumor cell-injection and that, as early as 7 d post tumor cell-injection, the tumor-induced stimulation of NK cell activity had declined significantly from thatseen at 3 d. These data suggested that the optimum time to investigate NK cell activity of tumor-infiltrating lymphocytes is in the first wk post injection. This presents a problem for studyingtumor-infiltrating lymphocyte activity as, at 7 d, the tumor is barely palpable. It would beimpossible to collect sufficient amounts of tumor tissue to isolate the required numbers oflymphocytes to conduct the study. To circumvent this problem, the sponge allograph modeldeveloped by Roberts and Hayry (1976) was adapted for use in the SC 115 model. The spongeallograft model involves implanting a small polyurethane sponge in the peritoneal cavity of amouse. After several days, the sponge becomes infiltrated with host connective tissue andimmune cells. The sponge is then removed and implanted subcutaneously on the dorsal surface ofan allogeneic mouse. The sponge can be removed at different time points and the immune cellsinfiltrating this allogeneic graft can be isolated by gently squeezing the sponge. Using this92procedure, Hoffman (1988) demonstrated that NK cells infiltrate the sponge matrix if the matrixcontains allogeneic cells but not if it contains syngeneic cells. Roberts and Hayry (1976) alsodemonstrated that tumor cells are capable of growing in the sponge matrix of subcutaneouslyimplanted sponges. Thus, the sponge is capable of supporting tumor growth and allowing theinfiltration of immune cells, yet the sponge itself is not immunogenic. The present study wasdesigned to investigate the NK cell activity of lymphocytes infiltrating sponges injected either withtumor cells or with vehicle alone. Mice from only GI (who develop the largest tumors) and IG(who develop the smallest tumors) were tested in this study.Materials and MethodsAnimals: One hundred fifty-four young adult DD/S mice (2-4 months of age) were used in thisstudy. In experiments designed to develop and test the sponge model, mice were housedindividually for the course of the experiments and were not exposed to the acute daily noveltystress regimen. In the experiment to investigate the effects of differential housing on NK cellactivity of lymphocytes infiltrating tumor cell- and vehicle-injected sponges, mice from the twoexperimental housing groups (GI and IG) were reared and housed as described previously in theGeneral Methods, were used and these mice were subjected to acute daily novelty stress.Tumor Cells:SC 115 Cells: All SC 115 cells used in this study were taken from frozen stocks aspreviously described in General Methods. For each experiment in the time course study of NKcell activity in mice from the GI and IG groups, 2 additional mice were injected with tumor cellss.c. as is typically done for tumor propagation, and monitored for 21 d to ensure that the cellswere viable. All tumors grew as expected in these control mice.Yac 1 Cells: These cells were used in the NK cell assays as previously described.Polyurethane Sponges:Preparation: Sponges (1 x 1 x 1.5 cm) were cut from a polyurethane foam block andsterilized as previously described (Hoffman et al., 1988). Briefly, sponges were washed 3 timeswith distilled water (1 I/wash), taking care to remove the air from the sponges. The sponges werethen boiled 3 times in distilled water (1 I/boiling), soaked for 30 min in approximately 100 ml ofacetone, soaked for 30 min in 95% Et0H (500 ml), washed 3 times with distilled water (1 I/wash)and finally boiled 3 times in distilled water (1 I/boiling). The sponges were then carefully sealed inaluminum foil so as not to compress the sponge and autoclaved.Implantation: The sponges were surgically implanted into mice anesthetized withhalothane (Ayerst Laboratories Inc., Montreal, Que) - oxygen mixture (Medigas Inc., Vancouver,B.C.). Briefly, a mouse's back was swabbed well with 70% ethanol to sterilize it. A 1 cm longcut was made in the skin with sterile scissors and a pocket was made under the skin by bluntdissection using the scissors. The sponge was wet in sterile phosphate buffered saline,compressed using sterile forceps and inserted into the subcutaneous pocket. The wound wasclosed with three sutures (Ethicon Inc., Somerville, NJ). The animals were allowed 5 d torecover and for the sponge to vascularize before the next experimental procedure was initiated.Tumor/Vehicle Injection: An animal was lightly anesthetized with ether for injection. Thesponge was gently compressed and the needle inserted into the sponge. As the 0.1 ml of fluid(tumor cell suspension or vehicle) was injected, the sponge was gently released.Histology of 5C115 Tumor Growth in Polyurethane Sponge Matrices: Sponges were implantedinto 2 male and 2 female DD/S mice. After 5 d, sponges in the male mice were injected withSC 115 cells (2 x 106 cells); mice were terminated at 19 d post injection. Four additional male939 4mice were injected with tumor cells in the usual fashion as controls for tumor growth. The femalemice were not injected and were terminated at 21 d post sponge-implantation. The sponges wereremoved, cut in half and fixed in 10% formalin for 1 wk. The sponges were then paraffin-embedded, sectioned and stained with hematoxylin and eosin for microscopy.Time Course Study of White Blood Cells Infiltrating Polyurethane Sponges: Twelve male mice(raised in groups of 3) were implanted with sponges and individually housed. After 5 d, spongeswere injected with either tumor cells (n=6) or vehicle (n=6). Mice were terminated at 1, 3 or 7 dpost injection and the sponges removed. Sponge-infiltrating lymphocytes were isolated by cuttinga sponge in half and gently squeezing the contents of the sponge into 3 changes of RPM"(approximately 35 mls in total) in a 60 mm petii dish. The lymphocytes were centrifuged (600 xg, 2 min), the supernatant decanted and the cells resuspended in 1 ml RMPI. Several 100 glaliquots were spun onto glass slides using a cytospin apparatus (Shanndon) at 120 x g for 1 min.The slides were stained with a Wright's stain and differential counts were made by counting 200cells from each of 2 slides for each animal. The relative proportions of the different white bloodcells in the sponge infiltrates of each animal were thus determined.Pilot Study of the Time Course of NK Cell Activity in Sponge -Infiltrating Lymphocytes: In aninitial experiment, 4 male mice were implanted with sponges and 5 d later were injected witheither tumor cells (n=2) or vehicle (n=2). All mice were terminated 3 d post injection and sponge-infiltrating cells isolated as previously described. This time point was chosen for termination as 3d was when maximal stimulation of splenic NK cell activity was observed in a previous study.The cells isolated from the sponges were used in the NK cell assay previously described with theexception that, due to the low numbers of cells isolated from each sponge, effector to target cellratios of 15:1 and lower were used in this study. The cells from each mouse were processed andused in the NK cell assay separately.9 5In a second experiment, 12 male mice were implanted with sponges and the spongesinjected with SC 115 tumor cells 5 d later. Due to the low cell yields observed in the first study,sponges from 2 mice were pooled for each data point. Mice were terminated at 1, 3 and 7 d postinjection (4 mice/day) and the NK cell activity of their tumor-infiltrating cells was determined.In a third experiment, 12 mice were implanted with sponges and 5 d later sponges wereinjected with either SC 115 tumor cells (n=8) or vehicle alone (n=4). Half the mice in eachcondition were terminated at 10 d post injection and half at 17 d post injection. The lymphocytesfrom the sponges of 2 mice in the same condition were pooled for each data point and NK cellactivity was measured.NK Cell Activity of Sponge-Infiltrating Lymphocytes at 3 d and 7 din Mice from theExperimental Groups: Ninty-six male mice were implanted with sponges and 5 d later spongeswere injected with either tumor cells (n=26) or vehicle (19). The mice were then rehoused aseither GI (T, n=10, V. n=7) or IG (T, n=16, V, n=12) groups and subjected to the acute dailynovelty stress regimen as described previously in General Methods. The mice were terminated at3 d or 7 d post injection and the sponges were removed. Sponge-infiltrating lymphocytes from 2mice of the same housing group were combined for a single data point in assays of NK cellactivity.ResultsHistology of SCI 15 Tumor Growth in Polyurethane Sponge Matrices.Both tumor-injected and noninjected sponges became invested with a thick connectivetissue capsule, not unlike that observed surrounding the SC 115 tumor, and were firmly attached96to the overlying skin. Microscopic examination of the sponges revealed that both tumor cell-injected and noninjected sponges contained many small blood vessels. In the tumor-injectedsponges, colonies of tumor cells were observed growing within the matrix of the sponge. Suchcolonies were not observed in the noninjected control sponges. Although tumor growth wasobserved in the sponges, it appeared to be delayed compared to the growth of tumors in animalsinjected at the same time by the usual subcutaneous procedure. At 19 d, tumors growing in thesponges were small foci of cells whereas, in the s.c. injected control mice, tumors were all largerthan 1 g. In addition, noninjected sponges exhibited a mild granulocytic invasion which was notobserved in the tumor cell-injected sponges.Time Course Study of White Blood Cells Infiltrating Polyurethane Sponges.Analysis of cytospin slides of cells harvested from sponges at 1, 3 and 7 d revealed that onall 3 test days, both tumor cell- and vehicle-injected sponges contained notable numbers ofneutrophils, monocytes, lymphocytes and eosinophils (Figure 9). However, there were markeddifferences across days in the relative percentages of these cell types in tumor cell-injectedsponges. The relative proportions of different white blood cells were similar in vehicle-injectedsponges at all time points. They consisted of approximately 60 % neutrophils, 25 monocyte,15 % lymphocytes and a small percentage of eosinophils . A similar distribution of white bloodcells was observed in the infiltrate of tumor cell-injected sponges at 1 and 3 d post injection.Importantly, at 7 d post-injection, there was a marked change in the relative proportions ofdifferent white blood cells infiltrating tumor cell-injected sponges. There was a decline in therelative numbers of neutrophils in the infiltrate and a corresponding increase in the relativenumbers of lymphocytes and monocytes compared with that seen at 1 and 3 d. As well, noeosinophils were observed in the cell infiltrates of tumor cell-injected sponges at 7 d, whereas theywere present in low numbers in all other groups. Although this was a pilot study and the groupsize was small (2 mice/group, 2 slides/animal), T-tests were conducted on this data. Analysis9 7Figure 9. Time Course Study of White Blood Cells Infiltrating Polyurethane Sponges. Micewere implanted s.c. with polyurethane sponges (1 x 1 x 1.5 cm); 5 d later sponges were injectedwith either tumor cells (T)(2 x 106) or vehicle (V), and mice terminated at 1, 3 and 7 d postinjection. White blood cells infiltrating the sponges were isolated, stained with Wright's stain andcounted. In infiltrates from vehicle-injected sponges, no differences in the relative numbers ofwhite blood cell types was observed on 1, 3 and 7 d post injection. In infiltrates from tumor cell-injected sponges, the relative numbers of white blood cell types on d 1 and 3 were similar to thosein vehicle-injected sponges. At 7 d post-injection, there was a marked decline in the relativenumbers of neutrophils in the infiltrate and a corresponding increase in the relative numbers oflymphocytes and monocytes compared with that at 1 d and 3 d (p's<0.01). No eosinophils wereobserved in the infiltrates of tumor cell-injected sponges at 7 d, whereas they were present in lownumbers in all other sponges. (n = 2 mice/condition, 2 x 200 cells counted/mouse).98100753(1)^50a)4-)^25ld Tld NT3d T3d NT7d T^I 7d NT1D 30 70Neutrophil Monocyte Lymphocyte EosinophilCell Type9 9revealed that the white blood cells infiltrating tumor cell-injected sponges at 7 d post injectiondiffered significantly from both that of both tumor cell-injected sponges at 1 and 3 d post injectionand from vehicle-injected sponges at all 3 d measured (p<0.01). It is understood that, due to thesmall group size, the power of this analysis is low. However, the results clearly support theconclusion that there is a significant change in the composition of the sponge-infiltrating whiteblood cells in tumor cell-injected sponges at 7 d post injection relative to those of tumor cell-injected sponges at 1 and 3 d and of vehicle-injected sponges at all 3 days.Pilot Study of the Time Course of NK Cell Activity in Sponge -Infiltrating Lymphocytes.There were marked differences in levels of NK cell activity of sponge-infiltratinglymphocytes of tumor cell- and vehicle-injected mice (Figure 10). However, due to the smallnumbers of mice per group in this pilot study, no statistical analysis of the data was possible.Consistent with NK cell activity observed in the spleen, NK cell activity of sponge-infiltratinglymphocytes did not differ in tumor cell- and vehicle-injected mice at 1 d post injection.However, by 3 d post injection, lymphocytes infiltrating tumor cell-injected sponges showed anincrease in NK cell activity compared with that of vehicle-injected mice. This increase wasmaintained through 7 d post tumor cell-injection. Although NK cell activity began to decline after7 d post tumor cell-injection, it was still elevated in lymphocytes infiltrating tumor cell-injectedsponges compared with that in their vehicle-injected counterparts at 10 d and 17 d post injection.Interestingly, the tumor-induced stimulation of NK cell activity at 3 d post injection observed insponge-infiltrating lymphocytes was similar to that observed previously in splenic lymphocytes.1 0 0Figure 10. Pilot Study of the Time Course of NK Cell Activity in Sponge -InfiltratingLymphocytes. Lymphocytes infiltrating tumor cell- and vehicle-injected sponges were isolated asdescribed in figure 9, mice were terminated at 1, 3, 7, 10 and 17 d post injection. NK cell activityof vehicle-injected mice did not differ across days. At 1 d post injection, NK cell activity (effectorto taget cell (E:T) ratio of 15 to 1) of lymphocytes infiltrating tumor cell-injected sponges did notdiffer from that of vehicle-injected sponges. By 3 d, lymphocytes infiltrating tumor cell-injectedsponges showed an increase in NK cell activity compared with that of vehicle-injected spongesand, this increase was maintained through 7 d. Although NK cell activity began to decline after 7d, it was still elevated in lymphocytes infiltrating tumor cell-injected sponges compared with thatin their vehicle-injected counterparts at 10 d and 17 d post injection. (N's: 1 d, T(2); 3 d, T(4),V(2); 7 d, T(2); 10 d, T(2), V(1); 17 d, T(2), V(1)).o 5^10 15 2070Tumor Injected =Vehicle Injected =60o10101Days Post InjectionNK Cell Activity in Sponge-Infiltrating Lymphocytes at 3 d and 7 din Mice from theExperimental Groups.This experiment investigated NK cell activity of sponge-infiltrating lymphocytes in micefrom the GI group (who develop the largest tumors) and IG group (who develop the smallesttumors) following either tumor cell- or vehicle-injection. Based on the findings of the previouspilot study, NK cell activity of sponge-infiltrating lymphocytes was assayed at 3 d and 7 d postinjection and group formation when NK cell stimulation in tumor cell-injected mice was likely tobe the greatest. At 3 d post-injection (Figure 11), ANOVAs revealed that there was a significanteffect of Group (F(1,20) = 7.548, P<0.02). Overall, NK cell activity in sponge-infiltratinglymphocytes was significantly greater in mice of the GI group than in mice of the IG group(p<0.05). However, it is clear from Figure 11 that the overall group effect was due primarily tothe fact that Tumor cell-injected mice of the GI group showed greater NK cell stimulation than allother mice. However, the main effect of Tumor was not significant. By 7 d post-injection (Figure12), the effect of Tumor was significant (F(1,17) = 29.293, P<0.001) and a Group x Tumorinteraction was found (F(1,17) = 5.852, P<0.03). Post-hoc analysis revealed that, overall, levelsof NK cell activity in sponge-infiltrating lymphocytes were significantly greater in tumor cell-injected mice than in vehicle-injected mice at 7 d (p<0.005). Also, at this time, tumor cell-injectedmice of the GI group had significantly greater levels of NK cell activity than did mice in all othergroups (p<0.05). In addition, tumor cell-injected mice of the IG group had significantly greaterNK cell activity than vehicle-injected mice of both the GI and IG groups (p<0.05).DiscussionThis study demonstrates three important findings. First, the study established a procedurefor utilizing the polyurethane sponge model to investigate tumor-infiltrating lymphocytes in theSC 115 mouse mammary tumor. It was demonstrated that polyurethane sponges implanted s.c. in1 021 03Figure 11. NK Cell Activity in Sponge-Infiltrating Lymphocytes at 3 d Post Injection in Micefrom the Experimental Groups. GI and IG mice (as described in figure 1) were implanted withpolyurethane sponges and injected with either tumor cells (2 x 106) or vehicle as described infigure 9. Mice were terminated at 3 d post injection and NK cell activity of lymphocytesinfiltrating tumor cell- and vehicle-injected sponges was determined. Points represent mean ±SEM. Overall, NK cell activity in sponge-infiltrating lymphocytes was significantly greater inmice of the GI group than in mice of the IG group (p<0.05). The overall group effect appears tobe due primarily to the greater NK cell stimulation in GI-T mice than in mice of all groups(p<0.05). (N's: GI, T(5), V(4); IG, T(8), V(7)). Spontaneous release was usually less than 15%of the total release and always less than 25%.100GI TumorGI VehicleIG TumorIG VehicleInjected =Injected =Injected =Injected =20o1 040.5:1^0.9:1^1.9:1^3.8:1^7.5:1^15:1Effector to Target Ratio1 0 5Figure 12. NK Cell Activity in Sponge-Infiltrating Lymphocytes at 7 d Post Injection in Micefrom the Experimental Groups. GI and IG mice (as described in figure 1) were implanted withpolyurethane sponges and injected with either tumor cells (2 x 106) or vehicle as described infigure 9. Mice were terminated at 7 d post injection and NK cell activity of lymphocytesinfiltrating tumor cell- and vehicle-injected sponges was determined. Points represent mean ±SEM. At 7 d, NK cell activity in sponge-infiltrating lymphocytes was significantly greater intumor cell-injected sponges than in vehicle-injected sponges (p<0.005). Also, at this time, tumorcell-injected mice of the GI group had significantly greater levels of NK cell activity than did micein all other groups (p's<0.05). In addition, tumor cell-injected mice of the IG group hadsignificantly greater NK cell activity than vehicle-injected mice of both the GI and IG groups(p's<0.05). (N's: GI, T(5), V(3); IG, T(8), V(5)). Spontaneous release was usually less than 15%of the total release and always less than 20%.100GI Tumor Injected =GI Vehicle Injected =IG Tumor Injected =IG Vehicle Injected =o1 060.5:1^0.9:1^1.9:1^3.8:1^7.5:1^15:1Effector to Target Ratio1 07DD/S mice could support the growth of the SC 115 tumor and that it was possible to isolatelymphocytes infiltrating the sponge. Secondly, we have shown that the SC 115 tumor significantlystimulates the activity of NK cells at the tumor site. Finally, it was demonstrated that mice of theGI group had significantly greater levels of NK cell activity at the tumor site than did mice of theIG group.This study utilized polyurethane sponges implanted s.c. in the mouse as a solid supportmatrix in which tumor cells could grow. This facilitated the isolation of immune cells infiltratingthe tumor at early time points post tumor cell-injection. To our knowledge, this is the first studyto utilize this model for the study of the SC 115 tumor, although it has previously has been used inthe study of several other tumor systems (Roberts and Hayry, 1976, Zangemeister-Wittke et al.,1989). It was demonstrated that the SC 115 tumor is capable of growing in the sponge matrix andthat blood vessels infiltrate the sponge to nourish the developing tumor. Tumor growth in thesponge was delayed compared with the tumor growth observed in control mice injected s.c. (thestandard injection protocol).White blood cells were isolated from both tumor cell- and vehicle-injected sponges. Cellsfrom vehicle-injected sponges demonstrated a white blood cell infiltrate indicative of a mildgranulocytic response. This is similar to previous reports that the presence of the polyurethanesponge itself does not induce significant immune reactivity (Hoffinan at al, 1988, Roberts andHayry, 1976). However, in tumor cell-injected sponges, there was an increase in the relativenumbers of monocytes and lymphocytes and a corresponding decrease in the number ofneutrophils at 7 d post injection compared with those at 1 d and 3 d post injection as well as withall vehicle-injected sponges. This change in relative cell numbers suggests that a specific immuneresponse might be occurring in the tumor cell-injected sponges at 7 d. A similar shift from agranulocytic infiltrate to a monocytic infiltrate has previously been reported to occur at 8-11 dpost implantation of sponges impregnated with allogeneic cells (Hoffman eta!., 1988). The1 08conclusion that the SC 115 tumor may be immunogenic is also supported by the finding that whiteblood cells from tumor cell-injected sponges at 7 d post-injection completely lacked the small butpersistent component of eosinophils found in all other sponges. It has been suggested thateosinophils are stimulated by the cytokines IL-4 and IL-5, which are primarily associated with ahumoral immune response, and that such a response may act to inhibit the stimulation of CTL andNK cells by TH1 cells (Street and Mosman, 1991). Although this finding that the SC 115 tumormay be immunogenic is not conclusive, it is supported by the demonstration that priorimmunopotentiation of mice by the injection of heat-killed Staphylococcus aureus virus decreasesSC115 tumor growth (Nohno eta!., 1986) and that the presence of the SC115 tumor stimulatessplenic NK cell activity (Rowse eta!., 1990).Importantly, this study demonstrates that the presence of the SC 115 tumor was capable ofstimulating NK cell activity, not only in the spleen, but also in tumor/sponge-infiltratinglymphocytes. At 3 d post tumor cell-injection, the level of NK cell activity at the tumor site wassimilar to that observed in the spleen at 3 d. Interestingly, NK cell activity remained stimulatedfor a longer period of time at the tumor site than at the spleen. Whereas NK cell activity in thespleen had declined significantly by 7 d post tumor cell-injection, NK cell activity at the tumor sitedid not decrease at 7 d post tumor cell-injection and at, 17 d post injection, NK cell activity oftumor-infiltrating lymphocytes was still greater than that of lymphocytes infiltrating vehicle-injected sponges. These data are consistent with the report that in the allogeneic sponge modellymphocytes infiltrating a sponge impregnated with allogeneic peritoneal cells demonstrate asignificant increase in NK cell activity at 5-11 d (Hoffman et al., 1988).There was a significant difference in the level of NK cell activity in tumor-infiltratinglymphocytes between mice in the GI and IG groups. At both 3 d and 7 d post injection, mice ofthe GI group, who would develop the largest tumors, had greater NK cell activity than mice ofthe IG group, who would develop the smallest tumors. A similar result was observed in the1 09spleens of tumor cell-injected mice at 3 d post injection. Thus, tumor-induced stimulation of NKcell activity is greater in mice of the GI group than in mice of the IG group both at the spleen andat the tumor site. This finding is contrary to what one would expect if NK cells are involved indecreasing the tumor growth observed in this model.There are a number of possible explanation for this apparent contradiction. Firstly, it ispossible that NK cells are not involved in mediating the differential tumor growth rates observedin this model. NK cells may be stimulated by cytokines released by another immune reaction (i.e.a CTL response to the tumor) and may not in fact be able to recognize and/or lyse the SC 115tumor cells. Alternately, NK cells may be able to recognize and lyse the SC 115 tumor, but thelevel of lysis may not be sufficient to inhibit the growth of the tumor to any great extent.Secondly it is possible that NK cells may act to modulate SC 115 tumor growth in a positivemanner. This could occur if, as a result of activation, NK cells released a cytokine or hormone-like molecule (i.e. IL-2) which was capable of stimulating SC 115 tumor growth. A study by Weiand Heppner (1989) supports this latter possibility. They demonstrate that the neoplastictransformation of preneoplastic hyperplastic alveolar nodule cells in the mammary glands of miceis stimulated by a factor released by activated NK cells which are attracted to the glands by thehyperplastic lesion. It has also been demonstrated that in early human breast cancer, levels of NKcell activity are higher than those found in normal control subjects (Brenner and Margolese, 1991,Pross eta!., 1984, Zielinski eta!., 1989). It is not clear if this increase in NK cell activity plays arole in inhibiting the spread of metastatic disease or if this activity is in some way facilitating thegrowth of these tumors. The subsequent study is designed to distinguish between the hypothesesthat in DD/S mice of our model NK cells act to increase the growth rate of the SC 115 tumor, orthat NK cell activity does not affect the growth rate of the SC 115 tumor. The effect of in vivomodulation of NK cell activity, both up-regulation and down-regulation, on the differentialgrowth of the SC 115 tumor as observed in this model is investigated.110C. In Vivo Modulation of NK Cell Activity and Its Effects on the Differential Growth ofthe SC! 15 Tumor Observed in the Model.IntroductionOur previous studies demonstrated that the presence of the SC 115 tumor markedlystimulated NK cell activity both at the spleen and at the tumor site. Further, there were significantdifferences in the levels of NK cell activity among mice from the different housing groups of ourmodel. Interestingly, mice of the GI group, that develop the largest tumors, showed the greateststimulation of NK cell activity at both sites, whereas mice of the IG group, that develop thesmallest tumors, showed significantly lower levels of NK cell activity than mice of the GI group.It was concluded that although NK cell activity is clearly stimulated by the SC 115 tumor, eitherNK cells are not involved in mediating the differential tumor growth observed in this model or NKcells may actually be facilitating the growth of the SC 115 tumor in some manner. However, it hasbeen suggested that immune functioning observed in in vitro studies does not necessarily reflectthe situation in vivo due to loss of modulatory influences which are found in the in vivoenvironment (Ben-Eliyahu and Page, 1992). The present study was designed to investigate if NKcells are likely to play an important role in mediating the effects of psychosocial stressors ontumor growth rate in this model. This question was examined by directly modulating (stimulatingor suppressing) in vivo NK cell activity of mice from the GI (largest tumors) and the IG (smallesttumors) groups and observing the effect of these manipulations on tumor growth rate.It has been demonstrated that NK cell activity is stimulated by a variety of endogenouscompounds, of which 1L-2 and molecules of the IFN family are the most potent (Ben Arribia etal., 1989, McGinnes eta!., 1988, Robertson and Ritz, 1992a). NK cells have been shown topossess specific receptors for IL-2 (Ben Arribia eta!., 1989, Robertson and Ritz, 1990) and forIFN (Welsh, 1984). A number of exogenous molecules also stimulate NK cell activity.1 1 1Polyinosinic-polycytidylic acid (poly I:C), a double stranded synthetic DNA polymer has beenshown to be a potent stimulator of NK cell activity (Talmadge eta!., 1985a). An importantphysiological effect of poly I:C is that it stimulates the production of IFN (Fresa eta!., 1985,Korngold and Doherty, 1985), and the stimulation of NK cells by the injection of poly I:C isblocked by the injection of anti-IFN antibodies (Fresa et al., 1985). Interestingly, althoughrepeated injections of IL-2 or IFN induce hypo-responsiveness of splenic NK cells to furtherstimulation by these compounds, multiple injections of poly I:C do not induce hypo-responsiveness to further stimulation by poly I:C (Talmadge et al., 1985a, Talmadge et al.,1985b). Thus, poly I:C is a useful compound for inducing long lasting alterations in in vivo NKcell activity. It should be noted, however, that injections of poly I:C may have other effects onthe animal in addition to stimulating NK cell activity (Talmadge et al., 1985a). The induction ofIFN in the animal (by poly I:C) can have many effects on the antitumor immunity of the animal,including stimulation of tumoricidal macrophages (Schwamberger eta!., 1991) and CTL (Chen etal., 1986, Knop, 1990) and an increase in the expression of MHC class I antigens on the tumorcells (which could decrease NK cell lysis of the tumor, Dawson et al., 1992). Thus, the in vivoeffects of poly I:C on NK cell mediated immunity must be interpreted with caution.Elimination or suppression of in vivo NK cell activity is generally accomplished by theinjection into the animal of antibodies specific for NK cells. Several such antibodies have beendeveloped. One of the earliest antibodies shown to react specifically with NK cells is the anti-asialo GM1 antibody (ASGM1). This polyclonal antibody is produced in rabbits by immunizationwith the ganglioside asialo GM1 isolated from bovine brain (Kassai et al., 1980), and has beendemonstrated to recognize NK cells preferentially over T and B lymphocytes (Kasai et al., 1980,Habu eta!., 1981). ASGM1 is known to recognize NK cells in many strains of mice and rats(Habu eta!., 1981, Mason eta!., 1990, Pelletier eta!., 1987). More recently, it has beendemonstrated that several other cell types also express the asialo GM1 epitope, includingthymocytes, macrophages and, at a lower density, CTL (Suttles et al., 1987). Data generally112indicate that in vitro and in vivo treatment with low doses of ASGM1 result in the selectivedepletion of NK cells over CTL, due to the higher expression of the asialo GM1 epitope of NKcells than of CTL (Habu et al., 1981). However, several studies have suggested that, undercertain conditions, CTL are quite sensitive to lysis by ASGM1 (Doherty and Allen, 1987, Stitz etal., 1986, Stout et al., 1987). Thus, caution should be exercised in attributing alteration inimmune functioning induced by anti-asialo GM1 antibody solely to NK cells. A second antibodythat has been produced against the NK cell is the monoclonal antibody NK 1.1 which wasgenerated by immunizing (Balb/c x C3H)F1 mice with NK cells from CE mice (Koo and Peppard,1984, Seaman eta!., 1987). It has been demonstrated that in C57BL/6 mice, the NK1.1 antibodyis expressed on most functional NK cells as defined by morphological and functional analysis(Lemieux eta!., 1991, Seaman eta!., 1987). However, in many common strains of mice otherthan the C57BL, the NK1.1 antigen is reported not to be present (Burton eta!., 1991, Lemieux etal., 1991). Thus, a number of studies have continued to use the ASGM1 antibody due to the lackof a more specific antibody which will universally recognize NK cells in different strains of mice.The ASGM1 antibody was used in the present study as our preliminary experiments demonstratedthat the anti-NK1.1 antibody did not recognize NK cells from DD/S mice.The present study was designed to test the hypothesis that in vivo NK cell activity plays arole in modulating the differential tumor growth rates observed in our model. ASGM1 and polyI:C were administered to mice from GI (largest tumors) and IG (smallest tumors) groups toincrease (poly I:C) or decrease (ASGM1) their in vivo NK cell activity. The tumor growth ratesof these mice were monitored over 21 d post tumor cell injection.Materials and MethodsAnimals: Two hundred twenty-six male DD/S mice (2-4 months of age) were used in thisstudy. In most experiments, mice were housed individually for the course of the experiment and1 1 3were not exposed to the acute daily novelty stress regimen. The notable exception was the studyto investigate the effects of in vivo modulation of NK cell activity on the differential tumorgrowth of mice from the GI (largest tumors) and IG (smallest tumors) housing groups. For thisstudy. mice were raised and housed as described previously in General Methods, and these micewere subjected to acute daily novelty stress during the experimental period.Tumor cells:SC 115: SC 115 cells used in this study were fresh cells continuously propagated in maleDD/S mice as described in General Methods. For experiments involving tumors, tumor cells wereinjected into unanesthetized male mice as it has been suggested that anesthesia may inhibit theability of subsequent poly I:C injections to stimulate NK cell activity (Markovic and Murasko,1990). As this injection protocol was different from that used in previous studies in which micewere briefly anesthetized for the injection procedure, 6 mice raised according to our standardprotocol were injected in the usual fashion with anesthesia, as described in General Methods. Alltumors grew as expected in these control mice and reached a mass of approximately 3 g by 21 d.Yac 1 Cells: These cells were used in NK cell assays as previously described.Modulation of NK Cell Activity: Initial experiments investigated ASGMI-inducedsuppression and Poly LC-induced stimulation of in vivo levels of NK cell activity to determine thedose, injection schedule and time course of suppression or stimulation of NK cell activity. ForASGMi, doses of 20 and 50 ill of ASGM1 (diluted to a final volume of 100 IA: 1:5 and 1:2dilution) were tested as single doses, repeated doses and with or without prior (18 h) injection ofpoly I:C. Animals were terminated 1 to 15 d post injection. For poly I:C, a dose of 100 lig wasused to investigate the time course of stimulation of NK cell activity induced by repeated doses ofpoly I:C. Anti-asialoGMi (Wako Chemicals Inc., Texas) was purchased in lyophilized form.Antibody titer was estimated at 1:1000 by immunoflocculation. Each vial was reconstituted with1 1 41 ml distilled water and stored at 4°C. The ASGM1 was used within one month of reconstitution.Poly I:C (Sigma) was obtained as a sodium salt, diluted to a concentration of 1 mg/ml and storedat -20°C in 1 ml aliquots.Methods for determining the effect of in vivo modulation of NK cell activity on thedifferential tumor growth rates of mice from GI and IG groups were based on the results of theinitial experiments. To inhibit NK cell activity, mice were injected (ip) with 100 ill ASGM1 (1:5dilution). To stimulate NK cell activity, mice were injected (ip) with 100 gl poly I:C (1 mg/ml).Mice in each housing group were randomly assigned to one of 4 experimental treatmentconditions: injected ip with either poly I:C, ASGM1 or saline (to control for the stress of ipinjection), or not ip injected at all. The ip injections were given every 5 d (-1, 4, 9 and 14 d).Mice were injected with tumor cells without anesthesia on day zero and tumor growth wasmonitored for 21 d.Assays of Splenic NK Cell Activity: Splenic NK cell activity was measured as previouslydescribed.ResultsAnti-AsialoGM t-Induced Decreases in In Vivo Levels of NK Cell Activity: In the firstexperiment, the ability of ASGM1 to decrease NK cell activity in vivo following prior stimulationwith poly I:C was measured. All mice were injected ip with poly I:C (100 jtl , 1 mg/ml) and, 18 hlater, half the mice were injected ip with ASGM1 (100 IA, 1:5 dilution) and half the mice were ipinjected with vehicle. Three days after the injection of ASGMi, splenic NK cell activity wasmeasured (Figure 13). At this time, stimulation of splenic NK cell activity by poly I:C was not1 1 5Figure 13. Effect of a Single Injection of Anti-AsialoGMI on In Vivo NK Cell Activity in MicePreviously Treated With Poly I.C. All mice (individually housed) were injected ip with 100 p.1 ofpolyinosinic:polycytidylic acid (Poly I: C, 1mg/m1). Eighteen h later, mice were injected ip with100 ul of anti-asialoGMi antibody (ASGMi, 100 gl, 1:5 dilution) (n=3) or vehicle (n=3). Threedays later, mice were terminated and splenic NK cell activity was assessed. Points represent mean± SEM. In mice receiving ASGMi, NK cell activity decreased to approximately 50 % of thatobserved in mice receiving poly I:C alone. In mice stimulated by poly I:C alone , NK cell activitywas comparable to that previously observed in vehicle-injected mice (Rowse eta!., 1990).Spontaneous release was less than 10% of the total release.10080cn771) -"-cr.)^60u÷-)a)excrsH 40ga)c.)a)ca-4200 4.5:1^9:1^18:1 37.5:1 75:1 150:1Effector to Target Ratio117very remarkable. In mice stimulated by poly I:C alone, NK cell activity was comparable to thatpreviously observed in vehicle-injected mice (Rowse eta!., 1990). Treatment with ASGM1decreased NK cell activity to approximately 50 % of that observed in mice receiving poly I:Calone. This decrease in activity was less than had previously been reported for ASGM1 in otherstrains of mice (Habu eta!., 1981).A second experiment was conducted to investigate the effects of different dosages ofASGM1 (1:5 vs 1:2 dilutions) and terminating the mice at different times post injection (1 and 3d) on the ability of a single injection of ASGM1 (mice were not previously injected with poly I:C)to decrease splenic NK cell activity (Figure 14). Overall, there was approximately a 60 %suppression of splenic NK cell activity compared with that in vehicle-injected controls. Dataindicate that the effect of ASGM1 was greater at 3 d than at 1d, but that increasing the dose ofASGM1 from 1:5 to 1:2 dilutions did not appreciably increase the suppression of NK cell activityobserved.In a third experiment, the effect of 2 injections of ASGM1 (100 ul each, 1:2 dilution), 1 dapart, was investigated. This procedure has been reported to increase the effectiveness of theantibody (Ben-Eliyahu and Page, 1992). The data indicate that, at 3 d following the secondinjection of antibody, there was approximately a 60 % decrease in splenic NK cell activitycompared with that in vehicle-injected controls. This was similar to the inhibition observed with asingle injection of ASGMi. Thus, in the DD/S strain of mice, there did not appear to be asignificant advantage of multiple injections of ASGMl.A final experiment investigated the effects of repeated injections of ASGMi, with doses of1:5 and 1:2 dilutions (100 III each), on NK cell activity over a 2 wk period. Antibody wasinjected every 5 d (the distributor's recommended optimum) for 2 wk and mice were terminated atvarying times (2, 5, 7, 10, 12, 15 d). Data indicate that both doses of ASGM1 caused a lasting1 1 8Figure 14. Effect of Dosage of Anti-AsialoGMI and Time Post Injection on In Vivo NK CellActivity in Mice. Mice were injected ip with 100 ill ASGM1 (1:5 or 1:2 dilution), or saline. Micewere terminated at 1 d or 3 d post injection and splenic NK cell activity was assessed. Barsrepresent mean ± the range (dots). The suppression of NK cell activity (E:T ratio of 150:1)induced by ASGM1 was greater at 3 d than at ld. However, ASGM1 doses of 1:5 and 1:2dilutions appeared to be equally effective at inducing suppression of NK cell activity. (N's: 2 mice/treatment/day). Spontaneous release was less than 10% of the total release.1^ 3Days Post Injection120L 100o^8060  4.:       t. ::ASGM1 (1:2)ASGM1 (1:5)Control01191 20suppression of NK cell activity over the 15 d time course when injected every 5 d (Figure 15).There appeared to be a small degree of recovery of NK cell activity by 5 d following eachinjection. However, suppression of NK cell activity was greater than 60 % in most cases. Onceagain, there did not appear to be a marked difference between the 200 lig and the 500 1.tg doses ofASGM1 in the level of inhibition of splenic NK cell activity observed.Poly PC-Induced Stimulation of In Vivo Levels of NK Cell Activity: In the precedingpilot study, a single injection of poly I:C did not produce a large increase in splenic NK cellactivity at 90 h (approximately 3 1/2 d) post injection (figure 16). Thus, an experiment wasundertaken to investigate the effects of repeated doses of poly I:C in stimulating and maintainingthe stimulation of NK cell activity over time. Animals were injected with poly I:C or saline on day0, 5 and 10 and mice were terminated at day 1, 3, 5, 8, 10, 12 and 15.Poly I:C injection induced a marked increase in splenic NK cell activity at 1 d postinjection compared with that in vehicle-injected mice (approximately 300 %, Figure 16).However, by 3 d post injection, NK cell activity in poly I:C injected mice had declined noticeably( to approximately 210 %) and it declined further (to approximately 160 %) by 5 d post injection.Subsequent injections of poly I:C at 5 and 10 d did not increase splenic NK cell activity observedat 8 and 10 d and at 12 and 15 d, respectively, above the levels observed at 5 d.In light of the apparent lack of effect of the 5 and 10 d injections of poly I:C, a secondexperiment was undertaken to investigate the possibility that anesthesia might block thestimulatory effects of poly I:C on NK cell activity. Mice were injected with either poly I:C orsaline on day 0 and 5 and terminated at 1, 6 or 7 d following the first injection. Half of the polyLC-injected mice were briefly exposed to anesthesia (ether) during the injection. It was foundthat poly I:C induced a strong stimulation of splenic NK cell activity at all 3 time points examined(Figure 17).1 2 1Figure 15. Effect of Repeated Injections of Anti-AsialoGM1 on In Vivo NK Cell Activity in MiceOver Time. Mice were injected ip with either 100 ill of ASGM1 (1:2 and 1:5 dilutions) or salineon d 0, 5 and 10 and terminated either on d 2, 5, 7, 10, 12 or 15 (1:2 dilution ASGM1 or saline),or on d 2, 5 or 10 (1:5 dilution ASGM1). Splenic NK cell activity was assessed. Points representmean ± SEM, the cross-hatched bar represents the mean of the controls 1 standard deviation.Repeated injection of both doses of ASGM1 induced lasting suppression (greater than 60 %) ofNK cell activity (E:T ratio of 150:1) over time, although there appeared to be a small degree ofrecovery of NK cell activity by 5 d following each injection. Once again, ASGM1 doses of 1:2and 1:5 dilutions appeared to be equally effective at inducing suppression of NK cell activity.(N's: ASGM1=3 mice /dose/day, Vehicle= 2 mice /day). Spontaneous release was less than 15%of the total release.120100806040200-2 0 ASGM 1 Injections• ASGM11:5—A-- ASGM11:2Or • Control2^4^6^8 10 12 14 16Days1 22123Figure 16. Effect of Repeated Injections of Poly I:C on In Vivo NK Cell Activity in Mice OverTime. Mice were injected ip with either 100 pl poly I:C (1mg/m1) or saline on d 0, 5 and 10 andwere terminated on days 1, 3, 5, 8, 10, 12 and 15. Splenic NK cell activity was assessed. Pointsrepresent mean ± SEM, the cross-hatched bar represents the mean of all controls ± 1 standarddeviation. Poly I:C injection induced a marked increase in splenic NK cell activity (E:T ratio of150:1) at 1 d post injection compared with that in saline-injected mice. There was a progressivedecrease in NK cell activity in poly I:C injected mice from d 3 to 5. Subsequent injections of polyI:C at 5 and 10 d did not increase splenic NK cell activity observed at 8 and 10 d and at 12 and 15d, respectively, above the levels observed at 5 d. (N's: poly I:C, 2 mice /day (1, 3, 8, 12 d), 3mice /day (5, 10, 15 d); Vehicle = 2 mice /day). Spontaneous release was usually less than 10%of the total release and always less than 15%.-2 0^2^4^6^8 10 12 14 16Days400350300250--111-- Poly I:C200Control15010050o1 24125Figure 17. Effect of Anesthesia on the Ability of Repeated Injections of Poly I.-C to Stimulate InVivo NK Cell Activity in Mice Over Time. Mice were injected ip with either poly I:C (1mg/m1) orsaline on day 0 and 5 and terminated at 1, 6 or 7 d following the first injection. Poly LC-injectedmice were either briefly exposed to ether anesthesia during the injection (n=2) or received noanesthesia (n=2). Points represent mean ± range (dots), the cross-hatched bar represents themean of all controls ± 1 standard deviation. Poly I:C induced a strong stimulation of splenic NKcell activity at all 3 time points examined. There were no apparent differences in the level of NKcell stimulation observed in mice injected with or without anesthesia. (N's: 2 mice /treatment/day). Dots indicate the range. Spontaneous release was less than 10% of the total release.12645040035030025020015010050--e— Poly I:C NA0^ Poly I:C AControlI"  Days Post Injection1 27There were no apparent differences in the level of NK cell stimulation observed in mice injectedwith or without anesthesia.The results of the first 2 experiments present an apparent contradiction concerning theability of repeated injections of poly I:C to stimulate an increase in the levels of splenic NK cellactivity. The first experiment demonstrated that the second and third injections, at 5 and 10 drespectively, failed to induce an increase in NK cell activity 2 d later. However, in the secondexperiment, a second injection of poly I:C at 5 d caused a stimulation of NK cell activity on both6 and 7 d that was of equal magnitude to that observed 1 d following the initial injection. A thirdexperiment was undertaken to resolve this apparent conflict in the results. Mice were injectedwithout anesthesia, with either poly I:C or vehicle, on day 0, 5 and 10, and were terminated ondays 1, 3, 5, 6, 8, 10, 11, 13 and 15 post injection. With this more complete time course, wefound marked stimulation of splenic NK cell activity on days 1, 6 and 11, that is one day aftereach injection of poly I:C (Figure 18). NK cell activity subsequently declined back to controllevels by 5 d after each injection.Effect of In Vivo Modulation of NK Cell Activity on The Differential Tumor GrowthRates Observed in Our Model: The effects of in vivo modulation of NK cell activity on thedifferential tumor growth rates observed in our model were examined. Only mice from GI(largest tumors) and IG (smallest tumors) groups were included. Mice in each housing groupwere randomly assigned to one of 4 experimental treatment conditions, mice were injected ip witheither poly I:C, ASGM1 or saline (to control for the stress of ip injection), or were not injected atall. The ip injections were given every 5 d (-1, 4, 9 and 14 d). Mice were injected with tumorcells without anesthesia on day zero and tumor growth was monitored for 21 d.An overall ANOVA indicated that control mice injected ip with saline and control micewho were not injected did not differ significantly from each other in tumor growth rate (F(3,58) =128Figure 18. Effect of Repeated Injections of Poly I:C on In Vivo NK Cell Activity in Mice OverTime. Mice were injected ip without anesthesia, with either 100 IA of poly I:C (1mg/m1) or salineon days 0, 5 and 10 and terminated on days 1, 3, 5, 6, 8, 10, 11, 13 and 15. Splenic NK cellactivity was assessed. Points represent mean ± SEM, the cross-hatched bar represents the meanof all controls ± 1 standard deviation. There was marked stimulation of splenic NK cell activityon days 1, 6 and 11 (i.e. 1 d after each injection of poly I:C). NK cell activity decreased back tocontrol levels by 5 d after each injection. (N's: poly I:C = 3 mice /day, Vehicle = 2 mice /day).Spontaneous release was less than 10% of the total release.Days12935030025020015010050--11— Poly I:CControlr A2^4^6^8 10 12 14 161 302.106, P = 0.109). Therefore data were collapsed across these 2 control conditions. Thesubsequent ANOVA revealed significant main effects of Group (F(1,60) = 41.574, P<0.001) andDay (F(4,240) = 175.949, P<0.001). Further, effects of Treatment (F(2,60) = 2.953, P=0.060)and the Day x Treatment interaction (F(8,240) = 1.884, P=0.063) approached significance. Post-hoc tests revealed that, overall, mice of the GI group had significantly larger tumors than mice ofthe IG group (p<0.001, Figure 19). Interestingly, simple main effects analysis of the Day xTreatment interaction revealed that on days 7 through 15, mice in the 3 treatment conditions(ASGMi, poly I:C and control) were not significantly different from each other, whereas on day18 and 21 significant treatment effects emerged (p's<0.01). Post-hoc analysis revealed that at 18d, mice treated with poly I:C had significantly larger tumors than mice treated with ASGIVIi(p<0.01) or control mice (p<0.05, Figure 19). At 21 d post injection, both mice treated with polyI:C (p<0.001) and control mice (p<0.05) had significantly larger tumors than mice treated withASGMl.The data were further analysed with the 21 d time point omitted. Our previousexperiments demonstrated that the effects of ASGM1 and poly I:C gradually decrease and aresignificantly reduced by 5 d following the last injection. We reasoned that because the effects ofASGIsAi and poly I:C decrease over time and because the last injection of these agents was at 14d, it was possible that, by 21 d, tumor growth was no longer affected by these agents. With thissubset of data, the ANOVA revealed significant effects of Group (F(1,60) = 45.822, P<0.001),Treatment (F(2,60) = 4.166, P=0.020) and Day (F(3,180) = 129.506, P<0.001). Further, theGroup x Treatment x Day interaction was now highly significant (F(3,180) = 2.726, P=0.015).Post-hoc analysis revealed that at this time, mice treated with poly I:C had significantly largertumors than mice treated with ASGM1 and control mice (p's<0.001, Figure 20). In contrast, formice in the IG group tumor growth did not differ among mice in the 3 treatment conditions at anytime point (p>0.50).1 3 1Figure 19. Effect of In Vivo Modulation of NK Cell Activity on The Differential Tumor GrowthRates Observed in GI and IG. Mice from GI (open symbols) and IG (closed symbols) (asdescribed in figure 4) were injected ip with 100 gl of either poly I:C (lmg /ml), ASGM1 (1:5dilution), saline or received no injection. Injections were given on d -1, 4, 9 and 14. Mice wereinjected with tumor cells without anesthesia on d 0 and tumor growth was monitored for 21 d.Points represent mean ± SEM. Overall, mice of the GI group had significantly larger tumors thanmice of the IG group (p<0.001). At 18 d, mice treated with poly I:C had significantly largertumors than mice treated with ASGM1 (p<0.01) or control mice (combined data for saline anduninjected conditions, p<0.05). At 21 d post injection, both mice treated with poly I:C andcontrol mice had significantly larger tumors than mice treated with ASGM1 (p<0.05). (N's: GI,ASGM1(8), poly I:C(9), control(12), IG, ASGM1(9), poly I:C(9), control(19)).DAY3.53.02.542.0c_.C 1.5H 1.00.55^10^150.00 20 251 32ASGM1= 0  Poly I:C= A  Control= o  1 3 3Figure 20. Effect of In Vivo Modulation of NK Cell Activity on The Differential Tumor GrowthRates Observed in GI and IG Mice at 18 Days Post Injection. Mice from GI and IG (asdescribed in figure 4) were injected ip with either poly I:C (lmg /m1), ASGM1 (100 g of a 1:5dilution), saline or received no injection. Injections were given on d -1, 4, 9 and 14. Mice wereinjected with tumor cells without anesthesia on d 0 and tumor growth was monitored for 18 d.Bars represent mean ± SEM. Mice treated with poly I:C had significantly larger tumors than micetreated with ASGM1 and control mice (combined data for saline and uninjected conditions)(p's<0.001). In contrast, for mice in the IG group tumor growth did not differ among mice in the3 treatment conditions at any time point (p>0.50). 1Poly I:C^Control1 342.502.00/ /IGGI0.500.001 3 5DiscussionThis study investigated the effect of in vivo modulation of NK cell activity on thedifferential growth rates of the SC 115 tumor in mice from the experimental housing groups of ourmodel. To accomplish this, we utilized agents that produce prolonged suppression or stimulationof in vivo NK cell activity. Suppression of NK cell activity was induced by ip injection of theantibody ASGMi. It was found that injection of ASGM1 resulted in an approximately 60 %reduction of in vivo splenic NK cell activity. Further, this reduction was maintained over time bythe repeated administration of the antibody every 5 d. Interestingly, the reduction of NK cellactivity observed in DD/S mice was less than that previously reported for this dose of antibody inother strains of mice (Habu et al., 1981). It is possible that the route of administration might beresponsible for the differences between our study and other studies. Most studies using ASGM1administer the antibody by tail vein injection. This route of injection was unacceptable for thisstudy as it would have involved an increase in the level of stress experienced by the animalsinjected with ASGM1 compared with animals injected with poly I:C. A second possibility is thatNK cells of DD/S mice do not uniformly express the asialo GM1 antigen, do not express highlevels of the antigen or are for some reason less susceptible to antibody- and complement-mediated lysis using this antibody. A third possibility is that DD/S mice develop an immuneresponse to the rabbit ASGM1 antibody. However, this last possibility is unlikely as one wouldexpect to see a decrease in the effectiveness of the antibody with repeated injections and no suchdecrease was observed. Despite the less than complete reduction of NK cell activity, thistreatment was demonstrated reliably to cause a greater than 60 % reduction in NK. cell activity fora period of up to 2 wk.Stimulation of NK cell activity was accomplished by repeated injection of poly I:C. It wasfound that 1 d after each injection of poly I:C, there was a significant stimulation of splenic NKcell activity. This finding supports previous reports that the effects of poly I:C do not attenuate1 3 6with repeated injections (Talmadge eta!., 1985a, Talmadge eta!., 1985b) and is in contrast withthe reports of the hyporesponsiveness of splenic NK cell activity following repeated injections ofIL-2 (Talmadge eta!., 1985a, Talmadge eta!., 1985b). However, the stimulation of NK cellactivity induced by an injection of poly I:C was observed to decline rapidly and, by 3 d postinjection, was approximately half that observed at 1 d post injection. By 5 d post injection thelevels of splenic NK cell activity were close to the levels observed in vehicle-injected mice. Theapparently conflicting results on the effects of repeated injections of poly I:C observed in the first2 poly I:C time course experiments were resolved by the third experiment which revealed that thediscrepancy was caused by the differences in the time of sampling (8 d vs 6 and 7 d). This declinein the stimulation of NK cell activity with time following an injection of poly I:C is more rapidthan has previously been reported in the literature. One study reported that, in C3H mice, a singleip injection of 0.5 mg/kg poly I:C resulted in a significant stimulation of splenic NK cell activityfor at least 7 d post injection (Talmadge eta!., 1985a). This dose of poly I:C is 5 times lowerthan that used in our study (approximately 2.5 mg/kg, a commonly used dosage), but comparablelevels of NK cell stimulation were observed.It is assumed that splenic NK cell activity is reflective of overall levels of NK cell activity.This assumption is supported by our previously finding (Chapter 1B, part 4) that the stimulationof NK cell activity by the SC 115 tumor is of similar magnitude in the spleen and at the tumor site.In contrast, it has been reported that following repeated injections of IL-2, stimulation of NK cellactivity declines in the spleen but is maintained in nonlymphoid target organs (Talmadge et al.,1985a). However, as attenuation of the NK response to poly I:C following repeated stimulationwas not observed, it appears that the attenuation of NK cell responses to poly I:C is not an issuein this study.Finally, the effect of modulating in vivo NK cell activity on the differential growth of theSC 115 tumor in mice housed in different experimental groups was examined. We postulated that1 3 7if NK cell activity played an important role in modulating the differential tumor growth ratesobserved in our model, then in vivo modulation of NK cell activity would result in an alteration ofthe differential tumor growth rates observed in mice from the IG and GI groups. That is, sincepoly I:C increases NK cell activity and ASGM1 decreases NK cell activity, one would expect thatthese treatments should produce opposite effects on tumor growth relative to each other.The data revealed that modulation of NK cell activity did result in a significant alterationin tumor growth rate. Importantly, poly LC-induced stimulation of NK cell activity resulted in asignificant increase in tumor growth rate at 18 d and 21 d post tumor cell-injection compared withthat in ASGMi-injected mice. As expected, the tumor growth rate of control mice was found tobe intermediate to that of poly I:C- and ASGMi-injected mice. The finding that poly I:Cstimulates the growth rate of the SC 115 tumor compared with that observed in ASGMi-injectedmice suggests that NK cells may actually facilitate the growth of the SC 115 tumor. These dataare consistent with our previous finding that mice of the GI group (who develop the largesttumors) have significantly increased levels of NK cell activity both in the spleen and at the tumorsite compared with mice of the IG condition (who develop the smallest tumors).It has been demonstrated that NK cells are active early in the immune response and thattheir activity declines at later stages. For example, NK cell activity of lymphocytes infiltratinghighly immunogenic allografts declines by about 8 d (Hoffman eta!., 1988). We havedemonstrated that NK cell activity of tumor-infiltrating lymphocytes has begun to decline by 10 dpost tumor injection, although it is still elevated at 17 d post injection when compared with that incontrol mice. In our experiment, the last injections of poly I:C and ASGM1 were given at 14 dpost-tumor injection. As we have demonstrated that the stimulatory and inhibitory effects ofthese compounds decline markedly by 5 d following each injection, it is likely that the growth ofthe tumor observed at 21 d post injection was not influenced by the treatment conditions to thesame extent as was observed at 18 d. When only the first 18 d of tumor growth were considered,1 3 8it was found that the effects of poly I:C and ASGM1 were observed only in mice of the GIcondition and that these effects were highly significant. The observation that the modulation ofNK cell activity only affects the growth of tumors in mice of the GI group and not in mice of theIG group suggests that several different processes may be involved in controlling the growth ofthe SC 115 tumor. This possibility is strengthened by the finding that modulation of in vivo NKcell activity, either stimulation of inhibition, does not result in similar tumor growth rates in miceof the GI and IG groups. Obviously, although NK cells appear to play a role in modulating thegrowth of the SC 115 tumor, other mechanisms must also be involved in mediating the effects ofpsychosocial stressors on the growth of tumors in this model.In conclusion, this study demonstrates that NK cells do play a role in modulating thedifferential tumor growth rates observed in our animal tumor model. Surprisingly, NK cellsappear to be involved in increasing SC 115 tumor growth rate through as yet undeterminedmechanisms and appear to be involved selectively in mice of the GI group. Further, it is apparentthat mechanisms other than those related to NK cell activity are also involved in mediating theeffects of psychosocial stressors on the differential tumor growth rates observed in this model.CHAPTER 4 ENDOCRINE STUDIES.In addition to the immune system, the endocrine system may also play a role in mediatingthe effects of psychosocial stressors on the differential tumor growth observed in our model. TheSC115 tumor is a hormone-responsive tumor and several hormones and growth factors areknown to affect its growth (Bruchovski and Rennie, 1978, Emerman and Siemiatkowski, 1984,Furuya eta!., 1990, Kitamura eta!., 1979, Tanaka eta!., 1990). Further, the endocrine systemhas been shown to be exquisitely sensitive to stressors. Thus, the endocrine system may also be amediator of the differential tumor growth rates observed in this model.A) Morphological Studies of the SC! 15 Tumor Grown in Male and Female Mice.IntroductionBreast cancers are composed of a heterogeneous population of cells including hormone-responsive and hormone-independent cells (Heppner et al., 1981). This heterogeneity results inmetabolic and functional variability within a tumor and subsequent variation in response totreatments such as chemotherapy and hormone therapy (Emerman and Siemiatkowski, 1984,Heppner eta!., 1981, Miller eta!., 1981). The environment of the tumor may modulate thecomposition of subpopulations in the tumor in several ways. There may be negative selectionagainst some cell populations if required growth factors or nutrients are absent from theenvironment. Alternately, there may be positive selection of some cell populations if the growthof these cells is stimulated by a factor present in the environment to which other cellsubpopulations do not respond. The SC 115 tumor is an example of a tumor that is heterogeneousin its response to hormones (Bruchovski and Rennie, 1978, Emerman and Siemiatkowski, 1984,1 391 4 0Yates eta!., 1980). The predominant subpopulation of the SC 115 tumor grown in intact malemice is androgen-responsive, whereas the predominant subpopulation is androgen-independentwhen SC 115 cells are grown in female or castrated male mice (Bruchovski and Rennie, 1978,Emerman and Siemiatkowski, 1984, Yates et at, 1980). It is possible that the development of theandrogen-independent tumors in female and castrated male mice may result from the selectivegrowth advantage of androgen-independent cell subpopulations in an environment where theconcentrations of androgens are not sufficient to stimulate the proliferation of the androgen-responsive cell populations. Both androgen-responsive and androgen-independent cell cloneshave been isolated from androgen-responsive SC 115 tumors (Yamaguchi et cd., 1992, Darbe andKing, 1987). The androgen-responsive cell clones exhibit low rates of proliferation in the absenceof testosterone and a 15 fold increase in the rate of proliferation in the presence of testosterone(Yamaguchi, 1992, Darbe and King, 1987). In contrast, androgen-independent cell clones do notrespond to the presence of testosterone with increased cell proliferation rates (Yamaguchi et al.,1992). Thus, when androgens are not present at the levels required to stimulate the growth of theandrogen-responsive cell subpopulations, cells of the androgen-independent subpopulations havebeen shown to have a selective growth advantage and become the predominate subpopulationwithin the tumor.The heterogeneity of tumors is often manifest as an alteration of the morphology of thecells. For example, hormones are known to alter the morphology of the SC 115 tumor (Emermanand Worth, 1985, Yates eta!., 1980). Tumors grown in intact male mice (maintained in ourstandard laboratory conditions) have been shown to exhibit significant phenotypic differencesfrom tumors grown in castrated male mice and female mice (Emerman and Worth, 1985).Androgen-responsive tumors grown in intact male mice have a sheet-like growth pattern. Thetumors are highly vascularized but there are considerable areas of necrosis in the center of thetumor. Androgen-independent tumors grown in female and castrated male mice lose this cohesive141growth pattern. The cells form loose sheets and irregular cords dispersed within large amounts ofloose connective tissue stroma.In our experimental model, the slow growth rate of tumors observed in male mice movedfrom the individual to the group condition (IG) was of particular interest to us as it approximatesthe growth rate of the SC 115 tumor maintained in female mice (Weinberg and Emerman, 1989).Tumors from female and castrated male mice develop in a low androgen environment whereandrogen is provided solely by the adrenal gland. It has been demonstrated that, in male mice,crowded housing conditions may induce the suppression of testosterone secretion at all levels ofthe hypothalamic-pituitary-gonadal axis. Thus, in mice of the IG group, crowding could induce alow androgen environment. It is possible that in environments which provide suboptimal levels ofandrogens, androgen-independent cells may have a growth advantage relative to the androgen-responsive cells such that there is selective outgrowth of the androgen-independentsubpopulation. Considering the dramatic morphological differences between tumors grown inintact male and female mice raised in our standard housing conditions, the present study wasdesigned to investigate: 1) if the slow-growing tumors seen in male mice moved from individualto group housing would display histological characteristics similar to those seen in the androgen-independent tumors grown in female mice and 2) to examine further the histologicalcharacteristics of the SC 115 tumor grown in androgen-rich (male) and androgen-deprived(female) environments.METHODS AND MATERIALSAnimals: Male (n=17) and female (n=2) mice of the DD/S strain, 2 to 4 months of age,were used in this study. In accordance with our model, fourteen males were used to form the 4experimental housing groups (GG, GI, II, IG) and were subjected to acute daily novelty stress as1 4 2previously described in the General Methods. Two additional groups of animals were includedfor comparison; males (n=3) and females (n=2) raised in the standard sibling rearing groups weremaintained in their groups following tumor cell injection and were not subjected to acute dailynovelty stress.Tumors: The SC 115 tumor was used in this study as described in General Methods. Malemice were terminated 3 wk post tumor cell-injection and the tumors were excised. Female micewere terminated when the tumors reached a mass of 1.5 g (approximately 50 d).Fixation and Staining:Tissue blocks approximately 0.125 cm3 were cut from the center and periphery of thetumors and placed immediately into 10% formalin (4% formaldehyde) containing 2% (w/v)calcium acetate. Blocks were routinely paraffin processed after one wk of fixation. Serialsections cut at a thickness of 5 pm were stained with the following techniques:I) Carbohydrate Histochemistrya) Selective periodate oxidation-Schiff (PA*/S). This permits the specific demonstration ofsialic acids without side chain 0-acyl substituents or with an 0-acyl substituent at position C7(Volz eta!., 1987).b) Saponification selective periodate oxidation-Schiff (KOH/PA*/S). This permits thespecific demonstration of all sialic acids (Volz et al., 1987).c) Saponification selective periodate oxidation-Alcian blue pH 1.0-Schiff(KOH/PA*/A131.0/S). With this stain, all sialic acids stain magenta, o-sulphate esters stainaquamarine blue; mixtures stain in various shades of purple (Reid eta!., 1987).d) KOH/AB2.5/PAS. This is the standard Alcian Blue pH 2.5 Periodic Acid Schiff ofMowry (Mowry, 1963) preceded by a saponification step to remove any 0-acyl esters blockingvicinal diols. This stain serves as a control, demonstrating the presence of carboxyl groups and o-sulphate esters in aqua marine blue and all sugars containing vicinal diols magenta; mixtures stainin various shades of purple (Mowry, 1963).1 4 3e) Allochrome (Lillie, 1954). This provides a rapid visualization of the connective tissuepresent in the tumor.0 Haematoxylin and Eosin, H and E (Culling 1974). This is used for standardmorphological analysis (Culling, 1974).2) Immunohistochemistrya) Representitive sections from tumors of animals in each condition were stained withmuscle specific actin (MS A) and visualized by the immunoperoxidase technique (Papotti et al.,1988) to determine if cells had characteristics of myoepithelial cells. Slides were counterstainedwith haematoxylin. Muscle tissue served as a positive control.b) Representitive sections from tumors of animals in each condition were stained with 5-100 protein antiserum and visualized by the itnmunoperoxidase technique (Dwarakanath eta!.,1987). S-100 has been shown to be associated with myoepithelial cells. Slides werecounterstained with haematoxylin. Peripheral nerves, which also stain with S-100, served as apositive control.RESULTS:H & E stained sections of the tumors grown in females and males maintained under ourstandard laboratory conditions conformed to previous morphological descriptions of these tumors(Figure 21). The tumors from males had a cohesive epithelial-like growth pattern, high degree ofvascularization and large areas of necrosis. The tumors from female mice contained cellsdispersed into loose sheets and irregular strands growing in loose connective tissue. Themorphology of tumors from the male mice exposed to the 4 experimental housing conditions werethe same as that of the males housed in the standard conditions. Lillie's allochrome stainconfirmed that the significant amounts of loose stromal connective tissue present throughout thefemale tumors were not present in any of the male tumors.1 4 4Figure 21. Morphology of SC] 15 Tumors Grown in Male and Female Mice. Male and femalemice raised under standard colony conditions were injected with SC 115 cells (2 x 106) andterminated at 21 d (males) or 50 d (females; when the tumors were approximately 1.5 g) postinjection. Tissue blocks approximately 0.125 cm3 were cut from the center and periphery of thetumors and placed immediately into 10% formalin containing 2% (w/v) calcium acetate. Blockswere routinely paraffin processed after one wk of fixation. Serial sections cut at a thickness of 5gm were stained with H&E. The morphology of SC 115 tumors grown in male (A) and female(B) mice conformed to previous descriptions. Tumors from males exhibited a cohesive epithelial-like growth pattern. Tumors from female mice contained cells dispersed into loose sheets andirregular strands growing in loose connective tissue. (x 812.8).146In addition, we observed that the tumors from female mice contained regions ofextracellular material which appeared osteoid-like in H & E stained sections. As it has beenshown that osteoid of cortical bone contains chondroitin sulphate and sialic acid-richglycoproteins (Andrews et al., 1969, Herring and Kent, 1963), all male and female tumors wereinvestigated histochemically for the presence of these moieties.The development of osteoid-like regions in female mice appears to proceed in severaldefined steps. First a cluster of cells lost their attachments with neighbouring cells forming alobule (Figure 22A). The cell clusters then began to secrete an extracellular product, (Figure228), spreading away from each other as they did so until they achieved an osteoid-likeappearance (Figure 22C). When these extracellular regions were investigated histochemically(Table 3), they stained moderately positive with the PA*/S stain, indicating the presence of sialicacids. There was a moderate increase in the staining intensity of these extracellular regions whenthe KOH/PA*/S technique was used indicating the presence of substituted sialic acids. Further,when stained with the KOH/PA*/AB1.0/S technique these regions stained purple, indicating thepresence of sulphated moieties as well as sialic acid moieties in the osteoid-like regions. No suchosteoid-like regions were observed in any of the tumors grown in male mice with any of the stainsused.It has been suggested that, under the influence of the endocrine environment, breast cancerdevelops from undifferentiated cells which have the ability to become either mammary secretoryepithelial cells or mammary myoepithelial cells (Hayashi eta!., 1984). Thus, it is possible that thedifferent hormonal milieu of the male and female mice induces different phenotypes in the breasttumor cells. Therefore, we hypothesized that the slow growing tumors of the IG male mice mighthave a phenotype similar to that of tumors from female mice. To investigate this possibility, nearserial sections of tumors were stained with the myoepithelial cell-specific markers S-100 andMSA to display myoepithelial cells present in the tumors. All tumors stained for MSA to some147Figure 22. Development of Osteoid-Like Regions in Tumors of Female Mice. The developmentof osteoid-like regions in tumors of female mice proceeded in several defined steps. A. First agroups of cells (arrowheads) lost their attachments with the surrounding cells. B. The cellclusters then began to secrete an extracellular product (arrowheads). C. Secretion continueduntil an osteoid-like appearance was acheived. (x812.8)149Table 3. Histochemical Staining of the Osteoid-like Pools in the Female Mice. Female mice(n=2) were injected with SC 115 cells (2 x 106) and terminated when the tumors reached a massof 1.5 g (approximately 50 d). Tissue blocks approximately 0.125 cm3 were cut from the centerand periphery of the tumors and placed immediately into 10% formalin containing 2% (w/v)calcium acetate. Blocks were routinely paraffin processed after one wk of fixation. Serialsections cut at a thickness of 5 gm were stained. Osteoid-like pools were examined.TABLE 3. Histochemical Staining of the Osteoid-likePools in the Female Mice.Stain^ Female Osteoid PoolsPA*S^ ++1KOH/PA*/S 4--FKOH/PA*/AB1.0/S^P,M,A2KOH/PA*/AB2.5/S P,M,APAS^ +++AB 1.0 +AB2.5^ ++Lilles Mlochrome^P,B31. +, weakly positive; ++, moderately positive;+++, strongly positive.2. P, purple; M, magenta; A, aquamarine blue.3. P, purple; B, blue.1 501 51degree and staining was not observed in negative controls. Tumors from both males and femalescontained small clusters of strongly positive cells within the regions of viable cells. In addition, intumors taken from female mice, a layer of MSA positive cells lined the osteoid-like regions(Figure 23A).Specific S-100 staining was observed in all tumors examined and was not observed innegative controls. In tumors from male mice, the S-100 staining was less extensive than the MSAstaining. In contrast, the tumors grown in female mice had large regions of viable cells whichstained intensely S-100 positive. S-100 stained the cells in areas surrounding the osteoid-likeregions, but did not stain the MSA positive cells immediately adjacent to the osteoid-like regions(Figure 23B). Further, in both males and females there appeared to be little overlap in the areasthat were stained by these two techniques using near serial sections.DISCUSSION:It has been demonstrated that the SC 115 mammary carcinoma consists of a heterogeneouspopulation of androgen-responsive and androgen-independent cells (Bruchovsky and Rennie,1978, Emerman and Siemiatkowski, 1984). The predominant subpopulation of the SC 115 tumorgrown in intact male mice is androgen-responsive, whereas an androgen-independentsubpopulation is selected for when SC 115 cells are grown in female or castrated male mice.These two subpopulations of cells have different growth rates and morphologies (Emerman andSiemiatkowski, 1984; Emerman and Worth, 1985). Based on previous data (Weinberg andEmerman, 1989) indicating that tumors grown in male mice moved from individual to grouphousing conditions (IG) and tumors grown in female mice have similar slow growth rates, it washypothesized that the slow growth rate in mice of the IG group could be due to selection for aslow growing androgen-independent cell subpopulation similar to that in tumors of female mice,1 52Figure 23. Comparison of MSA and S-100 Staining of Serial Sections From Osteoid-LikeRegions of a Tumor Grown in a Female Mouse. Comparison of MSA (A) and S-100 (B) stainingof serial sections from osteoid-like regions of a tumor grown in a female mouse. The osteoid-likeregions (arrows) are surrounded by MSA-positive cells. In contrast, cells positive for S-100(arrows) are found in the areas outside the MSA-positive regions. (x 203.2)153and thus that tumors grown in IG males might also share other characteristics of thepredominantly androgen-independent subpopulation of cells.This study confirmed previous work (Emerman and Worth, 1985) showing morphologicaldifferences between tumors grown in female and intact male mice housed under our standardconditions. Importantly, tumors grown in IG males had a morphology similar to tumors of themale controls.The presence of osteoid-like extracellular material was observed in tumors grown infemale mice but not in male mice. It is notable that the slow growing tumors in IG males wereagain similar to tumors grown in males of the other housing groups rather than to the tumorsgrown in female mice.The osteoid-like regions of the female tumors contained sulphate and sialic acid moietiessimilar to the osteoid of cortical bone (Andrews eta!., 1969, Herring and Kent, 1963). Althoughbreast carcinomas are reported not to undergo differentiation to bone, calcification is a welldocumented occurrence in human breast tumors (Bouropoulou eta!., 1984, Frappart eta!., 1986;Hatter eta!., 1969, Sickles, 1980). Such regions of calcification have been reported to containsialic acid moieties (Bouropoulou eta!., 1984). The production of extracellular material by theSC 115 tumor grown in female mice may be similar to the production of regions of calcificationreported in human breast cancers. Such osteoid-like regions have been previously reported inSC 115 tumors grown in female mice (Kitamura eta!., 1979).S-100 and MSA immunohistochemical staining patterns differed markedly from each otherin both male and female tumors. Myoepithelial cells, as demonstrated by the MSAimmunohistochemical stain, were found in small isolated clusters in all tumors investigated. Theseclusters were usually located in the center of sheets or cords of viable cells. Importantly, MSA1 541 55staining was more prominent in female tumors than in male tumors and it was associated with theosteiod-like regions in female tumors. It has been suggested that the differentiation of breasttumor stem cells into epithelial or myoepithelial cells is controlled by local environmentalconditions (Hayashi eta!., 1984). It appears that the SC115 tumor is capable of myoepithelial celldifferentiation and that this differentiation is promoted by the hormonal environment of thefemale. Furthermore, in tumors from female mice, myoepithelial cells appeared to be linked to theproduction of the osteoid-like extracellular material.The S-100 staining pattern in males was less extensive than that seen with MSA,whereasin tumors of females the S-100 staining was at least as extensive as seen with the MSA stain. Inboth males and females there was no overlap between the regions stained by the two techniques.The large regions of S-100 positive cells in tumors from female mice are of particular interest. Ithas been shown that the S-100 protein belongs to a family of structurally related proteins whichshare a high degree of sequence homology and exhibit extensive cross-reactivity immunologically(Kligman and Hilt, 1988). The S-100 family consists of calcium-binding proteins which probablyact as second messengers, similar to calmodulin (Kligman and Hilt, 1988), and may be involved inthe promotion of cell division and the calcium-induced depolymerization of microtubles. One S-100 protein, p9ka, has been isolated from rat myoepithelial cells (Barraclough eta!., 1987). Thelack of correlation between MSA and S-100 staining of tumors from female mice suggests thatthe S-100 protein being recognized is not the myoepithelial cell-associated p9ka, but rather, maybe an S-100 protein involved in regulation of cell division or microtubule formation. This view isconsistent with studies of primary human breast cancer, which find that both epithelial andmyoepithelial cells stain positive with the S-100 technique (Dwarakanath et at, 1987; Stroup andPinkus, 1988).In this study, it was demonstrated that although tumors grown in male mice moved fromindividual to group housing are similar to tumors grown in female mice with regard to theirgrowth rate, they clearly do not resemble tumors grown in female mice in their histology. Rather,slow growing tumors of IG males are histologically similar to the fast growing tumors of controlmales. These results suggest that the growth of tumors in IG mice may not be the result ofselection for an androgen-independent tumor phenotype as observed in female mice.1 561 57B) Selected Studies of Endocrine Functioning in Mice from the 4 Experimental HousingGroups.IntroductionOne possible mediator of the differential tumor growth rates observed in this model is analteration in endocrine functioning of the mice. Animals in our model experience a variety ofpotentially stressful stimuli, including brief anesthetization, tumor/vehicle-injection, and, for someanimals, a change in housing condition and/or individual housing. It is known that androgens andglucocorticoids are responsive to stress (Amario and Castellanos, 1984, Christian and Davis,1964, Frankel and Ryan, 1981) and also have significant effects on the growth rate of the SC 115tumor (Omukai et al., 1987). Thus, it is possible that changes in plasma levels of these hormonesmay be responsible for the altered tumor growth rates observed in this model.Previously we have demonstrated that, at 3 wk post injection and group formation, plasmalevels of testosterone, dihydrotestosterone and corticosterone were not significantly differentamong mice from the 4 housing conditions (Weinberg and Emerman, 1989). This suggests thathousing condition may not influence hormone levels. However, it is known that hormonalresponses to stressors are dynamic and adapt to the chronic application of stressor with time. Ithas recently been demonstrated that when male mice are moved from individual housing to a largegroup, there is an immediate rise in basal corticosterone which returns to normal levels within 2wk (Peng et al., 1989) . Further, acute stressors such as ether anesthesia (used in our modelduring tumor/vehicle injection) have been shown to cause transient alterations in bothcorticosterone and testosterone levels (Armario and Lopez-Calderon, 1986, Frankel and Ryan,1 581981). Thus, functionally important changes in the endocrine system may occur within the firstwk post tumor cell-/vehicle-injection and group formation and have returned to normal by 3 wk.Another possible mediator of differential tumor growth rates in our model could be a shiftin the responsiveness of the tumor cells themselves to hormones. As noted, the SC 115 tumor isheterogeneous, containing androgen-reponsive and androgen-independent cells (Emerman,1988,Emerman and Worth, 1985). Growing SC 115 tumor cells in an androgen-deprivedenvironment (in a female mouse or in vitro) results in the selection of androgen-independent cells(Emerman, 1988, Emerman and Worth, 1985). Thus, selection for cells with greater or lesserhormone sensitivity may occur in animals in the 4 housing conditions resulting in differentialtumor growth. In a previous study we demonstrated that the morphology of the slow growingtumors from mice of the IG group did not resemble morphologically the slow growing androgen-independent tumors from female mice. However, it was possible that morphology and hormoneresponsiveness of the tumor cells are not directly linked.The present study was designed to test the hypothesis that the differential tumor growthrates observed in this model are the result of differences in testosterone and/or corticosteronesecretion which occur in mice of the 4 experimental housing groups during the first wk postinjection and group formation. Also, this study investigated the hypothesis that the differentialtumor growth rates observed in this model result from the selection of a subpopulation of tumorcells which have altered hormonal responsiveness and associated changes in growth rate. Todetermine if the experimental conditions resulted in an early change in basal hormone levels, weexamined basal levels of plasma testosterone and corticosterone early in the experimental period:1, 3 and 7 d post tumor cell- /vehicle-injection and group formation. To determine if ourexperimental conditions resulted in a change in the hormone sensitivity of the tumor cellsthemselves, the in vitro hormone responsiveness of tumor cells from mice in the different housing1 59conditions was examined at 3 wk post tumor cell-injection (the time when the tumor mass waslarge enough for study).METHODS AND MATERIALSAnimals. Two hundred and twenty two male DD/S mice (2-4 months of age) were usedin this study. Animals were raised and housed as described in General Methods.Tumors. SC 115 cells used in this study were fresh cells continuously propagated in maleDD/S mice as previously described in General Methods.Plasma Hormone Levels. Two hundred and six animals were terminated 1 d (24 h), 3 dand 7 d following tumor cell- or vehicle-injection and group formation. Animals were notsubjected to novelty stress on the termination day so that basal hormone levels could bemeasured. Cages were quickly and quietly carried from the colony room to an adjacentlaboratory, animals were briefly anesthetized (15 sec) with ether and immediately decapitated.For cages with more than one mouse per cage, several experimenters were involved in thedecapitation procedure to ensure than all animals were terminated within 1 min of first touchingthe cage. Trunk blood was collected in heparinized tubes. Following centrifugation at 2200 x g,plasma was separated and stored at -200C for subsequent assay.Testosterone levels were measured by radioimmunoassay using a modification of themethods of Auletta et al. (Auletta et al., 1974). Briefly, 200 pl of plasma was extracted twicewith 4 mls of anhydrous diethyl ether. Samples were pooled and dried under nitrogen in a 50°Cwater bath. Testosterone was then measured by standard radioimmunoassay (RIA). Antiserumand tracer were obtained from Radioassay Systems Laboratories, Inc. (Carson, C.A.). Samples1 60were counted in Formula 989 (New England Nuclear, Lachine, Quebec). Cross reactivity of theantiserum with dihydrotestosterone (DHT) was less than 20% and we have previouslydemonstrated that, in DD/S mice, DHT comprises less than 10% of total plasma androgen(Weinberg and Emerman, 1989).Corticosterone was measured by radioinununoassay as previously described (Weinbergand Emerman, 1989). Briefly, 30 IA of plasma was extracted in 270 pi absolute ethanol and totalcorticosterone was then measured by standard RIA. Antiserum and tracer were obtained fromRadioassay Systems Laboratories, Inc. (Carson, C.A.). Samples were counted in Formula 989(New England Nuclear) .In Vitro Hormone Responsiveness of Tumor Cells. For these studies, mice from GI(largest tumors, n=6) and IG (smallest tumors, n=10) conditions were terminated at 17 d postinjection, when the fastest growing tumors had reached a mass of approximately 3 g. Tumorswere removed and dissociated to single cell suspensions as described in General Methods. Cellsfrom 5 mice (IG) or 3 mice (GI) were pooled and resuspended at a concentration of 1.6 x 105cells/ml in DMEM and either 2% dextran-coated charcoal-treated fetal bovine serum (DCC FBS;Grand Island Biology Co, Burlington, Ontario) alone or 2% DCC FBS plus either 10-7 M DHT(Sigma) or 10-6 M HC (Sigma). The hormone concentrations chosen were those shownpreviously to provide maximal stimulation of cell growth using tumors from mice housed understandard control conditions (Jiang et al., submitted). Cells were plated in 96 well plates, 100ml/well, 12 wells/condition, and incubated for 5 d at 37°C, 95% CO2, with complete mediachanges on d 2 and 4. On d 5, numbers of viable cells were determined by the MIT (344,5-dimethylthiazol-2-y1]-2,5-diphenyltetrazolium bromide; Sigma) colorimetric assay (Carmichael etal., 1987, Mosmann, 1983).161Statistical Analyses. Statistical analyses were performed using appropriate analyses ofvariance (ANOVA) for the factors of Group, Tumor and Day where appropriate. Significantmain effects and interactions were further analyzed by Tukey post-hoc tests (p<0.05).RESULTSPlasma Levels of Testosterone. The ANOVAs revealed significant main effects of Day(F(2,186) = 9.189, P<0.001) and Group (F(3,186) = 4.499, P<0.005) as well as a Group x Dayinteraction (F(6,186) = 3.102, P<0.01) (Figure. 24). Post hoc analysis of the Group x Dayinteraction revealed that, at 1 d post tumor cell- or vehicle-injection and group formation, GG andII animals had significantly greater basal testosterone levels than IG animals (p<0.05). Mice in theGI group also had marginally elevated plasma testosterone levels compared with mice in the IGgroup (p<0.10). At 3 d and 7 d, plasma testosterone levels in GI animals were significantlyelevated over levels of animals in all other groups (p<0.05). The presence of a tumor did notsignificantly affect plasma testosterone levels, (F(1,186) = 1.169, P=0.281).Plasma Levels of Corticosterone. The ANOVAs revealed significant main effects of Day(F(2,182) = 6.101, P<0.005) and Group (F(3,182) = 44.252, P<0.001). Post hoc comparisons ofgroup effects revealed that IG animals had significantly elevated basal levels of plasmacorticosterone compared with animals in all other groups at all time points measured, p<0.05(Figure. 25). Overall, plasma corticosterone levels showed significantly greater elevations at 1 dpost injection than at 1 wk post injection (p<0.05). As with testosterone, plasma corticosteronelevels were not affected by the presence of a tumor (F(1,182) = 0.051, P=0.822).Effect of Social Status (in Group Housed Mice) on Plasma Hormone Levels. Anovasrevealed that, in mice of the GG group there was no effect of dominance on plasma hormone1 62Figure 24. Plasma testosterone levels. Mice from the 4 experimental housing groups (asdescribed in figure 4) were injected with tumor cells (2 x 106) or vehicle and terminated on d 1, 3or 7 post injection. To insure that hormone levels measured were basal values, mice were rapidlydecapitated immediately following disturbance of their cage. Points represent mean ± SEM. At 1d, mice in the II and GO conditions had significantly increased basal testosterone levels comparedwith IG mice (p<0.05). On d 3 and 7, GI mice had significantly greater basal levels than all othermice (p<0.05).2000^1^2^3^4^5^6^7^8DAYS POST INJECTION1 64Figure 25. Plasma corticosterone levels. Mice from the 4 experimental housing groups (asdescribed in figure 4) were injected with SC 115 tumor cells (2 x 106) or vehicle and terminatedon d 1, 3 or 7 post injection (as described in figure 24). Points represent mean ± SEM. IG micehad significantly greater basal corticosterone levels than mice in all other conditions at all timepoints measured (p<0.05).6001650^1^2^3^4^5^6^7^8DAYS POST INJECTION1 6 6Table 4. Effect of Social Status (in Group Housed Mice) on Plasma Hormone Levels. Mice ofthe GG and GI groups (as described in figure 4) were injected with SC 115 tumor cells (2 x 106)or vehicle and terminated on d 1, 3 or 7 post injection (as described in figure 24). Dominancestatus was assessed by examining the relative frequency of tail wounds. The mouse with thefewest or no wounds, in each group, was assessed as dominant and all other mice in the groupwere classified as subordinant. In mice of the GG group there was no effect of dominance onplasma hormone levels. In mice of the IG group, subordinate mice had significantly lower levelsof plasma testosterone and higher levies of plasma corticosterone than did the dominant mice(p<0.05).Table 4Plasma Testosterone (n2/ml)DayDominantGG^ IGDominantSubordinate Subordinate1^7.561 12.67 5.08 1.19± 6.09 ± 3.00 ± 2.35 ± 0.323^6.22 3.49 4.03 2.09+ 3.76 ± 2.02 ± 1.82 ± 0.897^0.67 0.97 3.77 1.69± 0.13^± 0.24^± 3.15^+ 0.46Plasma Corticosterone (n2/m1)DayDominantGG^ IGDominantSubordinate Subordinate1^13.86 14.12 30.84 50.54± 5.82 ± 1.12 ± 7.66 ± 7.443^15.97 14.20 26.54 33.81± 2.53 ± 2.76 ± 6.27 ± 3.377^8.61 7.68 17.84 30.03± 0.32 ± 0.48 ± 3.90 ± 4.311671 Mean ± sem.1 68levels. In contrast, in mice of the IG group, there was a significant effect of dominance status onplasma levels of both corticosterone (F(1,72)=5.002, P<0.03) and testosterone(F(1,70)=6.299,P<0.02). Post hoc analysis revealed that, in mice of the IG group, subordinate mice hadsignificantly lower levels of plasma testosterone and higher levels of plasma corticosterone thandid the dominant mice (p<0.05, Table 4).In Vivo Growth of Tumor for the In Vitro Hormone Responsiveness Study. At 17 d postinjection, tumors of mice in the GI group were significantly larger than tumors of mice in the IGgroup (Figure. 26). These data are consistent with our previous findings using this model(Weinberg and Emerman, 1989).In Vitro Hormone Responsiveness. In vitro proliferation of tumor cells from GI and IGanimals in response to DHT and HC was examined. As shown in Figure 27, cells from both GIand IG mice were significantly stimulated by both DHT (F(1,68) = 386.630, P<0.001) and HC(F(1,68) = 673.335, P<0.001). Interestingly, a significant main effect of group for both DHT(F(1,68) = 52.598, P<0.001) and HC (F(1,68) =30.130, P<0.001) indicated that tumor cells fromIG mice had a significantly greater response to DHT and HC stimulation than did tumor cells fromGI mice.DiscussionThis study demonstrates that alterations in plasma levels of testosterone andcorticosterone in the first several days post tumor cell- /vehicle-injection and group formation mayplay a role in mediating the differential tumor growth rates observed in mice housed under thedifferent conditions of our model. As well, the study indicates that the slow growth rate of1 69Figure 26. Tumor Growth in Mice of The GI and IG Groups. Mice of the GI (n=12) and IG(n=20) groups were injected with SC 115 tumor cells (2 x 106). Points represent mean ± SEM.Tumors were significantly larger in GI mice than in IG mice on day 17 (p<0.05).-3.55^10^15^200.0 ^0DAYS POST INJECTIONIG 03.00.5GI A-1 7 1Figure 27. In vitro hormone response of tumor cells from IG mice and GI mice. Mice from theGI and IG housing conditions were injected with SC 115 tumor cells (2 x 106) and terminated ond 18. Tumors were removed, dissociated to a single cell suspension and cultured on collagen-coated 96-well tissue culture plates in medium containing 2 % dextran charcoal-treated fetalbovine serum alone or in combination with either 10 -8 M dihydrotestosterone or 10 -7 Mhydrocortisone (doses previously shown to provide optimal stimulation of SC 115 cells in ourhands (Jiang et al., submitted)). After 5 d, cultures were terminated and cell growth determinedby the MTT assay. Bars represent mean ± SEM. Cell growth was expressed as percent of cells incontrol cultures. Tumor cells from IG mice were significantly more responsive to both DHT andHC than cells from GI mice (p<0.05). ^1 72300GI=IGDHT (10-7 M)^HC (10-6 M)tumors in mice of the IG group was not due to the selection for a hormone-independentsubpopulation of tumor cells.Plasma testosterone levels were elevated in mice of the GG, GI and II groups, but not inmice of the IG group, at 1 d post injection and group formation. At 3 d and 7 d, testosteronelevels of mice in the GG and II condition had declined, whereas the levels of mice in the GIcondition remained elevated. The elevation of plasma testosterone levels 1 d post injection andgroup formation may be associated with an increase in level of arousal induced by this procedure.Frankel and Ryan (1981) have shown that, in rats, anesthetization causes a rise in plasmatestosterone levels which is followed by a suppression 4 to 8 h later. In our model, thestimulation of testosterone secretion was of a much longer duration, and the suppressive effectwas not apparent. However, our procedure involved transporting mice to an adjacent laboratory,a brief anesthetization, injection of tumor cells/vehicle, placing animals in a clean cage and, insome cases, a change of housing condition (GI and IG). While any of these factors alone wouldprobably be a mild stressor for the animal, their combination could result in a longer lasting effect.Although mice in both the GI and II conditions were housed individually following tumorcell- /vehicle-injection, they did not demonstrate similar patterns of testosterone secretion.Testosterone levels remained elevated for 7 d in mice of the GI condition, whereas they declinedby 3 d in mice of the II condition. It has been shown that isolated mice are hyper-responsive toexternal stressors (Bronson, 1967). It is possible that mice which were moved from group toindividual housing were more sensitive to external stressors than were mice raised as individuals.Thus, mice in the GI condition may adapt more slowly to the effects of the acute daily noveltystress and continue to display acute elevations of testosterone in response to the stress. To ourknowledge, differences in the endocrine responses to stress of mice moved from group toindividual housing compared to those of mice that remain individually housed from weaning havenot been explored previously. Interestingly, mice of the GI and II groups also differed in their1 731 74levels of splenic NK cell activity at 3 d post tumor cell-injection. Thus, it is likely that the priorhousing condition significantly affects mice housed individually for the experiment.Animals in the IG condition, which have the slowest tumor growth rates, did not show anincrease in plasma testosterone levels 24 h post injection and group formation. It is likely that, formice of the DD/S strain, a density of 5 mice per cage constitutes crowding. In rodents, crowdingcauses a suppression of testosterone secretion (Koike and Noumura, 1989) and decreases theinfluence of stressors on plasma testosterone levels (Armario and Lopez-Calderon, 1986).Although 5 mice per cage is a lower density than is traditionally used for crowding studies (Ortizeta!., 1985), the DD/S strain of mice appears to be more aggressive than other strains. Males ofother strains are routinely housed in groups of 5-7 per cage (Brain and Nowell, 1970); howeverwe are unable to house DD/S males in groups of 5 for periods longer than 6-8 wk since woundingand occasionally even death will result. Observation of male DD/S mice housed in groups of 5revealed that, in a 45 min period after lights out, there may be as many as 6-10 fights per cage(Sault eta!., in preparation). Furthermore, a period of isolation prior to crowding is known toincrease the aggression of male mice (Valzelli, 1973) and crowding effects have been reportedwith mice housed 4 per cage when the mice were isolated for several wk prior to group housing(Bronson, 1973). Thus, it is possible that mice of the IG condition did not exhibit increasedtestosterone levels at 1 d post group formation because the suppressive effects of crowdingcompete with the stimulatory effects of anesthetization and/or tumor cell- /vehicle-injection.During the first wk post tumor cell- /vehicle-injection and group formation, basal levels ofplasma corticosterone were significantly elevated in IG mice compared with mice in all othergroups. It has been reported that crowding elevates plasma glucocorticoid levels in male mice(Bronson, 1973, Peng eta!., 1989). This effect has been attributed to the fighting involved in theestablishment of dominance hierarchies (Brain, 1975, Peng eta!., 1989). Since mice of the IGcondition fight significantly more than mice of the GG condition (Sault eta!., in preparation) this175could explain their increased plasma corticosterone levels. Interestingly, further analysis of ourdata indicated that for mice of the IG condition, subordinate mice had significantly suppressedbasal testosterone and increased basal corticosterone levels compared to dominant mice. Suchdifferences were not observed in mice of the GG condition. The importance of dominance inmodulating tumor growth rate is currently being investigated.Importantly, glucocorticoid levels have been shown to exert a regulatory influence ontestosterone levels (Sapolsky, 1986) . Chronically elevated glucocorticoid levels are thought todepress plasma testosterone levels by inhibiting the actions of LH on Leydig cells (Sapolsky,1986). Thus, the pituitary-adrenal axis may be indirectly involved in the regulation of tumorgrowth rates by exerting a modulatory action on plasma testosterone secretion. In our model,basal corticosteroid levels were elevated in IG mice and testosterone levels were low.Conversely, basal corticosterone levels were low in all other groups and testosterone levels werehigh at 1 d post injection. Consistent with our results, Hiraoka eta! (1987) demonstrated in vivothat doses of corticosterone comparable to those seen in IG mice ( 10-7M) inhibited the growth-stimulating effect of physiological doses of androgen.Finally, our data indicate that the differential tumor growth rates observed in our modelare not due to altered hormone sensitivity of the tumor cells. In vivo, tumors in IG mice had aslower growth rate than tumors in GI mice, whereas in vitro, cells from tumors of IG miceconsistently grew faster and were actually more responsive to DHT and HC than cells fromtumors of GI mice. These data suggest that the slower growth rates of tumors in mice of the IGgroup result from alterations in the internal environment of the mice rather than from a decrease inthe tumor cell's ability to respond to hormones. These data are consistent with our previousfindings on tumor morphology (study 4A. of this chapter). The previous study indicated thattumors grown in male mice of the 4 housing groups were similar to each other in morphologicalcharacteristics and were also similar in morphology to SC 115 tumors grown in male mice raised1 7 6under the standard laboratory conditions. In contrast, tumors grown in female mice, a conditionknown to select for androgen-independent subpopulations of tumor cells, demonstrated amorphology which differed considerably from that observed in the male mice. Thus, these datasuggest that selection for an androgen-independent phenotype did not occur in mice of the IGgroup. Further, we have demonstrated that tumors cells grown in mice from the 4 housingconditions did not differ in terms of their of androgen or glucocorticoid receptor levels (Rowse etal., 1992). Together, these data provide strong evidence that the differential tumor growth ratesobserved in this model are not due to the selection of subpopulations of SC115 cells with differenthormone responsiveness but rather are due to alterations in the environment of the cells (i.e.altered plasma hormone levels and altered NI( cell activity).In summary, we have demonstrated that alterations in basal levels of plasma testosteroneand corticosterone occur in the first wk following tumor cell/vehicle injection and groupformation. These alterations likely represent an important factor modulating the differentialtumor growth rates observed in our model. Future studies will examine this issue using malesfrom the 4 housing conditions that have been castrated and/or adrenalectotnized and implantedwith osmotic pumps to maintain chronically high or chronically low basal hormone levels.CHAPTER 5 DISCUSSIONThis thesis examined several potential physiological mediators of psychosocial stressoreffects on the growth of the SC 115 tumor. In our animal tumor model, the growth of theandrogen-responsive SC 115 tumor is modulated by an animal's social housing condition. Micewhich are raised in groups from weaning age to adulthood and are subsequently housed asindividuals for the course of the experiment (GI) have significantly increased tumor growth ratescompared with those in mice who remain group housed for the experiment (GO). In contrast,mice who are raised as individuals and are subsequently rehoused in groups for the experiment(IG) have significantly decreased tumor growth rates compared with mice of the GO group.Previous evidence suggested that the growth of the SC 115 tumor may be influenced by bothimmune (Watanabe et al., 1982, Nohno et al., 1986) and endocrine (Emerman and Worth, 1985,Omukai et al., 1987) variables. Thus, this dissertation examined the role of specific immune andendocrine variables in modulating the differential tumor growth rates observed in this model. Thedata demonstrate that both immune and endocrine variables are likely to play an important role inmodulating the growth of the SC115 tumor.Our studies of the immune system demonstrated that the presence of the SC 115 tumorstimulated NK cell activity both in the spleen (Rowse et al., 1990) and at the tumor site.Interestingly, at both sites, NK cell activity was found to be greater in mice of the GI group, thosemice who develop the largest tumors, than in mice of the IG group, those mice who develop thesmallest tumors. One interpretation of these data is that NK cells may actually stimulate ratherthan inhibit the growth rate of the SC 115 tumor. This hypothesis was investigated by modulating(increasing or decreasing) in vivo NK cell activity in mice of the GI and IG groups and monitoringtumor growth rate. Modulation of in vivo NK cell activity did not affect the growth rate oftumors in mice of the IG condition. Importantly, however, it was discovered that in mice of the1 771 7 8GI group, in vivo modulation of NK cell activity was accompanied by a corresponding change intumor growth rate. That is, protocols which increased NK cell activity also stimulated the growthrate of the SC 115 tumor relative to that observed in mice with suppressed NK cell activity.Therefore, the data on NK cell activity in mice from the different housing groups support thehypothesis that NK cells play a role in stimulating growth of the SC 115 tumor in mice of the GIgroup. This finding is consistent with a study by Wei and Heppner (1987) which demonstratedthat NK cells increased the malignant transformation of preneoplasmic mouse mammary epithelialcells. The mechanism by which NK cells may stimulate the transformation or growth of malignantmammary epithelial cells in not known. However, it is possible that cytokines or peptidehormones secreted by activated NK cells may stimulate the growth of the mammary tumor cells.The endocrine system also appears to play a role in mediating the effects of psychosocialstressors on the differential tumor growth rates observed in our model. The data indicate that, inthe first 7 d post group formation, there is a strong correlation between plasma levels oftestosterone and tumor growth rate. Mice of the GI group, who develop the largest tumors, haveelevated basal levels of plasma testosterone throughout the first 7 d post group formation. Incontrast, mice of the IG group, who develop the smallest tumors, maintain low basal levels ofplasma testosterone through out the first 7 d of the experiment. Mice of the GG and II groups,who develop tumors of intermediate size, have elevated basal levels of plasma testosterone at 1 dpost group formation, but, by 3 d, plasma testosterone levels decline to the low basal levels ofmice of the IG group. Altered plasma levels of testosterone are likely to effect the growth rate ofthe SC 115 tumor as these cells possess functional androgen receptors and are stimulated toproliferate in the presence of physiological levels of testosterone in vitro (Bruchovsky and Rennie,1978, Emerman and Worth, 1985, Hiraoka et al., 1987). Interestingly, basal plasma levels ofcorticosterone, although elevated in all groups at 1 d compared with 3 and 7 d post groupformation, were significantly elevated in mice of the IG group compared with all other groupsduring the first 7 d of the experiment. The elevated plasma corticosterone levels could affect1 7 9tumor growth in several ways. These include the possibility that elevated levels of corticosteronecould interfere with the ability of testosterone to stimulate the growth of the SC 115 tumor(Hiraoka et al., 1987). These conditions occur in mice of the IG group who have low basal levelsof plasma testosterone and high basal plasma levels of corticosterone. In addition, elevatedplasma corticosterone levels have been shown to decrease the secretion of testosterone by a directeffect on the Leydig cells of the testis (Sapolsky, 1986). Thus, it appears very likely thatalterations in the secretion of testosterone and corticosterone play an important role in mediatingthe effects of psychosocial stressors on the differential tumor growth rates observed in this model.Interestingly, all available data suggest that selection for a slow-growing hormone-independent SC 115 cell subpopulation, similar to that observed in female DD/S mice, does notoccur in mice of the IG group. This suggestion is supported by 3 independent lines of evidence.First, data from this dissertation demonstrate that, in vitro, SC 115 cells from tumors grown inmice of the IG group exhibited greater rates of proliferation in response to DHT or HC than didcells from tumors grown in mice of the GI group (Rowse et al., 1992). Second, it was previouslydemonstrated by our laboratory that SC 115 cells from mice of the 4 experimental housing groupsdid not differ in their maximum binding capacity or binding affinity for both the androgen and theglucocorticoid receptor (Rowse et al, 1992). Finally, as demonstrated by the data in this thesis,tumors grown in mice from the 4 housing groups exhibited morphological characteristic similar toeach other and to tumors from mice reared in standard housing conditions and different from themorphology of androgen-independent tumors grown in female mice (Rowse eta!., 1990). Thus,we conclude that the differential tumor growth rates observed in this model probably do not ariseas a result of the selection for a subpopulation of SC 115 cells with altered hormone sensitivity.Rather, in vivo alterations in the environment of the cells, such as plasma hormone levels and NKcell activity, may be involved in modulating tumor growth rate. However, heterogeneity for othercharacteristics than hormone responsiveness may also be involved in mediating the differentialtumor growth rates observed in this model.1 80The data presented in this thesis suggest that several physiological mechanisms mayinteract to produce the differential tumor growth rates observed in this model. It is possible that,in mice of the GI group, the increased rate of tumor growth results from the combined stimulatoryactions of elevated plasma levels of testosterone and of activated NK cells at the tumor site. Incontrast, the decreased tumor growth rate of mice in the IG group may result from the lowplasma testosterone levels and an inhibitory action of the elevated plasma levels of corticosterone.Although these studies have revealed several physiological mechanisms by which psychosocialstressors may affect the growth rate of the SC 115 tumor, many questions still remain.1.) Further investigation is required to confirm the role of altered plasma levels oftestosterone and corticosterone in mediating the differential tumor growth rates. As mentionedpreviously, although the alterations in plasma levels of testosterone are strongly correlated withthe changes in tumor growth rate, it is not clear if the magnitude of these changes is sufficient tocause the differential tumor growth rates. This question could be investigated by monitoringtumor growth in adrenalectomized and castrated male mice implanted with osmotic pumpscontaining low or high doses testosterone and/or corticosterone. This would demonstrate theeffect of maintaining chronically high, chronically low or no plasma testosterone, both alone andwith either chronically high or chronically low basal corticosterone levels. Such an experimentwould reveal if the alterations in plasma hormone levels observed in the current study are of asufficient magnitude to induce the differential tumor growth rates observed with this model. Aswell, this experiment would allow the assessment of the relative contributions of plasmacorticosterone and testosterone to the differential tumor growth of the model.2.) It is possible that other hormones may also be involved in mediating the differentialtumor growth rates observed in this model. One interesting candidate is opioid peptides. Studiesin our laboratory have indicated that opioid peptides cause a suppression of DHT-, HC-, and1 8 1bFGF- stimulated SC 115 tumor cell growth in vitro (Jiang eta!., submitted). Further, it is knownthat various stressors induce the secretion of the opioid beta-endorphin from the pituitary andenkephalins from the adrenal gland (Rossier et al., 1977, Viveros et al., 1979). Thus, it is possiblethat endogenous opioids may mediate the suppression of SC 115 tumor growth rate observed inmice of the IG condition. Another category of hormones which could potentially play a role inmodulating the differential tumor growth rate observed in our model are the lactogenic hormones.As mentioned in the introduction, lactogenic hormones such as prolactin and growth hormone areknown to affect the growth and differentiation of normal mammary epithelial cells (Imagawa etal., 1990). As the SCI 15 tumor originated from the malignant transformation of breastepithelium of a female mouse and it may retain responsiveness to the growth promoting effects oflactogenic hormones. This possibility is supported by the finding that the androgen-responsiveSC 115 tumor subline used in these studies possesses functional estrogen receptors and respond toestrogen (Nohno et al., 1982). The effects of lactogenic hormones on the growth of the SC115tumor should be assessed both in vivo and in vitro.3.) It is possible that, in addition to NI( cells, other immune effector cell populations playa role in mediating the effects of pyscho social stressors on the differential tumor growth ratesobserved in our model. Data from several experiments in our laboratory suggest that CTL may bestimulated by the presence of the SC 115 tumor. First, as reported in this thesis, there is a markedincrease in the relative percentage of lymphocytes and monocytes in white blood cells infiltratingtumor cell-injected sponges 7 d post injection compared with that in vehicle-injected sponges.This finding suggests that a specific cell-mediated immune response could be occurring in theseanimals. This possibility is supported by the findings of Hoffman (1988), who demonstrated thatthere was a significant increase in the percentage of lymphocytes infiltrating polyurethane spongesduring an active immune response against allogeneic peritoneal cells. Furthermore, our laboratoryhas demonstrated that antitumor immune-activity against [51Cr1-labeled SC 115 cells may begenerated by culturing splenic lymphocytes with mitomycin C inactivated SC 115 cells for 6 d in1 82vitro (G. Rowse, unpublished observations). Importantly, this antitumor immune-activity wasonly observed when spleens of mice previously injected with SC 115 cells 9 or 14 d prior toharvesting of the splenocytes were used as the effector cells in the 6 d in vitro culture. Suchactivity was not generated in 6 d cultures of splenocytes from vehicle-injected mice or from miceinjected with SC 115 cells 21 d prior to the harvesting of the splenocytes (G. Rowse, unpublishedobservations). These findings strongly suggest that CTL may play a role in modulating thegrowth of the SC 115 tumor. Future studies could examine the role of CTL in mediating theeffects of psychosocial stressors on tumor growth rate in our model by measuring the ability oftumor-infiltrating lymphocytes to lyse [51Cd-labelled SC 115 cells and by directly inhibiting invivo CTL activity (using Lyt 2 antibody).In addition to these future lines of research, it is of interest to examine the implications ofthe current studies for human breast cancer. Of course, caution must be exercised whenextrapolating data from animal models to the human condition. Biological differences may existbetween humans and rodents in the regulation of physiological mechanisms. Further, it is difficultto compare directly the psychological experiences of experimental animals with the psychologicalexperiences of humans. Nevertheless, with these cautions in mind, certain correlations betweenthe current studies and findings in human studies are evident. Many studies in humans havesuggested the existence of a link between stressors and the altered growth rate of cancer. Themost compelling of these studies are the works of Grossarth-Maticek (1989) and Spiegel (1989),who independently demonstrated that psychological counselling could significantly extend the lifeexpectancy of women with advanced breast cancer. Several large prospective studies providesupport for these findings, indicating that decreased levels of meaningful social contact areassociated with increased susceptibility to cancer and poorer prognosis (Kaplan and Reynolds,1990, Ell, 1992, Waxler-Morrison, 1991). Our animal model also demonstrates that the level andtype of social contacts can significantly influence the growth rate of tumors in mice. Thephysiological mechanisms by which psychological counseling affects the survival of women with1 83advanced breast cancer are not known. Our model indicates that, in mice, NK cell activity andaltered plasma hormone levels are involved in mediating the effects of psychosocial stressors onbreast tumor growth rate. It is possible that similar physiological processes mediate the effects ofpsychological councelling in women with breast cancer as well.We observed that stimulation of NK cell activity is accompanied by increased growth ofthe SC 115 tumor. Interestingly, it has been shown that NK cells stimulate the malignanttransformation of hyperplastic alveolar nodules in the mammary glands of mice to a neoplasticstate (Wei and Heppner, 1989). These studies suggest that NK cells may stimulate the growth oftumor cells in the early stages of breast tumor development. However, many studies, in bothhumans and animals, suggest that NK cells are important in the destruction of tumor cells andespecially in the control of blood-borne metastases (Aslakson et al., 1991, Johnson et al., 1990,Greenberg et al., 1987). Prospective studies in humans demonstrate that greater levels of NK cellactivity in peripheral blood lymphocytes are a positive prognostic indicator for increased survivalof women with breast cancer (Levy et al., 1985). Further, studies in humans demonstrate thatsuppression of immune reactivity of lymphocytes in tumor-draining lymph nodes occurs as tumorsgrow in size (Reiss et al., 1983) and studies of lymphocytes infiltrating human breast cancersindicate that there is a suppression of NK cell activity in this cell population compared with that ofthe peripheral blood (Bonilla et al.,1988). Thus, studies in humans indicate that NK cells may beinvolved in suppressing tumor growth and metastases, i.e. the development of the tumor may beassociated with suppression of NK cell activity. These data are consistent with our findings in theSC 115 model, however, as we have demonstrated that, in addition to a positive correlationbetween NK cell activity and tumor growth rate in early tumor development, NK cell activitydeclines as the tumor grows in size.Currently it is not known if our observation that NK cells stimulate the growth of atransplantable murine breast tumor early in the development of the tumor is also applicable to1 8 4other tumors in rodents and or human tumors. It is possible that NK cells may play different rolesin modulating the growth of breast tumors depending on the site of action considered. In theprimary tumor, NK cells could stimulate the early development of the tumor. This effect couldresult if the tumor cells were relatively resistant to lysis by the NK cells but were stimulated toproliferate by a cytokine or hormone-like peptide produced by the activated NK cells. Thus, forthe tumor, the proliferation of the tumor cells in response to cytokines or hormones released byNK cells may outweigh the lysis of tumor cells by NK cells. In contrast, in the blood, theconcentration of tumor cells may be much lower allowing more NK cells to bind to these tumorcells, thus increasing the ability of the NK cells to lyse the tumor. Furthermore, the concentrationof cytokines and NK-derived hormones in the blood would be diluted and may be insufficient tostimulate the growth of the tumor cells.Studies examining the effects of in vivo NK cell activity on the early stages of tumordevelopment in humans have not been performed. However, the effect of NI( cells on the growthof early tumors in humans could be assessed by growing normal human breast epithelial cells incollagen gels to produce epithelial organoids (Imagawa 1990, Lawler et al., 1983) and thentransforming the cells by in vitro treatment with carcinogens (Ethier, 1987, Ganguly et al., 1982).These transformed breast epithelial organoids could then be cultured with NK cells isolated fromthe donor's peripheral blood lymphocytes.Similarly, it is also likely that alterations in plasma hormone levels play a significant role inmediating the stressor-induced alteration in tumor incidence and growth rate observed in humanstudies. We have demonstrated that, in our murine model, the alteration of plasma testosteroneand corticosterone levels are likely to play an important role in mediating the effects ofpsychosocial stressors on tumor growth rate. As discussed previously, stressors are known toalter the secretion of many hormones in humans. Further, most human breast cancers are thoughtto develop initially as hormone-responsive tumors (Dickson et al., 1992) and the alteration of185plasma hormone levels is thought to play an important role in the induction of breast cancer inhumans (Dickson eta!., 1992, Secreto eta!., 1991). Thus, stressor-induced alterations in plasmahormone levels may be universally important in mediating the effect of stressors on tumor growth.In summary, these experiments have demonstrated that stressors may alter the growth rateof tumors in animals. Further, in our model it was shown that alterations in NK cell activity andplasma hormone levels may play an important part in mediating the effects of psychosocialstressors on the growth of the SC 115 tumor. Studies in humans suggest that similar mechanismsmay be operating in the stressor-induced alteration in tumor incidence and progression.ReferencesAdams, D.O., Weiel, J.E., Becton, D.L., Somers, S.D. and Hamilton, T.A. (1985) RegulatoryMechanisms in Macrophage Tumoricidal Activation. in Macrophage Biology. S. Reichardand M. Kojima (Eds), Alan R Liss Inc., New York, pp. 3 - 14.Adams, D.O. and Johnson, W.J. (1982) Activation of Murine Mononuclear Phagocytes ForDestroying Tumor Cells: Analysis of Effector Mechanisms and Development. inMacrophages and Natural Killer Cells. S.J. Normann and E. Sorkin (Eds), Plenum Press,New York, pp. 707 - 718.Ader R. and Cohen N. (1975) Behaviorally Conditioned Immunosuppression. Psychosom. Med.37: 333-340.Amkraut, A. and Soloman, G.F. (1972) Stress and Murine Sarcoma Virus (Maloney)-inducedTumors. Cancer Res. 32:1428 - 1433.Anderson, S.K. (1992) Molecular Cloning of Human and Murine Natural Killer Cell TumorRecognition Proteins. in NK Cell Mediated Cytotoxicity: Receptors, Signaling andMechanisms. E. Lotzova and R.B. Herberman (Eds). CRC Press Inc., Boca Raton, pp. 1 -14.Andrews, A.T.deB., Herring, G.W. and Kent, P.W. (1969) The Periodate Oxidation of BovineBone Sialoprotein and Some Observations on its Structure. Biochem. J. 111:621 - 627.Armario, A. and Castellanos, J.M. (1984) A comparison of corticoadrenal and gonadal responsesto acute immobilization stress in rats and mice. Physiology and Behavior 32, 517-519.Armario, A., Castellanos, J.M. and Balasch, J. (1984a) Effect of Acute and Chronic PsychogenicStress on Corticoadrenal and Pituitary-Thyroid Hormones in Male Rats. Hormone Res.20:241 -.Armario, A., Castellanos, J.M. and Balasch, J. (1984b) Dissociation Between Corticosterone andGrowth Hormone Adaptation to Chronic Stress in the Rat. Horm. Metabol. Res. 16:142 -145.Armario, A., Restrepo, C., Castellanos, J.M. and Balasch, J. (1985) Dissociation BetweenAdrenocorticotropin and Corticosterone Responses to Restraint After Previous ChronicExposure to Stress. Life Sci. 36:2085 - 2092.Armario, A. and Lopez-Calderon, A. (1986) Pituitary gonadal function in adult male ratssubjected to crowding. Endocrine Res. 12, 115-122.186187Armario, A., Hidalgo, J. and Giralt, M. (1988) Evidence that the Pituitary-Adrenal Axis Does NotCross-Adapt to Stressors: Comparison to Other Physiological Variables.Neuroendocrinology 47:263 - 267.Armario, A, Marti, J. and Gil, M. (1990) The Serum Glucose Response to Acute Stress isSensitive to the Intensity of the Stressor and to Habituation. Psychoneuroendocrinology15:341 - 347.Aslakson, C.J., Mceachern, D., Conaway, D.H. and Miller, F.R. (1991) Inhibition of LungColonization at Two Different Steps in the Metastic Sequence. Clin. Exp. Metast. 9:139 -150.Asterita, M.F. (1985) The Physiology of Stress, With Special Reference to the NeuroendocrineSystem. Human Science Press Inc., pp.264.Auletta, F.J., Caldwell, B.V. and Hamilton, G.L. (1974) Androgens: testosterone anddihydrotestosterone. in Methods of Hormone Radioimmumoassay, B.M. Jaffe and H.R.Behrman (Eds.). Academic Press, New York, pp. 359-370.Axelrod J. and Reisine T.D. (1984) Stress Hormones: Their Interaction and Regulation. Science224: 452 - 459.Barraclough, R., Savin, J., Dube, S.K. and Rudland, P.S. (1987) Molecuoar Cloning andSequence of the Gene for p9Ka. A Cultured Myoepithelial Cell Protein with StrongHomology to S100, a Calcium-binding Protein. J. Mol. Biol. 198:13-20.Barrett J.C. (1992) Molecular Mechanisms of Hormonal Carcinogenesis. in Li, J.J., Nandi, S. andLi, S.A. (Eds) Hormonal Carcinogenesis. Springer-Verlag, New York. pp 159 -163.Beden,S.N. and Brain, P.F. (1982) Studies on the Effect of Social Stress on Measures of DiseaseResistance in Laboratory Mice. Aggressive Behaviour 8:126 - 129.Ben Arribia, M.-H., Moire, N., Metivier, D., Vaquero, C., Lantz, 0., Olive, D., Sharpentier, B.and Senik, A. (1989) IL-2 Receptors on Circulating Natural Killer Cells and T Lymphocytes.Similarity in Number and Affinity but Difference in Transmission of the Proliferation Signal.J. Immunol. 142:490 - 499.Ben-Eliyahu, S. and Page, G.G. (1992) In Vitro Assessment of Natural Killer Cell Activity.Progress in Neuroendocrine Immunology 5:199 - 214.Ben-Nathan, D. and Feuerstein, G. (1990) The Influence of Cold or Isolation Stress on Resistanceof Mice to West Nile Virus Encephalitis. Experientia 46:285 - 90.Bernton, E.W., Bryant, H.U. and Holaday, J.W. (1991) Prolactin and Immune Function. in Ader,R. Felton, D.L. and Cohen, N. (Eds) Psychoneuroimmunology 2nd edition. Academic PressInc., San Diego. pp. 403 - 423.Berthois, Y., Dong, X.-F. and Martin, P.M. (1989) Regulation of Epidermal Growth Factor-Receptor by Estrogen and Antiestrogen in the Human Breast Cancer Cell Line MCF-7.Biochem. Biophys. Res. Comm. 159:126 - 131.Bieliauskas, L.A. and Garron, D.C. (1982) Psychological Depression and Cancer. Gen. Hosp.Psychiatry 4:187-195.Biron, C.A, and Welsh, R.M. (1982) Proliferation and Role of Natural Killer Cells During ViralInfection. in NK Cells and Other Natural Effector Cells. R.B. Herberman (Ed.), AcademicPress Inc., New York, pp. 493 - 498.Blackburn, C.M. and Albert, A. (1959) Effects of 2-Methyl-Dyhydrotestosterone and ofTestosterone on Human Pituitary Gonadotropin. J. Clin. Endocrinol. 19:603 - 607.Blalock, J.E. and smith, E.M. (1980) Human Leukocyte Interferon: Structural and BiologicalRelatedness to Adrenocorticotropic Hormone and Endorphins. Proc. Natl. Acad. Sci. USA77: 5972 - 5974.Blalock, J.E. (1992) Production of Peptide Hormones and Neurotransmitters by the ImmuneSystem. in Blalock, J.E. (Ed) Neuroimmunology 2nd Edition. Chem. Immunol. 52: pp. 1 -24.Blanck, A., Gustafsson J.A. and Porsch-Hallstrom, I. (1992) Pituitary Regulation of Sex-Differentiated Promotion of Rat Liver Carcinogenesis. in Li, J.J., Nandi, S. and Li, S.A.(Eds) Hormonal Carcinogenesis. Springer-Verlag, New York. pp 251 -253.Blumberg, E.M., West, P.M. and Ellis, F.W. (1954) A Possible Relationship BetweenPsychological Factors and Human Cancer. Psychosom. Med. 16: 277 - 286.Bonilla ,(1988) Cancer 61:629 - 634.Boot, L.M., Kwa, H.G. and Ropcke, G. (1981) Hormonal Induction of Mouse MammaryTumors. in Mammary Tumors in the Mouse, Elsevier/North-Holland Biomedical Press,Amsterdam, pp. 117 - 199.Bouropoulou, V., Anastassiades, 0.Th., Kontogeorgos, G. and Rachmanides, M. (1984)Microcalcifications in Breast Carcinomas - A Histochemical Study. Path. Res. Pract. 179:51-58.Bovbjerb, D., Ader, R. and Cohen, N. (1982) Behaviorally Conditioned Inununosuppression of aGraft-vs-Host Response. Proc. Natl. Acad. Sci. USA 79: 583 - 585.188Brain, P.F. and Nowell, N.W. (1970) The Effects of Differential Grouping on Endocrine Functionof Mature Male Albino Mice. Physiol. Behav. 5:907 - 10.189Brain, P.F. and Nowell, N.W. (1971) Effect of Prior Housing on Adrenal Response to IsolationGrouping in Male Albino Mice. Psychonomic Sci.22:183 -.Brain, P.F. (1972) Endocrine and Behavioural Differences Between Dominant and SubordinateMale House Mice Housed in Pairs. Psychon. Sci. 28:260 - 262.Brain, P.F. (1975) What Does Individual Housing Mean to a Mouse? Minireview. Life Sci.16:187 - 200.Brann, D.W. and Mahesh, V.B. (1991) Role of Corticoids in Female Reproduction. FASEB J.5:2691 - 2698.Brayton, A.R. and Brain, P.F. (1974) Studies of Effect of Differential Housing on Some Measuresof Disease Resistance in Male and Female Laboratory Mice. J Endocrinol. 61:R48 - R49.Brenner, B.G. and Margolese, R.G. (1991) The Relationship of Chemotherapeutic and EndocrineIntervention on Natural Killer Cell Activity in Human Breast Cancer Cancer 68:482 - 488.Briski, K.P. and Sylvester, P.W. (1987) Effects of Repetitive Daily Acute Stress on Pituitary LHand Prolactin Release During Exposure to the Same Stressor of a Second Novel Stress.Psychoneuroendocrinol. 12:429 - 437.Briski, K.P. and Sylvester, P.W. (1988) Effect of Specific Acute Stressors on LuteinizingHormone Release in Ovarectomized and Ovarectomized Estrogen-Treated Female Rats.Neuroendocrinology 47:194 - 202.Bronson, F.H. (1967) Effects of social stimulation on adrenal and reproductive physiology ofrodents. In: Husbandry of Laboratory Animals, pp. 513-542. Editor: M.L. Conalty.Academic Press, London.Bronson, F.H. (1973) Establishment of Social Rank Among Grouped Male Mice: Relative Effectson Circulating FSH, LH and Corticosterone. Physiol. Behav. 10:947 - 951.Bruchovsky, N. and Rennie, P.S. (1978) Classification of dependant and autonomous variants ofthe Shionogi mammary carcinoma based on heterogeneous patterns of androgen binding.Cell 13, 272-280.Brunet, F.M. (1970) The Concept of Immunological Surveillance. Prog. Exp. Tumor Res. 13:1 -27.Bullock, L.P., Bardin, C.W. and Sherman, M.R. (1978) Androgenic, Antiandrogenic andSynandrogenic Actions of Progestins: Role of Steric and Allosteric Interactions WithAndrogen Receptors. Endocrinol. 103:1768 - 1782.190Burton, R.C., Smart, Y.C., Koo, G.C. and Winn, H.J. (1991) Studies on Murine Natural Killer(NK) Cells. V. Genetic Analysis of NK Cell Markers. Cellular Immunology 135:445 - 453.Carmichael, J., DeGraff, W.G., Gazdar, A.F., Minna, J.D. and Mitchell, J.B. (1987) Evaluation ofa tetrazolium-based semiautomated colorimetric assay: assessment of chemo sensitivitytesting. Cancer Res. 47, 936-942.Carney, D.N., Bunn, P.A., Gazdar, A.F., Pagan, J.A. and John, D.M. (1981) Selective Growth inSerum-Free Hormone-Supplemented Medium of Tumor Cells Obtained by Biopsy FromPatients with Small Cell Carcinoma of the Lung. Proc. Natl. Acad. Sci. USA 78:3185 -3189.Carr, D.J.J. and Blalock, J.E. (1991) Neuropeptide Hormones and Receptors Common to theImmune and Neuroendocrine Systems: Bidirectional Pathway of Intersystem Communication.in Ader, R. Felton, D.L. and Cohen, N. (Eds) Psychoneuroimmunology 2nd edition.Academic Press Inc., San Diego. pp. 573 - 588.Casselith, BR, Walsh, W.P. and Lask, E.J. (1988) Psychosocial Correlates of Cancer Survival: ASubsequent Report 3 to 8 Years after Cancer Diagnosis. J. Clin. Onco. 6: 1753 - 1759.Cato, A.B.C., Konig, H., Ponta, H. and Herrlich P. (1992) Steroids and Growth Factors in theRegulation of Expression of Genes and Gene Networks. J. Steroid BioChem. Molec. Biol.43: 63 - 68.Cerottini, J.C. and Brunner, K.T. (1974) Cell Mediated Cytotoxicity, Mlograft Rejection andTumor Immunity. Adv. Immunol. 18, 67-132.Chen, L., Tourvielle, B., Burns, G.F., Bach, F.H., Mathiou-Mahul, D., Sasportes, M. andBensussan, A. (1986) Interferon: A Cytotoxic T Lymphocyte Differentiation Signal. Eur. J.Immunol. 16:767 - 770.Chong A. S.-F., Scuderi, P., Grimes, W.J. and Hershi, E.M. (1989) Tumor Targets Stimulate IL-2 Activated Killer Cells to Produce Interferon-y and Tumor Necrosis Factor. J. Immunol.142:2133 -2139.Christian, J.J. and Davis, D.E. (1964) Endocrines, behavior and population. Science 146, 1550-1560.Christian, J.J., Lloyd, J.A. and Davis, D.E. (1965) The Role of Endocrines in the Self-Regulationof Mammalian Populations. Recent Prog. Horm. Res. 21:501 - 578.Christian, J. (1970) Social Subordination, Population Density and Mammalian Evolution. Science168:84 - 90.191Clarke, B. and Bost, K.L. (1989) Differential Expression of Functional AdrenocorticotropicHormone Receptors by Subpopulations of Lymphocytes. J. Immunol. 143: 464 - 469.Colomb, E., Berthon, P., Dussert, C., Calvo, F. and Martin P.-M. (1991) Estradiol and EGFRequirements for Cell-Cycle Progression of Normal Human Mammary Epithelial Cells inCulture. Int. J. Cancer 49:932 - 937.Conner, R., Vernikos-Danellis, J. and Levine, S. (1971) Stress, Fighting and NeuroendocrineFunction. Nature 234:564 - 566.Cooper, C.L., Cooper, R. and Faragher, E.B. (1989) Incidence and Perception of PsychologicalStress: The Relationship with Breast Cancer. Psychol. Med. 19:415 - 422.Cox, R.H., Hubbard, J.W., Lawler, J.E., Sanders, B.J. and Mitchell, V.P. (1985) Cardiovascularand Sympathoadrenal Responses to Stress in Swim-Trained Rats. J. Appl. Physiol. 58:1207 -1214.Crowcroft, P. and Rowe, F.P. (1963) Social Organization and Territorial Behaviour in the WildHouse Mouse (Mus Musculus L.). Proc. Zoo. Soc. 140:517-31,Culling, C.F.A. (1974) Carbohydrates. In Handbook of Histopathological and HistochemicalTechnique. 3rd ed. Toronto, London: Butterworths.D'Andrea, A., Rengaraju, M., Valiante, N.M., Chehimi, J., Kubin, M., Aste, M., Chan, S.H.,Kobayashi, M., Young, D., Nickbarg, E. et al. (1992) Production of Natural Killer CellStimulatory Factor (Interleukin 12) by Peripheral Blood Mononuclear Cells. J. Exp. Med.175:1387 - 1398.Daniel, C.W. and Robinson, S.D. (1992) Regulation of Mammary Growth and Function by TGF-13. Mol. Reprod. Devel. 32: 145 - 151.Darbe, P.D. and King, R.J.B. (1987) Interaction of Different Steroid Hormones DuringProgression of Tumour Cells to Steroid Autonomy. Int. J. Cancer 40:802 - 806.Dawson, J.R., Wyatt, R.M. and Storkus, W.J. (1992) Natural Killer Cell - B LymphocyteInteractions: The Role of Target Cell Major Histocompatability Class I Molecules. in NKCell Mediated Cytotoxicity: Receptors, Signaling and Mechanisms. E. Lotzova and R.B.Herberman (Eds). CRC Press Inc., Boca Raton, pp. 409 - 421.Dechambre, R.P. and Grosse, C. (1973) Individual VersusGroup Caging of Mice with GraftedTumors. Cancer Res. 33:140 - 144.Dembinski, T.C. and Shiu, R.P.C. (1987) Growth Factors in Mammary Gland Development andFunction. in The Mammary Gland: Development, Regulation and Function. Neville, M.C.and Daniel, C.W. (Eds.), Plenum Press, New York, pp 355 - 381.192Demura, R., Suzuki, T., Nakamura, S., Komatus, H., Odagiri, E. and Demura, H. (1989) J.Andrology 10:210 -213.De Plaen, E., Lurquin, C., Van Pel, A., Mariame, B., Szikora, J.-P., Wolfe!, T., Sibille, C.,Chomez, P. and Boon, T. (1988) Immunogenic (turn-) Varients of Mouse Tumor P815:Cloning of the Gene of Turn- Antigen p91A and Identification of the Turn- Mutation. Proc.Natl. Acad. Sci. USA 85:2274 - 2278.Derogatis, L.R., Abeloff, M.D. and McBeth, C.D. (1976) Cancer Patients and Their Physicians inthe Perception of Psychological Symptoms. Psychosomatics 17: 197 - 201.DeTurck, K.H. and Vogel, W.H. (1980) Factors Influencing Plasma Catecholamine Levels in RatsDuring Immobilization. Pharm. Biochem. Behav. 13:129 - 131.DeWinter, J.A.R., Trapman, J., Vermey, M., Mulder, E., Zegers, N.D. and Vanderkw., T.H.(1991) Androgen Receptor Expression in Human Tissues - An Immunohistochemical Study.J. Hist. Cyto. 39:927 - 936.Dickson, R.B., Huff, H.K., Spencer, E.M. and Lippman, M.E. (1986) Induction of EpidermalGrowth Factor-Related Polypeptides by 17 Beta Estradiol in MCF-7 Human Breast CancerCells. Endocrinol. 118:138 - 142.Dickson, R.B., Thompson, E.W. and Lippman, M.E. (1990) Regulation of Proliferation, Invasionand Growth Factor Synthesis in Breast Cancer by Steroids. J. Steroid BioChem. Molec.Biol. 37: 305 -316.Dickson, R.B., Johnson, M.D., Mozeena, B., Shi, E., Kurebayashi J., Ziff; B., Martinez-Lacuci,I., Amundadottir, L.T. and Lippman, M.E. (1992) J. Steroid BioChem. Molec. Biol. 43: 69 -78.Djeu, J.Y., Timonen, T. and Herberman, R.B. (1982) Production of Interferon by Human NaturalKiller Cells in Response to Mitogens, Viruses and Bacteria. in NK Cells and Other NaturalEffector Cells. R.B. Herberman (Ed.), Academic Press Inc., New York, pp. 669 - 674.Doherty, P.C. and Allen, J.E. (1987) Anti-Asialo-GM1 Eliminates Both Inflammatory Processand Cytotoxic T-Cell Function in the Lymphocytic Choriomeningitis Adoptive TransferModel. Cell. Immunol. 107:1 - 7.Dunn, A.J. (1989) Psychoneuroimmunology For the Psychoneuroimmunologist: A Review ofAnimal Studies of Nervous System-Immune System Interactions. Psychoneuroendocrinology14: 251 -274.Dwarakanath, S., Lee, A.K.C., Delellis, R.A., Silverman, M.L., Frasca, L. and Wolfe, H.J. (1987)S100 Protein Positivity in Breast Carcinomas: A Potential Pitfall in DiagnosticImmunohistochemistry. Human Pathol. 18: 1144-1148.193Ebbesen, P., Villadsen, J.A., Villadsen, H.D. and Heller, K.E. (1991) Effect of Subordinance,Lack of Social Hierarchy, and Restricted Feeding on Murine Survival and Virus Leukemia.Exp. Gerontol. 26:479 - 486.Edwards, E.A. and Dean, L.M. (1977) Effects of Crowding of Mice on Humoral AntibodyFormation and Protection to Lethal Antigenic Challenge. Psychosom. Med. 39:19 - 24.Ell, K., Nishimoto, R., Mediansky, L., Mantel!, J. and Hamovitch, M. (1992) Social Relations,Social Support and Survival Among Patients with Cancer. J. Psychosom. Res. 36:531 - 541.Ellis T.M., McKenzie, R.S., Simms, P.E., Helfrich, B.A. and Fischer, R.I. (1989) Induction ofHuman Lymphokine-Activated Killer Cells by IFN-a and IFN-y. J. Immunol. 143:4282 -4286.Emerman, J.T. and Pitelka, D.R. (1977) Maintenance and Induction of MorphologicalDifferentiation in Dissociated Mammary Epithelium on Floating Collagen Membranes. InVitro 13:316 - 328.Emerman, J.T., Enami, J., Pitelka, D.R. and Nandi, S. (1977) Hormonal Effects on Intracellularand Secreted Casein in Cultures of Mouse Mammary Epithelial Cells on Floating CollagenMembranes. Proc. Natl. Acad. Sci. USA 74:4466 - .Emerman, J.T., and Siemiatkowski, T. (1984) Effects of endocrine regulation of growth of amouse mammary tumor on it's sensitivity to chemotherapy. Cancer Res. 44, 1237- 1332.Emerman, J.T., and Worth, J. (1985) Phenotypic Stability of Mouse Mammary Tumor CellsCultured on Collagen Gels. In Vitro Cellular and Developmental Biology 21(1), 49-56.Emerman, J.T. (1988) Effects of hormonal modulation on cytotoxicity of chemotherapeuticagents in mouse mammary tumor cell cultures. Anticancer Res. 8, 205-208.Emerman, J.T., Tolcher, A.W. and Rebbeck, P.M. (1990) In Vitro Sensitivity Testing of HumanBreast Cancer Cells to Hormones and Chemotherapeutic Agents. Cancer Chemother.Pharmacol. 26:245 - 249.Ethier, S.P. and Cundiff, K.C. (1987) Importance of Extended Growth Potential and GrowthFactor Independence on in Vivo Neoplastic Potential of Primary Rat Mammary CarcinomaCells. Cancer Res. 47:5316 - 5322.Ethier, S.P. and Moorthy, R. (1991) Multiple Growth Factor Independence in Rat MammaryCarcinoma Cells. Breast Cancer Research and Treatment 18:73 - 81.Ewertz, M (1986) Bereavement and Breast Cancer. Br. J. Cancer 53: 701 - 703.194Eysenck, H.J. (1988) Personality, Stress and Cancer: Predictions and Prophylaxis. Br. J.Psychol. 61: 57 - 75Felton, S.Y. and Felton, D.L. (1991) Innervation of Lymphoid Tissue. in Ader, R. Felton, D.L.and Cohen, N. (Eds) Psychoneuroimmunology 2nd edition. Academic Press Inc., San Diego.pp. 27 -70.Felton, S.Y., Felton, D.L., Bellinger, D.L. and Olschowka (1992) Noradrenergic and PeptidergicInnervation of Lymphoid Organs. in Blalock, J.E. (Ed) Neuroimmunology 2nd Edition.Chem. Immunol. 52: pp. 25 - 48.Fidler, I.J. and Schroit, A.J. (1988) Recognition and Destruction of Neoplastic Cells by ActivatedMacrophages: Descrimination of Altered Self. Biochim. Biophys. Acta 948: 151-173.Folkman, J. and Haudenschild, C. (1980) Angiogenesis In Vitro. Nature 288:551 -556.Forsen, A. (1991) Psychosocial Stress as a Risk for Breast Cancer. Psychother. Psychosom.55:176 - 185.Frances, H., Khidichian, F. and Monier, C. (1990) Increases in the Isolation-Induced SocialBehvioral deficitby Agonists at 5-HT1A Receptors. Neuropharmicol. 29:103-107,Frankel, A.I. and Ryan, E.L. (1981) Testicular innervation is necessary for the response for theresponse of plasma testosterone levels to acute stress. Biol. Reprod. 24, 491-495.Frappart, L., Remy, I., Lin, H.C., Bremond, A., Raudrant, D., Grousson, B. and Vauzelle, J.L.(1986) Different Types of Microcalcifications Observed in Breast Pathology. Correlationswith Histopathological Diagnosis and Radiological Examination of Operative Specimens.Virchows. Arch. A. 410: 179-187.Frasure, N.S. and Prince, R. 1985) The Ischemic Heart Disease Life Stress Monitoring Program:Impact on Mortality. Psychosom. Med. 47:431-445.Fresa, K.L., Korngold, R. and Murasko, D.M. (1985) Induction of Natural Killer Cell Activity ofThoracic Duct Lymphocytes by Polyinosinic-Polycytidylic Acid (Poly I:C) or Interferon.Cellular Inununol. 91:336 - 343.Friedman, S.B., Glasgow, L.A. and Ader, R. (1969) Psychosocial Factors Modifying HostResistance to Experimental Infections. Ann. New York Acad. Sci. 164:381-392.Fulton, A.M. (1987) Interactions of Natural Effector Cells and Prostaglandins in the Control ofMetastasis. J.N.C.I. 78(4), 735-741.Furuya, Y., Sato, N., Akakura, K., Ichikawa, T., Suzuki, N., Sato, R. and Shimazaki, J. (1990)Paracrine Growth Stimulation of Androgen - Responsive Shionogi Carcinoma 115 by ItsAutonomous Subline (Chiba Subline 2). Cancer Res. 50:4979-7983.195Gabelman, B.M. and Emerman, J.T. (1992) Effects of Estrogen, Epidermal Growth Factor, andTransforming Growth Factor-a on the Growth of Human Breast Epithelial Cells in PrimaryCulture. Exp. Cell Res. 201:113 - 118.Galin F.S., LeBoeuf R.D. and Blalock J.E. (1990) A Lymphocyte mRNA Encodes theAdrenocorticotropin/13-Lipotropin region of the Pro-Opiomelanocortin gene. Prog. Neuro.Endocrin. Immunol. #: 205 - 208.Ganguly, N., Ganguly, R., Mehta, N.M. and Banerjee, M.R. (1982) Growth and Differentiationof Hyperplastic Outgrowths Derived From Mouse Mammary Tumor Cells Transformed inOrgan Culture. JNCI 69:453 - 463.Gehde, E. and Baltrusch, H.J.F. (1990) Early Experience and Development of Cancer in LaterLife: Implications for Psychoneuroimmunologic Research. Int. J. NeuroSci. 51:257-260.Gentsch, C., Lichtsteiner, M. and Feer, H. (1990) Competition for Sucrose-Pellets in Triads ofMale Wistar Rats: Disinhibitory Effect of Individual Housing in Poor-Performing Rats.Behav. Brain Res. 38:19 - 24.Gershben, L.L., Benuck, I. and Shurrager, P.S. (1974) Influence of Stress on Lesion Growth andon Survival of Animals Bearing Parenteral and Intracerebral Leukemia L1210 and WalkerTumors. Oncology 30:429 - 435.Gerson, J.M., Tagliabue, A. and Herberman, R.B. (1981) Systemic and in situ Natural KillerActivity in Transplanted and Spontaneous Mammary Tumors in C3H/HeN Mice. J.Reticuloendothel. Soc. 29(1), 15-22.Geyer, S. (1992) Artifacts in "Limited Prospective" Designs? Testing Confounding Effects onResponse Behaviour of Women Prior to Breast Surgery. J. Psychosom. Res. 36:107 - 116.Ghoneun, M., Gill, G., Assanah, P. and Stevens, W. (1987) Susceptibility of Natural Killer CellActivity of Old Rats to Stress. Immunology 60:461 - 465.Criorda, R., Wiesberg, E.P., Tagge, E.P. and Trucco, M. (1992) Molecular Characterization of aNovel NK Cell Signal Transduction Molecule: NKR-P1. in NK Cell Mediated Cytotoxicity:Receptors, Signaling and Mechanisms. E. Lotzova and R.B. Herberman (Eds). CRC PressInc., Boca Raton, pp. 25 - 51.Giralt, M. and Armario, A. (1987) Individual Housing Does Not Influence the Adaptation of thePituitary Adrenal Axis and Other Physiological Variables to Chronic Stress in Adult MaleRats. Physiol. Behav. 45:477 - 481.Gorezynski R.M. (1987) Analysis of Lymphocytes in, and host environment of, mice showingconditioned immunosuppression to cyclophosphamide. Brain Behav. Immune. 1: 21-35.Greenberg, A.H. and Playfair, J.H. (1974) Spontaneously Arising Cytotoxicity to P-815-YMastocytoma in NZB Mice. Clin. Exp. Iinmunol. 15:251 - 259.Greenberg A.H., Dyck, D.G. and Osachuk, T.A.G. (1986) Tolerance to Drug-Induced (Poly I:C)Natural Killer Activation: Congruence with a Pavlovian Conditioning Model. J. Exp.Psycho!. 12:25 -31..Greenberg, A.H., Khalil, N., Pohajdak, B., Talgoy, M., Henkart, P. and Orr, F.W. (1986) NK-Leukocyte Chemotactic Factor (NK-LCF): A Large Granular Lymphocyte (LGL) Granule-Associated Chemotactic Factor. J. Immunol. 137(10), 3224-3230.Greenberg, A.H., Egan, S.E., Jarolim, L., Gingras, M. and Wright, J.A. (1987) NK CellRegulation of Implantation and Early Long Growth of H-ras Transformed 10T1/2 FibroblastsCancer Res. 47:4801 - 4805.Greer, S. and Morris, T. (1975) Pshchological Attributes of Women Who Develop Breast Cancer:A Controlled Study. J. Psychosom. Res. 19:147 - 153.Grossarth-Maticek, R., Bastiaans, J. and Kanazir, D.T. (1985) Psychosocial Factors as StrongPredictors of Mortality From Cancer, Ischaemic Heart Disease and Stroke: The YugosavProspective Study. J Psychosom Res 29:167 - 176.Grossarth-Maticek, R, and Eysenck, H.J. (1989) Length of Survival and Lymphocyte Percentagein Women with Mammary Cancer as a Function of Psychotherapy. Psychological Reports65: 315 -321.Habu, S., Fukui, H., Shimamura, K., Kasai, M., Nagai, Y., Okumura, K. and Tamaoki, N. (1981)In Vivo Effects of Anti-Asialo GMl. 1. Reduction of NK Activity and Enhancement ofTransplanted Tumor Growth in Nude Mice. J. Inununol. 127:34 - 38.Hahn, R.C. & Petitti, D.B. (1988) Minnesota Multiphasic Personality Inventory: RatedDepression and the Incidence of Breast Cancer. Cancer 61:845-848.Haller, 0., Hansson, M., Kiessling, R., and Wigzell, H. (1977) Role of Nonconventional NaturalKiller Cells in Resistance Against Syngeneic Tumor Cells In Vivo. Nature 270, 609-611.Hamed, H., Caleffi, M., Fentiman, I.S., Thomas, B. and Balbrook, R.D. (1991) Steroid Hormonesin Lymph and Blood From Women with Early Breast Cancer. Eur. J. Cancer 27:42 - 44.Harris D.T., Jaso-Friedmann, L. and Evans, D.L. (1992) Modulation of Function-AssociatedMolecules on Natural Killer Cells Alters Recognition and Cytotoxicity. in NK Cell MediatedCytotoxicity: Receptors, Signaling and Mechanisms. E. Lotzova and R.B. Herberman (Eds).CRC Press Inc., Boca Raton, pp. 54 - 65.196197Haslam, S.Z. (1987) Role of Sex Steroid Hormones in Normal Mammary Gland Functions. inThe Mammary Gland: Development, Regulation and Function. Neville, M.C. and Daniel,C. W. (Eds.), Plenum Press, New York, pp 355 - 381.Hatch, A.M., Wibery, G.S., Zawidzka, Z., Cann, M., Airth, J.M.and Grice, H.C. (1965) IsolationSyndrome in the Rat. Toxicol. Appl. Pharmacol. 7:737-745.Hatter, V.P.R., Snyder, R., Lucas, C.J., Foote, W. and Farrow, H.J. (1969) Clinical andPathological Correlation with Mammographic Findings in Lobular Carcinoma in Situ.Cancer 23: 826-839.Hayashi, I. and Sato, G.H. (1976) Replacement of Serum by Hormones Permits Growth of Cellsin a Defined Medium. Nature 259:132 - 134.Hayashi, Y., Aoki, Y., Eto, R. and Tokuoka, S. (1984) Findings of Myoepithelial Cells in HumanBreast Cancer. Ultrastructural and Immunohistochemical Study by Means of Anti-myosinAntibody. Acta. Pathol. Jpn. 34(3): 537-552.Hedrick, S.M. (1988) Specificity of the T Cell Receptor for Antigen. Advances in Immunology43: 193-234.Hennessey, M.B. and Levin, S. (1977) Effects of Various Habituation Procedures on Pituitary-Adrenal Responsiveness in the Mouse. Physiol. Behav. 18:799 - 802.Hennessy, J.W. and Levine, S. (1979) Stress, Arousal, and the Pituitary-Adrenal System: APsychoendocrine Hypothesis. in Progress in Psychobiology and Physiological Psychology,vol. 8. J. M. Sprague and A. N. Epstein (Eds), Academic Press Inc., New York, pp. 133 -178.Heppner, G., Miller, B., Cooper, D.N. and Miller, F.R. (1981) Growth Interactions BetweenMammary Tumor Cells. in N.J. Brenner, C.M. McGrath and M.A. Rich (Eds.), Cell Biologyof Breast Cancer, Academic Press, Inc., New York, pp. 161 - 172.Herberman, R.B. and Ortaldo, J.R. (1981) Natural Killer Cells: Their Role in Defence AgainstDisease. Science 214, 24-30.Herberman, R.B. (1985) in Mechanisms of Cytotoxicity by NK Cells, R.B. Herberman and D.M.Callewaert (Eds). Academic Press Inc., Orlando, pp. 1 - 13.Herberman, R.B. and Ortaldo, J.R. (1981) Natural Killer Cells: Their Role in Defense AgainstDisease. Science 214:24 - 30.Herring, G.M. and Kent, P.W. (1963) Some Studies on Mucosubstances of Bovine CorticalBone. Biochem. J. 89: 405-414.198Hiraoka, D., Nakamura, N., Nishizawa, Y., Uchida, N., Noguchi, S., Matsumoto, K. and Sato, B.(1987) Inhibitory and stimulatory effects of glucocorticoid on androgen-induced growth ofmurine Shionogi carcinoma 115 in vivo and in cell culture. Cancer Res. 47, 6560-6564.Ho, S-M., Yu, M., Leav, I. and Viccione, T. (1992) The Conjoint Actions of Androgens andEstrogens in the Induction of Proliferative Lesions in the Rat Prostate. in Li, J.J., Nandi, S.and Li, S.A. (Eds) Hormonal Carcinogenesis. Springer-Verlag, New York. pp 18 -26.Ho^man, R.A. (1988) Characterization of Natural Killer Activity in Sponge Matrix Allografts. J.Immunol. 140:1702 - 1710.House, J.S., Landis, K.R. and Umberson, D. (1988) Social Relationships and Health. Science241: 540 - 545Imagawa, W., Bandyopadhyay, G.K. and Nandi, S. (1990) Regulation of Mammary EpithelialCell Growth in Mice and Rats. Endocrine Reviews 11:494 - 523.Imagawa, W., Tomooka, Y., Hamamoto, S. and Nandi, S (1985) Stimulation of MammaryEpithelial Cell Growth in Vitro: Interaction of Epidermal Growth Factor and MammogenicHormones. Endocrinology 116:1514 - 1524.Imagawa, W., Bandyopadhyay, G.K., Garcia, M., Matsuzawa, A. and Nandi, S. (1992)Pregnancy-Dependent to Ovarian-Independent Progression in Mammary Tumors Delineatedin Primary Culture: Changes in Signal Transduction, Growth Factor Regulation and MatrixInteraction. Cancer Res. 52:6531 - 6538.Irwin, J., Ahluwali., P., Zacharko, R.M. and Anisman, H. (1986) Central Norepinephrine andPlasma-Corticosterone Following Acute and Chronic Stressors - Influence of Social-Isolationand Handling. Pharm. Biol. Behav. 24:1151 - 1154.Ito, Y., Mine, K., Ago, Y., Nakagawa, T., Fujiwara, M. and Ueki, S. (1983) Attack Stress andIgE Antibody Production in Rats. Pharmacol. Biochem. Behav. 19:883 - 886.Rosenman, R.H., Friedman, M., Straus, R., Wurm, M., Kositchek, R., Hahu, W. and Werthessen,N.T. (1964) A Predictive Study of Coronary Heart Disease. J Am Med Ass 189: 15 - 23.Jaaskelainen, J., Kalliomaki, P., Paetau, A., and Timonen, T. (1989) Effect of LAK Cells AgainstThree-Dimensional Tumor Tissue. In vitro Study Using Multi-Cellular Human GliomaSpheroids as Targets. J. Immunol. 142(3), 1036-1045.Jacobson L. and Sapolsky R. (1991) The Role of the Hippocampus in Feedback Regulation of theHypothalamic - Pituitary - Adrenocortical Axis. Endocrine Reviews 12: 118 - 134.199Jeng, M-H. and Jordan, V.C. (1991) Growth Stimulation and Differential Regulation ofTransforming Growth Factor-01 (TGF-131), TGF-132 and TGF-03 Messenger RNA Levels byNorethindrone in MCF-7 Human Breast Cancer Cells. Mol. Endocrinol. 5:1120 - 1128.Jensen, E.V. (1992) Current Concepts Sex Hormone Action (An Overview). in Li, J.J., Nandi, S.and Li, S.A. (Eds) Hormonal Carcinogenesis. Springer-Verlag, New York. pp 43 - 50.Johnson, H.M., Downs, M.O. and Pontzer, C.H. (1992) Neuroendocrine Peptide HormoneRegulation of Immunity. Chem. Immunol. 52: 49 - 83.Johnson, P.W., Stankova, J., Dexter, D. and Roder, J.C. (1990) The In Vivo Clearance of Ha-rasTransformants by Natural Killer Cells. Clin. Exp. Metast. 8:13 - 25.Julius, M.H., Simpson, E. and Herzenberg, L.A. (1973) A Rapid Method for the Isolation ofFunctional Thymus Derived Murine Lymphocytes. Eur. J. Immunol. 3:645 - 649.Justice, A. (1985) Review of the Effects of Stress on Cancer in Laboratory Animals: Importanceof Time of Stress Application and Type of Tumor. Psychol. Bull. 98:108 - 138.Kant, G.J., Bunnel, B.N., Mougey, E.H., Pennington, L.L. and Meyerhoff, J.L. (1983) Effects ofRepeated Stress on Pituitary Cyclic AMP, and Plasma Prolactin, Corticosterone and GrowthHormone in Male Rats. Pharm. Biochem. Behav. 18:967 - 971.Kant, G.J., Eggleston, T., Landman-Roberts, L., Kenion, C.C., Driver, G.C. and Meyerhoff, J.L.(1985) Habituation to Repeated Stress is Stressor Specific. Pharm. Biochem. Behav. 22:631-634.Kaplan, P. and Reynolds, G.A. (1990) Social Connections and Risk for Cancer: ProspectiveEvidence From the Alameda County Study. Behav. Med. 16:101 - 110.Kaplan, P. and Reynolds, G.A. (1988) Depression and Cancer Mortality and Morbidity:Prospective Evidence From the Alamada County Study. J. Behav. Med. 11:1 - 13.Kasai, M., Iwamori, M., Nagai, Y., Okumura, K. and Tada, T. (1980) A Glycolipid on theSurface of Mouse Natural Killer Cells. Eur. J. Immunol. 10, 175-180.Keehn, R.J., Goldberg, L.D. and Beebe, G.W. (1974) Twenty-four Year Mortality Follow-up ofArmy Veterans with Disability Separations for Psychoneurosisin 1944. Psychosom. Med.36:27 -46.Keehn, R.J. (1980) Follow-up Studies of World War II and Korean Conflict Prisoners. Am. J.Epidemiol. 111:194 - 211.Kelley K.W. (1991) Growth Hormone in Immunobiology. in Ader, R. Felton, D.L. and Cohen, N.(Eds) Psychoneuroimmunology 2nd edition. Academic Press Inc., San Diego. pp. 377 - 402.200Kiessling R., Klein, E. and Wigzell, H. (1975) "Natural" Killer Cells in the Mouse. 1. CytotoxicCells with Specificity for Mouse Moloney Leukemia Cells. Specificity and DistributionAccording to Genotype. Eur. J. Immunol. 5:112 - 117.Kissen,D.M. and Eysenck, H.J. (1962) Personality in Male Lung Cancer Patients. J. PsychosomRes. 6: 123 - 127.Kitamura, Y., Okamoto, S.,Hayata, I., Uchida, N., Yamaguchi, K. and Matsumoto,K. (1979)Development of androgen-indepent spindle cell tumors from androgen-dependent medullaryShionogi carcinoma 115 in androgen-depleted nude mice. Cancer Res. 39, 4713-4719Kitamura, Y., Okamoto, S., Hayata, I., Uchida, N., Yamaguchi, K. and Matsumoto, K. (1980)Effects of Serial Passage in Female Nude Athymic Mice on Androgen Dependency ofShionogi Carcinoma 115. Cancer Res. 40:4781-4785.Kittrell, F.S., Oborn, C.J. and Medina, D. (1992) Development of Mammary Preneoplasias inVivo From Mouse Mammary Epithelial Cell Lines in Vitro. Cancer Res. 52:1924 - 1932.Klein, J. S. and Kearns, R.J. (1989) Analysis of the Kinetics of Natural Killer Cell Activity in MiceDuring an Active Infection with Pseudomonas Aeruginosa. Nat. Inunun. Cell Growth Regul.8, 37-47.Kligman, D. and Hilt, D.C. (1988) The S100 Protein Family. Trends BioC. 13(11): 437-443.Knop, J. (1990) Immunologic Effects of Interferon. J. Invest. Dermatol. 95:72s - 74s.Koike, S. and Noumura, T. (1989) Effects of Post Weaning Differential Housing on SerumTestosterone Levels in Male Mice Through-Out Aging. Zool. Sci. 6: 351 - 358.Koo, G.C. and Peppard, J.R. (1984) Establishment of Monoclonal Anti-NK-1.1 Antibody.Hybredoma 3:301 -7.Korngold, R. and Doherty, P.C. (1985) Treatment of Mice with Polyinocinic-PolycytidylicPolyribonucleotide Reduces T-Cell Involvement in a Localized Inflamitory Response toVaccinia Virus Challenge. J. Virology 53:489 - 494.Kraut, R.P. and Greenberg, A.H. (1986) Effects of Endogenous and Exogenous Opioids onSplenic Natural Killer Cell Activity. Nat. Ifnmun. Cell Growth Regul. 5, 28-40.Kushner, B.H. and Cheung, N.-K. V. (1992) Absolute Requirement of CD11/CD18 AdhesionMolecules, FcRII and the Phosphatidylinositol-Linked FcRIII for Monoclonal Antibody-Mediated Neutrophil Antitumor Cytotoxicity. Blood 76:1484 - 1490.201Kuttenn, F., Gompel, A., Maler, C., Leygue, E., Baudot, N., Plu, G., Thalabard, J.-C. andMauvais-Jarvis, P. (1992) The Influence of Steroid Hormones and Antihormones on theGrowth and Differentiation of Normal Human Breast Cells in Culture. in Li, J.J., Nandi, S.and Li, S.A. (Eds) Hormonal Carcinogenesis. Springer-Verlag, New York. pp 145 - 156.Kvetnansky, R., McMarty, R., Thoa, N.B., Lake, C.R. and Kopin, I.J. (1979) Sympatho-adrenalResponses of Spontaneously Hypertensive Rats to Immobilization Stress. Am. J. Physiol.236:H457 - H462.Kvetnansky, R., Sun, C.L., Torda, T. and Kopin, I.J. (1977) Plasma Epinephrine andNorepinephrine Levels in Stressed Rats - Effects of Adrenalectomy. Pharmacol. 19:241.Labrie, F., Simard, J., de Launoit, Y., Poulin, R., Theriault, C., Dumont, M., Dauvois, S., Martel,C. and Li, S. (1992) Androgens and Breast Cancer. Cancer Detect. Prevent. 16:31 - 38.Lai, L.C., Ghilchik, M.W., Shaikh, N.A., Reed, M.J. and James, V.H.T. (1989) RelationshipBetween Epidermal Growth Factor and Dehydroepiandrosterone and It's Sulphate in BreastCyst Fluid. Br. J. Cancer 60:320 - ?.Lawler, E.M., Miller, F.R. and Heppner, G.H. (1983) Significance of Three-Dimensional GrowthPatterns of Mammary Tissues in Collagen Gels. In Vitro 19:600 - 610.Lazarus, R.S. (1971) The Concepts of Stress and Disease. in Society, Stress and Disease Volume1. L. Levi (Ed.), ??, pp. 53 - 58.Lea, 0.A., Kvinnsland, S. and Thorsen, T. (1989) Improved Measurement of AndrogenReceptors in Human Breast Cancer. Cancer Res. 49:7162 - 7167.Lemieux, S., Ouellet-Talbot, F., Lusignan, Y., Morelli, L., Labreche, N., Gosselin, P. andLecomte, J. (1991) Identification of Murine Natural Killer Cell Subsets with MonoclonalAntibodies Derived From 129 Anti-057BL/6 Immune Spleen Cells. Cellular Immunology134:191 -204.LeShan, L. (1959) Psychological States as Factors in the Development of Malignant Disease: ACritical Review. JNCI 22:1 - 18.Levenson, J.L. and Bemis, C. (1991) The Role of Psychological Factors in Cancer Onset andProgression. Psychosom. 32:124 - 132.Levine, S. and Coover, G.D. (1976) Environmental Control of Suppression of the Pituitary-Adrenal System. Physiol. Behav. 17:35 - 37.Levine, S., Weinberg, J. and Ursin, H. (1978) Definition of the Coping Process and Statement ofthe Problem. in Psychobiology of Stress: A Study of Coping Men. ?? (Ed), Academic PressInc., pp. 3 - 21.Levine, S., Weinberg, J. and Brett, L.P. (1979) Inhibition of Pituitary-Adrenal Activity as aConsequence of Consummatory Behavior. Psychoneuroendocrinology 4:275 - 286.Levy, S.M., Herberman, R., Maluish, A., Schlien, B. and Lippman, M. (1985) Prognostic RiskAssessment in Primary Breast Cancer by Behavioral and Immunological Parameters. HealthPsychol. 4:99 - 113.Levy, S.M., Lee, J., Bagley, C. and Lippman, M. (1988) Survival Hazards Analysis on FirstRecurrent Breast Cancer Patients: Seven-Year Follow-Up. Psychosomatic Med. 50:520 -528Levy, S.M., Herberman, R.B., Lippman, M., D'Angelo, T. and Lee, J. (1991) Immunological andPsychosocial Predictors of Disease Recurrence in Patients With Early-Stage Breast Cancer.Behavioral Medicine ??:67 - 75.Li, B.L., Zhao, X.X., Liu, X.Y., Kim, H.S., Raska, K, Ortaldo, J.R., Schwartz, B. and Pestka, S.(1990) a Interferon Structure and Natural Killer Cell Stimulatory Activity. Cancer Res.50:5328 - ??.Li, J.J., Mueller, G.C. and Sekely, L.I. (1991) Workshop Report From the Division of CancerEtiology, National Cancer Institute, National Institute of Health. Current Perspectives andFuture Trends in Hormonal Carcinogenesis. Cancer Res 51:3636 - 3629Liao S. (1992) Androgen Receptors and Regulation of Androgen Action in Normal and AbnormalCells. in Li, J.J., Nandi, S. and Li, S.A. (Eds) Hormonal Carcinogenesis. Springer-Verlag,New York. pp 11 - 17.Lillie (1954) Histopathologic Technic and Practical Histochemistry. Blakiston co. Inc. N.Y.Lippman, M.E. (1985) Psychosocial Factors and the Hormonal Regulation of Tumor Growth. inLevy S. (Ed) Behaviour and Cancer. Jossey-Bass Inc., San Franscico. pp 134-147.Lippman, M.E., Dickson. R.B., Gelmann, E.P., Rosen, N., Knabbe, C., Bates, S., Bonzert, D.,Huff, K. and Kasid, A. (1987) Growth-Regulation of Human-Breast Carcinoma OccursThrough Regulated Growth-Factor Secretion. J. Cell Biochem. 35:1 - 16.Liesveld, J.L., Frediani, K.F., Winslow, J.M., Duerst, R.E. and Abboud, C.N. (1991) CytokineEffects and Role of Adhesive Proteins and Fc Receptors in Human Macrophage-MediatedAntibody Dependent Cellular Cytotoxicity. J. Cell. Biochem. 45:381 - 390.Ljunggren, H.-G. and Kane, K. (1986) J. Immunogenet. 13:141 - 151.Ljunggren, H.-G. and Karre, K. (1990) In Search of Missing Self-MI-IC Molecules and NK CellRecognition. Immunol. Today 11:237 - 244.202203Luthy, I.A., Begin, D. and Labrie, F. (1988) Mediation by the Androgen Receptor of theStimulatory and Antiandrogenic Actions of 170 - Estradiol on the Growth of Androgen -Sensitive Shionogi Mammary Carcinoma Cells in Culture. Endocrinol. 123:1418 - 1424.Lyte, M., Nelson, S.G. and Baissa, B. (1990) Examination of the Neuroendocrine Basis for theSocial-Conflict-Induced Enhancement of Immunity in Mice. Physiol. Behav. 48:685 - 691.Mackintosh JH (1970) Territory Formation by Laboratory Mice. Animal Behav. 18:177 - 183.Madden, K.S. and Livnat, S. (1991) Catecholamine Action and Immunologic Reactivity. in Ader,R. Felton, D.L. and Cohen, N. (Eds) Psychoneuroimmunology 2nd edition. Academic PressInc., San Diego. pp.Markovic, S.N. and Murasko, D.M. (1990) Anesthesia Inhibits Poly I:C Induced Stimulation ofNatural Killer Cell Cytotoxicity in Mice. Clinical. Immunol. Immunopath. 56:202 - 209.Marsh, J.T., Miller, B.E. and Lamson, B.G. (1959) Effect of Repeated Breif Stress on Growth ofEhrlich Carcinoma in the Mouse. JNCI 22:961 - 977.Mason, J.W. (1975a) A Historical View of the Stress Field: Part I. J. Human Stress 1:6 - 12.Mason, J.W. (1975b) A Historical View of the Stress Field: Part II. J. Human Stress 1:22 - 36.Mason, J.W. (1959) Recent Prog. Horm. Res. 15:345 - 389.Mason, L.L., Mathieson, B.J. and Ortaldo, J.R. (1990) Natural Killer (NK) Cell Subsets in theMouse. NK-1.1+/LGL-1+ Cells Restricted to Lysing NK Targets, Whereas NK-1.1+/LGL-Cells Generate Lymphokine-Activated Killer Cells. J. Immunol. 145:751 - 759.Matsuzawa, A. (1986) Hormone Dependence and Independence of Mammary Tumors in Mice.Int. Rev. Cytol. 103:303 - 340.McGillis, J.P., Mastato, M. and Payan, D.G. (1991) Immunologic Properties of Substance P. inAder, R. Felton, D.L. and Cohen, N. (Eds) Psychoneuroimmunology 2nd edition. AcademicPress Inc., San Diego. pp. 209 - 224.McGinnes, K, Chapman, G. and Penny, R. (1988) Effects of Interferon on Natural Killer (NK)Cells Assessed by Fluorescent Probes and Flow Cytometry. J. Immunol. Meth. 107:129 -136.Metzler M., Pfeiffer E., Kohl W. and Schnitzler R. (1992) Interactions of Carcinogenic Estrogenswith Microtubular Proteins. in Li, J.J., Nandi, S. and Li, S.A. (Eds) HormonalCarcinogenesis. Springer-Verlag, New York. pp 86 -94.204Miczek, K.A., Thompson, M.L. and Shuster, L. (1982) Opioid-Like Analgesia in Defeated Mice.Science 215:1520 - 1522.Migliori, R.J., Gruber, S.A., Sawyer, M.D., Hoffman, R., Ochoa, A., Bach, F.H. and Simmons,R.L. (1987) Lymphokine-Activated Killer (LAK) Cells Can be Focused at Sites of TumorGrowth by Products of Macrophage Activation. Surgery 102(2), 155-161.Milas, L. and Scott, M.T. (1978) Antitumor Activity of Corynebacterium parvum. Adv. CancerRes. 26:257 - 306.Miller, B.E., Miller, F.R. and Heppner, G.H. (1981) Interactions Between Tumor SubpopulationsAffecting Their Sensitivity to the Antineoplastic Agents Cyclophosphamide andMethotrexate. Cancer Res 41:4378 - 4381.Miller, W.R. and O'Neill, J.S. (1990) The Significance of Steroid Metabolism in Human Cancer.J. Steroid BioChem. Molec. Biol. 37:317 - 325.Molomut N., Lazere, F. and Smith, L.W. (1963) Effect of Audiogenic Stress UponMethylcholanthrene-Induced Carcinogenesis in Mice. Cancer Res. 23:575 -577.Moore, J.W., Thomas, B.S. and Wang, D.Y. (1986) Endocrine Status and the Epidemiology andClinical Course of Breast Cancer. Cancer Surveys 5:537 - 559.Mosmann, T. (1983) Rapid colorimetric assay for cellular growth and survival: application toproliferation and cytotoxicity assays. J. Immunol. Meth. 65, 55-63.Mowry (1963) Special Value of methods that colour both acidic and vicenal hydroxyl groups inthe histochemical study of mucins with revised directions for the colloidal vein stain, the useof Alcian Blue 8GX and their combination with the periodic acid Schiff reaction. Ann. N.Y.Acad. Sci. 106 402-423.Munck A., Guyre P.M. and Holbrook N.J. (1984) Physiological Functions of Glucocorticoids inStress and Their Relation to Pharmacological Actions. Endocrine Reviews 5: 25 - 44.Nagler A., Lanier, L.L. and Phillips, J.H. (1988) The Effect of IL-4 on Human Natural KillerCells. A Potent Regulator of IL-2 Activation and Proliferation. J. Inununol. 141:2349 -2351.Naume, B., Gately, M. and Espevik, T. (1992) A Comparative Study of IL-12 (CytotoxicLymphocyte Maturation Factor)-, IL-2- and IL-7-Induced Effects on ImmunomagneticallyPurified CD56+ NK Cells. J. Immunol. 148:2429 - 2436.Newberry, B.H. (1978) Restraint-induced Inhibition of DMBA-induced Mammary Tumors:Relation to Stages of Tumor Development. JNCI 61:725 - 729.205Noguchi, S., Nishizawa, Y., Nakamura, N., Uchida, N., Yamaguchi, K., Sato, B., Kitamura, Y.and Matsumoto, K. (1987) Growth-stimulating effect of pharmacological doses of estrogenon androgen-dependent Shionogi carcinoma 115 in vivo but not in cell culture. Cancer Res.47, 263-268.Nohno, T., Omukai, Y., Watanabe, S., Saito, T. and Senoo, T. (1982) Effects of Estrogens andAntiestrogens on Androgen - Dependent Growth of Shionogi Carcinoma 115: Role ofEstrogen Receptor. Cancer Lett. 15:237 - 244.Nohno, T., Watanabe, S. and Saito, T. (1986) Evaluation of Effect of Host Immunity on Growthof Androgen-Dependant Shionogi Carcinoma 115 in the Mouse. Cancer Lett. 33, 125-130.Nutt, J.E., Harris, A.L. and Lunee, J. (1991) Phorbol Ester and Bryostatin Effects on Growth andthe Expression of Oestrogen Responsive and TGF-131 Genes in Breast Tumor Cells. Br. J.Cancer 64:671 - 676.Olivier B., Mos, J., van der Heyden, J. and Hartog, J. (1989) Serotonergic Modulation of SocialInteractions in Isolated Male Mice. Psychopharmicol. 97:154-6,Omukai, Y., Nakamura, N., Hiraoka, D., Nishizawa, Y., Uchida, N., Noguchi, S., Sato, B. andMatsumoto, K. (1987) Growth-stimulating effect of pharmacological doses of glucocorticoidon androgen-responsive Shionogi carcinoma 115 in vivo in mice and in cell culture. CancerRes. 47, 4329-4334.Ormandy, C.J., Clarke, C.L., Kelly, P.A. and Sutherland, R.L. (1992) Androgen Regulation ofProlactin-Receptor Gene Expression in MCF-7 and MDA-MB-453 Human Breast CancerCells. IM. J. Cancer 50:777 - 782.Ortaldo, J.R., Mason, A., Rehberg, E., Moschera, J., Kelder, B., Pstka, S. and Herberman, R.B.(1983) Effects of Recombinant and Hybred Recombinant Human Leukocyte Interferons onCytotoxic Activities of Natural Killer Cells. J. BiolChem. 258:15011 - 15015.Ortaldo J.R., Mason, A.T., Gerard, J.P., Henderson, L.E., Farrar, W., Hopkins, R.F., Herberman,R.B. and Rabin, H. (1984) Effects of Natural and Recombinant IL-2 on Regulation of IFN-Gamma Production and Natural Killer Cell Activity. Lack of Involvement of the Tac Antigenfor These Immunoregulatory Effects. J. Immunol. 133:779 - 783.Ortaldo, J.R., Young, H.A., and Varesio, L. (1989) Modulation of CD3- Large GranularLymphocyte Functions by Agonists and Antagonists of Protein Kinase C: Effects on NK andLymphokine Activated Killer Activity and Production of IFN-_ J. Immunol. 143(1), 366-371.Ortiz, R, Armario, A. and Castellanos, J.M. (1984) Post-Weaning Differential Housing andTestosterone Secretion in Male Mice. Experientia 40:1428 - 1429.206Ortiz, R., Armario, A., Castellanos, J.M. and Balasch, J. (1985) Post-weaning crowding inducescorticoadrenal hyperreactivity in male mice. Physiol. Behav. 34, 857 - 860.Ottaway, C.A. (1991) Vasoactive Intestinal Peptide and Immune Function. in Ader, R. Felton,D.L. and Cohen, N. (Eds) Psychoneuroimmunology 21 edition. Academic Press Inc., SanDiego. pp. 225 - 262.Papotti, M., Coda, R., Ottinetti, A. and Bussolati, G. (1988) Dual Secretory and MyoepithelialDifferentiation in the Transplantable R3230AC Rat Mammary Carcinoma. Virchows ArchivB Cell Pathol. 55: 39-45.Pelletier, H., Olsson, N.O., Shimizu, T., Lagadec, P., Fady, C., Reisser, D. and Jeannin, J.-F.(1987) In Vitro Natural Killer Activity Against Progressive and Regressive Variants of a RatColon Adenocarcinoma. Effect of Treatments with Anti-Asialo GM1 Plus Complement.Immunobiol 175:202 - 213.Peng, X, Lang, C.M., Drozdowicz, C.K. and Ohlsson-Wilhelm, B.M. (1989) Effect of CagePopulation Density on Plasma Corticosterone and Peripheral Lymphocyte Populations ofLaboratory Mice. Lab. Animals 23:302 - 306.Platica, M., Doody, J., Mandeli, J.P. and Hollander, V.P. (1991) Role of Prolactin in Growth ofthe Rat Mammary Tumor MTW9. Int. J. Cancer 48:109 - 112.Plaut, P., Freidman, S.B., Ader, R., Ritterson, A.L. and Grota, L.J. (1969) Social Factors andResistance to Malaria in Mouse. Psychosom. Med. 31:453 -.Porritt, D. (1979) Social Support in Crisis: Quantity of Quality? Soc. Sci. Med. 13A:715-721.Prehn, R.T. and Main, J.M. (1957) Immunity to Methylcholanthrene-Induced Sarcomas. J. Nat.Inst. Cancer 18:769 - 778.Pross, H.F., Sterns, E. and MacGillis, (1984) Natural Killer Cell Activity in Women at High Riskfor Breast Cancer, With and Without Benign Breast Syndrome. Int. J. Cancer 34:303 - 308.Raab, A, Dantzer, R., Michaud, B, Mormede, P., Taghzouti, K., Simon, H. and LeMoal, M.(1986) Behavioural, Physiological and Immunological Consequences of Social Status andAggression in Chronically Coexisting Resident-Intruder Dyads of Male Rats. Physiol. Behav.36:223 - 228.Rabin, B.S., Lyte, M., Epstein, L.H. and Caggiula, A.R. (1987a) Alteration of ImmuneCompetency by Number of Mice Housed per Cage. Ann. New York Acad. Sci. 496:492 -500.Rabin, B.S., Lyte, M. and Hamill, E. (1987b) The Influence jof MOuse Strain and Housing on theImmune Response. J. Neuro Immun. 17:11 - 16.207Rabinowich, H., Vitolo, D., Altarac, S., Herberman, R.B. and Whiteside, T.L. (1992) Role ofCytokines in the Adoptive Immunotherapy of an Experimental Model of Human Head andNeck Cancer by Human IL-2-Activated Natural Killer Cells. J. Immunol. 149:340 - 349.Ramirez, A.J., Craig, T.K.J., Watson, J.P., Fentiman, I.S., North, W.R.S. and Rubens, R.D.(1989) Stress and Relapse of Breast Cancer. Br. Med. J. 298:291 - 293.Ray, P. and Pradhan, S.N. (1974) Breif Communication: Growth of Transplanted and InducedTumors in Rats Under a Shedule of Punished Behaviour. JNCI 52:575 - 577.Reid, P.E., Volz, D., Park, C.M., Owen, D.A. and Dunn, W.L. (1987) Methods for theIdentification of Side Chain 0-acyl Substituted Sialic Acids and for the SimultaneousVisualization of Sialic Acid, its Side Chain 0-acyl Varients and 0-sulphate Esters. ShortCommunication. Histochem. J. 19: 396-398.Reisine T., Affolter H.U., Rougon G. and Barbet J. (1986) New Insights into the MolecularMechanisms of Stress. Trends in Neuroscience 9:574 - 579.Reiss et al., 1983) J. Surg. Oncol. 22:249 - 253.Restrepo, C. and Armario, A. (1989) Comparison of Crowding and Food Restriction Effects onGrowth, Body Weight Gain and Endocrine Status in the Rat. Reprod. Nutr. Dev. 29:339 -345.Riley, V. (1981) Psychoendocrine influences on immunocompetence and neoplasia. Science 212,1100-1109.Riley, V., Fitzmaurice, M.A. and Spackman, D.H. (1981) Psychoneuroimmunologic Factors inNeoplasia: Studies in Animals. in Psychoneuroimmunology. R. Ader (Ed.), Academic PressInd., New York, pp. 31 - 101.Rivier, C. and Rivest, S. (1991) Effect of Stress on the Activity of the Hypothalamic-Pituitary-Gonadal Axis: Peripheral and Central Mechanisms. Biol. Reprod. 45:523 - 532.Rizzino, A. and Sato, G.H. (1978) Growth of Embryonal Carcinoma Cellsin Serum-Free Medium.Proc. Natl. Acad. Sci. USA 75:1844 - 1848.Roberts, P.J. and Hayry, P. (1976) Sponge Matrix Mlografts. A Model for Analysis of KillerCells Infiltrating Mouse Allografts. Transplantation 21:437 - 445.Robertson, M.J. and Ritz, J. (1990) Biology and Clinical Relevance of Human Natural KillerCells. Blood 76:2421 - 2438.Robertson, M.J. and Ritz, J. (1992) Role of IL-2 Receptors in NK Cell Activation andProliferation. in NK Cell Mediated Cytotoxicity: Receptors, Signaling and Mechanisms. E.Lotzova and R.B. Herberman (Eds). CRC Press Inc., Boca Raton, pp. 183 - 207.208Robertson, M.J., Soiffer, R.J., Wolf, S.F., Manley, T.J., Donahue, C., Young, D., Herrmann, S.H.and Ritz, J. (1992) Response of Human Natural Killer (NK) Cells to NK Cell StimulatoryFactor (NKSF): Cytolytic Activity and Proliferation of NI( Cells Are Differentially Regulatedby NKSF. J. Exp. Med. 175:779 - 788.Rogers, M.P., Reich, P., Strom, T.B. and Carpenter, C.B. (1975) Behaviorally conditionedimmunosuppression: replication of a recent study. Psychosom. Med. 38: 447-451.Rosenman, R.H., Friedman, M., Straus, R., Wurm, M., Kositchek, R., Hahn, W. and Wethessen,N.T. (1964) A Predictive Study of Coronary Heart Disease. J. Amer. Med. Assoc. 189:15 -23.Rossier, J., French, E.D., Rivier, C., Ling, N., Guillemin, R. and Bloom, F.E. (1977) Foot-ShockInduced Stress Increases 0-Endorpin levels in Blood but not in Brain. Nature 270:618 - 620.Rowan, R. (1992) Alterations in Spleen Cell Proliferation and Splenic Natural Killer Cell ActivityInduced by Changes in the Shionogi Mouse Mammary Carcinoma (SC 115) in Response toPsychosocial Stressors. M.Sc. Thesis, University of British Columbia.Rowse, G.J., Rowan, R.E., Worth, A.J., Reid, P.E., Weinberg, J., and Emerman, J.T. (1990) Ahistological study of the Shionogi adenocarcinoma 115 grown in male and female mice.Histology and Histopathology 5, 485-491.Rowse, G.J.., Rowan, R.E., Weinberg, J. and Emerman, J.T. (1990) Alterations in SplenicNatural Killer Cell Activity Induced by the Shionogi Mouse Mammary Tumor. Cancer Lett.54:81 - 87.Rowse, G.J., Weinberg, J., Bellward, G.D. and Emerman, J.T. (1992) Endocrine Mediation ofPsychosocial Stressor Effects on Mouse Mammary Tumor Growth. Cancer Lett. 65:85 - 93.Sakukura, T. (1991) New Aspects of Stroma-Parenchyma Relations in Mammary GlandDifferentiation. Int. Rev. Cytol. 125:165 - 202.Salvin, S.B., Rabin, B.S. and Neta, R. (1990) Evaluation of Immunological Assays to Determinethe Effects of Differential Housing on Immune Reactivity. Brain, Behav. Immun. 4:180 -188..Sapolsky, R.M. (1986) Stress-Induced Elevation of Testosterone Concentrations in High RankingBaboons: Role of Catecholamines. Endocrinol. 118:1630 - 1635.Savary, C.A. and Lotzova, E. (1987) Mechanism of Decline of Natural Killer Cell Activity inCorynebacterium parvum-Treated Mice: Inhibition by Erythroblasts and Thy 1.2+Lymphocytes. J.N.C.I. 79(3), 533-541.209Sayegh, J.F., Kobor, G., Lajtha, A. and Vadasz, C. (1990) Effects of Social Isolation and theTime of Day on Testosterone Levels in Plasma of C57BL/6By and Balb/cBy Mice. Steroids55:79 - 82.Schmale, A.H. and Iker, H.P. (1966) The Affect of Hopelessness and the Development of Cancer.1. Identification of Uterine Cervical Cancer in Women with Atypical Cytology. Psychosom.Med. 28: 714 - 721.Schreiber, H., Ward, P.L., Rowley, D.A. and Stauss, H.J. (1988) Unique Tumor-SpecificAntigens. Ann. Rev. Immunol. 6:465 - 483.Schwamberger, G., Flexch, I. and Ferber, E. (1991) Tumoricidal Effector Molecules of MurineMacrophages. Pathobiol. 59:248 - 253.Seaman, W.E., Sleisenger, M., Eriksson, E. and Koo, G.C. (1987) Depletion of Natural KillerCells in Mice by Monoclonal Antibody to NK-1.1. Reduction in Host Defense AgainstMalignancy Without Loss of Cellular or Humoral Immunity. J. Inununol. 138:4539 - 4544.Secreto, G., Toniolo, P, Berrino, F., Recchione, C., Cavalleri, A., Pisani, P., Totis, A., Fariselli,G. and Di Pietro, S. (1991) Serum and Urinary Androgens and Risk of Breast Cancer inPostmenopausal Women. Cancer Res. 51:2572 - 2576.Secreto, G., Toniolo, P., Pisani, P., Rechione, C., Cavalleri, A., Fariselli, G., Totis, A., Di Pietro,S. and Berrino, F. (1989) Androgens and Breast Cancer in Premenopausal Women. CancerRes. 49:471 - 476.Selye, H. (1975) Confusion and Controversy in the Stress Field. J. Human Stress 1:37 - 44.Shekelle, R.B., Raynor, W.J., Ostfeld, A.M. (1981) Psychological Depression and 17-Year Riskof Death From Cancer. Psychosom. Med. 43:435 - 449.Sickles, E.A. (1980) Further Experience with Microfocal Spot Magnification Mammography inthe Assesment of Clustered Microcalcifications. Radiology 137: 9-14.Simpson-Hen-en, L. and Lloyd, H.H. (1970) Kinetic parameters and growth curves forexperimental tumor systems. Cancer Chemother. Rep. 54, 143-174.Sklar, L.S. and Anisman, H. (1980) Social Stress Influences Tumor Growth. Psychosom. Med.42:347 - 365.Sklar, L.S. and Anisman, H. (1981) Stress and Cancer. Psycho]. Bull. 89:369 - 406.Sluyser, M (1987) Oncogenes and Hormones in Mouse Mammary Tumors. in Growth Factorsand Oncogenes in Breast Cancer, Sluyser, M. (Ed.), VCH, Weinheim, pp 123 - 141.210Smith, EM., Brosnan, M.P., Meyer W.J. and Blalock J.E. (1987) A Corticotropin Receptor onHuman Mononuclear Lymphocytes: Correlation with Adrenal ACTH Receptor Activity. N.Engl. J. Med. 317: 1266 - 1269.Smith E.M., Galin F.S., LeBouef R.D., Coppenhaver D.H., Harbour D.V. and Blalock J.E.(1990) Nucleotide and Amino Acid Sequence of Lymphocyte-Derived Corticotropin:Endotoxin Induction of a Truncated Peptide. Proc. Natl, Acad. Sci USA 87: 1057-1060.Soloman, G.F. and Amaukraut, A.A. (1981) Psychoneuroendocrinological Effects on the ImmuneResponse. Ann. Rev. Microbiol. 35:155 - 184.Solvason, H.B., Ghanta, V.K. and Horamoto, R.N. (1988) conditioned Augmentation of NaturalKiller Cell Activity. Independence from Nociceptive Effects and Dependence on Interferon-n.J. Immunol. 140: 661 - 665.Spiegel, D., Kraemer, H.C., Bloom, J.R. and Gottheil, E. (1989) Effect of PsychosocialTreatment on Survival of Patients with Metastic Breast Cancer. Lancet 2:888 - 891.Steplewski, Z., Goldman, P.R. and Vogel, W.H. (1987) Effect of Housing Stress on theFormation and Development of Tumors in Rats. Cancer Lett. 34, 257-261.Steplewski, Z., Vogel, W.H., Ehya, H., Porvpatich, C. and Smith, J.M. (1985) Effects of RestaintStress on Inoculated Tumor Growth and Immune Responses in Rats. Cancer Res. 45:5128 -5133.Stitz, L., Baenziger, J., Pircher, H., Hengartner, H. and Zinkernagel, R.M. (1986) Effect of RabbitAnti-Asialo GM1 Treatment In Vivo or with Anti-Asialo GM1 Plus Complement In Vitro onCytotoxic T Cell Activities. J. Immunol. 136:4674 - 4680.Stout R.D., Schwarting, G.A. and Suttles, J. (1987) Evidence that Expression of Asialo-GM1May Be Associated with Cell Activation. Correlation of Asialo-GM1 Expression withIncreased Total Cellular RNA and Protein Content in Normal Thymocyte and Spleen CellPopulations. J Immunol. 139:2123 - 2129.Street,N.E. and Mosman, T.R. (1991) Function Diversity of T Lymphocytes Due to Secretion ofDifferent Cytokine Patterns. FASEB 5:171 - 177.Stroup, R.M. and Pinkus, G.S. (1988) S-100 Immunoreactivity in Primary and MetastaticCarcinoma of the Breast: A Potential Source of Error in Immunodiagnosis. Human Pathol.19: 949-953.Suttles, J., Schwarting, G.A. and Stout, R.D. (1986) Flow Cytometric Analysis Reveals thePresence of Asialo GMion the Surface Membrane of Alloimmune Cytotxic T Lymphocytes.J. Immunol. 136:1591 -.211Suttles, J., Schwarting, G.A., Hougland, M.W. and Stout, R.D. (1987) Expression of Asialo GM1on a Subset of Adult Murine Thymocytes: Histological Localization and Demonstration Thatthe Asialo GM1-Postive Subset Contains Both the Functionally Mature and the ProliferatingThymocyte Subpopulations. J. Inununol. 138:364 - 372.Talmadge, J.E., Meyers, K.M., Prieur, D.J. and Starkey, J.R. (1980) Role of NK Cells in TumorGrowth and Metastasis in Beige Mice. Nature 284:622 - 624.Talmadge, J.E., Adams, J., Philips, H., Collins, M., Lenz, B., Schneider, M., Schlick, E.,Ruffmann, R., Wiltrout, R.H. and Chirigos, M.A. (1985) Immunomodulatory Effects in Miceof Polyinosinic-Polycytidylic Acid Complexed with Poly-L-Lysine andCarboxymethylcellulose. Cancer Res. 45:1058 - 1065.Talmadge, J.E., Herberman, R.B., Chirigos, M.A., Maluish, A.E., Schneider, M.A., Adams, J.S.,Philips, H., Thurman, G.B., Varesio, L., Long, C., Oldham, R.K. and Wiltrout, R.H. (1985)Hyporesponsiveness to Augmentation of Murine Natural Killer Cell Activity in DifferentAnatomical Compartments by Multiple Injections of Various Immunomodulators IncludingRecombinant Interferons and Interleukin 2. J. Immunol. 135, 2483-2491.Tanaka, A., Matsumoto, K., Nishizawa, Y., Lu, J., Yamanishi, H., Maeyama, M., Nonomura, N.,Uchida, N. and Sato, B. (1990) Growth Stimulation by Androgens, Glucocorticoids orFibroblast Growth Factors and the Blocking of the Stimulated Growth by Antibody AgainstBasic Fibroblast Growth Factor in Protein - Free Culture of Shionogi Carcinoma 115 Cells.J. Steroid Biochem. Molec. Biol. 37:23 - 29.Taylor, P., Abrams, D. and Hewstone, M. (1988) Cancer, Stress and Personality: A CorrelationalInvestigation of Life-Events, Repression-Sensitization and Locus of Control. Br. J. Med.Psychol. 61:179 - 183.Tejwani, G.A., Gudehithlu, K.P., Hanissian, S.H., Gienapp, I.E., Whitacre, C.C. and Malarkey,W.B. (1991) Facilitation of Dimethylbenz[a]anthracene-Induced Rat MammaryTumorigenesis by Restraint Stress: Role of 0-Endorphin, Prolactin and Naltrexone.Carcinogen. 12:637 - 641.Teller, M.N., Stock, C.C., Stohr, G., Merker, P.C., Kaufman, R.J., Escher, G.C. and Bowie, M.(1966) Biologic Characteristics and Chemotherapy of 7,12-Dimethylbenz[a]anthracene-Induced Tumors in Rats. Cancer Res. 26:245 - 252.Temosholc, L. and Fox, B.H. (1984) Coping Styles and Other Psychosocial Factors Related toMedical Status and to Prognosis in Patients with Cutaneous Malignant Melanoma. in Impactof Psychoendocrine Systems in Cancer and Immunity. B.H. Fox and B.H. Newberry (Eds),C.J. Hogrefe Inc., Lewiston, pp.258 - 287.The Cooperative Breast Cancer Group (1961). Progress report: results of studies by thecooperative breast cancer group 1956-1961. Cancer Chemother. Rep. 11:109 - 141.212Thomas, C.B., Duszyhski, K.R. and Shaffer, J.W. (1979) Family Attitudes Reported in Youth asPotential Predictors of Cancer. Psychosom. Med. 41: 287 - 302.Thompson, T.C., Egawa, S., Kadmon, D., Miller, G.J., Tinune, T.L., Scardino, P.T. and Park,S.H. (1992) Androgen Sensitivity and Gene Expression in Ras + Myc -Induced MouseProstate Carcinomas. J. Steroid BioChem. Molec. Biol. 43: 79 - 85.Topper, Y.J. and Freeman, C.S. (1980) Multiple Hormone Interactions in the DevelopmentalBiology of the Mammary Gland. Physiol. Rev. 60:1049 - 1106.Trinchieri, G., Matsumoto-Kobayashi, M., Clark, S.C., Seehra, J., London, L. and Perussia, B.(1984) Response of Resting Human Peripheral Blood Natural Killer Cells to Interleukin-2. J.Exp. Med. 160:1147 - 1169.Trinchieri,G. (1989) Biology of Natural Killer Cells. Adv. Immunol. 47:187 - 376.Unanue, E.R. and Cerottini, J.C. (1989) Antigen Presentation. FASEB Journal 3: 2496-2502.Valzelli, L. (1973) The isolation syndrome in mice. Psychopharmacolgia 31, 305-320.van der Ploeg, H.M., Kleijn, W.C., Mook, J., van Donge, M., Pieters, A.M.J. and Leer, J.W.H.(1989) Rationality and Antiemotionality as a Risk Factor for Cancer: Concept Differentiation.J Psychosom Res 33:217,Vander Burg, B., De Groot, R.P., Isbrucker, L., Kruijen, W. and DeLact, S.W. (1992) DirectStimulation by Estrogen of Growth Factor Signal Transduction Pathways in Human BreastCancer Cells. J. Steroid BioChem. Molec. Biol. 43: 111 - 115.Vessey, S.H. (1964) Effects of Grouping on Levels of Circulating Antibodies in Mice. Soc. Exp.Biol. Med. 115:252 - 255.Viveros 0.H., Diliberto, E.J., Hazum, E. and Chang, K.-J. (1979) Opiate-Like Materials in theAdrenal Medula: Evidence for Storage and Secretion with Catecholamines. Molec.Pharmacol. 16:1101 - 1108.Vogelstein, B. (1989) Allotype of Colorectal Cancer. Science 244:207 - 211.Volz, D., Reid, P.E., Park, C.M., Owen, D.A. and Dunn, W.L. (1987) A New HistochemicalMethod for the Selective Periodate Oxidation of Total Tissue Sialic Acids. Histochem. J. 19:396-398.Wagner, R.K. and Jungblut, P.W. (1976) Oestradiol- and Dihydrotestosterone-Receptors inNormal and Neoplastic Human Mammary Tissue. Acta Endocrinol. 82:105 - 120.213Watanabe, S., NoImo, T., Omukai, Y.M., Saito, T. and Senoo, T. (1982) Stimulatory Effects ofDexamethasone and Indomethacen on Growth of Androgen-Dependent Shionogi Carcinoma115 in the Mouse. Cancer Lett. 16, 261-266.Waxier-Morrison, N., Hislop, T.G., Mears, B. and Kan, L. (1991) Effects of Social Relationshipon Survival for Women with Breast Cancer: A Prospective Study. Soc. Sci. Med. 33:177 -183.Wayner, E.A., Flannery, G.R. and Singer, G. (1978) The Effects of Taste Aversion Conditioningon the Primary Antibody Response to Sheep Red Blood Cells and Brucella abortus in theAlbino Rat. Physiol. Behav. 21: 995 - 1000.Wei, W.-Z. and Heppner, G. (1987) Natural Killer Cell Activity of Lymphocytic Infiltrates inMouse Mammary Lesions. Br. J. Cancer 55, 589-594.Wei, W.-Z., Fulton, A. Winkelhake, J. and Heppner, G. (1989) Correlation of Natural KillActivity with Tumorigenesis of a Preneoplastic Mouse Mammay Lesion. Cancer Res.49:2709 -2715.Weinberg, J., Erskine, M. and Levine, S. (1980) Shock-Induced Fighting Attenuates the Effectsof Prior Shock Experience in Rats. Physiol. Behav. 25:9 - 16.Weinberg, J. and Levine, S. (1980) Psychobiology of Coping in Animals: The Effects ofPredictability. in Coping and Health. S. Levine and H. Ursin (Eds), Plenum Press, NewYork, pp. 39 - 59.Weinberg, J. and Wong, R. (1983) Consummatory Behavior and Adrenocortical Responsivenessin the Hamster. Physiol. Behav. 31:7 - 12.Weinberg, J. and Bezio, S., (1987) Alcohol-Induced Changes in Pituitary-Adrenal Activity DuringPregnancy. Alcoholism: Clin. Expt. Res. 11:274 - 280.Weinberg, J. and Emerman, J.T. (1989) Effects of Psychosocial Stressors on Mouse MammaryTumor Growth. Brain, Behavior, and Immunity 3, 234-246.Weiss, J.M., Pohorecky, L.A., Salman, S. and Gruenthal, M. (1976) Attenuation of GastricLesions by Psychological Aspects of Aggression in Rats. J. Comp. Physiol. Psycho!. 90:252-259.Welsh, R.M. (1984) Natural Killer Cells and Interferon. CRC C.R. Immun. 5:55 - 93.Welsh, C.W. (1985) Host Factors Affecting the Growth of Carcinogen-Induced Rat MammaryCarcinomas: A Reveiw and Tribute to Charles Brenton Huggins. Cancer Res. 45:3415 -3443.214Whitlock F & Siskand M (1979) Depression and Cancer : a follow-up study PsychologicalMedicine 9:747-752.Williams, R.B. and Eichelman, B. (1971) Social Setting: Influence on the Physiological Responseto Electric Shock in the Rat. Science 174:613 - 614.Wirsching, M, Stierlin, H., Hoffinan, F., Weber, G. and Wirsching, B. (1982) PsychologicalIdentification of Breast Cancer Patients Before Biopsy. J. Psychosom. Res. 26: 1 - 10.Yamaguchi N. (1992) Sympathoadrenal System in Neuroendocrine Control of Glucose:Mechanisms Involved in the Liver, Pancreas and Adrenal Gland Under Hemorrhagic andHypoglycemic Stress. Canadian Journal of Physiology and Pharmacology 70: 167 - 206.Yamanishi, H., Nonomura, N., Tanaka, A., Yasui, T., Nishizawa, Y., Matsumoto, K. and Sato, B.(1990) Roles of Transforming Growth Factor 13 in Inhibition of Androgen - Induced Growthof Shionogi Carcinoma Cells in Serum - Free Medium. Cancer Res. 50:6179 - 6183.Yates, J. and King, R.J.B. (1978) Multiple sensitivities of mammary tumor cells in culture.Cancer Res. 38, 4135-4137.Yates, J., Couchman, J.R. and King, R.J.B. (1980) Androgen Effects on Growth, Morphologyand Sensitivity of SC 115 Mouse Mammary Tumor Cells in Culture. in S. Iacobelli et al.(Eds.). Hormones and Cancer. Raven Press, New York, pp. 31 - 39.Zangemeister-Wittke, U., Kyewski, B. and Schirrmacher, V. (1989) Recruitment and Activationof Tumor-Specific Immune T Cells in Situ. CD8+ Cells Predominate the SecondaryResponse in Sponge Matrices and Exert Both Delayed-Type Hypersensitivity-Like andCytotoxic T Lymphocyte Activity. J. Immunol. 143(1), 379-385.Zielinski, C. C., Tichatschek, E., Muller, C. 1(989) Association of Increased Lytic Effector CellFunction with High Estrogen Receptor Levels in Tumor Bearing Patients with Breast Cancer.Cancer 63:1985 - 1989.

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 9 1
Japan 7 0
China 6 1
Germany 4 9
India 2 0
United Kingdom 2 0
Zambia 1 0
Morocco 1 0
Nigeria 1 0
City Views Downloads
Tokyo 6 0
Unknown 6 6
Ashburn 4 0
Putian 4 0
London 2 0
Sunnyvale 2 0
Delhi 2 0
Mountain View 1 0
Shenzhen 1 1
Kansas City 1 0
Cambridge 1 0
Zaria 1 0
Sesheke 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086495/manifest

Comment

Related Items