@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Medicine, Faculty of"@en, "Medical Genetics, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Orban, Paul C."@en ; dcterms:issued "2008-09-18T16:40:46Z"@en, "1993"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "This thesis is in two discrete parts. The first part deals with the mechanism of auto-stimulatory growth in the pathogenesis of malignant neoplasia. Experiments were undertaken to investigate the question of whether antagonism of growth-factor activity can lead to the death of the cells of a murine model of auto-stimulatory leukemia. Although many previous workers have examined the ability of growth-factor antagonists to inhibit the growth of the cells of human leukemias and of animal leukemia models, none have documented the complete blockage of growth, and none have documented the death of such cell populations. A model of auto-stimulatory leukemia was generated by transfecting a mouse IL-2-dependent cell line with vectors designed to cause expression of IL-2 in these cells. One series of clones was derived which grew in the absence of exogenous IL-2, and produced tumours in syngeneic mice. Cells of these clones produced very small amounts of IL-2, but their growth was not completely inhibitable by antibodies to IL-2 or the IL-2 receptor. Another clone was derived, which produced no detectable IL-2, but grew independently of exogenous IL-2. The growth of cells of this clone was completely inhibited by antibody, and death of the cells resulted. The experiments described here represent the first demonstration that antibody antagonists of the growth factor can induce the death of cells that grow by auto-stimulatory mechanisms. They support the hope that cytokine antagonists may find use as therapeutic reagents in the treatment of auto-stimulatory neoplasms. The second part of the thesis presents a new technique in gene targetting, which has broad applications in the study of gene function. Using a bacterial recombinase under the control of a tissue-specific developmentally regulated promoter, transgenic animals were derived, in which a target gene was deleted in thymocytes and their daughter cells, but no other tissues. This technique circumvents the impediment embryonic lethality may present in some gene deletion experiments, and allows questions of tissue-specifc function of genes in pathogenesis and normal development to be addressed."@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/2247?expand=metadata"@en ; dcterms:extent "9322313 bytes"@en ; dc:format "application/pdf"@en ; skos:note "TWO INVESTIGATIONS IN MOLECULAR PATHOPHYSIOLOGYbyPAUL CHRISTOPHER ORBANB. Med. Sc., The University of New South Wales, 1981M.B., B.S., The University of New South Wales, 1984BA., The University of Sydney, 1984A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES( Genetics Programme )We accept thi thesis as conformingto the ^required standardTHE UNIVERSITY OF BRITISH COLUMBIAAugust 1993© Paul Christopher Orban, 1993In 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.(Signature) Department of ^itA-c_occAL^C.rt 4_3 The University of British ColumbiaVancouver, CanadaDate^3 ,o^tq /3DE-6 (2/88)iiABSTRACTThis thesis is in two discrete parts. The first part deals with the mechanism of auto-stimulatorygrowth in the pathogenesis of malignant neoplasia. Experiments were undertaken to investigatethe question of whether antagonism of growth-factor activity can lead to the death of the cells of amurine model of auto-stimulatory leukemia. Although many previous workers have examined theability of growth-factor antagonists to inhibit the growth of the cells of human leukemias and ofanimal leukemia models, none have documented the complete blockage of growth, and nonehave documented the death of such cell populations. A model of auto-stimulatory leukemia wasgenerated by transfecting a mouse IL-2-dependent cell line with vectors designed to causeexpression of IL-2 in these cells. One series of clones was derived which grew in the absence ofexogenous IL-2, and produced tumours in syngeneic mice. Cells of these clones produced verysmall amounts of IL-2, but their growth was not completely inhibitable by antibodies to IL-2 or theIL-2 receptor. Another clone was derived, which produced no detectable IL-2, but grewindependently of exogenous IL-2. The growth of cells of this clone was completely inhibited byantibody, and death of the cells resulted. The experiments described here represent the firstdemonstration that antibody antagonists of the growth factor can induce the death of cells thatgrow by auto-stimulatory mechanisms. They support the hope that cytokine antagonists may finduse as therapeutic reagents in the treatment of auto-stimulatory neoplasms.The second part of the thesis presents a new technique in gene targetting, which has broadapplications in the study of gene function. Using a bacterial recombinase under the control of atissue-specific developmentally regulated promoter, transgenic animals were derived, in which atarget gene was deleted in thymocytes and their daughter cells, but no other tissues. Thistechnique circumvents the impediment embryonic lethality may present in some gene deletionexperiments, and allows questions of tissue-specifc function of genes in pathogenesis and normaldevelopment to be addressed.iiiTABLE OF CONTENTSAbstract^ iiTable of contents^ iiiList of figures ivList of abbreviations^ viiAcknowledgement ixForeword^ ixPart 1: Autostimulatory mechanisms in leukemogenesisIntroduction^ 1Results 17Discussion^ 68Bibliography 78Appendices^ 90Part 2: A new technique in gene deletionForeword^ 99Introduction 99Results^ 118Discussion 153Bibliography^ 158Appendices 165Materials and Methods^ 173ivLIST OF FIGURESFigure^ TITLE PAGE1.1. Factor responses of FD.C/2 cells maintained in IL-2 and IL-3 181.2. Factor responses of FD.C/2 cells maintained in IL-2 or IL-3 191.3. The proviral construct used to make I2-3.1 cells 221.4. FD.C/2T cells showed some response to IL-2 and IL-3 241.5. Response of FD.C/2T cells to IL-2 and IL-3 was more obviousat lower initial cell density 251.6. Comparison of responses of FD.C/01 1.1 cells with those of theparental FD.C/2 cells 261.7. Density-dependence curves for 2 FD.C/2T clones 281.8. Plating efficiency of FD.C/2 cells 291.9. IL-2 activity in the supernatants of FD.0/2'P cells 301.10. Response of FD.C/2 cells to FD.C/2T cell supernatants 321.11. Re-assay of FD.C/2T supernatants 321.12. IL-3 activity in the supernatants of FD.C/011 cells 331.13. IL-3-mediated response of FDC-P1 cells to FD.C/2W.1 supematant 341.14. GM-CSF-mediated response of FDC-P1 cells to FD.0/2'P.1 supernatant 351.15. Rabbit antibodies to IL-3 and GM-CSF partly abolish the effect ofFD.0/2'P.1 supernatant on FDC-P1 cells 371.16. IL-6 activity in the supernatant of FD.C/2 111.1 cells 381.17. Supematant of FD.C/2 cells contains no IL-2 detectable IL-2 activity 391.18. Specificity assay for the preparation of PC61 antibodies used in these studies 421.19. Specificity assay for the preparation of DMS antibodies used in these studies 421.20. PC61 antibody significantly and specifically inhibits the growth of FD.C/2T cells 441.21. DMS antibody significantly and specifically inhibits the growth of FD.C/2T cells 441.22. Growth-factor response of stained FD.C/2 cells is similar to that of unstained cells 461.23. Assessment of the ability of FD.C/2W.1 cells to support the growth of FD.C/2 cells 461.24. The vector used to confer hygromycin resistance upon FD.C/2Hyg cells 481.25. Growth-factor responsiveness of FD.C/2Hyg cells is similar to that of FD.C/2 cells 501.26. FD.C/2T cells support the survival of FD.C/2Hyg cells at highstimulator:responder ratios 521.27. FD.C/2T cells have a dramatic advantage over FD.C/2 cells in antibody resistance 541.28. FACS analyses of IL-2-receptor p55 on FD.C/2 and FD.C/2T cells 55V1.29. The integrating expression vector, pPO1 LhIL2, used to attempt to makeauto-stimulatory derivatives of FD.C/2 cells 581.30. Effect of zinc concentration on growth of FD.C/2Hyg cells 611.31. The episomally replicating expression vector, pBMGNeoLhIL2, used tomake auto-stimulatory derivatives of FD.C/2 cells 621.32. Zinc-dependence of FD.C/2pBLhIL2 cells 621.33. Density-dependence of FD.C/2pBLhIL2 cells 651.34. IL-2 activity is not detectable in the supematant of FD.C/2pBLhIL2 cells 661.35. IL-3 activity in the supernatant of FD.C/2pBLhIL2 cells 661.36. Antibody antagonists of IL-2 can completely inhibit the growth ofFD.C/2pBLhIL2 cells 671.37 Typical growth-factor responses of R6X cells 911.38. Typical growth-factor response of HT-2 cells 911.39. Typical growth-factor responses of FDC-P1 cells 921.40. Typical growth-factor response of 41 E5 921.41. Apparent decrease of bio-activity with duration of assay 931.42. ELISA monitoring of affinity-column purification of DMS antibodies 961.43. Specific inhibtory activity of the 11B11 antibody preparation used in these studies 961.44. Specific inhibtory activity of the 6B4 antibody preparation used in these studies 971.45. Specific inhibtory activity of the Rab7 antibody preparation used in these studies 971.46. Specific inhibtory activity of the Rab39 antibody preparation used in these studies 982.1. Topology of recombinations mediated by the conservative site-specificrecombinases 1042.2. Sequences of various Lox sites found in the bacteriophage P1 life cycle 1092.3. Transfection construct of Sauer and Henderson 1132.4. Potential stem-loop structure in the RNA resulting from transcription through aLoxP site 1132.5. p1017CRE 1192.6. Creation of pLOX2 1212.7. Determined sequence of a portion of the vector pLOX2 1222.8. pLOX2BGAL 1222.9. p1017L0X2BGAL 1232.10. Sub-cloned fragments used for probes 1252.11. Controls for dot-blots 1272.12. An alignment of the mouse and human growth hormone cDNA sequences 1282.13. Examples of dot-blots used to genotype transgenic progeny 130vi2.14. A portion of the genealogical tree of the mice used in these experiments^1302.15. Southern analysis of tail DNA from a CRE and a LOX-f3GAL-LOX transgenicmouse^ 1322.16. Illustration of head-to-tail integration arrays of the two transgene constructs^1332.17. Two of the Southern analyses used for approximate densitometric quantitationof transgene copy number^ 1342.18. Illustration showing expected band sizes from Southern analyses of transgeneconstructs and the expected recombination result^ 1352.19. Southern analysis of thymocyte DNA samples 1372.20. Southern analyses showing the results of recombination in doubly transgenicCRE/L0X-BGAL-LOX thymocyte DNA^ 1382.21. Southern analysis of tail and thymocyte DNAs from another doubly transgenicmouse^ 1392.22. Southern analysis of DNA showing the recombination result with the Stul/Mscldigest^ 1402.23. Southern analysis of PCR amplified fragments from CRE transgenic and doublytransgenic mice^ 1422.24. Restriction enzyme analysis of PCR-amplified fragment fromCRE/LOX-BGAL-LOX doubly transgenic mouse thymocyte DNA^ 1432.25. Tissue distribution of CRE expression^ 1442.26. Tissue distribution of recombination 1462.27. Southern analysis of DNA from spleen and spleen-derived T-cells of a doublytransgenic mouse^ 1472.28. Southern analysis of thymocyte DNAs showing recombination results in threemore doubly transgenic mice^ 1492.29. Southern analyses showing CRE-mediated recombination of target sequence inanother chromosomal context^ 1512.30. Southern analyses of the nature of the LOX-I3GAL-LOX integration in the 87 line^1522.31. PCR amplification of RNA from LOX-BGAL-LOX thymocytes^ 1662.32. Western analysis of thymocyte samples for detection of hGH 166VIILIST OF ABBREVIATIONSAb.^ antibodyAML acute myelogenous leukemiaby^ base pair(s)cDNA complementary DNACM^ conditioned mediumCML chronic myelogenous leukemiaDNA^ deoxyribonucleic acidEBV Epstein-Barr virusE. coli^ Escherichia coliES cell embryo stem cellG-CSF^ granulocyte colony-stimulating factorGM-CSF granulocyte/macrophage colony-stimulating factorhGH^ human growth hormonehIL-2 human IL-2IL^ interleukinkbp kilo base pairsLTR^ long terminal repeatM molarM.A.^ medium alone ( no growth factor added )mIL-2 murine IL-2ml^ millilitre(s)mM millimolarmRNA^ messenger RNAPCR polymerase chain reactionPDGF^ platelet-derived growth factorRNA ribonucleic acidVIII°GAL^ beta-galactosidase.1-19 microgram(s)1-1,1^microliter(s)°C degrees CelsiusAlthough the convention of naming genes in italics has been adhered to, in general, all transgenefragments have been named in upright capital letters in this work.ixACKNOWLEDGEMENTI wish to express my gratitude to my supervisor, Dr. John Schrader, whose enthusiasm andbreadth of knowledge were a source of inspiration, and without whose open-mindedness andgenerosity of spirit, project two would not have been undertaken. In addition, I am deeplyindebted to Dr. Jamey Marth of the Biomedical Research Centre, who supervised the workdescribed in the second part of the thesis, and to Daniel Chiu, who assisted with much of thework reported in the second part. I should also like to acknowledge Dr. Hermann Ziltener of theBiomedical Research Centre, who supplied several antibodies, and taught me some of the basictechniques of working with them, Dr. Ian Clark-Lewis, of the Biomedical Research Centre, whosupplied synthetic growth factors, and Dr. Allen Delaney of the Biomedical Research Centre whoprovided computer software and hardware support for this work.FOREWORDMaterials and methods aside, this thesis is in two parts, each dealing with a hypothesis and theexperiments carried out to test it. The first part of the thesis deals with a project proposed by mysupervisor. The second part deals with a project devised by myself. The link between the two isnot obvious from the purely scientific point of view, but the historical path from the first to thesecond was quite simple. The coincidence of three factors influenced me in conceiving thesecond project: 1) reading a review relating to the first project ( see introduction to Part 2 ), 2)general reading of scientific journals, and 3) the arrival of a new scientific technology at theBiomedical Research Centre ( where this work was carried out ). The materials and methodshave been presented as a single section, since many of the techniques there described wereused in both parts of the work.During the initial stages of work on project two, it became clear that there would be considerablecompetition from the group which donated the reagents essential to this project. The author'sefforts were therefore diverted towards completing project two as rapidly as possible.1AUTOSTIMULATORY MECHANISMS INLEUKEMOGENESISINTRODUCTIONThere is an abundant literature on the subject of autostimulatory tumours. Many relevant reviews havebeen published ( e.g. Sporn and Roberts, 1985; Lang and Burgess, 1990 ). Rather than attempting acomprehensive review of all the relevant publications, or simply providing an additional general review,the aim of this introduction is to make clear the reason for undertaking yet another study ofautostimulatory neoplasia - to identify the unanswered questions that prompted the present work.1) BACKGROUNDIn vitro, proliferation of mammalian cells depends on the presence of poly-peptide growth factors, andit is thought that the same growth factors regulate cell growth in vivo. More than 30 years ago, it wassuggested that production of a growth factor by a cell that would respond to it might be a componentof tumourigenesis ( Hsu, 1961 ). Secretion of growth factor that could then act on the secreting cellsreceptors was termed \"autocrine secretion\" by Sporn and Todaro, 1980. In this original sense, theterm \"autocrine\" strictly implied secretion of a factor which acted directly on the cell secreting it. Sincethe coining of this term, others have proposed alternative mechanisms of growth factor mediated self-support in tumours. In particular, the \"paracrine\" mechanism is an indirect mechanism involvingsecretion of a factor which induces surrounding cells to in turn secrete a factor that stimulates growthof the tumour cells ( e.g. Griffin et al., 1987; Leslie et al., 1991 ). The first suggestions regardingautostimulatory mechanisms implied action of growth factor on receptors displayed at the cell surface,however, many workers ( see below ) have more recently proposed that growth factors may functionfrom within the cell on receptors that are not exposed to the extra-cellular environment making specificantagonsim of the autostimulatory factor more difficult. Growth factor antagonists could clearly have arole in intervention in paracrine loops, however, the present work concerns the possibly more difficultproblem of direct mechanisms, in which factor and receptor are present in ( or on ) the same cell. Toallow for the possibility of intracellular function, but make clear the distinction from the paracrinemechanism of self-supporting growth factor loops, the term \"autostimulation\" is used here.Tumourigenesis has long been considered to be a multi-step process ( Foulds, 1958; Hunter, 1991 ),involving several genetic lesions in a cell that gives rise to a malignancy. Any given genetic lesion or2phenotypic trait of a tumour ( such as autostimulatory growth factor production ) may, however, beepiphenomenal with respect to the mechanism of tumourignesis. It is only by examining thefrequency of the occurence of the trait in tumours of a given type, or by observing the effect ontumourigenesis when the trait is reversed, that one could make claims about the involvement of thetrait in the tumourigenic process. The first part of this introduction will be concerned with the basis ofthe claim that autostimulatory mechansims are significant in tumourigenesis.Early suggestions as to the presence of autostimulatory factors in tumourigenic cells came fromTodaro and De Larco ( 1978 ). These workers found that mouse fibroblastoid cells transformed to thetumourigenic phenotype by various means, including transfer of DNA from human tumour tissue,released \"transforming growth factors\" into their conditioned medium. These factors would in turnstimulate growth of non-transformed cells. Since then, however, it has become apparent that theproduction of these two factors, termed TGF alpha and TGF beta, is not specific for tumour cells, andmany cell types in normal animals produce them ( Sporn and Roberts, 1985 ).One of the first direct pieces of evidence that production of an autostimulatory factor could have acausal role in oncogenesis involved platelet-derived growth factor ( PDGF ). As its name suggests, thisfactor is normally derived from platelets, which release it when activated. It is therefore present inserum used for much mammalian tissue culture, many cell types in culture requiring the presence ofthis factor for growth. The transforming protein of the simian sarcoma virus ( SSV ), encoded by theoncogene v-sis, is very closely related to a portion of one form of PDGF, and the presence of thisoncogene is essential to the tumourigenic activity of the virus ( Doolittle et al., 1983; Waterfield et al.,1983 ). Production of PDGF-like molecules occurs in several human tumours, e.g. osteosarcomas( Graves et al., 1983; Betsholtz et al., 1984) glioma ( Nister et al., 1984 ), and the bladder carcinomaline T24 ( Bowen-Pope et al., 1984 ), many of which also have functional PDGF receptors. In additionsome simian virus 40 ( SV40 ) -transformed cells as well as mouse fibroblasts transformed by murinesarcoma viruses produce PDGF-like molecules, and in some of these latter cells, neutralising antibodyto PDGF reduces the growth of the lines in vitro ( Bowen-Pope et al., 1984 ). Additionally, acorrelation was observed between the amount of PDGF produced by simian sarcoma virus ( SSV )-transformed cells and their growth rate in nude mice ( Huang et al., 1984 ). Johnsson et al. ( 1985 )were able to demonstrate some reduction in the ability of SSV to transform fibroblasts in the presenceof high concentrations of neutralising antibody to PDGF. It has since been shown that arterial smoothmuscle cells isolated from normal young rats also release biologically significant amounts of thesefactors into their culture medium ( reviewed in Sporn and Roberts, 1985 ). This puts into question thesuggestion that the production of PDGF-like substances is always aberrant in a tumour cell, rather thana simple part of the physiology of the corresponding normal cell type.3Autocrine loops in other tumours and the tumour models discussed below may indeed be reflectionsof physiological events. Nevertheless, one may still propose the autocrine loop as a mechanism in thepathogenesis and maintenance of these tumours. Cells which would otherwise pass through a stageof differentiation in which autostimulation was physiological might, in the process of mutation resultingin the genesis of the tumourigenic clone, be \"frozen\" at this stage of differentiation. Alternatively,mature cells might lose their normal ability to down-regulate a physiological autostimulaton. T-cells, forexample, when stimulated by antigen in the presence of appropriate cooperative signals, will up-regulate interleukin-2 ( IL-2) receptor levels, and secrete IL-2 ( reviewed in Dinarello, 1991 ). Variousmechanisms subsequently operate to down-regulate both the IL-2 production and the level ofreceptor, and the loss of such mechanisms would lead to pathological autostimulation. Of course,tumour cells may well acquire alternative mechanisms of autonomous growth, irrespective of whetherthey have passed through an autostimulatory stage during differentiation, or lost their ability to down-regulate an autostimulatory loop. Such cells may be producing a potentially autostimulatory growthfactor as an epiphenomenon.This consideration stresses the importance of showing more than thesimple presence of growth factor production and of the corresponding receptor on the cells of atumour, in order to establish the autocrine loop as essential to the tumour's viability. Studies aimed atinhibiting growth with growth factor antagonists therefore acquire greater importance in suggestingthe involvement of autostimulatory mechanisms in tumourigenesis.An example of spontaneous tumours that displayed autostimulation and susceptibility to inhibitionwas that of the human small-cell carcinomas of the lung that produce gastrin-releasing-peptide, themammalian equivalent of bombesin ( reported by Cuttitta et al.,1985, and others ). Gastrin-releasing-peptide ( GRP) is a poly-peptide hormonal factor usually produced by foregut-derived tissue, butaberrant production of hormones is a quite common feature of small-cell lung cancers, and manyinstances involve the production of GRP. Neutralising monoclonal antibodies to bombesin signifcantlyinhibited growth of the tumours in vitro. When transplanted into nude ( geneticallyimmunocompromised ) mice, cells of this tumour are able to proliferate and metastasise. Treatment ofsuch mice with neutralising antibodies to GRP results in a significant, although temporary inhibition ofthe proliferation of these cells. The tumours eventually adapt to the presence of antibody, possibly byoutgrowth of some of the cells with additional or pre-existing mutations which allow growthindependently of GRP.Several examples of the role of insulin-like growth factor-1 ( IGF-I ), in human tumours have beenreported, in which antibodies to this factor have slowed growth of the tumour cells in vitro. Theseinvolved neoplasms of the lung ( Minuto et al., 1988 ) and breast ( Huff et al., 1986 ), andosteosarcoma ( Blatt et al., 1984 ). IGF-I was also more recently the subject of another experimental4strategy to interfere with an autostimulatory loop in a spontaneously arising tumour ( Trojan et al,1992 ). Cells of the transplantable rat glioma C6 were found to produce IGF-I, and the receptors for thisfactor. Cells of this glioma were transfected with a plasmid that produced an anti-sense IGF-I transcriptand thereby caused a significant reduction of the amount of IGF-1 produced in the cells. C6 cellstransfected with the anti-sense plasmid were unable to produce tumours in syngeneic rats, whereascells transfected with empty vector were tumourigenic.Trojan et al. ( 1993) subsequently observedthat injection of anti-sense transfected C6 cells could cause regression of existing control C6 tumoursand found lymphocytic infiltration at the sites of tumour regression. This suggested that an immuneresponse was responsible for the death of tumour cells, and that the production of IGF-I may haveacted to overcome such an immune response. Since these authors make no claim as to the effect ofthe anti-sense plasmids on in vitro growth, it is at least conceivable that the IGF-I produced by thesecells was acting not as a growth factor, but simply as a mechanism of avoiding the immune response. Amore likely explanation of these findings might be that the growth-promoting activity of IGF-I wasallowing the tumour cells to overwhelm the response. The autocrine action of IGF-1 in these tumourcells may also be an indirect reflection of physiological events - there is some evidence to suggest thatcentral nevous system precursors of glial cells may produce IGF-1 in their normal proliferation anddifferentiation during fetal ontogeny ( Drago et al., 1991 ).Human melanomas liberate a factor, termed MGSA ( for \"melanoma growth stimulating activty\" ) whichstimulates the growth of normal melanocytes and melanoma cells. Indeed this factor was purified andcharacterised from the supernatant of a melanoma line. ( Richmond et al., 1988; Richmond andThomas, 1988 ). In the case of one such tumour, changes in growth rate of the cells could becoorrelated with changes in the level of messenger RNA for the factor present. ( Bordoni et al., 1989 ).In addition to the data reviewed above, there is a further body of evidence from which one could inferthat autostimulatory mechanisms might be involved in the development of malignant neoplasms.Several oncogenes have been described that act as constitutively active receptors for growth factors.The v-erb-B gene, for instance, encodes a truncated and activated epidermal growth factor receptor( Downward et al.,1984 ), and the v-fms gene is also constitutively activated by virtue of a mutation thatdistinguishes it from its cellular counterpart c-fms ( Sher et al., 1985 ), the receptor for macrophagecolony-stimulating factor ( CSF-1). A constitutively activating mutation also makes oncogenic anothergrowth factor receptor, the c-kit gene product ( Geissler et al., 1988 ). Constitutively activated receptorsignal transducers, such as mutated ras gene products, are also found widely distributed amongstspontaneous tumours ( reviewed in Barbacid, 1987, and Hunter, 1991 ). These examples support thenotion that the constitutive activity of a growth factor response pathway is a common mechanism intumourigenesis.5With the exception of the in vivo work of Trojan et al., no instance of autostimulation in a solid tumourhas been reported in which tumour cells have died as the result of growth factor antagonism.Eradication of malignancies in man generally requires destroying tumour cells - most tumours do notdisplay any evidence of being immunogenic, and the occurence of spontaneous regression oftumours is so rare as to merit report in a journal in every instance where documentation is sufficient.Normal cells of the hemopoietic system, however, display the unique property of being absolutelydependent on growth factors for not only proliferation, but also survival. This is true of cells in vivo aswell as in vitro ( Crapper et al., 1984; Savill et al., 1989; Koury and Bondurant, 1990 ). When thegrowth factor is completely removed from culture, or not available to the cells in vivo, such cells die.Moreover, in contrast to cells of solid tumours, malignant cells of hemopoietic origin are commonlyexposed to the circulation directly, rather than occuring in masses with some cells only poorlyaccessible from the circulation. For these reasons, hemopoietic tumours are more likely than solidtumours to be amenable to therapy with growth factor antagonists.2) AUTOSTIMULATION IN HEMOPOIETIC NEOPLASIAHemopoietic precursors as well as mature cells are relatively easy to obtain from patients andexperimental animals. In the right conditions, precursors can be made to differentiate in vitro. As aresult, much is known about the factors that support the growth, differentiation and function ofhemopoietic cells. Amongst the known factors are the granulocyte-macrophage, granulocyte, andmacrophage colony stimulating factors ( GM-CSF, G-CSF, and M-CSF or CSF-1 ), erythropoietin,interleukins ( ILs ) 1 to 13, interferons, and steel-factor ( for a recent overview, see Schrader, 1992 ).There is also considerable evidence for the role of autostimulatory loops both in malignancies and thenormal physiology of the hemopoietic system.The first evidence of an autostimulatory loop in the pathogenesis of a leukemia was the discovery of aspontaneously factor-independent derivative of an interleukin-3 ( IL-3) -dependent mouse cell line. Incontrast to cells of human or even other rodent origins, mouse cells have a marked tendency to giverise to immortal clones of cells in culture. Such cell lines, if of hemopoietic origin, generally remaindependent on hemopoietic growth factors for survival and proliferation. In this instance, culture of alarge number of cells in the absence of factor permitted the emergence of a single clone which grewautonomously. Whereas the parental IL-3-dependent cells died on injection into syngeneic mice inthe absence of an artificial source of IL-3, the autonomous cells gave rise to leukemia in mice withoutan IL-3 source. In vitro, the autonomous cells grew more rapidly at higher cell density, suggesting the6presence of an autostimulatory mechanism. IL-3 was detectable in the supernatant of the cells inculture, and in the serum of the leukemic mice ( Schrader and Crapper, 1983 ). Since theseobservations were made, numerous examples of autocrine loops in human tumours have beenreported.Interleukin-1 ( IL-1 ), a cytokine present in many mammalian cells appears to be responsible forautostimulation of some freshly explanted adult T-cell leukemic cells ( Shirakawa et al., 1989 ). Whencultured at sufficiently low density for the effect to be observed, such cells proliferate in response to afactor present in their own conditioned medium. ( At higher densities no effect was observedsuggesting that the concentration of autostimulatory factor was then saturating with respect to theproliferative capacity of the cells.) A rabbit neutralising anti-serum to IL-1 specifically inhibited thisresponse, although it did not completely inhibit proliferation of the cells. This autostimulatory loop mayreflect a physiological phenomenon, since activated T-cells produce IL-1, and respond to IL-1 inconjunction with antigenic stimulation, by upregulating IL-2 and IL-2 receptor production, in thecourse of an immune response.Several laboratories have reported that IL-1 is involved in the autostimulatory growth of Epstein-BarrVirus ( EBV ) -transformed cells. Though not tumourigenic in healthy hosts, EBV-infected cells,present in the circulation of more than 90% of the general population over 30, may give rise to B-cellleukemias in immunocompromised patients. When removed from the circulation and placed intoculture, some EBV infected cells give rise to immortal \"B-lymphoblastoid\" cell lines, which continue togrow apparently independently of hemopoietic growth factors. Scala et al. ( 1987 ), Wakasugi et al.( 1987 ), and Vandenabeele et al. ( 1988) have reported observations on three B-lymphoblastoidlines which produced a factor that promoted their own growth ( again, at lower densities ). Thesefactors had physical and biological characteristics of IL-1, and in two of the three instances,neutralising antibody to IL-1 was able to inhibit, to some extent, the growth of the cells. Bertoglio et al.( 1989 ) studied a further 12 B-lymphoblastoid lines, and found a strict correlation between thepresence of IL-1 mRNA and IL-1 receptor, consistent with an autostimulatory role of the IL-1.Many groups have attempted to account for the apparent factor-independence of EBV-tansformedcells. Interleukin 5, ( IL-5 ), originally described as a growth and differentiation factor produced by T-cells and acting on eosinophils and B-cells ( Takatsu et al., 1988 ; Sanderson et al., 1988 ), has beenshown to be involved in an autostimulatory loop in some lymphoblastoid lines ( Paul et al., 1990;Baumann and Paul, 1992 ). These workers demonstrated production and high-affinity binding of IL-5by the EBV-transformed cells, and showed that neutralising antibodies to IL -5 partially inhibitedgrowth of these cells. Interleukin 6 ( IL-6 ), known for its B-cell growth promting activities ( reviewed in7Kishimoto et al., 1992 ), was used by another group in transfection of EBV-transformed lines ( Scala etal., 1990 ). Expression of IL-6 in these cells caused a striking increase in growth rate at very low celldensities, secretion of detectable IL-6 into the culture medium, and the development of tumours inimmunocompromised mice from these cells, which were not tumourigenic in the absence of IL-6expression.IL-6 has come to prominence within the field of autostimulatory loops, as a growth factor involved inthe pathogenesis of multiple myeloma, a tumour of plasma cell origin. Kawano et al ( 1988) reportedan IL-6-dependent autostimulatory loop in cells freshly isolated from patients. They observedproduction of and response to IL-6, as well as some inhibition of in vitro growth with neutralisingantibody to IL-6. However, the presence of contaminating marrow cells in these myeloma samplescould not be excluded, thus admitting of the possibilty that the IL-6 detected in these samples wasnot produced by the myeloma cells themselves, but by such contaminating cells. Another group,( Klein et al., 1989 ) found that neither of two established human myeloma lines expressing receptorsfor IL-6 produced biologically detectable IL-6 or IL-6 mRNA. Moreover, anti-IL-6 antibody did not effectthe growth of these cells. When these workers compared bone marrow samples of patients withfulminant multiple myeloma and control marrow samples, they found that stromal cells in myelomapatients, rather than the myeloma cells themselves were secreting IL-6, suggesting a possibleparacrine growth loop. These workers also showed a correlation between in vitro responsivenes to IL-6, and in vivo proliferative status in 13 patients studied ( Zhang et al., 1989 ). However, Schwab et al,using one of the same lines ( nominally ) used by Klein et al., were able to show secretion andresponse to IL-6, as well as the presence of IL-6 message. They also showed a 70% inhibition of thegrowth of the cells in the presence of a neutralising antibody to IL-6. The findings of severallaboratories in relation to IL-6 and multiple myeloma have led to clinical trials ( e.g. Klein et al., 1991) ofneutralising monoclonal antibodies in myeloma patients, with some success in inhibition of myelomagrowth in vivo.Although no reports have described an activating mutation in the IL-6 gene in human multiplemyeloma cells, Blankenstein et al. ( 1990) have described the insertion of an intracisternal A particle( IAP ) in the IL-6 gene of a mouse myeloma-like tumour, the plasmacytoma MPC11. An IAP is atransposable DNA sequence with some features of the retroviral long terminal repeats ( LTRs ), and iscapable of activating genes by insertion 5' of the coding sequence. The presence of thisrearrangement supported the hypothesis that activation of IL-6 production was involved in thepathogenesis of the plasmacytoma. Additionally, plasmacytosis occurs in transgenic mice expressingIL-6 specifically in B-cells, and in the presence of an immortalising mutation ( such as an activation ofthe c-myc gene ), these mice develop plasmacytomas ( Suematsu et al., 1989; and 1992 ).8The human immunodeficiency virus ( HIV) has also provided an example of autostimulatory growth intumour development. The HIV tat gene product has a growth-promoting role in Kaposi's sarcomacells. These tumour cells produce IL-6 and the growth-promoting activity of the tat product can bespecifically reversed, at least in part, by anti-sense oligonucleotides to IL-6 ( Miles et al., 1990 ).In view of the physiological autostimulatory loop mediated by IL-2 in T-cells ( mentioned above ), it isnot surprising that such a mechanism appears to operate in some T-cell leukemias. Indeed one of theearliest reports suggesting an autostimulatory loop in hemopoietic neoplasms ( Gootenberg et al.,1981 ), described cells of several human cutaneous T-cell lymphoma lines producing and respondingto what was then known as \"T-cell growth factor\". Subsequently, Duprez et al. ( 1985) reported theestablishment of a line from a T-cell lymphoma, which secreted and responded to human IL-2. Thegrowth of this line was strictly density-dependent, and the cells prodcued very little IL-2 ( 0.6 units perml in a 25 x concentrated cell-culture supernatant ). Neutralising antibody to IL-2 was able to inhibit thegrowth of these cells by 90%. The same workers later showed that the autostimulatory loop in thesecells could be blocked by Cyclosporin A, which blocked IL-2 transcription ( Dautry-Varsat et al., 1988 ).Physiological production of myeloid growth factors has been demonstrated for several myeloid celltypes. Macrophage/monocyte lineage cells secrete GM-CSF, CSF-1 and G-CSF in response toinflammatory stimuli, and can respond to GM-CSF and CSF-1 ( reviewed in Moore, 1991 ). Eosinophilsand neutrophils can secrete GM-CSF and IL-3 when stimulated with other cytokines ( Kita et al., 1991;Moqbel et al., 1991 ). Mast cells can be stimulated, by cross-linking of cell-surface antibodies withantigen, to produce IL-3, GM-CSF, IL-4, and IL-5 ( Burd et al, 1989; Plaut et al, 1989; Wodnar-Fillipowicz et al., 1989; Razin et al., 1991 ), and can respond to IL-3 and IL-4. In all these situations,some form of ligand-mediated physiological stimulation of cell-surface receptors is required to inducesecretion of growth factor. Such stimuli function through intracellular pathways that share commoncomponents and some components of these pathways are activated by oncogenic mutations.Activating mutations of the N-ms gene, for example, are found in many acute myeloid leukemias ( Boset al., 1987 ). Therefore, it is conceivable that oncogenic mutations may secondarily activate theproduction of potentially autostimulatory growth factors, which do not participate in the survival orproliferation of the tumour cells. This highlights the importance of studies using growth factorinhibition before drawing conclusions as to the role of autostimulatory factors in the genesis oftumours in cell types that display a variety of physiological autostimulatory mechanisms.For many years, it was thought that culture of human acute and chronic myeloid leukemic ( AML andCML ) cells required the addition of exogenous growth factors ( reviewed in Metcalf, 1989 ). Culture ofAML blasts at high density, however, ( e.g.over 106 per ml ) permits growth of purified leukemic blast9populations in the absence of exogenous factor. Such density-dependence is consistent with theactivity of an autostimulatory mechanism in these cells. Many workers have described the production,in blast-cell enriched populations of myeloid leukemias, of factors known to stimulate the growth of thecorresponding normal cells ( e.g.Young and Griffin, 1986; Young et al., 1987; Young et al, 1988;Sakai et al., 1987; Oster et al, 1989; Reilly et al., 1989; Freedman et al., 1992 ). Several of thesereports simply described this finding of growth factor production, or of growth factor message, infreshly isolated cells, without attempting to establish the requirement for autostimulation inmaintenance of growth and many reports were subject to the criticism that it was possible that thegrowth factor detected was present not in the leukemic cells themselves, but in contaminating normalcells present in the isolates. Nevertheless, it is clear that some populations enriched for leukemic cellsproduce GM-CSF in vitro, and that neutralising antibody to GM-CSF reduces the growth rate of suchcells ( Young and Griffin, 1986 ). Similarly, there is good evidence for the presence of IL-1 in suchpopulations, and again, in vitro inhibition of IL-1 activity with antibodies, or more recently, with thenaturally occurring IL-1 antagonist IL-1ra has slowed growth in cultured cells ( Cozzolino et al., 1989;Rodriguez-Cimadevilla et al., 1990, Bradbury et al., 1990; Rambaldi et al., 1991; Estrov et al, 1992 ).With respect to IL-1, at least, there is evidence from in situ hybridisation studies showing that mRNAfor IL-1 is present in the leukemic blasts themselves ( Nakamura et al., 1990 ). In the case of GM-CSF,however, there is some evidence that the growth factor is produced by non-proliferating progeny ofthe blasts present in leukemic isolates ( Murohashi et al., 1989 ). This would be consistent with aparacrine rather than an autocrine mechanism, in which IL-1 stimulates production of GM-CSF in suchpopulations, and evidence for this has been reported in human AML ( Delwel et al., 1989; Estrov etal., 1992 ), and in a spontaneous mouse myelo-monocytic leukemia ( Leslie et al., 1991 ).The techniques used for isolating leukemic blasts themselves may induce the production of growthfactors ( Kaufman et al., 1988 ). These blasts represent cells at various stages of differentiation of themyeloid pathway, and it has not yet been determined what features of the apparent autostimulatoryand paracrine loops are physiological in cells at these various differentiation stages. Although generearrangements, particularly translocations, are common in leukemic cells, there are no data tosuggest that any such rearrangement has directly activated a growth factor gene. The inability todetect autostimulatory factors need not necessarily imply that they are functionally absent - very smallamounts may suffice to induce proliferation ( reviewed in Lang and Burgess, 1990, and Thomson,1991; and the data presented here ). For these reasons, it will be important in subsequent studies ofleukemic ( and normal ) hemopoiesis to attempt, perhaps by sensitive in situ polymerase chainreaction ( PCR ) techniques to demonstrate the presence of cytokine production in cells that are well-characterised with respect to differentation and function.10With respect to CML, there is no good evidence to suggest the involvement of autostimulatorymechanisms in the pathogenesis of the disease. The hallmark of this disease is the presence of thePhiladelphia chromosome ( an abnormal chromosome 22 resulting from a balanced translocationbetween chrmosomes 14 and 22 ), This results in the formation of a protein-encoding gene fusionbetween the bcr (for breakpoint cluster region ) and abl genes. This latter is the cellular equivalent ofthe oncogene found in the Abelson murine leukemia virus. Although its precise function is unclear,the c-abl product is a tyrosine kinase thought to be involved in pathways of signal transduction fromcell surface receptors. The fusion product function in a constitutively activated manner with respect toits tyrosine kinase activity. Thus, if a growth factor pathway were involved in leukemogenesis resultingfrom expression of a constitutively activated c-abl product, the mechanism would be more likely toinvolve short-circuiting the growth factor response, bypassing the need for ligand to bind to receptor,rather than inducing a truly autostimulatory loop. When the bcr/abl product was expressed in immortalmurine factor-dependent myeloid cells, these cells acquired the ability to produce IL-3, and becameindependent of exogenous factor ( Hariharan et al., 1988 ). This may simply reflect the action of theactivated kinase in mimicking a physiological stimulus, which would normally induce IL-3 production.Moreover, others have reported similar experiments in which no autostimulatory factor production wasdetected ( Laneuville et al., 1991 ). The relevance of hemopoietic growth factor production to thepathogenesis of CML has not been established, indeed one group has reported negative evidencewith respect to autostimulatory factors in CML cells ( Otsuka et al., 1991 ).3) ANIMAL MODELS OF AUTOSTIMULATORY HEMOPOIETIC NEOPLASIAMuch work in autostimulatory neoplasia has been carried out in animal model systems. These have thegreat advantage that autostimulatory loops can be constructed so that the existence and requirementof autostimulation for proliferation of cells can be ascertained. These autostimulatory loops can bemanipulated in vitro and the effects of such manipulations can be studied in cells returned to animals.Amongst the earliest work with animal models of autostimulatory leukemia is that of Adkins et al.( 1984 ). These workers were investigating transformation of normal chicken myeloid cells byoncogenic retroviruses. A virus carrying a single oncogene ( either v-myb or v-myc ) was able toimmortalise primary cells, which still required a growth factor ( chicken myelomonocytic growth factor,cMGF, a distant relative of IL-6 and G-CSF ) for continued proliferation in vitro. When super-infectedwith a second virus carrying a src-family oncogene ( a constitutively activated form of a signal-transducing tyrosine kinase ), the cells started to produce cMGF, and became independent ofexogenous factor. The subsequent growth of such cells could be inhibited, to some extent, by11neutralising anti-serum to cMGF. Graf et al. ( 1986 ) extended these results by showing that the v-miloncogene product ( also involved in signal-transduction pathways ) would similarly induce v-myc-transformed cells to produce cMGF, and would cooperate with v-myc in forming leukemias in vivo.Somewhat related are the observations of Baumbach et al. ( 1987 ) on tumours arising after infectionof mice with a retrovirus containing the c-myc gene. All the tumours obtained were of the monocyte-macrophage lineage, and they showed various phenotypes with respect to production of GM-CSFand CSF-1. The growth of one of these tumours was partially inhibited by neutralising antibody toCSF-1. This tumour was shown to have a CSF-1 gene rearrangement, suggesting that this mutation,resulting in the aberrant expression of CSF-1, cooperated with the myc gene in the genesis of thetumour. Another interesting set of experiments relating to the idea of cooperative events intumourigenesis were those of Andrejauskas and Moroni ( 1989 ), who transfected a mouse immortalIL-3-dependent line with a vector containing an activated v-H-ras gene under the control of aninducible promoter. In transfected cells, induction of expression of the activated ras gene correlatedwith loss of dependence on exogenous IL-3 and with production of IL-3 by the cells, a provocativeobservation in the light of the role of the ras gene products in many signal-transduction pathways thatmight be involved in the physiological induction of IL-3 production in reponse to appropriate stimuli.Strong support was given to the notion that acquisition of an autostimulatory mechanism might be afrequent cooperative factor in the formation of tumours from otherwise immortal cells by theobservations of Stocking et al. ( 1988 ). These workers produced 11 autonomous variants of a factor-dependent mouse cell line by random insertional mutagenesis with a retrovirus. Ten of the 11 clonesproduced one of the two growth factors to which the parental line responded ( GM-CSF or IL-3 ), andin most cases these workers were able to demonstrate the insertion of the retrovirus in proximity to theGM-CSF or IL-3 genes, which induced production of the factors. Similarly, Duhrsen et al. ( 1990 )observed that another mouse factor-dependent myeloid line, also responsive to both IL-3 and GM-CSF, spontaneously gave rise to tumours when injected into irradiated syngeneic mice. The cells ofabout one third of the tumours they isolated displayed production of either GM-CSF or IL-3, and in themajority of these cases it could be shown that this growth factor production was due to the insertion ofan IAP in the proximity of the relevant gene.When transfected with vectors that determine expression of the product of cDNAs of the relevantgrowth factors, mouse immortal factor-dependent cell lines are reproducibly converted toindependence of exogenous growth factor. Such experiments have been performed by manyworkers, and in some cases, density-dependence and the ability of neutralising antibody to growthfactor or receptor to inhibit growth has been examined.12Lang et al. ( 1985 ) conferred the ability to express GM-CSF on cells of a GM-CSF-dependent immortalline, and showed that the resultant cells were autonomous and tumourigenic. Although the amountof GM-CSF in the supernatant of these cells cultured at low density was too small to be able to supportthe growth of untransfected factor-dependent cells, the autostimulatory cells showed no density-dependence and were not inhibited by antibody to GM-CSF. These latter observations led theauthors to suggest the possibility that the engagement of ligand and receptor took place intracellularlyin these autostimulatory cells. ( This suggestion and other evidence relating to it are dealt with in somedetail in the discussion section of this portion of the thesis.) Laker et al. ( 1987 ), however, usingsimilar expression vectors and the same cell-line, observed that GM-CSF-producing autostimulatorycells were indeed density-dependent, at least for some period following transfection. Neutralisinganti-serum to GM-CSF significantly inhibited the growth of these cells, although cell death was notdocumented. These workers made the additional observation that such clones underwent a transitionto a state in which their growth was no longer density-dependent, nor inhibitable by antibody. Thistransition occured more rapidly in clones initially secreting higher amounts of GM-CSF, although noexplanation for this was apparent. Clearly, although an autostimulatory mechanism was involved in thegeneration of these cells, in the post-transition state they could not be shown to be autostimulatory,despite continued production of GM-CSF. These observations highlight the possibility of suchprogressions in the pathogenesis of spontaneous leukemias.At least two groups have made IL-2 dependent models of autostimulatory growth. Taniguchi et al.( 1987 ) using a retroviral vector and a human IL-2 cDNA, obtained clones that produced IL-2, from anIL-2-dependent cytotoxic T-cell line. These cells grew autonomously, but appeared no longer torespond to exogenous IL-2 ( except for some inhibition of growth at higher concentrations of IL-2 ).These workers were able to achieve some inhibition of growth of such cells with a neutralisingantibody to the IL-2 receptor. Cells of these clones were able to induce lymphomas in syngeneicmice. Karusayama et al., using a different vector system, and another IL-2 dependent line, obtainedIL-2 producing clones that, while capable of growth without exogenous IL-2, were still responsive toexogenous IL-2. Moreover, these clones proliferated in a density-dependent manner, and theirgrowth could be significantly, but not completely, inhibited by neutralising antibodies to IL-2 or to theIL-2 receptor. It is worth noting that this group used the mouse anti-human monoclonal antibody DMS-1 to apparently inhibit the action of mouse IL-2. The DMS-1 monoclonal antibody preparationgenerated by the present author (see results section ) did not appear to be able to neutralise mouseIL-2 ( data not shown ). These authors also observed that all of the autostimulatory clones they hadgenerated ,except the clone producing the least IL-2, were capable of inducing tumours in mice, andthat the latency of tumour induction was inversely correlated with the level of IL-2 production by theclones in vitro.13Several groups have generated autostimulatory cells in mouse IL-3-dependent cell lines, ( Hapel et al,1986; Jirik et al., 1987; Wong et al., 1987 ). In these cases, autostimulatory clones were tumourigenic,but density-dependence and the effects of neutralising antibodies to IL-3 were not documented.Dunbar et al, ( 1989) performed an elegant set of experiments in an IL-3 dependent autostimulatorymodel, which are discussed later. Cells of an IL-5-dependent line spontaneously acquired the abilityto produce IL-5 during culture, as a result of the activating insertion of an IAP 5' of the IL-5 gene.Proliferation of these autostimulatory cells was partially inhibited by a neutralising monoclonal antibodyto IL-5 ( Tohyama et al., 1990 ).Some salient observations resulting from this considerable body of investigation in the field ofautostimulatory hemopoietic tumourigenesis are as follows:1. The weight of the evidence suggests that autostimulatory mechanisms are probably involved in thepathogenesis of acute myeloid leukemia, and may be involved in the development of otherneoplasms.2. In instances where attempts have been made to inhibit the growth of autostimulatory cells withantagonists of the growth factor or its receptor, investigators have either failed to see any inhibition, orobtained incomplete inhibition. In no instance has growth factor antagonism been reported to result inthe death of the autostimulatory cells.These observations have prompted the present investigation into the inhibition of autostimulatorygrowth with growth factor antagonists. The answer to the question of whether blocking the action ofgrowth factor can lead to the death of the cells of an autostimulatory tumour has clear implications forthe feasibility of developing growth factor antagonists as therapeutic agents in some forms ofmalignant neoplasia.4) EXPERIMENTAL STRATEGYIn its most general form, the question that this project was designed to address can be stated simplyas: can specific antagonism of the relevant growth factor result in the death of growth factor-dependent autostimulatory cells? Underlying this question is the hope that a positive answer mightencourage the development of substances that could find clinical application in inhibiting the growth14of human tumours that have an autostimulatory component, with minimal non-specific toxicity. Inchoosing an experimental model, the following criteria were considered:1) The model should be as near to clinical relevance as practicable. The cells should be ofmammalian origin, preferably, and the autostimulatory cells should produce malignant tumoursin animals.2) The cells should for practical reasons be \"immortal\" in vitro, allowing repeated studies ofcells of essentially homogeneous phenotype.3) The growth factor requirements of the cells used should be well understood ( in so far as itis possible to say this of any mammalian cells ), and the required growth factor(s) should not bephysiologically present in the circulation in concentrations that would support growth of thesecells.4) In order to facilitate the interpretation of results, the nature of the cells should be such thatinterference with the factor-receptor interaction to the extent of factor starvation would lead tocell death, rather than merely growth arrest, i.e. the cells should be factor-dependent ratherthan merely factor-responsive.5) The proposed autostimulatory growth factor cDNA should be available, so that it would bepossible, given an appropriate expression vector, to introduce this cDNA into the chosencells and thereby render them autostimulatory. This would offer the advantage that one mightbe able to regulate the level of production of growth factor by using an inducible promoter,permitting one to address the question of whether varying the amount of factor producedwould vary the resistance of the autostimulatory loop to antagonists.6) The proposed autostimulatory growth factor should be one to which neutralisingantagonists that function by interfering with the interaction between the factor and its cell-surface receptor would be readily available.Of all mammalian cell types that have been studied and manipulated in vitro, those best conforming tocriteria 1-4 are those of the hemopoietic system. Since almost all reports of putative pathologicalautostimulatory loops in hemopoietic neoplasias have concerned myeloid cells, the criterion of clinicalrelevance determined the choice of a myeloid cell line. Given the availability of cell lines, and therelative ease of manipulation of the animal in the laboratory, cells of murine origin were deemed to be15most suitable for these studies. At the time of initiating these studies, the only known specifichemopoietic growth factor antagonists were antibodies. In view of the size of these molecules, relativeto most pharmacological therapeutic reagents, hemopoietic cells provided an ideal model, since theyare, in general, readily accessible to substances present in the circulation.The following subsidiary criteria were established to determine the system finally selected for study:a) The cells should be responsive to more than one growth factor so that optimal controls fornon-specific toxicity of antibody preparations could be performed. Thus, an antibodypreparation which inhibited proliferation of the cells in the presence of growth factor A, couldbe shown to have no effect on the same cells in the presence of growth factor B atconcentrations at which its inhibitory effect was specific.b) The growth factor requirement of the cells should be such that one of the factors to which itwould be responsive could be anatagonised by antibodies of a species in which one couldproduce tumours using the auto-stimulatory cells. In the assessment of the inhibition oftumour growth in vivo, one might otherwise expect significant interference with the activity ofinjected antibodies due to immune response to xenogeneic protein.c) It would be desirable to select a system in which inhibitory antibodies to the receptor, as wellas to the growth factor were available, as this would allow a second line of attack shouldantagonism by anti-factor antibody alone prove insufficient to prevent survival ofautostimulatory cells.A system was designed conforming to all these criteria, in which cells of the imortal cell line FD.C/2were made autostimulatory using a human IL-2 cDNA. Le Gros et al. ( 1985 ) derived the IL-2responsive clone, FD.C/2, from an IL-3-dependent murine myeloid cell line ( FDC-P2, Dexter et al.,1980 ). The cells of this line can be grown in either IL-2 or IL-3 and human IL-2 is an effective ligand forthe mouse IL-2 receptor ( see, e.g., Taniguchi et al, 1986, and references therein ). Hybridomasproducing murine neutralising monoclonal antibodies to human IL-2 ( DMS-1 and DMS-2, Smith et al.,1983 ), and rat neutralising monoclonal antibodies to the mouse IL-2 receptor ( PC61, Lowenthal etal., 1985) were readily available. A human IL-2 ( hIL-2) cDNA had been cloned and expressed, using aretroviral vector, in an immortal mouse IL-2-dependent T-cell line, rendering infected cellsautostimulatory and tumourigenic ( Yamada et al., 1987 ). Neither IL-2 nor IL-3 are present in themouse circulation at levels sufficient to support the proliferation of physiologically responsive cellsother than those in the immediate micro-vicinity of the source of these factors, except in conditions of16severe immunological stress such as hyperimmunisation or graft-versus host disease ( Kuziel andGreene, 1991, Schrader, 1991 ). Thus, autostimulatory derivatives of FD.C/2 cells would most likelygrow as tumours by an autostimulatory mechanism in mice of the strain of origin ( DBA/2 ). Althoughthis system would involve stimulation of myeloid cells with IL-2, which is traditionally thought of as alymphoid-cell stimulatory factor ( see e.g. Smith, 1988, Kuziel and Greene, 1991 ), severalpublications have reported data on the presence of IL-2 receptor molecules on myeloid cell lines ( LeGros et al., 1985, Koyasu et al., 1986, Le Gros et al., 1987) on monocytes ( Hermann et al., 1985,Armitage et al., 1986, Rambaldi et al, 1987, Hotter et al., 1987 ), and on the blasts of acute myeloidleukaemias ( Hermann et al., 1985, Armitage et al., 1986, Yamamoto et al., 1986) by the time ofcommencement of this project. These data suggested at least the possibility that IL-2 might serve asan autostimulatory growth factor in some stage of the pathogenesis of myeloid leukaemias, increasingthe clinical relevance of the proposed model system.17RESULTS1. CREATION AND CHARACTERISATION OF AUTOSTIMULATORY CLONESA) CHARACTERISTICS OF FD.C/2 CELLS.i) IL-2 and IL-3 growth responses.Prior to any attempt to create autostimulatory clones, the growth of FD.C/2 cells in response to variousconcentrations of IL-2 and IL-3 was determined. When cells were maintained in medium containingboth IL-2 and IL-3, growth responses were as represented in figure 1.1. Ample responses to both IL-3and IL-2 were observed, but the IL-3 response tended to be stronger than the IL-2 response, in termsof maximal thymidine incorporation. When cells were grown in IL-3 in the absence of IL-2 for 10passages ( 50 days ), the proliferative response to IL-2 was very poor in comparison to the response toIL-3 ( fig. 1.2A ). When such IL-3-passaged cells were washed free of IL-3 and replated in IL-2 alone,the cell population was observed to undergo a crisis such that there was a net decrease of viable cellnumbers, with death of some 90% of the population over 2 weeks. Most of this death was apparent inthe first 48 hours, followed by the outgrowth of the remaining cells. Cells that had been \"converted\" toIL-2-dependent growth in this manner showed similar maximal responses to IL-2 and IL-3 in 3H-thymidine incorporation assays ( fig. 1.2B) . Moreover such IL-2-passaged cells did not undergo acrisis when reconverted to IL-3-dependent growth, but continued to proliferate in the absence of IL-2( but the presence of IL-3 ) without any evidence of cell death, or decline in proliferation rate. Theseresults are consistent with the findings of Le Gros et al. ( 1985 ), who examined the conversion of theparent IL-3-dependent cells ( FD.C/1 ) into FD.C/2 cells. These authors were unable to detect the p55chain of the IL-2 receptor on the surface of the FD.C/1 cells, but after the conversion crisis, theresulting growing cells, which they designated FD.C/2, displayed \"readily detectable\" levels of p55 asdetermined by FACS analysis. Furthermore, addition of IL-3 to these FD.C/2 cells growing in thepresence of IL-2 did not affect levels of cell-surface p55, whereas withdrawal of IL-2 and replacementwith IL-3, while allowing continued growth, resulted in a 50% decline of cell-surface p55 within 20hours ( ibid. ).Very high concentrations of murine IL-4 permitted the survival of some FD.C/2 cells for a few days,yielding a 3H-thymidine-incorporation curve as seen in figure 1.1, part B. The cells could not,however, be maintained in passage medium containing IL-4 alone.120000100000 -—0-- thIL-2 (100U/ml)CACA (44M°20000 -CPMA18Factor dilutionB 30002000 -CPM —a— miIL-4 (10000/m1)1000 -I^12^4^16 32 64 128 256 512 MA.Factor dilution2^4^8 16 32 64 128 256 512 M.A.Fig. 1.1. Factor responses of FD.C/2 cells maintained in IL-2 and IL-3 A) FD.C/2 cells that had beenpassaged in a mixture of IL-2 and IL-3 were plated at 500 cells per 10 per well in Terasaki HLAmicrotiter plates on titrations of IL-3 ( CACA - a synthetic murine IL-3 kindly provided by Dr. IanClark-Lewis of the Biomedical Research Centre ) or IL-2 ( recombinant human IL-2 ). After 48 hoursof incubation, wells were pulsed with 3H-thymidine and incubation continued a further 12 hours beforeharvesting. B) FD.C/2 cells that had been passaged in a mixture of IL-2 and IL-3 were assayed forproliferative response to mouse IL-4. 48 hour incubation followed by 12 hour 3H-thymidine pulse. Theconcentrations of factor indicated in the legend represent the final concentration in the first well of thetitration. The units given are as calculated from the manufacturers specifications ( see Appendix 1 ).Error bars in this and all subsequent 3H-thymidine assays show standard error of the mean of .triplicates. M.A. indicates medium alone - i.e. without growth-factor.B--a-- CACA (2114/m1)(100U/n1)CPM14000 ^12000 -10000 - 4000 -2000 -019A20000CPM 10000 - CACA (211g/m1)--e r-- rhtL-2 (100U/m1)2^4^8 16 32 64 128 256 512 M.A.Factor dilution2^4^8 16 32 64 128 256 512 MA.Factor dilutionFig. 1.2. Factor responses of FD.C/2 cells maintained in IL-2 or IL-3 A) FD.C/2 cells that had beenpassaged in IL-3 were assayed for response to IL-2 and IL-3 as in figure 1.1. After 36 hours ofcultivation, wells were pulsed with 3H-thymidine and incubation continued a further 6 hours beforeharvesting. B) FD.C/2 cells that had been passaged in IL-2 were assayed for response to IL-2 and IL-3.20In the absence of exogenous IL-2 or IL-3, FD.C/2 cells were not only unable to incorporate 3H -thymidine ( M.A. in figure 1.1 ), but died within 24 hours of withdrawal of growth factor. This conclusionwas supported by phase-contrast microscopic inspection of proliferation assays such as that shown infigure 1.1, and by a separate experiment in which cells were washed free of factor-containing passagemedium, and replated in 5 ml cultures in medium without factor at 10 5 cells/ml. After 24 hours, thesecultures were harvested by centrifugation and the cells resuspended in 10 pi of medium. Ten pl ofeosin ( 2 cultures ) or trypan blue ( 2 cultures ) were added and the suspensions were examined bylight microscopy. In no instance were dye-excluding cells seen, confirming the absolute dependenceon FD.C/2 cells for exogenous growth factor for maintenance of viability. FD.C/2 cells did notdemonstrate detectable responses to GM-CSF or IL-6.While the FD.C/2 cells grow in \"suspension\", i.e. do not adhere to tissue culture dishes, their densityand physical characteristics are such that they will roll or otherwise accumulate under the influence ofgravity at the lowest point in a culture well or plate, and proliferation from a single cell at the edge of anundisturbed culture vessel will give rise to a \"carpet\" of adjacent cells. These characteristics facilitatedassays involving assessment of growth in 96-well culture plates, and cloning procedures, describedbelow.ii) Determination of G418 resistance of FD.C/2 cells.Preliminary experiments to determine the level of G418 to be used in selection of transfectants weredesigned as follows: 2 x 103 cells were plated in 1 ml of medium containing IL-2 in the wells of a 24 wellculture plate. G418 was added to the wells so as to achieve final concentrations ranging from 0 to2000 pg per ml in increments of 200 pg/ml. The cells were observed daily with a phase-contrastmicroscope. At concentrations of 1000 p.g/m1 and above, there was no apparent proliferation and allcells appeared dead by 48 hours. At concentrations of 400 pg/m1 and below, surviving cells wereapparent at up to 4 days of culture. By day 6, however, all cells appeared dead at 400 p.g/ml. On thebasis of this titration, it was decided that 500 pg/m1 G418 be used for subsequent selection oftransfectants.B) ISOLATION OF TRANSFECTED CLONES.The 112-3.1 cell line is a mouse fibroblastoid cell line that produces infectious but packaging-defectiveretrovirus, which contains hIL-2 and neomycin-resistance sequence, from the transposon Tn5 ( fig.211.3, Yamada et al., 1987 ). Cells infected with this virus can be selected in the neomycin analogueG418, and the expression of G418-resistance and of hIL-2 are under the control of Moloney leukemiavirus LTR in such cells. T2-3.1 cells were cocultivated with FD.C/2 cells in tissue-culture treated Petridishes in conditions such that a given dish might be left for 5 days without overgrowth of either cellpopulation. 2 x 105 cells of the 'F2-3.1 line were seeded onto dishes, and allowed to adhereovernight. The following day, dishes were gently agitated by hand, and the medium was removedalong with any dead or otherwise non-adherent cells. The medium was replaced with mediumcontaining mIL-2, in which were suspended 4 x 10 5 FD.C/2 cells. After 5 days, the non-adherent cellswere aspirated, washed by centrifugation twice, and plated at a density of 50 cells/m1 in soft agar withmedium containing mIL-2 and G418 at 500 µg/ml. Colony growth was observed over the next 10 - 14days, and 8 individual colonies were plucked with a pulled Pasteur pipette, and plated into individualwells of a 24-well plate, each containing 1 ml of mIL-2 medium and G418 at 50011g/ml.Cells were allowed to grow out of the agar into the liquid medium, and when the cells formed a carpetof over half the floor of the well, the contents of each well were gently resuspended and 500 ttl of thecontents of each well were transferred to an empty well, and 5000 of medium containing G418, butno IL-2, was added to these fresh wells. This process of weaning from exogenous IL-2 in thepresence of G418, was continued until, between two and three weeks after the initial plucking ofcolonies, the contents of a well were resuspended and washed free of any remaining exogenous IL-2that might have been carried over from the original plating, and replated in fresh medium containingG418, but no IL-2, at a density of 2 x 10 5 cells per ml. This high density was selected on the basis ofthe density tolerated by parental cells in super-saturating levels of IL-2, and in the expectation that theclones would show density-dependent growth ( see Hoffer et al., 1987, and section D, below ). Of the8 colonies initially plucked, 5 survived this weaning process, the remaining 3 dying within the firstweek of weaning. The cells of these surviving populations resembled the parental FD.C/2 cellsmorphologically ( under phase-contrast microscopy ), and showed no tendency to adhere to tissue-culture plastics. These 5 clones were designated FD.0/2'P.1 through '11.5.After this weaning process, cells were passaged in medium without IL-2 by 1:1 dilution every secondor third day, following inspection of the plates. While such passages were routinely performed in theabsence of G418, a succession of three passages in the presence of 500 µg/ml G418 wereperformed approximately once every two months, with little or no apparent cell death attributable tothe presence of G418.H h11.-2 cDNA 1Fig. 1.3. The proviral construct used to make IF2-3.1 cells. LTR - Moloney Murine Leukaemia VirusLTR, NeoR - Neomycin resistance gene from Tn5, SV40 on - origin of replication of SV40 virus, pBRon - bacterial origin of replication.23C) GROWTH CHARACTERISTICS OF AUTOSTIMULATORY CLONES.i) Confirmation of FD.C/2 derivation of FD.C/2T cells - response to exogenous growth factors.The morphology of the resultant cloned cells, and the process in which they were created ( since non-transformed fibroblasts, such as the 11,2-3.1 cells, are unable to grow in agar ), made it unlikely that theclones were of other than FD.C/2 origin. Assays of the response of these clones to exogenous IL-2and IL-3 were performed to confirm that these cells were FD.C/2 derived and that IL-2 and IL-3responsiveness were maintained, despite growth in the absence of exogenous IL-2 or IL-3. Initialexperiments using the same cell-density as used in bio-assays of parental FD.C/2 cells ( 500 cells in10111 per well of a Terasaki microtiter tray ), while revealing some apparent response to IL-2 and IL-3,showed that such response might be largely concealed by the degree of proliferation evident in theabsence of added factors ( fig. 1.4 ). The bio-assays were repeated at lower cell-densities ( 200cells/100well ) , and over a broader range of IL-2 and IL-3 doses to increase the likelihood of coveringthe range of concentrations at which an effect would be apparent. The results showed that all 5 lineswere responsive to exogenous IL-2 and IL-3 ( fig. 1.5 ). A comparison with IL-2-passaged parentalFD.C/2 cells of the responsiveness to IL-2 and IL-3 was undertaken. The results of this comparison inone clone are shown in figure 1.6. Interestingly, while responses to exogenous IL-3 were qualitativelysimilar, the shape of the IL-2 response curve of the FD.0/2'P.1 cells was very different from that of theparental FD.C/2 cells - significantly less IL-2 being needed to achieve maximal growth stimulation ofthe 'PA cells. Figures 1.4, 1.5 and 1.6 show that the FD.0/2'P cells are capable of incorporating 3H-thymidine in the absence of any exogenous factor.ii) Density dependence of FD.C/2 111 clones.Although proliferation continued in the absence of exogenous IL-2 for many passages ( over 100 forsome of the clones ), a few dead cells ( up to 5% of the total cell number ) were often observed at thetime of passage. When populations were depleted of the majority of dead cells, by Ficoll gradientcentrifugation, it was possible to maintain a higher level of viability ( > 98% viable cells, as judged byeosin uptake ) by passaging daily on a 2:1 dilution basis ( 2 volumes of existing culture to 1 volume offresh medium ). Parental cells maintained in IL-2 ( or IL-3, or IL-3 + IL-2) could be maintained at >99.5% viability when cell density was kept low enough to keep cells in log-phase growth ( belowapproximately 5 x 105 cells/ml ). Allowing cells of autostimulatory clones to fall to lower densities thanthose produced by the above passage conditions ( i.e. to below approximately 1 x 10 4 cells per ml ),resulted in decreased viability of these cell populations.ACPM2^4^8^16 32 64 128 256 512 M.A.Factor dilutionB90000 ^80000 -70000 -60000 -50000 -40000 -30000 -20000 -10000 -0 CPMPsi.4Psi.52^4^8^16 32 64 128 256 512 M.A.Factor dilutionFig. 1.4. FD.C/2`1' cells showed some response to IL-2 and IL-3 Cells of two of the FD.C/2T clones( Psi.4 and Psi.5 ) were assayed for responsiveness to exogenous IL-2 ( A ) and IL-3 ( B ). 500 cellsper well. 48 hours incubation followed by 12 hour 3H-thymidine pulse. The growth-factors were usedin the form of supematants and the highest concentrations were above the level required to giveoptimal responses in the FD.C/2 line ( and other IL-2 responsive cells ). In this and subsequentfigures where units of activity are not given, growth-factor has been applied from sources withoutprevious calculation of the units of activity, since interpretations of the results shown rely on internalcomparisons ( see Appendix 1 ).Psi.4Psi.5coO a) <0CO 0.<0 OfFactor dilutiondilutio(\"1 .^.^.0 2 2 2C;<0COCO04 sr <0 CO 01 sr CO 10 OA stl7 10 Of 4114- 01 40 0CO CO`41- Of0 004 srA25PSI.2PSI.3PSI.4PSI.5B• -■^-PSI.1PSI.2PSI.3PS1.4PS1.504 st <0 410 01 sr cotn co os(0 N^co ID 01 ..1* co cD oi111^04 st 03 CO CO CO CI^. . .• OMMM01 10 0 0 0(0▪ 40 s-01 sr 173 40 12 kb ). In a simple head-to tail array integration,.a single such \"flanking\" band would beexpected at a position determined by the nearest chromosomal EcoRI site on the 5' side of theintegration ( with respect to the orientation of the transgene ). Similarly, an Xbal flanking band wouldbe expected at the 3' side of the integration, and such a band may be visible ( depending on thequality of the reproduction ) in lane 3. That this Xbal \"flanking\" band appears of similar size to the EcoRI\"flanking\" band, is unimportant, since size discrimination is very poor at this \"upper\" end of gels, andan actual similarity of size would be coincidental. Since this hGH probe might be expected to hybridiseweakly to mouse growth hormone, sequnces, no firm concluisons can be drawn from any additionalbands that may be apparent on this Southern analysis.ii) Figure 2.17, part A, lanes 3 and 4.This analysis is of tail DNA, so no recombination products are present. The LCK probe has picked up asingle band in wild-type mouse DNA, due to the presence of the 2 copies of the enodogenous LCKgene. Digestion appears complete. The band resulting from the presence of the 1017BGAL169transgene ( lane 5) is, as expected, of approximately the same size as that from the 1017CREconstruct ( fig. 2.16 ). Therefore, no difference is seen between the bands from the doubly transgenicanimal ( lane 3) and the 1017CRE transgenic ( lane 4 ). In both lanes 3 and 4 two \"minor\" bandsappear at sizes of approximately 4 kb and 8-9 kb. The intensity of these bands suggests a copynumber of the corresponding DNA fragment of the order of that of the endogenous lck gene. Thesimplest explanation for these bands is that they represent true flanking sequences at EITHER end ofthe integration array, involving a truncation of the terminal copies of the transgene construct at pointscentral to the outer Sacl sites at each end ( i.e 3' to the most 5' expected Sacl site, and 5' to the most 3'expected Sacl site ) Such true flanking sequences would, however, be expected to produce bands ofapproximately HALF the intensity of the endogenous Ick band. Although more elaborate explanationsfor these additional bands might be entertained, it is not clear that the appearances of this Southernwarrant further interpretive effort.iii) Figure. 2.17 part B, lane 3 ( and figure 2.18, lanes 3,4, and 5; figure 2.26, part A; figure 2.27, lanes2 and 3; figure 2.28, lanes 2,5, 6, 7.).These Southerns show thymocyte DNA from wild-type and singly transgenic mice, probed with theLCK probe, and doubly digested with Stul and Mscl. The appearances suggest close to completedigestion. The endogenous lck gene fragment runs at between 1kb and 1.6 kb. In this instance, inaddition to the endogenous lck band, and a predominant band near 3kb that would be expected fromthe 1017CRE array ( fig. 2.16 ), a band of somewhat lesser intensity than that of the endogenous Ickband is present at between 2 and 3 kb. Such a fragment is consistent with the hypothesis of both3'and 5' true flanking sequences, suggested above, and, in the simplest interpretation, would be dueto the truncation at the 5' end, since a truncation here at a point 3' to the Sad site would also be 3' tothe unique Stul site in the construct. At the 3' end Sad site, it is possible that a truncation existsbetween the Mscl site and the Sad site, so that no flanking sequence from this end would beapparent with the StuVMscl double digest.iv) Figure.2.21, lanes 2 and 3.These lanes contain DNA from a doubly transgenic animal, digested with BgIII. Digestion appears nearto complete. The explanation of minor bands offered above leads one to expect two \"flanking\" bandsof the 1017CRE array to be present. The additional LCK-hybridising band immediately below the170expected 1017CRE band at between 3 and 4 kb ( see fig. 2.16) may represent one such band, andthe blurry signal at about 7-8 kb the other.B) The 86 line 1017L0X-BGAL-LOX array.i) Figure 2.15, lanes 6,7,8.The hGH probe of singly transgenic 86 line DNA shows the majority of the integration structure to beof the form illustrated in fig. 2.16 part B. The appearance of the Southern suggests that the Xbaldigest was incomplete. There is a faint suggestion of a flanking band in the EcoRI lane. Such a band-would be expected at the 5' end of the integration array.ii) Figure 2.17, part A, lane 5; figure 2.17, part B, lane 4; figure 228, lanes 3, and 5-7.Both Sac! digest and Stul/Mscl double digests probed with the LCK probe show bands of severaltimes the intensity of the endogenous Ick band at the sizes expected. No other bands are visible withthese digestsiii) Figure 2.20, part C, lanes 3 and 4.This Southern, although a Sacl digest probed with the LCK fragment, shows a band of very lowintensity in comparision to that of the endogenous lck fragment, at about 4-5 kb in lane 3. Thisappearance suggests it may be due to incomplete digestion. The same applies to the DNA fromdoubly transgenic tail in lane 4. In this case the additional bands at between 4 and 5 kb, and above 7kb, may be due to incomplete digestion of the CRE array.iv) Figure 2.26, part B, lanes 2-6.This tissue survey from a doubly-transgenic 86 line animal shows almost complete disappearance ofthe BGAL array fragment from thymocytes compared with the band of expected size at between 5 and6 kb in DNA derived from other tissues. This Bglll digest, probed with the LCK fragment, also showsthe endogenous Ick band at about1 kb. However, the expected recombination product at between 2171and 3 kb is NOT seen in this autoradiograph ( nor in a longer exposure of the same blot ). This contrastswith the Southern of figure 2.21, another Bglll digest probed with the LCK fragment,.which doesshow a band of this size in thymocyte DNA from another doubly transgenic 86 line animal. Bglll digestsof thymocyte DNA from a third doubly transgenic 86 line animal ( not shown ) failed to reveal the bandof the expected size of the recombination product. A possible explanation for this anomaly is asfollows: a truncation of the 1017L0X-BGAL-LOX construct has occured between the Sad and Bglllsites at either the 5' or 3' ends of the array, such that the recombination product of Bglll digestion INanimals in which the whole array has been collapsed by recombination is either too large to transferfrom the gel to nitrocellulose at sufficient levels to be visible, or is hidden by another band. The reasonthat a recombination product of the expected size is nevertheless visible in the Southern of figure2.21, is that in this animal the array was not completely collapsed, and the product seen is the result ofrecombination in copies of the transgene that are internal to the terminal copy which suffered thetruncation. Thus, the recombination product of expected size is still present, whereas in mice in whichthe array recombination has gone to completion ( in the majority of thymocytes ), the copiesresponsible for the visibility of this band have been \"recombined out\".C) The 87 line 1017L0X-BGAL-LOX array.i) Figure 2.29, part A, lanes2-5.Ths Sacl digest of DNA from two doubly transgenic animals of the 87 line, probed with the LCKfragment, shows a band of near 1.6 kb present at equal intensity in both tail and thymocyte samples.This band has never been seen in digests of DNA from singly transgenic 1017CRE animals, andappears despite the presence of the expected recombination product in thymocytes of both animals.It mayhave been that this anomaly was due to a truncation or other integration \"pathology\" thatremoved a LOX site, since these appearances suggest this Sacl fragment was not amenable torecombination.ii) Figure 229, part B, lanes 2-5.The Bglll digest of the same DNA samples as discussed in section i, above, probed with the LCKfragment, also suggests an integration anomaly. Some degradation is evident, especially in tail DNA,but digestion appears near to complete. Although the expected recombination product is present, atbetween 2 and 3 kb, there is a hint of a band in tail DNA at nearer to 3 kb, which seems fainter in172thymocyte DNA, and the additional appearance in thymocyte DNA of a band at about 10kb Theintensities of the new bands in relation to that of the endogenous lck band, and of the original andremnant BGAL bands, suggest that the array collapse may not have gone to completion in theseanimals.iii) Figure 2.30, part A, lanes 2 and 3; part B, lanes 1 and 2.A Stul/Mscl digest was performed on the tail and thymocyte DNAs derived from one of the two animalsof sections i and ii, above. Aliquots of these digests were analysed by probing either with the LCKfragment ( part A) , or the LOX probe ( part B ). The LCK probe shows the expected recombinationproduct at near 1.6 kb, and an additional product in thymocyte DNA of about 7-9 kb. The LOX-probedblot of the same samples shows at least three LOX-hybridising fragments in tail, which seem todiappear or diminish in intensity,in thymocyte DNA: that of the expected size ( at 5-6 kb), and muchmore faint bands at near 4kb and near 2 kb. Correspondingly, as the recombination product ofexpected size ( near 1.6 kb ) appears in thymocyte DNA, so also does a faint signal at the 7-9 kb area.An additional ver faint signal is apprent in both tail and thymocyte DNAs, unchanged, at about 3-4 kb.This latter is unlikely to correspond to the band of about 1.6 kb seen in the 87 line Sac! digest ( sectioni, above ), since the intensity relative to that of the LOX-BGAL-LOX array band is much less than thatseen in the Sac! digest. ( On the other hand, use of the LOX probe would be expected to give a muchlower relative intensity due to relative exaggeration of the intensity of the LOX-BGAL-LOX array band,as every copy of the transgene construct in this array harbours TWO LOX sites. ) This very faint signalmay reflect weak hybridisation of the LOX probe to some pUC sequence present in a few copies ofthe CRE transgene construct, due to the presence of some polylinker sequence in the LOX fragmentused as probe. The complexity of the 1017L0X-BGAL-LOX integration structure in the 87 line seemsto defy a simple explanation ialong the lines of the common integration structure anomalies ( Brinsterand Palmiter, 1986 ).173MATERIALS AND METHODS1) TISSUE CULTURE A) Passaging and proliferation assaysAll cell culture was performed in RPM' 1640 medium ( Terry Fox Laboratories, Vancouver B.C. )supplemented with 10% v/v fetal or newborn bovine serum ( Hyclone, Logan, Utah ) and 50 p,M 2-mercaptoethanol ( Aldrich, St. Louis, MO ). Cells were incubated in humidified 37° C incubators in anatmosphere of 5% CO2 in air. Factor dependent cell lines were passaged by making 5 serial 1 volumeto 5 volumes dilutions in fresh medium on the day of passage ( every 5 - 7 days ). Passaging wasusually performed in 60 x 15 mm culture dishes ( Corning, Corning, NY, ), and occasionally in the 1 mlwells of 24-well tissue culture plates ( Falcon, BD Labware, Lincoln Park NJ ). Pipetting of cells wasdone with calibrated serological pipettes from Baxter ( Burnaby, B.C. ) and a Pipetaid pipettor( Drummond Scientific Co., Brodwall, PA ). Adherent cells were passaged and assayed in tissue-culture treated dishes and trypsinised as necessary after washng the cells free of serum in twochanges of 10 mls of PBS. All washing of mammalian cells ( other than prior to electroporations ) wasdone at room temperature in a Canlab ( Edmonton, Alberta ) Labofuge B centrifuge, at 1200 rpmapproximate average relative centrifugal force of 400 ) for 4 to 5 minutes. Centrifugation wasperformed in 10 ml polypropylene or polystyrene tubes ( Disposable Products, Adelaide, Australia ) or15 ml polystyrene or 50 ml polypropylene tubes ( Falcon ). Passaging medium for factor-dependentcells contained super-saturating amounts of the appropriate growth-factor, generally between 4 and10 units as determined on the cell line being passaged ( Part 1, Appendix 1 and section B, below ).Cells were thus maintained in log-phase growth, generally at below 2 x 10 5 cells per ml.AUTOSTIMULATORY cells were passaged as described in the text.Cells were counted in a 1:1 mixture of suspended cells in PBS containing eosin ( 0.2% w/v - BDH,Vancouver, B.C.) or trypan-blue ( 0.4% BDH ) by examination in a hemacytometer with an Olympusmicroscope. At least 200 cells were counted for each enumeration. Cells were washed prior to allassays by repeated centrifugation with 3 changes of 8 mls of medium or Hanks' buffered salt solutioncontaining 2% v/v newborn bovine serum, and one centrifugation in assay medium, prior toenumeration. Medium for proliferation assays was always supplemented with fetal rather than newbornbovine serum. Titrations were performed by making serial 2-fold dilutions of growth-factor, supernate,antibody etc. in the wells of round-bottomed 96-well plates, or in the wells of 24-well tissue-culture174plates ( Falcon ), and applying 5g! of each dilution to the wells of 60-well Terasaki microtiter plates( Disposable Products, ), in triplicate. Resuspended cells were then applied in 5111 of medium. Platingin Terasaki wells was performed with an EDP automated pipettor ( Wobum, MA ). When the assay hadbeen plated out, approximately 200 -300 of medium was applied to the inner circumference of theTerasaki plates to maintain adequate hydration of the culture wells. After the initial period ofincubation, 250 nCi ( approximately 9.25 kBq ) of 3H-thymidine ( ICN, St. Laurent, QUE ) was added in5 RI of medium to each well, and incubation allowed to proceed further. In the last 30 - 60 minutesbefore harvesting the assay, Terasakis were inverted to encourage cells to gather in the culture liquiddrop, facilitating their harvesting onto pre-punched glass-fibre sheets ( Titertek, Flow Laboratories Inc,Mississauga, ONT ). Harvesting was performed using tap water, a Venturi device and a specialisedharvesting platform allowing culture liquid to be washed through the sheets with double-distilledwater, and subsequent fixing of DNA to the sheets with methanol. Scintillation cocktail ( BetaMax,ICN ) was applied, and counting was performed using a protocl of 2 minutes count per sample in ahumidified room, which minimised interference due to accumulation of static electrical currents on theglass-fibre or the beta-counter ( Taurus Autopmatic Scintillation Counter, Micromedics - ICN ).Pipetting of volumes of less than one milliliter, for cell-biology, biochemistry, and molecular biologyapplications, except for the plating of Terasaki assays ( above ), was performed with Socorex ( Renes,Switzerland ) or Gilson ( Villiers-Le-Bel, France ) pipettors. Pipette tips were from Island Scientific( Seattle, WA )or Costar ( Cambridge, MA ).Soft agar ( Difco, Detroit, MI) cloning was performed by making a 3% w/v slolution of agar in steriledouble-distilled water, heating the agar by flame or in a micro-wave oven until it had dissolved and thenallowing it to cool to 37° C prior to adding one volume of agar to nine volumes of medium containingcells and appropriate growth factor. The mixture was then mixed by pipetting and plated into 60 x 15mm tissue culture dishes which were rapidly transferred to a metal surface in a 4° C environment, todiscourage cells from sinking to the bottom of the dishes before the agar had gelled. Plucking ofcolonies from agar was performed on the stage of an Olympus dissecting microscope with a pulledPasteur pipette, in a sterile laminar outflow cabinet. Ficoll density-gradient centrifugation wasperformed with Ficoll-Hypaque ( Pharmacia, Baie d'Urfe, QUE ), according to the manufacturersinstructions.B) Sources of growth-factorsSeveral of the growth-factors used were produced as symthetic poly-peptides by Dr. Ian Clark Lewis ofthe Biomedical Research Centre, as indicated in the text and figures. Recombinant human IL-2 was175the kind gift of Cetus Corporation. Murine IL-2 was obtained as the supernatant of a hybridoma cell line( X63-Ag8/653 ) that had been engineered to secrete IL-2 by transfection with an expression vectorcarrying the murine IL-2 cDNA., the kind gift of Dr. Fritz Melchers of the Basel Institue of Immunology.Similar supernatant was also used as a source of IL-3. These supemates were obtained from 24 - 48hour growth at high densities ( approximately 10 6 per ml ), and were filtered through 022 gm Millex GVsyringe filters ( Millipore, Bedford, MA) prior to use. Supematant of the myeloid leukemia cell WEHI-3Bwas also used as a source of mIL-3, in this instance concentrated 10-fold prior to use as a 2% v/vsolution in passaging medium. Recombinant mIL-4 was from Cetus corporation. Mouse IL-6 wasobtained as the supernatant of the fibroblast cell line Psi2, and human IL-6 as supernate of thelymphoblastoid cell line LK28, derived by Dr. John Schrader at the Walter and Eliza Hall institute ofMedical Research, Melbourne, Australia. All concentrations were performed with Amiconconcentrators ( Amicon Canada, Oakville, ONT ).G418 was used as Geneticin powder ( Gibco BRL, Burlington, ONT ). The manufacturer's specificationof active weight was used to calculate the concentrations used. Hygromycin B was in powder formfrom Boehringer Mannheim ( Boehringer Mannheim Canada, Laval, QUE ), and concentrations werecalculated on a simple weight per volume basis. The fluoresecent dye PKH2 was from MolecularProbes Inc. ( Eugene, OR ). Polybrene was obtained from Sigma. ( St. Louis, MO ).C) AntibodiesMonoclonal antibodies were derived from tissue-culture supernatants of hybridoma cells or ascites.For production of ascites of mouse monoclonal antibodies ( DMS1, DMS2 and Mouse Anti-Rat Kappa- MAR 18.5 ), Balb/C mice were primed by intraperitoneal injection of pristane ( Aldrich ) 5 days to 5weeks before intraperitoneal injection of 1 - 5 x 10 6 hybridoma cells. For production of ascites of ratantibodies ( PC61, 11811, and 6B4 ), Balb/C mice were pristane primed and treated with sub-lethalgamma irradiation and hydrocortisone injection, prior to introduction of hybridoma cells, according tothe protocol of Weissman et al., 1985. Mice were sacrificed in carbon dioxide when abdominalswelling due mto ascites was apparent, but before the animals displayed other signs of the tumourburden. Rabbit antibodies Rab7 and Rab39, were used as serum and were the kind gift of Dr.Hermann Ziltener. The rabbit anti-rat immunoglobulin anti-serum was raised as indicated in Part 1,Appendix 2 by initial subcutaneous injection of approximately 50 gg of partially purified ratimmunoglobulins in Freund's complete adjuvant ( Difco ) , followed by booster injections containingapproximately the same amount of protein in incomplete adjuvant after 2 weeks, and every 4 weeksthereafter, for a total of 6 injections.176Fluorescence activated cell sorting and analysis was done by incubating approximately 10 6 target cellsin 1 ml of a solution of 2% newborn bovine serum in PBS, containing primary antibody for 30 minutesto 1 hour at 4° C with periodic agitation to prevent settling of cells. Primary antibody was used at 1 jigper ml ( in the case of PC61 ) or 1:2000 v/v dilutions of commercial CD4 and CD8 antibodies ( Becton-Dickinson, San Jose, CA) rehydrated according to the manufacturers specifications. After incubation,cells were washed twice in 10 mls of the incubation solution without antibody, and then resuspendedin a similar 1:2000 v/v solution of appropriate fluorescein-coupled secondary antibodies ( KierkegaardPerry Laboratories, Gaithersburg, MD) rehydrated according to the manufacturers instructions. Afterfurther incubation at 4° C for an hour, cells were again washed three times, and resuspended in 2%serum in PBS prior to sorting or analysis in the University Hospital Acute Care Unit facility with aFACStar sortet ( Becton-Dickinson ) by Mr. Dan Zecchini.D) Gene transfer into cells.Retroviral gene transfer was performed as described in the text. Cells were prepared forelectroporation as follows: 1 to 2 x 107 cells were washed by centrifugation in PBS twice, to removeserum, and then resuspended in 0.8 mls of PBS. If plasmid was linearised prior to electroporation,restriction digest was followed by standard phenoVchloroform extraction, precipitation with ethanol,and washing by centrifugation in 70% ethanol ( as per Maniatis et al., 1982 ). All plasmids wereresuspended at a concentration of 15 to 20 p.g in 20 .d of sterile double-distilled water. The cellsuspension was pipetted into the electroporation cuvette ( Bio-rad, Hercules, CA ), and the 20 plplasmid suspension added to it. The mixture was allowed to sit in ice for 5 to 10 minutes and thengently agitated to resuspend cells prior to electropration. Electroporations were performed asspecified in the text, using a Bio-rad Gene Pulser. After electroporation the cuvette was returned toice for 5 to 10 minutes prior to transfer with a disposable soft pipette into 10 mls of pre-warmedmedium in a 10 ml tissue culture flask. After 48 hours of incubation, cells were harvested andresuspended at a concentration of 5 x 104 to 1 x 105 viable cells per ml in medium containing G418 orhygromycin, and the suspension distributed into the wells of 96-well tissue culture plates. Cuvetteswere reused after soaking overnight in double-distilled waterN, rinsing with 70 % ethanol, andsubjecting to 2.5 x 106 rads of gamma irradiation in a Gammacell unit ( Nordion ).1772) BIOCHEMISTRYA) AntibodiesRabbit antibodies were used as serum, or purified by protein-A sepharose ( Pharmacia ) affinitychromatography ( as per Goding, 1986 ). Ascites and supemates were concentrated by ammoniumsulphate precipitation prior to further purification by antibody affinity chromatography ( ibid. ). Affinitycolumns were prepared with cyanogen bromide activated sepharose beads ( Pharmacia ) according tothe manufacturers instructions. Columns were run in PBS and eluted in 0.1M glycine pH 2.5. Proteinwas quantitated for coupling to beads, by reading absorption at 260 nm wavelength using a Hewlett-Packard ( Edmonton, Alberta ) -spectrophotometer. Other quantifications were performed using theBradford modification of the Lowry method ( Bradford, 1976) with reagent from Bio-rad, and astandard curve made with bovine serum albumin ( Sigma ). Readings were performed in a Phillips PyeUnlearn spectrophotometer ( Cambridge, England ).Enzyme linked immuno-sorbent assay ( ELISA ) was performed at room temperature in round-bottom96-well polyvinlychloride trays ( Falcon ) coated overnight with the appropriate capture reagentdiluted in PBS supplemented with 0.5% sodium azide. Incubation of sample and developingantibodies were for 1 to 2 hours each at room temperature. Sample titrations ( serial two-fold dilutionsin triplicate ) and subsequent antibodies were applied in 5% skim-milk powder in PBS with azideexcept the distilled water wash prior to addition of peroxidase substrate. Commercial entibodies wereused at concentration recommended by the manufacturers ( Kierkegard Perry Laboratories, orCalbiochem, La Jolla, CA ). The reactions were developed with 0.5 mg/ml of the chromogenicsubstrate 2,2u-azinobis(3-ethylbenzthiazoline sulfonic acid) ( Sigma, ) in citrate buffer ( pH 6.5 )containing 0.006% v/v hydrogen peroxide ( 1/500 dilution of 3% stock kept in the dark at 4° C ).Colour was allowed to develop at 37° C for 20 minutes and results were read as comparisons ofabsorption at 405 vs 490 nm for each well, in an EL309 Microplate Reader ( Biotek Instruments Inc.,Burlington, VT ).B) Protein analysesAnalytical protein gels were run according to the manufacturers instructions using the PharmaciaPhast mini-gel system with reagents from BDH.178Protein extraction from thymocytes was performed by resuspending PBS-washed cells at 4° Cin anEppendorf tube ( Brinkmann Instruments Inc., Rexfdale, ONT) at a concentration of 2 x 10 6 cells per80111in a solution of 20 mM tris(hydroxymethyl)aminomethane ( Tris, pH 8.0 ), 2 mM ethylenediamine-tetraacetic acid ( EDTA ), 1% VN Triton X-100, 137 mM sodium chloride, 1 mMphenylmethylsulphonylfluoride, 10 µg/ml leupeptin, 10 gg/m1 soybean trypsin inhibitor, 1 pMpepstatin. Buffer reagents for this procedure were from Sigma and BDH, and protease inhibitors fromSigma and Boehringer Mannheim. The cell solution was then centrifuged for 1 minute at 4° C in adesktop refrigerated Eppendorf centrifuge at 14,000 rpm ( approximate average relative centrifugalforce of 14,000 ). Supernatant was transferred to another Eppendorf tube, and 20 pi of solubilisationbuffer added. Solubilisation buffer was made as follows: to 31 mls of 50 mM Tris ( pH 6.8) was added 5g of sodium dodecyl sulphate ( SDS ), 25 ml glycerol, 0.5 ml of a saturated solution of bromophenolblue ( all from BDH ). This solution was stored frozen and prior to use 15 mg/ml of dithiothreitol( Boehringer Mannheim ) was added.3) MOLECULAR BIOLOGYA) Plasmid manipulationsAll plasmids were grown in E. coli strain DH5a. For both liquid and agar cultures, bacteria were grownin either Luria Broth ( Gibco-BRL ) or 2 x YT ( ingredients from Difco ), made from powder ( as describedin Sambrook et al., 1989 ). Competent cells were prepared using a variant of the method of Maniatis etal. ( 1982 ). After the initial incubation in 100 mM calcium chloride, cells were centrifuged,resuspended in 50 mM calcium chloride, and incubated at 4° C overnight. Glycerol ( BDH) was thenadded ( from autoclaved 75% w/v stock ) to a final concentration of 15%, and 210 ptl volumes of thecell suspension were aliquoted into 1.5m1 Eppendorf tubes for storage at -70° C. For transformations,100 or 200 p1 of thawed competent cells were added to 10 or 20 pcl of ligation reaction respectively,and the mixture was left in ice for 30 to 45 minutes. Then the mixture was transferred for 1.5 minutesto a 42° C waterbath, and returned to ice for 10 minutes prior to incubation with 1.3 ml of broth in ashaking incubator at 37° C for one hour. Cells were recoverd by centrifugation and plated onto agar( 1.2 % w/v in 2 x YT) in 100 x 15 mm plastic Petri dishes ( Fischer Scientific, Ottawa, ONT. ). Agarcontained 100 pg/ml ampicillin ( Sigma ) when selecting for ampicillin-resistant transformants, and 12.5i.t.g /ml kanamycin ( Boehringer Mannheim ) when selecting derivatives of the pBMGNeo plasmid.All small scale manipulations of bacteria, DNA and plasmids were performed in Eppendorf tubes.Centrifugations were performed in desktop Eppendorf high-speed centrifuges at an average relative179centrifugal force of approximately 14,000. Small scale plasmid preparations were performed byalkaline lysis and standard phenoVchlorofom extraction ( Maniatis et al., 1982) or the TELT method asdescribed in He et al. ( 1989 ), involving lysis of bacteria in a 1:1 mixture of a) standardphenoVchlorofom and b) a solution of 2.5M lithium chloride, 50 mM Tris ( pH 8.0 ), 4% v/v Triton X-100,62.5 mM EDTA , prior to preciptation of the aqueous phase with ethanol. Large scale preparationswere performed by alkaline lysis followed by a variation of the cesium chloride method as described inManiatis et al. ( 1982 ), in which centrifugation was performed in 6 ml Sorvall ( Du Pont, Wilmington,DE) polyallomer tubes at 15 ° C at 60,000 rpm in a Sorvall TV865 rotor in a Sorvall RC50 ultra-centrifuge ( with an average relative centrifugal force of approximately 315,000 ). The iso-propanolprecipitate from a 1 litre culture was divided between two 6 ml tubes, centrifuged for 4 hours, theethidium-stained bands pooled, and the result centrifuged for 4 to 16 hours. Bands were extracted ina final volume of less than 4 ml, and ethidium was extracted 3 to 4 times with TE-saturated 1-butanol.The volume of the DNA solution was then brought up to 10 mls with the addition of TE, 200 RI of a 5Msolution of sodium chloride in double-distilled water was added, and 20 mls of ethanol added. Thissolution was then incubated for 3 to 16 hours in a -20° C freezer, prior to precipitation of the plasmidby centrifugation at 19,000 rpm ( average relative centrifugal force of 33,000) at 4° C in an SS34fixed-angle rotor in an RC5 centrifuge ( Sorvall ). Plasmid pellets were resuspended in TE, and furtherprecipitated with ethanol, and washed by centrifugation in 70% ethanol to reduce volume and partiallysterilise the plasmid suspension prior to electropration. Plasmids were stored at 4° C in TE.Agarose ( Gibco BRL ) gels for plasmid analysis were run in TBE for ethidium bromide visualisation, orTAE for Southern analysis, as described in Maniatis et al. ( 1982 ). These were made as 10 x stock (TAE) or 5 x stock ( TBE) and diluted prior to use. Gels were run in International Biotechnologies Inc.gel boxes, using Pharmacia power supplies. Ethidium bromide staining was performed by soaking thegel for 5 to 10 minutes in a solution of approximately 5 1.19/m1 ethidium bromide in running buffer( made by adding 20 III of a 10 mg/ml solution in water to approximately 40 ml of running buffer. Thisstaining solution was reused as often as practicable for analytical gels other than those subsequentlytransferred for Southern analysis, and ethidium was added to the solution periodically, as required.Bands were visualised using a short-wavelength UV transilluminator ( UVP Inc., San Gabriel CA ), andgels photographed using an MP4 Land Camera ( Polaroid, Cambridege, MA ), and Polaroid 667 filmcartridges. Reagents for these manipulations were from BDH ( generally AnalaR grade ) except foragarose and cesium chloride ( Gibco BRL ), and ethidium bromide ( Boehringer Mannheim ). DNA wasquantitated by fluorometry with a Hoefer DNA Mini-Fluorometer ( San Francisco, CA ), by comparingfluorescence at 460 nm with that of a supplied standard DNA preparation ( Hoefer ) in a solution of 0.1g.g/mIHoechst dye 33258 ( Aldrich ) in water.180Ligations were performed at 4 0 C over 12 to 16 hours in buffers supplied by the manufacturers of theligase ( New England Biolabs, Mississauga, ONT, or BRL ). Fragment ratios for ligations weredetermined by ethidium bromide visualisation of fragments electrophoresed in agarose, and molarligation ratios were generally of the order of 3:1 to 5:1 ( insert:vector ) for cohesive-end fragmentligations, and 5:1 to 10:1 for blunt-end ligations. Restriction digests were performed with enzymesfrom Pharmacia, BRL, or New England Biolabs, with supplied buffers, according to the manufacturersinstructions. Incubation for analytical digests were for 30 minutes to 1 hour in a 37° C waterbath. Alldigestions were stopped by the addition of sucrose-based ( BRL) or Ficoll-based ( Pharmacia )loading buffer prepared as described in Maniatis et al. ( 1982 ).Phosphatase treatment of vector - ends was performed with alkaline phophatase ( BoehringerMannheim ) using between 18 and 24 units per treatment ( 1 pi of the supplied preparation ).Restriction digests were first heat inactivated in a 70° C waterbath for 10 minutes, cooled to roomtemperature by brief incubation in ice, and centrifuged for a few seconds to precipitate condensation.To every 20 pl of digest was then added 1 pl of 10% w/v SDS, 3.5 gl of 1M Tris ( pH 9.0) and thephophatase. The mixuture was incubated in a 37° C waterbath for 30 minutes, and 1 p.I of 0.5M EDTAwas added, and the mixture then subjected to phenol/choloroform extraction. After ethanolprecipitation and 70% ethanol washing, the DNA was resuspended in an appropriate volume ofdouble-distilled water for gel quantitation or fill-in of overhangs with Klenow prior to ligation.Filling of restriction enzyme ends was accomplished by resuspending plasmid DNA in 15 gl of double-distilled water, adding 2 pi of New England Biolabs restriction buffer 2 ( 50 mM sodium chloride, 10 mMmagnesium chloride, 1 mM dithiothreitol, 10 mM Tris pH7.9 ), 2 pi of mixed deoxynucleotidetriphosphates ( 10 x stock prepared according to Sambrook et al., 1989) and 1 pl ( 5 units ) of Klenowfragment of DNA polymerase I ( New England Biolabs ). Incubation was allowed to proceed for 30minutes in a waterbath at 37° C, and the mixture was then incubated for 15 minutes in a 70° Cwaterbath to inactivate the Klenow, and an aliquot then taken for gel quantitation prior to ligation.B) Genomic DNA manipulations and Southern analysesFragments for labelling as probes were prepared by gel separation, visualised after ethidium stainingin fresh solution, using a 365 nm wavelength UV transilluminator ( Spectroline, Westbury, NY )andpurified using the Geneclean ( Bio 101 Inc., La Jolla, CA) or Sephaglas ( Pharmacia ) kits, according tothe manufactures instructions. 16 p.I aliquots containing 30 to 50 ng of fragment each were made indouble-distilled water in Eppendorf tubes, and stored at -20° C. Labelling was carried out by the181random priming method. An aliquot was thawed, the tube immersed in boiling water for 5 to 10minutes, and cooled to room temperature. To the aliquot was added 15 jal of \"oligo-mix\" ( see below ),2 of Klenow fragment of DNA polymerase I, and 3 to 4 of a-32P dCTP solution ( NEN Du Pont,Mississauga, ONT - 111 TBq/mmol, 370 MBq/ml, used before or no more than 14 days after thesupplied activity date ). After incubation at room temperature for 2 hours, 20121 of a 1% w/v solution ofbromophenol blue ( BDH ) and 170 ttl of TE was added to the reaction. The mixture was then passedby centrifugation through a \"spin column\" prepared by compacting by centrifugation a slurry ofSephadex G50 ( Pharmacia ) in TE into a 1 ml syringe over a small wad of glass-wool, to separateprobe from unincorporated nucleotides. The labelled probe was collected in a screw-top 1.5 ml tube( Fischer Scientific ) and to the contents were added 20 ill of 2M sodium hydroxide ( BDH ) and 760 p.1of SET ( 1% w/v SDS, 10 mM Tris, 5 mM EDTA, pH 7.5 ). The labelled probe was heated for 10minutes in a boiling water bath prior to addition to hybridisation reactions. All handling of radioactivematerials was done with a sheet of plexiglass between the experimenter and the material.\"Oligo-mix\" was prepared as follows: solution A consisted of 1 ml of oligo buffer ( 1.25M Tris pH 8.0,0.125M magnesium chloride - stored at 4° C) 18 III of 2-mercaptoethanol ( Sigma ) and 10 pi of eachof dATP, dGTP, and dTTP ( 100mM stocks, Pharmacia ). Solution B was 2 M Hepes pH 6.6 ( - TerryFox Laboratories ), and solution C consisted of random hexanucleotides ( Pharmacia, supplied as \"50A260 units\" ) reconstituted in 500 pi of sterile double-distilled water ( and stored at -20° C ). The mixwas prepared by combining A,B, and C in ratios of 1:2.5:1.5 ( generally 100 ILI : 250 pi : 150 ji,1 ).\"Oligo-mix\" was aliquoted in 50 Al lots in Eppendorf tubes and stored at -20° C.Mice were obtained from the UBC Animal Facility ( DBA/2 ) or from Harlan-Sprague-Dawley,Indianapolis, Indiana ( ICR for transgenic experiments ). Mice were anaesthetised in a mixture offluorothane in air, or Avertin and sacrificed in carbon dioxide. Avertin was made by mixing lOg oftribromoethyl alcohol ( Sigma ) with 10 ml of tertiary amyl alcohol ( Sigma ), and stored at 4° C wrappedin foil. Transgenic animals were produced by pronuclear injection of zygotes that were subsequentlyre-implanted into pseudo-pregnant females ( obtained by mating with vasectomised males ).Genomic DNA was extracted from various tissues other than thymus ( 1 to 3 cm of tail, and anequivalent amount of other tissues ) by first cutting the tissue to pieces of less than3 mm in anydimension, and then suspending tissue in 400 ill of extraction buffer ( 100mM Tris pH 8.0, 50 mMEDTA) in an Eppendorf tube. To this was added 401.11of a 1 mg/ml stock of Proteinase K ( BoehringerMannheim ) and the mixture then incubated overnight in a 55° C waterbath. The mixture was thensubjected to phenol/chloroform extraction and the DNA precipitated with ethanol by centrifugation atroom temperature for 10 minutes. After a further ethanol wash, DNA was resupended in standard TE,182allowed to go into solution and vigorously pipetted prior to quantitation by fluorometry. Thymus cellswere obtained by cutting whole thymus with scissors, and using a syringe plunger to gently force itthrough a wire mesh into 10 to 20 mls of RPMI or PBS, and centrifuging the cells into a smaller volume.Genomic DNA was then extracted in a similar manner to that described for other tissues. Genomic DNArestriction digests were peformed overnight in a 37° C bacterial incubator in up to 50 Ofinal volume.For dot-blots, 4 - 5 lig of undigested tail DNA was spotted in as small a volume as allowed adequatedissolution of DNA for quantitation ( generally 4 to 8 III) onto dry nitrocellulose ( Schleicher andSchuell BA 85, Keene, NH ) and the spots allowed to air-dry. The nitrocellulose was then soaked twicefor 1 minute each time, in denaturing solution, and then twice for 1 minute in neutralising solution( solutions as described in Maniatis et al., 1982 ). After allowing to air-dry, the DNA was cross-linked tothe nitrocellulose with a UV Stratalinker ( Stratagene, Cambridge, MA ). The nitrocellulose was thenplaced between sheets of clear plastic ( BelArt Products, Pequannock, NJ) which was sealed aroundit with an Audion Elektro heat-sealer ( Packing Aids Corp. San Francisco, CA ), so as to form a plasticbag, and sufficient hybridisation solution ( pre-warmed in a 55° C water bath, see below ) was added towet the entire surface of the nitrocellulose. The probe was then added and the bag sealed. The bagwas massaged to distribute the probe evenly through the liquid, and to force bubbles to accumulate ina corner so they could be heat-sealed away from the nitrocellulose. The bag was then incubatedovernight at 42° C between two plates of plexiglass.For Southern analyses, DNA was quantitated after digestion by fluorometry, and 5 gg was loaded perlane in loading buffer as previously described. Gels were run as for analyses, but in TAE, and comb-sizes used for pouring gels were such as to allow loading of 40 ill per lane. After running gels theywere stained with fresh solutions of ethidium bromide, and photographed on a low-wavelength UVbox, and allowed to sit over the UV source for 1 to 2 minutes to nick the DNA and improve theefficiency of transfer. The gel was twice immersed for 20 minutes each time in alkaline denaturationsolution and then twice for 20 minutes in neutralisation solution, in a plastic lunch box on a rockingplatform ( Bellco Biotechnology, Vineland. NJ ). Capillary transfer to nitrocellulose was then performedas described in Sambrook et al. ( 1989 ). The nitrocellulose was allowed to air-dry and transferred to adry oven at 80° C for 3 hours. The nitrocellulose was then bagged as described above for dot-blots,but pre-hybridisation solution ( see below ) added without probe. The bag was placed betweenplexiglass sheets and incubated at 42° C overnight. The following day, the prehybridisation solutionwas drained from the bag, and replaced with hybridisation solution, probe added, and the bag sealedas before for a further overnight incubation.183After probe hybridisations, nitrocellulose was removed form the bags and rinsed briefly inapproximately 50 mis of \"blot wash\" ( see below ) in a flat-bottomed Pyrex vessel. The nitrocellulosewas then washed twice, adding approximately 100mIs of fresh \"blot-wash\" each time to the vessel andallowing the contents to rise to 60° C by partially immersing the bowl in a 70° C water bath. Thenitrocellulose was removed from the vessel, allowed to air-dry on a piece of Whatman ( 3MM paper,and then wrapped in Saran Wrap ( Dowbrands Canada, Paris, ONT ). Adhesive paper was placed onthe Wrap, and this was then marked with a 35S-containing ink to assure correct orientation of the auto-radiograph. The ink was made by injecting approximately 10 pCi ( 370 kBq ) of 35S-methionine fromICN into the refill cartridge of a Schaefer fountain-pen. The blot was then exposed to XAR5 film in a anX-omatic cassette ( Eastmann-Kodak, Rochester, NY) for several hours to several days at -70° C. Theauto-radiograph was developed in a Kodak X-omat processor. Densitometry was performed with aComputer Densitometer from Molecular Dynamics Inc., in the Biotechnology Laboratories of theUniversity of British Columbia.Prehybridisation solution was made by mixing: 80 mis of SET ( see above ), 160 mis of deionisedformamide ( nucleic acid grade, Gibco BRL ), 100 mis of 20 x SSC, 40 mis of 50 x Denhardt's solution( both as described in Sambrook et al., 1989 ), 20 mis of 1M sodium phosphate buffer ( pH 6.0 ), 40mis of 10% w/v glycine, and 40 mis of a 1.5 mg/ml solution of herring sperm DNA. The mixture wasstored at 4° C. Hybridisation solution was made by mixing: 167 mis of deionised formamide, 100 mis of20 x SSC, 12.5 mis of 1M sodium phosphate buffer ( pH 6.0 ), 100 mis of 50% dextran sulphate, 34mis of 1.5 mg/ml herring sperm DNA, 40 mis of 50 x Denhardt's solution, and 6.25 mis of 10 x SET.This solution was also stored at 4° C. Blot wash, was stored at room temperature, and consisted of 0.1x SSC and 0.1% w/v SDS in double-distilled water. Ingredients for these reagents were obtained asfollows: Ficoll 400 - Pharmacia, polyvinylpyrrolidone - BDH, bovine serum albumin - ICN, deionisedformamide - Gibco BRL, sodium citrate - BDH, glycine - Gibco BRL, monobasic and dibasic sodiumphophate - BDH, dextran sulphate - Pharmacia, herring sperm DNA - Sigma.Molecular weight markers for Southerns were radio-labelled as follows. A 10 x stock of buffer wasmade by mixing 500 p.I of 1M tris ( pH 7.5 ), 150 Ill of 1M magnesium chloride, 20 gl of 0.5Mdithiothreitol, and 330 ill of double-distilled water. The buffer was stored at -20° C. For the labellingreaction, 2 pi of this buffer was added to 1 pl of the 1kb ladder ( BRL ), 3 pl of a solution of dATP, dGTPand dTTP ( 3mM with respect to each ), 7µI double-distilled water, 5 pi of a- 32P dCTP, and 2 p.I of Ecoli DNA Polymerase I ( New England Biolabs ). This mixture was incubated for between 1 and 3 hoursat room temperature, and free label separated from labelled fragments by spin column, as above.184C) Sequencing, PCR, and Northern blotsSequencing of the pLOX2 plasmid was performed with a commercial primer ( Pharmacia ) recognisingthe bacterial SP6 promoter. Sequencing was performed by the dideoxy-nucleotide chain terminationmethod ( Sanger et al., 1977) using a Sequenase kit ( United States Biochemical, Cleveland, Ohio ),according to the manufacturers instructions.For polymerase chain reactions ( PCRs ), a 10 x buffer was prepared, consisting of: 200 mM Tris ( pH8.0 ), 15 mM magnesium chloride, 500 mM potassium chloride ( BDH ), and 0.1% w/v gelatin ( Sigma ).The solution was stored at 4° C. A 20 x stock of deoxynucleotide triphosphates ( Pharmacia ) wasprepared consisting of a solution that was 5 mM with respect to each dNTP.Oligonucleotide primers were synthesised by Mr John Babcook at the Biomedical Resaearch Centrewith an Applied Biosystems PCRMate synthesiser ( Mississauga, ONT ), using reagents fromPharmacia and Applied Biosystems. Oligonucelotides were cleaved and purified as follows:ammonium hydoxide solution ( BDH, stored at 4° C) was drawn into the synthesis cartridge with a 1 mlsyringe, and allowed to soak the matrix for 20 minutes at room temperature. Then the solution wasexpelled into a screw-top 1.5 ml tube, and this procedure was repeated 2 more times. The tube wasthen capped, sealed with Parafilm ( American National Can, Greenwich, CT ), and incubated in a 55° Cwater bath for 8 to 15 hours. The cap was removed and replaced with Parafilm, into which a few holeswere punched with an 18 gauge needle. The contents of the tube were then dried in a SavantSpeedvac concentrator ( Savant Instrument Inc., Farmingdale, NY ). The contents were thenresuspended in 200 gl of TE, and passed through a spin-column ( as above without dye ). Oligos werethen quantitated by measuring their absorbance at 260 nm, and a fomula allowing the calculation ofthe expected absorbance of a 1M solution of the oligo according to base composition ( 16,000A +12,000G + 7,000C + 9,600T = absorbance of a 1M solution, where A, G,C, and T are the numbers ofeach of these nucleotides residues in the oligo ).PCR was then performed by first mixing 5 jil of 10 x buffer, with 5 ill of 20 x dNTP stock, 4 III of eachprimer solution ( containing approximately 1 p.mol of primer ), 52 pi of double-distilled water, and 20 illof solution containing the template DNA in a PCR tube ( Perkin Elmer Canada, Etobicoke, ONT ) andadding 50 pi of paraffin oil ( BDH ). The tube was then incubated at 94° C for 5 to 10 minutes in a DNAThermal Cycler ( Perkin Elmer Cetus, Norwalk, CT ). While the contents were warming, a solution wasmade containing 5 p1 of 10 x PCR buffer, 3 or 4 ill of double-distilled water, and 1p.I of Taq Polymerase( 5 units, Promega, Madison, WI ). 30 cycles of PCR were then performed as follows 94° C for 1185minute, 550 C for 1 minute and 30 seconds, 72° C for 2 minutes. Reaction products were then kept at40 C until 10 - 20 p1 aliquots were analysed by gel electrophoresis.Reverse-transcription PCR was performed using Moloney murine leukemia virus reverse transcriptasefrom New England Biolabs, with supplied buffer, according to the manufacturers instructions, in a finalvolume of 20111, in the thermal cycler. The entire reaction was then used as template for thesubsequent PCR step.Total cellular RNA was prepared from feshly homogenised tissues or thymus cell suspensions in 6 Mguanidium thiocyanate ( Pharmacia ) and fractionated through a cesium chloride ( Gibco BRL )cushion, as described in Chirgwin et al. ( 1979 ). RNA was denatured with formamide/formaledehyde,for agarose gel electrophoresis, blotting and hybridisation, as described in Marth et al. (1985 ).BIBLIOGRAPHYBradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quamtities ofprotein utilizing the principle of dye binding. Anal. Biochem. 72, pp. 420-.Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. & Rutter, W.J. ( 1979) Biochemistry 18, pp.5294-5299.Goding, J.W. ( 1986) Monoclonal antibodies: principles and practice. ( Second edition, AcademicPress, London ).He, M., Wilde, A. & Kaderbhai, M.A. ( 1989 ) A simple single-step procedure for small-scalepreparation of Eschericia soli plasmids. NAR 18, p1660.Laemmli, U.K. ( 1970) Cleavage of structural proteins during the assembly of bacteriophage T4.Nature 227, pp 680 -Maniatis. T., Fritsch., E.F. & Sambrook, J. ( 1982 ) Molecular Cloning A Laboratory Manual ( Cold SpringHarbour Laboroatory, New York ).Marth, J.D., Peet, R., Krebs, E.G. & Perlmutter, R.M. ( 1985) Cell 43, pp. 393-404Sambrook, J., Fritsch., E.F. & Maniatis. T. ( 1989 ) Molecular Cloning A Laboratory Manual ( Secondedition, Cold Spring Harbour Laboroatory, New York ).186Sanger, F., Nicklen, S. & Coulson, A.R. ( 1977) DNA sequencing with chain-terminating inhibitors.Proc. Natl. Acad. Sci. U.S.A. 74, 5463-5467Weissman, D., Parker, D.J., Rothstein, T.L. & Marshak-Rothstein, A. ( 1985 ) Methods for theproduction of xenogeneic monoclonal antibodies in murine ascites. J. lmmunol. 135, pp. 1001-1003."@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1993-11"@en ; edm:isShownAt "10.14288/1.0098814"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Genetics"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Two investigations in molecular pathophysiology"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/2247"@en .