Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

In vitro screening of crude extracts and pure metabolites obtained from marine invertebrates for the.. Stingl, John 1992

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

Item Metadata


ubc_1993_spring_stingl_john.pdf [ 4.62MB ]
JSON: 1.0086168.json
JSON-LD: 1.0086168+ld.json
RDF/XML (Pretty): 1.0086168.xml
RDF/JSON: 1.0086168+rdf.json
Turtle: 1.0086168+rdf-turtle.txt
N-Triples: 1.0086168+rdf-ntriples.txt
Original Record: 1.0086168 +original-record.json
Full Text

Full Text

We accept this thesis as conformingIN VITRO SCREENING OF CRUDE EXTRACTS AND PURE METABOLITESOBTAINED FROM MARINE INVERTEBRATES FOR THE TREATMENT OFBREAST CANCERbyJOHN STINGLB.Sc. (Biochemistry), University of British ColumbiaA THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THEREQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(DEPARTMENT OF ANATOMY)THE UNIVERSITYUNIVERSITY OF BRITISH COLUMBIANOVEMBER, 1992©John Stingl, 1992In 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 ^/1-401T044 yThe University of British ColumbiaVancouver, CanadaDate  14AI vyl-y I^'193DE-6 (2/88)iiTHESIS ABSTRACTFifteen samples of crude extracts and pure metabolitesobtained from marine invertebrates collected from theoffshore waters of British Columbia, Papua New Guinea andSri Lanka had previously been shown to exert cytotoxicactivity in the in vitro L1210 leukemic bioassay. For thisthesis, these samples were screened for in vitro cytotoxicactivity against the drug-sensitive breast tumor cell linesMCF-7, T-47D, ZR-75-1 and MDA-MB-231; the multidrug-resistant and P-glycoprotein (Pgp)-positive breast tumorcell lines MCF-7 Adr and MDA-A1r; and normal and malignanthuman breast epithelial cells (HBEC) in primary culture.Eight samples exhibited significant [drug concentrationresulting in a 50% decrease in cell growth as compared withcontrols (ED50), < 25 gg/ml] dose-dependent cytotoxicityagainst the drug-sensitive cell lines; the ED50 values wereas low as 0.004 gg/ml. Five of the 8 samples exhibitedsignificant cytotoxicity against the multidrug-resistantcell lines; the ED50 values were as low as 0.0006 pg/ml.Incubation of the MCF-7 Adr cells with varyingconcentrations of compounds in the presence of Adriamycindemonstrated that none of the compounds tested appeared tointerfere with Pgp function. Results obtained using HBEC inprimary culture showed a wide range of chemosensitivitiesfor a given drug against tissue taken from differentpatients, demonstrating the uniqueness of the response ofdifferent individuals to chemotherapy.iiiTABLE OF CONTENTSPageTHESIS ABSTRACT ^ iiLIST OF TABLES viiLIST OF ILLUSTRATIONS ^ viiiLIST OF ABBREVIATIONS xACKNOWLEDGMENTS ^ xiCHAPTER 1. INTRODUCTION ^ 11. Treatment of Cancer with Chemotherapy ^ 1la. Treatment of Breast Cancer withChemotherapy ^ 22. The Basis of Chemotherapy ^ 43. The Kinetics of Tumor Growth 53a. Skipper's Laws ^ 53b. The Goldie-Coldman Model ^ 73c. Gompertzian Growth and the Norton-SimonModel ^ 104. Drug Resistance 134a. P-Glycoprotein and Multidrug Resistance ^ 145. Historical Background of Anticancer DrugDevelopment ^ 175a. Preclinical Drug Discovery and Development ^ 175b. Clinical Evaluation of New Anticancer Drugs ^ 215c. Evaluation of the National Cancer Institute'sDrug Discovery and Development Program ^ 236. Current Status of the National Cancer Institute'sivDrug Discovery and Development Program ^ 277. Methods to Assess Chemosensitivity 328. Predictive Value of Chemosensitivity Assays ^ 379. Natural Products as a Source of New AnticancerDrugs ^ 4210. Thesis Objectives ^ 47CHAPTER 2. MATERIALS AND METHODS1. Isolation and Purification of Compounds ^ 492. Cell Line Culture ^ 503. Human Breast Epithelial Cells in PrimaryCulture ^ 534. Compound-Testing Protocol ^ 565. Preparation of Standard Curves for each Cell Line ^ 576. MTT Assay ^ 586a. Preparation of MTT ^ 586b. MTT Assay Protocol 586c. Selection of Solvent for the MTT Assay:i. With No Removal of MTT SupernatantPrior to Solvent Addition ^ 61ii. With Removal of MTT SupernatantPrior to Solvent Addition ^ 626d. Selection of Optimal MTT Concentration ^ 636e. Selection of Optimal MTT Incubation Time ^ 636f. Removal of MTT Supernatant ^ 646g. Absorbance Limit of the MTT Assay ^ 65CHAPTER 3. RESULTS ^ 661. Optimization of MTT Assay Parameters ^ 66vla. Selection of Solvent for the MTT Assay ^ 66lb. Selection of Optimal MTT Concentration ^ 70lc. Selection of Optimal MTT Incubation Time ^ 72ld. Removal of MTT Supernatant ^ 75le. Absorbance Limit of the MTT Assay ^ 752. Compounds Screened Against Drug-SensitiveCell Lines ^ 793. Compounds Screened Against Drug-ResistantCell Lines ^ 824. Compounds Screened Against MCF-7 Adr Cellsin the Presence of Adriamycin ^ 865. Compounds Screened Against Normal and MalignantHuman Breast Epithelial Cells in Primary Culture ^ 88CHAPTER 4. DISCUSSION ^ 89REFERENCES ^ 101APPENDIX 1: Growth Medium for the MCF-7, T-47D andZR-75-1 Cell Lines for Routine Passaging ^ 116APPENDIX 2: Growth Medium for the MDA-MB-231 CellLine for Routine Passaging ^ 116APPENDIX 3: Growth Medium for the MCF-7 Adr CellLine for Routine Passaging ^ 116APPENDIX 4: Growth Medium for the MDA-Alr Cell Linefor Routine Passaging ^ 116APPENDIX 5: Preparation of Saline-Trypsin-Versene ^ 116APPENDIX 6: Preparation of Calcium and Magnesium-FreePhosphate-Buffered Saline ^ 117viAPPENDIX 7: Growth Medium for the MCF-7, MCF-7 Adr,T-47D and ZR-75-1 Cell Lines forDrug Testing ^ 117APPENDIX 8: Growth Medium for the MDA-MB-231and MDA-Alr Cell Lines for Drug Testing ^ 117APPENDIX 9: Transport Medium ^ 117APPENDIX 10: Freezing Medium 118APPENDIX 11: Dissociation Medium ^ 118APPENDIX 12: Growth Medium for Human Breast EpithelialCells (Normal and Malignant) inPrimary Culture ^ 118APPENDIX 13: Preparation of Human Serum Samples ^ 118APPENDIX 14: Preparation of Rat Tail Collagen 118viiLIST OF TABLESTable 1: Curability of disseminated cancerwith treatment involving chemotherapy ^ 2Table 2: Known mechanisms of action ofchemotherapeutic agents ^ 5Table 3: Pathways to the discovery of currentanticancer drugs ^ 18Table 4: Historical evolution of in vivo screensat the NCI ^ 19Table 5: Sources and number of samples collectedby the NCI (1986-1991) ^ 45Table 6: Coefficient of variation associated withdifferent methods of MTT supernatantremoval ^ 75Table 7: ED50 values obtained for compounds screenedagainst the drug-sensitive breast cancercell lines MCF-7, T-47D, ZR-75-1and MDA-MB-231 ^ 81Table 8: ED50 values obtained for compounds screenedagainst the multidrug-resistant breastcancer cell lines MCF-7 Adr and MDA-Alr ^ 83Table 9: ED50 values obtained for the geodiamolidesand jaspamide tested against normal andmalignant breast epithelial cellsin primary culture ^ 88viiiLIST OF ILLUSTRATIONS Figure 1: Total survival of premenopausal patientswith advanced breast cancer after CMFchemotherapy ^ 4Figure 2: Skipper-Shabel curve for tumor growth ^ 6Figure 3: Gompertzian curve for tumor growth ^ 11Figure 4: NCI preclinical drug developmentprotocol (1975-1985) ^ 21Figure 5: NCI preclinical drug developmentprotocol (1985-present) 30Figure 6: Human tumor colony forming assay ^ 34Figure 7: NCI MTT assay protocol ^ 35Figure 8: NCI SRB assay protocol 37Figure 9: Molecular structure of Didemnin B ^ 44Figure 10: Molecular structure of palytoxin 45Figure 11: MTT assay protocol ^ 60Figure 12: Wavelength scans of cell generated formazandissolved in DMSO, isopropanol and ethanolin the presence of the MTT supernatant ^ 68AFigure 13: Wavelength scans of cell generated formazandissolved in DMSO, isopropanol andethanol ^ 69AFigure 14: Formazan production as a function of MTTincubation concentration ^ 71AFigure 15: Formazan production as a function of MTTincubation time ^ 73AFigure 16: Coefficient of variation as a function ofMTT incubation time 74AFigure 17: Absorbance limit of the MTT assay ^ 77AFigure 18: Wavelength scans of pure formazan andcell generated formazan dissolved inisopropanol ^ 78AixFigure 19: Structures of pure metabolites isolatedfrom marine organisms ^ 80AFigure 20: Standard absorbance curve for the MCF-7cell line ^ 85AFigure 21: Standard absorbance curve for the T-47Dcell line 85BFigure 22: Standard absorbance curve for the ZR-75-1cell line ^ 85CFigure 23: Standard absorbance curve for theMDA-MB-231 cell line ^ 85DFigure 24: Standard absorbance curve for the MCF-7Adrcell line ^ 85EFigure 25: Standard absorbance curve for the MDA-Alrcell line 85FFigure 26: Dose response curve of MCF-7 Adr cells tovarying concentrations of the geodiamolidesin the presence of the ED50 concentrationof Adriamycin ^ 87AFigure 27: Dose response curve of MCF-7 Adr cells tovarying concentrations of the Garveaextract in the presence of the ED50concentration of Adriamycin ^ 87BFigure 28: Dose response curve of MCF-7 Adr cells tovarying concentrations of the PNG-137Aextract in the presence of the ED50concentration of Adriamycin ^ 87CFigure 29: Dose response curve of MCF-7 Ad r cells tovarying concentrations of glaciasterolin the presence of the ED50 concentrationof Adriamycin ^ 87DFigure 30: Dose response curve of MCF-7 Ad r cells tovarying concentrations of jaspamide inthe presence of the ED50 concentrationof Adriamycin ^ 87ExLIST OF ABBREVIATIONSA540^Absorbance at 540 nmBSA^Bovine serum albuminDMSO^Dimethyl sulfoxideED50^Effective dose which results in a 50% decrease incell viabilityFBS^Fetal bovine serumHBEC^Human breast epithelial cellsHTCFA^Human tumor colony forming assayILS^Increase in life spanMDR^Multidrug resistanceMTT^3-[4,5-dimethyl(thialzol-2-yl)-3,5-depheryl]tetradium bromideNCI^National Cancer Institute (USA)PALA^N-(phohonacety1)-L-aspartatePgp^P-glycoproteinSRB^Sulforhodamine BxiACKNOWLEDGEMENTSThere are many people I wish to thank for thecompletion of this thesis. The first and foremost is mysupervisor Dr. Joanne Emerman, whose patience, guidance andsupport was invaluable during my studies. I look forward tocontinuing my PhD studies in her laboratory. I also thankDr. Raymond Andersen for giving me the opportunity tocollaborate with him in this field of research and forparticipating on my committe. I would also like to thank Dr.Wayne Vogl for taking the time to be on my committee.I would also like to thank Darcy Wilkinson for teachingme tissue culture methods and for her assistance inlaboratory procedures, Dr. Mark Elliot for his usefultechnical discussions regarding the MTT assay and Dr. NellyAuersperg for the use of her MacIntosh computer and ELIZAreader.Finally, I would like to thank my mother and mygirlfriend Janet for tolerating my seclusion during mythesis writing.This research was supported by grants to Dr. JoanneEmerman and Dr. Raymond Andersen from the National CancerInstitute of Canada.1CHAPTER 1INTRODUCTION1. Treatment of Cancer with ChemotherapyOver the last half century, approximately forty drugssuitable for the treatment of cancer have been developed(148). As a result, great strides have been made with theuse of chemotherapy in the treatment of cancer. The curerates of different types of disseminated malignancies withchemotherapy, alone or in combination with surgery and/orradiotherapy, is summarized in Table 1. Moderate to highcure rates are seen in the treatment of malignancies such astesticular carcinoma, Wilms tumor, childhood sarcomas andmany of the leukemias and lymphomas. Unfortunately manycommon solid tumors, especially the epithelial neoplasms ofadulthood still have a poor response rate to chemotherapy,thereby indicating that there is still a dire need for newand more effective chemotherapeutic agents and regimes (21,62, 73, 138).2Table 1. Curability of disseminated cancer with treatmentinvolving chemotherapy.DISEASE PROBABLE CURE RATEADULTSDiffuse histiocytic lymphoma (III & IV) 50%Hodgkin's disease (III & IV) 50%Testicular cancer (III) 75%Gestational choriocarcinoma 90%Acute myelocytic leukemia 20%CHILDRENAcute lymphocytic leukemia 50%Non Hodgkin's lymphoma 50%Burkitt lymphoma 50%Wilm's tumor 50%Childhood sarcomas 50%*Adapted from: Chabner B.A. (1990) Clinical strategies forcancer treatment: The role of drugs. In: CancerChemotherapy, Principles and Practice. (Chabner B. andCollins J., eds.), JB Lippincot Co, New York. pp Treatment of Breast Cancer with ChemotherapyBreast cancer afflicts one in 9 women in North America.It arises from the epithelial lining of the ducts andlobules of the glandular portion of the breast. Usually atthe time of clinical presentation, the cancer hasmetastasized to distant sites in the body, thereby makingtreatment of breast cancer a systemic problem. Currently thebest form of treatment of advanced breast cancer istreatment of the primary tumor with either surgery orradiation followed by a short course of high-dose adjuvantchemotherapy. This chemotherapy is given to kill anyresidual cancer cells. Metastatic breast cancer isresponsive to all major classes of chemotherapeutic agents,with response rates falling in the 20%-40% range (77, 155).3Response rates of approximately 55% are observed whencombination chemotherapy is used instead of single-agenttherapy (77). Currently the best chemotherapy regimes arethe cyclophosphamide-methotrexate-5-fluorouracil-vincristine-prednisone (CMFVP) variants (18, 27, 71, 77, 79,113) and Adriamycin (doxorubicin) based variants (ie:Adriamycin-cyclophosphamide and Adriamycin-cyclophosphamide-5-fluorouracil) (53, 77, 79, 119). The chances of survivalof the patient with metastatic breast cancer treated withadjuvant chemotherapy is best illustrated in a reportreleased in 1986 by the medical statistician Dr. R Peto(121, 155). This report analyzed the cumulative survivaldata of approximately 15,000 women who have participated inadjuvant chemotherapy trials over the years. It is verystatistically powerful because of the high number ofparticipants involved. The report states that adjuvantcombination chemotherapy in premenopausal patients withmetastatic breast cancer results in approximately a 30-35%decrease in risk of death as compared to patients notreceiving therapy. This 30-35% decrease in risk of deathtranslates into a benefit to approximately 10-20% of thepatients. For example, instead of 55% rate of mortalityafter 10 years, combination chemotherapy will result in a38% rate of mortality. Postmenopausal patients experiencedlittle significant benefit (12±5%) in risk of death (116).The long term survival of the first successful CMF trials isillustrated in Figure 1.4Figure 1. Total survival of premenopausal patients withadvanced breast cancer after CMF chemotherapy.(From: Bonadonna G. (1989) Conceptual and practical advancesin the management of breast cancer. J. Clin. Oncol. 7:1380-1397).2. The Basis of ChemotherapyChemotherapeutic drugs, for the most part, arecytostatic and/or cytotoxic agents. The exception to theseare those agents which act as biological response modifiers.Examples of biological response modifiers are theinterferons and the interleukins, which act to enhance theimmume system (172). Most of the cytostatic/cytotoxicanticancer drugs can be categorized as agents that damageDNA, antimetabolites, mitotic spindle poisons, hormones andantihormones. Table 2 lists the mechanisms of action of manyof the currently available anticancer drugs. The basis ofthe treatment of malignancy with cytostatic and cytotoxicdrugs is that tumor cells are growing and dividing fasterthan nonmalignant cells and anticancer drugs tend to be more5effective against rapidly growing and dividing cells thanquiescent cells (163). This accounts for the hair loss,myelosuppression and gastrointestinal problems associatedwith chemotherapy, since the epithelial cells of the hairfollicle, the bone marrow stem cells and the epitheliumlining the gut are all rapidly dividing cells. Unfortunatelynot all cancer cells are rapidly dividing (158); some cellshave slow rates of division and some cells are quiescent,thereby making erradication of the tumor difficult. Inaddition, not all tumor cells are sensitive tochemotherapeutic drugs. These topics are discussed furtherin the Kinetics of Tumor Growth and the Drug Resistancesections.Table 21 Known mechanisms of action of chemotherapeuticagents.1. Damage to the DNA template2. Interference with normal microtubule function3. Enzyme inhibition4. Enzyme stimulation5. Growth inhibition via a hormone receptor6. Modification of the immune system*Adapted from: Yarbro J.W. (1992) The scientific basis ofcancer chemotherapy. In: The Chemotherapy Source Book (PerryM.C., ed). Williams and Wilkins, Baltimore. pp 3.3. The Kinetics of Tumor Growth 3a. Skipper's Laws In 1964, Skipper, Schabel and Wilcox using the in vivoL1210 mouse leukemia model made two historical observations(150). The first was that this tumor's growth rate isi1^I^I^1^I^I 0 2 ♦ G   10 2 14 IS6logarithmic. This means that the doubling time of apopulation of cells from this tumor is constant, regardlessof the number of cells in that population. For example, ifit takes 2 cells 24 hours to divide and form 4 cells, thenit will take 24 hours for 2000 cells to divide and form 4000cells. This logarithmic growth implies that all the cells inthe population are actively growing and dividing. A plot oftumor cell number vs time of this type of growth will yielda straight line on a semilog plot (Figure 2).OATSFigure 2. Skipper-Schabel curve for tumor growth.From: Skipper H.E., Schabel F.M., Jr. and Wilcox W.S. (1964)Experimental evaluation of potentail anticancer agents.XIII. On the criteria and kinetics associated with"curability" of experimental leukemia. Cancer Chemother.Rep. 35: 1-111.The second observation Skipper and colleagues made wasthat treatment of this L1210 tumor with a cytotoxic agentresulted in a specific percentage of the tumor cells being7killed (depending on dose of drug administered), regardlessof the original tumor cell number. For example, if the drugkilled 90% of the cells in an original population of 1000cells, after the first treatment 100 cells would survive.After the second treatment with the same dose, 10 cellswould survive, and after the third treatment 1 cell wouldsurvive. Furth and Kahn in 1937 demonstrated that leukemiacan be transmitted from one mouse to another with a singlecell (56). Therefore the survival of a single malignant cellafter treatment can result in relapse. In the examplepresented earlier, 4 treatments with our chemotherapeuticagent would be required to erradicate the tumor (ie: to getthe surviving cell number < 1, assuming that all cells aredrug sensitive).3b. The Goldie-Coidman Model One of the greatest obstacles in the treatment ofmalignancy is the development of drug resistance. Themechanisms of drug resistance are varied and are discussedlater under the Drug Resistance section. In 1952, Lawdemonstrated that resistance to a chemotherapeutic agent isa trait that can be acquired randomly through geneticmutation prior to exposure to that agent (88). Amathematical model relating drug sensitivity of tumors tothis random acquisition of resistance has been constructedby Goldie and Coidman (58). According to their model, theproportion and the absolute number of drug-resistant cells8within a tumor will increase with time. The fraction of thetumor that is composed of drug-resistant cells will varybetween tumors, depending on how early or late the mutationevent occured. Goldie and Coldman have calculated that theprobability (P) of a tumor consisting of N cells not havinga mutation that results in resistance to a given drug is:P = exp {-x(N-1)}where x = spontaneous mutation rateStudies have indicated that the probability of agenetic mutation occuring in a nonmalignant cell is in theorder of magnitude of 10-6 (1, 32, 41, 143), whereas inmalignant cells the probability is increased to 10-5 (31,41, 143). The minimum size of early detection of a tumor isconsidered to be approximately 1 gram of tumor whichcorresponds to roughly 109 cells (174). Thus the probabilityof a tumor at the time of detection not having any cellsresistant to a given drug is:P = exp {-10-5 (109 - 1)} _ 0The Goldie-Coldman model is based on the followingassumptions:1. The growth of the tumor is based on Skipper'smodel of exponential growth.2. Drug-resistant cells are resistant to allconcentrations of drug.3. Treatment is based on single-agent chemotherapy.Although the Goldie-Coldman model is a rather9simplistic representation of the evolution of drug-resistance within a tumor, one can still conclude withreasonable certainty that the appearance of drug-resistantcells is a major reason for treatment failure (58, 151).Several recommendations for improved chemotherapy can be putforth as a result of the Goldie-Coldman hypothesis:1. The earlier chemotherapy is initiated, the increasedprobability of a cure being obtained.2. Combination chemotherapy would be more effective inobtaining a cure than single-agent chemotherapysince the probability of a cell mutating to becomeresistant to several drugs is much less than itbecoming resistant to a single agent.The treatment of many different patients representing avariety of neoplasms has demonstrated that Skipper's modelof exponential growth and cell kill is not a true reflectionof tumor growth. Patients who have been treated withcombination chemotherapy to induce remission are responsiveto the same chemotherapeutic regime following relapse (78,162). The achievement of a second remission is opposite ofwhat one would expect if adhering to Skipper's model ofexponential growth and the Goldie-Coldman hypothesis. Itwould be expected that a second round of chemotherapy wouldresult in very little or no response since most or all thedrug-sensitive cells should have been erradicated by thefirst drug treatment. Assuming that the drug-sensitive anddrug-resistant cells have equal cycling times, the ratio of10drug-sensitive cells to drug-resistant cells should be lessthan one (152). Consequently no response should be observedsince more drug-resistant cells are being added to thepopulation than drug-sensitive cells being killed. Thisobservation of a second remission indicates that drug-resistant cells are reverting back to a drug-sensitivephenotype after treatment and/or a substantial number ofdrug-sensitive cells are not being killed by the firsttreatment of chemotherapy. There is experimental evidencethat the former can occur (36, 80), but a kineticexplanation, the Norton-Simon model, has been put forward toaccount for the treatment failure (110, 111).3c. Gompertzian Growth and the Norton-Simon Model The growth of experimental animal tumors (87) andclinically observed tumors (68, 109) when plotted (cellnumber vs time) yields a Gompertzian curve (Figure 3). Thisshape of growth curve indicates that this type of tumorgrowth is much more complex than that of Skipper's in vivoL1210 tumor model. In order to explain this observedGompertzian growth, the concept of "growth fraction" must beintroduced. Mendelsohn in 1960 suggested the notion thattumors are mixtures of proliferating and nonproliferatingcells (97). He defined the growth fraction of a tumor as theratio of proliferating cells to total cells. In 1968 Tannockdemonstrated that vascularity of a tumor influenced thegrowth fraction (158). Using a unique mouse mammary tumor11model in which the tumor cells grew uniformly aroundstraight blood vessels, he observed that tumor cells in theimmediate vicinity of the oxygen and nutrient supply had thefastest growth rates and the highest growth fraction,whereas those cells further away from the blood supply hadslower rates of division and lower growth fractions. Thecells farthest from the blood vessels were observed to benecrotic.Figure 3. Gompertzian Curve for Tumor Growth.The initial portion of the Gompertzian curve representsa small number of rapidly dividing tumor cells (171). Thisportion of the Gompertzian curve resembles Skipper's modelof exponential growth. Since cell number is so low, theoverall slope of the curve is low, despite the fact that thegrowth fraction of the tumor is high. As the tumor expandsand begins to outgrow it's blood supply, the growth fractionof the neoplasm decreases (158, 111). At the other end of12the curve, where the tumor becomes lethal to the patient,tumor cell number is maximal and growth fraction is minimal,thereby resulting in a low slope of the curve. Maximal tumorgrowth (slope) is observed in between these two extremeswhere neither cell number nor growth fraction is maximal,but when their product is. The Norton-Simon model forGompertzian tumor growth can be represented by the followingequation (111):dN(t)/dt = GF(N) x N(t) - K x L(t) x GF(N) x N(t)where N = cell numbert = timeGF = growth fractionK = constantL = level of chemotherapyor more simply stated:change in tumor cell number = cell generation - cell death.Cell death is dependent on the growth fraction of thetumor, since chemotherapeutic agents are more effectiveagainst mitotic cells than quiescent cells (163). The mainimplications of this equation is that the rate of tumor cellkill and the rate of regrowth increases as tumor sizedecreases and thus the level of therapy required to initiatea regression may not be adequate to maintain the regressionand to produce a cure. A greater level of cell kill iscounterbalanced by a greater fractional regrowth, so that13the only way a cure can be achieved is to use more intensivechemotherapy against very small tumors. To be curative, suchintense chemotherapy will require the patient to haveautologous bone marrow transplants and hematopoietic growthfactor support to counteract the myelosuppresive action ofthe intense chemotherapy. In summary, it now appears thatchemotherapy treatment failure is not entirely due to thedevelopment of drug resistance, but is also due in part tothe kinetic paremeters of Gompertzian tumor growth and thelimitation in dosage of chemotherapy due to host toxicity.4. Drug ResistanceOne of the reasons for chemotherapy treatment failureis the acquisition of drug resistance by the tumor cells.This acquisition of drug resistance can occur randomly priorto treatment (88) or it can evolve during the course ofchemotherapy (15, 90). The mechanisms by which cancer cellscan become refractory to chemotherapy are diverse (35, 173)and some of the more important methods are summarized asfollows:1. Increased expression of the plasma membrane proteinP-glycoprotein (Pgp).2. Changes in expression of drug detoxifying enzymes(glutathione-S-transferase, dihydrofolate reductase)3. Qualitative and/or quantitative changes in DNA repairenzymes (topoisomerase I and II, polymerase alpha and beta,thymidylate synthase).144. Genomic alterations.4a. P-Glycoprotein and Multidrug Resistance One of the factors responsible for unsuccessful cancerchemotherapy is the acquisition of multidrug resistance(MDR). Multidrug resistance is a phenotype characterized byresistance to a drug which arises upon exposure of the cellto that specific agent (15, 90). Multidrug-resistant cellsalso exhibit a cross resistance to a number of compoundsthat are structurally and functionally unrelated to the drugthat induced the mdr phenotype (6, 10, 92). Multidrugresistance has been observed to be associated with increasedlevels of a 170 kDa plasma membrane protein, P-glycoprotein(Pgp)(15, 164). Significant levels of Pgp have beenidentified in a wide variety of malignancies, includingsarcomas, carcinomas, lymphomas and leukemias (59, 76).Structural analysis of Pgp predicts that it is an ATP-binding pore forming pump, and that it brings about drugresistance by actively pumping drugs out of the cell (28,63, 76). The gene coding Pgp has been termed the mdrl gene.Pgp has been detected at low levels in a variety of normaltissues and is believed to have a role in normal physiology(54, 69). Amino acid sequencing confirms that Pgp sharesextensive sequence homology with numerous bacterial andeukaryotic transport proteins (22, 76). This conservation ofstructure of Pgp throughout evolution indicates that thisprotein is a fundamental membrane transport protein. Pgplevels are regulated by gene copy number (81, 164), by rates15of transcription (15) and by post transcriptionalmodifications (15).From a clinical point of view, the two most importantquestions are 1) does Pgp predict response to chemotherapyand 2) can the mdr phenotype be circumvented? The answer tothe first question at this point in time in not entirelyclear, since there is little data relating Pgp expressionand patient response to chemotherapy (89). Two preliminarystudies involving breast cancer have observed a correlationbetween Pgp expression (as determined by immunohistologicalmethods) and response to chemotherapy (131, 166). Aconfounding problem associated with using Pgp to predict aresponse to chemotherapy is that many normal tissues expressPgp and it may be difficult to discriminate between normaland neoplastic tissue (30). Some in vitro studies usingcancer cell lines (47, 140, 164) and malignant tissue inprimary culture (5, 59, 81, 144) have correlated increasedexpression of Pgp with decreased sensitivity tochemotherapy, but other in vitro and in vivo experimentsutilizing both cell lines and fresh tissue could notestablish a strong correlation between Pgp expression andchemosensitivity (59, 65, 126). Although Pgp expression maybe high, pump function may be zero.In vitro experiments have established that inhibitionof the energy-dependent Pgp-mediated drug efflux can resultin circumvention of mdr. Calcium antagonists such asverapamil, diltiazem, nifedipine and others have enhanced16the cytotoxicity of the Vinca alkaloids (vincristine andvinblastine) and the anthracyclines (Adriamycin) againstcell lines representing a variety of tumor types (98, 101,159, 176). This enhanced cytotoxicity is a result ofcompetition of binding to Pgp between the calciumantagonists and the antitumor agents (137, 176). The calciumchannel blockers act to circumvent the mdr phenotype in acalcium-independent manner (19). A problem encountered inthe use of calcium antagonists in circumvention of mdr isthat the maximal clinically achievable dose of theantagonist is limited because calcium antagonists causedetrimental cardiovascular side effects in vivo (8). As aresult, the maximal in vivo dose is below that required toreverse the mdr phenotype (98).Calcium influx blockers are not the only agents thatcompete with binding of antitumor agents to increaseintracellular drug concentration. Calmodulin inhibitors (85,160), monoclonal antibodies (7, 16), dihydropyridines (175),several of the steroid hormones, especially progesterone(70, 108), and the antiestrogen tamoxifen (50) have all beenshown to interfere with Pgp-mediated efflux of anticancerdrugs. With respect to mdr and breast cancer, the use of theprogestins and tamoxifen may prove to be interesting sinceboth are relatively nontoxic (75, 77) and both are used fortreatment of breast cancer (24, 125).175. Historical Background of Anticancer Drug Development The first nonhormonal antitumor agents to demonstrateactivity in humans were the nitrogen mustards (60, 72, 128).These alkylating agents, which were used in the chemicalwarfare programs of World War I, demonstrated significantantitumor activity in patients with lymphomas. Theseobservations signalled the beginning of the search for newanticancer agents. The pursuit for anticancer drugs prior tothe 1950's was performed mainly by the pharmaceuticalindustry (38). Sixteen of our current armament ofapproximately forty drugs were developed during this era. In1955, the Developmental Therapeutics Program at the NationalCancer Institute (NCI), based in Bethesda Md, initiated itsdrug discovery and development program (38, 62). Thisprogram has been the main driving force behind the searchfor new anticancer drugs since its inception. Of the sixteencommercially available drugs introduced between 1955 and1985, the NCI discovered eight of them. In addition, the NCIwas involved in the clinical evaluation of all sixteen ofthese drugs (38).5a. Preclinical Drug Discovery and DevelopmentThere were two methods by which the NCI approached newdrug discovery: rational drug design and mass screening ofsynthetic and natural products. Anticancer drugs such asasparaginase, cisplatin, mitotane, levamisole andantimetabolites such as fluorouracil and N-(phohonacetyl)-L-18aspartate (PALA) were discovered both by logical reasoningand serendipity (25, 38, 48, 66, 84, 132, 148). Althoughrational drug design had an important role in the discoveryof anticancer drugs, it was more commonly used in improvingleads discovered via mass screening. Biological screening ofcompounds so far has accounted for approximately half of thecurrent repertoire of chemotherapeutic agents (Table 3).Screening of acquired compounds was performed using a seriesof in vivo tumor models. These models measured the increasein life span (ILS) of tumor-bearing mice receiving treatmentwith a potential anticancer drug vs tumor-bearing mice notreceiving treatment. An ILS of at least 25% was required forthe agent to be considered "active" so it can be forwardedto the next stage of testing (62). The evolution of these invivo tumor models from 1955 to 1982 is presented in Table 4.Table 1. Pathways to the discovery of current anticancerdrugs.Targeted synthesisScreening ofnatural productsScreening ofchemicalsAnaloguesynthesisSerendipity andrational applicatiotAntimetabolites: Dactinomycin Busulfan Cyclophospharnide MechlorettuunineMethotrexate Vineristine Dacarbazine Chlorarnbucil AsparaginaseThioguanine Vinblastine Procarbaz int Melphalan MitotaneMercapiopurine Plicarnycin Hydroxyurea Ifosfamode CisplatinFlutwouracil Daunoruhicin T7note pa Etoposide LevarnisoleCytarabine Doxorubie in Carmustme Ten iposide lmerferonsPALA Mitomycin C Lomustine CarboplatinHormones: Bleomycin MitoxantroneSteroids S treptozocin Al tretamineTarnox i fen rand Pentostat inFlutam WeLeuprolideOctreotide*From: Sikic B.I. (1991) Anticancer Drug Discovery. J. Natl.Cancer Inst. (U.S.A.) 83: 738-742.19Tab1 4. Historical Evolution of In Vivo Screens at theNCI.YEAR^IN VIVO SCREEN1955^sarcoma 180**, carcinoma 755,1eukemia L12101960 L1210 + 2 models from a pool of 211965 L1210 + Walker carcinoma 2561968^L1210 for synthetics, L1210 and P388 leukemiafor natural products1972 L1210 for synthetics, P388 for naturalproducts, B16 melanoma and Lewis lungcarcinoma for special testing1975^P388 prescreen, followed by a panel of 8models, including human xenografts*From: Staquet M.J., Byar D.P., Green S.B. and Rozencweig M.(1983) Clinical predictivity of transplantable tumor systemsin the selection of new drugs for solid tumors: Rationalefor a three-stage strategy. Cancer Treat. Rep. 67: 753-765.for tumor models are murine except where noted.The NCI's preclinical drug development protocol for theyears 1975 to 1985 is illustrated in Figure 4. The firststep (Stage I) of this protocol was a P-388 murine leukemiaprescreen to eliminate the supposedly large number ofclinically inactive compounds. Any agent considered "active"(ILS > 25%) in this prescreen was then forwarded to the invivo tumor panel (Stage II) which consisted of a variety ofhuman tumor xenographs and murine malignancies. If the agentdemonstrated any activity (ILS > 50%) in Stage II, it wasfurther investigated (formulation/toxicology stage) todetermine structure, production, mechanism of action,dosage, toxic effects on animals and so on. The animaltoxicology studies, which were usually performed on two20species of animals, were performed to characterize thedrug's absorption (if oral), clearance, half-life,metabolism, excretion and either or both the maximumtolerated dose (MTD: dose which produces severe reversibletoxicity) and the LD10 (dose which is lethal to 10% of thepopulation given that dose). These studies helped todetermine if the agent was suitable for clinical trials andwhich dose to administer in clinical trials. The initialdose is usually a fraction of either the MTD or the LD10(26).21designer drugs^unknown compoundsP388 murine leukemia (in vivo)^STAGE IMURINE TUMORSILS > 25%HUMAN TUMOR XENOGRAFTL1210 leukemia^LX-1B-16 melanoma CX-1lungcolonMID 70's Lewis lung^MX-1 mammaryC-38 colonCD8F1 breastSTAGE IIL1210^MX - 1MID 80's M5076 sarcomaB-16ILS > 50%formulation and toxicologydisease oriented clinical trials PHASE IFigure 4. NCI preclinical drug development strategy (1975-1985).5b. Clinical Evaluation of New Anticancer Drugs Once the preclinical drug research phase is completed,the compound is introduced into clinical trials. Clinicaltrials are composed of three phases (1-111)(26, 94). Thegoal of Phase I is to assess the drug's qualitative andquantitative (maximum tolerated dose) toxicity upon thepatient. The patients involved in Phase I are those who have22tumor types in which there is no suitable treatmentavailable. In addition, patients are selected so that abroad spectrum of tumor types are represented by the patientpanel. The assumption in cancer chemotherapy is thatanticancer drugs are most effective when administered nearto or at their MTD (55). For most clinically active drugsthe bone marrow is usually the dose limiting factor, but theCNS, kidney, liver and other organs may also be doselimiting (94). The maximum tolerated dose for patients witha common tumor type is determined by using a "modifiedFibonacci scheme" (26). The initial dose is extrapolatedfrom the animal toxicology studies and is administered tothree patients. If acceptable toxicity is observed withthese patients, the dose is increased for the next threepatients who enter the study. As long as the toxicityremains acceptable, the dosage will be escalated withdecreasing increments for new patients entering the study.The main goal of Phase II studies is to establish whichtumor types the drug is effective in treating and to furtherdefine the drug's side effects. The patients in this phasecomprise the Human Tumor Panel. This panel was created in1975 and consists of patients with the same types ofmalignancies (lung, colon, breast, lymphoma, leukemia andmelanoma) as those in the preclinical in vivo tumor panel.The purpose of selecting these two panels is to determine ifresults obtained from the in vivo preclinical tumor modelscorrelate with those obtained using the Human Tumor Panel.23The goals of Phase III are to determine the responserate, response duration, survival, toxicity and relativeactivity (when compared to standard treatment) of the drugin patients with tumor types that were responsive in PhaseII. In addition, Phase III studies explore the incorporationof the agent with other drug combinations, alternate routesof administration and high dose therapy in combination withrescue procedures.5c. Evaluation of the National Cancer Institute's DrugDiscovery and Development ProgramIn 1984 it became evident that the NCI's criteria foridentifying new chemotherapeutic drugs was unsatisfactory.This situation was illustrated in a report released in 1987which evaluated the Phase II activity of all cytotoxic drugsintroduced into clinical trials by the NCI between the years1970 and 1985 (95). During this time period, 83 drugs wereintroduced into Phase I trials. Of these 83, 11 wereterminated at the end of Phase I and 5 remained in Phase Iat the time of the report. Of the 67 drugs that proceededonto Phase II, 13 were still in Phase II and 7 were notevaluated properly. After Phase II testing, 24 of theremaining 47 agents demonstrated significant activityagainst at least one type of malignancy. Only 14 of 24active agents were new chemical structures; the remaining 10were analogues of already known chemotherapeutic drugs. Withthe NCI drug discovery program screening up to 40,00024compounds annually with the cost in the tens of millions ofdollars per year ($29 M in 1984)(21), only 14 novelcompounds discovered after 15 years was highlyunsatisfactory. Of a more serious concern was that of theactive agents, 74% were rated active against lymphoma, 35%against leukemia and only marginal activity was observedagainst the solid tumors. It became apparent that the P-388murine leukemia prescreen was not a valid representation ofall types of malignancies. It is believed that the P-388prescreen was filtering out agents that are potentiallyeffective in treating solid tumors. The P-388 model wasassumed to be a valid representation of all tumor typesbecause of the P-388's and the L-1210's (P-388'spredecessor) historical predominance in anticancer drugscreening and of the fact that more agents discovered by theNCI program were effective against the in vivo L-1210 and P-388 leukemic bioassays than any other in vivo bioassay (156,165), which of course is faulty logic. When the NCI's drugscreening program was initiated, the L1210 and P-388 in vivotumor models seemed to be the appropriate choices because oftheir technical feasibility. The NCI at the time wasscreening 7,000-12,000 compounds per year. At this level ofscreening, the use of a single in vivo preliminary screenrequired that the tumor model used have a doubling time of10-15 hours so that the antitumor assay could be completedin 30 days (157). The NCI eventually dropped the colon andlung human tumor xenografts from their Stage II screen25because they were deemed too insensitive for drug screening(62, 114). This is somewhat ironic since lung and especiallycolon neoplasms are known to be clinically resistant tochemotherapy (59, 106). In fact, it would seem that the mostimportant tumor models would be those that are nonresponsiveto drugs currently in use. This is supported by theobservation that out of 979 compounds rated not active in anin vivo L1210 leukemic prescreen, 119 were later found tohave activity in at least one of the in vivo solid tumorsfrom the Stage II secondary screen (165).Another flaw in the NCI's use of murine tumor modelswas that for the P-388 Stage I screen and many of the murinetumor models in the Stage II screen, both the tumor and thedrug were injected intraperitoneally (62). This form oftumor model, nicknamed "test tube in the mouse", does notrepresent the true in vivo environment since nophysiological barriers have to be passed by the drug. As aresult, drug activity is falsely magnified and the assaytends to isolate drugs that do not cross physiologicalbarriers well (62).In the mid 1980's the NCI reevaluated its drugdiscovery program. Several alternatives to the drug-screening system were proposed (21):1. Abandonment of all drug discovery contracted out by theNCI. Approximately 50% of all compounds acquired for the NCIdrug-screening program have been via a chemical synthesis26contractor and a collection contractor (38). Instead the NCIwould redirect these funds towards drug discovery grants andclinical development of compounds submitted from outside theNCI. Active compound acquisition by the NCI and voluntarysubmission of agents to the NCI, which normally account forapproximately 20% and 30% respectively of all compoundsaquired, would then become the major sources of agentsentering the greatly scaled down NCI drug-screening program.This proposal was rejected because it was felt that thiswould be too much of a negative signal to the few remainingprivate sector companies involved in drug-discoveryprograms.2. Development of new molecular screening targets. With thescience communities' ever-expanding knowledge of tumorbiology and scientific technique, we are identifying moreand more intracellular targets involved in induction andmaintenance of neoplasia. These intracellular targets can beutilized in a screening protocol to identify agents that mayinhibit the neoplastic process. Some examples of thesetargets are protein kinases, growth-factor inhibitors,oncogene products and proteins involved in multidrugresistance. The NCI rejected the development of this type ofscreen for two reasons. The first is that this type ofscreen utilizes a cell free system and thus is too farremoved from the in vivo environment. Secondly, theimportance of any one specific intracellular target involved27in the malignant process is still not clear (21). Manyindependent laboratories have adopted this screening processand, as our understanding of molecular biology increases,this format of drug discovery will become more prominent(73, 154). The NCI recognizing this, has established theNational Cooperative Drug Discovery Groups, an organizationthat brings together molecular biologists, chemists andbiochemists so as to expand the role of rational drugdiscovery (154, 157, 165).3. Creation of an in vitro human tumor cell line panel forinitial testing of compounds. To avoid the "filtering"effect observed with the use of the P-388 leukemiaprescreen, the NCI decided to utilize an initial screeningpanel representing all the major categories of tumor types(14, 21, 103). An in vitro human tumor cell line panel wasselected instead of an in vivo panel because it iseconomically and technically impractical to use in vivomodels for mass screening.6. Current Status of the National Cancer Institute's Preclinical Drug Discovery and Development ProgramFigure 5 is a schematic representation of thepreclinical antitumor drug screening protocol implemented atthe NCI in 1985 (14, 62). In order to avoid the disease-specific Stage I screen and the "filtering" effect observedwith the screening protocol used in earlier prescreens, the28NCI switched to a disease-oriented in vitro human tumor cellline panel. The ultimate goal is to have the human tumorcell line panel consisting of 120 cell lines representing10-12 different tumor types (14, 103). Currently the pilot-scale cell line panel utilizes 60 human tumor cell linesorganized into subpanels representing melanoma, leukemia andcancers of the lung, colon, kidney, ovary and centralnervous system (103). Each tumor type is represented bymultiple cell lines because the use of subpanels will allowfor detection of agents that are specific for certaincategories of malignancies (ie: histiospecificity) or,conversely, agents that are active but not tumor-specific(14, 103, 148). In addition, the use of just 1 or 2 celllines originating from 1 type of cancer would not emulatethe heterogeneity observed in naturally occuring tumors (51,122). Cell lines are used instead of fresh tumor tissue inprimary culture because the cells are plentiful, easy togrow and well characterized (ie: for mechanisms resulting indrug resistance). Previous attempts at using fresh tumortissue from a variety of solid tumors (in conjunction withthe human tumor colony-forming assay) was shown to beimpractical for high volume screening (145, 146). The use ofcell lines in the primary screen is also advantageous,because many human tumor cell lines can be propagatedsubcutaneously in athymic nude mice (14, 114), therebyproviding a means of evaluating agents in vivo (Stage II offigure 5). This Stage II screening involves the testing of29agents in athymic nude mice with subcutaneous implants ofthe most chemosensitive cell lines from Stage I. In additionto the use of these subcutaneous human tumor xenografts, theNCI can further investigate promising compounds morethoroughly in vivo by using an orthotopic tumor model andthe microencapsulated tumor assay (META)(14, 61, 157). Inthe orthotopic tumor model, the tumor cells are implanted invivo in sites corresponding to the presumed origin of themalignancy rather than subcutaneously. For example, if anagent was observed to be active against renal carcinoma, thetumor cells would be implanted in the kidney. In the META,tumor cells are encapsulated in microcapsules, which in turnare implanted in the peritoneal cavity of nude mice. Thecapsules are designed to confine the tumor cells while stillexposing them to an in vivo environment. Following drugexposure, the capsules are removed and the surviving cellnumber counted and compared to controls. Neither the METAnor the orthotopic xenograft methods are to be used forfirst line in vivo screening since they are too technicallyimpractical (14). Instead, these tumor models are to be usedfor more specialized preclinical investigation of promisinganticancer compounds.30designer drugs^unknown compoundsdisease oriented, in vitro primary screen^STAGE Iusing the human tumor cell line panelsignificant activityin vivo testing in selected^STAGE IIhuman tumor subcutaneous xenograftsformulation and toxicologydisease specific clinical trials^PHASE IFigure 5. NCI preclinical drug development protocol (1985-present).A drawback of using murine tumor models forchemosensitiviy testing is that clinically achievable dosesare 2- to 10- fold higher than that achievable in mice(148). In humans, we have ways we can support themyeloproliferative cells in the bone marrow (ie: growthfactor and bone marrow transfusions and gene therapy) andthus we can increase the dose intensity (57, 82). But inmice, myelosuppression is the dose-limiting factor,therefore agents that are myelosuppresive in mice but not inhumans could potentially be filtered out. A possible way toget around this problem is to use a transgenic mouse thathas its bone marrow cells expressing the mdrl gene (102).31Inclusion of non-neoplastic myeloid colony formingcells as a control in the human tumor panel may give anindication of the therapeutic index (minimum toxicdose/minimum effective dose) of the test agent (21).Once preclinical drug development is completed, theagent under study is introduced into disease-specificclinical trials. This strategy is based on the assumptionthat in vitro activity against specific tumor types willtranslate to in vivo activty in patients who have thosetumor types. There is some evidence that histiospecificactivity does exist for certain antitumor agents. Mitotane,asparaginase and ipomeanol demonstrate preferential activityfor adrenocortical carcinomas, lymphoblastic leukemias andlung cancers respectively (23, 148). Several earlier studiesinvestigating the feasibility of alternative drug screeningmethodologies have demonstrated that agents previously shownnot to be active in a leukemic prescreen demonstratedactivity in non-leukemic screens. An NCI pilot screeninvestigating the feasibility of the (in vitro) human tumorcolony forming assay for large-scale drug screening detectedactivity for 14 of 79 compounds that were negative in theP388 leukemic prescreen (145). Of 1085 compounds testedagainst a panel of in vivo murine tumors representing avariety of murine and human tumor types, 106 were active inL1210 leukemia. Of the 979 compounds that were inactive inthe L1210 tumor model, 78 were active against melanoma, 45against human breast carcinomas, 18 against murine breast32carcinomas, 10 against murine lung cancers, 9 against murinecolon cancers, 5 against human colon cancers and 1 againsthuman lung cancer (156, 165).To aid detection of potential histiospecific activityof anticancer agents, the NCI has developed the mean graphand the COMPARE computer program (118). The mean graph,which is a type of bar graph, displays cytotoxic activity ofthe test agent against each cell line in relation to theaverage activity observed against all cell lines of thetumor panel. The graph is constructed by projecting barshorizontally, to the right or left of the mean, depending onwhether or not cell sensitivity to the test agent is greateror less than average. The length of the projection bars isdirectly proportional to the difference from the mean.Consequently, a "fingerprint" pattern is generated for eachcompound tested against the human tumor cell line panel.This COMPARE computer program can then compare thisfingerprint pattern to patterns generated with other testagents, thereby identifying similarities in biologicalproperties and/or chemical structure and properties (118).7. Methods to Assess ChemosensitivityOne of the difficulties associated with in vitro drugscreening is how to accurately measure tumor cell viabilityfollowing exposure to a test agent. Numerous assays havebeen developed, but unfortunately each has its limitations.These assays include the human tumor colony forming33(clonogenic) assay (HTCFA)(64, 139, 145, 146), thesulforhodamine B (SRB) assay (149), tritiated thymidineuptake (74, 161), neutral red assay (12, 13, 49), theCoomassie blue protein assay (147), the Hoechst 33342 assay(39, 129), the fluorescien diacetate assay (133), themethylene blue assay (112), the XTT tetrazolium dyereduction assay (141), the MTT tetrazolium dye reductionassay (2, 20, 33, 67, 107, 134, 167) and many others.The HTCFA, MTT, XTT, and SRB were all investigated aspossible tools in the NCI's drug-discovery and developmentprogram. The HTCFA (Figure 6) was deemed not suitable forlarge-scale screening because it is labour intensive, isplagued by low plating efficiencies for certain tumor typesand is difficult to obtaining single-cell suspensions (145,146).34tumordissociate to generate a single cell suspensionincubate cell suspension with drugcells are seeded as isolatedcells in a semisolid medium14-21 days in culturecount coloniesFigure 6. The human tumor colony forming assay.The assay originally selected by the NCI to use in thein vitro Stage I screen was the calorimetric MTT (1-[4,5-dimethylthiazol-2-y1]-2,5-diphenyl tetrazolium bromide)tetrazolium dye reduction assay (Figure 7)(2). The MTT assaywas observed to be a rapid, sensitive, and reproduciblemethod for in vitro anticancer drug screening (2, 20, 67,134). MTT is a soluble yellow substrate that is cleaved bythe dehydrogenase enzymes of mitochondria in living cells toform an insoluble purple formazan product (153). The amountof formazan produced by a population of cells isproportional to the number of viable cells in the population(107). Formazan levels are measured with a spectrophotometerdesigned to read 96-well microculture plates (ELIZA reader),which is interfaced with a computer to analyze results. Theuse of the 96-well microtiter plates in conjunction with the35ELIZA reader and the computer makes this procedure highlyefficient.Figure 7. NCI MTT assay protocol.From: Rubenstein L.V., Shoemaker R.H., Paull K.D., SimonR.M., Tosini S., Skehan P., Scudiero D.A., Monks A. and BoydM.R. (1990) Comparison of in vitro anticancer-drug-screeningdata generated with a tetrazolium assay versus a protein .assay against a diverse panel of human tumor cell lines. J.Natl Cancer Inst. (U.S.A.) 82: 1113-1118.A slight variation of this assay was developed toeliminate the aspiration/solubilization step (the insoluableformazan must be dissoled in solvent prior to colorimetry).The reduction of XTT (2,3-bis[2-mehtoxy-4-nitro-5-sulfopheny1]-5-[(phenylamino)carbony1]-2H-tetrazoliumhydroxide) results in a formazan product that is watersoluble, thereby eliminating the need for additionalsolvents. Some serious drawbacks associated with the MTT andXTT assays are i) agents that reduce tetrazolium salts bydirect chemical action may yield erroneous results (2), ii)agents that block cell-mediated MTT reduction indirectly byinhibiting cellular respiration may confound results (2),iii) MTT formazan production varies between cell lines and3 6culture age, iv) agents that alter the number ofmitochondria per cell may confound results and v) the colorintensity of the dissolved formazan is not stable, and hasto be measured at a precise time point (136). The variationof formazan production between cell lines and culture age isrelated to the availability of glucose at the time of theassay. Low levels of available glucose is related todecreased mitochondrial activity, which in turn is relatedto decreased formazan production (167). With the NCI's StageI screen operating at 1,000 microtiter plates/day, it becamedifficult to coordinate microculture plate reading. As aresult the NCI was prompted to develop the SRB assay (Figure8)(149). SRB is a aminoxanthene dye with two sulfonic groupsthat binds to basic amino acid residues of protein (91). TheSRB assay was determined to be a sensitive and rapid methodof measuring the cellular protein content of adherent andsuspension cultures. Unlike the MTT and XTT assays, theSRB's colorimetric endpoint is indefinitely stable (136,149). A study comparing the SRB assay and the MTT assayobserved that the two perform similarly (136).A00ORU 0 3^ WITH TCAFI X CELLS^ SOL U St1.121SOUND STAINIISMOV ISUP ERNATAN TWISH IXTWO OATS 1-2 WASAEU OV ASU• I RN ATAN TWISH GXA00SA•• •DAY^10-20 YIN^• DAY37• L AT CCELLSI /OATREADAT III NM• let•• .ay be N.I. •t I It•• • Wets tor tater or ^ In•Figure 8. NCI SRB assay protocol.From: Rubenstein L.V., Shoemaker R.H., Paull K.D., SimonR.M., Tosini S., Skehan P., Scudiero D.A., Monks A. and BoydM.R. (1990) Comparison of in vitro anticancer-drug-screeningdata generated with a tetrazolium assay versus a proteinassay against a diverse panel of human tumor cell lines. J.Natl. Cancer Inst. (U.S.A.) 82: 1113-1118.8. Predictive Value of In Vitro Chemosensitivitv Assays The most paramount question to be answered with regardto in vitro assays is whether or not in vitro activitytranslates into in vivo activity. In vitro chemosensitivityassays can be categorized as patient-oriented or as drug-oriented. The former is used to predict the response of apatient to cytotoxic drugs. The latter is used for measuringthe chemosensitivities of compounds. Although most of thestudies addressing whether in vitro activity correlates within vivo activity have been performed with respect to patientoriented in vitro assays, many of the conclusions arerelevant to in vitro drug screening. The HTCFA (Figure 6),first described by Hamburger and Salmon (64) has been themost studied of the in vitro assays. In this assay, stemcells are exposed to various concentrations of drugs and the38proportion of colonies that develop relative to a control isplotted against the dose level of the agent. A review of2,300 different in vitro-in vivo clinical correlations usingclonogenic based assays was presented by Von Hoff in 1990(170). The results of this review are as follows:Sensitive in vitro and in vivo^512Sensitive in vitro but resistant in vivo 226Resistant in vitro but sensitive in vivo 135Resistant in vitro and in vivo^1,427Total^ 2,300An analysis of these numbers reveals that if an agentis active in vivo, an in vitro assay will predict it 79% ofthe time. Conversely if the agent is not active in vivo, anin vitro assay will predict it 86% of the time. Since thenumber of correlations involved in generating these numbersis very large, these results are statistically powerful.Another study, in which patients with advanced metastaticcancer were randomized to receive single-agent chemotherapyas decided by an oncologist or by a clonogenic assay,demonstrated that an in vitro clonogenic assay could improvethe response rates to chemotherapy (169). Patients whoreceived chemotherapy as decided by the assay were observedto have a response rate of 21%, whereas those who wereadministered chemotherapy as selected by a clinician wereobserved to have a response rate of 3%. No difference in39survival was observed between the two arms of the study, butthis can be attributed to the current stock of anticancerdrugs and the single-agent chemotherapy regime administered.Although the results of these two reports confirm that invitro activity can reasonably predict in vivo activity,several comments must be made in order to keep these resultsin perspective: 1) Only certain tumor types are suitable forthe HTCFA (142, 145, 170). Von Hoff in his review of invitro-in vivo correlations only examined solid tumors. Ofthe 13,932 tumors involved in the prospective randomizedstudy mentioned above, approximately 30% of the tumorsdisplayed adequate growth in vitro. Adequate growth in vitrohas been considered to be •30 colonies per 5 x 105 cellsplated (142). The tumor types which were found to be themost suitable for the HTCFA are mesothelioma, ovary, corpusuteri and kidney, whereas leukemias, lymphomas and myelomaswere observed to be not suitable for the HTCFA. 2)Clonogenic assays have been plagued by quality controlproblems (ie: cell clumping) which have resulted inerroneous results (145, 146). Recent technical advances andthe establishment of stringent assay parameters and controlshave minimized these quality control problems (145).There are several limitations associated with thepredictive value of all in vitro assays. The most obvious isthe inability of in vitro assays to detect agents thatrequire metabolic activation, factors that could inhibitangiogenesis, invasion or metastases or agents that act via40effector cells of the immune system such as macrophages andnatural killer cells. One method proposed to minimize theformer of these is inclusion of liver microsomalpreparations in the incubation mixture (3, 86).Another problem associated with in vitro assays is thatit is difficult to determine the appropriate drugconcentrations and exposure times in culture that reflectthose attainable in vivo. This problem is further compoundedby the fact that drug penetration barriers exist within thetumor, thus different sites within the tumor may be exposedto different concentrations of drug (96). For patient-oriented predictive in vitro assays this is a seriousproblem. For drug-screening oriented in vitro assays, thisproblem is not as important since the purpose of these testsis to determine if a given agent has the potential forantitumor activity. Usually the drug concentrations used inpreliminary drug-screening studies are five ten-folddilutions starting with the highest soluble concentration oran arbitrarily selected concentration (usually 10-3 M forpure compounds or 250 gg/mL for crude extracts)(14, 103). Aknowledge of the potency of the test compound in vitro mayprovide a reference point on which to base in vivo drugdosages (122).The basis of all in vitro chemosensitivity assays isthe hypothesis that response in vivo is determined by theinherent chemosensitivity of the cells in vitro.Experimental evidence has demonstrated that this is not41always the case and that other factors determine tumorresponses in vivo (4, 122). A study by Phillips et a/.(122)demonstrated that a murine colon adenocarcinoma cell linegrown in vivo exhibited varying responses to chemotherapydepending on the site of tumor growth (subcutaneous orintraperitoneal). The reasons for this site-dependentresponse to chemotherapy is believed to be due to the drug'sability to access different tumor sites, tumor hypoxia andcell kinetics (4, 96). Furthermore, factors affectingchemosensitivity such as the proliferative status of tumorcells, the 3-dimensional tumor geometry, the presence of anextracellular matrix and variations in pH, 02 tension andosmolarity observed in vivo may not necessarily bereproduced in vitro (37, 122).The validity of the HTCFA has been examined by testingthe assay against a panel of standard agents (10 clinicalagents, 5 highly toxic compounds with no therapeuticefficacy and 5 nontoxic clinically ineffectivecompounds)(145). Using a 20% in vitro response rate(decrease in colony number as compared to control) asactivity criterion, 9 of the 10 clinical agents wereobserved to be active, as were all 5 of the highly toxiccompounds. None of the nontoxic clinically ineffectivecompounds demonstrated significant activity. A studycomparing the MTT and the HTCFA using a range of anticancerdrugs on a variety of cell lines demonstrated that the twoassays produce similar results (2, 20), as do both the MTT42and SRB assays with one another (149). Differences inrelative chemosensitivities between cell lines weremaintained when comparing the MTT and HTCFA assays (20).Other experiments comparing in vitro chemosensitivities ofleukemia cell lines with carcinoma cell lines against apanel of cytotoxic agents has demonstrated that carcinomacell lines tend to be more chemoresistant than leukemic celllines, especially when exposed to DNA binding drugs (4, 52,115). These results suggest that the refractoriness ofcarcinomas as compared with leukemias is due to inherentresistance of the tumor cells themselves. Ultimately,however, only results from clinical evaluations will tellwhether or not an in vitro human tumor cell line panel is avalid means to detect clinically effective anticancer agents(14, 62, 122, 148).9. Natural Products as a Source of New Anticancer Drugs The NCI has not only changed the way they investigatenew potential anticancer drugs, but also the way it selectscompounds for screening. In addition to screening someacquired synthetic compounds, the NCI is putting a greateremphasis on screening natural products because theyrepresent a vast untapped source of agents with novel modesof action and unique structures. Currently approximately 20-25% of U.S. prescriptions contain natural plant products,even though only a handful of a quarter of a million species43of tropical plants have been examined for pharmaceuticalpurposes (127). The yield of cytotoxic agents from naturalsources is generally considered to be several orders ofmagnitude greater than randomly acquired synthetic agents(21, 83). Table 3 lists currently available anticanceragents that are derived from natural sources. In 1959, theNCI initiated its fermentation broth and plant screeningprograms. In 1972, the marine acquisition program wasinitiated. By 1980, approximately 200 000 fermentation brothextracts, 114 000 plant extracts and 17 000 marine extractswere tested (127, 157). Of these, samples from marinesponges and tunicates exhibited exceptional activity (157).These marine organisms produce toxins as a mechanism ofself-defence (168). Many interesting and important compoundshave been isolated from marine organisms. For example, thedevelopment of cytarabine, a widely used antileukemicchemotherapeutic agent, was based on leads provided by theisolation of 1-B-arabinofuranosyl derivatives of uracil fromthe Caribbean sponge Cryptotethya crypta (9, 89). Didemnin-B(Figure 9), a metabolite of the tunicate Trididemninsolidium (130), has entered Phase II human cancer trials(157, 168). The bryostatins, a family of metabolitesisolated from the bryozoan Bugula neritina (120), is alsoundergoing clinical trials (157, 168). Palytoxin (Figure10), whose structural analysis represents a considerabletriumph in natural product research, is derived from thePalythoa species and is the most toxic non-proteinous agent•^44ever discovered (104). As the structures of these agentsillustrate, the structural diversity of compounds isolatedfrom marine organisms appears boundless.Figure 9. Molecular structure of Didemnin B.From: Suffness M. and Thompson J.E. (1988) National CancerInstitute's role in the discovery of new antineoplasticagents. In: Biomedical Importance of Marine Organisms(Fautin D.G., ed). California Academy of Sciences, SanFrancisco. pp 155.Figure 10. Molecular structure of palytoxin.From: Ireland C.M., Roll D.M., Molinski T.F., McKee T.,Zabriski T.M. and Swersey J.C. (1988) Uniqueness of themarine chemical environment: Categories of marine naturalproducts from invertebrates. In: The Biomedical Importanceof Marine Organisms (Fautin D.G., ed), California Academy ofSciences, San Francisco. pp 54.The sources and number of samples collected by the NCIfrom 1986 to 1991 is summarized in Table 5. This acquisitionprogram is devised so that the greatest possible diversityof plant and marine life is acquired.Table 5. Sources and number of samples collected by the NCI(1986-1991).MARINE MICROORGANISMS5 000 shallow water, Indo-Pacific5 000 deep water, Caribbean and AtlanticTROPICAL RAINFOREST PLANTS7 500 Southeast Asia7 500 Africa and Madagascar7 500 Central and South AmericaMICROORGANISMS1 500 fungi, soil origin1 500 cyanobacteria, soil and other origins2 000 marine microorganisms*From: Suffness M. and Thompson J.E. (1988) National CancerInstitute's role in the discovery of new antineoplasticagents. In: Biomedical Importance of Marine Organisms(Fautin D.G., ed). California Academy of Sciences, SanFrancisco. pp 155.46There are several problems associated with naturalproduct screening:1. Crude extracts may not be readily soluble.2. The active ingredient in a crude extract may be presentin exceedingly small amounts, so that even though it maypotentially be extremely active, it may not be detected.3. It may be difficult to isolate the active ingredient inthe crude extract.4. It may be difficult to acquire sufficient quantities ofthe active agent once it is identified.This last problem is usually considered to be the mostsevere. Several options to resolve this obstacle have beenproposed (11, 157):1. Use tissue culture to grow the cell(s) responsible forthe production of the agent of interest.2. Identify a precursor to the agent in question thatis easier to isolate.3. Maintain rare marine invertebrates in tanks and plants ingreenhouses.4. Collect close relatives of the plant or animal inquestion.5. Utilize genetic engineering techniques.4710. Thesis Objectives The first objective of this thesis was to select andimplement an in vitro assay suitable for evaluating theeffects of various test agents on the cellular viability ofhuman breast cancer cell lines and normal and malignanthuman breast epithelial cells (HBEC) in primary culture.This laboratory has, in the past, relied on theincorporation of radiolabelled thymidine as an indicator ofcell growth. It was felt that this assay, althoughreproducible, was not suitable for larger scale screening.After a review of the literature, the MTT tetrazolium dyereduction assay was selected. Unfortunately, there was agreat variation in the literature regarding the protocol forthe assay, especially for assay parameters such as solvent,MTT incubation concentration and MTT incubation time. An MTTassay protocol had to be elucidated that would optimize thesensitivity and the accuracy of this assay for use on humanbreast cells.The principal objective of this thesis was to determinethe effects of 7 crude extracts and 8 pure metabolitesderived from marine organisms on the cellular viability of 4drug-sensitive and 2 drug-resistant cell lines. In addition,it was a perogative to see if any agents that demonstratedsignificant cytotoxicity against the drug-resistant humanbreast cancer cell lines may circumvent the MDR phenotype.This thesis also investigated the cytotoxicity of themore potent test agent on normal and malignant human breast48epithelial cells (HBEC) in primary culture, which moreclosely resemble counterpart cells in vivo than do celllines. The normal HBEC were included in the screeningprotocol to detect tumor-specific agents.49CHAPTER 2 MATERIALS AND METHODS1. Isolation and Purification of Compounds The procedures used for the isolation and purificationof the geodiamolides (34), bromotopsentin (105),xestospongin E (99), bastidin-4 (100), imbricatine (17),chromodorolide A (40), jaspamide (29), glaciasterol A (123)and the Garvea extract (46) have previously been described.The crude extracts of the other organisms were prepared byhomogenizing the organisms in methanol, filtering andevaporating the methanol in vacuo. The isolation andpurification of compounds used for cytotoxicity testing didnot constitute part of this thesis work. They were providedby Dr. Raymond Andersen, Departments of Chemistry andOceanography, U.B.C.All samples were stored lyophilized in a refrigeratoruntil ready for use. When ready for use, some of thecompound was transferred to a 250 gl microfuge tube, weighedand then dissolved in enough solvent so that all of thesample is dissolved but the volume is still workable(typically about 100 gl). The solvent used for all samplesexcept for imbricatine was ethanol. Sterile saline was usedfor dissolving imbricatine. The 250 gl microfuge tubescontaining dissolved samples were then placed within 1.5 mlscrew cap microfuge tubes and these were then wrapped inparafilm (to minimize any evaporation of solvent) and storedin the freezer at -5°C.502. Cell Line Culture The MCF-7, T-47D, ZR-75-1 and MDA-MB-231 cell lines,which originated from effusions from women with metastaticbreast cancer, were obtained from the American Type CultureCollection (ATCC). The MDA-Alr cell line was furnished byDr. W. McGuire at the University of Texas, and the MCF-7 Adrcell line was obtained from Dr. K. H. Cowan at the NationalInstitute of Health (NIH). The cell lines were maintained in25 cm2 Corning tissue culture flasks containing 10 ml ofgrowth medium consisting of Ham's F12/Dulbecco's modifiedEagle medium [F12/DME; 1:1 (v/v); Terry Fox MediaPreparation Services] supplemented with 10 mM HEPES (H) andeither 10% fetal bovine serum (FBS; Gibco Laboratories; forMDA-MB-231), 5% FBS and 5 pg insulin/ml (Sigma; for MCF-7,T-47D and ZR-75-1), 10% FBS and 1 pg Adriamycin/ml (ADR;Adria Laboratories; for MDA-Alr), or 5% FBS, 5 pg insulin/mland 0.5 pg ADR/ml (for MCF-7 Adr)(Appendices 1-4). All ofthese cell lines grow adherent to the tissue cultureplastic.Cultures were maintained in a humidified atmosphereconsisting of 5% CO2 and 95% air at 37°C. Media changes wereevery second to third day (ie: M-W-F). Multiple flasks ofeach cell line were maintained. The rationale for this wasthat the drug samples were dissolved in very small amountsof ethanol (_ 100 gl). As a result of this small workingvolume, it was felt that repeated opening and closing of thedrug-containing microfuge tubes would result in excessive51evaporation of the solvent, thus altering the concentrationof the drug. To minimize the number of opening and closingof the tube, multiple flasks of cells were used toindependently set up (ie: no cell pooling) separate drug-testing cultures on microtiter plates at one time. If at anytime the stock drug concentrations became suspicious, eitherthe vehicle was evaporated in a N2 atmosphere and the samplewas reweighed and redissolved, or a fresh stock solution wasprepared from previously undissolved drug.Cells were subcultured when approaching confluence byremoving the growth medium, washing the culture with 5 ml ofF12/DME/H (to remove any residual FBS), and adding 7 ml ofeither warmed saline-trypsin-versene (MCF-7, T-47D, ZR-75-1,MDA-MB-231 and MCF-7 Adr)(Appendix 5) or warmed calcium andmagnesium-free phosphate-buffered saline (PBS) supplementedwith 1 mM ethylenediaminetetraacetic acid (EDTA; MDA-Alr)(Appendix 6). Gentle agitation of the tissue cultureflasks for 3-5 min resulted in a single-cell suspension. Toomuch or too little agitation resulted in cell clumping,which was found to be irreversible. The cell suspension wasthen centrifuged in a clinical centrifuge at 100 x g for 3min. The supernatant was then removed and 2 ml of theappropriate growth medium (Appendices 1-4) were then addedto resuspend the cells. A small fraction (0.4 ml) of thissuspension was then used to seed a new tissue culture plate.At the time of the experiment the cells were resuspended andcentrifuged at 100 x g for 3 min. The supernatant was then52removed and the pellet was resuspended in 2 ml of F12/DME/H.An aliquot of the cell suspension (0.1 ml) was placed in asmall glass test tube and a drop of trypan blue dye wasadded. Viable cells are able to prevent the trypan blue dyefrom penetrating the plasma membrane, whereas dead cellstake up the dye and thus stain blue. The viable cell numberwas then determined by counting on a hemacytometer. Aftercell counting, the remaing 1.9 ml of cells suspended inF12/DME/H was centrifuged at 100 x g for 3 min, thesupernatant removed and the cell pellet resuspended inF12/DME/H supplemented either 5% FBS and 5gg insulin/ml(MCF-7, T-47D, ZR-75-1 and MCF-7 Adr) or 10% FBS (MDA-MB-231and MDA-Alr) to result in a final viable cell concentrationof 4 x 104 cells/ml. Cells were then seeded at 4 x 10 3cells/well onto 96-well microtiter plates in 100 gl volumeof the appropriate medium (Appendices 7 and 8) and thenreincubated. Cells were only seeded in the inner 10 x 6matrix of the 12 x 8 microtiter plate, whereas cell-freeF12/DME/H was placed in the wells bordering the edge of themicrotiter plate. This was done to minimize anydiscrepancies in the microenviroments between wellsbordering the plate edge and those that do not. In addition,only the 8 x 6 matrix towards the left side of themicrotiter plate was used for the T-47D cell line since itwas observed in early experiments that alterations in pH ofthe growth medium stimulated the growth of this cell line.Since I work right to left when removing and adding media,53the cells to the right side of the microtiter plate wereexposed to the air of the laminar flow hood longer thanthose towards the left side of the plate, and thus had thepH of their growth media altered to a greater extent.3. Human Breast Epithelial Cells in Primary CultureNormal tissue cells from reduction mammoplasties andmalignant tissue from biopsies and mastectomies wereobtained from collaborating surgeons at local hospitals.Malignant breast tissue was used only from patients who hadno previous chemotherapy. During surgery, the tissue wasaseptically placed into plastic sterile cups containingtransport media (Appendix 9). Immediately following surgery,the cups were placed on crushed ice in a cooler andtransported to the tissue culture room. Under sterileconditions, excess fat was removed from the tissue samplesand the samples were finely minced using 2 scalpels. Theminced tissue was then transferred to 1.7 ml cryotubes sothat the tubes were half filled. Freezing medium (Appendix10) was then added to bring the volume to 1 ml. The tissuewas then gradually frozen in liquid nitrogen to be used at alater date. At the time when the tissue was needed, the vialwas removed from liquid nitrogen and quick-thawed byagitating in a 37°C waterbath. The vial was wiped with aKim-wipe saturated with 70% ethanol prior to opening. Onceopened, the minced tissue was transferred to a 15 mlcentrifuge tube. The vial was then rinsed with a few ml of54F12/DME/H which was also placed into the centrifuge tube.The tube was then centrifuged at 100 x g for 3 min and thesupernatant discarded. Following this, the pellet was washedwith 5 ml of F12/DME/H, recentrifuged at 100 x g for 3 minand then resuspended in 5 ml warmed dissociation medium(Appendix 11). This mixture was then transferred to a 125 mlBellco dissociation flask. The 15 ml centrifuge tube wasrinsed several times with warmed dissociation medium thatwas also transferred to the dissociation flask. The totalvolume of dissociation medium in the flask was then broughtup to 50 ml. The flask was covered first with sterile tinfoil followed by parafilm, then placed in a gyrating shakerinside a 37°C incubator for 16-18 h (until no large tissuefragments remained). After dissociation, the solution had acloudy appearance with a layer of liquid fat floating ontop. The dissociation mixture was transferred to a 50 mlcentrifuge tube. The flask was rinsed twice with 5 ml ofF12/DME/H, with each rinsing being added to the 50 mlcentrifuge tube. The tube was then centrifuged at 40 x g for30 sec to separate undigested tissue fragments from the cellsuspension. The supernatant was transferred to a second 50ml centrifuge tube. The pellet in the first tube was washedwith 5 ml of F12/DME/H, spun at 40 x g for 30 sec, with thesupernatant being added to the second 50 ml centrifuge tube.This washing and centrifuging of the pellet was repeated,with the supernatant being added to the second centrifugetube and the pellet being discarded. The cell suspension of55the second tube was centrifuged at 80 x g for 4 min toseparate the epithelial fraction (pellet) from the stromalcells, undigested stroma and cellular debris (supernatant).After discarding of the supernatant, the pellet wasresuspended in 1 ml of F12/DME/H and viable cells determinedby trypan blue exclusion and counted on a hemacytometer.Following cell counting, the suspension was centrifuged at100 x g for 5 min, the supernatant discarded and theepithelial cell pellet resuspended in growth medium(Appendix 12) such that the final cell concentration was 10 5cell/ml. These cells were plated into as many wells aspossible in the inner 10 x 6 matrix of collagen coated 96-well microtiter plates. The wells were collagen coated byadding one drop of rat tail collagen (Appendix 14) to eachwell, swirling the plates to ensure complete coverage of thewell and then allowing the plate to dry for 1 h in a laminarflow hood under UV light. Since the cell yield from eachtissue sample was very small, usually only a portion of eachmicrotiter plate was utilized. Tissue samples which hadinsufficient cell yields were discarded. All wells notreceiving cells contained F12/DME/H. The cell concentrationused was 104 cells/well in 0.1 ml of growth medium. Thecultures were incubated at 37°C in 5% CO2 and 95% air. Thegrowth medium of these cultures contained the patient's ownserum (Appendix 13). Culturing cells in the serum of thepatient from whom the tissue was received is the routineprotocol for chemosensitivity testing in our laboratory (42,5643). It has been observed that fibroblasts in serum-containing media are stimulated to grow (44). As a result ofthis situation, the tissue culture plates during their 5 dayculturing period were monitored using a phase contrastmicroscope to ensure that the dominant cell population(>80%) had an epithelial-like morphology (ie: polygonalshaped).4. Compound-Testing Protocol On day 0, the cells (cell lines and HBEC in primaryculture) were seeded onto the microtiter plates in theappropriate growth medium (Appendices 7, 8 and 13) asdescribed earlier. After 24 h, the medium was removed and100 gl of fresh medium containing either 5 serial dilutions(typically 100, 10, 0.1, 0.01 and 0.001 pg/m1) of the agentto be tested or 5 serial dilutions of the corresponding drugvehicle (ethanol or saline) were added to the wells. Drugcontrol cultures received the appropriate growth medium withno drug or vehicle. Each drug concentration, drug vehicleconcentration and control condition (no drug or vehicle) wasrepresented by 4 to 6 wells within each experiment. Eachexperiment was repeated 2-4 times, using independentcultures each time. Cultures were incubated with the drugsor drug vehicles for 48 h. The cells were then washed oncewith F12/DME/H to remove any residual drug and cultured inthe appropriate drug-free medium (Appendices 7 and 8) for 48h. A 48 h drug incubation period was selected to ensure that57all cells passed through at least one cell cycle, thusallowing phase-specific and cell-cycle-specific agents toexert their effects. The cell-cycle times of MCF-7, T-47D,ZR-75-1, MDA-MB-231, MCF-7 Adr and MDA-Alr cells used inthis laboratory are 30, 22, 44, 21, 23 and 18 h,respectively. A 48 h drug recovery period was selected toensure enough time had been allowed for the cytotoxiceffects of the drug to be realized. Inhibitors of proteinsynthesis usually require a drug recovery period ofapproximately 24 h, whereas inhibitors of DNA synthesisrequire a recovery period of 48-96 h (135). After therecovery period, viable cells were quantified by the MTTassay.In some experiments performed to detect agents thatmight circumvent the MDR phenotype, MCF-7 Adr cells wereincubated in medium containing the concentration ofAdriamycin that results in a 50% decrease in cell growth forthis line (ED50, 5 µg/ml) as well as serial dilutions of thedrug to be tested.5. Preparation of Standard Curves for Each Cell Line Since the relationship between absorbance at 540 nm andcell number is not necessarily linear when using the MTTassay (124), a standard curve for each cell line wasprepared. Serial dilutions (2 x 103 to 65 x 103 cells/well)of each cell line were seeded onto 96-well microtiter platesin the appropriate growth medium (Appendices 7 and 8).58Fifteen hours after cell seeding, the cells were quantifiedusing the MTT assay.6. MTT Assaya. Preparation of MTTMTT (Sigma) powder was weighed, placed in a 17 x 100 mmpolypropylene test tube and dissolved in 8 ml of warmedserum-free and phenol red-free F12/DME/H (Sigma). With theuse of a small spatula, the aggregates of MTT were broken upand the solution stirred until all the MTT was dissolved.The solution was filter sterilized through a Millex GV 0.22gm syringe filter to remove any pathogens, undissolved MTTand any spontaneously formed formazan crystals. Warmedserum-free and phenol red-free F12/DME/H was then added tobring the final concentration to 1 mg MTT/ml.b. MTT Assay Protocol The procedure for the MTT assay used in all drugchemosensitivity experiments is summarized in Figure 11. Themedium was removed and 100 gl of MTT (1 mg/ml) dissolved inserum-free and phenol red-free F12/DME/H were added to eachwell. Serum-free medium is used so that there is noprecipitation of any residual serum protein upon addition ofthe solvent, thereby decreasing background absorbancelevels. Similarily, phenol red-free medium is used toprevent interference of the red pH indicator during theabsorbance readings. Following a 5 h incubation period at5937°C in 5% CO2 and 95% air, 100 Al of 20% formol in salinewere added to each well. Following 30 min at roomtemperature, the medium was removed using a multichannelpipette (Titertek) and 100 Al of anhydrous isopropanol wereadded to each well to dissolve the formazan crystals. After1 h, the plates were gently agitated, and absorbance valuesof each well were determined at 540 nm using a 96-wellmicrotiter plate reader (model EL 311, Biotek Instruments)interfaced with a MacIntosh Apple computer.60Figure 11. MTT assay protocol.cells on a 96-wellmicrotiter plateremove mediumadd 100 Al of 1 mg MTT/mlserum-free and phenol red-freeF12/DME/H5 h at 37°Cgently add 100 Al of 20% formolin saline30 min at room temperatureremove supernatant with amultichannel pipettoradd 100 Al of isopropanol1 h at room temperaturegently agitate platemeasure A540To take into account any absorbance due to residualMTT, MTT control wells (which do not contain cells) wereincorporated into the experiment. Absorbance values obtainedfor these MTT control wells were subtracted from absorbance61values obtained for cell-containing wells. The correctedabsorbance values were then converted to cell numbers viastandard curves specific for each cell line. Cytotoxicitywas determined by comparing drug-treated cultures withcontrol cultures (which are assigned a value of 100%). TheED50 value was then determined from the dose response curve.A control to monitor reduction of MTT to formazan bydrugs that have reducing properties was not included in theexperiments. Previous studies have demonstrated thatreducing agents do not influence results generated by theassay (20).The elucidation of the various assay parameters(solvent, MTT incubation concentration and MTT incubationtime) presented in this assay protocol were determined asdescribed below (c-e).c. Selection of Solvent for the MTT Assayi) With No Removal of Supernatant Prior to Solvent AdditionMCF-7 cells were seeded at 1.3 x 105 cells/ml in 430 Alof 1 mg MTT/ml serum-free and phenol red-free F12/DME/H in 312 x 75 mm polypropylene test tubes. Following a 4 hincubation at 37°C in 5% CO2 and 95% air, 870 Al of eitheranhydrous isopropanol, ethanol or spectrophotometric gradedimethyl sulfoxide (DMSO) were added to each tube. The cellgenerated formazan was allowed to dissolve in the solventfor 2 h at room temperature. The formazan solvent mixturewas then transferred to a glass cuvette and a wavelength62scan from 400 to 640 nm was generated for each solventmixture using a UV/visible scanning spectrophotometer(Beckman Instruments). The reference cell for each of the 3wavelength scans was the solvent containing 1.3 x 105 dead(lysed) MCF-7 cells and 430 gl of 1 mg MTT/ml in F12/DME/H.ii) With Removal of MTT Supernatant Prior to SolventAdditionMCF-7 cells were seeded at 2 x 105 cells/ml in 1.3 mlof 1 mg MTT/ml serum-free and phenol red-free F12/DME/H in 312 x 75 mm polypropylene test tubes. Following a 4 hincubation at 37°C in 5% CO2 and 95% air, the tubes werecentrifuged at 100 x g for 3 min, the supernatant wasremoved and 1.3 ml of either isopropanol, ethanol or DMSOwere added to each tube. The cell generated formazan wasthen allowed to dissolve in the solvent for 1 h at roomtemperature. The formazan-solvent mixture was thentransferred to a glass cuvette and a wavelength scan from400 to 640 nm was generated for each solvent mixture using aUV/visible scanning spectrophotometer. The reference cellsused for each of the 3 wavelength scans was the solventcontaining 2.6 x 105 dead (lysed) MCF-7 cells (cells thatwere not exposed to MTT).63d. Selection of Optimal MTT ConcentrationMultiple wells of serial dilutions of MCF-7 cells (312to 5 x 105) were seeded onto 96-well microtiter plates in100 gl of growth medium (Appendix 1). Following a 24 hincubation at 37°C in 5% CO2 and 95% air, the growth mediumwas removed and serial concentrations (0.125 - 5 mg/ml) ofMTT dissolved in 100 gl serum-free and phenol red-freeF12/DME/H were added to the wells. MTT addition was suchthat every concentration of MTT tested was exposed to theentire spectrum of cell densities. Following a 4 hincubation period at 37°C, the medium was removed and 100 glof isopropanol were added to each well. The formazan wasallowed to dissolve for 1 h at room temperature and then theplates were read on a 96-well microplate reader at 540 nm.e. Selection of Optimal MTT Incubation Time MCF-7 cells were seeded at 4 x 103 cells/well in growthmedium (Appendix 1) on 7 different 96-well microtiterplates. Following a 24 h incubation at 37°C in 5% CO2 and95% air, the growth medium was removed and 100 gl of 1 mgMTT/ml serum-free and phenol red-free F12DME/H were added toall the wells. The cells were reincubated at 37°C and at 1 hintervals for 7 h, a single microtiter plate was taken outof the incubator and the mean absorbance of each plate wasmeasured at 540 nm.64f. Removal of MTT Supernatant In the initial experiments outlined above (c-e), it wasobserved that removal of the MTT-F12/DME/H supernatant witha multichannel pipettor (Titertek) resulted in cell loss andtherefore increased assay error. It became apparent thatMCF-7 cells, once they have formed intracellular formazan,become easily detached from the tissue culture plastic. Tocorrect for this problem, a variety of alternative methodswere tested for removing the MTT supernatant. Briefly, MCF-7cells were seeded at 4 x 103 cells/well in 42 wells on 5different 96-well microtiter plates. Following a 24 hincubation period at 37°C in 5% CO2 and 95% air, the growthmedium (Appendix 1) was removed and 100 gl of 1 mg MTT/mlserum-free and phenol red-free F12/DME/H were added to eachwell. Following a 5 h MTT incubation period, the MTTsupernatant was removed by a variety of methods: 1) Thecells, prior to MTT supernatant removal, were fixed to thetissue culture plate by the addition of 100 gl of 20% formolin saline (to result in a final concentration of 10%formol). After 30 min at room temperature the MTTsupernatant was gently removed using a multichannel pipette.2) The MTT supernatant was removed by submersing the platein saline and then blotting the plate on paper towels. 3)The MTT supernatant was removed by first inverting the plateand "flicking" it with a sudden wrist motion, then dippingthe plate in saline and finally by blotting on paper towels.4) The MTT supernatant was removed by pipetting with a65multichannel pipette.Following removal of the MTT supernatant, 100 gl ofisopropanol were added to each well and the absorbance valueat 540 nm of each cell-containing well was measured. Fromthese values, the coefficient of variation (the ratio of thestandard deviation (S) relative to the mean (X) expressed asa percentage; S - X • 100%) was determined for each plate.g. Absorbance Limit of the MTT AssayTo determine the upper limit of absorbance at 540 nm inwhich the assay becomes insensitive to differences informazan concentration (ie: cell number), the absorbances ofserial dilutions (1-800 gg/ml) of pure MTT formazan (Sigma)dissolved in isopropanol were measured using aspectrophotometer. The absorbance values were used toconstruct a graph relating absorbance to formazanconcentration.To relate the absorbance spectrum of pure formazan tocell-generated formazan, a wavelength scan (400-640 nm) ofpure formazan dissolved in isopropanol at 30 gg/ml wasgenerated using a UV/visible scanning spectrophotometer.66CHAPTER 3 RESULTS1. Optimization of MTT Assay Parameters a. Selection of Solvent for the MTT AssayThe MTT tetrazolium dye reduction assay utilizesformazan production as an index of cell populationviability. Figure 12 shows the wavelength scan from 400 to640 nm of cell-generated formazan dissolved in isopropanol,ethanol and DMSO in the presence of serum-free and phenolred-free F12/DME/H. A peak in absorbance was observed forisopropanol and ethanol at approximately 560 nm. The heightof the isopropanol peak (0.71) was slightly higher than thatfor the ethanol peak (0.68). A peak in absorbance for theDMSO curve was observed at a wavelength of 520 nm. Theheight of this DMSO peak (0.53) was substantially lower thanthat for isopropanol and ethanol. It was observed in thisexperiment that it was difficult to get the formazancrystals to dissolve in DMSO in the presence of F12/DME/H,even when the formazan dissolving period was extended to 2h. As a consequence, the experiment was repeated with themodification that the MTT supernatant was removed prior tosolvent addition.Figure 13 shows the wavelength scan from 400 to 640 nmof cell-generated formazan dissolved in isopropanol, ethanoland DMSO. A peak in absorbance was observed for isopropanolat 540 nm, for ethanol at 580 nm and for DMSO at 500 nm. Theheight of the isopropanol peak (2.67) was slightly higher67than that for the ethanol peak (2.39) and the DMSO peak(2.35). Since the ELIZA reader was equipped with a 540 nmfilter, these results indicated that isopropanol was themost suitable solvent for the MTT assay. No difficulty indissolving the formazan crystals was observed when usingDMSO as the solvent in this second experiment. The heightsof all the peaks for the 3 different solvents in the secondexperiment (removal of MTT supernatant) were severalmagnitudes higher than those generated in the firstexperiment (no removal of MTT supernatant).68Figure 12. Wavelength scans of cell-generated formazandissolved in DMSO, isopropanol and ethanol in the presenceof the MTT supernatant. MCF-7 cells were incubated in 1 mg MTT/ml serum-free andphenol red-free F12/DME/H for 4 h. Following the incubationperiod, 100 Al of either DMSO, isopropanol or ethanol wereadded to the cells to dissolve the cell-generated formazan.Two hours later, a wavelength scan from 400 - 640 nm wasgenerated for each formazan-solvent mixture. The resultsindicated that isopropanol was the most suitable solvent forreading A540 since it generated the highest peak (0.71) atthis wavelength.FIGURE 12^ 68A....■13....■ DMS0-0-- ETFIANOL-a-- ISOPROPANOLWAVELENGTH (nm)69Figure 13. Wavelength scans of cell-generated formazandissolved in DMSO, isopropanol and ethanol. MCF-7 cells were incubated in 1 mg MTT/ml serum-free andphenol red-free F12/DME/H for 4 h. Following the incubationperiod, the supernatant was removed and 100 gl of eitherDMSO, isopropanol or ethanol were added to the cells todissolve the cell-generated formazan. One hour later, awavelength scan from 400 - 640 nm was generated for eachformazan-solvent mixture. The results indicated thatisopropanol was the most suitable solvent for reading A540,since it generated the highest peak (2.67) at thiswavelength.FIGURE 136 9A- h-- ISOPROPANOL- 0-- ETHANOL- 13-- DIASOWAVELENGTH (nm)70b. Selection of Optimal MTT ConcentrationFigure 14 illustrates the relationship between MTTconcentration and absorbance over a variety of celldensities. At lower cell densities, slight differences inabsorbance were observed between all the concentrations ofMTT, with 1 mg MTT/ml being optimal. As cell numberincreases, the difference in absorbance between thedifferent concentrations of MTT also increases. At thehighest cell density (5 x 10 5 cells/well), 5 mg MTT/ mlresulted in the greatest absorbance (0.871), whereas 0.125mg MTT/ml resulted in the lowest absorbance (0.239). It wasexpected that cell density at the end of the drug testingexperiments would generally be _ 2 x 10 4 cells/well.Consequently, 1 mg MTT/ml was selected as the most suitableMTT concentration, because it resulted in the highest A540below this cell concentration.71Figure 14. Formazan production as a function of MTT incubation concentrationSerial dilutions of MCF-7 cells (312 - 5 x 10 5) were exposedto serial concentrations of MTT (0.125 - 5 mg/ml) such thatevery cell density was exposed to every MTT concentration.Following the MTT incubation period, the supernatant wasremoved, the formazan crystals were dissolved in isopvopanoland the A540 was measured. At cell densities below 10cells/well, 1 mg MTT/ml resulted in slightly higherabsorbance than the other MTT concentrations. Above 104cells/well the higher MTT concentrations resulted in greaterformazan production, with 5 mg MTT/ml being the optimal atthe very high cell densities. Since it was expected that thefinal cell density following the drug incubation periodwould generally be 2 x 104 cells/well, 1 mg MTT/ml wasselected as the optimal MTT concentration.FIGURE 14^ 71A- 13- 5 mg/ml-*- 4 mg/ml- 111-- . 2 mg/ml--40- 1 mg/ml0.5 mg/ml--D-- 0.25 mg/ml0.125 mg/ml2^ 4^5LOG CELL NUMBER72c. Selection of Optimal MTT Incubation TimeFigure 15 illustrates the relationship between lengthof MTT incubation and absorbance at 540 nm. Absorbance(0.302) was highest when the MCF-7 cells were incubated withthe MTT solution for 5 h, and consequently a 5 h MTTincubation time was selected for all further experiments. Itis interesting to note that the longer the MCF-7 cells wereexposed to the MTT solution after 5 h MTT incubation, thegreater the value was for the coefficient of variation(which is directly proportional to error). This relationshipis illustrated in Figure 16.73Figure 15. Formazan production as a function of MTTincubation time. MCF-7 cells were seeded at 4 x 104 cells/well on 7 differentmicrotiter plates. Twenty-four hours later, the medium(Appendix 1) was removed and 1 mg MTT/ml was added to allthe cells. The cells were reincubated at 37°C and at 1 hintervals, a single microtiter plate was removed from theincubator, and the mean A540 of its' cell containing wellswas determined. A peak in absorbance (0.302) is observed forcells that were incubated for 5 h, thereby indicating that a5 h MTT incubation time results in optimal formazanproduction. The error bars represent the standard error ofthe mean (standard deviation/ -n) of 2 experiments.FIGURE 15^ 7 3A0^2^4^6^8TIME (h)74Figure 16. Coefficient of variation as a function of MTTincubation time. The error associated with Figure 15 is represented here asthe coefficient of variation (standard deviation/mean x100%). The results demonstrate that the error associatd withthe assay increases significantly after 5 h MTT incubationtime.FIGURE 16Method of MTTsupernatant removalCoefficient ofvariation (%)Formol fixing andmultichannel pipetting8.8 Saline dipping withno flickingSaline dipping withflickingMultichannel pipetting15.811.016.575d. Removal of MTT Supernatant Table 6 lists the coefficients of variation fordifferent methods of removal of the MTT supernatant. Formol-fixing the cells to the microtiter plate and then removingthe MTT supernatant with a mulitchannel pipettor resulted inthe lowest coefficient of variation (8.8%), and consequentlyformol-fixing the cells followed by multichannel pipettingwas used for all subsequent experiments. The control studiesfor this experiment demonstrated that formalin did notinterfere with absorbance at 540 nm.Table 6. Coefficient of variation associated with differentmethods of MTT supernatant removal.e. Absorbance Limit of the MTT AssayFigure 17 illustrates the relationship between theconcentration of pure formazan and absorbance at 540 nm. Theupper limit of absorbance was approximately 3, whichcorresponded to a concentration of roughly 200 ugformazan/ml. Absorbance readings from the drug-screening76experiments very rarely exceeded 1.0.Figure 18 is the wavelength scan of pure formazan from400 to 640 nm. Included in the Figure is the correspondingwavelength scan of cell-generated formazan (from Figure 13).The peaks of the pure formazan and the culture-generatedformazan were both in the 540 to 560 nm range, therebyindicating it is valid to use pure formazan in thedetermination of the upper absorbance limit of the MTTassay.77Figure 17. Absorbance limit of the MTT assay. Serial dilutions (1 - 800 µg/ml) of pure formazan weredissolved in isopropanol and the A540 of each dilution wasmeasured. The upper limit of absorbance is observed to beapproximately 3.FIGURE 17 77A78Figure 18. Wavelength scans of pure formazan and cell-generated formazan dissolved in isopropanol. Pure MTT formazan dissolved in isopropanol was used togenerate a wavelength scan from 400 - 640 nm. The wavelengthscan of cell-generated formazan dissolved in isopropanol wasalso included in Figure 18. Both absorption spectrums had apeak in the 540 - 560 nm range.FIGURE 18792. Compounds Screened Against Drug-Sensitive Cell Lines The MTT assay protocol determined from the experimentsdescribed above is briefly described as follows. Followingthe 48 h drug recovery period, the medium was removed and100 gl of MTT (1 mg/ml) were added to each well. Following a5 h incubation period at 37°C, 100 gl of 20% formol insaline was added to each well to fix the cells. Following 30minutes at room temperature, the supernatant was removed and100 gl of isopropanol was added to each well. One hourlater, the plates were gently agitated and the absorbancevalues of each well was determined using a 96-wellmicrotiter plate reader.Figure 19 illustrates the wide range of chemicalstructures of all the pure samples tested in our in vitroscreening protocol. Table 7 summarizes the ED50 valuesobtained for compounds screened against the MCF-7, T-47D,ZR-75-1 and MDA-MB-231 cell lines. Of the 15 samplesscreened, 8 demonstrated significant cytotoxicity (ED50, _25 gg/ml) against either the MCF-7 or the T-47D cell line.Only these 8 samples were screened against ZR-75-1 and MDA-MB-231 cells; all 8 compounds were significantly cytotoxicagainst the former cell line, whereas 6 were significantlycytotoxic against the latter. The ED50 values for Adriamycinwere determined in all of these cell lines to provide astandard with which to compare the samples. The ED50 valuesrecorded for the two pure samples, the geodiamolides andjaspamide, were comparable with or lower than those noted80Figure 19. Molecular structures of pure metabolites isolated frommarine organisms.OH8 OAHOHOlinteicadneOHCe  Glaciasteml AFIGURE 19 Xescospongin E.4"--N" COgiNChromodorolide AJasparrideBromotopsendnBastadirk-481TABLE 7. ED50 values obtained for compounds screened against thedrug-sensitive breast cancer cell lines MCF-7, T-47D, ZR-75-1 andMDA-MB-231.AGENT ED50 values (µg/ml)MCF -7^T-47D ZR-75-1 MDA-MB-231GEODIAMOLIDES 0.0068±.0003 0.044±.003 0.033±.024 0.004XESTOSPONGIN E >>50 >50 ND NDBROMOTOPSENTIN >>100 32±5 ND NDPNG-31Aa 5.4±.5 0.6 7 24±1BASTIDIN-4 6.0±.6 0.58±.05 2.9±.1 2.6±.1IMBRICATINE >50 23±9 ND NDCHROMODOROLIDE A >>100 >>50 ND NDGARVEAa 33±9 19+3 21±1 23±1QCI-32a 1.2±.4 0.09±.02 0.09±.01 0.018±.001QCI 4-16-89a >100 50±22 ND NDPNG-137Aa 30+3 13+3 16±3 >50PNG-137Ba 98±2 35±12 ND NDQCI-117a »100 >>100 ND NDGLACIASTEROL A 19±1 21±3 26±1 >>50JASPAMIDE^0.027±.004 0.019±.009 0.042±.002 0.024±.001ADRIAMYCIN^0.10±.01 0.065±.009 0.005±.001 0.038±.006Data represent mean values ± standard error of the mean (standarddeviation/in). Only the compounds that exhibited significantcytotoxicity towards the MCF-7 and T-47D cell lines were screenedagainst the ZR-75-1 and MDA-MB-231 cell lines. ND, not done.a Crude extract82for Adriamycin. The values obtained for another sample, QCI-32, which is a crude extract, were 12-, 1.4-, 18- and 0.47-fold higher than those found for Adriamycin in screeningagainst the MCF-7, T-47D, ZR-75-1 and MDA-MB-231 cell lines,respectively. ED50 values as low as 0.6 µg/ml were noted forthe crude extract PNG-31A in the T-47D cell line; this valueis 10-fold higher than that determined for Adriamycin.Values recorded for the 5 remaining compounds that exhibitedsignificant cytotoxicity ranged from 0.58 to 33 µg/ml andwere at least 9-fold higher than those found for Adriamycin.Figures 20-23 represent the standard curves relatingabsorbance at 540 nm to cell number for the MCF-7, T-47D,ZR-75-1 and the MDA-MB-231 cell lines. Neither ethanol norsaline had any effect on drug-sensitive cell line viabilityin the concentrations used in the compound-testingexperiments.3. Compounds Screened Against Drug-Resistant Cell Lines Table 8 summarizes the ED50 values obtained for thesamples and for Adriamycin during screening against the Pgp-positive cell lines MCF-7 Adr and MDA-A1r. Of the 8 samplesscreened, 6 demonstrated significant cytotoxicity againsteither the MCF-7 Adr or the MDA-Alr cell line. Thegeodiamolides were 54- and 11 600- times more cytotoxic thanAdriamycin when screened against the MCF-7 Adr and MDA-Alrcell lines, respectively. Jaspamide was 6- and 318- timesmore cytotoxic than Adrimycin when screened against the83TABLE 8. ED50 values obtained for compounds screened against themultidrug-resistant breast cancer cell lines MCF-7 Adr and MDA-Alr.AGENT ED50 values (µg/ml)MCF-7 Adr MDA-AlrGEODIAMOLIDES 0.092 ±^.015^(14) 0.0006 ±^.0001^(0.15)PNG-31Aa 51 ± 2 (9.4) 41 ±^4^(1.7)BASTIDIN-4 >>20 (>3.3) >100 (>38)GARVEAa 27 ± 4 (0.82) 25 ±^3^(1.1)QCI-32a 10 ± 4 (8.3) 1.3 ±^.4^(72)PNG-137Aa >50 (>1.7) 25 ± 2^(<0.5)GLACIASTEROL A 18 ± 1 (0.95) 19 ± 2^(<0.38)JASPAMIDE 0.77 ±^.02 (29) 0.022 ±^.005^(0.92)ADRIAMYCIN 5 ± 2 (50) 7 ± 4^(184)Data represent mean values ± standard error of the mean (standarddeviation/in) for 2 - 4 experiments. Values in parenthesisindicate the magnitude of increase in resistance noted for theMDR daughter cell lines as compared with the corresponding drug-sensitive parent lines (ED50 of the resistant daughter cellline/ED50 of the sensitive parent line).aCrude extract84MCF-7 Adr and MDA-Alr cell lines, respectively. MCF-7 cellswere 50-times more resistant to Adriamycin than was theparent MCF-7 cell line, whereas MDA-Air cells were 184-timesmore resistant than was the parent line MDA-MB-231. Table 8also compares the ED50 values found for compounds screenedagainst the MDR cell lines MCF-7 Adr and MDA-Alr with thoseobtained using the drug-sensitive parent cell lines (ED50drug-resistant/ED50 drug-sensitive). The geodiamolidesGarvea, PNG-137-A, glaciasterol A and jaspamide wereobserved to have ED50 ratios g 1 for one or both cell types.Figures 24 and 25 represent the standard curvesrelating absorbance at 540 nm to cell number for the MCF-7Adr and MDA-Alr cell lines. Neither ethanol nor saline hadany effect on cell viability of the drug-resistant celllines in the concentrations used in the compound-testingexperiments.85Figures 20 - 25. Standard absorbance curves for the MCF-7, T-47D, ZR-75-1, MDA-MB-231, MCF-7 Adl.  and MDA-A1K cell lines. Serial dilutions of the cell lines were seeded in theappropriate growth medium (Appendices 7 and 8). Followingattachment of the cells to the tissue culture plate (15 h),the A540 value of each cell density was determined (Figure11). An A540 versus cell number graph was plotted for eachcell line (Figure 20: MCF-7, Figure 21: T-47D, Figure 22:ZR-75-1, Figure 23: MDA-MB-231, Figure 24: MCF-7 Adr andFigure 25: MDA-A1r) and a computer generated equation forthe best-fit line was determined (MCF-7: y = 8.060e-4 *xA0.7038, T-47D: y= 0.0048 * xA0.4852, ZR-75-1: y= 8.924e-4 * xA0.6849, MDA-MB-231: y = 4.944e-4 * xA0.6968, MCF-7Adr: y = 1.343e-4 * xA0.7953 and MDA-Alr: y = 0.001 *xA0.6477). Error bars represent standard error of the mean(standard deviation/"n) of 2 - 4 experiments.FIGURE 2085AFIGURE 21^85B85CFIGURE 22FIGURE 23 85DFIGURE 2485EFIGURE 25^ 85F864. Compounds Screened Against MCF-7 Adr- Cells in the Presence of AdriamycinTo determine whether compounds whose ED50 ratios wereg 1 might circumvent the MDR phenotype by interfering withPgp, MCF-7 Adr cells were incubated in the presence of theED50 concentration (5 gg/ml) of Adriamycin and in varyingconcentrations of the geodiamolides, Garvea, PNG-137A,glaciasterol A and jaspamide. No synergy was observedbetween any of the samples and Adriamycin (Figures 26-30),suggesting that these compounds do not interfere with Pgpfunction. The ED50 values of the various test agents in thepresence of Adriamycin are 0.95, 11, >50, 13 and 1 pg/m1 forthe geodiamolides, Garvea, PNG-137A, glaciasterol A andjaspamide, respectively. Additive cytotoxicity was observedbetween some agents (Garvea, PNG-137A, glaciasterol A andJaspamide) and Adriamycin (Figures 27 - 30). A 10-foldreduction in the cytotoxic activity of the geodiamolides wasobserved when MCF-7 Adr cells were exposed to thegeodiamolides in the presence of Adriamycin (Figure 26).87Figures 26 - 30. Dose response curves of MCF-7 Adr cells tovarying concentrations of the geodiamolides, Garvea, PNG-137A, glaciasterol A and jaspamide in the presence of theED50  concentration of Adriamycin. MCF-7 Adr cells were exposed to serial dilutions of eitherthe geodiamolides, Garvea, PNG-137A, glaciasterol A orjaspamide in the presence of 5 gg/ml Adriamycin. Followingthe drug recovery period the surviving cells were quantified(Figure 11) and a dose response curve for each compound wasconstructed (Figure 26: geodiamolides, Figure 27: Garvea,Figure 28: PNG-137A, Figure 29: glaciasterol A and Figure30: jaspamide). Dose response to 5 pg/m1 Adriamycin in thepresence of no test compound is plotted as 50%. The doseresponse of the cells to the various test agents is plottedrelative to this 50% level. No synergy was observed betweenany of the samples and Adriamycin. The ED50 values of thevarious test agents in the presence of Adriamycin were 0.95,11, >50, 13 and 1 gg/ml for the geodiamolides, Garvea, PNG-137A, glaciasterol A and jaspamide, respectively. A 10-folddecrease in the cytotoxicity of the geodiamolides isobserved. Error bars represent the standard error of themean (standard deviation/"n) of 2 experiments.FIGURE 2687AFIGURE 2787BFIGURE 2887CFIGURE 2987DFIGURE 3087E885. Compounds Screened Against Normal and Malignant Human Breast Epithelial Cells in Primary Culture Table 9 summarizes the cytotoxicities of thegeodiamolides and jaspamide as tested against malignant andnormal HBEC in primary culture. No significant trendindicating tumor specificity was observed for thegeodiamolides. Minimal data is available for the jaspamidesample due to limited quantity of drug. Large variation wasseen in the cytotoxicity for a given agent for differenttissue samples: the ED50 values recorded for thegeodiamolides ranged from 0.018 to 5.5 pg/m1 and from 0.21to > 10 gg/ml in malignant and normal tissue, respectively,whereas the corresponding values determined for jaspamideranged from 0.0005 to 3.8 pg/ml and from < 0.05 to 0.5 pg/mlrespectively. Adriamycin ED50 values were included forreference in cases in which sufficient tissue was available.Table 9. ED50 values obtained for the geodiamolides andjaspamide tested against normal and malignant HBEC inprimary culture.Agent ED50 values (µg/ml)Ni N2 N3 Cal Ca2 Ca3 Ca4 Ca5Geodiamolides 0.21 >10 >6 0.1 0.018 5.5 0.5 NDJaspamide <0.05 0.08 0.5 ND ND ND 0.0005 3.8Adriamycin ND 0.25 0.29 ND ND ND ND NDN, Normal cells; Ca, malignant cells; ND, not done.89CHAPTER 4 DISCUSSIONPast drug-screening programs have illustrated that theincorporation of solid tumors into a drug-screening protocolis essential, since agents that are effective in treatingleukemias are not necessarily effective against solid tumors(21, 62, 95). The crude extracts and pure metabolitesevaluated in this thesis had previously been screened andfound to be cytotoxic in vitro in the leukemic L1210bioassay ( 34, 100, 105, 117 and personal communicationwith Dr. R Andersen). In the present study, mammarycarcinoma was used as a solid-tumor model. Breast cancercell lines were selected for initial screening because thecells are plentiful and easy to grow. However, cell linesresult from cell selection and the populations are veryhomogenous in their characteristics. Therefore, four drug-sensitive and two drug-resistant cell lines were used torepresent the heterogeneity seen in naturally-occuringtumors. Primary cultures of epithelial cells from malignantbreast tissue were incorporated into the screening protocol,because they more closely resemble counterpart cells invivo. Epithelial cells from normal tissue were also includedto determine the tumor specificity of the geodiamolides. Asthe results illustrate, there was broad variation in thebiological activity of cytotoxic compounds tested ondifferent cell lines and on different primary cultures.90Several assays were considered to measure the cytotoxiceffects of the test compounds on the cell lines and the HBECin primary culture. Incorporation of radiolabelledprecursors, protein binding assays, the neutral red assayand the MTT assay were all reviewed. This laboratory hastraditionally used incorporation of radioactive thymidine asa measure of cell viability, but this method was rejectedbecause it is impractical for large-scale screening. Also, apotential problem that may arise when using this method toscreen drugs that have unknown mechanisms of action is thatthe drug under study and the radiolabelled precursor mayshare the same biochemical pathways and/or transportmechanisms, thus there may be competition between the labeland the drug resulting in irregularities in the results(Ruben 1988). Protein binding assays measure the totalprotein content of a population of cells. Consequently, apotential problem with protein binding assays is that theycould incorporate the protein content of nonviable cells inthe final results (136). This is especially so if the timeinterval between drug addition and the assay is short. Theneutral red assay was deemed inadequate, because it relieson the accumulation of neutral red dye into lysosomes and itwas this investigator's opinion that lysosome number wouldbe more susceptible to change when exposing cells to unknownpotentially cytotoxic agents. The MTT assay was selected,because it is a measure of cell viability and it hascorrelated well with the SRB assay (136), a 51Cr release91assay (67), the radiolabelled thymidine incorporation assay(unpublished results from our laboratory), the HTCFA and adye exclusion assay (20).Unfortunately, there is are a large amount of variationin the literature regarding the exact procedure of the MTTassay. The cause of this variation has to do with the factthat in order for the MTT assay to be a reproducible measureof cell survival, the parameters of the assay have to beoptimized for the type of cells used and the cultureconditions used to grow these cells. The parameters that hadto be optimized for the drug screening experiments outlinedin this thesis were the type of solvent, the concentrationof the MTT solution and the length of the MTT incubationtime. The original MTT protocol as described by Mosmann(107) called for the MTT-containing medium not to be removedbefore addition of the solvent. MTT assays performed forthis thesis had the MTT supernatant removed prior to solventaddition, because the results of the two wavelength scans ofcell-generated formazan dissolved in different solvents(Figures 12 and 13) indicate that greater assay sensitivitycan be achieved when the MTT supernatant is removed. Thesewavelength scans also demonstrate that isopropanol is themost suitable solvent for determing formazan absorbance at540 nm. The NCI, when investigating the feasibility of theMTT assay for its large scale drug-screening program,observed that DMSO is the most suitable solvent. Thisdiscrepancy can be explained by the fact that the NCI92cultures most of their cell lines in media supplemented with10% FBS (2) and, at the time of assay, they add their MTTdirectly to the existing drug-containing growth medium(Figure 7). Thus, when the MTT supernatant is removed, thereis a relatively high concentration of residual serumclinging to the cell monolayers and the walls of the wells.High concentrations of residual serum have been shown toenhance the absorption of formazan dissolved in DMSO (2).The MTT assay protocol outlined in this thesis (Figure 11)requires that the growth medium first be removed prior tothe addition of MTT dissolved in serum-free and phenol red-free F12/DME/H. Since there is relatively little residualserum present, formazan dissolved in isopropanol illicitsthe highest peak in the wavelength scan.MTT was added to the cell monolayer dissolved in freshserum-free and phenol red-free medium rather than addeddirectly to the existing growth medium for 2 additionalreasons. The first is that a positive correlation has beenestablished between the D-glucose concentration of theculture medium at the time of assay and the production offormazan (167). Thus, adding MTT to nutrient-depleted growthmedium runs the risk of decreasing the sensitivity of theassay. The relationship between D-glucose availability andformazan production may also account for the non-linearityof the cell number vs A540 curves, since one would expect D-glucose availability to be decreased when cell density ishigh. The second reason is that the pH indicator present in93the normal growth medium increases absorbance background(33), and thus phenol red-free medium is used instead.A possible explanation for the observation that thecoefficient of variation increases proportionally with theMTT incubation time (Figure 16) is that, as the cellsgenerate more and more intracellular formazan crystals, theybegin to lose their attatchment to the microtiter plates.When the MTT supernatant is removed, these cells are lost.To minimize cell loss in all subsequent experiments,formalin was used to fix the cells to the tissue cultureplastic.Of the the 15 compounds tested against the drug-sensitive cell lines MCF-7 and T-47D, 8 demonstratesignificant activity. Agents with an ED50 value of up to 25gg/ml, except for imbricatine, were selected for furthertesting against the ZR-75-1 and MDA-MB-231 cell lines andthe drug-resistant cell lines. There was an insufficientamount of imbricatine available for further testing in thesecell lines. An ED50 of 25 gg/ml was selected as a cutoffvalue because many of the compounds tested were crudeextracts, and a lower cutoff point might have resulted inthe exclusion of potentially toxic compounds from furthertesting. The cutoff value used by other researchers in thedrug-screening field is typically 1 gg/ml. In addition,glaciasterol, which exhibited ED50 values in the 25 gg/mlrange, also exhibited low in vivo host toxicity (personalcommunication with Dr. R Andersen).94The geodiamolides (a mixture of two cyclodepsipeptidesthat are identical except that one is brominated and theother iodinated) and jaspamide are very promising compounds,since they exhibit cytotoxicity comparable with that ofAdriamycin when tested against the drug-sensitive cell linesand are several orders of magnitude more cytotoxic thanAdriamycin when screened against the MDR cell lines.Jaspamide, when screened against the NCI's human tumor cellline panel, also exhibits selective cytotoxicity againstrenal and rectal cancer cell lines (personal communicationwith Dr. R Andersen). This selective activity suggests thatjaspamide may have a novel mode of action. Adriamycin wasselected as a standard by which the test samples could becompared because it is commonly used, in conjunction withcyclophosphamide and 5-fluorouracil, in the treatment ofbreast cancer.The observation that compounds such as thegeodiamolides are much less cytotoxic against MCF-7 Adrcells than the parent line MCF-7 whereas they were morecytotoxic against MDA-Alr cells than the parent line MDA-MB-231 (Tables 7 and 8) suggests either that cellularmechanisms in the MDA-Alr cells may enable the drugs toovercome the effects of Pgp overexpression or that thesedrugs are simply not affected by this efflux pump. Thepresence of Pgp in these cell lines is known since otherresearchers in this laboratory have demonstrated usingimmunohistochemical techniques that the MCF-7 Adr and MDA-95Air cell lines are Pgp-positive, whereas the MCF-7 cell lineis Pgp-negative. The Pgp staining status of the MDA-MB-231is currently being determined in this laboratory.It is observed for the Garvea extract, thegeodiamolides, PNG-137A, glaciasterol and jaspamide that theED50 values generated for either the MCF-7 or MDA-MB-231cells were not much different than those obtained using thedrug-resistant daughter lines. This suggests that eitherthese agents, in addition to being cytotoxic, may circumventthe MDR phenotype, or they may not be affected by thepresence of the Pgp pump. Glaciasterol is particularilyinteresting since its' molecular structure is analogous tothat of steroid hormones. Steroid hormones, particularilyprogesterone and the antiestrogen tamoxifen, have been shownto interfere with Pgp-mediated efflux of anticancer drugs(50, 70, 108). To gain an understanding of the relationshipbetween these 5 samples and the Pgp pump, serial dilutionsof these agents were exposed to the MCF-7 Adr cells in thepresence of Adriamycin. Adriamycin is a known substrate forthe Pgp pump. The failure to observe cytotoxic synergybetween the samples and Adriamycin suggests that these drugsdo not interfere with Pgp function.In retrospect, with regard to these particularexperiments, it may have been a more suitable experimentaldesign if varying concentrations of Adriamycin and the ED50concentration of our test agent were incubated with the MDRcells rather than varying concentrations of our test agent96and the ED50 concentration of Adriamycin. The reason forthis is that one wants to see if the test agent can modulatethe activity of Adriamycin (ie: decrease ED50), rather thanvice-versa. However, the experiments actually performed arenot invalid, since ultimately the ED50 concentrations ofboth Adriamycin and the test agent are exposed to the sameMDR cells. The in vitro tumor model used in theseexperiments resembles a Skipper-Schabel model of tumorgrowth much more than a Gompertzian model, since the cellsare growing at a low cell density (subconfluent) and thereis easy access to 02 and nutrients. Consequently one wouldexpect Skipper's constant-cell-kill law to prevail. If the 2agents in the test system worked independently from oneanother, then exposure of the MDR cells to the ED50concentration of both agents should result in 25% of thecells surviving treatment. This observation appears to holdtrue for all the samples (Figures 27 - 30) tested except thegeodiamolides (Figure 26). If the agents worksynergistically with one another (ie: the test agentinterfers with Pgp function), then one would expect >75%cell kill. Conversely, if the 2 agents interfere with oneanother then decreased cell kill (<75%) would be observed.The geodiamolides appear to fit into this category (Figure26).It should be stressed that although the MTT assay hasbeen shown to be a reproducible and accurate assay forgenerating dose response curves, there was great difficulty97in accurately weighing the compounds and maintaining theappropriate drug concentrations. For some of the puresamples, only a very small amount (- 1 mg) of drug wasavailable for testing. To keep drug vehicle (ethanol andsaline) concentrations to a minimum, the drugs had to bedissolved in a minimal volume (- 100 41) of vehicle.Consequently it was difficult to maintain a constant drugconcentration when working with such a small volume ofvolatile organic solvent, especially when repeated openingof the drug-containing vials was required.Previous experiments in this laboratory havedemonstrated that cells from the same tumor cultured indifferent media vary in their responses to chemotherapeuticagents (43). These previous experiments also demonstratedthat medium supplemented with the patient's serum promotesthe attachment and growth of HBEC. Since medium containingthe patient's serum more closely simulates the in vivoenvironment than defined medium and medium supplemented withFBS, the most appropriate nutrient environment fordetermining the effects of therapeutic agents should includeserum of the patient from whom the tumor is removed (43). Awide range of chemosensitivities for the geodiamolides isobserved when tested against normal and malignant breasttissue taken from different patients, although there appearsto be no significant trend indicating tumor specificity.Several notes of caution should be mentioned regarding theinterpretation of these results obtained from HBEC in98primary culture. Naturally occuring breast tumors in vivofollow Gompertzian growth kinetics (109). This is likely dueto the fact that they are very heterogenous, being composedof both drug-resistant and drug-responsive cells. All themalignant tissue samples used in these experiments were frombiopsies and mastectomies. Furthermore, the tissue samplereceived from the hospital is likely composed of normal aswell as malignant cells. Another confounding factor in theinterpretation of these results is that the growth mediaused to culture these cells also promotes the growth offibroblasts (44). The presence of fibroblasts in the cultureprobably helps mirror the in vivo environment since at leastsome epithelial stromal interactions are maintained.However, in the in vivo environment the malignant epithelialcells are much more mitotically active than the stromalcells, whereas in the in vitro condition both populationsare highly mitotic. Chemotherapeutic drugs are generallymore effective on mitotic cells than on quiescent cells(163) and thus the presence of mitotically activefibroblasts could influence the interpretation of an invitro assay. However, previous studies in this laboratoryhave demonstrated that the HBEC culturing methods outlinedin this thesis select for - 85% epithelial cells at the timeof cell seeding (44). It was also observed in theseexperiments that fibroblast overgrowth occurs 7 - 10 daysafter seeding. The drug screening protocol outlined in thisthesis requires cultures to grow for only 5 days. A defined99medium which inhibits the growth of stromal cells could beused instead of the patient's serum for the drug sensitivitytesting, but this has previously been shown to alter theresponse of HBEC in primary culture to cytotoxic agents(43).The geodiamolides, jaspamide and the five crudeextracts PNG-31A, Garvea, BC-3, QCI-32 and PNG-137-A warrantfurther investigation, since they demonstrate significantcytotoxicity in the preliminary cell-line screens. The fullpotential of the crude extracts will not be realized untilthe pure active metabolites have been isolated. The nextstep in the identification of these agents as cytotoxicwould be additional testing against other in vitro humantumor models representing a variety of tumor types, followedby in vivo testing of the tumor types that demonstrateactivity in the in vitro screens. This laboratory has amouse colony for maintaining the transplantable Shionogimammary carcinoma (45). In conjunction with these in vitroand in vivo screens, future experiments will also includemechanism based testing to determine modes of action ofthese agents.Drug resistance and host toxicity, as outlined by theGoldie-Coldman model and the Norton-Simon hypothesis, arethe reasons for chemotherapy failure. As the structures ofthe pure metabolites illustrate, marine organisms representa vast reservoir of new and interesting chemical familiesthat potentially provide novel modes of action, better100therapeutic indices and fewer side effects thanchemotherapeutic agents in clinical use against currentlyuncurable solid tumors.101REFERENCES1. Albertini R.J., Castle K.L., and Borcherding W.R. (1982)T-cell cloning to detect the mutant 6-thioguanine-resistantlymphocytes present in human peripheral blood. Proc. Natl.Acad. Sci. U.S.A. 79: 6617-6621.2. Alley M.C., Scudiero D.A., Monks A., Hursey M.L.,Czerwinski M.J., Fine D.L., Abbott B.J., Mayo J.G.,Shoemaker R. and Boyd M.R. (1988) Feasibility of drugscreening with panels of human tumor cell lines using amicroculture tetrazolium assay. Cancer Res. 48: 589-601.3. Alley M.C., Powis G., Appel P.L., Kooistra K.L. andLieber M.M. (1984) Activation and inactivatio of cancerchemotherapeutic agents by rat hepatocytes cocultured withhuman tumor cell lines. Cancer Res. 44: 549-556.4. Baguley B.C. and Wilson W.R. (1987) Comparison of in vivoand in vitro drug sensitivities of Lewis lung carcinoma andP388 leukemia to analogues of amsacrine. Eur. J. CancerClin. Oncol. 23: 607-613.5. Bak M., Efferth T., Mickisch G., Mattern J. and Volm M.(1990) Detection of drug resistance and P-glycoprotein inhuman renal cell carcinomas. Eur. Urol. 17: 7275.6. Beck W.T. (1983) Vinca alkaloid-resistant phenotype incultured human leukemic lymphoblasts. Cancer Treat. Rep. 67:875-882.7. Beck W.T. (1991) Do anti-P-glycoprotein antibodies have afuture in the circumvention of multidrug resistance? J.Natl. Cancer Inst. (U.S.A.) 83: 1364-1366.8. Benson A.B., Trump D.L., Koeller J.M. and 5 others (1985)Phase I study of vinblastine and verapamil given byconcurrent iv infusion. Cancer Treat. Rep. 69: 795-799.9. Bergmann W. and Feeney R.J. (1951) Contributions to thestudy of marine products. The nucleosides of sponges. J.Org. Chem. 16: 981.10. Biedler J.L., Chang T., Meyers M.B., Peterson R.H.F. andSpengler B.A. (1983). Drug resistance in Chinese hamsterlung and mouse tumor cells. Cancer Treat. Rep. 67: 859-867.11. Blume E. (1989) Investigators seek to increase taxolsupply J. Natl. Cancer Inst. (U.S.A.) 81: 1122-1123.12. Borenfreund E. and Puerner J.A. (1984) Toxicitydetermined in vitro by morphological alterations and neutralred absorption. Tox. Lett. 24: 119-124.10213. Borenfreund E. and Puerner J.A. (1987) Short-termquantitative in vitro cytotoxicity assay involving an s-9activating system. Cancer Lett. 34: 243-248.14. Boyd M.R. (1989) Status of the NCI preclinical antitumordrug discovery screen. In: Cancer: Principles and practiceof oncology update (DeVita V.T., Hellman S. and RosenbergS.A., eds), vol 3, Lippincott, Philadelphia. pp 1-12.15. Bradley G., Naik M. and Ling V. (1989) P-Glycoproteinexpression in multidrug-resistant human ovarian carcinomacell lines. Cancer Res. 49: 2790-2796.16. Broxterman H.J., Kuiper C.M., Schuurhuis G.J., TsuruoT., Pinedo H.M. and Lankelma J. (1988) Increase ofduanorubicin and vincristine accumulation in multidrugresistant human ovarian carcinoma cells by a monoclonalantibody reacting with P-glycoprotein. Biochem. Pharm. 37:2389-2393.17. Burgoyne D.L., Miao S., Pathirana C., Andersen R.J.,Ayer W.A., Singer P.P., Kokke W.C.M.C. and Ross D.M. (1991)The structure and partial synthesis of imbricatine, abenzyltetrahydroisoquinoline alkaloid from the starfishDermasterias Imbricata. Can. J. Chem. 69: 20-27.18. Canellos G.P., De Vita V.T., Gold G.L., Chabner B.A.,Schein P.S. and Young R.C. (1976) Combination chemotherapyfor advanced breast cancer: respnse and effect on survival.Ann. Inter. Med. 84: 389-392.19. Cano-Gauci D.F. and Riordan J.R. (1987) Action ofcalcium antagonists on multidrug resistant cells. Specificcytotoxicity independent of increased cancer drugaccumulation. Biochem. Pharm. 36: 2115-2123.20. Carmichael J., DeGraff W.G., Gazdar A.F., Minna J.D. andMitchell J.B. (1987) Evaluation of a tetrazolium-basedsemiautomated colorimetric assay: Assessment ofchemosensitivity testing. Cancer Res. 47: 936-942.21. Chabner B.A. (1990) In defence of cell line screening.J. Natl. Cancer Inst. (U.S.A.) 82: 1083-1085.22. Chen C.-J., Chin J.E., Ueda K., Clark D.P., Pastan I.,Gottesman M.M. and Roninson I.B. (1986) Internalduplication and homology with bacterial transport proteinsin the mdr-1 (P-glycoprotein) gene from multidrug-resistanthuman cells. Cell 47: 381-389.23. Christian M.C., Wittes R.E., Leyland-Jones B., McLemoreT.L., Smith A.C., Grieshaber C.K., Chabner B.A. and BoydM.R. (1989). 4-Ipomenaol: A novel investigational new drug103for lung cancer. J. Natl. Cancer Inst. (U.S.A.) 81: 1333-1143.24. Clark R., Lippman M.E. and Dickson R.B. (1990)Mechanisms of hormone and cytotoxic drug interactions in thedevelopment and treatment of breast cancer. In: Molecularendocrinology and steroid hormone action (Sato G.H. andStevens J.L., eds). Alan R. Liss Inc., New York. pp 243-278.25. Collins K.D. and Stark G.R. (1971) Aspartatetranscarbamylase: interaction with the transition stateanalogue N-(phosphonacetyl)-L-aspartate. J. Biol. Chem. 246:6599-6605.26. Conley B.A. and Van Echo D.A. (1992) Antineoplatic drugdevelopment. In: The chemotherapy source book (Perry M.C.ed.), Williams and Wilkins, Baltimore, Maryland. pp 15-21.27. Cooper R. (1969) Combination chemotherapy in hormoneresistant breast cancer. Proc. Amer. Assoc. Cancer Res. 10:15.28. Cornwell M.M., Tsuruo T., Gottesman M.M. and Pastan I.(1987) ATP-binding properties of P-glycoprotein frommultidrug-resistant KB cells. FASEB J. 1: 51-54.29. Crews P., Manes L.V. and Boehler M. (1986)Jaspalkilnotide, a cyclodepsipeptide from the marine spongeJaspis sp. Tetrahedron Lett. 31: 2797-2798.30. Dalton W.S. and Grogan T.M. (1991) Does P-glycoproteinpredict response to chemotherapy, and if so, is there areliable way to detect it? J. Natl. Cancer Inst. (U.S.A.)83: 80-81.31. DeFazio A., Musgrove E.A., and Tattersall M.H. (1988)Flow cytometric enumeration of drug-resistant tumor cells.Cancer Res. 48: 6037-6043.32. Dempsey J.L., Seshadri R.S., and Morley A.A. (1985)Increased mutation frequency following treatment with cancerchemotherapy. Cancer Res. 45: 2873-2877.33. Denizot F. and Lang R. (1986) Rapid colorimetric assayfor cell growth and survival. Modifications to thetetrazolium dye procedure giving improved sensitivity andreliability. J. Immun. Meth. 89: 271-277.34. deSilva E.D., Andersen R.J. and Allen T.M. (1990)Geodiamolides C to F, new cytotoxic cyclodepsipeptides fromthe marine sponge Pseudaxinyssa sp. Tetrahedron Lett. 31:489-492.35. Dietel M. (1991) What's new in cytostatic drug104resistance and pathology. Path. Res. Pract. 187: 892-905.36. Dolnick B.J., Berenson R.J., Bertino J.R., Kaufman R.J.,Nunberg J.H. and Schimke R.T. (1979) Correlation ofkehydrofolate reductase elevation with gene amplification ina homogeneously staining chromosomal region in L5178Y cells.J. Cell. Biol. 83: 399-402.37. Double J.A. and Bibby M.C. (1989) Therapeutic index: Avital component in selection of anticancer agents forclinical trial. J. Natl. Cancer Inst. (U.S.A.) 81: 988-994.38. Driscoll J.S. (1984) The preclinical new drug researchprogram of the National Cancer Institute. Cancer Treat. Rep.68: 63-76.39. Duerst R.E. and Frantz C.N. (1985) A sensitive assay ofcytotxicity applicable to mixed cell populations. J. Immun.Meth. 82: 39-46.40. Dumdei E.J., deSilva E.D., Andersen R.J., Iqbal ChoudryM. and Clardy J. (1989) Chromodorolide A, a rearrangedditerpene with a new carbon skeleton from the Indian Oceannudibranch Chromodoris cavae. J. Amer. Chem. Soc. 111: 2712-2713.41. Elmore E., Kakunaga T., and Barrrett J.C. (1983)Comparison of spontaneous mutation rates of normal andchemically transformed human skin fibroblasts. Cancer Res.43: 1650-1655.42. Emerman J.T., Tolcher A.W. and Rebbeck P.M. (1990) Invitro sensitivity testing of human breast cancer cells tohormones and chemotherapeutic agents. Cancer Chemother.Pharm. 26: 245-249.43. Emerman J.T., Fiedler E.E., Tolcher A.W. and RebbeckP.M. (1987) Effects of defined medium, fetal bovine serum,and human serum on growth and chemosensitivities of humanbreast cancer cells in primary culture: inference for invitro assays. In Vitro Cell. Devel. Biol. 23: 134-140.44. Emerman J.T. and Wilkinson D.A. (1990) Routine culturingof normal, dysplastic, and malignant human mammaryepithelial cells from small tissue samples. In Vitro Cell.Devel. Biol. 26: 1186-1194.45. Emerman J.T. and Siemiatkowski J. (1984) Effects ofendocrine regulation of growth of a mouse mammary tumor ontis sensitivity to chemotherapy. Cancer Res. 44: 1327-1332.46. Fahy E. and Andersen R.J. (1987) Minor metabolites ofthe marine hydroid Garvea annulate. Can. J. Chem. 65: 376-383.10547. Fairchild C.R., Ivy S.P., Kap-Shan C.S., Whang-Peng J.,Rosen N., Isreal M.A., Melea P.W., Cowan C.H. and GoldsmithM.E. (1987) Isolation of amplified and overexpressed DNAsequences from Adriamycin-resistant human breast cancercells. Cancer Res. 47:5141-5148.48. Farber S., Diamond L.K., Mercer R.D. et al. (1948)Temporary remissions in acute leukemia in children producedby folic acid antagonist, 4-aminopteroyl-glutamic acid(aminopterin). N. Engl. J. Med. 238: 787-793.49. Filman R.J., Brawn R.J. and Dandliker W.B. (1975)Intracellular supravital stain delocalization as an assayfor antibody-dependent complement-mediated cell damage. J.Immun. Meth. 6: 189-207.50. Fine R.L., Sachs C.W., Albers M. and Williams A. (1991)Tamoxifen potentiates the cytotoxicity of vinblastine byincreasing intracellular drug accumulation. Proc. Amer.Assoc. Cancer Res. 32: 375.51. Finlay G.J. and Bagulaey B.C. (1984) The use of humancancer cell lines as a primary screening system forantineoplastic compounds. Eur. J. Cancer Clin. Oncol. 20:947-954.52. Finlay G.J., Wilson W.R. and Baguley B.C. (1986)Comparison of in vitro activity of cytotoxic drugs towardshuman carcinoma and leukemia cell lines. Eur. J. Cancer 22:655-662.53. Fisher B., Redmon C., Legault-Poisson S. et al. (1990)Postoperative chemotherapy and tamoxifen compared withtamoxifen alone in the treatment of positive-node breastcancer patients with tumors responsive to tamoxifen: resultsfrom NSABP B-16. J. Clin. Oncol. 8: 1005-1018.54. Fojo A.T., Ueda K., Slamon D.J., Poplack D.G., GottesmanM.M. and Pastan I. (1987) Expression of a multidrug-resistance gene in human tumors and tissues. Proc. Natl.Acad. Sci. U.S.A. 84: 265-269.55. Frie E., III, and Canellos G.P. (1980) Dose: a criticalfactor in cancer chemotherapy. Am. J. Med. 69: 585-594.56. Furth J., Kahn M.C. (1937) The transmission of leukemiaof mice with a single cell. Am. J. Cancer. 31: 276-282.57. Gianni A.M., Siena S., Bregni M., Tarella C., SternA.C., Pileri A. and Bonadonna G. (1989). Granulocyte-macrophage colony-stimulating factor to harvest circulatinghaematopoietic stem cells for autotransplantation. Lancet 2:580-585.10658. Goldie J.H. and Coldman A.J. (1979) A mathematical modelfor relating the drug sensitivity of tumors to theirspontaneous mutation rates. Cancer Treat. Rep. 63: 1727-1733.59. Goldstein L.J., Galski H., Fojo A., Willingham M. and 10others (1989) Expression of a multidrug resistance gene inhuman cancers. J. Natl. Cancer Inst. (U.S.A.) 81: 116-124.60. Goodman L.S., Wintrobe M.M., Dameshek W., et al. (1946)Use of methyl-bis(beta-chloroethyl)amine hydrochloride forHodkin's disease, lymphomsarcoma, leukemia. JAMA 132: 126.61. Gorelik E., Alley M. and Shoemaker R. (1986) A new invivo short-term assay for evaluation of antitumorchemotherapeutic drugs. Proc. Amer. Assoc. Cancer Res. 27:389.62. Grindey G.B. (1990) Current status of cancer drugdevelopment: Failure or limited success? Cancer Cells 2:163-171.63. Hamada H. and Tsuruo T. (1988) Characterization of theATPase activity of the Mr 170,000 to 180,000 membraneglycoprotein (P-glycoprotein) associated with multidrugresistance in K562/ADM cells. Cancer Res. 48: 4926-4932.64. Hamburger A.W. and S.E. Salmon (1977) Primary bioassayof human tumor stem cells. Science 197: 461-463.65. Hegewisch-Becker S., Fliegner M., Hossfeld D. and ZanderA. (1991) Clincial relevance of P-glycoprotein expressionand drug uptake. Proc. Amer. Assoc. Cancer. Res. 32: 434.66. Heidelberger C., Chaudhuari N.K., Danenberg P. et al.(1957) Fluorinated pyrimidines: a new class of tumorinhibitory compounds. Nature 179: 663-666.67. Heo D.S., Park J.-G., Hata K., Day R., Herberman R.B.and Whiteside T.L. (1990) Evaluation of tetrazolium-basedsemiautomatic colorimetric assay for measurement of humanantitumor cytotoxicity. Cancer Res. 50: 3681-3690.68. Heuser L., Spratt J. and Polk H. (1979) Growth rates ofprimary breast cancer. Cancer 43: 1888-1894.69. Hitchins R.N., Harman D.H., Davey R.A. and Bell D.R.(1988) Identification of a multidrug resistance associatedantigen (P-glycoprotein) in normal human tissues. Eur. J.Cancer Clin. Oncol. 24: 449-454.70. Huang Yang C.-P., Goei DePinho S., Greenberger L.M.,Arceci R.J. and Band Horwitz S. (1989) Progesterone107interacts with P-glycoprotein in multidrug-resistant cellsand in the endometrium of gravid usterus. J. Biol. Chem.264: 782-788.71. Irwin L.E., Pugh R., Sadoff L., Hestorff R., Weiner J.M.and Bateman J.R. (1975) The influence on survival ofcontinuous combination vs sequential single agentchemotherapy in disseminated breast cancer. Presented at theAmerican College of Physicians 56th annual meeting, SanFrancisco.72. Jacobson L.P., Spurr C.L., Barron E.S.Q., et al. (1946)Studies on the effect of methyl-bis(beta-chloroethyl)aminehydrochloride for neoplastic deseases and allied disordersof the hemapoitic system. JAMA 132: 263.73. Johnson R.K. (1990) Screening methods in antineoplasticdrug discovery. J. Natl Cancer Inst. (U.S.A.) 82: 1082-1083.74. Jones C.A., Tsukamoto T., O'Brien P.C., Uhl C.B. andAlley M.C. (1985) Soft agarose culture human 'Llmour colonyforming assay for drug sensitivity testing: [ H]-thymidineincorporation vs colony counting. Br. J. Cancer 52: 303-310.75. Jordan V.C. and Murphy C.S. (1990) Endocrinepharmacology of antiestrogens as antitumor agents. EndocrineRev. 11: 578-601.76. Juranka P.F., Zastawny R.L. and Ling V. (1989) P-glycoprotein: Multidrug-resistance and a superfamily ofmembrane-associated transport proteins. FASEB J. 3: 2583-2592.77. Kardinal C.G. (1992) Chemotherapy of breast cancer. In:The chemotherapy source book (Perry M.C., ed), Williams andWilkins, Baltimore, Maryland. pp 948-988.78. Kardinal C.G., Perry M.C., Korzun A.H., Rice M.A.,Ginsberg S. and Wood W. (1988) Responses to chemotherapy ofchemohormonal therapy in advanced breast cancer patientstreated previously with adjuvant chemothrapy. Cancer 61:415-419.79. Kardinal C.G. (1979) Chemotherapy. In: Cancer of thebreast, 2nd edition (Donegan W.L. and Spratt J.S., eds),W.B. Saunders, Philadelphia. pp 405-447.80. Kaufman R.J., Brown P.C., and Schimke R.T. (1979)Amplified dihydrofolate reductase genes in unstablymethotrexate-resistant cells are associated with doubleminute chromosomes. Proc. Natl. Acad. Sci. U.S.A. 76: 5669-5673.10881. Keith W.N., Stallard S. and Brown R. (1990) Expressionof mdrl and gst-D in human breast tumours: Comparison to invitro chemosensitivity. Br. J. Cancer 61: 712-716.82. Kennedy J.M., Beveridge R.A., Rowley S.D., Gordon G.B.,Abeloff M.D. and Davidson N.E. (1991) High-dose chemotherapywith reinfusion of purged autologous bone marrow followingdos-intense induction as initial therapy for metastaticbreast cancer. J. Natl. Cancer Inst (U.S.A.) 83: 920-926.83. Kerkvliet G.J. (1990) Scientists search rain forests fornovel chemicals. J. Natl. Cancer Inst. (U.S.A.) 82; 1000-1001.84. Kidd J.G. (1953) Regression of transplanted lymphomasinduced in vivo by means of normal guinea pig serum: I.Course of transplanted cancers of various kinds in mice andrats given guinea pig serm, horse serum, or rabbit serum. J.Exp. Med. 98: 565-582.85. Klohs W.D., Steinkampf R.W., Havlick M.J. et al. (1986)Resistance to anthracyclines in multidrug-resistant P388murine leukemia cells: Reversal by calcium blockers andcalmodulin antagonists. Cancer Res. 46: 4352-4356.86. Kovach J.S., Ames M.M., Powis G. et al. (1980) Use of aliver microsome system in testing drug sensitivity of tumorcells in soft agar. Proc. Amer. Assoc. Cancer Res. 21: 257.87. Laird A.K. (1969) Dynamics of tumor growth in tumors andnormal organisms. Natl. Cancer Inst. Mono. 30: 15-28.88. Law L.W. (1952) Origin of the resistance of leukemiccells to folic acid antagonists. Nature 169: 628-629.89. Lee W.W., Benitez A., Goodman L. and Baker B.R. (1960)Potential anticancer agents. Synthesis of the B-anomer of 9-(D-arabinofuranosyl)-adenine. J. Am. Chem. Soc. 82: 2648.90. Licht T.H., Fiebig H.H., Bross K.J., Berger D.P., DreherC. and Shoemaker R. (1991) Induction of multidrug-resistanceduring antineoplastic chemotherapy. Proc. Amer. Assoc.Cancer Res. 32: 367.91. Lillie R.D. (1977) In: H.J. Conn's biological stains,9th ed. Williams and Wilkins, Baltimore, Maryland.92. Ling V., Kartner N., Sudo T., Siminovicth L and RiordanJ.R. (1983) Multidrug-resistance phenotype in Chinesehamster ovary cells. Cancer Treat. Rep. 67: 869-874.93. Ling V. (1989) Does P-glycoprotein predict response tochemotherapy? J. Natl. Cancer Inst. (U.S.A.) 81: 84-85.10994. Marsoni S. and Wittes R.E. (1984) Clinical developmentof anticancer agents - A National Cancer InstitutePerspective. Cancer Treat. Rep. 68: 77-85.95. Marsoni S., Hoth D., Simon R., Leyland-Jones B., De RosaM. and Wittes R.E. (1987) Clinical drug development: Ananalysis of phase II trials, 1970-1985. Cancer Treat. Rep.71: 71-80.96. Martin W.M.C. and McNally N.J. (1980) The cytotoxiceffects of Adriamycin on tumour cells in vitro and in vivo.Br. J. Cancer 41 (suppl IV): 306.97. Mendelsohn M.L. (9160) The growth fraction: A newconcept applied to tumors. Science 132: 1496.98. Merry S., Flanigan P., Schlick E., Freshney R.I. andKaye S.B. (1989) Inherent adriamycin resistance in a murinetumour line: Circumvention with verapamil and norverapamil.Br. J. Cancer 59: 895-897.99. Miao S. (1991) Novel secondary metabolites from selectedmarine invertebrates. PhD Thesis, University of BritishColumbia.100. Miao S., Andersen R.J. and Allen T.M. (1990) Cytotoxicmetabolites from the sponge lanthella Basta collected inPapua New Guinea. J. Naturall Prod. 53: 1441-1446.101. Mickisch G.H., Kossig J., Keilhauer G.K., Schlick E.,Tschada R.K. and Alken P.M. (1990) Effects of calciumantagonists in multidrug resistant primary human renal cellcarcinomas. Cancer Res. 50: 3670-3674.102. Mickisch G.H., Merlino G.T., Galski H., Gottesman M.M.and Pastan I. (1991) Transgenic mice that express the humanmultidrug-resistance gene in bone marrow enable a rapididentification of agents that reverse multidrug resistance.Proc. Natl. Acad. Sci. U.S.A. 88: 547-551.103. Monks A., Scudiero D., Skehan P., Shoemaker R. and 10others (1991) Feasibility of a high-flux anticancer drugscreen using a diverse panel of cultured human tumor celllines. J. Natl. Cancer Inst. 83: 757-766.104. Moore R.E. and Scheuer P.J. (1971) Palytoxin: A newtoxin from a coelenterate. Science 172: 494-498.105. Morris S.A. and Andersen R.J. (1989) Nitrogenousmetabolites fro the deep water sponge Hexadella sp. Can. J.Chem. 67: 677-681.106. Morrison V.A. and Luikart S.D. (1992) Chemotherapy oflung cancer. In: The chemotherapy source book (Perry M.C.,110ed), Williams and Wilkins, Baltimore, Maryland. pp 932-947.107. Mosmann T. (1983) Rapid colorimetric assay for cellulargrowth and survival: Application to proliferation andcytotoxicity assays. J. Immun. Meth. 65: 55-63.108. Naito M., Yusa K. and Tsuruo T. (1989) Steroid homonesinhibit binding of Vinca alkaloids to multidrug resistancerelated P-glycoprotein. Biochem. Biophys. Res. Comm. 158:1066-1071.109. Norton L. (1988) A Gompertzian model of human breastcancer. Cancer Res. 48: 7067-7071.110. Norton L. (1992) The Norton-Simon hypothesis. In: Thechemotherapy source book (Perry M.C., ed), Williams andWilkins, Baltimore. pp 36-53.111. Norton L. and Simon R. (1986) The Norton-Simonhypothesis revisited. Cancer Treat. Rep. 70: 163-169.112. Oliver M.H., Harrison N.K., Bishop J.E., Cole P.J. andLaurent G.J. (1989) A rapid and convenient assay forcounting cells cultured in microwell plates: application forassessment of growth factors. J. Cell Science 92: 513-518.113. Otis P.T. and Armentrout S.A. (1975) Combinationchemotherapy in metastatic carcinoma of the breast: resultswith a three-drug combination. Cancer 36: 311-317.114. Ovejera A.A. and Houchens D.P. (1981) Human tumorxenografts in athymic nude mice predlinical screen foranticancer agents. Semin. Oncol. 8: 386-393.115. Ozawa S., Yasuda T. and Inaba M. (1988) Comparison ofcellular basis of drug sensitivity of human colon,pancreatic, and renal carcinoma cell lines with that ofleukemia cell lines. Cancer Chemother. Pharm. 22: 41-46.116. Padmanabhan N., Howell A., Rubens R.D. (1986) Mechanismof action of adjuvant chemotherapy for breast cancer. Lancetii: 411-414.117. Pathirana C. and Andersen R.J. (1986) Imbricatine, anunusual benzyltetrahydroisoquinoline alkaloid isolated fromthe starfish Dermasterias imbricata. J. Am. Chem. Soc. 108:8228-8229.118. Paull K.D., Shoemaker R.H., Hodes L., Monks A.,Scudiero D.A., Rubinstein L., Plowman J. and Boyd M.R.(1989) Display and analysis of patterns of differentialactivity of drugs against human tumor cell lines:Development of mean graph and COMPARE algorithm. J. Natl.Cancer Inst. (U.S.A.) 81: 1088-1092.111119. Perry M.C., Kardinal C.G., Korzun A.H. et al. (1987)Chemohormonal theray in advanced carcinoma of the breast:Cancer and Leukemia Group B protocol 8081. J. Clin. Oncol.5: 1534-1545.120. Petit G.R., Herald C.L., Doubek D.L., Herald D.L.,Arnold E. and Clardy J. (1982) Isolation and structure ofbryostatin 1. J. Am. Chem. Soc. 104: 6846-6848.121. Peto R. Data presented at King's Fund Forum onTreatment of Primary Breast Cancer. London, 1 October, 1986.122. Phillips R.M., Bibby M.C. and Double J.A. (1990) Acritical appraisal of the predictive value of in vitrochemosensitivity assays. J. Natl. Cancer Inst. (U.S.A.) 82:1457-1468.123. Pika J., Tischler M. and Andersen R.J. (1992)Glaciasterols A and B, 9, 11-secosteroids from the marinesponge Aplysilla glacialis. Can. J. Chem. 70: 1506-1510.124. Price P. and T.J. McMillan (1990) Use of thetetrazolium assay in measuring the response of human tumorcells to ionizing radiation. Cancer Res. 50: 1392-1396.125. Ragaz J. and Ariel I.M. (1989) Diagnostic andtherapeutic highlights of high risk breast cancer-Part one.Comments related to surgery, radiotherapy and hormonaltherapy. In: High risk breast cancer-therapy. (Ragaz J. andAriel I.M., eds), Springer-Verlag, Heidelberg. pp 17-39.126. Ramachandran C., Wellham L., Sridhar K.S. and KrishanA. (1991) Mdr-1 gene, P-glycoprotein, and doxorubicincytotoxicity in human lung tumor cell lines. Proc. Amer.Assoc. Cancer Res. 32: 365.127. Reynolds T. (1991) Tropical rain forest conservationtied to drug development. J. Natl. Cancer Inst. (U.S.A.) 83;594-596.128. Rhoads C.P. (1946) Nitrogen mustards in treatment ofneoplastic disease. JAMA 131: 656.129. Richards W.L., Song M.-K., Krutzsch H., Evarts R.P.,Marsden E. and Thorgeirsson S.S. (1985) Measurement of cellproliferation in microculture using Hoechst 33342 for therapid semiautomated microfluorimetric determination ofchromatin DNA. Exp. Cell. Res. 159: 235-246.130. Rinehart K.L., Gloer J.B., Cook J.C., Mizsak S.A. andScahill T.A. (1981) Structures of the didemnins, antiviraland cytotoxic depsipeptides from a Caribbean tunicate. J.Am. Chem. Soc. 103: 1857-1859.112131. Ro J., Sahin A., Ro J.Y., Fritsche H., Hortobagyi G.and Blick M. (1990) Immunohistochemical analysis of P-glycoprotein expression correlated with chemotherapyresistance in locally advanced breast cancer. Hum. Path. 21:787-791.132. Rosenberg B. (1985) Fundamental studies with cisplatin.Cancer 55: 2303-2316.133. Ross D.D., Joneckis C.C., Wu R., Hambuerger A., CondonM., Ordonez J.V. and Sisk A. (1987) Estimation of leukemiccell kill by flow cytometric quantification of fluoresceindiacetate (FDA) viable cell number. Proc. Amer. Assoc.Cancer Res. 28: 428.134. Ruben R.L. and Neubauer R.H. (1987) Semiautomatedcolorimetric assay for in vitro screening of anticancercompounds. Cancer Treat. Rep. 71: 1141-1149.135. Ruben R.L. (1988) Cell culture for testing anticancercompounds. In: Advances in cell culture, vol 6 (MaramoroschK., ed), Academic Press, Inc., New York. pp 161-197.136. Rubinstein L.V., Shoemaker R.H., Paull K.D., SimonR.M., Tosini S., Skehan P., Scudiero D.A., Monks A. and BoydM.R. (1990) Comparison of in vitro anticancer-drug-screeningdata generated with a tetrazolium assay versus a proteinassay against a diverse panel of human tumor cell lines. J.Natl. Cancer Inst. (U.S.A.) 82: 1113-1118.137. Safa A.R., Glover C.J., Sewell J.L., Meyers M.B.,Biedler J.L. and Felsted R.L. (1987) Identification of themultidrug resistance-related membrane glycoprotein as anacceptor for calcium channel blockers. J. Biol. Chem. 262:7884-7888.138. Salmon S.E. (1989) Chemosensitivity testing: Anotherchapter. J. Natl. Cancer Inst. (U.S.A.) 82: 82-83.139. Salmon S.E., Hamburger A.W., Soehnlen B., Durie B.G.M.,Alberts D.S. and Moon T.E. (1978) Quantitation ofdifferential sensitivity of human-tumor stem cells toanticancer drugs. New Engl. J. Med. 298: 1321-1327.140. Schneider S.L., Fuqua S.A.W., Speeg C.V., Tandon A. andMcGuire W. (1990) Isolation and characterization of anAdriamycin-resistant breast tumor cell line. In Vitro Cell.& Devel. Biol. 26: 621-628.141. Scudiero D., Shoemaker R., Paull K., Alley M., MonksA., Fine D. and Boyd M. (1987) A new tetrazolium reagent fora simplified growth and drug sensitivity assay of humantumor cell lines. Proc. Amer. Assoc. Cancer Res. 28: 421.113142. Selby P., Buick R.N. and Tannock I. (1983) A criticalappraisal of the "human tumor stem-cell assay". New Engl. J.Med. 308: 129-134.143. Seshadri R., Kutlaca R.J., Trainor K., Matthews C. andMorley A.A. (1987) Mutation rates of normal and malignanthuman lymphocytes. Cancer Res. 47: 407-409.144. Shieh C.Y., Fojo A.T. and Bates S.E. (1989) Expressionof a multidrug resistance gene (mdr-1/Pgp) in human breastcarcinoma. Proc. Amer. Assoc. Cancer Res. 30: 521.145. Shoemaker R.H., Wolpert-DeFilippes M.K., Kern D.H.,Lieber M.M., Makuch R.W., Melnick N.R., Miller W.T., SalmonS.E., Simon R.M., Venditti J.M. and Vom Hoff D.D. (1985)Application of a human tumor colony-forming assay to newdrug screening. Cancer Res. 45: 2145-2153.146. Shoemaker R.H. (1986) New approaches to antitumor drugscreening: The human tumor colony forming assay. CancerTreat. Rep. 70: 9-12.147. Shopsis C. and Eng B. (1985) Rapid cytotoxicity testingusing a semi-automated protein determination on culturedcells. Tox. Lett. 26: 1-8.148. Sikic B.I. (1991) Anticancer drug discovery. J. Natl.Cancer Inst. (U.S.A.) 83: 738-742.149. Skehan P., Storeng R., Scudiero D., Monks A., McMahonJ., Vistica D., Warren J.T., Bokesch H., Kenney S. and BoydM.R. (1990) New colorimetric assay for anticancer-drugscreening. J. Natl. Cancer Inst. (U.S.A.) 82: 1107-1112.150. Skipper H.E., Schabel F.M., Jr, and Wilcox W.S. (1964)Experimental evaluation of potential anticancer agents.XIII. On the criteria and kinetics associated with"curability" of experimental leukemia. Cancer Chemother.Rep. 35: 1-111.151. Skipper H.E. (1979) Historic milestones in cancerbiology: A few that are important in cancer treatment.Semin. Oncol. 6: 506-514.152. Skipper H.E. (1986) Laboratory models: the historicalperspective. Cancer Treat. Rep. 70: 3-7.153. Slater T.F., Sawyer B. and Strauli U.D. (1963) Studieson succinate-tetrazolium reductase systems. III. Points ofcoupling of four different tetrazolium salts. Biochem.Biophys. Acta. 77: 383-393.154. Smigel K. (1991) Scientists fing better ways to find114better drugs. J. Natl. Cancer Inst. (U.S.A.) 83: 1370-1372.155. Smith I.E. (1989) Adjuvant chemotherapy in early breastcancer. Update 15 February: 359-365.156. Staquet M.J., Byar D.P., Green S.B. and Rozencweig M.(1983) Clinical predictivity of transplantabe tumor systemsin the selection of new drugs for solid tumors: Rationalefor a three-stage strategy. Cancer Treat. Rep. 67: 753-765.157. Suffness M. and Thompson J.E. (1988) National CancerInstitute's role in the discovery of new antineoplasticagents. In: Biomedical importance of marine organisms(Fautin D.G., ed), California Academy of Sciences, SanFrancisco. pp 151-157.158. Tannock I.F. (1968) The relationship betweenproliferation and the vascular system in a transplantedmouse mammary tumor. Br. J. Cancer 22: 258-273.159. Tsuruo T., Iida H., Nojiri M., Tsukagoshi S. andSakurai Y. (1983) Circumvention of vincristine andAdriamycin resistance in vitro and in vivo by calcium influxblockers. Cancer Res. 43: 2905-2910.160. Tsuruo T., Iida H., Tsukagoshi S. et al. (1982)Increased accumulation of vincristine and Adriamycin indrug-resistant P388 tumor cells following incubation withcalcium antagonists and calmodulin inhibitors. Cancer Res.42: 4730-4733.161. Twentyman P.R., Walls G.A. and Wright K.A. (1984) Theresponse of tumour cells to radiation and cytotoxic drugs-acomparison of clonogenic and isotope uptake assays. Br. J.Cancer 50: 625-631.162. Valagussa P., Tancini G. and Bonadonna G. (1986)Salvage treatment of pateints suffering relapse afteradjuvant CMF chemotherapy. Cancer 58: 1411-1417.163. Valeriote F. and Vanputten L. (1975) Proliferation-dependent cytotxicity of anticancer agents: A review. CancerRes. 35: 2619-2630.164. Van der Bliek A.M., Baas F., Van der Velde-Koerts T.,Biedler J.L., Meyers M.B., Ozols R.F., Hamilton T.C., JoenjeH. and Borst P. (1988) Genes amplified and overexpressed inhuman multidrug-resistant cell lines. Cancer Res. 48: 5927-5932.165. Venditti J.M. (1983) The National Cancer Instituteantitumor drug discovery program, current and futureperspectives: A commentary. Cancer Treat. Rep. 67: 767-772.115166. Verrelle P., Meissonnier F., Fonck Y., Feillel V.,Dionet C., Kwiatkowski F., Plagne R. and Chassagne J. (1991)Clinical relevance of immunohistochemical detection ofmultidrug resistance P-glycoprotein in breast carcinoma. J.Natl. Cancer Inst. (U.S.A.) 83: 111-116.167. Vistica D.T., Skehan P., Scudiero D., Monks A., PittmanA. and Boyd M.R. (1991) Tetrazolium-based assays of cellularviability: A critical examination of selected parametersaffecting formazan production. Cancer Res. 51: 2515-2520.168. Volkers N. (1992) Diving for drugs: Scientists searchthe sea. J. Natl. Cancer Inst. (U.S.A.) 84: 1062-1063.169. Von Hoff D.D., Sandbach J.F., Clark G.M., Turner J.N.,Forseth B.F., Piccart M.J., Colombo N. and Muggia F.M.(1990) Selection of cancer chemotherapy for a patient by anin vitro assay versus a clinician. J. Natl. Cancer Inst.(U.S.A.) 82: 110-124.170. Von Hoff D.D. (1990) He's not going to talk about invitro predictive assays again, is he? J. Natl. Cancer Inst.(U.S.A.) 82: 96-101.171. Watson J.V. (1976) The cell proliferation kinetics ofthe EMT6/M/AC mouse tumours at four volumes duringunperturbed growth in vivo. Cell Tissue Kinetics 9: 147-156.172. Wiltrout R.H., Talmadge J.E. and Herberman R.B. (1987)Role of NK cells in prevention and treatment of metastasesby biological response modifiers. In: Immune response tometastases, Vol II (Herberman R.B., Wiltrout R.H. andGorelik E. eds.), CRC Press, Bocca Raton, Florida. pp 26-41.173. Wooley P.V. and Tew K.D. (1988) Mechanisms of drugresistance in neoplastic cells. Academic Press, San Diego,California. pp 1-390.174. Yarbro J.W. (1992) The scientific basis of cancerchemotherapy. In: The chemotherapy source book (Perry M.C.,ed.), Williams and Wilkins, Baltimore. pp 7.175. Yoshinari T., Iwasawa Y., Miura K., Takahashi I.S.,Fukuroda T., Suzuki K. and Okura A. (1989) Reversal ofmultidrug resistance by new dihydropyridines with lowercalcium antagonistic activity. Cancer Chemother. Pharm. 24:367-370.176. Yusa K. and Tsuruo T. (1989) Reversal mechanism ofmultidrug resistance by verapamil: Direct binding ofverapamil to P-glycoprotein on specific sites and transportof veramail outward across the plasma membrane of K562/ADMcells. Cancer Res. 49: 5002-5006.116APPENDICES APPENDIX 1: Growth Medium for MCF-7, T-47D and ZR-75-1 CellLines for Routine PassagingF12/DME^(1:1)Hepes buffer^10 mMFBS 5%Insulin^5 gg/mlAPPENDIX 2: Growth Medium for the MDA-MB-231 Cell Line forRoutine PassagingF12/DME^(1:1)Hepes buffer^10 mMFBS 10%APPENDIX 3: Growth Medium for the MCF-7 AdK Cell Line forRoutine PassagingF12/DME^(1:1)Hepes buffer^10 mMFBS 5%Insulin^5 pg/m1Adriamycin 0.5 Ag/m1APPENDIX 4: Growth Medium for the MDA-A1K Cell Line forRoutine PassagingF12/DME^(1:1)Hepes buffer^10 mMFBS 10%Adriamycin 1 pg/miAPPENDIX 5: Preparation of Saline-Trypsin-VerseneNaHCO3NaClKC1GlucoseEDTATrypsin (Gibco)Distilled water0.35 g8.0 g0.4 g0.35 g0.25 g0.5 g1000 mlThe first 5 ingredients were weighed out and placed in alarge Erlenmeyer flask. Nine hundred and fifty ml of117distilled water was added and the solution was stirred untilthe ingredients were dissolved. Following this, the trypsinwas added and stirred until dissolved (_ 1h). The pH wasadjusted to 7.4 by the addition of either NaOH or HC1, andthen the remaining distilled water was added. Under aesepticconditions, the solution was filter sterilized using aMillipore peristaltic pump. The sterilized solution was thenstored as 10 ml aliquots in 17 x 100 mm propylene plastictest tubes in a -20°C freezer.APPENDIX 6: Preparation of Calcium and Magnesium-Free PBSNaC1^0.8 gKC1 0.03 gNa2HPO4.7H20^0.0138 gKH2PO4^0.02 gGlucose 0.2 gEDTA 0.038 gDeionizeddistilled water^100 mlAll the above ingredients are mixed into the deionizeddistilled water until dissolved. The solution is then filtersterilized using a Millipore peristaltic pump. The finalsolution is stored in the refrigerator at 5°C.APPENDIX 7: Growth Medium for the MCF-7, MCF-7 Adr, T-47Dand ZR-75-1 Cell Lines for Drug TestingF12/DME^(1:1)Hepes buffer^10 mMFBS 5%Insulin^5 µg/mlAPPENDIX 8: Growth Medium for the MDA-MB-231 and MDA-A11:Cell Lines for Drug TestingF12/DME^(1:1)Hepes buffer^10 mMFBS 10%APPENDIX 9: Transport MediumF12/DME^(1:1)Hepes buffer^10 mMBovine serum 5%Insulin^5 µg/ml118APPENDIX 10: Freezing MediumDME^ 50%Bovine serum^44%Dimethylsulfoxide 6%APPENDIX 11: Dissociation MediumF12/DME^(1:1), 18 ml/sampleHepes buffer^10 mMBSA 0.4 g/sampleInsulin^5 gg/mlCollagenase^300 U/mlHyaluronidase 100 U/mlAPPENDIX 12: Growth Medium for Human Breast Epithelial Cells(Normal and Malignant) in Primary CultureF12/DME^(1:1)Hepes buffer^10 mMInsulin 5 µg/mlPatients' serum^5%APPENDIX 13: Preparation of Human Serum SamplesSerum samples were collected in the mornings from patientswho had fasted over the previous 8-12 h. Blood was receivedin nonheparinized tubes, incubated for 30 min at 37°C,centrifuged at 100 x g and the serum was collected. If notused immediately, the serum was stored at -20°C.APPENDIX 14: Preparation of Rat Tail CollagenThe collagen solution was prepared by placing rat tails in95% ethanol for 15 min. The tendons were dissected out,weighed and then submersed in sterile deionized water whileunder UV radiation in a laminar flow hood. The fibers werethen submersed in a 0.01N acetic acid solution and stirredfor 48 h at 4°C. Subsequently they were left in the acidsolution at 4°C for 24 h with no stirring. The solution wasthen transferred into 50 ml centrifuge tubes and centrifugedin a Sorvall unitracentrifuge for 30 min at 10 000 x g. Thesupernatant consisted of the collagen solution and wasbottled and stored at 4°C.


Citation Scheme:


Usage Statistics

Country Views Downloads
United States 78 6
China 50 8
India 17 9
Saudi Arabia 3 0
Pakistan 3 0
Japan 3 0
United Kingdom 2 0
Malaysia 2 6
Ireland 1 0
Turkey 1 0
City Views Downloads
Washington 62 0
Hangzhou 20 0
Unknown 18 23
Beijing 8 0
Ashburn 7 0
Bangalore 7 2
Shenzhen 4 8
Mountain View 4 0
Shanghai 3 0
Guangzhou 3 0
Tianjin 3 0
Tokyo 3 0
George Town 2 0

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



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items