Chemical and biological studies of secondary metabolites from Lissoclinum patella

The investigation of the molecular biology of cancer and the discovery of novel anticancer agents are mutually beneficial research topics. Understanding the biology of cancer leads to improved drug design and target selection while describing new drugs and mechanisms of action may result in an increased knowledge of the workings of cancer. Marine organisms continue to be a source of bioactive secondary metabolites. Investigations of <italic>Lissoclinum patella

CHEMICAL AND BIOLOGICAL STUDIES OF SECONDARY METABOLITES FROM LISSOCLINUM PATELLA by Adam David Richardson A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Medicinal Chemistry The University of Utah August 2003 Copyright © Adam David Richardson 2003 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Adam D. Richardson This dissertation has been read by each member of the fol1owing supervisory committee and by majority vote has been found to be satisfactory. ( , I !", ''v " I 'J ! I' I· .'i. / \ , Date ' Frank A. Fitzpatrick David A. Jones J a~es A. McCloskey THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Adam D. Richardson in its final form and have found that (1) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. ~~/J . _)b d/~L YJ Date ! r ris M. Ireland Chair: Supervisory Committee Chair Approved for the Graduate Council David S. Chapm' n Dean of The Graduate School ABSTRACT The investigation of the molecular biology of cancer and the discovery of novel anticancer agents are mutually beneficial research topics. Understanding the biology of cancer leads to improved drug design and target selection while describing new drugs and mechanisms of action may result in an increased knowledge of the workings of cancer. Marine organisms continue to be a source of bioactive secondary metabolites. Investigations of Lissoclinum patella, a marine ascidian, have yielded many interesting compounds. The patellazoles are a family of compounds in which the known natural products consist of a 24-member macrolide ring with a thiazole-epoxide tail. The opening of this epoxide does not greatly affect the bioactivity of these compounds, although the general cellular toxicity is generally decreased. The patellazoles are extremely cytotoxic towards HCT 116 human colon tumor cells. Treatment with nanomolar amounts of these compounds results in immediate inhibition of protein synthesis and cell cycle arrest at G. and S phase. Interestingly, DNA synthesis is initially stimulated by patellazole treatment but the patellazoles do not interact with DNA. HCT 116 wild type cells underwent apoptosis after extended patellazole treatment. Although treatment with the patellazoles resulted in an increased amount of p53, the p53 null cells were still strongly affected by treatment. The inhibition of translation by patellazole treatment is linked to the inhibition of the mTORlp70 pathway. Like the mTOR inhibitor rapamycin, the patellazoles inhibit translation through the 4EBPI and S6 Kinase pathways. However, the cytotoxicity and inhibition profile of rapamycin and the patellazoles differ greatly in HCT 116 cells. The cellular target of the patellazoles is still unknown; the patellazole-induced inhibition of this pathway occurs either downstream or parallel to AKT. Lissoclinolide has been described to possess antimicrobial activity but its effect upon mammalian cells was previously unknown. This small, non-nitrogenous L. patella metabolite is cytotoxic towards a range of tumor cell lines. Treatment with lissoclinolide results in GiM arrest. However, neither p53 nor p21 are involved in the cellular response and apoptosis is not induced. COMPARE analysis in the NCI 60 tumor cell line panel revealed broad toxicity with some specificity towards colon tumor cell lines. v To Robyn, for the support, assistance and encouragement, scientific and otherwise. TABLE OF CONTENTS ABSTRACT ........................................................................................... iv LIST OF FIGURES ................................................................................... ix LIST OF TABLES .................................................................................... xii LIST OF ABBREVIATIONS ..................................................................... xiv ACKNOWLEDGMENTS ........................................................................ xvi Chapter I. INTRODUCTION ............................................................................ 1 Cancer Biology ................................................................................ 1 Cancer Therapies ............................................................................. 6 Marine Natural Products ................................................................... 11 Lissoclinum patella ..................... ..................................................... 17 II. THE PA TELLAZOLES ..................................................................... 20 Background ................................................................................... 20 Chemistry ....................................................................................... 22 Biology ........................................................................................ 35 Target Identification .......................................................................... 60 Discussion ..................................................................................... 66 III. LISSOCLINOLIDE ......................................................................... 72 Background .................................................................................. 72 Chemistry .................................................................................... 73 Biology ......................................................................................... 74 Discussion .................................................................................... 85 IV. CONCLUSIONS ............................................................................. 86 V. EXPERIMENTAL .......................................................................... 90 General ........................................................................................ 90 Patellazole Chemistry ....................................................................... 91 Patellazole Biology ......................................................................... 96 Lissoclinolide ................................................................................ 104 Appendices A. PATELLAZOLECHEMISTRY ........................................................ 108 B. PATELLAZOLEBIOLOGY ............................................................ 121 C. LISSOCLINOLIDE ........................................................................ 174 REFERENCES ...................................................................................... 197 viii LIST OF FIGURES H~ P~ 1. Schematic of the EGF pathway ............................................................... 2 2. Schematic of the p70 translational control pathway ..................................... .4 3. Naamidine A ................................................................................... 9 4. Raparnycin ..................................................................................... 9 5. CP-3139S ..................................................................................... 11 6. Dolastatin 10 .................................................................................. 13 7. Bryostatin 1 ................................................................................... 13 S. Didemnin B ................................................................................... 16 9. Aplidine ....................................................................................... 16 10. Ecteinascidin 743 ............................................................................ IS 11. Bistramide A ................................................................................. IS 12. Ulithiacyclamide ............................................................................. 19 13. Patellazoles A, Band C .................................................................... 21 14. Patellazoles H, I and 1 ...... ................................................................ 24 15. Possible mechanism of epoxide opening for patellazole B ............................ 34 16. Comparison of carbon NMR shifts ........................................................ 34 17. Selected HMBC correlations of patellazole H ........................................... 35 18. Relative cytotoxicities of the patellazoles towards the HCT 116 wild type cell line ................................................................................. 37 19. The effect of patellazole H on the expression level of p53 ............................. 38 20. The effect ofpatellazole H on PARP ..................................................... 39 21. The effect of patellazole B (pat B) vs. control (DMSO) on the cell cycle of AA8 (CHO) cells ................................................................... 40 22. The effect of patellazole H on the expression level of cyclin D in HCT 116 p53+1+ cells ................................................................................ 52 23. The effect of patellazole H on the expression level of cyclin D in HCT 116 p53-1 - cells ............................ 0 ••••••••••• 0 ....................................... 52 24. The effect of patellazole Hand rapamycin on p70 ........................................ 54 25. The effect ofpatellazole H on p-MAPK ................................................... 54 26. The effect of patellazole Hand rapamycin on p-MAPK ............................... 55 27. Patellazole Hand rapamycin vs. AKT ............... 0 0 •••••• 0 000 •• 0 •••••••• 0 •• 00 ••• 0.0 .. 55 28. Patellazole Hand rapamycin vs. 4EBPl. ..... 0 .......................................... 56 29. Patellazole H and rapamycin vs. S6 ...................................... ................. 56 30. The effect of 30 nM patellazole H (H) vs. DMSO (C) on the phosphorylation state of 4EBPl. ......................... 0 ••• ,'0 ........................... 57 31. The effect of 30 nM patellazole H (H) vs. DMSO (C) on the phosphorylation state of ribosomal protein S6 kinase ................................... 57 32. Patellazoles Band H vs. p70 ............................. 0 .................................... 58 33. Patellazoles Band H vs. 4EBPl. ........................................................... 59 34. Reaction scheme 1 .......................................................................... 61 35. Reaction scheme 2 .......................................................................... 63 x 36. Reaction scheme 3 ........................................................................... 63 37. Reaction scheme 4 ..................... 0 ••••••••••••••••••••••••• 0 .......................... 64 38. Lissoclinolide ................................................................................. 73 39. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on P ARP cleavage ............................................................... 77 40. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on ubiquitin distribution .......................................................... 78 41. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on p53 and expression levels ................................................ 78 42. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on p21 expression levels ....................................................... 80 43. The effect of lissoclinolide upon the NCI 60 cell line panel. ......................... 82 44. Standard curve used in protein normalization ......................................... 100 xi LIST OF TABLES 1. HRFAB Masses of the Patellazole Analogs ............................................ 27 2. IH and l3C NMR Assignments of Patellazole H ........................................ 27 3. IH and l3C NMR Assignments of Patellazole 1. ......................................... 29 4. IH and I3C NMR Assignments of Patellazole J ........................................ .31 5. IC50 Values for the Patellazoles in Three HCT 116 Cell Lines (nM) ................ 36 6. Fractional Survival of Various Cell Lines When Treated with 1 Jl.glmL of Patellazole .................................................................................. 39 7. Cell Cycle Effect of ICso Concentrations of Patellazoles on HCT 116 Wild Type Cells ............................................................................. 42 8. Cell Cycle Effect of IC50 Concentrations of Patellazoles on HCf 116 p53'" Cells .................................................................................... 43 9. Cell Cycle Effect of IC50 Concentrations of Patellazoles on HCT 116 p21'" Cells .................................................................................... 43 10. Cell Cycle Effect of IC50 Concentrations of Natural Products on HCT 116 p53+'+ Cells ............................................................................... 45 11. Cell Cycle Effect of ICso Concentrations of Natural Products on HCT 116 p53'" Cells ............................................................................... 45 12. Cell Cycle Effect of 30 nM Patellazole on HCT 116 p53+'+ Cells ................... .47 13. Cell Cycle Effect of 30 nM Patellazole on HCT 116 p53'" Cells ................... .47 14. Effect of Patellazole H on Macromolecular Precursor Incorporation .............. .49 15. Treated vs. Control Ratios of Selected mRNA Transcripts ........................... 50 16. IC~ Values of Lissoclinolide Versus Various Cell Lines ............................. 75 17. Effect of Lissoclinolide on the Cell Cycle in RCT 116 p53+1+ Cells ................ 76 18. Effect of Lissoclinolide on the Cell Cycle in RCT 116 p53-'- Cells ................. 76 19. Reversibility of Lissoclinolide Induced Cell Cycle Arrest ............................ 80 20. Antibodies Used: Epitope, Source, Dilution and Company ......................... 101 xiii 4EBPl ATP cdk cDNA CIP COSy DEPC DEPT DMSO DPM eIF4e EGF EGFR ERK ESI FAB FBS FACS FKBP12 GFR LIST OF ABBREVIATIONS eIF4E binding protein 1 Adenosine triphosphate Cyclin dependent kinase Complementary deoxyribonucleic acid Cyclin-dependent kinase inhibitor Correlated spectroscopy Diethyl pyrocarbonate Distortionless enhancement by polarization transfer Dimethylsulfoxide Decay per minute elongation initiation factor 4e Epidermal growth factor Epidermal growth factor receptor Extracellular-regulated kinase Electrospray Ionization Fast atom bombardment Fetal bovine serum Fluorescent acti vated cell sorter FK506 binding protein 1 Growth factor receptor HCT HMBC HPLC HRP HRFAB ICso MAPK MEK mRNA mTOR MTT NMR PARP PBS PBST PIK3 PKB PTEN PVDF SDS TFA TLC TOR Human colon tumor Heteronuc1ear mUltiple bond correlation High-pressure liquid chromatography Horseradish peroxidase High resolution FAB Inhibitory concentration (50%) Mitogen-activated protein kinase MAPKIERK kinase Messenger ribonucleic acid mammalian TOR 3-( 4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide Nuclear magnetic resonance Poly-(ADP-ribose) polymerase Phosphate buffered saline Phosphate buffered saline with 0.1 % (v/v) Tween-20 Phosphoinositol kinase 3 Protein kinase B; (AKT) Phosphatase and tensin homolog Polyvinylidene fluoride Sodium dodecyl sulfate Trifluoroacetic acid Thin-layer chromatography Target of rapamycin xv ACKNOWLEDGMENTS In recognizing all the people that made this doctorate possible, I'd like to start with my parents. Without their efforts and support over the last ten years of all of this would have been much more difficult and probably still not done. They have put more work into my life than most people put into their own. However, over the last few years no one has done more for me than Robyn. Rob, you have helped make this work possible more than you know. I deeply appreciate your scientific assistance with both my experiments and this dissertation, and even more greatly value the emotional support and encouragement that enabled me to get this done. My scientific career, including this dissertation, exists because of my advisor, Chris Ireland. I have learned much about science because of his continued support of my ideas and experiments, no matter how unlikely, difficult or expensive. I have had more scientific freedom and opportunity in the last five years than any graduate student could hope for (and he has the bills to prove it). Chris, thank you for the encouragement and trust. I would also like to thank the other members of my dissertation committee for their advice, ideas and (often) lab space, especially Lou Barrows and David Jones. Much of the work presented here was made possible or performed by the Core Facilities at the University of Utah. The Flow Cytometry Facility, the High Field NMR Facility, the Mass Spectrometry Facilities of the Departments of both Medicinal Chemistry and Chemistry, and the DNA Microarray Facility all provided excellent work and deserve more credit than I have the space to give. Additionally, I'd like to thank Jim Mullally for performing the lissoclinolide isopeptidase experiment, Mike Edler for performing the lissoclinolide tubuIin experiments and Doug Roberts for performing the patellazole yeast experiment. I'd like to thank Dr. Bert Vogelstein of Johns Hopkins University for generously donating the HCT 116 cell lines used in the majority of the cell-based experiments. I would also like to thank the government and marine scientists of Fiji for their permission and assistance in collecting the Lissoclinum patella that was the basis for everything presented here. This work has been supported by fellowships from the American Foundation for Pharmaceutical Education, the American Chemical Society (Division of Medicinal Chemistry) and Wyeth-Ayerst Research, as well as NIH grant CA36622. Last but certainly not least, I'd like to say thanks to everyone who made the last five years the amazing experience that it was. However much one loves science, graduate school can be overwhelming without good people around. The Ireland lab has always provided a supportive environment, both scientifically and socially, so thank you to all the members of the lab that helped me over the course of the years. Most importantly, this all would have been too much without Rob, Toaster, Kate, BenMajor, Diane, Cicely, Jay, Scott and Sandi. Respect is due. I can't imagine the last five years without you. See you Friday at pool. xvii CHAPTER I INTRODUCTION Cancer Biolo~y Cancer is the second leading cause of death in the United States, after heart disease. As the population ages and the average lifespan lengthens, the incidence of cancer is also increasing. The number of cancer deaths has risen throughout the later half of this century. Although cancer survival rates have also increased, improved treatments are needed. l Cancer is characterized by unregulated cell growth and genomic instability. There are several ways in which a cell may become cancerous. Growth signaling pathways may become overly active, whether by overexpression of growth factor receptors or through mutations which cause the signaling pathways to become constituitivelyactive. Cells may conversely become resistant to growth-inhibitory signals, again through either receptor or signaling mutations. Cells may also gain the ability to avoid apoptotic signaling that would normally cause a genomically unstable cell to undergo programmed death. 2 One example of abnormal growth signaling may be found in studies involving the epidermal growth factor (EGF) signaling pathway (Figure l). The EGF receptor (EGFR) is overexpressed in a number of cancers, including breast, head and neck, stomach, and epidermal cancers.3 When EGF activates EGFR, cell signaling cascades occur including G translation Figure 1. Schematic of the EGF pathway. 1 8 1 8 ~ 8 .. ·· .. ..... .. . . . . ~ G "- transcription 2 those which regulate transcription, division, adhesion, and death.4 One of these pathways involves Ras. Mutations in the ras gene are present in approximately 30% of 3 all human cancers.5 Ras activates Raf (a serine-threonine kinase),6 whose downstream effects include the activation of many transcription factors including Elk-I, c-Jun, c-Myc, c-Fos, and ribosomal 56 protein kinase, all of which cause cell proliferation.5 Another signaling pathway comnlonly mutated in cancer is the p70/mTOR pathway (Figure 2). This pathway helps to control translation through a phosphorylation cascade in response to external signals such as growth factors and cytokines. PIK3 is phosphorylated and activated in response to cellular stimulus by insulin, platelet-derived growth factor (PDGF), epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF).1,8 PIK3 then mediates the activation of AKT.9 The p70 56 kinase and mTOR are two focal points in this pathway. TOR was first identified in yeast as the Target Of Rapamycin.10 This protein is conserved in mammalian cells as mTOR (mammalian Target Of Rapamycin).11 p70 56 kinase (henceforth referred to as p70) is phosphorylated by the mTOR at threonine 389. Additionally, p70 threonine 229 is phosphorylated by AKT and the MAPK pathway controls the phosphorylation of threonine 421 and serine 424.12 There is some evidence that p70 may be directly phosphorylated and activated by either ERK or MEK.13, 14 Phosphorylating p70 activates its kinase activity, resulting in the phosphorylation and activation of ribosomal protein 56. Active 56 results in increased translation of protein synthesis machinery. mTOR is also able to phosphorylate 4E-BP1. 15 When not 4 1 G~ Thr308 .....---.l..... 1 Thr389 Ser424 1 8 G 1 ~ ------I •• translation Figure 2. Schematic of the p70 translational control pathway. 5 phosphorylated 4E-BPI binds to eIF4E, an initiator of translation. After phosphorylation by mTOR, 4E-BPI releases from eIF4E allowing translation to occur. 16, 17 The most well-known tumor suppressor gene is the "guardian of the genome," p53. p53 has multiple roles within a cell, including regulating gene transcription, activating cell cycle checkpoints in response to genome damage and inducing apoptosis. Thus a mutation or loss of p53 may cause a cell to become cancerous or lead to further mutations which complete the transformation. Over half of all tumors have mutated or deleted p53, making p53-controlled signaling pathways an interesting area of anticancer investigation.18 Additionally, interactions have been observed between p53 and ERK, as inhibition ofERK activity results in a decrease in the half-life ofp53.19 The roles of p21 WAFIICIPI and p53 in the control of cell growth and cell death are becoming increasingly clear, although their responses to anticancer therapies are complex.20 p53 is able to induce both cell cycle arrest and apoptosis.21 , 22 p53 is the primary director of cell cycle arrest and DNA repair following damage to genetic material. p21 WAFIICIPI is the primary mediator of cell cycle arrest at the G/S checkpoint, allowing the cell to attempt to repair DNA damage.21 , 22 p21 WAFlICIPl also has a role in sustaining arrest at the G/M checkpoint following certain types of cellular stress.23 Additionally, p53 can activate apoptotic pathways, including the induction of Bax and the caspases. The apoptosis-promoting characteristics of p53 are enhanced when p21 is not induced, whether due to the inability of p53 to increase p21 transcription or the absence of the p21WAFlICIPl gene.24 Cancer Therapies Most classic anticancer drugs work by attempting to activate apoptosis through the inhibition of cell cycle progression. Nucleotide analogs such as 5-fluorouracil interfere with DNA synthesis through inhibition of enzymes involved in DNA replication, by causing base-pair mismatches, or both. Although these drugs were first introduced many years ago, 5-fluorouracil remains the most prescribed agent for 6 colorectal cancer and novel analogs are still being developed.25 More recently, a number of drugs have been developed that act as spindle poisons, inhibiting the normal function of tubulin or actin.26 The primary rationale behind these drugs is that cancer cells are undergoing rapid cell growth and replication; thus a compound which interfers with these processes will be cytotoxic toward tumor cells. These drugs are also toxic to nontransformed cells, however, leading to a relatively small therapeutic ratio and rather poor selectivity. New paradigms in cancer treatment are currently being developed in which the molecular biology of cancer cells is much more well defined and disease-specific drug targets are identified. There are two possible strategies when treating cancer. The first and most common is to attempt to selectively kill the cancer cells while minimizing damage to healthy cells. This is difficult due to the lack of markers which may be used to identify and target cancer cells. However, as the physiology of cancer becomes more well defined, new targets are being identified. If it is possible to target a toxin to only cancerous cells and not normal cells, aggressive chemotherapy may be able to eradicate the cancer. 7 A possible strategy for treating p53 deficient tumors is to identify small molecules that are more toxic toward a p53 deficient cell line than the p53 competent parental line, while not showing differential toxicity when p21 WAFIICIPI deficient and competent lines are examined. This toxicity profile may indicate that the compound in question acts in a manner other than damaging DNA or inhibiting DNA synthesis, or affecting microtubule formation during mitosis. These two mechanisms of action are among the most common for antitumor chemotherapies and, while effective, often result in severe side effects for the patient. A compound which is selectively cytotoxic toward transformed (p53 deficient) cells by a mechanism other than the above may have significant medicinal value. A second potentially interesting profile in this small panel of ReT 116 cell lines is increased toxicity toward the p53 competent line while also exhibiting increased toxicity toward the p21 WAFlICIPI deficient line. That is, identifying the compound most toxic to the p21 WAFIICIPI deficient line and least toxic toward the p53 deficient line. The two major signaling cascades activated by p53 are the p21-dependent arrest and repair pathway and the apoptotic pathway. A compound that is preferentially toxic toward p53 competent and p21 deficient cells may be able to activate p53 dependent apoptotic pathways. One example of such an approach has been applied to EGF pathway overactivity. A small molecule from the sponge Leucetta chagosensis, naamidine A (shown in Figure 3), has been shown to specifically bind to and modulate the acti vity of ERKI and ERK2.27 Treating cells over expressing EGFR and displaying abnormally high mitogenesis with naamidine A resulted in complete growth arrest and cell death. Interestingly, naamidine A is not an inhibitor of the ERKs, but rather stimulates their kinase activity. This leads to an amplification of pro-growth signaling pathways to the point that the effected cell must compensate by shutting down. Another interesting example of a small molecule modulator of growth signaling 8 pathways is rapamycin. Rapamycin (Figure 4) was first discovered in 197528,29 and first used clinically as an immunosupressent.30 The newly discovered anticancer activity of rapamycin was recently reviewed by Hidalgo and Rowinsky.31 Its activity is derived from its ability to inhibit translation, including the production of proteins required for cell cycle progression from Ol to S phase. When rapamycin binds to the immunophilin FKBP12 (FK506 binding protein 1) the resultant complex directly inhibits mTOR. The inhibition of mTOR results in the inhibition of translation through both the 4EBP1 and p70 branches of this pathway, as discussed above. Additionally, the G1 phase cell cycle arrest is mediated by the inhibition of cyclin-dependent kinase activation, the phosphorylation of retinoblastoma protein and a reduction in the amount of the protein cycIin D1.31 There are two potential complications with any cell-death-dependent anticancer chemotherapy, however. The first is that one of the defining characteristics of cancer is genomic instability. 2 An unstable genome is more likely to acquire a mutation that leads to drug resistance. Although the initial treatment may kill most of the tumor cells, the resistant cells may be able to form a new drug-resistant tumor.32 Additionally, there is a second route to the same end point. Most tumors contain a heterogeneous population of cells. A drug that acts upon a specific target may be useless against the small percentage of the tumor with does not possess that target. Thus, 9 H OCH3 Figure 3. Naarnidine A. ,.OH HO I OMe OMe Figure 4. Rapamycin. although the treatment initially appears to cure the tumor, the small percentage of surviving cells may be able to reform the tumor.33 Another potential method of treating cancer is to rescue tumor cells by restoring 10 normal cell function and growth.34 For example, if a growth factor receptor mutation is causing a signaling pathway to be constituitively active it may be possible to partially inhibit the pathway and restore normal growth. This approach would be less likely to result in a population of drug resistant cells, although it would require treatment for the lifetime of the patient. Another strategy utilizing this approach focuses on identifying the mutations which have caused the cancer and restoring normal function of the mutated proteins. This approach has shown encouraging results in high-throughput screening format. A group at Pfizer screened a library of> 1 00,000 compounds and identified several classes of compounds which were able to thermodynamically stabilize the active conformation of p53 and restore transcriptional activity.35 This group identified a small molecule, CP-31398 (Figure 5), that is able to restore the normal function of p53. p53 can acquire mutations which destabilize its active conformation, abrogating the transcription activity which is central to the ability of p53 to act as a tumor suppressor. Treatment with CP-31398 lead to the restoration of p53 structure and function, includingtranscription. This resulted in the suppression of tumor growth both in cell culture and mouse models, demonstrating the potential power of restorative cancer treatments. .-. 11 N~N/ I OMe Figure 5. CP-31398. Marine Natural Products Not only have natural medicinal products been used by man over thousands of years, but many of the drugs being used today are derived from natural sources. 36 A review of anticancer and anti-infective agents reported between the years 1984 and 1995 revealed that over 60% of the approved drugs were of natural origin. Of the 87 anticancer drugs approved as of 1997,62% were natural products or natural product-derived. 37 Plants have been the historical source for natural medicines used by man, but in recent years marine natural products have become increasingly important. With respect to secondary metabolites, the flora and fauna of the marine environment are relatively unexplored when compared to terrestrial systems. Many of the marine invertebrates, such as sponges, cnidarians, macroalgae and tunicates, possess only primitive immune systems and limited mobility. These organisms rely on chemical defenses to both fight infection and to ward off attack. This reliance on chemical defenses has created diverse and numerous bioactive secondary metabolites, few of which have been evaluated as potential medicines.38, 39 However, even at this early stage, marine natural products have shown properties of interest to both medical and basic science research fields.37-39 A thorough review of marine chemical ecology has been compiled by McClintock and Baker40 and the Ireland research group presented a summary of marine natural products in the introductory chapter of that book.41 12 Natural product drugs from marine sources are currently in clinical development for use as anticancer agents. The marine peptide dolastatin 10 (Figure 6) is a potent inhibitor of tubulin polymerization.42 It was well tolerated in Phase I clinical trials.43 A phase II clinical trial investigating the potential of dolastatin 10 as a treatment for metastatic melanoma produced disappointing results44 as did a trial involving hormone­refractory prostate cancer.45 Dolastatin 10 was again well tolerated in both studies, however, and continues to be of interest as a potential combination therapy. Bryostatin 1 (Figure 7) has the ability to inhibit the activation of Protein Kinase C (PKC), an attractive target for cancer therapies due to its role in regulating cell growth.46 Bryostatin 1 has been evaluated in phase II clinical trials for possible use in treating advanced sarcoma,47 squamous cell carcinoma48 and relapsed multiple myeloma.49 However, like dolastatin 10, this compounds is well tolerated but disappointing as a single chemotherapeutic agent. Other marine natural products are important to the medical and research fields. Marine metabolites have shown potential as anti-infective and anti-inflammatory compounds. 38 Other compounds are used as biochemical and cellular research tools, as marine natural products have been shown to bind to membrane receptors and promote 13 o ( Figure 6. Dolastatin 10. o \ OH Figure 7. Bryostatin 1. 14 tumors. 38, 39 Due partially to the relative youth of the field of marine pharmaceuticals, no compound isolated from a marine source is currently in use as an approved anticancer agent. However, as research progresses, the biodiversity of marine secondary metabolites will produce an increasing number of medicinal compounds. Many compounds of clinical interest have been derived from natural product lead-compounds. 37 The first successful examples of this type of marine compound are the arabinosides.50 The nucleosides spongouridine and spongothymidine were isolated from sponges in the early 1950s. Today, the semisynthetic derivatives cytosine arabinoside and adenine arabinoside are approved as antitumor and antiviral agents, respectively. 39 One of the strengths of marine natural products in drug discovery is that the identification of a few bioactive conlpounds may lead to the development of multiple pharmaceutical products. Ascidians (phylum Chordata, subphylum Urochordata, class Ascidiacea) are a rich source of bioactive compounds, especially amino acid-derived, nitrogenous secondary metabolites.51 Class Ascidiacea is a relatively close relative of mammals; both have nerve cords and notochords, and so both belong to phylum Chordata. Only the larval stage of the ascidian life cycle possesses these traits, however. The appearance of the tunicates differs greatly from that of the other chordates, especially the vertebrates. Ascidians are sessile and well adapted to suspension feeding, and have lost the defined nerve cord over the course of their evolution.52 Ascidians are host to a range of organisms. This relationship may be commensal, symbiotic or parasitic. In regard to the chemical ecology of ascidians, the symbiotic relationships are the most important, as an exchange of metabolic products occurs 15 between the host and guest. Symbiotic algae are the most common and most consequential of these guests. The two algal genera observed as symbionts are the unicellular procyanophyte Synechocystis and the unicellular prochlorophyte Prochloron. Although the nature of this symbiosis is unclear (the extent and the nature of the materials exchanged have not been defined), ascidian parents pass symbionts onto their offspring and the ascidians and algae involved have not been cultivated independently of each other. Compounds isolated from ascidians, such the didemnin B, aplidine and ecteinascidin-743 have displayed sufficient biological activity to warrant clinical trials. The didemnins are a series of depsipeptides isolated from the tunicate Trididemnum solidum. Didemnin B (Figure 8) has shown both antitumor and antiviral properties.53, 54 Its antitumor activity is intriguing enough that it went into phase n clinical trials as a potential treatment for central nervous system tumors,55 Glioblastoma multiforme56 and non-Hodgkin's lymphoma (NHL).57 Unfortunately, all three of these studies revealed that the dose-limiting toxicity was too high in relation to the moderate-to-Iow antitumor activity. A similar compound, aplidine (Figure 9), is currently in phase n clinical development. 58 Aplidine displays significant cytotoxicity in breast, melanoma and non­small- cell lung cancer cell lines, although some nonspecific toxicity was again observed. 59 This compound can cause cell cycle arrest in both GJG 1 and GiM stages of the cycle and cell death but the molecular target is still unknown. 59 16 o ~~I o 0 OMe Figure 8. Didemnin B. o Yy~1 o 0 OMe Figure 9. Aplidine. 17 Ecteinascidin 743 (ET-743) is produced by the ascidian Ecteinascidia turbinata.60,61 This compound, shown in Figure 10, can bind to the minor groove of DNA and effect transcription in a complex manner.62 Early phase I clinical trials observed that prolonged exposure to ET-743 was the most efficacious schedule63 and subsequent phase I trials observed antitumor activity, encouraging pharmacokinetics, and acceptable toxicity.64, 65 ET-743 is currently undergoing phase II clinical trials. The macrolides bistramides A and B (previously known as the bistratenes) were isolated from the tunicate Lissoclinum bistratum and have been observed to enhance the phospholipid-dependent activity of protein kinase C.66 The structure of bistramide A (shown in Figure 11) was subsequently corrected after extensive NMR investigations.67 This activity may be related to the ability of bistramide A to effect the polymerization of microtubules.68 Additionally, while the activation of PKC is undoubtedly important in the mechanism of action of the bistramides, the cell cycle arrest and apoptotic results of bistramide treatment appear to be independent of PKC.69 Lissoclinum patella The ascidian Lissoclinum patella has been an especially prolific source of bioactive secondary metabolites. Classes of compounds include the lissoclinamides,70 patellamides,71 ulithiacyclamides,72 tawicyclamides,73 patellins 74, patellazoles 75, 76 and lissoclinolide.77 All are cyclic peptides except the patellazoles, which are macrolides with a thiazole side chain, and lissoclinolide, a non-nitrogenous lactone. These two compounds will be further introduced in Chapter II and III of this dissertation. o ,)lo Figure 10. Ecteinascidin 743. H Figure 11. Bistramide A. N H "0 OH 0 18 19 The cyclic peptides range in size from heptapeptides to octapeptides and are characterized by the presence of thiazole and/or oxazoline amino acids, and have a wide range of cytotoxicity. The most biologically interesting of these molecules is ulithiacyclamide.72 Ulithiacyclamide (Figure 12) displayed significant cytotoxicity toward the L1210 murine leukemia cell line (ICso = 0.35 nglmL), the CEM cell line (0.01 ~glmL) and the HeLa cell line (0.1 ~glmL) 71 and a therapeutic index (TIC) of 188 in an in vivo leukemia model. 78 S N '\ '\ o Figure 12. Ulithiacyclamide. CHAPTER II THE PA TELLAZOLES Background The patellazoles are a family of marine natural products which contain a 24 member macrolide ring and a thiazole-containing tai1.75, 76 These compounds, shown in Figure 13, were first described in 1988 in concurrent papers by Zabriskie 75 and Corley.16 The ascidian L. patella (described above) was collected in Fiji by the Ireland group and in Guam by the Paul group. To date, only L. patella collected in Fiji and Guam has yielded any patellazoles. Investigations of L. patella collected in the Philippines, Australia and Indonesia have so far failed to yield the patellazoles. Zabriskie and Ireland provided the structure elucidation of patellazole C while Corley and Paul described patellazole B. In total, the Ireland group isolated 97 mg of patellazole A, 143 mg of patellazole Band 313 mg of patellazole C from 220 g of freeze-dried organism during this investigation. Additionally, Zabriskie partially described another four members of the patellazole family (patellazoles D - G) in his dissertation.78 The patellazoles are extremely cytotoxic, particularly patellazole B, but neither the mechanism of action nor the cellular target of these compounds is known.78 Some interesting data has been identified, however. The mean ICso of patellazole A, Band C against the NCI human cell line panel was 3 x 10-4 /lglmL, <10-6 /lglmL and 3 x 10-3 Figure 13. Patellazoles A, Band C. A: RJ = H, R2 = H; B: Rl = H, R2 = OH; C: Rl = OH,R2=OH. ILglmL, respectively.78 Similar values were reported for these compounds against the L1210 murine leukemia cell line. Patellazole B was also reported to having 20 X toxicity toward p21 deficient cells as compared to the parental wild type cell line. Interestingly, treatment with the patellazoles causes an increase in DNA 21 synthesis78, but the patellazoles do not interact with DNA or any topoisomerase.79 An increase in DNA synthesis is often the result of DNA damage; however, this does not seem to be the case with the patellazoles. Additionally, patellazoles Band C were found to have excellent antiviral activity against Herpes simplex viruses I and 2 in vitro but not in vivo. Due to the interesting biological activity profile of the patellazoles L. patella 22 was recollected from Fiji many times between 1984 and 2001. The patellazoles isolated from these organisms were used to further investigate the mechanism of action of these compounds. The L. patella collected in 1996 was examined by Ashantai Yungai. During this investigation 5.4 mg of patellazole B was partially purified. This material was stored at 40 C for approximately 14 months before being used in two of the experiments which are described in the following sections. The flow cytometry experiment described on p 44 described the effect of this material upon AA8 Chinese Rampster Ovary (CRO) cells. The DNA microarray experiment described on p 54 also utilizes this material, as the subsequent recollections in 2001 had not yet been performed. Chemistry Reisolating the patellazoles was complicated by unexpected variations in patellazole biosynthesis. The Ireland research group has analyzed L. patella collected from the Fijian Islands between 1984 and 2001. Collections in November 1984, October 1987 and December 1988 were investigated by Dr. T. Mark Zabriskie and the collection in July 1997 by Ashantai Yungai and myself. I analyzed additional collections from January, July, August and October of 2001. This issue first came into question in January of 2001 when L. patella collected from the Astrolobe Reef near Kandavu failed to yield any patellazoles. The collections performed in 1984, 1987, 1988 and 1997 were at or near the Astrolobe Reef in the Fijian island group of Kandavu. Kandavu is located south of the main island of Viti Levu. Analysis of the 1984 collection by Dr. Zabriskie yielded 96.5 mg patellazole A (0.04% (w/w», 144.3 mg patellazole B (0.06% (w/w», and 312 mg 23 patellazole C (0.14% (w/w». The percent yields are the percentage of the mass of compound isolated compared to the dry mass of 220 g lyophilized anima1.78 Also, the Kandavu collection in October 1987 (183 g lyophilized material) yielded large amounts of patellazole. The exact yields are not available, but at least 78 mg, and probably hundreds of mg (based upon the masses of other crude fractions), of patellazoles were present. 80 The 1988 collection also yielded large-but-undescribed amounts of patellazoles.80 The examination of each of these three collections relied upon methanol/water C 18 High Pressure Liquid chromatography (HPLC ) at some step of their purificati on. The L. patella collected in 1997 was first examined by Ashantai Yungai. In this study, 5.4 mg of patellazole B was partially purified from 200 g of lyophilized animal material. The remaining organism remained frozen until is was lyophilized and extracted by me in November 1998. Exhaustive extraction followed by flash chromatography and HPLC yielded the three new patellazole shown in Figure 14. However, the purification yielded only 1.7 mg of patellazole H, 0.7 mg of patellazole I and 0.6 mg of patellazole 1. These compounds were isolated as only 0.00085% (w/w), 0.00035% (w/w) and 0.0003% (w/w) mass of the 200 g of dry organism, respectively. Although there appeared to be significant amounts of patellazole present in this organism purification proved difficult and the extracted material was chromatographed many times. During this process it appears that the epoxide present in the natural products was opened by the addition of either methanol or isopropanol into the ring at carbon 31. Methanol was present in purification steps involving methanol:water solvent 24 Figure 14. Patellazoles H, I and J. H: R2 = OH, R3 = OCH3; I: R2 = H, R3 = OCH3; J: R2 = OH, R3 = OCH(CH3)2' systems over a C18 solid phase and isopropanol was present in steps involving hexane:isoproanol solvent systems over a silica solid phase. This reaction was previously observed by Dr. Zabriskie when patellazole C remained in 85: 15 methanol:water for 3 days.80 BC and DEPT NMR experiments revealed three new carbon peaks at 51, 74 and 83 ppm coinciding with the disappearance of peaks at 60 and 65 ppm (C~6)' At that time the product was not further characterized and the epoxide was hypothesized to be opened at carbon 32. It is not clear what caused the reaction of occur as the patellazoles were exposed to water/methanol mixtures in many other experiments by both Dr. Zabriskie and myself without incident. 25 L. patella was collected from the Astrolobe Reef again in January, 200l. The animal material was lyophilized, resulting in 200 g dried material. No patellazoles were observed or isolated from this sample. In July 2001 a test collection was performed by the laboratory group of Dr. Bill Aalbersberg (University of the South Pacific, Fiji) from Seashell Cove in Navula Pass. Seashell Cove is located just off the western shore of Viti Levu near Nadi. L. patella was collected here, lyophilized and sent to me for analysis. From the 20 g of dry organism approximately SO mg of crude, mixed patellazoles were partially purified. Considering that L. patella had not been collected from the Kandavu site in January before 2001, the possibility that the lack of patellazoles was due to a seasonal variation in biosynthesis was considered. A small collection of L. patella (5 g dried material) was again performed by the Aalbersberg group in late Aug 2001 at Kandavu site. However, the patellazoles were again not observed. Thus it appears that a localized event occurred in the Astrolobe Reef between July 1997 and January 2001 which resulted in the loss of patellazole biosynthesis in the L. patella population. A large collection of L. patella yielding significant amounts of patellazoles A, B and C was gathered from Navula Pass in October of 200l. The 135 g dried animal material yielded 27.S mg patellazole A (0.02% (w/w», 40.0 mg patellazole B (0.03% (w/w» and 190.0 mg patellazole C (0.14% (w/w». The final purification step for patellazoles A and B was 60:40 hexane:chloroform over a diol HPLC column while the final purification step for patellazole C was 9:1 methanol:water over CIS. It is not known why these purification steps did not result in the production of solvent adducts when the extremely similar experiments described above did. The structures of the analogs were determined by NMR and mass spectrometry. 26 Due to the small amount of compounds isolated, the structure determination partially relied upon comparison of the acquired spectral data to that reported for patellazoles A, B and C.78 Patellazole H was produced by the addition of methanol to patellazole B, patellazole I from the addition of methanol to patellazole A and patellazole J from the addition of isopropanol to patellazole B. High resolution positive FAB mass spectrometry resulted in the masses and formulas shown in Table 1. Low resolution positive F AB mass spectrometry also yielded signals possibly corresponding to products resulting from the addition of isopropanol to patellazole A (MH+ = 948) and for each of the solvents to patellazole C (MH+ = 952 for the potential methanol adduct and MH+ = 980 for the potential isopropanol adduct). While patellazole A, B and C were not isolated it is reasonable to conclude that they were produced by this L. patella sample in and that the repeated chromatographic separation resulted in the formation of the solvent adducts. The NMR shifts of the protons and carbons of patellazoles H, I and J are shown in Tables 2, 3 and 4, respectively. The spectra were obtained on a 500 mHz instrument and the compounds dissolved in deuterated chloroform. The chemical shifts are presented as parts per million shifts, as referenced to the residual chloroform signal. Additional information regarding specific NMR experiments may be found in the EXPERIMENTAL chapter (V) of this dissertation. patellazole H I J MH+ 936.5514 920.5526 964.5841 Table 1 HRFAB Masses of the Patellazole Analogs fonnula C~oH82NOI3S C~82N012S CS2Hs6N013 S Table 2 required mass 936.5507 920.5558 964.5820 IH and 13C NMR Assignments of Patellazole H carbon # l3C (ppm)a IH (ppm)b 1 171.4 2 81.1 3 32.6 1.72 4 31.5 1.01 5 28.5 1.51 6 43.1 1.25 7 72.4 3.66 8 49.1 3.04 9 216.9 10 47.9 3.64 11 129.5 5.29 12 131.7 5.84 13 31.7 3.09 14 38.1 1.63 15 74.1 3.32 16 69.0 3.80 17 86.9 3.74 18 132.0 19 132.8 6.11 20 125.1 6.29 21 135.7 5.80 27 error (mmu) +0.7 -3.2 +2.1 28 Table 2 Continued carbon # l3C (ppm)a IH (ppm)b 22 35.4 2.96 23 85.1 4.54 24 75.2 25 128.9 5.37 26 134.6 27 34.6 3.54 28 153.7 29 115.3 6.98 30 174.2 31 83.0 32 74.5 3.85 33 16.4 0.94 34 175.3 35 48.9 2.33 36 68.9 3.57 37 20.0 1.16 38 23.9 1.37 39 17.9 0.83 40 13.5 0.80 41 15.3 1.07 42 15.8 0.99 43 56.3 3.19 44 10.8 1.73 45 19.0 1.33 46 27.9 1.31 47 25.1 1.69 48 15.0 1.56 49 14.7 1.11 50 51.6 3.20 a measured at 125 MHz; referenced to CDCl3 (77.0 ppm) b measured at 500 MHz; referenced to residual CHCl3 (7.26 ppm) 29 Table 3 IH and I3e NMR Assignments ofPatellazole I carbon # I3e (ppm)a IH (ppm)b 1 171.52 2 80.42 3 33.46 2.13 4 31.39 1.02 5 28.86 1.54 6 42.83 1.23 7 72.72 3.65 8 49.32 2.95 9 216.4 10 47.78 3.69 11 129.23 5.19 12 129.54 6.04 13 31.32 2.94 14 37.44 1.81 15 74.12 3.41 16 68.54 3.60 17 88.16 3.80 18 131.69 19 133.73 5.77 20 124.69 4.07 21 135.85 5.96 22 35.97 2.99 23 84.31 4.67 24 74.96 25 128.94 5.44 26 134.74 27 34.49 3.73 28 153.45 29 115.31 6.95 30 174.31 31 82.54 32 74.31 3.76 33 15.83 0.91 34 175.39 35 41.01 2.38 36 29.3 1.26 37 11.51 0.94 38 22.51 1.45 carbon # 39 40 41 42 43 44 45 46 47 48 49 50 Table 3 Continued 18.38 13.65 14.3 16.42 56.42 11.94 19.48 27.26 24.65 14.97 14.08 51.38 0.89 0.85 1.09 0.98 3.23 1.70 1.28 1.33 1.69 1.59 0.92 3.22 a measured at 125 MHz; referenced to CDCl) (77.0 ppm) b measured at 500 MHz; referenced to residual CHCl3 (7.26 ppm) 30 31 Table 4 IH and l3e NMR Assignments of Patellazole J carbon # I3e (ppm)a IH (ppm)b 1 171.30 2 80.90 3 32.37 2.25 4 31.19 1.04 5 28.40 1.68 6 42.72 1.28 7 72.34 3.67 8 48.86 3.06 9 216.33 10 47.37 3.67 11 129.61 5.28 12 131.38 5.89 13 31.38 3.12 14 37.95 1.85 15 73.90 3.33 16 68.86 3.76 17 86.66 3.81 18 132.20 19 132.36 6.14 20 125.06 6.21 21 135.30 5.89 22 35.13 2.92 23 84.81 4.59 24 75.00 25 128.74 5.39 26 134.40 27 34.46 3.53 28 153.33 29 114.98 6.93 30 175.50 31 82.50 32 74.50 3.89 33 16.21 0.93 34 175.23 35 48.51 2.38 36 68.79 3.59 37 19.85 1.19 38 23.49 1.39 carbon # 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Table 4 Continued 17.64 13.24 14.89 15.51 55.79 10.63 18.75 27.64 24.76 16.29 14.48 66.74 24.52 23.92 0.86 0.84 1.10 1.02 3.22 1.75 1.35 1.33 1.70 1.62 1.14 3.73 1.15 1.02 a measured at 125 MHz; referenced to CDC}3 (77.0 ppm) b measured at 500 MHz; referenced to residual CRC}3 (7.26 ppm) 32 The nucleophilic attack of solvent at the carbon alpha to the thiazole ring is supported by both the data and by theory. A quaternary carbon will be more able to support a carbocation or positive dipole charge than the adjacent methine, as will a 33 carbon close to the aromatic ring.81 This reaction is illustrated in Figure 15. The overall reaction results in the conservation of the patellazole and solvent masses as the epoxide oxygen is protonated and the solvent oxygen attacks the carbon alpha to the thiazole ring. Additionally, the carbon shift of 83.0 ppm for carbon 31 is more appropriate for a quaternary ether than a secondary ether. Figure 16 compares the carbon NMR shifts of the relevant carbons in patellazoles Band H. The liMBC correlations illustrated below (Figure 17) provide spectra) evidence of the quaternary location of the solvent addition. The experiment was optimized to observe three-bond correlations. Additionally, a COSY correlation is observed between the proton attached to carbon 32 and the methyl protons of carbon 33. No liMBC correlations involving H-32 were observed; however, this is most likely due to the low signal provided by the methine as compared to the methyls. The opening of the patellazole epoxide was previously observed by Mark Zabriskie when patellazole C reacted with methanol after remaining in 85: 15 methanol:water for three days. It is not clear what caused these reactions to occur, because water/solvent mixtures were used in other situations without incident. However, this unique reactive center may be able to be used in the site-selective derivatization of the patellazoles. Some such attempts were made to duplicate this reaction under controlled conditions and are described later in this chapter. 34 / fJ:Jz N /bC H~3 s AOH Figure 15. Possible mechanism of epoxide opening for patellazole B. Methanol is shown as the example solvent; isopropanol is proposed to act in the same manner. carbon .c H 48 15.6 15.0 31 60.0 83.0 32 65.0 74.5 33 14.0 16.4 33 50 51.6 48 33 48 Figure 16. Comparison of carbon NMR shifts. Patellazole C is on the left and patellazole H in on the right (both in ppm). 35 s-~ 33 Figure 17. Selected HMBC correlations of patellazole H. Biology The ability of both the natural products and the analogs to inhibit cellular growth was examined in HCT 116 cell lines using the MIT assay. This assay measures the amount of total growth inhibition through observing the activity of cellular mitochondria.82 MIT (3-(4, 5-dimethy Ithiazol-2-y 1)-2, 5-dipheny 1-2H -tetrazoli urn bromide), a soluble dye, is converted to an insoluble crystal by succinate dehydrogenase in actively respiring mitochondria; the lack of production of this compound correlates to decreased succinate dehydrogenase activity in tumor cells.83 The amount of enzyme activity is directly proportional to the total amount of cellular growth in the observed culture. The patellazoles are very potent inhibitors of cellular growth, as shown in Table 5. patellazole A B C H I J Table 5 ICso Values for the Patellazoles in Three RCT 116 Cell Lines (nM) RCT 116 wt RCT 116 p53-'- 0.62 0.66 0.39 0.62 4.70 5.60 30.0 34.0 8.80 8.30 2.60 3.00 RCT 116 p21-'- a a -a 9.0 2.4 1.0 apatellazoles A, Band C were not tested in the HCf 116 p21-'- cell line Patellazole B has been previously reported as the most cytotoxic of the three 36 patellazole natural products.78 Analyzing the relative cytotoxicity of the natural products and the analogs resulted in an interesting pattern, as shown in Figure 18. Patellazoles A and B were found to be similarly cytotoxic. Patellazole J, the isopropanol adduct of patellazole B, showed the next greatest cytotoxicity, followed by patellazole C, patellazole I and patellazole R. Patellazole J is approximately an order of magnitude less cytotoxic than the natural product. Patellazole H, the addition of methanol to patellazole B, is approximately two orders of magnitude less cytotoxic than the natural product. The addition of methanol to patellazole A, creating patellazole I, causes an order of magnitude loss of cytotoxicity. The two main possibilities for this pattern of altered cytotoxicity are a) that the opening of the epoxide and resulting functional groups have an altered affinity for the patellazole target, or b) the solvent addition changes either the ability of the compound to enter the cells or the distribution of the compound within the cells. Without further 37 B > A > J > C > I > H (A + MeOH) (B + MeOH) Figure 18. Relative cytotoxicities of the patellazoles toward the HCT 116 wild type cell line. Patellazole B is the most cytotoxic; patellazole H the least. The parent natural product and sol vent adduct is listed below each of the artifacts. understanding the mechanism of action of the patellazoles or detennining the cellular target; it is not possible to determine the structure-activity relationships between these compounds. However, the opening of the epoxide by no means abrogates the cytotoxicity of the patellazoles, and as will be discussed below, the overall biological profile of the analogs and the natural products is very similar. The role of the tumor suppressor gene p53 in the response of HCT 116 cells to the patellazoles was examined. As shown in Table 5, HCT 116 p53-'- cell lines did not show increased sensitivity toward either the natural products or the analogs as compared to the p53+'+ parental cell line. However, treatment of HCT 116 wild type cells with an ICso value of patellazole H resulted in a significant increase in the amount of p53 present in the cells after both 6 and 18 h of treatment. This western blot is show in Figure 19. Patellazoles H, I and J were used to examine whether the presence or absence of the cyclin inhibiting protein p21 effected the ability of HCT 116 cells to respond to patellazole treatment. HCT 116 p21-'- cells were approximately three times more sensitive toward each of the three analogs than the parental p21-competent cell line. This is a typical response for a compound which causes GJGl or S phase arrest of the cell cycle. 38 c H c H p53 6h ISh Figure 19. The effect of patellazole H on the expression level of p53. Treatment with the ICso value of patellazole H (H) after 6 h or 18 h as compared to the vehicle-treated control (C). The cleavage of poly(ADP-ribose) polymerase (PARP) from a native 116 kD form to 85 kD and 31 kD fragments is an early event during caspase-mediated apoptosis.84 Treatment ofHCT 116 wild type cells with 30 nM patellazole H for 18 h caused PARP cleavage indicating that apoptotic pathways had been activated. This western blot is shown in Figure 20. The ability of the patellazoles to inhibit the growth of Cac02 cells (a colon tumor cell line that is able to differentiate) and two breast cancer cell lines is shown in Table 6. The two breast cancer cells lines are MDA-MB-435S (which over expresses EGFR) and a daughter cell line, MDA-MB-468, in which both copies of PTEN have been disrupted. Surprisingly, none of the patellazoles showed significant cytotoxicity toward these cell lines, although another compound isolated from this organism, lissoclinolide, displayed similar toxicity toward both the HCT 116 cell lines and the Cac02 and MDA-MB cell lines. c H C H PARP 6h ISh Figure 20. The effect of patellazole H on P ARP. Treatment with an IC50 value of patellazole H (H) after 6 h or 18 h as compared to the vehicle-treated control (C). Table 6 Fractional Survival of Various Cell Lines When Treated with 1 JLglmL of Patellazole patellazole Cac02 MDA-MB-435S MDA-MB-468 A 0.86 0.72 0.62 B 0.89 0.92 0.87 C 0.74 0.79 1.10 H 0.85 0.83 0.71 I 0.83 0.56 0.46 J 0.72 0.77 1.01 39 40 The inability of the patellazoles to significantly affect these cell lines, especially the Cac02line, may indicate that the patellazoles affect the Ras pathway since RCT 116 wild type cells have constituitively active Ras.85 Patellazoles A, B, C, R, I and J all have the ability to arrest cells in the Go/Gt or S phase of the cell cycle. The first flow cytometry experiment performed examined the effect of a 28 nM treatment of patellazole B on the AA8 CRO cell line, as show in Figure 21. This treatment caused an increase in the percentage of cells in the Go/Gl and S phase of the cell cycle while eliminating the percentage of cells in G/M. Interestingly, 18 h of treatment did not result in an increase in the amount of cellular debris as compared to the control culture. This may be because the cultures were washed with versene before harvesting, resulting in the loss of dead cells and dead cell fragments. DMSO (control) GJGt : 36.33% S: 45.07% G11M: 18.6% debris: 6.18% 28nMpatB GJGt": 44.35% S: 55.65% G/M:O.OO% debris: 0.29% Figure 21. The effect of patellazole B (pat B) vs. control (DMSO) on the cell cycle of AA8 (CRO) cells. 41 A more thorough study of the effect of the patellazoles on the cell cycle was undertaken using the panel of HCT 116 cells. Patellazoles H, I and J were examined in the HCT 116 wild type, p53 null and p21 null cells at multiple concentrations while the natural products were tested in the wild type and p53 null daughter line. Generally, treatment with any of the patellazoles resulted in arrest in the GoIGl and S phase of the cell cycle in the wild type cells. This arrest was abrogated in the p21 null cell cultures. This is a reasonable result considering the central role p21 plays in GoIG 1 arrest and the fact that p53 is the major regulator of p21 expression. The data presented below are representative of the results obtained from multiple experiments. The cell cycle arrest in the HCT 116 cell lines is not as striking as the AA8 cell cycle arrest shown in Figure 21. This is most likely due to a specific trait or traits in the AA8 CHO cell line; identifiying this trait will require further understanding of the mechanism of action of the patellazoles. As shown in Table 7, treatment of wild type HCT 116 cells with the ICso concentration of each patellazole artifact resulted in an increase in the percentage of cells in GoIG I after 24 h of exposure. The amount of cellular debris increased as compared to the control, indicating that the patellazoles were inducing cell death. At this time point, the percentage of cells in S phase did not increase. After 48 h of exposure to an ICso concentration of patellazole H, the percentage of cells in GoIGl decreased while the percentage in S rose. However, the amount of cellular debris was also greatly increased. Most likely, cells that had been arrested in GoIGl had died, resulting in the large amount of cellular debris. The increase in the percentage of cells in the S phase may be due to cells slowly progressing across the GoIG1 arrest point and rearresting in S phase, or it may be an artifact due to the low amount of S phase cells in the contro] culture. 42 Table 7 Cell Cycle Effect of ICso Concentrations of Patellazoles on HCT 116 Wild Type Cells treatment duration GIG I S GiM debris control 24h 48.81 33.87 17.32 1.95 patellazole H 24h 58.69 32.46 8.84 9.48 patellazole I 24h 57.13 33.93 8.94 7.44 patellazole J 24h 59.19 28.65 12.15 9.31 control 48 h 76.87 13.92 9.21 7.83 ,eatellazole H 48 h 54.68 39.76 5.55 84.09 HCT 116 p53-1 - cells show a similar profile in that treatment with ICso concentrations of the patellazole H, I or J causes arrest in GoIG I after 24 h of exposure (Table 8). The amount of cellular debris does not increase at this time point, however. p53 is a major mediator of apoptosis; the loss of this protein would be expected to decrease the amount of cell death. Interestingly, after 48 h exposure to patellazole H, there remains an increase in the percentage of cells in GoIGl' possibly because cells that underwent apoptosis in the wild type cell line remained viable in the p53 null cell line. However, the amount of cellular debris begins to increase after 48 h exposure, indicating that p53-independent cell death is occurring. Table 9 reveals that knocking out p21 from the HCT 116 cell line results in the abrogation of the GoIG I arrest. p21 is the major mediator of GoIG 1 arrest, so this result is not surprising. The percentage of cells in S phase is increased after 24 h exposure to the patellazole artifacts. This may indicate that the patellazoles activate cell cycle arrest checkpoints in both the Go/G I and S phases of the cell cycle. If p21 is active, cells 43 Table 8 Cell Cycle Effect of ICso Concentrations of Patellazoles on HCT 116 p53 -1- Cells treatment duration GJG, S G/M Debris control 24h 35.57 40.52 23.91 2.85 patellazole H 24h 47.06 33.84 19.1 5.61 patellazole I 24h 44.5 31.17 24.33 7.36 patellazole J 24h 44.61 32.48 22.91 4.78 control 48 h 65.38 17.58 17.04 4.00 Eatellazole H 48 h 73.2 10.64 16.16 24.28 Table 9 Cell Cycle Effect of ICso Concentrations of Patellazoles on HCT 116 p21-1 - Cells treatment duration GJG) S G/M debris control 24h 39.66 40.61 19.74 29.11 patellazole H 24h 29.55 47.71 22.73 39.15 patellazole I 24h 28.22 49.62 22.16 38.1 patellazole J 24h 30.99 48.81 20.2 41.66 control 48 h 53.12 35.06 11.82 12.35 Eatellazole H 48 h 27.03 38.04 34.93 47.66 44 arresting in 00101 would not advance to S phase, masking this potential checkpoint arrest. Also of note is the relatively large amount of cellular debris in both the treated and control cells after 24 h. This illustrates the importance of p21 and general cell cycle arrest in cellular survival. If cellular replication and division is not regulated properly cell death may result, especially when the cell is challenged with a toxin. The HCT 116 wild type cells display a similar response to treatment with ICso values of the patellazole natural products (Table 10). The percentage of cells in 00101 increases after 24 h of exposure to the drugs; after 48 h of exposure there is a decrease in GoIOI arrested cells with a corresponding increase in the amount of cellular debris. Like patellazole H, patellazoles A and C also cause an increase in the percentage of cells in S phase at this time point. It is not known why treatment with patellazole B did not cause a similar response. Most likely there was an error made during treatment; a similar result occurred with patellazole B at the 24 h time point in Table 11. While the overall HCT 116 wild type response to the natural products is similar to that of the artifacts, the cell cycle arrest and increase in cell debris were not as dramatic as when the cells were exposed to the patellazoles H, lor J. This may be due to the lesser amount of drug applied when investigating the natural products - because both the natural products and artifacts were treated at the IC50 concentration determined by the MIT assay (Table 5), the natural products generally necessitated a lesser dose. This issue will be further discussed below. The result of treatment with the natural products in the p53 null HCT 116 cell line as compared to the HCT 116 wild type line is similar to that of the analogs (Table 11), The 00101 arrest is generally conserved while the amount of cell debris is generally 45 Table 10 Cell Cycle Effect of ICso Concentrations of Natural Products on HCT 116 p53+'+ Cells treatment duration GJG1 S GlM debris control 24h 50.1 34.5 15.4 4.0 patellazole A 24h 60.6 29.0 10.4 5.0 patellazole B 24 h 67.7 25.3 7.0 4.9 patellazole C 24h 71.2 19.3 9.6 6.7 control 48 h 68.0 25.9 6.2 7.8 patellazole A 48 h 60.5 32.1 7.4 16.0 patellazole B 48 h 69.8 24.7 5.5 21.5 Eatellazole C 48 h 62.3 31.2 6.4 17.5 Table 11 Cell Cycle Effect of 1C.50 Concentrations of Natural Products on HCT 116 p53-'- Cells Treatment duration GJG) S G,,/M t debris control 24h 45.4 34.2 20.4 5.0 patellazole A 24 h 62.2 20.5 17.3 4.4 patellazole B 24h 40.8 37.3 21.8 4.2 patellazole C 24h 66.1 20.6 13.3 6.7 control 48 h 74.3 17.3 8.4 18.8 patellazole A 48 h 66.9 23.4 9.8 8.6 patellazole B 48 h 70.9 22.8 6.3 16.9 Qatellazole C 48 h 75.4 18.3 6.4 12.6 46 decreased. As with the overall effect, the difference between the p53 competent and null lines is muted when compared to the results obtained from treatment with patellazoles A, Band C. This again raises the point that the cells are generally exposed to a greater concentration of drug when treated with ICso concentrations of the analogs than with ICso concentrations of the natural products. The five experiments described above involve treating the cells with an ICso amount of drug as determined by the MTT cytotoxicity assay. However, this results in levels of patellazole H, for example, 100 X that of patellazole B. Patellazole H is the methanol adduct of patellazole B. In an attempt to examine the effects of such differences in dose, HCT 116 wild type and p53 null cells were exposed to 30 nM patellazoles A, Band C for 24 h or 48 h. This is the ICso value of patellazole H, the patellazole family member with the least cytotoxicity. Incubation of HCT 116 wild type cells with 30 nM patellazoles A, B or C for 24 h or 48 h all resulted in a decrease in the percentage of cells in 00101 (Table 12). This effect is particularly strong after 48 h exposure. As observed in the ICso experiments, this decrease in 00101 arrest cells correlates with an increase in cells arrested in S phase and in the amount of cellular debris. Apparently, as exposure time or dose of the patellazoles increases, HCT 116 wild type cells transition from Go/01 arrest to either S phase arrest or cell death. Knocking out p53 alters the response of the HCT 116 cells to high levels of the patellazole natural products (Table 13). While the p53 competent cells displayed an increase in both the percentage of cells in the S phase and in cell debris after 24 h of 47 Table 12 Cell Cycle Effect of 30 nM Patellazole on RCT 116 p53+'+ Cells treatment duration Gc!G1 S G/M debris control 24h 42.6 41.5 15.9 23.5 patellazole A 24 h 33.5 58.0 8.5 36.8 patellazole B 24h 33.4 57.2 9.4 31.4 patellazole C 24h 33.3 58.2 8.6 38.2 control 48 h 70.0 24.0 5.9 18.1 patellazole A 48 h 36.0 49.7 14.2 64.0 patellazole B 48 h 40.3 47.8 12.0 66.8 ~atellazole C 48 h 41.8 44.4 13.8 62.8 Table 13 Cell Cycle Effect of 30 nM Patellazole on RCT 116 p53-'- Cells treatment duration Gc!G) S G/M debris control 24h 45.0 43.2 11.8 31.9 patellazole A 24h 50.4 41.3 8.3 37.8 patellazole B 24h 48.8 41.4 9.8 27.0 patellazole C 24h 45.1 44.5 10.4 39.8 control 48 h 74.7 20.5 4.8 17.3 patellazole A 48 h 51.0 32.2 16.9 65.7 patellazole B 48 h 52.1 39.1 8.9 62.8 ~atellazole C 48 h 49.6 41.5 8.9 59.2 48 exposure, the p53 null cells did not display a difference between the treated and control cells. However, after 48 h of exposure there was a great increase in both the percentage of cells in the S phase and in cell debris. This may indicate that p53 is an important early mediator of the cellular response to high-dose patellazole treatment. Additionally, as discussed above, after longer (48 h) exposure to patellazoles the HCT 116 wild type cells are able to undergo non-p53 mediated cell death. The ability of patellazole H to inhibit the cellular synthesis of macromolecules was examined using radiolabeled precursors. DNA damaging agents often cause an increase in thymidine uptake due to increased DNA repair, while topoisomerase inhibitors cause a decrease in both DNA and RNA synthesis.86 Treatment of cells with ribosome inhibitors such as MG 115 will cause a decrease in protein synthesis. A pulse-chase experiment was performed in which HeT 116 wild type cells were treated with 30 nM patellazole H for 3, 18 or 30 h before being exposed to the tritium­labeled precursors. 3H-thymidine, 3H-uracil and 3H-Iysine were used to observe DNA, RNA and protein synthesis, respectively. The level of radioactivity for each precursor was then determined and the amount of macromolecular synthesis compared between the treated and control cultures. The results are shown in Table 14. Protein synthesis began to decrease after only 3 h of exposure to 30 nM patellazole H (as compared to 65% that of the control culture) and continued to decrease to 13% that of the control culture after 30 h of treatment. RNA synthesis also decreased, although not as rapidly. The first inhibition of RNA synthesis was observed after 18 h of exposure (61 % that of the control culture) and continued to 30 h (23% that of the control 49 Table 14 Effect of Patellazole H on Macromolecular Precursor Incorporation precursor 3h 18 h 30h 3H-thymidine 1.02 2.80 1.22 3H-uracil 1.14 0.61 0.23 3H-Iysine 0.65 0.31 0.13 culture). DNA synthesis increased dramatically after 18 h of treatment to 280% that of the control culture and then returned to near the control rate of synthesis after 30 h of exposure. It is not clear whether the amount of DNA synthesis would have continued to decrease to less than that of the control culture if a longer time point had been observed. The effect of patellazole H on protein synthesis was further illustrated by a DNA microarray experiment. In this experiment~ HCT 116 wild type cells were treated with either 25 nglmL patellazole B (isolated by Ashantai Yungai) or DMSO (vehicle) for 18 hours. mRNA was then isolated from both treated and control cells and the mRNA submitted to the Microarray Core Facility at the Huntsman Cancer Institute (University of Utah). The Microarray Core Facility compares the relative amount of rrtRNA transcript through hybridization to cDNA spotted on a chip~ creating a profile of transcriptional activity in response to drug treatment. There were 74 genes for which a significant increase in transcription was observed. "Significanf~ was defined at having an increase in transcription at least 2 sigmas different from the mean of all gene changes.87 The genes encoding the proteins are listed in Table 15. Table 15 Treated VS. Control Ratios of Selected mRNA Transcripts gene product 60S Ribosomal protein L13 60S Ribosomal protein L6 60S Ribosomal protein, large P2 60S Ribosomal protein L32 60S Ribosomal protein Lll 60S Ribosomal protein L13A 60S Ribosomal protein LISA 60S Ribosomal protein LS 40S Ribosomal protein S5 40S Ribosomal protein S19 40S Ribosomal protein S9 40S Ribosomal protein S23 Apoptotic cysteine protease Mch4 Apoptotic cysteine protease Mch2 expression 5.31 5.60 5.67 6.16 6.23 6.S6 7.39 7.49 4.44 4.S4 5.69 6.06 3.46 4.32 In response to patellazole-induced protein synthesis inhibition, genes encoding translational machinery are increasingly transcribed. "Upregulated" here is used to describe genes which are present in the treated culture at least 2 sigmas greater than the 50 mean expression ratio of the treated cells VS. the control cells. Stated another way, each of the described genes is transcribed at least three times as much in the treated cells vs. the control cells. Twelve of the 74 genes observed to be upregulated after treatment of RCT 116 wild type cells with 25 nglmL patellazole B encode ribosomal proteins. (See Appendix B for the full list of upregulated genes.) This may be an attempt by the cell to compensate for the decrease in protein synthesis after patellazole treatment. Two other increasingly transcribed genes encode caspases involved in apoptosis. This, along with the observation of PARP cleavage after treatment and the accumulation of cell debris, illustrates the ability of the patellazoles to induce cell death. 51 The effect of the patellazoles on various cell signaling pathways, protein expression levels and protein phosphorylation states was examined through western blotting. A list of antibodies used and their source, dilution, etc., may be found in Chapter V (EXPERIMENTAL). The protocol for protein quatitation, nonnalization and loading may also be found there. Although treatment with the patellazoles resulted in arrest in the GoIG I and S phases of the cell cycle, the cyclins E, B 1, and D as well as CDKs 2 and 4 were not observed to be effected. The lack of effect by treatment with the ICso value of patellazole H on cyclin D is shown in Figures 22 and 23 as representative western blot. As described above, treatment with the patellazoles causes cell cycle arrest in GoIG1 followed by apoptosis. Additionally, the patellazoles effect intracellular macromoleculars through stimulating DNA synthesis and inhibiting protein synthesis. A review of the literature led to the investigation of the p70/mTOR pathway, a regulator of translation, as a potential target of the patellazoles. The three most well-known drugs that affect this pathway are raparnycin, wortmanin and L Y294002. Analogous to the patellazoles, rapamycin also inhibits protein translation and causes cells to arrest in GoIG1 in many cell types. Additionally, treatment of HCT wild type cells resulting in the upregulation of FK506-binding protein 1 to levels 4.4 times that of the vehicle-treated cells. As discussed in the introduction of this dissertation, rapamycin must bind to this protein in order to inhibit mTOR. 52 c H c H eyelin D 6h ISh Figure 22. The effect of patellazole H on the expression level of cyclin D in HCT 116 p53+'+cells. Treatment with patellazole H (H) after 6 h or 18 h as compared to the vehicle-treated control (C). c H c H eyelin D 6h ISh Figure 23. The effect of patellazole H on the expression level of cyclin D in HCT 116 p53-'-cells. Treatment with patellazole H (H) after 6 h or 18 h as compared to the vehicle-treated control (C). However, HCT 116 cells were relatively resistant to rapamycin compared to the patellazoles. The IC50 of rapamycin was observed to be 10 JLM in the HCT 116 wild type cell line used here, as compared to the low nanomolar ICsos for the patellazoles. Another difference is that rapamycin-induced 0 1 phase cell cycle arrest is mediated by a reduction is cyclin D 131 while treatment with the patellazoles does not affect the level of cyclin D 1 53 (Figures 22 and 23). There are similarities in the effect of these two drugs upon the p70/mTOR pathway, however. The following western blots utilized the ICso value of both patellazole Hand rapamycin; thus the cells were exposed to approximately 30 times more rapamycin than patellazole H. Rapamycin is known to strongly inhibit the phosphorylation of p70 at threonine 389.31 This position is phosphorylated by mTOR. The phosphorylation of serine 421 is under the control of the MAPK pathway 12 which has not been previously described to be affected by rapamycin. However, the western blot depicted in Figure 24 shows that rapamycin has the ability in HCT 116 cells to reduce the amount of phosphorylation at both sites, although to a lesser extent than at Thr 389. Patellazole H also inhibits the phosphorylation at Thr 389, though less quickly than rapamycin. Patellazole H does not appear to affect the phosphorylation state of Ser 421, nor does it effect the phosphorylation state of MAPK (Figures 25 and 26). This observation is reinforced by the previously discussed up-regulation of p53, as the inhibition of ERK activity results in a decrease in the half-life of p53. 19 The data indicate that the decrease in phosphorylation at Thr 389 is not mediated by signaling factors known to be upstream of p70. AKT is a key intermediary between growth factor receptors and p70 and was not affected by treatment with patellazole H (Figure 27). Rapamycin seems to have caused an initial increase in the amount of phosphorylation of AKT after 1 h of treatment. One hypothesis possibly explaining this observation is that a feedback mechanism occurs in which the cell is trying to compensate for the loss of signaling downstream of p70. p-p70 (Ser 421) p70 p-p70 (Thr 389) control pat H RAP Ih II , ·~ - ------ control pat H RAP 2h 54 Figure 24. The effect of patellazole H and rapamycin on p70. Examination of 30 nM patellazole H (pat H) or 10 /LM rapamycin (RAP) vs. DMSO (control) after either 1 h or 2 h exposure on the phosphorylation state and total expression of p70. c H C H p-MAPK Figure 25. The effect of patellazole H on p-MAPK. Examination of patellazole (H) treatment on the phosphorylation level MAPK after 6 h as compared to the vehicle­treated control (C) in HCT 116 p53+'+ and p53-'-cells. 55 p-MAPK control pat H RAP Figure 26. The effect of patellazole Hand rapamycin on p-MAPK. Examination of 30 nM patellazole H (pat H) and 10 JLM rapamycin (RAP) vs. DMSO (control) after either 1 h exposure on the phosphorylation state of MAPK. p-AKT control pat H RAP 1 hr control pat H RAP 2hr Figure 27. Patellazole H and rapamycin vs. AKT. The effect of 30 nM patellazole H (pat H) and 10 JLM rapamycin (RAP) vs. DMSO (control) after either 1 h or 2 h exposure on the phosphorylation state of AKT. The effect of both drugs is also observed downstream of p70. 4EBPI is a regulatory protein that binds to eIF4E when not phosphorylated. When phosphorylated, the binding is disrupted and eIF4E associates with a large protein complex that initiates translation. Thus the decrease in the phosphorylation of 4EBPI causes a decrease in translation. Rapamycin inhibits this phosphorylation after 1 hand 2 h of treatment; patellazole H begins to have an effect only after 2 h (Figure 28). Another branch of the p70 pathway involves the ribosomal protein S6 kinase. This protein also assists in protein translation but only when phosphorylated by p70. As p-4EBPl control pat H RAP 1 hr control pat H RAP 2hr 56 Figure 28. Patellazole Hand rapamycin vs. 4EBPI. The effect of 30 nM patellazole H (pat H) and 10 JLM rapamycin (RAP) vs. DMSO (control) after either 1 h or 2 h exposure on the phosphorylation state of 4EBPl. shown in Figure 29, rapamycin greatly inhibits this phosphorylation after 1 hand 2 h of exposure while patellazole H does not. However, after 6 h of exposure to patellazole H the phosphorylation of both 4EBPI and S6 kinase are decreased (Figures 30 and 31). Interestingly, the inhibition is sustained in 4EBPI after 18 h of drug exposure while the phosphorylation of S6 kinase actually increases above that of the control levels. This may indicate that patellazole H preferentially affects one branch of the p70 pathway. p-S6 Kinase control pat H RAP 1 hr "i;b" control pat H RAP 2hr Figure 29. Patellazole Hand rapamycin vs. S6. The effect of 30 nM patellazole H (pat H) and 10 JLM rapamycin (RAP) vs. DMSO (control) after either 1 h or 2 h exposure on the phosphorylation state of S6 kinase. 57 c H c H p-4EBPI 6h ISh Figure 30. The effect of 30 nM patellazole H (H) vs. DMSO (C) on the phosphorylation state of 4EBP1. c H c H p-S6 6h ISh Figure 31. The effect of 30 nM patellazole H (H) vs. DMSO (C) on the phosphorylation state of ribosomal protein S6 kinase. Treatment of HCT 116 wild type cells with patellazole B produced a similar effect upon the p70 pathway (Figure 32). An ICso amount of patellazole B (0.39 nM) produced a slightly weaker effect than 30 nM of patellazole B or H. This may indicate that the inhibition of the p70 pathway is a secondary effect rather than the primary target of the patellazoles. Similar results were observed when examining the effects of the same treatments upon the phosphorylation state to 4EBPI (Figure 33). p-p70 p-p70 p-p70 control pat B pat B ICso 30 nM control pat B pat B ICso 30 nM control pat B pat B ICso 30 oM patH ICso patH ICso patH ICso 58 Figure 32. Patellazoles Band H vs. p70. The effect of patellazole B (pat B) and patellazole H (pat H) vs. DMSO (control) after 2 h (top), 6 h (middle) or 24 h (bottom) of exposure to drug on the phosphorylation state of p70. p-4EBPI p-4EBPI p-4EBPI control pat B pat B ICso 30 nM control pat B pat B ICso 30 nM control pat B pat B ICso 30 nM patH ICso patH ICso patH ICso 59 Figure 33. Patellazoles Band H vs. 4EBPl. The effect of patellazole B (pat B) and patellazole H (pat H) vs. DMSO (control) after 2 h (top), 6 h (middle) or 24 h (bottom) of exposure to drug on the phosphorylation state of 4EBPl. 60 Tariet Identification Although the above data provide a description of the response of HCT 116 cell to the patellazoles, the molecular target is still unknown. Several approaches may be used to identify the cellular target of a cytotoxin. These include genetics (especially using model organisms), affinity chromatography, photoaffinity labeling, and high throughput screening. Two methods were used in attempts to identify the targets of the patellazoles, namely yeast mutation analysis and affinity chromatography using biotin-streptavidin to link a patellazole to a solid phase support. The signaling pathways that control translation in mammalian cells are generally well conserved in yeast. TOR was first identified as the Target Of Rapamycin using yeast genetics. 1 0 This protein is conserved in mammalian cells and is thus known as mTOR.ll The ability of patellazole B to affect yeast was examined as a first step to using a genetic approach to identify the molecular target of the patellazoles. A serial dilution of patellazole B up to approximately 10,000 times the ICso in mammalian cells was applied to two yeast strains (W303 and S288C) grown in suspension. However, even 1 rnM patellazole B did not result in any reduction in culture growth after four doubling times as measured by optical density (600 nm). It is not yet known whether this indicated that the mammalian target of the patellazoles is not conserved in yeast or if patellazole B is not able to cross the yeast cell wall. Synthetic studies were undertaken in an effort to link patellazole B to a solid phase support for use in affinity chromatography. Although there are differences in the cytotoxicities of patellazole Band patellazole H, opening the epoxide of the natural products appears to only minimally affect the bioactivity. Therefore attempts were made 61 to link a biotin moiety to patellazole B through the epoxide. Since the analogs appear to have been formed from silica-catalyzed solvent attack on the epoxide, attempts were made to duplicate this reaction under controlled condi tions (Figure 34). Biotinamidocaproate N-hydroxysuccinimide ester was reacted with 4- aminobutanol, producing biotinamidocaproate N-butanol amide in nearly quantitative yields. The biotinamidocaproate N-butanol amide was purified from the reaction mixture using reverse phase HPLC. In an attempt to react the free hydroxyl of the biotinamidocaproate N-butanol amide with the epoxide of patellazole C, a slight excess of biotinamidocaproate N-butanol amide was combined with patellazole C with 200 mg of silica gel and stirred overnight at room temperature. The reaction was periodically checked using ESI mass step A stepB ~-.. Figure 34. Reaction scheme 1. Steps A and B vary and are described in the text. R = patellazole. 62 spectrometry and after 1 week no reaction had occurred. TF A (30 ILL) was then added to the reaction mixture and the reaction stirred at room temperature. After 24 h there appeared to be a slight degradation of the patellazole by ESI mass spectrometry but none of the desired reaction product. Because the biotinamidocaproate N-butanol amide may have been sterically hindered from reacting with the patellazole epoxide, an alternative reaction scheme was attempted (Figure 35). In scheme 2 a 10-fold molar excess of 4-aminobutanol was combined with patellazole C and silica for 72 h at room temperature with stirring. Analysis by ESI revealed a large peak at MH+ = 1009, which agrees with the mass of the desired reaction product (the free hydroxyl having attacked and open the epoxide). However, all attempts to react this compound with biotinamidocaproate N­hydroxysuccinimide ester failed. Infrared comparison of patellazole C and the reaction product, along with extensive HPLC and mass spectrometry experiments, revealed that no reaction had actually occurred. Variations upon the above reactions were also attempted utilizing TF A and/or acetic acid, different solvents, extended reaction times and slight heating. The most common result was no reaction or patellazole degradations. Thus another reaction scheme was designed. Most commonly, thiols have been used to attack and open epoxides under basic conditions.S1 In reaction schemes 3 and 4 (Figures 35 and 36), 2-mecraptoethlyamine was substituted for 4-aminobutanol in the above reactions. Reaction scheme 3 combined 25 mg of biotinamidocaproate N-hydroxysuccinimide ester and 4.7 mg of 63 R,A +H~ ~ NH2 step A .. stepB Figure 35. Reaction scheme 2. Steps A and B vary and are described in the text. step A • stepB .. Figure 36. Reaction scheme 3. Steps A and B vary and are described in the text. 64 )L\+ H( _st_e_p_A_ _• JL<0H step B .. Figure 37. Reaction scheme 4. Steps A and B vary and are described in the text. 2-mercaptoethylamine in 2 mL 9: 1 CHCI3:MeOH at room tenlperature with stirring resulted in both the desired reaction product (biotinamidocaproate N-mercaptoethylamide) and a dimer resulting in the formation of a disulfide bridge. An attempt was made to reduce the disulfide bond using triphenylphosphine and TFA in 1:1 THF:H20 at 40° C. Reduction of the disulfide did occur after 6 h as observed by ESI mass spectrometry. The monomer form of biotinamidocaproate N-mercaptoethylamide was reacted with patellazole B under basic conditions. One mg of patellazole B was combined with 0.97 mg of the monomer produced in the reaction described above (1:2 molar ratio) in 1 mL 1:1 THF:0.1 M NaHC03 and stirred at room temperature. A second reaction was also attempted in which 1.2 mg of Ph3P was added to an otherwise identical mixture. After 8 days of reaction time no significant reaction had occurred with either scheme. Heating the reaction mixtures to 50° C overnight did not alter the results. 65 Two additional reactions were attempted in which the sulfide dimer was substituted for the monomer in the scheme described above. Again, no reaction occurred in either reaction mixture and heating the reaction mixtures to 50° C overnight did not improve the reaction. A final attempt to open the epoxide in a controlled manner was made by directly combining patellazole Band mercaptoethylamine. Two similar reactions were performed. Each contained 1 mg patellazole Band 85 mg mercaptoethylamine (1:1000 molar ratio) in 1 mL 1: 1 THF:O.l M NaC03H. One reaction was incubated at room temperature for 72 h; the other was incubated at 70° C for 6 h followed by 32° C overnight (both with stirring). The room temperature reaction did not appear to contain the desired product by ESI. The heated reaction displayed two major ESI peaks corresponding to the mass of the desired product (mlz = 1003) and patellazole B, as well as a smaller peak at mlz = 1078. This mass corresponds to a conserved addition of mercaptoethylamine (such as opening the epoxide or the reduction of a carbonyl) plus the formation of a disulfide bridge with another mercaptoethylamine unit (903 + 77 + 77 - 2 + 23 (sodium peak) = 1078). Thus whether the amine or the thiol portion of the mercaptoethylamine reacted with patellazole B is unclear, but the possible formation of a disulfide bridge hints that the amine may have reacted before the thiol under these conditions. 66 Discussion The Ireland research group has analyzed L. patella collected from the Fijian Islands between 1984 and 2001. The collections performed in November 1984, October 1987, December 1988 and July 1997 at or near the Astrolobe Reef in the Fijian island group of Kandavu. All four collections yielded large amounts of patellazoles. Recollections performed at the this site in January and August of 2001 failed to yield any patellazoles. It was considered that this variation may have been seasonal and cycle with an environmental factor such as nutrient availability, temperature, predation, or there may be a parallel seasonal variation among the Prochloron symbionts. Further collections were performed to test this hypothesis. Collections performed in July and September of 2001 from Seashell Cove in Navula Pass did produce patellazoles in yields similar to those from the 1984 to 1997 Kandavu collections. Although this is a relatively small sample size, it appears that a localized event occurred in the Astrolobe Reef between July 1997 and January 2001 that resulted in the loss of patellazole biosynthesis in the L. patella population. The nature of this event is unknown, however. Further characterizing the biosynthesis of the patellazoles and identifying factors that influence their production will be important both as basic research and in the production of these potentially useful compounds. Especially interesting is the role of the Prochloron symbiont in L. patella biosynthesis as these organisms are known to exchange material to an unknown extent. Understanding this symbiotic relationship may provide answers to the biosynthetic questions posed above and could potentially be a tool in the further investigation of patellazole chemistry. 67 The patellazoles are highly cytotoxic thiazole-containing macrolides with a range of bioactivity. Opening the epoxide by solvent addition does not greatly alter the bioactivity of the patellazoles. This reaction was previously observed when patellazole C reacted with methanol after remaining in 85: 15 methanol:water for 3 days. It is not clear what caused any of these reactions to occur, because water/methanol mixtures were used in other experiments without incident. Isolation of the adducts proved difficult and the compounds were isolated as only 0.00085% (w/w), 0.00035% (w/w) and 0.0003% (w/w) mass of the dry organism, respectively. The formation of patellazoles H, I and J from patellazoles A and B decreased cytotoxicity by either 10 X or 100 X, depending upon the natural product and solvent involved. Patellazole J, the isopropanol adduct of patellazole B, showed the next greatest cytotoxicity, followed by patellazole C, patellazole I and patellazole H. Patellazole J, the addition of isopropanol to patellazole B, is approximately an order of magnitude less cytotoxic as compared to the natural product. Patellazole H, the addition of methanol to patellazole B, is approximately two orders of magnitude less cytotoxic than the natural product. The addition of methanol to patellazole A, creating patellazole I, causes an order of magnitude loss of cytotoxicity. It is not known whether these shifts in cytotoxicity are due to decreased target affinity or to alterations in the cellular uptake or distribution of the patellazoles. Three major responses to treatment with the patellazoles were identified. First, from a general cellular perspective, treatment with nanomolar concentration of the patellazoles resulted in the activation of apoptotic pathways and cell death. Second, both the natural products and the analogs had the ability to arrest HCT 116 cells in the GJGJ 68 phase of the cell cycle, followed by either an increase in the S phase population or apoptosis, depending upon treatment dose and duration. Interestingly, the patellazoles are much more cytotoxic toward the HCT 116 colon tumor cell lines than the Cac02 colon tumor cell line. Elucidating the reason behind this difference may provide a clue to the mechanism of action of these compounds. The third major cellular response to the patellazoles is the strong effect on macromolecule synthesis. HCT 116 wild type cells displayed increased DNA synthesis after treatment with patellazole, as has been previously described.78 RNA synthesis decreased over time after treatment. Protein synthesis was dramatically inhibited, dropping to 31 % of control cells after an 18 h treatment. The effect of the patellazoles on protein synthesis is further illustrated in a DNA microarray experiment. Of the 74 genes observed to be increasingly transcribed after treatment, 12 encode ribosomal proteins. Two others encoded caspases involved in apoptosis, which, along with the observation of P ARP cleavage after treatment and the accumulation of cell debris, illustrates the ability of the patellazoles to induce cell death. The deletion of p53 from the HCT 116 wild type cell line did not alter the overall cytotoxicity of the patellazoles. However, the amount of p53 did increase in wild type cells after treatment and the deletion of p53 decreased the amount of cellular debris after treatment. Additionally, DNA synthesis increases after patellazole treatment; p53 is a major mediator of cellular responses to agents which effect DNA. p21 null HCT 116 cells were approximately three times more sensitive to the patellazoles than the wild type cells and the deletion of p21 abrogated GclG 1 arrest. 69 One pathway that plays a major roll in regulating transcription in mammalian cells is the p70/mTOR pathway. In response to extracellular stimuli, a phosphorylation cascade occurs that activates ribosomal transcription. The patellazoles are able to inhibit this pathway at several steps, although the primary cellular target is not known. The inhibition point appears to be either downstream of AKT or in a parallel pathway. Downstream proteins 4EBP 1 and S6 kinase are both affected by patellazole treatment, as well as the central protein p70 S6 kinase. Rapamycin is a large hydroxylated macrolide that exhibits similar effects upon the p70/mTOR pathway in RCT 116 cells. Interestingly, treatment of HCT wild type cells results in the increased transcription of the rapamycin binding partner FKBP 1 to levels 4.4 times that of the vehicle-treated cells. However, HCT 116 cells are much more resistant to rapamycin than to the patellazoles. The IC50 of rapamycin growth inhibition is 10 ILM, 100 to 10,000 x less cytotoxic than the patellazoles in the HeT 116 cell lines. Rapamycin more strongly inhibits the p70 pathway, but is not as cytotoxic, implying differing targets or mechanisms of action for these two compounds. Rapamycin and the patellazoles do share many similar characteristics, however, and it would be interesting to further compare the biological activity of these two compounds. The target and mechanism of rapamycin was first described in yeast. Surprisingly, patellazole C is not able to inhibit the growth of yeast strains W303 and S288C. Possible mechanisms of resistance include the exclusion of patellazole C by the yeast cell wall, increased metabolism of patellazole C, or the mammalian target not being conserved in yeast. Further investigation into the nature of the yeast-patellazole interaction may provide clues to the mammalian target or targets. 70 One of the most direct methods to determine a cellular target is to use affinity chromatography. Neither a complete nor a partial synthetic route to any patellazole is currently known and therefore the attachment of patellazole to a solid phase support will have to be accomplished through the derivation of the natural product. The formation of the solvent adducts was somewhat fortuitous in that it has now been demonstrated that opening the patellazole epoxide does not abrogate or greatly alter the bioactivity of the molecule. The epoxide is one of the few unique portions of the patellazoles; the presence of multiple chemically-similar hydroxyls and carboxyl functionalities makes selectively reacting one of these functional groups infeasible. Unfortunately, the direct attack of a thiol or hydroxy I into the epoxide proved infeasible. Further complicating matters, the presence of a free amine in the nucleophile appears to cause the formation of side products under common epoxide-opening conditions. The patellazoles continue to be both chemically and biologically interesting. Further chemical studies could include the elucidation of a crystal structure, which would provide both the stereochemistry of the 16 chiral carbons and information regarding global patellazole conformation. This information would be crucial in developing a synthetic route to the patellazoles, which would allow both more extensive biological testing and a potentially better method for accomplishing affinity chromatography or photoaffinity labeling. A synthetic scheme would also allow a library of patellazole­derived compounds to be synthesized for use in determining structure-activity relationships for the patellazoles and their cellular target or targets. A further evaluation of the cytotoxicity of the patellazoles in a variety of cell lines may provide a starting point for further mechanism of action studies as cell line 71 genotypes and phenotypes are increasingly well described. Genetic approaches to determining the target of cytotoxins are becoming increasingly powerful. Examining the effect of the patellazole in model organisms such as yeast or Drosophila may provide infonnation about how the patellazoles interact with mammalian cell components. Defining the molecular target of the patellazoles and understanding the mechanism of action of these compounds is a first step toward evaluating them as a potential anticancer treatment. CHAPTER III LISSOCLINOLIDE Back1:round Lissoclinolide was first reported from L. patella in 1990 and is the first nonnitrogenous secondary metabolite identified from this organism.77 Several complete syntheses of lissoclinolide have been described, with varying yields.88-91 There has been some debate over the relationship between lissoclinolide and tetrenolin, which was identified as a metabolite of the fungus Micropolyspora venezuelensis in 196992 and as an Actinomycetales metabolite in 1973.93 These two compounds were speculated to have different configurations at the /14 ,5 double bond, partially due to reported differences in their biological profiles. Lissoclinolide was reported to be active against only Gram negative Escherichia coli77 whereas tetrenolin was reported active against only Gram positive bacteria.92 Recently, Gorth and Bruckner stereoselecti vely synthesized both compounds and showed that tetrenolin has the same stereochemistry as lissoclinolide, but the stereoisomer with the E configuration is a different compound.89 Interestingly, Gorth and Bruckner's biological assays described the Z configured compound (lissoclinolide/tetrenolin) as possessing antibiotic activity against both Gram positive and Gram negative bacteria, while the E configuration had only marginal activity. 73 Chemistry L. patella collected from the Astrolobe Reef in the Fijian Islands in January, 2001. The animal material was lyophilized resulting in 200 g dried material. The chloroform solvent partition separated by elution with ethyl acetate and methanol over a silica solid phase. The front of the 9: 1 EtOAc:MeOH fraction yielded pure lissoclinolide. Ultimately, the 200 g dry L. patella yielded 279.3 mg of lissoclinolide(0.14% (w/w». The structure of lissoclinolide was confirmed through NMR and mass spectrometry by comparison to the literature data.?7 The molecular formula is CllHI20 4, resulting in a molecular mass of 208 AMU, and the structure is shown in Figure 38. The initial purification steps for lissoclinolide were the same as for the patellazoles - solvent partitioning followed by a methanol:ethyl acetate silica flash column. When either compound was isolated, it was easily observable in proton spectra of the silica column fractions. Interestingly, lissoclinolide and the patellazoles were not co-observed in the four collections performed in 2001 and investigated during this course of this dissertation. HO OH o o Figure 38. Lissoclinolide. 74 Lissoclinolide was first described by Dr. Bradley Davidson in 1990.77 L. patella collected in the Yasawa Island group, Fiji, which is located northwest of Viti Levu. In this investigation, 218 g of lyophilized animal yielded 20 mg of lissoclinolide but the patellazoles were not reported. Although there are not enough data at the current time to make a conclusive statement, it appears that lissoclinolide production may be exclusive of patellazole production. Due to the lack of information regarding the biological activity of lissoclinolide in mammalian cells pharmacological studies were performed. Biolo~y The ability of lissoclinolide to inhibit the growth of mammalian cells was examined using an MTT assay. This assay measures the inhibition of cellular growth as observed by the relative activity of mitochondrial succinate dehydrogenase to reduce MTT to an insoluble purple formazan dye. The average IC50 of lissoclinolide against nine cell lines was 395 nM. Lissoclinolide showed similar cytotoxicity against both p53 competent and null HCT 116 cell lines and was approximately twice as cytotoxic toward the p21 null HCT 116 line compared to the p21 competent parental cell line. The PTEN null MDA-MB-468 breast tumor cell line was approximately twice as resistant to lissoclinolide as the parental PTEN competent MDA-MB-435S cell line. A two-fold decrease in cytotoxicity is a relatively small change; thus this difference is most likely due to the fact the loss of PTEN results in generally more rapid cell growth rather than the deletion effecting the specific activity or target of lissoclinolide. These IC50s are shown in Table 16. 75 Table 16 ICso Values of Lissoclinolide Versus Various Cell Lines cell line ICsQ(nM) HCT 116 p53+'+ 339 HCT 116 p53-'- 526 HCT 116 p21 +1+ 423 HCT 116 p21-'- 173 MDA-MB-435S 258 MDA-MB-468 596 HeLa 194 293 509 Cac02 539 Treatment of HCT 116 p53+'+ with 2.4 p,M lissoclinolide resulted in a strong OzlM arrest after 24 h of exposure and was sustained through 48 h of exposure. Interestingly, there was not a large increase in cellular debris after even 48 h of treatment at this dose. Treatment with 481 nM lissoclinolide resulted in a much smaller GzlM arrest and a greater accumulation of cells in S phase after 24 h of treatment. The S phase population decreased to control levels after 48 h exposure to 481 nM lissoclinolide. Additionally, more cell debris was observed at this lower dose than at the higher 2.4 ILM dose. Thus data are presented in Table 17. As shown in Table 18, the overall effect of both doses of lissoclinolide was similar in HCT 116 p53-'- cells. The GzlM arrest was observed with 2.4 ILM treatment; S phase arrest and the accumulation of cell debris was observed with 481 nM treatment. There was more cell debris observed after treatment of the p53 null cells as compared to 76 Table 17 Effect of Lissoclinolide on the Cell Cycle in HCT 116 p53+'+ Cells treatment dose duration G/Gt S G.i. M debris control 24h 63.10 26.95 9.95 13.55 lissoclinolide 481 nM 24h 44.99 41.24 13.77 22.27 lissoclinolide 2.4#tM 24h 18.27 25.15 56.58 6.21 control 48h 71.2 16.09 12.71 17.53 lissoclinolide 481 nM 48 h 63.10 13.46 23.44 38.77 lissoclinolide 2.4#tM 48 h 20.32 27.71 51.97 8.36 Table 18 Effect of Lissoclinolide on the Cell Cycle in HCT 116 p53-'- Cells treatment dose duration G/Gt S G.i. M debris lissoclinolide 481 nM 24h 31.82 49.32 18.86 48.16 lissoclinolide 2.4#tM 24h 20.86 28.15 50.98 26.84 control 48h 52.72 24.84 22.44 37.47 lissoclinolide 481 nM 48h 52.57 30.77 16.66 17.93 lissoclinolide 2.4#tM 48h 24.87 32.63 42.50 29.48 the p53 competent cells, however, indicating that lissoclinolide-induced cell death is most likely p53 independent. Western blotting presented in Figure 39 revealed that after 24 h of 481 nM lissoclinolide treatment poly(ADP-ribose) polymerase was not cleaved. The cleavage of PARP from its native 116 kD form to 85 kD and 31 kD fragments is an early event during caspase-mediated apoptosis.84 This supports the flow cytometry data presented 77 C L C L PARP Figure 39. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on P ARP cleavage. above as p53 is often an activator of caspase-mediated apoptosis. The effect of 2.4 p.M is not clear, however, because it was difficult to harvest enough protein when HCT 116 cells were dosed at that level. This may be due to either growth inhibition and stasis, apoptosis or protein degradation. Figure 40 shows the effect of 481 nM lissoclinolide on ubiquitin in HCT 116 wild type cells. Since this treatment did not cause significant changes in ubiquitin levels or distribution, the proteosome is most likely not significantly involved in the response of HCT 116 cells to lissoclinolide treatment. Additionally, the low amount of protein observed in 2.4 p.M treated cells is most likely due to growth inhibition rather than protein degradation by the proteosome. The noninvolvement of p53 in the response of HCT 116 cells to lissoclinolide is further illustrated in the western blot shown in Figure 41. The treatment of HCT 116 wild type cells with 481 nM of lissoclinolide did not affect the levels of p53 within the cells. The p53 null cells did not produce any p53, of course. 78 c L c L Vb Figure 40. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on ubiquitin distribution. c L C L p53 Figure 41. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on p53 and expression levels. 79 p21 null RCT 116 cells were slightly more sensitive to lissoclinolide than the HCT 116 wild type cells. Although p21 often plays a role in G/M arrest, treatment with 481 nM lissoclinolide did not result in p21 accumulation (Figure 42). The extremely low levels of p21 present in the p53 null RCT 116 cultures is not surprising considering that p53 is a major transcriptional regulator of p21. Because treatment of HCT 116 cells with lissoclinolide did not result in caspase­mediated cell death, the possibility of cell-rescue from lissoclinolide treatment was examined. After 24 h of incubation with lissoclinolide as described above, treated HCT 116 wild type cells were washed three times with 37° C versene in an attempt to remove lissoclinolide from the cells. After another 24 h of incubation with fresh media, the washed cells were harvested and submitted for FACS analysis. The G/M arrest was slightly decreased compared with the nonwashed control cultures but was maintained in the majority of the arrested cells (Table 19). A significant amount of cell debris was observed in the 481 nM treated cell culture even after the wash, and again the amount of debris was greater than that in the 2.4 #LM treated cells. Whether lissolinolide was successfully removed from the cells and the effect is irreversible or lissoclinolide could not be washed out is unknown; regardless, the cells arrested by lissoclinolide treatment are not easily rescued after 24 h of exposure to the drug. Treatment of multiple HCT 116 colon tumor cell lines with 2.4 #LM lissoclinolide results in a strong G/M arrest. This cell cycle interruption is characteristic of a mechanism of action involving tubulin. However, biochemical assays showed that lissoclinolide neither inhibited or stabilized microtubule interactions.94 80 c L C L p21 Figure 42. The effect of 481 nM of lissoclinolide (L) vs. DMSO (C) after 24 h exposure on p21 expression levels. Table 19 Reversibility of Lissoclinolide Induced Cell Cycle Arrest treatment dose duration GIG, S GiM debris control 24 h 63.10 26.95 9.95 13.55 lissoclinolide 481 nM 24h 68.01 24.55 7.44 23.87 lissoclinolide 2.4 J1.M 24 h 27.75 36.83 35.42 25.77 control 48 h 59.91 25.47 14.61 16.41 lissoclinolide 481 nM 48 h 56.92 24.70 18.39 49.65 lissoclinolide 2.4 J1.M 48 h 30.98 49.79 19.22 13.64 control 48 h 69.51 22.50 7.99 7.91 washout 481 nM 48 h 50.23 25.94 15.87 25.24 washout 2.4~M 48 h 29.80 45.28 24.92 5.66 Another molecular mechanism that would result in GiM arrest is the inhibition of ubiquitin isopeptidases. The conjugated double bonds exocyclic to the lactone ring may potentially act as Michael receptors. The ability of alpha-beta conjugatedlactones to act as electrophiles has been implicated in the inhibition of ubiquitin isopeptidases and the proteosome in general.95 However, lissoclinolide was also shown to not inhibit ubiquitin isopeptidases in an in vitro assay. Lissoclinolide was screened in the Developmental Therapeutics Program's in vitro 60 cell line antitumor panel at the NCI. Briefly, lissoclinolide was tested in 48 h continuous doses at multiple concentrations and cell viability/growth was quantitated 81 using a sulforhodamine B protein assay. 96 This assay compares the amount of total protein after 48 h of drug treatment to the amount of proteins in both the control cells and the cells present at the time of drug administration. COMPARE analysis of lissoclinolide data revealed a broad range of activity against the 60 tumor cell line panel at the NCI, marked by a general ability to inhibit cellular proliferation but a limited ability to cause cell death. Lissoclinolide inhibited the growth of the cell lines in the colon tumor panel more than the other tumor cell lines. Lissoc1inolide was approximately 10 times more efficacious in the colon tumor cell line COLO 205 compared to the mean total growth inhibition (TGI) observed in the other 59 cell lines. The colon tumor cell lines HCC-2998, HCT -116 and HCT -15 were also more sensitive than average. The leukemia, CNS and breast tumor cell lines were the least growth inhi bi ted. Using the TGI results, a search of the Molecular Target Database yielded the strongest correlation with MT30, in which ErbB2 (receptor protein-tyrosine kinase) had been disrupted (PCC = 0.619; number of cell lines = 34; variance = 0.029). The correlation is not particularly strong, however. Additionally, the lissoclinolide TGI profile did not correspond well to any known mechanisms of action or activity of any standard drugs. Panel/CeO Line Leukemia CCRF-CEM HL-6O(TB) K-562 MOLT-4 RPMI-8226 SR Non-Small Cell LWlg Cancer A549/ATCC EKVX NCI·H226 NCI-H23 NCI-H322M NCI-H460 NCI-H522 Colon Cancer COLO 205 HCC-2998 HCf-1l6 HCT-15 HT29 KMI2 SW-620 CNS Cancer SF-268 SF-295 SF-539 SNB-19 SNB-75 U251 Melanoma LOXIMVI MALME-3M M14 SK-MEL-2 SK-MEL-28 SK-MEL-5 UACC-62 Ovarian Cancer fGROVl OVCAR-3 OVCAR-4 OVCAR-5 OVCAR·8 SK-OV-3 Renal Cancer 786-0 ACHN CAKf-l RXF393 SN12C TK-IO UO-31 Prostate Cancer PC-3 DU-145 Breast Cancer MCF7 NCUADR-RES MDA-MB-2311 ATCC MDA-MB-435 MDA-N BT-549 T-47D Log 0 GI50 -4.64 -4.53 -4.51 -4.30 -4.79 -4.46 -4.44 -4.81 -4.44 -4.74 -4.51 -4.49 -4.99 -5.65 -4.85 -4.91 -4.64 -4.43 -4.74 -4.61 -4.41 -4.62 -4.63 > -4.00 -4.39 -4.45 -4.86 -4.69 -4.71 -4.61 -4.66 -4.72 -4.80 -4.62 -4.34 -4.70 -4.72 -4.21 -4.90 -4.62 -4.84 -4.72 -4.55 -4.72 -4.78 -4.84 -4.68 -4.44 -4.48 -4.67 -4.45 -4.67 -4.70 -4.68 -4.64 1.02 1.65 Gl50 +3 +2 +1 o ·1 ·2 Figure 43. The effect of lissoc1inolide upon the NCI 60 cell line panel. 82 -3 Panel/CeU Lioe Leukemia CCRF-CEM HL-60(TB) K-562 MOLT-4 RPMI-8226 SR Non-Small Cell Lung Cancer A549/ATCC EKVX NCI-H226 NCI-H23 NCI-H322M NCI-H460 NCI-H522 Colon Cancer COLO 205 HCC-2998 HCT-116 HCT-15 1IT29 KM12 SW-620 CNSCancer SF-26S SF-295 SF-539 SNB-19 SNB-75 U251 Melanoma LOX IMVI MALME-3M M14 SK-MEL-2 SK-MEL-28 SK-MEL-5 UACC-62 Ovarian Cancer IGROVI OVCAR-3 OVCAR-4 OVCAR-5 OVCAR-8 SK-OV-3 Renal Cancer 786-0 ACHN CAKI-l RXF393 SN12C TK-IO UO-3f Prostate Cancer P<>3 DU-145 Breast Cancer MCF7 NCIIADR-RES MDA-MB-231/ATCC MDA-MB-435 MDA-N BT-549 T-47D Figure 43. Continued. Log () TGI -4.22 -4.16 > -4.00 > -4.00 -4.32 -4.15 TGI ............................................................................................................................................ -... > -4.00 -4.45 -4.02 -4.38 -4.02 > -4.00 -4.65 .............................................................................. ---- ................................................................ . -5.29 -4.70 -4.52 -4.51 -4.05 > -4.00 -4.34 ..................................... -"to ................ _. _,..* ........ ....... ........ • ................................. _ ........ _ ....................... .. -4.24 > -4.00 -4.26 -4.29 > -4.00 > -4.00 .............................. -.............................................. -........... . ............................. -_ .... _ .... "' ...................... . > -4.00 -4.49 -4.37 -4.36 -4.30 -4.34 -4.40 ............................................................................................................................. -4.41 -4.31 > -4.00 -4.38 -4.41 > -4.00 , ............................................................... .. .............................................. .. -4.47 -4.16 -4.55 -4.42 -4.09 -4.39 -4.5 1 ............................................................................ . ................................................................. ... -4.37 -4.29 ........................................... _ .............. _ .................. -.................................... -.......................... --.. . > -4.00 -4.05 -4.28 > -4.00 -4.22 -4.29 -4.29 ..... -.......................................... -........................................................................................................ .. -4.27 1.02 1.29 +3 +2 +1 o ·1 ·2 ·3 83 Panel/CeU Line Leukemia CCRF..cEM HL-60(TB) K-S62 MOLT-4 RPMI-8226 SR Non-Small Cell LlUlg Cancer A549/ATCC EKVX NCI-H226 NCI-H23 NCI-H322M NCI-H460 NCI-H522 Colon Cancer COLO 205 HCC-2998 HCT-116 Her-15 HT29 KMI2 SW-620 CNS Cancer SF·268 SF-295 SF-539 SNB-19 SNB-75 U251 Melanoma LOXIMVI MALME-3M M14 SK-MEL-2 SK-MEL-28 SK-MEL-S UACC-62 Ovarian Cancer IGROVl OVCAR-3 OVCAR-4 OVCAR-5 OVCAR-8 SK-OV-3 Renal Cancer 786-0 ACHN CAKI-l RXF393 SN12C TK-IO UO-31 Prostate Cancer PC-3 DU-145 Breast Cancer MCF7 NCUADR-RES MDA-MB-231/ATCC MDA-MB-435 MDA-N BT-549 T-47D MG_MID Delta Range Figure 43. Continued. LothoLCSO > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 -4.09 > -4.00 -4.02 > -4.00 > -4.00 -4.31 -4.74 -4.31 -4.20 -4.11 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 -4.12 -4.04 -4.01 > -4.00 -4.02 -4.08 -4.03 -4.01 > -4.00 -4.06 -4.11 > -4.00 -4.04 > -4.00 -4.26 -4.11 > -4.00 -4.05 -4.24 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 > -4.00 -4.05 0.68 0.74 I +3 I +2 I +1 LCSO -~ o 84 I I ·1 ·2 ·3 85 Discussion Marine organisms, especially sponges and ascidians, continue to yield bioactive compounds. Lissoclinolide is an interesting example of such a molecule in that it is both non-nitrogenous and has also been isolated from terrestrial sources; neither of which is common for an ascidian secondary metabolite. Although most notably possessing antimicrobial activity, this compound has now been shown to also act upon human tumor cell lines. The increased cytotoxicity of lissoclinolide toward colon tumor-derived cell lines compared to other tumor cell lines reveals that this compound acts preferentially on a cellular target or class of targets. Further investigation into the mechanism of action of this compound may reveal not only its target but insight into the nature of colon tumors and potential cancer treatments. It is also interesting to note that the biosynthetic production of lissoclinolide appears to vary inversely with patellazole production. Lissoclinolide was not observed in the patellazole-yielding L. patella collections described in Chapter II. A review of the available data from previous collections (1984 through 1987), it does not appear that lissoclinolide was prevalent in these patellazole-rich samples. The relationship between the biosynthesis and function of these two molecules in L. patella is intriguing. CHAPTER IV CONCLUSIONS The investigation of the chemical and biological characteristics of L. patella metabolites has relevance to two seemingly unrelated scientific fields - marine ecology and anticancer drug discovery. Although there are not enough data at the current time to make a conclusive statement, it appears that patellazole production and lissoclinolide production may be mutually exclusive. This raises interesting questions regarding the biosynthesis and function of these compounds within L. patella. One hypothesis is that lissoclinolide, a relatively simple molecule, is less biosynthetically demanding to produce than the complex patellazoles and is created during times of stress. This variation could be seasonal and cycle with an environmental factor such as nutrient availability, temperature, predation, or there may be a parallel seasonal variation among the Prochloron symbionts. The loss of patellazole production in L. patella in the Astrolobe Reef appears to have been an isolated event between 1997 and 200 1. It will be interesting to see if future collections from this site yield patellazoles or if the abrogation of patellazole biosynthesis in this L. patella population is a permanent event. Further characterizing the biosynthesis of both the patellazoles and lissoclinolide, while identifying factors that influence production, will be interesting both as basic research and as a production method for 87 potentially useful compounds. Discovering how the seasonal biosynthetic variation occurs could provide insight into the biochemistry and genetics of marine biosynthesis; discovering why this occurs may provide answers to ecological questions regarding coral reefs. Although the complexity of marine biosynthesis may seem an obstacle to working with a specific natural product, the same complexity is an overall strength in that a wide range of natural products are available for investigation. This study of L. patella illustrates several aspects of modem cancer drug discovery. Marine natural products continue to be a source of biologically active compounds. Ascidians seem to be particularly prolific sources of molecules which are able to modulate mammalian cellular activity. This may be due to the fact that, like mammals, ascidians are members of phylum Chordata. A review of marine natural products reveals that phylogenitically similar organisms often produce biosynthetically similar compounds. Also, phylogenitically related organisms are likely to have similar genetics. Therefore it is not surprising that animals closely related to mammals would feasibly produce metabolites that are biologically active within mammalian cells. The patellazole family of compounds continues to be a medically intriguing class of molecules. The patellazoles are potent modulators of human colon tumor cell growth and cell cycle progression, as well as DNA and protein synthesis. This activity seems to be related to the ability of these compounds to effect the p70/mTOR pathway, a pathway that regulates protein synthesis and cell growth. The molecular target of these compounds continues to elude identification, however. As a large, complex natural product, neither a partial nor complete synthesis has yet been described. A synthetic route to any patellazole is unlikely until the complete stereochemistry of at least one member of the family has been convincingly described. 88 This lack of a synthetic route to the patellazoles illustrates the importance of investigating and understanding marine natural products and marine biosynthesis. Without access to such unique and potent molecules, anticancer drug discovery will suffer in two ways. Most obviously, there will be fewer novel lead compounds from which to develop effective therapeutic agents. Terrestrial sources have been well investigated and combinatorial chemistry is still relatively limited in the range of functionalities it can present. Marine organisms have had millions of years to develop a wide range of small molecules and have shown great flexibility in modifying these molecules in response to environmental challenges. This range of diversity is further illustrated in the profiling of the bioactivity of lissoclinolide. Although the desired compounds, the patellazoles, were not produced in large yields when investigating L. patella, the isolation of lissoclinolide led to the identification of previously undescribed bioactivity. At least in the foreseeable future, the chemical diversity present in marine invertebrates will be an important source of bioactive small molecules. The investigation of marine natural products may also lead to further understanding of the molecular biology of cancer. It is important to note that the targets of neither the patellazoles nor lissoclinolide have yet been identified. A review of the literature led to the investigation of a known pathway, mTORJp70, which is affected by patellazole treatment in HCT 116 colon tumor cells. This pathway is most likely not the main target, however. Additionally, lissoclinolide does not act upon tubulin or the proteosome, as many GiM-arresting compounds do. The characterization of the cellular 89 targets of these small molecule marine metabolites may lead to the identification of new potential targets for anticancer therapies. Understanding the molecular biology of cancer and improving chemotherapeutic target selection is intertwined with identifying novel small molecule modulators of those targets. The desire for advancements in anticancer therapies will grow both scientifically and sociologically in the coming decades. By combining the development of new therapeutic strategies with the search