elloff and colleagues have been key players in the recent development of chemoprevention strategies--as initiators of their own studies and minders of others. The succinct summary of their approach is of particular value to oncologists, both because it provides a great deal of data on the current state of chemoprevention research itself and because it draws some useful distinctions between chemoprevention and chemotherapy.
Chemoprevention: Pharmacology or Biology?
elloff and colleagues have been key players in the recent developmentof chemoprevention strategies--as initiators of their own studiesand minders of others. The succinct summary of their approachis of particular value to oncologists, both because it providesa great deal of data on the current state of chemoprevention researchitself and because it draws some useful distinctions between chemopreventionand chemotherapy.
One role for a commentary on such a useful contribution is toset it in a wider context and to ask some questions that Kelloffand coauthors, by their very commitment to the enterprise, areunable to pose. Such a commentary can ask several related questions:
Do We Know Enough About Cancer?
Asking whether we know enough about cancer to do chemopreventionis not to subvert the current enterprise, nor even to questionthe enthusiastic pursuit of new and better agents. It is, however,intended to remind us that our models of cancer have been evolvingrapidly, and that there have been many surprises over the last10 years. Among these, we can count the semiorganized and cancer-specificnature of the sequence of somatic genetic changes;[1-3] the roleof DNA repair and microsatellite instability;[4-8] the tumor-suppressorrole of inherited genes like BRCA1 with little evidence of
somatic mutation in sporadic tumors;[9,10] the mimicking of asecond "hit" by hypermethylation;[11] clear evidencethat agents thought to cause mutation actually act by selection;[12]the place of apoptosis in carcinogenesis;[13-15] the possibilityof roles for transposons[16-18] and imprint-ing;[19-21] and theheterogeneity of consequences when a gene mutates at differentloci or in the germ line as opposed to somatically.[22-24]
At the very least, this incomplete catalog suggests that we shouldconsider agents beyond those that suppress proliferation or preventmutagenesis. More importantly, it reminds us that we will discoverwhole new processes relevant to carcinogenesis that may undermineour confidence in existing models. Such undiscovered processesmay even explain the failure of agents that, until recently, wefelt secure enough to test in large-scale trials.
In the face of growing evidence of heterogeneity in the cancerprocess, we should adopt an attitude of appropriate humility whilelooking for opportunities both to interrupt the sequence and torevise our models of it. While cultivating such an attitude, itprobably makes sense to eschew the concepts of "toxicity"and "side effects," as these are derived solely fromthe pharmacologic notion that we know what an agent ought to do,and not from a biologic understanding of what it actually does.There are no "main effects" and "side effects,"only effects.
What Have We Learned From Recent Chemoprevention Trials?
The results of recent chemoprevention studies are so at variancewith expectation that it is already clear that some of our modelsare wrong. Human experiments show that modest doses of b-carotenedo not prevent cancer in average-risk[25] or high-risk individuals.[26,27] But what are we to make of an increase in lung cancer[26,27]and large polyps[28] following b-carotene supplementation in high-riskindividuals? A null answer to the chemoprevention question canmean: good hypothesis, good science, but wrong agent. An increasein risk, however, strongly suggests that the model, not the agent,is wrong.
Can We Draw Parallels Between Chemotherapy and Chemoprevention?
One of the parallels between chemotherapy and chemopreventionthat the article does not draw is the outcome of using a singleagent. Extensive clinical trials, in both adults and children,have proven the efficacy of multiple agents, in combination andsequentially. The rationale is so deeply embedded in the way thatwe think about chemotherapy that it is almost subliminal.
Single agents select for clones of cancer cells capable of surviving.Polypharmacy reduces the likelihood of such clones emerging. Selectionoccurs because specific cells are capable, by a variety of means,of resisting the cytotoxic, apoptosis-inducing, and other actionsof the therapeutic agents. Chemopreventive agents, by inducingdifferentiation or reducing proliferation or mutation, are equallycapable of acting as agents of selection, however. Furthermore,the cells that survive chemopreventive agents may be as capableof clonal expansion as any survivor of a cytotoxic agent.
Indeed, it is this kind of selection that may explain why agentsas apparently benign and beneficial as b-carotene can actuallyincrease the risk of lung cancer in individuals known to havelarge numbers of initiated cells. One plausible mechanism forthis outcome is that bcl-2, an inhibitor of apoptosis, may operatevia an antioxidant pathway.[29] Although, on the face of it, b-carotenemay not be a sufficiently powerful antioxidant to induce comparableinhibition of apoptosis, nonetheless, the presence of high concentrationsof even a weak antioxidant may be sufficient to prevent the normalapoptotic death of abnormal cells. The agent then is not workingto prevent mutation-inducing oxidative damage in the DNA of normalcells, but rather, to inhibit the very mechanism by which theorganism sheds abnormal cells.
This is a testable hypothesis. If it is true, what it argues formost strongly is the use of chemopreventive polypharmacy (or diet,of course [30]). If we begin with a single agent and a high-riskgroup, we may indict an agent or class of agents as being uselessor even cancer-promoting when what we have actually discoveredis that our models of cancer are not very good. Some principlesthat derive from both evolutionary biology and chemotherapy applyalso to chemoprevention.
1. Rowley J: Chromosome abnormalities in cancer. Cancer GenetCytogenet 2:175-198, 1980.
2. Solomon E, Borrow J, Goddard A: Chromosome aberrations andcancer. Science 254:1153-1160, 1991.
3. Vogelstein B, Fearon E, Kern S, et al: Allelotype of colorectalcarcinomas. Science 244:207-211, 1989.
4. Fishel R, Lescoe M, Rao MRS, et al: The human mutator genehomolog MSH2 and its association with hereditary nonpolyposiscolon cancer. Cell 75:1027-1038, 1993.
5. Leach F, Nicolaides N, Papadopoulos N, et al: Mutations ofmutS homolog in hereditary nonpolyposis colorectal cancer. Cell75:1215-1225, 1993.
6. Ionov Y, Peinado M, Malkhosyan, et al: Ubiquitous somatic mutationsin simple repeated sequences reveal a new mechanism for coloniccarcinogenesis. Nature 363:558-561, 1993.
7. Parsons R, Li GM, Longley M, et al: Hypermutability and mismatchrepair deficiency in RER+ tumor cells. Cell 75:1227-1236, 1993.
8. Thibodeau S, Bren G, Schaid D: Microsatellite instability incancer of the proximal colon. Science 260:816-819, 1993.
9. Futreal PA, Liu Q, Shattuck-Eidens D, et al: BRCA1 mutationsin primary breast and ovarian carcinomas. Science 266:120-122,1994.
10. Takahashi H, Behkakht K, McGovern P, et al: Mutation analysisof the BRCA1 gene in ovarian cancers. Cancer Res 55:2998-3002,1995.
11. Baylin S, Herman J, Wales MM, et al: Hypermethylation of CpGislands and inactivation of tumor suppressor genes. Proc Am AssnCancer Res 36:691-692, 1995.
12. Jin Z, Zarbl H: Carcinogen induced mechanisms in mammary tumorigenesis.Proc Am Assn Cancer Res 36:658-659, 1995.
13. Hockenbery D: Bcl-2 in cancer, development and apoptosis.J Cell Sci 18:51-55, 1995.
14. Sinicrope F, Raun S, Cleary K, et al: Bcl-2 and p53 oncoproteinexpression during colorectal tumorigenesis. Cancer Res 55:237-241,1995.
15. Bronner M, Culin C, Reed J, et al: Bcl-2 protooncogene andthe gastrointestinal mucosal epithelial tumor progression model.Am J Pathol 146:20-26, 1995.
16. Sager R: Transposable elements and chromosomal rearrangementsin cancer--a possible link. Nature 282:447-448, 1979.
17. Morse B, Rothberg P, South V, et al: Insertional mutagenesisof the myc locus by a LINE-1 sequence in a human breast carcinoma.Nature 333:87-90, 1988.
18. Miki Y, Nishisho I, Horii A, et al: Disruption of the APCgene by a retrotransposal insertion of L1 sequence in a coloncancer. Cancer Res 52:643-645, 1992.
19. Reik W: Imprinting in leukemia. Nature 359:362-363, 1992.
20. Mannens M, Hoovers JMN, Redeker E, et al: Parental imprintingof human chromosome region 11p15.3-pter involved in the Beckwith-Wiedemannsyndrome and various human neoplasia. Eur J Hum Genet 2:3-23,1994.
21. Rachmilewitz J, Goshen R, Ariel I, et al: Parental imprintingof the human H19 gene. FEBS Lett 309:25-28, 1992.
22. van Heyningen V: One gene--four syndromes. Nature 367:319-320,1994.
23. Eng C, Mulligan L, Healey C, et al: Heterogeneous mutationof the RET proto-oncogene in subpopulations of medullary thyroidcarcinoma. Cancer Res 56:2167-2170, 1996.
24. Goodfellow P, Wells S: RET gene and its implications for cancer.J Natl Cancer Inst 87:1515-1523, 1995.
25. Hennekens C, Buring J, Manson J, et al: Lack of effect oflong-term supplementation with beta carotene on the incidenceof malignant neoplasms and cardiovascular disease. N Engl J Med334:1145-1149, 1996.
26. Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group:The effect of vitamin E and beta carotene on the incidence oflung cancer and other cancers in male smokers. N Engl J Med 330:1029-1035,1994.
27. Omenn G, Goodman G, Thornquist M, et al: Effects of a combinationof beta carotene and vitamin A on lung cancer and cardiovasculardisease. N Engl J Med 334:1150-1155, 1996.
28. MacLennan R, Macrae F, Bain C, et al: Randomized trial ofintake of fat, fiber, and beta carotene to prevent colorectaladenomas. J Natl Cancer Inst 87:1760-1766, 1995.
29. Hockenbery D, Oltvai Z, Yin M, et al: Bcl-2 functions in anantioxidant pathway to prevent apoptosis. Cell 75:241-251, 1993.
30. Potter J, Steinmetz K: Vegetables, fruit and phytoestrogensas preventive agents. IARC Monographs, 1996 (in press).