Special Communication

JAMA, Vol. 278. No. 16, pp. 1345-1348 (October 22/29, 1997)

From the Richard & Rhoda Goldman School of Public Policy & Center for the Economics and Demography of Aging, University of California, Berkeley (Dr. Banks); and Michigan State University College of Human Medicine, East Lansing (Dr. Fossel).

HISTORICALLY, there has been an-- as yet well-founded-assumption that although we may alter the course of age related diseases, such as emphysema by smoking cessation and atherosclerosis by lowering cholesterol and blood pressure, the underlying process of aging cannot itself be altered [1]. The actual life span of any single individual is determined by a multitude of factors including lifestyle, socioeconomic status, diet, environmental conditions, and genetic endowments [2,3]. Some of these factors can be altered with consequent alteration of the individual (or the mean) life span, but not with any known increase in the maximum human life span. Currently, the longest anyone has ever lived is 122 years [4]. The biological origin of this maximum life span has been controversial. Is there a "clock" regulating the aging process? Or is there a biological limit that cannot be surpassed-- irrespective of our genes? These are questions that gerontologists, geneticists, and molecular and cell biologists have been pondering for decades.

Recently, there has been a conceptual shift in our understanding of aging. The possibility of extending the maximum human life span has gone from legend to laboratory [5]. This change has been prompted by a growing academic literature that suggests, that the aging process itself, as well as the consequent and fundamental cellular changes that occur in age-related diseases, is modifiable. This revision in views, if borne out, has profound clinical implications for the incidence of age-related diseases. Furthermore, the fiscal impact on government expenditures in the areas of health and social security will be perplexing even if only a modicum of these views hold true.

Aging is often seen from two vantage points: damage theories and programmed theories. Damage theories view aging as the result of accumulated errors, from free radicals for example. Programmed theories view aging as the result of genetic regulation. These paradigms are not contradictory, but complementary. Changes in gene expression that occur with cellular aging permit damage that does not occur (or accumulate) in immortal cell lines such as germ cells. The issue is not whether free radical damage underlies much of aging-- it clearly does-- but the more complex question of whether we may learn to control the onset and timing of such damage, which ultimately determines our health and our maximum life span.

The mean human life span has been extended remarkably over the past two centuries; the maximum life span has not [6]. This has resulted in a "rectangularization" of life span, in which mortality is increasingly becoming compressed against an apparently fixed end point: the maximum human life span of approximately 120 years [7]. Over the past decade, however, several laboratories have successfully extended the maximum life spans of at least two multicellular species genetically (Drosophila [3] and Caenorhabditis elegans [8-10]) and several more by dietary restriction [11-13]. That this increase in maximum life span can occur at all invalidates the dogma that maximum life span is fixed, and it thus invalidates the inevitability of the rectangularization model described above. That it occurs with such minimal genetic manipulation and that the effect is so substantial (two gene mutations increase the maximum life span of C. elegans six-fold) raise the provocative question of what might lie in store for clinical medicine. To a first approximation, the increased life span occurs predominantly through the genetic control of free-radical metabolism. This is not unexpected: cellular aging (and modifications of maximum life span) operates predominantly through the damaging effects of free radicals [14,15]. Not only does free radical production increase substantially as cells age, but the free radicals produced are less well contained and less efficiently mopped up by free radical scavengers, and their consequent damage is less efficiently repaired. This simplification leaves unanswered the questions of how cancer cells and the germ cell lines avoid aging damage, though they have comparable genomes, mitochondria, and free radicals. Telomere shortening (and its consequent effects on gene expression and cell cycle mechanisms) is implicated in allowing cancer and other cell lines to avoid cell senescence [16,17]. Free radical damage appears to be the major cause of the damage that occurs in aging cells [18,19], but the more complex issue lies in the control of the timing and release of such damage [20].

What is the age of a cell? Unless we measure age from every new division (as we do in Saccharomyces, for example, in which the asexual form shows aging in that "maternal" cells have a limited number of divisions, but the number is "reset" in each new "daughter" cell), newly divided cells each inherit the same age as the single cell from which they derived. All cells-- as unbroken cell lines-- have, reductio ad absurdum, the same age. Life on this planet is 3.5 billion years old, and all life and every cell are equally old in a peculiar sense. But then how is the onset of free radical damage timed, if not simply by years of a cell's life? Aging does not occur for 3.5 billion years in the germ cell line, yet it clearly does so beginning sometime after fertilization. Free radical aging is not timed by the profoundly archaic age of the entire cell lineage, but rather by a clock that begins running only after fertilization (in multicellular organisms; in the asexual phase of unicellular organisms such as yeasts, aging is considered to commence at the asymmetric division of a daughter cell as she buds off the larger mother cell). Humans and other multicellular organisms derive from cells that do not show the loss of replicative ability and the morphologic and gene expression changes that are characteristic of senescent cells [1,21,22] until after the organism achieves multicellularity. Age per se does not determine aging; it is altered genetic expression that somehow permits the onset of aging to occur.

The onset and progression of aging is strongly affected by chromosomal structure [23] and gene expression [24]. Evidence from work on early aging syndromes (progerias) supports this observation. Patients with Werner's syndrome have an average life span of 47 years and a known defect in helicase metabolism that results in abnormal chromosomal "unwinding" and replication [25]. Patients with Hutchinson-Gilford syndrome have an average life span of 12.7 years, and their skin fibroblasts have telomere lengths characteristic of cells from far older patients [26-27].

The cloning of mature ovine mammary cells likewise raises the question of whether age can be reset. Wilmut, et al, [28] have shown that gene expression can be reset: in this case at least the six-year-old adult mammary cells from the donor have not yet been irreversibly aged by time per se or by the aging of the donor organism. To the contrary, successful cloning demonstrates that these cells still have sufficient intact genetic information available to allow reconstitution of an apparently healthy multicellular organism. But was Dolly-the cloned organism-- six years old at birth, or had her cells been reset to age zero? The most tempting and certainly viable hypothesis is that genes can be reset to reflect the cellular age consistent with a developing organism [29], a hypothesis that is now (in Dolly's ease) finally being tested. A considerable body of evidence suggests that cell aging does not occur because of accumulated DNA damage or poor transcription fidelity [30]; rather, the DNA is intact and transcription is intact, but transcription occurs at lower rates and in significantly altered patterns as cells age [21,23]. If Dolly's cells show normal ovine aging for her "new" age rather than being six years old at birth, then this will provide further support for the hypothesis that age can be reset even after it has been in progress. The nature of this putative resetting mechanism will be a prime target for further research.

Cancer research has also prompted revision of our views of cell aging [22]. The problem in oncology is not that cells age and die, but that they do not. Normal somatic cells have well-defined limits to replication. Young skin fibroblasts typically have 50 divisions before they show cellular senescence. Within the past seven years, there has been increasing support for the model that cellular senescence in normal (noncancerous) human cells is the result of an altered (and, for the cell, dysfunctional) pattern of gene expression, and that the onset and progression of this pattern of senescent gene expression is regulated (through insufficiently understood mechanisms likely to involve heterochromatin changes) regulated by the telomeres [31-33]. Most cancer cells, however, are biologically immortal and will continue dividing indefinitely under appropriate conditions. This potential is linked to their expression of telomerase [34-37]. Additionally, it appears that most, perhaps all, cancer cells are fundamentally different with regard to their aberrant cell-cycle braking systems. Typically, 60% of combined human cancers (70% of colon cancers, 40% of breast cancers) have an aberrant p53. If we include p2l, D cyclins, cyclin-dependent kinases, and other parts of the cell cycle "machinery," the percentage of cell cycle abnormalities might approach 100% of cancers [38,39]. More recently, evidence is accumulating that at least 90% of cancer cells from human malignancies are also capable of resetting their telomere lengths and evading cell senescence [40]. This model has proven a powerful predictor not only of cell aging, but also of cancer cell survival [41].

This realization-- that the aging process is mutable at the cellular level and the implications of this for age-related diseases-- has itself spawned more than a dozen targeted biotechnology companies over the past four years [42]. Not only does this realization have implications for cellular aging, but it has also prompted rethinking of the historically warranted pessimism about curing cancer. As a Science editorial put it at the end of 1996, predicting whether we might cure cancer, "the standard answer-- 'No'--- may be up for revision [43]. Such optimism is not novel, but a comprehensive conceptual base, good supporting data, and promising clinical trials are new and very welcome.

The recent cloning of the protein component of human telomerase [36] will only accelerate this already rapid progress in understanding cancer and cell senescence. More importantly, to the extent that we can affect cell senescence therapeutically, we may stand to alter age-related diseases and thereby their economic, social, and human outcomes.

 AN EVOLVING MODEL

Previously, aging was seen as an immutable, passive accumulation of entropic damage in the cell. The evolving paradigm, in which cell aging results from and is coordinate on altered gene expression [29,32,33] is growing in acceptance. Although still nascent, the Senescent Gene Expression (SGE) model is a model that encompasses both the programmed and damage schools-of-thought on aging. It suggests that aging is a complex cascade of processes that include repeated cell division, telomere effects at the chromosome ends (due to both their lengths and their associated heterochromatin), a significantly altered pattern of senescent gene expression in "old" cells, and consequent alterations in cell metabolism in general and free-radical metabolism (production, sequestration, trapping, and damage repair) in particular. In addition, even cells that do not directly demonstrate aging changes (e.g., myocardial cells) depend on those that do (e.g., vascular endothelial cells). Together this cascade of processes play a coordinated but complex role in aging and-- far more importantly-- identification of each step could provide a basis for developing therapeutic interventions to affect either a single step or the whole. This model-- that the process of aging in the cell can be altered or reset (as may be true in the cloning of adult sheep cells and as is demonstrably true in senescent fibroblast hybridomas [44]-- undermines the historical certainty that aging and age-related diseases are immutable at the cellular level. Coordinate and similar shifts are occurring in our understanding of cancer. Differences at the heart of cancer cells, such as cell cycle errors and telomerase expression, may provide leverage to draw far clearer therapeutic lines than we now can between cancers and normal cells. While none of our currently altering views of cancer or aging promise a cure, enormous clinical potentials may be opening for us where none existed previously [43,45,46].

These changes in our theoretical understanding of cell aging and the availability of data to support the mutability of cell aging together call into question our ability to provide accurate extrapolations of the social and economic consequences of projected aging in developed countries. In 1950, four years before the Salk vaccine, if we were to have predicted the medical costs (e.g., iron lungs) that would accrue from polio over the next decade, we would have risked similar inaccuracies. Predictions of the costs of aging will be hollow unless they factor in the ongoing, nascent, and fundamental changes in our understanding of aging and age-related diseases.

We are currently incapable of altering the aging process in any meaningful way, and any assertion to the contrary is misleading and not supported by fact. It is, however, equally misleading and disingenuous as well to ignore the ongoing shift in our knowledge of cell biology as we attempt to predict the future of health care and social policy in the United States and other developed countries. The assumptions on which we base our predictions need to be clearly understood, and there exists a reasonable degree of uncertainty in assuming that aging and age related diseases will remain immutable in the coming decades.

The fiscal and social implications of this possibility cannot be underestimated. Any increase in the life span will be accompanied by concomitant shifts in social spending on the aged, but we cannot reasonably estimate the timing, magnitude, or fiscal outcome of altering the maximum human life span and age-related diseases. Even if one denies the possibility of such alteration, the upward trend in median age of the world's population is substantial but of uncertain magnitude. Even assuming that the lifespan is fixed, the consequent rectangularization of the survival curve, described by Fries [7], will exert fiscal and social pressures on our ability to care for the elderly [47,48]. The magnitude of these pressures, their global distribution, and their effect on well-being will depend on several factors.

First, while there exists some variation internationally in the health status of elderly cohorts [49,50], the social costs of increasing the maximum human life span will largely depend on the level of functional impairments and chronic disabilities among the aged [51,52]. Recent research suggests that, at least in the United States, increases in the mean life span have been accompanied by a concomitant decline in the prevalence of morbidity and disability among the aged [53,54]. The magnitude of any such decline and its overall impact on health care costs, relative to prior elderly cohorts, will vary internationally as a function of a nation's health care system, preventive health measures, rate of technological diffusion, advances in the treatment of acute and chronic diseases, social and political structures, and economic systems-- to mention a few. The possibility that we might delay or prevent aging and associated disabilities only increases the already considerable variance in estimating such costs. However, the impact of such innovations would be determined by the political, social, regulatory, and economic structures of each country.

Second, innovations in medical technology (therapeutic, diagnostic, or organizational) for age-related conditions need not coincide with the health care needs or necessarily promote the well-being of those in developing countries. The immediate and future medical and public health concerns of sub-Saharan Africa are not in the research, development, and diffusion of age-related technologies, but in the development of effective mechanisms for controlling the spread of infectious diseases (such as the Acquired Immunodeficiency Syndrome [55], malaria, and tuberculosis), and the fatal consequences of drought, famine, and civil wars, which lead to mean life spans in these countries of approximately 48 years [56]. Therefore, when considering the international fiscal effects of increasing the maximum healthy human life span and the impact of these innovations on functional impairment among the aged, it is important to acknowledge that these advances are likely to benefit developed nations preferentially.

Third, the continued decline in the level of functional impairment among the elderly populations in developed countries, along with increases in active life expectancy [57,58], has profound implications for work and retirement years. Even though the effect of longevity on work and retirement choice remains unclear [59], developed nations can expect years of productive life among elderly age-specific cohorts to increase. How altering the maximum human life span-- by whatever magnitude-- would alter this effect cannot be reliably projected, but contrary to what one might expect, increased life expectancy in the United States has been associated with an accelerated decline in the labor force participation rates of older persons for the past 50 years [60]. To sustain current levels of well-being, however, longer life spans will have to be accompanied by either correspondingly longer work years, or higher premiums to enhance private contributions to retirement accounts, or higher taxes to enhance contributions into social insurance accounts. The magnitude of the latter will be determined by the willingness of current generations to subsidize future generations, as well as the ratio of workers to retirees. Further, the structure of public and private pension funds can in themselves affect the labor force participation rates of individuals. The retirement income provided by the funds allows older workers to leave the labor force at younger ages and still support themselves in retirement years [60]. Hence, it is the interplay among mortality, health status, work, and retirement and the structure of retirement accounts that will determine the overall impact of the burgeoning aging population and any increase in the maximum human life span on the fiscal health of nations.

Finally, with increasing relative numbers of the aged, along with the growth in the proportion of the oldest old by the year 2040 [61], the international demand for long-term care could impose unexpected pressures on governments and individuals. This will be exacerbated if the proportion of offspring willing to live with their elderly parents continues to decline internationally [62], therefore placing greater pressures on the public and private sectors for the establishment of innovative alternatives to long-term care-such as home or community-based care. The magnitude and direction of these pressures depend critically on the extent to which the maximum life span may be altered and the degree to which the prevalence of disabilities will be modified. Neither of these effects can be reliably estimated, yet current research suggests there is a reasonable possibility that both might alter within the time frame of current attempts to project future trends in international health and social welfare expenditures.

1. Hayflick L. How and Why We Age. New York, NY: Ballantine Books; 1994:313.

2. Finch CE. Longevity, Senescence and the Genome. Chicago, Ill: University of Chicago Press; 1990.

3. Jazwinski SM. Longevity, genes, and aging. Science. 1996;273:54-59.

4. Jeanne Calment, born 21 February 1875 in Arles, died 4 August 1997 in France: world's oldest woman dies at 122. Baltimore Sun. August 5,1997:B5.

5. National Institute on Aging, National Institutes of Health. In Search of the Secrets of Aging. Bethesda, Md: National Institute on Aging, National Institutes of Health; 1997.

6. Smith DWE. Human Longevity. New York, NY: Oxford University Press; 1993.

7. Fries JF, Crapo LF. Vitality and Aging: Implications of the Rectangular Curve. San Francisco, Calif. WH Freeman; 1981.

8. Lakowski B, Hekinii S. Determination of lifespan in Caenorhabditis elegans by four clock genes. Science. 1996;272:1010-1013.

9. Ewbank JJ, Barnes TM, Lakowski B, Lussier M, Bussey H, Hekimi S. Structural and functional conservation of the Caenorhabditis elegans timing gene elk-1. Science. 1997;275:980-983.

10. Larsen PL, Albert PS, Riddle DL. Genes that regulate both development and longevity in Caenorhabditis elegans. Genetics. 1995;139:1567-1583.

11. Weindruch R, Walford RL. The Retardation of Aging and Disease by Dietary Restriction. Springfield, Ill: Charles C. Thomas Publisher Inc; 1988.

12. Sohal RS, Weindruch R. Oxidative stress, caloric restriction, and aging. Science. 1996;273:59-63.

13. Lane MA, Ingram DK, Roth GS. Beyond the rodent model: calorie restriction in rhesus monkeys. Age. 1997;200:45-46.

14. Harman D. The biologic clock: the mitochondria? J Am Geriatr Soc. 1972;20:145-147.

15. Orr WC, Sojal RJ. Extension of life-span by overexpression of super oxide dismutase and catalase in Drosophila melanogaster. Science. 1994;263: 1128-1138.

16. Lundblad V, Wright WE. Telomeres and telomerase: a simple picture becomes complex. Cell. 1996;87:369-375.

17. Holt SE, Wright WE, Shay JW. Regulation of telomerase activity in immortal cell lines. Mol Cell Biol. 1996;16:2932-2939.

18. Harman D. Aging: a theory based on free radical and radiation biology. J Gerontol. 1956;11:298-300.

19. Sohal RS, Weindruch R. Oxidative stress, caloric restriction, and aging. Science. 1996;273:59-63.

20. Fossel M. Reversing Human Aging. New York, NY: William Morrow & Co; 1996.

21. Dimri GP, Lee X, Basile G, et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci U S A. 1995;92:9363-9367.

22. Smith JR, Pereira-Smith OM. Replicative senescence: implications for in vivo aging and tumor suppression. Science. 1996;273:63-67.

23. Guarente L. Do changes in chromosomes cause aging? Cell. 1996;86:9-12.

24. Campisi J. Replicative senescence: an old live's tale. Cell. 1996;84:497-500.

25. Yu CE, Oshima J, Fu YH, et al. Positional cloning of the Werner's syndrome gene. Science. 1996; 272:258-262.

26. Allsopp RC, Vaziri H, Patterson C, et al. Telomere length predicts replicative capacity of human fibroblasts. Proc Natl Acad Sci U S A. 1992;89: 10114-10118.

27. Oshima J, Brown WT, Martin GM. No detectable mutations at Werner helicase locus in progeria. Lancet. 1996;348:1106.

28. Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KHS. Viable offspring derived from fetal and adult mammalian cells. Nature. 1997;385:810-813.

29. Stewart C. An udder way of making lambs. Nature. 1997;385:769-771.

30. Harley CB, Pollard JW, Chamberlain JW, Stunner CP, Goldstein S. Protein synthetic errors do not increase during aging of cultured human fibroblasts. Proc Natl Acad Sci U S A 1980;77:1885-1889.

31. Harley CB, Flutter AB, Greedier CW. Telomeres shorten during aging of human fibroblasts. Nature. 1990;345:458460.

32. Guarantee L. Do changes in chromosomes cause aging? Cell. 1996;86:9-12.

33. Marinade S, Gasser SM, and Wilson E. Chromatin: a sticky silence. Cur Biol. 1996;6:1222-1225.

34. Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PLC. Specific association of human telomerase activity with immortal cells and cancer. Science. 1994;2?_:2011-2015.

35. Shay JW, Bacchett S. A survey of telomerase activity in human cancer. Eur J Cancer. 1997;33: 787-791.

36. Nakamura TM, Morin GB, Chapman KB, et al. Telomerase catalytic subunit homologs from fission yeast and human. Science. 1997;277:955-959.

37. Meyerson M, Counter C, Eaton EN, et al. hEST2, the putative human telomerase catalytic subunit gene, is up-regulated in tumor cells and during immortalization. Cell. 1997;90:785-795.

38. Rao RN. Targets for cancer therapy in the cell cycle pathway. Curr Opin Oncol. 1996;8:516-524.

39. Sherr CJ. Cancer cell cycles. Science. 1996;274: 1672-1677.

40. Shay JW, Gazdar AF. Telomerase in the early detection of cancer. J Clin Pathol. 1997;50:106-109.

41. Feng J, Funk WD, Wang SS, et al. The RNA component of human telomerase. Science. 1995;269: 1236-1241.

42. Jazwinski SM, Martin GM, Schellenberg GD, et al. Molecular and genetic strategies for treatment of age-related diseases. Presented at the National Managed Health Care Congress: Bio/technology Conferences; July 14-15, 1997; Seattle, Wash.

43. Scanning the research horizon. Science. 1996; 274:1989. Editorial box.

44. Wright WE, Brasiskyte D, Piatyszek MA, Shay JE. Experimental elongation of telomeres extends the life span of immortal x normal cell hybrids. EMBO J. 1996;15:1734-1741.

45. Hartwell LH, Kastan MB. Cell cycle control and cancer. Science. 1994;266:1821-1828.

46. Jacks T, Weinberg RA. Cell-cycle control and its watchman. Nature. 1996;381:643-644.

47. Winker MA, Glass RM. Aging global population: a call for papers. JAMA 1996;276:1758.

48. Marwick C. Longevity requires policy revolution. JAMA. 1995;273:1319-1321.

49. Manton KG, James VW. Survival after the age of 80 in the United States, Sweden, France, England, and Japan. N Engl J Med. 1995;333:1232-1235.

50. Sallar AM, Hogg RS, Schethter MT. Survival after age 80. N Engl J Med. 1996;&M:537-538.

51 . Lubitz J, Beebe J, Baker C. Longevity and Medicare expenditures. N Engl J Med. 1995;332: 999-1003.

52. Schneider EL, Guralnik JM. The aging of America: impact on health care costs. JAMA 1990;263:2335-2340.

53. Manton KG, Corder L, Stallard E. Chronic disability trends in the elderly United States populations: 1982-1994. Proc Natl Acad Sci U S A. 1997; 94:2593-2598.

54. Alliance for Aging Research. Seven Deadly Myths: Uncovering the Facts About the High Cost of the Last Year of Life. Washington, DC: Alliance for Aging Research; May 1997.

55. Van De Walle E. The social impact of AIDS in sub-Saharan Africa. Milbank Q. 1990;68(l):10-32.

56. US Bureau of the Census. Statistical Abstract of the United States, 115th Edition. Washington, DC: US Bureau of the Census; 1995.

57. Manton KG, Stallard E, Liu K. Forecasts of active life expectancy: policy and fiscal implications. J Gerontol. 1993;48:11-26.

58. Sullivan DF. A single index of mortality and morbidity. HSMHA Health Rep. 1971;86:347-354.

59. Lee R, Tubapurkar S. Death and taxes: longer life, consumption, and social security. Demography. 1997;34(l):67-81.

60. Wise DA. Retirement against the demographic trend: more older people living longer, working less, and saving less. Demography. 1997;34(l):83-95.

61. US Bureau of the Census. Statistical Abstract of the United States, 1995. Washington, DC: US Bureau of the Census; 1996.

62. Okamoto Y. Healthcare for the elderly in Japan: medicine and welfare in an aging society facing a crisis in long term care. BMJ. 1992;306:403-405.