Human Evolutionary Biology

28 Complex Chronic Diseases in Evolutionary Perspective S. Boyd Eaton From an evolutionary point of view, most complex Firstly, what is the evidence that we are still genetic chronic diseases appear to be the result of imbalance, Stone Agers? Secondly, if our ancestral lifestyle was so mismatch, between our genetic makeup and the condi- much healthier, why do we now live longer? And tions of life in Westernized twenty-first century thirdly, chronic illnesses ordinarily become apparent nations. The basic contentions (Eaton and Eaton, late in life. Isn’t it just because we live longer that we 1999a; Eaton et al., 2002a) are that: develop these disorders? Ten thousand years is plenty of time for at least 1. The contemporary human genome was selected over some genetic evolution to occur and, indeed, scientists thousands of millennia during which our ancestral have identified a number of genetic changes that have line existed as prehuman primates who became emerged since agriculture initiated the watershed increasingly human-like until, during the period alteration in human life conditions which has charac- between 100 and 50 thousand years ago, they became terized the last 10 millennia. The majority of these deal behaviorly modern and lived as near equivalents of with defenses against infectious diseases, the nature hunter-gatherers studied during the last century. of which has changed greatly since our ancestors lived 2. Our genetic makeup, especially that regarding our as nomadic hunter-gatherers. Settled communities core metabolic and physiologic characteristics, has (with attendant sanitation problems), increased popu- changed very little between the emergence of agri- lation density, and intimate association with domesti- culture, roughly 10 000 years ago, and the present. cated animals promoted epidemic “crowd” diseases 3. On the other hand, cultural change during these past which had previously been of far lesser importance. 10 millennia has progressed at an ever-accelerating Genetic evolutionary responses included the sickle cell rate. The resulting dissonance between what amount condition and similar hemoglobinopathies (defenses to Stone Age genes and Space Age lives fosters devel- against malaria), and also alterations in the genetics opment of multiple health disorders ranging from of our immune system. The latter explain why, after the potentially life threatening (cancer, pulmonary initial contact with Westerners, Native Americans, emphysema, heart attacks) to the more mundane, Hawaiians, and other such groups experienced disas- but still costly and uncomfortable conditions such trous death rates from “childhood” infectious illnesses as acne, high frequency hearing loss, myopia (near which, in Old World cultures, had come to be infre- sightedness), and dental caries. quently fatal. Those populations who adopted dairying These contentions make up the “discordance hypoth- and milk consumption developed persistence of intes- esis” and its corollary is that the “afflictions of afflu- tinal lactase into adulthood. Almost all mammals ence” might best be prevented by reincorporating the make the enzyme lactase during infancy because it essentials of our ancestral living pattern into our con- allows digestion of lactose, the sugar in mothers’ milk. temporary lives – ideally blending the best from the After weaning, ancestral humans, like other mammals, past with best from the present. no longer needed lactase because they did not consume dairy products. Animal husbandry made milk available to adults; hence lactase persistence became selectively OBJECTIONS TO THE EVOLUTIONARY beneficial. APPROACH As our ancestors migrated from Africa to other parts of the world, “races,” as we know them today, On initially encountering these propositions, educated differentiated. However, it is important to remember health-conscious individuals usually raise one or more that the identifying features that distinguish current of three potential reservations (Eaton et al., 2002b): humans from varying geographical regions are largely Human Evolutionary Biology, ed. Michael P. Muehlenbein. Published by Cambridge University Press. # Cambridge University Press 2010. 491

492 S. Boyd Eaton superficial. To our eyes, Finns and Australian centuries, from around 40 in 1800 to almost 80 in Aborigines look much more unlike than do mountain 2000. This unprecedented increase results from soci- and lowland gorillas, but there is far greater genetic etal affluence (better housing, more affordable food, dissimilarity between the gorillas. It appears that cheaper clothing), which has increased host resistance cheetahs are the only mammalian species whose indi- to infectious agents, and from effective public health vidual members are more genetically similar to each measures (immunizations, sanitary engineering, other than are humans. safer food) which decrease microbial transmission. A few other examples of “recent” genetic evolution Both factors work against contagious illnesses which have been proposed: a decreased ability to smell, were the major causes of human mortality until the last changes in genes affecting the liver’s capacity to detox- century. (In 1900, infectious diseases caused more ify natural poisons, and, most controversially, genes deaths in the United States than did cancer and affecting brain size and/or function. Some of the latter heart disease combined.) Increasingly sophisticated are analogous to sickle cell condition: when a person medical care has also played a role, although less than has only one such gene, the situation may be benefi- might be assumed. cial, however, when two copies of the mutant gene are On the other hand, personal lifestyle choices present, one from each parent, illness develops. affecting health – those targeted by health promotion An example is Tay–Sachs disease, which is caused by advocates – have actually had a negative impact. a single mutation and inherited in a fairly straightfor- Compared with the early 1900s, per capita cigarette ward Mendelian pattern. Tay–Sachs and conditions smoking has increased 10-fold; daily physical activity, like it are, apparently, unrelated to the common previously over 1000 kcal/day, has decreased to chronic degenerative diseases like diabetes and athero- 500 kcal/day. For adult men, average body mass sclerosis, which are influenced by many relevant genes, indices have increased from 23 to 28, 25 being con- and whose inheritance is much more complicated. sidered the upper limit of normal. The failure rate These genetic changes that have evolved during among candidates taking the Army’s induction physi- the past 10 000 years are exceptions which prove a cal fitness test has become so high that the Pentagon general rule: they are few in number and have little has seriously considered relaxing standards so that effect on chronic illness susceptibility. Even though more candidates can achieve eligibility. Intake of diet- some modifications have occurred, biologists, evolu- ary fiber has decreased while intake of sugars and tionary theorists, paleontologists, and human geneti- sweeteners has doubled. The interval between menar- cists concur that, genetically, we remain almost che and first full-term pregnancy has risen from identical to our late Paleolithic ancestors. If agricul- 5 years to 15 and breast-feeding at 12 months ture and industrialization had produced substantial (19.4%) is now one third what it was in 1912–1919 alteration, there should be categorical, “taxonomic” (58%). All these changes increase risk for one or more genetic differences between peoples whose ancestors chronic illnesses. have been farmers for thousands of years (e.g. Near The effect of increasing life expectancy on age- Easterners, Mayans, Chinese, etc.) and groups whose related chronic diseases is another potential stumbling ancestors have been foragers until the past gener- block. A population with average life expectancy of ation or two (e.g. !Kung San Bushmen, Paraguayan 40 years will inevitably have much lower mortality Ache ´, Australian Aborigines, etc.). Apart from genes from cancer, heart disease, diabetes, and stroke than affecting resistance to infectious disease, no such will a population with an average life expectancy of 75. markers have been identified. To this extent, comparisons between recently studied The second question, regarding life expectancy hunter-gatherers (imperfect, but the best available sur- now versus the Stone Age, disturbs nearly everyone rogates for prehistoric Stone Agers) and citizens of on initial consideration of the discordance hypothesis. affluent Western nations are invalid. Of course dis- The best available estimates are that until about 1800 eases whose frequency increases with age will cause average human life expectancy was relatively constant more deaths in a society with a greater proportion of at 30 to 40 years for as far back as can be determined. older individuals. The figure was likely less in urban centers until the This coin has another side, however. While chronic past century: estimates for London in the 1660s range degenerative diseases generally produce mortality in between 18 and 27 years! Infectious illnesses were later life, they begin much earlier, often in childhood. the major killers, especially those affecting children. This allows comparison between age-matched younger Individuals who reached age 40 are thought to have members of industrial and technologically primitive had an expectation of further life only moderately less societies. Biomarkers of developing abnormality such than that of contemporary humans at the same age. as obesity, rising blood pressure, nonobstructive cor- In affluent Westernized nations, average life onary atherosclerosis, and insulin resistance are expectancy has nearly doubled over the past two common among the former, but rare in the latter.

Complex Chronic Diseases in Evolutionary Perspective 493 Measurements of muscular strength and aerobic to, chronic diseases. This expectation has been well- power (endurance) reveal similar discrepancies, again substantiated: dozens to hundreds of genes have been favoring individuals whose lives more closely resemble found to play a role for most illnesses in this category: the ancestral pattern. About 20% of hunter-gatherers 250 and counting for obesity alone. There are excep- reach age 60 or beyond, but even in this age-bracket, tions such as hemoglobinopathies (e.g. sickle cell members of foraging and other technologically primi- anemia), retinoblastoma, and phenylketonuria, but tive cultures appear almost completely free from mani- the rule holds for the most widespread, numerically festations of most chronic degenerative diseases important chronic conditions. (osteoarthritis is an exception). Together, these obser- vations strongly suggest that it is current Western life- style rather than age alone that promotes those CATEGORIES OF DISCORDANCE “afflictions of affluence” the prevention of which is a major goal of contemporary health promotion efforts. It is through their impact on risk factors that the cultural changes which have accumulated since agri- culture’s emergence influence chronic disease propa- WHAT ARE COMPLEX CHRONIC gation. These cultural novelties fall into several DEGENERATIVE DISEASES? categories, but in all cases the new conditions appear to adversely affect human health. Illnesses in this category account for the overwhelming majority of deaths and disability in Westernized Nutrition nations. Such disorders are generally characterized by an uncertain etiology, multiple risk factors, a long There is surprisingly little overlap between current latency period, a protracted course of illness, a non- foods and those of the Paleolithic (Cordain et al., contagious origin, functional impairment or disability, 2002, 2005; Eaton, 2006). We get most of our calories and, in many cases, incurability. Examples include from grains, domesticated livestock, dairy products, asthma, arthritis, cancer, cardiovascular diseases and refined sugars, but preagricultural humans ate (including stroke and heart attack), and diabetes. naturally occurring plant foods and wild game. Of course, generalizations of this nature are subject They used almost no cereal grains, no dairy foods, to many exceptions and qualifications. Fortunately, no separated oils, no commercial processing, and many cancers are curable. There is increasing evidence they had no sources of “empty calories.” People in that chronic, low-grade infections may actually be the Stone Age consumed about twice the animal a prerequisite for several “noncontagious” chronic ill- protein current Westerners do. The proportion of nesses, including both malignancies and atheroscler- total fat in Paleolithic diets was roughly equal to that osis. The relationship appears to be that infection is a for contemporary Americans – about 35%; however, necessary, but not sufficient, causal factor. Sudden intake of serum-cholesterol raising fat was nearly death is the initial clinical manifestation for many always far less than at present, and there was individuals with coronary heart disease. The under- more dietary long-chain (C20 and above) polyunsatur- lying arterial plaque formation may have taken ated fatty acids. These are thought to be the most decades to reach the critical level, but from the biologically significant of the fatty acids because patient’s subjective viewpoint, the disorder is hardly they are necessary components of neurons and protracted and there has been no premonitory func- required for eicosanoid synthesis. The preagricultural tional impairment. essential fatty acid ratio (o-6:o-3) was perhaps 3:1; for One almost invariable attribute of chronic degen- average Americans it approximates 10:1 or more – an erative diseases is the necessary interaction of multiple imbalance likely to promote atherogenesis. Dietary causative influences – risk factors. Tobacco abuse, cholesterol content exceeded current US levels because heavy alcohol ingestion, and inadequate dietary both lean wild game and fatty supermarket steak meat antioxidant intake are all implicated for esophageal contain cholesterol. Carbohydrate consumption was cancer. These elements potentiate each other’s effects less than at present, but came almost exclusively in a multiplicative rather than a simple additive from fruits and vegetables, not from cereals, refined fashion. Similarly, abnormal body composition (hyper- sugars, and dairy products. The latter provide more adiposity, sarcopenia), excessive intake of high glycemic- readily absorbable “insulinogenic” glucose – thus load foods, and physical inactivity combine to induce increasing dietary glycemic load. That ancestral insulin resistance and, increasingly often, type II humans ate thrice the fruits and vegetables most afflu- diabetes. That multiple risk factors affect chronic dis- ent Westerners do should have enhanced their antioxi- ease incidence implies the existence of numerous genes dant capacity relative to ours. Compared with the which are involved in the development of, or resistance typical American pattern, Paleolithic diets generally

494 S. Boyd Eaton provided less sodium, but more potassium, fiber to first birth) interval was only 3 years, versus about (soluble and insoluble), micronutrients, and, probably, 12 years for average Americans and Europeans. phytochemicals. Before agriculture the net metabolic Foragers who lived through their full reproductive impact of human foods was slightly alkalotic, tending span had high parity: typically 6 live births vs. 1.8 for to raise bodily pH. Contemporary foods, especially Americans. Nursing was obligatory, intensive dairy products and those made from grains, have (on demand, not on schedule), and commonly lasted the opposite – acidotic – effect and tend to slightly three years. Only about 50% of American babies are lower bodily pH (Sebastian, 2005). Stabilizing nursed at all and mean nursing duration is barely homeostatic biochemical mechanisms largely correct 3 months. Age at menopause is hard to ascertain for for this influence, but over decades, the result is exces- forager women, but menses apparently cease some- sive skeletal calcium loss and, consequently, what earlier than in affluent societies. osteoporosis. New reproductive patterns and the associated These differences are pertinent to several areas ovulatory differential (three times as many ovulations of current nutrition-related research: e.g., o-3 fatty for Westerners not using oral contraceptives) are acids and depression; o-6:o-3 ratios and coronary associated with increased risk for cancers of the heart disease; fruits, vegetables, and phytochemicals breast, endometrium, and ovary. For example, imma- as cancer preventive agents; optimal versus minimal ture breast lobules form at puberty; their rapidly requirements of vitamins and minerals; dietary dividing cells are relatively susceptible to natural sodium, hypertension, and overall mortality; and the mutation, genotoxic carcinogens, and clonal promo- appropriate contribution of fats to dietary energy. tion. At first full-term pregnancy most lobules differ- entiate into mature forms whose cells divide more slowly and are hence more resistant. Prolonged nubil- Physical exertion ity thus extends a period of high susceptibility to Through nearly all of human evolution, physical carcinogenesis. exertion and food procurement have been inextricably linked; in the past people had to work, physically, Infection in order to eat. Hierarchical social stratification uncoupled this relationship for elites; industrializa- Relationships between humans and microbes were tion and mechanization have completed the dissoci- altered by the rise of agriculture. Higher population ation for practically everyone. Prior to the industrial density, frequent long-distance contacts, settled living, era humans are estimated to have required a total and interactions with domesticated animals vastly of about 3000 kcal (12 MJ) daily; for current affluent increased pathogen transmission. As a result, certain populations comparable estimates are 2000 kcal infections assumed greater importance, becoming (8 MJ) or less. This change has resulted from selective forces that have subsequently affected the decreased energy expenditure through physical human genome (e.g., malaria, typhoid fever). More exertion: about 20 kcal/kg/day (84 kJ) for hunter- recently, improved sanitation has reduced transmis- gatherers versus <5 kcal/kg/day (21 kJ) for sedentary sion, a pivotal contribution to the past two centuries’ Westerners – a four-fold differential (Eaton and increase in average life-expectancy. Discovery of Eaton, 2003). antibiotics had dramatic impact, but intensive usage, Exercise has important effects on aerobic power, including incorporation into animal feeds, has led to muscular strength, and skeletal robusticity, all of the emergence of resistant organisms. Consequently, which were substantially greater for ancestral popula- “preventive” anti-infective chemotherapy must now tions. Exercise likely affects the incidence of age- aim at minimizing resistance as well as attaining related fractures, some cancers, and atherosclerosis. clinical efficacy. To this end, mathematical models Obligatory exertion promoted greater lean body mass integrating classic pharmacological approaches while attenuating adipose tissue, thereby reducing type with the principles of evolutionary biology may help II diabetes risk for our ancestors. optimize treatment protocols. Attempts to reduce pathogen virulence may also benefit from Darwinian considerations. For example, vaccines directed against Reproduction virulence-enhancing microbial antigens might dispro- Studies of women in foraging and other traditional portionately affect dangerous strains and promote settings suggest substantial differences between pat- their displacement by milder variants (Eaton et al., terns of ancestral and modern reproduction (Eaton 2002a). and Eaton, 1999b Ellison, 2001). For preindustrial While adequate food, public health measures, and women menarche was later (16 vs. 12.5 years) and first medical interventions have lowered infectious disease birth earlier (~19 years) so that the nubility (menarche mortality during the past century, the megapolitan

Complex Chronic Diseases in Evolutionary Perspective 495 crowding and unparalleled mobility in current affluent Growth and development nations have probably increased transmission of In Western populations, less frequent and severe child- certain organisms, especially those spread by sexual hood infection, sharply reduced exercise requirements, and respiratory contact. This phenomenon could unprecedented caloric availability, and epigenetic affect chronic disease prevalence: there are well- influences on maternal–fetal relationships (which are established relationships between viral infections only now becoming understood) result in rapid bodily and certain cancers as well as intriguing hints of growth and early sexual maturation. Average adult a causal link between microbes and atherosclerosis. height is asymptotically approaching a maximum Epidemiological correlation between infectious expos- while age at menarche has fallen to about 12.5 years, ure rates and incidence of chronic “noninfectious” probably near the population’s genetic limit. Most degenerative diseases might ultimately open new recent hunter-gatherers have been short-statured, avenues for preventive intervention via evolution- reflecting the nutritional stress of foraging in marginal based antibiotic prophylaxis and/or vaccine develop- environments, but average height for Paleolithic ment. Immunization against human papillomavirus, humans (who lived in more fertile regions) appears to which is strongly linked to development of cervical have roughly equaled that at present. Nevertheless, cancer, is an example of this approach. maturation may have been slower, as it is for athletic The impact of public health measures and per- young women in Western nations. Traditional North sonal cleanliness may not have been altogether posi- African pastoralists – who have sufficient dietary pro- tive: the hygiene hypothesis posits that allergic tein, little access to empty calories, and high levels of conditions and the so-called autoimmune diseases physical exertion – may simulate the ancestral stand- are undesirable side effects (Yazdanbaksh et al., ard. They experience later puberty and slower growth 2002). For our ancestors, genetically determined in height than do Westerners, attaining full stature host defences – white blood cells, antibodies, etc. only in their early 20s; still, their average adult height (see Chapter 27 of this volume) – resisted the near- equals that of Europeans. continuous onslaught of bacteria, parasites, hel- Rapid growth is usually interpreted as a sign of minths, and viruses. The human immune system has societal health, but maximal is not necessarily optimal been designed through evolutionary selection for (Eaton et al., 2002a). The current experience of puberty ongoing battle with our microscopic adversaries. three years earlier than the hunter-gatherer average But now the terms of engagement have changed so may result in dissociation between psychological that, in a very real sense, our immune system has and sexual maturation, thus contributing to unwanted become under-employed. Previously prevailing teenage pregnancies as well as suboptimal parenting. circumstances (repetitive, unrelenting exposure to Both early menarche and youthful attainment of infectious agents) fostered an immunoregulatory adult stature are associated with increased breast atmosphere which maintained focus on deleterious cancer risk. Rapid bodily growth may also affect blood microorganisms while protecting against injury to pressure regulation if renal development is unable to our own tissues. In contrast, contemporary conditions keep pace allometrically, thus requiring compensatory apparently fail to activate appropriate immunoregula- blood pressure elevation to maintain homeostasis, and tion so the immune system has become prone to possibly establishing a pathophysiological trajectory attacking our lungs (asthma), paranasal sinuses towards subsequent hypertension. Also, in laboratory (sinusitis, rhinitis), joints (various arthridides), bone animals at least, slower growth during adolescence and marrow (childhood leukemias), intestines (Crohn’s early adulthood is associated with increased longevity – disease, ulcerative colitis), brain (multiple sclerosis), apparently independent of any effect on chronic disease and pancreas (type I, “childhood” diabetes). These susceptibility. allergic or autoimmune disorders occur disproportio- nately among affluent, urban, and perhaps otherwise optimally hygienic populations for whom exposure to Psychosocial factors microorganisms is (presumably) especially reduced as compared with prior human experience. Genes affecting human behavior are ancient and For immunologists, rheumatologists, and others probably coevolved with our life history characteris- concerned with diseases of this nature, the aim is tics. For example, prolongation of childhood during to isolate those microbial constituents (antigens) hominid evolution may have facilitated learning that previously acted to induce appropriate immuno- and correlated with brain expansion occurring regulation and to create from meticulously selected over the same period. But, like current sedentism and components vaccines that will guard against develop- diet, the social circumstances of contemporary exist- ment of allergic/autoimmune diseases while maintain- ence are novel. Many factors believed to exert import- ing adequate host defense against infection. ant influence on psychological development and

496 S. Boyd Eaton interpersonal relations are profoundly different from TABLE 28.1. A partial list of discordance-related what they are thought to have been during our evolu- diseases. tionary past (Eaton et al., 2002a). Average birth spacing is now closer, while nursing and physical Acne contact between infants and adults is much reduced. Age-related fractures In most affluent societies, babies do not sleep Alcoholism with their mothers – a break from general primate Asthma experience dating back many million years. Ancestral Cancer childhood and adolescence were almost certainly char- “Childhood” infections acterized by multiage play groups, less restrictive Cirrhosis supervision, and intense small group interpersonal Coronary heart disease dynamics quite different from the age-segregated, Dental caries more structured routines of contemporary schools Dental malocclusion (impacted teeth) and little leagues. Based on what we know about Depression hunter-gatherers, Paleolithic teenagers had relatively Diabetes clear societal expectations, not the exciting-but- Emphysema daunting array of life choices which confronts young Epidemic infectious disease people today. For adults, a global society has advan- Gingivitis (gum disease) tages, but it differs radically from the more human- High frequency hearing loss scale experience of our ancestors who lived, found Hypertension their roles, and developed self-esteem in bands of Lactose intolerance 15–50 people, most of whom were relatives. We have Low back pain syndrome little concrete evidence, but it seems likely that these Myopia (near sightedness) differences and others – frequent contact with Obesity strangers, conflicting social roles, wage labor, working Osteoporosis in bureaucracies, reduced support from kin, socio- Peripheral vascular disease economic societal stratification, and education that Sickle cell anemia questions social beliefs and ideologies – may contribute to syndromes such as attention deficit/hyperactivity, Stroke depression, anxiety disorders, and substance abuse. of Europeans have BPs above 140/90 mmHg or are on REPRESENTATIVE ILLNESSES anti-hypertensive medication. Also HT is becoming increasingly common: in the United States, its preva- Table 28.1 lists many chronic disorders, all of which lence rose 15% between 1990 and 2000. can be considered discordance-related conditions. Across the globe, many cultures studied during the Discussing each of these in detail would greatly exceed last century were found to have no HT (Table 28.2). the length allowed for this chapter; consequently, four Their BPs averaged 110/70 mmHg – within the range disparate ailments, whose ties to discordance vary newly designated as normal by the NIH. Members of from the simple and straightforward to complicated normotensive cultures were not immune from HT and indirect, will be analyzed as illustrative examples. because, when they adopted a Western lifestyle, either by migration or by acculturation within their home- land, their BP experience soon began to parallel that of Hypertension Americans and Europeans. These normal BP cultures For decades blood pressure (BP) readings of 120/80 existed in varied climatic circumstances – in the arctic, mmHg were considered normal, but recent studies the rainforest, the desert, and the savanna – but they have shown that adverse health effects begin at pres- shared a number of essential similarities, each of sures exceeding 115/75 mmHg. Accordingly, in 2003, which reprised, or continued, the experience of our an National Institutes of Health (NIH) select commit- late Paleolithic ancestors and which were the recipro- tee redefined “normal” as under 120/80 mmHg (Joint cal of postulated causal factors for HT. National Committee, 2003). Blood pressures between High BP is a treacherous condition because, of 120/80 and139/89 mmHg are now considered prehy- itself, it produces no symptoms. It has become pertensive, while BPs of 140/90 mmHg or above indi- known as the “silent killer” because its complications, cate that hypertension (HT), or high BP is present. affecting the eyes, kidneys, brain, and heart, can be Hypertension is rife in Westernized nations: 25% of deadly. Persons with HT double their risk of heart North American adults and an even higher proportion attack and have four times more risk of stroke than

Complex Chronic Diseases in Evolutionary Perspective 497 potassium (~2500 mg) each day. In contrast, for TABLE 28.2. Normotensive populations. Stone Agers, like all other terrestrial mammals, Hunter-gatherers daily sodium intake (1000 mg/day) was far less Kalahari San (Bushmen) than that of potassium (7500 mg/day). Alaskan Aleuts • Exercise 30 minutes or more on most days. Paleo- Greenland Eskimos lithic humans, who had no motorized equipment or Australian Aborigines draft animals, are estimated to have expended about Congo Pygmies 1240 kcal as physical activity each day. (The 1240 Tanzanian Hadza kcal figure is the averaged value for adult males and Agriculturalists females). The corresponding estimate for Ameri- Mexican Tarahumara Indians cans is 555 kcal per day. Panamanian Cuna Indians • Limit alcohol intake. Hunter-gatherers studied in Kenyan Suba Tribesmen the last century were unable to make alcoholic bev- Ugandan Tesot Tribesmen erages. Regular manufacture of such drinks almost Guatamalan Maya Indians certainly postdates the agricultural revolution. Gambian Tribesmen • Increase fruit and vegetable consumption. In East Africa, the likely human homeland, fruit Horticulturalists Cook Island Polynesians and vegetables provided about 50% of our ances- tors’ food energy intake during the period around Highland New Guinea Tukisenta 50 000 years ago. In the United States, currently, Venezuelan Yanamamo Indians fruits and vegetables provide only about 16% of our Pukapukan Polynesians daily food energy and the contribution in Europe is Solomon Islanders even less. Malaysian Aborigines New Guinea Chimbu The take home message should be inescapable. The Brazilian Caraja Indians life patterns of Stone Age humans protected against, Chinese Noso (Lolo) Aborigines while our own lead to, development of high BP. Only a Surinam Indians few of the normotensive populations in Table 28.2 New Guinea Bomai were hunter-gatherers, but, except for access to alco- holic beverages, they all shared those characteristics Brazilian Kren-Akorore Indians which typified ancestral life and which recent epi- Pastoralists Kenyan Samburu demiological investigations have shown to reduce risk for HT. Kenyan Masai Dental caries do individuals with BPs of < 120/< 80. Untreated HT is We cannot be sure about the frequency of HT for a common cause of congestive heart failure; about half the builders of Stonehenge nor for the Tong Dynasty the cases of end-stage renal disease (kidney failure) are Chinese, but teeth are the best preserved of all human attributable to HT; and hypertensive retinopathy, very remains at archaeological sites and, because of this, common in persons with untreated HT, can lead to the prevalence of dental caries (cavities) can be very blindness. reliably established at different time periods through- In 2002, the NIH published a list of lifestyle modifi- out human experience. While HT causation is almost cations which, together, can greatly reduce a person’s certainly a multifactorial process, the historical correl- risk of developing HT (Whelton et al., 2002). Compar- ation of dental caries with sugar consumption indi- ing these recommendations with the corresponding cates a less complicated genes  risk factor causal features of ancestral life illustrates how deviations relationship. from the life ways for which our genes were selected Researchers have found that, for late Paleolithic can lead to chronic disease: humans before the emergence of agriculture, about • Maintain normal weight. Body mass indices 1–2% of adult teeth showed caries (Brothwell, 1963). (weight/height) between 18.5 and 25 are considered At that time there were no refined sugars. Studies of normal. Studies of multiple hunter-gatherer groups recent hunter-gatherers revealed that they were very reveal an average BMI of 21.5; for Americans, the fond of honey, but because it was available seasonally 2002 average was 28! and was both difficult and painful to obtain, the aver- • Increase potassium and decrease sodium intake. age individual in such cultures consumed only about Americans consume more sodium (~4000 mg) than 2 kg/year. It is likely that intake during the Stone Age

498 S. Boyd Eaton was similar. During the Middle Ages bees had been curves for lung cancer, which in the early 1900s was domesticated, so honey was more available and about considered one of the rarest forms of primary neo- 10% of teeth from medieval times have cavities. plasm, parallel those for cigarette consumption with Cane sugar appears to have been developed around astonishing exactitude. There is a lag period of from 500 BCE in India, but it remained expensive – available 20 to 30 years (which indicates the time necessary almost exclusively to the upper classes, until sugar for tobacco’s carcinogens to induce malignancy), cane cultivation began in the West Indies and large- but both the rising and falling rates of pulmonary scale production was initiated. Even in 1815, average carcinoma follow the cigarette usage rate with near consumption in Britain, of honey and sugar, was only perfect alignment (Witschi, 2001). about 7 kg/year. The second most important risk factor for lung Industrialization made sugar cheap. By 1900 it had cancer is indoor exposure to radon, a radioactive gas, become a staple and, at that period, 60–70% of teeth in which may account for about 10% of American lung Britain were carious. In 2000, American sugar con- cancer deaths (Franklin and Samet, 2001). Radon sumption was nearly 70 kg/year, but because of fluorid- exposure is another example of discordance: despite ation, fluoride-containing toothpaste, and generally being called “cavemen,” Stone Agers spent far less of improved dental care, the dental cavity rate has fallen their lives indoors than do affluent Westerners. to 15–20% – only about 10 times what it was for our Recently studied hunger-gatherers, even those in remote ancestors. colder, upper-latitude climates, spent very little time in caves, rock shelters, or other places where radon might have accumulated. Nearly all their lives, espe- Lung cancer cially during daytime, were spent outdoors. Living, The relationship of lung cancer to cigarette smoking studying, and working indoors allows exposure to is almost as straightforward as that of dental caries radon gas, which originates from the ground and from to sugar. Tobacco is naturally indigenous to the building materials (stone, concrete, etc.) and accumu- Americas, Australia, and a few Pacific Islands. lates in houses, stores, schools, office buildings, and Australian Aborigines have chewed tobacco for many the like. millennia, but for most humans, tobacco exposure postdated the sixteenth-century voyages of explor- Insulin resistance and diabetes ation. Thereafter tobacco availability spread with amazing rapidity, but it was mainly utilized as snuff, Most everyone knows that the prevalence of diabetes for pipes and cigars, and as chewing tobacco. These is climbing rapidly, up nearly 5-fold, 500%, since 1960, forms of tobacco usage fostered cancers of the throat, but the relationship between insulin resistance and larynx, and mouth: Presidents Ulysses Grant and diabetes is less widely recognized. People who are Grover Cleveland, and also Kaiser Frederick III of insulin resistant must secrete more than the usual Germany, were victims illustrating this effect. amount of insulin (from the pancreas) to metabolize Cigarettes apparently originated in Turkey and a given amount of ingested carbohydrate. Insulin spread to Western Europe only after the Crimean War resistant individuals have a high risk of developing (1854–1856) in which Turkey, Britain, and France were type II (adult onset) diabetes and they are also more allied against Russia. Turkish officers introduced cig- likely to manifest HT, obesity, coronary heart disease, arette smoking to their English and French counter- peripheral vascular disease, and the polycystic ovary parts from whom it spread to upper-class Westerners. syndrome. Still, cigarettes remained a minimal component of Like type II diabetes, insulin resistance frequency overall tobacco usage (3% of all tobacco consumption is soaring and, also like diabetes, it is an affliction in 1900) until World War I when cigarettes were issued, of affluence. Australian Aborigines, Japanese Ainu, free, to soldiers as morale boosters. Thereafter cigar- Brazilian Amerindians, Efe pygmies from the Congo, ette usage skyrocketed to a peak in the mid twentieth and !Kung San Bushmen from Africa’s Kalahari Desert century. Subsequent to the US Surgeon General’s were all hunter-gatherers who exhibited exceptional report in 1964, cigarette smoking in the United States insulin sensitivity – the converse of insulin resistance. gradually declined. About 23% of US adults smoked These groups inhabited widely separated geographic in 2006 versus about 46% in 1949, the peak year of regions, but they shared the characteristics of being consumption. lean, exercising a great deal, and consuming few or Cigars, pipes, chewing tobacco, and snuff expose no foods that have excessive insulin-raising (insulino- the nose, oral cavity, throat, and larynx to tobacco’s tropic) effects. effects, but generally spare the lower portions of the As obesity is so strongly linked to type II diabetes, respiratory tract. Conversely, cigarette smoke is usu- weight, and especially weight for height, BMI, is com- ally inhaled, so that it affects the lungs. The incidence monly thought of as particularly important. Actually

Complex Chronic Diseases in Evolutionary Perspective 499 neither weight nor BMI is the critical factor. Rather, Carbohydrate consumption body composition seems most closely related to both diabetes and insulin resistance. Individuals with a Blood glucose elevation greater proportion of fat are more likely, and those Insulin release Insulin release with a greater proportion of muscle, less likely to develop insulin resistance. A descriptive expression is: % body fat Fat Muscle Insulin Resistance  Muscle Fat % muscle mass This relationship is easily explained (Eaton et al., Appropriate blood glucose Insufficient blood glucose 2002a). Fat cells and muscle cells have identical insulin reduction reduction: more insulin receptors which “compete” for circulating insulin required molecules. However, a given amount of fat, say 1 g, 28.1. Carbohydrate metabolism. can extract much less glucose from the blood than can a similar gram of muscle despite equivalent insulin insulin receptors of visceral fat cells are much better stimulation. In effect, insulin molecules interacting positioned than are those of subcutaneous fat cells with adipocyte receptors are being underutilized in when it comes to competing with muscle cell insulin comparison with the blood sugar lowering impact they receptors for circulating insulin molecules. The apple- could have achieved had they interacted with myocyte shaped person has more visceral and the pear-shaped insulin receptors. person fewer visceral fat cells as a proportion of total Physical activity is beneficial because it tends to body fat content. increase muscle tissue and decrease fat as proportions Insulin resistance appears to be a biphasic process. of body weight. Also, exercise increases muscle fitness, Phase I (Figure 28.2a) is a classic example of a correlate of muscle metabolic rate. A gram of fit discordance at work. Body composition differing from muscle extracts more glucose than a gram of unfit the ancestral pattern (too much fat, too little muscle), muscle, given equal exposure to insulin. When muscle exacerbated by low-level muscle fitness produces insu- fitness is taken into account, a more complete descrip- lin receptor imbalance (too many fat cell receptors; too tive expression is: few and insufficiently active muscle cell receptors). % body fat mass These factors interact with our recently adopted high Insulin Resistance % body muscle mass  muscle fitness glycemic load diet to induce repetitive hyperinsuline- mia. These phase I factors are responsible for the rising Studies of recent hunter-gatherers showed them to be insulin resistance and type II diabetes prevalence much fitter and far leaner than age-matched Western- during the past few decades. ers. Also, the skeletal remains of Stone Age humans It is phase II (Figure 28.2b) that fascinates geneti- reveal evidence of impressive muscularity, comparable cists and molecular biologists. Repetitive hyperinsuli- to that of today’s superior athletes (such as Olym- nemia activates genetically determined intracellular pians). These key factors affecting insulin resistance biochemical mechanisms (such as GLUT 4 trafficking) nicely illustrate the discordance hypothesis: our gen- whose cumulative effect is to create intrinsic cellular etic makeup is designed for ancestral circumstances resistance to the actions of insulin. Muscle, fat, and and deviation from the Stone Age pattern promotes liver cells from phase I individuals react normally to dysfunction. insulin (at least early in the condition); the problem at Figure 28.1 diagrams carbohydrate metabolism this stage is tissue imbalance and muscle unfitness. for Stone Age populations, whose end result was However, in phase II these tissues become less sensi- appropriate blood glucose reduction, and for modern tive to insulin stimulation. The adaptive mechanisms couch potatoes who require extra insulin secretion to induced by phase I’s repetitive hyperinsulinemia achieve adequate blood glucose regulation. somehow make body tissues subnormally responsive. Most health-conscious individuals have heard that The genetic and biomolecular phenomena involved in a pear-shaped body configuration (big hips and upper this segue are complex and the subject of intense inves- thighs) is preferable to an apple-shape (big midsec- tigative scrutiny, but cannot, in themselves, account for tion). This difference relates largely to regional blood rising rates of insulin resistance and type II diabetes. flow patterns. Intra-abdominal (visceral) blood flow is It remains to discuss insulinotropic foods 21% of total cardiac output at rest and much more (Lindeberg et al., 2007). These share the characteristics following a meal. In contrast, cutaneous and subcuta- of being newly adopted dietary constituents, in the neous blood flow is only 6% of total cardiac output evolutionary time frame, and of having special potency at rest and less following a meal. It follows that the in raising serum insulin levels. The carbohydrate

500 S. Boyd Eaton (a) (b) Insulin resistance: Insulin resistance: A biphasic process A biphasic process Phase I Phase II Intrinsic cellular Unfit muscle resistance Abnormal Insulin (muscle, fat, liver) body composition receptor Hyperinsulinemia (hyperadiposity, imbalance sarcopenia) Phase I underlies the recent secular increase in diabetes Hyperinsulinemia Biomolecular prevalence. adaptive mechanisms (genetically determined) Phase II cannot explain recent secular increase in diabetes prevalence. 28.2(a and b). The development of insulin resistance. available from sugars and refined cereal grains comes and, as a discipline within the biological sciences, in a form which is rapidly absorbed from the intestinal medicine should also be informed by evolutionary tract, that is, these foods have a high “glycemic index.” awareness. The discordance hypothesis is an attempt Also, both sugars and refined grains contain a high to connect the areas of disease causation and preven- proportion of carbohydrate as a component of their tion with our growing knowledge of human evolution; total weight so they are high “glycemic load” foods. in Kuhnian terms it represents a candidate paradigm (Glycemic load ¼ glycemic index  carbohydrate con- (Kuhn, 1996). As we learn more about the specifics of tent per serving.) Rapid entry of glucose (from digested Paleolithic experience and as our understanding of carbohydrate) – in large amounts – into the blood- human pathophysiology becomes better refined, some stream causes excessive insulin release relative to the of the material presented in this chapter will undoubt- ancestral condition, so both refined cereals and sugars edly require modification, but the basic points will are insulinotropic. Dairy products are different: they almost certainly be confirmed: have relatively low glycemic indices, yet produce dis- • Our genetic makeup is ancient. proportionately high insulin release. The mechanism • “Rapid” cultural change has created discordance involved is currently unknown, but they are also con- between our genes and our lives. sidered insulinotropic. • This discordance, or mismatch, fosters develop- From our genes’ standpoint all three food categor- ment of most complex chronic degenerative ies are Johnny-come-lately’s. Refined cereal grains diseases. were unavailable to Stone Agers who lacked the neces- • Preventing these diseases, and health promotion sary technology and who, in any case, consumed in general, involves reversion towards the basic cereals only during times of food shortage. Whole essentials of ancestral existence (while, ideally, pre- grains have been dietary staples since the advent of serving the positive health effects of cultural agriculture, but truly efficient milling appeared no evolution). earlier than the late nineteenth century. Now 85% of the cereals consumed in the United States are refined. We estimate that Paleolithic humans consumed about 2 kg of sugar (as honey) a year. In 2000 average DISCUSSION POINTS Americans consumed 70 kg/year. And Stone Agers had no domesticated animals (except, perhaps, dogs in the 1. How similar are we genetically to preagricultural latest phases). Consequently, adults during the entire humans? extent of human and prehuman evolutionary experi- 2. Why is the rate of type II diabetes rising around the ence had no dairy foods whatever. world? 3. Is there a “natural” diet for humans? CONCLUSIONS 4. How do NIH recommendations for preventing high blood pressure compare with our understanding of Evolution is increasingly accepted as the core around the ancestral human lifestyle? which our understanding of biology must be constructed, 5. What is the “Discordance Hypothesis”?

Complex Chronic Diseases in Evolutionary Perspective 501 REFERENCES Eaton, S. B., Strassman B. I. and Nesse R. M. (2002a). Evolutionary health promotion. Preventive Medicine, 34, Brothwell, D. R. (1963). The macroscopic dental pathology 109–118. of some earlier human populations. In Dental Anthropol- Eaton, S. B., Cordain, L. and Lindeberg, S. (2002b). Evolu- ogy, Brothwell, D. R. (ed.). New York: Pergamon Press, tionary health promotion: a consideration of common pp. 271–318. counterarguments. Preventive Medicine, 34, 119–123. Cordain, L., Brand-Miller, J., Eaton, S. B.et al. (2002). Plant- Ellison P. T. (2001). On Fertile Ground. A Natural History of animal subsistence ratios and macro-nutrient energy Human Reproduction. Cambridge, MA: Harvard Univer- estimations in worldwide hunter-gatherer diets. American sity Press. Journal of Clinical Nutrition, 56, 181–191. Franklin, H. and Samet, J. M. (2001). Radon. CA: A Cancer Cordain, L., Eaton, S. B., Sebastian, A. et al. (2005). Origins Journal for Clinicians, 51, 337–344. and evolution of the Western diet: health implications Joint National Committee (2003). The seventh report of the for the twenty-first century. American Journal of Clinical joint national committee on prevention, detection, evalu- Nutrition, 81, 341–354. ation, and treatment of high blood pressure. Journal of the Eaton, S. B. (2006). The ancestral human diet: what was it American Medical Association, 289, 2560–2572. and should it be a paradigm for contemporary nutrition? Kuhn, T. S. (1996). The Structure of Scientific Revolutions, Proceedings of the Nutrition Society, 65, 1–7. 3rd edn. Chicago: University of Chicago Press, p. 148. Eaton, S. B. and Eaton, S. B. III (1999a). The evolutionary Sebastian, A. (2005). Dietary protein content and the diet’s context of chronic degenerative diseases. In Evolution in net acid load: opposing effects on bone health. American Health and Disease, Stearns S. C. (ed.). Oxford: Oxford Journal of Clinical Nutrition, 82, 921–922. University Press, pp. 251–259. Whelton, P. K., et al. (2002). Primary prevention of hyperten- Eaton, S. B. and Eaton, S. B. III (1999b). Breast cancer in sion: clinical and public health advisory from the National evolutionary context. In Evolutionary Medicine, Trevathan High Blood Pressure Education Program. Journal of the W. R., Smith E. O., McKenna J. J. (eds). Oxford: Oxford American Medical Association, 288, 1882–1888. University Press, pp. 429–442. Witschi, H. (2001). A short history of lung cancer. Toxological Eaton, S. B. and Eaton, S. B. III. (2003). An evolutionary Sciences, 64,4–6. perspective on human physical activity: implications for Yazdanbaksh, M., Kremser, P. G. and van Ree, R. (2002). health. Comparative Biochemistry and Physiology. Part A, Allergy, parasites, and the hygiene hypothesis. Science, 136, 153–159. 296, 490–494.

29 Evolutionary Medicine and the Causes of Chronic Disease Paul W. Ewald The current integration of evolution with medicine is If pathogens have adaptations that allow them to evade artificially narrow because it reflects the biases of the immune system, they can persist indefinitely. Ill- medicine as a whole and the specializations of particu- nesses caused by such persistent pathogens tend to be lar investigators. If evolutionary principles are to offer chronic. The three-month dividing line between acute a fundamental framework for understanding medical and chronic diseases therefore corresponds roughly issues they should help identify these biases and the with an immunological basis for rapid resolution of areas over which integration needs to be broadened. infectious challenges versus protracted conflicts. The chapter discusses this problem as it relates to The causes of chronic diseases are not well under- evolutionary interpretations of chronic diseases. It also stood largely because the processes of causation tend illustrates how an evolutionary perspective can provide to be inconspicuous and difficult to evaluate (Cochran a broader framework for resolving the causes and con- et al., 2000). For the past century the prevailing view of trol of chronic diseases, focusing on the greatest killers modern medicine has been that chronic diseases result in prosperous societies – atherosclerosis and cancer. largely from interactions between a person&s genetic vulnerabilities and environmental influences. These interactions determine the rate at which the body OVERVIEW OF CHRONIC DISEASES succumbs to chronic disease. Evolutionary medicine has recast this problem Chronic diseases account for about 70% of the mortal- by asking why the process of natural selection has ity in the United States and other wealthy countries. resulted in the existing collection of vulnerabilities to Most of this disease-induced mortality is attributable chronic disease (Nesse and Stearns, 2008). The investi- to heart attacks, strokes, and cancer. In spite of this gative framework of evolutionary medicine integrates importance, the causes of chronic diseases remain the mechanistic (proximate) explanations, which largely unresolved. This situation represents a major address how biological processes work, with evolutio- short-coming of modern medicine because under- nary (ultimate) explanations, which address why standing the causes of disease enables prevention, biological evolution has resulted in these processes. and prevention is the most effective way of eliminating Proximate causes of human diseases include human the damage inflicted by disease. We can therefore alleles and environmental factors that contribute to expect that a better understanding of the causes of pathogenesis. Ultimate causes explain why humans chronic diseases will yield some of the most valuable have evolved to be vulnerable to such influences, and contributions that any discipline can make to improve- if an environmental cause is a parasite why the parasite ments in health. has evolved to cause this damage. Chronic diseases can be defined broadly as diseases In evolutionary medicine investigations of the that persist within individuals for a long time. The US causes of chronic disease generally focus on the differ- National Center for Health Statistics considers “long” ences between ancestral and modern environments. to mean persistence for at least three months. Although The prevailing emphasis is on the mismatch between this dividing line between acute and chronic is some- our modern environment and our evolved biology. what arbitrary, a biological basis for distinguishing Aspects of modern diets that differ from ancestral acute and chronic diseases can be made for infectious diets, for example, are raised as explanations for diseases. If an immune response can readily eliminate diseases such as atherosclerosis and diabetes (Eaton an infection, it generally does so within several weeks. et al., 1988, Nesse and Williams, 1995; Gerber and Infectious diseases that are controlled (or kill the host) Crews, 1999; Leonard, 2008). Modern changes in during this period are generally considered acute. nondietary exposures to cigarette smoke, ultraviolet Human Evolutionary Biology, ed. Michael P. Muehlenbein. Published by Cambridge University Press. # Cambridge University Press 2010. 502

Evolutionary Medicine and the Causes of Chronic Disease 503 radiation in light skinned people, or birth control, are acute diseases emphasize parasitic causes, whereas suggested as reasons for increased rates of cancers studies of chronic diseases emphasize the evolved (Williams and Nesse, 1991; Eaton et al., 1994; Nesse characteristics of humans and environmental factors and Williams, 1994; Diamond, 2005). Increased life other than parasites. A central challenge for evolution- span in modern environments is suggested as a reason ary medicine is to determine whether this discordance for chronic diseases associated with old age (Nesse reflects a real difference between acute and chronic and Williams, 1995; Austad 1997; Finch, 2007; Austad diseases, biases against considering infectious causes and Finch, 2008). Human bodies are presumed to be of chronic diseases, or both. vulnerable to such proximate causes of chronic dis- Although infectious causes of chronic diseases have eases because exposure to these proximate causes been accepted for over a century, the clarity of hind- has been too recent to purge these vulnerabilities sight shows that this category of causation has been by natural selection. Evolutionary medicine also underappreciated in evolutionary medicine. Illnesses considers the inability of natural selection to perfect such as peptic ulcers and cancers, for example, were adaptations, due, for example, to constraints on deve- explained by the mismatches between modern and lopmental processes or the limited power of natural ancient environments without consideration of infec- selection to perfect adaptations (Nesse, 2008). tious causation, even when evidence implicating These arguments of evolutionary medicine gener- infectious causation was well developed (e.g., Nesse ally accept the consensus of nonevolutionary medicine and Williams, 1995). It is now broadly accepted that on the proximate causes of disease and develop evolu- peptic ulcers, cancers of the stomach, liver, cervix, and tionary explanations for these proximate causes. This oropharyngeal, and nasopharyngeal regions are caused consensus of nonevolutionary medicine is generally largely by infectious processes (Ewald, 2000; Cochran that the proximate causes of chronic diseases result et al., 2000; Greaves, 2000). The overlooking of infec- from an interaction between human genetics and non- tious causation may be much more significant than infectious environmental factors, the familiar gene by these examples indicate because, as will be discussed environment explanations. But the proximate causes of in this chapter, evidence indicates that many other chronic illness are still largely uncertain. If evolution- important chronic diseases, such as atherosclerosis, ary medicine generates ultimate explanations for the Alzheimer&s disease, breast cancer, and schizophrenia, wrong proximate explanations, the ultimate explan- may be caused largely by infectious processes. ations will be wrong as well. It is therefore important To avoid a bias against identifying infectious for research in evolutionary medicine to consider the causes, disease causation can be considered schemat- full range of feasible proximate explanations of disease ically as a causal triangle with each apex of the triangle to safeguard against offering incorrect ultimate expla- corresponding to one of the three categories of cau- nations. Moreover, evolutionary insights may facilitate sation (Figure 29.1). The placement of any disease evaluation of alternative proximate mechanisms of within the triangle emphasizes the relative importance pathogenesis by distinguishing those mechanisms are of the three categories of disease causation. Diseases evolutionarily feasible from those that are not. are placed closest to the vertex that corresponds to the These arguments apply in principle to acute dis- primary cause of the disease, but the location within eases, but the causes of acute diseases are understood the triangle (instead of at the vertex itself), emphasizes better than are the causes of chronic diseases; there that other categories contribute to pathogenesis. is therefore less danger of proposing an evolutionary A primary cause is defined as one that is necessary explanation for an incorrect proximate mechanism of for the disease to occur. Prevention of the primary acute diseases, and little need for new evolutionary cause(s) of a disease prevents the disease. Prevention insights to identify the causal agents of acute diseases. of a secondary cause will reduce the frequency or The overwhelming majority of acute diseases are severity of disease but will not prevent the disease caused by parasitism, broadly defined to include multi- itself. Mycobacterium tuberculosis, for example, is a cellular, cellular, and subcellular parasites (infectious primary cause of tuberculosis, because prevention of diseases being defined as parasites that live in intimate Mycobacterium tuberculosis infection will prevent association with host cells and tissues). The evolutio- tuberculosis. Poor nutrition and genetic vulnerabilities nary study of acute disease has therefore focused on to M. tuberculosis are secondary (or exacerbating) the degree of harmfulness (virulence) to which host/ causes of tuberculosis, because tuberculosis can still parasite associations evolve and the evolutionary func- occur in individuals who have good nutrition or are tion of disease manifestations – do the manifestations genetically less vulnerable if other conditions such as increase the evolutionary fitness of the host, the para- immune suppression or intense exposure to M. tuber- site, both, or neither (Ewald, 1980, 1994). culosis occur. Sometimes more than one category of A discordance between the evolutionary study of causation must be present for a disease to occur. acute and chronic disease arises because studies of Cancers for which pathogens are accepted primary

504 Paul W. Ewald Genetic 29.1. The triad of disease causation. The diagram empha- sizes joint influences of the three categories of disease caus- ation. The placement of a disease corresponds to the relative importance of the three categories with the closest Cystic fibrosis apex indicating the primary cause as defined in the text. The vertices that are more distant represent exacerbating (i.e., secondary) causes or coprimary causes of secondary import- ance. Cystic fibrosis, for example, is a genetic disease that is exacerbated by a variety of different pulmonary infections and low-salt diet. Skin cancer refers to squamous cell skin cancers. The placement of tuberculosis and skin cancer is Skin cancer ? Tuberculosis Noninfectious discussed in the text. environmental Infectious causes, for example, generally also require additional lead to several paradoxes that suggest major shortcom- mutations. In this case environmental mutagens and ings of these explanations. Each of these paradoxes is the infectious agents are coprimary causes, and the resolved when hypotheses of infectious causation are cancer would be located roughly equidistant from integrated into the analysis (Ewald, 2008). the parasitic and environmental vertices. By consider- One paradox arises from variation in the frequency ing this triangle of causation researchers can avoid of the E4 allele of the APOE gene in humans and other falling into the trap of failing to consider the validity primates. Its frequency ranges from about 5–40% in of one category of primary or secondary causation, different human populations (Corbo and Scacchi, just because the validity of another category has been 1999; Fullerton et al., 2000). Comparisons of nucleo- identified. tide sequences show that it is the ancestral allele of The rest of this chapter will apply this perspective Homo sapiens and of primates generally (Fullerton to specific chronic illnesses that have been discussed et al., 2000) and that its relative frequency has declined in the literature of evolutionary medicine. I emphasize over the past 200 000 years of human evolution relative feasible hypotheses of disease causation that have not to the other common alleles, E2 and especially E3 been considered, especially when supportive evidence (Fullerton et al., 2000). The E4 allele therefore is not exists in the literature. As a result of the bias against an aberrant mutant allele. Some selection pressure in infectious causation of chronic disease mentioned humans must have disfavored E4 for a long time but above, the overlooked hypotheses generally involve a differently in different human populations. failure to consider hypotheses of infectious causation A diet-based hypothesis proposes that rich agricul- and the interplay between infectious causation and the tural diets are responsible (Corbo and Scacchi, 1999). other categories of causation. Accordingly, the E4 allele frequency is about 5–15% in populations with several thousand years of agricultural experience and about 20–40% in populations that have Atherosclerosis been living largely as hunter-gatherers until the twen- By effecting heart attacks and strokes, atherosclerosis tieth century (Corbo and Scacchi, 1999). These figures is the leading cause of mortality in the United States indicate that the allele has been declining more rapidly and other wealthy countries (http://www.cdc.gov/ in agricultural settings. But E4 is a minority allele even nccdphp/overview.htm). In accordance with its import- among recent hunter-gatherers. The most significant ance, vast economic and intellectual resources have decline in E4 frequency therefore occurred prior to the been spent to identify risk factors for atherosclerosis. onset of agriculture. Dietary constituents have been implicated as exacer- Finch and Stanford (2004) suggest that this shift bating influences (cholesterol, saturated fats, dietary may have resulted from a shift toward increased meat iron) and ameliorating influences (unsaturated fats, eating associated with the early evolution of H. sapiens. omega 3 fatty acids, garlic). The most important Although this timing is consistent with the reduction in genetic risk factor for atherosclerosis that has been E4 frequencies, it remains an untested hypothesis identified is the epsilon 4 (E4) allele of the apolipopro- because the change in the makeup of the human diet tein E (APOE) gene (Ilveskoski et al., 1999; Mahley and during this time is unknown, as is the way in which Huang, 1999). any such change might alter the role of E4 in lipid Almost all of the literature on the causes of athero- deposition. sclerosis in evolutionary medicine attempts to explain The APOE proteins transport fat and cholesterol. human vulnerability to atherosclerosis in the context Proximate explanations have therefore presumed that of such vulnerability to genetic and noninfectious the association between E4 and risk of atherosclerosis environmental risk factors. These analyses, however, results from some difference between E4 and the other

Evolutionary Medicine and the Causes of Chronic Disease 505 APOE proteins in transport of lipids. But a decade ago, At first glance, the beneficial effects of statins Ge ´rard et al. (1999) showed that people harboring seems to offer strong support for the cholesterol the E4 allele were more likely to be infected with the hypothesis, because statins lower serum cholesterol major candidate infectious cause of atherosclerosis: levels and reduce rates of heart attacks. This reduction Chlamydophila (Chlamydia) pneumoniae. Recently the in heart attacks, however, was more powerful than same research group has shown that the E4 protein could be accounted for by the cholesterol-lowering facilitates attachment of C. pneumoniae to human cells effects of statins (Mays and Dujovne, 2008). Importantly, (Ge ´rard et al., 2008). The evidence therefore supports statins are associated with reductions in inflammation the hypothesis that a fitness cost of the E4 allele is and infection-induced mortality in atherosclerosis increased vulnerability to C. pneumoniae. Unlike the patients (Almog et al., 2007). These findings suggest that agricultural diet hypothesis, the hypothesis that E4 the beneficial effects of statins may arise not from their exacerbates atherosclerosis by increasing vulnerability cholesterol-lowering effects, but from anti-inflammatory to C. pneumoniae is consistent with the timing of the or antimicrobial effects (Jerwood and Cohen, 2008; evolutionary decline in E4 frequency, so long as Mays and Dujovne, 2008). C. pneumoniae has been a pathogen of humans throug- Taubes (2008) has pointed out another paradox ® hout the evolution of H. sapiens (Ewald and Cochran, of the cholesterol hypothesis. The drug Vytorin was 2000; Ewald, 2008). The extent to which such a long- approved for lowering the risk of heart attacks based standing association has occurred should be testable on its cholesterol-lowering effects. These effects arise ® using molecular phylogenetic reconstructions of from the two active components of Vytorin . The first ® Chlamydophila. The hypothesis that a shift toward an component is a statin called Zocor . The second is a ® atherosclerosis-inducing diet occurred early during drug, called Zetia , which lowers cholesterol by a dif- the evolution of H. sapiens is also consistent with this ferent mechanism. The addition of Zetia ® to Zocor ® early shift away from E4 (Finch and Stanford, 2004), was approved because it lowered cholesterol more than ® ® but this hypothesis will be more difficult to test. Unlike Zocor alone. The addition of Zetia , however, did not ® the infectious causation hypothesis, this early sapiens lower the risk of heart attacks more than Zocor alone. diet hypothesis does not have empirical support from Taubes (2008) rightly points out that this finding serves a proximate pathogenic mechanism, because we do as a critical test of the cholesterol hypothesis and that not know whether the diet of early H. sapiens involved the cholesterol hypothesis failed the test. a shift toward constituents that fostered atheroscler- These considerations suggest that the paradoxes osis. This argument illustrates the need for a broader over the role of cholesterol resulted from premature consideration of alternative hypotheses in evolutionary acceptance of a reasonable hypothesis rather than a medicine, because the proponents of the early sapiens more comprehensive consideration of the evidence in diet hypothesis made no reference to the possibility the context of the full range of feasible hypotheses. that C. pneumoniae could be responsible for the Integrating hypotheses of infectious causation resolves reduction in the E4 frequency (Finch and Stamford, these paradoxes, by revealing that the incongruous 2004; Finch, 2007; Austad and Finch, 2008), even findings are consistent with a more cohesive theory though the vulnerability of E4 individuals to C. pneu- of pathogenesis. The accumulation of cholesterol moniae and the relevance of this vulnerability to ath- within the walls of arteries, for example, is prece- erosclerosis has been in the literature for about a ded by the accumulation of cholesterol-laden macro- decade (Ge ´rard et al., 1999, Ewald, 2000; Ewald and phages, called foam cells, within the arterial walls Cochran, 2000). (Crowther, 2005). Chlamydophila (Chlamydia) pneumo- Another paradox concerns the widely accepted niae can infect macrophages and transform them into hypothesis that cholesterol is a cause of atheroscle- foam cells by fostering the uptake of fat and choles- ® rosis. Cholesterol accumulates within arterial walls terol. Zetia may fail to decrease cardiovascular events early during the pathogenesis of atherosclerosis because elevated serum cholesterol is a correlate and (Crowther, 2005). When it was thought that this accu- perhaps a minor exacerbating cause of these events mulation was on the inside surface of the arteries it rather than a primary cause. was reasonable to hypothesize that high cholesterol A similar resolution applies to dietary components diets would lead to high levels of cholesterol in the that appear to reduce the risk of heart attacks. Con- blood, which in turn would lead to higher deposition sider garlic. Over the past two decades, investigations of cholesterol. The link between elevated serum choles- of apparently beneficial effects of garlic on atheroscle- terol and deposition of cholesterol within the arterial rosis have focused almost entirely on alterations of walls is less convincing, but the elevated cholesterol lipids by a component of garlic: allicin. But published hypothesis can still be rescued if the cholesterol is studies have not documented consistent associations. somehow selectively transported to sites within arterial Recently the controversy was largely put to rest by the wall without being deposited on the lining of arteries. most thorough study yet on the relationship between

506 Paul W. Ewald garlic and lipid composition, which showed no signifi- arguments tend to implicate the mismatch between cant effects (Gardner et al., 2007). The results of this ancestral and modern environments. It is argued, for study may be invoked by some to dismiss the hypothe- example, that the increased life spans of modern sized beneficial effects of garlic against atheroscler- humans increases the prevalence of cancer because osis. Yet the results of that study were predicted by cancers tend to occur during the later decades of life. a hypothesis that invokes the antibiotic effects of The proximate part of this argument focuses on the garlic instead of a lipid-altering effect (Ewald, 2008). greater chance for the accumulation of oncogenic This antibiotic hypothesis has not yet been specifically mutations and disregulation of cellular processes as tested because researchers have been so narrowly life span increases; the evolutionary part focuses on focused on the lipid hypothesis of atherosclerosis the weakness of natural selection to counteract cancers pathogenesis. Studies of the role of garlic on athero- that occur late in life (Greaves, 2000). To explain why sclerosis need to take a more balanced approach by genetic predispositions to cancer might be spreading in investigating the possibility that garlic may ameliorate human populations, evolutionary hypotheses raise the atherosclerosis by inhibiting pathogens that are impli- possibility that the genetic predispositions are alleles cated as primary causes of atherosclerosis. that have been favored recently in human evolution Unlike the other hypotheses that have been pro- and an that increased cancer risk has tagged along as posed to explain the atherosclerosis, the infectious a pleiotropic cost (Crespi and Summers, 2006). hypotheses of atherosclerosis do not generate such Analyses of oncogenesis in evolutionary medicine interpretive paradoxes. Moreover, predictions of the focus on environmental mutagens and inherited pre- infectious causation hypothesis have been confirmed dispositions (Crespi and Summers, 2005; Merlo et al., by the most direct tests: the presence of C. pneumoniae 2006; Frank, 2007; Greaves, 2008; Komarova and and other pathogens in the foci of disease, elevated Wodarz, 2008). A testament to this oversight is found positivity in afflicted patients, and mechanisms that in the most comprehensive text of biological topics in link infection with the other risk factors in a causal evolutionary medicine (Stearns and Koella, 2008), network (e.g., C. pneumoniae fostering foam cell where cancer is included in the section on “Noninfec- formation in early stages of atherosclerosis, and E4 tious and degenerative diseases,” even though infec- increasing vulnerability to C. pneumoniae; see Ewald tious causation can be definitively ruled out for only a and Cochran, 2000 and Ewald, 2008 for more details). miniscule proportion of human cancer. When infec- These considerations of atherosclerosis illustrate tions are considered (Greaves, 2000, 2006), infection the value of using the triangle of causation as a basis has not been integrated with mutation, and inheritance for analyzing disease causation. The evidence pertain- into a unified theory of oncogenesis. ing to E4 illustrates how paradoxes associated with the As was the case with peptic ulcers and atheroscler- most important genetic vulnerability to atherosclerosis osis, considerations of cancer within evolutionary are resolved when the analysis is broadened to include medicine has been lagging behind rather than advan- the infection vertex. The evidence pertaining to choles- cing demonstrations of infectious causation. Over the terol shows how paradoxes associated with one of the past 30 years a major trend in the understanding of most widely incriminated dietary agents of atheroscler- oncogenesis has been recognition of an increasing role osis are resolved when the analysis includes infection. for infectious causes of cancer, mirroring the trend for The evidence pertaining to garlic shows how an ameli- chronic diseases in general (Ewald, 2009). In the mid orating environmental influence could be inappro- 1970s unicellular or subcelluar parasites were gener- priately rejected because the spectrum of hypotheses ally accepted as a cause of only one human cancer: under consideration was inappropriately narrow. In Burkitt&s lymphoma, which is caused by simultaneous each case evolutionary considerations are useful in gen- infection of Epstein–Barr virus and Plasmodium falci- erating and evaluating alternative causal explanations. parum. With regard to multicellular parasites, some trematodes were recognized as contributing to liver and bladder cancer. The total amount of cancer Cancer accepted as being caused by parasitism was therefore Cancer is characterized by the pathological prolifer- less than 1% of all human cancer. Over the past three ation and systemic spread of cells. It is widely accepted decades, this percentage has increased steadily. Now that cancers result primarily from mutations that about 20% of all human cancer is accepted by the promote cell proliferation and inhibit cell adhesion World Health Organization as being caused by parasit- (leading to metastasis). Consequently, researchers ism. Although overviews of cancer often presume that writing from an evolutionary perspective have attem- the remaining 80% are caused by something other than pted to explain why the human body is vulnerable to parasitism, this conclusion is not justified by the evi- such mutation-induced disregulation of proliferation dence. Because all three categories of disease caus- and adhesion. As is the case with atherosclerosis, ation may act in concert, identification of a cause in

Evolutionary Medicine and the Causes of Chronic Disease 507 one category cannot be used as evidence that any other cancer. Rather, the breakdown of the barriers to cancer category is invalid. The documented importance of favor persistence within the host but nudge infected inherited predispositions, environmental mutagens, cells toward cancer. An infection with any of the known and the genes they mutate therefore does not imply oncogenic viruses is by itself insufficient to cause that infectious agents are playing no causal role. human cancer, because only a small proportion of the This idea is well illustrated by skin cancer. There is people who are infected with any one of these viruses general agreement that ultraviolet rays contribute to skin will develop cancer. Other causes must therefore cancer (Greaves, 2000; Kleinsmith, 2005; Karagas et al., contribute. 2007). One inherited genetic vulnerability for skin Consideration of breast cancer offers an illustra- cancer is low amounts of protective skin pigment tion of the need for an integrated approach to oncogen- (Greaves, 2000, Kleinsmith, 2005). This vulnerability esis. The most widely accepted evolutionary explanation can be attributed largely to a mismatch between ances- for breast cancer has been advanced by Boyd Eaton and tral and modern environments – the ancestry of people his colleagues (Eaton et al., 1994; Eaton and Eaton, 1999; with light skin color can be traced to northern latitudes, Greaves, 2008; Stearns et al.,2008; Trevathan et al.,2008). where exposure to ultraviolet radiation has tended to be This explanation, which can be labeled the hormonal low. Light skinned people tend to have high cancer rates proliferation hypothesis, focuses on oncogenic effects of when they live in more southerly environments with estrogen and, to a lesser extent, progesterone. It proposes higher exposure to ultraviolet radiation. Both evolution- that elevated rates of breast cancer in economically ary and proximate analyses of the causes of skin cancer wealthy societies may result from enhanced cyclic expos- have been largely restricted to these two categories with ure to these hormones. little consideration of infectious causation. Yet evidence Thehormonalproliferationhypothesispointsoutthat implicating nongenital serotypes of human papilloma- estrogen and perhaps progesterone contribute to the virus (HPV) as a cause of squamous cell skin cancer has proliferation of cells in breast tissue (Potten et al., 1988) been published over the past decade. The evidence is and attributes elevated rates of breast cancer in modern particularly strong for people with suppressed immunity societies to a mismatch between modern and ancestral and the rare hereditary skin disease epidermodysplasia environments. It proposes that the mismatch results from verruciformis, and probably involves disregulation of modern birth control and modern diets. Birth control control over cellular division (Jenson et al., 2001; Karagas increases the number of menstrual cycles during a life- et al., 2006; Andersson et al., 2008). These considerations time because women do not cycle during pregnancy and suggest that HPVs may be coprimary causes or exacer- much of lactation. Rich diets apparently increase the bating causes of squamous cell skin cancer (hence the number of cycles by advancing menarche and delaying arrow with the question mark in Figure 29.1). Consider- menopause. Studies within post-industrial societies show ing only epidermodysplasia verruciformis and HPV the that women who begin menstruation at younger ages and evidence supports a central location in the diagram. end menstruation at older ages have higher risks ofbreast Cells have four critical barriers to cancer. Cell cycle cancer than women whose menstrual cycles occur over arrest keeps the cell from dividing. Apoptosis (cell suicide) a more restricted range of years (Key and Pike, 1988; can destroy proliferating cells before they progress to Trichopoulos et al., 1996). As a result of these differences, metastatic cancer. Restriction of telomerase can block Eaton et al. (1994) estimate that women in the United oncogenesis by placing an upper limit on the total States have nearly three times as many menstrual cycles number of divisions that a cell lineage undergoes. Cell during their lifetimes as women in hunter-gather soci- adhesion prevents metastasis. eties. They estimate that US women have about 100 times Some viruses have evolved intricate mechanisms to more breast cancer. interfere with these barriers, apparently to foster per- As Eaton and his colleagues note, the risk of breast sistence of the viruses within their hosts. By causing cancer is greater during pregnancy (Bruzzi et al., 1988; the cells that they infect to divide, the genetic material Williams et al., 1990). The increased risk persists during of a virus can replicate in concert, while incurring little the first year after childbirth and declines gradually exposure to the immune system. By interfering with thereafter. Prior to menopause the declining risk of apoptosis the virus can keep the cell from destroying breast cancer with increasing time since last birth the virus via cell suicide. By increasing telomerase approaches but does not drop below that of nulliparous activity, the virus can push the infected cell toward women (Bruzzi et al., 1988; Williams et al., 1990; Lambe immortality, perpetuating this profitable exploitation et al., 1994). Protection associated with a greater number of the host cell for both resources and protection. By of births was significant only for women in their 40s interfering with cell-to-cell adhesion, infected cells can (Bruzzi et al., 1988; Williams et al., 1990). These associ- spread to other parts of the body to facilitate further ations raise a paradox if one assumes that pregnancy viral proliferation and transmission. This argument protects against breast cancer: no protective effect of does not suggest that pathogens benefit from lethal pregnancy is apparent after birth among premenopausal

508 Paul W. Ewald women. To the contrary, parous women have greater risks proliferation. But this reduced control of proliferation of premenopausal breast cancer than nulliparous women would have to occur after menopause but not detectably of the same age, but lower risks of cancer during the before menopause. decade of life in which menopause generally occurs. The protective association with parity for postmeno- Eaton and his colleagues accommodate this paradox by pausal breast cancer and the exacerbating association of restricting their hypothesis to postmenopausal breast parity prior to menopause is a paradox that needs to be cancer. At best this restriction leaves much of the elevated resolved by a general theory of breast cancer. If parity breast cancer in post-industrial societies – that occurring protects against breast cancer, why does it protect only premenopausally – without an adequate evolutionary after menopause? Higher rates of estrogen during preg- explanation. nancy may play a role, but the paradox is not resolved The restriction of the hormonal proliferation by direct effects of estrogen on proliferation rates, hypothesis to postmenopausal breast cancer raises because a large proportion of these cancers do not another paradox. How can the proliferative effects of expressthe receptorsfor estrogen and progesterone. Also, estrogen cause increased cancer rates only after estro- the persistence of elevated breast cancer risk during the gen levels have declined? Cognizant of this paradox, first year after parturition is not explainable as a direct Eaton and Eaton (1999) conclude that the main onco- effect of estrogen, because estrogen levels decline within genic effects of estrogen result from effects of prolifer- a week or so after pregnancy (Doyle et al., 2007). ation on mutation rather than the nudging of Consideration of infectious causation resolves proliferation higher. This resolution represents a major these paradoxes. An immune suppression/infection restriction of the estrogen proliferation hypothesis to hypothesis proposes that infectious agents may act in the proliferative effects on mutation, thus abandoning concert with mutations and inherited predispositions any direct contribution of hormone-induced prolifer- to cause breast cancer. It does not deny the potential ation on the runaway proliferation that characterizes applicability of estrogen and progesterone. But rather cancer. than focusing on the proliferative effects of these hor- Another paradox arises because protection against mones, it emphasizes their suppressive effects on cell peri- and postmenopausal breast cancer appears to be mediated immunity (Doyle et al., 2007, 2008), which associated with pregnancies during the first decade or may increase vulnerability to persistent viruses. so after menarche. Eaton and Eaton (1999) resolve Three persistent viruses have been strongly associ- this paradox by assuming that the relevant prolifer- ated with breast cancer: Epstein–Barr virus (EBV), ation is occurring restricted to this time period. As a HPV (particularly serotypes 16 and 18), and mouse result of earlier menarche, women in post-industrial mammary tumor virus (MMTV; human isolates are societies have more menstrual cycles prior to first generally referred to as MMTV-like). Each of these pregnancies than women in hunter-gatherer societies, viruses is found in about 25–50% of breast cancers but and thus may have more breast cell proliferation that in less than about 10% of the normal breast tissue from could predispose them to a greater incidence of breast the same patients (Wang et al., 1995; Bonnet et al., 1999; cancer. However, because the elevated rates of breast Fina et al., 2001; Lawson et al., 2001; Kleer et al., 2002; cancer in post-industrial societies generally occur Damin et al., 2004; Kan et al., 2005; de Villiers postmenopausally, decades after this proliferation, et al., 2005). Each virus stimulates proliferation of the proliferation itself does not seem to be causing infected cells and inhibits apoptosis (Subramanian and breast cancer directly by nudging up proliferation Robertson, 2002; Katz et al., 2005; Knight et al., 2005; rates. Eaton and his colleagues therefore suggest that Mileo et al., 2006; Guasparri et al., 2008; Hino et al., the mechanism by which proliferation increases 2008). Both EBV and HPV are also known to promote cancer is mainly through increased mutation prior to cellular immortalization by enhancing telomerase activ- the first birth. ity (Ding et al., 2007; Liu, 2008; Tungteakkhun and After menopause proliferation of breast cells is twice Duerksen-Hughes, 2008; investigations of effects of as great for nulliparous women than for parous women. MMTV on telomerase activity have not yet, to my know- Butthisdifferencedoesnotaccordwellwiththe ledge, been conducted). hormonal proliferation hypothesis, because women are Each of these viruses interferes with cell-to-cell adhe- not experiencing the elevated levels of estrogen after sion, thereby promoting metastasis. Expression of the menopause. Nor does this difference accord with the cellular adhesion molecule, nm23H1, declines as breast histological changes that are associated with parity, and other cancers progress to lethal metastatic out- because at menopause breast cell types revert to the com- comes (Branca et al., 2006; Mileo et al., 2006; Sgouros position found in nulliparous woman (Russo, 1992). et al., 2007). The MMTV inhibits expression of the One can save the hormonal proliferation hypothesis by nm23H1 (Ouatas et al., 2002) and EBV inhibits the abil- arguing that mutations that occur soon after puberty ity of nm23H1 to suppress cell migration (Subramanian are particularly associated with reduced control of and Robertson, 2002; Murakami et al., 2005). Human

Evolutionary Medicine and the Causes of Chronic Disease 509 papillomavirus downregulated and inactivated nm23H1 such pathogens because women who are raising families in keratinocytes (Mileo et al., 2006). In a study using tend to have fewer sexual partners than single women. esophageal cells, HPV16 did not inhibit nm23H1 but Moreover, nulliparous women may be nulliparous did inhibit a different cellular adhesion molecule, because they have had a relatively high exposure to CD44v6 (Liu et al., 2005), decreased levels of which are sexually transmitted pathogens which can reduce fertil- associated with progression to metastatic cancer (Maula ity (as is the case with Chlamydia trachomatis and et al., 2001). Neisseria gonorrheae). Whether any one of these viruses must act with any The immune suppression/infectious hypothesis other to generate cancer is unknown. These associations predicts a temporal pattern of mutations that is not could therefore explain anywhere from about half to predicted by the hormonal proliferation hypothesis. nearly 100% of human breast cancer. Regardless of this Specifically, if simultaneous inhibition of the cancer number, evidence indicates that any oncogenic action of barriers by viruses is needed to initiate oncogenesis, these viruses occurs together with oncogenic mutations, the additional mutations of barriers to cancer should because oncogenic mutations are found in a substantial tend to occur later during oncogenesis. In contrast the proportion of breast cancers. But consideration of the hormonal proliferation hypothesis proposes that such interdependence of the different steps of oncogenesis mutations are initiating events and should therefore suggests that viral sabotaging of barriers to cancer gener- often at the onset of oncogenesis. ally occurs before oncogenic mutations. Activation of This prediction is testable, as illustrated by a cell replication without inhibition of cellular senescence, recent study (Saal et al., 2008) of a gene referred to as for example, would have low oncogenic potential because PTEN (for Phosphatase and TENsin homolog) in it would generate only a limited number of cellular women who have inherited an mutation in a breast divisions. Similarly, inhibition of cellular senescence cancer susceptibility gene called BRCA1 (for BReast would be of limited value without inhibition of apoptosis, CAncer susceptibility gene type 1). The PTEN gene which would have the potential of destroying the infected fosters cellular replication and inhibits apoptosis. The cell. Without infectious causation, a specific sequence BRCA1 gene encodes a protein that repairs mutations. of very specific mutations would have to compromise Mutations in the BRCA1 gene therefore contribute to these barriers without destroying the cell&s viability. cancer by reducing the repair of mutations in other In contrast, infection by the three viruses mentioned genes that encode barriers to cancer, and women who above typically inhibit three or four of the key barriers inherit a mutant BRCA1 allele have a higher net rate to metastatic cancer simultaneously. This simultaneous of mutation in their somatic cells. Understanding inhibition would allow each infected cell to generate large the function of these two genes led researchers to numbers of infected cells through cellular replication. expect that women with the BRCA1 mutation would Once large numbers of precancerous cells are generated, be susceptible to PTEN mutations, which could then the standard arguments about cancer evolution would lead the mutated cells down the path to cancer. PTEN apply: additional oncogenic mutations that further mutations do tend to occur in the breast cancers. But inhibit these barriers or other barriers to cancer would in contrast with the expectations of researchers study- favor evolution of subsets of the precancerous cells ing this association, PTEN mutations occur late during toward cancer. Even if large numbers of precancerous oncogenesis, as expected if viral infection initiates cells in this population of precancerous cells lose their oncogenesis and mutations complete the process after viability because of damaging mutations, oncogenesis a sufficiently large population precancerous cells has can continue because many other infected cells remain become established. to generate additional oncogenic mutations. This example illustrates how hypotheses of infec- The immune suppression/infection hypothesis offers tious causation of cancer can be broadly distinguished resolutions of the paradoxes raised by the hormonal from hypotheses that rely solely on mutation. Infections proliferation hypothesis. The elevated risk of breast will be found early during oncogenesis, but oncogenic cancer during pregnancy may result from effects of the mutations will not be found in the earliest stages. elevated reproductive hormones on the part of the Because infection initiated cancers still depend on immune system that controls virally infected cells. This mutations for cancer progression evidence that broadly effect might occur synergistically with any hormonal implicates oncogenic mutation (e.g., age dependent enhancement of the proliferation of infected cells. The risks of cancer; Frank, 2007) are consistent with infec- associations with parity can be explained by effects of tion-initiated oncogenesis. But despite this consistency mate fidelity and duration of sexual activity on exposure the distinction between mutation-only hypotheses and to infection. Women who begin menses earlier will tend infection-initiated hypotheses is critical because the to have a longer exposure to sexually transmitted patho- infection-initiated hypotheses suggest that cancers gens such as HPV and kissing transmitted pathogens can be prevented by preventing infection. The great such as EBV. Having children may reduce exposure to practical benefit from cancer prevention relative to

510 Paul W. Ewald cancer treatment emphasizes the importance of testing seem to be a particularly difficult evolutionary hurdle. the central prediction from the infection-initiated Natural selection would favor regulation of the response hypothesis, namely that infectious agents tend to be so that women gain the protective benefits without present and oncogenic mutations absent during the risking death. Although it is appropriate to consider the earliest phases of oncogenesis. overzealous response hypothesis, it certainly should not be accepted without testing, especially when other more compelling hypothesis have not been considered or Pregnancy sickness tested. In some chronic diseases the failure to consider Consideration of infectious causation offers such infectious causation has led to study of hypotheses that an alternative hypothesis. Among infectious diseases may explain the baseline occurrence of a phenomona, one frequently finds manifestations that sometimes func- but not the most damaging occurrences. Pregnancy tion as a defense but can be disregulated by infectious sickness offers an example. Pregnancy sickness, also agents to such an extent that host death ensues. Diarrhea, called “morning sickness,” is a sensation of nausea that for example, is known to facilitate the recovery by is particularly common during the first trimester of expelling the causal pathogens – suppression of diarrhea pregnancy. Evolutionary considerations suggest that in patients experimentally infected with the diarrhea- debilitating nausea would be selected against if it did causing pathogen, Shigella sonnei, prolonged recovery not provide a compensating fitness benefit. from infection (Dupont and Hornick, 1973). When Hook (1976), Profet (1992), and Flaxman and Sher- humans are infected with Vibrio cholerae, however, the man (2000) developed the hypothesis that pregnancy diarrhea-induced dehydration is as a rule the cause of sickness is an adaptation that reduces ingestion of death. Unlike S. sonnei, V. cholerae does not invade the noxious agents (toxins and pathogens) and thereby intestinallining. Moreover, because V. cholerae is a strong protects the developing embryo and the expectant swimmer and can adhere to the lining of the small intes- mother. Such noxious agents tend to be dangerous or tine, it can persist in the intestine in the face of the rapid indicators of dangerous substances within food. Nausea fluid movements associated with diarrhea. In fact, in response to the smell or taste of secondary compounds V. cholerae induces the diarrhea by releasing a toxin may therefore cause the avoidance or regurgitation of that attaches to the cells that line the small intestine. toxin-laden and microbially contaminated foods. Fetal This attachment results the entrance of part of the toxin damage during the first trimester of pregnancy could molecule into the cell, which eventually leads to the have marked consequences because damage to small cell&s release of fluid into the lumen of the intestine. numbers of cells early during development may cause The resultant diarrhea washes out competing bacteria substantial damage to tissues or organs that develop from from the entire intestinal tract. Strains of V. cholerae that those cells, leading to birth defects or miscarriages. produce large amounts of toxin thus produce a watery Heightened nausea in response to secondary chemicals fecal material with large numbers of V. cholerae and few in food may protect developing embryo and fetus from other micro-organisms. This watery Vibrio-laden stool such damage. This hypothesis is supported by the timing can broadly contaminate the external environment, of symptoms during pregnancy and reduced risks of fostering transmission to other humans. For V. cholerae miscarriage among women experiencing pregnancy the generation of diarrhea is thus a sometimes lethal sickness (Profet, 1992; Flaxman and Sherman, 2000). manipulation (sensu Ewald, 1980) of what is often a This fetal protection hypothesis, however, raises a defensive response for its own benefit (summarized from paradox. Sometimes the vomiting associated with Ewald, 1994). pregnancy sickness is so severe that women will die of Consideration of infectious causation therefore dehydration. One would not expect natural selection to raises the possibility that a similar hypothesis might favor lethal nausea as a defense against damage that is resolve the paradox of lethal pregnancy sickness. much less costly to fitness. When such paradoxes are Specifically, it suggests the possibility that an infec- addressed in the literature of evolutionary medicine, they tious organism could manipulate the nausea defense, are often explained away by the weakness of natural exacerbating it to the point that it may cause life- selection. Specifically, it is argued that natural selection threatening dehydration. Helicobacter pylori is a is too weak to eliminate the damaging effects of an logical candidate for such an organism because it overzealous response, which might be maintained in the infects the lining of the stomach and regurgitation population as a result of the positive effects of a much would provide one of the most direct transmission more common, controlled response. This overzealous routes. Epidemiological evidence implicates this response hypothesis is not compelling because maternal route (Perry et al., 2006), as well as transmission from death seems like too great a price to reduce exposure of mother to offspring (Sinha et al., 2004; Yang et al., babies to the secondary compounds in food and because 2005, Delport and van der Merwe, 2007) and wife to control of nausea below life-threatening levels does not husband (Fujimoto et al., 2007).

Evolutionary Medicine and the Causes of Chronic Disease 511 TABLE 29.1. Illnesses of uncertain cause, their associations with infection, and publications within evolutionary medicine that did not consider infectious causation of the specified illness. Causal hypothesis presented in Infectious agent or Illness evolutionary medicine literature infectious disease Publications Skin cancer Ultraviolet radiation, light skin color HPV Greaves (2000*, 2008*) Breast cancer Increased menstrual cycling MMTV, EBV, HPV Eaton et al. (1994) Eaton and Eaton (1999*) Pre-eclampsia Parent–offspring genetic conflict Periodontitis, urinary Haig (1993, 2008*) tract infection Pregnancy sickness Protection from dietary mutagens Helicobacter pylori Profet (1992); Sherman and Flaxman (2002*) Schizophrenia Mutations and environmental stress Toxoplasma gondii Nesse and Stearns (2008*); Trevathan et al. (2008*) Atherosclerosis Rich diet and lipid transport of APOE4 Chlamydia pneumoniae Nesse and Williams (1995*) Gerber and Crews (1999*) Finch and Stanford (2004*) Alzheimer&s disease Immunological correlates and Chlamydophila Nesse and Williams (1995) (nonfamilial) compensating benefits of APOE4 (Chlamydia) Austad (1997), Finch (2007*) pneumoniae Austad and Finch (2008*) Note: *Signifies works that were published after the infectious link to the illness was published. EBV, Epstein–Barr virus; HPV, human papillomavirus; MMTV, mouse mammary tumor virus. Investigations of associations between H. pylori and but rather emphasizes conditions that have been severe pregnancy sickness have generated mixed results. considered often in the literature on evolutionary Abouthalfofthestudiesdocumentsignificantassociations medicine or pose major unsolved health problems. (Koc¸aketal.,1999;Reymundeetal.,2001;Kazeroonietal., The pathogens implicated are not yet broadly accepted 2002;Cevriogluetal.,2004;Shirinetal.,2004;Karaeretal., as causes of the listed conditions, even though strong 2008) and about half do not (Wu et al., 2000; Berker evidence of infectious causation exists. et al., 2003; Jacobson et al., 2003; McKenna et al., 2003; Discussionofschizophreniainevolutionarymedicine- Weyermannetal.,2003;Leeetal.,2005).Thisdiscrepancy literature (e.g., Nesse and Stearns, 2008), generally may result from differences in the strains in different has not incorporated the increasingly strong evidence populations of study subjects. In a study of 143 pregnant of infectious causation (Ledgerwoood et al., 2003; Turkish women (Noyan et al., 2004) the elevated positivity Brown et al., 2004; Crespi et al., 2007; Hinze-Selch for H. pylori among dyspepsic women (75% vs. 64%) was et al., 2007; Mortensen et al., 2007; Niebuhr et al., not statistically significant. But among those women 2008). Longstanding claims of genetic causation based who tested positive for H. pylori, those who complained on familial studies have been accepted even though of dyspepsia were significantly more often infected with these associations can be explained by in utero infec- strains possessing a major virulence determinant – the tious causation (Ledgerwoood et al., 2003). Specific cytotoxin-associated gene A (66% vs. 34%). genetic associations generally explain only vanishingly Taken together these results suggest that some small proportions of schizophrenia. If primary causes strains of H. pylori may facilitate their own transmis- of schizophrenia are largely infectious we can expect sion by exaggerating a nausea response that evolved to that the most common genetic predispositions to protect the fetuses of uninfected women. As is the case schizophrenia will tend to be vulnerabilities to such with V. cholerae, the conflict of interest between infectious causes (Ledgerwood et al., 2003). Accordingly H. pylori and its human host may sometimes lead to one of the strongest genetic associations identified to death through dehydration. date was recently found to encode an interleukin recep- tor, even though the researchers were not searching for a gene associated with the immune system and con- IMPLICATIONS FOR EVOLUTIONARY sidered this association surprising (Lencz et al., 2007). MEDICINE Pre-eclampsia is that occurs in about 3% of preg- nancies (Conde-Agudelo et al., 2008). In the evolution- Table 29.1 lists several other chronic illnesses of uncer- ary medical literature explanations of pre-eclampsia tain cause that have been discussed in the literature have been almost entirely restricted to genetic conflicts of evolutionary medicine. The list is not exhaustive, between mother and fetus (Haig, 1993, 2008). Yet a

512 Paul W. Ewald large literature on associations between infection and drawing evolutionary interpretations of the current pre-eclampsia has accumulated. Particularly reliable often incorrect consensuses? If evolutionary medicine are the associations between pre-eclampsia and follows the latter route, it is destined to present inter- both periodontal disease, which is caused largely by pretations that are faulty because they are based on Porphyromonas gingivalis, and urinary tract infection the faulty consensuses. (Conde-Agudelo and Belizan, 2000; Conde-Agudelo Taking the lead on these issues also offers the et al., 2008; Siqueira et al., 2008). promise of practical benefits. If infectious causation is Athoroughevolutionaryexaminationofpre-eclampsia more pervasive than is currently accepted by modern needs to resolve the maternal–fetal hypothesis with medicine and evolutionary biology expedites the recogni- these infectious associations. Analyses need to consider tion of this pervasiveness then it will also expedite the whether there is any evidence implicating infectious recognition of ways to prevent these diseases. In many causation that cannot be explained by maternal–fetus cases the practical benefits are obvious. Prevention of conflict and vice versa. Similarly, analyses need to chronic diseases such as cancers, heart attacks, and consider whether the two categories of causation are strokes, would provide vast health benefits because such both acting and perhaps interacting. Both categories chronic diseases are the major sources of mortality in of hypotheses need to be considered in light of other economically prosperous countries. It is reasonable to risk factors (Conde-Agudelo and Belizan, 2000) to expect that such diseases could be prevented, for address key characteristics of pre-eclampsia, such as example, by vaccination because the track record of the its relatively low prevalence, greater damage in poor health sciences at controlling or preventing infectious countries, and associations with parity. diseases has been very strong. The understanding of nonfamilial Alzheimer&s A broadened perspective within evolutionary medi- disease parallels that of atherosclerosis. Chlamydo- cine also provides more subtle insights into how dis- phila (Chlamydia) pneumoniae has been identified eases should be treated. Recognition that pregnancy in a large portion of Alzheimer&s disease patients sickness may often protect fetal development suggests (Balin et al., 1998). The association of E4 with Alzhei- that the symptoms of pregnancy sickness may often be mer&s disease and the vulnerability to C. pneumoniae best left untreated, but this suggestion does not apply to infection conferred by E4 indicates that C. pneumoniae symptoms that could cause life-threatening dehydra- is probably a cause of sporadic Alzheimer&s disease. Yet tion (Flaxman and Sherman, 2000). Without consider- evolutionary discussions of Alzheimer&s disease gener- ation of infectious causation, evolutionary medicine ally do not integrate the evidence implicating C. pneumo- may not be able to specify the boundary between appro- niae (Finch, 2007; Austad and Finch, 2008). priate and inappropriate treatment. If, however, Without consideration of infectious causation the H. pylori is responsible for the life-threatening symp- relative roles of infectious causation, inherited genes, toms of pregnancy sickness, this difficulty may be and noninfectious environmental risk factors cannot resolvable. The evolutionarily informed advice would be ascertained. Because the value of the evolutionary be for women to be tested for H. pylori and treated with explanations depends on the validity of the proximate antibiotics if the test comes back positive. Ideally this causes that they are attempting to explain, a failure to testing should be done prior to pregnancy to avoid any consider the full spectrum of the proximate causes raises negative effects of the antibiotics on fetal development. concern about the validity of the evolutionary explan- Once H. pylori is eliminated, symptoms of pregnancy ations. If such diseases are caused largely by infection, sickness should be ameliorated. evolutionary explanations that do not incorporate infec- Medicine has developed largely without a funda- tious causation will be largely incorrect. This problem mental theoretical framework. This situation has made is particularly important if infectious causation is intim- medicine vulnerable to the predispositions that are ately related to genetic causation and noninfectious present among the majority of experts at any given time. environmental causation, as appears to be the case with From the early nineteenth century onward these predis- breast cancer and atherosclerosis. If, for example, the positions have involved the unwarranted rejection of primary role of E4 in atherosclerosis is to increase the infectious diseases. Decade by decade new categories infection rate of C. pneumoniae then the attempts to of diseases have been gradually added to the spectrum explain the association between E4 allele and atheroscler- of infectious disease. Over the past half century this osis or Alzheimer&s disease strictly in terms of transport trend has continued with chronic diseases such as of lipids are futile. peptic ulcers, cervical cancer, and liver cancer being Much is at stake. Will evolutionary medicine lead recognized as infectious diseases. Infectious causation the health sciences to better understandings of med- of diseases such as breast cancer, atherosclerosis, ical issues and new discoveries? Or will it simply and schizophrenia are on the threshold of acceptance. follow circuitous routes dictated by the biases of Evolutionary medicine is uniquely positioned to nonevolutionary medicine at any particular time, advance these trends because, unlike other areas of

Evolutionary Medicine and the Causes of Chronic Disease 513 medicine, it is inclusive of all areas of medicine. It seeks valuable discussions on the topic. Holly A. Swain Ewald to integrate conceptually all proximate and ultimate first drew my attention to viral disruption of cellular explanations of health and disease. Molecular, bio- adhesion molecules, and has provided broadly insightful chemical, cellular, developmental, morphological, and suggestions throughout the development of this work. physiological processes provide the evidence on which this integration is based. But to accomplish this integra- tive goal, researchers in evolutionary medicine must REFERENCES broaden their perspectives not only to include these Almog, Y., Novack, V., Eisinger, M., et al. (2007). The effect areas but to recognize what is being overlooked and of statin therapy on infection-related mortality in patients unexplored. Evolutionary medicine has not yet taken with atherosclerotic diseases. Critical Care Medicine, 35, on this challenge. Instead it has been characterized by 1635–1636. narrowly circumscribed analyses that have failed to Andersson, K., Waterboer, T., Kirnbauer, R., et al. (2008). incorporate significant bodies of relevant evidence. In Seroreactivity to cutaneous human papillomaviruses among this chapter I have attempted to illustrate this problem patients with nonmelanoma skin cancer or benign skin and show how a more comprehensive and integrated lesions. Cancer Epidemiology, Biomarkers and Prevention, consideration of the feasible categories of causation 17, 189–195. helps to resolve paradoxes, provide a more cohesive Austad, S. N. (1997). Why We Age. What Science is Discovering understanding of chronic illnesses, and suggest critical About the Body&s Journey Through Life. New York: Wiley. Austad, S.N. and Finch, C.E. (2008). The evolutionary context areas for future study. of human aging and degenerative disease. In Evolution in Health and Disease,S.S.StearnsandJ.C.Koella(eds). New York: Oxford University Press, pp. 301–311. DISCUSSION POINTS Balin, B. J., Gerard, H. C., Arking, E. J., et al. (1998). Identi- fication and localization of Chlamydia pneumoniae in the 1. How does the triangle of disease causation provide Alzheimer&s brain. Medical Microbiology and Immun- a conceptual basis for an integrated understanding ology, 187, 23–42. the role of various risk factors in chronic disease? Berker, B., Soylemez, F., Cengiz, S.D., et al. (2003). Serologic 2. What is the distinction between primary and second- assay of Helicobacter pylori infection. Is it useful in hyper- ary causes of diseases and why is this distinction emesis gravidarum? Journal of Reproductive Medicine, 48, important to efforts to control disease? 809–812. 3. What is the main difficulty with interpreting onco- Bonnet, M., Guinebretiere, J. -M., Kremmer, E., et al. (1999). genesis solely as a mutation-driven process, and how Detection of Epstein–Barr virus in invasive breast cancers. does integration of infectious causation ameliorate Journal of the National Cancer Institute, 91, 1376–1381. this problem? Branca,M.,Giorgi,C.,Ciotti,M.,etal.(2006).Down-regulated nucleoside diphosphate (NDP) kinase nm23-H1 expression 4. How does the action of natural selection on infec- is unrelated to high-risk human papillomavirus (HPV) tious agents favor adaptations that nudge infected but associated with progression of cin and unfavourable cells toward cancer? prognosis in cervical cancer. JournalofClinicalPathology, 5. Why do evolutionary considerations and mechanism 59, 1044–1051. of cellular invasion by Chlamydophila pneumoniae Brown, A.S., Begg, M.D., Gravenstein, S., et al. (2004). invoke infectious causation as cause of the epsilon Serologic evidence of prenatal influenza in the etiology of 4 diseases: atherosclerosis, sporadic Alzheimer&s schizophrenia. Archives of General Psychiatry, 61, 774–780. disease, and multiple sclerosis? Bruzzi, P., Negri, E., La Vecchia, C., et al. (1988). Short term 6. What evolutionary paradox arises from interpreting increase in risk of breast cancer after full term pregnancy. pregnancy sickness solely as an adaptation to British Medical Journal, 297, 1096–1098. reduce exposure of the fetus to damaging chem- Cevrioglu, A. S., Altindis, M., Yilmazer, M., et al. (2004). Effi- icals, and how does consideration of Helicobacter cient and non-invasive method for investigating Helicobacter pylori resolve this paradox? pylori in gravida with hyperemesis gravidarum: Helicobacter pyloristoolantigentest.JournalofObstetricsandGynaecology Research, 30, 136–141. Cochran, G. M., Ewald, P. W. and Cochran, K. D. (2000). Infec- ACKNOWLEDGEMENTS tious causation of disease: an evolutionary perspective. Per- spectives in Biology and Medicine, 43,406–448. The Rena Shulsky Foundation has generously supported Conde-Agudelo, A. and Belizan, J.M. (2000). Risk factors for the work on cancer as part of a project to develop a preeclampsia in a large cohort of Latin American and Carib- unified theory of oncogenesis. Discussions with Gregory bean women. British Journal of Obstetrics and Gynaecology, Cochran over the past decade have profoundly influ- 107, 75–83. enced my thinking about the causes of chronic diseases. Conde-Agudelo, A., Villar, J. and Lindheimer, M. (2008). Carolyn Doyle provided references on breast cancer and Maternal infection and risk of preeclampsia: systematic

514 Paul W. Ewald review and metaanalysis. American Journal of Obstetrics Ewald, P. W. (2000). Plague Time. How Stealth Infections and Gynecology, 198, 7–22. Cause Cancer, Heart Disease, and Other Deadly Ailments. Corbo, R. M. and Scacchi, R. (1999). Apolipoprotein E (APOE) New York: Free Press. allele distribution in the world. Is APOE4 a “thrifty” allele? Ewald, P. W. (2008). An evolutionary perspective on the Annals of Human Genetics, 63, 301–310. causes of chronic diseases: atherosclerosis as an illustra- Crespi, B. J. and Summers, K. (2005). Evolutionary biology tion. In Evolutionary Medicine and Health. New Perspec- of cancer. Trends in Ecology and Evolution, 20, 545–552. tives, W. R. Trevathan, E. O. Smith and J. J. Mckenna Crespi, B. J. and Summers, K. (2006). Positive selection in the (eds). New York: Oxford University Press, pp. 350–367. evolution of cancer. Biological Reviews of the Cambridge Ewald, P.W. (2009). An evolutionary perspective on parasitism Philosophical Society, 81, 407–424. as a cause of cancer. Advances in Parasitology, 61, 28–43. Crespi, B. J., Summers, K. and Dorus, S. (2007). Adaptive Ewald, P. W. and Cochran, G. M. (2000). Chlamydia pneumo- evolution of genes underlying schizophrenia. Proceedings niae and cardiovascular disease: an evolutionary perspec- of the Royal Society of London. Series B, 274, 2801–2810. tive on infectious causation and antibiotic treatment. Crowther, M. A. (2005). Pathogenesis of atherosclerosis. Journal of Infectious Diseases, 181, S394–S401. Hematology. Education Program of the American Society Fina, F., Romain, S., Ouafik, L. H., et al. (2001). Frequency of Hematology, 1, 436–441. and genome load of Epstein–Barr virus in 509 breast Damin, A., Karam, R., Zettler, C., et al. (2004). Evidence for an cancers from different geographical areas. British Journal association of human papillomavirus and breast carcin- of Cancer, 84, 783–790. omas. Breast Cancer Research and Treatment, 84, 131–137. Finch, C. E. (2007). The Biology of Human Longevity: Inflam- De Villiers, E. -M., Sandstrom, R., Zur Hausen, H., et al. mation, Nutrition, and Aging in the Evolution of Lifespans. (2005). Presence of papillomavirus sequences in condylo- Boston: Academic. matous lesions of the mamillae and in invasive carcinoma Finch, C. E. and Stanford, C. B. (2004). Meat-adaptive genes of the breast. Breast Cancer Research, 7, R1–R11. and the evolution of slower aging in humans. Quarterly Delport, W. and Van Der Merwe, S. W. (2007). The transmis- Review of Biology, 79, 3–50. sion of Helicobacter pylori: the effects of analysis method Flaxman, S. M. and Sherman, P. W. (2000). Morning sickness: and study population on inference. Best Practice and a mechanism for protecting mother and embryo. Quarterly Research: Clinical Gastroenterology, 21, 215–236. Review of Biology, 75, 113–148. Diamond, J. (2005). Evolutionary biology: Geography and Frank, S. A. (2007). Dynamics of Cancer. Incidence, Inherit- skin colour. Nature, 435, 283–284. ance, and Evolution. Princeton: Princeton University Press. Ding, L., Li, L., Yang, J., et al. (2007). Latent membrane protein Fujimoto, Y., Furusyo, N., Toyoda, K., et al. (2007). Intrafa- 1 encoded by Epstein–Barr virus induces telomerase activity milial transmission of Helicobacter pylori among the viap16INK4A/Rb/E2F1andJNKsignalingpathways.Journal population of endemic areas in Japan. Helicobacter, 12, of Medical Virology, 79, 1153–1163. 170–176. Doyle, C., Swain Ewald, H. A. and Ewald, P. W. (2007). Pre- Fullerton, S. M., Clark, A. G., Weiss, K. M., et al. (2000). menstrual syndrome: an evolutionary perspective on its Apolipoprotein E variation at the sequence haplotype causes and treatment. Perspectives in Biology and Medicine, level: implications for the origin and maintenance of a 50, 181–202. major human polymorphism. American Journal of Human Doyle, C., Swain Ewald, H.A. and Ewald, P.W. (2008). An Genetics, 67, 881–900. evolutionary perspective on premenstrual syndrome and its Gardner, C.D., Lawson, L.D., Block, E., et al. (2007). Effect of implications for investigating infectious causes of chronic raw garlic vs. commercial garlic supplements on plasma disease. In Evolutionary Medicine and Health. New Perspec- lipid concentrations in adults with moderate hypercholes- tives, W.R. Trevathan, E. O. Smith and J.J. Mckenna (eds). terolemia: a randomized clinical trial. Archives of Internal New York: Oxford University Press, pp. 196–215. Medicine, 167, 346–353. Dupont, H. L. and Hornick, R. B. (1973). Adverse effect of Ge ´rard, H. C., Wang, G. F., Balin, B. J., et al. (1999). Fre- Lomotil therapy in shigellosis. Journal of the American quency of apolipoprotein E (APOE) allele types in patients Medical Association, 226, 1525–1528. with Chlamydia-associated arthritis and other arthritides. Eaton, S. B. and Eaton, S. B. I I I (1999). Breast cancer in Microbial Pathogenesis, 26, 35–43. evolutionary context. In Evolutionary Medicine, W. R. Tre- Ge ´rard, H. C., Fomicheva, E., Whittum-Hudson, J.A., et al. vathan, E. O. Smith and J. J. Mckenna (eds). New York: (2008). Apolipoprotein E4 enhances attachment of Chlamy- Oxford University Press, pp. 429–442. dophila (Chlamydia) pneumoniae elementary bodies to host Eaton, S. B., Shostak, M. and Konner, M. (1988). The Paleo- cells. Microbial Pathogenesis, 44, 279–285. lithic Prescription. New York: Harper and Row. Gerber, L. M. and Crews, D. E. (1999). Evolutionary perspec- Eaton, S. B., Pike, M. C., Short, R. V., et al. (1994). Women&s tives on chronic degenerative diseases. In Evolutionary reproductive cancers in evolutionary context. Quarterly Medicine, W. R. Trevathan, E. O. Smith, J. J. McKenna Review of Biology, 69, 353–367. (eds). New York: Oxford University Press, pp. 443–469. Ewald, P. W. (1980). Evolutionary biology and treatment of Greaves, M. (2000). Cancer. The Evolutionary Legacy.Oxford: signs and symptoms of infectious disease. Journal of The- Oxford University Press. oretical Biology, 86, 169–176. Greaves, M. (2006). Infection, immune responses and the Ewald, P. W. (1994). Evolution of Infectious Disease.New aetiology of childhood leukaemia. Nature Reviews Cancer, York: Oxford University Press. 3, 193–203.

Evolutionary Medicine and the Causes of Chronic Disease 515 Greaves, M. (2008). Cancer: evolutionary origins of vulner- Kazerooni, T., Taallom, M. and Ghaderi, A. A. (2002). Heli- ability. In Evolution in Health and Disease, S. S. Stearns cobacter pylori seropositivity in patients with hyperemesis and J. C. Koella (eds). New York: Oxford University Press, gravidarum. International Journal of Gynaecology and pp. 277–287. Obstetrics, 79, 217–220. Guasparri, I., Bubman, D. and Cesarman, E. (2008). EBV Key, T. J. A. and Pike, M. C. (1988). The role of oestrogens LMP2Aaffects LMP1-mediatedNF-kBsignalingand survival and progestagens in the epidemiology and prevention of of lymphoma cells by regulating TRAF2 expression. Blood, breast cancer. European Journal of Cancer and Clinical 111, 3813–3820. Oncology, 24, 29–43. Haig, D. (1993). Genetic conflicts in human pregnancy. Kleer, C., Tseng, M., Gutsch, D., et al. (2002). Detection of Quarterly Review of Biology, 68, 495–532. Epstein–Barr virus in rapidly growing fibroadenomas of Haig, D. (2008). Intimate relations: evolutionary conflicts of the breast in immunosuppressed hosts. Modern Pathology, pregnancy and childhood. In Evolution in Health and Dis- 15, 759–764. ease, S. S. Stearns and J. C. Koella (eds). New York: Oxford Kleinsmith, L. J. (2005). Principles of Cancer Biology.San University Press, pp. 65–76. Francisco: Pearson/Benjamin Cummings. Hino, R., Uozaki, H., Inoue, Y., et al. (2008). Survival advantage Knight, J., Sharma, N. and Robertson, E. (2005). Epstein–Barr ofEBV-associatedgastriccarcinoma:survivingup-regulation virus latent antigen 3C can mediate the degradation of the by viral latent membrane protein 2A. Cancer Research, 68, retinoblastoma protein through an SCF cellular ubiqitin 1427–1435. ligase. Proceedings of the National Academy of Sciences of Hinze-Selch, D., Da ¨ubener, W., Eggert, L., et al. (2007). the United States of America, 102, 18562–18566. A controlled prospective study of Toxoplasma gondii infec- Koc¸ak, I., Akcan, Y., Ustu ¨n, C., et al. (1999). Helicobacter pylori tion in individuals with schizophrenia: beyond seropreva- seropositivity in patients with hyperemesis gravidarum. lence. Schizophrenia Bulletin, 33, 782–788. International Journal of Gynaecology and Obstetrics, 66, Hook, E.B. (1976). Changes in tobacco smoking and ingestion 251–254. of alcohol and caffienated beverages during early pregnancy: Komarova, N. L. and Wodarz, D. (2008). Cancer as a micro- are these consequences in part of feto-protective mechan- evolutionary process. In Evolution in Health and Disease, isms diminishing maternal exposure to embryo toxins. S. S. Stearns and J. C. Koella (eds). New York: Oxford In Birth Defects, Risks and Consequences,S.Kelly(ed.). University Press, pp. 289–299. New York: Academic, pp. 173–183. Lambe, M., Hsieh, C., Trichopoulos, D., et al. (1994). Transient Ilveskoski, E., Perola, M., Lehtimaki, T., et al. (1999). Age- increase in risk of breast cancer after giving birth. New dependent association of apolipoprotein E genotype with England Journal of Medicine, 331,5–9. coronary and aortic atherosclerosis in middle-aged men. Lawson, J. S., Tran, D. and Rawlinson, W. D. (2001). From An autopsy study. Circulation, 100, 608–613. Bittner to Barr: a viral, diet and hormone breast cancer Jacobson, G. F., Autry, A. M., Somer-Shely, T. L., et al. (2003). aetiology hypothesis. Breast Cancer Research, 3, 81–85. Helicobacter pylori seropositivity and hyperemesis gravi- Ledgerwood, L. G., Ewald, P. W. and Cochran, G. M. (2003). darum. Journal of Reproductive Medicine, 48, 578–582. Genes, germs, and schizophrenia: an evolutionary perspec- Jenson, A. B., Geyer, S., Sundberg, J. P., et al. (2001). Human tive. Perspectives in Biology and Medicine, 46, 317–348. papillomavirus and skin cancer. Journal of Investigative Lee, R. H., Pan, V. L. and Wing, D. A. (2005). The prevalence Dermatology Symposium Proceedings, 6, 203–206. of Helicobacter pylori in the Hispanic population affected Jerwood, S. and Cohen, J. (2008). Unexpected antimicrobial by hyperemesis gravidarum. American Journal of Obstet- effect of statins. Journal of Antimicrobial Chemotherapy, rics Gynecology, 193, 1024–1027. 61, 362–364. Lencz, T., Morgan, T. V., Athanasiou, M., et al. (2007). Con- Kan, C. -Y., Iacopetta, B., Lawson, J., et al. (2005). Identifica- verging evidence for a pseudoautosomal cytokine receptor tion of human papillomavirus DNA gene sequences in gene locus in schizophrenia. Molecular Psychiatry, 12, human breast cancer. British Journal of Cancer, 93, 946–948. 572–580. Karaer, A., Ozkan, O., Ozer, S., et al. (2008). Gastrointestinal Leonard, W. R. (2008). Lifestyle, diet, and disease: compara- symptoms and Helicobacter pylori infection in early preg- tive perspectives on the determinants of chronic health nancy. A seroepidemiologic study. Gynecologic and Obstetric risks. In Evolution in Health and Disease, S. S. Stearns Investigation, 66,44–46. and J. C. Koella (eds). New York: Oxford University Press, Karagas, M. R., Nelson, H. H., Sehr, P., et al. (2006). Human pp. 265–276. papillomavirus infection and incidence of squamous cell Liu, W. K., Chu, Y. L., Zhang, F., et al. (2005). The relation- and basal cell carcinomas of the skin. Journal of the ship between HPV16 and expression of CD44v6, nm23H1 National Cancer Institute, 98, 1425–1426. in esophageal squamous cell carcinoma. Archives of Karagas, M. R., Nelson, H. H., Zens, M. S., et al. (2007). Virology, 150, 991–1001. Squamous cell and basal cell carcinoma of the skin in Liu, X., Roberts, J., Dakic, A., et al. (2008). HPV E7 contributes relation to radiation therapy and potential modification to the telomerase activity of immortalized and tumorigenic of risk by sun exposure. Epidemiology, 18, 776–784. cells and augments E6-induced hTERT promoter function. Katz, E., Lareef, M. H., Rassa, J. C., et al. (2005). MMTV Virology, 375, 611 –623. Env encodes an ITAM responsible for transformation of Mahley, R. W. and Huang, Y. D. (1999). Apolipoprotein E: mammary epithelial cells in three-dimensional culture. from atherosclerosis to Alzheimer&s disease and beyond. Journal of Experimental Medicine, 201, 431–439. Current Opinion in Lipidology, 10, 207–217.

516 Paul W. Ewald Maula, S., Huuhtanen, R. L., Blomqvist, C. P., et al. (2001). Mind, J. E. Barkow, L. Cosmides, et al. (eds). New York: The adhesion molecule CD44v6 is associated with a high Oxford University Press, pp. 327–365. risk for local recurrence in adult soft tissue sarcomas. Reeth, K. V., Brown, I., Essen, S., et al. (2004). Genetic British Journal of Cancer, 84, 244–252. relationships, serological cross-reaction and cross- Mays, M. E. and Dujovne, C. A. (2008). Pleiotropic effects: protection between H1N2 and other influenza A virus should statins be considered an essential component in subtypes endemic in European pigs. Virus Research, 103, the treatment of dyslipidemia? Current Atherosclerosis 115–124. Reports, 10, 45–52. Reymunde, A., Santiago, N. and Perez, L. (2001). Helicobac- McKenna, D., Watson, P. and Dornan, J. (2003). Helicobacter ter pylori and severe morning sickness. American Journal pylori infection and dyspepsia in pregnancy. Obstetrics of Gastroenterology, 96, 2279–2280. and Gynecology, 102, 845–849. Russo, J. (1992). Influence of age and parity on the develop- Merlo, L. M., Pepper, J. W., Reid, B. J., et al. (2006). Cancer ment of the human breast. Breast Cancer Research and as an evolutionary and ecological process. Nature Reviews Treatment, 23, 211–218. Cancer, 6, 924–935. Saal, L. H., Gruvberger-Saal, S. K., Persson, C., et al. (2008). Mileo, A. M., Piombino, E., Severino, A., et al. (2006). Multiple Recurrent gross mutations of the PTEN tumor suppressor interference of the human papillomavirus-16 E7 oncopro- gene in breast cancers with deficient DSB repair. Nature tein with the functional role of the metastasis suppressor Genetics, 40, 102–107. Nm23-H1 protein. Journal of Bioenergetics and Biomem- Sgouros, J., Galani, E., Gonos, E., et al. (2007). Correlation branes, 38, 215–225. of nm23-H1 gene expression with clinical outcome Mortensen, P. B., Nrgaard-Pedersen, B., Waltoft, B. L., et al. in patients with advanced breast cancer. In Vivo, 21, (2007). Early infections of Toxoplasma gondii and the later 519–522. development of schizophrenia. Schizophrenia Bulletin, Sherman, P. W. and Flaxman, S. M. (2002). Nausea and 33, 741–744. vomiting of pregnancy in an evolutionary perspective. Murakami,M.,Lan,K.,Subramanian,C.,etal.(2005).Epstein– American Journal of Obstetrics and Gynecology, 186, Barr virus nuclear antigen 1 interacts with Nm23-H1 in S190–S197. lymphoblastoid cell lines and inhibits its ability to suppress Shirin, H., Sadan, O., Shevah, O. et al. (2004). Positive ser- cell migration. Journal of Virology, 79, 1559–1568. ology for Helicobacter pylori and vomiting in the pregnancy. Nesse, R.M. (2008). The importance of evolution for medicine. Archives of Gynecology and Obstetrics, 270, 10–14. In Evolutionary Medicine and Health. New Perspectives, Sinha, S. K., Martin, B., Gold, B. D., et al. (2004). The inci- W. R. Trevathan, E. O. Smith and J.J. Mckenna (eds). New dence of Helicobacter pylori acquisition in children of a York: Oxford University Press, pp. 416–433. Canadian First Nations community and the potential for Nesse, R.M. and Stearns, S.C. (2008). The great opportunity: parent-to-child transmission. Helicobacter, 9, 59–68. evolutionary applications to medicine and public health. Siqueira, F. M., Cota, L. O., Costa, J. E., et al. (2008). Mater- Evolutionary Applications, 1, 28–48. nal periodontitis as a potential risk variable for pree- Nesse, R. M. and Williams, G. C. (1994). Why We Get Sick. clampsia: a case-control study. Journal of Periodontology, The New Science of Darwinian Medicine. New York: Times 79, 207–215. Books. Stearns, S. S. and Koella, J. C. (eds) (2008). Evolution in Niebuhr,D.W.,Millikan,A.M.,Cowan,D.N.,etal.(2008). Health and Disease, 2nd edn. New York: Oxford University Selected infectious agents and risk of schizophrenia among Press. US military personnel. American Journal of Psychiatry, 165, Stearns, S. C., Nesse, R. M. and Haig, D. (2008). Introducing 99–106. evolutionary thinking for medicine. In Evolution in Health Noyan,V.,Apan,T.Z.,Yucel,A.,etal.(2004).Cytotoxinasso- and Disease, S. S. Stearns and J. C. Koella (eds). New York: ciated gene A-positive Helicobacter pylori strains in Oxford University Press, pp. 1–15. dyspeptic pregnant women. European Journal of Obstetrics, Subramanian, C. and Robertson, E. S. (2002). The meta- Gynecology and Reproductive Biology, 116, 186–189. static suppressor Nm23-H1 interacts with EBNA3C at Ouatas, T., Clare, S.E., Hartsough, M.T., et al. (2002). MMTV- sequences located between the glutamine- and proline- associated transcription factor binding sites increase rich domains and can cooperate in activation of transcrip- nm23-H1metastasis suppressor gene expression in human tion. Journal of Virology, 76, 8702–8709. breast carcinoma cell lines. Clinical and Experimental Taubes, G. (2008). What&s cholesterol got to do with it? Metastasis, 19, 35–42. New York Times. January 27, p. 18. Perry, S., De La Luz Sanchez, M., Yang, S., et al. (2006). Trevathan, W. R., Smith, E. O. and Mckenna, J. J. (2008). Gastroenteritis and transmission of Helicobacter pylori Background. In Evolutionary Medicine and Health. infection in households. Emerging Infectious Diseases, New Perspectives, W. R. Trevathan, E. O. Smith and 12, 1701–1708. J. J. Mckenna (eds). New York: Oxford University Press, Potten, C. S., Watson, R. J., Williams, G. T., et al. (1988). The pp. 1–54. effect of age and menstrual cycle upon proliferative activ- Trichopoulos, D., Li, F. and Hunter, D. (1996). What causes ity in the normal human breast. British Journal of Cancer, cancer? Scientific American, 275, 80–87. 58, 163–170. Tungteakkhun, S. S. and Duerksen-Hughes, P. J. (2008). Profet, M. (1992). Pregnancy sickness as adaptation: a deter- Cellular binding partners of the human papillomavirus rent to maternal ingestion of teratogens. In The Adapted E6 protein. Archives of Virology, 153, 397–408.

Evolutionary Medicine and the Causes of Chronic Disease 517 Wang, Y., Holland, J., Bleiweiss, I., et al. (1995). Detection of Williams, G. C. and Nesse, R. M. (1991). The dawn of mammary tumor virus ENV gene-like sequences in human Darwinian medicine. Quarterly Review of Biology, 66,1–22. breast cancer. Cancer Research, 55, 5173–5179. Wu, C. Y., Tseng, J. J., Chou, M.M., et al. (2000). Correlation Weyermann, M., Brenner, H., Adler, G., et al. (2003). Helico- between Helicobacter pylori infection and gastrointestinal bacter pylori infection and the occurrence and severity of symptoms in pregnancy. Advances in Therapy, 17, 152–158. gastrointestinal symptoms during pregnancy. American Yang, Y. J., Sheu, B. S., Lee, S. C., et al. (2005). Children Journal of Obstetrics and Gynecology, 189, 526–531. of Helicobacter pylori-infected dyspeptic mothers are Williams, E., Jones, L., Vessey, M., et al. (1990). Short term predisposed to H. pylori acquisition with subsequent increase in risk of breast cancer associated with full term iron deficiency and growth retardation. Helicobacter, pregnancy. British Medical Journal, 300, 578–579. 10, 249–255.

30 Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome Christopher W. Kuzawa EXPLAINING THE MODERN METABOLIC 2008). The idea that obesity and related diseases result DISEASE EPIDEMIC: THE THRIFTY GENOTYPE from a “discordance” or “mismatch” between our HYPOTHESIS AND ITS LIMITATIONS ancient genes and our rapidly changing lifestyle and diet is now widely accepted, and it seems clear that Today, more than 1 billion people are overweight or obesity must be more common today in large part obese, and the related condition of cardiovascular dis- because we are eating more and expending less than ease (CVD) accounts for more deaths than any other our ancestors traditionally did (Eaton and Konner, cause (Mackay et al., 2004). Why this epidemic of 1985; see Chapter 28 of this volume). Despite the intui- metabolic disease has emerged so rapidly in recent tive appeal of these ideas, the hypothesis is not without history is a classic problem for anthropologists con- limitation. For one, it helps explain why we gain weight cerned with the role of culture change in disease in a modern environment of nutritional abundance, transition (Ulijaszek and Lofink, 2006). In 1962, the but says very little about the syndrome of metabolic geneticist James Neel proposed an explanation for this changes that account for the diseases that accompany phenomenon that looked for clues in the “feast– obesity (Vague, 1955; Kissebah et al., 1982). While famine” conditions that he believed our nomadic, for- excess weight gain in general is unhealthy, it is specif- aging ancestors faced in the past. Given the unpredict- ically fat deposition in the visceral or abdominal region ability of food resources in natural ecologies, Neel that accounts for the bulk of its adverse effects on suggested that a “thrifty” metabolism capable of effi- health. Visceral fat has unique metabolic properties ciently storing excess dietary energy as body fat when that contribute to prediabetes states like insulin resist- food was abundant would have provided a survival ance when deposition in this depot is excessive advantage during later periods of shortage. In the wake (Reaven, 1988; Wellen and Hotamisligil, 2003). The of the rapid dietary and lifestyle change in recent gen- simple idea that our bodies are famine adapted may erations, and the comparatively slow pace of genetic help explain why we are prone to gaining weight when change, these foraging-adapted genes would now be we eat too much, but it says little about the metabolic “rendered detrimental by progress” (Neel, 1962), symptoms that make weight gain unhealthy. leading to obesity and diabetes. While a useful way to The sole focus of the hypothesis on genes is also think about obesity generally, variations on this idea outdated in light of newer evidence that biological sus- have been proposed to help explain the high rates of ceptibility of developing these adult diseases is also metabolic disease among populations believed to have elevated among individuals who experienced poor experienced especially severe nutritional conditions in nutrition during early life (Barker et al., 1989; Gluck- the past, including the diabetes-prone Pima Indians of man and Hanson, 2006). A large research literature has the American Southwest (Knowler et al., 1982), and demonstrated that the body’s responses to early life several groups of South Pacific Islanders who have nutrition subsequently influences that individual’s risk among the highest rates of obesity in the world (Dowse for developing adult diseases like diabetes and CVD, and Zimmet, 1993; McGarvey, 1994). a process described as nutritional “programming” The thrifty gene hypothesis was one of the first uses (Lucas, 1991) or “induction” (Bateson, 2001). For of evolutionary reasoning to shed light on a human instance, CVD and the rate of CVD mortality in adult- disease, and is thus a classic example of evolutionary hood are inversely related to size at birth – a measure or “Darwinian” medicine (see Nesse and Williams, of fetal nutrition – in places like the UK, Sweden, 1994; Stearns and Koella, 2008; Trevathan et al., India, and the Philippines (Kuzawa and Adair, 2003; Human Evolutionary Biology, ed. Michael P. Muehlenbein. Published by Cambridge University Press. # Cambridge University Press 2010. 518

Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome 519 16 14 12 10 % body fat 8 6 4 2 0 Human Harp seal Fur seal Reindeer Baboon Lamb Calf Black bear Elephant seal Rabbit Cat Caribou Pig Rat Hamster Sea Lions Mouse Guinea pig Foal 30.1. Percentage of body fat at birth in mammals. Adapted from Kuzawa (1998). Yajnik, 2004; Lawlor et al., 2005). Individuals who 30 were born small also experience durable changes in biological systems that contribute to CVD, as reflected 25 Male in their higher risk of depositing fat in the visceral Female region, or of developing high blood pressure, impaired 20 glucose tolerance or diabetes, or high cholesterol (Adair et al., 2001; Kuzawa and Adair, 2003). When the dietary % body fat 15 intake of pregnant rats and sheep is restricted, their 10 offspring have much the same set of outcomes as we see in humans born small (McMillen and Robinson, 5 2005; Fernandez-Twinn and Ozanne, 2006). This finding suggests that the body has an ability to induce a “thrifty 0 0 12243648607284 phenotype” that is better suited to survival in nutrition- Age (months) ally challenging environments. It also illustrates why a 30.2. Age changes in percentage body fat in humans. Data purely gene-based model of metabolic disease evolution from Davies and Preece (1989), pp. 95–97. is incomplete (Hales and Barker, 1992). A third limitation of the thrifty genotype hypothesis is that it merely assumes that famine was the key source of selection on human metabolism without con- mammal studied thus far (Kuzawa, 1998). This is sidering the actual causes of that stress and the ages at followed by a continued period of rapid fat deposition which it is most severe (Kuzawa, 1997). There are during the early postnatal months. In well-nourished reasons to question whether our distant foraging populations, adiposity reaches peak levels during the ancestors experienced a major burden of famine, first year of life before gradually declining to a nadir in which archaeological and contemporary ethnographic childhood, when humans reach their lowest level of data suggest only emerged as an important problem body fat in the lifecycle (Figure 30.2). If the threat of after the evolutionarily recent development of agri- famine is what drove the human tendency to build up cultural subsistence (Cohen and Armelagos, 1984; fat reserves, it is not obvious why children’s bodies Benyshek and Watson, 2006). Given this, famine may should do so little to prepare for these difficult periods. have been a relatively weak force of selection on the The lower priority placed upon maintaining an energy human gene pool (Speakman, 2006). reserve by middle childhood suggests that the back- Additional support for this interpretation comes ground risk of starvation faced by our ancestors – from the growth pattern of body fat in humans, shown “famine” – was small in comparison to the nutritional in Figure 30.1. In humans, body fat makes up a larger stress experienced during the preceding developmental percentage of weight at birth than in any other period. A prereproductive life history stage that does

520 Christopher W. Kuzawa not prioritize energy storage is difficult to reconcile 100 with the assumption that famine was the major nutri- tional challenge faced by human populations. 80 In this chapter, I review the causes, consequences, and adaptive solutions to the nutritional stress of 60 infancy and childhood in hopes of shedding light on % RMR to brain the evolution of human metabolism. Viewing human 40 nutritional stress through a developmental lens helps identify ages when human metabolism has likely been 20 under strongest selection. I will show how two traits that are central to the adverse health consequences 0 0 5 10 15 20 of obesity in overnourished adults – visceral fat and Age (years) insulin resistance – likely evolved as components of 30.3. Percentage of resting metabolic rate (RMR) devoted to an energy backup system that reprioritizes the alloca- the brain by age in humans. Data adapted from Holliday (1986). tion of prized energy and glucose during a crisis, thus allowing the body to shunt resources from nonessen- tial functions to critical organs like the brain (for dis- which requires a constant redistribution of ions across cussion of the function of insulin resistance during the cell membrane using energy-dependent ion pumps. pregnancy, see Haig, 1993). Demands placed on this Humans are exceptional in the quantity of this costly system are greater during infancy and early childhood tissue that they must sustain, and it has been estimated than at any other age of the lifecycle, suggesting that that greater than 80% of the body’s metabolism is some of the evolutionary seeds of adult metabolic dis- devoted to the brain in the newborn (Figure 30.3). ease may trace to selection pressures operative during It is a notable feature of human metabolism that early life. In addition, evidence for developmental total metabolic rate measured in adulthood (for plasticity in this backup system and the adult diseases instance, calories expended per day) is not increased that it contributes to suggests that its priorities may be above other mammals of similar body size despite adjusted in response to nutritional and other stressors our highly encephalized state (Armstrong, 1983). experienced early in the life cycle. A developmental Thus, brain expansion must have been matched by a approach to the evolution of human metabolic disease reduction in other energetically expensive tissues underscores the importance of the body’s internal (Armstrong, 1983). Human evolution was marked by strategies of allocating finite resources – in addition a movement to more energy-dense and easy to digest to the external stress of famine – as key to understand- foods (Leonard and Roberton, 1992, 1994), which may ing the contemporary epidemic of metabolic disease. have allowed a reduction in the size and energy requirements of the metabolically costly gut (Aiello and Wheeler, 1995). This reduction in gut expenditure A DEVELOPMENTAL PERSPECTIVE ON may have helped offset our increased brain needs in ENERGY STRESS IN HUMAN EVOLUTION: adult life (Aiello and Wheeler, 1995), but this seems THE DEVELOPMENTAL BOTTLENECK less likely during infancy and childhood. The size of the brain relative to the body is far larger in the human Humans have been described as naturally obese (Pond, neonate and infant than in the adult, and as a result, 1997), a characterization which is particularly fitting at the body’s total energy requirements are likely elevated birth as human newborns are born with more body fat at this age relative to other similarly sized mammalian than any other species (Kuzawa, 1998). Attempts to and primate newborns (Foley and Lee, 1991). More- explain our unusual “baby fat” have traditionally over, unlike energy expended on other tissues or looked to our hairlessness for clues, and it is widely systems, brain metabolism may not be reduced to con- assumed that natural selection compensated for our serve energy during a period of shortage, but instead loss of fur with a layer of insulative body fat (e.g., must be maintained within narrow limits to avoid per- Hardy, 1960; Morris, 1967). A competing perspective manent damage (Owen et al., 1967). Thus, our large notes that this excess adipose tissue is well-suited to brains impose a double burden on metabolism during serve as a backup energy supply for another distinctive infancy: they increase demand for energy while human trait: our large brains (Kuzawa, 1998). Brains restricting flexibility in metabolic expenditure when have among the highest metabolic rates of any tissue or nutritional supply is disrupted. organ in the body, and they are quickly damaged in the Other factors that are commonly experienced event of even temporary disruption in energy supply. during infancy can impede the supply of nutrients, Neuronal tissues are costly because they must be main- ensuring that negative energy balance is a regular tained in a state far from thermodynamic equilibrium, occurrence at this age in most populations. We are

Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome 521 born with a naı ¨ve adaptive immune system, and must infections and periods of negative energy balance therefore come into contact with (and be infected by) decline to a small fraction of their high prevalence specific pathogens to acquire the repertoire of anti- during the postweaning period. Because older children bodies necessary to protect us from future infection. are far less likely to have to rely upon energy reserves Initially, newborns enjoy immune protection from for survival, it is easy to see why the human body several maternal sources. The first is in the form of places lower priority on maintaining sizeable body fat maternal immunoglobulin G (IgG) antibodies acquired stores by this age. across the placenta. In addition, exclusively breast-fed Thus, the unique energetic stressors of early life infants are shielded from exposure to pathogens, and have left an imprint on the developmental pattern of also enjoy passive immune protection from maternal fat deposition in humans, represented by intensive secretory antibodies (sIgA) in breast milk. As a result, investment in the tissue in the run-up to the stresses newborns are often quite healthy in the early postnatal of weaning, and a gradual decline in the priority given months. However, both sources of passive immune to maintaining this reserve at older and more resilient protection eventually wane. As energy requirements ages. But this merely shows that the body is prioritiz- outstrip the supply capacity of breast milk by roughly ing the allocation of energy to an energetic buffer. This 6 months of age, less sterile complementary foods must energy backup must have also been accompanied by be introduced, and infectious disease becomes the evolution of an efficient distribution and delivery unavoidable in all but the most sanitary environments. system. Because the brain accounts for 50–80% of the These childhood infections, in turn, are a source of body’s energy usage during the first few years of life, nutritional stress, and indeed, it is primarily through we should expect that this delivery system – human their effects on nutritional status that they comprom- metabolism – will be organized around the goal of ise health and contribute to mortality during infancy ensuring a constant supply of energy to this fragile and childhood (see Scrimshaw, 1989, for review). Once and costly organ. The availability of dietary nutrients sick, a child loses appetite and this may be com- is variable and often unpredictable, and as we will see pounded in some cultures by the withholding of food next, human metabolism manages this risk by rapidly by caretakers (Scrimshaw, 1989). The common diar- modifying energy allocation in response to changes in rheal diseases reduce nutrient absorption and diges- intake and expenditure. Multiple cues signaling nutri- tion, while the fevers associated with many viral tional stress induce a similar response that achieves a infections can increase metabolic rate and thus energy common end: stored fats are mobilized for use, while expenditure. The specific symptoms may vary by ill- glucose is spared for the brain at the expense of less ness, but the pattern of nutritional depletion that critical functions like muscle. accompanies infections has the effect of suppressing immune function, leaving the infant more prone to future infection and a compounding cycle of nutri- TISSUE-SPECIFIC INSULIN RESISTANCE tional stress (Pelletier et al., 1995). AS A LIFE HISTORY STRATEGY The human infant thus faces a profound energetic dilemma: at precisely the age when they are most Although long-term changes in fat stores are moni- dependent upon provisioning by caretakers to main- tored by the brain through the effects of fat-derived tain the high and obligatory energy needs of their large hormones like leptin, short-term changes in energy brains, they are likely to be cut off from that supply metabolism are regulated in a more distributed fashion chain as a result of illness and the nutritional stresses by organs like the pancreas, and also by the tissues of weaning. It is this confluence of factors, and the themselves, which alter their own uptake and use of synergy between nutritional stress and compromised glucose in response to changes in circulating hormones immunity, that accounts for much of the high infant and the supply of nutrients. A sudden glut of glucose or mortality in many societies (Pelletier et al., 1995). other substrate in the blood stream is sensed by the beta Natural selection likely favored neonatal adiposity as cells of the pancreas, which secrete insulin into the a strategy to prepare for these difficult periods. It is circulation. Insulin stimulates glucose uptake by tissues easy to imagine how infants with a genetic predispos- throughout the body, initiating a negative feedback pro- ition to deposit copious quantities of energy as fat prior cess that helps re-establish normal basal glucose con- to weaning would be better represented among the centrations. The tissues that are not responsive to subset who survive to adulthood to reproduce and pass insulin are few, but prominently include the organ most on their genes (Kuzawa, 1998). vulnerable to energy shortage – the brain (Seaquist It is important to note that these sources of energy et al., 2001). While this may at first seem paradoxical, stress have largely receded by mid childhood: children this tissue-specific pattern of insulin sensitivity provides have already acquired antibodies against the major the body with the ability to modify partitioning of glu- pathogens that they are likely to confront. As a result, cose between the brain and other tissues by increasing

522 Christopher W. Kuzawa 30.4. How body fat and insulin resistance protect the brain. During a crisis, stored triglycerides are broken down into free fatty acids (FFA) and glycerol. Glycerol enters the liver as a gluconeogenic substrate, thus Infection, increasing the pool of available glucose. Free fatty acids Starvation, Inflammatory enter the liver where they are converted to ketone Energy stress (+) bodies, which the brain uses as a substitute energy cytokines ketones substrate. Free fatty acids also induce insulin resistance (+) in the liver and muscle, which spares glucose for the Lipolysis brain. When the nutritional stress is caused by infection, inflammatory cytokines reach the liver and also spare FFA glucose by inducing insulin resistance in liver and muscle. Body fat Liver glycerol GLUCOSE visceral fat FFA (–) insulin resistance Muscle or decreasing insulin-mediated glucose uptake in eventually deplete – its production of insulin in type II the periphery (e.g. muscle, liver, body fat). diabetes (formally noninsulin dependent diabetes, or How this system of resource partitioning works is adult-onset diabetes). However, adopting instead the well illustrated by cases of accidental insulin overdose. perspective of a young organism managing a finite When insulin-dependent diabetics inject themselves energy supply, insulin resistance in a tissue-like skeletal with too much insulin, this can be fatal, because it muscle allows the body to shift its priorities of glucose “overfeeds” the insulin-sensitive periphery, leaving allocation – away from peripheral tissues and toward nothing to fuel the insulin-independent brain (Waring the fragile and energy demanding brain (Figure 30.4). and Alexander, 2004). Reducing insulin-mediated uptake in the periphery has the opposite effect, as it reduces glucose use in tissues like muscle, leaving HOW VISCERAL FAT AND INSULIN more to be delivered to the brain. The body has two RESISTANCE HELP PROTECT THE means of decreasing peripheral glucose use to spare BRAIN DURING A CRISIS the brain when confronted with nutritional stress. The first is to reduce insulin production, which reduces Given these design features, it is not surprising that glucose uptake in the periphery. Secondly, the body insulin resistance is triggered during states associated can influence where glucose is used with greater preci- with negative energy balance, such as starvation, infec- sion by increasing or decreasing insulin sensitivity on a tion, or trauma (Jensen et al., 1987; Childs et al., 1990). tissue-by-tissue basis. Skeletal muscle is a key player in During a fast, free fatty acids (FFA) are first mobilized this flexible system. Because it is the largest consumer from the metabolically active visceral adipose depot, of glucose, the body’s response to changes in energy which is secreted into the portal circulation that drains status prominently involves modulating muscle glucose into the liver. Fat mobilization from this depot is uptake by changing insulin sensitivity in this tissue. The achieved by innervation of the tissue by sympathetic brain is an important arbiter of this response. The brain nerve fibers and, unlike other fat depots, these effects may not be insulin sensitive, but it does express insulin are relatively insensitive to the anti-fat mobilizing receptors and can also sense the adequacy of the supply effects of insulin (van Harmelen et al., 2002; Hucking of glucose that it receives. In the event that cerebral et al., 2003). The liver interprets the sudden appear- glucose flux is attenuated, the brain helps ensure its ance of FFA in the portal circulation as a signal that the own glucose supply by inducing peripheral insulin body is being forced to mobilize fat from this depot, resistance via effects on the sympathetic nervous and reprioritizes metabolism to protect the supply of system and other pathways. In addition to these direct glucose to the brain. This is achieved by increasing effects of the brain, some of the change in glucose hepatic glucose production while also inducing insulin partitioning is co-ordinated directly by the affected resistance in the liver and in skeletal muscle, thus peripheral tissues, as discussed in greater detail below. reducing glucose uptake (Kabir et al., 2005). In add- From the perspective of adult metabolic disease, ition, FFA in the circulation are taken up to be used as this reduced insulin-stimulated uptake of glucose by energy by tissues like muscle, which directly induces skeletal muscle forces the body to increase – and insulin resistance (Roden et al., 1996).

Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome 523 Similar changes in how the body uses glucose are tissue, which saves the more easily metabolized and seen in response to the stress of infection or trauma. desirable energy substrate of glucose for use by the The body’s first response to an infection includes brain. As peripheral insulin resistance helps the body inflammation, which helps mobilize immune and non- control where glucose is used, it seems likely that these immune resources as a first line of defense against general features of human metabolism were forged in tissue invasion and injury (Fernandez-Real and Ricart, the evolutionary crucible of early life nutritional stress. 1999, 2003). Inflammation is initiated in the liver in As emphasized above, the challenge of fueling the response to signals indicating tissue injury, such as the brain, and thus the need to reduce insulin-mediated appearance of proteins that signal the presence of bac- uptake in the periphery during undernutrition or infec- teria, or by molecules called cytokines produced by tion, is most acute, common, and life-threatening at immune cells like macrophages. Interestingly, in obese this age. The need to buffer brain glucose must place individuals visceral adipose tissue is an important an important constraint on any response to nutritional source of these cytokines, which, like FFA, are secreted stress during fetal life, infancy, and early childhood, from fat cells into the portal circulation where they whenthebraindemandstheequivalentof50%ormoreof induce a similar constellation of changes in energy the body’s total available energy (e.g. Childs et al., 1990). metabolism as they reach the liver (Tsigos et al., 1999). Why these overfilled fat cells secrete proinflam- matory cytokines is currently a focus of intensive HOW PRENATAL NUTRITION HELPS FINE research, and there are interesting leads. One is that TUNE THE SETTINGS OF THE BODY’S adipocytes and the immune cells that secrete inflam- ENERGY BACKUP SYSTEM matory cytokines (macrophages) are closely related cells that share similar patterns of gene expression As previously alluded to, there is now a great deal of (Wellen and Hotamisligil, 2003). It is also of interest evidence that individuals born small have higher risk of that fat cells secrete some cytokines as they swell in developing conditions like diabetes and CVD. The pre- size, perhaps helping the cell avoid overfilling or rup- sent discussion provides an opportunity to speculate turing (McCarty, 1999). Regardless of the mechanism on the function of these biological changes, for they underlying the inflammatory effects of excessive body prominently involve resetting how the body prioritizes fat, it seems very likely that these cells are now sending traits like visceral adiposity and insulin resistance signals that, in the past, would have been reliable indi- (for a review see Kuzawa et al., 2007). Although most cators of infection or trauma rather than obesity. It is human research in this field has documented relation- ironic to think that the metabolism of the overfed ships between early life nutrition and metabolic dis- individual today, whose swollen fat cells produce high ease risk factors in later childhood, adolescence, or levels of cytokines, may be tricked by their obesity into adulthood, a handful of studies have now investigated sensing an inflammatory challenge or threat, inappro- these same outcomes in infants and young children – priately initiating the same constellation of metabolic the age when the allocation of scarce glucose may be adjustments designed to cope with starvation or most critical for survival, and thus, when these meta- infection. bolic pathways may have been under particularly What is clear is that multiple signals of nutritional strong evolutionary pressure. What these studies find stress or trauma – whether FFA or cytokines – help link is that, compared to their age mates who were better- the highly labile fat depot of the abdomen with nourished prior to birth, individuals born small tend to metabolic adjustments co-ordinated by the liver put on more visceral fat and become more insulin (Figure 30.4). Not unlike the effects of FFA in the portal resistant as they gain weight (Soto et al., 2003; Ibanez circulation, these cytokines induce a state of peripheral et al., 2006). Metabolic changes in glucose use and insulin resistance, while also mobilizing FFA stored in visceral fat are already detectable in the first year of visceral fat for use as energy. These FFA also bathe the life, showing that a body with a prior history of liver, and thus further contribute to insulin resistance prenatal nutritional stress is not only prone to adult through the pathways described above (Pickup and disease, but also handles its energy and substrate Crook, 1998; Orban et al., 1999). differently during late infancy and early childhood. The evolutionary perspective adopted here, empha- Why the body adopts this strategy when prenatal sizing the importance of the brain metabolism and the nutrition is scarce is uncertain, but two possibilities pattern and sources of nutritional stress, helps clarify seem plausible. Firstly, glucose-sparing adjustments the logic underlying metabolic adjustments that lie at could be important for protecting the brain during the heart of the metabolic syndrome: when the body is prenatal life when intrauterine nutrition is comprom- confronted with a challenge to energy balance, such as ised (Hales and Barker, 1992). The challenge of pro- starvation or infection, the less critical, peripheral tecting the brain, as outlined above, does not begin at tissues shift to using fats mobilized from adipose birth, and nutritional shortfall in utero can result when

524 Christopher W. Kuzawa fetal demand for energy outstrips maternal supply stress, thereby increasing survival and thus genetic across the placenta (Harding, 2001). By this reasoning, fitness. Because the nutritional stress of weaning then, an adjustment made in utero to boost immediate occurs prior to reproductive maturity, it represents a survival may have unintended side-effects after birth “developmental bottleneck” through which any meta- (Hales and Barker, 1992). bolic genes must first pass before being passed on to The second and not mutually exclusive possibility offspring (Kuzawa, 1998). This selection would help is that some of the adjustments have been shaped for ensure that traits that increase early life survival would benefits accrued after birth (Gluckman et al., 2005). To stabilize into the pattern of human ontogeny across the extent that poor prenatal nutrition or maternal many generations of differential survival. I have argued stress are correlated with the world that the fetus will that body fat, especially the rapidly mobilized visceral be born into, these signals could allow it to modify the fat depot, and the ability to reprioritize glucose alloca- priorities of energy use, including the settings of this tion by inducing insulin resistance, may be two backup system, to match the severity of postnatal examples of metabolic traits that could be beneficial nutritional stress or challenge that it is likely to face to an infant faced with the challenge of buffering its (for more on fetal nutrition as an anticipatory cue see most critical functions like the brain during the Bateson, 2001; Kuzawa, 2005; Wells, 2007; Kuzawa, common nutritional stressors confronted at this age. 2008). These adjustments could help the infant manage The developmental influences on these metabolic its precarious metabolic state and survive this traits show that the genes favored by natural selection developmental bottleneck of early life nutritional stress do not determine the organism’s metabolic state in a when conditions are difficult. But this shift in meta- simple one-to-one fashion. Not only does adult health bolic priorities would also have the effect of accentu- depend upon the interaction between one’s genome ating the metabolic derangements that accompany and lifestyle, as long appreciated, but the body also weight gain when nutrition is chronically abundant. appears to have a capacity to adjust the settings and When a previously undernourished body is confronted priorities of its metabolism in response to early nutri- with excess nutrition, a strategy of prioritizing visceral tional experiences and to cues conveyed by the mother fat deposition and sparing glucose becomes a liability, across the placenta. These biological responses may be increasing risk of developing the metabolic syndrome, viewed as examples of the well-known importance of diabetes, and CVD as an adult. In other words, if there developmental plasticity as a mode of human adapt- are “thrifty genes” they code not for static properties ability (Lasker, 1969; Frisancho, 1993; also see Chap- but for flexible strategies, or reaction norms. This abil- ter 2 of this volume). Organisms must cope with ity to flex may help match the body’s metabolism to ecological and social change on a variety of time scales, nutritional stress early in life, but this may plant the spanning rapid and reversible changes (e.g., breakfast seeds for metabolic diseases caused when nutrition is followed by a fast until lunch) to much longer trends chronically abundant later in life. that take years, decades, generations, or longer to unfold (e.g., extended droughts, migrating to a new ecology, an ice age) (Potts, 1998). Genes are effective BEYOND FEAST–FAMINE: FRAGILE BRAINS at tracking the slowest, most gradual changes in ecol- AND METABOLIC PLASTICITY ogy and diet, while homeostasis buffers rapid and reversible changes. Developmental adjustments made In proposing the thrifty gene hypothesis, Neel made in response to early life nutrition operate on an inter- the elegant point that we gain insights into modern mediate time scale, allowing the organism to change its human diseases by reconstructing the environments settings more rapidly than could be achieved by nat- and stressors that our ancestors faced, as this provides ural selection, but in a fashion that is more durable and a sense for what our bodies and biology are designed to stable than what the body achieves via homeostasis expect. I have tried to show how we can approach (Kuzawa, 2005). human metabolism from a developmental perspective In this sense, our metabolisms may not be all that to hone this exercise in reverse-engineering. Rather different from other systems that have a capacity to than inferring past selection from the default clinical adjust long-term settings to local conditions via devel- perspective of adult obesity or diabetes, the develop- opmental responses, illustrated classically by the mental approach developed in this chapter began by important effects of developmental experience on isolating ages of high mortality to identify when systems like the brain, the immune system, and the metabolism is most critical for survival. Identifying skeleton. Many of the body’s systems are built from a the sources of nutritional stress at the ages of peak genetic architecture that evolved through natural selec- nutritional mortality provides clues into the types of tion, but that – by design – rely upon information about biological responses or strategies that would have been local ecology acquired during early development to available to natural selection to work around this complete their construction. There is little doubt that

Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome 525 the size, configuration, and attachments of the body’s 4. This chapter argues that developmental adapta- skeletal elements evolved through natural selection tions made by the fetus to nutritional stress could operating on gene frequencies. However, because indi- contribute to a higher risk for diabetes later in life. viduals are idiosyncratic in their behavior and thus the What are these adaptations and how might they be mechanical loadings that they will impose on their beneficial early in life? skeletons, natural selection constructed a system that 5. Chronic diseases often negatively impact health has an exquisite capacity to compensate for biomecha- and survival late in the lifecycle – when one’s nical strain, and to organize developmentally around genome has already been passed on via offspring the pattern of use and disuse in the individual. Perhaps to the next generation. Discuss how selection pres- it should come as no surprise that the body’s priorities sures operating early in the lifecycle might influ- of energy allocation, which are so critical for survival, ence the evolution of diseases with late life negative should have a similar capacity. After all, humans do impacts. not merely vary in their culture, pathogen ecology, and 6. If we ask “What are the causes of the diabetes pattern of physical activity, but also in their diet, nutri- epidemic?”, a public-health or medicine-inspired tional sufficiency, and exposure to metabolically answer to this question might note that many demanding stressors. These physiological and meta- people are eating too much and gaining weight, bolic “loadings” may be hidden from view, but they among other factors, which leads to insulin resist- are just as critical as components of the individual’s ance. What types of answers does an evolutionary adaptive strategy. The ability to adjust metabolic set- approach to this problem inspire, and how are tings to local conditions might help humans prepare they different from a public health or medical for important adaptive challenges, like the severity of approach? infant and early childhood nutritional stress outlined here. However, this same flexibility may also carry longer-term costs to health in modern environments, ACKNOWLEDGEMENTS especially if scarcity-adapted metabolic priorities adopted early in life are later confronted with chronic Dan Benyshek, Dan Eisenberg, Peter Ellison, Lee overnutrition and weight gain (Gluckman et al., 2005). Getler, David Haig, Elizabeth Quinn, Melanie Vento, It remains to be determined which of these ideas and two anonymous reviewers provided helpful com- about the function of these traits are correct in detail. ments on various drafts of this paper. What does seem clear is that our metabolisms are designed to do more than survive the crisis of famine. Identifying the ages and causes of nutritional stress REFERENCES and nutritional mortality – both external and internal Adair, L. S., Kuzawa, C. W. and Borja, J. (2001). Maternal to the body itself – will be critical if we hope to under- energy stores and diet composition during pregnancy stand how natural selection responded to this stress, program adolescent blood pressure. Circulation, 104, and the legacy of these adaptations for the current 1034–1039. global plague of metabolic disease. Aiello, L. and Wheeler, P. (1995). The expensive-tissue hypothesis; the brain and the digestive system in human and primate evolution. Current Anthropology, 36, 199–221. DISCUSSION POINTS Armstrong, E. (1983). Relative brain size and metabolism in mammals. Science, 220, 1302–1304. 1. Human babies are born with more body fat than Barker, D. J., Osmond, C., Golding, J., et al. (1989). Growth any other mammal – including seals. What might in utero, blood pressure in childhood and adult life, and explain the evolution of this unusual trait? mortality from cardiovascular disease. British Medical 2. The human brain is energetically very costly, espe- Journal, 298, 564–567. cially early in the lifecycle. What are some of the Bateson, P. (2001). Fetal experience and good adult design. ways that evolution may have changed human biol- International Journal of Epidemiology, 30, 928–934. Benyshek, D. C. and Watson, J. T. (2006). Exploring the thrifty ogy to allow the evolution of a large brain size in genotype’s food-shortage assumptions: a cross-cultural humans? comparison of ethnographic accounts of food security 3. It is well known that diseases like diabetes and among foraging and agricultural societies. American cardiovascular disease “run in families,” reflecting Journal of Physical Anthropology, 131, 121 –126. a genetic contribution. Yet, these diseases have Childs, C., Heath, D. F., Little, R. A., et al. (1990). Glucose emerged as major public health problems in a few metabolism in children during the first day after burn short generations – which is not enough time for injury. Archives of Emergency Medicine, 7, 135–147. gene frequencies to change substantially. How do Cohen, M. N. and Armelagos, G. J. (1984). Paleopathology we reconcile these observations? at the Origins of Agriculture. New York: Academic Press.

526 Christopher W. Kuzawa Davies, P. S. W. and Preece, M. A. (1989). Body composition Kabir, M., Catalano, K. J., Ananthnarayan, S., et al. (2005). methods in children: methods of assessment. In The Physi- Molecular evidence supporting the portal theory: a causa- ology of Human Growth,J.M.TannerandM.A.Preece tive link between visceral adiposity and hepatic insulin (eds). Cambridge: Cambridge University Press, pp. 95–107. resistance. American Journal of Physiology. Endocrinology Dowse, G. and Zimmet, P. (1993). The thrifty genotype in and Metabolism, 288, E454–E461. non-insulin dependent diabetes. British Medical Journal, Kissebah, A. H., Vydelingum, N., Murray, R., et al. (1982). 306, 532–533. Relation of body fat distribution to metabolic complica- Eaton, S. B. and Konner, M. (1985). Paleolithic nutrition. tions of obesity. Journal of Clinical Endocrinology and A consideration of its nature and current implications. Metabolism, 54, 254–260. New England Journal of Medicine, 312, 283–289. Knowler, W., Savage, P., Nagulesparan, M., et al. (1982). Fernandez-Real, J. M. and Ricart, W. (1999). Insulin resist- Obesity, insulin resistance and diabetes mellitus in the ance and inflammation in an evolutionary perspective: the Pima Indians. In The Genetics of Diabetes Mellitus, contribution of cytokine genotype/phenotype to thrifti- J. Kobberling and R. Tattersall (eds). London: Academic ness. Diabetologia, 42, 1367–1374. Press, pp. 243–250. Fernandez-Real, J. M. and Ricart, W. (2003). Insulin resist- Kuzawa, C. W. (1997). Body fat as system: an evolutionary ance and chronic cardiovascular inflammatory syndrome. and developmental consideration of the growth and func- Endocrine Reviews, 24, 278–301. tion of body fat (abstract). American Journal of Physical Fernandez-Twinn, D. S. and Ozanne, S. E. (2006). Mechan- Anthropology, 24(suppl.), 148. isms by which poor early growth programs type-2 Kuzawa, C. W. (1998). Adipose tissue in human infancy and diabetes, obesity and the metabolic syndrome. Physiology childhood: an evolutionary perspective. American Journal and Behavior, 88, 234–243. of Physical Anthropology, 27(suppl.), 177–209. Foley, R. and Lee, P. (1991). Ecology and energetics of ence- Kuzawa, C. W. (2005). Fetal origins of developmental phalization in hominid evolution. Proceedings of the Royal plasticity: are fetal cues reliable predictors of future nutri- Society of London. Series B, 334, 223–232. tional environments? American Journal of Human Biology, Frisancho, A. (1993). Human Adaptation and Accommoda- 17, 5–21. tion. Ann Arbor, MI: University of Michigan Press. Kuzawa, C. W. (2008). The developmental origins of adult Gluckman, P. D. and Hanson, M. A. (2006). Developmental health: Intergenerational inertia in adaptation and dis- Origins of Health and Disease. Cambridge: Cambridge Uni- ease. In Evolutionary Medicine and Health: New Perspec- versity Press. tives, W. Trevathan, E. Smith and J. McKenna (eds). New Gluckman, P. D., Hanson, M. A., Morton, S. M., et al. (2005). York: Oxford University Press, pp. 325–349. Life-long echoes – a critical analysis of the developmental Kuzawa, C. W. and Adair, L. S. (2003). Lipid profiles in origins of adult disease model. Biology of the Neonate, 87, adolescent Filipinos: relation to birth weight and maternal 127–139. energy status during pregnancy. American Journal of Haig, D. (1993). Genetic conflicts in human pregnancy. Clinical Nutrition, 77, 960–966. Quarterly Review of Biology, 68, 495–532. Kuzawa, C. W., Gluckman, P. D., Hanson, M. A., et al. Hales, C. N. and Barker, D. J. (1992). Type 2 (non-insulin- (2007). Evolution, developmental plasticity and metabolic dependent) diabetes mellitus: the thrifty phenotype disease. In Evolution in Health and Disease, S. C. Stearns hypothesis. Diabetologia, 35, 595–601. and J. C. Koella (eds), 2nd edn. Oxford: Oxford University Harding, J. E. (2001). The nutritional basis of the fetal Press, pp. 253–264. origins of adult disease. International Journal of Epidemi- Lasker, G. W. (1969). Human biological adaptability. The ology, 30, 15–23. ecological approach in physical anthropology. Science, Hardy, A. (1960). Was man more aquatic in the past? New 166, 1480–1486. Scientist, 7, 642. Lawlor, D. A., Ronalds, G., Clark, H., et al. (2005). Birth Holliday, M. (1986). Body composition and energy needs weight is inversely associated with incident coronary during growth. In Human Growth: a Comprehensive Trea- heart disease and stroke among individuals born in the tise, F. Falkner and J. M. Tanner (eds). New York: Plenum 1950s: findings from the Aberdeen Children of the 1950s Press, pp. 117–139. Prospective Cohort Study. Circulation, 112, 1414–1418. Hucking, K., Hamilton-Wessler, M., Ellmerer, M., et al. Leonard, W. and Robertson, M. (1992). Nutritional require- (2003). Burst-like control of lipolysis by the sympathetic ments and human evolution: a bioenergetics model. nervous system in vivo. Journal of Clinical Investigation, American Journal of Human Biology, 4, 179–195. 111, 257–264. Leonard, W. and Robertson, M. (1994). Evolutionary per- Ibanez, L., Ong, K., Dunger, D. B., et al. (2006). Early devel- spectives on human nutrition: the influence of brain and opment of adiposity and insulin resistance following body size on diet and metabolism. American Journal of catch-up weight gain in small-for-gestational-age chil- Human Biology, 6, 77–88. dren. Journal of Clinical Endocrinology and Metabolism, Lucas, A. (1991). Programming by early nutrition in man. 91, 2153–2158. Ciba Foundation Symposium, 156, 38–50, discussion Jensen, M. D., Haymond, M. W., Gerich, J. E., et al. (1987). 50–55. Lipolysis during fasting. Decreased suppression by insulin Mackay, J., Mensah, G. A., Mendis, S., et al. (2004). The Atlas and increased stimulation by epinephrine. Journal of of Heart Disease and Stroke. Geneva: World Health Clinical Investigation, 79, 207–213. Organization.

Beyond Feast–Famine: Brain Evolution, Human Life History, and the Metabolic Syndrome 527 McCarty, M. F. (1999). Interleukin-6 as a central mediator of Expenditure, and Energy Requirements of Infants and Chil- cardiovascular risk associated with chronic inflammation, dren, B. Schurch and N. Scrimshaw (eds). Switzerland: smoking, diabetes, and visceral obesity: down-regulation Nestle ´ Foundation, pp. 215–238. with essential fatty acids, ethanol and pentoxifylline. Seaquist, E. R., Damberg, G. S., Tkac, I., et al. (2001). The Medical Hypotheses, 52, 465–477. effect of insulin on in vivo cerebral glucose concentrations McGarvey, S. (1994). The thrifty gene concept and adiposity and rates of glucose transport/metabolism in humans. studies in biological anthropology. Journal of Polynesian Diabetes, 50, 2203–2209. Society, 103, 29–42. Soto, N., Bazaes, R. A., Pena, V., et al. (2003). Insulin sensi- McMillen, I. C. and Robinson, J. S. (2005). Developmental tivity and secretion are related to catch-up growth in origins of the metabolic syndrome: prediction, plasticity, small-for-gestational-age infants at age one year: results and programming. Physiological Reviews, 85, 571–633. from a prospective cohort. Journal of Clinical Endocrin- Morris, D. (1967). The Naked Ape. New York: McGraw-Hill. ology and Metabolism, 88, 3645–3650. Neel, J. (1962). Diabetes mellitus: A “thrifty” genotype Speakman, J. R. (2006). Thrifty genes for obesity and the rendered detrimental by “progress”? American Journal of metabolic syndrome–time to call off the search? Diabetes Human Genetics, 14, 353–362. and Vascular Disease Research, 3, 7–11. Nesse, R. M. and Williams, G. C. (1994). Why We Get Sick: the Stearns, S. C. and Koella, J. C. (2008). Evolution in Health New Science of Darwinian Medicine. New York: Times Books. and Disease. Oxford: Oxford University Press. Orban, Z., Remaley, A. T., Sampson, M., et al. (1999). The Trevathan, W., Smith, E. O. and McKenna, J. J. (2008). differential effect of food intake and b-adrenergic sti- Evolutionary Medicine and Health: New Perspectives.New mulation on adipose-derived hormones and cytokines in York: Oxford University Press. man. Journal of Clinical Endocrinology and Metabolism, Tsigos, C., Kyrou, I., Chala, E., et al. (1999). Circulating 84, 2126–2133. tumor necrosis factor alpha concentrations are higher Owen, O. E., Morgan, A. P., Kemp, H. G., et al. (1967). Brain in abdominal versus peripheral obesity. Metabolism, metabolism during fasting. Journal of Clinical Investiga- 48, 1332–1335. tion, 46, 1589–1595. Ulijaszek, S. and Lofink, H. (2006). Obesity in bio- Pelletier, D., Frongillo, E., Shroeder, D., et al. (1995). The cultural perspective. Annual Review of Anthropology, 35, effects of malnutrition on child mortality in developing 337–360. countries. Bulletin of the World Health Organization, Vague, J. (1955). [The distinction of android and gynoid 73, 443–448. obesities as the base for their prognosis.] La semaine des Pickup, J. C. and Crook, M. A. (1998). Is type II diabetes ho ˆpitaux, 31, 3503–3509. mellitus a disease of the innate immune system? Diabeto- van Harmelen, V., Dicker, A., Ryden, M., et al. (2002). logia, 41, 1241–1248. Increased lipolysis and decreased leptin production Pond, C. (1997). The biological origins of adipose tissue in by human omental as compared with subcutaneous pre- humans. In The Evolving Female. A Life History Perspec- adipocytes. Diabetes, 51, 2029–2036. tive, M. Morbeck, A. Galloway and A. Zihlman (eds). Waring, W. S. and Alexander, W. D. (2004). Emergency Princeton: Princeton University Press, pp. 147–162. presentation of an elderly female patient with profound Potts, R. (1998). Environmental hypotheses of hominin evo- hypoglycaemia. Scottish Medical Journal, 49, 105–107. lution. American Journal of Physical Anthropology, 27 Wellen, K. E. and Hotamisligil, G. S. (2003). Obesity- (suppl.), 93–136. induced inflammatory changes in adipose tissue. Journal Reaven, G. (1988). Role of insulin resistance in human of Clinical Investigation, 112, 1785–1788. disease. Diabetes, 37, 1595–1607. Wells, J. (2007). Environmental quality, developmental Roden, M., Price, T. B., Perseghin, G., et al. (1996). Mechan- plasticity and the thrifty phenotype: a review of evolution- ism of free fatty acid-induced insulin resistance in ary models. Evolutionary Bioinformatics, 3, 109–120. humans. Journal of Clinical Investigation, 97, 2859–2865. Yajnik, C. S. (2004). Early life origins of insulin resistance Scrimshaw, N. (1989). Energy cost of communicable dis- and type 2 diabetes in India and other Asian countries. eases in infancy and childhood. In Activity, Energy Journal of Nutrition, 134, 205–210.

31 Human Longevity and Senescence Douglas E. Crews and James A. Stewart INTRODUCTION biological time using physiological biomarkers, cas- cades of interrelated metabolic events (that contrary From a life history perspective, growth and develop- to developmental integration, lead toward disintegra- ment are processes that enhance the functional tion), cell and transcription cycles (the fundamental capacities of an organism through attainments of com- units of biological time), gene expression patterns, petencies and physiological signaling across somatic changes in the proteome, epigenetic effects, and other systems, the “growth span” (Arking, 2006). The period factors (Arking, 2006). Measures of senescence do not following growth and development is the mature state. depend upon the passage of absolute time. Rather they This may usefully be described as the “health span” are secondary to current competencies, organism-wide during which the soma is most able to maintain itself, integration via messenger molecules, and the ability of while also reproducing and investing in reproductive receptors to receive various signals and modulate cel- effort. Senescence follows the health span. It is a lular functions (Finch and Rose, 1995). Here we exam- later occurring age-independent, event-driven process ine human senescence within the context of overall measured in biological time as cascades of physical organismic, large-bodied mammalian, and primate and chemical events that increase risks of death in evolution, and humankind’s unique biocultural adap- the next time interval, the “senescent span” (Crews, tations. Biocultural interactions extensively altered 2003; Arking, 2006). how today’s humans and earlier hominins interacted Senescence is a biological process that only the with their environments by modifying multiple eco- living experience; it is the final period of life history logical relationships. We start with a brief background 1 (LH) for most organisms whether unicellular or pos- on the development of evolutionary theories of life sessing both a soma and a germline. span and how human longevity intrigued early evolu- Senescent changes tend to be cumulative (increas- tionary thinkers and continues to intrigue modern ing vulnerability to challenge over time), progressive biologists. Following sections explore general evolu- (losses are gradual), intrinsic (not resulting from modi- tionary theories, the biology of senescence across fiable environmental factors), and deleterious (reduced organisms, some senescent mechanisms, aspects of function increases mortality risk in the next time inter- comparative biology, and briefly review similarities val) (Arking, 2006). Senescence may be measured in in primate LH, before reviewing trends in later Homo and human LH and life span, biocultural evolution 1 in constructed environments, and our species’ path to Life history (LH): In its simplest form, LH has been defined simply as things to which r, the rate of population growth, is modern life styles and longevity. most sensitive (Stearns, 1992, p. 32). It includes all aspects of a species growth, reproduction, and survivorship. Individual and population traits, including length of gestation, litter/clutch size, maturity at birth/hatching, ages at first menses (menar- che), first reproduction, weaning, last reproduction, and meno- EVOLUTIONARY MODELS pause, birth-spacing, interbirth intervals, age at puberty, length of growth and development period, age-specific fecundity/fertil- Historical evolutionists ity, and age at death are the data for studying LH evolution and examined as LH covariates. Understanding the coevolution of Darwin (Charles Robert, 1809–1882) noted that senes- these traits, along with age-specific mortality, survivorship, and cence presents multiple problems for evolutionary reproduction is the goal of LH research. Major components of theory (Darwin, 1859). However, it was in papers LH research includes understanding trade-offs between somatic maintenance and reproductive efforts, variable selective pres- presented between 1881 and 1889 that Weismann sures on LH traits, and biocultural and behavioral influences on (Friedrich Leopold August, 1834–1914) first proposed human LH. Life history theory is the study of the evolution and an integrated evolutionary theory of senescence and function of life states and behaviors related to these stages (Stearns, 1992). longevity. Weismann (1881–1889) suggested that death Human Evolutionary Biology, ed. Michael P. Muehlenbein. Published by Cambridge University Press. # Cambridge University Press 2010. 528

Human Longevity and Senescence 529 3 was adaptive in environments with limited resources reproductive potential (MRP : see Crews, 2003, 2007; even if organisms began with intrinsically long life Larke and Crews, 2006) and ages when the majority of spans of indeterminate length. In such a setting, nat- reproductive effort occurs, alleles with deleterious ural selection (NS) would favor organisms that passed effects at older ages, but with little or no effects at on their germlines before others and before environ- younger ones, may remain in the gene pool (Williams, mental insults and trauma could kill them, while 1957; Kirkwood, 2002). Alleles with late-acting detri- investing less in their somas than longer-lived, slow mental effects may then accumulate beyond ages at reproducing individuals. According to Weismann, which NS is capable of eliminating them. Conse- those with bodies better able to procure scarce quently, broad senescent changes may result from resources for investing in reproduction early in life these deleterious effects at ages when reproductive would create an array of living organisms whose life potential declines toward zero. Some may doubt that spans were inversely proportional to their rates of such late-acting mutations are responsible for popula- maturation and their prereproductive and reproductive- tion-level senescent processes. However, as pointed out age mortality rates. He saw that the quicker a species by Lasker and Crews (1996), variable mutations and achieved LH landmarks, the shorter its average and mutation rates across populations and environments, maximum life spans need be. coupled with linkage disequilibrium, will lead to differ- Many extant species show a semelparous LH, death ent alleles accumulating in various populations. As occurs after a single lifetime reproductive event. This is long as these are not linked to other alleles that reduce, particularly so for insects, annual plants, and some fish or they reduce fitness themselves, gene flow will spread species (salmon). Unless an organism must contribute these alleles and, importantly, their linked haplotypes, to raising its young, there is little push for somatic across a species. [For example multiple viral DNA seg- maintenance after reproduction (Weismann, 1881– ments have hitchhiked within the human genome for 1889; Crews, 2003, 2005a, 2007; Larke and Crews, millions of years, up to 8% of the total (Feschotte 2006). As larger organisms commonly develop more 2010).] When linked haplotypes show benefits to repro- 2 complex somas, provide parental investment (PI) to ductive success (RS) in many environments, NS may their offspring, who develop slowly, and have more maintain or increase their frequency in local settings. than a single reproductive cycle (iteroparous), their Population bottlenecks, founder effects, and various deaths are postponed to later ages by NS. Weismann forms of genetic drift also may increase frequencies suggested that for most organisms, there is no advan- of late-acting detrimental alleles within local popula- tage to keeping ancient, less competitive individuals tions. Thus late-acting detrimental alleles may “hitch- around to deplete resources needed for younger gene- hike” along with beneficial alleles within linked rations. As with many of Weismann’s conjectures, haplotypes, hidden till life spans increase to ages at such group-selection models are not accepted today. which they are expressed. However, he did presage Dawkin’s (1989) selfish genes In addition to addressing mutation accumulation and Wilson’s (1975) group selection approaches following Medawar (1952), Williams (1957) recognized and conceptualized a disposable soma, as was later that pleiotropy also might lead to senescence. Alleles developed more fully by Kirkwood (1977). Additionally, beneficial during the growth span or that enhance many later theoretical perspectives on senescence, reproduction and MRP during the health span may be including those of P. Medawar (1952), G. C. Williams incompatible with long life. Alleles enhancing senes- (1957), T. B. L. Kirkwood (1977), R. Holliday (1997), and cence in mid- to late-life, but improving survival during W. D. Hamilton (1966), were presaged by Weismann’s infancy, childhood, and attainment of MRP exhibit (1881–1889) insights into evolutionary processes and antagonistic pleiotropy (early benefits with late-acting senescence. deleterious effects). Due to trade-offs between repro- duction and somatic maintenance, once an organism begins reproducing and fledging offspring, NS Modern evolutionists gradually loses its ability to eliminate alleles that are Age-specific gene action and pleiotropy deleterious to the organism’s extended somatic main- Medawar (1952) first suggested age-specific gene tenance. For example, apoptosis is crucial during action and mutation accumulation contributed to sen- embryonic development for sculpting the developing escence. As NS diminishes at ages beyond maximum organism as stem cells create more new cells than are 3 Maximum reproductive potential (MRP): Is the point in life at 2 Parental investment (PI): All expenditures of parents’ time, which an organism is sufficiently mature to not only bear/sire energy, and future reproductive effort on one offspring at a cost offspring, but also best able to rear and fledge any offspring to parents’ abilities to invest in their own somas or other pos- produced with maximum efficiency (Crews and Gerber, 1994; sible offspring and enhance their own fitness (Hamilton, 1964, Gerber and Crews, 1999; Harper and Crews, 2000; Crews, 2003; 1966). Larke and Crews, 2006).

530 Douglas E. Crews and James A. Stewart needed. While the soma is sculpted, external markers are minimized. However, once sexual maturation and (ligands) regulate apoptosis of excess cells. Once adult- MRP are attained, a trade-off or competition between hood is achieved, animals are no longer growing or investing in maintaining the “disposable soma” or developing and rates of apoptosis are adjusted accord- reproducing the “immortal germline” through off- ingly as morphogenesis is completed. During adult- spring ensues (Kirkwood, 1990, 2002, 2005). Models hood most cells that are damaged, malfunctioning or proposing LH trade-offs between the soma and repro- show damaged DNA are marked for apoptosis by ductive effort mainly focus upon the postgrowth span internal cues (e.g., mitochondrial factors, p53). How- of LH. The soma is not disposable prior to achieving ever, proliferation of new cells to replace those that are reproductive maturation and RS. Trade-offs expressed worn out and/or lost through apoptosis may not be during the growth phase are associated with early LH balanced and tissue deterioration may occur. Loss of strategies of high versus low early life reproduction cells from apoptosis and accumulation of senescent and rapid versus slow growth and development. In cells may jointly contribute to senescence and tissue some species LH patterns have deep evolutionary roots deterioration (Campisi, 2003). Williams (1957) also reflecting phylogenetic inertia. For example, mayflies, suggested that as life span increases NS bumps up fruit flies, annual plants, many insects, amphibians, against detrimental alleles that have been retained and fish show rapid life histories, little to no PI, rapid through mutation accumulation, at what was previ- growth, and high numbers of eggs, seeds, and sperm ously the upper limit of reproductive effort. When this upon reproductive effort; along with rapid senescence occurs, alleles at other loci that modify the late-acting and high intrinsic mortality. Conversely, other species detrimental effects of accumulated alleles, or alternate of insects, plants, amphibians, reptiles, fish, and alleles at the same loci, may increase in frequency if mammals, notably cicadas, wasps and other hymenop- they promote longer life along with greater RS. How- tera species, redwood and bristlecone pine trees, ever, any alleles that improve prereproductive survival turtles, sharks, whales, and elephants show a different or improve RS in early life will not be eliminated via LH pattern. There are no phylogenetic continuums NS, regardless of their late life effects. of LH strategies nor does phylogeny necessarily predict LH, different species within even a single genus may Disposable somas and immortal genomes show different LH patterns and the same species may Accumulated evidence now suggests that senescence show markedly different LH patterns under different may occur in most organisms. For many decades fissile environmental constraints. organisms were thought to not show senescence and Early in life and during the period of maximum therefore be in a sense immortal. However, recent reproductive effort, NS favors somatic maintenance research on Escherichia coli and fissile single celled in species that reproduce multiple times (iteroparous) prokaryotes suggest they do senesce (Stewart et al., or that show PI in their offspring after birth/hatching 2005). Not only do eukaryotic, sexually reproducing (Hamilton, 1964, 1966; Crews, 2003, 2005a, 2008; species with telomeres to shorten with each round of Larke and Crews, 2006). As life continues, investment mitosis, LH stages to grow and develop through, antag- in reproductive effort eventually takes priority over onistic pleiotropy, vital macromolecules to oxidize, somatic maintenance and senescent processes may and proteomes that are degraded over time show sen- not be eliminated, because additional somatic invest- escence, it now seems that all living organisms may ments yield trivial reproductive benefits. Eventually, show senescence. the net effect of NS declines to zero. Metazoans have Kirkwood’s (1977, 1990, 2002, 2005) and Kirkwood complex, interrelated and interdependent cells, organs, and Holliday’s (1979) elaborations of Weismann’s dis- and systems within their somas. Senescence of cells posable soma theory of senescence proposes a distinc- begins early in the life span as strategic resources are tion between evolutionary theories and physiological directed toward replicating the immortal germline at descriptions of somatic senescence (see also Crews, the expense of the soma (Austad, 2005a; Kirkwood, 2003). An organism’s resources, such as energy, are 1977; Hayflick, 1979). Among species with high PI limited and must be divided between self-preservation and multiple reproductive cycles, NS works to main- (the soma) and continuance of the immortal germline tain the parental soma because it is the sole reproduct- across generations (reproduction). During the growth ive organ and source of continued investment in span resources are invested into the soma. Throughout reproductive effort and PI. Simple mathematical for- the growth span, the soma and germline share the mulas, such as the Gompertz or Weibull equations, same goal, producing a mature soma to pass on the model the increase in somatic susceptibility over time germline. Because the soma is the organ by which the from multiple, interrelated, and unrelated intrinsic and immortal germline ultimately must reproduce itself, extrinsic factors that combine to eventually kill all over the growth span trade-offs between somatic organisms. All metazoans senesce after they attain growth/development and transmitting the germline MRP. Although some processes of senescence even

Human Longevity and Senescence 531 may be initiated as early as fetal life, most occur after evolutionary theories of senescence advanced in MRP. For some species, a previously unidentified LH parallel with molecular biology. Cellular and compara- stage may lie just beyond the senescent phase. During tive biology, breeding experiments with animal models this “late-life stage,” the rate of increase in senescence and, more recently, growth of cells in laboratories, and mortality hazards may slow among some experi- analyses of DNA and transfers of DNA into and out mental and free-living populations (Carey et al., 1992; of animal models have improved understandings of Carnes and Olshansky, 2001; Rose et al., 2005b; Perls, cellular senescent biology. It is now clear that across 2006; Rauser et al., 2006). organisms, multiple alleles at many loci and differen- Rauser et al. (2006) suggest that such a late-life tial processing of proteins produce complex molecular phase may begin when an organism’s survivorship pathways modulating physiological functioning. and fecundity deteriorations cease and mortality rate Some of these are incompatible with longevity; others increases approach zero. The Gompertz equation are associated with high somatic maintenance and x (ðxÞ¼Ae ) predicts that as organisms age, their mor- extended life spans. Among humans both culture and tality rates increase exponentially (Hamilton, 1966, environment influence and alter these multiple genetic/ reviewed in Box 13.1 in Crews and Ice, 2010). However, protein-based predispositions. Across species and the slope of the Gompertz equation decreases slightly populations average age at first reproduction and after about age 85, suggesting decreasing increases average life span are closely associated. Variation in mortality may happen in late life. Late-life survival in life spans and reproductive potential within and is better modeled by Hamilton (1966): between populations and species shows that environ- ments modulate and extensive genetic variation Ae x X ry sðxÞ¼ e lðyÞmðyÞ and  ðxÞ¼ underlies LH (Perls, 2001, 2006; Carey and Judge, 2 x  1Þ 1 þ  Aðe½Š 1 y¼xþ1 2001; Perls et al., 2002; Crews, 2003; Perls and Terry, Where s(x) is the variance in survival and  (x) is the 2003; Pak et al., 2003; Bredesen, 2004; Holliday, 2004; average age at death. Other variables in the equations Austad, 2005a, 2005b; Kirkwood et al., 2005; Rose are, x representing age, r the Malthusian growth rate, et al., 2005a, 2005b; Christensen, et al., 2006; Harper l(y) the specified survivorship to age x, and m(y) et al., 2006a). age-specific fecundity, where y is the net expected reproduction at all ages x and higher. A is an age- Cellular biology independent parameter representing baseline mortal- ity within the population, and a is an age-independent Cell replicative ability parameter representing senescence (rate of aging in Hayflick (1965) reported that mitosis by dividing cells Hamilton’s terms). Data from some laboratories and (fibroblasts) cultured in vitro was limited. Contrary to experimental species suggest mortality rates at later the then prevailing opinion, that dividing cells pos- ages may follow patterns represented by Hamilton’s sessed an unlimited capacity for self-reproduction, (1966) equations. This leads to suggestions that a late- Hayflick observed that fibroblasts from newborns life phase with nonincreasing mortality rates may doubled in vitro only about 50–60 times. How, and if, occur late in the LH of some multicellular organisms Hayflick’s eponymous limit is associated with human (Hamiliton, 1964, 1966; Rose et al., 2005b; Rauser senescence was then, and remains today, largely et al., 2006). Hamilton (1966) suggested such a unknown. Some researchers suggest that dividing cells phenomenon might reflect declining NS over time. may come close to their in vitro doubling limit when Humans may show some late-life stabilization of age- they produce and maintain a mature adult reproduct- specific mortality rates as age-related deterioration ive human. However, current estimates are that only slows or ceases after 85 or 100 years (Perls, 2001; about 42 cell doublings by the zygote are needed to Doblhammer and Kytir, 2001; Rose et al., 2005b; produce a mature infant and only another 4–6 to grow Crimmins and Finch, 2006; Rauser et al., 2006). an adult soma and maintain it thereafter (Dennison et al., 1997). This would leave the average person with more than sufficient cell doublings to maintain their THE MODERN SYNTHESIS AND soma over 10 decades. Limited cell doublings as SENESCENT BIOLOGY defined by Hayflick (1965) may still be related to cell culture methods. By dividing the cell population, the The synthesis of evolutionary theory with genetics number of primary fibroblasts left to produce new cells combined with developments in multivariate statistical is halved also (Cristofalo et al., 1998, 1999, 2003; Clark, analyses of genes and population genetics during 1999). Among the criteria for an eponymous limit is a the mid twentieth century allowed some analyses of specified time during which the culture must cover the complex gene–environment interactions produ- its environment; without a time limit, cultures may cing senescence. During the late twentieth century, continue dividing slowly to confluence.

532 Douglas E. Crews and James A. Stewart Animals do not appear to “run out” of cells over mass (Aviv, 2002; Blasco, 2005). However, other bio- their life spans due to limited cell proliferations in vivo, markers of senescence are not highly correlated with nor do doublings in cell culture correlate as strongly telomere shortening (Blasco, 2005). with age as was originally suggested (Hart and It is now clear that shortening of telomeres limits Tuturro, 1983). For example, total doublings for cells cellular replicative life spans and leads to the Hayflick cultured from the same individual may show an aver- limit of human fibroblasts in vitro (Olovnikov et al., age of about 50 doublings, but a range from 10 to 90 1996; Shay, 1997; Aviv, 2002; deLange, 2002). Dividing doublings (Hart and Tuturro, 1983). Cells from differ- cells that reach their replicative limit in vivo become ent individuals of the same age also show wide vari- senescent and tend to accumulate in somatic tissues. ation in their total numbers of doublings; cells from Such data suggest that chromosome shortening meas- infants average about 50–60, but range from 25–120, ures or represents intrinsic cellular senescence, an while those from persons aged 65 years average about internal biological clock in dividing cell clones. Still, 20 doublings, with a range from 5–50. Across individ- cell replicative capacity is not closely associated uals of all ages the average is about 35 doublings, but with individual life spans. An unknown number of again the range is large, from 0–120 (Hart and Tuturro, variables – including cell type, donor age, oxidative 1983). Additionally, no significant correlation of fibro- stress, and heredity – influence correlations between blast doublings and age of cell donors, was observed in telomere shortening and age, both across and within a small sample of healthy Baltimore Longitudinal individuals. Telomere length varies across peripheral Study of Aging participants (n ¼ 42) (Cristofalo et al., lymphocytes from the same individuals, between simi- 1998, 1999). Also problematic is the finding that fibro- lar cells from different individuals of the same age, and blasts from some very old individuals show no across cell cultures of fibroblasts from the same donor doubling potential at all, yet these individuals continue (Harley et al., 1990; Hastie et al., 1990; Slagboom et al., to survive. Given these accumulated data, relationships 1994). Studies of human peripheral lymphocytes sug- between in vitro doublings and in vivo senescence gest an average loss of telomeric DNA of about 31–33 are unclear and may be less important than other base pairs (bp)/year (Hastie et al., 1990; Slagboon et al., processes (Cristofalo et al., 2003). 1994). The average, across various types of human During terminal passages in vitro cells do tend to cells, ranges from 40 to 200 bp/division (Allsopp, show multiple alterations that are similar to those of 1996), compared with an average of 50 bp/doubling in senescent cells in vivo (Clark, 1999; Campisi, 2003; cultured cells. Telomere length is sexually dimorphic Aviv, 2002). Data on telomere shortening suggest that in humans at birth, women’s telomeres are longer then in vivo skin, thyroid, immune, and digestive cells may men’s (Aviv, 2002). Men also appear to senesce faster fail to replace themselves adequately with increasing than women until late life (Foley, 1986; Perls, 1995; age (Clark, 1999). In general, loss of proliferate cap- Anderson-Ranberg et al., 1999), although this may acity in vitro is not thought to produce or reflect pro- not be so (Graves et al., 2006). Men may require more cesses of senescence. Loss of replicative capacity shows cell doublings to accommodate their 15–20% larger how a complex molecular system fails to maintain body sizes, exhausting their shorter telomeres earlier itself infinitely due to intrinsic fragility and random in life. Although attrition of telomeric DNA tandem losses of function. Average numbers of doublings in repeats (TTAGGG) paces in vitro replicative history of vitro are correlated loosely with age of cell donors but primate, including human, cells, a similar association are not highly predictive of individual life spans. Loss is not observed in rodent cells (Aviv, 2002). Available of replicative capacity does not appear to underlie sen- evidence does not suggest a causal relationship escence in humans or mammals. However, senescent between telomere attrition and biological senescence cells with traits similar to those of cells experiencing in humans (Aviv, 2002). One cause, oxidative stress, is replicative senescence do accumulate with time in a promoter of cellular senescence, thus telomere length human organs. also may indicate exposures to reactive oxygen species. Telomeres Apoptosis Cells of patients with syndromes mimicking some Senescence occurs at the molecular level (Arking, aspects of somatic senescence, e.g., ataxia telangiecta- 2006), but is expressed as somatic aging, cellular dys- sia, Werner’s syndrome, Bloom’s syndrome, trisomy function and death (Campisi, 2003). Senescent cells 21, dyskeratosis congetialia, aplastic anemia (Blasco, accumulate in most tissues with increasing age. This 2005), show progressive telomere shortening. Telo- may partly reflect sluggish apoptosis (programmed cell mere shortening also associates with heart failure, death). Tissues composed of dividing and nondividing immunosenecence, digestive tract atrophies, infertility, cells both accumulate senescent cells with age: reduced viability of stem cells, reduced angiogenic the former partly due to the Hayflick limit and loss potential, reduced wound-healing, and loss of body of telomeric DNA, the latter secondary to intrinsic

Human Longevity and Senescence 533 molecular stress, wear-and-tear, oxidative stress, and composed of short-lived, litter-baring, low PI species, other processes. Apoptosis is associated with the they do not show a similar proportional timing or pat- appearance of distinctive molecular markers. Among tern of LH to even the smallest of primates. Multiple LH most sexually reproducing organisms the force of NS features observed among insects and rodents are not continually decreases over the life span after matur- found in primate and humans (e.g., dauer stages, ation. It seems likely that over evolutionary history, litters). Such laboratory species tend to reveal basic alleles beneficially affecting cellular processes and processes of cellular senescence. What they do not illus- leading to greater cell stability would have evolved via trate is how senescence is expressed as aging and NS. However, increasing cell and organ dysfunction is longevity in humans. Humankind uses biocultural and observed with increasing age in most modern species. genetic adaptations, physiological and biological This appears to be due to intrinsic molecular alter- systems, behaviors and constructed environments in ations that disrupt cellular processes rather than to their responses to the age-independent, event-driven processes at the organ level that lead to systemic organ processes of senescence that occur over biological time. failure. In complex organisms, it seems that both apop- The same cascades of senescent physical and chemical tosis and cellular senescence may have evolved to events likely occur in all iteroparous sexually reprodu- suppress the transformation of functional cells to can- cing species (Arking, 2006). How evolution and pheno- cerous cells or cells that might harm the entire organ in typic adaptability have slowed these processes and risks other ways (e.g., autoimmunity, reduced OXPHOS, of death from altered gene expression, proteomic disin- mutant mtDNA). Multiple alleles influence apoptosis tegration, and disruptive metabolic events is what dif- (Bredesen, 2004) and create conditions that mimic ferentiate species-specific and population-specific LH. many of the effects of senescence seen in humans, In general, phylogeny is poorly associated with lon- e.g., Werner and Bloom syndromes, and other patholo- gevity (Rose, 1991). However, comparative biology has gies associated with shortened telomeres (Blasco, produced multiple insights on human senescence, and 2005). Continued apoptosis is one method for main- likely will continue to do so (Austad, 1997, 2003; Crews, taining the soma in response to continued cellular 2003). However, Austad (2003, p. 1327) points out that damage and deterioration. However, inappropriate “All of the best known and most widely employed apoptosis may limit life spans via antagonistic models for aging are short-lived taxa . . .” Although pleiotropy and loss of nonsenescent cells. Lack of commonly studied animals provide limited informa- apoptosis may allow senescent cells to survive and tion on how human LH, aging, and longevity are struc- deteriorate or compromise tissue and organ function tured (Austad, 2003), they do show that high “lifelong or become cancerous. Apoptosis may be delicately demographic heterogeneity” exists between and within balanced to eliminate most senescent and damaged species (Vaupel, 2001; Rose et al., 2005a, 2005b). Life- cells, but over the life span allows some to survive long demographic heterogeneity is measured by the and damage the organism. variability in mortality/survival patterns observed between cohorts and groups within populations; each requires a different Gompertz equation to fit its vari- ANIMAL MODELS AND able LH experiences (Carnes and Olshansky, 2001). COMPARATIVE BIOLOGY Experimental data from insects, rodents, and humans indicate broad LH variation within populations large Humans are complex organisms. They show several and small, and, by extension, different cohorts within LH attributes that seem to differ significantly from populations (Dobelhanner, 2000; Stearns et al., 2000; those of most other organisms (Austad, 1991; Hawkes Charlesworth, 2001; Arking et al., 2002; Kirkwood, et al., 1997, 1998; Alvarez, 2000; Smith and Bogin, 2005; Kirkwood et al., 2005; Rose et al., 2005a, 2005b; 2000; Carey and Judge, 2001; Bird and Bird, 2002; Perls, 2006). This illustrates the multiple influences of Crews, 2003; Crews and Gerber, 2003; Howell and environments and genes on LH. They also illustrate Pfeiffer, 2006; Larke and Crews, 2006; Christensen one pitfall of comparing demography, genes, or biology et al., 2006). One problem when applying results from across, within, and between species. Populations experimental animal models and molecular experi- within species inhabit a variety of local and regional ments to humans is that, thus far, only a limited environments (Rose et al., 2005b), producing multiple number of organisms have been studied (Austad, adaptive landscapes and variable adaptations. 1994, 1997). Yeasts, nematodes, fruit flies, and mice are the most frequently used animal models for study- ing longevity enhancement and senescent process. All Drosophila are “demonstrably inept at aging successfully” (Austad, Lacking a close evolutionarily relationship, fruit flies 2003; p. 1327). As mammals, mice share many homolo- do not recapitulate age-related declines and diseases gous gene loci with humans. However, as a clade observed among humans. However, these organisms

534 Douglas E. Crews and James A. Stewart show commonalities in their patterns of molecular and diet, mice with wild grandsires show changes in corti- cellular senescence, illustrating genetic and environ- costerone and testosterone that resemble those of mental influences on diminishing cellular functions standard laboratory animals, however their mean lon- with age (Arking et al., 2002; Grotewiel et al., 2005). gevity does not differ significantly from mice-eating Artificial selection produces populations showing standard laboratory chow (Harper et al., 2006b). variable longevity in multiple Drosophila species Calorie-restricted laboratory mice show higher mortal- (Luckinbill et al., 1984; Rose 1991; Pletcher and ity during early life, lower mortality in later life, and Curtsinger, 1998; Service, 2000; Charlesworth, 2001; have fewer cancers than ad libitum fed mice. Under- Arking et al., 2002; Rose et al. 2005a). Selecting long- standing how CR changes physiology and longevity of lived or late-reproducing fruit flies produces significant laboratory-bred mice may aid in understanding pro- increases in mean and maximum life spans after a few cesses of human senescence. Laboratory selection for generations. Given the controlled laboratory environ- either long-life or late reproduction in rodents extends ment, allelic variation must contribute to improved lon- average and maximum longevity. As with fruit flies, a gevity. As yet, directional artificial selection on longevity reservoir of longevity-enhancing alleles is exposed in has not shown an upper limit to life span (Curtsinger laboratory rodents. Wild rodents show a range of LH, et al., 1992; Arking et al., 2002). Among a single strain of body sizes, and longevities with the largest and longest- Drosophila, a variety of stressful external stimuli may lived inhabiting isolated islands and tropical settings. induce a variety of longevity phenotypes, suggesting Without major predators and when isolated from com- that multiple interrelated mechanisms modulate their petition, even rodents show longer lives and larger LH and life span (Arking et al., 2002). Combinations body sizes. This suggests that as populations develop of longevity-enhancing alleles also may recombine, mechanisms that reduce predation or competition predisposing latter generations to increasing life spans. for resources, isolation for rodents and culture for These results show that simple mechanisms influence humans, longer lives and increased body sizes ensue. longevity, and that life span is not intrinsically limited in most species. Current environments, evolutionary Nonhuman primates history, and multiple evolutionarily stable strategies Most primates show average life spans for their body across populations within species may produce inertial sizes above those predicted by regression models based barriers that limit life span. Genotypic variation for on extant mammals (Finch, 1990; Carey and Judge, survival, initial, and age-specific mortality rates is high 2001). Based upon longevity of 587 captive mammalian in fruit flies. However, specific DNA segments do not species, primates are not necessarily the longest-lived correlate highly with demographic and functional for their body sizes (Austad and Fischer, 2005). Even measures of senescence (Burger and Promislow, 2006). small- and medium-build primates tend toward longer Responses of young flies to early life stressors may alter life spans than predicted for mammals of their size. their homeobox genes rather than other DNA segments They also tend to have larger brains for their body sizes (Burger and Promislow, 2006), possibly modulating pre- than do most other mammals. Primates generally birth cursors of senescence during early life. single offspring, show high PI, arboreal adaptations, Drosophila species illustrate the relative simplicity and live in social groups, traits reducing susceptibility of modulating longevity in a laboratory setting (Arking to environmentally imposed mortality and predation et al., 2002). They also reveal a reservoir of genetic (Shea, 1998; Carey and Judge, 2001; Austad and variation for senescence-slowing alleles not expressed Fischer, 2005). Among prosimians, monkeys, and apes in wild populations. Senescence-slowing alleles are not multiple LH strategies are observed. However, being advantaged in wild populations because longevity long-lived and slow-reproducing mammals for their enhancement often is associated with reduced RS. In body sizes with large brains are common traits, sug- natural settings NS limits frequencies of “longevity- gesting that these reflect some similar evolutionary enhancing” alleles because they reduce RS. Unlike pressures that have influenced LH. Drosophila, humans, particularly men, may increase In general, primates progress through their LH their RS by living longer; women also may, but not as phases and transitions more slowly then similar-sized much as men (Marlowe, 2000; Crews, 2003, 2008). The nonprimate mammals (Austad, 1997; Shea, 1998; reservoir of longevity-enhancing alleles seen in many Carey and Judge, 2001; Leigh, 2004; Austad and wild-living populations already may be expressed as Fischer, 2005; Crews and Bogin, 2010). Most primates extended longevity in humans (Crews, 2003). lie above regressions lines predicting average life spans of mammals from average birthweight, adult Rodentia encephalization index (ratio of brain size to body size, Harper et al. (2006b) question the applicability of see Austad and Fischer, 2005), adult body size, or cra- research on laboratory-bred animals to human longev- nial capacity (Finch, 1990; Carey and Judge, 2001; ity (see also Austad, 2003). On a calorie-restricted (CR) Leigh, 2004). This is most obvious among large-bodied

Human Longevity and Senescence 535 4 apes. They may fall above such regression lines by a evolution” (Goodman and Leatherman, 1999; Carey standard deviation or more. Encephalization appears and Judge, 2001; Aiello and Wells, 2002; Crews, 2003, to characterize primates, along with extended off- 2005a, 2008; Crews and Gerber, 2003; Larke and spring dependency, high PI, and extended longevity. Crews, 2006; Arking, 2007). Physical development in This suite of LH characteristics has been identified as utero, physical maturity at birth, length and timing of a possible base for the modern human LH strategy postnatal growth and development, ages at menarche, (Shea, 1998; Bogin, 1999; Smith and Bogin, 2000; first and last reproduction, and menopause, reproduct- Crews, 2003; Crews and Bogin, 2010). ive effort, fertility, RS, rates of senescence, and longevity A high encephalization index is associated with continue to vary across the four remaining large-bodied greater behavioral adaptability. This allows organ- primate species and local populations (Finch, 1990; isms to use more self-motivated and opportunistic Carey and Judge, 2001; Crews, 2003; Leigh, 2004; behaviors when responding to their environments. Larke and Crews, 2006). Alterations in early LH par- Even early hominins had and modern apes show high ameters may alter all later aspects of LH, eventually encephalization. Encephalization among Homo pro- leading to changes in late-life survival and reproductive vided a basis for developing cultural activities leading patterns (Crews, 2003; Finch and Crimmins, 2004; to biocultural evolution. Primates generally birth a Crimmins and Finch, 2005, 2006; Finch, 2005; Larke single offspring. Although twining is the norm in and Crews, 2006). For example, within cohorts of some New World monkeys (marmosets and tam- modern populations, low childhood mortality and arins), it is rare in others (Cebids). Twinning among infectious disease morbidity tend to produce lower humans (about 3–5% of births) is relatively common. adult mortality and correlate with extended longevity Providing high PI to one offspring over an extended (Finch and Crimmins, 2004; Crimmins and Finch, period of time is energetically more efficient for parents 2005, 2006; Finch, 2005). This was predicted by than providing for two or more over a shorter period. Weismann (1881–1889) who surmised that life spans All primates share a suite of LH traits found across are inversely proportional to rates of maturation and long-lived species (Shea, 1998). Today humans express prereproductive mortality. These and other results sug- a high PI evolutionarily stable strategy (ESS) (Crews, gest that models based upon retention of reserve 2003, 2008; Larke and Crews, 2006). Human infants capacity (RC) into older ages and life in cushier environ- are among the most physically altricial of primate ments partly explain increases in later Homo and and mammalian offspring. Even so, humans show particularly modern human longevity (Crews, 2003, the shortest birth intervals, highest RS, and youngest 2005a; Caspari and Lee, 2004; Finch and Crimmins, weaning ages of extant large-bodied apes. These 2004; Crimmins and Finch, 2006; Larke and Crews, are LH traits correlated with fitness and evolutionary 2006; Bogin, 2009; Crews and Bogin, 2010). Such success. Since the Miocene, monkey genera have results do not support models suggesting that adult or been more successful than have apes at expanding postreproductive mortality rates influence patterns of their ranges and competing for survival (Leigh, 2004). growth and development. Rather the opposite, mortal- Monkeys, like humans, have shorter birth intervals, ity and infectious disease rates during the growth younger ages at weaning, and higher RS than do extant phase potentials lower adult and late-life mortality large-bodied apes. They also are smaller and show less and improved survival. postweaning PI and provisioning. Among the Homini- dae, humans show more monkey-like LH adaptations than do other large-bodied apes. Genes, evolution, and cell biology Biological, observational, and laboratory research have HUMAN LONGEVITY AND SENESCENCE advanced understandings of mechanisms by which many species (e.g., elephants, whales, sharks, tortoises, Foraging, scavenging, gathering, hunting, horti- humans) have extended their life spans past 50 years, culture, and eventually agriculture altered patterns while other species have not. However, specific genetic of LH, particularly survival, reproduction, and PI modulators and biomarkers have not shown strong among hominins. Changing subsistence strategies associations with longevity. Genetic variability and technologies altered energy availability and the between populations of individuals within species and human niche. Cultural developments and niche con- struction modified human environmental exposures 4 Biocultural evolution: The interplay of culture and biology on (Odling-Smee et al., 2003; Armelagos et al., 2005) human evolutionary processes. The scientific study of relation- producing differential longevity across populations ships between culture and human biology that have shaped modern human variation and the resources available to humans and groups (Crews, 2003, 2005a, 2008). Sociocultural across cultures. Only among humans are biocultural pressures adaptations by humans promoted “biocultural on evolution observed.

536 Douglas E. Crews and James A. Stewart across species in the organization of their DNA and lead to about 300 000–400 000 different protein forms genes hinders comparisons across humans and other in humans (Jaenisch and Bird, 2003). As understand- organisms (Arking et al., 2002; Hahn et al., 2007). For ings of cellular senescence expanded, mitochondrial example, chimpanzees and humans show relatively (mtDNA) came under closer scrutiny. Some senescent high population-level differences in their genetic struc- processes, e.g., loss of muscle mass (sarcopenia), are tures (Fischer, et al., 2006). Additionally, among associated directly with mtDNA mutations. However, humans, genetic variation is structured by cultural no single mutation is associated with a majority of behaviors. For example, there is little Y-chromosome senescent processes (Pak et al., 2003). A number of variation among members of the Cohen priesthood polymorphisms (e.g., mtDNA, APOE, HLA alleles/ (Hammer et al., 1997). Alleles at the BRCA 1 and 2 haplotypes, presenilins, BRCA 1 and 2 alleles) are and the apolipoprotein (APO) E loci, along with human associated with senescent-related diseases (e.g., sarco- leukocyte antigen (HLA) haplotypes show variable fre- penia, autoimmunity, dementias, cancer) that shorten quencies across populations. Some genes are associ- life span. Others, such as low levels of APOC3 are ated with early mortality, while others are found at associated with longer life spans (Atzmon et al., high frequencies among long-lived individuals of most 2006). A monotonic over-representation of the populations. For example, across samples from differ- APOC3–641CC genotype is observed at older ages in ent populations, different HLA haplotypes are associ- humans. Relatively infrequent (10%) at age 60 years, ated with longevity. Genetic variation within and this allele is found at 25% among survivors to 100 years across human samples limits any utility of comparing (Atzmon et al., 2006). APOC3–641CC predisposes to a alleles associated with longevity to nonhuman pri- lipid phenotype compatible with long-life in modern mates or other animals (Fischer et al., 2006). There settings. This association may not have occurred in are multiple human longevity genotypes and pheno- other populations at other times during hominin types. Local selection differentials, population size, evolution. Among some populations with low dietary genetic/phenotypic heterogeneity, and multiple genetic fat and cholesterol intakes, such as the Yanomami and and population processes structure correlations of LH other American Indian populations, a high frequency characteristics with specific alleles (Rose et al., 2005b). of the APOE*4 allele, apparently with no APOE*2 Population sizes, mating patterns, migration and sex alleles prior to European contact, suggest the APOE*4 distributions influence genetic variation in all species. allele may have been advantaged because it maintains Along with demographic, evolutionary, and environ- cholesterol levels (Crews et al., 1993). In such low mental factors, sociocultural behaviors have struc- cholesterol settings, APOE*4 alleles might have helped tured human genetic variability for LH. maintain cholesterol levels for generating steroid Humans appear to show some unique and many hormones during gestation, growth, development, similar LH traits compared to nonhuman primates and and reproductive effort. The HLA haplotypes and other long-lived animals. Humans are not the most APOE alleles show age-related frequency differences long-lived mammals, whales probably are. Verifying across populations. Different HLA haplotypes are asso- that a LH trait is unique to a species requires broad ciates with late-life survival in each population exam- comparative surveys, quantitative assessments, and ined. In modern Westernized populations, the APOE*4 measurements of variation within many different allele is associated with high lipid levels, predisposes to species. Uniqueness can not be assumed or dictated cardiovascular diseases in middle-age, and Alzheimer’s by arbitrary definitions. Sometimes researchers define dementia in old age. The APOE*4 allele shows a a trait in a manner that precludes it from applying to decreasing frequency with increasing age through at more than one species. Such definitional fiats hinder least the eighth decade of life. Although some associ- comparative research in LH evolution. In addition, ations are similar, multiple different alleles at single some LH associations observed in modern populations loci and variable haplotypes are associated with may not have been crafted by NS. Rather they may extended longevity in different human populations represent only current bioculturally constructed cor- in variable environments and cultures. Such a high relations. That is, evolutionary origins of some LH degree of variation defies attempts to identify species- traits may be unrelated to current patterns and distri- wide genetic modulators of senescence, although some butions, representing instead selective pressures on must exist. other traits or at previous points in human evolution. Associations with longevity suggest that many The human genome project shows that only about alleles may similarly predispose to long-term survival 23 000 segments of DNA code for proteins in humans as does APOC3–641CC. Besides specific alleles, (http://www.ornl.gov/sci/techresources/Human-Genome/ epigenetic regulation of gene expression affects mul- project/info.shtml). Subsequent research indicates tiple loci (Jaenisch and Bird, 2003). Epigenetic alter- that variable reading frames, pre- and post-translation ations affecting gene regulation occur during editing, and different lengths of final polypeptides embryonic and fetal development, but may continue

Human Longevity and Senescence 537 throughout life as cellular mitosis replicates imprinted 58 and females 53 years (without observable meno- loci (Barker, 1998; Cameron and Demerath, 2002; pause and an interbirth interval of 9.3 years). Survival Kuzawa et al. 2005). Multiple epigenetic factors influ- through 5–6 decades, about a 30–35 year female repro- ence health and life span. Associations of diet with ductive span, and multiple birth cycles is the basic LH cancers and the metabolic syndrome and of smoking pattern among not only living large-bodied apes, but with lung cancer are examples. All cells have multiple, many large-bodied mammals. Among men a 60-year- often overlapping, and highly conserved housekeeping plus reproductive span, highly variable RS, high pater- gene systems (e.g., albumin, ubiquitin, SOD, uric acid, nal PI, extended life spans, and survival past 7 decades glutathione, and other antioxidants) (Eisenberg and of life are not shared with other large-bodied primates, Levanon, 2003). Proteolytic systems, for example the nor are life spans of 100þ years and female survival to proteasome, are important for cellular housekeeping. twice their age of last menses. Along with altricial The proteasome degrades normal, misfolded, and offspring and high PI, these traits differentiate humans oxidized proteins recycling their amino acids. It also from other hominins and hominoids. degrades old proteins that accumulate in cells with Others have identified suites of unique human LH age. Both in vitro and in vivo, senescent alterations characteristics (Shea, 1998; Smith and Bogin, 2000; are correlated with impaired proteasome function Carey and Judge, 2001; Crews, 2003). Humans are (Chondrogianni and Gonos, 2005). Proteasome generalists with some LH traits linking them to other function regulates cellular homeostasis, helping to mammals, nonhuman primates, and extant and extinct determine cellular senescence along with hundreds of hominins and hominoids. Evolutionary and allometric other loci. constraints on the size of ovaries and number of avail- able primary oocytes may lead to reproductive decline with increasing age among all large-bodied mammals Uniquely human life history traits (Jones, 1975; Graham et al., 1979; Austad, 1994; Leidy, Humans, like a number of other mammalian and 1994; Packer et al., 1998). Although some have nonmammalian species show long lives. Comparisons attempted to make a special case for human reproduct- with other long-lived species are one avenue for deter- ive decline and menopause by suggesting it is uniquely mining unique and conserved aspects of human LH human (Pavelka et al., 1991; Turke, 1997; Hawkes (Austad, 1997; Carey and Judge, 2001; Crews, 2003; et al., 1997, 1998), oocyte depletion over the life course Austad and Fischer, 2005). Concepts such as a appears universal among female mammals (Graham minimum necessary life span (MNLS: the amount of et al 1979; Leidy, 1994; Austad, 1997; Packer et al., time members of a species must survive to complete 1998). Austad (1994, 1997) describes reproductive the “necessary tasks of life” as described by Weismann decline with increasing age among mammalian [1881–1889]: gestation, growth, development, matur- females in such a fashion that one must conclude it is ation, and reproductive effort), PI, and the costs of a pleisomorphic LH trait (see also Packer et al., 1998). reproduction help focus this research. Elsewhere, we The main reason few female mammals show meno- estimated a MNLS of about 45 years for humans and pause as currently defined is that few commonly sur- about 35–42 years for chimps and gorillas (Crews, vive past their fifth decade of life, at which age almost 2003; Larke and Crews, 2006). Based upon Weismann’s all women are postmenopausal (see Figure 3.5 in (1881–1889) concepts, a MNLS is similar to a Crews, 2003, p. 114). Darwinian necessary life span. Similar to oocyte depletion, gestation lengths of One woman already has lived three times the esti- about 9 months and completion of 95% brain growth mated human MNLS (about 122.5 years). As yet, no by about 5–6 years are shared among living large- chimp or gorilla has doubled their estimated MNLS. bodied apes. These commonalities suggest similar trait However, the chimp who played Cheeta in the 1950s states may have characterized earlier hominins and Tarzan films celebrated his 75th birthday on April 9, hominoids, perhaps representing basic large-bodied 2007. Among other large-bodied apes, on December 22, ape characteristics. Gestation length may be a symple- 2006, the world’s first captive-born gorilla (Colo) cele- siomorphic or synapomorphic LH trait within the brated her 50th birthday at the Columbus Zoo (Colum- large-bodied ape clade. Elephants show longer gesta- bus Dispatch, Ohio); on January 17, 2009 the oldest tion lengths, while marsupials have utilized shorter male gorilla in a North American zoo (Timmy) cele- gestation lengths and pouches for extended infant care. brated his 50th birthday at the Louisville Zoo (Colum- Such examples show that gestation length is an evolu- bus Dispatch, Ohio); the world’s oldest captive tionarily malleable trait. Both human and marsupial orangutan (Nonja) died at age 55 in December 2007 newborns are altricial. Evolving hominins and marsu- at the Miami Metropolitan Zoo. Based upon estimated pials may have faced a similar problem – gestating ages in a wild population observed over 32 years, Wich neonates in a limited amount of time – and hit upon a et al. (2004) suggest orangutan males may survive similar solution, altricial newborns. Although their

538 Douglas E. Crews and James A. Stewart developmental processes are distinctive, their solution, subsistence strategies practiced by each today likely high PI postbirthing, is the same. Sociality and food reflect the unique and marginal subsistence strategies sharing characterize many animal species. However, adopted by each in the late Miocene (Leigh, 2004). they are not associated with extended life spans in Rounding out the suite of coadapted human LH most. Extended life spans and multigeneration social traits are provisioning and care of nonkin, constant groups are shared by humans, whales, elephants, food-sharing, and shared childcare (Crews, 2003, chimps, gorillas, lions, and other social animals. 2005a; Crews and Gerber, 2003; Bogin 2009; Crews Neither is uniquely human, both are pleisomorphic and Bogin, 2010). Modern-day humans are not exact and synapomorphic mammalian LH traits. copies of earlier hominins and simple linear models Humans invest early in life in their neurological based upon earlier forms poorly fit LH traits of extant development by slowing their somatic development in primates, as do models based upon extant human for- utero and producing secondarily altricial offspring agers poorly fit our ancient ancestors (Crews 2003, (Smith, 1989). Born less physically developed, com- 2005a; Crews and Gerber, 2003). Earlier hominins did pared to other large-bodied apes, humans require an not possess the current suite of human biocultural extended growth period. Investments in neurological adaptations or our highly constructed niches and tissues in utero and the first five years of life delay built environments. Additionally, the four extant physical development. Altricial newborns, early neuro- large-bodied ape species are not an ideal comparative logical development, and slow physical development sample of all large-bodied primates that ever existed require late attainment of reproductive maturity com- for extrapolating Hominin evolution. They are just pared to more rapidly developing large-bodied apes the ones who managed to survive into the present due (Crews, 2003; Crews and Bogin, 2010). They also neces- to their adoption of unique subsistence strategies and sitate extensive postweaning PI, thereby extending the LH patterns (Leigh, 2004 presents some possible human MNLS. Slow developing secondarily altricial reasons). All four have followed separate evolutionary offspring with early neurological development are pathways since bipedal apes formed their own lineage uniquely human, an autapomorphic trait (Smith, some 7 mya. 1989; Peccei, 1995; Crews, 2003; Crews and Bogin, 2010). Secondarily altricial offspring are not seen in Human life span any other primate genus or species. These secondarily altricial offspring are the basis of human LH evolution Based upon recovered burials and bioarchaelogical and likely evolved with bipedalism and culture. Thus, reconstructions, Neolithic agriculturalists averaged they require their own evolutionary explanation per- about 25 years of life. However, nineteenth-century haps as a response to early bipeds’ limited pelvic width. industrialists could expect to survive about 45 years. Freeloading altricial offspring are dependent on their In some modern populations, women average 85 years parents’ abilities to provide continual PI over 15–20 of life, while men approach 80 years (Crews, 2005b). years to offspring. Although PI occurs in all primates, Based upon estimated ages and survival patterns its expansion to 15–20þ years is uniquely human. observed in living foraging/hunting populations, some Modern humans also show adaptations in their gut suggest that once people survive childhood and their morphology that differ from other large-bodied apes. second decade of life (to age 20 years) in such settings, Other apes have their greatest gut volume in their they are likely to live into their 5–6th decade of life colons, 45%, with only 14–29% small intestine (Milton, (40–50 years) (Hawkes et al., 1997, 1998; Carey and 1999). In humans, 56% of the total gut is small intes- Judge, 2001). Since the late Paleolithic, it is likely that tine, while the colon accounts for only 17–23% (Milton, in most human groups some individuals may have sur- 1999). Based upon gut proportions, large-brained vived to complete 50 years of life or more. Nevertheless, hominins adapted to high calorie diets digested mainly it is unlikely that many survived to 70 or 80 years; ages in the small intestine. Most other large-bodied apes that few survive to today in many populations. Among favor low-energy foods digested mainly in the colon. preagricultural groups of foragers, gathers, and hunters Humans show greater plasticity in diet than other of the Paleolithic. Among preagricultural groups of large-bodied apes. Their morphology supports the foragers, gathers, and hunters of the Paleolithic with high-energy diet model of hominin evolution (Leonard, small local populations (N  500), it is not probable that 2002; Aiello and Wells, 2002), as do isotopic analyses of many survived even to their 50s. Estimates of survival human remains (Sponheimer and Lee-Thorp, 1999). probabilities based upon estimated birth and death Early Homo and later humans likely relied upon a rates among today’s foraging populations are not variety of protein-rich foods as our modern LH pattern reliable. Actual birthdates seldom are known and data developed; likely including easily obtained marine for constructing stable life tables are not available resources (Kingdon, 1993). Only four great ape species for recent generations. Multiple external factors influ- survived the Miocene ape extinction. Different ence such estimates. In contemporary cosmopolitan

Human Longevity and Senescence 539 settings, life expectancy exceeds 85 years among women entered into multiple regression models with age at in only a couple of settings (Sweden and Japan) (Crews, first birth and age at mean childbearing, wealth’s effect 2005b; Larke and Crews, 2006). Some expect that max- on life expectancy reduces below statistical signifi- imum life span will achieve 130–150 years by the mid cance. These results support suggestions by Crews twenty-second century. A total wager of $300 between (2003, 2008; see also Larke and Crews, 2006; Bogin, two well-known gerontological researchers, Steven Aus- 2009; Crews and Bogin, 2010) that later reproduction tad and S. Jay Olshansky, in 2000, was widely reported. leads to increased RC and, ultimately, greater longev- Austad expects someone alive today to survive over 14 ity. Across nations, delayed reproduction is associated decades of life. If someone does, in 2150 his heirs will with longer life spans, suggesting that accumulating receive the original wager and accumulated interest; if resources before reproducing produces greater RC. not, Olshansky’s descendents win. This bet is significant High correlations of per capita income and mean not only for what it says about predicted life expectancy age at first birth with life expectancy suggest that cul- in the future, but also for what it implies about PIs in tural mechanisms provide opportunities to accumulate future descendents by provisioning them via social RC and increase women’s reproductive opportunities structures current in modern societies. In a recent at later ages. More benign, culturally constructed report, Vaupel and colleagues (Christensen et al., 2009) environments lead to greater longevity by providing suggest that if current trends continue, those born since women opportunities to increase their RC and their 2000 in cosmopolitan settings may achieve an expect- RS at later ages. ation of life of 100 years. Benign/cushy environments Reproduction and life span Human niche construction has created built environ- Across human populations, the number of offspring ments for humans to reside and reproduce in for over successfully birthed and reared to reproductive age 200 000 years (Odling-Smee et al., 2003; Crews, 2007, varies positively, negatively, and not at all with longev- 2008; Crews and Bogin, 2010). Built environments ity of women (Larke and Crews, 2006). Confounding allow people to live longer than ever before. Today’s arises partly because both early and late fecundity are constructed environments allow humans to retain associated with longevity, while total parity correlates their high functioning stress responses, immune, and poorly with longevity (Larke and Crews, 2006). Vari- housekeeping systems well into their later years. A less able associations of RS with longevity in women indi- positive result of constructed environments is that cate multiple confounding factors (Larke and Crews, children in modern environments have high asthma 2006). For one, RS is determined less by fecundity, and and allergy rates. Exposures to pathogens and allergens more by the quality/quantity of PI a women has to among individuals growing and developing in less- invest in offspring. Human PI is more complex, exten- constructed environments may protect children sive, and longer lasting (both relatively and absolutely) from such aliments. Modern urban slums may be than for any other mammal. This supports a model “dirtier” and produce more infections and parasitic that human LH coevolved with altricial offspring exposures than traditional farms and tropical settings. and their need for extensive long-term PI. Shaped In cosmopolitan settings, less exposure to childhood by biocultural factors, this need influenced all of diseases produces fewer immune responses and less human physiology, LH variation, reproduction, senes- inflammation, contributing to retention of RC and cence, and patterns of longevity (see also Crews and extended life spans (Crews, 2003, 2007, 2008; Drenos Bogin, 2010). and Kirkwood, 2004; Finch and Crimmins 2004; Analyzing wealth, longevity, reproductive, and Crimmins and Finch, 2005, 2006; Finch, 2005; Arking, other LH variables, Larke and Crews (2006) showed 2006; Drenos et al., 2006). that resources and the timing of reproductive events Well-nourished humans living in built environ- significantly influence life expectancy. Wealth and age ments experience more benign living conditions than at menopause correlated, while mean age at childbear- did their evolutionary forbearers. Laboratory, domesti- ing did not significantly correlate, with life span. When cated, some island, and provisioned animal popula- entered jointly in a regression model with age at first tions live better than their wild-living counterparts birth and wealth, mean age at childbearing signifi- and less isolated populations. Benign settings provided cantly positively predicts life expectancy. Colinearity opportunities for slowing senescence and enhancing with mean age at first birth partly explains reports that longevity not available in more stressful environments. mean age at childbearing is associated with longevity. Large increases in life span should not occur in high Across national populations, the strongest correlates stress settings (Crews and Bogin, 2010). In such cir- with life expectancy are wealth and age at first birth cumstances, organisms are frequently at or close to (see Figures 2 and 4 in Larke and Crews, 2006). When their limits of survival, metabolic efficiency is at a

540 Douglas E. Crews and James A. Stewart premium, and calories are best invested in reproduct- mortality and increased longevity. Inflammation is ive effort (Parsons, 2002, 2003). Multiple genes for described as the “fire of frailty” (Walston and Fried, cellular stress resistance, some predisposing to high 1999; Fried et al., 2004; Walston, 2005). Reducing life- metabolic efficiency, enabling adaptation to energy- long inflammatory responses reduces secondary som- poor environments, were adapted to harsher environ- atic damage, thereby improving RC and stamina. ments with different selective pressures than prevail Caruso et al. (2005) suggest individuals genetically pre- today (Parsons, 1996, 2002, 2003; Crews, 2003, 2008; disposed to weak inflammatory activity have a better Arking, 2006; Drenos et al., 2006). In natural-living chance of living longer, provided infectious diseases do populations, high vitality, stress resistance, and resili- not kill them. Variable alleles at multiple cytokine loci ence enhance RC and survival, and continued survival influence the type and intensity of immune–inflamma- depends upon substantial investments in allostatic tory responses, thereby affecting individual life expect- responses. Genes for stress resistance, immune compe- ancy. At least one proinflammatory protein (APOE4) is tence, and energetic and metabolic efficiency select- associated with earlier mortality among middle-aged ively advantaged in environments of evolutionary and older adults. It appears that cushier environments adaptation underlie modern human allostatic pro- with fewer pathogens and genetic predispositions cesses. Today LH and longevity vary with ecological to weaker inflammatory activity jointly predispose to circumstances. As culturally based niche construction longer life. created cushy built environments, extended life spans Cohort morbidity phenotypes also may explain sea- evolved in hominins. Similarly, only after defenses sonal variation in mortality (Finch and Crimmins, against predation (i.e., large body size, body armor) 2004) in agricultural populations. Agricultural diets were established did extended life spans appear in have ratios of omega-6 to omega-3 fatty acids (2:1) that whales, tortoises/turtles, sharks, and elephants. differ from the diet proposed for ancient hominins Older humans express their innate inclinations (10:1) (FAOUN/WHO, 1980). An altered omega-6 toward slower senescence and later mortality when to omega-3 fatty acid ratio from agricultural diets living in modern-day cushy environments (Perls 2001, may produce nutritional inflammation. Conversely, 2006; Crews, 2005a, 2007, 2008; Drenos and Kirkwood, by increasing access to fresh fruits and vegetables 2004; Drenos et al., 2006). A “late-life” phase of LH throughout the year, agriculture may have contributed following the senescent stage may occur among some to reduced stress from ROS and deficiency diseases species, laboratory organisms, and very old humans, in and slowed some senescent processes. cushy/benign environments (Rose et al., 2005a; Drenos et al., 2006; Rauser et al., 2006). Social inequality likely increased as the Neolithic revolution led to settled agri- DISCUSSION culture and large-scale niche construction. Not all par- ticipate equally in benefits from cushy environments. Humans, like all living beings, resulted from evolution- Inequalities within cosmopolitan societies and ary processes by which radiations of earlier taxonomic between them and traditional ones contributed to the groups created new ones. As the kingdom Animalea rapid spread of the third epidemiological transition arose from earlier precursors, basic aspects of cellular worldwide (Armelagos et al., 2005). Variable disease function were highly conserved through successive patterns and life spans even among individuals living radiations. The same evolutionary forces shaped rock- in the same environment reflect underlying demo- eating lichens, rock lobsters, rock-tool makers, and graphic and physiological heterogeneity. rocket scientists. They also lead to social organisms, nest builders, niche constructors, food sharers, nut- crackers, fire starters, and both microbial and human Frailty and inflammation wine and cheese makers. Evolution produced multiple As cushier environments developed over recent human species dependent upon complex patterns of food history, reduced disease exposures were a cohort phe- acquisition – ants, bees, chimpanzees, humans, lions, nomenon (Finch and Crimmins 2004; Crews, 2005b). wild dogs – using the same basic DNA and cellular Swedish demographic data from 1751 to the present structures. This is why animal models and their rates show that low early life mortality rates are associated of senescence reveal biomolecular strata occurring with low adult mortality and greater longevity of the within human cells and somas that lead to biological same cohort (Finch and Crimmins, 2004). This may be senescence. a general population phenomenon (Crimmins and Short-lived laboratory organisms are incompetent Finch, 2006). Finch and Crimmins (2004; Crimmins at successfully aging (Austad, 2003). Still, their DNA, and Finch, 2005, 2006) suggest reduced lifetime expos- proteins, macromolecular structures, and processes of ures to infectious diseases and inflammation contrib- cellular degradation are similar to those of most organ- uted to large-scale historical declines in old-age isms. Drosphilia show broad additive and nonadditive