Summary Notes: Diet and Nutrition Workshop

Workshop Rapporteurs: Jay T Stock (U Cambridge) Claudia R Valeggia (U Pennsylvania)

Workshop Leader: William R Leonard (Northwestern)

One of five workshops in a conference on
Evolution and Diseases of Modern Environments
Organized by Randolph Nesse, at the Berlin Charité October 13-14, 2009
In conjunction with The World Health Summit
Sponsored by the Volkswagen Foundation

The following provides a brief summary of the discussions of the Diet and Nutrition Workshop at the Symposium “Evolution and Diseases of Modern Environments”.  Discussions were broadly based around two themes: (1) diet and nutrition in earlier human evolution, and (2) recent human evolution, dietary adaptation and the origins of “diseases of the modern world”.  A summary of these is provided below, followed by a discussion of the relevance of dietary trends in human evolution to understanding the etiology of diseases of the modern world, some general points of agreement between the participants, and proposed areas of future research.

Theme I: Diet in Earlier Human Evolution

The group was in broad agreement that during much of human evolutionary history there has been selection for increased dietary flexibility.  There is evidence for occasional dietary specialization in hominin evolution (e.g., the Neandertals, being top-level carnivores based on isotope data); however, much of our dietary evolution suggests a broad range of food items.

The importance of so-called “fallback foods” was also widely discussed. Peter Ungar and Matt Sponheimer noted that these are secondary, low ranking food resources that are often relied on during periods of seasonal stress.  There was wide agreement that it is these lower-ranking (often poorer quality) resources that most strongly shape nutritional ‘adaptations’ (i.e., dietary minima) over evolutionary time, rather than the “average dietary conditions”.

In early hominin evolution, there is evidence for exploitation of tough foods, producing rapid tooth wear and selection for large molar tooth size, thick enamel and craniofacial robusticity.  Recent chemical and microwear analyses suggest that the australopithecines were exploiting more cereals / grasses / sedges than previously thought.  The emergence of early Homo, is associated with a reduction in mandibular size, robusticity, tooth enamel thickness, and buttressing of mastication. These changes suggest a shift in food use patterns and/or food processing.  Early Homo specimens in different parts of the old world are variable in body size, raising questions about when human body size and proportions fully emerged.  Many interpretations of the greater importance of animal foods with Homo are due to the ubiquity of stone tools and animal bones; however, more fine-grained analyses (e.g. bone chemistry) are required to provide clear evidence of a dietary shift to include more animal foods.  Such resolution may be possible in the next 5 years, according to the paleontologists in the group. In addition to meeting nutritional needs, meat/animal foods also may have been associated with assisting in the colonization of new environments for hominins, by allowing them to converge on a common dietary niche under different environmental conditions.

By the later stages of hominin evolution, we have compelling evidence of high levels of meat consumption in Neandertals, based on carbon/nitrogen isotopic analyses of bone collagen. In contrast to the Neandertals, Upper Paleolithic humans appeared to forage on a broader range of plant and animal resources, that varied widely based on local ecology.

With the origins of agriculture, there was a shift again to more extensive plant exploitation, including grasses – cereals / starch.  There was some disagreement within the group about the extent to which humans have adapted/adjusted to the exploitation of cereal grains.

Overall, the shift from a ‘paleolithic’ to modern diet was clearly characterized by a reduction in protein content of the diet and a marked increase in carbohydrates, particularly simple carbohydrates (sugars). Many have also noted that within the last 50 years, there has been a particularly rapid increase in the amount of processed sugar added to the modern American diet. Whether this increase is largely responsible for explaining the dramatic increase in obesity rates in the US since the 1960s remains a point of debate.

Theme II:  Recent Human Evolution & Dietary Adaptation

Genetic adaptation to human dietary regimes. A number of examples (and evidence for) genetic adaptations to distinctive human dietary practices were discussed. These included:

Lactase persistence: There was clear consensus that the ability to digest lactose past the weaning age is certainly a product of natural selection. However, there were differences of opinion and broad discussion on the nature of selection for lactase persistence.  Mark Thomas argued that selection is episodic, acting mostly during the bad times and not so much during the good times.  He proposed that consumption of milk and dairy products act as “buffering” for the bad times during farming lows.  When the going got tough, lactase persistence allowed farming people to drink milk without getting diarrhea.

Lactase persistence appears to have originated in central Europe (Balkans) but, because of demographic changes associated with farming/dairy mixed techniques, its distribution gets shifted up North (e.g., Scandinavia). This is consistent with the archaeological evidence of the origin and spread of  Linear Band Keramik [LBK] pottery in the European Neolitic.  An additional point noted by Mark Thomas is that the spread the lactase persistence allele appears to be driven by demographic factors – that is population growth/expansion – rather than by differential changes in food use over geographic space.

In addition to the cultural hypothesis for the selection of lactase persistence, Loren Cordain presented a novel hypothesis on the possible role of lactase having an indirect effect on resistance to malaria.  The malaria plasmodium needs to utilize the host’s PABA (para-aminobutiric acid) to survive.  Milk is very PABA-deficient, therefore a diet rich in dairy products may favor resistance to malaria.  Lactase is preserved as a way to promote consumption of milk (which would protect the person) into adulthood.

Copies of amylase gene. The number of copies of the amylase gene has been offered as another example of human genetic adaptation to distinct foods (either wild or domesticated foods with high starch content, in this case). There was some disagreement as to whether this represents a valid example. The reported association between the number of copies of the amylase gene and starch content in the diet is not significant.  There is a trend, but statistically the results are not robust.  The question here is why is variation in number of the amylase gene copies maintained?  Starch consumption could possibly be one reason, but starch is mostly digested in the intestine, not in the mouth (and this example is for salivary amylase).

Alleles for the Alanine-glyoxylate aminotransferase (AGT): This enzyme targets 2 different intracellular organelles.  In carnivorous species, AGT targets mainly mitochondria; in herbivores, it targets the the peroxisomes.  Humans, being omnivorous, show polymorphisms that seem to correlate with the relative proportion of plant or meat in the diet. Some evidence to suggest that populations with greater recent ancestry of meat eating (eg. the Saami) have higher frequencies of the allele favoring the “retargeting” of enzymatic activity to the mitochondria.

Detoxification genes (Cytochrome P450). These include genes that code for enzymes with role in a wide range of metabolic processes. They are highly variable and have been hypothesized to be associated with diet breadth in human diets.

Tolerance to alcohol. Variation in the ability to metabolize alcohol has been proposed as another interesting case for genetic adaptation; however, there was not much consensus about whether it represented a good example. Part of the problem here surrounds the fact that the “primitive [ancestral] condition” for humans appears to be the ability to metabolize alcohol, with various groups [eg. Asian populations] then losing that ability. Hence, unlike the “lactose/milk use” model, there is not a clear “gene-culture” story with the ability to metabolize ethanol.
Specifically, it is hard to develop a compelling adaptive argument for losing the ability to metabolize alcohol in certain populations, but not others. Is drunkenness, for example, considered such a negative social behavior that low tolerance was favored (selected for)?  There was not much enthusiasm about this hypothesis.

Uricase. In human and great apes it appears there has been selection for genetic reduction in the activity of uricase, the enzyme that metabolizes uric acid (a byproduct of meat digestion).  Some evidence to suggest that reduction in uricase activity was selected for under conditions of seasonal variation in food availability that characterized environments for our hominin ancestors.  It was pointed out that humans have high uric acid concentrations, but they also have a higher flux of metabolism.

Skin color/Vitamin D Metabolism.  Robert Perlman posed the question of whether genetic differences in skin color (& implications for vitamin D metabolism) likely re\presented a ‘dietary adaptation’. A discussion of the group ensued about this about the association between skin color and vitamin D.  Lighter pigmentation and lactase persistence go together.  Are they selected for together? There was some disagreement on whether higher latitudes get enough vitamin D.

Origin and nature of “Diseases of the Modern World”.

Finally, we discussed the nature of  “diseases of the modern world” and the role the dietary factors play in their origins.  There was wide agreement that the process of lifestyle and dietary ‘modernization’ does not produce the same health effects in all human populations. Rather, there seems to be remarkable disparities across different ethnic groups.  Some populations respond with increases in the incidence of diabetes and metabolic syndrome in general (e.g., the Pima of North America), where others (e.g., Siberian populations) show low incidence of diabetes with westernization of diets, but show high incidence of hypertension.

We felt that understanding the nature of this variation is an important task for researchers in human nutritional evolution.  Key unanswered questions include: What are the determinants of these differences in responses to diet changes?  Is this due to genetic differences, social differences, epigenetic effects or a combination of all of these?

The first example of possible determinants was variation due to ancestry among women regarding fat oxidation: Caucasian women have better oxidative responses than African American.  Discussion then ensued about differences in fat deposition patterns.  Having fat in the wrong place seems to be what is causing health problems, Andreas Pfeiffer commented that some obese women do not develop metabolic syndrome and they are the ones that do not have large abdominal fat deposits.  Low abdominal fat is correlated with better health outcomes, however Jonathan Wells pointed out that a certain amount of abdominal fat is essential for immune function.  Fat has macrophages and may be important to fight disease in high disease-load populations.  This can be what drives differences in fat deposition patterns across different populations.

Mike Power pointed out that adipose tissue, in addition to act as energy storage, contains enzymes that transform steroid hormones, which in turn affects reproduction and that has crucial fitness implications. This recent recognition of fat as more than simply a storage organ (playing important roles in both endocrine and immune regulation) has led us to see fat as playing a more dynamic role in shaping physiology and the origin of nutrition-related diseases.  An important question identified about the role of fat in influence human health was: (a) how do such factors as sex differences, reproductive strategies, disease loads and ethnicity (biological and social dimensions) shape the variation in fat storage and the activity of fat?

There was agreement on the general statement that energy balance seems to be the key, i.e. too many calories and too little physical activity puts the body in a state of positive energy balance that drives the body metabolism to store the surplus of energy as fat.  In fact, the amount of calories consumed does not seem to have changed much in the last 40 or so years, but the physical activity levels are dramatically low.

However, a few people in the group stressed the importance of differences in the quality of diet as also having a central role in the epidemics of obesity.  Jonathan Wells proposed that the obesity epidemics in the US started with the American Heart Association promoting a low fat diet to prevent risks of cardiovascular disease, pushing people to eat a high carbohydrate diet.  In addition, the processing of sugars, in particular sucrose, in food manufacturing also changed the glycemic indices and the energy availability of diet items for the public.

Discussion went back to energy balance and the difficulty of measuring physical activity levels in different populations, as noted by Josh Snodgrass.

On an applied perspective, what is the best exercise regime for losing weight? Bill Leonard discussed the utility of looking at energy expenditure & workloads in traditional societies as a window onto both the amount of activity/energy expenditure in earlier human societies and the pacing of daily activities (i.e., The “tortoise vs. hare” phenomenon). Traditionally societies often display a slow and steady pace (as in walking long distances frequently) in maintained sustained increased in metabolic rate. Work in nutrition suggests that this strategy works better in promoting energy balance and weight stability in the modern world. Andreas Pfeiffer noted, in contrast, that high intensity exercise is good to develop muscle mass and increase BMR, but sustained exercise burns fat.

There was agreement in that human populations experienced a major shift in macronutrient composition during modern times (what is modern?). We are now eating much more processed carbohydrates than before. This has the potential to skew things towards high prevalence of obesity.

General points of agreement across the discussions were:

There is no ‘ideal’ Palaeolithic diet for humans.  Dietary recommendations should include a knowledge of the evolution of the human diet and recent adaptations, but rather than there being an‘optimal’ human diet, there are a range of adequate diets which depend upon individual biological and cultural variation.

There is considerable variation between populations, which are based upon genetics, developmental and epigenetic factors, and cultural context.  It is the combination of these factors that interact with modern diets to lead to different health outcomes with the dietary transition.  Proposed human diets should be targeted to both biological and cultural factors, collectively considered as ethnicity.

We should not consider caloric content of diets in isolation of total lifestyle, both factors are important in combating modern metabolic diseases.

Epigenetics and early development are likely to be key factors in the origin of modern metabolic diseases.  Further research is needed on the relationship between developmental factors and adult health outcomes. Additionally, interventions are best targeted in early life.

Questions that remain unanswered include: What explains the variation in obesity levels across and within populations? Particularly, what are the relative roles of  (a) genetics, (b) epigenetic and developmental factors, and (c) social/political/economic factors.   Clarification of these issues should be the focus of future research.

Explaining population variation in changes in nutritional health with modernization

Does industrialization largely have the same impact regardless of the setting? Or do we see particular clustering in different populations in transition?  There is mounting evidence coming from different settings that although overweight and obesity tend to increase dramatically with a change in diet and lifestyle, the impact that this has on health varies both within and between populations. That is, not everybody is experiencing and responding to nutritional and lifestyle transitions in the same way.

The observed ethnic variation in how the nutritional transition impact health arises from genetic differences, environmental differences, and the combination of both.  Strict genetic control seems to be unlikely in the majority of obesity cases.  In fact, genetic differences can explain only about 5% of the variation in morbid obesity. Genes may explain the disease, but cannot explain the development of the disease in a particular person.  Furthermore, it is clear that the metabolic response to increased energy availability is a polygenic phenomenon: there are lots of genes with small effects involved.  For example, at least 20 genes have been identified as being related to diabetes, but each of them explains very little of the variation in the development of diabetes.  Epigenetic effects are the most likely culprits.

Environmental factors are contributing relatively more to the equation. Given the pivotal role of nutrition during early development, a life course approach to the study of this variation generally yields a clearer picture of the possible determinants of population and individual differences.

There are certain “life course experience” variables that have been shown to be associated with adult health, particularly with risk for obesity: among others, birth weight (as a proxy of prenatal nutrition), pattern of infant and childhood growth, amount of exercise, and infectious disease burden.

Our group went on to discuss how our understanding of population variation in responses to the nutritional transition can be applied to concrete medical advice[i].  Ethnicity was mentioned as an important determinant, but what do we mean by ethnicity?.  It was agreed that ethnicity, by all accounts, should be interpreted as individual/personal history, which includes your genetic background, your culture, and your life experience.  A given nutritional status is, therefore, the embodiment of all these variables.  Cultural practices, particularly those related to food preferences and taboos, were mentioned as crucial elements in the design of recommended healthy diets for individuals, as they are intimately related to compliance issues.  Food is indeed identity and cultural values must be incorporated.  In sum, there is no single paleodiet that will be the magic bullet. Rather, individual diets need to be tailored to recent past history and individual life course experiences.

A life course approach is paramount.  Nutrition has the highest “programming” impact when human growth is most plastic: during prenatal growth and early postnatal growth. Thus the importance of maternal nutrition during pregnancy and of infant feeding practices.  Growth restriction in utero coupled with rapid catch up growth has been strongly associated with higher risk of cardiovascular disease.  Breastfeeding has been shown to have programming effects as well. Breastfed infants develop better insulin management mechanisms than bottle-fed ones. The immunological effects of breastfeeding are as important as the nutritional ones for explaining the long term developmental consequences.

What aspects of the evolutionary perspective are particularly important to incorporate in mainstream western medicine and public health?

The concepts of trade-offs and of a life course approach are key for reorganizing the medical paradigm.

Ancestry, understood as a result of biocultural forces.

Nutriomics and nutrigenetics (Andreas Pfeiffer). However, most in the group said that developmental history is more important than knowing the genetics of an individual.

What research can be emphasized to reinforce those aspects?

Studies of migrant populations in which parents, children and grandchildren have different developmental histories, particularly with respect to nutritional availability and disease load.

Studies of populations in transition in different settings (circumpolar, tropical, subtropical) and with different original subsistence practices (hunter-gatherers, agriculturalist, pastoralists).

Studies that look closely at infant feeding practices and weaning food choices (both in present populations and in historical and pre-historical ones via bioarchaeological analysis).


[i] We called this “Richards’ conundrum” in reference to a question posed by Michael Richards during our discussion. Michael will give a presentation about the evolution of diet to an audience that includes First Nations communities.  He wondered whether the findings about ethnic differences in the response to dietary shifts can allow us to safely say to an First Nations individual that, because of his/her ethnicity, he/she had certain higher risks for nutritional problems.