The Global Distribution of Type Two Diabetes Risk Alleles (from Chen, et al. (2012)
Type Two Diabetes And Genotypes
Diabetes is a disease characterized by the inability of the body to use insulin to properly manage blood glucose (i.e. blood sugar) levels, primarily associated with pancreas function, although some new chemical pathways that play a part in the conditions have been discovered. It can be managed with insulin shots and careful restriction of sugar in one's diet (or foods that quickly metabolize to sugar), and in worst case scenarios when not managed well, can lead to diabetes shock and comas, to poor circulation leading to loss of limb function or even limbs themselves, and to kidney failure that must be treated with dialysis (i.e. having a machine mechanically treat your blood in the way that internal organs should, typically for many hours several times a week). Mismanaged diabetes is deadly.
There are two main kinds of diabetes. Type one diabetes is associated with poor insulin production that most often manifests early childhood and is essentially treatable but incurable, although scientists keep trying. Type two diabetes is the adult onset form that is strongly associated with obesity and other particular dietary imbalances. Diabetes is also sometimes a complication of pregnancy. A number of common genetic variants have been associated with diabetes risk.
A new open access paper at PLoS Genetics (Chen R, Corona E, Sikora M, Dudley JT, Morgan AA, et al. (2012) Type 2 Diabetes Risk Alleles Demonstrate Extreme Directional Differentiation among Human Populations, Compared to Other Diseases. PLoS Genet 8(4): e1002621. doi:10.1371/journal.pgen.1002621) shows that essentially every known gene associated with risk for Type Two Diabetes is more common in Africans than in Asians. As the abstract explains:
Many disease-susceptible SNPs exhibit significant disparity in ancestral and derived allele frequencies across worldwide populations. While previous studies have examined population differentiation of alleles at specific SNPs, global ethnic patterns of ensembles of disease risk alleles across human diseases are unexamined.
To examine these patterns, we manually curated ethnic disease association data from 5,065 papers on human genetic studies representing 1,495 diseases, recording the precise risk alleles and their measured population frequencies and estimated effect sizes. We systematically compared the population frequencies of cross-ethnic risk alleles for each disease across 1,397 individuals from 11 HapMap populations, 1,064 individuals from 53 HGDP populations, and 49 individuals with whole-genome sequences from 10 populations.
Type 2 diabetes (T2D) demonstrated extreme directional differentiation of risk allele frequencies across human populations, compared with null distributions of European-frequency matched control genomic alleles and risk alleles for other diseases. Most T2D risk alleles share a consistent pattern of decreasing frequencies along human migration into East Asia. Furthermore, we show that these patterns contribute to disparities in predicted genetic risk across 1,397 HapMap individuals, T2D genetic risk being consistently higher for individuals in the African populations and lower in the Asian populations, irrespective of the ethnicity considered in the initial discovery of risk alleles.
We observed a similar pattern in the distribution of T2D Genetic Risk Scores, which are associated with an increased risk of developing diabetes in the Diabetes Prevention Program cohort, for the same individuals. This disparity may be attributable to the promotion of energy storage and usage appropriate to environments and inconsistent energy intake. Our results indicate that the differential frequencies of T2D risk alleles may contribute to the observed disparity in T2D incidence rates across ethnic populations.A New Paradigm For Population Level Epidemiology
In most studies of disease risk, disease incidence is known, and a potential ancestry based risk which could be due to either nature or nurture is inferred from disease incidence data after controlling for known environmental risk factors (e.g. diet), but genotype is not directly measured. This paper is direct study of known genotypes.
It is beyond reasonable dispute that Africans (and people with African descent) have greater frequencies of genes believed to present a risk for type two diabetes, while East Asians, and people of East Asian descent have a lower frequency of these genes.
It is possible that there are unknown confounding genes or dietary practices or cultural practices that prevent these genes from manifesting in a Type Two Diabetes phenotype, or alternatively make low risk individuals more prone to developing Type Two Diabetes than the known Two Two Diabetes risk genes would suggest.
An analogous case involving an unknown confounding gene occurs for lactase persistence. A significant number of Africans who are lactase persistent (i.e. continue to have no problem drinking cows milk as adults), which is closely associated with specific known genes in European, lack the main European LP genes. This is because there are other African variants of the LP gene (some not specifically identified) that serve the same purpose. This scenario is plausible because African gene-disease associations are less well studied, in general, than European and Asian gene-disease associations.
An analogous case of a dietary practice confound is the case of cardiovascular disease in Parisians. The residents of Paris, France have diets that are high in fats known to be important risk factors for cardiovascular disease. But, they don't have an incidence of those diseases that reflects that level of fat consumption. The main reason for the disparity appears to be that Parsians also drink lots of red wine and that some combination of the alcohol and the other contents of the wine counteract the dangers of a high fat diet.
Wider Implications Related To Diabetes
But, the null hypothesis would be that a significant share of the known racial variation in type two diabetes incidence flows from the rates at which type two diabetes risk alleles are present, and since the impact of these type two diabetes risk alleles has been quantified in most cases, it should be possible to statistically distinguish between the racial variation in type two diabetes incidence due to known genetic risks from the variation due to environmental factors and undiscovered hereditary factors: the reverse of the usual epidemiological paradigm.
A study conducted with the methodology used in this paper would likely find, for example, that a significant share of of the type two diabetes incidence among African-Americans in the American South which has been previously attributed with old paradigm epidemiological methods to poor choices in diet and exercise in epidemiological studies conducted without genotype information may actually be due to differences in genotypes between African-Americans and whites.
Of course, while the proportion of the two two diabetes incidence rates attributable to heredity rather than diet and exercise may shift, the practical response is much the same.
In essence, a person with type two diabetes risk genes is someone who can't escape the disease consequence of suboptimal diet and exercise choices to the extent as someone who lacks those genes. For example, a full blooded Korean American woman with the same sedentary office worker lifestyle and moderately high sugar and fat and calorie diet is probably quite a bit less likely to get Type Two Diabetes than an African American woman with a typical level of African and non-African admixture, whose activity and diet are precisely the same. Proof, once again, that life is not in the least bit fair.
Put another way, on average people with African descent need to pay more attention to lifestyle risk factors for type two diabetes and obesity to avoid ill health effects than the average person.
Implications For Racial Disparity In Other Disease Risk Genotypes
On the other hand, with a few other minor exceptions (sickle cell anemia vulnerability, for example, which is more common in Africans because the same gene that causes the disease also provides resistance of malaria), a not very clearly stated implication of this study (given that it looked at genotype studies of 1,495 diseases and that Type Two Diabetes stood out as the most noteworthy of them) is that almost no other common diseases with known common SNP allele genetic risk factors have such a starkly ancestry linked pattern of genotypic risk. Type Two Diabetes appears to be something of a worst case scenario.
The default assumption when looking for geographic pattern in different diseases which are not associated with diet and metabolism or infectious diseases of a geographically constricted range, is that should be that genotype is much more loosely linked to geography, race or deep ancestry.
Also, since the source populations of particular areas of Europe or Asia often have much smaller effective populations and have shown serial founder effects, even very large populations in these areas are likely to have quite homogeneous patterns of disease vulnerability risk, so epidemiological models that focus on environmental factors may be more viable in these situations than populations in or near African, and multi-continental mixing pot populations in the New World.
Implications For Modern Human Evolution
Which adaptations conferred the greatest fitness advantages?
If diet, metabolism and infectious disease risk are the primary distinctions in disease risk genotypes between Africans and non-Africans (and one can obviously add skin, eye and hair coloration and type to the list), this implies that infectious disease and food supply have been some of the most powerful factors in the evolutionary selection on modern humans in the post-Out of Africa era, while many other hypothetically plausible targets of evolutionary selection in modern humans in this era turn out to have been largely irrelevant to evolutionary fitness for modern humans in this era.
When and where did these adaptations become common?
Of course, knowing that there is a distinction doesn't necessarily tell us when that distinction arose. Did it arise in the Middle Paleolithic, when modern humans left Africa; did it arise in the Upper Paleolithic, when modern humans settled in Europe, Australia, Papua New Guinea, Japan and the Americas for the first time; or did it arise in the Neolithic together with the shift from a hunting and gathering mode of food production to a farming and herding mode of food production, or could it be an even more recent metal age development?
The maps, which show Papua New Guinea and the Americas to be largely congruent with the Asian pattern suggest that these genotype differences arose at least as far back as the Upper Paleolithic and prior to the Neolithic revolution in Asia. Otherwise, American and Papuan populations which were genetically isolated from Neolithic populations until a few hundred years ago, would look more like the African populations and less like the Asian ones.
Although it is harder to eyeball, there appear to be (and the charts in the body of the paper support the conclusion that there are) lower frequencies of type two diabetes risk genes in areas that have "Southern Route" population histories in Asia, and intermediate frequencies of type two diabetes risk genes in Europe and areas with ties to Central Asia that were at some point or another experienced a significant and lasting Indo-European presence.
This suggests a two step selection process - one common to all non-Africans and possibily diluted at the African fringe by short distance migration, and a second one particular to non-Africans who got far enough along on a Southern route to make it past South Asia and were subject to heightened selective pressure.
Alternately, one could also imagine a scenario in which many of the common type two diabetes alleles emerge somewhere Asia where they approach fixation. The presence of these alleles at all in other parts of the Old World could be entirely due to back migration from Asia.
Indeed, while the distribution of these genes disfavors an association with Denisovian admixture which is much more narrowly distributed, although one could also fit a Neanderthal source to these genes to their distribution quite easily so long as one assumed that they conferred more selective advantage in Asia than in Europe.
To the extent that these were a disease resistance alleles that had selective advantage for non-Africans other than Asians (even if the advantage was not as great for Europeans than for East Asians, for example), it is worth noting that the overall genetic contribution of the back migrating population could have been much smaller than their percentage contribution of specific alleles at these specific locations in the genome where the percentages would be amplified over tens of thousands of years by the fitness advantage that they conferred.
For example, even though 40% of Europeans have some type two diabetes risk reduction allele that has reached near fixation in China today, that could easily have emerged from a back migration that is a source for only 4% or less of the overall whole genome of Europeans.
Such a back migration could easily have happened, for example, in a period that was some time after the Toba erruption (ca. 74,000 years ago) which is a plausible geoclimatological event that might have coincided with the arrival of modern humans in Southeast Asia from South Asia, but was long before modern humans started to displace Neanderthals in Europe (ca. 42,000 years ago). Hence, the presence of these disease risk reducing alleles may date back to a period when the back migration was to a proto-European population in South Asia, Iran or the Middle East, rather than to modern humans who actually lived in Europe at the time.
What kind of events could have caused these genes to confer a fitness advantage?
Perhaps the genes approach fixation because at some point in ancient prehistory in Asia only holders of the genes could survive some genetic bottleneck in significant numbers, or because the fitness enhancement was greater given the foods available in Asia than in Europe and the Middle East.
For example, perhaps there are more plants with natural sweet sugars in Asia than Europe and the Middle East, so an ability to manage blood sugar levels became more important.
What would a bottleneck like that look like? I'd imagine one where the ability to be both obese and healthy in the long term would provide an advantage. Thus, you'd imagine perhaps a hunting and gathering population that experienced both "fat times" and "lean times" where the best adapted individuals got quite obese in the fat times, without developing disadvantageous diabetes, and then were able to survive lean times as a result of having these great food reserves that the peers who either got fat and died from diabetes, or didn't get fat in the first place, lacked. These genes may have served a purpose analogous to the one served by a camel's ability to store water internally for long periods of time after gorging itself.
At a generalized level, type two diabetes risk reduction alleles might confer fitness primarily by providing a survival advantage in circumstances when food supplies are unreliable, something that may not have been nearly so much of a selective pressure for Africans many of whom may have enjoyed (and still enjoy) a more stable tropical and subtropical climate and seasons that had less of an impact on food supplies for hunters and gatherers. Those inclined to put things in Biblical metaphor could describe Africa as an Eden and the Eurasians as exiles from Eden who faced greater struggles to meet their needs from nature that made it fitness enhancing to have these alleles.
And, the Edenic nature of the food supply in Africa wouldn't have been coincidental. Modern humans, our hominin predecessors, and our primate predecessors evolved over millions of years to be ideally suited to African life. We may have been better able to find food all year around, year after year in Africa, because only the ancestors who learned to hunt, gather and eat a wide enough range of foods to do so thrived and were rewarded by the evolutionary process. Had modern humans evolved in Southeast Asia instead of Africa, for example, perhaps we would, like pandas, be able to digest bamboo, for example. But, since we evolved in Africa, rather than in Europe or Asia, even our omnivorous diet may been able to secure nourishment from a smaller percentage of the biomass that could have been food there than it did in Africa. And, the smaller the percentge of biomass one can digest, the more unstable one's food supply will be and the more one needs to be able to store calories in fat in good times so one can survive later in hard times.
One More Piece of Evidence Added To The Clues That Reveal Pre-History
The selectively driven type two diabetes risk allele distributions (and despite my lumping together of these alleles in the discussion above, there are really at least six separate distributions to consider which could have spread to world populations in two or more separate events) provide another tool, on top of uniparental genetic markers, autosomal genome data, and other genes that have known functions with geographically distinctive distributions (like lactase persistence genes and blood types) that allow us to make inferences (bound by the limitations that the inferences be consistent somehow), about human pre-history that we are pressed to understand with the very sparse available archaeological evidence.