Archaeic Admixture & Metabolic Variants

berkley.edu

With the ever growing emphasis on understanding how human nutrient needs evolved, I think it's pertinent to remember all of the forces of evolution -- I said back when I first started this blog that there's a strong chance that specific human populations likely have nutrition/metabolism variants related to the forces of evolution besides selection in a specific nutritope. Recent evidence suggests that gene flow may have played quite an important role in the evolution of metabolism. This evidence stems from the recent sequencing of a high quality Neanderthal genome by Svaante Paabo's group at the Max Planck Institute for Evolutionary Anthropology, leading evolutionary biologists and anthropologists to suggest that ancestral populations interbred with Neanderthals.

Neanderthals and humans likely diverged from their last common ancestor, which has eluded paleoanthropologists (1), about 1 mya; however, molecular clock data suggests a more recent LCA (2). Neanderthals disappeared around 30,000 years ago, but co-existed with 'humans' for a period of time, and likely interbred somewhere around 47,000-65,000 years ago (3). (i use human loosely as the debate still rages amongst Anthropologists as to whether Neanderthals should be considered 'human'). Because of this interbreeding, anywhere from 1-4% of the genomes from Non-Africans is derived from Neanderthals. Recognizing that some of us are walking around with Neanderthal alleles can potentially explain some of the phenotypic disparities seen between historically geographically distinct populations.

The first report to link gene flow from Neanderthals to metabolic disease was in the Mexican and other Latin American populations in the form of SLC16A11 sequence variant (4). Mexican populations have nearly twice the prevalence of type 2 diabetes mellitus (5), and this report, a GWAS study, identified variants in the SLC16A11 and SLC16A13 genes to be novel risk loci associated with T2DM. Other GWAS looked at individuals primarily of European descent, and did not identify these loci. 5 variants (1 silent mutation, 4 missense SNPs) at the A11 locus, showing strong linkage disequilibrium, were strongly associated with T2DM (even after adjustments for age/BMI), and co-segregated into a common haplotype that is common amongst individuals of Latin American ancestry. The variant has also been shown to alter lipid metabolism in HeLa cells, increasing intracellular triacylglycerol levels. What's most interesting about this finding is how rare the 5SNP haplotype is throughout populations of European and African ancestry, while being found in the genome of a Neanderthal from the Denisova Cave (but not found in the reference sequences for Neanderthals or Denisovans). This 73kb spanning haplotype is nearly identical for both Neanderthals and individuals in the 1000 Genomes Project that are homozygous for the 5 SNPs. The authors conclude that it is highly likely that the haplotype was introduced into modern humans via archaic admixture; thus, gene flow from Neandethals conferred a relatively higher risk of developing T2DM (2.1 years earlier, at a .9 lower BMI) to this population.

Nature.com.
GW=Genome Wide, LCP= Lipid Catabolism

More recently, researchers have questioned whether genomic regions that were obtained from Neanderthals are distributed randomly throughout contemporary modern human genomes (as one would expect), or, if certain regions remain more/less influenced, indicating positive/purifying selection. To answer this, Paabo and a group of bioinformaticians analyzed the genomes of 11 contemporary humans (3african, 3 asian, 5 euro), Neanderthals, and chimpanzees (6). After some D-stats, the authors parsed what they refer to as Neanderthal-like sites (NLSs) in the different human populations, and looked for enrichment of specific gene sets based on ontology terms. Interestingly, Asian populations showed the most enrichment in sites linked to immune and haematopoietic pathways. Europeans showed the strongest signal in lipid catabolic processes (LCPs). One interpretation here is that Neanderthals accumulated some beneficial variants related to lipid catabolism that conferred survival benefits when passed to modern humans living in the same nutritope. The authors went on to confirm evidence of selection at these lipid catabolism sites in Europeans (no evidence was found for Asians), using a composite of multiple signals score based on long-range haplotypes, differentiated alleles and high-frequency-derived alleles.  Next, the authors looked at a number of lipid brain metabolites and broke these down into 16 metabolic categories, 7 of which were associated with lipid catabolism. In Europeans, metabolites within these 7 categories were more diverged from chimps than the other 9, and Africans/Asians showed no divergence of lipids associated with any of the 16 categories. Of note, the brain lipid metabolites with the greatest number of peaks, out of the 16, were phosphatidylcholine, phosphatidylethanolamine and sphingomyelin. 7 of these metabolites were linked with 6 genes involved with lipid catabolism, identified using RNA-seq. The expression divergence of Europeans was, again, greatest in divergence from chimpanzees, as opposed to intermediately in Asian and no difference in African. The authors found that the 6 genes with divergent expression related to lipid catabolism also contained a significantly higher number of NLSs than all LCP genes, and the NLSs were concentrated at the transcription start sites of these genes (SNPs likely affecting gene expression).  As a final analysis, the authors investigated whether these same lipid catabolism changes were seen in Denisovans; they were not, implicating that this effect is unique to Neanderthals and Europeans.

As much I think it'd be really interesting to conclude this is the result of gene flow, other options must be considered. It's clear that there's some selection on LCPs that are affecting gene expression + brain lipids concentrations in Europeans and Neanderthals, but we can't entirely assume this is from gene flow. Is this some sort of incomplete lineage sorting? Were these variants present in the most recent common ancestor of Neanderthals and humans, and driven to a higher frequency due to some (elusive) selective factor, living in the same environment? Until the 1000 genomes project gets a larger sample size, we can't entirely conclude that these variants aren't present in other populations at varying prevalences, drifting about unselected for. I'll be interested to see more paleoanthropologists weighing in on the potential environmental conditions that may have led to the positive selection on this haplotype. Having only looked at one body system (brain), I'd take every interpretation put forth with several grains of salt.

The authors controlled for environmental effects on metabolites by normalizing the metabolic divergence in the LCP term to the divergence based on all other metabolic pathways within the same population. However, looking at the metabolite categories that are effected most and match to divergent gene expression (trigylcerides, cholesterol esters, lecithin, ceramide, sphingomyelin), diet/nutrition likely play a major role in affecting these metabolites concentrations.  One wonders how much of the noted divergence is due by nutrition (though I'm doubtful the trends seen would actually be significantly affected). Time to start hypothesizing how available nutrients in the shared European environment may have interacted with this haplotype! I personally find the choline stuff really interesting.

These analyses are prime examples of the need for focusing on identifying nutrient/metabolism-related haplotypes. A lot of individual gene variants grab the spotlight in nutrigenetics (ApoE4, Cytochrome P450 CYP, FADS, MTHFR), but it's more than likely that profound phenotypes have polygenic origins. As the study of nutrigenomics continues, i hope to see more evolutionary anthropology and population genetics playing a role.

I'm also excited to see what other variants may have flowed from ancestral hominids like Neanderthals, Denisovans, and these unknown hominids we're finding evidence of (7,8)... We need more studies of this Paleo Diet!

1. http://www.smithsonianmag.com/science-nature/homo-antecessor-common-ancestor-of-humans-and-neanderthals-143357767/?no-ist
2. http://www.pnas.org/content/early/2013/10/15/1302653110
3. http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002947
4. http://www.nature.com/nature/journal/vaop/ncurrent/full/nature12828.html
5.http://www.ncbi.nlm.nih.gov/pubmed/20585724?dopt=Abstract&holding=npg
6. http://www.nature.com/ncomms/2014/140401/ncomms4584/full/ncomms4584.html
7. http://www.nature.com/nature/journal/v464/n7290/full/nature08976.html
8. http://www.upi.com/Science_News/2013/11/19/DNA-of-early-hominid-found-to-include-mystery-early-genes/UPI-70651384895737/