Homeostasis

It's been a while since the last post - business travel and then the inevitable catching up.

I have been thinking about homeostasis lately and why GWAS results don't describe all that much regarding the total amount of variability. Perhaps it takes many hits ("disease" or "risk" alleles) in the same operational unit (e.g., a pathway) to see disease or biomarker thereof as a measurable phenotype. And so I wonder if the organism has a much stronger drive to maintain homeostasis than we realize. In other words, the body can absorb many small defects to a given pathway so long as the environmental conditions do not go awry such as might happen with years of poor nutrition. The body can even operate well within a certain range and that brings to mind a term I heard the other day - homeodynamics. It is true. Running a marathon won't make someone collapse, neither will a hot day and neither will a low intake of fluids. The combination, however, will greatly increase the risk of that taking place.

So, with respect to my research and examining GWAS data, I am more convinced than ever that we must look at an enrichment of pathways hit by the association data. I'm excited to give that a try.

Migraine and cardiovascular disease risk

A few recent papers report associations between migraine and cardiovascular disease (CVD) risk. This is particularly true for migraine with aura. Another study shows that this link is found in women. What is most exciting is the recent publication by Markus Schürks and colleagues reporting findings from a candidate gene study on the genetic basis of migraine. The data imply that polymorphisms of genes encoding constituents of inflammation pathways and migraine in women are associated. These genes include TNF, CCR2, TGFB1, NOS3, and IL9. We have looked at a few of these genes with respect to dyslipidemia, obesity and metabolic syndrome.

While it would be nice to look for direct connections between biomarkers of CVD and migraine, the phenotype collection of two of the largest and most useful populations with which we work do not contain information from the subjects on migraine or pain in general. These populations are GOLDN and BPRHS. This is where the proposed Nutrition Phenotype database project could come in handy. This group is calling for more extensive phenotyping of subjects. The project is in its infancy but an introductory paper has been accepted at Genes and Nutrition. The paper is titled "The Nutritional Phenotype database to store, share and evaluate nutritional systems biology studies" and was authored by B. van Ommen, J. Bouwman, L. Dragsted, C. A. Drevon, R. Elliott, P. de Groot, J. Kaput, J. C. Mathers, M. Müller, F. Pepping, J. Saito, A. Scalbert, M. Radonjic, P. Rocca-Serra, T. Travis , S. Wopereis and C. Evelo.

TXNIP - lots of data

A new report by Zhou, et al. (Nature Immunolog, http://www.nature.com/ni/journal/vaop/ncurrent/abs/ni.1831.html) links thioredoxin (TRX)-interacting protein (TXNIP) to oxidative stress and inflammation. This gene has been connected to insulin resistance and a host of other activities. These include:

mouse knockouts show increased fat to muscle ratio

TXNIP is a rat hyperlipidemia gene (rat genome database)

highly expressed in mouse adipose (3.23-fold over average of all other tissues)

TXNIP expression was inversely correlated to total body measures of glucose uptake (Parikh, Mootha 2007 PLos Med)

forced expression of TXNIP in cultured adipocytes significantly reduced glucose uptake, while silencing with RNA interference in adipocytes and in skeletal muscle enhanced glucose uptake, confirming a role as a regulator of glucose uptake (Parikh, Mootha 2007 PLos Med)

TXNIP expression is consistently elevated in the muscle of prediabetics and diabetics, although in a panel of 4,450 Scandinavian individuals, we found no evidence for association between common genetic variation in the TXNIP gene and T2DM (Parikh, Mootha 2007 PLos Med)

TXNIP regulates both insulin-dependent and insulin-independent pathways of glucose uptake in human skeletal muscle (Parikh, Mootha 2007 PLos Med)

this is gene #135 in a published PPARA signaling network analysis (Cavalieri, Muller 2009 BMC Genomics 10:596)

TXNIP is 1.5-fold increased expression in young vs. old mouse skeletal muscle. In old muscle, calorie restriction turns up the gene (young phenotype) also ~1.5-fold over normal diet (GDS2612)

So, perhaps the newest paper is not surprising, but welcome data to integrate into the big picture.

Blaine at The Genetic Geneologist (www.thegeneticgenealogist.com/2010/01/07/personalized-genomics-a-very-personal-post/) has posted a very personal story about his genotype at three loci for risk of type 2 diabetes. The SNPs are rs5219 in KCNJ11, rs1801282 in PPARG and rs7903146 in TCF7L2. With risk alleles for all six alleles, his risk, as calculated by the services he used, is significantly elevated.

A genotype is, as he writes, merely a suggestion of a possible future outcome. Lifestyle intervention before disease onset is clearly warranted. We have shown, for example, that variants at TCF7L2 associate with postprandial lipemia (after-meal levels of blood lipids) but are modulated by PUFA (polyunsaturated fatty acid) intake. See www.ncbi.nlm.nih.gov/pubmed/19141698, where my colleagues wrote "high (n-6) PUFA intakes (> or = 6.62% of energy intake) were associated with atherogenic dyslipidemia in carriers of the minor T allele at the TCF7L2 rs7903146 SNP and may predispose them to MetS, diabetes, and cardiovascular disease." Links between dyslipidemia and T2DM are well established.

In other words, the interplay between environment and genome is an important consideration that is often overlooked.

Positive selection and risk of heart disease

We have just submitted a review article discussing the role of genetic variants under positive selection that have roles in dyslipidemia, obesity or heart disease. Many of the identified SNPs show interactions with environmental or dietary factors, which means that one allele of the SNP, typically the risk allele, shows an effect on the phenotype only when the environment or diet passes a certain threshold. Specifically, SNP 3111 T>C (rs1801260) of the CLOCK gene shows significantly increased waist circumference for the C allele only when the saturated fat constituent of the diet is above 11.8% of total energy. Our results from variants in the CLOCK gene are published at Am J Clin Nutr 2009;90:1466–75 (www.ncbi.nlm.nih.gov/pubmed/19846548). This SNP is also under positive selection or can be called an adaptive variant. Think of natural selection.

The implication that genetic variants under positive selection have a role in human disease is profound. These polymorphisms sense the local environment - diet, sunlight, sleep, exercise, alochol and tobacco use, exposure to pathogen, etc. - and participate in an allele-specific response of the cell, tissue, organism to that stimulus if you will. Thus, modification of one's lifestyle choices based on genotypes may be a viable choice to maintain a heathly condition. Before that, though, it will be necessary to define the effects of many, many more genetic polymorphisms, both SNPs and copy number variants (CNVs). To date we've compiled over 650 such examples where the SNP-phenotype association is modified by environment/diet. These pertain mostly to blood lipids such as LDL-cholesterol, HDL-cholesterol, triglycerides and total cholesterol. CNVs are generally under-studied in such work.

A summary of the review article is here:
Purpose of review: Obesity, cardiovascular disease and unhealthy levels of blood lipids known as dyslipidemia are complex and arise from both genetic and environmental factors as well as their interrelationships. Both large-scale genetic experiments and approaches focused on a single gene have resulted in identification of many genetic variants that are related to lipid and obesity measures. However, these variants still account for only a small fraction of the observed differences in these measures.

Recent findings: That many such genetic associations to obesity and lipid measures involve genetic variants under natural selection and that those variants are sensitive to environmental cues together suggest prominent roles for both natural selection and the interaction with the environment. Genetic variants under natural selection which interact with the environment modulate susceptibility to disease but the level to which those variants contribute to dyslipidemia and obesity and how environmental factors, especially diet, alter the genetic relationship is not yet completely described.

Summary: It is evident that genetic variants under natural selection make important contributions to obesity and heart disease risk. Advances in resequencing the entire human genome will enable accurate identification of genetic variants under positive selection. This will add power to large-scale genetic studies and allow for characterization of the relationship between natural selection and the obese and dyslipidemic conditions.