Rebuttal to the Analysis that the Cause of Diabetes is Not Genetic

I think that your uncritical and unbalanced presentation of a non-peer reviewed analysis of the role of genes in diabetes, by total scientific unknowns with an agenda, was a great disservice to your audience.

Genetic analysis will not tell us all about the causes of diabetes but it is putting in place a very important piece of this multi-piece puzzle.  

The statements made in this article are blatantly false for both types of diabetes, but especially for Type 1, where genetic analysis has been a resounding success.

In Type 1 diabetes, the concordance in identical twins is >50% if you follow them long enough (Redondo et al., N Engl J Med. 2008 Dec 25;359(26):2849-50), while in dizygotic twins, exposed to exactly the same environment, it is ten times less (and not different from ordinary siblings, shooting down intrauterine environment as a possible difference between mono- and di-zygotic twins). Currently known genetic loci can identify, at birth, from a drop of blood, more than 90% of diabetics within the 20% of the general population with the highest risk. These are percentages that begin to be useful at the individual level. Suppose that one of the immune interventions currently tried by TrialNet and others to prevent T1DM in antibody-positive individuals is proven effective. Are we going to screen the entire population of children with antibodies yearly? This is logistically impossible but if we can confine this screening to only 20% of the population it is brought to the realm of the feasible.

By contrast, although we know that environment is also important in Type 1 diabetes, not a single environmental factor, modifiable or otherwise, has been identified to date. The recent publication of the TRIGR pilot data (N Engl J Med. 2010 Nov 11;363(20):1900-8) pours a lot of cold water on the cow’s milk theory. Intervention to remove cow’s milk from infant feeding made no difference in rates of diabetes in (genetically) high-risk infants and the (barely significant) difference in antibody positivity can be trivially explained by an unusually high rate of missing antibody data from children who ended up developing diabetes anyway. The literature on viral causes is a mess of half the papers contradicting the other half.

Here, I might also mention that genetic study of a very rare form of neonatal diabetes has discovered RFX6, a previously unstudied gene that is absolutely necessary for the development of the pancreatic beta cells and, likely, has been a crucial “missing link” in efforts to cure common types of diabetes by generating new beta cells. (See https://www.nature.com/nature/journal/v463/n7282/full/7282708a.html)

For Type 2 diabetes, it is true that environment is supremely important. There, genetics is called to explain why, from individuals exposed to exactly the same environment, some get diabetes but most don’t. The problem in T2DM is that genetic risk is due to a very large number (thousands) of genes, each contributing very small effects. This is not at all the same thing as saying that genes have no role in determining the disease. It will be closer to truth to say that, unlike the situation in T1DM, no single gene has an effect large enough to be useful in genetic prediction. This is not because these genes are not important, but simply because they are not different enough amongst individuals to account for relative risk. Their knowledge can still serve in the development of new treatments.

I think you owe it to your audience to publish a more balanced view of the issue.

Constantin Polychronakos M.D., F.R.C.P.C.
Professor, Departments of Paediatrics and Human Genetics
Director, Division of Endocrinology
McGill University Health Centre (Children’s Hospital)
Editor, Journal of Medical Genetics

https://www.montreal-diabetes-research-center.org