Thread by @StephenORahilly

Delighted to see our work on post-prandial insulin resistance published nature.com/articles/s4158… Its the result of a great collaboration with, among others, #ClaudiaLangenberg @omicscience, @DPhaz @klmohlke, #EleanorWheeler and #NickWareham

The first author is Alice @ AC_Williamson, a terrific PhD student with Claudia and me. In 2000 with @satya Dash009 , #David Savage we described a family where a mutation in a Rab-GAP for GLUT4, TBC1D4, impaired GLUT4 translocation in response to insulin in muscle and fat/2.

The mutation co-segregated with extremely high post-meal insulin levels. Clearly, the normal pancreas was somehow sensing the abnormally slow rate of disappearance of glucose into muscle and fat tissues, where both GLUT4 and TBC1D4 are normally expressed/3

As these genes are not expressed in liver, fasting insulin levels were normal (actually sl. low as the mutation causes a constant trickle of GLUT4 to the pl. membrane, depleting the intracellular pool so little is available to be translocated in resp. to insulin)/4.

In 2014 #TorbenHansen reported a similar mutation (in this case disrupting an isoform of TBC1D4 only expressed in skeletal muscle) at high frequency in Greenland Inuit population, where it is associated with post prandial (but not fasting) hyperinsulinemia and hyperglycaemia/5

So, people can be differentially sensitive to the glucose lowering effects of insulin in the post prandial vs the fasting state. Insulin resistance (and/or the compensatory hyperinsulinemia that results from it) is a key precursor to many cardio-metabolic illnesses/6

There have been many GWAS of insulin resistance. But such studies have had a principal focus on the fasting state, during which insulin controls glucose metabolism largely through its effects on glucose production by the liver (a process which doesn’t involve GLUT4)/7

The gold standard test for muscle insulin sensitivity is a hyperinsulinemic clamp with stable isotope measures of glucose disappearance. Such studies are very intensive and not suitable for large scale population genetic studies/8

So we used as a surrogate the fold increase in plasma insulin between the fasting state and 2 hrs after an oral glucose tolerance test. We also used the Stumvoll index as a sanity check, though this does reflect both fasting and post prandial insulin status)/9

Relevant data was available from ~55K ppl from multiple international studies and Alice and my gen. epi colleagues did their magic. An intronic polymorphism in the GLUT4 gene itself was highly significantly associated with increased Insulin Fold Change (IFC). Which was cool! /10

@klmohlke’s lab showed that this variant was associated with lower GLUT4 expression in human skeletal muscle and lower GLUT4 gene transcription in cells transfected with a construct containing that intronic variant/11

GLUT4 has been studied for decades and what we found was sort of obvious from previous mouse work, but it was gratifying to see human biology doing what we expected it to do. We found several other loci associated with IFC at genome wide levels of significance/12

Some of these encompassed “usual suspects” in ins. action e.g. IRS-1. To explore other loci functionally we turned to a system developed by @DPhaz in the @IMSMRL who has been using 3T3L1 adipocytes in which GLUT4 trafficking can be studied by two independent methodologies/13

Alice knocked down all the candidate genes in the IFC-associated loci and found 14 genes, interference with which caused highly significant alterations in either GLUT4 expression or trafficking. Please see the paper for the indiv. genes; there are some fascinating ones/14

Some look particularly druggable. In some cases inhibition increases GLUT4 movement to the cell surface, suggesting that blocking these could be insulinomimetic. Like all exciting studies this work has raised lots of great questions for follow up/ end.

A special thanks to @InesBarroso4 who kept us on the straight and narrow when writing the MS thereby greatly enhancing its quality