
and this is a summary of some of the points which came up in the comments (there is a lot of information and links from the comments about possible causes and management for those interested) added to the figure:

While I was raiding this paper I thought I would also put up Table 1, the characteristics of the control and diabetic offspring groups:

Now you have to be very, very careful with these groups. They have been exceedingly carefully preselected. Fortunately the pre selection process is laid out in some detail in a previous paper from 2004:
"All subjects were recruited by means of local advertising over a two-year period (2001 to 2003) and were prescreened to confirm that they were in excellent health, lean, nonsmoking, and taking no medications. A birth weight above 2.3 kg (5 lb) and a sedentary lifestyle, as defined by an activity index questionnaire,(21) were also required. Qualifying subjects (more than 150 persons) underwent a three-hour oral glucose-tolerance test (with a 75-g oral glucose load), after which two subgroups of subjects were consecutively selected to identify extreme phenotypes for insulin resistance and increased insulin sensitivity.
Insulin-resistant subjects (3 men and 11 women) were defined as having an insulin sensitivity index (22) of less than 4.0 (indicating insulin resistance; lower values indicate greater insulin resistance), at least one parent or grandparent with type 2 diabetes, and at least one other family member with type 2 diabetes. Insulin-sensitive control subjects (five men and seven women) were defined by an insulin sensitivity index of greater than 6.3 (with or without a family history of type 2 diabetes)."
As I read this it looks like over 150 people were screened for insulin resistance. Of those 150 those with the best and worst insulin sensitivity were selected as controls or subjects respectively. BUT you were only allowed in to the insulin resistant group if you did have a relative with T2DM. We get no idea of how many people had awful ISI and no diabetic relatives, ie if there were any index cases who might represent the red scrible on the second picture. Maybe there were loads, maybe not. I can't find that information in the paper. So we have to be very careful, the T2DM-relatives aspects MIGHT be a complete red herring. The insulin resistance is not.
Soooooo, with that caveat in place, we can see that (completely observationally) the insulin resistant group had, by definition, elevated insulin (and poor ISI) and the control group didn't. The control group weighed 61kg, the insulin resistant group weighed 66kg. Hmmmm. Observational, cross sectional. Fascinating.
You could simply say that the insulin resistant group were only hyperisulinaemic BECAUSE they were 10% porkier than the svelt control group. Indeed, if you consider porkiness to be a result of simple overconsumption of calories, for whatever reason, this would have to be an effect, not a cause.
Shulman's group extend the concept of mitochondrial failure in muscles to a potentially related mitochondrial failure in beta cells during the discussion. That's an interesting idea. Let's go one further and think about mitochondrial failure in adipocytes... I know it's odd to think that adipocytes (or indeed their mitochondria) might have anything to do with obesity, but stranger ideas have been floated.
Peter
BTW they also mentioned genes which control mitochondrial biogenesis:
"In this regard it is of interest that a common Gly482Ser polymorphism of the peroxisome proliferator–activated receptor γ coactivator 1, a transcriptional regulator of genes responsible for mitochondrial biogenesis and fat oxidation, has been linked to an increased relative risk of type 2 diabetes in Danish populations as well as to altered lipid oxidation and insulin secretion in Pima Indians."
I just noticed that the alpha form of peroxisome proliferator–activated receptor γ coactivator 1 got an honourable mention back in one of the Fiaf posts on the control of host metabolism by the gut mircobiota...