The findings of the DiRECT trial of a low-calorie diet for type 2 diabetes remission are well known and are being applied to further our understanding of the pathophysiology of obesity. At one year, 46% of participants were in remission and at two years, 36% were still in remission. “This is way above what we thought would be a clinically useful outcome,” said Professor Roy Taylor of Newcastle University, who led the study with Professor Mike Lean of Glasgow University. “Our findings surpassed expectations and, in fact, this has tended to distract from the pathophysiological studies that led up to it.” 

These started with the ‘twin cycle’ hypothesis that excess food over a period of time leads to excess liver fat, which is exported to the pancreas where beta cells are exposed to it, leading to beta cell dysfunction and type 2 diabetes. Professor Taylor and his team investigated this hypothesis in the Counterpoint study, in which rapid weight loss through a low-calorie diet led to a decrease in liver and pancreatic fat (confirmed by MRI), normalisation of liver insulin sensitivity and a decrease in plasma triglycerides. There was then a gradual increase in beta cell function and insulin sensitivity, confirming the twin cycle hypothesis. There were similar findings in the DiRECT study in a range of body mass index (BMI) from 27 to 45. 

“The earliest sign of type 2 diabetes is muscle insulin resistance, which causes de novo lipogenesis after meals that then shifts fat to the liver,” explained Professor Taylor. “There is then insulin resistance in the liver, which is tightly coupled to liver fat accumulation. Plasma glucose goes up and insulin tries to compensate, which stimulates lipogenesis and the whole lipogenesis circuit reverberates.” 

But, as the American physician Roger Unger pointed out in animal experiments, it was only in susceptible mice that overfeeding caused diabetes. The fat would just ‘bounce off’ the islets of rats who were not susceptible. Corresponding to that monogenic effect in rats, there must be a polygenic effect in humans. “We must bear in mind that the vast majority of people who are very large do not have type 2 diabetes,” said Professor Taylor. “Does the twin cycle hypothesis alone really explain type 2 diabetes? What about normal weight people? We know from the UK Prospective Diabetes Study – the very study that has taught us all we know about type 2 diabetes – that 35% of participants had a BMI of less than 25.” 

Professor Taylor and his team published their work on the personal fat threshold in 2015. This explains that individuals have differing susceptibility to fat excess, due to a differing capacity of subcutaneous tissue to hold it. They tested this hypothesis in the recently published ReTUNE study, which involved people with type 2 diabetes and BMI 21 to 27. The study was designed for stepwise weight loss in 5% steps and this only went to 15% for those not in remission. 

Studies were done at each point and, at 52 weeks, there was no average weight gain. Body fat was 32% at baseline and 28% at one year, while the thickness of visceral fat also decreased, though it was still more than in the controls.  

Liver fat, which was three times greater than in the controls, and very-low-density lipoprotein (VLDL) triglycerides also fell to normal with weight loss. And 70% of these ‘slim’ people achieved remission. Put simply, data from DiRECT and ReTUNE suggest that liver fat is increased in type 2 diabetes over a range of BMI. There is more information in Professor Taylor’s book ‘Life without Diabetes’ (all profits to Diabetes UK) and at go.ncl.ac.uk/reverse-diabetes.

“So type 2 diabetes is a state of excess fat inside the liver and pancreas, irrespective of BMI,” concluded Professor Taylor. “The person with type 2 diabetes has become too heavy for their own body and the condition is caused by having more fat in the liver and pancreas than can be tolerated by the individual. The aetiology of type 2 diabetes does not, therefore, relate to the conventional definition of obesity based on BMI.”  

The puzzle of metabolically healthy obesity

As Professor Taylor observed, most people with obesity do not have type 2 diabetes and obesity does not necessarily go hand-in-hand with other comorbidities. So is it possible to be healthy and have obesity? 

“Obesity receives a lot of attention nowadays as the driver for the pathophysiology of type 2 diabetes,” said Dr Matthias Blüher at the Helmholz Diabetes Center, Germany. “If you speak to people in the outpatient clinic, they tell you the main goal for the next year is to lose weight. It’s very hard to find people who are happy with their body weight.” 

Obesity can kill because there are so many obesity-related conditions and comorbidities, most of which are improved by weight loss. There is some indirect evidence from the Swedish Obese Subjects trial that significant weight loss through surgery – say 20 to 30% – is associated with increased survival. Until recently, though, there was no evidence that weight loss reduces cardiovascular endpoints. Now we have top-line results from the SELECT trial of weekly semaglutide, which shows a 20% reduction in major adverse cardiovascular events (MACE) with 16% weight loss. “That suggests you need double-digit weight loss to get to these hard endpoints like cardiovascular events and mortality,” said Dr Blüher.

“However, we are all aware that there are people – 15 to 20% in some cohorts – who are obese but protected from cardiovascular and metabolic diseases,” he continued. This metabolically healthy obesity (MHO) is defined by having blood pressure < 130/85 mmHg; HDL-C > 1.03 mmol/l (men); HDL-C > 1.3 mmol/l (women); triglycerides < 1.7 mmol/l; HbA1c < 6% and fasting glucose < 6.1 mmol/l.

The prevalence of MHO is 2-19% for men and 7-28% for women, depending on how strict a definition is used. Data from the UK shows that those with MHO still have a higher risk of cardiovascular disease. Looking at underweight, normal weight, overweight and obese individuals with respect to the number of metabolic syndrome components (0-3) they had, researchers found a higher risk of heart failure, stroke and myocardial infarction by both BMI and number of components they had. So an underweight person with more abnormalities had a higher risk within a weight category and people with, say, one abnormality had a higher risk the greater their BMI. “This shows that obesity and overweight in themselves increase the risk of cardiovascular disease, even in the absence of metabolic abnormalities,” said Dr Blüher. 

To discover whether MHO Is it a distinct phenotype, Dr Blüher and his team carried out RNA sequencing of the adipose tissue of 1479 people and used artificial intelligence to identify six clusters, four of which have been characterised as being metabolically healthy, metabolically unhealthy, having high basal cortisol or hyperinsulinaemia. 

They then carried out clamp experiments that showed how insulin sensitivity varied with BMI. With some, as expected, insulin sensitivity increased with lower BMI and vice-versa. However, for a given BMI, there were some outliers who were very insulin resistant with a raised HbA1c to match. “One explanation might lie in the adipose tissue, for those with low insulin sensitivity have immune cells in adipose tissue and more hypertrophy,” said Dr Blühers. 

He went on to describe MHO in terms of size, sites and cytes. Size refers to adipocyte size – the larger they are, the more metabolically unhealthy. If you compare people of the same BMI and age, some have bigger adipocytes than others. Is there, then, a threshold in adipocyte size where the risk of type 2 diabetes increases? One of his studies has shown a steep rise in adipocyte size and diabetes risk in both French and German cohorts.  

Site refers to accumulation of liver and pancreatic fat, independent of BMI, as discussed by Professor Taylor. Those with high insulin sensitivity have low visceral fat. Finally, cyte refers to inflammation of adipose tissue. Large adipocytes are more likely to go into apoptosis and attract macrophages to get rid of the remnants.

From these data, Dr Blüher and his team have developed a model which aligns with Professor Taylor’s work. “There are safe depots for fat and weight gain like subcutaneous adipose tissue, but these may expand by increase in adipocyte size and number,” he said. Hypertrophy of adipocytes leads to abnormal adipose tissue function via autophagy, inflammation, hypoxia and apoptosis.   

“From genome-wide association studies, we know that there are genes that decide whether adipocytes expand and if you are more likely to develop stress in adipose tissue. So there may be genetic factors involved and this is, of course, under investigation,” he said. “There may also be environmental factors, which are poorly understood, such as persistent organic pollutants, plasticisers and pesticides that may contribute to the pathophysiology of adipose tissue function.” He and his team are investigating these. The adipose tissue is a ‘safe’ place to store these compounds, which may be good for the body as a whole, but maybe not for adipose tissue. 

Defining obesity

“BMI is not a good indicator of obesity-related outcomes,” said Dr Blüher. He prefers the Edmonton Obesity Staging System, with stages 0-4 in terms of medical, mental or functional problems. Here MHO falls into stage 0 and stage 4 is end-stage obesity. This system is independent of BMI and there are other staging systems that say it’s more important to take account of the comorbidities and risk factors than BMI, waist circumference or body fat mass. For instance, when it comes to the effect of obesity on mortality, research shows that the Edmonton system does predict mortality, while BMI does not.   

“My therapeutic goal for someone in stage 0 obesity would be to make sure they don’t gain more weight,” he said. “In stages 1 and 2, I think we can achieve remission, like remission of type 2 diabetes, and remove the risk of obesity-related disease, where with weight loss of 10 to 15% we can bring people back to a lower stage of obesity. But weight loss would be less useful in stage 3 or 4 when someone might already have had a myocardial infarction or be on dialysis.” 

Similarly, weight loss can be used to convert metabolically unhealthy obesity (MUO) with risk factors as explained above. “We have used bariatric surgery to do this and hope in future to use pharmacotherapy, so MUO is a reversible phenotype,” said Dr Blüher. With this approach, the person might still be obese in terms of BMI. “A useful target for weight loss need not be leanness or a BMI of less than 25,” he continued. “The achievable goal is to have a person with obesity, but without the cardiometabolic challenges.”

They also found significant changes in cell type in visceral and subcutaneous fat on weight loss, with visceral fat rebuilding adipose tissue and fewer immune cells in subcutaneous fat, as if they were trying to go back to their original composition. 

To learn more, enrol on the EASD e-Learning course ‘Obesity and diabetes’.

Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.