The
genetics of obesity
Obesity is a very complex condition and involves
a multitude of metabolic pathways, each regulated by several genes. A disorder of one gene can lead
to severe obesity but the rate of occurrence of these single gene defects is so
low that they cannot explain the current pandemic of obesity. For example, a rare defect in the gene which
encodes for a protein, bearing the awkward title melanocortin-4 receptor
(MC4-R) can lead to obesity. But even among morbidly obese persons, this gene
defect is only found in 4% of this population. Thus a single gene defect is
never going to explain the obesity issue.
However, twin studies have repeatedly shown
that obesity is highly heritable to a level of about 70%. So if one gene defect
can’t explain this inherited susceptibility to obesity, maybe data on multiple
genes might assist us. A recent study by UK and US scientists[1], has
begun to throw some light on how we might begin to predict the susceptibility to
obesity based on our genes.
The first step in their quest was to explore a published database on 300,000 individuals whose body-weight was known and who had their
entire genetic code searched for over 2.1 million common genetic variants. No single
variation explained any significant risk of obesity. Thus the next step was to
take all of these 2.1 million common genetic variants and see if they could
compute a polygenic score which would be predictive of obesity. They used advanced
computational methodologies to devise 6 candidate scoring systems. To determine
which was best, they turned to the UK Biobank which has full genetic data and
body weight data on over 120,000 subjects. Their best performing scoring system
could now be put to the test.
They used four different data sets to explore
the predictability of their polygenic score in quite different groups. To understand their
scoring system, they could first dismiss all those common gene variants which
had simply no link with obesity. Thus our hair and eye colours are genetically
determined and different genetic variants explained redheads from blondes to
brunettes. So they were scrapped along with lots of others. They were left with
genes that had a very, very minor effect on obesity right up to some genetic
variants with a more significant link with obesity but still, on their own, would have no real predictive power. The more of the higher linked variants you
had, the higher your polygenic score. So every individual in the UK Biobank got
a score ranging from low to high. They divided the 300,000 subjects into ten
groups (deciles) with increasing polygenic scores and they showed a very strong
link between the score and body weight. Half
the people with the lowest score had a normal body mass index whereas among those
in the top 10% of the polygenic score, only 17% were of normal weight.
Conversely, among those classified as obese, only 9% belonged to the lowest scoring
decile while 38% were at the top end of the score. Across the ten deciles of
polygenic score there was a linear increase in bodyweight, BMI and the incidence of severe
obesity.
The next data sets they turned to were ones
that included subjects who underwent bariatric surgery for the treatment of
their morbid obesity. A high polygenic score was associated with a 5.0-fold
increased risk of severe obesity treated with bariatric surgery. To understand
the evolution of obesity over time, they then turned to the Framingham Offspring
and Coronary Artery Risk Development in Young Adults (CARDIA). This gave data
on 3,722 individuals who at baseline had no case of severe obesity. Over the
next 27 years they were weighed 8 times. Over that period some subjects went on
to develop severe obesity. Among those in the lowest 10% of the polygenic score
just over 1% went on to develop severe obesity. In contrast, those in the top
10% of the polygenic score, 16% went on the become severely obese. The final
database they used was one of UK babies born in the years 1991-1992 who were
followed up to their 18th birthday (Avon Longitudinal Study of
Parents and Children). The difference in birthweight between the top and bottom
10% of the polygenic score was just 0.06 kg. By 8 years of age the difference
was 3.5 kg and by 18 years of age the difference had risen to 12.3 kg.
This has shown that we are fast becoming
capable of defining genetic profiles which predicts the likelihood development of a normal
weight or a severe obesity problem. But before you jump up and down and blame
your genes, remember the key saying: “Genes load the obesity gun. But only the
environment pulls the trigger”. That said, a knowledge of the likely risk of
sever obesity might motivate people with a high score to watch their calories
more closely than most
