Damaging variants in fat-storage gene PLIN4 show up nearly four times as often in severe anorexia

The ratio holds up on the figures in Michigan State's account, as reported by Neuroscience News ^[9]. Seventeen of 154 is 11.0% ^[10]. At the roughly 3% rate seen in reference databases ^[3], a group of 154 would be expected to hold four or five carriers ^[11]. This one held 17, and 11 divided by 3 is about 3.7 ^[12].

The comparison group needs the closest scrutiny, because a rare-variant excess has to survive a comparison with controls processed the same way before the fold-change can be taken at face value. The 3% comes from large-scale reference databases ^[3], and the press account does not describe a control group sequenced alongside the patients or give a statistical test.

The cohort was also built to find this kind of signal. Patients were picked for strong multigenerational family histories, multiple failed treatments or atypical clinical courses ^[2]. The design suits a gene hunt, because rare variants of large effect tend to gather in heavily affected families. It also means 11% describes the hardest cases.

On these numbers a damaged copy is neither necessary for severe anorexia nor sufficient to cause it. Even in that enriched group, 137 of the 154 patients, about 89%, carry no damaging PLIN4 variant ^[13]. About 3 in 100 people in a healthy reference population do carry one ^[3].

The family results are the most suggestive part, and the smallest. Six patients carried severe loss-of-function variants, and in four families the identical variant appeared in an affected first-degree relative ^[5]. Close relatives share much of their DNA, so one shared variant in one affected pair is weak evidence alone. Here it happened in four families.

A germline variant does offer something blood work in patients cannot. It was present before any weight loss, so it cannot be a product of starvation. Exome sequencing identifies which variants patients carry, but this study does not test what those variants do to fat cells or appetite. The team's hypothesis is that faulty lipid storage could produce abnormal satiety or starvation signals from the body ^[16].

The researchers are plainly excited about where the gene works. "Probably the single most remarkable thing about this study is that this gene is mostly expressed in body fat and in muscle," said Michael Lutter, the psychiatrist who led the work ^{[8] [1]}. A.J. Robison, the Michigan State physiologist who advised the study, said the result "flies in the face of the dogma that the genes that contribute to this disease are expressed in the brain" ^[15]. They describe anorexia as possibly a complex metabolic-psychiatric disorder, in which defects in fat storage outside the brain contribute to susceptibility ^[14].

The treatment gap is large. Anorexia has one of the highest mortality rates of any psychiatric disorder, and care rests almost entirely on psychological therapy and inpatient nutritional refeeding, with no FDA-approved drug ^[7]. I think the result justifies a larger replication against matched controls, and any change in how patients are treated should wait for it.