Leptin: The Discovery That Did Not Become a Drug

Why 1995 looked like the end of the problem

Halaas and colleagues showed that the product of the ob locus was present as a 16 kilodalton protein in mouse and human plasma and was undetectable in the plasma of ob/ob mice, the strain that becomes massively obese. Daily injections of recombinant protein reduced the body weight of ob/ob mice by 30 percent in two weeks with no apparent toxicity, reducing food intake and increasing energy expenditure. The result in normal animals was more startling still: wild-type mice injected twice daily lost 12 percent of body weight and their body fat fell from 12.2 percent to 0.7 percent. A hormone that takes a normal animal to essentially no fat reserve is not a modest finding, and the reaction at the time was correspondingly large.

The result in the same paper that predicted the failure

Two sentences in that paper contain the entire subsequent history of leptin as a drug. Plasma levels of the protein were increased in db/db mice, a strain thought to be resistant to the effects of ob - and the injected protein had no effect on those animals. A mutant with high circulating hormone and no response to more of it is the definition of resistance. Common human obesity turned out to resemble db/db far more than ob/ob: people with obesity generally have high circulating leptin, in proportion to fat mass, and giving them more produces little. The hormone reports how much fat is stored, and in obesity that report is being made loudly and ignored.

What leptin is actually for

The framing that survives is that leptin is not a satiety signal but a starvation signal, and that its physiology is asymmetric. Falling leptin, as fat stores decline, triggers a powerful defensive response - increased hunger, reduced energy expenditure, altered reproductive and thyroid function - because from an evolutionary standpoint the emergency is not having too much stored energy but too little. Rising leptin produces a far weaker response in the other direction, because there was rarely any need for one. A system built to defend a floor does not have an equally capable ceiling, and that asymmetry is why supplying more of the signal does little.

Where leptin did become a medicine

Metreleptin, a recombinant leptin analogue, is licensed - but for generalised lipodystrophy, a rare condition in which adipose tissue is absent or severely deficient and circulating leptin is consequently very low. That is the human equivalent of ob/ob rather than of common obesity: the signal is genuinely missing, so replacing it works. The same logic applies to the very rare cases of congenital leptin deficiency, where treatment is transformative. Leptin is an effective drug in populations defined by not having any, and an ineffective one in the far larger population that has plenty.

Why this history matters for reading current claims

Leptin is the cautionary case the whole field learned from, and its shape recurs. A hormone is identified; administering it to a deficient animal produces spectacular results; the deficiency turns out not to be the human problem; resistance or desensitisation blunts the response in the population that matters. GDF15's apparent susceptibility to desensitisation on prior exposure is a version of the same question. So is the observation that appetite-suppressing signals tend to run out of headroom. When a new appetite hormone is announced, the useful question is not whether it reduces intake in a deficient model but whether the human population is deficient in it.

What incretins did differently

GLP-1 receptor agonism did not succeed by replacing a missing hormone. It succeeded by administering a pharmacological exposure well above anything physiology produces, to a receptor that does not appear to develop the same resistance, sustained by engineering that keeps concentrations high for a week at a time. That is a different strategy from hormone replacement, and it is the reason the incretin class cleared an obstacle that stopped leptin, PYY and several others. Semaglutide and tirzepatide are licensed medicines and nothing supplied on this site is a version of or an alternative to either.

Quick reference