Case Study 1 — The Leptin Decade

What happens when a field finds exactly what it was looking for

Type: Real, public, historical · Tier 1 discovery facts, Tier 2 trial detail · Relevance: §13.2, §13.7

Background: fifty years of a mouse

The ob/ob mouse was described in the 1950s: a recessive mutation producing an animal that ate enormously and became severely obese. A related strain, db/db, looked almost identical.

Classic experiments in the 1960s and 1970s connected the circulations of pairs of mice — parabiosis — and produced a striking result. When an ob/ob mouse was joined to a normal mouse, the ob/ob mouse lost weight. When a db/db mouse was joined to a normal mouse, the normal mouse stopped eating and starved.

The interpretation, worked out well before anyone could identify the molecule: ob/ob mice lack a circulating satiety factor, which the normal partner supplies. db/db mice produce the factor in excess but cannot respond to it — so they flood the shared circulation with a signal their partner can hear and they cannot.

That inference was correct, and it was made decades before the molecule was found. It is one of the more impressive pieces of reasoning in twentieth-century physiology, and it deserves noting that the resistance phenomenon — the entire complication of §13.2 — was visible in the db/db mouse from the beginning.

1994

The ob gene was cloned and the protein identified: leptin, produced by adipose tissue in proportion to fat mass.

Everything fell into place. Fat tissue reports how much energy is stored. The hypothalamus receives the report. Appetite adjusts. A negative feedback loop for body weight, exactly the architecture Chapter 3 describes for every other regulated variable.

Administer leptin to ob/ob mice and they normalize — eating falls, weight falls, and the associated metabolic and endocrine abnormalities correct.

The commercial and scientific response was enormous. Rights to the discovery reportedly changed hands for a very large sum, the field expanded rapidly, and the expectation that an obesity treatment would follow was widespread and, at the time, entirely reasonable.

The trials

Leptin was given to people with common obesity.

The results were modest and inconsistent. Some weight loss was reported, generally small, variable between individuals, and requiring high doses. Nothing resembling the mouse result appeared.

The explanation was already available in the db/db mouse. People with common obesity have elevated leptin — appropriately, since it tracks fat mass. They are not deficient. They are resistant, in the same way the db/db mouse was resistant, and supplying more of a signal that is already being ignored does not help.

🔬 Read the Study — the leptin obesity trials

text FIGURE 13.CS1 — "Giving people more of what they already had" [real published trials] THE STUDY Randomized trials of recombinant leptin administration in adults with common obesity, following the 1994 discovery. Various doses and durations; industry- and academically-funded. THE QUESTION Does leptin administration produce meaningful weight loss in common obesity, as it does in leptin-deficient animals? WHAT IT SHOWS Modest, inconsistent, dose-dependent weight reduction — far below what the animal model predicted and below what would be clinically useful. WHAT IT DOESN'T It does not show leptin is inactive, that the pathway is unimportant, or that leptin fails in every population. It specifically does not address leptin DEFICIENCY, where the answer is entirely different. THE VERDICT ❌ for common obesity. ✅ for congenital leptin deficiency — a separate claim in a separate population. THE LESSON AN ANIMAL MODEL OF A DEFICIENCY IS A MODEL OF THE DEFICIENCY, NOT OF THE COMMON DISEASE THAT RESEMBLES IT. The ob/ob mouse was a perfect model of congenital leptin deficiency. It was never a model of common human obesity, and the difference was visible in the db/db mouse the whole time.

The part that worked

Congenital leptin deficiency exists in humans — rarely, and identifiable by genetic testing. Affected children present with extreme hyperphagia and severe early obesity.

Leptin replacement in these children is transformative. Appetite normalizes. Weight falls substantially. Associated abnormalities in puberty and immune function correct. The published descriptions are among the more striking in endocrinology.

This is a genuine cure-shaped treatment for a population of a few dozen people worldwide.

And it is the same shape as setmelanotide (§13.5), and the same shape as mecasermin for severe IGF-1 deficiency (Chapter 16), and the same shape as insulin for type 1 diabetes (Chapter 11): total deficiency, exact replacement, dramatic response. Chapter 2's override/replace rule predicts every one of them.

What the decade actually produced

It is worth listing, because "leptin failed" is the usual summary and it is wrong.

Adipose tissue was established as an endocrine organ. Before leptin, fat was understood primarily as storage. Afterward it was a signaling tissue, and the entire field of adipokine biology followed.

The hypothalamic appetite circuitry was mapped. POMC and AgRP neurons, the melanocortin pathway, MC4R — §13.4's diagram exists because researchers went looking for what leptin acts on.

The monogenic obesities were identified. Leptin, leptin receptor, POMC, and MC4R deficiencies were found because the pathway was known to look in.

Setmelanotide exists because of it. A drug acting downstream of the broken step only becomes conceivable once the pathway is mapped.

And the concept of a defended set point was made concrete. The idea that body weight is regulated rather than merely accumulated — which underlies everything in Chapters 8 through 13 — has its modern form here.

Judged as "did it produce an obesity drug," the leptin decade failed. Judged as "did it produce the knowledge that later produced drugs," it is one of the most productive periods in metabolic medicine. Both are true, and which one you report is a choice.

Discussion questions

1. The parabiosis experiments correctly inferred both a circulating factor and a resistance phenomenon decades before the molecule was found. What made that inference possible without molecular tools?

2. The db/db mouse showed resistance from the beginning. Why do you think the field's expectations after 1994 focused on deficiency rather than resistance? Is there a general lesson about which model a field attends to?

3. "An animal model of a deficiency is a model of the deficiency, not of the common disease that resembles it." Apply this to at least one other case in this book.

4. Leptin works dramatically in a few dozen people and modestly at best in hundreds of millions. How should a health system value a drug like that? Does the answer change if it is expensive?

5. Judged one way the leptin decade failed; judged another it was extraordinarily productive. Which judgment should govern funding decisions, and how would you defend that to someone who wanted a drug?

6. Forward. Chapter 16 covers myostatin inhibitors, where human trials increased muscle mass and did not improve function. Compare that failure with leptin's. Are they the same kind of failure?