The Arizona Lab and the Drug That Escaped

Melanotan came out of a university laboratory in the Sonoran Desert. In the late 1970s and early 1980s, at the University of Arizona in Tucson, a peptide chemist named Victor Hruby and a physiologist named Mac Hadley were pulling apart alpha-melanocyte-stimulating hormone — the small signalling molecule, usually shortened to alpha-MSH, that instructs the pigment cells in skin to make more pigment. Their question was a practical one, and it was being asked in one of the sunniest cities in the United States: if you could persuade skin to tan without the ultraviolet light that ordinarily triggers it, could you take some of the danger out of a tan? Everything that followed — an approved prescription medicine, and one of the most widely misused unlicensed peptides in circulation — grew out of that question.

A tan without the sun

It is worth taking the motivation seriously, because the usual telling does not. By the late 1970s the epidemiology of skin cancer was moving in one direction and everybody in the field could see it. Rates were climbing across fair-skinned populations, most steeply in the places where the sun is fiercest and the culture most committed to being out in it — Australia, the American Southwest. The public health advice on offer was avoidance, and avoidance is a famously weak intervention against something people actively enjoy.

Melanin is a genuine, if modest, filter: more heavily pigmented skin burns less readily and carries lower melanoma risk. The trouble is the route by which most people acquire it. Ultraviolet light damages DNA in the outer layer of the skin, and the tan is the damage response — the repair crew arriving, not a substitute for the accident. You cannot tan in the ordinary way without first sustaining the injury that makes the tan happen.

So the idea had a certain elegance to it. If the pigment cells could be instructed directly — chemically, at the last step of the pathway rather than the first — a person might arrive at the protective end state without passing through the damaging one. Whether that would translate into fewer cancers was a hypothesis rather than a finding, and it is fair to say it was never tested in the form that would settle it. But it was a real hypothesis, held by serious people, and it is why the work attracted funding and attention. This was not a cosmetic venture that acquired a public-health justification afterwards.

What the chemistry actually did

Alpha-MSH is a short peptide, thirteen amino acids long, and like most short peptides it has a scheduling problem: it is potent, and it is gone almost immediately, dismantled by enzymes within minutes of release. A hormone built to act in pulses makes an awkward drug. The Arizona group's contribution was to fix that with two substitutions, each aimed at a specific weakness.

The first swapped the methionine at position four for norleucine, a residue of near-identical shape that lacks methionine's reactive sulfur atom and therefore cannot be oxidised. The second was the clever one. The phenylalanine at position seven was replaced by its mirror image — the D form, rather than the L form that all natural proteins are built from. The enzymes that chew through peptides are stereospecific: they are shaped to recognise L-amino acids and largely fail to grip the D-handed version. Flipping a single residue leaves the molecule's message legible to its receptor while making the sentence much harder to interrupt.

The resulting compound, NDP-MSH, was published in 1980 and described in the title of the paper itself as a highly potent alpha-melanotropin with ultralong biological activity 1. Both halves of that phrase matter. Superpotency on its own would have been a curiosity. Superpotency combined with duration turned a physiological pulse into a signal that stayed switched on — which was the achievement, and also, in hindsight, the property that made this whole family of molecules attractive to people who wanted something other than a medicine.

Two molecules, one receptor family

From this line of work came a cyclic analogue: Melanotan II. Cyclisation means joining part of the chain into a ring, which pins a floppy peptide into a fixed three-dimensional shape and leaves the enzymes far less to grip. It works. But the shape it locks in is not fussy about which receptor it activates.

That is the fact from which most of Melanotan II's later history follows. The melanocortin system is not a pigmentation system with side interests. It is a family of five related receptors distributed across skin, brain, adrenal gland and immune tissue, involved between them in pigmentation, appetite, sexual function, inflammation and blood pressure. A compound that does not discriminate between them does several things at once, whether or not anybody wants it to.

Its human record amounts to one small study: a pilot phase-I evaluation published in 1996, which confirmed that the cyclic peptide was active in people and recorded the acute effects that came with it 2. That is what a pilot study is for, and it did its job. Regulated development then stopped. Ordinarily that would be the end of the story, because most compounds that stall at phase I simply subside into the literature. This one did not. It moved into unregulated distribution instead, and the clinical writing that has accumulated since is not the literature of a drug programme but of dermatology clinics: moles darkening and changing outline, new ones appearing in unusual numbers, melanoma diagnosed in users — collated in a review of the risks of unregulated alpha-MSH analogue use 3.

The effect nobody was looking for

The turn in this story came from something no one had designed an experiment to find. During early clinical work on melanocortin agonists, spontaneous erection was recorded as an unanticipated effect — not the endpoint under study, not a pigmentation phenomenon, and not obviously connected to anything the researchers were in the room to measure. It is documented in the melanocortin therapeutics literature as part of how the field developed 4, and in the published account of the compound that came out of it 5. The mechanism, reconstructed afterwards, sits in the brain rather than the skin: the melanocortin receptor subtype concentrated in the central nervous system turns out to have a role in sexual function that had not been appreciated.

This is the point at which the popular version of the story takes over, and it is worth being blunt about that. Retellings circulate featuring a self-experimenting researcher, a named individual, an accident, a date, sometimes dialogue. They are vivid, they are widely repeated, and the published record does not support them. What the record supports is duller and more useful: an unanticipated pharmacological effect, observed during early clinical work, noticed by people who were paying attention, and then deliberately pursued. That is not a lesser story. Serendipity in pharmacology almost never arrives as a thunderbolt. It usually looks like a line in a results table that does not belong there, read by somebody who declines to ignore it.

Pursued is the operative word. A related compound, later named bremelanotide, was taken into a full development programme aimed at sexual function rather than pigmentation 5. That meant the unglamorous sequence: dose-ranging work, randomised placebo-controlled trials, prespecified endpoints, a regulatory submission, review. It emerged at the other end as an approved prescription medicine for hypoactive sexual desire disorder in premenopausal women — a specific, narrow indication, arrived at slowly and at considerable expense. Same laboratory lineage, same receptor family, entirely different destination.

Why one escaped and one did not

• The gap between what a compound does and what a lot of people want it to do.
• The ease of making a short cyclic peptide once its sequence has been published.
• The absence of any mechanism capable of making an existing molecule unavailable.

Take those in order. Bremelanotide treats a diagnosable condition, which means it comes with a gatekeeper by default: somebody has to assess you and write a prescription. Melanotan II's headline effect is a tan. An enormous number of people want one, will pay for one, and do not regard wanting one as a medical matter at all. There is no clinic to route them through, because nobody has ever needed a clinician's permission to be brown. Demand of that shape does not wait for a licence. It goes looking for a supplier.

Then the chemistry, which is the uncomfortable part. A short cyclic peptide is not a difficult molecule. By the 1990s solid-phase synthesis had become routine work for any competent peptide facility, and the sequences were in the public literature, where they belonged. Nothing about making the compound requires the insight that went into designing it. This is not a failure of science; it is science functioning exactly as intended. A structure that is not published cannot be checked, replicated, argued with or improved upon, and a finding only its authors can see is not really a finding.

It is tempting to imagine a counterfactual in which the Arizona group had been more careful about what it let out. That version is worse than it first appears. Restrict the structure and you also restrict the two decades of melanocortin pharmacology built on top of it, including the work that produced an approved medicine. The world in which Melanotan II never escapes is largely the same world in which bremelanotide is never made. There is no obvious way to have kept one and lost the other, and pretending otherwise is a way of avoiding the actual problem.

The actual problem is the third item on the list, and it is the one with no solution on offer. Once a compound exists, has been described in the literature and is straightforward to synthesise, there is no lever anywhere that un-invents it. Regulators can decline to approve it, issue public warnings, seize consignments and prosecute suppliers, and they have done all four. None of that reaches into a facility somewhere and stops the molecule being made. What holds it back is nothing more than the willingness of people to make and sell it, which has proved considerable.

The difference is usually not the molecule

Set the two side by side and the striking thing is how little separates them. The same laboratory lineage, the same receptor family, both short synthetic peptides built on the same natural hormone, both reached by the same intellectual route out of the same desert building. One of them has an approved label, a defined indication, a system that collects adverse events and a denominator to express them against. The other has a pilot study from 1996 and a scattering of case reports written by dermatologists who met its users only after something had gone wrong.

That gap was not created by the molecules. It was created by a process: years of trials, a submission, a review, a decision, and a monitoring system that keeps running long afterwards. The process is slow and expensive and easy to be cynical about, right up to the moment you need to know what a compound does to people over time. The difference between a licensed medicine and an unlicensed research chemical is very often not the molecule. It is whether anyone ran the trials.