Endorphins, enkephalins and the failure to develop opioid-free pain therapies
In 1975 two groups — in Aberdeen and in San Francisco — isolated the first endogenous opioids: enkephalins and endorphins. The hope of the following decades: a peptide-based pain therapy without addiction potential. 50 years and hundreds of preclinical studies later, no single approved peptide-based opioid alternative exists. A story about mechanistic plausibility and clinical failure — and the complexity of using a pathway that is biologically central.
The shock of 1975: the brain has its own opioids
In the early 1970s Solomon Snyder (Johns Hopkins), Candace Pert and their colleagues, and independently Lars Terenius (Uppsala), established that the brain has specific binding sites for morphine. From a pharmacological view that was a sensation: why should the brain have receptors for a substance that occurs only in the opium poppy capsule? The obvious hypothesis: there must be endogenous ligands that physiologically serve these receptors.
In December 1975 Hans Kosterlitz and John Hughes published in Nature the isolation and sequencing of two pentapeptides from pig brains: Met-enkephalin (Tyr-Gly-Gly-Phe-Met) and Leu-enkephalin (Tyr-Gly-Gly-Phe-Leu). Practically simultaneously the Goldstein group at Stanford and Choh Hao Li in San Francisco isolated β-endorphin from pituitary tissue — a 31-amino-acid peptide with the same Tyr-Gly-Gly-Phe-Met start. These endogenous opioids opened a new research field and a pharmacological hope.
The hope: a painkiller without addiction
The clinical logic seemed clear. Morphine and other opioids act highly effectively against pain, but cause tolerance, respiratory depression, constipation and — most importantly — addictive behaviour. If the brain uses endogenous opioids without the person becoming addicted, pharmacological substitution with body-like peptides should be able to deliver similar analgesia without the addiction component. Research of the late 1970s and 1980s focused on two strategies: stabilised enkephalin analogs and receptor-specific ligands (μ- vs. δ- vs. κ-opioid receptor).
Both strategies produced hundreds of preclinical studies, multiple clinical phase 1 and phase 2 programmes — and not a single FDA approval. The failure was consistent: the substances were either too short-lived (native and lightly-modified enkephalins), not brain-penetrant (charged, polar peptides), or they reproduced exactly the problems of classical opioids (tolerance, addiction, respiratory depression). Selective δ-agonists proved unsuitable in several clinical programmes because of seizure risk; κ-agonists unsuitable because of dysphoria and psychotomimetic effects.
Why does this path fail?
Three structural observations are methodologically valuable. First: opioid receptors are in their physiological function coupled to reward and learning — not only to pain. Every potent μ-agonist activates both axes because they are biologically entangled. A pharmacological separation of analgesia and reward is not achievable by minor sequence variation; it would require fundamental molecular architecture innovation that has not succeeded in 50 years.
Second: peptides have the general problems of their substance class (short half-life, poor CNS penetration, no oral bioavailability, high production cost) in an indication setting that makes these disadvantages particularly unfavourable. A painkiller must be fast, oral, long-acting and cheap — exactly what peptides cannot be. Third: regulatory hurdles for pain indications have substantially risen after the opioid crisis of the 2010s. A new substance mechanistically in the same family as morphine has a much higher safety threshold requirement than 30 years ago.
„We knew in 1975 that the brain has its own opioids. We knew in 1985 that their substitution alone does not solve the addiction problem. We still don't know in 2026 how to pharmacologically separate analgesia from reward."
What the endorphin story shows
Three methodological lessons from this line are valuable for assessing new peptide stories. First: the discovery of an endogenous hormone or neurotransmitter does not automatically mean a therapy axis. Endorphins and enkephalins are biologically real, mechanistically central and pharmacologically highly interesting — but that has not translated into a single approved medicine in 50 years. Second: mechanistic plausibility of a therapeutic concept is a necessary but not sufficient condition. If the endogenous pathway is biologically entangled with unwanted effects, even the endogenous mimic cannot separate these effects. Third: pharmacological research programmes can be productive without leading to clinical approvals. Endorphin research has delivered fundamental insights into reward circuits, addiction biology and pain modulation — even without a single resulting medicine.
In contrast to the GLP-1 success story, the endorphin story is the anti-success — an equally biologically grounded path that has not pharmacologically led to a market. Both stories are necessary for a realistic understanding of peptide pharmacology. The question 'will this peptide become a medicine?' has no predictable answer from biology alone.
Open questions
- Can analgesia be pharmacologically separated from reward — through functionally selective agonists or allosteric modulators?
- What role do biased agonists at the μ-opioid receptor (oliceridine, FDA-approved 2020) play as a compromise solution?
- Are non-opioid pain peptides (ziconotide from cone snail toxin, calcitonin) a better line than modified enkephalins?
- Which lessons from the endorphin failure should be transferred to other 'biologically central pathways' (substance P, bradykinin, galanin) before clinical programmes are launched?