The signal
Deep in the lateral hypothalamus sits a small population of neurons — on the order of tens of thousands, out of roughly 86 billion in the brain — that make a neuropeptide called orexin. It was found twice in 1998, by two labs working independently, which is why it has two names: the group that named it orexin was studying feeding behaviour, and the group that named it hypocretin was cataloguing hypothalamic transcripts. The literature still uses both.
Orexin's job turned out to have less to do with appetite than with holding states in place. The brain has a wake system and a sleep system, and orexin acts as the latch that keeps whichever one is running from flipping. Without it, the boundary between wake and sleep stops being a wall and becomes a curtain.
The receptors
Orexin acts on two receptors, OX1R and OX2R. OX2R is the one that matters most for sleep and wake stability, and it is the one nearly every drug in this class is built to hit. That is why you will see the phrase "OX2R-selective agonist" throughout the pipeline: selectivity for OX2R is a design choice intended to get the wake-stabilising effect without whatever else OX1R does.
An agonist is a drug that switches a receptor on. That single word is the whole distinction between this class and the insomnia drugs with confusingly similar names — those are antagonists, and they switch the same receptors off.
Where the idea came from
The link between orexin and narcolepsy was established in 1999 and 2000 through an unusually clean sequence of results. Dogs from a colony of narcoleptic Dobermans turned out to carry a mutation in the OX2R gene. Mice engineered to lack orexin developed a narcolepsy-like phenotype. Then human patients with narcolepsy were found to have almost no orexin in their cerebrospinal fluid, and post-mortem work showed the orexin-producing neurons themselves were largely gone.
Three species, three methods, one answer. Narcolepsy type 1 is what happens when the orexin neurons die. That made the therapeutic question obvious — put the signal back — and the following twenty-five years were spent on the hard part: finding a molecule that reaches the receptor, can be swallowed rather than infused, and is safe enough to take every day.
Why it took so long
Peptides do not survive the gut and do not cross into the brain, so a straightforward orexin replacement was never an option. The class needed small molecules that mimic a peptide's action at the receptor, which is a genuinely difficult piece of medicinal chemistry.
The first proof that it works in humans came from an intravenous compound, danavorexton, which was never meant to be a product — it was meant to answer whether stimulating OX2R in a narcoleptic person restores wakefulness. It did. What followed was the race to do the same thing in a pill.
That race has already produced one public failure. Takeda's TAK-994 reached Phase 2 and was stopped when liver safety signals emerged. Its successor, oveporexton, became the first drug in the class approved anywhere.
What the class is being aimed at
Narcolepsy type 1 is the obvious target, because that is where the deficiency is. But the pipeline has already spread wider: narcolepsy type 2 and idiopathic hypersomnia, where patients are not orexin-deficient by current measures, and further out to fatigue in multiple sclerosis and Parkinson's disease, to attention-deficit hyperactivity disorder, and to neurodegenerative and neuropsychiatric conditions.
That spread is a bet that orexin signalling is a general lever on wakefulness and attention, not a narcolepsy-specific repair. Whether the bet pays off is the open question this pipeline will answer over the next few years.