The Book of PeptidesThe Book of Peptidesthe sleep factor with no gene
1964 (Monnier & Hösli hemodialysate transfer in rabbits, Basel); 1971-1975 (purification and quantification of "sleep factor delta"); 1977-1978 (sequencing, synthesis and naming of DSIP)
DSIP was named in 1977 for the brainwave it was supposed to produce, and when the question was last reviewed, nearly thirty years on, no gene, precursor or receptor had been found.
In 1964, in Basel, the physiologist Marcel Monnier and his colleague L. Hösli went hunting for a sleep substance. Rather than wait for a rabbit to doze off, they induced it, running current into the donor animal's thalamus until its EEG slid into slow waves. The donor's blood was passed through a dialyser and the small-molecule filtrate, a hemodialysate, a dialysate of the blood rather than the blood itself, was transfused into a rested recipient. The recipient slept more. What made the result worth taking seriously is the arm most retellings drop: filtrate from sham-stimulated donors behaved differently. A 2025 history calls it perhaps the first sleep-factor experiment run with adequate controls.
Thirteen years of grinding chemistry followed. In 1977 Gerhard Schoenenberger and Monnier published the sequence in PNAS: a nine-amino-acid chain, Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu, from rabbit blood. Nine synthetic peptides were infused intraventricularly into 58 rabbits under double-blind conditions; only the full sequence significantly enhanced delta and spindle activity. They named it for that EEG trace, not for behaviour: delta-sleep-inducing peptide. Human work followed fast, though every early positive came from Basel itself. A 1981 double-blind study in six volunteers, co-authored by Monnier and Schoenenberger, reported median total sleep time up 59% against placebo after a 25 nmol/kg intravenous infusion, and by 1983 the same authors claimed 'complete normalization of disturbed sleep.'
Then other laboratories tried. Tobler and Borbély reported in 1980 that systemic DSIP in rats (40-160 nmol/kg) failed to increase sleep or EEG delta power, which in fact fell, and concluded it did not qualify as a specific sleep-promoting substance. The only two double-blind chronic-insomnia studies run outside the Basel circle landed in the same unglamorous place: Monti and colleagues (1987) called the improvement 'of little clinical significance'; Bes and colleagues (1992) recorded higher sleep efficiency and shorter sleep latency than placebo in 16 chronic insomniacs, but judged the effects statistically weak, possibly a placebo-group artefact, with no improvement in subjective sleep quality, and short-term DSIP 'not likely to be of major therapeutic benefit.'
The experiment is real, and better controlled than its retelling, but the popular version misdates it: the transfer was 1964, not 1974, and the naming came in 1977-78. What moved between rabbits was a dialysate of the donor's blood, not whole blood, and the donor's sleep was induced with a thalamic electrode. Nor was it ever confirmed as a sleep factor: no gene, precursor or receptor had been isolated as of the last substantive review, in 2006.
DSIP has never been approved as a medicine in the United States, the European Union or Australia, and no modern clinical trial programme has been completed for it; a 2026 review of therapeutic peptides in orthopaedics lists it among "recovery-enhancing agents" while noting that the whole group lacks trials. Its living literature is largely Russian and preclinical, built around a DSIP-containing preparation called Deltaran: lifespan and tumour-incidence work in mice (2003), an 11-patient pilot in elderly diabetic subjects (2004), motor recovery after focal stroke in rats (2021). Meanwhile the name, coined for an EEG band in a sedated rabbit, circulates through research-chemical channels labelled "not for human use" and marketed for sleep, the one application the two independent insomnia trials judged clinically trivial.
The deeper problem was structural. In a 2006 Journal of Neurochemistry review subtitled 'a still unresolved riddle,' Kovalzon and Strekalova reported that no DSIP gene, precursor protein or receptor had ever been isolated, that DSIP-like immunoreactivity mapped onto hypothalamic neurosecretory nuclei with no obvious sleep role, and that in their own rabbit and rat work certain artificial analogues promoted slow-wave sleep while DSIP itself did not. That remains the last substantive review of the question.