A specially tweaked version of the humanin peptide (S14G‑HN) stopped memory loss caused by amyloid‑beta in mice, but it had to be injected straight into the brain, so it isn’t something you can take as a supplement today.

Humanin (HN) is a 24-amino acid peptide that protects neuronal cells from death caused by Alzheimer's disease (AD)-related genes and amyloid-beta (Abeta). Multiple studies have revealed its biochemical and neuroprotective characteristics in vitro; however, little has been known regarding whether HN is effective in vivo in AD model systems. We examined the effect of S14G-HN, a 1,000-fold more potent derivative of HN in vitro, on amnesia induced by Abeta25-35 in mice. The Y-maze test revealed that at least 50 pmol of S14G-HN by intracerebroventricular injection prevented Abeta-induced impairment of short-term/spatial working memory; however, 5 nmol of S14A-HN, a neuroprotection-defective mutant in vitro, did not prevent Abeta-induced amnesia. These results are in agreement with the structure-function correlation shown previously in vitro. In the water-finding task, S14G-HN prevented prolongation of finding latency (the time to find water) observed in Abeta-amnesic mice, indicating that S14G-HN also blocked Abeta-induced impairment of latent learning. In accordance with these observations, immunohistochemical analysis showed that S14G-HN sustained the number of cholinergic neurons in the basal forebrain and the striata nearly to the normal level. Furthermore, genistein, a specific inhibitor of tyrosine kinases, blocked recovery from scopolamine-induced amnesia by S14G-HN, suggesting that certain tyrosine kinase(s) are involved in the inhibitory function of S14G-HN in vivo. Taking these findings together, we conclude that S14G-HN has rescue activity against memory impairment caused by AD-related insults in vivo by activating the same intracellular neuroprotective machinery as elucidated previously in vitro.