NSI-189

NSI-189 is a small synthetic molecule of the benzylpiperazine-aminopyridine type, not a peptide. The US company Neuralstem developed it as a neurogenic compound and supported its first trial in adults with major depressive disorder, which used the salt form NSI-189 phosphate (Fava et al., 2016).

It came from a screen against human fetal hippocampal neural stem cells, and its phase 2 trial did not meet its primary endpoint (Papakostas et al., 2020).

The other identity data come from PubChem CIDs 50922681 and 50922680, except ALTO-100 (Magaraggia et al., 2026).

Published work on NSI-189 falls into six areas. Each summary below states the model used. Apart from two published trials, the evidence is animal and cell data; every animal study of NSI-189 cited here has Neuralstem co-authors.

Discovery and neurogenesis claims

The phase 2 report states that NSI-189 came from screening a chemical library against a human fetal hippocampal neural stem cell line and raised dentate gyrus neurogenesis and hippocampal volume in healthy young adult mice (Papakostas et al., 2020). The phase 1b report cites those mouse findings as Neuralstem data on file, not peer reviewed (Fava et al., 2016). An independent 2026 screen in human neural stem cells found statistically significant but modest proliferation increases (Nakane et al., 2026). Models: human neural stem cell lines and mice.

Stroke and brain irradiation in rats

In male Sprague-Dawley rats after middle cerebral artery occlusion, NSI-189 phosphate by gavage daily for 12 weeks from 6 hours after the stroke (30 mg/kg as free base) significantly reduced motor and neurological deficits through 24 weeks and increased MAP2-labeled neurite outgrowth, mainly in the hippocampus; in cultured hippocampal cells it reduced cell death after oxygen-glucose deprivation (Tajiri et al., 2017).

After fractionated head-only irradiation, rats given NSI-189 daily for four weeks scored significantly better than irradiated controls on four recognition memory tasks, with more neurogenesis, fewer activated microglia and no major change in hippocampal volume (Allen et al., 2018). Models: rats and cultured cells.

Diabetic neuropathy models

In cultured adult rat sensory neurons, NSI-189 increased neurite outgrowth and mitochondrial function. Given orally to female mice with type 1 and male mice with type 2 diabetes, it prevented multiple signs of small and large fiber neuropathy, increased hippocampal neurogenesis, synaptic markers and volume, protected long-term memory and halted progression of established neuropathy (Jolivalt et al., 2019).

In male Zucker diabetic fatty rats, 16 weeks of NSI-189 phosphate begun after 16 weeks of diabetes raised cortical mitochondrial complex activity, reversed several nerve indices and lessened memory deficits (Jolivalt et al., 2022). Models: neuron cultures, diabetic mice and rats.

Synaptic plasticity and the TrkB question

In mouse hippocampal slices, NSI-189 increased long-term potentiation, and a short daily course reversed cognitive and motor impairments in a mouse model of Angelman syndrome; the authors linked the effects to TrkB and Akt activation (Liu et al., 2019). When a drug-discovery group retested published TrkB activators, NSI-189 did not bind the TrkB extracellular domain, trigger TrkB phosphorylation or modulate it allosterically in TrkB-expressing cells (Pankiewicz et al., 2021). No paper cited here names a direct molecular target. Models: brain slices, mice, receptor protein and cell lines.

Trials in major depressive disorder

The phase 1b study randomized 24 inpatients with recurrent major depressive disorder to three ascending cohorts of six on NSI-189 phosphate and two on placebo for 28 days. Its main aims were safety and pharmacokinetics; it was reported as relatively well tolerated, and on exploratory scales it separated significantly from placebo on the self-rated SDQ and CPFQ but not on the Montgomery-Asberg Depression Rating Scale (MADRS), with no significant difference in MRI hippocampal volume; the authors flagged the small, unbalanced sample (Fava et al., 2016).

In the phase 2 trial, 220 outpatients were randomized to placebo or one of two daily amounts for 12 weeks. The primary endpoint was not met: MADRS change did not differ significantly from placebo for either amount (p = 0.22 and 0.34). The lower amount separated on the SDQ, CPFQ and some CogScreen measures, and both amounts were well tolerated (Papakostas et al., 2020). Models: randomized human trials.