Peptides for Spinal Cord Injury and Stroke Recovery: Phase 3, Gray Market, and the Evidence Map

We replaced unrelated references, corrected the CARS endpoint and neutral CASTA result, distinguished rat BPC-157 experiments from canine claims, and corrected the mistaken description of Exenatide-PD3 as a positive trial.

Walk into any spinal cord injury (SCI) support group online and the same questions come up. "Anyone trying BPC-157 for nerve recovery?" "Is TB-500 worth it for paralysis after a year?" "Cerebrolysin protocols for chronic stroke?" "What about the new NervGen drug?" The same thread plays out in stroke support communities and traumatic brain injury (TBI) forums.

These aren't fringe questions. SCI affects roughly 300,000 Americans, with about 18,000 new cases each year. Stroke affects 795,000 Americans annually, and a third of survivors live with significant long-term motor or cognitive deficits. TBI affects another 1.5 million Americans each year. Standard rehab improves outcomes but rarely restores function once the acute window closes. And "once the acute window closes" is exactly when patients start searching for anything that might still help.

The peptide space for neural recovery has one program with genuinely strong recent clinical data, one drug that's been approved abroad for decades, a couple of compounds with intriguing preclinical biology but no Phase 3 trial in this space, and a long tail of gray-market wellness stacks. This article maps that picture against what the published evidence actually shows.

The peptide strategies differ significantly by injury type. Being specific matters:

• Acute ischemic stroke: Blood clot blocks a brain artery, neurons die from oxygen deprivation. The therapeutic window is 0-24 hours (tPA, thrombectomy). Peptide strategies target neuroprotection during this window.
• Chronic post-stroke recovery: Weeks to years after the initial event. Goal is to support neuroplasticity, axon sprouting, and functional reorganization. This is where most peptide research now sits.
• Acute spinal cord injury: Mechanical trauma severs or crushes the cord. Secondary injury (inflammation, glial scar) unfolds over days to weeks. Peptide strategies aim to limit secondary injury.
• Chronic spinal cord injury: Months to years post-injury. The blocking biology is the glial scar and the chondroitin sulfate proteoglycans that prevent axon regrowth. This is the NVG-291 indication.
• Traumatic brain injury: Diffuse axonal injury plus regional damage. Recovery biology overlaps both stroke and SCI.
• Peripheral nerve injury: Carpal tunnel, sciatic, brachial plexus. Different biology (axons can regrow more readily here). Some peptide work in DoD-funded animal studies.

The peptide most relevant to chronic SCI is not the peptide most relevant to acute stroke, and the wellness forum often blurs these together. The sections below try to keep the mapping clear.

The single strongest peptide-for-neural-recovery clinical result is from NervGen Pharma's NVG-291, a subcutaneous peptide mimetic designed to inhibit protein tyrosine phosphatase sigma (PTPσ). The mechanism is straightforward. Chondroitin sulfate proteoglycans (CSPGs) accumulate in the glial scar after spinal cord injury and bind to PTPσ on axons, locking the axons in place and preventing regrowth. NVG-291 mimics the intracellular sigma peptide (ISP) of NgR1, blocking that interaction and freeing the axons to sprout.

The Phase 1b/2a CONNECT SCI Study readout in chronic SCI showed something the field has not seen before. In the chronic cohort (n=10 active, n=10 placebo), NVG-291-treated subjects had a three-fold increase in motor connectivity to the first dorsal interosseus hand muscle (a sensitive marker of corticospinal tract function), measured by transcranial magnetic stimulation. Owner-rated and clinician-rated function scores moved in the same direction.

NervGen completed its End-of-Phase 2 FDA meeting and is on track to initiate the RESTORE registrational Phase 3 study in chronic tetraplegia in mid-2026. Blinded biomechanical gait analysis from the chronic CONNECT SCI cohort is expected in Q2 2026 to disentangle genuine neural recovery from compensatory movement. NervGen's Q1 2026 cash was $16.6M (Nasdaq: NGEN, listed January 8, 2026). The DoD has also funded preclinical work on NVG-291-R in traumatic hearing loss and peripheral nerve injury.

The caveats are real. The CONNECT SCI Phase 1b/2a is a small study (n=10 active in the chronic cohort), and the RESTORE Phase 3 has not yet started. Phase 3 trials have buried plenty of promising peptide programs that looked strong at Phase 2. But for chronic tetraplegia, where no pharmacological intervention has ever shown durable motor recovery in humans, this is the most credible peptide program in development.

Cerebrolysin trials have produced different results in different stroke settings. The CARS trial studied early rehabilitation after stroke. Its primary endpoint was the Action Research Arm Test at day 90, not the NIH Stroke Scale. The authors reported improvement but called for confirmation in a larger trial.

The CASTA trial enrolled 1,070 patients with acute ischemic stroke. Its prespecified combined endpoint showed no significant difference between Cerebrolysin and placebo. Favorable findings in more severely affected patients came from post hoc analysis and cannot replace the neutral primary result.

These trials do not establish efficacy for chronic spinal-cord injury. The previous blanket statement that importing a 90-day supply is legal was also unsupported and has been removed.

Thymosin beta-4 is a naturally occurring 43-amino-acid peptide with well-documented roles in cell migration, angiogenesis, anti-inflammation, and tissue repair. TB-500 is the synthetic version of the active fragment, sold widely through research-chemical and wellness channels.

The preclinical neural-recovery biology is interesting. Thymosin beta-4 has been shown to promote neurogenesis, support oligodendrocyte progenitor cell migration (important for remyelination), and reduce neuroinflammation in animal models of stroke, TBI, and SCI. Multiple rodent stroke studies have shown improved functional outcomes with thymosin beta-4 administration after middle cerebral artery occlusion.

The clinical evidence is much thinner. HLB Therapeutics' US subsidiary ReGenTree developed RGN-259, a thymosin beta-4 eye drop, for neurotrophic keratitis (an ocular indication, not neural recovery in the brain or cord). The European Phase 3 SEER-3 trial missed its primary endpoint of complete corneal healing at 4 weeks, with a stronger-than-expected placebo response carrying the control arm. The US Phase 3 SEER-2 is ongoing with H2 2026 readouts. RegeneRx (the originator) tested thymosin beta-4 in acute MI for stem cell priming, a cardiac application rather than neural recovery. No completed Phase 2 or Phase 3 trial of thymosin beta-4 in humans exists for spinal cord injury, ischemic stroke, or traumatic brain injury.

Use pattern in the wellness market is heavy. TB-500 protocols circulate widely among working-dog trainers, recovery-focused athletes, and SCI patient communities, usually at 5-20 mg per week subcutaneous. The forum case-report sentiment is favorable. The published clinical-trial evidence for neural recovery in humans is absent. If you use TB-500 for neural recovery, you are operating on rodent data and extrapolation, and the safety profile of chronic administration is not well characterized.

The verified BPC-157 spinal-cord injury paper studied rats. The previous version described canine results that this source does not support. Animal findings cannot establish recovery of motor or bladder function in people with spinal-cord injury.

Small human BPC-157 reports in other settings also cannot establish neural recovery. The 2025 intravenous pilot enrolled only two adults and did not assess stroke or spinal-cord outcomes. Its brief observations cannot establish general safety.

The 2017 exenatide trial was an earlier Parkinson's study, not Exenatide-PD3. The subsequent Phase 3 trial published in 2025 found no evidence supporting exenatide as a disease-modifying treatment for Parkinson's disease.

Neither study establishes a treatment for spinal-cord injury or post-stroke rehabilitation. A cardiovascular composite endpoint in a separate semaglutide trial also cannot establish recovery from an existing stroke.

Semax and Selank come from a different research tradition. Both were developed at the Institute of Molecular Genetics of the Russian Academy of Sciences and are approved in Russia for stroke recovery, cognitive enhancement, and anxiety disorders. Neither has FDA approval.

Semax is a 7-amino-acid analog of ACTH (4-10) with neuroprotective and nootropic claims. Russian and Eastern European clinical trials have reported improvements in stroke recovery, attention, and cognitive function. The trial quality varies, the sample sizes are usually small, and the work has not been replicated in Western multicenter trials. Western neuroscience laboratories have validated some of the mechanistic claims (BDNF upregulation, modulation of cholinergic and serotonergic systems), but the clinical claims rest on a Russian literature that doesn't easily translate into FDA-approval territory.

Selank is a 7-amino-acid analog of tuftsin with anxiolytic and immunomodulatory claims. Similar pattern: approved in Russia, used in Eastern European clinical practice, limited Western clinical-trial data.

Both peptides have a following among biohacker and nootropic communities. The sourcing problem is severe. Most Semax and Selank sold in Western markets comes through research-chemical channels with no quality control. The FDA does not affirmatively permit either for any indication. They are on the broader peptide-reclassification radar but not on the specific July 23-24, 2026 PCAC docket.

If you use Semax or Selank for post-stroke recovery, you're operating on a real but non-FDA-aligned evidence base, with significant sourcing-quality risk and no medical-supervision framework in the US.

Several peptides circulate in SCI, stroke, and TBI forums with thinner evidence than the compounds above.

IGF-1 LR3 is a modified insulin-like growth factor with extended half-life. Some preclinical neural-protection data exists, but it's used primarily in bodybuilding for hypertrophy. The theoretical risk of tumor promotion has kept it out of serious neurology development pipelines.

GH secretagogues (sermorelin, CJC-1295, ipamorelin) come up in wellness stacks for general 'nerve health.' The clinical neural-recovery evidence is weak. Acromegaly (excess GH) is associated with peripheral neuropathy, suggesting the relationship between GH and nerve function is more complicated than 'more GH = more recovery.'

KPV (the C-terminal tripeptide of α-MSH) has anti-inflammatory activity and shows up in stacks for general neural recovery. Most of the preclinical evidence is in IBD and skin rather than neural tissue. No human SCI or stroke trial.

NAP / Davunetide is an 8-amino-acid peptide derived from the ADNP gene that showed cognitive improvement in early Alzheimer's and MCI trials but failed Phase 3 in progressive supranuclear palsy in 2013. Not currently in active clinical development for stroke or SCI.

Mostly, the wellness peptide stack assembled for neural recovery is a collection of compounds with reasonable general repair mechanisms and minimal direct evidence in the neural-injury indication that brought the patient to the conversation.

Neural recovery is harder than musculoskeletal recovery for biological reasons that haven't changed. Mature neurons largely don't replicate. Axons can sometimes regrow in the peripheral nervous system but face powerful inhibitory signals in the central nervous system (myelin debris, chondroitin sulfate proteoglycans, the glial scar). The blood-brain barrier limits drug delivery to brain tissue. The blood-spinal-cord barrier creates similar problems for SCI.

Plasticity windows matter. The acute injury phase (hours to weeks) has different therapeutic targets than the subacute phase (weeks to months) and the chronic phase (months to years). Most rehabilitation protocols target subacute plasticity. NVG-291's chronic-SCI focus is novel because the biology of overcoming established glial scarring is different from the biology of supporting subacute plasticity.

Functional outcomes are hard to measure. A small motor connectivity improvement might or might not translate to real-world function. Blinded biomechanical gait analysis (NervGen's Q2 2026 deliverable for CONNECT SCI chronic cohort) is one attempt to disentangle genuine recovery from compensatory movement. Patient-reported outcomes are critical but vulnerable to placebo effects, as the SEER-3 trial of RGN-259 showed.

The research-chemical wellness market has filled the gap left by slow clinical development. That market sometimes captures real biology that academic medicine has neglected, and it sometimes captures hopeful speculation that crumbles under scrutiny. Telling those apart in real time is the hardest part of being a patient looking for help.

If you or someone you know is dealing with neural injury and considering peptides, the right call depends on the clinical situation.

For someone with chronic spinal cord injury (cervical or thoracic, more than a year post-injury), NVG-291 trial enrollment is the strongest available option. The RESTORE registrational Phase 3 in chronic tetraplegia is set to initiate mid-2026. ClinicalTrials.gov should list enrollment criteria as the trial opens. Outside trial enrollment, the options are conservative: continued rehab, electrical stimulation protocols (epidural stimulation has its own promising literature), and pre-clinical-rationale peptide use under medical supervision rather than research-chemical sourcing.

For someone with chronic post-stroke deficits, Cerebrolysin under a knowledgeable functional-medicine or integrative-neurology clinician is the option with the most clinical evidence. Personal-importation logistics and quality sourcing matter. GLP-1 RAs may offer secondary neuroprotective benefits if cardiometabolic indications also apply. BPC-157, TB-500, and Semax sit at the edges of the evidence base.

For someone with traumatic brain injury, the case for Cerebrolysin is somewhat stronger than for stroke because more TBI trials have been positive, though the methodological quality is debated. Standard cognitive rehabilitation, hyperbaric oxygen (mixed evidence), and treatment of comorbidities (sleep, mood, hormonal) matter more than peptide layering.

For someone with peripheral nerve injury (post-surgical, post-trauma, carpal tunnel), the biology is friendlier because peripheral nerves can regrow. NVG-291's DoD-funded preclinical work is exploring this. BPC-157 has more preclinical peripheral-nerve data than central nervous system data.

A few universal principles. Medical supervision matters even more than in arthritis: neural injuries are heterogeneous and the diagnosis often determines treatment. Self-directed wellness stacks without a neurologist's input are a higher-risk strategy here than in musculoskeletal recovery. Sourcing quality matters: research-chemical-grade compounds for chronic injection are a documented contamination risk. Track functional outcomes objectively where possible (strength tests, walking velocity, validated cognitive batteries, patient-reported scales like SCIM or modified Rankin) to distinguish real change from hope.

Research needs to be matched to the injury and outcome it studied. An exploratory stroke-rehabilitation trial, a neutral acute-stroke trial, a rat spinal-cord experiment, and a Parkinson's trial answer different questions. They do not establish a general peptide treatment for neural recovery.

Company announcements about investigational programs should be read as development updates. Decisions about rehabilitation or trial participation belong with the treating neurological and rehabilitation team.