FGL (FG loop) — the NCAM-derived peptide for memory

What those early connectivity problems may reflect, at a structural level, is a shift in the molecular machinery that holds synapses together and shapes them — and one of the most important molecules in that machinery is one that almost nobody outside of neuroscience has heard of.

Neural cell adhesion molecule — NCAM — is a large glycoprotein expressed broadly in the nervous system. Its name describes one of its functions: it mediates adhesion between cells, helping neurons recognize one another and maintain the physical relationships that underlie circuit formation. But NCAM does considerably more than glue cells together. It participates in synaptic plasticity, neuronal survival, learning and memory consolidation, and the molecular signaling that underlies long-term potentiation — the cellular process most closely associated with how memories are formed and strengthened. NCAM is, in this sense, not just structural scaffolding. It is an active participant in cognition.

The challenge with NCAM, as with most large membrane-bound proteins, is that you cannot simply administer it as a drug. It is enormous, it doesn't cross the blood-brain barrier, and its adhesion functions are deeply entangled with its signaling functions in ways that make targeted intervention complicated. What Elisabeth Bock's laboratory at the University of Copenhagen recognized was that NCAM's effects on synaptic plasticity were not uniformly distributed across the molecule. Specific regions — loops and segments of the protein — were responsible for specific signaling activities. One region in particular, the FG loop in the second fibronectin type III domain of NCAM, was identified as a key mediator of the molecule's interaction with fibroblast growth factor receptor 1, known as FGFR1.

FGFR1 is a receptor tyrosine kinase — a molecular switch that, when activated, triggers a cascade of intracellular signaling events related to cell growth, differentiation, and survival. In neurons, FGFR1 activation promotes synaptic plasticity, supports the expression of neurotrophins including BDNF (brain-derived neurotrophic factor), and contributes to the molecular processes underlying memory formation. NCAM normally activates FGFR1 through the FG loop, but the full NCAM molecule comes with all of its adhesion functions attached, which makes it unsuitable as a selective therapeutic tool.

The peptide that the Bock lab derived from this insight is FGL — a 15-amino-acid sequence corresponding to the FG loop region of NCAM. FGL is small enough to cross the blood-brain barrier, which immediately distinguishes it from the parent molecule. More importantly, FGL binds to and activates FGFR1 directly, bypassing the cell-adhesion function of full-length NCAM and engaging the downstream signaling cascade that supports synaptic plasticity. It is, in effect, a key shaped to fit one lock out of the many that NCAM normally touches. That specificity is scientifically useful and potentially therapeutically relevant, because it allows researchers to ask what FGFR1 activation alone does — without the noise of adhesion effects.

In animal models, the answer has been consistently encouraging across a range of conditions. FGL-treated rodents show enhanced performance on spatial learning and memory tasks. The compound has been studied in models of Alzheimer's disease, where it has demonstrated neuroprotective effects — reduced neuronal loss, attenuated amyloid-related pathology, preserved memory function relative to untreated animals. In ischemia models — designed to simulate the neurological damage of stroke — FGL administration has been shown to reduce the extent of injury and support functional recovery. In traumatic brain injury research, similar neuroprotective effects have been reported. The breadth of those findings is notable: it suggests that FGFR1 activation via FGL is not narrowly relevant to one disease state but may represent a more general mechanism for supporting neuronal health under conditions of stress or injury.

The delivery question is one of the more practically interesting aspects of FGL research. Many peptides that show promise in preclinical studies run into a fundamental problem: getting them into the central nervous system. Intravenous or oral delivery often means the peptide degrades before reaching meaningful CNS concentrations, and intracranial injection is not a practical clinical approach. FGL research has included work on intranasal delivery — administration via the nasal route, which allows peptides to travel along olfactory and trigeminal pathways into the central nervous system without crossing the blood-brain barrier in the traditional sense. Intranasal delivery is an elegant workaround that has generated interest not just for FGL but for a range of neuropeptides, because it represents a non-invasive route to CNS drug delivery that could be practical in a clinical setting. Studies in rodents have demonstrated that intranasally administered FGL reaches the brain and produces the expected biological effects.

The post-stroke recovery research context is worth addressing specifically because it represents one of the more immediate potential applications. Stroke remains one of the leading causes of acquired disability, and the window for most currently approved interventions is extremely narrow. The observation that FGL supports neuronal recovery in ischemia models — not just at the time of injury but in the post-acute period — positions it as a candidate for a neuroprotection and recovery indication. That is a clinical scenario where the unmet need is clear and the biological rationale for a FGFR1-activating compound is well-grounded.

Human data for FGL, however, is sparse. There are preclinical studies in multiple species, and the biological rationale is solid. But the compound has not progressed through clinical trials in a way that produces published human pharmacokinetic, safety, or efficacy data. This is where honesty about the evidence requires a pause. The animal data is encouraging and, in some respects, mechanistically elegant: a peptide derived directly from a molecule that the brain already uses, targeted to a receptor that normally participates in memory formation, delivered via a route that bypasses the blood-brain barrier. The scientific logic is clear. But scientific logic and clinical evidence are not the same thing, and FGL has not yet generated the latter.

The intranasal delivery route is the most practical research direction for human investigation, and it is the route that researchers interested in FGL's clinical potential have focused on. If FGL is ever to move into human trials at scale, it will likely be via intranasal formulation — a delivery system that is practically feasible, already used for other neuropeptides in clinical contexts, and better tolerated by any conceivable future trial participant than intracranial injection.

What the FGL story illustrates is something broader about how the field of cognitive pharmacology is developing. The most interesting leads are increasingly not coming from broad-spectrum receptor agonists or antagonists — the blunt instruments of an earlier era of neuropharmacology — but from molecules derived from the brain's own signaling vocabulary. FGL is a peptide that the brain, in some sense, invented. The laboratory contribution was to isolate it, shorten it, and figure out how to deliver it. Whether that contribution eventually translates into a clinical therapy depends on the slow, expensive, uncertain machinery of drug development. The science, at the preclinical level, is genuinely interesting. The translation remains open.