AppliedXL Response to FDA Drug Repurposing Initiative: 5 Drugs With Clinical Evidence and No Commercial Champion

Binds the alpha-2-delta subunit of voltage-gated calcium channels in thermoregulatory neurons of the hypothalamus, dampening the aberrant calcium channel activity that drives the neurogenic vasodilation and perspiration cascade underlying hot flashes. Distinct from the NK3/neurokinin mechanism (fezolinetant) and the serotonergic mechanism (paroxetine) — mechanistically complementary in patients who fail those approaches.

The mechanistic bridge from neuropathic pain to vasomotor symptoms runs through a shared neuropeptide: Substance P. Gabapentin binds CACNA2D1 with nanomolar affinity (Kd ~100–200 nM; Gee et al., 1996), reducing calcium-dependent neurotransmitter release from presynaptic terminals. Among the neurotransmitters suppressed is Substance P, which plays a well-characterized role in nociception — but also participates directly in hypothalamic thermoregulation. Substance P-expressing neurons in the arcuate nucleus are activated by estrogen withdrawal; their projections to the median preoptic nucleus directly trigger vasodilation and sweating. Gabapentin's calcium channel inhibition reduces Substance P release at both nociceptive and thermoregulatory synapses — a single mechanism serving two indications through the same intermediate node.

The pathway is traceable: gabapentin inhibits CACNA2D1, which reduces calcium ion import, which modulates vasodilation — the same vascular instability that drives vasomotor flushing. CACNA2D1 is expressed in arterial tissue (GTEx: aorta 3.88, coronary 3.79, tibial 4.21 TPM), though at lower levels than skeletal muscle (5.26 TPM), suggesting the central hypothalamic mechanism likely dominates over direct peripheral vascular effects.

Potentiates GABA-A receptor activity and antagonizes AMPA/kainate (glutamate) receptors, directly suppressing the mesolimbic dopamine reward pathway that mediates alcohol craving and reinforcement. This dual mechanism — enhancing inhibitory tone while dampening excitatory glutamatergic drive — targets the neurobiological substrate of compulsive alcohol seeking more directly than the approved AUD agents (naltrexone, acamprosate, disulfiram).

These two mechanisms converge on the mesolimbic reward circuit through parallel but independent pathways. The GABAergic arm enhances inhibitory tone, reducing the anxiety that drives stress-induced drinking. The glutamatergic arm, through kainate receptor antagonism, dampens excitatory drive to dopaminergic neurons, directly suppressing the reward response to alcohol. Both pathways terminate in altered dopamine and glutamate secretion, the core neurotransmitter systems of the mesolimbic circuit — and both are constitutively active, meaning topiramate's effects accumulate across drinking episodes rather than requiring acute dosing coincident with craving.

The GRIK1 rs2832407 variant adds a molecular basis for treatment response: GRIK1 encodes the GluK1 kainate receptor subunit, which is a direct target of topiramate's glutamate antagonism. Patients carrying specific GRIK1 variants may have altered GluK1 receptor function that increases sensitivity to topiramate's antagonism — a pharmacogenomically enriched population that could define a precision label.

One mechanistic nuance warrants attention. GABRA1 expression is highest in the cerebellum (5.82 TPM) and cortex (5.01 TPM), but lowest in the caudate (2.51 TPM) — a key mesolimbic structure. This expression pattern suggests topiramate's therapeutic effects in alcohol use disorder may operate more through cortical cognitive control and anxiety reduction than through direct suppression of striatal reward signaling. The clinical implication: topiramate may be most effective in patients where cortical dysregulation — impulsivity, anxiety-driven drinking — dominates over purely reward-driven compulsion.

PPARγ agonist; enhances insulin sensitivity, reduces neuroinflammation via microglial modulation, decreases tau phosphorylation, improves mitochondrial function. PPARγ activation corrects cerebral insulin signaling and suppresses microglial-driven neuroinflammation through NF-κB pathway inhibition, directly addressing the insulin-resistance/cerebrovascular axis biologically enriched in the post-stroke population.

Molecular docking confirms that pioglitazone binds PPARγ with the strongest affinity of any drug-target pair in this analysis (8.86 kcal/mol). PPARγ is not a narrow single-pathway target — it functions as a master transcriptional regulator with unusually high network connectivity (degree 6.7, the highest of any target examined here), meaning its activation cascades through multiple downstream pathways simultaneously.

The metabolic-neurological bridge operates through at least three convergent mechanisms. First, PPARγ activation shifts microglial phenotypes from pro-inflammatory M1 to anti-inflammatory M2 states (Heneka et al., Nature 2015), reducing production of neurotoxic IL-1β and TNF-α. Second, improved insulin signaling inhibits GSK-3β, directly reducing tau phosphorylation at pathogenic Ser396/404 epitopes — with 40–60% reductions demonstrated in 3xTg-AD mice (Sato et al., J Neurosci 2011). Third, PPARγ activation upregulates PGC-1α, restoring mitochondrial biogenesis in neurons — a mechanism independent of, and additive to, the vascular and neuroinflammatory pathways.

PPARγ is already linked to cerebrovascular disease in established mechanistic databases — the bridge to the post-stroke population is not a theoretical leap but a documented connection. The further link from cerebrovascular pathology to dementia, through shared nodes including memory impairment and acetylcholine neurotransmission, is where the IRIS trial's clinical signal provides the human evidence for what the biology predicts.

Selective estrogen receptor modulator (SERM) blocking hypothalamic estrogen negative feedback, increasing GnRH pulsatility and downstream LH/FSH secretion, stimulating Leydig cell testosterone production. In obese men, peripheral aromatization of testosterone to estradiol creates excessive estrogen negative feedback — a functional hypogonadotropic hypogonadism mechanistically identical to the pituitary axis dysregulation clomiphene corrects in anovulatory women.

The biological plausibility of this cross-sex application is stronger than it might appear. ESR1, the estrogen receptor alpha that clomiphene modulates, has the highest protein interaction network connectivity of any target examined here (degree 7.0) and a probability of loss-of-function intolerance (pLI) of 1.00 — the maximum possible score, indicating extreme evolutionary conservation. This is a gene the body cannot afford to lose, consistent with its master regulatory role in reproductive biology across vertebrates.

Tissue expression data confirms the anatomical basis: ESR1 is expressed in the pituitary at 2.77 TPM, comparable to ovarian tissue at 3.12 TPM. The receptor is present in the same hypothalamic-pituitary structures in both sexes — the difference is not the receptor, but the hormonal milieu acting on it. Clomiphene's high lipophilicity (LogP 6.56) ensures blood-brain barrier penetration for hypothalamic access. The literature evidence linking ESR1 modulation to hypogonadism is the strongest of any candidate-indication pair in this analysis (score 0.874) — a formal assessment that the biological connection between the drug's target and the proposed disease is supported by published literature.

The absence of this mechanism from curated pharmacological databases — despite decades of clinical use and AUA guideline endorsement — reflects a systematic blind spot: cross-sex-indication mechanisms are underrepresented in databases built primarily from indication-specific drug development programs.

AMPK activator; mTORC1 inhibition suppresses anabolic tumor-growth signaling; reduction of IGF-1 signaling and circulating insulin levels (a cancer growth factor). The T2DM population was the index studied population — cancer incidence was observed to be significantly lower than expected across multiple independent cohorts.

Docking metformin against AMPK yields a binding energy of 4.08 kcal/mol — the weakest target engagement of any drug-target pair in this analysis. For context, pioglitazone scores 8.86 against PPARγ. Binding energies below 6 kcal/mol typically indicate weak or non-specific interactions.

This weak binding is itself the finding. The emerging consensus is that metformin does not directly bind and activate AMPK. Instead, the mechanism operates through an indirect cascade: metformin enters the cell via organic cation transporters and inhibits mitochondrial complex I of the electron transport chain. Complex I inhibition disrupts ATP synthesis, raising the cellular AMP:ATP ratio. The elevated AMP:ATP ratio activates AMPK allosterically — as a metabolic stress signal, not as a direct ligand interaction. Wheaton et al. (eLife 2014) confirmed this by showing that metformin's anti-tumor effects are abolished in cancer cells expressing a metformin-resistant form of complex I (NDI1), establishing mitochondrial complex I as the necessary upstream target.

This distinction matters. If metformin's anti-cancer effect operated through direct AMPK binding — a single receptor interaction — one would expect organ-specific activity dependent on local AMPK density. Instead, the indirect mechanism through mitochondrial energetics explains why metformin's cancer risk reduction spans multiple organ sites (colorectal −25–40%, breast −15–25%, endometrial −30–40%): mitochondrial complex I is ubiquitous, and the downstream metabolic signaling it disrupts is constitutively active in proliferating cancer cells regardless of tissue type.