Allercur

Clemizole hydrochloride is a first-generation antihistamine from the 1950s (once sold as Allercur) that is being repurposed as EPX-100 for Lennox-Gastaut syndrome (LGS), a severe childhood epilepsy where patients suffer dozens of drop seizures a day (sudden atonic falls) and never respond fully to standard drugs [1]. The asset now sits inside Harmony Biosciences, which acquired originator Epygenix in April 2024 for up to $535M in upfront and milestone payments and added EPX-100 to its commercial CNS portfolio alongside Wakix (pitolisant) for narcolepsy [11]. The Phase 3 study (NCT05066217) is recruiting 260 patients on top of their existing anti-epileptic regimen and reads out on percent change in countable motor major seizures over 28 days [3]. The commercial thesis rests on a mechanism (5-HT2 serotonin receptor activation) that a competitor, UCB's fenfluramine (Fintepla), already validated in the same indication in 2022 [2]. That precedent both de-risks the biology and defines the bar clemizole has to clear.

Clemizole is not a novel molecule. It has decades of human exposure as an oral antihistamine and lost commercial relevance once second-generation, non-sedating antihistamines took over. Harmony Biosciences (NASDAQ: HRMY) acquired Epygenix Therapeutics in April 2024, gaining full rights to EPX-100 and pulling the asset from private-biotech ownership into a public commercial-stage CNS company [11]. Harmony is running two development programs in parallel: NCT05066217, a Phase 3 in LGS enrolling ~260 patients, and NCT04462770 (the ARGUS trial), a Phase 2 potentially-key study in Dravet syndrome enrolling ~150 patients aged 2 and older who carry SCN1A mutations [3][4]. The FDA has granted both Orphan Drug Designation and Rare Pediatric Disease Designation to EPX-100 for both indications, which together lock in seven years of market exclusivity and create eligibility for a Priority Review Voucher on approval [5][11]. A Phase 1 healthy-volunteer PK study (NCT04069689) has already completed [6]. Harmony has publicly guided to topline data from the Dravet ARGUS study in 2026, which will be the first controlled efficacy readout for clemizole and the key input to the LGS Phase 3 investment case [11]. As of mid-2026 Harmony has not disclosed positive Phase 2 efficacy data from ARGUS in press releases, so the LGS Phase 3 is currently running on preclinical zebrafish rationale plus mechanism analogy to fenfluramine, without a public Phase 2 human efficacy readout to anchor it.

In Lennox-Gastaut syndrome, a subset of neurons fire in runaway bursts that the brain's normal braking systems fail to stop, producing multiple seizure types (tonic, atonic, absence) that resist most drugs. Clemizole was rediscovered as an anticonvulsant through a zebrafish screen: Scott Baraban's UCSF lab put larval zebrafish carrying a mutation in scn1a through a chemical library, and clemizole suppressed their seizures [7]. SCN1A encodes a voltage-gated sodium channel; loss-of-function mutations in this gene impair GABAergic inhibitory interneurons in the cortex, leaving excitatory circuits without adequate braking and producing the Dravet phenotype. Follow-up work from Baraban and Griffin showed the anti-seizure effect does not come from H1 antihistamine blockade at all. It comes from partial agonism at serotonin 5-HT2 receptors [8]. Critically for the cardiac safety story below, clemizole is not cleanly 5-HT2A-selective: published pharmacology shows appreciable affinity at both 5-HT2A (the receptor most directly implicated in the anti-seizure effect) and 5-HT2B (the receptor implicated in valvulopathy) [8]. This is the central unresolved scientific question for the program. If clemizole's functional 5-HT2B activity is meaningfully lower than fenfluramine's in intact human tissue, it could differentiate on cardiac safety. If it is comparable, EPX-100 lands in the same REMS bucket as Fintepla. Harmony has not published head-to-head selectivity data addressing this. The relevance of the 5-HT2 target itself is not speculative anymore: fenfluramine, the appetite suppressant pulled off the market in 1997 for cardiac side effects, was rehabilitated as an anti-seizure drug precisely because it hits the same 5-HT2 receptors, and it earned FDA approval in Dravet in 2020 and Lennox-Gastaut in 2022 [2][9]. So the target is well-validated in this exact indication.

NCT05066217 is a randomized, double-blind, placebo-controlled Phase 3 trial in patients aged 2 and older with LGS who are inadequately controlled on at least one background anti-epileptic drug. The design is adjunctive, meaning EPX-100 is added on top of whatever anti-seizure regimen the patient is already taking rather than replacing it [3]. Enrollment target is 260, and the study is actively recruiting as of mid-2026. The primary endpoint is percent change from baseline in CMMS-28 (countable motor major seizures per 28 days) during a titration-plus-maintenance period versus placebo. This is the same style of endpoint the FDA accepted for fenfluramine and cannabidiol in LGS, so there is no novel regulatory risk in the design itself [2][9]. Adjunctive add-on design also mirrors precedent. The Phase 2 ARGUS trial in Dravet (NCT04462770) is the more informative near-term readout: it is a 16-week double-blind period on top of a 4-week observational baseline, in a genetically defined SCN1A-positive population where signal-to-noise is typically better than in LGS [4]. Harmony guides to ARGUS topline in 2026. That number is the first real human efficacy data for clemizole and will heavily determine whether the LGS Phase 3 has a plausible readout at all. Two structural concerns are worth naming for the LGS program. First, LGS is heterogeneous: patients arrive with different underlying etiologies (genetic, structural, cryptogenic), and small trials can be swamped by responder-population effects. Second, placebo response in pediatric epilepsy trials tends to run 15-25% seizure reduction, so clemizole needs a clean separation to hit statistical significance.

Our model estimates a 14% chance this drug is eventually approved. It starts from the historical base rate for Phase 3 drugs in this area (about 51%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is helped by more secondary endpoints than usual; it is held back by the sponsor's thin or weak approval record, weak or limited earlier-phase results, and heavier-than-usual blinding. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

The dominant safety risk is mechanism-based. 5-HT2B receptor activation on heart valve tissue is what drove fenfluramine's 1997 withdrawal for valvulopathy, and Fintepla's current label requires echocardiogram monitoring at baseline, every six months during treatment, and after discontinuation [10]. Because clemizole retains measurable 5-HT2B affinity, comparable REMS-style monitoring is a plausible outcome, which hurts uptake. Sedation is a separate concern given clemizole's antihistamine heritage, and pediatric epilepsy patients tolerate sedation poorly on top of clobazam and other GABAergic drugs. Efficacy risk sits in the implicit head-to-head with Fintepla. In the key Knupp/Lagae 2022 JAMA Neurology Phase 3, the fenfluramine 0.7 mg/kg/day arm produced a median 26.5% reduction in drop seizures versus 7.6% for placebo, for a placebo-adjusted median difference of roughly 19-20 percentage points [12]. Clemizole needs a comparable or better number. If it lands at 10-15% placebo-adjusted with cleaner cardiac safety, the commercial case is still viable but harder. Execution risk has been materially reduced by the Harmony acquisition: Harmony is publicly traded, has an active commercial CNS franchise (Wakix), and has an FDA regulatory track record, replacing the small-private-sponsor risk that applied to Epygenix. Commercial risk remains real. LGS has three approved options already (Fintepla, Epidiolex, rufinamide) plus off-label workhorses, so payers will demand differentiation on either efficacy magnitude or tolerability before granting favorable formulary placement. IP risk is meaningful: clemizole itself is off-patent as a chemical entity, so Harmony's protection relies on the seven-year orphan exclusivity plus any formulation, dosing, or method-of-use patents Epygenix filed. Public disclosures do not detail those filings in depth.

Worth watching, and the acquisition changes the frame. This is no longer a small-private-sponsor repurposing bet. Harmony bought this asset in April 2024 for up to $535M in upfront and milestone payments, and it is now a strategic priority inside a commercial CNS company [11]. The first data point that matters is the ARGUS Phase 2 topline in Dravet in 2026. That reads before the LGS Phase 3 and will drive market perception of the whole program. Specifically watch for: placebo-adjusted seizure reduction of 20% or higher combined with no signal of valvular thickening on echocardiogram. Anything less than that, and clemizole ends up as a second-line alternative to Fintepla rather than a first-line displacer, and the LGS Phase 3 investment case weakens. Harmony's existing Wakix commercial infrastructure gives it a real CNS channel into pediatric neurology and specialty pharmacy, which is a genuine differentiator versus what Epygenix could have executed alone. The M&A section of the thesis is closed, not open: the buyer already arrived. UCB (Fintepla) and Jazz (Epidiolex) remain relevant as competitors defending market share, not as potential acquirers. Check back on this node around the ARGUS 2026 readout, and specifically watch Harmony's investor communications for selectivity data on 5-HT2A versus 5-HT2B that would either substantiate or undercut the cardiac-differentiation thesis. Until then, the score reflects reasonable base-rate uncertainty for a repurposed compound in a validated mechanism where the key human efficacy data are still pending.