ABX-002

ABX-002 (elunetirom) is Autobahn Therapeutics' brain-penetrant thyroid hormone receptor beta (THRβ) agonist, being tested as add-on therapy for depression. Two Phase 2 trials are running in parallel: a 35-patient open-label study in bipolar depression (AMPLIFY-BD) [1] and the 230-patient randomized AMPLIFY trial in major depressive disorder [2]. The FDA granted Fast Track designation for the bipolar indication on May 26, 2026 [8]. Topline guidance is Q2 2026 for AMPLIFY-BD and Q3 2026 for AMPLIFY [8]. The bet is that brain-selective thyroid hormone mimicry can deliver the antidepressant boost clinicians have seen with off-label T3 augmentation, without the cardiac and bone effects of dosing systemic thyroid hormone.

Elunetirom is a novel compound, first-in-class brain-penetrant THRβ-selective agonist. The molecule cleared Phase 1 (NCT05528315, 48 healthy volunteers, completed) [3] and is now running two Phase 2 programs in parallel under Autobahn Therapeutics. The bipolar depression study (NCT06869187, AMPLIFY-BD) is open-label, n=35, primary endpoint HAMD-17 change from baseline, status active-not-recruiting; Autobahn guides topline for Q2 2026 [1][8]. The AMPLIFY trial (NCT06633016) in MDD is double-blind, 1:1 randomized to placebo on top of stable SSRI/SNRI, n=230, 6-week (42-day) treatment period, same HAMD-17 primary endpoint; Autobahn guides topline for Q3 2026 [2][8][9]. FDA granted Fast Track designation for elunetirom in adjunctive treatment of bipolar I/II depression on May 26, 2026 [8]. No breakthrough therapy or orphan designation has been disclosed. Fast Track is a procedural tailwind - more frequent FDA interaction, rolling review eligibility - but does not signal efficacy. Autobahn is privately held with no public guidance cadence and no analyst coverage. Funding is not the concern an early read would suggest: the company closed an oversubscribed $100M Series C in July 2024, led by Newpath Partners with Canaan, Monograph, and Insight Partners joining existing backers including ARCH Venture Partners, Biogen, Bristol Myers Squibb, Pfizer Ventures, and BVF [10]. Total capital raised across four rounds is approximately $228M [10]. Runway through both Phase 2 readouts is plausibly covered, though Autobahn does not disclose burn or cash position publicly. The structurally closest approved analog is resmetirom (Rezdiffra, Madrigal Pharmaceuticals), approved March 2024 for MASH [4]. Resmetirom doesn't meaningfully penetrate the brain - elunetirom is designed to. Same receptor class, different organ; the safety read-through is partial at best.

Thyroid hormone receptor beta (THRβ) is a nuclear receptor - it sits in the cell nucleus, and when thyroid hormone binds it switches on a defined set of genes [5]. There are two receptor isoforms, alpha (THRα) and beta (THRβ). The tissue distribution is more overlapping than often portrayed: THRα dominates in heart, skeletal muscle, and broad regions of adult CNS including cerebellum, cortex, and striatum; THRβ is the predominant isoform in liver, pituitary, hippocampus, and prefrontal cortex [5]. Both are expressed in the adult brain. A brain-penetrant THRβ-selective agonist therefore reduces but does not eliminate CNS THRα exposure - which is part of why the Phase 2 psychiatric safety read (mood instability, agitation, palpitations) matters. The case for THRβ in depression rests on three threads. Clinical: psychiatrists have used T3 (triiodothyronine) for decades as an add-on for patients who don't respond to standard antidepressants, and the STAR*D trial showed T3 augmentation produced modest but measurable benefit in treatment-resistant depression [6]. Genetics: patients with thyroid hormone receptor mutations show neuropsychiatric symptoms, and Open Targets gives THRB an evidence score of 0.79 for selective pituitary resistance and 0.74 for generalized resistance to thyroid hormone. Preclinical: thyroid hormone signaling has been linked to antidepressant-like behavior in rodent models, but Autobahn has not publicly disclosed elunetirom-specific preclinical efficacy data (e.g., forced swim, tail suspension, or chronic unpredictable stress readouts with the compound). The animal-model rationale therefore rests on older thyromimetic literature rather than compound-specific evidence in the public record. The link is real but not overwhelming. T3 augmentation never became a first-line strategy because effect sizes are modest and dosing is finicky - too much and patients get hyperthyroid symptoms; too little and there's no benefit. Elunetirom's pitch is that brain selectivity unlocks a cleaner dose-response curve: mimic T3 augmentation in brain tissue while sparing heart and bone, get the benefit without systemic hyperthyroid baggage. Whether that translates to a Hamilton Depression Rating Scale change patients actually feel is what these two Phase 2 trials are asking.

Quick HAMD-17 primer: scores run 0-52. Above 20 is moderate-to-severe depression; below 7 is remission. A 3-point drug-vs-placebo gap is the rough floor regulators treat as clinically meaningful - roughly the difference between a patient still reporting poor sleep and one sleeping adequately, or between persistent anhedonia and a partial return of interest. The bipolar depression study (NCT06869187, AMPLIFY-BD) is open-label, single-arm, n=35, primary endpoint change from baseline on HAMD-17 [1]. There is no placebo control. Open-label depression trials reliably show large improvements that vanish against placebo - patient expectation, regression to the mean, and clinician rating bias all push scores down. Use this readout for tolerability and biological signal, not for efficacy. AMPLIFY (NCT06633016) is the trial that matters. Phase 2, n=230, randomized 1:1 to elunetirom or placebo as adjunctive to a stable background SSRI/SNRI, 6-week (42-day) treatment period, 2-week post-dose safety follow-up, same HAMD-17 primary endpoint [2][9]. This is where elunetirom either beats placebo on top of standard antidepressants or doesn't. Adjunctive depression studies typically enroll patients with inadequate response to one or more prior antidepressants - exactly the population where T3 augmentation has historical support. A 230-patient Phase 2 is reasonably sized for psychiatry. Big enough to detect a 2-3 point HAMD-17 separation with adequate power; small enough that subgroup analyses will be noisy. Design strengths: validated primary endpoint, adjunctive structure (matches how T3 augmentation has worked historically), 6-week duration consistent with FDA precedent for adjunctive depression RCTs, reasonable sample size. Concerns: Autobahn has not publicly disclosed the elunetirom dose(s) used in AMPLIFY, whether the trial is single-fixed-dose or dose-ranging, or whether biomarker stratification (thyroid panel cutoffs, fMRI substudy, peripheral TSH) is built into the design. Phase 1 SAD/MAD almost certainly included pituitary TSH suppression as a pharmacodynamic readout - TSH suppression at therapeutic exposures is the cleanest target-engagement signal available for a thyromimetic - but Autobahn has not made those PD data public, which is a transparency gap worth flagging.

The model puts this drug's approval odds at 9%. It starts from the historical rate for Phase 2 drugs in this area - about 20% - then adjusts based on ten facts about the trial and sponsor. The biggest drags are the sponsor's thin approval record and weak earlier-phase results; the biggest boosts come from the trial's non-randomized design and open-label setup. The remaining factors are close to average for this stage, so they don't shift the estimate much.

Efficacy is the dominant risk. Adjunctive depression trials are a graveyard. Placebo response in depression studies runs 30-40%, and even well-grounded mechanisms have repeatedly missed. The T3 augmentation literature anchoring elunetirom's thesis is modest in effect size [6]. Translating that to a Hamilton score separation in 230 randomized MDD patients over a 6-week treatment period is genuinely uncertain. Safety risk is mechanism-based. Thyroid hormone affects mood, sleep, appetite, heart rate, and bone turnover. Even with THRβ selectivity, brain penetration means CNS thyroid effects: possible insomnia, agitation, anxiety. In bipolar patients specifically, anything that nudges mood upward must be watched for mania induction. The Phase 1 SAD/MAD study (n=48 healthy volunteers) ruled out gross acute toxicity [3], but 48 subjects cannot characterize rare events or chronic-exposure problems, and Autobahn has not disclosed Phase 1 TSH or peripheral thyroid PD data publicly. Off-target THRα activation is the other safety concern. Selectivity is never absolute, and because THRα is abundant in adult brain regions, residual CNS THRα exposure is a real possibility for any brain-penetrant THR agonist. Sustained THRα activity in the heart causes tachycardia risk; in bone, it accelerates resorption. Resmetirom's MASH program has shown clean cardiac signals [4], but resmetirom is peripheral; brain-penetrant compounds with broader tissue distribution carry more exposure risk. Execution: Autobahn closed a $100M Series C in July 2024 [10], which materially reduces - but does not eliminate - runway concern. Total raised across four rounds is approximately $228M [10]. The company does not disclose cash position or burn rate, so runway through both Phase 2 readouts is plausible but not verifiable from public sources. A private-company funding squeeze before AMPLIFY topline remains a tail risk rather than a base case. Commercial: even with a positive readout, adjunctive psychiatric drugs face brutal payer scrutiny. Spravato (esketamine) launched with breakthrough designation and a novel mechanism and still had a slow uptake curve. Generic SSRIs and SNRIs anchor the formulary at cents on the dollar. A new oral adjunct needs a clear efficacy edge to justify branded pricing - and would still compete with cheap off-patent options like bupropion, mirtazapine, and atypical antipsychotics.

Worth watching closely, not betting on yet. The thesis is mechanistically sound, the sponsor is better funded than it first appears [10], and FDA Fast Track is now in hand for the bipolar indication [8] - but human efficacy data is still zero. AMPLIFY-BD (NCT06869187, n=35, open-label) [1] is the first real efficacy look, with Autobahn guiding topline to Q2 2026 [8]. Don't read too much into a 5-point HAMD-17 drop in an open-label trial; that's normal regardless of drug effect. Pay attention to tolerability, the absence of manic switching, and any biomarker or PD work Autobahn discloses alongside the readout. AMPLIFY (NCT06633016, n=230, randomized 1:1 vs placebo on stable SSRI/SNRI, 6-week treatment) [2][9] is the real data point. Topline guidance is Q3 2026 [8]. If it hits with a clean placebo separation of ≥3 HAMD-17 points, Autobahn becomes a credible BD or IPO candidate and elunetirom enters the depression pipeline conversation seriously. If it misses or shows a 1-2 point trend, the program joins the long tail of plausible-but-unproven psychiatric assets. Check back when: (a) AMPLIFY-BD topline reads out (Q2 2026); (b) AMPLIFY topline reads out (Q3 2026); (c) any conference data - likely ASCP or ECNP - shows brain THRβ engagement in humans via fMRI, EEG, or PET markers; (d) Autobahn discloses Phase 1 TSH suppression data, which is the cleanest target-engagement signal currently absent from the public record. Autobahn is the kind of CNS biotech that either delivers a clean signal and gets bought or quietly winds down. The mechanism is more interesting than most psychiatric pipeline assets - there's an actual genetic and pharmacological rationale, and Fast Track suggests FDA sees the unmet need. But psychiatry's track record means default skepticism is the right posture until placebo-controlled data lands.