AB-1002

AB-1002 is an AAV gene therapy from AskBio (a Bayer subsidiary) that delivers a hyperactive version of a natural protein called inhibitor-1 into heart muscle cells to restore their calcium handling in advanced heart failure. The Phase 1 trial (NCT04179643) in 17 patients with NYHA Class III heart failure (symptoms with light activity like dressing or walking short distances, the second-most severe functional category and one step from end-stage) has completed enrollment, with results published in Nature Medicine in 2025 [1]. The Phase 2 study (NCT05598333, called GenePHIT) enrolled 150 non-ischemic heart failure patients into a randomized, double-blind, placebo-controlled design with four co-primary endpoints at 52 weeks including cardiovascular-related death; primary completion is estimated October 31, 2026 [2]. The commercial stakes are meaningful because this is one of the few remaining serious attempts to make cardiac gene therapy work after MYDICAR (AAV1-SERCA2a) collapsed in Phase 2b, and Bayer paid approximately $2 billion upfront with up to $2 billion in contingent milestone payments for AskBio in 2020 partly to own programs like this [3].

AB-1002 is a novel biologic, first-in-human, with no prior approvals anywhere. The Phase 1 trial (NCT04179643) is a non-randomized, sequential dose-escalation study of 17 patients that is now listed as active but not recruiting, with intracoronary infusion into cardiomyocytes as the delivery route [4]. The ClinicalTrials.gov primary outcome measures are change from baseline in peak VO2 (a measure of how much oxygen a patient can use during peak exercise, and the standard functional endpoint in heart failure trials) and echocardiographic left ventricular ejection fraction (LVEF), with safety and tolerability inherent to any dose-escalation design [4]. The Phase 2 study (NCT05598333, GenePHIT) is a randomized, double-blind, placebo-controlled trial with 150 patients allocated 1:1:1 across two AB-1002 dose arms and placebo. It is now listed as active, not recruiting, meaning enrollment is complete, and it has four co-primary endpoints assessed at 52 weeks: cardiovascular-related death, NYHA classification change, LVEF change, and 6-minute walk test distance [2]. Estimated primary completion is October 31, 2026. No breakthrough therapy, fast track, RMAT (Regenerative Medicine Advanced Therapy, an FDA designation that accelerates development and review of cell and gene therapies for serious conditions), or orphan designations have been publicly disclosed for AB-1002. Nature Medicine published the Phase 1 results in 2025, which is the first public data drop and matters because gene therapy programs often stall between Phase 1 and Phase 2 publication [1]. A key readout timeline has not been publicly guided by AskBio or Bayer beyond the ClinicalTrials.gov primary completion estimate.

In a failing heart, muscle cells cannot pump calcium in and out of their internal storage compartment (the sarcoplasmic reticulum) fast enough. That calcium recycling is what makes each heartbeat strong. The pump that does the work is called SERCA2a, and it sits with a brake pedal on top of it called phospholamban. When phospholamban is phosphorylated (has a phosphate group stuck on it), the brake releases and SERCA2a runs hard. An enzyme called protein phosphatase 1 (PP1) removes that phosphate and slams the brake back on. Inhibitor-1 is the body's natural block on PP1. In heart failure, inhibitor-1 activity drops, PP1 runs unchecked, phospholamban stays unphosphorylated, and SERCA2a stays braked [5]. AB-1002 uses an AAV2i8 virus (a bioengineered viral shell that preferentially delivers cargo to heart tissue) to install a gene for a constitutively active form of inhibitor-1 (I-1c) that cannot be shut off. The result should be sustained inhibition of PP1, sustained SERCA2a activity, better calcium cycling, and stronger contractions [5]. The mechanistic case is biologically clean and supported by decades of calcium-handling work, but human validation is thin because the closest prior attempt, MYDICAR from Celladon (AAV1 delivering SERCA2a itself), failed its Phase 2b CUPID-2 trial in 2016 [6].

Phase 1 (NCT04179643) is a small, non-randomized, sequential dose-escalation study in 17 patients with NYHA Class III heart failure receiving intracoronary AB-1002 infusion at two dose levels, with a third arm for patients carrying the PLN-R14Del mutation. The ClinicalTrials.gov primary outcomes are change from baseline in peak VO2 and echocardiographic LVEF, with 6-minute walk test as a key secondary [4]. There is no comparator arm and no randomization, which is standard for a Phase 1 gene therapy but limits how much can be inferred about efficacy. Phase 2 (NCT05598333, GenePHIT) is a substantially more serious study: 150 patients with non-ischemic (meaning heart failure not caused by blocked coronary arteries) NYHA Class III disease, randomized 1:1:1 to low-dose AB-1002 (7.15E13 vg), high-dose AB-1002 (1.43E14 vg), or placebo intracoronary infusion, with the study double-blinded [2]. Cardiovascular-related death is one of four co-primary endpoints at 52 weeks alongside NYHA class change, LVEF change, and 6-minute walk distance. Using a hard clinical event as a co-primary at Phase 2 is unusually ambitious; with only about 50 patients per arm, the trial will need either a strong effect size or a very sick population to hit statistical significance on the mortality endpoint, though the multi-component primary and the 52-week window help. AskBio is the sponsor of both trials. The non-ischemic restriction is smart because prior cardiac gene therapy work suggests ischemic scarred myocardium takes up AAV poorly, so limiting to non-ischemic patients biases toward responders.

Our model estimates a 4% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 27%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is held back by heavier-than-usual blinding, the sponsor's thin or weak approval record, weak or limited earlier-phase results, and a randomized design. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Efficacy risk is the largest. MYDICAR (AAV1-SERCA2a) failed Phase 2b in 2016 targeting the same calcium-handling pathway, and the community consensus was that too little transgene actually reached enough cardiomyocytes to matter [6]. AB-1002 uses a different capsid (AAV2i8, engineered for cardiotropism) and a different transgene (I-1c amplifies the effect by inhibiting PP1 rather than directly delivering SERCA2a), but the delivery-efficiency problem is not solved just because the biology is smarter. Safety risk includes the standard AAV liabilities: transaminase elevations, complement activation, and rare but serious hepatic and thrombotic microangiopathy events at high doses. Cardiac AAV dosing is intracoronary rather than systemic, which theoretically lowers off-target exposure. Immunogenicity is a hard exclusion criterion because roughly a third of adults carry pre-existing neutralizing antibodies to common AAV serotypes, shrinking the eligible population. Re-dosing is not feasible once neutralizing antibodies develop post-administration, so if the therapeutic effect attenuates over years, patients have no retreatment option. That is a structural liability oral heart failure drugs (SGLT2i, ARNIs, MRAs, beta-blockers) do not share. Phase 1 durability data specifically addressing how long any functional improvement persists is a key gap given this re-dosing constraint. Execution risk is real because enrolling advanced heart failure patients who meet AAV-naive criteria is slow, further shrinking the addressable population beyond the non-ischemic NYHA III subset. Commercial risk is severe even if AB-1002 works. Guideline-directed medical therapy for heart failure is cheap, oral, and increasingly effective. A one-time gene therapy priced in the multiple-hundreds-of-thousands to low-millions range will face aggressive payer pushback unless it demonstrates a mortality benefit that oral drugs cannot match. For scale, NYHA Class III heart failure carries roughly 50% 5-year mortality without transformative treatment, which is what makes a mortality endpoint clinically interpretable in this population.

Worth watching, but at low intensity until Phase 2 data. The signal that matters is the NCT05598333 readout: a positive cardiovascular death effect (or a clear composite signal across the four co-primary endpoints) in 150 non-ischemic NYHA III patients would be the first genuine win for cardiac gene therapy after two decades of failure, and it would revalue the entire AAV-cardiology space. With full enrollment complete and a 52-week co-primary window requiring event accrual, the ClinicalTrials.gov estimated primary completion is October 31, 2026, so a top-line readout in late 2026 or 2027 is plausible depending on data lock and analysis timing [2]. The specific data points to look for in the interim are any safety and biomarker updates from AskBio or Bayer. The Nature Medicine Phase 1 publication in 2025 is a meaningful positive, because sponsors with disappointing Phase 1 gene therapy data typically bury results in a poster rather than publish in a top-tier journal [1]. Commercial context: AskBio was acquired by Bayer in 2020 for approximately $2 billion upfront with up to $2 billion in contingent milestone payments, and Bayer's continued willingness to fund a 150-patient randomized Phase 2 outcomes trial is a soft signal that internal readouts have not derailed the program [3]. The counterweight is that Bayer has been culling programs elsewhere in its pharma pipeline under cost pressure, so AB-1002 is not guaranteed indefinite runway. No listed competitor is currently in Phase 2 or later for AAV heart failure gene therapy, which means a positive readout would land in an empty field. Check back after Bayer's next quarterly cardiovascular franchise update or any AskBio investor day disclosure on 2026 readout guidance.