ALN-HSD

ALN-HSD is a Regeneron/Alnylam siRNA injection that silences HSD17B13, a liver enzyme whose natural loss-of-function (LOF) carriers are protected from cirrhosis and chronic liver disease [2]. Phase 2 (NCT05519475) is enrolling 120 MASH (metabolic dysfunction-associated steatohepatitis, the current medical term for what was called NASH - a fatty liver disease driven by lipid accumulation and inflammation) patients without the protective allele, with primary endpoint a quantitative fibrosis score read from biopsy [1]. The thesis is precision medicine: identify patients whose biology mirrors the variant-negative population and knock their HSD17B13 down to phenocopy the protected genotype. If knockdown delivers on the human genetics, ALN-HSD enters a MASH market that Madrigal's resmetirom just cracked open - the first-ever approved MASH drug - opening a multi-billion-dollar opportunity if treatment becomes standard of care over the next decade [3].

Novel siRNA in Phase 2 (NCT05519475), enrolling 120 MASH patients since 2022 with no placebo arm in the disclosed design [1]. No FDA breakthrough or fast-track designation has been disclosed. The compound originated at Alnylam and moved under Regeneron clinical leadership through the April 2019 expansion of the Alnylam-Regeneron collaboration, which broadened the original 2014 deal to cover additional liver-targeted siRNA programs including HSD17B13 [5]. Under that deal Alnylam receives tiered royalties on net sales plus development and commercialization milestones, but Regeneron controls clinical development including Phase 3 go/no-go decisions - a structural point that matters for anyone modeling Alnylam's royalty book against Regeneron's pipeline prioritization [5]. ALN-HSD uses Alnylam's GalNAc conjugation platform - a sugar tag that drags the siRNA into hepatocytes through the asialoglycoprotein receptor on the liver cell surface. That delivery technology is past first-approval risk: it powers Novartis's approved Leqvio (inclisiran) for LDL cholesterol and Alnylam's own Givlaari and Onpattro. A Phase 1 dose-finding study established the dose used in Phase 2; published Phase 1 disclosures have been limited, so the realistic depth and durability of hepatic HSD17B13 knockdown in MASH patients remains incompletely characterized in public data. Readout timing has not been formally guided. Recruiting status as of mid-2026 implies top-line data 2026-2027 if enrollment closes this year, later if biopsy logistics stay slow. The next public catalysts are Regeneron's quarterly updates and any AASLD or AHA late-2026 abstract on early biomarker data.

HSD17B13 is a liver enzyme that lives on lipid droplets, the fat-storage bubbles inside liver cells. What it does day-to-day in MASH biology is still being worked out - it oxidizes retinol/retinaldehyde and acts on several lipid substrates, but its disease-relevant substrate profile and full enzymology remain under active investigation. The drug case rests on human genetics, not enzymology. A common loss-of-function variant (rs72613567) cuts the risk of NASH, cirrhosis, and hepatocellular carcinoma by roughly 17-30% across hundreds of thousands of sequenced people, discovered by Regeneron Genetics Center in 2018 [2]. Carriers have lower liver enzymes and less fibrosis without obvious downside. The drug logic: if losing HSD17B13 protects you for free, knocking it down with siRNA should mimic that protection therapeutically. siRNA tells the cell to destroy a specific mRNA before it makes protein. Dosing intervals run months. The modality is past first-approval risk: Leqvio, Onpattro, and Givlaari are on market. This is one of the cleanest 'human genetics to drug target' calls in hepatology. The empirical question is whether somatic knockdown in adults phenocopies a germline LOF that has been protecting carriers since conception.

NCT05519475 is Phase 2, n=120, recruiting, sponsored by Regeneron [1]. Primary endpoint: change in quantitative liver fibrosis measured by qFibrosis, an AI-scored continuous fibrosis readout from second-harmonic-generation imaging of biopsies. That is less mature than NASH CRN staging - the FDA-standard histological scoring system for liver biopsy that requires two-point fibrosis improvement or NASH resolution for registrational claims - but offers continuous resolution to detect sub-stage signal in a small trial. Translation to a registrational endpoint will be a Phase 3 problem. Patient selection is the interesting part. Enrolled patients have MASH with confirmed fibrosis and are selected for 'increased genetic risk' - meaning they do not carry the protective HSD17B13 LOF variant. Precision medicine done correctly: target the patients whose biology you can plausibly fix. The design concern is the lack of a clear placebo control in the public protocol. MASH fibrosis varies session-to-session and biopsy-to-biopsy; an open-label single-arm readout will need careful baseline-controlled analysis and probably a parallel placebo arm before anyone takes this to Phase 3. Enrollment pace is the other concern. A 120-patient trial opened in 2022 that is still recruiting in mid-2026 implies four years to fill 120 slots - that's slow even for biopsy-gated MASH studies and suggests genotype screening (excluding ~25-30% of MASH patients who carry the protective allele) plus biopsy logistics together are a serious throughput tax. The same screening problem will scale into Phase 3.

About 7 in 100 drugs at this stage are eventually approved, and our model estimates this one has a 7% chance. It adjusts that starting point using ten facts about the trial and its sponsor. Two things push the number up: more secondary endpoints than usual and a sponsor with a strong approval track record. Two things pull it down: heavier-than-usual blinding and weak earlier-phase results.

Efficacy risk first. Lifelong reduced HSD17B13 in LOF carriers is not the same physiology as deep acute knockdown in middle-aged adults with established fibrosis. Even if the genetics translates, reversal of existing fibrosis is a higher bar than prevention - every MASH program has run into this gap [2]. The qFibrosis endpoint reduces measurement noise but is not FDA-validated for registration, so Phase 3 will need NASH resolution or NASH CRN fibrosis improvement. Safety. HSD17B13 metabolizes retinol and several lipid substrates; its full substrate profile and disease-relevant activity remain under investigation. Sustained suppression of retinoid metabolism in the liver is a theoretical concern that has not shown up in LOF carriers but could behave differently under abrupt deep knockdown. siRNA durability cuts both ways: months-long target suppression means if something emerges late, you cannot titrate down quickly. Competitive risk is the loudest one. Two flavors. Same-target: Arrowhead's ARO-HSD targets the identical gene (HSD17B13) using a different RNAi delivery chemistry, so both programs will produce reads on whether somatic HSD17B13 knockdown actually moves human fibrosis. Parallel programs mean partial derisking from positive ARO-HSD data and harder enrollment competition for genotype-eligible patients - the same screening tax that's already slowing ALN-HSD recruitment. Different-target: Madrigal's Rezdiffra (resmetirom) is approved and launching with quarterly sales ramping through 2025 [3][6]. Akero's efruxifermin and 89Bio's pegozafermin (both FGF21 analogs) have strong Phase 2b fibrosis data. Boehringer's survodutide and Lilly's tirzepatide show MASH benefit driven by weight loss - tirzepatide's Phase 2 SYNERGY-NASH readout (NEJM 2024) showed roughly 62% NASH resolution at the top dose and is the highest-impact MASH data point from the metabolic-disease side to date [8]. Genotype-selected precision dosing carves a defensible niche only if the effect size is large; modest fibrosis improvement in a sub-population is hard to sell against drugs that work broadly.

Watching, not betting. ALN-HSD is the cleanest 'human-genetics-to-drug' shot in MASH right now, and Regeneron is the right shop to run it. The only data point that matters until readout is Phase 2 top-line - likely late 2026 if enrollment closes this year [1]. The signal worth waiting for: biopsy-confirmed fibrosis improvement on qFibrosis with a clean safety profile and a dose-response curve. That would justify Phase 2b/3 with placebo control and reset the commercial conversation. The single-arm design caps what any positive readout can claim. Even a good result needs controlled follow-up before Regeneron commits Phase 3 capital, so watch for Regeneron's 8-Ks or AASLD/AHA late-2026 presentations carrying the data. A program acceleration announcement (Phase 2b with placebo, broader genotype expansion, or a real Phase 3 design) is the call that says they believe the readout. ARO-HSD readouts are an exogenous signal worth tracking on the same calendar - same target, different chemistry, partial cross-read. Commercial frame: Regeneron's $14.3B revenue base [4] means MASH is upside, not survival - they can fund Phase 3 either way. Alnylam's economics are tiered royalties on net sales plus milestone payments at development and commercial gates [5]; ALN-HSD is one royalty stream in a partnered portfolio rather than a needle-mover on its own, but Phase 2 success would unlock a milestone tranche and reset the option value Alnylam books on the program. The real bear case is not failure of ALN-HSD but irrelevance: if GLP-1/GIP agonists like tirzepatide deliver MASH benefit at scale [8], a genotype-selected siRNA fights for a narrow slice of a market the metabolic giants own.