Rentosertib (TNIK Inhibitor): First AI-Discovered IPF Drug
Insilico's Rentosertib (ISM001-055): first-in-class AI-discovered TNIK inhibitor. GENESIS-IPF Ph2a FVC +98.4 mL vs -20.3 mL placebo (Nat Med 2025).
Introduction: The First AI-Discovered IPF Drug Clears Phase 2a
In June 2025, Nature Medicine published the Phase 2a results (GENESIS-IPF) of Insilico Medicine's oral TNIK inhibitor Rentosertib (development code ISM001-055 / INS018_055)[1]. This is the first drug in which both the target (TNIK) and the lead compound (ISM001-055) were identified and designed by a generative AI platform. The trial — 71 IPF patients, 12 weeks, safety-primary — showed a TEAE rate broadly comparable to placebo and a favorable secondary FVC signal (early clinical proof-of-concept).
The result matters twice over: first, as a potential new mechanistic class alongside pirfenidone and nintedanib; second, as a validation case for generative AI drug discovery. This article summarizes Rentosertib's mechanism, trial data, competitive context, and preclinical footprint using public sources only.
1. Compound Profile
| Item | Detail |
|---|---|
| Generic name (USAN) | Rentosertib (assigned January 2024) |
| Development codes | ISM001-055, INS018_055 |
| Sponsor | Insilico Medicine (Hong Kong / US / Suzhou) |
| Modality | Oral small-molecule, ATP-competitive TNIK inhibitor |
| Indication | Idiopathic pulmonary fibrosis (IPF) |
| Status | Phase 2a completed (GENESIS-IPF, China, NCT05938920); US Phase 2a ongoing (NCT05975983, up to 40 patients, ClinicalTrials.gov lists RECRUITING with the registered primary completion date of Feb 2026 unchanged; actual completion and readout not yet confirmed); oral rentosertib Phase III on track for H2 2026 initiation, inhaled formulation cleared by China CDE for Phase I (per Insilico press release, 28 April 2026) |
| Notable | First IPF candidate whose target and lead compound were both identified via generative AI (PandaOmics / Chemistry42) |
Insilico surfaced TNIK as an IPF target using PandaOmics (multi-omics plus knowledge graph), and nominated INS018_055 by running 30+ generative models in parallel in Chemistry42. Target-to-preclinical-candidate reportedly took ~18 months and target-to-Phase 1 about ~30 months — roughly half of traditional timelines[2].
For researchers tracking fibrosis & inflammation R&D
FDA approval alerts, trial readouts, preclinical model selection, and assay optimization — curated signal for bench-to-pipeline readers. 2 emails/month max.
2. Disease Context: The Unmet Need Beyond Pirfenidone / Nintedanib
IPF is a chronic progressive fibrosing interstitial lung disease in which repetitive alveolar epithelial injury triggers aberrant wound healing and myofibroblast accumulation. The established antifibrotics pirfenidone and nintedanib slow FVC decline by roughly 50% but do not halt or reverse disease, and tolerability issues (GI, hepatic) limit long-term adherence[1]. On 7 October 2025, nerandomilast (Jascayd, selective PDE4B inhibitor) received FDA approval for IPF (and for PPF on 19 December 2025), adding a new approved option that is starting to be considered in treatment sequencing.
The clinical field therefore continues to demand drugs that act through mechanisms distinct from current SOC, either as monotherapy or add-on. Rentosertib enters this gap by directly blocking a shared signaling hub downstream of Wnt and TGF-β.
3. Mechanism of Action: TNIK as a Wnt / TGF-β / Hippo Crosstalk Hub
TNIK Biology
TNIK (TRAF2- and NCK-interacting kinase) is a Ste20-family serine/threonine kinase (GCK family) that sits at the intersection of Wnt/β-catenin, TGF-β/SMAD, Hippo/YAP-TAZ, JNK, and NF-κB signaling[2]. It had been studied in colorectal cancer as an essential co-activator of TCF4/β-catenin transcription, but was essentially unreported as an IPF target prior to PandaOmics' nomination.
Antifibrotic Effect of TNIK Blockade
Ren et al. (Nat Biotechnol 2025, Epub 2024-03-08) showed that TNIK knockdown and ISM001-055 treatment attenuate TGF-β–driven EMT (reduced fibronectin, N-cadherin, p-SMAD2, p-FAK) and dose-dependently suppress nuclear/chromatin relocalization of β-catenin[2] — consistent with blockade of Wnt–TGF-β crosstalk at the myofibroblast level.
Differentiation vs SOC
- Pirfenidone: broad, weak multi-pathway suppression (MOA not fully defined)
- Nintedanib: multi-targeted receptor tyrosine kinase inhibitor (PDGFR / FGFR / VEGFR)
- Rentosertib: direct inhibition of TNIK, a downstream shared hub of Wnt and TGF-β
Because the pharmacological point of action does not overlap with SOC, combination development remains theoretically viable.
4. Clinical Evidence
Phase 1 (Healthy Volunteers)
78 healthy volunteers across 10 cohorts (5 SAD + 5 MAD) in New Zealand. Tolerability was favorable and the PK profile supported once-daily dosing[2].
GENESIS-IPF Phase 2a (NCT05938920)[1]
- Design: 22 sites in China, randomized double-blind placebo-controlled, 12-week treatment
- Population: 71 IPF patients
- Arms: placebo (n=17) / 30 mg QD (n=18) / 30 mg BID (n=18) / 60 mg QD (n=18)
- Primary endpoint: proportion with ≥1 TEAE (safety/tolerability)
- Key secondary endpoints: PK, FVC, DLCO, FEV1, Leicester Cough Questionnaire, 6MWD, acute exacerbation/hospitalization
Key Results
| Endpoint | 60 mg QD | Placebo |
|---|---|---|
| TEAE rate (primary) | comparable to placebo, mostly mild–moderate | — |
| FVC change, Week 12 | +98.4 mL | −20.3 mL |
| Serum proteomics (fibrosis markers) | COL1A1 / MMP10 / FAP / FN1 significantly reduced, IL-10 increased | — |
Changes in COL1A1 / MMP10 / FAP / FN1 correlated inversely with ΔFVC, providing exploratory support for a biomarker-efficacy linkage[1]. Because Phase 2a was safety-primary, small, and only 12 weeks, the TEAE rate was broadly comparable to placebo but treatment-related AEs and discontinuations (including hepatic and GI events) were reported, and the long-term safety of the 60 mg arm in particular is not yet established — confirmation in larger and longer trials is required. Separately, Qureight's AI-based HRCT analysis showed directionally consistent improvement in quantitative fibrosis indices[3].
Late-Stage Plans
Insilico had publicly targeted Q4 2025 for both a Phase 3 (China, ~500+ patients) and a Phase 2b (US, ~200+ patients)[4]. In its 28 April 2026 press release, the company updated the program: oral rentosertib Phase III is "on track" for H2 2026 initiation, and an inhaled rentosertib formulation has received IND clearance from China's CDE for a Phase I program enrolling ~80 healthy volunteers and IPF patients. The US Phase 2a (NCT05975983, 12 US sites across 8 states, up to 40 patients, INS018_055) is publicly listed on ClinicalTrials.gov as RECRUITING (last update 12 November 2025; the registered primary completion date of Feb 2026 has not been updated and actual completion / readout is not yet confirmed). Company guidance and registry status should be read separately.
5. Competitive Landscape
Late-stage IPF pipelines have entered a mechanism-diversification phase. Rentosertib's positioning:
- Nerandomilast (BI 1015550, Boehringer Ingelheim, brand Jascayd): selective PDE4B inhibitor. FIBRONEER-IPF / FIBRONEER-ILD Phase 3 positive; FDA approved for IPF on 2025-10-07 and for PPF on 2025-12-19 (China NMPA also approved PPF on 2025-12-10). A newly approved option, with combination and sequencing against existing antifibrotics now under consideration.
- Admilparant (BMS-986278, Bristol Myers Squibb): LPA1 receptor antagonist. ALOFT-IPF Phase 3 (NCT06003426, primary completion Oct 2026) ongoing.
- Bexotegrast (PLN-74809, Pliant Therapeutics): αvβ6/αvβ1 integrin inhibitor. Phase 2b/3 BEACON-IPF (NCT06097260) was discontinued on 2025-03-03 on DSMB recommendation (CT.gov: TERMINATED, primary completion 4 April 2025, 320 participants enrolled). The initial announcement framed the stop as a safety/tolerability signal — an imbalance in unadjudicated IPF-related adverse events vs placebo, despite an early FVC efficacy signal — and the full data review in June 2025 then characterized the overall risk-benefit profile as unfavorable, leading to program-level IPF discontinuation.
- Buloxibutid (C21, Vicore Pharma): AT2 receptor agonist. ASPIRE Phase 2b ongoing.
- ENV-101 (taladegib, Endeavor BioMedicines): Hedgehog/SMO inhibitor. WHISTLE-PF Phase 2b ongoing.
To the best of publicly available information, no other TNIK-selective inhibitor has been confirmed in IPF clinical development, so Rentosertib retains a first-in-class position within the TNIK subclass.
6. Preclinical Evaluation
Ren et al. (Nat Biotechnol 2025, Epub 2024) centered their in vivo package on the bleomycin mouse lung fibrosis model[2].
- Oral 3 mg/kg BID reduced fibrotic area by ~50%; 10 / 30 mg/kg BID by ~75%.
- Penh improvement was comparable to nintedanib.
- Activity was retained via the inhalation route, supporting route flexibility.
- Antifibrotic activity was also reported in skin and kidney fibrosis models, suggesting a tissue-agnostic MOA.
For preclinical candidate evaluation, bleomycin-induced lung fibrosis remains the industry standard. See Bleomycin Lung Fibrosis Model Pitfalls for practical design considerations, and Hydroxyproline Assay / Ashcroft Score Guide for endpoint selection.
7. Points to Watch
- Phase 3 timing and scale: whether the 12-week FVC benefit persists over 52+ weeks will determine real clinical value.
- Combination development with existing SOC: the non-overlapping MOA invites nintedanib/pirfenidone combination arms.
- Reproducibility of AI drug discovery: whether Insilico's follow-on candidates and competitor AI-discovered compounds achieve similar timeline compression.
- Biomarker-driven patient selection: whether serum COL1A1 / FAP / FN1 are incorporated into Phase 3 enrichment strategy.
- Global development integration: regulatory convergence (FDA / NMPA) for parallel China and US trials.
References
1. Xu Z, et al. A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial. Nat Med. 2025. Nature Medicine / PubMed 40461817 / ClinicalTrials.gov: NCT05938920
2. Ren F, et al. A small-molecule TNIK inhibitor targets fibrosis in preclinical and clinical models. Nat Biotechnol. 2025 (Epub 2024-03-08). Nature Biotechnology / PubMed 38459338
3. Drug Discovery News. Qureight analyses of Insilico Medicine's Phase IIa rentosertib data support preliminary efficacy results and future trial expansion. Article
4. Genetic Engineering & Biotechnology News. Insilico Eyes Q4 Start for Late-Stage Trials of IPF Candidate. Article