Novel randomized evidence for TNIK inhibition
The study supplies phase 2a randomized patient data for a first-in-class TNIK inhibitor and extends the cited phase 0/1 development program.
↳ Introduction; Abstract; Methods, Study Design
Abstract
Despite substantial progress in artificial intelligence (AI) for generative chemistry, few novel AI-discovered or AI-designed drugs have reached human clinical trials. Here we present the results of the first phase 2a multicenter, double-blind, randomized, placebo-controlled trial testing the safety and efficacy of rentosertib (formerly ISM001-055), a first-in-class AI-generated small-molecule inhibitor of TNIK, a first-in-class target in idiopathic pulmonary fibrosis (IPF) discovered using generative AI. IPF is an age-related progressive lung condition with no current therapies available that reverse the degenerative course of disease. Patients were randomized to 12 weeks of treatment with 30 mg rentosertib once daily (QD, n = 18), 30 mg rentosertib twice daily (BID, n = 18), 60 mg rentosertib QD (n = 18) or placebo (n = 17). The primary endpoint was the percentage of patients who have at least one treatment-emergent adverse event, which was similar across all treatment arms (72.2% in patients receiving 30 mg rentosertib QD (n = 13/18), 83.3% for 30 mg rentosertib BID (n = 15/18), 83.3% for 60 mg rentosertib QD (n = 15/18) and 70.6% for placebo (n = 12/17)). Treatment-related serious adverse event rates were low and comparable across treatment groups, with the most common events leading to treatment discontinuation related to liver toxicity or diarrhea. Secondary endpoints included pharmacokinetic dynamics (Cmax, Ctrough, tmax, AUC0-t/τ/∞ and t1/2), changes in lung function as measured by forced vital capacity, diffusion capacity of the lung for carbon monoxide, forced expiry in 1 s and change in the Leicester Cough Questionnaire score, change in 6-min walk distance and the number and hospitalization duration of acute exacerbations of IPF. We observed increased forced vital capacity at the highest dosage with a mean change of +98.4 ml (95% confidence interval 10.9 to 185.9) for patients in the 60 mg rentosertib QD group, compared with -20.3 ml (95% confidence interval -116.1 to 75.6) for the placebo group. These results suggest that targeting TNIK with rentosertib is safe and well tolerated and warrants further investigation in larger-scale clinical trials of longer duration. ClinicalTrials.gov registration number: NCT05938920 .
Key Claims
Primary findingTentative
We observed increased forced vital capacity at the highest dosage with a mean change of +98.4 ml (95% confidence interval 10.9 to 185.9) for patients in the 60 mg rentosertib QD group
Fragile — the claim's weakest requirement is that the effect exists at all. Experimental. Every point below is traceable to a quoted passage, but this forecast has not yet been validated against known replication outcomes.
secondary efficacy endpoint with small arms, differential missingness and MAR-only imputation
Abstract; Results, Secondary Lung Function Endpoints
SecondaryTentative
Targeting TNIK with rentosertib is safe and well tolerated and warrants further investigation in larger-scale clinical trials of longer duration.
systematic short-term monitoring offset by dose-related harms and unresolved ALT counts
Abstract conclusions
Derived from the full evaluation — not a separate score.
Strengths · Limitations
Strengths
Controlled and broadly blinded trial design
Allocation used interactive response technology with controlled code access, and participants, investigators, site staff and analysts remained blinded through data freeze.
↳ Methods, Study Design
Transparent positioning of clinical limitations
The Discussion identifies small treatment arms, geographic and demographic homogeneity, short follow-up and the need for larger global trials.
↳ Discussion, limitations and final paragraphs
Limitations
Conflicting hepatic safety event counts
The Primary Safety Endpoint narrative reports ALT increases of both 6 and 3 in the 60 mg QD group, whereas Table 2 reports 1. This unresolved discrepancy affects interpretation of a clinically salient dose-related harm.
↳ Results, Primary Safety Endpoint; Table 2
Differential missingness weakens efficacy inference
Week-12 spirometry was missing for 8 of 18 participants in the 60 mg QD arm versus 3 of 17 on placebo. Multiple imputation assumed MAR without a reported tipping-point or other missing-not-at-random sensitivity analysis.
↳ Methods, Statistical Analysis; Results, Patients; Fig. 2b
Safety language exceeds reported balance
The Abstract and Discussion call serious treatment-related events similarly low or comparable despite 0% in placebo and 5.6–11.1% across active arms, alongside dose-related treatment AEs and discontinuations.
↳ Abstract; Discussion, paragraph 1; Table 2
Nabu Commentary
The randomized, placebo-controlled design and systematic harms collection support a competent phase 2a evaluation of rentosertib. The principal efficacy signal is nevertheless secondary, not efficacy-powered, and depends on MAR imputation amid substantially greater missingness in the higher-dose arms. Safety interpretation is further constrained by incompatible ALT counts in the Primary Safety Endpoint narrative and Table 2, as well as dose-related treatment AEs and discontinuations. The study therefore contributes meaningful early clinical evidence while supporting further trials rather than confirmatory efficacy or broad tolerability conclusions.
Check another paper
Author Note / Field Feedback
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Dimension breakdown
Study Quality
Sound3.4
Confidence mediumContribution
3.7
The trial provides new randomized phase 2a evidence for a first-in-class TNIK inhibitor in IPF and meaningfully extends the cited phase 0/1 program. Its contribution remains preliminary rather than practice-changing because efficacy evidence comes from a small, short trial.
Introduction; Results, Secondary Lung Function Endpoints; Discussion
“Treatment with 60 mg rentosertib QD over 12 weeks was associated with a trend toward an increase in FVC”
Methodological Rigour
3.3
Randomization, allocation control, blinding, ITT analysis and systematic harms collection are described, but 22 week-12 spirometry values were missing and imputed under MAR without a reported sensitivity analysis. The hepatic-event count inconsistency is separately reflected in the Red reliability flag rather than a numerical reliability cap.
Methods, Study Design and Statistical Analysis; Results, Patients and Primary Safety Endpoint; Table 2
“Missing values at week 12 were imputed using a multiple imputation method assuming missing at random.”
Reporting
3.2
The report is logically organized and labels post-hoc analyses, but its prominent safety language is not calibrated to dose-related treatment AEs, discontinuations and active-arm serious events. The MCID wording also gives a stronger impression than the preliminary efficacy design supports.
Abstract; Discussion, paragraphs 1–2; Table 2
“Treatment-related serious adverse event rates were low and comparable across treatment groups”
Positioning
3.4
The paper substantively situates rentosertib against current IPF therapies and the limited clinical record of AI-discovered drugs, while acknowledging sample size, duration and demographic homogeneity. Interpretation is less complete where differential withdrawal and absent efficacy powering are not fully carried into the MCID claim.
Introduction; Methods, Statistical Analysis; Discussion, paragraphs 2 and 6
“The limitations of this study include the small cohort size of each arm”
Reliability
Concerns4 of 4 checks
Internal CoherenceConcern
A clinically salient hepatic-safety count is internally inconsistent, and prominent safety language is not fully aligned with the arm-level adverse-event pattern. The ALT discrepancy prevents confident reconciliation of the narrative with Table 2.
Research ConductClean
Ethics approval, written informed consent and prospective trial registration are declared, and sponsor involvement in design, analysis and interpretation is disclosed. No text-grounded research-conduct concern was identified.
Flags: 4 declared / 5 total
Reference IntegrityClean
70 of 70 checkable references verified
70 references in manuscript 2 references confirmed by manual review
References not verified automatically are resolved manually within 24 hours.
Post-Publication RecordClean
No retraction or expression of concern found in Crossref (includes Retraction Watch).
Last checked: 17 Oct 2026
Sources: Crossref (incl. Retraction Watch) ✓
Readiness
Where this paper’s evidence sits on the path from initial observation to real-world use.
Demonstrated
The work has progressed through phase 0/1 into a randomized phase 2a patient trial, placing it at clinical proof-of-concept. Its 12-week duration, small arms and preliminary secondary efficacy endpoint keep it well below deployment readiness.
Abstract; Methods, Study Design; Discussion, final paragraph
“warrants further investigation in larger-scale clinical trials of longer duration”
AI-generated, human-governed. Something look off? Contact us to request a review.