Azenosertib and gemcitabine: the toxicity did not come from the chemotherapy

What was tried

Azenosertib, an oral inhibitor of the WEE1 kinase, was combined with gemcitabine in a single arm, open label Phase 2 trial in second line advanced pancreatic adenocarcinoma. The study was investigator initiated and sponsored by Brandon Huffman, MD, with the Lustgarten Foundation, Stand Up To Cancer, and K-Group Beta, a Zentalis Pharmaceuticals subsidiary, as collaborators.

Participants had advanced pancreatic adenocarcinoma that had progressed on or had not tolerated a prior platinum based regimen. Azenosertib was given at 150 mg daily on days 1 to 5, 8 to 12, and 15 to 19 of a 21 day cycle, so the drug was on board for 15 of every 21 days. Gemcitabine was infused on days 1 and 8. The two dose levels differed only in the gemcitabine dose, 800 against 600 mg/m2.

The primary endpoint was the 6 month progression free survival rate by RECIST 1.1. The trial opened on 16 November 2023 and was recorded complete on 23 May 2026. Enrolment stopped at 12 participants, 5 at the lower gemcitabine dose and 7 at the higher one. The sponsor gave one reason: "Study stopped early due to reaching the minimum number of DLTs required for halting the study due to toxicity."

The biological hypothesis

WEE1 is a G2/M checkpoint kinase. It phosphorylates CDK1 on tyrosine 15, holding a damaged cell out of mitosis until repair is done. Pancreatic ductal adenocarcinoma is close to universally TP53 mutant, which removes the G1 checkpoint and leaves G2/M arrest as the main remaining brake. Inhibiting WEE1 removes that brake too, so a cell carrying unrepaired replication damage is pushed into mitosis and dies there.

Gemcitabine supplies the damage. It stalls replication forks, and pairing a replication stressor with a checkpoint inhibitor is one of the better worked out synthetic lethal designs in solid tumour oncology. The pancreatic precedent was specific: a dose escalation trial of adavosertib with gemcitabine and radiation in locally advanced pancreatic cancer was published in the Journal of Clinical Oncology in 2019.

The genetic case was thinner than the pharmacological one. Open Targets scores the WEE1 to pancreatic ductal adenocarcinoma association at 0.062, assembled from a literature component of 0.469 and a clinical component of 0.030, with no genetic association component at all. The target was chosen for what the pathway does in a TP53 null cell, not because human genetics pointed at WEE1 in this disease.

What actually happened

All 12 participants completed the study as defined. The primary endpoint came in at a 6 month progression free survival rate of 16.7 percent, with a 95 percent confidence interval from 4.7 to 59.1. That interval spans almost the entire plausible range. No values were posted for median overall survival, median progression free survival, objective response rate, or the grade 3 to 5 toxicity rate, so those endpoints cannot be reported here.

The safety record speaks more clearly. Across the 18 month reporting period, deaths were recorded in 5 of 5 participants at the lower gemcitabine dose and 7 of 7 at the higher one. At least one serious adverse event occurred in 3 of 5 and 5 of 7, which is 8 of 12. Decreased platelet count was reported in 5 of 5 and 5 of 7, raised alanine aminotransferase in 5 of 5 and 4 of 7, hypoalbuminaemia in 5 of 5 and 4 of 7, and anaemia in 4 of 5 and 5 of 7.

The dose levels carry the most useful signal. Cutting gemcitabine by 25 percent did not reduce the haematologic or hepatic burden. Every participant at the reduced gemcitabine dose had a decreased platelet count and a raised alanine aminotransferase, against 5 of 7 and 4 of 7 at the full dose.

Failure mechanism, best guess

This was a safety signal produced by overlapping organ toxicity rather than by either agent alone. Both drugs deplete the same marrow. In openFDA FAERS, gemcitabine carries 57,694 reports, and after disease progression and off label use the highest frequency terms are thrombocytopenia at 4,129, neutropenia at 3,692, and anaemia at 3,460. Azenosertib arrived with its own haematologic and infection liability on the record: the FDA placed a partial clinical hold on three azenosertib studies in June 2024 after two deaths from presumed sepsis in the DENALI ovarian cancer trial, and lifted it on 16 September 2024.

The de-escalation result is the part worth sitting with. Had gemcitabine been the dominant contributor, cutting it by a quarter should have moved the toxicity rates. It did not. That points at the azenosertib schedule, 15 days of dosing in every 21, as the larger contributor to cumulative haematologic burden. This is an inference from 12 participants across two dose levels, with no formal comparison appropriate at this size, so it is a direction rather than a measurement.

How to prevent this next time

The posted dataset supports qualitative design levers rather than quantitative recalibration. No subgroup, biomarker or exposure data were posted, so there is nothing to fit a Bayesian update or a power calculation to, and forcing one onto 12 patients would invent precision the record lacks.

Three levers apply. The first is schedule separation ahead of dose reduction. Work in Cancer Research in 2018 showed that mechanistic differences between CHK1 and WEE1 inhibition dictate how each should be scheduled against gemcitabine, and that sequencing, not dose, sets the therapeutic window. Reducing the chemotherapy partner while holding a near continuous checkpoint inhibitor schedule adjusts the variable that mattered least.

The second is treating a prior clinical hold as a design input. The sepsis deaths that paused the azenosertib programme were public well before this trial completed, and a combination protocol in a heavily pretreated population could have carried a predefined intermittent dosing arm.

The third is enrolment enrichment. No biomarker selection was applied, and Open Targets records no genetic association linking WEE1 to this disease. A replication stress marker defined in advance would have given the 6 month rate something to be conditional on.

The single highest leverage change would have been to vary the azenosertib schedule rather than the gemcitabine dose when the first dose limiting toxicities appeared.

What this means for similar programs

The Claidex mechanistic risk score for WEE1 in pancreatic ductal adenocarcinoma is 52 of 100, in the orange band, from a phase burden of 11.8, archetype severity of 14.1, recency of 7.1, genetic deficit of 14.1, and saturation of 4.9. The genetic deficit dominates, which is the honest summary of this target: WEE1 is tractable and mechanistically reasoned, and not genetically anchored in this indication.

Saturation is low. Open Targets lists two clinical stage molecules against WEE1, adavosertib and azenosertib. A thin field around a target with two safety driven stoppages in the Claidex graph, both involving azenosertib, suggests the constraint here is tolerability in combination rather than target validity.

Open questions

Which agent drove the dose limiting toxicities is not resolvable from the posted record, because no attribution breakdown was published. Whether an intermittent azenosertib schedule preserves the synthetic lethal effect in a TP53 null background at tolerable marrow cost is open, as is whether the unpopulated secondary endpoints will appear later.

Sources

1. • ClinicalTrials.gov record and posted results, NCT06015659. https://clinicaltrials.gov/study/NCT06015659 - Open Targets Platform, WEE1 (ENSG00000166483) and pancreatic ductal adenocarcinoma (MONDO_0005184). https://platform.opentargets.org/target/ENSG00000166483 - ChEMBL, azenosertib, CHEMBL5095036. https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL5095036/ - openFDA FAERS, gemcitabine and azenosertib report counts, retrieved 2026-10-28. https://open.fda.gov/apis/drug/event/ - Cuneo KC, et al. Dose Escalation Trial of the Wee1 Inhibitor Adavosertib (AZD1775) in Combination With Gemcitabine and Radiation for Patients With Locally Advanced Pancreatic Cancer. J Clin Oncol.

2. - Koh SB, et al. Mechanistic Distinctions between CHK1 and WEE1 Inhibition Guide the Scheduling of Triple Therapy with Gemcitabine. Cancer Res.

3. - Adavosertib and beyond: Biomarkers, drug combination and toxicity of WEE1 inhibitors. Crit Rev Oncol Hematol.

4. - FDA Lifts Partial Clinical Hold on Trials Evaluating Azenosertib in Advanced Solid Tumors. OncLive, 2024. https://www.onclive.com/view/fda-lifts-partial-clinical-hold-on-trials-evaluating-azenosertib-in-advanced-solid-tumors.