ASP1012: an eight-arm oncolytic vaccinia programme that ended with 15 patients dosed and nothing posted

What was tried

Astellas Pharma Global Development ran study 1012-CL-0101 (NCT06171178), a Phase 1 open-label, non-randomised dose escalation and expansion study of ASP1012 in adults with locally advanced or metastatic solid tumours. Dosing opened 15 May 2024 across eight United States sites. The record lists an actual primary completion date of 11 September 2025, an actual enrolment of 15 participants, and a status of Terminated. The posted reason is "Due to strategic decision made by the sponsor the study was terminated". No results are posted.

The protocol was written in three parts and eight arms. Part 1 escalated intravenous ASP1012 on a 21 day cycle in previously treated solid tumours. Part 2 opened three expansions: ASP1012 alone in previously treated cutaneous melanoma, ASP1012 with pembrolizumab in previously treated solid tumours, and the same combination in treatment-naive melanoma. Part 3 added gastric, colorectal and ovarian expansions. Melanoma eligibility required prior anti-PD-1 or anti-PD-L1 therapy.

Every primary outcome was a safety measure: dose limiting toxicities through 28 days, adverse events through 19 months, and laboratory, vital sign, electrocardiogram and performance status abnormalities. No efficacy endpoint was primary. Fifteen participants against an eight arm design means the study ended inside dose escalation, so none of the seven expansions can have read out.

The biological hypothesis

ASP1012 is VET2-L2, the first candidate from KaliVir Immunotherapeutics' Vaccinia Enhanced Template platform, licensed worldwide to Astellas in December 2020 and cleared for a Phase 1 IND in October 2023. It is an oncolytic vaccinia virus engineered to express a leptin-IL-2 fusion protein, given intravenously.

The design intent was not simply to add IL-2. Systemic interleukin-2 acts through the trimeric IL2RA, IL2RB and IL2RG receptor or the dimeric IL2RB and IL2RG complex, and its therapeutic window has always been set by what that activation does to vascular endothelium and to regulatory T cells. The KaliVir approach was to keep the cytokine where the tumour is. Work from the platform's founding laboratory showed that an IL-2 construct combining a glycosylphosphatidylinositol anchor with a rigid peptide linker produced functional IL-2 on the tumour cell surface, treated murine models without systemic toxicity, and cured most animals with high tumour burden when combined with PD-1 or PD-L1 blockade (DOI 10.1038/s41467-018-06954-z). The pembrolizumab arms translated that last result.

The delivery argument rests on preclinical evidence. The target argument is weaker. Open Targets scores IL2 to melanoma at 0.11223 with literature as the only contributing datatype at 0.9231, and IL2 to cancer at 0.11453 on literature alone. Cutaneous melanoma scores 0.19010, carried by somatic mutation at 0.3040. No genetic association datatype contributes to any of the three. IL-2 in solid tumours has never been a genetically supported target. It has been a pharmacological tool with a toxicity problem, and recent programmes re-engineered the tool rather than validated the target.

What actually happened

Nothing was published about what happened. The trial ran 16 months of dosing, enrolled 15 participants against a design needing several times that number to reach its first expansion, posted no results, and closed with a one sentence sponsor statement. Trade coverage in June 2026 reported the termination and noted no Astellas comment beyond the registry text.

That absence is the finding. A first-in-human oncolytic virus study that reaches 15 patients and reports nothing leaves no dose, no tolerability profile, no viral kinetics, and no answer to the question the construct was built to ask, which is whether tumour-anchored IL-2 expression is achievable in humans at a tolerated dose.

Failure mechanism, best guess

The archetype is strategic reprioritisation and the registry text supports it. The trial did not stop on a data monitoring committee recommendation, a toxicity cluster, or a futility boundary.

The mechanism worth naming is the one that made the programme easy to cut. Intravenous oncolytic virotherapy carries a delivery risk separate from its target risk. Circulating vaccinia is subject to complement-mediated lysis and to pre-existing and induced neutralising antibody, and how much virus reaches tumour after systemic dosing has been the recurring constraint on the modality (DOI 10.1038/mt.2015.49, DOI 10.1136/jitc-2022-006024). A programme resting on literature-only target evidence, using a route with a known biodistribution problem, and reporting only safety endpoints generates no interim signal a portfolio review can weigh.

How to prevent this next time

No endpoint-level data were released, so no Bayesian recomputation, power calculation or subgroup analysis is supported here. Three qualitative levers are.

Build the pharmacodynamic readout into the primary endpoints rather than around them. Every primary outcome here was a toxicity measure. For a construct whose thesis is that IL-2 stays in the tumour, paired biopsy evidence of anchored IL-2 expression and viral replication decides whether the platform works, and it should not sit among exploratory endpoints where it disappears on termination.

Size the registered design to the decision it can reach. Three parts and eight arms committed the record to a scope 15 patients could never test. Adding expansions only after an escalation gate would have left a cleaner record.

Separate delivery risk from target risk before escalating. The intravenous route carries a documented neutralisation and complement barrier, and a biodistribution substudy answers a modality question that outlives any single payload.

The single highest leverage change would have been to make intratumoural IL-2 display and viral replication in paired biopsies a co-primary endpoint of dose escalation, so that 15 dosed patients would have produced a publishable answer about the platform rather than a terminated record with no data.

What this means for similar programs

IL2 now carries two recorded terminations in the Claidex graph and a mechanistic risk score of 37, in the yellow band. The other is [[slc-3010-il2-cd122-biased-solid-tumor-phase1-2-strategic-termination]], a CD122-biased IL-2 complex shelved in advanced solid tumours, and both stopped on sponsor decisions rather than on data. The score is driven by the genetic deficit term at 13.32 of 15, not by phase burden or any safety finding. Adjacent nodes tell the same story: [[cln-617-il12b-il2-fusion-solid-tumors-phase1-efficacy-termination]] and [[eciskafusp-alfa-pd1-il2v-il2rb-solid-tumors-phase1-2-orr-futility]] both engineered the IL-2 axis for tumour localisation and both stopped in Phase 1.

Across these four the engineering problem keeps being solved and the target question keeps not being asked. Oncolytic virus combinations with checkpoint blockade have produced real clinical results in melanoma (DOI 10.1200/JCO-25-01346), so the modality is not the constraint. A new IL-2 payload programme should state which human evidence, other than literature co-occurrence, supports IL-2 as the right cargo for its tumour.

Open questions

Was ASP1012 tolerated, and was a recommended Phase 2 dose identified? Did paired biopsies show viral replication and anchored IL-2 expression after intravenous dosing? Did any combination arm open before closure? None of this is public.

Sources

1. • ClinicalTrials.gov, NCT06171178, accessed 30 August 2026. https://clinicaltrials.gov/study/NCT06171178 - NCI Drug Dictionary, ASP1012. https://www.cancer.gov/publications/dictionaries/cancer-drug/def/oncolytic-vaccinia-virus-expressing-leptin-il-2-fusion-protein-asp1012 - KaliVir Immunotherapeutics, ASP1012 IND clearance, 24 October 2023. https://www.businesswire.com/news/home/20231024827071/en - Vaccinia virus expressing re-designed interleukin-2. Nat Commun, 2018. https://- Complement inhibition prevents oncolytic vaccinia neutralization in immune humans. Mol Ther, 2015. https://- Oncolytic vaccinia with improved intravenous efficacy. J Immunother Cancer, 2023. https://- RP1 with nivolumab in anti-PD-1-failed melanoma (IGNYTE). J Clin Oncol, 2025. https://- Open Targets Platform, IL2 and melanoma. https://platform.opentargets.org/evidence/ENSG00000109471/MONDO_0005105 - openFDA FAERS, aldesleukin class reference, 30 August 2026. https://api.fda.gov/drug/event.json.