Sudhir Tauro

Therapeutic progress in higher-risk myelodysplastic syndromes (HR-MDS) has stalled. Despite repeated attempts to improve outcomes through incremental refinements to classification systems and hypomethylating agent (HMA)-based backbones, no phase 3 study since AZA-001 has replicated even the modest survival benefit observed with azacitidine. We propose that this stagnation results from a fundamental mismatch between the biological realities of HR-MDS and the assumptions underpinning contemporary drug development. HR-MDS encompasses biologically diverse ecosystems characterised by distinct clonal architectures, evolutionary trajectories and marrow microenvironmental dysfunction, despite overlapping phenotypic and genotypic features. Treating this heterogeneity as a single entity anchored to a hypomethylating agent backbone, risks obscuring therapeutic signals and misinterpreting responses. Here, we examine the potential biological, microenvironmental, and methodological drivers of therapeutic insufficiency in HR-MDS, to advocate for biologically coherent adaptive platform trials using biomarker-enriched patient selection, and propose shifting away from default HMA backbones. Only by redesigning strategy around biology, rather than prioritising ease of recruitment, or historical precedent, can we restore momentum and lift in HR-MDS drug development.

Supplementary Movie 2 from Elraglusib Induces Cytotoxicity via Direct Microtubule Destabilization Independently of GSK3 Inhibition

Supplementary Movie 1 from Elraglusib Induces Cytotoxicity via Direct Microtubule Destabilization Independently of GSK3 Inhibition

Supplementary Movie 2 from Elraglusib Induces Cytotoxicity via Direct Microtubule Destabilization Independently of GSK3 Inhibition

Supplementary Movie 1 from Elraglusib Induces Cytotoxicity via Direct Microtubule Destabilization Independently of GSK3 Inhibition

Elraglusib (9-ING-41) is an ATP-competitive inhibitor of glycogen synthase kinase-3 (GSK3) with preclinical studies demonstrating broad activity against many tumor types. Promising early-phase clinical trial data led to FDA orphan drug status, and a randomized phase II study in combination with cytotoxic chemotherapy in pancreatic cancer has recently completed its recruitment. Similarly, single-agent responses in adult T-cell leukemia/lymphoma and melanoma and combination treatment data in several other tumor types have been encouraging. The elraglusib mechanism of action is unknown, but it is unlikely to act through GSK3 inhibition because cytotoxicity is observed below the IC50 for GSK3, and other small molecule GSK3 inhibitors do not produce cytotoxic effects, at least in lymphoma cells. We show here that elraglusib perturbs chromosomal alignment to cause a mitotic arrest in multiple tumor lines. This arrest is caused by direct microtubule (MT) depolymerization, which prevents the attachment of kinetochores to MTs. At clinically relevant doses, these mitotically arrested cells eventually undergo mitotic slippage, leading to gross chromosome missegregation, DNA damage, and apoptosis. These effects explain the cytotoxicity of elraglusib because temporarily pausing cell-cycle progression with the CDK4/6 inhibitor palbociclib abolishes any drug-induced genotoxicity and apoptosis. In summary, elraglusib acts as a direct MT destabilizer both in vitro and across multiple cancer types, resulting in mitotic arrest, DNA damage, and apoptosis. These effects likely account for its broad pan-cancer activity, which does not rely upon GSK3 inhibition as they are not replicated by other GSK3 inhibitors.Significance:Elraglusib was designed as a GSK3 inhibitor and is currently in clinical trials for several cancers. We show conclusively that the target of elraglusib that leads to cytotoxicity is MTs and not GSK3. This has significant implications for ongoing clinical trials of the compound and will help in understanding off-target side effects, inform future clinical trial design, and facilitate the development of biomarkers to predict response.

Elraglusib impairs proliferation and reduces cell viability in stimulated primary human T cells but does not affect unstimulated cells (A and B). Primary human T cells were isolated from PBMCs from healthy volunteers. Unstimulated controls (A) remained in CM, and stimulated cells (B) were treated with a CD3/CD28-co-stimulation kit. Cells were treated with DMSO control or elraglusib (Elr). Cell viability was assayed using Promega RealTime-Glo MT assay and the plate read daily for 120 hours. Data combine three independent experiments with T cells from three individual volunteers. Data normalized to the experimental control peak luminescence. Elraglusib increases the soluble tubulin fraction and decreases the insoluble tubulin fraction in stimulated primary human T cells (C and D). Proliferating stimulated primary human T cells were treated for 5 hours and assayed for soluble and insoluble tubulin as described above. Blots are representative of three independent experiments, and quantification combines three independent experiments. Fold change calculated vs. control, error bars represent SEM, and statistical significance was determined using one-way ANOVA, with P ≤ 0.05. ns, not significant; *, P < 0.05; ****, P < 0.0001.

Elraglusib causes a mitotic arrest. A and B, Example flow cytometry plots (A) and quantifications (B) showing the effect of 24-hour elraglusib (Elr) treatment on the percentage of 2N, 4N, and 4N-mitotic K299 WT cells. C and D, Example flow cytometry plots (C) and quantifications (D) following 24-hour treatments with two structurally distinct GSK3 inhibitors, CT99021 and LY2090314 in K299 WT cells. E and F, Example flow cytometry plots (E) and quantifications (F) showing the effect of 24-hour elraglusib (Elr) treatment on the percentage of 2N, 4N, and 4N-mitotic K299 GSK3A/B dual paralog shRNA knockdown cells (DKD). G and H, Total GSK3 immunoblot (G) with quantification confirming knockdown efficiency in K299 DKD compared with K299 WT cells (H). Cells were treated for 24 hours prior to cell-cycle analysis by flow cytometry using propidium iodide and anti–phospho-histone-H3 (Ser10) staining. Ten thousand events were analyzed per sample. Cell-cycle histograms are representative of three independent experiments, and quantifications combine three independent experiments. Immunoblot representative of three independent experiments; quantification was performed by normalizing to actin loading control and combines three independent experiments. Fold change calculated vs. control, error bars represent SEM, and statistical significance was determined using one-way ANOVA, with P ≤ 0.05. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Elraglusib acts via direct MT destabilization (A–J). Example images (A) of MCF7 cells treated with nocodazole (Noco), paclitaxel (Taxol), elraglusib (Elr), and CT99021 (CT). MCF7 cells were treated for 24 hours before fixation and IF microscopy. Ten mitotic cells were imaged per coverslip; images are representative of three independent experiments. Tubulin fractionation experiments comparing soluble (S) and insoluble (I) tubulin by immunoblot in MCF7 (B and C), K299 (D and E) and HH (F and G) cells. Cells were treated for 5 hours and lysed in MT stabilization buffer before fractionation into soluble (S) and insoluble (I) tubulin and analysis by immunoblot. Quantification was performed by normalizing to actin loading control and then to the experimental control. Blots are representative of three independent experiments, and quantification combines three independent experiments. Fold change calculated vs. control, error bars represent SEM, and statistical significance was determined using one-way ANOVA, with P ≤ 0.05. Scale bars, 10 μm. H, Kymographs illustrating dynamics of MT growth in vitro in the absence of drug (control) or the presence of 3.3 μmol/L nocodazole, 1, 2, 3, or 10 μmol/L elraglusib. Rhodamine-labeled short stable MTs were seeded on the coverslip, and 12 μmol/L tubulins (containing 4% fluorescent HiLyte 488 tubulins) were added to the reaction to generate and visualize dynamic MTs. Scale bars, 2 μm. I and J, Quantification of MT growth rate (I) and catastrophe frequency (J) with no drug treatment (control) or with 3.3 μmol/L nocodazole, 1, 2, 3 or 10 μmol/L elraglusib. The number of analyzed MT growth events: n = 34 for control; n = 42 for 3.3 μmol/L nocodazole; and n = 43, 24, and 21 for 1, 2, and 3 μmol/L elraglusib, respectively. MT catastrophe frequency was analyzed for dynamic MT on individual MT seeds–the number of analyzed MT seeds: n = 13 for control; n = 20 for 3.3 μmol/L nocodazole; and n = 13, 18, and 15 for 1, 2, and 3 μmol/L elraglusib, respectively. Error bars represent SD. Statistical significance was determined using one-way ANOVA, with P ≤ 0.05. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.