Treating ICB-resistant glioma with anti-CD40 and mitotic spindle checkpoint controller BAL101553 (lisavanbulin)

Abstract

Glioblastoma is a highly malignant brain tumor with no curative treatment options, and immune checkpoint blockade has not yet shown major impact. We hypothesized that drugs targeting mitosis might affect the tumor microenvironment and sensitize cancer cells to immunotherapy. We used 2 glioblastoma mouse models with different immunogenicity profiles, GL261 and SB28, to test the efficacy of antineoplastic and immunotherapy combinations. The spindle assembly checkpoint activator BAL101553 (lisavanbulin), agonistic anti-CD40 antibody, and double immune checkpoint blockade (anti-programmed cell death 1 and anti-cytotoxic T lymphocyte-associated protein 4; anti-PD-1 and anti-CTLA-4) were evaluated individually or in combination for treating orthotopic GL261 and SB28 tumors. Genomic and immunological analyses were used to predict and interpret therapy responsiveness. BAL101553 monotherapy increased survival in immune checkpoint blockade-resistant SB28 glioblastoma tumors and synergized with anti-CD40 antibody, in a T cell-independent manner. In contrast, the more immunogenic and highly mutated GL261 model responded best to anti-PD-1 and anti-CTLA-4 therapy and more modestly to BAL101553 and anti-CD40 combination. Our results show that BAL101553 is a promising therapeutic agent for glioblastoma and could synergize with innate immune stimulation. Overall, these data strongly support immune profiling of glioblastoma patients and preclinical testing of combination therapies with appropriate models for particular patient groups.

Keywords: Brain cancer; Cancer immunotherapy; Immunology; Innate immunity; Oncology.

Figure 1. BAL27862 affects the number of viable SB28 cells in vitro but is a modest immunogenic cell death inducer.

(A) Number of viable SB28 cells based on total ATP content by luminescent (in relative light units) cell viability/proliferation assay; SB28 sensitivity to BAL27862 (BAL) corresponds to an IC₅₀ of 5.466 nM. (B–D) Evaluation of BAL27862 induction of key ICD features: (B) extracellular ATP release, (C) membrane translocation of calreticulin (CALR, calculated as MFI ratio), (D) HMGB1 release. Statistics: 2-way ANOVA with Tukey’s multiple-comparison test. ****: P < 0.0001. Box-and-whisker plot with the bounds of the box representing lower and upper quartiles, the line within the box showing the median, and the whiskers showing minimum and maximum values. All results include 3 biological replicates. Ab/iso, ratio of Ab MFI over isotype control MFI.

Figure 2. BAL101553 but not TMZ significantly improves survival of SB28-implanted mice.

(A) Treatment schedule of mice IC implanted with SB28 at day 0. (B) Symptom-free survival curve of mice treated with temozolomide (TMZ) or vehicle control (Ctrl). (C) Symptom-free survival curve of mice treated with anti–PD-1 and anti–CTLA-4 (ICB), BAL101553 (BAL), or a combination of both treatments (ICB+BAL). Treatments were injected intraperitoneally except for BAL101553, which was administered by oral gavage. Median survival (MS) is displayed in days for survival curves. Statistics: Log-rank (Mantel-Cox): nonsignificant: P > 0.05; **: P < 0.01; ***: P < 0.001. n = 10 mice per group.

Figure 3. IFN-γ signature and immunomodulatory molecule gene expression by SB28 glioma cells after in vivo growth.

(A and B) Gene expression level of genes comprising (A) an IFN-γ signature and (B) key immunomodulatory molecules, based on RNA sequencing between in vitro and ex vivo SB28 cells, respectively. Two biological replicates are analyzed for each group.

Figure 4. Early anti-CD40 combination with BAL101553 therapy offers a significant survival increase in SB28-implanted T/B cell–deficient mice.

(A) Treatment schedule of mice IC implanted with SB28 at day 0. (B and C) Symptom-free survival after treatment with anti-CD40 (aCD40), BAL101553 (BAL), or a combination of both (BAL+aCD40) of (B) wild-type, immunocompetent C57BL/6 mice or (C) immunodeficient B6.129S7-Rag1^tm1Mom/J (RAG1 KO) mice. MS is indicated in days, and the percentage increase (% Incr.) in MS is calculated relative to Ctrl group. Statistics: Log-rank (Mantel-Cox): nonsignificant P > 0.05; *: P < 0.05; **: P < 0.01; ****: P < 0.0001. n = 8–10 mice per group.

Figure 5. BAL101553-treated tumors are less infiltrated by BM-derived macrophages.

(A and B) Proportion of brain-infiltrating myeloid cells, and their corresponding PD-L1 expression (C and D), from mice IC implanted with SB28 and treated with vehicle control (Ctrl), anti-CD40 (aCD40), BAL101553 (BAL), or a combination of BAL and aCD40 were collected at day 23 (A and C) or at time of terminal symptoms (B and D). Statistics: 2-way ANOVA with Tukey’s multiple-comparison test. *: P < 0.05; ***: P < 0.001; ****: P < 0.0001. Error bars indicate SD. BMDM cells are defined as CD11b⁺CD49d⁺, microglia as CD11b⁺CD49d^–. All cells are gated on CD45⁺ cells after exclusion of doublets and dead cells. Box-and-whisker plot with the bounds of the box representing lower and upper quartiles, the line within the box showing the median, and the whiskers showing minimum and maximum values. n = 3–4 mice per group.

Figure 6. Combination of BAL101553 and aCD40 significantly improves survival of GL261-implanted mice.

(A) Treatment schedule of mice IC implanted with GL261 at day 0. (B) Symptom-free survival of mice treated with vehicle control (Ctrl), anti-CD40 (aCD40), BAL101553 (BAL), immune checkpoint blockade (ICB), a combination of BAL+ICB, a combination of BAL+aCD40, a combination of ICB+aCD40, and a combination of BAL+ICB+aCD40. MS is displayed in days. Statistics: Log-rank (Mantel-Cox): nonsignificant P > 0.05; *: P < 0.05; **: P < 0.01; ***: P < 0.001; ****: P < 0.001. n = 10 mice per group.