Onvansertib in Combination With Chemotherapy and Bevacizumab in Second-Line Treatment of KRAS-Mutant Metastatic Colorectal Cancer: A Single-Arm, Phase II Trial

Abstract

Purpose: This phase II study evaluated the efficacy and tolerability of onvansertib, a polo-like kinase 1 (PLK1) inhibitor, in combination with fluorouracil, leucovorin, and irinotecan (FOLFIRI) + bevacizumab for the second-line treatment of KRAS-mutant metastatic colorectal cancer (mCRC).

Patients and methods: This multicenter, open-label, single-arm study enrolled patients with KRAS-mutated mCRC previously treated with oxaliplatin and fluorouracil with or without bevacizumab. Patients received onvansertib (15 mg/m² once daily on days 1-5 and 15-19 of a 28-day cycle) and FOLFIRI + bevacizumab (days 1 and 15). The primary end point was the objective response rate (ORR), and secondary endpoints included progression-free survival (PFS), duration of response (DOR), and tolerability. Translational and preclinical studies were conducted in KRAS-mutant CRC.

Results: Among the 53 patients treated, the confirmed ORR was 26.4% (95% CI, 15.3 to 40.3). The median DOR was 11.7 months (95% CI, 9.4 to not reached). Grade 3/4 adverse events were reported in 62% of patients. A post hoc analysis revealed that patients with no prior bevacizumab treatment had a significantly higher ORR and longer PFS compared with patients with prior bevacizumab treatment: ORR of 76.9% versus 10.0% (odds ratio of 30.0, P < .001) and median PFS of 14.9 months versus 6.6 months (hazard ratio of 0.16, P < .001). Our translational findings support that prior bevacizumab exposure contributes to onvansertib resistance. Preclinically, we showed that onvansertib inhibited the hypoxia pathway and exhibited robust antitumor activity in combination with bevacizumab through the inhibition of angiogenesis.

Conclusion: Onvansertib in combination with FOLFIRI + bevacizumab showed significant activity in the second-line treatment of patients with KRAS-mutant mCRC, particularly in patients with no prior bevacizumab treatment. These findings led to the evaluation of the combination in the first-line setting (ClinicalTrails.gov identifier: NCT06106308).

Trial registration: ClinicalTrials.gov NCT03829410 NCT06106308 NCT03829410.

FIG 1.

Flow diagram of the phase Ib and phase II. ^aOnce daily on days 1-5 and 15-19 of a 28-day cycle. ctDNA, circulating tumor DNA; Onv, onvansertib.

FIG 2.

Antitumor activity according to prior bevacizumab treatment. (A) Best percentage change from baseline in target lesions. The dashed line at –30% change represents the RECIST v1.1 cutoff to define PR. Confirmed responses are indicated. (B) Kaplan-Meier curve of the PFS. Bev, bevacizumab, CR, complete response; PFS, progression-free survival; PR, partial response.

FIG 3.

Transcriptomic changes associated with bevacizumab treatment. (A) Schematic representation of analysis workflow. (B) Gene set enrichment analysis of differentially expressed genes in patients treated with oxaliplatin + bevacizumab (bevacizumab-exposed) compared with oxaliplatin (bevacizumab-naïve). Bar plots show NES of pathways significantly upregulated in bevacizumab-exposed patients (FDR P < .05). Pathways in bold indicate pathways potentially involved in onvansertib and bevacizumab resistance. Bev, bevacizumab; FDR, false discovery rate; mCRC, metastatic colorectal cancer; NES, normalized enrichment scores.

FIG 4.

Antitumor activity of onvansertib and bevacizumab in KRAS-mutant CRC xenografts and regulation of the hypoxia pathway by PLK1. (A-C) Tumor-bearing mice were treated with vehicle, onvansertib, and/or bevacizumab for 30-33 days. (A) Tumor growth curves. (B) SW620 and LoVo tumor photographs, scale bar: 10 mm. (C) Representative immunohistochemical images of SW620 tumors stained with anti-CD31 antibody (LoVo tumors are shown in Appendix Fig A3), scale bar: 100 μm. CD31+ blood vessels were quantified; graphs represent the mean score ± SEM for each treatment group (n = 6/group). Statistical analyses were performed using one-way ANOVA. **P < .01, ***P < .001, ****P < .0001. (D, E) KRAS-mutant CRC cell lines were treated with DMSO or onvansertib at the indicated doses for 20 hours and then exposed to hypoxia for 4 hours. (D) Representative Simple Western images of HIF1α and β-actin and quantification of HIF1α expression normalized to β-actin and normoxia control sample. Graphs represent mean ± SEM of ≥3 independent experiments. (E) Dot plot showing hallmarks of cancer hypoxia and glycolysis gene set enrichment analyses from comparisons of the indicated groups. The size of each dot corresponds to significance (–log₁₀(P)). ANOVA, analysis of variance; Bev, bevacizumab; FDR, false discovery rate; HIF1α, hypoxia-inducible factor 1α; Hx, hypoxia; Nx, normoxia; Onv, onvansertib; PLK1, polo-like kinase 1.

FIG A1.

Antitumor activity (n = 53). (A) Best percentage change from baseline in target lesions. The dashed line at –30% change represents the RECIST v1.1 cutoff to define PR. Confirmed responses are indicated. (B) Kaplan-Meier curve of the PFS. CR, complete response; PFS, progression-free survival; PR, partial response.

FIG A2.

Association between changes in KRAS-mutant ctDNA and clinical benefit. (A) Percentage change in KRAS-mutant ctDNA from baseline after one cycle of treatment according to best response. Data analyzed by unpaired t-test with Welch correction, **** corresponds to P < .0001. (B) PFS of patients who exhibited a ≥90% decrease in KRAS-mutant ctDNA after 1 cycle of treatment and in patients with less than 90% decrease. CR, complete response; ctDNA, circulating tumor DNA; HR, hazard ratio; PD, progressive disease; PFS, progression-free survival; PR, partial response; SD, stable disease.

FIG A3.

Antitumor activity of Onv and Bev in KRAS-mutant CRC xenografts, related to Figure 4. SW620 and LoVo tumors after treatment with vehicle, Onv, and/or Bev for 30-33 days. (A) SW620 and LoVo tumor photographs, scale bar 10 mm. (B) Representative immunohistochemical images of LoVo tumors stained with anti-CD31 antibody, scale bar 100 μm. (C) Representative immunohistochemical images of SW620 and LoVo tumors stained with anti-BrdU antibody, scale bar 100 μm. Graph represents the % of BrdU+ proliferating tumor cells (mean ± SEM) for each treatment group (n = 6/group). (D) Representative H&E images of SW620 and LoVo tumors, scale bar 50 μm. Tumor cells with apoptotic morphology were scored; graph represents the score (mean ± SEM) for each treatment group (SW620: n = 9/group, LoVo: n = 6/group). Statistical analyses were performed using one-way ANOVA. *P < .05, **P < .01, ***P < .001, ****P < .0001. ANOVA, analysis of variance; Bev, bevacizumab; H&E, hematoxylin and eosin; Onv, onvansertib.

FIG A4.

PLK1 regulates the hypoxia pathway, related to Figure 4. (A, B) KRAS-mutant CRC cells were treated with DMSO or Onv at the indicated doses for 20 hours and then exposed to hypoxia or kept in normoxia for 4 hours. (A) Heatmap of hypoxia-related genes significantly regulated by Onv in HCT116 and SW620 cells based on the RNA-seq analysis. (B) Expression of hypoxia-related genes in LoVo and DLD-1 cells, assessed by RT-qPCR and normalized to the housekeeping gene RPLP0. Bar graphs represent expression relative to Normoxia_DMSO sample. (C, D) SW620 and HCT116 cells were transfected with nontargeting control siRNA (siNTC) or PLK1 targeting siRNA (siPLK1) for 20 hours and then exposed to hypoxia for 4 hours. (C) Left: Simple Western images of PLK1, HIF1α, and β-actin. Right: HIF1α and PLK1 protein expression normalized to β-actin. (D) Expression of hypoxia-related genes assessed by RT-qPCR and normalized to the housekeeping gene RPLP0. Bar graphs represent expression relative to Normoxia_siNTC. (B-D) Data are shown as mean ± SEM of at least three independent biological replicates. HIF1α, hypoxia-inducible factor 1α; Hx, hypoxia; Nx, normoxia; Onv, onvansertib; PLK1, polo-like kinase 1; RT-qPCR, real-time quantitative polymerase chain reaction.

FIG A5.

Graphical summary of the proposed mechanisms of Onv and Bev combination therapy in Bev-naïve and Bev-exposed tumors. (A) In Bev-naïve tumors, the combination of Onv and Bev effectively inhibits tumor cell survival, proliferation, and angiogenesis through the indicated mechanisms. (B) In Bev-exposed tumors, Bev exposure leads to upregulation of mitotic and hypoxia pathways resulting in resistance to both Onv and Bev. Figure was created with BioRender.com. Bev, bevacizumab; HIF1α, hypoxia-inducible factor 1α; Onv, onvansertib.