The WEE1 inhibitor azenosertib broadly enhances efficacy of antibody-drug conjugates with topoisomerase I and microtubule inhibitor payloads

Summary

Antibody-drug conjugates (ADCs) have transformed targeted cancer therapy, yet strategies to overcome resistance and enhance efficacy remain needed. Since ADCs exert anti-tumor effects via DNA damage or mitotic disruption, combining them with WEE1 inhibitors represents a rational approach. We investigated the selective WEE1 inhibitor azenosertib in combination with ADCs carrying topoisomerase I inhibitor (TOP1i) or microtubule inhibitor (MTI) payloads. Azenosertib enhanced the activity of free TOP1i agents and TOP1i-based ADCs (trastuzumab deruxtecan [T-DXd] and sacituzumab govitecan), increasing DNA damage and apoptosis, and extended the duration of response while overcoming T-DXd resistance in patient-derived xenografts. Synergistic effects were also observed with MTI agents and MTI-based ADCs (mirvetuximab soravtansine, tisotumab vedotin, and enfortumab vedotin), associated with exacerbated mitotic defects and prolonged mitotic arrest. All combinations enhanced efficacy and were well tolerated in vivo. These findings position azenosertib as a broadly applicable enhancer of cytotoxic-payload ADCs, offering a promising strategy for patients with advanced solid tumors.

Graphical abstract

Highlights

• •Azenosertib is a broadly applicable enhancer of cytotoxic-payload ADCs in vivo
• •Combining azenosertib with antibody-drug conjugates is well tolerated in solid tumor models
• •Azenosertib combined with trastuzumab deruxtecan extends response and overcomes resistance

Therapeutics; Molecular biology; Biotechnology

Introduction

Precision oncology therapeutic development seeks to capitalize on tumor-specific features as vulnerabilities for targeted anti-cancer agents. Antibody-drug conjugates (ADCs) have emerged as a promising therapeutic modality in this effort as they are based on delivery of anti-tumor payloads via antibodies that recognize tumor-specific antigens. ADCs are typically composed of a monoclonal antibody, a potent cytotoxic drug (payload), and a chemical linker that connects the two. This approach leverages the targeted delivery of cytotoxic agents through monoclonal antibodies, enabling selective tumor cell killing while mitigating off-target toxicity associated with traditional cytotoxic chemotherapy.^1,2

The mechanism of action of an ADC is primarily driven by the cytotoxic payload that is released within the target cell upon ADC internalization and linker cleavage, ultimately inducing cell death. Although various payload types, including degraders, immunoactivators, and RNA polymerases, are being explored, currently approved ADCs utilize very potent payloads that either disrupt mitosis (microtubule inhibitors, MTIs) or induce DNA damage, such as topoisomerase I inhibitors (TOP1is).^1,3,4,5 As of March 2025, a total of 16 ADCs targeting both hematological cancers and solid tumors have received FDA approval, and hundreds more are in various stages of research and development, highlighting their growing importance in cancer treatment.^3,6 More than half of the approved ADCs utilize MTIs, while the remainder carry DNA-damaging payloads.^3,6 Therefore, these ADCs mainly exert their anticancer effects either by disrupting microtubule dynamics to cause mitosis defects or by inducing replication stress and DNA damage through mechanisms such as double-strand breaks, alkylation, and crosslinking. Commonly used MTI payloads include auristatins (e.g., MMAE and MMAF) and maytansinoids (e.g., DM1 and DM4), both of which are highly potent compounds with picomolar IC₅₀ values.⁷ Among DNA-damaging agents, TOP1is (e.g., SN-38 and exatecan derivatives) have gained significant prominence as payloads following the clinical successes of trastuzumab deruxtecan (T-DXd) and sacituzumab govitecan (SG).^{2,8,9,10,11,12,13,14,15}

Despite the initial success of ADCs in treating certain cancers, their full potential is often limited by issues such as drug resistance, side effects, and the complex nature of tumor biology.^16,17 To enhance their effectiveness and overcome resistance, ADCs are being explored in combination with other therapeutic modalities, including immune checkpoint inhibitors, chemotherapy, small-molecule inhibitors, and anti-angiogenic agents.^17,18 Since most ADCs exert anti-tumor activity by inducing DNA damage or mitotic defects, combining them with agents that exacerbate replication stress or disrupt DNA repair and cell cycle checkpoints, such as WEE1 inhibitors, presents a rational strategy to enhance the efficacy of ADCs.

WEE1 kinase is a key regulator of the cell cycle, acting as a brake by phosphorylating and inhibiting CDK1 and CDK2. It plays a crucial role at the G1-S and G2-M checkpoints, ensuring proper DNA replication and repair before cell cycle progression.^19,20 Thus, WEE1 is essential for maintaining genomic stability and preventing premature mitotic entry. Inhibition of WEE1 disrupts these regulatory mechanisms, leading to stalled replication forks, DNA damage, and mitotic catastrophe, ultimately driving cancer cell death.^21,22,23,24 As a result, WEE1 inhibitors are being actively explored as promising cancer therapeutics.^20,23 Azenosertib, an orally available and highly selective WEE1 inhibitor,²⁵ has demonstrated potent preclinical anti-tumor activity across multiple cancer types²⁶ and is currently being evaluated in clinical trials for patients with solid tumors. Encouraging preliminary efficacy has been observed, particularly in gynecological cancers, likely due to high levels of replication stress driven by factors such as Cyclin E1 overexpression, which can arise through various mechanisms including CCNE1 amplification.^27,28 As a monotherapy, azenosertib also exhibits meaningful anti-tumor activity in patients with various other solid tumors.²⁶ Opportunities to further improve responses remain, and one strategy is through combination with other anti-cancer agents. Notably, combining azenosertib with DNA-damaging chemotherapeutics such as carboplatin or microtubule-targeting agents like paclitaxel has demonstrated improved efficacy both preclinically and clinically.^27,28 Given that ADCs with DNA-damaging or microtubule-inhibiting payloads share mechanistic similarities with traditional cytotoxic chemotherapeutic agents and can induce replication stress, we hypothesized that combining azenosertib with these ADCs would enhance anti-tumor efficacy.

Herein, we evaluated the combination of azenosertib and ADCs carrying either MTI or TOP1i payloads in various tumor cell lines and in vivo models. Our data first confirmed that azenosertib improved the anti-tumor efficacy of ADCs with both TOP1i and MTI payloads. We then further characterized the underlying mechanisms of action for combining azenosertib with TOP1i or TOP1i-based ADCs and paclitaxel or MTI-based ADCs. Collectively, our preclinical findings strongly suggest that azenosertib holds promise as a generalizable combination partner to enhance the efficacy of both TOP1i- and MTI-based ADCs in patients with advanced solid tumors.

Results

Azenosertib and TOP1 inhibitors demonstrate synergistic effects in vitro and in vivo

TOP1is, such as irinotecan and topotecan, are established cancer treatments that have been widely used in colorectal, ovarian, and small cell lung cancers. These drugs function by disrupting the interaction between DNA and TOP1, an enzyme essential for DNA replication and repair.^9,29 The interplay between DNA repair and cell cycle regulation suggests that combining the WEE1 inhibitor azenosertib with TOP1is could be beneficial. To verify this hypothesis, we evaluated the combination of azenosertib and irinotecan in the colorectal cancer cell line HT-29 and the breast cancer cell line HCC1569. As shown in Figures 1A, S1A, and S1B, strong synergy (ZIP synergy scores >10) was observed across various doses in matrixed cell viability assays in both cell lines. To elucidate the mechanism underlying this observed synergy, we examined changes in key downstream signaling markers associated with cell cycle progression and DNA damage response (Figure 1B). Consistent with our earlier findings,^26,27 azenosertib treatment led to a decrease in phosphorylated CDK1 at tyrosine 15 (pCDK1^Y15), a direct downstream target of WEE1 kinase. Reduction in Cyclin E1 levels was also observed and has been previously described as an indicator of the accelerated transition through G1/S toward M phase caused by inhibition of WEE1 by azenosertib. Treatment with azenosertib consequently causes cells to accumulate in the M phase, where Cyclin E1 levels are lowest.²⁷ Furthermore, while irinotecan and azenosertib alone induced mild to modest DNA damage and apoptosis in both cell lines, combination of the two drugs significantly augmented these effects, as shown by the substantial elevation of γH2AX and cleaved caspase-3 levels (Figure 1B). These effects were more pronounced in HT-29 cells compared to HCC1569 cells, aligning with the higher synergy scores observed in the viability assays (Figures 1A and 1B). We further analyzed cell cycle changes in HT-29 cells by using flow cytometry to measure DNA content and incorporation of 5-ethynyl-2′-deoxyuridine (EdU). Treatment with either azenosertib or irinotecan led to substantial accumulation of cells in the EdU-negative S phase, with 51- and 54-fold increases relative to DMSO control, respectively, indicating a defect in nucleotide incorporation during DNA replication. This effect was further worsened by combination of the two drugs, leading to a 63-fold increase compared to DMSO (Figures S2A and S2B). A similar trend was observed in HCC1569 cells, where azenosertib and irinotecan monotherapies caused 10- and 13-fold increases in EdU-negative S phase cells, respectively, while the combination treatment led to a 17-fold increase relative to DMSO (Figures S2C and S2D).

Figure 1.

Azenosertib and irinotecan demonstrate synergistic effects in vitro and in vivo

(A) Synergy dose matrices for azenosertib and irinotecan in HT-29 (3-day treatment) and HCC1569 (6-day treatment) cells. ZIP synergy model is depicted, where values > 10 indicate the drug combination is synergistic, −10 to 10 indicate additivity, and <−10 indicate antagonism. CRC, colorectal cancer; BC, breast cancer.

(B) Western blot quantification of cell cycle and DNA damage proteins after treating HT-29 cells with DMSO, 350 nM azenosertib, and/or 5 μM irinotecan (left), or after treating HCC1569 cells with DMSO, 150 nM azenosertib, and/or 30 nM irinotecan (right). Arrowheads point to substantial increases in γH2AX and cleaved caspase-3 after combination treatment.

(C) Mean tumor volume ± SEM of HCC1569 xenografts in NOG mice treated for 25 days (n = 9/group). Percent TGI relative to vehicle control is indicated. ∗∗∗∗p ≤ 0.0001 for all comparisons to vehicle and for combination compared to azenosertib; ∗p ≤ 0.05 for combination compared to irinotecan, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(D) Mean percent change in body weight relative to day 0 (ΔBW) ± SEM. Black dashed line indicates baseline; red dashed line indicates cutoff of −15% change. ∗∗∗∗p ≤ 0.0001 for combination compared to vehicle; p > 0.05 (ns) for all other group comparisons, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(E) Mean tumor volume ± SEM of OV-90 xenografts in NOD/SCID mice treated for 25 days (n = 8/group). Percent TGI relative to vehicle control is indicated. ∗∗∗∗p ≤ 0.0001 for all comparisons to vehicle and for combination compared to azenosertib; ∗∗∗p ≤ 0.001 for combination compared with irinotecan, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(F) Mean percent change in body weight relative to day 0 (ΔBW) ± SEM. Black dashed line indicates baseline; red dashed line indicates cutoff of −15% change. p > 0.05 (ns) for all group comparisons, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

We next evaluated if this in vitro synergy could be translated into in vivo anti-tumor efficacy. In the HCC1569 breast cancer xenograft model, azenosertib combined with irinotecan achieved 99% tumor growth inhibition (TGI), which was significantly superior to azenosertib (80% TGI) or irinotecan (42% TGI) alone (p ≤ 0.0001 and p ≤ 0.05, respectively; Figure 1C; Table S1). Similarly, in the OV-90 ovarian cancer model, the combination significantly improved efficacy, resulting in 89% TGI, compared to moderate TGI from azenosertib (53%) or irinotecan (60%) alone (p ≤ 0.0001 and p ≤ 0.001, respectively; Figure 1E; Table S1). Combining azenosertib with topotecan, another approved TOP1i, in the OV-90 model also improved TGI (72%) over either monotherapy (10% with topotecan, 53% with azenosertib), although the overall efficacy was less pronounced at the tested doses, and only the comparison to topotecan alone was statistically significant (p ≤ 0.0001; Figure S3A; Table S1). Body weight loss for azenosertib in combination with irinotecan was more prominent in the HCC1569 study (Figure 1D; maximum average weight loss of 14% in NOG mice); however, the same treatment did not exceed 8% loss in the OV-90 model (Figure 1F; NOD SCID mice), suggesting that the increased weight loss in the HCC1569 study may be due to the use of a more immunocompromised host strain (NOG mice). In the OV-90 model, both combinations with irinotecan and topotecan were generally tolerated, with less than 10% average body weight loss observed (Figures 1F and S3B). Together, these data strongly suggest that azenosertib synergizes with TOP1i, both in vitro and in vivo, by increasing DNA damage and cell cycle dysregulation.

Azenosertib improves the anti-tumor activity of a TROP2 ADC carrying a TOP1 inhibitor payload

SG is the first approved anti-TROP2 ADC for treating triple-negative breast cancer (TNBC) and hormone receptor (HR)-positive, HER2-negative breast cancer.^10,11 We assessed the effects of combining azenosertib and SG or its payload SN38 on in vitro cell viability of a TROP2-expressing TNBC cell line MDA-MB-468 (Figures 2A, S1C, S1D, and S4A). Consistent with what was observed with irinotecan, azenosertib in combination with SG or SN38 showed synergistic effects across various doses (Figure 2A). Pharmacodynamic analysis revealed that both SN38 and SG as single agents led to moderately increased pCHK1, a marker of replication stress,^22,30 as early as 8 h after treatment (Figure 2B). Both treatments led to increased DNA damage at 48–72 h as shown by increased γH2AX. These effects were further augmented when combined with azenosertib (Figure 2B), suggesting that ADC treatment sensitizes to WEE1 inhibition through increased replication stress. Of note, treatment with SN38 and SG also increased pCDK1^Y15 levels, a direct downstream target of WEE1 kinase, implying that the replication stress and DNA damage induced by the ADC or its TOP1i payload led to activation of WEE1.

Figure 2.

Azenosertib and sacituzumab govitecan demonstrate synergistic effects in vitro and in vivo

(A) Synergy dose matrices for azenosertib and SN38 (4-day treatment, left) or azenosertib and SG (5-day treatment, right) in MDA-MB-436 cells. ZIP synergy model is depicted, where values > 10 indicate the drug combination is synergistic, −10 to 10 indicate additivity, and <−10 indicate antagonism. SG, sacituzumab govitecan.

(B) Western blot quantification of cell cycle and DNA damage proteins after treating MDA-MB-468 cells with DMSO, 650 nM azenosertib, and/or 2 nM SN38 (left), or with DMSO, 300 nM azenosertib, and/or 1 nM SG (right). TNBC, triple-negative breast cancer

(C) Mean tumor volume ± SEM of MDA-MB-231 xenografts in BALB/c nude mice treated for 46 days (n = 8/group). Percent TGI relative to vehicle control is indicated. p ≤ 0.0001 for SG and combination compared to vehicle; ∗∗∗∗p ≤ 0.0001 for combination compared to azenosertib alone; p > 0.05 (ns) for azenosertib compared to vehicle and for combination compared to SG alone; two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(D) Spider plots depicting individual MDA-MB-231 tumor volumes for drug-treated group. Dotted line at 500 mm³ indicates when tumors doubled in size relative to the starting volume.

(E) Mean percent body weight change (ΔBW) ± SEM. Black dashed line indicates baseline; red dashed line indicates cutoff of −15% change. ∗p ≤ 0.05 for combination compared to vehicle, although overall weight gain indicates good tolerability; p > 0.05 (ns) for all other group comparisons, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

In vivo efficacy was assessed in the MDA-MB-231 TNBC xenograft model, which was reported to have low-to-medium TROP2 expression (Figure S4A) and lower sensitivity to SG monotherapy.^8,31 In this model, azenosertib and SG alone at the tested doses resulted in TGI of 28% and 71%, respectively. The combination, as expected, enhanced efficacy, achieving 91% TGI (p ≤ 0.0001 compared to azenosertib alone; Figure 2C). Although the improvement over SG alone did not reach statistical significance (p = 0.34), further analysis of individual tumor volume changes indicated that by end of study (day 46), all tumors in the combination treatment group remained small (<500 mm³) or exhibited regression, whereas 63% (5/8) or 50% (4/8) of tumors in the azenosertib or SG monotherapy groups, respectively, had more than doubled in size (>500 mm³; Figure 2D). The treatments were well tolerated, with no body weight loss and all groups gaining weight over the course of the study (Figure 2E).

Combination of azenosertib with T-DXd improves anti-tumor efficacy and overcomes T-DXd resistance

Based on positive clinical data in the pivotal DESTINY-Breast01 trial (NCT03248492),¹² T-DXd is the first approved anti-HER2 ADC carrying the TOP1i payload deruxtecan (DXd). Since then, it has continued to achieve remarkable clinical success and recently received tumor-agnostic approval, offering a promising treatment option for patients with various HER2-expressing cancers.³² However, improving responses in HER2-low-expressing solid tumors and overcoming resistance remain important goals. We investigated the combination of azenosertib and T-DXd in multiple preclinical models, including a HER2-positive breast cancer model (HCC1569), a TNBC patient-derived xenograft (PDX) with HER2-low expression (CTG-3103), and a HER2-positive esophageal cancer PDX with clinically derived resistance to T-DXd (STM023).

In the HCC1569 HER2+ breast cancer xenograft model, the combination of azenosertib (60 or 80 mg/kg) with T-DXd resulted in greater anti-tumor efficacy (111% and 114% TGI, respectively, and higher incidence of tumor regression compared to T-DXd monotherapy (Figures 3A and S5A; Table S1). Analysis of individual tumor volume changes at end of study (day 25) showed a higher percentage of tumors achieving partial response ([PR], ≥30% regression [reduction in tumor volume]) and complete response ([CR], 100% regression) in the combination groups, especially when combined with azenosertib at 80 mg/kg (100% PR, 50% CR), compared to T-DXd alone (22% PR, 0% CR) (Figure 3A). Although body weight changes of the T-DXd combination groups were statistically different from that of the vehicle group, the average weight loss remained less than 10%, indicating that azenosertib in combination with T-DXd was generally well-tolerated even in the more sensitive and highly immunocompromised NOG animals (Figure S5B). Given T-DXd’s approval in HER2-low breast cancer, we also tested the combination of azenosertib and T-DXd in CTG-3103, a PDX model of HER2-low TNBC. While T-DXd monotherapy was able to induce regression in most of the tumors, achieving 50% PR on the last day of treatment (day 59) (Figure 3B), tumors started to regrow approximately 2 weeks after treatment cessation (Figure 3C). In contrast, the combination of azenosertib and T-DXd not only led to deeper regressions in all tumors, achieving 100% PR by day 59 (Figure 3B), but also extended the duration of response with no tumor regrowth observed in 88% (7/8) of mice up to 27 days after treatment termination (day 86) (Figures 3C and 3D). Even with the longer duration of dosing in the CTG-3103 study, no body weight loss was observed and weights increased across all treatment groups, further confirming that the combination with T-DXd was well tolerated (Figure S5C).

Figure 3.

Combination of azenosertib with T-DXd improves anti-tumor efficacy and overcomes T-DXd resistance

(A) Percent change in tumor volume (ΔTV) of individual HCC1569 xenografts on day 25 of treatment (n = 9/group). Bars below 0 indicate tumor regression relative to day 0. Dashed lines at −30% and −100% indicate cutoffs used to classify partial response (PR) or complete response (CR), respectively.

(B) Percent change in tumor volume (ΔTV) of individual CTG-3103 patient-derived xenografts (PDX) in nude mice on day 59 of treatment (n = 8/group). Bars below 0 indicate tumor regression relative to day 0. Dashed line at −30% indicates cutoff used to classify PRs.

(C) Mean tumor volume ± SEM of CTG-3103 (n = 8/group) comparing duration of response to T-DXd monotherapy and T-DXd combination with azenosertib. On-treatment period is indicated by the shaded yellow background. Off-treatment monitoring started from day 59 until day 86. Vertical dotted lines denote 7-day intervals since treatment stopped. Red arrow indicates approximate time when tumor regrowth was observed.

(D) Distribution of individual tumor volumes on last day of treatment (day 59) compared to last day of monitoring for tumor regrowth (day 86). Horizontal line denotes the median.

(E) Mean tumor volume ± SEM of T-DXd-resistant STM023 PDX tumors in nude mice treated for 27 days (n = 8/group). Percent TGI relative to vehicle control is indicated. ∗∗∗∗p ≤ 0.0001 for azenosertib and combination compared to vehicle; ∗∗∗p ≤ 0.001 for T-DXd compared to vehicle; ∗∗∗∗p ≤ 0.0001 for combination compared to either monotherapy, two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(F) Percent change in tumor volume (ΔTV) of individual STM023 PDX tumors on day 27 of treatment. Bars below 0 indicate tumor regression relative to day 0. Dashed line at −30% indicates cutoff used to classify PRs.

Even though there is a high rate of initial response to T-DXd in the clinic, patients inevitably stop responding and acquire resistance to treatment.^33,34,35 Overcoming T-DXd resistance is an unmet clinical need and presents an opportunity for the development of new therapeutic strategies. The combination of azenosertib and T-DXd was thus examined in the STM023 PDX model, established from a HER2-positive esophageal cancer patient who was treated with T-DXd for 24 months and eventually progressed on therapy. Treatment with T-DXd monotherapy resulted in minimal efficacy (27% TGI), confirming that the model was indeed T-DXd-resistant, while azenosertib alone showed moderate single-agent activity (56% TGI). The combination of azenosertib and T-DXd, however, strongly inhibited tumor growth, achieving significantly higher TGI of 99% and inducing tumor regression in 25% (2/8) of animals by day 27 (p ≤ 0.0001 compared to either monotherapy; Figures 3E and 3F; Table S1). Consistent with the other tested models, limited body weight loss was observed, and the combination was well-tolerated (Figure S5D). Tumors collected at end of study were subsequently evaluated for pharmacodynamic changes by immunochemistry (Figure S5E). pCDK1^Y15 was downregulated in the azenosertib-treated group as expected, confirming target engagement, but a decrease was not observed in the combination-treated group (Figure S5F). A trend in increased ɣHAX was observed, but differences in cleaved caspase-3 were not detected in the single agent- or combination-treated groups compared to vehicle (Figure S5F). While evidence of WEE1 inhibition and increased DNA damage were overall confirmed in vivo, the apparent lack of increase in apoptosis by drug treatment may be due to the end-of-study collection time point (after 27 days of treatment) when cell death can occur for various other reasons and may no longer be a direct reflection of the treatments themselves.

Collectively, these data from diverse in vivo models suggest that combining azenosertib with T-DXd provides meaningful benefit in both T-DXd-sensitive and -resistant tumors. In T-DXd-sensitive settings, this combination further enhances the depth of anti-tumor activity and extends duration of response. In T-DXd-resistant settings, our data imply that combination with azenosertib has the potential to overcome acquired T-DXd resistance, which is a growing challenge in the clinic.

Azenosertib and microtubule inhibitors or ADCs with microtubule inhibitor payloads demonstrate synergistic effects in vitro

MTIs are the most widely used and commercially successful cytotoxic payloads in ADCs. Several ADCs with MTI payloads have been approved for the treatment of solid tumors, including mirvetuximab soravtansine (MIRV) targeting folate receptor alpha (FRα) in ovarian cancer, tisotumab vedotin (TISO) targeting tissue factor in cervical cancer, and enfortumab vedotin (EV) targeting nectin-4 in urothelial cancer.³

We first checked the combination effects of azenosertib with an MTI payload and corresponding ADC on ovarian cancer cell viability in vitro. As shown in Figures S1E–S1H and S4A, azenosertib demonstrated synergistic activity when combined with the MTI DM4 across various concentrations in FRα-expressing OVCAR3 and OV-90 ovarian cancer cell lines (Figure S4B). Similarly, synergy was also observed in both cell lines when azenosertib was combined with MIRV, which carries DM4 as its payload (Figure 4B). The mechanism underlying the synergy between azenosertib and DM4 or MIRV was analyzed by western blot in OVCAR3. Compared to each single agent, the combination treatments led to marked increases in the levels of phospho-Histone H3 (pHH3), a marker of mitosis. Concurrently, the combinations resulted in elevated levels of cleaved caspase-3, an indicator of apoptosis (Figures 4C and 4D). These data demonstrated that azenosertib in combination with an MTI or MTI-based ADC increased mitotic arrest compared to each single agent. Interestingly, compared with DM4, the pharmacodynamic effects of combining azenosertib with MIRV tended to appear later, as both pHH3 and cleaved caspase-3 increased gradually and became more obvious at later time points (Figures 4C and 4D). This is likely due to MIRV exerting its cell killing function through a process of antigen binding, internalization, lysosomal processing, and payload release, while DM4 may exert its function more immediately as a small molecule.

Figure 4.

Azenosertib demonstrates synergistic effects with MTI-based ADCs or their payloads in vitro

(A and B) Synergy dose matrices for (A) azenosertib and DM4 or (B) azenosertib and mirvetuximab soravtansine in OVCAR3 (left panels) and OV-90 (right panels) cells treated for 7 days. ZIP synergy model is depicted, where values > 10 indicate the drug combination is synergistic, −10 to 10 indicate additivity, and <−10 indicate antagonism. HGSOC, high grade serous ovarian cancer.

(C and D) Western blot quantification of cell cycle and DNA damage proteins after treating OVCAR3 cells with (C) DMSO, 250 nM azenosertib, and/or 0.25 nM DM4 or (D) DMSO, 250 nM azenosertib, and/or 5 nM MIRV.

(E) Synergy dose matrices for azenosertib and MMAE (2-day treatment, left) or azenosertib and TISO (3-day treatment, right) in HeLa cells. ZIP synergy model is depicted, where values > 10 indicate the drug combination is synergistic, −10 to 10 indicate additivity, and <−10 indicate antagonism.

(F) Western blot quantification of cell cycle and DNA damage proteins after treating HeLa cells with DMSO, 300 nM azenosertib, and/or 0.5 nM MMAE (left) or with DMSO, 250 nM azenosertib, and/or 250 nM TISO (right).

We next looked at the cell viability and pharmacodynamic effects of the combination of azenosertib with a different type of MTI, MMAE, and a corresponding ADC, TISO (Figures S1I, S1J, S4E, and S4F). In agreement with what was observed for DM4 and MIRV, the combination of azenosertib with both MMAE and TISO resulted in synergistic inhibition of cell viability in the HeLa cervical cancer cell line (Figure 4E), which was confirmed to express tissue factor (Figure S4C). Increased levels of cleaved caspase-3 were observed with both combinations, suggesting that they ultimately led to increased cell death. Changes in pHH3 levels between the combination with MMAE or with TISO were inconsistent (Figure 4F), however, and further analysis is warranted to fully understand this.

Together, these data demonstrated that azenosertib synergizes in vitro with MTI-based ADCs (MIRV and TISO) and their respective payloads (DM4 and MMAE) in relevant cancer cell lines. This synergy may be associated with increased mitotic arrest and induction of apoptosis.

Azenosertib synergizes with MTIs or ADCs with MTI payloads by exacerbating mitotic defects

To further understand the underlying mechanism of synergy between azenosertib and MTIs or ADCs with MTI payloads, we analyzed cell cycle changes in OVCAR3 cells by looking at DNA content, incorporation of EdU, and pHH3 staining using flow cytometry (Figure 5A). Cells were treated with azenosertib, DM4, or MIRV as single agents or in combination. Paclitaxel, a widely used anti-microtubule chemotherapy agent, was used as a positive control. Each single-agent treatment resulted in an increase of M phase cells compared with DMSO control, as indicated by positive staining of pHH3, a marker of mitosis. Compared to single-agent treatments, the combination of azenosertib with paclitaxel, DM4, or MIRV further increased the proportion of pHH3-positive cells by 4- to 9-fold, 5-fold, or 2 -to 20-fold, respectively, suggesting a potent induction of M phase arrest (Figures 5A–5D).

Figure 5.

Azenosertib in combination with MTIs or mirvetuximab soravtansine exacerbates cell-cycle arrest and mitotic defects

(A) Flow cytometric analysis of cell cycle in OVCAR3 cells treated with DMSO, 300 nM azenosertib, and/or 5 nM paclitaxel for 24 h (top row); with DMSO, 300 nM azenosertib, and/or 0.25 nM DM4 for 24 h (middle row); or with DMSO, 300 nM azenosertib, and/or 5 nM MIRV for 48 h (bottom row). MIRV, mirvetuximab soravtansine.

(B–D) Quantification of the percentage of cells in different cell cycle phases detected in (A). M phase is defined by pHH3+ cells. G1 and G2 are defined by DNA content of 2n and 4n, respectively. S phase is defined by DNA content between 2n and 4n and further classified by EdU status.

(E) Representative immunofluorescence images of OVCAR3 cells (40× magnification) exhibiting normal or abnormal mitosis. Scale bars, 10 μM.

(F–H) Quantification of percentage of pHH3+ cells exhibiting normal or abnormal mitosis upon treatment with (F) DMSO, 300 nM azenosertib, and/or 5 nM paclitaxel for 24 h; with (G) DMSO, 300 nM azenosertib for 24 h, and/or 5 nM MIRV for 48 h; or with (H) DMSO, 300 nM azenosertib, and/or 0.5 nM DM4 for 24 h. Data are represented as means + SEM, derived from analyzing two technical replicates per treatment condition.

To better interpret how these combinations impacted cancer cells and eventually caused cell death, cells were analyzed by high-resolution imaging after treatment. Consistent with flow cytometry data, azenosertib in combination with paclitaxel, DM4, or MIRV increased the percentage of pHH3+ cells to 36%, 47%, and 44%, respectively (Figures S6A–S6D). Furthermore, markedly abnormal mitosis was found in these pHH3+ cells, characterized by different mitotic abnormalities, including misaligned chromosomes, multipolar spindles, and monopolar spindles (Figure 5E). While azenosertib, paclitaxel, MIRV, and DM4 each induced mitotic defects individually, the combination of azenosertib with these MTI agents further increased the proportion of mitotic cells displaying abnormal mitosis to nearly 100% (94.2%–99.8%) (Figures 5F–5H). Moreover, it was noticed that the combination-treated groups showed reduced cell counts, likely due to increased cell death or loss of cells with abnormal mitosis during the wash step of the assay. Based on the distribution of mitotic abnormalities observed, there was no clear pattern to indicate that these treatments preferentially induced a specific type of mitotic defect (Figures S6E–S6G). We next evaluated the azenosertib combination with paclitaxel or DM4 in another ovarian cancer cell line, OV-90. As expected, the combination of azenosertib with paclitaxel or DM4 increased the percentage of pHH3+ cells (Figures S6H–S6J) to ∼20% compared to 6%–12% that was induced by the respective monotherapies. Again, a higher percentage of abnormal mitosis was found within the pHH3-positive cells in the combination treatment groups (Figures S6K and S6L).

To further validate these findings and exclude the possibility of inhibitor-specific effects, we conducted cell cycle analysis on HeLa cells after treatment with single agents or the combination of azenosertib with MMAE or TISO. While both MMAE and TISO only induced M phase arrest (pHH3+) in 5%–8% of cells, the addition of azenosertib further enhanced this effect and increased the percentage of M phase cells to 12%–15% (Figures S7A and S7B). Similarly, azenosertib combination with EV, the anti-nectin-4 ADC carrying an MMAE payload, resulted in a moderate synergistic effect in the nectin-4-expressing MDA-MB-468 cells (Figure S4D), as demonstrated by the cell viability synergy assay (Figure S8A). Western blot analysis showed increased apoptosis in the combination-treated cells compared to the monotherapy-treated cells (Figure S8B). Last, azenosertib combination with EV resulted in a higher proportion of M phase cells by flow cytometry analysis (Figure S8C).

In summary, these data indicate that azenosertib synergizes with MTIs or ADCs with MTI payloads by exacerbating the mitotic defects and mitotic arrest induced by each single agent, ultimately leading to cell death.

Azenosertib significantly enhances the efficacy of ADCs with MTI payloads in vivo

Data from our previous preclinical studies have shown that azenosertib in combination with paclitaxel (MTI) improved anti-tumor efficacy, especially in models with higher levels of Cyclin E1 expression.²⁷ To determine if the in vitro synergy observed between azenosertib and ADCs with MTI payloads translates to in vivo efficacy, we evaluated combinations with MIRV (DM4 payload) and EV (MMAE payload) in relevant xenograft models (Figure 6).

Figure 6.

Azenosertib significantly enhances the efficacy of MIRV and enfortumab vedotin in vivo

(A) Mean tumor volume ± SEM of OV-90 xenografts in NOD/SCID mice treated for 25 days (n = 8/group). Percent TGI relative to vehicle control is indicated. ∗∗∗∗p ≤ 0.0001 for all comparisons to vehicle and for the combination compared to either monotherapy; two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(B) Mean percent change in body weight (ΔBW) ± SEM of mice in efficacy study bearing OV-90 tumors. Black dashed line indicates baseline; red dashed line indicates cutoff of −15% change. p > 0.05 (ns) for all group comparisons; two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(C) Percent change in tumor volume (ΔTV) of individual OV-90 xenografts on day 25 of treatment. Bars below 0 indicate regression relative to day 0. Dashed line at −30% indicates cutoff used to classify partial response (PR).

(D) Mean tumor volume ± SEM of HT-1376 xenografts in NOD/SCID mice treated for 25 days (n = 8/group). Percent TGI relative to vehicle control is indicated. ∗∗∗∗p ≤ 0.0001 for all comparisons to vehicle; ∗∗p ≤ 0.01 for combination compared to azenosertib alone; ∗∗∗∗p ≤ 0.0001 for combination compared to EV alone; two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(E) Mean percent change in body weight (ΔBW) ± SEM of mice in efficacy study bearing HT-1376 tumors. Black dashed line indicates baseline; red dashed line indicates cutoff of −15% change. p > 0.05 (ns) for all group comparisons; two-way repeated measures ANOVA with Tukey’s multiple comparisons.

(F) Percent change in tumor volume change (ΔTV) of individual HT-1376 xenografts on day 25 of treatment. Bars below 0 indicate tumor regression relative to day 0. Dashed line at −30% indicates cutoff used to classify PRs.

The combination of azenosertib with MIRV was first tested in the OV-90 ovarian cancer model, which expresses low-to-medium levels of FRα (Figure S4B). Both azenosertib and MIRV (2.5 mg/kg) alone resulted in low-to-moderate efficacy (TGI 53% and 41%, respectively). The combination, however, significantly inhibited tumor growth, achieving a TGI of 96% (p ≤ 0.0001 compared to vehicle and either monotherapy; Figure 6A; Table S1) and inducing tumor regression in 50% (4/8) of the mice by end-of-study (day 25) (Figure 6C). We also tested the combination of azenosertib and a lower dose of MIRV in OVCAR3, which was demonstrated by others to be more sensitive to MIRV than the OV-90 model.³⁶ Interestingly, MIRV monotherapy at a low dose of 1.2 mg/kg had no effect on OVCAR3 tumor growth, while azenosertib alone resulted in significant efficacy (TGI of 86%, p ≤ 0.0001). Despite the lack of monotherapy activity with the lower dose of MIRV, combining it with azenosertib still improved efficacy, resulting in a TGI of 99% and inducing tumor regression in 44% (4/9) of tumors by day 22 (Figures S9A and S9B). To assess what azenosertib and a higher dose of MIRV could achieve in OVCAR3, we conducted a follow-up test and subsequently increased the MIRV dose to 2.5 mg/kg after one cycle (21 days) of 1.2 mg/kg. The combination with azenosertib and MIRV at 2.5 mg/kg now resulted in much deeper and more sustained regressions in 100% (9/9) of tumors (Figure S9C). Combination treatments with MIRV at either dose level were well tolerated in the OV-90 and OVCAR3 models, as indicated by limited body weight loss, which was not significantly different from that of the vehicle groups (Figures 6B and S9D).

Next, we tested the combination of azenosertib with EV in the HT-1376 bladder urothelial cancer model, which was confirmed to express nectin-4 (Figure S4D). Not surprisingly, the combination significantly improved efficacy, generating a TGI of 98% compared to 72% and 61% by azenosertib and EV alone, respectively (p ≤ 0.01 and p ≤ 0.0001 for combination compared to azenosertib or EV, respectively; Figure 6D; Table S1). Individual tumor volume changes showed that 50% (4/8) of tumors in the combination group demonstrated regression at end of study (day 25) (Figure 6F). Combination treatment with azenosertib and EV was also well tolerated in this model (Figure 6E). Immunohistochemistry conducted on end-of-study HT-1376 tumors demonstrated the expected pharmacodynamic changes in response to drug treatment (Figure S10). Downregulation of pCDK1^Y15 was detected after both azenosertib monotherapy and combination treatment in vivo. γH2AX levels representative of DNA damage were significantly increased after treatment with azenosertib or EV as single agents and were further increased by the combination. Downstream apoptosis and cell death was also significantly higher in the combination group, demonstrating the benefit of combining azenosertib and EV (Figure S10B). Notably, treatment with azenosertib did not result in any significant change in surface expression of nectin-4 (Figure S11A), implying that activity of the combination with EV is likely not due to antigen upregulation by azenosertib. This observation was likewise confirmed in models expressing HER2, TROP2, and FRɑ (Figures S11C–S11H).

Altogether, the combination of azenosertib and MTI-based ADCs in models of clinically relevant indications demonstrated robust anti-tumor activity in vivo, which was accompanied by increased DNA damage and apoptosis.

Discussion

ADCs represent a significant advancement in targeted cancer therapy, providing an antigen-specific method of delivering potent cytotoxic payloads directly to tumor cells. Despite their clinical success, ongoing research to develop novel treatment strategies, including rational combinations, is aimed at improving efficacy across broader patient populations, overcoming drug resistance, and minimizing toxicity.

To this end, our study systematically evaluated the combination of a WEE1 kinase inhibitor, azenosertib, with clinically relevant ADCs (SG, T-DXd, MIRV, TISO, and EV), as well as with their respective payload classes (TOP1is: irinotecan, topotecan, and SN38 and MTIs: paclitaxel, DM4, and MMAE). We consistently observed synergistic anti-tumor activity in vitro across multiple cancer cell lines derived from diverse origins, including colorectal, breast, ovarian, cervical, and bladder. Mechanistically, the synergy between azenosertib and TOP1is and TOP1i-based ADCs was associated with significantly increased DNA damage, replication stress, and subsequent apoptosis compared to the single agents. For combinations with MTIs or MTI-based ADCs, synergy was accompanied by exacerbated mitotic defects followed by increased mitotic arrest, ultimately leading to apoptosis. Importantly, the synergistic effects translated well, resulting in enhanced in vivo efficacy across various xenograft models and demonstrating pharmacodynamic effects in line with in vitro findings. Azenosertib combined with all the ADCs tested in this study consistently resulted in superior TGI, often achieving deep tumor regressions compared to the respective monotherapies. Furthermore, the combination strategy also displayed the ability to extend the duration of response (azenosertib + T-DXd, Figures 3C and 3D) and overcome clinically acquired resistance to an ADC (azenosertib + T-DXd, Figures 3E and 3F).

Our findings are corroborated by an earlier study from DiPeri et al., where the authors showed that another WEE1 inhibitor, adavosertib, enhanced the anti-tumor activity of T-DXd in HER2-expressing cancer models.³⁷ Moreover, both studies tested WEE1 inhibitor and T-DXd combination in PDX models with clinically derived resistance to T-DXd, demonstrating the combination’s potential to still be effective in a post-T-DXd setting. While the work by DiPeri and colleagues indicated that WEE1 inhibition with T-DXd may be especially beneficial in HER2-positive models with co-occurring CCNE1 amplifications, our study revealed that azenosertib improved T-DXd’s efficacy in both CCNE1-amplified (HCC1569, CTG-3103) and Cyclin E1 protein-expressing, non-amplified models (STM023). Recent studies in high-grade serous ovarian cancer indicate that CCNE1 amplification is only one of several mechanisms that can induce Cyclin E1 protein overexpression; consequently, tumors may lack gene-level amplification but still express high levels of Cyclin E1 protein comparable to amplified counterparts as an indicator of replicative stress.^27,38,39 Furthermore, although Cyclin E1 expression is associated with azenosertib’s monotherapy efficacy in platinum-resistant ovarian cancer,^27,40 its importance in the context of combination treatments and in indications other than ovarian cancer warrants further study. To expand on the scope of these findings, and in addition to validating WEE1 inhibitor combination with T-DXd, our study importantly builds on this treatment strategy by establishing preclinical proof-of-concept with azenosertib and various other approved ADCs (SG, MIRV, TISO, and EV) utilizing diverse targeting antibodies and different cytotoxic payloads.

Results from our studies suggest that combining azenosertib with ADCs carrying TOP1i or MTI payloads may extend its therapeutic reach beyond ovarian cancers with baseline Cyclin E1 elevation. Delivering cytotoxic payloads to cancer cells can induce replication stress in ways that parallel the effects of having high Cyclin E1 levels, which prematurely accelerate cells into S phase and drive excessive origin firing, leading to nucleotide depletion and stalled replication forks. TOP1i and other DNA-damaging agents directly introduce DNA strand breaks and impede replication forks,^41,42,43 while MTIs primarily disrupt cell division and induce chromosomal abnormalities that can trigger replication stress in subsequent cell cycles.⁴⁴ Like Cyclin E1 overexpression, both treatment scenarios lead to replication fork stalling and DNA damage that activate and increase reliance on cell cycle checkpoint regulators, which include WEE1. A review of the models used in our study revealed that several are both CCNE1 non-amplified and have low levels of Cyclin E1 protein (Table S2), yet azenosertib combination with ADCs still strongly inhibited tumor growth in Cyclin E1-low models of various indications (OV-90, MDA-MB-231, and HT-1376), demonstrating efficacy outside the context of Cyclin E1-elevated tumors. Although this study did not directly examine replication fork progression, surrogate measures of replication stress including pCHK1 (Figure 2B) and ɣH2AX (Figure 4D) confirmed that the payloads or ADCs themselves induce replication stress. This was exemplified by the MDA-MB-468 Cyclin E1-low cell line where treatment with SN38 or SG as single agents both increased pCHK1 and ɣH2AX, which was also accompanied by an upregulation of WEE1 activity as seen by increased pCDK1^Y15 (Figure 2B). Similar effects on ɣH2AX and pCDK1^Y15 were observed with the MTI-based ADCs, as seen with MIRV in the OVCAR3 Cyclin E1-high cell line (Figure 4D) and with EV in the HT-1376 Cyclin E1-low in vivo xenografts (Figure S10). Based on these data, treatment with cytotoxic ADCs effectively “primes” cancer cells into a state of high replication stress, rendering them exquisitely sensitive to cell cycle checkpoint inhibitors such as azenosertib and implying that the combination therapies could benefit tumors independent of baseline Cyclin E1 expression. Future investigation will continue to elucidate the role of markers of replication stress in predicting sensitivity to these combinations.

Beyond broadening the patient population for azenosertib, our proposed combination strategy has the potential to significantly expand the addressable patient populations and clinical impact of ADCs themselves. For example, MIRV is currently approved for patients with FRα-positive (defined as ≥ 75% viable tumor cells with ≥2+ membrane staining by immunohistochemistry) platinum-resistant ovarian cancer,⁴⁵ a subset representing approximately one-third of the total patient group.⁴⁶ Our data demonstrate that combining azenosertib with MIRV significantly enhanced anti-tumor activity in the OV-90 ovarian cancer model, which exhibits low-to-medium levels of FRα and low Cyclin E1 expression.²⁷ This suggests that the combination approach could extend the clinical benefit of MIRV to patients whose tumors have lower FRα expression, thereby encompassing a larger portion of the ovarian cancer population. In the case of T-DXd, which has broad, tumor-agnostic approval based on HER2 expression, achieving deeper, more durable responses or overcoming resistance via combination with azenosertib could significantly enhance its clinical success across various HER2-expressing tumor types. In particular, combination with azenosertib could conceivably potentiate the activity of T-DXd in patients with HER2-low and -ultralow tumors. The observed efficacy improvements with SG and EV combinations likewise suggest that azenosertib combination therapy could improve patient outcomes in their approved indications, TNBC/HR+ breast cancer and urothelial cancer, respectively. Furthermore, considering the many other ADCs targeting various tumor antigens with TOP1i or MTI payloads that are currently in different stages of development, azenosertib may serve as a broadly applicable combination partner to expand the patient population for ADCs and range of tumor types amenable to these therapies.

Despite the clinical success of ADCs, drug resistance remains a significant challenge. Key resistance mechanisms include loss of surface antigen expression, impaired antibody binding, defective drug transport (e.g., lysosomal dysfunction and increased drug efflux), payload-specific resistance, enhanced cancer cell survival, and tumor microenvironment alterations.^34,47,48,49 Strategies to address ADC resistance mainly include developing novel ADCs or combining ADCs with other therapies.^49,50 We explored combination of azenosertib and T-DXd in a PDX model from a HER2-positive esophageal cancer patient who progressed after 24 months of T-DXd treatment. Monotherapies with T-DXd or azenosertib exhibited limited to moderate activity, whereas their combination almost completely suppressed tumor growth, indicating the potential to overcome acquired T-DXd resistance. While the exact resistance mechanisms in this model are unknown, loss of surface antigen expression was ruled out, as PDX characterization verified that HER2 expression was retained, and likewise, HER2 expression did not significantly change after treatment with azenosertib (Figures S11A and S11B). In addition to HER2, we also explored whether azenosertib treatment may upregulate other surface antigen receptors, but no significant increases in antigen were detected in TROP2, FRɑ, or nectin-4 (Figure S11). Further studies are required to elucidate the resistance mechanisms and better understand how the combination can overcome insensitivity to T-DXd. Given the diversity of potential resistance mechanisms to ADCs, azenosertib combination with T-DXd and other ADCs should be tested in additional resistant models to explore the broader applicability. Importantly, identifying and validating biomarkers of ADC resistance may be crucial for predicting patient response and identifying the most effective combination strategies. In total, these findings provide strong rationale for continued evaluation of azenosertib with T-DXd, as well as other ADCs, to address the critical and emerging unmet need of ADC resistance.

While combination therapy offers the potential for enhanced efficacy, it also raises concerns for potential overlapping toxicities. Previous interim data presented from clinical studies combining azenosertib with conventional chemotherapy, such as carboplatin or paclitaxel, showed acceptable toxicity profiles.²⁸ It is hypothesized that combining azenosertib with ADCs, which leverage targeted delivery, might offer an improved therapeutic window compared to systemic chemotherapy combinations.⁵¹ Indeed, we observed favorable tolerability for the various azenosertib and ADC combinations across all preclinical models tested. Relative to the combination with conventional TOP1i irinotecan, the combination of azenosertib with T-DXd appeared to be better tolerated (less body weight loss) in the same HCC1569 model, while simultaneously achieving deeper tumor response (Figures 1C and 1D vs. Figures 3A and 3B). However, translating these preclinical tolerability observations to clinical toxicity predictions requires additional validation, as differences between animal models and human physiology can impact drug response and toxicity.⁵² Based on reported safety profiles, hematological toxicities such as neutropenia are associated with monotherapy azenosertib, as well as with some ADCs like T-DXd and SG. Therefore, careful evaluation of the safety profile in clinical trials combining azenosertib with specific ADCs will be essential to identify and manage any overlapping or newly emerged toxicities. Given that the mechanisms of action of azenosertib and ADCs are likely to interact, it may be necessary to modify the dosage of both drugs to optimize the treatment regimen. Combining two drugs that are synergistic allows for potential reduction in toxicity by using lower doses of each drug, which should be determined through dose optimization studies.⁵³ Another way to mitigate toxicity is to identify patient populations most likely to benefit from treatment through incorporation of predictive biomarkers. Biomarker-guided patient selection can effectively enrich for tumors with the highest dependency on WEE1 or ADC target expression, thereby maximizing efficacy at lower or more tolerable doses. This precision-based approach has the potential to widen the therapeutic window, reduce unnecessary exposure in non-responders, and inform rational dose optimization strategies in early clinical testing.

While this study focused on ADCs with TOP1i and MTI payloads, the underlying mechanisms suggest that ADCs carrying other DNA-damaging payloads, such as calicheamicins or pyrrolobenzodiazepine (PBD) dimers, should also benefit from combination with azenosertib. These payloads induce distinct types of DNA lesions, with direct double-strand breaks by calicheamicin and highly cytotoxic DNA cross-links by PBD.^54,55 The resulting DNA damage and reliance on cell cycle checkpoints for repair create vulnerabilities that can be exploited by WEE1 inhibition. Multiple studies have demonstrated that WEE1 inhibition enhances anti-tumor immunity by activating the cGAS-STING pathway, promoting CD8⁺ T cell recruitment and killing, thereby enhancing the efficacy of immune checkpoint inhibitors.^56,57,58,59 Thus, combining azenosertib with ADCs having immunomodulator payloads or pursuing triple combination with ADCs and immune checkpoint inhibitors could also be a promising new strategy.^17,18,60,61 Exploring combinations with ADCs carrying other novel payloads, such as targeted protein degraders and apoptosis inducers, is warranted, although the rationale may depend more on specific payload biology and potential cell cycle dependencies.

In conclusion, this study demonstrates that the selective WEE1 inhibitor azenosertib synergizes broadly with ADCs carrying either TOP1i inhibitor or MTI payloads across multiple solid tumor models, including breast, ovarian, bladder, and cervical. This synergy, driven by the potentiation of DNA damage or exacerbation of mitotic defects depending on the payload, was well translated into significantly enhanced anti-tumor effects in vivo, as well as the potential to overcome acquired ADC resistance. These findings strongly support the clinical investigation of azenosertib or other WEE1 inhibitors in combination with various ADCs as a strategy to improve outcomes for patients with advanced solid tumors, potentially expanding the utility and impact of both therapeutic classes. Indeed, the synergistic relationships between WEE1 inhibitors and ADCs like T-DXd or SG are currently being explored in the clinic in HER2+ solid tumors (NCT06364410) and in TNBCs or HR+/HER2− breast cancers (NCT06612203).

Limitations of the study

This study comprehensively investigates the synergistic nature of combining azenosertib with various ADCs and their TOP1i or MTI payloads; however, several limitations should be considered. In vivo combination studies were conducted to demonstrate efficacy proof of concept, and while dose levels were selected to be clinically relevant based on our knowledge of azenosertib and on published literature of ADCs, additional work should be done to test whether lower doses of the ADCs can be used in combination settings. In particular, the dose levels of SG and T-DXd may have been at the higher end of the active range, as they already exhibited strong single agent activity in some of the models tested. Preclinical dose optimization of the combination with azenosertib could further clarify the therapeutic window and improve translation to the clinic. Toxicity assessments in the current study were also limited to monitoring body weight changes, and incorporating broader toxicity readouts like complete blood counts and liver function tests could provide more insight into the safety profile of the combination before entering the clinic. Mechanism-of-action studies largely used in vitro cell culture systems accompanied by western blot, flow cytometry, and microscopy to demonstrate treatment effects on DNA damage, cell death, and cell cycle. Although end-of-study tumor samples from in vivo efficacy studies were used to confirm the in vitro findings, including dedicated in vivo pharmacodynamic studies could strengthen mechanistic conclusions and resolve some of the inconsistent biomarker changes that were noted in the STM023 study. The observation that azenosertib and TOP1i- or MTI-bearing ADCs resulted in robust combination activity across a diverse panel of models with varying mutational and expression profiles suggests that efficacy of the combination may not need a biomarker selection strategy, but mechanistic studies such as knockdown or overexpression of potential sensitizing biomarkers would be informative. Finally, given that clinical landscapes are rapidly evolving as ADCs are moving into earlier lines of therapy, future studies should investigate more ADC-resistant models and consider how mechanisms of ADC resistance may influence subsequent treatment strategies.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Catherine Lee (calee@zentalis.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All raw data that can be shared by the lead contact upon request.
• •This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

The authors thank Pharmaron Inc., Champions Oncology, and XenoSTART for supporting our animal studies, as well as SD HistoPath and Crown Bioscience for IHC support. Funding for all experiments was provided by Zentalis Pharmaceuticals.

Author contributions

X.G. and J.M. conceived the project; X.G., M.R.L., C.L., J.M., and H.C. supervised the research; X.G., E.C., C.L., and J.M. designed the experiments; X.G. and E.C. performed the experiments; X.G., E.C., C.L., and H.C. analyzed the data and made the figures; X.G., C.L., and J.M. wrote the manuscript; all co-authors reviewed and/or revised the manuscript.

Declaration of interests

X.G. is a former employee and shareholder of Zentalis Pharmaceuticals and has filed patent applications related to the subject matter of this manuscript. D.K. is an employee and shareholder of Zentalis Pharmaceuticals. O.H. is an employee and shareholder of Zentalis Pharmaceuticals. E.C. is a former employee of Zentalis Pharmaceuticals. H.C. is an employee and shareholder of Zentalis Pharmaceuticals. F.M.-B. is a consultant for AstraZeneca Pharmaceuticals, Becton Dickinson, Biocartis NV, Calibr (a division of Scripps Research), Daiichi Sankyo, Dava Oncology, Debiopharm, EcoR1 Capital, eFFECTOR Therapeutics, Elevation Oncology, Exelixis, GT Aperion, Incyte, Jazz Pharmaceuticals, LigaChem Biosciences, Lengo Therapeutics, Menarini Group, Molecular Templates, Protai Bio, Ribometrix, SystImmune, Tallac Therapeutics, Tempus, Vir Biotechnology, and Zymeworks; is in the advisory committee for Cybrexa, go Therapeutics, Guardant Health, Harbinger Health, Illumen Therapeutics, Kivu Biosciences, LOXO-Oncology, Mersana Therapeutics, OnCusp Therapeutics, Sanofi Pharmaceuticals, Seagen, Theratechnologies, and Zentalis Pharmaceuticals; receives sponsored research (to the institution) from Jazz Pharmaceuticals, Zymeworks, Aileron Therapeutics, Inc., AstraZeneca, Bayer Healthcare Pharmaceutical, Calithera Biosciences Inc., Curis Inc., CytomX Therapeutics Inc., Daiichi Sankyo Co. Ltd., Debiopharm International, eFFECTOR Therapeutics, Genentech Inc., Guardant Health Inc., Klus Pharma, Takeda Pharmaceutical, Novartis, Puma Biotechnology Inc., and Taiho Pharmaceutical Co.; receives honoraria for Dava Oncology; and receives travel-related funding and reimbursement from European Organisation for Research and Treatment of Cancer (EORTC), European Society for Medical Oncology (ESMO), Cholangiocarcinoma Foundation, and Dava Oncology. M.R.L. is a consultant and shareholder of Zentalis Pharmaceuticals. C.L. is an employee and shareholder of Zentalis Pharmaceuticals and has filed patent applications related to the subject matter of this manuscript. J.M. is a consultant and shareholder of Zentalis Pharmaceuticals and has filed patent applications related to the subject matter of this manuscript.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-Cyclin E1 (CCNE1/2460)	Abcam	Cat# ab238081; RRID: AB_3096040
Rabbit polyclonal anti-cdc2 (CDK1)	Cell Signaling Technology	Cat# 77055S; RRID: AB_2716331
Rabbit monoclonal anti-WEE1	Cell Signaling Technology	Cat# 13084S; RRID: AB_2713924
Rabbit monoclonal anti-Phospho-Histone H2A.X (Ser139) (20E3)	Cell Signaling Technology	Cat# 9718S; RRID: AB_211800
Rabbit monoclonal anti-Phospho-Histone H3 (Ser10) (D7N8E)	Cell Signaling Technology	Cat# 53348S; RRID: AB_2799431
Rabbit monoclonal anti-Phospho Chk1 (Ser345) (133D3)	Cell Signaling Technology	Cat# 2348L; RRID: AB_331212
Rabbit polyclonal anti-Phospho-Chk2 (Thr68)	Cell Signaling Technology	Cat# 2661; RRID: AB_331479
Rabbit monoclonal anti-Chk2 (EPR4325)	Abcam	Cat# ab109413; RRID: AB_10863751
Rabbit monoclonal anti-Phospho cdc2 (CDK1) (Tyr15) (10A11)	Cell Signaling Technology	Cat# 4539S; RRID: AB_560953
Rabbit monoclonal anti-Cleaved Caspase-3 (Asp175) (5A1E)	Cell Signaling Technology	Cat# 9664L; RRID: AB_2070042
Mouse monoclonal anti-GAPDH (6C5)	Thermo Fisher Scientific	Cat# AM4300; RRID: AB_2536381
Rat monoclonal APC anti-human Folate Receptors α and β (FR-αβ)	BioLegend	Cat# 391803; RRID: AB_2721304
Mouse monoclonal PE anti-human TACSTD2 (TROP2)	BioLegend	Cat# 363803; RRID: AB_2572021
Rabbit monoclonal anti-Nectin-4/PVRL-4	Proteintech	Cat# 83223-4-RR; RRID: AB_3670907
Alexa Fluor® 488 Donkey anti-rabbit IgG (minimal x-reactivity)	BioLegend	Cat# 406416; RRID: AB_2563203
Mouse monoclonal APC anti-human CD142	BioLegend	Cat# 365205; RRID: AB_2564567
Mouse monoclonal Alexa Fluor® 488 anti-Histone H3 Phospho (Ser10)	BioLegend	Cat# 650804; RRID: AB_10918435
Mouse monoclonal anti-α-tubulin (DM1A)	Cell Signaling Technology	Cat# 3873; RRID: AB_1904178
Rabbit monoclonal anti-Phospho-Histone H3 (Ser10) (D2C8)	Cell Signaling Technology	Cat# 3377; RRID: AB_1549592
Goat anti-Mouse IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ Plus 488	Thermo Fisher Scientific	Cat# A32723; RRID: AB_2633275
Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 647	Thermo Fisher Scientific	Cat# A31573; RRID: AB_2536183
Rabbit monoclonal anti-FOLR1/FBP/FRa (EPR20277)	Abcam	Cat# ab221543; RRID: AB_2920895
Rabbit monoclonal anti-HER2/ErbB2 (D8F12)	Cell Signaling Technology	Cat# 4290; RRID: AB_10557104
Rabbit monoclonal anti-Trop2 (EPR20043)	Abcam	Cat# ab214488; RRID: AB_2811182
Rabbit monoclonal anti-Nectin-4 (E5Q1W)	Cell Signaling Technology	Cat# 41798
Rabbit monoclonal IgG (EPR25A)	Abcam	Cat# ab172730; RRID: AB_2687931
Biological samples
CTG-3103 PDX	Champions Oncology Inc.	N/A
STM023 PDX	XenoSTART	N/A
Chemicals, peptides, and recombinant proteins
DM4	MedChemExpress	HY-12454
Enfortumab vedotin	MedChemExpress	HY-P99016B
Irinotecan	MedChemExpress	HY-16562
Mirvetuximab soravtansine	MedChemExpress	HY-132258A
MMAE	MedChemExpress	HY-15162
Sacituzumab govitecan	MedChemExpress	HY-132254
SN38	MedChemExpress	HY-13704
Tisotumab vedotin	MedChemExpress	HY-152963
Topotecan hydrochloride	MedChemExpress	HY-13768A
Trastuzumab deruxtecan	MedChemExpress	HY-138298A
Critical commercial assays
CellTiter-Glo (CTG) Luminescent Cell Viability Assay	Promega	G7573
Pierce BCA Protein Assay kit	Thermo Fisher Scientific	23227
Jess Simple Western System	ProteinSimple	SM-W004
Click-iT™ Plus EdU Alexa Fluor™ 647 Flow Cytometry Assay Kit	Thermo Fisher Scientific	C10634
Experimental models: Cell lines
HCC1569	ATCC	Cat# CRL-2330; RRID: CVCL_1255
HeLa	ATCC	Cat# CCL-2; RRID: CVCL_0030
HT-1376	ATCC	Cat# CRL-1472; RRID: CVCL_1292
HT-29	ATCC	Cat# HTB-38; RRID: CVCL_0320
MDA-MB-231	ATCC	Cat# CRM-HTB-26; RRID: CVCL_0062
MDA-MB-468	ATCC	Cat# HTB-132; RRID: CVCL_0419
OVCAR3	ATCC	Cat# HTB-161; RRID: CVCL_0465
OV-90	ATCC	Cat3 CRL-3585; RRID: CVCL_3768
Experimental models: Organisms/strains
BALB/c nude mice (BALB/cNj-Foxn1nu/Gpt)	GemPharmatech Biotech Co., Ltd.	RRID: IMSR_GPT:D000521
NOD SCID mice (NOD/ShiLtJGpt-Prkdcem26Cd52/Gpt)	GemPharmatech Biotech Co., Ltd.	Strain T001492; RRID: IMSR_GPT:T001492
NOD SCID mice (NOD.Cg-Prkdcscid/Jnifdc)	Beijing Anikeeper Biotech Co., Ltd.	N/A
NOG mice (NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl)	Beijing Vital River Laboratory Animal Technology Co., Ltd.	Strain 408
Athymic nude mice (Crl:NU(NCr)-Foxn1nu)	Charles River Laboratories	RRID: IMSR_CRL:490
Software and algorithms
GraphPad Prism v8	GraphPad software	https://www.graphpad.com/
SynergyFinder	SynergyFinder	http://www.synergyfinderplus.org/
FlowJo	FlowJo software	https://www.flowjo.com/
Compass software v6.2	ProteinSimple	https://www.bio-techne.com/resources/instrument-software-download-center
Harmony high-content analysis software	Revvity	https://www.revvity.com/product/harmony-5-2-office-revvity-hh17000019

Experimental model and study participant details

Cell lines and cell culture

HT-29 cells purchased in 2022 were cultured in McCoy’s 5A (ATCC #30–2007, Manassas, VA) supplemented with 10% FBS (Gibco #A38400-02, Waltham, MA) and 1% penicillin-streptomycin (Gibco #15140–122). HCC1569 cells purchased in 2022 were cultured in RPMI 1640 medium (ATCC modification) (Gibco #A1049101) supplemented with 10% FBS and 1% penicillin-streptomycin. MDA-MB-231 and MDA-MB-468 cells purchased in 2021 were cultured in DMEM/F12 medium (Gibco #11320–033) supplemented with 10% FBS and 1% penicillin-streptomycin. OVCAR3 cells purchased in 2020 were cultured in RPMI 1640 medium (ATCC modification) supplemented with 20% FBS, 10 μg/mL insulin (Sigma #I9278, St. Louis, MO), and 1% penicillin-streptomycin. OV-90 cells purchased in 2021 were cultured in 1:1 mixture of MCDB 105 (Cell Applications #117–500, San Diego, CA) containing 1.5 g/L sodium bicarbonate (Gibco #25080–094) and Medium 199 (Gibco #11150–059) supplemented with 15% FBS and 1% penicillin-streptomycin. HeLa and HT-1376 cells purchased in 2019 and 2024, respectively, were cultured in EMEM medium (ATCC #30–2003) supplemented with 10% FBS and 1% penicillin-streptomycin. All cancer cell lines were maintained at 37°C in a humidified incubator containing 5% CO₂. Cell lines were carried for no more than 20 passages. Integrity (identity and Mycoplasma testing) of all cell lines was authenticated by the vendor at the time of purchase. Except for MDA-MB-468, HeLa, and HT-1376, all cell lines were re-tested and validated using short tandem repeat (STR) and Mycoplasma PCR most recently on April 26, 2024 (IDEXX BioAnalytics, Columbia, MO).

Mouse models

All mouse studies were approved and conducted in compliance with regulations and guidelines of the Institutional Animal Care and Use Committee (IACUC) of each vendor, including Pharmaron (China), Champions Oncology (Rockville, MD), XenoSTART (San Antonio, TX), and our in-house facility (San Diego, CA). IACUCl approval numbers: ON-CELL-XEN-06012023 (OV90, HCC1569); ON-CELL-XEN-06012024 (HT-1376); (CTG-3103, 2023-TOS-001); (STM023, 09–001). For cell line-derived xenograft (CDX) models, 6–8-week-old female BALB/c nude, NOD/SCID, or NOG mice were injected subcutaneously with 0.1–2×10⁷ tumor cells resuspended in a 1:1 mixture of PBS or corresponding medium and Matrigel. For patient-derived xenograft (PDX) models, 6–12-week-old female athymic nude mice were implanted subcutaneously with tumor fragments. Mouse health was monitored daily, and tumor volume and body weights were measured two-three times per week.

Method details

Flow cytometry surface antigen expression

Single cell suspensions were incubated with fluorophore-conjugated antibodies on ice for 20 min, protected from light. For surface nectin-4 staining, single cell suspensions were incubated with a primary antibody against nectin-4 (1:400) for 20 min, then incubated with secondary antibody on ice for 20 min, protected from light. Staining and wash steps used Cell Staining Buffer (CSB) (BioLegend #420201, San Diego, CA). Analysis was performed using the Attune™ NxT flow cytometer (Thermo Fisher Scientific, Waltham, MA) and FlowJo software (Ashland, OR).

Flow cytometry cell cycle analysis

Cell cycle analysis was performed with Click-iT™ Plus EdU Alexa Fluor™ 647 Flow Cytometry Assay Kit (Thermo Fisher Scientific #C10634) followed by FxCycle Violet (Thermo Fisher Scientific #F10347) staining for DNA content. Cells were seeded in 12-well plates and treated the following day. After 24 or 48 h of drug exposure, cells were EdU labeled and fixed/permeabilized according to manufacturer instructions. For phospho-histone H3 staining, cells were incubated with phospho-histone H3 (Ser10) Alexa Fluor 488 conjugated antibody (1:25) for 30–60 min at room temperature in the dark, before Click-iT detection reaction was performed according to manufacturer instructions. Lastly, cells were stained with FxCycle Violet solution for 30 min at room temperature in the dark. Samples were analyzed with the Attune™ NxT flow cytometer. Cell cycle progression was analyzed using FlowJo and plotted using GraphPad Prism. EdU+ cells were assigned to S phase and pHH3+ cells were assigned to M phase, irrespective of DNA content.

In vitro cell viability synergy analyses

Cells were plated in 96-well white-walled, clear-bottom plates (Corning Life Sciences #3903, Corning, NY) in standard tissue culture conditions. Due to different drug sensitivities of cell lines, concentration ranges tested in synergy assays were chosen to approximately cover each drug’s single agent IC₅₀ in each respective cell line. Treatments with these different concentrations of azenosertib and/or the indicated drugs were performed the following day. Effects on cell viability were assessed using CellTiter-Glo (CTG) Luminescent Cell Viability Assay (Promega #G7573, Madison, WI) according to manufacturer instructions. Synergy between drugs was then calculated by inputting cell viability data into the Zero Interaction Potency (ZIP) model from SynergyFinder (http://www.synergyfinderplus.org/). ZIP scores >10 indicated synergistic effects, ZIP scores from −10 to 10 indicated additive effects, and ZIP scores < −10 indicated antagonism.

Western blotting

Cells were seeded in 6-well or 6 cm plates for evaluating biomarker changes after drug treatment. Doses were determined based on results from combination cell viability assays, and monotherapy or combination treatments were administered the following day. Cells were harvested at multiple timepoints, lysed and sonicated in cold RIPA buffer (Sigma-Aldrich #R0278, St. Louis, MO) containing protease and phosphatase inhibitors for 15 min, and centrifuged at 4°C, 15,000 rpm for 15 min. Protein concentrations were measured using the Pierce BCA Protein Assay kit (Thermo Fisher Scientific #23227, Waltham, MA). Western blotting was performed using the Jess Simple Western System (ProteinSimple, San Jose, CA). Expression levels and phosphorylation of target proteins were assessed according to the manufacturer’s standard method for 12–230 kDa Jess separation module (SM-W004). Protein lysates were mixed with 0.1X sample buffer and fluorescent 5X master mix to achieve a final sample concentration of 1–2 μg/μL. Protein samples were denatured at 95°C for 5 min 5 μL of biotinylated ladder and samples were loaded into individual wells in the sample plate. Primary antibodies were diluted according to manufacturer instructions, and HRP-conjugated anti-rabbit or anti-mouse secondary antibodies were applied according to the Simple Western kit instructions. All subsequent separation, immunodetection, and analysis steps were performed automatically by the instrument. Chemiluminescence reactions with antibodies were measured, and digital blot images were constructed by the Compass software (Version 6.2; ProteinSimple).

High-content confocal imaging

OVCAR3 and OV-90 cells were seeded in 96-well plates overnight and treated for 24 or 48 h. Cells were fixed in 4% paraformaldehyde for 20 min, washed, permeabilized in 0.5% Triton X-100 in PBS for 10 min on ice, and blocked with 3% BSA in TBS for 1 h. Cells were then incubated with primary antibodies against α-tubulin (1:500) and phospho-Histone H3 (1:500) in TBS containing 3% BSA at room temperature for 1 h, followed by secondary antibodies (1:1000) in TBS containing 3% BSA for 1 h at room temperature, protected from light. After washing with TBST, cell nuclei were stained with DAPI (Invitrogen R37606, Waltham, MA) in PBS for 20 min at room temperature. Images were acquired in confocal mode by the Operetta CLS High-Content Imaging System (PerkinElmer Life Sciences, Boston, MA) with a 40× water objective. Approximately 0.4–1.1×10⁴ cells per condition were analyzed for mitotic arrest and spindle phenotypes. Harmony high-content analysis software (PerkinElmer Life Sciences, Boston, MA) was used for detecting total number of cells and phospho-Histone H3-positive cells. PhenoLOGIC Machine Learning algorithm (PerkinElmer Life Sciences, Boston, MA) was then applied to identify abnormal mitotic cells from total mitotic cells via calculation of signal intensity and texture of α-tubulin, phospho-Histone H3, and DAPI in the phospho-Histone H3-positive cells. This study was performed by BioDuro-Sundia (Shanghai, China).

In vivo xenograft studies

Mice were randomized into treatment groups once tumors reached an average tumor volume (TV) of 150–250 mm³. Azenosertib was administered orally in 20% HP-β-CD in water. ADCs (MedChemExpress #HY-132258A, #HY-138298A, #HY-132254, #HY-P99016B, Monmouth Junction, NJ) were administered i.v. in sterile PBS. Topotecan hydrochloride (MedChemExpress #HY-13768A, Monmouth Junction, NJ) and irinotecan (MedChemExpress #HY-16562, Monmouth Junction, NJ) were administered i.p. in sterile PBS pH 4–9 or 5% DMSO +95% saline pH 4–9, respectively. Percent tumor growth inhibition (TGI) was calculated using the formula: (1-(T_d – T₀)/(C_d – C₀)) × 100, where T_d and C_d are the mean tumor volumes of the Treated and Control animals on the day of analysis, and T₀ and C₀ are the mean tumor volumes of the Treated and Control animals at the start of treatment, respectively. Percent change in tumor volume (ΔTV) was calculated as: ((TV_d – TV₀)/TV₀) × 100, and percent change in BW (ΔBW) was calculated as: ((BW_d – BW₀)/BW₀) x 100. When tumor regression was observed in studies, efficacy data was preferentially visualized using bar graphs depicting percent ΔTV of individual mice instead of line (mean TV) or spider (individual TV) plots. Additional experimental details are available in Table S1.

Immunohistochemistry

Tumor tissues collected from efficacy studies were fixed in 10% neutral buffered formalin and paraffin embedded. Blocks were cut into 5 μm sections using a Leica RM2135 microtome. For PD biomarker analysis, sections were dewaxed and rehydrated in distilled water before antigen retrieval with 200 mL citrate buffer (pH 6.0) and heating for 10 min. Once the citrate buffer cooled, slides were removed and placed into an autostainer (ThermoFisher, Waltham, MA). Tissues were blocked for 10 min using 3% H₂O₂ and 3% normal goat serum, then primary antibodies for pCDK1 (1:100), ɣH2AX (1:200), Cleaved Caspase-3 (1:250) were incubated for 1 h. Primaries were detected using goat-anti-rabbit polymer conjugated to HRP, and visualizations were achieved by incubation of DAB substrate and counterstained with hematoxylin. Slides were dehydrated in ethanol and mounted in xylene based mounting media before scanning at 40× magnification using the NanoZoomer Digital Slide System (Hamamatsu NDP2.0-HT, Bridgewater, NJ). A pathologist scored tissues for positive staining by examining three 40× magnification fields per sample. For surface antigen analysis, slides were processed and stained using the Bond RX automatic IHC system (Leica Biosystems, Nussloch, Germany). Slides were incubated with antibodies against HER2 (1:200), TROP2 (1:500), FRɑ (1:1000), nectin-4 (1:200), or rabbit IgG isotype control (1:3000). Slides were scanned at 40× magnification using the NanoZoomer Digital Slide System (Hamamatsu NDP2.0-HT, Bridgewater, NJ) and analyzed with the HALO^TM Quantitative Pathology Image Analysis platform (Indica Labs, Albuquerque, NM).

Quantification and statistical analyses

For in vivo studies, grouped tumor volume and body weight change data were analyzed by two-way repeated measures (RM) ANOVA with Tukey’s multiple comparisons test using GraphPad Prism. Results were statistically significant if p ≤ 0.05 (∗), p ≤ 0.01 (∗∗), p ≤ 0.001 (∗∗∗), or p ≤ 0.0001 (∗∗∗∗). A minimum of 8 animals per group were used in each study. Results were represented as mean ± standard error of the mean (SEM).