Cell cycle inhibitors, particularly MTAs, are standard chemotherapy drugs for breast cancer⁵⁴. To investigate whether CMPD1 induces G2/M arrest in breast cancer cells, we treated MDA-MB-231 cells, a triple-negative breast cancer (TNBC) cell line, with varying concentrations of CMPD1 and analyzed its effects on cell cycle progression using flow cytometry. The results demonstrated that CMPD1 effectively arrested MDA-MB-231 cells in the G2/M phase at concentrations ranging from 1 to 10 µM (Fig. 1B), consistent with previous observations in glioblastoma cells⁵³. Given that flow cytometry cannot distinguish between G2 phase and mitosis, we performed high-temporal resolution live-cell imaging to further dissect the impact of CMPD1 on cell cycle progression across various breast cancer cell lines, including MDA-MB-231 (TNBC), CAL-51 (TNBC), and T-47D (luminal A)^{56, 57}. Our findings revealed that CMPD1 induced a severe prometaphase arrest across all tested breast cancer cell lines, regardless of subtypes, with most cells remaining arrested in prometaphase for over 10 hours, while control cells divided within 30-60 min (Fig. 1C-D and Supplementary Movies 1-2). Since CAL-51 cells harbor wild-type TP53, whereas the other two cell lines do not^{58, 59}, these results suggest that CMPD1 induces a robust prometaphase arrest in breast cancer cells through a p53-independent mechanism.

We demonstrated that concentrations exceeding 1 µM of CMPD1 effectively induced a robust prometaphase arrest in multiple breast cancer cell lines (Fig. 1D). In contrast, recent studies on MTA treatment for breast cancer have identified that the clinically relevant concentration of PTX in tissue culture cells ranges from 5 to 50 nM⁶⁰. Within this range of concentration, PTX does not induce severe mitotic arrest as observed at higher concentrations (> 1 µM), but it significantly increases the incidence of mitotic errors, thereby promoting chromosomal instability (CIN) and ultimately leading to cancer cell death⁶⁰. Given these findings, we next explored whether sub-µM concentrations of CMPD1 could induce CIN in breast cancer cells. To this end, we treated MDA-MB-231 and CAL-51 cells, along with RPE1 non-transformed retinal pigment epithelial cells (serving as a normal cell control), with either 10 or 50 nM CMPD1. We then compared the impact of CMPD1 on chromosome segregation using live-cell imaging. In the absence of CMPD1, no significant differences in mitotic duration were observed among RPE1, MDA-MB-231, and CAL-51 cells (Fig. 2A-C). However, upon treatment with 10 or 50 nM CMPD1, all cell lines were arrested in prometaphase (Fig. 2A-C). Notably, MDA-MB-231 and CAL-51 cells displayed significantly prolonged mitotic duration compared to RPE1 cells (Fig. 2A-C), indicating greater sensitivity of breast cancer cells to CMPD1 treatment relative to normal cells. Given that mitotic errors are a direct cause of CIN, we next assessed the frequency of mitotic errors including misaligned chromosome, chromosome bridge, lagging chromosome and multipolar division under these conditions (Fig. 2D). Both MDA-MB-231 and CAL-51 cells exhibited a significantly higher rate of mitotic errors when exposed to 10 or 50 nM CMPD1 (Fig. 2E). For instance, approximately 90% and 100% of CAL-51 cells experienced mitotic errors in the presence of 10 and 50 nM CMPD1, respectively (Fig. 2E). Conversely, nearly 100% of RPE1 cells underwent faithful cell division even at the same concentrations of CMPD1 (Fig. 2E). To determine whether CMPD1 selectively targets cancer cells compared to other MTAs, we evaluated mitotic error rates in the same cell lines treated with a clinically relevant concentration of PTX (10 nM). Our results demonstrated that both RPE1 and breast cancer cell lines displayed comparably high rates of mitotic errors when treated with 10 nM PTX treatment (Fig. 2F), consistent with previous observations⁶⁰. In summary, unlike PTX, CMPD1 induces CIN with selective toxicity toward breast cancer cells.

To further investigate the selectivity of CMPD1 in breast cancer cells, we conducted a CMPD1 washout assay using the above cell lines, aiming to recapitulate the clinical condition where the concentration of chemotherapy drugs in patients is diluted due to the “drug holiday” between regular treatments. In this assay, we treated the cells with 2 µM CMPD1 for 12 hours, followed by a washout before imaging. We observed that the mitotic duration was significantly prolonged in both MDA-MB-231 and CAL-51 cells upon CMPD1 washout (85 min and 151 min, respectively) compared to RPE1 cells (14 min) (Fig. 2G-H). This suggests that breast cancer cells struggle to recover from CMPD1-induced mitotic arrest, whereas normal cells can immediately resume progression to proper anaphase. Notably, 90% and 81% of mitotic cells in MDA-MB-231 and CAL-51 cells lines, respectively, exhibited mitotic errors, whereas 50% of RPE1 cells displayed normal chromosome segregation (Fig. 2G, I). These findings reinforce the hypothesis that breast cancer cells are more sensitive to CMPD1 treatment under clinically relevant conditions. Interestingly, approximately 60% of MDA-MB-231 cells that entered mitosis after CMPD1 washout still exhibited mitotic errors, while only ∼10% of RPE1 cells showed errors. This suggests a prolonged impact of CMPD1 on mitosis in breast cancer cells (Fig. S1C). Consistent with the increased sensitivity in breast cancer cells to CMPD1, MDA-MB-231 cells exhibited a significantly higher frequency of apoptotic cell death during or shortly after mitosis within 24 of CMPD1 washout, compared to RPE1 cells (Fig. S1D-E). Collectively, these results underscore that breast cancer cells are more sensitive to CMPD1 treatment than normal RPE1 cells, particularly in a clinically relevant context.

We subsequently evaluated the efficacy of CMPD1 in inhibiting anchorage-independent growth, a critical hallmark of tumorigenesis, in MDA-MB-231 cells. CMPD1 demonstrated a potent dose-dependent inhibition, with 500 nM being sufficient to completely suppress colony formation (Fig. 3A-B, Fig. S2A-B). Remarkably, even at a relatively lower concentration (250 nM), CMPD1 significantly reduced colony formation, surpassing the inhibitory effects of 10 µM PTX. These findings suggest that CMPD1 is more effective than PTX in inhibiting the anchorage-independent growth of MDA-MB-231 cells.

We demonstrated that CMPD1 effectively inhibits cancer cell growth in a tissue culture model. Next, we investigated whether CMPD1 could also inhibit tumor growth in vivo. For this purpose, we employed an MDA-MB-231 xenograft mouse model and found that CMPD1 significantly suppressed tumor growth (Fig. 3C-E). Notably, while both PTX and CMPD1 suppressed tumor growth in mice, CMPD1 achieved comparable efficacy at concentrations 10-100 times lower than PTX. Furthermore, two mice treated with PTX died during the experiment, whereas all CMPD1-treated mice survived and remained healthy throughout the treatment regimen. (Fig. S2C). These findings suggest that CMPD1 is not only effective in inhibiting tumor growth in both in vitro and in vivo models but also appears to be more potent and safer than PTX, a standard chemotherapeutic agent for breast cancer treatment.

We demonstrated that at least ≥1 µM CMPD1 induced a robust prometaphase arrest in breast cancer cell lines (Fig. 1D). A previous study suggests that CMPD1 may possess the ability to inhibit tubulin polymerization, as indicated by in vitro tubulin polymerization assay⁵³. To elucidate the mechanism underlying the CMPD1-induced prometaphase arrest, we performed high spatiotemporal resolution live-cell imaging to monitor microtubule dynamics in metaphase CAL-51 cells (see Methods). In these cells, α-tubulin and histone H2B genes were endogenously labeled with mNeonGreen and mScarlet, respectively, using CRISPR-Cas9 technology⁶⁰. Our results revealed that the signal intensity of mitotic spindles dramatically decreased to less than 10% within 8 min following CMPD1 treatment. (Fig. 4A-B and Supplementary Movies 3-4). Concurrently, the formation of multipolar spindle poles was observed (Fig. 4A). The decrease in spindle signal intensity correlated with a gradual disruption of the metaphase plate, likely due to the loss of kinetochore-microtubule attachments (Fig. 4A). Moreover, we observed that CMPD1-treated cells exited G2 phase and underwent nuclear envelope breakdown (NEBD), even though these cells were unable to assemble mitotic spindles (Fig. 4C). This finding further supports that CMPD1 specifically arrests cells in prometaphase rather than G2 phase. To determine whether CMPD1 induces rapid microtubule depolymerization or the degradation of tubulin proteins, we pre-treated the cells with MG132, a proteasome inhibitor, for 1 hour prior to CMPD1 treatment (Fig. 4A-B and Supplementary movie 5). We found that proteasome inhibition did not prevent the CMPD1-mediated reduction in spindle signals (Fig. 4A-B), suggesting that CMPD1 directly induces microtubule depolymerization in breast cancer cells.

To further investigate the mechanism underlying CMPD1-induced microtubule depolymerization, we conducted TIRF (Total Internal Reflection Fluorescence Microscopy) experiments to observe microtubule behaviors in response to CMPD1 treatment. The conventional TIRF assay utilizing PTX-stabilized immobilized microtubules is widely used due to its ease of use. However, this approach lacks the dynamic nature of microtubules as seen in living cells, limiting the study of CMPD1’s effects under more physiological conditions. To overcome this limitation, we utilized GMPCPP-stabilized microtubule seeds combined with rhodamine-labeled tubulin, which allows for the observation of microtubule growth and shrinkage at both ends, thereby better mimicking the dynamic behavior of microtubules in mitotic cells. As expected, microtubules exhibited repeated cycles of growth and shrinkage at both plus and minus ends, with a higher stability at minus ends (Fig. 4D). Upon treatment with CMPD1, there was a marked reduction in the proportion of microtubules with plus-end extensions within just 1 min (Fig. 4D-F), along with an increased frequency of microtubule catastrophes, reduced growth rates, and decreased maximum lengths of plus-end extensions (Fig. 4G-H). Interestingly, CMPD1’s effects on the minus ends of microtubules were notably different. Neither the fraction of microtubules with minus end extensions nor the catastrophe frequency at minus ends were significantly different from control (Fig. 4D-H). However, minor reductions were observed in the average maximum length of minus-end extensions and their growth rates (Fig. 4H). In contrast, vinblastine, a well-characterized microtubule destabilizer, rapidly depolymerized microtubules from both ends (Fig. 4D-E), highlighting that CMPD1 uniquely and preferentially depolymerizes microtubules from the plus ends. Collectively, our cell biology and biochemical data demonstrate that CMPD1 alone can depolymerize microtubules, with a pronounced preference for plus-end depolymerization.

Since proper microtubule dynamics is essential for regulating cell locomotion^61-63, we hypothesized that the improper microtubule dynamics induced by CMPD1 might inhibit cell migration. To investigate this, we evaluated the migratory capacity of CAL-51 cells using a wound healing assay following treatment with CMPD1. Our findings revealed that CMPD1, at concentrations ranging from 100 nM to 10 µM, significantly inhibited wound closure compared to the control (Fig. 5A-B), demonstrating its potential to suppress cancer cell migration and invasion. To further examine this hypothesis, we conducted a transwell-invasion assay using MDA-MB-231 cells. Treatment with CMPD1 at concentrations greater than 100 nM significantly suppressed breast cancer cell invasion, and at 1 µM, CMPD1 completely abolished this cancer invasion (Fig. 5C-D).

To determine whether CMPD1 inhibits cancer cell invasion in vivo, we assessed the frequency of invasion of cancer cells into blood vessels by examining the tissue sections of mice with cancer cell-derived xenografts described in Fig. 3C. CMPD1-treated mice displayed a significantly reduced frequency of cancer cell-infiltrated vessels compared to vehicle-treated mice, suggesting that CMPD1 significantly inhibits metastasis in vivo (Fig. 5E). Together, these results show that CMPD1 suppresses cancer cell migration and invasion in vitro and in vivo.

CMPD1’s distinct ability to inhibit tumor growth may stem from its dual inhibitory effect on both the p38-MK2 signaling pathway and microtubule dynamics. To rigorously test the hypothesis that inhibition of the p38-MK2 pathway could potentiate the efficacy of MTAs, we assessed the combinatorial effects of MK2-IN-3 (hereafter referred to as MK2i), a selective MK2 inhibitor, and clinically relevant concentrations of vinblastine (1 or 5 nM) on mitotic progression in CAL-51 cells. Individual treatment with either 10 µM MK2i or 1 nM vinblastine did not significantly alter the frequency of mitotic slippage or mitotic cell death compared to the control, although both treatments modestly prolonged mitotic duration and increased the frequency of mitotic errors (approximately 2 to 3-fold compared to control) (Fig. 6A). Strikingly, the combination of 10 µM MK2i with 1 nM vinblastine resulted in a profound synergistic effect, extending mitotic duration by approximately 13-fold and markedly increasing the frequency of mitotic errors and cell death (Fig. 6A, Condition 4). A similar synergistic effect was also observed when cells were simultaneously treated with 10 µM MK2i and 5 nM vinblastine, although the 5 nM vinblastine alone induced a more pronounced mitotic arrest compared to the control and 1 nM vinblastine (Fig. 6A, Condition 6). To better understand the impact of CMPD1 treatment on cancer cells at the gene expression level, we conducted RNA-seq analysis. When an FDR < 0.05 was applied using DESeq2 for differential gene expression analysis, 351 genes were found to be up-regulated, and 425 genes were down-regulated 24 hours after treatment with 10 μM CMPD1 (Fig. 6B). Gene ontology (GO) biological process (BP) pathway enrichment analysis revealed that the most significantly enriched pathways in up-regulated genes relate to cell migration, while down-regulated genes are predominantly associated with mitosis and chromosome segregation (Fig. 6C). To explore whether unique pathways were up- or down-regulated specifically in cancer cells, RNA-seq analysis was also performed on RPE1 cells. Comparing differentially expressed genes between MDA-MB-231 cells and RPE1 cells, cell death and apoptosis pathways were significantly enriched in genes uniquely upregulated in MDA-MB-231 cells (Fig. 6D). Genes specifically downregulated in MDA-MB-231 cells were again enriched in pathways related to mitosis and chromosome segregation (Fig. 6D), consistent to the results that cancer cells are more sensitive to CMPD1 treatment (Fig. 2). Collectively, CMPD1 upregulates cell death pathways while selectively downregulates mitotic genes in cancer cells, highlighting its potent cancer cell specificity. The pivotal role of the p38-MK2 signaling pathway in enhancing the efficacy of microtubule destabilizers likely contributes to these observed gene expression changes.