Low-Molecular Weight Cyclin E Confers a Vulnerability to PKMYT1 Inhibition in Triple-Negative Breast Cancer

Abstract

Cyclin E is a regulatory subunit of CDK2 that mediates S phase entry and progression. The cleavage of full-length cyclin E (FL-cycE) to low-molecular weight isoforms (LMW-E) dramatically alters substrate specificity, promoting G1-S cell cycle transition and accelerating mitotic exit. Approximately 70% of triple-negative breast cancers (TNBC) express LMW-E, which correlates with poor prognosis. PKMYT1 also plays an important role in mitosis by inhibiting CDK1 to block premature mitotic entry, suggesting it could be a therapeutic target in TNBC expressing LMW-E. In this study, analysis of tumor samples of patients with TNBC revealed that coexpression of LMW-E and PKMYT1-catalyzed CDK1 phosphorylation predicted poor response to neoadjuvant chemotherapy. Compared with FL-cycE, LMW-E specifically upregulates PKMYT1 expression and protein stability, thereby increasing CDK1 phosphorylation. Inhibiting PKMYT1 with the selective inhibitor RP-6306 (lunresertib) elicited LMW-E-dependent antitumor effects, accelerating premature mitotic entry, inhibiting replication fork restart, and enhancing DNA damage, chromosomal breakage, apoptosis, and replication stress. Importantly, TNBC cell line xenografts expressing LMW-E showed greater sensitivity to RP-6306 than tumors with empty vector or FL-cycE. Furthermore, RP-6306 exerted tumor suppressive effects in LMW-E transgenic murine mammary tumors and patient-derived xenografts of LMW-E-high TNBC but not in the LMW-E null models examined in parallel. Lastly, transcriptomic and immune profiling demonstrated that RP-6306 treatment induced interferon responses and T-cell infiltration in the LMW-E-high tumor microenvironment, enhancing the antitumor immune response. These findings highlight the LMW-E/PKMYT1/CDK1 regulatory axis as a promising therapeutic target in TNBC, providing the rationale for further clinical development of PKMYT1 inhibitors in this aggressive breast cancer subtype. Significance: PKMYT1 upregulation and CDK1 phosphorylation in triple-negative breast cancer expressing low-molecular weight cyclin E leads to suboptimal responses to chemotherapy but sensitizes tumors to PKMYT1 inhibitors, proposing a personalized treatment strategy.

Fig. 1 |. Co-expression of LMW-E and CDK1 pT14 in TNBC patient samples prior to neoadjuvant chemotherapy is predictive of poor clinical response.

a Representative immunohistochemical images showing the status of LMW-E (top panels) and CDK1 pT14 (bottom panels). Scale bars, 500 μm. b The number of patients grouped by the status of IHC staining for antibodies against CDK1 pT14 and cyclin E (LMW-E). c and d The number of patients with complete pathological response and with partial response or stable disease grouped by CDK1 pT14 (c, n=36) and LMW-E (d, n=40) status. e The number of TNBC patients with complete pathological response and with partial response or stable disease grouped by the combination of CDK1 pT14 and LMW-E status. The differences in the patient distributions in b-d were analyzed using the Fisher exact test, and those in e were analyzed using a chi-square test.

Fig. 2 |. LMW-E upregulates PKMYT1 in hMECs.

a Schematic of the cell model system to identify transcriptomic signatures following LMW-E or FL-cycE induction. The CCNE1 gene was knocked out in 76NE6 hMECs. Inducible LMW-E or FL-cycE driven by a Tet-On/rtTA system were introduced into the cells. Stable clones were treated with 100 ng/mL Dox (or DMSO [control]) for 36 h to induce the expression of LMW-E or FL-cycE. Gene expression patterns altered by LMW-E or FL-cycE were identified via RNA sequencing. b Results of GSEA of reactome pathways showing the significantly enriched pathways in hMECs with and without induced expression of LMW-E. c The 10 genes most significantly upregulated or downregulated by LMW-E and FL-cycE induction (adjusted P < 0.05) ranked by their fold changes. d hMECs were treated with the indicated concentrations of Dox for 24 h to induce LMW-E expression. A western blot of the levels of LMW-E, PKMYT1, CDK1 pT14, and total CDK1 is shown. e Quantitation of the relative PKMYT1 protein levels in panel d. The western blot densitometry of PKMYT1 under each condition was normalized to that for cells treated with DMSO (control; Dox = 0 ng/mL) and the loading control vinculin (n = 3 biological repeats; mean [± SD]). f. Side-by-side comparison of FL-cycE and LMW-E inducible hMECs for the expression of PKMYT1, CDK1, CDK2, RB, and phosphorylation of CDK1 T14, CDK2 T160 and RB S807/811 g and h Relative intensity of PKMYT1(panel g) and CDK2(panel h) in S-protein and streptavidin bead pull-down precipitates collected from HEK293T cells expressing an SFP-tagged empty vector, SFP-tagged FL-cycE, or SFB-tagged LMWE identified using mass spectrometry. i Western blot–based validation of PKMYT1 in S-bead pull-down precipitates collected from HEK293T cells with and without SFB-tagged LMWE overexpression. j Analysis of the binding of LMW-E and PKMYT1 in CCNE1 knock-out hMECs with inducible LMW-E. These cells were treated with 100 ng/mL Dox for 24 h to induce the expression of LMW-E, followed by co-immunoprecipitation (co-IP) and western blot analysis using the indicated antibodies. Uninduced cells (DMSO-treated) and co-IP samples using IgG were used as negative controls. k Inducible FL-cycE or LMW-E hMECs with or without Dox treatment were analyzed for PKMYT1 expression by immunofluorescence (IF) and for the expression of FL-cycE or LMW-E via fused GFP-tag. l TNBC cell line MDA-MB-157 cells were subjected to co-immunoprecipitation (co-IP)/WB analysis for the interaction between PKMYT1 and cyclin E. Western blot signal for cyclin E at the band size of LMW-E was detected in the co-IP precipitates using PKMYT1 as bait but not IgG control. m Western blot analysis for cyclin E and PKMYT1 using total cell lysates, cytoplasmic and nuclear fraction lysates of MDA-MB-157 cells. The results show LMW-E and PKMYT1 are predominantly located in cytoplasm while FL-cycE is mostly a nuclear protein. n-q Analysis of PKMYT1 protein stability via CHX chase assay. The inducible hMECs were subjected to western blot analysis for PKMYT1 to compare the effect of LMW-E over-expression with uninduced control (panel n) and FL-cycE (panel p). The cells were treated with 100ng/mL Dox or DMSO for 24 h, followed by 50 mM CHX treatment for the indicated times. The western blot densitometry of PKMYT1 under each condition is shown in panels o and q, which were normalized to that of the untreated control (CHX = 0 h) and loading control vinculin (n > 3 biological replicates; mean [± SD]). For all statistical analyses: **P < 0.01, ***P < 0.001, ****P < 0.0001 (Student t-test).

Fig. 3 |. Treatment with RP-6306 attenuates cell survival and induces DNA damage, chromosome breaks, and premature mitotic entry in an LMW-E–dependent manner.

a Western blot analysis of cyclin E (FL-cycE and LMW-E) expression in a panel of seven TNBC cell lines. b IC₅₀s of the PKMYT1 inhibitor RP-6306 in the TNBC cells in a. c Pearson correlation analysis of RP-6306 IC₅₀s and the FL-cycE and LMW-E protein levels shown in a and b. d Percentages of annexin V–positive MDA-MB-231 (endogenous LMW-E–low) and MDA-MB-157 (endogenous LMW-E–high) TNBC cells treated with RP-6306. Representative graphs of FACS are shown in Supplementary Fig. 1d. e Cell cycle distributions for MDA-MB-231 and MDA-MB-157 cells treated with RP-6306. Representative FACS histograms are shown in Supplementary Fig. 1e. f Representative images of chromosomal breakages (black arrows) in MDA-MB-231 and MDA-MB-157 cells treated with or without RP-6306. g Quantitation of chromosomal breakage ratios in cells shown in panel f. h and i DNA damage in MDA-MB-231 and MDA-MB-157 cells treated with RP-6306 assessed by IF analysis of 53BP1 and γ-H2AX and graphed according to foci/cell. Representative images of 53BP1 and γ-H2AX nuclear foci are shown in Supplementary Fig. 1f. j Western blot analysis of the indicated antibodies against DNA damage and apoptosis markers in MDA-MB-231 and MDA-MB-157 cells treated with RP-6306. k and l IF microscopic analysis of premature mitotic entry in cells treated with increasing concentrations of RP-6306. Representative images for EdU+ (green) and cyclin B pS126 (red) in cell nucleus (DAPI) are shown in k, whereas quantitation of the percentage of EdU+ cells with nuclear cyclin B pS126 is depicted in l. All results were representative of three independent biological replicates. All results are representative of three independent biological replicates. Statistical analyses were performed using one- or two-way analysis of variance (ANOVA). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n.s., not significant.

Fig. 4 |. Treatment of LMW-E–high cells with RP-6306 leads to induction of DNA damage and inhibition of replication fork restart.

a and b MDA-MB-231 cells with or without stable LMW-E overexpression were generated, treated with 600 nM RP-6306 for 24 h, and subjected to western blot analysis (a) followed by a cell viability assay with and without RP-6306 treatment (600 nM for 96 h; b). c and d MDA-MB-157 cells transfected with control siRNA or siRNAs targeting CCNE1 were analyzed using western blotting before and after RP-6306–based treatment (20 nM for 24 h; c) and a MTT cell viability assay with and without the treatment (20 nM for 96 h; d). e and f Analysis of DNA damage intensity by γ-H2AX IF in MDA-MB-231 and MDA-MB-157 cells (scale bars, 10 μm). Parental cells, MDA-MB-231 cells with LMW-E overexpression, and MDA-MB-157 cells with CCNE1 knockdown were treated with different doses of RP-6306 (0, 20, or 600 nM) for 48 h. Representative IF images of γ-H2AX (green) and DAPI (blue) are shown in e, and quantification of γ-H2AX–positive cells is shown in f (cell number >400). g-i HCC1806 cells with or without stable CCNE1 knockdown were generated, and their levels of LMW-E, CDK1, and CDK1 pT14 were analyzed via western blotting before and after treatment with RP-6306 (600 nM for 24 h, panel g); IC₅₀s of RP-6306 are shown in panel h; the effect of RP-6306 (0, 100, or 500 nM for 48 h) on DNA damage marker γ-H2AX are shown in panel i (cell number >400, representative images are shown in Supplementary Fig. 3d). j and k. DNA fiber assays to examine replication fork restart in shCtrl and shCCNE1 HCC1806 cells. LMW-E–knockdown and control cells were treated with 1 μM RP-6306 for 24 h. This was followed by sequential treatment with IdU (20 min) to label the progressing forks, hydroxyurea (HU; 1 h) to induce fork stalling, and CldU (20 min) to label when fork progression has restarted. Representative DNA fibers (panel j) and the calculated CldU/IdU ratio (panel k) are shown. l-n T47D cells with and without stable overexpression of FL-cycE or LMW-E were analyzed for RP-6306 sensitivity. Overexpression of FL-cycE and LMW-E was confirmed via western blotting (panel l); the IC₅₀s of RP-6306 are shown in panel m; analysis of DNA damage intensity by IF staining for γ-H2AX are shown in panel n (0, 0.5 or 1μM RP-6306 treatment for 48 h, cell number >400, representative images are shown in Supplementary Fig. 3e) o and p DNA fiber assays were used to examine replication fork restart in inducible T47D cells. Parental, FL-cycE–overexpressing, and LMW-E–overexpressing cells were treated with 1 μM RP-6306 for 24 h. This was followed by the treatment strategy shown in j. Representative DNA fibers are shown in o, and the calculated CldU/IdU ratio is shown in p. For all statistical analyses: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n.s., not significant.

Fig. 5 |. RP-6306 inhibits tumor growth and prolongs survival in LMW-E–high PDX models from TNBC patients.

a and b Tumor growth curves for female nude mice bearing the BCX070 (LMW-E–high) or XC5172013 (LMW-E–low) PDX and treated with RP-6306 (20 mg/kg orally b.i.d.) or vehicle for 24–30 days. Before treatment, tumors were allowed to grow until they reached a volume of 100 mm³. Mice were then given the treatment, and their tumors were collected when the tumor volume reached 1200 mm³. The length and width of the tumor xenografts were measured using calipers two to three times a week, and the tumor volumes were calculated using the equation (length X width²)/2. All tumors were collected at the same time and processed for biomarker analysis. The differences between groups were evaluated using two-way ANOVA or the Tukey multiple comparisons test. ****P ≤ 0.0001. n.s., not significant. The data represent the mean (± SEM) tumor volume. BCX-070: n = 8 mice for vehicle, n = 11 mice for RP-6306; XC5172013: n = 6 mice for vehicle, n = 7 mice for RP-6306 mice. c and d Survival analysis of female nude mice bearing the BCX070 or XC5172013 PDX and given treatment as described in a and b. Survival was calculated based on the time of treatment when the tumor volume reached 1000 mm³ for each arm of the study. The difference in survival curves was calculated using the log-rank (Mantel-Cox) test. e-g Immunohistochemical analysis of CDK1 pT14, γ-H2AX, and Ki67 in BCX070 and XC5172013 PDX tumor tissues from the experiment shown in a and b at the endpoint. Representative images (magnification, 20X; scale bars, 50 μm) are shown for each antibody. Quantification of all three markers was performed using QuPath software (and independently by a pathologist) by assessing multiple areas in each tumor for each treatment to cover the entire tumor. Data are presented as mean (± SEM) values. The difference between groups was evaluated using two-way ANOVA or the Tukey multiple comparisons test.

Fig. 6 |. Selective response to RP-6306 in LMW-E–driven transgenic murine mammary tumors.

a Study schema showing the treatment strategy for transgenic models of MMTV-LMW-E-driven tumors. Tumor fragments isolated from MMTV-LMW-E transgenic models were surgically implanted into the T4 mammary fat pad in 12-week-old FVB mice. When tumors reached 30–50 mm³, mice were randomized into four treatment arms: vehicle, 10 mg/kg RP-6306 (low dose), 20 mg/kg RP-6306 (high dose), and a survival arm (20 mg/kg RP-6306). Mice in the low-dose and high-dose groups were given treatments b.i.d., and their tumors were collected when tumor volumes in the vehicle-treated mice reached 1000 mm³. Mice in the survival arm were given continuous treatment b.i.d. until their tumors (in both vehicle and RP-6306 treatment arms) reached 1000 mm³. b Tumor growth curves for the low-dose and high-dose groups were compared with the vehicle group. The length and width of the tumor xenografts were measured using calipers two to three times a week, and the tumor volume was calculated using the equation (length x width²)/2. All tumors were collected at the same time and processed for biomarker analysis. The differences between groups were evaluated using two-way ANOVA or the Tukey multiple comparisons test. ****, P ≤ 0.0001; n.s., not significant. Data represent the mean (± SEM) tumor volume. n = 9 mice for vehicle, n = 10 mice for 10 mg/kg RP-6306, n = 14 mice for 20 mg/kg RP-6306. c Survival analysis of female FVB mice bearing MMTV-LMWE tumors and given treatment as described in a. Survival was calculated based on the time of treatment when the tumor volume reached 1000 mm³ in each arm of the study. The difference in survival curves was calculated using the log-rank (Mantel-Cox) test. d and e Immunohistochemical analysis of cyclin E, CDK1 pT14, γ-H2AX, and p-HH3 in MMTV-LMW-E tumor tissues from the experiment in b and c at endpoint. Representative images (magnification, 20X except for p-HH3 [40X]; scale bars, 50 μm) are shown for each antibody. Quantification of the markers was performed using QuPath software (and independently by a pathologist) by assessing multiple areas in each tumor for each treatment to cover the entire tumor. Data are presented as mean (± SEM) values. The difference between groups was evaluated using two-way ANOVA or the Tukey multiple comparisons test. ****, P ≤ 0.0001; n.s., not significant.

Fig. 7 |. Induction of T-cell infiltration following RP-6306 treatment of LMW-E–high tumors.

a and b Using the study schema shown in Fig. 6a, mice were randomized into four groups: vehicle, 20 mg/kg RP-6306 administered for 7 days, 20 mg/kg RP-6306 administered for 19 days, and a survival arm. Mice were treatment b.i.d., and their tumors were collected on days 7 and 19. Mice in the survival arm were treated daily for up to 42 days or until all tumors reached 1000 mm³. Tumors collected from the 7 and 19 days treatment and survival groups were subjected to RNA sequencing. DEGs from the vehicle and treatment (7 days, 19 days, and survival [LT] arm) groups were subjected to GSEA using (a) Hallmark and (b) KEGG data sets. Corresponding genes in the Hallmark and KEGG pathways are represented in the heat maps comparing day 7, 19 and survival (LT) arms (all three were compared with the vehicle group). c Bubble plot summarizing the common and unique Hallmark and KEGG pathways in the cell cycle, DNA damage, DNA replication, and interferon α and γ response pathways. d Mice were randomized into four arms—vehicle and 20 mg/kg RP-6306 administered b.i.d. for 7 or 14 days—and subjected to immune profiling (3–5 mice per treatment group) with a focused panel of antibodies specific to T-cell subsets of immune cells. The percentages of CD3, CD4, and CD8 cells (normalized to the percentage of CD45 cells) are shown. The percentage of CD45 cells was normalized to the live cell population in each tissue. e-g Multiplex immunofluorescence was performed using samples obtained from mice treated with RP-6306 for 19 days. The fractions of CD4, CD8, and γ-H2AX cells per square millimeter of tissue quantitated are shown in e and quantitated in g. Pan-cytokeratin staining, representing the number of tumor cells, was comparable in the vehicle (top panels) and RP-6306 (bottom panels) treatment groups (f). Co-localization algorithms were applied to the samples used in f (g). For all statistical analyses: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n.s., not significant.