CDC25A protein stability represents a previously unrecognized target of HER2 signaling in human breast cancer: implication for a potential clinical relevance in trastuzumab treatment

Abstract

The CDC25A-CDK2 pathway has been proposed as critical for the oncogenic action of human epidermal growth factor receptor 2 (HER2) in mammary epithelial cells. In particular, transgenic expression of CDC25A cooperates with HER2 in promoting mammary tumors, whereas CDC25A hemizygous loss attenuates the HER2-induced tumorigenesis penetrance. On the basis of this evidence of a synergism between HER2 and the cell cycle regulator CDC25A in a mouse model of mammary tumorigenesis, we investigated the role of CDC25A in human HER2-positive breast cancer and its possible implications in therapeutic response. HER2 status and CDC25A expression were assessed in 313 breast cancer patients and we found statistically significant correlation between HER2 and CDC25A (P = .007). Moreover, an HER2-positive breast cancer subgroup with high levels of CDC25A and very aggressive phenotype was identified (P = .005). Importantly, our in vitro studies on breast cancer cell lines showed that the HER2 inhibitor efficacy on cell growth and viability relied also on CDC25A expression and that such inhibition induces CDC25A down-regulation through phosphatidylinositol 3-kinase/protein kinase B pathway and DNA damage response activation. In line with this observation, we found a statistical significant association between CDC25A overexpression and trastuzumab-combined therapy response rate in two different HER2-positive cohorts of trastuzumab-treated patients in either metastatic or neoadjuvant setting (P = .018 for the metastatic cohort and P = .021 for the neoadjuvant cohort). Our findings highlight a link between HER2 and CDC25A that positively modulates HER2-targeted therapy response, suggesting that, in HER2-positive breast cancer patients, CDC25A overexpression affects trastuzumab sensitivity.

Figure 1

(A) CDC25A (144) sc-97 antibody (Santa Cruz Biotechnology) validation on the breast cancer cell line MCF-7. Immunofluorescence with CDC25A antibody in the MCF-7 breast cancer cell line transfected with either pCMV-AC-CDC25A-GFP or pCMV-AC-GFP: The antibody detected a strong CDC25A expression in the CDC25A-GFP-transfected MCF-7 cells compared to the pCMV-AC-GFP-transfected MCF-7 cells. DAPI counterstaining underlined the nuclear localization of CDC25A (red signal, CDC25A; green signal, GFP; blue signal, DAPI; yellow signal, merge). (B) CDC25A (144) sc-97 antibody IHC validation on FFPE breast cancer tissue samples. CDC25A nuclear localization was detected by IHC on FFPE breast cancer tissue sample (left); CDC25A antibody preabsorption testing using the antigenic peptide (CDC25A 144 P sc-97P; Santa Cruz Biotechnology) showed a complete block in the immunoreactivity (right).

Figure 2

(A) HER2 status by FISH and CDC25A expression by IHC in human breast cancer. Representative example of HER2-amplified tumor (FISH HER2 positive): ratio Spectrum Orange-HER2 signals and Spectrum Green-centromere 17 signals > 2; the same case is an example of CDC25A-overexpressing tumor (IHC CDC25A positive): all infiltrating ductal carcinoma cells show strong staining with CDC25A antibody. Representative example of HER2-nonamplified tumor (FISH HER2 negative): ratio Spectrum Orange-HER2 signals and Spectrum Green-centromere 17 signals < 2; the same case is an example of CDC25A-nonexpressing tumor (IHC CDC25A negative): no expression of CDC25A is seen in the infiltrating ductal carcinoma cells. (B) HER2 gene status and CDC25A expression histogram; 63.5% of HER2-positive breast cancer patients showed CDC25A overexpression and 54.4% of patients with HER2-negative breast cancer were negative for CDC25A overexpression. (C and D) CDC25A expression significantly correlates with prognosis. Kaplan-Meier curves showed that overexpression of CDC25A was associated with the decrease of both OS (P = .045) and DFS (P = .032). (E and F) Significative stratification of mortality risk by combinations of HER2 status and CDC25A expression. The group of patients with the highest mortality was the one with HER2/CDC25A double-positive breast carcinomas in terms of either OS (P = .005) or DFS (P = .002; purple curves).

Figure 3

(A) CDC25A expression and HER2 expression/gene amplification in breast cancer cell lines. Immunofluorescence with CDC25A and cERBB2 (Dako) antibodies revealed CDC25A overexpression (nuclear red signal) and HER2 overexpression (cytoplasmic red signal) in SKBR3 cells and basal levels of CDC25A and HER2 overexpression in BT474 cells. Green signal is the immunofluorescence staining of cytokeratin pair 8 and 18 (CK8.18) highlighting the cell cytoplasm; blue signal is DAPI counterstaining for nuclei. FISH analysis showed HER2 amplification (red signals) in SKBR3 and BT474 metaphase spreads. (B) HER2 inhibition induces CDC25A down-regulation of both protein expression and functional activity. Cell lines were treated with AG825 (45 µM) for 24 hours. Control cells were treated with DMSO. Immunoblot data showed a decrease in CDC25A protein levels in HER2/CDC25A double-positive SKBR3 cells treated with AG825. SKBR3 cells showed also a decreased in CDC25A activity after HER2 abrogation as measured by the increase of the inhibitory phosphorylation of its CDK substrates, pCDK1/2Y15. No such strong effect was seen in BT474. AG825 treatment could inhibit the kinase activity of HER2 as assessed by a decrease in HER2Y1248 phosphorylation with no effect on total HER2 levels in the HER2-positive cell lines, SKBR3 and BT474. (C) HER2 inhibition affects CDC25A protein stability in SKBR3. SKBR3 cells were treated with the protein synthesis inhibitor cycloheximide (5 µg/ml) at different time points (0, 15, 30, 60, 240, and 360 minutes) in the presence or absence of AG825: Immunoblot data showed that AG825 decreased the half-life of CDC25A protein from 60 to less than 15 minutes. β-Tubulin was used as a loading control. (D) Ubiquitin/proteasome pathway is involved in the increased turnover of CDC25A after SKBR3 AG825 treatment. To inhibit proteasome-dependent degradation of proteins, SKBR3 cells, either treated or not with AG825, were incubated with 50 µM proteasome inhibitor MG132 for 6 hours. CDC25A immunoblot analysis showed that inhibition of the proteasome by treatment with MG132 rescued CDC25A down-regulation. β-Tubulin was used as a loading control. (E) HER2 inhibition leads to CDC25A down-regulation through the PI3K/AKT pathway and DDR activation. Cell lines were treated with AG825 for 24 hours. Control cells were treated with DMSO. Immunoblot data showed, in AG825-treated SKBR3 cells, a decrease in AKT activity, as measured by the decrease in its phosphorylation on serine 473 (pAKTS473) and no reduction of total AKT observed; CHK1 and CHK2 activation, through phosphorylation of their key sites, serine 345 and threonine 68, respectively (pCHK1S345, pCHK2T68), with no effect on total CHK1 and CHK2 levels; and histone H2AX gamma phosphorylation (γH2AX). No differences were observed in BT474 cells either with or without AG825 treatment. β-Tubulin was used as a loading control.

Figure 4

(A and B) HER2 down-regulation affects both viability and cell death in the CDC25A-overexpressing cell line. SKBR3 and BT474 cells were treated or not with AG825 (45 µM) for 96 hours. Cell viability changes were measured by MTT assay. Results were expressed as mean percentage of cell viability ± SE of six replicates from a representative experiment that was repeated three independent times. HER2 down-regulation resulted in a lower viability of the CDC25A-overexpressing cell line (SKBR3) compared to the CDC25A basal expressing cells (BT474) (A). Cell death was evaluated by immunoblot for cleaved PARP: PARP cleavage was greater in SKBR3 compared to BT474. β-Tubulin was used as a loading control (B). (C and D) CDC25A inhibition by PM-20 induces breast cancer cell death. SKBR3 and BT474 were treated or not with PM-20 (3 µM) for 96 hours. Cell viability changes were measured by MTT assay as performed for HER2 inhibition studies: A marked reduction in SKBR3 viability was observed (C); PM-20 significantly inhibited CDC25A phosphatase activity, as evinced by increased pCDK1/2Y15 levels, and associated to increased apoptotic response, as deduced by increased PARP cleavage. β-Tubulin was used as a loading control (D).

Figure 5

Statistically significant correlation between CDC25A overexpression and trastuzumab response. (A) Metastatic cohort of patients. (B) Neoadjuvant cohort of patients. Chi-squared test showed a statistically significant correlation between CDC25A overexpression and trastuzumab response in the metastatic cohort (P = .018), where 90% of patients with CDC25A overexpression showed partial or complete response to trastuzumab, while the 66.7% of CDC25A-negative patients had stable or progressive disease (C). Chi-squared test showed a statistically significant correlation between CDC25A overexpression and trastuzumab response in the neoadjuvant cohort (P = .021), where 100% of patients with CDC25A overexpression showed partial or complete response to trastuzumab (D).

Figure 6

HER2 inhibition in a CDC25A overexpression context restores the DDR through a PI3K/AKT-dependent mechanism. In CDC25A-positive breast cancer cells, high levels of CDC25A, pushing the cell throughout the cell cycle transition, may induce an impaired DDR machinery, bypassing the DNA damage-induced cell cycle arrest. In HER2/CDC25A double-positive breast cancer cells, such DDR evasion could be enhanced by HER2 through CDC25A stabilization (A). In this context, HER2 down-regulation leads to AKT inhibition that stops its inhibitory action on CHK1/CHK2; the activation of the checkpoint kinases promotes CDC25A degradation and restores the DDR in checkpoint-proficient cells (B).