c-Jun N-Terminal Kinase Inactivation by Mitogen-Activated Protein Kinase Phosphatase 1 Determines Resistance to Taxanes and Anthracyclines in Breast Cancer

Improving breast cancer therapy requires novel prognostic and predictive markers (1, 2). The prognostic labeling of this heterogeneous disease is still based on conventional tumor–node–metastasis (TNM) staging and histopathologic features (3, 4). While estrogen receptor (ER)–dependent and progesterone receptor (PR)–dependent tumors and HER2-positive subtypes have been approved for noncytotoxic regimens, the triple-negative breast cancer subtype lacks a therapy alternative to chemotherapy (5). Hence, the molecular dynamics that rule breast cancer pathogenesis need to be elucidated to develop precise therapies to enhance survival and overcome resistance to standard chemotherapy regimens (6, 7).

MAPKs have been extensively described as some of the key molecular events driving breast cancer progression (8, 9). Part of cancer-related MAPK regulation consists of a dual dephosphorylation performed by MAPK phosphatases (MKP; ref. 8). Among them, MKP-1 plays a relevant role in tumorigenesis, being able to dephosphorylate all MAPKs, with substrate preference for p38 MAPK and c-Jun N-terminal kinase (JNK). The activation of the JNK pathway has been linked to the apoptosis induced by several chemotherapeutic agents. In breast cancer, JNK dephosphorylation has been correlated with cancer progression and tumor survival against different stress conditions, such as chemotherapy or oxidative damage (10–13). Further experimental research has related MKP-1 with tumor response to stress in breast cancer. Of importance, a tuned MAPKs/MKPs balance regulates cellular response to cancer therapy, as revealed by experimental evidence. For instance, in breast cancer cells, doxorubicin is able to activate JNK pathway to achieve its antitumoral effect (14, 15). On the contrary, MKP-1 mediates different tumor responses to anticancer therapy, depending on its activity or inactivity: MKP-1 activates antiapoptotic pathways in response to proteasome inhibitors (16, 17) as well as antiproliferative activity to PR in breast cancer cells (18). Of relevance, MKP-1 inhibition by small molecules enhanced the antitumoral effect of paclitaxel (19), and transient expression modulation of MKP-1 defined breast cancer cells' ability to survive after exposure to different cytotoxic agents (13). Given the reported apoptotic role of JNK in response to taxanes- and anthracyclines-based therapies, as well as the ability of MKP-1 to decrease JNK activation, we decided to explore the interplay of this kinase/phosphatase pair in the resistance of breast cancer cells to docetaxel and doxorubicin.

Previously, we showed that MKP-1 was overexpressed during the malignant transformation of the breast, thus affecting MAPK expression, and its activation could be inhibited by doxorubicin treatment (15). Nevertheless, we discovered that breast tumors overexpressing MKP-1 did not show this MAPK alteration after this treatment (15). In the present study, we confirm the importance of MKP-1 overexpression as a negative prognostic marker of response to chemotherapy. Of importance, we demonstrate that docetaxel and doxorubicin regulate ERK1/2 and JNK1/2 activation in part through MKP-1 modulation. Further, MKP-1 overexpression dephosphorylates JNK1/2 and results in higher cell growth and lower apoptotic rates in the tumor cells.

MDA-MB-231 (ATCC HTB-26) and BT-474 (ATCC HTB-20) cell lines were purchased from the American Type Culture Collection (ATCC) and authenticated according to the required standards (LGC Standards). Docetaxel, doxorubicin, anisomycin, and SP6000125 were purchased from Sigma Aldrich. Human MKP-1 cDNA was purchased from Open Biosystems and cloned into a pBluescriptR vector (clone ID 4794895; Open Biosystems, Dharmacon), digested with EcoRI and KpnI enzymes, and inserted into a pCMV-HA plasmid. Transfections were carried out using Lipofectamine 2000 (Life Technologies) and following the manufacturer's indications.

Cell proliferation was measured in triplicate by MTS cell proliferation assay using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega), following the manufacturer's indications. Cell growth was analyzed in triplicate by crystal violet assay as previously reported (20). Apoptosis was measured using Annexin V FITC Apoptosis Detection Kit I (BD Biosciences) and quantified in a FACS CANTO II cytometer (BD Biosciences).

Western blotting (WB) analysis in protein extracts from cultured cells was done as previously reported (21). Antibodies were as follows: anti–phospho-ERK1/2 (p-ERK1/2; Thr202/Tyr204), anti-ERK1/2, and anti-JNK (Cell Signaling Technology); anti-active JNK pAb (p-JNK; Thr183/Tyr185; Promega); anti-MKP-1 (Santa Cruz Biotechnology); anti–α-tubulin and anti-GAPDH (Sigma-Aldrich); and anti-Rabbit IgG (GE Healthcare).

Total RNA from the cell lines was isolated using the RNeasy mini Kit (Qiagen). RNA purity and integrity were assessed both by spectrophotometry (NanoDrop ND-2000, NanoDrop Technologies) and electrophoresis (2100 Bioanalyzer, Agilent Technologies); for microarray experiments, the minimal requirements for RNA purity were A260/280>2.0 and A260/230>1.4 and RIN > 9.4. Microarray expression profiles were obtained using the Affymetrix GeneChip Human Exon 1.0 ST Array (Affymetrix Inc.). Following hybridization, the array was stained in the Affymetrix GeneChip Fluidics Station 450 and scanned using a GeneChip Scanner 3000 7G.

Data were processed following the methodology previously described (22). Briefly, after quality control of raw data, background was corrected, quantile-normalized, and summarized to the gene level using the robust multi-chip average. Only those transcripts with an intensity signal of more than 10% of all intensities of the mean of studied groups and then over 50% of variance from total resting variance were considered for further analysis. Linear Models for Microarray (LIMMA) were used for detecting differentially expressed genes between conditions. Correction for multiple comparisons was performed using the FDR, and only genes with an adjusted P value <0.05 were selected as significant. For the purposes of functional analysis, genes were selected to have an unadjusted P value <0.05. Hierarchical cluster analysis was also performed. All data analysis was performed in R with the packages aroma.affymetrix, Biobase, LIMMA, and genefilter. Functional analysis was performed with Ingenuity Pathway Analysis software (Ingenuity Systems). All microarray procedures were performed by the IMIM Microarray Core Facility (SAM).

cDNA was produced using the Universal Transcriptor cDNA synthesis Kit (Roche Diagnostics) according to the manufacturer's recommendations. MKP-1 gene expression levels were determined using a quantitative real-time PCR (qPCR) assay, with ATP5E as a housekeeping gene. Primers were designed using the Lasergene Primer design software (DNASTAR Inc.) and based on the following genomic sequences: MKP-1 (NM_004417.3) and ATP5E (NM_006886.3). qPCRs were performed using the LightCycler480 II system (Roche Applied Science) for 45 cycles with the following set of primers: MKP-1, Fw, 5′-GAGGCCATTGACTTCATAGAC-3′ and Rv 5′-GTAAGCAAGGCAGATGGTG-3′; ATP5E, Fw, 5′-GTAGCTGAGTCCAGCCTGTC-3′ and Rv 5′- GATCTGGGAGTATCGGATG-3′. Specific probes from the Universal Probe Library (Roche Applied Science) were selected. Relative gene expression levels (RQ) were calculated in accordance with the MIQE guidelines (23).

Three-hundred and fifty surgically resected specimens from primary breast tumors were obtained from Parc de Salut Mar Biobank (MARBiobanc), Fundación Jiménez Díaz Biobank, and Valencia Clinic Hospital Biobank. Tumor specimens from formalin-fixed paraffin-embedded (FFPE) blocks were retrospectively selected from consecutive breast cancer patients diagnosed between 1998 and 2000, following these criteria: infiltrating carcinomas, operable, no neoadjuvant therapy, sufficient available tissue, and clinical follow-up. TNM staging was classified using the American Joint Committee on Cancer staging system. Histologic grade was defined according to the Elston-Ellis modification of the Scarff–Bloom–Richardson grading system (24). An independent cohort of 64 patients with locally advanced breast cancer who had been treated with neoadjuvant taxane–based chemotherapy was also included in the study. Pretreatment tumor specimens were histologically evaluated. For all cases, clinical data were collected from patients' medical records by oncologists. The ethical committees and Institutional Review Boards of the participating hospitals approved the project.

Clinical tumor response to primary chemotherapy was evaluated according to the International Union Against Cancer Criteria (25). Clinical complete response (cCR) was defined as the disappearance of all detectable malignant disease within the breast by physical examination. A reduction greater than 50% in the product of the two maximum perpendicular diameters of the tumor was classified as clinical partial response (cPR). Clinical progressive disease (cPD) was considered as an increase of at least 25%. Clinical stable disease (cSD) was defined as situations in which clinical breast cancer response did not meet the criteria for cCR, cPR, or cPD. Postchemotherapy specimens were evaluated for pathologic response. A pathologic complete response (pCR) was defined as no histologic evidence of invasive disease in the tumor specimen (26).

Immunostainings were performed on tissue sections (3 μm) obtained from FFPE tumors as previously described (15). All stainings were performed in a Dako Autostainer. Sections incubated with normal nonimmunized rabbit immunoglobulins were used as negative controls. Sections of breast tumor with known expression of targets were used as positive controls. Antibody sensitivity was calculated in a range of crescent dilutions of primary antibody (MKP-1, 1:50-1:200; p-JNK, 1:10-1:200; JNK, 1:100-1:1,000). Specificity was confirmed in a set of paired fresh frozen and FFPE samples processed by WB and IHC. Antigen preservation in tissues was confirmed by expression of phospho-tyrosines using a monoclonal antibody to tyrosine-phosphorylated proteins (clone 4G10, 1:500; Millipore). High proliferation in breast cancer based on Ki67 labeling by IHC was defined following the 13th St. Gallen International Breast Cancer Conference (2013) criteria based on a proliferation threshold ≥20% (5). Only the membrane of epithelial cells, but not stromal cells, was evaluated for MKP-1, p-JNK, and JNK. Expression blinded to clinical data was evaluated by two pathologists (F. Rojo and S. Zazo). A semiquantitative histoscore (Hscore) was calculated by estimating the percentage of tumor cells positively stained with low, medium, or high staining intensity. The formula used was Hscore = (low %) × 1 + (medium %) × 2 + (high %) × 3, and the results ranged from 0 to 300.

Statistical analyses were performed using the software SPSS 20 (SPSS Inc.). For in vitro studies, we included at least independent triplicates for all the cases. Receiver operating curve (ROC) analysis was used to determine the optimal cutoff point based on progression end point for MKP-1, p-JNK, and JNK expression as previously described (27). Overall survival (OS) was defined as the time from diagnosis to the date of death from any cause or last follow-up. Disease-free survival (DFS) was defined as the time from diagnosis until the first event, in which relapse at any location, death, or end of follow-up were considered events. Survivals were analyzed by the Kaplan–Meier method using the log-rank test. Multivariate analyses were carried out using the Cox proportional hazards model. Analysis of experimental conditions was done by paired t test. All statistical tests were conducted at the two-sided 0.05 level of significance. This work was carried out in accordance with Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) guidelines (28).

Preliminary cell viability assays were performed in MDA-MB-231 and BT-474 breast cancer cell lines to evaluate the effects of docetaxel and doxorubicin on MAPK activation (IC₅₀ of 48.3 nmol/L and 17.3 μmol/L were calculated, respectively, data not shown). Gene expression changes associated with the drug treatments were revealed, after BT-474 cells were exposed to docetaxel for 4 hours, by differentially expressed levels in more than 300 genes as compared with nontreated cells (GEO accession number: #16789213; Fig. 1A). Noticeably, several MKPs (such as MKP-1, MKP-2, and MKP-3) were among the 25 top downregulated genes. Similar results were found when MDA-MB-231 cells were treated with doxorubicin (Fig. 1A). MKP-1 downregulation resulting from chemotherapy treatment was confirmed by qPCR in both cell lines (Fig. 1B): in both cases, the effect of docetaxel was significant only at concentrations double of IC₅₀; doxorubicin, on the other hand, repressed the levels of MKP-1 even at concentrations below its IC₅₀, with a more drastic effect when a concentration double of IC₅₀ was used.

We postulated that this inhibitory effect of docetaxel and doxorubicin in several MKPs was channeled by different MAPKs. Therefore, we treated MDA-MB-231 and BT-474 cells with docetaxel 50 nmol/L and doxorubicin 10 μmol/L for 24 hours to further quantify the phosphorylation status of MAPKs. As expected, WB analyses revealed higher phosphorylation levels of JNK1/2 and ERK1/2 in both breast cancer cell lines (Fig. 1C). In addition, we confirmed the inhibition of MKP-1 after doxorubicin treatment in both cell lines. The apparent lack of modulation of MKP-1 protein expression by docetaxel was discarded when the effect was measured at larger times (i.e., by 48 hours of docetaxel treatment, the decrease of MKP-1 signal was evident, and by 72 hours, there was no detectable signal; Fig. 1D; Supplementary Fig. S1). Collectively, these data suggest that MKP-1 could mediate the responses of MAPKs to chemotherapy in breast cancer cells.

By transfecting both cell lines with a plasmid construction containing the MKP-1 clone cDNA, we observed a sustained increase in MKP-1 transcript levels. These levels were practically unaltered despite docetaxel or doxorubicin treatment (Fig. 2A). At the protein level, however, MKP-1 expression was abolished by doxorubicin in both cell lines (both endogenous and ectopic expression), indicating that the drug triggers a posttranscriptional mechanism on the cells. Docetaxel, on the other hand, did not show any effect on MKP-1 protein levels (Fig. 2B).

The overexpression of MKP-1 provided the breast cancer cell lines with a higher resistance to docetaxel or doxorubicin, as demonstrated by a significantly increased capacity for cell growth (Fig. 2C). The ability of transfected cells to resist exposure to chemotherapy was quantified from crystal violet images as the relative cell area in colorimetric growth assays (Supplementary Fig. S2). Cell viability using MTS assays confirmed higher viability values in transfected cells than in control cells, both in MDA-MB-231 and BT-474 cells after 24 hours of docetaxel (50 nmol/L) or doxorubicin (10 μmol/L), with statistical significance reached in all conditions (Fig. 2D). Finally, MKP-1 overexpression elicited a noticeable strong survival increase in breast cancer cells. Changes in apoptotic ratios increased cell survival by an order of magnitude (Fig. 2E), and although docetaxel and doxorubicin treatments were effective in bringing about a significant reduction in cell viability, they did not reach the levels of parental cells.

Given that alteration in MKP-1 protein-expression level was apparently conditioning the viability and apoptotic behavior of breast cancer lines, and considering JNK1/2 as one of the main proapoptotic regulators of the cell, we evaluated the activation levels of JNK1/2 in MDA-MB-231 and BT-474 cells transiently overexpressing MKP-1. WB analysis showed that MKP-1 overexpression was able to dephosphorylate JNK1/2 in both cell lines (Fig. 3A). To reveal the key role of JNK1/2 in the regulation of MKP-1–mediated response to docetaxel or doxorubicin treatment, we assessed the response of the transfected cells in the presence of well-known JNK1/2 modulators: anisomycin as an activator and SP6000125 as an inhibitor agent. Accordingly, the effect of doxorubicin treatment in MKP-1–overexpressing cells was highly marked in combination with the addition of anisomycin, revealing both an increase in JNK1/2 activation (phosphorylation) and the disappearance of the MKP-1 endogenous protein band (but not the ectopic band). The effect was more evident in MDA-MB-231 than in BT-474 cells. This point is in agreement with the rise of apoptotic rates in the presence of anisomycin. On the other hand, the overexpression of MKP-1 cleared the phosphorylation signaling mediated by JNK1/2 when the cells were exposed to the JNK1/2 inhibitor, and this effect could not be reverted by docetaxel or doxorubicin (Fig. 3A). In general, the effect of the overexpression of MKP-1 exceeded the modulation of MAPKs by activator or inhibitory agents.

Coinciding with protein-expression alterations, MKP-1 overexpression resulted in increased viability of the transfected cells after docetaxel or doxorubicin treatments, despite the presence of a JNK1/2 modulator (Fig. 3B). As expected, the net balance of cell proliferation was lower in anisomycin-pretreated cells, and above the balance for non-pretreated control cells in the SP6000125-treated cells. At the same time, the MKP-1–transfected cells showed higher viability rates than the cells with the empty vector. This double combination of MKP-1 overexpression and JNK1/2 chemical premodulation when cells are treated with either docetaxel or doxorubicin led us to the conclusion that the cellular response to chemotherapeutic treatments is mediated by JNK1/2, and that the response is conditioned by the expression level of MKP-1. Similar conclusions were drawn when apoptosis was measured in cells under different conditions (Fig. 3C): MKP-1 overexpression resulted in partial inhibition of JNK1/2-induced apoptosis after chemotherapeutic treatment. The proapoptotic effect of docetaxel or doxorubicin treatments was significantly repressed by the expression of extra quantities of MKP-1, even in the presence of anisomycin, when the induction of JNK1/2 should be activating the proapoptotic pathway. On the contrary, the addition of SP6000125 prior to the treatment nearly abolished the apoptotic response to docetaxel in both cell lines—a fact not altogether striking for doxorubicin—confirming the key role of JNK/MKP-1 interplay in the cellular response to chemotherapy.

In order to understand the clinical implications of these findings, we investigated the prevalence and clinical significance of MKP-1 overexpression and its relation with JNK1/2 activation. To do this, we quantified MKP-1 and p-JNK1/2 expression in a cohort of 350 tumors obtained from patients with early breast cancer treated with adjuvant anthracycline–based chemotherapy. Patient characteristics are shown in Supplementary Table S1. The IHC analysis of MKP-1 and p-JNK1/2 (Fig. 4A) showed that MKP-1 and p-JNK1/2 were diffusely distributed throughout the tumor, with primary expression located in the nucleus of tumor cells. Faint levels of MKP-1 and moderate signal for p-JNK1/2 were detected in normal breast epithelium and stromal cells (Supplementary Fig. S3).

High MKP-1 levels were detected in 31% of the samples. The elevated expression of MKP-1 was associated with the size of the tumors (P = 0.013) and with relapse (P < 0.001), but was not dependent on the molecular subtype of the tumors. Among tumors with high MKP-1 expression, 80% of the samples presented low levels of p-JNK1/2. It was therefore not a surprise that JNK1/2 inhibition was associated with the same parameters as high MKP-1 expression (tumor size and relapse), the clinical behavior reinforcing our previous understanding about the molecular relationship between MKP-1 and JNK1/2. The association between MKP-1 and p-JNK1/2 expression levels, as well as the molecular and clinical parameters of this series, is included in Table 1. 

Complete data from clinical follow-up were available for all the 350 patients included in the study. Of relevance, MKP-1 overexpression was found in those patients who relapsed (P < 0.001). Moreover, the subgroup of patients with MKP-1 overexpression showed substantially shorter OS (P < 0.001) and DFS (P < 0.001); among these patients, those with p-JNK1/2 inhibition presented the worst survival prognosis (Fig. 4B). Multivariate Cox analysis revealed that the combination of MKP-1(+) and p-JNK1/2(–) determinations provided an independent marker for adverse outcome associated with OS [HR, 26.1; 95% confidence interval (CI), 10.1–67.4; P < 0.001; Table 2] and DFS (HR, 33.4; 95% CI, 14.8–75.4; P < 0.001; Supplementary Table S2) in early breast cancer.

In order to provide clinical evidence to prove that MKP-1 overexpression determines docetaxel resistance, we analyzed MKP-1 and p-JNK1/2 expression in an independent set of 64 patients with locally advanced breast cancer who received neoadjuvant taxane–based chemotherapy. Patient characteristics are shown in Supplementary Table S3. The mean time from diagnosis to the beginning of chemotherapy was 21.3 days (range, 1–48 days). During this period, patients underwent standard clinical and radiological tumor staging. Patients received a median of four cycles of chemotherapy (range, 2–6 cycles). After recovering from the effects of the chemotherapy, the patients underwent surgery. The mean time between the last dose of chemotherapy and acquisition of the postchemotherapy specimen from surgery was 30.3 days (range, 8–59 days). Almost 30% of patients achieved a complete pathologic response (as analyzed in the surgical specimen) according to the histopathologic evaluation. Interestingly, we observed that MKP-1 and p-JNK1/2 expression correlated with pathologic response (P = 0.008; Supplementary Table S4).

MKP-1 has long been reported to act as an oncoprotein in breast cancer progression, also inducing antitumor response to several chemotherapeutic drugs (13, 15). Its capability to dephosphorylate p38, JNK, and ERK1/2—specifically in this order of affinity (29)—has been proven to be context- and stimulus-dependent. Conversely, it is well known that, under specific circumstances, MAPKs have the ability to control MKP-1 expression through complex regulatory loops (30, 31), although the molecular mechanisms regulating these routes merit further clarification.

In a previous work, we revealed that MKP-1 is overexpressed during the malignant transformation of the breast and independently predicts poor prognosis. Furthermore, we demonstrated that MKP-1 is repressed by doxorubicin in many human breast cancers (15). In the present study, we demonstrate that MKP-1 overexpression can be a crucial event in breast cancer, as breast cancer cells overexpressing MKP-1 increase their proliferation rate and acquire the capability to inhibit apoptosis activation, even following doxorubicin or docetaxel treatments. In addition, we prove that JNK1/2 dephosphorylation is a leading molecular event that correlates with MKP-1 expression to promote tumor-cell survival after chemotherapy treatment. MDA-MB-231 and BT-474 breast cancer cell lines overexpressing MKP-1 withstand doxorubicin and docetaxel treatments (e.g., enhanced cell proliferation, improved cell viability, reduced apoptotic rates) by dephosphorylating JNK1/2. On the contrary, enforced JNK1/2 activation (phosphorylation) by anisomycin increases proapoptotic signals after drug addition to almost parental levels, even in the presence of high MKP-1 protein abundance.

From a clinical perspective, we report here that MKP-1 overexpression is a crucial event in almost a third of breast cancer patients; further, we define JNK1/2 dephosphorylation as a widespread molecular event in this subset (80% of MKP-1–overexpressing tumors). Moreover, in our series, patients with MKP-1 overexpression showed significantly worse outcome, and multivariate analysis suggests that MKP-1 overexpression has an independent prognostic value for OS and DFS in adjuvant anthracycline–based chemotherapy. In particular, the dephosphorylation of JNK1/2 appears to be a significant molecular mechanism correlated to MKP-1–induced resistance to the cytotoxic treatment in those tumors. Furthermore, we proved, in an independent cohort of samples with locally advanced breast cancer, that patients with MKP-1 overexpression and JNK1/2 inhibition are significantly more resistant to neoadjuvant taxane–based chemotherapy. The IHC determinations of MKP-1 and p-JNK1/2 were associated with pathologic response in patients treated in neoadjuvancy, suggesting a key role of the MKP-1/p-JNK1/2 interplay in the acquisition of malignant traits. Consequently, these results suggest that the combination of MKP-1 overexpression and JNK1/2 inhibition is a common and relevant molecular event with high clinical importance in breast cancer, as it defines a subset of tumors with predicted lack of clinical success for some of the usual chemotherapeutic regimens. Finally, in our cohort of patients, MKP-1 overexpression did not correlate with biological subtype (Table 1), which would imply that the consequences of MKP-1 overexpression may be as heterogeneous as the breast cancer heterogeneity itself.

Although the link between MKP-1 and chemoresistance has been previously reported in an in vitro cellular model in breast cancer (13), no association with clinical findings had been described to date. Our results confirmed the ability of MKP-1 overexpression to inhibit JNK1/2 activation in tumor cell lines. Furthermore, we describe the clinical significance this molecular event may have in a particular subset of breast tumors, where the overexpression of MKP-1 would hinder the effect of adjuvant chemotherapy and enable tumor relapse.

In our subset of early breast cancer samples, most of the MKP-1–overexpressing tumors showed low levels of p-JNK1/2. However, approximately 20% of these samples presented p-JNK1/2 overexpression. Though it is widely accepted that elevated MKP-1 expression is linked to dephosphorylation of JNK1/2, there are reports in the literature elucidating the contribution of JNK1/2 phosphorylation to tumor progression in specific contexts: For instance, JNK pathway activation was linked with the upregulation of Ras in certain human tumors, including HER2-positive breast cancer (32); casein kinase 1 epsilon–mutant breast cancer has been reported to lead to the activation of the Wnt/Rac1/JNK/AP1 pathway instead of canonical Wnt/β-catenin, thus mediating higher invasion ability and aggressiveness of breast cancer cells (33); IL33 provoked epithelial cell transformation and breast tumorigenesis by activation of MEK-ERK, JNK-cJun, and STAT3 through the IL33/ST2/COT cascade (34). In summary, the different isoforms of JNK can exert anti- and protumor modulations in different cell types and stages of cancer, including breast tumors, as has been reviewed elsewhere (35).

Despite several molecular issues yet to be clarified regarding the molecular interplay of MKP-1 in breast cancer, it is clear that the overexpression of MKP-1 is a key molecular event that should be considered as a potential predictive biomarker in breast cancer. This alteration accompanying JNK1/2 dephosphorylation needs further study in order to understand the breast cancer prosurvival mechanisms associated to them. Our clinical results revealed a high prevalence of this breast cancer subtype (MKP-1 overexpression/JNK1/2 dephosphorylation) as nearly 1 of 4 breast cancer patients with this molecular profile would not benefit from conventional adjuvant chemotherapy treatment. What is more, most of these patients may suffer relapse after treatment. Therefore, the incorporation of MKP-1 determination to the screening panel of prognostic and predictive biomarkers in breast cancer would facilitate the management of such patients, aiding in the decision on which therapeutic regimens can best improve their survival.

No potential conflicts of interest were disclosed.

Conception and design: R. Rincón, R. Perona, J. Albanell, J. Madoz-Gúrpide, F. Rojo

Development of methodology: R. Rincón, S. Zazo, C. Chamizo, R. Manso, P. González-Alonso, I. Cristóbal, J. Albanell, A. Rovira, J. Madoz-Gúrpide, F. Rojo

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Rincón, S. Zazo, R. Manso, P. González-Alonso, C. Cañadas, A. Lluch, P. Eroles, J. García-Foncillas, F. Rojo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Rincón, S. Zazo, R. Manso, P. González-Alonso, P. Eroles, J. García-Foncillas, J. Madoz-Gúrpide, F. Rojo

Writing, review, and/or revision of the manuscript: R. Rincón, I. Cristóbal, P. Eroles, J. García-Foncillas, J. Albanell, A. Rovira, J. Madoz-Gúrpide, F. Rojo

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Rincón, E. Martín-Aparicio, J. Albanell, J. Madoz-Gúrpide, F. Rojo

Study supervision: I. Cristóbal, A. Lluch, J. Madoz-Gúrpide, F. Rojo

We thank Oliver Shaw for linguistic correction of the article.

The current work was supported by grants from the Spanish Ministry of Economy and Competitiveness (MINECO; AES Program, grants PI12/01552; PI12/00680; PI12/01421); the Ministry of Health (Cancer Network); the Community of Madrid (S2010/BMD-2344 grant); and the Government of Catalonia (2014/SGR/740 grant). The biobanks are funded by grants from the MINECO (Institute of Health Carlos III, RETICS Biobanks Network, with FEDER funds: Fundación Jiménez Díaz Biobank, RD12/0036/0021; Parc de Salut Mar Biobank, RD12/0036/0051; Valencia Clinic Hospital Biobank, RD12/0036/0070). S. Zazo and C. Chamizo are supported by grants from the Biobanks initiative. J. Albanell and F. Rojo are recipients of an intensification program ISCIII/FEDER. R. Manso and P. González-Alonso are supported by Fundación Conchita Rábago de Jiménez Díaz grants.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.