The tumor microenvironment supplies proinflammatory cytokines favoring a permissive milieu for cancer cell growth and invasive behavior. Here we show how breast cancer progression is facilitated by IL4 secreted by adipose tissue and estrogen receptor–positive and triple-negative breast cancer cell types. Blocking autocrine and paracrine IL4 signaling with the IL4Rα antagonist IL4DM compromised breast cancer cell proliferation, invasion, and tumor growth by downregulating MAPK pathway activity. IL4DM reduced numbers of CD44+/CD24 cancer stem-like cells and elevated expression of the dual specificity phosphatase DUSP4 by inhibiting NF-κB. Enforced expression of DUSP4 drove conversion of metastatic cells to nonmetastatic cells. Mechanistically, RNAi-mediated attenuation of DUSP4 activated the ERK and p38 MAPK pathways, increased stem-like properties, and spawned metastatic capacity. Targeting IL4 signaling sensitized breast cancer cells to anticancer therapy and strengthened immune responses by enhancing the number of IFNγ-positive CTLs. Our results showed the role of IL4 in promoting breast cancer aggressiveness and how its targeting may improve the efficacy of current therapies. Cancer Res; 77(12); 3268–79. ©2017 AACR.

Despite the advent of efficacious treatment of primary breast cancer lesions, metastatic disease is poorly sensitive to the common therapeutic regimens (1).

Adipose tissue is the most abundant constituent of the breast cancer microenvironment and is mainly composed of mature adipocytes, preadipocytes, and adipose-derived stem cells (ADSC; ref. 2). Obesity and overweight have recently been suggested as being meaningful risk factors for the development of breast cancer and its therapy resistance (3). Indeed, compelling evidence shows that paracrine signals, provided by the adipose tissue and surrounding cancer cells, contribute to tumorigenesis and cancer progression (2, 4). Secreted protumorigenic cytokines and hormones feed breast cancer cells regardless of their hormone status and are also responsible for the acquisition of an aggressive cell phenotype (5).

Furthermore, we have identified autocrine and paracrine production of IL4 as a survival signal and tool to protect colorectal cancer cells from anticancer therapy, through the upregulation of antiapoptotic molecules (6, 7). Indeed, IL4 is a pleiotropic cytokine secreted by fibroblasts, immune, adipose, and a wide range of epithelial cells, including breast cancer cells (8). IL4 cognate receptors comprise two types: (i) type I IL4R (IL4RI) and (ii) type II IL4R (IL4RII). The first is mainly present on immune cells and characterized by the heterodimerization of the IL4Rα and the common γ-chain subunits. IL4RII, on the other hand, is present on nonhematopoietic cells and composed of IL4Rα and the IL13Rα1 subunits. IL4RII is expressed on the surface of many cancer cells and lacks intrinsic kinase activity, thus it requires further associated kinases for the initiation of signal transduction (8). Upon binding of IL4, the tyrosine kinases Jak1/2 and Tyk2 are indeed recruited on the transmembrane domain of IL4RII and mediate its phosphorylation, leading to the activation of PI3K/AKT, MAPK, and Jak/STAT6 downstream pathways (8).

In breast cancer, the secretion of the protumorigenic cytokines, IL6 and IL8, is controlled by the dual specificity phosphatase 4 (DUSP4; ref. 9). In physiologic conditions, the transcription of DUSP4 is MEK-dependent and its expression in turn suppresses ERK, along with p38, JNK1, NF-κB, and Rb (9–12), ensuring a proper negative feedback control of cellular proliferative stimuli. DUSP4 is differentially expressed among luminal and basal breast cancers. Specifically, the most aggressive tumors hold DUSP4 under-expression, due to methylation events or genomic loss (9, 10).

It has been clearly demonstrated that a subpopulation of cancer cells, named tumor-initiating cells (TIC), is endowed with the capability of self-renewal and tumor initiation (13–16). Indeed, this cell compartment is refractory to the common anticancer drugs and responsible for recurrence (13–16).

An inflammatory microenvironment has been shown to favor the maintenance of the breast TICs and their invasive behavior (17, 18). However, insufficient data are available on the mechanisms regulating this phenomenon. Recently, it has been demonstrated that, in a syngeneic breast cancer mouse model, the IL4/IL4R interaction promotes metastatic spreading by activating the MAPK pathway (19).

Taken together, these findings suggest that IL4 may conceivably play a role in the progression of breast cancer and resistance to standard therapeutic regimens driven by TICs. In this study, we establish for the first time the molecular mechanisms elicited by IL4, which augment proliferation and invasiveness of breast cancer sphere cells (BCSC), notoriously largely composed of TICs (15). Blocking autocrine and paracrine IL4 signaling via the attenuation of the MAPK pathway, counteracts the protumorigenic effect of all the proinflammatory cytokines released by ADSCs. Moreover, we unveil that IL4 acts through NF-κB to lower DUSP4 expression levels.

Here, we emphasize the role of IL4 in the metastatic potential of BCSCs and its neutralization as a useful strategy for therapeutic intervention.

Tissue collection

Breast cancer tissues were collected at the Department of Surgical, Oncological and Stomatological Sciences (University of Palermo, Palermo, Italy), in accordance with the ethical standards of the Institutional Committee on Human Experimentation and their molecular subtypes. Breast cancer characterization was established using the following IHC markers (20): ER+, PR+/−, HER2, Ki67low (luminal A); ER+, PR+/−, HER2, Ki67high or ER+, PR+/−, HER2+ (luminal B); ER, PR, HER2+ (HER2 amplified); ER, PR, HER2 (triple negative breast cancer, TNBC; Table 1). Cancer staging was determined according to the 7th Edition AJCC classification of malignant tumors.

Isolation and culture of cancer cells

Breast cancer tissue was mechanically and enzymatically digested using collagenase (1.5 mg/mL; Gibco) and hyaluronidase (20 mg/mL; Sigma Aldrich) in DMEM and shaken for 1 hour at 37°C. The recovered cell suspension was plated in ultra-low attachment flasks in serum-free medium (SFM) supplemented with basic fibroblast growth factor (bFGF; 10 ng/mL; PeproTech) and EGF (20 ng/mL; PeproTech), as previously described (21, 22). SFM allows breast cancer cells to propagate as sphere structures, which are enriched in cells that harbor stem-like and tumor-initiating properties (6, 13, 15). We used ER+-BCSCs derived from luminal B patients #4 and #18 and TN-BCSCs, derived from TNBC patients #10 and #30 (Table 1) for all the experiments performed. BCSCs and their relative tumor tissues were authenticated using a highly informative short tandem repeat (STR) system (GlobalFiler STR Kit; Applied Biosystems) and then sequenced using the ABIPRISM 3130 genetic analyzer (Applied Biosystems). CD90 staining was used to exclude the presence of stromal cells within the ER+- and TN-BCSC compartment. Human ADSCs were purchased from STEMPRO and cultured for less than six passages according to manufacturer's instructions. MCF7 cell line was purchased from CLS cell line service in July 2013, authenticated from the cell bank by DNA profiling (STR analysis), and maintained in 10% FBS DMEM. Cells, expanded for two passages after receipt, were then frozen and used within 6 months after thawing. SUM159 cell line was kindly provided by Prof. Max Wicha (University of Michigan, Ann Arbor, MI) and authenticated in our laboratory by STR as described above. SUM159 cells were cultured in Ham F-12, 5% FBS, hydrocortisone (1 μg/mL; Sigma), insulin (5 μg/mL; Sigma-Aldrich). Cells were routinely tested for mycoplasma infection with MycoAlert PLUS Mycoplasma Detection Kit (Lonza), revealed using Infinite F500 (Tecan).

Both BCSCs and established breast cancer cell lines were exposed to ADSC conditioned medium (ADSC-CM), IL4 (20 ng/mL; PeproTech), and the IL4Rα antagonist (IL4DM; 15 μg/mL), kindly donated by Apogenix. IL4DM was administered for a total of 72 hours. All treatments were replenished every 48 hours.

Withaferin A (Wit A, 2 μmol/L, Tocris) and 5-aminosalicylic acid (5-ASA, 26 mmol/L; Sigma-Aldrich) were used as NF-κB inhibitors and added 1 hour and 30 minutes before IL4 treatment, respectively.

For the cell viability assay, BCSCs were exposed to fulvestrant (1 μmol/L; Selleckchem), docetaxel (100 nmol/L; Selleckchem) or BKM120 (5 μmol/L; Selleckchem). BCSCs were exposed to IFNγ (100 ng/mL; Novus Biologicals), as positive control for PD-L1 expression, for 4 and 24 hours.

To evaluate the expression of PD-1 or IFNγ in gated CD8+ T cells, 5 × 105 PBMCs/mL were activated (Activated) for 4 days in 24-well plates coated with purified anti-human CD3ϵ (OKT3, IgG2a; Biolegend), anti-human CD28 (CD28.2, IgG1κ; Biolegend), and in presence of human rIL2 (100 IU/mL Proleukin; Novartis Pharmaceuticals) in 10% FBS RPMI. Activated cells were then treated, for additional 4 days, with medium (medium), IL4, IL4DM alone, or in combination with IL4 (IL4+IL4DM). Prior flow cytometry analysis, for the intracellular staining of cytokines, untreated and treated activated cells were cultured in presence of PMA (20 ng/mL; BD Biosciences) and ionomycin (1 μmol/L; BD Biosciences) for 4 hours and for the last 3 hours in combination with monensin (2 μmol/L; Sigma-Aldrich), for blocking cellular protein transport.

CM production and Luminex cytokine quantification

CM from ADSCs (ADSCs-CM), ER+-BCSCs (ER+-CM) and TN-BCSCs (TN-CM) were obtained from cells plated at 70% confluence and cultured in SFM for 48 hours. CM was filtered through a 0.22-μm filter to eliminate cell debris.

Quantification of cytokine production was assessed by using multiplex Bio-Plex Pro Assays (Human Cytokine 21 and 27-plex Assay; Bio-Rad). Raw data (mean fluorescence intensity) were analyzed by Bio-Plex Software (Bio-Rad).

Cell motility and invasion assay

A total of 1 × 103 BCSCs were plated into 6-well attachment plates in DMEM with 10% FBS to allow cell attachment. The spreading of cells, which were treated with medium alone, ADSC-CM, or IL4 in presence of 5% FBS, was determined by phase contrast microscopy at 12 hours.

A total of 2 × 103 breast cancer cells, treated with IL4DM for 24 hours and subsequently exposed to ADSCs-CM or IL4 for another 48 hours in SFM, were plated onto growth factor–reduced matrigel (BD Biosciences)-coated transwell of 8-μm pore size. SFM in presence of 10% human serum AB was used as a chemoattractant in the bottom part of the chamber (600 μL/well). Cells invading Matrigel were monitored and counted using an optical microscope for up to 48 hours.

Cell viability

Cell viability was performed using the CellTiter-Glo Luminescent Cell Viability Assay Kit (Promega) according to the manufacturer's instruction. Cell proliferation was assessed using CellTiter 96 AQueous One Solution cell proliferation assay (MTS) according to the manufacturer's instruction. Detection of both luminescence and absorbance was measured by using Infinite F500 (Tecan).

Stem cell frequency and colony-forming efficiency

BCSCs were plated at a concentration of a single cell per well. Wells containing more than three cells were excluded. The stem cell frequency was statistically evaluated after 3 weeks by using ELDA analysis program (23). For the colony-forming assay, 500 dissociated BCSCs were mixed with 0.3% agarose (SeaPlaque Agarose Lonza) in SFM and seeded onto a layer of 0.4% agarose. After 20 days, colonies were stained with 0.01% crystal violet in 1% methanol. Colonies were first distinguished on the basis of their size: micro < 30 μm; small 30–60 μm; medium 60–90 μm; and large >90 μm and then counted using ImageJ software.

Animals and tumor models

Mice experiments were performed according to the ARRIVE and animal care committee guidelines of the University of Palermo. A total of 4 × 103 BCSCs, treated with medium alone or IL4 for 24 hours, were suspended in 100 μL of SFM 1:1 Matrigel (BD Biosciences) and injected subcutaneously into 6-week-old female NOD/SCID mice (Charles River Laboratories). After 1 week, mice were intraperitoneally treated with vehicle (PBS) or IL4DM (3 mg/kg) 5 days/week for 10 weeks (6, 24). One week after the end of treatment, at 12 weeks, mice were sacrificed following IACUC guidelines and tumor xenografts were collected. Tumor growth was monitored weekly with an electronic caliper. Tumor volume was calculated using the formula: largest diameter × (smallest diameter)2 × π/6. For tail vein experiments, 1.5 × 105 luciferase (LUC)-transduced BCSCs (Supplementary Information) were suspended in 30 μL of PBS and injected into 6-week-old female NOD/SCID mice. After injection of VivoGlo Luciferin (150 mg/kg; Promega), in vivo cell spreading was monitored by the detection of bioluminescence intensity using a Photon IMAGER (Biospace Lab), for up to 9 weeks. The photon count (photons/s/sr, photons per second per steradian), emitted by mice metastasis, was calculated by using M3 Vision (Biospace Lab). Mice were sacrificed when lesions reached 4 × 103 photons/s/sr, corresponding to a ≤0.5 cm2 tumor area, and liver and lungs were analyzed ex vivo to detect metastasis formation.

Flow cytometry

ER+- and TN-BCSCs were washed twice in PBS and stained at 4°C with purified CD44-PE (G44-26, mouse IgG2b; BD Biosciences), CD24-APC (ML5, mouse IgG2a; R&D Systems), CD10-APC (97C5, mouse IgG1; Miltenyi Biotec), MUC1-PE (604804, mouse IgG2b; R&D Systems), CD49f-APC (GoH3, rat IgG2a; Miltenyi Biotec), EPCAM-PerCP (EBA-1, mouse IgG1; BD Biosciences), or CD90-PE-CF594 (5E10, IgG1κ; BD Biosciences) antibodies or corresponding IMC CD14-PE (MφP9, IgG2b; BD Biosciences), CD4-APC (11830, IgG2a; R&D Systems), CD4-APC (REA623, IgG1; Miltenyi Biotec), CD8-PE (37006, IgG2b; R&D Systems), CD3-APC (BW264/56, IgG2a; Miltenyi Biotec), CD3-PERCP (SP34-2, IgG1; BD Biosciences), CD3-PE-CF594 (UCHT1, IgG1κ; BD Biosciences). Dead cells were excluded on the basis of light scatter and the uptake of 7-AAD (0.25 μg/L × 106 cells; BD Biosciences) detected in FL3 channel. Single cells were gated in FSC-A versus FSC-H dot plots. Samples were analyzed by FACSCanto II (BD Biosciences) or Accuri C6 (BD Biosciences) flow cytometer and data analyzed using FlowJo software (Tree Star).

PBMCs were incubated on ice with FcR blocking reagent (Miltenyi Biotec) and stained with CD3-PeCy7 (UCHT1, mouse IgG1κ; Biolegend), CD8-PE (RPA-T8, mouse IgG1κ; BD Biosciences), IFNγ-APC (B27, mouse IgG1κ; BD Biosciences), PD-1-FITC (EH12.2H7, mouse IgG1κ; Biolegend) antibodies or corresponding isotype matched controls (IMC), specifically APC mouse IgG1κ (MOPC-21; BD Biosciences) and FITC mouse IgG1κ (MOPC-21; Biolegend). For intracellular staining, cells were previously subjected to Cytofix/Cytoperm protocol following the manufacturer's instructions (BD Biosciences).

Western blot analysis

Cells were washed in ice-cold PBS and suspended in TPER Reagent (Pierce) in presence of NaCl (300 mmol/L; Sigma Aldrich), sodium orthovanadate (1 mmol/L; Sigma Aldrich), pefabloc (2 mmol/L; Roche), and proteinase inhibitor cocktail (5 μg/mL; Sigma Aldrich). Extracted proteins were loaded, separated by SDS-PAGE, and blotted onto nitrocellulose membranes (Hybond-C Extra, nitrocellulose; Amersham Biosciences). Membranes were blocked in 0.1% Tween 20 and 5% non-fat dry milk for 1 hour at room temperature and then exposed to DUSP4 (D9A5, rabbit IgG, Cell Signaling Technology), P-MEK (41G9, rabbit IgG, Cell Signaling Technology), MEK (rabbit polyclonal, Cell Signaling Technology), P-ERK (E-4, mouse IgG2a, Santa Cruz Biotechnology), ERK (rabbit polyclonal, Santa Cruz Biotechnology), RAS (27H5, rabbit IgG, Cell Signaling Technology), P-NF-κB (T.849.2, rabbit IgG, Thermo Fisher Scientific), NF-κB (112A1021, mouse IgG1κ; Thermo Fisher Scientific), P-P38 (D3F9, rabbit IgG; CST), P38 (rabbit polyclonal; Cell Signaling Technology), IL4Rα (25463.111, mouse IgG2a; R&D Systems), IL13Rα2 (goat polyclonal; R&D Systems), P-STAT6 (rabbit polyclonal; Cell Signaling Technology), STAT6 (rabbit polyclonal; Cell Signaling Technology), PD-L1 (E1L3N, rabbit IgG; Cell Signaling Technology), or β-actin (8H10D10, mouse IgG2b; Cell Signaling Technology) antibody. Primary antibodies were revealed using anti-mouse or anti-rabbit HRP–conjugated (goat IgG; Thermo Fisher Scientific) and detected by a chemiluminescence imager (GE Healthcare).

Statistical analysis

Kaplan–Meier curves of relapse-free survival based on the gene expression ratio of IL4/DUSP4 in breast cancer patients were obtained by interrogating the PROGgeneV2 - Pan Cancer Prognostics Database-GSE10893-GPL887. Statistical analysis was calculated using ANOVA with Bonferroni post test. Significance was indicated as P values.

ADSC-CM and IL4 promote expansion of BCSCs

The breast cancer microenvironment is abundantly composed of adipose tissue that constitutes a reservoir of protumorigenic molecules (25, 26). In the attempt to identify the main contributor to breast cancer cells expansion, we examined the composition of the CM from ADSCs (ADSCs-CM) as well as from ER+- and TN-BCSCs (ER+-CM and TN-CM, respectively). Several studies demonstrate that ADSCs promote proliferation and invasive capacity in breast cancer cells (27) and, interestingly, the maintenance of the breast TIC pool (28). Moreover, the use of ADSCs overcomes the difficulties experienced in the isolation and long-term culture of mature adipocytes from adipose tissue (4). ER+- and TN-BCSCs, enriched with cells having stem-like properties, were obtained from surgical resection of breast cancers. ER+-BCSCs were derived from patient #4 and #18, whereas TN-BCSCs from patient #10 and #30 (see Materials and Methods and Table 1), as described previously (22). ER+ cells were positive for MUC1 and EPCAM, and expressed low levels of CD49f, whereas TN cells displayed epithelial-to-mesenchymal transition (EMT) markers such as CD10, CD49f, and VIMENTIN (Supplementary Fig. S1A and S1B).

Exploring the levels of a panel of cytokines across the CMs of ADSCs and ER+- and TN-BCSCs, we detected a higher secretion of proinflammatory and protumorigenic cytokines in CMs of ADSCs and TN-BCSCs. These data suggest that TN-, as opposed to ER+-BCSCs, support the generation of an inflammatory microenvironment, linked to enhanced tumorigenic features (Fig. 1A and B).

In the past, we have reported that IL4 is secreted by primary cancer cells isolated from breast cancer specimens (29) and constitutes an autocrine and paracrine prosurvival signal for breast and colorectal TICs (6, 7, 29). In ADSCs-CM, the secreted IL4 levels were comparable with those in the TN-CM and greater than in the ER+-CM (Fig. 1A). Likewise, ER+ MCF7 cells expressed significant lower IL4 mRNA levels than TN SUM159 established cell line (∼40-fold difference; Supplementary Fig. S1C). Of note, the exposure to ADSCs-CM, for 6 hours, boosted the IL4 mRNA levels of ER+-BCSCs and MCF7 cells (Supplementary Fig. S1D).

Although the ER+-BCSCs morphologic shape was not affected by the exposure to ADSCs-CM or IL4 (data not shown), TN-BCSCs acquired elongated protrusions early on, expanding in all directions (Fig. 1C). Thus, this phenomenon indicates a susceptibility of TN-BCSCs to acquire a migratory phenotype upon microenvironmental stimuli.

We also assessed the capability of breast cancer cells to expand in response to ADSCs-CM and IL4 stimuli and found a stepwise increase in the proliferation of both ER+- and TN-BCSCs as well as bulk MCF7 and SUM159 cells (Fig. 1D and Supplementary Fig. S1E). In presence of ADSCs-CM and IL4, stem cell frequency resulted in a 2- to 3-fold increase of in ER+- and TN-BCSCs (Fig. 1E), and was paralleled by a significant increase in the formation of colonies (Fig. 1F) large in size (>90 μm; Fig. 1G; Supplementary Fig. S1F).

ADSCs boost BCSCs proliferation and invasion in an IL4-dependent manner

To investigate whether IL4 is the main player within ADSCs-secreted cytokines that enhance the aggressiveness of breast cancer cells, we evaluated its receptor expression on ER+- and TN-BCSCs. Both BCSC subtypes displayed IL4Rα along with IL13Rα2, a receptor inhibitor of IL13 signaling (Supplementary Fig. S2A and S2B; ref. 30).

Accordingly, we interfered with IL4 ligand/receptor interaction by using an IL4Rα antagonist, which consists of a double mutant IL4 (IL4DM) generated by the replacement of two amino acids (R121D and Y124D; refs. 24, 31). This molecule binds to IL4Rα and IL13Rα1 with the same affinity as the wild-type IL4 but lacks the agonistic activity (Supplementary Fig. S2C). Because the in vitro exposure to IL4DM for 72 hours impaired the proliferation of ER+- and TN-BCSCs, MCF7, and SUM159 cells (Fig. 2A; Supplementary Fig. S2D), we sought to explore whether IL4 targeting was necessary and sufficient to limit the effect of the ADSCs-CM.

Both ADSCs-CM and IL4 strongly increased the formation of colonies in the 3D culture (Fig. 2B), which highlights that the effect of other paracrine cytokines, including IL13, is only partial. IL4DM was not only able to affect the ability to form colonies (Fig. 2B), but also to impair the invasive potential of ER+- and TN-BCSCs and of the established cell lines MCF7 and SUM159, which counteracted the effect of both IL4 and ADSCs-CM (Fig. 2C; Supplementary Fig. S2E).

We found that BCSCs, pretreated with medium or IL4 and subcutaneously injected into mice, delayed the tumor outgrowth when IL4DM was intraperitoneally administered (Fig. 2D; ref. 6). These findings suggest that IL4DM delays the in vivo growth rate of BCSCs.

IL4DM inhibits MAPK pathway activation, upregulates DUSP4 expression, and sensitizes to anticancer therapy

We wondered whether the administration of IL4DM could be implicated in the modification of the CD44+/CD24 phenotype of breast TICs with a mesenchymal-like state (16). In TN-BCSCs, CD24 expression increased in response to IL4DM, suggesting that tumor cells with stem-like features may transit between a mesenchymal-like and a less aggressive epithelial-like phenotype (Fig. 3A; Supplementary Fig. S3A; refs. 9, 32). We did not notice any significant change in CD24 expression in the exiguous compartment of CD24 cell fraction present in the ER+-BCSCs, following exposure to IL4DM (data not shown).

It has recently been demonstrated that the depletion of CD44+/CD24 population is determined by the exogenous expression of DUSP4 (9). DUSP4 is highly expressed in ER+-BCSCs, whereas it was barely detectable in the TNBC subtype. ER+- and TN-BCSCs, in presence of IL4DM and ADSCs-CM, gradually increased the expression of DUSP4 for up to 24 hours (Fig. 3B).

It was reported that the MAPK pathway activity, including in breast cancer, is partially modulated by DUSP4 (9–12). We hypothesized that the effect of IL4DM in impairing proliferation and invasive potential of BCSCs could be due to the DUSP4-mediated downregulation of MAPK pathway. IL4DM decreased the activity of MEK, ERK, and RAS, at 45 minutes and 24 hours (Fig. 3C), even in presence of ADSCs-CM and IL4 in both ER+- and TN-BCSCs (Fig. 3D and E; Supplementary Fig. S3B–S3D). In ER+-BCSCs, which retain high levels of DUSP4, IL4DM slightly augmented its levels in the absence or presence of ADSCs-CM or IL4 (Fig. 3C–E). Similarly, IL4DM prompted DUSP4 expression in MCF7 and SUM159 cells (Supplementary Fig. S3E). Although IL4DM boosted DUSP4 expression in TN-BCSCs (Fig. 3C), this effect was weaker in presence of IL4 and ADSCs-CM (Fig. 3D and E; Supplementary Fig. S3B–S3D). NF-κB has been reported to be modulated by IL4 treatment in B lymphocytes (33) and to promote the expansion of breast TICs, which contribute to cancer progression (34, 35). To understand how the IL4 signaling blockade is capable of modulating expression levels of DUSP4, we analyzed the activation of NF-κB. The phosphorylation of NF-κB was enhanced by IL4 exposure of both ER+- and TN-BCSCs, whereas its expression levels were barely present in ER+-BCSCs and potently lowered in TN-BCSCs by the addiction of IL4DM to the cell culture (Fig. 3F; Supplementary Fig. S3F). Withaferin A (Wit A), which selectively inhibits IKKβ (36, 37), overcame the effects of IL4, reducing the NF-κB activation to a greater extent in TN-BCSCs (Fig. 3G; Supplementary Fig. S3G). Wit A restored in ER+- and upregulated in TN-BCSCs the DUSP4 expression levels in the presence of IL4 (Fig. 3H). 5-Aminosalicylic acid (5-ASA), another NF-κB inhibitor (38), behaved like Wit A in boosting the expression of DUSP4 in both ER+- and TN-BCSCs (Supplementary Fig. S3H).

Because IL4 blockade circumvents the resistance to conventional antitumor treatment (6, 7), we aimed to couple the inhibition of IL4 with conventional therapies used in breast cancer management. In this context, several clinical trials are evaluating the effectiveness of standard therapy in combination with PI3K inhibitors, such as BKM120, in both ERα-negative and -positive breast cancers (39, 40).

IL4DM in combination with BKM120 and the ERα inhibitor, fulvestrant (Fulv), for ER+-BCSCs, or the chemotherapeutic compound docetaxel (DTX), for TNBC subtype, in presence of microenvironmental cytokines present in the ADSCs-CM, significantly potentiated cell death induced by treatment alone in both ER+- and TN-BCSCs, reaching from 40% to 80% (Fig. 3I).

To determine the specific contribution of IL4 in mediating the refractoriness to therapy, we exposed BCSCs to the combination of IL4DM with the indicated anticancer compounds in presence of IL4. Although ER+-BCSCs exposed to fulvestrant plus BKM120 (Fulv+BKM120) did not benefit from the addition to the treatment of IL4DM (Fig. 3J), the latter significantly increased the efficacy of docetaxel plus BKM120 (DTX+BKM120) in TN-BCSCs at both 24 and 48 hours in the presence or absence of IL4 (Fig. 3J). These data suggest that the inhibition of IL4 achieves only a partial response in ER+-BCSC model due to their extreme sensitivity to fulvestrant and BKM120, whereas enhances drug efficacy in TNBC model.

DUSP4 reduces proliferation, invasion, and metastatic potential of BCSCs

To assess the role of DUSP4 with regard to tumorigenic and metastatic potential, we stably knocked-down DUSP4 (shDUSP4) in ER+-BCSCs and MCF7 cells and ectopically expressed it (DUSP4) in TN-BCSCs and SUM159 cells (Fig. 4A; Supplementary Fig. S4A–S4C). shDUSP4 provoked a slight increase in ER+-BCSC proliferation, which was delayed in TN-BCSCs overexpressing DUSP4 (Fig. 4B). The analysis of the invasion assay revealed a potent capability of ER+-BCSCs harboring shDUSP4 to invade the Matrigel in response to the chemoattractant human serum AB. The ectopic expression of DUSP4 curtailed the invasive potential of TN-BCSCs (Fig. 4C). Moreover, the knockdown of DUSP4 led to an enrichment in cells able to form large colonies, whereas their reduction was dictated by DUSP4 ectopic expression (Fig. 4D).

In accordance with already published data, DUSP4 knockdown, in ER+-BCSCs, increased the CD44+/CD24/CD90 compartment, conversely TN-BCSCs, ectopically overexpressing DUSP4, showed a significant decrease of this cell fraction (Fig. 4E; Supplementary Fig. S4D and S4E) by the modulation of MAPK pathway activation (Fig. 4F; Supplementary Fig. S4F; ref. 9). Likewise, shDUSP4 in MCF7 cells and DUSP4 overexpression in SUM159 cells, respectively, enhanced or decreased the activation of MEK, ERK, and P38 MAPKs (Supplementary Fig. S4G).

By modulating EMT effectors and molecules, which control breast cancer invasion, P38 has been implicated in multiple steps of the metastatic process (41). Thus, we reasoned that DUSP4 could be involved in the metastatic dissemination of BCSCs. In vivo imaging analysis of cells transduced with luciferase showed that shDUSP4 rendered ER+-BCSCs able to colonize the liver and the lung when injected in the tail vein of NOD/SCID mice (Fig. 4G). Conversely, TN-BCSCs overexpressing DUSP4 lost their metastatic potential (Fig. 4H). Collectively, these findings confirm the tumor suppressor role of DUSP4 and unveil a novel role in inhibiting metastasis.

IL4 favors a more permissive microenvironment

We subsequently evaluated the clinical relevance of IL4 and DUSP4 in breast cancer patients by interrogating a publically available database (PROGgeneV2—Pan Cancer Prognostics Database-GSE10893-GPL887). High expression of IL4 and low expression levels of DUSP4 resulted to be associated with a decreased relapse-free survival of patients affected by breast cancer (HR 2.36; 95% CI, 1.15–4.85; P = 0.019; Fig. 5A). IL4 biological main function relies on the modulation of immune response (42). We investigated whether adipose tissue, aided by the infiltrating T lymphocytes, could foster a protumorigenic microenvironment. Tumor cells escape from the cytotoxic activity of CD8+ cells through the expression of PD-L1, which by binding to its receptor, PD-1, promotes the apoptosis of cytotoxic T cells (43). IL4DM markedly diminished the expression of PD-1 on the surface of activated cytotoxic CD8+ T cells, as compared with its control (medium) and regardless of the presence of IL4 (Fig. 5B; Supplementary Fig. S5A and S5B). We found that the treatment with IL4 reduced the percentage (of about 15% vs. medium) of activated cytotoxic CD8+ T cells expressing IFNγ. Although IL4DM alone did not alter T-cell activation, it was notably able to counteract the effect of IL4, restoring the number of cytotoxic T cells expressing IFNγ (Fig. 5C; Supplementary Fig. S5B). As ER+- and TN-BCSCs express PD-L1 (Supplementary Fig. S5C), we hypothesize that IL4DM could potentially limit the activation of the PD-1 pathway by reducing the number of cytotoxic PD-1/CD8+ T-cell compartment. These phenomena suggest the existence of a negative DUSP4 feedback loop that inhibits the MAPK pathway, as well as the production of cytokines involved in the priming of microenvironment that fuels cancer progression (Fig. 5D).

Here, we provided evidence that BCSCs with a metastatic propensity express high levels of IL4 along with a downregulation of DUSP4. The latter is confirmed by the inverse correlation between relapse-free survival of breast cancer patients and IL4 expression. Targeting inflammatory mediators in breast cancer has pointed out appealing endpoints in vitro and in preclinical models (31, 44). Notwithstanding the ongoing development of innovative therapeutic agents that block inflammatory cytokines released by tumor microenvironment and involved in the progression of breast cancer, nowadays the clinical use of these therapies does not show improvement in patient outcome (45). On the basis of our previous findings in colorectal cancer (6), we here investigated the IL4-mediated mechanism that regulates the tumorigenic and metastatic potential of BCSCs.

Targeting IL4 signaling depleted the tumorigenic and metastatic CD44+/CD24 cell fraction, thus delaying the proliferation and invasion capability of BCSCs through the inhibition of MEK, ERK, and RAS activity. Notably, ER+-BCSCs expressed elevated levels of IL4Rα and were promoted to produce IL4 following microenvironmental cytokine stimuli, although to a lesser extent than TN-BCSCs. Our data indicated that nonmetastatic ER+-BCSCs, under tumor microenvironmental influence, acquire an invasive phenotype by strengthening IL4 signaling activity, which is overcome by the blockade of IL4Rα with IL4DM.

We observed that blocking IL4 signaling, by IL4DM, increases the expression of DUSP4 concomitantly with the downregulation of RAS–MAPK pathway. It has recently been demonstrated that in physiologic conditions, the activation of MAPK promotes DUSP4 expression, which in turn suppresses ERK, ensuring breast cell turnover. Nevertheless, loss of DUSP4 and its methylation lead to an aberrant cell proliferation in basal-like breast cancers (9, 10). In addition, the activation of NF-κB signaling plays a crucial role in downregulating the expression levels of DUSP4 in endothelial cells (46). Strikingly, our data revealed that DUSP4 is inhibited by NF-κB activation in an IL4-dependent manner.

Compelling data indicate that P38 and ERK activation promotes cancer cell dormancy, allowing their survival in a hostile microenvironment. In addition, P38 induces EMT awakening cancer cells from dormancy (41, 47). The overexpression of DUSP4 potently impaired ERK and P38 MAPK pathway, preventing the intrinsic metastatic capability of BCSCs.

Within tumor microenvironment, aside from ADSCs, infiltrating CD4+ Th2 lymphocytes are the other main source of IL4 (42). IL4 mediates the switching of CD4+ Th1 cells to CD4+ Th2 cells affecting their innate antitumor immunity (42). Although the activation of IL4/STAT6 signal boosts both the tumorigenic and metastatic activity of breast cancer cells, its loss exerts an antitumor activity through the activation of CD8+ T cells, in a CD4+ LTh cell-independent manner (19, 48). Although the role of Th2 cytokines in cancer is still contradictory and context dependent (49), we here demonstrated that targeting IL4 pathway fostered the cytotoxicity of CD8+ T cells, by increasing intracellular IFNγ content and decreasing their PD-1 expression, expression that is related with a shorter overall survival in epithelial-originated cancer patients (50).

Finally, our findings propose the targeting of IL4 signaling as a powerful approach in reducing tumor burden and metastatic colonization. Several commercially available compounds that block IL4 signaling are presently being utilized in the treatment of asthma and some have already been tested in clinical trials as anticancer agents (8). Their administration, in combination with other novel more effective therapies such as MEK, ERK, and PI3K inhibitors, warrants further investigation to establish their effective doses and efficacy.

No potential conflicts of interest were disclosed.

Conception and design: M. Gaggianesi, A. Turdo, S. Vieni, G. Stassi, M. Todaro

Development of methodology: M. Gaggianesi, A. Turdo, A. Chinnici, E. Lipari, I. Sperduti, E. Lo Presti, S. Vieni

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Apuzzo, A. Benfante, S. Di Franco, V. Caputo, S. Vieni

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Chinnici, E. Lipari, T. Apuzzo, A. Benfante, I. Sperduti, S. Di Franco, S. Meraviglia, E. Lo Presti, G. Militello

Writing, review, and/or revision of the manuscript: M. Gaggianesi, A. Turdo, I. Sperduti, F. Dieli, S. Vieni, G. Stassi, M. Todaro

Study supervision: M. Todaro

The authors thank Prof. Max Wicha for kindly providing SUM159 cell line and Apogenix GmbH, Heidelberg for supplying IL4DM. We thank Francesco Calò for his help on the graphic images, Alessandro Gorgone and Giovanni Tomaselli for their technical assistance, Tatiana Terranova for editing the manuscript, Emanuela Scavo for STR genotyping analysis, Alessandro Giammona for providing assistance in experiments involving ADSCs-CM. A. Turdo and S. Di Franco are recipients of an Associazione Italiana per la Ricerca sul Cancro (AIRC) fellowship. A. Chinnici is a PhD student of molecular and experimental medicine PhD programme at University of Palermo in partnership with Humanitas University.

This work was supported by grants from Associazione Italiana per la Ricerca sul Cancro (AIRC) 14415 to M. Todaro and AIRC 16746 to G. Stassi.

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.

1.
Arnedos
M
,
Vicier
C
,
Loi
S
,
Lefebvre
C
,
Michiels
S
,
Bonnefoi
H
, et al
Precision medicine for metastatic breast cancer—limitations and solutions
.
Nat Rev Clin Oncol
2015
;
12
:
693
704
.
2.
Wang
YY
,
Lehuede
C
,
Laurent
V
,
Dirat
B
,
Dauvillier
S
,
Bochet
L
, et al
Adipose tissue and breast epithelial cells: a dangerous dynamic duo in breast cancer
.
Cancer Lett
2012
;
324
:
142
51
.
3.
Sinicrope
FA
,
Dannenberg
AJ
. 
Obesity and breast cancer prognosis: weight of the evidence
.
J Clin Oncol
2011
;
29
:
4
7
.
4.
Laurent
V
,
Guerard
A
,
Mazerolles
C
,
Le Gonidec
S
,
Toulet
A
,
Nieto
L
, et al
Periprostatic adipocytes act as a driving force for prostate cancer progression in obesity
.
Nat Commun
2016
;
7
:
10230
.
5.
Iyengar
NM
,
Hudis
CA
,
Dannenberg
AJ
. 
Obesity and inflammation: new insights into breast cancer development and progression
.
Am Soc Clin Oncol Educ Book
2013
:
46
51
.
6.
Todaro
M
,
Alea
MP
,
Di Stefano
AB
,
Cammareri
P
,
Vermeulen
L
,
Iovino
F
, et al
Colon cancer stem cells dictate tumor growth and resist cell death by production of interleukin-4
.
Cell Stem Cell
2007
;
1
:
389
402
.
7.
Francipane
MG
,
Alea
MP
,
Lombardo
Y
,
Todaro
M
,
Medema
JP
,
Stassi
G
. 
Crucial role of interleukin-4 in the survival of colon cancer stem cells
.
Cancer Res
2008
;
68
:
4022
5
.
8.
Bankaitis
KV
,
Fingleton
B
. 
Targeting IL4/IL4R for the treatment of epithelial cancer metastasis
.
Clin Exp Metastasis
2015
;
32
:
847
56
.
9.
Balko
JM
,
Schwarz
LJ
,
Bhola
NE
,
Kurupi
R
,
Owens
P
,
Miller
TW
, et al
Activation of MAPK pathways due to DUSP4 loss promotes cancer stem cell-like phenotypes in basal-like breast cancer
.
Cancer Res
2013
;
73
:
6346
58
.
10.
Mazumdar
A
,
Poage
GM
,
Shepherd
J
,
Tsimelzon
A
,
Hartman
ZC
,
Den Hollander
P
, et al
Analysis of phosphatases in ER-negative breast cancers identifies DUSP4 as a critical regulator of growth and invasion
.
Breast Cancer Res Treat
2016
;
158
:
441
54
.
11.
Cagnol
S
,
Rivard
N
. 
Oncogenic KRAS and BRAF activation of the MEK/ERK signaling pathway promotes expression of dual-specificity phosphatase 4 (DUSP4/MKP2) resulting in nuclear ERK1/2 inhibition
.
Oncogene
2013
;
32
:
564
76
.
12.
Hijiya
N
,
Tsukamoto
Y
,
Nakada
C
,
Tung Nguyen
L
,
Kai
T
,
Matsuura
K
, et al
Genomic Loss of DUSP4 contributes to the progression of intraepithelial neoplasm of pancreas to invasive carcinoma
.
Cancer Res
2016
;
76
:
2612
25
.
13.
Zhou
BB
,
Zhang
H
,
Damelin
M
,
Geles
KG
,
Grindley
JC
,
Dirks
PB
. 
Tumour-initiating cells: challenges and opportunities for anticancer drug discovery
.
Nat Rev Drug Discov
2009
;
8
:
806
23
.
14.
Zeuner
A
,
Todaro
M
,
Stassi
G
,
De Maria
R
. 
Colorectal cancer stem cells: from the crypt to the clinic
.
Cell Stem Cell
2014
;
15
:
692
705
.
15.
Valent
P
,
Bonnet
D
,
De Maria
R
,
Lapidot
T
,
Copland
M
,
Melo
JV
, et al
Cancer stem cell definitions and terminology: the devil is in the details
.
Nat Rev Cancer
2012
;
12
:
767
75
.
16.
Al-Hajj
M
,
Wicha
MS
,
Benito-Hernandez
A
,
Morrison
SJ
,
Clarke
MF
. 
Prospective identification of tumorigenic breast cancer cells
.
Proc Natl Acad Sci U S A
2003
;
100
:
3983
8
.
17.
Korkaya
H
,
Kim
GI
,
Davis
A
,
Malik
F
,
Henry
NL
,
Ithimakin
S
, et al
Activation of an IL6 inflammatory loop mediates trastuzumab resistance in HER2+ breast cancer by expanding the cancer stem cell population
.
Mol Cell
2012
;
47
:
570
84
.
18.
Saha
S
,
Mukherjee
S
,
Khan
P
,
Kajal
K
,
Mazumdar
M
,
Manna
A
, et al
Aspirin suppresses the acquisition of chemoresistance in breast cancer by disrupting an NF-κB-IL6 signaling axis responsible for the generation of cancer stem cells
.
Cancer Res
2016
;
76
:
2000
12
.
19.
Venmar
KT
,
Carter
KJ
,
Hwang
DG
,
Dozier
EA
,
Fingleton
B
. 
IL4 receptor ILR4α regulates metastatic colonization by mammary tumors through multiple signaling pathways
.
Cancer Res
2014
;
74
:
4329
40
.
20.
Schnitt
SJ.
Classification and prognosis of invasive breast cancer: from morphology to molecular taxonomy
.
Mod Pathol
2010
;
23
:
S60
4
.
21.
Todaro
M
,
Gaggianesi
M
,
Catalano
V
,
Benfante
A
,
Iovino
F
,
Biffoni
M
, et al
CD44v6 is a marker of constitutive and reprogrammed cancer stem cells driving colon cancer metastasis
.
Cell Stem Cell
2014
;
14
:
342
56
.
22.
Di Franco
S
,
Turdo
A
,
Benfante
A
,
Colorito
ML
,
Gaggianesi
M
,
Apuzzo
T
, et al
DeltaNp63 drives metastasis in breast cancer cells via PI3K/CD44v6 axis
.
Oncotarget
2016
;
7
:
54157
73
.
23.
Hu
Y
,
Smyth
GK
. 
ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays
.
J Immunol Methods
2009
;
347
:
70
8
.
24.
Zurawski
SM
,
Vega
F
 Jr
,
Huyghe
B
,
Zurawski
G
. 
Receptors for interleukin-13 and interleukin-4 are complex and share a novel component that functions in signal transduction
.
EMBO J
1993
;
12
:
2663
70
.
25.
Picon-Ruiz
M
,
Pan
C
,
Drews-Elger
K
,
Jang
K
,
Besser
AH
,
Zhao
D
, et al
Interactions between adipocytes and breast cancer cells stimulate cytokine production and drive Src/Sox2/miR-302b-mediated malignant progression
.
Cancer Res
2016
;
76
:
491
504
.
26.
Gilbert
CA
,
Slingerland
JM
. 
Cytokines, obesity, and cancer: new insights on mechanisms linking obesity to cancer risk and progression
.
Annu Rev Med
2013
;
64
:
45
57
.
27.
Walter
M
,
Liang
S
,
Ghosh
S
,
Hornsby
PJ
,
Li
R
. 
Interleukin 6 secreted from adipose stromal cells promotes migration and invasion of breast cancer cells
.
Oncogene
2009
;
28
:
2745
55
.
28.
Devarajan
E
,
Song
YH
,
Krishnappa
S
,
Alt
E
. 
Epithelial-mesenchymal transition in breast cancer lines is mediated through PDGF-D released by tissue-resident stem cells
.
Int J Cancer
2012
;
131
:
1023
31
.
29.
Todaro
M
,
Lombardo
Y
,
Francipane
MG
,
Alea
MP
,
Cammareri
P
,
Iovino
F
, et al
Apoptosis resistance in epithelial tumors is mediated by tumor-cell-derived interleukin-4
.
Cell Death Differ
2008
;
15
:
762
72
.
30.
Rahaman
SO
,
Sharma
P
,
Harbor
PC
,
Aman
MJ
,
Vogelbaum
MA
,
Haque
SJ
. 
IL-13R(α)2, a decoy receptor for IL-13 acts as an inhibitor of IL-4-dependent signal transduction in glioblastoma cells
.
Cancer Res
2002
;
62
:
1103
9
.
31.
Andrews
AL
,
Holloway
JW
,
Holgate
ST
,
Davies
DE
. 
IL-4 receptor α is an important modulator of IL-4 and IL-13 receptor binding: implications for the development of therapeutic targets
.
J Immunol
2006
;
176
:
7456
61
.
32.
Liu
S
,
Cong
Y
,
Wang
D
,
Sun
Y
,
Deng
L
,
Liu
Y
, et al
Breast cancer stem cells transition between epithelial and mesenchymal states reflective of their normal counterparts
.
Stem Cell Rep
2014
;
2
:
78
91
.
33.
Thieu
VT
,
Nguyen
ET
,
McCarthy
BP
,
Bruns
HA
,
Kapur
R
,
Chang
CH
, et al
IL-4-stimulated NF-κB activity is required for Stat6 DNA binding
.
J Leukocyte Biol
2007
;
82
:
370
9
.
34.
Kendellen
MF
,
Bradford
JW
,
Lawrence
CL
,
Clark
KS
,
Baldwin
AS
. 
Canonical and non-canonical NF-κB signaling promotes breast cancer tumor-initiating cells
.
Oncogene
2014
;
33
:
1297
305
.
35.
Biswas
DK
,
Shi
Q
,
Baily
S
,
Strickland
I
,
Ghosh
S
,
Pardee
AB
, et al
NF-κ B activation in human breast cancer specimens and its role in cell proliferation and apoptosis
.
Proc Natl Acad Sci U S A
2004
;
101
:
10137
42
.
36.
Kaileh
M
,
Vanden Berghe
W
,
Heyerick
A
,
Horion
J
,
Piette
J
,
Libert
C
, et al
Withaferin a strongly elicits IκB kinase β hyperphosphorylation concomitant with potent inhibition of its kinase activity
.
J Biol Chem
2007
;
282
:
4253
64
.
37.
Heyninck
K
,
Lahtela-Kakkonen
M
,
Van der Veken
P
,
Haegeman
G
,
Vanden Berghe
W
. 
Withaferin A inhibits NF-κB activation by targeting cysteine 179 in IKKβ
.
Biochem Pharmacol
2014
;
91
:
501
9
.
38.
Egan
LJ
,
Mays
DC
,
Huntoon
CJ
,
Bell
MP
,
Pike
MG
,
Sandborn
WJ
, et al
Inhibition of interleukin-1-stimulated NF-κB RelA/p65 phosphorylation by mesalamine is accompanied by decreased transcriptional activity
.
J Biol Chem
1999
;
274
:
26448
53
.
39.
Wallin
JJ
,
Guan
J
,
Prior
WW
,
Lee
LB
,
Berry
L
,
Belmont
LD
, et al
GDC-0941, a novel class I selective PI3K inhibitor, enhances the efficacy of docetaxel in human breast cancer models by increasing cell death in vitro and in vivo
.
Clin Cancer Res
2012
;
18
:
3901
11
.
40.
Yamamoto-Ibusuki
M
,
Arnedos
M
,
Andre
F
. 
Targeted therapies for ER+/HER2- metastatic breast cancer
.
BMC Med
2015
;
13
:
137
.
41.
del Barco Barrantes
I
,
Nebreda
AR
. 
Roles of p38 MAPKs in invasion and metastasis
.
Biochem Soc Transact
2012
;
40
:
79
84
.
42.
Kacha
AK
,
Fallarino
F
,
Markiewicz
MA
,
Gajewski
TF
. 
Cutting edge: spontaneous rejection of poorly immunogenic P1.HTR tumors by Stat6-deficient mice
.
J Immunol
2000
;
165
:
6024
8
.
43.
Chen
L
,
Han
X
. 
Anti-PD-1/PD-L1 therapy of human cancer: past, present, and future
.
J Clin Invest
2015
;
125
:
3384
91
.
44.
Ginestier
C
,
Liu
S
,
Diebel
ME
,
Korkaya
H
,
Luo
M
,
Brown
M
, et al
CXCR1 blockade selectively targets human breast cancer stem cells in vitro and in xenografts
.
J Clin Invest
2010
;
120
:
485
97
.
45.
Crusz
SM
,
Balkwill
FR
. 
Inflammation and cancer: advances and new agents
.
Nat Rev Clin Oncol
2015
;
12
:
584
96
.
46.
Kao
DD
,
Oldebeken
SR
,
Rai
A
,
Lubos
E
,
Leopold
JA
,
Loscalzo
J
, et al
Tumor necrosis factor-α-mediated suppression of dual-specificity phosphatase 4: crosstalk between NF-κB and MAPK regulates endothelial cell survival
.
Mol Cell Biochem
2013
;
382
:
153
62
.
47.
Sosa
MS
,
Avivar-Valderas
A
,
Bragado
P
,
Wen
HC
,
Aguirre-Ghiso
JA
. 
ERK1/2 and p38α/β signaling in tumor cell quiescence: opportunities to control dormant residual disease
.
Clin Cancer Res
2011
;
17
:
5850
7
.
48.
Ostrand-Rosenberg
S
,
Grusby
MJ
,
Clements
VK
. 
Cutting edge: STAT6-deficient mice have enhanced tumor immunity to primary and metastatic mammary carcinoma
.
J Immunol
2000
;
165
:
6015
9
.
49.
Yu
P
,
Fu
YX
. 
Tumor-infiltrating T lymphocytes: friends or foes?
Lab Invest
2006
;
86
:
231
45
.
50.
Zhang
Y
,
Kang
S
,
Shen
J
,
He
J
,
Jiang
L
,
Wang
W
, et al
Prognostic significance of programmed cell death 1 (PD-1) or PD-1 ligand 1 (PD-L1) Expression in epithelial-originated cancer: a meta-analysis
.
Medicine
2015
;
94
:
e515
.