- Email: [email protected]
CXCR2 promotes breast cancer metastasis and chemoresistance via suppression of AKT1 and activation of COX2
Cancer Letters 412 (2018) 69e80
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
Original Article
CXCR2 promotes breast cancer metastasis and chemoresistance via suppression of AKT1 and activation of COX2 Han Xu a, b, c, 1, Fengjuan Lin a, b, c, 1, Ziliang Wang a, c, Lina Yang a, c, Jiao Meng a, c, Zhouluo Ou b, c, Zhimin Shao a, b, c, **, Genhong Di b, c, ***, Gong Yang a, c, d, * a
Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China Department of Breast Surgery, Fudan University Shanghai Cancer Center, Shanghai, 200032, China Department of Oncology, Shanghai Medical College, Fudan University, 200032, China d Central Laboratory, The Fifth People's Hospital of Shanghai, Fudan University, 200240, China b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 April 2017 Received in revised form 17 September 2017 Accepted 21 September 2017
Metastasis and chemoresistance are two major causes of breast cancer death. We show here that the chemokine receptor CXCR2 was overexpressed in breast cancer cell lines and tissues. CXCR2 promoted anti-apoptosis, anti-senescence, and epithelial-to-mesenchymal transition (EMT) of breast cancer cells, leading to the enhanced metastasis and chemoresistance. Further study suggested that AKT1 and cyclooxygenase-2 (COX2; PTGS2) might mediate the CXCR2 signaling to inversely control the breast cancer metastasis and chemoresistance through the regulation of EMT, apoptosis, and senescence. Analyses of clinical data indicate that the high expression of CXCR2 was correlated with the high expression of COX2 and the low expression of AKT1, P85a, E-cadherin, and b-catenin in cancer tissues. Poor outcomes were associated with the high expression of CXCR2 or COX2 while favorable survivals were associated with the high expression of P85a, AKT1, or E-cadherin in all cancer patients. Cox multivariate analysis demonstrated that CXCR2, COX2, and AKT1 could be independent predictors for disease free survivals. All these data suggest that CXCR2 promotes breast cancer metastasis and chemoresistance via suppressing AKT1 and activating COX2. Thus, antagonists of the CXCR2 signaling molecules may be used to treat breast cancer patients particularly with high metastasis and chemoresistance. © 2017 Elsevier B.V. All rights reserved.
Keywords: CXCR2 AKT1 COX2 Breast cancer Metastasis
Introduction Although the overall survival of breast cancer at early stage is improved by standard chemotherapy following surgery, the number of death is still climbing because of distant metastasis and chemoresistance. Studies have shown that chemokines and their receptors mediate breast cancer metastasis [1,2]. Interleukin 8 (IL8) has been proved to be an essential molecule regulating breast
* Corresponding author. Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. ** Corresponding author. Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. *** Corresponding author. Department of Breast Surgery, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. E-mail addresses: [email protected] (Z. Shao), [email protected] (G. Di), [email protected] (G. Yang). 1 These authors equally contributed to this work. https://doi.org/10.1016/j.canlet.2017.09.030 0304-3835/© 2017 Elsevier B.V. All rights reserved.
cancer metastasis and chemoresistance. A recent study from Britschgi et al. suggests that IL-8 mediates resistance to the PI3Kpathway-targeted therapy in breast cancer [3]. Some studies have shown that CXCR2, the co-receptor of IL-8 and Gro-a, may enhance breast cancer migration, invasion, and metastasis [4] or mediate breast cancer bone metastasis through promoting the migration of breast mesenchymal stem cells [5]. Inhibition of both CXCR1/2 reduces the activity of breast cancer stem cells and improves the survival of HER2-positive patients in combination with the HER2targeted chemotherapies [6]. However, how the CXCR2 signaling regulates breast cancer metastasis and chemoresistance is largely unclear. Studies have shown that AKT regulates chemoresistance through PED, a broad anti-apoptotic molecule [7], and that the phosphorylation of AKT at Ser 473 promotes breast cancer metastasis [8]. However, a clinical study suggests that breast cancer patients with the phosphorylation of AKT at Ser473 appear to be sensitive to treatment with paclitaxel [9]. These controversial
70
H. Xu et al. / Cancer Letters 412 (2018) 69e80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
results indicate that AKT may have different functions in terms of tumor progression, metastasis and chemoresistance, but the detailed signaling has yet to be defined. A few studies have shown that the PI3K/AKT pathway may be involved in the CXCR2 signaling through its ligands IL-8 and Gro-a [10], but the essential mechanism how CXCR2 modulates breast cancer metastasis and chemoresistance through the PI3K/AKT pathway is unclear. Cyclooxygenase 2 (COX2, also known as the prostaglandin H synthase-2, PGHS-2) is usually up-regulated in many cancers. In breast cancer, COX2 can be used as a biomarker to stratify breast cancer risk in women with atypical hyperplasia [11]. COX2 also promotes progression of ductal carcinoma in situ (DCIS) to invasive breast carcinomas [12], osteolytic bone metastasis of breast cancer [13] and tumor angiogenesis [14]. Overexpression of COX2 confers chemoresistance in breast cancer through PGH2 generation and NF-kB activation [15]. Although a number of studies have reported that COX2 stimulates IL-8 expression to enhance breast cancer invasion [16] or bone metastasis [17], no associated studies between COX2 and CXCR2 have been reported in breast cancer. In this study, we employed breast cancer cell lines, animals, and human specimens to investigate how the CXCR2-associated signaling is involved in breast cancer metastasis and chemoresistance. Materials and methods Cell lines and cell culture Normal breast epithelial cell lines HMEC207, HMEC212, immortalized breast epithelial cell line MCF-10A, low-metastatic breast cancer cell lines MCF-7,T47D, BT20, BT474, BT483, SKBR3, high-metastatic breast cancer cell lines MDA-MB-231 (simplified as 231), MDA-MB-231HM (high lung metastasis, simplified as 231HM), MDA-MB-231BO (high bone metastasis, simplified as 231BO), and the gemcitabineresistant breast cancer cell line MDA-MB-231Gem (simplified as 231Gem) were used in this study. MCF-7, BT474, BT483, SKBR3, 231 and 231BO cell lines were obtained from the American Tissue Culture Collection (ATCC). 231HM and 231Gem cell lines were established in our institute. Cells were routinely maintained in the recommended medium supplemented with 10% fetal bovine serum, 100 mg/mL penicillin, and 100 mg/mL streptomycin. All cultures were incubated at 37 C in a humidified 5% CO2 atmosphere. Drug treatment and apoptosis detection Cells were treated with either taxol at 50 mg/ml, gemcitabine at 4 mM (for cells), Cell apoptosis induced by taxol or gemcitabine was assessed by using flow cytometry after cells were stained with propidium iodide (PI) and Annexin V. Briefly, 1 106 cells seeded in 60 mm dish for 24 h were treated with taxol or gemcitabine for 12 h, and then trypsinized and washed twice in ice-cold phosphate-buffered saline (PBS). A total of 1 105 cells were resuspended in 100 mL binding buffer to which was added 5 mL 2 mg/ml Annexin V and 5 ml 50 mg/ml PI. Following 15 min incubation in dark, the cells were detected by flow cytometry under different channels. All tests were repeated three times. Examination of cellular senescence To detect senescence-associated b-galacatosidase activity, cells were fixed with 2% formaldehyde and 0.2% glutaraldehyde at room temperature for 4 min and then stained with the buffer containing 150 mM NaCI, 20 mM MgCI2,10 mg/ml X-gal, 50 mM Potassium Ferricyanide, 50 mM Potassium Ferrocyanide, and 1.6 mM citric acid (pH 6.0) at 37 C for 2e15 h. Western blot was used to detect the expression of senescence-associated molecules including p16, p21, and HP-1g. Statistical analysis Statistical analysis was performed by t-test at different time points between the mean tumor sizes of each group. The statistical significance of the differences in cell growth, migration, colony formation, tumor volume and the rate of spontaneous metastasis was calculated by using the Student's t-test. The relation between protein
71
expression in tissue arrays and patient pathological characteristics was analyzed by X2 test or Fisher's exact test. Disease-specific survival rates were determined by using the Kaplan-Meier method and the log-rank test from SPSS 18.0 software (SPSS, Inc). The P value smaller than 0.05 was considered statistically significant. All statistical tests were two-sided. Graphs were created with GraphPad Prism 5 and Adobe Photoshop CS6. Plasmid construction and delivery to cells, qPCR, Western blotting, immunofluorescence, cell proliferation, migration, invasion and anchorage-independent assay, Xenograft tumor growth, animal treatment, immunostaining and analysis of animal tissues and human tissue microarray were available in Supplemental Materials and Methods.
Results CXCR2 stimulates breast cancer cell proliferation and tumorigenesis High level of CXCR2 was clearly detected in breast cancer cell lines 231, 231HM, 231BO, SKBR3, and T47D, while low level of CXCR2 was detected in breast cancer cell lines BT20, BT474, BT483, MCF-7, and in immortalized breast epithelial cell line MCF-10A, as well as in normal breast epithelial cell lines HMEC207 and HMEC212. Six of nine (6/9, 66.7%) breast cancer tissues were found with high expression of CXCR2, whereas low CXCR2 was detected in normal human mammary tissues (Fig. 1A and B). Overexpression or silencing of CXCR2 (Fig. 1C and D) significantly promoted or repressed cell proliferation (SFig. 1A), soft agar colony formation (Fig. 1E and F), and the mammary fat pad tumor growth in animals (Fig. 1G, SFig. 1B), compared with controls, although MCF-7 and BT474 cell lines failed to generate tumor burdens in animals without regard to the CXCR2 status. These data suggest that CXCR2 promotes breast cancer cell proliferation and tumorigenesis. CXCR2 promotes breast cancer cell migration, invasion and metastasis via the P85a/AKT1-mediated EMT signaling Overexpression or knockdown of CXCR2 increased or decreased cell migration and invasion in vitro (Fig. 2A and SFig. 2A and B). Moreover, cells expressing CXCR2 shRNA induced fewer number of micrometastasis in lung tissues than those expressing control vector in vivo (Fig. 2B and C). In addition, CXCR2 knockdown also reduced the number of micrometastatic sites in other organs including bone, liver, brain and axillary lymph nodes (STable 1). All these data suggest that CXCR2 promotes breast cancer metastasis. It is well-known that the epithelial to mesenchymal transition (EMT) may play an important role in cancer metastasis. Immunofluorescent staining showed that MCF-7 cells overexpressing CXCR2 became spindle, while 231 cells expressing CXCR2 shRNA appeared round (Fig. 2D), suggesting that CXCR2 might mediate the EMT signaling. Analysis of the gene expression profiles showed that mRNAs of the genes associated with EMT were altered by downregulation of CXCR2 in 231 and 231HM cells, compared with control cells (Fig. 2E). By Western blotting, we found that CXCR2 downregulated the expression of PI3K-P85a, total AKT, and AKT1 (Fig. 2F), but not that of PI3K-P110a, AKT2 and AKT3 (data not shown). However, CXCR2 up-regulated the phosphorylated AKT (Ser473), but not pAKTThr308 (Fig. 2F), suggesting that CXCR2 may promote metastasis through suppressing P85a and AKT1 expression and enhancing the phosphorylation of AKT at Ser473. We also found that the expression of E-cadherin and b-catenin was either
Fig. 1. CXCR2 stimulates breast cancer cell proliferation and tumor growth. A, CXCR2 expression in normal breast epithelial cells, breast cancer cell lines (upper panel), and tissues (lower panel). b-actin was used as a loading control. B, CXCR2 expression in normal breast tissues and breast tumor tissues (left panel, 200; right panel, 400). Tissues were stained with a rabbit anti-CXCR2 antibody and visualized following a donkey anti-rabbit secondary antibody. C-D, Analyses of CXCR2 by Western blot (C) and realtime-PCR (D) in breast cancer cells transfected with CXCR2 cDNA (CXCR2), shRNA(CXCR2i), or control vectors. E-F, In vitro tumorigenicity tested by soft agar assay. G, Fat-pad tumor growth of mice injected with cells expressing either CXCR2i or scrambled shRNA (Scr). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
72
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 2. CXCR2 promotes breast cancer cell migration, invasion and metastasis via the P85a/AKT1 signaling-mediated EMT. A, Quantitative analysis of migration speed by wound healing scratch assay (left panel) and invaded cells by Matrigel cell invasion assay (right panel). B, Detection of tumor micrometastasis by using H&E staining of lung tissues derived from xenograft nude mice injected with breast cancer cells expressing either CXCR2 shRNA or control vectors (left panel, 200; right panel, 400). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. C, Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). D, EMT-like morphological changes of breast cancer cells with CXCR2 overexpression or silencing detected by immunofluorescence (Magnification 400). Scar bars, 10 mm. E, Gene expression profiles of EMT-associated molecules in 231/Scr, 231/CXCR2i, 231HM/Scr and 231HM/CXCR2i cell lines. The green and red colors represent genes down-regulated and up-regulated by CXCR2 silencing, respectively. F, Immunoblotting analysis of PI3K(P85a), PI3K(P110a), pAKT (Ser473), pAKT (Thr 308), AKT and AKT1, E-cadherin, b-catenin and caveolin-1 in breast cancer cells transfected with CXCR2 cDNA or shRNA and their control vectors. G, Immunostaining analyses of lung tissues suggest a slightly increased cytokeratin but decreased vimentin in breast cancer cells with CXCR2 silencing (Magnification 200). H, Immunoblotting analysis of P85a, AKT1 and the metastasis-associated proteins. All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
decreased or increased by CXCR2 overexpression or knock down, and that caveolin-1 was reduced after CXCR2 was silenced (Fig. 2F). Furthermore, we showed that silencing of CXCR2 slightly increased the cytokeratin 8/18 expression, but decreased the expression of vimentin in micrometastatic tumor cells of lung tissues from tumor bearing mice (Fig. 2G). Additionally, by overexpression or silencing of P85a, we found that P85a could up-regulate AKT1, E-cadherin and b-catenin expression, but down-regulate caveolin 1 expression to counteract the CXCR2-mediated signaling (Fig. 2H). These data indicate that CXCR2 may promote breast cancer metastasis mainly through the P85a/AKT1-mediated EMT signaling.
CXCR2 induces breast cancer chemoresistance via the anti-apoptosis and pro-metastasis signaling
or
By flow cytometer, we first found that CXCR2 overexpression silencing reduced or promoted Taxol-induced cellular
apoptosis (Fig. 3A and B). We then performed gene expression profile analysis and found a clear alteration of the p53-associated apoptosis signaling molecules particularly after CXCR2 was disrupted in 231 and 231HM cells (Fig. 3C). The expression of the apoptosis-related proteins including Bax, Bak, Bad, Bid, Bcl-2 and Bcl-xL were confirmed by Western blot (Fig. 3D). Thus, CXCR2 appears to promote breast cancer chemoresistance possibly by suppressing cellular apoptosis. Silencing of CXCR2 also markedly enhanced the sensitivity of xenograft tumors to Taxol treatment, compared with controls (Fig. 3E and F). Although treatment of animals with taxol did not significantly reduce the volume of tumors formed by control cells (231/Scr, 231HM/Scr), the number of micrometastasis in lung tissues from all animals was significantly reduced by taxol treatment. (Fig. 3G). These data suggest that CXCR2 induces breast cancer cell chemoresistance likely through suppression of the p53-mediated apoptosis signaling.
H. Xu et al. / Cancer Letters 412 (2018) 69e80
73
Fig. 3. CXCR2 induces breast cancer chemoresistance via the anti-apoptosis and pro-metastasis signaling. A, Taxol-induced apoptosis detected in CXCR2 overexpressing cell lines (MCF-7 and BT474) and in CXCR2 silencing cell lines (231, 231HM and 231BO) by flow cytometry. B, Quantitative analysis of apoptotic cells induced by taxol. C, Gene expression profiles of apoptosis molecules in 231/Scr, 231/CXCR2i, 231HM/Scr and 231HM/CXCR2i cell lines. The green and red colors represent genes down-regulated and up-regulated by CXCR2 silencing, respectively. D, Apoptosis-related proteins detected by Western blotting. b-actin was used as a loading control. E-F, Tumor growth following subcutaneous implantation of control and CXCR2 knockdown cells in vivo treated with taxol or placebo. G, Detection of tumor micrometastasis in lung tissues of xenograft nude mice treated with taxol (left panel). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed) (right panel). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
74
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 4. AKT1 counteracts CXCR2 to suppress breast cancer tumorigenesis, metastasis and chemoresistance. A, Fat-pad tumor growth from nude mice injected with cells expressing 231/CXCR2i/Scr or 231/CXCR2i/AKT1i. BeC, Quantitative analysis of migration and invasion with the migration index and the number of invaded cells. D, Detection of tumor micrometastasis in lung tissues of nude mice injected with 231/CXCR2i/AKT1i cells and control cells (left panel, 200; right panel, 400). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). E, Taxol-induced apoptosis detected by flow cytometry in cell lines (left panel). Quantitative analysis of apoptotic cells (right panel). F, Gene expression profiles of metastasis and apoptosis molecules in 231/Scr, 231/CXCR2i, 231/AKT1 and 231/CXCR2i/AKT1i cell lines. The green and red colors represent genes downregulated and up-regulated by AKT1 overexpression and silencing, respectively. G, Immunoblotting analysis of EMT and apoptosis-associated proteins in MCF-7/CXCR2 cells transfected with AKT1 cDNA or vector (left panel), and in 231/CXCR2i cells transfected with AKT1 shRNA or scramble shRNA (right panel). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
H. Xu et al. / Cancer Letters 412 (2018) 69e80
AKT1 counteracts CXCR2 to suppress breast cancer tumorigenesis, metastasis and chemoresistance To verify the function of AKT1 in the CXCR2-mediated signaling, we delivered AKT1 cDNAs or AKT1 shRNAs into MCF-7/CXCR2 or 231/CXCR2i cells (Fig. 4G) and found that AKT1 counteracted the CXCR2 signaling to block cancer growth/metastasis in vivo (Fig. 4AeD) and to attenuate cell migration/invasion in vitro (Fig. 4B and C, SFig. 2C and D). In addition, treatment of AKT1 overexpressing or silencing cells with taxol induced more or less apoptosis than did treatment of control cells with taxol (Fig. 4E). The gene expression profile data showed that AKT1 might reverse both the CXCR2-mediated EMT and apoptosis associated gene expression (Fig. 4F). Western blot analysis confirmed that AKT1 could abrogate the CXCR2-induced alteration of the EMT-related markers including E-cadherin, b-catenin, caveolin-1 and the apoptosis-associated proteins such as Bax, Bak, Bid, Bcl-2, Xiap at the protein levels (Fig. 4G). These data suggest that AKT1 suppresses the CXCR2-mediated cell growth, migration, invasion, metastasis, and chemoresistance, which is partially consistent with a previous study [18]. COX2 mediates the CXCR2 signaling to promote breast cancer growth, metastasis and chemoresistance We found that COX2 was positively regulated by CXCR2 in breast cancer cell lines (Fig. 5A) and differently expressed in breast cancer cell lines (SFig. 3A). And we found that COX2 promoted cell growth, tumor growth, cell migration, invasion and metastasis (SFig. 3BeH), which might be resulted from the altered expression of P85a, AKT1, E-cadherin, b-catenin, and pAKT (Ser473) (Fig. 5B). Moreover, overexpression of COX2 rescued the tumorigenic and metastatic abilities of the CXCR2 shRNA-treated cells in animals (Fig. 5C and D). These data suggest that COX2 may control breast cancer metastasis at the downstream signaling of CXCR2. Since gemcitabine is frequently used to treat breast cancer, we also treated COX2 overexpression or silencing cells. We found that COX2 overexpression or repression inhibited or promoted cellular apoptosis induced by gemcitabine treatment (SFig. 4AeC). This result was supported by treatment of the gemcitabine resistant 231/Gem cell line generated from 231 cells in our institute [19] with gemcitabine (Fig. 5E and F). Silencing of COX2 markedly enhanced the sensitivity of 231/Gem cells to gemcitabine treatment (Fig. 5F). COX2 also altered the expression of Bax, Bak, and Bcl-2 (Fig. 5G), and knockdown of COX2 significantly increased sensitivity of tumor bearing animals to gemcitabine treatment (Fig. 5H, upper panel). Overexpression of COX2 in CXCR2 silencing cells restored the chemoresistance of tumors to both gemcitabine and taxol treatment (Fig. 5H, lower panel). Tumor micrometastasis was highly blocked or enhanced by COX2 shRNA or cDNA, although both gemcitabine and taxol treatments also reduced the number of micrometastasis (Fig. 5I and J). These data suggest that COX2 may mediate the CXCR2 signaling to promote breast cancer metastasis and chemoresistance.
75
down-regulated or up-regulated by silencing of CXCR2 (Fig. 6C, upper panel). In addition, overexpression of AKT1 in 231 cells or silencing of AKT1 in 231/CXCR2i cells also induced the senescenceassociated gene expression alteration (Fig. 6C, lower panel). The summary of gene expression profiles is shown in STables 2 and 3. Furthermore, the senescence-associated markers p16, p21, and HP1g (heterochromatin protein gamma) were differentially altered in terms of the CXCR2, COX2 and AKT1 expression levels although some inconsistence was present in different cell lines possibly due to the different genetic backgrounds of these cells (Fig. 6D), suggesting that the CXCR2/AKT1/COX2 axis may regulate cellular senescence. Clinical association of the CXCR2 signaling with patient survivals Immunohistochemical staining was performed with a tissue microarray (TMA) consisting of 250 invasive ductal breast cancer specimens, and 225 patients with definite outcome were analyzed. As shown in STable 4, patient characteristics were analyzed in the CXCR2 low-expression group against the highexpression one. The distribution of age, menstrual status, T stage, LNM, anatomic stage, histological grade, ER status, PR status, HER2 status, molecular subtype, and E-cadherin expression level exhibited no statistical significance between CXCR2 low- and high-expression group, while the correlation between the other characteristics and CXCR2 expression were further evaluated. In normal breast tissues, CXCR2 and COX2 were undetectable, while moderate expression of AKT1 and P85a or strong expression of b-catenin and E-cadherin were detected (Fig. 7A). In cancer tissues, a positive correlation between the expression of CXCR2 and the expression of COX2 or negative correlations between the expression of CXCR2 and the expression of P85a, AKT1, or b-catenin were conceived (Fig. 7BeF). High expression of CXCR2 or COX2 predicted poor outcomes for DFS (disease free survival) in all breast cancer patients, compared with low-expression group by KaplaneMeier analysis (Fig. 7G, Table 1). Better survivals for DFS were found in patients with high expression of AKT1C, AKT1N, P85aN, or E-cadherin, compared with low expression of each marker (Fig. 7H and I, Table 1). Cox multivariate analysis illustrated that the CXCR2 expression (P ¼ 0.005, HR 5.445, 95%CI 1.670e17.753) was an independent predictor of disease free survival (Table 1). So were the expression of COX2 (P ¼ 0.004, HR 5.753, 95%CI 1.755e18.867), the expression of AKT1C (P ¼ 0.026, HR 0.473, 95%CI 0.245e0.915), and the expression of P85aN expression (P ¼ 0.029, HR 0.503, 95%CI 0.271e0.934) (Table 1). High expression of CXCR2 or COX2 predicted poor outcomes for OS in all breast cancer patients, compared with low-expression group by KaplaneMeier analysis (Fig. 7G, Table 1). Better survivals for OS were found in patients with high expression of AKT1C, AKT1N, or E-cadherin, compared with low expression of each marker (Fig. 7H and I, Table 1). Cox multivariate analysis illustrated that regional lymph node metastasis (P ¼ 0.019, HR 3.287, 95%CI 1.220e8.856), or AKT1N expression (P ¼ 0.047, HR 0.413, 95%CI 0.173e0.987) was an independent predictor of overall survival (Table 1).
AKT1 and COX2 inversely regulate the CXCR2-associated antisenescence
Discussion
CXCR2 is reported to regulate cellular senescence, so we tested the cellular senescence in the CXCR2-mediated signaling. We found that cell senescence was repressed by CXCR2/COX2 overexpression, but enhanced by AKT1 overexpression (Fig. 6A and B). Analysis of the gene expression profiles showed that the mRNA levels of genes associated with anti-senescence or pro-senescence were either
It has been well-established that the cytokines IL-8 and Gro-1 (CXCL1) and their cognate receptor CXCR2 mediate breast cancer initiation, development, and chemoresistance via multiple signal pathways [20e22]. Therefore, targeting the IL-8 or Gro-1/CXCR2 signaling has been considered effective in repression of breast cancer tumor growth, angiogenesis, lung metastasis, and stem cell
76
H. Xu et al. / Cancer Letters 412 (2018) 69e80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
77
Fig. 6. AKT1 and COX2 inversely regulate the CXCR2-associated anti-senescence. A, Detection of cell senescence by senescence-associated b-gal staining in cells expressing the different combinations of CXCR2 cDNA/shRNA, COX2 cDNA/shRNA and AKT1 cDNA/shRNA, and their control cells. B, Quantitative analysis of senescent cells in Fig. A. C, Gene expression profiles of senescence-associated molecules in 231/Scr, 231/CXCR2i, 231HM/Scr, 231HM/CXCR2i, 231/Vector, 231/AKT1, 231/CXCR2i, and 231/CXCR2i/AKT1i cell lines. The green and red colors represent genes down-regulated and up-regulated by silencing of CXCR2 and overexpression or silencing of AKT1, respectively. D, Detection of p16, p21 and HP1-g by Western blotting. b-actin was used as the loading control. All error bars ¼ 95% CIs. *P < 0.05, ** P < 0.01.
activity, and in enhancement of HER2-associated therapy or chemotherapeutic response in breast cancer patients [6,23]. We show here that the phosphorylated AKT and COX2 might be two major mediators in the CXCR2 signaling to promote breast cancer cell EMT, anti-apoptosis, and anti-senescence. However, the nonphosphorylated AKT1 abrogated EMT but activated cellular apoptosis and senescence. Thus, a mechanism that AKT1 and pAKT inversely regulate breast cancer chemoresistance and metastasis has been uncovered in association with the CXCR2 signaling. Although AKT has been reported to promote breast cancer tumorigenesis and metastasis in mouse mammary tumor [24,25], a different study revealed that the hyperactivation of AKT1 inhibits breast cancer cell migration, invasion and metastasis [26]. In our study, the up-regulation of pAKT (Ser473) or COX2 by CXCR2 facilitated EMT, anti-apoptosis, and anti-senescence, whereas the up-regulation of AKT1 at the basal level blocked these activities, which subsequently attenuated chemoresistance and metastasis. Therefore, the regulation of the balanced expression between AKT
and pAKT (Ser473) is essential for breast cancer chemoresistance and metastasis. It has been well-documented that either CXCR2, AKT, or COX2 is associated with cellular apoptosis, which is also supported by our data. However, our data also suggested that cellular senescence could be inhibited by the high level of CXCR2 or COX2, but be induced by the high expression of AKT1. A study from Acosta et al. suggested that in normal human primary cells, overexpression of CXCR2 induced premature senescence, while knockdown of CXCR2 extended the lifespan of cultured cells through the p53, NF-kB, or C/EBPb associated pathways [27], which is consistent with our previous finding [28]. However, more studies have documented that CXCR2 in cancer cells is an oncogenic receptor, which is highly detected in multiple cancers as a poor prognostic marker [29e32]. Therefore, we infer that the CXCR2-associated signaling may function to promote cellular senescence of normal/preneoplastic cells, but to inhibit senescence of cancer cells. We show that AKT1 overexpression induced
Fig. 5. COX2 mediates the CXCR2 signaling to promote breast cancer growth, metastasis and chemoresistance. A, Analysis of COX2 in cell lines expressing CXCR2 cDNA, shRNA, or control vectors. B, Immunoblotting analysis of PI3K (P85a), AKT, pAKT (Ser473), AKT1, and EMT-associated proteins. b-actin was used as the loading control. C, Fat-pad tumor growth from mice injected with cells expressing CXCR2 shRNA/COX2 cDNA or their control cells. D, Quantitative analysis of tumor micrometastasis in lung tissues of xenograft animals. E, Immunoblotting analysis of COX2 in 231, 231/Gem and 231/Gem/COX2i cell lines. F, Gemcitabine-induced apoptosis detected by flow cytometry (left panel). Quantitative analysis of apoptotic cells induced by gemcitabine (right panel). G, Immunoblotting analysis of apoptosis-associated proteins. b-actin was used as the loading control. H, Tumor growth of animals injected with 231HM/Scr, 231HM/COX2i, and 231HM/CXCR2i/COX2 cells and treatment with gemcitabine, taxol, or placebo. I, H&E staining of lung tissues from the animals treated with gemcitabine, taxol, or placebo. Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. J, Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
78
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 7. Clinical associations of the CXCR2 signaling with patient survivals. A, Immunostaining of representative normal human breast tissues showing low levels of CXCR2 and COX2, moderate expression of AKT1 and P85a, and high expression of b-catenin and E-cadherin in epithelial cells. BeF, Immunostaining of breast cancer tissues indicating a positive correlation between CXCR2 and COX2 (B), but a negative correlation between CXCR2 and P85a (C), AKT1(D), E-cadherin (E), or b-catenin (F). G, Poor disease-free survival (DFS) and overall survival (OS) were associated with strong staining of CXCR2 (P ¼ 0.003 and P ¼ 0.007) or COX2 (P ¼ 0.002 and P ¼ 0.018). H, Favorable DFS and OS were associated with the
H. Xu et al. / Cancer Letters 412 (2018) 69e80
79
Table 1 Univariate and multivariate analyses of predictors of overall survival and disease free survival in breast cancer patients. Variables
Disease free survival
Overall survival
Recurrence
Age <50 50 Menstrual status No Yes T stage T1 & T2 T3 & T4 LNM Negative Positive Anatomic stage Phase I & II Phase III Histological grade Grade 1 & 2 Grade 3 ER Negative Positive PR Negative Positive HER2 Negative Positive CXCR2 Low expression High expression COX2 Low expression High expression AKT1C Low expression High expression AKT1N Low expression High expression E-cadherin Low expression High expression P85 aC Low expression High expression P85 aN Low expression High expression b-catenin Low expression High expression
Univariable analysis
Multivariable analysis HR (95% CI)
Yes
No
N
%
Pa
26 27
82 90
108 117
48.0 52.0
18 35
69 103
87 138
46 6
165 6
24 29
Death
Multivariable analysis HR (95% CI)
Pb
Yes
No
N
%
Pa
0.809
10 17
98 100
108 117
48.0 52.0
0.233
38.7 61.3
0.489
6 21
81 117
87 138
38.7 61.3
0.064
211 12
94.6 5.4
0.007
2.403 (0.973, 5.932)
0.057
22 5
189 7
211 12
94.6 5.4
<0.001
2.302 (0.743, 7.137)
0.149
114 58
138 87
61.3 38.7
0.006
1.847 (0.953, 3.579)
0.069
8 19
130 68
138 87
61.3 38.7
<0.001
3.287 (1.220, 8.856)
0.019
39 13
148 23
187 36
83.9 16.1
0.035
1.120 (0.504, 2.485)
0.781
16 11
171 25
187 36
83.9 16.1
<0.001
1.616 (0.571, 4.573)
0.366
28 19
104 38
132 57
69.8 30.2
0.052
13 12
119 45
132 57
69.8 30.2
0.031
1.695 (0.729, 3.942)
0.221
33 20
100 71
133 91
59.4 40.6
0.464
18 9
115 82
133 91
59.4 40.6
0.420
45 8
122 47
167 55
75.2 24.8
0.058
24 3
143 52
167 55
75.2 24.8
0.089
33 20
103 67
136 87
61.0 39.0
0.815
20 7
116 80
136 87
61.0 39.0
0.297
3 50
42 130
45 180
20.0 80.0
0.003
5.445 (1.670, 17.753)
0.005
0 27
45 153
45 180
20.0 80.0
0.007
4.270E þ 05 (5.929E-211, 3.076E þ 221)
0.959
3 50
46 126
49 176
21.8 78.2
0.002
5.753 (1.755, 18.867)
0.004
1 26
48 150
49 176
21.8 78.2
0.018
6.667 (0.889, 50.024)
0.065
35 18
86 86
121 104
53.8 46.2
0.029
0.473 (0.245, 0.915)
0.026
20 7
101 97
121 104
53.8 46.2
0.026
0.426 (0.165, 1.098)
0.077
29 24
69 103
98 127
43.6 56.4
0.039
0.624 (0.326, 1.197)
0.156
17 10
81 117
98 127
43.6 56.4
0.03
0.413 (0.173, 0.987)
0.047
34 19
78 94
112 113
49.8 50.2
0.017
0.599 (0.328, 1.094)
0.095
19 8
93 105
112 113
49.8 50.2
0.026
0.668 (0.273, 1.632)
0.376
24 29
54 118
78 147
34.7 65.3
0.086
13 14
65 133
78 147
34.7 65.3
0.125
35 18
85 87
120 105
53.3 46.7
0.037
19 8
101 97
120 105
53.3 46.7
0.065
15 38
38 134
53 172
23.6 76.4
0.441
10 17
43 155
53 172
23.6 76.4
0.089
0.503 (0.271, 0.934)
Pb
Univariable analysis
0.029
a and b indicate the P values derived from univariable analysis and multivariable analysis, respectively. Numbers in bold indicate the statistical significance of P values (<0.05).
cancer cell senescence, which has been documented by numerous reports [33e35]. Although no significant association of COX2 with cell senescence has been reported in human cancers, COX2 may indirectly manipulate cell aging through down-regulation of senescence prone molecules including p16, p21, and HP-1g according to our study. Moreover, we found that the examined molecules were significantly associated with patient survivals in all enrolled breast cancer tissues, and most of them could also predict outcomes for
breast cancer patients. While favorable outcomes were significantly conceived in patients with high expression of AKT1, P85a or Ecadherin, poor prognosis was found in patients with high expression of CXCR2 or COX2. Thus, specific inhibitors against CXCR2 or COX2 may effectively boost the chemotherapeutic efficacy of breast cancer. However, traditional therapy against the PI3K/AKT signaling may have dilemmas if used for breast cancer treatment because inhibition of AKT1 may induce chemoresistance and metastasis based on our study.
cytoplasmic staining of AKT1 (P ¼ 0.029 and P ¼ 0.026), the accumulated nucleus staining of AKT1 (P ¼ 0.039 and P ¼ 0.030). I, Favorable DFS was associated with the nucleus staining of P85a (P ¼ 0.037), and favorable DFS and OS were associated with the strong staining of E-cadherin (P ¼ 0.017 and P ¼ 0.026). J, Correlation of CXCR2 expression with other molecules in TMA.
80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Funding This study was supported by grants from the National Natural Science Foundation of China (No. 81372797, 81572553, and 81772789 for G. Yang); by the Shanghai Pujiang Program (11PJ1402200) from the Shanghai Municipal Government of China for G. Yang.
[16]
[17]
[18]
Conflict of interests [19]
The authors declare that they have no competing interests. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.canlet.2017.09.030. References [1] A. Muller, B. Homey, H. Soto, N. Ge, D. Catron, M.E. Buchanan, et al., Involvement of chemokine receptors in breast cancer metastasis, Nature 410 (2001) 50e56. [2] A.E. Karnoub, R.A. Weinberg, Chemokine networks and breast cancer metastasis, Breast Dis. 26 (2006) 75e85. [3] A. Britschgi, R. Andraos, H. Brinkhaus, I. Klebba, V. Romanet, U. Muller, et al., JAK2/STAT5 inhibition circumvents resistance to PI3K/mTOR blockade: a rationale for cotargeting these pathways in metastatic breast cancer, Cancer cell 22 (2012) 796e811. [4] K.C. Nannuru, B. Sharma, M.L. Varney, R.K. Singh, Role of chemokine receptor CXCR2 expression in mammary tumor growth, angiogenesis and metastasis, J. Carcinog. 10 (2011) 40. [5] J.L. Halpern, A. Kilbarger, C.C. Lynch, Mesenchymal stem cells promote mammary cancer cell migration in vitro via the CXCR2 receptor, Cancer Lett. 308 (2011) 91e99. [6] J.K. Singh, G. Farnie, N.J. Bundred, B.M. Simoes, A. Shergill, G. Landberg, et al., Targeting CXCR1/2 significantly reduces breast cancer stem cell activity and increases the efficacy of inhibiting HER2 via HER2-dependent and -independent mechanisms, Clin. cancer Res. official J. Am. Assoc. Cancer Res. 19 (2013) 643e656. [7] G. Stassi, M. Garofalo, M. Zerilli, L. Ricci-Vitiani, C. Zanca, M. Todaro, et al., PED mediates AKT-dependent chemoresistance in human breast cancer cells, Cancer Res. 65 (2005) 6668e6675. [8] M. Qiao, J.D. Iglehart, A.B. Pardee, Metastatic potential of 21T human breast cancer cells depends on Akt/protein kinase B activation, Cancer Res. 67 (2007) 5293e5299. [9] S.X. Yang, J.P. Costantino, C. Kim, E.P. Mamounas, D. Nguyen, J.H. Jeong, et al., Akt phosphorylation at Ser473 predicts benefit of paclitaxel chemotherapy in node-positive breast cancer, J. Clin. Oncol. official J. Am. Soc. Clin. Oncol. 28 (2010) 2974e2981. [10] K. Hinohara, S. Kobayashi, H. Kanauchi, S. Shimizu, K. Nishioka, E. Tsuji, et al., ErbB receptor tyrosine kinase/NF-kappaB signaling controls mammosphere formation in human breast cancer, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 6584e6589. [11] D.W. Visscher, V.S. Pankratz, M. Santisteban, C. Reynolds, A. Ristimaki, R.A. Vierkant, et al., Association between cyclooxygenase-2 expression in atypical hyperplasia and risk of breast cancer, J. Natl. Cancer Inst. 100 (2008) 421e427. [12] T.R. Lyons, J. O'Brien, V.F. Borges, M.W. Conklin, P.J. Keely, K.W. Eliceiri, et al., Postpartum mammary gland involution drives progression of ductal carcinoma in situ through collagen and COX-2, Nat. Med. 17 (2011) 1109e1115. [13] M. Hu, G. Peluffo, H. Chen, R. Gelman, S. Schnitt, K. Polyak, Role of COX-2 in epithelial-stromal cell interactions and progression of ductal carcinoma in situ of the breast, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 3372e3377. [14] B. Singh, J.A. Berry, A. Shoher, G.D. Ayers, C. Wei, A. Lucci, COX-2 involvement in breast cancer metastasis to bone, Oncogene 26 (2007) 3789e3796. [15] M.C. Zatelli, D. Mole, F. Tagliati, M. Minoia, M.R. Ambrosio, E. degli Uberti, Cyclo-oxygenase 2 modulates chemoresistance in breast cancer cells
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33] [34]
[35]
involving NF-kappaB, Cell. Oncol. Official J. Int. Soc. Cell. Oncol. 31 (2009) 457e465. A.M. Simeone, R. Nieves-Alicea, V.C. McMurtry, S. Colella, R. Krahe, A.M. Tari, Cyclooxygenase-2 uses the protein kinase C/interleukin-8/urokinase-type plasminogen activator pathway to increase the invasiveness of breast cancer cells, Int. J. Oncol. 30 (2007) 785e792. B. Singh, J.A. Berry, L.E. Vincent, A. Lucci, Involvement of IL-8 in COX-2mediated bone metastases from breast cancer, J. Surg. Res. 134 (2006) 44e51. M. Yoeli-Lerner, G.K. Yiu, I. Rabinovitz, P. Erhardt, S. Jauliac, A. Toker, Akt blocks breast cancer cell motility and invasion through the transcription factor NFAT, Mol. Cell 20 (2005) 539e550. X.L. Yang, F.J. Lin, Y.J. Guo, Z.M. Shao, Z.L. Ou, Gemcitabine resistance in breast cancer cells regulated by PI3K/AKT-mediated cellular proliferation exerts negative feedback via the MEK/MAPK and mTOR pathways, Onco Targets Ther. 7 (2014) 1033e1042. R.I. Fernando, M.D. Castillo, M. Litzinger, D.H. Hamilton, C. Palena, IL-8 signaling plays a critical role in the epithelial-mesenchymal transition of human carcinoma cells, Cancer Res. 71 (2011) 5296e5306. R. Divella, A. Daniele, E. Savino, F. Palma, A. Bellizzi, F. Giotta, et al., Circulating levels of transforming growth factor-betaeta (TGF-beta) and chemokine (C-XC motif) ligand-1 (CXCL1) as predictors of distant seeding of circulating tumor cells in patients with metastatic breast cancer, Anticancer Res. 33 (2013) 1491e1497. B. Sharma, K.C. Nannuru, M.L. Varney, R.K. Singh, Host Cxcr2-dependent regulation of mammary tumor growth and metastasis, Clin. Exp. Metastasis 32 (2015) 65e72. B. Sharma, D.M. Nawandar, K.C. Nannuru, M.L. Varney, R.K. Singh, Targeting CXCR2 enhances chemotherapeutic response, inhibits mammary tumor growth, angiogenesis, and lung metastasis, Mol. Cancer Ther. 12 (2013) 799e808. X. Ju, S. Katiyar, C. Wang, M. Liu, X. Jiao, S. Li, et al., Akt1 governs breast cancer progression in vivo, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 7438e7443. I.G. Maroulakou, W. Oemler, S.P. Naber, P.N. Tsichlis, Akt1 ablation inhibits, whereas Akt2 ablation accelerates, the development of mammary adenocarcinomas in mouse mammary tumor virus (MMTV)-ErbB2/neu and MMTVpolyoma middle T transgenic mice, Cancer Res. 67 (2007) 167e177. H. Liu, D.C. Radisky, C.M. Nelson, H. Zhang, J.E. Fata, R.A. Roth, et al., Mechanism of Akt1 inhibition of breast cancer cell invasion reveals a protumorigenic role for TSC2, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 4134e4139. J.C. Acosta, A. O'Loghlen, A. Banito, M.V. Guijarro, A. Augert, S. Raguz, et al., Gil, Chemokine signaling via the CXCR2 receptor reinforces senescence, Cell 133 (2008) 1006e1018. G. Yang, D.G. Rosen, Z. Zhang, R.C. Bast Jr., G.B. Mills, J.A. Colacino, et al., The chemokine growth-regulated oncogene 1 (Gro-1) links RAS signaling to the senescence of stromal fibroblasts and ovarian tumorigenesis, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 16472e16477. C. Murphy, M. McGurk, J. Pettigrew, A. Santinelli, R. Mazzucchelli, P.G. Johnston, et al., Nonapical and cytoplasmic expression of interleukin-8, CXCR1, and CXCR2 correlates with cell proliferation and microvessel density in prostate cancer, Clin. cancer Res. Official J. Am. Assoc. Cancer Res. 11 (2005) 4117e4127. G. Yang, D.G. Rosen, G. Liu, F. Yang, X. Guo, X. Xiao, et al., CXCR2 promotes ovarian cancer growth through dysregulated cell cycle, diminished apoptosis, and enhanced angiogenesis, Clin. cancer Res. Official J. Am. Assoc. Cancer Res. 16 (2010) 3875e3886. Y.S. Lee, I. Choi, Y. Ning, N.Y. Kim, V. Khatchadourian, D. Yang, et al., Interleukin-8 and its receptor CXCR2 in the tumour microenvironment promote colon cancer growth, progression and metastasis, Br. J. Cancer 106 (2012) 1833e1841. T. Desurmont, N. Skrypek, A. Duhamel, N. Jonckheere, G. Millet, E. Leteurtre, et al., Overexpression of chemokine receptor CXCR2 and ligand CXCL7 in liver metastases from colon cancer is correlated to shorter disease-free and overall survival, Cancer Sci. 106 (2015) 262e269. M. Wu, B. Wang, J. Fei, N. Santanam, E.R. Blough, Important roles of Akt/PKB signaling in the aging process, Front. Biosci. 2 (2010) 1169e1188. M.V. Astle, K.M. Hannan, P.Y. Ng, R.S. Lee, A.J. George, A.K. Hsu, et al., AKT induces senescence in human cells via mTORC1 and p53 in the absence of DNA damage: implications for targeting mTOR during malignancy, Oncogene 31 (2012) 1949e1962. H. Miyauchi, T. Minamino, K. Tateno, T. Kunieda, H. Toko, I. Komuro, Akt negatively regulates the in vitro lifespan of human endothelial cells via a p53/ p21-dependent pathway, EMBO J. 23 (2004) 212e220.
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
Original Article
CXCR2 promotes breast cancer metastasis and chemoresistance via suppression of AKT1 and activation of COX2 Han Xu a, b, c, 1, Fengjuan Lin a, b, c, 1, Ziliang Wang a, c, Lina Yang a, c, Jiao Meng a, c, Zhouluo Ou b, c, Zhimin Shao a, b, c, **, Genhong Di b, c, ***, Gong Yang a, c, d, * a
Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China Department of Breast Surgery, Fudan University Shanghai Cancer Center, Shanghai, 200032, China Department of Oncology, Shanghai Medical College, Fudan University, 200032, China d Central Laboratory, The Fifth People's Hospital of Shanghai, Fudan University, 200240, China b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 April 2017 Received in revised form 17 September 2017 Accepted 21 September 2017
Metastasis and chemoresistance are two major causes of breast cancer death. We show here that the chemokine receptor CXCR2 was overexpressed in breast cancer cell lines and tissues. CXCR2 promoted anti-apoptosis, anti-senescence, and epithelial-to-mesenchymal transition (EMT) of breast cancer cells, leading to the enhanced metastasis and chemoresistance. Further study suggested that AKT1 and cyclooxygenase-2 (COX2; PTGS2) might mediate the CXCR2 signaling to inversely control the breast cancer metastasis and chemoresistance through the regulation of EMT, apoptosis, and senescence. Analyses of clinical data indicate that the high expression of CXCR2 was correlated with the high expression of COX2 and the low expression of AKT1, P85a, E-cadherin, and b-catenin in cancer tissues. Poor outcomes were associated with the high expression of CXCR2 or COX2 while favorable survivals were associated with the high expression of P85a, AKT1, or E-cadherin in all cancer patients. Cox multivariate analysis demonstrated that CXCR2, COX2, and AKT1 could be independent predictors for disease free survivals. All these data suggest that CXCR2 promotes breast cancer metastasis and chemoresistance via suppressing AKT1 and activating COX2. Thus, antagonists of the CXCR2 signaling molecules may be used to treat breast cancer patients particularly with high metastasis and chemoresistance. © 2017 Elsevier B.V. All rights reserved.
Keywords: CXCR2 AKT1 COX2 Breast cancer Metastasis
Introduction Although the overall survival of breast cancer at early stage is improved by standard chemotherapy following surgery, the number of death is still climbing because of distant metastasis and chemoresistance. Studies have shown that chemokines and their receptors mediate breast cancer metastasis [1,2]. Interleukin 8 (IL8) has been proved to be an essential molecule regulating breast
* Corresponding author. Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. ** Corresponding author. Cancer Institute, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. *** Corresponding author. Department of Breast Surgery, Fudan University Shanghai Cancer Center, Shanghai, 200032, China. E-mail addresses: [email protected] (Z. Shao), [email protected] (G. Di), [email protected] (G. Yang). 1 These authors equally contributed to this work. https://doi.org/10.1016/j.canlet.2017.09.030 0304-3835/© 2017 Elsevier B.V. All rights reserved.
cancer metastasis and chemoresistance. A recent study from Britschgi et al. suggests that IL-8 mediates resistance to the PI3Kpathway-targeted therapy in breast cancer [3]. Some studies have shown that CXCR2, the co-receptor of IL-8 and Gro-a, may enhance breast cancer migration, invasion, and metastasis [4] or mediate breast cancer bone metastasis through promoting the migration of breast mesenchymal stem cells [5]. Inhibition of both CXCR1/2 reduces the activity of breast cancer stem cells and improves the survival of HER2-positive patients in combination with the HER2targeted chemotherapies [6]. However, how the CXCR2 signaling regulates breast cancer metastasis and chemoresistance is largely unclear. Studies have shown that AKT regulates chemoresistance through PED, a broad anti-apoptotic molecule [7], and that the phosphorylation of AKT at Ser 473 promotes breast cancer metastasis [8]. However, a clinical study suggests that breast cancer patients with the phosphorylation of AKT at Ser473 appear to be sensitive to treatment with paclitaxel [9]. These controversial
70
H. Xu et al. / Cancer Letters 412 (2018) 69e80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
results indicate that AKT may have different functions in terms of tumor progression, metastasis and chemoresistance, but the detailed signaling has yet to be defined. A few studies have shown that the PI3K/AKT pathway may be involved in the CXCR2 signaling through its ligands IL-8 and Gro-a [10], but the essential mechanism how CXCR2 modulates breast cancer metastasis and chemoresistance through the PI3K/AKT pathway is unclear. Cyclooxygenase 2 (COX2, also known as the prostaglandin H synthase-2, PGHS-2) is usually up-regulated in many cancers. In breast cancer, COX2 can be used as a biomarker to stratify breast cancer risk in women with atypical hyperplasia [11]. COX2 also promotes progression of ductal carcinoma in situ (DCIS) to invasive breast carcinomas [12], osteolytic bone metastasis of breast cancer [13] and tumor angiogenesis [14]. Overexpression of COX2 confers chemoresistance in breast cancer through PGH2 generation and NF-kB activation [15]. Although a number of studies have reported that COX2 stimulates IL-8 expression to enhance breast cancer invasion [16] or bone metastasis [17], no associated studies between COX2 and CXCR2 have been reported in breast cancer. In this study, we employed breast cancer cell lines, animals, and human specimens to investigate how the CXCR2-associated signaling is involved in breast cancer metastasis and chemoresistance. Materials and methods Cell lines and cell culture Normal breast epithelial cell lines HMEC207, HMEC212, immortalized breast epithelial cell line MCF-10A, low-metastatic breast cancer cell lines MCF-7,T47D, BT20, BT474, BT483, SKBR3, high-metastatic breast cancer cell lines MDA-MB-231 (simplified as 231), MDA-MB-231HM (high lung metastasis, simplified as 231HM), MDA-MB-231BO (high bone metastasis, simplified as 231BO), and the gemcitabineresistant breast cancer cell line MDA-MB-231Gem (simplified as 231Gem) were used in this study. MCF-7, BT474, BT483, SKBR3, 231 and 231BO cell lines were obtained from the American Tissue Culture Collection (ATCC). 231HM and 231Gem cell lines were established in our institute. Cells were routinely maintained in the recommended medium supplemented with 10% fetal bovine serum, 100 mg/mL penicillin, and 100 mg/mL streptomycin. All cultures were incubated at 37 C in a humidified 5% CO2 atmosphere. Drug treatment and apoptosis detection Cells were treated with either taxol at 50 mg/ml, gemcitabine at 4 mM (for cells), Cell apoptosis induced by taxol or gemcitabine was assessed by using flow cytometry after cells were stained with propidium iodide (PI) and Annexin V. Briefly, 1 106 cells seeded in 60 mm dish for 24 h were treated with taxol or gemcitabine for 12 h, and then trypsinized and washed twice in ice-cold phosphate-buffered saline (PBS). A total of 1 105 cells were resuspended in 100 mL binding buffer to which was added 5 mL 2 mg/ml Annexin V and 5 ml 50 mg/ml PI. Following 15 min incubation in dark, the cells were detected by flow cytometry under different channels. All tests were repeated three times. Examination of cellular senescence To detect senescence-associated b-galacatosidase activity, cells were fixed with 2% formaldehyde and 0.2% glutaraldehyde at room temperature for 4 min and then stained with the buffer containing 150 mM NaCI, 20 mM MgCI2,10 mg/ml X-gal, 50 mM Potassium Ferricyanide, 50 mM Potassium Ferrocyanide, and 1.6 mM citric acid (pH 6.0) at 37 C for 2e15 h. Western blot was used to detect the expression of senescence-associated molecules including p16, p21, and HP-1g. Statistical analysis Statistical analysis was performed by t-test at different time points between the mean tumor sizes of each group. The statistical significance of the differences in cell growth, migration, colony formation, tumor volume and the rate of spontaneous metastasis was calculated by using the Student's t-test. The relation between protein
71
expression in tissue arrays and patient pathological characteristics was analyzed by X2 test or Fisher's exact test. Disease-specific survival rates were determined by using the Kaplan-Meier method and the log-rank test from SPSS 18.0 software (SPSS, Inc). The P value smaller than 0.05 was considered statistically significant. All statistical tests were two-sided. Graphs were created with GraphPad Prism 5 and Adobe Photoshop CS6. Plasmid construction and delivery to cells, qPCR, Western blotting, immunofluorescence, cell proliferation, migration, invasion and anchorage-independent assay, Xenograft tumor growth, animal treatment, immunostaining and analysis of animal tissues and human tissue microarray were available in Supplemental Materials and Methods.
Results CXCR2 stimulates breast cancer cell proliferation and tumorigenesis High level of CXCR2 was clearly detected in breast cancer cell lines 231, 231HM, 231BO, SKBR3, and T47D, while low level of CXCR2 was detected in breast cancer cell lines BT20, BT474, BT483, MCF-7, and in immortalized breast epithelial cell line MCF-10A, as well as in normal breast epithelial cell lines HMEC207 and HMEC212. Six of nine (6/9, 66.7%) breast cancer tissues were found with high expression of CXCR2, whereas low CXCR2 was detected in normal human mammary tissues (Fig. 1A and B). Overexpression or silencing of CXCR2 (Fig. 1C and D) significantly promoted or repressed cell proliferation (SFig. 1A), soft agar colony formation (Fig. 1E and F), and the mammary fat pad tumor growth in animals (Fig. 1G, SFig. 1B), compared with controls, although MCF-7 and BT474 cell lines failed to generate tumor burdens in animals without regard to the CXCR2 status. These data suggest that CXCR2 promotes breast cancer cell proliferation and tumorigenesis. CXCR2 promotes breast cancer cell migration, invasion and metastasis via the P85a/AKT1-mediated EMT signaling Overexpression or knockdown of CXCR2 increased or decreased cell migration and invasion in vitro (Fig. 2A and SFig. 2A and B). Moreover, cells expressing CXCR2 shRNA induced fewer number of micrometastasis in lung tissues than those expressing control vector in vivo (Fig. 2B and C). In addition, CXCR2 knockdown also reduced the number of micrometastatic sites in other organs including bone, liver, brain and axillary lymph nodes (STable 1). All these data suggest that CXCR2 promotes breast cancer metastasis. It is well-known that the epithelial to mesenchymal transition (EMT) may play an important role in cancer metastasis. Immunofluorescent staining showed that MCF-7 cells overexpressing CXCR2 became spindle, while 231 cells expressing CXCR2 shRNA appeared round (Fig. 2D), suggesting that CXCR2 might mediate the EMT signaling. Analysis of the gene expression profiles showed that mRNAs of the genes associated with EMT were altered by downregulation of CXCR2 in 231 and 231HM cells, compared with control cells (Fig. 2E). By Western blotting, we found that CXCR2 downregulated the expression of PI3K-P85a, total AKT, and AKT1 (Fig. 2F), but not that of PI3K-P110a, AKT2 and AKT3 (data not shown). However, CXCR2 up-regulated the phosphorylated AKT (Ser473), but not pAKTThr308 (Fig. 2F), suggesting that CXCR2 may promote metastasis through suppressing P85a and AKT1 expression and enhancing the phosphorylation of AKT at Ser473. We also found that the expression of E-cadherin and b-catenin was either
Fig. 1. CXCR2 stimulates breast cancer cell proliferation and tumor growth. A, CXCR2 expression in normal breast epithelial cells, breast cancer cell lines (upper panel), and tissues (lower panel). b-actin was used as a loading control. B, CXCR2 expression in normal breast tissues and breast tumor tissues (left panel, 200; right panel, 400). Tissues were stained with a rabbit anti-CXCR2 antibody and visualized following a donkey anti-rabbit secondary antibody. C-D, Analyses of CXCR2 by Western blot (C) and realtime-PCR (D) in breast cancer cells transfected with CXCR2 cDNA (CXCR2), shRNA(CXCR2i), or control vectors. E-F, In vitro tumorigenicity tested by soft agar assay. G, Fat-pad tumor growth of mice injected with cells expressing either CXCR2i or scrambled shRNA (Scr). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
72
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 2. CXCR2 promotes breast cancer cell migration, invasion and metastasis via the P85a/AKT1 signaling-mediated EMT. A, Quantitative analysis of migration speed by wound healing scratch assay (left panel) and invaded cells by Matrigel cell invasion assay (right panel). B, Detection of tumor micrometastasis by using H&E staining of lung tissues derived from xenograft nude mice injected with breast cancer cells expressing either CXCR2 shRNA or control vectors (left panel, 200; right panel, 400). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. C, Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). D, EMT-like morphological changes of breast cancer cells with CXCR2 overexpression or silencing detected by immunofluorescence (Magnification 400). Scar bars, 10 mm. E, Gene expression profiles of EMT-associated molecules in 231/Scr, 231/CXCR2i, 231HM/Scr and 231HM/CXCR2i cell lines. The green and red colors represent genes down-regulated and up-regulated by CXCR2 silencing, respectively. F, Immunoblotting analysis of PI3K(P85a), PI3K(P110a), pAKT (Ser473), pAKT (Thr 308), AKT and AKT1, E-cadherin, b-catenin and caveolin-1 in breast cancer cells transfected with CXCR2 cDNA or shRNA and their control vectors. G, Immunostaining analyses of lung tissues suggest a slightly increased cytokeratin but decreased vimentin in breast cancer cells with CXCR2 silencing (Magnification 200). H, Immunoblotting analysis of P85a, AKT1 and the metastasis-associated proteins. All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
decreased or increased by CXCR2 overexpression or knock down, and that caveolin-1 was reduced after CXCR2 was silenced (Fig. 2F). Furthermore, we showed that silencing of CXCR2 slightly increased the cytokeratin 8/18 expression, but decreased the expression of vimentin in micrometastatic tumor cells of lung tissues from tumor bearing mice (Fig. 2G). Additionally, by overexpression or silencing of P85a, we found that P85a could up-regulate AKT1, E-cadherin and b-catenin expression, but down-regulate caveolin 1 expression to counteract the CXCR2-mediated signaling (Fig. 2H). These data indicate that CXCR2 may promote breast cancer metastasis mainly through the P85a/AKT1-mediated EMT signaling.
CXCR2 induces breast cancer chemoresistance via the anti-apoptosis and pro-metastasis signaling
or
By flow cytometer, we first found that CXCR2 overexpression silencing reduced or promoted Taxol-induced cellular
apoptosis (Fig. 3A and B). We then performed gene expression profile analysis and found a clear alteration of the p53-associated apoptosis signaling molecules particularly after CXCR2 was disrupted in 231 and 231HM cells (Fig. 3C). The expression of the apoptosis-related proteins including Bax, Bak, Bad, Bid, Bcl-2 and Bcl-xL were confirmed by Western blot (Fig. 3D). Thus, CXCR2 appears to promote breast cancer chemoresistance possibly by suppressing cellular apoptosis. Silencing of CXCR2 also markedly enhanced the sensitivity of xenograft tumors to Taxol treatment, compared with controls (Fig. 3E and F). Although treatment of animals with taxol did not significantly reduce the volume of tumors formed by control cells (231/Scr, 231HM/Scr), the number of micrometastasis in lung tissues from all animals was significantly reduced by taxol treatment. (Fig. 3G). These data suggest that CXCR2 induces breast cancer cell chemoresistance likely through suppression of the p53-mediated apoptosis signaling.
H. Xu et al. / Cancer Letters 412 (2018) 69e80
73
Fig. 3. CXCR2 induces breast cancer chemoresistance via the anti-apoptosis and pro-metastasis signaling. A, Taxol-induced apoptosis detected in CXCR2 overexpressing cell lines (MCF-7 and BT474) and in CXCR2 silencing cell lines (231, 231HM and 231BO) by flow cytometry. B, Quantitative analysis of apoptotic cells induced by taxol. C, Gene expression profiles of apoptosis molecules in 231/Scr, 231/CXCR2i, 231HM/Scr and 231HM/CXCR2i cell lines. The green and red colors represent genes down-regulated and up-regulated by CXCR2 silencing, respectively. D, Apoptosis-related proteins detected by Western blotting. b-actin was used as a loading control. E-F, Tumor growth following subcutaneous implantation of control and CXCR2 knockdown cells in vivo treated with taxol or placebo. G, Detection of tumor micrometastasis in lung tissues of xenograft nude mice treated with taxol (left panel). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed) (right panel). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
74
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 4. AKT1 counteracts CXCR2 to suppress breast cancer tumorigenesis, metastasis and chemoresistance. A, Fat-pad tumor growth from nude mice injected with cells expressing 231/CXCR2i/Scr or 231/CXCR2i/AKT1i. BeC, Quantitative analysis of migration and invasion with the migration index and the number of invaded cells. D, Detection of tumor micrometastasis in lung tissues of nude mice injected with 231/CXCR2i/AKT1i cells and control cells (left panel, 200; right panel, 400). Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). E, Taxol-induced apoptosis detected by flow cytometry in cell lines (left panel). Quantitative analysis of apoptotic cells (right panel). F, Gene expression profiles of metastasis and apoptosis molecules in 231/Scr, 231/CXCR2i, 231/AKT1 and 231/CXCR2i/AKT1i cell lines. The green and red colors represent genes downregulated and up-regulated by AKT1 overexpression and silencing, respectively. G, Immunoblotting analysis of EMT and apoptosis-associated proteins in MCF-7/CXCR2 cells transfected with AKT1 cDNA or vector (left panel), and in 231/CXCR2i cells transfected with AKT1 shRNA or scramble shRNA (right panel). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
H. Xu et al. / Cancer Letters 412 (2018) 69e80
AKT1 counteracts CXCR2 to suppress breast cancer tumorigenesis, metastasis and chemoresistance To verify the function of AKT1 in the CXCR2-mediated signaling, we delivered AKT1 cDNAs or AKT1 shRNAs into MCF-7/CXCR2 or 231/CXCR2i cells (Fig. 4G) and found that AKT1 counteracted the CXCR2 signaling to block cancer growth/metastasis in vivo (Fig. 4AeD) and to attenuate cell migration/invasion in vitro (Fig. 4B and C, SFig. 2C and D). In addition, treatment of AKT1 overexpressing or silencing cells with taxol induced more or less apoptosis than did treatment of control cells with taxol (Fig. 4E). The gene expression profile data showed that AKT1 might reverse both the CXCR2-mediated EMT and apoptosis associated gene expression (Fig. 4F). Western blot analysis confirmed that AKT1 could abrogate the CXCR2-induced alteration of the EMT-related markers including E-cadherin, b-catenin, caveolin-1 and the apoptosis-associated proteins such as Bax, Bak, Bid, Bcl-2, Xiap at the protein levels (Fig. 4G). These data suggest that AKT1 suppresses the CXCR2-mediated cell growth, migration, invasion, metastasis, and chemoresistance, which is partially consistent with a previous study [18]. COX2 mediates the CXCR2 signaling to promote breast cancer growth, metastasis and chemoresistance We found that COX2 was positively regulated by CXCR2 in breast cancer cell lines (Fig. 5A) and differently expressed in breast cancer cell lines (SFig. 3A). And we found that COX2 promoted cell growth, tumor growth, cell migration, invasion and metastasis (SFig. 3BeH), which might be resulted from the altered expression of P85a, AKT1, E-cadherin, b-catenin, and pAKT (Ser473) (Fig. 5B). Moreover, overexpression of COX2 rescued the tumorigenic and metastatic abilities of the CXCR2 shRNA-treated cells in animals (Fig. 5C and D). These data suggest that COX2 may control breast cancer metastasis at the downstream signaling of CXCR2. Since gemcitabine is frequently used to treat breast cancer, we also treated COX2 overexpression or silencing cells. We found that COX2 overexpression or repression inhibited or promoted cellular apoptosis induced by gemcitabine treatment (SFig. 4AeC). This result was supported by treatment of the gemcitabine resistant 231/Gem cell line generated from 231 cells in our institute [19] with gemcitabine (Fig. 5E and F). Silencing of COX2 markedly enhanced the sensitivity of 231/Gem cells to gemcitabine treatment (Fig. 5F). COX2 also altered the expression of Bax, Bak, and Bcl-2 (Fig. 5G), and knockdown of COX2 significantly increased sensitivity of tumor bearing animals to gemcitabine treatment (Fig. 5H, upper panel). Overexpression of COX2 in CXCR2 silencing cells restored the chemoresistance of tumors to both gemcitabine and taxol treatment (Fig. 5H, lower panel). Tumor micrometastasis was highly blocked or enhanced by COX2 shRNA or cDNA, although both gemcitabine and taxol treatments also reduced the number of micrometastasis (Fig. 5I and J). These data suggest that COX2 may mediate the CXCR2 signaling to promote breast cancer metastasis and chemoresistance.
75
down-regulated or up-regulated by silencing of CXCR2 (Fig. 6C, upper panel). In addition, overexpression of AKT1 in 231 cells or silencing of AKT1 in 231/CXCR2i cells also induced the senescenceassociated gene expression alteration (Fig. 6C, lower panel). The summary of gene expression profiles is shown in STables 2 and 3. Furthermore, the senescence-associated markers p16, p21, and HP1g (heterochromatin protein gamma) were differentially altered in terms of the CXCR2, COX2 and AKT1 expression levels although some inconsistence was present in different cell lines possibly due to the different genetic backgrounds of these cells (Fig. 6D), suggesting that the CXCR2/AKT1/COX2 axis may regulate cellular senescence. Clinical association of the CXCR2 signaling with patient survivals Immunohistochemical staining was performed with a tissue microarray (TMA) consisting of 250 invasive ductal breast cancer specimens, and 225 patients with definite outcome were analyzed. As shown in STable 4, patient characteristics were analyzed in the CXCR2 low-expression group against the highexpression one. The distribution of age, menstrual status, T stage, LNM, anatomic stage, histological grade, ER status, PR status, HER2 status, molecular subtype, and E-cadherin expression level exhibited no statistical significance between CXCR2 low- and high-expression group, while the correlation between the other characteristics and CXCR2 expression were further evaluated. In normal breast tissues, CXCR2 and COX2 were undetectable, while moderate expression of AKT1 and P85a or strong expression of b-catenin and E-cadherin were detected (Fig. 7A). In cancer tissues, a positive correlation between the expression of CXCR2 and the expression of COX2 or negative correlations between the expression of CXCR2 and the expression of P85a, AKT1, or b-catenin were conceived (Fig. 7BeF). High expression of CXCR2 or COX2 predicted poor outcomes for DFS (disease free survival) in all breast cancer patients, compared with low-expression group by KaplaneMeier analysis (Fig. 7G, Table 1). Better survivals for DFS were found in patients with high expression of AKT1C, AKT1N, P85aN, or E-cadherin, compared with low expression of each marker (Fig. 7H and I, Table 1). Cox multivariate analysis illustrated that the CXCR2 expression (P ¼ 0.005, HR 5.445, 95%CI 1.670e17.753) was an independent predictor of disease free survival (Table 1). So were the expression of COX2 (P ¼ 0.004, HR 5.753, 95%CI 1.755e18.867), the expression of AKT1C (P ¼ 0.026, HR 0.473, 95%CI 0.245e0.915), and the expression of P85aN expression (P ¼ 0.029, HR 0.503, 95%CI 0.271e0.934) (Table 1). High expression of CXCR2 or COX2 predicted poor outcomes for OS in all breast cancer patients, compared with low-expression group by KaplaneMeier analysis (Fig. 7G, Table 1). Better survivals for OS were found in patients with high expression of AKT1C, AKT1N, or E-cadherin, compared with low expression of each marker (Fig. 7H and I, Table 1). Cox multivariate analysis illustrated that regional lymph node metastasis (P ¼ 0.019, HR 3.287, 95%CI 1.220e8.856), or AKT1N expression (P ¼ 0.047, HR 0.413, 95%CI 0.173e0.987) was an independent predictor of overall survival (Table 1).
AKT1 and COX2 inversely regulate the CXCR2-associated antisenescence
Discussion
CXCR2 is reported to regulate cellular senescence, so we tested the cellular senescence in the CXCR2-mediated signaling. We found that cell senescence was repressed by CXCR2/COX2 overexpression, but enhanced by AKT1 overexpression (Fig. 6A and B). Analysis of the gene expression profiles showed that the mRNA levels of genes associated with anti-senescence or pro-senescence were either
It has been well-established that the cytokines IL-8 and Gro-1 (CXCL1) and their cognate receptor CXCR2 mediate breast cancer initiation, development, and chemoresistance via multiple signal pathways [20e22]. Therefore, targeting the IL-8 or Gro-1/CXCR2 signaling has been considered effective in repression of breast cancer tumor growth, angiogenesis, lung metastasis, and stem cell
76
H. Xu et al. / Cancer Letters 412 (2018) 69e80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
77
Fig. 6. AKT1 and COX2 inversely regulate the CXCR2-associated anti-senescence. A, Detection of cell senescence by senescence-associated b-gal staining in cells expressing the different combinations of CXCR2 cDNA/shRNA, COX2 cDNA/shRNA and AKT1 cDNA/shRNA, and their control cells. B, Quantitative analysis of senescent cells in Fig. A. C, Gene expression profiles of senescence-associated molecules in 231/Scr, 231/CXCR2i, 231HM/Scr, 231HM/CXCR2i, 231/Vector, 231/AKT1, 231/CXCR2i, and 231/CXCR2i/AKT1i cell lines. The green and red colors represent genes down-regulated and up-regulated by silencing of CXCR2 and overexpression or silencing of AKT1, respectively. D, Detection of p16, p21 and HP1-g by Western blotting. b-actin was used as the loading control. All error bars ¼ 95% CIs. *P < 0.05, ** P < 0.01.
activity, and in enhancement of HER2-associated therapy or chemotherapeutic response in breast cancer patients [6,23]. We show here that the phosphorylated AKT and COX2 might be two major mediators in the CXCR2 signaling to promote breast cancer cell EMT, anti-apoptosis, and anti-senescence. However, the nonphosphorylated AKT1 abrogated EMT but activated cellular apoptosis and senescence. Thus, a mechanism that AKT1 and pAKT inversely regulate breast cancer chemoresistance and metastasis has been uncovered in association with the CXCR2 signaling. Although AKT has been reported to promote breast cancer tumorigenesis and metastasis in mouse mammary tumor [24,25], a different study revealed that the hyperactivation of AKT1 inhibits breast cancer cell migration, invasion and metastasis [26]. In our study, the up-regulation of pAKT (Ser473) or COX2 by CXCR2 facilitated EMT, anti-apoptosis, and anti-senescence, whereas the up-regulation of AKT1 at the basal level blocked these activities, which subsequently attenuated chemoresistance and metastasis. Therefore, the regulation of the balanced expression between AKT
and pAKT (Ser473) is essential for breast cancer chemoresistance and metastasis. It has been well-documented that either CXCR2, AKT, or COX2 is associated with cellular apoptosis, which is also supported by our data. However, our data also suggested that cellular senescence could be inhibited by the high level of CXCR2 or COX2, but be induced by the high expression of AKT1. A study from Acosta et al. suggested that in normal human primary cells, overexpression of CXCR2 induced premature senescence, while knockdown of CXCR2 extended the lifespan of cultured cells through the p53, NF-kB, or C/EBPb associated pathways [27], which is consistent with our previous finding [28]. However, more studies have documented that CXCR2 in cancer cells is an oncogenic receptor, which is highly detected in multiple cancers as a poor prognostic marker [29e32]. Therefore, we infer that the CXCR2-associated signaling may function to promote cellular senescence of normal/preneoplastic cells, but to inhibit senescence of cancer cells. We show that AKT1 overexpression induced
Fig. 5. COX2 mediates the CXCR2 signaling to promote breast cancer growth, metastasis and chemoresistance. A, Analysis of COX2 in cell lines expressing CXCR2 cDNA, shRNA, or control vectors. B, Immunoblotting analysis of PI3K (P85a), AKT, pAKT (Ser473), AKT1, and EMT-associated proteins. b-actin was used as the loading control. C, Fat-pad tumor growth from mice injected with cells expressing CXCR2 shRNA/COX2 cDNA or their control cells. D, Quantitative analysis of tumor micrometastasis in lung tissues of xenograft animals. E, Immunoblotting analysis of COX2 in 231, 231/Gem and 231/Gem/COX2i cell lines. F, Gemcitabine-induced apoptosis detected by flow cytometry (left panel). Quantitative analysis of apoptotic cells induced by gemcitabine (right panel). G, Immunoblotting analysis of apoptosis-associated proteins. b-actin was used as the loading control. H, Tumor growth of animals injected with 231HM/Scr, 231HM/COX2i, and 231HM/CXCR2i/COX2 cells and treatment with gemcitabine, taxol, or placebo. I, H&E staining of lung tissues from the animals treated with gemcitabine, taxol, or placebo. Yellow arrows and circled dash lines indicate the micronodules formed by metastasis of human breast cancer cells. J, Number of micrometastasis indicates the average number of micronodules per slide based on analysis of total 25 slides from 5 different positions of lung tissues of each animal (total 5 xenograft animals were analyzed). All error bars ¼ 95% CIs. *P < 0.05, **P < 0.01.
78
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Fig. 7. Clinical associations of the CXCR2 signaling with patient survivals. A, Immunostaining of representative normal human breast tissues showing low levels of CXCR2 and COX2, moderate expression of AKT1 and P85a, and high expression of b-catenin and E-cadherin in epithelial cells. BeF, Immunostaining of breast cancer tissues indicating a positive correlation between CXCR2 and COX2 (B), but a negative correlation between CXCR2 and P85a (C), AKT1(D), E-cadherin (E), or b-catenin (F). G, Poor disease-free survival (DFS) and overall survival (OS) were associated with strong staining of CXCR2 (P ¼ 0.003 and P ¼ 0.007) or COX2 (P ¼ 0.002 and P ¼ 0.018). H, Favorable DFS and OS were associated with the
H. Xu et al. / Cancer Letters 412 (2018) 69e80
79
Table 1 Univariate and multivariate analyses of predictors of overall survival and disease free survival in breast cancer patients. Variables
Disease free survival
Overall survival
Recurrence
Age <50 50 Menstrual status No Yes T stage T1 & T2 T3 & T4 LNM Negative Positive Anatomic stage Phase I & II Phase III Histological grade Grade 1 & 2 Grade 3 ER Negative Positive PR Negative Positive HER2 Negative Positive CXCR2 Low expression High expression COX2 Low expression High expression AKT1C Low expression High expression AKT1N Low expression High expression E-cadherin Low expression High expression P85 aC Low expression High expression P85 aN Low expression High expression b-catenin Low expression High expression
Univariable analysis
Multivariable analysis HR (95% CI)
Yes
No
N
%
Pa
26 27
82 90
108 117
48.0 52.0
18 35
69 103
87 138
46 6
165 6
24 29
Death
Multivariable analysis HR (95% CI)
Pb
Yes
No
N
%
Pa
0.809
10 17
98 100
108 117
48.0 52.0
0.233
38.7 61.3
0.489
6 21
81 117
87 138
38.7 61.3
0.064
211 12
94.6 5.4
0.007
2.403 (0.973, 5.932)
0.057
22 5
189 7
211 12
94.6 5.4
<0.001
2.302 (0.743, 7.137)
0.149
114 58
138 87
61.3 38.7
0.006
1.847 (0.953, 3.579)
0.069
8 19
130 68
138 87
61.3 38.7
<0.001
3.287 (1.220, 8.856)
0.019
39 13
148 23
187 36
83.9 16.1
0.035
1.120 (0.504, 2.485)
0.781
16 11
171 25
187 36
83.9 16.1
<0.001
1.616 (0.571, 4.573)
0.366
28 19
104 38
132 57
69.8 30.2
0.052
13 12
119 45
132 57
69.8 30.2
0.031
1.695 (0.729, 3.942)
0.221
33 20
100 71
133 91
59.4 40.6
0.464
18 9
115 82
133 91
59.4 40.6
0.420
45 8
122 47
167 55
75.2 24.8
0.058
24 3
143 52
167 55
75.2 24.8
0.089
33 20
103 67
136 87
61.0 39.0
0.815
20 7
116 80
136 87
61.0 39.0
0.297
3 50
42 130
45 180
20.0 80.0
0.003
5.445 (1.670, 17.753)
0.005
0 27
45 153
45 180
20.0 80.0
0.007
4.270E þ 05 (5.929E-211, 3.076E þ 221)
0.959
3 50
46 126
49 176
21.8 78.2
0.002
5.753 (1.755, 18.867)
0.004
1 26
48 150
49 176
21.8 78.2
0.018
6.667 (0.889, 50.024)
0.065
35 18
86 86
121 104
53.8 46.2
0.029
0.473 (0.245, 0.915)
0.026
20 7
101 97
121 104
53.8 46.2
0.026
0.426 (0.165, 1.098)
0.077
29 24
69 103
98 127
43.6 56.4
0.039
0.624 (0.326, 1.197)
0.156
17 10
81 117
98 127
43.6 56.4
0.03
0.413 (0.173, 0.987)
0.047
34 19
78 94
112 113
49.8 50.2
0.017
0.599 (0.328, 1.094)
0.095
19 8
93 105
112 113
49.8 50.2
0.026
0.668 (0.273, 1.632)
0.376
24 29
54 118
78 147
34.7 65.3
0.086
13 14
65 133
78 147
34.7 65.3
0.125
35 18
85 87
120 105
53.3 46.7
0.037
19 8
101 97
120 105
53.3 46.7
0.065
15 38
38 134
53 172
23.6 76.4
0.441
10 17
43 155
53 172
23.6 76.4
0.089
0.503 (0.271, 0.934)
Pb
Univariable analysis
0.029
a and b indicate the P values derived from univariable analysis and multivariable analysis, respectively. Numbers in bold indicate the statistical significance of P values (<0.05).
cancer cell senescence, which has been documented by numerous reports [33e35]. Although no significant association of COX2 with cell senescence has been reported in human cancers, COX2 may indirectly manipulate cell aging through down-regulation of senescence prone molecules including p16, p21, and HP-1g according to our study. Moreover, we found that the examined molecules were significantly associated with patient survivals in all enrolled breast cancer tissues, and most of them could also predict outcomes for
breast cancer patients. While favorable outcomes were significantly conceived in patients with high expression of AKT1, P85a or Ecadherin, poor prognosis was found in patients with high expression of CXCR2 or COX2. Thus, specific inhibitors against CXCR2 or COX2 may effectively boost the chemotherapeutic efficacy of breast cancer. However, traditional therapy against the PI3K/AKT signaling may have dilemmas if used for breast cancer treatment because inhibition of AKT1 may induce chemoresistance and metastasis based on our study.
cytoplasmic staining of AKT1 (P ¼ 0.029 and P ¼ 0.026), the accumulated nucleus staining of AKT1 (P ¼ 0.039 and P ¼ 0.030). I, Favorable DFS was associated with the nucleus staining of P85a (P ¼ 0.037), and favorable DFS and OS were associated with the strong staining of E-cadherin (P ¼ 0.017 and P ¼ 0.026). J, Correlation of CXCR2 expression with other molecules in TMA.
80
H. Xu et al. / Cancer Letters 412 (2018) 69e80
Funding This study was supported by grants from the National Natural Science Foundation of China (No. 81372797, 81572553, and 81772789 for G. Yang); by the Shanghai Pujiang Program (11PJ1402200) from the Shanghai Municipal Government of China for G. Yang.
[16]
[17]
[18]
Conflict of interests [19]
The authors declare that they have no competing interests. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.canlet.2017.09.030. References [1] A. Muller, B. Homey, H. Soto, N. Ge, D. Catron, M.E. Buchanan, et al., Involvement of chemokine receptors in breast cancer metastasis, Nature 410 (2001) 50e56. [2] A.E. Karnoub, R.A. Weinberg, Chemokine networks and breast cancer metastasis, Breast Dis. 26 (2006) 75e85. [3] A. Britschgi, R. Andraos, H. Brinkhaus, I. Klebba, V. Romanet, U. Muller, et al., JAK2/STAT5 inhibition circumvents resistance to PI3K/mTOR blockade: a rationale for cotargeting these pathways in metastatic breast cancer, Cancer cell 22 (2012) 796e811. [4] K.C. Nannuru, B. Sharma, M.L. Varney, R.K. Singh, Role of chemokine receptor CXCR2 expression in mammary tumor growth, angiogenesis and metastasis, J. Carcinog. 10 (2011) 40. [5] J.L. Halpern, A. Kilbarger, C.C. Lynch, Mesenchymal stem cells promote mammary cancer cell migration in vitro via the CXCR2 receptor, Cancer Lett. 308 (2011) 91e99. [6] J.K. Singh, G. Farnie, N.J. Bundred, B.M. Simoes, A. Shergill, G. Landberg, et al., Targeting CXCR1/2 significantly reduces breast cancer stem cell activity and increases the efficacy of inhibiting HER2 via HER2-dependent and -independent mechanisms, Clin. cancer Res. official J. Am. Assoc. Cancer Res. 19 (2013) 643e656. [7] G. Stassi, M. Garofalo, M. Zerilli, L. Ricci-Vitiani, C. Zanca, M. Todaro, et al., PED mediates AKT-dependent chemoresistance in human breast cancer cells, Cancer Res. 65 (2005) 6668e6675. [8] M. Qiao, J.D. Iglehart, A.B. Pardee, Metastatic potential of 21T human breast cancer cells depends on Akt/protein kinase B activation, Cancer Res. 67 (2007) 5293e5299. [9] S.X. Yang, J.P. Costantino, C. Kim, E.P. Mamounas, D. Nguyen, J.H. Jeong, et al., Akt phosphorylation at Ser473 predicts benefit of paclitaxel chemotherapy in node-positive breast cancer, J. Clin. Oncol. official J. Am. Soc. Clin. Oncol. 28 (2010) 2974e2981. [10] K. Hinohara, S. Kobayashi, H. Kanauchi, S. Shimizu, K. Nishioka, E. Tsuji, et al., ErbB receptor tyrosine kinase/NF-kappaB signaling controls mammosphere formation in human breast cancer, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 6584e6589. [11] D.W. Visscher, V.S. Pankratz, M. Santisteban, C. Reynolds, A. Ristimaki, R.A. Vierkant, et al., Association between cyclooxygenase-2 expression in atypical hyperplasia and risk of breast cancer, J. Natl. Cancer Inst. 100 (2008) 421e427. [12] T.R. Lyons, J. O'Brien, V.F. Borges, M.W. Conklin, P.J. Keely, K.W. Eliceiri, et al., Postpartum mammary gland involution drives progression of ductal carcinoma in situ through collagen and COX-2, Nat. Med. 17 (2011) 1109e1115. [13] M. Hu, G. Peluffo, H. Chen, R. Gelman, S. Schnitt, K. Polyak, Role of COX-2 in epithelial-stromal cell interactions and progression of ductal carcinoma in situ of the breast, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 3372e3377. [14] B. Singh, J.A. Berry, A. Shoher, G.D. Ayers, C. Wei, A. Lucci, COX-2 involvement in breast cancer metastasis to bone, Oncogene 26 (2007) 3789e3796. [15] M.C. Zatelli, D. Mole, F. Tagliati, M. Minoia, M.R. Ambrosio, E. degli Uberti, Cyclo-oxygenase 2 modulates chemoresistance in breast cancer cells
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33] [34]
[35]
involving NF-kappaB, Cell. Oncol. Official J. Int. Soc. Cell. Oncol. 31 (2009) 457e465. A.M. Simeone, R. Nieves-Alicea, V.C. McMurtry, S. Colella, R. Krahe, A.M. Tari, Cyclooxygenase-2 uses the protein kinase C/interleukin-8/urokinase-type plasminogen activator pathway to increase the invasiveness of breast cancer cells, Int. J. Oncol. 30 (2007) 785e792. B. Singh, J.A. Berry, L.E. Vincent, A. Lucci, Involvement of IL-8 in COX-2mediated bone metastases from breast cancer, J. Surg. Res. 134 (2006) 44e51. M. Yoeli-Lerner, G.K. Yiu, I. Rabinovitz, P. Erhardt, S. Jauliac, A. Toker, Akt blocks breast cancer cell motility and invasion through the transcription factor NFAT, Mol. Cell 20 (2005) 539e550. X.L. Yang, F.J. Lin, Y.J. Guo, Z.M. Shao, Z.L. Ou, Gemcitabine resistance in breast cancer cells regulated by PI3K/AKT-mediated cellular proliferation exerts negative feedback via the MEK/MAPK and mTOR pathways, Onco Targets Ther. 7 (2014) 1033e1042. R.I. Fernando, M.D. Castillo, M. Litzinger, D.H. Hamilton, C. Palena, IL-8 signaling plays a critical role in the epithelial-mesenchymal transition of human carcinoma cells, Cancer Res. 71 (2011) 5296e5306. R. Divella, A. Daniele, E. Savino, F. Palma, A. Bellizzi, F. Giotta, et al., Circulating levels of transforming growth factor-betaeta (TGF-beta) and chemokine (C-XC motif) ligand-1 (CXCL1) as predictors of distant seeding of circulating tumor cells in patients with metastatic breast cancer, Anticancer Res. 33 (2013) 1491e1497. B. Sharma, K.C. Nannuru, M.L. Varney, R.K. Singh, Host Cxcr2-dependent regulation of mammary tumor growth and metastasis, Clin. Exp. Metastasis 32 (2015) 65e72. B. Sharma, D.M. Nawandar, K.C. Nannuru, M.L. Varney, R.K. Singh, Targeting CXCR2 enhances chemotherapeutic response, inhibits mammary tumor growth, angiogenesis, and lung metastasis, Mol. Cancer Ther. 12 (2013) 799e808. X. Ju, S. Katiyar, C. Wang, M. Liu, X. Jiao, S. Li, et al., Akt1 governs breast cancer progression in vivo, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 7438e7443. I.G. Maroulakou, W. Oemler, S.P. Naber, P.N. Tsichlis, Akt1 ablation inhibits, whereas Akt2 ablation accelerates, the development of mammary adenocarcinomas in mouse mammary tumor virus (MMTV)-ErbB2/neu and MMTVpolyoma middle T transgenic mice, Cancer Res. 67 (2007) 167e177. H. Liu, D.C. Radisky, C.M. Nelson, H. Zhang, J.E. Fata, R.A. Roth, et al., Mechanism of Akt1 inhibition of breast cancer cell invasion reveals a protumorigenic role for TSC2, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 4134e4139. J.C. Acosta, A. O'Loghlen, A. Banito, M.V. Guijarro, A. Augert, S. Raguz, et al., Gil, Chemokine signaling via the CXCR2 receptor reinforces senescence, Cell 133 (2008) 1006e1018. G. Yang, D.G. Rosen, Z. Zhang, R.C. Bast Jr., G.B. Mills, J.A. Colacino, et al., The chemokine growth-regulated oncogene 1 (Gro-1) links RAS signaling to the senescence of stromal fibroblasts and ovarian tumorigenesis, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 16472e16477. C. Murphy, M. McGurk, J. Pettigrew, A. Santinelli, R. Mazzucchelli, P.G. Johnston, et al., Nonapical and cytoplasmic expression of interleukin-8, CXCR1, and CXCR2 correlates with cell proliferation and microvessel density in prostate cancer, Clin. cancer Res. Official J. Am. Assoc. Cancer Res. 11 (2005) 4117e4127. G. Yang, D.G. Rosen, G. Liu, F. Yang, X. Guo, X. Xiao, et al., CXCR2 promotes ovarian cancer growth through dysregulated cell cycle, diminished apoptosis, and enhanced angiogenesis, Clin. cancer Res. Official J. Am. Assoc. Cancer Res. 16 (2010) 3875e3886. Y.S. Lee, I. Choi, Y. Ning, N.Y. Kim, V. Khatchadourian, D. Yang, et al., Interleukin-8 and its receptor CXCR2 in the tumour microenvironment promote colon cancer growth, progression and metastasis, Br. J. Cancer 106 (2012) 1833e1841. T. Desurmont, N. Skrypek, A. Duhamel, N. Jonckheere, G. Millet, E. Leteurtre, et al., Overexpression of chemokine receptor CXCR2 and ligand CXCL7 in liver metastases from colon cancer is correlated to shorter disease-free and overall survival, Cancer Sci. 106 (2015) 262e269. M. Wu, B. Wang, J. Fei, N. Santanam, E.R. Blough, Important roles of Akt/PKB signaling in the aging process, Front. Biosci. 2 (2010) 1169e1188. M.V. Astle, K.M. Hannan, P.Y. Ng, R.S. Lee, A.J. George, A.K. Hsu, et al., AKT induces senescence in human cells via mTORC1 and p53 in the absence of DNA damage: implications for targeting mTOR during malignancy, Oncogene 31 (2012) 1949e1962. H. Miyauchi, T. Minamino, K. Tateno, T. Kunieda, H. Toko, I. Komuro, Akt negatively regulates the in vitro lifespan of human endothelial cells via a p53/ p21-dependent pathway, EMBO J. 23 (2004) 212e220.