Oncogene (2015) 34, 165–176 & 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc

ORIGINAL ARTICLE -17 produced by tumor microenvironment promotes self-renewal of CD133 þ cancer stem-like cells in ovarian cancer

T Xiang1,5, H Long1,5,LHe1, X Han1, K Lin1, Z Liang2, W Zhuo1, R Xie1,3 and B Zhu1,4

Inflammatory , components of cancer stem cells (CSCs) niche, could affect the characteristics of CSCs such as self-renewal and metastasis. Interleukin-17 (IL-17) is a new pro-inflammatory mainly produced by T-helper (Th17) cells and . The effects of IL-17 on the characteristics of CSCs remain to be explored. Here we first demonstrated a role of IL-17 in promoting the self-renewal of ovarian CD133 þ cancer stem-like cells (CSLCs). We detected IL-17-producing cells (CD4 þ cells and CD68 þ macrophages) in the niche of CD133 þ CSLCs. Meanwhile, there was IL-17 receptor expression on CD133 þ CSLCs derived from A2780 cell line and primary ovarian cancer tissues. By recombinant human IL-17 stimulation and IL-17 transfection, the growth and sphere formation capacities of ovarian CD133 þ CSLCs were significantly enhanced in a dose-dependent manner. Moreover, ovarian CD133 þ CSLCs transfected with IL-17 showed greater tumorigenesis capacity in nude mice. These data suggest that IL-17 promoted the self-renewal of ovarian CD133 þ CSLCs. Further investigation through profiling revealed that the stimulation function of IL-17 on self-renewal of ovarian CD133 þ CSLCs might be mediated by the nuclear factor (NF)-kB and p38 mitogen- activated kinases (MAPK) signaling pathway. NF-kB and p38 MAPK were activated by IL-17. More importantly, IL-17- promoted self-renewal was inhibited by specific inhibitors of NF-kB and p38 MAPK. Taken together, our data indicate that IL-17 contributed to ovarian cancer malignancy through promoting the self-renewal of CD133 þ CSLCs and that IL-17 and its signaling pathway might serve as therapeutic targets for the treatment of ovarian cancer.

Oncogene (2015) 34, 165–176; doi:10.1038/onc.2013.537; published online 23 December 2013 Keywords: cancer stem cells; IL-17; self-renewal; ovarian cancer

INTRODUCTION between CSCs niche and inflammation.10 Indeed, mounting Ovarian cancer is a complex tumor that displays cellular evidence has shown that and cytokines, such as 11 12 13 14 15 heterogeneity within the bulk tumor. Recent studies suggest that interleukin (IL)-1, IL-4, IL-6, IL-8, and IL-15, produced ovarian cancer contained a small subset of tumor cells with potent by CSCs itself or tumor microenvironment could exert tumorigenesis and stem cell-like properties, which are so-called direct effects on CSCs. However, the effect of inflammation on cancer stem cells (CSCs) or cancer stem-like cells (CSLCs).1–3 There CSCs characteristics in ovarian cancer remains to be detailed is now growing evidence that ovarian cancer, like many cancers, is explored. driven by CSCs.4 These cells are characterized by their self-renewal IL-17 is a newly identified pro-inflammatory cytokine that is þ capacity, multi-lineage differentiation properties and highly mainly produced by activated CD4 T-helper cells (also known as þ 16 oncogenic potential. Besides the capacity of tumorigenesis, Th17 cells), macrophages and CD8 T cells. Accumulating ovarian CSCs may mediate tumor metastasis and tumor relapses evidences have shown that IL-17-positive cells were frequently by virtue of their resistance to radio- and chemotherapies.5 Thus, involved in multiple inflammation-associated cancers, including 17 18 19,20 elucidating characteristics of ovarian CSCs is of great importance breast cancer, colorectal cancer, prostate cancer and 21,22 and may facilitate the development of novel agents targeting ovarian cancer. Although IL-17 may promote the growth of these cell populations. ovarian cancer, little is known about the specific biological Substantial evidences have suggested that CSCs were regulated mechanisms through which IL-17 contributes to ovarian cancer by intrinsic cellular pathways6 as well as extrinsic signals initiation or progression. In other cancers, it has been suggested generated by the tumor microenvironment. Cancer-related that IL-17 contributed to tumor malignancy by promoting inflammation is the hallmark of tumor microenvironment and capabilities of chemoresistance, angiogenesis and invasion of has powerful effects on tumor development.7 The tumor cancer cells.23 Interestingly, these cellular behaviors induced by microenvironment in which ovarian cancer develops has been IL-17 were also characteristics of CSCs, suggesting that IL-17 described as one enriched with a broad spectrum of pro- may contribute to the development of CSCs. Taken together, inflammatory cytokines and chemokines, which have been these data suggest that the role of IL-17 signaling in ovarian shown to influence clinical disease status and prognosis.8,9 cancer should be evaluated in the subpopulation of the ovarian Recent studies have shown that there may be a direct link CSCs.

1Institute of Cancer, Xinqiao Hospital, Third Military Medical University, Chongqing, China; 2Department of Obstetrics and Gynecology, Southwest Hospital, Third Military Medical University, Chongqing, China; 3Department of Obstetrics and Gynecology, Xinqiao Hospital, Third Military Medical University, Chongqing, China and 4Biomedical Analysis Center, Third Military Medical University, Chongqing, China. Correspondence: Dr R Xie, Department of Obstetrics and Gynecology, Xinqiao Hospital, Third Military Medical University, Chongqing 400037, China or Dr Professor B Zhu, Institute of Cancer, Xinqiao Hospital, Third Military Medical University, Chongqing 400037, China. E-mail: [email protected] or [email protected] 5These authors contributed equally to this work. Received 10 June 2013; revised 12 October 2013; accepted 1 November 2013; published online 23 December 2013 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 166 IL-17 signaling is transduced by IL-17 receptor (IL-17R), which (Figure 2e), indicating that IL-17 was expressed in the niche of activates the factor nuclear factor (NF)-kB. However, CSCs in ovarian cancer. Collectively, we found that IL-17- albeit weakly, IL-17R signaling can also be transduced by p38 producing cells locate in the niche of CSCs in ovarian cancer MAPK, resulting in the activation of a potential molecular AP1 tissues, suggesting the potential role of IL-17 on CD133 þ CSLCs in bridge between inflammation and cancer.24–27 Interestingly, these ovarian cancer. two signaling pathways are involved in the self-renewal of normal 28,29 stem cells and CSCs. Together, based on these evidences, we IL-17 promotes ovarian CD133 þ CSLCs self-renewal in vitro aimed to investigate the effect of IL-17 on the ovarian CSCs and the underlying mechanisms. Self-renewal is the most important characteristic of CSCs. And, NF-kB and MAPK, which could be activated by IL-17/IL-17R signaling, are involved in self-renewal of stem cells and CSCs.28,29 Thus, we hypothesized that IL-17 might promote self-renewal of ovarian RESULTS þ 1,3 þ CD133 CSLCs. Sphere formation reflects stem cell self-renewal. Expression of IL-17R on the ovarian CD133 CSLCs We first dissociated A2780-derived CD133 þ CSLCs spheres into It is known that inflammatory cytokines are components of CSCs 100 cells and stimulated cells with different concentrations of niche. Thus, we first used the human inflammatory response PCR recombinant human IL-17 (rhIL-17) (25, 50, 75, 100, 250, 500, 750 array to analyze the of inflammation-associated and 1000 ng/ml). After incubation for 7 days, the number of newly þ in A2780-derived ovarian CD133 CSLCs, which have been formed spheres were counted. As shown in Figure 3a, the number proved to be CSCs characteristics including self-renewal, multi- of spheres significantly increased by rhIL-17 stimulation from lineage differentiation properties and highly oncogenic potential concentrations of 25–250 ng/ml (P 0.01) when compared with 30 o in our previous study. Differences in gene expression profiles IL-17-untreated group, and the sphere number reached the þ between the A2780-derived ovarian CD133 CSLCs and non- maximum at 250 ng/ml. Then we dissociated A2780-derived À CSLCs (CD133 A2780 cells) were compared. Figure 1a shows the CD133 þ CSLCs spheres into 1, 10 and 100 cells and treated with þ scatter plot of expression level of each gene in ovarian CD133 250 ng/ml of rhIL-17 for 7 days. Similarly, treatment with rhIL-17 CSLCs and non-CSLCs. Interestingly, expression levels of many significantly increased the number of spheres (Po0.01; Figures 3b inflammation cytokines and their receptors were upregulated in and c). In addition, the effect of rhIL-17 was blocked by an IL-17R- þ the A2780-derived ovarian CD133 CSLCs (Figure 1b and neutralizing antibody at the concentration of 100 ng/ml (Po0.01; Supplementary Table 2). Some of them (such as IL-1R, IL-4R, Figure 3d and Supplementary Figure S1A). Importantly, similar 11–14 IL-6R and IL-8) were previous reported, whereas some of results were obtained in ovarian CD133 þ CSLCs derived from them were newly observed. Notably, the expression of IL-17R in ovarian tumor tissues (P 0.01, Figures 3e and f). In contrast to its þ o A2780-derived ovarian CD133 CSLCs was first observed. effects on CD133 þ CSLCs, rhIL-17 did not have any effect on the Expression of IL-17R was further verified by flow cytometric proliferation capacity of CD133-negative cells (non-CSLCs) analysis and immunofluorescence assay. As shown by flow (P40.05; Supplementary Figure S1B). However, CD133 þ CSLCs cytometric analysis, the percentage of IL-17R-expressing cells in percentage increased threefold more than control group in the þ the A2780-derived ovarian CD133 CSLCs was 73.80 þ 9.36%, serum-free culture medium when 100 ng/ml rhIL-17 was added in whereas it was just 24.30±3.4% in non-CSLCs, at the individual the culture medium (Po0.05; Figure 3g). cell level, determined by mean fluorescence intensity measured in Sphere size is another indicator of progenitor cell proliferation. FACS assays; the average expression level of IL-17R is fivefold Therefore, we dissociated the spheres of patient-derived CSLCs higher on the surface of CSLCs than non–CSLCs (Figure 1c). To (ova 4) into single cells and plated them into 96-well plates at the þ determine whether primary ovarian CD133 CSLCs could also density of one cell per well. The size of the sphere generated from express IL-17R, we performed immunofluorescence assay on a single cell was observed at day 5, 10 and 15. As shown in Figures þ CD133 CSLCs isolated from primary ovarian cancer tissues and 3h and i, treatment of rhIL-17 significantly increased the sphere on in situ ovarian cancer tissues, respectively. Consistent with size of CD133 þ CSLCs (P 0.05), suggesting that treatment with þ o A2780-derived CD133 CSLCs, IL-17R expression was both found rhIL-17 also enhance the size of spheres. CD133 þ CSLCs from þ on primary tumor-derived ovarian CD133 CSLCs spheres A2780 cell line showed the similar results to the cells from the þ (Figure 1d) and on CD133 CSLCs in situ ovarian cancer tissues patient (Supplementary Figure S1C and D). (Figure 1e). These data strongly proved that there was IL-17R Taken together, these results suggest that IL-17 could promote þ expression in ovarian CD133 CSLCs. ovarian CD133 þ CSLCs self-renewal in vitro.

IL-17-producing cells locate in the niche of ovarian CD133 þ CSLCs IL-17 enhances the tumorigenic potential of ovarian þ IL-17R functions through binding with its ligand IL-17. Therefore, CD133 CSLCs in vivo we first detected IL-17-producing cells in ovarian cancer using We have shown that IL-17 could promote the self-renewal of immunohistochemistry. The IL-17-positive staining cells were ovarian CD133 þ CSLCs in vitro. Tumorigenic potential represents located mainly in the stroma of tumor section (Figure 2a, left). the self-renewal ability of CD133 þ CSLCs in vivo. To further The majority of IL-17-positive staining cells showed classical evaluate the role of IL-17 on ovarian CD133 þ CSLCs self-renewal morphological characteristics of (with a medium- in vivo, we determined to acquire ectopic expression of IL-17 in sized oval nucleus or kidney-shaped nucleus) (Figure 2a, middle) CD133 þ CSLCs using lentivirus-mediated transduction to mimic and lymphocyte (with a small round nucleus) (Figure 2a, right). IL-17 in vivo effect on CD133 þ CSLCs. Therefore, we transduced To further confirm this, we performed immunofluorescence to human IL-17 gene into CD133 þ CSLCs and got stable expression check the localization of IL-17 with macrophages and lymphocytes of IL-17 (74.5±4.80 ng/ml/104 cells) in lenti-green fluorescent in ovarian cancer tissues. As shown in Figures 2b and c, IL-17- protein (GFP)-IL-17-transduced CD133 þ CSLCs, whereas there positive staining cells co-localized with CD4 þ lymphocytes and was no expression of IL-17 in lenti-GFP-transduced cells CD68 þ macrophages, respectively. To determine whether IL-17- (Supplementary Figure S2A and S2B). Secondly, we analyzed the producting cells locate in the niche of ovarian CD133 þ CSLCs, we effect of IL-17 overexpression on sphere formation of ovarian further stained ovarian cancer tissue sections with CSCs marker CD133 þ CSLCs. Consistent with the treatment of rhIL-17, IL-17- CD133 and IL-17 antibody. As shown in Figure 2d, there were transduced ovarian CD133 þ CSLCs also showed increased sphere IL-17-producing cells around CD133 þ CSLCs, the distance of IL-17- formation when compared with GFP-transduced CD133 þ CSLCs producing cells from CD133 þ CSLCs was about 25–75 mm (Supplementary Figure S2C and D); however, this effect caused by

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 167

Figure 1. Expression analysis of IL-17R on the ovarian CD133 þ CSLCs. (a) Scatter plot of expression level of each gene in ovarian CD133 þ CSLCs and A2780 cell line. The black lines represented a onefold change and the pink lines indicated a fivefold change in baseline. (b) The fold induction of inflammatory cytokine genes and receptor genes between ovarian CD133 þ CSLCs and A2780 cell line was shown. (c) Flow cytometry analysis of IL-17R-expressing cells in A2780-derived CD133 þ CSLCs and non-CSLCs. (d) Immunofluorescence detection of IL-17R- positive and CD133 þ cells in CD133 þ CSLCs isolated from primary ovarian cancer tissues. Scale bars, 25 mm. (e) Immunofluorescence detection of IL-17R-positive and CD133 þ cells in ovarian cancer tissues. Scale bars, 50 mm. ectopic IL-17 expression on CD133 þ CSLCs could be blocked by anti- (Figure 4d; Po0.05) and in the tumors (Figure 4f; Po0.05) from IL-17R antibody (Supplementary Figure S2E and F). Lenti-GFP-IL-17 CD133 þ CSLC-grafted mice than those in lenti-GFP Then we subcutaneously injected IL-17- or GFP-infected ovarian CD133 þ CSLC-grafted mice, suggesting the involvement of IL-17 CD133 þ CSLCs cells (1 Â 104) into 6–8-week-old female nude mice. in the maintenance of CSCs. Tumor volume and weight were measured at the day of 70 days Recent studies have suggested that CD133 þ CSCs in ovarian after implantation. As shown in Figures 4a and b, the volume of cancer could differentiate into endothelial cells and contribute to tumors from the Lenti-GFP-IL-17 group were much bigger than the angiogenesis by serving as vascular progenitors.31 On the the ones from the Lenti-GFP group. Similarly, the weight of tumors other hand, IL-17 has been shown to promote tumor angiogenesis in the Lenti-GFP-IL-17 group was also significantly higher than in animal models.32 Thus, we hypothesized that the promotion of those in Lenti-GFP group (Figure 4c; Po0.05). These data indicate IL-17 on tumor growth may in part caused by increasing that ectopic IL-17 expression in CD133 þ CSLCs significantly angiogenesis. To test this possibility, we analyzed tumor enhanced the tumorigenic potential in vivo. In addition, the vasculature by labeling the endothelial marker CD31. percentages of CD133 þ cells in tumors from grafted mice were Immunofluorescence results of tumor sections demonstrated also assessed by immunofluorescence and flow cytometry. There that vascular density was increased in Lenti-GFP-IL-17 were significantly more CD133 þ cells percentage both in per field CD133 þ CSLC-grafted group than that in Lenti-GFP

& 2015 Macmillan Publishers Limited Oncogene (2015) 165 – 176 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 168

Figure 2. IL-17-producing cells infiltrated in ovarian cancer tissues and located in the niche of CD133 þ CSLCs. (a) Immunohistochemical detection of IL-17-positive cells in ovarian cancer tissue. Left: IL-17-positive cells located in the peritumoral areas (arrows and inset), scale bar, 100 mm; middle: IL-17-positive macrophages characterized by an oval- or a kidney-shaped nucleus (arrows and inset), scale bar, 25 mm; right: IL-17-positive lymphocytes characterized by a small round nucleus (arrows and inset), scale bar, 25 mm. (b) Immunofluorescence detection of IL-17-positive and CD4 þ cells in ovarian cancer tissue. Nuclei were stained with DAPI. Scale bars, 10 mm. (c) Immunofluorescence detection of IL-17-positive and CD68 þ cells in ovarian cancer tissue. Nuclei were stained with DAPI. Scale bars, 7.5 mm. (d) Immunofluorescence detection of IL-17-positive and CD133 þ cells in ovarian cancer specimens. Nuclei were stained with DAPI. Scale bars, 50 mm. (e) Graph reports the distance of IL-17-producing cells from CD133 þ cells.

CD133 þ CSLCs group (Supplementary Figure S3A and B). Previous of vessel-like structures was observed when cells were treated literatures have shown that CSCs could differentiate into vascular with rhIL-17 or transfected with Lenti-GFP-IL-17 (Supplementary endothelial cells via autocrine and paracrine production of Figure S3D and F). angiogenic factors such as vascular endothelial (VEGF).33 Therefore, increased CD133 þ CSLCs self-renewal might contribute to the enhanced angiogenesis that we found in Lenti- IL-17 increase NF-kB and AP1 target genes expression in ovarian þ GFP-IL-17 CD133 þ CSLC-grafted tumors. To test these possibilities, CD133 CSLCs we used a three-dimensional in vitro system with high-density To investigate the underlying mechanisms of IL-17 on ovarian Matrigel. Our results showed that ovarian CD133 þ CSLCs could CD133 þ CSLCs self-renewal, we compared the differential gene form vessel-like structures similar to those observed with HUVEC expression profiles between IL-17- and PBS-treated A2780-derived structures (Supplementary Figure S3C). Interestingly, no increase CD133 þ CSLCs by microarray analysis. The gene expression

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 169

Figure 3. Sphere formation analysis of ovarian CD133 þ CSLCs stimulated by rhIL-17. (a) A2780-derived spheres were dissociated into 100 cells and stimulated by different concentrations of rhIL-17 (0–1000 ng/ml) for 7 days. The number of spheres were counted. (b, c) A2780-derived spheres were dissociated into 1, 10 and 100 cells and treated with rhIL-17 (250 ng/ml) or untreated for 7 days in suspension culture. Representative images and number of the spheres were shown. Scale bars in b,50mm. (d) Dissociated A2780-derived spheres were incubated with IL-17R-neutralizing antibody (anti-IL-17R) for 7 day. The number of spheres were counted. (e, f) Dissociated cells from the patient-derived ovarian CD133 þ CSLCs (ova1, ova2 and ova3) were incubated with IL-17(100 ng/ml) and/or anti-IL-17R for 7 days. (g) CD133 þ CSLCs percentage change in the serum-free culture medium added with 100 ng/ml rhIL-17 detected by Flow cytometry analysis. (h, i) Size of spheres generated from single-cell cultures of patient-derived CD133 þ CSLCs (ova 4). Scale bar in h,50mm. All experiments were performed three times and data were expressed as mean±s.d. *Po0.05, **Po0.01. profiles of IL-17-stimulated spheres at time points of 4 and 6 h genes in the NF-kB and p38 MAPK response signature, and the top were analyzed (which were published on the NCBI web site). We response signature (CSNK2A1) belonged to the NF-kB pathway identified 2096/2028 genes whose expression were twofold (Supplementary Figure S4). Next, we screened eight differentially down-/upregulated in the IL-17-stimulated ovarian CD133 þ CSLCs expressed self-renewal-related genes (Figure 5b). Quantitative PCR compared with PBS stimulated at 4 h of treatment. After 6 h of further validated the mRNA changes of the eight self-renewal- treatment, the down-/upregulated genes were significantly related genes at 4 h on CD133 þ CSLCs both from A2780 cell changed and bounced back to the twofold change lines line (Figure 5c, upper panel) and human ovarian cancer (Figure 5c, (Figure 5a). Signaling Net Array showed differentially regulated bottom panel). Interestingly, most of these genes except

& 2015 Macmillan Publishers Limited Oncogene (2015) 165 – 176 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 170

Figure 4. The effect of IL-17 on the growth of ovarian cancer xenografts in vivo. Ovarian A2780-derived CD133 þ CSLCs (1 Â 104 cells per mouse) transduced with Lenti-GFP or Lenti-GFP-IL-17 were injected subcutaneously into female nude mice, and were killed at day 70 after transplantation. (a) Images of xenografts tumor at day 70. (b) The curve of xenografts tumor volumes during 70 days. (c) Tumor weight of xenografts tumor at day 70. (d) Immunofluorescence detection of CD133 þ cells in xenografts tumors. Scale bars, 50 mm. (e) Percentage of CD133 þ cells in xenografts tumors. (f) Flow cytometry analysis of CD133 þ cells percentage in xenografts tumors. Experiments were performed three times and data were expressed as mean±s.d. *Po0.05, **Po0.01.

TCL1A were identified as NF-kB and/or AP1 target genes e). Collectively, these results suggested that NF-kB pathway was (Figure 5d) using the online Chip Transcription Factor Search activated by IL-17, and this pathway had a pivotal role in IL-17- (http://www.sabiosciences.com/chipqpcrsearch.php). These results enhanced self-renewal in ovarian CD133 þ CSLCs. suggest that IL-17 may promote self-renewal through activation of As mentioned above, p38 MAPK was another response NF-kB and p38 MAPK signaling pathways. signature after stimulation of IL-17. Interestingly, p38 MAPK pathway is also a downstream signaling pathway of IL-17/IL-17R. Therefore, we detected the activation of p38 MAPK after IL-17 NF-kB and p38 MAPK signaling pathways are involved in IL-17- stimulation. There was significant increase of phospho-p38 MAPK þ enhanced self-renewal of ovarian CD133 CSLCs in ovarian CD133 þ CSLCs after IL-17 stimulation (Figure 7a). NF-kB is the best known mediator of IL-17/IL-17R-associated To further validate the activation, we did immunofluorescence cellular responses24 and is associated with the regulation of analysis at 15 min of IL-17 stimulation, and found that strong CSCs.28 Therefore, we checked whether NF-kB pathway mediated signals of phospho-p38 MAPK were located in the cytoplasm and the IL-17-induced self-renewal of ovarian CD133 þ CSLCs. The nucleus (Figure 7b). To confirm the role of p38 MAPK in IL-17- results showed that IL-17 stimulation (100 ng/ml) was associated mediated self-renewal, we used p38 MAPK inhibitor SB203580 with a significant increase in the level of p65 in the nuclear (5 mM) to block its activation. Our results found that SB203580 extracts at 5 min, whereas the level of p65 in the cytoplasm could partially weaken the increase in self-renewal by IL-17 decreased (Figure 6a). Immunofluorescence analysis showed treatment (Figures 7c and d). similar results (Figure 6b). These results indicate that NF-kB was Furthermore, we determined NF-kB or the p38 MAPK activation activated after IL-17 (100 ng/ml) treatment. Furthermore, we in non-CD133 þ CSLCs after IL-17 stimulation. The results showed determined the DNA binding activity of NF-kB p65 in the nuclear that slight NF-kB and no p38 MAPK activation occurred when non- extracts using enzyme-linked immunosorbent assay (ELISA) assay. CSLCs were treated with rhIL-17 at the concentration of 100 ng/ml As shown in Figure 6c, the DNA binding activity of NF-kB p65 in (Supplementary Figure S5). These results were consistent with the IL-17 stimulation group at 5 min of treatment was significantly results that few non-CSLCs expressed IL-17R (Figure 1c), and there higher than that in the control group (Po0.05). To further verify was no effect of IL-17 on non-CD133 þ CSLCs proliferation the role of NF-kB signaling in IL-17-enhanced self-renewal, we (Supplementary Figure S1B). used NF-kB inhibitor pyrrolidine dithiocarbamate (PDTC) (5 mM)to In addition, the number of spheres in the group treated with block its activation during the treatment of rhIL-17. As expected, both PDTC and SB203580 were significantly decreased when we found that the increased number of spheres caused by IL-17 compared with groups treated with PDTC or SB203580 alone treatment could be abolished by PDTC treatment (Figures 6d and (Figures 7c and d). This result suggests that the roles of p38 MAPK

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 171

Figure 5. Microarray analysis of A2780-derived CD133 þ CSLCs after treatment with rhIL-17 (100 ng/ml). (a) Scatter plot of median log2 intensities of the rhIL-17-stimulated A2780-derived CD133 þ CSLCs and PBS-stimulated cells at 4 and 6 h. (b) Hierarchical cluster analysis of differentially expressed self-renewal genes. (1) control, (2) A2780-derived CD133 þ CSLCs stimulated by rhIL-17 at 4 h, (3) A2780-derived CD133 þ CSLCs stimulated by rhIL-17 at 6 h. (c) Fold changes between rhIL-17-stimulated A2780-derived and patient-derived ovarian (ova5) CD133 þ CSLCs at 4 h determined by quantitative PCR. Experiments were performed three times and data were expressed as mean±s.d. *Po0.05, **Po0.01. (d) The relationship between the self-renewal genes and transcription factors (NF-kB/AP1). ‘O’: the transcription factor could activate the gene. ‘ Â ’: transcription factor could not activate the gene.

and NF-kB signaling were not redundant in mediating the effect corresponding receptors on their surface. We reasoned that of IL-17 on self-renewal of ovarian CD133 þ CSLCs. Using the two the expression profile of these receptors in CSCs may imply the inhibitor above (PDTC and SB203580, the same effects occurred spectrum of signals received from the niche. With CD133 þ CSLCs on patient-derived ovarian CD133 þ CSLCs (Figures 7e and f). derived from A2780 ovarian cancer cell line, we observed that Taken together, these data demonstrate that IL-17 was capable multiple cytokine receptors are highly upregulated. Among them, of promoting ovarian CD133 þ CSLCs self-renewal through activa- the regulatory function of IL-1R,11 IL-4R,12 IL-6R13 and IL-15R15 on tion of NF-kB and P38 MAPK signaling pathways. various characteristics of CSCs has been previously reported. Herein, we focused on an under-studied , IL-17R. Interestingly, in comparison with their CD133 À counterparts, DISCUSSION although the elevation of IL-17R at the mRNA level is moderate Since the discovery of CSCs, major efforts of researches were (Figure 1b), the augmentation of this protein on CSLCs’ surface is focused on the elucidation of intrinsic molecular mechanisms much more pronounced (Figure 1c). In coordination, IL-17- underlying the stem cell properties and the tumorigenicity of producing cells, specifically CD4 þ T cells and macrophages, were CSCs. Increasing evidences showed that the intrinsic properties of located in the niche adjacent to ovarian CD133 þ CSLCs. More CSCs are tightly regulated by specific signals from the micro- importantly, IL-17 not only promoted sphere formation of ovarian environments. It is well known that infiltrated leukocytes and CD133 þ CSLCs in vitro, but also enhanced the tumorigenic inflammatory cytokines are important constituents of the tumor potential of ovarian CD133 þ CSLCs in vivo, by p38 MAPK and microenvironment.9 However, the impacts of inflammatory NF-kB activation (Figure 7g). cytokines on CSCs, especially CSCs in ovarian cancer, are not We initiated this study with A2780 cell line as a model system well understood. Recent study showed a direct interaction for ovarian tumor. A2780 is a cell line derived from an untreated between CSCs and inflammatory cells and indicated that stage III–IV ovarian cancer patient.35 Although there is no specific inflammatory cells represented an important component of the histological origin recorded in the original publication, this cell line CSC niche.34 Cytokines and chemokines produced by has been well accepted as a model for high-grade serous inflammatory cells influence the fate of CSCs through the adenocarcinoma.36 However, very recently, we did notice that,

& 2015 Macmillan Publishers Limited Oncogene (2015) 165 – 176 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 172

Figure 6. NF-kB signaling pathways were involved in IL-17-promoted self-renewal of ovarian A2780-derived CD133 þ CSLCs. (a) Western blot analysis of p65 levels in cytoplasmic extract and nuclear extract at the indicated time points. (b) Immunofluorescence detection of the nuclear translocation of p65 in A2780-derived spheres at 5 min after IL-17 treatment. Scale bars, 5.0 mm. (c) DNA binding activity analysis of NF-kB p65 in A2780-derived spheres at indicated time after IL-17 treatment. (d, e) Sphere formation analysis. Representative macrographs on day 7 and number of A2780-derived spheres after treatment with rhIL-17 (100 ng/ml) and p65 inhibitor PDTC (5 mM) were shown. Scale bars in d, 100 mm. Experiments were conducted three times and data were expressed as mean±s.d. *Po0.05.

based on current genomic profiling data, A2780 cells does not our in vitro assays support a direct role of IL-17 on CSLC expansion, harbor any mutation or cope number change for TP53 gene, which we cannot rule out the proangiogenesis effect of IL-17 in our is a key characteristics of high-grade serous ovarian cancer.37 in vivo tumor model, in which the enhanced tumorigenic potential Therefore, A2780 may not be able to represent high-grade serous and tumor growth were observed with IL-17-expressing CD133 þ adenocarcinoma. We examined this limitation with six tumor CSLCs. Indeed, we found increased vascular density in these IL-17 samples collected from patients categorized as stage III serous ectopic-expressing ovarian CD133 þ CSLC-grafted tumors. One adenocarcinomas. Our results validated that, both phenotypically explanation for this observation is that the more ovarian CD133 þ and mechanistically, the role of IL-17 on CD133 þ CSLCs is very CSLCs promoted by IL-17, the more angiogenic properties the similar, no matter these cells were derived from A2780 or from tumors will have, which is supported by our findings that CD133 þ patients. This suggested that the IL-17 signaling axis may CSLCs form vessel-like structures (Supplementary Figure S3C). The represent a fundamental regulatory mechanism for various alternative possibility is that IL-17 stimulates endothelial cells or ovarian CSCs. fibroblasts to produce proangiogeneic factors, thereby promoting Previous studies have illustrated the dual role of IL-17 in cancer angiogenesis, which has been demonstrated previously.42,43 development. On one hand, IL-17 has been shown to induce Nevertheless, in our in vivo animal model, CD133 þ cells are tumor-specific cytotoxic T-lymphocyte activation through induc- significantly enriched in grafted tumors with ectopic IL-17 tion of IL-6 and IL-12;38 it can facilitate the maturation of dendritic expression (Figures 4d and f), suggesting that the enhanced CSLC cells;39 and also IL-17 is necessary for the efficient killing of ovarian self-renewal is a plausible mechanism, even though may not CSCs by gd T cells.40 Clinically, some reports showed that high being the single mechanism. density of IL-17-producing cells was associated with improved It is important to note that IL-17 should not be considered as a prognosis for advanced epithelial ovarian cancer.41 On the other general tumor growth factor: we and others44,45 have previously hand, IL-17 could promote angiogenesis within the tumor, mainly demonstrated that IL-17 treatment could not change the by stimulating surrounding endothelial cells and fibroblasts to proliferation capacity of cancer cells in vitro even though they produce proangiogenic factors such as VEGF.42,43 In this study, also express IL-17R. This target-specific functional style of IL-17 is we unveiled another mechanism for IL-17’s pro-tumor effect-to not limited to tumor cells: in a mouse model of promote the self-renewal of tumor stem cells. However, although (experimental autoimmune encephalomyelitis), IL-17 specifically

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 173

Figure 7. MAPK signaling pathways were involved in IL-17-promoted self-renewal of ovarian A2780-derived and patient-derived CD133 þ CSLCs. (a) Western blot analysis of the levels of the phospho-p38 protein in A2780-derived CD133 þ CSLCs at the indicated time points after rhIL-17 treatment (100 ng/ml). (b) Immunofluorescence detection of the phospho-p38 protein in A2780-derived spheres at 5 min after rhIL-17 treatment (100 ng/ml). Scale bars, 7.5 mm. (c, d) Sphere formation analysis. Representative macrographs on day 7 and number of A2780- derived spheres after treatment with rhIL-17 (100 ng/ml), p65 inhibitor PDTC (5 mM) and p38 MAPK inhibitor SB203580 (5 mM) were shown. Scale bars in c, 100 mm. (e, f) Representative macrographs on day 7 and number of patient-derived spheres (ova6) after treatment with rhIL-17 (100 ng/ml), p65 inhibitor PDTC (5 mM) and p38 MAPK inhibitor SB203580 (5 mM) were shown. Scale bars in e, 100 mm. (g) The working model illustrating the mechanism by which IL-17 (produced by tumor microenvironment) promoted self-renewal of ovarian CD133 þ CSLCs in ovarian cancer. Experiments were conducted three times and data were expressed as mean±s.d. *Po0.05. acts upon the progenitor of oligodendrocytes, not mature CD133 þ CSLCs. Signaling Net Array showed differentially regu- oligodendrocytes, although IL-17Rs are expressed on the surface lated genes in the NF-kB and p38 MAPK response signature. NF-kB of both cell types. Although blocking of IL-17 signaling in mature was an important mediator of IL-17/IL-17R-associated cellular oligodendrocytes fail to deliver any protection from responses,24 and was involved in the regulation of CSCs.28 p38 autoimmunity, blocking IL-17R signaling on the surface of MAPK also has been shown as a downstream signaling pathway of progenitor oligodendrocytes significantly rescues their survival IL-17.25–27 Previously, we found the activation of p38 MAPK in and neuron myelination.46 Taken together, for both ovarian cancer NSCLC cell lines after treatment with IL-17, leading to the cells and oligodendrocytes, a comprehensive signaling study is in production of VEGF-C.44 Moreover, we also found that many urgent needs to dissect mechanisms carrying differential effects of upregulated genes (such as STAT3, HOXD9, FOXO3, PAX6, TCL1A, IL-17 on progenitors versus differentiated progenies. NFATC1, ETV5 and FGF7) were involved in the maintenance of self- To elucidate the mechanisms underlying the effects of IL-17 on renewal either in stem cells or CSCs. For example, STAT3 was self-renewal of ovarian CD133 þ CSCLs, we compared the gene essential for the maintenance of neurosphere-initiating tumor expression profiles between IL-17- and PBS-treated ovarian cells in patients with glioblastomas.47 FOXO3 was essential for the

& 2015 Macmillan Publishers Limited Oncogene (2015) 165 – 176 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 174 maintenance of leukemia-initiating cells.48 Collectively, based on conjugated specific for human IL-17R (IL-17R, mouse the findings, we hypothesize that activations of NF-kB and p38 IgG1, BD Biosciences, East Rutherford, NJ, USA) for 30 min on ice. Isotype- MAPK may mediate IL-17-promoted self-renewal of ovarian matched monoclonal antibody was used as a control antibody (mouse CD133 þ CSLCs. This hypothesis was confirmed by the result that IgG1, APC). After incubation, cells were washed twice before analysis. IL-17 treatment activated NF-kB and p38 MAPK in ovarian CD133 þ CSLCs. More importantly, inhibition of NF-kB and/or p38 Immunohistochemistry þ MAPK significantly weakened the effect of IL-17 on CD133 CSLCs The samples of ovarian cancer sections were incubated in 0.01 mol/l citrate self-renewal, which provided further evidence for our hypothesis. buffer (pH 6.0) for 25 min in microwave oven to unmask antigens. After However, the pivotal target genes of NF-kB and p38 MAPK cooling to room temperature, samples were treated successively with 1% activations have not been identified in the present research methanol/30% H2O2 to block endogenous peroxidase and 5% bovine and remain to be determined. Besides NF-kB and p38 MAPK serum albumin to block nonspecific binding sites. After rinsing with PBS, pathway, another self-renewal-related gene TCL1 was also found they were incubated overnight at 4 1C with the goat-anti-rabbit IL-17 to be significantly upregulated by IL-17. TCL1 is an Akt kinase (BD Biosciences, Inc.) primary antibody. The sections were then rinsed with PBS and incubated with the biotinylated secondary antibody (Bios, Ltd, co-activator49 that has been proved to be involved in CSCs self- 50 Shanghai, China) for 30 min at room temperature. Then the sections were renewal. Therefore, our present data could not exclude washed and incubated with DAB (Real Envision, Gene Tech Company, the possibility that another (NF-kB- and AP1-independent) Shanghai, China), and was terminated by rinsing with distilled H2O. Finally, pathway might be involved in the IL-17 effects on ovarian the samples were counterstained with hematoxylin. CD133 þ CSLCs. In summary, our results showed that IL-17-producing cells were Immunofluorescence components of CD133 þ CSLCs niche in ovarian cancer. IL-17 could promote self-renewal in vitro and enhance tumorigenic potential Samples of spheres and frozen ovarian cancer sections were attached to þ the coverslips, fixed in 4% paraformaldehyde solution for 30 min and of ovarian CD133 CSLCs in xenograft mice. And these effects washed three times with PBS. Then, samples were permeabilized with ice- were mediated by NF-kB and p38 MAPK signaling. Owing to the cold 0.25% Triton X-100 for 30 min at room temperature. Samples used for differential effect of IL-17 on CSLCs and non-CSLCs, the IL-17-NF- the analysis of p65 and p38 were not permeabilized. After permeabiliza- kB/p38 signaling network may be an ideal candidate for antitumor tion, samples were washed three times with PBS and blocked with 5% therapy against ovarian CSC. bovine serum albumin in PBS for 30 min at room temperature. After washing with PBS, the samples were incubated at 4 1C overnight with the following antibodies, respectively: goat anti-IL-17 monoclonal antibody MATERIALS AND METHODS (1:100; BD Biosciences), rabbit anti-CD4 polyclonal antibody (1:100; Bios, þ Ltd), mouse anti-CD68 monoclonal antibody (1:50; Bios, Ltd), rabbit anti-IL- Generation and culture of ovarian CD133 CSLCs from cell line 17R monoclonal antibody (1:100; Santa Cruz Biotechnology, Dallas, TX, and primary ovarian cancer tissues USA), mouse anti-CD133 monoclonal antibody(1:100; Santa Cruz Biotech- The human ovarian cell line A2780 was obtained from the American Type nology), rabbit anti-P38 MAPK monoclonal antibody (1:1000; , Culture Collection (Mannassas, VA, USA) that was lastly tested for Danvers, MA, USA), rabbit anti-p65 monoclonal antibody (1:100; Santa Cruz authenticity by short tandem repeat profiling in September 2012. Ovarian Biotechnology) and mouse anti-CD31 monoclonal antibody (1:100; Cell þ 30 CD133 CSLCs were generated as previously described. A2780-derived Signaling). Then the samples were washed and incubated at room þ ovarian CD133 CSLCs were maintained in DMEM supplemented with 10% temperature for 30 min with secondary fluorescein (FITC)-conjugated IgG fetal bovine serum and penicillin/streptomycin. For A2780-derived stem and CY3-conjugated IgG, respectively. After washing, nuclei were then cell spheres, A2780 cells were seeded in six-well plates (Corning, New York, counterstained with DAPI (40,6-Diamidino-2-Phenylindole, dilactate). Cover- NY, USA) and cultured in serum-free DMEM/F12 (Hyclone, Logan, UT, USA) slips were viewed under fluorescence microscopy (LAS SP5-6000, supplemented with 10 ng/ml of basic fibroblast growth factor Wendenstrasse, Hamburg, Germany). (PeproTech, Inc, Rocky Hill, NJ, USA), 20 ng/ml of epidermal growth factor (PeproTech, Inc), 5 mg/ml of insulin (Sigma, St Louis, MO, USA) and 0.4% bovine serum albumin (Sigma). Under these culture conditions, cell Sphere formation assay spheres were formed within 3 days after seeding. Then they were CD133 þ CSLCs with or without rhIL-17 stimulation were seeded at mechanically dissociated and re-seeded at 3–4 day intervals. The four concentrations of 1, 10, 100 and 1000 cells per well on 96-well plates. tumor tissues (designated as ova1, 2, 3 and 4) and two ascites (designated After incubation at 37 1C in a humidified incubator with 5% CO2 for 2 as ova 5 and 6) from ovarian cancer patients used in this study were weeks, spheres containing X3 cells were counted under an inverted categorized as stage III serous adenocarcinomas. The patient-derived microscope. During the sphere formation assay, cells were treated with þ CD133 CSLCs were isolated by magnetic bead sorting (AC133, mouse IgG, rhIL-17, p65 inhibitor PDTC (5 mM) and p38 MAPK inhibitor SB203580 (5 mM). cell isolation ; Miltenyi, Bergisch Gladbach, Cologne, Germany) using the MidiMACS system as described previously.30 Ovarian cancer tissues were collected by surgical resection upon patient informed consent and Lentivirus infection approval by the Institutional Review Boards. IL-17-expressing lentivirus and GFP-expressing lentivirus was obtained commercially from Neuron Bio (Shanghai, China). The lentivirus was transfected into ovarian CD133 þ CSLCs according to the manufacturer’s Gene expression profiling using quantitative PCR array protocol. At 48 h after transduction, the medium was replaced and cells Total RNAs were extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, were harvested. The efficiency of IL-17 expression was determined by USA). First strand cDNA was prepared using SuperscriptIII RNase Reverse immunofluorescence and ELISA. Transcriptase (Invitrogen). Expression of 384 inflammation-associated genes was analyzed using the Human Inflammatory Response and Autoimmunity PCR Array (PAHS-3803E, SABiosciences, Frederick, MD, USA) according to the In vivo xenograft experiments manufacturer’s protocol. Gene expression levels were determined using the Severe combined immunodeficient mice (female, 4–6-week old) were Ct (DDCt) method, with b-actin as a reference gene. Fold changes were purchased from the Chinese Academy of Medical Sciences (Beijing, China). calculated by comparing the expression in ovarian CD133 þ CSLCs and non- Mice were housed and maintained in laminar flow cabinets under specific ovarian CD133þ CSLCs derived from A2780 cells. pathogen-free conditions. For xenograft experiments, 1 Â 104 ovarian CD133 þ CSLCs transfected with IL-17-expressing or GFP-expressing lentivirus were resuspended in 200 ml PBS/Matrigel (1:1; BD Biosciences) Flow cytometric analysis and injected into the left and right flank of mice. By visual observation and For flow cytometric analysis, tumor spheres were dissociated into single palpation, engrafted mice were inspected biweekly for tumors. At 3-month cells. Then the single cells were washed and incubated with post transplantation, mice were killed by cervical dislocation for further allophycocyanin (APC)-conjugated monoclonal antibody specific for evaluation. Mice care and use was performed in accordance with local human CD133/1 (AC133, mouse IgG1, Miltenyi) or phycoerythrin (PE)- ethical guidelines.

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 175 Microarray and quantitative real-time PCR 4 Curley MD, Therrien VA, Cummings CL, Sergent PA, Koulouris CR, Friel AM et al. Gene expression profiling of control group and IL-17-stimulated spheres CD133 expression defines a tumor initiating cell population in primary human (at 4 and 6 h of stimulation) were determined by microarray. For microarray ovarian cancer. Stem Cells 2009; 27: 2875–2883. analysis, Agilent Array platform was used. The sample preparation and 5 Baba T, Convery PA, Matsumura N, Whitaker RS, Kondoh E, Perry T et al. Epigenetic microarray hybridization were performed by Kangcheng Company regulation of CD133 and tumorigenicity of CD133 þ ovarian cancer cells. (Shanghai, China). Differentially expressed genes were identified through Oncogene 2009; 28: 209–218. fold-change filtering. The scatter plot was used for assessing the gene 6 O’Brien CA, Kreso A, Jamieson CH. Cancer stem cells and self-renewal. Clin Cancer expression variation (or reproducibility) between the compared arrays. Res 2010; 16: 3113–3120. Finally, Hierarchical Clustering was performed to show the distinguishable 7 Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011; gene expression pattern among samples. The differentially expressed 144: 646–674. genes after IL-17 treatment were identified using empirical criterion of 8 Pasquier J, Rafii A. Role of the microenvironment in ovarian cancer stem cell more than twofold change (order of change) when compared with the maintenance. BioMed Res Int 2013; 2013: 630782. control sample. 9 Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature Quantitative PCR was performed to validate the enriched genes 2008; 454: 436–444. observed in microarray experiments. The individual gene expression was 10 Yi L, Xiao H, Xu M, Ye X, Hu J, Li F et al. Glioma-initiating cells: a predominant analyzed with a PCR Kit (TaKaRa, Tokyo, Japan), and carried out in triplicate role in microglia/macrophages tropism to glioma. J Neuroimmunol 2011; 232: with an ABI 7300 Prism Sequence Detection System (Applied Biosystems, 75–82. Foster City, CA, USA). The PCR conditions were as follows: 95 1C for 30 s, 11 Sheridan C, Kishimoto H, Fuchs RK, Mehrotra S, Bhat-Nakshatri P, Turner CH et al. followed by 35 cycles of 95 1C for 5 s, 60 1C for 34 s, and 72 1C for 45 s. The CD44 þ /CD24- breast cancer cells exhibit enhanced invasive properties: an early relative gene expression levels were calculated using the comparative Ct step necessary for metastasis. Breast Cancer Res 2006; 8: R59. (DDCt) method, with GAPDH as a reference gene. The primers used for PCR 12 Todaro M, Alea MP, Di Stefano AB, Cammareri P, Vermeulen L, Iovino F et al. Colon were shown in Supplementary Table S1. cancer stem cells dictate tumor growth and resist cell death by production of interleukin-4. Cell Stem Cell 2007; 1: 389–402. 13 Iliopoulos D, Hirsch HA, Struhl K. An epigenetic switch involving NF-kappaB, Lin28, Western blotting Let-7 MicroRNA, and IL6 links inflammation to cell transformation. Cell 2009; 139: Spheres and cell pellets were resuspended and lysed in ice-cold NP-40 lysis 693–706. buffer containing protease inhibitors. The Nuclear-Cytosol Extraction Kit 14 Ginestier C, Liu S, Diebel ME, Korkaya H, Luo M, Brown M et al. CXCR1 blockade (KeyGen, Nanjing, China) and Phospho-Protein Purification Kit (KeyGen) selectively targets human breast cancer stem cells in vitro and in xenografts. were used for the extraction of p65 and p38 protein, respectively. For J Clin Invest 2010; 120: 485–497. western blotting analysis, membranes were first labeled with primary 15 Azzi S, Bruno S, Giron-Michel J, Clay D, Devocelle A, Croce M et al. Differentiation antibodies specific for p65 (1:100; Santa Cruz Biotechnology), p38 (1:500; therapy: targeting human renal cancer stem cells with . J Natl Cancer Inst Cell Signaling) and b-actin (Ab-1, mouse IgM, Calbiochem, San Diego, CA, 2011; 103: 1884–1898. USA). Then the membranes were labeled with HRP-conjugated anti-mouse 16 Rouvier E, Luciani MF, Mattei MG, Denizot F, Golstein P. CTLA-8, cloned from an and HRP-conjugated anti-rabbit secondary antibodies (Amersham, Piscat- activated , bearing AU-rich messenger RNA instability sequences, and away, NJ, USA). The signals were visualized by chemiluminescence homologous to a herpesvirus saimiri gene. J Immunol 1993; 150: 5445–5456. detection system (chemilmagerTM 5500, Alpha Innotech, San Leandro, 17 Zhu X, Mulcahy LA, Mohammed RA, Lee AH, Franks HA, Kilpatrick L et al. IL-17 CA, USA). expression by breast-cancer-associated macrophages: IL-17 promotes invasive- ness of breast cancer cell lines. Breast Cancer Res 2008; 10: R95. ELISA 18 Le Gouvello S, Bastuji-Garin S, Aloulou N, Mansour H, Chaumette MT, Berrehar F et al. High prevalence of Foxp3 and IL17 in MMR-proficient colorectal carcinomas. The DNA binding activity of NF-kB in nuclear protein extracts was assessed Gut 2008; 57: 772–779. using a commercially available NF-kB p65 transcription factor ELISA kit 19 Sfanos KS, Bruno TC, Maris CH, Xu L, Thoburn CJ, DeMarzo AM et al. Phenotypic according to the manufacturer’s instructions (Cayman Chemical Company, analysis of prostate-infiltrating lymphocytes reveals TH17 and Treg skewing. Ann Arbor, MI, USA). Clin Cancer Res 2008; 14: 3254–3261. 20 Steiner GE, Newman ME, Paikl D, Stix U, Memaran-Dagda N, Lee C et al. Expression Statistical analysis and function of pro-inflammatory interleukin IL-17 and IL-17 receptor in normal, All the data from quantitative assays were expressed as the mean±s.d. benign hyperplastic, and malignant prostate. Prostate 2003; 56: 171–182. Statistical analyses were performed using the independent samples t-test 21 Miyahara Y, Odunsi K, Chen W, Peng G, Matsuzaki J, Wang RF. Generation and or one-way analysis of variance. The difference was considered statistically regulation of human CD4 þ IL-17-producing T cells in ovarian cancer. Proc Natl Acad Sci USA 2008; 105: 15505–15510. significant when Po0.05. All statistical analyses were carried out with SPSS 13.0 software (IBM, Armonk, New York, NY, USA). 22 Kato T, Furumoto H, Ogura T, Onishi Y, Irahara M, Yamano S et al. Expression of IL-17 mRNA in ovarian cancer. Biochem Biophys Res Commun 2001; 282: 735–738. 23 Murugaiyan G, Saha B. Protumor vs antitumor functions of IL-17. J Immunol 2009; 183: 4169–4175. CONFLICT OF INTEREST 24 Shalom-Barak T, Quach J, Lotz M. Interleukin-17-induced gene expression in The authors declare no conflict of interest. articular chondrocytes is associated with activation of mitogen-activated protein kinases and NF-kappaB. J Biol Chem 1998; 273: 27467–27473. 25 Subramaniam SV, Cooper RS, Adunyah SE. Evidence for the involvement of JAK/ ACKNOWLEDGEMENTS STAT pathway in the signaling mechanism of interleukin-17. Biochem Biophys Res Commun 1999; 262: 14–19. This work was supported by National Nature Science Foundation of China (grant nos. 26 Subramaniam SV, Pearson LL, Adunyah SE. Interleukin-17 induces rapid tyrosine 81070018, 81071772 and 81222031), by the outstanding Youth Scientist Foundation and activation of raf-1 kinase in human monocytic progenitor of Chongqing (no. CSTC, 2008BA5035) and by National Key Basic Research Program cell line U937. Biochem Biophys Res Commun 1999; 259: 172–177. of China (973 program, grant nos. 2010CB529404 and 2012CB526603). 27 Chang SH, Park H, Dong C. Act1 adaptor protein is an immediate and essential signaling component of interleukin-17 receptor. J Biol Chem 2006; 281: 35603–35607. REFERENCES 28 Shostak K, Chariot A. NF-kappaB, stem cells and breast cancer: the links get 1 Zhang S, Balch C, Chan MW, Lai HC, Matei D, Schilder JM et al. Identification and stronger. Breast Cancer Res 2011; 13: 214. characterization of ovarian cancer-initiating cells from primary human tumors. 29 Armstrong L, Hughes O, Yung S, Hyslop L, Stewart R, Wappler I et al. The role of Cancer Res 2008; 68: 4311–4320. PI3K/AKT, MAPK/ERK and NFkappabeta signalling in the maintenance of human 2 Hu L, McArthur C, Jaffe RB. Ovarian cancer stem-like side-population cells are embryonic stem cell pluripotency and viability highlighted by transcriptional tumourigenic and chemoresistant. Br J Cancer 2010; 102: 1276–1283. profiling and functional analysis. Hum Mol Genet 2006; 15: 1894–1913. 3 Silva IA, Bai S, McLean K, Yang K, Griffith K, Thomas D et al. Aldehyde 30 Long H, Xie R, Xiang T, Zhao Z, Lin S, Liang Z et al. Autocrine CCL5 signaling dehydrogenase in combination with CD133 defines angiogenic ovarian cancer promotes invasion and migration of CD133 þ ovarian cancer stem-like cells via stem cells that portend poor patient survival. Cancer Res 2011; 71: 3991–4001. NF-kappaB-mediated MMP-9 upregulation. Stem Cells 2012; 30: 2309–2319.

& 2015 Macmillan Publishers Limited Oncogene (2015) 165 – 176 IL-17 promotes self-renewal of CD133 þ CSLCs T Xiang et al 176 31 Kusumbe AP, Mali AM, Bapat SA. CD133-expressing stem cells associated with 41 Lan C, Huang X, Lin S, Huang H, Cai Q, Lu J et al. High density of IL-17-producing ovarian metastases establish an endothelial hierarchy and contribute to tumor cells is associated with improved prognosis for advanced epithelial ovarian vasculature. Stem Cells 2009; 27: 498–508. cancer. Cell Tissue Res 2013; 352: 351–359. 32 Numasaki M, Watanabe M, Suzuki T, Takahashi H, Nakamura A, McAllister F et al. 42 Takahashi H, Numasaki M, Lotze MT, Sasaki H. Interleukin-17 enhances bFGF-, IL-17 enhances the net angiogenic activity and in vivo growth of human non- HGF- and VEGF-induced growth of vascular endothelial cells. Immunol Lett 2005; small cell lung cancer in SCID mice through promoting CXCR-2-dependent 98: 189–193. angiogenesis. J Immunol 2005; 175: 6177–6189. 43 Numasaki M, Lotze MT, Sasaki H. Interleukin-17 augments - 33 Wang R, Chadalavada K, Wilshire J, Kowalik U, Hovinga KE, Geber A et al. alpha-induced elaboration of proangiogenic factors from fibroblasts. Immunol Lett Glioblastoma stem-like cells give rise to tumour . Nature 2010; 468: 2004; 93: 39–43. 829–833. 44 Chen X, Xie Q, Cheng X, Diao X, Cheng Y, Liu J et al. Role of interleukin-17 in 34 Filatova A, Acker T, Garvalov BK. The cancer stem cell niche(s): the crosstalk lymphangiogenesis in non-small-cell lung cancer: enhanced production of between glioma stem cells and their microenvironment. Biochim Biophys Acta vascular endothelial growth factor C in non-small-cell lung carcinoma cells. 2013; 1830: 2496–2508. Cancer Sci 2010; 101: 2384–2390. 35 Hamilton TC, Foster BJ, Grotzinger KR, McKoy W, Young RC, Ozols RF. Develop- 45 Du JW, Xu KY, Fang LY, Qi XL. Interleukin-17, produced by lymphocytes, promotes ment of drug-sensitive and -resistant human ovarian cancer cell lines. Proc Am tumor growth and angiogenesis in a mouse model of breast cancer. Mol Med Rep Assoc Cancer Res 1983; 24: 313. 2012; 6: 1099–1102. 36 Ciucci A, De Stefano I, Vellone VG, Lisi L, Bottoni C, Scambia G et al. Expression of 46 Kang Z, Wang C, Zepp J, Wu L, Sun K, Zhao J et al. Act1 mediates IL-17-induced the glioma-associated oncogene homolog 1 (gli1) in advanced serous ovarian EAE pathogenesis selectively in NG2( þ ) glial cells. Nat Neurosci 2013; 16: cancer is associated with unfavorable overall survival. PLoS One 2013; 8: e60145. 1401–1408. 37 Domcke S, Sinha R, Levine DA, Sander C, Schultz N. Evaluating cell lines as tumour 47 Villalva C, Martin-Lanneree S, Cortes U, Dkhissi F, Wager M, Le Corf A et al. STAT3 models by comparison of genomic profiles. Nat Commun 2013; 4: 2126. is essential for the maintenance of neurosphere-initiating tumor cells in patients 38 Jovanovic DV, Di Battista JA, Martel-Pelletier J, Jolicoeur FC, He Y, Zhang M et al. with glioblastomas: a potential for targeted therapy? Int J Cancer 2011; 128: IL-17 stimulates the production and expression of proinflammatory 826–838. cytokines, IL-beta and TNF-alpha, by human macrophages. J Immunol 1998; 160: 48 Naka K, Hoshii T, Muraguchi T, Tadokoro Y, Ooshio T, Kondo Y et al. 3513–3521. TGF-beta-FOXO signalling maintains leukaemia-initiating cells in chronic myeloid 39 Antonysamy MA, Fanslow WC, Fu F, Li W, Qian S, Troutt AB et al. Evidence for a leukaemia. Nature 2010; 463: 676–680. role of IL-17 in organ allograft rejection: IL-17 promotes the functional 49 Laine J, Kunstle G, Obata T, Sha M, Noguchi M. The protooncogene TCL1 is an Akt differentiation of progenitors. J Immunol 1999; 162: 577–584. kinase coactivator. Mol Cell 2000; 6: 395–407. 40 Lai D, Wang F, Chen Y, Wang C, Liu S, Lu B et al. Human ovarian cancer stem-like 50 Bleau AM, Hambardzumyan D, Ozawa T, Fomchenko EI, Huse JT, Brennan CW cells can be efficiently killed by gammadelta T lymphocytes. Cancer Immunol et al. PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 Immunother 2012; 61: 979–989. activity in glioma tumor stem-like cells. Cell Stem Cell 2009; 4: 226–235.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2015) 165 – 176 & 2015 Macmillan Publishers Limited