Activation of CXCL5-CXCR2 axis promotes proliferation and accelerates G1 to S phase transition of papillary thyroid carcinoma cells and activates JNK and p38 pathways

ABSTRACT

C-X-C motif chemokine ligand 5 (CXCL5) is initially identified to recruit neutrophils by interacting with its receptor, C-X-C motif chemokine receptor 2 (CXCR2). Our prior work demonstrated that the expression levels of CXCL5 and CXCR2 were higher in the papillary thyroid carcinoma (PTC) tumors than that in the non-tumors. This study was performed to further investigate how this axis regulates the growth of PTC cells. B-CPAP cells (BRAF^V600E) and TPC-1 cells (RET/PTC rearrangement) expressing CXCR-2 were used as in vitro cell models. Our results showed that the recombinant human CXCL5 (rhCXCL5) promoted the proliferation of PTC cells. rhCXCL5 accelerated the G1/S transition, upregulated the expression of a group of S (DNA synthesis) or M (mitosis)-promoting cyclins and cyclin-dependent kinases (CDKs), and downregulated CDK inhibitors in PTC cells. The CDS region of homo sapiens CXCL5 gene was inserted into an eukaryotic expression vector to mediate the overexpression of CXCL5 in PTC cells. The phosphorylation of c-Jun N-terminal kinases (JNK) and p38, and the nuclear translocation of c-Jun were enhanced by CXCL5 overexpression, whereas attenuated by CXCR2 antagonist SB225002. Additionally, CXCL5/CXCR2 axis, JNK and p38 pathway inhibitors, SB225002, SP600125 and SB203580, suppressed the growth of PTC cells overexpressing CXCL5 in nude mice, respectively. Collectively, our study demonstrates a growth-promoting effect of CXCL5-CXCR2 axis in PTC cells in vitro and in vivo.

Keywords:

• papillary thyroid carcinoma
• CXCL5
• CXCR2
• cell proliferation
• cell cycle
• JNK pathway
• p38 pathway

Introduction

The worldwide incidence of thyroid carcinomas (TCs) is steadily rising over the last decade, and the papillary thyroid carcinoma (PTC) accounts for approximately 85% of all TCs¹. Although PTC is a well-differentiated tumor and can be removed by surgical resection², its recurrence risk still threatens the PTC patients’ life post the surgery. Identification of molecules essential for tumor cell survival and metastasis will provide novel insights in PTC therapy.

C-X-C motif chemokine ligand 5 (CXCL5) gene (the alias for CXCL5 gene is neutrophil-activating protein 78) encodes a protein consisted of 114 amino acids³. It is initially identified as a C-X-C chemokine that recruits neutrophils by binding to C-X-C motif chemokine receptor 2 (CXCR2)⁴. Our group prior demonstrated that the exogenous CXCL5 augmented the mobility and induced the epithelial-mesenchymal transition of two PTC cell lines expressing CXCR2⁵. Several previous studies are consistent with our work showing a similar pro-metastatic role of CXCL5-CXCR2 axis in various types of cancers, such as prostate cancer and breast cancer^6,7. However, the effects of this axis on tumor cell proliferation differ. While the exogenous CXCL5 of 1 nM promoted the proliferation of prostate cancer cells⁶, an even higher concentration of CXCL5 (10 nM) hardly affected the proliferation of breast cancer cells⁷. Although the discrepant phenomena may result from different purities and activities of the recombinant human CXCL5 (rhCXCL5) used in these two studies, it is plausible that CXCL5-CXCR2 axis has inconsistent effects on cell proliferation in different types of cancers.

Mitogen-activated protein kinase (MAPK) cascades regulate multiple cellular activities involved in cancer progression, including cell proliferation and cell cycle, with c-Jun N-terminal kinases (JNKs) and p38 as the two major molecule hubs⁸. Earlier studies revealing the roles of JNKs in the tumorigenesis are related to their ability to regulate the stress-induced cell apoptosis and antitumor therapy-induced cytotoxic effects⁹. Paradoxically, the phosphorylated JNKs were found to be overexpressed in human PTCs, and this elevation was associated with poor prognosis¹⁰. A broad spectrum inhibitor for JNK pathways, SP600125, can suppress the growth of TPC-1 cells, a PTC cell line harboring RET/PTC rearrangement¹¹. In addition, it has been reported that the proliferation decline of PTC cells is often accompanied by the deactivation of p38 pathway¹². Although both pro-survival and pro-death effects of JNK and p38 pathways have been reported in cancer cells before, the aforementioned studies suggest that the activation of these two MAPK pathways may contribute to the survival of PTC cells. Interestingly, both JNK and p38 pathways in the lung tissues of CXCL5 knockout mice were deactivated by comparison with the corresponding wild type mice¹³. The study at least implies that the CXCL5-CXCR2 axis is required for maintaining the signaling transduction of these two MAPK pathways.

In this study, we propose that CXCL5 can promote the proliferation of PTC cells by activating JNK and p38 MAPK pathways. The underlying mechanisms were explored in PTC cell lines harboring mutated oncoprotein BRAF^V600E (B-CPAP) and RET/PTC rearrangement (TPC-1).

Materials and methods

Cell culture and transfection

B-CPAP and KTC-1 cells were obtained from the Cell bank of Chinese Academy of Sciences (Shanghai, China), whilst TPC-1 cells were obtained from Zhong Qiao Xin Zhou Biotechnology Co.,Ltd. The cell lines were authenticated with short tandem repeat analysis, and had no mycoplasma infection detected in our lab. Cells were allowed to grow in RPMI medium (Gibco) plus 10% FBS (HyClone) under a standard cell culture condition.

The complete CDS region of homo sapiens CXCL5 (NM_002994) was amplified and inserted into the eukaryotic expression vector pEGFP-N1 between XhoI (upstream) and BamHI (downstream). The construction of CXCL5 overexpression (OE) plasmid was validated via DNA sequencing. Empty pEGFP-N1 vector (NC plasmid) was used as control. B-CPAP and TPC-1 cells were transfected with CXCL5 OE or NC plasmid by using Lipofectamine™ 3000 transfection reagent (#L3000075; Invitrogen) according the supplier’s instruction. For stable transfection, B-CPAP cells were first transfected with a plasmid and then selected with G418 (400 μg/mL) for over 2 weeks. Cells with the resistance to G418 were used in the following xenograft tumor assay.

Cell viability and cell cycle

PTC cells were seeded in the 96-well plates (5,000 cells per well), and after the attachment, cells were treated with 1, 5, 10 nM rhCXCL5 (#RPA860Hu01; USCN Business Co., Ltd.) up to 96 hrs, or with 5, 10 μM ribociclib (a CDK4/6 inhibitor) up to 120 hrs (Selleck Chemicals, Houston, TX, USA). Cell vitality was determined with a CCK-8 kit (#C0037; Beyotime) according to the manufacturer’s protocols. In addition, PTC cells were transfected with CXCL5 OE or NC plasmid, and 48 hrs later, these cells were treated with SB225002 of 5 μM (#SML0716; Sigma) for another 24 hrs before CCK-8 assay. Flow cytometry assay using propidium (#C1052; Beyotime) to stain the double-stranded DNA was performed to determine the cell cycle according to the manufacturer’s protocols.

Real-time quantitative PCR

Using the total RNAs as templates, cDNAs were processed via Super M-MLV reverse transcriptase (#PR6502; BioTeke). The relative mRNA expression levels of CXCL5 gene were analyzed on the ExicyclerTM 96 by using the SYBR Green I (#SY1020; Solarbio) and calculated via 2^−ΔΔCt (with β-actin as control). The primer sequences for CXCL5 gene were as following: F-5ʹ cgctgctgtgttgagaga 3ʹ/R-5ʹ ggctaccacttccaccttg 3ʹ.

Antibodies

Anti-CXCL5 antibody (#PAA860Hu01) and anti-CXCR2 antibody (#BA0732-2) were obtained from Wuhan USCN Business Co., Ltd. and Boster, respectively. Anti-phosphorylated cyclin dependent kinase (CDK)-1^Thr161 antibody (#ab183554) was from Abcam. Anti-cyclinD1 (#2922), anti-cyclinB1 (#4138), anti-CDK-4 (#12790), anti-phosphorylated JNK1/2^{Thr183/Tyr185} (#4668) and anti-JNK1/2 (#9252) antibodies were purchased from Cell Signaling Technology, Inc. Anti-cyclinE1 (#11554-1-AP), anti-CDK inhibitor 1A/p21 (CDKN1A) (#10355-1-AP), anti-CDKN1B/p27 (#25614-1-AP), and anti-proliferating cell nuclear antigen (PCNA) (#10205-2-AP) antibodies were obtained from SANYING Proteintech. Anti-c-JUN (#bs-20067R), anti-phosphorylated p38^{Thr180/Tyr182} (#bs-0636R) and anti-p38 (#bs-0637R) antibodies were obtained from Bioss.

Western blot

Total proteins were isolated by using RIPA buffer containing 1% phenylmethanesulfonyl fluoride (#P0013B; Beyotime). Equal protein samples were first separated with 10 % SDS-PAGE, and then were transferred to PVDF membranes (#IPVH00010; Millipore). After blocking with 5% (M/V) skim milk, the membranes were incubated with one of the primary antibodies at 4℃ overnight. The dilutions for anti-CXCL5 and anti-CXCR2 were 1: 300 and 1: 200, respectively. The dilutions for anti-cyclinD1, anti-CDK4 antibody, anti-cyclinE1, anti-CDKN1A, anti-CDKN1B, anti-cyclinB1, anti-p-CDK1, anti-p-JNK1/2, anti-JNK1/2 and anti-PCNA were 1: 1000. The dilutions for anti-c-jun, p-p38 and p38 were 1: 500. Then, the membranes were incubated with a secondary antibody, and the blot signals were visualized with ECL reagent (#P0018; Beyotime). For some experiments, the cells were first transfected with CXCL5 OE or NC plasmid, and 48 hrs later, cells were treated with 10 μM SP600125 (#S1876; Beyotime) or 20 μM SB203580 (#S1863; Beyotime) for another 24 hrs.

Immunofluorescence

After being fixed with 4 % paraformaldehyde, cells on the glass slides were permeabilizated with 0.1% tritonX-100 (#ST795; Beyotime), and blocked with normal goat serum. Thereafter, cells were incubated with anti-c-Jun (1:200) or anti-PCNA (1: 200) overnight, and then with Cy3-labled goat anti-rabbit secondary antibody (1: 200; #A0516; Beyotime) for 60 min in the dark. Cell nuclei were stained with DAPI (#C1002; Beyotime).

Xenografted tumor

The animal procedures complied with the Guide for the Care and Use of Laboratory Animals, and were approved by the Ethics Committee of Dalian Medical University. B-CPAP cells were resuspended in 0.1 mL 1× PBS, mixed with 0.1 mL Metrigel and then injected into the axillary skin of nude mice. Nude mice injected with B-CPAP cells overexpressing CXCL5 were given CXCR2 antagonist SB225002 (intraperitoneal injection of 10 mg/kg for three times a week), JNK pathway inhibitor SP600125 (intraperitoneal injection of 5 mg/kg for three times a week) or p38 inhibitor SB203580 (subcutaneous injection of 4 mg/kg for five times a week) for four weeks. The xenografted tumors were isolated and photographed on day 28 post the injection of cancer cells. The tumor volume was measured every three day before sacrificing on day 28.

Statistical analysis

Data were expressed as mean ± standard deviation or standard error, and analyzed with soft SPSS 20.0 package (IBM, Armonk, NY, USA). All experiments were repeated for three times. One-way ANOVA followed by Bonferroni multiple test was used to evaluate the statistical differences over three groups. Two-way ANOVA followed by Bonferroni multiple test was performed to compare the differences from different time points. Statistical differences between groups were considered significant when the p values < 0.05.

Results

Exogenous rhCXCL5 promotes proliferation and G1 to s transition of PTC cells

The expression level of CXCR2 was determined in TPC-1, B-CPAP and KTC-1 cells with western blot analysis. We found that the basal expression of CXCR2 was TPC-1 > B-CPAP > KTC-1 (Figure 1(a)). TPC-1 is a cell line of RET/PTC rearrangement, while B-CPAP and KTC-1 are cell lines carrying BRAF^V600E. Therefore, TPC-1 and B-CPAP cells with higher CXCR2 level and respectively representing RET/PTC rearrangement and BRAF^V600E were used in the following study.

Figure 1. Exogenous rhCXCL5 promotes PTC cell proliferation (a) The basal expression of CXCR2 in B-CPAP, TPC-1 and KTC-1 cells were determined with Western blot analysis. (b-c) B-CPAP cells and (d-e) TPC-1 cells were treated with 1, 5, 10 nM rhCXCL5 protein for 24, 48, 72 or 96 hrs before CCK-8 (n = 5), immunofluorescence and Western blot assays using anti-PCNA antibody (bars, 50 μm; 48 hrs post rhCXCL5 treatment). **P < 0.01 and *P < 0.05 (1 nM rhCXCL5 versus Control), ^##P < 0.01 and ^#P < 0.05 (5 nM rhCXCL5 versus Control), ^§§P < 0.01 (10 nM rhCXCL5 versus Control). PTC, papillary thyroid carcinoma; rhCXCL5, recombinant human C-X-C motif chemokine ligand 5; CXCR2, C-X-C motif chemokine receptor 2; PCNA, proliferating cell nuclear antigen.

B-CPAP and TPC-1 cells, grown in complete culture medium, were treated with 1, 5 or 10 nM rhCXCL5. Results from IF staining and western blot for PCNA and from CCK-8 assay indicated that rhCXCL5 promoted the PTC cell proliferation in a dose-dependent manner (Figure 1(b-e)). Data from flow cytometry assay confirmed the pro-proliferative effects of CXCL5 by showing that more PTC cells entered S phase when treated with 10 nM rhCXCL5 (Figure 2(a & d)). Further, the expression levels of a group of cell cycle-associated proteins in B-CPAP and TPC-1 cells were determined with western blot. rhCXCL5 upregulated the expression of cyclins (cyclin D1, E1 and B1) and enhanced the activity of cyclin kinases (CDK4 and CDK1), and downregulated the expression of cyclin dependent kinase inhibitors (CDKN1A and CDKN1B) (Figure 2(b-c) & (e-f)). Furthermore, 5 or 10 μM ribociclib, a CDK4/6 inhibitor, was used to treat B-CPAP and TPC-1 cells for −120 hrs. Ribociclib reduced cell proliferation as revealed by CCK-8 data (sFigure 1). B-CPAP cells responded to ribociclib more quickly than TPC-1 cells: 24 hrs versus 48 hrs post the ribociclib treatment (sFigure 1).

Figure 2. Exogenous rhCXCL5 promotes the G1 to S transition in PTC cells (a) B-CPAP cells and (D) TPC-1 cells were treated with or without 10 nM rhCXCL5 protein for 48 hrs before flow cytometry assay (Propidium staining). The levels of a group of cell cycle-associated proteins in (b-c) B-CPAP cells and (e-f) TPC-1 cells were detected with western blot (with β-actin as control; n = 3). **P < 0.01 (1 nM rhCXCL5 versus Control), ^##P < 0.01 (5 nM rhCXCL5 versus Control), ^§§P < 0.01 (10 nM rhCXCL5 versus Control). PTC, papillary thyroid carcinoma; rhCXCL5, recombinant human C-X-C motif chemokine ligand 5; CXCR2, C-X-C motif chemokine receptor 2.

MAPK pathways contribute to PTC cell proliferation induced by CXCL5-CXCR2 activation

Forty-eight hours after the CXCL5 OE or NC plasmid transfection, the mRNA and protein expression levels of CXCL5 gene were determined via real-time PCR and western blot, respectively. The results indicated a significant elevation of CXCL5 in cells transfected with CXCL5 OE plasmid (Figure 3(a-b) & (d-e)). In addition, SB225002, a CXCR2 selective inhibitor, suppressed the proliferation induced by CXCL5 overexpression (Figure 3(c & f)). In response to SB225002 of 5 μM, B-CPAP cells were more sensitive than TPC-1 cells (Figure 3(c & f)).

Figure 3. Forced overexpression of CXCL5-induced cell proliferation and activation of JNK and p38 pathways are inhibited by CXCR2 inhibitor SB225002.

The complete CDS region of homo sapiens CXCL5 (NM_002994) was amplified and inserted into the eukaryotic expression vector pEGFP-N1 (CXCL5 OE plasmid). B-CPAP cells and TPC-1 cells were transfected with CXCL5 OE plasmid or negative control (NC) plasmid by using Lipofectamine™ 3000 transfection reagent, and 48 hrs later, the mRNA and protein expression levels of CXCL5 were respectively determined with (a & d) real-time quantitative PCR and (b & e) western blot. Additionally, forty-eight hours post the transfection, CXCR2 antagonist SB225002 (5 μM) was used to treat PTC cells for another 24 hrs. (c & f) Cell viability was determined with CCK-8 as well (n = 3). (g-h & j-k) The levels of nuclear c-Jun, phosphorylated or total JNK1/2, and phosphorylated or total p38 were determined with western blot analysis. (i & l) Immunofluorescence assay using anti-c-Jun antibody was performed to probe the expression of c-Jun in PTC cells (bars, 50 μm). All data were expressed in mean ± standard deviation (SD) (n = 3). **P < 0.01; *P < 0.05. CXCL5, C-X-C motif chemokine ligand 5; CXCR2, C-X-C motif chemokine receptor 2; JNK, c-Jun N-terminal kinases

Next, B-CPAP cells and TPC-1 cells overexpressing CXCL5 were treated with 5 μM SB225002 for additional 24 hrs. We found that the overexpression of CXLC5-induced nuclear accumulation of c-jun and phosphorylation of JNK1/2 at ¹⁸³Thr/¹⁸⁵Tyr and p38 at ¹⁸⁰Thr/¹⁸²Tyr in B-CPAP cells and TPC-1 cells were suppressed by SB225002 (Figure 3(g-l)). Furthermore, JNK pathway inhibitor (SP600125; 10 μM) or p38 pathway inhibitor (SB203580; 20 μM) was used to treat PTC cells transfected with CXCL5 OE plasmid for 24 hrs. The expression levels of the proliferation marker PCNA and the essential cyclins E1 and B1 were analyzed with western blot. Additional data showed that the elevation in PCNA, cyclins E1 and B1 was partly suppressed by SP600125 or SB225002 (Figure 4). These results suggested that the pro-proliferative effects of CXCL5-CXCR2 axis at least involved JNK or p38 MAPK pathway activation.

Figure 4. Blockade of JNK or p38 pathway suppresses CXCL5 overexpression-induced cell proliferation.

Forty-eight hours post the transfection, 10 μM SP600125 (JNK pathway inhibitor) or 20 μM SB203580 (p38 pathway inhibitor) was used to treat PTC cells for another 24 hrs, and the expression levels of PCNA, cyclin E1 and B1 were determined with western blot assay. **P < 0.01; *P < 0.05. CXCL5, C-X-C motif chemokine ligand 5; JNK, c-Jun N-terminal kinases; PCNA, proliferating cell nuclear antigen

Forced overexpression of CXCL5 promotes the growth of B-CPAP xenografted tumors

Finally, the in vivo growth of B-CPAP cells was evaluated. As shown in Figure 5(a), the xenografted tumor formed by B-CPAP cells overexpressing CXCL5 was larger than that generated by NC plasmid transfected cells (Figure 5(a); left column). More PCNA-positive cells presented within the tumor when CXCL5 was overexpressed (Figure 5(a); middle and right columns). CXCL5 overexpression promoted the nuclear translocation of c-Jun and augmented the phosphorylation of p38 in tumor (Figure 5(b-c)). Furthermore, nude mice injected with CXCL5 OE B-CPAP cells were given CXCR2 antagonist SB225002, JNK pathway inhibitor SP600125 or p38 inhibitor SB203580 for four weeks. We found that blockade of CXCL5/CXCR2 axis with SB225002 restricted tumor growth without affecting tumor formation rate (Figure 5(d)). Similar anti-tumor effects were also observed in mice treated with SP600125 (Figure 5(e)) or p38 inhibitor (Figure 5(f)).

Figure 5. Forced overexpression of CXCL5 promotes the growth of B-CPAP xenografted tumors via activation of JNK or p38 pathway.

BCPAP cells were resuspended in 0.1 mL serum-free RPMI medium, mixed with 0.1 mL Metrigel and then injected into the axillary skin of nude mice. (a) The xenografted tumors were isolated and photographed on day 28 post the injection of cancer cells (left column). Immunofluorescence assay using PCNA antibody was conducted to probe PCNA-positive cells within the tumors (bars, 50 μm; middle and right columns). (b) The levels of nuclear c-Jun and phosphorylated or total p38 were determined with western blot analysis, and (c) calculated by comparing to H3 histone or β-actin. (d-f) Nude mice injected with CXCL5 OE BCPAP cells were given CXCR2 antagonist SB225002 (intraperitoneal injection of 10 mg/kg for three times a week), JNK pathway inhibitor SP600125 (intraperitoneal injection of 5 mg/kg for three times a week) or p38 inhibitor SB203580 (subcutaneous injection of 4 mg/kg for five times a week) for four weeks. The xenografted tumors were isolated and photographed on day 28 post the injection of cancer cells. The tumor volume was measured every three day before sacrificing on day 28. All data were expressed in mean ± standard error (n = 6). **P < 0.01; *P < 0.05. CXCL5, C-X-C motif chemokine ligand 5; PCNA, anti-proliferating cell nuclear antigen; OE, overexpression

Discussion

^BRAFV600E mutation and RET/PTC rearrangement are the two major oncogenic events that frequently occur in patients with PTCs ². To evaluate the role of CXCL5-CXCR2 axis in the proliferation of PTC cells, B-CPAP (BRAF^V600E mutation) and TPC-1 (RET/PTC rearrangement) cells were used as the in vitro cell models. The expression of CXCR2 on the surface of these two cell lines enables the delivery of CXCL-5-mediated cellular signals. We found that, regardless of the discrepancy in genetic background of the two PTC cell lines, both rhCXCL5 treatment and CXCL5 overexpression enhanced the cell proliferation.

Deregulated cell cycle is one of the most consistently observed events in human cancers, including PTC^14–16. A previous study from Kuo et al. showed that exogenous CXCL5 induced a decrease in CDKN1B and an increase in CDK4 in prostate cancer cells¹⁷, promoting us to decipher the role of CXCL5-CXCR2 in the cell cycle progress of PTC cells. Our data showed that rhCXCL5 treatment accelerated the transition from G1 phase to S phase (DNA synthesis) in B-CPAP and TPC-1 cells, which was accompanied with an upregulation in cyclins and CDKs, and a downregulation in CDK inhibitors. Cell cycle-associated proteins are considered as attractive targets for treating human cancers^18,19. Our work revealing that rhCXCL5 effectively triggers changes in cyclins and CDKs implies that PTC cells may be sensitive to CDK inhibitors. This assumption was then tested by treating PTC cells with ribociclib, a CDK4/CDK6 inhibitor. Interestingly, both B-CPAP and TPC-1 cells exposed to ribociclib displayed a reduced proliferation. Two recent studies have respectively demonstrated that ribociclib and palbociclib (another CDK4/CDK6 inhibitor) can inhibit the proliferation of anaplastic thyroid cancer cells^20,21. These previous studies together with our findings support the rationale for clinical application of the CDK4/6 inhibitor in patients with thyroid cancer.

In addition to treating PTC cells with rhCXCL5, eukaryotic expression vector carrying the complete CDS region of homo sapiens CXCL5 gene was constructed to mediate the overexpression of CXCL5 in PTC cells. Like exogenous rhCXCL5, genetic overexpression of CXCL5 also promoted the growth of B-CPAP and TPC-1 cells. The pro-proliferative effect of CXCL5 was attenuated when CXCR2 was selectively antagonized by SB225002 in vitro and in vivo. The growth-inhibiting effects induced by SB225002 in CXCL5-overexpressing B-CPAP cells were stronger than that in TPC-1 cells. Although this phenomenon may result from the discrepant basal states of the two cell lines, it at least suggests that PTC cells harboring BRAF^V600E mutation are more sensitive to CXCL5-CXCR2-induced alterations than cells carrying RET/PTC rearrangement.

In vivo, how CXCL5-CXCR2 axis influences the growth of cancer cells remains controversial. Zhou et al. demonstrated that overexpression of CXCL5 promoted the growth and lung metastasis of hepatocellular carcinoma cells in nude mice²², whilst Speetjens et al. found that re-expression of CXCL5 had opposite effects on colorectal cancer cells in Wag/Rij rats²³. Such controversial phenomena may result from the different immune backgrounds of the experimental animals. Thymic aplasia leads to immunodeficiency in nude mice. In order to avoid the interference from the host immune system, we and Zhou et al.²² chose to inject the cancer cells into nude mice. As a chemokine, CXCL5 can be released by various cell types, including the intratumor cells, and it links the outer microenvironment to the tumor and influences the tumorigenic progress. The intratumoral infiltration of CXCR2-expressing neutrophils, natural killer cells and T cells in response to CXCL5-CXCR2 interaction may trigger antitumor immune responses²³. Unlike the nude mice, Wag/Rij rats have intact immune system. The growth-inhibiting effects of CXCL5 on colorectal cancer cells in Wag/Rij rats may result from an enhanced antitumor immune attack. It is possible that when the infiltration of anti-tumor immune cells induced by CXCL5-CXCR2 interaction overwhelms the net pro-survival effects of this axis within tumor masses, tumor growth will be suppressed in vivo. To verify this hypothesis during thyroid tumorigenesis, further study is needed.

The role of p38 and JNK MAPK pathways in cancer cells is controversial^24,25. Both activation and deactivation of p38 and JNK pathways have been reported to contribute to aggressive tumor behaviors of thyroid cancer cells^26–28. Our current results indicated that genetic overexpression of CXCL5 activated p38 and JNK signaling pathways in PTC cells, which was blunted by CXCR2 antagonist SB225002. Blockade of p38 pathway with SB203580 and JNK pathway with SP600125 suppressed CXCL5 overexpression-evoked PTC cell growth in vitro and in vivo. It is generally considered that the function of MAPK pathways at least depends on the nature of stimuli. Our data thus suggest that the signals delivered by CXCL5-CXCR2 axis are more likely to promote cancer cell survival. Besides p38 and JNK pathways, CXCL5 also promotes the activation of ERK1/2 in colorectal cancer cells²⁹. ERK pathway is one of the three critical MAPK pathways, the activation of which can stabilize and allow the c-Fos protein to interact with c-Jun to form activating protein-1 (AP-1) complexes³⁰. Our results imply an involvement of CXCL5-CXCR2 in the formation of AP-1 dimers. Moreover, it should be noted that the regulation of p38 and JNK signaling by CXCL5/CXCR2 activity is linear. Gene or protein expression profiling using high throughput detection and screening method will provide more information to how this axis influences the development and progress of PTCs.

Exogenous CXCL5 binds to CXCR2 and triggers its intercellular internalization ^31,32 without significantly affecting its expression in cancer cells^29,33. CXCR2 is the only identified receptor for CXCL5 up to date⁴. It is plausible that cells lacking CXCR2 will not respond to addition of rhCXCL5 or CXCL5 overexpression. Several non-synonymous mutations have been identified in CXCR2 protein, including Arg153His, Arg236Cys, Arg248Gln and Ala249Leu^34–36. These mutations are located within the CXCR2 intracellular loop, a structure that is important for G-protein interactions and receptor activation^34,35. Lys327 residue is essential for the poly-ubiquitination of CXCR2³⁷. However, to our knowledge, none of these CXCR2 mutations have been reported to be associated with carcinogenesis of PTC or any other malignancies. As mutations within the key positions may alter the intracellular signaling transduction of CXCR2, there is a need to further investigate potential nucleotide polymorphisms of CXCR2 gene in PTC clinical samples, and analyze the association of the CXCR2 mutations with the clinical outcome of PTC patients.

In summary, our study demonstrates a growth-promoting effect of CXCL5-CXCR2 axis in PTC cells in vitro and in vivo. The activation of JNK and p38 pathways is involved in this process.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.