Tissue samples of 83 BCa patients were obtained from the Department of Urology, First Affiliated Hospital of Xi’an Jiaotong University. The samples comprised 52 males and 31 females (age 39–78 years, average age 63.7±3.4). Briefly, the samples were fixed within 4% formalin, paraffin-embedded and 4-μm thick sections used for subsequent histological analysis conducted at the Department of Pathology, First Affiliated Hospital of Xi’an Jiaotong University.

IHC was conducted using a Dako Autostainer Plus system (Dako Corp., Carpinteria, CA, USA). Tissue sections were de-paraffinized, rehydrated and subjected to 5-min pressure-cooker antigen retrieval methods, 15-min endogenous enzyme block, 60-min primary antibody incubation and 30-min Dako Cytomation EnVision-HRP reagent incubation for rabbit antibodies. Signals were detected by adding substrate hydrogen peroxide using DAB (diaminobenzidine) as a chromogen followed by hematoxylin counterstaining. Negative control slices were prepared by omitting the primary antibodies. Stained (brown) cells were quantified by (number of positive cells × 100 / total number of cells) in 10 random microscopic (x400) fields in each slice.

Cells were harvested at 80% confluency, and washed with cold PBS three times. Total cellular protein lysates were prepared with RIPA buffer (50 mM Tris (pH 8.0), 150 mM NaCl, 0.1% SDS, 1% NP40 and 0.5% sodium deoxycholate) containing proteinase inhibitors [1% inhibitor cocktail and 1 mM PMSF, both from Sigma, (St. Louis, MO, USA)]. Protein (30 μg) was separated on 10% SDS-PAGE gels and transferred to nitrocellulose membranes. The membranes were blocked at room temperature for 1 h with 5% skim milk in Tris-buffered saline (pH 7.6, TBS). Polyclonal primary antibodies were applied at different dilutions (Table I) in 5% skim milk in TBS at 4°C overnight, followed by TBST (with Tween-20) washes. Membranes were incubated with fluorescent secondary antibodies (Licor, Rockford, IL, USA) coupled to the first antibody at room temperature in the dark for 1 h, followed by being TBST washes, dried with neutral absorbent paper and scanned by Odyssey Detection system (Licor). MG-132 (Sigma-Aldrich, USA) was used to inhibit the proteasome-dependent degradation when necessary (10 μM, 4 h before the protein harvest). GAPDH was used as loading control (for total cell fraction).

Cellular total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and quantitated by absorbance at 260 nm. RNA (2 μg) was reverse transcribed using Revert Aid™ First Strand cDNA Synthesis kit (MBI Fermentas, StLeon-Rot, Germany) according to the manufacturer’s protocol. For real-time PCR, we used the SYBRR Premix Ex Taq™ II system (Takara Biotechnology Co., Ltd., Dalian, China) and the Bio-Rad CFX96™ Real-time system (Bio-Rad, CA, USA). SYBRR Premix Ex Taq II (12.5 μl), 1 μl primer (10 μM, primers, Table II), 200 ng cDNA and 9.5 μl distilled and deionized water were mixed together, followed by two stage, pre-degeneration for 95°C, 30 sec, one repeat; and PCR reaction, 95°C 5 sec followed by 60°C, 30 sec, 30 repeats; and the third stage as dissociation, 95°C, 15 sec followed by 60°C, 30 sec, and another 95°C, 15 sec. GAPDH was used as the loading control.

Human bladder cancer cell lines T24, J82, 253J, 5637, UM-UC-2, RT4, TCCSUP were obtained from ATCC (American Type Culture Collection, Manassas, VA, USA) and cultured in DMEM (for T24, J82, 253J, RT4, and UM-UC-2) or 1640 (for 5637) supplemented by 10% FBS (Invitrogen). Cells were cultured in an atmosphere with 5% CO₂ at 37°C (incubators: Thermo Scientific, Germany).

In order to demonstrate the roles of ELR⁺ CXC-chemokine-stimulation in BCa cell lines, we initially used all the ELR⁺ CXC-chemokines (including CXCL1, CXCL2, CXCL3, CXCL5, CXCL6, CXCL7, and CXCL8, Sigma-Aldrich) to stimulate BCa cell lines for 4 h prior to conducting subsequent studies.

The ELISA kits for detecting CXCL5 in BCa cell line medium were obtained from Sigma (RAB0130, Sigma-Aldrich). Briefly, ELISA was performed in a blinded manner with triplicate of each sample strictly following the protocol.

Migration and invasion were tested by Boyden chamber assay, obtained from Millipore (Millipore, Switzerland). For migration assays, 0.2 ml FBS-free DMEM medium suspension with 10,000 cells was added to the upper chamber in the 24-well plate, and 0.8 ml FBS-free DMEM was added to the lower chamber. Following 12-h incubation, the chambers were washed with PBS (pH 7.4) three times to remove the cells in the upper chamber and were fixed with 4% formalin for 15 min, then stained with crystal violet (0.01% in the ethanol) for 25 min followed by washing three times with PBS. The cells were counted using an inverted microscope, and five visions were randomly taken in the ×200 magnification, and the average number of cells analyzed. For the invasion assay, the cell suspension (10,000 cells/well) in the upper chamber contained 0.2 ml mixture of FBS-free DMEM/Matrigel at an 8/1 ratio (Matrigel, Sigma, USA). Cells were incubated for 36 h and the rest of the protocol conducted in a similar manner as the migration assay.

SiRNA transfections (for sequence see Table II) were used to silence the expression of CXCL5 in 5637 BCa cell line, and pCEFL-CXCL5/Vec plasmid to overexpress CXCL5 in UM-UC-2. SiRNA for knocking down CXCR2 and Akt (both supplied by Santa Cruz Biotechnology) in 5637 and UM-UC-2 were used to further understand the roles of these two genes. Briefly, the mixture of the CXCR2/Akt1 siRNA (20 mM) and Lipofectamine 2000 (Invitrogen 1:1) were added into the cell medium for transfection over 24 h. Fresh medium with 5% FBS replaced the transfection medium and cells were incubated for no more than 72-h post-transfection. As a control of CXCR2 siRNA, the scramble sequence (Invitrogen) was transfected by the same method.

PT3-myr-Akt/Vec (Addgene plasmid 31789, http://www.addgene.org) was used to demonstrate the elevated phosphorylation of AKT (both T308 and S473 sites). PT3-myr-Akt/Vec and pCEFL-CXCL5/Vec were transfetced into target cell lines to generate stable clones, which had a high stable expression of p-Akt or CXCL5 compared with vector controls, respectively. Lipofectamine™ 2000 (Life Technologies, USA) was used to transfect plasmids into the target cell lines strictly accordance with the manufacturer’s protocol. Western blot and real-time-PCR analyses were used to monitor transfection efficiency.

Parallel to the CXCR2 siRNA experiments, SB225002, the specific inhibitor of CXCR2, was used to functionally inhibit CXCR2 at a final concentration of 4 μM.

5637 and UM-UC-2 cells were plated on 24-well plates to 50–70% confluency and then stimulated with CXCL5 (10 nM) for 4 h. BrdU was added into the medium (3 μg/ml). Cells were incubated for an additional 4 h and then rinsed 3× with PBS over 10 min to remove residual free BrdU. Cells were then fixed by 4% paraformaldehyde for 45 min, followed by rinsing 5× with PBS over 20 min. Triton X-100 (0.1%) was used to permeabilize the cell membrane for 15 min and 2N HCl added for 25 min to separate DNA into single strands to allow primary antibody access to the incorporated BrdU. Cells were then rinsed 3× with PBS over 10 min and non-specific epitopes blocked by 10% BSA for 20 min. Anti-BrdU antibody (1:200) in 10% BSA was added and incubated overnight in 4°C. Cells were rinsed 5× with PBS, followed by incubation with TRTIC-labeled secondary antibody for 1 h in RT, and finally rinsed 3× with PBS to remove the free antibody. The fluorescence intensity of TRITC was monitored by SuperMicro Orifice Plate Spectrophotometer (BioTek, USA) in 547 nm.

ANOVA test was used to analyze the statistical discrepancy in ≥3 groups. Student’s t-test was used to detect any statistically significant difference between 2 groups. p-values ≤0.05 were considered statistically significant.

Expression of CXCL5 in BCa tissue has previously been reported (22), indicating the expression of this chemokine is elevated in positive correlation to BCa grade. Our IHC analysis of tumor samples derived from 83 BCa patients confirm these previous findings (Fig. 1A). Although the discrepancy between grades II and III is not so significant, herein, we focus on the discrepancy between grade III versus grade I, and tumor tissue versus normal control. This increased expression of CXCL5 in BCa tissue suggested that CXCL5 likely plays a vital role in BCa progression.

In order to further understand the role of CXCL5 in BCa, we assessed the expression of CXCL5 in BCa cell lines derived from different grades of BCa tissues using western blot and real-time PCR assays. As presented in Fig. 1B (top panel), among all the BCa cell lines, 5637 exhibited the highest CXCL5 expression, while UM-UC-2 had the lowest expression of CXCL5. These results were replicated in real-time PCR analysis (Fig. 1D, left), and ELISA analysis (Fig. 1C). Our data suggested that the expression of CXCL5 in medium grade BCa varied, but 5637 and UM-UC-2 were consistently the highest and lowest expressors, respectively. As previously reported, CXCL5 executes its functionality by binding to, and activating its receptor CXCR2. Therefore, we wanted to determine the level of expression of CXCR2 in BCa cell lines. As shown in Fig. 1B (bottom panel), we did not find a significant difference in the protein expression levels of CXCR2, although there was a difference at the mRNA level (Fig. 1D, right).

Thus far, we concluded that the expression of CXCL5 in BCa tissue was higher than normal tissue, and this expression was increased in positive correlation to the BCa grade. In addition, while the expression of CXCL5 differed in various BCa cell lines, the expression of CXCR2 was not significantly different in BCa cells.

A previous study reported that a knock-down in the expression of CXCL5 in BCa T24 cell resulted in decreased cell migration (22). In addition, CXCL5 has been shown to promote tumor cell migration and invasion in breast cancer, and non-small cell lung cancer (NSCLC) (29). However, it still unknown whether increased levels of CXCL5 contribute to migration and invasion of BCa cells directly. To address this, we used exogenous CXCL5 to stimulate BCa for 4 h and assessed migration and invasion. As shown in Fig. 2A, the capacity of both migration and invasion of the two cell lines was significantly enhanced after stimulation by exogenous CXCL5. This CXCL5-stimulatory effect on migration and invasion was more enhanced in UM-UC-2 cells than in 5637 cells, which is likely due to the lower basal expression of CXCL5 in UM-UC-2 cells. In addition, due to the different basal expression of CXCL5 in 5637 cells (high expressor cells) and UM-UC-2 cells (low expressors) (Fig. 1B and D), we used siRNA transfections to silence endogenous CXCL5 in 5637 cells, and plasmid transfections to overexpress CXCL5 in UM-UC-2 cells, and to demonstrate the role of CXCL5 in tumor cell migration and invasion. Fig. 2B shows the transfection efficiency of the siRNA knock-down in 5637 or overexpression in UM-UC-2 cells. As expected, this endogenous alteration of CXCL5 significantly decreased the migration and invasion of 5637 cells (Fig. 2C, top panel) or significantly increased this effect in UM-UC-2 cells (Fig. 2C, bottom panel). Moreover, this phenomenon was more pronounced in UM-UC-2 cells.

Many investigators have shown that angiogenic chemokines, which includes ELR⁺ chemokines, recruit endothelial cells which express the cognate receptors to these chemokines, which ultimately contributes to tumor cell proliferation and metastasis. We therefore investigated whether the increased migration and invasion we had observed thus far involved CXCR2. In order to demonstrate this point, we used siRNA transfection to silence CXCR2, or administration of SB225002, a specific CXCR2 inhibitor. As indicated in Fig. 3A, siRNA induced the decreased expression of CXCR2 at both the protein and mRNA level. Fig. 3B suggested that knocking down the expression of CXCR2 significantly attenuated the CXCL5-stimulatory effect of BCa tumor cell migration and invasion; with the basal migration and invasion also attenuated, especially in 5637 cells (top panel). As expected, SB225002 also significantly inhibited the basal (Fig. 3C) and CXCL5-stimulatory induced tumor cell migration and invasion (data not shown).

In BCa, MMP2 and MMP9 are reported to contribute to tumor cell migration/invasion in various ways. As we had found that CXCL5/CXCR2 axis induces enhanced capacity of migration and invasion in 5637 and UM-UC-2 cells, we assessed the expression of MMP2 and MMP9 in CXCL5-stimulated 5637 and UM-UC-2 cells in the presence or absence of SB225002. We found that CXCL5-stimulated 5637 and UM-UC-2 cells manifested higher expression of MMP2 and MMP9 compared to controls (Fig. 4A). Additionally, inhibition of CXCR2 by SB225002 (Fig. 4A) or siRNA targeting CXCR2 (data not shown) significantly inhibited the CXCL5-stimulatory induced expression of MMP2 and MMP9. We also assessed the expression of E-cadherin and vimentin, key markers of EMT and known to be important in tumor cell migration and invasion, and unexpectedly there was no significant difference in the expressions of either E-cadherin or vimentin under SB225002 treatment (Fig. 4A), or CXCR2-siRNA transfection (data not shown).

Elevated expression of MMPs can be monitored by various signaling pathways, including MAPK/ERK, PI3K/AKT and NF-κB pathway, among which, PI3K/AKT signaling is of interest to us. PI3K/AKT is an evolutionary-conserved pathway that plays irreplaceable and critical roles in tumor development, including tumor metastasis, proliferation, angiogenesis, and cell cycle control. In BCa, PI3K/AKT signaling has been reported to drive tumor progression. Emerging of investigation has provided evidence that MMP2 and MMP9 can be induced in many signal pathways, of which the interesting one is the PI3K/AKT signaling. As we had found that 5637 and UM-UC-2 stimulated by CXCL5 led to significantly elevated expression of MMP2 and MMP9, therefore, we hypothesized that PI3K/AKT signaling may play a role in this process. Indeed, as indicated in Fig. 4A, CXCL5 stimulation of 5637 and UM-UC-2 cells manifested higher expression of p-AKT (Th308), indicating the activation of PI3K/AKT pathway, compared with the unstimulated controls. Moreover, this elevated phosphorylation of AKT was significantly inhibited in the presence of SB225002, and accompanied by decreased expression of MMP2 and MMP9 (Fig. 4A).

To further understand the role of AKT in CXCL5 stimulation-induced elevation of MMP2 and MMP9, we use siRNA transfections to knock down AKT expression in the two cell lines, and the siRNA interference led to decreased expression of AKT at the protein and mRNA levels indicated by western blot (Fig. 4B, top) and real-time PCR analyses (Fig. 4B; bottom). The silencing of AKT led to the attenuation of CXCL5 stimulation-induced expression of MMP2 and MMP9 (Fig. 4C), accompanied by a decreased ability of migration and invasion of these two BCa cell lines (Fig. 4D). However, this AKT silencing by siRNA still had no effect on the expression of E-cadherin and vimentin.

One of the important sequences of activating of PI3K/AKT signaling is the recruitment of AKT to the cell membrane to bind, via its PH domain, to PIP3 following exogenous stimulations. We had seen that 5637 and UM-UC-2 stimulated by CXCL5 resulted in the elevated activation of AKT (pAKT), accompanied by elevated expression of MMP2 and MMP9, accompanied by enhanced ability of migration and invasion (Fig. 4). In order to further dissect whether the elevated expression of MMP2 and MMP9 induced by CXCL5 stimulation involved the activation of PI3K/AKT signal, we used the plasmid-tagged myristoylated sequence, which can automatically bind to cell membrane independent of the stimulating signal, leading to the activation of PI3K/AKT signaling. As noted, pAKT was very low in parental 5637 and UM-UC-2 (Fig. 4), indicating the inhibition of this signaling in parental cells. However, transfection with pT3-myr-Akt-HA significantly induced pAKT (Fig. 5A) compared to vector controls. The basic expression of MMP2 and MMP9 were also increased in myr-Akt-transfected cells compared to vector controls (Fig. 5B). In addition, the expression of MMP2 and MMP9 was decreased in the presence of SB225002, whereas, this inhibitor had no effect on the expression of MMP2 and MMP9 in the myr-Akt-transfected cells. This suggests that the CXCL5-stimulation-induced elevation of MMP2 and MMP9 is dependent on the activation of PI3K/AKT signaling. The functional role of this transfection is shown in Fig. 5C.

ELR⁺ CXC-chemokines induce angiogenesis by contributing to endothelial cell recruitment, and proliferation, through interacting with CXCR2. As other ELR⁺ chemokines can also bind and activate CXCR2, we investigated whether other ELR⁺ CXC-chemokines could induce the enhanced migration and invasion seen thus far in the two BCa cell lines. We stimulated 5637 and UM-UC-2 cells with CXCL1, CXCL2, CXCL3, CXCL6, CXCL7, and CXCL8 in a similar manner as with CXCL5. Boyden chamber assay showed that there were significant discrepancies in the migration and invasion capacity of these two cell lines upon stimulation by other ELR⁺ chemokines. Only CXCL5 consistently enhanced migration and invasion of both 5637 and UM-UC-2 cells, as indicated in Fig. 6A.