Targeting the CXCR4/CXCL12 axis in treating epithelial ovarian cancer

Abstract
Ovarian carcinoma is the most crucial and difficult target for available therapeutic treatments among gynecological malignancies, and great efforts are required to find an effective solution. Molecular studies showed that the chemokine stromal cell- derived factor (SDF)- 1 (also known as CXCL12) and its receptor, CXCR4, are key determinants of tumor initiation, progression, and metastasis in ovarian carcinomas. Hence, it is generally believed that blocking the CXCR4/CXCL12 pathway could serve as a potential therapy for patients with ovarian cancer. Herein, we investigated the role of the CXCR4/CXCL12 axis in regulating ovarian cancer progression. Using flow cytometry, a real-time PCR, and Western blot analyses, we showed that the chemokine receptor CXCR4 protein and mRNA were overexpressed in human epithelial ovarian cancer cell lines, and these were closely correlated with poor outcomes. Moreover, silencing CXCR4 by small hairpin RNA in HTB75 cells reduced cell proliferation, migration, and invasion and significantly reduced RhoA and Rac- 1/Cdc42 expressions, whereas, overexpression of CXCR4 in SKOV3 cells significantly increased cell migration and markedly increased RhoA, Rac- 1/Cdc42 levels. Silencing CXCR4 also led to decreased in vitro cytotoxicity of AMD3100, a specific antagonist of CXCR4, which exerts its effect upon CXCR4 expression. Remarkably, knockdown of CXCR4 in HTB75 cells led to a significantly decreased capability to form tumors in vivo, and the Ki67 proliferation index of xenograft tumors showed a dramatic reduction. Our results revealed that the CXCR4/CXCL12 pathway represents a promising therapeutic target for epithelial ovarian carcinoma.

Keywords: CXCR4, CXCL12, AMD3100, Ovarian Cancer, Proliferation

Introduction
Ovarian carcinoma (OC) is the most lethal malignancy of the female reproductive system and is responsible for high levels of cancer-related deaths in women1. The traditional therapeutic strategy is surgery plus first-line chemotherapy for high-grade OCs, but the recurrence rate after treatment remains high. The overall cure rate is <40% across all stages, and with current treatments the 5-year survival rate is <50%2. Among the four major histologic types, around 70% of all OCs are the “high-grade serous type”. Most ovarian high-grade serous carcinomas (HGSCs) are diagnosed at advanced stages, at which time, tumor cells have already metastasized3. Unfortunately, our understanding of the chemoresistance and recurrence of this highly lethal malignancy remains rather limited. Therefore, elucidating the unknown molecular mechanisms, and tumor progression and recurrence of HGSCs, and developing an effective therapeutic strategy have attracted much research interest for years.CXCL12 (C-X-C chemokine ligand 12), also known as stromal-derived factor (SDF)- 1, is secreted by tumor stromal fibroblasts in OC cells, as well as by mesothelial cells of humans and mice4,5. CXCR4 (also known as fusin), one of two known CXCL12 receptors, participates in the early stage of malignant transformation of OCs6. CXCR4 overexpression was detected in many different types of human cancer7, whereas the highly expressed CXCL12 is considered a crucial indicator of metastasis in various types of tumors8,9. Moreover, CXCL12 signaling through CXCR4 enhances the proliferation, invasion, and metastasis of tumor cells, which causes high aggressiveness in human malignances, including OC10. Aprevious study revealed that tumor progression is influenced by the “survival niche”, a microenvironment which allows CXCL12 to recruit tumor stem cells or immunosuppressive cells to become a malignant mesothelioma, indicating that CXCR4/CXCL12 signaling might be a potential target for therapeutic purposes11.

AMD3100 (Plerixafor), abicyclam derivative, is a specific antagonist of CXCR4 which mobilizes progenitor cells between the bone marrow and the peripheral blood 12. Mechanistically, AMD3100 significantly inhibits signaling of phosphatidylinositol 3-kinase (PI3K)/Akt phosphorylation and extracellular signal-regulated kinase (ERK) pathways and enhances poly (ADP-ribose) polymerase (PARP) cleavage13,14. Furthermore, AMD3100 was reported to be a therapeutic regimen for reducing metastasis in various types of cancer15-18, and its effects on downstream pathways following silencing of the CXCR4/CXCL12 axis were investigated19. Although studies showed that blocking the CXCR4/CXCL12 axis by either AMD3100 or CXCR4 short hairpin (sh)RNA can suppress cell growth in some cancers19,20, little is known about the effects of inhibition of the CXCL12/CXCR4 axis target in OCs, and its in vivo therapeutic efficiencies need to be investigated.In this present study, we demonstrated that two human epithelial OC cell lines, HTB75 (Caov3) and OVCAR3, expressed high levels of CXCR4 compared to two other OC lines, SKOV3 and TOV112D, which showed low CXCR4 expression. Moreover, the former two cell lines were specifically sensitive to AMD3100 treatment through targeted inhibition of the CXCR4/CXCL12 axis. Silencing of CXCR4 in HTB75 and OVCAR3 cells significantly suppressed cell migration and invasion, whereas overexpression of CXCR4 in SKOV3 cells significantly increased the cell migratory ability in vitro. For the in vivo study, knockdown of CXCR4 by shRNA in HTB75 cells led to significant inhibition of tumor cell growth in mouse xenografts. Furthermore, tumor tissue sections from CXCR4-shRNA-transfected HTB75 mouse xenografts showed substantial reductions in expression of the Ki67 proliferation marker. Our results suggest that CXCR4 blockade is a potential treatment target in OC.

Results
Expressions of CXCR4 and CXCL12 in OCs
We detected CXCR4 protein and mRNA expressions in four different OC cell lines (TOV112D, OVCAR3, SKOV3, and HTB75) using flow cytometry, Western blotting, and RT-PCR analyses. The level of CXCR4 was highest in HTB75 cells among the four OC cell lines according to the flow cytometric analysis (Fig. 1A, B). The Western blot analysis revealed CXCR4 overexpression in both HTB75 and OVCAR3 cells but not in TOV112D or SKOV3 cells (Fig. 1D). Likewise, CXCR4 messenger (m)RNA was highest in the HTB75 cell line (Fig. 1C). We then chose these four cells for further experiments in this study. Using an ELISA, we showed that the amount of the CXCL12 chemokine released into harvested culture medium was insignificant for all cell lines (Fig. 1E).CXCR4/CXCL12 signaling promotes epithelial OC growth It was hypothesized that the CXCL12 chemokinebinds to its receptor, CXCR4, to enhance the carcinogenetic potential in OC10,21. We therefore investigated if CXCL12 could increase tumor cell proliferation through the CXCR4/CXCL12 pathway. As shown in Figure 2A, HTB75 cell numbers significantly increased at different CXCL12 doses compared to the other three cell lines, which showed no significant difference in cell numbers following induction with CXCL12 (Fig. 2A). Moreover, when treated with the CXCR4 antagonist, AMD3100, for 48 h,both HTB75 and OVCAR3 cells revealed significantly reduced survival rates of 74% and 66%, respectively, whereas no significant difference in cell survival was seen in SKOV3 and TOV112D cells (Fig. 2B). These results suggest that the CXCL12 chemokine can trigger cell proliferation via CXCR4/CXCL12 signaling. Furthermore, blocking CXCR4/CXCL12 signaling by AMD3100 resulted in reduced tumor growth, particularly in epithelial OCs (HTB75 and OVCAR3) expressing CXCR4 (Fig. 2B).

CXCR4 knockdown in HTB75 and OVCAR3 decreases cytotoxicity elicited by AMD3100 To further confirm that the CXCR4/CXCL12 signaling can affect the progression of OCs, we knocked down the expression of CXCR4 by two shRNAs (clones B and C), and used one shNC (containing the LacZ reporter gene) as anegative control to transfect HTB75 cells. Results showed that CXCR4 mRNA expression levels were reduced by 82% and 40% in HTB75 cells transfected with shRNA-C (clone C) and shRNA-B (clone B), respectively, compared to untransfected HTB75 cells (P) (Fig. 3A). To investigate the efficiency of CXCR4 down regulated by shRNA, Western blotting was used to detect the CXCR4 protein level at 72 h after transfection. We found that shRNA-C and shRNA-B reduced CXCR4 expression by 85% and 50%, respectively (Fig. 3B), while shNC (clone Z) and parental (P) cells showed normal CXCR4 protein levels. Following CXCR4 silencing, cells were then treated with the CXCR4 antagonist, AMD3100, for 48 h and were then analyzed with an MTT assay. Results showed that shRNA-B- and -C-transfected cells became less sensitive to AMD3100 treatment compared to shNC-transfected control (LacZ) and parental (P) cells which were significantly (p<0.05) more sensitive to AMD3100 (Fig. 3C). We also conducted cell number counts over a 4-day period, and found that knockdown of CXCR4 in HTB75 cells reduced cell proliferation with different doses of CXCL12 of 0, 20, 50, and 100 ng/ml compared to the mock transfection control (HTB75-Z) (Fig. 3D). In addition,knockdown of CXCR4 in HTB75 (HTB75-C) cells also reduced levels of RhoA and Rac1/Cdc42, which likely contribute to the invasiveness of cancer cells, whereas, RhoA and Rac1/Cdc42 levels Supervivencia libre de enfermedad had recovered following 72 h of incubation with CXCL12 stimulation compared to control HTB75-Z cells (Fig. 3E). These results confirmed that CXCR4 levels could affect cell proliferation by targeting the CXCR4/CXCL12 pathway in vitro. Similar results were also shown in another cell line (OVCAR3) using the same CXCR4-knockdown approach (Supplemental Fig. S3).

Inhibition of tumor cell migration and invasion through downregulating CXCR4 in vitro Compared to CXCR4-overexpressing HTB75 cells, downregulation of CXCR4 by CXCR4-shRNA led to a significant decrease (p<0.05) in the cell proliferation rate in clone C cells, which indicated that the ability of HTB75 cells to proliferate was closely associated with CXCR4 expression. More importantly, the decreases in cell migration and invasion in CXCR4-knockdown cells (clone C) were more dramatic upon CXCL12 stimulation (Fig. 4). As revealed by a wound-healing assay, CXCR4- knockdown cells (clone C) migrated significantly more slowly during an incubation period of 24~48 h (bar scale is 100 μm, p<0.001; Fig. 4A, B) compared to the other controls. Similar results were also obtained by performing wound-healing assays with and without stimulation with 100 ng/ml CXCL12 (Supplemental Fig. S1a, b). In the meantime, transwell assays also confirmed that HTB75 cells with CXCR4 downregulation showed a dramatic decrease in the cell invasion status compared to the other controls (bar scale is 50 μm; Fig. 4C, Supplemental Fig. S2A). Invasive cells were counted following 24 h of stimulation with 100 ng/ml CXCL12, where CXCR4 shRNA clone (C)-transfected cells showed significantly reduced cell numbers compared to mock-transfected (Z) and parental (P) cells (p<0.001; Fig. 4D, Supplemental Fig. S2B).Overexpression of CXCR4 in SKOV3 stimulates cell migration but has non- significant effect on cell proliferation in vitro To further investigate the role of CXCR4/CXCL12 axis involved in the OCs progression, we thus investigated the effects of CXCR4 overexpression in low CXCR4-expressing SKOV3 (atypical non-serous) cells.

CXCR4 mRNA (Fig. 5A) and protein (Fig. 5B) levels were markedly increased in SKOV3 and TOV112D (data not shown) cells transduced with a CXCR4 recombinant lentivirus. In the wound-healing assay, markedly increased migration was shown in SKOV3 cells overexpressing CXCR4 (CXCR4) during the incubation period of 16~24 h (bar scale is 100 μm, p<0.001; Fig. 5D) compared to the mock transduction (Mock) and parental (P) controls regardless of stimulation by CXCL12 (100 ng/ml). These results suggest that CXCR4 interacts with CXCL12 showed difference in SKOV3 cells, thus CXCL12 did not encourage the proliferation of SKOV3-CXCR4 cells. Similar results were also shown in the MTT cytotoxicity analysis of SKOV3-CXCR4 (CXCR4) cells not being sensitive to AMD3100 compared to the mock (Mock) and parental (P) controls (Fig. 5C).CXCR4-knockdown limits tumor growth in mouse xenografts via reducing the proliferative ability We next investigated the relationship between the CXCR4/CXCL12 axis and ovarian tumor growth in vivo using tumor xenografts in mice. The CXCR4 shRNA-knockdown clone (C), mock control shNC (Z), and parental (P) HTB75 cells were subcutaneously (sc) injected into NOD/SCID mice (n=4/group) at 107 cells/mouse, as shown in Figure 6A. After 8 weeks,two of four parental HTB75-injected mice had died due to an overgrown tumor mass (data not shown). Therefore, tumors dissected from mock-transfected (Ctrl shRNA-Z) HTB75 cells were used for comparison, and these showed a larger tumor size than that of tumors from mice treated with shRNA- C-transfected HTB75 cells (n=4/group), with respective mean tumor sizes of 1055.5 vs. 360 mm3 (p<0.01) (Fig. 6B-D). Additionally, individual tumor sizes from each xenograft mouse group are also shown (Fig. 6D). To further investigate the potential mechanism which blocked the CXCR4/CXCL12 pathway to the reduction of OC cells growth, we examined tumor cell proliferation by Ki67 IHC staining from dissected tumor tissues. Tumor sections from control mice (Ctrl shRNA-Z) showed a much-higher Ki67 index than those from CXCR4-shRNA (C) xenografts, in which Ki67+ cells were significantly reduced by 50% (Fig. 7A). Using ImmunoRatio software counting, the mean number of Ki67+ cells from Ctrl shRNA (Z) sections was higher than that from CXCR4 shRNA (C) sections, at 84.9% vs. 27.4%, respectively (Fig. 7B, C). This finding demonstrated that limiting CXCR4/CXCL12 signaling by knocking down CXCR4 can lead to reduced tumor cell proliferation and hence reduced OC growth.

Discussion
The SDF1 (CXCL12)/CXCR4 pathway plays a critical role in tumor invasion and metastasis22, which are corrected with the high mortality inpatients with OCs10. Our study attempted to investigate the role of CXCR4 in OC progression by RNA interference and protein overexpression and has provided in vitro evidence that the CXCR4/CXCL12 pathway is functionally linked to OC cell proliferation, migration, and invasion. Notably, shRNA knockdown of CXCR4 reduced in vivo tumor growth. Our results highly recommend that the CXCR4/CXCL12 axis can serve as a useful therapeutic target for treating epithelial OC.It was reported that high CXCR4 levels are correlated with a poor prognoses in many cancers23-25, especially in HGSC, in which only CXCR4 was detected among 14 chemokine receptors checking10. Herein, we analyzed mRNA and protein expression levels of CXCR4 in four malignant OC cell Student remediation lines with two histotypes,atypical non-serous cell (CCC/ENOCa) and HGSC. Our results revealed that CXCR4 was mostly overexpressed in HGSC, in which the HTB75 (CAOV3) cells showed the highest level of CXCR4 expression (Fig. 1)21. Hence the HTB75 cell line was chosen for a further functional study to elucidate the roles of CXCR4 and CXCL12 and their interactions in the CXCR4/CXCL12 pathway for regulating cell proliferation. We showed that CXCL12 stimulation led to increased cell proliferation in the CXCR4- overexpressing HTB75 cell line, but not in cells with low CXCR4 expression. Moreover, the CXCR4 antagonist, AMD3100, inhibited cell proliferation, being significantly more dramatic in the two cell lines, HTB75 and OVCAR3,which had higher CXCR4 expression than did the SKOV3 cell line, which had a negligible level of CXCR4 expression (Fig. 2). These findings suggest that CXCR4 plays an important role in tumor cell proliferation, and that the effect of CXCR4 was sensitive to an antagonist.

The crucial role of CXCR4 signaling in cell proliferation was further supported by shRNA targeting CXCR4 to disrupt the CXCL12/CXCR4 signal transduction pathway in HTB75 and OVCAR3 cells (Fig. 3, Supplemental Fig. S3). CXCR4 silencing in HTB75 cells led to reduced cell growth as well as reduction of RhoA, Rac- 1/Cdc42 molecules that involved in the CXCR4-CXCL12/AKT axis, which were found to regulate cell invasion and tumor metastasis26,27. On the other hand, CXCR4-silenced HTB75 cells were less sensitive to AMD3100-targeted cytotoxicity (Fig. 3) compared to the controls. These findings were consistent with described observations that AMD3100 exerts cytotoxicity against cells with CXCR4 expression. More importantly, tumor invasion and metastasis are the main factors responsible for poor prognoses in OCs. To investigate the roles of the CXCR4/CXCL12 axis in these critical processes, we examined CXCL12-induced CXCR4 signaling molecules including RhoA and Rac- 1/Cdc42 in CXCR4-expressing cells and CXCR4-knockdown cells. In addition to cell proliferation, HTB75 cells highly sensitive to CXCL12 stimulation, showing increases in tumor cell migration,invasion through Matrigel, and recovery of RhoA and Rac- 1/Cdc42 levels (Figs. 2, 3, 4, Supplemental Figs. S1, S2). Similar to AMD3100, CXCR4-knockdown HTB75 cells also did not respond to CXCL12 stimulation. Similar results were also shown in another HGSC cell line (OVCAR3) using the same CXCR4-knockdown investigation (Supplemental Fig. S3). These results indicate that silencing of the CXCR4/CXCL12 signaling reduced OC growth and metastasis, and reduced sensitivity to AMD3100 cytotoxicity and CXCL12 stimulation. Hence, CXCR4 could be a potential selective target for OC patients’ therapy, particularly the HGSC type. Usually, CXCR4 is either lacking or only weakly presented in benign lesions and primary tumors, whereas CXCR4 positivity was correlated with metastatic tumors28. In OC, we also found that CXCR4 was over-expressed in HGSCs rather than in other histotypes.

Therefore, in contrast to the CXCR4-knockdown investigation in HGSCs, HTB75 and OVCAR3 cells, in vitro experiments also applied CXCR4-overexpression in atypical non-serous cell lines, i.e., TOV112D and SKOV3 cells. Accordingly, our results also showed that downstream molecules, RhoA and Rac- 1/Cdc42, of the AKT/ ERK pathways significantly increased in SKOV3 cells overexpressing CXCR4 (CXCR4) (Fig. 5). Therefore, the CXCR4/CXCL12 pathway seems to promote SKOV3-CXCR4 (CXCR4) cell migration by activating the AKT and ERK signaling pathways. Nevertheless, CXCL12 failed to promote the migration of SKOV3 cells overexpressing CXCR4 (CXCR4) probably due to the co-receptor for CXCR4 which was reported by Zhu et al.28It was also demonstrated that SKOV3-CXCR4 (CXCR4) cells were not sensitive to AMD3100 in the MTT analysis (Fig. 5).For further validation, we silenced CXCR4 by shRNA to investigate the effects of CXCR4-knockdown on in vivo tumor growth in NOD/SCID mice. Tumor growth of xenografts decreased with blockade of CXCR4 (Fig. 6) in mice, in accordance with previous reports which demonstrated that activation of the CXCL12/CXCR4 axis induced tumor cell growth, migration, and invasion29,30. Nevertheless, the mechanism by which the CXCL12/CXCR4 pathway regulates tumorigenesis remains poorly understood, particularly in the in vivo model. We showed that CXCR4-knockdown led to reduced tumor growth accompanied by a reduction LGH447 in Ki67+ cells in xenograft tumor sections compared to control tumor sections (Fig. 7). These results demonstrated that the SDF- 1/CXCR4 pathway induces cell proliferation in OC cells. Indeed, previous reports revealed that silencing of CXCR4 blocked Wnt/β -catenin signaling31-33and ERK1/229pathways in cancers; both pathways were correlated with cell proliferation34and tumor progression35,36.

In conclusion, we demonstrated that CXCR4 was significantly correlated with the progression of OCs. Our results confirmed that silencing of CXCR4 to limit SDF- 1/CXCR4 signaling greatly suppressed tumor growth by reducing cell proliferation. Our findings highlight the critical role of the SDF- 1/CXCR4 pathway in OC progression and imply that CXCR4 represents a potential and selective target for OCs treatment. The human HTB75 (Caov3), OVCAR3, SKOV3, and TOV112D OC cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and were cultured at 37 °C in ahumidified incubator with 5% CO2 in Dulbecco’s modified Eagle medium (DMEM) or RPMI- 1640 medium as previously described21. OC cells were cultured using standard protocols, and CXCR4 expression was determined with a FACS CantoTM II cell analyzer (Becton Dickinson Biosciences,San Jose, CA, USA) using an anti-human CXCR4-PerCP mAb (R/D Systems, MN, USA).The gating strategy is shown in Figure 1, and fractions of CXCR4+ cells among gated lived cells were calculated based on dot plot diagrams.RT-qPCR (Quantitative real-time reverse transcription polymerase chain reaction) analysis Cancer cell lines and shRNA-knockdown cells were seeded in six-well plates and incubated at 37 °C with 5% CO2. Cells were used for RNA isolation.

Total RNA was isolated using the Trizol reagent (Life Technologies, NY, USA). The RNA quantity was analyzed with a NanoDrop 2000C spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Two micrograms of total RNA was subjected to RT using a Quant script complementary (c)DNA synthesis kit (Bio-Rad, Richmond, CA, USA). Briefly, 3 μl of synthesized cDNA was used to perform a real-time RT-PCR using the premixiQ SYBR green QRT-PCR kit (Bio-Rad) with human CXCR4-specific oligonucleotide primers (forward: 5’-CCACCATCTACTCCATCATCTTC-3’ sense and reverse: 5’-ACTTGTCCGTCATGCTTCTC-3’ antisense; Integrated DNA Technologies, Coralville, IA, USA). The crossing threshold (Ct) value of the transcripts assessed by the RT-PCR was normalized with human GAPDH gene. Changes in (m)RNA expression levels are expressed as multiples of change relative to the control ± standard deviation (SD). Each sample was run in triplicate. OC cells, CXCR4 shRNA knockdown, and CXCR4-overexpressing cells (105 cells/well) were seeded and collected for western blotting. Cell lysates were prepared as previously described37. Primary antibodies against CXCR4, RhoA, Rac- 1/Cdc42, and α-tubulin were purchased from Novus Biologicals (Littleton, CO, USA), and α- tubulin was taken as a loading control. The appropriate secondary antibodies were used to detect proteins as previously described37. An ImageQuantTM LAS 4000 analyzer (GE Healthcare Life Science, Pittsburgh, PA, USA) was used to the protein expression levels quantification. Cancer cell lines were seeded at a density of 105 cells/well on six-well plates supplemented with DMEM/F12 and/or RPMI medium with 10% FBS (fetal bovine serum). After 24, 48, and 72 h of incubation, the medium was collected and analyzed with a CXCL12/SDF- 1 DuoSet® ELISA kit (R/D Systems), ELISA assays were according to the manufacturer’sprotocol.

MTT (Sigma-Aldrich, St. Louis, MO, USA) cytotoxicityassays were performed triplicated in 96-well plates. Tumor cells (104 cells per well) were plated in triplicate wells with 100 ng/ml CXCL12 supplementation. After 24 h, AMD3100 was added at the indicated dosage of 0, 1, 10, 100, or 200 (μg/ml) and treated for 72 h. The MTT assay was quantified using the absorbance at 570 nm. In total, 3×104 cells were seeded incomplete medium in 24-well plates. After being cultured for 24 h, CXCL12 (R/D Systems) was added to cells at different concentrations (0, 20, 50, and 100 ng/ml). Cell growth was analyzed at 0, 2, 4, and 5 days by counting cell numbers. Two CXCR4 shRNA clones in different regions of CXCR4 (GenBank: NM_003467) were purchased from the National RNAi Core Facility (Taipei, Taiwan) and were used to knock down CXCR4 in this study. The shCXCR4-B clone-targeted sequence was 5’-AGATAACTACACCGAGGAAAT-3’, the shCXCR4-C clone-targeted sequence was 5’ -TCCTGTCCTGCTATTGCATTA-3’, and a negative control LacZ gene that did not targeted to any region of the human genome was used. These shRNA fragments were cloned into the pLKO.1 lentiviral vector. A sequencing analysis was used to verify all the constructs. Stable transfection was performed using PolyJetTM (SignaGen, Rockville, MD, USA) according to the manufacturer’s instructions. CXCR4 (GeneBank: AF005058.1) gene was used in a recombinant lentivirus expressing vector, pLAS2w.Pbsd (Invitrogen), construction. SKOV3 and TOV112D cells at a density of 105 cells/well were seeded in 24-well plates. After overnight incubation, 2 ml of fresh media containing 8 μg/ml polybrene (Sigma, St. Louis, MO, USA) and the recombinant lentivirus were add. The green fluorescent protein (GFP) expressing was used as the indicator for lentiviral transduction efficiency determination. Obvious GFP expressions were shown 48 h after transduction and reached a peak level at 72 h following transduction. SKOV3 and TOV112D cells with CXCR4-overexpression were then used for the RT-PCR, Western blot, cytotoxicity, and migration assays. Initially, 1.3×105 HTB75 cells per well were seeded in 24-well plates.

After cells had reached confluence, the surface area of a plate was lightly scraped with a sterile micropipette tip (Sorenson Bioscience, Salt Lake City, UT, USA). Floating cells were carefully washed with 1x PBS (phosphate-buffered saline). An inverted optical microscope (Carl Zeiss, Jena, Germany) was used to detect the healing of the wound at 0, 24, and 48 h. ImageJ software was used to measure the wound area. The membranes of 24-well transwell chambers (Merck Millipore, Darmstadt, Germany) coated with Matrigel were used for the invasion assays. Parental HTB75 (P) and transfected HTB75 cells (clones Z and C) were seeded in the upper chamber which contained DMEM without serum, whereas DMEM combined with 10% FBS,which acted as a chemoattractant, was added to the lower chamber.Following 72 h of incubation, cells that had not migrated across the membrane were removed with a cotton swab, while cells that had adhered to the lower surface of the inserts were stained using 0.1% crystal violet for 20 min. Briefly, complete filters were washed twice with water before the observation. All mice were reared according to National Institutes of Health guidelines for animal care and guidelines of the Animal Center at National Taiwan University College of Medicine (NTUMC), and all animals were maintained in a SPF (specific pathogen- free) facility. Six to 8-week-old female NOD/SCID (non-obese diabetic mice with severe combined immunodeficiency disease) mice were used in this study. In total,107 HTB75 cells transfected with either sh-CXCR4 (clone C) or the negative control (clone LacZ) were subcutaneously (s.c.) injected into the groin of each mouse. Accordingly, all animal protocols were approved by the Animal Care and Use Committee of NTUMC. Solid tumors were excised from sacrificed NSG mice under sterile conditions at the indicated times after treatment. For IHC analyses, serial tissue sections (4 μm thick)were prepared from formalin-fixed tumor samples and mounted on glass slides. After rehydration, sample sections were stained with mAbs against human Ki67. After 10 min of DAB (3,3’-diaminobenzidine) incubation, tumors were counterstained with hematoxylin. Images were acquired using an Olympus microscope (BX50; Tokyo, Japan). Results arepresented as the mean ± standard deviation (SD). Statistical significance of the difference between means was measured using a two-tailed Student’s t-test. Significant differences were considered at * p<0.05, ** p<0.01, and *** p<0.001.