Therapeutic Targeting of Tetraspanin8in Epithelial Ovarian Cancer Invasion and Metastasis
Running title: TSPAN8 as a therapeutic target in epithelial ovarian cancer
Chang Sik Park1,†,§, Taek-Keun Kim1,†, Han Gyul Kim2, Youn-Jae Kim3, Mee Hyun Jeoung1, Woo Ran Lee1, Nam Kyung Go1, Kyun Heo2,*, Sukmook Lee1,*
1Laboratory of Molecular Cancer Therapeutics, Scripps Korea Antibody Institute, 1 Gangwondaehak-gil, Chuncheon-si, Gangwon-do, 200-701, Korea; 2New Experimental Therapeutics Branch, Research Institute, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do, 410-769, Korea; 3Specific Organs Cancer Branch, Research Institute, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do, 410-769, Korea
†These authors contributed equally to this work.
*To whom correspondence should be addressed:
Dr. Sukmook Lee, Laboratory of Molecular Cancer Therapeutics, Scripps Korea Antibody Institute, Hyoja-2-dong, Chuncheon-si, Gangwon-do, 200-701, Korea, Tel: 82-33-250-8096; Fax: 82-33-250-8088; E-mail: Dr. Kyun Heo, New Experimental Therapeutics Branch, Research Institute, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do, 410-769, Korea, Tel: 82-31-920-2430; Fax: 82-31-920-2542; E-mail:
§Current address: Eco & Bio Convergence Team, Korea Institute of Ceramic Engineering and Technology, 101 Soho-ro, Jinju-si, Gyeongsangnam-do, 661-031, Korea
Abstract
Epithelial ovarian cancer (EOC) invasion and metastasis are complex phenomena that result from the coordinated action of many metastatic regulatorsand must be overcome to improve clinical outcomes for patients with these cancers. The identification of novel therapeutic targets is critical because of the limited success of current treatment regimens, particularly in advanced-stage ovarian cancers.In this study, we found that tetraspanin 8 (TSPAN8)isoverexpressed in about 52% (14/27) of EOC tissues and correlates with poor survival. Using siRNA-mediated TSPAN8 knockdown and a competition assay with purified TSPAN8 large extracellular loop (TSPAN8-LEL) protein, we identified TSPAN8-LEL as a key regulator of EOC cell invasion. Furthermore, monotherapy with TSPAN8-blocking antibody we developed shows that antibody-based modulation of TSPAN8-LEL can significantly reduce the incidence of EOC metastasis without severe toxicity in vivo. Finally, we demonstrated that the TSPAN8-blocking antibody promotesthe internalization and concomitant downregulation of cell surface TSPAN8. Collectively, our data suggest TSPAN8 as a potential novel therapeutic target in EOCs and antibodytargeting of TSPAN8 as an effective strategy for inhibiting invasion and metastasis of TSPAN8-expressing EOCs.
Keywords
TSPAN8, Therapeutic target, Epithelial ovarian cancer,Invasion,Metastasis
Introduction
Epithelial ovarian cancer (EOC) is the most common type of ovarian cancer and is the fifth leading cause of cancer-related deaths among women worldwide. This cancer arises from epithelial cells of the ovary, which are important for hormonal regulation of female reproduction (1). Because of a lack of characteristic symptoms and early detection strategies, most ovarian cancer patients are diagnosed at stages III and IV, after the cancer has already metastasized to other organs(2). The high mortality rate associated with this cancer is largely explained by the fact that the majority (around75%) of patients present at advanced stages with widely metastatic disease within the peritoneal cavity. These cancers grow rapidly, metastasize early, and have a very aggressive disease course (3). Thus, ovarian cancer invasion and metastasis still represent a major hurdle that must be overcome to improve patient outcomes.
Over the course of several decades, a number of chemotherapeutic agents that target DNA and microtubule structures have been developed for treating ovarian cancer. Despite their clinical efficacy, these agents are not targeted therapies and result in widespread cytotoxicity and side effects, including vomiting, diarrhea, hair loss, bleeding, and bone marrow suppression (4). Furthermore, the 5-year survival rates for stages III and IV ovarian cancer patients are extremely low, at 21.9% and 5.6%, respectively (5). Recently, bevacizumab, a humanized antibody targeting vascular endothelial growth factor (VEGF), received European Medicines Agency (EMA) approval as a first-line therapy for advanced ovarian cancer and recurrent, platinum-resistant ovarian cancer, in combination with chemotherapy. However, bevacizumab only increases progression-free survival by approximately 3–4 months, compared to standard chemotherapeutic agents, including paclitaxel, carboplatin, and gemcitabine (6). Accordingly, there remains a need to identify novel therapeutic targets for ovarian cancer therapy.
Tetraspanins are a family of small proteins that consist of four transmembrane domains, two extracellular domains, including the small and large extracellular loops (SEL and LEL), and three cytosolic domains. They form complexes by interacting with themselves and a variety of other transmembrane and cytosolic proteins, building a network of interactions referred to as tetraspanin webs or tetraspanin-enriched microdomains (TEMs) (7). These TEMs provide a signaling platform that is involved in many important cellular functions and malignant processes (8).
Tetraspanin 8 (TSPAN8), a member of the tetraspanin superfamily, is a tumor-associated antigen. It is highly overexpressed during the progression of colorectal, liver, pancreatic, and gastric cancers (9-11), and its increased expression promotes liver and lung metastasis (12-14). TSPAN8 may also act as an adaptor molecule, forming a complex with various membrane proteins, including CD151, EpCAM, claudin-7, E-cadherin, and CD44v6, that has been shown to promote cancer progression and metastasis (15, 16). However, the relevance and role of TSPAN8 are yet to be investigated in EOC.
In this study, we examined therelevance and function of TSPAN8 in EOC in vitro and in vivo. We showed that antibody targeting of TSPAN8 reduced EOC invasion and metastasis by internalizing and concomitant downregulation of cell surface TSPAN8. Thus, these findings indicate that targeting of TSPAN8 may potentially be effective against TSPAN8-expressing EOCs. Therefore,TSPAN8 is not only aprognostic biomarkerof EOCs,but also a therapeutic target for antibody therapy.
Results
Analysis of TSPAN8 Expression in EOCPatient Samples - We performed immunohistochemistry to compare TSPAN8 expression between normal ovarian and EOC tissues (Figures 1a and b). Normal ovarian tissue (n=4) lacked TSPAN8 expression, whereas high expression of TSPAN8(over 2-fold) was observed in about 52%(14/27) of EOC tissues. Furthermore, using EOC patient gene expression profiling data (GSE14764) obtained from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO), bioinformatics-based survival analysis also indicated that there was a statistically significant association between high TSPAN8 expression and poor survival.The5-year overall survival in the low-expression group (n=25)of the cohortwas 63.7% compared to 32.1% in the high-expression group (n=49)(Figure 1c). Collectively, these results suggest that TSPAN8 may be closely associated with EOCs.
Identification of TSPAN8-LELas a Key Target inEOC Invasion - We examined the role of TSPAN8 in EOC invasion at the molecular level by silencing TSPAN8 in the SK-OV3 EOC cell line using siRNA (Figure 2a). A tumor cell invasion assay demonstrated that TSPAN8 knockdown reduced SK-OV3 cell invasionby 50.71%withstatistical significance(Figure 2b).To investigate the functional relationship between the two TSPAN8 extracellular domains (TSPAN8-SEL and TSPAN8-LEL) in EOC invasion, we generated Fc-fusion proteins, including TSPAN1-LEL-Fc, TSPAN8-SEL-Fc, and TSPAN8-LEL-Fc, and determined their effects on SK-OV3 cell invasion (Figure 2c). With a competitive blocking experiment, we found thatSK-OV3 cell invasion was specifically and significantly inhibited by TSPAN8-LEL-Fc but not by TSPAN8-SEL-Fc; TSPAN1-LEL-Fc, used as a negative control, was also without any inhibitory effect, suggesting that TSPAN8-LEL-Fc may interrupt the TSPAN8-LEL-mediated interactions in SK-OV3 cell invasion.Taken together, these data suggest that the TSPAN8-LEL domain may play a keyrole in the regulation of TSPAN8-mediated EOC invasion.
Effect of TSPAN8-LEL IgG on Inhibition of EOC Invasionand Metastasis - Using phage display technology,we generated a novel human antibody specific to TSPAN8-LEL (TSPAN8-LEL IgG) that has a dissociation constant (Kd)of approximately 0.35nM(Supplementary Figures1a-d, available online). To investigate the effect of TSPAN8-LEL IgG on EOC invasion, we evaluated TSPAN8 expression in SNU-8, SNU-251, and SK-OV3 EOC cell lines. TSPAN8 was expressed specifically in all of these cell lines (Figure 3a). Here, HUVECs were used as negative cells that TSPAN8 does not express.Next, we performed Transwellinvasion assays using these cell lines in the absence or presence of TSPAN8-LEL IgG. The antibody significantly inhibited the invasion of all threeEOC cell lines to a similar extent(Figures3b-d), suggesting a generalized inhibitory effect of the TSPAN8-LEL antibody on EOC invasion, whereas bevacizumab does not significantly inhibited SK-OV3 cell invasion (Supplementary Figure 2, available online).
To investigate the effect of TSPAN8-LEL IgG on EOC metastasis, we established anEOC metastasis animal model. Control IgG or TSPAN8-LEL IgG was then injected intravenously twice weekly, starting from one day prior to, and continuing for 42 days after, SK-OV3-luc cell injection (Figure4a). Metastasis was monitored using bioluminescence imaging (Figure4b). The incidence of cell metastasis was determined as the number of mice with a detectable luminescence signal in removed organs, including the ovary, pancreas, colon, heart, liver, spleen, and kidney. We found that in the TSPAN8-LEL IgG-treated group (15/30), the SK-OV3-luc cell metastasis was observed in 15 of 30 mice injected with the SK-OV3 cells, whereas in the control IgG-treated group (24/31), the cell metastasis was observed in 24 of 31 mice injected with the SK-OV3 cells. The results indicated that the incidence of SK-OV3-luc cell metastasis was suppressedsignificantly by approximately 35%, with a single dose in the TSPAN8-LEL IgG-treated group, compared with the control IgG-treated group(Figure4c). These results suggest that the targeting of TSPAN8may be effective in thesuppression ofEOCmetastasisin vivo.
Influence of TSPAN8-LEL IgG on In Vitro or In VivoToxicity - To evaluate the in vitro cytotoxicity of TSPAN8-LEL IgG, we determined the viability of HUVECs and TSPAN8-overexpressing COS-7 cells after treatment with TSPAN8-LEL IgG. We found that TSPAN8-LEL IgG had no cytotoxic effect onall of these cells, whereas 5-fluorouracil (5-FU) significantly reduced the viability of HUVECsand TSPAN8-overexpressing COS-7 cells(Figure 5a and b). Here, TSPAN8-overexpressing COS-7 cells were representative of other TSPAN8-expressing cells and were used to further confirm that the TSPAN8-LEL IgGs had little effect on the viability of other TSPAN8-expressing cells.We also evaluated HUVEC morphology in the absence or presence of TSPAN8-LEL IgG using immunocytochemistry. TSPAN8-LEL IgG did not alter the morphology of HUVECs (Figure5c). To investigate the effect of TSPAN8-LEL IgG on endothelial cell activation—an initial inflammatory response to harmful stimuli—we treated HUVECs with TSPAN8-LEL IgG and monitored HUVEC activation by measuring the expression of endothelial cell activation markers, including vascular cell adhesion molecule-1 (VCAM-1) and intercellular cell adhesion molecule-1 (ICAM-1). We used human tumor necrosis factor- (hTNF) as a positive control for endothelial cell activation. TSPAN8-LEL IgG had little effect on HUVEC activation, whereas hTNF, as expected, induced HUVEC activation (Figure5d).
To evaluate the in vivo toxicity of the antibody, we performed our immunohistochemistry-based tissue cross-reactivity study and found that TSPAN8-LEL IgG specifically bound to ovarian cancer tissues but had weak or no affinity for normal ovarian or other tissue types (Supplementary Figures 3a and b, available online). We also administered control IgG or TSPAN8-LEL IgG into mice via intravenous injection and then monitored the liver and kidney function and body weight both prior to, and 42 days after, antibody injection. Liver function was determined by measuring serum concentrations of glutamic-oxaloacetic transaminase (GOT), glutamic pyruvic transaminase (GPT), and total bilirubin (TBIL); and kidney function was determined by measuring blood urea nitrogen (BUN) and creatinine (CRE) concentrations. No significant changes in liver function, kidney function, or body weight were observed (Figure5e). Collectively, these data suggest that the TSPAN8-LEL antibody is not significantly toxic in vitro or in vivo.
Effect of TSPAN8-LEL IgG on Internalization and Downregulation of Cell Surface Expression of TSPAN8 - To determine the effect of TSPAN8-LEL IgG on the downregulationof TSPAN8 expression on EOC cells, we performed a cell ELISA, using horseradish peroxidase (HRP)-conjugated TSPAN8-LEL IgG, to measure TSPAN8 expression on the surface of SK-OV3 cells following TSPAN8-LEL IgG treatment. TSPAN8-LEL IgG significantly downregulated the surface expression of TSPAN8 in a time-dependent manner, whereas control IgG had no effect(Figure6a). The time-dependent downregulation of TSPAN8 expression was also confirmed by immunoblot analysis (Figures6b and c). To exclude the possibility that the lower signal could be attributed to steric hindrance, we treated SK-OV3 cells in the absence or presence of HRP-conjugated TSPAN8-LEL IgGs or naked TSPAN8-LEL IgGs, respectively, and then monitored the cell surface TSPAN8 on SK-OV3 cells using cell ELISA. The results indicated that HRP-conjugated TSPAN8-LEL IgGs and naked TSPAN8-LEL IgGs could also induce the down-regulation of cell surface TSPAN8 on SK-OV3 cells, suggesting specific down-regulation of TSPAN8 by TSPAN8-LEL IgGs (Supplementary Figure 4). To verify antibody-induced internalization of TSPAN8, we treated SK-OV3 cells with fluorescein isothiocyanate (FITC)-labeled TSPAN8-LEL IgG and then monitored internalization by immunocytochemistry. LysoTracker was also used to label lysosomes. TSPAN8-LEL IgG rapidly colocalized with lysosomes in SK-OV3 cells, demonstrating rapid internalization and lysosomal targeting of TSPAN8-LEL IgG (Figure6d). These data indicate that the TSPAN8-LEL antibody induces rapid internalization and concomitant downregulation of TSPAN8 on the surface of EOC cells.
Discussion
EOC cell invasion and metastasis are complicated processes regulated by the coordinated action of multiple metastatic regulators (17). Despite the availability of a number of cancer therapeutics, invasion and metastasis still occur at high frequency and are major hurdles that must be overcome to improve outcomes for patients with ovarian cancers (2, 3). To this end, it is important to identify potential therapeutic targets and therapeutics for the treatment of ovarian cancer.In this study, we propose TSPAN8 as a novel therapeutic target for antibody therapy, andantibody targeting of TSPAN8 as an effective strategy for inhibiting invasion and metastasis of TSPAN8-expressing EOCs.
Cancer biomarkers are indicators of the severity or presence of cancer and are useful for evaluating the efficacy of therapeutic regimens. Human epidermal growth factor 2 (HER2), which is overexpressed in around 18–20% of breast cancer patients, is currently a useful biomarker for identifying patients who could benefit from treatment with trastuzumab, a humanized antibody that targets HER2 (18-20). Another biomarker, wild-type KRAS, is used to identify patients with epidermal growth factor receptor (EGFR)-positive colorectal cancers who could benefit from cetuximab, a chimeric antibody that targets EGFR (21, 22). However, an EOC biomarker is yet to be identified. Our resultsshowed a high expression of TSPAN8 in around 52% of EOC patients and its correlation with poor survival,suggesting that TSPAN8 might be a useful biomarker for EOC and thus could be exploited for therapeutic targeting.
Despite the development of many ovarian cancer therapeutics, the 5-year survival rate for patients remains relatively low (4, 5), reinforcing the importance of developing novel therapeutics. A number of lines of evidence suggest that TSPAN8-LEL IgG may have therapeutic potential. We verified thatthe IgG antibody binds specifically, and with subnanomolar affinity, to TSPAN8-LEL. Our in vitro and in vivo efficacy testing showed that TSPAN8-LEL IgG suppressed the invasion and metastasis of TSPAN8-expressing EOC. In addition, we demonstrated that the IgG antibody had little effect on the HUVEC viability, morphology, and activation. in vivo, the IgG antibody did not induce any changes in liver or kidney function, or body weight in a mouse model. Furthermore, our immunohistochemistry-based tissue cross-reactivity study suggests the specific targeting of the IgG antibody to ovarian cancer tissues in vivo. Thus, the TSPAN8-LEL antibody may specifically inhibit the invasion and metastasis of TSPAN8-expressing EOCs without causing severe toxicity to normal tissue.
TSPAN8 is a tumor-associated antigen that forms complexes with itself and with other factors involved in intracellular signal transduction (7, 15, 16). Using immunocytochemistry, we showed that TSPAN8-LEL IgG induced rapid internalization of TSPAN8 from the surface of SK-OV3 cells,along with TSPAN8 translocation to lysosomes, which are cellular organelles involved in protein degradation. ELISA and immunoblot analyses showed that treatment with TSPAN8-LEL IgG also significantly downregulated TSPAN8 in SK-OV3 cells in a time-dependent manner. Thus, the binding of TSPAN8-LEL IgG to TSPAN8 on the surface of EOC cells leads to rapid internalization of TSPAN8 and a consequent reduction in its surface expression, thereby suppressing TSPAN8-mediated signaling that promotes EOC metastasis. In this context, Ailane et al. recently reported that a mouse monoclonal antibody to TSPAN8 suppresses the growth of TSPAN8-expressing colorectal cancer cell lines in vivo(23). Collectively, these observations suggest that antibody-based modulation of TSPAN8 may suppress TSPAN8-mediated signaling in tumor cells.
Bevacizumab was the first therapeutic antibody available for treating patients with ovarian cancers (6). Previously, several groups reported that, although bevacizumab could prolong life in a peritoneal model of human ovarian cancer by inhibiting tumor growth, it did not suppress the incidence of tumor metastasis(24, 25). Intriguingly, in the current study, we found that TSPAN8-LEL IgG alone reduced the incidence of metastasis in a peritoneal model of human ovarian cancer. We also found that TSPAN8-LEL IgG, but not bevacizumab, inhibited SK-OV3 cell invasion in vitro. Therefore, these results lead us to speculate that the TSPAN8-LEL antibody suppresses more efficiently the invasion and metastasis of EOC, withadifferentmode of action from that of bevacizumab. Finally, we also suggest that, for better clinical outcome in ovarian cancer therapy,the TSPAN8-LEL antibody may be used not only in combination with chemotherapeutic agents, including paclitaxel and carboplatin, but also as an antibody platform for an antibody–drug conjugate or radioimmunotherapy,although additional studies are required. Taken together, these findings support the therapeutic potential and possible application of the TSPAN8-LEL antibody in the treatment of EOC.
In conclusion, we have shown that TSPAN8 is a novel therapeutic target in EOC,and antibody targeting of TSPAN8 may be an effective strategyfor suppressing the invasion and metastasis of TSPAN8-expressing EOC. On the basis of currently available evidence, we suggest a mode of action whereby the TSPAN8-LEL antibody binds to TSPAN8 on the surface of EOC cells and rapidly induces TSPAN8 internalization and translocation to lysosomes, resulting in a reduction in TSPAN8 surface levels and suppression of TSPAN8-mediated signaling implicated in EOC invasion and metastasis. In future studies, we plan to investigate the mechanism of action of the TSPAN8-LEL antibody in more detail and evaluate its in vivo efficacy, in combination with chemotherapeutic agents, against TSPAN8-mediated EOC metastasis.