The frequency of renal cell carcinoma (RCC) in the Japanese population is 6 out of 100,000, accounting for approximately 1% of all cancers [1
] and 2% of all cancers in North America [2
]. Thus, RCC is not a frequent disease; however, up to 30% of RCC cases have synchronous or metachronous metastases [2
]. Furthermore, RCC cells were detected in the bone marrow aspiration in 25% of RCC cases without metastasis [3
], suggesting that RCCs have a high metastatic potential. In cases with clear cell RCC (CCC), which is the most common RCC, 90% or more showed chromosome 3P deletion and inactivation of the von Hippel-Lindau (VHL) gene, leading to hypoxia-inducible factor (HIF)-1 activation [4
]. Epithelial-mesenchymal transition (EMT) is a factor related to the metastatic potential of CCC [5
], and HIF1 activation is considered a factor promoting EMT [6
]. Claudin-4 (CLDN4) is a tight junction protein and an epithelial marker [7
], and its decreased expression correlates with EMT [9
Attention has recently been paid to the function of the non-tight junction protein claudins (CLDNs). CLDNs are normally localized in the cell membrane as a tight junction-forming protein, and they regulate the diffusion of solutes through the intercellular space by their homotypic binding [7
]. In contrast, it has been reported that CLDNs that do not form a tight junction take part in intracellular signaling [10
]. CLDN7 binds to integrin β1 and suppresses the growth and movement of colon cancer cell lines [11
]. CLDN4, which does not form tight junctions in undifferentiated gastric cancer, becomes a ligand for integrin β1, activates focal adhesion kinase, and promotes stemness [13
]. Furthermore, we have reported that impairment of tight junctions caused by Clostridium perfringens
enterotoxin (CPE) leads to alterations in CLDN4 localization, activates Yes-associated protein (YAP), and induces EMT [14
]. However, significance of nuclear CLDN4 has not been reported in RCC.
In this study, we investigated the role of CLDN4 in RCC, focusing on nuclear CLDN4.
Previous reports have suggested that CLDN4 is expressed at low levels in RCC and in the bladder, colon, stomach, pancreatic, breast, and oral cancers [13
]. In CCCs, which account for most of the cases examined in this study, VHL inactivation is widely observed [24
]. VHL gene mutation reduces occludin and CLDN expression [25
]. Furthermore, HIF1 activation caused by VHL gene inactivation suppresses E-cadherin expression [25
In RCC, expression of CLDNs 1, 2, 3, 4, 5, 7, and 16 has been reported [27
], and CLDN2 was observed to show high expression [27
]. CLDN1 and 16 are downregulated with cancer progression [28
]. CLDN1 and 2 expression correlates with tumor grade [27
], and CLDN1, 3, and 4 expression is considered a poor prognostic factor [29
]. In contrast, in our study, CLDN4 expression did not correlate with clinicopathologic factors of RCC.
In the present study, CLDN4 expression was observed in the nucleus in some RCC cases. In the previous reports, CLDN3, 4, 7, and 8 were found to be expressed in the cell membrane and cytoplasm, and CLDN1 was expressed in the cell membrane [30
]. In our previous study, intranuclear CLDN4 was observed in approximately 30% of oral cancer cases. In these cases, Clostridium perfringens
infection was detected to impair tight junction through CPE, which led to the intracytoplasmic translocation of CLDN4 in these tumors [14
]. In the large intestine mucosa, tight junction impairment by CPE also causes CLDN4 translocation [21
]. Furthermore, not only Clostridium perfringens
but also Shigella
infections impair the tight junction of the large intestine mucosa and alter CLDN2 and 4 localization, resulting in their accumulation in the cytoplasm [31
]. From these findings, it is considered that release of CLDN4 from tight junctions induces alteration in the intracellular localization of CLDN4. The nuclear localization signal of CLDN4 is not clear. However, nuclear translocation of CLDN1, which lacks a nuclear localization signal, might translocate with binding with APC, ZO-1, or ZO-2 as shuttles [32
Bacteria such as Clostridium perfringens
that impair tight junctions were not found in the RCC cases that we examined (data not shown). Therefore, we investigated the phosphorylation of CLDN, which is known to impair tight junctions [33
]. Our data showed that in SN12L1 cells showing CLDN4 nuclear localization, phosphorylation of serine and tyrosine residues was observed in cytoplasmic CLDN4, whereas only serine phosphorylation was observed in nuclear CLDN4. This suggests that tyrosine phosphorylation might be associated with cytoplasmic translocation and serine phosphorylation might be related to nuclear translocation.
EphA2 is an enzyme known to phosphorylate the tyrosine residue of CLDN4 [18
]. EphA2 binds to CLDN4 of tight junction, phosphorylates Tyr208, which is in an intracellular domain near the N-terminus of CLDN4, and reduces the binding of CLDN4 to ZO-1; this leads to the translocation of CLDN4 from the tight junction to the cytoplasm [18
]. Our data showed that SN12L1 cells highly express EphA2 and its ligand, Ephrin A1. Inhibition of EphA2 reduced cytoplasmic CLDN4 levels. In contrast, in SN12C cells that show low expression of EphA2 and Ephrin A1, treatment with Ephrin A1 led to increased CLDN4 in the cytoplasm. EphA2 promotes RCC invasion and survival and is correlated with RCC grade, tumor size, and poor prognosis [34
]. EphA2 also induces EMT in cancer cells [37
PKCε is an enzyme that causes phosphorylation of serine/threonine residues of CLDN [33
]. PKCε phosphorylates Thr189 and Ser194 in ovarian cancer, resulting in a decrease in tight junction binding [19
]. There are no reports of other PKC isozymes using CLDN4 as a substrate. However, classical PKCs phosphorylate CLDN1 to induce nuclear localization in melanoma cells [38
]. Our data showed that PKCε expression was high in SN12L1 cells, and inhibition of PKCε reduced nuclear CLDN4 expression levels. In contrast, SN12C cells with low PKCε expression also showed increased nuclear CLDN4 upon activation of PKCε. In these experiments, because cytoplasmic CLDN4 expression was also altered, PKCε-mediated serine phosphorylation might be associated with cytoplasmic translocation of CLDN4.
PKCε is overexpressed in RCC, especially in CCC, and correlates with Fuhrman grade and tumor size [39
]. PKCε is one of the novel PKCs and is known as a transforming oncogene and a tumor biomarker. PKCε activates phosphoinositide 3-kinase/Akt, extracellular-signal-regulated kinase signaling, and integrin β1 [40
], which is associated with increased RCC stemness [40
]. PKCε is activated by stress such as ultraviolet and radiation in addition to phorbol ester (12-O-tetradecanoylphorbol 13-acetate) [42
]. Furthermore, PKCε and other novel PKCs are activated by 14-3-3ζ [33
], which is highly expressed in RCC and is associated with metastasis and poor prognosis [45
From these findings, CLDN4 phosphorylation by EphA2 and PKCε might cause tight junction impairment and release CLDN4 from tight junction and might enhance binding with YAP and ZO-1 to form nuclear translocating complex [15
We have previously shown that the nuclear translocation of YAP bound to CLDN4 occurs with the nuclear translocation of CLDN4, resulting in the activation of YAP and EMT [14
]. In this study, nuclear transfer of YAP bound to CLDN4 was observed in SN12L1 cells, and nuclear YAP was decreased by PKCε inhibition. In contrast, in SN12C cells, nuclear translocation of CLDN4 and YAP was not observed, whereas PKCε activation induced their nuclear translocation and EMT.
With the nuclear translocation of CLDN4 and YAP, the induction of the EMT phenotype was observed, and the invasive and metastatic abilities of the cells were also increased. In our previous studies, induction of the EMT phenotype by YAP activation was found in oral and colorectal cancers [14
]. In addition, increased CD44 expression and enhanced sphere formation were observed, which suggest increased stemness in cancer cells. Furthermore, when a nude mouse lung metastasis model was examined, SN12L1 cell metastasis was suppressed by PKCε inhibition and, conversely, SN12C cell metastasis was promoted by PKCε activation. The SN12L1 cell line is a highly metastatic strain established from the SN12C cell line [46
]. The EMT phenotype induced by the EphA2-PKCε-CLDN4-YAP axis is considered one of the mechanisms underlying the metastatic ability of SN12L1 cells. Thus, it was considered that the YAP activation associated with CLDN4 nuclear translocation plays an important role in the acquisition of a malignant phenotype in RCC. These findings are expected to provide new molecular targets for RCC treatment.
We found that CLDN4 phosphorylation by EphA2/Ephrin A1 and PKCε induced nuclear translocation of YAP with CLDN4 to provided EMT in RCC cells. Sarcomatoid RCC is not included in the 202 cases examined this time. Examining the nuclear translocation of CLDN4 in tumors exhibiting EMT phenotype such as sarcomatoid RCC is an interesting issue. It should be examined in the future. The frequency of nuclear translocation of CLDN4 is low, probably because it uses a method with low detection sensitivity by immunostaining. It is necessary to study using a highly sensitive method such as ELISA. The release of CLDN4 by impaired tight junction might be one of the mechanisms of malignant properties of RCC, which is expected to be a novel therapeutic target for RCC treatment.
4. Materials and Methods
4.1. Surgical Specimens
We reviewed the pathological diagnosis and clinical data of 202 patients with surgically resected RCC, reviewed at the Department of Molecular Pathology, Nara Medical University, during 2006–2015. As written informed consent was not obtained, any identifying information was removed from the samples before analysis to ensure strict privacy protection (unlinkable anonymization). All procedures were performed in accordance with the Ethical Guidelines for Human Genome/Gene Research enacted by the Japanese Government and were approved by the Ethics Committee of Nara Medical University (Approval Number 937, 2020/4/1).
4.2. Cell Lines
SN12C and SN12L1 human RCC cell lines were kindly provided by Professor Isaiah J Fidler (MD Anderson Cancer Center, TX, USA) [47
]. Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum at 37 °C in 5% CO2
An in vitro invasion assay was performed using a type IV collagen-coated insert. The number of cells invading into the collagen membrane was measured after 48 h.
4.3. Sphere Assay
SN12C or SN12L1 cells (1000 cells per well) were seeded on uncoated bacteriological 35 mm-dish (Coning Inc., Coning, NY, USA) with 3D Tumorsphere Medium XF (Sigma-Aldrich Inc., St. Louis, MO, USA). After 5 days, the sphere number was counted.
4.4. Antibody and Reagents
Anti-human CLDN4 extracellular domain antibody, 4D3, was developed by immunizing rats with a plasmid vector encoding human CLDN4 [20
]. Anti-EphA2 antibody (clone 1A9C3, 1 µg/mL for blocking concentration, Proteintech Group Inc., Rosemont, IL, USA), PKCε inhibitor peptide (10 μM for working concentration, sc-3095, Santa-Cruz Biotechnology, Santa-Cruz, CA, USA), L-alpha-phosphatidylinositol-3,4,5-trisphosphate sodium salt (PKCδ/ε/η activator, 5 µM for working concentration, ab145221, Abcam, Cambridge, MA, USA), and recombinant human ephrin A1 protein (10 µg/mL for working concentration, ab181919, Abcam) were purchased.
Consecutive 4-mm sections were immunohistochemically stained using anti-CLDN4 antibody (0.2 µg/mL, clone 4D3), which was established in our laboratory [20
], and a previously described immunoperoxidase technique [48
] was performed. Secondary antibodies for peroxidase-conjugated mouse IgG and alkaline phosphatase-conjugated rabbit IgG (Medical and Biological Laboratories, Nagoya, Japan) were used at a concentration of 0.2 µg/mL. Tissue sections were color-developed with diamine benzidine hydrochloride (DAKO, Glastrup, Denmark). Slides were counterstained with Meyer’s hematoxylin (Sigma). We counted immunopositive cells at the cytoplasmic membrane. Staining strength was scored from 0 to 3 (a score of 1 was used to describe the expression level in normal renal tubule epithelium). The staining index was calculated as the staining strength score multiplied by the staining area (%). As negative control, non-immunized rat IgG (Santa-Cruz) was used as the primary antibody.
4.6. Protein Extraction
For preparing whole-cell lysate, SN12C and SN12L1 cells were washed twice with cold phosphate-buffered saline (PBS), harvested, and lysed with 0.1% sodium dodecyl sulfate (SDS)-added radioimmunoprecipitation assay buffer (Thermo Fisher Scientific, Tokyo, Japan) [49
]. Cell fractions were extracted using a Cell Fractionation Kit (Abcam), according to the manufacturer’s instructions [50
]. Protein assay was performed using a Protein Assay Rapid Kit (Wako Pure Chemical Corporation, Osaka, Japan).
4.7. Immunoblot Analysis
Lysates (20 μg) were subjected to immunoblot analysis using SDS–polyacrylamide gel electrophoresis (12.5%), followed by electrotransfer onto nitrocellulose filters. The filters were incubated with primary antibodies, followed by peroxidase-conjugated IgG antibodies (Medical and Biological Laboratories). Anti-tubulin antibody was used to assess the protein levels loaded per lane (Oncogene Research Products, Cambridge, MA, USA). The immune complex was visualized using an enhanced chemiluminescence Western blot detection system (Amersham, Aylesbury, UK). Antibodies for E-cadherin (DAKO), CLDN4 (clone 4D3) [20
], EphA2, Ephrin A1, vimentin (Proteintech), snail (Biorbyt, St Louis, MO, USA), YAP1, CD44 (Abcam, Cambridge, UK), and PKCε (Enzo Lifesciences, Inc., Farmingdale, NY, USA) were used as primary antibodies. Tubulin (Zymed Laboratories Inc., South San Francisco, CA, USA) and lamin (Proteintech) were used as the loading control.
Immunoprecipitation was performed according to the method described previously [51
]. Briefly, whole-cell lysates were pre-cleaned in lysis buffer with protein A/G agarose (Santa-Cruz) for 1 h at 4 °C and subsequently centrifuged. The supernatants were incubated with antibody against CLDN4 (4D3) and protein A/G agarose for 3 h at 4 °C. Precipitates were collected by centrifugation, washed five times with lysis buffer, solubilized with sample buffer (40 µL; Sigma), and subjected to immunoblot analysis with antibodies against YAP1 (Abcam) or phosphoserine (Abcam, ab9332), or phosphotyrosine (Abcam, ab179530). Loading protein volume was confirmed by slot blot analysis of 10 µL of the samples by Coomassie blue staining (Bio-Rad, Hercules, CA, USA).
4.9. Enzyme-Linked Immunosorbent Assay (ELISA) and Colorimetric Assay
An ELISA kit was used to measure the concentration of human CLDN4 (Cusabio Biotech Co., Ltd., Houston, TX, USA). The assay was performed according to the manufacturers’ instructions, and whole-cell lysates were used for the measurements.
BALB/c nude mice (4 weeks old, male) were purchased from SLC Japan (Shizuoka, Japan). The mice were maintained according to the institutional guidelines approved by the Committee for Animal Experimentation of Nara Medical University, in accordance with the current regulations and standards of the Ministry of Health, Labor, and Welfare (Approval number 12047, 2017/7/20).
4.11. Lung Metastasis Model
SN12L1 and SN12C cells were pretreated with or without PKC-I (10 μM) and PKC-A (5 µM), respectively, for 24 h. Suspensions (1 × 106 cells/50 μL of PBS) of SN12L1 and SN12C cells labeled with VivoTrack 680 (PerkinElmer Inc., Waltham, MA, USA) were injected into the caudal vein. The mice were euthanized, and the lungs were observed using the Clairvivo OPT in vivo imager (Shimazu, Kyoto, Japan) two weeks after inoculation.
4.12. Statistical Analysis
Statistical significance was calculated using two-tailed Fisher’s exact test, an ordinary analysis of variance, and InStat software version 3.0 (GraphPad, Los Angeles, CA, USA). A two-sided p value of <0.05 was considered to indicate statistical significance.