The growth and progression of cancers are crucially regulated by the tumor microenvironment where tumor cells and stromal cells are mutually associated. In this study, we found that stomatin expression was markedly upregulated by the interaction between prostate cancer cells and stromal cells. Stomatin suppressed cancer cell proliferation and enhanced apoptosis in vitro and inhibited xenograft tumor growth in vivo. Stomatin inhibited Akt activation, which is mediated by phosphoinositide-dependent protein kinase 1 (PDPK1). PDPK1 protein stability was maintained by its binding to HSP90. Stomatin interacted with PDPK1 and interfered with the PDPK1–HSP90 complex formation, resulting in decreased PDPK1 expression. Knockdown of stomatin in cancer cells elevated Akt activation and promoted cell increase by promoting the interaction between PDPK1 and HSP90. Clinically, stomatin expression levels were significantly decreased in human prostate cancer samples with high Gleason scores, and lower expression of stomatin was associated with higher recurrence of prostate cancer after the operation. Collectively, these findings demonstrate the tumor-suppressive effect of stromal-induced stomatin on cancer cells.
These findings reveal that interactions with stromal cells induce expression of stomatin in prostate cancer cells, which suppresses tumor growth via attenuation of the Akt signaling axis.
Tumor growth and progression are affected by a number of factors, such as abnormal gene expressions as well as aberrant cellular functions of both the tumor and the surrounding cells. The growth and progression of tumors are profoundly related to tumor malignancy and are critical for the prognosis of patients with cancer. The tumor microenvironment that is generated by the reciprocal cross-talk among cancer cells, several types of host stromal cells, and various extracellular matrices plays a key role in tumor behavior (1, 2). The behavior appears to be generally determined by extracellular matrix- and cytokine-mediated communications between cancer cells and by the direct cell-to-cell interactions of cancer cells with stromal cells (3, 4). However, how the direct interactions in the tumor microenvironment promote or suppress tumor growth is not known.
To explore the relationship between interactions in the tumor microenvironment with tumor behavior, we previously investigated genes that showed changes in expression in response to the direct physical association of tumor cells with stromal cells. We established a coculture assay system using prostate cancer cells and primary human prostate stromal (PrS) cells, and identified 30 genes that were significantly increased in the cocultured prostate cancer cells compared with control prostate cancer cells that were cultured alone (5). Among the increased genes, the STOMATIN gene was included.
Stomatin belongs to the SPFH superfamily of proteins consisting of stomatin, stomatin-like proteins, prohibitins, flotillin/reggie proteins, and HflK/C proteins (6). SPFH protein family members localize in the plasma membrane of many cell types, including red blood cells, fibroblasts, and cancer cells, to regulate diverse cellular functions (7, 8). The absence of stomatin in humans may cause hereditary stomatocytosis, a form of hemolytic anemia (9–11). However, the homologous deletion of stomatin in mice shows no obvious phenotype in the physiologic state (12). The subcellular localization of stomatin is important for several channel activities (6, 13, 14). Stomatin also contributes to the determination of cell morphology by binding to cortical actin in epithelial cells (15, 16). However, little is understood about the function of stomatin in cancer cells.
In this study, we examined the effect of stomatin on cancer cells. High expression of stomatin inhibited cancer cell proliferation and enhanced apoptosis. Tumor growth in the xenograft tumor model was almost completely inhibited by the induction of stomatin expression in cancer cells. Stomatin attenuated the Akt-mediated signaling pathways in both in vitro and in vivo models, which was caused by stomatin-induced loss of protein stability of phosphoinositide-dependent protein kinase 1 (PDPK1), a major activator of Akt. Regarding clinical implications, prostate cancer tissues from patients with higher Gleason scores (GS) expressed a lower level of stomatin, which was associated with higher recurrence of prostate cancer after the operation. These results suggest that stomatin functions as a tumor suppressor.
Materials and Methods
LNCaP cells (RRID: CVCL_0395) were purchased from Dainippon Sumitomo Pharma, PC3M cells (RRID: CVCL_9555) were from Elabscience, HEK293T (RRID: CVCL_0063) and HepG2 cells (RRID: CVCL_0027) were purchased from ATCC, and human prostate stromal cells were purchased from Lonza. 22Rv1 cells (RRID: CVCL_1045) and VCaP cells (RRID: CVCL_2235) were gifted by Dr. M. Nagasawa at Department of Urology, Shiga University of Medical Science. All cell lines were authenticated by the supplier, tested to be free of Mycoplasma by PCR, and routinely examined for contamination, morphology, and growth profile. Once cell lines were thawed from authenticated cell stocks, we used them within 2 months.
Male NOD/ShiJic-scid (NOD.CB17-Prkdcscid; NOD/SCID) mice (RRID: MGI_2173972) were purchased from CLEA Japan, Inc., and housed in specific pathogen-free conditions in the Research Center for Animal Life Science of Shiga University of Medical Science. The animal experiments conducted in this study were approved by Shiga University of Medical Science Animal Care and Use Committee (No. 2017-12-1), and were performed in accordance with relevant guidelines and regulations including Animal Research Reporting of In Vivo Experiments (ARRIVE) guidelines.
After total RNA extraction, cDNA was synthesized by the ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo). qPCR was performed using the LightCycler Instrument (Roche Diagnostics) as described previously (17). The samples were run in duplicate, and the data were quantified by the standard curve method. The primers used were as follows: STOMATIN (forward 5′-GATTTTGGTGGCGTTCTCAT-3′ and reverse 5′-CACACCATCCACGCTAATTG-3′), PDPK1 (forward 5′-AAAGAATTCATGGCCAGGACCACCAGCCAGCTG-3′ and reverse 5′-TTTGTCGACCTGCACAGCGGCGTCCGGGTGGCTC-3′), and β-ACTIN (forward 5′-TCCTCCCTGGAGAAGAGCTA-3′ and reverse 5′-GTGGATGCCACAGGACTC-3′).
All antibodies used in this study are listed in Supplementary Table S1.
Coculture of LNCaP cells with PrS cells
This experiment was performed as described previously (5). Briefly, LNCaP cells were labeled by the Vybrant CFDA SE Cell Tracer Kit (Invitrogen). The fluorescence-labeled LNCaP cells were cocultured with the same number of PrS cells for 48 hours, or cultured alone as a control. Cell culture conditioned media were mixed between dishes every 6 hours to cancel the effect of culture dish-specific different growth factors on the stomatin expression. After the coculture, fluorescence-labeled LNCaP cells were isolated from PrS cells by flow cytometry (FACSAria; BD Biosciences). Isolated LNCaP cells were lysed and the extracts were used for qPCR and Western blotting to assess the stomatin expression.
Tumor formation in mice
LNCaP cells (1 × 107 cells) or PC3M cells (2 × 106 cells) with Dox-inducible expression of control GFP or stomatin-GFP in 200 μL of PBS containing Matrigel (50%, v/v; Corning Inc.) were subcutaneously injected into the dorsal surfaces of male NOD/SCID mice aged 8 to 12 weeks through a 25-gauge needle (day 0). For the coculture xenograft model, LNCaP cells (1 × 107 cells) with doxycycline (Dox)-inducible expression of GFP were subcutaneously injected in the mice together with PrS cells (3 × 106 cells) in 500 μL of PBS containing Matrigel (50%, v/v). The mice were treated with 0.2% Dox (Tokyo Chemical Industry) in the drinking water just after transplantation or after formation of tumors. 22Rv1 cells stably expressing shControl or shStomatin (2 × 107 cells each) in 200 μL of PBS containing Matrigel (50%, v/v) were subcutaneously injected in the mice. Tumor volume was calculated every week as follows: volume (mm3) = 0.52 × [short axis (mm)]2 × long axis (mm). After sacrificing the mice, the tumors were measured, weighed, and processed for histologic analysis.
Total RNA extraction from formalin-fixed paraffin-embedded human tissues
Formalin-fixed paraffin-embedded prostate cancer samples were sectioned and deparaffinized. Total RNA in the samples was extracted using the ReliaPrep FFPE Total RNA Miniprep System (Promega). The experiments and procedures using human prostate cancer samples were approved by the Research Ethics Committee at Shiga University of Medical Science (No. R2017-059). This study is an observational study using existing tissue samples and information. The informed consent was obtained from the subjects whose samples and information were used for analysis after the operation in the form of opt-out on the website of Shiga University of Medical Science Hospital. This opt-out procedure is described in Chapter 5, Ethical Guidelines for Medical and Health Research Involving Human Subjects (https://www.mhlw.go.jp/file/06-Seisakujouhou-10600000-Daijinkanboukouseikagakuka/0000080278.pdf), developed by The Ministry of Health, Labour and Welfare, Japan. The Research Ethics Committee at Shiga University of Medical Science follows the above guidelines, and did not require the written informed consent in this study. The experiments were performed in accordance with relevant guidelines and regulations including Declaration of Helsinki principles.
Formalin-fixed paraffin-embedded prostate cancer samples were sectioned and deparaffinized. The samples were soaked in preheated antigen retrieval solution [10 mmol/L Tris-HCl (pH 8.3) and 1 mmol/L EDTA] at 80°C for 10 minutes. The samples were cooled down at room temperature, quenched by 3% hydrogen peroxidase: 100% methanol for 15 minutes, and then blocked with 3% bovine serum albumin prior to overnight incubation with a mouse monoclonal anti-stomatin antibody (1:1,000 dilution), anti-PDPK1 antibody (1:500 dilution), or anti-phospho-Akt antibody (1:500 dilution). After washing with PBS, samples were incubated with a biotin-conjugated anti-mouse IgG (H+L) secondary antibody (1:200 dilution; Thermo Fisher Scientific) for 1 hour at room temperature, followed by incubation with streptavidin-peroxidase (Nacalai Tesque) for 30 minutes to form the avidin–biotin complex, which was detected with the Strep ABC Peroxidase Kit (Nacalai Tesque). Samples were counterstained with hematoxylin, followed by dehydration steps. Chromogenic staining was evaluated by determining H-score. The score is based on the predominant staining intensity of 3,3′-diaminobenzidine in tumor cells (0, negative; 1+, weak; 2+, moderate; 3+, intense; ref. 18). The total number of cells and the positively stained tumor cells at each intensity score were counted, and the percentage of the cells was obtained. The H-score (ranging from 0 to 300) was calculated using the following formula: H-score = 1 × (percentage of score 1 cells) + 2 × (percentage of score 2 cells) + 3 × (percentage of score 3 cells).
All data are expressed as the mean ± SEM. The experiments in each figure were performed at least three times, and statistical differences between experimental groups were evaluated by the two-tailed Student t test, a one-way ANOVA or two-way repeated measures ANOVA. If ANOVA indicated overall significance, individual differences were evaluated using the Bonferroni posttest. Patients’ clinical data were statistically examined by logistic regression analysis. We used Kaplan–Meier analysis for evaluation of the recurrence of prostate cancer in two groups. The recurrence-free curves were compared by the log-rank test. P value < 0.05 was considered to be statistically different.
Additional materials and methods
Additional materials and methods are described in the Supplementary Materials and Methods of Supplementary Data.
Expression and cellular function of stomatin in cancer cells
We first confirmed that the expression of stomatin was increased in prostate cancer cells by the cell-to-cell contact between cancer cells and PrS cells. After coculture of fluorescence labeled-LNCaP cells with PrS cells, LNCaP cells were isolated by fluorescence-activated cell sorting. qPCR and Western blotting showed that the stomatin expression was increased in LNCaP cells cocultured with PrS cells compared with LNCaP cells cultured alone (Fig. 1A and B).
Next, we generated an expression vector of GFP-tagged stomatin to examine stomatin functions in cancer cells. Stomatin-GFP was partially localized on the plasma membrane in transfected LNCaP cells (Fig. 1C). We for the first time found that the increase in the number of LNCaP cells overexpressing stomatin-GFP was almost completely inhibited (Fig. 1D). This was also observed in another cell line, HEK293T cells. Upon transfection of stomatin-GFP in LNCaP cells, the expression of Ki67, a cell proliferation marker, was reduced, and TUNEL-positive cells were increased, compared with GFP-transfected controls (Fig. 1E and F). These results suggest that the inhibition of cancer cell growth by stomatin is mediated by both suppressed cell proliferation and enhanced apoptosis.
Stomatin-induced inhibition of tumor growth in vivo
The results from the in vitro experiments showed that the transient expression of stomatin in cancer cells strongly suppressed the increase of cancer cells. To further examine the tumor suppressive role of stomatin in tumor growth in a xenograft model in vivo, we generated LNCaP and PC3M cells with Dox-inducible expression of stomatin-GFP or control GFP, because endogenous expression of stomatin was not detected in these prostate cancer cells (Supplementary Fig. S1A). In these cells, neither stomatin-GFP nor GFP was expressed in the absence of Dox; upon treatment with Dox, the expression of stomatin-GFP with some degradative products and GFP were induced (Fig. 2A; Supplementary Fig. S1B). The increase in the number of both LNCaP and PC3M cells in response to induced stomatin-GFP was significantly inhibited (Fig. 2B; Supplementary Fig. S1C). We injected these cells subcutaneously into NOD/SCID mice that then received Dox in drinking water at three days after injection. The injected cancer cells that expressed control GFP by the induction of Dox were highly engrafted, and tumor volume increased during the observation period, whereas cells induced for stomatin-GFP expression hardly formed tumors (Fig. 2C and D; Supplementary Fig. S1D). IHC demonstrated that the Ki67-positive cells were reduced and that the level of cleaved caspase-3 was enhanced in the tumors formed by LNCaP and PC3M cells with induced stomatin-GFP expression (Fig. 2E and F; Supplementary Figs. S1E and S1F). These results suggest that stomatin suppressed tumor growth by inhibition of cell proliferation and promotion of apoptosis in vivo.
To further examine whether in the in vivo xenograft model, stomatin expression is endogenously induced in prostate cancer cells by association of the cancer cells with stromal cells and modulates tumor progression, we injected LNCaP cells, in which Dox-inducible GFP expression is observed as a marker for the prostate cancer cells, together with PrS cells subcutaneously into NOD/SCID mice, and treated the mice with Dox. The tumor growth was significantly attenuated by coinjection of LNCaP and PrS cells, compared with that of LNCaP cells alone (Supplementary Fig. S2A). Stomatin expression was actually induced in GFP-expressing LNCaP cells coinjected with and neighbored to PrS cells, whereas the expression was undetectable in LNCaP cells injected alone (Supplementary Fig. S2B). Ki67- and cleaved caspase-3-positive cells were decreased and increased, respectively, in LNCaP cells coinjected with PrS cells, compared with LNCaP cells injected alone (Supplementary Figs. S2C and S2D).
Considering the potential clinical implications of the tumor suppressive role of stomatin, we then examined whether the induction of stomatin expression could inhibit or reverse tumor growth and progression in existing tumors. First, we performed in vitro assays and examined the effect of stomatin on the number of LNCaP cells that had been cultured for 6 days in the absence of Dox. The growth rate of LNCaP cells expressing GFP with Dox was slightly reduced compared with LNCaP cells with GFP cells cultured in the absence of Dox (Fig. 3A). This is most likely due to the toxicity of Dox. In LNCaP cells expressing stomatin-GFP, the cell growth was significantly inhibited during the administration of Dox, and more strikingly, the number of cells was gradually reduced (Fig. 3A). We then conducted subcutaneous injection of these LNCaP cells into NOD/SCID mice, and monitored the mice without Dox administration. The mice received Dox in drinking water once the volume of xenograft tumors reached approximately 350 mm3. The tumors derived from GFP-expressing LNCaP cells became very large, whereas the growth of tumors expressing stomatin-GFP was mild (Fig. 3B). The external appearance of tumors was clearly different between the groups (Fig. 3C). GFP-expressing tumors were aggressively expanded and relatively softer, whereas stomatin-GFP-expressing tumors were discolored, smaller and more solid, indicating the stagnation of tumor growth. Hematoxylin and eosin staining of tumor sections showed abnormal protein depositions and plenty of empty cavities between neighboring cells in stomatin-GFP-expressing tumors, compared with GFP-expressing tumors (Fig. 3D), suggesting the induction of intratumoral apoptosis and necrosis by stomatin. Furthermore, similar to the results in Fig. 2 and Supplementary Fig. S1, Ki67-positive cells were reduced, and cleaved caspase-3-positive cells were increased in the tumors derived from LNCaP cells expressing stomatin-GFP after Dox treatment (Fig. 3E and F). Collectively, these results in the xenograft tumor model suggest that stomatin exhibits strong tumor suppressive activity in vivo.
Inhibition of the Akt signaling by stomatin
To investigate the molecular mechanism of stomatin in inhibition of cancer cell growth, we examined intracellular signaling pathways that critically define the phenotype of cancer cells. The PI3K–Akt pathway is crucial for cell proliferation and survival, and dysregulation of this pathway results in the abnormality of cell growth, leading to carcinogenesis and cancer progression (19). In LNCaP cells, Akt activation (phosphorylation) was inhibited upon expression of stomatin-GFP compared with control cells expressing GFP (Fig. 4A). The inhibition of Akt activation was further confirmed in LNCaP and PC3M cells with induced stomatin-GFP expression both in vitro and in vivo (Fig. 4B and C; Supplementary Figs. S3A and S3B). We next investigated how the inactivation of Akt by stomatin suppresses both cancer cell proliferation and survival. The forkhead box class O (FOXO) family of transcription factors function as tumor suppressors in the nucleus by inhibiting cell proliferation; upon phosphorylation by Akt, FOXO proteins translocate into the cytoplasm, resulting in loss of their inhibitory effect on cell proliferation (20, 21). The expression of stomatin-GFP in LNCaP cells decreased the phosphorylation levels of FOXO3a (Fig. 4D). Furthermore, stomatin-GFP expression increased the cleavage of PARP, an indicator of apoptosis, and decreased the protein levels of antiapoptotic factors Bcl-2 and Bcl-XL (Fig. 4E), indicating the enhancement of apoptosis by stomatin. Collectively, these results suggest that stomatin attenuates cell proliferation and induces apoptosis by inhibition of Akt.
We performed several following experiments to confirm that Akt inactivation by stomatin is essential for the inhibitory effects of stomatin on cancer cell growth and that both apoptosis and FOXO-modulated cell proliferation play a combinational role in the stomatin–Akt signaling axis. First, when the HA-tagged constitutively active form of Akt (HA-Akt-DD) was expressed in the presence of stomatin, the stomatin-mediated inhibition of cell proliferation activity through FOXO3a phosphorylation and enhancement of apoptosis assessed by cleaved PARP levels were completely canceled (Supplementary Figs. S4A and S4B). In addition, the increase in the cell number, which was suppressed by inducible expression of stomatin in LNCaP, was fully recovered by co-expression of HA-Akt-DD (Supplementary Fig. S4C). Next, when LNCaP cells expressing stomatin were treated with an apoptosis inhibitor Z-VAD, the level of cleaved PARP was completely suppressed, but inactivation of Akt, phosphorylation of FOXO3a, and frequency of Ki67-positive proliferative cells were not affected (Supplementary Figs. S4D and S4E).The Z-VAD treatment partially rescued the suppression of cell number increase by stomatin (Supplementary Fig. S4F). Furthermore, knockdown of FOXO3a did not affect the stomatin-mediated suppression of Akt phosphorylation or induction of cleaved PARP level (Supplementary Fig. S4G). However, the knockdown promoted the cell proliferation activity, resulting in the partial recovery of the LNCaP cell number increase (Supplementary Figs. S4H and S4I). Finally, we found that both Z-VAD treatment and FOXO3a knockdown fully recovered the stomatin-mediated inhibition of the increase in LNCaP cell number (Supplementary Fig. S4J), confirming that tumor suppressive role of stomatin crucially depends on both enhancement of apoptosis and inhibition of cell proliferation through Akt function.
Stomatin-mediated instability of PDPK1 for inhibition of Akt
Akt is phosphorylated and activated by PDPK1, which functions downstream of PI3K (22). We examined the expression level of PDPK1, and found that PDPK1 expression was significantly reduced by transient and induced expression of stomatin-GFP in LNCaP and PC3M cells (Fig. 5A and B; Supplementary Fig. S3B). PDPK1 is stabilized by its association with HSP90, one of the chaperone proteins (23). Before examining the HSP90-dependent PDPK1 stabilization, we confirmed that the mRNA level of PDPK1 was not affected by stomatin expression (Fig. 5C). When HSP90 was knocked down in LNCaP cells by transfection with HSP90 siRNA (siHSP90), PDPK1 expression was decreased, and Akt phosphorylation was also reduced (Fig. 5D). Furthermore, the increase in the number of LNCaP cells was attenuated in siHSP90-transfected cells, compared with control scramble RNA-transfected cells (Fig. 5E).
As described above, the interaction of PDPK1 with HSP90 is required for PDPK1 protein stability (23), and thus we tested whether stomatin affects the HSP90–PDPK1 interaction. Expression of stomatin-GFP inhibited the interaction of HSP90 with PDPK1 at endogenous protein levels in LNCaP cells (Fig. 5F). The binding of stomatin-GFP to PDPK1 was increased even though PDPK1 expression was reduced under the condition of stomatin-GFP expression (Fig. 5G). The binding was also confirmed using the purified proteins in the GST pull-down assay. Endogenous PDPK1 in LNCaP cells and purified His-tagged PDPK1 were bound to GST-stomatin (Fig. 5H). The HSP90 binding site in PDPK1 is at residues 156 to 223 (23), and we hypothesized that stomatin might also bind to this region to interfere with the complex formation of HSP90 and PDPK1. We generated three expression vectors: FLAG-tagged full-length PDPK1 (PDPK1-FL: aa 1–556), FLAG-tagged N-terminal PDPK1 (PDPK1-N: aa 1–241) including the HSP90-binding site, and FLAG-tagged C-terminal PDPK1 (PDPK1-C: aa 224–556; Fig. 5I). In immunoprecipitation assays using HEK293T cells transfected with stomatin-GFP and either of the PDPK1-FLAG vectors, stomatin-GFP co-immunoprecipitated with PDPK1-N-FLAG as well as PDPK1-FL-FLAG, but not with PDPK1-C-FLAG (Fig. 5J). Furthermore, expression of stomatin-GFP reduced the interaction of PDPK1-N-FLAG with endogenous HSP90 in HEK293T cells (Fig. 5K). These results suggest that stomatin interacts with PDPK1 and inhibits PDPK1 interaction with HSP90.
Enhancement of PDPK1–Akt signaling and cell growth by knockdown of stomatin
We next investigated the intrinsic role of stomatin in cancer cells. Evaluation of the expression levels of endogenous stomatin in several cancer cell lines by Western blotting revealed higher expression of stomatin in hepatocellular carcinoma HepG2 cells and prostate cancer 22Rv1 cells, compared with other cell lines (Fig. 6A; Supplementary Fig. S1A). Endogenous stomatin was knocked down in HepG2 and 22Rv1 cells by siRNA, and the effectiveness of the knockdown was confirmed by Western blotting and immunocytochemistry (Fig. 6B and C; Supplementary Figs. S5A and S5B). Knockdown of stomatin in both cell lines enhanced the phosphorylation of Akt through the increased expression of PDPK1, and promoted cell growth (Fig. 6C–E; Supplementary Fig. S5A and S5C). In addition, stomatin knockdown augmented the association of PDPK1 with HSP90 (Fig. 6E). Consistent with the results from the in vitro experiments, the enhanced tumor growth was also observed in the mice subcutaneously injected with 22Rv1 cells stably expressing shStomatin, compared with those with the cells expressing shControl (Supplementary Figs. S5D and S5E). The expression of stomatin was hardly detected in the tumor formed by shStomatin-expressing 22Rv1 cells during 12 weeks of the experimental period (Supplementary Fig. S5F). Akt phosphorylation and PDPK1 expression levels were higher in stomatin-knockdown 22Rv1 cells than control cells (Supplementary Fig. S5G).
To further examine whether the enhancement of Akt phosphorylation and PDPK1 expression by knockdown of stomatin is dependent on HSP90 action in HepG2 cells and vice versa, both stomatin and HSP90 were knocked down in HepG2 cells and evaluated in comparison with cells with knockdown of either stomatin or HSP90. Knockdown of stomatin or HSP90 induced an increase or decrease in Akt phosphorylation, respectively, and these changes were reversed by double knockdown of stomatin and HSP90 (Fig. 6F). Similar results were also observed in the expression of PDPK1 (Fig. 6F).
Decrease in stomatin expression in malignant human prostate cancer
As stomatin was shown to possess inhibitory effects on tumor growth in vitro and in vivo, we further examined the relationship between stomatin expression in human prostate cancer and the cancer malignancy. The mRNA level of STOMATIN in prostate cancer samples with a lower GS was higher than in those with a higher GS (Fig. 7A). IHC analyses showed that stomatin-positive signals frequently existed along the contact sites between the cancer lesion and the stromal tissue, and that the signals were strongly observed in the lower GS samples compared with the higher GS samples (Fig. 7B). Detailed IHC of consecutive sections of the lower GS samples further confirmed that the expression of stomatin in the E-cadherin-positive prostate cancer cells was in juxtaposition to the α-SMA-positive stroma cells (Fig. 7C).
Next, 32 patients of which prostate cancer samples were examined above were divided into the two groups based on the median of the stomatin mRNA levels for further analyses: Stomatin-High and Stomatin-Low groups. The patients’ clinical characteristics in each group were shown in Supplementary Table S2. The prostate cancer recurrence-free rate after the operation was significantly better in the Stomatin-High group than the Stomatin-Low group (Fig. 7D), whereas the rate was not statistically different between the patients with higher GS and lower GS groups (Supplementary Fig. S6). These results suggest that the low expression of stomatin in the prostate cancer might be a good predictive marker for poor outcome in cancer recurrence, although the sample number is limited. Moreover, the signals of PDPK1 and phosphorylated Akt were strongly observed mainly at the cancer–stroma contact area in the samples from the Stomatin-Low group, compared with those from the Stomatin-High group (Fig. 7E). These results suggest that decreased expression of stomatin is clinically related to the malignancy of human prostate cancer. Combined with all of the results in this study, our findings indicate that stomatin may function as a novel tumor suppressor through inhibition of the PDPK1–Akt signaling pathway (Fig. 7F).
In this study, we demonstrated the stomatin-mediated novel signaling pathway to suppress tumor growth and progression. Stomatin expression was elevated by the interaction between prostate cancer cells and prostate stromal cells in vitro and in vivo, which mimics the tumor microenvironment that occurs upon cancer cell invasion into the surrounding connective tissue. The mechanism by which stomatin suppresses tumor progression was dependent on the inhibition of Akt activity through the instability of PDPK1, resulting in both the significant attenuation of cancer cell proliferation activity and the enhancement of apoptosis. Consistent with the results of this study, it is reported that stomatin expression is decreased in HER2-positive malignant breast cancer (24). Another recent study shows that stomatin inhibits tumor metastasis in non–small cell lung cancer (25), supporting our conclusion that stomatin exhibits tumor suppressive effects. However, these previous studies did not elucidate the detailed molecular functions of stomatin in cancer cells. In this context, our findings have advances in the field of cancer biology.
The tumor microenvironment critically influences the characteristics of cancer cells (26). Similar to the binding of soluble factors to their receptors on cancer cells, the direct cell-to-cell contact between invading cancer cells and neighboring stromal cells is an important factor for the tumor microenvironment (27). Our previous study shows that the expression of epithelial membrane protein 1 (EMP1) is elevated by the cancer cell–stromal cell interaction (5). In contrast to stomatin, EMP1 promotes cancer invasion and metastasis. Thus, the tumor microenvironment has both pro- and antitumor progressive functions, which may determine the benign or malignant property of tumors in their developmental processes. Regarding the clinical implications, stromal cells as well as cancer cells consist of heterogenous cell populations, in contrast to those used in our in vitro assays. We thus also examined human prostate cancer samples, and found an inverse correlation between stomatin expression levels and GS. Although stomatin-positive cancer cells were present along the interface of cancer and stroma areas, the reason why stomatin expression was different between samples of higher and lower GS is not yet known. These suggest that the interaction between cancer cells and stromal cells does not completely govern stomatin expression in cancer cells. Differences in the genetic background and gene expression mechanism in human prostate cancer cells of each patient might additionally determine the degree of stomatin expression levels when cancer cells contact to stromal cells. This point will be explored in a future study. Nonetheless, the lower stomatin expression in the prostate cancer was significantly associated with the higher recurrence of the cancer after the operation.
The SPFH domain protein family in mammals includes stomatin-like protein, prohibitin and flotillin as well as stomatin. Stomatin-like protein 2 (SLP-2) plays a role in the enhancement of tumor progression. SLP-2 is overexpressed in esophageal carcinoma and promotes cell proliferation and adhesion (28). The overexpression of SLP-2 is associated with poor patient prognosis in ovarian cancer (29). SLP-2 also inhibits apoptosis in cervical cancer (30), and induces cell survival in non–small lung cancer (31). Prohibitin localized in the nucleus acts as a component of a transcription repressor in prostate cancer cells by recruiting Rb to inhibit the E2F transcription factor, resulting in the suppression of cell cycle (32, 33). In contrast, prohibitin on the plasma membrane is required for c-Raf activation and promotes cell adhesion and migration (34), whereas prohibitin in the mitochondria prevents apoptosis together with SLP-2 in T-cell lymphoma (35). Thus, prohibitin has both tumor promoting and suppressive effects that are determined by its intracellular localization. Flotillin is overexpressed in multiple types of cancers and promotes tumorigenesis (36). Together these studies show that each member of the SPFH family has a different role and unique property in cancer cells in its expression context.
Although this study revealed that stomatin acts as a tumor suppressor by inhibiting cancer cell proliferation and enhancing apoptosis in both in vitro and in vivo models, how the murine homologue of stomatin affects tumor progression remains unclear. As stomatin knockout mice show no obvious phenotype in the physiological state (12), it might be interesting to examine whether the knockout mice are vulnerable to tumorigenesis and/or whether tumor progression including invasion and metastasis is rapid in the stomatin knockout mice compared with wild-type mice. In addition, better understanding of the mechanism regulating stomatin gene expression might be attractive for future perspectives of cancer treatment. Another study demonstrates that hypoxia and dexamethasone treatment induce stomatin expression in lung adenocarcinoma cells (15). Because dexamethasone has a wide range of pharmacologic effects against many diseases including cancers (37), dexamethasone may be useful for anticancer therapy.
Continuous efforts to identify the detailed mechanism for the efficient induction of stomatin in cancer cells are required for the development of potential novel anticancer therapies related to stomatin. These findings could contribute to the improvement of prognosis in patients with cancer.
No disclosures were reported.
N.I.A. Rahman: Conceptualization, formal analysis, investigation, writing–original draft. A. Sato: Conceptualization, formal analysis, supervision, funding acquisition, investigation, writing–original draft. K. Tsevelnorov: Formal analysis, investigation. A. Shimizu: Formal analysis, investigation, writing–review and editing. M. Komeno: Investigation. M.K.B. Ahmat Amin: Investigation. M.R. Molla: Investigation. J.E.C. Soh: Investigation. L.K.C. Nguyen: Investigation. A. Wada: Data curation, formal analysis, writing–review and editing. A. Kawauchi: Conceptualization, data curation, formal analysis, supervision, investigation, writing–review and editing. H. Ogita: Conceptualization, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing.
The authors thank Dr. Sachiko Tanaka, Dr. Masatsugu Ema, Dr. Suzuko Moritani, Dr. Atsushi Takano, Mr. Takefumi Yamamoto, Mr. Michiharu Nasu, Mr. Yasuhiro Mori, and Ms. Nao Kougami at Shiga University of Medical Science for their excellent technical assistance in this study. This study was supported by grants-in-aid for Scientific Research <KAKENHI> from Japan Society for the Promotion of Science (16K08623 to A. Sato; 23112508, 26460389, 17K08627, and 20K08489 to H. Ogita), Shiga University of Medical Science In-House Grant to N.I.A. Rahman, The Uehara Memorial Foundation to H. Ogita, Takeda Science Foundation to H. Ogita, and The Naito Foundation to H. Ogita.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.