Ese-3 contributes to colon cancer progression by downregulating EHD2 and transactivating INPP4B

Abstract

Epithelium-specific Ets protein 3 (Ese-3), a member of the Ets family of transcription factors, plays an important role in the development of cancers. However, little is known concerning its role in colon cancer (CC). In this study, we demonstrate that the expression of Ese-3 is upregulated in CC tissues and elevated Ese-3 expression is relationship with advanced T stage (P=0.037) and poor disease-free survival (DFS, P=0.044). Univariate and multivariate cox regression analyses show that Ese-3 expression may be an independent prognostic value for CC patients. Moreover, Ese-3 knockdown suppresses CC cell proliferation in vitro and in vivo, while Ese-3 overexpression has the opposite result. Further, we first demonstrate that EHD2 and INPP4B are the downstream genes of Ese-3. Subsequent investigation find that EHD2 is downregulated in CC tissues and knockdown of EHD2 significantly increase CC cell proliferation in vitro and vivo. Our findings reveal that Ese-3 promotes CC cell proliferation by downregulating EHD2 and transactivating INPP4B, and targeting the pathway may be a promising therapeutic target for CC patients.

Introduction

Colon cancer (CC), which has one of the highest incidence rates and one of the mortality rates among gastrointestinal tumors worldwide [1,2]. Despite with the development of various treatment methods including chemotherapy, radiation, and biotherapy, the 5 years survival rate for the patients with CC is no more than 40% due to the proliferation, metastasis and recurrence of CC [3]. Therefore, the molecular mechanisms of CC need to be further explored and novel therapeutic targets need to be identified.

Ese-3/EHF, a member of the Ets family of transcription factors, is widely expressed in various epithelial cells [4]. All members of the Ets family contain a highly conserved DNA-binding domain and recognize a specific binding site (GGAA/T) of within their target genes regions [5]. Increasing evidence indicates that Ese-3 could be a tumor oncogene or tumor suppressor gene in the developmental programs of various tumors. Domenico and Giuseppina reported that the loss of Ese-3 could promote prostate tumor progression by activating the Lin28/let-7 axis and deregulating the miR-424/COP1/STAT3 axis pathways [6,7]. Ese-3, as a negative tumor regulator, is downregulated and inhibits cell invasion and metastasis by upregulating E-Cadherin in pancreatic cancer [8,9]. However, Ese-3 as an oncogene in some other tumor types, can induce tumor proliferation and metastasis in gastric cancer [10], oral squamous cell carcinoma [11] and ovarian Cancer [12]. Although, Ese-3 plays an important role in the development of tumors by certain specific behaviors and signaling pathways, its roles and molecular mechanism remain unclear in CC.

Eps15 homology domain-containing 2 (EHD2), a member of the EHD family, was initially reported as a plasma membrane-associated protein, but its precise function has been mysterious [13]. A series of reports have shown that EHD2 is a negative regulatory gene and EHD2 expression is significantly correlated with clinicopathologic parameters in tumors [14]. Decreased expression of EHD2 facilitated tumor invasion and metastasis in breast cancer [15], papillary thyroid carcinoma [16], esophageal squamous cell carcinoma [17] and induced the proliferation of clear cell renal cell carcinoma (ccRCC) [18] and hepatocellular carcinoma [19]. However, the role and function of EHD2 and its link to Ese-3 in CC have not been reported.

In this study, we demonstrated that Ese-3, a member of the Ets family of transcription factors, was upregulated in CC and that elevated expression of Ese-3 was dramatically correlated with poor disease free survival and advanced clinical T stages. Next, in vitro and in vivo assays results demonstrated that Ese-3 could accelerate the proliferation of CC. Further results showed Ese-3 could promote the activation of AKT by restraining the expression of EHD2 and targeting the INPP4B promoter region. Thus, this novel pathway may be valuable as a therapeutic approach for patients with CC.

Materials and methods

Patient specimens

Five paired CC and peritumoral tissue samples were obtained from CC patients who underwents urgical excision in Tangdu hospital affiliated with Air Force Military University (Xi’an, China). Every tissue sample were divided into two parts, one part of them were frozened in liquid nitrogen until use, the other were made into paraffin sections. Characterization of CC patients were described (Supplementary Table 1). Our study was supported by the Ethics Committee of Tangdu hospital. All subjects patients with CC gave the informed consent to participate in this study.

Cell lines and cell culture

In our study, we purchased CC cells (HCT116, HT29, LoVo, WiDr, and SW620) from Procell Life Science & Technology Company (Wuhan, China) and BeNa Culture Collection (BNCC, Beijing, China). All these cells have the STR authentication report. HCT116 and HT29 cell lines were cultured in McCoy’s 5A media (Gibco, USA), WiDr and SW620 cell lines were grown in Dulbecco’s Modified Eagle Medium media (Gibco, USA), and LoVo cell lines were maintained in Dulbecco’s Modified Eagle Medium/F12 media. All CC cells were cultured in media with 10% fetal bovine serum (FBS), penicillin-streptomycin (Gibco, USA) and were incubated at 37°C in 5% CO₂.

Immunohistochemistry (IHC)

We purchased the human CC tissue microarray from Outdo Biotech Company (HColA180su16, Shanghai, China). Tissue microarray and tumor tissue sections were deparaffinized using the xylene, then rehydrated using the different concentration of ethanol. 3% hydrogen peroxide was used to inhibit endogenous peroxide activities and immediately to repair antigens through heating in citrate buffer (pH 6.0) using a microwave. Then, the 5% normal goat serum was used to block the tissue microarray and sections for 30 minutes. After being washed in PBS, the microarray and sections were incubated overnight at 4°C with the primary antibodys against Ese-3 (1:100, NBP2-14942, Novus), EHD2 (1:100, bs-14526R, Bioss), and Ki67 (1:1000, 27309-1-AP, Proteintech). The next day, after washing in PBS solution for 10 minutes, the peroxidase-conjugated secondary antibody was used to incubate the microarray and sections for 30 minutes, and then these tissue microarray and tumor tissue sections were stained with diaminobenzidine for 3 minutes. Finally, the slides were counterstained using hematoxylin.

The IHC staining results were assessed based on the stained cells percentage of (0, 0-5%; 1, 6-25%; 2, 26-50%; 3, 51-75%; 4, 76-100%) and staining intensity (0, no staining; 1, faint yellow; 2, reddish; 3, brown). Then the two scores were multiplied to obtain the final result. A final staining score <6 was suggested as low Ese-3 expression, while a score of staining >6 was considered as high Ese-3 expression.

Lentivirus infection and stable cell line construction

The pLent-U6 lentiviral plasmid (Vigene Biosciences, Jinan, China) was used to generate lentivirus with shEse-3#1, shEse-3#2, shEHD2 and shControl sequences. The shControl sequence was used as the negative control (Supplementary Table 2). Lentiviral vectors encoding the human Ese-3 and EHD2 genes were constructed in the pLent-EF1a plasmid (Vigene Biosciences, Jinan, China) and named Ese-3 and EHD2. All the plasmids with packaging plasmids (pMD2G and psPAX2) were cotransfected into HEK293T cells, respectively. After transfecting for 48-72 hours, we collected the viral supernatants. The CC cells were infected with lentivirus. After 12 hours, 4 μg/ml puromycin was used to select these cells (OriGene, USA) for one week to establish stable CC cell lines.

Western blotting

All CC cells, CC and matched control tissues were lysed using lysis buffer (Applygen, Beijing, China) containing a protease and phosphatase inhibitor cocktail (Roche, Branchburg, USA). SDS-PAGE was used to fractionate the protein samples, then the protein bands were transferred onto PVDF membranes. 3% nonfat milk was used to block the PVDF membranes for 3 hours, then the PVDF membranes were incubated using appropriate antibodies overnight at 4°C (Supplementary Table 3). The next day, the PVDF membranes were incubated with HRP-conjugated secondary antibodies. Finally, the protein bands were detected with a Tanon-5200 chemiluminescent system (Shanghai, China).

Cell proliferation assay

The cell counting chamber was used to assess the CC cell proliferation ability. Briefly, CC cells were plated into 6-well plates in triplicate at the density of 3×10⁴/well. After digestion, each well cell number was counted using the cell counting chamber at the indicated time point.

Colony-formation assay

The cell colony formation assay was performed using 6-well plates. A total of 800 cells/well were seeded on the plates in triplicates. After 10 days, the cells of 6-well plate were fixed using the anhydrous methanol and stained using the 1% crystal violet buffer for 20 minutes, then the plates were washed with running water to visualize the colony-formation ability.

Flow cytometry analysis

For flow cytometry, the CC cells (4×10⁵ cells/plate) were plated in 6-well plates in triplicate and cultured for 24 hours. Then, the CC cells were collected and fixed with 70% ethanol for 30 minutes. Finally, the cells were stained using the PI/RNase solution for 20 minutes (Thermo, USA) and subjected to flow cytometry to perform cell cycle detection (BD FACSCalibur, USA).

Label-free quantitative LC/MS proteomics analysis

Knockdown of Ese-3 in HCT116 cells and control group cells were seeded on 100 mm cell culture dishes in triplicate, respectively. When the cells reached 80-90% of the cell culture dish, the cells were digested with trypsinization and washed with cold PBS for two times. Label-free quantitative LC/MS proteomics was performed by Mhelix Biotech Company (Shanghai, China).

Luciferase reporter assay

Luciferase reporter assay was used to detect the luciferase activity (Promega, E1910, USA) according to the experimental protocol of the manufacturer. The CC cells were transfected with plasmids. After two days, the cells were lysed with lysis solution, and then the relative luciferase activity was detected by a GloMaxTM 20/20 Microplate Luminometer (Promega, E1910, USA). All experiments were performed three times independently.

Ch-IP and real-time PCR assays

Ch-IP assays were performed using the Pierce™ Agarose ChIP Kit (Thermo, USA). Briefly, CC cells were cross-linked for 20 minutes with 1% formaldehyde, and then cross-linked chromatin was sonicated into small DNA fragments. Sonicated DNA fragments were immunoprecipitated with Ese-3 antibody (Proteintech, 27195-1-AP) and subjected to real-time PCR to amplify the corresponding fragment. The primers are displayed in Supplementary Table 4.

Animal studies

We purchased male BALB/c nude mice (4-6 weeks) from Beijing Vital River Laboratory Animal Technologies company (Beijing, China), and all nude mice were kept in an SPF environment. A total of 2-3×10⁶ CC cells were injected into subcutaneous of the nude mice (n=4 or 5). Every three days, the tumor volume of the nude mice was monitored (volume = length × width² ×0.5). At the end of the assay, the nude mice were sacrificed, and their subcutaneous tumors were weighed and imaged. Lastly, the nude mice tumors were fixed with paraformaldehyde or frozen in liquid nitrogen until use. All animal experiments were approved by the Institutional Animal Care and Use Committee of Tangdu Hospital.

Statistical analysis

All quantitative experimental data are showed as the mean ± SD of three independent replicates and were compared between two groups using Student’s t-test, while ANOVA was used to assess differences among more than two groups. The correlation between Ese-3 expression levels and clinicopathological parameters was assessed using the chi-square test. Pearson correlation coefficient analysis was used to assess the association of Ese-3 mRNA levels and EHD2, INPP4B mRNA levels in prospective_CPTAC_COAD samples. Kaplan-Meier curves and log-rank tests were used to assess the correlation between Ese-3 expression and disease-free survival. Univariate or multivariate cox regression analyses were performed. The diagnostic value of Ese-3 was assessed by ROC curve analysis. SPSS 19.0 or GraphPad Prism 5.0 were used to analyze these experimental data (SPSS, Chicago, Illinois, USA) (La Jolla, CA, USA). P value <0.05 was considered as statistically significant.

Results

Enhanced Ese-3 expression is associated with poor prognosis in CC

To illuminate the potential function of Ese-3 in CC, we first analyzed the mRNA expression levels of Ese-3 using TCGA and GTEx databases in CC samples (n=275) and the normal control group (n=349). Ese-3 mRNA expression levels in CC cancer tissues were significantly upregulated compared with normal control tissues (P<0.05, Figure 1A). Western blotting results showed that Ese-3 expression levels also were elevated in CC tissues compared with paired peritumor tissues (Figure 1B). This result was further confirmed by IHC staining. In CC cancer tissues Ese-3 staining was strongly positive, but Ese-3 staining was weakly positive in matched peritumor tissues (Figure 1C). Then IHC staining was used to analyze the association between Ese-3 expression and clinical parameters using a tissue microarray. Statistical analysis revealed that Ese-3 expression significantly correlated with advanced T stage (P=0.037), but not age, gender, clinical grade or N stage, TNM stage (Table 1, Supplementary Table 5). Furthermore, Kaplan-Meier curves and the long-rank test analysis showed that higher Ese-3 mRNA expression levels were correlated with poorer DFS (P=0.044, Figure 1D). Next, TCGA samples were used to assess Ese-3 prognostic value by univariate and multivariate Cox regression analyses. Univariate analysis showed that Ese-3 expression, T stage, N stage, M stage and TNM stage could be used to predict the disease free survival time (Figure 1E). Importantly, multivariate analysis showed that the expression of Ese-3 was an independent predictor value of DFS in CC (HR=1.013, 95% CI (1.005-1.022, P=0.002) (Figure 1F). Taken together, these results reveals that the Ese-3 expression level is up-regulated in CC and might as an independent prognostic value in patients with CC.

Figure 1.

Elevated Ese-3 expression is associates with poor prognosis in CC. A. Ese-3 mRNA expression levels in CC and normal tissue samples were analyzed using TCGA and GTEx samples (P<0.05). B. Western blotting was used to analyze Ese-3 protein expression levels in 5 matched pairs of CC and peritumoral tissues (Original blots were showed in Supplementary Figure 3). C. Representative IHC staining of Ese-3 in matched CC and peritumoral tissues. Scale bar =500 μm (left) or Scale bar =50 μm (right). D. Kaplan-Meier disease-free survival analysis of Ese-3 expression in patients with CC using TCGA samples (n=317, P=0.044, log-rank test). E, F. Univariate (E) and multivariate (F) analyses of the relationship of disease free survival with clinicopathological characteristics and Ese-3 expression in TCGA samples.

Table 1.

Correlation of Ese-3 expression with clinicopathological features of CC

Parameters	Total (n)	Ese-3 expression (IHC)
		Low	High	χ 2	P value	r
Gender
Male	49	10	39	0.654	0.419	0.077
Female	51	15	36
Age (years)
<60	17	3	14	0.213	0.645	0.103
≥60	83	22	61
T
I	1	0	1	8.478	0.037*	0.3078
II	5	2	3
III	75	14	61
IV	18	9	9
N
N0	60	13	47	0.500	0.480	0.094
N1-2	40	12	28
TNM stage
I	6	2	4	3.957	0.266	0.1872
II	54	11	43
III	49	11	38
IV	1	1	0
Pathological grade
I	22	2	20	4.064	0.131	0.1976
II	73	22	51
III	5	1	4

^*Statistically significant (P<0.05).

Ese-3 promotes CC cell proliferation in vitro and in vivo

To further assess the functional role of Ese-3 in CC in vitro, firstly, the protein expression levels were detected in different tumor cells (Supplementary Figures 1A, 10H1, 10H2). Then, we established stable Ese-3 knockdown CC cells (HCT116 and HT29) using two different shRNA sequences, and stable Ese-3-overexpressing WiDr and HCT116 cells. Western blotting was performed to confirm the successful knockdown and overexpression of Ese-3 in CC cells (Figure 2A-C, Supplementary Figures 1B, 10H3-H5). The knockdown of Ese-3 obviously decreased CC cell proliferation (Figure 2D, 2E) and clonogenic ability (Figure 2G, 2H). Flow cytometry analysis of the cell cycle revealed that knockdown of Ese-3 significantly increased the number of G0/1, and G2/M phase cells and concomitantly decreased the number of S phase cells in HCT116 and HT29 cells (Figure 2J, 2K). In contrast, enriched Ese-3 expression remarkably promoted CC cell growth (Figure 2F, Supplementary Figure 1C) and colony formation (Figure 2I, Supplementary Figure 1D). Overexpression of Ese-3 resulted in an increase in G0/1 phase cells and a decrease in S phase cells in WiDr cells (Figure 2L). Then we further explored the impact of Ese-3 on CC cells in vivo. Ese-3-silenced HCT116 cells with luciferase and control cells, Ese-3-overexpressing HCT116 cells and the corresponding control group were subcutaneously into the subcutaneous nude mice, the results showed that decreased Ese-3 cells reduced the bioluminescence intensity, tumor volume and tumor weight, while increasing Ese-3 expression increased tumor volume and tumor weight (Figure 3A, 3B, Supplementary Figure 1E-G). Moreover, ki67 expression was decreased in Ese-3-silenced HCT116 cells compared with control group cells (Figure 3C).

Figure 2.

Ese-3 promotes CC cell proliferation in vitro. A-C. Western blotting analyzed the Ese-3 protein expression levels in Ese-3-silenced HCT116 cells, HT29 cells and Ese-3-overexpressing WiDr cells (Original blots were showed in Supplementary Figure 4). D-F. Cell proliferation assay analysis of the impact of Ese-3 on growth in Ese-3-silenced HCT116 and HT29 cells and Ese-3-overexpressing WiDr cells. G-I. Ese-3-silenced HCT116 and HT29 cells and Ese-3-overexpressing WiDr cells were subjected to colony formation assay. J-L. Ese-3-silenced HCT116 and HT29cells and Ese-3-overexpressing WiDr cells were used to evaluate the cell cycle by flow cytometry assay. Data are displayed as the mean ± SD of three in dependent replicates (*P<0.05; **P<0.01; ***P<0.001).

Figure 3.

Ese-3 promotes CC cell growth in vivo and regulates the expression of EHD2, INPP4B and p-AKT. A. Representative bioluminescence imaging showed that knockdown of Ese-3 suppressed the growth of subcutaneous tumors in nude mice (n=4). B. Total flux, volume and weight of tumors were quantified and measured. C. IHC staining was used to measure the Ese-3 and Ki67 expression in nude mice tumor tissues (scale bars =50 µm). D, E. Volcano plot and heat map showing some differentially expressed proteins between Ese-3-silenced HCT116 and control group cells was identified by label-free quantitative proteomics technology. F. The expression level of Ese-3 was negatively correlated with that of EHD2 (P=0.021, r=-0.224) and positively correlated with that of INPP4B (P=0.42, r=0.198) (linear regression). G. Western blot analysis of the indicated proteins expression levels in Ese-3-silenced and Ese-3-overexpressing cells (Original blots were showed in Supplementary Figure 5).

Ese-3 regulates two tumor-related genes, EHD2 and INPP4B

To explore the underlying mechanisms of Ese-3 promoting CC proliferation, label-free quantitative proteomics technology was used to detect the changes of protein expression levels mediated by Ese-3. The results showed that 231 protein expression levels were upregulated and 234 protein expression level were downregulated in Ese-3 silenced HCT116 cells compared with control group cells (Figure 3D; Supplementary Table 6). Among these proteins, EHD2 protein levels were upregulated and INPP4B protein levels were reduced in Ese-3-silenced HCT116 cells compared with control group cells (Figure 3E). We further confirmed that Ese-3 expression levels were significantly negatively associated with EHD2 expression levels (R²=0.05, r=-0.224, P=0.021) and obviously positively correlated with INPP4B expression levels (R²=0.039, r=0.198, P=0.042) in 106 prospective_CPTAC_COAD samples (Figure 3F; Supplementary Table 7). In addition, Western blotting assays further revealed that EHD2 was significantly downregulated and INPP4B was remarkably upregulated in Ese-3 knockdown HCT116 and HT29 cells. Meanwhile, we also found that the phosphorylation levels of AKT were decreased in Ese-3-silenced cells, while Ese-3 overexpression in WiDr cells increased INPP4B expression and AKT phosphorylation levels and reduced EHD2 expression (Figure 3G).

EHD2 is downregulated in CC

In our current study, EHD2 protein expression levels were assessed in 5 matched pairs of CC tissues and normal tissues and EHD2 expression levels were significantly downregulated in CC tissues (Figure 4A). Further IHC staining analysis confirmed that EHD2 protein expression levels were also reduced in CC tissues compared with peritumoral tissues (Figure 4B). In agreement with these results, EHD2 mRNA levels were obviously lower in CC samples than that in normal control group by analyzing CPTCA and TCGA samples (Figure 4C, 4D, P<0.001). Importantly, the mRNA expression levels of EHD2 in the normal control sample were significantly higher than that in the other T stages, including T1 (P<0.05), T2 (P<0.001), T3 (P<0.001) and T4 (P<0.01) (Figure 4E). In addition, EHD2 mRNA levels were remarkably downregulated in TNM stage I (P<0.001), TNM stage II (P<0.001), TNM stage III (P<0.001), and TNM stage IV (P<0.001) compared with the normal control in CPTCA and TCGA samples (Figure 4F, 4G). ROC curve analysis showed that EHD2 as a good diagnostic value in patients with CC (AUC=0.798; P<0.0001) (Figure 4H). These findings suggest that EHD2 could be used as a promising biomarker in patients with CC.

Figure 4.

EHD2 is downregulated in CC. A. EHD2 protein expression levels were measured in 5 matched pairs of CC and peritumoral tissues (Original blots were showed in Supplementary Figure 6D1, 6D2). B. Representative IHC staining analysis of EHD2 expression in CC and peritumoral tissues. Scale bar =500 μm (left) or Scale bar =50 μm (right). C, D. EHD2 expression levels were analyzed in the CC and normal control groups using TCGA and CPTAC samples. E. The expression levels of EHD2 were analyzed in T1, T2, T3, T4 and normal tissues using TCGA samples. F, G. CPTAC and TCGA sample analysis of EHD2 expression levels in TNM stage I, TNM stage II, TNM stage III, TNM stage IV and normal control samples. H. The diagnostic value of EHD2 for CC was evaluated using the ROC curve. I. EHD2 protein expression levels were measured in different CC cells (Original blots were showed in Supplementary Figure 6D3, 6D4). J. Knockdown of EHD2 in HCT116 cells were performed cells proliferation assay. K. Colony formation assay was used to analyze the impact of EHD2 on growth in EHD2-silenced HCT116 cells. L. The cell cycle was examined in EHD2-silenced HCT116 cells using the flow cytometry assay. Data are shown as the mean ± SD of three independent replicates (*P<0.05; **P<0.01; ***P<0.001).

EHD2 inhibits CC cell proliferation in vitro and in vivo

Firstly, the protein expression levels of EHD2 were assessed in different CC cells. The results indicated that EHD2 expression was lower in SW620 cells than in other CC cancer cells (Figure 4I). We then established stable EHD2-overexpressing SW620 cells, and EHD2-silenced HCT116 and HT29 cells with lentivirus infection. The downregulation of EHD2 significantly increased tumor cell proliferation and colony formation ability in HCT116 and HT29 cells, while EHD2 overexpression inhibited tumor cell growth and capacity of colony formation in SW620 cells (Figures 4J, 4K, 5A-D). Furthermore, flow cytometry analysis showed that EHD2-silenced in HCT116 and HT29 cells significantly decreased G0/1 phase cells and increased S phase cells (Figures 4L, 5E). Consistent with these results, increased EHD2 expression remarkably increased the number of G0/1 phase cells and reduced the number of S-phase cells (Figure 5F). Next, we performed a nude mice tumorigenicity assay. Knockdown of EHD2 significantly increased the volume and weight of nude mice tumors (Figure 5G-I). In addition, IHC staining analysis indicated that ki67 expression levels were upregulated in EHD2-knockdown tumors compared with the control group (Figure 5J). Taken together, these experimental results suggest that EHD2 is a negative regulatory gene in CC.

Figure 5.

Effects of EHD2 on CC cell growth in vitro and vivo. A, B. Cell proliferation assay was performed to assess the impact of EHD2 knockdown (A) and overexpression (B) on CC cell growth. C, D. Colony formation assay was used to detect EHD2-silenced HT29 cells and EHD2-overexpressing SW620 cells. E, F. Flow cytometry assay was performed to examine the cell cycle in EHD2-silenced HT29 cells and EHD2-overexpressing SW620 cells. G. The subcutaneous tumors of nude mice were separated and measured. H. The volume of xenograft tumors was monitored every 3 days in EHD2-silenced HCT116 and control cells. I. The tumors of the two groups were weighed. J. EHD2 and ki67 expression levels in the two groups were measured by IHC staining. Data are presented as the mean ± SD of three independent replicates (*P<0.05; **P<0.01; ***P<0.001).

EHD2 and INPP4B contribute to Ese-3-mediated CC cell proliferation

We further focused on examining the molecular mechanism of Ese-3-mediated CC cell proliferation. We first found that EHD2 knockdown can activate the phosphorylation of AKT in HCT116, HT29 and LoVo cells, while upregulated EHD2 decreased AKT phosphorylation levels in SW620 cell lines (Figure 6A, Supplementary Figures 2A, 11I1-I4), further implying that EHD2 is a tumor suppressor gene in CC by regulating AKT phosphorylation. To confirm the importance of EHD2 and INPP4B in Ese-3-mediated oncogenic function in CC, we knocked down the EHD2 gene in Ese-3-silenced HCT116, HT29, and LoVo cells, and the results showed that knockdown of EHD2 rescued the phosphorylation levels in Ese-3-knockdown HCT116, HT29, and LoVo cells (Figure 6B, Supplementary Figures 2B, 11I5-I9). Then, we transplanted Ese-3 plus EHD2-knockdown HCT116 cells with control Ese-3-silenced HCT116 cells into the back of nude mice, and the results showed that knockdown of EHD2 restored the growth capacity of HCT116 cells in the Ese-3-silenced xenograft group, which led to increase tumor volume and tumor weight (Figure 6C-E).

Figure 6.

Ese-3 promotes CC cell proliferation by downregulating EHD2 and transactivating INPP4B. A. The indicated protein expression levels were detected in EHD2-knockdown and EHD2-overexpressing cells (Original blots were showed in Supplementary Figure 7). B. Western blotting analysis of the indicated protein expression levels in HCT116 and HT29 cells transduced with shControl, shEse-3 or shEse-3 plus shEHD2 (Original blots were showed in Supplementary Figure 8). C. ShEse-3#1-silenced HCT116 cells with or without co-transduction with shEHD2 were injected into the subcutaneous of nude mice (n=5). The tumors that developed in nude mice were isolated and measured at the end of this cell injection experiment. D. The growth curves of the two groups were compared. E. Comparison of the weight between the two groups. F. Relative luciferase activity was measured to analyze the effect of Ese-3 on INPP4B promoter activity in HCT116 cells. G. Luciferase reporter assays were performed to analyze the impact of Ese-3 on the serially truncated INPP4B promoter in HCT116 cells. H. The effect of Ese-3 on truncated INPP4B-4 wild-type or mutant promoters were detected by luciferase reporter assays. I. Ch-IP assays were used to detect the binding site of Ese-3 to INPP4B in HCT116 and HT-29 cells (Original gels were showed in Supplementary Figure 9). J. The ability of Ese-3 bind to the INPP4B promoter was determined using qRT-PCR in HCT116 and HT29 cells. Data are presented as the mean ± SD of three independent replicate experiments (*P<0.05; **P<0.01; ***P<0.001).

In our previous study, chromatin immunoprecipitation (ChIP) sequence assay analysis found that INPP4B might be a target gene of Ese-3. Further, we explored the molecular mechanism of Ese-3-mediated INPP4B. First, luciferase reporter assay results indicated that upregulated Ese-3 significantly transactivated the promoter activity of INPP4B (Figure 6F). Then, the INPP4B promoter sequence was analyzed, finding seven potential Ese-3 binding sites at the INPP4B promoter region. To further confirm which site of Ese-3 binds to the INPP4B promoter region, we established a serially truncated INPP4B promoter construct. We found that when Ese-3 binds to the region between -1043 to -640 bp, the luciferase activity was much higher in Ese-3- overexpressing HCT116 cells than in the control group (Figure 6G). In agreement with the results, the mutation between the -648 to -641 INPP4B promoter site significantly decreased the activity of the luciferase reporter compared with the wild-type group (Figure 6H). Ch-IP and qRT-PCR assays further confirmed that Ese-3 binds to the -648 to -641 bp site (Figure 6I, 6J). The results suggest that the -648 to -641 bp region is vital for Ese-3-mediated INPP4B promoter transactivation.

Discussion

Great progress has been achieved in the diagnosis and treatment of CC, one of the most common malignant tumors of the digestive tract, but the duration of survival remains unfavorable due to a lack of good biomarkers as well as high rates of tumor metastasis and recurrence [20]. Thus, in order to identify new therapeutic targets, there is an urgent need to explore the molecular mechanisms underlying the development of CC. In our study, we found that Ese-3 knockdown inhibited the growth and colony formation of CC cells and induced cell cycle arrest in vitro. We also verified that knockdown of Ese-3 inhibited the tumorigenicity of nude mice in vivo. Furthermore, decreased Ese-3 expression in CC cells significantly inhibited the AKT phosphorylation. These results suggest that Ese-3 might be a potential therapeutic target in CC patients.

Ese-3/EHF, which belongs to the ESE subfamily, alongside Ese-1 and Ese-2, has a highly conserved 85-amino acid ETS DNA binding region that binds to the promoters of target genes [21]. A series of studies have shown that Ese-3 is expressed in differentiated epithelial cells of various organs and participates in cell development, differentiation and homeostasis [22-24]. Studies on ovarian cancer have shown that decreased Ese-3 expression can inhibit tumor proliferation and invasion by activating the ERK and AKT signaling pathways [12]. Knockdown of Ese-3 expression can increase the conversion and expansion of regulatory T cells by decreasing the secretion of TGFβ1 and GM-CSF by the tumor, thereby increasing the effect of checkpoint immunotherapy [25]. In the present study, we found that Ese-3 was upregulated in CC tissues compared with normal control tissues using the TCGA and GTEx samples, and this result was corroborated by Western blotting and immunohistochemistry (IHC) assays. Our data also indicated that higher expression of Ese-3 predicted poorer disease-free survival for CC patients. Furthermore, univariate and multivariate analyses showed that Ese-3 could improve the predictive capacity of the clinical hazard ratio. However, the underlying molecular mechanism of Ese-3 in CC remains unclear.

To identify the underlying molecular mechanisms by which Ese-3 promote CC proliferation, we performed label-free quantitative proteomics analysis of differentially expressed proteins in control cells compared with Ese-3 silenced CC cells. Mass spectrometry (MS) has been developing rapidly, and modern proteomics can not only identify the types of proteins but also quantify their expression levels. Label-free quantitative proteomics technology has some advantages compared with tandem mass tags (TMT) and isobaric tags for relative and absolute quantitation (iTRAQ) in terms of throughput and the ability to bypass a great deal of manual operation and expensive isotope and metabolic labeling [26,27]. Labelfree quantitative proteomics analysis found that a total of 465 proteins were differentially regulated (231 up-regulated and 234 down-regulated) in Ese-3 silenced HCT116 cells compared with control cells, of which two novel genes involved in the Ese-3-mediated CC proliferation, namely EHD2 and INPP4B. Knockdown of Ese-3 in CC cells decreased the expression levels of EHD2 and increased the expression levels of INPP4B compared with control group cells. These results were further verified by Western blotting analysis and CC samples.

EHD2 is a member of Eps15 homology domain-containing 2, so far, only a little papers have reported that EHD2 is involved in the regulation of tumor development, such as cell proliferation, and metastasis. A series of researchs have shown that EHD2 as an oncogene in ccRCC, and high expression levels of EHD2 in ccRCC contributed to tumor proliferation, invasion, and migration and inhibited tumor cell apoptosis [18]. However, EHD2 has been reported as a negative regulatory gene, and elevated EHD2 expression can inhibit tumor metastasis, invasion and indicate poor prognosis in papillary thyroid carcinoma, hepatocellular carcinoma and esophageal squamous cell carcinoma [15-17]. In Colorectal Cancer, a reporter has shown that the expression of EHD2 is downregulated, but the function and underlying molecular mechanisms remain unclear [28]. In this study, we showed that EHD2 expression was downregulated in CC tissues, and the result was verified using the TCGA and CPTAC samples. Then we demonstrated for the first time that the downregulation of EHD2 in CC cells significantly improved tumor cell proliferation and colony formation ability in vitro assay. Conversely, overexpression of EHD2 markedly inhibited CC cell proliferation and colony formation. We also established for the first time that the mechanism by which EHD2 inhibit CC cell proliferation is mediated by the regulation of AKT phosphorylation. In accordance with the in vitro assay, decreased expression of EHD2 promoted tumor growth in nude mice xenograft models. EHD2 depletion resulted in AKT activation in Ese-3-knockdown CC cells, and restored their abolished proliferation capacity in Ese-3-knockdown xenograft nude mice. These results suggest that Ese-3 can promote CC cell proliferation by suppressing its downstream target gene, EHD2.

Inositol polyphosphate 4-phosphatase type II (INPP4B), a phosphoinositide phosphatase, plays an important role in the development of various cancers by the PI3K/Akt signaling pathway [29,30]. Previous works showed that INPP4B expression levels were upregulated and that elevated INPP4B expression could accelerate cell proliferation, metastasis, and invasion by regulating the AKT and SGK signaling pathways in cervical cancer [31], prostate cancer [32,33], and hepatocellular carcinoma [34], suggesting that INPP4B is a negative tumor regulatory gene in these cancers. Interestingly, some reports have shown that INPP4B-mediated autophagy or SGK3, and AKT activation can promote cell survival and suppresses apoptosis in some other types of cancers, including leukemia [35,36] and pancreatic cancer [37]. In CC, a series of studies have verified that INPP4B can promote tumor cell proliferation by AKT signaling of INPP4B-mediated PTEN [38] and contribute to colorectal cancer proliferation through the mTORC1 signaling pathway [39]. These results showed that INPP4B plays a vital role in promoting CC cell proliferation. In our study, elevated Ese-3 expression promoted the expression of INPP4B. Meanwhile, our Ch-IP and qRT-PCR results showed that INPP4B was the target gene of Ese-3. Furthermore, our results showed that Ese-3 directly binds to the INPP4B promoter at -648 to -641 bp to promote its transcription. These results suggest that Ese-3 promotes CC cell proliferation by promoting INPP4B transcription, and then activating AKT phosphorylation.

In conclusion, our study demonstrates that Ese-3 is upregulated in CC tissues, and is correlated with poor prognosis in patients with CC. Moreover, Ese-3 is an oncogene that promotes CC cell proliferation by downregulating the expression of EHD2 and promoting the transcription of INPP4B to activate the AKT signaling pathway (Figure 7). This new pathway may provide novel insights into CC, and targeting this new pathway could be a promising therapeutic strategy for patients with CC.

Figure 7.

Ese-3 promotes CC cell proliferation. EHD2 as a tumor suppressor gene that suppresses CC cell proliferation by inactivating AKT signaling pathway. Ese-3 facilitates CC progression through downregulating the expression of EHD2 and promoting the transcription of INPP4B to activate AKT signaling pathway.

Disclosure of conflict of interest

None.

Abbreviations

Ese-3

Epithelium-specific Ets protein 3

EHD2

Eps15 homology domain-containing 2

CC

Colon Cancer

INPP4B

Inositol polyphosphate 4-phosphatase type II

TCGA

The Cancer Genome Atlas

GTEx

Genoytpe-Tissue Expression Program

CPTAC

Cancer Proteomic Tumor Analysis Consortium

iTRAQ

Isobaric Tags for Relative and Absolute Quantitation

TMT

Tandem Mass Tags

ROC

Receiver operating characteristic

SDS PAGE

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis

PVDF

Polyvinylidene fluoride

IHC

Immunohistochemistry

Ch-IP

Chromatin Immunoprecipitation

qRT-PCR

Reverse transcription quantitative real-time PCR

DFS

Disease Free Survival