DUSP7 inhibits cervical cancer progression by inactivating the RAS pathway

Abstract To determine the differentially expressed proteins (DEPs) between paired samples of cervical cancer (CC) and paracancerous tissue by quantitative proteomics and to examine the effects of DUSP7 expression on the tumorigenesis and progression of CC. Proteomic profiles of three paired samples of CC and paracancerous tissue were quantitatively analysed to identify DEPs. The relationship between DEP expression and patient clinicopathological characteristics and prognosis was evaluated. The effects of the selected DEPs on CC progression were examined in SIHA cells. A total of 129 DEPs were found. Western blot and immunohistochemistry (IHC) staining analyses confirmed the results from quantitative proteomic analysis showing that the selected DEP, HRAS, P‐ERK1/2, and PLD1 levels were increased, whereas the DUSP7 level was decreased in CC tissue compared with the paired normal paracancerous tissues. The IHC results from the CC TMA analysis showed that the decreased expression of DUSP7 (p = 0.045 and 0.044) was significantly associated with a tumour size >2 cm and parametrial infiltration. In addition, the decreased expression of DUSP7 and increased expression of p‐ERK1/2 were adversely related to patient relapse (p = 0.003 and 0.001) and survival (p = 0.034 and 0.006). The expression of HRAS and p‐ERK1/2 was decreased in DUSP7‐SIHA cells compared with NC‐SIHA cells (p = 0.0003 and 0.0026). Biological functions in vitro, including invasion, migration and proliferation and tumour formation in vivo were decreased in DUSP7‐SIHA cells (all p < 0.05) but increased in shDUSP7‐SIHA cells (all p < 0.05). DUSP7 inhibits cervical cancer progression by inactivating the RAS pathway.


| INTRODUC TI ON
Human cervical carcinoma (CC) is the third most common cancer worldwide and is the most frequent gynaecological cancer in developing countries. China has approximately 150,000 new cases of CC each year, and CC tends to occur in younger people 1 . CC remains a serious health problem in women. It is therefore crucial to explore the pathogenesis of CC and discover effective therapeutic targets for this lethal disease. The key aetiological role of human papillomavirus (HPV) in the development of CC and its precursors has been well documented 2 , and the use of HPV vaccination in women has the potential to reduce the incidence of CC in the future. However, due to the high price of the vaccine, coverage rates are still low, especially in low-or middle-income countries 3 . Furthermore, viral presence is not sufficient to induce CC 4 , suggesting that a distinct molecular mechanism could play a key role in its transformation and progression.
Proteomics is defined as the comprehensive global analysis of a specific proteome, the set of all proteins expressed in a cell or a biological system or organism at a given point in time and under certain conditions. Proteomics has been widely used to identify certain proteins with complex biological functions related to the pathogenesis of various diseases, including human malignancies. Wang et al. 5 compared the proteomes of the primary tumours of CC patients with and without lymph node metastasis and revealed that patients with high FABP5, HspB1 and MnSOD expression have a high risk of lymph node metastasis and adverse prognosis. By comparing the proteomes of primary CC tissues and corresponding adjacent normal tissues, Zhang et al 6 found that Notch signalling, viral carcinogenesis, RNA transport and Jak-STAT signalling play an important role in tumour progression. In these studies, the study and control groups had the same genetic backgrounds. Differentially expressed proteins (DEPs) identified through this manner may reflect the process of tumorigenesis and progression of CC to some extent. However, no further details on the preparation of tissue specimens were provided. In this analysis, the tissue specimens available for proteomic examination were subjected to a stricter pathological evaluation so the results would more accurately reflect the factors involved in the invasion and progression of cervical cancer. The protein profiles between paired samples of CC and paracancerous tissues were compared and analysed, with the goal of providing useful information about diagnostic biomarker or molecular therapeutic targets for patients with CC.

| Materials and cell culture
The main materials, including reagents, instruments and antibodies for Western blot (WB) and immunochemistry (IHC), used in this work are shown in Table S1 and S2, respectively. The human CC cell line SIHA was cultured in RPMI-1640 supplemented with 10% foetal bovine serum (FBS) and antibiotics (100 U/ml penicillin and 100 μg/ ml streptomycin) at 37°C in a humidified atmosphere in a 5% CO 2 incubator.

| Specimen collection of CC and paracancerous tissue
CC and paracancerous tissue samples were all collected from patients who underwent surgery at the Department of Obstetrics and Gynecology, Beijing Chao-yang Hospital, Capital Medical University.
Those who received preoperative chemotherapy or radiotherapy or had concurrent or successive primary malignancies were excluded.
Patients, with extensive cancerous areas occupying the cervix, were also excluded. Samples were collected according to the 'sandwich' method ( Figure 1A) and obtained from 3 consecutive sites of suspicious lesions (C1/C2/C3) and normal-looking areas (N1/N2/N3).
Specimens at the middle site (C1 and N1) were immediately washed 3 times with ice-cold PBS solution and stored in liquid nitrogen. If the specimens at both ends (C2/C3 and N2/N3) were consistently confirmed by pathological examination as cervical invasive carcinoma and normal cervical tissue, respectively, the middle specimens (C1 and N1) were regarded as qualified and analysed. A total of 13 pairs of qualified specimens were analysed. The cancerous and adjacent normal tissue samples were named C1-13 and N1-13, respectively.
Tissue collection was performed under the approval of the Beijing Chao-yang Hospital Ethics Committee and under the patients' informed consent.

| Tissue microarray (TMA) construction
A TMA was constructed based on the formalin-fixed, paraffin- TMAs were constructed according to a method described in our previous studies 7, 8 . The slides were reviewed, and the pathological diagnosis of cervical SCC of all the included patients was confirmed by two independent gynaecological pathologists who were blinded F I G U R E 1 Proteomic profiles of CC and paracancerous tissue and the identification of DEPs. Sample was collected according to the 'sandwich' method and obtained from 3 consecutive sites of suspicious lesions (C1/C2/C3) and normal-looking areas (N1/N2/N3). If the specimens at both ends (C2/C3, N2/N3) were consistently confirmed by pathological examination as cervical invasive carcinoma and normal cervical tissue, respectively, the middle specimens (C1 and N1) were qualified ( Figure 1A). The interaction of HRAS, DUSP7, PLD1 and p-ERK1/2 is shown in Figure 1B. WB ( Figure 1C) and IHC ( Figure 1D) staining analyses consistently confirmed the results of the quantitative proteomic analysis that DEPs (HRAS, P-ERK1/2 and PLD1) levels were increased, whereas the DUSP7 level was decreased in CC tissue compared with the paired normal paracancerous tissue. A total of 102 patients' FFPE samples were included in the TMA ( Figure 1E; optical magnification*20). The IHC results from the CC TMA analysis showed that the decreased expression of DUSP7 and increased expression of PLD1 and p-ERK1/2 were adversely related to patients' relapse (p = 0.003, 0.040 and 0.001, respectively; Figure 1F) and survival (p = 0.034, 0.001 and 0.006, respectively). *p < 0.05, **p < 0.01 to the clinical data. Accurate locations of the tumours were marked on the FFPE samples. Two tissue cores, 1 mm in diameter, were taken from a donor block and placed in a recipient block (10 × 12 arrays) using a manual tissue array instrument. Sequential 4 µm-thick sections were cut from the FFPE TMA blocks and mounted on blank slides.  Real-time PCR and WB were used to validate the transfection efficiency of the candidate DEP at the protein and mRNA levels, respectively. Independent sample t test was used for statistical evaluation. Primers for real-time PCR are shown in Table S3.

| The effect of candidate DEPs on the biological function of SIHA
Based on the reports on the candidate DEPs in the literature, the following experimental procedures were performed on the candidate DEP-SIHA and NC-SIHA, as well as sh-candidate DEP-SIHA and shNC-SIHA cells. The Cell Counting Kit-8 (CCK-8) assay and FCM were used to detect cell proliferation and cell cycle distribution, respectively. A colony formation assay was used to assess colony formation. The migration ability was detected through wound healing and Transwell migration assays. Transwell invasion assays were used to assess the invasion ability. Independent sample t test was used for statistical evaluation. Immunofluorescence assays were used to detect the expression of E-cadherin and vimentin in the target cells.
The role of the candidate DEPs in epithelial-to-mesenchymal transition (EMT) was also evaluated in this manner.

| Ethics approval
Patient records and information were anonymized and deidentified prior to analysis; therefore, consent was not necessary. The study protocol was approved by the Ethics Committees at Beijing Chao-Yang Hospital.

| Proteomic profile of CC and paracancerous tissue and the identified DEPs
A total of 7811 proteins were identified in all samples, with less than a 1% false discovery rate (FDR) ( Table S4). According to the TMT ratios (≥1.5 or ≤0.5), 129 proteins were found to be differentially expressed in all 3 pairs of samples, with a p value <0.05. The gene function of 97 out of the 129 DEPs, including 89 upregulated and 8 downregulated DEPs, was associated with the pathogenesis and progression of tumours (Table 1).
To investigate the functions of the DEP-related genes in signal transduction in tumour cells, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed. Three DEPs, HRAS, DUSP7 and PLD1, are RAS pathway components. ERK, also a RAS pathway component, was not included in the DEPs. Total ERK is not necessarily highly expressed in human malignancies 9 ; however, persistent ERK1/2 activation through phosphorylation ultimately promotes cell proliferation and malignant transformation 10 .
The interaction of HRAS, DUSP7, PLD1 and phosphorylated ERK1/2 (p-ERK1/2) is illustrated in Figure 1B. These four genes were thus selected for further verification. WB and IHC staining analyses confirmed the results of the quantitative proteomic analysis that the levels of HRAS, PLD1 and p-ERK1/2 were increased, whereas the DUSP7 level was decreased in CC tissue compared with the paired normal paracancerous tissue ( Figure 1C,D).

| TMA construction and patients' clinicopathological information
During the study period, a total of 102 patients' FFPE samples were included in the TMA (Table 2, Figure 1E At the last contact, 10 patients died of the disease. The 5 years relapse-free survival rate was 86.9%, and the 5 years overall survival rate was 90.5%.

| The expression of DUSP7, PLD1 and p-Erk1/2 in CC TMA
The IHC results from the CC TMA analysis showed that the de-

| Correlation between DUSP7 expression and the Ras pathway
The expression of HRAS and p-ERK1/2 was decreased in DUSP7-SIHA cells compared with NC-SIHA cells (p = 0.0003 and 0.0026, respectively; Figure 3H). In contrast, the expression of HRAS and p-ERK1/2 was significantly upregulated in shDUSP7-SIHA cells compared with control cells (p = 0.034 and 0.0026, respectively).
The difference in the expression level of PLD1 between the two cell groups was not statistically significant (p = 0.0947 and 0.307, respectively).

| DISCUSS ION
Proteomics, as the leading technology in the postgenomic era, plays an important role in screening diagnostic and therapeutic markers for many human malignancies 11 . In this analysis, the 'sandwich' sampling method was adopted to ensure the accuracy of the histopathological results of the sampling. Cervical SCC and paired adjacent cervical tissues with the same genetic backgrounds and high comparability were used as the experimental and control groups, respectively. DEPs identified in this way could relatively objectively reflect the process of tumorigenesis and progression of CC. In this study, a total of 7811 proteins were identified through quantitative proteomics. The number of proteins identified in this study was large, and the quality of proteome detection was satisfactory.
In this study, 97 out of 129 DEPs were found to be related to tu-  Figure 2C). Three specific shRNAs targeting DUSP7 were designed and synthesized. The qRT-PCR results showed that the silencing efficiencies of these shRNAs were 55%, 54% and 47% (all p < 0.0001) when shNC was used as a reference. Cells infected with the most effective (55%) shRNA sequence (GCAUCAAGUAUAUCCUCAATT) were named shDUSP7-SIHA and used for subsequent experiments.
DUSP1 plays different roles in human tumorigenesis; specifically, it acts as a cancer-promoting factor in lung cancer and leukaemia 21,22 , and as a tumour suppressor in head and neck SCC, prostate cancer and urothelial bladder cancer 23,24 .
The DUSP7 gene is located on human chromosome 3p21 25 .
DUSP7 has a MAP kinase-binding domain/kinase-acting region that can specifically bind to p-ERK1/2, thus leading to its dephosphorylation. In this manner, DUSP7 can promote the meiosis of

| Strengths and Limitations
This study has several strengths compared with other contemporary studies. First, CC and paracancerous tissue samples used for proteomics testing were collected according to the 'sandwich' method and underwent relatively strict pathological examination. DEPs identified in this manner accurately reflect the tumorigenesis of CC. Moreover, The effect of DUSP7 on the biological function of SIHA cells. The CCK-8 assay growth curves showed that DUSP7-SIHA cells proliferated significantly slower than NC-SIHA cells, based on a clear delay in the doubling time (47.72 ± 1.14 h vs. 23.99 ± 0.47 h, p = 0.0001; Figure 3A). Cell cycle analysis indicated that the DUSP7-SIHA cells displayed a concomitant decrease in the percentage of cells in S phase (37.71 ± 0.53% vs. 46.96 ± 0.59%, p < 0.0001) and a significant increase in the percentage of cells in G0/G1 phase (52.50 ± 3.49% vs. 44.04 ± 0.71%, p = 0.0473; Figure 3B). Colony formation assays showed that the number of colonies formed by DUSP7-SIHA cells was significantly less than that formed by NC-SIHA cells (44.67 ± 9.0 vs. 75.33 ± 14.47, p = 0.0121; Figure 3C). In the Matrigel invasion/ migration assay, DUSP7-SIHA cells demonstrated a significantly weaker ability to invade (0.34 ± 0.05 vs.1, p = 0.0207; Figure 3D; optical magnification*20) and migrate (0.56 ± 0.14 vs.1, p = 0.0059) through the membrane than control cells. Wound-healing assays showed that the migration area of DUSP7-SIHA cells was significantly smaller than that of NC-SIHA cells (0.55 ± 0.03 vs.1, p = 0.049; Figure 3E). E-cadherin expression was significantly increased, but vimentin expression was significantly reduced in DUSP7-SIHA cells ( Figure 3F). In contrast, the results of CCK-8 assay growth curves indicated that the doubling time of shDUSP7-SIHA cells was significantly shorter than that of shNC-SIHA cells (49.12 ± 1.14 h vs. 64.14 ± 0.47 h, p = 0.0001; Figure 3A). Cell cycle analysis indicated that shDUSP7-SIHA cells displayed a concomitant increase in the percentage of cells in S phase (49.54 ± 1.53% vs. 46.4 ± 0.97%, p = 0.019) and a significant decrease in the percentage of cells in G0/G1 phase (41.8 ± 0.38% vs. 45.2 ± 0.80%, p = 0.020; Figure 3B). Colony formation assays showed that the number of colonies formed by shDUSP7-SIHA cells was significantly greater than that formed by shNC-SIHA cells (13.33 ± 3.4 vs. 30.33 ± 16.50, p = 0.049; Figure 3C). In the Matrigel invasion/migration assay, DUSP7-SIHA cells demonstrated a significantly weaker ability to invade (2.29 ± 0.38 vs. 1, p = 0.0007; Figure 3D) and migrate (1.87 ± 0.28 vs.1, p = 0.0426) through the membrane than control cells (p = 0.0207 and 0.0059, respectively; Figure 3D). Wound-healing assays showed that the migration area of shDUSP7-SIHA cells was larger than that of shNC-SIHA cells (1.95 ± 0.19 vs.1, p = 0.0313; Figure 3E). E-cadherin expression was significantly reduced, while vimentin expression was increased, in shDUSP7-SIHA cells ( Figure 3F; optical magnification*20). After subcutaneous injection, DUSP7-SIHA tumours were observed much later than NC-SIHA tumours (13 ± 4 vs 6 ± 1 days; p = 0.0122; Figure 3G). On the study end date, the DUSP7-SIHA tumours were significantly smaller than the NC-SIHA tumours (0.74 ± 0.38 vs 3.25 ± 1.68 cm 3 ; p = 0.0183). In contrast, the shDUSP7-SIHA tumours were observed much earlier than the shNC-SIHA tumours (7 ± 1 vs 12 ± 2 days; p = 0.0303). On the study end date, the shDUSP7-SIHA tumours were significantly larger than the shNC-SIHA tumours (3.66 ± 1.33 vs 0.75 ± 0.41 cm 3 ; p = 0.0201). The expression of HRAS and p-ERK1/2 was decreased in DUSP7-SIHA cells compared with NC-SIHA cells (p = 0.0003 and 0.0026, respectively; Figure 3H). In contrast, the expression of HRAS and p-ERK1/2 was significantly upregulated in shDUSP7-SIHA cells compared with control cells (p = 0.034 and 0.0026, respectively). The difference in the expression level of PLD1 between the two cell groups was not statistically significant (p = 0.0947 and 0.307, respectively) this study provides the first evidence that DUSP7 inhibits CC progression by inactivating the RAS pathway. However, some important limitations also exist in this study. First, IHC staining for DUSP7 in human CC tissues must be interpreted carefully due to the drawbacks of TMAs. TMAs including the limited tissues collected in the present study may not be representative of the whole tumour. In addition, due to the limitations of basic research, the effects of DUSP7 and its correlation with the RAS pathway should be verified in the future.

| CON CLUS IONS
DUSP7 is decreased in cervical cancer tissues compared with normal tissues. Increasing or decreasing DUSP7 expression was found to significantly reduce or enhance the anchorage-independent growth of SIHA cells, respectively. The biological function of DUSP7 is possibly achieved through dephosphorylation of ERK1/2 and inactivation of the RAS pathway. Upregulating the expression of DUSP7 may be useful for the prevention or treatment of CC.

ACK N OWLED G EM ENTS
This work is supported by Natural Science Foundation of China (No.81972424, 81702551).

CO N FLI C T O F I NTE R E S T S
The authors have no conflicts of interest to declare.

CO N S ENT FO R PU B LI C ATI O N
All the authors have reviewed the manuscript and the related files and consented to its publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets supporting the results of this article are included within the article and its additional files.