A total of 59 pairs of tumor and adjacent normal tissues were obtained from the Department of Urology, Union Hospital, Tongji Medical College (Wuhan, China) between January 2018 and January 2019. The normal kidney tissues were obtained from 2 cm away from the edge of lesions. The clinical information of these samples was presented in Table 1. The samples were divided into two groups. The first group was immediately stored in liquid nitrogen for RNA and protein extraction. The second group was fixed in formalin and embedded in paraffin, then used for immunohistochemistry assays. Among these samples, we randomly selected 8 pairs of tissues for protein detection, 20 pairs of samples for quantitative real-time PCR and 3 pairs of samples for immunohistochemistry. No patients received anticancer therapy before surgery. All patients gave written informed consent before inclusion in this study, and the study was approved by the Human Research Ethics Committee of Huazhong University of Science and Technology. The study complies with the guidelines of the Declaration of Helsinki.

The human renal proximal tubular epithelial cell line HK-2, and 5 kinds of human renal cell carcinoma cell lines: 786-O, ACHN, A498, CAKI-1, and OSRC-2, were used in this study and were obtained from the American Type Culture Collection. These cell lines were used for RT-qPCR and western blotting. The cells were grown in Dulbecco’s modified Eagles medium containing high glucose (4.5 g/L), fetal bovine serum (10%), and penicillin/streptomycin solution (1%). All cells were cultured under standard conditions: at 37°C in a 5% CO₂ atmosphere.

The paired RCC tissues and adjacent normal tissues were first fixed in 4% formalin at room temperature for 12 h, dehydrated, and embedded in paraffin. Then the tissue sections were incubated with rabbit TIMP1 monoclonal antibody (Abcam, ab109125, 1:1,000) overnight at 4°C. Tissue sections were washed three times with phosphate-buffered saline and incubated with secondary antibodies that were conjugated to horseradish peroxidase at room temperature for 2 h. The sections were scanned by a NanoZoomer S360 (Hamamastu Corporation) and observed with NDP.view2 software (Hamamastu Corporation). Random fields were selected to interpret the expression of TIMP1 in tissue sections under 100x and 400x magnification.

Total RNA was isolated from tissues or cells using TRIzol® reagent (Thermo Fisher Scientific, Inc.). The concentration and purity of the RNA solution were detected using a NanoDrop 2000 spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc.). Total RNA was then reverse transcribed into cDNA using a Superscript II reverse transcription kit (Takara Bio, Inc.) according to the manufacturer’s protocols. All the experiments were repeated thrice for all the samples. The primers used to amplify TIMP1 and GAPDH were synthesized by TSINGKE Inc. The sequences of forward and reverse primers were as follows: TIMP1-forward: 5′-CGC AGC GAG GAG GTT TCT CAT-3’; TIMP1-reverse: 5′-GGC AGT GAT GTG CAA ATT TCC-3’. GAPDH-forward: 5′-CGT GGA AGG ACT CAT GAC CA-3’; GAPDH -reverse: 5′-GCC ATC ACG CCA CAG TTT C-3’.

Total protein was extracted from RCC tissues and corresponding adjacent normal tissues of 12 patients using RIPA lysis buffer (Servicbio.Inc.) with protease inhibitor phenyl methane sulfonyl fluoride (PMSF, 1%), and then the concentration was measured with BCA protein assay kit (Beyotime Biotechnology, Jiangsu, China). Primary rabbit polyclonal antibody against primary antibodies (TIMP1 1:1,000, Abcam,Inc., ab109125; N-cadherin 1:5000, Abcam.Inc., ab76011; E-cadherin 1:10000, Abcam,Inc., ab40772) and β-actin (1:10000; Abclonal,Inc., cat.AC026) were incubated overnight at 4°C. All the procedures were performed according to the manufacturer’s instructions.

Small interfering RNA (siRNA) oligonucleotide sequences specifically targeting TIMP1 (si-TIMP1) and negative control (si-NC) siRNA (cat. no. siBDM 1999A) were obtained from Guangzhou RiboBio Co., Ltd. The plasmids harboring TIMP1 (ov-TIMP1) and negative control (ov-NC) were constructed and supplied by Vigene Biology (Vigene, China). Cells were collected for subsequent experiments 48 h post transfection. The si-TIMP1 sequence was as follows: 5′-GCC AAT GTG ATG GTG GAC A-3'. The information about the plasmid that has been used for over expression of TIMP1 was provided in Supplementary Figure S1.

One thousand transfected cells were added to each hole in the 96-well plates. Cell proliferation was assessed by the CCK-8 assay (CCK8; Dojindo Molceular Technologies, Inc.) at 24, 48, 72, and 96 h following treatments, according to the manufacturer’s instructions. Ten μl of CCK8 was added to each hole and incubated with cells for 2 h. Then the optical density values were measured by a spectrophotometer at 450 nm to estimate the number of living cells.

We planted 1 × 10⁴ cells into the upper chambers in serum-free medium for migration and 2 × 10⁴ cells for invasion. Sixty μl Matrigel (Thermo Fisher Scientific; Waltham, USA) had been added into the upper chambers for invasion assays. The lower chambers were filled with 600 μl DMEM added with 10% FBS. After 24 h of incubation, the cells were fixed in 100% methanol, then stained with 0.05% crystal violet. Finally, the results were observed under a light microscope at 100x magnification, and the cells passed through the membrane were counted in 3 randomly chosen fields.

TIMP1, TIMP2, TIMP3, and TIMP4 mRNA expression and clinical information of The Cancer Genome Atlas (TCGA) clear cell renal cell carcinoma dataset (TCGA_KIRC) were downloaded from the Xena Functional Genomics Explorer of University of California Santa Cruz (https://xenabrowser.net/). Beroukhim renal, Jones renal, and Yusenko renal datasets (Jones et al., 2005; Beroukhim et al., 2009; Yusenko et al., 2009) were obtained from the Oncomine database (https://www.oncomine.org). The gene set enrichment analysis (GSEA) platform with the Kyoto Encylopedia of Genes and Genomes and Gene Ontology databases (c2.all.v6.2.symbols.gmt) was employed to find pathways enriched in the gene set, based on the pathway Enrichment Score (ES). STRING (Version11.0) was used to explore the protein-protein reaction and biological function of TIMP1(https://string-db.org/).

All statistical analyses were processed by GraphPad Prism 7.0 (GraphPad Software, Inc., USA) and SPSS Statistics 22.0 software (IBM SPSS, Chicago, IL, United States). Data of paired cases were analyzed using a paired student t-test, while analysis of unpaired cases was performed using a one-way analysis of variance (ANOVA) or t-test. Pearson’s χ2 test was applied to analyze the relationship between TIMP1 and TIMP3 expression and clinical parameters. The Kaplan-Meier analysis was used to estimate the correlation between TIMP1 and TIMP3 mRNA expression with overall survival (OS) and disease-free survival (DFS) times with the log-rank test. The TIMP1 mRNA levels downloaded from the TCGA_KIRC dataset were first divided into two groups according to different clinical parameters and then applied to draw receiver operating characteristic (ROC) curves and analyze the area under the curve (AUC) with GraphPad Prism 7.0. The diagnostic value of TIMP1 mRNA expression in RCC was evaluated by ROC curves and AUC. Finally, univariate and multivariate Cox proportional hazard regressions were applied to determine the prognostic significance of TIMP1 and TIMP3. All experiments were repeated thrice independently and all data were represented as mean ± SEM. A confidence threshold, p < 0.05, was considered to be statistically significant. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

To determine whether TIMP expression was related to the occurrence and progress of RCC, we downloaded the mRNA of 4 members in this family (TIMP1, TIMP2, TIMP3, and TIMP4) from TCGA_KIRC and drew a heatmap of them according to their mRNA expression. The result showed that TIMP1 was upregulated in RCC tissues (Figure 1). It was further verified that TIMP1 was found in higher expression levels in paired comparison (Figure 2A). To confirm these results, we compared the mRNA expression of TIMP1 in Oncomine datasets (Beroukhim, Jones, and Yusenko) and it was shown that TIMP1 was highly expressed in renal cancer samples (Figure 2B). All these results indicated that TIMP1 might play an important role in RCC progression, which raised our interest to further study.

Kaplan-Meier survival analysis and log-rank test were applied to determine the OS and DFS in patients with RCC. The result showed that patients with higher levels of TIMP1 had a poorer outcome (Figure 3A). Further Kaplan-Meier survival analysis for subgroups of patients with different clinical parameters demonstrated that TIMP1 was an ideal prognostic biomarker for patients with the following characteristics: T1+T2 stage (Figure 3B), N0 (Figure 3C), M0 (Figure 3D), G1+G2 stage (Figure 3E), age<60 (Figure 3F), age≥60 (Figure 3G), male (Figure 3H) and female (Figure 3I). The DFS survival analysis revealed that patients with higher TIMP1 had a shorter disease-free time (Figure 4A), and TIMP1 could act as a biomarker for patients with the following characteristics: T1+T2 (Figure 4B), N0 (Figure 4C), age≥60 (Figure 4D), and female (Figure 4E). Furthermore, we applied univariate and multivariate regression models to assess the integrated prognostic value of TIMP1. The results suggested that TIMP1 independently correlated with the OS and DFS status of RCC patients (Tables 2, 3). In conclusion, the above results demonstrated that the high level of TIMP1 indicated a poor clinical outcome and TIMP1 could serve as an ideal prognostic biomarker for RCC.

To clarify the expression pattern of TIMP1 in patients with different clinical parameters, the present study analyzed the TIMP1 expression levels of 522 cases from the TCGA database. The results confirmed that the high expression level of TIMP1 is associated with patients’ gender, higher T stage, N stage, M stage, TNM stage, and histological grade (Figures 5A–F). Besides, patients with worse OS status and DFS status had a higher level of TIMP1 expression (Figures 5G,H). However, there were no obvious differences between patients aged ≥60 years and those aged <60 years (Table 4). These results demonstrated that TIMP1 was upregulated and closely related to gender, T stage, N stage, M stage, TNM stage, and histological grade in RCC.

To access the value of TIMP1 mRNA expression level in the diagnosis of RCC, receiver operating characteristic (ROC) curves were applied for patients with different clinicopathological variables. Area under curve (AUC) was used to evaluate the diagnostic efficiency. The results indicated that TIMP1 could adequately distinguish RCC patients with an AUC of 0.8858 (p < 0.0001; Figure 6A). Additionally, the TIMP1 expression level also exhibited diagnostic value for subgroups of patients with RCC as follows: T1+T2 vs. T3+T4 (AUC = 0.6414, p < 0.0001; Figure 6B), G1+G2 vs. G3+G4 (AUC = 0.6307, p < 0.0001; Figure 6C), M0 vs. M1 stage (AUC = 0.6375, p = 0.0001; Figure 6D), TNM I + II vs TNM III + IV stage (AUC = 0.6546, p < 0.0001; Figure 6E), OS good vs. OS poor (AUC = 0.6617, p < 0.0001; Figure 6F), DFS good vs DFS poor (AUC = 0.7056, p < 0.0001; Figure 6G) and male vs. female (AUC = 0.5875, p = 0.001; Figure 6H). Therefore, TIMP1 might act as a potential biomarker for RCC diagnosis.

To further confirm the results from bioinformatics analysis, we performed a quantitative real-time polymerase chain reaction (qRT-PCR) analysis and western blot in RCC cell lines and tissues. We found a significantly higher level of TIMP1 mRNA in RCC cells (786-O, ACHN, A498, CAKI-1, OSRC-2) relative to HK-2 (Figure 7A). Western blot revealed that the expression level of TIMP1 in RCC was higher compared with HK-2 (Figure 7B). Then we studied the expression of TIMP1 mRNA and protein in tissues. We detected 20 pairs of RCC tissues and corresponding adjacent normal tissues for qRT-PCR and found TIMP1 over-expression in RCC in 17 pairs of tissues (Figure 7C). Western blot for tissues showed a similar result. TIMP1 expression was significantly higher in RCC compared with adjacent normal tissues (Figure 7D). Furthermore, IHC was conducted in 3 pairs of RCC tissues and adjacent normal tissues. TIMP1 was primarily located in the membranes and cytoplasm of cancer cells and renal tubular epithelial cells (Figure 7E). The results proved that the expression of TIMP1 was higher in RCC tissues again. These findings confirmed that TIMP1 was upregulated in RCC cells and tissues.

In order to elucidate the function of TIMP1 in RCC, TIMP1 was knocked down by TIMP1-siRNA and was overexpressed by transfecting plasmid into ACHN and 786-O cell lines. Then we used qRT-PCR and western blot to test the efficiency of transfection. The results showed that TIMP1-siRNA and plasmid carrying TIMP1 could raise corresponding effects on TIMP1 expression in cells (Figures 8A,B). CCK8 assays revealed that TIMP1 knockdown could significantly repress the proliferation rates of ACHN and 786-O, while overexpression of TIMP1 accelerated cell proliferation (Figures 8C,D). We applied transwell assays to detect the influence of TIMP1 on RCC cells’ migration and invasion ability. The results showed that TIMP1 knockdown suppressed the migration and invasion ability of ACHN and 786-O, while TIMP1 overexpression promoted these features (Figures 9E,F). In conclusion, TIMP1 facilitated the progression of RCC by promoting the proliferation, migration, and invasion of RCC cells.

To clarify the specific function of TIMP1 in RCC, GSEA based on the TCGA database was performed. Besides, we retrieved the biological process of TIMP1 in STRING (https://string-db.org/). The results of GSEA have shown that TIMP1 was closely related to tumorigenesis, metastasis, cell cycles, and epithelial-to-mesenchymal transition (Figures 9A–F). The protein-protein interaction network calculated by STRING displayed the proteins interacting with TIMP1 (Figure 9G). The major GO term included signaling transduction and cell migration (Figure 9H; Table 5). The KEGG pathways TIMP1 participated in involved PI3K-Akt and JAK-STAT signaling pathways (Figure 9I; Table 6). These results indicated that TIMP1 might play a key role in the metastasis of RCC and participate in extracellular signal transduction. According to the information above, we examined the EMT pathway which was proved to be critical in tumor metastasis by silencing TIMP1. The results revealed that TIMP1 silencing led to the downregulation of N-cadherin and upregulation of E-cadherin (Figure 9J). These findings illustrated that TIMP1 facilitated the progression of RCC via EMT transition.