High Cytoplasmic and Membranous Versus Nuclear Talin-1 Expression Associated With Malignancy Potential and Poor Survival in Clear Cell Renal Cell Carcinoma

Using bioinformatics analysis Talin-1 protein was indicated as a potential prognostic focal adhesion factor in renal cell carcinoma (RCC). We, therefore, examined its protein expression levels and prognostic signicant of Talin-1 with clinical follow-up in total of 269 tissue specimens from three important subtypes of RCC and 30 adjacent normal samples using tissue microarrays (TMAs). Then, we used combined analysis with B7-H3 to conrm that Talin-1 may be related to metastasis. The results showed that high membranous and cytoplasmic expression of Talin-1 was signicantly associated with advanced nucleolar grade (P < 0.001, all), microvascular invasion (P = 0.007, P = 0.004, respectively), histological tumor necrosis (P = 0.020, P = 0.018), and invasion to Gerota’s fascia (P = 0.025, P = 0.050) in ccRCC. In addition, high membranous and cytoplasmic expression of Talin-1 was associated with signicantly poorer disease-specic survival (P = 0.043, P = 0.024, respectively) and progression- free survival (P = 0.046, in cytoplasm). Moreover, increased cytoplasmic expression of Talin-1 High /B7-H3 High compared to the other phenotypes was associated with poor prognosis and progression of the disease in patients with distant metastasis. Collectively, these observations indicate that Talin-1 is an important molecule involved in the spread and progression of ccRCC when expressed particularly in cytoplasm and may serve as a novel prognostic biomarker in this subtype. (7.43%) cytoplasmic DSS or PFS low Talin-1 level and 5-year DSS PFS time, we showed that membranous, cytoplasmic, and nuclear Talin-1 protein expression in other important subtypes of RCC. We found no signicant association between Talin-1 protein and clinicopathological features and patients' outcomes in pRCC cases. In chRCCs, a signicant association was observed between membranous Talin-1 protein expression and increased tumor stage. More investigations need to be performed to support these ndings using a larger number of cases of each pRCC and chRCC subtypes. cytoplasmic cytoplasmic of association and correlation between Talin-1 protein expression and clinicopathological parameters, Pearson's chi-square and Spearman’s correlation tests were performed. Kruskal-Wallis and Mann-Whitney U tests were applied for pairwise comparison between groups. Survival curves by Talin-1 status were estimated using the Kaplan–Meier method with 95%


Introduction
Renal cell carcinoma (RCC) is the most common type of kidney cancer found in adults. It represents approximately 3% of adult malignancies and 90-95% of all neoplasms arising from the kidney 1,2 . The RCC incidence rates have been increasing during the decade 3 . According to the American Cancer Society, an estimated 73,750 new cases (45,520 in males and 28,230 in females) and 14,830 deaths (9,860 in males and 4,970 in females) from RCC will occur in 2020 4 . If cancer has not metastasized, the 5-year survival rate is 65-90%, but this is lowered considerably when cancer has spread so that the 5-year survival rates for patients with stage IV disease are less than 20% 5 . Unfortunately, ~ 20-30% of RCC patients have distant metastasis at initial diagnosis. In addition, about 25-40% of RCCs would relapse and ultimately develop into metastatic disease after resection of the primary tumor 6 . Moreover, increasing resistance to targeted therapies have been occurred in advanced RCC patients 7 . Therefore, prognostic and predictive biomarkers will be an urgent requirement to assist in early diagnosis and identify novel therapeutic targets for RCC.
To date, several subtypes of RCC have been de ned, of which clear cell RCC (ccRCC) is the most common subtype (75-80%), followed by papillary RCC (pRCC) 15%, and chromophobe RCC (chRCC) around 5% 8 . In this study, we have focused on all these three important histological subtypes of RCC.
High throughput proteomics was introduced as an innovative novel approach that paves the way for cancer diagnostic or prognostic biomarker discovery 9 . The advantages of these types of studies can be the identi cation of effective proteins with a different expression and cellular functions in the samples in various diseases 10 . Although a high level of variance exists in proteomics data, network, and enrichment analysis explore the hub genes and their related biological signi cance among results 11 . In the present study, protein-protein interaction (PPI) network analysis was performed for proteomics data of available studies that had reported proteins with signi cant differential expression in RCC tissue samples compared to adjacent normal tissues. Enrichment analysis of hub genes was obtained for a better understanding of in uential genes in the RCC network. It was found that TLN1 (Talin-1) is one of the hub genes that could be practical as a prognostic marker in renal cancer.
Talin-1 is a ubiquitous cytoskeleton-associated protein with a high-molecular-weight which has been shown to play an essential role in regulating the activity of the integrin family of cell adhesion proteins using coupling integrins to F-actin 12,13 .
Also, Talin-1 is responsible for binding to critical adhesion molecules, including integrins, vinculin, actin, and other focal adhesion kinases (FAKs) and induce cell cytoskeletal remodeling. Therefore increased expression of Talin-1 could lead to the development of tumor and migration 14,15 . A number of studies have demonstrated that overexpression of Talin-1 was associated with advanced disease and reduced overall survival (OS) in patients, including oral squamous cell carcinoma (OSCC) 16 , prostate cancer 17 , colon cancer 18 , and nasopharyngeal carcinoma (NPC) 19 . To date, the expression levels and clinical signi cance of Talin-1 in the RCC tumors remain unclear. Thus, in the current study, to validate and con rm the Talin-1 protein obtained from the network analysis on results of proteomics studies in RCC samples, we have investigated the membranous, cytoplasmic, and nuclear expression levels of Talin-1 protein in three important subtypes of RCC from formalinxed para n-embedded (FFPE) tissue samples which assembled on tissue microarrays (TMAs) using the immunohistochemistry (IHC) technique. Then, we sought to determine the association between expression levels of Talin-1 protein and clinicopathological parameters, including nucleolar grade, tumor stage, microvascular invasion, and other important clinicopathological characteristics of RCC as well as survival outcomes. Finally, we used combined analysis with another marker of our previous study, namely B7-H3 20 which is useful for predicting progression and metastasis in ccRCC to con rm that Talin-1 may be related to metastasis.

Bioinformatics approach
Signi cant differential expression of proteins from tumor tissues in comparison to adjacent normal tissues was obtained of selected previous studies that include 1055, 589, and 983 proteins for ccRCC, pRCC, and chRCC, respectively (Supplementary Table 1). The results of the Venn diagram analysis illustrated 110 common differential expression proteins between subtypes of RCC ( Fig. 1). PPI network analysis result explored 52 hub genes with the highest con dence > 0. 9 Table 1). Pathway analysis displays that most key genes contribute to metabolism pathways and several genes involved in FAK and degradation of extracellular matrix (ECM) pathways (Fig. 3). The major part of GO analysis results for 24 key genes indicates these genes are present in the microbody and play role in fatty acid catabolic process. Also, the basement membrane organization is a remarkable part of gene annotation results ( Supplementary Fig. 1). According to these results, key genes involved in FAK and pathways related to ECM, as key drivers for both development and cancer progression 21 , were highlighted for selection marker. Our key genes list, including COL4A1, LAMB2, NID1, and TLN1 (Talin-1) were evaluated among literature and led to the selection of Talin-1 as our nal biomarker for investigation in RCC. Talin-1 as a master integrin-associated protein is a focal adhesion (FK) complex that mediates ECM proteolysis through different matrix metallopeptidases (MMPs) expression 22 . FAK activation triggers the following processes, including survival, angiogenesis, cell growth, migration, and invasion. Since FAK functions potentially regulate gene expression to affect cancer progression 23 , it seems that its signaling to be vital in the neoplastic cells invasion, and its activity inhibition decrease RCC metastasis both in vitro and in vivo [24][25][26] . Although the FAK signaling pathway is deregulated in different cancer types and its importance approved in the regulation of cell adhesion and motility due to the tumor invasive behavior [27][28][29] , its evidence in RCC is few and it is necessary to explore which is crucial 22 . Further, UALCAN database revealed a higher expression of Talin-1 protein in ccRCC tissue samples compared to normal samples ( Fig. 4).

Characteristics of RCC patients
Two hundred and ninety-ve RCC patients were included in this study, while technical problems led to a loss of some cases, and nally, 269 cases remained for evaluating. These samples include 195 (73.58%) ccRCC, 20 (7.43%) pRCC type I, 20 (7.43%) pRCC type II, and 34 (12.63%) chRCCs. The patients' clinicopathological characteristics were summarized in Table 1 based on whole RCC samples and histological subtypes of RCC.    cytoplasmic Talin-1 protein expression levels in different RCC subtypes (P < 0.001, P = 0.003, respectively). Also, Mann-Whitney U test showed a statistically signi cant difference in the median levels of membranous and cytoplasmic Talin-1 protein expression between ccRCC and pRCC (type I & II) (P < 0.001, P = 0.001, respectively) and also pRCC and chRCC (P = 0.006, P = 0.029, respectively) ( Fig. 6A, B). In contrast, the Kruskal-Wallis test showed that there is no statistically signi cant difference between the median expression level of nuclear Talin-1 protein in different subtypes of RCC (P = 0.904) ( Table 3).
Moreover, we did not nd a statistically signi cant difference in the median expression level of nuclear Talin-1 protein between the RCC subtypes (Fig. 6C). The results of Pearson's chi-square test showed a highly signi cant association between expression of membranous and cytoplasmic Talin-1 protein and increased nucleolar grades (I + II versus III + IV) (Pearson's chi-square, P < 0.001, all) ( Table 4). Kruskal-Wallis and Mann-Whitney U tests also showed that statistically signi cant differences in the median levels of membranous and cytoplasmic Talin-1 protein expression in various nucleolar grades (P = 0.001, P < 0.001, respectively) ( Fig. 7A, B). Moreover, Pearson's chi-square test revealed that signi cant differences between higher levels of expression of membranous and cytoplasmic Talin-1 protein and MVI (P = 0.007, P = 0.004, respectively), histological TN (P = 0.020, P = 0.018, respectively) as well as Gerota's fascia invasion (P = 0.025, P = 0.050, respectively). Furthermore, there was a statistically signi cant difference between higher level of membranous Talin-1 protein expression and invasion to renal pelvis (P = 0.038). Kruskal-Wallis and Mann-Whitney U tests also indicated a statistically signi cant difference between the median expression levels of membranous and cytoplasmic Talin-1 protein and MVI (P = 0.022, P = 0.007, respectively) ( Fig. 7C, D). Bivariate analysis (Spearman's correlation) showed a signi cant direct correlation between membranous and cytoplasmic expression and advanced in nucleolar grades (Spearman's rho, P < 0.001, P = 0.001, respectively), MVI (P = 0.007, P = 0.004, respectively), histological TN (P = 0.015, P = 0.014, respectively), and invasion to Gerota's fascia (P = 0.025, P = 0.050, respectively). The results of Spearman's correlation analysis also revealed that membranous Talin-1 protein expression and renal pelvis (P = 0.038) were directly correlated. In this study, we did not nd any signi cant association and correlation between nuclear Talin-1 protein expression and clinicopathological features in ccRCC cases (Table 4).  Tables 5, 6). In chRCC, we observed a statistically signi cant association between membranous expression of Talin-1 and tumor stage (P = 0.050) ( Table 7). Moreover, Spearman's correlation analysis showed a statistically signi cant positive correlation between increased Talin-1 protein expression and higher tumor stage (Spearman's rho, P = 0.023).       (Table 8). Survival outcomes based on membranous Talin-1 protein expression The results of the Kaplan-Meier curve demonstrated that signi cant differences between DSS and the patients with high and low membranous expression of Talin-1 protein (Log-rank test: P = 0.043) (Fig. 8A). The mean DSS time for patients with high and low membranous expression of Talin-1 protein was 91 (SD = 4.6) and 103 (SD = 3.5) months, respectively. Additionally, the 5-year survival rates for DSS in patients whose specimens expressed high and low membranous expression of Talin-1 protein were 78 and 88%, respectively (Log-rank test: P = 0.050). Moreover, the results of the Kaplan-Meier analysis for PFS showed that there are no signi cant differences between PFS and the patients with high and low expression of membranous Talin-1 protein (Log-rank test: P = 0.081) (Fig. 8B).

Survival outcomes based on cytoplasmic Talin-1 protein expression
Our nding of Kaplan-Meier curves analysis in the cytoplasmic expression of Talin-1 revealed that signi cant differences between DSS (Log-rank test: P = 0.024) or PFS (Log-rank test: P = 0.046) and the patients with high and low cytoplasmic Talin-1 protein expression, respectively (Fig. 8C, D). Further, the mean DSS or PFS time for patients with high and low cytoplasmic expression of Talin-1 protein were 90 (SD = 4.7), 87 (SD = 5.0) and 103 (SD = 3.4), 99 (SD = 3.9) months, respectively. The 5-year survival rates for DSS or PFS of patients with high cytoplasmic Talin-1 protein expression were 77 and 72% and with low expression were 89 and 85% (Log-rank test: P = 0.044, P = 0.050, respectively).

Survival outcomes based on nuclear Talin-1 protein expression
Kaplan Meier survival curves showed that there are no signi cant differences between DSS or PFS and the patients with high and low nuclear expression of Talin-1 protein in ccRCC cases (Log-rank test: P = 0.160, P = 0.219, respectively) (Fig. 9A, B).
The results from univariate and multivariate analysis were summarized in Tables 9, 10. As demonstrated in Tables 9 and 10, nucleolar grade was the only independent prognostic factor for DSS or PFS in the multivariate analysis. Expression of membranous and cytoplasmic Talin-1 for DSS and cytoplasmic expression of Talin-1 for PFS were not signi cant risk factors for prognosis in the multivariate analysis.

Combined analysis of Talin-1/B7-H3 populations
In this part of the study, we rst chose the samples which have stained in two markers. The results showed that a statistically signi cant correlation between expression of cytoplasmic Talin-1 and cytoplasmic B7-H3 proteins (Spearman's rho, 0.318; P < 0.001), however, in membranous expression, we did not nd any signi cant correlation between two markers (Spearman's rho, 0.026; P = 0.756). Therefore, we examined the association between co-expression of cytoplasmic Talin-1 and cytoplasmic B7-H3 proteins with clinicopathologic parameters. The expression levels of Talin-1 and B7-H3 were divided into two categories based on median expression and four phenotypes including Talin-1 High /B7-H3 High , Talin-1 High /B7-H3 Low , Talin-1 Low /B7-H3 High , and Talin-1 Low /B7-H3 Low (Table 11). The analysis showed that a highly signi cant association between cytoplasmic Talin-1 High /B7-H3 High phenotype and nucleolar grade (P = 0.002). The cytoplasmic Talin-1 High /B7-H3 High phenotype was more often found in patients with MVI present (P = 0.002), histological TN present (P = 0.026), Gerota's fascia invasion (P = 0.043), and distant metastasis (P = 0.022) compared to patients without MVI and histological TN, as well as no invasion to Gerota's fascia and no distant metastasis (Table 11).  The survival rate of patients with co-expression of cytoplasmic Talin-1 High /B7-H3 High phenotype was compared with other phenotypes. Higher cytoplasmic expression of Talin-1 High /B7-H3 High phenotype was associated with worse DSS and PFS than another phenotype (Log-rank test; P = 0.007, P = 0.021, respectively) (Fig. 10A, B). Further, the 5-year survival rates for DSS and PFS were 87.0% and 83.0% in cytoplasmic Talin-1 High /B7-H3 High phenotype expression, and 67.0% and 65.0% in another phenotype of cytoplasmic Talin-1/B7-H3 expression (Log-rank test; P = 0.010, P = 0.029, respectively). Moreover, the strati ed analysis indicated that in the patients with distant metastasis, the differences in DSS between patients with cytoplasmic Talin-1 High /B7-H3 High phenotype and other phenotypes of Talin-1/B7-H3 expression is signi cant (P = 0.050) (Fig. 10C, D). Tumor size and nucleolar grade were the signi cant risk factors affecting the DSS or PFS of patients with ccRCC in univariate analysis.

Discussion
Although much progress has been made big changes in the treatment of RCC, drug resistance to targeted therapies brings more failures in advanced RCC patients' treatment 7 . Thus, it is urgently needed to explore novel and practical clinical prognostic markers of RCC in the patient's early diagnosis in future renal cancer therapy.
The bioinformatics analysis would be conducive to molecular markers research in clinical RCC. This study set out with the aim to identify the novel biomarker that was potentially involved in the various RCC subtypes. In this regard, we focused on PPI network analysis for proteomics results of other studies related to RCC subtypes and used bioinformatics software investigations to identify hub genes. The hub genes were generated based on the highest degree of connectivity and were screened on Enricher for cancer disease. Besides, enrichment analysis determined the involved pathways and molecular function of hub genes. These results suggested several proteins as hub genes related to cancer, which among them we selected Talin-1 and warranted further investigation. Therefore, to validate the Talin-1 protein as a prognostic marker for RCC, for the rst time, expression levels of Talin-1 protein was investigated in a well-characterized series of 269 tissues specimen from patients treated with radical nephrectomy in three main subtypes of RCC including 195 (73.58%) ccRCC, 20 (7.43%) pRCC type I, 20 (7.43%) pRCC type II, and 34 (12.63%) chRCC tissue samples through IHC on TMA slides and evaluation of the association between its expression and clinicopathological characteristics as well as patient survival outcomes. The pattern of Talin-1 protein expression in RCCs also was classi ed into membranous, cytoplasmic, and nuclear expression.
It is important to investigate histological subtypes of speci c cancer, giving consideration to various subtypes which may be associated with different biologic behavior, pattern, and prognosis, hence each of them needs different treatment. The evaluation of the staining pattern in each subtype of RCC displayed differential expression of Talin-1 protein in membrane, cytoplasm, and nucleus with a range of intensities from weak to strong as well as the percentage of staining area. Further, there was a statistically signi cant association between the membranous and cytoplasmic expression of Talin-1 protein but not in nuclear expression and the RCC subtypes. Subsequently, a statistically signi cant difference was observed in the median levels of membranous and cytoplasmic Talin-1 protein expression between ccRCC and pRCC (type I & II) as well as pRCC and chRCC subtype, indicating that expression of Talin-1 protein in membrane and cytoplasm versus nucleus are more important in subtypes of RCC and also these expression patterns vary among the histological RCC subtypes, thus this may affect on the prognostic value of them and treatment options.
IHC analysis of human RCCs compared to adjacent normal tissue samples demonstrated that Talin-1 protein expression is upregulated in RCCs. The present study also considered the expression of Talin-1 on UALCAN database, containing a large amount of proteomics data which showed that increased expression of Talin-1 in ccRCCs rather than normal samples. Thus, these results are in line with our ndings of Talin-1 protein expression using IHC on ccRCCs.
The advantage of analyzing the expression and subcellular distribution of Talin-1 in RCCs is, the protein in different subcellular localization may have a speci c function. Our nding showed that membranous and cytoplasmic expression of Talin-1 protein is positively associated with the important clinicopathological parameters, including advanced nucleolar grade, MVI, histological TN, invasion to Gerota's fascia, and renal pelvis, while in nuclear expression, there was no any association between expression of Talin-1 protein and clinicopathological features in ccRCC cases. Importantly, our results exhibited the increased expression of the median membranous and cytoplasmic of Talin-1 protein in the high grade of RCC compared with the low grade of RCC and also increased expression in patients with MVI present rather than MVI absent which exhibited the association of membranous and cytoplasmic Talin-1 protein expression with the aggressiveness of ccRCC. In addition, in this study nucleolar grade was found as an independent prognostic factor for DSS and PFS in membranous and cytoplasmic Talin-1 protein expression in multivariate analysis. After the tumor stage, one the strongest prognosticators of survival in patients with RCC is the nucleolar grade, and tumors with a high nucleolar grade display a more aggressive phenotype; thus they are associated with local invasion and distant metastasis 30 . Previous studies demonstrated that MVI is related to cancer progression and survival in RCC and is the most signi cant prognostic variable for prognostication of metastatic spread and survival outcome independent from macrovascular invasion 31,32 . Histological TN also has proposed to be a sign of tumor aggressiveness that generally leads to poor survival outcomes 33 . A recently systematic review and meta-analysis study suggested that histological TN is associated with DSS, OS, RFS (recurrence-free survival), and PFS of RCC patients and may serve as a predictor of poor prognosis in RCC patients 34 . Moreover, in this study tumor size and tumor stage were found as prognostic variables in univariate analysis that depicted the associations between these parameters and more aggressive tumor behaviors. The tumor stage is one of the most factors for predicting tumor progression and recurrence in RCC 35 and tumor size has shown signi cantly associated with the risk of metastasis 36 . Therefore, these results indicated that membranous and cytoplasmic Talin-1 protein expression compared with nuclear expression are related to the degree of malignancy and progression of disease in ccRCC cases.
Our results from Kaplan-Meier survival curves in ccRCC cases showed that higher membranous and cytoplasmic Talin-1 protein expression is associated with signi cantly worsened DSS as well as worsened DSS or PFS compared to low Talin-1 protein expression, respectively. In addition, ccRCC patients, who expressed a high level of membranous and cytoplasmic Talin-1, had shorter 5-year DSS or PFS compared with those with low expression. However, the pattern of Talin-1 protein expression was not a predictor of survival in multivariate analysis, indicating the long term follow-up is needed to increase the events and prognosis.
Investigations have shown that increased expression of Talin-1 could lead to the progression of tumor cell adhesion and migration because of Talin-1 as a FA protein is able to mediate integrin interaction with ECM and can bind to a variety of adhesion molecules and induce cell cytoskeleton remodeling 37 . B7-H3, also a member of the B7 family of immunoregulatory proteins which is overexpressed in many types of malignancies and is linked to poor prognosis, increased tumor grade, and metastasis 38 . Previous studies have been reported a correlation between expression levels of B7-H3 and progression of disease in ccRCC 20,39 . To con rm that whether Talin-1 is related to metastasis, we used the combined analysis of both Talin-1 protein and our previous data of B7-H3 protein in ccRCC. The co-expression of cytoplasmic Talin-1 High /B7-H3 High was observed to be associated with more aggressive tumor behavior and metastasis. Further, Tumor size and nucleolar grade were the signi cant risk factors affecting the DSS or PFS which depict tumor aggressiveness. Moreover, increased cytoplasmic expression of Talin-1 High /B7-H3 High compared to the other phenotypes was associated with poor prognosis and progression of the disease, especially in patients with distant metastasis, therefore, our ndings revealed that increased cytoplasmic expression of Talin-1 is associated with invasiveness and metastasis in ccRCC.
Several previous studies revealed that overexpression of Talin-1 is associated with advanced disease and poor prognosis. Lai et al con rmed that increased expression of Talin-1 in OSCC is associated with poor prognosis and knockdown of Talin-1 expression inhibited ability of proliferation and invasion 16 . In a recent study by Ji Ling et al. using IHC analysis on colorectal cancer (CRC) and adjacent normal tissue demonstrated that Talin-1 protein expression is upregulated in CRC and knockdown of Talin-1 reduces the proliferation and migration via the epithelial-mesenchymal transition (EMT) signaling pathway 40 . Xu N et al. showed that the upregulation of Talin-1 protein in prostate cancer is signi cantly associated with advanced pathological parameters and predicts lymph node metastases and biochemical recurrence 17 . In NPC, authors exhibited that high Talin-1 expression is associated with signi cantly poorer OS 19 . Our present data further is in line with these previous studies and suggest that Talin-1 protein is an important molecule involved in the spread and progression of ccRCC. However, Talin-1 has a different expression in other cancers such as hepatocellular carcinoma (HCC). A previous study by Kanamori et al. 13 showed that high-level expression of Talin-1 in HCC tissues while another study demonstrated that Talin-1 is downregulated in HCC 41 , while others showed that expression levels of Talin-1 in serum is signi cantly higher 42 . In addition, Talin was found to be completely absent in endometriosis and endometrioid carcinomas 43 . Therefore, it seems that expression levels of Talin-1 are different in various cancers and more investigations are needed to explore the exact mechanism and function of Talin-1 for nding the new ways for therapeutic targeting.
In the present study, for the rst time, we showed that membranous, cytoplasmic, and nuclear Talin-1 protein expression in other important subtypes of RCC. We found no signi cant association between Talin-1 protein and clinicopathological features and patients' outcomes in pRCC cases. In chRCCs, a signi cant association was observed between membranous Talin-1 protein expression and increased tumor stage. More investigations need to be performed to support these ndings using a larger number of cases of each pRCC and chRCC subtypes.

Conclusion
In summary, bioinformatics analysis indicated Talin-1 protein as a potential focal adhesion prognostic marker in RCC. Our nding also demonstrated that Talin-1 protein is upregulated in human RCC tissues compared to adjacent normal tissues. Further, our results for the rst time revealed that the statistically signi cant differences between membranous and cytoplasmic Talin-1 expression in various histological subtypes of RCC. A better understanding of the histological subtypes in RCC is urgently needed for further development of treatment options. Moreover, the current study indicates that membranous and cytoplasmic Talin-1 protein expression, particularly cytoplasmic expression compared to nuclear expression plays an important role in more aggressive tumor behaviors, metastasis, and progression of ccRCC. In addition, high membranous and cytoplasmic expression of Talin-1 protein was identi ed as a worse prognostic variable affecting DSS or PFS in univariate analysis, indicating that Talin-1 may serve as a novel prognostic biomarker in ccRCC if follow up time more extended.
However, further studies on the function and mechanism of action of Talin-1 is a requirement to provide new opportunities for therapeutic targeting of RCC.

Data source and network analysis
Search on PRoteomics IDEnti cations database (PRIDE) (https://www.ebi.ac.uk/pride/) 44 form 2019 to 2020 led to identifying three RCC studies that each of them included the proteomics information of one of the RCC subtypes and adjacent normal tissues. The results of proteomics analysis were downloaded from the supplementary information of their published articles 37,45,46 . Venn diagram analysis was performed to nd common signi cant differential proteins in three subtypes of RCC by OriginPro 2019 (https://www.originlab.com/2019, OriginLab Corporation, Northampton, MA, USA). PPI network was constructed using srtringapp (https://string-db.org/) 47 with medium con dence > 0.4 in Cytoscape (https://cytoscape.org/) 48 . Hub genes were identi ed with the highest degree of connectivity of PPI network analysis (highest con dence > 0.9). Next, hub genes highly associated with cancer (P-value > 0.5).
In both Jensen 49 55 for key genes. GO analysis was contained cellular component (CC), biological process (BP), and molecular function (MF) domains. Finally, one hub gene was selected for the evaluation of expression using the IHC method in RCC tissues. Also, protein expression level for the selected hub gene was considered in divers tumors on UALCAN database (http://ualcan.path.uab.edu/) which provides protein expression analysis option using data from Clinical Proteomic Tumor Analysis Consortium (CPTAC) (https://proteomics.cancer.gov/programs/cptac) 56 .

Study population
Two hundred-ninety ve FFPE tissues from RCC tumor specimen who underwent radical nephrectomy were collected from a major referral university-based urology-nephrology center, the Hasheminejad Hospital, in Tehran, Iran, during 2008-2015. It was con rmed that none of the patients had received radiotherapy or chemotherapy. Three important subtypes of RCC were comprised in this study including ccRCC, pRCC (type I & II), and chRCC. Hematoxylin and eosin (H&E) stained pathology slides and clinical and pathological information were obtained from medical archival records. The following variables were included age, gender, the maximum diameter of tumor size, nucleolar grade, tumor stage, microvascular invasion (MVI), lymph node invasion (LNI), renal vein invasion, histological tumor necrosis (TN), renal sinus fat, and renal sinus invasion, perirenal fat invasion, Gerota's fascia invasion, distant metastasis, and tumor recurrence.
Clinical follow-up was available for all patients until July 2015. Information of disease-speci c survival (DSS) was de ned as the time from the rst operation to death due to renal cancer. Progression-free survival (PFS) was de ned as the interval between primary operation and the last follow-up visit without any disease, evidence of tumor recurrence, or metastasis. In addition, tumor stage and nucleolar grade were de ned according to the pTNM classi cation for RCC [57][58][59] . Moreover, it was agreed that chRCC should not be graded 58 .

Tissue microarray (TMA) construction
The tissue samples were brought into a TMA format as follow, brie y, H & E slides were examined and the most representative areas of the tumor were marked by pathologists. 0.6-mm tissue cylinders were punched out from the selected regions of each donor tumor tissue blocks and transferred into a recipient para n block using tissue arrayer MiniCore (ALPHELYS, Plaisir, France). The TMA blocks were constructed in triplicate for each specimen. The mean of the three scores was calculated as the nal score. Previous studies have shown that three cores are highly representative of the whole Sect. 60,61 . Finally, 4-µm sections were cut from array blocks and transferred to adhesive slides. In each TMA block, adjacent normal renal tissue samples (totally, 30 samples) were included to compare the expression pattern of Talin-1 protein compared to the tumor tissue samples. Immunohistochemistry (IHC) for Talin-1 protein expression IHC was applied to detect the expression of Talin-1 protein. Brie y, all TMA sections were depara nized at 60•C for 20 min and were transferred in xylene and rehydrated via a series of graded alcohols. 3% H2O2 was used to block endogenous peroxides and non-reactive staining for 20 min at room temperature. Antigen retrieval was performed by immersing the tissues in citrate buffer (pH 6.0) for 10 min in an autoclave. Sections were then incubated with a primary rabbit polyclonal antibody to Talin-1 (1:100, ab71333, Abcam Inc., Cambridge, MA, UK) at 4•C overnight. For isotype control, rabbit immunoglobulin G (IgG) (1:100, Invitrogen, Thermo Fisher Scienti c, USA) was used. Then, TMA slides were incubated with anti-rabbit/anti-mouse Envision (Dako, Glostrup, Denmark) as a secondary antibody for 30 min. Afterward, the sections were visualized by diaminobenzidine (DAB) (Dako, Glostrup, Denmark) and stained with hematoxylin (Dako, Glostrup, Denmark). Finally, the slides were dehydrated in alcohol, cleared in xylene, and mounted for examination. In each run of the experiment, human normal kidney tissue was used as a positive control and replacement of the primary antibody with Tris-buffered saline (TBS) as a negative control to con rm the nonspeci c bindings of the secondary antibody.

Immunohistochemistry evaluation
The evaluation procedure of the IHC was reviewed by one independent pathologist (M.A) without any related clinical datainformed. The intensity and percentage of the staining area were considered as two important components of the scoring system for Talin-1 protein expression. The intensity of staining was scored by applying a semi-quantitative system, ranging from negative to strong as follows: 0 (no staining), 1 (weak), 2 (moderate), and 3 (strong). The percentage of staining area was categorized according to the positive tumor cells: less than 25% positive cells, 25-50% positive cells, 51-75% positive cells, and more than 75% positive cells. The nal Talin-1 protein expression score was calculated by multiplying the value of the intensity score by the percentage of positive tumor cells which ranged from 0 to 300, namely histochemical score (Hscore). In this study, the median H-score was chosen to categorize samples with high or low Talin-1 protein expression.

Combined analysis of Talin-1/B7-H3 markers
Previous studies have shown that an association between increased expression of B7-H3 and potential to metastasize 20,62 , therefore we applied our previous data of B7-H3 in ccRCC 20 to used combined analysis to con rm that Talin-1 may be related to metastasis.

Statistical analysis
Analyses were performed using the "statistical software SPSS, version 22.0. Armonk, NY: IBM Corp". To verify the signi cance of association and correlation between Talin-1 protein expression and clinicopathological parameters, Pearson's chi-square and Spearman's correlation tests were performed. Kruskal-Wallis and Mann-Whitney U tests were applied for pairwise comparison between groups. Survival curves by Talin-1 status were estimated using the Kaplan-Meier method with 95% con dence interval (CI) and compared by the log-rank test. Univariate Cox proportional hazard models were t to identify factors signi cantly related to DSS or PFS. To assess whether Talin-1 and the other variables were an independent predictor of survival, multivariate Cox models were also assessed. The data were expressed as the mean; standard deviation (SD) and median; quartile (Q1, Q3). The difference was considered statistically signi cant at a P-value < 0.05.

Declarations
Authors' contributions LS designed and supervised the work, gathered the para n embedded tissues, collected the patient data, prepared the information of patient survival outcomes, analyzed and interpreted the data as well as wrote the manuscript; SV performed the immunohistochemistry examinations and contributed to write the some sections of the manuscript; MA examined hematoxylin and eosin slides, marked the most representative areas in different parts of the tumor for preparing the TMAs blocks, and scored TMAs slides after immunohistochemical staining; FF performed the bioinformatic analysis and wrote the bioinformatic sections in the manuscript. ZM provided for us the laboratory equipment and some materials in this project. All authors read and approved the nal manuscript.