Talin1 is downregulated in testicular germ cell tumors according to combined bioinformatics and experimental approaches

Abstract

Talin1 is a focal adhesion protein involved in cell adhesion and migration, with abnormal expression linked to cancer progression. However, its role in testicular germ cell tumors (TGCTs) remains unclear. This study aimed to evaluate Talin1 expression in TGCTs using integrated bioinformatics and immunohistochemical approaches. Differentially expressed genes were identified through GEO and proteomics datasets. Venn diagram, Gene Ontology (GO), and protein-protein interaction (PPI) analyses revealed Talin1 as a key gene in cell adhesion and migration pathways. Prognostic relevance was assessed using TCGA and GTEx data. Talin1 expression was further examined via immunohistochemistry on 191 TGCT tissues. Results showed that reduced Talin1 expression was associated with higher pT-stage in seminomas (P = 0.036), embryonal carcinoma (P = 0.021), and teratomas (P = 0.044). It was also significantly linked to venous invasion (P = 0.021) and tunica vaginalis invasion (P = 0.049) in embryonal carcinoma, as well as hilum involvement and the presence of tumor-infiltrating lymphocytes in yolk sac tumors. These findings suggest that decreased cytoplasmic Talin1 expression correlates with aggressive tumor behavior and disease progression in TGCTs. Talin1 may have potential as a prognostic biomarker in TGCTs, though further functional studies are necessary to elucidate its mechanistic role and therapeutic significance.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37569-w.

Introduction

Testicular germ cell tumors (TGCTs) are the most commonly diagnosed cancer in young men, particularly in the age range of 20 to 40 years, with the prevalence steadily increasing globally, according to the World Health Organization’s Global Cancer Observatory (GLOBOCAN)^1,2.

TGCTs account for more than 90% of all testicular cancers and are divided into two categories: those developed from germ cell neoplasia in situ (GCNIS) and those unrelated to GCNIS. The GCNIS subtypes include seminomas and non-seminomatous germ cell tumors (NSGCTs), which comprise embryonal carcinoma, yolk sac tumor, teratoma, choriocarcinoma, and mixed germ cell tumors^3,4. While TGCTs are generally highly treatable and respond well to cisplatin-based chemotherapy, challenges persist in the form of short- and long-term complications, and a subset of patients eventually develop resistance to chemotherapy. Furthermore, the molecular mechanisms underlying TGCT development, recurrence, and metastasis remain poorly understood⁵.

The identification of new biomarkers is essential, as they can improve diagnosis, prognosis, and monitoring of diseases^6,7. Currently, α-fetoprotein (AFP), the beta subunit of human chorionic gonadotropin (β-hCG), and lactate dehydrogenase (LDH) are used for TGCT follow-up, but they lack sensitivity and reliable predictive value^8–10 .

The development of more specific advances in bioinformatics, particularly transcriptomic and proteomic analyses, provides opportunities to identify such biomarkers and clarify disease mechanisms. Recent multi-omics and machine-learning approaches have also been successfully applied to predict tumor recurrence and progression in malignancies, further demonstrating the power of integrative bioinformatics in biomarker discovery^7,11–13.

Talin1 is a large cytoskeletal adaptor protein, approximately 270 kDa in size, encoded by the TLN1 gene. It connects integrins to the actin cytoskeleton and plays a crucial role in cell adhesion, migration, and signal transduction. Structurally, Talin1 consists of an N-terminal head domain that binds to β-integrin tails, which activate the integrins, and a C-terminal rod domain that interacts with vinculin, actin, and focal adhesion kinase (FAK)^14,15. Through these interactions, Talin1 regulates mechanotransduction and the dynamics of focal adhesions, both of which are essential for maintaining cell polarity, motility, and tissue architecture. Dysregulation of Talin1 can disrupt cell-matrix interactions and lead to abnormal signaling through pathways such as FAK, PI3K/AKT, and MAPK mechanisms that are commonly associated with tumor invasion and metastasis. Given its pivotal role at the integrin-cytoskeleton interface, Talin1 has emerged as a key modulator of cancer cell behavior^16–18.

Previous studies have revealed that Talin1 exhibits a dual and context-dependent role in cancer biology, functioning either as an oncogenic driver or a tumor suppressor depending on the tumor type. For instance, reduced expression of Talin1 has been linked to a poor prognosis and increased aggressiveness in hepatocellular carcinoma^19,20. Conversely, high levels of Talin1 expression have been associated with tumor progression and unfavorable outcomes in several types of cancer, including oral squamous cell carcinoma²¹, prostate cancer²², colon cancer²³, and nasopharyngeal carcinoma²⁴. Research conducted by our group has also demonstrated that high Talin1 expression correlates with unfavorable clinicopathological features and decreased survival rates in colorectal cancer²⁵, renal cell carcinoma²⁶, melanoma²⁷, and ovarian cancers²⁸. In contrast, in gastric signet-ring cell carcinoma, overexpression of Talin1 was associated with less invasive behavior²⁹. These observations collectively highlight the dual, tissue-specific behavior of Talin1 in human cancers and underscore the importance of defining its biological and prognostic relevance in TGCTs, where its role remains largely unexplored.

In this study, Gene Expression Omnibus (GEO) and proteomic datasets were analyzed to identify key candidate genes for TGCTs, and Talin1 emerged as a promising biomarker associated with tumor progression³⁰. To summarize, this study aims to assess the expression pattern of Talin1 as a potential biomarker in cancer progression and possible prognostic significance in TGCTs. Importantly, the analysis was performed separately for each histological subtype of TGCTs, including seminoma, embryonal carcinoma, yolk sac tumor, and teratoma, to uncover subtype-specific associations. The bioinformatics findings were validated through tissue microarray (TMA) slides using immunohistochemistry (IHC) in TGCT specimens, analyzing their association with clinicopathological factors and survival outcomes.

Results

Bioinformatics approaches

Using Venn diagram analysis, four studies from the GEO database were included to identify common genes among testicular cancer studies. GSE12630 (GPL96) and GSE12630 (GPL7280) were selected, comprising three metastatic versus thirteen non-metastatic cancer samples and one metastatic versus ten non-metastatic cancer samples, respectively. Additionally, GSE15220 and GSE1818 included three cancerous versus three normal samples and twenty cancerous versus three normal samples, respectively. All included GEO datasets represented independent biopsy samples from different patients. None of the selected datasets contained matched samples from the same individuals (e.g., tumor vs. normal from the same patient or primary vs. metastatic from the same tumor). GSE12630 includes expression data from TGCT tissues classified as metastatic vs. non-metastatic tumors collected from different patients. GSE15220 and GSE1818 similarly comprise tumor vs. normal testis tissues from distinct individuals, rather than paired samples. In total, 1,149 common genes (common1) were identified across these studies and are displayed in Fig. 1A.

Fig. 1.

Identification of common genes in testicular cancer datasets. (A) Venn diagram showing 1149 common genes (COMMON1) identified across four microarray datasets: GSE12630-GPL7280, GSE12630-GPL96, GSE15220, and GSE1818 from the GEO database. (B) Venn diagram illustrating the overlap between two proteomic datasets: one derived from testicular germ cell tumor (TGCT) cell lines (PXD010275, in vitro) and the other from TGCT patient tissues (PXD013658, in vivo). A total of 25 proteins were identified in both studies. This comparison represents an exploratory integration of published proteomic findings rather than a cross-normalized meta-analysis.

Proteomics analysis identified two valuable datasets containing proteomic information and adjacent normal tissues. These datasets were downloaded from the supplementary information of the corresponding published articles^31,32. The datasets, PXD010275 (Proteomics1) and PXD013658 (Proteomics2), were included to explore protein-level findings in TGCT. From these sources, we extracted lists of differentially expressed or reported proteins. Twenty-five overlapping proteins were identified, as shown in Fig. 1B. This comparison served as an exploratory integration rather than a formal differential protein abundance analysis. Full lists and references are provided in Supplementary Table 1. Next, the selected genes were analyzed using Enrichr based on Gene Ontology (GO). The cellular components identified included cell-substrate junction (GO:0030055), focal adhesion (GO:0005925), intermediate filament (GO:0005882), tertiary granule lumen (GO:1904724), specific granule lumen (GO:0035580), and intracellular organelle lumen (GO:0070013), all with a significant adjusted p-value < 0.05. The significant pathways related to these genes included positive regulation of protein localization to membrane (GO:1905477), positive regulation of protein localization to the cell periphery (GO:1904377), positive regulation of protein localization to plasma membrane (GO:1903078), renal filtration (GO:0097205), cellular response to hormone stimulus (GO:0032870), regulation of protein localization to plasma membrane (GO:1903076), positive regulation of gene silencing by miRNA (GO:2000637), glial cell differentiation (GO:0010001), negative regulation of cell migration (GO:0030336), response to cytokine (GO:0034097), and response to estradiol (GO:0032355).

The significant molecular functions of the genes were cadherin binding (GO:0045296), LIM domain binding (GO:0030274), cysteine-type peptidase activity (GO:0008234), RNA binding (GO:0003723), BH domain binding (GO:0051400), BH3 domain binding (GO:0051434), primary miRNA binding (GO:0070878), dihydropyrimidinase activity (GO:0004157), keratin filament binding (GO:1990254), ubiquitin-protein transferase activator activity (GO:0097027), retinol-binding (GO:0019841), phosphatase binding (GO:0019902), protein phosphatase binding (GO:0019903), clathrin heavy chain binding (GO:0032050), RAGE receptor binding (GO:0050786), O-methyltransferase activity (GO:0008171), hydrolase activity acting on carbon-nitrogen (but not peptide) bonds, in cyclic amides (GO:0016812), MAP kinase activity (GO:0004709), filamin binding (GO:0031005), death domain binding (GO:0070513), and vinculin binding (GO:0017166), all of which are shown in Supplementary Table 2.

A protein–protein interaction (PPI) network was constructed using the STRING database, employing k-means clustering with a high confidence level (≥ 0.7), as shown in Fig. 2.

Fig. 2.

Protein–protein interaction (PPI) network of the 25 common genes (COMMON2) identified from transcriptomic and proteomic datasets. The 25 commonly deregulated proteins were grouped into distinct functional clusters. Notably, TLN1 (Talin1) was positioned in a separate subcluster from the main interaction hub centered around EGFR, connected by a dotted line indicating a lower-confidence or indirect association. This clustering pattern suggests that Talin1 may participate in a functionally specialized module related to cytoskeletal remodeling and focal adhesion dynamics, distinct from the receptor tyrosine kinase signaling network. The PPI network was generated using the STRING database (confidence score ≥ 0.7) and visualized with k-means clustering. Nodes represent proteins encoded by the identified genes, and edges indicate known or predicted protein associations. Different node colors represent distinct clusters of functionally associated proteins.

Based on the integrated bioinformatics analyses, TLN1 (Talin1) was selected as the focus of this study due to its biological significance and consistent dysregulation observed across multiple independent transcriptomic and proteomic datasets related to TGCTs. Although the direction and magnitude of the Talin1 fold change varied among individual datasets, likely due to the inherent heterogeneity in the composition of cohorts and analytical platforms, it consistently ranked among the top differentially expressed genes linked to cell migration, focal adhesion, and cytoskeletal regulation. Functionally, Talin1 acts as a core mechanosensitive adaptor that connects integrins to the actin cytoskeleton, facilitating focal adhesion turnover, vinculin binding, and downstream signaling through the FAK–PI3K/AKT pathways. These processes are vital for cell motility and tumor invasiveness^17,18. Within the STRING-derived PPI network, Talin1 localized to an adhesion and cytoskeleton-enriched subcluster, distinct from canonical growth factor signaling hubs. This unique positioning, together with its recurrent dysregulation and lack of prior characterization in TGCTs, supported its selection for further validation.

A confirmatory bioinformatics analysis was performed using the Gene Expression Profiling Interactive Analysis (GEPIA) database by integrating TCGA and Genotype-Tissue Expression (GTEx) datasets to compare Talin1 expression between TGCTs and normal testicular tissues (Fig. 3A). The results demonstrated that Talin1 expression was overall lower in tumor samples than in normal tissues. Kaplan–Meier survival analyses were performed to explore potential associations between Talin1 expression and overall survival (Fig. 3B) or disease-free survival (Fig. 3C) in TGCTs. No statistically significant differences were observed (log-rank P > 0.05) in these analyses. Moreover, Talin1 transcript levels, evaluated through the UALCAN database on TCGA data, showed a gradual decline from stage I to stage III TGCTs (Fig. 3D), indicating a loss of adhesion-related regulation during malignancy and supporting the association between lower Talin1 expression and disease progression. The pan-cancer analysis (Fig. 3E) further illustrates Talin1 expression across various human cancers, providing a comprehensive view of its role in tumor biology. Collectively, the bioinformatics findings suggest that the downregulation of Talin1 in TGCTs is correlated with tumor progression, as reflected by its lower expression in tumors compared with normal tissues and its gradual decrease with advancing pathological stage. However, it is important to note that these bioinformatics analyses were based on pooled TGCT datasets that included all histological subtypes. Therefore, the results reflect overall expression patterns rather than those specific to individual subtypes. Additionally, the lack of a significant association between survival and Talin1 expression in the combined cohort may be due to this biological heterogeneity. To overcome this limitation, we performed IHC analyses categorized by histological subtype: specifically, seminoma, embryonal carcinoma, yolk sac tumor, and teratoma. This approach aimed to clarify Talin1’s expression patterns and assess its potential prognostic significance within each subtype.

Fig. 3.

Expression analysis and prognostic evaluation of TLN1 (Talin1) in testicular cancer and other The Cancer Genome Atlas (TCGA) tumors. (A) Comparison of Talin1 expression levels between TGCT tissues (T) and normal testicular tissues (N) using the Gene Expression Profiling Interactive Analysis (GEPIA) database, showing overall downregulation in tumors. (B) Kaplan–Meier plot for overall survival (OS) and (C) disease-free survival (DFS) of TGCT patients based on high and low Talin1 expression. However, no statistically significant difference was observed (log-rank P > 0.05). (D) Analysis of Talin1 mRNA expression in TGCTs at various pathological stages in the UALCAN database reveals a gradual decrease as the stage progresses, indicating a loss of adhesion-related regulation during malignancy and supporting the association between lower Talin1 expression and disease progression. (E) Talin1 expression levels across various TCGA tumor types are presented as a boxplot.

Patients’ characteristics

In this study, 191 formalin-fixed paraffin-embedded (FFPE) tissue specimens were included, some of which were pure, and others contained mixed histological subtypes, including seminoma, embryonal carcinoma, yolk sac tumor, and teratoma. The distribution of the specimens was as follows: seminoma (64 pure, 20 mixed), embryonal carcinoma (5 pure, 41 mixed), yolk sac tumor (2 pure, 44 mixed), and teratoma (8 pure, 30 mixed). In total, 70 seminomas, 42 embryonal carcinomas, 41 yolk sac tumors, and 24 teratomas were identified among the TGCT histological subtypes. Supplementary Table 3 summarizes the clinicopathological characteristics of tumor samples.

Expression levels of Talin 1 in TGCT histological subtypes and adjacent normal tissue samples

The IHC technique was used to assess the expression levels of Talin1 protein on TMA slides of TGCTs, evaluating the intensity of staining, percentage of positive tumor cells, and histochemical score (H-score). Of the 191 TGCT specimens initially included, 177 were evaluable for analysis after excluding tissue cores that were lost or insufficient during staining.

Cytoplasmic and membranous expressions of Talin1 were observed in various histological subtypes of TGCTs at different intensities in the cell membrane and cytoplasm of tissue samples, so these results were analyzed separately (Fig. 4; Table 1). The results of the normality tests, including Kolmogorov–Smirnov and Shapiro–Wilk tests, revealed that the data were not normally distributed (all, P < 0.001). Therefore, the median of the H-scores was used for classifying the data. Considering that the median of H-scores for membranous Talin1 expression was 0.00 for three subtypes, the median was applied as a cut-off only for cytoplasmic Talin1 expression. For membranous expression in these cases, H-scores were classified into two groups: one based on the expression of Talin1 protein and the other for no expression of Talin1. This approach was biologically reasonable because Talin1 staining at the cell membrane indicates the presence or absence of protein localization, rather than presenting a continuous intensity pattern. This makes binary classification (positive vs. negative) more appropriate. This classification, based on these cut-offs, was consistently applied for both the analysis of clinicopathological characteristics and survival outcomes. Similar stratification approaches have been used in prior IHC-based biomarker studies^10,29,33. Our findings from the Mann-Whitney U test revealed that Talin1 expression in TGCTs was significantly lower compared to its adjacent normal tissues (Figs. 4 and 5). Pearson’s chi-squared test was applied to examine the association between Talin1 protein expression and the histological subtypes of TGCTs. The results indicated a statistically significant association between cytoplasmic expression of Talin1 and the intensity of staining, as well as H-score across the different histological subtypes (Table 1). Moreover, the nonparametric Kruskal-Wallis and Mann-Whitney U tests were used to compare the median expression levels of Talin1 protein among the TGCT subtypes. The Kruskal-Wallis test showed a statistically significant difference in cytoplasmic Talin1 expression among the TGCT subtypes (P = 0.001), while no significant difference was found in membranous expression (P = 0.067). The Mann-Whitney U test further confirmed statistically significant differences in cytoplasmic Talin1 expression between histological subtypes of TGCTs (Fig. 6).

Fig. 4.

Immunohistochemical staining of cytoplasmic and membranous expression of Talin1 in testicular germ cell tumors (TGCTs). Representative images showing varying levels of cytoplasmic and membranous Talin1 expression in different TGCT subtypes. (A,B) Seminoma with strong and weak Talin1 expression, respectively. (C,D) Embryonal carcinoma showing strong (C) and weak (D) Talin1 staining. (E,F) Yolk sac tumor with high (E) and low (F) expression. (G,H) Teratoma samples exhibiting variable Talin1 staining. (I) Adjacent normal tissue showed strong Talin1 expression. (J) Positive control from renal tissue showing Talin1 expression. (K) Negative control with no primary antibody.

Table 1.

Association of membranous and cytoplasmic Talin-1 expressions among subtypes of testicular germ cell tumors (TGCTs).

Expression	Membranous Talin-1 expression
	Seminoma N (%)	Embryonal carcinomas N (%)	Yolk sac tumors N (%)	Teratomas N (%)	P-value
Intensity of staining Negative (0) Weak (+ 1) Moderate (+ 2) Strong (+ 3)	37 (52.9) 0 (0.0) 1 (1.4) 32 (45.7)	10 (23.8) 2 (4.8) 8 (19.0) 22 (52.4)	20 (48.8) 0 (0.0) 4 (9.8) 17 (41.5)	14 (58.3) 0 (0.0) 0 (0.0) 10 (41.7)	0.002
Percentage of positive tumor cells < 25% 25–50% 75%- 51 > 75%	70 (100.0) 0 (0.0) 0 (0.0) 0 (0.0)	0 (0.0) 41 (97.6) 0 (0.0) 1 (2.4)	40 (97.6) 0 (0.0) 1 (2.4) 0 (0.0)	23 (95.8) 0 (0.0) 1 (4.2) 0 (0.0)	0.505
H-score (cut off) Positive/negative or Median (Low ≤, High > )	37 (52.9) 33 (47.1)	27 (64.3) 15 (35.7)	22 (53.7) 19 (46.3)	14 (58.3) 10 (41.7)	0.219
Expression	Cytoplasmic Talin-1 expression
Intensity of staining Negative (0) Weak (+ 1) Moderate (+ 2) Strong (+ 3)	1 (1.4) 1 (1.4) 14 (20.0) 54 (77.1)	0 (0.0) 3 (7.1) 18 (42.9) 21 (50.0)	0 (0.0) 0 (0.0) 6 (14.6) 35 (85.4)	1 (4.2) 0 (0.0) 4 (16.7) 19 (79.2)	0.009
Percentage of positive tumor cells < 25% 25–50% 51%- 75% > 75%	2 (2.9) 0 (0.0) 2 (2.9) 66 (94.3)	0 (0.0) 2 (4.8) 3 (7.1) 37 (88.1)	1 (2.4) 2 (4.9) 1 (2.4) 37 (90.2)	2 (8.3) 6 (25.0) 3 (12.5) 13 (54.2)	< 0.001
H-score (cut off) Median (Low ≤, High > )	38 (54.3) 32 (45.7)	22 (52.4) 20 (47.6)	25 (61.0) 16 (39.0)	12 (50.0) 12 (50.0)	0.049
Total	70	48	41	24

H-score indicates histological score.

P values are based on Pearson’s χ2 test.

Values in bold and italic are statistically significant.

Fig. 5.

Box plot analysis of cytoplasmic and membranous Talin1 expression between TGCT subtypes and adjacent normal tissues. Box plots illustrate the differential expression of Talin1 in tumor versus adjacent normal testicular tissues across four major histological subtypes of TGCTs: (A,B) seminoma, (C,D) embryonal carcinoma, (E,F) yolk sac tumor, and (G,H) teratoma. Cytoplasmic (left panels) and membranous (right panels) expression levels were quantified by H-scores. Statistical comparisons were performed using the non-parametric Mann–Whitney U test. In all subtypes, both cytoplasmic and membranous Talin1 expression was significantly lower in tumor tissues than in adjacent normal tissues.

Fig. 6.

Box plot analysis of cytoplasmic Talin1 expression across histological subtypes of testicular germ cell tumors (TGCTs). Cytoplasmic expression levels of Talin1 were compared among seminomas, embryonal carcinomas, yolk sac tumors, and teratomas using the Mann–Whitney U test. Statistically significant differences were observed between seminomas and embryonal carcinomas (P = 0.005), embryonal carcinomas and yolk sac tumors (P = 0.037), yolk sac tumors and teratomas (P = 0.006), and seminomas and teratomas (P = 0.001).

Associations between Talin1 protein expression (membranous and cytoplasmic expression) and clinicopathological features in histological subtypes of TGCTs

Seminomas

Pearson’s chi-squared test showed a significant association between the cytoplasmic expression of Talin1 protein and the advancement of primary tumor (pT) stage (P = 0.036) (Table 2). Spearman’s correlation test was performed to assess the correlation between Talin1 protein expression and clinicopathological features. The results revealed a significant negative correlation between cytoplasmic expression of Talin1 and pT stages (P = 0.036). Moreover, no significant association was found between cytoplasmic and membranous Talin1 expressions and other clinicopathological parameters (Table 2).

Table 2.

The association between membranous and cytoplasmic Talin-1 expression and clinicopathological characteristic of seminoma samples.

Characteristics of tumors	Total cases N (%)	Membranous expression N (%)		P-value	Cytoplasmic expression N (%)		P- value
Seminomas	70	Positive	Negative		Low (≤ 270)	High (> 270)
Median age, years (Range) ≤Median > Median	33 (20–56) 38 (54.3) 32 (45.7)	20 (54.1) 17 (45.9)	18 (54.5) 15 (45.5)	0.967	21 (55.3) 17 (44.7)	17 (53.1) 15 (46.9)	0.858
Median tumor size (cm)(Range) ≤Median > Median	4.5 (1-19.5) 37 (52.9) 33 (47.1)	21 (56.8) 16 (43.2)	16 (48.5) 17 (51.5)	0.489	17 (44.7) 21 (55.3)	20 (62.5) 12 (37.5)	0.138
Primary tumor (PT) stage pT0 + pT1 pT2 + pT3	48 (68.6) 19 (27.1)	25 (67.6) 12 (32.4)	23 (69.7) 10 (30.3)	0.848	22 (57.9) 16 (42.1)	26 (81.3) 6 (18.8)	0.036
Venous invasion Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Germ cell neoplasia in situ Present Not identified	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Rete testis involvement Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Hilum involvement Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Tumor infiltrating lymphocyte Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Epididymis involvement Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Tunica vaginalis invasion Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Tunica albuginea invasion Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Spermatic cord invasion Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Spermatic cord margin invasion Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Tumor necrosis Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Scar Present Absent	0 (0.0) 70 (100.0)	0 (0.0) 37 (100.0)	0 (0.0) 33 (100.0)	-*	0 (0.0) 38 (100.0)	0 (0.0) 32 (100.0)	-*
Distant metastasis Present Absent	1 (1.4) 69 (98.6)	1 (2.7) 36 (97.3)	0 (0.0) 33 (100.0)	0.341	1 (2.6) 37 (97.4)	0 (0.0) 32 (100.0)	0.355
Tumor recurrence Yes No	3 (4.3) 67 (95.7)	2 (5.4) 35 (94.6)	1 (3.0) 32 (97.0)	0.624	1 (2.6) 37 (97.4)	2 (6.3) 30 (93.8)	0.456

H-score indicates Histological score.

P values are based on Pearson’s χ2 test.

*No statistical analyses are computed because the parameter is constant.

Values in bold are statistically significant.

Embryonal carcinoma

Using Pearson’s chi-squared test, we found a statistically significant association between cytoplasmic expression of Talin1 protein and advanced pT stage (P = 0.021), venous invasion (P = 0.021), and invasion of the tunica vaginalis (P = 0.049) (Table 3). Furthermore, Spearman’s correlation test revealed a significant negative correlation between cytoplasmic expression of Talin1 and advanced pT stage (P = 0.021), as well as venous invasion (P = 0.021). However, in patients with embryonal carcinoma, no significant association was observed between membranous Talin1 expression and clinicopathological features (Table 3).

Table 3.

The association between membranous and cytoplasmic Talin-1 expression and clinicopathological characteristic of embryonal carcinoma samples.

Characteristics of tumors	Total cases N (%)	Membranous expression N (%)		P-value	Cytoplasmic expression N (%)		P- value
Embryonal carcinomas	42	Low (≤ 15)	High (> 15)		Low (≤ 200)	High (> 200)
Median age, years (Range) ≤Median > Median	28 (18–59) 22 (52.4) 20 (47.6)	15 (55.6) 12 (44.4)	7 (46.7) 8 (53.3)	0.580	13 (59.1) 9 (40.9)	9 (45.0) 11 (55.0)	0.361
Median tumor size (cm)(Range) ≤Median > Median	4 (0-11.5) 25 (59.5) 17 (40.5)	19 (70.4) 8 (29.6)	6 (40.0) 9 (60.0)	0.055	11 (50.0) 11 (50.0)	14 (70.0) 6 (30.0)	0.187
Primary tumor (PT) stage pT0 + pT1 pT2 + pT3	26 (61.9) 16 (38.1)	17 (63.0) 10 (37.0)	9 (60.0) 6 (40.0)	0.850	10 (45.5) 12 (54.5)	16 (80.0) 4 (20.0)	0.021
Venous invasion Present Absent	16 (38.1) 26 (61.9)	11 (40.7) 16 (59.3)	5 (33.3) 10 (66.7)	0.636	12 (54.5) 10 (45.5)	4 (20.0) 16 (80.0)	0.021
Germ cell neoplasia in situ Present Not identified	36 (85.7) 6 (14.3)	23 (85.2) 4 (14.8)	13 (86.7) 2 (13.3)	0.895	19 (86.4) 3 (13.6)	17 (85.0) 3 (15.0)	0.900
Rete testis involvement Present Absent	14 (33.3) 28 (66.7)	8 (29.6) 19 (70.4)	6 (40.0) 9 (60.0)	0.495	9 (40.9) 13 (59.1)	5 (25.0) 15 (75.0)	0.275
Hilum involvement Present Absent	7 (16.7) 35 (83.3)	5 (18.5) 22 (81.5)	2 (13.3) 13 (86.7)	0.666	5 (22.7) 17 (77.3)	2 (10.0) 18 (90.0)	0.269
Tumor infiltrating lymphocyte Present Absent	8 (19.0) 34 (81.0)	6 (22.2) 21 (77.8)	2 (13.3) 13 (86.7)	0.482	4 (18.2) 18 (81.8)	4 (20.0) 16 (80.0)	0.881
Epididymis involvement Present Absent	4 (9.5) 38 (90.5)	2 (7.4) 25 (92.6)	2 (13.3) 13 (86.7)	0.531	2 (9.1) 20 (90.9)	2 910.0) 18 (90.0)	0.920
Tunica vaginalis invasion Present Absent	3 (7.1) 39 (92.9)	2 (7.4) 25 (92.6)	1 (6.7) 14 (93.3)	0.929	0 (0.0) 22 (100.0)	17 (85.0) 3 (15.0)	0.049
Tunica albuginea invasion Present Absent	25 (59.5) 17 (40.5)	16 (59.3) 11 (40.7)	9 (60.0) 6 (40.0)	0.963	12 (54.5) 10 (45.5)	13 (65.0) 7 (35.0)	0.491
Spermatic cord invasion Present Absent	2 (4.8) 40 (95.2)	1 (3.7) 26 (96.3)	1 (6.7) 14 (93.3)	0.666	1 (4.5) 21 (95.5)	1 (5.0) 19 (95.0)	0.945
Spermatic cord margin invasion Present Absent	1 (2.4) 41 (97.6)	1 (3.7) 26 (96.3)	0 (0.0) 15 (100.0)	0.451	1 (4.5) 21 (95.5)	0 (0.0) 20 (100.0)	0.335
Tumor necrosis Present Absent	37 (88.1) 5 (11.9)	24 (88.9) 3 (11.1)	13 (86.7) 2 (13.3)	0.831	20 (90.9) 2 (9.1)	17 (85.0) 3 (15.0)	0.555
Scar Present Absent	0 (0.0) 42 (100.0)	0 (0.0) 2 (100.0)	0 (0.0) 15 (100.0)	-*	0 (0.0) 22 (100.0)	0 (0.0) 20 (100.0)	-*
Distant metastasis Present Absent	1 (2.4) 41 (97.6)	1 (3.7) 26 (96.3)	0 (0.0) 15 (100.0)	0.451	1 (4.5) 21 (95.5)	0 (0.0) 20 (100.0)	0.335
Tumor recurrence Yes No	3 (7.1) 39 (92.9)	1 (3.7) 26 (96.3)	2 (13.3) 13 (86.7)	0.246	1 (4.5) 21 (95.5)	2 (10.0) 18 (90.0)	0.493

H-score indicates Histological score.

P values are based on Pearson’s χ2 test.

*No statistical analyses are computed because the parameter is constant.

Values in bold are statistically significant.

Yolk sac tumor

Our analysis revealed a statistically significant association between cytoplasmic expression of Talin1 and increased age (P = 0.041), hilum involvement (P = 0.034), and tumor-infiltrating lymphocytes (P = 0.017) (Table 4). Moreover, Spearman’s correlation test showed a significant negative correlation between low levels of cytoplasmic Talin1 expression and increased age (P = 0.042), hilum involvement (P = 0.034), and tumor-infiltrating lymphocytes (P = 0.017). However, no association was found between membranous Talin1 protein expression and clinicopathological features in these patients (Table 4).

Table 4.

The association between membranous and cytoplasmic Talin-1 expression and clinicopathological characteristic of yolk sac tumor samples.

Characteristics of tumors	Total cases N (%)	Membranous expression N (%)		P-value	Cytoplasmic expression N (%)		P- value
Yolk sac tumors	41	Positive	Negative		Low (≤ 285)	High (> 285)
Median age, years (Range) ≤Median > Median	27 (16–59) 21 (51.2) 20 (48.8)	11 (50.0) 11 (50.0)	10 (52.6) 9 (47.4)	0.867	16 (64.0) 9 (36.0)	5 (31.3) 11 (68.8)	0.041
Median tumor size (cm)(Range) ≤Median > Median	4 (1.5–11.5) 22 (53.7) 19 (46.3)	13 (59.1) 9 (40.9)	9 (47.4) 10 (52.6)	0.453	14 (56.0) 11 (44.0)	8 (50.0) 8 (50.0)	0.707
Primary tumor (PT) stage pT0 + pT1 pT2 + pT3	25 (61.0) 16 (39.0)	12 (54.5) 10 (45.5)	13 (68.4) 6 (31.6)	0.364	15 (60.0) 10 (40.0)	10 (62.5) 6 (37.5)	0.873
Venous invasion Present Absent	16 (39.0) 25 (61.0)	10 (45.5) 12 (54.5)	6 (31.6) 13 (68.4)	0.364	10 (40.0) 15 (60.0)	6 (37.5) 10 (62.5)	0.873
Germ cell neoplasia in situ Present Not identified	35 (85.4) 6 (14.6)	19 (86.4) 3 (13.6)	16 (84.2) 3 (15.8)	0.846	22 (88.0) 3 (12.0)	13 (81.3) 3 (18.8)	0.551
Rete testis involvement Present Absent	15 (36.6) 26 (63.4)	9 (40.9) 13 (59.1)	6 (31.6) 13 (68.4)	0.536	11 (44.0) 14 (56.0)	4 (25.0) 12 (75.0)	0.218
Hilum involvement Present Absent	6 (14.6) 35 (85.4)	5 (22.7) 17 (77.3)	1 (5.3) 18 (94.7)	0.115	6 (24.0) 19 (76.0)	0 (0.0) 16 (100.0)	0.034
Tumor infiltrating lymphocyte Present Absent	11 (26.8) 30 (73.2)	8 (36.4) 14 (63.6)	3 (15.8) 16 (84.2)	0.138	10 (40.0) 15 (60.0)	1 (6.3) 15 (93.8)	0.017
Epididymis involvement Present Absent	5 (12.2) 36 (87.8)	3 (13.6) 19 (86.4)	2 (10.5) 17 (89.5)	0.762	2 (8.0) 23 (92.0)	3 (18.8) 13 (81.3)	0.305
Tunica vaginalis invasion Present Absent	3 (7.3) 38 (92.7)	1 (4.5) 21 (95.5)	2 (10.5) 17 (89.5)	0.463	2 (8.0) 23 (92.0)	1 (6.3) 15 (93.8)	0.834
Tunica albuginea invasion Present Absent	23 (56.1) 18 (43.9)	14 (63.6) 8 (36.4)	9 (47.4) 10 (52.6)	0.295	17 (68.0) 8 (32.0)	6 (37.5) 10 (62.5)	0.055
Spermatic cord invasion Present Absent	4 (9.8) 37 (90.2)	2 (9.1) 20 (90.9)	2 (10.5) 17 (89.5)	0.877	1 (4.0) 24 (96.0)	3 (18.8) 13 (81.3)	0.120
Spermatic cord margin invasion Present Absent	2 (4.9) 39 (95.1)	1 (4.5) 21 (95.5)	1 (5.3) 18 (94.7)	0.915	0 (0.0) 25 (100.0)	2 (12.5) 14 (87.5)	0.070
Tumor necrosis Present Absent	37 (90.2) 4 (9.8)	20 (90.9) 2 (9.1)	17 (89.5) 2 (10.5)	0.877	21 (84.0) 4 (16.0)	16 (100.0) 0 (0.0)	0.092
Scar Present Absent	0 (0.0) 41 (100.0)	0 (0.0) 22 (100.0)	0 (0.0) 19 (100.0)	-*	0 (0.0) 25 (100.0)	0 (0.0) 16 (100.0)	-*
Distant metastasis Present Absent	2 (4.9) 39 (95.1)	1 (4.5) 21 (95.5)	1 (5.3) 18 (94.7)	0.915	1 (4.0) 24 (96.0)	1 (6.3) 15 (93.8)	0.744
Tumor recurrence Yes No	6 (14.6) 35 (85.4)	3 (13.6) 19 (86.4)	16 (84.2) 3 (15.8)	0.846	3 (12.0) 22 (88.0)	3 (18.8) 13 (81.3)	0.551

H-score indicates Histological score.

P values are based on Pearson’s χ2 test.

*No statistical analyses are computed because the parameter is constant.

Values in bold are statistically significant.

Teratoma

In cytoplasmic expression, a low level of Talin1 was associated with an advanced pT stage (P = 0.044) (Supplementary Table 4). No significant association was observed between membranous Talin1 expression and clinicopathological parameters.

Clinical outcomes in histological subtypes of TGCTs

Of the 191 TGCT specimens included in this study, 177 were eligible for survival analysis after excluding cases with insufficient or damaged tissue sections. In the total cohort, metastasis was observed in 4 patients (2.3%), recurrence in 17 patients (9.6%), and cancer-related deaths in 5 patients (2.8%) during follow-up. The median follow-up duration was 57 months for disease-specific survival (DSS) (Q1, Q3: 42, 57), and 51 months for progression-free survival (PFS) (Q1, Q3: 36, 51), with a range of 16–109 months for DSS and 1-109 months for PFS. The main characteristics of the patients included in the survival analysis are summarized in Supplementary Table 5.

Survival analysis based on Talin1 protein expression in histological subtypes of TGCTs

Seminomas

The product-limit estimator, also known as the Kaplan–Meier estimator, was used to calculate the survival function from lifetime data. No cancer-related deaths occurred in patients with seminomas; therefore, the Kaplan–Meier survival analysis curves could not be drawn.

Embryonal carcinomas

Kaplan–Meier survival curves showed no significant differences in DSS or PFS between patients with high or low cytoplasmic and membranous expression of Talin1 protein (DSS: P = 0.967, P = 0.716; PFS: P = 0.959, P = 0.651, respectively) (Supplementary Fig. 1A, B, C, D).

Yolk sac tumor

Kaplan–Meier survival analysis revealed no significant differences in DSS or PFS between patients with high or low cytoplasmic and membranous Talin1 expression (DSS: P = 0.749, 0.927; PFS: P = 0.766, 0.955, respectively) (Supplementary Fig. 2A, B, C, D).

Teratoma

No significant differences were found in DSS or PFS between patients with high or low cytoplasmic and membranous Talin1 expression in teratomas (DSS: P = 0.317, P = 0.317; PFS: P = 0.157, P = 0.157, respectively) (Supplementary Fig. 3A, B, C, D).

Discussion

TGCTs are not fully curable, with recurrence observed in approximately 15% to 30% of patients undergoing chemotherapy^34,35. These patients often experience significant therapeutic side effects, including secondary leukemia^36,37, cerebrovascular accidents^38,39, renal disease⁴⁰, chemotherapy-induced stroke, and carotid artery occlusion resulting from long-term treatments⁴¹. Given the limitations of current treatments and the potential for recurrence, identifying novel prognostic biomarkers is essential to improving patient outcomes in TGCTs.

In this study, we first conducted comprehensive bioinformatics analyses utilizing multiple online platforms and databases, including GEO microarrays, to identify genes implicated in the pathogenesis of TGCTs. These analyses yielded valuable insights into the molecular pathways and biological processes underlying TGCT progression. Among the identified candidates, Talin1 emerged as a promising molecular marker and a potential prognostic target. Talin1 plays a critical function in regulating cellular adhesion and migration, which are fundamental processes in tumor dissemination. Notably, Talin1 functionally interacts with FAK, orchestrating dynamic modulation of adhesion sites and thereby influencing downstream signal transduction pathways. Dysregulation of FAK–Talin1 interactions can profoundly affect cancer cell motility and invasiveness, ultimately contributing to tumor progression^17,42,43.

Although Talin1 showed variable transcriptional regulation across different GEO and proteomic datasets, this variability likely reflects the biological and technical heterogeneity inherent to TGCT cohorts rather than a lack of biological significance. TGCTs are a varied group of neoplasms that arise from pluripotent germ cells. They are distinguished by unique histological subtypes, differences in cellular differentiation, tumor purity, and variations in their microenvironment^44,45. These factors, along with differences in analytical platforms and normalization methods, can lead to apparent inconsistencies in transcriptomic data. Importantly, our integrative bioinformatics analysis identified Talin1 as a repeatedly dysregulated and highly connected node within the focal adhesion and cytoskeletal remodeling pathways. It interacts with integrins, FAK, and vinculin, key regulators of cell adhesion, motility, and signal transduction. The consistent alterations of Talin1 across multiple datasets suggest that its perturbation is not random but a stable feature of TGCT biology. Due to the origin of TGCTs, they exhibit significant biological flexibility and changes in adhesion-related pathways. Recent studies indicate that germ cell tumors exhibit epithelial-mesenchymal plasticity and show disruptions in key adhesion molecules. For example, there is a notable decrease in E-cadherin/β-catenin expression in seminomas, suggesting a shift toward a less stable adhesion phenotype between the cells and their surrounding matrix^46,47. These characteristics differ from those of most traditional epithelial cancer types, providing a biologically relevant context in which alterations in focal adhesion regulators, such as Talin1, may promote invasive behavior in TGCTs and underscore a potential role as a prognostic biomarker.

To further examine this hypothesis, we evaluated Talin1 protein levels in 191 TGCT specimens and adjacent normal tissues across the four major histological subtypes of TGCTs, including seminoma, embryonal carcinoma, yolk sac tumor, and teratoma. Our findings demonstrated that both cytoplasmic and membranous Talin1 expression levels were significantly lower in TGCTs compared to adjacent normal tissues, suggesting that downregulation of Talin1 may be associated with tumor progression rather than serving a protective function. Given Talin1’s established function in cytoskeletal remodeling and integrin-mediated adhesion, its decreased expression in more advanced TGCTs may reflect a shift toward increased cellular motility, invasiveness, and reduced adhesion^17,48. Moreover, our analysis revealed significant variations in Talin1 expression among different histological subtypes of TGCTs. Notably, cytoplasmic Talin1 expression demonstrated statistically significant differences (P = 0.001), indicating that Talin1 may play a more prominent role in the biological behavior of specific TGCT subtypes, particularly those with higher metastatic potential, such as embryonal carcinoma and yolk sac tumors⁴⁹. In contrast, membranous Talin1 expression showed no statistically significant differences among TGCT subtypes (P = 0.067). However, variations in staining intensity were observed, which may reflect differences in the activation status of Talin1 rather than changes in its overall expression levels. These findings suggest potential differences in integrin-mediated signaling dynamics across distinct TGCT subtypes.

We further observed that lower cytoplasmic Talin1 expression was significantly associated with more advanced pT stages across TGCT subtypes, including seminomas, embryonal carcinomas, and teratomas, corroborating the findings of our bioinformatics analyses. In yolk sac tumor specimens, a statistically significant negative correlation was identified between cytoplasmic Talin1 expression and the presence of tumor-infiltrating lymphocytes (TILs). Given the well-established role of TILs in mediating anti-tumor immunity and their recognized prognostic and predictive significance⁵⁰, these observations suggest that reduced Talin1 expression may contribute to tumor immune evasion as well as enhanced tumor progression. Collectively, these results indicate that cytoplasmic Talin1 plays a more active role in promoting tumor aggressiveness and metastatic behavior. In contrast, its membranous expression remains relatively stable across the various subtypes of TGCT. This prognostic relevance may be explained by its molecular mechanisms. Mechanistically, Talin1 orchestrates focal adhesion assembly by interacting with FAK and integrins, thereby regulating cellular adhesion, migration, and survival^17,18. Disruption of Talin1-mediated adhesion complexes leads to alterations in critical downstream signaling cascades, including the FAK/PI3K/AKT pathway, ultimately enhancing cancer cell motility and invasiveness¹⁷. In addition to its function in adhesion dynamics, Talin1 has been implicated in the regulation of epithelial-mesenchymal transition (EMT), a pivotal event in tumor metastasis. Downregulation of Talin1 compromises epithelial integrity, fostering the acquisition of a mesenchymal phenotype characterized by increased migratory and invasive capabilities^20,51. Given its multifaceted roles in tumor biology, the biological function of Talin1 appears to be highly context-dependent across cancer types, with either tumor-promoting or tumor-suppressive effects reported in various malignancies^19–29,51. In this context, our findings highlight the unique biological behavior of Talin1 in TGCTs, where reduced cytoplasmic expression correlates with higher tumor stage and aggressiveness, contrasting with trends observed in several epithelial cancers.

However, previous studies have demonstrated that Talin1 overexpression is associated with tumor aggressiveness and poor prognosis in several epithelial cancers¹⁷. Our findings suggest a different pattern in TGCTs, which display unique biological behaviors associated with their germ cell origin, including dedifferentiation, cytoskeletal disorganization, and high cellular plasticity. TGCTs are also characterized by their high sensitivity to platinum-based chemotherapy and a unique tumor microenvironment, which includes immune cell infiltration and variable stromal responses. Certain subtypes, such as embryonal carcinoma, exhibit stronger invasive and metastatic behaviors, which may be regulated differently by adhesion-related pathways^44,45. In this context, reduced expression of Talin1 could lead to a loss of epithelial characteristics, increased motility, and evasion of immune responses, all of which contribute to a higher metastatic potential. Therefore, the observed association between low Talin1 expression and metastatic behavior in TGCTs aligns with the distinct tumor biology of germ cell-derived malignancies.

Biomarkers play a critical role in TGCT management, including diagnosis, risk stratification, and surveillance. Traditional serum tumor markers such as AFP, β-hCG, and LDH, though clinically valuable, show low sensitivity, being elevated in only about half of patients, and often remain normal in early-stage disease or histologic subtypes such as seminoma and teratoma. Moreover, false-positive elevations may occur in benign or non-germ cell malignancies, limiting their specificity^8,9. Considering our findings, unlike these systemic markers that primarily reflect overall tumor burden, Talin1 provides tissue-level insight into adhesion and cytoskeletal regulation. Such mechanistic specificity suggests that Talin1 may complement existing serum biomarkers and serve as a more informative prognostic indicator of tumor aggressiveness and progression in TGCTs.

Importantly, as Talin1’s interaction with FAK and activation of downstream PI3K/AKT signaling have been implicated in cancer cell survival, migration, and EMT, future studies focusing on this signaling axis in TGCTs may uncover novel mechanisms of tumor progression and identify therapeutic vulnerabilities specific to Talin1-regulated pathways.

While previous studies have linked Talin1 expression to survival outcomes in multiple cancer types^20,48, our analysis in TGCTs did not reveal statistically significant differences in DSS or PFS between high and low expression groups. This outcome may primarily reflect the relatively limited follow-up period, which could have constrained the detection of long-term survival differences. The relatively low number of cancer-related deaths during a median follow-up of approximately five years likely limited the statistical power of our survival analyses, and this limitation should be considered when interpreting the prognostic value of Talin1. Extending the follow-up duration in future studies may provide a more definitive understanding of the prognostic significance of Talin1 in each subtype of TGCTs.

Given the limited cohort size and number of outcome events, the use of median-based cut-offs represents a pragmatic compromise; therefore, the present findings should be interpreted as exploratory and require validation in larger, independent cohorts.

Alternative cut-off strategies, including tertile-based stratification or sensitivity analyses, were considered. However, the available sample size and event distribution would likely render such analyses statistically unstable and potentially misleading; therefore, these approaches were not pursued in the present study.

A key limitation of the present study is the absence of functional validation to evaluate the mechanistic contribution of Talin1 in TGCT biology. While our bioinformatics and IHC analyses support its association with tumor aggressiveness, the absence of in vitro or in vivo experiments (e.g., siRNA knockdown or CRISPR/Cas9 gene editing) limits the ability to establish causality or therapeutic implications. Future studies should prioritize functional assays such as proliferation, apoptosis, and migration analyses to confirm the biological relevance of Talin1 in TGCT progression. These investigations could be facilitated through collaborative research networks providing access to TGCT-derived cell lines or patient-derived organoids. Alternatively, pluripotent embryonal carcinoma cell lines such as NCCIT and NTERA-2, which serve as well-established in vitro models of germ cell tumors, may provide practical systems to explore Talin1-mediated adhesion and signaling pathways.

Conclusions

This study is the first to demonstrate a link between reduced Talin1 protein levels and advanced pT stage across subtypes of TGCTs. Bioinformatics analyses initially identified Talin1 as a pivotal gene involved in cell migration and invasion pathways, highlighting its potential prognostic significance in TGCTs. Subsequent IHC validation confirmed that lower cytoplasmic Talin1 expression is directly associated with increased invasiveness and disease progression among different TGCT subtypes. However, Kaplan–Meier survival analysis did not show a statistically significant association between Talin1 expression and patient survival within each subtype, likely due to the small number of cancer-related deaths and the relatively short follow-up period. Future studies involving larger, well-characterized patient cohorts with extended follow-up are needed to further evaluate the prognostic relevance of Talin1 and confirm its clinical utility in TGCT management.

Methods

Bioinformatics study

The GEO database (https://www.ncbi.nlm.nih.gov/geo/) was searched to identify TGCT studies⁵². Additionally, the PRIDE database was queried for relevant datasets⁵³. Venn diagram analysis was then performed using Venny (https://bioinfogp.cnb.csic.es/tools/venny/index2.0.2.html)⁵⁴ to identify commonly expressed genes across the selected studies. Each GEO dataset was analyzed individually to identify differentially expressed genes using the normalized data and study-specific criteria. No cross-dataset normalization was performed due to differences in array platforms and preprocessing protocols. The resulting lists of genes were used for intersection analysis through Venn diagrams to identify commonly dysregulated genes across various datasets. Venn diagram analysis was conducted using Venny to identify overlapping genes among the differentially expressed gene lists obtained independently from each dataset. The common genes were analyzed further using Enrichr (amp.pharm.mssm.edu/Enrichr/) based on gene ontology (GO) terms (http://geneontology.org/)⁵⁵. GO analysis encompassed the categories of cellular component (CC), biological process (BP), and molecular function (MF). Moreover, a Protein-Protein Interaction (PPI) network was constructed to explore the connections between the identified genes, using the STRING database (https://string-db.org/).

Furthermore, several online tools were utilized to gain a deeper understanding of the genes within the network. Two proteomics datasets relevant to TGCT were selected based on the availability of protein expression profiles in TGCT models. The first dataset (PXD010275) contains proteomic data from TGCT cell lines, serving as an in vitro model of the disease. The second dataset (PXD013658) provides protein expression profiles from primary TGCT patient tissues and adjacent normal controls. These datasets were obtained from the supplementary information of previously published studies. Rather than performing a formal meta-analysis or normalization across datasets, we conducted an exploratory comparative integration to identify overlapping proteins reported in both in vitro and patient-derived TGCT samples. Common proteins were visualized using Venn diagram analysis via Venny (https://bioinfogp.cnb.csic.es/tools/venny/index2.0.2.html).

The GEPIA tool (http://gepia.cancerpku.cn/index.html) was utilized to analyze RNA sequencing expression data and assess the gene’s prognostic value. All samples in the GEPIA database were derived from the GTEx and TCGA projects. Finally, the mRNA expression levels of the selected genes in TGCTs were evaluated using the UALCAN database (http://ualcan.path.uab.edu/), which provides gene expression analysis based on TCGA transcriptomic data. All these enrichment analyses were performed to assess potential novel biomarkers in tissues obtained from patients with TGCT using the IHC method.

Patient’s characteristics and tumor samples

This study obtained 191 paraffin-embedded TGCT clinical specimens from Hasheminejad Urology-Nephrology Hospital, Tehran, Iran, between 2008 and 2011. No patients had undergone chemotherapy or radiotherapy before surgery. The study included four subgroups of TGCTs: seminoma, embryonal carcinoma, yolk sac tumor, and teratoma. Some specimens were pure, while others contained histopathological subtypes. Hematoxylin and eosin-stained tissue sections and archived clinical records were reviewed to collect clinical and pathological characteristics, including age, tumor size (maximum diameter), pT stage, venous invasion, GCNIS, hilum involvement, rete testis involvement, epididymis involvement, tumor-infiltrating lymphocytes, tunica vaginalis and tunica albuginea invasion, spermatic cord and spermatic cord margin invasion, scar, tumor necrosis, serum tumor marker levels, distant metastasis, and tumor recurrence. Twenty adjacent normal tissues for each subtype were also included to compare Talin1 expression with cancerous samples. DSS was calculated in survival analyses based on the time between radical orchiectomy and cancer-related death. In contrast, PFS was assessed based on the duration between the primary surgery and the final follow-up survey for cases free of symptoms, metastasis, or recurrence. Additionally, pathological grading was determined using the pTNM classification system released by the American Joint Committee on Cancer (AJCC) in 2018.

TMA construction

TMAs of testicular cancer specimens were prepared according to our previous studies^26,56,57. Briefly, a pathologist reviewed H&E-stained sections to identify three representative areas in morphologically distinct regions of each block. The Tissue Arrayer Minicore was used to punch out the selected areas (0.6 mm diameter) and deposit them into a recipient paraffin block (ALPHELYS, Plaisir, France). Given the heterogeneity of tissue samples, all specimens were analyzed at three distinct points and multiple areas to ensure more accurate and reliable results. Finally, the mean expression from each point was considered the result for each sample.

IHC staining

Talin1 expression was assessed using our previously published methodology^25–28. To block native peroxidase activity, TMA sections were deparaffinized (Dako Glostrup, Denmark), rehydrated, washed, and incubated with 3% H2O2 for 20 min. The TMA sections were then rinsed three times with Tris-Buffered Saline (TBS) and autoclaved for 10 min in Tris-EDTA Buffer (pH 6) for antigen retrieval. Afterward, the sections were washed in TBS and treated with 5% sheep serum dissolved in a blocking protein solution (Dako, Denmark) for 20 min. Next, the sections were incubated overnight at 4 °C with a 1:100 dilution of the primary antibody (anti-Talin1 antibody, ab71333, Abcam, USA). Rabbit immunoglobulin IgG (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) was used as an isotype control at a 1:100 dilution. The TMA slides were then washed three times in TBS and incubated for 40 min with a secondary antibody (Medaysis USA’s Polymer-HRP Anti-Mouse & Rabbit System). Following this, the specimens were rinsed and exposed to the chromogen substrate 3,3′-diaminobenzidine (DAB) (Dako, Denmark) for 20 min. Finally, the sections were cleaned, stained with hematoxylin (Dako, Denmark), dehydrated, cleared with xylene, and mounted for examination.

Evaluation of immunostaining

The pathologist, M.R., assessed the expression levels of Talin1 without any prior knowledge of the pathological data, using a semi-quantitative scoring method. A consensus was reached for the entire section. Staining intensity was graded as follows:

0 indicates no staining; 1 indicates weak staining; 2 indicates moderate staining; and 3 indicates strong staining. The proportion of immunoreactive cells was measured on a scale from 0 to 100%. Positive cancer cells were classified in the following manner: 1–25% as Group 1, 26–50% as Group 2, 51–75% as Group 3, and more than 75% as Group 4. To calculate the histochemical (H) score, the intensity grades were multiplied by the proportion of positive cancer cells, yielding a value between 0 and 300.

Statistical analysis

The analyses were performed using IBM Corp.‘s SPSS software version 22.0. Categorical variables were represented as frequency (N) and percentage, while continuous variables were described using mean (SD) and median (Q1, Q3). The validity of connections and correlations between Talin1 expression and clinicopathological factors was assessed using Pearson’s chi-squared test and Spearman’s correlation coefficient. Paired comparisons across categories were evaluated with the Kruskal-Wallis and Mann-Whitney U tests. DSS and PFS curves were created using the Kaplan-Meier method. Comparisons of the survival curves were performed using the log-rank test, with 95% confidence intervals (CI). A p-value of < 0.05 was considered statistically significant for all analyses. Furthermore, normality tests (Kolmogorov–Smirnov and Shapiro–Wilk) were conducted to determine the appropriate cut-offs for data classification for each subtype. As previously mentioned, all quantitative data were obtained from three independent replicates.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, methodology, and supervision, Z.M., L.SZ.; Formal analysis & Data curation, L.SZ., S.V.; Investigation, A.Y., M.R.; Visualization, S.S., L.SZ.; Writing original draft preparation, L.SZ., M.R., Validation, R.GH. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by a grant from the Oncopathology Research Center, Iran University of Medical Sciences, Tehran, Iran (Grant No. 99-3-28-19735). The authors disclose receipt of this financial support for the conduct of the research.

Data availability

The bioinformatic data supporting the findings of this study have been included in the main text and Supplementary Tables 1 and 2. Raw data analyzed using SPSS are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

This study was approved by the Ethics Committee of the Iran University of Medical Sciences Human Research, Iran (Ref no: IR.IUMS.REC.1399.1189). All procedures performed in this study were conducted in accordance with the 1964 Helsinki Declaration and its subsequent amendments.

Informed consent

Informed consent was obtained from all subjects involved in the study.