Immunohistochemical expression of GATA3, CK5/6 and CK20 in molecular subtypes of bladder carcinoma: correlation with clinicopathological features

Abstract

Background

Bladder urothelial carcinoma, is considered the 7th most common cancer in males. It is classified into luminal and basal subtypes depending on molecular markers, influencing prognosis and treatment. Identifying reliable biomarkers like GATA3, CK20, and CK5/6 through immunohistochemical methods can aid in early detection, risk stratification, and personalized treatment strategies. This study aims for evaluation prognostic role of these mentioned markers in correlation with clinicopathological parameters in urothelial carcinomas.

Methods

Tumor samples of forty cases were immunohistochemically stained for GATA3, CK5/6, and CK20. A cutoff of 20% positivity was used to determine subtype classifications, with staining patterns guiding the categorization into basal, luminal, double positive, or double negative groups.

Results

In this study of 40 urothelial carcinoma patients tumors were classified into basal and luminal subtypes using GATA3, CK5/6 and CK20 markers. GATA3 expression showed no significant association with clinicopathological parameters; while, CK20 was associated with tumor size, and CK5/6 with T, N classification, and lymphovascular invasion. Significant differences in clinicopathological parameters were observed when subtypes were defined by CK5/6, GATA3 or CK20, particularly in tumor grade, T and N classification, and gender. Basal molecular subtypes was correlated with poor prognostic parameters.

Conclusions

This study documented that use of triple markers could define the luminal and basal subtypes of urothelial carcinoma. Basal tumors have shown to be associated with the aggressive behavior and future studies may allow the development of new therapies in the context of molecular subtypes.

Background

Bladder urothelial carcinoma ranks the tenth most frequent cancer world-wide overall the both genders (Available online 2023; Mukherjee et al. 2024). In Egypt, urinary bladder cancer was the most frequent cancer in urinary tract as about 14.2% of male’s malignancies of urothelial origin according to Global Cancer Observatory (Ferlay et al. 2024). About 25% of urothelial carcinomas were muscle-invasive when initially diagnosed (Smith et al. 2014) while about 75% of patients present with non-invasive disease that could be cured through non-invasive therapy. However, as much as 20% of non-invasive cases proceed to muscle-invasive carcinoma that could be treated with local intervention as surgery or radiotherapy as well as systemic treatment as immunotherapy or chemotherapy. Difference in prognosis depending on cancer extent and in turn influences therapeutics options, with muscle-invasive carcinomas is potentially life-threating with median survival of about 48 months (Babjuk et al. 2022; Jazvi´c 2019; Witjes et al. 2020, 2021, 2022; WHO 2022).

Staging is clinically determined by trans-urothelial resection (TURB), then by histopathological examination for proper assessment of tumor invasion, accurate staging and grading, hence histopathology has predicting prognosis (Agarwal et al. 2019). However, it is subjective to individual variability and may have poor specify. So, the identification of dependable biomarkers for invasive carcinoma may be important for early detection and patients risk classification which in turn could help in selecting personalized treatment strategies and predicting prognosis (Tamalunas et al. 2020).

According to “Bladder Cancer Molecular Taxonomy Group,” molecular classification of muscle-invasive bladder carcinoma can be categorized to two main groups, luminal and basal with difference in biological and histological patterns and clinical manifestation (Robertson et al. 2017; Kamoun et al. 2020). It has been reported as immunohistochemical antibodies are useful indicators for both luminal and basal tumors. Luminal bladder carcinomas express markers of terminal differentiation as CK20, GATA3 and uroplakins; whereas, basal carcinomas can express basal types cytokeratin like CK5, CK6 and CK14 which act as markers of basal urothelial cells progenitor /stem cells (Sanguedolce et al. 2022a, b). These attempts for molecular sub-typing of muscle-invasive carcinomas based on transcriptional analysis and use immunohistochemical method is important to translate research findings into routine clinical practice.

In this study we aimed to use immunohistochemical markers GATA3, CK20, CK5/6 to define basal and luminal subtypes in cases of urothelial carcinoma. Their expression will be correlated with patient clinicopathological parameters in attempt to explore their prognostic role.

Methods

The specimens of our current study were collected randomly for cystectomy specimens of urothelial carcinoma (UC) cases from The Tissue Research and Image Analysis Lab, the Medical Research Center of Excellence in the National Research Center, and the archives of private centers during the period from January 2021 to September 2023. The study was approved by The Ethical Committee of the National Research Center (Approval No. 01430624).

Forty cases were included in our current study, each case was stained by routine H&E stained slides for evaluation of diagnosis, grading according to “World Health Organization classification of tumors 2022” (Sanguedolce et al. 2022b) and staging according to “American Joint Committee on Cancer 2020” (Olawaiye et al. 2021

In immunohistochemical staining, three 5µ thick sections from each formalin fixed paraffin-embedded blocks of all specimens. Sections were mounted on positive charged glass slides, then incubated at 37 °C overnight for accurate adhesion of the section of the slide. Deparaffinization, rehydration then epitope retrieval were performed in Pt link retrieval system (Dako, Copenhagen, Denmark). Staining steps and incubation times were pre-programmed into auto-stainer Link software (Dakoomnis, Copenhagen, Denmark). Endogenous peroxidase was blocked with 3% H₂O₂ in methanol. Non-specific binding was blocked for 10 min using protein-blocking buffer. Sections were washed in phosphate-buffered saline (PBS). Diluted primary antibodies against GATA3 (Biocare, cat#CN405,1:200,USA), CK 5/6 (Biocare,cat#CM105,1:200,USA) and CK20 (Labome,cat#M7019,1:200,USA) were added to tissue and incubated overnight at 4ºC. For negative control; primary antibody was replaced by normal mouse serum. Then, Samples were incubated with horseradish peroxidase labeled secondary antibody (Thermo Fisher Scientific Tudor Road, Manor Park, Runcorn, CheshireWA7 1TA, UK) for 30 min at room temperature. Diaminobenzidine was used for color development and hematoxylin as counterstain.

Evaluation of immunohistochemistry

Both of CK5/6 and CK20 cytoplasmic/membrane expression and of GATA3 nuclear expression were inspected visually under microscope by two pathologists. As recommended by previous study (Dadhania et al. 2016 Cut-off value of 20% tumor cells cytoplasmic/membrane (for CK5/6 and CK20) or nuclear (for GATA3) positivity was employed; if expression of the target protein in IHC was less than or equal to 20% samples were considered negative and if greater than 20% were considered positive. Patients were grouped in basal subtype when CK5/6 positive and CK20 negative; or CK5/6 positive and GATA3 negative. While Patients who were CK5/6 negative and CK20 positive; or CK5/6 negative and GATA3 positive, were grouped in luminal subtype (Warrick et al. 2017) Patients who were positive or negative for both marker sets were grouped as double positive or double negative subtypes (Ravanini et al. 2021).

Statistical methods

Using the Windows version of SPSS 20.0, statistical analysis was carried out (New York, IBM, USA). In order to examine association between protein expression and clinicopathological indices,

Fisher's exactor Pearson Chi-square tests were utilized. Using these tests perform a univariate study comparing lymph node metastasis, Lymphovascular invasion, and other clinicopathological indices was carried out. For analysis, beta coefficients and 95% confidence intervals (CI) were utilized. P value was regarded as statistically significant, when its value is ≤ 0.05.

Results

In this study, 40urothelialcarcinoma patients were evaluated. Most patients were males representing 82.5% of all cases. The mean of age was 67 ± 4.7 years with range 45–77 years, 10 patients (25%) were < 60 years and the remaining 30 patients (75%) were ≥ 60 years. Based on pathological report of surgical specimens, 26 cases were larger than 3 cm in tumor size. From the forty cases included in this study, sixteen (40%) cases were classified as UC of low-grade; while, 24 (60%) cases were diagnosed as high-grade UC. Most cases (70%) were T3 and T4. Regional lymph nodes were involved in 10 (25%) cases and lymphovascular invasions were detected in 12 (30%) cases (Table 1).

Table 1 Clinicopathological features of urothelial carcinoma patients

Fifty-five % of cases were positively stained for GATA3, CK20 was positive in 50% of cases; while, 45% of cases were CK5/6 positive (Table 2, Fig. 1).

Table 2 Immunohistochemical Expression of GATA3, CK20 and CK5/6 in Urothelial Carcinoma

Fig. 1

A Strong cytoplasmic & membranous expression CK5/6 of moderately differentiated UC, B Strong cytoplasmic & membranous expression CK20 of well differentiated UC, C Moderate nuclear expression GATA3 of well differentiated UC (Immunohistochemistry X200)

When immunohistochemical subtypes were defined by the 3 markers, they were categorized as luminal, basal, double positive and double negative. By using GAT3, sixteen cases were classified as luminal where they were GATA3 + ve and CK5/6 –ve (Fig. 2A, C), 6 cases were double positive and 6 cases were double negative for both markers (GAT3 and CK 5/6). By Using CK20, 12 cases only were classified as luminal where they were positive for CK20 and negative four CK5/6 (Fig. 2B, C), 8 cases were positive and 10 cases were negative for two markers (CK20 and CK5/6). On the contrary, the cases were categorized as basal when they were CK5/6 positive and negative for both luminal markers GATA3 or CK20 (Table 3, Fig. 3).

Fig. 2

A case of muscle-invasive urothelial carcinoma, Luminal molecular subtype showing; high grade moderately differentiated revealed A Strong nuclear expression of GATA3, B CK5/6 negative expression and C Moderate cytoplasmic and membranous expression of CK20 (Immunohistochemistry X200)

Table 3 Immunoscore of Luminal, Basal, Double Positive and Double Negative

Fig. 3

A case of muscle-invasive urothelial carcinoma, basal molecular subtype; high-grade poorly differentiated revealed showing A GATA3 negative expression, B CK20 negative expression and C Strong cytoplasmic and membranous expression of CK5/6 (Immunohistochemistry X200)

On correlating immunohistochemical expression of the three markers and clinicopathological patients’ parameters with urothelial carcinoma, It was found that GATA3 expression was not significantly associated with any of the parameters. Expression of CK20 was significantly associated with only tumor size; while, CK5/6 was associated with T, N classification and lymphovascular invasion of urothelial carcinoma (Table 4).

Table 4 Correlation between GATA3, CK20, CK5/6 and clinicopathological parameters in Urothelial Carcinoma

When IHC subtypes were defined by CK5/6 and CK20, only tumor grades and T classification were significantly different between the four groups (Table 5). On the contrary, when the IHC subtypes were defined by CK5 and GATA3, there were significant difference between the four groups as regard gender and tumor grade also T and N classification and lymphovascular invasion showed significant difference (Table 6).

Table 5 Association between clinicopathological features of urothelial carcinoma patients and molecular subtypes in relation to CK20 and CK5/6

Table 6 Association between Clinicopathological features of Urothelial Carcinoma patients and molecular subtypes in relation to GATA3 and CK5/6

In our study, we focused in the main two molecular subtypes of urothelial carcinoma; Basal and luminal types and correlated them with the clinicopathological parameters in (Tables 7, 8).

Table 7 Association between Clinicopathological features of Urothelial Carcinoma patients and Basal, Luminal subtypes in relation to GATA3 and CK5/6

Table 8 Association between Clinicopathological features of Urothelial carcinoma patients and Basal, Luminal subtypes in relation to CK20 and CK5/6

When Basal and luminal subtypes were defined by GATA3 and CK5/6, Gender, T and N classification were significantly different between both groups (Table 7).

When luminal and basal subtypes were defined by CK5/6 and CK20, there were significant difference of both groups as regard tumor grade, T and N classification. (Table 8).

Discussion

Bladder urothelial carcinoma is considered the 7th most commonly diagnosed male tumor; with about 25% of patients have muscle invasion at the time of initial diagnosis (Ravanini et al. 2021). Grading of bladder Carcinoma is an important prognostic factor and the treatment regimen based on both tumor stage and grade (Weyerer et al. 2017; Ying et al. 2023While radical cystectomy is considered the standard treatment for muscle-invasive tumors, adjuvant chemotherapy and immunotherapy have provided opportunities to improve the patients’ survival (Khalife et al. 2021 Using clinicopathological parameters for identifying high risk group patients may be insufficient and those have the same histopathological features may belong to distinct molecular subtypes with different prognosis (Dyrskjøt et al. 2017; Wu et al. 2021).

The luminal/basal of molecular subtypes of bladder urothelial carcinoma has been reported to predict prognosis of the tumor. This molecular sub-typing of muscle-invasive cases with high-cost of sequencing, limits its clinical application (Ying et al. 2023). Our present work, three immunohistochemical markers including GATA 3, CK 20 as luminal markers, and Ck5/6 as basal marker were used to identify molecular subtypes. Their expression was correlated with clinicopathological parameters to evaluate their diagnostic and Prognostic implication in urothelial carcinoma.

In our present study, 40 urothelial carcinomas were included. Studied cases mean age was 67 years and range 45–77 years with a male predominance (82.5%). Most of the studied cases were high histologic grade UC (60%). Seventy percent of cases were stage T3 and T4. Regional lymph nodes were involved in 25% of cases; while, lymphovascular invasions were detected in 30% of cases. These results were comparable with the study of Helal et al., (2023 which showed that median age of 60 studied cases was 66.5 years with strong male predominance. Most of these cases were stage T3 (41.7%), N2 (39.3%), and all cases was of high grade. Almost similar results were shown both of Ying et al., (2023) and Terlevi´c et al. study (Terlevi´c et al. 2023).

In the current study GATA 3 and Ck20, the markers for identifying luminal subtype were evaluated for their expression in urothelial carcinoma. GATA3 was expressed in 22 cases (55%); while, CK20 was positive in 20 cases (50%) which was in agreement with the previous studies. By using both markers with CK5/6 for sub-typing of studied cases, they were classified as luminal when they were positive for GATA3 and CK 20 and negative for CK 5/6. On the contrary they were categorized as basal when they were positive for CK5/6 and negative for both luminal markers GATA3 and CK20.

On correlating GATA3 expression with the clinicopathological parameters, it was found no significand association with any of the parameters. Several studies came in line with our study, However, a previous study showed that GATA3 high expression was associated with larger size (Mohammed et al. 2016). Regarding histological grade and stage, no significant association was of found in our study. However, in previous study of Agarwal et al., (2019) there were statistically significant correlation between expression of GATA3 and histological grade (p < 0.001) as all cases (100%) of low-grade tumors were immunoreactive for GATA3; while, 70% of high-grade tumors were negative. In the study of Hoang et al., (2015) they showed no correlation between grade of the tumor and GATA3 expression; while, El kholy et al., (2021) showed significant correlation between expression of GATA3 and histological stage and grade.

On correlating expression of CK20 as one of the markers identifying luminal subtype with the clinicopathological parameters, it showed only significant association with tumor size. On the other hand, CK5/6 which identified as basal marker was found to be significantly correlated with T, N classification and lymphovascular invasion of our cases of urothelial carcinoma. Results were in accordance with previous report of Hashmi et al., (2018) which noticed CK 5/6 low expression in urothelial carcinomas and its expression is correlated with higher tumor grade and muscularis mucosa invasion. Parallel to that, the study of El kholy et al., (2021) reported that all cases of non- invasive carcinomas (86.7%) were negative for CK5/6. Our study suggested that CK5/6 as a basal marker was correlated with bad prognostic parameters, and most basal subtypes could be identified as aggressive tumors with patients at high risk of progression.

In the present study we correlate clinicopathological features and all molecular subtypes in relation to CK20 and CK5/6, only tumor grade and T classification were significant between the studied four groups including double positive and double negative cases. On the other hand in relation to GATA3 and CK5/6 significant difference were found between the four studied groups and clinicopathological parameters as regard gender, tumor grade, T classification, N classification and lymphovascular invasion.

On correlating the features of patient in our study with the main molecular subtypes (basal and luminal) subtypes in relation to GATA3 and CK5/6, it was found that female gender was significantly different between both subtypes where all cases showed positive expression of CK 5/6 was female while only 50% of males showed CK5/6 expression. Significant difference was also found between luminal and basal subtypes as regard T and N classification. On the other hand when luminal and basal subtype were defined by CK20 and CK5/6, a significant difference were found between the subtypes as regard tumor grade, T and N classification.

Our Finding regarding the association between basal subtype and female gender on using GATA3/ CK 5/6 is similar to the result of Ravonini et al., (2021) that also showed a greater percentage of female patients who highly expressed basal marker CK5/6. They suggested that females are more likely to develop bladder urothelial carcinoma with keratinized squamous phenotype associated with poor prognosis.

On a previous result of Langner et al., at 2004 (2004) showed that advanced cases of urothelial carcinoma can show independently differentiation including squamous differentiation in about 50% of cases of U.C, and it is associated with poor progression. They suggested that prognostic value of Keratin sub-typing in U.C may be associated with poor survival and were more sensitive to chemotherapy and immunotherapy. Jackson et al., (2022) studied 681 patients which were classified into basal and luminal subtypes using IHC. They validate diagnostic and prognostic significance of molecular subtypes, and reported that usage of different IHC markers for sub-typing urothelial carcinoma in clinical practice may help to explain diagnostic problems and improve therapeutic options for patients. Although identifying the molecular subtypes of U.C with gene expression analysis by sequencing is ideal, it is not economically feasible for routine clinical diagnosis, In the contrary IHC markers could permit cost effective and simple classification (Hodgson et al. 2018; Jangir et al. 2019; Rodriguez Pena et al. 2019; Al-Ahmadie and Netto 2020.

Conclusions

In conclusion, our study confirmed that luminal and based molecular subtypes of urothelial carcinoma, assessed by triple markers GATA3, CK5/6 and CK20 that proved to be a practical clinical approach to diagnostic prediction. Unlike luminal subtype, basal tumors were aggressive and significantly correlated with poor prognostic parameters. Our work paves the way for future studies with large numbers of patients using IHC to clarity diagnostic dilemmas and develop treatment strategies in the context of molecular subtypes.