Introduction

Urothelial carcinoma is the most common morphologic type of bladder cancer. Chromosome 9p21 is the region which is most commonly altered in the onco-pathogenesis of bladder cancer. p16 is the most important gene which is either deleted or mutated in this process. p16 is a tumor suppressor gene; therefore loss of its function results in abnormal cell proliferation leading to cancer development (Cairns et al., 1995; Williamson et al., 1995). Prognostic parameters of bladder cancer include histologic differentiation, grade, lamina propria invasion and deep muscle invasion. Various prognostic biomarkers have been studied in bladder cancer to help in prognostic stratification of patients (Hashmi et al., 2018a; Hashmi et al., 2018b). These biomarkers include epidermal growth factor receptor (EGFR), p53 and retinoblastoma (Rb) (Sarkis et al., 1993; Cordon-Cardo et al., 1992; Mumtaz et al., 2014). p16 can be lost or abnormally expressed in bladder cancer as a result of deletion or mutations respectively; however the prognostic significance of this biomarker is still unclear. Moreover, p16 mutations or protein expression patterns have not been studied in our population in bladder cancer; therefore in the present study, we aimed to investigate the expression pattern of p16 in bladder cancer and to evaluate the prognostic significance of this biomarker in our setup.

Methods

We retrospectively evaluated 121 diagnosed cases of urothelial carcinomas from January 2010 till December 2014 over a period of 5 years from records of pathology department archives. All patients had undergone elective surgeries at Liaquat National hospital, Karachi. The study was approved by research and ethical review committee of Liaquat National Hospital. Informed written consent was taken from all patients antecedent to surgery. Specimens included 108 (89.3%) transuretheral resections (TUR) and 13 (10.7%) radical cystectomies specimens. Slides of all cases were retrieved and reviewed by senior two histopathologists to evaluate pathologic characteristics. Clinical records of 54 patients were available and were thus evaluated from institutional records to determine recurrence and survival status of the patients. Moreover, representative tissue blocks were selected for p16 immunohistochemistry. For TUR specimens, if the specimen is entirely submitted in 2–3 cassettes then, IHC was applied on all tissue blocks, whereas if the tissue is submitted in more than 3 cassettes then 3 representative tissue blocks were selected for IHC. For radical cystectomy specimens, IHC was applied on 3 representative tissue blocks.

p16 antibody was purchased from Roche Ventana and IHC was performed using antibody CINtec R p16INK4a, clone E6H4™ according to manufacturers protocol. Tonsils and carcinoma cervix was used as positive controls. Both nuclear and cytoplasmic staining was evaluated. Intensity of staining was scored into no staining (0), weak (1+), intermediate (2+), strong (3+) while percentage of positively stained cells were scored from 0 to 100. Intensity and percentage scores were multiplied to calculate a total H-score. An H-score of less than 200 was considered as low p16 expression while an H-score of more than 200 was taken as high p16 expression i.e., moderate staining (2+) in 100% tumor cells (2 × 100 = 200) or strong staining (3+) in more than > 67% tumor cells (3 × 67 = 201) (Fig. 1).

Fig. 1
figure 1

p16 expression in Urothelial carcinoma. a Complete loss of p16 expression, H-score 0; (b) Low p16 expression, weak expression (1+) in 10% of tumor cells, H-score 10 (1 × 10), (c) Low p16 expression, Intermediate expression (2+) in 25% of tumor cells, H-score 50 (2 × 25); (d) High expression p16 expression, strong expression (3+) in 95% of tumor cells, H-score 285 (3 × 95). Magnification = 400×

Full size image

Recurrence status and follow-up were determined by evaluating hospital records. Overall survival was defined as time from surgical excision till death or last follow-up.

Statistical package for social sciences (SPSS 21) was used for data compilation and analysis. Mean and standard deviation were calculated for quantitative variables. Frequency and percentage were calculated for qualitative variables. Independent t-test and ANOVA were used to compare mean difference. Chi-square and fisher exact were applied to determine association. Significant variables were included in univariate and multivariate binary logistic regression analysis. Survival curves were plotted using Kaplan-Meier method and the significance of difference between survival curves were determined using log-rank test. P-value of ≤0.05 was taken as significant.

Results

Mean age of patients was 63.43 + 14.88 years with male to female ratio of 2.66:1. 52.1% (63 cases) were of high grade morphology, whereas 47.9% (58 cases) showed low grade histology. All cases were those of papillary urothelial carcinoma. Out of total 121 cases, 33 (27.3%) cases were those of invasive urothelial carcinoma, while 88 (72.7%) cases were of non-invasive urothelial carcinoma. There was no case of isolated carcinoma insitu (CIN). Lamina propria invasion was seen in 27.3% (33 cases), while muscularis propria invasion was noted in 17.4% (21 cases). Mean follow up of patients involved in the study was 23.11 + 13.574 months and recurrence was seen in 42.6% (23 cases) as shown in Table 1. Among 54 cases in which follow-up were available, 16.7% (9 cases) died of disease.

Table 1 Association of p16 expression with Clinicopathologic features of Urothelial carcinoma
Full size table

Low expression of p16 was noted in 86% (104 cases), whereas 14% (17 cases) revealed high p16 expression as shown in Fig. 2. Significant association of p16 expression with clinicopathologic parameters was noted. We found significant association of p16 expression with tumor grade (p = 0.000), muscularis propria invasion (p = 0.001), lamina propria invasion (p = 0.001) and survival status (p = 0.020). Univariate binary logistics showed that low grade tumors were less likely to express high p16 expression as compared to high grade tumors. Similarly, patients with lamina propria and muscularis propria invasion were more likely to exhibit high p16 expression. Detailed results of univariate and multivariate binary logistic regression are presented in Table 2. Significant association of high p16 expression was noted with worse long term survival (Table 1 and Fig. 3). In our study, most of cases were of conventional histology, there were only two cases of micropapillary variant, one case showing microcystic pattern, while there was one case each with squamoid and glandular differentiation. Among these cases, only one case with micropapillary and one case with squamoid differentiation showed positive p16 expression. Cases with variant histology were not sufficient enough to get evaluated for statistical correlation.

Fig. 2
figure 2

p16 expression among studied population

Full size image
Table 2 Odds ratio for patients with high expression of p16 expression
Full size table
Fig. 3
figure 3

Kalpien–Meier for Survival status

Full size image

We found significant association between survival status of patients and p16 expression (p = 0.020) as shown in Table 3 while univariate logistic regression showed that patients with low p16 expression were at low risk (HR = 0.194) to die of disease as compared to patients with high p16 expression (Table 4).

Table 3 Association of Clinicopathologic features with survival status of the patients
Full size table
Table 4 Hazard Ratio for patients died of disease
Full size table

Table 5 shows mean H-scores and cases with complete loss of p16 expression and its comparison with clinico-pathologic features. Overall 33 cases (27.3%) showed complete loss of p16 expression and mean H-score was 64.61 ± 93.13.

Table 5 Mean p16 expression (mean H-score) and complete loss of p16 expression in urothelial carcinoma
Full size table

We also compared p16 expression in urothelial carcinomas with 25 cases of normal urothelium. 44% of normal urothelium showed complete loss of p16 expression while 56% showed low p16 expression. None of the cases revealed high p16 expression as shown in Figs. 4 and 5.

Fig. 4
figure 4

Comparison of p16 expression between normal and malignant specimens

Full size image
Fig. 5
figure 5

p16 expression in normal urothelium, (a) Complete loss of p16 expression, H-score 0; (b) Low p16 expression in normal urothelium, Intermediate expression (2+) in 10% of tumor cells, H-score 20 (2 × 10)

Full size image

Discussion

In the current study, we found that high expression of p16 to be associated with higher grade and poor prognostic factors including muscularis propria invasion and poor long term survival.

p16 expression in bladder cancer has been studied previously (Kruger et al., 2005; Bartoletti et al., 2007; Yang et al., 2002); however, its prognostic significance is still unclear. In a meta-analysis of 37 studies including 2246 patients revealed that downregulated p16 expression was linked to poor prognosis in bladder cancer (Gan et al., 2016). In most of the studies analyzed, IHC markers were used to evaluate p16 expression and marked heterogeneity exists in the definition of abnormal IHC expression of p16. p16 is a tumor suppressor gene, which is most commonly involved in bladder carcinogenesis. Mutations involving p16 gene can be in the form of deletions of chromosome 9, in that case there would be complete loss of IHC expression of p16 protein. Conversely, inactivated or mutated gene product is generally not easily digestible leading to abnormally high expression as it is noted in the case of p53 (another tumor suppressor gene product) which is known to be overexpressed (immunohistochemically) in endometrial and ovarian serous carcinoma. Many studies evaluating p16 IHC expression didn’t take into account this important fact and therefore didn’t categorize p16 expression on the basis of intensity and percentage. Significant association of high p16 expression by IHC with poor prognostic parameters of bladder cancer in our study may be due to aberrant protein expression as a result of underlying gene mutations.

Yin M et al., found that strong p16 expression was seen in 100% of cases of urothelial carcinoma insitu (CIS) that can be used in differentiating CIS from adjacent normal epithelium showing normal or focal loss of p16 expression (Yin et al., 2008). Similarly, Raspollini MR et al., found a statistically significant association of p16 expression with disease stage (Raspollini et al., 2006). They found strong p16 expression in 28.2% cases of urothelial carcinoma. Moreover, they reported a statistically significant association of p16 expression with disease stage, however no significant association was noted with tumor grade or disease progression. On the contrary, we found a significant association of p16 expression with recurrence status and tumor grade. p16 expression is considered as a surrogate marker of HPV infection in oral and cervical squamous cell carcinoma, however studies revealed that no such association was found in case of bladder cancer (Alexander et al., 2012).

One of the major limitations of our study was that molecular studies were not performed to evaluate patterns of mutations in p16 gene; therefore, we suggest that correlation of abnormal p16 expression with underlying gene mutations should be performed in our setup for better understanding of disease pathogenesis in our population.

Conclusion

p16 is an important biomarker in bladder cancer as it can be used for prognostic stratification of patients with bladder cancer. Moreover, we suggest that molecular studies should be performed in our population in order to correlate abnormal p16 expression with underlying gene mutations.