Abstract
Background/Aim: Nuclear factor I (NFI) A and NFIB are transcription factors involved in the regulation of cell differentiation and organ development. More recently, they have been implicated in the pathogenesis of cancer, acting as context-dependent tumor promoters or suppressors. Materials and Methods: Expression of NFIA and NFIB was assessed by immunohistochemistry in 136 primary urothelial bladder cancers. Results: Progressive down-regulation of NFIA was observed with increasing pT stages and higher grade of analyzed tumors. Consequently, muscle invasive cancers exhibited lower NFIA expression compared with non-muscle invasive cases. Analogous comparisons yielded negative results in the case of NFIB. Expression of neither protein was associated with patient survival. Conclusion: NFIA may act as a suppressor of urothelial carcinogenesis, but functional studies and understanding of post-transcriptional regulation of NFI expression is necessary to dissect its role in bladder malignancies.
Bladder cancer is among the most prevalent types of cancer, ranked the 5th and the 6th most common malignancy in Europe and the US, respectively (1). Superficial tumors that do not invade the muscularis propria of the bladder wall [non-muscle invasive bladder cancers (NMIBCs)] are typically treated by transurethral resection of bladder tumor (TURBT) (2). Despite the overall recurrence rate for such tumors is high, the majority represent low grade neoplasms that only rarely transform into more aggressive entities with metastatic potential (3). On the other hand, muscle invasive bladder cancers (MIBCs), which may arise de novo or by progression of selected NMIBCs, pose a substantial threat to patients and therefore, require more aggressive treatment that typically involves radical cystectomy (2, 4). However, even after a successful surgery, approximately 50% of patients suffer a recurrence and succumb to the disease within 5 years (4). Recent years have brought new hope related to the advent of immunotherapy, but only some patients benefit from this form of treatment and the overall survival rates remain unsatisfactory (2). While mechanisms that determine a durable remission are only partially understood, there is a need for new prognostic factors that could help select patients at high risk of bladder cancer recurrence.
Members of the nuclear factor I (NFI) family have been recently proposed as markers of tumor behavior and potential novel therapeutic targets (5-7). The family consists of four proteins, namely NFIA, NFIB, NFIC, and NFIX, that form homo- and heterodimers interacting with DNA in a site-specific manner (8). Years of research have shown their indispensable regulatory function during differentiation of cells and development of tissues and organs (8). More recently, NFI genes and their products have been demonstrated as promoters or suppressors of various cancers (5-7). NFIA and NFIB remain the most extensively investigated members of the family. Several studies on glial tumors provided evidence for an oncogenic function of NFIA, mediated by down-regulation of p53, p21 and activation of MAPK pathway (9-11). On the contrary, NFIB exerts tumor suppressor activity in glioblastoma and lung adenocarcinoma, while promoting growth and progression of colorectal, breast or small cell lung cancers (12-15). Thus, previous research has demonstrated a context-dependent functional engagement of NFIA and NFIB in the pathogenesis of selected tumors and a therapeutic potential of modulating these transcription factors. In bladder cancer, mRNA levels of NFIA and NFIB are down-regulated in either superficial and infiltrating tumors as compared to the healthy mucosa (5). According to the Kaplan–Meier plotter RNA-seq dataset, high expression of NFIA, but not NFIB, predicts unfavorable prognosis in bladder cancer patents (16). This and other high-throughput mRNA databases suggest that NFIA and NFIB may be involved in urothelial carcinogenesis (5). However, neither the putative prognostic utility of expression of NFIs nor other potential clinicopathological associations have been tested in bladder cancer on the protein level. To fill this gap, we analyzed immunohistochemical reactivity for NFIA and NFIB in a group of variably advanced bladder carcinomas and investigated its relationship with other clinicopathological characteristics.
Materials and Methods
Patient samples and characteristics. Archival tissue samples of muscle-invasive tumors were obtained from 111 consecutive patients who underwent radical cystectomy for bladder cancer in Wroclaw Medical University Hospital between 2011 and 2014. Seven of these patients underwent known neoadjuvant chemotherapy. Clinicopathological parameters retrieved and analyzed in this study included patient age, sex, TNM classification, information about concurrent prostate cancer in postcystectomy specimens, tumor grading according to WHO 1973 and 2004 classifications (17, 18), presence of lymphovascular invasion and area of tumor necrosis. Histologically, all cases were diagnosed as urothelial carcinomas, although divergent differentiation and variant morphology were not infrequent. Follow-up data regarding overall survival were available for 97 patients from the cystectomy group and information about disease recurrence was available in 88 cases. Additionally, early-stage cases of bladder cancer were included in some analyses. To this end, a random group of 25 transurethral resection-derived specimens of NMIBCs was selected from archives. The study was carried out in accordance with the Declaration of Helsinki and approved by the Bioethics Committee of Wroclaw Medical University (Approval #: KB-580/2019).
Immunohistochemical reaction and evaluation of staining. Paraffin blocks with tissues of primary bladder tumors were cut with a microtome to prepare 4 μm-thick sections, subsequently mounted on silanized slides (Agilent DAKO, Santa Clara, CA, USA). Automated dewaxing, rehydration and heat-induced epitope retrieval through a 30-min incubation at 97°C was processed in PT Link Pre-Treatment Module for Tissue Specimens (DAKO) with the use of EnVision Target Retrieval Solution (Agilent DAKO). Immunohistochemical staining with anti-NFIA (rabbit polyclonal; ab228897; dilution 1:500; Abcam, Cambridge, UK) and anti-NFIB (rabbit polyclonal; HPA003956; dilution 1:500; Sigma Aldrich, St. Louis, MO, USA) antibodies was performed in Autostainer Link 48 (DAKO) and visualized with diaminobenzidine. Human tissues of the liver and the prostate were used as positive controls for NFIA and NFIB, respectively. FLEX Rabbit Negative Control, Ready-to-Use (Agilent DAKO) instead of primary antibodies was utilized for negative controls. Nuclear expression of NFIA and NFIB was evaluated in cancerous compartments according to a modified Remmele scale by combining the intensity of staining (0: none, weak: 1 point, medium: 2 points, strong: 3 points) and the area of reactive tumor tissue (0%: 0 points, 1-25%: 1 point, 26%-50%: 2 points, 51%-75%: 3 points and 76%-100%: 4 points) (19). The product of these values gave the final immunoreactive score (IRS) ranging from 0 to 12 points.
Statistical analysis. All statistical analyses were carried out using R language [R. 4.1.0 (2021)]. NFIA and NFIB expressions were split into two classes: low (IRS≤4) and high (IRS>4); consequently, they were handled as dichotomous variables. The relationship of binary expression levels with age and tumor necrosis (which were treated as continuous variable) were performed by Welch two sample t-test and Mann–Whitney (Wilcoxon) test, respectively. Associations of dichotomous NFIA and NFIB expression levels with categorical variables were examined by Fisher exact test in case of binary variables and chi-squared test for the remaining non-binary variables. Kaplan–Meier curves and corresponding log-rank tests were generated with survminer package for R in order to compare survival times between the groups with low and high expression of NFIA and NFIB. For all statistical analyses the p-values were specified and the value of 0.05 was considered as the threshold of significance.
Results
Expression of NFIA and NFIB in bladder cancer. In the entire group of 136 bladder cancers, high NFIA expression was observed in 61 cases and low NFIA was seen in the remaining 75 cases (Figure 1), whereas high NFIB expression was seen in 66 cases and low in 70 cases (Figure 2). NFIA and NFIB expression scores were not correlated (data not shown). When analyzed separately, TURBT-treated NMIBCs showed considerably more robust NFIA staining compared to more advanced tumors from the cystectomy group (84% and 36% cases with high expression, respectively). High NFIB expression was observed in approximately half of either NMIBCs or MIBCs. Interestingly, some cases showed cytoplasmic staining for NFIA and/or NFIB, ranging from weak to intense, in addition to the typical nuclear expression pattern. Exceptionally, such cytoplasmic staining was not accompanied by any appreciable nuclear reactivity for NFIA/NFIB and these tumors were considered negative (Figure 1D and Figure 2D).
Expression of NFIA in urothelial carcinomas of the urinary bladder. A: Negative staining of a high-grade invasive bladder cancer; surface epithelium serves as an internal positive control (×100). B: Weak nuclear reactivity for NFIA in cancer cells (×100). C: Strong and diffuse nuclear staining of a noninvasive papillary urothelial carcinoma (×100). D: Aberrant cytoplasmic staining of bladder cancer cells with only scarce nuclear positivity (×200).
Expression of NFIB in urothelial carcinomas of the urinary bladder. A: Focal staining is seen only at the periphery of cancerous nests (×100). B: Medium intensity of staining in tumor cells (×100). C: Infiltration of urothelial carcinoma expressing high level of NFIB (×100). D: Distinct cytoplasmic positivity for NFIB with no nuclear expression (×200).
NFIA and NFIB expression and clinicopathological characteristics. NFIA staining was unrelated to patient sex and age in our cohort. By analyzing bladder cancers of various pT stages, a clear shift was observed from increased NFIA expression in early tumors to decreased staining in advanced cases. When pathological grading was considered, high NFIA expression prevailed among lower-graded tumors in either the newer two-tiered (LG/HG) or older three-tiered (G1/G2/G3) system. Also, tumor necrosis (area estimated for cystectomy specimens) was less extensive in tissues with high NFIA expression. Lymph node metastases were found in 44% of patients who underwent regional lymphadenectomy and were more frequent in patients with low NFIA expression (51% pN+ cases) compared to those with increased NFIA staining (32% pN+ cases); however, this difference did not reach a threshold of statistical significance. The abovementioned analyses yielded no apparent correlations in the case of NFIB expression. Interestingly, among 82 male patients who underwent cystoprostatectomy for bladder cancer, 15 had a concurrent, occult prostate cancer; 12 of 15 (80%) bladder tumors with concurrent prostate cancer exhibited high NFIB expression. Clinicopathological correlations are summarized in Table I.
Clinicopathological characteristics of the studied cases of bladder cancer according to the expression of NFIA and NFIB.
NFIA and NFIB expression and survival of bladder-cancer patients. Kaplan–Meier analysis of follow-up data for patients who underwent cystectomy revealed no significant differences with regard to overall survival (Figure 3) and recurrence-free survival (data not shown) between patients with low and high NFIA and NFIB expression.
Kaplan–Meier curves show no significant differences in overall survival among bladder cancer patients after cystectomy, stratified according to tumoral NFIA (A) and NFIB (B) expression.
Discussion
NFI proteins constitute a family of transcription factors known to be indispensable players during cell differentiation and organ development (8). Recent evidence has shown their functional involvement and prognostic significance in several cancers (5-7). NFIA and NFIB are the best studied members of the family, nevertheless little is known about their expression and clinicopathological implications in bladder cancer. In this study, we demonstrated that protein expression of NFIA, but not NFIB, is markedly down-regulated among MIBCs compared with NMIBCs. Likewise, NFIA immunoreactivity was decreased in tumors with higher pT stages. Moreover, tumors with low NIFA expression were more necrotic and generally of higher grade than tumors with abundant NFIA staining. There was also a tendency toward decreased expression of NFIA among cases with nodal metastases, but it did not reach a statistically significant level in our limited cohort. To the best of our knowledge, this is the first clinicopathological analysis of NFI protein expression in bladder cancer.
Our results correspond well with previous finding that NFIA mRNA level is decreased in infiltrating bladder cancer (5). However, molecular datasets show that analogous down-regulation occurs in the case of other NFI family members, including NFIB (5). This seeming inconsistency with the current study, which did not show such analogy in profiles of NFIA and NFIB expression on the protein level, may be related to methodological differences. While mRNA data from sequencing of bladder cancers were confronted with samples of healthy bladder mucosa (5), we compared NFIA/NFIB expression between MIBCs and NMIBCs as well as among various pT stages, and did not use normal urothelium as a reference. Consequently to the marked down-regulation of the mRNA of all NFIs in bladder cancer, these proteins might be perceived as tumor suppressors that are gradually lost during disease progression. However, Kaplan– Meier plotter shows that low mRNA levels of NFIA, NFIC, and NFIX, but not NFIB, counterintuitively imply a better prognosis (16). Moreover, Sharron Lin et al. performed comparative profiling of progressive vs. nonprogressive pT1 bladder carcinomas and discovered that NFIA mRNA was decreased in the latter (20). However, their analysis was limited to three nonprogressive and four progressive tumors, which impedes drawing definitive conclusions. On the protein level, we did not detect any relationship between NFIA and NFIB expression and disease recurrences or overall survival. The rationale for this discrepancy is unclear, but it is possible that the concentration of NFI proteins does not accurately mirror the expression of NFI genes due to intricate mechanisms of post-transcriptional regulation. At least two reports have been published that show either NFIA or NFIB mRNA is targeted by microRNAs in bladder cancer (21, 22), and multiple more such interactions are predicted by bioinformatic tools (23). Further research into the post-transcriptional regulation of NFI expression, as well as functional studies involving targeted silencing of NFI mRNAs will be necessary to pinpoint the impact of these transcription factors on bladder carcinogenesis.
Recent review by Chen et al. highlights a number of similarities between tissue-specific function of NFIs during development and carcinogenesis (7). For example, NFIB seems essential for mesenchymal-to-epithelial differentiation in lung development as well as for epithelial cell maintenance, while down-regulation of this transcription factor is related to a reverse process, namely epithelial-to-mesenchymal transition in non-small cell lung cancer (24, 25). In the central nervous system, NFIA orchestrates the onset of gliogenesis and induces differentiation of precursor cells towards astrocytes (7). This role as a lineage determinant is somewhat conserved during gliomagenesis – astrocytomas consistently show more NFIA-expressing cells than oligodendrogliomas and induced expression of NFIA in oligodendroglial tumors in mice results in their transdifferentiation to an astrocytoma-like phenotype (26, 27). Interestingly, forced over-expression of NFIA or NFIB in glioblastomas drives astrocytic differentiation of tumor cells, which has been suggested as a potential mechanism for therapy (28). With regard to the urinary bladder, clinical observations of human patients as well as experimental data from mouse models provide evidence that disruption of NFIA gene results in urinary tract defects and central nervous system malformations (29). Thus, NFIA plays a nonredundant role during urinary tract development, but it has yet to be explored on a mechanistic and molecular level. Future investigations in this field may give valuable insight into the significance of NFI proteins in bladder carcinogenesis.
Although clinicopathological correlations of NFIB expression did not emulate those of NFIA staining, we observed an intriguing convergence between high NFIB immunoreactivity in bladder cancer and the presence of occult prostatic carcinomas. It seems unlikely that a limited focus of an early prostate cancer would affect the biology of a concurrent bladder tumor in such a way to induce NFIB expression. Alternatively, this association could be an epiphenomenon in which NFIB would be upregulated in bladder cancer by some microenvironmental factors that drive carcinogenesis in the prostate. Of note, NFIB is typically over-expressed in prostatic carcinomas compared with benign controls (30). Lastly, a chance nature of this finding cannot be ruled out, because of the limited number of prostatic cancer cases in our cohort.
The cytoplasmic localization of NFIA and NFIB transcription factors demonstrated in our experiment was an unexpected finding. Thus far, such a pattern has been reported in esophageal squamous cell carcinoma for NIFA and in colorectal, breast, and prostate cancers for NFIB (13, 14, 30, 31). In the case of prostate cancer, cytoplasmic staining for NFIB was accompanied by nuclear reactivity (30). Moreover, cytoplasmic-to-nuclear ratio of NFIB expression predicted earlier biochemical recurrence (30). The mechanistic background of cytoplasmic localization of NFIs remains unclear. Nanda et al. suggested that in prostatic carcinomas, NFIB may accumulate in the cytoplasm simply due to its over-expression (30). However, this hypothesis seems rather unlikely an explanation for our observations because 1) intense cytoplasmic expression in bladder cancer cells was sometimes present without appreciable nuclear reactivity and 2) some tumors expressed cytoplasmic NFIA or NFIB even when the overall intensity was weak. Recently, Chen et al. demonstrated that subcellular localization of NFIB in breast cancer may be actively controlled by circularRNA (32). One of these novel molecules, namely circHIF1A, promoted expression and nuclear translocation of NFIB by posttranscriptional and posttranslational modulation (32). Another possibility is that cytoplasmic staining for NFIA/NFIB indicates, at least in some cases, their mutation disrupting the nuclear localization signal, in analogy to the aberrant cytoplasmic pattern of p53 staining (33). Molecular interrogation of NFIA and NFIB genes in bladder cancer in correlation with various expression patterns of NFIA and NFIB proteins might shed more light on the nature of their cytoplasmic localization. Whether these proteins play any functional role outside the nucleus remains to be determined.
In summary, this is the first characterization of NFIA and NFIB protein expression in bladder cancers in relation to clinicopathological parameters. We report that NFIA is down-regulated in cases with higher histologic grade and more invasive tumors, either in the context of pT stages or detrusor muscle involvement. On the other hand, the expression level of NFIB seems generally unrelated to various features of bladder tumors. These results indicate a potential tumor-suppressive activity of NFIA in urothelial carcinoma, however, this interpretation is influenced by a couple of limitations. While the majority of patients in our cohort were neoadjuvant treatment-naïve, some did undergo pre-cystectomy chemotherapy, which might have altered NFIA and NFIB expression in these cases and diminished the maximal extent of cancer according to pTNM categories. Second, the studied population is restricted and therefore our observations need to be validated on a larger scale. Subsequent investigations that would integrate data on mRNA as well as protein expression of NFIs and possibly how they are influenced by microRNAs in bladder cancer are warranted.
Acknowledgements
The study was funded by a statutory subsidy by the Polish Ministry of Science and Higher Education as a part of grant STM.B130.20.005 (record number in the Simple system).
Footnotes
Authors’ Contributions
MK: study design, resources, evaluation of immunohistochemistry, analysis of the results, drafting the manuscript; JM: statistical analysis, preparation of figures, drafting the manuscript; PK: clinical data, analysis of the results; BM: clinical data, analysis of the results; AH: resources, analysis of the results, supervision; All Authors read, revised and approved the final manuscript.
Conflicts of Interest
The Authors report no conflicts of interest in relation to this study.
- Received October 11, 2021.
- Revision received November 3, 2021.
- Accepted November 4, 2021.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.