TERT Promoter Mutation as a Potential Predictive Biomarker in BCG-Treated Bladder Cancer Patients

Telomerase reverse transcriptase gene promoter (TERTp) mutations are recognized as one of the most frequent genetic events in bladder cancer (BC). No studies have focused on the relevance of TERTp mutations in the specific group of tumors treated with Bacillus Calmette–Guérin (BCG) intravesical therapy. Methods — 125 non muscle invasive BC treated with BCG therapy (BCG-NMIBC) were screened for TERTp mutations, TERT rs2853669 single nucleotide polymorphism, and Fibroblast Growth Factor Receptor 3 (FGFR3) hotspot mutations. Results — TERTp mutations were found in 56.0% of BCG-NMIBC and were not associated with tumor stage or grade. FGFR3 mutations were found in 44.9% of the cases and were not associated with tumor stage or grade nor with TERTp mutations. The TERT rs2853669 single nucleotide polymorphism was associated with tumors of higher grade. The specific c.1-146G>A TERTp mutation was an independent predictor of nonrecurrence after BCG therapy (hazard ratio—0.382; 95% confidence interval—0.150–0.971, p = 0.048). Conclusions — TERTp mutations are frequent in BCG-NMIBC and -146G>A appears to be an independent predictive marker of response to BCG treatment with an impact in recurrence-free survival.


Introduction
Bladder cancer (BC) ranks as the fifth most common cancer in western society and the sixth most prevalent in the world, with an increasing incidence in the past years [1]. The increased incidence, along with the high costs in surveillance per BC patient, results in a high burden for public health systems [2,3]. BC can be divided in non muscle invasive (NMI) and muscle invasive (MI) tumors. NMI bladder cancer (NMIBC) accounts for 70% to 80% of all BC and is present as superficial and recurrent lesions that only seldom progress to an MI phenotype. Prompt treatment, usually with complete transurethral tumor resection, grants a 5-year survival rate that can surpass 90%. However, up to 70%-80% of them may relapse, making recurrence the main challenge in clinical management [3,4]. Present in approximately 70% of cases, Fibroblast Growth Factor Receptor 3 (FGFR3) activating mutations are the most frequent genetic event in the NMI phenotype [4]. MI bladder cancer (MIBC) accounts for the remaining 20% to 30% of BC cases and presents as an invasive tumor at diagnosis. Characterized by a high risk of distant metastasis, MIBC prognosis is considerably worse, with 5-year survival rates often described as lower than 40% [5]. MI tumors are genetically more heterogeneous than NMI tumours; present in approximately half of the cases, TP53 mutations are identified as the most frequent genetic alteration in these tumors [4].
Cell immortalization is a classic hallmark of cancer cells and telomerase reactivation is proposed to be involved in the underlying process. In a large part of cancer models, the intervening mechanisms remained elusive, until in 2013, mutations of the promoter of the telomerase (TERT) were described in melanoma [6,7]. We and others reported for the first time the presence of recurrent somatic mutations in the TERT promoter (TERTp) in numerous types of cancer, including BC [8][9][10][11][12][13][14][15]. Studies focusing on BC have described a prevalence of TERTp mutations ranging from 52% to 85% of the cases [10,13,14,[16][17][18][19]. Conflicting results have been obtained on the association between TERTp mutations and BC clinical outcome [13,14]. A common polymorphism in TERTp (rs2853669 single nucleotide polymorphism) is also accountable to act as a modifier of the promoter mutations' effect on survival and tumor recurrence in several cancers, such as glioblastoma, liver, and bladder cancer [18,20,21].
Clinicopathological features are the central determinants of recurrence, and according to the European Organization for Research and Treatment of Cancer (EORTC), the NMIBC high-risk group includes high-grade papillary tumors, carcinoma in situ, and those with multifocal or recurrent lesions [22]. Tumor resection followed by a schedule of intravesical instillations with Bacillus Calmette-Guérin (BCG) is the standard adjuvant therapy for this high-risk group (henceforth referred to as BCG-NMIBC) [22,23]. Nonetheless, 30% to 40% of patients present either intolerance or recurrence following BCG treatment, demanding a life-long follow-up and repeated courses of treatment [24]. This clinical relevance is recognized and there is a shortage of dedicated genetic markers predicting BCG-NMIBC subgroup outcomes, in particular, the now recognized two most common genetic events in NMIBC-TERT promoter and FGFR3 mutations.
In this study, we screened a series comprising 125 BCG-NMIBCs resected before BCG therapy initiation for TERTp mutations, FGFR3 mutations, and for the TERTp rs2853669 polymorphism. This represents a unique report of TERTp and FGFR3 mutation genotyping dedicated to the BCG-NMI group of BC. To investigate the significance of TERTp mutations in the BCG-treated tumor response, we compared the obtained results with the available clinicopathological data, including recurrence-free survival following BCG therapy.

Clinicopathological Characteristics and Genetic Alterations
A comparison between the clinicopathological characteristics of TERTp wild type and mutated cases was performed ( Table 2). An association was found between TERTp mutations and recurrence status prior to BCG therapy, where an over-representation of TERTp mutations in primary tumors when compared with recurrent tumors can be detected (61.4% vs. 38.6%, p = 0.048).
In the BCG-NMIBC cases a statistically significant association between tumor size and FGFR3 p.R248C mutations was found (p = 0.048). There was an over-representation of the mutation presence among tumors larger than 3 cm in comparison with the smaller ones (27.9% vs. 11.5%), Table S2. However, multivariate analysis revealed that FGFR3 p.R248C is not independently associated with tumor size.
The stratification of tumors in two groups, those wild type for both TERTp and FGFR3 and those mutated for any, did not present statistically significant differences in the clinicopathological characteristics. Regarding the relationship of the studied polymorphism and clinicopathological features, an over-representation of the rs2853669 AA genotype was found in high-grade tumors when compared with low-grade tumors (77.4% vs. 22.6%, p = 0.018) ( Table 3).

Clinicopathological and Molecular Characteristics with BCG Therapy Success
Prior to tumor sampling, the BCG-NMIBC patients were treated with a scheme of BCG intravesical therapy. We evaluated how the clinicopathological characteristics affected BCG therapy outcome. Success was defined as no recurrence detected until the last surveillance check-up. Failure was defined as any recurrence after BCG treatment. After a univariate analysis, the age group ≥65 years (hazard ratio (HR): 2.827; 95% CI: 1.481-5.398; p = 0.002), multifocality (HR: 2.000; 95% CI: 1.096-3.649; p = 0.024) and maintenance BCG (mBCG) schedule (HR: 0.505; 95% CI: 0.282-0.902; p = 0.021) were the only variables significantly associated with the outcome, Table S1.
Next, we evaluated if the molecular characteristics have an effect on BCG therapy success. We performed a univariate analysis considering TERTp and FGFR3 mutations and BCG therapy success. No statistically significant association was found on univariate analysis. To adjust for the effect of age group, multifocality, and BCG schedule on treatment success, we then performed a multivariate Cox regression analysis. When adjusted, the effect of status (wild type vs. mutated) for TERTp and FGFR3 remained nonsignificant (Tables 4 and 5). However, when we considered TERTp c.1-146G>A carriers against TERTp non c.1-146G>A carriers (either TERTp wild type or c.1-124G>A), the c.1-146G>A mutation was significantly associated with therapy success (HR: 0.382; 95% CI: 0.150-0.971; p = 0.043) ( Tables 4 and 5). In our series, the TERTp mutation c.1-146G>A was an independent predictor of therapy sucess following BCG intravesical therapy.
We further investigated the possible role of TERTp genetic events in predicting BCG therapy success by evaluating the presence of the single nucleotide polymorphism rs2853669 in the BCG-NMIBC series. We characterized cases as either carrier or noncarrier. No significant association was found for rs2853669 per se, or for TERTp mutation effect after splitting for rs2853669.

TERTp Mutations and Recurrence-Free Survival
The recurrence-free survival function of all 125 BCG-treated NMIBC patients, grouped according to the existence of a TERTp mutation, was evaluated and log-rank testing revealed no statistically significant difference for either group ( Figure S1). When considering TERTp c.1-146G>A carriers against TERTp non c.1-146G>A carriers (either TERTp wild type or c.1-124G>A), the TERTp c.1-146G>A patients presented a longer recurrence-free survival in comparison with the noncarriers (mean 126 months vs. mean 100 months, log rank p = 0.035) ( Figure 1).

TERTp Mutations and Recurrence-Free Survival
The recurrence-free survival function of all 125 BCG-treated NMIBC patients, grouped according to the existence of a TERTp mutation, was evaluated and log-rank testing revealed no statistically significant difference for either group ( Figure S1). When considering TERTp c.1-146G>A carriers against TERTp non c.1-146G>A carriers (either TERTp wild type or c.1-124G>A), the TERTp c.1-146G>A patients presented a longer recurrence-free survival in comparison with the noncarriers (mean 126 months vs. mean 100 months, log rank p = 0.035) (Figure 1).
Conflicting results have been reported on the association between TERTp mutations and clinical stage and/or grade of bladder tumors. Wu et al. [19] found that TERTp mutations were more prevalent in MI tumors than in NMI tumors and in patients with advanced tumor stages. On the other hand, other studies reported no association between mutation status and stage or grade [13,14]. The results we present here support that TERTp mutations rates are not significantly different across grades or stages in this subset of BCG-treated NMIBC. However, it should be taken into account that this subset represents a group of particularly aggressive NMI tumors (BCG-NMIBC), and comparisons with NMIBC subseries in other studies must be made with caution.
Found in approximately 70% of tumors, FGFR3 activating mutations are regarded as important genetic events in the NMI phenotype [4]. We found that 44.9% of the BCG-NMIBC cases were

Discussion
TERTp mutations were reported in 52% to 85% of bladder cancer (BC) cases, depending on the series [10,13,14,[16][17][18][19]. These results rank TERTp mutations as one of the most common genomic events observed in BC and possibly as the most frequent. Of all the TERTp mutations, c.1-124G>A has been consistently reported as the most frequent, detected in 88% to 95% of the positive cases [10,13,14,[16][17][18][19]. In this study composed of BCG-treated NMIBC tumors, we report an overall TERTp mutation prevalence of 56.0%, in accordance with previously reported studies [10,13,14,[16][17][18][19]. Conflicting results have been reported on the association between TERTp mutations and clinical stage and/or grade of bladder tumors. Wu et al. [19] found that TERTp mutations were more prevalent in MI tumors than in NMI tumors and in patients with advanced tumor stages. On the other hand, other studies reported no association between mutation status and stage or grade [13,14]. The results we present here support that TERTp mutations rates are not significantly different across grades or stages in this subset of BCG-treated NMIBC. However, it should be taken into account that this subset represents a group of particularly aggressive NMI tumors (BCG-NMIBC), and comparisons with NMIBC subseries in other studies must be made with caution.
Found in approximately 70% of tumors, FGFR3 activating mutations are regarded as important genetic events in the NMI phenotype [4]. We found that 44.9% of the BCG-NMIBC cases were mutated for FGFR3. As FGFR3 mutations are associated with low-grade and low-stage tumors and seem to predict a more favorable clinical outcome among patients with NMI tumors [3,25], it was expectable that the more aggressive BCG-NMIBC tumors presented lower mutation rates than those reported in other series [4]. Analyzing the specific FGFR3 mutation distribution, a novel pattern emerges; previously, in NMIBC series, the most frequent mutations were at the exon 7 p.S249C (66.6% overall, 87.3% of the exon 7 mutations) and p.R248C (9.7% overall, 12.0% of the exon 7 mutations), and mutations in exon 10 and 15 were infrequent [26]. The FGFR3 mutations detected in this study were mostly present on exon 7 (91.7%), but we observed a different prevalence of the specific mutations, with p.S249C accounting for 52.0%, and an enrichment for p.R248C, with 39.6%. In the BCG-NMIBC cases wild type and mutated for FGFR3, a statistically significant association between tumor size and FGFR3 p.R248C mutation was found (p = 0.048). However, multivariate analysis revealed that FGFR3 p.R248C is not independently associated with tumor size. One can discuss if FGFR3 p.R248C mutations are associated with more aggressive tumors, since it is particularly enriched in this subset of tumors and associated with larger tumor size in a univariate analysis. Loss of association in multivariate analysis may indicate that this association may be due to the influence of other clinicopathological features or, more likely, the analyzed cohort is too low to robustly perform this analysis. Further studies are required to elucidate these assumptions and findings. Finally, no significant association was found between the presence of TERTp and FGFR3 mutations.
Intravesical BCG therapy is used as prophylaxis against NMIBC recurrences after tumor resection and it is in fact regarded as one of the first and most successful of all oncological immunotherapies [27,28]. BCG intravesical instillation results in multiple immune reactions. Although the precise immunological mechanism of BCG therapy is not clear, it appears to act through three main actions-infection of urothelial cells or bladder cancer cells, induction of immune responses, and induction of anti-tumour effects [27]. Although effective, 30% to 40% of the cases still show either intolerance or recurrence after BCG treatment, demanding life-long follow-up and repeated courses of treatment [24]. This results in extreme discomfort for the patients and exceedingly high financial costs, and ranks BC as the most expensive cancer per patient [3]. Biomarkers that could help identify which patients were more likely to respond to BCG versus those with risk of recurrence -those who would benefit the most from either a tighter surveillance or a different treatment -would be very useful in optimizing the clinical care offered to BC patients. In this study, we report the effects of TERTp and FGFR3 mutations in BCG therapy success (recurrence or nonrecurrence) and recurrence-free survival. Age at BCG treatment, multifocality, and BCG schedule were independent predictors of BCG therapy success (defined as no recurrence), with the age group ≥65 years and multifocal tumours associated with a higher risk of recurrence, whereas mBCG schedule was associated with a lower risk of recurrence. These results are concordant with previous reports [29]. After adjusting for age, multifocality, and BCG schedule, we found no association between FGFR3 mutations and BGC therapy success. Similarly, TERTp-mutated cases as a whole showed no difference when compared to wild type cases. However, when we compared carriers of the TERTp c.1-146G > A mutation against those without this mutation, we observed that this specific mutation was an independent predictor of better outcome (delayed or nonrecurrence).
Recently, Rachakonda et al. reported that a common polymorphism within a pre-existing Ets2 binding site in TERTp, rs2853669, acts as a modifier of the mutations' effect on survival and tumor recurrence [18]. The patients with the TERTp mutations presented a poorer survival in the absence, but not in the presence of the variant allele (G) of the polymorphism [18]. TERTp mutations in the absence of the variant allele were highly associated with disease recurrence in patients with Tis, Ta, and T1 tumors [18]. To further investigate this, we screened BCG-NMIBC tumors for this common SNP. We found that rs2853669 carrier status did not modulate TERTp effect on BCG therapy success in our series, however, there was an association of rs2853669 AA carriers with tumors of higher grade. This association is in accordance with what Rachakonda and others described. In patients that harbor the germline rs2853669 AA genotype, the TERTp mutation effect is not reverted in the BC tumor. As Rachakonda described, patients with this combination (germline rs2853669 G absence, TERTp tumor positive) present poorer survival and increased disease recurrence, which are features compatible with the presence of more aggressive tumors, such as higher grade.
Finally, we analyzed recurrence-free survival after BCG treatment, comparing TERTp and FGFR3 mutation status and specific mutations and rs2853669 carrier status. Kaplan-Meier survival analysis showed a promising recurrence-free survival advantage for those c.1-146G>A mutation carriers. Our results demonstrate that BCG-NMIBC c.1-146G>A mutation carriers are three time less likely to recur after BCG therapy and may have more favorable recurrence-free survival rates when compared to both TERTp wild type and c.1-124G>A cases. To interpret these findings, it is important to note that what we are evaluating is how TERTp mutations modulate the tumor response after BCG therapy. It has previously been suggested that the mechanism of BCG therapeutic effects on BC is related to its ability to reduce telomerase (TERT) activity [30] -we may speculate that c.1-146G>A mutated tumors might be more susceptible to the reduction of telomerase activity by BCG. As reported by Huang et al., TERTp mutations are associated with higher TERT transcription levels compared to wild type promoters but TERTp c.1-146G>A carriers have lower transcriptional capacity than those with the c.1-124G>A mutation [7]. We can speculate that the higher TERT expression induced by the c.1-124G>A mutation partially impairs BCG capability to sufficiently reduce telomeric activity to a therapeutic level -a level that could be achieved in a c.1-146G>A setting. Also, the lower frequency of TERTp mutations in BCG-treated recurrent tumors when compared to primary tumors could be explained by the enhanced BCG action on tumor cells harboring TERTp mutations, leading to clonal selection pressure towards cells harboring other alterations (such as FGFR3 mutations) in recurrent tumors, hence shifting the prevalence of recurrent tumors towards TERTp-negative tumors. More studies comparing TERT expression and telomerase activity before and after BCG therapy with the different TERTp mutations are required to further interpret our results. To our knowledge, this study is one of the first studies addressing TERTp and FGFR3 mutations in a BCG-NMI series of BC patients. We found no association between TERTp mutations, as a whole, and tumor grade or stage. However, we observed that the specific TERTp c.1-146G>A mutation was an independent predictor of nonrecurrence after BCG therapy in the BCG-NMI tumors. Our results suggest that it might be relevant to further study the role of TERTp mutations in tumor recurrence and as predictive markers of response to BCG therapy.

Human Cancer Samples and Clinicopathological Data
Formalin-fixed, paraffin-embedded (FFPE) tissues were obtained from 125 patients with NMI bladder urothelial cell carcinoma treated with intravesical BCG therapy, with samples being collected at the time of transurethral resection before any BCG therapy administration. Patients underwent resection of the tumors in the Portuguese Institute of Oncology -Francisco Gentil (IPO) Porto. Hematoxylin-eosin-stained sections were reviewed according to the standard histopathological examination by two independent pathologists. Staging and grading were conducted according to the American Joint Committee on Cancer [31], and the 2004 WHO classification system [32]. Clinicopathological and follow-up data were retrieved from the files of IPO databases. Age refers to age at BCG treatment initiation in the BCG-NMIBC group. Recurrence status characterizes the BCG-NMIBC cases as either a primary newly diagnosed tumor selected for BCG therapy or, alternatively, as a recurrence of a previously resected NMI tumor (that did not fill the criteria for being included in the BCG-NMIBC group before) that is only now selected for BCG therapy. BCG therapy selection was performed according to the EORTC criteria previously described [22]. BCG schedule characterizes the treatment regimen used as maintenance (mBCG) or induction-only (iBCG) intravesical BCG instillation. BCG therapy success was defined as no recurrence and failure was defined as any recurrence after BCG treatment. Analysis of patients' age by age groups (<65 years and ≥65 years) was recognized as an informative analysis and has been used previously [33]. All the procedures described in this study were in accordance with national and institutional ethical standards and previously approved by Local Ethical Review Committees (Ethics Committee of the Portuguese Institute of Oncology of Porto with the number CES IPOPFG-EPE 586/08 in 25 of September of 2008). According to Portuguese law, informed consent is not required for retrospective studies.

DNA Extraction, PCR, and Sanger Sequencing
DNA was obtained from FFPE (10-micron sections) after careful microdissection. DNA extraction was performed using an Ultraprep Tissue DNA Kit (AHN Biotechnologie, Nordhaussen, Germany) following manufacturer's instructions.
To screen for TERTp mutations, we analyzed by PCR followed by Sanger sequencing of the hotspots previously identified [10]. TERTp mutation analysis was performed with the pair of primers Fw TERT-5 -CAGCGCTGCCTGAAACTC-3 and Rv TERT-5 -GTCCTGCCCCTTCACCTT-3 . Amplification of genomic DNA (25-100ng) was performed by PCR using the Qiagen Multiplex PCR kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Sequencing reaction was performed with the ABI Prism BigDye Terminator Kit (Perkin Elmer, Foster City, CA, USA), and the fragments were run in an ABI prism 3100 Genetic Analyzer (Perkin-Elmer). The sequencing reaction was performed in a forward direction, and an independent PCR amplification/sequencing, both in a forward and reverse direction, was performed in positive samples or samples that were inconclusive. To screen for FGFR3, we analyzed the hotspots previously identified in exon 7, 10, and 15 in 107 BCG-NMIBC cases (18 cases have been excluded due to insufficient DNA for the analysis) by PCR followed by Sanger sequencing. FGFR3 exon 7, 10, and 15 mutation analysis was performed with the respective pairs of primers Fw Exon 7-5 -AGTGGCGGTGGTGGTGAGGGAG-3 and Rv Exon 7-5 -GCACCGCCGTCTGGTTGG-3 ; Fw Exon 10-5 -CAACGCCCATGTCTTTGCAG-3 and Rv Exon 10-5 -AGGCGGCAGAGCGTCACAG-3 ; Fw Exon 15-5 -GACCGAGGACAACGTGATG-3 and Rv Exon 15-5 -GTGTGGGAAGGCGGTGTTG-3 . Subsequent steps followed the same methodology as outlined for the TERT promoter mutation screening.

Single Nucleotide Polymorphism Assay
Screening for the rs2853669 polymorphism was performed in 98 BCG-NMIBC cases (27 cases were excluded due to insufficient DNA for the analysis) using the rs2853669 TaqMan ® SNP Genotyping Assay (Applied Biosystems, Foster City, USA). Peripheral blood DNA was extracted using a genomic DNA extraction kit (Qiagen). The purified genomic DNA was used for the assay. The procedure was performed according to manufacturer's instructions.

Uromonitor Real-Time PCR screening Assay
Screening of 125 nonmuscle invasive BC tumors treated with BCG therapy (BCG-NMIBC) for TERTp mutations and FGFR3 hotspot mutations were confirmed by using a specific IVD commercial kit Uromonitor ® -Real-Time PCR kit for the amplification and detection of TERTp and FGFR3 hotspot mutations (U-Monitor, Porto, Portugal), according to manufacturer's instructions.

Statistical Analysis
The statistical analysis was performed using IBM SPSS statistics software version 25.0. For the analysis of the relationship between patients' age, we used the independent-samples t-test. Pearson's Chi-square and Fisher's exact test were used in the statistical analysis of the other parameters, according to sample size. Cox proportional hazard ratios were estimated to obtain risks of recurrence for cases in each molecular factor stratum before and after adjusting for other confounding variables. Kaplan-Meier survival curves were computed by each category of the potential prognostic factors and the log-rank and Breslow tests were applied to compare curves. Means were used instead of medians because some survival curves did not fall under 50%. Results were considered statistically significant if p < 0.05.