Development of a Custom NGS Panel for the Determination of Bladder Cancer Risk

Bladder cancer (BCa) is a heterogeneous disease caused by the interaction between environmental and genetic risk factors. The objective of this study was to design a panel that evaluates the role of some selected variants in BCa susceptibility. We are also interested in studying the interaction between environmental and genetic risk factors. The case/controls cohort was composed with 249 BCa cases and 255 controls. The designed Bladder cancer hereditary panel (BCHP) is composed of 139 selected variants. These variants were genotyped by an amplication-based targeted Next-Generation Sequencing (NGS) on the Ion Torrent Proton sequencer (Life Technologies, Ion Torrent technology). We have found that rs162555, rs2228000, rs10936599, rs710521, rs3752645, rs804276, rs4639, rs4881400 and rs288980 were signicantly associated with decreased risk of bladder cancer. However the homozygous genotypes for VPS37C (rs7104333, A/A), MPG (rs1013358, C/C) genes or the heterozygous genotype for ARNT gene (rs1889740, rs2228099, rs2256355, rs2864873), GSTA4 (rs17614751) and APOBR/IL27 (rs17855750) were signicantly associated with increased risk of bladder cancer development compared to reference group (OR=2.53, 2.34, 1.99, 2.00, 2.00, 1.47, 1.96 and 2.27 respectively). We have also found that non–smokers patients harboring heterozygous genotypes for ARNT/rs2864873 (A>G), ARNT/ rs1889740 (C>T) or GSTA4/rs17614751 (G to A) were respectively at 2.775, 3.069 and 6.608 –folds increased risk of Bca development compared to non-smokers controls with wild genotypes. Moreover the ARNT CT (rs1889740), normalized 5ng/ul 20 plates validated in-house protocol perform multiplex PCRs reaction HotStarTaq out format plates conditions 22 paramagnetic OT2 sequencing reaction Sequencing This gene encodes for an enzyme implicated in the control of cell cycle and plays an important role in apoptosis. Abnormal expression of TP63 is associated with a loss of urothelial differentiation [54]. In our study the TP63 T/C genotype (rs710521) was reported to be inversely associated with BCa (p-value = 0.035 and an OR = 0.62; CI95% 0.42–0.89). Our results seem to agree with those of Kimeney L. et al. and Lehmann M.-L et al. who conrmed that this variant plays a role in decreasing the risk of BCa in European patients [55, 56]. We have also found that ROCK1 TT genotype (rs288980) was associated with 0.75 fold-decreased risk of BCa. ROCK1 enzyme is a necessary effector kinase downstream of Rho GTPases, a very important pathway involved in cell migration and has been identied as a possible therapeutic target [57]. The Cancer Genome Project identied three non-synonymous mutations within the ROCK1 gene [58] but the effect of these variations in protein activity or expression was not more elucidated. Finally ACTRT3/ MYNN (Actin Related Protein T3; 3q26.2) gene is poorly described in the literature. The minor T allele of rs10936599 in ACTRT3/ MYNN has been described in the Genome-wide association study conducted by Figueroa J. et al. in 2014, to be highly associated with decreased risk of BCa [59], which was conrmed by Wang M. et al. [60]. However, the effect of rs10936599 polymorphism on the MYNN activity is not well known and more biological functional studies are needed to draw more concise conclusions regarding the underlying molecular mechanism of this variation.


Biological Samples And Dna Extraction
Peripheral blood samples were collected into tubes with ethylene diamine tetra-acetic acid EDTA (PH 8). Genomic DNA was extracted from leukocytes using a phenol /chloroform procedure. First, the integrity of the genomic DNA was visualized by electrophoresis on a 1% agarose gel stained with ethidium bromide. In a second step, DNA samples were quanti ed by Nanodrop and Qubit (High Sensitivity HS / Broad Rang BR) and were dried with the speedvac (Eppendorf™ Vacufuge™ Concentrator) to be normalized into 6*96-well plates.

Panel Design
We have selected a panel of 139 polymorphisms from 97 genes. These SNP were selected according to their eventual implications in poor-prognosis BCa such as those affecting the xenobiotic metabolism, DNA repair, treatment response, and in ammatory reaction. The design of the panel was done, after consulting published data and available databases (dbSNP, UCSC Genome Browser, and GWAS).

Results
All assays were performed using DNA samples from a total of 544 cases and controls and duplicated samples for quality control. Only 40 samples (40/544; 7.35%) failed the library preparation and were excluded from the analysis. In this series, two runs led to 75% and 88% ISP loading and to the generation of total bases of 4.33 and 11.4 Gigabases for total bases of 33, 329, 31 and 84,548,378 respectively, 99% of which aligned to the reference genome (Hg19) (Fig.   4). This analysis provided high analytical sensitivity and allowed detecting SNPs. Using this approach, we have generated molecular pro les of a large number of individuals and identi ed speci c BCa variations.
The N-acetyltransferase 2 (NAT2) gene encodes for an important phase II xenobiotic-metabolizing enzyme frequently present in liver and intestinal mucosa. It catalyzes the reaction of aromatic and heterocyclic amine carcinogens via O-acetylation and N-acetylation [28][29][30]. In our case-control study we have found an inverse association between NAT2 A/A genotype (rs1799930) and BCa development. This result concord with those of Lei Quan who reported that NAT2 A/A carriers (rs1799930) were at 50% decreased risk of bladder cancer development compared to rapid acetylator [31]. Catechol-O-methyltransferase (COMT) encodes a phase II enzyme mainly liable for the degradation of catecholamines, like dopamine and noradrenalinev [32]. It catalyzes the O-methylation of 2hydroxyestradiol to yield 2-ME2 [32]. Additionally, it is involved within the inactivation of potential carcinogenic compounds that produce in ammation and catechol estrogens and thus might protect DNA from oxidative damage. Rs4680 (Ex4-12 G > A, or val158met) is the most studied COMT single nucleotide polymorphism (SNP). In our cohort we explain the signi cant inverse association of heterozygous genotype against BCa development by the modi cation of enzyme activity associated to this genotype. Indeed it has been reported that valine (val) variant enzyme is 3-4 times more active than the methionine (met) variant [33,34]. Moreover we observed that PRKAR2B A/A genotype (rs3752645) is inversely correlated with BCa risk. The rs3752645 G/A SNP lies within the PRKAR2B gene, which encodes a regulatory subunit for cyclic adenosine 3′, 5′-monophosphate kinase. Our result con rms previous ndings of a GWAS study where the authors showed that the rs3752645 SNP had a strong inverse association with BCa risk [35].
In contrast we have found that homozygous minor genotype or heterozygous genotype for rs1889740 C/T, rs2228099 C/G, rs2256355 T/C, rs2864873 A/G in ARNT gene (Aryl hydrocarbon receptor nuclear translocator) were signi cantly associated to increased risk of bladder cancer. These SNPs are described for the rst time in association with increased risk of BCa development. ARNT gene encodes a protein that binds to ligand-bound aryl hydrocarbon receptor and promotes the expression of genes involved in xenobiotic metabolism [36]. We also found that the inheritance of rare homozygous genotype of GSS (rs7260770) is associated with an increased risk of BCa which con rms ndings reported in other studies [37]. These ndings may be explained by the functional impact of this variant on the enzyme activity. Indeed it has been reported that GSS de ciency resulted in decreased GSH levels and caused dramatic metabolic consequences. However, the role of GSS in cancers has not been studied in details. Finally we showed that GSTA4 G/A genotype (rs17614751) was associated with an increased risk of BCa in our population. This result could be explained by the perturbation of enzyme activity or expression (the rs17614751 SNP is a downstream gene variant). Moreover, it has been reported that the deregulation of GSTA4 enzyme activity was implicated in cell proliferation in nasopharyngeal carcinoma, breast cancer, and liver cancer cells [38].

DNA repair pathway
When considering the DNA repair pathway we have found that the minor alleles of the rs4639 and rs2228000 SNPs in NEIL2 and XPC genes respectively were associated with a decreased risk of BCa development compared to reference groups carrying major alleles. However the minor alleles observed in the studied population for the rs1013358 and rs1799801 SNPs in respectively MPG and ERCC4/ XPF genes increased the risk of bladder cancer.
The NEIL2 (Nei-like DNA glycosylase 2) gene encodes for an enzyme implicated in the rst step of the base excision repair (BER) mechanism which consists of cleaving oxidatively damaged bases and introducing a DNA strand break via the associated lyase reaction [39]. The rs4639, is located on 3′UTR of NEIL2 [40]. In our study, we found that NEIL2 G/G genotype for rs4639 was associated with 0, 49-fold decreased risk of BCa. This inverse association could be explained by the fact that the minor allele interacts with some speci c miRNA, which activates the BER pathway [41]. Moreover, we have found that rs4639 and rs804276 in NEIL2 gene were in linkage disequilibrium (p-value = 0.03). Beside NEIL2, we found that XPC AA genotype (rs2228000) was associated with a decreased risk of BCa compared to reference group. The XPC gene is located on 3p25 and encodes for an enzyme involved in global genome repair. This enzyme represents the earliest damage detector by initiating the NER pathway and eliminating the DNA damages induced by chemical and environmental exposures such as aromatic amines and UV light [42]. Our result con rms the study of Zhu Y. et al. which demonstrates a signi cant association between the presence of XPC rs2228000 AA genotype and a decreased risk of BCa [43]. However, the recent study of Dai Y. et al. suggests that XPC AA genotype may be linked to an increased risk of bladder and breast cancer [44]. These con icting results and differences in risk associations may be explained by the different etiology and mechanisms of BCa in study populations with different ethnic backgrounds [15].
On the other hand, ERCC4/XPF and MPG enzymes represent the headmaster in DNA metabolic process, DNA repair, and cellular response to DNA damage stimulus. ERCC4 (16p13.12) is a NER gene that plays a key role in DNA repair that protects against genetic instability and carcinogenesis [45]. The ERCC4-rs1799801 is a synonymous variation (Ser835Ser). A signi cant association between the ERCC4-rs1799801 polymorphisms and increased risk of Bca in Tunisian population could be explained by the effect of this variation on the enzyme activity. Indeed a previously genotype-phenotype correlation analysis indicated that the ERCC4-rs1799801 rare homozygous C/C genotype carriers had an increased trend of ERCC4 expression levels [46]. In contradiction with our results, the meta-analysis of Shi T.Y. et al. also did not provide statistical evidence for an association between ERCC4 gene and the overall risk of several human cancers, they also report that strati cation analysis showed an inverse correlation with cancer in Caucasians [47]. These con icting results could be explained by a difference in both ethnic origins of patients and exposure to environmental risk factors. MPG, a BER gene, also plays an important role in DNA repair.

Others cellular pathway
In addition to xenobiotic metabolic pathway and DNA repair process we are interested in this study to investigate the impact of others variations in gene known as a negative regulator of T-and B-cell receptor signaling. Its expression has been shown to be implicated in T lymphocytes tolerance toward tumor cells [49]. In contrast we have found that genetic variations in BLNK (rs3789928), TP63 (rs710521), ROCK1 (rs288980) and ACTRT3/MYNN (rs10936599) were associated with a decreased risk of bladder cancer. BLNK gene encodes for a cytoplasmic linker or adaptor protein that plays a critical role in B cell development [50]. It plays an important role in the pro-B cell to pre-B cell transition and B-cell apoptosis via BCR signaling pathway and reported to have a tumor-suppressive function in various hematologic malignancies [51]. Moreover the Human Protein Atlas consortium showed that high expression of BLNK had a favorable prognosis value in urothelial cancer [52]. TP63 (3q27-28) has a homolog sequence to TP53 (tumor suppressor) and TP73 [53]. This gene encodes for an enzyme implicated in the control of cell cycle and plays an important role in apoptosis. Abnormal expression of TP63 is associated with a loss of urothelial differentiation [54]. In our study the TP63 T/C genotype (rs710521) was reported to be inversely associated with BCa (p-value = 0.035 and an OR  [55,56]. We have also found that ROCK1 TT genotype (rs288980) was associated with 0.75 fold-decreased risk of BCa. ROCK1 enzyme is a necessary effector kinase downstream of Rho GTPases, a very important pathway involved in cell migration and has been identi ed as a possible therapeutic target [57]. The Cancer Genome Project identi ed three non-synonymous mutations within the ROCK1 gene [58] but the effect of these variations in protein activity or expression was not more elucidated. Finally ACTRT3/ MYNN (Actin Related Protein T3; 3q26.2) gene is poorly described in the literature. The minor T allele of rs10936599 in ACTRT3/ MYNN has been described in the Genome-wide association study conducted by Figueroa J. et al. in 2014, to be highly associated with decreased risk of BCa [59], which was con rmed by Wang M. et al. [60]. However, the effect of rs10936599 polymorphism on the MYNN activity is not well known and more biological functional studies are needed to draw more concise conclusions regarding the underlying molecular mechanism of this variation.

Interaction between genetic and environmental risk factors
When we compare the distribution of unfavorable genotypes [GSS G/A (rs7260770), ARNT A/G (rs2864873), ARNT C/T (rs1889740) ARNT C/G (rs2228099), ARNT T/C (rs2256355), GSTA4 G/A (rs17614751), ERCC4 C/C (rs1799801), MPG C/C (rs1013358), VPS37C A/A (rs7104333), APOBR/IL27 A/C (rs17855750)] between exposed bladder cancer cases and controls according to environmental risk factors we found that ARNT C/T (rs1889740), ARNT A/G (rs2864873) and GSTA4 G/A (rs17614751) genotypes were associated with an increased risk of BCa even in the absence of tobacco risk factors, or professional risk factors. Moreover when we compared exposed cases to exposed controls, we have not found a signi cant additive effect between the majority of unfavorable genotypes and exposure to environmental risk factors (smokers' ≥ 20PY or exposed to professional risk factors). The absence of signi cant additive effect between the majority of unfavorable genotypes and BCa risk could be explained by the fact that the targeted SNPs encoding enzymes were not directly implicated in the metabolism of xenobiotics. For example, the proin ammatory role of APOBR/IL27 enzyme could explain the role of rs17855750 in BCa development independently to exposition to environmental risk factors. However this additive effect reached more than 12 fold-increased risk of BCa development for ARNT rs1889740 T/T, ARNT rs2228099 G/G and ARNT rs2256355 C/C genotypes in subjects exposed to professional risk factors. This association may be explained by the fact that ARNT gene encodes a protein thats binds to ligand-bound aryl hydrocarbon receptor and promotes the expression of genes involved in xenobiotic metabolism [36]. Whether this association is caused by speci c bladder carcinogens present in the work environment warrants further investigations. To conclude our analysis, a decision tree was implemented to create a disease prediction model. This tree has allowed us to de ne the risk groups most genetically likely to develop BCa. According to the established decision tree, we note the importance of both smoking status (≥ 20PY) and the genotype of CYP1A2 (rs762551) in the development of BCa. The decision-tree analysis produced a three major BCa class. The rst major class (58/249) is de ned by the presence of CYP1A2 CC or CYP1A2 CA genotype and also de ned by others 8 variations: ARNT C > G (rs2228099), MAP2K4 C > T (rs4791489). The second major class of BCa group (31/249) is only de ned by the intensity of tobacco use ≥ 20PY and the inheritance of the homozygous genotype for rs762551 in CYP1A2 (CYP1A2 A/A), XPC GG or AG genotype (rs2228000, G > A), and MAP2K4 CC or CT genotype (rs4791489, C > T). This decision tree con rms the crucial role of tobacco consumption in the etiology of BCa. Indeed BCa is considered as a smoking-related cancer [61]. This risk was attributed to many compounds of tobacco such as 4-Aminobiphenyl, 3-Amino-1,4-dimethyl-5H-pyrido [4,3-b] indole (Trp-P-1), Toluene, Benzo[a]pyrene, Benzene… [62]. In the other hand CYP1A2 is an enzyme responsible for the metabolism of caffeine and some tobacco compounds. Rs762551 variation in CYP1A2 gene encodes for the CYP1A2*1F allele. The baseline activity of the enzyme is similar in CYP1A2*1F allele carriers and non-carriers. Moreover it has been reported that he presence of rs762551 (A) codes for the "high inducibility" form the CYP1A2 enzyme, characterized by higher enzyme activity in the presence of an inducer such as smoking or heavy coffee consumption [63]. To explain our result we have reanalyzed the CYP1A2 C > A genotype distribution between cases and controls and according to tobacco status. As result we have found that the inheritance of CYP1A2 (CC) or CYP1 A2 (CA) genotype were respectively associated with 6.89 and 9.04 -fold increased risk of BCa in only heavy smokers patients (≥ 20PY) compared to heavy smokers controls. However we have not data about coffee consumption.

Conclusion
We have conducted the rst study in Tunisian population to evaluate systematically the association between genetic variations in BCa and environmental factors. We also determined the effect of studied pathway SNPs in comparison with environmental exposition. Once validated, these ndings may provide urologists additional genetic information that may help for clinical assessment and treatment decisions. Nevertheless, the underlying mechanisms through which these genes or SNPs affect the clinical behavior of BCas require further studies. Future investigations in our populations and detailed functional characterization are needed to establish predictive or prognostic markers for BCa.  -Informed consent was obtained from all participants (in both languages: Arabic/French) prior to enrollment and participation in this study..

Abbreviations
-All samples are coded and no patient names appear in the study. All data whether clinical or personal remains anonymous and secret.

Consent for publication
Not applicable. This study did not use identifying images and any clinical and personal details despite written informed consent for publication of their clinical details and/or clinical Availability of data and material The dataset is available upon reasonable request to the corresponding author Overview of data analysis work ow Run2 Report for Auto user CIRC-PROTON-121-TERT Imen2 100418 299