Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Mutational analysis of field cancerization in bladder cancer

Mutational analysis of field cancerization in bladder cancer bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 1 Mutational analysis of field cancerization in bladder cancer 1,2 1 1 1,2 3 Trine Strandgaard , Iver Nordentoft , Philippe Lamy , Emil Christensen , Mathilde 1 2,3 1,2* 4 Borg Houlberg Thomsen , Jørgen Bjerggaard Jensen , and Lars Dyrskjøt 6 Department of Molecular Medicine, Aarhus University Hospital, 8200 Aarhus N, 7 Denmark 8 Department of Clinical Medicine, Health, Aarhus University, 8000 Aarhus C, Denmark 9 Department of Urology, Aarhus University Hospital, 8200 Aarhus N, Denmark 11 *Corresponding author: Correspondence and requests for materials should be 12 addressed Lars Dyrskjøt, PhD, Department of Molecular Medicine, Aarhus University 13 Hospital, Denmark. Email: lars@clin.au.dk. bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 14 The multifocal and recurrent nature of bladder cancer has been explained by field 15 cancerization of the bladder urothelium. To shed light on field cancerization in the 16 bladder, we investigated the mutational landscape of normal appearing urothelium and 17 paired bladder tumors from four patients. Sequencing of 509 cancer driver genes 18 revealed the presence of 2-16 mutations exclusively localized in normal tissue 19 (average target read depth 634x). Furthermore, 6-13 mutations were shared between 20 tumor and normal samples and 8-75 mutations were exclusively detected in tumor 21 samples. More mutations were observed in normal samples from patients with 22 multifocal disease compared to patients with unifocal disease. Mutations in normal -16 23 samples had low allele frequencies compared to tumor mutations (p<2.2*10 ). 24 Furthermore, significant differences in the type of nucleotide changes between tumor, -8 25 normal and shared mutations (p=2.7*10 ) were observed, and mutations in APOBEC 26 context were observed primarily among tumor mutations (p=0.026). No differences in 27 functional impact between normal, shared and tumor mutations were observed 28 (p=0.23). Overall, these findings support the theory of multiple fields in the bladder, 29 and document non-tumor specific driver mutations to be present in normal appearing 30 bladder tissue. bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 31 Introduction 32 By applying whole exome sequencing and deep targeted sequencing on bladder 33 tumors, it was recently shown that tumors developed years apart in the same patients 1–3 34 share multiple mutations and hence are clonally related . Furthermore, apparently 35 normal urothelium has been documented to contain mutations with low allele 36 frequencies (~3%) that are typically observed at high frequencies in tumors (clonal 1–3 37 mutations) . Multiple studies have investigated genomic alterations in normal 38 appearing bladder tissue from cystectomy specimens, however using technologies 39 that do not allow detection of low-frequency mutations. The genomic alterations 40 observed in these studies include copy number alterations of chromosome 5, 9, 13, 4–9 41 16, and 17 as well as mutations or loss of RB1 and TP53 . These findings 42 corroborates the suggestions of the presence of field cancerization in the bladder. 43 Similar results have been reported in other tissue types, where studies have revealed 44 the presence of mutations in well-characterized cancer driver genes in apparently 10–13 45 healthy tissue and pre-cancer lesions . 47 Bladder cancer (BC) is multifocal in almost half of the cases with primary tumour and 48 in more than 50% of the patients with recurrent non-muscle invasive BC (NMIBC) . 49 Moreover, recurrent BC is common as the majority of the patients with non-muscle 15,16 50 invasive BC (NMIBC) relapse within five years . Approximately 75% of patients with 51 BC present with NMIBC, and 5-25% of these will progress to muscle-invasive bladder 16,17 52 cancer (MIBC) . Multifocality and the frequent recurrences of BC are hypothesized 53 to originate from field cancerization of the bladder urothelium . This concept was first 54 described in oral squamous epithelium in 1953 by Slaughter et al. as an explanation 55 of the high local recurrence rate of oral cancers . More recent, field cancerization has 56 been described as an underlying mechanism for tumor development in various cancer 57 types, including BC . 59 Field cancerization is understood as one or more areas, or fields, with mutated cells. 60 Normal cell lineages acquire mutations that are positively selected for in the 61 microenvironment of an otherwise healthy organ. Consequently, the mutant clone can 62 grow to produce fields of a monoclonal origin that predispose to malignant growth 63 within these transformed areas. The transformed cells may appear normal or 20,21 64 dysplastic . Thomsen et al proposed a theory of multiple fields being present in the bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 65 bladder where parallel expansion of different mutated stem cells might lead to multiple 66 transformed fields intermixed in the bladder urothelium. Tumors will mirror the genetic 67 alterations from the field from which it arose. This theory may explain the low 68 frequencies of mutations observed in normal samples . 70 In our previous study of bladder cancer field cancerization we analyzed mutations in 71 adjacent normal tissue restricted to mutations observed in the tumor samples, and 72 consequently, non-tumor specific mutations were not investigated . In this study, we 73 characterized mutations in normal appearing urothelium adjacent to tumors by deep 74 targeted sequencing. We detected high-impact mutations in known driver genes that 75 were not observed in the tumor. Furthermore, we observed mutations shared between 76 tumor and normal samples (tumor field effect) as well as mutations specific to the 77 tumors (mutations acquired later in development). 79 Results 80 We performed deep-targeted sequencing of DNA obtained from four patients (patients 81 1 to 4) with advanced bladder cancer, treated with radical cystectomy (see 82 Supplementary Fig. S2 and Supplementary Table S3 for detailed disease courses). 83 From each patient, DNA was procured from bulk tumor biopsies (n=2-7) and laser 84 microdissected (LMD) biopsies of normal appearing urothelium (n=6-11) (See 85 Supplementary Table S1 for overview of samples and sequencing information). 86 Individual bulk tumor samples were previously analyzed by whole exome sequencing 87 (WES) followed by deep targeted amplicon sequencing of LMD tumor and normal 88 samples guided by the original WES of bulk tumor . In this present study, we expand 89 on our previous study to include the analysis of mutations uniquely present in normal 90 appearing adjacent tissue by deep targeted sequencing (Figure 1a). 92 Deep targeted sequencing. Extracted DNA from tumors and LMD normal samples 93 was pooled resulting in one pool of tumor DNA (tumor pool) and one pool of normal 94 DNA (normal pool) from each of the four patients. We performed deep targeted 95 amplicon sequencing of 509 cancer genes on both pools and on matched leukocyte 96 DNA as reference. We obtained an average target read depth of 634x (range: 360- 97 1073). Following sequence read consolidation (UID error correction) the average 98 target read depth was 69x (range: 36-129). In total, after filtering, we identified 30-93 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 99 mutations in the samples from the four patients. Of these, 2-16 were unique for pools 100 of normal samples (N-Mutations), 8-75 were unique for tumor pools (T-Mutations), and 101 6-13 were shared between tumor pools and normal pools (S-Mutations)(Figure 1b). 103 Analysis of field cancerization. Patients 1 and 2 presented with multifocal disease, 104 whereas patients 3 and 4 had unifocal disease. In patients 1 and 2, 39% (25/64) of the 105 mutations were N-Mutations, and 34% (22/64) were S-Mutations. Mutations called in 106 patients 3 and 4 were mainly T-Mutations, with only 5% (7/143) being N-Mutations and 107 13% (19/143) S-Mutations - indicating that uni- and multifocal patients may show 108 different levels of field cancerization. Mutations in known BC driver genes were 109 detected in both N-, S- and T-Mutation groups, most of them being among T- 110 Mutations. However, in patient 1, two N-mutations were observed in bladder cancer 111 driver genes. Damaging mutations were present in all N-, S- and T-Mutation groups. 112 We detected the introduction of premature stop codons, mainly in the T-Mutation 113 group. However, for patient 1 premature stop codons were solely observed within the 114 N- and S-Mutations. Mutation allele frequencies (AFs) varied for the different 115 mutations detected but were generally low for N-Mutations and high for T-Mutations. 116 See Figure 1b and Table 1 for details. 118 Interestingly, we observed N-Mutations in genes known to have a role in cancer 119 development. To corroborate our findings, we investigated the genes affected by non- 120 synonymous mutations in 1889 patients with a total of 1934 samples from 11 different 121 BC studies using cBioPortal. In total, 0.6% to 23% (mean 4%) of the bladder tumors 122 harbored mutations in the same set of genes. The six most frequently non- 123 synonymous mutated N-Mutation genes in the BC datasets were KMT2D (23%), 124 SPTA1 (8%), TRRAP (7%), PRKDC (6%), POLE (4%), and KDM5A (4%). 126 Validation of mutations by WES and ddPCR. Validation of mutations was performed 127 in a two-step process. Firstly, WES data of tumor samples was used to validate 128 mutations detected by our deep targeted sequencing approach. In general, we 129 observed consistency in AFs measured by the two platforms, and most positions were -16 130 covered across all samples (Spearman correlation=0.77, p-val=2.2*10 ) (Figure 2a 131 and 2b). bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 133 Secondly, we used ddPCR to validate the presence/absence of selected alterations in 134 normal and tumor samples. Eight mutations previously observed in tumor and normal 135 samples and three additional N-Mutations were chosen for ddPCR validation. For 136 every patient, tumor mutations were analyzed by ddPCR in 6-11 samples from the 137 normal appearing urothelium. Except for a deletion in RBM10, the tumor alterations 138 were detected at low frequencies in normal samples (Figure 2c). AFs from ddPCR 139 were compared to deep targeted amplicon sequencing of the same samples and a 140 correlation coefficient of 0.93 was observed. For N-Mutation analysis, DNA extracted 141 from 4-7 tumor areas were analyzed and none of the mutations were detected in any 142 of the tumor samples analyzed by ddPCR (Figure 2d), which validated the normal 143 tissue specificity 145 Analysis of mutational context. We performed a combined analysis of the mutations 146 detected in the four patients as the individual patients harbored too few mutations for 147 robust statistical analyses. We observed a significant difference in the six single-base -8 148 substitutions between the three groups of mutations (p=2.7*10 , Fisher’s Exact Test): 149 58% of T-Mutations were C>T changes compared to 40% of both N- and S-Mutations. 150 Furthermore, we observed no T>G mutations in N-Mutations, whereas 40% of S- 151 Mutations and 1.5% of T-Mutations were T>G base pair substitutions. C>G mutations 152 were present among N-Mutations and T-Mutations at 25% and 22% frequency, 153 respectively, compared to 3% in S-Mutations (Figure 3a). C>T mutations have been 154 associated with various signatures, including the age-dependent signature 1 and the 155 APOBEC-related signature 2. C>G substitutions have been attributed to signature 13 13,23–25 156 (APOBEC related), which is commonly observed in BC . 157 We observed no difference in the functional impact of the mutations observed in the 158 three mutation categories. This was observed both when assessing mutations 159 categorized as being of high, moderate, or low/modifier impact by the SNPEff software 160 (p=0.23, Fisher’s Exact Test), and when analyzing synonymous and non-synonymous 161 mutations (p=0.77, Fisher’s Exact Test) (Figure 3b and 3c). 162 Next, we assessed the proportion of APOBEC related mutagenesis. C>T/G mutations 163 in a TCW context, where W is either T or A, were evaluated as representing the 164 APOBEC signature . We observed a significant difference between the proportion of 165 N-, S-, and T-Mutations in APOBEC related context (p=0.0011, Fisher’s Exact Test). 166 In addition, we observed a significant difference when comparing C>T/C>G mutations bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 167 in an APOBEC-related context and C>T/C>G in non-APOBEC related context in N-, 168 S-, and T-Mutations (p=0.026, Fisher’s Exact Test) (Figure 3d). 169 Finally, AFs for mutations in normal samples were significantly lower than for -16 170 mutations in tumor samples (p<2.2*10 , Unpaired T-test)(Figure 3d). There was no 171 significant difference between AFs for T-Mutations and S-Mutations measured in the 172 tumor pool (p=0.09, Unpaired T-test). 174 Discussion 175 Here we characterized the field cancerization in four patients with advanced BC and 176 addressed the question of multiple mutated fields being present within the bladder. 177 Field cancerization was observed in all four patients analyzed, being more pronounced 178 in patients with multifocal disease compared to patients with unifocal disease. We 179 found that the normal appearing urothelium harbored private mutations not detected 180 in the tumor samples. We suggest that these mutations represent one or more fields 181 that have not lead to tumor development. Additionally, we detected mutations that 182 were shared between normal and tumor samples, representing mutations from the 183 field developing into a tumor. Mutations unique for tumor samples were also present, 184 indicating further genomic evolution of the tumor after initial development from the 185 field. 187 Different origins of these mutated cells have been proposed . These include 188 intraepithelial migration and/or luminal seeding of carcinoma cells from existing tumors 189 followed by implantation of the carcinoma cells – eventually giving rise to recurrent 190 tumors. Another theory is that the field develops before the tumor from an altered stem 191 cell embedded in the urothelium. Following this, the altered clone can expand, leading 20,21 192 to a population of mutated daughter cells forming a cancerized field . 194 Our analysis showed that mutations were present at low frequencies in the normal 195 appearing samples. This could be explained by the seeding of tumor cells from existing 196 tumors, resulting in the presence of a few mutated tumor cells in normal samples. Also, 197 it could be due to some tumor cells migrating through the epithelial layer . However, 198 these explanations do not explain the presence of mutations unique for the normal 199 samples. Therefore, another possible explanation for the presence of low frequency 200 mutations in normal samples is that a few mutated cells are intermixed either with bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 201 normal cells or with other differently transformed cells. Different mutated cell lineages 202 can arise if more self renewing cells (e.g. stem cells) are mutated in different ways and 2,27,28 203 expand in parallel, creating multiple transformed fields . This theory may explain 204 the presence of normal specific mutations. If recurrent tumors develop from fields that 205 arose from the same mutated stem cell, these will be clonally related . This could 206 hence explain the clonal origin of metachronous bladder tumors as well as paired 207 upper tract and bladder urothelial tumors . 11,13 209 Two studies from Martincorena et al. have revealed the presence of non-tumor 210 specific mutations in normal tissue from esophagus and skin, respectively. These 211 results indicate that the field arise prior to eventual tumor development and that normal 212 cells harbor mutations without necessarily developing into a tumor. To our knowledge, 213 no previously published studies have focused on mutations in normal appearing 214 bladder tissue without being restricted to mutations observed in the tumor. Our study 215 was performed on normal appearing bladder tissue for non-tumor guided detection of 216 mutations. In order to detect these low-frequency mutations in normal samples, it is 217 necessary to perform deep sequencing. Furthermore, to differentiate low frequency 218 mutations from common sequencing errors, error correction methods, such as the 219 inclusion of UIDs , should be included in the sequencing and subsequent analyses. 221 We observed that the expected impact of N-, S- and T-Mutations was the same across 222 all three groups. We would expect S-Mutations and T-Mutations to have a higher 223 impact than N-Mutations, as these two groups drive initial tumor formation and later 224 tumor evolution. In the Martincorena et al studies, high impact mutations, missense 225 mutations, and cancer driver mutations were observed in normal tissue from non- 11,13 226 cancerous individuals . Consequently, these findings may imply that tumor 227 formation is more dependent on the affected genes, combination of genes, and the 228 order in which mutations occur . Additionally, from our analysis it is not possible to 229 know how many mutations are present in the individual cell, and future studies utilizing 230 single cell sequencing are needed to delineate the genomic changes per cell. 232 In addition, we observed that mutations in APOBEC context were mainly present in 233 the T-Mutation group. This is in concordance with other studies that have suggested 31,32 234 that APOBEC mediated mutagenesis is a late event in tumor evolution . bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 235 Furthermore, most of the non-APOBEC related C>T mutations observed in the normal 236 samples were found in a CpG context (7/11) and may hence be related to the age 237 related signature 1, in accordance with the fact that mutations accumulate in normal 238 cells over time . 240 We hypothesize that field cancerization may have prognostic and predictive value. 241 However, as stated previously, results from our and other studies have shown that 242 mutations do indeed occur in normal cells without leading to cancer development. This 243 may affect screening initiatives for early detection of cancer using e.g. analysis of 244 mutated DNA in urine and plasma. Detection of high impact mutations might not imply 245 that patients have cancer. A recent study detected mutations in cfDNA from individuals 246 without cancer, documenting the need for using tumor guided approaches . 248 In conclusion, this study sheds light on the field cancerization in BC, and documents 249 that non-tumor specific mutations are present in normal appearing tissue. It will be 250 necessary to analyze tissue from additional patients to be able to better describe the 251 field cancerization and its role in tumor development, disease recurrence and 252 aggressiveness, and e.g. BCG treatment efficacy. Moreover, novel methods for single 253 cell analysis may be powerful supplements to better understand the biology of field 254 cancerization. 256 Patients and methods 257 Clinical samples. Patients included in the study were diagnosed with primary BC and 258 underwent open radical cystectomy and extended lymph node dissection to the aortic 259 bifurcation. The patients had not received neoadjuvant chemotherapy or radiation TM 260 therapy before cystectomy. Tissue biopsies were embedded in TissueTek OCT 261 Compound (Sakura, Finetek, Vaerloese, Denmark), snap-frozen in liquid nitrogen and 262 stored at -80 C. Two to seven biopsies were obtained from tumors from each patient 263 together with six to 12 biopsies taken throughout the normal appearing urothelium. 264 Blood samples were stored in EDTA tubes at -80 C. Areas of tumor and normal 265 urothelium were LMD for all patients to ensure cell content specificity of the samples. 266 LMD and DNA extraction from bulk and LMD samples and blood samples were 267 performed as described previously . Patients were treated at Aarhus University 268 Hospital in 2014 and provided informed written consent. The study was approved by bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 269 The Danish National Committees on Health Research Ethics (#1300174). All methods 270 in the study were carried out in accordance with the approved guidelines and 271 regulations. 273 Targeted sequencing and data processing. Targeted sequencing was performed 274 on pools of normal samples and pools of tumor samples using the NuGEN Ovation® 275 Cancer Panel 2.0 Target Enrichment System (509 genes; NuGEN Technologies) . 276 DNA from normal samples and tumor samples from each patient was pooled prior to 277 library generation in order to obtain enough input material. Tumor pools for all patients 278 consisted of 1:1 amounts of bulk tumor DNA. Libraries were prepared from 500 ng 279 DNA (Qubit), as previously described . Libraries were amplified using 21 PCR cycles 280 and subsequently pooled eight at a time and single-end sequenced (150 bp) on an 281 Illumina NextSeq 500 (High output). 282 Sequencing data was aligned and mapped, as previously described . In brief, reads 283 with identical UIDs and mapping positions were collapsed to create high confidence 284 consensus reads. If less than three reads shared UIDs and mapping positions, they 285 were discarded. Mutations were called using MuTect2. 286 Mutations identified in pools of normal samples and/or pools of tumor samples were 287 assessed using bam-readcount in previously generated WES data. WES data was 288 obtained from tumor and leukocyte samples from the same patients and processed as 1,2 289 previously described . Moreover, mutations identified in pools of normal samples 290 were assessed in the associated pools of tumor samples and vice versa. 292 Filtering of mutations. Initially, mutations were categorized in three different sets 293 based on whether they were called (MuTect2) or observed (pileup tools) only in normal 294 samples (Normal specific mutations - N-Mutations), only in tumor samples (Tumor 295 specific mutations or T-Mutations) or in both pools (Shared mutations or S-Mutations) 296 using the cancer panel sequencing (Supplementary Fig. S1). To ensure normal 297 sample specificity, initial N-Mutations were evaluated in previously generated WES 298 data. Mutations were discarded if present with two or more alternate reads in any of 299 the corresponding tumor samples. 300 Any positions with more than two alleles were excluded and all remaining mutations 301 were reviewed manually using the Integrative Genomics Viewer (IGV) . bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 303 Functional assignment. We identified mutations in known BC driver genes defined 304 in IntOGen (BBGLab) and assigned the functional impact to mutations using 37,38 305 PolyPhen-2 and snpEff v4.3 . 307 Digital Droplet PCR (ddPCR). For the validation of N-Mutations, an oligo covering 308 the whole mutated amplicon of interest (positive control) was designed due to 309 insufficient sample amounts. ddPCR and data analysis were performed as previously 310 described . Assays targeting regions on chromosome 16 and 3 were used for 311 quantification of total DNA copies as these regions are rarely subject to copy number 312 alterations in BC . Primer and probe sequences are listed in Supplementary Table 313 S2. 315 Statistical analysis. The Shapiro-Wilk test or Quantile-Quantile plot (QQ-plot) was 316 used to test for normality of the data. Statistical analyses were performed using 317 unpaired t-test on log-transformed parametric data with Welch correction for data with 318 significantly different standard deviations. For categorical variables, Fisher’s Exact test 319 was used. Correlation was calculated using Spearman. Statistical significance was set 320 at p<0.05. All statistical analyses were performed using R (R version 3.5.1). 322 Data availability 323 The raw sequencing datasets generated during the current study are not publicly 324 available due to local Danish legislation on data sharing. However, processed 325 datasets are available from the corresponding author on reasonable request. 327 References 328 1. Lamy, P. et al. Paired Exome Analysis Reveals Clonal Evolution and Potential 329 Therapeutic Targets in Urothelial Carcinoma. Cancer Res. 76, 5894–5906 330 (2016). 331 2. Thomsen, M. B. H. et al. Comprehensive multiregional analysis of molecular 332 heterogeneity in bladder cancer. Sci. Rep. 7, 11702 (2017). 333 3. Nordentoft, I. et al. Mutational context and diverse clonal development in early 334 and late bladder cancer. Cell Rep. 7, 1649–1663 (2014). 335 4. Chaturvedi, V. et al. Superimposed histologic and genetic mapping of 336 chromosome 17 alterations in human urinary bladder neoplasia. Oncogene 14, 337 2059–2070 (1997). 338 5. Czerniak, B. et al. Superimposed histologic and genetic mapping of bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 339 chromosome 9 in progression of human urinary bladder neoplasia: implications 340 for a genetic model of multistep urothelial carcinogenesis and early detection of 341 urinary bladder cancer. Oncogene 18, 1185–1196 (1999). 342 6. Czerniak, B. et al. Genetic modeling of human urinary bladder carcinogenesis. 343 Genes Chromosomes Cancer 27, 392–402 (2000). 344 7. Kram, A. et al. Mapping and genome sequence analysis of chromosome 5 345 regions involved in bladder cancer progression. Lab. Invest. 81, 1039–1048 346 (2001). 347 8. Majewski, T. et al. Understanding the development of human bladder cancer by 348 using a whole-organ genomic mapping strategy. Lab. Invest. 88, 694–721 349 (2008). 350 9. Yoon, D. S. et al. Genetic mapping and DNA sequence-based analysis of 351 deleted regions on chromosome 16 involved in progression of bladder cancer 352 from occult preneoplastic conditions to invasive disease. Oncogene 20, 5005– 353 5014 (2001). 354 10. Weaver, J. M. J. et al. Ordering of mutations in preinvasive disease stages of 355 esophageal carcinogenesis. Nat. Genet. 46, 837–843 (2014). 356 11. Martincorena, I. et al. Tumor evolution. High burden and pervasive positive 357 selection of somatic mutations in normal human skin. Science 348, 880–886 358 (2015). 359 12. Wood, H. M. et al. The clonal relationships between pre-cancer and cancer 360 revealed by ultra-deep sequencing. J. Pathol. 237, 296–306 (2015). 361 13. Martincorena, I. et al. Somatic mutant clones colonize the human esophagus 362 with age. Science (2018). doi:10.1126/science.aau3879 363 14. Ströck, V. & Holmäng, S. A Prospective Study of the Size, Number and 364 Histopathology of New and Recurrent Bladder Tumors. Urology Practice 2, 260– 365 264 (2015). 366 15. Sylvester, R. J. et al. Systematic Review and Individual Patient Data Meta- 367 analysis of Randomized Trials Comparing a Single Immediate Instillation of 368 Chemotherapy After Transurethral Resection with Transurethral Resection Alone 369 in Patients with Stage pTa-pT1 Urothelial Carcinoma of the Bladder: Which 370 Patients Benefit from the Instillation? Eur. Urol. 69, 231–244 (2016). 371 16. Kamat, A. M., Bağcıoğlu, M. & Huri, E. What is new in non-muscle-invasive 372 bladder cancer in 2016? Turk J Urol 43, 9–13 (2017). 373 17. Babjuk, M. et al. EAU Guidelines on Non–Muscle-invasive Urothelial Carcinoma 374 of the Bladder: Update 2016. Eur. Urol. 71, 447–461 (2017). 375 18. Höglund, M. Bladder cancer, a two phased disease? Semin. Cancer Biol. 17, 376 225–232 (2007). 377 19. Slaughter, D. P., Southwick, H. W. & Smejkal, W. ‘Field cancerization’ in oral 378 stratified squamous epithelium. Clinical implications of multicentric origin. Cancer 379 6, 963–968 (1953). 380 20. Curtius, K., Wright, N. A. & Graham, T. A. An evolutionary perspective on field 381 cancerization. Nat. Rev. Cancer 18, 19–32 (2018). 382 21. Braakhuis, B. J. M., Tabor, M. P., Kummer, J. A., Leemans, C. R. & Brakenhoff, 383 R. H. A genetic explanation of Slaughter’s concept of field cancerization: 384 evidence and clinical implications. Cancer Res. 63, 1727–1730 (2003). 385 22. Christensen, E. et al. Optimized targeted sequencing of cell-free plasma DNA 386 from bladder cancer patients. Sci. Rep. 8, 1917 (2018). 387 23. Alexandrov, L. B. et al. Clock-like mutational processes in human somatic cells. 388 Nat. Genet. 47, 1402–1407 (2015). bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 389 24. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. 390 Nature 500, 415–421 (2013). 391 25. Ju, Y. S. The mutational signatures and molecular alterations of bladder cancer. 392 Transl. Cancer Res. 6, S689–S701 (2017). 393 26. Roberts, S. A. et al. An APOBEC cytidine deaminase mutagenesis pattern is 394 widespread in human cancers. Nat. Genet. 45, 970–976 (2013). 395 27. Höglund, M. On the origin of syn- and metachronous urothelial carcinomas. Eur. 396 Urol. 51, 1185–93; discussion 1193 (2007). 397 28. Pardal, R., Clarke, M. F. & Morrison, S. J. Applying the principles of stem-cell 398 biology to cancer. Nat. Rev. Cancer 3, 895–902 (2003). 399 29. Audenet, F. et al. Clonal Relatedness and Mutational Differences between 400 Upper Tract and Bladder Urothelial Carcinoma. Clin. Cancer Res. (2018). 401 doi:10.1158/1078-0432.CCR-18-2039 402 30. Levine, A. J., Jenkins, N. A. & Copeland, N. G. The Roles of Initiating Truncal 403 Mutations in Human Cancers: The Order of Mutations and Tumor Cell Type 404 Matters. Cancer Cell 35, 10–15 (2019). 405 31. Jamal-Hanjani, M. et al. Tracking the Evolution of Non-Small-Cell Lung Cancer. 406 N. Engl. J. Med. 376, 2109–2121 (2017). 407 32. Jolly, C. & Van Loo, P. Timing somatic events in the evolution of cancer. 408 Genome Biol. 19, 95 (2018). 409 33. Phallen, J. et al. Direct detection of early-stage cancers using circulating tumor 410 DNA. Sci. Transl. Med. 9, (2017). 411 34. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows- 412 Wheeler transform. Bioinformatics 25, 1754–1760 (2009). 413 35. Robinson, J. T., Thorvaldsdóttir, H., Wenger, A. M., Zehir, A. & Mesirov, J. P. 414 Variant Review with the Integrative Genomics Viewer. Cancer Res. 77, e31–e34 415 (2017). 416 36. Gonzalez-Perez, A. et al. IntOGen-mutations identifies cancer drivers across 417 tumor types. Nat. Methods 10, 1081–1082 (2013). 418 37. Schepeler, T. et al. A high resolution genomic portrait of bladder cancer: 419 correlation between genomic aberrations and the DNA damage response. 420 Oncogene 32, 3577–3586 (2013). 421 38. Cingolani, P. et al. A program for annotating and predicting the effects of single 422 nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila 423 melanogaster strain w1118; iso-2; iso-3. Fly 6, 80–92 (2012). 424 39. Birkenkamp-Demtröder, K. et al. Monitoring Treatment Response and Metastatic 425 Relapse in Advanced Bladder Cancer by Liquid Biopsy Analysis. Eur. Urol. 73, 426 535–540 (2018). 428 Acknowledgements 429 This work was funded by the Danish Cancer Research Foundation, the Danish Cancer 430 Society, and the Danish Cancer Biobank (grant numbers are not applicable). 432 Contributions 433 M.B.H.T. and J.B.J. collected patient material. T.S. performed clinical follow up. 434 M.B.H.T. performed LMD experiments and DNA-extraction. T.S. and I.N. performed bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 435 experimental work. T.S., E.C., and P.L. performed bioinformatic analyses. L.D., I.N., 436 P.L., and T.S. designed the study and interpreted data. T.S. drafted the manuscript 437 with input from all authors. 438 Competing interests 439 No authors have competing interests in this study. 440 Corresponding author 441 Correspondence to Lars Dyrskjøt. 442 Table 1: Patient 1 Patient 2 Patient 3 Patient 4 Table 1: Analysis of mutations Focality Multifocal Multifocal Unifocal Unifocal T Stage (Clinical) T3b T3b T3b T2b Grade High High High High N Status 1 0 0 0 Total number of mutations 30 34 50 93 No. N-Mutations 16 (53%) 9 (26%) 2 (4%) 5 (5%) 8 (27%) 14 (41%) 42 (84%) 75 (81%) No. T-Mutations No. S-Mutations 6 (20%) 11 (32%) 6 (12%) 13 (14%) Mutated bladder cancer driver genes N-Mutations BCOR, TBX3 - - - T-Mutations BAP1, TP53 CDKN1A, CHEK2, BRCA1, FAT1, HSP90AA1, APC NOTCH1, CDH1, KDM6A, CDKN1A, TBX3, NRAS TBX3, MAP3K1, FBXW7, GNAS S-Mutations FGFR3, EP300, IRS2 TBX3 EP300 BAP1 Premature stop codons N-Mutations TRRAP, EPHB4 - - - T-Mutations - CHEK2 NF2, CDH1 BIRC3, HSP90AA1, KDM6A, RPTOR, EPHA5 S-Mutations BAP1 - - - Allele frequencies (median (min-max)) N-Mutations 0.042 (0.029-0.091) 0.067 (0.035-0.15) 0.070 (0.049-0.091) 0.059 (0.049-0.091) T-Mutations 0.23 (0.049-0.31) 0.13 (0.031-0.40) 0.17 (0.039-0.50) 0.16 (0.032-0.67) S-Mutations (Normal pool) 0.022 (0.0074-0.14) 0.033 (0.0064-0.063) 0.10 (0.014-0.13) 0.025 (0.0065-0.13) 0.20 (0.067-0.44) 0.077 (0.016-0.56) 0.13 (0.057-0.19) 0.15 (0.020-0.61) S-Mutations (Tumor pool) bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 443 Figure legends 444 Figure 1: Analysis of field cancerization in four patients. (a) Study design. Upper 445 part: analyses performed previously. WES was performed on bulk tumor samples. 446 Multiple tumor and normal biopsies were laser microdissected (LMD) and subjected 447 to deep targeted amplicon sequencing guided by the bulk tumor WES. Lower part: 448 present study (black box). Tumor and normal DNA samples were pooled and 449 subjected to deep targeted amplicon sequencing. Mutation calls were analyzed and 450 grouped into T-Mutations, N-Mutations, and S-Mutations. (b) Analysis of patients 1-4. 451 Field cancerization visualized using T-Mutations, N-Mutations, and S-Mutations. Gene 452 names and allele frequencies (AF) are displayed. AFs are illustrated as light grey bars 453 (AF measured in tumor) and dark grey (AF measured in normal). 455 Figure 2: Validation of mutations. (a) All mutations were evaluated in previously 456 generated WES data from tumors, recurrences, and metastases from the four patients 457 (patients 1 and 2 shown, patients 3 and 4 in Supplementary Fig. S3). Obtained AFs 458 are marked (yellow to red ranging from >0 to 0.6). For WES data, a minimum of five 459 reads at a given position were required for validation (indicated in grey). Dark blue 460 indicates no alternate alleles on the position. LN = lymph node. Targ. seq. = Targeted 461 sequencing.(b) AFs obtained by cancer panel sequencing of tumor compared to mean 462 AFs from WES on tumor samples from all four patients. Recurrences and metastases 463 were excluded from calculation of the mean as these samples were not included in the 464 tumor pools. Spearman correlation was calculated. (c) Validation of previously 465 identified tumor mutations by ddPCR on DNA from normal samples. Multiple assays 466 for specific mutations were included for the four patients and the fraction of mutated 467 sequences identified using ddPCR is shown (%). * indicates that the value is out of 468 scale (max value = 14.8%). (d) Validation of absence of N-Mutations in DNA from 469 tumor samples by ddPCR analysis. A positive control (synthesized oligo) for each 470 assay was included as well as negative controls (H2O and HT1197 bladder cancer cell 471 line). The purple line indicates cutoff set for positive droplets. Droplets positive for 472 mutation are marked in blue and negative droplets are indicated by grey. 474 Figure 3: Analysis of mutational context, impact and frequency. All analyses 475 were performed on the combined set of mutations from all patients. The total number bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 476 of mutations in each category is indicated. (a) The six single-base substitutions 477 counted among N-, T-, and S-Mutations. (b) Predicted impact of mutations among N- 478 , T-, and S-Mutations grouped into high, moderate, low/modifier impact. (c) Predicted 479 impact of mutations in N-, T-, and S-Mutations grouped into synonymous and non- 480 synonymous (mutations predicted to have a high or moderate impact) mutations (d) 481 Number of C>G and C>T mutations among N-, T-, and S-Mutations in APOBEC 482 context. (e) Allele frequencies from N-, T-, and S-Mutations. For S-Mutations, allele 483 frequencies are measured both in the normal samples and in the tumor samples and 484 both are indicated. 0.8 0.6 0.4 0.2 0.4 0.3 0.15 0.2 0.1 0.1 0.05 0.5 0.4 0.3 0.2 0.1 0.09 0.3 0.06 0.15 0.03 0 0 0.1 0.05 Figure 1 Tumor Biopsies from Normal Biopsies Appearing Urothelium WES of Bulk Tumor Biopsies Tumor guided Deep Bulk Tumor Samples LMD Normal Samples Targeted Amplicon Sequencing NuGEN Deep Pool of all Tumor Pool of all Normal Targeted Samples Samples Sequencing Mutations unique for Tumor: T-Mutations Mutations unique for Normals: N-Mutations Mutations shared between Tumor and Normal: S-Mutations PolyPhen Probably damaging PolyPhen Possibly damaging PolyPhen Benign IntOGen Bladder Cancer Driver Stop Codon Patient 1 Patient 2 SMAD2 GATA6 CDKN1A* CDX2 KAT6A PAFAH1B2 CHEK2* FGFR3* IRS2* BAP1* APC* EP300* PPARA KMT2D PRKCA BAP1* KALRN MYH9 CBFB EPHA2 GATA2 FANCD2 MAGI2 MPL ERBB2 HDAC2 MCL1 Tumor TP53* FLNC ABL2 ADGRB3 JUN Tumor PIK3CA NBN bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint KALRN Field giving rise to tumor EP400 (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. TRRAP All rights reserved. No reuse allowed without permission. Field giving rise to tumor CARD11 KAT6A PLCG1 PRKDC CCND1 EPHB4 CCND2 Multiple other elds TCF12 MCL1 PML LZTR1 BCOR* Multiple other elds EPHB6 BCORL1 KDM5A BCR EXT2 CDC25A CD79B FANCD2 PLCG2 FES KMT2D BCORL1 EPHB2 SPTA1 TBX3* 16 (53%) Mutations only found in normals (Field Eect Specic Mutations) 9 (26%) Mutations only found in normals (Field Eect Specic Mutations) 6 (20%) Mutations shared between Normals and Tumor 11 (32%) Mutations shared between Normals and Tumor 8 (27%) Mutations only found in Tumor (Tumor Specic Mutations) 14 (41%) Mutations only found in Tumor (Tumor Specic Mutations) RBM10 ATRX IGF1R NF2 MED12 CDKN2B RAD51B BIRC3 FLT1 Patient 3 Patient 4 RBM10 PIK3CG NTRK3 BRCA1* FGFR4 GRIN2A FANCF HSP90AA1*COL1A1 NF2 KDM6A* SOX9 FLNC GRM3 TBX3* FAT1* TPO THBS1 TNK2 PIK3C2B MSH2 NOTCH1* IDH2 BRCA2 CDK8 NTRK3 MAP3K1* BMP4 RAD51B SMAD3 VEGFA JAK3 PTCH1 KDM5A SOX2 CDKN1A* POLE EP300* NTRK3 RAD51B NUP214 INPP4B EGFR IGF1R SUFU ZNF703 IGF2 FOS NF2 FLT1 RUNX1T1 FOXP1 Tumor EPHA3 SF3B1 GRIN2A CBFA2T3 VDR Tumor NUP214 GRIN2A DDR2 PPP2R1A GATA2 Field giving rise CDH1* SPEN ABL1 MST1R CDX2 to tumor APC2 PTPRF ROS1 IRF2 TBX3* Field giving rise to tumor DOT1L FOXO3 ASXL1 JUN TSPAN31 BCL3 CREBBP ADCY9 PLCG1 IGFBP3 FBXW7* PPP2R2B NF1 RET CBL NSD1 CUL3 CDK6 CBFA2T3 BCL2L2 RPTOR LAMA1 Other eld(s) PAK5 MAPK3 DGKB Other eld(s) ERCC3 KEL EPHA8 MYCL EGF EPHA5 ADGRB3 KAT6B EPHA8 MUC1 CDK2 LOC101928120 TCF3 NOS3 RPS6KB1 EGFR TET2 PRKDC ZNF331 NCOA2 NCOA2 WT1 NSD1 APC2 POLE RET FANCA FANCF GNAS* DNMT1 CDH1 PLCG2 KDM5C TBX3* AGAP2 2 (4%) Mutations only found in normals (Field Eect Specic Mutations) 5 (5%) Mutations only found in normals (Field Eect Specic Mutations) NRAS* DGKG 6 (12%) Mutations shared between Normals and Tumor 13 (14%) Mutations shared between Normals and Tumor 42 (84%) Mutations only found in Tumor (Tumor Specic Mutations) 75 (81%) Mutations only found in Tumor (Tumor Specic Mutations) 0.45 0.2 0.3 0.1 0.15 0.15 0.15 0 0 0.6 0.5 0.6 0.4 0.45 0.3 0.3 0.2 0.15 0.1 0.08 0.04 0 Figure 2 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. a b Patient 2 Patient 1 0.7 Gene Targ. seq. WES Targ. seq. Gene WES Rho = 0.77 MCL1 TRRAP -16 P-val < 2.2*10 LZTR1 KAT6A EPHB6 PRKDC KDM5A 0.6 EPHB4 EXT2 CCND2 CD79B TCF12 PLCG2 PML KMT2D BCOR 0.5 EPHB2 BCORL1 PIK3CA BCR MCL1 CDC25A IRS2 FANCD2 GATA2 FES 0.4 ABL2 BCORL1 GATA6 SPTA1 KAT6A TBX3 PPARA CDX2 KALRN FGFR3 0.3 MPL EP300 KALRN BAP1 CDKN1A EPHA2 CHEK2 SMAD2 APC 0.2 PAFAH1B2 PRKCA BAP1 CBFB KMT2D MAGI2 MYH9 HDAC2 FANCD2 FLNC 0.1 ERBB2 JUN TP53 NBN ADGRB3 EP400 CARD11 No a lterna te a llele 0.0 PLCG1 # Rea ds < 5 CCND1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 >0 0.6 AF targeted sequencing c d Patient 1 Patient 2 Patient 1 - DDB2 Patient 1 - EPHB4 Cell Cell Tumor samples Pos. H O Tumor samples Pos. H O 2 2 line line 5.00 5.00 FGFR3 PIK3CA 4.50 4.50 CDKN1A 8 8 STAG2 4.00 4.00 3.50 3.50 3.00 3.00 2.50 2.50 4 4 2.00 2.00 1.50 1.50 1.00 1.00 2 2 0.50 0.50 0.00 0.00 N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N1 N2 N3 N4 N5 N6 N7 N8 N9 Patient 3 Patient 4 Patient 2 - EPHB2 Cell * Tumor samples Pos. H O line 5.00 5.00 RBM10 PFKP C11orf70 CDH11 4.50 4.50 * Max =14.8 % 4.00 4.00 3.50 3.50 3.00 3.00 2.50 2.50 2.00 2.00 1.50 1.50 1.00 1.00 0.50 0.50 0.00 0.00 N1 N2 N3 N4 N5 N6 N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N11 Normal Pool Tumor Pool Tumor 1 Tumor 2 LN metastasis Normal Pool Tumor Pool Tumor 1 Tumor 2 Tumor 3 Tumor 4 Tumor 5 Tumor 6 Tumor 7 Recurrence AF (%) AF (%) AF (%) AF (%) Tumor Shared Normal Tumor Shared Normal AF WES Amplitude x 1,000 Amplitude x 1,000 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. Figure 3 a b c 100% 100% 100% 90% 90% 90% 80% 80% 14 80% 13 13 70% 70% 70% 60% 60% 60% 50% 50% 50% 40% 40% 40% 14 29 30% 30% 3 30% 20% 20% 20% 2 18 10% 10% 10% 3 11 1 4 0% 0% 0% T>A T>C T>G C>A C>G C>T High Moderate Modier/low Synonymous Non-Synonymous d e 100% -16 P-val < 2.2*10 0.8 90% 80% 0.6 70% 9 60% 50% 0.4 40% 30% 0.2 20% 10% 0% C>G/C>T non-APOBEC context N-Mutations T-Mutations C>G/C>T APOBEC context S-Mutations S-Mutations (Normal pool) (Tumor pool) AF http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png bioRxiv bioRxiv

Loading next page...
 
/lp/biorxiv/mutational-analysis-of-field-cancerization-in-bladder-cancer-AlJK8O8cW0

References (54)

Publisher
bioRxiv
Copyright
© 2019, Posted by Cold Spring Harbor Laboratory. The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission.
DOI
10.1101/536466
Publisher site
See Article on Publisher Site

Abstract

bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 1 Mutational analysis of field cancerization in bladder cancer 1,2 1 1 1,2 3 Trine Strandgaard , Iver Nordentoft , Philippe Lamy , Emil Christensen , Mathilde 1 2,3 1,2* 4 Borg Houlberg Thomsen , Jørgen Bjerggaard Jensen , and Lars Dyrskjøt 6 Department of Molecular Medicine, Aarhus University Hospital, 8200 Aarhus N, 7 Denmark 8 Department of Clinical Medicine, Health, Aarhus University, 8000 Aarhus C, Denmark 9 Department of Urology, Aarhus University Hospital, 8200 Aarhus N, Denmark 11 *Corresponding author: Correspondence and requests for materials should be 12 addressed Lars Dyrskjøt, PhD, Department of Molecular Medicine, Aarhus University 13 Hospital, Denmark. Email: lars@clin.au.dk. bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 14 The multifocal and recurrent nature of bladder cancer has been explained by field 15 cancerization of the bladder urothelium. To shed light on field cancerization in the 16 bladder, we investigated the mutational landscape of normal appearing urothelium and 17 paired bladder tumors from four patients. Sequencing of 509 cancer driver genes 18 revealed the presence of 2-16 mutations exclusively localized in normal tissue 19 (average target read depth 634x). Furthermore, 6-13 mutations were shared between 20 tumor and normal samples and 8-75 mutations were exclusively detected in tumor 21 samples. More mutations were observed in normal samples from patients with 22 multifocal disease compared to patients with unifocal disease. Mutations in normal -16 23 samples had low allele frequencies compared to tumor mutations (p<2.2*10 ). 24 Furthermore, significant differences in the type of nucleotide changes between tumor, -8 25 normal and shared mutations (p=2.7*10 ) were observed, and mutations in APOBEC 26 context were observed primarily among tumor mutations (p=0.026). No differences in 27 functional impact between normal, shared and tumor mutations were observed 28 (p=0.23). Overall, these findings support the theory of multiple fields in the bladder, 29 and document non-tumor specific driver mutations to be present in normal appearing 30 bladder tissue. bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 31 Introduction 32 By applying whole exome sequencing and deep targeted sequencing on bladder 33 tumors, it was recently shown that tumors developed years apart in the same patients 1–3 34 share multiple mutations and hence are clonally related . Furthermore, apparently 35 normal urothelium has been documented to contain mutations with low allele 36 frequencies (~3%) that are typically observed at high frequencies in tumors (clonal 1–3 37 mutations) . Multiple studies have investigated genomic alterations in normal 38 appearing bladder tissue from cystectomy specimens, however using technologies 39 that do not allow detection of low-frequency mutations. The genomic alterations 40 observed in these studies include copy number alterations of chromosome 5, 9, 13, 4–9 41 16, and 17 as well as mutations or loss of RB1 and TP53 . These findings 42 corroborates the suggestions of the presence of field cancerization in the bladder. 43 Similar results have been reported in other tissue types, where studies have revealed 44 the presence of mutations in well-characterized cancer driver genes in apparently 10–13 45 healthy tissue and pre-cancer lesions . 47 Bladder cancer (BC) is multifocal in almost half of the cases with primary tumour and 48 in more than 50% of the patients with recurrent non-muscle invasive BC (NMIBC) . 49 Moreover, recurrent BC is common as the majority of the patients with non-muscle 15,16 50 invasive BC (NMIBC) relapse within five years . Approximately 75% of patients with 51 BC present with NMIBC, and 5-25% of these will progress to muscle-invasive bladder 16,17 52 cancer (MIBC) . Multifocality and the frequent recurrences of BC are hypothesized 53 to originate from field cancerization of the bladder urothelium . This concept was first 54 described in oral squamous epithelium in 1953 by Slaughter et al. as an explanation 55 of the high local recurrence rate of oral cancers . More recent, field cancerization has 56 been described as an underlying mechanism for tumor development in various cancer 57 types, including BC . 59 Field cancerization is understood as one or more areas, or fields, with mutated cells. 60 Normal cell lineages acquire mutations that are positively selected for in the 61 microenvironment of an otherwise healthy organ. Consequently, the mutant clone can 62 grow to produce fields of a monoclonal origin that predispose to malignant growth 63 within these transformed areas. The transformed cells may appear normal or 20,21 64 dysplastic . Thomsen et al proposed a theory of multiple fields being present in the bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 65 bladder where parallel expansion of different mutated stem cells might lead to multiple 66 transformed fields intermixed in the bladder urothelium. Tumors will mirror the genetic 67 alterations from the field from which it arose. This theory may explain the low 68 frequencies of mutations observed in normal samples . 70 In our previous study of bladder cancer field cancerization we analyzed mutations in 71 adjacent normal tissue restricted to mutations observed in the tumor samples, and 72 consequently, non-tumor specific mutations were not investigated . In this study, we 73 characterized mutations in normal appearing urothelium adjacent to tumors by deep 74 targeted sequencing. We detected high-impact mutations in known driver genes that 75 were not observed in the tumor. Furthermore, we observed mutations shared between 76 tumor and normal samples (tumor field effect) as well as mutations specific to the 77 tumors (mutations acquired later in development). 79 Results 80 We performed deep-targeted sequencing of DNA obtained from four patients (patients 81 1 to 4) with advanced bladder cancer, treated with radical cystectomy (see 82 Supplementary Fig. S2 and Supplementary Table S3 for detailed disease courses). 83 From each patient, DNA was procured from bulk tumor biopsies (n=2-7) and laser 84 microdissected (LMD) biopsies of normal appearing urothelium (n=6-11) (See 85 Supplementary Table S1 for overview of samples and sequencing information). 86 Individual bulk tumor samples were previously analyzed by whole exome sequencing 87 (WES) followed by deep targeted amplicon sequencing of LMD tumor and normal 88 samples guided by the original WES of bulk tumor . In this present study, we expand 89 on our previous study to include the analysis of mutations uniquely present in normal 90 appearing adjacent tissue by deep targeted sequencing (Figure 1a). 92 Deep targeted sequencing. Extracted DNA from tumors and LMD normal samples 93 was pooled resulting in one pool of tumor DNA (tumor pool) and one pool of normal 94 DNA (normal pool) from each of the four patients. We performed deep targeted 95 amplicon sequencing of 509 cancer genes on both pools and on matched leukocyte 96 DNA as reference. We obtained an average target read depth of 634x (range: 360- 97 1073). Following sequence read consolidation (UID error correction) the average 98 target read depth was 69x (range: 36-129). In total, after filtering, we identified 30-93 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 99 mutations in the samples from the four patients. Of these, 2-16 were unique for pools 100 of normal samples (N-Mutations), 8-75 were unique for tumor pools (T-Mutations), and 101 6-13 were shared between tumor pools and normal pools (S-Mutations)(Figure 1b). 103 Analysis of field cancerization. Patients 1 and 2 presented with multifocal disease, 104 whereas patients 3 and 4 had unifocal disease. In patients 1 and 2, 39% (25/64) of the 105 mutations were N-Mutations, and 34% (22/64) were S-Mutations. Mutations called in 106 patients 3 and 4 were mainly T-Mutations, with only 5% (7/143) being N-Mutations and 107 13% (19/143) S-Mutations - indicating that uni- and multifocal patients may show 108 different levels of field cancerization. Mutations in known BC driver genes were 109 detected in both N-, S- and T-Mutation groups, most of them being among T- 110 Mutations. However, in patient 1, two N-mutations were observed in bladder cancer 111 driver genes. Damaging mutations were present in all N-, S- and T-Mutation groups. 112 We detected the introduction of premature stop codons, mainly in the T-Mutation 113 group. However, for patient 1 premature stop codons were solely observed within the 114 N- and S-Mutations. Mutation allele frequencies (AFs) varied for the different 115 mutations detected but were generally low for N-Mutations and high for T-Mutations. 116 See Figure 1b and Table 1 for details. 118 Interestingly, we observed N-Mutations in genes known to have a role in cancer 119 development. To corroborate our findings, we investigated the genes affected by non- 120 synonymous mutations in 1889 patients with a total of 1934 samples from 11 different 121 BC studies using cBioPortal. In total, 0.6% to 23% (mean 4%) of the bladder tumors 122 harbored mutations in the same set of genes. The six most frequently non- 123 synonymous mutated N-Mutation genes in the BC datasets were KMT2D (23%), 124 SPTA1 (8%), TRRAP (7%), PRKDC (6%), POLE (4%), and KDM5A (4%). 126 Validation of mutations by WES and ddPCR. Validation of mutations was performed 127 in a two-step process. Firstly, WES data of tumor samples was used to validate 128 mutations detected by our deep targeted sequencing approach. In general, we 129 observed consistency in AFs measured by the two platforms, and most positions were -16 130 covered across all samples (Spearman correlation=0.77, p-val=2.2*10 ) (Figure 2a 131 and 2b). bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 133 Secondly, we used ddPCR to validate the presence/absence of selected alterations in 134 normal and tumor samples. Eight mutations previously observed in tumor and normal 135 samples and three additional N-Mutations were chosen for ddPCR validation. For 136 every patient, tumor mutations were analyzed by ddPCR in 6-11 samples from the 137 normal appearing urothelium. Except for a deletion in RBM10, the tumor alterations 138 were detected at low frequencies in normal samples (Figure 2c). AFs from ddPCR 139 were compared to deep targeted amplicon sequencing of the same samples and a 140 correlation coefficient of 0.93 was observed. For N-Mutation analysis, DNA extracted 141 from 4-7 tumor areas were analyzed and none of the mutations were detected in any 142 of the tumor samples analyzed by ddPCR (Figure 2d), which validated the normal 143 tissue specificity 145 Analysis of mutational context. We performed a combined analysis of the mutations 146 detected in the four patients as the individual patients harbored too few mutations for 147 robust statistical analyses. We observed a significant difference in the six single-base -8 148 substitutions between the three groups of mutations (p=2.7*10 , Fisher’s Exact Test): 149 58% of T-Mutations were C>T changes compared to 40% of both N- and S-Mutations. 150 Furthermore, we observed no T>G mutations in N-Mutations, whereas 40% of S- 151 Mutations and 1.5% of T-Mutations were T>G base pair substitutions. C>G mutations 152 were present among N-Mutations and T-Mutations at 25% and 22% frequency, 153 respectively, compared to 3% in S-Mutations (Figure 3a). C>T mutations have been 154 associated with various signatures, including the age-dependent signature 1 and the 155 APOBEC-related signature 2. C>G substitutions have been attributed to signature 13 13,23–25 156 (APOBEC related), which is commonly observed in BC . 157 We observed no difference in the functional impact of the mutations observed in the 158 three mutation categories. This was observed both when assessing mutations 159 categorized as being of high, moderate, or low/modifier impact by the SNPEff software 160 (p=0.23, Fisher’s Exact Test), and when analyzing synonymous and non-synonymous 161 mutations (p=0.77, Fisher’s Exact Test) (Figure 3b and 3c). 162 Next, we assessed the proportion of APOBEC related mutagenesis. C>T/G mutations 163 in a TCW context, where W is either T or A, were evaluated as representing the 164 APOBEC signature . We observed a significant difference between the proportion of 165 N-, S-, and T-Mutations in APOBEC related context (p=0.0011, Fisher’s Exact Test). 166 In addition, we observed a significant difference when comparing C>T/C>G mutations bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 167 in an APOBEC-related context and C>T/C>G in non-APOBEC related context in N-, 168 S-, and T-Mutations (p=0.026, Fisher’s Exact Test) (Figure 3d). 169 Finally, AFs for mutations in normal samples were significantly lower than for -16 170 mutations in tumor samples (p<2.2*10 , Unpaired T-test)(Figure 3d). There was no 171 significant difference between AFs for T-Mutations and S-Mutations measured in the 172 tumor pool (p=0.09, Unpaired T-test). 174 Discussion 175 Here we characterized the field cancerization in four patients with advanced BC and 176 addressed the question of multiple mutated fields being present within the bladder. 177 Field cancerization was observed in all four patients analyzed, being more pronounced 178 in patients with multifocal disease compared to patients with unifocal disease. We 179 found that the normal appearing urothelium harbored private mutations not detected 180 in the tumor samples. We suggest that these mutations represent one or more fields 181 that have not lead to tumor development. Additionally, we detected mutations that 182 were shared between normal and tumor samples, representing mutations from the 183 field developing into a tumor. Mutations unique for tumor samples were also present, 184 indicating further genomic evolution of the tumor after initial development from the 185 field. 187 Different origins of these mutated cells have been proposed . These include 188 intraepithelial migration and/or luminal seeding of carcinoma cells from existing tumors 189 followed by implantation of the carcinoma cells – eventually giving rise to recurrent 190 tumors. Another theory is that the field develops before the tumor from an altered stem 191 cell embedded in the urothelium. Following this, the altered clone can expand, leading 20,21 192 to a population of mutated daughter cells forming a cancerized field . 194 Our analysis showed that mutations were present at low frequencies in the normal 195 appearing samples. This could be explained by the seeding of tumor cells from existing 196 tumors, resulting in the presence of a few mutated tumor cells in normal samples. Also, 197 it could be due to some tumor cells migrating through the epithelial layer . However, 198 these explanations do not explain the presence of mutations unique for the normal 199 samples. Therefore, another possible explanation for the presence of low frequency 200 mutations in normal samples is that a few mutated cells are intermixed either with bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 201 normal cells or with other differently transformed cells. Different mutated cell lineages 202 can arise if more self renewing cells (e.g. stem cells) are mutated in different ways and 2,27,28 203 expand in parallel, creating multiple transformed fields . This theory may explain 204 the presence of normal specific mutations. If recurrent tumors develop from fields that 205 arose from the same mutated stem cell, these will be clonally related . This could 206 hence explain the clonal origin of metachronous bladder tumors as well as paired 207 upper tract and bladder urothelial tumors . 11,13 209 Two studies from Martincorena et al. have revealed the presence of non-tumor 210 specific mutations in normal tissue from esophagus and skin, respectively. These 211 results indicate that the field arise prior to eventual tumor development and that normal 212 cells harbor mutations without necessarily developing into a tumor. To our knowledge, 213 no previously published studies have focused on mutations in normal appearing 214 bladder tissue without being restricted to mutations observed in the tumor. Our study 215 was performed on normal appearing bladder tissue for non-tumor guided detection of 216 mutations. In order to detect these low-frequency mutations in normal samples, it is 217 necessary to perform deep sequencing. Furthermore, to differentiate low frequency 218 mutations from common sequencing errors, error correction methods, such as the 219 inclusion of UIDs , should be included in the sequencing and subsequent analyses. 221 We observed that the expected impact of N-, S- and T-Mutations was the same across 222 all three groups. We would expect S-Mutations and T-Mutations to have a higher 223 impact than N-Mutations, as these two groups drive initial tumor formation and later 224 tumor evolution. In the Martincorena et al studies, high impact mutations, missense 225 mutations, and cancer driver mutations were observed in normal tissue from non- 11,13 226 cancerous individuals . Consequently, these findings may imply that tumor 227 formation is more dependent on the affected genes, combination of genes, and the 228 order in which mutations occur . Additionally, from our analysis it is not possible to 229 know how many mutations are present in the individual cell, and future studies utilizing 230 single cell sequencing are needed to delineate the genomic changes per cell. 232 In addition, we observed that mutations in APOBEC context were mainly present in 233 the T-Mutation group. This is in concordance with other studies that have suggested 31,32 234 that APOBEC mediated mutagenesis is a late event in tumor evolution . bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 235 Furthermore, most of the non-APOBEC related C>T mutations observed in the normal 236 samples were found in a CpG context (7/11) and may hence be related to the age 237 related signature 1, in accordance with the fact that mutations accumulate in normal 238 cells over time . 240 We hypothesize that field cancerization may have prognostic and predictive value. 241 However, as stated previously, results from our and other studies have shown that 242 mutations do indeed occur in normal cells without leading to cancer development. This 243 may affect screening initiatives for early detection of cancer using e.g. analysis of 244 mutated DNA in urine and plasma. Detection of high impact mutations might not imply 245 that patients have cancer. A recent study detected mutations in cfDNA from individuals 246 without cancer, documenting the need for using tumor guided approaches . 248 In conclusion, this study sheds light on the field cancerization in BC, and documents 249 that non-tumor specific mutations are present in normal appearing tissue. It will be 250 necessary to analyze tissue from additional patients to be able to better describe the 251 field cancerization and its role in tumor development, disease recurrence and 252 aggressiveness, and e.g. BCG treatment efficacy. Moreover, novel methods for single 253 cell analysis may be powerful supplements to better understand the biology of field 254 cancerization. 256 Patients and methods 257 Clinical samples. Patients included in the study were diagnosed with primary BC and 258 underwent open radical cystectomy and extended lymph node dissection to the aortic 259 bifurcation. The patients had not received neoadjuvant chemotherapy or radiation TM 260 therapy before cystectomy. Tissue biopsies were embedded in TissueTek OCT 261 Compound (Sakura, Finetek, Vaerloese, Denmark), snap-frozen in liquid nitrogen and 262 stored at -80 C. Two to seven biopsies were obtained from tumors from each patient 263 together with six to 12 biopsies taken throughout the normal appearing urothelium. 264 Blood samples were stored in EDTA tubes at -80 C. Areas of tumor and normal 265 urothelium were LMD for all patients to ensure cell content specificity of the samples. 266 LMD and DNA extraction from bulk and LMD samples and blood samples were 267 performed as described previously . Patients were treated at Aarhus University 268 Hospital in 2014 and provided informed written consent. The study was approved by bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 269 The Danish National Committees on Health Research Ethics (#1300174). All methods 270 in the study were carried out in accordance with the approved guidelines and 271 regulations. 273 Targeted sequencing and data processing. Targeted sequencing was performed 274 on pools of normal samples and pools of tumor samples using the NuGEN Ovation® 275 Cancer Panel 2.0 Target Enrichment System (509 genes; NuGEN Technologies) . 276 DNA from normal samples and tumor samples from each patient was pooled prior to 277 library generation in order to obtain enough input material. Tumor pools for all patients 278 consisted of 1:1 amounts of bulk tumor DNA. Libraries were prepared from 500 ng 279 DNA (Qubit), as previously described . Libraries were amplified using 21 PCR cycles 280 and subsequently pooled eight at a time and single-end sequenced (150 bp) on an 281 Illumina NextSeq 500 (High output). 282 Sequencing data was aligned and mapped, as previously described . In brief, reads 283 with identical UIDs and mapping positions were collapsed to create high confidence 284 consensus reads. If less than three reads shared UIDs and mapping positions, they 285 were discarded. Mutations were called using MuTect2. 286 Mutations identified in pools of normal samples and/or pools of tumor samples were 287 assessed using bam-readcount in previously generated WES data. WES data was 288 obtained from tumor and leukocyte samples from the same patients and processed as 1,2 289 previously described . Moreover, mutations identified in pools of normal samples 290 were assessed in the associated pools of tumor samples and vice versa. 292 Filtering of mutations. Initially, mutations were categorized in three different sets 293 based on whether they were called (MuTect2) or observed (pileup tools) only in normal 294 samples (Normal specific mutations - N-Mutations), only in tumor samples (Tumor 295 specific mutations or T-Mutations) or in both pools (Shared mutations or S-Mutations) 296 using the cancer panel sequencing (Supplementary Fig. S1). To ensure normal 297 sample specificity, initial N-Mutations were evaluated in previously generated WES 298 data. Mutations were discarded if present with two or more alternate reads in any of 299 the corresponding tumor samples. 300 Any positions with more than two alleles were excluded and all remaining mutations 301 were reviewed manually using the Integrative Genomics Viewer (IGV) . bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 303 Functional assignment. We identified mutations in known BC driver genes defined 304 in IntOGen (BBGLab) and assigned the functional impact to mutations using 37,38 305 PolyPhen-2 and snpEff v4.3 . 307 Digital Droplet PCR (ddPCR). For the validation of N-Mutations, an oligo covering 308 the whole mutated amplicon of interest (positive control) was designed due to 309 insufficient sample amounts. ddPCR and data analysis were performed as previously 310 described . Assays targeting regions on chromosome 16 and 3 were used for 311 quantification of total DNA copies as these regions are rarely subject to copy number 312 alterations in BC . Primer and probe sequences are listed in Supplementary Table 313 S2. 315 Statistical analysis. The Shapiro-Wilk test or Quantile-Quantile plot (QQ-plot) was 316 used to test for normality of the data. Statistical analyses were performed using 317 unpaired t-test on log-transformed parametric data with Welch correction for data with 318 significantly different standard deviations. For categorical variables, Fisher’s Exact test 319 was used. Correlation was calculated using Spearman. Statistical significance was set 320 at p<0.05. All statistical analyses were performed using R (R version 3.5.1). 322 Data availability 323 The raw sequencing datasets generated during the current study are not publicly 324 available due to local Danish legislation on data sharing. However, processed 325 datasets are available from the corresponding author on reasonable request. 327 References 328 1. Lamy, P. et al. Paired Exome Analysis Reveals Clonal Evolution and Potential 329 Therapeutic Targets in Urothelial Carcinoma. Cancer Res. 76, 5894–5906 330 (2016). 331 2. Thomsen, M. B. H. et al. Comprehensive multiregional analysis of molecular 332 heterogeneity in bladder cancer. Sci. Rep. 7, 11702 (2017). 333 3. Nordentoft, I. et al. Mutational context and diverse clonal development in early 334 and late bladder cancer. Cell Rep. 7, 1649–1663 (2014). 335 4. Chaturvedi, V. et al. Superimposed histologic and genetic mapping of 336 chromosome 17 alterations in human urinary bladder neoplasia. Oncogene 14, 337 2059–2070 (1997). 338 5. Czerniak, B. et al. Superimposed histologic and genetic mapping of bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 339 chromosome 9 in progression of human urinary bladder neoplasia: implications 340 for a genetic model of multistep urothelial carcinogenesis and early detection of 341 urinary bladder cancer. Oncogene 18, 1185–1196 (1999). 342 6. Czerniak, B. et al. Genetic modeling of human urinary bladder carcinogenesis. 343 Genes Chromosomes Cancer 27, 392–402 (2000). 344 7. Kram, A. et al. Mapping and genome sequence analysis of chromosome 5 345 regions involved in bladder cancer progression. Lab. Invest. 81, 1039–1048 346 (2001). 347 8. Majewski, T. et al. Understanding the development of human bladder cancer by 348 using a whole-organ genomic mapping strategy. Lab. Invest. 88, 694–721 349 (2008). 350 9. Yoon, D. S. et al. Genetic mapping and DNA sequence-based analysis of 351 deleted regions on chromosome 16 involved in progression of bladder cancer 352 from occult preneoplastic conditions to invasive disease. Oncogene 20, 5005– 353 5014 (2001). 354 10. Weaver, J. M. J. et al. Ordering of mutations in preinvasive disease stages of 355 esophageal carcinogenesis. Nat. Genet. 46, 837–843 (2014). 356 11. Martincorena, I. et al. Tumor evolution. High burden and pervasive positive 357 selection of somatic mutations in normal human skin. Science 348, 880–886 358 (2015). 359 12. Wood, H. M. et al. The clonal relationships between pre-cancer and cancer 360 revealed by ultra-deep sequencing. J. Pathol. 237, 296–306 (2015). 361 13. Martincorena, I. et al. Somatic mutant clones colonize the human esophagus 362 with age. Science (2018). doi:10.1126/science.aau3879 363 14. Ströck, V. & Holmäng, S. A Prospective Study of the Size, Number and 364 Histopathology of New and Recurrent Bladder Tumors. Urology Practice 2, 260– 365 264 (2015). 366 15. Sylvester, R. J. et al. Systematic Review and Individual Patient Data Meta- 367 analysis of Randomized Trials Comparing a Single Immediate Instillation of 368 Chemotherapy After Transurethral Resection with Transurethral Resection Alone 369 in Patients with Stage pTa-pT1 Urothelial Carcinoma of the Bladder: Which 370 Patients Benefit from the Instillation? Eur. Urol. 69, 231–244 (2016). 371 16. Kamat, A. M., Bağcıoğlu, M. & Huri, E. What is new in non-muscle-invasive 372 bladder cancer in 2016? Turk J Urol 43, 9–13 (2017). 373 17. Babjuk, M. et al. EAU Guidelines on Non–Muscle-invasive Urothelial Carcinoma 374 of the Bladder: Update 2016. Eur. Urol. 71, 447–461 (2017). 375 18. Höglund, M. Bladder cancer, a two phased disease? Semin. Cancer Biol. 17, 376 225–232 (2007). 377 19. Slaughter, D. P., Southwick, H. W. & Smejkal, W. ‘Field cancerization’ in oral 378 stratified squamous epithelium. Clinical implications of multicentric origin. Cancer 379 6, 963–968 (1953). 380 20. Curtius, K., Wright, N. A. & Graham, T. A. An evolutionary perspective on field 381 cancerization. Nat. Rev. Cancer 18, 19–32 (2018). 382 21. Braakhuis, B. J. M., Tabor, M. P., Kummer, J. A., Leemans, C. R. & Brakenhoff, 383 R. H. A genetic explanation of Slaughter’s concept of field cancerization: 384 evidence and clinical implications. Cancer Res. 63, 1727–1730 (2003). 385 22. Christensen, E. et al. Optimized targeted sequencing of cell-free plasma DNA 386 from bladder cancer patients. Sci. Rep. 8, 1917 (2018). 387 23. Alexandrov, L. B. et al. Clock-like mutational processes in human somatic cells. 388 Nat. Genet. 47, 1402–1407 (2015). bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 389 24. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. 390 Nature 500, 415–421 (2013). 391 25. Ju, Y. S. The mutational signatures and molecular alterations of bladder cancer. 392 Transl. Cancer Res. 6, S689–S701 (2017). 393 26. Roberts, S. A. et al. An APOBEC cytidine deaminase mutagenesis pattern is 394 widespread in human cancers. Nat. Genet. 45, 970–976 (2013). 395 27. Höglund, M. On the origin of syn- and metachronous urothelial carcinomas. Eur. 396 Urol. 51, 1185–93; discussion 1193 (2007). 397 28. Pardal, R., Clarke, M. F. & Morrison, S. J. Applying the principles of stem-cell 398 biology to cancer. Nat. Rev. Cancer 3, 895–902 (2003). 399 29. Audenet, F. et al. Clonal Relatedness and Mutational Differences between 400 Upper Tract and Bladder Urothelial Carcinoma. Clin. Cancer Res. (2018). 401 doi:10.1158/1078-0432.CCR-18-2039 402 30. Levine, A. J., Jenkins, N. A. & Copeland, N. G. The Roles of Initiating Truncal 403 Mutations in Human Cancers: The Order of Mutations and Tumor Cell Type 404 Matters. Cancer Cell 35, 10–15 (2019). 405 31. Jamal-Hanjani, M. et al. Tracking the Evolution of Non-Small-Cell Lung Cancer. 406 N. Engl. J. Med. 376, 2109–2121 (2017). 407 32. Jolly, C. & Van Loo, P. Timing somatic events in the evolution of cancer. 408 Genome Biol. 19, 95 (2018). 409 33. Phallen, J. et al. Direct detection of early-stage cancers using circulating tumor 410 DNA. Sci. Transl. Med. 9, (2017). 411 34. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows- 412 Wheeler transform. Bioinformatics 25, 1754–1760 (2009). 413 35. Robinson, J. T., Thorvaldsdóttir, H., Wenger, A. M., Zehir, A. & Mesirov, J. P. 414 Variant Review with the Integrative Genomics Viewer. Cancer Res. 77, e31–e34 415 (2017). 416 36. Gonzalez-Perez, A. et al. IntOGen-mutations identifies cancer drivers across 417 tumor types. Nat. Methods 10, 1081–1082 (2013). 418 37. Schepeler, T. et al. A high resolution genomic portrait of bladder cancer: 419 correlation between genomic aberrations and the DNA damage response. 420 Oncogene 32, 3577–3586 (2013). 421 38. Cingolani, P. et al. A program for annotating and predicting the effects of single 422 nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila 423 melanogaster strain w1118; iso-2; iso-3. Fly 6, 80–92 (2012). 424 39. Birkenkamp-Demtröder, K. et al. Monitoring Treatment Response and Metastatic 425 Relapse in Advanced Bladder Cancer by Liquid Biopsy Analysis. Eur. Urol. 73, 426 535–540 (2018). 428 Acknowledgements 429 This work was funded by the Danish Cancer Research Foundation, the Danish Cancer 430 Society, and the Danish Cancer Biobank (grant numbers are not applicable). 432 Contributions 433 M.B.H.T. and J.B.J. collected patient material. T.S. performed clinical follow up. 434 M.B.H.T. performed LMD experiments and DNA-extraction. T.S. and I.N. performed bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 435 experimental work. T.S., E.C., and P.L. performed bioinformatic analyses. L.D., I.N., 436 P.L., and T.S. designed the study and interpreted data. T.S. drafted the manuscript 437 with input from all authors. 438 Competing interests 439 No authors have competing interests in this study. 440 Corresponding author 441 Correspondence to Lars Dyrskjøt. 442 Table 1: Patient 1 Patient 2 Patient 3 Patient 4 Table 1: Analysis of mutations Focality Multifocal Multifocal Unifocal Unifocal T Stage (Clinical) T3b T3b T3b T2b Grade High High High High N Status 1 0 0 0 Total number of mutations 30 34 50 93 No. N-Mutations 16 (53%) 9 (26%) 2 (4%) 5 (5%) 8 (27%) 14 (41%) 42 (84%) 75 (81%) No. T-Mutations No. S-Mutations 6 (20%) 11 (32%) 6 (12%) 13 (14%) Mutated bladder cancer driver genes N-Mutations BCOR, TBX3 - - - T-Mutations BAP1, TP53 CDKN1A, CHEK2, BRCA1, FAT1, HSP90AA1, APC NOTCH1, CDH1, KDM6A, CDKN1A, TBX3, NRAS TBX3, MAP3K1, FBXW7, GNAS S-Mutations FGFR3, EP300, IRS2 TBX3 EP300 BAP1 Premature stop codons N-Mutations TRRAP, EPHB4 - - - T-Mutations - CHEK2 NF2, CDH1 BIRC3, HSP90AA1, KDM6A, RPTOR, EPHA5 S-Mutations BAP1 - - - Allele frequencies (median (min-max)) N-Mutations 0.042 (0.029-0.091) 0.067 (0.035-0.15) 0.070 (0.049-0.091) 0.059 (0.049-0.091) T-Mutations 0.23 (0.049-0.31) 0.13 (0.031-0.40) 0.17 (0.039-0.50) 0.16 (0.032-0.67) S-Mutations (Normal pool) 0.022 (0.0074-0.14) 0.033 (0.0064-0.063) 0.10 (0.014-0.13) 0.025 (0.0065-0.13) 0.20 (0.067-0.44) 0.077 (0.016-0.56) 0.13 (0.057-0.19) 0.15 (0.020-0.61) S-Mutations (Tumor pool) bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 443 Figure legends 444 Figure 1: Analysis of field cancerization in four patients. (a) Study design. Upper 445 part: analyses performed previously. WES was performed on bulk tumor samples. 446 Multiple tumor and normal biopsies were laser microdissected (LMD) and subjected 447 to deep targeted amplicon sequencing guided by the bulk tumor WES. Lower part: 448 present study (black box). Tumor and normal DNA samples were pooled and 449 subjected to deep targeted amplicon sequencing. Mutation calls were analyzed and 450 grouped into T-Mutations, N-Mutations, and S-Mutations. (b) Analysis of patients 1-4. 451 Field cancerization visualized using T-Mutations, N-Mutations, and S-Mutations. Gene 452 names and allele frequencies (AF) are displayed. AFs are illustrated as light grey bars 453 (AF measured in tumor) and dark grey (AF measured in normal). 455 Figure 2: Validation of mutations. (a) All mutations were evaluated in previously 456 generated WES data from tumors, recurrences, and metastases from the four patients 457 (patients 1 and 2 shown, patients 3 and 4 in Supplementary Fig. S3). Obtained AFs 458 are marked (yellow to red ranging from >0 to 0.6). For WES data, a minimum of five 459 reads at a given position were required for validation (indicated in grey). Dark blue 460 indicates no alternate alleles on the position. LN = lymph node. Targ. seq. = Targeted 461 sequencing.(b) AFs obtained by cancer panel sequencing of tumor compared to mean 462 AFs from WES on tumor samples from all four patients. Recurrences and metastases 463 were excluded from calculation of the mean as these samples were not included in the 464 tumor pools. Spearman correlation was calculated. (c) Validation of previously 465 identified tumor mutations by ddPCR on DNA from normal samples. Multiple assays 466 for specific mutations were included for the four patients and the fraction of mutated 467 sequences identified using ddPCR is shown (%). * indicates that the value is out of 468 scale (max value = 14.8%). (d) Validation of absence of N-Mutations in DNA from 469 tumor samples by ddPCR analysis. A positive control (synthesized oligo) for each 470 assay was included as well as negative controls (H2O and HT1197 bladder cancer cell 471 line). The purple line indicates cutoff set for positive droplets. Droplets positive for 472 mutation are marked in blue and negative droplets are indicated by grey. 474 Figure 3: Analysis of mutational context, impact and frequency. All analyses 475 were performed on the combined set of mutations from all patients. The total number bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. 476 of mutations in each category is indicated. (a) The six single-base substitutions 477 counted among N-, T-, and S-Mutations. (b) Predicted impact of mutations among N- 478 , T-, and S-Mutations grouped into high, moderate, low/modifier impact. (c) Predicted 479 impact of mutations in N-, T-, and S-Mutations grouped into synonymous and non- 480 synonymous (mutations predicted to have a high or moderate impact) mutations (d) 481 Number of C>G and C>T mutations among N-, T-, and S-Mutations in APOBEC 482 context. (e) Allele frequencies from N-, T-, and S-Mutations. For S-Mutations, allele 483 frequencies are measured both in the normal samples and in the tumor samples and 484 both are indicated. 0.8 0.6 0.4 0.2 0.4 0.3 0.15 0.2 0.1 0.1 0.05 0.5 0.4 0.3 0.2 0.1 0.09 0.3 0.06 0.15 0.03 0 0 0.1 0.05 Figure 1 Tumor Biopsies from Normal Biopsies Appearing Urothelium WES of Bulk Tumor Biopsies Tumor guided Deep Bulk Tumor Samples LMD Normal Samples Targeted Amplicon Sequencing NuGEN Deep Pool of all Tumor Pool of all Normal Targeted Samples Samples Sequencing Mutations unique for Tumor: T-Mutations Mutations unique for Normals: N-Mutations Mutations shared between Tumor and Normal: S-Mutations PolyPhen Probably damaging PolyPhen Possibly damaging PolyPhen Benign IntOGen Bladder Cancer Driver Stop Codon Patient 1 Patient 2 SMAD2 GATA6 CDKN1A* CDX2 KAT6A PAFAH1B2 CHEK2* FGFR3* IRS2* BAP1* APC* EP300* PPARA KMT2D PRKCA BAP1* KALRN MYH9 CBFB EPHA2 GATA2 FANCD2 MAGI2 MPL ERBB2 HDAC2 MCL1 Tumor TP53* FLNC ABL2 ADGRB3 JUN Tumor PIK3CA NBN bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint KALRN Field giving rise to tumor EP400 (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. TRRAP All rights reserved. No reuse allowed without permission. Field giving rise to tumor CARD11 KAT6A PLCG1 PRKDC CCND1 EPHB4 CCND2 Multiple other elds TCF12 MCL1 PML LZTR1 BCOR* Multiple other elds EPHB6 BCORL1 KDM5A BCR EXT2 CDC25A CD79B FANCD2 PLCG2 FES KMT2D BCORL1 EPHB2 SPTA1 TBX3* 16 (53%) Mutations only found in normals (Field Eect Specic Mutations) 9 (26%) Mutations only found in normals (Field Eect Specic Mutations) 6 (20%) Mutations shared between Normals and Tumor 11 (32%) Mutations shared between Normals and Tumor 8 (27%) Mutations only found in Tumor (Tumor Specic Mutations) 14 (41%) Mutations only found in Tumor (Tumor Specic Mutations) RBM10 ATRX IGF1R NF2 MED12 CDKN2B RAD51B BIRC3 FLT1 Patient 3 Patient 4 RBM10 PIK3CG NTRK3 BRCA1* FGFR4 GRIN2A FANCF HSP90AA1*COL1A1 NF2 KDM6A* SOX9 FLNC GRM3 TBX3* FAT1* TPO THBS1 TNK2 PIK3C2B MSH2 NOTCH1* IDH2 BRCA2 CDK8 NTRK3 MAP3K1* BMP4 RAD51B SMAD3 VEGFA JAK3 PTCH1 KDM5A SOX2 CDKN1A* POLE EP300* NTRK3 RAD51B NUP214 INPP4B EGFR IGF1R SUFU ZNF703 IGF2 FOS NF2 FLT1 RUNX1T1 FOXP1 Tumor EPHA3 SF3B1 GRIN2A CBFA2T3 VDR Tumor NUP214 GRIN2A DDR2 PPP2R1A GATA2 Field giving rise CDH1* SPEN ABL1 MST1R CDX2 to tumor APC2 PTPRF ROS1 IRF2 TBX3* Field giving rise to tumor DOT1L FOXO3 ASXL1 JUN TSPAN31 BCL3 CREBBP ADCY9 PLCG1 IGFBP3 FBXW7* PPP2R2B NF1 RET CBL NSD1 CUL3 CDK6 CBFA2T3 BCL2L2 RPTOR LAMA1 Other eld(s) PAK5 MAPK3 DGKB Other eld(s) ERCC3 KEL EPHA8 MYCL EGF EPHA5 ADGRB3 KAT6B EPHA8 MUC1 CDK2 LOC101928120 TCF3 NOS3 RPS6KB1 EGFR TET2 PRKDC ZNF331 NCOA2 NCOA2 WT1 NSD1 APC2 POLE RET FANCA FANCF GNAS* DNMT1 CDH1 PLCG2 KDM5C TBX3* AGAP2 2 (4%) Mutations only found in normals (Field Eect Specic Mutations) 5 (5%) Mutations only found in normals (Field Eect Specic Mutations) NRAS* DGKG 6 (12%) Mutations shared between Normals and Tumor 13 (14%) Mutations shared between Normals and Tumor 42 (84%) Mutations only found in Tumor (Tumor Specic Mutations) 75 (81%) Mutations only found in Tumor (Tumor Specic Mutations) 0.45 0.2 0.3 0.1 0.15 0.15 0.15 0 0 0.6 0.5 0.6 0.4 0.45 0.3 0.3 0.2 0.15 0.1 0.08 0.04 0 Figure 2 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. a b Patient 2 Patient 1 0.7 Gene Targ. seq. WES Targ. seq. Gene WES Rho = 0.77 MCL1 TRRAP -16 P-val < 2.2*10 LZTR1 KAT6A EPHB6 PRKDC KDM5A 0.6 EPHB4 EXT2 CCND2 CD79B TCF12 PLCG2 PML KMT2D BCOR 0.5 EPHB2 BCORL1 PIK3CA BCR MCL1 CDC25A IRS2 FANCD2 GATA2 FES 0.4 ABL2 BCORL1 GATA6 SPTA1 KAT6A TBX3 PPARA CDX2 KALRN FGFR3 0.3 MPL EP300 KALRN BAP1 CDKN1A EPHA2 CHEK2 SMAD2 APC 0.2 PAFAH1B2 PRKCA BAP1 CBFB KMT2D MAGI2 MYH9 HDAC2 FANCD2 FLNC 0.1 ERBB2 JUN TP53 NBN ADGRB3 EP400 CARD11 No a lterna te a llele 0.0 PLCG1 # Rea ds < 5 CCND1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 >0 0.6 AF targeted sequencing c d Patient 1 Patient 2 Patient 1 - DDB2 Patient 1 - EPHB4 Cell Cell Tumor samples Pos. H O Tumor samples Pos. H O 2 2 line line 5.00 5.00 FGFR3 PIK3CA 4.50 4.50 CDKN1A 8 8 STAG2 4.00 4.00 3.50 3.50 3.00 3.00 2.50 2.50 4 4 2.00 2.00 1.50 1.50 1.00 1.00 2 2 0.50 0.50 0.00 0.00 N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N1 N2 N3 N4 N5 N6 N7 N8 N9 Patient 3 Patient 4 Patient 2 - EPHB2 Cell * Tumor samples Pos. H O line 5.00 5.00 RBM10 PFKP C11orf70 CDH11 4.50 4.50 * Max =14.8 % 4.00 4.00 3.50 3.50 3.00 3.00 2.50 2.50 2.00 2.00 1.50 1.50 1.00 1.00 0.50 0.50 0.00 0.00 N1 N2 N3 N4 N5 N6 N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N11 Normal Pool Tumor Pool Tumor 1 Tumor 2 LN metastasis Normal Pool Tumor Pool Tumor 1 Tumor 2 Tumor 3 Tumor 4 Tumor 5 Tumor 6 Tumor 7 Recurrence AF (%) AF (%) AF (%) AF (%) Tumor Shared Normal Tumor Shared Normal AF WES Amplitude x 1,000 Amplitude x 1,000 bioRxiv preprint first posted online Jan. 31, 2019; doi: http://dx.doi.org/10.1101/536466. The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. Figure 3 a b c 100% 100% 100% 90% 90% 90% 80% 80% 14 80% 13 13 70% 70% 70% 60% 60% 60% 50% 50% 50% 40% 40% 40% 14 29 30% 30% 3 30% 20% 20% 20% 2 18 10% 10% 10% 3 11 1 4 0% 0% 0% T>A T>C T>G C>A C>G C>T High Moderate Modier/low Synonymous Non-Synonymous d e 100% -16 P-val < 2.2*10 0.8 90% 80% 0.6 70% 9 60% 50% 0.4 40% 30% 0.2 20% 10% 0% C>G/C>T non-APOBEC context N-Mutations T-Mutations C>G/C>T APOBEC context S-Mutations S-Mutations (Normal pool) (Tumor pool) AF

Journal

bioRxivbioRxiv

Published: Jan 31, 2019

There are no references for this article.