Edinburgh Research Explorer Assessment of the InSiGHT Interpretation Criteria for the Clinical Classification of 24 MLH1 and MSH2 Gene Variants Citation for published version: Tricarico, R, Kasela, M, Mareni, C, Thompson, BA, Drouet, A, Staderini, L, Gorelli, G, Crucianelli, F, Ingrosso, V, Kantelinen, J, Papi, L, Angioletti, MD, Berardi, M, Gaildrat, P, Soukarieh, O, Turchetti, D, Martins, A, Spurdle, AB, Nyström, M, Genuardi, M, InSiGHT Variant Intepretation Committee & Farrington, S 2017, 'Assessment of the InSiGHT Interpretation Criteria for the Clinical Classification of 24 MLH1 and MSH2 Gene Variants', Human Mutation, vol. 38, no. 1. https://doi.org/10.1002/humu.23117 Digital Object Identifier (DOI): 10.1002/humu.23117 Link: Link to publication record in Edinburgh Research Explorer Document Version: Peer reviewed version Published In: Human Mutation General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately and investigate your claim. Download date: 18. Jun. 2020 1 Assessment of the InSiGHT Interpretation Criteria for the Clinical Classification of 24 MLH1 2 and MSH2 Gene Variants 3 4 Rossella Tricarico1,2, Mariann Kasela3, Cristina Mareni4, Bryony A. Thompson5,6, Aurélie Drouet7, 5 Lucia Staderini1,8, Greta Gorelli1, Francesca Crucianelli1, Valentina Ingrosso1, Jukka Kantelinen3, 6 Laura Papi1, Maria De Angioletti9, Margherita Berardi9, Pascaline Gaildrat7, Omar Soukarieh7 Daniela 7 Turchetti10, Alexandra Martins7, Amanda B. Spurdle11, Minna Nyström3, Maurizio Genuardi1,12*, and 8 InSiGHT Variant Intepretation Committee13 9 10 1 11 of Florence, Florence, Italy 12 2 13 Philadelphia, United States 14 3 Department of Biosciences, Division of Genetics, University of Helsinki, Helsinki, Finland 15 4 Tuscan Tumor Institute, Florence, Italy 16 5 Department of Oncological Sciences, Huntsman Cancer Institute, University of Utah, Salt Lake City, 17 United States 18 6 19 University of Melbourne, Victoria, Australia Department of Biomedical, Experimental and Clinical Sciences, Medical Genetics Unit, University Current address: Cancer Epigenetics and Cancer Biology Programs, Fox Chase Cancer Center, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/humu.23117. This article is protected by copyright. All rights reserved. 20 7 21 Medicine, Rouen, France 22 8 23 Matteo, Pavia 24 9 25 Italy 26 10 27 Bologna, Italy 28 11 29 Brisbane, Australia 30 12 31 University of the Sacred Heart; and Medical Genetics Unit, Fondazione Policlinico Universitario “A. 32 Gemelli”, Rome, Italy 33 13 34 contributors see footnote after the Acknowledgments section Inserm-U1079-IRIB, University of Rouen, Normandy Centre for Genomic and Personalized Current address: Microbiology and Virology Department, Fondazione IRCCS Policlinico San Cancer Genetics and Gene Transfer - Core Research Laboratory, Istituto Toscano Tumori, Florence, Medical Genetics; Department of Medical and Surgical Sciences (DIMEC); University of Bologna, Genetics and Computational Biology Department, QIMR Berghofer Medical Research Institute, Current affiliations: Institute of Genomic Medicine, "A. Gemelli" School of Medicine, Catholic InSiGHT: International Society for Gastrointestinal Hereditary Tumors; for a full list of InSiGHT 35 36 37 *Corresponding author: 38 E-mail address: maurizio.genuardi@unicatt.it 39 40 Short title: Interpretation of MMR Gene Variants 2 This article is protected by copyright. All rights reserved. 41 Key words: Lynch Syndrome; Functional assays; Splicing; Variants of Uncertain Significance (VUS); 42 Multifactorial analysis; Microsatellite instability. 43 44 ABSTRACT 45 46 Pathogenicity assessment of DNA variants in disease genes to explain their clinical consequences is 47 an integral component of diagnostic molecular testing. The International Society for Gastrointestinal 48 Hereditary Tumors (InSiGHT) has developed specific criteria for the interpretation of mismatch repair 49 (MMR) gene variants. Here, we performed a systematic investigation of 24 MLH1 and MSH2 50 variants. The assessments were done by analyzing population frequency, segregation, tumor 51 molecular characteristics, RNA effects, protein expression levels and in vitro MMR activity. 52 Classifications were confirmed for 15 variants and changed for 3, and for the first time determined for 53 6 novel variants. Overall, based on our results we propose the introduction of some refinements to the 54 InSiGHT classification rules. The proposed changes have the advantage of homogenizing the 55 InSIGHT interpretation criteria with those set out by the Evidence-based Network for the 56 Interpretation of Germline Mutant Alleles (ENIGMA) consortium for the BRCA1/BRCA2 genes. We 57 also observed that the addition of only few clinical data was sufficient to obtain a more stable 58 classification for variants considered as “likely pathogenic” or “likely non pathogenic”. This shows 59 the importance of obtaining as many as possible points of evidence for variant interpretation, 60 especially from the clinical setting. 61 62 63 INTRODUCTION 3 This article is protected by copyright. All rights reserved. 64 65 Lynch syndrome (LS) (MIM# 120435) is the most common form of inherited colorectal and 66 endometrial cancer, predisposing also to other gastrointestinal (GI) (stomach, small bowel, biliary 67 tract, pancreas) and non-GI cancers (urinary tract, ovary and others). The syndrome is transmitted as 68 an autosomal dominant trait, and caused by constitutional defects in the mismatch repair (MMR) 69 genes MLH1 (MIM# 120436), MSH2 (MIM# 609309) MSH6 (MIM# 600678) and PMS2 (MIM# 70 600259) [Lucci-Cordisco et al., 2003; Lynch & De la Chapelle, 2003]. Detection of a constitutional 71 loss-of-function variant in an MMR gene provides diagnostic confirmation of LS and is essential for 72 the identification of at-risk members in LS families through predictive testing. This is especially 73 important for LS care, since carriers of MMR gene pathogenic variants can benefit from different risk 74 reducing 75 chemoprevention [Vasen et al., 2013]. options, including stringent surveillance protocols, prophylactic surgery, and 76 77 A major challenge in the diagnosis and management of LS is the frequent occurrence of variants of 78 uncertain significance (VUS) in the MMR genes. Depending on the gene, about 1/5th to 1/3rd of DNA 79 sequence variants identified during the course of LS clinical testing are of uncertain significance 80 [Sijmons et al., 2013], limiting risk reduction and management options in probands and preventing 81 their use in predictive gene testing in relatives. 82 83 The classification of DNA sequence variants identified in MMR and other cancer predisposition genes 84 is recommended to be based on data from multiple sources, including clinical observations, tumor 85 pathology studies and several RNA and protein based functional assays [Couch et al., 2008; Hofstra et 86 al., 2008; Spurdle et al., 2008; Tavtigian et al., 2008a; Richards et al., 2015]. A number of in silico 4 This article is protected by copyright. All rights reserved. 87 programs have also been devised to assist with the prediction of functional consequences of inherited 88 MMR gene alterations [Tavtigian et al., 2008b; Thompson et al., 2013b; Niroula & Vihinen, 2015]. 89 90 Recently, the International Society for Gastrointestinal Hereditary Tumors (InSiGHT) has developed 91 criteria for the interpretation of MMR gene variants, with the aim to improve the clinical utility of 92 genetic testing for LS. A systematic clinical classification of all variants contained in InSiGHT locus 93 specific databases (http://insight-group.org/variants/database/) 94 multifactorial bayesian quantitative approach and/or on stringent combinations of qualitative clinical 95 and functional data [Thompson et al., 2014]. Variants were classified using a 5-tier system devised for 96 cancer predisposing genes [Plon et al., 2008]. was performed, based on a 97 98 In the present study, we assessed the pathogenicity of 24 MMR gene variants using an extensive 99 combination of RNA and protein-based functional assays, segregation studies, and tumor analyses. 100 We were able to classify 6 novel variants as well as to confirm or refine the classification of 18 101 variants previously assessed by InSiGHT. Overall, we show the necessity of using different analyses 102 in VUS classification and discuss their specific value and status in the interpretation process. 103 104 105 MATERIALS AND METHODS 106 107 Genetic variants, patients and samples 108 The variants assessed in this study were detected in a single laboratory in families fulfilling the 5 This article is protected by copyright. All rights reserved. 109 Bethesda criteria [Vasen, 2005] ascertained through cancer family clinics from 2002 to 2011. Overall, 110 57 MMR gene variants (25 MLH1 and 32 MSH2), excluding well established polymorphisms, were 111 detected in a total of 56 families. Variants that were clearly disease causing (ie truncating, splicing 112 alterations, or large rearrangements), as well as established polymorphisms, were not considered. 113 Exceptions were some alleles that, despite being currently considered polymorphic, were not clearly 114 classified at the time of their detection; these included MLH1 c.1558+14G>A, and MSH2 c.380A>G 115 and c.1511-9A>T, whose assessment was useful for the validation of the variables investigated for 116 classification. Overall, we evaluated 24 MMR gene variants, 13 MLH1 and 11 MSH2, identified in 37 117 unrelated families. All variants were single nucleotide substitutions at the genomic level. Based on 118 their positions in the DNA sequence and predicted effects, they could be divided into the following 119 groups: 12 aminoacid substitutions (8 in MLH1 and 4 in MSH2), 1 MLH1 potential splice site change 120 affecting the first exonic position and also potentially causing an aminoacid substitution, 4 121 synonymous exonic nucleotide substitutions (1 MLH1, 3 MSH2), and 7 intronic changes outside the 122 most conserved positions of the splice site consensus sequences (4 MLH1 and 3 MSH2). Six variants 123 (4 MLH1 and 2 MSH2) were previously unreported and therefore not assessed by InSiGHT. 124 125 Peripheral blood leukocyte (PBL) samples were collected from 76 and 16 subjects for DNA and RNA 126 extractions, respectively. Furthermore, lymphoblastoid cell lines (LCLs) were established from 7 127 variant carriers. Forty-nine paraffin-embedded tumor specimens were obtained from 42 carriers of 23 128 different variants. 129 130 The study was approved by the Institutional Ethical Board of the Careggi University Hospital, 131 Florence. Informed consent was obtained from all patients for the use of specimens and 132 clinical/pathological data for research purposes. 6 This article is protected by copyright. All rights reserved. 133 134 Molecular analyses 135 The complete coding sequence and flanking exon–intron borders of the MLH1 and MSH2 genes were 136 investigated by direct sequencing on genomic DNA. The presence of MLH1 and MSH2, as well as of 137 EPCAM (MIM# 185535), genomic rearrangements and the methylation status of the MLH1 promoter 138 were investigated by Multiplex Ligation-dependent Probe Amplification (MLPA), as previously 139 described [Crucianelli et al., 2014]. Values lower than 0.15 were assumed as a cut-off for normal 140 methylation levels according to previous studies [Gylling A et al., 2007; Crucianelli et al., 2014]. 141 142 The identified MLH1 and MSH2 variants have been submitted to the InSiGHT MMR gene variant 143 database (http://www.insight-group.org/variants/database/). Variants were defined according to the 144 Human Genome Variation Society (HGVS) recommendations [den Dunnen et al., 2016]. DNA variant 145 numbering is based on the MLH1 and MSH2 cDNA sequences (GenBank accession numbers 146 NM_000249.2 and NM_000251.1, respectively) with the A of the ATG translation–initiation codon 147 numbered as +1. Aminoacid numbering starts with the translation initiator methionine as +1. 148 149 To investigate the presence of the p.Val600Glu (V600E) hotspot mutation, BRAF (MIM# 164757) 150 exon 15 was directly sequenced in 7 tumor samples of MLH1 variant carriers, using previously 151 reported primers and conditions [Mancini et al., 2010]. 152 153 Loss of heterozygosity (LOH) analysis of the regions containing the identified variants was performed 154 on matched leukocyte and tumor tissues from 19 probands by analysis of direct sequencing 155 electropherograms [Janssen et al., 2011; Janssen et al., 2012]. 7 This article is protected by copyright. All rights reserved. 156 157 Microsatellite instability and immunohistochemical analyses 158 A total of 47 tumor samples and matched normal mucosa or PBLs from 40 patients were evaluated for 159 microsatellite instability (MSI) using a 5-mononucleotide marker panel [Suraweera et al., 2002; 160 Buhard et al., 2006; Giunti et al., 2009]. Tumors were classified into three categories according to the 161 proportions of markers showing instability: MSI-H (high-level MSI), MSI-L (low-level MSI) and 162 MSS (microsatellite stable), which have ≥ 30 – 40 %, > 0 % – < 30 – 40 %, and 0 % unstable markers, 163 respectively [Boland et al., 1998]. Immunohistochemical (IHC) analysis of MMR protein expression 164 was performed on paraffin-embedded tumor tissue sections from 42 samples, as previously described 165 [Roncari et al., 2007]. 166 167 Allelic frequencies in control chromosomes 168 To assess frequencies of the 24 MLH1 and MSH2 variants in control chromosomes, one hundred and 169 sixty DNA samples from anonymized healthy Italian blood donors with no history of colorectal 170 cancer among 1st degree relatives and from the same region of origin (Tuscany) of most of the patients 171 were analyzed by direct sequencing. In addition, the Exome Aggregation Consortium database (ExAC 172 Browser (Beta), http://exac.broadinstitute.org/, 04/2016 accessed) was interrogated, excluding the 173 Cancer Genome Atlas (TGCA) data. 174 175 Co-segregation with phenotype and multifactorial likelihood analysis 176 Co-segregation analysis was performed for 11 variants in 16 families; in these, 24 affected carrier 177 relatives, in addition to probands, were identified. The variants were detected by direct sequencing. 178 Multifactorial likelihood analysis was performed for 14 variants for which sufficient data were 8 This article is protected by copyright. All rights reserved. 179 available, as described previously [Thompson et al. 2013a; Thompson et al. 2013b]. Briefly, a 180 probability of pathogenicity based on variant location or in silico scoring of missense substitutions is 181 combined with likelihood ratios (LR) for segregation and tumour characteristics (MSI/BRAF status) 182 to derive a posterior probability of pathogenicity. 183 184 mRNA splicing analysis 185 Total RNA was extracted from the 7 cycloheximide-treated and untreated LCLs established from 186 MLH1 or MSH2 variant carriers, using RNeasy® Plus Mini Kit (Qiagen, Hilden, D). Cycloheximide 187 (Sigma-Aldrich, Saint Louis, MO, USA) was added at 10µg/ml to the medium 4 hr before harvesting 188 the cells to prevent degradation of unstable transcripts by nonsense-mediated decay (NMD). cDNA 189 was synthesized using TaqMan Reverse Trascription Kit (Applied Biosystems, Foster City, CA, 190 USA). Primers and conditions used for cDNA amplification are available upon request. PCR products 191 were analyzed on agarose gels, and individual bands, corresponding either to the full length or to the 192 aberrantly spliced transcripts were excised and eluted using the QIAquick Gel Extraction Kit (Qiagen, 193 Hilden, Germany) before amplification and direct sequencing. All RT-PCR experiments were 194 performed in duplicate. Since alternative splicing is commonly observed for MLH1 and MSH2 195 [Genuardi et al. 1998], to improve the detection and interpretation of splicing aberrations, eight 196 control samples were also analyzed [Thompson et al., 2015]. 197 198 Allele-specific expression (ASE) analysis 199 Allele-specific expression (ASE) was investigated by Single Nucleotide Primer Extension (SNuPE) in 200 10 patients heterozygotes for the coding SNPs rs1799977 (MLH1 c.655G>A) or rs4987188 (MSH2 201 c.965G>A, as previously described [Crucianelli et al., 2014]. Total RNA was extracted from blood 202 samples collected into PAXgene Blood RNA tubes (PreAnalytiX, Qiagen, Hilden, Germany), using 9 This article is protected by copyright. All rights reserved. 203 the PAXgene Blood RNA Kit (Qiagen, Hilden, Germany) according to the manufacturer’s 204 instructions. Samples from heterozygotes for the same SNPs who had no additional MLH1 and MSH2 205 sequence change were used as controls [Perera et al., 2010; Pastrello et al., 2011]. ASE was calculated 206 after measuring peak heights in heterozygous samples [Castellsagué et al., 2010]. Values included in 207 the 0.8-1.2 range were assumed as a cut-off for normal ASE according to previous studies [Renkonen 208 et al., 2003; Castellsagué et al., 2010; Perera et al., 2010]. All experiments were carried out in 209 triplicate and two independent replicates of all experiments were performed. Control heterozygotes 210 for the MLH1 or MSH2 exonic polymorphisms (rs1799977 and rs4987188, respectively) were 211 included in each experiment. 212 213 Minigene splicing assay 214 Splicing assays were performed by comparing the splicing pattern of WT and mutant minigenes 215 transiently expressed in HeLa cells [Soukarieh et al., 2016]. Two different vectors were used in the 216 minigene splicing assay: pCAS2 or pSPL3m [Soukarieh et al., 2016], as specified. Except for 2 217 contructs (MSH2 c.2006-6T>C and c.2081T>C), minigenes were prepared by first PCR-amplifying 218 wild-type (WT) and mutant genomic segments from patients’ DNA using forward and reverse primers 219 mapping approximately at 150 nucleotides upstream and downstream the exon of interest, 220 respectively. Primer sequences are available upon request. The PCR products were then inserted into 221 the intron of pCAS2 to generate splicing reporter minigenes as previously described [Tournier et al., 222 2008]. Minigenes carrying the single variants MSH2 c.2006-6T>C and MSH2 c.2081T>C (present in 223 cis in patient genomic DNA) were prepared by site-directed mutagenesis by using a two-stage overlap 224 extension PCR method [Ho et al., 1989]. The psPL3m construct carrying MLH1 c.301G>A was 225 generated 226 pCAS1.MLH1.exon3.c.301G>A [Tournier et al., 2008]. by transferring the insert from the 10 This article is protected by copyright. All rights reserved. previously described minigene 227 228 Protein stability and vitro MMR activity analyses 229 Altogether 9 MLH1 and MSH2 missense variations were introduced into the MLH1 and MSH2 230 cDNAs cloned into a pFastBac1 vector (Invitrogen, Carlsbad, CA, USA), using a PCR-based site- 231 directed mutagenesis kit according to manufacturer´s instructions (QuikChange Lightning®Site- 232 directed mutagenesis Kit, Stratagene, La Jolla, CA, USA). The mutated constructs were sequenced 233 (ABIPrism 3100 Genetic Analyzer; Applied Biosystems, Foster City, CA, USA) prior to protein 234 production. Primer sequences and PCR parameters are available upon request. 235 236 Recombinant baculoviruses were generated by Bac-to-Bac system (Invitrogen, Carlsbad, CA, USA) 237 and proteins were produced in Spodoptera frugiperda (Sf9) insect cells as described earlier [Nyström- 238 Lahti et al., 2002; Kariola et al, 2002; Ollila et al., 2006]. For protein production Sf9 cells were co- 239 infected with MLH1 and PMS2 viruses to yield MutLα heterodimers, or MSH2 and MSH6 viruses to 240 yield MutSα heterodimers. The total protein extracts (TE) were prepared as previously described 241 [Nyström-Lahti et al., 2002; Kariola et al., 2002]. 242 243 The expression levels of produced protein variants were studied by Western blot analysis. The 244 proteins were blotted onto nitrocellulose membranes (Hybond, PVDF, Amersham Pharmacia biotech, 245 Uppsala, Sweden) and visualized with anti-MLH1 (BD Biosciences/Pharmingen, San Diego, CA, 246 USA, clone 168-15) (0.5 µg/ml), anti-PMS2 (Calbiochem/Oncogene Research, San Diego, CA, USA, 247 Ab-1) (0.2 µg/ml), anti-MSH2 (Calbiochem, San Diego, CA, USA, MSH2- Ab1, NA-26) (0.2 μg/ml) 248 and anti-MSH6 (BD Transduction Laboratories, Lexington, KY, USA, clone 44) (0.02 μg/ml) 249 monoclonal antibodies. To estimate the MMR protein level in the extracts, α-tubulin was used as a 250 loading control (anti-α-tubulin; Sigma, Louis, MO, USA, DM1A) (0.2 µg/ml). 11 This article is protected by copyright. All rights reserved. 251 252 The repair efficiencies (R%) of the recombinant protein variants were analyzed by complementing the 253 MMR-deficient nuclear extracts (NE) of HCT116 or LoVo cells (American Type Culture Collection, 254 Manassas, VA, USA) with the Sf9 TEs containing overexpressed MLH1 or MSH2 proteins [Nyström- 255 Lahti et al., 2002; Kantelinen et al., 2012] according to the protocol previously described [Kantelinen 256 et al., 2012]. Seventy-five µg of NE was incubated with TE including in each sample comparable 257 amounts of MutLα or MutSα, respectively, and with an excess amount (100 ng) of the heteroduplex 258 DNA substrate. R% was calculated as an average of three independent experiments using GeneTools 259 3.08 (SynGene, Cambridge, England). The relative R% was calculated in respect to the WT control 260 [Drost et al., 2010; Kantelinen et al., 2012]. 261 262 263 264 Clinical classification of variants 265 The 5 class system for clinical classification recommended by the International Agency for Research 266 on Cancer (IARC) working group on the interpretation of DNA sequence variants in cancer 267 predisposition genes was used [Plon et al., 2008]. Class 5 and 4 include definitely pathogenic and 268 likely pathogenic variants, respectively; when observed in a proband, they provide confirmation of LS 269 diagnosis, so that relatives can be consequently offered predictive testing. Variants assigned to class 1 270 and 2 correspond to definitely and likely neutral/not pathogenic (or of low clinical impact) sequence 271 changes, respectively; their detection is not followed by further clinical testing in the family. Finally, 272 class 3 includes all those variants whose clinical and/or functional effects cannot be determined, due 273 to either insufficient (e.g. detection in a single family) or contradictory evidence (ie discordant results 12 This article is protected by copyright. All rights reserved. 274 from similar in vitro assays performed in different laboratories); these are also called variants of 275 uncertain (or unknown) significance (VUS). 276 277 278 RESULTS 279 280 Population frequency 281 We first verified variant allele frequencies in control populations (Table 1). Previous classifications 282 performed by InSiGHT relied on frequencies reported in the 1000 Genomes Project database 283 (http://www.1000genomes.org/), in addition to data published or reported by single centers on local 284 populations. In this study we used values from the Exome Aggregation Consortium (ExAC), which is 285 a more comprehensive dataset and which also incorporates the 1000 Genomes data. We also tested 286 160 Italian control subjects for 20 variants. 287 288 Four variants (MLH1 c.1558+14G>A; MSH2 c.380A>G, c.1511-9A>T and c.2006-6T>C) that 289 reached minor allele frequencies > 0.01 had previously been classified as Class 1-not pathogenic 290 based on population frequency data alone [Thompson et al., 2014], and were subjected to further 291 analyses to verify consistency across different points of evidence. The frequency of MSH2 c.380A>G 292 in phase 1 of the 1000 Genomes Project, which had been originally used for classification, was 0.02, 293 while the currently reported frequency in ExAC is slightly lower, 0.00692. Of note, the frequencies of 294 MLH1 c.1039-8T>A were 0.00155 in ExAC and 0.02187 in 160 Italian controls, respectively. This 295 difference is likely accounted for by a low quality of calling in the ExAC population, as suggested by 296 the following observations: (i) it is called in less than 80% of individuals in ExAC; (ii) the variant was 13 This article is protected by copyright. All rights reserved. 297 found to be in linkage disequilibrium with MLH1 c.1558+14G>A in the Italian population; and (iii) 298 MLH1 c.1558+14G>A had similar polymorphic frequencies in ExAC and in Italian controls (0.03948 299 versus 0.02187). 300 301 Tumor pathology data 302 Analyses 303 immunohistochemistry (IHC), BRAF p.Val600Glu somatic mutation, MLH1 promoter methylation, 304 and loss of heterozygosity (LOH) (Table 1). performed on tumors included microsatellite instability (MSI), MMR 305 306 MSI and/or IHC data were available for 23/24 and 22/24 variants, respectively (Table 1). These 307 included 12/13 predicted missense changes, for 6 of which > 2 tumors were investigated. Results 308 indicative of in vivo MMR inactivation (MSI-H status and/or lack of expression of the protein 309 encoded by the gene carrying the constitutional variant) were observed for 7 predicted missense 310 substitution variants: MLH1 c.301G>A p.(Gly101Ser), c.779T>G p.(Leu260Arg), c.1421G>C 311 p.(Arg474Pro), c.1814A>C p.(Glu605Ala), and MSH2 c.2081T>C p.(Phe694Ser) and c.2087C>T 312 p.(Pro696Leu), as well as for the potential splice variant MSH2 c.2006G>T. Results indicative of in 313 vivo and in vitro MMR proficiency (MSS status, normal MMR protein expression and proficient 314 functional assays) were obtained on 2 missense variants, MLH1 c.1043T>C p.(Leu348Ser) and MSH2 315 c.244A>G p.(Lys82Glu). Discordant MSI and IHC results were observed for the missense variants 316 MLH1 c.2041G>A p.(Ala681Thr) and c.2059C>T p.(Arg687Trp) in tumors from different 317 individuals; for both variants 1 MSI-H tumor showed normal IHC staining, while 1 MSS sample was 318 associated with lack of MLH1 expression. 319 14 This article is protected by copyright. All rights reserved. 320 BRAF and/or MLH1 promoter analyses were performed on tumor samples for 9 variants (Table 1). 321 The BRAF p.Val600Glu mutation was detected in two MLH1-negative tumors from carriers of the 322 MLH1 variants c.1421G>C p.(Arg474Pro) and c.1743G>A p.(Pro581=), respectively. MLH1 323 epigenetic defects could be tested only for the tumor from the carrier of the c.1743G>A variant, which 324 however did not show MLH1 promoter hypermethylation. Four additional samples had both BRAF 325 p.Val600Glu and MLH1 promoter methylation tested: both analyses were negative in 3 MSI-H 326 samples from carriers of MLH1 c.779T>G p.(Leu260Arg) carriers and in 1 MSI-H sample from a 327 c.2041G>A p.(Ala681Thr) carrier. 328 329 LOH analysis was performed for 14 variants, 8 MLH1 and 6 MSH2 (Table 1). Loss of the variant 330 allele was detected in tumors from carriers of 3 different MLH1 variants: c.1217G>A p.(Ser406Asn), 331 c.1421G>C p.(Arg474Pro) and c.1732-19T>C. 332 333 RNA analyses 334 The MSH2 variant c.2006G>T, which is currently assigned to Class 5 based on evidence of a major 335 splicing defect, was associated with complete exclusion of exon 13 in the minigene assay (Fig. 1), but 336 with only partial skipping in the LCL from a carrier (Fig. 2A and 2B). Both alleles at position 2006 337 were detected in the full-length cDNA product from the LCL (Fig. 2C and 2D). Partial exon 13 loss 338 was also detected in blood samples drawn in PAXgene tubes from the patient above and from 2 339 additional individuals carrying MSH2 c.2006G>T investigated in another laboratory in France, as well 340 as in a LCL established from one of these French carriers (data not shown). 341 15 This article is protected by copyright. All rights reserved. 342 Three other variants (MLH1 c.301G>A and c.1039-8T>A; MSH2 c.2006-6T>C) were associated with 343 partial exon skipping (Table 2; Fig. 1; Supp. Fig. S1-S2) in patient samples, in the ex vivo minigene 344 assay, or in both. These involved in all cases known alternatively spliced transcripts [Genuardi et al., 345 1998; Thompson et al., 2015]. Interestingly, for MSH2 c.2006-6T>C partial exon 13 exclusion was 346 only detected by the minigene assay (Fig. 1) but not in the patient sample (Tournier et al., 2008), 347 further suggesting that the minigene assay may overestimate the splicing defect for this exon. On the 348 other hand, in this study partial skipping of exon 12 was observed in lymphoblastoid cells (data not 349 shown) but not with the minigene assay for MLH1 c.1039-8T>A (Supp Fig. S2). The latter result is in 350 accordance with previous findings obtained for this variant by minigene assay [Petersen et al., 2013]. 351 352 In addition, the minigene assay showed that Class 5-pathogenic MLH1 c.301G>A p.(Gly101Ser) is 353 associated with loss of the use of an alternative splice site (Supp Fig. S3); while the clinical 354 significance of this finding cannot be established based on the minigene result only, the variant allele 355 should produce only the canonical transcript. 356 357 None of the 10 variants tested by the SNuPE assay showed evidence of allelic expression imbalance, 358 consistent with the splicing assay results (Table 2). 359 360 Mismatch repair activity and protein expression analyses 361 An in vitro MMR complementation assay based on the synthesis of MMR protein in Spodoptera 362 frugiperda (Sf9) insect cells was performed in MMR-defective human cell lines for 9 coding variants 363 (Table 2). Three of the five MMR defective protein variants (MLH1 p.Leu260Arg; MSH2 364 p.Phe694Ser and p.Pro696Leu were found to be unstable in vitro (Fig. 3). Loss of MMR activity 16 This article is protected by copyright. All rights reserved. 365 (relative activity < 1%) was demonstrated for 5 variants: MLH1 p.Gly101Ser and p.Leu260Arg, and 366 MSH2 p.Gly669Val, p.Phe694Ser and p.Pro696Leu. The remaining four variants, MLH1 367 p.Leu348Ser, p.Arg474Pro, p.Glu605Ala, and MSH2 p.Lys82Glu, were all stable in the transient 368 expression assay and MMR proficient. 369 370 Four other missense variants, MLH1 p.Ser406Asn, p.Ala681Thr, p.Arg687Trp, and MSH2 371 p.Asn127Ser, had been previously shown to be proficient in the MMR activity assay, although two of 372 them, MLH1 p.Ala681Thr and p.Arg687Trp, showed discordant protein instability results across 373 different studies (Table 2). 374 375 Clinical data, multifactorial analysis and variant classification 376 Family history types, co-occurrence of other MMR gene variants, and the components and results 377 multifactorial analysis are shown in Supp. Tables S1-S2. 378 379 Multifactorial analysis was performed for 14 variants. Using quantitative analysis (based on 380 multifactorial posterior probability) and/or assessment of qualitative criteria, variants were classified 381 according to the 5-tier system proposed by InSiGHT (Table 3) [Thompson et al., 2014]. Four of the 382 six novel variants (MLH1, c.1732-19T>C and c.1743G>A; MSH2 c. 244A>G and c.2442T>G) were 383 assigned to Class 2-likely not pathogenic. Of note, one novel variant, MLH1 c.1814A>C 384 p.(Glu605Ala), is in class 4-likely pathogenic according to multifactorial analysis based only on 2 385 available values, a 0.7 prior probability calculated in silico, and a single MSI-H CRC not expressing 386 the MLH1 protein. The remaining novel variant, MLH1 c.1043T>C p.(Leu348Ser) is in Class 3-VUS 387 due to insufficient evidence. 17 This article is protected by copyright. All rights reserved. 388 389 The previous InSiGHT assignment of MSH2 c.2006G>T to Class 5-pathogenic based on RNA 390 splicing data was confirmed by the results of multifactorial analysis in this study (Table 3). 391 392 The classification made by InSiGHT was changed for 3 variants after the addition of novel 393 segregation and molecular tumor data. MLH1 c.301G>A p.(Gly101Ser), originally in Class 4-likely 394 pathogenic, was upgraded to Class 5-pathogenic, while variants MSH2 c.1387-8G>T and c.1737A>G 395 p.(Lys579=) were moved from Class 2-likely not pathogenic to Class 1-not pathogenic. 396 397 Previous classifications of the other 15 variants were supported by the data obtained. Results of novel 398 RNA analyses performed in this study were in agreement with the assignment to Class 1-not 399 pathogenic of intronic MLH1 variants c.1039-8T>A and c.1558+14G>A. Insufficient evidence to 400 attain a clinically actionable category was available for MLH1 c.1421G>C p.(Arg474Pro), even 401 though novel data from tumor studies brought down the posterior probability of pathogenicity from 402 0.51 to 0.094. 403 404 405 DISCUSSION 406 407 The ultimate purpose of genetic testing for LS and other cancer predisposition syndromes is to reduce 408 cancer morbidity and mortality through the implementation of specific preventive options for carriers 409 of disease causing variants. Interpreting the significance of DNA variants identified in the diagnostic 18 This article is protected by copyright. All rights reserved. 410 laboratory is an integral component of clinical DNA testing. The interpretation process is complex, as 411 several independent datasets must be taken into account. Recently, recommendations for clinical 412 classification of MMR gene variants have been formulated [Thompson et al., 2014]. We have 413 performed a thorough investigation of 24 MMR gene sequence variants identified in a single center in 414 order to assess their clinical relevance, using points of evidence that are included in the InSiGHT 415 recommendations, as well as additional potential classification components. Our findings confirm the 416 overall validity of the InSiGHT criteria and suggest that the interpretation process could be improved 417 by introducing some adjustments. 418 419 Overall, our results provide support to or improve previous classifications for the 18 variants that had 420 already been assessed by InSiGHT (http://insight-group.org/variants/classifications). For 3 of these 421 variants (MLH1 c.301G>A; MSH2 c.1387-8G>T and c.1737A>G), a more stable classification, either 422 from Class 4-likely pathogenic to Class 5-pathogenic or from Class 2-likely not pathogenic to Class 1- 423 not pathogenic, was achieved using novel clinical and molecular data. These changes do not affect 424 cancer prevention strategies, since the same clinical recommendations apply to Class 5 and 4 and 425 Class 2 and 1, respectively [Plon et al., 2008]. However, assignments to Class 5 and 1 can be 426 considered definitive, since the likelihood that a variant in either of these categories will be moved to 427 a class associated with different clinical advice is very low [Plon et al., 2008]. The IARC/InSiGHT 428 interpretation criteria advise to consider research testing of further samples/relatives to try and obtain 429 definitive classifications for Class 4-likely pathogenic and Class 2-likely not pathogenic variants 430 [Plon et al. 2008; Thompson et al. 2014], and our results demonstrate the practical importance of this 431 recommendation. For all 3 variants the classification was based on multifactorial analysis, and in all 432 cases the class switch was made possible by the incorporation of few novel data on tumor 433 characteristics and/or segregation, highlighting the relevance of collecting these types of information. 434 19 This article is protected by copyright. All rights reserved. 435 The novel variants MLH1 c.1814A>C p.(Glu605Ala) and MSH2 c.244A>G p.(Lys82Glu) were in 436 Class 4-likely pathogenic and Class 2-likely not pathogenic, respectively, following multifactorial 437 analysis. For both variants only one clinical observation, that is, molecular information obtained on a 438 single tumor sample (Table 1 and Supp. Table S2), is available. The Evidence-based Network for the 439 Interpretation of Germline Mutant Alleles (ENIGMA) in the BRCA1/BRCA2 (MIM# 113705 and 440 MIM# 600185) genes recommends that variants attaining thresholds for assignment to clinically 441 actionable classes by multifactorial analysis with limited contribution from clinical or laboratory 442 evidence 443 (http://enigma.consortium.org/documents/ENIGMA_Rules_2015_03_26.pdf). We propose to adopt 444 this recommendation also for the MMR genes, especially when there is apparent discordance between 445 functional and clinical evidence, such as for MLH1 p.(Glu605Ala). In particular, for MLH1 variants 446 additional evidence from BRAF and/or promoter methylation tumor testing could be used to reinforce 447 the evidence in favor of pathogenicity. be considered of uncertain significance until further evidence is accrued 448 449 The partially discordant RNA splicing results between the minigene assay and analyses of patient 450 derived samples obtained in this study for MSH2 c.2006G>T suggest that the splicing alteration may 451 not be the only or the major inactivation mechanism for MSH2 c.2006T. Indeed, the functional in 452 vitro assay showed reduced repair activity of the protein encoded by the variant allele, p.669Val and 453 the variant could be assigned to Class 5 also based on multifactorial analysis. However, complete 454 absence of c.2006T allele in full-length transcript.in patient RNA, together with total exon 13 455 exclusion in a minigene assay, was observed in another study [van der Klift et al., 2015]. Therefore, 456 further studies will be needed to clarify the mechanisms underlying pathogenicity of MSH2 457 c.2006G>T. At the same time, the interpretation criteria for RNA analyses should be reconsidered 458 based on these apparently inconsistent results. 459 20 This article is protected by copyright. All rights reserved. 460 Minor effects on splicing were observed either on patient RNA or by the minigene assay in this study 461 for the Class 1-not pathogenic variants MLH1 c.1039-8T>A and MSH2 c.380A>G and c.2006-6T>C. 462 All are in Class 1 based on population frequency only, confirming that they have no major clinical 463 effects [Genuardi et al., 1998; Thompson et al., 2015]. 464 465 None of the other variants were found to be associated with significant splicing anomalies. Lack of 466 abnormal splicing products was important to assign MLH1 c.307-19A>G, c.1732-19T>C and 467 c.1743G>A, and MSH2 c. 2442T>G p.(Leu814=) to Class 2-likely not pathogenic. Two of them, 468 MLH1 c.307-19A>G and c.1743G>A also had population frequencies ~1/5,000 and ~1/10,000, 469 respectively. According to the InSiGHT criteria, synonymous or deep intronic variants for which 470 splicing assays do not show alterations should be considered as Class 2-likely not pathogenic. One of 471 the combinations required for assignment of a variant to Class 1-not pathogenic includes all of the 472 following points of evidence: allelic frequency 0.01%-1%, lack of co-segregation with disease, 473 estimated risk <1.5 determined by case-control studies, and presence of molecular features not 474 compatible with involvement of the gene carrying the variant in ≥ 3 tumors; this criterion applies to 475 all types of variants, regardless of their nature and prior probability of altering gene function. Since 476 intronic and synonymous variants have a low a priori likelihood of affecting gene processing, 477 combinations of any of the above evidences (ie, lack of segregation, population frequency, low 478 estimated risk, and molecular characteristics) and normal splicing patterns could reasonably be 479 considered sufficient for assignment to Class 1. Interestingly, the association of intronic location or 480 synonymous coding nucleotide substitution and absence of mRNA aberrations demonstrated by in 481 vitro assays has been proposed by the ENIGMA consortium as a criterion for assignment of variants 482 in 483 (http://enigmaconsortium.org/documents/ENIGMA_Rules_2015-03-26.pdf). Data from our study the BRCA1/BRCA2 genes to 21 This article is protected by copyright. All rights reserved. Class 1-not pathogenic 484 indicate that it would be justified to consider homogenization of the Class 1 criteria for 485 intronic/synonymous substitutions between the ENIGMA and InSiGHT consortia. 486 487 Nine missense variants were investigated by in vitro MMR assay based on the production of MSH2 or 488 MLH1 proteins in Sf9 insect cells and subsequent complementation of human MMR deficient cell 489 lines. The same assay had been previously used for three other missense substitutions found in our 490 series [Raevaara et al., 2005; Ollila et al., 2008; Christensen et al., 2009; Kansikas et al., 2011], while 491 one variant - MLH1 c.1217G>A p.(Ser406Asn) - had been tested with two different mammalian repair 492 assays. All five MMR deficient variants (MLH1 p.(Gly101Ser) and p.(Leu260Arg); MSH2 493 p.(Gly669Val), p.(Phe694Ser) and p.(Pro696Leu)) are in Class 5-pathogenic, supporting the 494 classifications based on multifactorial analysis. Of note, the aminoacids replaced in MSH2 495 p.(Phe694Ser) and p.(Pro696Leu) are located nearby in the ATPase domain, indicating that this 496 region is particularly sensitive to structural changes; this suggestion is reinforced by the observation 497 that none of the 29 reported MSH2 exon 13 missense changes have been so far assigned to Class 1-not 498 pathogenic 499 (http://chromium.lovd.nl/LOVD2/colon_cancer/variants.php?select_db=MSH2&action=search_all&s 500 earch_Variant%2FExon=13&search_MutCol=%3E&search_Variant%2FDNA=&search_Variant%2F 501 RNA=&search_Variant%2FProtein=&search_Patient%2FPhenotype%2FDisease=&search_Patient% 502 2FReference=). or Class 2-likely not pathogenic by InSiGHT 503 504 MLH1 c.2041G>A p.(Ala681Thr) and c.2059C>T p.(Arg687Trp) are assigned to Class 5-pathogenic 505 despite the results of the functional assays, which show inconclusive data on protein expression and 506 normal MMR activity, with discordant observations across different studies. However, both are 507 associated with an abundance of clinical data allowing them to overcome the Class 5-pathogenic 22 This article is protected by copyright. All rights reserved. 508 posterior probability threshold using multifactorial analysis. At the same time discordant tumor 509 pathology findings, including samples that were MSS and/or expressed MLH1, have also been 510 reported for both variants. It will be interesting to verify the degree of phenotypic expression 511 associated with these two variants. By analogy with equivocal functional results obtained on the 512 BRCA1 variant p.Arg1699Gln [Spurdle et al., 2012] they might be considered as candidate 513 intermediate risk variants. Notably, other MLH1 missense substitutions located in proximity of these 514 variants are associated with proficient repair but reduced or inconclusive protein expression data; 515 these include for instance the Class 5-pathogenic c.1942C>T p.(Pro648Ser) and c.1943C>T 516 p.(Pro648Leu), and the Class 3-VUS c.1918C>T p.(Pro640Ser), c.1919C>T p.(Pro640Leu), 517 c.1976G>A p.(Arg659Gln), c.2027T>G p.(Leu676Arg), and c.2027T>C p.(Leu676Pro). Therefore 518 variants located in this region of the MLH1 protein may cause functional impairment through reduced 519 expression/stability and/or other as yet to be determined mechanisms not directly affecting repair 520 activity. 521 522 While LOH is an important silencing mechanism of the wild type allele [Alemayehu et al., 2007], so 523 far it has not been considered as a point of evidence for MMR gene variant classification by InSiGHT. 524 This is due to several reasons, including multiple observations of loss of variant pathogenic alleles in 525 cancers from MMR gene carriers and technical difficulties, ie, due to the potential presence of MSI 526 hampering analysis of LOH using microsatellite markers [Hofstra et al., 2008]. The findings from this 527 study, especially the observation of loss of the variant allele in samples from carriers of Class 1-not 528 pathogenic and Class 2–likely not pathogenic variants confirm that LOH should be considered with 529 caution for the interpretation of variant pathogenicity in the MMR genes. Studies on large series are 530 needed to assess the usefulness of this marker and its predictive value. 531 23 This article is protected by copyright. All rights reserved. 532 The evaluation of multiple clinical parameters and functional assays undertaken in this study allows 533 refining the strategy for the clinical classification of MMR gene variants. Intronic and synonymous 534 variants that cannot be tested in the in vitro MMR assay should be assessed for effects on RNA 535 processing, by detection of aberrant transcripts (in the presence of NMD inhibitors) and allele-specific 536 expression 537 observed, the variant could be assigned to Class 1-not pathogenic, even without further evidence 538 (from ie, segregation and tumor characteristics), as stated by ENIGMA for BRCA1 and BRCA2. The 539 underlying rationale is that the probability that an intronic variant with no documented splicing 540 aberration will cause high tumor risk is very low, < 1/1,000. (in the absence of NMD inhibitors). We suggest that, when no major alteration is 541 542 For potential missense variants, concordant evidence in favour or against pathogenicity should be 543 derived both from functional assays - RNA first, and if normal, protein - and clinical data. Given the 544 importance of obtaining segregation and molecular tumor results for the purpose of variant 545 classification, any attempt should be made to test additional patients and samples, especially from 546 carriers of missense variants which are usually more difficult to classify compared to silent and 547 intronic changes. 548 549 Finally, classifications obtained by multifactorial analysis should be supported by multiple data 550 points; this could be achieved by requiring a minimum threshold or different points of evidence from 551 clinical and tumor data to allow assignment to a clinically actionable class. 552 553 humu23117-sup-0001-SuppMat.docx 554 555 Supplementary Information Figure S1. Identification of MSH2 splicing effects by using a splicing minigene reporter assay. (A) Structure of the pCAS2-MSH2 minigenes. Boxes represent 24 This article is protected by copyright. All rights reserved. 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 exons and lines in between indicate introns. MSH2 segments are shown in dark colour. Splicing events detected in the minigene assay are indicated by the dotted lines and further described on the right. (B) Analysis of the splicing pattern of wild-type and mutant pCAS2-MSH2 minigene transcripts. Wild-type and mutant minigenes, as indicated, were transfected into HeLa cells and the minigene transcripts were analyzed by RT-PCR. The image shows a 2% agarose gel stained with ethidium bromide, visualized by exposure to ultraviolet light under conditions of non-saturating exposure. The identities of the RT-PCR products are shown both on the left and on the right of the gel. Results are representative of 3 independent experiments. M, size marker (100 bp DNA ladder, New England Biolabs); pCAS2, empty vector; WT, wild-type. Supplementary Information Figure S2. Analysis of the impact on splicing of MLH1 variants by using a minigene reporter assay. (A) Structure of the pCAS2-MLH1 minigenes. Boxes represent exons and horizontal lines in between indicate introns. MLH1 segments are shown in dark colour. Splicing events detected in the minigene assay are indicated by the dotted lines and further described on the right. (B) Analysis of the splicing pattern of wild-type and mutant pCAS2-MLH1 minigene transcripts. Wild-type and mutant minigenes, as indicated, were transfected into HeLa cells and the minigene transcripts were analyzed by RT-PCR. The image shows a 2% agarose gel stained with ethidium bromide, visualized by exposure to ultraviolet light under conditions of non-saturating exposure. Results are representative of 3 independent experiments. M, size marker (100 bp DNA ladder, New England Biolabs); pCAS2, empty vector; WT, wild-type. Supplementary Information Fig. S3. MLH1 c.301G>A alters the alternative splicing pattern of MLH1 exon 3 in the minigene splicing assay. (A) Structure of the pSPL3m-MLH1ex3 minigene. Boxes represent exons and horizontal lines in between indicate introns. The MLH1 segment is shown in dark colour. Splicing events detected in the minigene assay are indicated by the dotted lines. (B) Analysis of the splicing pattern of wild-type and mutant pSPL3m-MLH1ex3 minigene transcripts. Wild-type and mutant minigenes, as indicated, were transfected into HeLa cells and the minigene transcripts were analyzed by RT-PCR as described under Materials and Methods. The image shows a 2% agarose gel stained with ethidium bromide, visualized by exposure to ultraviolet light under conditions of non-saturating exposure. The identities of the RT-PCR products are shown on the left and below the gel. (C) Usage of the reference 5’ss and a5’ss of MLH1 exon 3 (NM_000249.3 and NM_001167617.1, respectively) in the WT and mutant contexts. The upper panel shows in silico predictions for the effect of c.301G>A on the strength of the 5’ splice site of MLH1exon 3 (predictions obtained with 5 different algorithms, as described in Soukarieh et al., 2016). The bottom panel shows the sequence of the RT-PCR products indicated by the star (heteroduplexes) and purified from the gel shown in B. 5’ss, 5’splice site; a5’ss, alternative 5’ splice site; Δ5 nts, deletion of the last 5 nucleotides of MLH1 exon 3. Supplementary Information Table S1. Clinical data and co-occurrence of multiple variants in families with MMR gene variants. Supplementary Information Table S2. Segregation and multifactorial likelihood analysis for the investigated MMR gene variants. 595 596 597 ACKNOWLEDGMENTS 25 This article is protected by copyright. All rights reserved. 598 599 MG has been supported by a grant from Istituto Toscano Tumori (ITT). BAT is supported by an 600 NHMRC Early Career Fellowship (ID1091211). ABS is supported by an NHMRC Senior Research 601 Fellowship (ID1061779). Aspects of this research (bioinformatic interpretation) were supported by an 602 NIH subcontract (grant ID NIH R01CA164944). MN has been supported by a grant from the 603 European Research Council (2008-AdG-232635). Part of this project was supported by a grant from 604 the French Institut National du Cancer/Direction Générale de l’Offre de Soins (INCa/DGOS) and the 605 Fondation ARC pour la Recherche sur le Cancer to AM. OS was funded by a fellowship from the 606 French Ministry of Education. The authors declare they have no conflict of interest. 607 608 InSiGHT VIC CONTRIBUTORS. Kiwamu Akagi, Div. Molecular Diagnosis & Cancer Prevention 609 Saitama Cancer Center, Saitama, Japan. Fahd Al-Mullah, Molecular Pathology Unit, Health Sciences 610 Center, Faculty of Medicine, Kuwait University, Safat Kuwait, Kuwait. Ian R Berry, Leeds Genetics 611 Laboratory, Leeds, UK. Michael Farrell, Department of Cancer Genetics, Mater Private Hospital, 612 Dublin, Ireland. Susan Farrington, Institute of Genetics and Molecular Medicine, University of 613 Edinburgh, UK. Ian Frayling, Institute of Cancer & Genetics, Cardiff University, Cardiff, UK. Elke 614 Holinski-Feder, Medizinische Klinik und Poliklinik IV, Campus Innenstadt, Klinikum der Universität 615 München; and MGZ – Medizinisch Genetisches Zentrum, Munich, Germany. Maija Kohonen-Corish, 616 The Kinghorn Cancer Centre, Garvan Institute of Medical Research, Sydney; St Vincent's Clinical 617 School, UNSW Australia, Sydney; and School of Medicine, Western Sydney University, Sydney, 618 Australia. Kristina Lagerstedt-Robinson, Department of Molecular Medicine and Surgery, Karolinska 619 Institutet, and Department of Clinical Genetics, Karolinska University Hospital, Solna, Stockholm, 620 Sweden. Finlay Macrae, Dept of Colorectal Medicine and Genetics, The Royal Melbourne Hospital, 621 Victoria, Australia. Pål Møller, Research Group Inherited Cancer, Department of Medical Genetics, 622 The Norwegian Radium Hospital, Oslo University Hospital, Norway. Monika Morak, Medizinische 26 This article is protected by copyright. All rights reserved. 623 Klinik und Poliklinik IV, Campus Innenstadt, Klinikum der Universität München; and MGZ – 624 Medizinisch Genetisches Zentrum, Munich, Germany. John-Paul Plazzer, Department of Colorectal 625 Medicine and Genetics, Royal Melbourne Hospital, Melbourne, Victoria, Australia. Lene Rasmussen, 626 Faculty of Health Sciences, University of Copenhagen. Brigitte Royer-Pokora, Institut für 627 Humangenetik und Anthropologie Heinrich Heine Universität Düsseldorf, Düsseldorf, Germany. 628 Rodney J. Scott, School of Biomedical Sciences and Pharmacy, Faculty of Health, University of 629 Newcastle; Information Based Medicine Program, Hunter Medical Research Institute, Newcastle; 630 Division of Molecular Medicine, Pathology North (Newcastle), John Hunter Hospital, Australia. Rolf 631 H. Sijmons, Dept of Genetics, University Medical Center Groningen, University of Groningen, 632 Groningen, the Netherlands. 633 634 635 REFERENCES 636 637 Alemayehu A, Tomkova K, Zavodna K, Ventusova K, Krivulcik T, Bujalkova M, Bartosova 638 Z, Fridrichova I. 2007. The role of clinical criteria, genetic and epigenetic alterations in 639 Lynch-syndrome diagnosis. Neoplasma 54:391-401. 640 641 Arnold S, Buchanan DD, Barker M, Jaskowski L, Walsh MD, Birney G, Woods MO, Hopper 642 JL, Jenkins MA, Brown MA, Tavtigian SV, Goldgar DE, Young JP, Spurdle AB. 2009. 643 Classifying MLH1 and MSH2 variants using bioinformatic prediction, splicing assays, 644 segregation, and tumor characteristics. Hum Mutat 30:757-770. 27 This article is protected by copyright. All rights reserved. 645 646 Auclair J, Busine MP, Navarro C, Ruano E, Montmain G, Desseigne F, Saurin JC, Lasset C, 647 Bonadona V, Giraud S, Puisieux A, Wang Q. 2006. Systematic mRNA analysis for the effect 648 of MLH1 and MSH2 missense and silent mutations on aberrant splicing. Hum Mutat 27:145- 649 154. 650 651 Betz B, Theiss S, Aktas M, Konermann C, Goecke TO, Möslein G, Schaal H, Royer-Pokora 652 B. 2010. Comparative in silico analyses and experimental validation of novel splice site and 653 missense mutations in the genes MLH1 and MSH2. J Cancer Res Clin Oncol 136:123-134. 654 655 Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, Meltzer SJ, 656 Rodriguez-Bigas MA, Fodde R, Ranzani GN, Srivastava S. 1998. A National Cancer Institute 657 Workshop on Microsatellite Instability for cancer detection and familial predisposition: development 658 of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer 659 Res 58:5248-5257. 660 661 Buhard O, Cattaneo F, Wong YF, Yim SF, Friedman E, Flejou JF, Duval A, Hamelin R. 2006. 662 Multipopulation analysis of polymorphisms in five mononucleotide repeats used to determine the 663 microsatellite instability status of human tumors. J Clin Oncol 24:241-251. 664 28 This article is protected by copyright. All rights reserved. 665 Castellsagué E, González S, Guinó E, Stevens KN, Borràs E, Raymond VM, Lázaro C, Blanco I, 666 Gruber SB, Capellá G. 2010. Allele-specific expression of APC in adenomatous polyposis families. 667 Gastroenterology 139:439-447. 668 669 Christensen LL, Kariola R, Korhonen MK, Wikman FP, Sunde L, Gerdes AM, Okkels H, 670 Brandt CA, Bernstein I, Hansen TV, Hagemann-Madsen R, Andersen CL, Nyström M, 671 Ørntoft TF. 2009. Functional characterization of rare missense mutations in MLH1 and 672 MSH2 identified in Danish colorectal cancer patients. Fam Cancer 8:489-500. 673 674 Couch FJ, Rasmussen LJ, Hofstra R, Monteiro AN, Greenblatt MS, de Wind N. IARC 675 Unclassified Genetic Variants Working Group. 2008. Assessment of functional effects of 676 unclassified genetic variants. Hum Mutat 29:1314-1326. 677 678 Crucianelli F, Tricarico R, Turchetti D, Gorelli G, Gensini F, Sestini R, Giunti L, Pedroni M, Ponz de 679 Leon M, Civitelli S, Genuardi M. 2014. MLH1 constitutional and somatic methylation in patients 680 with MLH1 negative tumors fulfilling the revised Bethesda criteria. Epigenetics 9(10):1431-1438. 681 682 den Dunnen JT, Dalgleish R, Maglott DR, Hart RK, Greenblatt MS, McGowan-Jordan J, Roux AF, 683 Smith T, Antonarakis SE, Taschner PE. 2016. HGVS Recommendations for the description of 684 sequence variants: 2016 update. Hum Mutat. 37:564-569. 685 29 This article is protected by copyright. All rights reserved. 686 Di Giacomo D, Gaildrat P, Abuli A, Abdat J, Frébourg T, Tosi M, Martins A. 2013. Functional 687 analysis of a large set of BRCA2 exon 7 variants highlights the predictive value of hexamer scores in 688 detecting alterations of exonic splicing regulatory elements. Hum Mutat 34:1547-1557. 689 690 Drost M1, Zonneveld Je, van Dijk L, Morreau H, Tops CM, Vasen HF, Wijnen JT, de Wind 691 N. 2010. A cell-free assay for the functional analysis of variants of the mismatch repair 692 protein MLH1. Hum Mutat 31(3):247-253. 693 694 Furukawa T, Konishi F, Masubuchi S, Shitoh K, Nagai H, Tsukamoto T. 2002. Densely 695 methylated MLH1 promoter correlates with decreased mRNA expression in sporadic 696 colorectal cancers. Genes Chromosomes Cancer 35:1-10. 697 698 Gaildrat P, Killian A, Martins A, Tournier I, Frébourg T, Tosi M. 2010. Use of splicing reporter 699 minigene assay to evaluate the effect on splicing of unclassified genetic variants. Methods Mol Biol 700 653:249-257. 701 702 Genuardi M, Viel A, Bonora D, Capozzi E, Bellacosa A, Leonardi F, Valle R, Ventura A, Pedroni M, 703 Boiocchi M, Neri G. 1998. Characterization of MLH1 and MSH2 alternative splicing and its 704 relevance to molecular testing of colorectal cancer susceptibility. Hum Genet 102:15-20. 705 30 This article is protected by copyright. All rights reserved. 706 Giunti L, Cetica V, Ricci U, Giglio S, Sardi I, Paglierani M, Andreucci E, Sanzo M, Forni M, 707 Buccoliero AM, Genitori L, Genuardi M. 2009. Type A micro satellite instability in pediatric gliomas 708 as an indicator of Turcot syndrome. Eur J Hum Genet 17:919-927. 709 710 Gylling A, Abdel-Rahman WM, Juhola M, Nuorva K, Hautala E, Jarvinen HJ, Mecklin JP, Aarnio M, 711 Peltomaki P. 2007. Is gastric cancer part of the tumour spectrum of hereditary non-polyposis 712 colorectal cancer? A molecular genetic study. Gut 56: 926–33. 713 714 Hardt K, Heick SB, Betz B, Goecke T, Yazdanparast H, Küppers R, Servan K, Steinke V, 715 Rahner N, Morak M, Holinski-Feder E, Engel C et al. 2011. Missense variants in hMLH1 716 identified in patients from the German HNPCC consortium and functional studies. Fam 717 Cancer 10: 273-284. 718 719 Hinrichsen I, Brieger A, Trojan J, Zeuzem S, Nilbert M, Plotz G. 2013. Expression defect 720 size among unclassified MLH1 variants determines pathogenicity in Lynch syndrome 721 diagnosis. Clin Cancer Res 19: 2432-2441. 722 723 Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. 1989. Site-directed mutagenesis 724 byoverlap extension using the polymerase chain reaction. Gene 77:51–59. 725 31 This article is protected by copyright. All rights reserved. 726 Hofstra RM, Spurdle AB, Eccles D, Foulkes WD, de Wind N, Hoogerbrugge N, Hogervorst FB, 727 IARC Unclassified Genetic Variants Working Group. 2008. Tumor characteristics as an analytic tool 728 for classifying genetic variants of uncertain clinical significance. Hum Mutat 29:1292-1303. 729 730 Jakubowska A, Górski B, Kurzawski G, Debniak T, Hadaczek P, Cybulski C, Kladny J, 731 Oszurek O, Scott RJ, Lubinski J. 2001. Optimization of experimental conditions for RNA- 732 based sequencing of MLH1 and MSH2 genes. Hum Mutat 17:52-60. 733 734 Janssen MJ, Waanders E, Te Morsche RH, Xing R, Dijkman HB, Woudenberg J, Drenth JP. 2011. 735 Secondary, somatic mutations might promote cyst formation in patients with autosomal dominant 736 polycystic liver disease. Gastroenterology 141:2056-2063. 737 738 Janssen MJ, Salomon J, Te Morsche RH, Drenth JP. 2012. Loss of heterozygosity is present in SEC63 739 germline carriers with polycystic liver disease. PLoS One 7(11):e50324. 740 741 Kansikas M, Kariola R, Nyström M. 2011. Verification of the three-step model in assessing 742 the pathogenicity of mismatch repair gene variants. Hum Mutat 32:107-115. 743 744 Kantelinen J, Kansikas M, Candelin S, Hampel H, Smith B, Holm L, Kariola R, Nyström M. 745 2012. Mismatch repair analysis of inherited MSH2 and/or MSH6 variation pairs found in 746 cancer patients. Hum Mutat 33:1294-1301. 32 This article is protected by copyright. All rights reserved. 747 748 Kariola R, Raevaara TE, Lönnqvist KE, Nyström-Lahti M. 2002. Functional analysis of 749 MSH6 mutations linked to kindreds with putative hereditary non-polyposis colorectal cancer 750 syndrome. Hum Mol Genet 11:1303-1310. 751 752 Lucci-Cordisco E, Zito I, Gensini F, Genuardi M. 2003. Hereditary nonpolyposis colorectal 753 cancer and related conditions. Am J Med Genet A 122A:325-334. 754 755 Lynch HT, de la Chapelle A. 2003. Hereditary colorectal cancer. N Engl J Med 348:919-932. 756 757 Mancini I, Santucci C, Sestini R, Simi L, Pratesi N, Cianchi F, Valanzano R, Pinzani P, Orlando C. 758 2010. The use of COLD-PCR and high-resolution melting analysis improves the limit of detection of 759 KRAS and BRAF mutations in colorectal cancer. J Mol Diagn 12:705-711. 760 761 Montera M, Resta N, Simone C, Guanti G, Marchese C, Civitelli S, Mancini A, Pozzi S, De 762 Salvo L, Bruzzone D, Donadini A, Romio L, et al. 2000. Mutational germline analysis of 763 hMSH2 and hMLH1 genes in early onset colorectal cancer patients. J Med Genet 37(7):E7. 764 765 Niroula A, Vihinen M. Classification of Aminoacid Substitutions in Mismatch Repair Proteins Using 766 PON-MMR2. 2015. Hum Mutat 36:1128–1134. 33 This article is protected by copyright. All rights reserved. 767 768 Nyström-Lahti M, Perrera C, Räschle M, Panyushkina-Seiler E, Marra G, Curci A, 769 Quaresima B, Costanzo F, D'Urso M, Venuta S, Jiricny J. 2002. Functional analysis of 770 MLH1 mutations linked to hereditary nonpolyposis colon cancer. Genes Chromosomes 771 Cancer 33:160-167. 772 773 Ollila S, Sarantaus L, Kariola R, Chan P, Hampel H, Holinski-Feder E, Macrae F, Kohonen- 774 Corish M, Gerdes AM, Peltomäki P, Mangold E, de la Chapelle A, Greenblatt M, Nyström 775 M. 2006. Pathogenicity of MSH2 missense mutations is typically associated with impaired 776 repair capability of the mutated protein. Gastroenterology 131:1408-1417. 777 778 Ollila S, Dermadi Bebek D, Greenblatt M, Nyström M. 2008. Uncertain pathogenicity of 779 MSH2 variants N127S and G322D challenges their classification. Int J Cancer 123(3):720- 780 724. 781 782 Pastrello C, Pin E, Marroni F, Bedin C, Fornasarig M, Tibiletti MG, Oliani C, Ponz de Leon M, Urso 783 ED, Della Puppa L, Agostini M, Viel A. 2011. Integrated analysis of unclassified variants in 784 mismatch repair genes. Genet Med 13:115-124. 785 34 This article is protected by copyright. All rights reserved. 786 Perera S, Li B, Tsitsikotas S, Ramyar L, Pollett A, Semotiuk K, Bapat B. 2010. A novel and rapid 787 method of determining the effect of unclassified MLH1 genetic variants on differential allelic 788 expression. J Mol Diagn 12:757-764. 789 790 Petersen SM, Dandanell M, Rasmussen LJ, Gerdes AM, Krogh LN, Bernstein I, Okkels H, Wikman 791 F, Nielsen FC, Hansen TV. 2013. Functional examination of MLH1, MSH2, and MSH6 intronic 792 mutations identified in Danish colorectal cancer patients. BMC Med Genet 14:103. 793 794 Plon SE, Eccles DM, Easton D, Foulkes WD, Genuardi M, Greenblatt MS, Hogervorst FB, 795 Hoogerbrugge N, Spurdle AB, Tavtigian SV; IARC Unclassified Genetic Variants Working Group. 796 2008. Sequence variant classification and reporting: recommendations for improving the 797 interpretation of cancer susceptibility genetic test results. Hum Mutat 29:1282-1291. 798 799 Raevaara TE, Korhonen MK, Lohi H, Hampel H, Lynch E, Lönnqvist KE, Holinski-Feder E, 800 Sutter C, McKinnon W, Duraisamy S, Gerdes AM, Peltomäki P, Kohonen-Ccorish M, 801 Mangold E, Macrae F, Greenblatt M, de la Chapelle A, Nyström M. 2005. Functional 802 significance and clinical phenotype of nontruncating mismatch repair variants of MLH1. 803 Gastroenterology 129:537-549. 804 805 Renkonen E, Zhang Y, Lohi H, Salovaara R, Abdel-Rahman WM, Nilbert M, Aittomaki K, Jarvinen 806 HJ, Mecklin JP, Lindblom A, Peltomaki P. 2003. Altered expression of MLH1, MSH2, and MSH6 in 807 predisposition to hereditary nonpolyposis colorectal cancer. J Clin Oncol 21:3629-3637. 35 This article is protected by copyright. All rights reserved. 808 809 Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, Grody WW, Hegde M, Lyon E, 810 Spector E, Voelkerding K, Rehm HL; ACMG Laboratory Quality Assurance Committee. 811 2015. Standards and guidelines for the interpretation of sequence variants: a joint consensus 812 recommendation of the American College of Medical Genetics and Genomics and the 813 Association for Molecular Pathology. Genet Med 17:405-424. 814 815 Roncari B, Pedroni M, Maffei S, Di Gregorio C, Ponti G, Scarselli A, Losi L, Benatti P, Roncucci L, 816 De Gaetani C, Camellini L, Lucci-Cordisco E, Tricarico R, Genuardi M, Ponz de Leon M. 2007. 817 Frequency of constitutional MSH6 mutations in a consecutive series of families with clinical 818 suspicion of HNPCC. Clin Genet 72:230-237. 819 820 Sijmons RH, Greenblatt MS, Genuardi M. Gene variants of unknown clinical significance in 821 Lynch syndrome. An introduction for clinicians. 2013. Fam Cancer 12:181-187. 822 823 Soukarieh O, Gaildrat P, Hamieh M, Drouet A, Baert-Desurmont S, Frébourg T, Tosi M, 824 Martins A. 2016. Exonic splicing mutations are more prevalent than currently estimated and 825 can be predicted by using in silico tools. PLoS Genet 12:e1005756. 826 36 This article is protected by copyright. All rights reserved. 827 Spurdle AB, Couch FJ, Hogervorst FB, Radice P, Sinilnikova OM; IARC Unclassified 828 Genetic Variants Working Group. 2008. Prediction and assessment of splicing alterations: 829 implications for clinical testing. Hum Mutat 29:1304-1313. 830 831 Suraweera N, Duval A, Reperant M, Vaury C, Furlan D, Leroy K, Seruca R, Iacopetta B, Hamelin R. 832 2002. Evaluation of tumor microsatellite instability using five quasimonomorphic mononucleotide 833 repeats and pentaplex PCR. Gastroenterology 123:1804-1811. 834 835 Takahashi M, Shimodaira H, Andreutti-Zaugg C, Iggo R, Kolodner RD, Ishioka C. 2007. 836 Functional analysis of human MLH1 variants using yeast and in vitro mismatch repair assays. 837 Cancer Res 67: 4595-4604. 838 839 Tavtigian SV, Byrnes GB, Goldgar DE, Thomas A. 2008. Classification of rare missense 840 substitutions, using risk surfaces, with genetic- and molecular-epidemiology applications. 841 Hum Mutat 29:1342-54. 842 843 Tavtigian SV, Greenblatt MS, Goldgar DE, Boffetta P; IARC Unclassified Genetic Variants 844 Working Group. 2008. Assessing pathogenicity: overview of results from the IARC 845 Unclassified Genetic Variants Working Group. Hum Mutat 29:1261-1264. 846 37 This article is protected by copyright. All rights reserved. 847 Thompson D, Easton DF, Goldgar DE. 2003. A full-likelihood method for the evaluation of causality 848 of sequence variants from family data. Am J Hum Genet 73:652-655. 849 850 Thompson BA, Goldgar DE, Paterson C, Clendenning M, Walters R, Arnold S, Parsons MT, Michael 851 D W, Gallinger S, Haile RW, Hopper JL, Jenkins MA et al. 2013a. A multifactorial likelihood model 852 for MMR gene variant classification incorporating probabilities based on sequence bioinformatics and 853 tumor characteristics: a report from the Colon Cancer Family Registry. Hum Mutat 34:200-209. 854 855 Thompson BA, Greenblatt MS, Vallee MP, Herkert JC, Tessereau C, Young EL, Adzhubey IA, Li B, 856 Bell R, Feng B, Mooney SD, Radivojac P et al. 2013b. Calibration of multiple in silico tools for 857 predicting pathogenicity of mismatch repair gene missense substitutions. Hum Mutat 34:255-265. 858 859 Thompson BA, Spurdle AB, Plazzer JP, Greenblatt MS, Akagi K, Al-Mulla F, Bapat B, Bernstein I, 860 Capellá G, den Dunnen JT, du Sart D, Fabre A et al. 2014. Application of a 5-tiered scheme for 861 standardized classification of 2,360 unique mismatch repair gene variants in the InSiGHT locus- 862 specific database. Nat Genet 46:107-115. 863 864 Thompson BA, Martins A, Spurdle AB. 2015. A review of mismatch repair gene transcripts: issues 865 for interpretation of mRNA splicing assays. Clin Genet 87:100-108. 866 867 Tournier I, Vezain M, Martins A, Charbonnier F, Baert-Desurmont S, Olschwang S, Wang Q, Buisine 868 MP, Soret J, Tazi J, Frébourg T, Tosi M. 2008. A large fraction of unclassified variants of the 38 This article is protected by copyright. All rights reserved. 869 mismatch repair genes MLH1 and MSH2 is associated with splicing defects. Hum Mutat 29:1412- 870 1424. 871 872 van der Klift H, Jansen AML, van der Steenstraten N, Bik EC, Tops CMJ, Devilee P, Wijnen JT. 873 2015. Splicing analysis for exonic and intronic mismatch repair gene variants associated with Lynch 874 syndrome confirms high concordance between minigene assays and patient RNA analyses. Mol Genet 875 Genom Med 3:327-345. 876 877 Vasen HF. 2005. Clinical description of the Lynch syndrome [hereditary nonpolyposis 878 colorectal cancer (HNPCC)]. Fam Cancer 4:219-225. 879 880 Vasen HF, Blanco I, Aktan-Collan K, Gopie JP, Alonso A, Aretz S, Bernstein I, Bertario L, Burn J, 881 Capella G, Colas C, Engel C et al. 2013. Revised guidelines for the clinical management of Lynch 882 syndrome (HNPCC): recommendations by a group of European experts. Gut 62:812-823. 883 884 Xie J, Guillemette S, Peng M, Gilbert C, Buermeyer A, Cantor SB. 2010. An MLH1 mutation 885 links BACH1/FANCJ to colon cancer, signaling, and insight toward directed therapy. Cancer 886 Prev Res (Phila) 3:1409-1416. 887 888 889 FIGURE LEGENDS 39 This article is protected by copyright. All rights reserved. 890 891 Figure 1. Identification of MSH2 exon 13 splicing alterations by using a splicing minigene 892 reporter assay. (A) Structure of the pCAS2-MSH2ex13 minigene. Boxes represent exons and 893 horizontal lines in between indicate introns. The MSH2 segment is shown in dark colour. Splicing 894 events detected in the minigene assay are indicated by the dotted lines and further described on the 895 right. (B) Analysis of the splicing pattern of wild-type and mutant pCAS2-MSH2ex13 minigene 896 transcripts. Wild-type and mutant minigenes, as indicated, were transfected into HeLa cells and the 897 minigene transcripts were analyzed by RT-PCR as described under Materials and Methods. The 898 image shows a 2% agarose gel stained with ethidium bromide, visualized by exposure to ultraviolet 899 light under conditions of non-saturating exposure. The identities of the RT-PCR products are shown 900 on the right. Results are representative of 3 independent experiments. Marker, 100 bp DNA ladder 901 (New England Biolabs); pCAS2, empty vector; WT, wild-type; *, heteroduplexes. 902 40 This article is protected by copyright. All rights reserved. 903 Figure 2. Splicing analysis by RT-PCR on cDNA from a LCL established from a carrier of 904 MSH2 variant c.2006G>T. (A) Gel electrophoresis of cDNA PCR products obtained using primers 905 located in MSH2 exons 12 and 14. MW: molecular weight marker (100 bp ladder). The upper band 906 corresponds to the full length mRNA product, the fainter lower band to the isoform lacking exon 13. 907 The arrow next to the asterisk shows the heteroduplex formed by the two PCR products. (B) 908 Schematic representation of MSH2 mRNA encompassing exons 12-14 and of the two splicing 909 products detected in the LCL sample. (C) Electropherogram of the sequence of the ∆13 cDNA 910 product. (D) Electropherogram of the sequence of the full-length cDNA product showing presnece of 911 both alleles at c.2006G>T (the reverse strand is shown). 912 913 Figure 3. Expression and functional analyses of the 5 MLH1 and 4 MSH2 missense variants. (A- 914 B) Western Blot analysis of total protein extracts from Sf9 cells coinfected with baculovirus 915 constructs expressing PMS2 wild-type protein (PMS2 WT) with either MLH1 WT or with MLH1 916 variant, and MSH6 WT with either MSH2 WT or MSH2 variant, and showing instability of proteins 41 This article is protected by copyright. All rights reserved. 917 MLH1 p.Leu260Arg, MSH2 p.Phe694Ser, and MSH2 p.Pro696Leu. α-tubulin was used as a loading 918 control. (C-D) Repair efficiency (R%) of the recombinant MutLα (MLH1+PMS2) and Mutsα 919 (MSH2+MSH6) protein variants measured in the in vitro MMR assay and calculated as the ratio of 920 double digested DNA relative to total DNA added to the reaction, and showing functional deficiency 921 in MLH1 p.Gly101Ser, MLH1 p.Leu260Arg, MSH2 p.Gly669Val, MSH2 p.Phe694Ser, and MSH2 922 p.Pro696Leu. R% corresponds to the assay results shown in the figure, X3R% denotes the average of 923 three independent experiments with standard deviations (±). Nuclear extract free MOCK and 924 uncomplemented MMR-deficient HCT116 NE and LoVo NE serve as negative controls and HCT116 925 and LoVo NE complemented by MutLα and Mutsα, respectively, serve as positive controls. The top 926 fragment (3193 bp) shows the migration of unrepaired linearized G·T mismatch containing construct 927 and the two smaller fragments (1833 bp and 1360 bp) represent the repaired and double digested 928 fragments. (E) The relative repair % was calculated in respect to the wild type control (MutLα or 929 MutSα respectively) set as 100 (100, X3R%) according to Drost et al., 2010, and Kantelinen et al., 930 2012. 42 This article is protected by copyright. All rights reserved. 931 932 Tabel 1.Population frequency and tumor molecular and immunohistochemical characteristics of MMR gene variants. Sequence variant Population frequency Tumor data Ge ne Loca tion Nucleoti de change1 Predicted aminoaci d change1 ExA C2,3 Italian control chromos omes4 MSI stat us6 IHC6 BRAF p.Val60 0Glu6 MLH1 promot er methyla tion6,7 LO H6 ML 3 c.301G> p.(Gly101 nr 0 MSI nt nt wt nt 43 This article is protected by copyright. All rights reserved. A H1 3i c.30719A>G 9 c.779T> G 11i c.10398T>A 12 c.1043T >C 12 c.1217G >A Ser) -H (1) 0.00 020 0 MS S (1) normal (1) nt nt no (1) nr 0 MSI -H (6); MS S (1) MLH1 loss (5: all MSIH tumors ) wt (3) wt (3: all with MLH1 loss) no (3) 0.00 155 0.02187 5 MSI -H (1) MLH1 loss (1: tumor with MSI status unkno wn); MSH2 loss (1: MSI-H tumor) nt nt nt p.(Leu34 8Ser) 0.00 002 0 MS S (1) normal (1) nt nt no (1) p.(Ser406 Asn) 0.00 089 0.00333 MSI -H (2) MLH1 loss (1); MSH2/ MSH6 loss (1) nt wt (1: tumor with MLH1 loss) vari ant allel e (1: tum or wit h ML H1 loss ) p.(Leu26 0Arg) 44 This article is protected by copyright. All rights reserved. 13 c.1421G >C nr 0 MSI -H MLH1 loss mut nt vari ant allel e 13i c.1558+ 14G>A 0.03 948 0.02187 MSI -H (1) MLH1 loss (1: tumor with MSI status unkno wn); MSH2 loss(1: MSI-H tumor) nt nt nt 15i c.173219T>C nr 0 MS S (1) normal nt nt vari ant allel e 16 c.1743G >A p.(Pro581 =) 0.00 008 0 MSI -H (1) MLH1 loss mut wt nt 16 c.1814A >C p.(Glu605 Ala) nr 0 MSI -H (1) MLH1 loss nt nt no 18 c.2041G >A p.(Ala681 Thr) nr 0 MS S (1); MSI -H (1) MLH1 loss (1: MSS tumor) ; normal (1: MSI-H tumor) wt (1: MSI-H tumor) wt (1: MSI-H tumor) no (1: MSI -H tum or) 18 c.2059C >T p.(Arg687 Trp) 0.00 003 0 MSI -H (2); MS S normal (1: MSI-H tumor) ; MLH1 wt (1: MSI-H tumor) nt nt p.(Arg474 Pro) 6 45 This article is protected by copyright. All rights reserved. MS H2 (1) loss (1: MSS tumor) 2 c.244A> G p.(Lys82G lu) nr 0 MS S (1) normal nt nt no (1) 3 c.380A> G p.(Asn12 7Ser) 0.00 692 nt nt nt nt nt nt 81 c.13878G>T 0.00 194 nt MSI -L (1); MS S (1) normal (1: MSI-L tumor) ; MLH1 loss (1: MSS tumor) nt nt no (1: MS S tum or) 9i c.15119A>T 0.08 400 0.07333 MSI -H (5); MS S (2) normal (3: 1 MSI-H and 2 MSS tumors ); MSH2/ MSH6 loss (4: all MSIH tumors ) nt meth (1: MSIH tumor) nt 11 c.1666T >C 0.00 437 0 MSI -H (1); MS S (1) normal (2) nt wt (1: MSI-H tumor) nt p.(Pro556 =) 46 This article is protected by copyright. All rights reserved. 11 c.1737A >G 0.00 192 0 MSI -L (1); MS S (1) normal (1: MSI-L tumor) ; MLH1 loss (1: MSS tumor) nt nt no (1: MS S tum or) 12i c.20066T>C 0.11 500 nt MSI -H (2) MSH2/ MSH6 loss (2) nt nt nt 13 c.2006G >T reported as p.(Pro670 Leufs*) (predicte d missense change: p.(Gly669 Val) nr 0 MSI -H (1) MSH2/ MSH6 loss (1) nt nt no (1) 13 c.2081T >C p.(Phe69 4Ser) nr 0 MSI -H (3) MSH2/ MSH6 loss (3) nt nt nt 13 c.2087C >T p.(Pro696 Leu) nr 0 MSI -H (3) MSH2/ MSH6 loss (1) nt nt no (1) 14 c.2442T >G p.(Leu81 4=) nr nt MSI -H (1) MSH2/ MSH6 loss (1) nt nt no (1) p.(Lys579 =) 933 934 1 Previously unclassified variants are indicated in bold. 935 2 nr = not reported. 936 3 ExAC: http://exac.broadinstitute.org/; TGCA allele frequencies are excluded. 937 4 nt = not tested. 938 5 c.1039-8T>A and c.1558+14G>A are in linkage disequilibrium in the Italian population. 47 This article is protected by copyright. All rights reserved. 939 6 In brackets number of samples; nt = not tested. 940 7 wt = not methylated, meth = methylated. 941 942 943 Table 9442. Effects of MMR gene variants on RNA processing, protein stability and in vitro MMR activity. 945 Gen e Sequence variant Locati on ML H1 3 Nucleotid e change1 c.301G>A mRNA analysis2,3 Predicted aminoacid change1 p.(Gly101Se r) Studies on patient samples Splicing analysis SNUPE assay nt no allelic imbala nce 48 This article is protected by copyright. All rights reserved. Functional analysis3,4 Minigen e splicing assay4 Mamma lian protein stability Mamma lian mmr activity total inclusio n of referenc e exon 3 (Tournie r et al., 2008; this study) with concomi tant loss of alternati ve 5’ss five nucleoti des upstrea m the referenc e5’ ss Stable Deficien t (this study) 3i c.30719A>G 9 c.779T>G 11i c.10398T>A 12 c.1043T>C 12 nt nt no effect (Tournie r et al. 2008) na na no allelic imbala nce no effect Unstabl e Deficien t partial loss of exon 12 nt no effect (Peterse n et al., 2013; this study) na na p.(Leu348Se r) no effect no allelic imbala nce no effect Stable Proficie nt c.1217G> A p.(Ser406As n) no effect no allelic imbala nce (Pastre llo et al. 2011 and this study) no effect Stable (Takaha shi et al. 2007) Proficie nt (Takaha shi et al. 2007; Drost et al. 2010) 13 c.1421G>C p.(Arg474Pr o) nt nt no effect Stable Proficie nt 13i c.1558+14 G>A nt nt no effect na na 15i c.1732- no effect nt no effect na na p.(Leu260Ar no effect g) (Montera et al., 2000; this study) 49 This article is protected by copyright. All rights reserved. 19T>C MS 16 c.1743G> A p.(Pro581=) nt nt no effect na na 16 c.1814A> C p.(Glu605Al a) nt no allelic imbala nce no effect Stable Proficie nt 18 c.2041G> A p.(Ala681Th r) no effect (Jakubo wska et al., 2001; Betz et al., 2010) no allelic imbala nce no effect (Tournie r et al., 2008) Discord ant results (Raevaa ra et al., 2005; Takahas hi et al., 2007; Xie et al., 2010; Hardt et al., 2011; Hinrichs en et al., 2013) Proficie nt (Raevaa ra et al., 2005; Takahas hi et al., 2007; Hinrichs en et al., 2013) 18 c.2059C>T p.(Arg687Tr p) no effect (Jakubo wska et al., 2001; Furukaw a et al., 2002; Auclair et al., 2006; Arnold et al., 2009; this study) no allelic imbala nce no effect Discord ant results (Takaha shi et al., 2007; Christen sen et al., 2009) Proficie nt (Takaha shi et al., 2007; Christen sen et al., 2009) 2 c. 244A>G p.(Lys82Glu nt no no effect Stable Proficie 50 This article is protected by copyright. All rights reserved. H2 nt allelic imbala nce 3 c.380A>G nt nt partial exclusio n of exon 3 Stable (Kansik as et al., 2011) Proficie nt (Ollila et al., 2008) 8i c.13878G>T nt nt no effect (Tournie r et al., 2008) na na 9i c.15119A>T nt nt no effect (Tournie r et al., 2008) na na 11 c.1666T>C p.(Pro556=) no effect (Auclair et al., 2006) nt no effect (Tournie r et al., 2008) na na 11 c.1737A> G p.(Lys579=) no effect (Auclair et al., 2006) nt no effect (Tournie r et al., 2008) na na 12i c.20066T>C5 no effect (Tournie r et al., 2008) na partial exon 13 exclusio n (Tournie r et al., 2008; this study) na na 13 c.2006G> T partial exon 13 exclusion (this study) nt complet e exon 13 exclusio n (van der Klift Stable Deficien t p.(Asn127Se r) reported as p.(Pro670Le ufs*) (predicted missense 51 This article is protected by copyright. All rights reserved. 13 c.2081T>C 5 change p.(Gly669Va l) complete exon 13 exclusion (van der Klift et al., 2015) et al., 2015; this study) p.(Phe694Se r) nt no allelic imbala nce no effect Unstabl e Deficien t 13 c.2087C>T p.(Pro696Le u) nt no allellic imbala nce no effect (Tournie r et al., 2008) Unstabl e Deficien t 14 c.2442T> G p.(Leu814=) nt nt no effect na na 946 1 947 Previously unclassified variants are shown in bold. 2 nt = not tested. 3 The results shown are from this study, unless otherwise indicated. 948 4 In italics: results of studies using in vitro mammalian assays different from that used in the present study.5 These two variants 949 (MSH2 c.2006-6T>C and c.2081T>C ) were also tested in combination in the minigene assay, since they were found in linkage disequilibrium. 950 951 952 953 Table 3. Clinical classification of MLH1 and MSH2 variants. 954 GENE DNA and predicted 1 protein change MLH 1 c.301G>A; p.(Gly101Ser) InSiGHT Proposed 2, classificatio classification 3 n Rationale for classification4 4 0.99740 5 52 This article is protected by copyright. All rights reserved. Posterior probability of pathogenicit y by multifactori al analysis3 Qualitative criteria  Co-segregation  MSI/IHC data  Deficient MMR c.307-19A>G 2 2 na c.779T>G; p.(Leu260Arg) 5 5 1 c.1039-8T>A 1 1 na 3 0.64637 1 < 0.00100 c.1043T>C; p.(Leu348Ser) c.1217G>A; p.(Ser406Asn) 1 53 This article is protected by copyright. All rights reserved. function  Allelic frequency: 0 (this study)  Intronic location  Normal minigene splicing assay  MSI/IHC data  Allelic frequency: 0.0002 (ExAc); 0 (this study)  Co-segregation  MSI/IHC data  Deficient MMR function  Allelic frequency: 0 (this study)  Intronic location  Allelic frequency: 0.00155 (ExAc); 0.02187 (this study)  MSI/IHC data  No major splicing abnormalities Insufficient data (proficient MMR function; 1 MSS tumor; no major splicing alteration)  Allele frequency: 0.00089 (ExAc); c.1421G>C; p.(Arg474Pro) 3 3 0.09448 c.1558+14G>A 1 1 na 54 This article is protected by copyright. All rights reserved. 0.00333 (this study)  MSI/IHC data  Lack of cosegregation with phenotype (combined segregation likelihood ratio < 0.01)  Co-occurrence of MSH2 truncating variant that segregates with the phenotype in the family  Proficient MMR function  No major splicing abnormalities  Estimated risk from casecontrol studies (1.5) Insufficient data (no major splicing alteration; proficient MMR function; 1 tumor MSI-H BRAF p.Val600Glu positive)  Intronic location  Allelic frequency: 0.03948 (ExAc); 0.02187 (this study)  MSI/IHC data MSH 2  No major splicing abnormalities  Intronic location  1 MSS tumor  No major splicing abnormalities  Synonymous coding change  No major splicing abnormalities by minigene assay  Allelic frequency: 0.00008 (ExAc); 0 (this study)  1 MSI-H MLH1neg BRAF p.Val600Glupos tumor MSI/IHC data c.1732-19T>C na 2 0.01386 c.1743G>A; p.(Pro581=) na 2 na c.1814A>C; p.(Glu605Ala) na 4 0.95294 c.2041G>A; p.(Ala681Thr) 5 5 0.99708  Co-segregation  MSI/IHC data c.2059C>T; p.(Arg687Trp) 5 5 0.99999 c. 244A>G; p.(Lys82Glu) na 2 0.00980  Co-segregation  MSI/IHC data  Homozygosity associated with constitutional mismatch repair deficiency syndrome  MSI/IHC data  Proficient MMR function 55 This article is protected by copyright. All rights reserved. c.380A>G; p.(Asn127Ser) 1 1 na c.1387-8G>T 2 1 0.00088 c.1511-9A>T 1 1 na 56 This article is protected by copyright. All rights reserved.  No major splicing abnormalities  Allelic frequency: 0.0692 (ExAc)  No major splicing abnormalities by minigene assay  Proficient MMR function  Intronic location  Allelic frequency: 0.00194 (ExAc)  MSI/IHC data (> 3 tumors not showing features of MMR deficiency and/or MSH2 inactivation)  No major splicing abnormalities  Intronic location  Allelic frequency: 0.08400 (ExAc); 0.07333 (this study)  MSI/IHC data  No major splicing abnormalities by minigene c.1666T>C; p.(Pro556=) 1 1 < 0.00010     c.1737A>G; p.(Lys579=) 2 1 0.00021     c.2006-6T>C 1 1 na   c.2006G>T; reported as p.(Pro670Leufs *) (predicted missense change p.(Gly669Val) 5 5 0.99906     57 This article is protected by copyright. All rights reserved. assay Synonymous coding change Allele frequency: 0.00437 (ExAc); 0 (this study) MSI/IHC data No major splicing abnormalities Synonymous coding change Allelic frequency: 0.0019 (ExAc); 0 (this study) MSI/IHC data (> 3 tumors not showing features of MMR deficiency and/or MSH2 inactivation) No major splicing abnormalities Intronic location No major splicing abnormalities Co-segregation MSI/IHC data Deficient MMR functional test Contrasting results of RNA splicing analyses c.2081T>C; p.(Phe694Ser) 5 5 0.99990 c.2087C>T; p.(Pro696Leu) 5 5 1 c.2442T>G; p.(Leu814=) na 2 na  Co-segregation  MSI/IHC data  Deficient MMR function  Co-segregation  MSI/IHC data  Deficient MMR function  Synonymous coding change  No major splicing abnormalities by minigene splicing assay  Co-observation of MSH2 Class 5-pathogenic variant (phase unknown) 955 956 957 958 959 1Variants not yet classified by InSiGHT are shown in bold. 2 For previously classified variants, the classification corresponds to that reported on http://insight-group.org/variants/classifications/. 3 na = not available. 4 Classification was achieved by multifactorial analysis, qualitative criteria or both; in italics data obtained at least in part from the present study. 960 58 This article is protected by copyright. All rights reserved.