RNA Analysis Identifies Pathogenic Duplications in MSH2 in Patients With Lynch Syndrome

Abstract

Clinical DNA genetic testing (DGT) for Lynch syndrome (LS) has increased in recent years, helping clarify cancer risks with one important caveat: variants of unknown significance (VUS). VUS pose challenges to physicians because there are no well-defined guidelines for the clinical management of LS patients with VUS.1NCCN. NCCN clinical practice guidelines in oncology V.3.2017: Genetic/Familial High-Risk Assessment: Colorectal. NCCN Clinical Practice Guidelines 2017.Google Scholar Among the LS genes, MSH2 has the highest number of germline structural variants (SV), such as exonic inversions, deletions, and duplications.2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar The vast majority of deletions and inversions are pathogenic.2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar However, little is known about the pathogenicity of exonic duplications in MSH2, and consequently, they are often classified as VUS.2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar Classification of duplications requires functional analysis to show whether they are located adjacent to the original sequence (in tandem) and disrupt gene expression, or if they are located elsewhere in the genome.2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar, 3Liccardo R. et al.Cancer Genet. 2018; 221: 19-24Abstract Full Text Full Text PDF PubMed Scopus (12) Google Scholar, 4Mazzarella R. et al.Genome Res. 1998; 8: 1007-1021Crossref PubMed Scopus (83) Google Scholar, 5Richardson M.E. et al.Genet Med. 2018; 21: 683-693Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar We performed RNA genetic testing (RGT) to evaluate a series of duplications identified in a clinical cohort of more than 185,000 individuals. RNA evidence proved that these duplications were in tandem and was used to reclassify them from VUS to clinically actionable pathogenic/likely pathogenic variants (PV). This shows the diagnostic value of RGT, because reclassification to PV empowers clinicians to offer preventive measures and lifesaving screenings to LS patients. Classification of germline variants was performed by following the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) guidelines.6Richards S. et al.Genet Med. 2015; 17: 405-424Abstract Full Text Full Text PDF PubMed Scopus (5113) Google Scholar DGT, including multi-gene panel testing, array comparative genomic hybridization, and/or multiplex ligation-dependent probe amplification, and RGT, including RNA sequencing7Mercer T.R. et al.Nat Protoc. 2014; 9: 989-1009Crossref PubMed Scopus (72) Google Scholar (RNA-Seq), tandem reverse transcription polymerase chain reaction (RT-PCR), and complementary DNA (cDNA) Sanger sequencing, were performed as described in the supplementary methods. All individuals consented to participate in this study. Clinical grade DGT was performed on a consecutive series of 185,943 individuals with personal and/or family history of cancer. Of those, 2% carried a reportable MSH2 variant (PV or VUS), of which 0.7% were exonic duplications. Five individuals heterozygous for the VUS MSH2 EX2_3dup, MSH2 EX2_6dup, MSH2 EX5_6dup, MSH2 EX9_11dup, or MSH2 EX14dup sent additional samples to participate in this study. A summary of the clinical presentation, family history, tumor pathology, and DGT and RGT results of these patients is presented in Table 1.Table 1Identification of Germline Pathogenic In Tandem MSH2 Duplications With RGTVariantDNA detection methodFamily historyTumor (age at diagnosis, y)IHC MSIRGT results (predicted protein change)ACMG classification (before > after RGT)MSH2 EX2_3dupMLPA—Adenocarcinoma: sigmoid colon (10)MSH2-MSH6-MSI-Hin tandem (MSH2 p.I216Efs*13)VUS to pathogenicMSH2 EX2_6dupMLPAAmsterdam IAdenocarcinoma: ascending colon (56)MSH2-MSH6-in tandem (MSH2 p.L360Sfs*10)VUS to pathogenicMSH2 EX5_6dupMLPA—17 Tubular adenoma polyps (39)N/Ain tandem (MSH2 p.N361Qfs*8)VUS to pathogenicMSH2 EX9_11dupaCGHAmsterdam IAdenocarcinoma: right-sided colon (44)MSH2-MSH6-in tandem (MSH2 p.Y588Gfs*4)VUS to pathogenicMSH2 EX14dupaCGHAmsterdam IIAdenocarcinoma: ileocecal (62)MSH2-in tandem (MSH2 p.V821Lfs*7)VUS to pathogenicNOTE. Five exonic germline MSH2 duplications were first identified by DGT and classified as VUS after ACMG/AMP guidelines. Subsequent RGT showed that the duplications are in tandem, creating out-of-frame transcripts. This resulted in reclassification from VUS to clinically actionable PV.aCGH, array comparative genomic hybridization; IHC, immunohistochemistry; MLPA, multiplex ligation-dependent probe amplification; MSI, microsatellite instability; N/A, not applicable. Open table in a new tab NOTE. Five exonic germline MSH2 duplications were first identified by DGT and classified as VUS after ACMG/AMP guidelines. Subsequent RGT showed that the duplications are in tandem, creating out-of-frame transcripts. This resulted in reclassification from VUS to clinically actionable PV. aCGH, array comparative genomic hybridization; IHC, immunohistochemistry; MLPA, multiplex ligation-dependent probe amplification; MSI, microsatellite instability; N/A, not applicable. RNA-Seq showed that all tested duplications were in tandem and resulted in out-of-frame transcripts containing the duplicated exons (Table 1 and Supplementary Figure 1A–E). RNA-Seq results were validated by tandem RT-PCR, which requires the duplication to be in tandem to result in amplification (Supplementary Figure 2A and B), and by Sanger sequencing of tandem RT-PCR amplicons to confirm the abnormal exon–exon junctions in the cDNA (Supplementary Figure 2C). Following the ACMG/AMP guidelines6Richards S. et al.Genet Med. 2015; 17: 405-424Abstract Full Text Full Text PDF PubMed Scopus (5113) Google Scholar, we then used RNA evidence to reclassify these duplications from VUS to PV (Table 1). Overall, the number of MSH2 exonic duplications classified as PV was significantly higher in the group of individuals who received DGT followed by RGT than in the group who received only DGT (P = .0048). Currently, one of the major issues in the interpretation of DGT is exonic duplications, usually classified as VUS due to their unknown impact on gene expression. DGT followed by RGT was performed in the blood of patients heterozygous for different germline VUS MSH2 exonic duplications, showing that these SV result in abnormal transcripts and cause LS. RNA data were used as evidence toward the assessment of these germline SV based on the ACMG/AMP guidelines,6Richards S. et al.Genet Med. 2015; 17: 405-424Abstract Full Text Full Text PDF PubMed Scopus (5113) Google Scholar which resulted in the reclassification of all tested MSH2 VUS into clinically actionable PV. Exonic duplications account for approximately 33% of all SV in cancer predisposition genes,2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar indicating that this workflow could also be used to reduce VUS in other genes. DNA-based assays may be able to show whether a duplication is in tandem, but they cannot provide any direct functional evidence that the alteration is disrupting gene expression.2Mu W. et al.Genet Med. 2018; https://doi.org/10.1038/s41436-018-0397-6Abstract Full Text Full Text PDF Scopus (15) Google Scholar, 3Liccardo R. et al.Cancer Genet. 2018; 221: 19-24Abstract Full Text Full Text PDF PubMed Scopus (12) Google Scholar, 5Richardson M.E. et al.Genet Med. 2018; 21: 683-693Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar Furthermore, RNA-Seq overcame the shortcomings of low-throughput RNA methodologies (e.g., RT-PCR and cDNA Sanger sequencing), allowing large-scale acquisition of functional evidence on a myriad of other VUS (e.g., variants affecting RNA splicing) while maintaining the turnaround time necessary for clinical testing. These attributes support the benefit of implementing RNA-Seq analysis to the diagnostic workflow of clinical laboratories performing LS genetic testing. Upon identification of a pathogenic variant in a LS gene, the American Gastroenterological Association recommends routine screenings, including colonoscopy, every 1–2 years starting at age 20–25 years.8Rubenstein J.H. et al.Gastroenterology. 2015; 149: 777-782Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar Furthermore, the National Comprehensive Cancer Network (NCCN) recommends consideration of risk-reducing salpingo-oophorectomy and hysterectomy upon completion of childbearing in women.1NCCN. NCCN clinical practice guidelines in oncology V.3.2017: Genetic/Familial High-Risk Assessment: Colorectal. NCCN Clinical Practice Guidelines 2017.Google Scholar Therefore, clarifying the pathogenicity of germline exonic duplications identified in LS genes is of utmost importance for the proper management of these patients. Author contributions: Blair R. Conner and Rachid Karam are responsible for the study concept and design. Blair R. Conner and Tyler Landrith are responsible for the acquisition of data. Blair R. Conner, Rachid Karam, Felicia Hernandez, Beth Souders, Tyler Landrith, and C. Richard Boland performed analysis and interpretation of data and participated in the writing of the manuscript. RGT was offered to individuals from a consecutive series of 185,943 patients referred to Ambry Genetics (Aliso Viejo, CA) for DNA multi-gene panel testing (MGPT) between June 2012 and October 2018 who met all of the following criteria: 1) individual was diagnosed with an intragenic exonic MSH2 variant of unknown significance (VUS), and 2) individual was negative for additional germline variants in colorectal cancer predisposition genes (APC, BMPR1A, CDH1, CHEK2, EPCAM, GREM1, MLH1, MSH2, MSH6, MUTYH, PMS2, POLD1, POLE, PTEN, SMAD4, STK11, and TP53), and 3) individual was available to submit an additional blood sample for RNA extraction. Reclassification reports were issued when RGT evidence was sufficient to reclassify a VUS to either pathogenic or variant likely pathogenic. This study was approved by the Western Institutional Review Board. Data regarding each patient’s clinical and familial histories were extracted from the original test requisition form. If more details were required, the medical provider was contacted. All information has been de-identified. Tumor immunohistochemistry results were extracted from external pathology reports regarding nuclear staining of mismatch repair proteins (MSH2, MSH6, PMS2, and MLH1) and microsatellite instability status. Genomic DNA was isolated from whole blood or saliva via QiaSymphony (Qiagen, Redwood City, CA) and purified by using a column-based method (Zymo, Irvine, CA). Isolated DNA was quantified with a NanoDrop spectrophotometer or a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA). Next-generation sequencing multi-gene panel testing (MGPT) was performed on isolated DNA, as described previously.1Mu W. et al.Genet Med. 2018; Google Scholar Briefly, dual-indexed libraries generated with the Kapa Hyper-prep kit with enzymatic fragmentation (Kapa Biosystems, Wilmington, MA) were capture-enriched with xGen Lockdown probes (IDT, San Jose, CA) and sequenced on a NextSeq500 (Illumina, San Diego, CA) with 150-base pair paired-end reads. Ambry’s custom bioinformatics pipeline was used to identify germline variants, as described previously.1Mu W. et al.Genet Med. 2018; Google Scholar For array comparative genomic hybridization, 0.5 μg of reference DNA was labeled with Cy3-9mer and 0.25 μg of patient DNA was labeled with Cy5-9mer. Samples were hybridized to the custom array comparative genomic hybridization microarray for 16 hours. Boundaries of duplicated regions were identified by the extent of elevated probes in the patient sample when compared with the reference by NxClinical software (BioDiscovery, El Segundo, CA). For multiplex ligation-dependent probe amplification, 100–200 ng of reference and patient DNA was hybridized to SALSA-labeled probes spanning exonic regions in MSH2 by using a commercially available kit (MRC-Holland, Amsterdam, The Netherlands) and separated by size via capillary electrophoresis (ABI 3300, Thermo Fisher Scientific). Signal intensity of the patient sample was compared with the reference to determine exons involved in duplications by using Coffalyser (MRC-Holland). Patients’ blood was drawn into a tube containing RNA stabilizers (PAXgene, Qiagen). Total RNA was isolated from whole blood (PAXgene) and quantified by using the TapeStation 2200 HSRNA kit (Agilent Technologies, Santa Clara, CA). RNA integrity was determined with RNA integrity number equivalent (RINe), with values above 5 considered to be sufficient quality for downstream analysis. RNA-Seq with oligonucleotide probes to capture selected regions of interest for high-throughput targeted cDNA sequencing was performed with the Kapa RNA Hyper Prep kit with RiboErase (Kapa Biosystems). Ribosomal RNA was depleted enzymatically from 500 ng of total RNA. Ribo-depleted RNA was then fragmented and underwent first-strand synthesis followed by second-strand synthesis and A-tailing. A-tailed cDNA libraries were then ligated to standard Illumina dual index adapters and amplified using a 16-cycle PCR. Adapter-ligated cDNA libraries were quantified with a TapeStation 2200 D1000 kit (Agilent Technologies). Next, 125 ng each of 8 samples were pooled together and blocked with human COT-1 DNA (Invitrogen, Carlsbad, CA) and xGen universal blocking oligos (IDT) in preparation for hybrid capture, which was performed overnight at 65°C with custom xGen Lockdown probes (IDT). Nonspecific targets were removed, and final captured libraries were amplified with a 12-cycle PCR and then quantified with a TapeStation 2200 D1000 kit (Agilent Technologies). The 150-base pair paired-end sequencing was performed with the NextSeq500 High Output kit (Illumina). BAM files were generated as described previously2Mercer T.R. et al.Nat Protoc. 2014; 9: 989-1009Crossref PubMed Scopus (110) Google Scholar and were examined manually for the presence of abnormal splicing in Integrative Genomics Viewer (IGV) by viewing soft-clipped bases on the 3′ end of the duplicated region. Sanger sequencing of tandem RT-PCR products was performed as an orthogonal method to RNA-Seq. Total RNA was used to synthesize cDNA with a 500-ng input by using Oligo-dT to select for mRNA (SuperScript IV, Invitrogen). Primers were designed on MSH2 cDNA (NM_000251.1) so that the forward primer was downstream of the reverse primer within the duplicated region as follows: MSH2 EX2_3dup forward 5′-ATGTCAGCTTCCATTGGTGT-3′, reverse 5′-GCCAAATACCAATCATTCTCC-3′; MSH2 EX2_6dup forward 5′-GTCTCTGGCTGCCTTGCT-3′, reverse 5′-GCAACCTGATTCTCCATTTC-3′; MSH2 EX5_6dup forward 5′-GTCTCTGGCTGCCTTGCT-3′, reverse 5′-GAGAGCCAGTGGTATCTTCA-3′; MSH2 EX9_11dup forward 5′-TTGGACCCTGGCAAACA-3′, reverse 5′-TAATGTTGACTGCATCTTCTTTTC-3′; MSH2 EX14dup forward 5′-CCTACGATGGATTTGGGTTAG-3′, reverse 5′-AGTTCCTCTTCCCAATTCATCT-3′. The proband, blood control, commercially available colon control (Life Technologies, Carlsbad, CA), and a nontemplate control (data not shown) were assayed with each respective primer set under the same conditions (500 ng cDNA, 0.5 μmol/L primer, Qiagen HotstarTaq with 1.5 mmol/L MgCl2, 35 cycles, 45-second elongation, and Ta of 58°C). Amplicons were visualized and quantified (TapeStation 2200, Aglient Technologies). Sanger sequencing was performed on cDNA amplicons from tandem RT-PCR present only in the probands (20 ng tandem RT-PCR product, 2.5 μmol/L primer; Genewiz, La Jolla, CA). Electropherograms were visualized with a Sequence Scanner2 (Thermo Fisher Scientific), and FASTA sequences were entered into an open reading frame finder (NCBI, Bethesda, MD) to determine reading frame alterations. Classification of germline duplications was performed by following the ACMG/AMP guidelines.3Richards S. et al.Genet Med. 2015; 17: 405-424Abstract Full Text Full Text PDF PubMed Scopus (14616) Google Scholar Multiple lines of evidence were used to reach 1 of the following classifications: pathogenic variant, variant likely pathogenic, VUS, variant likely benign, or benign variant. The Fisher exact test was used to test for an association between the MSH2 duplication classification (VUS or variant likely pathogenic/pathogenic variant) in patients in which RGT was performed in addition to DNA MGPT compared with classification in patients with only DNA MGPT. Significance was set to P < .05.Supplementary Figure 2Tandem RT-PCR and cDNA Sanger sequencing confirm abnormal transcripts from RNA-Seq. (A) Schematic diagram of duplicated MSH2 transcripts, with the exons involved in the duplication colored white. Tandem RT-PCR primers (dark blue arrows) and the premature stop codon (red asterisk) are shown on the duplicated transcript. (B) TapeStation gel image of tandem RT-PCR amplicons present in the proband (+) only and none of the (–) blood or (–) colon controls. (C) Sanger sequencing of tandem RT-PCR products in each proband confirms abnormal splicing observed in RNA-Seq. Codons are marked with brackets, and the premature stop codons are marked with red asterisks. bp, base pair.View Large Image Figure ViewerDownload Hi-res image Download (PPT)