Colorectal cancer (CRC) is the second most frequent cancer among men and women in Brazil, corresponding approximately to more than 45 thousand new cases per year (1). Approximately 5% of CRCs are associated with germline variants, and Lynch syndrome (LS) is the most prevalent cause of hereditary CRC and is an autosomal dominant disorder related to monoallelic germline pathogenic variants (GPVs) in DNA mismatch repair (MMR) genes MLH1, MSH2, MSH6, and PMS2, and deletions in the EPCAM gene (2–4). It is clinically characterized by predisposition to a broad spectrum of tumors, including early-onset CRC and extracolonic tumors, including endometrial, ovarian, gastric, ureter, renal pelvis, pancreatic, prostate, biliary tract, central nervous system, and small bowel (5, 6).

While GPV in MLH1 and MSH2 genes account for almost 70% of LS cases, mutations in PMS2 contribute to <15% (7, 8). Molecular testing of PMS2 is challenging due to high homology of PMS2 gene to its counterpart PMS2CL pseudogene, which is considered biologically inactive. Both are located on chromosome 7, and interpreting the clinical relevance of variants detected in these regions is essential for patients’ follow-up and is considered an important challenge nowadays (9).

The PMS2CL pseudogene presents high homology (>98%) with PMS2, with the greatest identity being found over exon 9 and between exons 11 and 15 (10, 11). NGS is able to identify variants along all the coding segments of this gene; however, mapping and variant calling pipelines struggle to differentiate whether a variant is present in the gene or in the pseudogene. Because of the high methodological complexity required to distinguish variants between PMS2 and PMS2CL, most laboratories do not clearly report the origin of this molecular finding.

Thus, due to the extreme importance of correctly reporting reliable variants in the PMS2 gene, the aim of this study was to confirm the GPVs detected by NGS in regions of high homology segments of the PMS2 gene in a Brazilian sample using the orthogonal and gold standard long-range PCR (LR-PCR) methodology to separate variants detected in the PMS2 gene from those detected in the pseudogene. This strategy will prompt reliable results that will directly contribute to appropriate clinical management.

A total of 74 samples with a PMS2 GPV detected by NGS in exons with high homology with the PMS2CL pseudogene were evaluated. Four different GPVs were identified in exons 11 and 13 ( Figure 1 ). Of these, the most common variant detected in 68 samples was NM_000535.6: c.2182_2184delinsG (p.Thr728Alafs*7) (ClinVar ID: 231999, VCV000231999.10), located in exon 13 of the PMS2 gene, according to the NGS mapping pipeline. This variant causes a translational frameshift with a predicted stop codon and has been reported in the literature to be associated with LS (16). All detected GPVs are described in Table 3 .

It was not possible to correlate the molecular findings with the tumor MMR status (immunohistochemical) or tissue microsatellite instability analysis.

We did not have access to the correlated data or clinical information of the patients evaluated in this study. All the 68 patients harboring variant c.2182_2184delinsG, detected through NGS, have failed to confirm it by LR-PCR, indicating the absence of this variant in PMS2. The mean variant allele frequency (VAF) of these variants was 24.7% (ranging from 10.7 to 39.7%), and the median was 22.2% ( Table 3 ).

Of the other six GPV detected, four are located in exon 11 and two in exon 13 ( Table 3 ). Five samples evaluated by LR-PCR were confirmed and considered positive result. Three patients had the following confirmed variant NM_000535.7:c.1239dup (ClinVar ID: 216072, VCV000216072.32), located in exon 11 of PMS2. The mean VAF of these variants was 51.18% (ranging from 48.1 to 53.91%), and the median was 51.55%. The variant NM_000535.6:c.1687C>T (ClinVar ID: 135067, VCV000135067.29) was detected in one patient with VAF of 46.9% and was confirmed by LR-PCR.

The variant NM_000535.7:c.2192_2196del (ClinVar ID: 91331, VCV000091331.38) was detected in two patients and confirmed in one case. The mean VAF was 26%.

In summary, of the 74 GPV identified, five (6.8%) were confirmed by LR-PCR. Conversely, the other 69 patients (93.2%) who did not confirm the presence of the variant after LR-PCR had the diagnosis of LS ruled out by molecular mechanisms associated with the PMS2 gene, not excluding the possibility of other clinical criteria involved with this diagnosis.

Considering the five variants confirmed by LR-PCR, we found a mean VAF of 46.3% versus 24.6% of the other 69 unconfirmed variants.

NGS has some limitations, and the analysis of genes with high identity to pseudogenes is one of them. This is an important issue because the presence of the pseudogene can result in false positive or negative tests, thereby affecting clinical practice and genetic counseling. Thus, the use of different approaches is necessary to avoid interference with data interpretation, which could lead to misleading conduct.

The analysis of PMS2 variants by NGS is very complex. Even considering that the NGS approach can detect variants along all the coding segments of this gene, when a pathogenic variant is detected, the correct clinical interpretation is very challenging because of the high homology of PMS2 to its counterparts, such as the non-expressed PMS2CL pseudogene (17, 18).

The evaluation of PMS2 gene is neglected by many laboratories, due to the methodological difficulties in reporting reliable variants and not ensuring that the variant has been detected in the gene. Many commercial laboratories do not analyze the regions of these PMS2 pseudogenes, consequently generating incomplete analysis of this gene. However, some laboratories report GPV in these regions of high homology, but without confirmation using other techniques.

Although the PMS2 gene has low penetrance, a reliable identification of the presence of pathogenic variants in this gene is fundamental for the correct management of LS and genetic counseling.

In our experience, more than 90% of the pathogenic variants identified by NGS in the PMS2 gene in exons with high identity to pseudogenes were not confirmed using LR-PCR. We observed that variants confirmed by LR-PCR presented a higher VAF, near 50%. In contrast, unconfirmed variants had lower VAF, indicating that variants in the actual gene have higher VAF.

It is important to note that c.2182_2184delinsG variant identified in our cohort was not confirmed to be in the PMS2 gene in any of the patients. This variant is reported in ClinVar database as conflicting, was previously described as a deleterious mutation in a study of African-American patients with LS, and has been widely reported by laboratories as a pathogenic variant associated with the LS phenotype (16). According to Chong et al. (2020), this variant was incorrectly assigned to PMS2 in a sample of patients, suggesting reclassification and caution when interpreting these variants (19). Although we studied variants only in exons 11 and 13, this methodology was developed to confirm variants in other regions with homology to PMS2 gene.

Our findings strongly support this suggestion, and we recommend a pathogenic variant classification only if the variant is at the PMS2 gene evaluated by a LR-PCR. Thus, there are benefits for patients because the diagnosis of LS is excluded, avoiding unnecessary screening and even unequivocal indication of hysterectomy and risk-reducing salpingo-oophorectomy (20).

In conclusion, the use of LR-PCR was demonstrated to be a reliable approach for accurate molecular analysis of PMS2 gene variants in segments with high homology with the PMS2CL pseudogene. We highlight that our laboratory is a pioneer in the diagnostic complementation of the PMS2 gene in Brazil, directly contributing to more assertive molecular diagnosis. Our results indicate that using confirmation strategies such as LR-PCR in those segments is essential to avoid misdiagnosis of LS, directly impacting the genetic counseling of these patients and their families, since the correct molecular diagnosis can avoid inappropriate clinical management.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

The studies involving humans were approved by Human Research Ethics Committee (protocol number NP_614) - Fleury S.A. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

DP: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. TL: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. RS: Methodology, Validation, Writing – review & editing. JC: Writing – review & editing. CP-A: Writing – review & editing. IS-F: Writing – review & editing. PS: Writing – review & editing. MM-N: Methodology, Validation, Writing – review & editing. CM: Project administration, Supervision, Writing – review & editing. WB: Project administration, Supervision, Writing – review & editing.