Lynch syndrome (LS), also known as hereditary nonpolyposis colorectal carcinoma (HNPCC), is the most common hereditary colorectal cancer (CRC) syndrome, accounting for approximately 2%–4% of all CRCs. LS is an autosomal dominant disorder caused by germline mutations in mismatch repair (MMR) genes (MLH1, MSH2, MSH6, PMS2) or the epithelial cell adhesion molecule (EpCAM) gene (Vasen et al., 2010). These mutations lead to MMR deficiency (dMMR) and high microsatellite instability (MSI-H), resulting in genomic instability and increased tumor susceptibility. In addition to CRC, LS is associated with increased risks of endometrial, gastric, ovarian, and brain cancers. The spectrum and magnitude of cancer risk among LS families vary depending on which specific MMR or EPCAM gene is mutated.

Population-based prospective data further highlight gene-specific penetrance. According to the Prospective Lynch Syndrome Database (PLSD), the cumulative cancer risks by age 70 are estimated at 71.9% for MLH1, 74.5% for MSH2, 46.3% for MSH6, and 21.7% for PMS2 mutation carriers (Vasen et al., 2007; Prospective Lynch Syndrome Database, 2025). CRC risk is particularly high in MLH1 and MSH2 carriers (48.2% and 43.7%, respectively), whereas endometrial cancer (EC) risk is most elevated in female MSH2 and MSH6 mutation carriers (46% and 41%, respectively). Although brain tumors are relatively uncommon in LS, their relative risk is highest in MSH2 carriers (3.7%).

EpCAM deletions are rare, accounting for only 1%–3% of LS cases (Tiwari et al., 2016). A well-established mechanism is that 3′EpCAM deletions can lead to MSH2 promoter hypermethylation, thereby silencing MSH2 expression and triggering dMMR (Niessen et al., 2009). In a cohort of 41 LS families, EpCAM deletion carriers had a cumulative CRC risk of 75% by age 70, which was higher than that of MSH6 carriers (50%) and similar to that of MSH2 (77%) and MLH1 mutation carriers (79%) (Kempers et al., 2011). Notably, female carriers with EpCAM-MSH2 co-deletions showed a markedly increased EC risk (55%) compared with EpCAM-only deletions (12%) (Kempers et al., 2011). These findings suggest that EpCAM-related alterations not only confer substantial CRC risk but also modify extracolonic cancer susceptibility depending on their interaction with MSH2.

At the molecular level, MSH2 with MSH6 forms MutSα complex that recognizes DNA mismatches, while PMS2 with MLH1 forms MutLα complex that executes the excision and repair step of MMR (Li, 2008-01). This division of labor highlights their complementary roles in genomic maintenance. APC, a gatekeeper of the Wnt/β-catenin pathway, initiates colorectal tumorigenesis (Kinzler and Vogelstein, 1996-10). In this context, MMR deficiency may accelerate APC-related mutational events, while APC loss in a dMMR background can further lower the threshold for malignant transformation. Thus, concurrent EpCAM/MSH2, PMS2, and APC alterations provide a biological rationale for synergistic tumorigenesis.

Despite these insights, most prior studies focus on single-gene mutations or EpCAM–MSH2 co-deletions, often limited to case reports or small series. Important knowledge gaps remain: (i) how concurrent alterations spanning multiple loci (e.g., EpCAM, MSH2, PMS2, and APC) jointly shape penetrance, tumor spectrum, and age at onset; (ii) whether point mutations and structural variants interact to produce tissue-specific silencing and phenotype modification; and (iii) to what extent current surveillance guidelines, which are organized on a gene-by-gene basis, adequately capture multi-locus constellations.

To address these gaps, we report a LS family harboring a previously undocumented co-mutation subtype: EpCAM c.344T>C, MSH2 c.2744A>G, PMS2 c. 1408 C>T and APC c.5465T>A. Notably, we identified novel large fragment deletions spanning EpCAM exons 8–9 and MSH2 exons 1–16 in the proband’s CRC tissues. This case underscores the complexity of LS genetics and highlights the need for comprehensive evaluation incorporating both point mutations and structural variations, to better inform clinical management and surveillance strategies.

LS is an autosomal dominant syndrome with high cancer susceptibility, as 54%–61% of patients develop a second primary tumor and 15%–23% develop three or more (Stoffel et al., 2009). Diagnosis is based on family history (Amsterdam II Criteria, the Revised Bethesda Guidelines, and the Chinese Criteria) and confirmed by genetic testing of MMR, EpCAM, and BRAF V600E. Among LS families, MSH2 mutations account for ∼50%, MLH1 for 30%–40%, MSH6 for 7%–10%, and PMS2 for <5% (Silva et al., 2009).

Co-mutations involving EpCAM and MSH2 are rare. A cohort study of 41 LS families, including 194 EpCAM deletion carriers, identified 42 patients with both EpCAM and MSH2 co-deletions (Kempers et al., 2011). Additionally, only five case reports document LS patients with EpCAM and MSH2 co-mutations (Table 1).

Case 1 describes a 44-year-old woman with three episodes of CRC, whose family history includes her mother (CRC and EC) and sister (CRC and ovarian cancer) (Sekine et al., 2017).

To our knowledge, co-mutations involving EpCAM, MSH2, PMS2 and APC have not been reported. We present an LS family in which all six tested members carried EpCAM c.344T>C, MSH2 c.2744A>G, PMS2 c.1408C>T, and APC c.5465T>A co-mutations, with the proband’s CRC tissue showing large deletions of EpCAM (exons 8–9) and MSH2 (exons 1–16). Tumor testing corroborated the sequencing results: IHC of the proband’s tumor showed loss of EpCAM, MSH2, and MSH6 with retention of MLH1 and PMS2, consistent with MSH2-pathway deficiency; MSI testing demonstrated 30.43% unstable loci (MSI-H). These findings confirm dMMR and link EpCAM-associated MSH2 disruption to the observed immunophenotype (Figures 1B,E). Tumor testing was limited to the proband, restricting assessment across relatives.

In this family, eight members developed cancers at ages 25–32, including three CRCs before 40 (youngest 25), two brain cancers, one EC, one breast cancer, and one pancreatic cancer. Among the six carriers tested, two (33%) had developed cancer while four (67%) remained unaffected, illustrating incomplete penetrance. This rate is lower than PLSD estimates (∼45% for MLH1/MSH2 carriers by age 70), likely reflecting the relatively young age of unaffected carriers in this family (Prospective Lynch Syndrome Database, 2025). Compared with published cohorts, this family shows unusually early CRC onset and excess brain cancer. CRC onset was earlier than reported by the German Consortium (42–69 years for MSH6, 61–66 years for PMS2, ∼44 years for MLH1/MSH2) and PLSD (49–50 years for MSH2) (Gupta et al., 2019). The five reported EpCAM/MSH2 co-mutation cases also showed early CRC (9–54 years), and Danish data reported brain cancers in 14% of LS families (mainly MSH2) (Therkildsen et al., 2015), compared to 3.7% in PLSD male MSH2 carriers. In contrast, brain tumors occurred in 25% of affected members here. Compared to recent findings (Niessen et al., 2009; Markowitz and Bertagnolli, 2009), the co-mutation of EpCAM, MSH2, PMS2 and APC might be a pathogenic factor, though statistical validation is lacking. While most tumors were single primaries, metachronous cancers cannot be excluded. The strong family history facilitated early CRC detection by colonoscopy in the proband and his uncle, leading to favorable outcomes with standard management.

Genetic testing revealed multiple germline mutations together with large deletions in EpCAM and MSH2, suggesting a hereditary multi-tumor etiology. Review of public databases (NCBI, MGeND, LOVD, OncoKB; accessed July 2025) classified EpCAM c.344T>C and APC c.5465T>A as benign, PMS2 c.1408 C>T as benign or neutral, and MSH2 c.2744A>G as VUS. Based on ACMG/AMP criteria (Richards et al., 2015), EpCAM c.344T>C and APC c.5465T>A fulfill BA1/BS1 owing to high allele frequency in gnomAD (Karczewski et al., 2020), and BP4 given consistent benign predictions. PMS2 c.1408C>T shows ∼35–40% frequency in gnomAD and 1,000 Genomes, meeting BA1 and BP4. In contrast, MSH2 c.2744A>G shows very low population frequency (<0.01) and inconsistent in silico results, thus remains VUS. Overall, these mutations appear unlikely to drive cancer (Table 2), though discordant predictions (e.g., MSH2 c.2744A>G was classified as pathogenic by SIFT) underscore the limitations of in silico tools and the need for functional validation.

Despite the overall benign or uncertain classification, several reports suggest possible biological effects of these mutations (Supplementary Table S1). EpCAM c.344T>C may weaken protein folding, thereby promoting cervical cancer progression (Hu et al., 2012; Sankpal et al., 2021), enhance tumor cell stemness to facilitate breast cancer growth (Jiang et al., 2011), and impair EpCAM’s inhibitory effect on cathepsin L, increasing the invasiveness of non-small cell lung cancer (Yang et al., 2014). APC c.5465T>A, though not disrupting the Wnt/β-catenin signaling pathway, may still promote colorectal adenoma formation. Wallis’s report indicates that this mutation could disrupt β-catenin regulation, leading to intestinal cancer or adenoma formation (Wallis et al., 1999). Moreover, Theodoratou et al. noted that this mutation might alter the response to environmental factors through Wnt pathway modulation, thus increasing CRC risk (Theodoratou et al., 2008). There is currently no evidence linking PMS2 c.1408 C>T to tumor development and progression (Yurgelun et al., 2017), and the significance of MSH2 c.2744A>G remains unclear. Overall, while ACMG/AMP classification favors benign or VUS categories, subtle functional effects and gene-environment interactions cannot be excluded, highlighting the need for further validation.

In this family, sequencing of six members revealed EpCAM, MSH2, PMS2 and APC co-mutations associated with early-onset CRC. These co-mutations may modify CRC risk under specific exposures. We employed multiple prediction systems to assess the clinical significance of each mutation. Only Polyphen-2 and MutationTaster classified these mutations as benign, while most other systems classified them as potentially benign. The CADD prediction system identified APC c.5465T>A as likely pathogenic, while SIFT prediction system classified MSH2 c.2744A>G as pathogenic (Table 2). Therefore, current prediction tools cannot fully capture the biological impact of co-mutations, underscoring the need for functional validation.

The unusual phenotype may reflect epistasis, as co-existing EpCAM, MSH2, PMS2, and APC mutations could interact across MMR and Wnt/β-catenin pathways. 3′ EpCAM deletions can silence MSH2, establishing an MMR-deficient background, and animal models show that MSH2 deficiency synergizes with APC loss to accelerate intestinal tumorigenesis (Reitmair et al., 1996).

V-3, a 6-year-old child, warrants attention as WES revealed EpCAM, MSH2, PMS2 and APC co-mutations along with large deletions in EpCAM (exons 8–9) and MSH2 (exons 1–16), paralleling his father’s CRC tissue and suggesting possible germline mosaicism. Mosaicism is increasingly recognized in cancer (Ogawa et al., 2022) but may be missed by blood-based genetic testing, complicating LS diagnosis and genetic counseling. Case reports have documented mosaic MSH2 mutations in Lynch (-like) syndromes—for example, Erell Guillerm et al. detected a low-frequency MSH2 mutation in tumor tissue but absent in blood, which was confirmed as germline mosaicism in her daughter (Guillerm et al., 2020), while Pastrello et al. described somatic mosaicism with uneven mutation distribution across tissues (Pastrello et al., 2009). Reviews emphasize that mosaic events are underdetected in LS with blood-only testing (Jansen and Goel, 2020), supporting multi-tissue or tumor-first approaches. Additional proband tissue could clarify this but was unavailable. Alternatively, recurrent EpCAM 3′ deletions with shared haplotypes in Dutch LS families (Kuiper et al., 2011-04) suggest ancestral inheritance as another plausible explanation for the clustering of EpCAM, MSH2, PMS2, and APC mutations in this pedigree.

LS family members carrying mutations require ongoing surveillance. According to current guidelines (Vasen et al., 2010), MMR mutation carriers should undergo colonoscopy every 1–2 years starting at age 20–25, or 2–5 years before the earliest CRC in the family. Based on the 25-year earliest onset, surveillance should begin by age 20, and potentially earlier for young carriers. Given the 25% incidence of brain cancer, annual neurological evaluations are warranted, with brain MRI performed when clinically indicated. Female carriers should also undergo endometrial and ovarian cancer surveillance, such as annual transvaginal ultrasound or endometrial biopsy from age 30–35. At the same time, cascade testing of at-risk relatives is essential to identify asymptomatic carriers and enroll them in surveillance programs at an appropriate age, thereby reducing cancer risk through early detection and intervention. This strategy facilitates personalized risk communication and supports informed decision-making for long-term management within families.

Constitutional mismatch repair deficiency (CMMRD) is a rare hereditary cancer syndrome caused by biallelic pathogenic MMR mutations, leading to childhood tumors in the brain, gastrointestinal tract, and blood. PMS2 mutations are the most predominant causative genes (Mar et al., 2025). Genetic testing in V-3 identified a heterozygous MSH2 mutation without biallelic mutations or CMMRD symptoms (e.g., early cancers, café-au-lait spots) (Wimmer and Kratz, 2010; Wimmer et al., 2014). V-3’s heterozygous status elevates LS-related cancer risk, distinct from CMMRD’s biallelic-driven childhood cancer predisposition (Wimmer et al., 2014). Wimmer et al. noted that CMMRD leads to early malignancies, whereas LS typically manifests later (Wimmer et al., 2014). Therefore, V-3 requires LS-specific surveillance, not CMMRD’s pediatric protocols. For offspring inheriting germline mosaic mutations, LS-related cancer screening is advisable before adulthood, as the youngest CRC patient identified was only 9 years old in published case reports (Li-Chang et al., 2013). Further testing of more samples in this family is essential, as the results hold significant importance for management and disease prediction.

This case report describes a LS family harboring a rare co-mutation of EpCAM c.344T>C, MSH2 c.2744A>G, PMS2 c. 1408C>T and APC c.5465T>A. Notably, we identified pathogenic large fragment deletions in EpCAM (exons 8–9) and MSH2 (exons 1–16) in the proband’s CRC tissue for the first time. These findings expand the LS genetic spectrum and highlight the complexity of interpretating co-mutations. Despite databases suggest these point mutations are benign or VUS, bioinformatic predictions and prior reports suggest possible functional relevance. Given these uncertainties, conclusion should be regarded as preliminary. Future research should include development of cellular or animal models to dissect the synergistic effects of these alterations, systematic exploration of mosaicism and founder effects, and integration of functional assays with genomic profiling. Clinically, surveillance and cascade testing remain essential, and therapeutic implications, including immunotherapy responsiveness in dMMR, warrant further investigation.