MLH1, MSH2, MSH6, & PMS2: Gene Function and Cancer Risk

MLH1, MSH2, MSH6, and PMS2 are the four genes at the heart of your body’s DNA mismatch repair system, a built-in proofreading mechanism that catches and corrects errors made during DNA copying. When any of these genes carries a harmful mutation, the repair system falters, and mistakes in DNA accumulate with each cell division. That accumulation is what drives an elevated risk of several cancers, most prominently colorectal and endometrial cancer. These four genes do not all carry the same weight, though. The specific gene involved shapes how high the risk climbs, which organs are most threatened, how early screening should begin, and even how well certain treatments work.

How the Repair System Works

Every time a cell divides, its entire genome has to be copied. The machinery that does this copying is remarkably accurate, but it still makes occasional mistakes, inserting the wrong base or slipping on a repetitive stretch of DNA. Mismatch repair is the cleanup crew that patrols freshly copied DNA, finds those errors, and fixes them before the cell divides again.

The system relies on two teams of proteins that work in sequence. The first team handles detection: MSH2 pairs with either MSH6 or MSH3 to form the complexes responsible for spotting errors. The MSH2-MSH6 pair is the workhorse, recognizing single-base mismatches and small insertion or deletion errors of one or two bases. The MSH2-MSH3 pair handles a complementary set, primarily targeting larger insertion and deletion loops of two to four or more extra bases.¹ The two complexes have overlapping territory on small errors but distinct specialties for different mismatch types, meaning both are needed for comprehensive surveillance.²

Once one of these detection complexes clamps onto a mismatch, it signals the second team. MLH1 pairs with PMS2 to form a complex that acts as the system’s executioner. This MLH1-PMS2 complex is a latent endonuclease, meaning it sits idle until the detection signal arrives, then activates and cuts the newly copied strand near the error. That nick marks the section of faulty DNA for removal and resynthesis.³ The cutting activity resides specifically in the PMS2 protein, but it requires MLH1 as a partner to function. This dependency explains why a mutation knocking out either MLH1 or PMS2 can disable the incision step entirely.

MSH2 is the common partner in both detection complexes, which is why MSH2 mutations tend to be particularly consequential: losing MSH2 knocks out both the MSH2-MSH6 and MSH2-MSH3 detection pathways at once. MLH1 similarly anchors the repair-execution step and also pairs with other partners for related functions. This central-hub role for MSH2 and MLH1 is why mutations in those two genes generally carry the highest cancer risks.

Colorectal Cancer Risk Varies Sharply by Gene

Inherited mutations in any of the four mismatch repair genes cause Lynch syndrome, the most common hereditary predisposition to colorectal cancer. But the risk is not uniform across genes. In a large prospective study tracking carriers to age 75, the cumulative risk of colorectal cancer was about 46% for MLH1 mutation carriers, 43% for MSH2 carriers, and 15% for MSH6 carriers.⁴ PMS2 carriers have the lowest colorectal cancer risk of the four, typically in the single digits to low teens by that age. This is a huge range, and it means that grouping all Lynch syndrome carriers together without specifying the gene involved can be misleading.

Even within a single gene, the penetrance varies widely. A large retrospective study from the International Mismatch Repair Consortium found that among MLH1 and MSH2 carriers, depending on sex and geographic region, anywhere from 7% to 56% of carriers had a colorectal cancer penetrance below 20%, while 9% to 44% had penetrance above 80%.⁵ That spread reflects the influence of family-specific factors: the particular variant involved, sex, environment, and probably other genes that modify overall risk.⁶ Two people carrying different MLH1 variants may have dramatically different lifetime odds of developing cancer.

Cancers Beyond the Colon

Colorectal cancer gets most of the attention, but Lynch syndrome raises the risk for a range of other cancers, and the pattern again depends on which gene is affected. In the same prospective study, cumulative risks by age 75 for endometrial cancer were roughly 43% for MLH1 carriers, 57% for MSH2, and 46% for MSH6. Ovarian cancer risk ranged from about 10% to 17% depending on the gene. MSH2 carriers faced the highest risks for urinary tract cancers (around 25%) and prostate cancer (about 32%), while MLH1 carriers had the highest cumulative risk for upper gastrointestinal cancers, including gastric, duodenal, bile duct, and pancreatic tumors (around 21%).⁴

A nationwide cohort study in Japan focusing on extracolonic tumors found a cumulative endometrial cancer risk of 66% by age 70, gastric cancer risk of 23%, and ovarian cancer risk of 9%. Gastric and urothelial cancers in that study were primarily linked to MLH1 and MSH2 mutations.⁷ The differences between studies partly reflect population-specific cancer patterns, but the consistent thread is that MSH2 mutations tend to cast the widest net across organ systems, while MSH6 and PMS2 mutations tend to produce narrower, lower-magnitude risks.

Inherited Mutations Versus Epigenetic Silencing

Not every mismatch repair failure in a tumor traces back to an inherited gene mutation. MLH1 in particular is frequently silenced by a chemical modification called promoter methylation, in which methyl groups are added to the gene’s control region, effectively switching it off without altering the DNA sequence itself. This epigenetic silencing is the most common cause of mismatch repair deficiency in sporadic (non-inherited) colorectal cancers, especially in older patients.

In one study of colorectal tumors, MLH1 promoter methylation was found in more than half of tumors with high microsatellite instability, and extensive methylation covering both regions of the promoter was present in all such tumors that had lost MLH1 protein expression.⁸ Distinguishing this sporadic methylation from an inherited MLH1 mutation is clinically important. If MLH1 loss in a tumor is caused by methylation rather than a germline mutation, the patient does not have Lynch syndrome and does not face the same elevated risks for future cancers in other organs, nor do their relatives. Testing for MLH1 methylation and for a common associated mutation in the BRAF gene helps sort this out and avoids unnecessary alarm and genetic testing in families.

How Mismatch Repair Status Is Tested

Clinicians use two main approaches to determine whether a tumor’s mismatch repair system is intact. Immunohistochemistry, or IHC, stains tumor tissue with antibodies against each of the four proteins. If a protein is absent, the corresponding gene is likely nonfunctional. The staining patterns also give clues about which gene is responsible: because the proteins work in pairs, loss of MSH6 typically means MSH6 is the culprit, while loss of both MLH1 and PMS2 suggests that MLH1 is defective, since PMS2 is unstable without its partner. However, the interpretation is not always clean. Some tumors show patchy or mixed staining rather than uniform loss, and this heterogeneous pattern has been shown to still correlate with high microsatellite instability in the regions where staining is absent.⁹

The second approach tests for microsatellite instability directly. Microsatellites are short, repetitive sequences scattered throughout the genome that are especially prone to errors during DNA copying. When mismatch repair is defective, these repeats accumulate uncorrected insertions and deletions, producing a measurable pattern called microsatellite instability-high, or MSI-H. Traditional testing uses PCR to examine a standard panel of microsatellite markers, but next-generation sequencing methods have been developed that can assess microsatellite instability from the same sequencing data already being generated for tumor profiling. These NGS approaches have shown high sensitivity and specificity compared to conventional PCR, and in some cases may detect instability that PCR-based methods miss.¹⁰¹¹

An important caveat: mismatch repair gene mutations and microsatellite instability do not always travel together. A large analysis of over 19,000 tumors across 11 cancer types found that about half of tumors carrying mismatch repair gene mutations did not show microsatellite instability, and the degree of concordance between the two varied dramatically by tumor type. Colorectal and endometrial cancers had reasonably high agreement, while pancreatic and lung cancers showed almost none.¹² This means that relying on a single test in isolation can miss the full picture, especially outside the tumor types where MSI testing has been most validated.

Why Mismatch Repair Status Matters for Treatment

The clinical relevance of mismatch repair deficiency extends well beyond diagnosis and risk estimation. When mismatch repair fails, the resulting flood of uncorrected DNA errors produces a high tumor mutational burden, meaning the cancer cells accumulate many abnormal proteins. These abnormal proteins act as flags that the immune system can recognize. Tumors with this profile tend to be heavily infiltrated by immune cells and respond exceptionally well to immune checkpoint inhibitors, drugs that release the brakes on the immune system’s anti-tumor response.¹³ The response rates in MSI-H colorectal cancer have been dramatic enough that checkpoint inhibitors targeting the PD-1 and CTLA-4 pathways have become standard treatment options, with some patients experiencing durable, long-lasting remissions.¹⁴

The relationship with traditional chemotherapy cuts the opposite direction. A commonly used drug called 5-fluorouracil (5-FU) partly works by creating errors in DNA that the mismatch repair system detects as lethal damage. When mismatch repair is defective, cells no longer recognize 5-FU-induced damage as fatal, and they survive treatment instead of dying. Laboratory work has confirmed a strong link between mismatch repair deficiency and 5-FU resistance across large panels of colorectal cancer cell lines.¹⁵ Some of this resistance appears to be mediated by specific molecular pathways, including downregulation of a small RNA molecule called miR-552 that normally promotes cell death in response to 5-FU.¹⁶ The practical consequence is that patients with MSI-H tumors may not benefit from 5-FU-based chemotherapy the way patients with intact mismatch repair do, a distinction that influences treatment decisions.¹⁷

Gene-Specific Surveillance Strategies

Because cancer risks differ so sharply among the four genes, there is a growing push to tailor surveillance schedules accordingly rather than applying one set of guidelines to all Lynch syndrome carriers. A cost-effectiveness analysis found that for MLH1 and MSH2 carriers, the optimal approach was colonoscopy beginning at age 25 with exams every one to two years. For MSH6 carriers, starting at age 35 with three-year intervals was cost-effective, and for PMS2 carriers, starting at age 40 with three-year intervals was sufficient.¹⁸ Real-world data support this stratification: the incidence of precancerous and advanced lesions found during colonoscopy surveillance differs meaningfully between gene groups, reinforcing the idea that a one-size-fits-all interval may be too aggressive for low-risk carriers and not aggressive enough for high-risk ones.¹⁹

These recommendations are evolving, and many current clinical guidelines still treat Lynch syndrome as a single entity. If you carry a known mismatch repair gene mutation, it is worth discussing with your care team whether your specific gene and variant warrant a modified schedule rather than the default.

Variants of Unknown Significance

Genetic testing for Lynch syndrome does not always produce a clean answer. Roughly 30% of variants identified in MLH1 and MSH2 are classified as variants of unknown significance, meaning the lab has detected a DNA change but cannot confidently say whether it is harmful or benign.²⁰ This is one of the most frustrating situations in clinical genetics: you know something is different, but not whether it matters.

Researchers are working to resolve these uncertainties through functional assays that test each variant’s effect on mismatch repair activity in the laboratory, and through large-scale saturation mutagenesis efforts that systematically create and score every possible single-amino-acid change in a protein. One such effort generated a near-complete functional map of MSH2 missense variants, covering more than 94% of possible substitutions. The encouraging finding was that a large majority of missense variants scored as tolerated, suggesting they do not impair the protein’s function.²¹ Maps like these should steadily reduce the number of variants left in the uncertain category, giving patients and clinicians clearer answers.

Constitutional Mismatch Repair Deficiency

Lynch syndrome involves inheriting one defective copy of a mismatch repair gene while retaining one working copy. In the extremely rare condition known as constitutional mismatch repair deficiency, or CMMRD, a child inherits defective copies from both parents, leaving no functional version of the gene at all. The consequences are severe: children with CMMRD develop cancers in early childhood across multiple organ systems, including brain tumors, blood cancers, and gastrointestinal malignancies.²² Any of the four mismatch repair genes can cause CMMRD when both copies are knocked out, though PMS2 is disproportionately represented in reported cases, likely because carriers of a single PMS2 mutation have relatively mild cancer risks and may not be identified before having children who inherit two copies.²³

CMMRD is rare enough that most oncologists will never encounter it, but it underscores a broader point: the four mismatch repair genes exist on a spectrum of clinical severity, from the relatively manageable surveillance needed for PMS2-associated Lynch syndrome all the way to the life-threatening childhood cancers of CMMRD. That spectrum is why precise identification of the gene and variant matters so much.

Somatic Mutations and Lynch-Like Syndrome

Not all mismatch repair deficiency in tumors is inherited. Some tumors acquire mutations in mismatch repair genes during their development, a phenomenon called somatic mutation, which is not present in the person’s normal cells and cannot be passed on to children. When a tumor shows mismatch repair deficiency but no germline mutation or MLH1 methylation can be found, it falls into a category sometimes called Lynch-like syndrome.

A cross-tumor analysis of more than 19,000 tumors found that MSH6 was the most frequently mutated mismatch repair gene at the somatic level, accounting for about 29% of all mismatch repair variants identified. MSH2 followed at roughly 21%, MLH1 at about 20%, and PMS2 at about 17%. The frequency of mismatch repair mutations and their clinical significance varied considerably by tumor type. Endometrial cancer had the highest rate of mismatch repair mutations (around 20%), followed by colorectal cancer (about 8%) and bladder cancer (roughly 9%).¹² A smaller study investigating the molecular details of Lynch-like tumors found that most carried one to three somatic mismatch repair gene variants that could explain the repair deficiency, and in some cases, other non-mismatch-repair genes were responsible.²⁴

The treatment implications of somatic mismatch repair deficiency are similar to those of inherited deficiency: these tumors still tend to respond well to immunotherapy and poorly to 5-FU. But the family implications are completely different. A somatic mutation means the rest of the family is not at elevated risk, and no cascade genetic testing of relatives is needed.

Liquid Biopsy and Monitoring

Emerging research is exploring whether mismatch repair status can be assessed from blood draws rather than tumor biopsies. Circulating tumor DNA, fragments of tumor genetic material shed into the bloodstream, can be tested for microsatellite instability using next-generation sequencing. This approach has shown clinical utility in monitoring treatment response in patients with mismatch repair-deficient colorectal cancer receiving immunotherapy.²⁵ Case reports have also demonstrated that liquid biopsy can identify microsatellite instability in cancers where tissue biopsy is difficult, such as pancreatic cancer, and can track responses to immunotherapy over time through serial blood tests.²⁶

Mismatch Repair Beyond Cancer

The mismatch repair system is not exclusively a tumor suppressor. Research has uncovered a counterintuitive role for these same genes in driving certain mutations rather than preventing them. In Huntington’s disease and several other neurodegenerative conditions caused by expanded stretches of repeated DNA sequences, mismatch repair proteins appear to actively promote the expansion of those repeats in non-dividing cells like neurons. The very system that protects against cancer in rapidly dividing tissues seems to worsen repeat-expansion diseases in the brain.²⁷ MSH3, the occasional partner of MSH2, has emerged as a particularly strong genetic modifier of Huntington’s disease progression, and clinical trials targeting mismatch repair-related pathways in neurodegenerative disease are in early stages. This dual identity of the mismatch repair system, protective in one context and damaging in another, is one of the more surprising findings in recent genetics research and a reminder that biological systems rarely have a single, simple function.