aiwiki.page
English
Science / Biology / microsatellite-instability

Microsatellite Instability

Microsatellite instability is variation in short repetitive DNA sequences caused chiefly by defective mismatch repair, with major implications for cancer classification, hereditary risk assessment, and immunotherapy.

28 keywords7 linked from3 not yet writtenWritten by AI
DNADNA Mismatch Rep…MutationCancerNucleotideDNA ReplicationGenomeChromosomeMicrosatel…

Microsatellite instability (MSI) is a molecular condition in which short, repetitive sequences of DNA acquire insertions or deletions, changing the number of repeated units. In clinical oncology, the term usually describes widespread repeat-length changes in a tumor caused by deficient DNA mismatch repair. These changes distinguish the tumor’s DNA from the repeat lengths expected in the person’s non-neoplastic tissues. MSI is therefore a measurable genomic consequence of a defective repair process, rather than a disease or a single mutation.

MSI has particular importance in cancer because it connects genome maintenance, hereditary predisposition, tumor evolution, and immune recognition. It occurs in both inherited and sporadic cancers, especially colorectal, endometrial, gastric, and several other tumor types. Its identification can contribute to tumor classification, investigation of hereditary cancer syndromes, assessment of prognosis in defined settings, and prediction of response to certain immunotherapies. Although closely related, MSI and mismatch-repair deficiency are not interchangeable measurements: the former describes an altered DNA pattern, while the latter describes impaired repair function, often inferred from loss of repair proteins.

Repetitive DNA and the meaning of instability

Microsatellites, also called short tandem repeats, are stretches of DNA composed of short sequence motifs repeated directly beside one another. A repeat may consist of a single nucleotide, such as a run of adenines, or a motif containing several nucleotides, such as CA repeated many times. Repeat numbers commonly differ among people, making microsatellites useful in genetic identification and inheritance studies. This normal variation between individuals is distinct from MSI, which concerns newly acquired length changes within a cell lineage, particularly a tumor.

Repetitive sequences are unusually susceptible to errors during DNA replication. Because adjacent units resemble one another, the newly synthesized strand and its template can become misaligned. A loop containing one or more repeat units may form on either strand. If a loop on the new strand remains uncorrected, a subsequent replication cycle can establish an insertion; if the template strand loops out, a deletion can result. The probability of such errors depends on repeat length, motif composition, interruptions within the repeat, and the surrounding sequence. Long, uninterrupted mononucleotide tracts are especially informative for detecting mismatch-repair failure.

Clinical tests classify a tumor by examining multiple repeat loci. Historically, the main categories have been microsatellite instability-high, or MSI-H; microsatellite instability-low, or MSI-L; and microsatellite stable, or MSS. With the original five-marker Bethesda panel, instability at two or more markers generally indicated MSI-H, instability at one indicated MSI-L, and no unstable markers indicated MSS. For larger historical panels, instability at roughly 30–40% or more of tested loci was commonly used to define MSI-H. These numerical rules are assay-specific conventions, not universal biological thresholds.

Modern mononucleotide panels and sequencing-based tests use their own validated criteria. Consequently, an MSI score or numerical threshold cannot necessarily be transferred from one platform to another. MSI-L has also proved less biologically distinct than MSI-H: in many clinical settings, MSI-L tumors resemble MSS tumors more closely than mismatch-repair-deficient tumors. Laboratories may therefore emphasize the distinction between MSI-H and non-MSI-H rather than treating all three categories as equivalent disease classes.

Microsatellite stability does not mean that a tumor has a normal genome. An MSS cancer may contain extensive point mutations, structural rearrangements, or abnormal chromosome numbers. Conversely, a strongly MSI-H tumor need not have extensive chromosome-level instability. These are different dimensions of genomic alteration, although they can coexist within the same cancer.

Mismatch repair and the molecular production of MSI

Mismatch repair corrects replication errors that escape the proofreading activity of DNA polymerases. Its substrates include incorrectly paired bases and small insertion–deletion loops. Repair requires coordinated recognition of the error, identification of the newly synthesized strand, removal of a DNA segment containing the error, and resynthesis using the intact strand as a template. Failure of this system raises mutation rates throughout the genome, with particularly conspicuous effects at repetitive sequences.

In human cells, the MSH2–MSH6 complex, often called MutSα, recognizes base mismatches and many small insertion–deletion loops. MSH2 can also pair with MSH3 to form MutSβ, which preferentially recognizes larger loops. Downstream repair is coordinated principally by the MLH1–PMS2 complex, known as MutLα. Other components of replication and DNA repair participate in strand discrimination, excision, and replacement synthesis. The proteins do not simply “measure” repeat lengths; they recognize abnormal DNA structures generated during copying.

Loss of one component can destabilize its binding partner. This explains common patterns in tumor protein testing: loss of MLH1 is frequently accompanied by loss of PMS2, whereas loss of MSH2 commonly produces loss of MSH6. Isolated loss of PMS2 or MSH6 also occurs. These patterns reflect protein-complex biology and help narrow the possible molecular defect, but none by itself establishes whether the cause is inherited or acquired.

Mismatch-repair deficiency can result from damaging sequence variants, deletions, or epigenetic silencing. A frequent sporadic mechanism is methylation of the MLH1 regulatory region, which suppresses transcription without changing the coding sequence. This links MSI to epigenetics: a stable alteration in gene expression can produce a mutator phenotype comparable to that caused by a disruptive coding mutation. Reduced repair allows replication errors to become fixed as heritable changes in descendant tumor cells.

The relationship between repair failure and detectable MSI is nevertheless not instantaneous or absolute. A tumor must accumulate sufficient errors at the loci measured by an assay. The affected repair gene, residual activity, number of cell divisions, and marker selection all influence detection. MSH6-deficient tumors, for example, can show less conspicuous instability with some older marker panels than tumors lacking MLH1 or MSH2.

Other forms of repeat instability also exist. Elevated microsatellite alterations at selected tetranucleotide repeats, commonly abbreviated EMAST, have been associated in some settings with altered MSH3 function. This pattern is not equivalent to conventional MSI-H and does not automatically carry the same established diagnostic or treatment implications. The broader concept of repeat instability therefore includes biologically distinct processes that should not be collapsed into a single clinical category.

Hereditary predisposition and sporadic repair failure

The best-known inherited condition associated with MSI is Lynch syndrome, caused by a constitutional pathogenic variant in MLH1, MSH2, MSH6, or PMS2, or by certain deletions affecting EPCAM that lead to epigenetic silencing of adjacent MSH2. Affected individuals inherit increased susceptibility to several cancers rather than an already unstable genome in every tissue. Most normal cells retain a functional copy of the relevant repair gene; tumor development commonly involves acquired inactivation of that remaining copy.

Lynch syndrome is inherited in an autosomal dominant pattern at the level of cancer predisposition. The cellular repair defect generally follows a two-hit pattern: both functional copies must be sufficiently impaired in the tumor lineage. The second event may be a mutation, loss of a chromosomal segment, or another inactivating alteration. Cancer risks and tumor distributions differ among the associated genes, so Lynch syndrome is not a uniform biological or clinical entity.

A distinct condition, constitutional mismatch repair deficiency, results from pathogenic variants affecting both copies of a mismatch-repair gene from conception. It is associated with childhood cancers, including brain tumors, hematologic malignancies, and gastrointestinal cancers, and may include nonmalignant features such as café-au-lait-like pigmentation. Because repair impairment is constitutional rather than restricted to a developing tumor, its biology differs fundamentally from the usual heterozygous predisposition in Lynch syndrome. Specialized assays may be needed to detect instability in non-neoplastic tissues.

Most MSI-H colorectal cancers, however, are not caused by Lynch syndrome. A major sporadic pathway involves acquired MLH1 promoter hypermethylation. In colorectal cancer, this often occurs within a broader pattern of extensive promoter methylation and is frequently associated with the BRAF V600E mutation. The association is useful for interpreting the origin of MLH1-deficient colorectal tumors, but it is not a universal rule, and BRAF testing does not have the same exclusionary role across all tumor types.

Another important explanation is the presence of two acquired, or somatic, alterations in a mismatch-repair gene. Such “double-somatic” tumors can resemble Lynch-associated cancers in both protein staining and MSI testing, despite the absence of a detectable inherited pathogenic variant. Rare constitutional MLH1 epimutations and mosaic alterations further complicate the distinction between inherited and sporadic disease.

Accordingly, MSI is evidence about a tumor’s repair phenotype, not a direct test of inheritance. Tumor analysis, constitutional genetic testing, family history, and sometimes paired tumor–normal sequencing address different parts of the causal question. A tumor can be MSI-H without an inherited predisposition, while an individual with Lynch syndrome can occasionally develop a tumor unrelated to the inherited repair defect.

Mutational consequences and tumor evolution

Mismatch-repair failure produces more than altered marker lengths. It increases small insertions and deletions, especially within repetitive DNA, and generates characteristic patterns of base substitutions. These patterns form part of the mutational record of a tumor’s development. Sequencing studies can identify combinations of mutation types associated with defective repair, although such signatures are not identical to a direct MSI measurement.

Most microsatellites lie outside protein-coding sequences, and many repeat-length changes have no demonstrated effect on tumor behavior. A smaller subset occurs within coding regions. An insertion or deletion whose length is not divisible by three changes the reading frame, producing a frameshift mutation. This can alter the downstream amino acid sequence and introduce a premature termination signal. The result may be loss of a functional protein, degradation of the altered messenger RNA, or production of an abnormal peptide.

Recurrent targets in MSI-H colorectal cancers include coding repeats in TGFBR2, ACVR2A, and BAX. TGFBR2 and ACVR2A participate in growth-regulatory signaling, while BAX contributes to apoptosis. Mutations in these genes can influence the balance between proliferation and cell death. Repeatedly observed mutations are not automatically equivalent in functional importance, however: some arise frequently because a repeat is intrinsically mutable, whereas others are favored because they confer a selective advantage.

MSI-H tumors illustrate the distinction between mutation supply and selection. Defective repair generates many variants, but the tumor population is shaped by selection acting on cell survival, growth, tissue interactions, and immune escape. Different subclones may accumulate different repeat lengths, producing heterogeneity within a single tumor. A bulk tissue test averages these populations and may not reveal the full range of subclonal instability.

MSI is also related to, but distinct from, tumor mutational burden, usually expressed as the number of qualifying somatic mutations per megabase of analyzed DNA. MSI-H tumors often have high mutational burden, yet some highly mutated tumors are microsatellite stable. Pathogenic proofreading-domain alterations in POLE or POLD1, for example, can produce a very high mutation burden through a mechanism different from primary mismatch-repair failure.

The biological effects of frameshifts extend beyond protein loss. Altered reading frames can create novel peptide sequences that the immune system has not previously encountered. Whether these sequences become immunologically visible depends on transcription, RNA surveillance, protein processing, and antigen presentation. Thus, the large number of mutations in MSI-H cancer is a substrate for both tumor evolution and immune recognition, rather than a simple measure of tumor aggressiveness.

Distribution across cancers and pathological features

MSI-H is unevenly distributed among cancer types. In colorectal cancer, it is found in approximately 10–15% of newly diagnosed cases overall, with variation by population, stage, and assay. Its frequency is lower in metastatic disease, commonly around 4–5%. Endometrial carcinoma has a higher proportion of mismatch-repair-deficient tumors, often approximately one-quarter to one-third in large series. Substantial MSI-H subsets also occur in gastric and small-intestinal cancers, whereas many other solid tumors show it only infrequently.

These figures require attention to definitions. Studies reporting mismatch-repair protein loss do not always produce exactly the same prevalence as studies measuring MSI directly. Case selection, histological subtype, disease stage, and technical thresholds also affect estimates. Pan-cancer sequencing cohorts may overrepresent advanced disease or patients referred for molecular testing and therefore need not reflect an unselected population.

In the colon, sporadic MSI-H cancers occur disproportionately in the proximal, or right-sided, colon and are more frequent in older patients. Recognized histological associations include prominent tumor-infiltrating lymphocytes, lymphoid aggregates near the tumor margin, mucinous differentiation, and medullary morphology. None of these features is specific enough to establish MSI status without molecular or protein-based testing. A conventionally appearing adenocarcinoma may be MSI-H, and a lymphocyte-rich tumor may be microsatellite stable.

MSI-H colorectal cancer also intersects with particular precursor pathways. Some sporadic cases arise through serrated lesions, with progressive epigenetic alteration and eventual MLH1 silencing. Lynch-associated colorectal cancers develop in a different inherited context and do not necessarily share the same age distribution, methylation profile, or accompanying driver mutations. Treating all MSI-H colorectal cancers as one etiological group would obscure these differences.

In endometrial cancer, mismatch-repair deficiency forms one of the major molecular categories used in contemporary classification, alongside POLE-mutated, p53-abnormal, and no-specific-molecular-profile groups. Interpretation can be complicated when a tumor has more than one defining molecular alteration, and classification systems use rules for such multiple-classifier cases. The significance of MSI therefore depends on its place within the broader molecular profile.

Rare MSI-H tumors have been documented in organs including the pancreas, prostate, biliary tract, and urinary tract. Their rarity makes large, site-specific outcome studies difficult. An established molecular finding can support a shared biological classification across organs, but it does not erase differences in tissue of origin, accompanying mutations, tumor microenvironment, or treatment history.

Laboratory detection and interpretation

Three major approaches are used to evaluate the mismatch-repair/MSI phenotype: repeat-length analysis by polymerase chain reaction, immunohistochemistry for repair proteins, and sequence-based computational analysis. They measure different features. PCR and sequencing evaluate DNA instability, whereas immunohistochemistry evaluates the presence and distribution of selected proteins. Close agreement is common in well-characterized colorectal cancers, but discordance is biologically and technically possible.

Traditional PCR testing amplifies a panel of microsatellite loci, with fragment lengths measured by capillary electrophoresis. Tumor peaks are compared with matched non-neoplastic DNA or with validated reference distributions. The historical Bethesda panel included two mononucleotide markers, BAT-25 and BAT-26, and three dinucleotide markers, D2S123, D5S346, and D17S250. Later panels emphasizing mononucleotide repeats improved discrimination of the classic mismatch-repair-deficient phenotype and reduced some interpretive difficulties associated with dinucleotide markers.

Many mononucleotide markers are nearly invariant across a reference population, allowing some validated assays to operate without matched normal tissue. Nevertheless, constitutional variation, including population-associated differences at individual loci, can resemble tumor-specific change. Repeat amplification also creates stutter peaks through polymerase slippage during the assay itself. Interpretation therefore depends on the overall panel, validated thresholds, and the distinction between technical background and genuine shifted tumor alleles.

Immunohistochemistry generally examines MLH1, PMS2, MSH2, and MSH6 in tumor-cell nuclei. Non-neoplastic lymphocytes, stromal cells, or other internal controls should retain staining when tumor staining is interpreted as lost. Poor fixation, weak staining, heterogeneous loss, or treatment-related changes can complicate assessment. Conversely, some pathogenic missense variants leave an antigenically detectable but functionally defective protein, producing retained staining despite impaired repair.

DNA sequencing can interrogate hundreds or thousands of repeats and use computational methods to infer MSI from length distributions or related sequence features. Assays may analyze paired tumor and normal samples or tumor DNA alone. Their performance depends on sequencing depth, the repeat regions covered, read alignment, tumor purity, and training or validation across relevant cancer types. A threshold established for one platform is not automatically appropriate for another.

False-negative or indeterminate results can arise when the specimen contains too few tumor cells, DNA is degraded, or the selected markers are insensitive to the underlying defect. Spatial heterogeneity can also make different tissue blocks appear discordant. Plasma-based testing introduces additional dependence on the fraction of circulating tumor DNA. Laboratory interpretation therefore distinguishes a confidently stable result from a technically inadequate or insufficiently informative test.

Hereditary assessment, classification, and prognosis

MSI testing initially gained clinical importance as a way to identify tumors associated with Lynch syndrome. Many health systems now use broad or universal tumor screening in newly diagnosed colorectal and endometrial cancers rather than restricting testing to patients with a conspicuous family history. This reflects the fact that age and pedigree criteria alone miss some affected individuals. Screening a tumor for repair deficiency and establishing a hereditary diagnosis remain separate processes.

The pattern of protein loss can guide investigation of the cause. MLH1/PMS2 loss often leads to assessment of MLH1 promoter methylation; in colorectal cancer, BRAF V600E status can also help distinguish common sporadic pathways. MSH2/MSH6 loss, isolated MSH6 loss, and isolated PMS2 loss suggest different underlying gene defects. These are probabilistic associations, not definitive assignments, because somatic mutations and unusual molecular mechanisms can produce similar staining patterns.

MSI also has prognostic associations independent of its hereditary implications. In localized colon cancer, particularly stage II disease, MSI-H or mismatch-repair deficiency is generally associated with more favorable outcomes than mismatch-repair proficiency. This association does not apply uniformly across every stage, anatomical site, or treatment setting. In advanced disease, accompanying alterations, including BRAF mutations in some colorectal cancers, and available therapies substantially influence observed outcomes.

The distinction between a prognostic and a predictive biomarker is essential. A prognostic marker is associated with outcome regardless of a particular treatment; a predictive marker is associated with differential treatment effect. Analyses of adjuvant studies have shown that stage II mismatch-repair-deficient colon cancers do not demonstrate the same benefit from fluoropyrimidine monotherapy as repair-proficient cancers. This finding cannot simply be extended to all chemotherapy combinations, all stages, or all gastrointestinal tumors.

MSI status also contributes to molecular classification. In colorectal cancer, many MSI-H tumors fall within an immune-rich molecular subgroup, although different classification frameworks measure different biological features. In endometrial cancer, repair deficiency is integrated with other molecular classifiers and pathological findings. Molecular categories supplement rather than replace information such as anatomical stage, histological type, and surgical findings.

A clinically meaningful interpretation therefore identifies the question being answered. The same MSI-H result may indicate a repair-defective tumor phenotype, support further assessment of hereditary predisposition, carry stage-specific prognostic information, or predict sensitivity to checkpoint blockade. These uses are related but not interchangeable. In particular, neither an MSI-H result alone proves Lynch syndrome nor an MSS result proves that all hereditary cancer susceptibility has been excluded.

Immune recognition and checkpoint-blockade therapy

MSI-H tumors often generate many altered peptides, including peptides produced by coding frameshifts. Some function as neoantigens: tumor-associated antigens arising from acquired genomic changes. For a neoantigen to be recognized, its source gene must be expressed, the peptide must be generated and presented by a major histocompatibility complex molecule, and a suitable T cell must encounter it. Mutation number alone therefore does not determine the strength of an antitumor immune response.

Many MSI-H cancers contain substantial lymphocytic infiltrates and show activation of inflammatory pathways. At the same time, they may express inhibitory molecules that restrain T-cell activity. The PD-1 pathway is one such immune checkpoint. PD-1 on immune cells can interact with ligands including PD-L1, reducing effector activity under appropriate conditions. Therapeutic antibodies that block this interaction can restore antitumor responses in a proportion of patients.

Clinical evidence established MSI-H and mismatch-repair deficiency as important predictive biomarkers for checkpoint blockade. A 2015 study led by Dung Le and colleagues demonstrated markedly different pembrolizumab activity between mismatch-repair-deficient and mismatch-repair-proficient colorectal cancers. In 2017, the United States Food and Drug Administration granted pembrolizumab a tissue-agnostic accelerated approval for specified previously treated, unresectable or metastatic MSI-H or mismatch-repair-deficient solid tumors. This was a landmark authorization based on a shared molecular feature rather than solely on the organ of origin.

The randomized KEYNOTE-177 trial subsequently compared first-line pembrolizumab with chemotherapy-based treatment in metastatic MSI-H or mismatch-repair-deficient colorectal cancer. The initial report showed median progression-free survival of 16.5 months with pembrolizumab versus 8.2 months with chemotherapy. The trial established a substantial treatment effect in this molecularly defined population, while also showing that some patients experienced early progression. Checkpoint combinations involving PD-1 and CTLA-4 blockade have also demonstrated activity in this disease group.

MSI-H is not a guarantee of response. Some tumors show primary resistance, and others progress after an initial response. Potential mechanisms include impaired antigen presentation, alterations affecting interferon signaling, an immunosuppressive microenvironment, and variation in clonal neoantigens. Loss of beta-2-microglobulin can disrupt aspects of class I antigen presentation, but the relationship between individual alterations and clinical resistance is context-dependent rather than absolute.

Treatment effects also depend on disease setting and regimen. Evidence from advanced cancers cannot automatically establish identical benefits in localized disease or justify pooling all tumor types without qualification. Checkpoint blockade can cause immune-mediated adverse effects in multiple organs, so its clinical significance includes both efficacy and toxicity. MSI testing identifies a biologically enriched population; it does not by itself specify every aspect of treatment outcome.

Discovery, standardization, and continuing research

The recognition of MSI as a cancer phenotype emerged in 1993, when independent groups reported widespread alterations in repetitive DNA in colorectal tumors. Studies associated with Stephen Thibodeau, Lauri Aaltonen, and Manuel Perucho’s group, including work by Yoram Ionov and colleagues, helped establish that some cancers followed a route distinct from the chromosome-loss patterns emphasized in earlier models. The unstable phenotype was initially described using terms such as replication error-positive, or RER-positive.

These observations converged rapidly with research on inherited colorectal cancer and DNA repair. During 1993–1994, defects in human mismatch-repair genes, including MSH2 and MLH1, were linked to hereditary nonpolyposis colorectal cancer, now generally discussed within the framework of Lynch syndrome. Experimental work in microbial systems had already established the role of mismatch repair in replication fidelity. The cancer findings connected that basic mechanism to human tumor predisposition and explained why short repeats became unstable.

Standardization followed because early studies used different marker panels and definitions. A National Cancer Institute workshop held in 1997, with recommendations published in 1998, established the influential Bethesda reference panel and reporting framework. Subsequent development of mononucleotide panels, four-protein immunohistochemistry, and sequencing-based classifiers expanded testing beyond the original assays. Modern standardization addresses not only marker choice but also specimen quality, tumor content, analytical validation, and the handling of discordant results.

Current research examines MSI as a dynamic evolutionary process rather than only a binary laboratory label. Single-cell and spatial approaches investigate when repair deficiency arises, how unstable subclones spread, and how immune selection changes the tumor population. Repeat-rich regions remain technically difficult for many short-read sequencing methods, making the accurate characterization of individual repeat alleles an ongoing analytical problem. Long-read technologies offer different capabilities but introduce their own requirements for error modeling and validation.

Another research direction concerns recurrent frameshift-derived neoantigens. Because unrelated MSI-H tumors can acquire mutations in the same coding repeats, some altered peptide sequences are shared across patients. Such shared targets are being investigated in therapeutic vaccine and immune-monitoring studies. Their usefulness depends on factors including mutation clonality, expression, peptide presentation, and variation in human leukocyte antigen types; recurrence of a DNA mutation alone does not establish an effective immune target.

Noninvasive measurement is also developing through circulating tumor DNA and other specimen types. Here, sensitivity is constrained by how much tumor-derived material reaches the sample, particularly in early-stage disease or low-shedding tumors. In parallel, computational models are being studied for predicting MSI from routine histological images. These approaches require independent validation across laboratories and patient populations because staining methods, scanners, tissue composition, and cancer type can all alter model performance.