Search

Background & Aims Fifty percent of colorectal cancers show elevated microsatellite alterations at selected tetranucleotide repeats (EMAST) and are associated with inflammation, metastasis, and poor patient outcome. EMAST results from interleukin 6–induced nuclear-to-cytosolic displacement of the DNA mismatch repair protein Mutated S Homolog 3, allowing frameshifts of dinucleotide and tetranucleotide but not mononucleotide microsatellites. Unlike mononucleotide frameshifts that universally shorten in length, we previously observed expansion and contraction frameshifts at tetranucleotide sequences. Here, we developed cell models to assess tetranucleotide frameshifts in real time. Methods We constructed plasmids containing native (AAAG)18 and altered-length ([AAAG]15 and [AAAG]12) human D9S242 locus that placed enhanced green fluorescent protein +1 bp/-1 bp out-of-frame for protein translation and stably transfected into DNA mismatch repair–deficient cells for clonal selection. We used flow cytometry to detect enhanced green fluorescent protein–positive cells to measure mutational behavior. Results Frameshift mutation rates were 31.6 to 71.1 × 10-4 mutations/cell/generation and correlated with microsatellite length (r2 = 0.986, P = .0375). Longer repeats showed modestly higher deletion over insertion rates, with both equivalent for shorter repeats. Accumulation of more deletion frameshifts contributed to a distinct mutational bias for each length (overall: 77.8% deletions vs 22.2% insertions), likely owing to continual deletional mutation of insertions. Approximately 78.9% of observed frameshifts were 1 AAAG repeat, 16.1% were 2 repeats, and 5.1% were 3 or more repeats, consistent with a slipped strand mispairing mutation model. Conclusions Tetranucleotide frameshifts show a deletion bias and undergo more than 1 deletion event via intermediates, with insertions converted into deletions. Tetranucleotide markers added to traditional microsatellite instability panels will be able to determine both EMAST and classic microsatellite instability, but needs to be assessed by multiple markers to account for mutational behavior and intermediates., Graphical abstract

Background. Oxidative DNA damage and inflammatory cytokines cause loss of function of the MMR protein MSH3. MSH3-MSH2 (MutSβ) function is required for correction of postreplicative polymerase errors in the form of frameshifts at tetranucleotide microsatellite sequences to prevent elevated microsatellite alterations at selected tetranucleotides (EMAST). EMAST is increasingly being used as a biomarker for MSH3 deficiency in subtyping colorectal cancers due to its association with metastasis and poor patient outcome. Here, we sought to understand frameshift mutational behavior of tetranucleotide sequences in the absence of MMR. Methods. We generated EGFP-based mutation reporter model systems containing the D9S242 tetranucleotide microsatellite at its native length of (AAAG)18 and at modified lengths of (AAAG)15 and (AAAG)12 to represent other endogenous microsatellite lengths. Plasmid vectors were constructed for each length that placed EGFP +1 bp or -1 bp out-of-frame for protein translation, enabling us to measure deletion and insertion frameshifts. These constructs were stably-integrated into MMR-deficient HCT116 cells, and monoclonal cell lines containing the constructs were isolated to measure mutation rates and spectra. Deletion and insertion frameshift mutations restored the EGFP reading frame to become detectable by flow cytometry. Genomic DNA from mutated cells were isolated and sequenced for insertion/deletion frameshift mutational biases. Results. Combined insertion/deletion frameshift mutation rates at D9S242 locus ranged from 31.6 X 10-4 to 71.1 X 10-4 mutations/cell/generation and was strongly correlated with longer length of tetranucleotide microsatellites (r2 = 0.986, P = 0.0375). We observed slightly higher deletion mutation rates over insertion rates at longer microsatellites, but equivalent insertion and deletion rates at shorter microsatellites. However, accumulation of more deletion frameshifts over time contributed to a distinct mutational bias for deletions for each length, likely due to continual frameshift mutation at insertions. About 79% of all observed frameshifts (insertion or deletion) were one- repeat (e.g., four nucleotides), 16% two-repeat, and 5% three or more repeat mutations that were consistent with a slipped strand mispairing mutation model. Conclusions. Our results suggest that intact MMR recognizes and repairs shorter length tetranucleotide frameshifts more efficiently than larger ones, and preferentially corrects loops on the template strand (preventing deletions) over loops on the newly-synthesized strand (preventing insertions) in human cells. MMR deficiency thus manifests in human tissue with a deletion mutation bias at tetranucleotide sequences (EMAST). Longer tetranucleotide microsatellites have increased propensities for larger DNA slippage errors. Citation Format: Maide O. Raeker, Jovan Pierre-Charles, John M. Carethers. Insertion and deletion frameshift rates and mutational spectra of tetranucleotide microsatellites in DNA mismatch repair-deficient human cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3368.