Portal de Periódicos da CAPES

Cell‐free DNA analysis for detection of MYD88 L265P and CXCR4 S338X mutations in W aldenström macroglobulinemia

2021; Wiley; Volume: 96; Issue: 7 Linguagem: Inglês

10.1002/ajh.26184

1096-8652

Maria Demos, Zachary R. Hunter, Lian Xu, Nicholas Tsakmaklis, Amanda Kofides, Manit Munshi, Xia Liu, Maria Luisa Guerrera, Carly Leventoff, Timothy P. White, Catherine Flynn, Kirsten Meid, Christopher J. Patterson, Guang Yang, Andrew R. Branagan, Shayna Sarosiek, Jorge J. Castillo, Steven P. Treon, Joshua Gustine,

Viral-associated cancers and disorders

Molecular testing for MYD88 and CXCR4 mutations is increasingly used in patients with Waldenström macroglobulinemia (WM). MYD88L265P is the most frequent variant (93–95%) in WM patients, while non-L265P MYD88 variants (1–2%) can rarely occur.1, 2 The presence of mutated MYD88 is associated with a decreased risk of histological transformation, longer overall survival, and sensitivity to ibrutinib.3, 4 Over 40 nonsense and frameshift CXCR4 variants have been identified in WM patients.5, 6 Nonsense CXCR4S338X variants are the most common, occurring in 50% of WM patients due to C > A or C > G nucleotide transversions.5, 6 Among CXCR4 mutations, CXCR4S338X has the largest clinical impact with higher serum IgM levels, symptomatic hyperviscosity, earlier time to initiation of frontline therapy, and shorter progression-free survival with ibrutinib.3 MYD88 and CXCR4 mutations can therefore provide important data regarding diagnosis, prognosis, and treatment response in WM.3 The use of bone marrow (BM) aspirate materials represents the current "gold standard" for molecular testing in WM.1-3, 5, 6 Although MYD88 and CXCR4 mutations can also be identified in peripheral blood (PB), the diagnostic yield in PB is inferior to BM, particularly for previously treated patients.7, 8 Tumor enrichment with B-cell selection can significantly improve testing sensitivity, but pre-sorting B cells is time-consuming and not feasible in most clinical laboratories.7, 9 Recent studies have demonstrated the feasibility of identifying MYD88 and CXCR4 mutations by using cell-free DNA (cfDNA) from WM patients.10-12 These findings prompted us to perform a comprehensive analysis comparing the use of cfDNA to matched BM and PB, with or without B-cell selection, for detection of the most common MYD88 (L265P) and CXCR4 (S338X) mutations in WM patients. We prospectively collected matched BM and PB samples from 28 consecutive WM patients. PB was collected in Streck Cell-Free DNA tubes (Streck, La Vista, Nevada) to preserve cfDNA. Tubes were centrifuged and plasma was isolated within 30 h of venipuncture. cfDNA was extracted using the QIAGEN Circulating Nucleic Acid Kit (Qiagen, Hilden, Germany) with all modifications employed by Kang et al.13 Both CD19-selected and unselected BM and PB mononuclear cells were isolated as before.2, 6, 7 Overall, five different tissue fractions were isolated for analysis: CD19-selected BM (BM19+), unselected BM (BMMC), CD19-selected PB (PB19+), unselected PB (PBMC), and cfDNA. Quantitative allele-specific polymerase chain reaction (AS-PCR) assays for MYD88L265P and CXCR4S338X mutations were performed for each tissue fraction using 2.5 ng of DNA as previously described.2, 6, 7 To avoid a potential batch effect, tissue fractions from the same patient were run on the same plate. Calculations were performed with R (R Foundation for Statistical Computing, Vienna, Austria). The p values < .05 were considered statistically significant. The Dana Farber/Harvard Cancer Center IRB approved this study, and all patients provided written consent for sample use. Baseline patient characteristics are shown in Table S1. Mutation testing for MYD88L265P and CXCR4S338X was performed in 28 and 23 patients, respectively; limited tumor DNA precluded evaluation of CXCR4S338X in five patients. Using BM19+ as the reference tissue,1-3, 5, 6 we benchmarked the test performance of MYD88L265P and CXCR4S338X detection for BMMC, PB19+, PBMC, and cfDNA (Table 1). The rates of concordance between the different tissue fractions and BM19+ were consistent with previous studies.7, 9-12 We then compared the clinicopathological characteristics between patients with concordant and discordant results for MYD88L266P by cfDNA. Discordant patients had a lower median BM involvement (9% vs. 45%; p = .04) and serum IgM level (985 vs. 1597 mg/dL; p = .02), as well as a higher hemoglobin level (12.3 vs. 10.7 g/dL; p = .03) versus concordant patients. The BM involvement significantly correlated with serum IgM (r = 0.44, p = .02) and hemoglobin (r = −0.39, p = .04) levels. Consistent with these findings, the concordance for MYD88L265P detection with cfDNA was significantly lower in those with a BM involvement .05 for all comparisons). Discordant patients by BMMC, PB19+, and PBMC tissue fractions similarly had a significantly lower BM involvement (data not shown), akin to previous studies.7, 9 These findings collectively demonstrate that a low BM tumor burden adversely impacts MYD88L265P detection with cfDNA in WM patients. An unexpected observation was in one patient with a "false positive" result for MYD88L265P. The patient had MYD88L265P detected in BMMC and cfDNA, but not the BM19+ fraction following treatment with ixazomib, dexamethasone, and rituximab (IDR). Flow cytometric analysis identified clonal CD38+ plasma cells, but not CD19+ B cells in this patient, suggesting that residual plasma cells unamenable to CD19-selection probably accounted for the detection of MYD88L265P in BMMC and cfDNA but not the CD19+ fraction.9 Indeed, MYD88L265P is present in both clonal B and plasma cells derived from individual WM patients, and clonal plasma cells may persist long after B-cell depleting therapy.1, 14 To account for this treatment effect, we re-calculated the test performance statistics for cfDNA using both the B and plasma cell compartments as reference tissue. The adjusted results yielded a concordance, sensitivity, and specificity of 82%, 80%, and 100%, respectively, for MYD88L265P by cfDNA in WM patients (Table S2). Overall, our findings strengthen the available evidence that PB cfDNA can reliably be used to identify the most common MYD88 (L265P) and CXCR4 (S338X) variants in WM patients. Our data expand upon previous studies by validating cfDNA for MYD88L265P and CXCR4S338X against CD19-selected and unselected BM and PB tissue fractions.10-12 This approach also revealed that cfDNA correctly identified the mutation status of a WM patient who underwent B-cell depleting therapy, a finding likely related to residual clonal plasma cells.9 While a cfDNA "liquid biopsy" does not supplant a BM biopsy to diagnose WM, it can offer a non-invasive, convenient, and potentially cost-effective method to genotype WM patients. MYD88L265P can help discriminate WM from other IgM-secreting malignancies, while both MYD88 and CXCR4 (particularly S338X) mutations may be useful in making treatment decisions, especially if considering therapy with BTK inhibitors.2-4 Because of the high specificity, cfDNA could represent an initial testing modality for WM patients in whom an invasive BM biopsy might be needed solely for genotyping purposes. Our study is not without limitations. The low number of patients with CXCR4S338X precluded an assessment of clinicopathological characteristics and CXCR4S338X concordance with cfDNA. Additionally, we focused our study on MYD88L265P and CXCR4S338X since these constitute the most common variants observed in WM.5, 6 The use of cfDNA with targeted ultra-deep next-generation sequencing may lead to the reliable identification of other MYD88 and CXCR4 variants. Further investigation is also warranted to evaluate cfDNA as a treatment monitoring tool, as acquired BTK, CARD11, and PLCγ2 mutations are associated with disease progression on BTK inhibitors in WM patients.15 In summary, MYD88L265P and CXCR4S338X can be identified with high sensitivity and specificity in cfDNA derived from the plasma of WM patients. The use of cfDNA represents a non-invasive, convenient, and potentially cost-effective method for genotyping WM patients. The authors would like to thank the Orszag Family Fund for WM Research, the D'Amato Family Fund for Genomic Discovery, the International Waldenström's Macroglobulinemia Foundation, the Leukemia and Lymphoma Society (Grant: R6507-18), the Kerry Robertson Fund for WM. S.P.T., Z.R.H. and G.Y. are supported by an NIH SPORE in Multiple Myeloma (Grant: 2P50CA100707-16A1). S.P.T., J.J.C., G.Y., Z.R.H. have received research funding and/or consulting fees from Pharmacyclics Inc., Janssen Pharmaceuticals Inc., the manufacturer of ibrutinib. S.P.T. has received research funding from Bristol Myers Squibb, X4 Pharmaceuticals, and Beigene. J.J.C. received research funding and/or consulting fees from Abbvie, Beigene, Kymera, and TG Therapeutics. Maria G. Demos, Joshua N. Gustine, Zachary R. Hunter, and Steven P. Treon designed the study and performed the analysis. Maria G. Demos, Lian Xu, Guang Yang, Xia Liu, Amanda Kofides, Nicholas Tsakmaklis, Manit Munshi, and Maria Luisa Guerrera prepared and performed molecular testing on patient samples. Kirsten Meid, Christopher J. Patterson, Shayna Sarosiek, Carly R. Leventoff, Timothy P. White, Catherine A. Flynn, Andrew R. Branagan, Jorge J. Castillo, and Steven P. Treon provided clinical care for the patients and collected the samples. Maria G. Demos, Joshua N. Gustine and Steven P. Treon drafted the manuscript. All authors critically reviewed and approved the manuscript. Supplemental Table 1. Patient characteristics at the time of mutation assessment. Supplemental Table 2. Adjusted test performance findings for cfDNA using both BM19+ and BMMC fractions as reference tissue. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.