Targeted Next Generation Sequencing Identifies Novel Genetic Mutations in Patients with Waldenstrom's Macroglobulinemia/Lymphoplasmacytic Lymphoma or IgM-Monoclonal Gammopathies of Undetermined Significance

Background. TheMYD88 (L265P) somatic mutation is present in more than 90% of patients (pts) with Waldenström's Macroglobulinemia (WM)/lymphoplasmacitic lymphoma (LPL). The second most common mutations are nonsense (NS) or frameshift (FS) mutations in the CXCR4 gene, detectable in approximately 25-30% of WM pts by Sanger sequencing. Limited data are available about other genetic mutations in WM/LPL and its precursor condition IgM-monoclonal gammopathy of undetermined significance (IgM-MGUS).

Pts and methods. Using targeted next generation sequencing (NGS), we evaluated the prevalence of somatic mutations of 11 genes selected on the basis of evidences available from the literature (MYD88, CXCR4, ARID1A, KMT2D, TP53, NOTCH2, PRDM1, CD79b, TRAF3,TNFAIP3, MYDBBP1A) in 119 pts, classified as WM/LPL (n=63) or IgM-MGUS (n=56) according to International Consensus Criteria. Median age of pts (67 males, 52 females) was 65 years (range: 38-82). Samples were collected at diagnosis (n=101), after diagnosis but before any treatment (n=9) or at progression after therapy (n=9). Paired tumor and germline DNA extracted respectively from CD19-selected and CD19-depleted bone marrow (BM) mononuclear cells was available in all pts. Mean resequencing depth across gene panel was 1009x. Only mutations tagged as oncogenic or possibly oncogenic based on information derived from the literature and on in silico prediction effect were considered in the analysis. For MYD88 (L265P) and CXCR4 mutations, results obtained with NGS were compared with those obtained respectively with allele-specific PCR (AS-PCR) and Sanger sequencing.

Results. Overall, we found 151 mutations in 88 pts (74%). The median number of mutations was significantly higher in WM/LPL as compared with IgM-MGUS and in pts previously treated as compared with untreated ones (median 2 versus 1, P < 0.001 for both comparisons). MYD88 mutations were found in 80/119 pts (67%), with a median allele burden of 34.2% (range: 2.5-93.3%). The prevalence of MYD88 mutations was significantly higher in WM/LPL as compared with IgM-MGUS (86% versus 46%, P <0.001). MYD88 mutations other than classical L265P (n=76) were found in 4 pts and were represented by V217F (n=2), S219C (n=1), M232T (n=1). Fifteen pts who were MYD88 (L265P) wild-type by NGS were found to be mutated by AS-PCR (K coefficient of concordance between NGS and AS-PCR: 70%, P < 0.001). CXCR4 mutations were found in 19/119 pts (16%), with a median allele burden of 34% (range: 4.2-84%). The prevalence of CXCR4 mutations was significantly higher in WM/LPL as compared with IgM-MGUS (24% versus 7%, P < 0.02). The K coefficient of concordance between NGS and Sanger was 83% (P < 0.001), with 2 pts mutated only by NGS and 2 pts mutated only by Sanger. Somatic mutations were also found in KMT2D (formerly known as MLL2) (16% of pts), TP53 (8%), NOTCH2 (7%), PRDM1 (4%), ARID1A (3%), CD79b (2%), and TRAF3 (1%). No mutations were found in MYBBP1A and TNFAI3. Overall, the prevalence of these mutations was significantly lower in pts wild-type either for MYD88 or CXCR4 as compared with those with MYD88 and/or CXCR4 mutations (15% versus 41%, P = 0.04). The prevalence of KMT2D mutations was significantly higher in WM/LPL as compared with MGUS (25% versus 5%), while for the other genes the distribution was not statistically different according to diagnosis. With a median follow-up of 20 months (range: 0-264), we did not find a statistically significant correlation between genetic mutations and pts' outcome in terms of overall survival or time to first treatment.

Conclusions. In this cohort of pts with WM/LPL and IgM-MGUS studied with NGS we could demonstrate that: i) NGS identifies MYD88 mutations other than L265P in a small proportion of pts; ii) the prevalence of CXCR4 mutations by Sanger is confirmed by NGS, despite the higher sensitivity of the latter method; iii) the subgroup of pts wild type either for MYD88 or CXCR4 shows a low incidence of other genetic mutations; iv) 25% of pts with WM/LPL were found to carry KMT2D mutations, a prevalence similar to that reported in marginal zone lymphoma; v) genetic mutations are more common in WM/LPL than in IgM-MGUS in agreement with the hypothesis that multiple genetic hits are required for progression from a pre-benign condition to a neoplastic disease; vi) due to the indolent nature of these disorders, longer follow-up is probably needed to see the prognostic impact of these mutations, if any.