The Genomic Landscape of MYD88 Wild-Type WaldenströM's Macroglobulinemia Ischaracterized By Somatic Mutations in TBL1XR1, the CBM Complex, and NFKB2

Abstract

Waldenstrom's Macroglobulinemia (WM) is characterized by bone marrow (BM) involvement of IgM secreting lymphoplasmacytic lymphoma. Whole genome sequencing identified somatic activating mutations in MYD88 and CXCR4 in 95% and 35-40% of WM patients, respectively. Somatic mutations were observed in ARID1A, CD79B, TP53, and MYBPB1A in MYD88 mutated WM patients. Deletions in in chromosome 6q were also observed in half of MYD88 mutated patients. The somatic events that drive MYD88 wild-type (WT) WM remain to be defined. Patients with MYD88 WT WM show an increased risk of death compared to their MYD88 mutated counterparts (Treon et al, Blood 2014). Moreover, WM patients with MYD88 WT disease show no major response to ibrutinib versus 81% in MYD88 mutated patients (Treon et al, NEJM 2015). These findings point to fundamental genomic differences in tumor biology between MYD88 mutated and wild-type patients. To better understand this rare and aggressive form of WM, we performed whole exome sequencing of 12 WM patients who met the clinicopathological diagnostic criteria for WM using WHO and WM consensus criteria. These patients were WT for MYD88 following genotyping by highly specific AS-PCR for MYD88 L265P and Sanger sequencing to exclude non-L265P MYD88 mutations. CD19-selected BM mononuclear cells were used for tumor sampling, and CD19-depleted cells for germline comparisons. This cohort was split evenly between males and females. Their median age was 58 (range 50-85 years), median BM disease involvement was 12.5% (range 2.5%-80%), and median serum IgM was 4,055 (range 1,070-5,620 mg/dL). Seven (58%) patients had adenopathy, and 3 (25%) had splenomegaly. Seven (58%) patients were previously treated. Next generation sequencing was performed by the Center for Cancer Computational Biology, and the data were analyzed following the Broad Institutes GATK Best Practice Guidelines. Small variants were analyzed using both Strelka and MuTect2. Somatic structural variants were detected using Manta, and copy number alterations were called using Control-FREEC. All somatic variants were validated by Sanger sequencing. The most recurrently mutated gene was TBL1XR1, which was observed in 3 (25%) patients, with one patient harboring two TBL1XR1 mutations. Two missense mutations were observed within distinct WD40 domains, while an early frameshift mutation is predicted to impact all WD40 domains. A fourth mutation is predicting to effect splicing. TBL1XR1 mutants in the WD40 domain are found in marginal zone lymphomas and are known to increase TBL1XR1 binding with nuclear receptor corepressor (NCoR), leading to increased degradation of NCoR and the activation of NF-κB and JUN target genes (Jung et al, Oncotarget 2017). Novel, truncating MALT1 mutations were present in two (17%) patients. Notably, these nonsense mutations were 23 base pairs apart and predicted for truncation of the c-terminal domain with loss of a TRAF6 binding site. Finally, a BCL10 nonsense mutation was observed just upstream of a critical regulatory phospho-serine site. Both MALT1 and BCL10 are members of the CARD11-BCL10- MALT1 (CBM) complex that signals downstream to NF-kB. Other validated somatic mutations included TIMELESS (n=3; 25%), SHROOM3 (n=2; 17%), BLM (n=1; two mutations found in the same patient); KCNB1 (n=1), IGF1-R (n=1), NOTCH1 (n=1). No CXCR4 mutations were detected. The genomic and protein coordinates for these somatic mutations are shown in Table 1.Our structural variant analysis also identified large deletions that removed the DEATH/PEST domain of NFKB2 in two patients covering amino acids 691-822 and 711- 839, respectively. Deletions in the DEATH/PEST domain of NFKB2 are associated with constitutive non-canonical NF-kB activation. All together, NF-kB related mutations were present in 6/12 (50%) MYD88 wild-type WM patients. Analysis of copy number alterations revealed no highly recurrent events, but was remarkable for the absence of deletions on chromosome 6q which are typically found in 50% of MYD88 mutated WM patients. Our findings highlight for the first time the genomic landscape that surrounds MYD88 wild-type WM, and point to a distinct mutational profile from MYD88 mutated WM. The findings also include several novel somatic mutations not previously reported in other B-cell lymphoproliferative disorders. [Display omitted] DisclosuresCastillo:Pharmacyclics: Consultancy, Research Funding; Millennium: Research Funding; Janssen: Consultancy, Research Funding; Abbvie: Research Funding. Treon:Pharmacyclics: Consultancy, Research Funding.