Distinct involvement of NF‐κB regulators by somatic mutation in ocular adnexal malt lymphoma

Mucosa-associated lymphoid tissue (MALT) lymphoma are characterized by t(11;18)(q21;q21)/BIRC3 (API2)-MALT1, t(1;14)(p22;q32)/BCL10-IGH@ and t(14;18)(q32;q21)/IGH@MALT1, which cause constitutive nuclear factor (NF)-jB activation. Intriguingly, these translocations are seen primarily in those of stomach and lung, but rarely or not at all in those of the ocular adnexa, thyroid, salivary gland and skin. We recently showed that TNFAIP3 was inactivated by somatic mutations in 28 6% of ocular adnexal MALT lymphomas (OAML) and its inactivation was associated with enhanced expression of the NF-jB target genes (Bi et al, 2012). The findings reinforce the notion that NF-jB activation may be a common molecular mechanism underlying the genesis of MALT lymphoma of various sites (Du, 2011) and that somatic mutation may be the major genetic events that activate NF-jB in translocation-negative MALT lymphomas. To further examine the extent of genetic bases of NF-jB activation in OAML, we investigatedMYD88, CARD11, CD79A/B and BIRC3 (API2) mutation in a large cohort previously studied for TNFAIP3 abnormalities (Bi et al, 2012). We also investigated PRDM1 mutation because the gene is commonly found within the region of 6q deletion in OAML (Chanudet et al, 2009) and also frequently mutated in activated B-cell like diffuse large B-cell lymphoma (ABC-DLBCL)(Mandelbaum et al, 2010). All the coding exons of MYD88 were studied in 105 cases of OAML by polymerase chain reaction (PCR) and Sanger sequencing. Mutation was found in 6 (5 7%) cases and included missense changes (L265P in 3 cases, S243N in 1 case) reported in previous studies (Li et al, 2012), and a novel in-frame deletion (delT285-T294 in 2 cases) (Fig 1, Supplementary Fig S1). Of the latter, it was possible to isolate non-neoplastic cells at approximately 60% of purity in one case, and sequence analysis of the DNA sample from the microdissected cells indicated that the deletion was a somatic event. MYD88 mutation was largely mutually exclusive from TNFAIP3 inactivating mutation, with only one case bearing both mutations. The two missense mutations previously showed gain of ability to activate NF-jB (Ngo et al, 2011). To examine the functional impact of the novel in-frame deletion, we first investigated the mutant using NF-jB reporter assay and demonstrated that the deletion mutant was even more potent in NF-jB activation than the MYD88 L265P mutant (Fig 2A). With the exception of TLR3, MYD88 is a universal adaptor for IL1R and TLR signalling. In addition to the canonical NF-jB pathway, MYD88 also couples the receptor signalling to AP-1 and IRF activation pathway. We next tested whether MYD88 mutants gained ability to activate the AP-1 transcriptional factor. As shown by the in vitro reporter assay, both the missense and in-frame deletion mutants gained ability to activate AP-1, with the deletion mutant being more potent than the missense mutant (Fig 2A). Previous transcriptomic analyses of MALT lymphoma showed enhanced expression of NF-jB target genes CCR2, TLR6, BCL2 and CD69 in cases with chromosome translocation (Hamoudi et al, 2010). To investigate whether MYD88 mutation lead to increased NF-jB activities, we measured the expression of the above NF-jB target genes in five cases with MYD88 mutation and 11 cases without MYD88, TNFAIP3, TNIP1 (ABIN1), TNIP2 (ABIN2) and CARD11 mutation, and MALT1 and IGH@ translocation. The expression of these NF-jB target genes was higher in cases with MYD88 mutation than those without the mutation, with TLR6 showing a significant difference (Fig 2B). The MYD88 missense mutations are clustered in the Toll/ IL-1 receptor (TIR) domain and these mutants can spontaneously recruit and activate its downstream molecules, such as IRAK1 and IRAK4 (Ngo et al, 2011). The deletion mutation seen in this study affects the region including the key amino acids such as D288 (D275 according to the second translation start site), known to confer dominant negative inhibitory function (Li et al, 2005). The above findings of both in vitro and in vivo studies indicate that these MYD88 mutants are most probably constitutively active. Sequencing of the coding exons 5–9 of CARD11 in 105 cases of OAML identified only a novel missense alteration (G1466A, Glu378Lys) (Fig S1) in one case. The mutation was somatic, as shown by analyses of the microdissected normal cells. The case with CARD11 mutation was also negative for BCL10 and MALT1 translocation, TNFAIP3 and MYD88 mutation. The coding exon 5 of CD79A and exons 5-6 of CD79B, where mutation hot points were seen in ABC-DLBCL, were sequenced in all 105 cases, but showed no mutation, in line with recently published work (Liu et al, 2012). All the coding exons of BIRC3 were sequenced in 52 selected cases that were negative for TNFAIP3, TNIP1/2, MYD88mutations and chromosome translocation. This identified a novel missense mutation (R411K) in one case. Due to lack of normal DNA, it was not possible to confirm whether this mutation was somatic.