MYD88 mutated and wild-type Waldenström’s Macroglobulinemia: characterization of chromosome 6q gene losses and their mutual exclusivity with mutations in CXCR4

Waldenström’s Macroglobulinemia (WM) is a lymphoplasmacytic lymphoma characterized by bone marrow (BM) infiltration of immunoglobulin M (IgM)-secreting lymphoplasmacytic cells.1 Activating mutations in MYD88 are present in 93-97% of WM and 50-70% of IgM monoclonal gammopathy of undetermined significance (MGUS) patients.2 IgM MGUS patients with MYD88 mutations may be at a higher risk of progression to WM.3 Mutated MYD88 triggers WM cell growth and survival by activation of nuclear factor κ-light-chainenhancer of activated B cells (NF-κB) pro-survival signaling through interleukin-1 receptor-associated kinase 1 (IRAK1)/IRAK4 and Bruton’s tyrosine kinase (BTK). WM patients with wild-type MYD88 (MYD88^WT) show recurring somatic mutations in TBL1XR1, NFKB2, and the CARD11-BCL10-MALT1 (CBM) complex genes BCL10 and MALT1, and show shorter survival and a higher incidence of associated diffuse large B-cell lymphoma events versus MYD88 mutated (MYD88^MUT) patients.4 MYD88^MUT WM patients also show higher levels of response and progression-free survival to ibrutinib in comparison to MYD88^WT patients.5 Activating CXCR4 mutations are present in 30-40% of WM patients, and are typically subclonal to MYD88 mutations.6 CXCR4 mutations trigger pro-survival protein kinase B (AKT) and extracellular signal-regulated kinase 1 (ERK1/2) signaling and are associated with inferior and/or delayed response to many WM therapeutics, including ibrutinib.75 Deletions in chromosome 6q (del6q) occur in about 50% of WM patients, and are associated with the transition from IgM-MGUS to WM.98 The functional role of del6q in this transition, and their relationship to MYD88 and CXCR4 mutations remain to be characterized. A minimal region of deletion (MDR) for 6q (6q14.1-6q27) in WM patients bearing the MYD88 mutation was previously reported by us,9 and included many genes with important regulatory functions for BTK/B-cell receptor (IBTK), apoptosis (FOXO3), (BCL2 BCLAF1) and NF-kB (TNFAIP3, HIVEP2), signaling (Figure 1A). In the study herein, we sought to delineate the gene losses related to del6q in asymptomatic and symptomatic WM, as defined by The Second International Workshop on WM (IWWM-2) criteria,1 and their association to MYD88 and CXCR4 mutations and signaling.

Figure 1.Characteristics of chromosome 6q deletions in Waldenström’s Macroglobulinemia. (A) Ideogram of chromosome 6q showing the location of all study genes. (B) Heatmap of the statistically significant copy number alterations based on TaqMan RT-PCR analysis for MYD88^MUTpatients. Dark blue indicates a fully clonal deletion (copy number=1) for that gene. Hierarchical clustering of the patients revealed two distinct patterns of chromosome 6q deletion, whereby one group demonstrates more clonal and contiguous deletions, and the other more focal and subclonal gene losses. The mean deletion clonality per patient was highly statistically significant between the two groups (P=0.0002). No differences were observed based on symptomatic status. (C) Contiguous deletions and CXCR4 mutation status stratified by the two populations (clonal and subclonal deleted) that were identified by hierarchical clustering in B. The analysis was restricted to patients with 6q deletions. (D) Heatmap of statistically significant copy number alterations based on TaqMan RT-PCR analysis for MYD88^WTpatients. No CXCR4 mutations were detected in this population, and all patients had symptomatic disease.

The Dana-Farber/Harvard Cancer Center (DF/HCC) Institutional Review Board approved this study and the samples were collected following informed consent. The study cohort included 33 untreated WM patients (21 males, 12 females). Patients had a median age of 62 (range: 35-91) years, BM involvement of 60% (range: 2.5-90%), serum IgM levels of 3,010 (range: 257-6910) mg/dl, and hemoglobin of 10.9 (range: 8.4-14.4) g/dl. MYD88 and CXCR4 mutations were assessed using allele-specific polymerase chain reaction (PCR) and Sanger sequencing as previously described.6 All MYD88 mutations corresponded to p.Leu265Pro and were present in 25 (76%) patients, 11 (44%) of whom also carried CXCR4 activating mutations (CXCR4^MUT). The eight asymptomatic WM patients were all MYD88^MUT, five of whom also carried CXCR4^MUT. The 25 symptomatic WM patients included 17 MYD88^MUT patients (six of whom were also CXCR4^MUT) and all eight MYD88^WTCXCR4^WT patients (Figure 1 B-D). For the five studied genes, copy number alterations (CNA) were measured in quadruplicate and gene expression in triplicate from CD19-selected BM lymphoplasmacytic cells with TaqMan real-time (RT) PCR protocols (Thermo Fisher Scientific, MA, USA) using the following assays for CNA and expression, respectively: IBTK: Hs01076984_cn/Hs00394118_m1; FOXO3: Hs01521732_cn/Hs00818121_m1; BCLAF1: Hs02855566_cn/Hs03004661_g1; TNFAIP3: Hs00548617_cn/Hs00234713_m1; HIVEP2: Hs01433181_cn/Hs00198801_m1. Paired CD19-depleted peripheral blood mononuclear cells (PBMC) were used as germline controls. Paired CD19 and CD19 PBMC from six healthy donors were included to rule out possible B-cell specific findings. Deletions affecting <20% of WM cells were considered to be under the RT-PCR detection threshold. Whole genome sequencing was previously performed in 17/33 (52%) patients and used to validate del6q TaqMan findings.9 Previously published ribonucleic acid sequencing (RNASeq) data10 were reanalyzed using Bioconductor in R (R Foundation for Statistical Computing, Vienna, Austria).

Comparing germline and tumor DNA by CNA assays revealed heterozygous somatic deletions for at least one 6q MDR gene evaluated in 20/25 (80%) MYD88^MUT WM patients. No CNA for any of the 6q MDR genes were observed in any of the healthy donors. In MYD88^MUT WM patients, BCLAF1 was the most frequently deleted gene (19/25; 76%), followed by TNFAIP3 (15/25; 60%). HIVEP2, IBTK, and FOXO3 were each deleted in 13/25 (52%) of cases. Deletions for at least one 6q MDR gene were detected in 7/8 (88%) asymptomatic patients compared to 13/17 (76%; P=not significant [NS]) symptomatic MYD88^MUT patients. Likewise, no individual gene demonstrated significantly different deletion rates between these groups. Our findings are therefore consistent with previous studies indicating that del6q was indicative of WM, regardless of symptomatic status.118

In MYD88^MUT WM patients, two distinct patterns of del6q were identified. One group was comprised of 8/20 (40%) patients and showed more clonal and contiguous losses spanning across all MDR genes, while a second group (12/20; 60%) had more focal and subclonal genes losses (Figure 1B,C). The mean copy number estimate for deleted genes per patient was significantly lower in the eight patients with contiguous deletions (median 1.02; range: 0.98-1.20) compared with the other del6q patients (median 1.65; range: 1.43-1.75; P=0.0002). No differences in MYD88^MUT clonality were noted between the groups, ruling out normal B-cell contamination differences that might affect these findings. All eight of the clonal patients were CXCR4^WT, while CXCR4 mutations were observed in 11/17 (65%) of the remaining patients (P=0.003). No other significant differences in the clinical features were noted. Contiguous del6q were not observed in the eight MYD88^WTCXCR4^WT patients. Non-contiguous deletions in the MYD88^WT cohort included FOXO3, BCLAF1, TNFAIP3 and HIVEP2 in three (38%), two (25%), two (25%) and one (13%), respectively, while IBTK remained intact (P=0.01 compared to MYD88^MUT WM, Figure 1D).

The number of patients in the MYD88^MUT cohort harboring at least one deleted gene was similar between the CXCR4^MUT (9/11; 82%) and CXCR4^WT (11/14; 79%; P=NS) populations. Because the nature of these deletions differed significantly with clonal contiguous deletions being mutually exclusive of CXCR4 mutations, we used RT-PCR to investigate the effect of del6q and CXCR4^MUT on CXCR4 transcript levels, and observed no differences in expression (data not shown). To investigate further, we performed principal component analysis on previously published RNASeq data from 57 WM patients using 131 genes that are differentially expressed in the presence of del6q.10 This analysis not only stratified patients by del6q, but also by MYD88/CXCR4 genotypes (Figure 2A) indicating that some of these genes are also modulated by MYD88/CXCR4 status. The 20 most influential genes from the rotation matrix are available for the first two component Online Supplementary Table S1. Intersecting the gene lists associated with del6q and CXCR4^MUT revealed 19 overlapping genes, all of which change in the same direction in response to these somatic events (Table 1). As both CXCR4^MUT and del6q are associated with the presence of MYD88^MUT, bootstrapped hierarchical clustering of the 19 genes was conducted on MYD88^MUT RNASeq data (Figure 2B). This generated three distinct clusters that significantly stratified patients based on del6q (P<0.005) and CXCR4^MUT (P<0.001) status. In context with previous studies that supported the acquisition of CXCR4^MUT and del6q after MYD88^MUT,1186 these genes may play a critical role in WM clonal evolution.

Figure 2.Transcriptional impact of chromosome 6q deletions in Waldenström’s Macroglobulinemia. (A) Principal Component Analysis for 131 genes affected by 6q deletions based on next-generation RNA sequencing. Samples were stratified based on 6q deletion status on principal component 1 (PC1), and on MYD88 and CXCR4 genotype on principal component 2 (PC2). (B) Bootstrapped hierarchical clustering of the 19 genes that were similarly impacted by chromosome 6q deletions and CXCR4 mutations in the MYD88^MUT RNASeq data. Approximate unbiased (AU) P-values are shown in red and represent the probability (%) of these samples clustering together under bootstrap simulations of similar populations. The three groups identified by this analysis stratified patients by 6q deletion and CXCR4 mutation status (P<0.001 and P<0.005, respectively). (C) Real-time PCR gene expression results of samples from the MYD88^MUT cohort for each of the 6q groups identified. Relative fold change was calculated based on the median expression value for 6q intact patients. Median values and range are shown for each group. * indicates P-values <0.05 based on the presence of clonal deletions of that gene. WM: Waldenström’s Macroglobulinemia; Chr: chromosome.

Table 1.List of 19 genes found to be dysregulated in MYD88^MUT patients with chromosome 6q deletions or CXCR4 mutations.

Via quantitative RT-PCR, we sought to determine those study genes that were transcriptionally impacted by the presence of CNA in our study cohort. This analysis included all study samples and revealed that clonal deletions of IBTK, BCLAF1 and HIVEP2 significantly reduced the corresponding gene transcriptional levels in the eight clonally 6qdel versus all the other MYD88^MUT patients (P=0.03, P=0.01, P=0.01, respectively; Figure 2C).

Inhibitor of Bruton tyrosine kinase (IBTK) is a negative regulator of BTK, which is located downstream of mutated MYD88 and triggers pro-survival NF-κB signaling in WM.12 The lack of IBTK deletions in MYD88^WT WM was notable as the BTK inhibitor ibrutinib shows poor activity in MYD88^WT WM patients, consistent with the notion that BTK is not essential for tumor survival in this patient population.5 BCLAF1 plays a pro-apoptotic role in the interaction with pro-survival BCL2 protein family members.13 Its decreased expression may contribute to the survival of WM cells, which universally express high levels of BCL2.10 IBTK and BCLAF1 could potentially help delineate those patients who are suited for BTK-inhibitors and BCL-2 antagonist therapies. Moreover, the BCL-2 inhibitor venetoclax has shown significant activity in a phase I trial conducted in relapsed/refractory non-Hodgkin lymphoma patients, including WM,14 and is currently under further investigation in relapsed/refractory WM. HIVEP2, which also showed decreased transcription levels in clonally deleted patients, blocks NF-κB nuclear signaling by binding to NF-κB consensus binding sites.15 Surprisingly, FOXO3 and TNFAIP3, which are important regulators of apoptosis and NF-κB signaling, respectively, were not impacted transcriptionally. Therefore, IBTK, BCLAF1 and HIVEP2 may serve as particularly important determinants of disease progression. However, the limited number of patients enrolled in the study herein precluded any investigation into the prognostic or predictive role for the 6q MDR genes examined, and further efforts into clarifying their clinical significance are warranted.

In MYD88 mutated patients, fully clonal 6qdel and CXCR4 mutation status showed mutual exclusivity, suggesting shared roles for the two genomic events. CXCR4^MUT was previously shown by us to downregulate tumor suppressors that are transcribed in response to mutated MYD88.10 Indeed, herein we identified 19 genes co-regulated by 6qdel and CXCR4 mutation status, which may be involved in WM clonal evolution. In summary, our findings provide new insights into WM pathogenesis, including loss of key regulators of BTK, apoptosis, BCL2 and NF-κB signaling in asymptomatic and symptomatic WM patients, and shared regulatory signaling for MYD88^MUT WM patients with either 6qdel or CXCR4^MUT disease.