Discovery of an undocumented pattern of intron retention that is abundant in healthy B-cells and circumvented in Waldenstrom's macroglobulinemia impacting genes driving cellular growth and survival

Abstract

Abstract Background: We previously developed a quasi-Poisson model to predict background coverage of intron segments with deviations from predicted results used to identify novel differentially retained introns. In addition to somatic retention events, this approach documented novel retentions in healthy donor (HD) memory B-cells (MB; CD19+CD27+). Many of these were identified in genes associated with Waldenstrom Macroglobulinemia (WM) and in the cases of BTK and NFKB2, appeared to account for close to 50% of transcription in HD MB and were largely absent in WM. Mechanistic insights of translational control could offer novel opportunities for therapy by restoring intron retentions and silencing of genes important to WM pathogenesis. Methods We performed transcriptome analysis on primary bone marrow (BM) CD19+ lymphoplasmacytic cells (LPCs) from 249 treatment-naïve MYD88 mutated patients, along with CD19+CD27+ MB and peripheral blood (PB; CD19+) cells from 13 HDs. Reads were realigned to Gencode augmented with novel isoforms identified using long read IsoSeq analysis of HDMB, HD plasma cells (CD138+), and primary BM CD19+ WM LPCs. HD CD19+ and CD19+CD27+ samples were obtained using gradient centrifugation and magnetic bead purification. Retentions were validated by PCR gel analysis and Sanger sequencing. Results Transcriptome analysis confirmed the predicted intron retentions in HD MB and PB samples that were absent in primary LPCs from WM patients. Retentions in genes relevant to WM growth and survival were further validated by PCR including NFKB2, IRAK4, BTK, CD19, CDK16 and FGR. The two most impacted genes were BTK and NFKB2. For BTK, the canonical transcript (ENST00000308731) constituted a median of 4.2% (range 0-53%) of BTK transcripts while it represented 72.2% (range 32-96.3%; p<0.0001) in WM. Likewise, retained isoforms collectively made up 40.5% (32.9-53%) of HD transcripts and 16.5% (0-46.5%; p<0.0001) in WM. Established protein coding isoforms of NFKB2 (ENST00000601386, ENST00000428099, ENST00000369966, ENST00000652277, and ENST00000189444) made up a median 3.5% (range 0-17.9%) and 27.1% (0 – 58.5%) for HD and WM, respectively (p<0.0001). NFKB2 isoforms with intronic retentions made up 75.5% (49.7-83.1%) and 37.9% (2.3-81%) of HD and WM samples, respectively (p<0.0001). Total transcription of both genes was higher in HD samples with median transcripts per million (TpM) of 144 (range 71.3-242) TpM and 95.2 (8.6-206) TpM for BTK and 74.7 (52.8-90.8) TpM and 32.4 (4.31-73.4) TpM for NFKB2 in HD and WM samples, respectively (p<0.0001 for both). NFKB2 retentions were time dependent and disappeared within two hours of cell selection with a corresponding increase in spliced transcript abundance. This was observed in PB and leukopak samples and in CD19+CD27-, CD19+CD27+ and unselected populations. It was also independent of time spent in incubation, though cellular stress following gradient centrifugation and/or exposure to red cell lysis buffer triggered intron removal within 2 hours. Nuclear extraction confirmed that spliced NFKB2 was in the cytoplasm while the retained isoforms were present equally in both compartments. To test if RNA degradation could explain the disappearance of these retentions within two-hours, cells were exposed to the transcriptional and splicing inhibitor actinomycin D. We observed the NFKB2 retentions were quite stable relative to the SGK1 quick decay control and were still present at the 3-hour mark long after they had disappeared in our DMSO controls. As poly-A selection was used in both PCR and RNASeq validations, these transcripts are likely polyadenylated. Given that many of the retained NFKB2 isoforms can be converted to coding isoforms, it suggests that retained forms are mass produced and spliced to coding isoforms if needed.Conclusions: Undocumented intron retentions that impact translation of genes important to WM growth and survival including NFKB2 and BTK are highly prevalent in HD B-cells but absent in WM LPCs. Our studies show that NFKB2, a key mediator of non-canonical NFKB pro-survival signaling, is primarily produced with retained introns that prevent their translation and are post transcriptionally spliced in the cytoplasm of HD B-cells. Our findings provide novel insights into the role of splicing and translation of key pro-survival proteins in WM pathogenesis.