Genomic determinants and differences of primary and secondary plasma cell leukemia Free

Introduction Plasma cell leukemia (PCL) is a rare, aggressive form of multiple myeloma (MM) defined by the presence of >5% circulating plasma cells (PCs) in the peripheral blood (PB). PCL can present de novo, known as primary PCL (pPCL), or emerge at relapse, referred to as secondary PCL (sPCL). Approximately 50% of patients with pPCL harbor the t(11;14) translocation, allowing classification into t(11;14) and non-t(11;14) pPCL phenotypes. The prognosis is poor, with a median overall survival <3 years, despite the use of immune effector therapies. Understanding the molecular alterations in PCL is key to exploring novel therapeutic approaches, but it remains to be well-characterized. Here, we determine tumor and non-tumor differences in biologic pathways and the genetic landscape in patients with pPCL and sPCL to understand treatment resistance and relapse, compared to matched MM patients without a leukemic component.

Methods We included bone marrow (BM) aspirates from 11 PCL patients at diagnosis, including pPCL, sPCL, and paired PB samples where available. We performed unbiased single-cell mRNA profiling (5' sequencing) of BM and PB mononuclear cells using 10x Genomics on the CD138+ and CD138- sorted non-tumor cells separately. We processed data via the Cell Ranger pipeline, which grouped T cells into clonotypes based on shared TCR α/β sequences. Further data processing was done with Seurat, scRepertoire and Immunarch. In total, 228,579 cells were sequenced from 28 samples, including 98,401 PC and 130,178 microenvironment non-tumor cells. Of these, 43,043 cells originated from the PB and 185,536 from the BM.

Results The median age was 52 years (range 23 - 81 years). Of 11 patients, 7 (64%) had pPCL and 4 (36%) sPCL. Median PC involvement was 80% BM (range 2-96%) and 15% PB (6-90%). Four pPCL and one sPCL had t(11;14), respectively. None of the patients had true extramedullary disease.

Paired PB and BM pPCL (n=6: 3 PB, 3 BM) comparison revealed distinct phenotypes and transcriptomic profiles within the same patient. BM PCs favored JAK STAT, MAPK, and phosphatidylinositol signaling, while PB PCs utilized oxidative phosphorylation, DNA repair (non-homologous joining, base excision and nucleotide excision, and replication), RNA processing, and proteosome pathways in Gene Set Enrichment Analysis (GSEA). CD38 and TNFRSF17 expression differed significantly, and was higher in PB clones versus the BM compartments (p < 1e-5), with no significant segregation by IgH translocation subtypes.

Transcriptomic analysis of randomly selected cells (n=500) and bootstrapped comparisons between pPCL and sPCL revealed both shared and distinct transcriptomic programs (TP). Three overlapping TPs (TP1, TP3, TP5) and one unique program each, TP4 in pPCL and TP2 in sPCL, were noted. Some shared programs had a differential contribution to each phenotype. TP4 (pPCL) featured upregulated differential gene expressions with at least a twofold change (FDR<0.05) for MALAT1, NEAT1, JUND, BTG1, H1FX, H3F3B, NFKBIA, and downregulated MZB1, PARK7, PPIA, PPIB, UBA52, TXN, and UQCRQ genes. TP2 (sPCL) showed upregulation of IL32, GZMB, FGFBP2, KLRF1, GAGE12H, PRF1, TXN, and downregulation of MALAT1, JUND, TRAM1, JCHAIN, MEST, PBX1, and PCLAF. These could explain differences in transcriptional regulation, NF-κB-mediated inflammatory survival signaling, antigen processing, protein isomerization and chaperoning, and ubiquitin-proteasome dynamics. PCs in sPCL exhibited a stemness-like phenotype (CD19⁺, CD33⁺, IL32⁺, IgM⁺), and a higher CD38 and TNFRSF17R expression as compared to pPCL PCs (adjusted p-value: <1e-05), supporting a more immature and aggressive phenotype.

In the tumor microenvironment, sPCL showed stronger inferred cell-cell interaction strength than pPCL. sPCL had increased MIF signaling toward PC & elevated MHC-I signaling away from malignant PCs. CypA and PECAM1/2 signaling were similar between groups. CD45 signaling toward B-cells was also higher in sPCL.

Conclusion PB and BM PCs in PCL are transcriptionally distinct, with BM cells enriched for MAPK/JAK-STAT signaling and PB cells for proteasome, DNA repair, and RNA processing pathways. sPCL displays greater tumor microenvironment interactions overall suggesting targetable signaling. sPCL shares stem-like features with early B cell markers and higher druggable target expression despite refractoriness, hence providing a framework for investigating newer therapies based on tumor evolution.