Richter Syndrome-Associated Genetic Lesions in TP53 and the Cell Cycle Inhibitors CDKN2A and CDKN2B Induce Transformation of a CD5+ B Cell Subset Characterized By a Restricted IGHV Repertoire, Autoreactivity and BCR Dependence

The Richter Syndrome (RS) genomic landscape is characterized by frequent loss-of-function (LOF) genetic lesions in negative cell-cycle regulators, such as the cyclin dependent kinase (CDK) inhibitors CDKN2A and CDKN2B, which are deleted in 40-50% of RS tumors. These genetic lesions usually co-occur with mutations/deletions in TP53, which also negatively regulates the cell cycle by inducing the CDK inhibitor CDKN1A. We previously showed that in CLL cells, B cell receptor (BCR) stimulation leads to increased expression of these three CDK inhibitors, resulting in a block of cell cycle progression at the G1/S phase. In addition, we showed that biallelic LOF mutations in TP53, CDKN2A and CDKN2B can overcome this block and allow for spontaneous proliferation of autoreactive murine Eμ-TCL1-derived CLL cells, suggesting a potential interplay between these genetic lesions and BCR signaling in the pathogenesis of RS.

To further understand the role of TP53, CDKN2A and CDKN2B abnormalities in RS, in the current study we used multiplex CRISPR/Cas9 editing to investigate whether the combination of these genetic lesions can induce transformation of nonleukemic B cells from healthy C57BL/6 mice or mice genetically predisposed to develop CLL (mice with deletion of 13q14-MDR). The TP53/CDKN2A/CDKN2B edited cells were generated by transfecting splenic or peritoneal cavity B cells from 2-3 month old C57BL/6 or del(13q14-MDR) mice with recombinant Cas9 and a pool of guide RNAs targeting TP53, CDKN2A and CDKN2B (n=8 separate experiments per group). The edited cells were transplanted in immunodeficient NSG mice and monitored for leukemia development by peripheral blood flow cytometry analysis. Remarkably, 14 of the 16 recipient mice developed a CD5+CD19+CD20+IgM+ leukemia within 3 to 6 months after transplantation [median time 3.5 months for the del(13q14-MDR) group and 4.5 months for the C57BL/6 group, P = 0.023]. The leukemic cells were serially propagated in both NSG and immunocompetent syngeneic mice, with the latter showing a delayed leukemia engraftment. Histological analysis of spleens obtained at euthanasia showed the presence of sheets of blast cells, consistent with Richter transformation, in 7 of the 10 NSG and 2 of the 7 immunocompetent recipients. IGHV sequencing analysis of the propagated leukemias revealed monoclonal or biclonal leukemic cell populations that expressed, in most cases, stereotyped BCRs encoded by the IGHV1-55 or IGHV11-2 gene (13 out of 20 IGHV sequences), with all of the IGHV11-2-encoded BCRs displaying reactivity with the autoantigen phosphatidylcholine (n=9). To investigate the BCR dependency of the malignant B cells, we performed adoptive transfer experiments (n=5) with mixtures of IGHM-wild type and CRISPR/Cas9-generated IGHM-knockout cells (50-77% sIgM-negative cells injected per experiment), and observed complete depletion of the sIgM-negative cells after 2 weeks of in vivo propagation.

To investigate whether multiplex CRISPR/Cas9 editing of other CLL driver genes that are frequently altered in RS tumors will result in transformation of the same B cell subsets, we simultaneously introduced LOF mutations in ATM, POT1, RB1, FBXW7, NFKBIE, BIRC3, TRAIL-R, MGA, CHD2, SETD2, ARID1A, and SAMHD1, and gain-of-function mutations in NOTCH1 into nonleukemic B cells from young del(13q14) mice (n=9). TP53 was also targeted in these experiments because it is altered in most RS tumors. In all cases the recipient mice developed a CD5+CD19+CD20+IgM+ leukemia within 4-8 months of transplantation. Mutations in TP53, NFKBIE and either NOTCH1 or FBXW7 were present in all leukemias, whereas mutations in the other targeted genes occurred at a lower frequency. Histological analysis showed features consistent with RS in 5/5 investigated spleens. Adoptive transfer experiments with IGHM-wild type and IGHM-knockout cells showed that these leukemias are also mostly BCR dependent (4 of 5 investigated cases) but differ from the leukemias generated by TP53/CDKN2A/CDKN2B editing in expressing a diverse IGHV gene repertoire (none expressed the IGHV1-55 or IGHV11-2 gene) and displaying no reactivity with phosphatidylcholine. Collectively, these data show that combined genetic lesions in TP53, CDKN2A and CDKN2B preferentially transform a subset of B cells with particular BCR features and further suggest a role for BCR signaling in the pathogenesis of RS.

No relevant conflicts of interest to declare.