In vivo CRISPR screening reveals cooperation of KMT2D and TP53 deficiencies in B-cell lymphomagenesis

Abstract

Although recent genetic studies have identified numerous genetic alterations in diffuse large B-cell lymphoma (DLBCL), their biological relevance remains elusive. Here, we performed in vivo CRISPR loss-of-function screening targeting 86 genes recurrently altered in DLBCL to examine oncogenicity of single-guide RNA (sgRNA)-targeted genes, association between genotype and lineage, occurrence of second-hit alterations, and cooperability among sgRNA-targeted genes and second-hit alterations. Transplantation of the CRISPR library-transduced hematopoietic stem/progenitor cells induces various hematologic malignancies, including B-cell lymphomas in mice. Enrichment analysis of sgRNA-targeted genes demonstrates significant overrepresentation of Kmt2d, Pax5, and Trp53 in B-cell lymphomas. Whole-exome sequencing identifies recurrent second-hit driver alterations, showing significant enrichment of Trp53 alterations in sgKmt2d-targeted B-cell lymphomas. Importantly, KMT2D and TP53 mutations are found to be the most prevalent co-occurring combination in human DLBCL, which is more prominent in relapsed/refractory DLBCL. Moreover, this combination confers significantly worse prognosis independent of clinical factors. Transcriptomic sequencing identifies overexpression of Yap1, the Hippo pathway component, in double sgKmt2d-targeted/Trp53-altered B-cell lymphomas. Furthermore, chromatin accessibility analysis demonstrates enrichment of transcriptional enhanced associate domain 1 binding motifs in regions that gained accessibility and increased expression of their nearest genes in these B-cell lymphomas. Most importantly, genetic and pharmacological inhibition of YAP1 suppresses in vitro cell proliferation and in vivo tumor growth of a human KMT2D/TP53-altered DLBCL cell line and prolongs survival of mice transplanted with double sgKmt2d-targeted/Trp53-altered B-cell lymphoma cells. Our findings demonstrate the utility of in vivo CRISPR screening to integrate human cancer genomics with mouse modeling and highlight the functional interplay between KMT2D and TP53 aberrations, providing insights into therapeutic strategies in DLBCL.

Graphical abstract

Figure 1.

CRISPR loss-of-function screening of recurrently mutated genes in DLBCL. (A) Schematic representation of CRISPR loss-of-function screening targeting 86 recurrently mutated genes in DLBCL. (B) Overlap of sgRNA-targeted genes with recurrently altered TSGs in B-ALL/LBL, and PTCL, NOS. Because of interspecies differences, 86 murine genes in the CRISPR library correspond to 84 orthologous human genes. (C) Survival curves of recipient mice transplanted with HSPCs transduced with mock plasmid (n = 6) or the CRISPR library (n = 199). Log-rank test. (D) Distribution of hematologic malignancies in recipient mice (n = 132) according to the Bethesda proposals. (E) Landscape of significantly recurrent (q < 0.01) sgRNA-mediated disruption in the entire cohort (n = 104). Disease type, involved organ, and CRISPR library pool (bottom) as well as q values (right) are shown. (F-K) Number of samples harboring sgRNAs for each gene in the entire cohort (F; n = 104), B-cell lymphoma (G; n = 20), B-ALL/LBL (H; n = 19), T-cell lymphoma (I; n = 24), T-ALL/LBL (J; n = 15), and AML (K; n = 20). Significantly recurrent genes (q < 0.01) are colored and marked with asterisks. AML, acute myeloid leukemia; AUL, acute undifferentiated leukemia; NGS, next-generation sequencing; NK-lymphoma, natural killer-cell lymphoma; PTCL NOS, peripheral T-cell lymphoma, not otherwise specified; TSGs, tumor suppressor genes.

Figure 2.

Second-hit driver alterations revealed by WES in CRISPR screening. (A) Landscape of significantly recurrent second-hit somatic alterations in the entire cohort. Disease type, involved organ, and CRISPR library pool (bottom) as well as q values (right) are shown. Significantly recurrent genes (q < 1 × 10⁻¹⁰) are marked with asterisks. (B) Number of samples harboring second-hit alterations for each gene in the entire cohort (n = 104). (C) Type and position of Trp53 second-hit alterations in the entire cohort. (D) Proportion of copy number–altered segments in samples with and without Trp53 alterations in B-cell lymphoma (n = 19) and T-ALL/LBL (n = 15). Brunner-Munzel test. (E) Stat5a/Stat5b amplification in a B-cell lymphoma case (SCR208). (F) Nup214::Abl1 fusion in a B-ALL/LBL case (SCR009). aa, amino acids; AUL, acute undifferentiated leukemia; ITD, internal tandem dupulication; NES, nuclear export signal; NLS, nuclear localization signal; SNV, single-nucleotide variant; SV, structural variation.

Figure 3.

Associations between second-hit alterations and disease type in CRISPR screening. (A) Proportion of samples harboring second-hit alterations for each gene according to sgRNA-targeted genes and disease type. Significantly recurrent genes (q < 1 × 10⁻¹⁰) are colored and marked with asterisks. (B) Type and position of second-hit alterations in Jak3, Jak1, Sh2b3, Flt3, Ptpn11, and Kit in the entire cohort (n = 104). (C) Proportion of samples harboring second-hit alterations involving FLT3, JAK/STAT, PI3K, and RAS pathway molecules in murine sgPax5-targeted B-cell malignancies (n = 26) and others (n = 78). Fisher exact test. (D) Proportion of samples harboring somatic alterations (mutations and structural variations) involving FLT3, JAK/STAT, PI3K, and RAS pathway molecules in human PAX5-altered subtypes of B-ALL/LBL (n = 85) and others (n = 738). Fisher exact test. aa, amino acids; FLT3, FMS-like tyrosine kinase 3; PI3K, phosphatidylinositol-3 kinase.

Figure 3.

Associations between second-hit alterations and disease type in CRISPR screening. (A) Proportion of samples harboring second-hit alterations for each gene according to sgRNA-targeted genes and disease type. Significantly recurrent genes (q < 1 × 10⁻¹⁰) are colored and marked with asterisks. (B) Type and position of second-hit alterations in Jak3, Jak1, Sh2b3, Flt3, Ptpn11, and Kit in the entire cohort (n = 104). (C) Proportion of samples harboring second-hit alterations involving FLT3, JAK/STAT, PI3K, and RAS pathway molecules in murine sgPax5-targeted B-cell malignancies (n = 26) and others (n = 78). Fisher exact test. (D) Proportion of samples harboring somatic alterations (mutations and structural variations) involving FLT3, JAK/STAT, PI3K, and RAS pathway molecules in human PAX5-altered subtypes of B-ALL/LBL (n = 85) and others (n = 738). Fisher exact test. aa, amino acids; FLT3, FMS-like tyrosine kinase 3; PI3K, phosphatidylinositol-3 kinase.

Figure 4.

Co-occurring KMT2D and TP53 mutations are prevalent and adversely prognostic in human DLBCL. (A) Proportion of Trp53 alterations in murine sgKmt2d-targeted B-cell lymphoma (n = 12) and others (n = 92). Fisher exact test. (B) Number of co-occurring combinations of driver alterations in 489 untreated patients with DLBCL from National Cancer Institute's Center for Cancer Research (NCICCR) cohort. (C) HRs with 95% CIs for overall survival evaluated by Cox proportional hazards model incorporating presence of KMT2D, TP53, or both mutations, IPI, and genetic subtype according to the LymphGen classification in 230 untreated patients with DLBCL from NCICCR cohort. (D) Proportion of KMT2D, TP53, or both mutations in 489 untreated and 117 patients with relapsed/refractory DLBCL. (E) Number of co-occurring combinations of driver mutations in 117 patients with relapsed/refractory DLBCL. In panels B,E, genetic subtype related to each altered gene is shown according to the LymphGen classification. Bars are colored by odds ratios calculated by Fisher exact test. Aberrant somatic hypermutation target genes are removed. CI, confidence interval; HR, hazard ratio; IPI, International Prognostic Index; re, rearrangement.

Figure 5.

Kmt2d and Trp53 deficiencies coordinately induce B-cell lymphoma with Yap1 overexpression. (A) Survival curves of recipient mice transplanted with mock (n = 7), sgKmt2d (n = 27), sgTrp53 (n = 9), and double-targeted (n = 9) HSPCs with constitutive Cas9 expression. Log-rank test. (B) Survival curves of primary (n = 9) and secondary (n = 17) recipient mice transplanted with double sgKmt2d/sgTrp53-targeted HSPCs with constitutive Cas9 expression. Log-rank test. (C) Survival curves of recipient mice transplanted with mock (n = 6), sgKmt2d (n = 6), sgTrp53 (n = 6), and double-targeted (n = 6) HSPCs with GC B-cell–specific Cas9 expression. Log-rank test. (D) Volcano plot showing differentially expressed genes between single sgKmt2d-targeted (n = 7) and double sgKmt2d-targeted/Trp53-altered (n = 17) B-cell lymphoma. Genes with q value of <0.2 and |log₂(fold change)| of >2 are considered significant and colored red (gained) or blue (lost). (E) Yap1 mRNA expression measured by RNA-seq in samples with (n = 17) or without (n = 7) Trp53 alterations in sgKmt2d-targeted B-cell lymphoma. Brunner-Munzel test. (F) YAP1 and pYAP1 protein expression measured by capillary electrophoresis-based immunoassay in representative samples with or without Trp53 alterations in sgKmt2d-targeted B-cell lymphoma. Lamin B1 is used as a loading control. Representative of 3 independent experiments. ∗P < .05; ∗∗P < .005; ∗∗∗P < .0005. mRNA, messenger RNA; pYAP1, phosphorylated YAP1; TPM, transcripts per million.

Figure 6.

Epigenetic changes in Kmt2d- and Trp53-altered B-cell lymphoma. (A) Principal component analysis plot of ATAC-seq data from normal splenic B cells from wild-type mice (n = 6) as well as tumor cells from single sgKmt2d-targeted (n = 7), single Trp53-altered (n = 1), and double sgKmt2d-targeted/Trp53-altered (n = 14) B-cell lymphomas and those without Kmt2d and Trp53 alterations (n = 7). (B) Volcano plot showing DARs between single sgKmt2d-targeted (n = 7) and double sgKmt2d-targeted/Trp53-altered (n = 14) B-cell lymphoma. Merged ATAC-seq peaks with q value of <0.2 and |fold change| of >1.5 are considered significant and colored red (gained) or blue (lost). (C) Heat map showing DARs between single sgKmt2d-targeted and double sgKmt2d/Trp53-altered B-cell lymphoma. Unsupervised hierarchical clustering was performed with Pearson correlation and Ward.D2 linkage algorithm. (D) Mean normalized ATAC-seq signal intensity for regions that gained accessibility in double sgKmt2d-targeted/Trp53-altered B-cell lymphoma compared with single sgKmt2d-targeted B-cell lymphoma. (E) Cumulative frequency of expression for genes nearest to regions that gained or lost accessibility or remained stable in double sgKmt2d-targeted/Trp53-altered B-cell lymphoma compared with single sgKmt2d-targeted B-cell lymphoma. Kolmogorov-Smirnov test. (F) Motif enrichment analysis for distal regions (−1000 base pairs [bp] to +100 bp of TSS) that gained or lost accessibility in double sgKmt2d-targeted/Trp53-altered B-cell lymphoma compared with single sgKmt2d-targeted B-cell lymphoma. (G) Gene set enrichment analysis of expression data comparing single sgKmt2d-targeted and double sgKmt2d-targeted/Trp53-altered B-cell lymphoma using genes nearest to distal regions that gained accessibility and contained the TEAD1 motif. ATAC-seq, assay for transposase-accessible chromatin with sequencing; DAR, differentially accessible region; EBF, early B cell factor; ES, enrichment score; Ets-1, ETS proto-oncogene 1; FDR, false discovery rate; Hoxa9, homeobox A9; mRNA, messenger RNA; NFIA, nuclear factor I A; Oct6, organic cation/carnitine transporter 6; PC1/2, principal component 1/2; Pft1a, pancreas associated transcription factor 1a; PGR, progesterone receptor; RUNX, runt-related transcription factor; TSS, transcriptional start site.

Figure 7.

Yap1 overexpression and its targeting in Kmt2d- and Trp53-mutated B-cell lymphoma. (A) YAP1 mRNA expression measured by real-time quantitative polymerase chain reaction in 6 human DLBCL cell lines (n = 3 for each) with and without KMT2D and/or TP53 mutations. (B) YAP1 and pYAP1 protein expression measured by capillary electrophoresis-based immunoassay in 6 human DLBCL cell lines. Lamin B1 is used as a loading control. Representative of 3 independent experiments. (C) Effect of CA3 treatment on viability of 6 human DLBCL cell lines (n = 3 for each) measured by CellTiter-Glo luminescent cell viability assay. (D) Cell proliferation of Toledo cells expressing short hairpin RNA (shRNA) targeting GFP (control) or YAP1 after puromycin selection. (E) YAP1 protein expression measured by capillary electrophoresis-based immunoassay in Toledo cells expressing shRNA targeting GFP or YAP1. Lamin B1 is used as a loading control. Representative of 2 independent experiments. (F) Tumor volume in Toledo xenograft mice administered with 4 mg/kg of CA3 (n = 7) or vehicle (n = 7) 3 times a week for 12 doses. (G) Survival curves of mice transplanted with 2.5 × 10³ murine double sgKmt2d-targeted/Trp53-altered B-cell lymphoma cells and administered with 4 mg/kg of CA3 (n = 16) or vehicle (n = 14) 3 times a week for 6 doses. Log-rank test. In panels C-D,F, data are represented as mean ± standard deviation. In panels A,D,F, Welch t test was performed. ∗P < .05; ∗∗P < .005; ∗∗∗P < .0005.