Synthetic Lethal and Convergent Biological Effects of Cancer-Associated Spliceosomal Gene Mutations

Abstract

Mutations affecting RNA splicing factors are the most common genetic alterations in myelodysplastic syndrome (MDS) patients and occur in a mutually exclusive manner. The basis for the mutual exclusivity of these mutations and how they contribute to MDS is not well understood. Here we report that although different spliceosome gene mutations impart distinct effects on splicing, they are negatively selected for when co-expressed due to aberrant splicing and downregulation of regulators of hematopoietic stem cell survival and quiescence. In addition to this synthetic lethal interaction, mutations in the splicing factors SF3B1 and SRSF2 share convergent effects on aberrant splicing of mRNAs that promote nuclear factor κB signaling. These data identify shared consequences of splicing-factor mutations and the basis for their mutual exclusivity.

Keywords: NF-κB; SF3B1; SRSF2; U2AF1; myelodysplastic syndromes; splicing.

Figure 1.. Simultaneous Expression of Mutations in Srsf2 and Sf3b1 or Expression in the Homozygous State Is Incompatible with Hematopoiesis

(A) Heat map of the four most commonly mutated gene sencoding RNA splicing factors across 11 studies in myeloid malignancies (Bejar et al., 2012; Damm et al., 2012; Haferlach et al., 2014; Lasho et al., 2012; Makishima et al., 2012; Meggendorfer et al., 2012; Papaemmanuil et al., 2013; Patnaik et al., 2013; Thol et al., 2012; Yoshida et al., 2011; Zhang et al., 2012). Each column represents a patient, and each colored bar represents the presence of the specified mutation. (B) Schema of competitive and noncompetitive bone marrow transplantation (BMT) using bone marrow mononuclear cells (BM MNCs) from 8-week-old Mx1-Cre⁺ wild-type (WT), Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+, Mx1-Cre⁺ Srsf2^P95H/P95H, and Mx1-Cre⁺ Srsf2^P95H/fl mice. Polyinosinic-polycytidylic acid (pIpC) was administered to recipients 4 weeks post BMT to induce expression of mutant alleles. (C) Percentage of CD45.2 chimerism in peripheral blood of recipients (n = 8–10 mice per genotype) in noncompetitive BMT. (D) Representative fluorescence-activated cell sorting (FACS) plots of CD45.2⁺ cells in peripheral blood of recipients in noncompetitive BMT 52 weeks post pIpC. (E) Percentage of CD45.2 chimerism in peripheral blood of recipients (n = 10 mice per genotype) in competitive BMT. (F) Analysis of CD45.2 chimerism in the BM, spleen, thymus, and blood of recipients (n = 5–10 mice per genotype) in competitive BMT 20 weeks post pIpC. (G) Percentage of CD45.2 chimerism in peripheral blood of recipients (n = 5–10 mice per genotype) in competitive BMT from Mx1-Cre⁺ WT, Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Srsf2^P95H/P95H, and Mx1-Cre⁺ Srsf2^P95H/fl mice. (H) Analysis of CD45.2 chimerism in BM, spleen, thymus, and blood of recipients (n = 5–10 mice per genotype) in competitive BMT 20 weeks post pIpC. Error bars represent mean ± SD. ANOVA and Tukey’s post hoc test were used to compare groups. *p < 0.05, **p < 0.01, ***p < 0.001 versus Mx1-Cre⁺ WT mice; ^{^}p < 0.05, ^{^^}p < 0.01, ^{^^^}p < 0.001 versus Mx1-Cre⁺ Sf3b1^K700E/+ mice; ^##p < 0.01, ^###p < 0.001 versus Mx1-Cre⁺ Srsf2^P95H/+ mice. See also Figure S1 and Table S1.

Figure 2.. Combined Expression of Mutations in Srsf2 and Sf3b1 Results in Hematopoietic Stem and Progenitor Cell Apoptosis and Loss of Quiescence

(A and B) Percentage of bromodeoxyuridine⁺ (BrdU⁺) (A) or annexin-V⁺ propidium iodide⁻ (PI⁻) (B) LSK cells from Mx1-Cre⁺ WT (n = 4), Mx1-Cre⁺ Srsf2^P95H/+ (n = 5), Mx1-Cre⁺ Sf3b1^K700E/+ (n = 4), and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ (n = 4) mice 2 weeks post pIpC administration. (C) Number of live mice at weaning from crossing Sf3b1^K700E/+ mice to Vav-Cre⁺ Srsf2^P95H/+ mice or by crossing Srsf2^P95H/+ mice to Vav-Cre⁺ Sf3b1^K700E/+ mice. **p = 0.0064; two-sided Chi-square test. (D and E) Percentage of LSK (D) and long-term hematopoietic stem cells (LT-HSC; LSK CD150⁺ CD48⁻) (E) in E14.5 fetal livers from Vav-Cre⁺ WT (WT; n = 17), Vav-Cre⁺ Srsf2^P95H/+ (P95H; n = 17), Vav-Cre⁺ Sf3b1^K700E/+ (K700E; n = 15), and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ double-knockin (DKI; n = 16) fetuses and Vav-Cre⁻ WT (n = 14), Vav-Cre⁻ Srsf2^P95H/+ (P95H; n = 17), Vav-Cre⁻ Sf3b1^K700E/+ (K700E; n = 19), and Vav-Cre⁻ Srsf2^P95H/+ Sf3b1^K700E/+ DKI (n = 17) fetuses. (F) Colony numbers from E14.5 fetal liver cells from Vav-Cre⁺ WT (n= 12), Vav-Cre⁺ Srsf2^P95H/+ (P95H; n = 7), Vav-Cre⁺ Sf3b1^K700E/+ (K700E; n = 12), and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ DKI (n = 7) fetuses and from Vav-Cre⁻ WT (n = 10), Vav-Cre⁻ Srsf2^P95H/+ (P95H; n = 10), Vav-Cre⁻ Sf3b1^K700E/+ (K700E; n = 11), and Vav-Cre⁻ Srsf2^P95H/+ Sf3b1^K700E/+ DKI (n = 13) fetuses. (G–I) Representative FACS plots (G) and quantitation of BrdU⁺ (H) and annexin-V⁺ PI⁻ (I) LSK cells from Vav-Cre⁺ WT (n = 5), Vav-Cre⁺ Srsf2^P95H/+ (P95H; n = 8), Vav-Cre⁺ Sf3b1^K700E/+ (K700E; n = 3), and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ DKI (n = 9) fetuses and from Vav-Cre⁻ WT (n = 4), Vav-Cre⁻ Srsf2^P95H/+ (P95H; n = 6), Vav-Cre⁻ Sf3b1^K700E/+ (K700E; n = 7), and Vav-Cre⁻ Srsf2^P95H/+ Sf3b1^K700E/+ DKI (n = 2) fetuses. Error bars represent mean ± SD. ANOVA and Tukey’s post hoc test was used to compare groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See also Figure S2.

Figure 3.. Srsf2 and Sf3b1 Mutations Have Distinct and Independent Effects on Gene Dysregulation

(A) Schema for RNA-seq. (B) Expression of Srsf2^P95H and Sf3b1^K700E alleles as percentage of mRNAs expressed from Srsf2 and Sf3b1. Color indicates genotype; the three biological replicates are A to C from left to right. (C) Scatterplots comparing coding gene expression in Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre+WT cells for replicate B. Red and blue indicate coding genes significantly up- or down regulated, respectively, in mutant relative to Mx1-Cre⁺WT cells. TPM, transcripts per million (TMM-normalized). (D) Bar plots comparing percentage of significantly dysregulated coding genes (percentage of total coding genes expressed) in Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre⁺ WT cells for replicate B. Error bars represent mean ± SD. A two-sided binomial proportion test was used to compare groups; ***p < 0.0001. (E) Venn diagram showing overlap between coding genes significantly dysregulated in Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre⁺ WT cells for replicate B. (F) GO enrichment analysis of Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre⁺ WT cells for replicate B. Circle size indicates the magnitude of the p value for each term and comparison. See also Figure S3; Tables S2 and S3.

Figure 4.. Srsf2 and Sf3b1 Mutations Have Distinct and Independent Effects on RNA Splicing

(A) Scatterplots of cassette exon inclusion in Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre⁺ WT cells. Axes indicate the fraction of mRNAs containing each cassette exon in the indicated sample. Red and blue indicate cassette exons whose inclusion is promoted or repressed, respectively, in mutant relative to WT cells. (B) As in (A), but for alternative 3′ splice site events. Axes indicate the fraction of mRNAs that use the intron-proximal 3′ splice site in the indicated sample. Red and blue indicate intron-proximal 3′ splice sites whose usage is promoted or repressed, respectively, in mutant relative to WT cells. (C) Plots illustrating the spatial distribution of the CCNG and GGNG (N = any nucleotide) exonic splicing enhancers adjacent to differentially spliced cassette exons promoted versus repressed in mutant relative to WT cells. Vertical axis indicates the relative frequency of each motif, averaged over all promoted versus repressed cassette exons for the indicated genotype comparisons. (D) Expression of Srsf2^P95H and Sf3b1^K700E alleles as percentage of mRNAs expressed from Srsf2 and Sf3b1 in lineage^- c-Kit⁺ cells from fetal livers of Vav-Cre mice at E14.5. Color indicates genotype; the three biological replicates are A to C from left to right. (E) Scatterplots comparing mean coding gene expression for all replicates in Vav-Cre⁺ Srsf2^P95H/+, Vav-Cre⁺ Sf3b1^K700E/+, and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Vav-Cre⁺ WT cells. Red and blue indicate coding genes significantly up- or down regulated, respectively, in mutant relative to Vav-Cre⁺ WT cells. TPM, transcripts per million (TMM-normalized). (F) Venn diagram showing the overlap between coding genes significantly dysregulated in Vav-Cre⁺ Srsf2^P95H/+, Vav-Cre⁺ Sf3b1^K700E/+, and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Vav-Cre⁺ WT cells in all replicates. See also Figure S4; Tables S4 and S5.

Figure 5.. Co-expression of Srsf2 and Sf3b1 Mutations Results in Aberrant Splicing and Expression of Key Regulators of Hematopoietic Stem Cell Survival and Quiescence

(A) Venn diagram of genes differentially expressed and spliced in Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E+ cells relative to Mx1-Cre⁺ WT cells in any replicate. (B) Expression of Mpl, Itga2b, and Pbx1 in lineage^- c-Kit⁺ (LK) cells from Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ mice relative to control groups (n = 8–10 mice per genotype). Error bars represent mean ± SD. ***p < 0.001, ****p < 0.0001 versus Mx1-Cre⁺ WT; ^####p < 0.0001 versus Mx1-Cre⁺ Srsf2^P95H/+; ^AAp < 0.01, ^AAAp < 0.001, ^AAAAp < 0.0001 versus Mx1-Cre⁺ Sf3b1^K700E/+. (C) GO enrichment analysis of Mx1-Cre⁺ Srsf2^P95H/+, Mx1-Cre⁺ Sf3b1^K700E/+, and Mx1-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ cells relative to Mx1-Cre⁺ WT cells. Circle size indicates the magnitude of the p value for each term and comparison. (D) Immunofluorescence of nuclear phosphorylated-p65 (p-p65) level in LK cells from Vav-Cre⁺ WT, Vav-Cre⁺ Srsf2^P95H/+, Vav-Cre⁺ Sf3b1^K700E/+, and Vav-Cre⁺ Srsf2^P95H/+ Sf3b1^K700E/+ mice following LPS stimulation ex vivo. Scale bars, 10 mm. (E) Violin plots quantifying nuclear p-p65 intensity of LK cells from (D). ANOVA and Kruskal-Wallis ranked test was performed and adjusted for false discovery rate. (F) Schema of competitive BMT using BM MNCs from Vav-Cre⁺ WT, Vav-Cre⁺ Srsf2^P95H/+, and Vav-Cre⁺ Sf3b1^K700E/+ mice after chronic LPS exposure. (G) Percentage of CD45.2 chimerism in peripheral blood of recipients (n = 4–5 mice per group) from primary and secondary BMT (both at 10 weeks post BMT). (H) Chimerism of LSK or myeloid progenitor (MP; lineage^- Sca-1^- c-Kit⁺) fractions from primary recipients 14 weeks post BMT. (I) Kaplan-Meier analysis of Mx1-Cre⁺ WT, Mx1-Cre⁺ Srsf2^P95H/+, and Mx1-Cre⁺ Sf3b1^K700E/+ mice after a single dose of LPS (15 mg/kg) in vivo. Log-rank Mantel-Cox test was performed. p = 0.0055. ANOVAand Tukey’s post hoc test were used to compare groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus Uav-Cre⁺WT mice(“#” denotes versus Vav-Cre⁺ Srsf2^P95H/+ mice and “^Λ” denotes versus Vav-Cre⁺ Sf3b1^K700E/+ mice). In (G) and (H), the top and bottom lines of the box represent the upper and lower quartiles, respectively; the line inside the box represents the median; the lines above and below the box represent the maximum and minimum values, respectively. See also Figure S5 and Table S6.

Figure 6.. SF3B1 Mutations Promote Mis-splicing of MAP3K7, Resulting in Hyperactivation of NF-κB Signaling

(A) Venn diagram illustrating overlap of differentially spliced genes in MDS patient samples mutant versus WT for SF3B1 and murine hematopoietic progenitors in Sf3b1^K700E/+ versus Sf3b1^+/+ mice. (B) From top to bottom, conservation of mouse and human MAP3K7 sequences adjacent to the competing 3 splice site affected by SF3B1 mutations, and RNA-seq coverage plots in human and mouse samples. (C) RT-PCR of the MAP3K7 competing 3′ splice site in MDS and CLL patient samples with or without SF3B1 mutations as well asisogenic human and mouse cells. (D and E) Immuno blot of phosphorylated p65 (p-p65), IκB-α, MAP3K7, and loading controls in isogenic K562 (D) and NALM-6 (E) cells. “Time (hr)” refers to hours following LPS (5 mg/mL) exposure. (F and G) Immunoblot analysis of p-p65, MAP3K7, and loading controls in BM MNCs from MDS (F) and peripheral blood MNCs from CLL patients (G) with or without SF3B1 mutations. See also Figure S6 and Table S7.

Figure 7.. MAP3K7 Loss Results in Hyperactive NF-κB Signaling in SF3B1-Mutant Cells

(A) Immunofluorescence of phosphorylated p65 (p-p65) in K562 cells with or without SF3B1^K700E mutation 2 hr following LPS stimulation. Quantitation of p-p65 intensity is shown on the right (n = 3 independent experiments). Scale bars, 10 mm. (B) Heat map of NF-κB reporter signal in NALM-6 SF3B1-isogenic cells (left two panels) or parental NALM-6 cells with MAP3K7 shRNAs (right two panels) following LPS or TNF a stimulation for 24 hr. (C) qRT-PCR analysis of IL-1β and TNF a 8 hr post LPS stimulation in NALM-6 SF3B1-isogenic cells (n = 2 independent experiments). (D) Immunoblot of p-p65 in K562 SF3B1-isogenic cells ± FLAG-MAP3K7 cDNA and/or LPS (5 mg/mL) exposure for 2 hr. (E) Quantitation of NF-κB reporter signal in cells from (D) (n = 3 independent experiments). (F) Immunofluorescence of nuclear p-p65 and FLAG (MAP3K7) in cells from (D) (quantitation of p-p65 intensity [n = 3 independent experiments] on the right). Scale bars, 10 μm. Error bars represent mean ± SD. ANOVA followed by Tukey’s post hoc test were used to compare groups. *p < 0.05, **p < 0.005, ***p < 0.0002, ****p < 0.0001, ^####p < 0.0001 versus Vehicle; ^###p < 0.001 versus Vehicle. See also Figure S7.

Figure 8.. SRSF2 Mutations Promote Aberrant Splicing of Caspase-8, Resulting in Expression of a Truncated Protein that Hyperactivates NF-κB Signaling

(A) RNA-seq coverage plots of caspase-8 splicing in CMML and AML patients WT or mutant (MUT) for SRSF2. (B) RT-PCR analysis of caspase-8 splicing in SRSF2 WT (TF1 and K562) or SRSF2-mutant (K052) cells. (C and D) Immunoblot of caspase-8 using an N-terminal anti-caspase-8 antibody in K562 SRSF2-isogenic cells (C) or human leukemia cells (left panel) and primary AML patient samples (right panel) WT or mutant for SRSF2 (D). (E) Immunoblot of phosphorylated p65 (p-p65), IκB-α, caspase-8 in K562 cells expressing empty vector (EV), full-length caspase-8 (CASP8^FL), or the truncated caspase-8 isoform (CASP8^TR) after exposure to TRAIL (50 ng/mL). (F) Heatmap of NF-κB reporter signal following TRAIL stimulation in K562, HAP1, or CASP8^KO HAP1 cells expressing EV, CASP8^FL, or CASP8^TR. (G) Immunoblot of p-p65 and IκB-α in K562 cells with or without Srsf2^P95H following exposure to TRAIL (50 ng/mL). See also Figure S8.