SF3B1 Mutations Are Associated with Resistance to Non-Genotoxic MDM2 Inhibition in Chronic Lymphocytic Leukemia

Abstract

Chronic lymphocytic leukemia (CLL) is a genetically and clinically heterogeneous malignancy affecting older individuals. There are a number of current treatment options for CLL, including monoclonal antibodies, targeted drugs, chemotherapy, and different combinations of these. However, for those patients who are intrinsically treatment resistant, or relapse following initial responses, novel targeted therapies are still needed. Targeting the mouse double-minute-2 human homolog (MDM2), a primary negative regulator of p53, is an appealing therapeutic strategy for non-genotoxic reactivation of p53, since the TP53 gene is in its wild-type state at diagnosis in approximately 90% of patients. Mutated SF3B1 and TP53 are both associated with more aggressive disease, resistance to therapies and poorer overall survival for CLL. In this study, we performed a screen for SF3B1 and TP53 mutations and tested RG7388 (idasanutlin), a second-generation MDM2 inhibitor, in a cohort of CLL primary patient samples. SF3B1 mutations were detected in 24 of 195 cases (12.3%) and found associated with poor overall survival (hazard ratio [HR] 2.12, p = 0.032) and high CD38 expression (median CD38 (%) 32 vs. 5; p = 0.0087). The novel striking finding of this study was an independent link between SF3B1 mutational status and poor response to RG7388. Overall, SF3B1 mutations in CLL patient samples were associated with resistance to treatment with RG7388 ex vivo, and patients with the wild type for both SF3B1 and TP53 are more likely to benefit from treatment with MDM2 inhibitors.

Keywords: MDM2-p53 antagonists; RG7388 (idasanutlin); SF3B1; chronic lymphocytic leukemia (CLL); splicing factor.

Figure 1

(A) The percentage distribution of the SF3B1 mutations and respective amino acid positions across all CLL patient samples are displayed in a pie chart. (B) SF3B1 mutations were seen with lower a allelic fraction in earlier samples taken from the same patient, indicating clonal expansion during leukemic progression. For the patient sample CLL199, the mutation in SF3B1 (c.1997A>C; p.K666T) was not detected in the earlier samples CLL33 and CLL120.

Figure 2

Survival plots of CLL patients depending on their SF3B1 status. Kaplan–Meier curves of (A) overall survival (OS) and (B) treatment-free survival plot defined as time to first treatment (TTFT). Relationship of SF3B1 mutations to (C) overall survival and (D) treatment-free survival by separating the subgroup of patients with del(17p) and/or TP53 mutation. Blue line: SF3B1^WT without del(17p). Red line: SF3B1^MUT without del(17p). Black line: All cases with del(17p) independent of SF3B1 status. (E) Overall survival and (F) treatment-free survival plot defined as time to first treatment. Blue line: Both SF3B1 and TP53 wild type (without del(17p)). Red line: SF3B1^MUT and TP53^WT (without del(17p)). Black line: All cases with del(17p) and/or TP53 mutation independent of SF3B1 status. The information next to the red and black lines shows the comparison with the blue lines in the corresponding graph. p values were calculated by log-rank statistics. HR: hazard ratio. CI: confidence interval.

Figure 3

Primary CLL samples are grouped as SF3B1^MUT and SF3B1^WT according to their CD38 expression. CD38 status of primary CLL patient samples. Horizontal bars show the median CLL cell % positivity for CD38 expression. The % CD38 positivity was compared between SF3B1 mutant and wild-type CLL samples. The p value shows a significant difference between the median values for the groups calculated using the Mann–Whitney U test.

Figure 4

Comparison of response to RG7388 in CLL samples with respect to SF3B1 mutational status. (A) Representative cytotoxicity curves for three SF3B1^MUT (CLL233, 240 and 288) and three SF3B1^WT (CLL265, 275 and 284) CLL samples exposed to increasing concentrations of RG7388 (from 0 to 10 µM) for 48 h. All six samples are wild type for the TP53 gene. The SF3B1^MUT samples were highly resistant to RG7388 (LC₅₀ > 10 µM). Cell viability was assessed by XTT assay. (B) LC₅₀ values showing the range of responses to RG7388 in TP53^WT CLL samples. SF3B1^MUT samples are marked with different colors and the mutated codons are indicated beside the graph. CLL233 is the only one resistant to RG7388 among the four CLL samples with the identical SF3B1^K700E mutation (shown in blue). (C) LC₅₀ values showing the range of responses to RG7388 in SF3B1^WT CLL samples. Horizontal bars show the median. There was no significant difference between the primary samples lacking both del13q and del11q, clinically relevant cytogenetic abnormalities in CLL, and samples carrying either or both of these abnormalities (unpaired t-test; p = 0.14). ns: not significant. (D) The majority of (4 out of 5) the highly resistant (LC₅₀ > 10 µM) CLL samples were found to be SF3B1^MUT. The chi-square test was used to measure whether the proportion of cases with LC₅₀ values greater than 10 µM is statistically significant between the SF3B1^WT and SF3B1^MUT groups. **: p = 0.0017. The table below the bar graph shows the total number of CLL samples belonging to the related group. (E) Western blots of SF3B1^MUT CLL samples showed low levels of p53 increase saturated at the lowest drug concentration without activation of the downstream protein MDM2. On the other hand, concentration-dependent stabilization of p53 and the activation of the downstream protein MDM2 was evident in the SF3B1^WT samples. Actin was used as loading control. D: DMSO-treated control cell lysate; +C: positive control lysate from NGP cells treated with 2.5 µM Nutlin-3a for 4 h; MUT: mutant; WT: wild type.

Figure 5

Comparative response of Nalm-6 isogenic cells expressing either SF3B1^K700K or SF3B1^K700E to different MDM2-p53 binding antagonists. (A) Growth curves of Nalm-6 isogenic cell lines over a period of four days in culture. The number of viable cells was assessed by trypan blue exclusion. Error bars show the mean ± standard error of the mean (SEM) of three independent counts of viable cells observed under trypan blue exclusion assay. (B) Sanger sequencing showing Nalm-6^K700E has a heterozygous point missense mutation c.2098A>G resulted in the amino acid change from lysine to glutamic acid at amino acid position 700 (K700E). Multiple silent mutations around the target point mutation were introduced to disrupt the PAM sequence and prevent re-cutting of the modified allele by Cas9. The sequencing was primed in the reverse direction of the coding strand. Red asterisks show the point mutation. (C,D) Pre-B isogenic Nalm-6 cells were treated with different concentrations of MDM2 inhibitors either RG7388 or HDM201 for 72 h. The % viability was measured by an XTT assay and normalized to DMSO solvent control treatment. Data are representative of three independent experiments and the error bars show the mean ± standard error of the mean (SEM) of at least three independent repeats. (E) Summary of IC₅₀ (mean ± SEM) values for RG7388 and HDM201 in Nalm-6 isogenic cell lines from three independent repeats. Statistically significant differences between the cell lines (* p ˂ 0.05; ** p ˂ 0.01) are indicated.

Figure 6

(A) Each cell line transfected with 500 nM siRNA targeting SF3B1 was compared with the corresponding cell line transfected with 500 nM scrambled siRNA control. SF3B1 protein levels measured 48 h post-transfection. Actin was used as loading control. (B) Cell count of viable cells for three cell lines with SF3B1 knockdown, compared with untreated and control siRNA transfected cells, as assessed by trypan blue exclusion assay. There was no statistically significant difference between the growth of siControl and siSF3B1 for OCI-Ly3 and Daudi cells (unpaired t-test; p > 0.05). Growth of HEL cells was inhibited with siSF3B1. Error bars show the mean ± SEM. Statistical analysis was performed by t-test and corrected for multiple comparisons using the Holm-Sidak method (* p < 0.05; ** p < 0.01 different from siControl control). (C) Schedule of treatment and immunoblot showing p53^WT OCI-Ly3 cells either non-transfected (UT) or transfected with 500 nM siControl or siSF3B1 24 h prior to treatment with 0, 0.1 or 0.5 μM of HDM201. HDM201 stabilized p53, with subsequent activation of downstream protein, p21^WAF1. Actin was used as loading control. (D) Summary of IC₅₀ values for HDM201 and RG7388 in OCI-Ly3 cells either non-transfected (UT) or transfected with 500 nM siControl or siSF3B1 24 h before treatment with the compounds. Viability measured by an XTT assay 48 h post-treatment with HDM201 or RG7388. There was no significant difference between the cells transfected with the siControl and siSF3B1 (p = 0.78 for HDM201; p = 0.80 for RG7388). Data are presented as the mean ± SEM for three independent repeats. M: only media control cell lysate; D: DMSO-treated control cell lysate. UT: untreated (non-transfected).