Prevalence, Distribution And Predictive Value Of Xpo1 Mutation In A Real-Life Chronic Lymphocytic Leukaemia Cohort

Over the past few years, large-scale sequencing studies have defined the mutational landscape of chronic lymphocytic leukaemia (CLL) and revealed that clonal heterogeneity contributes to the variability in clinical course among patients.1, 2 However, the clinical relevance of most recurrently mutated genes is unknown. To date, only evaluation of TP53 mutation status is recommended in clinical practice3 given its predictive and prognostic significance regardless of the therapeutic regimen.4-6 The clinical relevance of the other recurrent mutations is still being investigated. Among them, mutations of XPO1, encoding the nuclear export protein exportin-1, include the common E571 mutation.1, 2 Mutant XPO1 alters the nuclear export of numerous proteins,7 and it has been suggested that the tumour suppressor protein p53 and its negative regulators are substrates of XPO1.8 Although XPO1 overexpression due to gain in the short arm of chromosome 2 has been associated with drug resistance,9 the relative impact of XPO1 mutation and its prognostic value, compared with those of other recurrent mutations, is unknown. In this report, we evaluated the prevalence, distribution, clonal evolution and prognostic value of XPO1 mutations in a retrospective large cohort of CLL patients. The study included 246 progressive CLL patients with a median age of 60·2 years. IGHV sequencing for assessment of mutational status, conventional cytogenetics, fluorescence in situ hybridization (FISH) analysis and amplicon-based next-generation sequencing (NGS ) of CLL-related mutated genes were performed on blood samples (Data S1). Moreover, we conducted a longitudinal analysis of gene mutations in 5 XPO1-mutated patients. The genetic profile was analysed before initiating therapy (treatment naïve, TN) in 152 progressive patients and in 94 relapsed/refractory (R/R) CLL patients (Table SI). The study was conducted in accordance with national ethical recommendations and the Declaration of Helsinki and written informed patient consent was obtained. Molecular and clinical data are presented in Fig 1. The IGHV status was unmutated (IGHV-UM), mutated (IGHV-M) and unavailable (ND) in 157 (64%), 87 (35%) and 2 (1%) CLL patients respectively. Deletion chromosome 13q, trisomy 12, 17p deletion and 11q deletion were found in 59 of 189 patients (31%), 53 of 211 (25%), 36 of 224 (16%) and 26 of 219 (12%) patients respectively. Complex karyotypes (CK), defined by the presence of at least three chromosomal abnormalities, were present in 57 of 208 (27%) patients and were significantly more frequent in R/R patients than in TN patients [35 of 73 R/R patients (48%) vs. 22 of 145 (15%) TN patients; P < 0·01, chi-squared test]. Altogether, these cytogenetic characteristics are in line with those expected in a cohort of high-risk CLL patients. In this cohort, we identified a relatively high frequency of XPO1 mutations (19 of 246 patients, 8%) compared to other large-scale sequencing studies.2 We observed a strong association between XPO1 mutation and IGHV-UM status, as 18 of 19 XPO1-mutated patients were IGHV-UM (Figure S1). All XPO1-mutated patients harboured the common E571 mutation which was virtually always clonal with a high variant allelic frequency (VAF). The frequency of XPO1 mutations was similar in TN (7%) and R/R (10%) patients (P = 0·4). Likewise, the VAF distribution was similar in TN [median of 36% (6–60%)] and R/R [median of 40% (32–46%)] patients (Fig 2A). Interestingly, longitudinal analysis of five patients harbouring XPO1 mutations showed that the XPO1-mutated clone remained stable with a similar VAF before [40% (27–56%)] and after treatment [37% (20–46%); mean follow-up after treatment: 3·3 years; Fig 2B] suggesting an early appearance of XPO1 mutations and a driver role of mutated XPO1 in CLL. In line with these results, XPO1 mutation was evidenced in early haematopoietic progenitors of CLL patients10 and was demonstrated to be a driving event in B cell malignancies through alteration of the nuclear trafficking of proteins involved in inflammatory signalling, DNA repair, RNA export and chromatin remodelling pathways.7 When considering the other mutated genes, the most frequently mutated genes were TP53 (23%), NOTCH1 (22%), SF3B1 (18%) and ATM (9%). NOTCH1 and TP53 mutations were significantly more frequent in R/R patients compared to the TN patients (Figure S2), as reported previously,11 while no difference was observed for all other mutations tested between the two groups. XPO1 mutation was the sole alteration in six of 19 patients (32%) and variably associated with TP53, SF3B1, NOTCH1, BIRC3, ATM or FBXW7 mutations in 13 of 19 patients (68%) (Fig 1; Figure S3). Given the suggested role of XPO1 in the nuclear export of p53, we particularly focused on TP53 and XPO1 mutation co-occurrence. By re-analysing the data from three large CLL cohorts, we observed that concurrent TP53 and XPO1 mutations were mostly rare (4 of 1697 patients harboured XPO1 and TP53 mutations).1, 2, 12 In our cohort, the co-occurrence of XPO1 and TP53 mutations was found in three patients (Fig 1; Figure S3). However, in these patients, while the XPO1 mutation was clonal [mean VAF: 35% (11–56%)], we noticed that the TP53 mutations were mostly sub-clonal [mean VAF: 14% (2–41%), Fig 2C], suggesting that both mutations were present in independent clones. Interestingly, among the five XPO1-mutated patients included in the longitudinal analysis, we identified the onset of TP53 mutations after treatment in 2/5 patients. These mutations were not detected before treatment with a sensitive NGS technique (Fig 2D). Altogether, these results suggest that when both XPO1 and TP53 mutations are detected, TP53 abnormalities occur later in the course of the disease and therefore define an independent targetable pathway. Finally, consistent with data reported in a short-follow-up cohort (13·5 months),13 we observed that the presence of XPO1 mutation did not alter time to first treatment (TTFT), event-free survival (EFS) or overall survival (OS) of patients whether considering the whole cohort or only IGHV-UM patients (median follow up of 74 months) (Fig 2E). Interestingly, in line with previously published data,14 TTFT was shorter for SF3B1-mutated patients (Figure S4). Nonetheless, recent experimental studies suggest that XPO1-mutated cells are preferentially sensitive to XPO1 inhibitors.7 Moreover, the combination of XPO1 and BCL2 inhibitors blocks tumour progression and extends survival in a lymphoma mouse model.15 These experimental data suggest that combining XPO1 inhibitors with BH3-mimetic-based therapy could be of interest in R/R patients with XPO1 mutations. GT collected, analysed and interpreted the data and wrote the paper. GL analysed and interpreted the data and wrote the paper. VE, RL and CF analysed and interpreted the data. VL and JFC performed the sequencing. JMZ and CT provided blood samples and collected data from their patients. TS analysed the data and reviewed the study. FC and FBM designed the research, collected, analysed and interpreted the data and wrote the paper. The authors declare no competing financial interests. The data that support the findings of this study are available from the corresponding author upon reasonable request. Fig S1. Mutation frequencies based on the IGHV mutational status. Fig S2. Mutation distribution according to patient status (TN, treatment naïve; or R/R, relapsed/refractory). Fig S3. XPO1 mutation associations. Fig S4. Kaplan–Meier curves of time to first treatment (TTFT), comparing patients with SF3B1 (A), ATM (B), NOTCH1 (C) mutation versus patients without the corresponding mutation in the IGHV-mutated group. Table SI. Characteristics of the 246 CLL patients. Data S1. Supplementary methods Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.