Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic lymphocytic leukemia

Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic... OPEN Citation: Blood Cancer Journal (2015) 5, e342; doi:10.1038/bcj.2015.65 www.nature.com/bcj ORIGINAL ARTICLE Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic lymphocytic leukemia 1 1 2 1 1 2 1 1 1 AM Hurtado , T-H Chen-Liang , B Przychodzen , C Hamedi , J Muñoz-Ballester , B Dienes , MD García-Malo , AI Antón , F de Arriba , 1 1 1 2 1 R Teruel-Montoya , FJ Ortuño , V Vicente , JP Maciejewski and A Jerez An increasing numbers of patients are being diagnosed with asymptomatic early-stage chronic lymphocytic leukemia (CLL), with no treatment indication at baseline. We applied a high-throughput deep-targeted analysis, especially designed for covering widely TP53 and ATM genes, in 180 patients with inactive disease at diagnosis, to test the independent prognostic value of CLL somatic recurrent mutations. We found that 40/180 patients harbored at least one acquired variant with ATM (n = 17, 9.4%), NOTCH1 (n = 14, 7.7%), TP53 (n = 14, 7.7%) and SF3B1 (n = 10, 5.5%) as most prevalent mutated genes. Harboring one ‘sub-Sanger’ TP53 mutation granted an independent 3.5-fold increase of probability of needing treatment. Those patients with a double-hit ATM lesion (mutation+11q deletion) had the shorter median time to first treatment (17 months). We found that a genomic variable: TP53 mutations, most of them under the sensitivity of conventional techniques; a cell phenotypic factor: CD38-positive expression; and a classical marker as β2-microglobulin, remained as the unique independent predictors of outcome. The high-throughput determination of TP53 status, particularly in this set of patients frequently lacking high-risk chromosomal aberrations, emerges as a key step, not only for prediction modeling, but also for exploring mutation-specific therapeutic approaches and minimal residual disease monitoring. Blood Cancer Journal (2015) 5, e342; doi:10.1038/bcj.2015.65; published online 28 August 2015 INTRODUCTION In this study, we propose to refine and apply a method for high- throughput targeted analysis of somatic recurrent mutations in Nowadays, an increasing numbers of patients are being diagnosed CLL, especially designed for covering widely TP53 and ATM genes. with early-stage chronic lymphocytic leukemia (CLL), likely owing Our main aim is to assess the independency of the prognostic to the use of routine blood tests for health screening and the 1–3 value of those variants, related to time to first treatment and widespread availability of flow cytometry. Among this subset of survival, in patients with CLL and no indication for therapy at CLL patients, most with a non-active disease and no treatment diagnosis. indication at baseline, different prognostic modeling approaches, incorporating traditional (clinical and laboratory), cytogenetic, immunophenotypic and immunoglobulin heavy-chain variable 4,5 MATERIALS AND METHODS region gen (IgVH) status, have been proposed. The contemporary vision of neoplasm development is based on Patients a consecutive acquisition of genetic changes with a selection and From 2006 to 2012, presentation bone marrow aspirates or blood samples 6,7 expansion of the more fit population. The heterogeneous DNA was collected during the diagnostic workout from 265 consecutive course of CLL is, to a certain extent, driven by the diverse CLL patients, after informed consent, according to the protocols approved by the Institutional Review Board of Hospital Morales Meseguer (EST-32/13) combinations of clones with acquired chromosomal lesions and 8,9 and with the Declaration of Helsinki. Patients who met criteria for an active somatic mutations. Whole-exome studies (WES) have shown disease at baseline, did not reach a minimum treatment-free follow-up of that the number of somatic variants per case is lower in CLL than 3 months, or nucleic acids did not pass the quality control for either IgVH that of those described in solid tumors and other leukemias, and status or targeted sequencing, were excluded (Figure 1). Diagnosis and 10–12 that the set of genes affected is discreet. These two aspects definition of active disease, requiring therapy, were achieved according to make this disease a suitable candidate for deep-targeted sequen- the International Workshop on Chronic Lymphocytic Leukemia established cing, a technique focused on distinct genomic sites, which also criteria. Time-to-first-treatment (TTFT) was measured from diagnosis to enables reliable detection of subclonal mutations because of to its date of first treatment. Regular follow-up consisted of blood cell counts and clinical examinations every 3 months the first year after diagnosis, and higher depth of coverage compared with WES. Using this more henceforth, visits were carried out from 3 to 6 months, depending on affordable and rapid strategy, recent studies have determined the patient risk. prognostic impact of TP53 subclones. In addition, the status of ATM, NOTCH1 and SF3B1, also recurrently mutated in CLL, have been 14–16 associated with impaired overall and treatment-free survival. Diagnostic workout Nevertheless, a comprehensive high-throughput sequencing study Every patient underwent a flow cytometry characterization with a panel of these variants, assessing their clinical relevance, in the context of including CD45, pan B-cell markers (CD19, CD20, CD22, CD79b, and surface both traditional and newer factors, is lacking. immunoglobulin light chains), markers for differential diagnosis with other 1 2 Hematology and Medical Oncology Department, Hospital Morales Meseguer, IMIB, Murcia, Spain and Traslational Hematology and Oncology Research, Cleveland Clinic, Cleveland, OH, USA. Correspondence: Dr A Jerez, Hematology and Medical Oncology Department, Hospital Morales Meseguer, IMIB, Marques de los Velez s/n, Murcia E-30008, Spain. E-mail: andjercay@gmail.com Presented in part at the 57th ASH Annual, San Francisco, CA, December 6–9, 2014. Received 18 May 2015; revised 17 June 2015; accepted 30 June 2015 Outcome and clonal pattern in inactive CLL AM Hurtado et al B-cell chronic lymphoproliferative diseases (CD5, CD23, FMC7, CD10, CD81, (2 × 250 bp) was performed with MiSeq v2.2 chemistry, and a mean depth CD103, CD25 and CD11c) and prognosis markers (CD38 and ZAP70) of 938 reads/base within the regions of interest was obtained. Raw data (Antibodies from BD Biosciences, San Jose, CA, USA). were analyzed with IlluminaonJboard Real Time Analysis (RTA v.2.4.60.8) Fluorescence in situ hybridization (FISH) analysis was performed on software and MiSeq Reporter. interphase nuclei at diagnosis from directly harvested peripheral blood or bone marrow samples according to the manufacturer’s protocol and using Variant call requirements and validation the following commercially available probes (Abbott Molecular, Des The following conditions were established for a variant to be called: (i) to Plaines, IL, USA): LSI MYB (6q23), LSI P53 (17p13.1)/ LSI ATM (11q22.3), be non-synonymous; (ii) not to be listed in dbSNP database (NCBI Human LSI D13S319 (13q14.3)/CEP12, as reported. A minimum of 400 nuclei Build 141); and (iii) a cutoff for any nucleotide position of 30 or more were scored for each probe or probe combination. variant reads and a Q score 430 (see Variant call requirements: technique Immunoglobulin heavy-chain variable diversity (D)-joining (J) rearrange- accuracy in Results). The filtered variant lists were manually reviewed and ments were amplified from either reverse-transcribed total RNA (preferred BAM file examined in Integrated Genome Viewer (Broad Institute). source) or genomic DNA. Purified amplicons were sequenced either Every variant with a clonal size of, at least, 20% and 430 variant reads directly or on subcloning. Sequences were aligned to the ImMunoGe- were bi-directionally sequenced using an ABI 3730 DNA Analyzer (Life neTics for computation of mutational load. Sequences were considered Technologies, Carlsbad, CA, USA). mutated or not using the cutoff of 2% mismatch. Six TP53-mutated cases were selected for applying the whole-amplicon panel on germline DNA (four cases from CD3+ sorted cells, two cases from Targeted sequencing buccal mucosal swab) to test its somatic nature. We designed a TruSeq Custom Amplicon panel (Illumina, Inc. San Diego, CA, USA) containing 13 genes and covering 28.099 bases (Table 1). For Statistical analysis some genes known mutation hotspots were targeted; and for those with a Comparisons of proportions and ranks of variables between groups were widespread localization of the lesions, the entire coding sequence was analyzed. The average amplicon size was 238 base pairs and ~ 99.1% performed by χ -test, Fisher's exact test, t-test or Mann–Whitney U-test, as of the regions were covered on both strands. Library preparation was appropriate. We used the Kaplan–Meier and the Cox method to analyse performed according to manufacturer's instruction. Paired-end sequencing overall survival (OS) and progression-free survival, with a two-sided P- value⩽ 0.05 considered to be significant. In Cox models, examination of log (− log) survival plots and partial residuals was performed to assess that the underlying assumption of proportional hazards was met. Thresholds of o2.4 mg/dl for β2-microglobulin, an absolute B-cell count of 11 ×10 /l or over, and higher LDH levels than the upper normal limit (4UNL (that is, 23–25 378 U/l), were chosen as reported elsewhere. RESULTS Testing technique accuracy by resequencing To estimate the accuracy of the technique to reproduce a variant call, 221 variants found in the first run were resequenced in depth (average reads per variant × 583). Eighteen of these variants had a clonal size over Sanger sensitivity (420%), 58 variants with a clonal size between 10 and 20%, and 145 variants below 10% of clone size. For this experiment, new libraries were built to capture and amplify exclusively the amplicons covering the variants selected, and the same genomic DNA used in the first run was used. Receiver operating characteristic curves for both number of Figure 1. Study flow diagram. Visual representation of the exclusion variant reads and variant allelic frequency (VAF) were created by criteria (left) and the targeted sequencing process pipeline (right). plotting the true positive rate (sensitivity) against the false-positive Table 1. Targeted NGS panel characteristics Target gen Length covered % Gen exome % CLL mutations Hotspot references (bp) sequenced covered* TP53 2724 100% 100% 12 10 ATM 7411 56% 78% (28 /36) Wang et al., Puente et al. 12 10 NOTCH1 1487 19% 100% (75/75) Wang et al., Puente et al. 12 10 SF3B1 3915 14% 92% (22/24) Wang et al., Puente et al. 12 10 MYD88 296 31% 100% (11/11) Wang et al., Puente et al. 12 10 POT1 549 28% 100% (5/5) Wang et al., Puente et al. 12 10 45 BIRC3 963 53% 100% /(11/11) Wang et al., Puente et al., Rossi et al. KRAS 567 100% 100% NRAS 570 100% 100% U2AF1 572 100% 100% 12 10 BRAF 357 14% 100% (2/2) Wang et al., Puente et al., Jebaraj et al. 12 10 BCOR 5268 100% 100% Wang et al., Puente et al. 12 10 SETBP1 3420 71% 100% (2/2) Wang et al., Puente et al. TOTAL 28099 Abbreviations: bp, base pair; CLL, chronic lymphocytic leukemia. Genes included, total gene length covered, percentage of gene exon sequenced and percentage of CLL mutations covered using as reference those variants described in the whole-exome studies referred in the last column. Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al rate (1-specificity) at various threshold settings. The number of SF3B1 and NOTCH1 mutations were mutually exclusive, and a variant reads in the initial run showed to be more accurate to significant correlation between NOTCH1 mutations and the predict the reproducibility of the variant in a second run, with an presence of a trisomy 12 was found (Po0.01). area under the curve of 0.894 (P ⩽ 0.001; 95% CI, 0.817-0.970), than the clonal size (VAF) (area under the curve: 0.613; P = 0.037; 95% Clonal diversity of recurrently mutated genes in early-stage CLL.As CI, 0.487–0.738; Supplementary Figures 1A and 1B). A cutoff of 30 samples used for this sequencing study belong to the diagnostic reads was chosen as threshold to consider a variant reproducible immunophenotypic workout, we could estimate the clonal or with a sensibility of 0.85 and a specificity of 0.945. No discrepancy subclonal nature of the acquired mutations, adjusting the variant was found when Sanger sequencing 15 selected variants over allele burden in non-sorted blood or bone marrow DNA for the 20% of clonal size. None of the 13 variants harbored by six percentage of the immunophenotypically quantified CLL popula- TP53-mutated cases was called when applying the panel on tion. In eight patients with a del11q or del 17p and a ATM or TP53 germline DNA. mutation, VAF was adjusted considering the loss of heterozygosity (Figure 3). Cohort and distribution of mutations. Two hundred and sixty-five Thirty-five mutations were estimated to be clonal, that is, patients were diagnosed in our center of a CLL from 2006 to 2012. present in the whole tumor population, either in a heterozygous Patients who, without needing therapy, did not achieve a (n = 28) or hemizygous (n = 7) configuration. The allelic ratio minimum of 3 months of follow-up, were not contemplated of other mutations showed that they appeared only in a fraction (Figure 1). No monoclonal lymphocytosis cases were considered. of tumor cells, indicating that they were secondarily acquired or The baseline characteristics of the 180 patients finally included subclonal (n = 33). Certain genes showed predominantly clonal were in accordance with their indolent-no need for treatment mutations (71% of NOTCH1, 71% of ATM variants), whereas others status at diagnosis, with 93% of the cohort assigned to Rai Stages were mainly subclonal (80% of TP53 variants, 100% of Ras-family 0 and I (Table 2). and U2AF1 genes). We found that 40/180 (22.2%) patients harbored at least one mutation; ATM (n = 17, 9.4%), NOTCH1 (n = 14, 7.7%), TP53 (n = 14, Clinical correlates. When confronting the 40 patients with, at 7.7%), SF3B1 (n = 10, 5.5%), BCOR (n = 3, 1.6%), BIRC3 (n = 2, 1.1%), least, one mutation, with the 140 non-mutated patients, we found KRAS (n = 2, 1.1%), U2AF1 (n = 2, 1.1%), POT1 (n = 2, 1.1%), less patients stratified as stage 0 in the mutated group, in favor of MYD88 (n = 1, 0.6%), and SETBP1 (n = 1, 0.6%). No somatic stages I and II. Focusing on FISH abnormalities, both 13q and +12 variants were identified for BRAF and NRAS (Figure 2). Sixty-eight shown to be more frequent in mutated cases, though none of mutations were detected in the whole cohort with 18 deletions these differences reached the statistical significance (Table 2). causing a frameshift, 1 non-frameshift deletion, 1 non-frameshift Likewise, no significant disparity was found when considering insertion and 48 missense single-nucleotide variants. Forty-one leukocyte, lymphocyte and platelet counts or hemoglobin, lactate out of 68 mutations were already reported to the Catalog of Somatic Mutations in Cancer (COSMIC; http://cancer.sanger.ac. dehydrogenase and β -microglobulin levels. In addition, no specific CD38 or ZAP70 expression pattern was observed to be uk/cancergenome/projects/cosmic), as human cancers variants (Supplementary Table 1). characteristic of each group. Table 2. Characteristics of patients included in the study at baseline and according to the presence or absence of, at least, a mutation Variable Total (n = 180) (A) Mutated cases (n = 40) (B) Non-mutated cases (n = 140) P (A vs B) Age, years, (mean± s.d.) 69 ±170± 12 68 ± 11 0.9 Sex, male/female,% (n) 59 (107)/41 (74) 67 (27)/33 (13) 60 (85)/40 (55) 0.3 Rai Stage, % 0 45 38 49 0.3 I 48 52 47 0.5 II 7 10 4 0.3 Leukocytes, x10 /l (median, IQR) 15.3 (10.9–26.9) 16.6 (10.9–24.2) 14.6 (10.7–24) 0.2 Lymphocytes, x10 /l (median, IQR) 10.3 (7–19.8) 12.2 (7.3–20.6) 9.8 (7–16) 0.3 B-cell count, x10 /l (median, IQR) 9.1 (6.1–17.3) 11.2 (7.4–19.1) 8.9 (6–15.6) 0.3 Hemoglobin, g/dl (median, IQR) 13.8 (12.2–14.7) 13.7 (12–14.5) 13.8 (12.3–14.7) 0.6 Platelets, x10 /l (median, IQR) 182 (136–217) 168 (126–216) 187 (136–217) 0.6 LDH IU/l (median, IQR) 368 (323–425) 377 (318–453) 324 (288–390) 0.4 B microglobulin, mg/l (median, IQR) 2.2 (1.8–3) 2.4 (1.8–3.6) 2.2 (1.7–3) 0.3 ZAP70 +, % (n) 42 (76) 50 (20) 40 (56) 0.1 CD38 +, % (n) 28 (50) 38 (15) 22 (30) 0.1 IgHV-unmutated status, % (n) 34 (61) 43 (17) 27 (37) 0.07 FISH abnormalities, % (n) Del(13q) 46 (84) 50 (20) 47 (65) 0.7 +12 18 (32) 25 (10) 14 (20) 0.2 Del(11q) 5 (8) 17 (7) 1 (0.7) 0.01 Del(17p) 2 (4) 10 (4) 2 (1.4) 0.02 Abbreviations: CD38, cluster of differentiation 38; FISH, fluorescence in situ hibridization; IgVH, immunoglobulin heavy-chain variable region gen; IQR, interquartile range; LDH, lactate dehydrogenase; ZAP70, Zeta-chain-associated protein kinase 70; +12, trisomy 12. Bold face: del(11q) and del (17p) as variables distributed differently among mutated or not mutated patients with statistical significance (P= 0.014) . Note: quantitative variables with a normal distribution are expressed as mean ± s.d. Quantitative variables not following the normal distribution are expressed as median and interquartile ranges. Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al Figure 2. Distribution of mutations and chromosomal aberrations. Figure 3. Estimated variant allele frequencies in the tumor fraction. The vertical axis represents allele frequency. Patients mutated are depicted in the abscissa. Color-coded circles for each gene affected. The variant allele frequency in non-sorted blood or bone marrow DNA was adjusted for the percentage of the immunophenotypically quantified CLL population. Despite the low number of 17p and 11q deletions in our non- hazards ratio (HR)=5.8;95% CI,3.1–10.9; Figure 4a) than those cases aggressive cohort, they were found mostly in the mutated group; without a detected mutation, and a shorter median OS (54 months the only difference statistically significant. vs not reached; P = 0.01; HR = 3.9; 95% CI, 2.2–6.9; Figure 4f). With a median follow-up of 54 months (interquartile range, TP53-mutated cases showed both a shorter TTFT (median 42–85 months), the median OS of the whole cohort has not been 29 months vs not reached; P ⩽ 0.001; HR = 5.3; 95% CI, 2.6–10.8 CI; reached. Forty-two patients (23.3%) required therapy. Median time Figure 4b) and OS (median 48 months; P ⩽ 0.001; HR = 3.7; 95% CI, to treatment for all patients has not been reached. Considering 1.9–7.2; Figure 4g) than the wild-type cases (median not reached). only those patients who were treated, median time to first The presence of a mutation in ATM predicted both for a shorter treatment was 48 months (range, 5–96 months). TTFT (mean 60 months; P ⩽ 0.001; HR = 5; 95% CI, 2.5–10.2 CI; Patients with, at least, one mutation had a worse time to first Figure 4c) and OS (median 61 months; P = 0.016; HR = 2.5; 95% CI, treatment (median TTFT: 60 months vs not reached; P⩽ 0.001; 1.2–5.2 CI; Figure 4h). Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al 1.0 TP53 WT cases (n=166, events=31) 1.0 1.0 WT cases (n=140, events=21) ATM WT cases (n=162, events=31) 0.8 0.8 0.8 0.6 0.6 0.6 Mutated cases ATM mutated cases (n=40, events=21) 0.4 0.4 0.4 (n=18, events=11) TP53 mutated cases 0.2 0.2 0.2 (n=14, events=11) p≤0.001 p≤0.001 0.0 0.0 p≤0.001 0.0 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120 TTFT (months) TTFT (months) TTFT (months) 1.0 1.0 1.0 SF3B1 WT cases (n=170 events=36) NOTCH1 WT cases (n=166, events=35) WT cases (n=140, events=31) 0.8 0.8 0.8 0.6 0.6 0.6 Mutated cases (n=40, events=23) NOTCH1 mutated cases 0.4 0.4 0.4 (n=14, events=7) SF3B1 mutated cases 0.2 0.2 0.2 (n=10, events=6) p=0.006 p≤0.001 0.0 0.0 0.0 p≤0.001 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 TTFT (months) TTFT (months) Overall Survival (months) 1.0 1.0 1.0 TP53 WT cases (n=166, events=42) NOTCH1 WT cases (n=166, events=46) ATM WT cases 0.8 0.8 0.8 (n=162, events=45) 0.6 0.6 0.6 ATM mutated cases 0.4 0.4 0.4 (n=18, events=9) TP53 mutated cases NOTCH1 mutated cases (n=14, events=12) (n=14, events=8) 0.2 0.2 0.2 p=0.01 p=0.016 0.0 0.0 0.0 p=0.001 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Overall Survival (months) Overall Survival (months) Overall Survival (months) 1.0 SF3B1 WT cases (n=170, events=49) 0.8 0.6 0.4 SF3B1 mutated cases (n=10, events=5) 0.2 0.0 p=0.03 0 20 40 60 80 100 Overall Survival (months) Figure 4. Differences in time to first treatment (TTFT) and survival outcomes (OS) in patients with, at least a mutation vs non-mutated (a, f); TP53 mutated or WT (b, g); ATM mutated or WT (c, h); NOTCH1 mutated or WT (d, i); and SF3B1 mutated or WT (e, j). P-values presented correspond to the Cox regression between the groups indicated. NOTCH1-mutated cases also presented a decreased TTFT (that is, variables strongly related to each other, measuring the (66 months; P = 0.006; HR = 3.1; 95% CI, 1.4–7.1; Figure 4d) and a same effect), precluded us from using them in this step of the shorter median OS (61 months; P = 0.01; HR = 2.7; 95% CI, 1.3–5.7; analysis. Mutational status of TP53, ATM, NOTCH1, SF3B1, IgVH status, the Figure 4i) than the wild-type cases (not reached). SF3B1-mutated presence of +12, del(13q), the Rai Stage, the positive expression of patients also showed worse prognosis both in terms of TTFT CD38 and ZAP70, β2-microglobulin and LDH serum levels, and a (median 44 months; P ⩽ 0.001; HR = 5.3, 95% CI, 2.2-12; Figure 4e) B-cell count ⩾ 11 × 10 /l were evaluated in a univariate Cox and of OS (P = 0.05; HR = 2.7; 95% CI, 1–7.8; Figure 4j) than the regression, both for TTFT and OS (Supplementary Tables 2A and B). wild-type cases (mean not reached). Only those variables with a P⩽ 0.05 were included in a multi- The low number of del(11q) and del(17p) cases, and the potential correlation of regressors with ATM and TP53 mutations variate analysis (Tables 3A and B). The presence of a somatic Blood Cancer Journal Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Outcome and clonal pattern in inactive CLL AM Hurtado et al DISCUSSION Table 3A. Multivariate Cox regression for time to first treatment In the last few years a boost in the number of patients being diagnosed with early-stage CLL, not requiring therapy at diagnosis Variable RR (95% CI) P-value 1–3 has been stated. In this subset, we have shown that TP53 high- TP53 3.9 (1.4–10.9) 0.01 throughput mutational status emerges as an independent IgVH 2 (0.9–4.4) 0.08 predictor, even when adjusting for the other recurrent gene NOTCH1 1.9 (0.7–5.7) 0.2 variants in CLL, and for both traditional and recently reported ATM 1.9 (0.7–4.6) 0.17 prognostic factors. Surprisingly, even including newer genomic +12 1.3 (0.6–2.7) 0.4 prognostic factors, a classical serologic parameter as β2-microglobulin CD38 2.3 (1–5.3) 0.04 levels, remained an independent predictor both for TTFT and OS, SF3B1 2.6 (0.8–8) 0.09 Rai Stage40 1.1 (0.5–2.3) 0.7 whereas a positive CD38 expression also independently predicted B-cell count ⩾ 10 × 10 /l 1.9 (0.9–3.9) 0.08 a shorter TTFT. β2-microglobulin ⩾ 2.4 2.4 (1.2–4.8) 0.02 The baseline characteristics of the patients included in our study were in accordance with their indolent-no need for treatment Abbreviations: CD38, cluster of differentiation 38; CI, confidence interval; status. As other groups have shown, the new definition of CLL, IgVH, immunoglobulin heavy-chain variable region gen; RR, relative risk; +12, trisomy 12. Bold face: variables reaching the independent statistical excluding those cases with a B-monoclonal population of o5000/ significance. ul, determined in our cohort a shift toward higher Rai Stages and a 26,27 higher rate of patients progressing and needing therapy. Patients with deletions on chromosome 17p respond worse to treatment than do those without it, resulting in early relapse and 28–30 Table 3B. Overall survival shorter survival. This cytogenetic lesion can be found in up to 50% of relapsed and refractory patients, but is rare at baseline Variable RR (95% CI) P-value (5–10%). We have found a 5% of TP53 mutations in a subset of patients where a ‘wait and watch’ therapeutical strategy is TP53 2.5 (1.2–6.4) 0.02 currently recommended. The relevant frequency and prognostic IgVH 1.5 (0.8–2.6) 0.2 impact of TP53 variants represented in o20% leukemic cells was β2-microglobulin ⩾ 2.4 1.9 (1.1–3.4) 0.04 recently reported by Rossi, et al. Of note, we show here that the NOTCH1 1.8 (0.7–4.5) 0.2 majority of TP53 mutations, at baseline in our inactive CLL subset, SF3B1 2.5 (0.9–7.2) 0.08 Rai Stage 40 1.1 (0.5–2.3) 0.7 were under the sensitivity of conventional sequencing, and that ATM 0.9 (0.4–2.5) 0.9 they kept its independent impact on prognosis even when adjusting by high-throughput detected ATM, NOTCH1 and SF3B1 Abbreviations: CD38, cluster of differentiation 38; CI, confidence interval; variants. Given the markedly short TTFT and OS of our TP53- IgVH, immunoglobulin heavy-chain variable region gen; RR, relative risk; mutated patients, it would seem an attractive approach to +12, trisomy 12. Bold face: variables reaching the independent statistical significance. consider investigational studies directed to eradicate these subclones, in which these patients could be treated. Closely related, Farooqui et al. reported a remarkable 2-year survival and variant in TP53, a positive CD38 antigen expression and β2- excellent drug-side-effects profile in the largest series of Ibrutinib microglobulin serum levels above 2.4 mg/dl, prevailed as inde- therapy in treatment-naive patients with CLL and 17p deletions. In pendent variables linked to a shorter time for needing treatment, addition, identification of these subclones emerges as crucial as a whereas only TP53 mutational status and β2-microglobulin levels concern is raised for clonal evolution and selection of resistant remained as significant predictors for OS. clones due the use of conventional (p53-dependent DNA damage) 9,32 We next sought to explore the impact of TP53 lesions below the chemotherapeutic agents. sensitivity of Sanger. That is, to define the role of high-throughput The frequency of ATM mutations in large series of CLL patients it sequencing in defining the prognosis in our cohort, as TP53 is not so well understood, mostly owing to the fact that its size mutations remained as the only independent genomic variable. In and the scattered distribution of its somatic mutations precludes 33,34 this sub-analysis we excluded five patients with a TP53 lesion the drafting of an amplicon-limited design. We found ATM to detectable by conventional techniques (direct sequencing and be the most frequently mutated gene, with a predominance of FISH/cytogenetics): one patient with a subclone harboring a TP53 clonal variants, indicating the ancestral-founding nature of these mutation and a del(17p), two patients with a clone with both a lesions. In addition, we found that the group with a shorter TTFT in TP53 and a del(17p), a case with a del(17p) in a clonal fashion but our study was defined by those patients with a double-hit ATM. not detectable mutation, and one patient with a clonal TP53 Likely, the reduced number of cases accounts for not reaching a mutation and 37% of variant reads. The other 10 subclonal cases statistical significance in multivariate or when addressing OS. Our would have been missed by Sanger sequencing as they called in finding supports the notion of a biological and clinical separation o20% reads (seven of them would have been missed even in of double-hit ATM cases, from those where 11q deletion and the B-cell-sorted DNA). We then replicated the multivariate Cox presence of an undamaged ATM allele give rise to a functional 35–37 regression analyses shown in Tables 3A and B. Harboring one of protein. these ten ‘sub-Sanger’ TP53 mutations granted an independent Recent studies have suggested an important prognostic role for 15,38,39 3.5-fold increase of probability of needing treatment during the NOTCH1 and SF3B1 in CLL. However, no study has course of the disease than a wild-type patient (P = 0.04; HR = 3.5; performed a multivariate analysis including traditional clinical 95% CI, 1–12.2), but it did not reach the significance for predicting and laboratory markers, flow cytometry factors, IgVH status and, at OS (P = 0.15; HR = 2; 95% CI, 0.8–5.5). least, the presence of ATM, TP53, NOTCH1 and SF3B1 variants. In Finally, a second sub-analysis showed that those patients with a addition, using high-throughput sequencing seems essential to double-hit ATM lesion (mutation+11q deletion) had the shorter discern the independent prognostic value of these mutations, as median TTFT reported in this study (17 months) strikingly reduced sub-Sanger mutations accounted for 25% of the variants in our compared with one-hit ATM patients (60 months). This impact was study. Though strongly associated with a worse outcome and TTFT not seen in multivariate analysis or concerning OS (Supplementary in univariate analysis, NOTCH1 and SF3B1 did not kept its Figure 2). significance when including TP53 and ATM mutations. Five out Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al of seven, and four out of five NOTCH1 and SF3B1-mutated cases, of patients lacking treatment indication at baseline, adds another which needed therapy, also harbored a TP53/del 17p and/or cobblestone to the positioning of amplicon deep-sequencing ATM/del11q lesion. This co-occurrence with more aggressive assays in established CLL diagnostics algorithms. The high- alterations can explain, in part, the lack of independent predictive throughput determination of TP53 status, particularly in this set value of SF3B1 and NOTCH1 in our cohort. Consistent with of patients frequently lacking high-risk chromosomal aberrations, previous reports including aggressive disease cases, in our early- emerges as a key step, not only for prediction modeling, but also stage CLL cohort, SF3B1 and NOTCH1 were mutually exclusive and for exploring mutation-specific therapeutic approaches and 40,41 a correlation between NOTCH1 and trisomy 12 was found. minimal residual disease monitoring. Previous studies have indicated that the prognostic significance of IgVH status is independent from that of classical clinical stages, 5,42 markedly in patients with early-stage disease. In our work, the CONFLICT OF INTEREST inclusion of high-throughput mutational status ousted IgVH from The authors declare no conflict of interest. the group of independent predictors, though it showed a trend for a shorter TTFT. ATM-mutated cases were found in a significant higher proportion among IgVH-unmutated cases. We did not find ACKNOWLEDGEMENTS a different distribution of the IgVH status between TP53-mutated This study was supported by a grant from Instituto de Salud Carlos III (PI13/02099) cases. These mutated cases can occur in both CLL IgVH subgroups. with participation of funds from FEDER (European Union). The late acquisition, autonomous of the emergence of the ancestral clone, of most TP53 mutations (80% are subclones in our work), might justify this lack of association. Rossi et al. REFERENCES recently reported a similar frequency of IgVH non-mutated cases 1 Abrisqueta P, Pereira A, Rozman C, Aymerich M, Gine E, Moreno C et al. Improving in TP53 subclonal and wild-type cases. survival in patients with chronic lymphocytic leukemia (1980-2008): the Hospital Ours is a translational study, with the focus shifted to clinical Clinic of Barcelona experience. Blood 2009; 114: 2044–2050. pragmatism. In that sense, the clonal or subclonal nature of the 2 Rawstron AC, Bottcher S, Letestu R, Villamor N, Fazi C, Kartsios H et al. Improving mutations is not as crucial as its detectability by Sanger. Given that efficiency and sensitivity: European Research Initiative in CLL (ERIC) update on the international harmonised approach for flow cytometric residual disease most of our cohort patients were lacking high-risk chromosomal monitoring in CLL. Leukemia 2013; 27: 142–149. alterations, we aimed to define the clinical impact of those 3 van den Broek EC, Kater AP, van de Schans SA, Karim-Kos HE, Janssen-Heijnen ML, variants with an allele frequency o20%. Of note, 10 out of 14 Peters WG et al. Chronic lymphocytic leukaemia in the Netherlands: trends in patients with a TP53 lesion could be reported herein only by the incidence, treatment and survival, 1989-2008. Eur J Cancer 2012; 48:889–895. use of next-generation sequencing. Most of ATM mutations were 4 Wierda WG, O'Brien S, Wang X, Faderl S, Ferrajoli A, Do KA et al. Multivariable detectable by conventional PCR technique, but the extent and model for time to first treatment in patients with chronic lymphocytic leukemia. lack of hotspots in this gene makes capillary Sanger sequencing as J Clin Oncol 2011; 29: 4088–4095. time-consuming and labor-intensive as to preclude its use in the 5 Molica S, Giannarelli D, Gentile M, Cutrona G, Di Renzo N, Di Raimondo F et al. routine praxis. These two aspects strongly favor the use of The utility of two prognostic models for predicting time to first treatment in early targeted sequencing in routine CLL workout. chronic lymphocytic leukemia patients: results of a comparative analysis. Leuk Res Surprisingly, β2-microglobulin levels, associated for the first 2013; 37:943–947. 6 Tomlinson IP, Novelli MR, Bodmer WF. The mutation rate and cancer. Proc Natl time to CLL 25 years ago, showed its predictive value even Acad Sci USA 1996; 93: 14800–14803. compared with the newer biological markers of intrinsic cell 7 Landau DA, Carter SL, Getz G, Wu CJ. Clonal evolution in hematological phenotypic and genomic features. Of note, ours is not the first malignancies and therapeutic implications. Leukemia 2014; 28:34–43. study to observe this strong role of β2-microglobulin in predicting 8 Dohner H, Stilgenbauer S, Benner A, Leupolt E, Krober A, Bullinger L et al. 23,44 outcome in this particular early-stage CLL subset. The reason Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med why a significant proportion of patients show high β2-microglobuline 2000; 343: 1910–1916. levels in the context of a low CLL burden disease still remains 9 Landau DA, Carter SL, Stojanov P, McKenna A, Stevenson K, Lawrence MS et al. unclear. A positive expression of CD38 correlated with a shorter Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. TTFT independently, a more expected finding, given that it Cell 2013; 152:714–726. showed an independent prognostic value (which ZAP70 and FISH 10 Puente XS, Pinyol M, Quesada V, Conde L, Ordonez GR, Villamor N et al. findings failed to reach) in one of the largest study of prognostic Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 2011; 475:101–105. markers carried out in Binet stage A patients with more than a 11 Quesada V, Ramsay AJ, Lopez-Otin C. Chronic lymphocytic leukemia with SF3B1 thousand patients an 8 years of median follow-up. mutation. N Engl J Med 2012; 366: 2530. We acknowledge some caveats in our study. Considering FISH 12 Wang L, Lawrence MS, Wan Y, Stojanov P, Sougnez C, Stevenson K et al. SF3B1 data as a loss of heterozygosity measure for VAF adjusting, can and other novel cancer genes in chronic lymphocytic leukemia. N Engl J Med lead to underestimation of VAF in those cases where 17p or 11q 2011; 365: 2497–2506. clones might be under the sensibility threshold established for 13 Rossi D, Khiabanian H, Spina V, Ciardullo C, Bruscaggin A, Fama R et al. Clinical that probe. It is true that VAF would change, and TP53 or ATM VAF impact of small TP53 mutated subclones in chronic lymphocytic leukemia. Blood estimates of o10% should be taken cautiously. However, those 2014; 123: 2139–2147. mutations would remain defined as subclonal, as none of our TP53 14 Austen B, Powell JE, Alvi A, Edwards I, Hooper L, Starczynski J et al. Mutations or ATM-mutated patients had a CLL population of o25%. in the ATM gene lead to impaired overall and treatment-free survival that is independent of IGVH mutation status in patients with B-CLL. Blood 2005; 106: In addition, it could be argued we did not sequenced all ATM 3175–3182. exons, and that we are underestimating ATM mutational status. 15 Rossi D, Bruscaggin A, Spina V, Rasi S, Khiabanian H, Messina M et al. Certainly, though we tried to design a cost-effective amplicon panel, Mutations of the SF3B1 splicing factor in chronic lymphocytic leukemia: anditcoversmostof ATM mutations reported in CLL, only an ATM association with progression and fludarabine-refractoriness. Blood 2011; 118: whole-exome sequencing study will address the precise clinical 6904–6908. impact of these gene variants. Finally, the lack of clinical grade 16 Baliakas P, Hadzidimitriou A, Sutton LA, Rossi D, Minga E, Villamor N et al. availability of deep-sequencing data precludes its use presently. The Recurrent mutations refine prognosis in chronic lymphocytic leukemia. Leukemia implementation will need to go through standardization of methods 2015; 29:329–336. and validation of prognostic value in clinical trials. 17 Hallek M, Cheson BD, Catovsky D, Caligaris-Cappio F, Dighiero G, Dohner H et al. In sum, our finding of an independent prognostic value of TP53 Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: mutations, not detectable by conventional techniques in a subset a report from the International Workshop on Chronic Lymphocytic Leukemia Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al updating the National Cancer Institute-Working Group 1996 guidelines. Blood 33 Damm F, Mylonas E, Cosson A, Yoshida K, Della Valle V, Mouly E et al. Acquired 2008; 111: 5446–5456. initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov 18 Dohner H, Stilgenbauer S, Fischer K, Bentz M, Lichter P. Cytogenetic and 2014; 4: 1088–1101. molecular cytogenetic analysis of B cell chronic lymphocytic leukemia: specific 34 Jethwa A, Hullein J, Stolz T, Blume C, Sellner L, Jauch A et al. Targeted rese- chromosome aberrations identify prognostic subgroups of patients and point to quencing for analysis of clonal composition of recurrent gene mutations in loci of candidate genes. Leukemia 1997; 11:S19–S24. chronic lymphocytic leukaemia. Br J Haematol 2013; 163:496–500. 19 Capello D, Guarini A, Berra E, Mauro FR, Rossi D, Ghia E et al. Evidence of biased 35 Cuneo A, Bigoni R, Rigolin GM, Roberti MG, Bardi A, Cavazzini F et al. immunoglobulin variable gene usage in highly stable B-cell chronic lymphocytic Late appearance of the 11q22.3-23.1 deletion involving the ATM locus in leukemia. Leukemia 2004; 18:1941–1947. B-cell chronic lymphocytic leukemia and related disorders. Clinico-biological 20 Marasca R, Vaccari P, Luppi M, Zucchini P, Castelli I, Barozzi P et al. Immuno- significance. Haematologica 2002; 87:44–51. globulin gene mutations and frequent use of VH1-69 and VH4-34 segments in 36 Austen B, Skowronska A, Baker C, Powell JE, Gardiner A, Oscier D et al. Mutation hepatitis C virus-positive and hepatitis C virus-negative nodal marginal zone B-cell status of the residual ATM allele is an important determinant of the lymphoma. Am J Pathol 2001; 159:253–261. cellular response to chemotherapy and survival in patients with chronic 21 Giudicelli V, Chaume D, Lefranc MP. IMGT/V-QUEST, an integrated software lymphocytic leukemia containing an 11q deletion. J Clin Oncol 2007; 25: program for immunoglobulin and T cell receptor V-J and V-D-J rearrangement 5448–5457. analysis. Nucleic Acids Res 2004; 32: W435–W440. 37 Agathanggelou A, Weston VJ, Perry T, Davies NJ, Skowronska A, Payne DT et al. 22 Damle RN, Wasil T, Fais F, Ghiotto F, Valetto A, Allen SL et al. Ig V gene mutation Targeting the Ataxia Telangiectasia Mutated-null phenotype in chronic lympho- status and CD38 expression as novel prognostic indicators in chronic lymphocytic cytic leukemia with pro-oxidants. Haematologica 2015; e-pub ahead of print 3 leukemia. Blood 1999; 94: 1840–1847. April 2015. 23 Gentile M, Cutrona G, Neri A, Molica S, Ferrarini M, Morabito F. Predictive value of 38 Weissmann S, Roller A, Jeromin S, Hernandez M, Abaigar M, Hernandez-Rivas JM beta2-microglobulin (beta2-m) levels in chronic lymphocytic leukemia since Binet et al. Prognostic impact and landscape of NOTCH1 mutations in chronic lym- A stages. Haematologica 2009; 94:887–888. phocytic leukemia (CLL): a study on 852 patients. Leukemia 2013; 27: 2393–2396. 24 Molica S, Giannarelli D, Levato L, Mirabelli R, Gentile M, Lentini M et al. 39 Mansouri L, Cahill N, Gunnarsson R, Smedby KE, Tjonnfjord E, Hjalgrim H et al. A prognostic algorithm including a modified version of MD Anderson Cancer NOTCH1 and SF3B1 mutations can be added to the hierarchical prognostic Center (MDACC) score predicts time to first treatment of patients with clinical classification in chronic lymphocytic leukemia. Leukemia 2013; 27:512–514. monoclonal lymphocytosis (cMBL)/Rai stage 0 chronic lymphocytic leukemia (CLL). 40 Schnaiter A, Paschka P, Rossi M, Zenz T, Buhler A, Winkler D et al. NOTCH1, SF3B1, Int J Hematol 2014; 100:290–295. and TP53 mutations in fludarabine-refractory CLL patients treated with alemtu- 25 Shanafelt TD, Kay NE, Jenkins G, Call TG, Zent CS, Jelinek DF et al. B-cell count and zumab: results from the CLL2H trial of the GCLLSG. Blood 2013; 122: 1266–1270. survival: differentiating chronic lymphocytic leukemia from monoclonal B-cell 41 Balatti V, Bottoni A, Palamarchuk A, Alder H, Rassenti LZ, Kipps TJ et al. NOTCH1 lymphocytosis based on clinical outcome. Blood 2009; 113: 4188–4196. mutations in CLL associated with trisomy 12. Blood 2012; 119: 329–331. 26 Call TG, Norman AD, Hanson CA, Achenbach SJ, Kay NE, Zent CS et al. Incidence of 42 Pepper C, Majid A, Lin TT, Hewamana S, Pratt G, Walewska R et al. Defining the chronic lymphocytic leukemia and high-count monoclonal B-cell lymphocytosis prognosis of early stage chronic lymphocytic leukaemia patients. Br J Haematol using the 2008 guidelines. Cancer 2014; 120: 2000–2005. 2012; 156:499–507. 27 Molica S, Giannarelli D, Mirabelli R, Levato L, Gentile M, Lentini M et al. Changes in 43 Simonsson B, Wibell L, Nilsson K. Beta 2-microglobulin in chronic lymphocytic the incidence, pattern of presentation and clinical outcome of early chronic leukaemia. Scand J Haematol 1980; 24:174–180. lymphocytic leukemia patients using the 2008 International Workshop on CLL 44 Oliveira AC, Fernandez de Sevilla A, Domingo A, De La Banda E, guidelines. Exp Rev Hematol 2014; 7: 691–695. Domingo-Domenech E, Mercadal S et al. Prospective study of prognostic factors 28 Hallek M, Fischer K, Fingerle-Rowson G, Fink AM, Busch R, Mayer J et al. Addition in asymptomatic patients with B-cell chronic lymphocytic leukemia-like lym- of rituximab to fludarabine and cyclophosphamide in patients with chronic phocytosis: the cut-off of 11 x 10(9)/L monoclonal lymphocytes better identifies lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet 2010; 376: subgroups with different outcomes. Ann Hematol 2015; 94:627–632. 1164–1174. 45 Rossi D, Fangazio M, Rasi S, Vaisitti T, Monti S, Cresta S et al. Disruption of BIRC3 29 Catovsky D, Richards S, Matutes E, Oscier D, Dyer MJ, Bezares RF et al. Assessment associates with fludarabine chemorefractoriness in TP53 wild-type chronic lym- of fludarabine plus cyclophosphamide for patients with chronic lymphocytic phocytic leukemia. Blood 2012; 119: 2854–2862. leukaemia (the LRF CLL4 Trial): a randomised controlled trial. Lancet 2007; 370: 46 Jebaraj BM, Kienle D, Buhler A, Winkler D, Dohner H, Stilgenbauer S et al. BRAF 230–239. mutations in chronic lymphocytic leukemia. Leuk Lymphoma 2013; 54: 30 Strati P, Keating MJ, O'Brien SM, Ferrajoli A, Burger J, Faderl S et al. Outcomes of 1177–1182. first-line treatment for chronic lymphocytic leukemia with 17p deletion. Haema- tologica 2014; 99: 1350–1355. 31 Farooqui MZ, Valdez J, Martyr S, Aue G, Saba N, Niemann CU et al. Ibrutinib This work is licensed under a Creative Commons Attribution 4.0 for previously untreated and relapsed or refractory chronic lymphocytic International License. The images or other third party material in this leukaemia with TP53 aberrations: a phase 2, single-arm trial. Lancet Oncol 2015; article are included in the article’s Creative Commons license, unless indicated 16: 169–176. otherwise in the credit line; if the material is not included under the Creative Commons 32 Malcikova J, Stano-Kozubik K, Tichy B, Kantorova B, Pavlova S, Tom N et al. license, users will need to obtain permission from the license holder to reproduce the Detailed analysis of therapy-driven clonal evolution of TP53 mutations in chronic material. To view a copy of this license, visit http://creativecommons.org/licenses/ lymphocytic leukemia. Leukemia 2015; 29:877–885. by/4.0/ Supplementary Information accompanies this paper on Blood Cancer Journal website (http://www.nature.com/bcj) Blood Cancer Journal http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Blood Cancer Journal Springer Journals

Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic lymphocytic leukemia

Loading next page...
 
/lp/springer-journals/prognostic-signature-and-clonality-pattern-of-recurrently-mutated-sjD4Uorfzk

References (99)

Publisher
Springer Journals
Copyright
Copyright
Subject
Biomedicine; Biomedicine, general; Cancer Research; Oncology; Hematology
eISSN
2044-5385
DOI
10.1038/bcj.2015.65
Publisher site
See Article on Publisher Site

Abstract

OPEN Citation: Blood Cancer Journal (2015) 5, e342; doi:10.1038/bcj.2015.65 www.nature.com/bcj ORIGINAL ARTICLE Prognostic signature and clonality pattern of recurrently mutated genes in inactive chronic lymphocytic leukemia 1 1 2 1 1 2 1 1 1 AM Hurtado , T-H Chen-Liang , B Przychodzen , C Hamedi , J Muñoz-Ballester , B Dienes , MD García-Malo , AI Antón , F de Arriba , 1 1 1 2 1 R Teruel-Montoya , FJ Ortuño , V Vicente , JP Maciejewski and A Jerez An increasing numbers of patients are being diagnosed with asymptomatic early-stage chronic lymphocytic leukemia (CLL), with no treatment indication at baseline. We applied a high-throughput deep-targeted analysis, especially designed for covering widely TP53 and ATM genes, in 180 patients with inactive disease at diagnosis, to test the independent prognostic value of CLL somatic recurrent mutations. We found that 40/180 patients harbored at least one acquired variant with ATM (n = 17, 9.4%), NOTCH1 (n = 14, 7.7%), TP53 (n = 14, 7.7%) and SF3B1 (n = 10, 5.5%) as most prevalent mutated genes. Harboring one ‘sub-Sanger’ TP53 mutation granted an independent 3.5-fold increase of probability of needing treatment. Those patients with a double-hit ATM lesion (mutation+11q deletion) had the shorter median time to first treatment (17 months). We found that a genomic variable: TP53 mutations, most of them under the sensitivity of conventional techniques; a cell phenotypic factor: CD38-positive expression; and a classical marker as β2-microglobulin, remained as the unique independent predictors of outcome. The high-throughput determination of TP53 status, particularly in this set of patients frequently lacking high-risk chromosomal aberrations, emerges as a key step, not only for prediction modeling, but also for exploring mutation-specific therapeutic approaches and minimal residual disease monitoring. Blood Cancer Journal (2015) 5, e342; doi:10.1038/bcj.2015.65; published online 28 August 2015 INTRODUCTION In this study, we propose to refine and apply a method for high- throughput targeted analysis of somatic recurrent mutations in Nowadays, an increasing numbers of patients are being diagnosed CLL, especially designed for covering widely TP53 and ATM genes. with early-stage chronic lymphocytic leukemia (CLL), likely owing Our main aim is to assess the independency of the prognostic to the use of routine blood tests for health screening and the 1–3 value of those variants, related to time to first treatment and widespread availability of flow cytometry. Among this subset of survival, in patients with CLL and no indication for therapy at CLL patients, most with a non-active disease and no treatment diagnosis. indication at baseline, different prognostic modeling approaches, incorporating traditional (clinical and laboratory), cytogenetic, immunophenotypic and immunoglobulin heavy-chain variable 4,5 MATERIALS AND METHODS region gen (IgVH) status, have been proposed. The contemporary vision of neoplasm development is based on Patients a consecutive acquisition of genetic changes with a selection and From 2006 to 2012, presentation bone marrow aspirates or blood samples 6,7 expansion of the more fit population. The heterogeneous DNA was collected during the diagnostic workout from 265 consecutive course of CLL is, to a certain extent, driven by the diverse CLL patients, after informed consent, according to the protocols approved by the Institutional Review Board of Hospital Morales Meseguer (EST-32/13) combinations of clones with acquired chromosomal lesions and 8,9 and with the Declaration of Helsinki. Patients who met criteria for an active somatic mutations. Whole-exome studies (WES) have shown disease at baseline, did not reach a minimum treatment-free follow-up of that the number of somatic variants per case is lower in CLL than 3 months, or nucleic acids did not pass the quality control for either IgVH that of those described in solid tumors and other leukemias, and status or targeted sequencing, were excluded (Figure 1). Diagnosis and 10–12 that the set of genes affected is discreet. These two aspects definition of active disease, requiring therapy, were achieved according to make this disease a suitable candidate for deep-targeted sequen- the International Workshop on Chronic Lymphocytic Leukemia established cing, a technique focused on distinct genomic sites, which also criteria. Time-to-first-treatment (TTFT) was measured from diagnosis to enables reliable detection of subclonal mutations because of to its date of first treatment. Regular follow-up consisted of blood cell counts and clinical examinations every 3 months the first year after diagnosis, and higher depth of coverage compared with WES. Using this more henceforth, visits were carried out from 3 to 6 months, depending on affordable and rapid strategy, recent studies have determined the patient risk. prognostic impact of TP53 subclones. In addition, the status of ATM, NOTCH1 and SF3B1, also recurrently mutated in CLL, have been 14–16 associated with impaired overall and treatment-free survival. Diagnostic workout Nevertheless, a comprehensive high-throughput sequencing study Every patient underwent a flow cytometry characterization with a panel of these variants, assessing their clinical relevance, in the context of including CD45, pan B-cell markers (CD19, CD20, CD22, CD79b, and surface both traditional and newer factors, is lacking. immunoglobulin light chains), markers for differential diagnosis with other 1 2 Hematology and Medical Oncology Department, Hospital Morales Meseguer, IMIB, Murcia, Spain and Traslational Hematology and Oncology Research, Cleveland Clinic, Cleveland, OH, USA. Correspondence: Dr A Jerez, Hematology and Medical Oncology Department, Hospital Morales Meseguer, IMIB, Marques de los Velez s/n, Murcia E-30008, Spain. E-mail: andjercay@gmail.com Presented in part at the 57th ASH Annual, San Francisco, CA, December 6–9, 2014. Received 18 May 2015; revised 17 June 2015; accepted 30 June 2015 Outcome and clonal pattern in inactive CLL AM Hurtado et al B-cell chronic lymphoproliferative diseases (CD5, CD23, FMC7, CD10, CD81, (2 × 250 bp) was performed with MiSeq v2.2 chemistry, and a mean depth CD103, CD25 and CD11c) and prognosis markers (CD38 and ZAP70) of 938 reads/base within the regions of interest was obtained. Raw data (Antibodies from BD Biosciences, San Jose, CA, USA). were analyzed with IlluminaonJboard Real Time Analysis (RTA v.2.4.60.8) Fluorescence in situ hybridization (FISH) analysis was performed on software and MiSeq Reporter. interphase nuclei at diagnosis from directly harvested peripheral blood or bone marrow samples according to the manufacturer’s protocol and using Variant call requirements and validation the following commercially available probes (Abbott Molecular, Des The following conditions were established for a variant to be called: (i) to Plaines, IL, USA): LSI MYB (6q23), LSI P53 (17p13.1)/ LSI ATM (11q22.3), be non-synonymous; (ii) not to be listed in dbSNP database (NCBI Human LSI D13S319 (13q14.3)/CEP12, as reported. A minimum of 400 nuclei Build 141); and (iii) a cutoff for any nucleotide position of 30 or more were scored for each probe or probe combination. variant reads and a Q score 430 (see Variant call requirements: technique Immunoglobulin heavy-chain variable diversity (D)-joining (J) rearrange- accuracy in Results). The filtered variant lists were manually reviewed and ments were amplified from either reverse-transcribed total RNA (preferred BAM file examined in Integrated Genome Viewer (Broad Institute). source) or genomic DNA. Purified amplicons were sequenced either Every variant with a clonal size of, at least, 20% and 430 variant reads directly or on subcloning. Sequences were aligned to the ImMunoGe- were bi-directionally sequenced using an ABI 3730 DNA Analyzer (Life neTics for computation of mutational load. Sequences were considered Technologies, Carlsbad, CA, USA). mutated or not using the cutoff of 2% mismatch. Six TP53-mutated cases were selected for applying the whole-amplicon panel on germline DNA (four cases from CD3+ sorted cells, two cases from Targeted sequencing buccal mucosal swab) to test its somatic nature. We designed a TruSeq Custom Amplicon panel (Illumina, Inc. San Diego, CA, USA) containing 13 genes and covering 28.099 bases (Table 1). For Statistical analysis some genes known mutation hotspots were targeted; and for those with a Comparisons of proportions and ranks of variables between groups were widespread localization of the lesions, the entire coding sequence was analyzed. The average amplicon size was 238 base pairs and ~ 99.1% performed by χ -test, Fisher's exact test, t-test or Mann–Whitney U-test, as of the regions were covered on both strands. Library preparation was appropriate. We used the Kaplan–Meier and the Cox method to analyse performed according to manufacturer's instruction. Paired-end sequencing overall survival (OS) and progression-free survival, with a two-sided P- value⩽ 0.05 considered to be significant. In Cox models, examination of log (− log) survival plots and partial residuals was performed to assess that the underlying assumption of proportional hazards was met. Thresholds of o2.4 mg/dl for β2-microglobulin, an absolute B-cell count of 11 ×10 /l or over, and higher LDH levels than the upper normal limit (4UNL (that is, 23–25 378 U/l), were chosen as reported elsewhere. RESULTS Testing technique accuracy by resequencing To estimate the accuracy of the technique to reproduce a variant call, 221 variants found in the first run were resequenced in depth (average reads per variant × 583). Eighteen of these variants had a clonal size over Sanger sensitivity (420%), 58 variants with a clonal size between 10 and 20%, and 145 variants below 10% of clone size. For this experiment, new libraries were built to capture and amplify exclusively the amplicons covering the variants selected, and the same genomic DNA used in the first run was used. Receiver operating characteristic curves for both number of Figure 1. Study flow diagram. Visual representation of the exclusion variant reads and variant allelic frequency (VAF) were created by criteria (left) and the targeted sequencing process pipeline (right). plotting the true positive rate (sensitivity) against the false-positive Table 1. Targeted NGS panel characteristics Target gen Length covered % Gen exome % CLL mutations Hotspot references (bp) sequenced covered* TP53 2724 100% 100% 12 10 ATM 7411 56% 78% (28 /36) Wang et al., Puente et al. 12 10 NOTCH1 1487 19% 100% (75/75) Wang et al., Puente et al. 12 10 SF3B1 3915 14% 92% (22/24) Wang et al., Puente et al. 12 10 MYD88 296 31% 100% (11/11) Wang et al., Puente et al. 12 10 POT1 549 28% 100% (5/5) Wang et al., Puente et al. 12 10 45 BIRC3 963 53% 100% /(11/11) Wang et al., Puente et al., Rossi et al. KRAS 567 100% 100% NRAS 570 100% 100% U2AF1 572 100% 100% 12 10 BRAF 357 14% 100% (2/2) Wang et al., Puente et al., Jebaraj et al. 12 10 BCOR 5268 100% 100% Wang et al., Puente et al. 12 10 SETBP1 3420 71% 100% (2/2) Wang et al., Puente et al. TOTAL 28099 Abbreviations: bp, base pair; CLL, chronic lymphocytic leukemia. Genes included, total gene length covered, percentage of gene exon sequenced and percentage of CLL mutations covered using as reference those variants described in the whole-exome studies referred in the last column. Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al rate (1-specificity) at various threshold settings. The number of SF3B1 and NOTCH1 mutations were mutually exclusive, and a variant reads in the initial run showed to be more accurate to significant correlation between NOTCH1 mutations and the predict the reproducibility of the variant in a second run, with an presence of a trisomy 12 was found (Po0.01). area under the curve of 0.894 (P ⩽ 0.001; 95% CI, 0.817-0.970), than the clonal size (VAF) (area under the curve: 0.613; P = 0.037; 95% Clonal diversity of recurrently mutated genes in early-stage CLL.As CI, 0.487–0.738; Supplementary Figures 1A and 1B). A cutoff of 30 samples used for this sequencing study belong to the diagnostic reads was chosen as threshold to consider a variant reproducible immunophenotypic workout, we could estimate the clonal or with a sensibility of 0.85 and a specificity of 0.945. No discrepancy subclonal nature of the acquired mutations, adjusting the variant was found when Sanger sequencing 15 selected variants over allele burden in non-sorted blood or bone marrow DNA for the 20% of clonal size. None of the 13 variants harbored by six percentage of the immunophenotypically quantified CLL popula- TP53-mutated cases was called when applying the panel on tion. In eight patients with a del11q or del 17p and a ATM or TP53 germline DNA. mutation, VAF was adjusted considering the loss of heterozygosity (Figure 3). Cohort and distribution of mutations. Two hundred and sixty-five Thirty-five mutations were estimated to be clonal, that is, patients were diagnosed in our center of a CLL from 2006 to 2012. present in the whole tumor population, either in a heterozygous Patients who, without needing therapy, did not achieve a (n = 28) or hemizygous (n = 7) configuration. The allelic ratio minimum of 3 months of follow-up, were not contemplated of other mutations showed that they appeared only in a fraction (Figure 1). No monoclonal lymphocytosis cases were considered. of tumor cells, indicating that they were secondarily acquired or The baseline characteristics of the 180 patients finally included subclonal (n = 33). Certain genes showed predominantly clonal were in accordance with their indolent-no need for treatment mutations (71% of NOTCH1, 71% of ATM variants), whereas others status at diagnosis, with 93% of the cohort assigned to Rai Stages were mainly subclonal (80% of TP53 variants, 100% of Ras-family 0 and I (Table 2). and U2AF1 genes). We found that 40/180 (22.2%) patients harbored at least one mutation; ATM (n = 17, 9.4%), NOTCH1 (n = 14, 7.7%), TP53 (n = 14, Clinical correlates. When confronting the 40 patients with, at 7.7%), SF3B1 (n = 10, 5.5%), BCOR (n = 3, 1.6%), BIRC3 (n = 2, 1.1%), least, one mutation, with the 140 non-mutated patients, we found KRAS (n = 2, 1.1%), U2AF1 (n = 2, 1.1%), POT1 (n = 2, 1.1%), less patients stratified as stage 0 in the mutated group, in favor of MYD88 (n = 1, 0.6%), and SETBP1 (n = 1, 0.6%). No somatic stages I and II. Focusing on FISH abnormalities, both 13q and +12 variants were identified for BRAF and NRAS (Figure 2). Sixty-eight shown to be more frequent in mutated cases, though none of mutations were detected in the whole cohort with 18 deletions these differences reached the statistical significance (Table 2). causing a frameshift, 1 non-frameshift deletion, 1 non-frameshift Likewise, no significant disparity was found when considering insertion and 48 missense single-nucleotide variants. Forty-one leukocyte, lymphocyte and platelet counts or hemoglobin, lactate out of 68 mutations were already reported to the Catalog of Somatic Mutations in Cancer (COSMIC; http://cancer.sanger.ac. dehydrogenase and β -microglobulin levels. In addition, no specific CD38 or ZAP70 expression pattern was observed to be uk/cancergenome/projects/cosmic), as human cancers variants (Supplementary Table 1). characteristic of each group. Table 2. Characteristics of patients included in the study at baseline and according to the presence or absence of, at least, a mutation Variable Total (n = 180) (A) Mutated cases (n = 40) (B) Non-mutated cases (n = 140) P (A vs B) Age, years, (mean± s.d.) 69 ±170± 12 68 ± 11 0.9 Sex, male/female,% (n) 59 (107)/41 (74) 67 (27)/33 (13) 60 (85)/40 (55) 0.3 Rai Stage, % 0 45 38 49 0.3 I 48 52 47 0.5 II 7 10 4 0.3 Leukocytes, x10 /l (median, IQR) 15.3 (10.9–26.9) 16.6 (10.9–24.2) 14.6 (10.7–24) 0.2 Lymphocytes, x10 /l (median, IQR) 10.3 (7–19.8) 12.2 (7.3–20.6) 9.8 (7–16) 0.3 B-cell count, x10 /l (median, IQR) 9.1 (6.1–17.3) 11.2 (7.4–19.1) 8.9 (6–15.6) 0.3 Hemoglobin, g/dl (median, IQR) 13.8 (12.2–14.7) 13.7 (12–14.5) 13.8 (12.3–14.7) 0.6 Platelets, x10 /l (median, IQR) 182 (136–217) 168 (126–216) 187 (136–217) 0.6 LDH IU/l (median, IQR) 368 (323–425) 377 (318–453) 324 (288–390) 0.4 B microglobulin, mg/l (median, IQR) 2.2 (1.8–3) 2.4 (1.8–3.6) 2.2 (1.7–3) 0.3 ZAP70 +, % (n) 42 (76) 50 (20) 40 (56) 0.1 CD38 +, % (n) 28 (50) 38 (15) 22 (30) 0.1 IgHV-unmutated status, % (n) 34 (61) 43 (17) 27 (37) 0.07 FISH abnormalities, % (n) Del(13q) 46 (84) 50 (20) 47 (65) 0.7 +12 18 (32) 25 (10) 14 (20) 0.2 Del(11q) 5 (8) 17 (7) 1 (0.7) 0.01 Del(17p) 2 (4) 10 (4) 2 (1.4) 0.02 Abbreviations: CD38, cluster of differentiation 38; FISH, fluorescence in situ hibridization; IgVH, immunoglobulin heavy-chain variable region gen; IQR, interquartile range; LDH, lactate dehydrogenase; ZAP70, Zeta-chain-associated protein kinase 70; +12, trisomy 12. Bold face: del(11q) and del (17p) as variables distributed differently among mutated or not mutated patients with statistical significance (P= 0.014) . Note: quantitative variables with a normal distribution are expressed as mean ± s.d. Quantitative variables not following the normal distribution are expressed as median and interquartile ranges. Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al Figure 2. Distribution of mutations and chromosomal aberrations. Figure 3. Estimated variant allele frequencies in the tumor fraction. The vertical axis represents allele frequency. Patients mutated are depicted in the abscissa. Color-coded circles for each gene affected. The variant allele frequency in non-sorted blood or bone marrow DNA was adjusted for the percentage of the immunophenotypically quantified CLL population. Despite the low number of 17p and 11q deletions in our non- hazards ratio (HR)=5.8;95% CI,3.1–10.9; Figure 4a) than those cases aggressive cohort, they were found mostly in the mutated group; without a detected mutation, and a shorter median OS (54 months the only difference statistically significant. vs not reached; P = 0.01; HR = 3.9; 95% CI, 2.2–6.9; Figure 4f). With a median follow-up of 54 months (interquartile range, TP53-mutated cases showed both a shorter TTFT (median 42–85 months), the median OS of the whole cohort has not been 29 months vs not reached; P ⩽ 0.001; HR = 5.3; 95% CI, 2.6–10.8 CI; reached. Forty-two patients (23.3%) required therapy. Median time Figure 4b) and OS (median 48 months; P ⩽ 0.001; HR = 3.7; 95% CI, to treatment for all patients has not been reached. Considering 1.9–7.2; Figure 4g) than the wild-type cases (median not reached). only those patients who were treated, median time to first The presence of a mutation in ATM predicted both for a shorter treatment was 48 months (range, 5–96 months). TTFT (mean 60 months; P ⩽ 0.001; HR = 5; 95% CI, 2.5–10.2 CI; Patients with, at least, one mutation had a worse time to first Figure 4c) and OS (median 61 months; P = 0.016; HR = 2.5; 95% CI, treatment (median TTFT: 60 months vs not reached; P⩽ 0.001; 1.2–5.2 CI; Figure 4h). Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al 1.0 TP53 WT cases (n=166, events=31) 1.0 1.0 WT cases (n=140, events=21) ATM WT cases (n=162, events=31) 0.8 0.8 0.8 0.6 0.6 0.6 Mutated cases ATM mutated cases (n=40, events=21) 0.4 0.4 0.4 (n=18, events=11) TP53 mutated cases 0.2 0.2 0.2 (n=14, events=11) p≤0.001 p≤0.001 0.0 0.0 p≤0.001 0.0 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120 TTFT (months) TTFT (months) TTFT (months) 1.0 1.0 1.0 SF3B1 WT cases (n=170 events=36) NOTCH1 WT cases (n=166, events=35) WT cases (n=140, events=31) 0.8 0.8 0.8 0.6 0.6 0.6 Mutated cases (n=40, events=23) NOTCH1 mutated cases 0.4 0.4 0.4 (n=14, events=7) SF3B1 mutated cases 0.2 0.2 0.2 (n=10, events=6) p=0.006 p≤0.001 0.0 0.0 0.0 p≤0.001 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 TTFT (months) TTFT (months) Overall Survival (months) 1.0 1.0 1.0 TP53 WT cases (n=166, events=42) NOTCH1 WT cases (n=166, events=46) ATM WT cases 0.8 0.8 0.8 (n=162, events=45) 0.6 0.6 0.6 ATM mutated cases 0.4 0.4 0.4 (n=18, events=9) TP53 mutated cases NOTCH1 mutated cases (n=14, events=12) (n=14, events=8) 0.2 0.2 0.2 p=0.01 p=0.016 0.0 0.0 0.0 p=0.001 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Overall Survival (months) Overall Survival (months) Overall Survival (months) 1.0 SF3B1 WT cases (n=170, events=49) 0.8 0.6 0.4 SF3B1 mutated cases (n=10, events=5) 0.2 0.0 p=0.03 0 20 40 60 80 100 Overall Survival (months) Figure 4. Differences in time to first treatment (TTFT) and survival outcomes (OS) in patients with, at least a mutation vs non-mutated (a, f); TP53 mutated or WT (b, g); ATM mutated or WT (c, h); NOTCH1 mutated or WT (d, i); and SF3B1 mutated or WT (e, j). P-values presented correspond to the Cox regression between the groups indicated. NOTCH1-mutated cases also presented a decreased TTFT (that is, variables strongly related to each other, measuring the (66 months; P = 0.006; HR = 3.1; 95% CI, 1.4–7.1; Figure 4d) and a same effect), precluded us from using them in this step of the shorter median OS (61 months; P = 0.01; HR = 2.7; 95% CI, 1.3–5.7; analysis. Mutational status of TP53, ATM, NOTCH1, SF3B1, IgVH status, the Figure 4i) than the wild-type cases (not reached). SF3B1-mutated presence of +12, del(13q), the Rai Stage, the positive expression of patients also showed worse prognosis both in terms of TTFT CD38 and ZAP70, β2-microglobulin and LDH serum levels, and a (median 44 months; P ⩽ 0.001; HR = 5.3, 95% CI, 2.2-12; Figure 4e) B-cell count ⩾ 11 × 10 /l were evaluated in a univariate Cox and of OS (P = 0.05; HR = 2.7; 95% CI, 1–7.8; Figure 4j) than the regression, both for TTFT and OS (Supplementary Tables 2A and B). wild-type cases (mean not reached). Only those variables with a P⩽ 0.05 were included in a multi- The low number of del(11q) and del(17p) cases, and the potential correlation of regressors with ATM and TP53 mutations variate analysis (Tables 3A and B). The presence of a somatic Blood Cancer Journal Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Cumulative Survival Outcome and clonal pattern in inactive CLL AM Hurtado et al DISCUSSION Table 3A. Multivariate Cox regression for time to first treatment In the last few years a boost in the number of patients being diagnosed with early-stage CLL, not requiring therapy at diagnosis Variable RR (95% CI) P-value 1–3 has been stated. In this subset, we have shown that TP53 high- TP53 3.9 (1.4–10.9) 0.01 throughput mutational status emerges as an independent IgVH 2 (0.9–4.4) 0.08 predictor, even when adjusting for the other recurrent gene NOTCH1 1.9 (0.7–5.7) 0.2 variants in CLL, and for both traditional and recently reported ATM 1.9 (0.7–4.6) 0.17 prognostic factors. Surprisingly, even including newer genomic +12 1.3 (0.6–2.7) 0.4 prognostic factors, a classical serologic parameter as β2-microglobulin CD38 2.3 (1–5.3) 0.04 levels, remained an independent predictor both for TTFT and OS, SF3B1 2.6 (0.8–8) 0.09 Rai Stage40 1.1 (0.5–2.3) 0.7 whereas a positive CD38 expression also independently predicted B-cell count ⩾ 10 × 10 /l 1.9 (0.9–3.9) 0.08 a shorter TTFT. β2-microglobulin ⩾ 2.4 2.4 (1.2–4.8) 0.02 The baseline characteristics of the patients included in our study were in accordance with their indolent-no need for treatment Abbreviations: CD38, cluster of differentiation 38; CI, confidence interval; status. As other groups have shown, the new definition of CLL, IgVH, immunoglobulin heavy-chain variable region gen; RR, relative risk; +12, trisomy 12. Bold face: variables reaching the independent statistical excluding those cases with a B-monoclonal population of o5000/ significance. ul, determined in our cohort a shift toward higher Rai Stages and a 26,27 higher rate of patients progressing and needing therapy. Patients with deletions on chromosome 17p respond worse to treatment than do those without it, resulting in early relapse and 28–30 Table 3B. Overall survival shorter survival. This cytogenetic lesion can be found in up to 50% of relapsed and refractory patients, but is rare at baseline Variable RR (95% CI) P-value (5–10%). We have found a 5% of TP53 mutations in a subset of patients where a ‘wait and watch’ therapeutical strategy is TP53 2.5 (1.2–6.4) 0.02 currently recommended. The relevant frequency and prognostic IgVH 1.5 (0.8–2.6) 0.2 impact of TP53 variants represented in o20% leukemic cells was β2-microglobulin ⩾ 2.4 1.9 (1.1–3.4) 0.04 recently reported by Rossi, et al. Of note, we show here that the NOTCH1 1.8 (0.7–4.5) 0.2 majority of TP53 mutations, at baseline in our inactive CLL subset, SF3B1 2.5 (0.9–7.2) 0.08 Rai Stage 40 1.1 (0.5–2.3) 0.7 were under the sensitivity of conventional sequencing, and that ATM 0.9 (0.4–2.5) 0.9 they kept its independent impact on prognosis even when adjusting by high-throughput detected ATM, NOTCH1 and SF3B1 Abbreviations: CD38, cluster of differentiation 38; CI, confidence interval; variants. Given the markedly short TTFT and OS of our TP53- IgVH, immunoglobulin heavy-chain variable region gen; RR, relative risk; mutated patients, it would seem an attractive approach to +12, trisomy 12. Bold face: variables reaching the independent statistical significance. consider investigational studies directed to eradicate these subclones, in which these patients could be treated. Closely related, Farooqui et al. reported a remarkable 2-year survival and variant in TP53, a positive CD38 antigen expression and β2- excellent drug-side-effects profile in the largest series of Ibrutinib microglobulin serum levels above 2.4 mg/dl, prevailed as inde- therapy in treatment-naive patients with CLL and 17p deletions. In pendent variables linked to a shorter time for needing treatment, addition, identification of these subclones emerges as crucial as a whereas only TP53 mutational status and β2-microglobulin levels concern is raised for clonal evolution and selection of resistant remained as significant predictors for OS. clones due the use of conventional (p53-dependent DNA damage) 9,32 We next sought to explore the impact of TP53 lesions below the chemotherapeutic agents. sensitivity of Sanger. That is, to define the role of high-throughput The frequency of ATM mutations in large series of CLL patients it sequencing in defining the prognosis in our cohort, as TP53 is not so well understood, mostly owing to the fact that its size mutations remained as the only independent genomic variable. In and the scattered distribution of its somatic mutations precludes 33,34 this sub-analysis we excluded five patients with a TP53 lesion the drafting of an amplicon-limited design. We found ATM to detectable by conventional techniques (direct sequencing and be the most frequently mutated gene, with a predominance of FISH/cytogenetics): one patient with a subclone harboring a TP53 clonal variants, indicating the ancestral-founding nature of these mutation and a del(17p), two patients with a clone with both a lesions. In addition, we found that the group with a shorter TTFT in TP53 and a del(17p), a case with a del(17p) in a clonal fashion but our study was defined by those patients with a double-hit ATM. not detectable mutation, and one patient with a clonal TP53 Likely, the reduced number of cases accounts for not reaching a mutation and 37% of variant reads. The other 10 subclonal cases statistical significance in multivariate or when addressing OS. Our would have been missed by Sanger sequencing as they called in finding supports the notion of a biological and clinical separation o20% reads (seven of them would have been missed even in of double-hit ATM cases, from those where 11q deletion and the B-cell-sorted DNA). We then replicated the multivariate Cox presence of an undamaged ATM allele give rise to a functional 35–37 regression analyses shown in Tables 3A and B. Harboring one of protein. these ten ‘sub-Sanger’ TP53 mutations granted an independent Recent studies have suggested an important prognostic role for 15,38,39 3.5-fold increase of probability of needing treatment during the NOTCH1 and SF3B1 in CLL. However, no study has course of the disease than a wild-type patient (P = 0.04; HR = 3.5; performed a multivariate analysis including traditional clinical 95% CI, 1–12.2), but it did not reach the significance for predicting and laboratory markers, flow cytometry factors, IgVH status and, at OS (P = 0.15; HR = 2; 95% CI, 0.8–5.5). least, the presence of ATM, TP53, NOTCH1 and SF3B1 variants. In Finally, a second sub-analysis showed that those patients with a addition, using high-throughput sequencing seems essential to double-hit ATM lesion (mutation+11q deletion) had the shorter discern the independent prognostic value of these mutations, as median TTFT reported in this study (17 months) strikingly reduced sub-Sanger mutations accounted for 25% of the variants in our compared with one-hit ATM patients (60 months). This impact was study. Though strongly associated with a worse outcome and TTFT not seen in multivariate analysis or concerning OS (Supplementary in univariate analysis, NOTCH1 and SF3B1 did not kept its Figure 2). significance when including TP53 and ATM mutations. Five out Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al of seven, and four out of five NOTCH1 and SF3B1-mutated cases, of patients lacking treatment indication at baseline, adds another which needed therapy, also harbored a TP53/del 17p and/or cobblestone to the positioning of amplicon deep-sequencing ATM/del11q lesion. This co-occurrence with more aggressive assays in established CLL diagnostics algorithms. The high- alterations can explain, in part, the lack of independent predictive throughput determination of TP53 status, particularly in this set value of SF3B1 and NOTCH1 in our cohort. Consistent with of patients frequently lacking high-risk chromosomal aberrations, previous reports including aggressive disease cases, in our early- emerges as a key step, not only for prediction modeling, but also stage CLL cohort, SF3B1 and NOTCH1 were mutually exclusive and for exploring mutation-specific therapeutic approaches and 40,41 a correlation between NOTCH1 and trisomy 12 was found. minimal residual disease monitoring. Previous studies have indicated that the prognostic significance of IgVH status is independent from that of classical clinical stages, 5,42 markedly in patients with early-stage disease. In our work, the CONFLICT OF INTEREST inclusion of high-throughput mutational status ousted IgVH from The authors declare no conflict of interest. the group of independent predictors, though it showed a trend for a shorter TTFT. ATM-mutated cases were found in a significant higher proportion among IgVH-unmutated cases. We did not find ACKNOWLEDGEMENTS a different distribution of the IgVH status between TP53-mutated This study was supported by a grant from Instituto de Salud Carlos III (PI13/02099) cases. These mutated cases can occur in both CLL IgVH subgroups. with participation of funds from FEDER (European Union). The late acquisition, autonomous of the emergence of the ancestral clone, of most TP53 mutations (80% are subclones in our work), might justify this lack of association. Rossi et al. REFERENCES recently reported a similar frequency of IgVH non-mutated cases 1 Abrisqueta P, Pereira A, Rozman C, Aymerich M, Gine E, Moreno C et al. Improving in TP53 subclonal and wild-type cases. survival in patients with chronic lymphocytic leukemia (1980-2008): the Hospital Ours is a translational study, with the focus shifted to clinical Clinic of Barcelona experience. Blood 2009; 114: 2044–2050. pragmatism. In that sense, the clonal or subclonal nature of the 2 Rawstron AC, Bottcher S, Letestu R, Villamor N, Fazi C, Kartsios H et al. Improving mutations is not as crucial as its detectability by Sanger. Given that efficiency and sensitivity: European Research Initiative in CLL (ERIC) update on the international harmonised approach for flow cytometric residual disease most of our cohort patients were lacking high-risk chromosomal monitoring in CLL. Leukemia 2013; 27: 142–149. alterations, we aimed to define the clinical impact of those 3 van den Broek EC, Kater AP, van de Schans SA, Karim-Kos HE, Janssen-Heijnen ML, variants with an allele frequency o20%. Of note, 10 out of 14 Peters WG et al. Chronic lymphocytic leukaemia in the Netherlands: trends in patients with a TP53 lesion could be reported herein only by the incidence, treatment and survival, 1989-2008. Eur J Cancer 2012; 48:889–895. use of next-generation sequencing. Most of ATM mutations were 4 Wierda WG, O'Brien S, Wang X, Faderl S, Ferrajoli A, Do KA et al. Multivariable detectable by conventional PCR technique, but the extent and model for time to first treatment in patients with chronic lymphocytic leukemia. lack of hotspots in this gene makes capillary Sanger sequencing as J Clin Oncol 2011; 29: 4088–4095. time-consuming and labor-intensive as to preclude its use in the 5 Molica S, Giannarelli D, Gentile M, Cutrona G, Di Renzo N, Di Raimondo F et al. routine praxis. These two aspects strongly favor the use of The utility of two prognostic models for predicting time to first treatment in early targeted sequencing in routine CLL workout. chronic lymphocytic leukemia patients: results of a comparative analysis. Leuk Res Surprisingly, β2-microglobulin levels, associated for the first 2013; 37:943–947. 6 Tomlinson IP, Novelli MR, Bodmer WF. The mutation rate and cancer. Proc Natl time to CLL 25 years ago, showed its predictive value even Acad Sci USA 1996; 93: 14800–14803. compared with the newer biological markers of intrinsic cell 7 Landau DA, Carter SL, Getz G, Wu CJ. Clonal evolution in hematological phenotypic and genomic features. Of note, ours is not the first malignancies and therapeutic implications. Leukemia 2014; 28:34–43. study to observe this strong role of β2-microglobulin in predicting 8 Dohner H, Stilgenbauer S, Benner A, Leupolt E, Krober A, Bullinger L et al. 23,44 outcome in this particular early-stage CLL subset. The reason Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med why a significant proportion of patients show high β2-microglobuline 2000; 343: 1910–1916. levels in the context of a low CLL burden disease still remains 9 Landau DA, Carter SL, Stojanov P, McKenna A, Stevenson K, Lawrence MS et al. unclear. A positive expression of CD38 correlated with a shorter Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. TTFT independently, a more expected finding, given that it Cell 2013; 152:714–726. showed an independent prognostic value (which ZAP70 and FISH 10 Puente XS, Pinyol M, Quesada V, Conde L, Ordonez GR, Villamor N et al. findings failed to reach) in one of the largest study of prognostic Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 2011; 475:101–105. markers carried out in Binet stage A patients with more than a 11 Quesada V, Ramsay AJ, Lopez-Otin C. Chronic lymphocytic leukemia with SF3B1 thousand patients an 8 years of median follow-up. mutation. N Engl J Med 2012; 366: 2530. We acknowledge some caveats in our study. Considering FISH 12 Wang L, Lawrence MS, Wan Y, Stojanov P, Sougnez C, Stevenson K et al. SF3B1 data as a loss of heterozygosity measure for VAF adjusting, can and other novel cancer genes in chronic lymphocytic leukemia. N Engl J Med lead to underestimation of VAF in those cases where 17p or 11q 2011; 365: 2497–2506. clones might be under the sensibility threshold established for 13 Rossi D, Khiabanian H, Spina V, Ciardullo C, Bruscaggin A, Fama R et al. Clinical that probe. It is true that VAF would change, and TP53 or ATM VAF impact of small TP53 mutated subclones in chronic lymphocytic leukemia. Blood estimates of o10% should be taken cautiously. However, those 2014; 123: 2139–2147. mutations would remain defined as subclonal, as none of our TP53 14 Austen B, Powell JE, Alvi A, Edwards I, Hooper L, Starczynski J et al. Mutations or ATM-mutated patients had a CLL population of o25%. in the ATM gene lead to impaired overall and treatment-free survival that is independent of IGVH mutation status in patients with B-CLL. Blood 2005; 106: In addition, it could be argued we did not sequenced all ATM 3175–3182. exons, and that we are underestimating ATM mutational status. 15 Rossi D, Bruscaggin A, Spina V, Rasi S, Khiabanian H, Messina M et al. Certainly, though we tried to design a cost-effective amplicon panel, Mutations of the SF3B1 splicing factor in chronic lymphocytic leukemia: anditcoversmostof ATM mutations reported in CLL, only an ATM association with progression and fludarabine-refractoriness. Blood 2011; 118: whole-exome sequencing study will address the precise clinical 6904–6908. impact of these gene variants. Finally, the lack of clinical grade 16 Baliakas P, Hadzidimitriou A, Sutton LA, Rossi D, Minga E, Villamor N et al. availability of deep-sequencing data precludes its use presently. The Recurrent mutations refine prognosis in chronic lymphocytic leukemia. Leukemia implementation will need to go through standardization of methods 2015; 29:329–336. and validation of prognostic value in clinical trials. 17 Hallek M, Cheson BD, Catovsky D, Caligaris-Cappio F, Dighiero G, Dohner H et al. In sum, our finding of an independent prognostic value of TP53 Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: mutations, not detectable by conventional techniques in a subset a report from the International Workshop on Chronic Lymphocytic Leukemia Blood Cancer Journal Outcome and clonal pattern in inactive CLL AM Hurtado et al updating the National Cancer Institute-Working Group 1996 guidelines. Blood 33 Damm F, Mylonas E, Cosson A, Yoshida K, Della Valle V, Mouly E et al. Acquired 2008; 111: 5446–5456. initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov 18 Dohner H, Stilgenbauer S, Fischer K, Bentz M, Lichter P. Cytogenetic and 2014; 4: 1088–1101. molecular cytogenetic analysis of B cell chronic lymphocytic leukemia: specific 34 Jethwa A, Hullein J, Stolz T, Blume C, Sellner L, Jauch A et al. Targeted rese- chromosome aberrations identify prognostic subgroups of patients and point to quencing for analysis of clonal composition of recurrent gene mutations in loci of candidate genes. Leukemia 1997; 11:S19–S24. chronic lymphocytic leukaemia. Br J Haematol 2013; 163:496–500. 19 Capello D, Guarini A, Berra E, Mauro FR, Rossi D, Ghia E et al. Evidence of biased 35 Cuneo A, Bigoni R, Rigolin GM, Roberti MG, Bardi A, Cavazzini F et al. immunoglobulin variable gene usage in highly stable B-cell chronic lymphocytic Late appearance of the 11q22.3-23.1 deletion involving the ATM locus in leukemia. Leukemia 2004; 18:1941–1947. B-cell chronic lymphocytic leukemia and related disorders. Clinico-biological 20 Marasca R, Vaccari P, Luppi M, Zucchini P, Castelli I, Barozzi P et al. Immuno- significance. Haematologica 2002; 87:44–51. globulin gene mutations and frequent use of VH1-69 and VH4-34 segments in 36 Austen B, Skowronska A, Baker C, Powell JE, Gardiner A, Oscier D et al. Mutation hepatitis C virus-positive and hepatitis C virus-negative nodal marginal zone B-cell status of the residual ATM allele is an important determinant of the lymphoma. Am J Pathol 2001; 159:253–261. cellular response to chemotherapy and survival in patients with chronic 21 Giudicelli V, Chaume D, Lefranc MP. IMGT/V-QUEST, an integrated software lymphocytic leukemia containing an 11q deletion. J Clin Oncol 2007; 25: program for immunoglobulin and T cell receptor V-J and V-D-J rearrangement 5448–5457. analysis. Nucleic Acids Res 2004; 32: W435–W440. 37 Agathanggelou A, Weston VJ, Perry T, Davies NJ, Skowronska A, Payne DT et al. 22 Damle RN, Wasil T, Fais F, Ghiotto F, Valetto A, Allen SL et al. Ig V gene mutation Targeting the Ataxia Telangiectasia Mutated-null phenotype in chronic lympho- status and CD38 expression as novel prognostic indicators in chronic lymphocytic cytic leukemia with pro-oxidants. Haematologica 2015; e-pub ahead of print 3 leukemia. Blood 1999; 94: 1840–1847. April 2015. 23 Gentile M, Cutrona G, Neri A, Molica S, Ferrarini M, Morabito F. Predictive value of 38 Weissmann S, Roller A, Jeromin S, Hernandez M, Abaigar M, Hernandez-Rivas JM beta2-microglobulin (beta2-m) levels in chronic lymphocytic leukemia since Binet et al. Prognostic impact and landscape of NOTCH1 mutations in chronic lym- A stages. Haematologica 2009; 94:887–888. phocytic leukemia (CLL): a study on 852 patients. Leukemia 2013; 27: 2393–2396. 24 Molica S, Giannarelli D, Levato L, Mirabelli R, Gentile M, Lentini M et al. 39 Mansouri L, Cahill N, Gunnarsson R, Smedby KE, Tjonnfjord E, Hjalgrim H et al. A prognostic algorithm including a modified version of MD Anderson Cancer NOTCH1 and SF3B1 mutations can be added to the hierarchical prognostic Center (MDACC) score predicts time to first treatment of patients with clinical classification in chronic lymphocytic leukemia. Leukemia 2013; 27:512–514. monoclonal lymphocytosis (cMBL)/Rai stage 0 chronic lymphocytic leukemia (CLL). 40 Schnaiter A, Paschka P, Rossi M, Zenz T, Buhler A, Winkler D et al. NOTCH1, SF3B1, Int J Hematol 2014; 100:290–295. and TP53 mutations in fludarabine-refractory CLL patients treated with alemtu- 25 Shanafelt TD, Kay NE, Jenkins G, Call TG, Zent CS, Jelinek DF et al. B-cell count and zumab: results from the CLL2H trial of the GCLLSG. Blood 2013; 122: 1266–1270. survival: differentiating chronic lymphocytic leukemia from monoclonal B-cell 41 Balatti V, Bottoni A, Palamarchuk A, Alder H, Rassenti LZ, Kipps TJ et al. NOTCH1 lymphocytosis based on clinical outcome. Blood 2009; 113: 4188–4196. mutations in CLL associated with trisomy 12. Blood 2012; 119: 329–331. 26 Call TG, Norman AD, Hanson CA, Achenbach SJ, Kay NE, Zent CS et al. Incidence of 42 Pepper C, Majid A, Lin TT, Hewamana S, Pratt G, Walewska R et al. Defining the chronic lymphocytic leukemia and high-count monoclonal B-cell lymphocytosis prognosis of early stage chronic lymphocytic leukaemia patients. Br J Haematol using the 2008 guidelines. Cancer 2014; 120: 2000–2005. 2012; 156:499–507. 27 Molica S, Giannarelli D, Mirabelli R, Levato L, Gentile M, Lentini M et al. Changes in 43 Simonsson B, Wibell L, Nilsson K. Beta 2-microglobulin in chronic lymphocytic the incidence, pattern of presentation and clinical outcome of early chronic leukaemia. Scand J Haematol 1980; 24:174–180. lymphocytic leukemia patients using the 2008 International Workshop on CLL 44 Oliveira AC, Fernandez de Sevilla A, Domingo A, De La Banda E, guidelines. Exp Rev Hematol 2014; 7: 691–695. Domingo-Domenech E, Mercadal S et al. Prospective study of prognostic factors 28 Hallek M, Fischer K, Fingerle-Rowson G, Fink AM, Busch R, Mayer J et al. Addition in asymptomatic patients with B-cell chronic lymphocytic leukemia-like lym- of rituximab to fludarabine and cyclophosphamide in patients with chronic phocytosis: the cut-off of 11 x 10(9)/L monoclonal lymphocytes better identifies lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet 2010; 376: subgroups with different outcomes. Ann Hematol 2015; 94:627–632. 1164–1174. 45 Rossi D, Fangazio M, Rasi S, Vaisitti T, Monti S, Cresta S et al. Disruption of BIRC3 29 Catovsky D, Richards S, Matutes E, Oscier D, Dyer MJ, Bezares RF et al. Assessment associates with fludarabine chemorefractoriness in TP53 wild-type chronic lym- of fludarabine plus cyclophosphamide for patients with chronic lymphocytic phocytic leukemia. Blood 2012; 119: 2854–2862. leukaemia (the LRF CLL4 Trial): a randomised controlled trial. Lancet 2007; 370: 46 Jebaraj BM, Kienle D, Buhler A, Winkler D, Dohner H, Stilgenbauer S et al. BRAF 230–239. mutations in chronic lymphocytic leukemia. Leuk Lymphoma 2013; 54: 30 Strati P, Keating MJ, O'Brien SM, Ferrajoli A, Burger J, Faderl S et al. Outcomes of 1177–1182. first-line treatment for chronic lymphocytic leukemia with 17p deletion. Haema- tologica 2014; 99: 1350–1355. 31 Farooqui MZ, Valdez J, Martyr S, Aue G, Saba N, Niemann CU et al. Ibrutinib This work is licensed under a Creative Commons Attribution 4.0 for previously untreated and relapsed or refractory chronic lymphocytic International License. The images or other third party material in this leukaemia with TP53 aberrations: a phase 2, single-arm trial. Lancet Oncol 2015; article are included in the article’s Creative Commons license, unless indicated 16: 169–176. otherwise in the credit line; if the material is not included under the Creative Commons 32 Malcikova J, Stano-Kozubik K, Tichy B, Kantorova B, Pavlova S, Tom N et al. license, users will need to obtain permission from the license holder to reproduce the Detailed analysis of therapy-driven clonal evolution of TP53 mutations in chronic material. To view a copy of this license, visit http://creativecommons.org/licenses/ lymphocytic leukemia. Leukemia 2015; 29:877–885. by/4.0/ Supplementary Information accompanies this paper on Blood Cancer Journal website (http://www.nature.com/bcj) Blood Cancer Journal

Journal

Blood Cancer JournalSpringer Journals

Published: Aug 28, 2015

There are no references for this article.