Letters to the editor 490 3 Patel JP, Gonen M, Figueroa ME, Fernandez H, Sun Z, Racevskis J et al. Prognostic phase chronic myeloid leukemia treated with imatinib. J Clin Oncol 2010; 28: relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 2761–2767. 2012; 366: 1079–1089. 10 Lin LI, Lin TC, Chou WC, Tang JL, Lin DT, Tien HF. A novel fluorescence-based 4 Falini B, Mecucci C, Tiacci E, Alcalay M, Rosati R, Pasqualucci L et al. Cytoplasmic multiplex PCR assay for rapid simultaneous detection of CEBPA mutations and nucleophosmin in acute myelogenous leukemia with a normal karyotype. N Engl J NPM mutations in patients with acute myeloid leukemias. Leukemia 2006; 20: Med 2005; 352: 254–266. 1899–1903. 5 Georgiou G, Efthymiou A, Vardounioti I, Boutsikas G, Angelopoulou MK, Vassila- 11 Dvorak P, Hruba M, Subrt I. Development of acute myeloid leukemia associated kopoulos TP et al. Development of acute myeloid leukemia with NPM1 mutation, with Ph-negative clone with inv(3)(q21q26) during imatinib therapy for chronic in Ph-negative clone, during treatment of CML with imatinib. Leukemia 2012; 26: myeloid leukemia. Leuk Res 2009; 33: 860–861. 824–826. 12 Fava C, Cortes J. Philadelphia-negative acute myeloid leukemia with new 6 Konoplev S, Yin CC, Kornblau SM, Kantarjian HM, Konopleva M, Andreeff M et al. chromosomal abnormalities developing after first-line imatinib treatment for Molecular characterization of de novo Ph þ acute myeloid leukemia. chronic phase chronic myeloid leukemia. Am J Hematol 2008; 83: 755. Leuk Lymphoma. e-pub ahead of print 9 July 2012; doi:10.3109/ 13 Kovitz C, Kantarjian H, Garcia-Manero G, Abruzzo LV, Cortes J. Myelodysplastic 10428194.2012.701739. syndromes and acute leukemia developing after imatinib mesylate therapy for 7 Piccaluga PP, Sabattini E, Bacci F, Agostinelli C, Righi S, Salmi F et al. Cytoplasmic chronic myeloid leukemia. Blood 2006; 108: 2811–2813. mutated nucleophosmin (NPM1) in blast crisis of chronic myeloid leukaemia. 14 Pawarode A, Sait SN, Nganga A, Coignet LJ, Barcos M, Baer MR. Acute myeloid Leukemia 2009; 23: 1370–1371. leukemia developing during imatinib mesylate therapy for chronic myeloid leuke- 8 Oki Y, Jelinek J, Beran M, Verstovsek S, Kantarjian HM, Issa JP. Mutations mia in the absence of new cytogenetic abnormalities. Leuk Res 2007; 31: 1589–1592. and promoter methylation status of NPM1 in myeloproliferative disorders. 15 Schafhausen P, Dierlamm J, Bokemeyer C, Bruemmendorf TH, Bacher U, Zander Haematologica 2006; 91: 1147–1148. AR et al. Development of AML with t(8;21)(q22;q22) and RUNX1-RUNX1T1 fusion 9 White DL, Dang P, Engler J, Frede A, Zrim S, Osborn M et al. Functional activity of following Philadelphia-negative clonal evolution during treatment of CML with the OCT-1 protein is predictive of long-term outcome in patients with chronic- Imatinib. Cancer Genet Cytogenet 2009; 189: 63–67.

Telomeres and chromosomal instability in chronic lymphocytic leukemia

Leukemia (2013) 27, 490–493; doi:10.1038/leu.2012.194 data identified groups with distinct cytogenetic and telomere profiles. This prospective study included 77 patients (57 males and 20 females, median age of 67 years) diagnosed with CLL at Telomeres are protective chromosomal end structures composed the Clermont-Ferrand University Hospital. Blood samples were of G-rich nucleotide repeats and an associated protein complex obtained after informed consent. At the time of sampling, 34 termed shelterin. Because of incomplete replication, telomeric patients were in Binet stage A, 17 in stage B and 26 in stage C. repeats are lost with every cell division, and this telomere attrition Samples were obtained before any treatment in 60 (77.9%) is involved in cell senescence and cancer (reviewed in Artandi and patients. Among the remaining 17 cases, 16 received no treatment DePinho1). Short, unprotected telomeres are erroneously fused within 6 months prior to the date of sampling. by the DNA-repair system, which gives rise to aneuploidies or Karyotype was analyzed after immunostimulation of cell structural chromosome rearrangements through chromosome cultures with the CpG-oligonucleotide DSP30 and interleukin 2. fusion-bridge-breakage cycles. Thus, telomere dysfunction can Interphase fluorescence in situ hybridization (FISH) was performed drive genomic instability during tumorigenesis.1 with a panel of commercially available probes (Abbott Molecular, In chronic lymphocytic leukemia (CLL), short telomeres have Rungis, France) for the detection of trisomy 12 and deletions of been associated with poor-prognosis cytogenetics.2,3 Besides 13q14, 11q22.3 (ATM) and 17p13 (p53). telomere length, the functional state of telomeres depends on Telomere studies were performed on mononuclear cells (MNCs). several molecular factors. Telomeres can be regenerated by Average telomere length was evaluated with quantitative real- telomerase and are protected from fusions by shelterin proteins. time DNA PCR.2 The method measures the difference between a Detailed investigations assessing all together the expression of sample and a reference normal DNA in the ratio of telomere telomerase and shelterin genes, telomere length, recurrent aberra- repeat copy number to single-gene copy number (relative T/S tions and overall karyotype instability have not yet been reported ratio). This T/S ratio is proportional to the average telomere length in CLL. Here, we studied the telomere status and chromosomal assessed with a classical telomere restriction fragment analysis aberrations in a series of CLL patients. A clustering analysis of the (r ¼ 0.89; Supplementary Figure S1).

Figure 1. (a–c) Hierarchical clustering of 77 CLL patients according to the pattern of chromosomal aberrations and telomere characteristics. Clustering was performed on the combined telomeres and cytogenetics data set (a) and separately: cytogenetics alone (b) and telomeres alone (c). Qualitative and quantitative values were normalized (centered and reduced) before clusterization. Aneuploidies, translocations or deletions, del17p or del11q, and trisomy 12 are presented in green (absence of abnormality) and red (presence of abnormality) squares. Presence of the low-risk del13q is shown in green, the absence in dark red. The different color intensities of dichotomic parameters reflect normalized values and not original binary (0/1) values. High normalized quantitative values of telomere length and expression of telomere- related genes are shown in red, intermediate values in black and lower values in green (intensity scale is shown). These normalized values, which are not logarithmic, are necessary to calculate distances between subjects and items. Distances between clusters were calculated using 1-Pearson’s correlation coefficient values and the dendrogram was constructed according to Ward’s algorithm. Of note, the average proportion of tumor cells in patient samples belonging to different clusters was essentially the same. For instance, the mean tumor cell content was 75% in cluster I, 75% in cluster II and 72% in cluster III (Kruskal–Wallis H-test, P ¼ 0.68, not significant). (d) The proportions of cases with negative prognostic factors in different clusters. P-values were obtained with the Chi-square test showing the overall difference between the clusters.

Accepted article preview online 13 July 2012; advance online publication, 10 August 2012

Leukemia (2013) 482 – 516 & 2013 Macmillan Publishers Limited Letters to the editor 491 The expression of hTERT (human telomerase reverse transcrip- ratio between transcript copy numbers of the target and control tase) and core members of the shelterin complex, TRF1 and TRF2 (GUS) genes multiplied by 100. The mean percentage of CLL cells (telomeric repeat binding factors 1 and 2) and POT1 (protection determined in MNC samples as the CD19 þ CD5 þ population was of telomeres 1) genes were quantified using real-time quantitative 74±19% (s.d.). To evaluate the possible impact of normal cell PCR.4,5 The normalized copy numbers (NCNs) were expressed as the contamination, we analyzed several samples (n ¼ 10) before and

Low High

Cytogenetics & Telomeres Aneuploidies +12 del(13q) Translocations or deletions del(17p) or del(11q) hTERT POT1 TRF1 TRF2 Telomere length

Cluster I (n=35) Cluster II (n=17) Cluster III (n= 25)

Cytogenetics

Aneuploidies +12 del(13q) Translocations or deletions del(17p) or del(11q)

Cluster 1 (n= 15) Cluster 2 (n= 29) Cluster 3 (n= 33)

Telomeres

POT1 TRF1 TRF2 Telomere length hTERT

Cluster 1* (n= 22) Cluster 2* (n= 26) Cluster 3* (n= 29)

Binet stage C UM IgVH ( 98%) 100 100 80 P= 0.16 P= 0.28 80 P= 0.006 P= 0.02 P= 0.11 P= 0.4 60 60 40 40 % of cases 20 % of cases 20 0 0 I II III 1 2 3 1* 2* 3* I II III 1 2 3 1* 2* 3* Clusters Clusters CD38 positivity ( 30%) Short LDT (< 6 M) 100 100

80 P= 0.006 P= 0.04 80 P= 0.02 P= 0.15 P= 0.7 P= 0.16 60 60 40 40 % of cases % of cases 20 20 0 0 I II III 1 2 3 1* 2* 3* IIIIII 123 1*2*3* Clusters Clusters

& 2013 Macmillan Publishers Limited Leukemia (2013) 482 – 516 Letters to the editor 492 after selection of CD19 þ cells. The mean values of CD19 þ cells ATM deficiency on telomere length was previously reported in CLL were 75.6±13.9% before and 98.2±2.4% after selection. We did cells.2,6 not find any statistically significant difference between the values Separate clustering of cytogenetic (Figure 1b) and telomeric data of telomere length and gene expression obtained before and after (Figure 1c) also showed the relationship between the telomere status selection (data not shown). and chromosomal aberrations. In particular, cytogenetic cluster 1 Unsupervised hierarchical clustering analyses on the basis of without any chromosomal aberration was normal in terms of either a combination of cytogenetic and telomeric characteristics telomere characteristics (Table 1). Cytogenetic cluster 2 displays or cytogenetic and telomeric parameters alone were performed to structural chromosomal changes, including 17p or 11q deletions, identify subclasses with distinct chromosome aberration and and telomere shortening, high hTERT and decreased shelterin telomeric profiles (Figures 1a–c). Karyotype and FISH data were gene-expression levels. Clustering of telomere parameters alone qualitative. Karyotypes were characterized by the presence or revealed an interesting group (cluster 3*) with deficient telomeres absence of aneuploidies, translocations or deletions. FISH patterns with the highest hTERT level, presenting high incidences of various were as follows: normal, del(13q), trisomy 12, del(17p) or del(11q). aberrations, such as aneuploidies (52%), translocations or deletions Single FISH aberrations were found in 48 cases and multiple (79%), del(17p) and del(11q) (55%) and a trisomy 12 (24%). aberrations in 10 cases. In nine cases, deletions of 17p or 11q were Unlike telomere length, the level of shelterin complex genes in accompanied either by a del(13q) or a trisomy 12. One case CLL was not previously analyzed in relation to chromosomal showed the presence of a trisomy 12 and a del(13q). Telomere instability. The total expression of shelterin genes was found to be characteristics were included as quantitative values: relative T/S lower in clusters harboring an increased rate of aberrations ratios for telomere length and NCN of hTERT, TRF1, TRF2 and POT1 (clusters III, 2 and 3*, Table 1). This reduced expression was also transcripts. The differences in the incidence of chromosomal associated with shorter telomere length. Shelterin downregulation aberrations and in the values of telomere parameters across the may either be secondary to telomere shortening or, in contrast, clusters were analyzed statistically (Table 1). may drive telomere dysfunction, thus compromising chromoso- Clustering of the combined data set highlighted three mal integrity.7 The present data do not allow us to examine this groups. Cluster I demonstrated good-prognosis cytogenetics issue, although some evidence exists. In particular, cluster 1* (isolated del(13q) or normal cytogenetics), long telomeres, characterized by long telomeres and low level of chromosome negative or low hTERT and a high expression of the shelterin aberrations contains a small subcluster with significantly complex genes (TRF1, TRF2 and POT1 values were similar to those decreased shelterin gene expression (seven last patients of of normal B cells, data not shown). In contrast, clusters II and III cluster 1*, Figure 1). Interestingly, this subcluster did not harbor displayed multiple chromosome aberrations, mostly aneuploidies more chromosome aberrations than the second subcluster (first in cluster II and structural chromosomal changes in cluster III. All 15 patients) characterized by higher shelterin levels (data not poor-prognosis del(17p) and del(11q) cases were in cluster III, shown). Given the small patient number, this result should be which also showed multiple karyotypic abnormalities (50% of taken with caution, but it suggests that the shelterin reduction cases) and translocations (72% of cases), the majority of may be insufficient for the acquisition of substantial chromosomal translocations (86%) being unbalanced. Clusters II and III were instability when telomeres are (still) long. Follow-up studies in characterized by a striking decrease in telomere length and such patients could help to investigate if a sustained shelterin shelterin gene-expression levels. Thus, rising chromosomal downregulation would lead to more rapid telomere erosion and/ instability was associated with telomeric deficiency. The hTERT or accumulation of aberrations. In addition, interactions with other gene was activated in clusters II and III, consistently with its role in shelterin complex members could have a role in chromosomal stabilization of rearranged tumor genomes.1 The results also instability. Studies of RAP1, TPP1 and TIN2 gene expression are highlight the relationship of 17p and 11q deletions with more needed to clarify the involvement of shelterin alterations in sever telomere and chromosome instability. These deletions lead chromosomal instability in CLL. to the loss of p53 and ATM genes, respectively, which are involved We also evaluated well-established factors predictive of CLL in monitoring telomere and DNA integrity. An impact of p53 and progression in patients from different clusters to determine if the

Table 1. Frequencies of chromosomal aberrations and values of quantitative telomere characteristics in different clusters

Cytogenetics and telomeres Cytogenetics Telomeres

Cluster I, Cluster II, Cluster III, Pa Cluster 1, Cluster 2, Cluster 3, Pa Cluster 1*, Cluster 2*, Cluster 3*, Pa n ¼ 35 n ¼ 17 n ¼ 25 n ¼ 15 n ¼ 29 n ¼ 33 n ¼ 22 n ¼ 26 n ¼ 29

Chromosomal aberrations Aneuploidies 0.0% 94.1% 24.0% o10 À 6 0.0% 34.5% 36.4% 0.02 18.2% 11.5% 51.7% o10 À 2 Translocations 0.0% 29.4% 96.0% o10 À 6 0.0% 86.2% 12.1 o10 À 6 13.6% 11.5% 79.3% o10 À 5 or deletions Del(17p) or 0.0% 5.9% 64.0% o10 À 6 0.0% 51.7% 6.1% o10 À 4 0.0% 3.8% 55.2% o10 À 4 del(11q) þ 12 0.0% 70.6% 4.0% o10 À 6 0.0% 0.0% 39.4% o10 À 4 9.1% 15.4% 24.1% 0.35 Del(13q) 57.1% 35.3% 52.0% 0.33 0.0% 62.0% 63.6% o10 À 3 50.0% 57.7% 44.8% 0.64

Telomere parametersb Telomere length 1.6±0.1 0.9±0.1 0.6±0.1 o10 À 6 1.7±0.1 0.7±0.1 1.2±0.1 o10 À 6 1.6±0.1 1.2±0.1 0.6±0.04 o10 À 6 (T/S ratio) hTERT (NCN) 4.4±3.1 90±23 88±17 o10 À 5 0.0 96±17 56±9 o10 À 4 9.0±4.8 9.7±5.8 118±16 o10 À 6 TRF1 (NCN) 145±997±984±9 o10 À 3 142±10 83±8130±10 o10 À 3 114±9 145±12 88±8 o10 À 3 TRF2 (NCN) 193±12 183±21 153±16 0.11 192±14 155±15 192±14 0.15 133±8 235±14 161±15 o10 À 4 POT1 (NCN) 195±12 131±14 132±14 o10 À 2 208±16 130±12 165±13 o10 À 3 141±12 216±14 125±12 o10 À 4 Total shelterinc 534±26 413±36 369±31 o10 À 3 542±34 369±27 488±31 o10 À 3 389±26 596±27 376±27 o10 À 6 (NCN) Abbreviations: hTERT, human telomerase reverse transcriptase; NCN, normalized copy number; POT1, protection of telomeres 1; TRF, telomeric repeat binding factor. aAberration frequencies were compared using the Chi-square test and quantitative telomere characteristics using the Kruskal–Wallis H-test bMean values with s.e. cTotal shelterin gene expression (TRF1 þ TFR2 þ POT1).

Leukemia (2013) 482 – 516 & 2013 Macmillan Publishers Limited Letters to the editor 493 obtained classification may be clinically significant. Figure 1d 1CHU Clermont-Ferrand, Cytoge´ne´tique Me´dicale, shows the incidence of negative prognostic features, such as Binet Clermont-Ferrand, France; C stage, unmutated (UM) immunoglobulin variable region genes 2Clermont Universite´, Universite´ d’Auvergne, (IgVH), CD38 positivity (X30%) and short (o6 months) lympho- Faculte´ de Me´decine, Histologie Embryologie Cytoge´ne´tique, cyte doubling time (LDT), in different clusters. No significant Clermont-Ferrand, France; relationship was found with Binet C stage. In contrast, clusters 3EA 4677 ERTICa, Universite´ d’Auvergne, with enhanced telomere and chromosome instability showed Clermont-Ferrand, France; higher incidences of classical biological factors of poor prognosis. 4CHU Clermont-Ferrand, He´matologie Clinique Adulte, Of note, the strongest associations with UM IgVH and CD38 Clermont-Ferrand, France; positivity were found for clusters obtained by combining 5EA 7283 CREaT, Universite´ d’Auvergne, telomeric and cytogenetic parameters, whereas the links with Clermont-Ferrand, France; telomeric clusters were weaker but still significant and, finally, the 6Inserm U823, Institut Albert Bonniot, associations with cytogenetic clusters were not significant. Only Universite´ Joseph Fourier, Grenoble, France; the combined model showed a significant relationship with short 7CHU Grenoble, Laboratoire de Ge´ne´tique LDT. Cytogenetics and telomeres are known to represent valuable Onco-he´matologique, Grenoble, France; prognostic factors in CLL,2 and perhaps the combinations of both 8Centre Jean Perrin, Clermont-Ferrand, France and characteristics could provide a more accurate prediction of disease 9Service d’He´matologie Biologique, Hoˆpital Pitie´-Salpeˆtrie`re, progression. Our study was prospective, and so far we were not Paris, France able to compare survival of patients from different clusters. This E-mail: [email protected] issue can be addressed in a large cohort of CLL patients with adequate follow-up. REFERENCES In conclusion, the identification of groups with distinct 1 Artandi SE, DePinho RA. Telomeres and telomerase in cancer. Carcinogenesis 2010; telomeric and cytogenetic profiles supports the link between 31: 9–18. telomere status and genomic instability in CLL. Telomere short- 2 Roos G, Kro¨ber A, Grabowski P, Kienle D, Bu¨hler A, Do¨hner H et al. Short telomeres ening, changes in shelterin and hTERT gene expression coincided are associated with genetic complexity, high-risk genomic aberrations, and short with increasing levels of chromosome aberrations, and this was survival in chronic lymphocytic leukemia. Blood 2008; 111: 2246–2252. 3 Brugat T, Nguyen-Khac F, Grelier A, Merle-Be´ral H, Delic J. Telomere dysfunction- particularly affected by p53 or ATM gene deletions. Theses CLL induced foci arise with the onset of telomeric deletions and complex chromosomal groups are associated with established prognostic factors and may aberrations in resistant chronic lymphocytic leukemia cells. Blood 2010; 116:239–249. be clinically significant. 4 Tchirkov A, Chaleteix C, Magnac C, Vasconcelos Y, Davi F, Michel A et al. hTERT expres- sion and prognosis in B-chronic lymphocytic leukemia. Ann Oncol 2004; 10: 1476–1480. 5 Poncet D, Belleville A, t’kint de Roodenbeke C, Roborel de Climens A, Ben Simon E, CONFLICT OF INTEREST Merle-Beral H et al. Changes in the expression of telomere maintenance genes The authors declare no conflict of interest. suggest global telomere dysfunction in B-chronic lymphocytic leukemia. Blood 2008; 111: 2388–2391. 6 Britt-Compton B, Lin TT, Ahmed G, Weston V, Jones RE, Fegan C et al. Extreme LVe´rone`se1,2,3, O Tournilhac4,5, M Callanan6,7, N Prie1,2,3, 3,8 1,2,3 6,7 9 telomere erosion in ATM-mutated and 11q-deleted CLL patients is independent of F Kwiatkowski , P Combes , M Chauvet , F Davi , disease stage. Leukemia 2012; 26: 826–830. 1,2,3 5,8 4,5 1,2,3 L Gouas , P Verrelle , R Guie`ze , P Vago , 7 Cookson JC, Laughton CA. The levels of telomere-binding proteins in human 4,5 1,2,3,8 JO Bay and A Tchirkov tumours and therapeutic implications. Eur J Cancer 2009; 45: 536–550.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Partial 17q gain resulting from isochromosomes, unbalanced translocations and complex rearrangements is associated with gene overexpression, older age and shorter overall survival in high hyperdiploid childhood acute lymphoblastic leukemia

Leukemia (2013) 27, 493–496; doi:10.1038/leu.2012.198 present in 9% of high hyperdiploid ALL as detected by single- nucleotide polymorphism (SNP) array analysis.4 Surprisingly, the majority of these 17q gains did not seem to arise from i(17q). We High hyperdiploid acute lymphoblastic leukemia (ALL) is asso- have now further characterized rearrangements with breakpoints ciated with a B-cell precursor (BCP) immunophenotype, low white in chromosome 17, screened for TP53 mutations, and investigated blood cell count, age between 3 and 5 years, and a favorable the gene expression in high hyperdiploid pediatric BCP ALL prognosis.1 It is characterized by 51–67 chromosomes, with (Table 1; Supplementary Table 2). Cases were included if they had nonrandom trisomies and tetrasomies. Interestingly, partial gains uniparental disomy (UPD) or trisomy (UPT) involving 17p, had a of the common extra chromosomes are rare,1 with the exception 17p deletion, or gain of 17q not associated with trisomy 17, as of chromosome 17, which is trisomic in 75% of cases and forms an ascertained by SNP array analysis. For cases 1–11, which were isochromosome resulting in 17q gain and 17p loss in an additional included in a series of 74 high hyperdiploid patients,4 SNP array 2–5% of cases.2,3 We recently reported that partial gains of 17q is results have been previously published. Diagnostic bone

Accepted article preview online 17 July 2012; advance online publication, 24 August 2012

& 2013 Macmillan Publishers Limited Leukemia (2013) 482 – 516