Translocation t(14;18) is not associated with inferior outcome in chronic lymphocytic leukemia

Translocations, in particular, unbalanced subtypes, were suggested to carry a poor prognostic effect in chronic lymphocytic leukemia (CLL).1, 2 Balanced translocations, especially those involving immunoglobulin (IG) genes, are recurrent, but uncommon, in this disorder.2, 3, 4 Their significance remains poorly understood.5 In published series,4, 6 part of the cases have unknown partner genes. Recurrent partner genes include BCL2, BCL3, BCL11A and CMYC. The best described is the t(14;19) involving the BCL3 gene, often associated with atypical morphology, unmutated VH genes and inferior prognosis.4, 6 The t(14;18)(q32;q21) and its variants, the t(2;18) (p11;q21) and t(18;22)(q21;q11), lead to overexpression of the anti-apoptotic bcl2 protein. It is interesting that, even in the absence of t(14;18), bcl2 overexpression is a constant finding in CLL, and is not explained by BCL2 gene hypomethylation. Other mechanisms, such as deletion or downregulation of miR–15a and miR–16–1, correlate inversely with bcl2 expression. Hence, the pathogenetic role, if any, of IGH/BCL2 translocations in disease initiation and evolution of CLL remains hypothetical. In most studies, CLL cases with translocations involving IGH are analyzed as one group.3, 6 However, the target gene that becomes overexpressed as a result of the translocation may be relevant for the outcome. Here, the demographic features, clinical evolution, morphological and immunophenotypic characteristics of 40 consecutive patients, all showing an IG/BCL2 translocation, were analyzed as were the cytogenetic and molecular features. These patients were referred for cytogenetic characterization of CLL to the Centers for Human Genetics of the University Hospital Leuven, Gosselies (Belgium) and Le Kremlin–Bicêtre (France) between 1990 and 2008. Inclusion criteria were morphological diagnosis of B-CLL and typical immunophenotype, defined as a Matutes–Catovsky score 3 and presence of t(14;18), t(2;18) or t(18;22) by conventional cytogenetics, fluorescent in situ hybridization (FISH) and/or molecular analysis. Clinical data were provided by the referring clinicians. Forty patients fulfilled the inclusion criteria, representing less than 2% of the CLL cases who were referred during the study period. Population characteristics were comparable with those of CLL patients in general; males (ratio M/F: 28/12) were predominantly affected with a median age of 63 years (range 38–83 years) at the time of primary diagnosis, and of 67 years (range 40–88 years) at the time of the detection of the translocation. Staging data were available in 38 patients, 28 presented with early stage (Binet A/Rai 0 or 1), seven with intermediate stage (Binet B/Rai 2), and three with advanced disease (Binet C/Rai 3 or 4) (Table 1). Laboratory data at the time of diagnosis were available for 32 patients. Median lymphocytosis was 13.7 109/l (range 1.4–123.4 109/l). LDH level was slightly elevated (that is, above the upper normal value of 333 IU/l) in 12 patients, and median LDH was 331.5 IU/l (range 126–558 IU/l). Anemia and thrombocytopenia were rare; six male patients presented with hemoglobin level below 13 g/dl, whereas only one female patient presented with hemoglobin level below 12 g/dl. Of these, none were below 10 g/dl. In three patients, platelet counts were below 100 109/l. Reports on morphology at diagnosis were available for 39 patients. All indicated a classical morphology (typical CLL), except for five cases with atypical morphology. Additionally, morphology was independently reviewed on available stored peripheral blood or marrow smears, by two cytologists (PM and BC). Forty-four samples from 29 patients and 44 samples from control cases (CLL, n=39; aCLL, n=2; follicular lymphoma (FL), n=3) were analyzed in a blinded and randomized sequence. The following morphology was observed; typical (n=24), atypical (n=1), normal morphology in a sample taken after therapy (n=1). In one patient, peripheral blood was morphologically compatible with FL. However, in the latter patient, peripheral blood and bone marrow aspirate before detection of the t(14;18) were suggestive of CLL. Bone marrow trephine at that time was compatible with a small-cell non-Hodgkin's lymphoma, possibly atypical CLL. Moreover, the Matutes–Catovsky score was 5/5 and there was no expression of CD10. This case was therefore considered as CLL and not as FL. Two patient samples could not be evaluated, because of poor quality. Immunophenotype was typical for CLL in all 40 cases, showing a Matutes–Catovsky score of at least 3 (Table 1). Only two cases were CD5 negative. CD10 expression was determined in 34 cases and was negative. CD38 was negative in 21 and positive (that is, expressed 30%) in seven cases. Partial CD38 expression was seen in two cases (Table 1, cases 14 and 19). ZAP-70 was not routinely determined, except in case 23, in which expression was negative. Thus, FL could be excluded in all patients. Standard conventional chromosome analysis was carried out on peripheral blood or bone marrow lymphocytes cultured for 72 h. Detailed cytogenetic characteristics of the patients are shown in a Supplementary Table. Thirty-seven patients showed a t(14;18) or variant in the karyotype. In one the t(14;18) was observed in 1/5 metaphases only, but was confirmed by FISH. In three additional patients the t(14;18) or variant was identified by molecular analysis and/or FISH (Supplementary Table). The translocation most frequently involved IGH (n=32), but variant cases involving IGK or IGL were observed in three and five cases, respectively. In three cases the translocation was unbalanced. FISH confirmed the t(14;18) or variant translocation in 36 patients. There was no archived material for FISH in four patients. In 14 cases, the t(14;18) or variant was the sole aberration. In six cases with clonal evolution, the t(14;18) was the only aberration in the stemline. The translocation was associated with one additional change in 19 cases, and two or more additional changes in four patients, respectively. Recurrent associated changes were trisomy 12 (n=14) and del(13q) (n=5). It was noted in two cases that the del(13q) was present in a complex karyotype. Other recurrent abnormalities included del(6q) (n=2), and loss of 17p (n=2). FISH detected additional aberrations missed by conventional cytogenetic analysis in 12 cases (del(13q), n=11; del(17p), n=1). None of the cases showed del(11q). In two patients (cases 4 and 39), the t(14;18) or variant presented as a subclonal change, with trisomy 12 as the primary change. In case 4, a follow-up karyotype (not shown in Supplementary Table) showed 47,XY,+12[6]/47,idem,t(14;18)(q32;q21)[3]/46,XY[17]. This finding and earlier observations suggest that at least in some cases, the t(14;18) may represent a secondary or 'late' aberration, which is probably not important for disease onset. The latter hypothesis is further supported by the high incidence of clonal evolution in CLL.1 Recently, clonal evolution was reported to selectively occur in patients with unmutated CLL. VH was unmutated in case 4 and mutated in case 39 (Table 1). VH mutational status and family usage was determined in 25 patients, and D and J gene usage in 24 patients. This analysis failed in another 11 cases because the V–J sequences generated by the PCR could not be adequately interpreted. VH was mutated in 23 and unmutated in two cases (Table 1). In 25 patients, the following four VH families were used: VH3 in 19, VH5 in four (all 5.51), VH4 and VH1 each in one case. None of the patients showed VH3.21 usage. D3–10 was used in four patients, D2–8 in three, and D1–26, D3–9, D4, D4–17 and D6–19 in two patients each. The seven remaining patients all showed a different D usage. J4 was used in 11 cases, J6 in six, J3 in three, J5 in three and J1 in one case, respectively. There were no recurrent VDJ combinations. Follow-up data were available in 36 patients over a median period of 43.5 months (range 1–161 months) from the date of detection of the BCL2 translocation and of 92.5 months (range 1–261 months) from the date of clinical diagnosis (Table 1). Decisions for treatment were based on standardized criteria of the National Cancer Institute in each of the patients treated. Twenty-two patients (61.1%) required therapy because of progressive lymphocytosis with an increase of >50% over 2 months, or a lymphocyte doubling time of <12 months (n=13), lymphadenopathy (n=10), thrombocytopenia (n=8), anemia of central origin or auto-immune hemolytic anemia (n=5), B-symptoms (n=3), splenomegaly (n=2) and/or symptomatic cutaneous involvement (n=1). Two or more indications for therapy were present in 15 patients. Treatment-free survival (TFS) of the entire population was 48 months. First-line treatment consisted of chlorambucil (n=10), chlorambucil/prednisone (n=8), fludarabine (n=2), and other, that is, cyclophosphamide/doxorubicin/vincristine/prednisone (CHOP) (n=1), and cyclophosphamide/doxorubicin/teniposide/prednisone (n=1). Fifteen patients achieved response (partial response in 12, and complete response in three patients) to first-line therapy. Only three patients did not respond to first-line therapy; however, all of them responded well to subsequent therapies. In two patients a second-line therapy had to be administered because of chlorambucil intolerance. In two patients, no response data were available. Four patients died because of disease 48–192 months after initial diagnosis and 14–132 months after detection of the t(14;18) or variant. Six patients died because of unrelated disease (lung cancer (n=3), myelodysplasia with concomitant breast cancer (n=1), myelofibrosis (n=1), squamous cell carcinoma (n=1)), whereas 26 patients are still alive and four have been lost to follow-up. Overall survival (OS) and 0.95 confidence intervals (CI) were estimated by the Kaplan–Meier method from the date of clinical diagnosis and from the date of detection of the t(14;18)/variant until death because of any cause or until the last follow-up. Survival probability at 5 years from the date of clinical diagnosis and from the date of genetic detection of the t(14;18)/variant was 91% (CI 0.76–0.97) and 77% (CI 0.61–0.88), respectively. Median OS from the date of clinical diagnosis and from the date of genetic detection of the t(14;18) was 182 months (range 1–261) and 106 months (range 1–161), respectively. The estimated median OS for patients with versus without associated del(13q) was 174 and 182 months, respectively (P=0.39, not significant), and for patients with versus without associated trisomy was 12, 174 and 182 months, respectively (P=0.39, not significant). In this series, the t(14;18) was not associated with advanced disease; indeed, most patients (n=28/38) presented with early stages (Binet A/Rai 0–1). Although therapy was required in 61.1% of the patients, the median TFS was 48 months and all patients responded well to first-line (n=15) and/or subsequent therapy. This contrasts with earlier studies reporting a TFS of only 2 months in patients with 14q32/IGH translocations (n=18)6 and a progression-free survival (PFS) of 20.6 months in patients with IGH/BCL2 translocations (n=8).5 However, another study showed different PFS and time to treatment (TTT) according to the partner gene. Indeed, translocations involving BCL3 were correlated with a worse prognosis than those involving BCL2 with a median TTT of 1.5 and 3.6 years, respectively (P=0.017).4 Five-year survival probability from clinical diagnosis and from cytogenetic detection of the t(14;18) was 91 and 77%, respectively. This is higher, compared with the reported 5-year survival rate of 75.9% documented by the US SEER registry (http://seer.cancer.gov/statfacts/html/clyl.html). Moreover, the median OS of our cohort is longer than the median OS of the prognostically most favorable subgroup reported by Döhner et al., that is, the group with isolated del(13q) (182 vs 133 months).7, 8 Our findings also contrast with the data reported by Juliusson3 and Cavazzini et al.,6 who described shorter survival in patients with 14q abnormalities in general. However, the series of Juliusson probably included lymphoproliferative disorders other than CLL, as immunophenotypic inclusion criteria were not very stringent. In the cohort of Cavazinni et al., only 5 of 18 patients showed an IGH/BCL2 rearrangement, a finding that could have had an influence on the general outcome. Cytogenetic complexity (that is, the presence of at least three abnormalities) and loss of 17p/TP53, generally associated with poor prognosis, were uncommon in the present cases, whereas aberrations correlated with intermediate (+12) and good prognosis (13q-), were more prevalent.7, 8 In our series, there seems to be an association between the t(14;18)/variant and trisomy 12, since the prevalence of this chromosomal abnormality was higher than that observed in CLL in general (35 versus 16%, respectively).8 Prognostically unfavorable unmutated VH status and VH3.21 usage were uncommon or not observed, respectively, in this cohort. Thus, there could be an association between the t(14;18)/variant and mutated VH status, as the observed prevalence was higher compared with CLL in general (92% versus 40–70%, respectively). In conclusion, in this series, CLL with t(14;18) was not associated with clinically or biologically aggressive features and did not negatively affect outcome. We thank the following clinicians and clinical biologists for providing clinical data and material for morphological review: D Bauwens, T Braeckevelt, G Bries, D Brohée, C Chatelain, A Delannoy, A De Weweire, C Doyen, C Dubois, F Forget, O Hamdam, L Harlet, MF Louis, B Maes, S Marichal, P Mineur, S Neven, A Nijs, V Peeters, P Pierre, J Verschuere, L Waumans. We gratefully acknowledge D. Sartenaer, M. Stul and P. Vannuffel for excellent technical assistance and R Siebert for kindly providing probes for FISH. We also are indebted to A. Ferrant and A. Hagemeijer for critical reading of this manuscript. N Put is supported by Fonds voor Wetenschappelijk Onderzoek (FWO) Vlaanderen—Research Foundation Flanders. P Vandenberghe is a senior clinical investigator of FWO Vlaanderen. Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)