(2003) 17, 255–259  2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00 www.nature.com/leu BIO-TECHNICAL METHODS SECTION (BTS) BTS Leukemia

Is t(14;18)(q32;q21) a constant finding in follicular ? An interphase FISH study on 63 patients A Godon1,3, A Moreau2, P Talmant1, L Baranger-Papot3, F Genevie`ve3, N Milpied4, M Zandecki3 and H Avet-Loiseau1

1Hematology Laboratory , University Hospital, Nantes, France; 2Pathology Department, University Hospital, Nantes, France; 3Hematology Laboratory, University Hospital, Angers, France; and 4Clinical Department, University Hospital, Nantes, France

The translocation (14;18)(q32;q21) is the hallmark of follicular cytogenetics is the best way to detect karyotypic changes in lymphoma (FL). However, conventional cytogenetics and PCR tumor samples but several limitations, including the need of techniques fail to detect it in at least 10% of cases. In order to evaluate the true incidence of this translocation in FL, we ana- fresh samples, lead to demonstrating t(14;18) in only 75% to 3 lyzed 63 patients with FL, and 17 patients with diffuse large cell 85% of FL. Interphase fluorescence in situ hybridization lymphoma (DLCL) corresponding to suspected FL transform- (FISH) is an alternative for the diagnosis of cytogenetic ations using interphase fluorescence in situ hybridization changes, and several sets of probes have been described to (FISH). Colocalized signals related to the translocation were assess this translocation.6–8 Using these assays, bcl-2 = observed in 19–92% of cells (median 51%), corresponding to rearrangements were observed in 85% to 100% of the cases. positivity over the threshold in all (63/63) cases. Similarly, 16/17 possibly secondary DLCL displayed the translocation. Thus, the exact incidence of t(14;18) in FL remains unknown. Although some cytogenetic changes might be missed by this In this report, we developed a colocalization-based FISH FISH assay (such as rare insertion, or translocations with other assay using a new set of probes, hybridizing to the centrom- chromosomal partners), our results stress t(14;18)(q32;q21) as ericpart of IgH, and to the telomericregion of the bcl-2 loci, an almost constant finding in FL. Our sensitive interphase FISH respectively, in order to detect t(14;18) whatever the breakpo- assay should be of great value to define FL more accurately, ints within the bcl-2 gene. This assay was applied to frozen namely in patients included into therapeutic trials. Further- more, this approach could be of interest in (re)defining some samples from 63 histologically proven FL. Because a part of types of FL, especially the grade 3 FL which frequently lack diffuse large (DLBCL) are reported as t(14;18). resulting from a preceding FL phase,9,10 we tested these probes Leukemia (2003) 17, 255–259. doi:10.1038/sj.leu.2402739 in a limited number of DLBCL with a known or highly sus- Keywords: interphase FISH; follicular lymphoma; pected previous history of FL, to determine the accuracy of t(14;18)(q32;q21); bcl-2; diffuse lymphoma that assay in such instances.

Introduction Materials and methods

Follicular lymphoma (FL) accounts for 20–25% of all non- Tissue samples Hodgkin’s lymphomas (NHL)1 and is recognized as one single entity in the World Health Organization (WHO) classi- Retrospective interphase FISH study was performed on frozen fication.2 The hallmark of FL is a translocation involving the samples (−80°C) from 63 FL and 17 DLBCL stored at the tissue bcl-2 gene (Bcell leukemia/lymphoma 2 ) and the IgH libraries of the Pathology Departments from Nantes and (immunoglobulin heavy chain) gene, leading to Angers (France). Samples corresponded to biopsies of t(14;18)(q32;q21) and the overproduction of BCL2 protein, lymph nodes, , parotids, , , bone that prevents cells from undergoing apoptosis. marrow or peripheral blood, in 60, 2, 2, 1, 1, 9 and 5 Besides , various methods have been proposed to cases, respectively. ascertain diagnosis of FL. Overexpression of BCL2 protein is Diagnosis of FL (n = 63) was performed according to the detected in Ͼ90% of cases using ,3 but WHO classification criteria, including morphology it is far from being specific for FL.3,4 Also, breakpoints at bcl- (predominantly follicular pattern, without or with some diffuse 2 locus are not always located within either the major break- areas, cleaved and noncleaved follicle center cells) and point region (MBR) or the minor cluster region (mcr), and PCR immunophenotype. MonotypicIg chainexpression, methods using such primers cannot demonstrate molecular negativity for CD5, positivity for CD20 (or CD19) and for rearrangement in all the cases, 10–15% of the cases lacking CD10 was needed to include every case. Flow cytometry was any specific amplification with these primers.5 Conventional performed at diagnosis on suspended cells (peripheral blood, , or disruption of one part of the ) to allow determination of tumor involvement in 40 cases. All Correspondence: H Avet-Loiseau, Laboratoire d’He´matologie, Institut de Biologie, 9 quai Moncousu, 44093 Nantes Cedex 1, France; Fax: cases that met the above mentioned diagnostic criteria and (33) 240 08 41 14 that had sufficient frozen material were included in the study. Received 16 January 2002; accepted 2 July 2002 For DLBCL (n = 17), diffuse growth pattern, or/ t(14;18)(q32:q21) in follicular lymphoma A Godon et al 256 and immunoblasts, and immunophenotype of B cell lineage milliliters of 1X PBS (phosphate buffer saline), in order to were used as diagnosis criteria. Preceding history of FL or part obtain an homogeneous cell suspension. After incubation in of the tumor sample showing remnant architecture of FL an hypotonicsolution (KCl 0.075 M/37 °C/30 min), cells were (hypothesizing evolution from a previously undiagnosed FL) fixed (methanol/cold acetic acid, 3 vol/1 vol) and dropped on were present in 12 and five cases, respectively. slides. Nuclei were denaturated for 3 to 5 min in formamide Reactive lymph nodes (10 cases), bone marrow and periph- 70%/2 X SSC, pH 7, and then dehydrated in gradient alcohol eral blood mononuclear cells from normal donors (10 cases) baths. Denatured probes (10 ␮l) were dropped on the slides, were used as negative controls. covered with a coverslip and sealed with rubber cement. An overnight hybridization was performed at 37°C in a humid chamber. After washes for 5 min at 73°C in 2 X SSC baths DNA probe description and a quick rinse for 2 min in 2 X SSC 0.1% Triton bath, nuclei were counterstained with DAPI and mounted in an The first set of probes corresponded to cos Ig10 and PAC antifade solution. Slides were analyzed on an epifluorescence 210c12, labeled by nick translation with SpectrumGreen- microscope (Axioplan 2; Zeiss, Jena, Germany), and hybridiz- dUTP and SpectrumOrange-dUTP, respectively (Vysis, Vois- ation signals were evaluated on 100 to 200 nuclei on each ins-le-Bretonneux, France). Cos Ig10 (T Rabbits; MRC, Cam- slide. bridge, UK) hybridizes the C␣2 and C⑀ IgH sequences, located at the centromeric border of the IgH gene on , as previously reported.11 PAC 210c12 maps at the 5’ end R banding analysis and PCR of the bcl-2 gene, downstream from the FVT-1 locus12 on , as previously published6 (Figure 1). Cytogenetic analysis was performed in 21 patients, according The second set of probes used was LSI IGH/BCL2 Dual to Bastard et al13 Karyotypes were described according to the Color, Dual Fusion Translocation Probe (Vysis), a mixture of International System for Human GeneticNomenclature(ISCN, SpectrumGreen-labeled LSI IGH and Spectrum Orange-lab- 1995). Thirty-three patients were studied at diagnosis for PCR- eled LSI BCL2 probes. LSI IGH contains sequences homolo- MBR rearrangement, as previously described.14 All patients gous to the whole IgH locus, and sequences extending about corresponded to FL at diagnosis. 300 kb beyond the 3’ end. LSI BCL2 covers the whole bcl-2 gene region and sequences extending approximatively 250 kb both distal and proximal to the gene. Cos Ig10, PAC 210c12, Results LSI IGH and LSI BCL2 probes span approximately 60 kb, 120 kb, 1.5 Mb and 750 kb, respectively. PCR and R banding analysis Centromeric(a satellite) CEP 18 and CEP 8 DNA probes (Vysis) were used in some cases, in order to better ascertain PCR demonstrated a MBR-bcl2 rearrangement in 19/33 FL or polyploidy in some patients. samples (58%) (mcr not tested). Cytogenetic study was perfor- med in 21 cases and demonstrated t(14;18) in most (at least 80%) or in a few (5 to 40%) metaphases analyzed in 14 and Fluorescence in situ hybridization three cases, respectively. One case showed 18q− only (14q+ undetected), one case (AJ) presented with a complex kary- Interphase FISH was performed on nuclei isolated from uncul- otype, as follows: 50,XX,der(1)t(1;2)(p36;?),del(6)(q22),+7,der tured frozen biopsies or suspended cells, as follows. After fast (11)t(11;12)(p15;q12), der(12) t(1;12)(q25;p13),+del(12)(q thawing, each biopsy was mechanically disrupted in a few 21),+14,add(18)(q21),+21 [15], and the last two samples presented with a normal karyotype.

Interphase FISH

Controls: In non-malignant nuclei, the expected pattern was the same with both sets of probes: two Orange and two Green separated signals (2O2G pattern). Using cos Ig10 and PAC 210c12 probes, t(14;18) led to the colocalization of one Orange and one Green signal leading to a yellow (or Orange/Green) fusion signal (1O1G1F pattern). Using LSI IGH/BCL2 probes, the most common pattern was as follows : one Orange signal, one Green signal representing the normal chromosomal homologs, respectively, and two Orange/Green (yellow) fusion signals, corresponding to both derivative chro- mosomes (1O1G2F pattern). In normal patients and reactive lymph nodes some fusion signals were observed, correspond- ing to random colocalization of bcl-2 and IgH regions. Using Ig10/PAC210c12 probes, an average of 6.7% nuclei showed one fusion signal, and the threshold to ascertain translocation was defined over 11% (mean + 3 s.d.)-positive cells. Using Figure 1 The IgH/bcl-2 molecular rearrangement. Breakpoint clus- LSI IGH/BCL2 probes, an average of 0.5% nuclei showed two ters are indicated (arrows), as well as sequences hybridizing the cos fusion signals, and threshold was defined over 1%. Depending Ig10/PAC 210c12 set of probes. on the level of condensation of DNA and on the cycle phase,

Leukemia t(14;18)(q32:q21) in follicular lymphoma A Godon et al 257 Table 1 Atypical patterns observed using Ig10/210c12 and/or LSI IGH/BCL2 probes

n Cos Ig10 / PAC LSI IGH / BCL2 (Vysis) 210c12

Atypical pattern observed with both sets of probes Triploidy 2 2O 2G 1F 2O 2G 2F Tetraploidy 2 3O 3G 1F 3O 3G 2F or 2O 2G 2F or 2O 2G 4F +18 2 2O1G1F 2O1G2F i (18q) 1 2O 1G 1F 2O 1G 2F Atypical pattern observed only with LSI IGH/BCL2 (Vysis) isolated der(14)t(14;18) 1 1O 1G 1F 2O 1G 1F +der(18) 1 1O 1G 1F 1O 1G 3F

a few cells could exhibit up to three or four signals of the blood and nine bone marrow samples, ranging from 19% to same color: random colocalization led to two fusion signals 100% (median 52%). Results are reported in Table 2. in some of such cells. In DLBCL, among the 17 cases tested using Ig10/210c12 Using Ig10/210c12 probes, 63 of the 63 (100%) FL tested probes, only one presented with colocalized signals below the presented with at least one fusion signal in a number of cells threshold (one fusion signal in 8% of cells), whereas the 16 over the threshold (19 to 92% (mean 51%)). In 56/63 cases, other specimens presented with the classical 1O1G1F pattern the expected pattern corresponding to 1O, 1G, and 1F signal, (25 to 85% of nuclei (mean 54%)). The negative sample was respectively, was observed. In seven cases, atypical patterns tested using the LSI IGH/BCL2 probes and was also negative. could be observed, corresponding to an increased number of O and G signals: the use of centromeric probes (and karyotype if performed) led to an explanation in each instance (Table 1). Discussion All samples demonstrating atypical patterns were selected and studied using the LSI IGH/BCL2 probes. The translocation (14;18)(q32;q21) is the hallmark of FL, dis- Using LSI IGH/BCL2 (Vysis) probes, 39/40 FL tested covered in up to 85% of patients after conventional cytog- presented with two colocalized fusion signals over the thres- eneticstudy. 3,9 If breakpoints at 14q32 always lay within IgH hold (8 to 90% (mean 49%)). The expected 1O1G2F pattern J(D) segments,16 breakpoints are much more scattered at was observed in 30 cases and atypical patterns in the other 18q21: although mainly located within MBR or mcr regions nine cases (Table 1). Comparison of the results obtained using (50–75% and 5–25% of cases, respectively),9,10 various other both sets of probes (40 cases) demonstrated discrepancies in breakpoints, scattered between MBR and mcr, or at the 5’ side two cases (Table 2). In one case LSI IGH/BCL2 failed to detect in VCR (variant cluster region) or FVT-1 (follicular variant any abnormality although MBR rearrangement was demon- translocation) have been reported.12,17,18 The interphase FISH strated after PCR study. In the last case, the translocation was assay that we have developed here uses a set of probes ascertained using both Ig10/210c12 and LSI IGH/BCL2 hybridizing to DNA sequences located at distance from all probes, but the respective numbers of positive nuclei differed breakpoints reported in the literature to date: PAC 210c12, (30% and 8%). In that latter case, clear-cut positivity was mapping to the telomericregion (or 5’) of all the 18q21 break- observed using Ig10/210c12 probes whereas green signals points, and Ig10, mapping to the centromeric part (or 3’) of corresponding to IgH hybridization of Vysis probe were very all reported 14q32 breakpoints (Figure 1). This set of probes faint. As V(D)J rearrangements (within the variable region) or allowed us to ascertain t(14;18) in 63/63 (100%) FL tested. class-switch deletions (within the constant region) may occur Conventional cytogenetic study was performed in only one on one or both the translocated or non-translocated IgH part of our cases, but results were in agreement with literature alleles in FL,15 the green probe signal intensity from the rel- data (eg 85% positive cases).3,9 Several explanations dealing evant IgH region (on one or both alleles) may be diminished with normal karyotypes in FL have been suggested, including as a result of the LSI IGH probe target . complex karyotypes masquerading the translocation (case AJ), Flow cytometry using CD10 and CD19 allowed determi- fully balanced cryptic translocations (under the resolution of nation of tumor infiltration in 26 lymph nodes, five peripheral banding), low tumor cell infiltration, or low proliferative index

Table 2 Comparison of results obtained in 40 FL using both cos Ig10/PAC 210c12 and LSI IGH/BCL2 sets of probes

CosIg10/PAC210c12 LSI IGH/ BCL2 (Vysis) Estimated level of tumor PCR (No. of positive nuclei) (No. of positive nuclei) involvement using FCMa (MBR primer) n = 38 ++19–100% + or − (19–92%)b (20–90%)b FM 30% 8% 36% MBR+ (low green signal intensity) CG 25% 0% 29% MBR+ aPositivity was defined as the number of cells expressing CD10 and CD20 (or CD19) after flow cytometry (see text for details). b% corresponds to the number of cells demonstrating fusion signals.

Leukemia t(14;18)(q32:q21) in follicular lymphoma A Godon et al 258 of tumor cells.3,7 The latter two hypotheses cannot be ruled include patients into clinical trials: our interphase FISH assay out in some of our cases, and we found a good relationship could be used in order to ascertain histological diagnosis of between the number of cells exhibiting positive fluorescent FL, namely for those cases negative after cytogenetic or signal and the number of tumor cells infiltrating the sample, molecular studies. That will certainly lead to define more in each case studied using flow cytometry (table 2). accurately patients into prospective therapeutic trials. We observed two cases with normal karyotype although tumor infiltration was high (80% and 85% tumor cells). As molecular study failed to detect MBR rearrangement in either References case, this led to question about the presence of t(14;18). How- ever, in both cases, interphase FISH assays (using Ig10/210c12 1 The Non-Hodgkin’s Lymphoma Classification Project. A clinical and LSI IGH/BCL2 probes) were positive in a number of cells trial of the International Lymphoma Study Group classification of similar to tumor infiltration rate. That led to hypothesize in non-Hodgkin’s lymphoma. Blood 1997; 89: 3909–3918. both cases low tumor cell proliferation rate, although a cryptic 2 Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, translocation could not be fully ruled out, but the lack of Vardiman J, Lister TA, Bloomfield CD. The World Health Organi- zation Classification of Neoplasms of the Hematopoietic and associated cytogenetic abnormalities does not support the Lymphoid Tissues: Report of the Clinical Advisory Commitee hypothesis of FL with a cryptic translocation. Meeting – Airlie House, Virginia, November, 1997. The Hematol- LSI IGH/BCL2 dual fusion translocation probe (Vysis) also ogy Journal 2000; 1: 53–56. detects t(14;18) in FL, but comparison with our set of probes 3 Macintyre E, Willerford D, Morris SW. Non-Hodgkin’s Lym- demonstrated discrepancies in two of 40 cases tested (Table phoma: molecular features of B Cell Lymphoma. Education Pro- 2). In one instance, discrepancy was related to a difference in gram Book of the American Society of Hematology. American Society of Hematology: San Francisco, 2000, pp 180–204. the respective numbers of cells exhibiting positive signal, LSI 4 Skinnider BF, Horsman DE, Dupuis B, Gascoyne RD. Bcl-6 and IGH/BCL2 probes underestimating the level of the tumor cell Bcl-2 protein expression in diffuse large B-cell lymphoma and fol- infiltration as compared to both results using our probes and licular lymphoma: correlation with 3q27 and 18q21 chromosomal immunological methods. In one instance, low intensity of abnormalities. Hum Pathol 1999; 30: 803–808. fluorescent signal corresponding to the IgH probe might cor- 5 Lopez-Guillermo A, Cabanillas F, McDonnell TI, McLaughlin P, respond to a large deletion within the IgH loci, as reported.15 Smith T, Pugh W, Hagemeister F, Rodriguez MA, Romaguera JE, Younes A, Sarris AH, Preti HA, Lee MS. Correlation of bcl-2 Other interphase FISH assays were reported to study FL, and rearrangement with clinical characteristics and outcome in indol- a segregation-based interphase FISH assay using two probes ent follicular lymphoma. Blood 1999; 93: 3081–3087. flanking bcl-2 breakpoint regions was described by Vaand- 6 Vaandrager JW, Schuuring E, Raap T, Philippo K, Kleiverda K, rager et al.6 Thirty-five out of 35 (100%) samples with bcl-2 Kluin P. Interphase FISH detection of BCL2 rearrangement in fol- rearrangement were detected. As the partner chromosome licular lymphoma using breakpoint-flanking probes. Genes Chro- could not be identified using that interphase FISH assay, the mosomes 2000; 27: 85–94. 7 Poetsch M, Weber-Matthiesen K, Plendl HJ, Grote W, Schlegel- same authors proposed either the study of metaphase nuclei berger B. Detection of the t(14;18) chromosomal translocation by or to develop a three-color FISH assay, using two probes interphase cytogenetics with yeast-artificial-chromosome probes flanking the breakpoint region on one chromosome and a in follicular lymphoma and nonneoplastic lymphoproliferation. J third one for the partner chromosome, both proposals difficult Clin Oncol 1996; 14: 963–969. to apply in routine conditions. Using YAC yA153A6, spanning 8 Akiko T, Ikuo M, Shinsuke I, Shohei Y, Shigeo H, Kazuhiro N, the whole bcl-2 region, and J1.3, hybridizing within the C Hiroshi F, Shigeo N, Masao S, Ryuzo U, Masafumi T. Interphase detection of immunoglobulin heavy chain gene translocations region of the IgH gene, a colocalization-based interphase with specific loci in 173 patients with B-cell lymphoma. 7 FISH assay was described by Poetsh et al: a fusion signal was Cancer Genet Cytogenet 2001; 129: 1–9. detected in 30/30 FL. However, threshold for positivity was 9 Mitelman F, Johansson B, Mertens S. Mitelman database of cata- quite low, that led to ascertain positivity for all FL tested but logue of chromosome aberrations in cancer. 2000. also for one part of normal controls. Moreover, subsequent http://cgap.nci.nih.gov/Chromosomes/Mitelman studies failed to confirm that low threshold using both ident- 10 Knutsen T. Cytogeneticmechanismsin the pathogenesis and pro- 6,8 gression of follicular lymphoma. Cancer Surv 1997; 30: 163–192. ical probes, and a recent study showed that this set of 11 Avet-Loiseau H, Li JY, Facon T, Brigaudeau C, Morineau N, Malo- 8 probes was unable to detect some cases of FL, and the pro- isel F, Rapp MJ, Talmant P, Trimoreau F, Jaccard A, Harousseau posed explanation was that the bcl-2 probe used did not map JL, Bataille R. High incidence of translocations t(11;14)(q13;q32) at a sufficient distance of all possible breakpoints. and t(4;14)(;q32) in patients with plasma cells malignancies. In a preliminary study, we applied our interphase FISH Cancer Res 1998; 58: 5640–5645. assay in DLBCL cases following a FL: 16 of 17 cases tested 12 Rimokh R, Gadoux M, Bertheas MF, Berger F, Garoscio M, Dele- age G, Germain D, Magaud JP. FVT-1, a novel human transcrip- were found to bear t(14;18), showing that our assay is also tion unit affected by variant translocation t(2;18)(p11;q21) of fol- suitable to identify such a subgroup of patients within DLBCL. licular lymphoma. Blood 1993; 81: 136–142. Cytogeneticstudies reported some chromosomalrearrange- 13 Bastard C, Tilly H, Lenormand B, Bigorgne C, Boulet D, Kunlin ments in FL other than t(14;18)(q32;q21), including translo- A, Monconduit M, Piguet H. Translocations involving band 3q27 cations between bcl-2 gene and Ig␬ at 2p12 or Ig␭ at 22q11, and Ig gene regions in non-Hodgkin’s lymphoma. Blood 1992; or double rearrangement on the same allele (5’end transferred 79: 2527–2531. 14 Gribben JG, Saporito L, Barber M, Blake KW, Edwards RM, Griffin ␭ 6 19 to the Ig locus, VCR deletion ). All the above mentioned JD, Freedman AS, Nadler LM. Bone marrows of Non-Hodgkin’s cytogenetic changes cannot be detected using both our assay lymphoma patients with a bcl-2 translocation can be purged of and the commercially available one. However, as we demon- polymerase chain reaction-detectable lymphoma cells using strated t(14;18)(q32;q21) in all patients from a series of 63 FL monoclonal antibodies and immunomagnetic bead depletion. cases tested, incidence of such abnormalities is certainly low. Blood 1992; 80: 1083–1089. Moreover, several reports described FL cases lacking any bcl- 15 Vaandrager JW, Schuuring E, Philippo K, Kluin PM. V(D)J recom- 20–22 binase-mediated transposition of the BCL2 gene to the IgH locus 2 rearrangement, especially in grade 3 FL. These cases in follicular lymphoma. Blood 2000; 96: 1647–1652. frequently display bcl-6 rearrangements and may correspond 16 Ku¨ppers R, Klein U, Hansmann ML, Rajewsky K. Cellular origin to special forms of DLBCL. Stringent criteria are needed to of human B-cell lymphomas. N Engl J Med 1999; 11: 1520–1529.

Leukemia t(14;18)(q32:q21) in follicular lymphoma A Godon et al 259 17 Buchonnet G, Lenain P, Ruminy P, Lepretre S, Stamatoullas A, esen M. Multiple recurrent genomic defects in follicular lym- Parmentier F, Jardin F, Duval C, Tilly H, Bastard C. Characteriz- phoma. A possible model for cancer. N Engl J Med 1987; 316: ation of BCL2-JH rearrangements in follicular lymphoma: PCR 79–84. detection of 3’BCL2 breakpoints and evidence of a new cluster. 21 Offit K, Jhanwar SC, Ladanyi M, Filippa DA, Chaganti RS. Cytog- Leukemia 2000; 14: 1563–1569. enetic analysis of 434 consecutively ascertained specimens of 18 Yabumoto K, Akasaka T, Muramatsu M, Kadowaki M, Hayashi T, non-Hodgkin’s lymphoma: correlations between recurrent aber- Ohno H, Fukuhara S, Okuma M. Rearrangement of the 5’ cluster rations, histology, and exposure to cytotoxic treatment. Genes region of the BCL2 gene in lymphoid neoplasm: a summary of Chromosomes Cancer 1991; 3: 189–201. nine cases. Leukemia 1996; 10: 970–977. 22 Ott G, Katzenberger T, Lohr A, Kindelberger S, Rudiger T, Wilhelm 19 Seite P, Hillion J, D’Agay MF, Berger R, Larsen CJ. BCL2 complex M, Kalla J, Rosenwald A, Muller JG, Ott MM, Muller-Hermelink rearrangement in follicular lymphoma: translocation mbr/JH and HK. Cytomorphologic, immunohistochemical, and cytogenetic deletion in the vcr region of the same BCL2 allele. Oncogene profiles of follicular lymphoma: 2 types of follicular lymphoma 1993; 8: 3073–3080. grade 3. Blood 2002; 99: 3806–3812. 20 Yunis JJ, Frizzera G, Oken MM, McKenna J, Theologides A, Arn-

Leukemia