Gastric Low-Grade MALT Lymphoma, High-Grade MALT

Home , Gastric lymphoma, MALT lymphoma

Leukemia (1999) 13, 799–807  1999 Stockton Press All rights reserved 0887-6924/99 $12.00 http://www.stockton-press.co.uk/leu Gastric low-grade MALT lymphoma, high-grade MALT lymphoma and diffuse large B cell lymphoma show different frequencies of trisomy MA Hoeve1, IAM Gisbertz2, HC Schouten3, E Schuuring1, FJ Bot2, J Hermans4, A Hopman5, PhM Kluin1, J-W Arends2 and JHJM van Krieken1

Departments of 1Pathology and 4Medical Statistics, Leiden University Medical Center, Leiden; and Departments of 2Pathology, 3Internal Medicine, and 5Molecular Cellbiology and Genetics, University Hospital Maastricht, Maastricht, The Netherlands

Gastric MALT lymphoma is a distinct entity related to Helico- as diffuse large B cell lymphoma (DLBCL) of the stomach. In bacter pylori gastritis. Some studies suggest a role for trisomy the REAL classification both high-grade MALT and large B cell 3 in the genesis of these lymphomas, but they mainly focused lymphoma of the stomach are classified as the same entity, on low-grade MALT lymphoma. Gastric MALT lymphoma, how- 2 ever, comprises a spectrum from low- to high-grade cases. Fur- namely diffuse large cell lymphoma. thermore, its exact relation to primary diffuse large B cell lym- Genetic aberrations associated with gastric MALT lymphoma (DLBCL) of the stomach is not clear. We applied in situ phoma of either low- or high-grade type and with lympho- hybridisation (ISH) with centromeric probes on 43 samples of magenesis from inflammation to low-grade and (eventually) 39 patients with primary gastric lymphoma (13 samples with high-grade MALT lymphoma have not yet been fully iden- low-grade MALT lymphoma, 25 with high-grade MALT lymphoma and five with DLBCL) to detect numerical aberrations tified. So far, studies investigating the genetics of MALT lym- of 10 chromosomes. ISH was performed immunohistochem- phoma have mainly focused on low-grade type. Recurrent ically on nuclei isolated from paraffin-embedded resection aberrations include trisomies of chromosomes 3, 12 and 18, tissue and on whole paraffin sections using immunofluoresc- and to a lesser extent of chromosome 7.5–8 Furthermore, trans- ence. In six of 13 low-grade MALT lymphomas trisomy was location t(11;18) and less frequently t(1;14), t(3;14), t(8;14) detected (46%) and mostly involved chromosome 3 (33%). In and t(14;18) have been identified. Aberrations involving MYC high-grade MALT lymphomas, trisomies were found in 16 of 25 have so far been described in three studies on large cell cases (64%), mainly involving chromosomes 12 and 18. Tri- 9–11 somy 3 was present in only 13% of these cases. Of five DLBCL, lymphomas. only one showed trisomy. Nine of the 16 aberrant high-grade To investigate numerical chromosomal aberrations in the MALT lymphomas (56%) showed trisomy of more than one three different types of primary gastric lymphoma (low-grade chromosome per case vs two of six for low-grade cases. In lym- MALT, high-grade MALT and DLBCL) we performed in situ phomas with separate low- and high-grade tumour components hybridisation (ISH) with (peri)centromeric probes for multiple some trisomies were detected in both components, whereas others occurred only in the high-grade tumour cells. This sup- chromosomes. Per case, the percentage of nuclei with three ports the hypothesis that high-grade MALT lymphomas can spots for each hybridised chromosome was used to determine develop from a low-grade type and that this progression is the presence of trisomy. To ensure a large series of cases, ISH accompanied by the acquisition of more genetic aberrations. results of two different studies were combined, namely analy- However, trisomy 3 probably does not play a major role in sis of isolated nuclei and analysis of tissue sections. Both were this progression. equally powerful to detect trisomy, as was demonstrated in Keywords: gastric lymphoma; MALT lymphoma; in situ hybridisation; trisomy 3; numerical aberration lymphomas analysed by both techniques. ISH on isolated nuclei has the intrinsic advantage of avoiding cutting artefacts. On the other hand, ISH on intact tissue sections has the advantage that tumour cells can be distinguished from epi- Introduction thelial cells and reactive inflammatory cells present in the same tissue section. Moreover, in cases separately containing Since lsaacson and Wright1 in 1983 first described the low-grade and high-grade MALT lymphoma, analysis of tissue mucosa-associated lymphoid tissue (MALT) lymphoma con- sections allowed us to investigate both components indepen- cept, this clinicopathological entity has increasingly been dently. Analysis of 39 evaluable cases revealed different fre- studied. It is a distinct entity within the group of extra-nodal quencies of trisomies in low-grade and high-grade MALT lym- marginal zone B cell lymphomas.2 MALT lymphomas usually phoma and DLBCL, indicating different oncogenetic pathways arise in a background of chronic inflammation and are most for some of these tumours. common in the stomach, in the setting of Helicobacter pylori gastritis. Although first described as a low-grade lymphoma, transformation into a large cell lymphoma is now well estab- Materials and methods lished.3,4 In some studies, these lymphomas are designated as high-grade MALT lymphomas. Similarly, primary large cell Tissue specimens lymphoma with features of MALT lymphoma such as lym- phoepithelial lesions or an accompanying small cell compo- From the files of the pathology departments of the Leiden Uni- nent are often classified as high-grade MALT lymphoma. versity Medical Center and the Academic Hospital Maastricht, However, many gastric large cell lymphomas lack these fea- formalin-fixed and paraffin-embedded tissue of surgically tures, especially in small biopsy specimens, and are classified resected gastric lymphomas of 47 patients was retrieved. Thir- teen of these were classified as low-grade MALT lymphoma, 29 as high-grade MALT lymphoma (four of which contained Correspondence: MA Hoeve, Department of Pathology, Leiden Uni- separate fields of low- and high-grade tumour cells in the versity Medical Center, PO Box 9600, 2300 RC Leiden, The Nether- lands; Fax: 31 71 5248158 same tissue block) and five were classified as diffuse large B MA Hoeve and IAM Gisbertz contributed equally to this work cell lymphoma without the presence of features of a low-grade Received 23 November 1998; accepted 3 February 1999 MALT component (DLBCL). Classification was based on cri- Trisomy in different types of gastric lymphoma MA Hoeve et al 800 teria described by Isaacson et al;1,12,13 MALT origin was (D15Z1, Oncor),23 No. 17 (p17H8)24 and No. 18 (L1.84).25 defined by the presence of a small cell component showing Probe pAE0.68 was labelled with digoxigenin-11-dUTP while the typical features of MALT: reactive follicles surrounded by all other probes were labelled with biotin-11-dUTP by means centrocyte-like cells, plasma cell differentiation and lympho- of standard nick-translation of complete plasmid DNA, epithelial lesions. High-grade MALT lymphomas were charac- according to the manufacturer’s directions (Boehringer terised by the additional presence of sheets of large, blastic Mannheim). cells. DLBCL was defined as a tumour consisting of B cell blasts, but lacking low-grade MALT features. As controls, formalin-fixed, paraffin-embedded tissue of three Helicobacter ISH on isolated nuclei pylori-associated chronic gastritis samples, a tonsil and a lymph node involved by chronic lymphocytic leukaemia (CLL) For hybridisation on nuclear suspensions, probes were diluted were used. The presence of the tumour cells and tumour type in a hybridisation buffer (60% formamide, 2 × SSC pH 7.0 were evaluated in each tissue block on haematoxylin and (0.3 M NaCl/30 mM Na citrate), 10% dextran sulphate eosin-stained slides that were cut before and after the sections (Pharmacia, Uppsala, Sweden), 0.2 ␮g/␮l herring sperm DNA used for ISH. and 0.2 ␮g/␮l yeast tRNA) to a final concentration of 0.2 ng/␮l for pUC1.77 and 1 ng/␮l for the other probes. Ten microlitres of hybridisation mixture was added on to the slide and probe Isolation of nuclei and target DNA were denatured for 5 min in a metal box in an 80°C waterbath. After overnight hybridisation in a humidified Nuclei were isolated from the paraffin-embedded tissues by a chamber at 37°C samples were washed 2 × 5 min in method modified from Hedley et al14 and Schutte et al.15 2 × SSC/0.05% Tween-20 at 42°C, 2 × 5 min in 0.1 × SSC at Briefly, two sections of 30 ␮m were deparaffinised in xylene 60°C and 2 × 5 min in 4 × SSC/0.05% Tween-20 at room tem- and rehydrated in a graded ethanol series. Subsequently, perature. Detection of the biotinylated probes was performed samples were pre-treated by incubation for 20 min in 85% with mouse-␣-biotin (Dako, Glostrup, Denmark, 1:100), formic acid/0.3% peroxidase and rinsed in 0.01 M HCl. Enzy- detection of the digoxigenin-labelled probe with mouse-␣- matic digestion was performed with 4 mg/ml pepsin (Sigma, digoxygenin (Sigma, 1:2000). This step was followed, for all St Louis, MO, USA) in 0.2 M HCl for 30 min at 37°C. The slides, by the application of biotinylated horse-␣-mouse digest was spun on to poly-L-lysine (MW 150 000; Sigma)- (Vector, Burlingame, CA, USA, 1:200) and avidin-biotin-per- coated glass slides for 5 min at 700 r.p.m. with Cytospin3 oxidase (Vector, 1:100). All antibody incubations were perfor- (Shandon Scientific Limited, Arsmoor, UK). After the slides med for 30 min in a humidified chamber at 37°C and slides were air-dried, they were dehydrated with 70% ethanol in were washed with 1 × PBS/0.1% Tween-20/5% normal goat 0.01 M HCl to preserve nuclear morphology and further dehy- serum. When necessary, signals were amplified using CARD drated subsequently in 90% and 100% ethanol and air-dried. (Catalyzed Reporter Deposition:26 after the last washing step (that follows the avidin-biotin-peroxidase) biotinylated tyra- mine was applied on the slide, which was afterwards detected Section preparation by avidin-peroxidase). In all cases the peroxidase activity was visualised with 3,3Ј-diaminobenzidine tetrahydrochloride Three-␮m paraffin sections were stretched in a 40°C water (DAB, Sigma) and nuclei were counterstained with bath, mounted on APES-coated (3-aminopropyltriethoxysilan; haematoxylin. Sigma) glass slides and dried overnight in a 56°C oven. The sections were deparaffinised (at room temperature) in xylene (3 × 10 min), immersed in 100% ethanol (2 × 2 min) and air FISH on paraffin sections dried for 30 min. To facilitate penetration of the DNA probe into the nuclei, the slides were incubated in 1 M NaSCN For fluorescent in situ hybridisation (FISH) on tissue sections, (Sigma) at 80°C for 10 min, rinsed in deionized water (at 37°C) the labelled DNA probes were diluted at 0.5 ng/␮l (in 60% and digested in a solution of pepsin (4 mg/ml pepsin; Sigma) formamide, 2 × SSC, 10% dextran sulphate), pipetted on the in 0.2 M HCl, pH 1.2–2 at 37°C. The digestion time appeared digested, dehydrated tissue section (15–40 ␮l depending on to depend on cell type and tissue fixation and was evaluated the section size) and covered by a glass coverslip. Denatur- for each tissue block by means of a time series (5–10–15–20– ation of probe and target DNA was performed by placing the 25 min). The optimal digestion time was determined on the tissue section with the probe on a brass bar in an 80°C oven basis of signal intensity in combination with the quality of the for 10 min. Hybridisation took place overnight at 37°Cina morphology. After the digestion step, the sections were rinsed humidified chamber (deionized water). The following day in 1 × phosphate-buffered saline (PBS), dehydrated in an etha- slides were immersed briefly in 2 × SSC at 37°C to remove the nol series (70%, 90%, 100%, 5 min each) and air-dried at coverslip, and washed in 60% formamide/2 × SSC 80°C for 25 min. (2 × 10 min, 37°C) and 3 × 5 min in 1 × PBS at room temperature. The sections were preincubated with 4 × SSC with 5% non-fat dry milk for 30 min at 37°C in a humidified chamber. Probes for in situ hybridisation After removal of the coverslip by immersion in 1 × PBS/0.1% Tween-20 at room temperature, the slides were immuno- Hybridisation was performed with (peri)centromeric repetitive stained in three steps with FITC-avidin (Vector, 1:1000) and ␣ satellite DNA probes specific for chromosomes No. 1 (Puc a biotinylated goat-anti-avidin monoclonal antibody (Vector, 1.77),16 No. 3 (pAE0.6817 (Boehringer Mannheim, Mannheim, 1:200). Between the different antibody steps, slides were Germany), D3Z117 (Oncor, Gaithersburg, MD, USA) or washed three times with 1 × PBS/0.1% Tween-20. Nuclear p␣3.5)17, No. 7 (p7t1),18 No. 8 (D8Z2, Oncor),19 No. 9 DNA was counterstained with 1 ␮g/ml propidium-iodide in (pHUR98),20 No. 11 (pLC11A),21 No. 12 (p␣12H8),22 No. 15 antifade solution (Vectashield; Vector). Trisomy in different types of gastric lymphoma MA Hoeve et al 801 Counting hybridisation signals results obtained using different probes on consecutive sections of the same tissue block. The distribution of the percentage of Immunohistochemical (= DAB) and fluorescent signals were nuclei with a certain number of spots obtained after hybridis- counted using guidelines reported in the literature.27–29 The ing a particular probe did not differ much between different following criteria were taken into account: (1) only ‘intact’ cell types within one section, in spite of differences in nuclear (=spherical), non-overlapping nuclei with a clear counterstain size (see Figure 1 and Table 1A). Furthermore, different sec- were counted; (2) nuclei with paired spots (split spots) were tions of the same tissue block gave comparable distributions of not counted; (3) signals within one nucleus should have more the mean number of spots per nucleus for different hybridised or less the same size and intensity (thus excluding non-specific probes (Figure 1). Taken together, our findings allowed com- signals such as minor binding sites); and (4) parallel-cut sec- parison of the number of spots per chromosome in normal tions stained with HE were evaluated simultaneously to cells and tumour cells (both low and high grade) of one tissue identify cell types and tissue areas. block and calculation of a cut-off level for trisomy per investi- Per hybridised paraffin section, in at least 4 × 100 nuclei of gated case (see below; section on Statistics). Since the spot a particular cell type the number of fluorescent signals was distribution appeared to be mainly tissue block dependent, counted. Of the nuclear suspensions immunohistochemical thresholds for trisomy were, contrary to other ISH studies on signals were counted, also in 4 × 100 nuclei. The relative paraffin sections, determined for each individual case. numbers of nuclei showing nil, one, two, three or more spots were expressed as percentages. The intra-observer variability between different series of 100 counted cells was less than Immunohistochemical ISH on isolated nuclei vs 5%. fluorescent ISH on tissue sections To test the above mentioned guidelines and the quality of the probes for our FISH procedure on paraffin sections, the For comparison of both hybridisation methods (ISH on nuclear various probes were hybridised on sections of a tonsil, a gas- samples and FISH on sections), the results of counting DAB tritis sample and a CLL in which a trisomy 12 was found with signals in hybridised nuclei of different tissues, using probes previous ISH experiments in leukemic blood cells.30 Although for various chromosomes, were plotted in a similar way as the the various probes showed slight differences in intensity, they FISH results. The absolute numbers of nuclei containing nil, all produced evaluable signals in most of the samples investi- one, two or three DAB spots per nucleus in a sample differed gated. FISH on sections of the CLL resulted in nuclei with from the spot distribution in FISH samples (Table 1B). The three fluorescent spots per nucleus for chromosome 12, con- range of the percentages of cells with nil, one, two and three firming the presence of trisomy 12. The tonsil and gastritis spots in normal diploid cells and trisomic cells in our study case displayed diploidy for all probes that were evaluated (Table 1B) is comparable with the results obtained by others.31 (data not shown). To evaluate the quality of a FISH experiment DAB and fluorescent spot numbers in nuclei possibly contain- and establish optimal digestion time per tissue block, with ing a trisomy were both statistically evaluated using the each experiment a section of the same tonsil was hybridised binomial distribution (see Figure 2 and below; section on in parallel. Statistics). The efficiency of the immunohistochemical hybridisation procedure on isolated nuclei also showed differences for the various probes tested, but most could be evaluated in the Statistics majority of the cases. Of the two probes that were used for immunohistochemical detection of chromosome 3 (D3Z1 and To determine threshold levels for trisomy in individual pAE0.68) in situ hybridisation with probe pAE0.68 gave the tumours results of hybridisation with the different probes were best results. pooled per tumour and the frequencies of cells with three spots were plotted. This analysis results in a diagram, rep- resenting a binomial distribution, from which the mean Evaluation of hybridisation results (background) percentage of cells with three spots could be determined per case (Figure 2). For cells without chromo- Previous ISH studies on archival paraffin tissue sections of somal gains, the mean percentage of nuclei with three spots various tumours pinpointed certain factors that might influ- varied between 0 and 4.5% for ISH on nuclear suspensions ence quantitative analysis of chromosome copy number by and between 1 and 3% for FISH on tissue sections. By com- means of this method.29 Sectioning the tissue at 3 ␮m leads paring the percentage of three-spotted nuclei in cells with sus- to truncation of the nuclei. As a result, DNA content is often pected trisomy for a certain chromosome with these back- incomplete and chromosome numbers will be underesti- ground values, the cells could be classified as diploid or mated. When investigating both isolated nuclei and paraffin trisomic for this chromosome. For this comparison, the distri- sections two signals might be misinterpreted as one when bution of the number of nuclei with three spots, measuring N close together or overlapping. Moreover, various other factors nuclei, was addressed as a binomial distribution with para- might affect the average number of hybridisation signals per meters N (number of nuclei) and a mean percentage (q)of nucleus, such as quality of tissue fixation and probe pen- 4.5% for ISH on nuclei and 3% for FISH on sections. For a etration in relation to nuclear permeability obtained by diges- given experiment, N is known and the number of nuclei with tion. Together with aspects inherent in the quality of the three spots, X, is counted. The binomial probability of X or hybridisation experiment (signal intensity, morphology of more nuclei with three spots, when investigating N nuclei, nuclei) these factors influence the cut-off level for trisomy. can be approximated by a normal distribution, using To investigate the role of these aspects in our FISH study Z = (X − q*N)/[N*q*(1 − q)]1/2. on paraffin-sections we (1) plotted the number of nuclei with The tail probability of the normal variate Z can be found in a certain number of spots counted in both tumour cells and tables of the normal distribution. If this probability is small, normal cells in a single tissue section, and (2) compared the say Ͻ10−5, then the experimental finding (counted number of Trisomy in different types of gastric lymphoma MA Hoeve et al 802

Figure 1 Example of FISH analysis to detect chromosomal polyploidy. Parafﬁn sections containing both epithelial and MALT lymphoma cells of a patient (case 12) were hybridised with probes for Nos 1, 3, 7, 12 and 18, and the mean number of spots of 400 nuclei evaluated for each probe are represented (in percentages) in separate bars. Results of the epithelial cells are depicted on the left side, the lymphoma cells on the right side. The lymphoma cells show a relative increase of nuclei with two and three spots for No. 12 compared with No. 12 on epithelial cells and other numbers on lymphoma cells. Statistical analysis indicated the presence of trisomy 12 (P Ͻ 10−5) in this case.

nuclei with three spots) is unlikely under the assumed mean independently with both methods in four randomly chosen of 3% or 4.5% of nuclei with three spots. The assumption can gastric lymphomas. The samples were blindly evaluated by then be rejected, which means that the experimental finding two researchers (nuclear samples by IAMG and sections by significantly aberrates the expected number of three spots in MAH). Interpretation of the results of both methods resulted normal, i.e diploid cells and that trisomy is likely. In all situ- in similar outcomes for the various chromosomes tested ations where finally the classification ‘trisomy’ was made, this (Table 3). We therefore combined the results obtained with probability was Ͻ10−5. both methods. ISH revealed relative gains of one or more of The Fisher’s exact test (two-sided) on a 2 × 2 table was per- the investigated chromosomes (1, 3, 7, 8, 9, 11, 12, 15, 17 formed, using SPSS software for Windows, to compare the fre- and 18) in 21 of the 39 tumours (Table 2). Numerical aber- quencies of trisomy (per chromosome and of all chromosomes rations occurred more frequently in high-grade MALT lym- together) occurring in the different lymphoma groups (low- phoma (16/25; 64%) than in low-grade MALT lymphoma grade MALT, high-grade MALT and DLBCL). A value of (6/13; 46%) (Table 4). However, this difference was statisti- P Ͻ 0.05 was considered significant. cally not significant (P = 0.3). Furthermore, in high-grade MALT lymphomas, chromosomal gain involved most of the evaluable chromosomes, whereas in low-grade MALT lym- Results phoma only four of the 10 chromosomes that were tested showed trisomy. Moreover, in high-grade MALT lymphomas, In 39 of the 47 gastric lymphoma cases, the chromosome nine of 16 aberrant cases showed trisomy for two or more numbers of three or more chromosomes could be determined chromosomes (56%). One lymphoma (case 34) even showed after in situ hybridisation (ISH). In 18 of these, ISH was perfor- trisomy for all five chromosomes tested. This case appeared med on nuclear suspensions, in 17 cases ISH was performed to contain an aneuploid cell population, as was determined on tissue sections, and in four cases ISH was performed on by FACS analysis of isolated nuclei (according to Hedley et both nuclear suspensions and tissue sections. Table 2 shows al14) from the same tissue block (data not shown). Of the six the results for all 39 cases for chromosomes 1, 3, 7, 12, 15, aberrant low-grade MALT lymphomas, two showed simul- 17 and 18. Chromosomes 8, 9, 11, 15 and 17 were investi- taneous trisomy of two chromosomes. Only one of the five gated only in nuclear suspensions. Eight cases could not be DLBCL (20%) in the study showed a trisomy, namely of evaluated after ISH on isolated nuclei (three low-grade and chromosome 12. two high-grade MALT lymphomas) or on paraffin sections In low-grade MALT lymphomas, trisomy of chromosome 3 (two low-grade and one high-grade MALT lymphomas) due to was seen most often, namely in four of the 12 evaluable low- inferior hybridisation signals and/or poor nuclear morphology. grade cases (33%). Gain of chromosome 12 was seen in two of the 13 tumours (15%), while trisomy 11 and 18 were each found once (8% of evaluable cases) (Table 4). In the high- ISH on nuclear suspensions and sections from grade MALT lymphomas, trisomy of chromosome 12 and 18 paraffin-embedded gastric lymphomas, combined was found most frequently (respectively nine of 25 evaluable results cases, 36%; and seven of 23 cases, 30%) (Table 4). Trisomy 1 and 7 each occurred in four cases (respectively 4/25, 16%; To check whether the two hybridisation methods did indeed and 4/24, 17%), followed by trisomy 3 (3/23, 13%) and 15 yield the same results, various chromosomes were analysed (2/13, 15%). Trisomy of chromosome 17 was seen once (1/13, Trisomy in different types of gastric lymphoma MA Hoeve et al 803 Table 1 Spot distribution in different tissues of various lym- 8%). Case 17, a high-grade MALT lymphoma consisting of phoma cases various tumour fields of mixed small and large cells, showed trisomy 12 in some of these fields (Table 2). A Four cases contained separate low-grade and high-grade MALT lymphoma components that could be analysed inde- Spots/ Case 3 Case 11 Case 12 CLL pendently in tissue sections. This allowed us to investigate Nucleus possible tumour progression. Two tumours, cases 11 and 13, ED E D T E D T T showed trisomy 3 in both components. Case 13, moreover, showed gain of chromosome 18 in both low-and high-grade 0 116334427 lymphoma cells. In contrast, this case showed trisomy 1 in 1 383654523456553523 2 596039444939405252 exclusively the high-grade component. Similarly, case 12 3 23 1 1141 11118 showed trisomy 12 only in the high-grade cells (Figure 3).

B Discussion Spots/ ISH on nuclei % FISH on sections % Nucleus Recurrent chromosomal aberrations are specifically associa- 32–34 DT EDTted with different lymphoma types and can serve as diag- nostic or prognostic markers.35–37 Furthermore, they may be 0 0–10 0–8 1–10 1–16 1–6 useful for elucidating disease development and progression. 1 1–22 1–10 36–53 31–56 19–36 Although lymphomas resembling mucosa-associated lymph- 2 64–99 30–72 39–63 31–66 44–55 oid tissue (MALT lymphoma) were described already in 1983 3 0–9 11–53 1–3 0–5 8–28 as a distinct entity,1 no conclusive specific chromosomal aberrations associated with them have been identified as yet. Some (A) depicts the mean percentages of fluorescent spots per nucleus studies on extranodal MZBCL and/or MALT lymphomas of of a chromosome, counted in 400 nuclei of hybridised sections of various sites report gain of chromosome 3 in more than 50% three gastric and one lymphocytic lymphoma (No. 3 on case 3, No. of these lymphomas.6,38–40 However, other reports show lower 3 on case 11, No. 12 on case 12, and No. 12 on a CLL with trisomy 5,7,41,42 12). E stands for epithelial cells, D for lymphoma cells classified as incidences of trisomy 3 in this type of NHL. Further- diploid for the particular chromosome tested and T for lymphoma more, it is still not clear whether primary gastric large cell cells classified as trisomic for the particular chromosome. In case lymphoma originates from low-grade MALT lymphoma or that 3 no trisomic cells were present and in the CLL no epithelial and it arises de novo and is a separate entity. Some studies are diploid tumour cells were present. Although the various cases show supportive of a lymphomagenesis of high-grade MALT lym- variations in the distribution of the number of nuclei with nil, one, two and three spots, the spot distributions in cases 3, 11 and 12 phoma from low-grade MALT lymphoma, eg by showing of the epithelial and diploid tumour cells are comparable. Further- clonal relationship between low- and high-grade lymphoma more, the distribution of spots in the trisomic cells of cases 11 and cells.3,4,43–45 It might be hypothesised that MALT lymphocytes 12 are comparable. undergo a sequence of genetic changes that ultimately lead (B) shows the mean percentages of cells in all hybridised samples, to high-grade MALT lymphoma. In this case, one would containing nil, one, two and three spots. Numbers represent the expect to find aberrations of the low-grade tumour to be range of all countings, performed on nuclear samples (left) and on whole paraffin sections (right) hybridised with the different probes. present in the high-grade tumour cells as well. Moreover, one might expect additional, more complex genetic alterations in the high-grade cells as compared to the low-grade component. To investigate this possible relationship we analysed numerical chromosomal aberrations with in situ hybridisation (ISH) in a series comprising different types of gastric lymphoma, using probes for centromeric regions of 10 different chromosomes. Chromosomes 3, 7, 12 and 18 were included since these have been described to be aberrant in NHL in general and in gastric lymphomas in particular.5–8,38,46–48 Cytogenetic studies by Wotherspoon et al5 and Ott et al7 reported trisomy 3 in respectively 20% and 10% of gastric low-grade MALT lymphomas and in 21% of high-grade MALT lymphomas.7 Using FISH on isolated nuclei of 10 gastric low- grade MALT lymphomas, Ott et al7 confirmed a relatively low incidence of trisomy 3 (10%). Very recently, these investigators again reported low frequencies of trisomy 3 in MALT lymphomas, showing +3 in respectively 20% and 36% of low- Figure 2 Diagram showing a binomial distribution of the number grade and high-grade tumours using bicolour FISH on isolated of cells with three spots per nucleus for various chromosomes. Data nuclei.42 Furthermore, investigating gastrointestinal large B represent the results of counting spots in 5× a cohort of 100 nuclei cell lymphomas, including MALT-type, Chan and co-workers8 for each of the chromosomes 3 (black bars) and 1, 7, 12 and 18 (white found no trisomy 3 using comparative genomic hybridisation bars) hybridised on sections of a low-grade MALT lymphoma (case (CGH) in seven cases, and Barth and colleagues41 showed 6), and shows a binomial distribution. The mean percentage of three- spotted cells for Nos 1, 7, 12 and 18 lies at 1%, whereas for No. 3 gain of chromosome 3 in only 13% of 31 cases. However, in 6 the mean percentage lies at 10.6%. Statistical analysis confirmed the 1995 Wotherspoon and co-workers had found trisomy 3 in presence of trisomy 3 in this case. 62%, trisomy 18 in 28%, trisomy 12 in 14% and trisomy 7 in Trisomy in different types of gastric lymphoma MA Hoeve et al 804 Table 2 Overview of numerical aberrations detected in paraffin-embedded gastric lymphomas hybridised with various centromeric probes

Case Tumour Cell type I/F No. 1 No. 3 No. 7 No. 12 No. 15 No. 17 No. 18

1 MALT low small I N ne N N N N N 2 MALT low small I N N N N N N N 3 MALT low small F N N N N — — N 4 MALT low small I N N N N N N N 5 MALT low small I N N N N N N N 6 MALT low small FNTN N — — N 7 MALT low small I N N N T ne N N 8 MALT low small I N T N N N N N 9 MALT low small I N N N T N N N

10 MALT high small F N N N N — — N large F N N N N — — N 11 MALT high small FNTN N — — N large FNTN N — — N 12 MALT high small F N N N N — — N large F N N N T — — N 13 MALT high small FNTN N — — T large F T T N N — — T

14 MALT high mixed F N N N N — — N 15 MALT high mixed F N N N T — — T 16 MALT high mixed F N N T T — — N 17 MALT high mixed F N N N T+N— — T 18 MALT high mixed F T N N N — — N

19 MALT high large F N N N N — — ne 20 MALT high large F N N N N — — N 21 MALT high large I N ne N N N N N 22 MALT high large I N N N N N N N 23 MALT high large I N N N N N N N 24 MALT high large I N N N N N N N 25 MALT high large I N N N N N N N 26 MALT high large I N ne N N N N T 27 MALT high large I N N ne T N N N 28 MALT high large I N N N N T N ne 29 MALT high large I N N N T N N T 30 MALT high large I N N T T N T N 31 MALT high large I T N N N T N N 32 MALT high large I N N T N N N T 33 MALT high large I N N N T N N N 34 MALT high large F T T T T — — T

35 DLBCL large I N ne N N ne N N 36 DLBCL large I N N N N N N N 37 DLBCL large F N N ne N — — ne 38 DLBCL large F N N N N — — N 39 DLBCL large F N N N T — — N

I, hybridised immunohistochemically; F, hybridised ﬂuorescently; N, normal (no numerical aberrations detected); T, trisomy; −, not done, ne, not evaluable; mixed, sample consisted of intermingled low- and high-grade lymphoma cells. The results for chromosomes 8, 9 and 11, hybridised only on isolated nuclear samples, were only informative in respectively 12, six and four lymphomas and are therefore not shown in this Table.

3% of gastric low grade MALT-lymphomas using a similar aberrations in different haematological malignancies.49 Fur- FISH technique. A year later, Dierlamm et al38,47 reported tri- thermore, different research groups might have selected for somy 3 in respectively 50% and 67% of extranodal marginal different cases, so that the lymphomas investigated differ in zone B cell lymphomas (MZBCL) using cytogenetics and FISH. histology, tumour stage, or site of origin of the lymphoma. Recently, Zhang et al40 found trisomy 3 in 75% of gastric low- Dierlamm et al38,47 for instance, reported their high trisomy grade MALT lymphomas with ISH on tissue sections. The 3 frequencies in lymphomas with complex karyotypes when high-grade MALT lymphomas investigated in that study also compared with the cases studied by Ott and colleagues,42 showed a high incidence of trisomy 3, which was not found possibly indicating higher tumour grades of the cases used by by others. Our ISH data are in agreement with the former stud- the former investigators. Moreover, the incidences of other ies showing a relatively low incidence of trisomy 3.5,7,8,41,42 genetic aberrations like t(11;18)7,47 and bcl-6 involvement Various factors might contribute to this discrepancy in alone or in concert with the presence of trisomy 36,47,50 vary reported trisomy 3 incidences. For instance, geographical vari- as well in the reports of different researchers. Another variable ation might play a role. Various studies describe the inﬂuence in the different studies is the methodology used to determine of this variable on the occurrence of certain chromosomal the chromosome copy number. Differences in technical Trisomy in different types of gastric lymphoma MA Hoeve et al 805 Table 3 Comparison of results obtained with hybridisation of nuclear samples (n) and sections (s) of four MALT lymphomas

Case Tumour Cell type n/s Chrom 1 Chrom 3 Chrom 7 Chrom 12 Chrom 18

1 MALT low small s N (5/53/41/1) — — N (5/52/42/1) N (4/50/43/3) n N ne N N (0/4/87/9) N 2 MALT low small s N (3/46/50/1) N (7/47/45/1) — N (5/42/51/2) — n N N (1/4/87/8) N N N 29 MALT high large s N (1/38/60/1) N (5/38/55/22) — T (1/19/51/28) ne n N N N T (1/4/42/53) T (7/7/72/14) 32 MALT high large s N (3/37/57/3) N (4/54/40/2) T (4/39/45/12) N (5/51/41/3) T (2/36/44/18) n N N T (1/8/45/46) N (0/9/83/8) T (1/10/45/13/14)

N, normal (no numerical aberrations detected); T, trisomy; −, not done; ne, not evaluable. Between brackets the mean number of nuclei with nil/one/two/three(four) spots are indicated.

Table 4 Numerical aberrations per chromosome detected with ISH on nuclear samples and FISH on parafﬁn sections in the various types of gastric lymphoma

Tumour Trisomy 1 Trisomy 3 Trisomy 7 Trisomy 12 Trisomy 15 Trisomy 18 Total trisomy

MALT low 0/13 4/12 (33%) 0/13 2/13 (15%) 0/6 1/13 (8%) 6/13 (46%) MALT high 4/25 (16%) 3/23 (13%) 4/24 (17%) 9/25 (36%) 2/13 (15%) 7/23 (30%) 16/25 (64%) DLBCL 0/5 0/4 0/4 1/5 (20%) 0/1 0/4 1/5 (20%)

Chromosomes 8 and 9 are not shown since no numerical aberrations of these chromosomes were found. The column total trisomy represents data of all cases with one or more trisomies and include all chromosomes. approaches and interpretations of the results might lead to dif- nine of 16 (56%) aberrant high-grade lymphomas had tri- ferent outcomes, especially in studies based on the ISH tech- somies of two or more of the investigated chromosomes, while nique. Comparing conventional cytogenetics with ISH, some of the six aberrant low-grade tumours two cases (33%) had a authors found discordant results while others reported con- gain of more than one chromosome (Table 2). These data cordant findings. Wotherspoon and co-workers for instance resemble those of a small study by Robledo et al51 and indi- found a much higher frequency of trisomy 3 by ISH (62%)6 cate an increasing genetic instability by accumulation of than by karyotyping (20%).5 However, both Ott et al7,42 and chromosomal aberrations during transformation from low- to Dierlamm et al38,47 corroborated their cytogenetic data with high-grade MALT lymphoma. ISH. Yet, these groups found respectively low and high fre- Using in situ hybridisation this study shows that low-grade quencies of trisomy 3, so this also does not fully explain the MALT lymphomas carry trisomy 3 in 33% of cases, high-grade discrepancies. Differences in, for instance, the interpretation MALT lymphomas in 13% and DLBCL in none. Of the four 29 of ISH results, especially of so-called split-spots might con- high-grade MALT lymphomas with separately evaluated tribute to the discrepancy. Furthermore, there is some debate components two showed trisomy 3 in both the low- and high- whether ISH on paraffin-embedded material is equally reliable 8 38 40 grade component. In these cases the high-grade lymphoma as on fresh tissue, although both Dierlamm and Zhang likely developed from the low-grade component. Yet, contrary feel that ISH is an adequate tool in detecting chromosomal to trisomy 12 or 18, gain of chromosome 3 appeared to occur gains in archival lymphoma material. more frequently in low-grade cases than in high-grade lym- The results of our ISH study support the above mentioned phomas. Since disappearance of trisomy 3 during the disease assumption concerning disease progression, suggesting that at progression seems unlikely, one might speculate that, apart least some high-grade MALT lymphomas can evolve from a from lymphomagenesis of high-grade from low-grade, also low-grade tumour: (1) trisomy of chromosomes 3, 12 or 18 was found in both low-grade and high-grade MALT lym- other pathways of tumour development of high-grade MALT phoma (Tables 2 and 4); and (2) in two high-grade MALT lym- lymphoma exist. Possibly some low-grade MALT lymphomas phomas with separate low- and high-grade tumour fields tri- indeed do progress to high-grade MALT lymphomas, while somy of chromosome 3 could be demonstrated independently other, genetically distinct, low-grade MALT lymphomas stay in both the low- and the high-grade component of the tumour as they are. Unfortunately, our study group is too small to 52 (cases 11 and 13, Table 2). One of these cases also showed draw a firm conclusion. Yet, it was also noted by Du et al trisomy 18 in both the low-grade and the high-grade lym- that trisomy 3 is not important for high-grade transformation, 8 41 phoma cells (case 13). Furthermore, chromosomal gain was by Chan et al and by Barth et al who reported trisomy 3 in found more often in the high-grade MALT lymphomas than 2/7 (29%) of low-grade and in 2/24 (8%) of high-grade gastric in the low-grade lymphomas (64% vs 46%). Two high-grade lymphomas. Furthermore, other findings of Ott and co-work- MALT lymphomas, cases 12 and 13, consisting of separate ers7 also support this hypothesis. They found differences in fields of low- and high-grade tumour cells, demonstrate this chromosomal aberrations between low-grade and high-grade well: both showed certain trisomies exclusively in the high- MALT lymphomas as well, most apparently the t(11;18). This grade component (Table 2 and Figure 3). In addition, the high- translocation was seen only in low-grade cases (in four of 10 grade MALT lymphomas more often had simultaneously gastric MALT lymphomas, and in total in seven of 20 cases of occurring trisomies as compared to the low-grade tumours: various sites) but not in any of 20 aberrant high-grade lym- Trisomy in different types of gastric lymphoma MA Hoeve et al 806 Acknowledgements

This work was supported by the Dutch Cancer Society, KWF/NKB grant 96-1332.

References

1 Isaacson P, Wright DH. Malignant lymphoma of mucosa-associated lymphoid tissue. A distinctive type of B-cell lymphoma. Cancer 1983 52: 1410–1416. 2 Harris NL, Jaffe ES, Stein H, Banks PM, Chan JK, Cleary ML, Delsol G, de Wolf Peeters C, Falini B, Gatter KC. A revised European– American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 1994; 84: 1361–1392. 3 Chan JK, Ng CS, Isaacson PG. Relationship between high-grade lymphoma and low-grade B-cell mucosa-associated lymphoid tissue lymphoma (MALToma) of the stomach. Am J Pathol 1990; 136: 1153–1164. 4 Montalban C, Manzanal A, Castrillo JM, Escribano L, Bellas C. Low grade gastric B-cell MALT lymphoma progressing into high grade lymphoma. Clonal identity of the two stages of the tumour, unusual bone involvement and leukemic dissemination. Histopa- thology 1995; 27: 89–91. 5 Wotherspoon AC, Pan LX, Diss TC, Isaacson PG. Cytogenetic study of B-cell lymphoma of mucosa-associated lymphoid tissue. Cancer Genet Cytogenet 1992; 58: 35–38. 6 Wotherspoon AC, Finn TM, Isaacson PG. Trisomy 3 in low-grade B-cell lymphomas of mucosa-associated lymphoid tissue. Blood 1995; 85: 2000–2004. 7 Ott G, Katzenberger T, Greiner A, Kalla J, Rosenwald A, Heinrich U, Ott MM, Muller Hermelink HK. The t(11;18)(q21;q21) chromosome translocation is a frequent and specific aberration in low- grade but not high-grade malignant non-Hodgkin’s lymphomas of the mucosa-associated lymphoid tissue (MALT-) type. Cancer Res 1997; 57: 3944–3948. 8 Chan WY, Wong N, Chan AB, Chow JH, Lee JC. Consistent copy number gain in chromosome 12 in primary diffuse large cell lymphomas of the stomach. Am J Pathol 1998; 152: 11–16. 9 van Krieken JH, Raffeld M, Raghoebier S, Jaffe ES, van Ommen Figure 3 Example of fluorescent in situ hybridisation. Centromeric GJ, Kluin PM. Molecular genetics of gastrointestinal non-Hodg- probe for chromosome 12 hybridised on a tissue section of case 12, kin’s lymphomas: unusual prevalence and pattern of c-myc a high-grade MALT lymphoma with trisomy 12 in the high-grade but rearrangements in aggressive lymphomas. Blood 1990; 76: 797– not the low-grade tumour component. (A) depicts cells of the low- 800. grade component, showing nuclei with nil, one or two spots. (B) 10 Peng H, Diss T. Isaacson PG, Pan L. c-myc gene abnormalities in depicts cells of the high-grade component, in addition showing nuclei mucosa-associated lymphoid tissue (MALT) lymphomas. J Pathol with three spots. Images were captured using a COHU 4910 series 1997; 181: 381–386. monochrome CCD camera (COHU, San Diego, CA, USA) attached 11 Chong JM, Fukayama M, Hayashi Y, Hishima T, Funata N, Koike to a DM fluorescence microscope (Leica, Wetzlar, Germany) × M, Matsuya S, Konishi M, Miyaki M. Microsatellite instability and equipped with a PL Fluotar 100 , NA 1.30 to 0.60 objective and 13 loss of heterozygosity in gastric lymphoma. Lab Invest 1997; 77: and N2.1 filters (Leica), and Leica QFISH software (Leica Imaging Sys- 639–645. tems, Cambridge, UK). Images were processed using Adobe Photo- 12 Isaacson PG. Recent developments in our understanding of gastric shop (Version 3.0; Adobe Systems, Mountain View, CA, USA). lymphomas. Am J Surg Pathol 1996; 20 (Suppl. 1): S1–S7. 13 Isaacson PG, Spencer J. Malignant lymphoma of mucosa-associa- phomas. This strongly supports the hypothesis that high-grade ted lymphoid tissue. Histopathology 1987; 11: 445–462. 14 Hedley DW, Friedlander ML, Taylor IW. Application of DNA flow lymphomas may develop via different pathogenetic pathways. cytometry to paraffin-embedded archival material for the study of In conclusion, trisomy 3 could be detected in approxi- aneuploidy and its clinical significance. Cytometry 1985; 6: mately one-third of the low-grade MALT lymphomas. In high- 327–333. grade MALT lymphoma the incidence of trisomy 3 was not 15 Schutte B, Reynders MM, Bosman FT, Blijham GH. Flow cyto- increased, while trisomies of other chromosomes occurred metric determination of DNA ploidy level in nuclei isolated from more frequently. On the other hand, in cases with separate paraffin-embedded tissue. Cytometry 1985; 6: 26–30. 16 Cooke HJ, Hindley J. Cloning of human satellite III DNA: different low-and high-grade components, trisomy 3 was detected in components are on different chromosomes. Nucleic Acids Res both components. Taken together, these data indicate that at 1979; 6: 3177–3197. least some high-grade MALT lymphomas develop from low- 17 Waye JS, Willard HF. Chromosome specificity of satellite DNAs: grade tumour, but that trisomy 3 does not play a major role short- and long-range organization of a diverged dimeric subset in this progression. The finding that primary large cell lym- of human alpha satellite from chromosome 3. Chromosoma 1989; phomas without features of MALT lymphoma appear to con- 97: 475–480. 18 Waye JS, England SB, Willard HF. Genomic organization of alpha tain less numerical aberrations than high-grade MALT lym- satellite DNA on human chromosome 7: evidence for two distinct phoma might indicate that these two types of aggressive alphoid domains on a single chromosome. Mol Cell Biol 1987; lymphomas have distinct oncogenetic pathways. 7: 349–356. Trisomy in different types of gastric lymphoma MA Hoeve et al 807 19 Donlon T, Wyman AR, Mulholland J, Barker D, Bruns G, Latt S, phoma. A possible model for cancer. New Engl J Med 1987; 316: Botstein D. a satellite-like sequences at the centromere of chromo- 79–84. some No. 8. Am J Hum Pathol 1986; 39: (A196). 36 Offit K, Richardson ME, Ceng QQ, Hampton A, Koduru PR, Jhan- 20 Moyzis RK, Albright KL, Bartholdi MF, Cram LS, Deaven LL, Hilde- war SC, Filippa DA, Lieberman PH, Clarkson B, Chaganti RS. Non- brand CE, Joste NE, Longmire JL, Meyne J, Schwarzacher Robinson random chromosomal aberrations are associated with sites of T. Human chromosome-specific repetitive DNA sequences: novel tissue involvement in non-Hodgkin’s lymphoma. Cancer Genet markers for genetic analysis. Chromosoma 1987; 95: 375–386. Cytogenet 1989; 37: 85–93. 21 Waye JS, Creeper LA, Willard HF. Organization and evolution of 37 Offit K, Chaganti RS. Chromosomal aberrations in non-Hodgkin’s alpha satellite DNA from human chromosome 11. Chromosoma lymphoma. Biologic and clinical correlations. Hematol Oncol 1987; 95: 182–188. Clin North Am 1991; 5: 853–869. 22 Looijenga L, Smit V, Wessels J. Localization and polymorphism 38 Dierlamm J, Michaux L, Wlodarska I, Pittaluga S, Zeller W, Stul of a chromosome 12-specific alpha-satellite DNA sequence. Cyto- M, Criel A, Thomas J, Boogaerts M, Delaere P et al. Trisomy 3 in genet Cell Genet 1990; 54: 216–218. marginal zone B-cell lymphoma: a study based on cytogenetic 23 Higgins MJ, Wang HS, Shtromas I, Haliotis T, Roder JC, Holden analysis and fluorescence in situ hybridization. Br J Haematol JJ, White BN. Organization of a repetitive human 1.8 kb Kpnl 1996; 93: 242–249. sequence localized in the heterochromatin of chromosome 15. 39 Stolte M, Kroher G, Meining A, Morgner A, Bayerdorffer E, Bethke Chromosoma 1985; 93: 77–86. B. A comparison of Helicobacter pylori and H. heilmannii 24 Waye JS, Willard HF. Molecular analysis of a deletion polymor- gastritis. A matched control study involving 404 patients. Scand J phism in alpha satellite of human chromosome 17: evidence for Gastroenterol 1997; 32: 28–33. 40 Zhang Y, Cheung AN, Chan AC, Shen DH, Xu WS, Chung LP, Ho homologous unequal crossing-over and subsequent fixation. FC. Detection of trisomy 3 in primary gastric B-cell lymphoma by Nucleic Acids Res 1986; 14: 6915–6927. using chromosome in situ hybridization on paraffin sections. Am 25 Devilee P, Cremer T, Slagboom P, Bakker E, Scholl HP, Hager J Clin Pathol 1998; 110: 347–353. HD, Stevenson AF, Cornelisse CJ, Pearson PL. Two subsets of 41 Barth TF, Dohner H, Werner CA, Stilgenbauer S, Schlotter M, Paw- human alphoid repetitive DNA show distinct preferential localiz- lita M, Lichter P, Moller P, Bentz M. Characteristic pattern of ation in the pericentric regions of chromosomes 13, 18, and 21. chromosomal gains and losses in primary large B-cell lymphomas Cytogenet Cell Genet 1986; 41: 193–201. of the gastrointestinal tract. Blood 1998; 91: 4321–4330. 26 Bobrow MN, Harris TD, Shaughnessy KJ, Litt GJ. Catalyzed 42 Ott G, Kalla J, Steinhoff A, Rosenwald A, Katzenberger T, Roblick reporter deposition, a novel method of signal amplification. Appli- U, Ott MM, Muller Hermelink HK. Trisomy 3 is not a common cation to immunoassays. J Immunol Meth 1989; 125: 279–285. feature in malignant lymphomas of mucosa-associated lymphoid 27 Hopman AH, Ramaekers FC, Raap AK, Beck JL, Devilee P, van tissue type. Am J Pathol 1998; 153: 689–694. der Ploeg M, Vooijs GP. In situ hybridization as a tool to study 43 Peng H, Du M, Diss TC, Isaacson PG, Pan L. Genetic evidence numerical chromosome aberrations in solid bladder tumors. Histo- for a clonal link between low and high-grade components in gas- chemistry 1988; 89: 307–316. tric MALT B-cell lymphoma. Histopathology 1997; 30: 425–429. 28 Hopman AH, Poddighe PJ, Smeets AW, Moesker O, Beck JL, 44 de Jong D, Boot H, van Heerde P, Hart GA, Taal BG. Histological Vooijs GP, Ramaekers FC. Detection of numerical chromosome grading in gastric lymphoma: pretreatment criteria and clinical rel- aberrations in bladder cancer by in situ hybridization. Am J Pathol evance. Gastroenterology 1997; 112: 1466–1474. 1989; 135: 1105–1117. 45 Hallas C, Greiner A, Peters K, Muller Hermelink HK. Immunoglob- 29 Hopman AH, van Hooren E, van de Kaa CA, Vooijs PG, Ramaek- ulin VH genes of high-grade mucosa-associated lymphoid tissue ers FC. Detection of numerical chromosome aberrations using in lymphomas show a high load of somatic mutations and evidence situ hybridization in paraffin sections of routinely processed blad- of antigen-dependent affinity maturation. Lab Invest 1998; 78: der cancers. Mod Pathol 1991; 4: 503–513. 277–287. 30 Raghoebier S, Kibbelaar RE, Kleiverda JK, Kluin Nelemans JC, van 46 Whang Peng J, Knutsen T, Jaffe ES, Raffeld M, Zhao WP, Duffey Krieken JH, Kok F, Kluin PM. Mosaicism of trisomy 12 in chronic P, Longo DL. Cytogenetic study of two cases with lymphoma of lymphocytic leukemia detected by non-radioactive in situ mucosa-associated lymphoid tissue. Cancer Genet Cytogenet hybridization. Leukemia 1992; 6: 1220–1226. 1994; 77: 74–80. 31 van Dekken H, Bosman FT, Teijgeman R, Vissers CJ, Tersteeg TA, 47 Dierlamm J, Pittaluga S, Wlodarska I, Stul M, Thomas J, Boogaerts Kerstens HM, Vooijs GP, Verhofstad AA. Identification of numeri- M, Michaux L, Driessen A, Mecucci C, Cassiman JJ et al. Marginal cal chromosome aberrations in archival tumours by in situ zone B-cell lymphomas of different sites share similar cytogenetic hybridization to routine paraffin sections: evaluation of 23 phae- and morphologic features. Blood 1996; 87: 299–307. 48 Brynes RK, Almaguer PD, Leathery KE, McCourty A, Arber DA, ochromocytomas. J Pathol 1993; 171: 161–171. Medeiros LJ, Nathwani BN. Numerical cytogenetic abnormalities 32 Fifth International Workshop on Chromosomes in Leukemia–Lym- of chromosomes 3, 7, and 12 in marginal zone B-cell lymphomas. phoma. Correlation of chromosome abnormalities with histologic Mod Pathol 1996; 9: 995–1000. and immunologic characteristics in non-Hodgkin’s lymphoma and 49 Bjerrum OW, Phillip P, Muller Berat N, Hertz H, Killmann SA. adult T cell leukemia-lymphoma. Blood 1987; 70: 1554–1564. Acute lymphocytic leukaemia with t(4;11): a clinical subentity. 33 Schouten HC, Sanger WG, Weisenburger DD, Anderson J, Armit- Scand J Haematol 1985; 35: 96–101. age JO. Chromosomal abnormalities in untreated patients with 50 Gaidano G, Volpe G, Pastore C, Chiarle R, Capello D, Gloghini non-Hodgkin’s lymphoma: associations with histology, clinical A, Perissinotto E, Savinelli F, Bosco M, Mazza U et al. Detection characteristics, and treatment outcome. The Nebraska Lymphoma of BCL-6 rearrangements and p53 mutations in Malt-lymphomas. Study Group. Blood 1990; 75: 1841–1847. Am J Hematol 1997; 56: 206–213. 34 Offit K, Jhanwar SC, Ladanyi M, Filippa DA, Chaganti RS. Cyto- 51 Robledo M, Benitez J, Rivas C, Martinez Castro P. Cytogenetic genetic analysis of 434 consecutively ascertained specimens of study of B-cell lymphoma of mucosa-associated lymphoid tissue non-Hodgkin’s lymphoma: correlations between recurrent aber- (letter). Cancer Genet Cytogenet 1992; 62: 208–209. rations, histology, and exposure to cytotoxic treatment. Genes 52 Du M, Peng H, Singh N, Isaacson PG, Pan L. The accumulation Chromosomes Cancer 1991; 3: 189–201. of p53 abnormalities is associated with progression of mucosa- 35 Yunis JJ, Frizzera G, Oken MM, McKenna J, Theologides A, Arn- associated lymphoid tissue lymphoma. Blood 1995; 86: 4587– esen M. Multiple recurrent genomic defects in follicular lym- 4593.