Document zo49DG9pkpmvxxJ3OEo8vpa2m

cc x b hNTIFICATION OF COMMON GERMINAL-CENTER B-CELL PRECURSORS -.*-=----.HODGKIN'S DISEASE AND NON-HODGKIVS -.. .-.. ., MARTIN-LEOHANSMANN, M.D.,JOHN G. STRICKLERM, .D., REINHARD DUMMER, M.D., R BURG, M.D., KLAUS RAJEWSKY,M.D., AND RALF K~PPERSP, H.D. -;GI -*i- 1 u-ABSTRACT , Bac&vllttrt Hodgkin`s disease and non-Hodgkin's B-cell lymphoma occasionally occur in the same patient. The identification of a common precursor of 'the two types of lymphoma would show definitively `that Reed-Sternbergcells originate from B cells. Ma&ods We studied lymphomas from two pa- tients, one with a composite lymphoma (classic Hodgkin's disease and a follicular lymphoma in the same lymph node) and the other with a T-cell-rich B-cell lymphoma that was followed by classic Hodgkin`s disease. Single Reed-Sternberg cells and non-Hodgkin`s lymphoma cells from frozen sections were micromanipulated.The rearrangedimmunoglobulin variable-region genes ( V genes) of the heavy and light chains were amplified by the polymerase chain reaction from genomic DNA and sequenced. Rwrcla In both patients, the Reed-Sternberg cells were related clonally to the non-Hodgkin's lymphoma B cells. The V genes carried somatic mutations (a hallmark of germinal-center B cells and their descendants). In both patients, some somatic mutations were shared by the Reed-Sternberg and nonHodgkin`s lymphoma cells, whereas other somatic mutationswere found exclusively in one or the other cell type. Conclusions In two patients with classic Hodg- kin's disease and non-Hodgkin`s B-cell lymphoma, we identified a common B-cell precursor, probably a germinal-center B cell, for both lymphomas. This finding suggests that the two types of lymphoma un- derwent both shared and distincttransforming events and provides proof of the B-cell derivation of Reed- Sternberg cells in classic Hodgkin`s disease. (N Engl J Med 1999;340:1239-47.) Q1999, Massachusetts Medical Society. mune response within a germinal center; these mutations are the distinctive feature of germinal-center B cells and their descendants.9 In cl&c Hodgkin's disease, the origin of Reed-Stemberg cells from germinal-center B cells was indicated by the finding that in some cases the Vgenes were rendered nonfunctional by crippling mutations (such as mutations treating stop codons).2 Because such mutations usually cause the death of B cells by apoptosis within the germinal center, we surmised that the precursors of the Reed-Sternberg cells must have resided within the germinal center and been rescued from apoptosis by a transforming event.2.10 However, because Reed-Sternberg cells express several genes that are thought to be specific to dendritic or monocytic cells, or both, it has also been argued that Reed- Stemberg cells represent an unusual type of dendritic or monocytic cell."-14 In rare cases, Hodglun's disease and a non-Hodgkin's B-cell lymphoma occu in the same patient, either simultaneously or sequentially.15-'7 Definitive proof of the origin of Reed-Sternberg cells from B cells could be obtained by demonstrating that in such cases the B-cell lymphoma and the Reed-Stemberg cells are members of the same B-cell clone. In the current study, we used micromanipulation of sin- gle Reed-Sternberg cells and non-Hodgkin's B-cell lymphoma cells fiom twopatients with such cases, and we amplified rearranged immunoglobulin V genes from genomic DNA of the single cells. Because rearrangements of immunoglobulin genes are highly diverse among B cells, they represent ideal clonal markers for these cells. In both patients in our study, the Reed-Stemberg cells and the non-Hodgkin's B-cell lymphoma cells shared a common germinalcenter B-cell precursor. OLECULAR studies of single Hodgkin and Reed-Sternberg cells (hereafter referred to as Reed-Sternberg cells) in classic Hodglun's disease (i.e., nodu- lar sclerosis, mixed cellularity, and lymphocyte depletion) and lymphocyte-predominant Hodglun's disease have shown that these cells are clonal populations of transformed B lymphocytes with somatically mutated .immunoglobulin variable-region genes (V genes).1-8 Somatic mutations are introduced into rearranged Vgenes of B cells that are participating in an im- CASE REPORTS Patient 1 was a 75-year-old woman who presented with retroperitoneal lymphadenopathy in June 1990. A biopsy of retropentoned nodes showed a composite follicular small-cleaved-cell From the Dcparunent of Pathology, University of Frankfiur, Fnnkfurt am hhh, Germany (A.B., M:L.H.); the Department of Pathology, May0 Clinic, Rochester,Minn. (J.G.S.);the Department of Dermatology, Univcrsifitsspital Ziirich, Zurich, Switzerland (R.D., G.B.); and the Institute for Generics,Universityof Cologne, Cologne, Germany(K.R., R.K.).Addrcss reprint requests to Dr. Kiippcrs at the university of Cologne, University Hospital,L.FI E4 R706,Joseph-Stclzmannstr9. ,50931 Cologne,Germany, or at rkuppersBmac.gcnctik.uni-koc1n.de. Volume 340 Number 16 * 1239 The New England urnal of Medicine lymphoma and, in the same lymph node, mixcd-cellularity Hodgkin`s disease with abundant Reed-Stcrnbcrg cells. The RcedSternberg cells wcre CD30- and CD15-positive, whereas the tumor cells of the follicular lymphoma were negative for both m a r h . On a section of paraffin-embedded material, some ReedSternberg cells were CD20-positive. The patient was considered to have stagc IVB disease on the basis of the presence of follicular lymphoma in the bone marrow and liver. She was mated with cight cycles of cyclophosphamide, vinblastine, procarbazinc, and prednisone but the disease recumd (recumncc was not confirmed by biopsy) in December 1991. Patient 2 was a 50-par-old man who presented with a T-cellrich B-cell lymphoma ofthe skin in January 1994. The lymphoma was surgically removed, and no additional therapy was given. In January 1997, classic Hodgkin's disease was diagnosed in an axillary lymph node. The Reed-Stcrnberg cells were CD30- and CD15-positive. The patient also had Gardner's syndrome, and he underwent colectomy in 1992. METHODS Immunostaining and Micromanipulation We stained frozen tissue sections,5 to 10 pm thick, with antiCD20 ( U 6 ;Dako, Hamburg, Germany) or anti-CD3O (BcrH2; Dako) monoclonal antibodics.1*J9We visualized bound alkaline phosphatase with fast red. We overlaid stained sections with TRIS-buffered saline, and wc isolated single cdls using two hydraulic micromanipulators.~~I~so9lated cells were stored at -20C in 2 0 4 Expand high-fidelity polymerase-chain-rcaction (PCR) bu&r (BochringcrMannhcim, Mannhcim, Germany).From each section used for micromanipulation, we aspirated aliquots of the buffer covering the section and used them as negative controls in parallel with the analysis of the cdl samples. Single-Cell PCR Analysis We amplified rcarrangcd immunoglobulingenes from genomic DNA of single cells in a seminested PCR analysis using either Family-specific V, (variable-region heavy-chain gene) leader or framework region I primers and V, or V, (variable-region K or h light-chain) framework region I family-specific primers together with two setS of the respective joining-region ( J genc)-scgment ~ r i m c r s . ~ WJ Je~carried out the k t round of amplification using a collection of primers for V, (and V ,or V,) genes together with the respective J-segment primers. In the second round of amplification, we furrhcr amplified 1.5-4 aliquots (3 percent) fromthe first round with the same Vgenc primers in separate reactions and internal J-segment primer mixes. The PCR products wcre gel-purified and dircctly sequenced on an automatic sequencer (ABI377, Applied Biosystems, Wietcrstadt, Germany). We analyzed the sequences using the data bases of Immunogenetics and GenBank. Reed-Stcrnbcrg cells and non-Hodgkin's B-cell lymphoma cells wcrc analyzed in parallel with buffer controls. RESULTS PCR Analysis of Immunoglobulin Gene Rearrangements in Single Reed-Sternberg and Non-Hodgkin`s Lymphoma B Cells In both cases - the first a composite lymphoma consisting of classic Hodglun's disease and a follicular lymphoma in the same lymph node, and the second a T-cell-rich B-cell lymphoma of the skin followed by classic Hodgkin's disease in a lymph node three years later - the Reed-Stemberg cells were CD3O-positive and CD2O-negative, whereas the nonHodglun's lymphoma cells were CD30-negative and CD2O-positive(Fig. 1);in the composite lymphoma, some Reed-Stemberg cellswere CD2O-positive. Consequently, for the micromanipulation of single cells, Reed-Sternberg cells in frozen sectionswere stained with the anti-CD30 antibody, and the non-Hodgkin's lymphoma cells in fiozen sections were stained with the anti-CD20 antibody. In the case of Patient 1, we analyzed 22 ReedSternberg cells for immunoglobulin gene rearrangements (Table 1)(not all cells generated a PCRproduct). The same V, gene rearrangement was amplified from 12 of the 22 Reed-Sternberg cells, and the `, same V , (light chain) gene was amplified from 4 of the 12 cells analyzed for V , gene rearrangements. We obtained a clonal V, rearrangement from four of five cells analyzed for V, (light chain) gene rearrangements. The same three clonal V, V,, and V, gene rearrangements were also obtained from cells of the follicular lymphoma: the V, gene fiom 26 of 35 samples (each containing two cells), the V , gene from 5 of 12 samples, and the V, gene fiom 4 of 5 samples (Table 1).From 1 of 11 buffer controls, we amplified a V, gene rearrangement, which was most likely due to cellular contamination. One V, gene that differed from the clonal rearrangement was amplified from a CD20-positive cell; it may have represented a rare normal B cell present in the fol- licular lymphoma. In the case of Patient 2, we amplified clonal V, and clonal V ,gene rearrangementsfiom 18 and 21 of 60 Reed-Sternberg cells, respectively (Table 1).The same rearrangements were also amplified from T-cell-rich B-cell lymphoma cells, the V, gene fiom 23 of 60 cells and the V , gene from 9 of 60 cells. We am- plified three different V gene rearrangements fiom 1Reed-Sternberg cell and 2 of 48 buffer controls, We amplified one clonal V, gene fiom another buff- er control (Table 1).We assumed that these PCR products represented cellular contamination. The presence of the same clonal Vgene rearrangements in the Reed-Sternberg cells of Hodgkin's disease and the associated non-Hodglun's lymphoma .cells indicates that a single B cell is the common pre- cursor of both lymphomas. Analysis of Somatic-Mutation Patterns The V, and V, gene rearrangementsamplified h m the Reed-Sternberg cells of Patient 1were potent i d y functional (translZa61einto protein RNA) and somatically mutated with mutation ficquencies of 9.6 and 5.8 percent, respectively (Table 2). No intraclonal diversity was observed. Most somatic mutations in the V, and V, genes of the Reed-Sternberg cells were shared by the corresponding rearrangements of the follicular lymphoma. However, in the heavy-chain rearrangement, the ked-Stemberg c c b harbored two point mutations that were absent in the V, gene of the follicularlymphoma, and the cells lacked one point mutation present in the V, gene of 1240 * April 22, 1999 C O M M O N 6-CELL PRECURSORS IN PATIENTS W I T H H-O D G K I N ' S DISEASE A N D N O N - H O D G K I N ' S L Y M P H O M A . '1 ! 'I Figure 1. lmmunostaining of the Composite Lymphoma from Patient 1 and the T-Cell-Rich B-Cell Lymphoma and Hodgkin's Disease infiltrate from Patient 2. Panel A shows the composite lymphoma from Patient 1 stained for CD20 with use of peroxidase. Positive immunostaining of the follicular lymphoma is shown at the upper right. Reed-Sternberg cells (arrows) were CD2O-negative. Panel B shows the composite lymphoma stained for CD30 with use of alkaline phosphatase. Cells of the follicular lymphoma were negative, whereas Reed-Sternberg cells (arrows)were CD30-positive. Panel C shows antiLCD30 staining of a frozen section from the composite lymphoma. Panel D shows a skin-biopsy specimen of the T-cell-rich B-cell lymphoma from Patient 2. It shows numerous large CD20-positive blast cells in the corium (alkaline phosphatase). Panel E shows the T-cell-rich B-cell lymphoma with large CD20-positive B blasts surrounded by T lymphocytes. CD20 staining was visualized with alkaline phosphatase and fast red. Panel F shows Hodgkin's disease infiltrate in a lymph-node-biopsy specimen from Patient 2. Three CD30-positive Reed-Sternberg cells can be seen. CD30 staining was visualized with alkaline phosphatase and fast red. V o l u r n c 3 4 0 N u i i i h c r I h . 1241 The New England Journal of Medicine TABLE1. PCR AND SEQUENCE ANALYSIS OF V GENE REARRANGEMENTS OF SINGLE REED-STERNBERG ANTI NoN-HoDGKI"L~YMPHOMA CELLS. PATIEN7 No. No. OF POMM SAMPIES PCR Pmmucrs no. poritivd total no. 1 Hodgkin's disease Bu&r controlst Follicular lymphoma Bu& controlst 2 Hodgkin's discax Buffer controlst T-cdl-rich B - c ~ lymphoma Bu&r controlst 12/22 4/12 4/5 1/7 013 0/2 27/35$ 5/12$ 4/5$ 0/4 0/2 o/z 18/60 21/60 2/24 0/24 23/60 9/60 1/24 0/24 12 v,4 4 v,l 4 VJ 1v - 27 V,,4 5 v,l 4 V,l 18 V3, 21 v,2 1 V,l, 1 v,3 23 V3, 9 v9 1 v,3 12 4 4 1 26 1 5 4 17 1 21 2 2011 9 1 * S d primer collectionswere used for the analysis ofthe two cases: V, framework region I family-specific primers together with V, 6ammork region I family-spccilic primers; V, leader region funily-specificprimers alone (Patient 1)or in combination with V ,hneworlr region I funily-spedfic primers (Patient 2); and V, 6amavork region I funily-specific primers together with V,fnmcwort region I fimily-specific primers. In both cases,V, and V, genes wcrc copmpli6cdrepeatedlyfromthe same cclls. tBu&r conmls wcrc aiiquots ofthe b& cowing the tissue sectionsduring micromanipulation. Additional controlswith tubes lacking cclls or aspirated b&r were also included but arc not shown here. They wvc consistentlynegative. $Because the effidency of PCR for the amplifiution of a V gene from a single ccll is usually lcss than 50 percent, OM)naghboring CD2O cells were put inm one d o n tube m shorten mimmanipulation and PCR analysis. Eighteen of the samples were obtained by micromanipulating a group of 30 to 40 B cclls and dividing thc DNA ofthcsc cclls into 18 samples after proteinve K digestion, so that each tube contained on average the DNA of 2 cclls. p e n cclls of each lymphoma and four bufcr conmk each were analyzed with a V, leader region family-s@c primer mix and primers spcci6c for Al7K gene rcanangunents. (Thesequcaceofthc Al7-spCcific prim- wu SCTACCCITCCCITGAflWCAGGAC3'.) llThrcc V d PCR products were not sequenced. .? I" I .. ..., . .. -. the fbllicular lymphoma (Fig.2).Likewise, 15 point mutations ia the V, gene were shared by the Reed- Sternberg and fofi&lar-lyrnphoma ceh, but the fbrmer carried 2 additional D o i n t mutations. In the 'V and V, genes of the fobicular lymphoma, on- gomg somauc mutation was observed, as is typical of,follicular 'lymphorna.21.22 For the heavy-chain rearrangement, we observed 23 sequence variants among the 26 V, gene sequences (Fig.2 and Table 2). Most variants differed by one to four mutations . h mthe sequence with the smallest number of mu- iations. . < ' + The V, gene amplified from Reed-Sternberg and follicular-lymphoma cells of the composite ma was nonfimctional and unmutated, In B pressing imrnunoglobulhi-h light $ah, the &n: functional 'v: gene rearranemens that art oftm' present in su&t-cells are u s d y inactivated by 'Wk- tion of the C,gene and tion also abolishes somatic ing TJ,joints of the re genes.9 In Patient 2, the V, of the Reed-Sternberg cells fhctional and somatic frequenciesof 11.7and 8.3 percent, mpe&vely ( I April 22, 1999 t Q, r- ~ COMMON B-CELL P ~ i' s IN PATIENTS WITH HODGKIN'S DISEASE AND NON-HODGKIN'S LYMPHOMA TABL2E. V GENE SEQUENCEAND MUTATIONANALYSISOF IMMu"GENE &"GEMENlS FROM ~-STERN'BERGAND NON-HODGI~LNY'MSPHOMA CELLS. VGmE IVFurrv) wrm po7EN1wy MvrrnOrV IwnuaOnrU srouwcE hmcnowL Fuzauum Dwmaw Vuu~low no. with variations/ % total no.' &4-59 (VA) m a (K1) Vl-lg(v,l) Ycs No Ye5 Follicular lymphoma v,4-59 (VA) L24a(V,1) Vl-lg(V,l) Ycs No Ycs 2 Hodgkin'sdisusc VH3-30 ( V d ) Yq A17 (VZ) Ya T-cd-rich 3-cell VH3-30 (VH3) YCS lymphoma A17 (VZ) YCS 9.6 0 5.8 9.3 0 5.1 11.7 8.3 15.8 14.6 No -/12 No -/4 No -/4 YCS 23/26 No -/5 Yes 2/4 Ycs 2/17 Not 4 2 1 YCS 16/20 Yes 3/9 *The dash indicates that all sequencer werc identical. tTwo mixcd scqucnces with two nudcotidcs at one or two positions on the sequencing gels (rep- resenting variants of thc same clonal rearrangement) wue amphied, indicating thc pmncc of two copics of the rearranged immunoglobutingencs in the cdI of origin, Tq DNA polymerase errors in onc of the first PCR cydcs, or both. Thc Vgene scqucnccs reported in this u t i d e have been submitted to thc European Molc& Biology Laboratory data library under accession numbcrs NO11130 through AJ011139. For thc non-Hodgl;in's lymphoma, the V gene sequences with thc lowest numbers of mutations were submitted. ble 2). One V, sequence differed from the 17 others by a single nucleotide. The V gene rearrangements of the T-cell-rich B-cell lymphoma carried a higher load of somatic mutation than did the corresponding rearrangements of the Reed-Sternberg cells: 15.8 percent for the V, gene and 14.6 percent for the V , gene. In this patient, most of the mutations were unique to either the Reed-Sternberg cells or the T-cell-rich B-cell lymphoma cells. For the heavy- chain gene rearrangement, only 7 mutations were shared between the two lymphomas; 28 of the mutations in the V, gene of the Reed-Stemberg cells were not present in the T-cell-rich B-cell lymphoma cells, whereas the latter carried 39 mutations that were not present in the Reed-Stemberg celis (data not shown). The situationwas similar for the V ,lightI chain genes (data not shown). As in Patient 1, the non-Hodgkin'slymphoma cells showed considerable intraclonal diversity (Table 2). The pattern of somatic mutation in the rearranged V genes of the clonally related lymphomas is evidence of the reliability of our analysis. In both patients, the patterns of somatic mutations found in the clonally related immunoglobulin gene rearrangements of the Reed-Sternberg cells differed from those of the corresponding non-Hodgkin's lymphoma cells (Fig. 2). We found these distinct patterns of somatic mutation in different populations of tumor cells that were isolated and analyzed in parallel, ruling out the possibility of contamination. Because crippling mutations rendering previously functional rearrangements nonfunctional were detected in some cases of classic Hodglun's disease in our earlier studies,Z we amplified and sequenced additional upstream and downstream fragments of the Vgenes of the lymphomas analyzed here (data not shown). However, no obviously crippling mutations (such as stop codons) were detected. This finding does not argue against the possibility that ReedSternberg cells in these cases derive from crippled germinal-center B cells, because most types of crippling mutations (such as those that disturb heavy- or light-chain folding or reduce afhity to the selecting antigen) cannot be identified easily. DISCUSSION The occurrence of a non-Hodgkin's B-cell lymphoma and Hodgkin's disease in the same patient is rare.15.16 The most frequent combination of the two is a diffuse large-cell lymphoma that develops after lymphocyte-predominant Hodgkin's disease.15-17Because the B-cell origin of the Reed-Sternberg cells in the lymphocyte-predominant subtype of Hodgkin's disease has long been suspected,= a transformation to a high-grade lymphoma (as is well known to occur in several types of low-grade nonHodgkin's B-cell lymphoma) would not be surprising. However, studies of the clonal relation between lymphocyte-predominantHodgkin's disease and subsequent or concurrent non-Hodglun's B-cell lym- Volume 340 Number 16 * 1243 ... I. _. >.* .. A `>?~ ,I,C,..Pe .- The New England Journal of Medicine IIlIIIIlllll 11&111111111 1:11111::111 111111111111 I1 I1 I1 I1 I1 I1 I1 I1 I1 I1 I1 I1 ?rr?TTrrErrr 111111l11111 1 8 4 9YI YI YI YI YI YI YI YI I 1111111l1111 IIIIIIIIIIII IIIIIIIIIIII OOOOOOUOOQ~ 811111111111 IIIIIIIIIII 811111111111 11111111111 41 I I I I I I I I I I EII111II1111 e5 1l 1I 1I 1I 1I 1I 1I 1I 1I 1I 1I 111111111111 IIIIIIIIIIII I I I I I I II I I I I 11111lIIIIII 111111111111 IIIIIIIIIIII 111lIlIIllll IIIIIIIIIIII 111111111111 l~llllllllll 111111111111 IIIIIIIIIIII $2 I I h>.CI 11.1 t I 1 I I I I I 1 1-1 I I I, 1 . 1 . 1 I , I I & . I I 11.1 11.1 I I I I c I 1'1 I I I I I I I I 11111111111111 I I I I 1iI.I 1 1 . 1 I I L I I I I I 11.1 I I 1 I I & J J J J & ' dJ b J J J - . *e. 3' L.`COMMON B-CELL PRECURSORS IN PATIENTS WITH HODGKIN`S DISEASE AND NON-HODGKIN'S LYMPHOMA Figure 3. Scenariofor the Generation of a Composite Lymphoma. The horizontallines within the circles indicate a V gene rearrangement;vertical lines within the circles indicatesomatic mutations. phoma found a clonal relation between the two lym- to germinal-center B cells, these findings -together phomas in only some cases.24-27 In the example of with other morphologic, histologic, and immunohis- classic Hodgkin's disease in which the patient also tochemical features of the two types of lympho- had a non-Hodglun's lymphoma, only a single case mas30131 -indicate that the non-Hodglun's lympho- has been analyzed with molecular techniques.28 The mas in both patients were cancers derived from two lymphomas were clonally unrelated. In this in- germinal-center B cells. stance, the non-Hodgkin's lymphoma may have been In the composite lymphoma, most of the somatic related to the therapy received, because the patient mutations were the same in the Reed-Sternberg underwent radiotherapy and chemotherapy before cells as in the follicular-lymphoma cells. This finding the non-Hodgkin's lymphoma developed. indicates that a germinal-center B cell that had al- ,' Both lymphomas in each of the two patients we ready acquired a high load of somatic mutation gave studied were derived from a common B-cell precur- rise to two daughter cells, one of which was the or- sor. In Patient 2, in whom Hodglun's disease was di- igin of the Reed-Sternberg cell clone, and the other agnosed three years after the T-cell-rich B-cell lym- of the follicular lymphoma. In the other patient, the phoma, the Reed-Sternberg cells or their precursor number of shared mutations was lower (only 7 of a must have been present during those three years. total of 74 mutations in the heavy chain). Neverthe- In both patients, there was considerable ongoing less, in this case the pattern of Vgene mutations also mutation within the tumor cells of the non-Hodg- indicates that the tumor clones originated from a kin's lymphoma, as indicated by numerous variations common germinal-center B-cell progenitor. in the clonal sequences. This phenomenon is well In our previous work, the interpretation that the known in follicular lymphoma21J2 and also typical of Reed-Sternberg cells in classic Hodgkin's disease T-cell-rich B-cell lymphoma.29 Because the process derive from germinal-center B cells was based on the of somatichypermutation is thought to be restricted argument that transformed B cells with crippled i Volume 340 Number 16 * 1245 The New England Journal of Medicine - Vgenes reside within the germinal center and must have been rescued from apoptosis in this microenvironment by some transforming event.2.10 The two cases presented here address the issue of the cellular origin of Reed-Sternberg cells in classic Hodglun's disease independentlyand more directly. The h d i n g of a clonal relation and the pattern of Vgene mutations of the Reed-Stemberg and non-Hodgkin's lymphoma cells in the same patient unequivocally demonstrate the derivation of Reed-Sternberg cells fiom mature B cells and are further evidence of their germinal-center B-cell derivation. Because Reed-Sternberg cells express a number of genes typical of dendritic cells, monocytic cells, or both,ll-14 it has been argued that Reed-Stemberg cells derive h m dendriticcells in which immunoglobulin gene rearrangements accidentallyoccurred. The results presented here rule out thispossibility. Rather, the Reed-Sternberg cells are B cells that in the course of malignant transformation can acquire features of dendritic cells, an event that may be of key importance fbr the pathogenesis of Hodgkin's disease. The finding of a shared precursor of the lymphomas in each patient has implications for the pathogenesis of lymphomas. Because it is unlikely that two B cells belonging to the same germinal-center B-cell clone give rise to two lymphomas independently, we assume that both lymphomas in each patient shared one or more transforming events. These transform- ing events could have taken place before the tumor precursor entered the germinal center or in the course of the germinal-center reaction. Additional trans- forming events that are specific for the Reed-Sternberg or the non-Hodglun's lymphoma precursor cell presumably occurred later, probably within the ger- minal center, thus explaining the development of two distinct diseases fiom the same B-cell precursor (Fig. 3). Because the Epstein-Barrvirus is thought to participate in the pathogenesis of classic Hodgkin's dis- ease in some cases,32 the presence of Epstein-Barr virus was sought in the Reed-Sternberg and nonHodglun's lymphoma cells in both cases, but the results were negative (data not shown). Moreover, although most follicular lymphomas carry translocations of the bc2-2 oncogene into the immunoglobulin heavy-chain locus,33 we fiiled to detect such a translocation in the Reed-Stemberg and follicularlymphoma cells of the composite lymphoma using a PCR-based method (data not shown). For these reasons,the shared and unique transforming events involved in the development of these lymphomas remain to be identi6ed. In summary, the two combinations of Hodglun's disease and non-Hodgkin's lymphoma that we studied exemplify the dose relation between these diseases and provide definitive evidence of the origin of Reed-Stemberg cells from B cells in classic Hodg- kin's disease. Moreover, the development of two lp- phomas fiom members of a single germinal-center B-cell clone supports the emerging concept that molecular processes in germinal-center B cells play a decisive part in the malignant transformation of B lymphocytes and M e r supports the concept that Reed-Stemberg cells derive h m B cells resid- ing within the germinal center? 1 Supponxd by grants from the Dcutsche F i m c h u n g s g e m ~ (SFB502) and the Dcutsche Krcbshilfc. We are indebted to Mubaela Fah&, J d i a Jmdinsky, cbtistiarrc Gmbani, and Tanja Scbafijiw cxulkntncbnkal aahncc. REFERENCES ! 1. Bracuningcr A, K i i p p ~R, Saiddu JG,Wackcr HH, Rajcwsky K, Hansmann ML. Hod@ and Rccd-Stembergcells in lymphocyte prcdominant Hod@ disease rrprcscnt clonal populationsof geminal CUI- tcr-dcrived tumor B cells. Proc Nul Acad Sci U S A 1997949337-42. [Emcum,Proc Natl Acad Sci U S A 19979414211.1 2 Kanzlcr H, K i I p p R,Hansmann ML,Rajcwsky IC. Hod@ and Rccd-Stembergcells in Hodgkin's disease q m n t the outgrowth of a dominant tumor done dcrivcdfrom (crippled)pmbal center B cells. J Erp Mcd 1996;1841495-505. 3. K;mzlcerH, Hansmann ML,Xspp U,et al. Molecular single cell analysis demonstrates the derivationof aperipheral blood-derivedcdl line (Ll236) from the Hodgkin/Recd-Stcmbcrg c& of a Hodgkin's lymphoma patient. Blood 1996;873429-36. 4. K i i p p R,Rajcwdcy IC. Zhw M, et ai. Hod& dirclsc:Hod@and Reed-Stembcrgcdls picked from histological sections show donal immunoglobulin gene rearrangementsand appear to be derived from B cells at various smgu of development. Proc Natl Acad Sci U S A 1994;91:10%2- 6. 5. Mara6oti T,Hummel M, AnagnostopoulosI, et al. Origin of nodular lymphoqtc-predomhanc Hodgkin'sdisease from a donal expansion of highly m u d germinal-center B c&. N Engl J Med 1997;337453-8. S. Ohno T,Stribicy JA, Wu G, Hinrichs SH, Weiscnbtqcr DD, Chan WC. Clonality in nodular lymphocyte-predominantHodgkin's disease. N Engl J Med 1997;337459-65. 7. Vodrerodt M, Sovcs M, Kanzler H, et al. Detection of d o d H W and Reed-Stcmbcrgcellswith identicalsomafidly mutated and rearranged V H genes in &rent biopsies in relapsed Hodgkin's disarc. Blood 1998; 922899-907. 8. Irsch J, Nitsch S, Hansmaon M-L, et al. Isolation of viable Hodgkin and Reed-StcmbugfdLr from Hodgkindisease tispucs. Proc Nul Aud Sci U S A 1998;9510117-22. 9. Klein U,Goossens T,Fischer M,et al. Somatic hypcrmuncionin nord and W m e d human B cdls. Imm~noRl N 1998;162:261-80. 10. Kiippus R, Rajewsky K. The originof Hodgkin and Rccd/Stcmbcrg cells in Hodgkin's diecuc. Annu RCVImm~n011998;16:471-93. 11. Sorg UR,Morse TM,PattonWN,et 11. Hodgkin's cellsexpressCD83, a dendritic cdl lineage vsocivcd antigen. Pathology 199729294-9. 12. P i GS,P iJL,LanghoffE,et al. Fascin, ascnsitivcnew marker for Rccd-Stcmbugcclls of Hodgkin's disease: evidence for a dendriticor B cell derivation?Am J M o l 1997;150543-62. 13. Brown RE.Histogenesis of Reed-Stemberg and dendritic interdigitating cells in nodular sdmsing Hodgkin's disease: immunohistochemical cvidcnce for monoCy(0idprecursors.Ann Clin Lab Sa 1997;2732937. 14. Wcrstr6m KO,b e R. Hopsu-Haw VK,Jilrvincn M, Rinn; Hodgkin'sdisarc:a malignancy of folliculardendritic &il LvKet1 343:422-3. 16. Kim H. Compoaitc lymphomaand dated dirordcrs 1993;99:445-51. 16. JalEES,ZuUc-OsornoA, Kin in~oIuhipbctwccn~s Ann Oncol1994;5:SupplL7-11. 17. H a n s m a ~ML, Stein H, Fcllbaurn C, Hui PK,pywuach nut K. Nodular paragrandomacan puwfonnintohigh-grade lymphoma O f B tvpc. H u P~ath01 1989;201169-75. 18. K i I p p R, Zhao M, Hvumnnn ML,Rajewsky K. `Ihciag B cell vclopment in human gvminal CUIWby mole& analysisof sin picked fromhistological scaions. EMBO J 1993;124955-67. 1246 * April 22, 1999 -_den,Mas.:Blackdl Sc-i.e.nce, 1997: ' I -~ k n y H M , G o a a m v l K S . s e - l - c qs,et al. Clusteringof utensivr somatic Bwkitt's Gphoma arising&om lymphocywpradominant Hodghn'sdic-, ew.Br J Haemat01 19%:95:380-2. . .;? .'i clonal R c c d - S & b cell population and a subaequcnt B-cell small noncleaved cell Ivmuhoma. Mod Pathol 15 . rt*3"-,T2 nancc Hodgtias disZGc-a distinct cbiwpathologjcdentity. Am J Surg athd 1m,18526-30: 24. Greiner TC,GaMoyneRD, AndersonME, et al. Nodularlymphocytc- predo&t Hodgtin's disuseassochd with large-celllymphoma: analy- Sic of Ig gsm rurranpncnts by V-J polymerase chain rtaction. Blood 19%;88657-66. 25. Pan LX,Dirs TC,Pcng HZ,Norton 47,Isaacson PG.Nodular lym- phocyte predominance Hodgkds disax:a monoclonal or polydonal B-cell disordu)Blood 1996;87:2428-34. 26. Wickrt RS,W-b- DD, liucns A, Grciner TC,chvlWC. Clod rclatioaJhipbctwcmlymphoqic predominanceHodgkin's disease -non-Hod&n's 1ymphoG:a PrOgrrsMdform of follidc ccnm cell lym- phoma and lymphocyte predominance Hodgkin's disease. 1996;28:15-24. 31. Hvris NL,JdEES,Stcin H,et al. A rrvisedE ~ ~ w - A m e rdias~-n sification of lymphoid neoplasm: a pmposd from the InternationalLymphoma SIX+ Grow.Blood 1994;841361-92. 32. Oudejans JJ,JimNM,Majer CJLM.Epstein-Barrvirus in Hodgkin's Idisease: more rhan just an innocent bystander. J Pathol 1997;181:353-6. 33. Cotter FE. The role of the bd-2 gene in lymphoma. Br J H+cmud 1990;75449-53. FULL TEXT OF ALL JOURNAL ARTICLES ON THE WORLD WIDE WEB Access to the complete text of the Journnl on the Internet is fkcc to all subscribers. To use thisWeb site, subscribers should go to the Jowrnaf's home page (www.ncjm.org) and register by enteringtheir names and subscriber numbers as they appear on their mailing labels. Aftcr this one-time registration, subscribcrs can use their passwords to iog on for electronicacccss to the entire JounrnC &om any computer that is C O M C C ~ C ~ to the Internet. Fcaturrs include a library of all issues since January1993,a full-tat seardl capacity, a personal archive fbr saving artides and scarch d t s of interest, and fiec software for downloading articles so they can be printed in a format that is virmally identical to that of the typeset pages.