Normal Germinal Center Research Articles

Expression of the human germinal center–associated lymphoma (HGAL) protein, a new marker of germinal center B-cell derivation

AbstractWe identified the human germinal center-associated lymphoma (HGAL) in gene-expression profiling studies of diffuse large B-cell lymphoma (DLBCL). The expression of HGAL correlated with survival in patients with DLBCL. The HGAL gene is the human homolog of M17, a mouse gene expressed specifically in normal germinal center (GC) B cells. We generated a monoclonal antibody against the HGAL protein and show that HGAL is expressed in the cytoplasm of GC lymphocytes and in lymphomas of GC derivation. Among 727 lymphomas tested by immunohistochemistry on tissue microarrays, HGAL staining was found in follicular lymphomas (103 of 107), Burkitt lymphomas (40 of 40), mediastinal large B lymphomas (7 of 8), and in DLBCLs (103 of 151). Most marginal zone lymphomas lacked HGAL staining. Lymphocyte-predominant Hodgkin lymphomas (12 of 17) and, surprisingly, classical Hodgkin lymphomas (78 of 107) were found to be positive. Hierarchical clustering of comparative immunohistologic results in DLBCLs demonstrates that the expression of HGAL is similar to 2 other GC-associated proteins, BCL6 and CD10, but different from 2 markers associated with a non-GC phenotype, MUM1/IRF4 and BCL2. The restricted expression and GC specificity of HGAL protein suggest that it may have an important role in the diagnosis of specific lymphomas, and, potentially in the identification of subtypes associated with different prognoses.

PU.1 protein expression has a positive linear association with protein expression of germinal centre B cell genes including BCL‐6, CD10, CD20 and CD22: identification of PU.1 putative binding sites in the BCL‐6 promotor

The transcription factor PU.1 has been shown to be crucial for the early stages of B cell development but its function at later stages of B cell development is less well known. We observed previously that PU.1 is expressed uniformly throughout the mature pre-plasma cell B cell population, the only exception being a subpopulation of germinal centre (GC) cells which showed exceptionally high expression of PU.1. This suggested that PU.1 may also have a role in GC B cell biology. To test this hypothesis and to screen for possible genes regulated by PU.1, we first evaluated semi-quantitatively the possible co-expression of PU.1 with proteins known to be upregulated or downregulated during GC B cell development. Normal lymphoid tissues and 255 B cell non-Hodgkin lymphomas of putative GC B cell origin were evaluated. PU.1 expression was positively associated with CD10 (p < 0.0001), CD20 (p = 0.043), CD22 (p = 0.005), CD79a (p = 0.024) and Bcl-6 (p < 0.0001) and negatively associated with cytoplasmic immunoglobulin light-chain expression (p = 0.036) in diffuse large B cell lymphoma. Identical or nearly identical associations were found in follicular lymphoma. Since CD20 is known to be partly regulated by PU.1 and putative PU.1-binding sites have been described in the regulatory regions of the CD22, CD79a and CD10 genes, we looked for putative PU.1 binding sites in the BCL6 promotor. Four such putative PU.1 binding sites were identified. Further analysis by gel-shift electromobility essay showed that PU.1 protein binds to three of the four putative binding sites in the BCL6 promotor. PU.1 and Bcl-6 were also found to be upregulated in centroblasts in the normal GC, but jointly downregulated in a subpopulation of centrocytes. Our findings support the contention that PU.1 may also have an important role in GC B cell development.

MALT1 and the API2-MALT1 fusion act between CD40 and IKK and confer NF-κB-dependent proliferative advantage and resistance against FAS-induced cell death in B cells

AbstractThe most frequently recurring translocations in mucosa-associated lymphoid tissue (MALT) B-cell non-Hodgkin lymphoma, t(11;18)(q21;q21) and t(14;18)(q32; q21), lead to formation of an API2-MALT1 fusion or IgH-mediated MALT1 overexpression. Various approaches have implicated these proteins in nuclear factor κB (NF-κB) signaling, but this has not been shown experimentally in human B cells. Immunohistochemistry showed that MALT1 is predominantly expressed in normal and malignant germinal center B cells, corresponding to the differentiation stage of MALT lymphoma. We expressed MALT1 and apoptosis inhibitor-2 API2/MALT1 in human B-cell lymphoma BJAB cells and found both transgenes in membrane lipid rafts along with endogenous MALT1 and 2 binding partners involved in NF-κB signaling, B-cell lymphoma 10 (BCL10) and CARMA1 (caspase recruitment domain [CARD]-containing membrane-associated guanylate kinase [MAGUK] 1). API2-MALT1 and exogenous MALT1 increased constitutive NF-κB activity and enhanced IκB kinase (IKK) activation induced by CD40 stimulation. Both transgenes protected BJAB cells from FAS (CD95)-induced death, consistent with increases in NF-κB cytoprotective target gene expression, and increased their proliferation rate. Expression of a dominant-negative IκBα mutant showed that these survival and proliferative advantages are dependent on elevated constitutive NF-κB activity. Our findings support a model in which NF-κB signaling, once activated in a CD40-dependent immune response, is maintained and enhanced through deregulation of MALT1 or formation of an API2-MALT1 fusion.

Differential expression of activation‐induced cytidine deaminase (AID) in nodular lymphocyte‐predominant and classical Hodgkin lymphoma

Activation-induced cytidine deaminase (AID) is indispensable for class switch recombination and somatic hypermutation of immunoglobulin genes. Expression of AID has been detected in germinal centre centroblasts and in lymphomas derived from germinal centre cells. However, in situ studies of AID expression have until now been hampered by a lack of antibodies suitable for immunohistochemistry. To overcome this problem, an AID-specific monoclonal antibody suitable for immunohistochemical staining of formalin-fixed, paraffin wax-embedded tissue sections has been generated. This antibody was shown to detect AID expression in normal germinal centre B-cells as well as in non-Hodgkin lymphomas with a putative germinal centre origin. Using this antibody, a virtually exclusive cytoplasmic localization of AID in normal and neoplastic B-cells is shown. Employing a combination of immunohistochemistry and AID-specific in situ hybridization, it is demonstrated that AID is consistently expressed in the neoplastic cells of nodular lymphocyte-predominant Hodgkin lymphoma (HLnlp) but only infrequently in classical HL (cHL). This is in keeping with the notion that tumour cells of HLnlp represent transformed germinal centre B-cells showing evidence of somatic hypermutation. AID represents an additional marker useful in the differential diagnosis of HLnlp and cHL.

The Role of the Perifollicular Sinus in Determining the Complex Immunoarchitecture of Angioimmunoblastic T-cell Lymphoma

The growth of angioimmunoblastic T-cell lymphoma (AIL) in lymph node often produces complex patterns of neoplastic T cells and nonneoplastic B cells that complicate diagnosis. To understand better how these different patterns of B-cell expansion arise, we compared the microanatomic localization of B cells and T cells within the follicular, paracortical, and sinusoidal compartments in 30 patients with AIL (including 10 with multiple sequential samples) with that seen in 33 cases of other types of T-cell lymphoma. With early or partial nodal involvement in AIL, germinal center B-cell expansions were relatively undisturbed and often associated with a variably distended D2-40+ CD31+ perifollicular sinus that surrounded most of the follicular compartment. Identifiable tumor T cells resided mostly in the paracortex. In later stages of AIL with more complete nodal effacement, bcl-6+ follicular B-cell proliferations shifted to distorted FDC networks arrayed along patent trabecular sinuses and were more intermixed with tumor T cells. In both AIL and other T-cell lymphomas, the density and locations of follicular B cells as well as bcl-6-negative monocytoid B cells were largely related to the patency of adjacent sinuses, except in Epstein-Barr virus (EBV)+ and histiocyte-rich B-cell proliferations, which arose in paracortical locations. The prominence of the perifollicular sinus in early stages of AIL resembled that seen in reactive lymphadenitis during conditions of lymphatic engorgement and implicates cytokines within lymph fluid in maintaining both the normal and altered germinal center reactions. Patterns of sinus drainage largely explain the useful changes in B-cell distribution that occur in nodal T-cell lymphomas and represent an important tool in classification and diagnosis of these tumors.

Expression of the Human Germinal Center-Associated Lymphoma (HGAL) Protein, a New Marker of Germinal Center B Cell Derivation.

We identified the Human Germinal center-Associated Lymphoma (HGAL) gene in gene expression profiling studies of diffuse large B-cell lymphoma (DLBCL). The expression of HGAL was correlated with survival of patients with this diagnosis. The HGAL gene is the human homologue of M17, a mouse gene that had been previously identified to be expressed specifically in normal germinal center B cells. Here we generated a monoclonal antibody against the HGAL protein and show that HGAL is expressed in the cytoplasm of GC-lymphocytes and in lymphomas considered to be derived from the GC. Among 718 lymphomas tested by immunohistochemistry on tissue microarrays, staining for HGAL protein was present in 97% (103/107) of follicular lymphoma, 100% (40/40) of Burkitt lymphoma, 87% (7/8) of mediastinal large B-cell lymphoma and 70% (103/151) of DLBCL. Nodal and extranodal marginal zone lymphomas lacked HGAL expression. Lymphocyte predominant Hodgkin (70%, 12/17) and, surprisingly, a majority of cases of classical Hodgkin (73%, 78/107) lymphomas were found to be positive.Double-immunohistologic labeling experiments show that HGAL protein expression overlaps with the expression of BCL6 and CD10 proteins within GCs. Hierarchical clustering of comparative immunohistologic results in 151 cases of DLBCL demonstrates that the expression of HGAL is similar to two other GC-associated proteins BCL6 and CD10, but different from two markers associated with a non-GC phenotype, MUM1/IRF4 and BCL2. The restricted expression and GC-specificity of HGAL protein suggest that it may have an important role in the diagnosis of specific lymphomas, and, potentially in the identification of subtypes associated with different prognoses.

Aberrant Somatic Hypermutation Targets an Extensive Set of Genes in Diffuse Large B-Cell Lymphoma.

Genomic instability is a driving force in tumor development that can be achieved by a variety of mechanisms, such as defective chromosome segregation or inactivation of the DNA mismatch repair pathway. Although B-cell lymphomas are associated with chromosomal translocations deregulating oncogene expression, a mechanism for genome-wide instability during lymphomagenesis has long not been described. We have reported that the somatic hypermutation process (SHM), which normally targets the immunoglobulin variable region (IgV) and BCL6 genes in germinal center (GC) B-cells, functions aberrantly in >50% of diffuse large B-cell lymphoma (DLBCL), the most common type of B-cell non-Hodgkin lymphoma (Pasqualucci et al., Nature 412:341, 2001). As a consequence, multiple somatic mutations are introduced into the 5′ region of genes that do not represent physiologic SHM targets, including known proto-oncogenes such as PIM1, PAX5, RhoH/TTF and cMYC. To further define the extent of this phenomenon, termed aberrant somatic hypermutation (ASHM), and to identify additional hypermutated loci of possible pathogenetic significance in DLBCL, we screened 113 genes for the presence of mutations affecting their 5′ sequences (≥1.3 Kb from the transcription start site, the target region for SHM) in 10 DLBCL cell lines. Fifteen genes (13.3%) were found to harbor a significant number of mutations (p<0.05), with 70% of the cell lines being mutated in 7 or more genes; among these, six B-cell specific loci -BCL7A, CIITA, IRF4, LRMP, NCOA3 and SIAT1- carried 9–53 mutational events distributed in 20 to 70% of the cases, corresponding to an overall mutation frequency of 0.032–0.15% (frequency in the mutated cases: 0.07–0.25%). The same genes were found hypermutated in a panel of 20 primary DLBCL biopsies, which displayed an overall mutation load of 7 to 45 distinct events/gene (total N=125). Mutations were of somatic origin, independent of chromosomal translocations to the Ig loci and were restricted to the first 1.5–2 Kb from the promoter. In addition, analogous to previously identified SHM and ASHM targets, the mutations exhibited characteristic features, including a bias for transitions over transversions, preferential hotspot (RGYW/WRCY motifs) targeting, and higher frequencies at G:C pairs. However, in contrast to physiologic SHM targets such as IgV and BCL6, none of the 4 newly identified hypermutated genes that have been analyzed so far (BCL7A, CIITA, SIAT1, LRMP) displayed significant levels of mutations in purified normal GC B-cells as well as in other B-cell malignancies. This finding indicates that these genes represent aberrant hypermutation targets resulting from a tumor-associated malfunction, possibly a loss of target specificity of the physiologic SHM process. Considering previous results and the present survey, 17 (13%) out of 130 genes investigated have been found involved in ASHM, suggesting that this aberrant activity may involve an extensive set of target genes in DLBCL. Since the mutations affect both regulatory and coding sequences of the targeted genes, aberrant SHM may represent a major contributor to the pathogenesis of this disease and may explain in part its phenotypic and clinical heterogeneity.

The early transcription factor GATA‐2 is expressed in classical Hodgkin's lymphoma

Hodgkin/Reed-Sternberg (HRS) cells of classical Hodgkin's lymphoma (cHL) are thought to be derived from germinal centre B-cells in almost all cases. However, expression profiling has revealed that HRS cells do not show a germinal centre B-cell-like phenotype. Although the nature of this aberrant phenotype and the underlying molecular mechanisms remain largely unknown, it has been reported that the activity of NOTCH1 plays an important role in the growth and survival of HRS cells. In some leukaemic cell lines, the effect of Notch signalling is mediated by the early transcription factor GATA-2. This and the fact that HRS cells lack expression of PU.1, which can repress Gata-2, led to an investigation of GATA-2 expression in HRS cells. GATA-2 expression was found in all the cHL-derived cell lines studied, but not in a Burkitt lymphoma-derived cell line. In addition, 50% of biopsies from patients with cHL contained GATA-2-expressing HRS cells. In contrast, neither normal germinal centre B-cells nor malignant cells of nodular lymphocyte-predominant Hodgkin's lymphoma, Burkitt lymphoma or diffuse large B-cell lymphoma expressed GATA-2. Thus, GATA-2 expression was found specifically in HRS cells of cHL, suggesting that GATA-2 is important in establishing the abnormal B-cell phenotype of HRS cells.

Expression of the Cytoplasmic Tail of LMP1 in Mice Induces Hyperactivation of B Lymphocytes and Disordered Lymphoid Architecture

The oncogenic EBV protein LMP1 mimics a dysregulated CD40 receptor in vitro. To compare CD40 and LMP1-mediated events in vivo, transgenic mice were engineered to express mouse CD40 (mCD40tg) or a protein with extracellular mCD40 and cytoplasmic LMP1 (mCD40-LMP1tg). Transgenic and CD40 −/− mice were bred so that only the transgenic CD40 molecule is expressed in B cells, macrophages, and dendritic cells. mCD40-LMP1tg mice had normal lymphocyte subsets, and immunization elicited an antibody response featuring normal isotype switching, affinity maturation, and germinal center (GC) formation. However, unimmunized mCD40-LMP1tg mice had expanded immature and germinal center B cells, produced autoantibodies, exhibited marked splenomegaly and lymphadenopathy, and elevated serum IL-6. Thus, signaling through the LMP1 cytoplasmic tail results in amplified and abnormal mimicry of CD40 functions in vivo, indicating possible ways in which LMP1 contributes to the pathogenesis of EBV-associated human disease.

Silencing of the p18INK4c gene by promoter hypermethylation in Reed-Sternberg cells in Hodgkin lymphomas

Abstractp18INK4c is a cyclin-dependent kinase (CDK) inhibitor that interferes with the Rb-kinase activity of CDK6/CDK4. Disruption of p18INK4c in mice impairs B-cell terminal differentiation and confers increased susceptibility to tumor development; however, alterations of p18INK4c in human tumors have rarely been described. We used a tissue-microarray approach to analyze p18INK4c expression in 316 Hodgkin lymphomas (HLs). Nearly half of the HL cases showed absence of p18INK4c protein expression by Reed-Sternberg (RS) cells, in contrast with the regular expression of p18INK4c in normal germinal center cells. To investigate the cause of p18INK4c repression in RS cells, the methylation status of the p18INK4c promoter was analyzed by methylation-specific polymerase chain reaction (PCR) and bisulfite sequencing. Hypermethylation of the p18INK4c promoter was detected in 2 of 4 HL-derived cell lines, but in none of 7 non-Hodgkin lymphoma (NHL)–derived cell lines. We also detected p18INK4c hypermethylation, associated with absence of protein expression, in 5 of 26 HL tumors. The correlation of p18INK4c immunostaining with the follow-up of the patients showed shorter overall survival in negative cases, independent of the International Prognostic Score. These findings suggest that p18INK4c may function as a tumor suppressor gene in HL, and its inactivation may contribute to the cell cycle deregulation and defective terminal differentiation characteristic of the RS cells.

Development of Lymphoma in Autoimmune Lymphoproliferative Syndrome (ALPS) and its Relationship to Fas Gene Mutations

Autoimmune Lymphoproliferative Syndrome (ALPS) is generally the result of a mutation in genes associated with apoptosis, like Fas, Fas ligand, Casp 8 and Casp 10. As a result, the normal homeostasis of T- and B-lymphocytes is disturbed and a proliferation of polyclonal T lymphocytes occurs. This leads to hepatosplenomegaly and lymphadenopathy and in most patients also to autoimmune phenomena like anemia and thrombocytopenia. The proliferating T cells are TCRαβ and/or TCRγδ positive but lack both CD4 and CD8. Hence they are termed double negative (DN) T cells. In addition, there is an increase of CD5 positive B cells. Individuals with germline mutations in the Fas gene have a high risk to develop non Hodgkin lymphomas ( × 14) as well as Hodgkin lymphomas ( × 51), in particular NLP Hodgkin lymphoma. Somatic mutations of Fas are frequently acquired during the normal germinal center reaction. Non Hodgkin lymphomas carry somatic mutations of the Fas gene in 11% and of the Casp 10 gene in 14.5% of the patients. In Hodgkin lymphomas, Fas mutations can be demonstrated in Reed-Sternberg cells in 10 – 20% of the patients. These data implicate a role for Fas-mediated apoptosis in preventing lymphomas. Inherited defects in receptor-mediated lymphocyte apoptosis represent a risk factor for lymphomas and somatic mutations of these genes may also play a role in the development and/or progression of lymphomas.

Telomerase activity and proliferation index in aggressive mature B-cell lymphoma: comparison to germinal center phenotypic markers

Cell proliferation may be evaluated by various methods, including Ki-67 immunohistochemistry and measures of telomerase activity. Both methods would theoretically show comparable increases in a given case. To evaluate the relationship between these 2 markers of proliferation in aggressive mature B-cell lymphomas, 48 cases were studied. The study group included 5 cases of mantle cell lymphoma (MCL); 6 cases of Burkitt’s/Burkitt’s-like lymphoma (BL); 9 cases of follicular lymphoma, grade 3 (FLC); and 28 cases of diffuse large B-cell lymphoma (DLC). Telomerase activity was measured as total product generated (TPG) units, and TPG results for the aforementioned cases were compared to the TPG results for 10 cases of reactive follicular hyperplasia. An overlap in TPG scores between reactive cases and lymphoma cases was found. Significant differences in both log TPG ( P = 0.0443) and Ki-67 ( P = 0.0006) were seen in the different lymphoma types. A positive correlation between Ki-67 percentage and TPG score was identified in FLC ( r = 0.9281; P = 0.0003), but a poor correlation between these 2 indicators was seen in the other lymphoma types. Cluster analysis identified distinct patterns for MCL, FLC, and BL, but heterogeneous patterns for DLC. Because increases in both Ki-67 proliferation and telomerase activity are reported in normal germinal centers (GCs), these tests were also evaluated for usefulness as markers of a GC cell phenotype. Among the FLC and DLC cases, features of a GC phenotype significantly correlated with increased Ki-67 percentage ( P = 0.0152), but not with increased log TPG. An elevated log TPG correlated with CD10 expression, and elevated Ki-67 percentage correlated with both CD10 and BCL-6 expression. TPG level and Ki-67 percentage did not correlate with the presence of t(14;18) or BCL-2 protein expression. Although the proliferation patterns were fairly distinctive for MCL, FLC, and BL, these studies show that markers of cell proliferation do not by themselves,identify distinct subtypes of large cell lymphomas. With the exception of FLC, the tumors exhibited poor correlation between telomerase activity and Ki-67 proliferation index. These tests did show some correlation with expression of GC cell phenotypic markers, however.

Gene Expression Profiling of Diffuse Large B-Cell Lymphoma

Initial gene expression profiling studies of diffuse large B-cell lymphoma (DLBCL) revealed that this single diagnosis actually encompasses two distinct diseases that differ in the expression of hundreds of genes. One subtype, germinal center B-cell-like (GCB) DLBCL, strongly resembles normal germinal center B-cells and has a good prognosis following chemotherapy, whereas activated B-cell-like (ABC) DLBCL resembles mitogenically activated blood B cells and has a poor outcome. An expanded analysis of 274 DLBCL cases confirmed the existence of the GCB and ABC subgroups, but demonstrated that additional subgroups exist. Furthermore, two recurrent oncogenic events in DLBCL, t(14;18) and amplification of the c-rel locus on chromosome 2p, were only observed in GCB DLBCL, whereas constitutive activation of NF-κB was seen in ABC DLBCL, showing that the gene expression subgroups represent pathogenetically distinct diseases. Gene expression profiling has also been used to identify individual genes that predict overall survival in DLBCL, the majority coming from gene expression signatures that reflect the cell of origin, proliferation rate, and host immune response to the tumor. A multivariate model including 17 genes representing these biological features divided patients with DLBCL into quartiles with strikingly distinct 5-year survival rates, ranging from 73% to 15%. The use of gene expression profiling should eventually lead to an integration of molecular diagnosis and consequent selection of the most appropriate treatment.

Protein Gene Product 9.5 (PGP 9.5) Is Not a Specific Marker of Neural and Nerve Sheath Tumors: An Immunohistochemical Study of 95 Mesenchymal Neoplasms

In non-neoplastic tissues, the expression of protein gene product 9.5 (PGP 9.5), a member of the ubiquitin hydrolase family of proteins, is confined to neural and neuroendocrine cells. Although it has been claimed that PGP 9.5 is a specific marker of neural and/or nerve sheath differentiation in human tumors, careful review of the literature suggests that relatively few nonneural or nerve sheath tumors have been studied. Prompted by our recent observation of a PGP 9.5–positive malignant fibrous histiocytoma, we undertook a study of PGP 9.5 expression in a large group of well-characterized mesenchymal neoplasms. Sections from 95 mesenchymal tumors were retrieved from our archives and immunostained for PGP 9.5 using standard avidin-biotin complex technique and heat-induced epitope retrieval. Scoring was as follows: negative, 1+ (<10–25% of cells), 2+ (25–50% of cells), and 3+ (>50% of cells). Normal nerves and fibrous tissue were internal positive and negative controls, respectively. Positive immunostaining was seen in 80/95 (84%) of cases. Positive results by tumor subtype were as follows: (1) nerve sheath tumors: malignant peripheral nerve sheath tumor (7/10), neurofibromas (10/10), and perineuriomas (3/3); (2) (Myo) fibroblastic tumors: malignant fibrous histiocytoma (18/20), low-grade fibromyxoid sarcomas (8/9), fibromatoses (7/7), and desmoplastic fibroblastomas (2/2); (3) vascular tumors: angiosarcomas (4/4), hemangioendotheliomas (3/5), and hemangiomas (3/4); and (4) other non-nerve sheath tumors: pleomorphic liposarcoma (4/4), dermatofibrosarcoma protuberans (2/5), rhabdomyosarcomas (2/2), synovial sarcomas (8/8), melanomas (1/2). All positive cases were 2–3+ except 6 malignant peripheral nerve sheath tumor, 1 neurofibroma, 3 malignant fibrous histiocytoma, 2 low-grade fibromyxoid sarcoma, and 1 dermatofibrosarcoma protuberans. Positive staining was seen in normal smooth muscle and germinal centers in addition to nerves. We conclude that in this, the largest study to date of PGP 9.5 expression in mesenchymal neoplasms, we have found strong (2–3+) expression in the vast majority of nonneural or nerve sheath neoplasms studied. Although PGP 9.5 is a sensitive neural/nerve sheath marker, it is essentially totally nonspecific for diagnostic purposes. It is possible that our findings reflect cross-reactivity of the 13C4 clone with epitopes present on other ubiquitin hydrolases. Alternatively, PGP 9.5 expression may be aberrantly up-regulated in a variety of mesenchymal neoplasms.

Receptor revision of immunoglobulin heavy chain genes in human MALT lymphomas

Background/Aims: Rearrangement of immunoglobulin gene segments, leading to B cells with functional receptors, is thought to be largely restricted to developing immature B cells in bone marrow. However, accumulating evidence...

The expression of bcl-2 and bcl-6 protein in normal and malignant transitional epithelium.

The bcl-2 proto-oncogene plays a key role in cell longevity by preventing apoptosis. Bcl-2 is important in developing and maintaining the normal function of lymphoid and epithelial tissues. The bcl-6 protein is a 96 kDa nuclear protein selectively expressed in mature B cells within normal germinal centers as well as in their transformed counterparts in diffuse large B cell lymphoma. Recently, the bcl-6 protein has also been reported to be expressed in normal skin and epidermal neoplasms. In this study, 47 cases of transitional cell carcinomas (TCCs) were immunohistochemically studied for bcl-2 and bcl-6 protein expression. The results showed that bcl-2 was expressed only on basal layer cells, whereas bcl-6 expression was restricted to the superficial layers in the normal transitional epithelium. Von Brunn's nests showed strong immunostaining to bcl-2, but were negative to bcl-6. Among 47 TCCs, 15 (32.6%) and 29 (61.7%) cases were positive for bcl-2 and bcl-6, respectively. Compared with the normal transitional epithelium, the expression of bcl-2 was significantly decreased, whereas bcl-6 expression was significantly increased in TCCs. Additionally, the strong expression of bcl-6 had a positive correlation with the histopathologic grade of TCC. In conclusion, bcl-2 and bcl-6 proteins may play a role in the pathogenesis of TCCs, and bcl-6 expression reflects histopathologic grade.

BCL6 controls the expression of the B7-1/CD80 costimulatory receptor in germinal center B cells.

The BCL6 proto-oncogene encodes a transcriptional repressor required for the development of germinal centers (GCs) and implicated in the pathogenesis of GC-derived B cell lymphoma. Understanding the precise role of BCL6 in normal GC formation and in lymphomagenesis depends on the identification of genes that are direct targets of its transcriptional repression. Here we report that BCL6 directly controls the expression of B7–1/CD80, a costimulatory receptor involved in B–T cell interactions critical for the development of T cell–mediated antibody responses. Upon CD40 signaling, transcription of the CD80 gene is induced by the nuclear factor (NF)-κB transcription factor. Our results show that BCL6 prevents CD40-induced expression of CD80 by binding its promoter region in vivo and suppressing its transcriptional activation by NF-κB. Consistent with a physiologic role for BCL6 in suppressing CD80, the expression of these two genes is mutually exclusive in B cells, and BCL6-defective mice show increased expression of CD80 in B cells. The results suggest that BCL6 may directly control the ability of B cell to interact with T cells during normal GC development. In addition, these findings imply that T–B cell interactions may be disrupted in B cell lymphoma expressing deregulated BCL6 genes.

Two Newly Characterized Germinal Center B-Cell-Associated Genes, GCET1 and GCET2, Have Differential Expression in Normal and Neoplastic B Cells

A group of genes are highly expressed in normal germinal center (GC) B cells and GC B-cell-derived malignancies based on cDNA microarray analysis. Two new genes, GCET1 (germinal center B-cell expressed transcript 1) and GCET2, were cloned from selected expressed sequence tags (IMAGE clone 1334260 and 814622, respectively). GCET1 is located on chromosome 14q32 and has four splicing isoforms, of which the longest one is 1787 bp and encodes a 435-amino acid protein. GCET2 is located on 3q13.13, and the cloned fragment is 3270 bp, which encodes a protein of 178 amino acids. Blast search showed that GCET1 has a highly conserved serine proteinase inhibitor (SERPIN) domain and is located on a chromosomal locus containing seven other SERPIN family members. GCET2 is a likely homologue of the mouse gene M17, a GC-expressed transcript. Analysis of the GCET2 protein sequence indicated that it may be involved in signal transduction in the cytoplasm. Northern blot and real-time polymerase chain reaction analyses confirmed that GCET1 is highly restricted to normal GC B cells and GCB-cell-derived cell lines. Although GCET2 is also a useful marker for normal and neoplastic GC B cells, it has a wider range of expression including immature B and T cells. Real-time polymerase chain reaction assay showed that both GCET1 and GCET2 are preferentially expressed in follicular lymphoma and diffuse large B-cell lymphoma with GC B-cell differentiation, confirming previous microarray gene expression analysis, but neither one is entirely specific. Multiple markers are necessary to differentiate the GCB from the activated B-cell type of diffuse large B-cell lymphoma with a high degree of accuracy.

Epstein-Barr virus (EBV)-positive lymphoproliferations in post-transplant patients show immunoglobulin V gene mutation patterns suggesting interference of EBV with normal B cell differentiation processes.

In a model for persistent infection, Epstein-Barr virus (EBV) uses the germinal center (GC) reaction to establish persistence in memory B cells. To study whether EBV adopts to normal B cell differentiation processes also in EBV-associated lymphoproliferative diseases, we micromanipulated EBV(+) cells from biopsies of five patients with post-transplantation lymphoproliferative disease (PTLD) and one unusual Hodgkin lymphoma with many small EBV(+) cells, and analyzed rearranged V genes of single cells. In all cases clonal expansions of EBV(+) B cells were identified. The vast majority of these clones carried mutated V gene rearrangements and a fraction of clones showed ongoing hypermutation. Hence, PTLD likely derive from GC and/or post-GC B cells. In two clones hypermutation occurred in the absence of follicular dendritic and CD4(+) T cells, important interaction partners of normal GC B cells. Furthermore, in one case sustained somatic hypermutation occurred without expression of a functional antigen receptor. Hence, EBV(+) B cells in PTLD can retain or acquire features of GC B cells in an unphysiological setting and may continue to undergo somatic hypermutation uncoupled from normal selection processes, suggesting that EBV interferes with normal B cell differentiation and selection processes in PTLD.

Transcriptional Deregulation of Mutated BCL6 Alleles by Loss of Negative Autoregulation in Diffuse Large B Cell Lymphoma

The BCL6 proto-oncogene encodes a POZ/zinc finger transcriptional repressor expressed in germinal center (GC) B cells and required for GC formation. In ~35% of diffuse large B cell lymphomas (DLBCL) and 5–14% of follicular lymphomas (FL), the BCL6 locus is altered by chromosomal translocations that deregulate its expression by a mechanism known as promoter substitution. In addition, the BCL6 5′ noncoding sequences are targeted by multiple somatic mutations that cluster within ~1.5 kb from the transcription initiation site and are found in 30–40% of normal GC B cells (frequency: 0.14 × 10 –2 /bp) as well as in a fraction of all GC-derived B cell lymphomas. 1,2 The structural features, distribution, and specificity for GC B cells strongly suggest that BCL6 mutations are introduced by the same mechanism that targets immunoglobulin V genes in the GC. Nonetheless, the biological role of these mutations in normal and malignant B cell development remains unclear. The present study was aimed at investigating the effect of BCL6 mutations on gene transcriptional regulation. To this end, the BCL6 5′ noncoding region (~5 kb) from 24 tumor cases [6 DLBCL, 5 Burkitt lymphoma (BL), 4 FL, and 5 chronic lymphocytic leukemia (CLL), corresponding to 35 mutated alleles] or sorted GC B cells (N = 20 mutated alleles) was linked to the luciferase gene, and the activity of the corresponding constructs was evaluated in transient transfection/reporter gene assays in two BCL6 permissive lymphoma lines (Ly1 and MUTUI-BL59). We found that, while mutant alleles derived from normal GC cells, or from BL, FL, and CLL cases, displayed comparable activity to that of a wild-type construct, four (33%) of the 12 DLBCL-associated alleles were significantly overexpressed (4–18-fold) in both transfected lines. Among multiple mutations scattered throughout the BCL6 5′ noncoding sequences of these four alleles, single nucleotide substitutions, all located within the first noncoding exon of the gene, were found to be responsible for the observed overexpression. These mutations affect two adjacent BCL6 binding sites and were found to prevent BCL6 from binding to its own promoter, both in vitro in electrophoretic mobility shift