Increased Expression of Germinal Center–Associated Nuclear Protein RNA-Primase Is Associated with Lymphomagenesis

Malignant lymphomas originate via various oncogenic mechanisms from diverse cell types and at various differentiation stages, with ongoing variations during tumor growth. B lymphoid cells, generated from hematopoietic precursor cells with immunoglobulin gene (Ig) rearrangements of the B cell antigen receptor (BCR), undergo antigen-stimulated clonal expansion and differentiation into antibody-producing cells in peripheral lymphoid organs. During B cell differentiation, various types of extracellular stimulation cause dynamic alterations of Ig genes, with frequent somatic hypermutation (SHM) of variable (V) regions, and class switch recombination of Ig heavy chain (IgH) loci in the germinal centers of secondary lymphoid organs (1–9). These changes may therefore cause gene alterations triggering malignant transformation of B cells.

In general, mutations of signal transduction molecules result in spontaneous malignant transformation, as originally observed with the myc and abl genes (10–13), and more recently with the TCL1 (14) and STAT5 genes (15), and null mutation of the bad gene (16). Recent studies have characterized the abnormalities of the functional molecules involved in germinal center events that target Ig genes and Ig gene transcripts during malignant lymphoma generation, such as AID (17), UNG (18), DNA polymerase ζ (19), DNA polymerase η (20), DNA polymerase ι (21), and Rad51 (22). Viral infection, especially with EBV, has been reported to be an etiologic factor in oncogenesis of Hodgkin lymphoma, in which molecular expression of EBV components (e.g., LMP1 and LMP2) is involved (23, 24).

Germinal center–associated nuclear protein (GANP) was found to be normally up-regulated in GC-B cells after immunization with T cell–dependent antigens in mice (25). Similar up-regulation was also observed in human tonsil sections by monoclonal antibodies (mAb) and by in situ RNA hybridization (26). Expression was not specific to B cells but was markedly up-regulated in B cells stimulated with anti-BCR and anti-CD40 antibodies in vitro, suggesting GANP involvement in clonal expansion and differentiation of antigen-driven B cells in germinal centers during the immune response. GANP contains two domains potentially involved in DNA replication: the NH₂-terminal RNA-primase domain (27) and the COOH-terminal MCM3-binding domain with acetylating activity for MCM3 (28). DNA replication initiated by the origin recognition complex requires an MCM complex composed of MCM2, MCM3, MCM4, MCM5, MCM6, and MCM7, which has helicase activity to relieve double-stranded DNA torsional stress. This is followed by bidirectional DNA extension as the leading and lagging strands interact with specific DNA polymerases (29). Extension of lagging strand DNA solely depends on the α-primase and DNA polymerase α that are associated with the MCM complex through a bridge with p70 and Cdc45 (29). The middle of the GANP protein sequence is homologous to yeast Sac3 and has been characterized recently as a molecule involved in transcription-coupled DNA recombination (30, 31). We have shown the critical role of GANP in B-cell maturation by conditional targeting of the ganp gene in B cells under CD19-Cre regulation (32). Loss of the ganp gene impaired B-cell maturation, with a decreased frequency of SHM of Ig-VH after immunization with (4-hydroxy-3-nitrophenyl)acetyl conjugated with chicken γ-globulin, and resulted in the lack of generation of high-affinity antigen-specific B cells in vivo. These data suggested that the principal function of GANP is in regulation of DNA replication and repair and the survival of B cells.

We investigated whether GANP is expressed in human B-cell lymphomas and is associated with lymphomagenesis of GC-B cells in various human disorders. Microscopic observations clearly showed a higher level of GANP expression in tumor cells. Staining with anti-GANP mAb clearly demarcated malignant cells from surrounding tissues and is potentially useful for pathologic diagnosis of lymphoma cells.

Based on these finding, we hypothesized that continued up-regulation of GANP expression in mice causes an effect similar to continual viral infections or other kinds of stresses during life as an oncogenic factor for human B lymphomagenesis. To explore this hypothesis, we examined survival and lymphomagenesis of ganp-transgenic (ganp-Tg) mice that express exogenous GANP under control of the human Ig enhancer and mouse Ig gene promoter. The ganp-Tg mice showed spontaneous lymphomagenesis, indicating that abnormal GANP expression is an oncogenic factor for malignant transformation of B cells in vivo.

Antibody. Anti-mouse GANP (42-23) and phosphorylated GANP (PG/103) at residue 502 (pSer⁵⁰² GANP) were used for detection of human GANP (25, 27). Labeled mAbs were purchased as phycoerythrin-labeled anti-B220 (RA3/6B2), phycoerythrin/anti-IgM (R6-60.2), FITC-labeled anti-GL-7, FITC/anti-CD34 (RAM34), FITC-anti-CD45 (leukocyte common antigen; 30-F11), biotin-anti-IgM (R6-60.2), biotin-anti-CD43 (S7), biotin-anti-Syndecan-1 (281-2), and allophycocyanin-labeled anti-c-kit (2B8; BD PharMingen, San Diego, CA). Phycoerythrin/anti-IgD (11-26) mAb was obtained from Southern Biotechnology Associates (Birmingham, AL). FITC-anti-Mac-1 (M1/70), biotin-anti-CD3 (145-2C11), and biotin-anti-IL-7R (A7R34) mAbs were purchased (eBioscience, San Diego, CA). Rat anti-mouse CD40 mAb (LB429) was described previously (27).

Clinical samples. Immunohistochemical studies were carried out with lymph node specimens from 25 lymphoma cases and bone marrow smears from 13 leukemia cases that had been diagnosed clinically and pathologically at the Department of Pathology, School of Medicine, Fukuoka University, Japan. The lymph nodes were fixed in buffered formalin and embedded in paraffin. After deparaffinization, sections on slides were retrieved with 0.5 mmol/L EDTA and then stained with GANP mAb (42-23) or pSer⁵⁰² GANP mAb (PG/103) in combination with alkaline phosphatase-conjugated goat anti-rat IgG antibody (Southern Biotechnology Associates). The stained clinical samples were compared with controls from nonspecific lymphadenitis.

Generation of Ganp-Tg mice. A mouse ganp cDNA fragment (nucleotides 442-5757) with a FLAG-tag was inserted into the EcoRI site of the pLG vector with the XbaI-BamHI DNA fragment containing the intronic Eμ enhancer (667 bp) of human IgH and the 343 bp mouse IgH promoter (33). The ganp gene DNA fragment was purified after XbaI-SalI digestion of the vector construct and microinjected into C57BL/6 fertilized eggs. Two transgenic founder lines were maintained in specific pathogen-free (SPF) conditions by the Center for the Animal Resources and Development, Kumamoto University. The genotype of transgenic ganp gene was analyzed by PCR with the following primers: ganp1-5′ (5′-TCCCGCCTTCCAGCTGTGAC-3′) and ganp1-3′ (5′-GTGCTGCTGTG TTATGTCCT-3′). The ganp gene transcripts were examined by reverse transcriptase-PCR (RT-PCR) using two sets of primers as flag-5′ (5′-GATTACAAGGATGACGACGATAAG-3′) and ganp0-3′ (5′-GCGCACAGACTTTCCCCTGA-3′) for transgenic gene and ganp-5194-5′ (5′-CCCGTGGGATGACATCATCA-3′) and ganp-5913-3′ (5′-CATGTCCACCATCTCCAGCA-3′) for the endogenous gene. The internal control was β-actin transcripts.

Magnetic resonance imaging. Wild-type and ganp-Tg mice with abdominal distension were anesthetized with thiamylal sodium (Kyorin Pharmaceuticals, Tokyo, Japan). Magnetic resonance imaging (MRI) examination of whole mice, except for the heads, was done in the Department of Diagnostic Radiology, Kumamoto University Hospital.

Histology. Organ specimens from various lymphoma cases were fixed in 4% paraformaldehyde in PBS at 4°C for 4 hours, embedded in paraffin, and cut into 6-μm-thick sections. The smears were stained with anti-CD45R/B220 mAb (RA3-6B2) and anti-IgM mAb (AM/3), followed with Alexa 488–conjugated goat anti-rat IgG antibody (Molecular Probes, Eugene, OR; ref. 34). The images were analyzed by fluorescence microscopy (BX51; Olympus, Tokyo, Japan).

Cell cycle analysis. Spleen cells were suspended in hypotonic propidium iodide (PI) solution as described (32) and analyzed by FACSCalibur (BD Biosciences, Franklin Lakes, NJ) with MODFIT LT software.

Proliferation assays. Single cell suspensions of mouse spleens were cultured in vitro for 48 hours (32). Cells were pulse-labeled with 0.2 μCi [³H]-thymidine (ICN, Costa Mesa, CA) per well for 16 hours before harvesting, and the radioactivity incorporated into DNA was measured. Affinity-purified goat anti-mouse μ-chain specific antibody [F(ab′)₂] (ICN), rat anti-mouse CD40 mAb (LB429), and lipopolysaccharide (LPS; Sigma, St. Louis, MO) were used for stimulation in vitro.

Flow cytometry. Single-cell suspensions from spleen of ganp-Tg mice with lymphomas were stained for two-color analysis with biotin-labeled mAb in combination with PerCP-Cy5.5-conjugated streptavidin (BD PharMingen) and with phycoerythrin-conjugated mAb, FITC-labeled mAb, and phycoerythrin-labeled mAb, or FITC-labeled mAb and allophycocyanin-labeled mAb and analyzed by FACSCalibur using the CellQuest software.

PCR detection of Ig gene rearrangements. Genomic DNAs were isolated from spleens mostly composed of lymphoma cells (>70% as detected by flow cytometry) by the standard procedure. Reaction mixtures in a total volume of 25 μL contained 0.5 μg genomic DNA, 250 μmol/L each of the four deoxynucleotide triphosphates, 10 pmol primers, and 0.5 unit LA-Taq polymerase (Takara Bio, Inc., Otsu, Japan). PCR reaction mixtures were preheated for 1 minute at 94°C, processed by 30 PCR cycles (denaturation for 30 seconds at 94°C, annealing for 30 seconds at 66°C, and extension for 2 minutes at 72°C), and finally incubated for 5 minutes at 72°C. Primers for detection of Ig rearrangements of mouse VH genes of various families were determined previously by Wei et al. (35): DH-5′ (5′-ACAAGCTTCAAAGCACAATGCCTGGCT-3′), JH4-3′ (5′-CTCTCAGCCGG CTCCCTCAGGG-3′), VH7183-5′ (5′-GCAGCTGGTGGAGTCTGG-3′), VHQ52-5′ (5′-TCCAGACTGAGCATCAGCAA-3′), and VHJ558-5′ (5′-CAGGTCCAACTGCAGCAG-3′). Primers for mouse Vκ genes were described previously (36): Vκcon (5′-GGCTGCAGSTTCAGTGGCAGTGGRTCWGGRAC-3′) and Jκ5 (5′-TGCCACGTCAACTGATAATGAGCCCTCTC-3′).

Reverse transcriptase-PCR. Total RNA was extracted from tumors using Trizol. VDJ-Cμ and CD79α (Igα, mb-1) transcripts were amplified by RT-PCR. Primers were based on previous data on VH genes from studies of Ig gene rearrangements, as follows: VDJ-Cμ, VH7183-5′, VHQ52-5′, VHJ558-5′, and Cμ-3′ (5′-GAAGACATTTGGGAAGGACTGACT-3′); Igα-5′ (5′-GCCAGGGGGTCTAGAAGC-3′); and Igα-3′ (5′-TCACTTGGCACCCAGTACAA-3′). PCR amplification was carried out as described above using Z-Taq DNA polymerase (Takara Bio).

DNA sequencing. PCR-amplified DNA products of VH and Vκ transcripts and VH-genomic DNA were analyzed by electrophoresis on 1.0% agarose gels, and cloned into the pGEM-T Easy vector (Promega, Madison, WI) for sequencing by an automated sequencer (Applied Biosystems, Foster City, CA) using LA-Taq polymerase. The primers for Vκ transcripts were described previously (37).

Increased GANP expression in lymphoid malignancy. We carried out immunohistochemical studies with anti-GANP mAb to investigate GANP expression in the cells from various human lymphoid and hematologic malignancies. The clinical samples were diagnosed using the WHO classification protocol (38), based on clinical findings, laboratory data, and pathology results. Several lymphoma cells showed high levels of GANP expression (Fig. 1), summarized in Table 1. GANP expression was markedly up-regulated in B-cell lymphomas [i.e., diffuse large B-cell lymphoma, mantle cell lymphoma (MCL), and Burkitt lymphoma] and in other leukemic cells (i.e., acute myeloid leukemia and myelodysplastic syndrome). GANP expression was increased in other lymphomas of B-cell origin (i.e., Hodgkin lymphoma, follicular lymphoma, and multiple myeloma) and of T-cell origin (i.e., adult T-cell leukemia/lymphoma, Lennert lymphoma, and chronic myeloid leukemia). However, normal cells in control sections and bone marrow specimens from healthy volunteers did not show up-regulation of GANP expression (Fig. 1). Phosphorylation at GANP Ser⁵⁰² was induced in vitro by anti-CD40 stimulation. The phosphorylation could be catalyzed by Cdk2 in vitro and GANP primase activity was dependent on its phosphorylation at Ser⁵⁰² (27). We also examined whether the primase-active form of GANP was increased in lymphomas using anti-pSer⁵⁰² GANP mAb that binds to phosphorylated GANP at Ser⁵⁰². The pSer⁵⁰² GANP was found to be up-regulated to different levels in different types of lymphomas; for example, Hodgkin lymphoma and MCL cells showed higher GANP expression than DLCBL cells (Table 1). Interestingly, phosphorylation at GANP Ser⁵⁰² was up-regulated in particular malignant cell types, such as the Reed-Sternberg cells of Hodgkin lymphoma (Fig. 1). These results indicate that GANP and GANP pSer⁵⁰² may be useful markers for malignant cells.

Spontaneous lymphomagenesis in ganp-Tg mice. We used a Tg mouse model to study whether B lineage malignancies are generated spontaneously during aging with enhanced expression of the ganp gene. Most ganp-Tg mouse lines showed a similar increase of ganp transcripts; therefore, we selected two breeding lines. RT-PCR analysis clearly showed the expression of the transgenic ganp gene in primary and secondary lymphoid organs containing B lineage cells, but it also appeared in thymus cells. The transgenic ganp gene appeared selectively in lymphoid organs because the brain and liver did not express at detectable levels (Fig. 2A). However, there was no alteration in the expression of endogenous ganp gene transcripts shown in a previous report (26). The mice were born and grew normally until 50 weeks after birth. The lymphoid lineage cells differentiated normally until 16 weeks (data not shown). We found that about 30% of ganp-Tg mice suffered from lymphomas spontaneously under SPF conditions. Examination of a matched cohort of ganp-Tg mice and wild-type littermates of C57BL/6 background showed that the average life span of the ganp-Tg mice was similar to that of the wild-type littermates, in spite of tumorigenesis of the mutant mouse group (data not shown). However, we observed apparently shorter life spans of the tumor-bearing group of ganp-Tg mice (Fig. 2B). Tumor-bearing ganp-Tg mice showed an abrupt decrease in survival after 91 weeks with a peculiar histogram, which is in marked contrast to those of nontumor-bearing ganp-Tg mice or wild-type littermates. Interestingly, the nontumor-bearing group of ganp-Tg mice showed a longer life span of 96.0 ± 20.7 weeks in comparison with 86.1 ± 16.6 weeks for wild-type mice. Whereas we do not know the reason the nontumor-bearing ganp-Tg group showed a longer life span, this might have masked the difference in the average life span between ganp-Tg and wild-type mice.

The ganp-Tg mice showed abdominal dilation, emaciation, awoken hairs, and frequent generation of spontaneous tumors (18 of 61 mice, 29.5%) after the age of 60 weeks, but these characteristics were not seen in wild-type littermates (0 of 48 mice; data not shown). Eight of the 10 tumors in ganp-Tg mice were predominantly of the non-B/non-T cell surface phenotype. Most of the tumor-bearing mice showed a similar course of lymphomagenesis, as investigated in detail in lymphoma case 2 (Tumor-2; Fig. 2C; Table 2). MRI images showed massive invasion of tumor cells throughout the abdominal cavity with enlarged liver and spleen. The spleen was enlarged 1.5-fold; the liver showed a coarse surface with irregular appearance and a 1.6-fold enlargement compared with wild-type littermates. In the spleen, liver, lung, and lymph nodes from a ganp-Tg mouse with a lymphoma (Tumor-3), diffuse proliferation of atypical lymphocytes with irregular nuclei and cell shapes, and destruction of tissue-specific architectures were observed (Fig. 2D). In particular, the liver showed a collapsed sinusoid hepatic structure with massive and diffuse infiltration of lymphoma cells. Most cells were medium-sized lymphocytes possessing atypical nuclei with occasional mitosis (arrows). A few cells showed differentiation toward plasma cells. Most cells in the lymph nodes were B220-positive.

Flow cytometric analysis with PI staining showed increased S and G₂-M phases in tumor cells (17.5 ± 9.7%) compared with normal spleen cells (2.6 ± 5.1%; Fig. 3A). To examine whether lymphoma cells can proliferate in tissue culture conditions, the proliferation assay was compared with normal spleen B cells stimulated in vitro. Tumor-2 cells proliferated slowly in the absence of B cell stimulants and were less responsive to stimulation with anti-BCR or anti-CD40 antibody (Fig. 3B). LPS stimulation induced an acceleration of tumor cell proliferation in vitro. These results showed that lymphomas generated in ganp-Tg mice were less responsive than normal spleen cells.

Characterization of lymphomas generated in ganp-Tg mice. The surface phenotype of Tumor-5 is shown in Fig. 4A. The majority of the spleen cells of the ganp-Tg (Tumor-5) mouse did not show any surface markers detectable with the antibodies used. Nearly 90% of Tumor-5 cells expressed the B220⁻CD3⁻ phenotype, which is in marked contrast to the normal spleen with B220⁺ (68.0%) and CD3⁺ (28.6%) cells. The tumor cells did not express CD5, which indicated that they were not directly associated with the B-1 cell population. Most of the Tumor-5 cells did not express Mac-1, but a small proportion of Mac-1-positive cells were detected (7.2 ± 2.8% versus 0.9 ± 1.3% in the control spleen). Tumor-5 cells were surface IgM (sIgM)⁻ (B220⁻IgM⁻ = 92.9%) and sIgD⁻. The majority of the tumor cells did not express mature B cell differentiation markers such as CD40, GL-7, and Syndecan-1. We also examined the expression of the immature B cell markers, CD43, CD45, c-kit, CD34, and IL-7R, but these markers were absent on the surface of the tumor cells. A proportion of the cells expressed CD43 (13.8%), c-kit (7.0%), but they did not express CD34 or IL-7R. Tumor cells in another ganp-Tg mouse (Tumor-2) expressed low levels of CD45R/B220, but cytoplasmic IgM was not detected in these cells (Fig. 4B).

To further characterize the origin of lymphomas, we studied the expression of various B cell–specific gene transcripts by RT-PCR (Fig. 4C). The mRNA of CD79α (Igα/mb-1) and Ig VHDHJH-Cμ was expressed in lymphomas.

Ig gene rearrangements and V region sequences of lymphomas generated in ganp-Tg mice. A genomic DNA configuration study by PCR confirmed that the lymphomas showed rearrangements of the IgH and Igκ alleles as VHDHJH and VκJκ forms in the tumor (Fig. 5A), and some were further confirmed by Southern blot analysis with a JH3/JH4 probe (Supplementary Fig. S1). The rearranged bands showed monoclonal or oligoclonal patterns, suggesting that the lymphoma cells proliferated from a single or pauciclonal transformation of B cells, regardless of the non-B/non-T cell surface phenotype.

The V regions of lymphoma Ig genes were studied employing the primers that were used for the initial PCR detection of Ig rearrangements, and the individual sequence was compared with previously determined VH family sequences in the data base. The sequence analysis showed the V region involvement of the tumors in detail (Fig. 5B). Tumor 1 was composed of three rearranged IgH genes: AF246423 (European Molecular Biology Laboratory accession number), AF296428, and AF296426 (data not shown). Tumor-2, Tumor-3, and Tumor-4 each had a single V region as the rearranged IgH gene, although several mutations were observed in each IgH gene. Tumor-2 had a deletion of the G at nucleotide 22 in a V region sequence, generating a stop codon at protein codon 36. However, all other clones with multiple mutations did not cause a stop codon, indicating most V regions in the rearranged IgH genes are productive in their configurations (Fig. 5A and B). We also analyzed the rearranged Ig Vκ genes of Tumor-1 and Tumor-3, which showed similar results. Although it remains to be determined whether the V region mutations are associated with somatic hypermutation of the V region during the tumor growth, these results showed that lymphomas were derived from a single or oligoclones, as had been reported in human B lymphomas of GC-B cell origin (39).

We proposed that up-regulation of GANP could be a causative factor for malignant transformation. Alternatively, increased GANP expression may be a secondary effect of the accelerated proliferation of malignant cells and not necessarily linked to oncogenesis. To investigate this question, we used a ganp-Tg mouse model. The ganp-Tg mice showed spontaneous tumorigenesis in old age (80-91 weeks) with an incidence of 29.5%, which was quite high in comparison with that of wild-type littermates (0%). The tumor cells showed atypical lymphoid histologic characteristics and no apparent B lineage or T lineage cell surface markers. They also showed evidence of monoclonal or oligoclonal expansion of B lineage cells that had undergone Ig gene rearrangements. The ganp-Tg mice could provide useful information regarding lymphomagenesis of GC-B cells during severe or long-lasting infections with various kinds of microorganisms.

A simple comparison of the average life span did not reveal any difference in the ganp-Tg mice, but life span was clearly affected when the mice generated lymphomas in old age. Surprisingly, we observed the prolongation of the average life span of mutant mice. It would be interesting to know how the life span of ganp-Tg mice was affected. Continuous and high level expression of ganp transcripts might not always be a life-threatening factor but could be beneficial for survival of individuals infected with various microorganisms during long-term life. These results might suggest that ganp-Tg mice may survive but could be affected by various oncogenic factors during the hyperreactive state of B lineage cells in germinal center regions.

In Fig. 4C, Tumor-1 did not show an increased expression of ganp gene, whereas most of the tumor lines showed higher levels of ganp transcripts. Tumor-1 showed distinct oligoclonal configurations of Ig gene rearrangements (Fig. 5A; Supplementary Fig. S1), which strongly indicated that Tumor-1 cells were of malignant transformation but not of the mere hyperproliferation state. The possibility that Tumor-1 lymphoma is just a background lymphoma might be remained, but we have not observed malignancies of any other lineages and tissues throughout our observation of control mice. Another possibility is that the effect of the ganp transcripts can be nonautonomous, perhaps by stimulating the secretion of cytokines or affecting the costimulatory cells in the peripheral lymphoid tissues. Transgenic ganp gene was mainly detected in lymphoid tissues (Fig. 2A), suggesting that Tumor-1 cells were generated from the B lineage cells that most likely had ganp gene overexpression. The findings from Tumor-1 might also suggest that ganp overexpression is involved in tumorigenesis but is not always necessary for tumor growth, for which another genetic alteration could be a causative factor. We observed karyotypic abnormalities of Tumor-2 cells with X chromosome trisomy and chromosome translocations (data not shown). This might explain for the late onset of lymphomas, regardless of high levels of expression of the ganp gene in the mutant mice.

The phenotype of mature B cell lymphomas was first reported as sIg⁺B220⁺CD19⁺, and the molecular characteristics of mature B cell lymphomas, regardless of type, were presented as if both IgH and IgL alleles of Ig genes had undergone rearrangements. The ganp-Tg mice developed lymphomas of B cell origin, based on their Ig gene rearrangement, which could not be directly comparable in a mouse lymphoma model of the Bethesda proposed classification (40).

In 1954, Dunn proposed a classification of neoplastic and related conditions of the reticular tissue in mice (41). Type A included neoplasms composed primarily of the cells that are designated a monocyte, a histiocyte, or a phagocyte with morphologic variations. Type B neoplasms were pleomorphic and had some points of resemblance to human Hodgkin lymphoma, and were called the “Hodgkin lymphoma–like lesion of the mouse” or “reticulum cell sarcoma type B (RCSB)”. Type C was not thought neoplastic but had a proliferation of reticulum cells. In earlier reports (42), laboratory mice, presumably maintained under conventional conditions, bore spontaneous tumors with reticuloses as a borderline between inflammatory and neoplastic responses. Histologic examination revealed RCSB tumors in a C57BL/6 strain with an incidence of 7% to 14% at the age of 96 to 99 weeks (43). The etiology was considered viral or inflammatory factors. The tumors displayed pleomorphic and structureless tissues with the large, pale reticulum cells varying in size and shape, the lymphocytic cells, and a variable number of plasma cells, frequently containing Russell bodies. Many reports indicate considerable difficulties in the transplantation of the RCSB.

Tumors generated in ganp-Tg mice displayed several similarities to the RCSB, including the age of onset, the rate of tumor growth, and the difficulty in transplantation. Histologic findings were similar to pleomorphic tissues, in which the basophilic mononuclear cells, resembling normal lymphocytes, were intermingled with the reticulum cells and plasma cells. Under SPF conditions until 115 weeks, we have not observed tumorigenesis in control C57BL/6 mice (below 2%) but have observed a high incidence in ganp-Tg mice (29.5%). This may suggest that inflammatory factors are involved in the cause, or at least in the support of tumor proliferation in the mice, even under SPF conditions, and that the GANP overexpression plays some role in lymphomagenesis. Our attempt to transplant, whereas the trial was limited, failed to produce growth of tumors in the wild-type mice (data not shown).

As a germinal center–derived lymphoma, human Hodgkin lymphoma is classified as classic (c) and lymphocyte-predominant (lp) forms. Classic Reed-Sternberg cells, probably derived from preapoptotic GC-B cells, have hCD30⁺hCD15⁺hCD45⁻hCD20⁻ and an EBV-infected phenotype (44). The lp-type Hodgkin lymphoma cells have hCD30⁻hCD15⁻hCD45⁺hCD20⁺ and an EBV-noninfected phenotype, often showing ongoing SHM of Ig V region genes during tumor growth (45). The lymphomas of ganp-Tg mice showed lymphocyte-form tumors, but expression of CD30 or CD20 was undetectable (data not shown).

EBV is linked to the development of several malignancies, primarily of B-cell and epithelial cell origin, including post-transplant lymphomas in immunocompromised hosts, Hodgkin lymphoma, nasopharyngeal carcinomas, and AIDS-related lymphomas (24). Human B cells immortalized in vitro express at least nine proteins, including six EBV-encoded nuclear antigens (EBNA1-6) and three latent proteins (LMP1, LMP2A, and LMP2B). LMP1-Tg mice were generated with the gene under the control of an IgH promoter and enhancer. The transgenic mice had an increased incidence of lymphomas as they aged (42% over 18 months; ref. 46), which showed that constitutive LMP1 expression is a critical oncogenic factor in the development of EBV-associated malignancies. Gene rearrangement of IgH indicated that all lymphomas were derived from monoclonal or oligoclonal origins. LMP1 has been shown to interact with tumor necrosis factor receptor-associated factors (47). Triggering of CD40 and expression of LMP1 activate NF-κB, which induces the expression of many of the same cellular genes (48). In ganp-Tg mice, lymphomas developed with a similar incidence and onset as in LMP1-Tg mice, and lymphomas of both Tg mice were similarly monoclonal or oligoclonal. The ganp-Tg mouse could be useful for studies of the oncogenetic factors that affect the molecules involved in proliferation and differentiation of B cells during the immune response.

Grant support: Special coordination funds for promoting science and technology from the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Y. Kumamoto for technical assistance.