Metastasis-Associated Protein 1 Transgenic Mice: A New Model of Spontaneous B-Cell Lymphomas

The morbidity and mortality of cancer patients predominantly result from tumor invasion and metastasis of neoplastic cells from primary tumors to distant organ sites. Because distant metastasis represents an important factor affecting prognosis, identifying the physiologic role of metastasis-associated genes remains a key challenge in understanding and reducing metastasis. One such gene product is metastasis-associated protein 1 (MTA1; ref. 1). MTA1 is widely up-regulated in a variety of carcinomas such as breast, endometrial, colorectal, gastric, esophageal, pancreatic, ovarian, non–small cell lung, prostate, and hepatocellular carcinomas (2, 3). MTA1 has been identified as a component of chromatin remodeling complexes (4, 5). MTA1 was found to directly interact with histone deacetylases and repress the estrogen receptor α–driven transcription by recruiting histone deacetylase to the estrogen response element–containing target gene chromatin in breast cancer cells (5). MTA1 also interacts with a number of binding proteins that modulate its corepressor proteins (6–9). In addition, MTA1 acts as a coactivator of breast cancer–amplified sequence 3 (BCAS3), a gene amplified and overexpressed in breast cancers (10). Recently, MTA1 has been shown to be acetylated on lysine 626, and acetylation of MTA1 was essential for an effective transcriptional recruitment of RNA polymerase II complex to the BCAS3 enhancer sequence (10). These findings suggest that MTA1 functions may not be restricted to breast tumor cells. Although MTA1 has been shown to be up-regulated in various human carcinomas, its role in metastasis in a whole animal model remains unknown. Here, we show that MTA1 overexpression in MTA1-transgenic mice (MTA1-TG) was accompanied by an extremely high incidence of spontaneous B cell lymphomas in multiple organ systems, and that MTA1 was widely up-regulated in human diffuse large B-cell lymphomas (DLBCL) and in B-cell lymphoma cell lines.

Cell culture and antibodies. MCF-7-pcDNA, MCF-7-MTA1, KIS-1, SP53, IM9, DB, Jurkat, L1236, Karpass, and Daudi cells were maintained in RPMI 1640 supplemented with 10% FCS. Antibodies against MTA1, cyclin D1, and Bcl6 were purchased from Santa Cruz Biotech. Anti-Bcl2 was from DakoCytomation California, anti–B220-CD45 was from BD PharMingen, and anti-p27^Kip1 was from NeoMarkers.

Transgenic mice. The generation of MTA1-TG mice under the B6D2F1/J background has been previously described (11). Both female and male mice were monitored for lymphoma development, and for experimental purposes, female transgenic mice were used. Blood samples were collected in the mice from the retro-orbital sinus.

Human samples. Normal reactive lymph node and tonsil specimens were obtained from the Department of Hematopathology, M. D. Anderson Cancer Center, Houston, TX. The use of human residual tissues in this study was approved by the M.D. Anderson Cancer Center Institutional Human Research Committee. DLBCL and other lymphoma specimens were from Dr. Insun Kim, Department of Pathology, Korea University Medical College, Seoul, Korea (12).

Reverse transcription-PCR and Southern analysis. Reverse transcription-PCR (RT-PCR) and Southern analysis was done as previously described (5, 11). Genomic DNA was isolated from lymphoma samples, spleen, or thymus and Southern blotted using standard methods. Assays for Western blotting, immunofluorescence, B cell proliferation, spleen and tumor cell transplantation, ELISA, and immunohistochemical staining were done as described in the Supplementary Methods.

Statistical analysis. Groups were compared using two-tailed Student's t tests. P < 0.05 was considered significant.

To better understand the cellular functions of MTA1 in the context of an animal model, we recently generated transgenic mice (MTA1-TG) expressing MTA1 under the control of the mouse mammary tumor virus long-terminal repeat (11), which is also known to be expressed in lymphoid tissues (13). To examine the effects of MTA1 transgene on metastasis, we followed a cohort of animals for the potential presence of lymph node enlargements or tumors. Approximately 20% of the MTA1-TG animals autopsied at 4 to 18 months of age exhibited massive lymph node enlargement (1–3 cm in length) that was subsequently found to represent lymphomas (Fig. 1A and B; Supplementary Table S1). The incidence of pathologic lymph node enlargement increased with age, reaching values of 33% in MTA1-TG animals examined at 18 to 24 months and 55% at 30 months as compared with wild-type mice, which were tumor-free (Fig. 1B; Supplementary Table S1). In 66% to 80% of affected MTA1-TG animals, disease presented as extranodal lymphoma in the gastrointestinal tract (Supplementary Table S1). In 89% of affected transgenic animals, multiple disseminated lymph nodes (in the mesenteric, renal, mediastinal, and cervical areas) and spleen were involved, and lymphoid infiltrates were frequently observed in other tissues (Fig. 1A; Supplementary Fig. S1; Supplementary Table S1). Examination of 49 cases of disseminated disease revealed a spectrum of histologic subtypes, with DLBCL as the major subtype (40 out of 49 cases) along with other disseminated diseases such as Burkitt-like lymphoma (2 of 49) and follicular lymphoma (2 of 49; Supplementary Table S1). Other MTA1-TG mice showed localized diseases of different origins: unclassified sarcoma (1 of 49), endometrial stromal sarcoma (1 of 49), lung brochioalveolar carcinoma (2 of 49), and hepatocellular carcinoma (1 of 49; Supplementary Table S1). The lymphomas exhibited high immunohistochemical staining for transgenic T7-MTA1 (T7-Ab), B cell marker CD45 (B220 Ab), and the germinal center marker BCL6 (with the exception of follicular lymphoma; Fig. 1C). MTA1-TG mice showed widespread down-regulation of p27^Kip1, and up-regulation of Bcl2 and cyclin D1 as compared with spleen lysates from the wild-type mice (Fig. 1D). RT-PCR analysis also showed c-Myc up-regulation in lymphoma as compared with spleen from wild-type mice (Fig. 1D).

To evaluate whether the observed MTA1 deregulation–linked development of DLBCL was restricted to the MTA1-TG mouse model, we first carried out a data-mining analysis of public expression array data sets using Oncomine 2 (14, 15). MTA1 was found to be dysregulated in a total of 9 studies with a significance level of P < 0.005 from the 43 available studies queried on lymphomas as shown in Supplementary Fig. S2. Normalized MTA1 expression levels are shown as boxplots (Supplementary Fig. S2, top). To confirm these observations, we examined the expression of MTA1 in a well-characterized set of human DLBCL specimens (12). We found that 75 out of 76 specimens (98.6%) were strongly MTA1-positive. MTA1 expression was detected in the nuclei (32%) as well as in both the nucleus and cytoplasm (68%) of DLBCL cases (Fig. 2A, Supplementary Table S2). As controls, we used nine normal reactive lymph nodes and five tonsils which showed no evidence of MTA1 up-regulation (Fig. 2A). No significant difference in MTA1 expression was observed when the DLBCL cases were categorized into germinal center–derived DLBCL based on CD-10 or BCL6-positive and MUM1-negative immunostaining. All 23 of the germinal center cases showed MTA1 expression, and in 8 of these (34.8%), MTA1 was expressed in the nucleus, whereas in 15 cases (65.2%), expression was seen in both the nucleus and cytoplasm. Of the 53 non–germinal center cases, MTA1 expression was observed in 52 specimens (98.1%), with expression seen in the nucleus in 16 cases (30.8%) and in both compartments in 36 specimens (69.2%; Supplementary Table S2).

In reactive lymph nodes, MTA1 was found to be strongly expressed in a small minority of mantle zone B cells and rare cells in the germinal center, with minimal expression in the T cells of the paracortical regions, and moderate expression of MTA1 in mantle zone B cells. In the tonsil lymph nodes, MTA1 was weakly expressed in the B cells of the primary follicles as well as in the surrounding mantle B cells (Fig. 2A). These findings may be of physiologic significance as deregulation of MTA1 may be a common event in DLBCL in both humans as well as in the mouse model. MTA1-TG mice thus represent a unique model of spontaneous DLBCL associated with high tumor incidence.

To evaluate whether MTA1 expression is restricted to DLBCLs or is a widespread phenomenon in hematologic malignancies, MTA1 expression was examined in follicular lymphoma, mantle cell lymphoma, T-cell lymphoma, and NK/T tumors (Fig. 2B; Supplementary Table S3). MTA1 expression in follicular lymphoma was observed in five out of five cases (100%) and it was strongly cytoplasmic in the follicles (Fig. 2B). Among the tumors analyzed, MTA1 staining was observed in all cases of mantle cell lymphomas (11 of 11; 6 of 11 were nuclear, 5 of 11 were nuclear and cytoplasmic; Fig. 2B), NK/T tumors (12 of 12; 1 of 12 was cytoplasmic, 11 of 12 were nuclear and cytoplasmic; Fig. 2B), and in all the T-cell lymphomas (16 of 16; 2 of 16 were cytoplasmic, 14 of 16 were nuclear and cytoplasmic; Fig. 2B). These results suggest that misexpression of MTA1 is a more common phenomenon in lymphomas. In addition, we also analyzed MTA1 expression in various human lymphoma and leukemia cell lines by RT-PCR analysis. Four cell lines showed moderate to high levels of MTA1 expression, and one cell line (SP53) had very low levels of MTA1 (Fig. 2C). Western analysis also showed MTA1 expression in various cell lines analyzed (Fig. 2D).

To study the tumorigenic potential of spleen cells from MTA1-TG mice, splenocytes from age-matched wild-type and MTA1-TG mice were injected into two and five athymic nu/nu recipients, respectively. After 20 weeks, we found that the spleens of four out of five nu/nu mice from the MTA1-TG group were repopulated with MTA1-TG splenocytes and revealed lymphoid hyperplasia with a 40% increase in B220+ IgM cells (Fig. 3A), and none of the wild-type littermates developed lymphoid disease. Simultaneously, lymphoma cells from MTA1-TG mice were also injected into seven athymic nu/nu recipients. Interestingly, three of seven of these nu/nu mice developed lymphoma in the kidney and liver (Fig. 3B), clearly showing that these tumors were transplantable and retained their tumorigenic potential in mice.

The increase in B cell numbers in MTA1-TG mice upon progression to the symptomatic phase could result from either the outgrowth of neoplastic B cell clones or a gradual accumulation of polyclonal B cells. We next analyzed DNA from 21 MTA1-TG lymphoma tumors and DNA from the thymus or spleen of four age-matched wild-type mice for the presence of gene rearrangement of VDJ sequences from Igh genes. Sequencing analysis revealed the presence of somatic mutations in 9 of 10 (90%) of the cases (Supplementary Table S4), indicating that these tumors transited through the germinal center. Southern blotting analysis was carried out for a subset of the same samples to confirm the previous result as shown in Fig. 3C. To assess the major function of B lymphocytes, immunoglobulin levels were assayed by ELISA in the sera of MTA1-TG mice and wild-type mice. Baseline serum IgG_2a, IgG₃, IgM, and IgK levels were significantly lower in MTA1-TG animals compared with wild-type mice (Fig. 3D).

Upon autopsy, MTA1-TG mice presented enlargement of the spleen and exhibited increased organ weight (2.3-fold) as compared with wild-type mice. Histologic analysis of tissue sections from 12 apparently healthy 3-month-old MTA1-TG mice showed lymphoid hyperplasia and disorganization of white pulp regions in 6 of the 12 spleens analyzed. The fusion of the germinal centers and leukocyte infiltration into the surrounding red pulp were noted in these hyperplastic spleens (Supplementary Fig. S3). In addition, spleens from the MTA1-TG mice also exhibited increased bromodeoxyuridine and classical germinal center marker BCL6 immunostaining in germinal centers (Supplementary Fig. S4). BCL6 favors the sustained proliferation of B cells involved in germinal center formation (16), suggesting the hyperproliferation of splenic lymphocytes in the MTA1-TG mice. To investigate whether B lymphocytes in MTA1-TG mice were hyperproliferative, purified splenic B cells from wild-type and MTA1-TG mice were stimulated with lipopolysaccharide, and the differences in their proliferative rates were examined. B cells from MTA1-TG mice exhibited 3- to 4-fold increases in DNA synthesis as compared with B cells from age-matched wild-type mice after stimulation (Fig. 3E). The purity of the isolated B cells was confirmed by B220 staining (Fig. 3E , inset).

To further assess the effect of MTA1 on B cell development, we did multicolor immunofluorescence analyses of isolated bone marrow and spleen cells from age-matched wild-type and MTA1 mice. The percentage of CD19⁺ B cells within splenocytes increased from 24% in wild-type mice to 42% in MTA1-TG mice, suggesting that the accumulation of B cells occurred in the spleen of MTA1-TG mice (Fig. 4A). Remarkably, whereas CD19⁺CD43⁺IgM⁻ early B cell compartments or CD19⁺IgM⁺ immature B cells in bone marrow remained unaltered in the MTA1-TG mice (Fig. 4B), we found that these accumulated B cells from bone marrow or spleen in MTA1-TG mice showed no expression of surface IgD, indicating that the phenotype of transitional B cells were altered in MTA1-TG mice (Fig. 4B and C). However, splenic B cell compartments in MTA1-TG mice contained similar ratios of follicular (CD21^medCD23^hi) versus marginal zone (CD21^hiCD23^lo-neg) B cells compared with that of wild-type mice before and after antigen challenge (Fig. 4B and C; data not shown). These results indicate that recirculating IgD+ B cells were absent in MTA1-TG mice. The absence of surface IgD expression may result in their resistance to apoptosis and lead to abnormal cell proliferation. Put together, the results point to a scenario in which MTA1 and/or a subset of its targets could direct the development/maintenance of the malignant phenotype in lymphoma, and thus, identified a new role for MTA1 in lymphomagenesis. The noticed latency in tumor development in MTA1-TG mice suggests that MTA1 may cooperate with additional oncogenic events in tumorigenesis.

Grant support: Norman Brinkler Award for Research Excellence, and NIH grants CA098823 (R. Kumar).

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 T.J. McDonnell for the J_H region immunoglobulin heavy-chain gene probe, Riccardo Dalla-Favera for KIS-1 cells, and Biao Q. Zheng for providing antigen for immune challenge experiments.