Overexpression of C-Maf Contributes to T-Cell Lymphoma in Both Mice and Human

Research Article

Overexpression of c-Maf Contributes to T-Cell Lymphoma in Both Mice and Human

Naoki Morito,1 Keigyou Yoh,2 Yuki Fujioka,1 Takako Nakano,1 Homare Shimohata,1 Yuko Hashimoto,1 Akiko Yamada,1 Atsuko Maeda,1 Fumihiko Matsuno,5 Hiroyuki Hata,5 Atsushi Suzuki,6 Shigehiko Imagawa,3 Hiroaki Mitsuya,5 Hiroyasu Esumi,6 Akio Koyama,2 Masayuki Yamamoto,4 Naoyoshi Mori,7 and Satoru Takahashi1

1Department of Anatomy and Embryology, Biomolecular and Integrated Medical Sciences, 2Pathophysiology of Renal Diseases, Medical Sciences for Control of Pathological Processes, 3Clinical and Experimental Hematology, Major of Advanced Biomedical Applications, 4Graduate School of Comprehensive Human Sciences, Center for Tsukuba Advanced Research Alliance, University of Tsukuba, Tsukuba, Japan; 5Department of Internal Medicine II, Kumamoto University School of Medicine, Kumamoto, Japan; 6Cancer Physiology Project, National Cancer Center Research Institute East, Chiba, Japan; and 7Department of Pathology of Biological Response, Nagoya University Graduate School of Medicine, Nagoya, Japan

Abstract malignancies. c-Maf is expressed at high levels in multiple c-Maf translocation or overexpression has been observed in myeloma cells carrying the translocation t(14;16) (q32;q23), which human multiple myeloma. Although c-maf might function as results in the fusion of the immunoglobulin heavy chain to the an oncogene in multiple myeloma, a role for this gene in other c-maf gene locus (10–12). Moreover, using gene expression profiling cancers has not been shown. In this study, we have found that by DNA microarray analysis, Hurt et al. (13) detected over- mice transgenic for c-Maf whose expression was direct to the expression of c-Maf in 50% of multiple myeloma cell lines that they T-cell compartment developed T-cell lymphoma. Moreover, have analyzed. However, the t(14;16) (q32;q23) primary transloca- we showed that cyclin D2, integrin B , and ARK5 were up- tion is present in only 5% to 10% cases of multiple myeloma. Thus, 7 the molecular pathogenesis of multiple myeloma characterized by regulated in c-Maf transgenic lymphoma cells. Furthermore, f 60% of human T-cell lymphomas (11 of 18 cases), classified as c-Maf overexpression identifies 50% is not fully understood (14). Although c-maf might function as an oncogene in human multiple angioimmunoblastic T-cell lymphoma, were found to express c-Maf. These results suggest that c-Maf might cause a type of myeloma, there is no evidence to date for the involvement of this T-cell lymphoma in both mice and humans and that ARK5, in gene in the development of other malignancies. In this study, we have generated c-Maf transgenic mice that overexpress c-Maf in addition to cyclin D2 and integrin B7, might be downstream target genes of c-Maf leading to malignant transformation. the T-cell compartment using the human CD2 promoter and locus control region. c-maf transgenic mice developed T-cell lymphoma, (Cancer Res 2006; 66(2): 812-9) providing the first direct evidence that c-maf can function as an oncogene in T cells in vivo. In addition, we analyzed c-Maf Introduction expression in cases of human T-cell lymphoma, showing c-Maf The maf proto-oncogene was originally identified within the overexpression in angioimmunoblastic T-cell lymphoma (AITL) genome of the avian musculoaponeurotic fibrosarcoma virus, AS42 at high frequency. These results indicate that c-Maf can cause a (1). The product of the maf gene and other members of the Maf type of T-cell lymphoma with characteristics similar to multiple family share a conserved basic region and amphipathic helix (bZip) myeloma and suggest that the c-Maf transgenic mouse may be a motif that mediates dimmer formation and DNA binding to the good model to study how the oncogene c-maf contributes to the Maf recognition element (MARE; ref. 2). Large Maf proteins, such pathogenesis of lymphoid malignancies. as c-Maf, MafB, MafA/L-Maf/SMaf, and NRL, contain an acidic domain that mediates transcriptional activation and plays a key Materials and Methods role in cellular differentiation (1–5). Mice. A 1.5-kb full-length cDNA encoding the murine c-Maf protein was c-Maf encodes a Th2-specific transcription factor that activates inserted into a VA CD2 transgene cassette containing the upstream gene the expression of interleukin (IL)-4 and IL-10 in T cells (6). c-Maf regulatory region and locus control region of the human CD2 gene. The VA has also been known to be involved in the regulation of lens vector has been reported to direct expression of an inserted cDNA in fiber cell differentiation (7–9). Recently, it has been determined all T lymphocytes of transgenic mice, with expression being linearly that c-maf functions as an oncogene in multiple myeloma (10–14). proportional to the transgene copy number (15). This c-Maf construct was Multiple myeloma is an incurable neoplasm of terminally injected into BDF1 fertilized eggs to generate transgenic mice. The differentiated B cells characterized by monoclonal expansion of c-Maf transgenic allele was genotyped by PCR using a pair of primers: V V V malignant plasma cells, accounting for f20% of all hematologic sense 5-TGCAGCAGAGACACGTCCTG-3 and antisense 5-TGAGTTCGAA- GATGCAGAGG-3V. Mice were maintained in specific pathogen-free conditions in a Laboratory Animal Resource Center. All experiments were done according to the Guide for the Care and Use of Laboratory Animals at the University of Tsukuba. Note: N. Morito and K. Yoh contributed equally to this work. Requests for reprints: Satoru Takahashi, Department of Anatomy and Southern hybridization analysis. High molecular weight DNA was Embryology, Biomolecular and Integrated Medical Sciences, Graduate School of prepared from the tail of each mouse, and 10 Ag of this DNA were Comprehensive Human Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba digested with NcoI and then subjected to electrophoresis through 0.8% 305-8575, Japan. Phone: 81-298-53-7516; Fax: 81-298-53-6965; E-mail: satoruta@ agarose gels. After electrophoresis, the DNA was transferred to a nylon md.tsukuba.ac.jp. I2006 American Association for Cancer Research. membrane (Zeta-probe; Bio-Rad, Richmond, CA). To generate DNA probes doi:10.1158/0008-5472.CAN-05-2154 for Southern hybridization, we cloned the open reading frame region of

Cancer Res 2006; 66: (2). January 15, 2006 812 www.aacrjournals.org

Figure 1. Generation of c-Maf-overexpressing mice. A, structures of the mouse c-maf gene locus and the transgenic construct. The c-Maf cDNA was inserted into a vector (VA vector) containing the human CD2 transgene cassette. Southern blotting probe site, restriction sites, and predicted sizes of the endogenous gene and the transgene (NcoI). Arrowheads, positions of genotyping PCR primers (sense and antisense). LCR, locus control region. B, we obtained five transgenic mouse lines (Tg 50, 17, 78, 235, and 312) screened by PCR using tail DNA as templates. C, Southern blot analysis of the endogenous and transgenic c-maf genes in transgenic mice. The 0.3-kb fragment shown in (A) was used as the probe. The 2.8-kb endogenous and 1.1-kb transgenic genes are shown for transgenic lines 50 (Tg 50), 17 (Tg 17), and 78 (Tg 78) mice. The transgene copy numbers for Tg 50 were 6 copies, Tg 17 and 78 were 2 copies, respectively. D, analysis of c-Maf mRNA in thymocytes. c-Maf mRNA was detected in all samples tested. The amount of c-Maf mRNA in samples from Tg 50 mice was several-fold greater than that in samples from wild-type mice through densitometric analysis (Image J). E, immunofluorescence profile analysis of thymocyte subpopulations from c-Maf transgenic mice at 5 weeks of age. Analysis of the total thymocyte population for CD4 and CD8 expression revealed that CD4/CD8 double-negative thymocytes were increased and the thymus contained very few double-positive or single-positive mature thymocytes. Analysis of CD44 and CD25 expression showed that the numbers of CD44+CD25-, CD44+CD25+, and CD44ÀCD25+ thymocytes were increased and that numbers of CD44ÀCD25À thymocytes were decreased in the c-Maf transgenic mice. the c-maf gene. Transgene copy number was determined using Image J intensity and propidium iodide exclusion. The following phycoerythrin (PE), (NIH, Bethesda, MD). To investigate T-cell receptor (TCR) h-chain gene FITC, and peridinin chlorophyll protein (PerCP) labeled monoclonal rearrangement, tail and other DNAs were also examined by Southern antibodies were used: anti-CD4-PE, anti-CD8-FITC, anti-CD25-PerCP, and hybridization as described previously (16). The TCR probes were gel anti-CD44-FITC (BD PharMingen, San Diego, CA). purified and then labeled for hybridization. Hind III-digested DNAs were Histopathologic analysis. Each mouse was bled while under ether hybridized with the Jh2 fragment. anesthesia. At autopsy, organs were fixed with 10% formalin in 0.01 mol/L Flow cytometry analysis. Single-cell suspensions were prepared from phosphate buffer (pH 7.2) and embedded in paraffin. Sections were stained the thymi and splenic tumors of each mouse, which were then blocked with with H&E stain for histopathologic examination by light microscopy. The anti-FcR antibody (2.4G2) for 10 minutes on ice to inhibit the interaction of immunohistochemistry and analysis for CD3, B220, CD4, and CD8 antigens the staining reagents with the cell surface. Multicolor flow cytometric were done using the streptavidin and biotin technique. Anti-CD3 antibody analysis was done using LSR and CellQuest software (Becton Dickinson, was purchased from Serotec (Oxford, United Kingdom). Anti-B220, CD4, Franklin Lakes, NJ) on viable cells as determined by forward light scatter and CD8 antibodies were obtained from BD PharMingen. www.aacrjournals.org 813 Cancer Res 2006; 66: (2). January 15, 2006

Quantitative reverse transcription-PCR assay. Total RNA (1 Ag) was Statistical analysis. Results were expressed as mean F SE. Multiple data reverse transcribed into cDNA with SuperScript II reverse transcriptase comparisons were done using the one-way ANOVA routinely with the (Invitrogen, Carlsbad, CA). TaqMan PCR primers and TaqMan probes used to Bonferroni correction. Significant differences between the groups of mice quantify c-Maf and ARK5 cDNAs were as follows: c-Maf sense 5V-AAGGAG- were analyzed using the Wilcoxon test for paired samples, and Ps < 0.05 AAATACGAGAAGCTGGTGA-3V, antisense 5V-TGGGATCGCGTGTCACACTC- were considered statistically significant. Comparisons of incident rates were ACATG-3V, and probe 5V-CGACAACCCTTCCTCTCCCGAATTT-3Vand ARK5 done by the Kaplan-Meier method. sense 5V-GAGCCCACTCTATGCGTC-3V, antisense 5V-ATGTCCTCGATGGTGG- CT-3V, and probe 5V-CACTCTCATTTATGGAACGATGCCC-3V. Mixture of pri- Results mers and TaqMan probes for cyclin D2, integrin h7, and IL-4 were obtained from Assays-on-Demand Gene Expression products (Applied Biosystems, Generation of c-Maf transgenic mice in T cells. To generate Foster City, CA). Each reaction was done in duplicate. The cDNA quantities transgenic mice expressing high levels of c-Maf specifically in T were normalized to the amount of glyceraldehyde-3-phosphate dehydroge- cells, the mouse c-Maf cDNA was inserted into the VA vector nase (GAPDH) cDNA, which was quantified using Rodent GAPDH Control under the control of the human CD2 promoter and locus control Reagents (VIC probe, Applied Biosystems). PCR reaction mixtures were pre- region (Fig. 1A). We obtained five transgenic mouse lines (Tg 17, pared according to the manufacturer’s instructions and the amplification was 50, 78, 235, and 312), which we screened by PCR using tail DNA carried out in MicroAmp optical 96-well reaction plates (Applied Biosystems) as templates (Fig. 1A and B). Genomic Southern blotting analysis in the ABI Prism 7700 Sequence Detection System (Applied Biosystems). Tissue samples of human T-cell lymphoma. Formalin-fixed, paraffin- was done to confirm the integrity and copy number of the embedded samples of a series of previously diagnosed T-cell lymphoma transgene in each transgenic mouse line. The length of the NcoI tissues, comprising 7 unspecified peripheral T-cell lymphoma (PTL), 7 adult fragment containing the c-maf transgene was 1.1 kb, whereas the T-cell leukemia (ATL), and 18 AITL cases were obtained from the files of corresponding fragment for the endogenous c-maf gene was 2.8 the Department of Pathology of Biological Response of the Nagoya kb (Fig. 1A). The results of densitometric analyses revealed that University Graduate School of Medicine (Nagoya, Japan). Multiple myeloma Tg line 50 contained approximately six copies of the transgene, samples of human tissues were used as controls. Immunohistochemical whereas Tg lines 17 and 78 contained approximately two copies staining was done using the streptavidin and biotin technique. Anti-c-Maf (Fig. 1C). We also observed that Tg lines 235 and 312 seemed antibody was provided kindly by Dr. Masaharu Sakai (University of to contain six and eight copies, respectively (data not shown). Hokkaido, Sapporo, Japan). Lymph node samples from lymphoma patients These transgenes were stably transmitted to progeny. To confirm were obtained under approval of the Ethical Committee of Kumamoto University (Kumamoto, Japan). expression of the transgene, reverse transcription-PCR (RT-PCR) Reverse transcription-PCR Southern hybridization. Total RNA was was used to monitor c-Maf mRNA levels in thymi isolated from obtained from frozen human lymph nodes from patients diagnosed with the transgenic mice. We observed that c-Maf was overexpressed T-cell lymphoma. The samples were obtained from the archives of the in the transgenic mice (Fig. 1D). The level of c-Maf mRNA in the Department of Internal Medicine II, Kumamoto University School of thymus of Tg 50 was several-fold greater than that in Tg 17 or 78 Medicine. RNA (1 Ag) isolated from each sample was reverse transcribed or in wild-type thymus, indicating that the level of expression of into the corresponding cDNA as described previously. PCR products were the transgene mRNA was copy number dependent. Thymi from then hybridized on Southern blots under stringent hybridization conditions Tg 312 mice showed relatively high levels of expression of c-Maf to specific probes. PCR primers and specific probes were as follows: human (data not shown). hypoxanthine phosphoribosyltransferase sense 5V-CTATAGACTATCAGTTC- T-cell development was arrested at the double-negative CCT-3V, antisense 5V-CTGCTTCTTACTTTTCTAACAC-3V, and probe 5V-CATT- GAACTCATATCTGTAAGAAATAAAGA-3V; human c-Maf sense 5V-CCTGCC- CD4/CD8 stage in c-Maf transgenic mice. To investigate the GCTTCAAGAGGGTGC-3V, antisense 5V-TCGCGTGTCAGACTCACATG-3V,and effect of c-Maf overexpression on T-cell differentiation, we did flow probe 5V-CGACAACCCTTCCTCTCCCGAATTT-3V; human cyclin D2 sense cytometric analysis on thymocytes isolated from transgenic mice 5V-CGTCGATGATCGCAACTGGA-3V, antisense 5V-TCCGTCACGTTGGTCC- carrying high copy number (Tg 50) at 5 weeks of age (Fig. 1E). TGAC-3V, and probe 5V-ACTGAGCTGCTGGCTAAGATCACCA-3V; and human Levels of mature, single-positive (CD4+ or CD8+) and double- ARK5 sense 5V-GAGTCCACTCTATGCATC-3V, antisense 5V-ATGTCCTCAATAG- positive (CD4+CD8+) T cells were decreased in c-Maf transgenic TGGCC-3V,andprobe5V-TCATTCGGCAAATCAGCAGCGGAGA-3V. mice (Tg 50). On the other hand, double-negative (CD4ÀCD8À) T cells were increased significantly in these mice. These results suggested that c-Maf overexpression arrested T-cell development at the double-negative (CD4ÀCD8À) stage. During thymocyte development, immature double-negative (CD4ÀCD8À) precursor cells pass through four phenotypically distinct stages defined by expression of CD44 and CD25. Double-negative cells subsequently go through DN1 (CD44+CD25À), DN2 (CD44+CD25+), DN3 (CD44ÀCD25+), and DN4 (CD44ÀCD25À) stages before giving rise to CD4+CD8+ double-positive T cells (17). We found that numbers of CD44+CD25-,CD44+CD25+,andCD44ÀCD25+ cells were increased in c-Maf transgenic thymus, whereas CD44ÀCD25À cells were significantly decreased. These results suggested that T-cell development was arrested at the DN3 stage in these mice. We found that the CD4/CD8 distribution in the lower copy number Tg mice (Tg 17 and 78 lines) was similar to that observed in wild type mice (data not shown). c-Maf transgenic mice developed lymphoma. Mice from Tg Figure 2. Incident rate of lymphoma development by c-Maf transgenic mice. Cumulative incident rate curves of lymphomas obtained from three c-Maf lines 50, 312, and 235 appeared healthy up to 50 weeks of age. transgenic lines and control mice (Tg 50, 312, and 235 and wild-type mice). However, three lines of transgenic animals subsequently developed

Cancer Res 2006; 66: (2). January 15, 2006 814 www.aacrjournals.org

Figure 3. c-Maf transgenic mice develop T-cell lymphoma. A, gross enlargement of the spleens in c-Maf transgenic mice. Bottom, representative sample; top, wild-type mice. B, invasion of spleen by malignant cells resulting in the disruption of the normal areas. H&E stain, Â50. t, tumor lesion. C, infiltration of lung by tumor cells resulting in the disruption of alveolar and bronchial tissue. H&E stain, Â50. D, atypical cells were observed in the lung. H&E stain, Â400. E and F, immunohistochemical examination revealed that infiltrated cells were mainly CD3+ cells (E, Â50; F, Â400, CD3 immunohistochemical stain). G to I, CD3+ lymphoma cells also expressed B220 (G, Â400, FITC anti-CD3 antibody; H, Â400, PE anti-B220 antibody; I, Â400, G and H merged, lung).

tumors (Fig. 2). The mean age at diagnosis was 72.2 weeks for Tg (Fig. 3A). The spleens of those mice with lymphoma were line 50 and 78.4 weeks for Tg line 312. Mice from Tg lines 17 and 78, infiltrated by numerous cells that partially or completely destroyed which contained low copy numbers of the transgene, developed normal tissue architecture (Fig. 3B). Tumor-bearing mice showed tumors in very low frequencies. The level of expression of the evidence of tumor cell invasion into pulmonary (Fig. 3C), hepatic, transgene is determined both by the activity of the locus where and renal tissue (data not shown). The infiltrating lymphoma cells integration takes place (18) and by the copy number of the exhibited a large atypical cell morphology (Fig. 3D). Immunohis- integrated genes (19). The difference in the incidence of lymphoma tochemical examination revealed that the infiltrating cells were between Tg 50, 235, and 312 lines may have arisen due to CD3+ (Fig. 3E and F). These CD3+ lymphoma cells did not express differences in the sites of transgene integration. The tumors were CD4 or CD8 (data not shown). Moreover, we found that these classified as lymphomas due to the enlargement of lymphoid double-negative (CD4ÀCD8À) T cells also expressed the B220 tissues and/or the infiltration of the tumor cells into the liver, cell surface marker, usually restricted to B cells and immature kidneys, lungs, and other internal organs. thymocytes (ref. 20; Fig. 3G-I). Lymphomas observed in c-Maf transgenic mice were of Autonomous proliferation in nude mice of T-cell lymphoma T-cell origin. We next did histologic analysis to identify the origin cells from c-Maf transgenic animals. To ascertain whether the of the lymphoma cells. The spleens of the c-Maf transgenic mice infiltrating tumor cells in c-Maf transgenic mice could proliferate were enlarged f4- to 8-fold relative to those in the wild-type mice autonomously, we isolated mononuclear cells from the spleens of www.aacrjournals.org 815 Cancer Res 2006; 66: (2). January 15, 2006

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2006 American Association for Cancer Research. Cancer Research three individual mice (two from Tg line 50 and another from Tg line 17) and injected these cells into nude mice via the tail vein (106 cells per mouse). These studies showed that the tumor cells were transplantable from nude mouse to nude mouse. Recipient nude mice displayed prominent splenomegaly and enlarged lymph nodes throughout their bodies within 6 weeks of transplantation. The tumor cells were transplantable from nude mouse to nude mouse. A typical example is shown in Fig. 4A. To assess clonality, we examined the rearrangement of the TCR locus in the c-Maf transgenic lymphoma cells. A Southern blot hybridization analysis was carried out using a murine Jh2 probe (Fig. 4B). We extracted DNA from the tail, spleen, and lymph nodes of the recipient nude mice. Fig. 4B (lane 1) contains DNA isolated from control germ line cells. Rearrangement of the TCR locus was found in DNA isolated from spleens of the c-Maf transgenic mice (Fig. 4B, lanes 2-4). The same pattern of rearrangement observed in lane 4 was present in DNA samples isolated from the nude mice (lane 5) that received lymphoma cells isolated from the mice whose DNA was probed in lane 4. These results support our contention that clonal lymphoma cells committed to the T lymphoid lineage proliferate in the c-Maf transgenic mice. Expression of c-Maf target genes in T-cell lymphoma. To gain Figure 5. Analysis of c-Maf target genes by quantitative RT-PCR. Expression insight into the molecular mechanism of the development of the profiles of c-Maf and target genes were examined by real-time RT-PCR T-cell lymphoma, we measured the level of the c-Maf and target analysis. Total RNA (2 Ag) obtained from infiltrating lymphoma cells from the thymus and enlarged spleen. Columns, mean; bars, SE. 5, thymus mRNA from wild-type thymus; , thymus from c-Maf Tg 50 line; , spleen infiltrated lymphoma from Tg 50 line. *, P < 0.05.

genes mRNA in c-Maf transgenic mice using a quantitative RT-PCR assay (Fig. 5). As we described previously, c-Maf overexpression has been implicated in the development of multiple myeloma, a mature B-cell lymphoid tumor. However, a relationship between c-Maf overexpression and T-cell lymphomagenesis has yet to be established. Using gene expression profiling, Hurt et al. (13) identified cyclin D2 and integrin h7 as c-Maf target genes in myeloma. c-Maf transactivated the cyclin D2 promoter and enhanced myeloma cell proliferation, whereas dominant-negative c-Maf blocked tumor formation in immunodeficient mice. c-Maf-driven expression of integrin h7 enhanced myeloma adhesion to the bone marrow stroma and increased the production of vascular endothelial growth factor (VEGF), which led to effects on cellular proliferation and adhesion (13). Recently, we observed high expression of ARK5 in human myeloma cells expressing high levels of c-Maf (21). ARK5 is the fifth member of the AMP-activated protein kinase (AMPK) catalytic subunit family, which is also known as the cellular stress response factor (22–24). Furthermore, experiments done in vitro and in nude mice revealed that ARK5- expressing tumor cells were characterized by accelerated invasion and metastasis (25). These facts suggest a relationship between ARK5 and c-Maf in other cancers, such as T-cell lymphoma. c-Maf is also known as a Th2-specific transcription factor, which Figure 4. Autonomous proliferation of c-Maf-induced lymphoma cells in the recipient nude mice. A, nude mice transplanted with mononuclear cells isolated promotes the differentiation of Th2 cells mainly by an IL-4- from the spleens of c-Maf transgenic mice gave rise to infiltration of the spleens dependent mechanism (6). Thus, we next analyzed expression of the recipient mice. Moreover, we observed swelling of multiple lymph nodes of cyclin D2, integrin h , ARK5, and IL-4 in c-Maf-induced T-cell in the recipient mice. Arrowheads, lymph nodes. B, Southern blot analyses of 7 infiltrating lymphoma cells from the spleen of c-Maf transgenic mice and recipient lymphoma. The level of c-Maf expression in the thymus and nude mice analyzed with a Jh2 probe after digestion with Hind III. Lane 1 lymphoma-infiltrated spleens of c-Maf transgenic mice was 8- to (tail from wild-type mouse), lanes 2 to 4 [enlarged spleen from c-Maf transgenic mice (Tg 50 line), independent samples], and lane 5 (a sample from a recipient 10-fold greater than levels in wild-type mice (Fig. 5). We found nude mouse transplanted with cells isolated from a Tg mouse with an enlarged elevation of the expression of both integrin h7 and cyclin D2 in spleen, corresponding to lane 4) showed the presence of a DNA fragment c-Maf transgenic mice, consistent with what was reported corresponding to the germ line TCR gene (arrowheads). Lanes 2 to 5 also showed DNA fragments corresponding to a rearranged TCR gene (asterisk). previously in myeloma (13). Importantly, we observed a significant Lanes 4 and 5 showed the same rearranged fragment. elevation of ARK5 expression, which has been shown to accelerate

Cancer Res 2006; 66: (2). January 15, 2006 816 www.aacrjournals.org

Figure 6. Analyses of c-Maf expression in human T-cell lymphoma cases. A, histology of AITL. H&E stain. B, immunostaining analysis of the same case of (A) with anti-c-Maf antibody. C, anti-c-Maf immunostaining of a human multiple myeloma case. D, anti-c-Maf immunostaining of a human reactive lymphadenopathy case.

invasion and metastasis in tumor cells. There was no difference of c-Maf expression levels in AITL (Fig. 6A and B). We also in IL-4 expression between wild-type and transgenic mice, which stained multiple myeloma samples as a positive control (Fig. 6C). might be due to dysregulation of normal T-cell differentiation in For a nonlymphoma control, we analyzed reactive lymphoaden- c-Maf transgenic mice. opathy samples. We found that not many but some lymph node Overexpression of c-Maf is also identified in a type of human cells, including lymphocytes and macrophages expressed c-Maf T-cell lymphoma. The evaluation of c-Maf transgenic mice has (Fig. 6D). thus shown that overexpression of c-Maf is able to induce T-cell To confirm c-Maf overexpression in AITL, c-maf gene expression lymphoma in mice. To examine whether c-Maf overexpression was analyzed by RT-PCR (Fig. 7). RT-PCR analysis showed that might contribute to the development of human lymphomas, we c-Maf was overexpressed in AITL but not in other lymphoma cases. analyzed c-Maf expression in samples of human lymphoma by Moreover, ARK5 was also expressed at a high level in AITL. In immunohistochemical and RT-PCR analyses. We tested 7 cases addition, we did RT-PCR on RNA isolated from two additional of unspecified PTL, 7 cases of ATL, and 18 cases of AITL using cases of AITL. The results showed that these two cases also an anti-c-Maf antibody. Immunostaining analysis (Fig. 6A-D) overexpressed both c-Maf and ARK5 (data not shown). We thus revealed that 60% of AITL (11 of 18 cases) samples were positive confirmed that c-Maf and ARK5 were both overexpressed in 4 of 4 for nuclear staining by anti-c-Maf, whereas only 29% of PTL or cases of AITL. These results suggest that overexpression of c-Maf ATL (2 of 7 cases) expressed c-Maf. We have found variable levels and ARK5 may contribute to the development of AITL. www.aacrjournals.org 817 Cancer Res 2006; 66: (2). January 15, 2006

Discussion overexpression of c-Maf may not be sufficient for full transfor- It is important to emphasize that v-maf is a classic oncogene mation of lymphocytes. identified in an avian transforming virus (1). Furthermore, Kataoka In addition, the number of double-negative T cells was et al. have shown that overexpression of wild type c-maf is capable significantly increased in Tg 50 lines, which also developed tumors. of contributing to transformation of fibroblasts (26). Recently, it However, we found that the CD4/CD8 distribution in the lower has been reported that c-maf translocation and/or overexpression copy number Tg lines (Tg 17 and 78), which developed tumors at a are frequently identified in human multiple myeloma (10, 13). very low frequency, was similar to that observed in wild type mice. However, neither a direct oncogenic function of c-Maf in vivo nor a These results suggest there may be a similar threshold level of role for c-Maf in other cancers has yet been reported. In this study, c-Maf expression for both arrest of T-cell development and tumor we have shown that c-Maf contributes to T-cell lymphoma in vivo development in c-Maf transgenic mice. in c-Maf transgenic mouse models. Because c-Maf is a transcription factor, it is of interest to The c-Maf transgenic mouse model shows several interesting identify downstream target genes of its action that might features. First, c-Maf transgenic mice develop T-cell lymphoma or contribute to malignant transformation of T cells. We used RT- h plasmacytoma at older ages. The mean age at diagnosis was f80 PCR analysis to show that cyclin D2, integrin 7, and ARK5 were weeks. The advanced age at which the disease appeared in the up-regulated in c-Maf transgenic thymocytes and T lymphoma transgenic mice is reminiscent of the fact that both multiple cells. The transcriptional activation of cyclin D2 by c-Maf myeloma and AITL develop predominantly in elderly human provides a potential mechanism for the promotion of cell cycle patients. These findings suggest a link between the mechanism of progression in myeloma cells (13). It is not surprising that c-Maf c-Maf oncogenesis in humans and our mouse model. Interestingly, induces cyclin D2 expression in T-cell lymphoma, because this lymphoma cells from the c-Maf transgenic mice exhibited a rare effect has been already shown in multiple myeloma cells. We h CD3/B220 double-positive phenotype, suggesting a potential also observed overexpression of integrin 7 in T-cell lymphoma. h target population for c-Maf action. RBTN-2 has been shown to c-Maf-driven expression of integrin 7 enhances myeloma adhe- induce human T-cell acute leukemia in childhood. Interestingly, sion to the bone marrow stroma and increases the production of overexpression of Rbtn-2 in mice under control of the CD2 VEGF, which enhances cellular proliferation and adhesion (13). regulatory elements also leads to the development of T-cell Although the interaction of the bone marrow stroma with lymphoma. Lymphoma cells isolated from these mice also multiple myeloma cells is an important component of the exhibited a range of phenotypes, the most notable being CD3/ pathophysiology of this malignancy (31, 32), there have been few B220 double-positive cells (27). Interestingly, a subset of B220+ reports describing the involvement of adhesion molecules, such h thymocytes, which are thought to be T-cell progenitor cells (20), as integrin 7, in the development of T-cell lymphoma. A more h are particularly susceptible to malignant transformation by detailed analysis of the role of integrin 7 in T-cell function is Philadelphia chromosome–dependent BCR-ABL fusion proteins necessary to understand the contribution of this molecule to (28). Thus, CD3/B220 cells might be particularly sensitive to T-cell malignancy. Interestingly, we found that ARK5 mRNA is malignant transformation. Rbtn-2 transgenic mice develop acute up-regulated in c-Maf transgenic thymocytes and lymphoma. thymic lymphomas (27), whereas c-Maf transgenic mice develop ARK5 is also known as the cellular stress response factor peripheral, long-latency cell tumors probably by transformation of (22–24). Previously, it has been shown that overexpression of the same CD3+/B220+ precursor thymocyte. This is a significant ARK5 mRNA is associated with tumor progression in colon feature of c-maf, because most proto-oncogenes activated in the cancer and that metastatic foci of colon cancer in the liver show thymus [rbtn-2 (27), HMG-I (29), Rgr (30), etc.] cause T-cell acute markedly increased expression of ARK5 mRNA (33). Furthermore, lymphoblastic leukemia and not peripheral T-cell cancers. We do studies in vitro and in nude mice have revealed accelerated not know the reason behind the difference between c-maf and invasion and metastasis in ARK5-expressing tumor cells (25). other genes in promoting tumor formation, but one possible ARK5 is transcriptionally regulated by members of the large explanation is that c-maf is only weakly oncogenic and that Maf family through the MARE sequence in its promoter (21). Therefore, ARK5 gene transcription is tightly regulated by c-Maf and overexpression of ARK5 might be one of the causes of malignant transformation in c-Maf transgenic mice. c-Maf is a Th2-specific transcription factor that activates the IL-4 promoter (6). However, we did not observe an increase in IL-4 expression in our c-Maf transgenic mice. Flow cytometric analysis of thymocytes and lymphoma cells revealed an increase in levels of immature CD4ÀCD8À T cells in the c-Maf transgenic mice. This result suggests that c-Maf alone is insufficient to transactivate the endogenous IL-4 promoter in this immature T cell in vivo. Our results in mice suggested that c-Maf overexpression might contribute to the development of some cases of human lymphoma. On examination of human lymphoma cases, we found one type of T-cell lymphoma, classified as AITL, frequently expressed high levels of c-Maf. AITL is an uncommon T-cell lymphoma characterized by clear cell accumulation and vascular proliferation. Figure 7. RT-PCR analysis of human lymphomas. c-Maf and ARK5 were AITL show the most prominent vascular component among overexpressed in AITL patients (lanes 2 and 5). Lane 1, extranodal natural killer/ T-cell lymphoma, nasal type; lane 2, AITL; lane 3, anaplastic large T-cell lymphomas and their prognosis is difficult to predict. VEGF gene lymphoma; lane 4, PTL; lane 5, AITL. was overexpressed in both AITL lymphoma and endothelial cells.

Cancer Res 2006; 66: (2). January 15, 2006 818 www.aacrjournals.org

The vascular component, a critical pathologic characteristic in development of both human and mouse lymphomas. Clarification AITL, also contributes to lymphoma progression (34). Mentioned of the mechanism by which c-Maf contributes to oncogenesis may above, c-maf-driven expression of integrin h7 enhanced myeloma ultimately facilitate the discovery of more specific therapies to adhesion to bone marrow stroma and increased production of prevent the progression of the diseases. VEGF (13). c-Maf might contribute to hypervascularity and prognosis in AITL. The results presented in this article suggest that overexpression of c-Maf might contribute to the development Acknowledgments of this lymphoma. Although we have yet to examine the trans- Received 6/20/2005; revised 9/29/2005; accepted 11/11/2005. Grant support: Grant-in Aid for Scientific Research on Priority Areas and Genome location of the c-maf gene to the TCR locus in AITL cells, such a Network Project from the Ministry of Education, Culture, Sports, Science and translocation may yet be identified in these cells as has been Technology, Japan. already shown in some cases of multiple myeloma. In this study, we 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 have shown that c-Maf transgenic mice develop T-cell lymphoma. with 18 U.S.C. Section 1734 solely to indicate this fact. Further studies will be needed to define the role of c-Maf in the We thank Drs T. Sugawara and Y. Takahama for help and discussion.

References 12. Seidl S, Kaufmann H, Drach J. New insights into the 24. Kemp BE, Stapleton D, Campbell DJ, et al. AMP- pathophysiology of multiple myeloma. Lancet Oncol activated protein kinase, super metabolic regulator. 1. Nishizawa M, Kataoka K, Goto N, Fujiwara KT, 2003;4:557–64. Biochem Soc Trans 2003;31:162–8. Kawai S. v-maf, a viral oncogene that encodes a 13. Hurt EM, Wiestner A, Rosenwald A, et al. Over- 25. Suzuki A, Lu J, Kusakai G, Kishimoto A, Ogura T, ‘‘leucine zipper’’ motif. Proc Natl Acad Sci U S A 1989; expression of c-maf is a frequent oncogenic event in Esumi, H. ARK5 is a tumor invasion-associated factor 86:7711–5. multiple myeloma that promotes proliferation and downstream of Akt signaling. Mol Cell Biol 2004;24: 2. Kataoka K, Noda M, Nishizawa M. Maf nuclear pathological interactions with bone marrow stroma. 3526–35. oncoprotein recognizes sequences related to an AP-1 Cancer Cell 2004;5:191–9. 26. Kataoka K, Nishizawa M, Kawai S. Structure-function site and forms heterodimers with both Fos and Jun. 14. Kienast J, Berdel WE. c-maf in multiple myeloma: an analysis of the maf oncogene product, a member of the Mol Cell Biol 1994;14:700–12. oncogene enhancing tumor-stroma interactions. Cancer b-Zip protein family. J Virol 1993;67:2133–41. 3. Swaroop A, Xu J, Pawar H, Jackson A, Skolnick C, Cell 2004;5:109–10. 27. Larson RC, Fisch P, Larson TA, et al. T cell tumours of Agarwal N. A conserved retina-specific gene encodes a 15. Zhumabekov T, Corbella P, Tolaini M, Kioussis D. disparate phenotype in mice transgenic for Rbtn-2. basic motif/leucine zipper domain. Proc Natl Acad Sci Improved version of a human CD2 minigene based Oncogene 1994;9:3675–81. U S A 1992;89:266–70. vector for T cell–specific expression in transgenic mice. 28. Clark SS, Chen E, Fizzotti M, Witte ON, Malkovska V. 4. Ogino H, Yasuda K. Induction of lens differentiation by J Immunol Methods 1995;185:133–40. BCR-ABL and v-abl oncogenes induce distinct patterns activation of a bZIP transcription factor, L-Maf. Science 16. Rabbitts TH, Forster A, Larson R, Nathan P. Fusion of of thymic lymphoma involving different lymphocyte 1998;280:115–8. the dominant negative transcription regulator CHOP subsets. J Virol 1993;67:6033–46. 5. Kajihara M, Kawauchi S, Kobayashi M, Ogino H, with a novel gene FUS by translocation t(12;16) in 29. Xu Y, Sumter TF, Bhattacharya R, et al. The HMG-I Takahashi S, Yasuda K. Isolation, characterization, and malignant liposarcoma. Nat Genet 1993;4:175–80. oncogene causes highly penetrant, aggressive lymphoid expression analysis of zebrafish large Mafs. J Biochem 17. Shortman K, Wu L. Early T lymphocyte progenitors. malignancy in transgenic mice and is overexpressed in (Tokyo) 2001;129:139–46. Annu Rev Immunol 1996;14:29–47. human leukemia. Cancer Res 2004;64:3371–5. 6. Ho IC, Hodge MR, Rooney JW, Glimcher LH. The 18. Palmiter RD, Brinster RL. Germ-line transformation 30. Jimenez M, Perez de Castro I, Benet M, Garcia JF, proto-oncogene c-maf is responsible for tissue-specific of mice. Annu Rev Genet 1986;20:465–99. Inghirami G, Pellicer A. The Rgr oncogene induces expression of interleukin-4. Cell 1996;85:973–83. 19. Lang G, Wotton D, Owen MJ, et al. The structure of tumorigenesis in transgenic mice. Cancer Res 2004;64: 7. Kawauchi S, Takahashi S, Nakajima O, et al. Regulation the human CD2 gene and its expression in transgenic 6041–9. of lens fiber cell differentiation by transcription factor mice. EMBO J 1988;7:1675–82. 31. Anderson KC. Multiple Myeloma. Advances in c-Maf. J Biol Chem 1999;274:19254–60. 20. Kishihara K, Penninger J, Wallace VA, et al. Normal B disease biology: therapeutic implications. Semin Hema- 8. Kim JI, Li T, Ho IC, Grusby MJ, Glimcher LH. lymphocyte development but impaired T cell matura- tol 2001;38:6–10. Requirement for the c-Maf transcription factor in tion in CD45-exon6 protein tyrosine phosphatase- 32. Shain KH, Landowski TH, Buyuksal I, Cantor AB, crystallin gene regulation and lens development. Proc deficient mice. Cell 1993;74:143–56. Dalton WS. Clonal variability in CD95 expression is the Natl Acad Sci U S A 1999;96:3781–5. 21. Suzuki A, Iida S, Kato-Uranishi M, et al. ARK5 is major determinant in Fas-medicated, but not chemo- 9. Ring BZ, Cordes SP, Overbeek PA, Barsh GS. Regulation transcriptionally regulated by the Large-MAF family and therapy-medicated apoptosis in the RPMI 8226 multiple of mouse lens fiber cell development and differentiation mediates IGF-1-induced cell invasion in multiple myelo- myeloma cell line. Leukemia 2000;14:830–40. by the Maf gene. Development 2000;127:307–17. ma: ARK5 as a new molecular determinant of malignant 33. Kusakai G, Suzuki A, Ogura T, et al. ARK5 expression 10. Chesi M, Bergsagel PL, Shonukan OO, et al. Frequent multiple myeloma. Oncogene 2005;24:6936–44. in colorectal cancer and its implications for tumor dysregulation of the c-maf proto-oncogene at 16q23 by 22. Hardie DG, Carling D. The AMP-activated protein progression. Am J Pathol 2004;164:987–95. translocation to an Ig locus in multiple myeloma. Blood kinase-fuel gauge of the mammalian cell? Eur J Biochem 34. Zhao WL, Mourah S, Mounier N, et al. Vascular 1998;91:4457–63. 1997;246:259–73. endothelialgrowthfactor-Aisexpressedbothon 11. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving 23. Kemp BE, Mitchelhill KI, Stapleton D, et al. Dealing lymphoma cells and endothelial cells in angioimmuno- genetic events and host interactions. Nat Rev Cancer with energy demand: the AMP-activated protein kinase. blastic T-cell lymphoma and related to lymphoma 2002;2:175–87. Trends Biochem Sci 1999;24:22–5. progression. Lab Invest 2004;84:1512–9.

www.aacrjournals.org 819 Cancer Res 2006; 66: (2). January 15, 2006

Naoki Morito, Keigyou Yoh, Yuki Fujioka, et al.

Cancer Res 2006;66:812-819.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/2/812

Cited articles This article cites 34 articles, 13 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/2/812.full#ref-list-1

Citing articles This article has been cited by 16 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/2/812.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/66/2/812. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.