HMG-I/Y is overexpressed in human cancer, although a direct role for this gene in transformation has not been established. We generated transgenic mice with HMG-I targeted to lymphoid cells. All seven informative founder HMG-I mice developed aggressive lymphoma by a mean age of 4.8 months. Tumors express T-cell markers and are transplantable. We also demonstrate that HMG-I mRNA and protein are increased in human acute lymphocytic leukemia samples. Our results show that HMG-I functions as an oncogene and suggest that it contributes to the pathogenesis of leukemia and other cancers with increased HMG-I expression.
Although HMG-I/Y expression is increased in a variety of human cancers and portends a poor prognosis in some tumors (1, 2, 3, 4) , the contribution of this gene to malignant transformation is unknown. The HMG-I/Y gene encodes the HMG-I and HMG-Y protein isoforms (1, 2, 3) , which function as architectural chromatin-binding proteins involved in regulating gene expression (1, 2, 3 , 5, 6, 7, 8) . Interestingly, this gene is located on the short arm of chromosome 6, in a region known to be involved in rearrangements, translocations, and other abnormalities correlated with human cancers (1, 2, 3) . We showed previously (9) that HMG-I/Y is a c-Myc target gene involved in Burkitt’s lymphoma, an aggressive childhood malignancy that develops as a consequence of deregulated c-myc expression. HMG-I/Y genes have oncogenic properties in cell lines derived from various tissues (9, 10, 11, 12) . In addition, decreasing HMG-I/Y proteins blocks transformation in human cancer cell lines (9 , 12 , 13) . These studies suggest that HMG-I/Y may function as an oncogene important in those cancers associated with increased HMG-I/Y expression. To further explore the role of HMG-I/Y in neoplastic transformation, we generated transgenic mice overexpressing HMG-I in lymphoid tissue. Here, we report that 100% of the HMG-I transgenic mice develop lymphoid tumors. Our findings provide the most direct evidence that overexpression of HMG-I leads to malignancy.
Materials and Methods
Construction of Transgenic Mice.
cDNA encoding murine HMG-I was excised from pBS-I (9) with BamHI and SalI and inserted into the SalI restriction sites of the vector pHSE3′ (14 , 15) . This construct was cleaved with XhoI to release a DNA fragment containing the HMG-I coding sequence flanked by the H-2K promoter and immunoglobulin μ intronic enhancer. The transgene was injected into fertilized eggs from B6C57/SJL females and maintained by mating hemizygous animals with wild-type B6C57 mice. Germ-line transmission was assessed by Southern blot analysis of tail DNA.
Quantitative Reverse Transcription-PCR Analysis.
Tissues from transgenic mice were stored in TRIzol at −80°C, and total RNA was extracted from lymphoid organs according to the manufacturer’s instructions (Invitrogen). RNA (10 ng) was used in 20-μl reaction mixtures with SYBR Green PCR Master Kit and TaqMan One-Step reverse transcription-PCR kit (Applied Biosystems) using the 7700 model ABI PRISM sequence detector (Applied Biosystems). The sequences of the forward and reverse primers that amplify HMG-I are 5′-AAAGAGCCCAGTGAAGTGCCAAC and 5′-GTCAAAGAACCCAGACGGG, respectively. The sequences for the forward and reverse primers that amplify the internal control, GAPDH, are 5′-CATCTCCGCCCCTTCTGCTG and 5′-GCTAAGCAGTTGGTGGTGC, respectively. Reaction conditions were 30 min at 48°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, and 45 s at 60°C.
Total RNA was isolated from patient leukemia samples (106 to 108 cells) and mouse T and B cells using TRIzol and RNAEasy (Qiagen). RNA (100 ng) was used in 20-μl reactions using TaqMan One-Step reverse transcription-PCR kit (Applied Biosystems). The sequences for the forward and reverse primers that amplify HMG-I are 5′-AGGAAAAAGGATGGCACTGAGAA and 5′-CCCTGAGGTCTCTTAGGTTTTGG, respectively. The sequence for the specific HMG-I probe is 5′-CTCCTTCTGACTCCCTCCCAGCGC. Human phosphoprotein was used as a control with commercially available primers and probes (Applied Biosystems). Reaction conditions were 30 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, and 45 s at 60°C. Reverse transcription-PCR reactions were performed in triplicate, and experiments were performed at least twice.
Fluorescence-Activated Cell-Sorting Analysis.
Cells from lymphoid organs were analyzed using a FACScan analyzer (15) .
Transfer of Tumors to Mice.
Splenocytes (107 or 108) from transgenic mice or control, nontransgenic littermates were injected i.p. into nu−/− or rag−/− mice. Mice that died underwent necropsy. Mice that showed signs of tumor growth were sacrificed and underwent necropsy. Mice without evidence of tumor were observed for 20–24 weeks, sacrificed, and analyzed.
Tissues were fixed in 10% formalin and embedded in paraffin. Sections (5 μm) were stained with H&E.
Blood smears were obtained by tail vein puncture every 2–4 weeks after the age of 6–7 months and at necropsy. Complete blood counts were obtained at necropsy by Coulter counter calibrated for mice.
B- and T-Cell Sorting.
Splenocytes (5 × 107 cells) were separated into B and T cells using MACS CD45R (B220) and MACS CD90 (Thy1.2) antibodies and magnetic microbead sorting (Mitenyi Biotec).
Immunohistochemistry for terminal deoxynucleotidyltransferase was performed using commercially available antibodies (Supertech #C-99300) and a DAKO autostainer.
Construction of Transgenic Mice and Transgenic Expression.
The HMG-I transgene was targeted to lymphoid cells by the murine H-2K promoter and immunoglobulin μ enhancer, which drive expression in B and T cells (Fig. 1A ⇓ ; Refs. 14 and 15 ). We generated eight independent founders with 1–28 copies of the HMG-I transgene (Fig. 1B) ⇓ . Seven founders developed lymphoid tumors; the eighth died at 2 months, and tissue autolysis precluded further analysis.
HMG-I expression is increased in lymphoid tissues of the transgenic mice compared with that of control littermates (Fig. 1, C and D) ⇓ . Those organs with extensive invasion by malignant cells have the highest levels of HMG-I mRNA. We assessed HMG-I expression in T and B lymphocytes from a mouse with 20 copies of the transgene. We showed increased HMG-I mRNA in the T cells (18-fold) and B cells (3-fold) compared with those cells from a control littermate (Fig. 1E) ⇓ . HMG-I protein is also increased in the splenocytes of transgenic mice (Fig. 1F) ⇓ . The mouse with the most extensive disease had higher HMG-I protein levels, which probably reflects a higher proportion of tumor cells in the sample (Fig. 1F ⇓ , right lane).
Lymphoid Malignancy in 100% of HMG-I Transgenic Mice.
The seven founders all developed organomegaly and died between 1 and 8.5 months of age, with a mean age of death of 4.8 ± 2.6 months. Histopathological analysis showed lymphoma in all cases (Fig. 2, A–G) ⇓ . There was no correlation between onset of disease and copy number. The founder with 20 copies of the transgene was successfully bred to establish a transgenic line of mice; the rest died before sexual maturity or were infertile.
All of the descendants of the founder mouse with 20 copies of the transgene (1 second generation and 22 third generation) were sacrificed and analyzed histopathologically when they developed signs of disease (abnormal lymphoblasts in the peripheral blood, decreased activity, and/or organomegaly). Four mice also developed rectal (3) or cervical (1) prolapse. There were abnormal lymphoblasts infiltrating the prolapsed organ, which probably contributed to the prolapses. All transgenic mice with 20 copies of the HMG-I transgene from three generations developed lymphoid malignancy at a mean age of 9.6 ± 2.4 months. Histopathological analysis (Table 1) ⇓ showed lymphoma in the spleen of all mice analyzed (23 of 23). In most cases, lymphoma was found in the thymus (20 of 22), bone marrow (21 of 23), lymph nodes (21 of 24), and gut-associated lymphoid tissue (18 of 23). Blood smears (Fig. 2G) ⇓ showed abnormal, circulating lymphoblasts (19 of 21), and blood counts showed a pronounced leukocytosis (8 of 15). The presence of abnormal lymphoblasts in the bone marrow, spleen, and peripheral blood in most mice (17 of 18) suggests a leukemia-like process. Abnormal lymphoblasts were observed in the liver (11 of 23), kidneys (8 of 23), and lungs (5 of 24). All mice had at least two organs affected, and most had multiorgan involvement. These findings indicate aggressive, widespread disease in the HMG-I transgenic mice.
To determine the phenotype of the transformed cells from the line of mice with 20 copies of the transgene, fluorescence-activated cell-sorting analysis was performed and showed CD3+, CD4−, CD8+, and αβTCR+ T-cell markers (Fig. 3A) ⇓ . Cells are negative for the Natural Killer 1.1 (NK 1.1) and terminal deoxynucleotidyltransferase (data not shown), consistent with a mature T-cell phenotype (16) . A homogenous population of cells with a single T-cell receptor rearrangement was observed in four of seven mice, including Vβ2 (one of seven), Vβ8.2 (two of seven), and Vβ6 (one of seven), as expected for clonal disease (Fig. 3B) ⇓ . The founder with 1 copy of the transgene had no gross thymic abnormalities with lymphoma involving the spleen, bone marrow, and lymph nodes. HMG-I mRNA was increased in the spleen, bone marrow, and lymph nodes, but not the thymus (Fig. 1D) ⇓ . The absense of thymic involvement in this founder suggests B-cell disease. The remaining founders had thymic disease, consistent with a T-cell phenotype. The basis for the preponderance of T-cell disease in the HMG-I mice is not known, although it is possible that factors present in T cells facilitate or enhance expression of the HMG-I transgene.
Tumors from the HMG-I Mice Are Transplantable.
Mice that received injection with splenocytes from the transgenic mice developed lymphoma and died after 8–16 weeks, depending on the number of cells injected. Splenocytes (107) from an HMG-I mouse with lymphoma were injected into a nu−/− mouse, which developed lymphoma after 16 weeks. When 108 splenocytes from another transgenic mouse with lymphoma were injected into five rag−/− mice, all developed organomegaly and died after 8 weeks with lymphoma. Fluorescence-activated cell-sorting analysis from the tumor cell-injected mice showed that the lymphoid tumors had the same T-cell markers as the original tumor. None of the control mice injected with the same number of splenocytes from a nontransgenic littermate developed lymphoid disease.
HMG-I Expression in Human Acute Lymphocytic Leukemia.
We showed previously (9 , 10) that overexpression of HMG-I leads to a transformed phenotype in human lymphoid cells. HMG-I is also increased in Burkitt’s lymphoma (9) and five human acute lymphocytic leukemia samples (17) . Taken together with our HMG-I mice, these findings suggest that HMG-I is involved in the pathogenesis of human lymphocytic leukemia. Therefore, we assessed HMG-I expression in bone marrow samples from patients with acute lymphocytic leukemia and acute lymphocytic leukemia cell lines. HMG-I mRNA was increased in all acute lymphocytic leukemia samples (28 of 28) and both cell lines by 2–9-fold compared with normal lymphocytes (Fig. 3C) ⇓ . We also showed increased HMG-I protein levels in the patient samples (10 of 10 samples; Fig. 3D ⇓ ) and both cell lines. In all cases, the protein levels correlate with the mRNA.
The results of our studies clearly show that HMG-I overexpression is responsible for malignant transformation in vivo. The early onset of tumors, widespread disease, and 100% penetrance demonstrate that HMG-I overexpression causes aggressive disease in mice. The mechanisms involved in transformation by HMG-I are unknown. The onset of tumors after 4 months suggests that more than one genetic event is involved. Because HMG-I proteins function in transcriptional regulation (1, 2, 3 , 5, 6, 7, 8) , overexpression of HMG-I may lead to transformation by altering expression of genes involved in regulation of cell growth. Recent studies also show that HMG-I proteins may contribute to genomic instability (18) . Preliminary data with spectral karyotyping analysis of cell lines overexpressing HMG-I and tumor cells from our transgenic mice suggest that overexpression of HMG-I is correlated with an increased frequency of translocations, polyploidy, and other chromosomal abnormalities observed in cancer. 5
HMG-I/Y is a c-Myc target gene important in the transformed phenotype in Burkitt’s lymphoma (9) . Our HMG-I transgenic mice are phenotypically similar to transgenic mice overexpressing c-myc in lymphoid tissue (19 , 20) . c-myc transgenic mice develop lymphoma between 4 and 14 months of age. In contrast to the HMG-I mice, the c-myc transgene was targeted only to B cells, and these mice therefore develop B-cell disease. To our knowledge, HMG-I is the only c-Myc target gene to exhibit malignancy in a transgenic mouse model. These results further suggest that HMG-I is critical for c-Myc function in transformation.
HMG-I may also function independent of c-Myc in some malignancies. For example, there are cancers known to express high levels of HMG-I protein without concurrent elevations in the c-Myc protein (12) . Analysis of the HMG-I/Y promoter showed that HMG-I/Y is also regulated by transcription factors other than c-Myc (9) . Thus, in some cancers, factors other than c-Myc may enhance HMG-I/Y expression. It is possible that other mechanisms, such as gene amplification, may enhance HMG-I/Y expression in cancer, although this has not yet been demonstrated.
Because HMG-I overexpression occurs in aggressive human malignancy, including Burkitt’s lymphoma (9) , we investigated human acute lymphoid leukemia samples to determine whether our HMG-I mouse model is applicable to human disease. Notably, we observed that HMG-I expression was increased in all lymphoid leukemia samples studied. A larger, more comprehensive study is needed to determine the prevalence of HMG-I overexpression in human lymphoid malignancy and whether HMG-I expression correlates with prognosis in this disease.
In summary, we show that overexpressing HMG-I in lymphoid tissue leads to an aggressive T-cell leukemia and lymphoma in transgenic mice. We also show that HMG-I is overexpressed in human lymphoid leukemia. Our findings demonstrate that HMG-I functions as an oncogene in vivo and suggest that it is important in the pathogenesis of human leukemia. The HMG-I transgenic mouse should provide a valuable model for human lymphoid malignancy and a useful tool for identifying the mechanisms through which overexpression of HMG-I leads to transformation. Because increased HMG-I expression occurs in a wide range of aggressive human cancers, this model should enhance our understanding of cancer biology relevant to tumors of various origins. Understanding the mechanisms that underlie transformation may also identify novel therapeutic targets for malignancy associated with HMG-I overexpression.
We thank Steve Desiderio and Dominic Dordai for guidance with fluorescence-activated cell-sorting analysis, Chi Dang for advice and reagents, Donald Coffey for insightful discussions, Judy Hutchison for technical support, John Lauro and Manju Kaushal for the terminal deoxynucleotidyltransferase analysis, Barbara Migeon for thoughtful review of the manuscript, and The Johns Hopkins Transgenic Core Facility. The Sidney Kimmel Comprehensive Cancer Center Tumor and Cell Procurement Bank provided the leukemia samples.
Grant support: Grants from the National Cancer Institute and American Cancer Society (to L. M. S. Resar). F. Sumter, R. Bhattacharya, L. J. Wood, and A. Tesfaye were supported by a training grant from the NIH.
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.
Note: This work is dedicated to the memory of Dr. Daniel Nathans. L. J. Wood is currently at Oregon Health & Science University School of Nursing, Portland, OR 97329.
Requests for reprints: Linda M. S. Resar, The Johns Hopkins University School of Medicine, The Ross Research Building, Room 1125, 720 Rutland Avenue, Baltimore, MD 21205. Phone: (410) 955-6132; Fax: (410) 955-8208; E-mail:
↵5 L. M. S. Resar, unpublished results.
- Received January 7, 2004.
- Revision received February 18, 2004.
- Accepted March 12, 2004.
- ©2004 American Association for Cancer Research.