Cytomegalovirus Infection Leads to Pleomorphic Rhabdomyosarcomas in Trp53^+/− Mice

Rhabdomyosarcoma (RMS) is a lethal tumor believed to originate from immature striated muscle cells that primarily affects children (1). Approximately half of patients with RMS perish and survival has changed little in the last few decades (2). RMS is divided into 3 subtypes: embryonal, alveolar, and pleomorphic. Both embryonal and alveolar are more common in children, whereas the more lethal pleomorphic subtype occurs primarily in adults with a median survival of 2.25 years (3).

Human cytomegalovirus (HCMV) is a DNA β-herpes virus. Like all herpesviridae, acute infection is followed by a lifelong latent infection that is re-activatable. HCMV is present in 70% to 90% of adults and in 40% of one-year-olds (4). It infects several human cells including RMS (5) and myoblasts (6) in vitro, shown to be a cell of origin for RMS (7). Accumulating evidence suggests a link between persistent HCMV infection and cancer. Several reports have shown HCMV presence in solid tumors including glioma, medulloblastoma, breast, prostate, and colorectal cancer (8–10). This warrants research into how this virus is linked to cancer.

Although some have reported a high frequency of TP53 mutations in RMS (11, 12), this is not universally accepted (13). Heterozygous Trp53^+/− mice develop a mixed spectrum of tumors occurring at a late age (>9 months; ref. 14). Combining Trp53 heterozygosity with K-Ras oncogene activation in mice leads to pleomorphic RMS, suggesting a cooperative role between the 2 mutations (15). HCMV has been shown to interact with many cancer pathways, some of which are implicated in RMS. HCMV can interact with p53 (9); IE2 can bind directly to p53, inactivating p53-dependent apoptosis; UL44 inhibits p53 transcriptional activity (16); and IE1 reduces p53 expression in glioma cells.

There is no evidence for a HCMV role in RMS. Using Trp53^+/− mice, we report that MCMV infection leads to the development of murine pleomorphic RMS. We show HCMV gene expression in human RMS, implicating this virus in the pathogenesis of these neoplasms.

See Supplementary Methods.

The Institutional Animal Care and Use Committee at The Ohio State University (Columbus, OH) approved mouse experiments and care. Trp53^+/− mice (see Supplementary Methods) were intraperitoneally (i.p.) injected on postnatal day 2 (P2) with 10³ plaque-forming units or at 4 weeks of age with 10⁶ plaque-forming units with an MCMV strain that possesses an ablated m157 allele to enhance infectivity in B6 mice (17) in 100 μL of PBS or PBS only (Mock). MCMV infection in mice is considered a reasonable surrogate for modeling HCMV infection in humans. An additional cohort received 10³ plaque-forming units of HSV-1 at P2 in 100 μL of PBS. To prevent unintentional viral spread, mock-infected mice were housed separately from virus-infected mice and always handled before MCMV mice, and their cages were changed on different days than MCMV-infected mice. Mice were genotyped on P19 (18) and genotypes were reconfirmed. Mice were sacrificed after tumors became 1.6 cm in diameter or moribund. Tumors appeared throughout the body, although they favored the abdominal cavity or limb girdles (Supplementary Table S1A).

After anesthesia with i.p. injection of ketamine (100 mg/kg) and xylazine (20 mg/kg), 0.5 mL of blood was obtained via the facial vein. Mice were perfused via the intracardiac route with ice-cold PBS followed by organ harvest. Genomic DNA was obtained using a DNeasy Kit and RNA via RNeasy Kit (Qiagen) with on-column DNase treatment. MCMV-GB gene was amplified by PCR using the following primers (5′–3′): MCMV-GB forward: TGGGTGAGAACAACGAGAT and MCMV-GB reverse: CGCAGTCTCCCTTCGAGTA. β-Actin was used as control (mouse β-actin forward: AGCCTCGTCCCGTAGACAAAT; mouse β-actin reverse: GAAGACACCAGTAGACTCCACGACAT). For reverse transcriptase (RT)-PCR experiments, 1 μg of total RNA was transcribed using a SuperScript First-Strand cDNA Synthesis Kit (Invitrogen). Primers (5′–3′) for MCMV-IE1 were forward: TAGCCAATGATATCTTCGAGCG and reverse: TTAGCTGATCTGAGGAGCACCAGAT.

Initially visible tumors reached a large size within 1 week. After anesthesia, mice were examined for visible pathology. Tumor tissue or other tissues of interest were snap frozen at −80°C or placed in 4% paraformaldehyde, processed for paraffin embedding, and cut into 4-μm sections. A board-certified pathologist (O.H. Iwenofu) specializing in sarcomas was blinded to infection status and evaluated hematoxylin and eosin stained sections. In addition, immunohistochemistry was conducted. Antibodies were against MyoD1 (Santa Cruz Biotechnology, Inc.), Desmin (Sigma), smooth muscle actin (Millipore), and Croma101 (from Stipan Jonjic). MOM kit (Vector) was used for antibodies raised in mouse. Microscopic images were captured with an Olympus BX51 microscope (Olympus America Inc.) equipped with DP72 camera.

Human RMS and normal muscle biopsy samples were obtained using an approved Institutional Review Board protocol. Immunohistochemistry for HCMV in tumors has been previously described with minor modifications (8). HCMV-IE1 (Millipore), HCMV-pp65 (Leica Microsystems), or smooth muscle actin (Millipore) primary antibodies were incubated on sections overnight at 4°C. Secondary antibody was applied for 30 minutes followed by biotin link for 30 minutes (Biogenex). Positive staining was visualized with a 3,3′-diaminobenzidine (DAB) substrate.

Mice were infected with MCMV (10³ plaque-forming units) 2 days after birth (P2) or at 4 weeks of age (10⁶ plaque-forming units). Separate Trp53^+/− control cohorts were inoculated with PBS vehicle (mock infection) or with a nonlethal dose of another herpes virus, herpes simplex virus 1 (HSV-1), at the same time points, to test for nonspecific viral effects on tumor formation. Virus infection caused no distinguishing behavioral or clinical effects in mice. Because Trp53^+/− mice start developing tumors at a late age (>9 months), we followed mice for 9 months to detect MCMV-specific effects on tumor development. P2 MCMV infection was associated with a statistically significant decrease in survival (the point of veterinary need for euthanasia) compared with mock-infected mice (P < 0.05, log-rank test; Fig. 1A). About 42.8% (12 of 28) of P2 MCMV–infected Trp53^+/− mice developed large, visible tumors over 9 months (Fig. 1B). Instead, only 1 of 27 mock-infected and 1 of 12 HSV-1-infected mice developed a visible tumor during this time period. In addition, MCMV infection at 4 weeks of age did not result in visible tumors. Therefore, MCMV (but not HSV-1 or mock) infection of Trp53^+/− mice at P2, but not at 4 weeks, led to a significantly higher incidence of visible tumors over the first 9 months of animal life.

Next, we examined mice for MCMV distribution. Tissues from mice, infected with MCMV at P2 and analyzed 5 days postinfection, were positive for MCMV immediate, early (MCMV-IE1), and late (MCMV-GB) transcripts (Fig. 1C). In addition, quadriceps exhibited expression of MCMV-IE1 protein (Fig. 1D). MCMV RNA expression in muscle was not detected 2 weeks after infection (data not shown). In addition, tissues from mice infected at 4 weeks of age were positive for MCMV-specific glycoprotein B DNA (MCMV-GB; Supplementary Fig. S1A) and for MCMV-IE1 and GB transcripts (Supplementary Fig. S1B). Mock-infected mice were negative for MCMV-GB (data not shown). These findings confirm that MCMV infects the tissues from which pleomorphic RMS originate.

Tumors that developed in MCMV-infected Trp53^+/− mice revealed diagnostic features of pleomorphic RMS including pleomorphic features (Fig. 2A), rhabdomyoblasts (Fig. 2B), and infiltration into surrounding muscle (Fig. 2C). The immunophenotypic expression of MyoD1 and Desmin confirmed RMS (Fig. 2D). Analysis of a subset of MCMV-infected tumors (n = 4) revealed MCMV DNA (Fig. 3A), which was absent in the single tumor that had developed in one of the mock-infected Trp53^+/− mice. MCMV-IE1 transcripts were also present in the 4 tested tumors (Fig. 3B). Fidelity of PCR reactions was confirmed by DNA sequencing of PCR products (Fig. 3C). Furthermore, MCMV-infected tumors expressed IE1 protein, whereas this was not seen in the mock-infected tumor (Fig. 3D). Isotype control and no primary antibody confirmed immunohistochemical specificity (Supplementary Fig. S2A). Taken together, these data show transcriptionally active MCMV within tested RMS.

Published data for uninfected Trp53^+/− mice show variable tumor types occurring after the age of 9 months. In one study, 28% (28 of 100) of Trp53^+/− mice develop tumors by 17 months. These tumors were primarily lymphomas, osteosarcomas, and fibrosarcomas; only one was an RMS (14). Pathologic analyses of MCMV-infected tumors in this study revealed that 84.6% were RMS (10 of 12), whereas 15.4% (2 of 12) were lymphomas. Nine of 10 RMS tumors were pleomorphic (Supplementary Table S1A). Therefore, our data suggest that MCMV accelerates tumor formation in mice with heterozygous Trp53 mutation, leading preferentially to pleomorphic RMS.

Because the majority of tumors from MCMV-infected mice were pleomorphic RMS, we tested whether this finding was clinically relevant by examining human tumors for the presence of HCMV (Supplementary Table S1B). Pleomorphic RMS cases were first examined for HCMV-IE1 via immunohistochemistry. All human pleomorphic RMS cases (8 of 8) stained strongly and diffusely for HCMV-IE1 (Fig. 4A), with expected isotype and negative control immunohistochemistry (Supplementary Fig. S2B). As a control, HCMV-IE1 immunoreactivity was not observed in 5 of 5 normal human adult muscle (Fig. 4A). Staining was expanded to include embryonal and alveolar (collectively referred to as pediatric) RMS tumors (Fig. 4A). Nine of 10 (90%) of these tumors stained positive for HCMV-IE1 (Fig. 4B). Because HCMV-IE1 transcription can frequently be detected even after the virus becomes latent, we also stained for the presence of HCMV-pp65, a late protein expressed during active viral replication. All adult pleomorphic RMS cases and most (6 of 10) pediatric RMS tumors stained strongly for pp65 (Fig. 4B), and pp65-positive cells were abundant throughout the positive tumors (Fig. 4A). Altogether, 100% (8 of 8) of adult and 90% (9 of 10) of pediatric RMS cases expressed HCMV antigen (Fig. 4B). Only one pediatric case did not stain positive for HCMV-IE1 or HCMV-pp65 (Supplementary Fig. S2C and not shown). Rhabdomyoblasts stained strongly for HCMV antigen (Supplementary Fig. S2D, arrow), suggesting that the RMS-associated primitive cell type was indeed infected with the virus. In several cases, cells in muscle beyond the tumor margin were also positive for HCMV antigen (Supplementary Fig. S2E). One limitation of the presented immunopositive findings in human RMS cases results from our inability to secure serum from these same historical subjects to determine their cytomegalovirus serology. To confirm the selective presence of HCMV DNA in human RMS, quantitative PCR analysis showed that tumors expressed an average of 81.1 copies/ng of HCMV-UL83 and of 22.2 copies/ng of HCMV-UL146 compared with 0 and 0.49 copies/ng in normal muscle, respectively (P < 0.05, Student t test; Supplementary Fig. S3A). The amplified products' nucleotide sequence was greater than 95% identical to that of a clinical strain (Supplementary Fig. S3B).

To further validate the presence of HCMV in RMS, we next tested for the presence of HCMV RNA by using in situ hybridization in a subset of tumors. A commercially available RNA probe targeted to an unspecified early gene of HCMV revealed HCMV RNA in RMS (Fig. 4C). In situ hybridization with a negative control probe or no probe was negative (Fig. 4C and data not shown). Altogether 100% (5 of 5) of RMS cases tested, but none of the normal muscle biopsies, contained HCMV RNA. The one pediatric case that did not express HCMV-IE1 or HCMV-pp65 protein was positive for HCMV via in situ hybridization (data not shown). Taken together, 100% (18 of 18) of tested RMS cases contained HCMV genetic material.

Our data suggest a link between RMS and cytomegalovirus. During the 9 months of our study, wild-type mice infected with MCMV did not develop tumors (data not shown). However, when MCMV infection was introduced into a neonatal mouse with a Trp53^+/− tumor suppressor deficiency, mice develop tumors at a higher rate and earlier onset than mock-infected mice. Interestingly, tumor formation primarily occurred in Trp53^+/− mice infected at P2. In addition, mice preferentially develop pleomorphic RMS as opposed to a wide array of tumors. Molecular analyses reveal the presence of MCMV in these mouse tumors. In addition, examination of human RMS samples revealed that 100% of tumors contain HCMV genetic material. This implicates cytomegalovirus in the pathogenesis of the disease. Cytomegalovirus has never been shown to be a transforming virus. Our data shows for the first time that cytomegalovirus infection combined with Trp53 heterozygosity promotes pleomorphic RMS. Furthermore, the idea that cytomegalovirus can promote tumorigenesis in an organism with genetic aberrations may help to explain the difficulty of epidemiologically linking cytomegalovirus infection with relatively rare cancers.

No potential conflicts of interest were disclosed.

Conception and design: R.L. Price, C.-H. Kwon, C. Cook, E.A. Chiocca

Development of methodology: R.L. Price, L. Harkins, C.-H. Kwon

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.L. Price, K. Bingmer, O.H. Iwenofu, C. Pelloski

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.L. Price, O.H. Iwenofu, C.-H. Kwon, C. Cook, C. Pelloski, E.A. Chiocca

Writing, review, and/or revision of the manuscript: R.L. Price, O.H. Iwenofu, C.-H. Kwon, C. Cook, C. Pelloski, E.A. Chiocca

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.L. Price, L. Harkins, E.A. Chiocca

Study supervision: C.-H. Kwon, E.A. Chiocca

The authors thank Ulrich Koszinowski (Max von Pettenkofer-Institute, Munich, Germany) for kindly providing MCMV and Stipan Jonjic (University of Rijeka, Rijeka, Croatia) for providing Croma101 antibody. Denis Guttridge (Ohio State University, Columbus, OH) provided constructive guidance to the project.

R.L. Price was funded by a seed grant from the American Medical Association Foundation. This study was funded by a Viral Oncology Program Grant, the Dardinger Neuro-oncology Fund, the Jeffrey Thomas Hayden Foundation (E.A. Chiocca), OSU MBCG Program grant (E.A. Chiocca and C.-H. Kwon), and OSU CCC Start-up Fund (C.-H. Kwon).