Rhabdomyosarcoma is a malignant neoplasm of skeletal muscle, and is one of the most common soft tissue malignancies in the children and adolescent population. Histopathologically rhabdomyosarcoma can be divided into two major subtypes, embryonal rhabdomyosarcoma representing about 80% of the clinical cases, and alveolar rhabdomyosarcoma representing about 20% of the total cases 12. Extensive cytogenetic studies showed that the embryonal rhabdomyosarcomas are generally associated with loss of heterozygosity (LOH) for multiple closely linked loci at chromosomal 11p15.5, suggesting inactivation of one or more tumor-suppressor gene(s) in this region 23. But the gene involved in this region has yet to be identified. Many other associated mutation or amplification of oncogenes, tumor suppressor genes including p53, K-ras, N-ras and N-myc, etc. are frequently seen, but the roles of these gene mutations/amplifications in pathogenesis of the tumor remain largely unclear 4;5. The alveolar subtype of rhabdomyosarcoma, however, is consistently associated with two common chromosomal translocations. Most cases of alveolar rhabdomyosarcoma are associated with a chromosomal translocation t(2; 13) (q35; q14), and less commonly with t(1; 13)(p36; q14) 6789. The t(2; 13) translocation results in fusion of a portion of the Pax3 gene to a portion of the FKHR gene, leading to formation of fusion protein Pax3-FKHR. Pax3 gene encodes a transcriptional factor important in early development program, while FKHR gene encodes another generic transcriptional factor that is widely expressed in mammalian tissues. The t(1; 13) translocation results in formation of Pax7-FKHR. Both Pax3-FKHR and Pax7-FKHR were shown to possess more potent transcriptional activity than its original forms Pax3 and Pax7 1011. How the chimeric fusion proteins contribute to pathogenesis of rhabdomyosarcoma remains unclear.

Ataxia telangiectasia mutated gene (ATM) is a newly identified member of PI3 kinase family that is mutated in human autosomal recessive disease, ataxia telangiectasia (AT) 1213. ATM gene is large, spanning about 150 kbs of genomic DNA, and is located on chromosome 11q22-23. In response to DNA damage agents such as ionizing irradiation and chemical agent cisplatin, ATM kinase is activated, leading to a cascade of kinase reactions regulating cell cycle, apoptosis, and DNA damage repair 1415. Many downstream target molecules of ATM kinase were so far identified. These target proteins include the c-Abl tyrosine kinase, the p53 tumor suppressor, the Chk1 and Chk2 serine/threoine kinase, the p34 subunit of replication protein A (RPA), Chk1 serine/threoine kinase and NBS1 kinase which is mutated in human autosomal recessive disease Nijmegan breakage syndrome 1617181920212223. Clinical characteristics of AT patients include hypersensitivity to irradiation, thymic dysplasia, immune defects and higher tendency to develop malignant tumor, suggesting that ATM function is critical in DNA damage repair and tumor suppression. However, up to date, AT patients are not known to develop rhabdomyosarcoma.

In this study, we report an association of human rhabdomyosarcoma with deletion/mutation of ATM gene. This is the first report so far of ATM gene associated with solid malignancy except for the mantle cell lymphoma, a malignant tumor in lymphoid origin. The association of ATM gene mutation/deletion with rhabdomyosarcoma suggests a link of ATM gene with cancer risk, and ATM kinase may contribute to pathogenesis of rhabdomyosarcoma.

Totally 17 patients specimen were retrieved from the archived tissue blocks. The clinical characteristics of these patients were listed as in Table 1. Some patient's clinical information was not available at the time of the study. These patients all carried clinical histopathological diagnoses of rhabdomyosarcoma (embryonal or alveolar type), and showed immunostaining positivity for myogenin, a specific skeletal muscle marker. Immunofluorescent or immunohistochemical staining with anti-ATM antibody was performed on the total of 17 specimens, and 7 of the 17 rhabdomyosarcomas were negative for ATM protein (41%) (see Table 1). Totally 10 of the 17 were positive for ATM kinase (59%). ATM protein was predominantly located in cytoplasm in the tumor sections and interestingly, predominantly seen in the nuclei in the rhabdomyosarcoma cell line, as those seen with myoD staining (RD cells) (Fig 1 and 2). There is no difference in staining pattern in regard to the subtypes of the rhabdomyosarcomas, namely embryonal or alveolar subtypes.

It is surprising to note that 7 of the17 (41%) rhabdomyosarcoma specimens stained negatively for the presence of ATM protein. Under the normal condition, ATM expression is ubiquitous. To further confirm the immunostaining results, two different approaches were taken. One was to see if the ATM mRNA is properly transcribed, and another is to directly see the ATM protein itself by Western blotting analysis from the tumors and compare with the normal form. Due to the limited resource of the fresh tumor samples, we turned to the cell culture system and used the well-established rhabdomyosarcoma cell lines. First, we designed eight pairs of oligonucleotide primers encompassing the almost entire coding sequence of ATM mRNA based on the published cDNA sequence (Genbank accession number U33841)24(Table 2). The primer designing strategy is shown in Fig 3. ATM mRNA is approximately 9.4 kbs in length and contains multiple protein binding domains for its kinase activity and interaction with other cellular proteins (Fig 3). We used reverse transcription-PCR method to analyze five different rhabdomyosarcoma cell lines, RD cells (embryonal), Rh28, Rh30 (alveolar), Rh3, Rh4, and mouse normal myoblast C2C12 cells. The human foreskin fibroblasts were used as the control cells. The mRNA was isolated from these cells and used for RT-PCR assays. As shown in Fig 4, in normal fibroblasts, RD cells, Rh28 cells, the amplified DNA fragments were identical in regard to the lengths in comparing with the control fibroblasts, whereas in Rh30 cells, two fragments in #1 and #7 lanes (0.94 kbs in Lane #1, and 1.3 kbs in Lane #7) were absent, and in lane #6, the fragment amplified was approximately 200 bps smaller than that seen in the fibroblasts and other rhabdomyosarcoma cells (Fig 4). The combined length of missing DNA sequence is estimated 2.4 kb (fragment #1, plus fragment #7, and additional 200 bps in lane #6).

The next question is whether ATM protein is normally translated from ATM mRNA and expressed in the rhabdomyosarcoma cells. We analyzed the ATM protein in different cell lines. In human Hela cells and human normal foreskin fibroblasts, the ATM was estimated to be 360 kD, consistent with the results previously described1213. In three rhabdomyosarcoma cell lines (RD, Rh28 and RH30), there was one aberrant form of ATM protein (~260 kD) that was significantly smaller than that seen in the normal fibroblasts and the Hela cells (Fig 5). The levels of ATM expression in these three rhabdomyosarcoma cell lines are not identical, and Rh30 cells seemed to express less ATM than the other two Rhabdomyosarcoma cells. The murine ATM proteins from C2C12 and 10T 1/2 cells were significantly different in size and there appeared to be two species of ATM proteins in murine cells.

ATM gene is a member of PI3 kinase family. ATM kinase is critical in DNA damage signaling pathway. In response to DNA damage induced by ionizing radiation or chemical agents such as cisplatin, ATM kinase is activated and phosphorylates a number of downstream regulatory proteins important for cell cycle arrest, growth control, apoptosis and DNA repair 25262728. AT patients are known to be predisposed to a variety of cancers, and the fibroblasts of AT patients are hypersensitive to ionizing radiation 29. Although many reports showed the association of ATM mutation with risk of a number of different human malignancies, including prostate cancer, breast cancer, ovarian cancer, B-cell chronic lymphocytic leukemia and mantle cell lymphoma, the underlying molecular mechanism by which ATM gene causes these changes are not well understood 303132333435. There is so far no study showing the association of ATM gene with any solid tumor. There is enormous epidemiological interest in establishing the link of ATM gene abnormality with cancer risk, since the percentage of the ATM heterozygotes is high, and estimated to represent 1.5% of total population.

We studied the immunostaining characteristics of ATM protein in clinical rhabdomyosarcoma cases and found that a significantly high number of cases lacked the proper ATM gene product (41% in 17 cases). The clinical significance of this finding is unclear at this point. There is high percentage of clinical rhabdomyosarcoma cases associated with p53 gene mutation/deletion. The p53 mutant mice tend to develop rhabdomyosarcoma 29. P53 is a tumor suppressor regulating cell cycle control and apoptosis, and p53 is one of the many downstream targets of ATM kinase2527. ATM phosphorylate p53 at Serine 15 in response to DNA damage agents 17. It is unknown how the mutated ATM contributes to pathogenesis of rhabdomyosarcoma in the presence of the mutated p53. In fact, the rhabdomyosarcoma cells Rh30 is known to have p53 mutation, and is also positive for t(2; 13) chromosomal translocation 11. Further investigations are needed to address these important questions.

The finding of ATM protein predominantly present in the cytoplasm, but not the nucleus of the tumor sections suggests a different role of ATM in the tumor cells. In the tumor cell lines, the ATM protein is present predominantly in the nucleus. The underlying significance of this finding is unclear. Presumably ATM protein in the cytoplasm may function differently in the tumor cells in vivo from that in the nucleus seen in the cell culture.

The murine ATM was found significantly different in size, although the ATM knockout mice recapitulate the phenotypic features of A-T patients 36. This species difference may be important in interpreting the data from human cells and from murine cells.

PZ designed and analyzed the data, and performed large portion of the experiments. KB helped in protein extractions and Western blot analyses. PLP, JF and JW participated with extensive discussion of design and experiments. RON reviewed all the patient's tumor staining and patients cases for diagnoses. All authors read the final manuscripts.