Antibody specific to muscle actins in the diagnosis and classification of soft tissue tumors

A series of soft tissue tumors, melanomas, carcinomas, and lymphomas were studied immunohistochemically for the presence of muscle actins (MA) with the monoclonal antibody HHF-35, and for the presence of desmin for comparison. In nonneoplastic tissues, MA immunoreactivity was present in skeletal and smooth muscle cells, in the pericytes of small vessels, and in the myoepithelial cells. Desmin immunoreactivity had a similar distribution, except that the pericytes of small vessels and myoepithelial cells were negative. All 17 rhabdomyosarcomas were positive for both MA and desmin. Of leiomyosarcomas, 31/32 were positive for MA, and 29/32 for desmin. In pleomorphic undifferentiated sarcomas (malignant fibrous histiocytomas) MA and desmin-positive cells were present in 9/35 and 5/35 cases, respectively. Three of five pleomorphic liposarcomas showed MA-positive tumor cells, which were also desmin-positive in one case. Desmoid tumors often showed a moderate number of both desmin- and MA-positive cells. Hemangiopericytoma, Kaposi's sarcoma, and endometrial stromal sarcoma showed MA-positive staining only in the pericytes and not in the neoplastic cells. In various types of carcinomas, melanomas, and lymphomas, MA- or desmin-positive neoplastic cells were not identified. MA, but not desmin, was present in the desmoplastic stroma in many carcinomas. Both MA and desmin are good markers for muscle differentiation and especially serve to identify rhabdomyosarcomas and leiomyosarcomas. These markers are also present in some sarcomas currently regarded as nonmuscle tumors. This may suggest that some of these tumors have differentiation properties related to true myosarcomas. The absence of muscle actin, a pericytic marker, in hemangiopericytoma does not confirm the concept of pericytic nature of this tumor.     

Expression of muscle-specific actins and myosin in light microscopically undifferentiated small and dark cell malignancies of soft tissues

Eighteen light microscopically undifferentiated small and dark cell malignancies, previously studied ultrastructurally and immunohistochemically in terms of desmin, vimentin and myoglobin expression, were analyzed using mono- and polyclonal antibodies to muscle-specific isoforms of myosin and actin. For comparison, 10 characteristic rhabdomyosarcomas, 5 alveolar and 5 embryonal, were included in the study. A polyclonal antibody to skeletal myosin produced an indistinct staining in all the alveolar and embryonal rhabdomyosarcomas. The staining was most prominent in well-differentiated rhabdomyoblastic tumor cells. In the analysis of the small and dark cell malignancies, this antibody produced a weak and indistinct positivity which can not be interpreted with certainty as an expression of muscle-specific properties. An antibody directed at the alpha and gamma isoforms of actin, which are present in smooth and striated muscle, produced a distinct positive staining in all the alveolar and embryonal rhabdomyosarcomas and in 8/18 small and dark cell malignancies, 7 of which were also shown to express desmin. An antibody directed at the alfa-smooth muscle isoform of actin did not produce any positive staining in any of the tumors. The present study indicates that both muscle-specific actin and desmin can be expressed in tumors lacking ultrastructural evidence of a rhabdomyoblastic differentiation and that the combined use of monoclonal antibodies to desmin and muscle-specific actin is of value when it comes to recognizing rhabdomyosarcomas within the group of undifferentiated small and dark cell malignancies of soft tissue tumors.     

Passive stiffness changes in soleus muscles from desmin knockout mice are not due to titin modifications

Passive stiffness was found to be increased in mouse soleus muscles lacking desmin. Because titin is considered to be the major source of muscle elasticity, the stiffening might be explainable by titin adaptation. To test this, passive mechanical properties of single skinned fibres of soleus muscles from desmin knockout and control mice were analysed by using various extension tests. Titin expression was studied by SDS-gel electrophoresis. Absence of desmin did not modify either electrophoretic mobility of the titin band (3700 kDa) or optical density-unit ratios between bands for titin and nebulin (congruent with 0.3) and bands for titin and myosin heavy chain (congruent with 0.08). Elastic properties of fibres were not altered in the absence of desmin since passive tensions were similar under quasi-static (56-66 kN m(-2)) and dynamic (100-118 kN m(-2)) conditions whatever the kind of fibre. Thus, titin is unlikely to be responsible for the large increase in passive stiffness observed in whole soleus muscles when desmin is lacking.     

Sarcoplasmic reticulum (SR) Ca2(+)-ATPase as a marker of muscle cell differentiation: immunohistochemical investigations of rhabdomyosarcomas and enhancement of the immunostaining after sodium methoxide pretreatment.

An immunocytochemical investigation of sarcoplasmic reticulum (SR) Ca2(+)-ATPase (SR-Ca-ATPase) was performed on formalin-fixed paraffin-embedded specimens of different types of rhabdomyosarcomas such as variants of embryonal and pleomorphic forms. Immunostaining frequency of tumours using SR-Ca-ATPase was compared with that of traditionally used muscle specific markers myoglobin, and desmin. Utilizing the possible cleaving of ester bounds sodium methoxide pretreatment was found to be very effective in enhancement of SR-Ca-ATPase immunostaining reaction. In 11 of 15 tissue specimens of 5 cases round shaped and elongated rhabdomyoblasts with definite cytoplasm exhibited positive immunoreactions with all of the polyclonal antibodies tested, using the streptavidin-biotinylated peroxidase complex (S-ABC-method). In formalin-fixed and paraffin-embedded material of 2 cases of undifferentiated rhabdomyosarcomas composed of small round tumour cells with scanty cytoplasm pretreatment with sodium methoxide induced the immunostaining of SR-Ca-ATPase. After that pretreatment a staining of the paranuclear cytoplasm occurred in many of these undifferentiated tumour cells. In these 2 cases, neither myoglobin nor desmin antibodies could react. However, when frozen sections of one of the poorly differentiated tumours were used monoclonal and polyclonal desmin antibodies reacted immunocytochemically in all of the small cells. Sodium methoxide induced or enhanced SR-Ca-ATPase immunocytochemical reaction can be a further addition to the diagnosis of rhabdomyosarcomas in formalin-fixed paraffin-embedded sections, even when desmin antibody fails to react.     

Myofibrillogenesis in primary tissue cultures of adult human skeletal muscle: expression of desmin,titin, and nebulin

To investigate the in vitro development of myofibrils in skeletal muscle cells derived from adult human muscle biopsies, immunohistochemical analysis was performed using monoclonal antibodies against desmin, titin, and nebulin. Diffuse desmin reactivity was detected 48 h after plating in about 60% of all mononucleated cells. This supports the use of desmin as a marker for undifferentiated rhabdomyosarcomas in man. Titin was visible from day 4 onwards, while nebulin was not found in mononucleated cells. After 1 week polynucleated myotubes appeared, and grew up to 30 days. Desmin was distributed diffusely throughout the cytoplasm until day 21, when the pattern became patchy. Titin began to be organized in a predominantly longitudinal orientation at day 15, while nebulin, which appeared for the first time in fusing myoblasts on the fifth to the seventh day, was almost immediately organized in a dotted longitudinal pattern, which became a Z line connected striation in matured myotubes.Abbreviations FCS fetal calf serum - mAb monoclonal antibody - TBS Tris-buffered saline Correspondence to: T. Behr     

Proteins of intermediate filaments. An immunohistochemical and biochemical approach to the classification of soft tissue tumors

The intermediate filament cytoskeleton of various types of human soft tissue tumors was analyzed by immunofluorescence microscopy with the use of specific antibodies against cytokeratins, vimentin, and desmin, as well as by one- and two-dimensional gel electrophoresis of high-salt buffer- and detergent-resistant cytoskeletal preparations. All leiomyomas as well as a leiomyosarcoma contained desmin. Leiomyomas of both gastrointestinal and uterine derivation and the retroperitoneal leiomyosarcoma showed strong reaction for desmin in the smooth muscle cells, but the latter two exhibited also vimentin staining. In embryonal rhabdomyosarcomas, desmin prevailed in the large, apparently well-differentiated rhabdomyoblasts; whereas the smaller, less differentiated tumor cells preferentially contained vimentin. Cells of malignant fibrous histiocytomas were characterized by their content of vimentin as the only intermediate filament protein present. In alveolar soft part sarcoma, a rare tumor of hitherto unknown histogenesis, vimentin and desmin co-existed within the same tumor cells, indicating, together with chemical determinations, the myogenic derivation of this neoplasm. The results show that immunologic and biochemical analysis of proteins associated with the intermediate filament cytoskeleton is a useful adjunct in the diagnosis of diverse neoplasms, particularly those with equivocal histologic features, and thus aids in the histogenetic classification of soft tissue tumors.     

IMMUNOHISTOCHEMISTRY OF MYOGLOBIN IN RHABDOMYOSARCOMAS

Immunoperoxidase staining for myoglobin was performed on formalin-fixed, paraffin-embedded tissues of normal human skeletal muscle and rhabdomyosarcomas in order to assess its potential diagnostic value. Myoglobin-positive cells were observed in normal adult and fetal muscles, and in 13 of 26 rhabdomyosarcomas. In normal skeletal muscle, the intensity of staining for myoglobin varied in individual cells. Of the 13 myoglobin-positive rhabdomyosarcomas tested, 6 showed cross striations in tumor cells, whereas the others did not. Myoglobin-positive tumor cells were not demonstrated in two rhabdomyosarcomas in which cross striations were observed. In other types of tumors examined that sometimes resembled rhabdomyosarcomas, no myoglobin was detected. Thus tests for myoglobin by the immunoperoxidase method seem helpful in identifying rhabdomyosarcomas and confirming their histological diagnosis.     

Rhabdomyosarcoma in childhood. An immunohistological analysis with myoglobin, desmin and vimentin

A retrospective immunohistological analysis of 64 rhabdomyosarcomas in children was performed using antibodies against desmin and in 35 cases against myoglobin. In addition a group of 12 undifferentiated tumours in which the differential diagnosis included rhabdomyosarcomas was studied. Rhabdomyosarcomas were desmin positive in 57 cases (89%), 28 cases of which showed positivity of undifferentiated small cells (44%). Myoglobin was positive in 23 cases (66%), but only one case showed positivity of undifferentiated small cells. The results show the limited use of myoglobin in the diagnosis of rhabdomyosarcoma, especially of cases with a low degree of differentiation. Three out of 12 undifferentiated tumours were desmin positive and were reclassified as rhabdomyosarcomas. In 49 rhabdomyosarcomas the investigation was complemented by the demonstration of vimentin. Vimentin was shown to be present in 27 cases in tumour cells (55%). Undifferentiated cells were positive in 26 tumours (53%) and rhabdomyoblasts reacted in 9 cases (18%). Coexpression of vimentin and desmin in some cases reflects a situation in rhabdomyosarcomas that aberrantly mimics skeletal muscle embryogenesis. In three cases desmin and vimentin positive globular inclusions were observed. It is suggested that their formation is related to dystrophic changes of contractile and cytoskeletal filaments. From the diagnostic point of view a high percentage of desmin positive cases makes desmin a successful marker for rhabdomyoblastic tumours. It is pointed out, however, that even immunohistochemistry may not contribute to solving the problem of undifferentiated tumours and that each case must be evaluated comprehensively.     

Immunohistochemistry of myoglobin in rhabdomyosarcomas

Immunoperoxidase staining for myoglobin was performed on formalin-fixed, paraffin-embedded tissues of normal human skeletal muscle and rhabdomyosarcomas in order to assess its potential diagnostic value. Myoglobin-positive cells were observed in normal adult and fetal muscles, and in 13 of 26 rhabdomyosarcomas. In normal skeletal muscle, the intensity of staining for myoglobin varied in individual cells. Of the 13 myoglobin-positive rhabdomyosarcomas tested, 6 showed cross striations in tumor cells, whereas the others did not. Myoglobin-positive tumor cells were not demonstrated in two rhabdomyosarcomas in which cross striations were observed. In other types of tumors examined that sometimes resembled rhabdomyosarcomas, no myoglobin was detected. Thus tests for myoglobin by the immunoperoxidase method seem helpful in identifying rhabdomyosarcomas and confirming their histological diagnosis.     

Desmin expression in rhabdomyosarcoma: influence of the desmin clone and immunohistochemical method.

AIM--To determine which, if any, of five commercially available desmin clones is most reliable at labelling desmin filaments and whether the enhanced polymer one step (EPOS) method of labelling is of any advantage in the routine diagnostic laboratory. METHODS--Thirty four rhabdomyosarcomas from the files at The Hospital for Sick Children, Great Ormond Street, London, were studied. Four different desmin clones, DE-R-11, D33, DE-U-10, and PDE, were applied to each using the conventional extravidin biotin peroxidase method. The D33 clone was also applied using the EPOS method. RESULTS--The EPOS method incorporating D33 persistently scored more cells as desmin positive and was positive in four cases which were negative on staining with the other clones. CONCLUSIONS--The D33 desmin clone used with the EPOS method is more reliable for identifying desmin filaments in tumours than other desmin antibodies tested. Different desmin clones using a routine technique label different rhabdomyosarcoma cells and therefore it is justifiable to use more than one clone.     

Lectin histochemistry of experimental murine rhabdomyosarcomas

Methylcholanthrene-induced murine rhabdomyosarcomas and skeletal muscle of 10 and 18 d old murine embryos were investigated by lectin histochemistry (WGA, RCA-I, LCA, Con-A, PSA, UEA-I, PNA) and by immunohistochemistry (vimentin, desmin, myoglobulin). In rhabdomyosarcomas as well as in the developing skeletal muscle a clear trend was visible. A decrease of vimentin positivity and an increase of desmin positivity were associated with a diminution of binding sites for WGA, RCA, and LCA. No binding moieties for these lectin could be demonstrated in myoglobin positive normal and neoplastic rhabdomyomatous cells at all. The homologous expression or absence of markers reflected the cellular variability in rhabdomyosarcomas and may be explained as a phenomenon of different tumor cell maturation. The results show that rhabdomyosarcomatous cells are imitating the normal skeletal muscle development.     

Expression of desmin and myoglobin in rhabdomyosarcomas and in developing skeletal muscle

Immunohistochemical staining for desmin and myoglobin was investigated in 35 rhabdomyosarcomas from young people and in skeletal muscle from 16 human fetuses of known gestational age. Twenty-nine of the rhabdomyosarcomas expressed desmin but six undifferentiated or poorly-differentiated tumours were desmin negative. Of the desmin positive cases, most undifferentiated or poorly-differentiated sarcomas expressed desmin alone (12/35). Tumours with increasing rhabdomyoblastomic differentiation co-expressed myoglobin (9/35) and well-differentiated examples also contained cross-striations (7/35). Skeletal muscle from fetuses aged 8 weeks or less consisted mainly of primitive desmin negative round cells. As the cells began to differentiate they quickly expressed desmin and, at approximately 10 weeks, myoglobin was expressed and cross-striations were seen. The combined results strengthen the view that desmin (within a strictly defined context of round cell tumours in young people) is a reliable marker for rhabdomyoblastic differentiation. Support is also given to the notion that very primitive rhabdomyosarcomas may be desmin-negative, although the difficulties of establishing firm diagnoses for some of these tumours is emphasized.     

Titin expression as an early indication of heart and skeletal muscle differentiation in vitro. Developmental re-organisation in relation to cytoskeletal constituents

Summary Established myogenic cell lines of different species and tissue origin have been used to study expression and organisation of muscle-specific proteins during differentiation. Furthermore, primary cultures of rat myocard cells were used to examine these same processes during dedifferentiation. In particular, we were interested in the general mechanism that underlies the changes in the supramolecular organisation of titin during in vitro myogenesis. It became obvious that in the differentiating muscle cell cultures the redistribution of desmin, actin and myosin in a typical, differentiation state dependent fashion, always showed a certain delay when compared to titin. The sequence of changes in the assembly of cytoskeletal and sarcomeric structures observed during differentiation of the cell lines was reversed during the process of dedifferentiation in cultured rat myocard cells. These results all indicate that titin is an early marker of myogenic differentiation, both in vivo and in vitro, and that the typical reorganisation of this giant molecule is independent of species or muscle cell type.     

Mesenchymal and muscle-specific intermediate filaments (vimentin and desmin) in relation to differentiation in childhood rhabdomyosarcomas

Twenty-one childhood rhabdomyosarcomas were divided into three groups on the basis of cytologic composition. The tumors in group P consisted entirely of primitive mesenchymal cells, whereas those in groups M and W were characterized by the additional presence of numerous round rhabdomyoblasts and strap cells, respectively. The tumors were studied for the universal mesenchymal intermediate filament vimentin, and for the muscle-specific intermediate filament desmin. Vimentin positivity, which tended to be more prominent in primitive tumor cells, was found in all tumors, whereas desmin was found especially in round rhabdomyoblasts and strap cells. Desmin-positive primitive cells were found only in groups M and W, not in group P. It was concluded that the differentiation from primitive mesenchymal cells to morphologically recognizable myogenic tumor cells is accompanied by an increase in desmin positivity and, presumably, a decrease in vimentin positivity. Moreover, the observations suggest the existence of a group of "committed" cells that are morphologically primitive, but desmin-positive. These cells might play an important role in the observed further differentiation of rhabdomyosarcomas under chemotherapy.     

Cellular origin and differentiation of renal carcinomas. A fluorescence microscopic study with kidney-specific antibodies, antiintermediate filament antibodies, and lectins

Frozen sections of human renal carcinomas were studied in indirect immunofluorescence using antibodies against intermediate filaments of cytokeratin, desmin and vimentin type, and against proximal tubular brush border and distal tubular Tamm-Horsfall glycoprotein antigens, as well as with fluorochrome-labeled lectins in an attempt to study the origin and stage of differentiation of renal carcinomas. Eighty per cent of the renal carcinomas expressed the brush border antigens, whereas the Tamm-Horsfall glycoprotein could not be found. Antibodies against epidermal cytokeratins reacted only with collecting ducts in normal kidney, whereas antibodies against cytokeratins of Madin-Darby canine kidney epithelial cell line also reacted with glomerular and tubular epithelium. In 93% of the carcinomas tumor cells showed reactivity with both types of antikeratin antibodies. Vimentin, the cytoskeletal protein of mesenchymal cells, was present in the carcinoma cells of 53% of the tumors, although it was not present in normal tubular epithelium. Moreover, vimentin was expressed together with cytokeratin in the carcinoma cells in 57% of the keratin-positive samples as judged by double immunostaining, whereas the muscle type of intermediate filament protein, desmin, was not seen in the malignant cells. Binding sites for Lotus tetragonolobus agglutinin and soybean agglutinin, normally present in the cells of proximal tubules, were lacking or only faintly detectable in the neoplastic cells. Dolichos biflorus agglutinin, normally present in collecting ducts, was not detected in the tumors. The results show that most renal carcinomas express cytokeratin antigens as a sign of their epithelial origin and also show characteristics of proximal tubular cells. On the other hand, the results indicate that lectin-binding sites typical for normal differentiated tubular cells are profoundly modified in renal carcinomas. Ulex europaeus agglutinin did not bind to the malignant cells but decorated the endothelial cells of the tumors.     

Immunohistochemical distinction of human carcinomas by cytokeratin typing with monoclonal antibodies.

Carcinomas of different origin have been tested in immunofluorescence microscopy with the monoclonal murine antibodies CK1-CK4, which recognize a single cytokeratin polypeptide (human cytokeratin No. 18) present in simple but not in stratified squamous epithelia, and with the monoclonal antibody KG8.13 and guinea pig kerA antibodies, both of which recognize a variety of cytokeratins common to almost all epithelial cell types. Tumors derived from simple epithelia, including adenocarcinomas and some other tumors such as ductal breast carcinomas, were strongly stained by all three antibodies. So was a transitional carcinoma of the bladder. In contrast, basal cell epithelioma, cloacogenic carcinoma, and squamous cell carcinoma of skin, tongue, and esophagus appeared negative with CK1-CK4 but positive with the other two antibodies. Other squamous cell carcinomas derived from epiglottis and cervix uteri showed a mixture of positive and negative cells when tested with CK1-CK4, although all tumor cells were positive when tested with KG8.13 and with kerA. Thus, use of an appropriate collection of cytokeratin antibodies with different specificities not only allows tumors of epithelial origin to be distinguished from other tumor types but, in addition, allows a further subdivision of carcinomas in relation to their histologic origin.     

Diagnostic value of intermediate filament antibodies in clinical cytology

Summary Antibodies to intermediate filament (IF) proteins can distinguish the major tumor groups as shown by results with sectioned human material. In this study we evaluate the use of similar methods in the cytology of human tumors. Smears obtained from fine needle aspiration biopsies were investigated using well characterized antibodies, each specific for only one of the five types of intermediate filaments. Tumor cells of different carcinomas, thymomas, and the epithelial part of pulmonary blastomas were positive with antibodies recognizing cytokeratins. Tumor cells in non-muscle sarcomas, including lymphoma and Ewing's sarcoma, could be specifically identified with antibodies to vimentin. Tumor cells of muscle sarcomas were desmin-positive. Finally, tumor cells in pheochromocytoma and bronchus carcinoid were positive with antibodies specific for neurofilaments. Specimens were also examined in parallel using conventional cytochemical stains, such as May-Grünwald-Giemsa. In addition, in most cases sections of the tumor were examined both by histology and IF typing of frozen sections to confirm the diagnosis made on the cytologic specimens. The results show that IF typing is a valuable diagnostic aid in clinical cytology.     

Immunophenotypic diagnosis of leiomyosarcomas and rhabdomyosarcomas with monoclonal antibodies to muscle-specific actin and desmin in formalin-fixed tissue

Two fibrillary proteins, muscle-specific actin (MSA) and desmin, are found only in cells of smooth and skeletal muscle lineages. Among the monoclonal antibodies (MAbs) to these antigens which we have tested, we found several to be reactive in formalin-fixed, paraffin-embedded sections. This finding widened the possibility of using these MAbs in routine diagnostic surgical pathology for the immunodiagnosis of rhabdomyosarcomas (RMS) and leiomyosarcomas (LMS). We therefore conducted a comparative study of three such MAbs which are available commercially and which we applied to paraffin-embedded, formalin-fixed tissues from 15 patients with RMS and 19 patients with LMS. The case selection criteria included typical light-microscopic appearances as well as immunoreactivity with at least one of the MAbs. MSA was detected in all cases of RMS and LMS, whereas desmin was reactive in 12 of 13 RMS and 10 of the 19 LMS. (Desmin antigenicity was judged to be lost in two RMS, since the vascular smooth-muscle tissue present in the specimens failed to react with these antibodies.) In LMS, desmin tended to show focal positivity, whereas the MSA in the same specimens was diffusely positive. These results demonstrate the utility of MAbs for confirmation of the muscle lineage of LMS and RMS in formalin-fixed, paraffin-embedded tissue. The results also indicate that, with the MAbs tested, the antigenicity of MSA is preserved more consistently than that of desmin in formalin-fixed, paraffin-embedded tissue, and that MSA is a more sensitive marker for the detection of muscle differentiation than is desmin, especially in LMS.     

Inherited cardiomyopathies as a troponin disease

Troponin, one of the sarcomeric proteins, plays a central role in the Ca(2+) regulation of contraction in vertebrate skeletal and cardiac muscles. It consists of three subunits with distinct structure and function, troponin T, troponin I, and troponin C, and their accurate and complex intermolecular interaction in response to the rapid rise and fall of Ca(2+) in cardiomyocytes plays a key role in maintaining the normal cardiac pump function. More than 200 mutations in the cardiac sarcomeric proteins, including myosin heavy and light chains, actin, troponin, tropomyosin, myosin-binding protein-C, and titin/connectin, have been found to cause various types of cardiomyopathy in human since 1990, and more than 60 mutations in human cardiac troponin subunits have been identified in dilated, hypertrophic, and restrictive forms of cardiomyopathy. In this review, we have focused on the mutations in the genes for human cardiac troponin subunits and discussed their functional consequences that might be involved in the primary mechanisms for the pathogenesis of these different types of cardiomyopathy.     

Expression of cardiac ankyrin repeat protein, CARP, in malignant tumors: diagnostic use of CARP protein immunostaining in rhabdomyosarcoma

Cardiac ankyrin repeat protein (CARP) is highly expressed in cardiac muscles and detectable in normal skeletal muscles. Arpp, a close homolog of CARP, has been demonstrated to be useful for distinguishing rhabdomyosarcoma from other malignant tumors. However, the CARP distributions among malignant tumors have been poorly investigated. Here, we analyzed the comprehensive expression of CARP in malignant tumors and evaluated its potential use for rhabdomyosarcoma diagnosis. A total of 159 malignant tumors, including 34 rhabdomyosarcomas, 85 non-rhabdomyosarcomas, and 40 carcinomas, were immunohistochemically analyzed for CARP expression. Cytoplasmic expression of CARP was detected in 29 (85%) of 34 rhabdomyosarcomas. The immunoreactivity was observed in both small cells with little differentiation and differentiated tumor cells with abundant eosinophilic cytoplasm. In contrast, focal immunoreactivity for CARP was only observed in 5 (4%) of 125 non-rhabdomyosarcomas, comprising 2 malignant fibrous histiocytomas, 1 angiosarcoma, 1 epithelioid sarcoma, and 1 squamous cell carcinoma of the lung. Comparative analysis of the CARP expression profiles with those of myogenic markers in rhabdomyosarcomas revealed that myogenin (88%) and desmin (88%) exhibited the best sensitivity, followed by CARP (85%), MyoD (82%), muscle-specific actin (79%), and myoglobin (65%). MyoD (96%) and myoglobin (96%) had the best specificity, followed by CARP (95%), myogenin (95%), desmin (89%), and muscle-specific actin (86%). Our results indicate that CARP is a sensitive and specific marker for rhabdomyosarcoma and that it will be useful for the differential diagnosis of rhabdomyosarcoma.