- Email: [email protected]
Embryonic myosin heavy chain as a differentiation marker of developing human skeletal muscle and rhabdomyosarcoma
Experimental
Embryonic
Cell Research
163 (1986) 21 l-220
Myosin Heavy Chain as a Differentiation of Developing Human Skeletal Muscle and Rhabdomyosarcoma A Monoclonal
Antibody
Marker
Study
S. SCHIAFFINO,’ L. GORZA,’ S. SARIORE,’ L. SAGGIN’ and M. CARLI ‘Institute
of General University
Pathology of Padova,
and *Department Z-35100 Padua,
of Pediatrics, Ztaly
Hybridoma cell lines were obtained from the fusion of NS-0 myeloma cells with spleen cells of mice immunized with bovine fetal skeletal myosin. A stable hybridoma clone, BFG6, produced immunoglobulin Cl k antibodies reacting specitically with embryonic-type myosin heavy chains present in fetal but not in neonatal or adult human skeletal muscle, as determined by enzyme immunoassay and immunoblot analysis. Fetal but not adult skeletal muscle fibers were stained by this monoclonal antibody in indirect immunofluorescence assays; smooth muscle cells and cardiac muscle cells, as well as non-muscle cells were also unreactive. Solid tumors of infants and children were tested for reactivity with BF-G6 by immunofluorescence and immunoperoxidase staining. Embryonic myosin heavy chain was expressed in rhabdomyosarcomas but not in other types of tumor, except for Wilms’ tumor. Rhabdomyosarcoma cells isolated from a bone marrow metastasis and grown in vitro for several months were also labelled by BF-G6. Embryonic myosin heavy chain can thus be used as a specific differentiation marker of normal and neoplastic skeletal muscle tissue.
@ 1986 Academic
Press, Inc.
Rhabdomyosarcoma is a highly malignant myogenic neoplasm accounting for a major proportion of soft tissue sarcomas in infants and children [l, 21. Different types of this tumor have been described, embryonal and alveolar rhabdomyosarcomas being the most common cytological variants. The distinguishing feature of rhabdomyosarcoma is the presence of striated muscle cells at initial stages of differentiation (rhabdomyoblasts). However, histologic evidence of striated muscle differentiation may be lacking or ambiguous and diagnosis of poorly differentiated rhabdomyosarcoma is often difficult by routine histology [3]. Identification of specific striated muscle markers by immunocytochemical procedures can provide more precise criteria for differential diagnosis. Myoglobin, which has been used as a rhabdomyosarcoma marker [4-61, may lead to false-positive results; lymphoma cells invading skeletal muscle were found to stain for myoglobin, which was presumably released by necrotic muscle fibers and taken up by tumor cells [7]. Desmin is a good marker for rhabdomyosarcoma [8, 93 but does not allow differentiation with leiomyoma and leiomyosarcoma [lo, 111. Antibodies against myosin, used in previous studies on rhabdoCopyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/86 $03.00
212 Schiafjno
et al.
myosarcoma [12, 131, were not characterized with respect to specificity; some of these antibodies were certainly not specific for sarcomeric myosins and stained also breast carcinoma cells [13]. Furthermore, antimyosin antibodies were prepared against adult-type myosins and were thus not appropriate for identifying undifferentiated muscle cells, which are now known to express distinct myosin isoforms . The myosin molecule consists of two heavy and four light chains. In striated muscle there are multiple forms of myosin heavy chains (MHCs): these sarcomerit MHCs are encoded by different genes whose expression is tissue-specific and developmentally regulated [ 141. Developing rat muscle contains embryonictype MHCs that are progressively lost during post-natal development [ 151 but can be re-expressed during muscle regeneration [16]. Embryonic isoforms of MHCs have also been identified in human skeletal muscle [17-191 and evidence has been presented for the existence of an embryonic specific human MHC gene [20]. We describe here a monoclonal antibody which recognizes embryonic MHC in developing human muscle and in rhabdomyosarcoma. MATERIALS
AND
METHODS
Tissue Sources Fetal weeks. months Tumors years).
samples were from 6 fetuses resulting from termination of pregnancy at 8,9, 10, 11 [21 and 20 Post-natal samples were from an I-day-old infant (biopsy specimen), 2 babies aged 8 and 21 (biopsy specimens) and 8 individuals, aged 947 years (autopsy and biopsy specimens). samples were from different solid tumors of infants and children (age range 3 months to 13 All samples were frozen in liquid nitrogen and stored in a freezer at -80°C.
Tissue Culture of Rhabdomyosarcoma
Cells
A bone marrow aspiration from a rhabdomyosarcoma patient with bone marrow metastasis (patient no. 3) was first treated with ice-cold 0.83 % NH&l to eliminate red blood cells. Centrifuged cells were then plated in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10 % horse serum and 1% glutamine. Cells were either grown in plastic culture flasks or, for immunofluorescence studies, on gelatin-coated glass coverslips placed in plastic dishes. After various time intervals coverslips were treated with acetone at -20°C for 5 min and air-dried before processing for immunofluorescence 1211.
Production
of Monoclonal
Antibodies
Myosin was isolated from skeletal muscles of 3-month-old bovine fetus as described by Barany & Close 1221. BALB/c mice were immunized by two intradermal injections (administered two weeks apart) of 100 pg of bovine fetal myosin emulsified with complete Freund adjuvant. Two weeks after the second injection the animals were bled and the sera were screened for the presence of antimyosin antibodies by enzyme immunoassay and immunofluorescence. Animals showing the highest antibody titres were selected and one month later were injected intraperitoneally with 50 pg of myosin on one day and the same dose intravenously on the following day. Three days after the last injection spleen cells of the mice were fused with mouse myeloma cell line NS-0 using standard techniques 1231. The hybridomas were grown in microplates and were screened for production of antimyosin antibodies by indirect immunofluorescence on cryosections of fetal and adult skeletal muscle. Selected hybridomas were cloned twice by the limiting dilution method. The hybridoma line used for this study, called BFG6, was found to produce IgGl k antibody, as determined by enzyme immunoassay with antisera specific for the major mouse immunoglobulin classes and for the subclasses of IgG and light chains; culture supematant was used for immunochemical assays and immunofluorescence tests. Exp
Ceil
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
213
Fig. 1. Reactivity of monoclonal antibody BF-G6 with myosin isolated from 0, lOweek human fetal muscle; n , I-day newborn muscle; 0, adult skeletal muscle, as determined by solid-phase enzyme immunoassay. 0.3 Myosin
3
30
(pglmll
Enzyme Immunoassay Solid-phase enzyme immunoassays were performed on microplates following the procedure of Engvall & Perlman [24]. Microplate wells coated with different concentrations of fetal or adult muscle myosin were incubated with appropriate dilutions of monoclonal supematant (usually 1: 100 to 1 : 200). Bound antibody was revealed by rabbit anti-mouse IgG conjugated with horseradish peroxidase (Miles Lab.); enzyme concentration was determined with o-dianisidine as substrate and measuring the absorbance at 410 nm.
Zmmunoblotting SDS-PAGE of crude myosin preparations [25] from human fetal and adult skeletal muscle was followed by electrophoretic transfer to nitrocellulose paper as described by Towbin et al. [26]. Bound antibody was revealed by immunoperoxidase staining, using rabbit anti-mouse IgG conjugated with horseradish peroxidase and diaminobenzidine as substrate in the presence of imidazole [27].
Zmmunojluorescence
and Immunoperoxidase
Staining
Indirect immunofluorescence tests were performed on fresh frozen sections of fetal and adult human muscle and tumor biopsies, or acetone-fixed cultures of rhabdomyosarcoma. After incubation with appropriate dilutions of monoclonal supematant (usually 1: 8-l : 16) for 30 mitt at 37’C, sections were washed and incubated with fluorescein-labelled rabbit anti-mouse IgG (Miles Lab.) for 30 min at 37°C. Sections were then post-fared in 1.5 % paraformaldehyde in phosphate-buffered saline (PBS) and examined under epitluorescence illumination on a Leitz Dialux microscope. Immunoperoxidase staining was performed following the peroxidase-anti-peroxidase (PAP) technique as described by Stemberger [28].
RESULTS The specificity of monoclonal antibody BF-G6 is illustrated in figs 1-3. BF-G6 reacted with myosin from human fetal muscle and not with neonatal and adult myosin in solid-phase enzyme immunoassay (fig. 1) and reacted exclusively with fetal MHCs on immunoblots (fig. 2). Fetal muscle fibers showed specific immunofluorescent staining with BF-G6 in cryosections of human skeletal muscle (fig. 3). Comparable reactivity with BF-G6 was found in fetuses of 8-11 and 20 weeks by immunofluorescence, except that all fibers were brightly labelled in the Exp
Ceil
Res 163 (1986)
214
Schiaffino
et al.
12.3
4
2oOK-
Fig. 2. Immunoblotting of crude myosin preparations from 2, 4 adult human muscle; I, 3, IO-week fetal muscle with 3, 4, antibody BF-G6; 1, 2, Coomassie Blue-stained gel. Nitrocellulose paper strip reacted with antibody and processed for immunoperoxidase staining. MW markers are shown on the left.
younger fetuses, while a minor proportion of unlabelled fibers were seen in the older fetus. Only few weakly labelled fibers persisted in neonatal muscle and no reactivity was seen in muscle samples from 8- and 21-month-old infants and in adult samples. Vascular smooth muscle and cardiac muscle, as well as nonmuscle tissues were unstained by this antibody at any age. Cryosections from 17 solid neoplasms, including 8 rhabdomyosarcomas, occurring in infants and children were processed for indirect immunofluorescence with monoclonal antibody BF-G6. Reactive cells were detected in all cases of rhabdomyosarcoma, showing variable degree of differentiation, whereas they were absent from other tumors, except for Wilms’ tumor (table 1). Labelled cells were also present in specimens obtained from metastases or local recurrences of rhabdomyosarcoma, except for one specimen from case no. 3 (table 2). In this patient a biopsy from a lymph node metastasis was negative, whereas a second biopsy obtained 5 months later from the bone marrow revealed the presence of
Table 1. Embryonic childhood”
myosin
immunoreactivity
Type of tumor
Age range
Bhabdomyosarcoma Other tumors=
3m13y
1-13 y
in solid tumors of infancy and
No. of positive casesb No. of cases examined 818 l/9
a As determined by immunofluorescent staining with monoclonal antibody BF-G6. ’ Cases showing presence of labelled cells in at least one of the specimens examined. c Fibrosarcoma, neuroblastoma, ganglioneuroblastoma [Z], Ewing’s sarcoma [2], extraskeletal Ewing sarcoma, embryonal carcinoma, Wilms’ tumor. Wilms’ tumor was the only positive case. C-p
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
215
Fig. 3. Indirect immunofluorescence staining of human skeletal muscle with antibody BFG6. Cryosections of skeletal muscle from (a) lO-week fetus; (6) 20-week fetus; (c) E-day newborn; (4 E-month infant. X400.
216
Schiaffino
et al.
Fig. 4. (a) Immunofluorescence; (b) immunoperoxidase staining of cryosections of human rhabdomyosarcoma with antibody BF-G6. The section in (b) has been stained with hematoxylin after the immunoperoxidase reaction: note the presence of both labelled and unlabelled tumor cells. X4Cl&
Exp
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
217
Fig. 5. Immunofluorescence staining with antibody BF-G6 of (a) smear of bone marrow aspiration from child with rhabdomyosarcoma; (b, c) cells from the same metastasis cultured in vitro for 30 days. Occasional labelled cells showed cross-striation (b). x500.
labelled cells. As shown in fig. 4 the frequency of labelled cells varied from specimen to specimen and in different regions of the same specimen, and a large number of unlabelled tumor cells were present in all samples. Labelled cells were usually small in size and mononucleated, either round or elongated in shape, and displayed a diffuse pattern of antimyosin staining; large multinucleated myofibers with obvious cross-striations were also observed in some specimens. Rhabdomyosarcoma cells from the bone marrow metastasis of case no. 3 were grown in vitro for more than three months and tested for reactivity with BF-G6 by indirect immunofluorescence. As shown in fig. 5 a number of cultured cells were specifically labelled by the antibody; most labelled cells displayed a diffuse or fibrillar pattern of staining, only rare cells showed a typical cross-striated pattern with fluorescent A-bands. Exp Cell Res 163 (1986)
218
Schiaffino
et al.
Table 2. Embryonic
myosin immunoreactivity
in rhabdomyosarcoma” Immunoreactivityb
Patient
Age
Sex
Type
1 2 3 4 5 6 7 8
lY 2~ 2~ 3~ 4.5 y 6~ 7~ 13y
F F M F M F F M
Embryonal Embryonal Alveolar Embryonal Embryonal Alveolar Embryonal Embryonal
._:3cc
Primary tumor
Metastasis’
+ + NT + +
+
t t t
-
+
+
t
t
LI As determined by immunofluorescent staining with monoclonal antibody BF-G6. ’ +, presence of labelled cells; -, absence of labelled cells; NT, not tested. c Specimens from two different metastases were obtained in patients no. 3 and 8.
DISCUSSION The results of this study show that human rhabdomyosarcoma cells in situ, as well as cultured cells derived from them, express a MHC type antigenically similar to that present in human muscle during early stages of development. This appears to be a further example of retrograde expression of embryonic proteins by tumor cells. Embryonic myosin can thus be used as a specific marker for rhabdomyosarcoma and the monoclonal antibody described here appears to be a valuable diagnostic aid. Other solid tumors of infancy and childhood were unreactive with this antibody, except for Wilms’ tumor. The latter finding was not surprising, since the frequent presence of striated muscle cells in Wilms’ tumor has long been recognized. It is difficult to compare embryonic MHC with other markers previously used in the study of rhabdomyosarcomas. The great variability of positive results obtained by different groups, even those using the same marker, clearly show that a reliable evaluation of various markers will be attained only when standardized reagents and procedures will be compared on the same samples. Monoclonal antibodies should be particularly useful for such comparisons. The principal advantage of embryonic MHC as a marker of skeletal muscle differentiation appears to be the absolute specificity for skeletal muscle and particularly for developing skeletal muscle. However, it should be stressed that many rhabdomyosarcoma cells do not express embryonic MHCs and that the frequency of positive cells may vary in different specimens and in different regions of the same specimen. In this respect embryonic myosin seems to differ from desmin, which according to a recent study is expressed in more than 95% of tumor cells in established cases of rhabdomyosarcoma [29]. Thus it appears that desmin is constitutively expressed in rhabdomyosarcoma cells, whereas embryonic myosin expression reflects a subsequent stage of the myogenic pathway. A combined Exp
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
219
study of desmin and embryonic myosin distribution in rhabdomyosarcomas may contribute to understand the differentiation process in neoplastic and normal muscle cells and also yield information useful for diagnostic and prognostic purposes. Using specific polyclonal antibodies we have previously shown that embryonic type MHCs are re-expressed in regenerating fibers in cold-injured rat [16] and chicken muscle [30], and in human dystrophic muscle [193. These results have now been confirmed using the monoclonal antibody described here (our unpublished observations). Regenerating muscle cells are thought to derive from the undifferentiated satellite cells that lie under the basal lamina of mature skeletal muscle fibers [31, 321. It is not known whether rhabdomyosarcomas derive from the same myogenic precursors or from other stem cells; the latter possibility is supported by the finding that many primary rhabdomyosarcomas arise from organs and tissues not containing skeletal muscle [l, 21. It will be of interest to compare the developmental capacity of rhabdomyoblasts with that of embryonic myoblasts and of satellite myoblasts. Previous studies have indicated that early embryonic myoblasts differ from satellite myoblasts, which emerge at later developmental stages in the human fetal limb, with respect to clonal growth in vitro and response to medium composition [33, 341. The study of MHC transitions may contribute to define the developmental capacity of different myogenic precursors. In developing rat muscle a distinct MHC isoform, referred to as neonatal MHC, is expressed after the embryonic isoform and before the appearance of the definitive fast adult MHCs [15]. A similar sequence of MHC transitions seems to occur in human skeletal muscle. This is clearly seen by comparing the pattern of reactivity of BF-G6 with that of a polyclonal antibody to human fetal MHCs [IS]. Neither antibody reacts with adult myosin but, at variance with BF-G6, the polyclonal antibody reacts with similar intensity with fetal and neonatal myosin; the difference between the two antibodies can be explained by assuming that the polyclonal antibody crossreacts with a distinct neonatal isoform which is not recognized by BF-G6. Independent evidence for the existence in human skeletal muscle of a MHC isoform analogous to the neonatal MHC of rat muscle can be inferred from electrophoretic studies [ 171 and from preliminary observations with other monoclonal antibodies (our unpublished work). Specific probes for the different MHC types will help to elucidate the stages of muscle cell differentiation during normal ontogenesis and regeneration, and determine whether rhabdomyosarcomas correspond to an arrested stage in the differentiation pathway. The finding that embryonic MHCs can be expressed in cultured rhabdomyosarcoma cells is of particular interest in this respect. Rhabdomyosarcoma cell lines represent an attractive experimental system for studying the expression of MHC isoforms and the effect of inducers of muscle cell differentiation on myosin switching. This work was supported by grants from CNR (Special Project MPR, SP3, grant no. 83.0X323.56; Special Project Oncology, grant no. 84.00495.44), Minister0 Pubblica Instruzione, and Associazione Italiana Ricerca Cancro. Exp
Cell
Res
163 (1986)
220
Schiaffino
et al. REFERENCES
1. Isaacs, H, Path01 ann 18 (1983) 165. 2. Sutow, W W, Fembach, D J & Vietti, T J, Clinical pediatric oncology. C. V. Mosby, St. Louis (1973). 3. Gonzalez-Crussi, F & BlackSchaffer, S, Am j surg path01 3 (1979) 157. 4. Mukai, K, Rosai, J & Hallaway, B E, Am j surg path01 3 (1979) 373. 5. Corson, J M & Pinkus, G S, Am j path01 103 (1981) 384. 6. Kindblom, L G, Seidal, T & Karlsson, K, Acta path01 microbial immunol Stand, sect. A 90 (1982) 167. 7. Eusebi, V, Bondi, A & Rosai, J, Am j surg path01 8 (1984) 51. 8. Altmannsberger, M, Osbom, M, Treuner, J, Holscher, A, Weber, K & Schauer, A, Virchows arch cell path01 39 (1982) 203. 9. Miettinen, M, Lehto, V P, Badley, R A & Virtanen, I, Am j path01 108 (1982) 246. 10. Gabbiani, G, Kapanci, Y, Barazzone, P & Franke, W W, Am j path01 104 (1981) 206. 11. Osbom, M & Weber, K, Lab invest 48 (1983) 372. 12. Koh, S J & Johnson, W W, Arch path01 lab med 104 (1980) 118. 13. Tsokos, M, Howard, R & Costa, J, Lab invest 48 (1983) 148. 14. Wydro, R, Nguyen, H, Gubits, R & Nadal-Ginard, B, J biol them 258 (1983) 670. 15. Whalen, R G, Sell, S M, Butler-Browne, G S, Schwartz, K, Bouveret, P & Pinset-Harstrom, I, Nature 292 (1981) 805. 16. Sartore, S, Gorza, L & SchiaBino, S, Nature 298 (1982) 294. 17. Fitzsimons, R B & Hoh, J F, J neurol sci 52 (1982) 367. 18. Biral, D, Damiani, E, Margreth, A & Scarpini, E, Biochem j 224 (1984) 923. 19. SchialEno, S, Gorza, L, Dones, I, Comelio, F & Sat-tore, S, Muscle & nerve. In press. 20. Leinwand, L A, Saez, L, McNally, E & Nadal-Ginard, B, Proc natl acad sci US 80 (1983) 3716. 21. Cantini, M, &tore, S L Schiaffino, S, J cell bio185 (1980) 903. 22. Barany, M & Close, R I, J physiol 213 (1971) 455. 23. Galfre, G & Milstein, C, Meth enzymo173 (1981) 3. 24. Engvall, E & Perlmann, P, J immunol 109 (1972) 129. 25. SchiafBno, S, Gorza, L, Saggin, L, Valfre, C & Sartore, S, Eur heart j 75 (1984) 95. 26. Towbin, H, Stae.helin, T & Gordon, J, Proc natl acad sci US 76 (1979) 4350. 27. Trojanowski, J Q, Obrocka, M A & Lee, V M Y, J histochem cytochem 31 (1983) 1217. 28. Stemberger, L A, Immunocytochemistry. Wiley, New York (1979). 29. Altmannsberger, M, Weber, K, Droste, R & Osbom, M, Am j path01 118 (1985) 85. 30. Gorza, L, Sartore, S, ‘Biban, C & Schiaflino, S, Exp cell res 143 (1983) 395. 31. Mauro, A, J biophys biochem cytol9 (l%l) 493. 32. Snow, M H, Anat ret 188 (1977) 201. 33. Hauschka, S D, Dev biol 37 (1974) 345. 34. Cossu, G, Cicinelli, P, Fieri, C & Molinaro, M, Exp cell res 160 (1985) 403. Received July 22, 1985
Exp
Cell
Res
163 (1986)
Printed
in Sweden
Embryonic
Cell Research
163 (1986) 21 l-220
Myosin Heavy Chain as a Differentiation of Developing Human Skeletal Muscle and Rhabdomyosarcoma A Monoclonal
Antibody
Marker
Study
S. SCHIAFFINO,’ L. GORZA,’ S. SARIORE,’ L. SAGGIN’ and M. CARLI ‘Institute
of General University
Pathology of Padova,
and *Department Z-35100 Padua,
of Pediatrics, Ztaly
Hybridoma cell lines were obtained from the fusion of NS-0 myeloma cells with spleen cells of mice immunized with bovine fetal skeletal myosin. A stable hybridoma clone, BFG6, produced immunoglobulin Cl k antibodies reacting specitically with embryonic-type myosin heavy chains present in fetal but not in neonatal or adult human skeletal muscle, as determined by enzyme immunoassay and immunoblot analysis. Fetal but not adult skeletal muscle fibers were stained by this monoclonal antibody in indirect immunofluorescence assays; smooth muscle cells and cardiac muscle cells, as well as non-muscle cells were also unreactive. Solid tumors of infants and children were tested for reactivity with BF-G6 by immunofluorescence and immunoperoxidase staining. Embryonic myosin heavy chain was expressed in rhabdomyosarcomas but not in other types of tumor, except for Wilms’ tumor. Rhabdomyosarcoma cells isolated from a bone marrow metastasis and grown in vitro for several months were also labelled by BF-G6. Embryonic myosin heavy chain can thus be used as a specific differentiation marker of normal and neoplastic skeletal muscle tissue.
@ 1986 Academic
Press, Inc.
Rhabdomyosarcoma is a highly malignant myogenic neoplasm accounting for a major proportion of soft tissue sarcomas in infants and children [l, 21. Different types of this tumor have been described, embryonal and alveolar rhabdomyosarcomas being the most common cytological variants. The distinguishing feature of rhabdomyosarcoma is the presence of striated muscle cells at initial stages of differentiation (rhabdomyoblasts). However, histologic evidence of striated muscle differentiation may be lacking or ambiguous and diagnosis of poorly differentiated rhabdomyosarcoma is often difficult by routine histology [3]. Identification of specific striated muscle markers by immunocytochemical procedures can provide more precise criteria for differential diagnosis. Myoglobin, which has been used as a rhabdomyosarcoma marker [4-61, may lead to false-positive results; lymphoma cells invading skeletal muscle were found to stain for myoglobin, which was presumably released by necrotic muscle fibers and taken up by tumor cells [7]. Desmin is a good marker for rhabdomyosarcoma [8, 93 but does not allow differentiation with leiomyoma and leiomyosarcoma [lo, 111. Antibodies against myosin, used in previous studies on rhabdoCopyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/86 $03.00
212 Schiafjno
et al.
myosarcoma [12, 131, were not characterized with respect to specificity; some of these antibodies were certainly not specific for sarcomeric myosins and stained also breast carcinoma cells [13]. Furthermore, antimyosin antibodies were prepared against adult-type myosins and were thus not appropriate for identifying undifferentiated muscle cells, which are now known to express distinct myosin isoforms . The myosin molecule consists of two heavy and four light chains. In striated muscle there are multiple forms of myosin heavy chains (MHCs): these sarcomerit MHCs are encoded by different genes whose expression is tissue-specific and developmentally regulated [ 141. Developing rat muscle contains embryonictype MHCs that are progressively lost during post-natal development [ 151 but can be re-expressed during muscle regeneration [16]. Embryonic isoforms of MHCs have also been identified in human skeletal muscle [17-191 and evidence has been presented for the existence of an embryonic specific human MHC gene [20]. We describe here a monoclonal antibody which recognizes embryonic MHC in developing human muscle and in rhabdomyosarcoma. MATERIALS
AND
METHODS
Tissue Sources Fetal weeks. months Tumors years).
samples were from 6 fetuses resulting from termination of pregnancy at 8,9, 10, 11 [21 and 20 Post-natal samples were from an I-day-old infant (biopsy specimen), 2 babies aged 8 and 21 (biopsy specimens) and 8 individuals, aged 947 years (autopsy and biopsy specimens). samples were from different solid tumors of infants and children (age range 3 months to 13 All samples were frozen in liquid nitrogen and stored in a freezer at -80°C.
Tissue Culture of Rhabdomyosarcoma
Cells
A bone marrow aspiration from a rhabdomyosarcoma patient with bone marrow metastasis (patient no. 3) was first treated with ice-cold 0.83 % NH&l to eliminate red blood cells. Centrifuged cells were then plated in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10 % horse serum and 1% glutamine. Cells were either grown in plastic culture flasks or, for immunofluorescence studies, on gelatin-coated glass coverslips placed in plastic dishes. After various time intervals coverslips were treated with acetone at -20°C for 5 min and air-dried before processing for immunofluorescence 1211.
Production
of Monoclonal
Antibodies
Myosin was isolated from skeletal muscles of 3-month-old bovine fetus as described by Barany & Close 1221. BALB/c mice were immunized by two intradermal injections (administered two weeks apart) of 100 pg of bovine fetal myosin emulsified with complete Freund adjuvant. Two weeks after the second injection the animals were bled and the sera were screened for the presence of antimyosin antibodies by enzyme immunoassay and immunofluorescence. Animals showing the highest antibody titres were selected and one month later were injected intraperitoneally with 50 pg of myosin on one day and the same dose intravenously on the following day. Three days after the last injection spleen cells of the mice were fused with mouse myeloma cell line NS-0 using standard techniques 1231. The hybridomas were grown in microplates and were screened for production of antimyosin antibodies by indirect immunofluorescence on cryosections of fetal and adult skeletal muscle. Selected hybridomas were cloned twice by the limiting dilution method. The hybridoma line used for this study, called BFG6, was found to produce IgGl k antibody, as determined by enzyme immunoassay with antisera specific for the major mouse immunoglobulin classes and for the subclasses of IgG and light chains; culture supematant was used for immunochemical assays and immunofluorescence tests. Exp
Ceil
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
213
Fig. 1. Reactivity of monoclonal antibody BF-G6 with myosin isolated from 0, lOweek human fetal muscle; n , I-day newborn muscle; 0, adult skeletal muscle, as determined by solid-phase enzyme immunoassay. 0.3 Myosin
3
30
(pglmll
Enzyme Immunoassay Solid-phase enzyme immunoassays were performed on microplates following the procedure of Engvall & Perlman [24]. Microplate wells coated with different concentrations of fetal or adult muscle myosin were incubated with appropriate dilutions of monoclonal supematant (usually 1: 100 to 1 : 200). Bound antibody was revealed by rabbit anti-mouse IgG conjugated with horseradish peroxidase (Miles Lab.); enzyme concentration was determined with o-dianisidine as substrate and measuring the absorbance at 410 nm.
Zmmunoblotting SDS-PAGE of crude myosin preparations [25] from human fetal and adult skeletal muscle was followed by electrophoretic transfer to nitrocellulose paper as described by Towbin et al. [26]. Bound antibody was revealed by immunoperoxidase staining, using rabbit anti-mouse IgG conjugated with horseradish peroxidase and diaminobenzidine as substrate in the presence of imidazole [27].
Zmmunojluorescence
and Immunoperoxidase
Staining
Indirect immunofluorescence tests were performed on fresh frozen sections of fetal and adult human muscle and tumor biopsies, or acetone-fixed cultures of rhabdomyosarcoma. After incubation with appropriate dilutions of monoclonal supematant (usually 1: 8-l : 16) for 30 mitt at 37’C, sections were washed and incubated with fluorescein-labelled rabbit anti-mouse IgG (Miles Lab.) for 30 min at 37°C. Sections were then post-fared in 1.5 % paraformaldehyde in phosphate-buffered saline (PBS) and examined under epitluorescence illumination on a Leitz Dialux microscope. Immunoperoxidase staining was performed following the peroxidase-anti-peroxidase (PAP) technique as described by Stemberger [28].
RESULTS The specificity of monoclonal antibody BF-G6 is illustrated in figs 1-3. BF-G6 reacted with myosin from human fetal muscle and not with neonatal and adult myosin in solid-phase enzyme immunoassay (fig. 1) and reacted exclusively with fetal MHCs on immunoblots (fig. 2). Fetal muscle fibers showed specific immunofluorescent staining with BF-G6 in cryosections of human skeletal muscle (fig. 3). Comparable reactivity with BF-G6 was found in fetuses of 8-11 and 20 weeks by immunofluorescence, except that all fibers were brightly labelled in the Exp
Ceil
Res 163 (1986)
214
Schiaffino
et al.
12.3
4
2oOK-
Fig. 2. Immunoblotting of crude myosin preparations from 2, 4 adult human muscle; I, 3, IO-week fetal muscle with 3, 4, antibody BF-G6; 1, 2, Coomassie Blue-stained gel. Nitrocellulose paper strip reacted with antibody and processed for immunoperoxidase staining. MW markers are shown on the left.
younger fetuses, while a minor proportion of unlabelled fibers were seen in the older fetus. Only few weakly labelled fibers persisted in neonatal muscle and no reactivity was seen in muscle samples from 8- and 21-month-old infants and in adult samples. Vascular smooth muscle and cardiac muscle, as well as nonmuscle tissues were unstained by this antibody at any age. Cryosections from 17 solid neoplasms, including 8 rhabdomyosarcomas, occurring in infants and children were processed for indirect immunofluorescence with monoclonal antibody BF-G6. Reactive cells were detected in all cases of rhabdomyosarcoma, showing variable degree of differentiation, whereas they were absent from other tumors, except for Wilms’ tumor (table 1). Labelled cells were also present in specimens obtained from metastases or local recurrences of rhabdomyosarcoma, except for one specimen from case no. 3 (table 2). In this patient a biopsy from a lymph node metastasis was negative, whereas a second biopsy obtained 5 months later from the bone marrow revealed the presence of
Table 1. Embryonic childhood”
myosin
immunoreactivity
Type of tumor
Age range
Bhabdomyosarcoma Other tumors=
3m13y
1-13 y
in solid tumors of infancy and
No. of positive casesb No. of cases examined 818 l/9
a As determined by immunofluorescent staining with monoclonal antibody BF-G6. ’ Cases showing presence of labelled cells in at least one of the specimens examined. c Fibrosarcoma, neuroblastoma, ganglioneuroblastoma [Z], Ewing’s sarcoma [2], extraskeletal Ewing sarcoma, embryonal carcinoma, Wilms’ tumor. Wilms’ tumor was the only positive case. C-p
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
215
Fig. 3. Indirect immunofluorescence staining of human skeletal muscle with antibody BFG6. Cryosections of skeletal muscle from (a) lO-week fetus; (6) 20-week fetus; (c) E-day newborn; (4 E-month infant. X400.
216
Schiaffino
et al.
Fig. 4. (a) Immunofluorescence; (b) immunoperoxidase staining of cryosections of human rhabdomyosarcoma with antibody BF-G6. The section in (b) has been stained with hematoxylin after the immunoperoxidase reaction: note the presence of both labelled and unlabelled tumor cells. X4Cl&
Exp
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
217
Fig. 5. Immunofluorescence staining with antibody BF-G6 of (a) smear of bone marrow aspiration from child with rhabdomyosarcoma; (b, c) cells from the same metastasis cultured in vitro for 30 days. Occasional labelled cells showed cross-striation (b). x500.
labelled cells. As shown in fig. 4 the frequency of labelled cells varied from specimen to specimen and in different regions of the same specimen, and a large number of unlabelled tumor cells were present in all samples. Labelled cells were usually small in size and mononucleated, either round or elongated in shape, and displayed a diffuse pattern of antimyosin staining; large multinucleated myofibers with obvious cross-striations were also observed in some specimens. Rhabdomyosarcoma cells from the bone marrow metastasis of case no. 3 were grown in vitro for more than three months and tested for reactivity with BF-G6 by indirect immunofluorescence. As shown in fig. 5 a number of cultured cells were specifically labelled by the antibody; most labelled cells displayed a diffuse or fibrillar pattern of staining, only rare cells showed a typical cross-striated pattern with fluorescent A-bands. Exp Cell Res 163 (1986)
218
Schiaffino
et al.
Table 2. Embryonic
myosin immunoreactivity
in rhabdomyosarcoma” Immunoreactivityb
Patient
Age
Sex
Type
1 2 3 4 5 6 7 8
lY 2~ 2~ 3~ 4.5 y 6~ 7~ 13y
F F M F M F F M
Embryonal Embryonal Alveolar Embryonal Embryonal Alveolar Embryonal Embryonal
._:3cc
Primary tumor
Metastasis’
+ + NT + +
+
t t t
-
+
+
t
t
LI As determined by immunofluorescent staining with monoclonal antibody BF-G6. ’ +, presence of labelled cells; -, absence of labelled cells; NT, not tested. c Specimens from two different metastases were obtained in patients no. 3 and 8.
DISCUSSION The results of this study show that human rhabdomyosarcoma cells in situ, as well as cultured cells derived from them, express a MHC type antigenically similar to that present in human muscle during early stages of development. This appears to be a further example of retrograde expression of embryonic proteins by tumor cells. Embryonic myosin can thus be used as a specific marker for rhabdomyosarcoma and the monoclonal antibody described here appears to be a valuable diagnostic aid. Other solid tumors of infancy and childhood were unreactive with this antibody, except for Wilms’ tumor. The latter finding was not surprising, since the frequent presence of striated muscle cells in Wilms’ tumor has long been recognized. It is difficult to compare embryonic MHC with other markers previously used in the study of rhabdomyosarcomas. The great variability of positive results obtained by different groups, even those using the same marker, clearly show that a reliable evaluation of various markers will be attained only when standardized reagents and procedures will be compared on the same samples. Monoclonal antibodies should be particularly useful for such comparisons. The principal advantage of embryonic MHC as a marker of skeletal muscle differentiation appears to be the absolute specificity for skeletal muscle and particularly for developing skeletal muscle. However, it should be stressed that many rhabdomyosarcoma cells do not express embryonic MHCs and that the frequency of positive cells may vary in different specimens and in different regions of the same specimen. In this respect embryonic myosin seems to differ from desmin, which according to a recent study is expressed in more than 95% of tumor cells in established cases of rhabdomyosarcoma [29]. Thus it appears that desmin is constitutively expressed in rhabdomyosarcoma cells, whereas embryonic myosin expression reflects a subsequent stage of the myogenic pathway. A combined Exp
Cell
Res
163 (1986)
Embryonic
myosin in rhabdomyosarcoma
219
study of desmin and embryonic myosin distribution in rhabdomyosarcomas may contribute to understand the differentiation process in neoplastic and normal muscle cells and also yield information useful for diagnostic and prognostic purposes. Using specific polyclonal antibodies we have previously shown that embryonic type MHCs are re-expressed in regenerating fibers in cold-injured rat [16] and chicken muscle [30], and in human dystrophic muscle [193. These results have now been confirmed using the monoclonal antibody described here (our unpublished observations). Regenerating muscle cells are thought to derive from the undifferentiated satellite cells that lie under the basal lamina of mature skeletal muscle fibers [31, 321. It is not known whether rhabdomyosarcomas derive from the same myogenic precursors or from other stem cells; the latter possibility is supported by the finding that many primary rhabdomyosarcomas arise from organs and tissues not containing skeletal muscle [l, 21. It will be of interest to compare the developmental capacity of rhabdomyoblasts with that of embryonic myoblasts and of satellite myoblasts. Previous studies have indicated that early embryonic myoblasts differ from satellite myoblasts, which emerge at later developmental stages in the human fetal limb, with respect to clonal growth in vitro and response to medium composition [33, 341. The study of MHC transitions may contribute to define the developmental capacity of different myogenic precursors. In developing rat muscle a distinct MHC isoform, referred to as neonatal MHC, is expressed after the embryonic isoform and before the appearance of the definitive fast adult MHCs [15]. A similar sequence of MHC transitions seems to occur in human skeletal muscle. This is clearly seen by comparing the pattern of reactivity of BF-G6 with that of a polyclonal antibody to human fetal MHCs [IS]. Neither antibody reacts with adult myosin but, at variance with BF-G6, the polyclonal antibody reacts with similar intensity with fetal and neonatal myosin; the difference between the two antibodies can be explained by assuming that the polyclonal antibody crossreacts with a distinct neonatal isoform which is not recognized by BF-G6. Independent evidence for the existence in human skeletal muscle of a MHC isoform analogous to the neonatal MHC of rat muscle can be inferred from electrophoretic studies [ 171 and from preliminary observations with other monoclonal antibodies (our unpublished work). Specific probes for the different MHC types will help to elucidate the stages of muscle cell differentiation during normal ontogenesis and regeneration, and determine whether rhabdomyosarcomas correspond to an arrested stage in the differentiation pathway. The finding that embryonic MHCs can be expressed in cultured rhabdomyosarcoma cells is of particular interest in this respect. Rhabdomyosarcoma cell lines represent an attractive experimental system for studying the expression of MHC isoforms and the effect of inducers of muscle cell differentiation on myosin switching. This work was supported by grants from CNR (Special Project MPR, SP3, grant no. 83.0X323.56; Special Project Oncology, grant no. 84.00495.44), Minister0 Pubblica Instruzione, and Associazione Italiana Ricerca Cancro. Exp
Cell
Res
163 (1986)
220
Schiaffino
et al. REFERENCES
1. Isaacs, H, Path01 ann 18 (1983) 165. 2. Sutow, W W, Fembach, D J & Vietti, T J, Clinical pediatric oncology. C. V. Mosby, St. Louis (1973). 3. Gonzalez-Crussi, F & BlackSchaffer, S, Am j surg path01 3 (1979) 157. 4. Mukai, K, Rosai, J & Hallaway, B E, Am j surg path01 3 (1979) 373. 5. Corson, J M & Pinkus, G S, Am j path01 103 (1981) 384. 6. Kindblom, L G, Seidal, T & Karlsson, K, Acta path01 microbial immunol Stand, sect. A 90 (1982) 167. 7. Eusebi, V, Bondi, A & Rosai, J, Am j surg path01 8 (1984) 51. 8. Altmannsberger, M, Osbom, M, Treuner, J, Holscher, A, Weber, K & Schauer, A, Virchows arch cell path01 39 (1982) 203. 9. Miettinen, M, Lehto, V P, Badley, R A & Virtanen, I, Am j path01 108 (1982) 246. 10. Gabbiani, G, Kapanci, Y, Barazzone, P & Franke, W W, Am j path01 104 (1981) 206. 11. Osbom, M & Weber, K, Lab invest 48 (1983) 372. 12. Koh, S J & Johnson, W W, Arch path01 lab med 104 (1980) 118. 13. Tsokos, M, Howard, R & Costa, J, Lab invest 48 (1983) 148. 14. Wydro, R, Nguyen, H, Gubits, R & Nadal-Ginard, B, J biol them 258 (1983) 670. 15. Whalen, R G, Sell, S M, Butler-Browne, G S, Schwartz, K, Bouveret, P & Pinset-Harstrom, I, Nature 292 (1981) 805. 16. Sartore, S, Gorza, L & SchiaBino, S, Nature 298 (1982) 294. 17. Fitzsimons, R B & Hoh, J F, J neurol sci 52 (1982) 367. 18. Biral, D, Damiani, E, Margreth, A & Scarpini, E, Biochem j 224 (1984) 923. 19. SchialEno, S, Gorza, L, Dones, I, Comelio, F & Sat-tore, S, Muscle & nerve. In press. 20. Leinwand, L A, Saez, L, McNally, E & Nadal-Ginard, B, Proc natl acad sci US 80 (1983) 3716. 21. Cantini, M, &tore, S L Schiaffino, S, J cell bio185 (1980) 903. 22. Barany, M & Close, R I, J physiol 213 (1971) 455. 23. Galfre, G & Milstein, C, Meth enzymo173 (1981) 3. 24. Engvall, E & Perlmann, P, J immunol 109 (1972) 129. 25. SchiafBno, S, Gorza, L, Saggin, L, Valfre, C & Sartore, S, Eur heart j 75 (1984) 95. 26. Towbin, H, Stae.helin, T & Gordon, J, Proc natl acad sci US 76 (1979) 4350. 27. Trojanowski, J Q, Obrocka, M A & Lee, V M Y, J histochem cytochem 31 (1983) 1217. 28. Stemberger, L A, Immunocytochemistry. Wiley, New York (1979). 29. Altmannsberger, M, Weber, K, Droste, R & Osbom, M, Am j path01 118 (1985) 85. 30. Gorza, L, Sartore, S, ‘Biban, C & Schiaflino, S, Exp cell res 143 (1983) 395. 31. Mauro, A, J biophys biochem cytol9 (l%l) 493. 32. Snow, M H, Anat ret 188 (1977) 201. 33. Hauschka, S D, Dev biol 37 (1974) 345. 34. Cossu, G, Cicinelli, P, Fieri, C & Molinaro, M, Exp cell res 160 (1985) 403. Received July 22, 1985
Exp
Cell
Res
163 (1986)
Printed
in Sweden