Histol Histopathol (1997) 12: 755-760 and 001: 10.14670/HH-12.755 Histopathology http://www.hh.um.es From to Engineering

Invited Review

Unorthodox : possible developmental significance and implications for tissue and

G. Cossu Department of Histology and Medical , University of Rome "La Sapienza", Rome, Italy

Summary. During the last few years several reports have th ere is no such cell as an undiffe rentiated cell : in a described the occurrence of skeletal myogenesis in cell s given area of an , at a given developmental stage, derived from embryonic, fetal and perinatal tissues that each cell is already di ffe rent fro m cell s located in usuall y do not contribute to in the adult di ffe rent areas as well as from its ancestors and progeny. body. Although in most cases fa te may still be experimentall y Afte r a brie f descriptio n of c urre nt ideas on changed, cell s already express a subset of genes typical, myogenic determination in higher , three a ltho ug h pe rhaps no t unique, o f a pa rtic ula r examples of this uno rthodox m yogenesis will be develo pme nta l stage and locatio n. Up to few years described: 1) the occurrence of myogenesis in chick ago we knew very few of these genes. With the ce ll s, c ultured in isola tion in serum-f ree explosive progress of mo lecul ar e mbryolog y, new me dium ; 2) the presence of c e ll s endow e d with developmental genes are continuously identified and myogenic potential in the embryonic mouse ; their developmental pattern of expression described. An and 3) th e occurre nce of spo nt aneous o r induced emerging picture begin s to appear where each cell , or myogenesis in mesenchymal cells during fe tal and post­ ra the r each g ro up of ce ll s , e xpresses a unique natal life. combination of genes, such as ho meogenes and other A possible e mbryologica l basis fo r uno rthodox transcriptio n facto rs, g rowth facto rs a nd cyto kines, myogenesis, in relati on to gastrulati o n and mo rpho­ extracellular matri x and adhesion molecules, each of genetic fields, is then presented. It is also proposed that which is certainly al so expressed in many other pl aces unorthodox myogenesis may represent a compensatory and times. As a result, a progressive specificati on of mechanism for higher vertebrates that have lost much of positio n a nd fa te is o bta ined leading eventuall y to the regeneration potential of lower vertebrates. terminal di fferentiation. For example a newly fo rmed mesodermal cell (w hi ch already expresses ­ Key words: Myogenesis, Determination, Histogenesis, specific genes such as brachiury, , etc.) may find Differentiation, Myogenic genes itself closer to the ax ia l structures (no tocho rd and mesoderm) and thus adopt a paraxial fate, ending up in . Later the dorsal part of the is specified Introduction as dermomyoto me and finall y a choice between the fi broblastic and the myogenic lineage will be made. Di ffe rentiatio n is commonl y defined as a process The avail ability of new molecular marke rs has w hi c h leads a cell to express a reperto ire of genes allowed us to investigate the expression of specific gene specific and characteristi c of the tissue where the cell produc ts a t th e s ing le cell leve l, d uring ti ssue belongs. However we can presently study with some histogenesis and regeneration. Several of these studies confid ence onl y terminall y differentiated cells. In this have unexpectedl y revealed the occurrence of (or the c ontext it is impo rta nt to dis ting uis h te rmina l po te ntia l fo r) s ke le ta l muscle di ffe re ntiati o n in a di ffe rentiation from all the previous steps th at lead the significant number of cells belonging to tissues where progeny of a to progressive diversification muscle no rma ll y does no t form. Furthe rmo re this from the other cells of the embryo. Indeed, as pointed process has been observed at inappropriate time, either out by Holtzer many years ago (Holtzer et aI. , 1973), before or aft er the end of . In this review I will briefl y discuss current ideas on Offprint requests to: Dr. Gu iul io Cossu , Institute of Histology and myogenic determinati on in , then describe General Embryology, University of Rome .. La Sapienza», 14 Via A. three di ffe rent examples of unorthodox myogenesis, and Scarpa, 00161 Italy then I will discuss possible deveJopmental signi ficance Unorthodox myogenesis in mammals

in terms of plasticity and regenerative potential of epiblast cells with adjacent tissue will inhibit mammalian tissues. myogenesis (George-Weinstein et al., 1996). One Unorthodox myogenesis is conceptually distinct interpretation of these experiments therefore is that from trans-differentiation, a phenomenon by which an epiblast cells even prior to their entry into the primitive already differentiated cell can be induced to change the streak and specificatioq as mesoderm, are already repertoire of gene expressed and to express genes typical programrned for myogenesis and that in the absence of of a different tissues. In higher vertebrates, this situation repression exerted by the in vivo context, they will is mainly related to pathology (metaplasia), although differentiate into muscle (Fig. 1). Notch has been trans-differentiation from smooth to skeletal muscle has proposed as a possible mediator of this inhibitory been recently demonstrated at the single cell leve1 in the mechanism (Kopan et al., 1994). If this is the case, then post-natal mammalian esophagus (Patapoutian et al., the influence of structures surrounding the early somites 1995). Trans-differentiation is not discussed in this is not to induce myogenesis but to relieve its repression. review. By PCR analysis, MyoD messenger RNA is detectable in the chick epiblast and in the mouse embryo prior to The origin of myoblasts in mammals somitogenesis (Kopan et al., 1994; George-Weinstein et al., 1996). The recent claim that, in Drosophila, wingless Skeletal myoblasts of the vertebrate body (with the is another ligand of notch (Couso and Martinez-Arias, exception of the head) are derived from somites, 1994), potentially competing with delta, raises the segrnented blocks of which form in a possibility that in vertebrates the Wnts (homologues of cranio-caudal sequence along the neural tube and the wingless) produced by the neural tube or dorsal (Christ and Ordhal, 1995). Among the cells of , relieve repression of myogenesis by blocking the somite, only those which are located in the dorsal the notch receptor. Even if the receptor for Wnt has been domain, termed dermomyotome, are specified as recently identified as Fri.z.de (Bhanot et al., 1996), a myoblasts (and dermal ) while cells located in protein different from Notch, a cross-talk between the the ventral domain, the sclerotome, will form two receptors may anyway relieve this repressive action. and . Myogenesis depends upon activation of either Whatever the case in molecular terms, the natural myf-5 or MyoD , the two upstream genes of this family tendency of embryonic cells towards myogenesis may of muscle specific transcnption factors (Rudnicki et al., explain the peculiar capacity of MyoD to activate 1993). Recent work has shed light on the role of adjacent myogenesis in non muscle cells and should be kept in tissue such as neural tube, notochord and dorsal mind when discussing the unorthodox origin of ectoderm, which appear to activate different myogenic myogenic cells. programs in responding mesodermal cells. Specifically the neural tube activates myogenesis in the dorso-media1 Misplaced cells and lineage lnfidellty half of paraxial mesoderm (fated to give rise to epaxial muscles: Ordhal and Le Douann, 1991) through a myf-5- Spontaneous myogenic differentiation of cells from dependent pathway, while the dorsal ectoderm can the has been documented a number of times activate myogenesis in the dorso-lateral precursors of (examples quoted in Tajbakhsh et al., 1994) but it was hypaxial muscles (such as limb and body wall) through a only through insertion of the reporter gene LacZ into the Myo D-dependent pathway. In this case, lateral myf-5 locus that it was possible to unequivocally identify mesoderm delays the positive effect of the ectoderm myf-5 expressing cells in the and to show suggesting that this tissue produces an inhibitory signal that these cells co-express neural and muscle markers (Cossu et al., 1996a) to maintain the cells in a committed and yet undifferentiated state during the migration to their final destination. Candidate molecules for this Myogenesis by default complex signaling activity include and Wnts as positive signals (Munstenberg et al., 1995), and BMP4 as a possible inhibitor (reviewed in Cossu et al., 1996b).

Myogenesis as a default process In apparent contrast with what is exposed above, recent observations suggest a "default" tendency of embryonic cells toward myogenesis, which should be therefore repressed in vivo until proper time and place. In contrast with their normal fate, the great majority of cells from the chick epiblast layer, cultured in vitro, will Flg. 1. A schernatic representation of the repressive influence of cell-cell undergo myogenesis, even when grown at low clonal interaction on the default tendency to rnyogenesis of chick pre-gastrula density in protein-free medium. Co-culture of the cells. Unorthodox myogenesis in mammals

(Tajbakhsh et al., 1994). A possible explanation for the quiescent in the ectopic tissue. Expression of myf-5 in neural origin of a small population of myoblasts may be the neural tube in vivo would support the possibility that found in the process of gastrulation. Fate maps of at least a fraction of these cells neither die nor mammalian gastrulae have shown that the areas completely convert to muscle, and transient co- predestined to become mesoderm are contiguous and expression of neurona1 and muscle marken in culture overlapping with areas fated to ectoderm and, moreover suggests that a conversion to muscle can occur in vitro. these lineages are not separated at the beginning of gastrulation (Lawson et al., 1991). It is conceivable that Myogenlc conversion of mesodermal cells ' . a few cells, lying at the border between presumptive r >I' - . paraxial mesoderm and may be trapped A different embryological situation is represented by into an improper neurogenic field. If this is the case, the the recently reported myogenic conversion of "fibro- presence of B-galactosidase-positive cells in the neural blasts" originating from dermis and, to a different extent, tube (Fig. 2) suggests that some cells may express myf-5 other mesodermal tissues. The first example of this in the well before somitogenesis. phenomenon was correction by -myoblast Preliminary observations suggest that this may be the fusion of the genetic defect of the mdg mouse mutant case (Cossu et al., 1996b). If these cells can no longer be muscle fibers (Chaudari et al., 1989; Courbin et al., converted to a neurogenic fate, they may die or remain 1989). Recently Gibson et al., (1995) found that dermal

Fig. 2. Transverse section of a 10.5 dpc myf-5/L& mouse embryo. The arrow shows B-galactosidase-positiw, myf-6 expressing cells in the neural tube (NT). The majority of 0- galactosidase-positive, myf-5 expressing cells are located in the myotome (M), while a few myf-5-expressing cells are already , detectablein'the iimb bud (LB). Bar: 5qum. - --. --7-.- 7--- Unorthodox myogenesis in mammals

fibroblasts can give rise to muscle fibers when injected precursors lay at the border of a "progenitor field". Since into skeletal muscle of mdx mice. These studies showed any field must by definition possess sharp boundaries, evidence for the fusion of fibroblasts with myogenic those cells which find themselves at the border of a cells. In these cases myogenesis could be activated as "progenitor field", rnay escape from the "cornm~nity'~ occurs in heterokaryons where the fibroblast nucleus is since they are not completely surrounded by sister cells exposed to muscle transcription factors. However we and and, furthermore, they receive subthreshold levels of others reported that myogenesis can also be induced in a different signals (Fig. 3). These cells rnay therefore be mononucleated fibroblast by signals derived from frozen in a penultimate and possibly bi-potential state. neighboring myogenic cells: in this case fusion would be Other examples of lineage switching among mesodermal just a consequence of the acquired myogenic phenotype cells have been reported (Grigoriadis et al., 1988; (Breton et al., 1995; Salvatori et al., 1995). Katagiri et al., 1994) supporting the idea of It remains to be explained why these mesodermal mesenchymal cells which maintain a bi-potential or cells require a "muscle field" to undergo myogenesis and multipotential state. In this state, mesenchymal cells rnay have not already undergone spontaneous differentiation. be induced to adopt a specific pathway of terminal In vertebrates, organs or parts of them are established differentiation from signals derived from neighboring starting from "progenitor fields" which are dependent differentiating cells, as occurs in a forming muscle field upon complicated cross-talks between diffusible (both in vitro and in vivo during regeneration). The molecules and specific transcription factors (Davidson, 10T1/2 cells, which can differentiate into muscle, 1993). The recently reported "community effect" for cartilage and fat, rnay be the immortaiized progeny of amphibian myogenesis rnay explain in part how such one of these cells (Taylor and Jones, 1982). fields rnay be established (Gurdon, 1993). While receiving inductive signals, embryonic cells must check Conclusions and future perspectlves that they are surrounded by similar cells in order to be committed to a given fate. We have recently observed the The examples described above point to the existence existence of a "community effect" for mammalian of multiple mechanisms by which a cell of the vertebrate myogenesis as well (Cossu et al., 1995). In the examples body can be forced to adopt a myogenic fate. of myogenic conversion reported here, however, cells Even at the present leve1 of uncertainty, the come from tissues that have already completed following possibilities rnay be proposed: morphogenesis and therefore we rnay imagine a 1) a "default" tendency of pre-gastrula cells, which recruitment to myogenesis of cells which are not totally lacks a cellular explanation; it rnay only be uniquely determined, such as those cells whose tentatively linked to the so far unique property of myogenic bHLH transcription factors to act dominantly Origin of multipotential cells in a foreign cellular environment (Weintraub et al., 1991). 2) a misplacement at, or soon after, gastrulation that leads a cell belonging to an area fated to be paraxial mesoderm to migrate within a different territory and to be trapped in a different tissue. This rnay explain the frequent and unexplained presence of cell with myogenic potential within the central nervous tissue (Tajbakhsh et al., 1994 and examples therein) 3) a borderline position between two different progenitor fields, which exposes the same cell to -w - subhtreshold levels of two different signaling molecules which specify different fates (Le. those cells of newly e: AA specified formed somites, lying at the border between future 0: A specified sclerotome and dermomyotome). These cells rnay not adopt a final fate and thus remain as undifferentiated, @: ABunspecified possibly multipotential cells. Perhaps it is worth noting @: B specified that cells which maintain a developmental potency, such as satellite cells or osteoblasts, invariably lie at the Q) : BB specified border of a tissue, under the basa1 lamina, and rnay be experimentally induced to adopt different fates. Flg. 3. A simple model showing the possible origin of multipotential Borderline cells might have been exploited during cells. A group of cells is subjected to two different inducing signals (A mammalian evolution to compensate for the reduced and B) emanating frorn neighboring tissues. Cells closest to these regenerative capacity of mammalian as compared to tissues reoeive a high dose of the signais and are specified as AA or BB. The next layers of cells, receiving lower doses are specified as A lower vertebrate tissues (Muneoka and Sassoon, 1992). and B respectively. Cells which receive subthreshold doses of both A In the case of skeletal muscle, regeneration potential and B signals, are not specifed and remain as (AB) multipotential cells. rnay be rapidly exhausted in response to chronic injury, Unorthodox myogenesis in mammals

as occurs in Duchenne Muscular Dystrophy (Partridge, (1989). Rescue of excltation-contraction coupling in disgenic muscle 1991). Surprisingly, in situ analysis of MyoD and by addiüon of fibroblasis in vitro. 2, 1341-1350. myogenin expression in regenerating muscle, revealed Couso J.P. and Martínez-Mias A. (1994). Notch is required for Wingless an unexpectedly high number of positive cells near the signaling in the of Drosophlla. Cell79, 259-272. area of muscle necrosis (Grounds et al., 1991), far Davidson E.H. (1993). Later embryogenesis: regulatory circuitry in exceeding that of expected resident satellite cells. These morphogenetic fields. Deveiopment 118, 665-690. MyoD-expressing cells might derive from satellite cells George-WeinstBin M., Gerhart J., Flynn J., Callihan B.A., Miehle C., Foti of neighboring muscles which have rapidly migrated into G.J. and Lash J.W. (1996). lndependence of chich embryo myo- the damaged area, but may as well represent recruitment genesis from gastrulation, tissue interaction and replication. Dev. into myogenesis of resident mesodermal cells, which Biol. 173, 729-741. appear as fibroblasts, but may maintain a myogenic Gibson A.J., Karasinski J., Rehras J. Moss J., Sherratt T.G., Strong P.N. potential. and Watt D.J. (1995). Dermal fibroblasts convert to a myogenic Hopefully, the increasing understanding of the lineage in mdx mouse muscle. J. Cell Sci. 108, 207-214. cellular and molecular control of differentiation will Grigoriadis A.E., Heenche J.N.M. and Aubin J.E. (1988). Dmerentiation shed further light on the process of unorthodox myo- of muscle, fat, cartilage and bone from progenitor cells present in a genesis, leading to a better definition in embryological bone marrow derived clonal cell population: effect of terms. The understanding of this process may bear dexamethasone. J. Cell Biol. 106, 2139-2151. important implications for ex vivo cell therapy of Grounds M.D., Garrett K.L., Lai M.C., Wnght W.E. and Beilharz M.W. primary myopathies. The current limitation of this (1991). Identication of skeletal muscle precursor cells in vivo by use approach lies in the limited number and limited lifespan of MyoD 1 and myogenin probes. Cell Tissue Res. 267, 99-104. of myogenic cells which can be obtained from the Gurdon J.B. (1993). Community effect and related phenomena in biopsy of a patient. Fibroblasts from a non damaged site developrnent. Cell75,501-506. such as , may be more easily expanded in vivo, Holtzer H., Sanger J., lshikawa H. and Strahs K. (1973). Selected topics transduced with the vector carrying the therapeutic gene, in skeletal myogenesis. Cold Spring Harbor Symp. Quantit. Biol. and then converted to myogenesis, before reintroduction MIII. pp 549-566. in the patient's own muscle tissue. Experiments are in Katagiri T., Yamaguchi A., Komakl M., Abe E., Takahashi N., lkeda T., progress to test the feasibility of this approach. Rosen V., Wozney J. M., Fujisawa-Sehara A. and Suda T. (1994). Bone morphogenetic protein-2 coverts the differentiation pathway of Acknowledgments. Work in the Author's laboratory has been supported C2C12 myoblasts into the osteoblast lineage. J. Cell Biol. 127, by grants from Fondazione lstituto Pasteur-Cenci Bolognetti, EC B104 1755- 1766. CT95-0228 and MURST. I wish to thank Drs. S. Adamo, M. BouchB, M. Kopan R., Nye J.S. and Weintraub H. (1994). The intracellular domain Molinaro and C. Newi for critica1 reading of the manuscript. of mouse Notch: a costitutively activated repressor of myogenesis directed at the basis helix-loop-helix region of MyoD. Development 120, 2385-2396. References Lawson K.A., Meneses J.J. and Pedersen R.A. (1991) Clonal analysis of epiblast fate during formation in the mouse embryo. Banhot P., Brink M., Harryman Sammos C., Hsieh J.-C., Wang Y., Development 113, 891-91 1. Macke J.P., Andrew D., Nathans J. and Nusse R. (1996). A new Muneoka K. and Sassoon D. (1992). Molewlar aspects of regeneration member of the friuled family from drosophila functions as Wingless in developing vertebrate limbs. Dev. Biol. 152,37-49. receptor. Nature 382,225-230. Münsterberg A.E., Kitajewski J., Bumcrot D.A., McMahon A.P. and Breton M., Li Z., Paulin D., Harris J.A., Rieger F., Pincon-Raymond M. Lassar A. (1995) Combinatorial signaling by Sonic hedgehog and and Garcia L. (1995). Myotube driven myogenic recruitment d cells Wnt family mernbers induces myogenic bHLH gene expression in during in vitro myogenesis. Dev. Dynam. 202, 126-136. the somite. Genes Dev. 9,291 1-2922. Chaudary N,, Delay R. and Beam K.G. (1989). Restoration of normal Ordahl C.P. and Le Douarin N. (1991). Two myogenic lineages within function in genetically defective myotubes by spontaneous fusion the developing somite. Development 114,339-353. with fibroblasts. Nature 341, 445-447. Partridge T.A. (1991). Myoblast transfer: a possibie therapy for inherlted Christ B. and Ordahl C. (1995). Early stage of chick somite myopathies? Muscle Newe 14,197-199. development. Anat. Embryol. 191,381396. Patapoutian A., Wold B.J. and Wagner R. (1995). Evidence for Cossu G., Kelly R., Di Donna S., Vivarelli E. and Buckingham M. (1995). developmentally programmed transdifferentiation in mouse Myoblast differentiation during mammalian somitogenesis is esophageal rnuscle. Science 270, 181E1 821. dependent upon a community effect. Proc. Natl. Acad. Sci. USA 92, Rudnicki MA., Schnegelsberg P.N.J., Stead R.H., Braun T., Arnold H.H. 2254-2258. and Jaenisch R. (1993). MyoD or myf-5 is required for the fonnatüon Cossu G., Kelly R., Tajbakhsh S., Di Donna S., Vivarelli E. and of skeletal muscle. Cell71, 1351-1359. Buckingham M. (1996a). Activation of dierent myogenic pathways: Salvatori G., Laitanzi L., Coletla M., Aguanno S., Vivarelli E., Keily R., myf-5 is induced by the neural tube and MyoD by the dorsal Ferrari G., Harris AJ., Mavilio F., Molinaro M. and Co~suG. (1995). ectoderm in mouse paraxial mesodemi. Development 122,429-437. Myogenic conversion of mammalian fibroblasts induced by Cossu G., Tajbakhsh S. and Buckingham M. (1996b). How is dierentiating muscle cdls J. Cell Sci. 108, 2733-2739. myogenesis specified in the embryo? Trends Genet. 12,218-223. Taylor S.M. and Jones P.A. (1979). Multiple new phenotypes induced in Courbin P., Koenig J., Ressouches A,, Beam K.G. and Powell J.A. 10T112 and 3T3 cells treated with 5-azacytidine. Ceil 17,771-7i9. Unorthodox myogenesis in mammals

Tajbakhsh S., Vivarelli G., Cusella-De Angelis O., Rocancourt D., Azacytidine. Muscie Neme 18, 1417-1 426. Buckingham M., and Cossu G. (1994). A population of myogenic Weintraub H., Davis R., Tapscott S., Thayer M., Krause M., Benezra R., cells derived from the mouse neural tube. Neuron 13, 81 3-821. Blackwell T.K., Turner D., Rupp R., Hollenberg S., Zhuang Y. and Watanaki S., Seito T. and Caplan A.I. (1995). Myogenic cells derived Lassar A. (1991). The MyoD gene family: point during from rat bone marrow mesenchimal strem cells exposed to 5'- specification of the lineage. Science 251,761-766.