Christopher J. McGann*, Shannon J. Odelberg*^,@ and Mark T. Keating,§,¶,||,^@

* Department of Internal Medicine, Division of Cardiology, University of Utah, Salt Lake City, UT 84112; and
§ Howard Hughes Medical Institute, ¶ Department of Cell Biology, Harvard Medical School, and ^|| Department of Cardiology, Children's Hospital, Boston, MA 02115

C.J.M. and S.J.O. contributed equally to this work.

^@ To whom reprint requests should be addressed at:
University of Utah, 15 North 2030 East, Room 2100, Salt Lake City, UT 84112-5330 (for S.J.O.;
E-mail:  odelberg@howard.genetics.utah.edu) or
Children's Hospital, 320 Longwood Avenue, 1261 Enders Building, Boston, MA 02115 (for M.T.K.;
E-mail:  mkeating@genetics.med.harvard.edu)

Newts are capable of regenerating several anatomical structures and organs, including their limbs. This remarkable regenerative capacity is thought to depend on cellular dedifferentiation. Terminally differentiated mammalian cells, by contrast, are normally incapable of reversing the differentiation process. Several factors could explain the absence of cellular dedifferentiation in mammals: (i) inadequate expression of genes that initiate dedifferentiation; (ii) insufficient intracellular signaling pathways; (iii) irreversible expression of differentiation factors; and (iv) structural characteristics that make dedifferentiation impossible. To investigate the causes underlying the lack of cellular plasticity in mammalian cells, we examined the effect of an extract derived from newt regenerating limbs on terminally differentiated mouse C2C12 myotubes. Approximately 18% of murine myotubes reentered the cell cycle when treated with regeneration extract, whereas 25% of newt myotubes exhibited cell cycle reentry. The muscle differentiation proteins MyoD, myogenin, and troponin T were reduced to undetectable levels in 15-30% of treated murine myotubes. We observed cellular cleavage in 11% of the treated murine myotubes and approximately 50% of these myotubes continued to cleave to produce proliferating mononucleated cells. These data indicate that mammalian myotubes can dedifferentiate when stimulated with the appropriate factors and suggest that one mechanism preventing dedifferentiation of
mammalian cells is inadequate spatial or temporal expression of genes that initiate dedifferentiation.

Adult urodele amphibians and teleost fish can replace lost anatomical structures through a process known as epimorphic regeneration. Morgan (1) coined the term epimorphosis to refer to the regenerative process in which cellular proliferation precedes the development of a new anatomical structure. An adult urodele (e.g., a newt or axolotl) is capable of regenerating its limbs, spinal cord, heart ventricle, tail, retinas, eye lenses, and upper and lower jaws (2-4), whereas teleost fish can regenerate their fins and spinal cord (5, 6). These remarkable regenerative capabilities are lacking in mammals and other vertebrates.

The molecular and cellular mechanisms that govern epimorphic regeneration are incompletely defined. Following limb amputation, epithelial cells begin to migrate across the amputation site to form a wound epithelium (WE) a few layers thick within a single day of amputation. This wound epithelium thickens in response to continued epithelial cell migration and within days forms the mature apical epithelial cap (AEC) (7). The internal stump cells underlying the WE/AEC begin to dedifferentiate in response to undefined signals found in the early limb regenerate (8-15). These dedifferentiated cells then proliferate to form a mass of progenitor and pluripotent cells, known as the regeneration blastema, which harbors the cells that will later redifferentiate to form the regenerated limb. Although cellular dedifferentiation has been demonstrated in newts, terminally differentiated mammalian cells are normally thought to be incapable of reversing the differentiation process (16, 17). Several mechanisms could explain the lack of cellular plasticity in mammalian cells: (i) the extracellular factors that
initiate dedifferentiation are not adequately expressed following amputation; (ii) the intrinsic cellular signaling pathways for dedifferentiation are absent; (iii) differentiation factors are irreversibly expressed in mammalian cells; and (iv) structural characteristics of mammalian cells make dedifferentiation impossible.

Here, we demonstrate that mouse myotubes can dedifferentiate when stimulated with an extract prepared from newt regenerating limb tissue. Mouse myotubes reenter the cell cycle, exhibit reduced levels of muscle differentiation proteins, and cleave to form smaller myotubes or proliferating, mononucleated cells. We demonstrate that proteins are an essential component of this dedifferentiation signal. These results indicate that mammalian cells can dedifferentiate when exposed to the appropriate factors.
...
Discussion:

Cellular dedifferentiation, a phenomenon central to epimorphic regeneration in newts, is not normally observed in terminally differentiated mammalian myotubes. Mouse myotubes are incapable of reentering the cell cycle unless they have been genetically altered or treated with myoseverin, a microtubule-binding purine (20-27). We have previously demonstrated that ectopic expression of msx1, which encodes a homeobox-containing transcriptional repressor, can induce mouse myotubes to dedifferentiate to mononucleated cells that possess the properties of stem cells (28). Here, we show that unaltered mammalian myotubes can dedifferentiate when exposed to the same signals that induce dedifferentiation of newt cells during the early phases of limb regeneration. Mouse myotubes treated with a limb regeneration extract reenter the cell cycle, exhibit reduced levels of muscle differentiation proteins, and cleave to produce smaller myotubes or proliferating, mononucleated cells. We also demonstrate, for the first time, that cultured newt myotubes can respond to regeneration extract by reentering the cell cycle and by cleaving to form smaller myotubes or mononucleated cells. Thus, the cellular plasticity that was thought unique to urodeles can now be extended to mammals.

All of the nuclei within a given newt myotube responded concordantly to the newt limb regeneration extract by either reentering the cell cycle or by remaining in a quiescent state, whereas most often mouse myotube nuclei did not exhibit such concordance. In mouse myotubes that responded to the regeneration extract, ~64% of the nuclei reentered the S phase, whereas the remaining nuclei retained their quiescence. Although we do not yet understand the reason for this S phase discordance within a single myotube, it must be noted that similar results were obtained following serum stimulation of newt/mouse hybrid myotubes (29). In the hybrid myotube study, the presence of the newt nuclei within the hybrid myotube made some, but not all, of the neighboring mouse nuclei competent to respond to serum and thrombin stimulation. This difference between newt and mouse myotubes may reflect the fact that newt myotubes appear to be primed for cellular dedifferentiation, as suggested by their response to serum or thrombin stimulation (30-32) and their minimal response to differentiation medium or nonregenerating limb extract in this present study.

We did not observe a size difference between myotubes that exhibited cell cycle reentry and those that did not. However, the procedure that we used to isolate myotubes selected for smaller multinucleated cells containing 4-23 nuclei (mean = 7.4 ± 4.5_sd). Therefore, at this time, we do not know whether larger myotubes are capable of reentering the cell cycle in response to regeneration extract stimulation.

It has previously been demonstrated that cultured newt myotubes can dedifferentiate when transplanted back into the blastema of a regenerating newt limb (14, 15). Approximately 10-15% of the transplanted newt myotubes responded by cleaving to form mononucleated cells. In our study, newt regeneration extract induced cleavage of ~16% of newt myotubes and 11% of the mouse myotubes, a finding that is remarkably consistent with the previous in vivo studies. The fact that the in vitro results recapitulate the in vivo studies suggest that the limb regeneration extract may have retained all of the dedifferentiation activity normally inherent in the regenerating limb and that the in vitro system may provide a reliable method for studying the dedifferentiation process.

Our findings are inconsistent with the interpretation that the proliferating cells arise from contaminating reserve cells in the myotube preparation. First, a careful photographic record was kept of the entire dedifferentiation process and no unexplained mononucleated cells suddenly appeared, as might be expected if hidden mononucleated cells were responsible for our observations. Second, we observed myotubes as they were beginning to pull apart and cleave (Fig. 3A, day 1). Third, cleavage events often lead to smaller multinucleated myotubes, which cannot be mistaken for contaminating mononucleated cells (Fig. 3A, day 2). Finally, if unobserved mononucleated cells were responsible for our observations, we would expect approximately the same number of apparent cleavage events in the controls. As noted above, no cleavage events of control myotubes were observed, whereas 11% of the treated myotubes cleaved. This difference in cleavage
frequency is statistically significant at P = 0.0001. Therefore, we conclude that newt limb regeneration extract induces mouse myotubes to cleave to form smaller myotubes or proliferating, mononucleated cells.

These studies indicate that mammalian cells have retained the intracellular signaling pathways required for dedifferentiation. Therefore, it is likely that mammals fail to exhibit in vivo cellular dedifferentiation because they lack the signal that initiates the dedifferentiation process. These conclusions differ from those reached in a previous study (32). Based on the observation that newt myotubes reenter S phase when stimulated with serum or thrombin, whereas mammalian myotubes are refractory to these treatments (30-32), it was suggested that mammalian myotubes were intrinsically incapable of responding to dedifferentiation signals. Our results can be reconciled with these previous studies by recognizing that our work was performed by using crude extracts from the entire regenerating portion of the limb, whereas the previous studies (32) were performed by using either serum, crude thrombin, fractionated thrombin, or mixtures of thrombin and serum. The serum factor, which is thought to be a thrombin-activated ligand, induces newt myotubes to reenter the cell cycle but does not promote a continuation of the cell cycle through mitosis to produce mononucleated cells. Therefore, other factors found
in the regenerating limb are required for complete dedifferentiation and may either cooperate with the thrombin-activated ligand or entirely substitute for it.

Our data suggest that proteins are required components of the dedifferentiation signal. This conclusion is based on the following observations: (i) the signal is contained in the soluble fraction of the extract; (ii) it is heat labile; (iii) it is sensitive to repeated freeze-thaw cycles; and (iv) the signal is abolished by treatment with the protease trypsin. We speculate that these factors are soluble extracellular proteins acting as ligands to activate receptors that transduce dedifferentiation signals through receptive cells, such as myotubes.

During newt limb regeneration, cellular dedifferentiation is followed by proliferation and redifferentiation of the blastemal cells. The patterning and differentiation signals involved in reforming the lost limb may be regulated by the regeneration blastema. If so, the primary obstacle preventing induced epimorphic regeneration in mammals may be the inability to form a mass of dedifferentiated cells from local tissues. If this hypothesis is correct, identifying essential dedifferentiation factors may lead to methods for enhancing the regenerative process in mammals.

Acknowledgements:

We thank Jeremy Brockes for the newt A1 limb cells. We thank Diana Stafforini and Kirk Thomas for helpful suggestions. The Howard Hughes Medical Institute supported this work.

Abbreviation:

DM, differentiation medium.

References:
 
1. Morgan, T. H. (1901) Regeneration (Columbia University, New York).
 
2. Becker, R. O. , Chapin, S. & Sherry, R. (1974) Nature (London) 248, 145-147.
 
3. Davis, B. M. , Ayers, J. L. , Koran, L. , Carlson, J. , Anderson, M. C. & Simpson, S. B., Jr. (1990) Exp. Neurol. 108, 198-213.
 
4. Brockes, J. P. (1997) Science 276, 81-87.
 
5. Zottoli, S. J. , Bentley, A. P. , Feiner, D. G. , Hering, J. R. , Prendergast, B. J. & Rieff, H. I. (1994) Prog. Brain Res. 103, 219-228.
 
6. Johnson, S. L. & Weston, J. A. (1995) Genetics 141, 1583-1595.
 
7. Christensen, R. N. & Tassava, R. A. (2000) Dev. Dyn. 217, 216-224.
 
8. Chalkley, D. T. (1954) J. Morphol. 94, 21-70.
 
9. Thornton, C. S. (1957) J. Exp. Zool. 134, 357-382.
 
10. Bodemer, C. W. & Everett, N. B. (1959) Dev. Biol. 1, 327-342.
 
11. Hay, E. D. & Fischman, D. A. (1961) Dev. Biol. 3, 26-59.
 
12. Thornton, C. S. & Thornton, M. T. (1965) Experentia 21, 146-151.
 
13. Steen, T. P. (1968) J. Exp. Zool. 167, 49-78.
 
14. Lo, D. C. , Allen, F. & Brockes, J. P. (1993) Proc. Natl. Acad. Sci. USA 90, 7230-7234.
 
15. Kumar, A. , Velloso, C. P. , Imokawa, Y. & Brockes, J. P. (2000) Dev. Biol. 218, 125-136.
 
16. Andres, V. & Walsh, K. (1996) J. Cell Biol. 132, 657-666.
 
17. Walsh, K. & Perlman, H. (1997) Curr. Opin. Genet. Dev. 7, 597-602.
 
18. Ferretti, P. & Brockes, J. P. (1988) J. Exp. Zool. 247, 77-91.
 
19. Guo, K. , Wang, J. , Andres, V. , Smith, R. C. & Walsh, K. (1995) Mol. Cell. Biol. 15, 3823-3829.
 
20. Endo, T. & Nadal-Ginard, B. (1989) in Cellular and Molecular Biology of Muscle Development, eds. Stockdale, F. & Kedes, L. (Liss, New York), pp. 95-104.
 
21. Iujvidin, S. , Fuchs, O. , Nudel, U. & Yaffe, D. (1990) Differentiation 43, 192-203.
 
22. Schneider, J. W. , Gu, W. , Zhu, L. , Mahdavi, V. & Nadal-Ginard, B. (1994) Science 264, 1467-1471.
 
23. Tiainen, M. , Spitkovsky, D. , Jansen-Durr, P. , Sacchi, A. & Crescenzi, M. (1996) Mol. Cell. Biol. 16, 5302-5312.
 
24. Novitch, B. G. , Mulligan, G. J. , Jacks, T. & Lassar, A. B. (1996) J. Cell Biol. 135, 441-456.
 
25. Endo, T. & Nadal-Ginard, B. (1998) J. Cell Sci. 111, 1081-1093.
 
26. Latella, L. , Sacco, A. , Pajalunga, D. , Tiainen, M. , Macera, D. , D'Angelo, M. , Felici, A. , Sacchi, A. &
Crescenzi, M. (2001) Mol. Cell. Biol. 21, 5631-5643.
 
27. Rosania, G. R. , Chang, Y. T. , Perez, O. , Sutherlin, D. , Dong, H. , Lockhart, D. J. & Schultz, P. G. (2000)  Nat. Biotechnol. 18, 304-308.
 
28. Odelberg, S. J. , Kollhoff, A. & Keating, M. T. (2000) Cell 103, 1099-1109.
 
29. Velloso, C. P. , Simon, A. & Brockes, J. P. (2001) Curr. Biol. 11, 855-858.
 
30. Tanaka, E. M. , Gann, A. A. , Gates, P. B. & Brockes, J. P. (1997) J. Cell Biol. 136, 155-165.
 
31. Tanaka, E. M. & Brockes, J. P. (1998) Wound Repair Regen. 6, 371-381.
 
32. Tanaka, E. M. , Drechsel, D. N. & Brockes, J. P. (1999) Curr. Biol. 9, 792-799. 

1. Czihak G, "Evidence for Inductive Properties of the Micromere RNA in Sea Urchin Embryos", Naturwissenschaften, vol. 52, p. 14 (1965).

2. Frenster JH, "Nuclear Polyanions as Derepressors of Synthesis of Ribonucleic Acid", Nature 206: 680 (1965).

3. Kolodny GM, "Evidence for Transfer of Macromolecular RNA Between Mammalian Cells in Culture", Exp. Cell Res. 65: 313 (1971).

4. Frenster JH, "Activation of DNA Transcription Within Repressed Chromatin", 14th John Innes Symposium, September 5-8, 2001.