Experimental Cell Research 281, 101–106 (2002)doi:10.1006/excr.2002.5650

Thrombin Activation of S-Phase Reentry by Cultured PigmentedEpithelial Cells of Adult Newt Iris

Andras Simon1,2 and Jeremy P. Brockes

Department of Biochemistry and Molecular Biology, University College London, London, WC1E 6BT United Kingdom;

Following local injury or tissue removal, regenera-tion in urodele amphibians appears to be dependenton cell cycle reentry and dedifferentiation of postmi-totic, terminally differentiated cells in the remainingtissues. Regeneration of the lens of the eye occurs bythe dedifferentiation of pigmented epithelial cells(PEC) of the iris and their subsequent transdifferen-tiation into lens cells. A key question is how cell cyclereentry is regulated. Here we demonstrate that throm-bin activates S-phase reentry of newt PEC in vitro.Based on these findings, and on previous experimentsshowing that newt skeletal myotubes reenter the cellcycle following thrombin stimulation, we suggest thatthrombin is a critical signal for initiation of vertebrateregeneration. © 2002 Elsevier Science (USA)

INTRODUCTION

Urodele amphibians, such as newts and axolotls,have a unique capacity among vertebrates to regen-erate their body parts as adults. While loss or degen-eration of the neuroretina and other ocular struc-tures leads to visual handicap and eventuallyblindness in most vertebrates, a newt can regenerateits lens and neuroretina. Other newt structureswhich can be replaced after injury include the upperand lower jaws, spinal cord, parts of the centralnervous system, the cardiac musculature, and theintestine [1, 2].

Several studies emphasize the role of the plasticity ofthe differentiated state during amphibian regenera-tion and show that the progenitor cells of the regener-ate are derived from postmitotic, differentiated celltypes (for a review see [3]). Multinucleated skeletalmuscle cells reenter the cell cycle and undergo frag-mentation into proliferating, cycling progeny cells dur-ing limb and tail regeneration in urodeles. These prog-

1 Present address: Medical Nobel Institute, Department of Celland Molecular Biology, Karolinska Institute, Stockholm, Sweden.

2 To whom correspondence and reprint requests should be ad-

dressed. Fax: 46(0)8308374. E-mail: Andras.Simon@cmb.ki.se.

101

eny cells contribute to the growth zone or blastema,from which the cells of the regenerate originate [4–8].Similarly, regeneration of the heart depends on cellcycle reentry and proliferation of cardiomyocytes in thevicinity of the injury [9]. Regeneration of ocular struc-tures such as the lens and neuroretina depends ontransdifferentiation. Upon removal of the lens, pig-mented epithelial cells (PEC) in the dorsal margin ofthe iris enter S phase, lose their pigmentation, and giverise to a new lens. Transdifferentiation to lens is con-fined normally to PEC of the pupillary margin of thedorsal iris and the new lens does not arise from theventral iris [10].

The molecular mechanisms that induce cells toleave the postmitotic arrest have long remainedenigmatic. Recently it was shown that the postmi-totic arrest of newt skeletal myotubes in culture canbe undermined by stimulating them with serum [11].The active component in serum is not a known pro-tein growth factor but an as yet unidentified mole-cule, which is activated by thrombin but which isdistinct from the protease [12]. Thus, a subthresholdconcentration of serum can be activated by digestionwith thrombin, followed by inhibition of residual pro-tease activity. The thrombin-derived activity, whichleads to cell cycle reentry of the newt myotubes, isinactive on mouse myotubes, although it is presentin sera from various mammalian sources [12, 13]. Itis appealing to speculate that following injury, theactivation of thrombin generates a signal, whichnewt cells can transduce and which evokes a keycellular response in regeneration.

Here we have addressed the question of whetherthe thrombin-derived activity may represent a com-mon signal for cells involved in a different context ofregeneration. Since cell cycle reentry of PEC is thefirst step in lens regeneration, we established a cul-ture model in which the PEC are quiescent but re-spond to appropriate molecules that can induce cellcycle reentry. We found that the thrombin-derivedactivity is able to induce S-phase reentry of PEC in

and Medical Nobel Institute, Department of Cell and Molecular Biology, Karolinska Institute, Stockholm, Sweden

vitro.0014-4827/02 $35.00

© 2002 Elsevier Science (USA)All rights reserved.

FIG. 1. Morphology of PEC in culture. (A and D) Phase-contrast image of PEC at 20- and 40-fold magnifications, respectively. (B and E)DAPI (blue) staining of the nuclei at 20- and 40-fold magnifications, respectively. (C and F) BrdU staining of nuclei in S-phase (red) at 20-and 40-fold magnifications, respectively. Scale bar, 50 �m

102 SIMON AND BROCKES

MATERIAL AND METHODS

Animals. Red spotted newts, Notophthalmus viridescens, werepurchased from Charles Sullivan Co. (TN). Animals were kept in tapwater at 19°C and fed once a week.

Preparation of PEC. Animals were anesthetized by immersion in0.1% MS 222 (Sigma) dissolved in tap water. The eyeballs wereremoved and collected in a tissue culture dish containing PBS ad-justed to amphibian osmolarity. The eyeballs were washed in 70%ethanol for 2 min and transferred to Leibowitz-15 (L-15) medium(Gibco), adjusted to amphibian osmolarity, and supplemented with0.5% bovine serum albumin (BSA), penicillin/streptomycin, and L-glutamine. Corneal cells were removed and an anterior eye cup wascreated by cutting the eyeball along the iris. The lens was removedand the iris rings were placed in L-15 medium containing 0.5% BSAand 0.5% dispase grade II (La Roche). If dorsal and ventral cells wereanalyzed separately, the iris rings were cut into dorsal and ventralhalves at this stage. The tissues were incubated at 25°C for 4 h andthe sheet of PEC was separated from the underlying stromal cells.Dissociation of the PEC sheet was achieved by incubation with 0.25%trypsin–EDTA solution for 10 min at room temperature. Trypsin wasdiluted 1/20 in L-15 medium containing 0.5% BSA, and the cells werecentrifuged at 100g for 5 min. Cells were resuspended in L-15 me-dium containing 0.5% BSA, and approximately 600 cells were platedonto collagen type IV- (Sigma) coated wells in 96-well plates. Thirtyto fifty percent of the plated cells attached and spread. With thismethod over 99% of the cells were PEC.

S-phase reentry assay. Seven days after plating, the cells wereshifted to L-15 medium containing various concentrations of fetalcalf serum (FCS; Gibco; Lot F40F8814) with or without thrombin oradditional growth factors. After 3 days, bromodeoxyuridine (BrdU;La Roche) was added to a final concentration of 10 �M. After 18–20h the cells were fixed for 30 s in 2% paraformaldehyde and postfixedfor 5 min in ice-cold methanol. To test the effect of growth factors,

medium containing 0.25% FCS was supplemented with 40 ng/mlpurified mouse EGF (Collaborative Research), human recombinantFGF-2 (Gibco; both factors are kind gifts from Professor JonasFrisen), PDGF-C (a kind gift from Professor Ulf Eriksson), or allthree growth factors simultaneously. The effect of thrombin wasassayed in the presence of 0.25% FCS and 100 �g/ml crude prepa-ration of bovine thrombin (Calbiochem). The effect of the thrombin-derived activity was analyzed essentially as described in [12].Briefly, medium containing 1% FCS was incubated with purifiedbovine thrombin (Enzyme Research Laboratories) for 24 h at 25°C.Thrombin was inactivated using a 50-fold molar excess of D-Phe-Pro-Arg chloromethyl ketone (PPACK; Sigma). The activated mediumcontaining 1% FCS was diluted to a final concentration of 0.25% FCSand added to cells. In control experiments PPACK and purifiedthrombin were added simultaneously in order to prevent activation.Thrombin activity was measured spectrophotometrically using Chro-mosyme TH (La Roche) as substrate according to the manufacturer’srecommendations. Thrombin activity was completely inhibited byPPACK.

Immunocytochemistry and image analysis. BrdU staining wasperformed as described [11] and visualized using Alexa 546-conju-gated anti-mouse IgG-specific secondary antibody (MolecularProbes). The nuclei were stained by DAPI. Cells were observed usinga Nikon inverted microscope and pictures were captured by a colorCCD camera.

Data analysis. The data represent the average of results fromthree independent wells in a representative experiment. Between 90and 300 cells were counted in each well.

RESULTS AND DISCUSSION

We first determined the conditions under which PECare proliferating or quiescent. PEC were prepared and

FIG. 2. S-phase reentry of PEC induced by increasing concentration of FCS. Bars represent the averages of three independent wells ina representative experiment. Dots indicate the results in the individual wells.

103THROMBIN AND S-PHASE REENTRY

plated in serum-free medium. Figure 1 shows the mor-phology of the cells at two different magnifications. Thecells are heavily pigmented, often display a hexagonalshape with clearly visible nuclei, and tend to attach toeach other and to form islets (Figs. 1A, 1B, 1D, and 1E).In the presence of 10% FCS a substantial portion ofPEC reenter S phase as assayed by BrdU incorporation(Figs. 1C and 1F). Most of the cells are quiescent inserum-free medium and as little as 1% FCS inducedDNA replication in more than 20% of PEC (Fig. 2). Themaximal level of S-phase reentry was observed in thepresence of 10% FCS (Fig. 2). It should be noted thatdifferent sources of serum resulted in different levels ofS-phase reentry (data not shown), although the overallpattern was the same as shown in Fig. 2. We thereforeperformed all experiments in the same batch of serum,as specified under Material and Methods.

Next we asked whether thrombin could induce S-phase reentry as observed in the case of newt skeletalmyotubes. We assayed the effect of a crude preparationof bovine thrombin in the presence of 0.25% FCS, sincethe background level at this serum concentration wasrelatively low. Crude thrombin resulted in a 6.5-foldincrease in BrdU-positive cells compared to the control

(Fig. 3). To test whether this effect of thrombin wasdirect or indirect, we activated FCS-containing me-dium with purified thrombin and inactivated the activ-ity of the protease irreversibly by PPACK before addi-tion to the cells [12]. As a control we used mediumwhich was preincubated with thrombin in the presenceof thrombin inhibitor. As shown in Fig. 3, thrombin-activated serum induced a 5.5-fold increase in thenumber of PEC that entered S-phase. The crude prep-aration of thrombin also induced S-phase reentry inserum-free medium, but to a lower extent compared toactivation in the presence of serum (data not shown),suggesting that crude thrombin contains the thrombin-derived activity identified by Tanaka et al. [12].

Lens regeneration in situ is dependent on the PEC ofthe dorsal margin of the iris, while ventral PEC do notparticipate in formation of the new lens [14]. To testwhether this difference is correlated to thrombin re-sponsiveness, we separated the iris into dorsal andventral halves prior to the removal of PEC and seededdorsal and ventral cells in separate wells. We did notfind any difference between dorsal and ventral cellswith respect to responsiveness to subthreshold concen-tration of serum, which was activated with pure throm-

FIG. 3. Thrombin induces S-phase reentry of PEC. “Crude thr” indicates that medium containing 0.25% FCS was supplemented with acrude preparation of bovine thrombin as described under Material and Methods. “0.25% FCS non-act” indicates that medium containing0.25% FCS was preincubated with both purified bovine thrombin and thrombin inhibitor for 24 h before addition to cells. “0.25% FCS act”indicates that medium containing 0.25% FCS was preincubated with purified thrombin for 24 h and thrombin activity was irreversiblyinhibited before addition to cells. Bars represent the average of three independent wells in a representative experiment. Dots indicate theresults in the individual wells.

104 SIMON AND BROCKES

FIG. 4. Both dorsal and ventral PEC reenter S phase upon stimulation with thrombin-activated serum. Labels should be interpreted as in Fig.3. Bars represent the average of three independent wells in a representative experiment. Dots indicate the results in the individual wells.

FIG. 5. Growth factors induce S-phase reentry of PEC. “GF-MIX” indicates that all three growth factors were added to the medium. Barsrepresent the average of three independent wells in a representative experiment. Dots indicate the results in the individual wells.

105THROMBIN AND S-PHASE REENTRY

bin. Dorsal and ventral cells showed identical re-sponses compared to control (Fig. 4). These results arein agreement with previous results by Eguchi et al.[15], which showed that both dorsal and ventral cellsform lentoid bodies in vitro, despite the fact that onlydorsal cells participate in lens regeneration in vivo.

As serum also contains mitogenic growth factors, wewanted to see whether purified growth factors couldinduce S-phase reentry of PEC. We tested three differ-ent growth factors, EGF, PDGF-C, and FGF-2. Allthree induced cell cycle reentry to approximately thesame level either alone (PDGF-C, 9.4-fold; FGF-2, 11-fold; EGF, 10.7-fold), or in combination (10.5-fold), sug-gesting that PEC express receptors for these growthfactors (Fig. 5).

The phenomenon of Wolffian regeneration [16] is ademonstration of the reprogramming and transdiffer-entiation of a fully differentiated cell type, the PEC.Transdifferentiation begins with cell cycle reentry, andour results show that PEC respond to thrombin stim-ulation in culture in the same way as do skeletal myo-tubes. While the implications of this finding would stillrequire extension from in vivo experiments, our resultsindicate that the thrombin-activated pathway mayrepresent a common signal for postmitotic cells, whichare reactivated during regeneration in newts. Variousgrowth factors in addition to thrombin evoke DNAreplication in PEC. This contrasts with the case ofnewt skeletal myotubes, which are refractory to growthfactor stimulation [11, 12]. One simple interpretationof these data could be that PEC have receptors forthese growth factors, whereas the myotubes have re-ceptors for only the thrombin-activated factor. None-theless, these observations are in agreement with sev-eral previous studies which underline the role ofmitogenic growth factors in proliferation and transdif-ferentiation of PEC to lens or lentoid bodies [17–19].

The culture system described here provides an op-portunity for systematic analysis of the proliferativepotential of PEC under different conditions. It may beuseful in future experiments to investigate how cellcycle reentry and transdifferentiation are coupled toeach other and how specific agents may influence theircourse. One important task for the future is also todetermine whether thrombin is able to induce cells tocomplete the mitotic cycle.

We are most grateful to Y. Imokawa for help with initial prepara-tion of PEC and for helpful discussions. We thank K. Agata for adviceon culture conditions, E. Tanaka for suggestions, S. Sandberg forcomments on the manuscript, and J. Frisen and U. Eriksson forproviding growth factors. The financial support from an MRC pro-

gram grant to J.P.B. and to A.S. from The Wenner–Gren Foundation,The Swedish Research Council, Carl Tryggers Stiftelse, Åke WibergsStiftelse, Magnus Bergvalls Stiftelse, and Stiftelsen Lars HiertasMinne is greatly appreciated.

REFERENCES

1. Stocum, D. L. (1995). “Wound Repair, Regeneration, and Arti-ficial Tissues,” Landes, Austin, TX.

2. Wallace, H. (1981). “Vertebrate Limb Regeneration,” Wiley,New York.

3. Brockes, J. P. (1997). Amphibian limb regeneration: Rebuildinga complex structure. Science 276, 81–87.

4. Echeverri, K., Clarke, J. D., and Tanaka, E. M. (2001). In vivoimaging indicates muscle fiber dedifferentiation is a major con-tributor to the regenerating tail blastema. Dev. Biol. 236, 151–164.

5. Hay, E. D. (1959). Electron microscopic observations of musclededifferentiation in regenerating Amblystoma limbs. Dev. Biol.1, 555–585.

6. Hay, E. D., and Fischman, D. A. (1961). Origin of the blastemain regenerating limbs of the newt Triturus viridescens. Dev.Biol. 3, 26–59.

7. Kumar, A., Velloso, C. P., Imokawa, Y., and Brockes, J. P.(2000). Plasticity of retrovirus-labelled myotubes in the newtlimb regeneration blastema. Dev. Biol. 218, 125–136.

8. Lo, D. C., Allen, F., and Brockes, J. P. (1993). Reversal of muscledifferentiation during urodele limb regeneration. Proc. Natl.Acad. Sci. USA 90, 7230–7234.

9. Oberpriller, J. O., and Oberpriller, J. C. (1974). Response of theadult newt ventricle to injury. J. Exp. Zool. 187, 249–253.

10. Eguchi, G. (1988). Cellular and molecular background of Wolf-fian lens regeneration. Cell Differ. Dev. 25(Suppl.), 147–158.

11. Tanaka, E. M., Gann, A. A., Gates, P. B., and Brockes, J. P.(1997). Newt myotubes reenter the cell cycle by phosphoryla-tion of the retinoblastoma protein. J. Cell Biol. 136, 155–165.

12. Tanaka, E. M., Drechsel, D. N., and Brockes, J. P. (1999).Thrombin regulates S-phase re-entry by cultured newt myo-tubes. Curr. Biol. 9, 792–799.

13. Velloso, C. P., Simon, A., and Brockes, J. P. (2001). Mammalianpostmitotic nuclei reenter the cell cycle after serum stimulationin newt/mouse hybrid myotubes. Curr. Biol. 11, 855–888.

14. Yamada, T. (1977). Control mechanisms in cell-type conversionin newt lens regeneration. Monogr. Dev. Biol. 13, 1–126.

15. Eguchi, G., Abe, S. I., and Watanabe, K. (1974). Differentiationof lens-like structures from newt iris epithelial cells in vitro.Proc. Natl. Acad. Sci. USA 71, 5052–5056.

16. Wolff, G. (1895). Entwicklungsphysiologische Studien. I. DieRegeneration der Urodelelinse. Arch. Entw.-Mech. Org. 1, 380–390.

17. Del Rio-Tsonis, K., Jung, J. C., Chiu, I. M., and Tsonis, P. A.(1997). Conservation of fibroblast growth factor function in lensregeneration. Proc. Natl. Acad. Sci. USA 94, 13701–13706.

18. Kosaka, M., Kodama, R., and Eguchi, G. (1998). In vitro culturesystem for iris-pigmented epithelial cells for molecular analysisof transdifferentiation. Exp. Cell Res. 245, 245–251.

19. McDevitt, D. S., Brahma, S. K., Courtois, Y., and Jeanny, J. C.(1997). Fibroblast growth factor receptors and regeneration ofthe eye lens. Dev. Dyn. 208, 220–226.

Received June 18, 2002Revised version received August 15, 2002Published online October 11, 2002

106 SIMON AND BROCKES