paperpododo6

8/6/2019 paperpododo6

1/16

Mammalian Regeneration andRegenerative Medicine

Ken Muneoka,* Christopher H. Allan, Xiaodong Yang, Jangwoo Lee,and Manjong Han

Mammals are generally considered to be poor regenerators, yet there area handful of mammalian models that display a robust ability to regener-ate. One such system is the regenerating tips of digits in both humansand mice. In vitro studies of regenerating fetal human and mouse digittips display both anatomical and molecular similarities, indicating that themouse digit is a clinically relevant model. At the same time, genetic stud-ies on mouse digit tip regeneration have identified signaling pathwaysrequired for the regeneration response that parallel those known to beimportant for regeneration in lower vertebrates. In addition, recent stud-ies establish that digit tip regeneration involves the formation of a blas-tema that shares similarities with the amphibian blastema, thus estab-lishing a conceptual bridge between clinical application and basic researchin regeneration. In this review we discuss how the study of endogenousregenerating mammalian systems is enhancing our understanding ofregenerative mechanisms and helping to shed light on the developmentof therapeutic strategies in regenerative medicine. Birth DefectsResearch (Part C) 84:265280, 2008. VC 2008 Wiley-Liss, Inc.

Key words: regeneration; mammal;digit; finger; blastema; ossification

INTRODUCTION

In this posthuman genome/post-animal cloning era of modern biol-

ogy, many have turned theirattention to the prospect of con-trolling the regeneration of tissuesor organs that do not regeneratein humans. Successes in this newfield of Regenerative Medicinewould have enormous impact oncurrent medical practices and, aswell, on the general quality ofhuman life. Regenerative medicineis strongly influenced by break-throughs in our understanding oforgan and tissue formation duringembryogenesis on the one hand,

and on the bodys endogenousability to repair wounds followinginjury on the other. The goal of re-

generative medicine is to be ableto replace adult body parts ondemand, and to this end we canidentify three general avenuesbeing taken for the developmentof novel regeneration therapies.The first is a cell based approach.This approach has grown largelyfrom the successes in the use ofhematopoietic stem cells in cellreplacement therapies for the cureof blood diseases (Bhattacharyaet al., 2008). The potential toexpand into other organ systems

is driven by (1) the potential ofadult stem cells to participate inthe formation of various organsystems when introduced in earlyembryos (Jiang et al., 2002), (2)the feasibility of transformingadult cells into pluripotent stemcells (Yamanaka, 2008), and (3)

the isolation and characterizationof adult multipotent stem cellsfrom virtually every tissue of thebody (Crisan et al., 2008). Thesecond approach involves a bioen-gineering strategy in which a sub-strate or scaffold is introducedthat can either be infiltrated byhost cells (Badylak, 2007), orseeded with selected cells beforeimplantation (Howard et al.,2008). This approach includes thein vitro engineering of specific tis-sues for use in transplantations,

and in doing so sidesteps theproblems associated with tissuemorphogenesis and patterningthat are key to the successfulregeneration of injured bodyparts. However, it does introducea secondary problem of integrat-ing an engineered tissue with thehost that still needs to beaddressed (Khan et al., 2008).

The third approach is to studynaturally regenerating models forcomparison with nonregeneratinginjury wounds to discover critical

REVIEW

VC 2008 Wiley-Liss, Inc.

Birth Defects Research (Part C) 84:265280 (2008)

Ken Muneoka is from Division of Developmental Biology, Department of Cell and Molecular Biology, Tulane University, New Orleans,Louisiana and The Center for Bioenvironmental Research, Tulane University, New Orleans, Louisiana.Christopher H. Allan is from Department of Orthopaedics and Sports Medicine, University of Washington, Seattle, Washington.Xiaodong Yang, Jangwoo Lee, and Manjong Han are from Division of Developmental Biology, Department of Cell and MolecularBiology, Tulane University, New Orleans, Louisiana.

Grant sponsor: National Institutes of Health; Grant numbers: R01-HD043277; P01-HD022610

*Correspondence to: Ken Muneoka, Division of Developmental Biology, Department of Cell and Molecular Biology, Tulane University,New Orleans, LA 70118. E-mail: [email protected]

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/bdrc.20137


2/16

factors necessary for a regenera-tion response, and this is the pri-mary topic of this review. Thisapproach represents a long historyof experimental inquiry focusedlargely on invertebrate models that

have enhanced regenerative ability(Sanchez Alvarado and Tsonis,2006) and selected vertebrategroups that possess the ability toregenerate structures such as thelimb and tail (Brockes and Kumar,2005; Gardiner, 2005). Included inthis category are studies on thedeveloping appendages of mam-mals, birds, and frogs which pos-sess regenerative ability that is lostas the animal matures (Mulleret al., 1999). Although the leapbetween regenerating systems and

human therapies may seem large,there is substantial evidence thatsignaling pathways important forregeneration have been conservedthrough evolution (Sanchez Alvar-ado, 2000; Brockes et al., 2001),

and there are examples of specificsignaling pathways or genes thatare important for regeneration inboth traditionally regenerating andnonregenerating organisms (Tayloret al., 1994; Yokoyama et al.,2000; Beck et al., 2003; Han et al.,2003). In addition, there are a

handful of mammalian systemsthat can successfully regenerate.Studies of these systems are pro-ducing important insight into thefeasibility of an enhanced endoge-nous regenerative response inhumans, and also in the design ofalternative strategies in regenera-tive medicine, particularly toaddress the problem of integrationwith host tissues (see Pendegrasset al., 2006). In this review wehighlight studies on appendageregeneration in mammals, with

particular emphasis on the regen-erating digit tip, in the context ofhow such efforts may impact thedevelopment of successful thera-pies in regenerative medicine.

MAMMALIAN MODELS OF

APPENDAGE

REGENERATION

Appendage regeneration has beenstudied primarily in amphibians

with a focus on limb or tail regen-eration in adult urodeles (Brockesand Kumar, 2005; Tanaka, 2003)or larval anurans (Slack et al.,2008). While mammals lack simi-lar regenerative capabilities, there

are a handful of model systems inwhich a variation of appendageregeneration has been described,and these models provide aglimpse at the limitations andpotential for regeneration inhumans. These include the closureof excisional tissues in earsfollowing hole punch in rabbitsand mice (Metcalfe et al., 2006),the annual regeneration of antlersin deer (Price et al., 2005; Kier-dorf et al., 2007), and the regen-eration of amputated digit tips

known to occur in humans androdents (Han et al., 2005).Although the regenerative capa-bility of mammals does not com-pare with that of amphibians, it iscritical to keep in mind that these

mammalian models provideinsight into how successful regen-eration can be accomplishedwithin the context of a warm-blooded terrestrial animal withsimilarities to humans. Lessonslearned from such examples arelikely to provide important insight

into how to effectively modify thehuman wound environment toelicit an enhanced regenerativeresponse, or to establish a func-tional interface with a bioengi-neered or artificial organ or struc-ture. The mammalian ear punchand the deer antler models haverecently been reviewed (Priceet al., 2005; Metcalfe et al.,2006; Kierdorf et al., 2007), so avery brief introductory overviewis provided here and the reader isdirected to these excellent re-

views. We will focus most of ourattention on the human andmouse digit models that we haveexplored over the past few years.On the one hand the mouse digitmodels share anatomical and mo-lecular similarities with humanfingertip regeneration making itclinically relevant, and on theother hand, digit tip regeneration iscomparable to other well studied

regeneration models in that woundhealing culminates in the formation

of a blastema that mediates theregeneration response.

EARS AND DEERS

The ears of some mammals are

able to undergo scar-free healingand regeneration after an exci-sional hole punch that removes acylindrical mass of tissue includingepidermis, dermis, muscle andcartilage. This response in mam-mals was first characterized inrabbit ears and later shown to bea characteristic not restricted tolagomorphs (Williams-Boyce andDaniel, 1986). In recent yearsresearch on ear hole punch regen-eration has been stimulated by thefinding that different mouse

strains display variability in thisregeneration response (Clarket al., 1998; Kench et al., 1999;Li et al., 2001), raising the possi-bility that the genetic basis of thisvariation might be uncovered

(Heber-Katz, 1999). In mice, a 2-mm diameter hole punch under-goes re-epithelization that in-volves epidermal closure from thetwo opposite surfaces of the ear,and centripetal filling in of the holeis driven by growth of a blastema-like structure that forms between

the existing ear tissue and thewound epidermis. The MRL strain,also known as the healer strain,displays the highest level of regen-erative ability described (Heber-Katz, 1999); however, even in thisstrain complete regeneration doesnot always occur (Rajnoch et al.,2003). The regeneration process ischaracterized by a wound healingresponse that involves formationof a specialized wound epidermisthat integrates the epidermallayers from the inner and outer

ear surfaces. This wound healingresponse is influenced by the na-ture of the injury and the degreeof trauma ear tissues experience(Rajnoch et al., 2003). This hasled to the suggestion that a highdegree of trauma leads to a regen-erative response, whereas a lowdegree of trauma results in a rep-arative response (Metcalfe et al.,2006). The existence of a dichoto-

mous switch that triggers an epi-morphic regeneration response

266 MUNEOKA ET AL.

Birth Defects Research (Part C) 84:265280, (2008)


3/16

versus a wound healing responseis an important consideration inthe development of regenerationtherapies and needs to be furtherinvestigated. The wound healingprocess results in the formation of

a blastema-like structure that iscontinuous around the margin ofthe wound. The blastema is com-posed of proliferating cells and asit grows centripetally, the punchhole eventually closes and the epi-thelial surfaces fuse. Re-differen-tiation of ear cartilage occurs byextension of the existing cartilagesheet or by the differentiation ofcartilaginous islands at the base ofthe blastema (Rajnoch et al.,2003). In some cases ectopic boneformation has been described dur-

ing redifferentiation (William-Boyce and Daniel, 1986), suggest-ing that the cells involved in thisresponse are multipotent and re-sponsive to the wound environ-ment.

The annual regeneration of deerantlers represents another exam-ple of a naturally occurring regen-erative response in mammals.Deer, along with many of their rel-atives, shed their antlers annuallyonly to have them undergo a com-plex regenerative response that

involves outgrowth from the pedi-cle, a bilateral bony protrusion ofthe frontal bone. Primary antlerdevelopment occurs during pu-berty and in response to circulat-ing levels of sex steroids. Antlerdevelopment involves the initialformation of the pedicle from aspecialized periosteum associatedwith the frontal bone. Pedicle out-growth and elongation involvesmany developmental processesbeginning with intramembranousossification to initiate pedicle for-

mation, and is followed by the for-mation of a distal endochondralgrowth zone with proximal ossifi-cation that continues until the ant-ler is fully developed (Price et al.,2005). Antler shedding or castingnormally occurs in the spring andis mediated by enhanced osteo-clast activity at the distal region ofthe pedicle, leaving an open pedi-cle wound that forms a scab (Goss

et al., 1992). The regenerationprocess is initiated by the closure

of the epidermis over the pediclewound to form a wound epithe-lium. Beneath the wound epithe-lium is a dermal layer overlaying afibrous perichondral layer that iscontinuous with the periosteallayer of proximal bone. Just proxi-mal to the fibrous perichondriumat the distal end of the regenerat-ing antler is a mesenchymalgrowth zone (also called thereserve mesenchyme) where cells

are actively dividing, and is argu-ably the antler blastema (Fau-cheux et al., 2004; Li et al., 2005;Kierdorf et al., 2007). The cells ofthe mesenchymal growth zoneappear to be derived from thepedicle perichondrium and thesecells display characteristics ofmesenchymal stem cells (Rolfet al., 2008). Proximal to the mes-enchymal growth zone are regions

containing, in distal to proximalorder, chondroprogenitor cells,

chondroblasts, hypertrophic chon-drocytes, and bone, undergoingendochondral ossification similarto the developing antler. In addi-tion to bone tissue, antler regen-eration involves the regenerationof epidermis and its derivatives,dermis, and vasculature.

FINGERTIP REGENERATION

IN HUMANS

The prospect of developing strat-egies for enhancing regenerativeability in humans is encouraged byclinical observations that thehuman fingertip is capable of a re-generative response (Fig. 1A, B).While the initial descriptions of fin-gertip regeneration were made inchildren (Douglas, 1972; Illing-worth, 1974), they were followedby descriptions of fingertip regen-eration in adults as well (Leeet al., 1995). The key for human

Figure 1. Fingertip regeneration in humans. (A, B) A fingertip injury of a 7-year oldgirl resulted in an amputation at the base of the nail. The injury was treated conserva-tively with dressing changes and after 8 weeks the fingertip regenerated (From StocumDL. Regen Biol Med 2006, 394, copyright 2006, Elsevier, reproduced by permission.).(CE) A fingertip injury of a 2-year old child resulted in an amputation at a level proximalto the nail. (C) Radiograph at the time of injury indicated that the level of amputation wasthrough the proximal region of the terminal phalangeal bone. (D) The amputation injurywas treated conservatively with dressing changes and after 10 months the fingertiphealed without significant scarring and a nail rudiment was present. The fingertip had anormal contour and sensibility had returned. (E) Radiographic evidence after healingshowed that there was no re-growth of the terminal phalangeal bone and indicated that aregenerative response was not stimulated. (From Han M, Yang X, Lee J, et al. Dev Biol2008, 315:125135, Copyright 2008, Elsevier, reproduced by permission.).

MAMMALIAN REGENERATION 267



4/16

fingertip regeneration is to treatthe amputation wound in a con-servative manner, e.g., clean anddress the wound so as to allow itto heal by secondary intention(i.e., without assisted wound clo-

sure). It is thought that such con-servative treatment in humanspromotes the formation of awound epidermis that is requiredfor the initiation of a regenerativeresponse. Appendage regenerationin amphibians and in embryoniclimbs of birds has been shown tohave a similar requirement of aspecialized wound epidermis for asuccessful regenerative response(see Muller et al., 1999). Never-theless, the actual closure of theamputation wound itself in humans

is a very slow process, and muchof the regenerative growth andremodeling associated with theregeneration response occursbefore the completion of woundhealing. Thus, it is safe to conclude

that continuity of the wound epi-dermis is not a prerequisite forregeneration, but it is unclear nowwhat role the wound epidermisplays in human regeneration.

Fingertip regeneration inhumans is reported to be re-stricted to the distalmost, or ter-

minal phalangeal element andassociated with the nail organ(Illingworth, 1974). The clinicaluse of the term regeneration is notstrictly defined so the clinicaldescription of amputation injuriesoften involve cosmetic and neuro-logical assessment of soft tissuerepair, whereas regenerative stud-ies in animal models generallyfocus on the restoration of skeletaltissue in addition to soft tissue(Han et al., 2008). Using boneregrowth as definitive evidence for

a regenerative response, there is asubset of clinical reports thatdocument fingertip regenerationafter conservative management ofamputation wounds in both chil-dren (Vidal and Dickson, 1993)and adults (Lee et al., 1995).Thus, it is clear that human finger-tips display a true regenerativeresponse that establishes thefoundation upon which we can

begin to explore ways to enhancethe regenerative response. As a

first step, we provide evidencefrom a case report that begins todefine the proximal extent of re-generative capabilities (Han et al.,2008). This case report involvedan amputation injury in the proxi-

mal region of the terminal phalan-geal bone that was conservativelytreated, and because there was X-ray documentation at the time ofinjury and after the healingresponse was completed, there isclear indication that a regenerativeresponse that included boneregrowth did not occur (Fig. 1C, D,E). Thus, despite the fact that cos-metic healing and good sensibilityof the fingertip was restored, thiscase report begins to identify theproximal boundary of regenerative

ability in humans. Understandingthe physical boundaries of regen-erative potential in humans is animportant first step toward de-veloping a protocol that has pre-dictive value for the treatment ofamputation injuries. What isneeded is a concerted effort tobetter document the limits of thisamazing regenerative response inhumans that would entail radio-graphic analysis of amputationinjuries before and after healing toestablish a database that can be

used both for predicting clinicaloutcome and for experimentalstudies (see below). Since injuriesto the hand alone represent 30%of all reported injuries (Oleske andHahn, 1992; Angermann andLohmann, 1993), and there are19,000 reported digit amputa-tions per year in the United Statesalone (Sorock, 1993), it should bepossible to establish such a data-base in a relatively short timeframe. This would be the first steptoward developing an understand-

ing of the limitations of human re-generative ability and, as well, forthe development of therapies toenhance this amazing response.

DIGIT TIP REGENERATION

IN MICE

A valuable experimental model forhuman fingertip regeneration isthe digit tip in rodents. The mousedigit tip includes the terminal pha-langeal bone which is surrounded

by connective tissue and encasedwithin a nail (Fig. 2A). Like humanfingertips, regenerative ability islevel specific within the terminalphalanx. In adult mice amputa-tions midway through the terminal

phalanx result in a robust regener-ative response, whereas amputa-tion in which more than 3/4th ofthe bone is removed fails to mounta response (Neufeld and Zhao,1995; Han et al., 2008). Becausethe mouse digit is relatively small,the physical distance betweenthese two amputation injuries isless than 1 mm and the tissuecomposition of both regeneratingand nonregenerating wounds issimilar, making this a good modelto investigate the cause of regen-

erative failure. The fact that themouse digit tip is able to regener-ate makes it a very unique part ofthe body, and we have made con-siderable effort to characterize itsdevelopmental anatomy as well as

its regenerative potential.

Developmental Anatomy of the

Digit Tip

The primary structure of themature digit tip is the terminalphalangeal bone (P3), which is lat-

erally flattened and has a triangu-lar shape with its base articulatingwith the subterminal phalangealelement (P2) and its apex forminga sharp point (Fig. 2B, C). The P3bone contains a bone marrowregion localized to the base of theelement and associated with asmall oblong canal that is contigu-ous with the lateral connective tis-sues. There are insertion sites forthe dorsal extensor tendon andventral flexor tendon in the proxi-mal region, and collateral liga-

ments join P2 and P3 laterally. Theterminal phalangeal bone isencased within a nail organ thatcomes to a sharp distal pointextending well beyond the tip ofthe terminal phalanx. The nail cov-ers the dorsal and lateral surfacesof P3 but is not contiguous ven-trally. The nail organ consists ofthe nail matrix proximally thatsupplies cells to the nail bed whichis overlain by the nail plate. Mousenail growth is continuous through-

268 MUNEOKA ET AL.



5/16

out life with nail loss occurring api-cally. The ventral surface of thedigit tip is covered with a thinlayer of keratinized epidermis thatextends to the distal tip where it isthickened and contiguous with thedorsal nail epidermis. Proximally,

the ventral epidermis is contigu-ous with the thickened interdigi-tating epidermis of the ventral fatpad. Between the nail bed and theterminal phalanx is a layer of looseconnective tissue that surroundsthe bone. This layer of connective

tissue is more prominent along thedorsal and lateral surfaces bycomparison to the ventral, andthere are regions where the con-nective tissue cells are organizedperpendicular to the bone surfaceand appear to attach the nail bed

Figure 2. Developmental anatomy of the mouse digit tip. (AC) Mature mouse digit tips. (A) A digit tip showing that the terminalphalange is encased within a nail dorsally and laterally. The terminal phalangeal bone is visible through the nail (arrow). (B) Wholemount terminal phalangeal (P3) bone stained with Alizarin Red S showing a sharp point at the apex. Proximal end of the P3 bonearticulates with the subterminal phalangeal (P2) bone. (C) Sagittal sectioned sample stained with Mallorys triple stain. The looseconnective tissue between the nail bed and the bone contains cells that appear fibroblastic. ( DH) Gene expression in the developingdigit tips of E14.5 embryos. (D) Msx1 is expressed in the apical mesenchymal cells surrounding the forming terminal phalanx. (E)Msx2 is expressed in the apical epidermis and in mesenchymal cells subjacent to the epidermis. (F) Bmp4 is expressed in apical mes-enchymal cells in a domain similar to that of Msx1. (G) Ihh is expressed in digit tip cells, initiating endochondral ossification of theterminal phalanx. (H) The nail organ marker, Hoxc13, is expressed in the distal epidermis associated with presumptive nail tissue.(IM) Gene expression in the developing digit tips at birth. (IK) Expression of cartilage-specific genes. (I) Type II collagen (Col II),a marker for proliferating chondrocytes; (J) Indian hedgehog ( Ihh), a marker for prehypertrophic chondrocytes; (K) Type X Collagen(Col X), a marker for hypertrophic chondrocytes. (LM) Expression of osteoblast-specific genes. (L) Osteocalcin transcripts are firstexpressed at the apex of the terminal phalangeal bone then extend along the dorsal surface; (M) Type I collagen ( Col I) is firstexpressed in a similar pattern. (NP) Whole-mount skeletal staining of postnatal digit tips stained with Alizarin Red S and AlcianBlue. Chondrogenic tissue stains blue and osteogenic tissue stains red. Ossification begins from the distal tip at birth (N) and pro-

gresses in a proximal direction. (O) At postnatal day 7 (PN7), the distal 3=4 of the terminal phalangeal bone has initiated ossificationand by PN14 (P) ossification has commenced along the entire proximal-distal length of the bone. [B,C, IP (From Han M, Yang X, LeeJ, et al. Dev Biol 2008, 315:125135, Copyright 2008, Elsevier, reproduced by permission.); DH (From Han M, Yang X, FarringtonJE, et al. Development 2003, 130:51235132, Copyright 2003, Company of Biologist, reproduced by permission.)]




6/16

to the bone surface. The looseconnective tissue consists primar-ily of fibroblasts and cells associ-ated with the vascular system.

Developmentally, the digit tip isfirst identified as an autonomous

structure at embryonic day 14.5(E14.5) when the P2-P3 interpha-langeal joint becomes visible. Atthis stage, there are several genesthat are specifically expressed bycells of the digit tip that distin-guishes this region from the proxi-mal phalangeal elements (Fig. 2DH). For example, the expression ofHoxc13 in cells of the dorsal epi-dermis marks the forming nailorgan (Godwin and Capecchi,1998), Msx2 is expressed in theapical epidermis and also in cells of

the most distal mesenchyme(Reginelli et al, 1995), and Msx1and Bmp4 are expressed in a largerdomain of the distal mesenchyme(Reginelli et al., 1995; Han et al.,2003). Other genes known to be

expressed at the developing mousedigit tip include Dlx5 (Acamporaet al., 1999), Bambi (Grotewoldet al., 2000), and Dachshund(Hammond et al., 1998; Daviset al., 1999). The distal specificexpression of these genes reinfor-ces the conclusion that the digit tip

is a unique structure both in devel-opment and in regeneration. Theuniqueness of the digit tip in birdsand mammals has recently beenreviewed by Casanova and Sanz-Ezquerro (2007) in the context ofdigit evolution.

Ossification of the mouse digittip initiates just before birth (Hanet al., 2008). Gene expressionstudies indicate that hypertrophicchondrocytes begin to maturebetween E17.5 and E18.5 (birth),with the onset of ColX expression

at the distal tip. At this stage pro-liferating chondroblasts identifiedby ColII expression are localizedto the proximal half of P3 andthere is a band of prehypertrophiccells identified by Ihh expressionbetween the ColX and ColIIexpression domains (Fig. 2IK).During this same period ossifica-tion is initiated at the digit tip inassociation with the ColX domain

as evidenced by histological stain-ing and expression of marker

genes for ossification, Osteocalcinand ColI(Fig. 2L, M). As the digit tipmatures, ossification progresses

in a distal to proximal direction(Fig. 2NP).

In mice, elongation of the termi-nal phalanx continues until itreaches a mature length at 8

weeks of age (Fig. 3A). Elongationrate is rapid during the first 34

Figure 3. Ossification and growth of the terminal phalangeal (P3) bone. (A) Growthcurve of P3 bone. The terminal phalangeal bone continues to elongate during the post-natal period and reaches its mature length at 8 weeks. The lower curves show that fol-lowing amputation at PN3, the terminal phalangeal bone never catches up with unam-putated controls. (BG) Calcein incorporation into the P3 bone. (B) Calcein labeling forone day identifies two ossification centers at 3 weeks of age, one associated with theproximal growth plate and one at the distal tip. (C) By 5 weeks of age only the distalossification center is observed. (D) Calcein incorporation is used as a vital marker toidentify existing bone by long-term labeling studies. New bone deposition distal to thecalcein label identifies bone deposition that has occurred since the initial labeling(white arrows). Calcein was injected at PN1 and analyzed at 4 weeks (D) and 7 weeks

(E) of age. (F) By measuring the length of the terminal phalangeal bone and the proxi-mal-distal length of new bone deposition, we show that both proximal and distal ossifi-cation centers contribute equally to bone elongation from birth until 4 weeks of age.(G) Similar analysis of the period between 4 and 9 weeks of age indicate that length-ening of the terminal phalanx results solely from the distal ossification center. [AE(From Han M, Yang X, Lee J, et al. Dev Biol 2008, 315:125135, Copyright 2008,Elsevier, reproduced by permission.)]

270 MUNEOKA ET AL.



7/16

weeks following birth, after whichthe rate of elongation levels offuntil the mature length is reached(Han et al., 2008). Calcein is amarker that incorporates intonewly forming bone and pulse

labeling is used to identify regionsof the bone that are undergoingactive ossification (Suzuki andMathews, 1966). Calcein pulselabeling of the terminal phalanxindicates that the early stages ofbone growth (before 4 weeks) areassociated with two ossificationcenters, one proximal and one dis-tal (Fig. 3B), whereas by 5 weeksof age only the distal ossificationcenter remains (Fig. 3C). Theproximal ossification center islinked to the proximal endochon-

dral growth plate which closes at34 weeks of age, and the distalossification center results in appo-sitional bone growth similar tothat occurring during diametricalgrowth of the bone collar. Because

ossification by appositional growthis typically linked to growth of thebone collar in long bones, it seemsreasonable to consider the termi-nal phalangeal bone as equivalentto the proximal half of a subtermi-nal phalanx with its central collarregion constricted to form the api-

cal tip of P3. Calcein can also beintroduced as a vital marker ofexisting bone to quantitate ossifi-cation that takes place after itsintroduction; in this case newlyformed bone is unlabeled (Fig. 3D,E). Measurements of distal ossifi-cation by comparison to the elon-gation of the entire terminal pha-lanx show that during the periodfrom birth to 4 weeks of age abouthalf the length of the terminalphalanx results from distal ossifi-cation (Fig. 3F). After 4 weeks of

age the distal ossification center isresponsible for 100% of terminalphalanx elongation (Fig. 3G).

Embryonic Digit Tip as a Model

for Mammalian Regeneration

The developing limb bud anddigits of higher vertebrates pos-sess enhanced regenerative capa-bilities by comparison to adults,and represent models for investi-gating regenerative responses and

regenerative decline associatedwith maturation (Muneoka andSassoon, 1992; Muller et al.,1999). For example, the limb budsof rats and mice have been shownto partially regenerate followingamputation in vitro (Deuchar,1976; Lee et al., 1991) and in vivo(Wanek et al., 1989), and the em-bryonic mouse digit tip undergoesa rapid and complete regenerativeresponse in utero (Reginelli et al.,

1995) and in vitro (Han et al.,2003). In mice, the proximal limitof regenerative capability is asso-ciated with the proximal extent ofthe Msx1 expression domain dur-ing digit development (Reginelliet al., 1995). During embryonicdigit tip regeneration a number ofgenes specifically expressed at theapex of the developing digit areupregulated, suggesting thatthese genes, Msx1, Msx2, Bmp4,

Figure 4. Regeneration response in the fetal digit tips of mice and humans. (A): Thecentral digits (digits 2, 3 and 4) that were amputated at E14.5 exhibit normal regener-ation response. Note that the regenerated digit tips ( asterisk) are little bit shorter thana non-amputated control digit tip (arrowhead). (BD) In situ hybridization of frontalsections at 2 days after amputation showing expression of Msx1 (B), Msx2 (C), andBmp4 (D). (E) Digits of Msx12/2 mutant mice display a regeneration defect. (FH) Insitu hybridization of frontal sections at 2 days after amputation. (F) A nonfunctionaltranscript of Msx1 is expressed at the amputation wound. (G, H) Msx2 and Bmp4

transcripts are not upregulated in the failed regeneration response. ( I) Treatment ofwildtype and Msx12/2 mutant digits with the BMP antagonist, NOGGIN, inhibits theregeneration response. Despite the absence of a regeneration response by NOGGINtreatment, stump tissues maintain expression of Msx1 (J), Msx2 (K) and Bmp4 (L),suggesting that all three genes are upstream of the NOGGIN inhibitory effect in digitregeneration. (MP) Embryonic human digits initiate a regeneration response in vitro.(M) A fingertip from a human embryo with an estimated gestational age (EGA) of 57days displays MSX1 immunostaining that is localized to the nail forming region. (N)Amputated fingertips cultured for 7 days postamputation (DPA) initiate a regenerationresponse, forming a blastema at the wound site (cytokeratin-19 immunostaining). (O)Amputated fingertips cultured for 4 DPA and immunostained for MSX1 show upregula-tion in dorsal mesenchymal tissue. (P) At 7 DPA, blastemal cells stain positive forMSX1. [AH, JL (From Han M, Yang X, Farrington JE, et al. Development 2003,130:51235132, Copyright 2003, Company of Biologist.); MP (From Allan CH, Fleck-man P, Fernandes RJ, et al. Wound Repair Regen 2006, 14:398404, Copyright 2006,Wound Healing Society.)]




8/16

Hoxc13, and Ihh (Fig. 4AD), areimportant for the regenerationresponse. These genes are notexpressed at a proximal digitamputation wound that is normallynonregenerating (Reginelli et al.,

1995; Han et al., 2003). We haveanalyzed the Msx1 mutant (Sato-kata and Maas, 1994) and theMsx2 mutant (Satokata et al.,2000) to determine whether eitherof these genes are playing a func-tional role in the regenerationresponse (Han et al., 2003). Ourstudies show that the Msx1 mu-tant, but not the Msx2 mutant, dis-plays a regeneration phenotype(Fig. 4E), suggesting that Msx1 isplaying a critical role in the regen-eration response (Han et al.,

2003). The distal tip of the Msx1mutant digit following amputationbehaved like a proximal levelamputation, resulting in a trun-cated digit phenotype and a failureto upregulate the distal digit

marker genes (Fig. 4FH). Sincethe Msx1 mutant digit does nothave a limb phenotype in develop-ment, these results suggest thatMsx1 is functioning in a regenera-tion-specific manner. We alsofound that Bmp4 expression in thedigit tip was controlled by the com-

bined action of Msx1 and Msx2,and we therefore carried out stud-ies to determine whether exoge-nous BMP4 could rescue the Msx1regeneration phenotype. Indeed,we found that BMP4 rescued digittip regeneration in a dose-depend-ent manner and expression of alldistal digit marker genes wasupregulated in this induced regen-erative response. To corroboratethis finding, we used the BMP an-tagonist, Noggin, to determine ifBMP signaling was required for

digit regeneration in wildtypedigits, and found that treatingamputated digits with exogenousNoggin inhibited the regenerationresponse (Fig. 4I). Gene expres-sion studies demonstrated thatNoggin treatment did not affectMsx1, Msx2, or Bmp4 expression(Fig. 4JL), suggesting a linearpathway in which Msx1/Msx2regulated Bmp4 and that BMP4

played a key role in controlling theregeneration response (Han et al.,

2003). In other studies we haveinvestigated the role ofDlx5 in em-bryonic digit regeneration (J. Lee,unpublished data). Dlx5 isexpressed in the apical ectodermand mesenchyme in the E14.5

digit tip in a domain that overlapswith Msx2 (Acampora et al., 1999;Han et al., 2003). In amputationstudies we do not find a regenera-tion phenotype in amputatedhomozygous Dlx5 mutant embryodigit tips. In addition, we havetested mutant embryos lackingboth the Dlx5 and Msx2 genes,and in response to digit tip ampu-tation, these embryos can alsosuccessfully regenerate. Thus, wecan conclude that both Msx2and Dlx5 are not essential for

embryonic digit tip regenera-tion even though both are promi-nently expressed in the formingdigit tip.

In separate studies on embry-onic human digits, Allan et al.(2006) established that cultures ofembryonic human digits under se-rum-free conditions were able togo through early stages of a re-generative response. Digits testedwere from embryos with an esti-mated gestational age (EGA) of53117 days and were maintained

in culture from 4 to 28 days. MSX1expression was analyzed immuno-histochemically and was found incontrol digits to be expressed inthe connective tissue between thenail bed and the terminal phalan-geal bone in digits up to 70 daysEGA (Fig. 4M). An analogousexpression domain of Msx1 isobserved in the late mouseembryo and early neonatal digittips (Reginelli et al., 1995). Ampu-tated 57 day EGA digits initiated aregenerative response with the

formation of a blastema like struc-ture (Fig. 4N) and MSX1 expres-sion was found to be associatedwith the regenerating apical cells(Fig. 4O, P; Allan et al., 2006).These results provide evidencethat human digit tissues sharegene expression and injuryresponses with that of the mousedigit, and point to the MSX1 geneas a candidate regulator in thecontrol of a human regenerationresponse.

The Msx genes are implicated inother models of regeneration and/or cell renewal. In the regenerat-ing urodele limb, Msx genes aredown-regulated in the maturelimb, and following limb ampu-

tation both Msx1 and Msx2 arere-expressed during regenerationthen down-regulated after redif-ferentiation (Crews et al., 1995;Simon et al., 1995). Msx geneexpression during regeneration islargely similar to developmentalexpression: Msx1 is expressed bymesenchymal cells, whereas Msx2is expressed by both mesenchy-mal and apical epidermal cells(Carlson et al., 1998; Koshibaet al., 1998). Msx1 is alsoexpressed in association with limb

regeneration in developing Xeno-pus limbs (Endo et al., 2000) andin association with FGF-inducedregeneration of the amputatedchick wing bud (Taylor et al.,1994; Kostakopoulou et al.,1996). In these cases the re-expression of Msx1 is initiated atthe wound surface and establishesan expression domain in theregenerate that is similar to thedeveloping limb. Msx1 has beenshown to be required for tailregeneration in Xenopus (Beck

et al., 2003) and fin regenerationin Zebrafish (Thummel et al.,2006). Developmental studiesshow that Msx1 acts to inhibit dif-ferentiation of a variety of celltypes (Hu et al., 2001 ) and thereis evidence that over-expressionof Msx1 can induce myotube dedif-ferentiation in vitro (Odelberget al., 2000). Transcription studiesshow that the Msx1 protein func-tions as a transcriptional repressoracting with a TATA binding protein(Catron et al., 1995; Zhang et al.,

1996, 1997) and the linker his-tone, H1b, to inhibit differentia-tion-specific gene expression (Leeet al., 2004). Since we find thatMsx1 expression is required fordigit regeneration, is acting in aregeneration-specific manner, andis upregulated during the regener-ation process, we speculate that itis functioning to repress an activ-ity that is normally inhibitory for aregenerative response. Our BMP4rescue data suggests that the

272 MUNEOKA ET AL.



9/16

Bmp4 is a target gene for such ananti-regeneration activity.

Postnatal Digit Tip Regeneration

The regeneration of neonataland adult digit tips is thought tobe largely equivalent (Neufeld andZhao, 1995), although regenera-tion of neonatal digit tips occursduring a time of rapid skeletalelongation. Amputation midwaythrough the terminal phalanx

results in a regeneration response(Fig. 5AC), whereas amputationthrough the proximal region failsto mount a regenerative response(Fig. 5DF). These two amputationplanes transect similar tissues yetthe repair response differs dra-matically. Digit tip regenerationcan be described in terms of distinct phases that are character-istic of other regeneration modelssuch as the urodele amphibianlimb (Gardiner et al., 2002). Afterdistal amputation the regeneration

response is characterized by aninitial wound closure response thatrequires multiple days, despite therelatively small surface area of theamputation injury (Fig. 6A, B).The slow rate of epidermal closurefollowing amputation is similar tothe human response and unlikethe rapid re-epithelialization thatoccurs during amphibian limbregeneration (Carlson et al.,1998). During the wound healingphase there is considerable ero-

sion of the amputated stump boneresulting from enhanced osteo-clast activity (Revardel and Che-bouki, 1987), which extends theactual limit of regeneration to a

more proximal level so as toinclude the eroded bony tissues. Asimilar region of tissue remodelingduring amphibian limb regenera-tion is associated with the upregu-lation of MMPs that are thought tomediate this effect (Yang et al.,1999). Similarly, microarray stud-

ies of digit tip regeneration in miceidentify several upregulated histo-lytic genes, including MMPs, asso-ciated with this phase of regenera-tion (Chadwick et al., 2007).

The second phase involves blas-tema formation. The question ofwhether or not a blastema formsin mammals is addressed below,and for purposes of discussion theblastema is simply defined as anaggregation of proliferating cellsinvolved in the regeneration proc-ess. Based strictly on histological

observations and cell proliferationstudies the blastema appears toform from two sources: (1) themigration of connective tissue cellsacross the amputation wound(Neufeld et al., 2004), and (2)cells arising from the marrow cav-ity of the skeletal stump. In nonre-generating proximal digit amputa-tion, the skeletal stump forms aperiosteum across the injury site

and undergoes ossification result-ing in a truncated cap of the

amputated bone. The blastema,on the other hand, appears tointegrate cells derived from theconnective tissue surrounding thebone with cells derived from themarrow to form a proliferation

center associated with the regen-eration response (Fig. 6C, D; Hanet al., 2008). The cell contributionto the blastema has not yet beendocumented, and this is an area ofresearch that is critically importantfor our understanding of mamma-

lian regeneration. The third phaseof regeneration involves the differ-entiation of regenerated struc-tures, i.e., bone, connective tis-sue, and nail. The regrowth ofbone is most critical because itstructurally defines the regener-

ate. For example, nail regrowthcontinues in proximally amputateddigits but the regrown nail lacksanatomical structure and simplycovers the truncated stump (Fig.5F). In the neonate, bone re-growth following amputation occursbetween 7 and 14 days postampu-tation and proceeds by direct ossi-fication, i.e., without expression ofany endochondral marker genes inthe distal digit region where ossifi-cation is commencing (Fig. 6E).Osteoblasts present at the inter-

face between the blastema andthe bone stump at 7 days postam-putation suggest that the blas-tema is organizing and perhapscontributing to the regeneratednew bone (Fig. 6F). There is aburst of ossification that occursbetween 7 and 14 days postampu-tation that restores the character-istic pattern of the P3 bone andcompletes the regenerativeresponse (Fig. 6G).

THE MAMMALIAN

BLASTEMA

One of the hallmarks of limbregeneration in urodele amphib-ians is the formation of a blastemaof undifferentiated cells that prolif-erate, go through morphogenesis,and differentiate to replace struc-tures lost by amputation (Bryantet al., 2002; Brockes and Kumar,2005). The blastema is a transientphase in regeneration that hasbeen described in terms of the

Figure 5. Digit tip regeneration in neonatal mice. Amputations were carried out at adistal level through bone (A) and at a proximal level through cartilage (D) at postnatalday 3 (PN3). After 6 weeks, digits were analyzed using whole-mount bone stain withAlizarin Red S (B, E) and histological analysis with Mallorys triple stain (C, F). Distalamputations regenerate anatomically normal digit tips (B, C), however, proximalamputations show no signs of regeneration (E, F).




10/16

characteristics of cells with respectto both their tissue of origin aswell as their ultimate fate inregeneration. Thus, for example,we know that the amphibian blas-temal cells (1) arise from eitherdedifferentiation of, and/or stemcells present in, mature tissues,(2) appear undifferentiated andexpress developmental genes dur-

ing the blastema phase, (3) prolif-erate, and (4) differentiate in ei-ther a homotypic or heterotypic(metaplastic) manner (Brockesand Kumar, 2002; Han et al.,2005; Morrison et al., 2006).There is currently no mammaliancounterpart to the urodele blas-tema, and because there exists agrowing interest in developingstrategies to induce regenerativeresponses in mammals, particu-larly humans, it is both necessary

and important to identify parallelswith, as well as deviations from,the best characterized regenerat-ing systems.

Many models for mammalianappendage regeneration involvethe formation of a proliferating ag-gregate of undifferentiated cellsthat undergoes differentiation toform a regenerated structure. This

cell aggregate is often described as blastema-like because it doesnot have all of the characteristics ofthe classical amphibian blastemathat mediates appendage regener-ation. The term blastema is usedgenerally to describe a cell aggre-gate involved in development(e.g., blastema condensations dur-ing skeletal formation, metaneph-rogenic blastema during kidney de-velopment) or regeneration (e.g.,osteoblastic blastema in fracture

healing); however, in limb regener-ation, the term has been redefinedto include characteristics of urodeleblastema cells, and/or their inter-actions with the overlying epider-mal layer. For example, the regen-eration blastema has been definedas (1) a structure derived fromthe dedifferentiation of cells at theamputation wound, (2) arising

through epithelial-mesenchymalinteractions, and (3) contains in-trinsic morphogenetic information(see Carlson, 2005, 2007). Somehave even proposed that regenera-tion itself is defined as a responsethat must involve the dedifferentia-tion of cells at the amputationwound (Kostakopoulou et al.,1996). With respect to mammalianregeneration, this strict definition

of a regeneration blastema impairsthe utility of the blastema concept

Figure 6. Histological and gene expression analyses of regenerating digit tips. (AC) Histological sections of regenerating digit tipsstained with Mallorys triple stain at the time of amputation (A), 4 days postamputation (DPA) (B) and 6 DPA (C). Note the forma-tion of a blastema by 6 DPA. (D) BrdU incorporation at 7 DPA shows robust proliferation in the connective tissue and the bonestump. (E, F) In situ hybridization analyses documenting the expression patterns of a hypertrophic chondrocyte marker (E), Type XCollagen (Col X), and an osteoblast maker (F), Osteocalcin. Note the absence of chondrogenic marker gene transcripts associatedwith the blastema indicating that the regeneration response involves direct ossification. (G) Calcein was used as a vital label toidentify ossification in the stump 1 week after amputation and Alizarin Red S was used to stain bone after 2 weeks. Differentiationof the regenerate is largely completed by 2 weeks after amputation. ( H) Msx2 expression is induced in the dorsal connective tissueduring wound healing stages but is absent in the blastema (not shown). ( I) Bmp4 transcripts are present at the distal tip of theblastema and also in the dorsal connective tissue. (J) Dlx5 is expressed in cells at the base of the blastema and also in the marrowregion of the stump. (K, L) Pedf is specifically expressed in the bone marrow and distal apex of the blastema underneath the woundepidermis during regeneration.

274 MUNEOKA ET AL.



11/16

simply because it presumes thatmammalian regeneration mustproceed via mechanisms similar toamphibians. Whether or not this istrue remains to be seen. Indeed,we argue below that there are likely

multiple ways to regenerate a sin-gle structure, and if a blastema isinvolved, it follows that the charac-teristics of the blastema must notbe constant. Thus, there is value inbroadly defining the blastema asan anatomical structure so as tobring regeneration studies under acommon umbrella where we candiscover similarities and differen-ces when comparing regenerationevents of distinct animal groups(e.g., amphibians versus mam-mals). We therefore favor the sim-

ple definition of a regenerationblastema as an aggregate of prolif-erating undifferentiated cellsinvolved in the regeneration of alost body part.

All three mammalian regenera-

tion models described earlier (earpunch, antler, and digit tip) regen-erate via a blastema. In ear punchregeneration, the blastema formsaround the circumference of thehole punch and it grows inwardprogressively filling in the holewith tissue. Several genes and

proteins are expressed in the blas-tema, including BMP2 (Urist et al.,1997), the EGF family memberPref-1 (Samulewicz et al., 2002),and the matrix metalloproteinases(MMP), MMP2 and MMP9 (Goure-vitch et al., 2003). In addition, anangiopoietin-related growth factor(AGF) expressed in injured skinpromotes regeneration in thetransgenic mouse model (Oikeet al., 2003). The blastema of theregenerating deer antler is theproliferating mesenchymal region

called the mesenchymal growthzone or the reserve mesenchyme(Price et al., 2005; Kierdorf et al.,2007). These cells express themesenchymal stem cell markerSTRO-1, and have been shown tobe multipotent when challenged inin vitro differentiation assays;thus, it is proposed that the blas-tema is stem cell derived (Rolfet al., 2008). In vivo, these cells

give rise to chondrocytes thatundergo hypertrophy and are

invaded by osteocytes following anosteogenic process that recapitu-lates development (Faucheuxet al., 2004). Cultures of antlerblastemal cells provide for a way tocharacterize the control of blas-

tema growth and differentiation(Price et al., 1994; Sadighi et al.,1994). Factors shown to enhancecell proliferation in vitro includeIGF-I, IGF-II, FGF2, and PTHrR,and in vivo support for this prolifer-ative effect comes from immuno-histochemical studies showing thatthe corresponding receptors (IGFR,FGFR, and PPR) are expressed inthe blastema (Price et al., 1994;Sadighi et al., 1994; Barling et al.,2004; Faucheux et al., 2004;Lai et al., 2007). Other studies pro-

vide evidence that key develop-mental signaling pathways, suchas canonical Wnt signaling (Mountet al., 2006), BMP signaling(Barling et al., 2005) and retinoicacid signaling (Allen et al., 2002),are also playing a critical role inantler regeneration.

The regenerating mouse digit tipforms a blastema of proliferatingcells that later undergoes directossification to restore the distalregion of the terminal phalanx(Revardel and Chebouki, 1987;

Neufeld, 1992). While the cellularorigins of the blastema remain tobe explored, studies of the neona-tal blastema identify the looseconnective tissue surrounding theterminal phalangeal bone and themarrow-forming region as areaswhere enhanced cell proliferationis occurring in association with theregenerative response (Han et al.,2008; see Fig. 6D). The digit blas-tema is characterized by its conti-nuity both with the connective tis-sue surrounding the stump bone

and with the stump bone itself.Unlike digit amputation at a proxi-mal (nonregenerating) level wherea periosteum forms a distal capassociated with bone truncation(Neufeld, 1985), the smooth inte-gration of the regenerated bonetissue with the stump is a likelyoutcome of the stump-blastemacontinuity, and the extensiveremodeling of the stump thatoccurs during the wound healingphase appears to play a role. The

integration of the stump tissueswith the regenerate is a researchtopic that has not received muchattention, yet this interface isclearly important for success infunctional tissue engineering and

regenerative medicine. In regen-erating systems, this interfacerepresents a site where terminallydifferentiated cells of the stumpmust functionally interact with theundifferentiated cells of the regen-erate. Further studies on this topicwill prove to be important for thedevelopment of therapeutic strat-egies critical for functional tissueengineering.

Recently, we characterized theexpression of digit developmentmarkers during neonatal digit tip

regeneration using in situ hybrid-ization (Han et al., 2008). Msx1and Msx2 are expressed in the dis-tal digit tip during developmentand are prominently expressedduring embryonic digit tip regen-eration (Fig. 4). In the neonataldigit Msx1 is expressed in the dor-sal connective tissue, whereasMsx2 is expressed in the nail epi-dermis. Following digit tip amputa-tion we see Msx1 expression upre-gulated in association with thehealing dorsal connective tissue,

and in this same region we findMsx2 expression induced (Fig.6H). This upregulation of Msxgenes is transiently associatedwith the wound healing responsesince we do not observe expres-sion of either gene in the blas-tema. We note that the earlyinduction of Msx2 has also beenobserved following amputationduring amphibian limb regenera-tion (Carlson et al., 1998), thusperhaps Msx2 may be serving aparallel function. Once the blas-

tema has formed, we find Bmp4transcripts expressed in the dorsalconnective tissue, in the formingbone, and in the distal blastemalmesenchyme (Fig. 6I). In theunamputated digit, Bmp4 expres-sion is restricted to the formingbone, so it appears to be upregu-lated specifically in the blastemaand dorsal connective tissue. Wealso find the homeodomain con-taining gene Dlx5 expressed in theblastema. Dlx5 is expressed in




12/16

both the epidermis and mesen-chyme of the developing digit tipin a pattern similar to Msx2, andgenetic studies indicate that nei-ther Dlx5 nor Msx2 are requiredfor regeneration (see above). Dlx5

is also known to play a critical rolein osteogenesis (Holleville et al.,2007), thus its expression duringpostnatal digit regeneration isprobably associated with the onsetof ossification by regeneratingcells. In the blastema, Dlx5 ex-pression is restricted to the proxi-mal regions and lies between theBmp4 domain and the skeletalstump (Fig. 6J). We also findgenes expressed in the blastemathat are not expressed during digitdevelopment. For example, pig-

ment epithelium-derived factor(Pedf), a secreted protein withneurotrophic activity (Ramirez-Castillejo et al., 2006), is ex-pressed in the bone marrow, andis prominently expressed in the

blastema (Fig. 6K, L). To date,Pedf is the earliest gene that wehave specifically localized to theforming blastema. Cells express-ing Pedf localize to the apex of theblastema just underlying thewound epidermis. It is not ex-pressed in proximal amputation

wounds that fail to regenerate(not shown), suggesting that itmay be playing a key role in theregeneration response. BecausePedf is prominently expressed inboth bone marrow and blastema,one possibility is that Pedfexpressing cells in the bone mar-row contribute to the establish-ment of the blastema. In addition,the neurotrophic activity of PEDFmay be linked to maintainingdamaged neurons following ampu-tation and during the regeneration

process. In conclusion, the digittip blastema cells display charac-teristics similar to the developingdigit tip, but also characteristicsthat are unique to regeneration.

EVOLUTION OF

REGENERATION IN

MAMMALS

The evolution of regeneration inthe animal kingdom has been

addressed largely within the con-text of animals that possess a highdegree of regenerative capabil-ities, i.e., invertebrates and lowervertebrates (fish and amphibians).For these animals there is indirect

evidence that the ability to regen-erate is an ancient attribute ofmetazoans that is linked to devel-opmental mechanisms, and assuch, is non selectable (Brockeset al., 2001). It is further pro-posed that regeneration is an evo-lutionary remnant of the processof asexual reproduction (SanchezAlvarado, 2000). It seems likelyhowever, that the wound healingcomponent of the regenerationprocess that does not involve areiteration of development is adapt-

ive and has undergone consider-able evolution among metazoans.Since wound healing is the initialphase of any regenerationresponse, it would seem that animportant consideration for our

understanding of regeneration liesat the interface between an evolv-ing wound healing process andconserved developmental mecha-nisms. The process of intercalarygrowth (French et al., 1976; Bry-ant et al., 1981), for example,may represent a mechanism that

has successfully evolved inamphibians and invertebrates tospan this interface (Gardineret al., 1995).

The flip side of the view that re-generative ability is a conservedmechanism in animals that canregenerate is that negative selec-tion for regeneration can explainwhy many groups of animals,including mammals, lack the exten-sive regenerative capacity enjoyedby others. It is here where most dis-cussions on the evolution of regen-

eration ends; the animal kingdomis divided into regenerators andnonregenerators, and the discus-sion moves to addressing thequestion of what might be the cir-cumstances surrounding the lossof regenerative ability in mam-mals. Indeed, a prevalent view isthat the loss of regenerative abilityis linked to a more urgent selec-tive need of developing an adapt-

ive immune system to combatpathogens invading injured tissue

that evolved at the expense of re-generative ability (Mescher andNeff, 2005). However, this viewleaves no place for endogenousregeneration models in mammals,because how can we systemically

forfeit regenerative ability yetretain it in select parts of thebody? Regeneration in mammalsis real and how it fits into ourthinking about the evolution ofregeneration turns out to be criti-cal for understanding how to move

forward in regenerative medicine.For example, it is likely that re-generative ability evolved second-arily from a nonregenerative statein some mammals (see below), sowe can look upon these regenera-tion models as examples of how

nature got into the regenerativemedicine business and came upwith a successful product. Simi-larly, if regenerative ability wasselectively retained in some partsof the body, how is it that theseparts of the body can become re-fractory to a strong negativeselective force? The answers tothese questions may be quite pro-found and could have a majorimpact on the development ofstrategies for how we can success-fully implement regenerative

therapies.Deer antler regeneration is per-haps the best example of anevolved regenerative response asfossil records clearly indicate thatantler formation and regenerationevolved from a nonregenerativeprecondition (Goss, 1969). Theantler develops postnatally as anoutgrowth of the frontal bone andits formation and regeneration fol-low similar processes. The frontal

bone forms during embryogenesisby intramembranous ossification

and the initiation of antler out-growth involves direct ossification;however, antler elongation itselfinvolves an unusual form of endo-chondral ossification in which thegrowth plate is localized at theapex of the outgrowth. Theseobservations clearly demonstratethat the evolution of antler forma-tion/regeneration is not linked tothe developmental mechanismsinvolved in frontal bone formation,but to developmental mechanisms

276 MUNEOKA ET AL.



13/16

used to form elongated bonystructures. Thus, the evolution ofantlers involves the activationand modification of developmentalmechanisms utilized in other partsof the body (e.g., long bones) and

not those involved in frontal boneformation itself. One conclusionthat can be drawn from this is thatwhen regenerative ability evolved,it was not constrained by the de-velopmental history of the tissueresponding to the injury.

It is unclear whether the digit tipin mice represents a model of anevolved regenerative response,one in which regenerative abilityhas been maintained, or both. Thevertebrate digit has undergonetremendous anatomical modifica-

tion, including digit elongation,shortening, and reduction in num-ber, so it is clear that the digit isunder strong selective pressure.Bmp4 is expressed in developingand regenerating digit tips and

required for embryonic digitregeneration (Han et al., 2003),so its activity at the digit tip couldvery well be a target for adaptiveselection. It is worth noting thatBMP4 is known to play a criticalrole in the evolution of tooth for-mation in birds (Chen et al., 2000)

and, as well, is responsible forvariation in beak morphologies(Abzhanov et al., 2004; Wu et al.,2004). Thus, the expression ofBmp4 in both digit developmentand regeneration, along with itsclose association with other evolv-ing vertebrate structures, is con-sistent with the idea that digit tipregeneration has been maintainedduring vertebrate evolution. Onthe other hand, the regenerationof the postnatal digit tip involvesintramembranous ossification of

the terminal phalangeal bone (Hanet al., 2008), a process that is notrelated to the developmental proc-ess of endochondral ossificationthat originally forms the limb skel-eton. Thus, if digit tip regenerationis an example of a maintainedresponse, then why would it utilizea developmental process that iscompletely novel for digit forma-tion? Indeed, the intramembra-

nous ossification of the regener-ated bone suggests that digit tip

regeneration represents anevolved regenerative responsewith cells at the amputation injuryutilizing a developmental processthat is novel for the developinglimb. One possibility is that digit

tip regeneration has been evolu-tionarily conserved because it isadaptive, but the process itselfhas evolved as well. We note thatboth intramembranous ossificationand endochondral ossification areinvolved in bone formation duringlong bone fracture healing, thusboth processes have evolved dur-ing bone healing and remodeling(Schindeler et al., 2008). It is alsointeresting to note that deer antlerregeneration evolved from bonethat undergoes direct ossification

during embryogenesis to form theantler by endochondral ossifica-tion, so these two regenerationmodels appear to have evolved inopposite directions.

CONCLUSION

By comparison to other well stud-ied regeneration models such asthe amphibian limb, the study ofappendage regeneration in mam-mals remains at an immature

stage. Regeneration in mammalsis very real despite the fact that ithas clear limitations. In multiplemammalian models, the woundhealing response results in the for-mation of a blastema of proliferat-ing cells that mediates the regen-eration response. The source ofthe cells that forms the blastemaremains unknown, although thereis evidence that these cells displaycharacteristics associated withmesenchymal stem cells. At thesame time there is also evidence

that connective tissue fibroblastsalso participate in blastema forma-tion, thus the mammalian blas-tema may be composed of cellsderived from multiple sources.Although the formation of theblastema in different mammalianmodels appears to share similar-ities, the mechanisms guiding re-differentiation during regenerationseem to be quite diverse. Weinterpret this observation to sug-gest that the process of re-differ-

entiation during regeneration hasevolved away from the mecha-nisms that guided their initial de-velopment, i.e., regeneration itselfis not constrained by existing de-velopment models. This suggests

that in mammals, regeneration isnot refractory to adaptive selec-tion, but that it is a process thatcontinues to evolve. Sorting outaspects of regeneration that areevolutionarily conserved versusthose that have evolved will proveto be important as we begin toconsider human application basedon discoveries from animal mod-els.

The anatomy of the mammalianresponse suggests that a spatialsystem of positional information is

required to guide regeneration,although we have no understand-ing of the nature of such a systemin mammals. Because the regen-erative response is limited, we donot have anatomical assays avail-able to explore the role that posi-tion plays in regeneration. How-ever, we are encouraged fromstudies of human fibroblasts whichdisplay expression profiles thatvary with position in the body(Chang et al., 2002; Rinn et al.,2006). Since limb fibroblasts play

a critical role in establishing theurodele blastema and in organiz-ing the system of position infor-mation required to effect a suc-cessful regenerative response(Gardiner et al., 2002), the findingthat analogous cells in adulthumans maintain a similar posi-tional memory is suggestive thatthe most essential component fora complex regenerative responsein humans is largely intact. Finally,the role of fibroblasts in a mam-malian injury response is generally

linked to fibrosis and the produc-tion of scar tissue, yet in regener-ating models these same cells areviewed as important contributorsand play an organizing role in theregenerated pattern. Thus, it isour contention that understandingand re-directing the response ofthese cells, in particular, in a non-regenerating mammalian wound isa critical first step in transforminga wound healing response to a re-generative response.




14/16

REFERENCES

Abzhanov A, Protas M, Grant BR, et al.2004. Bmp4 and morphological vari-ation of beaks in Darwins finches.Science 305:14621465.

Acampora D, Merlo GR, Paleari L, et al.1999. Craniofacial, vestibular and

bone defects in mice lacking the Dis-tal-less-related gene Dlx5. Develop-ment 126:37953809.

Allan CH, Fleckman P, Fernandes RJ,et al. 2006. Tissue response andMsx1 expression after human fetaldigit tip amputation in vitro. WoundRepair Regen 14:398404.

Allen SP, Maden M, Price JS. 2002. Arole for retinoic acid in regulating theregeneration of deer antlers. DevBiol 251:409423.

Angermann P, Lohmann M. 1993. Inju-ries to the hand and wrist. A study of50,272 injuries. J Hand Surg [Br]18:642644.

Badylak SF. 2007. The extracellular

matrix as a biologic scaffold material.Biomaterials 28:35873593.

Barling PM, Liu H, Matich J, et al.2004. Expression of PTHrP and thePTH/PTHrP receptor in growing reddeer antler. Cell Biol Int 28:661673.

Barling PM, Lai AK, Tong AST, et al.2005. The distribution of growth fac-tors and their receptors in growingred deer antlers. In: Suttie JM, edi-tor. Advances in Antler Science andProduct Technology. Taieri Print,Dunedin: New Zealand. pp. 3744.

Beck CW, Christen B, Slack JM. 2003.Molecular pathways needed forregeneration of spinal cord and mus-cle in a vertebrate. Dev Cell 5:429

439.Bhattacharya D, Ehrlich LI, Weissman

IL. 2008. Space-time considerationsfor hematopoietic stem cell trans-plantation. Eur J Immunol 38:20602067.

Brockes JP, Kumar A. 2002. Plasticityand reprogramming of differentiatedcells in amphibian regeneration. NatRev Mol Cell Biol 3:566574.

Brockes JP, Kumar A, Velloso CP.2001. Regeneration as an evolution-ary variable. J Anat 199:311.

Brockes JP, Kumar A. 2005. Append-age regeneration in adult vertebratesand implications for regenerativemedicine. Science 310:19191923.

Bryant SV, French V, Bryant PJ. 1981.Distal Regeneration and Symmetry.Science 212:9931002.

Bryant SV, Endo T, Gardiner DM.2002. Vertebrate limb regenerationand the origin of limb stem cells. IntJ Dev Biol 46:887896.

Carlson BM. 2005. Some principles ofregeneration in mammalian systems.Anat Rec B New Anat 287:413.

Carlson BM. 2007. Principles of regen-erative biology. Burlington, MA: Aca-demic Press.

Carlson MR, Bryant SV, Gardiner DM.1998. Expression of Msx-2 during de-

velopment, regeneration, and woundhealing in axolotl limbs. J Exp Zool282:715723.

Casanova JC, Sanz-Ezquerro JJ. 2007.Digit morphogenesis: is the tip differ-ent? Dev Growth Differ 49:479491.

Catron KM, Zhang H, Marshall SC,et al. 1995. Transcriptional repres-

sion by Msx-1 does not requirehomeodomain DNA-binding sites. MolCell Biol 15:861871.

Chadwick RB, Bu L, Yu H, et al. 2007.Digit tip regrowth and differentialgene expression in MRL/Mpj, DBA/2,and C57BL/6 mice. Wound RepairRegen 15:275284.

Chang HY, Chi JT, Dudoit S, et al.2002. Diversity, topographic differ-entiation, and positional memory inhuman fibroblasts. Proc Natl Acad SciUSA 99:1287712882.

Chen Y, Zhang Y, Jiang TX, et al. 2000.Conservation of early odontogenicsignaling pathways in Aves. Proc NatlAcad Sci USA 97:1004410049.

Clark LD, Clark RK, Heber-Katz E.1998. A new murine model for mam-malian wound repair and regenera-tion. Clin Immunol Immunopathol88:3545.

Crews L, Gates PB, Brown R, et al.1995. Expression and activity of thenewt Msx-1 gene in relation to limbregeneration. Proc Biol Sci 259:161171.

Crisan M, Yap S, Casteilla L, et al.2008. A perivascular origin for mes-enchymal stem cells in multiplehuman organs. Cell Stem Cell3:301313.

Davis RJ, Shen W, Heanue TA, et al.1999. Mouse Dach, a homologue of

Drosophila dachshund, is expressedin the developing retina, brain andlimbs. Dev Genes Evol 209:526536.

Deuchar E. 1976. Regeneration ofamputated limb-buds in early ratembryos. J Embryol Exp Morphol35:345354.

Douglas BS. 1972. Conservative man-agement of guillotine amputation ofthe finger in children. Aust Paediatr J8:8689.

Endo T, Tamura K, Ide H. 2000. Analy-sis of gene expressions during Xeno-pus forelimb regeneration. Dev Biol220:296306.

Faucheux C, Nicholls BM, Allen S, et al.2004 Recapitulation of the parathy-

roid hormone-related peptide-Indianhedgehog pathway in the regenerat-ing deer antler. Dev Dyn 231:8897.

French V. 1976. Leg regeneration inthe cockroach, Blatella germanica.II. Regeneration from a non-congru-ent tibial graft/host junction. JEmbryol Exp Morphol 35:267301.

Gardiner DM. 2005. Ontogeneticdecline of regenerative ability andthe stimulation of human regenera-tion. Rejuvenation Res 8:141153.

Gardiner DM, Blumberg B, Komine Y,et al. 1995. Regulation of HoxAexpression in developing and regen-

erating axolotl limbs. Development121:17311741.

Gardiner DM, Endo T, Bryant SV.2002. The molecular basis of am-phibian limb regeneration: integrat-ing the old with the new. Semin CellDev Biol 13:345352.

Godwin AR, Capecchi MR. 1998.

Hoxc13 mutant mice lack externalhair. Genes Dev 12:1120.

Goss RJ. 1969. Principles of Regenera-tion. New York: Academic Press.Chapter 11, pp.223255.

Goss RJ, Van Praagh A, Brewer P.1992. The mechanism of antler cast-ing in the fallow deer. J Exp Zool264:429436.

Gourevitch D, Clark L, Chen P, et al.2003. Matrix metalloproteinase activ-ity correlates with blastema forma-tion in the regenerating MRL mouseear hole model. Dev Dyn 226:377387.

Grotewold L, Plum M, Dildrop R, et al.2000. Bambi is coexpressed with

Bmp-4 during mouse embryogenesis.Mech Dev 100:327330.

Hammond KL, Hanson IM, Brown AG,et al. 1998. Mammalian and Dro-sophila dachshund genes are relatedto the Ski proto-oncogene and areexpressed in eye and limb. Mech Dev74:121131.

Han M, Yang X, Farrington JE, et al.2003. Digit regeneration is regulatedby Msx1 and BMP4 in fetal mice.Development 130:51235132.

Han M, Yang X, Taylor G, et al. 2005.Limb regeneration in higher verte-brates: developing a roadmap. AnatRec B New Anat 287:1424.

Han M, Yang X, Lee J, et al. 2008. De-

velopment and regeneration of theneonatal digit tip in mice. Dev Biol315:125135.

Heber-Katz E. 1999. The regeneratingmouse ear. Semin Cell Dev Biol 10:415419.

Holleville N, Mateos S, Bontoux M,et al. 2007. Dlx5 drives Runx2expression and osteogenic differen-tiation in developing cranial suturemesenchyme. Dev Biol 304:860874.

Howard D, Buttery LD, Shakesheff KM,et al. 2008. Tissue engineering:strategies, stem cells and scaffolds.J Anat 213:6672.

Hu G, Lee H, Price SM, et al. 2001. Msx

homeobox genes inhibit differentia-tion through upregulation of cyclinD1. Development 128:23732384.

Illingworth CM. 1974. Trapped fingersand amputated finger tips in chil-dren. J Pediatr Surg 9:853858.

Jiang Y, Jahagirdar BN, Reinhardt RL,et al. 2002. Pluripotency of mesen-chymal stem cells derived from adultmarrow. Nature 418:4149.

Kench JA, Russell DM, Fadok VA, et al.1999. Aberrant wound healing andTGF-beta production in the autoimmune-prone MRL/1 mouse. ClinImmunol 92:300310.

278 MUNEOKA ET AL.



15/16

Khan IM, Gilbert SJ, Singhrao SK,et al. 2008. Cartilage integration:evaluation of the reasons for failureof integration during cartilage repair.A review. Eur Cell Mater 16:2639.

Kierdorf U, Kierdorf H, Szuwart T.2007. Deer antler regeneration:cells, concepts, and controversies.

J Morphol 268:726738.Koshiba K, Kuroiwa A, Yamamoto H,

et al. 1998. Expression of Msx genesin regenerating and developinglimbs of axolotl. J Exp Zool 282:703714.

Kostakopoulou K, Vogel A, Brickell P,et al. 1996. Regeneration of wingbud stumps of chick embryos andreactivation of Msx-1 and Shhexpression in response to FGF-4 andridge signals. Mech Dev 55:119131.

Lai AK, Hou WL, Verdon DJ, et al.2007. The distribution of the growthfactors FGF-2 and VEGF, and theirreceptors, in growing red deer antler.Tissue Cell 39:3546.

Lee H, Habas R, Abate-Shen C. 2004.MSX1 cooperates with histone H1bfor inhibitionof transcription and myo-genesis. Science 304:16751678.

Lee KK, Chan WY. 1991. A study onthe regenerative potential of partiallyexcised mouse embryonic fore-limbbud. Anat Embryol (Berl) 184:153157.

Lee LP, Lau PY, Chan CW. 1995. A sim-ple and efficient treatment for finger-tip injuries. J Hand Surg [Br] 20:6371.

Li C, Suttie JM, Clark DE. 2005. Histo-logical examination of antler regen-eration in red deer (Cervus elaphus).Anat Rec A Discov Mol Cell Evol Biol

282:163174.Li X, Gu W, Masinde G, et al. 2001.

Genetic control of the rate of woundhealing in mice. Heredity 86:668674.

Mescher AL, Neff AW. 2005. Regenera-tive capacity and the developingimmune system. Adv Biochem EngBiotechnol 93:3966.

Metcalfe AD, Willis H, Beare A, et al.2006. Characterizing regeneration inthe vertebrate ear. J Anat 209:439446.

Morrison JI, Loof S, He P, et al. 2006.Salamander limb regenerationinvolves the activation of a multipo-tent skeletal muscle satellite cell

population. J Cell Biol 172:433440.Mount JG, Muzylak M, Allen S, et al.

2006. Evidence that the canonicalWnt signalling pathway regulatesdeer antler regeneration. Dev Dyn235:13901399.

Muller TL, Ngo-Muller V, Reginelli A,et al. 1999. Regeneration in highervertebrates: limb buds and digittips. Semin Cell Dev Biol 10:405413.

Muneoka K, Sassoon D. 1992. Molecu-lar aspects of regeneration in devel-oping vertebrate limbs. Dev Biol152:3749.

Neufeld DA. 1985. Bone healing afteramputation of mouse digits and newtlimbs: implications for inducedregeneration in mammals. Anat Rec211:156165.

Neufeld DA. 1992. Digital regeneration inmammals. In: Bone, Vol. 5. CRC Press.Boca Raton, Florida pp. 287312.

Neufeld DA, Hosman S, Yescas T, et al.2004. Actin fiber patterns detectedby Alexafluor 488 phalloidin suggestsimilar cell migration in regeneratingand nonregenerating rodent toes.Anat Rec A Discov Mol Cell Evol Biol278:450453.

Neufeld DA, Zhao W. 1995. Boneregrowth after digit tip amputation inmice is equivalent in adults and neo-nates. Wound Repair Regen 3:461466.

Odelberg SJ, Kollhoff A, Keating MT.2000. Dedifferentiation of mamma-lian myotubes induced by msx1. Cell103:10991109.

Oike Y, Yasunaga K, Ito Y, et al. 2003.

Angiopoietin-related growth factor(AGF) promotes epidermal prolifera-tion, remodeling, and regeneration.Proc Natl Acad Sci USA 100:94949499.

Oleske DM, Hahn JJ. 1992. Work-related injuries of the hand: datafrom an occupational injury/illnesssurveillance system. J CommunityHealth 17:205219.

Pendegrass CJ, Goodship AE, Price JS,et al. 2006. Natures answer tobreaching the skin barrier: an inno-vative development for amputees.J Anat 209:5967.

Price JS, Oyajobi BO, Oreffo RO, et al.1994. Cells cultured from the grow-

ing tip of red deer antler expressalkaline phosphatase and proliferatein response to insulin-like growthfactor-I. J Endocrinol 143:R9R16.

Price JS, Allen S, Faucheux C, et al.2005. Deer antlers: a zoological curi-osity or the key to understandingorgan regeneration in mammals?J Anat 207:603618.

Rajnoch C, Ferguson S, Metcalfe AD,et al. 2003. Regeneration of the earafter wounding in different mousestrains is dependent on the severityof wound trauma. Dev Dyn 226:388397.

Ramirez-Castillejo C, Sanchez-SanchezF, Andreu-Agullo C, et al. 2006. Pig-

ment epithelium-derived factor is aniche signal for neural stem cellrenewal. Nat Neurosci 9:331339.

Reginelli AD, Wang YQ, Sassoon D,et al. 1995. Digit tip regenerationcorrelates with regions of Msx1(Hox 7) expression in fetal and new-born mice. Development 121:10651076.

Revardel JL, Chebouki F. 1987. Etudede la reponse a lamputation desphalanges chez la souris: role mor-phogenetique des epitheliums, stim-ulation de la chondrogenese. Can JZool 65:31663176.

Rinn JL, Bondre C, Gladstone HB, et al.2006. Anatomic demarcation by posi-tional variation in fibroblast geneexpression programs. PLoS Genet2:e119.

Rolf HJ, Kierdorf U, Kierdorf H, et al.2008. Localization and characteriza-tion of STRO-1 cells in the deer pedi-

cle and regenerating antler. PLoSONE 3:e2064.

Sadighi M, Haines SR, Skottner A,et al. 1994. Effects of insulin-likegrowth factor-I (IGF-I) and IGF-II onthe growth of antler cells in vitro.J Endocrinol 143:461469.

Samulewicz SJ, Seitz A, Clark L, et al.2002. Expression of preadipocytefactor-1(Pref-1), a delta-like protein,in healing mouse ears. Wound RepairRegen 10:215221.

Sanchez Alvarado A. 2000. Regenera-tion in the metazoans: why does ithappen? Bioessays 22:578590.

Sanchez Alvarado A, Tsonis PA. 2006.Bridging the regeneration gap:

genetic insights from diverse animalmodels. Nat Rev Genet 7:873884.

Satokata I, Maas R. 1994. Msx1 defi-cient mice exhibit cleft palate andabnormalities of craniofacial andtooth development. Nat Genet6:348356.

Satokata I, Ma L, Ohshima H, et al.2000. Msx2 deficiency in mice causespleiotropic defects in bone growthand ectodermal organ formation. NatGenet 24:391395.

Schindeler A, Morse A, Harry L, et al.2008. Models of tibial fracture heal-ing in normal and Nf1-deficient mice.J Orthop Res 26:10531060.

Simon HG, Nelson C, Goff D, et al.

1995. Differential expression of myo-genic regulatory genes and Msx-1during dedifferentiation and rediffer-entiation of regenerating amphibianlimbs. Dev Dyn 202:112.

Slack JM, Lin G, Chen Y. 2008. TheXenopus tadpole: a new model forregeneration research. Cell Mol LifeSci 65:5463.

Sorock GS, Smith E, Hall N. 1993. Hos-pitalized occupational finger amputa-tions, New Jersey, 1985 and 1986.Am J Ind Med 23:439447.

Suzuki HK, Mathews A. 1966. Two-color fluorescent labeling of mineral-izing tissues with tetracycline and2,4-bis[N,N0-di-(carbomethyl)amino-

methyl] fluorescein. Stain Technol41:5760.

Tanaka EM. 2003. Cell differentiationand cell fate during urodele tail andlimb regeneration. Curr Opin GenetDev 13:497501.

Taylor GP, Anderson R, Reginelli AD,et al. 1994. FGF-2 induces regenera-tion of the chick limb bud. Dev Biol163:282284.

Thummel R, Bai S, Sarras MP, et al.2006. Inhibition of zebrafish finregeneration using in vivo electropo-ration of morpholinos against fgfr1and msxb. Dev Dyn 235:336346.




16/16

Urist MR, Behnam K, Kerendi F, et al.1997. Lipids closely associated withbone morphogenetic protein (BMP)and induced heterotopic bone forma-tion. With preliminary observationsof deficiencies in lipid and osteoin-duction in lathyrism in rats. ConnectTissue Res 36:920.

Vidal P, Dickson MG. 1993. Regenera-tion of the distal phalanx. A casereport. J HandSurg [Br] 18:230233.

Wanek N, Muneoka K, Bryant SV.1989. Evidence for regulation follow-ing amputation and tissue grafting inthe developing mouse limb. J ExpZool 249:5561.

Williams-Boyce PK, Daniel JC Jr. 1986.Comparison of ear tissue regenera-tion in mammals. J Anat 149:5563.

Wu P, Jiang TX, Suksaweang S, et al.2004. Molecular shaping of the beak.Science 305:14651466.

Yamanaka S. 2008. Pluripotency andnuclear reprogramming. Philos Trans

R SocLondB Biol Sci363:20792087.Yang EV, Gardiner DM, Carlson MR,

et al. 1999. Expression of Mmp-9and related matrix metalloproteinasegenes during axolotl limb regenera-tion. Dev Dyn 216:29.

Yokoyama H, Yonei-Tamura S, Endo T,et al. 2000. Mesenchyme with fgf-10

expression is responsible for regen-erative capacity in Xenopus limbbuds. Dev Biol 219:1829.

Zhang H, Catron KM, Abate-Shen C.1996. A role for the Msx-1 homeodo-main in transcriptional regulation:residues in the N-terminal armmediate TATA binding protein inter-

action and transcriptional repression.Proc Natl Acad Sci USA 93:17641769.

Zhang H, Hu G, Wang H, et al. 1997.Heterodimerization of Msx and Dlxhomeoproteins results in functionalantagonism. Mol Cell Biol 17:29202932.

280 MUNEOKA ET AL.


Documents

paperpododo6