pluripotent neural crest stem cells in the adult hair...

RESEARCH ARTICLE

Pluripotent Neural Crest Stem Cells in the AdultHair FollicleM. Sieber-Blum,1* M. Grim,2 Y.F. Hu,1 and V. Szeder1,2

We report the presence of pluripotent neural crest stem cells in the adult mammalian hair follicle. Numerous neuralcrest cells reside in the outer root sheath from the bulge to the matrix at the base of the follicle. Bulge explants fromadult mouse whisker follicles yield migratory neural crest cells, which in clonal culture form colonies consisting ofover a thousand cells. Clones contain neurons, smooth muscle cells, rare Schwann cells and melanocytes,demonstrating pluripotency of the clone-forming cell. Targeted differentiation into Schwann cells and chondrocyteswas achieved with neuregulin-1 and bone morphogenetic protein-2, respectively. Serial cloning in vitro demonstratedself-renewal capability. Together, the data show that the adult mouse whisker follicle contains pluripotent neuralcrest stem cells, termed epidermal neural crest cells (eNCSC). eNCSC are promising candidates for diverse celltherapy paradigms because of their high degree of inherent plasticity and due to their easy accessibility in the skin.Developmental Dynamics 231:258–269, 2004. © 2004 Wiley-Liss, Inc.

Key words: neural crest; stem cell; adult stem cell; neuron; Schwann cell; smooth muscle cell; melanocyte; chondrocyte;epidermis; hair follicle; bulge

Received 21 January 2004; Revised 13 April 2004; Accepted 19 April 2004

INTRODUCTION

The neural crest is a transient embry-onic tissue that originates in the neu-ral folds, invades the embryo, anddifferentiates in distinct locationsinto a wide array of adult cell typesand tissues. Neural crest derivativesinclude neurons, Schwann cells, andglia of the autonomic and entericnervous systems, most primary sen-sory neurons, endocrine cells (e.g.,the adrenal medulla and C-cells ofthe thyroid), the smooth muscula-ture of the cardiac outflow tract andgreat vessels, pigment cells of theskin and internal organs, as well asbone, cartilage, and connective tis-sue of the face and ventral neck (LeDouarin and Kalcheim, 1999).

Wnt-1 is expressed transiently bypremigratory and early migratoryneural crest cells, and by dorsal neu-ral tube cells (Wilkinson et al., 1987;Davis et al., 1988; Molven et al., 1991;McMahon et al., 1992; Wolda et al.,1993; Echelard et al., 1994). R26R/�mice express �-galactosidase in acre-inducible manner (Soriano,1999), whereas Wnt1-cre mice ex-press cre under the control of theWnt1- promoter (Danielian et al.,1998). In Wnt1-cre/R26R compoundtransgenic mice; therefore, neuralcrest cells and their progeny perma-nently express �-galactosidase(Friederich and Soriano, 1991; Ech-elard et al., 1994; Danielian et al.,1998; Soriano, 1999; Chai et al., 2000;

Jiang et al., 2002; Szeder et al.,2003), providing a genetic ap-proach to neural crest cell tracing.

At the onset of migration, the neu-ral crest is a mixed population ofcells that consists of pluripotent stemcells, fate-restricted cells, and cellsthat are committed to a particularcell lineage (Sieber-Blum and Co-hen, 1980; Sieber-Blum and Sieber,1984; Bronner-Fraser and Fraser,1988; Stemple et al., 1988; Baroffio etal., 1988; Sieber-Blum, 1989; Gershonet al., 1993; Ito et al., 1993; Henionand Weston, 1997). Microenviron-mental cues as well as cell-autono-mous mechanisms affect neuralcrest stem cell survival, proliferation,and lineage commitment. Some

1Department of Cell Biology, Neurobiology and Anatomy, Medical College of Wisconsin, Milwaukee, Wisconsin2Institute of Anatomy, First Medical Faculty, Charles University Prague, Prague, Czech RepublicGrant sponsor: National Institute of Neurological Disorders and Stroke, NIH; USPHS; Grant number: NS38500; Rockefeller Brothers Fund,New York; Charles E. Culpeper Biomedical Pilot Initiative, Grant number: 03-128; Ministry of Education of the Czech Republic; Grant number:VZ 1111 00003-3G; Grant number: LN 00A065.*Correspondence to: Maya Sieber-Blum, Ph.D., Department of Cell Biology, Neurobiology, and Anatomy, Medical College of Wisconsin,Milwaukee, WI 53226. E-mail: [email protected]

DOI 10.1002/dvdy.20129Published online 28 July 2004 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 231:258–269, 2004

© 2004 Wiley-Liss, Inc.

neural crest stem cells elude thesesignals and persist within crest-de-rived tissues. Neural crest stem cellshave been identified in the embry-

onic dorsal root ganglia (Duff et al.,1991), sympathetic ganglia (Duff etal., 1991), somatic ectoderm (Rich-ardson and Sieber-Blum, 1993), car-

diac outflow tract (Ito and Sieber-Blum, 1993), and in the embryonicsciatic nerve (Bixby et al., 2002). Re-cently, Kruger et al. (2002) have iso-

Fig. 1. X-gal–positive cells in the Wnt1-cre/R26R prenatal and postnatal mouse whis-ker follicle. Blue, X-gal reaction product;red, nuclear red counterstain. A: Tangentialsection through whiskers of a gestationalday 16.5 mouse embryo. X-gal–positiveneural crest-derived cells are present in thedermis (d) between whiskers, in the dermalpapilla (dp), and in the blood sinuses (bs)that surround the whiskers. In addition, neu-ral crest cells are localized in the outer rootsheath within the follicle. ed, epidermis.B: Higher magnification of the area markedat the left in A. A stream of neural crest cellsextends from the bulge region to the matrixnear the dermal papilla (arrows), suggest-ing migratory behavior within the follicle.C: Higher magnification of area marked atthe right in A. This section shows the bulgearea. Many neural crest cells are present(e.g., arrow). D: Section through the base ofa whisker follicle of a newborn double-transgenic mouse. Several X-gal–positivemelanocyte are visible (e.g., arrows).E: Tangential section through a postnatalday 24 whisker follicle. The bulge region (b)contains several X-gal–positive neural crestcells (e.g., arrow). The hair is marked byan asterisk. rs; ring sinus. F: Cross-sectionthrough a whisker from an adult double-transgenic mouse at the level of the dermalpapilla (dp). Several neural crest-derivedX-gal–positive cells (e.g., arrows) arepresent in the matrix, which surrounds thedermal papilla. Scale bars � 100 �m in A,50 �m in C (applies to B,C), 50 �m in D,100 �m in E, 50 �m in F.

Fig. 2. Dissection of the bulge from adult whisker follicles. A: Dissected follicle surrounded by dermis and fat tissue. On the left, the darkmatrix and, on the right, the hair are visible. B: Dissected whisker follicle devoid of dermis. On the right, below the skin, the ring sinus, andon the left, the cavernous sinus are visible. C: Bulge within the connective tissue capsule. D: Isolated bulge. Scale bars � 100 �m in A–D.

EPIDERMAL NEURAL CREST STEM CELLS 259

lated neural crest-derived stem cellsalso from adult intestine. Here, wedescribe an additional adult tissuethat harbors pluripotent neural creststem cells, the epidermis of the hairfollicle.

RESULTS

Neural Crest-Derived Cells inthe Embryonic and PostnatalFacial Skin

Figure 1 shows sections through em-bryonic and postnatal Wnt1-cre/R26R facial skin of an embryonic day(E) 16.5 double transgenic embryo.Neural crest-derived cells are identi-fied by their blue 5-bromo-4-chloro-3-indolyl-�-D-galactopyranoside (X-gal) reaction product. In the face,many structures are of neural crestorigin. These cells include the facialdermis (Fig. 1A, d) and the dermalpapilla of hair follicles (Fig. 1A, dp)but not the epidermis (Fig. 1A, ed).The whisker follicle is surrounded byblood sinuses (Fig. 1A, bs), the con-nective tissue of which is also of neu-ral crest origin. The inner layers of theouter root sheath, which constitutesthe epidermal invagination of thehair follicle, contain neural crest cellsas well. Some neural crest cells withinhair follicles form streams that ex-tend from the bulge region to thebase of the follicle, which containsthe matrix that forms the new hair(Fig. 1B, arrows; higher magnificationof area marked at the left in A), sug-gesting that neural crest cells mi-grate within the hair follicle. A largenumber of neural crest cells reside inthe bulge region both of embryonic(Fig. 1C, arrow; E16.5) and postnatalwhiskers (Fig. 1D, arrows; postnatalday 24). Because melanocyteswithin hair follicles provide melaninto the growing hair, the matrix is ex-pected to contain neural crest cells.X-gal–positive neural crest cells areindeed present in the matrix of whis-kers both from newborn (Fig. 1E, ar-rows) and adult (Fig. 1F, arrows;cross-section) mice. Many of themcontain pigment granules (Fig. 1E,e.g., arrow). In summary, there areneural crest cells localized in the epi-dermis of the whisker follicle alongthe entire length of the follicle, bothprenatally and postnatally.

Neural Crest Cells EmigrateFrom Bulge Explants

The objective of this study was todetermine, by in vitro clonal analysis,the developmental potentials of theneural crest cells that reside in thebulge area, which is a known nichefor keratinocyte stem cells (Oshimaet al., 2001) and melanocyte stemcells (Nishimura et al., 2002). To thisend, we microdissected the bulgearea. Whiskers were dissected fromthe skin (Fig. 2A), and the dermis andfat tissue were then removed me-chanically and with buffer rinses (Fig.2B), exposing the ring sinus belowthe skin and the cavernous sinusnear the base of the follicle. Thecapsule of the whisker was cut lon-gitudinally, and subsequently the fol-licle was transected both above thecavernous sinus and below the skin,which yielded the bulge area withinthe collagen capsule (Fig. 2C). Fi-nally, the bulge region was rolled outof the capsule (Fig. 2C) and placedinto a collagen-coated cultureplate, where it adhered to the sub-stratum within 1 hr.

Figure 3 shows a bulge explant 4days postexplantation. In Figure 3A,the explant and the emigrated cellsare shown with phase contrast op-tics. Figure 3B shows a higher magni-fication with focus on the explant.Numerous X-gal–positive neuralcrest cells are present within thebulge. Figure 3C shows at highermagnification the area to the rightmarked in A. Figure 3D is the corre-sponding brightfield image, showingthat all cells are X-gal–positive. Fig-ure 3E shows with phase contrast op-tics the area to the left marked in(A). Figure 3F is the correspondingbrightfield image, which shows thatall cells are X-gal–positive. At 48 hrpostexplantation, 0–15 cells and, at72 hr, 126.8 � 41.5 cells per explantare present on average. The in-crease in cell numbers over time isdue to emigration of new cells andconcomitant rapid proliferation (ini-tial doubling time approximately 6hr). Bulge explants from wild typeand single-transgenic littermates areX-gal–negative. Emigrated cells alsoexpress Sox10, a marker for neuralcrest cells (Fig. 4; Kuhlbrodt et al.,1998; Rehberg et al., 2002), which

confirms the neural crest origin of thecells that emerge from bulge ex-plants.

Bulge Explant-Derived NeuralCrest Cells Are Pluripotent

In vitro clonal analysis establishedthat neural crest cells from bulge ex-plants are pluripotent. Figure 5A–Dshows the time course of a clone asit grows during the first 3 days: (Fig.5A) 6 hr, (Fig. 5B) 18 hr, (Fig. 5C) 48 hr,and (Fig. 5D) 72 hr after cloning. Sixhours after cloning, cells have a stel-late morphology (Fig. 5A) and an ini-tial doubling time of approximately 6hr, which in subsequent days in-creases to 12–24 hr. At 24 hr of clonalculture, clones contain 4–16 cells.The continuously changing shape ofthe clone illustrates the high cell mo-tility (Fig. 5). By 2 weeks of culture,this type of colony consists of thou-sands of cells and constitutes 83.0 �2.7% of all colonies (Table 1). The re-maining 17% of colonies are small,consisting of four to eight cells andcontain one of two types cells: eitherflattened cells that resemble smoothmuscle cells or unidentified smallelongated cells.

At 2 weeks of culture, many cells inclones exhibit long processes (Fig.6A–C). Others resemble morpholog-ically mature smooth muscle cells(Fig. 6B, arrowhead). Many prolifer-ating cells are still present at 2 weeksin clonal culture (see doublet of lateanaphase cells in Fig. 6B, arrow) andall cells are X-gal–positive (Fig. 6C).To identify different cell types, cloneswere first processed for X-gal reac-tion and subsequently for immuno-cytochemistry with cell type-specificantibodies. Cells with long processesare intensely immunoreactive forneuron-specific �-III tubulin (Fig. 7A).Figure 7B shows X-gal reactivity inthe areas marked in Figure 7A. TheX-gal reaction product is most in-tense in and often limited to inclu-sion bodies, as has been describedpreviously by Rico et al. (2002) andSzeder et al. (2003). Large flattenedcells are immunoreactive for smoothmuscle actin (Fig. 7A, fluorescein;Fig. 7C, Texas Red). Rarely, clonesalso contain Schwann cell progeni-tors as shown by SCIP immunoreac-tivity (Fig. 7D,E; Zorick et al., 1996)

260 SIEBER-BLUM ET AL.

Fig. 3. Bulge explant. X-gal–reacted explant 4 days postexplantation. A: Phase contrastimage of a bulge explant. The explant with the hair still inside and emigrated cells arevisible. The explant got displaced away from the cells during mounting of the cover slip. B:The same bulge explant with focus on the bulge. Numerous X-gal–positive neural crestcells are present within the bulge. C: Area marked at the right in A, shown with phasecontrast optics. D: Corresponding brightfield image. All emigrated cells are X-gal–positive.E: Area marked on the left in A shown with phase contrast optics. F: Corresponding imagewith brightfield optics. All cells are X-gal–positive. Scale bars � 100 �m in A, 100 �m in B, 100�m in F (applies to C–F).

Fig. 4. Sox10 expression in bulge explants.Sox10 is a marker for early migrating neuralcrest cells. All cells that have emigratedfrom bulge explants express Sox10 at highlevels, confirming that they are neural crestcells (arrows). Scale bar � 50 �m in B (ap-plies to A,B).

Fig. 5. Bulge-derived cells in clonal culture. A: One cell 6 hr after cloning. B: Four daughter cells from the same clone-forming cell as inA at 18 hr. C: The same clone at 48 hr consists of approximately 13 cells. D: The same clone at 72 hr. The changing shape of the clonereflects the high motility of the cells. Scale bars � 50 �m in B (applies to A,B), 50 �m in D (applies to C,D).

Fig. 6. Clones at 2 weeks in clonal cultures. At 2 weeks in clonal culture, colonies consist of thousands of cells. A: A small part of a2-week-old clone. Most cells are elongated. B: Higher magnification of a different area. Some cells are large and flattened (arrowhead).Proliferating cells are still present (see postmitotic doublet marked by the arrow). C: All cells in clones are X-gal–positive (Hoffmanncontrast optics), confirming that they are neural crest-derived cells. Scale bars � 50 �m (applies to A–C).

and S100 immunoreactivity (Fig.7F,G; Parkinson et al., 2001). Thepresence of pigment cells in clonesis documented by MelEM immuno-reactivity (Fig. 7H,I; Nataf et al., 1993;Alexanian and Sieber-Blum, 2003).The differentiation of multiple celltypes within clones demonstratesthat bulge-derived neural crest cellsare pluripotent.

Targeted Differentiation IntoSchwann Cell Progenitors

Shah et al. (1994) have reported theneuregulin promotes differentiation ofneural crest cells into Schwann cells. Inthe presence of neuregulin-1 (10 nM),clones contain large numbers of

Schwann cell progenitors, and neu-rons are present as well. Figure 8shows a quadruple stain that com-bines glial fibrillary acidic protein(GFAP; Schwann cell marker) immu-noreactivity (Texas Red) with �-III im-munoreactivity (fluorescein), DAPI(4�,6-diamidine-2-phenylidole-dihy-drochloride) nuclear stain, and X-galreaction. Figure 8A–D focuses on agroup of GFAP-positive Schwann cellprogenitors: GFAP, Texas Red, immu-noreactivity (Fig. 8A), DAPI (Fig. 8B),�-III tubulin (fluorescein; absent, Fig.8C). Figure 8D shows merged imagesin pseudocolor. Figure 8E–I shows agroup of neurons in a different area ofthe same clone: �-III tubulin immuno-reactivity (Fig. 8E), merged brightfield

and fluorescein images to better visu-alize the X-gal reaction product withinneurons (e.g., Fig. 8F, arrows), corre-sponding DAPI stain (Fig. 8G), andGFAP immunofluorescence (Fig. 8H,absent). Figure 8I shows merged im-ages in pseudocolor. Taken together,the data show that, while Schwanncell progenitors rarely develop in ourregular culture medium, their numberincreases greatly in the presence ofneuregulin-1, which suggests thatneuregulin-1 directs stem cells to dif-ferentiate along the Schwann cell lin-eage. Furthermore, the data in Figure8 show that neurons do not express aglia marker and Schwann cell pro-genitors do not express a neuronalmarker.

Targeted Differentiation IntoChondrocytes

To determine whether bulge-de-rived neural crest cells can generatethe full spectrum of cranial neuralcrest derivatives, we sought to differ-entiate them into chondrocytes.Sox9 is required for the commitmentof neural crest cells to the chondro-genic lineage (Mori-Akiyama et al.,2003) as it is a potent activator oftype II collagen expression in chon-drocytes (Kypriotou et al., 2003).BMP-2 causes robust up-regulation

TABLE 1. Percentage of Primary and Secondary Clones Formeda

Clone typeClones formed by stem cells(% of total � SEM)

Primary (from day 4 bulge explants) 83.0 � 2.7*Secondary (from day 3 primary clones 73.5 � 6.7**Secondary (from day 5 primary clones) 66.2 � 4.4***

aAverage number of cells per primary clone at the time of subcloning, 28.5 � 3(day 3) and 52.4 � 5.5 (day 5).

*P � 0.13.**P � 0.015.***P � 0.31.

Fig. 7. Differentiated cell types expressed in clones. A,B: Quadruple stain combining anti–neuron-specific �-III tubulin antibodies (TexasRed) with anti-smooth muscle actin monoclonal antibody (fluorescein), DAPI (4�,6-diamidine-2-phenylidole-dihydrochloride) nuclearstain (blue), and X-gal reactivity (shown in B). Several neurons and one smooth muscle cell are visible (focus is on neurons). B: X-galreaction of two areas marked in A. C: Smooth muscle cell (anti-smooth muscle actin antibody; Texas Red). D: Rare SCIP-immunoreactiveSchwann cell. E: Corresponding phase contrast image. F: Rare S100-immunoreactive Schwann cell progenitor. G: Corresponding phasecontrast image. H: Three MelEM-immunoreactive melanocyte progenitors. I: Corresponding phase contrast image. Scale bars � 50 �min A (applies to A,B), in C, in D (applies to D–G), in H (applies to H,I).

Fig. 8. Targeted differentiation of epidermal neural crest cells into Schwann cell progenitors. To obtain larger numbers of Schwann cellprogenitors, clones were grown in the presence of neuregulin-1. Quadruple stain combining �-III tubulin (fluorescein), glial fibrillary acidicprotein (GFAP, Texas Red), DAPI (4�,6-diamidine-2-phenylidole-dihydrochloride) nuclear stain (blue), and X-gal reaction (black). A: GFAPstain with X-gal reaction product (e.g., arrow). Several Schwann cell progenitors are visible. B: DAPI nuclear stain of same area as in A.C: �-III tubulin immunoreactivity (Texas Red) in the same area is absent. Arrow depicts same cell and X-gal reaction product as in A andB. D: Merged images of GFAP (Texas Red), �-III tubulin (fluorescein, absent), DAPI nuclear stain, and X-gal reaction. This series of imagesshows that Schwann cell progenitors are present in large numbers and that they do not express a neuronal marker. In a different area inthe same clone, neurons are present (arrow). E: �-III tubulin stain; several multipolar neurons are present. F: Merged brightfield and �-IIItubulin images of the same area as in E to show that neurons contain X-gal reaction product (arrows in E, F, G, and I). G: CorrespondingDAPI nuclear stain. H: GFAP immunoreactivity in the same area is absent (Texas Red). I: Merged images in pseudocolor of �-III tubulin(fluorescein), GFAP (Texas Red, absent), and DAPI nuclear stain (blue). The data show that neurons do not express a Schwann cell marker.Scale bar � 10 �m in I (applies to A–I).

Fig. 9. Targeted differentiation of epidermal neural crest cells (eNCSC) into chondrocytes. To determine whether eNCSC are able togenerate chondrocytes, we grew clones in the presence of BMP-2. After 2 weeks in culture, most cells in clones were immunoreactive forcollagen type II, a marker for chondrocytes. Triple stain combining collagen type II immunoreactivity (Texas Red), X-gal reaction (black),and DAPI (4�,6-diamidine-2-phenylidole-dihydrochloride) nuclear stain (blue). A: Collagen type II immunoreactivity. Six cells are visible,and they contain X-gal reaction product (e.g., arrows). B: Merged images in pseudocolor of collagen type II immunoreactivity (TexasRed), DAPI (blue), and X-gal (black). C: Brightfield image of the same area as in A and B to better visualize the X-gal reaction product(e.g., arrows). Scale bar � 50 �m in C (applies to A–C).


Fig. 7.

Fig. 8.

Fig. 9.

of Sox9 (Zehentner et al., 2002). We,therefore, cultured bulge-derivedneural crest cells in the presence ofBMP-2 (10 ng/ml) for 2 weeks inclonal culture. Under these condi-tions, most cells in clones becomecollagen type II immunoreactive, in-dicating that they have differenti-ated into chondrocytes. Figure 9Ashows a group of collagen type II-immunoreactive cells with X-gal re-action product (arrows). Figure 9Bshows merged images of collagentype II immunoreactivity (Texas Red),DAPI (blue), and X-gal (black, ar-rows). Figure 9C shows the samearea with brightfield optics to bettervisualize the X-gal reaction product.The data show that BMP-2 directsbulge-derived neural crest cells inclonal culture to differentiate alongthe chondrogenic cell lineage.

Verification of Neuronal andSchwann Cell Differentiation byReverse Transcriptase-Polymerase Chain Reaction

To establish the specificity of the an-tibody stains and to test additionalneuronal markers by differentmeans, we performed reverse tran-scriptase-polymerase chain reac-tion (RT-PCR) for three neuronal andthree Schwann cell markers. Bulge-derived cells grown for 2 weeks inculture medium supplemented withneuregulin-1 express the Schwanncell markers GFAP, SCIP/Oct6, andprotein zero (P0) abundantly (Fig.10). Cells that were grown in regularculture medium in the absence ofneuregulin-1 express �-III tubulin, pe-ripherin, and microtubule-associ-ated protein (MAP2; Fig. 10). The au-thenticity of the RT-PCR productswere verified by sequencing.

Bulge-Derived Neural CrestCells Can UndergoSelf-Renewal

We determined whether bulge-de-rived neural crest cells can undergoself-renewal by serial cloning in vitro.To this end, primary clones were pre-pared. Either 3 or 5 days later, pri-mary clones were resuspended bytrypsinization with the aid of a glasscloning ring. The resuspended clone

was seeded again at clonal density(20–50 cells per 35-mm cultureplate) and incubated for 2 weeks(Fig. 11A). Figure 11B shows a 5-day-old secondary clone. The insertshows the secondary clone-formingcell shortly after plating at highermagnification. Figure 11C shows ahigher magnification of the areamarked in Figure 11B. The morphol-ogy of cells in secondary clones issimilar to that in primary clones andin primary explants. At the end of theculture period, secondary cloneswere analyzed with cell type-spe-cific antibodies. Figure 11D–I shows atriple stain that combines X-gal (Fig.11F,I, black dots, arrows) with �-III tu-bulin (Fig. 11E, Texas Red) andsmooth muscle actin (Fig. 11G, fluo-rescein) antibodies. A �-III tubulin-im-munoreactive neuronal cell is shownin (Fig. 11E); it does not expresssmooth muscle actin (Fig. 11D). In adifferent area of the same second-ary clone, a group of smooth musclecells (Fig. 11G) is present, which donot express �-III tubulin (Fig. 11H). Thedata indicate that the primary clonecontained stem cells, which wereable to generate at least two dis-tinct differentiated cells types, whichfulfills the criterion for self-renewal.

Primary clones from day 4 bulgeexplants comprise 83.0 � 2.7% of allcolonies. The percentage of sec-ondary clones formed by pluripotentcells is 73.5 � 6.7% when taken fromday 3 primary clones and 66.2 �4.4% when prepared from day 5 pri-mary clones (Table 1). Thus, the por-tion of stem cells is maintained atrelatively high levels over an esti-mated total of 18 doublings in pri-mary and clonal culture, despitethat our culture medium was devel-oped to support the differentiationof neural crest cells (Ito and Takeu-chi, 1984; Ito et al., 1993). Taken to-gether, we have shown that bulge-derived neural crest cells arepluripotent and that they can un-dergo self-renewal. Thus, bulge-de-rived neural crest cells fulfill the crite-ria for pluripotent stem cells(epidermal neural crest stem cells,eNCSC).

eNCSC Are Distinctly DifferentFrom Schwann Cell Progenitorsof the Adult Sciatic Nerve

Because whiskers follicles are inner-vated by myelinated nerves, wesought to determine whethereNCSC are, in fact, Schwann cell

Fig. 10. Reverse transcriptase-polymerase chain reaction to detect Schwann cell- andneuron-specific gene expression. In cells grown in the presence of neuregulin (10 nM), glialfibrillary acidic protein (GFAP), SCIP/Oct6, and protein zero (P0) were abundantly ex-pressed. Likewise, in cultures that were grown in nonsupplemented culture medium, theneuronal genes �-III tubulin, peripherin, and microtubule-associated protein (MAP2) wereexpressed.


Fig. 11. Serial cloning. We have deter-mined that bulge-derived neural crest cellscan undergo self-renewal. A: Cells from pri-mary bulge explants were put into clonal cul-ture and let grow into a colony (primaryclone). Clones were subsequently isolatedby putting a glass cloning ring around them,and they were then resuspended by addingtwo drops of 0.025% trypsin in 0.02% ethyl-enediaminetetraacetic acid in phosphate-buffered saline. When the cells detached,trypsinization was stopped by 1 mg/ml oftrypsin inhibitor in culture medium. The cellswere aspirated, diluted with 1.5 ml of culturemedium, and placed into a new collagen-coated 35-mm plate (secondary clones).Plates with secondary clones were grown for2 weeks and then analyzed with cell type-specific antibodies. B: Five-day-old second-ary clone. Insert: The secondary clone-form-ing cell. C: Higher magnification of areamarked in B. Cell morphology in secondaryclones resembles that in primary clones. After2 weeks, secondary clones were fixed andprocessed for immunocytochemistry. D–I: Tri-ple stain combining �-III tubulin (Texas Red)with smooth muscle actin (fluorescein) andX-gal reaction (black). D: Smooth muscle ac-tin (fluorescein) immunoreactivity is absent.E: �-III tubulin stain in the same area as in Dshows one neuronal cell. F: The correspond-ing phase contrast image shows X-gal reac-tion product (arrows). G: Smooth muscle ac-tin immunoreactivity (fluorescein) showsseveral smooth muscle cells (arrows) in a dif-ferent area of the same secondary clone.H: Same area as in G, but with Texas Redillumination. Smooth muscle actin-immuno-reactive cells do not express the neuronalmarker. I: Corresponding phase contrast im-age with X-gal reaction product (arrows).Scale bars � 100 �m in B, 100 �m in C, 10 �min I (applies to D–I).

Fig. 12. Are epidermal neural crest cells (eNCSC) identical with adult Schwann cell progenitors? Whisker follicles are innervated bymyelinated nerves. Thus, our preparation could be contaminated by nerve endings. To determine whether eNCSC are adult Schwann cellprogenitors, we compared eNCSC with Schwann cell precursors from adult sciatic nerve explants. A: Nestin/DAPI (4�,6-diamidine-2-phenylidole-dihydrochloride) double labeling in day 5 bulge explant. Virtually all cells, in particular stellate cells (e.g., long arrow), areintensely nestin immunoreactive (Texas Red). Cells with the morphology of differentiating smooth muscle cells (short arrow) express lowerlevels of nestin. B: By contrast, cells at culture day 5 that had emigrated from adult sciatic nerve explants express nestin at low levels onlyor not at all (e.g., arrow; Texas Red). C: The same cells, however, were intensely immunoreactive for SCIP (fluorescein fluorescence, arrow),a marker for Schwann cell progenitors (Zorick et al., 1996), whereas eNCSC differentiate only rarely into SCIP-positive cells under theseculture conditions (Fig. 7D). D: Nestin/DAPI merged images of area shown in B and C. E: SCIP/DAPI merged images of area shown in B,C, and D. It can be concluded, thus, that eNCSC and adult Schwann cell progenitors are distinctly different types of cell. Scale bar � 50�m in A, 50 �m in E (applies to B–E).

progenitors derived from contami-nating nerve endings. We thereforedetermined similarities and differ-ences between cells that emigratefrom whisker bulge explants andfrom sciatic nerve explants fromadult mice. Both types of tissuewere cultured under the same con-ditions in regular culture medium.Nestin is a marker for both neuralstem cells (Lendahl et al., 1990)and neural crest stem cells (Lothianand Lendahl, 1997; Mujtaba et al.,1998; Josephson et al., 1998). As ex-pected, bulge-derived cells ex-press nestin at high levels (Fig. 12A;e.g., arrow). Differentiating cellswith smooth muscle cell morphol-ogy expressed nestin at lower lev-els (Fig. 12A, arrowhead). By con-trast, cells that have emigratedfrom adult sciatic nerve explantsexpress nestin at low levels only ornot at all (Fig. 12B, e.g., arrow).Whereas SCIP-positive cells are ob-served only very rarely in bulge-de-rived cultures (Fig. 7D,E) underthese culture conditions, all sciaticnerve-derived cells express SCIP athigh levels. Figure 12C,E shows SCIPexpression in the same cells shownin Figure 12B,D.

At 24 hr of clonal culture, 73.7 �4.9% of sciatic nerve-derived colo-nies consist of 1–2 cells with Schwanncell-like morphology. These cells diewithin the second 24 hr in clonal cul-ture, suggesting that they were de-pendent on axonal contact. The re-maining 26.3% of colonies contain9.3 � 0.9 cells per clone at 48 hr and12.8 � 1.6 cells at 72 hr in clonalculture, suggesting a low rate of pro-liferation that is possibly combinedwith cell death.

Thus, while eNCSC have a highrate of proliferation and express nes-tin at high levels but not SCIP, sciaticnerve-derived cells have a lowerrate of proliferation under the sameculture conditions, most of them diewithin the first 48 hr, and they areintensely immunoreactive for SCIP,but not for nestin. These observationsdemonstrate that adult bulge-de-rived neural crest cells and adult sci-atic nerve-derived Schwann cellprogenitors are distinctly differenttypes of cell.

DISCUSSION

In this study, we have documentedthe existence of pluripotent neuralcrest stem cells, termed eNCSC, inthe bulge region of the adult mousewhisker follicle. The presence of neu-ral crest-derived cells in the hair folli-cle is not surprising, as neural crest-derived melanocyte progenitorshave been described in this locationby Peters et al. (2002) and Nishimuraet al. (2002). Moreover, Merkel cells,which are also of neural crest origin(Szeder et al., 2003), are present inlarge numbers in the basal layer ofthe outer root sheath in whisker folli-cles. However, the persistence in theadult organism of pluripotent stemcells that are derived from a tran-sient embryonic tissue, the neuralcrest, is surprising.

In the mouse and in all other ver-tebrate embryos studied to date,Wnt1 expression during embryogen-esis is transient and limited to neuralcrest cells and dorsal neural tubecells (Wilkinson et al., 1987; Davis etal., 1988; Molven et al., 1991; McMa-hon et al., 1992; Wolda et al., 1993;Echelard et al., 1994). In E9.5 double-transgenic Wnt1-cre/R26R mice,dorsal neural tube cells, premigra-tory neural crest cells, and migratoryneural crest cells, but no other celltypes, express cre as detected withan anti-cre antibody (Sieber-Blum etal., 2003). At E16.5, epidermal cellsexpress the Wnt-3 and Wnt-5a genes(St-Jacques et al., 1998; Millar et al.,1999; Reddy et al., 2001; Fuchs et al.,2001) but not Wnt-1. As an importantnegative control, we have never-theless confirmed the absence ofWnt-1 expression in epidermal cellsof hair follicles and surface skin(Sieber-Blum et al., 2003). The expres-sion by bulge-derived cells of Sox-10,a neural crest marker, and of nestin,a neural crest stem cell marker(among other stem cells), supportthe notion that X-gal–positive cellsare indeed neural crest cells, as dotheir differentiated progeny inclones (neurons, Schwann cells,smooth muscle cells, melanocytes,and chondrocytes), all of which arephysiological neural crest deriva-tives.

eNCSC are located in the epider-mis of the hair follicle. Stem cells with

neurogenic and gliogenic potentialhave been isolated also from thedermis (SKP, skin-derived precursors;Toma et al., 2001). Comparisons be-tween the two studies are difficult,because there are fundamental dif-ferences in the techniques appliedto obtain eNCSC and SKP cells. Wehave isolated eNCSC from a dis-crete location, the epidermal follic-ular bulge. By contrast, SKP cells areobserved only when dermis ispresent and neural crest cell markersare not expressed (Toma et al.,2001). We have cloned minimally ex-panded eNCSC with the aim to pre-serve their innate characteristics. Incontrast, Toma et al. (2001) start withcultures of high cell density that areexpanded for several weeks. Cur-rently available data thus suggestthat eNCSC and SKP cells are un-likely to be of the same origin.

Our observation of neural creststem cells in the follicular epidermishas implications for the interpreta-tion of the work by Kobayashi et al.(1993), Oshima et al. (2001), Li et al.(2003), and Nishimura et al. (2002).By in vitro clonal analysis, Kobayashiet al. (1993) have identified keratin-ocyte precursors with stem cell prop-erties in the bulge region of the ratwhisker follicle. During hair growth,these stem cells migrate from thebulge region to the matrix at thebase of the follicle to participate inhair regeneration (Kobayashi et al.,1993; Oshima et al., 2001; Li et al.,2003). Combined, the different stud-ies suggest that the bulge cell pop-ulation is of mixed origin, containingkeratinocyte stem cells, pluripotentneural crest stem cells, and stemcells that are committed to the mel-anogenic lineage, all of which mayundergo hair cycle-related migra-tion within the follicle.

Nishimura et al. (2002) used theDct-lacZ mouse, in which �-galacto-sidase is expressed under the controlof the dopachrome tautomerasepromoter. Because this enzyme isconsidered to be specific for mela-nocytes, lacZ-positive cells in theNishimura study identify cells that arecommitted to the melanogenic lin-eage. It is likely that a subset of X-gal–positive cells observed in Wnt1-cre/R26R whisker follicles areidentical to the X-gal–positive cells in


Dct-lacZ follicles, in particular as83.0 � 2.7%, rather than 100%, ofclones in our cultures were formedby pluripotent stem cells (Table 1). Ifcommitted pigment cell progenitorsare present in the bulge, why did wenot observe pigmented colonies inour clonal cultures? There are twopossible explanations. First, the ma-jority of our Wnt1-cre/R26R mice arewhite and, therefore, do not containrecognizable melanized pigmentcells. The second reason concernsour culture medium. As formulatedoriginally by Ito et al. (Ito and Takeu-chi, 1984; Ito et al., 1993), the culturemedium contained the tumor pro-moter TPA and cholera toxin to pro-mote melanogenesis. Neurons donot develop in these culture condi-tions (M.S.-B., unpublished observa-tion). In contrast, our primary interestlies in neurogenesis. Therefore, wedid not supplement the culture me-dium with TPA and cholera toxin,which makes it conducive to neuro-nal differentiation but does not sup-port melanogenesis.

eNCSC are intriguing for severalreasons. First, like embryonic neuralcrest stem cells, eNCSC have an in-nate high degree of plasticity. Theycan give rise to the entire array ofcranial neural crest cells, includingneurons, Schwann cells, smoothmuscle cells, chondrocytes, melano-cytes, and possibly other cell typesthat originate from the cranial neu-ral crest. Second, they are abun-dant and easily accessible. Thus,eNCSC are attractive candidatesfor diverse cell therapy applications.Because they are located in an ac-cessible tissue, only minimally inva-sive surgery will be necessary to har-vest them. Furthermore, becausethey can be obtained from the livingorganism, they are good candi-dates for autologous transplanta-tion, which will avoid both transplantrejection and graft-versus-host dis-ease.

EXPERIMENTAL PROCEDURES

Animals and Genotyping

Heterozygous Wnt1-cre mice weremated with R26R heterozygotes.Genotyping was performed exactlyas described (Szeder et al., 2003).

Bulge Explants

Whiskers were dissected from the whis-ker pad of 8- to 10-week-old Wnt1-cre/R26R double-transgenic mice accord-ing to Baumann et al. (1996). The con-nective tissue was scraped from thefollicle with a bent electrolyticallysharpened tungsten needle andrinsed several times, thus exposing thering sinus and cavernous sinus (Fig.1B). The capsule was then cut longitu-dinally with a small scalpel, the bloodflushed with a stream of buffer, thefollicle cross-sectioned first at the levelabove the cavernous sinus and thenbelow the skin, yielding the bulge re-gion within the capsule (Fig. 1C). Thebulge was rolled out of the capsule,rinsed three times, and plated in anempty collagen-coated 35-mm cul-ture plate that had been preincu-bated for 3 hr with culture medium.The bulge explants adhered to thesubstratum within 1 hr, at which time1.5 ml of culture medium was added.The culture medium was designed toaccommodate the survival and pro-liferation of neural crest stem cells, aswell as their differentiation into multi-ple phenotypes, including neurons,smooth muscle cells, glia, chondro-cytes, and melanocytes (Ito et al.,1993; Sieber-Blum, 1999). It consistedof 75% alpha-modified MEM medium,5% day 11 chick embryo extract, and10% of fetal calf serum (HyClone), andit was supplemented with 1 �g/mlgentamicin as described previously(Sieber-Blum, 1999). Fifty percent ofthe culture medium was exchangedevery other day.

Clonal Cultures

Cells started to emigrate from ex-plants 48–72 hr postexplantation. At4 days postexplantation, the bulgeexplant was removed, leaving theemigrated cells on the collagen sub-stratum. The emigrated cells werethen resuspended by trypsin diges-tion exactly as described for mouseneural crest cells (Sieber-Blum, 1999).The percentage of single cells was100, due to the sparse arrangementwithin early primary explants (Fig.3A,C,E). Cells were plated at 20 cellsper cm2, which at a plating efficiencyof approximately 10% yielded 8–10clones per 35-mm culture plate. Thecells adhered within 30 min. One

hour after plating, single cells weremarked by circling the underside ofthe plate with a diamond marker (4mm diameter). Fifty percent of theculture medium was exchanged ev-ery other day.

Subclones were prepared by re-moving the culture medium andplacing a glass cloning ring aroundthe clone. The clones within the ringswere then rinsed with phosphate-buffered saline (PBS) and subse-quently detached by trypsinizationas described for primary explants.The clonal cell suspension consistingof 20–50 cells was subsequentlyplaced into a new 35-mm culturedish.

X-gal Reaction and IndirectImmunocytochemistry

Cultures were fixed with 4% parafor-maldehyde for 5 min at room tem-perature, and X-gal histochemistrywas performed according to Galileoet al. (1990) with the modificationthat potassium ferricyanide and po-tassium ferrocyanide were used at 5mM and incubation was overnightat 30°C. Subsequently, the cultureswere processed for indirect immuno-cytochemistry as follows. Cells werepost-fixed with 4% paraformalde-hyde for 30 min on ice, rinsed 3 � 10min with PBS, blocked with 2% nor-mal goat serum for 20 min and thenincubated with pooled primary anti-bodies overnight in the cold. Subse-quently, the plates were rinsed 3 �10 min with PBS, incubated withpooled secondary antibodies for 2hr at room temperature, rinsed 4 �20 min with PBS, stained with DAPInuclear stain (3 �M; MolecularProbes, Eugene, OR), rinsed again,and finally mounted with ProLongAntifade (Molecular Probes) andcover-slipped. The following primaryantibodies were used: mouse mono-clonal antibody against smoothmuscle actin (1:800; Sigma, St. Louis,MO); mouse monoclonal antibodyagainst neuron-specific �-III tubulin(1:200; Chemicon, Temecula, CA);polyclonal rabbit anti–�-III tubulinantibodies (1-400; gift of A. Frankfurt-er; Lee et al., 1990); rabbit polyclonalantibodies against S100 protein (1:200; Novocastra Laboratories, New-castle upon Tyne, UK); rabbit anti-


SCIP antibodies (1:300; gift of G.Lemke; Zorick et al., 1996); MelEM(melanocyte marker; 1:1; Hybrid-oma Bank; Nataf et al., 1993); anti-Sox10 rabbit serum (1:100; Chemi-con), nestin mouse monoclonalantibodies (1:400; BD Biosciences);and mouse anti-GFAP ascites fluid(1:500; Chemicon). The followingsecondary antibodies were used ata dilution of 1:200: Texas Red-conju-gated and fluorescein-conjugatedgoat-anti-mouse immunoglobulin G(IgG) and goat anti-rabbit IgG des-ignated for multiple labeling (Jack-son ImmunoResearch, West Grove,PA). The X-gal reaction in tissue sec-tions was performed exactly as de-scribed (Szeder et al., 2003).

RT-PCR

By using Trizol reagent (Invitrogen),total RNA was prepared from cellsthat were grown for 2 weeks either inthe presence of neuregulin1 (10 nM;for Schwann cell markers) or in itsabsence (for neuronal markers). Re-verse transcription was performedwith 3 �g of total RNA using Super-Script II reverse transcriptase (Invitro-gen). PCR amplification was carriedout with 50 ng of reverse transcribedDNA template, 10 pmol of primers,and 0.2 mM dNTP. The PCR reactionconsisted of denaturation at 94°Cfor 45 sec, annealing for 45 sec (tem-perature dependent on the primerpair), and extension at 72°C for 1min. Annealing temperatures wereas follows: P0, 59°C; MAP2 andGFAP, 58°C; beta-III tubulin and pe-ripherin, 55°C; SCIP/Oct6, 52°C. PCRproducts were electrophoresed on2% agarose gels and stained withethidium bromide. The following primerswere used. P0: forward, 5�-ACT-ATGCCAAGGGACAACCTTACATC-3�;reverse, 5�-ACATAGAGCGTGACCT-GAGAGGTC-3�; product size, 196 bp.MAP2: forward, GGCCCAAGCTAA-AGTTGG-3�; reverse, 5�-CAAGCCA-GACCTCACAGCG-3�; product size,215 bp. Neuron-specific �-III tubulin:forward, 5�-CCCGTGGGCTCAAAA-TGT-; reverse, 5�-TGGGGGCAGTGT-CAGTAGC-3�; product size, 380 bp.SCIP/Oct6: forward, 5�-AAGAACAT-GTGCAAGCTCAA-3�; reverse, 5�-ACAACAAAAAGAGTCCAGGC-3�;product size 528 bp. GFAP; forward,

5�-CAAGCCAGACCTCACAGCG-3�;reverse, 5�-GGTGTCCAGGCTGGTTT-CTC-3�; product size 508 bp. Peripherin:forward, 5�-ACAGCTGAAGGAAGA-GATGG-3�; reverse, 5�-GATTTGCTGTC-CTGGGTATC-3�; product size 538 bp.The RT-PCR products were se-quenced for verification.

ACKNOWLEDGMENTSWe thank A.P. McMahon, P. Soriano,and H. Sucov for providing the Wnt1-cre and R26R mouse lines and G.Lemke for the SCIP antibody. Wealso thank Solomon Senok for his ad-vice on follicle dissection and Ste-phen Duncan for his comments onthe manuscript. Additionally, wethank Eva Kluzakova and Joan Wardfor excellent technical assistance.M.S.-B. was funded by the NationalInstitute of Neurological Disordersand Stroke, NIH; USPHS; and theRockefeller Brothers Fund, New York;Charles E. Culpeper Biomedical PilotInitiative. M.G. was funded by theMinistry of Education of the CzechRepublic.

REFERENCES

Alexanian AR, Sieber-Blum M. 2003. Differ-entiating adult hippocampal stemcells into neural crest derivatives. Neu-roscience 118:1–5.

Baroffio A, Dupin E, Le Douarin NM. 1988.Clone-forming ability and differentiationpotential of migratory neural crest cells.Proc Natl Acad Sci U S A 85:5325–5329.

Baumann KI, Chan E, Halata Z, Senok SS,Yung WH. 1996. An isolated rat vibrissalpreparation with stable responses ofslowly adapting mechanoreceptors.Neurosci Lett 26:1–4.

Bixby S, Kruger GM, Mosher JT, JosephNM, Morrison SJ. 2002. Cell-intrinsic dif-ferences between stem cells from dif-ferent regions of the peripheral nervoussystem regulate the generation of neu-ral diversity. Neuron 35:643–656.

Bronner-Fraser M, Fraser SE. 1988. Cell lin-eage analysis reveals multipotency ofsome avian neural crest cells. Nature335:161–164.

Chai Y, Jiang X, Ito Y, Bringas P Jr, Han J,Rowitch DH, Soriano P, McMahon AP,Sucov HM. 2000. Fate of the mamma-lian cranial neural crest during toothand mandibular morphogenesis. De-velopment 127:1671–1679.

Danielian PS, Muccino D, Rowitch DH, Mi-chael SK, McMahon AP. 1998. Modifi-cation of gene activity in mouse em-bryos in utero by a tamoxifen-inducibleform of Cre recombinase. Curr Biol 8:1323–1326.

Davis CA, Noble-Topham SE, Rossant J,Joyner AL. 1988. Expression of the ho-meo box-containing gene En-2 delin-eates a specific region of the develop-ing mouse brain. Genes Dev 2:361–371.

Duff RS, Langtimm CJ, Richardson MK,Sieber-Blum M. 1991. In vitro clonalanalysis of progenitor cell patterns indorsal root and sympathetic ganglia ofthe quail embryo. Dev Biol 147:451–459.

Echelard Y, Vassileva G, McMahon AP.1994. Cis-acting regulatory sequencesgoverning Wnt1 expression in the de-velopingmouseCNS.Development120:2213–2224.

Friedrich G, Soriano P. 1991. Promotertraps in embryonic stem cells: a ge-netic screen to identify and mutate de-velopmental genes in mice. GenesDev 5:1513–1523.

Fuchs E, Merrill BJ, Jamora C, DasGuptaR. 2001. At the roots of a never-endingcycle. Dev Cell 1:13–25.

Galileo SA, Gray GE, Owens GC, MajorsJ, Sanes JR. 1990. Neurons and gliaarise from a common progenitor inchicken optic tectum: demonstrationwith two retroviruses and cell type-spe-cific antibodies. Proc Natl Acad Sci U SA 87:458–462.

Gershon MD, Chalazonitis A, Rothman TP.1993. From neural crest to bowel: de-velopment of the enteric nervous sys-tem. J Neurobiol 24:199–214.

Henion PD, Weston JA. 1997. Timing andpattern of cell fate restrictions in theneural crest lineage. Development 124:4351–4659.

Ito K, Sieber-Blum M. 1993. Pluripotentand developmentally restricted neuralcrest-derived cells in posterior visceralarches. Dev Biol 156:191–200.

Ito K, Takeuchi T. 1984. The differentiationin vitro of the neural crest cells of themouse embryos. J Embryol Exp Mor-phol 84:49–62.

Ito K, Morita T, Sieber-Blum M. 1993. Invitro clonal analysis of mouse neuralcrest development. Dev Biol 157:517–525.

Jiang X, Iseki S, Maxson RE, Sucov HM,Morriss-Kay GM. 2002. Tissue origins andinteractions in the mammalian skullvault. Dev Biol 241:106–116.

Josephson R, Muller T, Pickel J, Okabe S,Reynolds K, Turner PA, Zimmer A,McKay RDG. 1998. POU transcriptionfactors control expression of CNS stemcell-specific genes. Development 125:3087–3100.

Kobayashi K, Rochat A, Barrandon Y.1993. Segregation of keratinocyte col-ony-forming cells in the bulge of the ratvibrissa. Proc Natl Acad Sci U S A 90:7391–7395.

Kypriotou M, Fossard-Demoor M, Chadji-christos C, Ghayor C, de CrombruggheB, Pujol JP, Galera P. 2003. SOX9 exertsa bifunctional effect on type II colla-gen (COL2A1) expression in chondro-cytes depending on the differentiationstate DNA. Cell Biol 22:119–129.

Kruger GM, Mosher JT, Bixby S, Joseph N,Iwashita T, Morrison SJ. 2002. Neural


crest stem cells persist in the adult gutbut undergo changes in self-renewal,neuronal subtype potential, and factorresponsiveness. Neuron 35:657–669.

Kuhlbrodt K, Herbarth B, Sock E, Hermans-Borgmeyer I, Wegner M. 1998. Sox10, anovel transcription modulator in glialcells. J Neurosci 18:237–250.

Le Douarin NM, Kalcheim C, editors.1999. The neural crest. Cambridge, UK:Cambridge University Press.

Lee MK, Rebhun LI, Frankfurter A. 1990.Posttranslational modification of classIII beta-tubulin. Proc Natl Acad Sci U SA 87:7195–7199.

Lendahl U, Zimmerman LB, McKay RD.1990. CNS stem cells express a newclass of intermediate filament protein.Cell 60:585–595.

Li L, Mignone J, Yang M, Matic M, PernmaS, Enikolopov G, Hoffman RM. 2003.Nestin expression in hair follicle sheathprogenitor cells. Proc Natl Acad Sci U SA 100:9958–9961.

Lothian C, Lendahl U. 1997. An evolution-arily conserved region in the secondintron of the human nestin gene directsgene expression to CNS progenitorcells and to early neural crest cells. EurJ Neurosci 9:452–462.

McMahon AP, Joyner AL, Braddley A,McMahon JA. 1992. The midbrain-hind-brain phenotype of Wnt1-/Wnt1- miceresults from stepwise deletion of en-grailed-expressing cells by 95 dayspost-coitum. Cell 69:581–595.

Millar SE, Willert K, Salinas PC, Roelink H,Nusse R, Sussman DJ, Barsh GS. 1999.WNT signaling in the control of hairgrowth and structure. Dev Biol 207:133–149.

Molven A, Njolstad PR, Fjose A. 1991.Genomic structure and restricted neu-ral expression of the zebrafish Wnt1(int-1) gene. EMBO J 10:799–807.

Mori-Akiyama Y, Akiyama H, Rowitch DH,de Crombrugghe B. 2003. Sox9 is re-quired for determination of the chon-drogenic cell lineage in the cranialneural crest. Proc Natl Acad Sci U S A100:9360–9365.

Mujtaba T, Mayer-Proschel M, Rao MS.1998. A common neural progenitor forthe CNS and PNS. Dev Biol 200:1–15.

Nataf V, Mercier P, Ziller C, Le DouarinNM. 1993. Novel markers of melano-cyte differentiation in the avian em-bryo. Exp Cell Res 207:171–182.

Nishimura EK, Jordan SA, Oshima H, Yo-shida H, Osawa M, Moriyama M, Jack-

son J, Barrandon Y, Miyachi Y, Nish-ikawa S-I. 2002. Dominant role of theniche in melanocyte stem-cell fate de-termination. Nature 416:854–860.

Oshima H, Rochat A, Kedzia C, Koba-yashi K, Barrandon Y. 2001. Morpho-genesis and renewal of hair folliclesfrom adult multipotent stem cells. Cell104:233–245.

Parkinson DB, Dong Z, Bunting H, WhitfieldJ, Meier C, Marie H, Mirsky R, Jessen KR.2001. Transforming growth factor beta(TGFbeta) mediates Schwann celldeath in vitro and in vivo: examination ofc-Jun activation, interactions with sur-vival signals, and the relationship of TGF-beta-mediated death to Schwann celldifferentiation. J Neurosci 21:8572–8585.

Peters EMJ, Tobin DJ, Botchkareva N,Maurer M, Paus R. 2002. Migration ofmelanoblasts into the developing mu-rine hair follicle is accompanied bytransient c-kit expression. J HistochemCytochem 50:751–766.

Reddy S, Andl T, Bagasara A, Lu MM, Ep-stein DJ, Morrisey EE, Millar SE. 2001.Characterization of Wnt gene expres-sion in developing and postnatal hairfollicles and identification of Wnt5a asa target of Sonic hedgehog in hair fol-licle morphogenesis. Mech Dev 107:69–82.

Rehberg S, Lischka P, Glaser G, Stam-minger T, Wegner M, Rosorius O. 2002.Sox10 in an active nucleocytoplas-matic shuttle protein, and shuttling iscrucial for Sox10-mediated transacti-vation. Mol Cell Biol 22:5826–5834.

Richardson MK, Sieber-Blum M. 1993. Plu-ripotent neural crest cells in the devel-oping skin of the quail embryo. Dev Biol157:348–358.

Rico B, Xu B, Reichardt LF. 2002. TrkB re-ceptor signaling is required for estab-lishment of GABAergic synapses in thecerebellum. Nat Neurosci 5:225–233.

Shah NM, Marchionni MA, Isaacs I, Stroo-bant O, Anderson DJ. 1994. Glialgrowth factor restricts mammalianneural crest stem cells to a glial fate.Cell 77:349–360.

Sieber-Blum M. 1989 Commitment of neu-ral crest cells to the sensory neuron lin-eage. Science 243:1608–1611.

Sieber-Blum M. 1999. The neural crest col-ony assay: assessing molecular influ-ences on development in culture. In:Haynes LW, editor. The neuron in tissueculture. New York: John Wiley and SonsLtd. p 5–22.

Sieber-Blum M, Cohen AM. 1980. Clonalanalysis of quail neural crest cells: theyare pluripotent and differentiate invitro in the absence of noncrest cells.Dev Biol 80:96–106.

Sieber-Blum M, Sieber F. 1984 Heteroge-neity among early quail neural crestcells. Brain Res 316:241–246.

Sieber-Blum M, Szeder V, Grim M, HalataZ. 2003. Mammalian Merkel cells areneural crest derivatives In: Baumann KI,Halata Z, Moll I, editors. The Merkel cell;structure-development-function-carci-nogenesis. Berlin: Springer-Verlag.

Soriano P. 1999. Generalized lacZ expres-sion with the ROSA26 Cre reporterstrain. Nat Genet 21:70–71.

Stemple DL, Mahanthappa NK, AndersonDJ. 1988. Basic FGF induces neuronaldifferentiation, cell division, and NGFdependence in chromaffin cells: a se-quence of events in sympathetic de-velopment. Neuron 1:517–525.

St-Jacques B, Dassule HR, Karavanova I,Botchkarev VA, Li J, Danielian PS, Mc-Mahon JA, Lewis PM, Paus R, McMa-hon AP. 1998. Sonic hedgehog signal-ing is essential for hair development.Curr Biol 8:1058–1068.

Szeder V, Grim M, Halata Z, Sieber-BlumM. 2003. Neural crest origin of mamma-lian Merkel cells. Dev Biol 253:258–263.

Toma JG, Akhavan M, Fernandez KJL,Barnabe-Heider F, Sadikot A, KaplanDR, Miller FD. 2001. Isolation of multipo-tent adult stem cells from the dermis ofmammalian skin. Nat Cell Biol 3:778–784.

Wilkinson DG, Bailes JA, Champion JF,McMahon AP. 1987. Expression of theproto-oncogene int-1 is restricted tospecific neural cells in the developingmouse embryo. Cell 50:79–88.

Wolda SL, Moody CJ, Moon RT. 1993.Overlapping expression of Xwnt-3Aand XWnt1 in neural tissue of Xenopuslaevis embryos. Dev Biol 155:46–57.

Zehentner BK, Haussmann A, Burtscher H.2002. The bone morphogenetic proteinantagonist Noggin is regulated by Sox9during endochondral differentiation.Dev Growth Differ 44:1–9.

Zorick TS, Syroid DE, Arroyo E, Scherer SS,Lemke G. 1996. The transcription factorsSCIP and Krox-20 mark distinct stagesand cell fates in Schwann cell differenti-ation. Mol Cell Neurosci 8:129–145.


pluripotent neural crest stem cells in the adult hair...

Documents