Advertisement

Contact between Dermal Papilla Cells and Dermal Sheath Cells Enhances the Ability of DPCs to Induce Hair Growth

      We previously showed that cultured rat dermal papilla cells (DPCs) retain their hair-inducing capacity on afollicular epidermal cell (EPCs). Here, we examined the hair growth–inducing capacity of differently subcultured DPCs by transplanting them, along with rat EPCs, onto the backs of nude mice (graft chamber assay). DPCs at passage (p) ≤6 (DPCsp≤6 or, more generally, low-passage DPCs) induced hair formation. However, DPCsp>30 (high-passage DPCs) had no such activity and induced only subepidermal hair follicles (HFs) that were not encapsulated by the dermal sheath (DS). Thus, we examined the effect of DS cells (DSCsp=1) on the ability of DPCsp=60 to induce hair growth by testing a mixture of these two cell types (cotransplant) in the graft chamber assay, in which DSCsp=1 and DPCsp=60 were labeled with enhanced green fluorescent protein (EGFP) and 1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocyanine perchlorate (DiI), respectively. These cotransplants generated hairs as actively as did DPCsp=6 transplants. Their HFs were encapsulated with EGFP+-DS and had DPs consisting largely of EGFP+-DPCs (47%) and DiI+-DPCs (43%), indicating a major contribution of DSCp=1-derived DPCs to HF induction. In addition, the results of in vitro coculture of DPCsp=60 and DSCsp=1 suggest that high-passage DPCs stimulate the expression of certain trichogenic genes in DSCs.

      Abbreviations

      α-SMA
      α-smooth muscle actin
      ALP
      alkaline phosphatase
      Ab
      antibody
      BMP4
      bone morphogenetic protein 4
      DiI
      1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate
      DP
      dermal papilla
      DPC
      dermal papilla cell
      DPChigh passage
      high-passage DPC
      DPClow passage
      low-passage DPC
      DS
      dermal sheath
      DSC
      dermal sheath cell
      DSCsp=1-S
      DSCsp=1 in single culture
      EGFP
      enhanced green fluorescent protein
      EPC
      epidermal cell
      FB
      fibroblast
      HF
      hair follicle
      PDT
      population doubling time

      Introduction

      The hair follicle (HF) comprises dermal and epidermal compartments whose close interactions are critical to hair-related biological and pathological processes (
      • Hardy M.H.
      The secret life of the hair follicle.
      ;
      • Inui S.
      • Fukuzato Y.
      • Nakajima T.
      • et al.
      Androgen receptor co-activator Hic-5/ARA55 as a molecular regulator of androgen sensitivity in dermal papilla cells of human hair follicles.
      ). The dermal papilla (DP)—the major dermal compartment—has roles in hair formation during embryonic morphogenesis (
      • Millar S.E.
      Molecular mechanisms regulating hair follicle development.
      ;
      • Rendl M.
      • Lewis L.
      • Fuchs E.
      Molecular dissection of mesenchymal-epithelial interactions in the hair follicle.
      ) and hair cycling after birth (
      • Botchkarev N.A.
      • Ahluwalia G.
      • Shander D.
      Apoptosis in the hair follicle.
      ). Cultured rat DP cells (DPCs) have been shown to retain their original ability to induce hair growth until passage 3 (p=3) (
      • Jahoda C.A.B.
      • Horne K.A.
      • Oliver R.F.
      Induction of hair growth by implantation of cultured dermal papilla cells.
      ). In general, rat DPCs had been thought to lose this ability when cultured to p>10 (
      • Jahoda C.A.B.
      • Horne K.A.
      • Oliver R.F.
      Induction of hair growth by implantation of cultured dermal papilla cells.
      ). However, we demonstrated the hair-inducing ability of DPCs at p=56 (DPCsp=56) using the kidney capsule assay, in which DPCsp=56 were inserted into rat sole skin and then transplanted under the kidney capsule of homologous rats (
      • Inamatsu M.
      • Matsuzaki T.
      • Iwanari H.
      • et al.
      Establishment of rat dermal papilla cell lines that sustain the potency to induce hair follicles from afollicular skin.
      ). We also showed the defective hair-inducing ability of DPCsp=20 using the skin transplantation assay, in which the DPCsp=20 inserted into the rat sole skin were transplanted onto the backs of nude mice, but histological analysis of this case revealed that the transplanted DPCs were able to induce HFs under the epidermis that did not develop hair shafts (
      • Inamatsu M.
      • Tochio T.
      • Makabe A.
      • et al.
      Embryonic dermal condensation and adult dermal papilla induce hair follicles in adult glabrous epidermis through different mechanisms.
      ). In the present study, we term these follicles “incomplete HFs,” as previously described (
      • Chuong C.M.
      • Cotsarelis G.
      • Stenn K.
      Defining hair follicles in the age of stem cell bioengineering.
      ). Correspondingly, we refer to HFs that develop hair shafts and grow up over the epidermis as “complete HFs.” Thus, the hair-inducing potential of DPCs seems to depend on the assay employed.
      The second dermal compartment—the dermal sheath (DS)—is contiguous with the DP at its base. Studies suggest that it has a role in hair induction. When the lower third of a rat's whiskers, including the DP, was removed, the remaining tissues regenerated new DP and whiskers, supposedly via differentiation of DS cells (DSCs) into DPCs (
      • Oliver R.F.
      Histological studies of whisker regeneration in the hooded rat.
      ). In addition,
      • Horne K.A.
      • Jahoda C.A.B.
      Restoration of hair growth by surgical implantation of follicular dermal sheath.
      removed the lower half of HFs and then implanted DSs isolated from the base of other HFs into the removed edge, where they formed new DP. These results suggest that the DS may contain DPC progenitor cells.
      In the present study, we investigated the relationship between the hair-inducing ability of subcultured rat vibrissa DPCs and their passage number. The hair-inducing activity of DPCs was tested in a graft chamber assay (
      • Lichti U.
      • Weinberg W.C.
      • Goodman L.
      • et al.
      In vivo regulation of murine hair growth: Insights from grafting defined cell populations on nude mice.
      ;
      • Weinberg W.C.
      • Goodman L.V.
      • George C.
      • et al.
      Reconstitution of hair follicle development in vivo: determination of follicle formation, hair growth, and hair quality by dermal cells.
      ), in which cells were transplanted, along with freshly isolated newborn rat epidermal cells (EPCs), onto the backs of nude mice. DPCsp>30 markedly reduced the complete HF-inducing ability, which was associated with the inability of the induced HF to reconstitute the DS. Notably, DPCsp>30 (high-passage DPCs (DPCshigh passage)) regained the complete hair-inducing ability when cotransplanted with cultured DSCs. Histological examination suggested that DSCs participate in the process of DPC-induced hair formation and that DS formation is critical to the normal development of hairs with hair shafts. In addition, possible interactions between DPCshigh passage and DSCs were investigated through in vitro coculture, the results of which suggest that DPCshigh passage activate the expression of certain trichogenic genes in DSCs.

      Results

      Hair-inducing ability of subcultured DPCs

      The DPCs were serially subcultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 5ngml−1 fibroblast (FB) growth factor 2 (FGF2) (DMEM/FBS/FGF2). The cells grew in this medium as actively as in the previously used medium, which used keratinocyte-conditioned medium in place of FGF2 (
      • Inamatsu M.
      • Matsuzaki T.
      • Iwanari H.
      • et al.
      Establishment of rat dermal papilla cell lines that sustain the potency to induce hair follicles from afollicular skin.
      ), with 40–60hours for the population doubling time (PDT) (see Supplementary Figure S1a online). DPCs of different passage numbers (p=5, 6, 9, 15, 31, and 39) were used as HF-inducer cells. They were mixed with responder cells (rat EPCs) and filling cells (rat sole skin FBs) and transplanted onto the backs of nude mice for testing in the graft chamber assay. DPCsp≤6 were labeled with enhanced green fluorescent protein (EGFP) to track following transplantation (
      • Inamatsu M.
      • Tochio T.
      • Makabe A.
      • et al.
      Embryonic dermal condensation and adult dermal papilla induce hair follicles in adult glabrous epidermis through different mechanisms.
      ); DPCsp>6 were labeled with 1,1’-dioctadecyl-3,3,3’,3’- tetramethylindocarbocyanine perchlorate (DiI) immediately before transplantation, because EGFP fluorescence weakened at p>6 to an undetectable level.
      The graft sites were photographed 3 weeks posttransplantation. Hairs (complete HFs) were abundantly produced from transplants that included DPCsp=5 (DPCp=5 transplants, Figure 1a) and DPCp=9 transplants (Figure 1b). Hair growth, in terms of both numbers and hair-shaft length, decreased significantly in DPCp=15 transplants (Figure 1c): mean hair-shaft length of transplants of DPCp=5, DPCp=9, and DPCp=15 was 5.6±1.9mm (mean±SD, n=20), 5.3±1.6mm (n=16), and 2.1±0.6mm (n=12), respectively. No appreciable hair growth was observed with DPChigh passage (DPCp=31 or DPCp=39) transplants (Figure 1d for DPCp=39 transplants), but closer observation revealed the presence of some short hairs in these DPChigh passage transplants as described below.
      Figure thumbnail gr1
      Figure 1Correlation of the hair-inducing ability of cultured dermal papilla cells (DPCs) with dermal sheath (DS) formation in the induced hair follicles (HFs). Cell mixtures of epidermal cells (EPCs), fibroblasts (FBs), and DPCs with different passage numbers (PNs) (p=5, 6, 9, 15, 31, and 39) were tested by the graft chamber assay. DPCsp=5, 6 and DPCsp≥9 were EGFP+ and DiI+, respectively. The graft sites of DPCs at p=5 (a), 9 (b), 15 (c), and 39 (d) were photographed 3 weeks posttransplantation. The EGFP+-DPCsp=6 (eg) and DiI+-DPCsp=31 grafts (hj) were histologically examined with serial frozen and paraffin-embedded sections, respectively, by staining with hematoxylin and eosin (H&E) (e, h), Hoechst (f, i), and anti-α-smooth muscle actin (α-SMA) antibodies (g, j). # in e and h indicates the EGFP+- and DiI+-DPs, respectively. Only rectangular regions in e and h are shown for the EGFP signal (green) and Hoechst-stained nuclei (blue) (f), and for DiI signal (red) and Hoechst (blue) (i), respectively, and for α-SMA signal (brown) (g and j, respectively). EGFP+- and DiI+-DPs are enclosed with broken lines in f and i, respectively. HFs in f are enclosed in DSs that are EGFP+ and α-SMA+, but those in i are not enclosed in DS. DP, dermal papilla; Ec, erythrocyte; HM, hair matrix; and HS, hair shaft; *, capillary vessel. Scale bar (ad)=5mm, (ej)=100μm.

      Histological characteristics of induced HFs

      To assess the morphological characteristics of EGFP+-DPCp=6 transplants and track the transplanted cells, serial sections were prepared and stained with hematoxylin and eosin (H&E) (Figure 1e), Hoechst 33258 (Figure 1f), and an antibody (Ab) raised against α-smooth muscle actin (α-SMA), a marker of the DS (
      • Jahoda C.A.B.
      • Reynolds A.J.
      • Chaponnier C.
      • et al.
      Smooth muscle α-actin is a marker for hair follicle dermis in vivo and in vitro.
      ) (Figure 1g).
      In the EGFP+-DPCp=6 transplants, abundant HFs were formed (Figure 1e, marked with #), whose DP contained EGFP+ cells (Figure 1f, stained green), indicating that the transplanted DPCs formed DP and participated in HF induction. EGFP+ cells were also seen in the tissues surrounding the HFs (Figure 1f, arrows). These EGFP+ tissues were shown by the presence of α-SMA+ to be DS (Figure 1g, arrowheads), indicating that the transplanted DPCs converted to DSCs that participated in reconstructing the DS in the induced HFs. The cultured DPCs used in this series of transplantation experiments were α-SMA+ in culture, but the DP of the HFs induced by these DPCs was α-SMA (Figure 1g and j), in support of the findings of a previous study (
      • Jahoda C.A.B.
      • Reynolds A.J.
      • Chaponnier C.
      • et al.
      Smooth muscle α-actin is a marker for hair follicle dermis in vivo and in vitro.
      ).
      Serial sections from the DiI+-DPCp=31 transplants were similarly stained with H&E (Figure 1h), Hoechst (Figure 1i), and an anti-α-SMA Ab (Figure 1j). Considerable numbers of HFs were generated in the transplants (Figure 1h; #), although, as described above, the DPCp=31 transplants did not foster macroscopically appreciable hair growth. The DiI+-DP did yield HFs (Figure 1i, stained red), indicating the participation of the transplanted DPCsp=31 in forming DP. Note that the DiI+-DPCp=31-induced HFs were “bare,” that is, not encapsulated by α-SMA+ tissue (Figure 1j), in contrast to the DPCp=6-induced HFs. Thus, DPCshigh passage (DPCsp=31 in this case) were able to induce DS lacking incomplete HFs, suggesting that DS formation is a prerequisite for incomplete HFs to develop to hairs.

      Hair induction in transplants with both DPCshigh passage and DSCs

      The finding that DPCsp=31 induced the formation of DS lacking incomplete HFs prompted us to examine the effect of DSCs on the formation of HFs by DPCshigh passage in the graft chamber assay in which cultured DSCs were cotransplanted with DPCshigh passage.

      Growth potential of DSCs

      DSCs were subcultured in DMEM/FBS/FGF2 as DPCs. DSCs propagated up to p=3 with PDTs=40–50hours (Supplementary Figure S1a online), but PDT dramatically increased (to >110hours) at p>6, perhaps as a result of the medium used not being optimized for DSCs or the DSCs possessing an inherently poor replicative potential. FGF2 was essential for DSC proliferation: PDT of DSCsp=2 became ∼370hours in its absence. We found that early-passage DSCs exhibit the ability to support the hair-inducing potential of DPCshigh passage when cotransplanted with DPCshigh passage as detailed below, but DSCsp≥6 lost the ability (data not shown). DSCsp=1 were used in the following cotransplantation experiments.

      Experimental conditions of cotransplantation

      The graft chamber assay for the cotransplantation experiment was performed by making some modifications to the graft chamber assay shown in Figure 1: the “total cell number per transplantation” and “number ratio of DPCs to FBs” were reduced (as shown in Supplementary Table S1 online), which did not affect the hair-inducing capacity of the transplants. Five cell mixture types (I–V) were prepared (Supplementary Table S1 online), all of which contained EPCs and FBsp=3 as the responder cells and the filling cells, respectively. Thus, these cell types are not included in the following descriptions of the cell compositions of each transplant type (except in the case of type V, in which FBsp=3 were used as negative-control HF-inducer cells, as well as filling cells): type I (DPCsp=6), comprising DiI+-DPCsp=6; type II (DPCsp=60), DiI+-DPCsp=60; type III (DPCsp=60/DSCsp=1), DiI+-DPCsp=60 and EGFP+-DSCsp=1; type IV (DSCsp=1), EGFP+-DSCsp=1; and type V (FBsp=3) (control), EPCs and FBsp=3. In this study, each type of transplant was repeatedly assayed for HF development (that is, 6–10 times, one host mouse each time). The results obtained for each transplant type were reproducible. Representative results are shown here.

      Gross appearance of the HFs formed by the five types of transplants

      The above-mentioned five types of transplants were subjected to the graft chamber assay and the graft sites were photographed 3 weeks posttransplantation (Figure 2a–e). Type I (DPCsp=6) transplants frequently produced long hairs (Figure 2a), whereas those of type II (DPCsp=60) elicited sparse growth of thin, short hairs (Figure 2b). Such sparsely separated hairs were also formed from type V (FBsp=3) transplants (Figure 2e), which suggests that their formation is induced by the contaminating newborn rat dermal cells in the EPC preparations, because the FBs that were included in this transplant type as control cells have no HF-inducing ability (
      • Horne K.A.
      • Jahoda C.B.
      • Oliver R.F.
      Whisker growth induced by implantation of cultured vibrissa dermal papilla cells in the adult rat.
      ). Note that type III (DPCsp=60/DSCsp=1) transplants yielded long hairs at a rate comparable to that of type I (DPCsp=6), as shown in transplants (Figure 2c), which indicates that the presence of DSCs may restore the lost capacity of DPCshigh passage to induce hair growth. Type IV (DSCsp=1) transplants generated sparsely separated hairs that were shorter than those produced from type I (DPCsp=6) transplants but longer than those formed from type II (DPCsp=60) transplants (Figure 2d).
      Figure thumbnail gr2
      Figure 2Hair-inducing ability of dermal papilla cells at passage 60 (DPCsp=60) in the presence of dermal sheath cells (DSCs). Five types of cell mixture (I–V) were transplanted on the backs of nude mice: types I (DiI+-DPCsp=6); II (DiI+-DPCsp=60); III (DiI+-DPCsp=60/ EGFP+-DSCsp=1); IV (EGFP+-DSCsp=1); and V (FBsp=3). The graft sites were photographed (ae) 3 weeks posttransplantation and analyzed histologically (fw). The results of type I–V transplantations are shown in (a, f, k, p), (b, g, l, q, t), (c, h, m, r, u), (d, i, n, s), and (e, j, o), respectively. Serial paraffin sections were treated with H&E (fj), Hoechst (ko, blue), and anti-EGFP antibodies (s and u, brown, arrows). DiI signals were viewed through a fluorescence microscope (pr). Hoechst and DiI signals were superimposed (t and v). Images in fj were assembled from composite photos. The solid rectangular region marked by k/p in f is selected for the photos marked with k (Hoechst) and p (DiI). Similarly, the regions of l/q and t in g correspond to figures l and q, and t, respectively, and u in h, s in i, and o in j to u, s, and o, respectively. The broken rectangular regions with m/r in u and n in s correspond to m and r, and n, respectively. The fluorescence photo of v was taken from a section of type I (DiI+-DPCsp=6) graft, and the arrows point to host blood capillaries composed of cells with brightly fluorescent nucleoli. The fluorescence photo of w was obtained from a type IV (EGFP+-DSCsp=1) graft section. Induced HFs and their DPs are outlined by broken lines in photos (kr, t, and w) referring to the corresponding H&E and fluorescence photos. DP, dermal papilla; Ec, erythrocyte; HM, hair matrix; and IRS, inner root sheath. Scale bar (ae)=1mm, (fw)=100μm.
      Figure thumbnail gr3
      Figure 2Hair-inducing ability of dermal papilla cells at passage 60 (DPCsp=60) in the presence of dermal sheath cells (DSCs). Five types of cell mixture (I–V) were transplanted on the backs of nude mice: types I (DiI+-DPCsp=6); II (DiI+-DPCsp=60); III (DiI+-DPCsp=60/ EGFP+-DSCsp=1); IV (EGFP+-DSCsp=1); and V (FBsp=3). The graft sites were photographed (ae) 3 weeks posttransplantation and analyzed histologically (fw). The results of type I–V transplantations are shown in (a, f, k, p), (b, g, l, q, t), (c, h, m, r, u), (d, i, n, s), and (e, j, o), respectively. Serial paraffin sections were treated with H&E (fj), Hoechst (ko, blue), and anti-EGFP antibodies (s and u, brown, arrows). DiI signals were viewed through a fluorescence microscope (pr). Hoechst and DiI signals were superimposed (t and v). Images in fj were assembled from composite photos. The solid rectangular region marked by k/p in f is selected for the photos marked with k (Hoechst) and p (DiI). Similarly, the regions of l/q and t in g correspond to figures l and q, and t, respectively, and u in h, s in i, and o in j to u, s, and o, respectively. The broken rectangular regions with m/r in u and n in s correspond to m and r, and n, respectively. The fluorescence photo of v was taken from a section of type I (DiI+-DPCsp=6) graft, and the arrows point to host blood capillaries composed of cells with brightly fluorescent nucleoli. The fluorescence photo of w was obtained from a type IV (EGFP+-DSCsp=1) graft section. Induced HFs and their DPs are outlined by broken lines in photos (kr, t, and w) referring to the corresponding H&E and fluorescence photos. DP, dermal papilla; Ec, erythrocyte; HM, hair matrix; and IRS, inner root sheath. Scale bar (ae)=1mm, (fw)=100μm.

      Cell analysis of DP in type V transplants to test HF induction by contaminating rat dermal cells in EPC preparations

      We tested the hypothesis that the small hairs formed by type V (FBsp=3) transplants are derived from contaminating dermal cells in EPC preparations. First, we examined whether these hairs were actually induced by the transplanted cells and not by host cells. Mouse and rat cells were distinguished according to the criteria of
      • Cunha G.R.
      • Vanderslice K.D.
      Identification in histological sections of species origin of cells from mouse, rat and human.
      (CV criteria), which identify mouse cells as those that display small, discrete, strongly fluorescing intranuclear bodies after Hoechst staining. An H&E-stained section from a type V (FBsp=3) transplant is shown in Figure 2j: a few induced HFs are seen in the region indicated by “o”. Hoechst staining of this o-region is shown in Figure 2o, which demonstrates that no cells in the formed follicles contained such mouse cell–specific intranuclear bodies, indicating that these HFs were all of rat origin. Type V (FBsp=3) transplants contained two donor cell types—EPCs and FBsp=3—of which FBs lack the capacity to generate HFs (
      • Horne K.A.
      • Jahoda C.B.
      • Oliver R.F.
      Whisker growth induced by implantation of cultured vibrissa dermal papilla cells in the adult rat.
      ). On the basis of these results, we concluded that the small and sparse hairs generated from the type V (FBsp=3) transplants were induced by the contaminating rat dermal cells. In fact, immunocytochemical analysis of the EPC preparation before transplantation revealed 1.4% of cells to be vimentin+. Serial sectioning of type V (FBsp=3) transplant graft sites showed that the HFs formed by these contaminating cells are of the complete type. The other four transplant types all contained EPCs, indicating that the hairs formed (Figure 2a–e) include some derived from contaminating rat dermal cells. To discriminate these HFs from those induced by transplanted DPCs or DSC-derived DPCs, we term them “contaminating” and “authentic” HFs, respectively.
      Using the CV criteria, we examined the origin of cells present in host skin sections. A section of a type I (DiI+-DPCsp=6) graft was visualized under a fluorescence microscope and the images (Hoechst and DiI) were merged (Figure 2v). The arrows in the figure indicate host blood capillaries composed of cells with brightly fluorescent nucleoli, which are thus of mouse origin. Microscopic analysis of the Hoechst-stained sections of the transplants indicated that the follicular keratinocytes of the newly formed HFs were all of rat origin (Figure 2o). Similarly, those of the other types of transplants were all of rat origin (Figure 2k–n), which indicates that the transplanted EPCs participated in the formation of new HFs in all the examined types of transplants.

      Cell analysis of the DP in type I transplants

      Examination of serial H&E sections showed that almost all the induced HFs in type I (DiI+-DPCsp=6) grafts were complete. Examples of such HFs are shown in the region marked with “k/p” in Figure 2f. The Hoechst and DiI signals in the k/p region are shown in Figure 2k and p, respectively, and indicate that the DPCs were all DiI+ (that is, all the DPCs were derived from transplanted DiI+-DPCsp=6). Thus, we concluded that nearly all the HFs formed in type I (DiI+-DPCsp=6) transplants were authentic.

      Cell analysis of the DP in type II transplants

      Examinations of H&E serial sections of type II (DiI+-DPCsp=60) transplant sites showed that this type of transplant generated incomplete and complete HFs at an approximate ratio of 4:1. An example of an incomplete HF is shown in the region marked with “l/q” in Figure 2g. Hoechst and DiI signals in this l/q region are shown in Figure 2l and q, respectively, and indicate that the DiI+-DP contained both DiI+-DPCs and DiI-DPCs, the latter derived from rat dermal cells contaminating the EPC preparation. We termed such incomplete HFs, induced by DP composed of DiI+-DPCsp=60 and rat dermal cells, “l/q-type incomplete HFs.” Analysis of serial sections showed the ratio of DiI+-DPCsp=60 to rat dermal cells in the l/q-type incomplete HFs to be 78.5±26.3:21.5±26.3% (n=74) (Supplementary Table S2 online). There was the possibility that the observed red fluorescence in the DP in Figure 2q was caused not by DiI but by autofluorescence. This possibility was ruled out by the following observation. A serial section of a type IV (EGFP+-DSCsp=1) graft that did not contain DiI-labeled cells was visualized under a fluorescence microscope (Figure 2w). The findings showed that HF tissues such as the hair matrix and inner root sheaths yielded red signals due to autofluorescence, but the DP did not, indicating that DP was not autofluorescent under a DiI filter.
      As shown in Figure 2b, type II (DiI+-DPCsp=60) transplants yielded complete HFs. Histological examinations revealed that these complete HFs are grouped into two types. One type was the HF that, like l/q-type incomplete HFs, contained both DiI+-DPCsp=60 and DiI rat dermal cells (histological data not shown), but, unlike the l/q-type incomplete HFs, DiI rat dermal cells dominated DiI+-DPCsp=60: the ratio of DiI+-DPCsp=60 to DiI rat dermal cells was 7.0±8.3:93.0±8.3% (n=53). We term this type of complete HF “l/q-type complete HF.” Considering the high percentage of rat dermal cell-derived DPCs, it is most likely that the l/q-type complete HF was induced by contaminating rat dermal cells.
      The other type of complete HF was the HF that contained only the rat dermal cells contaminating the EPC preparations. An example of such HF is shown in the region marked “t” in Figure 2g, which represents the bulbous part of the hair shown in Figure 2b. A Hoechst and DiI–merged image of the DP of this type (“t-type”), presented in Figure 2t, clearly demonstrates that the DP did not contain DiI+ cells; instead, the cells of the DP were all contaminating dermal cells, indicating that this HF was similar to that described in type V (FBsp=3) transplants (Figure 2e). Of the 53 hairs visible in Figure 2b, 25 (47%) and 28 (52%) were l/q-type complete and “t-type complete” HFs, respectively. In the present study, both l/q-type and t-type complete HFs were ignored, because the transplanted DPC-derived cells did not have major roles in their formation or participate in their formation at all.
      From these results, we conclude that type II (DPCsp=60) transplants generated only incomplete, but authentic, HFs, ∼80% of whose DPCs were derived from transplanted DPCsp=60. The possibility remained that ∼20% contaminating rat dermal cell–derived DPCs have some role in induction of this type of HF.

      Cell analysis of the DP in type III transplants

      Serial H&E sections from type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) transplant sites revealed that, as in type I (DiI+-DPCsp=6) sites, almost all the induced HFs were complete and authentic. A representative H&E section is shown in Figure 2h. The EGFP signal in the highlighted region of Figure 2h indicated with a “u” is shown in Figure 2u. The DSs of the induced HFs were EGFP+ (Figure 2u, arrows), demonstrating that hair development was associated with the development of the DS and that the transplanted DSCs contributed to formation of the DS of the newly formed HFs. The highlighted region (“m/r”) of Figure 2u was examined for Hoechst and DiI signals, as shown in Figure 2m and r, respectively. The DP in the induced HFs was also found to contain DiI+ cells (Figure 2r), as was the case with the type I (DiI+-DPCsp=6) transplants (Figure 2p). In addition, comparison of Figure 2m, r, and u revealed the presence of HFs that contained unlabeled DPCs (that is, DiI- and EGFP-DPCs). Examination of nuclear brightness in Hoechst sections to distinguish host and donor cells indicated that the label cells were of rat origin. As with type II (DPCsp=60) transplants, these HFs were excluded from further analysis because their formation was not induced by DiI+-DPCsp=60/EGFP+-DSCsp=1.
      We quantified the contribution of transplanted DPCsp=60 and DSCsp=1 to the DP of the induced HFs in type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) transplants by counting Hoechst-, DiI-, and EGFP-labeled cells in the DP of the HFs in serial sections. The ratio of DiI+ to EGFP+ to label cells in the DP was 46.5±20.0:43.3±21.0:10.2±15.0% (n=120) (Supplementary Table S2 online). From these results, we concluded that cells derived from transplanted DiI+-DPCsp=60 and EGFP+-DSCsp=1 contribute significantly to the DP. This in turn led us to conclude that, like type I (DPCsp=6) grafts, type III (DPCsp=60/DSCsp=1) grafts induced the formation of authentic complete HFs. Similarly, we estimated the contribution of transplanted DSCsp=1 to the bulbous DS of the HFs induced by type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) grafts. Of the cells in the bulbous DS, 49.7±29.6% (n=81) were EGFP+, but, notably, no DiI+ cells were present, indicating that DPCshigh passage do not give rise to DSCs. The remaining DSCs in the induced HFs originated from contaminating dermal cells.

      Cell analysis of the DP in type IV transplants

      Fewer HFs were induced by type IV (EGFP+-DSCsp=1) transplants (Figure 2i). The region labeled “s” in Figure 2i was analyzed for EGFP signals (Figure 2s). The induced HFs were enclosed by EGFP+-DSCs. Hoechst staining of the indicated region (“n”) of Figure 2s is shown in Figure 2n. Comparison of Figure 2n and s showed that EGFP+ cells were also present in the DP (Figure 2s), indicating that transplanted DSCs differentiated into DPCs and strongly suggesting that the transplanted EGFP+ -DSCs formed the DP and induced HFs by interacting with transplanted EPCs. Similar differentiation of DSCs into DPCs occurred with type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) grafts (Figure 2u). We quantified the contribution of transplanted DSCsp=1 to the DP of the induced HFs in type IV (EGFP+-DSCsp=1) transplants by counting EGFP-labeled cells in the DP of the HFs in serial sections. The ratio of EGFP+ to label cells in the DP was 73.3±22.4:26.7±22.4% (n=76) (Supplementary Table S2 online). Considering that the majority of DPCs in the induced HF were derived from the transplanted EGFP+-DSCsp=1, we concluded that the induced HFs were authentic.

      Characterization of induced HFs using specific protein markers

      HFs induced in type I–IV grafts were analyzed for the expression of α-SMA, versican, and Ki67—markers of DSCs, anagen DP, and cellular proliferation, respectively. The HFs induced by type I (DPCsp=6), III (DPCsp=60/DSCsp=1), and IV (DSCsp=1) transplants were encapsulated with α-SMA+-DSCs (Figure 3a, g, and j, respectively), whereas those induced by type II (DPCsp=60) were α-SMA (Figure 3d; see also Figure 1j). The DSCs of the HFs of type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) transplants were both α-SMA+ (Figure 3g) and EGFP+ (Figure 3m), indicating that transplanted DSCs contributed to the formation of the DS of the newly formed hairs. These HFs contained both DiI+- (Figure 3o) and EGFP+-DPCs (Figure 3m; see also Figure 2r and u). As expected, versican was detected in the DP of the HFs of type I (DPCsp=6), III (DPCsp=60/DSCsp=1), and IV (DSCsp=1) transplants (Figure 3b, h, and k, respectively), but not type II (DPCsp=60) transplants (Figure 3e). Bulbous regions and the outer epithelium of the HFs from all types of transplants clearly contained Ki67+ cells (Figure 3c, f, i, and l).
      Figure thumbnail gr4
      Figure 3Characterization of induced hair follicles (HFs). Serial paraffin sections were prepared from the graft sites of type I (DPCsp=6), II (DPCsp=60), III (DPCsp=60/DSCsp=1), and IV (DSCsp=1) shown in for histological examination and are shown in (a, b, c), (d, e, f), (g, h, i, m), and (j, k, l), respectively. The sections were immunostained for α-SMA (a, d, g, j), versican (b, e, h, k), Ki67 (c, f, i, l), and EGFP (m). The broken rectangular region indicated by n/o in m was viewed for Hoechst (n, blue) and DiI signals (o, red). The arrowheads in a, g, and j point to the α-SMA+-DS. The arrows in m point to the EGFP+-DS. DP, dermal papilla; DS, dermal sheath; *, capillary vessel. Scale bar (ao)=100μm.

      Morphometry of the induced HFs in type I–V transplants

      Hair-shaft lengths in the same mice used to generate the data presented in Figure 2 were measured 3 weeks posttransplantation (Supplementary Figure S2a online). The hair shafts in type III (DPCsp=60/DSCsp=1) transplants were significantly longer (5.0- and 2.2-fold, respectively) than those in type II (DPCsp=60) and IV (DSCsp=1) transplants. Type III (DPCsp=60/DSCsp=1) hair shafts tended to be longer than those in type I (DPCsp=6) transplants (although not to a significant extent). The hair shafts in type I (DPCsp=6) graft sites were significantly longer (4.3- and 1.9-fold, respectively) than those in type II (DPCsp=60) and IV (DSCsp=1) transplants. The hair shafts in type IV (DSCsp=1) transplants were significantly longer (2.3-fold) than those in type II (DPCsp=60) transplants.
      The HFs induced by transplanted DPCs and/or DSCs were counted in serial sections of type I–IV transplants by identifying HFs with DPs containing DiI+- or EGFP+-DPCs. In the case of the type II (DiI+-DPCsp=60) transplants, we excluded l/q-type complete HFs. Contaminating dermal cells in the type V (FBsp=3) transplants generated 19.8±1.3 (n=3) HFs per graft (Supplementary Figure S2b online). More HFs tended to be generated from type III (DPCsp=60/DSCsp=1) transplants than from type I (DPCsp=6) transplants (although the difference was not significant). Significantly more HFs (2.8- and 5.6-fold more, respectively) arose from type III (DPCsp=60/DSCsp=1) transplants than from type II (DPCsp=60) and IV (DSCsp=1) transplants. Significantly more HFs (2.4- and 4.8-fold more, respectively) were generated from type I (DPCsp=6) transplants than from type II (DPCsp=60) and IV (DSCsp=1) transplants. Type II (DPCsp=60) and IV (DSCsp=1) transplants did not give rise to significantly different numbers of HFs.

      Expression of trichogenic genes in cocultures of DPCshigh passage and DSCslow passage

      Collectively, the findings that we have described support the idea that the cotransplantation of DSCs restores the hair-inducing ability of DPCshigh passage that had been lost as a result of repeated subculturing and strongly suggest that DPCshigh passage and low-passage DPCs (DSCslow passage) interact cooperatively with rat EPCs. These possible interactions were tested using an in vitro coculture system. DPCsp=60 and DSCsp=1 were cocultured for 5 days such that the two cell types did not come into direct contact with each other. As controls, DPCsp=6, DPCsp=60, and DSCsp=1 were also cultured alone (single cultures). Cells were harvested, and the expression of the genes encoding the following trichogenic genes was determined by reverse transcriptase–PCR: versican (Supplementary Figure S3a online), alkaline phosphatase (ALP) (Supplementary Figure S3b online), bone morphogenetic protein 4 (BMP4) (Supplementary Figure S3c online), and Noggin (Supplementary Figure S3d online). As expected, DSCsp=1 in single culture (DSCsp=1-S) expressed versican and BMP4 mRNAs at lower levels than DPCsp=6 in single culture (DPCsp=6-S) as well as DPCsp=60-S. DSCsp=1-S expressed ALP and Noggin mRNAs at comparable levels with DPCsp=6-S. Both displayed much higher expression than DPCsp=60-S. Furthermore, coculture of DPCsp=60 with DSCsp=1 (DPCsp=60-C) did not restore expression to the levels observed in DPCsp=6-S but instead tended to reduce levels of expression relative to DPCsp=60. Notably, DSCsp=1 cultured with DPCsp=60 (DSCsp=1-C) displayed higher expression of versican and BMP4 mRNAs compared with DSCsp=1-S. Indeed, expression was as high as that observed in DPCsp=6-S. However, coculture did not affect the expression of ALP and Noggin mRNAs in DSCsp=1-S. The results of these experiments suggest that DPCshigh passage and DSCslow passage mutually affected their expression of trichogenicity-related genes and that stimulatory factors secreted by DPCshigh passage influence expression of the genes encoding versican and BMP4 in DSCslow passage.

      Discussion

      This study confirmed that, as reported previously (
      • Horne K.A.
      • Jahoda C.B.
      • Oliver R.F.
      Whisker growth induced by implantation of cultured vibrissa dermal papilla cells in the adult rat.
      ), DPCslow passage are able to induce the formation of complete HFs. However, this ability was largely lost after p=30, although the ability to induce incomplete HFs was retained. We also showed that DPCshigh passage were unable to induce complete HFs but were able to induce incomplete HFs. Cells in the outer epithelium and bulbous regions of HFs induced by DPCshigh passage in type II (DPCsp=60) transplants were clearly Ki67+. However, these follicles did not develop hair shafts. The induced incomplete HFs were α-SMA (which explains our failure to identify the DSCs histologically), and their DPCs were versican. When cotransplanted with DSCs in type III (DPCsp=60/DSCsp=1) transplants, DPCshigh passage regained full HF-inducing ability, comparable to that displayed by DPCslow passage. Cotransplantation-induced restoration of complete HF formation was accompanied by encapsulation of the induced HFs with the α-SMA+-DSCs and by restoration of versican expression in their DPCs. These results suggest that DPCshigh passage and DSCs stimulate each other's activities and highlight a strong correlation between a lack of DS and incomplete HF formation. More investigation of the DS-dependent restoration of DPCshigh passage-induced HF formation will be needed to determine whether the HFs formed meet the requirements for classification as “typical” HFs according to the system proposed by
      • Chuong C.M.
      • Cotsarelis G.
      • Stenn K.
      Defining hair follicles in the age of stem cell bioengineering.
      .
      Interactions between DPCs and hair matrix cells have crucial roles in the formation and growth of HFs (
      • Oliver R.F.
      • Jahoda C.A.B.
      Dermal-epidermal interactions.
      ), in which signaling molecules including Wnt3a (
      • Kishimoto J.
      • Burgeson R.E.
      • Morgan B.A.
      Wnt signaling maintains the hair-inducing activity of the dermal papilla.
      ) and BMP6 (
      • Rendl M.
      • Polak L.
      • Fuchs E.
      BMP signaling in dermal papilla cells is required for their hair follicle-inductive properties.
      ) are considered to act as stimulators of the hair-inducing activity of DPCs, and other BMPs (
      • Blessing M.
      • Nanney L.B.
      • King L.E.
      • et al.
      Transgenic mice as a model to study the role of TGF-13-related molecules in hair follicles.
      ;
      • Kulessa H.
      • Turk G.
      • Hogan B.L.
      Inhibition of Bmp signaling affects growth and differentiation in the anagen hair follicle.
      ) and Noggin (
      • Kulessa H.
      • Turk G.
      • Hogan B.L.
      Inhibition of Bmp signaling affects growth and differentiation in the anagen hair follicle.
      ) as activators of matrix cells. Several studies have linked versican and ALP with the follicle-inducing ability of DPCs (
      • Kishimoto J.
      • Ehama R.
      • Wu L.
      • et al.
      Selective activation of the versican promoter by epithelial-mesenchymal interactions during hair follicle development.
      ;
      • McElwee K.J.
      • Kissling S.
      • Wenzel E.
      • et al.
      Cultured peribulbar dermal sheath cells can induce hair follicle development and contribute to the dermal sheath and dermal papilla.
      ;
      • Osada A.
      • Iwabuchi T.
      • Kishimoto J.
      • et al.
      Long-term culture of mouse vibrissal dermal papilla cells and de novo hair follicle induction.
      ). Indeed, we demonstrated high expression of versican in the DP of the HFs induced by type III (DPCsp=60/DSCsp=1) transplants.
      Considering the potential significance of the findings of the type III (DPCsp=60/DSCsp=1) transplant experiments, we analyzed in detail the individual contributions of DPCsp=60 and DSCsp=1 in this type of transplants to HF induction. The generated HFs were complete and authentic and were associated with the DS. The HFs formed from type III (DPCsp=60/DSCsp=1) transplants contained DPCs of three sources in their DP—DPCsp=60, DSCsp=1, and rat dermal cells—at an approximate ratio of 47:43:10. Thus, we concluded that DPCsp=60 and DSCsp=1 contributed significantly and comparably to DP formation in the induced HFs. Several possible explanations exist for the role of DSCs in HF formation. DSCp=1-derived DPCs may be entirely responsible for the induction of complete HFs; that is, DPCsp=60 are not themselves involved. However, this seems unlikely because type IV (DSCsp=1) transplants, in which most DPCs are derived from DSCsp=1, displayed much lower HF-inducing activity than type I (DPCsp=6) transplants. Alternatively, DSCp=1-derived DPCs or DSCsp=1 themselves may activate DPCsp=60 such that they function similarly to DPCslow passage. However, our in vitro coculture experiments do not support this hypothesis because expression of four trichogenic genes in DPCsp=60was significantly reduced by the presence of DSCsp=1 in the culture. In a third scenario, DPCsp=60 may activate DSCsp=1 to differentiate into active DPCs. This hypothesis appears to be supported, in part, by findings of our coculture experiments because DSCp=1 expression of the genes encoding versican and BMP4 was significantly stimulated by the presence of DPCsp=60. In addition, DSCp=1-derived DPCs and DPCshigh passage may activate each other during the reconstitution of HFs and together generate the DP. Finally, note that this study cannot completely exclude the possibility that the contaminating rat dermal cells contributed to the formation of complete HFs in type III (DPCsp=60/DSCsp=1) transplants.
      Our coculture experiments were performed under conditions in which individual cell types were allowed to proliferate to confluence and then bathed in the same culture medium to allow them to interact with each other by means of diffusible factors. Direct contact between cells of the two different types was prevented. These experiments yielded some noteworthy observations. First, expression of the genes encoding versican and BMP4 in DSCsp=1-S was lower than in DPCsp=6-S, but comparable to that of DPCsp=60. Moreover, coculture with DPCsp=60 increased the expression of these genes in DSCsp=1 to levels comparable to those in DPCsp=6-S. Second, DSCsp=1-S displayed high expression of the genes encoding ALP and Noggin, comparable to that observed in active DPCsp=6 in single culture. These expressions were not affected by the presence of DPCsp=60. Finally, as expected, expression of the marker genes was lower in DPCsp=60 than in DPCsp=6 when each cell type was grown in single culture. Unexpectedly, however, DPCp=60 reduced expression of all the studied trichogenic genes in the presence of DSCsp=1. These results strongly support the hypothesis that the trichogenicity of DSCslow passage is enhanced by the presence of DPCshigh passage, but they do not support the notion that DSCslow passage restore the hair-inducing capacity of DPCshigh passage. The fact that DPCshigh passage in the DP of type III (DPCsp=60/DSCsp=1) transplants did not convert to DSCs also appears to indirectly preclude the possibility of DSClow passage-mediated reactivation of DPCshigh passage, because active DPCs such as DPCslow passage may convert to DSCs. However, the present study cannot completely exclude the possibility that DSCslow passage reactivate DPCshigh passage, because DPCp=60-derived DPCs that, in type II (DPCsp=60) transplants, were versican, were versican+ in type III (DPCsp=60/DSCsp=1) transplants.
      Determining a realistic mechanism to explain the apparent reactivation of DPCshigh passage in type III (DPCsp=60/DSCsp=1) transplants will necessitate further investigation because some uncertainty exists as to whether the in vitro coculture model that we adopted faithfully reproduced the events that occur in the type III (DPCsp=60/DSCsp=1) transplants in vivo. For example, the stimulation of DPCsp=60 in type III (DPCsp=60/DSCsp=1) transplants may require direct contact between the two types of cells; also, the EPCs obtained from newborn rats and contaminating rat dermal cells might be involved in stimulatory interactions in the type III (DPCsp=60/DSCsp=1) transplants. To study the interactions between such cells was not an aim of this study. Future testing of our proposed mechanisms should further our understanding of the functional relationships that exist between DPCshigh passage and DSCslow passage.
      This study clearly demonstrated the interconvertibility of DPCs and DSCs. The DPCs and DSCs used in the present transplantation experiments were microscopically pure (∼100%) and appropriately labeled for monitoring following transplantation. When EGFP+-DPCslow passage alone were transplanted, the DSCs, as well as the DP, in the induced HFs were EGFP+, indicating that DPCs may convert to DSCs, as reported previously (
      • Oliver R.F.
      Histological studies of whisker regeneration in the hooded rat.
      ;
      • Jahoda C.A.B.
      Induction of follicle formation and hair growth by vibrissa dermal papillae implanted into rat ear wounds: vibrissa-type fibers are specified.
      ;
      • Reynolds A.J.
      • Lawrence C.
      • Cserhalmi-Friedman P.B.
      • et al.
      Trans-gender induction of hair follicles.
      ;
      • Horne K.A.
      • Jahoda C.A.B.
      Restoration of hair growth by surgical implantation of follicular dermal sheath.
      ). Our results imply that the DPCshigh passage in type II (DPCsp=60) transplants were unable, on their own, to convert to DSCs because the induced incomplete HFs lacked DS structures. Note that DSCsp=1 themselves were, on their own, able to induce complete HFs, directly indicating that some of the transplanted DSCs differentiated into DPCs. Type I (EGFP+-DPCsp=6) and IV (EGFP+-DSCsp=1) transplants gave rise to hairs containing EGFP+-DSCs and EGFP+-DPCs, respectively. Furthermore, type III (DiI+-DPCsp=60/EGFP+-DSCsp=1) transplants generated hairs containing EGFP+-DPCs. These results convincingly demonstrated the mutual convertibility of DPCs and DSCs. In the present study, we showed that the newly active DPCshigh passage in type III (DPCsp=60/DSCsp=1) transplants do not convert to DSCs, like the DPCslow passage in type I (DPCsp=6) transplants, which can convert to DSCs. A recent study showed that thrombin regulates the convertibility of DPCs to DSCs by activating the phosphoinositide 3-kinase–protein kinase B signaling pathway (
      • Feutz A.C.
      • Barrandon Y.
      • Monard D.
      Control of thrombin signaling through PI3K is a mechanism underlying plasticity between hair follicle dermal sheath and papilla cells.
      ). The inability of DPCshigh passage to convert to DSCs suggests that decreased DP–DS cell plasticity resulting from increased DPC passage number may involve a reduction in phosphoinositide 3-kinase–protein kinase B pathway–dependent thrombin signaling.

      Materials and Methods

      Animals and materials

      Adult (6 to 10 weeks old) and newborn rats of the Fisher 344 (F344) and Wistar varieties and BALB/C nu/nu mice (4 weeks old) were purchased from Charles River (Yokohama, Japan). EGFP-tg rats were produced from fertilized eggs of EGFP-tg rats (
      • Hakamata Y.
      • Tahara K.
      • Uchida H.
      • et al.
      Green fluorescent protein-transgenic rat: a tool for organ transplantation research.
      ) that had been obtained from the Health Science Research Resources Bank of the Japan Health Sciences Foundation (Osaka, Japan). All animal experiments were approved by the ethics and animal welfare committee of PhoenixBio. Chemicals and reagents were obtained as follows: dispase from Sanko Junyaku (Tokyo, Japan); FBS from Invitrogen (Carlsbad, CA); FGF2 from Upstate Biotechnology (Lake Placid, NY); DMEM, Hoechst 33258, and BSA from Sigma (St Louis, MO); and DiI from Wako Pure Chemical Industries (Osaka, Japan).

      Isolation and cultivation of DPCs, DSCs, and FBs

      DPs and DSs were isolated from the anagen vibrissa follicles in the upper lip of F344 or Wistar rats aged 6–10 weeks or EGFP-tg rats and cultured as described previously (
      • Inamatsu M.
      • Matsuzaki T.
      • Iwanari H.
      • et al.
      Establishment of rat dermal papilla cell lines that sustain the potency to induce hair follicles from afollicular skin.
      ;
      • McElwee K.J.
      • Kissling S.
      • Wenzel E.
      • et al.
      Cultured peribulbar dermal sheath cells can induce hair follicle development and contribute to the dermal sheath and dermal papilla.
      ). A total of 7–10 explants were placed in a 35-mm dish and cultured in 2ml of DMEM/FBS/FGF2. Once fully confluent, the DPCs and DSCs were detached from the plastic surface and subjected to the first subcultivation at a density of 3.4 × 103 cells per cm2 and 6.0 × 103 cells per cm2, respectively. These DPCs and DSCs were serially passaged weekly and every 5 days at 1:10 and 1:7.5 splits, respectively. FBs were obtained as the cells that grew out from the sole skin explants prepared from 10-week-old F344 or Wistar rats. The FBs were serially (p=3–6) subcultured in DMEM/FBS until used for the graft chamber assay. DPCsp=60 and DSCsp=1 were cocultured for 5 days, and then isolated for determining versican, ALP, BMP4, and Noggin mRNAs expression by RT-PCR as described in Supplementary Methods and Supplementary Table S3 online.

      Graft chamber assay

      Skins from newborn (1- to 2-day-old) F344 or Wistar rats were floated in dispase solution (1,000Uml−1) overnight at 4°C and separated into the dermal and epidermal fractions. To remove contaminating dermal cells, the epidermal layers were incubated in the same dispase solution for 5min at 37°C with gentle shaking, minced, and then incubated in 0.25% trypsin–1mM EDTA solution for 10min at 37°C with gentle shaking. The lysates were then filtered successively through 100- and 40-μm nylon filters. The resulting single cells were used as EPCs and were checked for the contamination of dermal cells by staining with anti-vimentin Abs (Santa Cruz Biotechnology, Santa Cruz, CA).
      The cell mixture for transplantation in the experiments depicted in Figure 1 was composed of three cell types: EPCs as a responder cell type for hair induction, DPCs as an inducer cell type, and FBs as a filler cell type to compensate the cell number of mesenchymal cells in relation to the number of EPCs. EGFP+-DPCsp≤6 were obtained from EGFP-tg rats. DPCsp>9 were labeled with DiI (
      • Inamatsu M.
      • Tochio T.
      • Makabe A.
      • et al.
      Embryonic dermal condensation and adult dermal papilla induce hair follicles in adult glabrous epidermis through different mechanisms.
      ). These cells were individually suspended, mixed, and centrifuged. The pellets of EPCs (1 × 107), DPCs (8 × 106), and FBs (2 × 106) were suspended in 50μl of carryover medium for implanting and transplanted onto the back of 4-week-old male nude mice for the graft chamber assay (
      • Lichti U.
      • Weinberg W.C.
      • Goodman L.
      • et al.
      In vivo regulation of murine hair growth: Insights from grafting defined cell populations on nude mice.
      ;
      • Weinberg W.C.
      • Goodman L.V.
      • George C.
      • et al.
      Reconstitution of hair follicle development in vivo: determination of follicle formation, hair growth, and hair quality by dermal cells.
      ).
      The ratio of DPCs to FBs and the total cell number per transplantation in the basic cell mixture were reducible to 0.4 and 20%, respectively, of the original preparation without affecting the hair-inducing capacity of the transplants. The cell numbers in the basic cell mixtures in experiments shown in Figure 2 were as follows: 2 × 106 EPCs, 6 × 105 DPCs, and 1.4 × 106 FBs. In experiments shown in Figure 2, five types (I–V) of cell mixtures were prepared for transplantation; the actual cell compositions are described in Supplementary Table S1 online. Note that the DPCsp=6 in type I (DPCsp=6) were labeled not with EGFP, but with DiI, because this transplantation was the positive control of a series of transplantation experiments to compare the inductive ability of DPCs cultured for various passage numbers in which DPCshigh passage were labeled with DiI. DPCs at p=31, 39, and 60 were used for transplantation in the experiments depicted in Figures 1 and 2. We established a DPC line from cultures of 116 pieces of DPs isolated from four rats. The DPCshigh passage utilized in this study were all from this line. Each cell type was suspended and mixed to prepare the transplants with the above compositions. These cell mixtures were transplanted onto the back of nude mice for graft chamber assay. The animals were killed 3 weeks posttransplantation. The transplanted sites were directly photographed using a Polaroid Macro 5 SLR camera (Polaroid, Minnetonka, MN) (Figure 1a–d) or an Olympus DP20 digital camera (Olympus, Tokyo, Japan) through a stereomicroscope (Figure 2a–e). Identical transplantation experiments using these five types of transplants were separately conducted three times (three series experiments): one series included three or four mice for each type of transplant. Some of the hosts failed to accept the transplants. Thus, in this study, altogether, each type of transplant was repeatedly (6 to 10 times) assayed for HF development. Similar results were obtained from these repeated assays for each of five types of transplants.

      Histology and immunohistochemistry

      Transplants were processed for histological examinations for H&E and Hoechst staining, and, when necessary, for detecting EGFP signals as previously reported (
      • Inamatsu M.
      • Tochio T.
      • Makabe A.
      • et al.
      Embryonic dermal condensation and adult dermal papilla induce hair follicles in adult glabrous epidermis through different mechanisms.
      ). DS in the induced HFs on the histological sections was identified with monoclonal anti-α-SMA Abs (Sigma) (
      • Jahoda C.A.B.
      • Reynolds A.J.
      • Chaponnier C.
      • et al.
      Smooth muscle α-actin is a marker for hair follicle dermis in vivo and in vitro.
      ) following standard procedures. When necessary, EGFP+ cells in the transplants were detected with polyclonal anti-EGFP Abs (Clontech Laboratories, Mountain View, CA). Some tissue sections were stained with anti-versican Abs (Millipore, Billerica, MA) and monoclonal anti-Ki67 Abs (Dako Cytomation Denmark A/S, Glostrup, Denmark). The bound Abs were visualized with an ABC system using diaminobenzidine as substrate. Sections were counterstained with hematoxylin.

      Quantification of donor-derived DPCs and DSCs in HFs in transplants

      Serial sagittal paraffin sections, 5μm thick, were prepared from each of three graft sites of type II (DiI+-DPCsp=60) and III (DiI+-DPCsp=60/EGFP+-DSCsp=1) transplants and stained with Hoechst or anti-EGFP Abs. The DiI+- and EGFP+-DPCs in the DP and the EGFP+-DSCs in the bulbous DS were counted on the sections that corresponded to the central region of hair bulbs by detecting DiI+ and EGFP+ cells through a fluorescence microscope and immunohistochemically as described above.

      Statistical analysis

      Analyses were performed using the t-test, Steel–Dwass test, or Tukey's test, as indicated. P-values <0.05 were considered statistically significant.

      ACKNOWLEDGMENTS

      Y Yonehara and S Furukawa assisted us in this study. This work was supported in part by the Yoshizato-Towa-Tokumen Project in Innovation Plaza Hiroshima of the Japan Science and Technology Agency.

      SUPPLEMENTARY MATERIAL

      Supplementary material is linked to the online version of the paper at http://www.nature.com/jid

      REFERENCES

        • Blessing M.
        • Nanney L.B.
        • King L.E.
        • et al.
        Transgenic mice as a model to study the role of TGF-13-related molecules in hair follicles.
        Genes Dev. 1993; 7: 204-215
        • Botchkarev N.A.
        • Ahluwalia G.
        • Shander D.
        Apoptosis in the hair follicle.
        J Invest Dermatol. 2006; 126: 258-264
        • Chuong C.M.
        • Cotsarelis G.
        • Stenn K.
        Defining hair follicles in the age of stem cell bioengineering.
        J Invest Dermatol. 2007; 127: 2098-2100
        • Cunha G.R.
        • Vanderslice K.D.
        Identification in histological sections of species origin of cells from mouse, rat and human.
        Stain Technol. 1984; 59: 7-12
        • Feutz A.C.
        • Barrandon Y.
        • Monard D.
        Control of thrombin signaling through PI3K is a mechanism underlying plasticity between hair follicle dermal sheath and papilla cells.
        J Cell Sci. 2008; 121: 1435-1443
        • Hakamata Y.
        • Tahara K.
        • Uchida H.
        • et al.
        Green fluorescent protein-transgenic rat: a tool for organ transplantation research.
        Biochem Biophys Res Commun. 2001; 286: 779-785
        • Hardy M.H.
        The secret life of the hair follicle.
        Trends Genet. 1992; 8: 55-61
        • Horne K.A.
        • Jahoda C.A.B.
        Restoration of hair growth by surgical implantation of follicular dermal sheath.
        Development. 1992; 116: 563-571
        • Horne K.A.
        • Jahoda C.B.
        • Oliver R.F.
        Whisker growth induced by implantation of cultured vibrissa dermal papilla cells in the adult rat.
        J Embryol Exp Morph. 1986; 97: 111-124
        • Inamatsu M.
        • Matsuzaki T.
        • Iwanari H.
        • et al.
        Establishment of rat dermal papilla cell lines that sustain the potency to induce hair follicles from afollicular skin.
        J Invest Dermatol. 1998; 111: 767-775
        • Inamatsu M.
        • Tochio T.
        • Makabe A.
        • et al.
        Embryonic dermal condensation and adult dermal papilla induce hair follicles in adult glabrous epidermis through different mechanisms.
        Develop Growth Differ. 2006; 48: 73-86
        • Inui S.
        • Fukuzato Y.
        • Nakajima T.
        • et al.
        Androgen receptor co-activator Hic-5/ARA55 as a molecular regulator of androgen sensitivity in dermal papilla cells of human hair follicles.
        J Invest Dermatol. 2007; 127: 2302-2306
        • Jahoda C.A.B.
        Induction of follicle formation and hair growth by vibrissa dermal papillae implanted into rat ear wounds: vibrissa-type fibers are specified.
        Development. 1992; 115: 1103-1109
        • Jahoda C.A.B.
        • Horne K.A.
        • Oliver R.F.
        Induction of hair growth by implantation of cultured dermal papilla cells.
        Nature. 1984; 311: 560-562
        • Jahoda C.A.B.
        • Reynolds A.J.
        • Chaponnier C.
        • et al.
        Smooth muscle α-actin is a marker for hair follicle dermis in vivo and in vitro.
        J Cell Sci. 1991; 99: 627-636
        • Kishimoto J.
        • Burgeson R.E.
        • Morgan B.A.
        Wnt signaling maintains the hair-inducing activity of the dermal papilla.
        Genes Dev. 2000; 14: 1181-1185
        • Kishimoto J.
        • Ehama R.
        • Wu L.
        • et al.
        Selective activation of the versican promoter by epithelial-mesenchymal interactions during hair follicle development.
        Proc Natl Acad Sci USA. 1999; 96: 7336-7341
        • Kulessa H.
        • Turk G.
        • Hogan B.L.
        Inhibition of Bmp signaling affects growth and differentiation in the anagen hair follicle.
        EMBO J. 2000; 19: 6664-6674
        • Lichti U.
        • Weinberg W.C.
        • Goodman L.
        • et al.
        In vivo regulation of murine hair growth: Insights from grafting defined cell populations on nude mice.
        J Invest Dermatol. 1993; 101: 124S-129S
        • McElwee K.J.
        • Kissling S.
        • Wenzel E.
        • et al.
        Cultured peribulbar dermal sheath cells can induce hair follicle development and contribute to the dermal sheath and dermal papilla.
        J Invest Dermatol. 2003; 121: 1267-1275
        • Millar S.E.
        Molecular mechanisms regulating hair follicle development.
        J Invest Dermatol. 2002; 118: 216-225
        • Oliver R.F.
        Histological studies of whisker regeneration in the hooded rat.
        J Embryol Exp Morph. 1966; 16: 231-244
        • Oliver R.F.
        • Jahoda C.A.B.
        Dermal-epidermal interactions.
        Clin Dermatol. 1988; 6: 74-82
        • Osada A.
        • Iwabuchi T.
        • Kishimoto J.
        • et al.
        Long-term culture of mouse vibrissal dermal papilla cells and de novo hair follicle induction.
        Tissue Eng. 2007; 13: 975-982
        • Rendl M.
        • Lewis L.
        • Fuchs E.
        Molecular dissection of mesenchymal-epithelial interactions in the hair follicle.
        PLoS Biol. 2005; 3: 1910-1924
        • Rendl M.
        • Polak L.
        • Fuchs E.
        BMP signaling in dermal papilla cells is required for their hair follicle-inductive properties.
        Genes Dev. 2008; 15: 543-557
        • Reynolds A.J.
        • Lawrence C.
        • Cserhalmi-Friedman P.B.
        • et al.
        Trans-gender induction of hair follicles.
        Nature. 1999; 402: 33-34
        • Weinberg W.C.
        • Goodman L.V.
        • George C.
        • et al.
        Reconstitution of hair follicle development in vivo: determination of follicle formation, hair growth, and hair quality by dermal cells.
        J Invest Dermatol. 1993; 100: 229-236

      Linked Article