Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Tracing the origin of hair follicle stem cells

Abstract

Tissue stem cells are generated from a population of embryonic progenitors through organ-specific morphogenetic events1,2. Although tissue stem cells are central to organ homeostasis and regeneration, it remains unclear how they are induced during development, mainly because of the lack of markers that exclusively label prospective stem cells. Here we combine marker-independent long-term 3D live imaging and single-cell transcriptomics to capture a dynamic lineage progression and transcriptome changes in the entire epithelium of the mouse hair follicle as it develops. We found that the precursors of different epithelial lineages were aligned in a 2D concentric manner in the basal layer of the hair placode. Each concentric ring acquired unique transcriptomes and extended to form longitudinally aligned, 3D cylindrical compartments. Prospective bulge stem cells were derived from the peripheral ring of the placode basal layer, but not from suprabasal cells (as was previously suggested3). The fate of placode cells is determined by the cell position, rather than by the orientation of cell division. We also identified 13 gene clusters: the ensemble expression dynamics of these clusters drew the entire transcriptional landscape of epithelial lineage diversification, consistent with cell lineage data. Combining these findings with previous work on the development of appendages in insects4,5, we describe the ‘telescope model’, a generalized model for the development of ectodermal organs in which 2D concentric zones in the placode telescope out to form 3D longitudinally aligned cylindrical compartments.

This is a preview of subscription content

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Bulge stem cells are derived from the periphery of the placode basal layer.
Fig. 2: Cell fate determination relies on cell position.
Fig. 3: Two-dimensional concentric transcriptional landscape characterizes stem cell origin.
Fig. 4: Transcriptional landscape that underlies coordinated diversification of epithelial lineages and stem cell induction in the developing hair follicle.

Data availability

Live imaging data in this study have been deposited in the SSBD repository at https://doi.org/10.24631/ssbd.repos.2020.06.002 and http://ssbd.qbic.riken.jp/set/20200602/. The scRNA-seq data in this study have been deposited in the Gene Expression Omnibus under accession code GSE147372. The mouse genome (mm10) used in this study is available at https://genome.ucsc.edu/. The web-based tool Enrichr is available at https://maayanlab.cloud/Enrichr/. The KEGG PATHWAY database is available at https://www.genome.jp/kegg/. Any other relevant data are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Code availability

Scripts used for scRNA-seq analysis are available at https://github.com/FujiwaraLab/Morita_et_al_2021 and source code for analysis of the cell division orientation in live imaging data is available at https://github.com/RIKEN-PHB/Morita-Paper-Spindle-Analysis.

References

  1. 1.

    Xin, T., Greco, V. & Myung, P. Hardwiring stem cell communication through tissue structure. Cell 164, 1212–1225 (2016).

    CAS  Article  Google Scholar 

  2. 2.

    Guiu, J. et al. Tracing the origin of adult intestinal stem cells. Nature 570, 107–111 (2019).

    ADS  CAS  Article  Google Scholar 

  3. 3.

    Ouspenskaia, T., Matos, I., Mertz, A. F., Fiore, V. F. & Fuchs, E. WNT–SHH antagonism specifies and expands stem cells prior to niche formation. Cell 164, 156–169 (2016).

    Article  Google Scholar 

  4. 4.

    Lecuit, T. & Cohen, S. M. Proximal–distal axis formation in the Drosophila leg. Nature 388, 139–145 (1997).

    Article  Google Scholar 

  5. 5.

    Ruiz-Losada, M., Blom-Dahl, D., Córdoba, S. & Estella, C. Specification and patterning of Drosophila appendages. J. Dev. Biol. 6, E17 (2018).

    Article  Google Scholar 

  6. 6.

    Pispa, J. & Thesleff, I. Mechanisms of ectodermal organogenesis. Dev. Biol. 262, 195–205 (2003).

    CAS  Article  Google Scholar 

  7. 7.

    Fujiwara, H., Tsutsui, K. & Morita, R. Multi-tasking epidermal stem cells: beyond epidermal maintenance. Dev. Growth Differ. 60, 531–541 (2018).

    Article  Google Scholar 

  8. 8.

    Solanas, G. & Benitah, S. A. Regenerating the skin: a task for the heterogeneous stem cell pool and surrounding niche. Nat. Rev. Mol. Cell Biol. 14, 737–748 (2013).

    CAS  Article  Google Scholar 

  9. 9.

    Nowak, J. A., Polak, L., Pasolli, H. A. & Fuchs, E. Hair follicle stem cells are specified and function in early skin morphogenesis. Cell Stem Cell 3, 33–43 (2008).

    Article  Google Scholar 

  10. 10.

    Liu, Y., Lyle, S., Yang, Z. & Cotsarelis, G. Keratin 15 promoter targets putative epithelial stem cells in the hair follicle bulge. J. Invest. Dermatol. 121, 963–968 (2003).

    CAS  Article  Google Scholar 

  11. 11.

    Rhee, H., Polak, L. & Fuchs, E. Lhx2 maintains stem cell character in hair follicles. Science 312, 1946–1949 (2006).

    ADS  CAS  Article  Google Scholar 

  12. 12.

    Horsley, V., Aliprantis, A. O., Polak, L., Glimcher, L. H. & Fuchs, E. NFATc1 balances quiescence and proliferation of skin stem cells. Cell 132, 299–310 (2008).

    CAS  Article  Google Scholar 

  13. 13.

    Xu, Z. et al. Embryonic attenuated Wnt/β-catenin signaling defines niche location and long-term stem cell fate in hair follicle. eLife 4, e10567 (2015).

    Article  Google Scholar 

  14. 14.

    Morita, R. et al. Coordination of cellular dynamics contributes to tooth epithelium deformations. PLoS ONE 11, e0161336 (2016).

    Article  Google Scholar 

  15. 15.

    Tsutsui, H., Karasawa, S., Shimizu, H., Nukina, N. & Miyawaki, A. Semi-rational engineering of a coral fluorescent protein into an efficient highlighter. EMBO Rep. 6, 233–238 (2005).

    CAS  Article  Google Scholar 

  16. 16.

    Hayashi, T. et al. Single-cell full-length total RNA sequencing uncovers dynamics of recursive splicing and enhancer RNAs. Nat. Commun. 9, 619 (2018).

    ADS  Article  Google Scholar 

  17. 17.

    Qiu, X. et al. Single-cell mRNA quantification and differential analysis with Census. Nat. Methods 14, 309–315 (2017).

    CAS  Article  Google Scholar 

  18. 18.

    Qiu, X. et al. Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods 14, 979–982 (2017).

    CAS  Article  Google Scholar 

  19. 19.

    Huelsken, J., Vogel, R., Erdmann, B., Cotsarelis, G. & Birchmeier, W. β-Catenin controls hair follicle morphogenesis and stem cell differentiation in the skin. Cell 105, 533–545 (2001).

    CAS  Article  Google Scholar 

  20. 20.

    Närhi, K. et al. Sostdc1 defines the size and number of skin appendage placodes. Dev. Biol. 364, 149–161 (2012).

    Article  Google Scholar 

  21. 21.

    Kandyba, E. et al. Competitive balance of intrabulge BMP/Wnt signaling reveals a robust gene network ruling stem cell homeostasis and cyclic activation. Proc. Natl Acad. Sci. USA 110, 1351–1356 (2013).

    ADS  CAS  Article  Google Scholar 

  22. 22.

    Genander, M. et al. BMP signaling and its pSMAD1/5 target genes differentially regulate hair follicle stem cell lineages. Cell Stem Cell 15, 619–633 (2014).

    CAS  Article  Google Scholar 

  23. 23.

    Joost, S. et al. Single-cell transcriptomics reveals that differentiation and spatial signatures shape epidermal and hair follicle heterogeneity. Cell Syst. 3, 221–237 (2016).

    CAS  Article  Google Scholar 

  24. 24.

    Kandyba, E., Hazen, V. M., Kobielak, A., Butler, S. J. & Kobielak, K. Smad1 and 5 but not Smad8 establish stem cell quiescence which is critical to transform the premature hair follicle during morphogenesis toward the postnatal state. Stem Cells 32, 534–547 (2014).

    CAS  Article  Google Scholar 

  25. 25.

    Saxena, N., Mok, K. W. & Rendl, M. An updated classification of hair follicle morphogenesis. Exp. Dermatol. 28, 332–344 (2019).

    Article  Google Scholar 

  26. 26.

    Maini, P. K., Baker, R. E. & Chuong, C. M. The Turing model comes of molecular age. Science 314, 1397–1398 (2006).

    CAS  Article  Google Scholar 

  27. 27.

    Ahtiainen, L. et al. Directional cell migration, but not proliferation, drives hair placode morphogenesis. Dev. Cell 28, 588–602 (2014).

    CAS  Article  Google Scholar 

  28. 28.

    Glover, J. D. et al. Hierarchical patterning modes orchestrate hair follicle morphogenesis. PLoS Biol. 15, e2002117 (2017).

    Article  Google Scholar 

  29. 29.

    Abe, T. et al. Establishment of conditional reporter mouse lines at ROSA26 locus for live cell imaging. Genesis 49, 579–590 (2011).

    CAS  Article  Google Scholar 

  30. 30.

    Vassar, R., Rosenberg, M., Ross, S., Tyner, A. & Fuchs, E. Tissue-specific and differentiation-specific expression of a human K14 keratin gene in transgenic mice. Proc. Natl Acad. Sci. USA 86, 1563–1567 (1989).

    ADS  CAS  Article  Google Scholar 

  31. 31.

    Sakaue-Sawano, A. et al. Visualizing spatiotemporal dynamics of multicellular cell-cycle progression. Cell 132, 487–498 (2008).

    CAS  Article  Google Scholar 

  32. 32.

    Biggs, L. C. et al. Hair follicle dermal condensation forms via Fgf20 primed cell cycle exit, cell motility, and aggregation. eLife 7, e36468 (2018).

    Article  Google Scholar 

  33. 33.

    Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902 (2019).

    CAS  Article  Google Scholar 

  34. 34.

    La Manno, G. et al. RNA velocity of single cells. Nature 560, 494–498 (2018).

    ADS  Article  Google Scholar 

  35. 35.

    Angerer, P. et al. destiny: diffusion maps for large-scale single-cell data in R. Bioinformatics 32, 1241–1243 (2016).

    CAS  Article  Google Scholar 

  36. 36.

    Street, K. et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics 19, 477 (2018).

    Article  Google Scholar 

  37. 37.

    Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 16, 284–287 (2012).

    CAS  Article  Google Scholar 

  38. 38.

    Nakao, K. et al. The development of a bioengineered organ germ method. Nat. Methods 4, 227–230 (2007).

    CAS  Article  Google Scholar 

Download references

Acknowledgements

We thank S. Kuraku and O. Nishimura for assistance with RNA sequencing and data processing; the RIKEN Genome Network Analysis Support Facility (GeNAS) for RNA sequencing; RIKEN Kobe Light Microscopy Facility, Laboratory for Animal Resources and Genetic Engineering and Unit for Four-Dimensional Tissue Analysis for technical assistance; F. Matsuzaki for critical reading of the manuscript; T. Matsuzaki for an important suggestion related to the telescope model; S. Onami for discussions; M. Morimoto for providing a mouse line; A. Iijima for help with cell tracking analysis; and A. Matsushima and M. Ishii for their assistance with the infrastructure for the data analysis. This work was supported by a RIKEN intramural grant, the RIKEN Single Cell Project, the Platform Project for Supporting in Drug Discovery and Life Science Research from MEXT and AMED and the JST CREST program (JPMJCR1926) to H.F. and JSPS Grant-in-Aid for Young Scientists (B) (15K19709 and 17K16361), JSPS Grant-in-Aid for Scientific Research (C) (19K08763), the RIKEN BDR-Otsuka Pharmaceutical Collaboration Center (RBOC) founding program and a Shiseido Female Researcher Science Grant to R.M. This work was partially supported by JST CREST grant number JPMJCR16G3 to I.N.

Author information

Affiliations

Authors

Contributions

R.M. and H.F. conceived the project, designed experiments and wrote the manuscript. R.M. performed most of the experiments and bioinformatics and analysed the data. N.S. and H.S. assisted with cell tracking analysis and histological analysis. T.H. and M.U. performed the library preparation for the RamDA-seq. M.Y. and R.M. performed the bioinformatics analyses. I.N. supervised scRNA-seq experiments and bioinformatics analyses. T.Y. and T.S. provided the custom Python program to analyse the orientation of cell division. T.A., H.K. and Y.F. generated R26-CAG-nKikGR, R26R-CAG-nKikGR and K14-H2B-eGFP mice, and assisted with mouse experiments. H.F. supervised the project.

Corresponding author

Correspondence to Hironobu Fujiwara.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Maria Kasper, Michael Rendl and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended Data Fig. 1 Origin and early lineage of bulge stem cells cannot be traced by expression of known adult stem cell markers.

a–g, Immunohistochemistry of developing whisker hair follicles (HFs) for known stem cell (SC) markers at E12.0 (a, early hair placode), E12.5 (b, late hair placode), E13.0 (c, hair germ), E14.0 (d), E15.0 (e), E17.0 (f) and post-natal day (P)5 (g). LHX2 was expressed in the basal cells of the placode, while K15 and SOX9 were detected in suprabasal and peripheral basal cells of the placode (a). At the late placode stage (E12.5), SOX9 was strongly expressed in the flat suprabasal cells of the placode (arrows in b) and also in some basal cells located near the placode periphery (arrowheads in b). NFATc1 was undetectable throughout these stages (a, b). Thus, at the onset of HF morphogenesis, the expression patterns of adult SC markers vary and there is no clear region in which all SC markers overlap. From the hair germ stage (E13.0) onward, NFATc1 appeared in whisker HFs, and its expression was always restricted to the upper region of the HF (cg). NFATc1-positive epithelial cells formed a compartment with pseudo-stratified (see h) and bulge-like morphology (dg) with gradual loss of the expression of mitotic marker Ki67 during follicle development (ag), which are typical characteristics of adult bulge HFSCs. SOX9 and LHX2 expression also became restricted to the middle to upper part of the HF after E14.0 (dg). The expression patterns of SC markers in E17.0 follicles were almost equivalent to those seen in the upper half of P5 mature follicles, in which a large bulge-like epithelial structure was evident (brackets in f, g). Taken together, these results indicate that basal epithelial cells acquire compartmentalized SC marker expression patterns of the mature bulge by E17.0 in whisker HFs. The origin and early lineage of bulge SCs cannot be traced by following the expression of known adult SC markers because of their wide variation in expression patterns of the hair placode, hair germ and bulbous peg stages. Thus, at present, there is no marker that can exclusively label the origin and early lineage of prospective SCs from the placode stage. Scale bars, 100 μm. h, Three-dimensional-reconstructed z-stack images from whisker HFs derived from KRT5-cre;R26R-Lyn-Venus;R26-H2B-mCherry at E14.0. Cell membranes were sparsely labelled by Venus. Arrowheads indicate the pseudo-stratified epithelium. Scale bar, 50 μm. i, Summary of known SC marker expressions in the developing whisker HF. jn, Immunohistochemistry of developing dorsal HFs for known SC markers at E14.5 (j, hair placode), E15.5 (k, hair germ), E16.5 (l), E17.5 (m) and E19.5 (n). SOX9-positive cells were localized in the suprabasal layer (arrows in j) and in the basal layer near the placode periphery (arrowheads in j) in the dorsal HF placode as observed in the whisker HF placode. Morphogenetic events and marker expression patterns closely resemble those of whisker HFs. Scale bars, 50 μm.

Extended Data Fig. 2 Ex vivo culture system of developing whisker HFs.

a, Strategy to identify the origin and lineage dynamics of HF epithelial cells. b, Photographs of ex vivo cultured whisker pad derived from E11.5 K14-rtTA;TetO-H2B-eGFP;Fucci-G1 mice on day 0 and day 7. Scale bars, 100 μm. c, Immunostaining of whisker HFs derived from E11.5 embryos on day 6 of ex vivo culture. Antibodies detected upper (NFATc1+SOX9+LHX2+) and lower (KRT15+SOX9+LHX2+) stem cell compartments, proliferative cells (Ki67+), sebaceous glands (SCD1+), hair matrix (CDH3+), dermal sheaths (SMA+) and dermal papilla cells (SOX2+). Scale bars, 50 μm. d, Immunostaining of day-6 ex vivo cultured whisker HFs derived from E11.5 embryos for hair cell layer markers shown in the left panel. Ex vivo developing whisker HFs had unique cellular layers characteristic of HFs, except for the medulla, which is formed in mature HFs. HS, hair shaft; IRS, inner root sheath; ORS, outer root sheath. Scale bars, 50 μm.

Extended Data Fig. 3 Identification of the origin and lineage dynamics of HF epithelial cells by long-term live imaging.

a, Schema of epithelial cell subpopulations at the hair germ stage, which we defined for cell tracking. Magenta, IFE basal cells; red, basal cells located in the upper half of the HF; yellow, basal cells located in the lower half of the HF; blue, cells adjacent to the dermal papilla; green, suprabasal cells in the IFE or HF. b, Lineage tree reconstructed from tracking of a hair germ cell and its progeny. The x axis shows the duration of imaging. Lineage is colour-coded based on cell fate, and cell fate was identified based on the cell position in tissue. Scale bars, 50 μm. c, Examples of lineage trees of tracked upper, lower and hair germ cells in Fig. 1a. d, e, HF development is accomplished by enlarging each earlier compartment longitudinally aligned in the follicle epithelium. Data of replicate whisker HFs related to Fig. 1a, b are shown. d, Snapshot images (top panels) and lineage tracking data (bottom panels) of long-term continuous imaging of whisker HF development from the hair germ to bulbous peg stage. Different epithelial lineages were longitudinally aligned as 3D cylindrical compartments in HFs, and prospective bulge SCs were located in the upper part of the HF as shown in Fig. 1a (replicate no. 1, represented in the main figure). Scale bars, 100 μm. e, Cell fates of epithelial cells in the IFE, upper, lower, hair germ and inner regions of the hair germ stage. Cell fates at the bulbous peg stage are shown. f, Bar plot converted from stacked bar plot of Fig. 1b. Statistical analysis was performed by one-way ANOVA followed by Tukey’s test. gi, Different epithelial lineages are aligned in a concentric manner in the placode. Replicate whisker HFs related to Fig. 1c–e are shown in gi. g, Snapshot images (top panels) and lineage tracking data (bottom panels) of long-term continuous imaging of whisker HF development from the placode stage to hair germ stage. Origin of prospective bulge SCs (red) was located at the periphery of the hair placode as shown in Fig. 1c (replicate no. 1, represented in the main figure). Scale bars, 50 μm. h, Bee swarm plot showing the distances of cells from the placode centre. Different epithelial lineages were aligned in a concentric manner in the placode. Statistical analysis was performed by two-sided unpaired t-test. i, Fate of basal and suprabasal cells of the hair placode stage. Cell fates at the hair germ stage are shown. j, Summary of the developmental origins and lineage dynamics of HF epithelial cells. The embryonic stage at which imaging started is indicated in each figure. Each value in the graph is the mean ± s.d. from three independent experiments, one HF each. Numbers of analysed cell lineages are summarized in Supplementary Table 1. See also ‘Statistical analysis and reproducibility’ in Methods and Source Data.

Extended Data Fig. 4 Measurement of cell division angles in the developing epithelium.

In 3D live imaging data, morphology of the developing epithelium was changing constantly. Accordingly, the basement membrane zone was also bending, but not flat. Therefore, we calculated the cell division orientation relative to the basement membrane zone in the developing epithelium three-dimensionally as shown here. a, We first marked the dividing cells in the placode and surrounding IFE in live imaging data, based on chromosome condensation. Examples of a perpendicular (left panels) and horizontal division (middle panels) relative to the basement membrane zone, which were pseudo-coloured in blue; the mitotic spindle axis is indicated by a cyan line connecting the daughter-cell nuclei. The top panels are a planar view images of the epithelium, and the bottom panels are sagittal view images of the epithelium. The placode region in the epithelium was distinguished by accumulation of Fucci-G1 signals accompanied by condensation and cell cycle arrest of the underlying dermal cells, as shown by the dashed circle in the right panel. Scale bars, 20 μm. b, To extract positional information of the basement membrane zone, we next obtained the surface of the developing epithelium based on expression of K14-H2B-eGFP using the surface rendering function in Imaris. A surface object in Imaris was constructed as a mesh object consisting of triangles and vertex normal vectors. Then, based on the direction of the normal vectors, we cut the surface of the opposite side of the basement membrane zone (outer surface of the epithelium) and substituted the remaining dermis-side surface for the basement membrane zone. c, We calculated the 3D orientation of cell division to the basement membrane zone using xyz axis coordinates of the mitotic spindle axis and basement membrane zone. In brief, we first found the closest vertex and its associated normal unit vector on the dermis-side surface by calculating the distance between the centre of a mitotic spindle and each vertex. The cell division angle to the basement membrane zone was then calculated from an inner product of the closest vertex normal unit vector and the unit vector corresponding to a mitotic spindle.

Extended Data Fig. 5 Placode cell fate is determined by cell position but not cell division orientation.

a, b, Bar plots were converted from the stacked bar plot in Fig. 2a, b, respectively. Statistical analysis was performed by one-way ANOVA followed by Tukey’s test. c, Upper and lower daughter cells after perpendicular division in late placode basal layer are marked by LHX2, while the SOX9+ suprabasal nuclei are not labelled with LHX2. This suggests that upper daughter cells hold transcriptional similarity to the basal layer and remain in the pseudo-stratified basal layer. E14.5 dorsal HFs in embryonic skin tissue (not explants) were immunostained for SOX9 and LHX2. Cell division was identified by chromosome condensation. Dashed box shows magnified region (right panels). BM, basement membrane; DP, dermal papilla; U, upper daughter cell; L, lower daughter cell. Scale bars, 50 μm. d, e, Localization of SOX9+ cells in HF placodes in vivo. E14.5 dorsal HFs (d) and E12.0 whisker HFs (e) in embryonic skin tissues (not explants) were immunostained for SOX9 and CDH3. Hair placodes were detected with CDH3 expression or Fucci-G1 probe fluorescence signals. SOX9 expressions were detected not only in flat suprabasal cells (cyan arrowheads) but also in basal cells (yellow arrowheads) located at the periphery of the hair placode. Scale bars, 50 μm. f, Photographs of ex vivo cultured dorsal skin of E12.5 K14-rtTA;TetO-H2B-eGFP;Fucci-G1 mice on days 0, 2, 3 and 10. Yellow open arrowheads and filled arrowheads indicate first-wave and second-wave hair placodes, respectively. Dashed box shows magnified region (right). Scale bars, 100 μm. g, Immunostaining of dorsal HFs in day-9 ex vivo culture of E12.5 dorsal skin. The following tissue compartments were detected: bulge stem cells (NFATc1+SOX9+NPNT+KRT15+), sebaceous glands (LipidTOX+), hair matrix (CDH3+), dermal sheaths (SMA+), dermal papilla cells (SOX2+) and melanocyte (TRP2+). Scale bars, 50 μm. h, Immunostaining of day-9 ex vivo cultured dorsal HFs derived from E12.5 embryos for hair cell layer markers shown in Extended Data Fig. 2d. Ex vivo developing dorsal HFs had distinct cellular layers characteristic of HFs except for the medulla, which is formed in mature HFs. Scale bars, 50 μm. i, Lineage tracking data of long-term continuous imaging of dorsal HF development from the placode stage to the hair germ stage, which correspond to the bottom panels in Fig. 2c. Origin of prospective bulge SCs (red) was located at the periphery of the hair placode as observed in the whisker HF placode. Scale bars, 50 μm. j, Bee swarm plot of the distances of dorsal hair placode cells from the placode centre. HFs used for measurement were from cultured dorsal HFs of K14-rtTA;TetO-H2B-eGFP;Fucci-G1 and Sox9IRES-eGFP/+;R26-H2B-mCherry  mice. Values were scaled based on the diameter of each placode. Different epithelial lineages were aligned in a concentric manner in the placode. Two-sided nested t-test was used. k, Fate of basal and suprabasal cells in the dorsal HF placode analysed in i. Cell fates at the hair germ stage are shown. Summarized data are shown in the left panel, and the data for corresponding replicate HFs are shown in the right panels. Two-sided Fisher’s exact test was used. l, Stacked bar plots showing the lineage distribution of placode basal cells in the dorsal HF at the hair germ stage. Cells grouped in the black bar in k were examined. Summarized data are shown in the left panel, and the data for corresponding replicate HFs are shown in the right panel. mo, Replicate HFs related to Fig. 2d–f are shown in mo. m, Fate of basal and suprabasal cells in pre-placodes of Sox9-IRES-eGFP reporter derived dorsal skin explants. Fate of GFP+ cells at hair placode stage are shown. GFP+ cell lineages were determined at hair germ stage. n, Fate of basal and suprabasal cells in placodes of Sox9-IRES-eGFP reporter-derived dorsal skin explants. Fate of GFP+ cells at the hair germ stage are shown. o, Lineage distribution of GFP+ basal cells at hair germ stage. Cells grouped in the black bar in n were examined. p, Lineage-tracing strategy of Sox9+ cells in ex vivo cultured dorsal skin derived from E12.5 Sox9creERt2/+;R26R-H2B-mCherryfl/+;K14-H2B-eGFP (top panel) and snapshot images of the culture (bottom panels). Yellow spots, one of tracked basal cell lineages; Cyan spots, one of tracked suprabasal cell lineages. 4-OHT, 4-hydroxytamoxifen. Scale bars, 50 μm. q, Fate of basal and suprabasal cells in the dorsal HF placode analysed by lineage-tracing with Sox9-creER. Fates of H2B–mCherry+ cells at the hair germ stage are shown. Summarized data are shown in the left panel, and the data for corresponding replicate HFs are shown in the right panels. Two-sided Fisher’s exact test was used. r, Lineage distribution of H2B–mCherry+ basal cells at hair germ stage. Cells grouped in the black bar in q were examined. Summarized data are shown in the left panel, and the data for corresponding replicate HFs are shown in the right panels. Each value in the graph is the mean ± s.d. from three independent experiments, one or two HF each. Numbers of analysed cell lineages are summarized in Supplementary Table 2. See also ‘Statistical analysis and reproducibility’ in Methods and Source Data.

Source data

Extended Data Fig. 6 Cell sorting of photo-converted developing whisker HF epithelial cells.

a, Photo-labelling of whisker HF epithelial cells in nKikGR mice at each embryonic stage. Scale bars, 50 μm. b, Single plane of E15.0 whisker HF showing the nKikGR signal intensities before and after the photo-conversion (Supplementary Video 9). Expression level of nKikGR varies between cells. There are two possible reasons for this: (1) the expression level of nKikGR varies for different cell types, and (2) the different mesenchymal tissue thickness of the explants around the whisker HFs affects the efficiency of detecting the fluorescent signals from the inside of the tissue. Despite this, the variation of nKikGR-green signal levels within the HF before photo-conversion closely correlates with the variation of nKikGR-red signal levels after photo-conversion, suggesting that the nKikGR signal was completely converted from green to red throughout the HF tissues (Supplementary Video 9). Scale bars, 50 μm. c, Experimental design for scRNA-seq of the developing whisker HF epithelium. d, Immunofluorescence staining of E14.0 whisker HF showed that the basal epithelial layer was marked by ITGA6+CD31 and that blood vessels in the mesenchyme were labelled by ITGA6+CD31+ (arrowhead). Box shows magnified region (right). Scale bars, 50 μm. eg, Isolation of photo-labelled whisker HF epithelial cells (E12.0–E17.0, DAPICD31ITGA6+KikGR-red+ cells; E11.5, DAPICD31ITGA6+ cells) by flow cytometry. hl, Immunolocalization of ITGA6 in developing whisker HFs. Whisker HFs at E12.0 (h), E13.0 (i), E14.0 (j), E15.0 (k) and E17.0 (l) were stained with an antibody against ITGA6, and sections were counterstained with DAPI. ITGA6 was detected in all basal epithelial cells of developing whisker HFs. Blood vessels in the dermis also were positive for ITGA6. Scale bars, 100 μm.

Extended Data Fig. 7 Spatial and temporal reconstruction of scRNA-seq data of the developing whisker HF epithelium.

a, b, t-SNE plot visualizing scRNA-seq data for 1,614 single cells from developing whisker HFs (E11.5, 94 cells; E12.0, 276 cells; E13.0, 267 cells; E13.5, 181 cells; E14.0, 177 cells; E15.0, 350 cells; and E17.0, 269 cells). Cell populations were categorized by embryonic stage (a) and cell type (b). c, Violin plot showing the expression patterns of known lineage markers in each cluster. Colours refer to t-SNE clusters in b. d, Subclustering of whisker HF epithelial cells identified in ac. t-SNE plot of 962 epithelial cells is coloured by embryonic stage. e, Low batch effects and high reproducibility of scRNA-seq data; scRNA-seq samples in this study consisted of cells from multiple batches: a plurality of plates and different dates and places of sampling, library preparation and sequencing. To investigate the batch effect caused by this technical handling, transcriptomes were divided into stages and projected onto the t-SNE plot corresponding to d. Grey spots in the t-SNE plot indicate all transcriptomes derived from E11.5–E17.0 epithelial cells. Cells derived from each stage were highlighted with different colours according to the batch of experiments. Cells from different batches were mixed stage-by-stage on the t-SNE plot and not clustered by batch. This suggests the low batch effects and high reproducibility of our scRNA-seq analysis. f, t-SNE plot in d is coloured by cluster annotations. g, Percentage cell cluster distribution per individual plate. Each plate derived from the same embryonic stage showed a similar distribution of the clusters. h, Heat map of the heterogeneity within and between clusters. i, Dot plot showing differentially expressed genes in each cluster identified in f. j, Expression of cell-type-specific marker genes projected onto the t-SNE plot in d. k, Expression of the marker genes projected onto the t-SNE plot are shown in the top panels, and the corresponding RNA ISH results are shown in the bottom panels. Grey spots in the t-SNE plot indicate all transcriptomes derived from E11.5–E17.0 epithelial cells. Only cells derived from E17.0 are highlighted and coloured according to relative expression levels of the markers. Brackets indicate the intended localization of ISH signals. Scale bars, 100 μm.

Source data

Extended Data Fig. 8 In vivo expression patterns of representative genes in each cell cluster of E13.0–E17.0 whisker HFs.

Feature plots of characteristic differentially expressed genes in each cluster identified in Extended Data Fig. 7f are shown in the left panels. Feature plots divided by developmental stages of the whisker HFs (E13.0–E17.0) and the corresponding RNA ISH images are shown on the top and bottom right panels, respectively. Grey spots in the t-SNE plot indicate all transcriptomes derived from E11.5–E17.0 epithelial cells. Only cells derived from the indicated stage are highlighted and coloured according to relative expression levels of the indicated markers. Arrowheads indicate the intended localization of ISH signals. A cell cluster located at the lower part of the t-SNE plot (clusters 5 and 6) corresponded to the upper part of the HF, such as the infundibulum and junctional zone. A bulge SC marker (Nfatc1) was strongly expressed in cell clusters 7–12 located in the centre of the t-SNE plot. These clusters were divided into the upper and lower bulge regions, based on the expression pattern of each region marker (such as Adamts20 and Shisa2). Vdr expression confirmed that clusters 7, 11 and 13 contained cells derived from the stalk region, the lower part of the HF. Shh-positive cells in cluster 14 were in the hair germ in vivo. These data indicated that cells were aligned from the bottom left to the top right on the t-SNE plot, reflecting the tissue architecture of the whisker HF. Scale bars, 100 μm.

Extended Data Fig. 9 Pseudospace analysis of scRNA-seq data derived from E12.0 placode cells.

a, In vivo expression patterns of representative genes in the pseudospace of E12.0 whisker HFs. Expression of the indicated marker genes projected onto the pseudospace are shown in the left panels, and the corresponding whole-mount RNA ISH results are shown in the right panels. Arrowheads indicate the intended localization of ISH signals. Scale bars, 100 μm. be, Expressions of indicated genes in an E14.5 primary dorsal HF placode were detected with whole-mount RNAscope fluorescent ISH. A merged image (left) and individual images are shown. Scale bars, 50 μm. fj, Expression of genes related to signalling pathways in pseudospace of the hair placode. Genes involved in WNT (f), TGF, BMP and activin (g), Hedgehog (h), FGF (i) and Notch signalling (j) were selected from the KEGG database and the expression of the genes is represented in a heat map. Direct target genes of WNT and β-catenin signalling, including Axin2 and Tcf4, showed high expression at the placode centre but suppression towards the IFE. However, the expression of Notch, FGF and TGF and BMP target genes, including Heyl, Etv genes and Id3, was more broadly elevated from the placode centre to the periphery. SHH target genes Igf2 and Ccnd2 were highly expressed from the periphery to the IFE regions. Thus, the placode basal epithelium forms gradients of signalling activities from the placode centre to the periphery, and the placode periphery is characterized, in part, as a WNTlowBMPhigh state, the characteristics of adult bulge SCs.

Extended Data Fig. 10 Reconstruction of epithelial lineage diversification and SC induction.

a, Timing of epithelial lineage divergence. Three-dimensional diffusion map showing pseudotemporally ordered whisker HF epithelial basal cells, related to Fig. 4c. Grey spots in the diffusion map indicate all transcriptomes derived from E11.5–E17.0 whisker HF epithelial basal cells. Cells derived from each stage are highlighted with different colours according to their embryonic stage. The arrow in the E12.0 diffusion map indicates the branching point of the hair germ lineage (branch 5). The open arrowhead in the E13.5 diffusion map indicates the branching point of infundibulum (branch 1), upper SCs (branch 2), lower SCs (branch 3) and stalk cells (branch 4). b, Cell clusters belonging to each trajectory are highlighted on the diffusion map (top panels) and t-SNE plot (bottom panels). Arrow and open arrowheads represent branching points. c, Expression of each lineage marker gene projected onto the diffusion map. d, RNA velocity field projected onto the t-SNE plot of epithelial cells (arrows represent the average RNA velocity). e, Selected terms of enrichment analysis associated with the gene categories identified in Fig. 4d. Only the gene list of gene cluster 6 raised no valid GO terms (P value < 0.03). f, HFSC signature genes and pSMAD1 targets in gene clusters 6, 9 and 10.

Supplementary information

Reporting Summary

Peer Review File

Supplementary Table 1

The detailed sample size information and analysed cell lineage numbers for Fig. 1 and Extended Data Fig. 3.

Supplementary Table 2

The detailed sample size information and analysed cell lineage numbers for Fig. 2 and Extended Data Fig. 5.

Supplementary Table 3

Information on scRNA-seq data used in this study.

Supplementary Table 4

List of resources and primers used for ISH probe synthesis.

Video 1 Time-lapse video of whisker HF development from the hair germ to bulbous peg stages with cell tracking data

Single plane time-lapse video of whisker HF development from the hair germ to bulbous peg stages with cell tracking data corresponding to Fig. 1a (replicate No. 1). Epithelial nuclei were visualised by K14-rtTA;TetO-H2B-eGFP and all nuclei were visualised by R26-H2B-mCherry.

Video 2 Another time-lapse video of whisker HF development from the hair placode to bulbous peg stages

Another single plane time-lapse video of whisker HF development from the hair placode to bulbous peg stages, related to Fig. 1a and corresponding to replicate No. 3 in Extended Data Fig. 3d and replicate No. 2 in Extended Data Fig. 3g. Epithelial nuclei were visualised by K14-rtTA;TetO-H2B-eGFP and G0/G1 phase nuclei were visualised by a Fucci-G1 probe (mKO2-hCdt1).

Video 3 Time-lapse video of whisker HF development from the hair placode to hair germ stages

Single plane time-lapse video of whisker HF development from the hair placode to hair germ stages, corresponding to Fig. 1c (replicate No. 1). Epithelial nuclei were visualised by K14-rtTA;TetO-H2B-eGFP and G0/G1 phase nuclei were visualised by Fucci-G1 probe (mKO2-hCdt1).

Video 4 Video 3 with cell tracking data

Supplementary Video 3 with cell tracking data, corresponding to Fig. 1c. Epithelial nuclei were visualised by K14-rtTA;TetO-H2B-eGFP and G0/G1 phase nuclei were visualised by Fucci-G1 probe (mKO2-hCdt1).

Video 5 Time-lapse video showing cell tracking results of suprabasal cells at the placode stage

Time-lapse video containing four representative clips showing cell tracking results of suprabasal cells at the placode stage. Tracked suprabasal cells, equivalent to SOX9+ suprabasal cells, did not contribute to the basal layer or HF formation. Tracked cells are marked by spots.

Video 6 Time-lapse video showing cell tracking results of an upper daughter cell that remained in the basal layer after perpendicular cell division

Time-lapse video containing four representative clips showing cell tracking results of an upper daughter cell that remained in the basal layer after perpendicular cell division. Tracked cells are marked by spots.

Video 7 Time-lapse video of dorsal HF development from the hair placode to hair peg stages

Single plane time-lapse video of dorsal HF development from the hair placode to hair peg stages, related to Extended Data Fig. 5i–l. Epithelial nuclei were visualised by K14-rtTA;TetO-H2B-eGFP.

Video 8 Another time-lapse video of dorsal HF development from the hair placode to hair peg stages

Single plane time-lapse video of dorsal HF development from the hair placode to hair peg stages, corresponding to Fig. 2c. Sox9 expression was visualised by Sox9-IRES-eGFP reporter, and all nuclei were visualised by R26-H2B-mCherry.

Video 9 nKikGR signal intensities in E15.0 whisker HF before and after photo-conversion

Representative 3D video showing nKikGR signal intensities in an E15.0 whisker HF before and after photo-conversion. The photo-conversion occurs throughout the tissue.

Source data

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Morita, R., Sanzen, N., Sasaki, H. et al. Tracing the origin of hair follicle stem cells. Nature 594, 547–552 (2021). https://doi.org/10.1038/s41586-021-03638-5

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links