Lrig1 expression defines a distinct multipotent stem cell population in mammalian epidermis

Abstract

Lrig1 is a marker of human interfollicular epidermal stem cells and helps maintain stem cell quiescence. We show that, in mouse epidermis, Lrig1 defines the hair follicle junctional zone adjacent to the sebaceous glands and infundibulum. Lrig1 is a Myc target gene; loss of Lrig1 increases the proliferative capacity of stem cells in culture and results in epidermal hyperproliferation in vivo. Lrig1-expressing cells can give rise to all of the adult epidermal lineages in skin reconstitution assays. However, during homeostasis and on retinoic acid stimulation, they are bipotent, contributing to the sebaceous gland and interfollicular epidermis. beta-catenin activation increases the size of the junctional zone compartment, and loss of Lrig1 causes a selective increase in beta-catenin-induced ectopic hair follicle formation in the interfollicular epidermis. Our results suggest that Lrig1-positive cells constitute a previously unidentified reservoir of adult mouse interfollicular epidermal stem cells.

Figure 1

Lrig1 Expression Defines a Distinct Population of Epidermal Cells in the Junctional Zone (A–H) Immunofluorescence labeling of sections of embryonic (A) and adult back (C) and tail (D) skin and whole mounts of embryonic (B) and adult (E–H) tail epidermis. Adult skin was in telogen (C–F and H) or early anagen (G). Skin was from wild-type (A, B, upper panel of C, D, E, G, and H) or Lrig1-null mice (lower panel of C, F). Insert in (G) shows the bulb at a higher magnification. Color coding indicates antibody labeling. Dashed lines represent the boundary between dermis and epidermis (A and C) or demarcate hair follicles (B and G). epi, epidermis; de, dermis; IFE, interfollicular epidermis; SG, sebaceous gland; JZ, junctional zone; BU, bulge; BB, bulb; HFSC, developing hair follicle stem cell compartment; PL, hair follicle placode; HG, hair germ. (A) is shown at a higher magnification in Figures S1F–S1H. Scale bars, 25 μm (A and D) and 100 μm (B and E–H). (I) Diagram illustrating the different regions of adult epidermis. Red, junctional zone; green, bulge.

Figure 2

Characterization of Lrig1-Expressing Cells in Adult Epidermis (A–D) Immunofluorescence labeling of tail epidermal whole mounts from wild-type mice. Skin was in anagen (A) or telogen (B–D). Color coding indicates antibody labeling. Scale bars, 100 μm. (E) Flow cytometry of telogen epidermis from wild-type (WT) and Lrig1-null (KO) epidermis disaggregated with thermolysin and labeled with antibodies to α6 integrin, Lrig1, CD34, and Sca1. Colored gates in top-right panel correspond to populations in lower-right panel and in (F) and (G). (F) Colony-forming efficiency of wild-type primary keratinocytes isolated as in (E). (G) Q-PCR of RNA from 10⁵ wild-type epidermal cells isolated as in (E). (F and G) Data are means ± SEM (n = 4).

Figure 3

Epidermal Reconstitution by Lrig1-Expressing Cells (A–C) Keratinocytes isolated from the back skin of wild-type adult telogen (7-week-old) mice were sorted into α6 integrin-high (I) and -low (II) populations (A). Populations I (B) and II (C) were gated into three further populations on the basis of Sca1 and CD34 expression, yielding a total of six discrete populations of keratinocytes (1–3 in [B]; 4–6 in [C]). (D) RNA from each of the six populations or total live cells (low forward/side scatter; all) were analyzed using Q-PCR for the genes indicated (mean ± SEM, n = 5). (E–P) Epidermal cells from telogen back skin of GFP-expressing mice were fractionated based on Sca1, α6 integrin, and CD34 expression as in (A)–(C) or Lrig1 (as in Figure 2E) for epidermal reconstitution experiments by mixing GFP-positive and GFP-negative epidermal cells. Four cell populations were compared: α6 high (ItgA6H), CD34 positive, Sca1 negative (population 1; enriched for bulge cells; [E], [I], and [M]); α6 high, CD34 negative, Sca1 negative (population 2; enriched for Lrig1-positive cells; [F], [J], and [N]); α6 high, CD34 negative, SCA1 positive (population 3; [G], [K], and [O]); or Lrig1 positive ([H], [L], [P]). (E–H) Dermal view of grafts; (I-P) Sections of grafts labeled with GFP antibody (brown) and hematoxylin counterstain (blue). Interfollicular epidermis (IFE), hair follicles (HF), and sebaceous glands (SG) are indicated. Scale bars, 100 μm (I–L) and 25 μm (M–P).

Figure 4

Effects of Lrig1 Loss In Vivo (A and B) Hematoxylin and eosin-stained sections of adult tail skin. (C and D) Macroscopic view of adult tails. (E, I, G, H, K, L, N, and O) Tail whole mounts of P1 (G and K) and adult telogen (E, I, N, and O) and anagen (H and L) epidermis. WT, wild-type; KO, Lrig1 null. Color coding indicates antibody labeling. (F and J) Flow cytometry of (F) WT and (J) KO Sca1-low cells with antibodies to CD34 and α6 integrin. The percentage of cells in each quadrant is indicated. (M) Quantitation of Ki67-positive cells in KO IFE expressed relative to WT controls. Data shows mean ± SEM; average of three fields per whole mount and five mice per sample. (N–P) BrdU label-retaining cells were examined following a 70 day chase period. (P) Cells were subjected to flow cytometry with antibodies to Sca1, CD34, and α6 integrin. The percentage of cells in each group that were BrdU labeled is shown. Error bars represent SD (WT n = 8; KO n = 6). WT, wild-type; KO, Lrig1 null. Scale bars, 100 μm (A, B, G, H, K, M, and O) and 200 μm (E and I).

Figure 5

Regulation of Lrig1 Expression by Myc (A) (Top) Immunoblot of protein lysates from tail skin of adult littermates lacking (KO) or heterozygous (Het) for Lrig1. Each lane contains protein from a separate mouse. Blot was probed for Myc or, as a loading control, β-tubulin (Tub). (Bottom) Q-PCR for cMyc from adult littermates lacking (KO) or heterozygous (Het) for Lrig1. Expression levels are relative to Gapdh, and error bars represent SD (n = 3). (B) Q-PCR for Lrig1 mRNA in back skin of WT and K14MycER-transgenic mice treated with 4OHT for 8 days. Expression levels are relative to Gapdh, and error bars represent SD (n = 3). (C) Luciferase assay using 1.6 kb of the Lrig1 promoter in murine keratinocytes from WT or K14MycER mice, represented as induction in K14MycER cells relative to WT cells. Cells were treated with the 4OHT concentrations indicated (nM). Error bars represent SD (n = 4). (D–G) Back skin sections (D and E) and tail epidermal whole mounts (F and G) of wild-type (WT) and K14MycER mice treated with 4OHT for the number of days shown. Color coding indicates antibody labeling. Scale bars, 100 μm. (H) Location of two E boxes in the Lrig1 promoter. (I) Quantitative ChIP from keratinocytes isolated from back skin of WT or K14MycER-transgenic mice treated with 4OHT for 4 days. (J) Knockdown of cMyc in primary cultures of WT keratinocytes causes a decrease in Lrig1 transcript levels. Error bars represent SD (n = 4).

Figure 6

Lineage Analysis of Junctional Zone Cells (A) Schematic summary of the two mouse lines that were crossed for lineage tracing. Leakiness of Cre expression in the absence of 4OHT was exploited for clonal analysis. (B and C) Sections of back skin stained for GFP, showing two examples of junctional zone clones. (B′) Insert in (B) shown at higher magnification. (D) Schematic diagram showing contribution of junctional zone clones (green) to IFE and SG in control (left) and ATRA-treated skin. (E) Cumulative frequency plots of the size distribution of GFP-positive clones originating in the IFE, junctional zone, infundibulum, SG, and HF (below the level of the SG). Clone size (volume) was determined from 3D reconstructions of confocal images. Diagrams represent ranking based on cumulative percentage of clones as a function of clone size (exponential scale) in control (blue) and ATRA-treated (red) skin. (F and G) Three-dimensional projections of whole-mount reconstructions stained for GFP and LRIG1. Color coding indicates antibody labeling. GFP expression is shown by isosurface labeling. (F) Control. (G) Treated with ATRA. (H) Clone size distribution in control (ctrl) and ATRA-treated skin. Data from (E) are replotted as box and whisker plot of GFP clone sizes (volume) in the different epidermal regions in control and ATRA-treated tail epidermis. Minimum and maximum clone sizes are marked by whiskers; upper and lower box boundaries indicate clones in the 25th to 75th percentile, and cross (+) indicates median clone size. Asterisk indicates statistically significant change in median clone size following ATRA treatment (p < 0.05, two-tailed unpaired Mann-Whitney test). Scale bars, 100 μm (B, C, F, and G) and 25 μm (B′).

Figure 7

Effect of Lrig1 Loss on Stem Cell Renewal and Responsiveness to β-Catenin Activation (A) Clonal growth assays of primary keratinocytes sorted based on CD34, Sca1, and α6 integrin (Figure 3A–C) from the skin of wild-type (WT) and Lrig1-null (KO) adult littermates. Error bars represent SD (n = 3). Six-hundred Sca1-negative cells and 2500 Sca1-positive and unfractionated cells were seeded; representative dishes are shown. (B and C) Tail epidermal whole mounts of WT and K14ΔNβ-cateninER mice treated with 4OHT for 3 weeks. Arrow and insert indicate Lrig1-positive ectopic follicles in interfollicular epidermis. (D and E) Flow cytometric analysis for Lrig1 and α6 integrin of cells from WT and K14ΔNβ-cateninER mice treated with 4OHT for 3 weeks. The proportion of cells in each fraction is indicated. (F) Q-PCR of Lrig1 and cMyc mRNA in back skin of WT and K14ΔNβ-cateninER-transgenic mice treated with 4OHT for 2 weeks. Expression levels are relative to Gapdh, and error bars represent SD (n = 3). (G) Q-PCR of levels of Lrig1, Axin2, and Jagged1 in primary murine epidermal keratinocytes treated with Wnt3A. Data are expressed relative to unstimulated cells. Error bars represent SD (n = 3). (H) ChIP analysis of endogenous cMyc on the Lrig1 promoter in WT and K14ΔNβ-cateninER mice treated with 4OHT for 10 days. Data represent two separate samples and show level of isolated genomic DNA relative to amount of input DNA. (I–M) Hematoxylin and eosin stained sections (I and J) and whole mounts (L and M) of adult tail epidermis from K14ΔNβ-cateninER × Lrig1 heterozygous (het) or knockout (KO) mice treated with 4OHT for 3 weeks. The number of ectopic HFs formed from the interfollicular epidermis was scored by morphology (K) and by the appearance of clusters of CDP-expressing cells (N) (two independent experiments; KO, n = 8; WT, n = 7). Arrows in (J) indicate expanded infundibulum with associated ectopic follicles. In (B), (C), (L), and (M), color coding indicates antibody labeling. Scale bars, 100 μm (B and C) and 200 μm (I, J, L, and M).