The disruption of Sox21-mediated hair shaft cuticle differentiation causes cyclic alopecia in mice

Abstract

Hair is maintained through a cyclic process that includes periodic regeneration of hair follicles in a stem cell-dependent manner. Little is known, however, about the cellular and molecular mechanisms that regulate the layered differentiation of the hair follicle. We have established a mutant mouse with a cyclic alopecia phenotype resulting from the targeted disruption of Sox21, a gene that encodes a HMG-box protein. These mice exhibit progressive hair loss after morphogenesis of the first hair follicle and become completely nude in appearance, but then show hair regrowth. Sox21 is expressed in the cuticle layer and the progenitor cells of the hair shaft in both mouse and human. The lack of this gene results in a loss of the interlocking structures required for anchoring the hair shaft in the hair follicle. Furthermore, the expression of genes encoding the keratins and keratin binding proteins in the hair shaft cuticle are also specifically down-regulated in the Sox21-null mouse. These results indicate that Sox21 is a master regulator of hair shaft cuticle differentiation and shed light on the possible causes of human hair disorders.

Fig. 1.

Cyclic alopecia phenotype and histological features of Sox21 KO mice. (A) Hair loss in the same mouse from P7 to P55; images were taken every 4 days. (B) Histological changes in the Sox21 mutant mice during different hair cycle stages. (C) Oil-red O staining of sections at P12 in Sox21-null mice. Arrows indicate sebaceous glands. Brackets indicate epidermal layers. Magnified images are provided to show epidermal layers and sebaceous glands. Cross-sections for sebaceous gland are also shown. (Scale bars: B, 200 μm; C, 50 μm.)

Fig. 2.

The restricted expression of Sox21 within the hair follicle. (A and B) Sox21 expressions detected by in situ hybridization using a Sox21 cRNA probe in Sox21^+/+ (A) and Sox21^−/− (B) at P28. (C) A schematic illustration of the layered structure and specific keratins expressed in each follicular layer. (D–H) Immunohistological detection of Sox21 in P5 Sox21^+/− hair follicles using K14 for the ORS (D), K71 for the IRS (E), K32 for the Cuh (F), K31 for the cortex (G), and the cell cycle marker PCNA (H). Enlarged images are provided with single confocal channels: F′ and F″ for F, G′ and G″ for G, and H′ and H″ for H. Arrowheads indicate Sox21 signals. (I–N) Immunohistological detection of Sox21-GFP in Sox21^+/− (I, K, and M) and Sox21^−/− (J, L, and N) mice using K14 (I and J), K71 (K and L), and K31 (M and N). Enlarged images are shown for double and single (GFP) channels in I′–N″. Arrows indicate ectopic Sox21-GFP expression in Sox21^−/− samples. (O–R) Immunohistological detection of the cuticle specific proteins K82 (O and P) and K32 (Q and R) in Sox21^+/− (O and Q) and Sox21^−/− hair follicles at P5. (S) Immunohistological detection of human Sox21 and K31 in the hair follicle of the calvarial skin. An enlarged image is shown in S′. (Scale bars: 25 μm in A–R, 250 μm in S.)

Fig. 3.

Defective Cuh in the Sox21-null mouse. (A–D) Comparison of hair samples during the second hair cycles from Sox21^+/− (A and B) and Sox21^−/− (C and D) mice by SEM analyses. (E–H) Transelectron micrographs of longitudinally-sectioned anagen hair follicles of P14 WT (E and F) and Sox21^−/− (G and H) mice. Micrographs were obtained from the same anatomical level in the suprabulbar area, where trichohyaline granules (Th) emerge in Huxley's layer. Note that the interface between Cuh and the Cui form an interlocking structure (arrows) to anchor the developing hair shaft at the base of hair follicle (E and F). Co: cortex; Hu: Huxley's layer, He: Henle's layer, Cl: companion layer, Th: trichohyaline granule. (Bars: 1 μm.) (I and J) Schema illustrating the ultrastructural anomalies observed in the Sox21 KO mouse.

Fig. 4.

Sox21 is a master regulator of normal keratinization in the cuticle of the hair shaft. (A) Summary of comparative microarray analyses of P5 back skin samples from WT, heterozygous, and homozygous mice. Gene ontology and expression profiling of the 119 probe sets are shown in Table S1 and a graph (Right). (B) Real-time Q-PCR analysis (blue) confirmed the results of the microarray analysis (black). All of the expression data were calculated relative to the WT levels.