Highly upregulated Lhx2 in the Foxn1-/- nude mouse phenotype reflects a dysregulated and expanded epidermal stem cell niche

Abstract

Hair cycling is a prime example of stem cell dependent tissue regeneration and replenishment, and its regulatory mechanisms remain poorly understood. In the present study, we evaluated the effect of a blockage in terminal keratinocytic lineage differentiation in the Foxn1(-/-) nude phenotype on the epithelial progeny. Most notably we found a constitutive upregulation of LIM homeobox protein 2 (Lhx2), a marker gene of epithelial stem cellness indispensible for hair cycle progression. However, histological evidence along with an erratic, acyclic rise of otherwise suppressed CyclinD1 levels along with several key markers of keratinocyte lineage differentiation indicate a frustrated expansion of epithelial stem cell niches in skin. In addition, CD49f/CD34/CD200-based profiling demonstrated highly significant shifts in subpopulations of epithelial progeny. Intriguingly this appeared to include the expansion of Oct4+ stem cells in dermal fractions of skin isolates in the Foxn1 knock-out opposed to wild type. Overall our findings indicate that the Foxn1(-/-) phenotype has a strong impact on epithelial progeny and thus offers a promising model to study maintenance and regulation of stem cell niches within skin not feasible in other in vitro or in vivo models.

Figure 1. Schematic of mature, anagen hair follicle (HF) morphology with typical localization of keratinocytic lineage progeny in wild type (WT).

CD49f/α6-integrin (blue) positive epithelium dominates within the basal layer of the surface epithelium. CD200 (purple) and CD34 (red) are both strongly expressed within the bulge region. In addition, CD200 is also expressed in upper parts of the outer root sheet (ORS) whereas CD34 expression may extend below the bulge region. Matrix cells of the hair bulb are considered CD49f/CD200/CD34 negative. In general, surface markers of epithelial progeny are not exclusive for each other in part depending on the stage of hair cycling. This study evaluates various effects of Foxn1 loss-of-function on telogen-anagen hair cycling and epithelial stem cell niche regulation in Nu/Nu mice.

Figure 2. Morphological features along with CyclinD1 levels suggest a continuous, yet frustrated activation of hair cycling within the nude mouse phenotype.

(A) The Foxn1^−/− (Nu/Nu) phenotype demonstrates spontaneous lanugo hair growth as opposed to a complete hairless condition. Histologically, superficial epithelial cysts/pseudo-infundibula which often fail to connect to the surface and associate with sebaceous gland cells (1) represent the equivalent of HF’s in wild type (WT). In addition, invaginations of a hyperplastic epithelium (2) resemble stages of de novo embryonic HF formation. With spontaneous lanugo hair formation, a dysmorphic root sheet (RS) (3) and infundibulum (4) containing a rudimentary hair shaft can be seen. (B) In WT telogen, a dermal papilla complex (5) gets activated by depilation and a newly forming RS expands both in width and length into the deeper dermis (6) in order to generate several terminal hair shafts. (C) Comparative CyclinD1 mRNA levels are suggestive for a telogen and anagen equivalent in the Nu/Nu phenotype (qPCR/ΔΔCt-method: n = 6; *P<0.05). The red horizontal line represents the telogen skin level used as reference. Scale bars: 4 mm (macroscopic); 50 µm (microscopic).

Figure 3. Comparative expression of stem cell markers in the Nu/Nu phenotype vs. WT.

(A) Following induction of a new anagen hair cycle by depilation, WT skin shows a distinct dynamic expression of the embryonic stem cell (SC) markers Sox2 and Oct4. (B) Conversely, in the Nu/Nu phenotype, both markers are permanently suppressed (qPCR/ΔΔCt-method: n = 4; *P<0.05). (C) Other epithelial lineage SC markers show distinct changes in expression compared to WT, notably of Lhx2. (D) IHC confirms a very localized Lhx2 expression in the bulge region (asterisk) of WT telogen (D1) HF’s that extents to the infundibulum and the epithelial surface (arrow) during anagen (D2/3). In the Nu/Nu phenotype Lhx2 is strongly expressed in both hyperplastic surface epithelium and bulge region proximity. Scale bars: 100 µm (D).

Figure 4. As Oct3/4+ population is expanded; CD49f+, CD34+ and CD200+ populations are reduced in epithelial isolates from newborn skin of the Foxn1^−/− phenotype vs. wild type.

(A) Flow cytometry summary table listing fractions positive for markers of progeny within epithelial isolates from newborns (%±SD, n>6). Distinct differences between the Foxn1^−/− phenotype and wild type are demonstrated by comparative sample flow cytometry data gated for (B) CD49f, CD34, CD200 and (C) Oct3/4 positive epithelial fractions.

Figure 5. Alterations in Wnt- and Shh-signaling component expression in the Nu/Nu phenotype.

(A) The canonical (Wnt5a/Cateninβ increased) and non-canonical (Wnt10b suppressed) Wnt-pathway is conversely altered in the Nu/Nu phenotype. (B) Also, with the exception of Shh itself, downstream components of Shh-signaling are upregulated in the Nu/Nu phenotype compared to WT telogen skin (qPCR/ΔΔCt-method: n = 4; *P<0.05).

Figure 6. Aspects of the Foxn1^−/− phenotype in epithelial isolates and in vitro.

(A) Newborn epithelial isolates on day one post birth (NBD1) of Foxn1^−/− background show strong upregulation of Lhx2 which is not maintained following three days of in vitro culture along with other marker genes of progeny (ND = non-detectable). (B) However, CD200 levels remain higher in cultured Foxn1^−/− isolates compared to WT (qPCR/ΔΔCt-method: n = 4; *P<0.05, ↓P<0.05). (C) Flow cytometry on day 5 of in vitro cultures reveals a shift from a dominant CD49f++ subpopulation in vivo in favor of a CD34+/CD49f+ positive epithelial subpopulation in vivo which is more pronounced in Nu/Nu isolates vs. WT.

Figure 7. Schematic of the effects of Foxn1 loss-of-function on keratinocytic lineage differentiation in nude mice.

(A) A severe disturbance in keratin synthesis and terminal differentiation in the Foxn1^−/− phenotype also affects upstream lineage progenitors with strong upregulation of Lhx2 as a key feature. (B) Altered Wnt- and Shh-signaling in the Foxn1^−/− phenotype.