Is the Hair Follicle Necessary for Normal Wound Healing?

Abstract

The hair follicle contributes cells to the interfollicular epidermis after wounding, but the functional role of these cells has not been resolved. To address this question, Langton et al. (this issue, 2008) take advantage of the Edaradd mutant mouse, which lacks hair follicles on its tail. They discover an initial sluggish response of the hairless tail epidermis to wounding that is rapidly compensated for by recruitment of epidermal cells from outside the normally responsive area. This suggests that the hair follicle is important but not necessary for normal wound healing.

Investigators and clinicians have noted for years that hair-bearing areas tend to heal more quickly than areas lacking follicles (Bishop, 1945; Brown and McDowell, 1942; Martinot et al., 1994). These observations suggested that optimal wound healing, which involves both an epidermal and a dermal response, depends on intact hair follicles. The exact relationship between the hair follicle and the epidermis is slowly being unraveled, although further work is needed to understand the role of hair follicle dermis during wound healing (Jahoda and Reynolds, 2001). For example, using genetic lineage analysis and transplantation techniques, several groups demonstrated that during normal homeostasis the follicle does not contribute cells to the epidermis (Claudinot et al., 2005; Ito et al., 2005; Levy et al., 2005). This indicates that the epidermis possesses enough proliferative capacity for self-renewal throughout the lifetime of the animal. But what happens when the system is stressed?

Wounding perturbs epidermal homeostasis by cell depletion and results in activation of epithelial cells in the hair follicle. These cells then provide progeny, which migrate to the epidermal defect and assist in re-epithelializa tion (Ito et al., 2005; Levy et al., 2007). Although movement of cells from the hair follicle to the epidermis following wounding clearly takes place, its exact functional contribution to epidermal healing has not been demonstrated. The work by Headon and colleagues in this issue demonstrates for the first time that hair follicles contribute to the acute phase of re-epithelialization (Langton et al., 2008).

Using a mouse carrying a genetic mutation in Edaradd—a member of the ectodermodysplasin receptor pathway, which prevents development of all appendageal structures in tail skin— these investigators were able to tease out the contribution of the hair follicle to wound healing. Linear incisional wounds on the tail skin healed differently in mice lacking hair follicles. In particular, the wound on the hairless skin did not close as quickly initially; however, after a lag period, the wound re-epithelialized at the same time as that of the control.

The outer root sheath of the hair follicle is continuous with the interfollicular epidermis, allowing keratinocytes to move from the hair follicle to the epidermis (Cotsarelis, 2006). The outer root sheath immediately adjacent to the interfollicular epidermis runs down to the sebaceous gland duct and forms the infundibulum. The area between the sebaceous gland duct and the bulge is called the isthmus. One question that remains is whether there are differences in response to epidermal injury between subpopulations of cells within the hair follicle.

From previous work using genetic lineage analysis, we obtained information about the contribution of cells from various portions of the follicle to the epidermis after wounding (Table 1). Using an inducible K-15 promoter to drive Cre recombinase to the bulge cells, we demonstrated that hair follicle bulge cells migrated to the epidermis following wounding (Ito et al., 2005). Approximately 25% of the cells in a reepithelialized wound originated in the bulge. These cells did not persist in the newly formed epidermis; rather, they behaved as transient amplifying cells and were lost after several weeks. Levy et al. (2007) used a sonic hedgehog promoter in a similar Cre-Lox reporter system to target the great majority of hair follicle cells, including the bulge cells, isthmus, and many infundibular cells (Levy et al., 2007). They found hair follicle progeny in the re-epithelialized epidermis after wounding, but many hair follicle-derived colonies persisted for long periods. Taking together the studies by Ito and Levy, we can conclude that nonbulge follicular keratinocytes, either in the upper isthmus or in the infundibulum, behave differently than bulge cells and possess the capacity to contribute permanent residents to the epidermis after wounding.

Table 1.

Hair follicle contribution to re-epithelialization following wounding

Labeling method	Initial labeled population	Lineage tracing	Location of the labeled cells under homeostatic conditions	Location of labeled cells above the bulge after wounding	Percentage of labeled cells in the re-epithelialized area		Reference
					Short-term analyses	Long-term analyses
BrdU-label retaining cells	Bulge cells	Transient	All layers of the hair follicle epithelium	Infundibulum	Not quantitated	Not done	Taylor et al., 2000
GFP-label retaining cells	Bulge cells	Transient	Hair follicle epithelium below the bulge	Infundibulum and epidermis	Not quantitated	Not done	Tumbar et al., 2004
K15crePR1/R26R transgenic mice	Bulge cells	Permanent	Hair follicle epithelium below the bulge	Infundibulum and epidermis	26%, 8 d after wounding	4%, 20 d after wounding	Ito et al., 2005
ShhGFPcre/R26R transgenic mice	All layers of the hair follicle epithelium	Permanent	All layers of the hair follicle epithelium	Infundibulum and epidermis	50% (23%*), 1–2 weeks after wounding	(37%*), 4 months after wounding	Levy et al., 2007

Results from several studies indicate cells from the hair follicle move to the epidermis after wounding. Cells originating above the bulge persist in the epidermis.

^*Percent labeled basel cells.

Is the hair follicle necessary for healing of the skin? According to the results of Langton et al. (2008), a hair follicle contribution is required to initiate rapid re-epithelialization but not to close the wound. A potential limitation of the study was the use of a mutant mouse that lacked Edaradd. It is possible that this mutation affected wound healing through different mechanisms. The authors confirmed that the major characteristics of epithelial stem cells, including the number of clonogenic cells and label-retaining cells in the interfollicular epidermis of mutant mice, did not differ significantly from those in control mice. This allowed the authors to presume that differences observed in the wound healing process could be attributed to the absence of hair follicles. However, the possibility exists that unrecognized cellular or molecular differences in mutant skin could delay the early steps of wound healing or accelerate the later stages.

Another point to keep in mind is that hair follicles are relatively sparse on the tail of a mouse compared with its back skin. Nonetheless, the delay of epithelialization at 4 days after an incisional wound indicates that the hair follicle does contribute to wound closure. In areas with higher follicle densities, such as the back skin, it is likely that the hair follicles contribute to an even greater extent. Extrapolating to humans is difficult, but the findings corroborate the observation that the scalp heals better than other non-hair-bearing areas, although other explanations—such as the higher level of vasculature in the scalp, which could also contribute to rapid healing—have not been ruled out.

As with most good studies, this work raises many questions. For example, the current study examined the effects of linear incisional wounds. In contrast, most previous lineage studies were carried out using full-thickness excisional wounds. Because of the spreading of the wound that occurs even after a linear incision, as well as the similar response of the bulge cells to linear and excisional wounds (Ito et al., 2005), it is likely that the findings reported in this article are relevant to excisional wounds as well. However, the possibility exists that the interfollicular epidermis would not have the proliferative capacity to close a larger excisional wound or that the re-epithelialization process would be markedly delayed. Keeping in mind that the defect created was through an incision, the main message of the work is that the skin relies on the hair follicle for its reparative response but that the epidermis can compensate for loss of follicles by calling on its vast proliferative potential.

The evolutionary pressure for wounds to heal quickly must have been great. Even small delays in healing would likely result in increased rates of infection, hemorrhage, desiccation, and even death. Perhaps it is not surprising, then, that even by eliminating large reservoirs of keratinocytes in the follicle, the compensatory ability of the remaining cells is great enough to overcome this deficiency, at least for small linear incisional wounds.

Another question that remains is whether the new epidermis that formed in the tail functions normally compared with epidermis derived from a combination of follicular and interfollicular epithelium. For example, it is possible that a new epidermis formed following the wounding of the mutant mice either would not persist or would not respond to further injury as robustly as epidermis derived from hair follicles?

Little is known about the homeostasis of the wound epidermis postclosure, whereas the behavior of the epidermal cells during wound closure has been extensively investigated. Considering that squamous-cell carcinomas arise at a higher rate within the re-epithelialized epidermis of healed burn wounds long after injury, it will be particularly important to understand whether and how stem/progenitor cells are re-established during re-epithelialization and how follicular cells are involved in this process. The contribution of cells within the infundibulum, isthmus, and bulge of the hair follicle to the overlying epidermis still needs to be determined. The difficulty lies in the fact that no promoters have been isolated that specifically target these distinct regions of the hair follicle.

To understand further the relationship between the hair follicle and the interfollicular epidermis, the investigators carefully examined the number of colony-forming units that arose in these two areas. Remarkably, they discovered that the embryonic epidermis possessed a large number of colony-forming units; after birth, however, the number of colonies formed from the interfollicular epidermis dropped dramatically. In the neonatal period and in adults, the greatest number of colony-forming units arose from cells isolated from the hair follicle, including the bulge area. Intriguingly, this shift in proliferative potential from the embryonic epidermis to the adult bulge correlates with the previously reported preferential expression of keratin-15 in the neonatal and embryonic epidermis and then to its retreat to the hair follicle bulge in adults (Liu et al., 2003). Thus, K-15 expression and K-15 promoter activity patterns seem to correlate with colony-forming ability.

This work indicates that skin injury triggers a massive migratory and proliferative response from adjacent cells in the epidermis and hair follicle. Whether the response of the follicle and epidermis is triggered by the same or different factors remains to be determined. Possibly, the cells simply sense the loss of their neighbors and begin a march toward the depleted area. Understanding signals influencing recruitment of keratinocytes from epidermis and follicle, as well as the sensors directing cessation of proliferation and migration will provide new insights into wound repair and other cutaneous responses associated with malignancies and other disorders.

Hair follicle cells accelerate the onset of wound healing.