The Effects of Ischemia and Hyperoxygenation on Hair Growth and Cycle

ABSTRACT

Alopecia has several causes, but its relationship with ischemia/hypoxia has not yet been investigated in detail. In this study, we studied the changes of hair follicles induced by ischemia and potential effects of normobaric hyperoxygenation (NBO) on the hair cycle and growth. We found that skin ischemia reduced hair growth rate, hair shaft size, and its pigmentation in the anagen phase of mice, which may reflect an aspect of pathophysiology of hair loss (alopecia) and depigmentation (gray/white hairs). Hyperoxygenation increased hair growth rate in organ culture of both human and murine hair follicles. Systemic NBO promoted hair growth in early anagen and mid-anagen, and delayed catagen onset in mice. However, telogen-to-anagen transition was not affected by NBO as far as non-ischemic skin is concerned. The results of this study indicated that the hair follicle is very sensitive to oxygen tension and oxygen tension affects the regulation of hair growth and cycle in vitro and in vivo. It was suggested that systemic NBO can be safely applied for a long period and can be a noninvasive therapeutic approach to alter hair growth and cycle by manipulating the microenvironment of hair follicles.

KEYWORDS:

• alopecia
• baldness
• depilation
• hypoxia
• ischemia
• normobaric hyperoxygenation

Introduction

Alopecia, otherwise known as hair thinning or baldness, has several contributing factors including genetics, hormone metabolism disorders, trauma, infections, chemotherapy, autoimmune.¹ However, ischemia has not yet been investigated in detail as an influential factor on hair growth. The reduced blood flow that is characteristic of ischemia causes both hypoxia and undernutrition. The effects of O₂ supply on hair growth rate and the hair cycle have not been carefully examined.

Among the limited studies conducted to date, one reports that areas of alopecia are associated with ischemia and hypoxia.² In this report of androgenic alopecia cases, the partial pressure of O₂ in the forehead tissue (where hair loss was occurring) was significantly lower than that in the temporal hairy region, and the difference in partial oxygen pressure between forehead and temporal area was greater in androgenic alopecia than those exhibiting normal hair growth. This finding suggests that transcutaneous oxygen pressure was significantly lower in the frontal area than in temporal area of hair loss subjects. This may be due to the lower vascularity and/or high mechanical tension of the skin. However, how ischemia or hypoxia relates to hair follicle growth and the hair cycle has not been examined in details before.

When performing basic research on hair growth rate and the hair cycle, the most commonly used experimental animal is the C57BL/6JJcl mouse, because correlations between the hair cycle and age, the accompanying histological changes, and other details have been thoroughly investigated.^{3 4} Stimuli, signals, and tissue changes have been investigated using these mice at 7 weeks of age, which are known to be in telogen phage.⁵ Using the same strain, we investigated the effects of tissue partial pressure of O₂ on hair growth and hair cycle.

If ischemia or hypoxia can cause alopecia, then it may be possible to treat or prevent alopecia by correcting the ischemic condition or improving tissue oxygen tension. Minoxidil, an approved hair-restoring agent, has a variety of effects (e.g., increasing the proportion of anagen-phase follicles, extending the anagen phase, and increasing the rate of hair growth by promoting the proliferation of dermal papilla and hair matrix cells) that are mediated by the vasodilatory activity of minoxidil and the production of adenosine after binding to the sulfonylurea receptors of cell membrane-bound potassium channels of dermal papilla cells.⁶ Other methods of improving ischemia that have been attempted previously include local injection of growth factors or platelet-rich plasma,⁷ fat grafts, or adipose-derived stem cells; and, in recent years, improving blood flow using botulinum toxin,⁸ which showed clinical improvement of hair growth.

We sought to examine if oxygenation induces anagen and promotes hair growth.

Hyperbaric oxygenation has been clinically used for some pathological conditions, but this approach can lead to cellular dysfunction due to oxygen toxicity and cannot be applied for more than 2 hours.^9–11 The relation between oxidative stress and androgenic alopecia is also reported.¹² Therefore, rather than hyperbaric hyperoxygenation, we used normobaric hyperoxygenation (NBO) as a treatment to improve hair growth and cycles. We previously compared the partial pressure of O₂ in various organs and tissues by housing mice under 20% or 60% O₂ and found that long-term application of 60% NBO is safe and maintain partial pressure of O₂ in skin at 178% compared to room air.¹³ In this study, we sought to clarify the influence of ischemia and oxygenation on hair growth and hair cycle.

Results

Organ culture of hair follicles in different O₂ concentrations

Human scalp hair follicles were organ cultured in 1% (hypoxia), 6% (normoxia), and 20% (hyperoxia) O₂. The physiological partial O₂ tension of human subcutaneous fat tissue is around 50–60 mmHg, which corresponds to 6% O₂ at 1 atmosphere. In all groups, anagen-phase follicles continued to grow for 7 days (Figure 1a). On Day 7, the 1% O₂ hair averaged 0.7 ± 0.6 mm (n = 5), the 6% O₂ hair averaged 1.0 ± 0.5 mm (n = 5), and the 20% O₂ hair averaged 2.1 ± 0.8 mm (n = 5) in length. Similar results were also seen in hairs in mice (Figure 1b). On Day 7, growth of the 1% O₂ strands was not observed. Nonetheless, hairs maintained under 6% O₂ were 0.8 mm and those under 20% O₂ were 1.8 mm.

Figure 1. Organ culture of human hair follicles under different oxygen tensions.

(a) Human scalp hair follicles were cultured in 1%, 6%, and 20% O₂ for 7 days. Representative microscopic views of organ-cultured hair follicles (Left). Hair shafts grew in length over time up to 7 days in all samples; under 1% (left), 6% (middle), and 20% (right panel) O₂. Scale bar = 300 µm. Quantitative change in hair shaft length (Right; n = 5). Significant difference was seen between 20% and 6% (p = .046), as well as between 20% and 1% (p = .015). (b) Organ cultures of mouse whiskers (Left). Mouse whisker follicles were organ-cultured in 1%, 6%, and 20% normobaric O₂. Hair growth rate was the highest with 20% O₂ (Right).

Influence of ischemia on mouse hair follicles

After ischemic and non-ischemic skin flaps were elevated in the groin region on each side, the tissue O₂ partial pressure was measured over time and compared. Values stayed at approximately 60 mmHg on the control side, whereas the oxygen pressure reached approximately 30 mmHg on the ischemic side until 2 weeks after flap elevation and depilation for anagen induction. Blood flow recovered starting from Week 3 and stayed relatively stable until Week 5 (Figure 2a).

Figure 2. Effects of ischemia on hair follicles in mice.

(a) Sequential changes in the partial pressure of O₂ in ischemic or non-ischemic groin skin flap tissue. Until 2 weeks, the O₂ partial pressures on the control and ischemic sides were approximately 60 and 30 mmHg, respectively. (b) Typical views of four types of body hair: guard, awl, auchen, and zigzag. Scale bar = 300 µm. (c) Hair length at 2 weeks. Hair on the control side was significantly longer than that on the ischemic side for all types. (d) Number of hair follicles per section. The number of hair follicles of the control flap is significantly larger than that of the ischemia flap (p = .0084).

C57BL/6JJcl mice have four types of body hair; guard, awl, auchen, and zigzag (Figure 2b). The guard hair type is the longest and straight, awl is short and straight, auchen has one bend, and zigzag has two bends. Hairs on the ischemic side were significantly shorter than those on the control side (Figure 2c). Guard hairs were 4,096 ± 127.9 µm (n = 7) on the control side and 3,504 ± 420.8 µm (n = 10) on the ischemic side (p = .0015). Awl hairs were 3,156 ± 261.0 µm (n = 21) on the control side and 2,642 ± 249.9 µm (n = 21) on the ischemic side (p = .000009). Auchen hairs were 2,683 ± 118.9 µm on the control side (n = 15) and 2,267 ± 130.1 µm on the ischemic side (n = 20) (p = .000003). Zigzag hairs were 2,574 ± 249.4 µm (n = 7) on the control side and 1,812 ± 512.2 µm (n = 7) on the ischemic side (p = .006).

The number of anagen-phase follicles per section was significantly higher on the control side (41.8 ± 7.0, n = 8) than on the ischemic side (27.7 ± 6.2, n = 8) (p = .0084) (Figure 2d). These findings suggest that ischemia makes the hair shafts thinner, decreases hair density, delays hair growth and affects the hair cycle in mice.

Effect of NBO on hair follicles in early anagen, mid-anagen, and late-anagen phases (Figure 3)

i) NBO in early anagen phase

After a 7-day treatment of NBO starting from Day 0, hair was harvested on Day 7. Several days after anagen induction, the dorsal skin started to turn gray in color in both 20% and 60% O₂ groups. Immunohistology for Ki67 indicated that a large number of Ki67-positive hair matrix cells were confirmed in both 20% and 60% O₂ groups on Days 3 and 4 (Figure 4a). The proportion of Ki67-positive cells among all hair matrix cells in the follicles was 49.0 ± 9% in the 60% O₂ group (n = 5) and 33.9 ± 7% in the 20% O₂ group (n = 5) on Day 3, whereas the proportion was 58.3 ± 8% in the 60% O₂ group (n = 5) and 43.6 ± 10% in the 20% O₂ group (n = 5) on Day 4 (Figure 4b). On both days, there were a significantly larger number of Ki67-positive hair matrix cells in the 60% O₂ group (Day 3, p = .008; Day 4, p = .037), indicating that NBO promoted hair growth in early anagen phase.

Figure 3. Treatment protocols of systemic normobaric hyperoxygenation (NBO) for hair follicles in early, mid, and late-anagen stages.

(a) The mice were kept in 20% or 60% (NBO) O₂ for 7 days immediately after hair removal (Day 0–7), and thereafter in 20% O₂. The effects on the early anagen phase were evaluated. (b) The mice were kept in 20% O₂ for 7 days after hair removal, followed by 20% or 60% O₂ for 7 days from Day 7 (Day 7–14), then again in 20% O₂. The effects on hair growth during the mid-anagen phase were evaluated. (c) The mice were kept in 20% O₂ for 14 days after hair removal, followed by 20% or 60% O₂ for 7 days from Day 14 (Day 14–21), then again in 20% O₂. The effects on the anagen-to-catagen transition were evaluated.

Figure 4. NBO effects on hair follicles in early anagen phase.

(a) Immunostained hair follicles. Proliferation of hair matrix cells was found in both groups. Scale bar = 50 µm. (b) Proportion of Ki67-positive hair matrix cells among all hair matrix cells. The proportion was significantly higher in the 60% O₂ group on Days 3 and 4. *p < .05.

ii) NBO in mid-anagen phase

In mid-anagen, the hair length increases rapidly. After a 7-day treatment of NBO starting from Day 7, hair was harvested on Day 14. The hair length of NBO-treated mice was 7.7 ± 0.50 mm (n = 64) and significantly longer than that from 20% O₂-treated mice was 6.9 ± 0.42 mm (n = 64), indicating that NBO promotes the hair shaft growth in mid-anagen.

iii) NBO in late anagen to catagen phase

After a 7-day treatment of NBO starting from Day 14, hair was harvested on Day 21. In control mice, anagen-catagen transition occurred on Day 17 to 19. Differences in macroscopic views on either the external or subcutaneous side were not apparent. In catagen-phase follicles, apoptotic epithelial cells, which are terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL)-positive, were found in the hair matrix. On Day 19, histological sections of NBO-treated mice showed 2.8 ± 1.0 (n = 5) TUNEL-positive cells per section, which was significantly (p = .007) smaller in number compared to control mice (9.2 ± 3.1, n = 5) (Figure 5), suggesting that NBO may delay anagen-to-catagen transition and extend the anagen phase.

Figure 5. NBO effects on hair follicles in late-anagen phase.

Number of TUNEL-positive cells in hair matrices. There were significantly fewer cells in the 60% O₂ group at 19 days after hair removal. *p < .05.

Effect of NBO on telogen hair follicles

In contrast to depilation, shaving of the back hairs of 7-week-old C57BL/6JJcl mice does not induce anagen phase, and the telogen phase continues. To verify whether NBO affects the length of telogen phase, the mice were maintained under 20% or 60% O₂ for 14 weeks after shaving. In one animal in the 20% O₂ group and 2 animals in the 60% O₂ group, anagen phase was induced probably by mechanical injury at the time of shaving. Eleven animals in the 20% O₂ group and 10 animals in the 60% O₂ group were evaluated. Among animals in which the next anagen occurred between 35 and 98 days after shaving, anagen was observed after 47.3 ± 14.4 days and 45.5 ± 9.65 days in the 20% and 60% O₂ group, respectively. However, the other animals in both groups did not show anagen onset until 98 days (Figure 6). Between 7 and 21 weeks, anagen induction occurred in 8 mice at 20% O2 and 5 mice at 60% O2 with no significant difference (Figure 7).

Figure 6. Transition to the anagen phase after shaving in the 60% and 20% O₂ groups.

Seven-week-old mice were maintained under 60% or 20% O₂ after shaving to verify transition from the telogen to anagen phase. Animal images with a red margin were evaluated as anagen was induced.

Figure 7. Transition to the anagen phase after shaving in the 60% and 20% O₂ groups.

Number of animals with anagen induction. No difference was seen between systemic NBO (60% O₂) and control treatment (20% O₂).

Discussion

This study is the first to quantify how much induction of ischemia and hyper-or hypo-oxygenation may affect on hair cycle and hair shaft growth. Regenerated hairs in ischemic flaps presented not only slower rate of hair growth but also some typical features of alopecia such as thinner hair shafts and lesser hair density. It is known that hair loss frequently occurs after temporary ischemia after facelift surgery or hair transplantation though followed by hair regeneration in part.¹⁴ One study on frontal fibrosing alopecia showed increasing vascular flow in the deep plexus, while superficial tissue became ischemia or fibrosis.¹⁵ Partial arterial occlusion by hyaluronic acid injection induced transient alopecia in the ischemic area.¹⁶

We also observed depigmented hairs in some ischemic flaps of mice, which may suggest that melanocyte stem cells may be more sensitive to ischemia/hypoxia than epithelial ones. In the case of early-stage androgenic alopecia, subcutaneous blood flow to the forehead has been reported to be about 2.6 times lower than that in those exhibiting normal hair growth.¹⁷ In our study, the O₂ partial pressure of the ischemic side was about one-third of that on the control side, which may be similar to the condition of androgenic alopecia.

In organ culture, hair growth was slower under hypoxic conditions (1% O₂) when compared to physiological (6% O₂) or hyperoxic (20% O₂) environments and there was no difference between the two latter treatments. It was shown that the rate of hair growth was highly dependent on O₂ concentration. A previous study using human hair follicles organ-cultured in 20% or 95% O₂ showed that hair matrix cell degeneration was suppressed and hair matrix degeneration was reduced and structural integrity of the matrix was maintained for a longer period of time when exposed to 95% O₂.¹⁸

NBO can be easily performed clinically, while hyperbaric hyperoxygenation cannot be applied for longer than 90 min due to its oxygen toxicity.^19–23 In this study, 7-day NBO applied either in the early anagen or mid-anagen promoted matrix cell proliferation and consequently accelerated hair growth. Our results suggested that oxygen tension is a very influential microenvironment factor on hair follicles and hyperoxygenation may provide beneficial effects to promote hair growth and regulate hair cycles in vitro and in vivo.

Transition to the catagen phase is known to occur approximately 16 to 19 days after the anagen phase is induced in C57BL/6JJcl mice.³ Mice maintained under NBO for 7 days during this catagen-phase transition showed fewer TUNEL-positive cells than those maintained under 20% O₂. When transition to the catagen phase occurs, hair matrix cells are known to become apoptotic (TUNEL-positive), so this result indicates that NBO can delay the anagen-to-catagen transition.

In addition to NBO effects on anagen, we examined NBO effects on a therapeutic potential of NBO for treating hair loss by inducing telogen-to-anagen transition. As the dorsal hairs of 7-week-old C57BL/6JJcl mice are known to be in the telogen phase, systemic NBO or control treatment was started in mice models with hair shaving (not inducing anagen by depilation) since at the age of 7 weeks. Under physiological condition, the next anagen was detected 6–7 weeks later (at the age of 13–14 weeks). A long-period NBO therapy did not significantly promote the transition from the telogen to anagen phase when applied to non-ischemic telogen skin. As pathological alopecia appears to be clinically associated with ischemic/hypoxic conditions of the scalp, our results may not always indicate that NBO is ineffective for treating alopecia in human. Further analyses using animal models with ischemic/hypoxic telogen skin may help to investigate the potentials of NBO more precisely.

Some signaling pathways such as TGF-β, BMP, and Wnt pathways have been shown to regulate hair development and cycles.^18–20 Our results suggested that oxygen tension is a very influential microenvironment factor on hair follicles and hyperoxygenation may provide beneficial effects to promote hair growth and regulate hair cycles in vitro and in vivo. One study of oxygen reports suggested that oxidative stress inhibits migration, proliferation, and secretory functions of dermal papilla cells from bald skin and may be related pathogenesis of alopecia.²⁴

In conclusion, our findings suggested that ischemia and/or low skin pO₂ reduces hair fiber growth rate, pigmentation and may contribute to the onset of catagen. These effects could contribute to alopecia as well as hair graying that occurs naturally with aging. Hyperoxygenation treatments increased hair fiber growth during anagen and delayed the onset of catagen. These findings support the hypothesis that ischemia and/or the resulting reduction in skin pO₂ and nutrient supply may contribute to impaired hair growth and hair loss.

Materials and methods

All experimental protocols involving human subjects in this study were performed in accordance with relevant government guidelines and approved by the institutional review board (IRB) of the University of Tokyo Hospital. Prior to the procedure, each patient provided her/his informed consent using an IRB approved protocol. The protocols of animal experiments followed the relevant guidelines and were approved by the animal experimental committee of the University of Tokyo School of Medicine.

Inguinal ischemic mouse model

Nine-week-old C57BL/6JJcl mice (male) were anesthetized by intraperitoneal injection of somnopentyl (Kyoritsu Seiyaku, Tokyo, Japan). A minimal amount of anesthesia was administered to avoid deep anesthesia and thereby ensure that the O2 partial pressure in the tissue did not decrease. The groin area was depilated using depilating tape (Veet, Reckitt Benckiser, United Kingdom), and a 1.5 × 1.5 cm skin flap was raised. To create an ischemic environment for the skin around the hair follicles in the skin flaps, the femoral artery was ligated and dissected proximal to the inferior epigastric artery so that blood supply to the skin flaps came only from the reverse flow of the femoral artery. The contralateral side served as the control. Both groin areas were monitored, and after 2 weeks, the body hair on the skin flaps was pulled out. The length of the hair was measured, and the skin flaps were evaluated histologically. In addition to hair and skin flap sample collection, a probe (200 µm in diameter) was inserted directly into the fat in the groin to measure O₂ partial pressure over time using an O₂ partial pressure monitor (Eiko Kagaku, Tokyo, Japan).

Organ cultures of mouse whiskers and human hair

Six-week-old C57BL/6JJcl mice (male) were placed under deep anesthesia and euthanized. Skin samples including whisker follicles were removed from the face, and only anagen-phase follicles that penetrated the subcutaneous tissue were isolated using a microscope.²⁵ Sections of human scalp containing hair that was to be discarded following facelift operations were used. Hair follicles that penetrated the subcutaneous tissue were isolated from the skin sample. The isolated mouse whiskers (n = 3) and human hair follicles (n = 5) were placed in a 24-well plate and immersed in 1 ml of Williams’ medium E (Gibco®, Life Technologies, Hamburg, Germany) to which 10 µg/ml insulin, 0.4 ng/ml hydrocortisone, 2.0 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin had been added.^26–28 The length in which the hair shaft extended was measured every two days. The hair follicles were randomly divided into three groups to be cultured under 1% (n = 5), 6% (n = 5), and 20% (n = 5) normobaric O₂, and hair growth rate was subsequently measured. To reproduce the physiological oxygen condition (40–60 mmHg), the oxygen content inside the incubator was set at 6% (= 46–50 mmHg).

Immunohistochemistry

The hair follicles were placed in OCT compound (Sakura Finetek, Tokyo, Japan), frozen with liquid nitrogen, and stored at −80°C until sectioning. Using a standard method, 10-µm thick, frozen, serial sections were cut and mounted on glass slides. Before staining, the samples were air dried at room temperature for 1 h, fixed with 4% paraformaldehyde, and washed with PBS for 5 min. All sections were then stained with HE using a standard method. The sections that underwent immunohistochemistry staining, were first incubated at room temperature for 30 min in 5% fetal bovine serum (FBS) to block nonspecific binding sites. Rabbit anti-human Ki67 (clone SP6, 1:200 dilution; Thermo Fisher Scientific, Fremont, CA) was used as the primary antibody to detect proliferating keratinocytes. Alexa Fluor 594-conjugated donkey anti-rabbit IgG (dilution 1:200; Invitrogen) was used as the second antibody. An isotope antibody was used as the negative control. Additionally, TUNEL was used to evaluate apoptotic cells using an In-Situ Cell Death Detection Kit (Roche Diagnostics, Mannheim, Germany). Hoechst 33342 was used as the nuclear stain.

In vivo experiments for NBO effects on the hair cycle

Seven-week-old C57BL/6JJcl mice (male) were anesthetized by intraperitoneal injection of somnopentyl, and hair was pulled out from approximately 2 × 2 cm dorsal sections. This age of mice was chosen because the dorsal hair of 7-week-old C57BL/6JJcl mice is in the telogen phase, and pulling out hair in the telogen phase is known to induce the anagen phase of hair follicles.² In the first experiment (Figure 3a), immediately after the hair was depilated, mice were kept in a 20% O₂ environment (20% O₂ group) or a 60% O₂ environment (60% O₂ group) for 7 days, which is equivalent to the early anagen phase. On Day 7, hair growth was compared between groups. In the second one (Figure 3b), to verify the effects of NBO on growing hair, mice were kept in a 20% O₂ environment (20% O₂ group) or a 60% O₂ environment (60% O₂ group) for 7 days from Day 7 (starting 1 week after hair removal), which is equivalent to the mid-anagen phase. On Day 14, hair growth was evaluated. Mice are known to have four variations of body hair (i.e., guard, awl, auchen, and zigzag),^{29 30} among which the guard type is the thickest and longest. To evaluate body hair length, strands of guard hair were selected from those randomly pulled out, and their length was measured under a microscope. In the third one (Figure 3c), we verified whether administering 60% O₂ caused a delay in transition to the catagen phase thereby producing an extended anagen phase. To do so, mice were kept in a 20% O₂ environment (20% O₂ group) or a 60% O₂ environment (60% O₂ group) for 7 days from Day 14 (starting at 2 weeks after hair removal), which is equivalent to the late-anagen phase, through the catagen phase, to the telogen phase. On Day 21, hair growth was evaluated. When moving from the anagen to the catagen phase, TUNEL-positive cells in the hair matrix showing apoptosis were confirmed. Histological evaluations were performed to determine whether 60% O₂ showed any effect. In all groups, the animals were anesthetized during sample collection and euthanized. Samples containing dorsal skin and subcutaneous tissue were taken en bloc.

In vivo experiment for NBO effects on physiological telogen-to-anagen transition

Seven-week-old C57BL/6JJcl mice (male) were anesthetized by intraperitoneal injection of somnopentyl, and 2.5 × 4.5 cm sections on the dorsa were shaved. Pulling hair out during the telogen phase induces the anagen phase, but anagen induction does not occur with shaving. After being shaved, the mice were randomly assigned to the 60% O₂ or 20% O₂ group and maintained for 8 weeks. Changes in the dorsal hair were evaluated every 7 days.

Statistical analyses

Data are presented as mean ± standard deviation, and p < .05 (two-tailed test) was considered a statistically significant difference. The ischemic groin models were compared using a paired t-test. The effects of NBO on the hair cycle, hair growth, and physiological anagen-phase induction were examined using the Mann–Whitney U test. A multiple comparison test (Tukey–Kramer method) was used for the organ cultures.

Competing financial interests

There are no competing financial interests to declare.

Acknowledgments

There were no grants, equipment, or research support.