The FASEB Journal 2006 – Towards a “free radical theory of graying” …

The FASEB Journal 2006;20:1567-1569.

Towards a “free radical theory of graying”: melanocyte apoptosis in the aging human hair follicle is an indicator of oxidative stress induced tissue damage

Petra Clara Arck†,1, Rupert Overall*,1, Katharina Spatz*, Christiane Liezman*, Bori Handjiski†, Burghard F. Klapp‡, Mark A. Birch-Machin§ and Eva Milena Johanne Peters*,1

* Cutaneous Neuroimmunology, Biomedical Research Center, University Medicine Charité, Virchow and Mitte Campus;

† Psychoneuroimmunology, Biomedical Research Center, University Medicine Charité, Virchow Campus; and

‡ Internal Medicine, Psychosomatics, University Medicine Charité, Mitte Campus, Humboldt-University of Berlin, Berlin, Germany; and

§ Dermatology, Medical School, University of Newcastle, UK

1Correspondence: Biomedical Research Center, Rm. Nr. 2.0549, University Medicine Charité, Virchow Campus, Humboldt University of Berlin, Augustenburger Platz 1, Berlin 13353, Germany. E-mail: or


Oxidative stress is generated by a multitude of environmental and endogenous challenges such as radiation, inflammation, or psychoemotional stress. It also speeds the aging process. Graying is a prominent but little understood feature of aging.

Intriguingly, the continuous melanin synthesis in the growing (anagen) hair follicle generates high oxidative stress. We therefore hypothesize that hair bulb melanocytes are especially susceptible to free radical-induced aging. To test this hypothesis, we subjected human scalp skin anagen hair follicles from graying individuals to macroscopic and immunohistomorphometric analysis and organ culture. We found evidence of melanocyte apoptosis and increased oxidative stress in the pigmentary unit of graying hair follicles.

The “common” deletion, a marker mitochondrial DNA-deletion for accumulating oxidative stress damage, occurred most prominently in graying hair follicles. Cultured unpigmented hair follicles grew better than pigmented follicles of the same donors.

Finally, cultured pigmented hair follicles exposed to exogenous oxidative stress (hydroquinone) showed increased melanocyte apoptosis in the hair bulb.

We conclude that oxidative stress is high in hair follicle melanocytes and leads to their selective premature aging and apoptosis. The graying hair follicle, therefore, offers a unique model system to study oxidative stress and aging and to test antiaging therapeutics in their ability to slow down or even stop this process.

—Arck, P. C., Overall, R., Spatz, K., Liezman, C., Handjiski, B., Klapp, B. F., Birch-Machin, M. A., Peters, E. M. J. Towards a “free radical theory of graying”: melanocyte apoptosis in the aging human hair follicle is an indicator of oxidative stress induced tissue damage.

Key Words: gray hair • premature aging

Full Text:

IT IS OBVIOUSthat the color of our hair has important socio-economic implications and propels a multimillion hair product industry. Common and premature graying (canities) are common and intriguing phenomena frequently discussed in the context of environmental or endogenous processes that lead to an accelerated aging process such as pollution, UV-exposure, inflammation, or even psychoemotional stress (6 , 28 , 31 , 48 , 58) . These and other challenges can be viewed as stressors that affect our health by generating free radicals, which alter proteins and nucleic acids important in the maintenance of our biological functions.

Clinical observation states that in senile canities by the age of 50, 50% of all hair follicles in 50% of all men have lost their pigment (see Ref 1 ). Premature canities is much more striking and often lead to rapid graying of rather young individuals. This dramatic change in physical appearance is frequently used in the arts as well as medical literature to demonstrate and witness severe disease, rapid aging, psychoemotional stress experiences, and radical changes in life (see ref 2 3 4 ). However, this condition may also serve as an indicator of biological response to a stressor. However, the process of graying and its link to environmental and endogenous stressors is seldomly studied in a scientific context.

The biological process appears to be associated with a progressive loss of the pigment-producing cells, the melanocytes, from the aging hair bulb and outer root sheath with physiological aging and premature aging syndromes (1 , 5 6 7) . To date, the cause of this loss has not been satisfactorily explained, though a number of interesting hypothesis have been brought forward (1 , 6 7 8 , 10 ). From animal experiments and mutation-reports, we know about predetermined braking points in the function of the hair follicle pigmentary unit, which is formed between the pigment-producing melanocytes, anchored to the basement membrane above the dermal papilla, and the hair shaft producing keratinocytes in the hair bulb (c.f. 1 , 6 , 9 ).

Pigmentation braking points include the exhaustion of growth factors and melanogenesis-related enzymes, impairment of DNA-repair mechanisms, loss of antioxidative enzymes and cofactors, loss of telomerase, or loss of antiapoptotic signals (1 , 5 , 6 , 9 10 11 12) . One prominent hypothesis features decreased stem cell factor (SCF) signaling through its tyrosinekinase receptor c-Kit, since anti-hypothesis c-Kit treatment can produce gray hair follicles (68) . Another focuses on Bcl-2, which scavenges oxygen radicals in the membranes of mitochondria, because its loss by knockout leads to premature graying of mice that otherwise lack a striking phenotype (10) . Analysis of the fate of melanocytes during the murine hair cycle revealed that melanocytes can be lost from the hair bulb by apoptosis during hair follicle involution (catagen) (13 , 14) . Ultimately, this may lead to the exhaustion of a melanocyte stem cell pool residing in the bulge region of the hair follicle (12 , 15) . However, conclusive proof that any of the above braking points of melanocyte function are relevant to common graying has not been brought forward to date, and it remains open whether apoptosis occurs in aging hair follicle melanocytes and what might cause it.

Oxidative stress appears to play a key role in many of the indicated pathways (6) . Accordingly, an anecdotal report has it that melanocytes in the graying human hair bulb show vacuoles prior to their loss, as an indicator of increased oxidative stress (Westerhof in 16 ). In this context, note that melanin-synthesis by itself generates cellular oxidative stress (17) . Several of the steps in melanin production yield H2O2 (18 , 19) and other free radicals (20 , 21) . This places melanocytes under a higher oxidative stress load than for example keratinocytes, which is most prominent in hair follicle melanocytes because they produce large quantities of melanin constitutively throughout the anagen phase of the hair cycle. Oxidative stress generated outside hair follicle melanocytes, e.g., by UV-light induced (1 , 46 , 56) , psychoemotional (22 , 23) , or inflammatory (24 , 25) stress may add to this endogenous oxidative stress, overwhelm the hair follicle melanocyte antioxidant capacity, and speed up terminal damage accumulating for example in the aging hair follicle.

Widely used markers to determine the degree of cellular oxidative damage are mitochondrial DNA deletions (26 27 28 29) . Deletions such as the so-called “common” deletion lead to defects in the respiratory chain, more oxidative stress, hydrops, and ultimately apoptosis of the damaged cells (30 , 31) . In theoretical concepts of aging, this chain of events is known as the “free radical theory of aging” (32) ; in analogy, we would like to propose a “free radical theory of graying”. Along this concept we attempt to answer the following questions:

1) Are decreased numbers, morphology and melanocyte apoptosis in the hair follicles of aging individuals associated with oxidative stress in the pigmentary-unit?

2) Are melanocytes in graying hair follicles protected from oxidative stress by endogenous scavengers such as Bcl-2 and supported by growth factors such as SCF?

3) Do graying hair follicles have higher levels of oxidative stress-induced permanent damage (mitochondrial DNA damage)?

4) Is exogenous oxidative stress leading to apoptosis selectively in hair follicle melanocytes?

To address these questions we performed morphological, immunohistochemical, and rtPCR analysis of pigmented, graying, and unpigmented human scalp skin anagen hair follicles to determine oxidative stress in the hair follicle pigmentary unit. We then used an in vitro culture system (33) for human scalp skin anagen hair follicles to analyze viability of pigmented vs. unpigmented hair follicles and the response of the hair follicle pigmentary unit to exogenous oxidative stress as an in vitro model of graying.


Scalp skin samples
Temporal scalp skin containing mainly anagen VI hair follicles from disposed excess skin samples of patients undergoing elective plastic surgery was obtained with informed consent. The study was conducted according to Declaration of Helsinki Principles. Available donor samples included in this study were derived from female donors between 50–72 years of age and past menopause.

After excision, tissue was maintained in Williams E Medium (Biochrom KG seromed, Berlin, Germany) for transportation at 4°C up to 24 h. On arrival, the samples were divided by two: One part was immediately snap-frozen in liquid nitrogen for immunohistochemical analysis, the second was processed for rtPCR analysis and hair follicle culturing as described below.

Hair follicle isolation
Single human anagen VI hair follicles (34 , 35) were obtained by microdissection from human skin biopsies following, with slight modifications, the protocol published by Philpot and colleagues (33 , 36) . Briefly, after separation of epidermis and dermis from subcutaneous (s.c.) fat, the dermis just above the dermis/subcutis border under a binocular dissecting microscope, the proximal two-thirds of anagen hair follicles located in the s.c. fat were isolated using watchmakers forceps and subsequently were collected in Petri dishes containing complete hair follicle culture medium (Williams E, Biochrom KG seromed); 1% penicillin-streptomycin (Life Technologies, Eggenstein, Germany); 1% L-glutamine 200 mM (Life Technologies); 0.02% hydrocortisone (Sigma, Taufkirchen, Germany); and 0.1% Insulin (Sigma).

Microdissected hair follicles consisted of the hair bulb and the distal hair follicle, up to the level where the arrector pili muscle inserts into the bulge region (including the hair shaft, the inner and outer root sheath, the connective tissue sheath, and the dermal papilla). This part of the hair follicle is constantly reconstructed during the hair-growth cycle and contains the dermal papilla, formed between the pigment producing melanocytes above the dermal papilla and the hair shaft producing keratinocytes (the pigmentary-unit, compare Fig. 1 ), and the proximal part of the hair shaft, inner root sheath, and outer root sheath (33) . It does not contain the constant part of the hair follicle that consists of the stem cell niche in the bulge region of the distal outer root sheath, the arrector pili muscle, the sebaceous gland, and the hair follicle ostium (33) .

View larger version (101K):
In this window
In a new window

Figure 1. Melanocytes disappear gradually and undergo apoptosis in the graying hair bulb. Graying differed morphologically from depigmentation in catagen hair follicles (a). Scalp skin biopsies from graying individuals contained hair follicles with varying degrees of depigmentation (f–o). Pigmentation status ranged from full pigmentation (b–e) to almost complete depigmentation (l–o). Schematic representations (c, h, m) summarize macroscopic and immunhistochemical findings and show localization of the dermal papilla, inner and outer root sheath, and the differentiated melanocytes above Auber’s line (dotted lines). These melanocytes are attached to the basement membrane to transfer their pigment to the hair shaft producing keratinocytes, forming the pigmentary-unit (gray areas). Hair bulbs may display any number of melanocytes in the pigmentary-unit within the hair matrix as can be observed macroscopically on isolated hair follicles (b, f, g, k, l) or by TUNEL (green label)/NKI-beteb (red label) double-labeling on full-thickness scalp skin biopsies (d, e, i, j, n, o). Nuclei are counterstained with DAPI (blue fluorescence). Abbreviations: bm, basement membrane; cts, connective tissue sheath; dp, dermal papilla; em, ectopic differenting melanocyte; es, epithelial strand; hc, hair canal; hm, hair matrix; hs, hair shaft; irs, inner root sheath; m, differentiated dendritic melanocyte in the pigmentary-unit; ors, outer root sheath; pu, pigmentary-unit; rm, rounded melanocyte dissociating from basement membrane. a) A catagen hair bulbs can be easily distinguished from graying anagen hair bulbs by their characteristic morphology (compare b–o). Melanocytes in the catagen hair bulb appear densely pigmented and rounded. They have no dendrites and are arranged around the condensed dermal papilla. These features resemble melanocytes in graying hair bulbs. However, melanocytes in catagen hair follicles are more pigmented and less dendritic and no oligodendritic lightly pigmented melanocytes can be detected (compare f–o). Note that the pigmentary-unit has disintegrated and no longer exists and that the hair shaft production has been discontinued, while an unpigmented epithelial strand forms between hair shaft and dermal papilla. b) In a fully pigmented hair bulb, no pigment dilution can be observed in the hair shaft and the dermal papilla is completely hidden by pigment. Individual melanocytes cannot be distinguished in the pigmentary-unit. c) Schematic representation of a fully pigmented anagen VI human hair bulb. Note that all melanocytes in the pigmentary-unit in the center of the hair matrix are in close contact with the basement membrane separating the hair follicle epithelium from the dermal papilla. d) A fully pigmented anagen hair follicle from a scalp skin biopsies of a graying individual displays numerous NKI-beteb+ melanocytes in the outer root sheath (insert). e) A fully pigmented anagen hair follicle from a scalp skin biopsies of a graying individual displays highly dendritic NKI-beteb+ melanocytes (inserts) in the hair bulb, located in the pigmentary-unit above Auber’s line (dotted white line). f, g) The hair shafts of two graying hair bulbs display pigment dilution and allow to distinguish individual keratinocytes packed with melanosomes, giving the hair shaft a spotted appearance. Melanocytes move away from the dermal papilla and appear rounded, oligodendritic, and stubby. Individual lightly pigmented oligodendritic melanocytes occur below Auber’s line (arrows). h) Schematic representation of a graying anagen VI human hair bulb. i, j) In a graying hair follicle, numerous TUNEL+ apoptotic melanocytes can be detected in the pigmentary-unit (i and inserts). Melanocytes are also detectable below Auber’s line (j). k, l) In two hair bulbs, which do not produce pigmented hair shafts, melanocytes appear triangular rather than dendritic, and small rounded pigmented fragments can be observed. In addition, single dendritic melanocytes with very light pigmentation are located in the most proximal hair bulb adjacent to the basement membrane surrounding the hair follicle epithelium and separating it from the connective tissue sheath. The low number of pigmented melanocytes in the pigmentary-unit allows full appreciation of the tear-shaped dermal papilla. m) Schematic representation of an almost depigmented anagen VI human hair bulb. n) In a graying hair follicle a single NKI-beteb+ melanocyte can be detected in the outer root sheath (insert). o) In the hair bulb of a graying hair follicle the number and dendricity of melanocytes is dramatically reduced (inserts), melanocytes can be detected below Auber’s line, and a single dendritic melanocyte is located in the most proximal hair bulb adjacent to the basement membrane surrounding the hair follicle (insert). TUNEL+ cell nuclei are also detectable in the terminally differentiating inner root sheath and serve as internal positive control (arrowheads).

Determination of pigmentation status
With the help of an inverted microscope, floating microdissected hair follicles were classified macroscopically as hair follicles, which produced a pigmented hair shaft when the dermal papilla was hidden in the strongly pigmented pigmentary unit and no pigment-dilution was evident in the hair shaft (Fig. 1) . Hair follicles were classified as graying when individual melanocytes could be distinguished in the pigmentary-unit and pigment dilution was evident in the hair shaft (Fig. 1) . Hair follicles were classified as unpigmented when no melanocytes could be detected macroscopically and the hair shaft appeared unpigmented (Fig. 1) . For simplification, hair follicles will be termed pigmented, graying, or unpigmented hair follicles throughout the manuscript, bearing in mind that pigmentation occurs only in the pigmentary-unit and hair shaft.

Only anagen VI hair follicles were included in the study (compare Fig. 1 ) (34 , 35) . Pigmentation status was photodocumented with the help of a digital camera attached to the ocular of the microscope (Nikon Coolpix 990, Torrance, CA). Hair follicles were then either transferred to a fresh 24-well tissue culture dish for culture or to an Eppendorf tube for polymerase chain reaction (PCR) analysis. For transfer, watchmaker forceps where used and great care was applied to touch the hair follicle only at the distal outer root sheath.

From each donor, the number of isolated pigmented, graying, and unpigmented hair follicles was determined, and a mean percentage of pigmented, graying, and unpigmented hair follicles in relation to total isolated hair follicle number as calculated (Table 1 ). We also assessed presence of melanocytes in the bulge region and morphology and distribution of melanocytes in the hair bulb (for details, see Table 2 ).

View this table:
In this window
In a new window

Table 1. Hair follicles in all states of pigmentation could be isolated from aging donors

View this table:
In this window
In a new window

Table 2. Melanocytes disappear gradually from the graying hair follicle

Immunohistochemistry I: TUNEL-assay
For the analysis of TUNEL+ melanocytes, cryostat sections of whole-mount scalp skin samples or of cultured hair follicles were processed for TUNEL-labeling (apoptotic cell nuclei) according to previously published protocols (37) adapted for human antigens. The TUNEL assay detects DNA fragmentation by enzymatic labeling of free 3‘-OH ends according to the manufacturer’s instructions (Apopdetect Fluorescein, QBiogene, Heidelberg, Germany). Subsequently, sections were incubated with mouse anti-human NKI-beteb antiserum, a pan-melanocyte marker (Anti-Human NKI-beteb antigen, monoclonal, mouse, Sanbio Biotechnology Research Products, Beutelsbach, Germany). Labeling was detected by a Rhodamine-conjugated goat antimouse secondary antibody (Ab). Cell nuclei were counterstained with 4‘,6-diamidin-2‘-phenylindol-dihydrochloride (4′,6′-diamidino-2-phenylidole (DAPI), Boehringer Mannheim, Mannheim, Germany) with a concentration of 1 µg/ml (1 min, room temperature). For negative controls, the terminal deoxynucleotidyl transferase-enzyme or primary Ab step was omitted. Finally, hair bulbs were photodocumented at 200x magnification with a Zeiss Axioscope 2 fluorescence microscope (Zeiss, Göttingen, Germany).

The number of samples derived from different donors of a whole-mount scalp skin biopsy that contained hair follicles with TUNEL+ melanocytes was determined. Also, for each donor the total number of hair follicles that allowed full appreciation of a longitudinal-sectioned dermal papilla and pigmentary-unit was determined as well as the number of such hair follicles with TUNEL+ melanocytes.

For cultured hair follicles, TUNEL+ melanocytes were counted per individual hair follicle and a mean number per group was calculated. The following tissue compartments were analyzed: the pigmentary unit, the proximal hair bulb, and the outer root sheath that allowed full appreciation of a longitudinal-sectioned dermal papilla and pigmentary unit. Since this is very difficult to achieve, the total number of hair bulbs analyzed was restricted to 3–12 hair bulbs per group.

Immunohistochemistry II: oxidative stress assays and growth factor analysis
For analysis of oxidative stress protection, oxidative stress damage, and growth factor support in the hair follicle pigmentary unit, we performed an immunohistochemical analysis of Bcl-2 (scavenger in mitochondrial membrane), 8-hydroxyguanosine (8OhDG) (oxidatively damaged DNA base), and c-Kit (tyrosinekinase receptor for SCF), adapting established staining protocols (38) . Briefly, 10-µm-thick cryosections fixed in acetone (at –20°C, 10 min) were preincubated with 10% normal goat serum (20 min, room temperature) and then incubated overnight at room temperature with a monoclonal primary Ab to Bcl-2 (BD Biosciences, San Diego, CA), 8OHdG (goat polyclonal, Acris Antibodies, Hiddenhausen, Germany), or c-Kit (rabbit polyclonal, DAKO, Carpinteria, CA) supplemented with 2% normal goat serum. This procedure was followed by an incubation of 60 min at room temperature with tetramethylrhodamine-isothiocyanate (TRITC)-conjugated F(ab)2 fragments of guinea pig anti-goat IgG (Dianova, Hamburg, Germany) diluted 1:200 in tris-buffered saline. Incubation steps were interspersed by three washes with tris-buffered saline (5 min each). Then sections were stained with DAPI for identification of cell nuclei and NKI-beteb for identification of melanocytes. For negative controls, slides were incubated with the secondary Ab alone.

Hair follicle culture
Per well, three pigmented or unpigmented hair follicles were randomly distributed and cultured in Costar® 24-well plates (Corning Life Sciences, Wiesbaden, Germany) containing 500 µl of complete hair follicle culture medium per well. Per experiment, a minimum of three wells (containing a total of nine hair follicles) was assigned to each group and were either left untreated to analyze growth rates of pigmented vs. unpigmented hair follicles or supplemented with different concentrations of hydroquinone (Sigma). Each experiment was repeated with hair follicles from different donors (20 donors for hair growth analysis, 4 donors for hydroquinone analysis).

Every second day, each well was photodocumented, the total length of each hair follicle was measured, medium was replaced, and fresh supplements were added. After 8 (hair growth analysis) or 3 (hydroquinone analysis) days hair follicles were snap-frozen in a drop of OCT Cryochrome (Shandon, Pittsburgh, PA) and stored at –80°C until cryosectioning.

PCR analysis of the “common” deletion
To detect mitochondrial oxidative stress damage, we performed a PCR-based assay that applied primers to a region of mitochondrial genome unaffected by rearrangements as positive control and to the regions flanking the mutation site of a 4977 bp DNA fragment deletion known as the “common” deletion. Numbering is given as in undeleted mitochondrial DNA region, L3108 and H3717; deleted mitochondrial DNA by standard PCR, L8282, and H13851, forward primer: 5′-ccc ctc tag agc cca ctg taa agc-3′, reverse primer: 5′ gtt gag gtc tag ggc tg tta-3′; (39) internal primers to amplify deleted DNA in a “nested” PCR in conjunction with the “Barron” primers as control of the standard PCR, L8377, and H13484.

As the name suggests, the “common” deletion is the most commonly found DNA rearrangement in the mitochondrial genome. Flanked by two direct repeats, the “common” deletion removes 4977 bp from the mitochondrial chromosome, including DNA coding for two components of the mitochondrial oxidative phosphorylation (OXPHOS) system, thereby initiating the breakdown of the respiratory chain with subsequent cell hydrops and death. This deletion has been shown to be a suitable marker for the overall degree of DNA damage (40) .

Twenty different donors were included into the study, each donating pigmented, graying, and unpigmented hair follicles. For pigmented hair follicles, a total of 51, for graying of 54, and for unpigmented of 48 hair follicles were analyzed. The PCR cycle used contains short extension times (less then 2 min) that do not allow amplification of the full-length, undeleted product (5.5 kb) and so selectively only amplify deleted DNA (592 bp).

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
An automatic insert of some related ads:

Thanks for your patronage. Article continues below:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

DNA was extracted as follows. The follicles were digested in proteinase K buffer (500 mM Tris, pH 8.5; 1 mM EDTA; 0.5% Tween-20; 200 µg proteinase K) for 16 h at 37°C and then ground using a plastic pestle. Another 50 µg proteinase K was added and incubated for 2 h. The DNA was then extracted using phenol/chloroform, the supernatant ethanol was precipitated, and the DNA was resuspended in a vol of 50 µl.

The above DNA preparation (1 µl) was amplified using the total mtDNA or the Δ4977 primer sets described above in a 25 µl reaction containing 1.5 mM MgCl2, 0.2 mM dNTP, 15 pmol of each primer, and 0.5 U Amplitaq Gold DNA polymerase (Perkin-Elmer, Rodgau – Jügesheim, Germany). Reactions were run using the following cycles. For the Total mtDNA primers: 12 min 94°C (45 s at 94°C, 30 s at 51°C, 1 min 72°C) for 34 cycles and a final extension of 8 min at 72°C. For the Δ4977 primers: 12 min 94°C (45 s at 94°C, 30 s at 51°C, 1 min 72°C) for 65 cycles and a final extension of 8 min at 72°C. Half of the reaction was visualized by electrophoresis in a 1% agarose gel containing EtBr.

TUNEL+ melanocyte numbers were pooled per group for each experiment. The results among different experiments using samples from different donors were highly comparable. Thus, the mean scores and numbers of TUNEL+ melanocytes per group and from all experiments were then pooled again and statistical differences between groups were determined by Mann-Whitney U test for unpaired samples.


Donor samples older than 50 years all contained graying hair follicles
We were able to isolate varying numbers of pigmented and unpigmented hair follicles (Fig. 1) and graying hair follicles with different grades of depigmentation from each donor (Tables 1 and 2) . The number of unpigmented hair follicles was lower (Table 1) than the average number of unpigmented hairs reported in donors aged 50 and older in the literature [c.f. (1) ]. In our experience microdissection selects for pigmented hair follicles, since the unpigmented hair follicles are difficult to locate at the dermis-subcutis border during isolation. Also, many hair follicles that produce a seemingly white hair still retain some melanocytes in their hair bulb and are therefore classified as graying under the dissection microscope.

Hair bulb melanocytes in the graying hair follicle display features of oxidatively stressed and apoptotic cells
In the search of morphological hints to pathways involved in graying, we first analyzed number, appearance, and distribution of melanocytes in graying hair follicles from aged donors. In contrast to fully pigmented anagen hair bulbs (Fig. 1B, C, E ), melanocytes in graying hair bulbs (Fig. 1F-O ) were reduced in number and showed features of oxidative stress and apoptosis by morphometric analysis (Fig. 1 , Table 2 ). They appeared rounded and oligodendritic and moved away from their designated site, the basement membrane, which separates the pigmentary unit of the hair bulb from the mesenchymal pace-maker of hair growth, the dermal papilla (13) . This process appeared to be progressive because the fewer melanocytes we detected above Auber’s line, the more rounded and less dendritic was their appearance (Table 2) (1) .

Outside the pigmentary unit, melanocytes were frequently detectable in the bulge region of the outer root sheath of pigmented hair follicles, where melanocytic stem cells reside (Fig. 1D ; Table 2 ). In graying or unpigmented hair follicles of the same donors, melanocytes were only rarely detected in this location (Fig. 1N ; Table 2 ). Additional single, lightly pigmented oligodendritic melanocytes became detectable in the proximal hair bulb below Auber’s line. In this compartment, unpigmented and undifferentiated putative melanocyte stem cells, but not pigmented differentiated melanocytes, are normally found. Interestingly however, comparable ectopic differentiation of melanocytes has been reported before in the graying hair follicle bulge region (Fig. 1G, H, K, L, M, O ; Table 2 ) (12) . In addition to the reduced number of melanocytes in the pigmentary unit, this finding indicates a diminished and prematurely differentiated stem cell pool selectively in graying and unpigmented hair follicles.

By immunohistochemical analysis, hair bulbs with an intermediate number of melanocytes in their pigmentary unit occasionally displayed TUNEL-positive labeling of pigment-producing melanocytes (Fig. 1I and J , Table 2 ). This finding demonstrates the loss of melanocytes from the graying hair bulb by apoptosis. Such apoptotic melanocytes were never seen in fully pigmented hair bulbs (Fig. 1E ) or in melanocytes outside the pigmentary unit (Fig. 1D, E, J, N, O ).

Immunohistochemistry confirms presence of oxidative stress in the pigmentary unit of graying individuals
By immunohistochemistry, the oxidatively damaged oligonucleotide 8OHdG could be detected in the pigmentary unit of anagen hair follicles in the scalp skin of graying individuals (Fig. 2 ). The strongest expression localized to the melanocytes (Fig. 2) was in the pigmentary unit. Strong expression was also detected in the terminally differentiating inner root sheath keratinocytes (not shown), a compartment devoid of melanocytes and characterized by the most rapid terminal differentiation within the hair follicle. These observations indicate high oxidative stress selectively in these compartments.

View larger version (47K):
In this window
In a new window

Figure 2. Oxidative stress in the pigmentary-unit is high. Immunohistochemistry for 8OHdG reveals strong expression in the pigmentary-unit and melanocytes of a hair bulb in a graying individual (b and c). Cell nuclei are counterstained with DAPI (a and c). Arrows point at melanocytes in the individual (a and b) and merged photomicrographs (c).

Hair follicles in graying individuals lack oxidative stress protection by Bcl-2 and growth factor support by c-Kit-signaling
Immunohistochemical analysis of full mount scalp skin biopsies of graying individuals revealed that Bcl-2 expression was weakly present in the melanocytes of the pigmentary unit of fully pigmented hair follicles (Fig. 3 ai). By contrast, melanocytes in the pigmentary unit of graying and almost white hair follicles lacked this expression (Fig. 3bi, ci ). Bcl-2 expression in melanocytes of graying and catagen hair bulbs was located to single, oligodendritic, and immature appearing melanocytes outside the pigmentary unit (Fig. 3: bi, d ). As an intrinsic positive control, Bcl-2 expression was always found in melanocytes of the outer root sheath and the epidermis (Fig. 3e, f ). Interestingly, hydrochinone treatment failed to induce Bcl-2 expression in hair bulb melanocytes (Fig. 3g ).

View larger version (66K):
In this window
In a new window

Figure 3. Oxidative stress protection and growth factor support is down-regulated in hair follicles from graying individuals. Abbreviations: bm, basement membrane; D, dermis; dp, dermal papilla; e, epidermis; em, ectopic differenting melanocyte; hs, hair shaft; irs, inner root sheath; m, differentiated dendritic melanocyte in the pigmentary-unit; ors, outer root sheath; pu, pigmentary-unit; rm, rounded melanocyte dissociating from basement membrane. Schematic representations of hair bulbs in graying individuals depict a fully pigmented hair bulb (a), a graying hair bulb (b), and an almost white hair bulb (c). Photomicrographs that correspond with the respective schematic have been labeled with the same letter and (i) for Bcl-2 immunolabeled bulbs or (ii) for c-Kit immunolabeled bulbs. Each photomicrograph is shown in triplicate showing the DAPI staining for cell nuclei on the left, the NKI-beteb staining for melanocytes in the center and the specific staining for Bcl-2 or c-Kit, respectively, on the right as indicated above the photomicrographs. Higher magnifications of individual regions of interest are indicated by white boxes. All photomicrograhs were taken from immunolabeled sections from full-thickness skin biopsies from graying individuals with the exception of (g) and (k), which were taken from pigmented anagen scalp skin hair follicles treated with hydrochinone as described in Materials and Methods. ai) Bcl-2 immunoreactivity is weak but readily detectable in melanocytes of the pigmentary-unit of a fully pigmented, nongraying hair bulb. bi) Bcl-2 immunoreactivity is absent from melanocytes of the pigmentary-unit of a graying hair bulb. By contrast, a single oligodendritic and weakly NKI-beteb+ melanocyte outside the pigmentary-unit is Bcl-2+. ci) Bcl-2 immunoreactivity is absent from an almost white hair bulb showing a single, rounded melanocyte dissociated from the basement membrane in the pigmentary-unit (arrow). d) Single, strongly NKI-beteb+ polydendritic melanocytes in a catagen hair follicle retain Bcl-2-expression. e, f) Dendritic melanocytes in the epidermis (e) and outer root sheath (f) are strongly labeled with Bcl-2 and NKI-beteb (arrowheads), while single, nondendritic putative melanocyte stem cells in the epidermis (e) are labeled only by Bcl-2 (arrows). g) Hydrochinone treatment induced the development of catagen-like morphology with a rounded dermal papilla in pigmented anagen hair bulb. After the treatment, Bcl-2 expression was no longer detectable in melanocytes of the pigmentary-unit (arrowheads) and greatly diminished in melanocytes moving out of the pigmentary-unit (arrows). aii–cii) c-Kit expression is absent from melanocytes in (arrowheads) and outside (arrows) of the pigmentary-unit in the hair bulb of a fully pigmented (aii), a graying (bii), and an almost white (cii) hair follicle. d) Melanocytes in a catagen hair follicle are not labeled by c-Kit, neither in (arrowheads) nor outside (arrows) the pigmentary-unit. e) Melanocytes in the epidermis (arrowheads) and outer root sheath (arrows) are c-Kit+. f) A large, granular, putative mast cell (arrows) in the dermis is labeled with c-Kit but not NKI-beteb. g) Hydrochinone treatment induced the development of catagen-like morphology with a rounded dermal papilla in pigmented anagen hair bulb but failed to induce c-Kit expression in melanocytes inside (arrowheads) and outside (arrows) of the pigmentary-unit.

Also, we could not detect any Kit expression in anagen scalp hair follicles from graying individuals (Fig. 3aii-cii ) or in catagen hair follicles (Fig. 3h ). By contrast, and serving as an intrinsic control, c-Kit expression was always found in epidermal melanocytes, in outer root sheath melanocytes and in dermal mast cells (Fig. 3i, j ). Again, hydrochinone treatment failed to induce c-Kit expression in pigmented hair follicles from graying scalp skin (Fig. 3k ).

Both Bcl-2 and c-Kit were absent from melanocytes in the bulge region of aging hair follicles (not shown).

As a result of increased oxidative stress, graying hair follicles show increased mitochondrial DNA deletion
Pigmented, graying, and unpigmented hair follicles that were harvested from the same donors were analyzed for the presence of the “common” deletion. A total number of 20 different donor samples could be included. We obtained 2–3 pigmented, graying and unpigmented hair follicles from each sample (total number was 153 hair follicles: 51 pigmented, 54 graying, 48 unpigmented). We found an increased number of deletions in hair follicles that were macroscopically classified as graying (8 out 20 donors=40%) and in unpigmented hair follicles (4 out of 20 donors=20%) (Fig. 4 ). These later hair follicles may still contain some unpigmented, macroscopically undetectable melanocytes. In macroscopically pigmented hair follicles, we found only one sample with a deletion (1 out of 20 donors=5%) (Fig. 4) . The presence of mitochondrial DNA deletions was independent of age within the sample group (all donors older than 50 years).

View larger version (27K):
In this window
In a new window

Figure 4. The “common” deletion is highly present in graying hair follicles.

Pigmented, graying, and unpigmented anagen hair follicles from scalp skin biopsies of graying individuals were analyzed for the presence of the “common” deletion as indicated in Materials and Methods. The graph shows data pooled from 20 different donors. The photomicrograph shows a representative example of the PCR analysis of one donor. Each donor’s sample was analyzed in triplicate.

Unpigmented hair follicles retain full viability and grow better in culture than pigmented hair follicles
Over a culture period of 8 d, unpigmented hair follicles showed efficient hair shaft elongation (Fig. 5 ). When compared with hair shaft elongation of pigmented hair follicles from the same donors, unpigmented hair follicles showed significantly higher growth rates than pigmented hair follicles (Fig. 5) . Both groups maintained an anagen-like morphology of the hair bulb and did not enter into a catagen-like stage over this culture period (not shown).

View larger version (78K):
In this window
In a new window

Figure 5. Hair shaft-elongation is higher in unpigmented than in pigmented hair follicles. Anagen hair follicles from scalp skin biopsies of graying individuals were separately cultured in two groups: pigmented hair follicles and unpigmented hair follicles. Examples of individual unpigmented and pigmented hair follicles in culture are given (hair follicles A and B). On day 8 in culture difference in hair shaft elongation is significant (P <0.05 by Mann-Whitney U analysis for two independent samples) between unpigmented and pigmented hair follicles. Data was pooled from 20 independent experiments/individuals. Each experiment included 9 hair follicles per group.

Pigmented hair follicles show melanocyte apoptosis in the pigmentary unit after additional exogenous oxidative stress
Pigmented hair follicles cultured over a period of 3 d in the presence of hydroquinone at concentrations between 10–3 to 10–7 M showed a dose dependent increase of TUNEL+ apoptotic melanocytes in their pigmentary-unit (Fig. 6 ). Hydroquinone (10–7) appeared to be without effect and showed no significant differences to control in TUNEL+ melanocyte number. Hydroquinone (10–4 and 10–3) showed toxic effects on the hair follicle, with increased apoptosis in the hair follicle keratinocyte population in the hair matrix and the inner root sheath (not shown). Hydroquinone (10–5 to 10–6 M) showed apoptosis of melanocytes selectively in the pigmentary-unit by TUNEL-labeling (Fig. 6) .

View larger version (45K):
In this window
In a new window

Figure 6. Oxidative stress induces apoptosis of melanocytes in the pigmentary-unit.

Pigmented anagen hair follicles from scalp skin biopsies of graying individuals were cultured in the presence of hydroquinone and immunohistochemical analysis was performed on the third day in culture. TUNEL+ (green fluorescence) apoptotic melanocytes (red fluorescence) were counted per hair follicle harvested from 4 different donors (3 to 9 hair follicles per donor) and mean and SEM were calculated. Hair follicles treated with 10–5 M hydroquinone showed significantly more apoptotic melanocytes in the pigmentary-unit (P <0.05 by Mann-Whitney U analysis for two independent samples). Abbreviations: hm, hair matrix; dp, dermal papilla.


Here we provide unique evidence for oxidative stress induced loss of melanocytes from the human hair follicle during aging. In detail, we show for the first time that:

1) a decreased number of viable melanocytes in the aging hair follicle bulge and bulb and an increased incidence of hair bulb melanocyte apoptosis in aging individuals are associated with oxidative stress in the pigmentary unit;

2) the aging hair follicle is characterized by the absence of oxidative stress-protectors, such as Bcl-2, and melanocyte growth factors, such as c-Kit;

3) a higher frequency of oxidative stress associated mitochondrial DNA damage occurs in graying hair follicles, while unpigmented hair follicles prove to be not “older” than pigmented hair follicles;

4) melanocytes of the pigmentary unit are highly and selectively susceptible to exogenous oxidative stress damage.

One major route, by which oxidative stress leads to permanent melanocyte damage, appears to pass by the mitochondria, since their DNA is not so well protected as genomic DNA. The accumulation of mutations, therefore, correlates with age and is indicative of general exposure and generation of oxidative stress, (39 , 40) caused for example by psychoemotional stress, inflammation, UV-light, and others. In summary, our findings support the proposed hypothesis of a “free radical theory of graying” and suggest that melanocytes in the hair follicle are highly susceptible to endogenous oxidative stress. Exogenous oxidative stress can trigger and hasten this process and provoke permanent damage selectively and prematurely in hair bulb melanocytes.

Clinical observations reported in the literature support the premature loss of hair bulb melanocytes by oxidative stress. Smokers for example gray early (41 , 42) and in alopecia arreata, a disease characterized by high psychoemotional and inflammatory stress, hair frequently grows back depigmented (43) . The morphological features we observed in melanocytes of aging and graying hair follicles further support this notion. Such features have been described before in melanocytes under oxidative stress for example in culture or in vitiligo (6 , 44 , 45) . Consequently, we could show that melanocytes in the graying hair bulb undergo apoptosis and show oxidative stress damage by 8OHdg labeling (28) .

We also observed ectopic differentiation of melanoblasts in the hair bulb below Auber’s line. This observation can be interpreted as an oxidative stress damage-response, as has been shown in the epidermis for example after UV-light (1 , 46) . Such ectopic melanogenesis in the aging hair follicle has been described before in the bulge region, (12) hosting melanocytic stem cells in the mouse (15) . It was therefore suggested that untimely ectopic differentiation of melanocyte stem cells in the aging hair follicle diminishes the stem cell pool (6) . However, this work did not consider oxidative stress as a cause.

The first model that established stem cells in the hair follicle utilized bromodeoxyuridine to detect label-retaining cells (64) . These studies focused on the bulge region of the hair follicle and did not differentiate, what type of cell maintained the label. They rather assumed that all label-retaining cells were keratinocytes, since they represent the majority of hair follicle constituting cells. Only years later it was shown that the label-retaining cell population in the bulge also hosted melanocytic stem cells (15) . Recently, data were published that show label-retaining putative stem cells in the outer root sheath, close to the outer basement membrane of the hair bulb (65) and in the hair follicle matrix (66) . We hypothesize that these newly defined stem-cell reservoirs of the hair follicle also contain melanocytic in addition to keratinocytic stem cells. This hypothesis is further supported by the observation that the outer root sheath and lateral hair bulb contain amelanotic, undifferentiated melanocytes that cannot be detected by most immunohistochemical labeling methods. However, these cell populations are evident from ultrastructural analysis by electronmicroscopy in the mouse (38) and await analysis by label retention. Ectopic differentiation of melanocytes in these niches may add to the progressive disability of the aging hair follicle to revive the pigmentary unit with every hair cycle.

The progressive loss of melanocytes from aging human hair follicles was previously characterized by decreased numbers of melanocytes immunoreactive to pigmentation-related markers, such as pMel-17, vimentin, and MITF, during anagen, telogen, and catagen of young vs. old donors. Expression of these markers in the ostium and sebaceous glands of young vs. old hair follicles did not differ (5 , 12) . Our findings basically confirm this progressive loss of melanocytes from the hair follicle bulge and bulb during aging. However, by using the pan-melanocyte marker NKI-beteb, we could show that the presence of hair bulb melanocytes may greatly vary even within one donor and that the decrease of hair bulb melanocytes corresponds with the disappearance of the oxidative stress protector Bcl-2. Moreover, the aging hair follicle, but not the aging epidermis, contains only melanocytes that lack expression of the SCF receptor c-Kit and thereby lose the ability to respond to SCF stimulation during migration and pigmentation. In addition, expression of NKI-beteb, Bcl-2, and c-Kit did not differ in the hair follicle ostium and sebaceous gland epithelium between pigmented and unpigmented hair follicles of the same donors (not shown).

Together, these observations emphasize the specific and permanent loss of melanocytes from the pigment-producing unit during graying, including it’s stem cell populations while the epidermal reservoirs and pigmented hair follicles appear to remain relatively untouched. However, we were unable to differentiate between pigmented and graying hair follicles in telogen and catagen, because these follicles are characterized by their low to nil expression of melanocytes independent of their pigmentation status and could therefore not be categorized in terms of their pigmentation status.

Increase in oxidative stress damage has been correlated with age in a variety of species (32 , 47 , 48) and 8Ohdg-labeling was detected alongside increased levels of the ‘common’ deletion in the aging epidermis (28) . We were able to detect TUNEL-immunoreactivity and 8Ohdg-labeling in association with the “common” deletion in graying individuals, indicating permanent oxidative stress damage in the aging hair follicle. However, pigmented and unpigmented hair follicles of the same donors showed less frequent occurrence of the deletion, and in culture, we found that unpigmented hair follicles of the same donors showed increased hair shaft elongation. Together these findings indicate that the hair bulb melanocytes acquire permanent damage first, while the hair shaft producing keratinocytes retain their ability to actively and effectively produce a hair shaft.

To our knowledge, this is first report that proves the long-suspected and often anecdotally described accelerated growth of white scalp hair (1) . The relative protection from the aging process may also be due to their higher antioxidant enzyme activities as compared to melanocytes (49) . The improved growth of unpigmented hair follicles may be due to lower overall oxidative stress and lower calcium, which is normally transferred to the keratinocytes with the melanosomes and may lead to earlier terminal differentiation of pigmented hair follicles for example in beard hair follicles (50) .

Progressive premature graying and loss of melanocytes from aging hair follicles has also been shown in mice deficient in Bcl-2 (12 , 51 , 52) . Bcl-2 is a molecule protecting mitochondria from oxidative stress and subsequent cell death (53 54 55) . Thus, progressive and selective loss of melanocytes from Bcl-2 deficient mouse hair follicles as opposed to melanocytes in the sebaceous gland, and in the absence of general features of premature aging, also supports that oxidative stress is a major cause of hair follicle melanocyte apoptosis. Along this line of evidence one study in mice has shown that topical application of an antioxidant enzyme can protect hair follicles from UV-light induced graying (56) .

Premature graying is a process with dramatic effect on appearance, socio-cultural status and self-esteem of the affected individual. Oxidative stress damage in hair follicle melanocytes appears to be at the tip of the iceberg and can explain the rapid graying process often reported in individuals experiencing environmental and endogenous challenges that generate oxidative stress (67) such as psychoemotional (57 , 58) or inflammatory disease-induced stress(24) . Oxidative stress generated by such challenges may add to the endogenous oxidative stress-load of the hair bulb melanocyte. Upon exposure to such stressors, the highly susceptible hair follicle melanocytes disappear and visibly mark their disappearance (Fig. 7) .

View larger version (30K):
In this window
In a new window

Figure 7. Hypothetical scenario indicating the involvement of oxidative stress in the graying process. Melanocytes in pigmented anagen hair follicles in young and healthy subjects can deal with the endogenous oxidative stress caused by melanin-production. In the old and diseased subject, however, stressors induce alterations in pigment-producing and antioxidative enzymes and cofactors and the production of endogenous antioxidants and repair-enzymes as well as growth factors, etc. decreases. This results in a breakdown of the hair follicle melanocyte redox-capacity and subsequent deleterious oxidative stress damage.

In the light of our findings the graying process may be understood as an easily accessible and quite obvious indicator of the overall oxidative stress an individual is exposed to and of its redox-capacities (28) . This indicator function is relevant to other organ-systems. Accordingly, premature graying was named as a risk factor and indicator in age- and increased oxidative stress-related diseases such as coronary heart disease (59) or osteoporosis (60 , 61) . Moreover, the common deletion was determined in hair follicles to monitor aging (62) and end-stage renal disease (63) . However, these later studies failed to determine hair follicle pigmentation status in the follicles used for their study.

The graying hair follicle therefore offers a unique model–system to study oxidative stress effects and aging and to test antioxidants and other antiaging therapeutics in their ability to slow down or even stop this process. For example, protection from oxidative stress-induced apoptosis by stimulation of the endogenous redox-capacity or treatment with exogenous antioxidants can be tested in hair follicle culture. Also, in this model-organ the activation and maintenance of the follicular melanocyte stem cell pool can be easily assessed. Moreover, in clinical trails the fate of hair follicle melanocytes can be monitored as a measure for oxidative stress-tissue damage and effectiveness of antiaging and antioxidant therapeutics. Alongside with the protection from endo- and exogenous oxidative stressors, this provides us with a plethora of tools to be used in the prevention of excessive oxidative stress in the hair follicle pigmentary-unit and other tissues.


This study was supported by a research grant of the University-Medicine Charité (UFF) to Dr. Peters as well as by the German Research Foundation (Pe890 4-1, Ar 232 14-2). Many thanks are owed to Dr. Hoting, Schwarzkopf, Hamburg, Germany; as well as Profs. Schallreuter and Tobin, Dept. of Biomedical Sciences, Bradford, UK, for introducing the senior author to the world of hair follicle pigmentation. Many thanks go also to our colleagues in plastic surgery that have generously supported our work with the supply of scalp skin biopsies: Dr. Johannes Bruck, Clinica Vita, Berlin, Germany, and Drs. Meyburg, Klinik für kosmetische Chirurgie, Berlin, Germany.

Received for publication June 2, 2005. Accepted for publication January 17, 2006.


An automatic insert of a related ad:

Thanks for your patronage.

If you enjoyed this article you may enjoy these as well

Tags: , , , , , , , , , ,

Leave a Reply

seven + = 15

Disclaimer: I must say this: The information presented herein is for informational purposes only. Consult your doctor, practitioner, and/or pharmacist for any health problem and before using any supplements, making dietary changes, or before making any changes in prescribed medications.
All posts are strictly opinions meant to foster debate, education, comment, teaching, scholarship and research under the "fair use doctrine" in Section 107 of U.S. Code Title 17. No statement of fact is made and/or should be implied. Please verify all the articles on this site for yourself. The Information found here should in no way to be construed as medical advice. If You have a health issue please consult your professional medical provider. Everything here is the authors own personal opinion as reported by authors based on their personal perception and interpretation as a part of authors freedom of speech. Nothing reported here should be taken as medical advice, diagnosis or prescription; medical advice should only be taken from your health care provider. Consume the information found on this web site under your own responsibility. Please, do your own research; reach your own conclusions, and take personal responsibility and personal control of your health.