| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
RESEARCH COMMUNICATION |
a Department of Dermatology, Charité, and
b Institute of Laboratory Medicine, Charité, Campus Virchow Clinic, Humboldt University, Berlin, D-10117, Germany;
c Growth Factor and Regeneration Group, Max-Delbrück Center for Molecular Medicine, Berlin, D-13122 Germany;
d Institute of Immunology, Free University, Berlin, D-12203, Germany; and
e Hoechst Marion Roussel Inc., Bridgewater, New Jersey 08807, USA
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: BDNF • NT-4 • TrkB • p75NTR • catagen
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the present study, we have explored nonneuronal functions of brain-derived neurotrophic factor (BDNF) and neurotrophin-4 (NT-4), two members of the neurotrophin family. These basic proteins exert their biological effects via the high-affinity tyrosine kinase receptor TrkB and the low-affinity p75 neurotrophin receptor (p75NTR) 18-26) . BDNF has become recognized for its critical role in the development of the sensory nervous system: for example, BDNF knockout (-/-) mice are characterized by selective sensory disorders and display a reduced number of neurons in sensory ganglia (27 , 28 ), whereas target-derived BDNF is responsible for the development of different subsets of skin nerve fibers 29-31) and for the function of cutaneous mechanoreceptors (32 , 33 ). In contrast to BDNF null animals, which do not survive longer than 3–4 wk after birth (27 , 28 ), NT-4 knockout mice are viable and fertile. They display a 50% neuron loss in the nodose-petrosal and geniculate ganglia, but lose all D-hair afferent neurons that innervate down hairs in the skin 34-37) . In mouse facial skin, NT-4 plays a suppressive role for the formation of free nerve endings in the epidermis (30) .
In embryonic rat skin, BDNF mRNA expression has been reported in the mesenchymal cells of developing vibrissa HF, and NT-4 mRNA was seen in the HF epithelium (38 , 39 ). The beginning of HF morphogenesis in rat embryos coincides with the presence of high levels of cutaneous BDNF and NT-4 mRNA (40) . In adolescent rat skin, BDNF mRNA expression was found in fibroblasts (41) , whereas TrkB-like immunoreactivity (IR) was reported in epidermal KCs of human skin (42 , 43 ). Together, this suggested that BDNF and NT-4 may have additional functions in normal skin physiology that go beyond their recognized roles in the development and normal function of skin innervation and cutaneous sensory organs.
Here we have exploited the HF, a prototypical neuroectodermal–mesenchymal interaction system, as a particularly instructive model for dissecting the general role of BDNF and NT-4 in epithelial tissue remodeling (44 , 45 ). Studying the murine hair cycle with its highly synchronized, rhythmic transformations between stages of resting (telogen), growth (anagen), and apoptosis-driven regression (catagen) (44 , 46 ), the distribution and expression levels of BDNF, NT-4, and TrkB in back skin of adolescent C57BL/6 mice was analyzed by semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR), in situ hybridization, and/or immunohistology during various stages of the depilation-induced hair cycle (47) . Next, spontaneous HF cycling in infantile BDNF-knockout mice (28) , BDNF-overexpressing mice [A. Subramaniam et al. (1997) Am. J. Resp. Crit. Care Med., Vol. 155, p. A484 (abstr.)], and adolescent NT-4 knockout mice (35) was compared by quantitative histomorphometry to that of corresponding, age-matched wild-type mice. Finally, the effects of BDNF and NT-4 on HF cycling in situ were studied, using a modification of established skin organ culture techniques (10 , 48 ). Using these mouse models, we present the first evidence that BDNF and NT-4 are indeed involved in the control of HF cycling.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
BDNF and NT-4 knockout mice were generated using conventional gene
targeting techniques (28
, 35
) and were
obtained from the corresponding laboratories. Genotyping of mutant
animals was performed using PCR protocols for the mutated alleles and
slot blot analysis of isolated tail DNA (49)
. BDNF
knockout (-/-) mice show vestibular abnormalities and die within
22–23 days after birth (28)
, whereas NT-4 knockout mice
are viable and fertile (35)
. BDNF-overexpressing
transgenic mice were generated using the 
-myosin heavy chain as
promoter [A. Subramaniam et al. (1997) Am. J. Resp. Crit. Care Med.,
Vol. 155, p. A484 (abstr.)], which ensures strong expression of BDNF
in heart, lung, and skeletal muscles (50
,
51
). These transgenic mice are characterized by the
ectopic expression of BDNF mRNA in the panniculus carnosus muscle in
murine subcutis and display an almost threefold increase in the
steady-state levels of BDNF protein in back skin, as determined by
enzyme-linked immunoassay (ELISA) (31)
. BDNF and NT-4
knockout as well as BDNF-overexpressing mice display no obvious
macroscopic hair growth abnormalities.
Active hair growth (anagen) was induced as previously described (47) in the back skin of 6- to 9-wk-old C57BL/6 female mice or in 8-wk-old NT-4 knockout (-/-) and their corresponding age-matched wild-type mice in the telogen phase of the hair cycle. In C57BL/6 mice, all key hair cycle stages (52) were studied, using at least five mice per time point: telogen (untreated skin); anagen II-VI [3–12 days after anagen induction by depilation (p.d.)]; and catagen (17–19 days p.d.) (47 , 53 ). In adolescent NT-4 knockout and corresponding wild-type mice, only the catagen stage (day 19 p.d.) of the induced hair cycle (53) was studied; four mice of each strain were analyzed.
To the analyze spontaneous hair follicle cycling in infantile BDNF knockout (-/-), BDNF-overexpressing, and corresponding age-matched, wild-type mice, skin was harvested 16–23 days after birth (P16–P23: postnatal days 16–23) and four to seven mice of each strain were analyzed. In all experiments, the neck region of the back skin was harvested parallel to the vertebral line and was embedded by using a special technique for obtaining longitudinal cryosections through the HFs from one defined site (54) .
RT-PCR
Total RNA was isolated from full thickness back skin
samples (homogenized in liquid nitrogen), using a single-step guanidine
thiocyanate-phenol-chloroform methodwith the
RNeasy-total-RNA-kit (Quiagen, Hilden, Germany). Skin samples included
the subcutaneous skeletal muscle layer (panniculus carnosus). cDNA was
synthesized by reverse transcription of 3 µg total RNA, using a cDNA
synthesis kit (Invitrogen, San Diego, Calif.). The following sets of
oligonucleotide primers were used to amplify specific cDNA: ß-actin:
5'-GAA AAC GCA GCT CAG TAA CAG TCC G and 5'-TAA AAC GCA GCT CAG TAA CAG
TCC G-3'; BDNF: 5'-GTG AGA AGC TTG ATG ACC ATC C and 5'-AAC AGA ATT CCA
CTA TCT TCC C (55)
; NT-4: 5'-CCC TGC GTC AGT ACT TCT TCG
AGA C and 5'-CTG GAC GTC AGG CAC GGC CTG TTC (56)
.
Amplification was performed using taq polymerase (Gibco, Grand Island,
N.Y.) over 24–34 cycles, using an automated thermal cycler (Perkin
Elmer Cetus, Norwalk, Conn.). Each cycle consisted of the following
steps: denaturing at 94°C (1 min), annealing at 60°C (45 s), and
extension at 72°C (45 s). PCR products were analyzed by gel
electrophoresis (2% agarose) and enzymatic digestion using standard
methods (57
, 58
). For semiquantitative
RT-PCR, linear correlation of signal intensity for ß-actin was found
between 24 and 27 cycles, and for the other markers, between 30 and 35
cycles, using a video scanner system Scan Pack 2.0 (Biometra,
Göttingen, Germany). This allowed for the demonstration of
clearly different signals as compared to control signals.
In situ hybridization
In situ hybridization (ISH) was carried out on mouse skin, as
described (59)
, with the following modifications. Cryostat
sections (8 µm) of paraformaldehyde-fixed skin were placed on
RNAse-free gelatin-coated slides and hybridized with a
digoxigenin-labeled synthetic riboprobe (labeled at the 3'-end with
Dig-dUTP) that is complementary to bases 224-734 of the BDNF gene
(60)
. ISH was performed at 60°C for 17 h with 160
µl of hybridization buffer containing 50% formamide, 4xSSC, and 150
ng/ml of riboprobe. After hybridization, the slides were washed first
in 2xSSC at 69°C for 1 h, then in 0.1xSSC at 69°C for 1 h. After washing, slides were incubated with alkaline
phosphatase-conjugated anti-digoxigenin antibody (Boehringer-Mannheim,
Mannheim, Germany; 3 h, room temperature) and processed for
reaction product development, as described (59)
.
Incubation of sections with ribonuclease or correspondent sense probe,
as well as skin sections from BDNF knockout mice (-/-), were used as
negative controls; cryosections from the heart of BDNF-overexpressing
mice were used as positive controls. These controls confirmed the
specificity and sensitivity of the used ISH methodology.
For analysis of TrkB mRNA, ISH was carried out according to a protocol that had been adapted for oligonucleotide probes complementary to NT and NT receptor mRNAs (17 , 61 ), using the following modifications. Cryostat sections (8 µm) were hybridized with the commercially available synthetic oligonucleotide probes complementary to bases 716-745 (62) of full-length rat TrkB (Biognostic, Göttingen, Germany), labeled at the 3'-end with TRITC-dUTP (Boehringer-Mannheim) using terminal deoxynucleotidyl transferase (TdT, Promega, Madison, Wis.). ISH was performed at 30°C for 16 h with 60 µl of hybridization buffer (Biognostic) containing 50% formamide, 4xSSC, and 10 pmol/ml of oligonucleotide probe. After hybridization, the slides were first washed in 1xSSC 5 min at room temperature, then in 0.1xSSC (4 changes, 15 min each) at 35°C. After washing, slides were rinsed rapidly in deionized water, mounted into Vectashield medium (Vector, Burlingame, Calif.), and examined under a Zeiss Axioscope fluorescence microscope. Hybridization specificity was controlled by the pretreatment of sections with ribonuclease A (100 µg/ml, 1 h at 37°C) as well as by abrogation of the hybridization signal by the addition of 100 pmol/ml of unlabeled oligonucleotide to the TRITC-labeled probe (negative controls). Cryostat sections of embryonic brain with their strong TrkB transcript signals served as positive controls. All controls yielded the expected negative or positive results, documenting hybridization specificity.
Immunohistochemistry
Incubation of cryostat sections of skin either with rabbit
antisera against BDNF, NT-4, TrkB, or chicken antisera against BDNF,
NT-4 (see Table 1
), followed by TRITC-conjugated F(ab)2 fragments of a
goat anti-rabbit immunoglobulin G (IgG) (Jackson ImmunoResearch,
West Grove, Pa.) or a rabbit-anti-chicken Ig (Sigma, St. Louis, Mo.),
was performed as described 63-65)
. For all antisera,
incubation of skin sections without primary antisera and cryostat
sections of embryonic brain were used as negative and positive
controls, respectively. For all antisera, preabsorption of primary
antisera with 100 µg/ml of corresponding antigenic peptides (37°C,
60 min) was used as an additional negative control.
|
For double immunodetection of TrkB- or PGP9.5 (pan-neuronal marker) -IR or BDNF- and NT-4-IR, the tyramide amplification method was used, which had been developed for the simultaneous immunofluorescence determination of two antigens with antisera obtained from the same species (66) . Briefly, after blocking of endogenous peroxidase and nonspecific avidin/biotin binding, sections were incubated in TNB buffer (0.1 M Tris-HCl, pH 7.5, 0.15 M NaCl, and 0.5% DuPont Blocking Reagent, DuPont NEN, Boston, Mass.) for 30 min. Next, rabbit antisera against TrkB (1:1000) or PGP9.5 (1:10000) were applied overnight. At this dilution, antigen–antibody complexes were undetectable for the fluorochrome-labeled secondary goat anti-rabbit antiserum, but became visible when using biotinylated goat anti-rabbit antiserum, diluted in TNB blocking buffer (DuPont NEN; 1:200, 30 min), and the tyramide amplification kit (DuPont NEN). Subsequently, sections were incubated in streptavidin-horseradish peroxidase (1:100 in TNB, 30 min). Three washes with TNT buffer (0.1 M Tris-HCl, pH 7.6, 0.15M NaCl, 0.05% Tween) were followed by a 10 min application of TRITC-tyramide (1:50 in Amplification Diluent, Du Pont NEN). After blocking nonspecific binding with 10% normal goat serum, sections were incubated with rabbit antisera against BDNF or NT-4 (1:50, overnight, room temperature), washed in TBS (3x5 min), followed by incubation with Cy2-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch; 1:20, 37°C, 1 h). Finally, sections were washed three times with Tris buffer and counterstained with Hoechst 33342 for identification of cell nuclei (46 , 67 ).
For the double immunovisualization of p75NTR- or Bcl-2-IR on the one hand, and TrkB-IR on the other, monoclonal antibodies against mouse p75NTR or Bcl-2 (Table 1) were used and the tyramide amplification method was applied for their visualization (see above), using biotinylated goat-anti-rat or goat-anti-hamster IgG, streptavidin-horseradish peroxidase, and TRITC-tyramide. Immunovisualization of TrkB-IR was performed using Cy2-conjugated goat anti-rabbit secondary antibody, as described above. All sections were examined under a Zeiss Axioscope microscope and photodocumented with the help of a digital image analysis system (ISIS MetaSystem, Altlussheim, Germany).
Double immunodetection of TUNEL-positive cells and TrkB-IR
To detect apoptotic cells, we used an established, commercially
available TUNEL kit (ApopTag, Oncor, Gaithersburg, Md.) as described
(46
, 67
). For double immunofluorescence
detection of TUNEL-positive cells and TrkB-IR, sections were incubated
with digoxigenin-dUTP in the presence of TdT, followed by incubation
with rabbit-anti-TrkB antiserum. Subsequently, TUNEL-positive cells
were visualized by anti-digoxigenin FITC-conjugated F(ab)2
fragments. TrkB-IR was detected by goat anti-rabbit TRITC-conjugated
antibody, and sections were counterstained with Hoechst 33342
(46)
. Negative controls for TUNEL staining were made by
omitting TdT, according to the manufacturer's protocol. Positive TUNEL
controls were run in comparison with tissue sections from the thymus of
young mice, which display a high degree of spontaneous thymocyte
apoptosis (46
, 67
).
Skin organ culture
Punch biopsies (4 mm) were prepared under sterile conditions
from adolescent C57BL/6 mouse back skin with all HFs in the late anagen
VI to early catagen stage of the induced hair cycle [i.e., day 17
after depilation (53)
], following previously described
protocols (10
, 48
) with some modifications.
For each experimental group, eight to ten randomized skin punches,
derived from the back skin of three different mice, were placed (dermis
down) on prehydrated gelatin sponges (Gelfoam, Upjohn Co., Kalamazoo,
Mich.) in 35 mm petri dishes containing 5 ml Dulbecco's modified
Eagle's medium, 10% fetal bovine serum, 50 mg/ml L-glutamine, and
antibiotic/antimycotic mixture (Gibco). After addition of 0.5–50 ng/ml
human recombinant BDNF or NT-4 (Promega), organ cultures were incubated
at the air–liquid interphase for 48 h at 37°C in 5%
CO2, and 100% humidity. At the end of incubation, all skin
fragments were washed repeatedly in phosphate-buffered saline buffer at
4°C, fixed in 4% paraformaldehyde, and embedded in paraffin for
routine histology and histomorphometry.
Histomorphometry and statistical analysis
In adolescent skin, IR patterns were examined by studying at
least 50 different HFs per mouse; five mice were assessed per hair
cycle stage. For each stage of HF cycling, the major IR patterns were
recorded in previously prepared, computer-generated schematic
representations of murine HF cycling that allow a standardized, easily
reproducible, and systematic comparison of different follicular IR
patterns (68)
. For precise identification of defined
stages of HF cycling, histochemical detection of endogenous alkaline
phosphatase activity was used since it allows one to visualize the
morphology of the dermal papilla as a useful morphological marker for
staging HF cycling (69)
.
The percentage of HFs in different stages of growth (anagen), regression (catagen), or resting (telogen) was assessed and calculated in BDNF-overexpressing and BDNF knockout (-/-) mice at P16 and P22, respectively, as well as in their corresponding age-matched wild-type littermates. During these days of postnatal development, the hair follicle, after completion of morphogenesis, begins its lifelong cycle of regression, resting, and growth by spontaneous entry into the first catagen stage (44 , 68 ). In normally cycling mice, this occurs around P17, and at P20–22 all hair follicles in back skin reach the resting stage of the hair cycle (telogen).
In adolescent NT-4 knockout (-/-) and corresponding wild-type mice, the percentage of hair follicles in different stages of catagen development was calculated on day 19 of the depilation-induced hair cycle (47 , 53 ). All evaluations were performed on the basis of accepted, well-defined morphologic criteria of HF classification (13 , 44 , 52 , 53 , 68-71 ). Only every tenth cryosection was used for analysis in order to exclude the repetitive evaluation of the same HF, and 2–3 cryosections were assessed from each animal. Altogether, 200–350 follicles in 50–60 microscopic fields, derived from four to seven animals (approximately 50–60 follicles per animal), were analyzed and compared to a corresponding number of HF from the appropriate, age-matched wild-type mice. In skin organ culture, the percentage of HFs in defined catagen stages was calculated in 8–10 biopsies per group at magnification x400 under a Zeiss Axioscope microscope, following accepted morphological criteria for the classification of defined stages of catagen development (53 , 69-71 ). All sections were analyzed at x200 magnification.
For the assessment of hair shaft length, hairs were plucked from the back skin of BDNF-overexpressing (n=8) and wild-type mice (n=4) at P22, i.e., in the telogen phase of the hair cycle. Hair length was measured as a distance from the proximal part of club hair to hair tip (10 hairs were analyzed from each animal) as assessed under x40 of Zeiss Axioscope microscope using a digital image analysis system (ISIS MetaSystem). The length of these hair shafts is directly proportional to the duration of active hair generation during the final stages of HF morphogenesis, and thus reflects the time point when hair shaft production is switched off by entry into the first catagen stage.
Means and SEM were calculated from pooled data. Differences were judged as significant if the P value was lower than 0.05, as determined by the independent Student's t test for unpaired samples.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
In contrast, NT-4 mRNA was present in telogen skin, and steady-state levels increased progressively from telogen to late anagen, peaking during late anagen and anagen-catagen transformation of the HF (days 12–18 after depilation, Fig. 1A, D ). During the development of catagen and the subsequent transformation to telogen (day 25 after depilation), NT-4 gene expression significantly declined again compared with late anagen levels (Fig. 1A, D ). Together, these transcription data are compatible with the concept that both BDNF and NT-4 gene expression is regulated in a hair cycle-dependent manner and that they may be specifically involved in the control of HF transformation from active growth (anagen) to regression (catagen).
Hair follicle expression of BDNF and NT-4 is up-regulated during
catagen
To correlate BDNF and NT-4 gene expression in full thickness skin
with the intrafollicular expression patterns of BDNF and NT-4 in situ
during different stages of HF cycling, in situ hybridization of BDNF
mRNA and immunohistochemistry for BDNF and NT-4 antigen expression were
assessed in the back skin of adolescent C57BL/6 mice. BDNF- and NT-4-IR
patterns were each assessed with two distinct antisera (Table 1)
, since
antibody specificity in discriminating between members of the
neurotrophin family may be unsatisfactory (61)
. However,
both antisera gave very similar results in all staining protocols as
well as in the positive and negative control experiments (not shown),
thus attesting to the specificity and sensitivity of our
immunohistological detection techniques. For more extensive analyses,
we selected an antiserum against amino acids 128-147 mapping at the
carboxy terminus of the human BDNF (which is fully homologous to the
corresponding mouse sequence) and an antiserum against amino acids
80-99 mapping at the amino terminus of the NT-4 precursor
(3)
because they provided the best signal/noise ratio. The
observed expression patterns of BDNF- and NT-4-IR are documented by the
representative examples in Fig. 2
and are schematically summarized in Fig. 3
.
|
|
In line with our RT-PCR results (Fig. 1A, C ), no expression of BDNF mRNA was found in telogen HFs (Fig. 2A ), whereas the only signal indicating the presence of BDNF was observed in dermal nerve bundles of telogen skin, which showed strong BDNF-IR (Fig. 2B ). Double immunovisualization of PGP9.5- and BDNF-IR confirmed the BDNF-positive structures shown in Fig. 2B to nerve fibers (not shown). In telogen epidermis, the BDNF mRNA signal was not above background levels (Fig. 2A ), which is consistent with RT-PCR data (Fig. 1A, C ). KCs in the outer root sheath (ORS) of the HF were weakly NT-4-IR in telogen skin (Fig. 2C ). During early and middle anagen, NT-4-IR was seen in single ORS KCs, but no expression of BDNF mRNA and IR was found (Fig. 3) . During progressing anagen development, BDNF mRNA and BDNF-IR appeared in the proximal ORS and hair matrix of anagen VI HFs (Fig. 2D, E ), which corresponds well to the RT-PCR results (Fig. 1A, C ). NT-4-IR was present not only in hair matrix KCs and the proximal ORS, but also in the inner root sheath (IRS) of anagen VI HFs (Fig. 2F ).
A marked enhancement of BDNF mRNA expression in situ, as well as of BDNF- and NT-4-IR, was noted in proximal hair matrix, ORS, and IRS KCs when the HF entered spontaneously into the early steps of regression (53 , 70 ) (catagen II, Fig. 2G–I ). During later stages of HF regression (catagen VI), KCs of the so-called secondary hair germ and ORS expressed BDNF mRNA and maximal BDNF-IR (Fig. 2J, K ), whereas isolated KCs in the regressing epithelial strand were maximally NT-4-IR (Fig. 2L ). The dermal papilla of murine HF never showed BDNF mRNA-related signals above background, BDNF-, or NT-4-IR at any time during the hair cycle. Sensitivity and specificity of our in situ hybridization and immunohistology protocols were confirmed as described in Materials and Methods and as clearly shown in Fig. 2M–O .
TrkB is prominently expressed in the dermal papilla and regressing
hair follicle epithelium during catagen
To define potential follicular targets for BDNF and NT-4
signaling, whose expression peaked in the regressing HF epithelium
during catagen (Fig. 2G–L
), ISH and immunohistochemistry
were performed to localize expression of TrkB, the common high-affinity
receptor for these neurotrophins, during the anagen-catagen transition.
Also, double immunostaining was performed to covisualize the expression
of TrkB and BDNF or TrkB and NT-4 antigens. Since we had previously
defined the follicular IR of the low-affinity neurotrophin receptor
p75NTR during the anagen-catagen transformation (46)
,
coexpression of TrkB and p75NTR was also assessed.
Again, the patterns of TrkB-IR were defined using two different antisera (Table 1) to assure specificity of the staining results. Both antisera yielded very similar staining results (not shown). Furthermore, the patterns of TrkB-IR were compared to those obtained with different antisera against other Trk receptors (full-length forms of TrkA and TrkC, truncated form of TrkB; ref 45 ) that we are using in related studies. Since the latter antibodies demarcated substantially different IR patterns from the ones described here for TrkB (not shown), this further attested to the specificity and sensitivity of our TrkB immunostaining results.
In anagen VI HFs, expression of TrkB mRNA and TrkB antigen was found in the central and proximal ORS (Fig. 3A ; Fig. 4A, B ). During early catagen, TrkB mRNA and protein were observed in hair matrix KCs and dermal papilla fibroblasts (Fig. 3A ; Fig. 4C, D ), which are widely recognized as the mesenchymal key element of hair cycle control (44 , 72 ). During laterstages of catagen development (catagen VI), expression of TrkB mRNA and IR disappeared from the dermal papilla, but became prominent in the epithelial strand and secondary hair germ (Fig. 3A ; Fig. 4E, F ).
During catagen VI, many TrkB-IR cells in the secondary hair germ coexpressed BDNF-IR (Fig. 4G ). TrkB-IR in the epithelial strand was localized closely to NT-4-IR cells, yet no coexpression of TrkB- and NT-4-IR was seen here (Fig. 4H ). Also, a subpopulation of cells in the secondary hair germ coexpressed TrkB-IR and p75NTR, but no such coexpression was found in the epithelial strand (Fig. 4I ).
|
Because apoptosis-driven HF regression is associated with characteristic patterns of Bcl-2-IR and TUNEL-positive cells (46) , TrkB-IR, on the one hand, and Bcl-2-IR or TUNEL positivity on the other were also covisualized. During catagen VI, numerous TrkB+ KCs in the secondary hair germ contained Bcl-2-IR nuclei and, in the epithelial strand, TrkB- and Bcl-2-IR also frequently colocalized (Fig. 4J ). In contrast, no coexpression of TrkB-IR and TUNEL was found in the secondary hair germ, and only single TrkB-IR KCs in the epithelial strand displayed TUNEL positivity, whereas most TrkB-IR cells in the epithelial strand were TUNEL negative (Fig. 4K ).
Together, these data suggest that TrkB-related signaling plays a role in the control of apoptosis-driven HF regression. This appears to be in line with recent reports that neurotrophins can both suppress and induce apoptosis in different model systems (cf. 73-80 ). Therefore, we used functional assays in order to further explore the role of BDNF and NT-4 in catagen control.
BDNF and NT-4 knockout mice show catagen retardation whereas BDNF
overexpression causes acceleration of catagen and shortening of hair
length
To define whether constitutive BDNF or NT-4 deletion or
overexpression alter spontaneous catagen development, neonatal BDNF
knockout (-/-) and BDNF transgenic mice were compared for the onset
and speed of spontaneous HF regression, using age-matched wild-type
mice as controls. Spontaneous catagen development was studied both in
adolescent NT-4 knockout (-/-) (35)
and corresponding
wild-type mice during the depilation-induced hair cycle
(47)
. For precise identification of defined stages of HF
cycling, namely, of the anagen-catagen transformation, in addition to
well-recognized morphological classification criteria (13
,
46
, 52
, 53
, 68
),
endogenous alkaline phosphatase activity was assessed histochemically,
which allows one to visualize changes in the dermal papilla morphology
as a useful marker for staging HF cycling (69)
.
After the completion of HF morphogenesis (which is often mislabeled as the `first hair cycle'), the hair follicle begins its lifelong cycle of regression, resting, and growth by spontaneous entry into the first catagen stage (14 , 16 , 44 ). In normally cycling mice, this occurs around P17; by P20–P22, most HFs are already in the resting stage of the cycle (telogen) (16) . In BDNF knockout (-/-) mice at P22, many HFs were still in later stages of regression (catagen VI) (Fig. 5 A–C), whereas practically all HFs of wild-type mice were already in the latest stage of catagen or in telogen.
|
Inverse dynamics of catagen development were found in transgenic mice
that overexpress BDNF under the -myosin heavy chain promoter. These
mice are characterized by ectopic expression of BDNF mRNA in the
subcutaneous striated muscle layer (panniculus carnosus) and by
threefold enhanced levels of BDNF protein in skin extracts, as
determined by ELISA (31)
. In transgenic skin, many HFs had
prematurely entered into advanced stages of catagen development on P16,
and a significant decline in the percentage of catagen I and III HFs
was found in BDNF-overexpressing mice compared to wild-type skin (Fig. 5F
). This was in contrast to wild-type animals, where
most back skin HFs were still in late anagen VI or early catagen (Fig. 5D–F
). Moreover, as a reliable additional indicator
of advanced catagen development (47
, 70
,
81
), transgenic skin was substantially thinner than
wild-type skin (Fig. 5D, E
). Hair length in
BDNF-overexpressingmice at P22 (7.29±0.4 mm) was also
significantly shorter compared to the corresponding age-matched
wild-type animals (8.81±0.31 mm, P<0.05), which confirms
premature catagen development (i.e., shortening of the hair
shaft-producing anagen phase) in BDNF transgenic mice.
Similar to infantile BDNF knockout animals, adolescent NT-4 knockout (-/-) mice also showed retardation of catagen development. On day 19 after depilation, many HFs were still in early or in middle catagen (catagen II-IV); a significant reduction in the number of catagen VI HFs was found compared to wild-type skin (Fig. 6 A–C), where only catagen VI-VIII HFs were seen. For comparison, in normally cycling mice the catagen-telogen transformation is usually completed around day 19 (46 , 53 , 68 , 70 ). Also, skin thickness was substantially higher in NT-4 knockout mice compared to that of wild-type animals, indicating a strong retardation of catagen (81) . These mutational analyses suggest that BDNF and NT-4 are both involved in the control of catagen.
|
BDNF and NT-4 accelerate catagen development in skin organ culture
To further probe the concept that BDNF and NT-4 are indeed among
the few catagen-promoting agents identified so far (14
,
44
, 82
, 83
), BDNF or NT-4 were
added to organ-cultured murine skin, with most HFs in the process of
initiating the anagen VI–catagen transformation of the hair cycle.
This assay also addressed the problem that alterations in HF cycling
observed in BDNF/NT-4 mutant mice in vivo might not
necessarily reflect a direct role of these neurotrophins in hair cycle
control, but might result from the consequences of neurotrophin
overexpression or deletion, such as differences in skin and hair
follicle innervation 29-31)
. For this purpose, biopsies
were taken from normal C57BL/6 mouse skin 17 days after anagen
induction by depilation (47)
so that skin contained
homogeneous, well-defined HF populations about to undergo spontaneous,
apoptosis-driven regression (46
, 53
,
70
) and were devoid of functional innervation or vascular
supply. These skin fragments and were cultured at the air–liquid
interphase on gelatin gels (`histoculture') (48)
for
48 h in the presence or absence of BDNF or NT-4 at concentrations
routinely used in neurotrophin studies (0.5–50 ng/ml)
(73-75
, 78
, 80
).
Quantitative histomorphometric analysis (71 , 83 , 84 ) revealed that BDNF or NT-4 treatment each significantly accelerate catagen development in vitro. Compared to vehicle controls, where early catagen HFs (catagen I-III) was dominant (Fig. 7 A), there was a significant increase in the number of HFs in mid-catagen (catagen IV-V) in those skin biopsies that had been cultured in the presence of 0.5 or 5 ng/ml of BDNF. As illustrated by the representative photomicrographs shown in Fig. 7B, C , this acceleration of catagen development corresponded to substantial, easily recognizable morphological differences between control and test biopsies in HF regression in organ culture of intact, normal mouse skin. NT-4 displayed even more potent catagen-promoting activity than BDNF: after incubation with 5 ng/ml of NT-4, some tests HFs were already in catagen VI and a significant increase in the number of catagen IV-V HFs was found (Fig. 7A, D ) compared to vehicle controls. This provides further evidence that BDNF and NT-4 can accelerate spontaneous, apoptosis-driven HF regression even in the absence of functional skin nerves.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although the underlying mechanisms of action remain to be dissected definitively in future studies, stimulation of the HF anagen-catagen transition by BDNF and NT-4 might be achieved via neurotrophin interaction with dermal papilla fibroblasts. These key elements of hair cycle control (72 , 85 ) strongly express TrkB receptors (Fig. 4C, D ) during the initiation of catagen, whereas hair matrix KCs produce potential ligands (i.e., express BDNF and NT-4 protein) (Fig. 2G–I ). This may result in a down-regulation of papilla-derived growth factors that normally stimulate hair matrix keratinocyte proliferation and suppress KC apoptosis in the HF (44 , 46 ), thus contributing to catagen induction.
During further catagen progression, BDNF and NT-4 may directly regulate the apoptosis and/or survival in those HF KCs that express TrkB receptors in the regressing epithelial strand and the secondary hair germ (Fig. 4E, F ). In the secondary hair germ, a unique HF compartment characterized by the predominance of Bcl-2- over Bax-IR, even during catagen (46) , TrkB-IR is often colocalized with Bcl-2-IR; no coexpression of TrkB-IR and TUNEL was observed here (Fig. 4J, K ). This suggests that TrkB and Bcl-2 double-positive cells in the secondary hair germ of catagen VI HFs might represent KCs, which are selected for survival during the catagen-telogen transformation (46 , 86 ).
Since catagen is essentially an apoptosis-driven phenomenon of organ regression (46) , our findings appear to contrast with numerous publications that have shown BDNF to prevent apoptosis and promote survival in many neuronal model systems (22 , 87-91 ). In fact, it has recently been shown that another member of the neurotrophin family, NGF, can suppress apoptosis in KCs, possibly via interacting with TrkA receptors and subsequent up-regulation of Bcl-2 (12) . BDNF often is coexpressed with TrkB in KCs of the secondary hair germ during catagen (Fig. 4G ), and coexpression of Trk receptors and p75NTR can enhance the anti-apoptotic effects of neurotrophins (76 , 77 , 92 ). In addition, someTrkB-IR KCs in the secondary hair germ also coexpress p75NTR (Fig. 4I ). Therefore, it is tempting to speculate that during later stages of catagen development, BDNF indeed acts as a survival factor for selected, individual KCs in the secondary hair germ.
In contrast, in the epithelial strand, which is a substantially larger regressing HF compartment than the secondary hair germ, single TUNEL+ KCs coexpress TrkB (Fig. 4K ). Also, numerous TrkB-positive cells in the epithelial strand are Bcl-2-IR (Fig. 4J ) but, in contrast to the secondary hair germ, do not coexpress p75NTR (Fig. 4I ). Therefore, BDNF (and possibly NT-4) may serve to induce or up-regulate apoptosis in the epithelial strand, whose rate of apoptosis appears to determine the speed of catagen development (46) . This concept of dual neurotrophin functions in different HF compartments is in line with recent findings that NGF and BDNF, which are primarily appreciated for their anti-apoptotic properties (24 , 25 , 88 ), can also induce apoptosis in several systems via interaction with p75NTR (73-76 , 80 ). Since NGF is even capable of inducing apoptosis in medulloblastoma cells via stimulation of TrkA (78 , 93 , 94 ), it should be considered that BDNF and NT-4 may operate as proapoptotic signals for epithelial strand KCs during catagen. Whether apoptosis in individual KCs of the regressing HF is stimulated by BDNF or NT-4 not only varies among follicular KC subpopulations, but most likely also depends on the local concentration of neurotrophins and the receptor expression pattern, namely, the coexpression of Trk and p75NTR on the target cells (76 , 77 , 79 ).
Until the molecular mechanisms of catagen acceleration of by BDNF and NT-4 have been clarified, it is probably wise to operate on the working hypothesis that BDNF- NT-4/TrkB signaling is just one element in the intrinsic `clock' that determines the switch-on of catagen as the net result of multiple (catagen-promoting and -suppressing) factors (for discussion, see ref 44 ).
Clearly, catagen develops even in the absence of BDNF and NT-4 gene products (Figs. 5, 6) . Yet even a mere fine-tuning role of these neurotrophins in catagen control would be of profound clinical interest, since most hair growth disorders seen in clinical practice (effluvium, alopecia, hypertrichosis, and hirsutism) essentially reflect or involve abnormalities in the time course of anagen termination/catagen initiation (44) . Therefore, TrkB receptor agonists and antagonists deserve to be systematically explored as novel hair growth-modulatory agents for the treatment of common hair growth disorders. Together, the roles of BDNF and NT-4 in hair cycle control described here offer intriguing new perspectives to developmental biologists, clinicians, and industry alike and call for a systematic evaluation of neurotrophins as regulators of epithelial tissue remodeling, since it can be studied in an exemplary manner in the HF (13 , 14 , 16 , 44 , 98 ).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Abbreviations: BDNF, brain-derived neurotrophic factor; ELISA, enzyme-linked immunoassay; HF, hair follicle; Ig, immunoglobulin; IR, immunoreactive or immunoreactivity; IRS, inner root sheath; ISH, in situ hybridization; KC, keratinocyte; NGF, nerve growth factor; NT, neurotrophin; NT-3, neurotrophin-3; NT-4, neurotrophin-4; ORS, outer root sheath; P, postnatal day; p75NTR, p75 low-affinity neurotrophin receptor; p.d., depilation; RT-PCR, reverse transcriptase-polymerase chain reaction; TdT, terminal deoxynucleotidyl transferase; TrkA, B, C, tyrosine kinase A, B, C (high-affinity neurotrophin receptors).
Received for publication April 15, 1998.
Revision received October 2, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
epidermal and upper dermal innervation on neurotrophins, trk receptors and p75LNGFR. Dev. Biol. 1998;198:57-81.[Medline]
