Trypsin-Induced Follicular PapillaApoptosis Results
in Delayed Hair Growth and Pigmentation
M. SEIBERG,*S. WISNIEWSKI, G. CAUWENBERGH, AND S.S. SHAPIRO
Skin Research Center, Johnson and Johnson CPWW, Skillman, New Jersey 08558
ABSTRACT Programmed cell death is a con-
trolled process that leads to the elimination of
singlecellsviaapoptosis. Programmedcelldeath
is fundamental to development, morphogenesis,
and homeostasis. Proteases play a major role in
the death process. We have previously shown that
a serine protease, secreted by a keratinocyte cell
line, can induce apoptosis in numerous cell lines.
Here we show that serine proteases can induce
cell death in vivo as well. Using a synchronized
hair growth mouse model, we show that topical
trypsin treatment following depilation induces
cell death at the follicular papilla. This results in
delayinghairgrowthandpigmentation.We specu-
late that trypsin might affect a receptor-medi-
ated signaling pathway that leads to follicular
papilla cell death. Dev. Dyn. 208:553–564, 1997.
r1997 Wiley-Liss, Inc.
Key words: apoptosis; hair cycle; hair follicle;
trypsin; liposomes
INTRODUCTION
The hair follicle is an epithelial structure that under-
goes cycles of active growth (anagen), regression (cata-
gen), and rest (telogen) (Panaretto, 1993). The anagen
phase involves the growth of the hair follicle down into
the dermis, forming a complex layered structure with a
pigmented shaft. Melanogenesis had been shown to be
coupled to anagen (Slominski and Paus, 1993). During
the catagen phase, the hair follicle shortens as its lower
two thirds undergoes programmed cell death and apop-
tosis. In telogen the epithelial cells are resting and the
hair shaft remains inside the short follicle until a new
hairshaftisproduced.Althoughthemorphologicalchanges
throughout the hair cycle are well documented (Chase,
1954), only limited information on the molecular biology of
that cycle has been described (see Panaretto, 1993;
Stenn et al., 1994b, 1996; Seiberg et al., 1995).
The hair follicle is one of the few organs that cycle
throughout adult life. This implies that a portion of the
follicle must be permanent and responsive to the signal
of cycle-reinitiation. Two regions of the follicle might
fulfill these requirements (Cotsarelis et al., 1990; Rey-
nolds and Jahoda, 1991a,b). One is the bulge area, an
epithelial structure that contains a keratinocyte stem
cellpopulationthatcangenerateanewfollicle(Cotsare-
liset al., 1990; Sun et al., 1991; Lavker et al.,1993). The
other is the follicular (dermal) papilla, a mesenchymal
structure that interacts with epithelial cells to induce
hair growth (Reynolds and Jahoda, 1991b; Oliver, 1966;
Jahoda et al., 1984; Oliver and Jahoda, 1988; Jahoda,
1992; Messenger, 1993). The follicular papilla dictates
the nature of the follicle, and it can induce hair growth
even from epithelia that are not normally associated
withhairformation(ReynoldsandJahoda,1990,1991b;
Jahoda, 1992). In order for the follicle to cycle, it is
assumed that the bulge and the papilla must be pro-
tected from cell death. We have previously shown that
bcl-2, a survival gene that rescues cells from pro-
grammed cell death, is expressed in the bulge region
during anagen, and in the follicular papilla throughout
the hair cycle (Stenn et al., 1994a).
Programmed cell death (PCD) is a fundamental
aspect of development, morphogenesis, and tissue ho-
meostasis. Many PCD pathways lead to apoptosis, a
mode of cell death involving cytoplasmic condensation
and specific DNA fragmentation (reviewed in Cohen,
1993; Fesus, 1993; Barr and Tomei, 1994; Martin et al.,
1994; Bellamy et al., 1995; Kroemer et al., 1995; Vaux
and Strasser, 1996). The controls and signals initiating
cell death are only partially shared between different
biological systems, but a common final pathway seems
to be shared by all apoptotic pathways.
Cytoplasmic proteases play a functional role in PCD.
The Caenorhabditis elegans protein Ced-3 is essential
for cell death and its mammalian-homologue cysteine
proteases act as vertebrate PCD genes. Cell granule prote-
ases (granzymes) induce apoptosis in permeabilized cells.
Viral proteins that inhibit apoptosis have protease inhibi-
tor activity. Experimental inhibition of cysteine or serine
proteases inhibits apoptotic cell death in many in vitro
systems (reviewed in Patel et al., 1996).
We have previously shown that a serine protease,
secreted by the keratinocyte cell line Pam212, can
induce apoptosis in numerous cell lines. Moreover, the
induction of apoptosis in vitro was reproducible using
trypsin (Marthinuss et al., 1995b). Here we show that
serine proteases were able to induce apoptosis in vivo
as well. Using a synchronized hair growth mouse model
we show that a topical trypsin treatment immediately
following depilation induces apoptosis in the follicular
papillae. Cell death within the papillae results in a
delay in hair growth and pigmentation.
*Correspondence to: Miri Seiberg, Skin Research Center, J&J
CPWW, 199 Grandview Rd., Skillman, NJ 08558.
Received 21 August 1996;Accepted 14 January 1997
DEVELOPMENTAL DYNAMICS 208:553–564 (1997)
r1997 WILEY-LISS, INC.
RESULTS
Synchronization of Hair Growth in C57Bl/6 Mice
To determine whether serine proteases could induce
apoptosis in vivo, we examined the effect of Trypsin on
the mouse hair cycle. C57Bl/6 mice, at 6–10 weeks of
age, are in the telogen phase of the hair cycle. Hair
growth is induced by wax depilation (plucking) of the
animal’sbackfur(Stenn et al., 1993). The growth phase
(anagen) starts synchronously in all hair follicles at the
time of depilation. The first histological changes are
observed after one day (early anagen), when a new
follicle starts to grow out from the bulge area. Several
days after depilation, the hair growth is visible as the
pink skin of the animal starts darkening. This is due to
pigmentation in the shaft, as the C57Bl/6 mouse con-
tains melanocytes only in their follicles and not in the
epidermis (Slominski and Paus, 1993). By 3–4 days the
hair follicle is fully developed, but the hair shaft is not
yet visible. By 8 days (late anagen) the mouse has a
very dark skin, and the hair shafts start to penetrate
through the epidermis at days 11–12. At day 14 the
mouse back is covered with short hairs. By days 19–21
the regression of the follicle (catagen) is observed
histologically,andby days 21–25 the hair follicle is back
to resting phase. A similar synchronized hair cycle
could also be induced by chemical depilation. In that
case, the lower portion of the follicle remains intact.
Hair shafts of the previous cycle remain intact in the
Fig. 1. Trypsin delivery into hair follicles. Untreated (a) and 12-day trypsin-treated mice (b) were painted
with fluorescently labeled trypsin. Animals were sacrificed after 4 hr, and skins were processed for frozen
sections. Bars 510 µm.
Fig. 2. Trypsin delays hair growth and pigmentation. C57BI/6 mice
were depilated and treated with vehicle or trypsin (1%) immediately after
depilation. (a) 8 days post depilation. Left is treated, right is untreated. (b)
11 days and (c) 14 days post depilation. Left is untreated, middle is
vehicle, and right is trypsin treated. Darker skin color indicates a more
progressive stage of the hair cycle, before hair shafts are visible.
554 SEIBERG ET AL.
dermis, until pushed out by the new hairs. It is impor-
tant to note that the murine hair cycle varies not only
between strains, but also among individual animals.
Therefore, each skin sample was examined histologi-
cally, to verify the phase of the hair cycle.
Delivery of Serine Proteases Into Hair Follicles
To examine the effect of trypsin on the hair cycle of
themouse, we needed to deliver serine proteases, 20–40
kD in size, into the hair follicles. The potential for the
use of non-ionic liposomes to target the pilosebaceous
unit has been previously demonstrated (Niemiec et al.,
1995; Lauer et al., 1996). Using that delivery system,
we were able to target proteases into the mouse hair
follicles. The specificity of the delivery system was
tested with fluorescently labeled trypsin using anagen
animals (most stringent condition). Mice were topically
treated with the labeled trypsin, sacrificed at 1 and 4 hr
after treatment, and their skin was analyzed histologi-
cally.
As shown in Figure 1a almost all of the fluorescent
labelingwasfound within the hair follicle. The 1-hr (not
shown) and 4-hr (Fig. 1a) treatments with the tagged
trypsin displayed an identical staining pattern, with no
apparent additional skin penetration at the later time
point. This observation rules out a possible non-specific
skin extracellular matrix digestion by the protease
(which would show as deeper penetration of the fluores-
cent stain into the stratum corneum at the later time
point).
The effect of long-term trypsin treatment on the
delivery system was also studied. Mice depilated (both
chemically and by wax) and treated with trypsin daily
for 12 days, were treated with the fluorescently labeled
trypsinfor 4 hr (Fig. 1b). No major change wasobserved
in the delivery route into the hair follicles of the treated
skin. A minimal staining at the outer portion of the
stratum corneum of the trypsin-treated skins indicated
some loss of barrier integrity. This was confirmed by
measuringtransepidermalwaterloss(TEWL,seeTable
4 and Discussion).
In contrast to the non-ionic liposomes, using aqueous
buffered solutions or lipid-based delivery vehicles
(Granulex) was shown to be ineffective. We were unable
todemonstrate the delivery of the protease into the hair
follicle and observed no biological effects (not shown).
Trypsin Delays Hair Growth and Pigmentation
A single topical application of trypsin (1%) immedi-
ately after depilation had a dramatic effect on the hair
cycle. As shown in Figure 2, both hair growth and
pigmentation were delayed. Untreated controls exhib-
ited dark skin at 7–8 days after hair growth induction,
while trypsin-treated animals remained pink (not pig-
mented) until day 8 (Fig. 2a). The hair shafts of control
and vehicle treated mice were visible at 11–13 days
after depilation. At that time, the skin color of the
trypsin-treated mice was darker, but still pink, and no
hair shafts were visible (Fig. 2b). By 14 days, the
control mice were covered with short fur, while trypsin-
treated animals exhibited gray skin with no hair shafts
(Fig. 2c). The hair shafts of the trypsin-treated mice
were first visible at days 16–19. These hair shafts were
of reduced quality (e.g., unequal shaft thickness), but
within 4–7 more days, except for length, they were
almost indistinguishable from the controls.
Histology of Trypsin-Treated Hair Follicles
Histological analysis of untreated, liposome control
(identical to untreated, not shown) and trypsin-treated
skins revealed major changes in the trypsin-treated
animals. As shown in Figure 3, the hair follicles were
delayed in their development. The characteristic lay-
ered structure, the expanded follicular papilla, and the
new pigmentation were observed 5–7 days later than
the controls, and the follicles displayed a dialated
infundibulum (compare untreated follicles, Fig. 3A-a,b
to trypsin-treated, Fig. 3B-a–e). At 7–8 days after
treatment about half of the treated follicles started to
overcome the trypsin effect and exhibited characteristic
follicular development of a 3–4-day control follicle. The
upper part of these follicles was still distorted, exhibit-
ing a dialated infundibulum (See Fig. 3B-d). By 11–12
days most of the treated follicles matured, but still
displayed reduced pigmentation and shorter shafts,
resembling a 4–5-day control follicle (delay of 7 days,
compare Fig. 3A-d to Fig. 3B-f). One fourth of the
maturefolliclesremainedhistologicallyatypical(bends,
kinks,unequalshaftthickness)throughouttheobserva-
tion period (14 days). The trypsin treatment results
also in epidermal hyperplasia (see Fig. 3B), as it
induces epidermal differentiation and increases the
number of cell layers of the epidermis.
Trypsin InducesApoptosis at the
Follicular Papilla
Following wax depilation, we expected trypsin to
induceapoptosiswithinthefollicularpapillaandaround
it, as the previous telogen follicle is removed. Upon
chemical depilation we could affect the lower epithelial
portion of the follicle as well. Terminal transferase
end-labeling (TUNEL staining) revealed an increase in
apoptotic figures in the trypsin-treated follicles, rela-
tive to untreated and vehicle-treated controls, regard-
less of the depilation system. As shown in Figure 4,
apoptoticbodiesweredetectedwithinthetreatedfollicu-
lar papilla throughout the first week after hair growth
induction (Fig. 4B-a–e). At a given time, only a few cells
within a single papilla were affected. No other portions
of the follicle, epidermis, or dermis were affected by the
serine protease treatment. While a minimal level of
apoptosis was occasionally detected in untreated early
anagen follicles (see Fig. 4A-a–c), it was always at the
isthmus of the follicle, well above the follicular papilla.
Most of the untreated follicles did not display any cell
death.
Continuous daily applications of trypsin also had a
later effect on the growing follicle. At day 8 post
555
FOLLICULAR PAPILLAAPOPTOSIS DELAYS HAIR GROWTH
Fig. 3. Trypsin delays hair follicle development. Mice treated as
indicated in Figure 2 were sacrificed daily and their skins were processed
for histology using H&E staining. A: Untreated, days 4 (a),5(b),8(c), and
12 (d). Bar 510 µm in a–c, 5 µm in d. B: Trypsin treated, days 4 (a),5(b),
6(c),8(d,e), and 12 (f). Bar 510 µm in a–e, 5 µm in f. Note that the
follicular papilla of treated follicles (dark arrowhead) remains condensed
up to 6 days post depilation (a–c). Layering structure and minimal shaft
pigmentation(white arrowhead) appearinsome, butnotall follicles,only8
days post depilation (d,e). C: Lower portion of a hair follicle, indicating the
papilla (DP), the epithelial layering structure (L), and the developing
pigmented shaft (P).
depilation, about half of the daily treated follicles were
able to overcome the protease effect and started to form
layers (see Fig. 3B-d,e). These follicles exhibited mini-
mal cell death at the upper follicular region, around the
bulge area (Fig. 4B-f). Surprisingly, cell death at such
an important region of the follicle had no effect on
follicular development.
Trypsin Can Induce Changes in Gene Expression
During the Hair Cycle
Several serine proteases have been recently impli-
cated as mediators of signal transduction and regula-
tors of gene expression (e.g., Patel et al., 1996). Our
previous work suggests that serine protease(s) induce
apoptosis in cultured cells via a signal transduction
Fig. 4. Trypsin induces apoptosis in the follicular papillae. Mice were
treatedasindicated in Figure2 and sacrificeddaily. Paraffinsections were
stained for apoptosis using a TUNEL stain with a peroxidase end point
(brown), and methyl green counter-stain. A: Untreated, days 1 (a),2(b),
and3(c) after depilation. Bar 510 µm. B: Trypsin treated. a–e, 1–5 days
after depilation, single trypsin application. f, daily treated for 8 days after
depilation. Bar 510 µm, in a,d,e and 5 µm in b,c,f. Apoptosis is detected
only in the treated follicular papillae (arrowheads). Note the black
pigmented shaft (P) forming above the follicular papilla, which is different
from the brown TUNEL staining (f, see also pigment localization in Fig. 3).
557FOLLICULAR PAPILLAAPOPTOSIS DELAYS HAIR GROWTH
mechanism (Marthinuss et al., 1995a,b), and we had
shown changes in gene expression following serine
protease treatment of cultured cells. Therefore, we
examined the pattern of expression of a series of genes,
throughout the hair growth-delay period. We chose the
RT-PCR amplification assay for its sensitivity, even
though it is semi-quantitative only. We clearly demon-
strate trypsin-induced changes in mRNA levels along
the hair cycle using this assay.
A strong increase was demonstrated for 1L-1band
IFNgmRNAs (Fig. 5A,B, Table 1), genes which are
upregulated in AlopeciaAreata and are associated with
the inhibition of hair growth. Amoderate upregulation
was observed in IL-1amRNA level (Fig. 5B, Table 1), a
gene associated with the inhibition of hair growth in
culture. Since IL-1ainduction could also result from
the loss of the epidermal barrier function, we analyzed
the barrier integrity by measuring TEWL. Amoderate
increase in TEWL, which did not correlate with trypsin
concentration, was always observed (see Table 4).
To verify that the delay in follicular development was
nottheresultofanon-specificirritationoraninflamma-
tory response, we analyzed the mRNA levels of genes
that are upregulated during such situations. We found
no change in the mRNA levels of IL-6, IL-10, and
GM-CSF, a slight upregulation of TNFa, a slight down-
regulation of TNFband TNF-RI, and a moderate
TABLE 1. Patterns of Gene Expression at Day 8
Post Depilationa
Gene Untreated Trypsin
IL-6 22
IL-10 66
GM-CSF 22
TNFa61
TNFb16
TNF-RI 61/22
TNF-RII 22
MIP 12
IL-1a111
IL-1b 6 111
IL-1R 22
IFNg 2 111
c-myc 12
c-myb 11 1
c-fos 16
c-jun 61
Collagenase 16
Tyrosinase 111 6
POMC 16
Transglutaminase 11
a
For experimental details see Figure 5. Note that RT-PCR is
semi-quantitative only. Comparisons are valid only for each
amplified sequence, within the different hair cycles, and not
amongthedifferentgenes. 2, nodetectableexpression;1/22,
a very weak band; 6, a weak band; 1, a strong band; 11,a
stronger band; 111, a very strong band.
Fig. 5. Gene expression during the delayed hair cycle. Mice were
treated as indicated in Figure 2 and sacrificed daily for RNA extraction.
RT-PCR was used to compare steady-state mRNA levels of the genes
indicated. A: mRNA levels (the RT-PCR product of 25 ng total RNA)
throughout the delayed hair cycle. B: mRNA levels (the RT-PCR product
of 250 ng total RNA, 5 ng for G3PDH) at day 8 of the cycle. For each gene
tested in panel B, left is untreated, right is trypsin treated. (See also Table
1.)
558 SEIBERG ET AL.
downregulationinmacrophageinducibleprotein(MIP).
This profile of gene expression (see Fig. 5B, Table 1)
rules out an inflammatory reaction or a response to
dermal irritation. Moreover, mice treated with known
irritants like hexadecane or SDS (1%), either at a single
dose or using daily applications for 2 weeks, showed no
effect on hair growth or pigmentation (not shown).
Their gene expression profile was very similar to an
irritation-induced gene expression profile (Kondo et al.,
1994).
A slight reduction was detected in the mRNAlevels of
c-myb, c-myc, and c-fos, while the c-jun level was
slightly increased throughout the delay period (see
Table 1). Collagenase, a gene regulated via an AP-1
response element, was slightly reduced as well.
Tyrosinase, a key enzyme for hair pigmentation, was
downregulated during the delay in hair growth period.
Its mRNA level increased as the follicles start to
overcome the delay (Fig. 5A). Proopiomelanocortin
(POMC), the precursor of the melanogenic peptide
melanocyte stimulating hormone (MSH), was moder-
ately downregulated throughout the delay period (Fig.
5A). This indicates that trypsin could, directly or indi-
rectly, affect the regulation of melanogenesis as well.
Transglutaminase, a gene induced in apoptosis and
during catagen, did not exhibit a change in mRNA level
during the delayed hair cycle (Fig. 5A, Table 1). This
reflects the limited number of cells within the skin that
are induced to death. At a given time, only a few of the
papilla cells are affected by the protease death signal
(see Fig. 4B). Changes in gene expression within such a
small fraction of the skin could not be detected in a
whole skin assay.
Correlations of gene expression patterns with the
hair cycle should not be confused with expression
within the hair follicle itself, since many cell types
withinskincouldcontributetoacycle-dependent expres-
sion pattern. Some gene expression in total skin is hair
cycle dependent. This includes changes in expression
both within the changing follicle and within the skin
itself (see Seiberg et al., 1995, and references therein).
Other Serine Proteases Have a Reduced
Delaying Effect on Hair Growth
Our delivery studies (Fig. 1) indicate that the effect of
trypsin on hair growth and pigmentation is not the
result of non-specific proteolytic digestion within skin.
To further address this question, the effect of several
serine proteases on the mouse hair growth was ana-
lyzed. Chromameter measurements were used to com-
pared skin color of mice, as skin darkness results from
follicular development (Slominski and Paus, 1993).
Mice treated with different serine proteases were ana-
lyzed for this pigmentation effect at 8 days post depila-
tion.
Both Carboxypeptidase-Y (hydrolyses L-amino acids
at the C-termini of proteins) and a nonspecific endopep-
tidase (Protease IV, cuts 56% of peptide bonds at
neutral PH) had only a minimal delaying effect on hair
growth. Subtilysin (nonspecific peptidase at alkaline
PH), on the other hand, slightly increased the rate of
hair growth. Trypsin (endopeptidase, cuts at the C-side
of Arg, Lys), in comparison, induced the longest delay in
hair growth. Chromameter readings of skin color fully
correlate with the delaying effect (Table 2). Trypsin-
treatedskinswerelesspigmentedthantheCarboxypep-
tidase-Y- or Protease IV–treated skins; untreated skins
(natural development of pigmentation) were slightly
brighter than the Subtilysin-treated skin. This clearly
demonstrates that non-specific proteolytic digestion is
not the major cause for the delayed growth.
TrypsinAffects an Early Step in Hair Growth
Induction
Additional daily trypsin treatments did not prolong
the delay in hair growth. Two, three, and seven treat-
ments per week, up to 2 weeks, resulted in an identical
hair growth profile, both morphologically and histologi-
cally,whencomparedtoasingleapplication(notshown).
Even though the follicular delivery is not changed (see
Fig. 1b), no further delay in the hair cycle is observed
with the additional treatments.
To analyze the timing of the papilla sensitivity to the
death signal, mice were treated with a single dose of
trypsin at different time points. Mice treated immedi-
ately post depilation showed the longest delay in hair
growth and pigmentation (see Table 3). Mice treated 2
and 4 hr after hair growth induction still exhibited a
delayed hair cycle, but progressively shorter. Mice
treated 6 hr after depilation or at later time points were
not delayed for hair growth and were indistinguishable
fromuntreatedcontrol.Skincolormeasurements(Table
3) demonstrated an increase in darkness (more pig-
ment, less of the delay) that correlates with the in-
creased time between depilation and trypsin applica-
tion.
Trypsin Effect Might Involve a
Receptor-Mediated Mechanism
Mice induced for hair growth were treated with
reduced concentrations of trypsin, from 1% down to
0.01% (4 31024M–4 31026M), and analyzed morpho-
logically and colorimetrically for the dose effect on hair
growth. Reducing the trypsin concentration prolonged
TABLE 2. Chromameter Measurements of Serine
Protease-Treated Skinsa
Treatment Sample size L*
Untreated 6 49 60.30
1% Trypsin 6 56.1 60.81
1% Subtilysin 4 42.9 62.87
1% Endo-pep 4 51.8 60.11
1% Carboxy-Y 4 51.9 60.47
aC57Bl/6 mice, 8 days post depilation, were analyzed for skin
brightness (L* scale: 0 5black, 100 5white). Animals were
treated with a single dose of protease immediately after
depilation. Carboxy-Y, carboxypeptidase Y; Endo-pep, non-
specific endopeptidase.
559FOLLICULAR PAPILLAAPOPTOSIS DELAYS HAIR GROWTH
the delay in hair growth and pigmentation by 1–2 days.
Chromameter measurements of the treated skins re-
vealed an increase in brightness (L*, more white, less
pigment) that correlates with the decrease in trypsin
concentration (down to 0.01%) and with the increase in
thedelayofthehair cycle (Table4). This couldsuggest a
receptor-mediated mechanism including desensitiza-
tion with higher doses.
A receptor-based mechanism could involve the occu-
pancy of the receptor by a ligand, or receptor activation
bya proteolytic cleavage. Tofurtheranalyzethe mecha-
nism of the trypsin death signal we used an inactivated
preparation of trypsin (1%, 48 hr at room temperature
in aqueous solution). This preparation was enzymati-
cally inactive, but was not completely denatured. Such
apreparationfurtherenhancedthedelayin hair growth.
Chromameter measurements of treated skins show
that animals treated with the inactive trypsin had the
highest L* score (Table 4). The inactive and native
trypsin preparations induced similar histological
changes, with an increased delay in hair growth for the
inactive preparation (not shown). This clearly indicates
that the proteolytic activity of trypsin does not play a
role in the delay of hair growth. Boiled trypsin, which is
inactive as well but is completely denatured, did not
have any effect on the hair cycle, morphologically or
histologically (not shown). This indicates that the 3-D
structure of trypsin, and not its proteolytic activity,
might be essential for the delaying effect.
DISCUSSION
Using a synchronized hair growth mouse model
(Slominski and Paus, 1993; Stenn et al., 1993) we show
that topical trypsin treatment, immediately after depi-
lation, induces cell death at the follicular papilla. This
death signal, which is independent of the proteolytic
activity of the protease, results in delaying hair growth
and pigmentation. We speculate that trypsin affects a
receptor-mediated signaling pathway that leads to fol-
licular papilla cell death.
We had previously shown that keratinocytes can
undergo spontaneous apoptosis in vitro (Marthinuss et
al., 1995a). We had further demonstrated that a serine
protease, secreted by the keratinocyte cell line Pam212,
can induce cell death of numerous cell lines (Marthi-
nuss et al., 1995b). Using cycloheximide, we demon-
strated that this serine protease activates the death
mechanism via a signal transduction pathway (Marthi-
nuss et al., 1995b). We excluded the proteolytic activa-
tion of the thrombin receptor, PAR-2 (a protease-
activated receptor expressed in keratinocytes; Santulli
et al., 1995) and urokinase plasminogen activator, as a
part of the death signaling pathway (Marthinuss et al.,
1995b).
Herewe demonstrate that serine proteases could also
induce apoptosis in vivo, and probably via a receptor-
mediated mechanism. We clearly demonstrate that the
proteolytic activity itself is not necessary to induce the
cell death. Reducing the trypsin concentration results
in an increase in the delaying activity, suggesting
receptor desensitization with higher doses. Since boiled
trypsin did not affect hair growth, we speculate that the
3-D structure of the molecule might be important for
this process. The timing of the protease application is
very critical. The follicular papilla is sensitive to the
death signal only during the early steps of hair growth
induction. Daily application of trypsin had no additive
effect on the hair cycle.
One possible mechanism for serine protease-induced
apoptosis is the perforin-granzyme mechanism, em-
ployed by cytotoxic lymphocytes. The combined effect of
perforin, a pore-forming protein, and granzymes, a
family of granule proteins that includes many serine
proteases, leads to apoptosis and DNA fragmentation of
target cells (reviewed in Patel et al., 1996). One might
claim that wax depilation causes enough damage to the
remaining cells to enable the entry of trypsin without
additional perforin activity. However, the waxing dam-
age is not unique to the follicular papilla, so one might
expect the induction of apoptosis in dermal and epider-
mal cells as well. Moreover, depilatory creams induce
hair growth of telogen mice without physically damag-
ing the lower portion of the follicle or the follicular
papillae. Trypsin could induce the same delay in hair
growth in the wax and chemically depilated animals.
TABLE 3. Chromameter Measurements of
Trypsin-Treated Skina
Treatment Sample size L*
Untreated 6 49 60.32
At depilation 6 56.1 61.54
After 2 hr 4 54.4 61.30
After 4 hr 4 53.3 61.23
After 6 hr 4 48.0 62.01
After 18 hr 4 49.2 61.07
After 48 hr 4 50.4 61.01
aC57Bl/6 mice, 8 days post depilation, were analyzed for skin
brightness (L* scale: 0 5black, 100 5white). Animals were
treated with a single dose of trypsin (1%) immediately follow-
ing depilation, or as indicated.
TABLE 4. Physical Properties of
Trypsin-Treated Skinsa
Treatment Sample size L* TEWL
Untreated 6 45 60.97 27.01 63.8
Liposomes 4 45 61.03 29.81 62.9
0.01% Trypsin 4 48.6 61.01 36.89 64.6
0.1% Trypsin 4 49.5 61.07 43.2 65.2
0.5% 4 47.7 60.3 34.53 64.9
1% Trypsin 4 47.9 60.08 37.5 63.9
1% Trypsin—inactive 3 51.2 60.17 33.99 66.2
aC57Bl/6 mice, 9 days post depilation, were analyzed for skin
brightness (L*scale: 0 5black,100 5white) and transepider-
mal water loss (TEWL; increases when barrier function of the
epidermis is disrupted). Animals were treated with a single
dose of trypsin, immediately after depilation. Trypsin was
inactivated by incubating at room temperature for 48 hr, in
aqueous solution.
560 SEIBERG ET AL.
The epithelial-mesenchymal interactions which lead
to skin appendage formation are well studied (Hardy,
1992; Jahoda, 1992), demonstrating the important role
of the follicular papilla in hair formation. The follicular
papilla could induce follicle formation from the epithe-
lium of the palm, which is usually not associated with
hair follicles (Reynolds and Jahoda, 1991b; Jahoda,
1992). The combination of follicular papilla fibroblasts
with epithelial cells results in pigmentation and hair
follicle formation in nude mice (see Prouty et al., 1996,
and references therein). Inducing cell death at the
follicular papilla has not yet been described. Here we
show that the induction of apoptosis in the follicular
papillae results in perturbation of the hair cycle, and in
delaying hair growth and pigmentation.
During the hair cycle the papilla changes its size by
the addition of extracellular matrix molecules. The
number of cells within this structure remains constant
throughout the hair cycle (Messenger et al., 1991). The
trypsin-induced apoptosis might result in a decrease in
thenumber of cells within the papilla. It ispossiblethat
the delay in hair growth equals the time needed to
regenerate these missing papillae cells. Maybe only
when a damaged papilla is recovered, then new hair
formation can take place. It is possible, also, that
atypical hair follicles rise from a papilla that is not
completely recovered. Such a mechanism could further
demonstratetheimportanceofthismesenchymalstruc-
ture in hair follicle formation. We cannot exclude the
possibility that amelanotic melanocytes or endothelial
cells were induced to death by the protease. The
distribution of the apoptotic bodies within the papilla,
however, does not resemble the localization of follicular
melanocytes or endothelial cells.
We had previously shown that the localization of
Bcl-2 (a protein that negatively regulates apoptosis)
within the hair follicle is hair cycle dependent (Stenn et
al., 1994a). The papilla continues to express bcl-2
throughout the cycle, as well as during telogen, while
the bulge area, where follicular stem cells reside,
expressesthisprotectiveproteinonlyduringthegrowth
phase. This supports the notion that the follicular
papilla is a long-lived structure, and that its function is
continuously necessary. The induction of papilla cell
death by the proteases, therefore, might reflect on a
specialproperty of the papilla at the time of hair growth
induction, which makes it sensitive to the death signal.
Thistiming could correlate with the reduced expression
of one or several members of the Bcl-2 family. Alterna-
tively, the protease signal could act downstream of the
bcl-2 protection check point, as was shown for epider-
mal keratinocytes apoptosis (Marthinuss et al., 1995b).
Nexin-1, a serine protease inhibitor, is expressed
within intact skin at the late anagen papilla only (Yu et
al., 1995). Nexin-1 provides a mechanistic control for
serine proteases at and near the cell surface of fibro-
blasts (Wagner et al., 1989). Nexin-1 had been shown to
rescue neuronal cells from apoptosis (Houenou et al.,
1995). It is conceivable that the expression of Nexin-1
enables the papillae to survive the protease death
signal, possibly by complexing with the catalytic site
serine residue and internalizing and degrading the
complex. This would have the effect of changing puta-
tive follicular death into delayed growth.
Some gene expression within skin is hair cycle-
dependent. We had demonstrated a change in patterns
of gene expression immediately before and during
catagen, which might be involved in the regulation,
initiation, or execution of the regression of the lower
follicle (Seiberg et al., 1995). The changes in gene
expression induced by trypsin do not reproduce the
pattern of gene expression during catagen. The general
reduction in mRNA levels of several genes observed
here is reflective of the overall slow down in hair
growth, and not of catagen. This implies that the serine
proteases do not induce premature catagen. Indeed,
trypsin treatments during anagen did not induce the
regression of the follicle. Moreover, while TNFa, EGF,
and IL-1bcan abrogate hair growth in organ culture, a
catagen-like morphology is formed with TNFaand
EGF, but not with IL-1b(Hoffmann et al., 1996). Here
we show that IL-1b, but not TNFa, is highly upregu-
lated by the serine protease induced papilla death,
suggestingamechanismdifferentfromprematurecata-
gen. Transglutaminase, a gene upregulated in apopto-
sis and catagen, is not upregulated by the trypsin
signal. This also supports the notion that the death
signal does not induce premature catagen. Changes in
mRNA level of a few of the papilla cells cannot be
detected by RT-PCR of total skin, while transglutami-
nase gene expression of regressing follicles is easily
detected (Seiberg et al., 1995).
IL-1aand IL-1bare related proteins with broad
biological activity, associated mainly with inflamma-
tion, but also expressed in non-immune cells including
epidermal keratinoytes and fibroblasts. Both IL-1aand
IL-1bcan inhibit follicular growth in organ culture
(Harmon and Nevins, 1993; Hoffmann et al., 1996), and
transgenic mice overexpressing IL-1ain the skin ex-
hibit patchy hair loss (Groves et al., 1995). In Alopecia
Areata,whena cascade of immunological events results
in hair loss, high levels of IL-1bexpression has been
reported (Hoffmann et al., 1994; Telegdy et al., 1994).
We observe the highest increase in gene expression of
the trypsin delayed hair cycle at the mRNA levels of
IL-1band IFNg. The increase observed in IFNglevel
also correlates with the IFNgupregulation in Alopecia
Areata (Hoffmann et al., 1994; Telegdy et al., 1994).
We demonstrate a significant increase in IL-1b
mRNA, and a moderate upregulation in IL-1amRNA
level. While both could affect hair growth, it is impor-
tant to note that IL-1aexpression is also stimulated by
epidermal barrier disruption (Wood et al., 1996). Our
observation of increased TEWL indicates that the mod-
erate increase in IL-1acould reflect the barrier effect,
and might not be related to the hair growth delay. The
increase in IL-1bduring the delayed hair cycle might
provide a clue to the mechanism of the perturbed
561
FOLLICULAR PAPILLAAPOPTOSIS DELAYS HAIR GROWTH
follicular growth. If IL-1bis involved in the mainte-
nance of the delayed growth, then one might expect the
follicleto remain dormant as long as IL-1bisexpressed.
This is, indeed, the situation in our study. When IL-1b
levels start to decrease, the follicles start to overcome
the inhibitory signal. This is observed histologically,
and later also morphologically. At day 14, when the
growth delay is over, the level of IL-1bmRNA is below
detection. It would be interesting to analyze the levels
of IL-1 receptor antagonist (IL-1ra) and melanocyte
stimulating hormone (aMSH), two potent inhibitors of
IL-1b, and find out whether an increase in one of these
molecules downregulates IL-1band abrogates the
growth inhibition. These two regulators are expressed
in skin: MSH has a major role in pigmentation (Wint-
zen et al., 1996), and IL-1ra gene polymorphism is
associated with the severity of AlopeciaAreata (Tarlow
et al., 1994; Cork et al., 1995). IL-1 expression is high in
undifferentiated keratinocytes, and is reduced when
they become terminally differentiated (Ansel et al.,
1988). Whether it is possible to draw an analogy to an
undifferentiated (telogen) follicle vs. the growing, more
differentiated (anagen) one, remains to be studied.
The increase in IL-1band IFNgcould also reflect an
inflammatory response or could result from epidermal
irritation. To verify that the delay in follicular develop-
ment is not the result of such non-specific processes, we
analyzed the mRNA levels of several other genes that
are upregulated during irritation and inflammatory
situations (Kondo et al., 1994). The profile of gene
expression demonstrated (see Table 1) rules out an
inflammatory reaction or a response to dermal irrita-
tion. Moreover, animals treated with known irritants
had no effect on hair growth and pigmentation. Their
gene expression profile was very different, resembling
the described irritant-induced gene expression profile
(Kondo et al., 1994).
Tyrosinase, the major regulator of hair pigmentation
(Sanchez-Ferreret al., 1995; Mishima, 1994) was down-
regulated during the delay in hair growth period. Its
mRNA level increased as the follicles started to over-
come the delay (Fig. 5). Proopiomelanocortin (POMC),
the precursor of several peptides including the melano-
genic peptide MSH (Jimbow, 1995; Wintzen et al.,
1996), was slightly downregulated throughout the de-
layed hair cycle. The cell death induced by the protease
at the papilla, therefore, affects the pigmentation of the
follicle. This implies that the follicular papilla has an
important role in the regulation of melanogenesis as
well.
In this study, we were able to manipulate the hair
cycle via trypsin-induced apoptosis at the follicular
papilla. This is the first demonstration of induced
apoptosis in the follicular papilla, a structure thought
to be protected from cell death. The proteolytic activity
of the protease is not necessary for the death signal. We
speculate that trypsin might affect a receptor-mediated
signaling pathway that leads to follicular cell death.
Whether trypsin blocks a survival signal or activates a
death receptor remains to be studied.
MATERIALS AND METHODS
Chemicals
Serine proteases and other chemicals were from
Sigma(St.Louis,MO).Granulex(DowHickamPharma-
ceuticals Inc., Sugar Land, TX) contains 1.25% trypsin
in Balsam Peru and Castor Oil. Trypsin was inacti-
vated by either incubating at room temperature for 48
hr, or boiling for 10 min, in 0.05M Hepes PH 7.4.
Proteolytic activity was analyzed by the PanVera kit
(Madison, WI). Trypsin was labeled fluorescently using
a Molecular Probes kit (Eugene, OR). GDL liposomes
(50 mg lipids/ml) were prepared as described in Ni-
emiec et al. (1995). The non-ionic liposomal formulation
containsglyceroldilaurate/cholesterol/polyoxyethylene-
10-stearyl ether ratio 58/15/27.
Animals and Hair Growth Induction
C57Bl/6 female mice, 7–8 weeks old, were purchased
from Charles River (Kingston, NY). Hair growth was
induced by wax depilation as previously described
(Stenn et al., 1993), or by chemical depilation (Neat,
Nair). Each experiment described was performed using
both chemical and wax depilation, with similar results.
Askin sample from each animal was examined histologi-
cally, using H&E stain. One hundred microliters of the
protease solution were applied to the back of each
animal. Each experiment was repeated at least four
times, with at least three animals per group. Transepi-
dermal water loss (TEWL) was measured with an
Evaporimeter using standard techniques (Evaporim-
eter EPI, Servomed AB, Stockholm, Sweden). Color
measurementswereperformedusingtheMinoltaChro-
mameter model CR300 using standard techniques
(Osaka, Japan).
TUNELAssay
Apoptotic staining was performed on paraffin sec-
tions using Apoptag (Oncor, Gaithersburg, MD, manu-
facturer’s protocol), a technique based on the labeling of
fragmented-DNA ends (Gavrieli et al., 1992). Each
experiment was repeated at least three times. Pictures
presented are of a single experiment. Apoptotic cells
were defined by both morphology (condensed or frag-
mented nuclei and cytoplasm or apoptotic bodies), and
staining (fragmented DNA within the condensed nuclei
or apoptotic bodies).
RT-PCR
At the time points indicated animals were sacrificed
and total RNAs were extracted from whole skins (RNA
Stat-60, Tel-Test B, Friendswood, TX, manufacturer’s
protocol). At least 3 animals were used for each time
point studied, and samples were analyzed individually.
Total skin DNased-RNA (200 ng) (Promega, Madison,
WI, RQ1 RNase-free DNase, manufacturer’s protocol)
from each sample was reverse transcribed (Gibco-BRL,
562 SEIBERG ET AL.
Gaithersburg, MD, Superscript II reverse transcrip-
tase, manufacturer’s protocol), using random hexamers
(Gibco-BRL). RT products were PCR-amplified (Taq
polymerase, Perkin-Elmer-Cetus, Branchburg, NJ), us-
ing Clontech primers, Clontech positive control and
Clontech PCR protocol (Clontech, Palo Alto, CA) for
mouse glyceraldehyde-3-phosphate-dehydrogenase
(G3PDH),transcriptionfactorsandcytokines.Transglu-
taminase primers were: sense: 58AACCCCAAGTTCCT-
GAAG and antisense: 58TTTGTGCTGGGCCACTTC.
Thereactioncontained 2.5 mM MgCl2and the cycle was
of 1 min at 94°C, 2 min 55°C, and 3 min at 72°C, for 35
cycles.Tyrosinaseprimers were: sense: 58TCAGCCCAG-
CATCCTTCTTC and antisense: 58CAGCCATTGTTC-
AAAAATACTGTCC. The reaction contained 5 mM
MgCl2andthecyclewasof 1 min at 94°C, 2 min at 45°C,
and 3 min at 72°C, for 35 cycles. POMC primers were:
sense: 58AAAAGAAGAGAGAAGAGCGAC and anti-
sense: 58AGAGCTGAGACACCCTTACC. The reaction
contained 2.5 mM MgCl2and the cycle was of 1 min at
94°C, 2 min at 55°C, and 3 min at 72°C, for 35 cycles.
Collagenase primers were: sense: 58AAGACCCCAAC-
CCTAAGCACand antisense: 58CAGCACTGACGGTTT-
TCACC. The reaction contained 2.5 mM MgCl2and the
cycle was of 1 min at 94°C, 2 min at 53°C, and 3 min at
72°C, for 35 cycles. PCR products were ethanol precipi-
tated when required. For G3PDH, only 10% of the PCR
reaction was used. PCR products were analyzed on 2%
agarose/ethidium bromide gels. An RNA sample that
was not reverse-transcribed was used as a negative
control for each PCR amplification. Obtaining no band
indicates the lack of genomic DNA contaminants. A
six-month-oldmouseskin(non-synchronizedhaircycle)
RT was used as a positive control when plasmids were
not available. The migration of the RT-PCR products on
the gels was always identical to the positive controls,
and to the reported amplimer sizes. To compare the
relative quality of the RT-PCR reactions, the transcrip-
tion level of G3PDH, a ‘‘housekeeping’’ gene, was used
as a control. G3PDH gene expression was found to be
similar at all the time points examined (see Fig. 5),
enabling analysis of the relative levels of gene expres-
sion for the desired genes.
ACKNOWLEDGMENTS
We thank Dr. S. Niemiec for the preparation of
liposomes and advising on delivery systems, Ms. A.G.
Johnson and Mr. P. Siock for excellent technical assis-
tance, Mr. J. Pote for assisting with chromameter
readings, Dr. S. Prouty for fruitful discussions through-
out this study, and Drs. M. Eisinger, A. Gosiewska, S.
Niemic, and S. Prouty for a critical reading of this
manuscript.
REFERENCES
Ansel, J.C., Luger, T.A., Lowry, D., Perry, P., Roop, D.S.R., Mountz,
J.D. (1988) The expression and modulation of IL-1 alpha in murine
keratinocytes. J. Immunol. 140:2274–2278.
Barr, P.J., Tomei, D.L. (1994) Apoptosis and its role in human disease.
Biotechnology 12:487–493.
Bellamy, C.O.C., Malcomson, R.D.G., Harrison, D.J., Wyllie, A.H.
(1995) Cell death in health and disease: The biology and regulation
of apoptosis. Semin. Cancer Biol. 6:3–16.
Chase, H.B. (1954) Growth of the hair. Physiol Rev. 34:113–126.
Cohen, J.J. (1993) Apoptosis. Immunol. Today 14:126–130.
Cork, M.J., Tarlow, J.K., Clay, F.E., Crane, A., Blakemore, A.I.F.,
McDonagh, A.J.G., Messenger, A.G., Duff, G.W. (1995) An allele of
the IL-1 receptor antagonist as a genetic severity factor in alopecia
areata. J. Invest. Dermatol. 104(Suppl. 5):15s–16s.
Cotsarelis, G., Sun, T.T., Lavker, R.M. (1990) Label-retaining cells
reside in the bulge area of pilosebaceous unit: Implications for
follicular stem cells, hair cycle, and skin carcinogenesis. Cell
61:1329–1337.
Fesus, L. (1993) Biochemical events in naturally occuring forms of cell
death. Fed. Eur. Biochem. Soc. 328:1–5.
Gavrieli, Y., Sherman, Y., Ben-Sasson, S.A. (1992) Identification of
programmed cell death in situ via specific labeling of nuclear DNA
fragmentation. J. Cell Biol. 119:493–501.
Groves,R.W.,Mizutani, H., Kieffer,D.J., Kupper,T.S.(1995) Inflamma-
tory skin disease in transgenic mice that express high levels of IL-1
alpha in basal epidermis. Proc. Natl. Acad. Sci. U.S.A. 92:11874–
11878.
Hardy, M.H. (1992) The secret life of the hair follicle. Trends Genet.
8:55–61.
Harmon, C.S., Nevins, T.D. (1993) IL-1 alpha inhibits human hair
follicle growth and hair fiber production in whole organ culture.
Lymphokine Cytokine Res. 12:197–303.
Hoffmann, R., Wenzel, E., Huth, A., Steen, P., Schaufele, M., Hen-
ninger, H.P., Happle, R. (1994) Cytokine mRNA levels in alopecia
areatabefore and aftertreatment with thecontact alergan diphenyl-
cyclopropenone. J. Invest. Dermatol. 103:530–533.
Hoffmann, R., Eicheler, W., Huth, A., Wenzel, E., Happle, R. (1996)
Cytokine and growth factors influence hair growth in vitro. Possible
implications for the pathogenesis and treatment of alopecia areata.
Arch. Dermatol. Res. 288:153–156.
Houenou, L.J., Turner, P.L., Li, L., Oppenheim, R.W., Festoff, B.W.
(1995)Aserine proteaseinhibitor, proteasenexin-1, rescuesmotoneu-
rons from naturally occurring and axotomy-induced cell death. Proc.
Natl. Acad. Sci. U.S.A. 92:895–899.
Jahoda, C.A.B. (1992) Induction of follicle formation and hair growth
by vibrissa dermal papillae implanted into rear wounds: Vibrissa-
type fibres are specified. Development 115:1103–1109.
Jahoda, C.A.B., Horne, K.A., Oliver, R.F. (1984) Induction of hair
growth by implantation of cultured dermal papilla cells. Nature
311:560–562.
Jimbow, K. (1995) Current update and trends in melanin pigmenta-
tion and melanin biology. Keio J. Med. 44:9–18.
Kondo, S., Pastore, S., Shivji, G.M., Mckenzie, R.C., Sauder, D.N.
(1994) Characterization of epidermal cytokine profiles in sensitiza-
tion and elicitation phases of allergic contact dermatitis as well as
irritant contact dermatitis in mouse skin. Lymphokine Cytokine
Res. 13:367–375.
Kroemer, G., Petit, P., Zamzami, N., Vayssiere, J.-L., Mignotte, B.
(1995) The biochemistry of programmed cell death. FASEB J.
9:1277–1287.
Lauer,A.C., Ramachandran, C., Lieb, L.M., Niemiec, S., Weiner, N.D.
(1996) Targeted delivery to the pilosebaceous unit via liposomes.
Adv. Drug Delivery Rev. 18:311–324.
Lavker, R.M., Miller, S., Wilson, C., Costarelis, G., Wei, Z.-G., Yang,
J.-S., Sun, T.-T. (1993) Hair follicle stem cells: Their location, role in
haircycle,and involvement intumor formation. J.Invest.Dermatol.
101:16s–26s.
Marthinuss,J., Lawrence, L.,Seiberg, M.(1995a)Apoptosis in Pam212,
an epidermal keratinocyte cell line: A possible role for bcl-2 in
epidermal differentiation. Cell Growth Differ. 6:239–250.
Marthinuss, J., Andrade-Gordon, P., Seiberg, M. (1995b) A secreted
serine protease can induce apoptosis in Pam212 keratinocytes. Cell
Growth Differ. 6:807–816.
Martin, S.J., Green, D.R., Cotter, T.G. (1994) Dicing with death:
Dissecting the components of the apoptosis machinery. Trends Biol.
Sci. 19:26–31.
563FOLLICULAR PAPILLAAPOPTOSIS DELAYS HAIR GROWTH
Messenger, A.G. (1993) The control of hair growth: An overview. J.
Invest. Dermatol. 101:4s–9s.
Messenger, A.G., Elliot, K., Westgate, G.E., Gibson, W.T. (1991)
Distribution of extracellular matrix molecules in human hair fol-
licles. Ann. N.Y. Acad. Sci. 642:253–262.
Mishima, Y. (1994) Molecular and biological control of melanogenesis
through tyrosinase genes and intrinsic and extrinsic regulatory
factors. Pigment Cell Res. 7:376–387.
Niemiec, S., Ramachandran, C., Wiener, N. (1995) Influence of non-
ionic liposomal composition on topical delivery of peptide drugs into
pilosebaceous units: An in vivo study using the hamster ear model.
Pharm. Res. 12:1184–1188.
Oliver, R.F. (1966) Whisker growth after removal of the dermal papilla
and lengths of follicle in the hooded rat. J. Embryol. Exp. Morphol.
3:331–347.
Oliver, R.F., Jahoda, C.A.B. (1988) The dermal papilla and mainte-
nance of hair growth. In: ‘‘The Biology of Wool and Hair,’’ Marshall,
R.C., Rogers, G.E., Reis, P.J. and Ward, K.A. (eds). London: Chap-
man and Hall, pp. 51–67.
Panaretto, B.A. (1993) Gene expression of potential morphogenes
during hair follicle and tooth formation: A review. Reprod. Fertil.
Dev. 5:345–360.
Patel, T., Gores, G.J., Kaufmann, S.H. (1996) The role of proteases
during apoptosis. FASEB J. 10:587–597.
Prouty, S.M., Lawrence, L., Stenn, K.S. (1996) Fibroblast dependent
induction of a murine skin lesion with similarity to human common
blue nevus. Am. J. Pathol. 148:1871–1885.
Reynolds, A.J., Jahoda, C.A.B. (1990)Adult rat pelage dermal papilla
cells induce type-specific follicle formation and hair growth in adult
footpad skin. J. Invest. Derm. 95:485.
Reynolds, A.J., Jahoda, C.A.B. (1991a) Hair follicle stem cells? A
distinct germinative epidermal cell population is activated in vitro
by the presence of hair dermal papilla cells. J. Cell Sci. 99:373–385.
Reynolds, A.J., Jahoda, C.A.B. (1991b) Inductive properties of hair
follicle cells. Ann. N.Y. Acad. Sci. 642:226–42.
Sanchez-Ferrer,A., Rodrigez-Lopez, J.N., Garcia-Canovas, F., Garcia-
Carmona, F. (1995) Tyrosinase: A comprehensive review of its
mechanism. Biochem. Biophys. Acta. 1247:1–11.
Santulli, R.J., Derian, C.K., Darrow,A.L., Tomko, K.A., Eckardt,A.J.,
Seiberg, M., Scarborough, R.M., Andrade-Gordon, P. (1995) Evi-
dence for the presence of a protease-activated receptor distinct from
thrombin receptor in human keratinocytes. Proc. Natl. Acad. Sci.
U.S.A. 92:9151–9155.
Seiberg, M., Marthinuss, J., Stenn, K.S. (1995) Changes in expression
of apoptosis-associated genes in skin mark early catagen. J. Invest.
Dermatol. 104:78–82.
Slominski, A., Paus, R. (1993) Melanogenesis is coupled to murine
anagen: Towards new concepts for the role of melanocytes and the
regulation of melanogenesis in hair growth. J. Invest. Dermatol.
101:90s–97s.
Stenn, K.S., Paus, R., Dutton, T., Sarba, B. (1993) Glucocorticoid effect
on hair growth initiation: A reconsideration. Skin Pharmacol.
6:125–134.
Stenn, K.S., Lawrence, L., Veis, D., Korsmeyer, S., Seiberg, M. (1994a)
Expression of the bcl-2 protooncogene in the cycling adult mouse
hair follicle. J. Invest. Dermatol. 103:107–111.
Stenn, K.S., Prouty, S., Seiberg, M. (1994b) Molecules of hair follicle
cycling: a tabulated review. J. Dermatol. Sci. 7:s109–s124.
Stenn, K.S., Combates, N.J., Eilertsen, K.J., Gordon, J.S., Pardinas,
J.R., Parimoo, S., Prouty, S.M. (1996) Hair follicle growth controls.
Dermatol. Clin. 14:543.
Sun, T.-T., Costarelis, G., Lavkar, R.M. (1991) Follicular stem cells:
The bulge-activation hypothesis. J. Invest. Dermatol. 96:77s–78s.
Tarlow, J.K., Clay, F.E., Cork, M.J., Blakemore, A.I.F., McDonagh,
A.J.G., Messenger, A.G., Duff, G.W. (1994) Severity of alopecia
areata is associated with a polymorphism in the IL-1 receptor
agonist gene. J. Invest. Dermatol. 103:387–390.
Telegdy, E., Lutz, G., Mojzes, J., Kreysel, H.W., Schneider, I. (1994)
Cytokine expression on keratinocytes in alopecia areata. Exp.
Dermatol. 3:147.
Vaux, D.L., Strasser, A. (1996) The molecular biology of apoptosis.
Proc. Natl. Acad. Sci. U.S.A. 93:2239–2244.
Wagner, S.L., Lau, A.L., Cunningham, D.D. (1989) Binding of protease
nexin-1 to the fibroblast surface alters its target proteinase specific-
ity. J. Biol. Chem. 264:611–615.
Wintzen, M., Yaar, M., Peter, J., Burbach, H., Gilchrest, B.A. (1996)
Proopiomelanocortin gene product regulation in keratinocytes. J.
Invest. Dermatol. 106:673–678.
Wood, L.C., Elias, P.M., Calhoum, C., Tsai, J.C., Grunfeld, C., Fein-
gold, K.R. (1996) Barrier disruption stimulates IL-1 alpha expres-
sion and release from a pre-formed pool in murine epidermis. J.
Invest. Dermatol. 106:397–403.
Yu, D.-W., Yang, T., Sonoda, T., Gaffney, K., Jensen, P.J., Dooley, T.,
Ledbetter, S., Freedberg, I.M., Lavker, R., Sun, T.-T. (1995) Message
of nexin 1, a serine protease inhibitor, is accumulated in the
follicular papilla during anagen of the hair cycle. J. Cell Sci.
108:3867–3874.
564 SEIBERG ET AL.
