Content uploaded by Elizabeth R Sharlow
Author content
All content in this area was uploaded by Elizabeth R Sharlow on Sep 19, 2018
Content may be subject to copyright.
Exp Dermatol 2001: 10: 405–413 Copyright CMunksgaard 2001
Printed in Denmark ¡All rights reserved
E
XPERIMENTAL
D
ERMATOLOGY
ISSN 0906-6705
Soymilk reduces hair growth and hair follicle
dimensions
Seiberg M, Liu J-C, Babiarz L, Sharlow E, Shapiro S. Soymilk reduces
M. Seiberg, J.-C. Liu,
hair growth and hair follicle dimensions.
L. Babiarz, E. Sharlow and
Exp Dermatol 2001: 10: 405–413. CMunksgaard, 2001
S. Shapiro
Johnson & Johnson – Consumer Products
Abstract: We have recently shown that soybean-derived serine protease
Worldwide, Skin Research Center,
inhibitors and soybean extracts alter skin pigmentation, suggesting that
199 Grandview Rd, Skillman, NJ 08558, USA
soymilk could be used as a natural alternative to skin lightening. The
present studies were initiated to examine the possible effect of STI, BBI
and soymilk on hair pigmentation. Interestingly, these agents were found
to affect not only hair pigmentation, but also the rate of hair growth, the
Key words: hair growth – hair follicle –
dimensions of the hair follicle and hair shaft, and the appearance of the
soymilk – STI
hair. The studies presented here provide first evidence, at the morphological
and histological level, that soymilk and the soybean-derived serine pro-
Miri Seiberg, Skin Research Center, J&J CPWW,
199 Grandview Rd, Skillman, NJ 08558, USA
tease inhibitors could be used as effective agents for hair care and manage-
e-mail: MSEIBER/CPCUS.JNJ.COM
ment. These agents could reduce the rate of hair growth, decrease hair
shaft dimensions and alter the pattern of melanogenic gene expression.
Accepted for publication 27 February 2001
Introduction
The hair follicle is an epithelial structure that un-
dergoes cycles of active growth (anagen), re-
gression (catagen) and rest (telogen) (1). During
anagen the hair follicle grows down into the der-
mis, forming a complex layered structure with a
pigmented shaft. During catagen the hair follicle
shortens, and in telogen the epithelial cells are
resting while the hair shaft remains inside the short
follicle. The morphological changes throughout
the hair cycle are well-documented (2), but the
regulation of the different phases of this cycle is
not completely understood. The synchronized hair
growth mouse model (3, 4) has been extensively
utilized to better understand the regulation of hair
growth and cycling.
One main function of mammalian hair is to pro-
vide environmental protection. However, that
function has been lost in humans, in whom hair is
kept or removed for social and cosmetic purposes.
Many procedures are used to remove unwanted
hair, from simple inexpensive home treatments like
shaving, plucking and waxing, to expensive and
potentially time-consuming methods like electroly-
sis, laser, light therapies and therapeutic antian-
drogens. These methods differ not only in the dur-
ation of hair removal, their price range and their
405
pain and discomfort levels, but also in their poss-
ible undesired effects (reviewed in (5)). Shaving
may result in nicks and cuts in the skin’s surface,
may increase the risk of infection, may leave a per-
ception of an increase in the rate of hair growth,
and may also leave undesirable stubble. Plucking
causes pain and discomfort, and often results in
poor removal of short hair. While electrolysis may
keep an area free of unwanted hair for a prolonged
period of time, the process is expensive and painful
and may further result in scarring. Promising clin-
ical results have been obtained with laser treat-
ments, but not without changes in skin texture and
pigmentation. Several unwanted side effects, such
as liver toxicity or effects on muscularity often ac-
company the use of antiandrogens. Alternatives to
hair removal are popular hair cosmetic treatments
like hair dying or bleaching, used to reduce hair
visibility in desired body parts. These treatments,
which are sometimes irritating, are not always sat-
isfying, as the emerging portions of the hair shafts
are always darker than the already treated parts.
The Kunitz-type trypsin inhibitor (soybean tryp-
sin inhibitor, STI), and the Bowman–Birk protease
inhibitor (BBI) are two serine protease inhibitors
isolated from soybeans (reviewed in (6, 7)). STI in-
hibits the proteolytic activity of trypsin (6), and
BBI inhibits trypsin and chymotrypsin (7, 8). Re-
Seiberg et al.
cently we demonstrated that soymilk, STI and BBI
inhibit PAR-2 activation and thus induce skin de-
pigmentation (9). The studies described here were
designed to explore the potential use of these
agents in the inhibition of hair pigmentation. Our
data demonstrate that STI, BBI and soymilk re-
duce not only hair pigmentation, but also the rate
of hair growth and the final dimensions of the hair
shaft. These data suggest that such agents may
serve as an inexpensive, natural alternative treat-
ment for undesired hair growth.
Materials and methods
Chemicals and test materials
All chemicals were from Sigma (St Louis, MO,
USA), unless otherwise described. STI and BBI
were dissolved in GDL liposomes (20 mg/ml, in
PBS) prepared according to (10) comprising gly-
cerol dilaurate/cholesterol/polyoxyethylene-10-ste-
aryl ether/polyoxyethylene-9-lauryl ether at a ratio
of 37.5:12.5:33.3:16.7. Soymilk was prepared as a
10% solution in deionized water. Soybeans (100 g)
(Oriental Mascot Soybeans, imported from China,
NY, NY) were hydrated overnight in 1 liter of water.
Soybeans were rinsed in water and then processed in
1 liter of water using a standard juice extractor. The
milk was collected and filtered through cheesecloth.
The preservative phenoxyethanol (Phenonip, NIPA
Hardwicke Inc., Wilmington, DE, USA) was added
as 1% of the total volume and the soymilk was
stored at 4 æC. A thickener, Sepigel 305, 3% (poly-
acrylamide/Laureth-7/C
13–14
Isoparrafin, from
Seppic Inc., Fairfield, NJ, USA) was occasionally
added to the soymilk in order to ease the in vivo ap-
plication. This addition resulted in no changes rela-
tive to soymilk treatment in its liquid form (not
shown). Flavosterone SE (Ichimaru, Japan), which
contains about 0.1% pure isoflavones, was used to
supplement the soymilk when indicated. Heat de-
natured soymilk was heated to 100 æC for 10 min.
The lack of trypsin inhibitory activity, which implies
the loss of STI and BBI activity, was demonstrated
(9) using EnzChek
TM
protease assay kit, following
manufacturer’s instructions (Molecular Probes, Eu-
gene, OR, USA).
Animals and hair growth induction
C57Bl/6 and C3H female mice, 7–8 weeks old, were
purchased from Charles River (Kingston, NY,
USA), and were housed in appropriately sized
cages in an environmentally controlled room with
a 12-h light–12-h dark photoperiod and supplied
with food and water ad libitum. Animal care was
based on the ‘‘Guide for the Care and Use of Lab-
oratory Animals’’, NIH Publication No. 85–23.
406
Animals were acclimated for a week, and studies
started at 8–9 weeks of age, when their hair was in
the telogen phase. Hair growth was induced by wax
depilation as previously described (4), or by com-
mercially available chemical depilatory creams,
and biopsies were taken from the center of the
mouse dorsum using standard procedures. Mice
were treated daily with 50 or 100 ml of test com-
pound or liposome vehicle alone. Each experiment
was repeated at least three times, with at least five
animals per each time point.
Histology and staining
Sections from biopsies taken at key time points
throughout the hair cycle, from identical sites on
the dorsum of the mice, were stained with Fon-
tana–Mason (F&M), using standard procedures
(11). F&M staining identifies silver nitrate reduc-
ing molecules. In skin, this non-specific stain ident-
ifies primarily melanin. At least six sections per bi-
opsy were processed and analyzed.
Image analysis and statistics
Digital pictures were taken using Hi-Scope Model
KH-2400R, with MX-MACROZ Zoom Lens (Hi-
rox Co. Ltd, Tokoyo, Japan). Histological images
were obtained using a Leitz Diaplan microscope
(Nurnberg, Germany). All images were captured
and analyzed with Image Pro Plus 3.0 software
(Media Cybernetics, Silver Spring, MD, USA),
and data were graphed using SigmaPlot
A
5.0
(SPSS Science, Chicago, IL, USA). Statistical
analysis (paired t-test) was performed using
SigmaStat
A
2.0 (SPSS Science) software. For hair
follicle size and stage analysis, slides from six sep-
arate experiments, at least three mice per group
per experiment, two independent sections per
mouse, at least 10 full follicles per mouse, were
measured. Measurements count hair shaft thick-
ness (at epidermal level) and the follicle diameter
at Auber’s critical line at ¿400 magnification.
There was no difference in these parameters be-
tween untreated and PBS plus thickener-treated
animals (not shown).
SDS-PAGE and Western Blotting
Biopsies from treated skins were homogenized and
lysed in RIPA buffer [1% nonidet-P40, 0.5% sodium
deoxycholate, 0.1% sodium dodecyl sulfate, and
Complete
TM
protease inhibitors (Boehringer
Mannheim, Indianapolis, IN, USA)] in PBS. Lys-
ates (10 mg per lane) were separated using 10% SDS-
PAGE gels and proteins were analyzed using stan-
dard procedures (12) and enhanced chemilumi-
Soymilk reduces hair growth and hair follicle dimensions
nescence (ECL, Amersham, Arlington Heights, IL,
USA). Antibodies to tyrosinase (anti PEP-1) and to
TRP-1 (anti PEP-7) were a generous gift of Dr V.
Hearing and were used as described in (13, 14).
Equal protein loading on the gels was verified using
Ponceau S solution (Sigma, St Louis, MO, USA)
following manufacturer’s protocol (not shown).
Human studies
Human facial studies were performed with in-
formed consent, employing three male volunteers,
40–60 years old, Fitzpatrick skin type II, with no
history of skin cancer, no skin disease or unusual
skin reaction, and no topical treatments within a
period of 30 days before enrollment. Volunteers
shaved daily and applied soymilk (with or without
a thickener) to one side of their face immediately
after shaving. No other after-shave products were
used during the treatment period. Hair growth was
evaluated visually, and images were taken at differ-
ent time points. Hair shaft length, width and total
area were measures using computerized image
analysis. Two volunteers were also evaluated 4
weeks after the end of the treatment period. Hu-
man leg studies used the same criteria, except both
genders were included. Legs were wax depilated at
symmetrical sites, using commercial wax prepara-
tions, and soymilk was applied immediately after
waxing and daily afterwards for 4 weeks.
Figure 1. STI and BBI affect hair growth. C57Bl/6 (a–c) and C3H (d–f) mice were wax depilated to induce a new hair cycle, and
were treated daily with liposome vehicle (a, d), STI (b, e) or BBI (c, f). At day 18 of the hair cycle, a visual difference was observed
in hair appearance, as shown in images of skin biopsies (a–c) at ¿8 magnification. Histological analysis (d–f, F&M staining)
revealed that STI and BBI treatments lead to reduced hair follicle dimensions (day 8).
407
Results
STI and BBI delay hair growth and reduce hair
follicle size and pigmentation
To determine the effect of STI and BBI on hair
growth and pigmentation, we examined their effect
on the mouse hair cycle. Hair growth of C57Bl/6
and C3H mice, ages 8–9 weeks, was induced by
wax depilation of the animal’s back fur as de-
scribed in (4), and animals were treated daily with
STI or BBI (1%, in GDL liposomes), or with lipo-
somes alone. Animals were observed daily for hair
appearance and biopsies were analyzed for histo-
logical changes following these treatments. Visual
observations throughout the hair cycle indicated
that both STI and BBI delayed hair growth and
reduced hair shaft length, while the vehicle alone
had no effect. Fig. 1 shows that the hair of STI
(Fig. 1b) or BBI (Fig. 1c) treated C57Bl/6 mice was
shorter and appeared more ‘‘orderly organized’’
relative to the untreated control (not shown) or
liposome vehicle treatment (Fig. 1a), when exam-
ined at low magnification. Histological analysis
confirmed these observations. Fig. 1d–f shows that
at day eight of the hair cycle STI (Fig. 1e) and
BBI (Fig. 1f) reduced hair follicle size in C3H mice,
relative to the liposome control (Fig. 1d). The in-
hibitory effect of STI and BBI was identical in
both mouse strains, but hair appearance was more
homogeneous using STI, relative to BBI.
Seiberg et al.
Soymilk delays hair growth and reduces hair
follicle size and pigmentation
To test the hypothesis that soybean extracts can
induce similar effects, mice induced for a new hair
cycle were treated daily with fresh soymilk and ob-
served as indicated earlier. Soymilk treated mice
showed delayed and reduced hair growth and their
hairs were visibly thinner, more ‘‘directionally or-
ganized’’, and smoother to touch, relative to un-
treated controls (Fig. 2a, day 21 of the hair cycle).
The treated mice did not show skin darkness at
days 7–8 of the hair cycle, as expected, and hair
shafts were not visible at days 11–12 as in the con-
trol animals. In average, the hair cycle of the soy-
milk treated mice was delayed by 3–6 days. Histo-
logical analysis confirmed these observations, and
documented a delayed and shorter hair cycle with
Figure 2. Soymilk affects hair growth. C57Bl/6 mice were wax
depilated to induce a new hair cycle, and were treated daily
with fresh soymilk. A visual difference was observed in hair
appearance (a, day 21 of the hair cycle), as the soymilk treated
mice (a, right) exhibited shorter, softer and more orderly organ-
ized fur, relative to the untreated controls (a, left). F&M stain-
ing of skin biopsies taken at days 4 (b, c), 7 (d, e) and 18
(f, g) of the hair cycle, demon-
strate delayed follicular devel-
opment, and reduced final
follicle dimensions and pig-
ment deposition in soymilk
treated mice (c, e, g) relative
to untreated controls (b, d, f).
408
smaller size follicles. Fig. 2 demonstrates delayed
follicular development at 4 days (Fig. 2b–c) and 7
days (Fig. 2d–e) of the hair cycle, and reduced final
follicle size (day 18, Fig. 2f–g) following soymilk
treatment. Table 1 documents the delayed and
shortened hair cycle, comparing the growth stages
of 300 control and treated follicles, from six differ-
ent experiments, at different days of the hair cycle.
Statistical analysis of hair follicle size was also
performed using histological data from six differ-
ent experiments. Hair shaft thickness and the di-
ameter of the hair bulb at Auber’s line were meas-
ured in follicles properly positioned within the sec-
tions, at ¿400 magnification, using computerized
image analysis. As shown in Table 2, Soymilk
treatment results in a statistically significant effect
on hair follicle dimensions. Hair shaft diameter
was reduced by an average of 42%, and the hair
bulb diameter was reduced by an average of 23.8%.
Similar effects were shown when chemically
depilated mice were treated with soymilk. Soymilk
treated shaved mice that were not synchronized for
their hair cycle, also showed slower hair growth,
reduced hair shaft length and nicer hair ‘‘appear-
ance’’, following soymilk treatment (not shown).
These data suggest that soymilk can reduce hair
growth and pigmentation regardless of the method
of hair removal, and not only when following a
synchronized hair growth induction. Heat de-
natured soymilk (100 æC, 10 min) or commercially
available pasteurized soymilk (Similac) had no ef-
fect on hair growth or appearance (not shown),
supporting the role of intact STI and BBI in the
inhibitory effect of soymilk.
Soymilk reduces melanogenic protein expression
The depigmenting effect of serine protease inhibi-
tors in epidermal equivalents containing melano-
cytes involves the reduced expression of TRP-1
mRNA, with no effect on tyrosinase mRNA ex-
pression. This suggests that the reduced TRP-1
level impairs tyrosinase stability, leading to re-
duced tyrosinase protein levels (15). Therefore, we
studied the effect of soymilk on tyrosinase and
TRP-1 protein levels throughout the mouse hair
cycle. Mice were treated as described above, and
samples were collected throughout the hair cycle
for protein analysis. As shown in Fig. 3, the
steady state of tyrosinase and TRP-1 proteins is
dramatically affected by the soymilk treatment.
Tyrosinase and TRP-1 protein levels are reduced,
and the duration of their expression throughout
the hair cycle is shortened. These two factors re-
duce hair pigmentation, due to the reduced level
and shorter duration of melanogenesis.
Soymilk reduces hair growth and hair follicle dimensions
Table 1.
Control Soy
Anagen Anagen Anagen Anagen Anagen Anagen
Stage I–II III–IV V–VI Catagen I–II III–IV V–VI Catagen
Day 2 198 102 – – 300 – – –
Day 4 3 235 62 – 204 96 – –
Day 5 – – 300 – – 253 47 –
Day 20 – – 161 139 – – 27 273
Table 2.
P
-value
Control (m)∫SD Soymilk (m)∫SD % decrease (
t
-test)
Shaft diameter 92.8∫21.8 53.26∫14.9 42.6 ⬍0.001
Auber’s line length 200.41∫27.40 153.33∫10.70 23.48 ⬍0.001
Long term treatment with soymilk reduces hair size
and pigmentation
Mice treated with soymilk for 5 and 7 months were
analyzed for hair appearance and hair shaft par-
ameters. As shown in Fig. 4a–b, soymilk treated
mice had shorter, thinner, and less dense fur, rela-
tive to untreated control. Histological analysis
demonstrated reduced thickness of the hair shaft
and reduced pigment deposition within the shaft
(Fig. 4c–d). No visual signs of irritation, scaling or
other negative effects were observed following the
5- or 7-month-long treatment, and no immune in-
filtrate or changes in skin architecture were ob-
served histologically (not shown).
STI and isoflavones affect hair growth via different
pathways
Since phytoestrogens were shown to reduce hair
growth in vitro (16), we studied the effect of the
Figure 3. Soymilk affects melanogenic
gene expression. Western blots of C57Bl/
6 mice skins harvested throughout the
hair cycle demonstrate reduced tyrosin-
ase and TRP-1 protein levels and dur-
ation of expression following soymilk
treatment. C57Bl/6 and C3H mice were
induced for a new hair cycle and were
treated daily with soymilk. Skin samples
were collected throughout the study for
protein analysis. Protein extraction and
Western blot analysis were performed
using standard procedures, using the
anti-tyrosinase antibody anti PEP1, and
the anti-TRP-1 antibody anti PEP7.
409
soybean-derived phytoestrogens, alone and in com-
bination with soymilk, on hair growth. Since the
soymilk preparations used in this experiment con-
tain about 0.001% isoflavones, we treated mice
throughout the hair cycle with SE flavosterone, a
soybean derived isoflavone extract (at about
0.001% isoflavones), soymilk (containing STI, BBI,
and about 0.001% isoflavones), or a combination
of both (containing STI, BBI and about 0.002%
isoflavones). To ease the application process, these
actives were formulated with a thickener (Poly-
acrylamide/Laureth-7/C
13–14
Isoparrafin, 3%). Fig.
5 shows the hair of C57Bl/6 mice following 3 weeks
of topical treatment. Both untreated control mice
and vehicle treated mice have long and ‘‘randomly
ordered’’ hair (Fig. 5a–b). As expected, soymilk re-
duced hair growth and led to a more uniform, dir-
ectionally ordered hair appearance (Fig. 5c). Iso-
flavones (0.001%) reduced hair growth, but not as
effectively as soymilk (Fig. 5d). A heat inactivated
soymilk preparation, with minimal or no trypsin
Seiberg et al.
Figure 4. Mice treated with soymilk for 5 months (b) have
shorter, thinner, and less dense fur, relative to untreated control
(a). Histological analysis (F&M staining) demonstrated re-
duced thickness of the hair shaft and reduced pigment depo-
sition within the shaft of soymilk treated mice (d), relative to
the untreated control (c).
inhibitory activity gave similar results to 0.001%
isoflavones (not shown). These observations sup-
port our data that STI and BBI contribute to the
reduced hair growth effect, and suggest that a dif-
ferent pathway than that of the estrogen receptor
is involved in this process. The combination of soy-
milk supplemented with 0.001 soybean-derived iso-
flavones reduced hair growth to a higher degree
than soymilk or isoflavones alone (Fig. 5e), further
supporting complementary roles for STI, BBI and
isoflavones in the hair inhibitory effect. Further
supplementation with isoflavones did not result in
additional effect (not shown).
Soymilk reduces human hair length and thickness
Three Caucasian males with dark facial hair who
shave daily treated the right side of their face with
soymilk, immediately after shaving, for 5 or 6
weeks. By the third week, and more noticeably by
the fourth week, the hair of the treated side was
visually lighter and felt smoother to touch. Digital
pictures taken throughout the treatment period
clearly demonstrate the reduced length and thick-
ness of the hair shafts at the treated side (see Fig.
6a–b). Since both sides of the face were shaved
daily, at the same time, and pictures of both sides
were taken at the same time, the difference in
length of the facial hair indicates slower growth
rate at the treated area. Computerized image
analysis of the facial hair length, thickness and
total area, following 4 weeks of soymilk treatment
documents a statistically significant difference in
all measured parameters, following soymilk treat-
410
ment (Fig. 6c). Four weeks after the end of the
treatment period, two of the volunteers were evalu-
ated for hair regrowth. At that time, there was no
difference between the treated and untreated sides
of their face (not shown), demonstrating a com-
plete reversibility of the soymilk effect.
Hair was wax-depilated of two symmetrical
areas of the medial part of the legs, below the knee,
in two individuals. One leg was treated daily, for 4
weeks, with soymilk. Visual observations indicate
slower hair growth on the treated site. Hair shafts
were minimally reduced in number, but were vis-
ibly shorter and thinner than those of the un-
treated site, as shown in Fig. 6d–e. These obser-
vations further confirm the effect of soymilk on
hair growth in humans, demonstrating that this ef-
fect is not related to the method of hair removal
or to the body part being treated.
Discussion
The inhibition of PAR-2 activation by synthetic
serine protease inhibitors (15, 17) results in skin
Figure 5. STI, BBI and soy isoflavones affect hair growth in
complementary pathways. C57Bl/6 mice were wax depilated to
induce a new hair cycle, and were treated daily with no treat-
ment (a), vehicle (b), soymilk (c), 0.001% soy-isoflavones (d),
and a combination of soymilk supplemented with 0.001
soybean-derived isoflavones (e). Hair pictures are shown at day
21 of the hair cycle. Soymilk treated mice (c) show shorter and
softer coat than isoflavone only treated mice (d, of similar iso-
flavone concentration as in the soymilk), indicating that isofla-
vones are not sufficient to reproduce the whole soymilk effect.
The combination of soymilk supplemented with 0.001 soybean-
derived isoflavones (e) had the highest inhibitory effect on hair
growth. Further addition of isoflavones (not shown) did not
enhance that effect.
Soymilk reduces hair growth and hair follicle dimensions
Figure 6. Soymilk reduces human hair length and thickness. A
Caucasian male with dark facial hair treated the right side of
his face with soymilk, immediately after shaving, for 5 weeks.
Digital pictures taken throughout the treatment period demon-
strate reduced growth rate and dimensions of the hair shafts at
the treated side (b), relative to the untreated one (a). A compu-
terized image analysis of the facial hair length, thickness and
total area, at 4 weeks of soymilk treatment documents a statisti-
cally significant difference (P⬍0.0001) in all measured par-
ameters, following soymilk treatment (c). Hair was wax-depi-
lated of two symmetrical areas of the legs, below the knee, in
one individual. One leg was treated daily, for 4 weeks, with
soymilk. Visual observations indicate reduced hair growth at
the treated site (e), and hair shafts were shorter and thinner
than those of the untreated site (d). Leg hair images (d, e, at 4
weeks’ treatment) were taken using a clear gel and a glass plate,
to compensate for leg curvature, therefore they do not represent
the direction of hair growth.
lightening, suggesting a new class of depigmenting
agents. Similarly, inhibition of the PAR-2 pathway
by soymilk leads to skin depigmentation, sug-
gesting that soymilk could be used as a natural
alternative to skin lightening (9). The present
studies were initiated to examine the possible effect
of STI, BBI and soymilk on hair pigmentation.
Interestingly, these agents were found to affect not
only hair pigmentation, but also the rate of hair
growth, the dimensions of the hair follicle and hair
shaft, and the appearance of the hair.
Soybeans were not utilized as food until precip-
itation and fermentation techniques were de-
veloped, because they produced serious gastric dis-
411
tress. These are due to the inhibitory activity of
STI and BBI, which block the action of trypsin
and other proteases needed for protein digestion
(18). It is the heat inactivation of STI and BBI
during soybean processing (reviewed in (19, 20))
that renders the soybean edible. It is important to
note that fresh soymilk, but not heat-denatured or
pasteurized soymilk, was the active inhibitor in our
hair growth studies. Since STI is heat labile, but
BBI is less affected by heat (6, 21), it is more likely
that STI is the major protein agent in soymilk that
affects hair growth, size and pigmentation.
Topical treatments with 17-beta-estradiol inhib-
ited hair growth in mice, while an estrogen recep-
tor antagonist initiated hair growth in this system
(22, 23). This treatment was more effective at the
site of application than at distant sites, indicating
a direct rather than systemic effect. These studies
implicate a skin-specific estrogen receptor path-
way in the regulation of the hair cycle (23). The
weak estrogenic effects of isoflavones are heavily
documented. Interestingly, the soy derived phy-
toestrogen genistein was shown to decrease hair
growth by 60%–80%, when hair follicles isolated
from scalp biopsies were treated in culture (16).
Our data clearly demonstrate a role for soy iso-
flavones in the regulation of hair growth. How-
ever, we also show that the soy-derived serine
protease inhibitors, STI and BBI, alone, affect
hair growth. Since these agents are not known to
act via the estrogen receptor, we suggest that they
affect hair growth via a different pathway. More-
over, the combinations of soy and isoflavones
tested show an additive effect, until saturation is
observed.
A serine protease inhibitor treatment of kera-
tinocyte-melanocyte co-cultures results in a de-
crease in TRP-1 gene expression, but in no changes
in tyrosinase gene expression (15). As TRP-1 is a
regulator of tyrosinase by stabilizing the protein
(24, 25), we suggested that the down regulation of
TRP-1 expression led to reduced tyrosinase half
life and activity, and reduced pigment production
(15). This suggests that while tyrosinase mRNA
levels were not affected by the serine protease in-
hibitor, its steady state protein level should be re-
duced. Our current hair studies support this sug-
gestion, demonstrating reduced tyrosinase and
TRP-1 protein levels following soymilk treatment.
These data could also imply that soymilk reduces
pigment deposition within the hair follicle via the
PAR-2 pathway. However, studies that induce
PAR-2 activation and enhance pigment deposition
are required to substantiate this hypothesis.
Using the synchronized hair growth mouse
model (3, 4) we showed that topical trypsin treat-
ment, immediately after depilation, induced cell
Seiberg et al.
death at the follicular papilla. This death signal,
which is independent of the proteolytic activity of
the protease, resulted in delaying hair growth and
pigmentation (26). Here we show that the trypsin
inhibitors STI and BBI also lead to delayed hair
growth, as well as to reduced follicle dimensions
and reduced pigment deposition within the hair
shaft. Since we could not detect apoptotic cells in
the papillae of soymilk, STI or BBI treated mice
(not shown), we suggest that STI and BBI do not
affect papillae cell death, but exert their effect on
hair growth and size via a different mechanism.
Humans keep or remove hair from different
body parts for social and cosmetic purposes. Un-
wanted hair is removed using home treatments or
professional services, or is bleached to reduce its
visibility. A safe and effective agent that could re-
duce hair growth, hair size and hair pigmentation
would add to the available home treatments, which
are not always satisfying. The finding that soymilk
and soymilk-derived serine protease inhibitors de-
lay hair growth and reduce hair shaft dimensions
and pigmentation provides a new concept in hair
growth and management, and could serve as an
effective natural way to manage unwanted hair
growth. The human studies presented here repre-
sent a positive proof-of-concept pilot study, which
led us to perform larger double blind studies. Pre-
liminary data of these studies (J.-C. Liu, in prep-
aration) indicate that treatment with soymilk-con-
taining formulations reduced the rate of hair
growth of women’s legs, affected the direction of
hair growth to look more homogeneous, and made
the hair softer, finer and less noticeable.
Hirsutism is a relatively frequent condition af-
fecting about 4% of women. Facial hirsutism often
interferes with personal and work activities, and
temporary hair removal is a major component in
the management of hirsute patients. Shaving is the
most frequently used temporary method for facial
hair, as plucking, waxing and depilatories are more
difficult to tolerate and care must be taken to avoid
folliculitis, pigmentary changes, and scarring. Cos-
metic cover-ups are usually used to hide cuts and
stubble (27, 28), and electrolysis and thermolysis
are used for permanent hair removal when af-
fordable (29). Daily treatment with soybean-de-
rived protease inhibitors or soymilk products
would be painless and of modest cost and could
serve to reduce hair growth and visibility in these
patients, enhancing their quality of life. Prelimi-
nary clinical data (J.-C. Liu, in preparation) indi-
cate that treatment with STI and soymilk-contain-
ing formulations significantly reduced women’s fa-
cial hair growth rate and visibility.
African-type hair is unique in its morphology –
a kinky hair shaft with variations in diameter. This
412
complex shaft structure creates the need for spe-
cialized grooming products and procedures to en-
sure that the African-type hair maintains cosmetic
desired properties. The addition of STI, BBI or
soymilk into hair care products could reduce this
complexity and make the African-type hair more
manageable, improving its appearance.
The studies presented here establish soymilk and
the soybean-derived serine protease inhibitors as
effective and inexpensive cosmetic agents for hair
care and management. These agents could reduce
the rate of hair growth, decrease hair shaft dimen-
sions and reduce hair pigmentation, resulting in
slower hair growth, softer and lighter hair shafts,
and more manageable hair care.
Acknowledgement
We would like to thank A. Harmon, A. Johnson and C. Paine
for technical assistance, Dr R. Gallagher for help with imaging
and Dr G. Cauwenbergh for fruitful discussions throughout
this study. Special thanks to Dr V. Hearing for the generous gift
of antibodies to tyrosinase and to TRP-1.
References
1. Panaretto B A. Gene expression of potential morphogenes
during hair follicle and tooth formation: A review. Repro-
duction Fertility Development 1993: 5: 345–360.
2. Chase H B. Growth of the hair. Physiol Rev 1954: 34: 113–
126.
3. Slominski A, Paus R. 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 1993: 101: 90S–97S.
4. Stenn K S, Paus R, Dutton T et al. Glucocorticoid effect on
hair growth initiation: A reconsideration. Skin Pharmacol
1993: 6: 125–134.
5. Olsen E A. Methods of hair removal. J Am Acad Dermatol
1999: 40: 143–155.
6. Birk Y. The Bowman–Birk inhibitor. Trypsin- and chymo-
trypsin-inhibitor from soybeans. Int J Pept Protein Res
1985: 25: 113–131.
7. Kennedy A R. The Bowman–Birk inhibitor from soybeans
as an anticarcinogenic agent. Am J Clin Nutr 1998: 68:
1406S–1412S.
8. Billings P C, Habres J M. A growth-regulated protease ac-
tivity that is inhibited by the anticarcinogenic Bowman–
Birk protease inhibitor. Proc Natl Acad Sci 1992: 89: 3120–
3124.
9. Paine C, Sharlow E, Liebel F, Eisinger M, Shapiro S, Seib-
erg M. An alternative approach to depigmentation by
Soybean extracts via inhibition of the PAR-2 pathway. JID
2001: (In press).
10. Niemiec S M, Ramachandran C, Weiner N et al. Influence
of nonionic liposomal composition on topical delivery of
peptide drugs into pilosebaceous units: An in vivo study
using the hamster ear model. Pharm Res 1995: 12: 1184–
1188.
11. Sheenan D C, Hrapckak B B, eds. Theory and practice of
histo-technology. St Louis: The C. V. Mosby Co., 1980:
222–227.
12. Bonifacino J S. Immunofluorescence staining. In: Current
protocols in cell biology on CD-ROM. (Teton Data Sys-
tems). Jackson, Wyoming: John Wiley & Sons, Inc., 1999.
Soymilk reduces hair growth and hair follicle dimensions
13. Jimenez M, Kameyama K, Maloy W L, Tomita Y, Hearing
V. Mammalian tyrosinase: biosynthesis, processing and
modulation by melanocyte stimulating hormone. Proc Natl
Acad Sci USA 1988: 85: 3830–3834.
14. Jimenez M, Kameyama K, Maloy W L, Hearing V. Specific
identification of an authentic tyrosinase clone. J Biol Chem
1989: 264: 3397–3403.
15. Seiberg M, Paine C, Sharlow E et al. The protease-activated
receptor 2 regulates pigmentation via keratinocyte-melano-
cyte interactions. Exp Cell Res 2000a: 254: 25–32.
16. Hoffmann R, Eicheler W, Wenzel E, Happle R. Interleukin-
1beta-induced inhibition of hair growth in vitro is mediated
by cyclic AMP. Invest Dermatol 1997: 108: 40–42.
17. Seiberg M, Paine C, Sharlow E et al. Inhibition of mel-
anosome transfer results in skin lightening. J Invest Derma-
tol 2000b: 115: 162–167.
18. Messina M J. Legumes and soybeans: overview of their nu-
tritional profiles and health effects. Am J Clin Nutr 1999:
70: 439S–450S.
19. Wallace G M, Bannatyne W R, Khaleque A. Studies on
the processing and properties of soymilk. II. Effect of pro-
cessing conditions on the trypsin inhibitor activity and the
digestibility in vitro of proteins in various soymilk prepara-
tions. J Sci Food Agric (England) 1971: 22: 526–531.
20. Kwok K C, Niranjan K. Effects of thermal processing on
soymilk. Int J Food Sci Technol 1995: 30: 263–265.
21. Liu K. Chemistry and nutritional value of soybean compo-
nents. In: Liu K, ed. Soybeans, chemistry, technology and
utilization. Gaithersburg, MD: Aspen publishers, 1999: 32–
35.
413
22. Oh H S, Smart R C. An estrogen receptor pathway regu-
lates the telogen–anagen hair follicle transition and influ-
ences epidermal cell proliferation. Proc Natl Acad Sci USA
1996: 93: 12525–12530.
23. Chanda S, Robinette C L, Couse J F, Smart R C. 17 beta-
estradiol and ICI-182780 regulate the hair follicle cycle in
mice through an estrogen receptor-alpha pathway. Am J
Physiol Endocrinol Metab 2000: 278: E202–E210.
24. Hearing V J, Tsukamoto K, Urabe K, Kameyama K, Mon-
tague P M, Jackson I J. Functional properties of cloned
melanogenic proteins. Pig Cell Res 1992: 5: 264–270.
25. Orlow S J, Zhou B K, Chakraborty A K, Drucker M, Pif-
ko-Hirst S, Pawelek J M. High-molecular-weight forms of
tyrosinase and the tyrosinase-related proteins: evidence for
a melanogenic complex. J Invest Dermatol 1994: 103: 196–
201.
26. Seiberg M, Wisniewski S, Cauwenbergh G, Shapiro S S.
Trypsin-induced follicular papilla apoptosis results in
delayed hair growth and pigmentation. Dev Dyn 1997: 208:
553–564.
27. Richards R N, Uy M, Meharg G. Temporary hair removal
in patients with hirsutism: a clinical study. Cutis 1990: 45:
199–202.
28. Young R, Sinclair R. Hirsutes. II: Treatment. Australas J
Dermatol 1998: 39: 151–157.
29. Wagner R F Jr. Physical methods for the management of
hirsutism. Cutis 1990: 45: 19–21.
