878
Integral Lipids of H u m a n Hair
Philip W. W e d z * a n d D o n a l d T. D o w n i n g
The Marshall Dermatology Research Laboratories,Department of Dermatology, Universityof Iowa College of Medicine, Iowa City,
Iowa 52242
It has long been recognized that hair is coated with
nonpolar lipids originating in the sebaceous glands,
and recently it has been shown that hair also contains
cholesterol sulfate and small amounts of ceramldes,
similar to those found in the keratinized portion of the
epidermis. In the present study, it is demonstrated
that significant amounts of several additional lipids
are tightly associated with hair in such a way as to be
highly resistant to solvent extraction.
These integral hair iipids included cholesterol sul-
fate (3.3 mg/g of extracted hair), cholesterol {0.6 mg/g),
fatty alcohols {0.2 mg/g) and free fatty acids (4.3 mg/g).
The principal fatty acid, comprising 40% of the total
fatty acids, was identified as 18-methyl-eieosanoic acid
by cochromatography with authentic standard on gas-
liquid chromatography (GLC) and by mass spectrometry
(MS).
Lipids 23. 878-881 I1988).
Human scalp hairs consist of a central medulla of loosely
connected keratinized cells surrounded by more-tightly-
packed, 80-100 w-long cortical cells with 6 to 10 outer
layers of flattened, rectangular cuticle cells (1). Late in
the sequence of differentiative events leading to for-
mation of the mature cuticle and cortex cells, a multi-
layered material is deposited between the bounding
membranes of the individual cells to form membrane
complexes (2-4). Although it is chemically undefined,
the intercellular material (3,4) has been referred to as a
cement substance (3). The chemical nature of the cell
periphery in hair, like that of the cornified epidermal
cell, also has been somewhat of a puzzle because hair
does not contain the usual membrane-forming phospho-
lipids (5).
Associated with the hair follicles are sebaceous glands
that provide abundant nonpolar lipids to the skin sur-
face and the hair (6,7). While these sebaceous lipids
may serve to lubricate and waterproof, they are not
membrane-forming lipids. Recently, it has been demon-
strated that hair from several species of mammals con-
tains cholesterol sulfate and several series of ceramides
similar to those found in the keratinized portion of the
epidermis, or stratum corneum, and it has been postulated
that these lipids are of structural significance in hair
(5). However, the amounts of these polar lipids were
low, and it was uncertain as to whether they provided
enough material to account for the observed membranes.
More recently, the occurrence of significant amounts
of covalently bound lipid in the stratum corneum has
been demonstrated (8). This material consists mainly of
an unusual r ester-linked to a
*To whom correspondence should be addressed at 270 Medical
Laboratories, Dept. of Dermatology, University of Iowa College
of Medicine, Iowa City, IA 52242 USA
Abbreviations: ECL, equivalentchain lengths; GLC,ga~liquidchro-
matography; MS, mass spectrometry; TLC, thin-layer chro-
matography.
LIPIDS,Vol. 23, No. 9 ('1988)
cross-linked protein envelope to form what appears like
a plasma membrane (8,9). It is actually an integral part
of the cell envelope and can be removed from the cell
surface only after saponification (8,9).
The purpose of the present investigation was to
determine whether human hair, a keratinized epidermal
appendage, contains covalently bound lipids as found
in the stratum corneum. To answer this question, one
pooled sample of hair clippings from several people and
samples from three individuals were obtained. After
extensive extraction with chloroform]methanol mixtures
to remove all of the extractable lipids, the hair samples
were saponified and reextracted. Additional lipids recov-
ered in this way included major amounts of cholesterol
sulfate (3.3 mg/g extracted hair) and free fatty acids
(4.3 mg/g) and small amounts of cholesterol, fatty alco-
hol and several unidentified components. The major
fatty acid, comprising 40% of the total fatty acids,
proved to be 18-methyl-eicosanoic acid.
MATERIALS AND METHODS
Preparation of delipidized hair. Samples of human hair
were repeatedly rinsed with distilled water. They were
then extracted, first with methanol and then with chloro-
form]methanol 2"1, 1:1 and 1:2, for 2 hr each, at room
temperature. Each of the chloroform/methanol extractions
was then repeated for 24 hr. Finally, the hair was
extracted with chloroform in a Soxhlet extractor, con-
tinuously for 24 hr. The extracted hair was thoroughly
dried en vacuo and weighed. One large sample (100 g)
of pooled hair from several individuals and smaller
(3.5-8.9 g) hair samples from three individuals were
processed by this method.
Saponification and recovery of constitutive lipids.
Extracted hair samples (3-5 g after extraction) were
saponified by heating for 2 hr at 60 C in 200 ml 1 M
NaOH in 90% methanol. The reaction was cooled to
room temperature, and 88 ml of water and 360 ml of
chloroform were added. The mixture was shaken and
the phases were separated in a separatory funnel. The
lower chloroform layer was transferred to a flask, and
the upper phase and hair residue were acidified by
addition of 50 ml of 6 N HC1. The acidified upper phase
was extracted with an additional 360 ml of chloroform,
and the combined chloroform extracts were taken to
dryness with a rotary evaporator. The dried residue
was dissolved in 5 ml chloroform:methanol, 2:1, and
washed with 1 ml of 2 M KC1 containing 0.1 M HC1.
The lower chloroform phase then was filtered through
a solvent-resistant, 0.5-micron millipore filter, and the
filtrate was collected and taken to dryness in a tared
glass tube. The wt of the dried lipid was determined,
and chloroform:methanol was added to produce a solu-
tion containing 20 mg lipid per ml.
Thin-layer chromatography (TLC}. Analytical TLC
was done on 0.25-mm layers of Silica Gel G (E.M. Rea-
gents, Darmstadt, West Germany} on 20 • 20 cm glass
plates. Samples were applied 2 cm from the bottom
 879
INTEGRAL LIPIDS OF HAIR

edge of the plate, which was developed first to 9 cm GLC with the CP-SIL 88 column, both isothermally at
with chloroform/methanol/acetic acid/water (40:10:1:1), 160 C and with the temperature program described
and then to 20 cm with hexane/ether/acetic acid (70:30:1). above. Standard fatty alcohol acetates included C14:0,
After drying, the plate was sprayed with 50% sulfuric C16:0, C18:0 and C20:0.
acid and charred by slowly heating to 220 C. The charred Mass spectrometry. F a t t y acid methyl esters were
chromatograms were quantitated by photodensitometry separated on a 25-m CP-SIL 5 quartz capillary column
(10) with a Shimadzu model CS-930 photodensitometer. operated with a temperature program, starting at 150
Known amounts of stearic acid, stearyl alcohol, ch~ C and increasing 5 C/min, until a final temperature of
lesterol and cholesterol sulfate were used as standards. 250 C was attained. Mass spectra were recorded with a
F a t t y acids, fatty alcohols and the sterol fraction Nermag R 10-10 C mass spectrometer with an ionizing
were isolated by preparative TLC on 0.5-mm Silica Gel voltage of 70 eV. A range of mass/charge from 40
H (E.M. Reagents) using a mobile phase of hex- through 600 was examined.
ane/ether/acetic acid {70:30:1). Cholesterol sulfate was
isolated using chloroform/methanol/acetic acid/water RESULTS
(40:10:1:1) as the mobile phase. Composition of constitutive hair lipids. A typical densit~
The isolated fractions, as well as the solvolysis product meter tracing of a thin-layer chromatogram of the lipids
derived from cholesterol sulfate, were checked for purity released by saponification of previously extracted hair
by TLC with solvent systems of hexane/ether/acetic is presented as Figure 1. As described below, the most
acid (70:30:1), chloroform]methanol/acetic acid (190:9:1), prominent components were identified as free fatty
chloroform]methanol]water (40:10:1) and chloroform/meth- acids and cholesterol sulfate, and minor components
anol]acetic acid/water (40:10:1:1). Each fraction was judged included cholesterol and fatty alcohols. Some lipid ma-
to be chromatographically pure. terial appearing between Rf 0.2-0.4 was not identified
Acetylated sterols, acetylated fatty alcohols and fatty and variable amounts of origin material not thought to
acid methyl esters were purified by preparative TLC be lipid were sometimes present. The quantitative results
using a mobile phase of toluene. of such TLC analyses are summarized in Table 1.
Chemical procedures. The isolated fatty acid fraction
was converted to methyl esters by treatment with 10%
BC13 in methanol at 50 C for 1 hr. The reaction mixture
was dried under nitrogen. E
Cholesterol sulfate was solvolyzed overnight by treat-
ment with 10% BCI3 in methanol at 55 C. The sample A
was dried under nitrogen.
The free sterol, fatty alcohol and the sterol obtained
after solvolysis of cholesterol sulfate were acetylated
by treatment with acetic anhydride in pyridine at room
temperature for 2 hr. Excess reagents were removed by
evaporation under a gentle s t r e a m of nitrogen.
Gas-liquid chromatography (GLC). The fatty acid B C
methyl esters were examined by GLC on a 50 m CP-
SIL 88 quartz capillary column (Chrompack, Inc., Bridge-
water, NJ). An initial temperature of 160 C was main-
tained for 5 min, after which the temperature was in- ' 0:2 0'.4 Rf 0:6 0:8 '
creased linearly at a rate of 5 C/min, until 220 C was
reached. The final temperature was maintained until no FIG. 1. Densitometer tracing of a thin-layer chromatogram of the
further peaks eluted. The fatty acid methyl esters also constitutive iipids of hair. The chromatogram was developed to
were examined isothermally at 160 C and 175 C on the Rf 0.30 with chloroform/methanol/water/acetic acid {40:10:1:1) and
CP-SIL 88 column, as well as at 200 C on a nonpolar then to the top with hexane/ether/acetie acid {70:30:1). The charred
chromatogram was scanned with a Shimedzu model CS-9~0 photo.
25-m BP-1 quartz capillary column (Scientific Glass denm'tometer. Peak identification: A, cholesterol sulfate; B, unidenti-
Engineering, Inc., Austin, TX). fied; C, cholesterol; D, fatty alcohols; E, fatty adds.
Standard fatty acid methyl ester mixtures included
16:0, 18:0, 18:1, 18:2, 18:3, 20:0 (15A, NuChek Prep,
Elysian, MN); 14:0, 16:0, 16:1, 18:0, 18:1, 18:2, 18:3 TABLE 1
(CE1-62, NuChek Prep); 20:0, 20:1, 20:4, 22:1, 22:6
(I.r209, Applied Science, State College, PA); and fatty Constitutive Lipids of Human Hair
acids isolated from wool wax (ll). Mg lipid per g extracted hair
GLC of fractions isolated by argentation-TLC was
Fatty Cholesterol Fatty
used to confirm the identification of saturated and Sample acid sulfate Cholesterol alcohol
monoenoic fatty acid methyl esters. A 4.7 3.8 0.7 0.1
The acetylated sterols derived from both the free B 4.2 3.2 0.4 0.2
sterol fraction and the sterol sulfate fraction were exam- C 3.2 1.8 0.5 0.1
ined by GLC with the CP-SIL 88 column at 220 C and Pool 5.2 4.2 0.6 0.3
on BP-1 at 290 C. Authentic cholesterol acetate and Mean 4.3 3.3 0.6 0.2
cholestanol acetate were used as standards. SD 0.7 0.9 0.1 0.1
The fatty alcohol acetates also were examined by Weights of lipid were estimated by quantitative TLC.

LIPIDS,Vol. 28, No. 9 (1988)

880

P.W. WERTZ AND D.T. DOWNING

Identification of cholesterol sulfate. The identifica-
tion of cholesterol sulfate was b a s e d on its chro-
matographic and chemical properties. The isolated ma-
terial behaved like authentic cholesterol sulfate in t e r m s
of b o t h its mobility on T L C in several solvent s y s t e m s
and the pink-violet color produced on heating with
sulfuric acid prior to charring. Also, the p r o d u c t s pro-
duced b y hydrolysis and hydrolysis plus acetylation of
the isolated lipid b e h a v e d during TLC and charring
like authentic cholesterol and cholesterol acetate respec-
tively. Finally, the acetylated p r o d u c t g a v e rise to a
single m a j o r {96.6% of the total} p e a k when examined
b y G L C with either a polar (CP-SIL 88) or nonpolar
(BP-1) column. This m a j o r c o m p o n e n t c o m i g r a t e d with
authentic cholesterol acetate. Several minor components
were not identified.
Identification of free steros The isolated free sterol
was identified as cholesterol b y essentially the s a m e
criteria used to identify the sterol released b y solvolysis
of the sterol sulfate fraction. GLC of the sterol acetate
g a v e one m a j o r p e a k (96.9%) t h a t c o c h r o m a t o g r a p h e d
with cholesterol acetate on b o t h the polar and the
nonpolar columns.
Identification of fatty alcohols. The f a t t y alcohols
had the s a m e mobility as authentic f a t t y alcohols on
T L C with mobile phases of hexane/ether/acetic acid
(70:30:1) and chloroform/methanol/acetic acid (190:9:1).
Like the free sterol, acetylation yielded a single less
polar p r o d u c t t h a t m i g r a t e d on TLC like authentic
f a t t y alcohol acetate.
E x a m i n a t i o n b y G L C revealed a complex mixture.
A l t h o u g h the m a j o r c o m p o n e n t s accounting for 63.3%
of the t o t a l were identified as m e m b e r s of the straight-
chain s a t u r a t e d series b y comparison with s t a n d a r d s
and construction of plots of carbon n u m b e r vs log
retention t i m e f r o m isothermal c h r o m a t o g r a m s , the
remainder consisted of components with fractional equiv-
alent chain lengths (ECL). The f a t t y alcohol acetates
were identified as either s a t u r a t e d (86%) or u n s a t u r a t e d
(14%) after fractionation of the acetates on silver nitrate-
i m p r e g n a t e d silicic acid. This information is summa-
rized in Table 2.
Identification of fatty acids. The f a t t y acids were
converted to m e t h y l esters and examined b y GLC and
b y G L C - m a s s s p e c t r o m e t r y (MS). A r e p r e s e n t a t i v e
c h r o m a t o g r a m is shown in Figure 2, and the composi-
tion is s u m m a r i z e d in Table 3.
The m a j o r f a t t y acid was identified as 18-methyl-
eicosanoic acid on the basis of its G L C behavior and its
m a s s spectrum, presented as Figure 3. E C L values of
20.74 and 20.67 were determined from isothermal chro-
m a t o g r a m s obtained with CP-SIL 88 and BP-1, respec-
tively. A value of 20.72 would be expected for 18-methyl-
eicosanoic acid (12). Furthermore, on b o t h the polar
and nonpolar columns, the f a t t y acid m e t h y l ester in
question c o c h r o m a t o g r a p h e d with the C-21 anteiso
methyl-branched f a t t y acid m e t h y l ester derived f r o m
wool wax (11).
21ai

TABLE 2

Composition of Alcohol Acetates

ECL Chain % ECL Chain %
FIG. 2. Gas-liquid chromatogram of methyl esters prepared from the
12.00 12:0 0.1 20.30 21:0br 0.2 fatty add fraction of hair constitutive lipids. The methyl esters were
13.00 13:0 0.1 20.38 21:0br 0.2 chromatographed on a CP~SIL 88 quartz capillary column with a
14.00 14:0 4.7 20.41u 20:1 6.0 temperature program as expl~ned under Materials and Methods.
15.00 15:0 2.1 20.50u 20:1 1.5 The stralght~,hMned saturated species and the 21~wbon anteiso-
15.53 16:0br 0.3 20.56 21:0br 1.4 branched component are indicated.
16.00 16:0 10.6 20.71 21:0br 4.0
16.30 17:0br 0.2 21.00 21:0 1.8
16.40 17:0br 0.1 21.08 22:0br 0.3 TABLE 3
16.53 17:0br 0.5 21.43 22:0br 0.2
16.70 17:0br 2.6 21.57 22:0br 2.5 Composition of Covalently Bound Fatty Adds
17.00 17:0 1.1 22.00 22:0 5.2
17.53u 17:1 0.6 22.28u 22:1 2.7 Chain Wt %
18.00 18:0 17.0 22.33u 22:1 2.6 14:0 0.8
18.40u 18:1 0.4 22.43 23:0br 0.7 15:0 1.1
18.50u 18:1 0.2 22.49 23:0br 0.2 16:0 18.3
18.54 19:0br 0.9 22.59 23:0br 0.4 16:1 1.9
18.71 19:0br 1.2 22.72 23:0br 1.8 17:0br 5.5
19.00 19:0 2.0 23.00 23:0 1.7 17:0 1.9
19.49 20:0br 0.1 23.36 24:0br 0.7 18:0 7.0
19.55 20:0br 2.3 23.59 24:0br 1.8 18:1 3.9
20.00 20:0 11.6 24.00 24:0 5.3 18:0br 8.4
Monounsaturated species, are designated by a u following the 20:0 2.3
E C L value, as well as by the N:I designation of the chain struc- 20:0br 3.5
ture. It is not known if any of the unsaturated chains also are 21:0 anteiso 40.5
branched. For the saturated species, branching is indicated by a others 4.9
br at the end of the chain assignment. Several saturated species indicated by br, appear to be branched.

LIPIDS, Vol. 23, No. 9 (1988)

 881

INTEGRAL LIPIDS OF HAIR

74

O
c3 OCH3

M
C7 199 M-43

f
I.
'
I ,,III ,L,I~
I"" I ' I
,I,.I ,I
l ;---r"I'-T"T'-;"T"'"'T
w ............
I ,, , , , ~s,
"~"T'"F'-I ' l ' ! ' I: ' I ' ~ ' I:' I 'III ' IIL' I II I' '-'I]"-"II ' I"~II
L"~'-,-,s
' ~h"r"J-"--l"r
,,
III' I"' I ' I ' 1 1 I '""l
40 bO flO 100 120 140 lbO IBO 200 220 240 2bO 280 300 320 340 3bO 3B0 400

FIG. 3. Electron impact mass spectrum of methyl-18-methyl~icosanoate. The molecular ion ~M) and other significant fragments are indicated.

The identification was confirmed b y the mass spec-
t r u m (Fig. 2). The spectrum included a molecular ion at
m/e 340, consistent with the formula C22H4402. The
spectrum also included a number of features charac-
teristic of f a t t y acid methyl esters. These included a
base peak at m/e 74, and prominent fragments at m]e
87 and 143. Also, characteristic chain-degradation frag-
ments (C3-C8) were present. Of particular significance
are the peaks at M-29 and M-31 (13). M-31 is present in
all methyl esters and is t h o u g h t to represent formation
of an acylium ion. The M-29 represents loss of ethyl,
and is more prominent t h a n M-31 only in cases of
anteiso m e t h y l branching. In the p r e s e n t spectra,
M-29/M-31 -- 2.2.
DISCUSSION
The present results demonstrate t h a t human scalp hair
contains constitutive lipids t h a t are not removable b y
exhaustive extraction with chloroform/methanol mix-
tures. These lipids, t h a t together represent 0.7-1.3% of
the weight of hair, were recovered only after t r e a t m e n t
with alkaline methanol. The major lipid class recovered
in this way is f a t t y acid, which is present at a level of
4.3 m g per g of hair. This f a t t y acid is likely attached
to the cell surface through ester or thioester linkages.
Cholesterol sulfate is the second most abundant of the
hair constitutive lipids and constitutes 3.3 m g of each
g of hair. The nature of the strong interaction of this
lipid with hair is uncertain; however, it m a y be present
in the form of an insoluble salt rather than a covalent
adduct. Cholesterol and f a t t y alcohol, both relatively
minor constitutive lipids, could be esterified to acidic
groups at the cell surface.
Recently, covalently bound lipids have been identi-
fied in mammalian epidermis (8,9,14). In contrast to the
present findings, the major bound lipid found in the
epidermis is an unusual ceramide consisting of a long-
chain {30-34 carbon) r acid in amide linkage
with sphingosine. This is the predominant bound lipid
in the s t r a t u m corneum, where it is t h o u g h t to com-
prise a lipid envelope on the exterior of each keratinized
cell. By analogy, the constitutive lipids of hair also may be
bound to the cell surface to make a pliable but environ-
mentally resistant lipid coat or envelope.
Perhaps the most striking result of the present inves-
tigation was the finding t h a t 40% of the f a t t y acid
bound to hair is 18-methyl-eicosanoic acid. This anteiso
methyl-branched f a t t y acid is present in small amounts
in human sebum and vernix caseesa {15) and in wool
wax (11), but is not abundant in any source other than
hair. Shorter 13-17 carbon anteiso f a t t y acids are major
lipid components of some microorganisms (16-18), where
it is t h o u g h t t h a t they are used, instead of u n s a t u r a t e d
f a t t y acids, to make fluid membranes. I t seems feasible
t h a t the 21-carbon anteiso f a t t y acid in hair also m a y
serve to fluidize membranes. In this capacity, the
branched-chain acid would have an advantage over
u n s a t u r a t e d f a t t y acids in being resistant to oxidative
damage on long exposure to the atmosphere.
ACKNOWLEDGMENTS
This study was supported in part by Grants AR 32374 and AR
01610 from the United States Public Health Service.
REFERENCES
1. Chapman, R.E. (1986) in Biology of the Integumen~ (Bereiter-
Hahn, J., Matoltsy, A.G., and Richards, K.S., eds.) pp. 293-311,
Springer-Verlag, Berlin, West Germany.
2. Birbeck, M.S.C., and Mercer, E.H. (1956)Nature 178, 985-986.
3. Rogers, G.E. {1959)Ann. N.Y. Acad. Sci. 83, 378-399.
4. Parakkal, P.F., and Matoltsy, A.G. {1964)J. Invest. DermatoL
43, 23-34.
5. Wix, M.A., Wertz, P.W., and Downing, D.T. (1987) Comp.
Biochen~ Physiol. 86B, 671-673.
6. Nicolaides, N., Fu, H.C., and Rice, G.R. (1968) J. Invest.
Dermatol. 51, 83-89.
7. Stewart, M.E., Downing, D.T., and Strauss, J.S. {1983) De~
matol. Clin. 1, 335-344.
8. Wertz, P.W., and Downing, D.T. {1987) Biochim. Biophys.
Acta 917, 108-111.
9. Swartzendruber, D.S., Wertz, P.W., Madison, K.C., and
Downing, D.T. {1987)J. Invest. DermatoL 88~ 709-713.
10. Downing, D.T. (1968)J. Chromatogr. 38, 91-99.
11. Downing, D.T., Kranz, F.H., and Murray, K.E. {1960)Aust.
J. Chem. 13, 80-94.
12. Nicolaides, N., and Apon, J.M.B. (1976) Lipids 11, 781-790.
13. Abrahamsson, S., Stallberg-Stenhagen, S., and Stenhagen,
E. (1963) Prog. Chem. Fats Other Lipids 7, 1-164.
14. Wertz, P.W., and Downing, D.T. (1986) Biochem. Biophys.
Res. Commun. 137, 992-997.
15. Nicolaides, N., Fu, H.C., Ansari, M.N.A., and Rice, G.R.
(1972) Lipids 7, 506-517.
16. Zuneda, M.C., Guillenea, J.J., Dominguez, J.B., Prado, A.,
and Goni, F.M. {1984)Lipids 19, 223-228.
17. Song, Y., Sawa, T., Tsuvhiya, M., Kondo, S., Hamada, M.,
and Umezawa, H. {1981)J. Antibiot. Tokyo 34, 980-983.
18. Kaneda, T., and Smith, E.J. {1980) Can. J. Microbiol. 26,
893-898.
[Received April 13, 1988; Revision accepted June 25, 1988]

UPIDS, VoL 23, No. 9 (1988)