Investigation by imaging mass spectrometry of biomarker candidates for aging in the hair cortex.

Abstract

Human hair is one of the essential components that define appearance and is a useful source of samples for non-invasive biomonitoring. We describe a novel application of imaging mass spectrometry (IMS) of hair biomolecules for advanced molecular characterization and a better understanding of hair aging. As a cosmetic and biomedical application, molecules whose levels in hair altered with aging were comprehensively investigated.
Human hair was collected from 15 young (20±5 years old) and 15 older (50±5 years old) volunteers. Matrix-free laser desorption/ionization IMS was used to visualize molecular distribution in the hair sections. Hair-specific ions displaying a significant difference in the intensities between the 2 age groups were extracted as candidate markers for aging. Tissue localization of the molecules and alterations in their levels in the cortex and medulla in the young and old groups were determined.
Among the 31 molecules detected specifically in hair sections, 2--one at m/z 153.00, tentatively assigned to be dihydrouracil, and the other at m/z 207.04, identified to be 3,4-dihydroxymandelic acid (DHMA)--exhibited a higher signal intensity in the young group than in the old, and 1 molecule at m/z 164.00, presumed to be O-phosphoethanolamine, displayed a higher intensity in the old group. Among the 3, putative O-phosphoethanolamine showed a cortex-specific distribution. The 3 molecules in cortex presented the same pattern of alteration in signal intensity with aging, whereas those in medulla did not exhibit significant alteration.
Three molecules whose levels in hair altered with age were extracted. While they are all possible markers for aging, putative dihydrouracil and DHMA, are also suspected to play a role in maintaining hair properties and could be targets for cosmetic supplementation. Mapping of ion localization in hair by IMS is a powerful method to extract biomolecules in specified regions and determine their tissue distribution.

Full text

Page 1

Investigation by Imaging Mass Spectrometry of
Biomarker Candidates for Aging in the Hair Cortex
Michihiko Luca Waki1, Kenji Onoue1, Tsukasa Takahashi1, Kensuke Goto1, Yusuke Saito1, Katsuaki
Inami1, Ippei Makita1, Yurika Angata1, Tomomi Suzuki1, Mihi Yamashita1, Narumi Sato1, Saki
Nakamura1, Dai Yuki1, Yuki Sugiura1, Nobuhiro Zaima1, Naoko Goto-Inoue1, Takahiro Hayasaka1,
Yutaka Shimomura2, Mitsutoshi Setou1*
1Department of Cell Biology and Anatomy, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka, Japan, 2Laboratory of Genetic Skin Diseases, Niigata
University Graduate School of Medical and Dental Sciences, Niigata, Niigata, Japan
Abstract
Background: Human hair is one of the essential components that define appearance and is a useful source of samples for
non-invasive biomonitoring. We describe a novel application of imaging mass spectrometry (IMS) of hair biomolecules for
advanced molecular characterization and a better understanding of hair aging. As a cosmetic and biomedical application,
molecules whose levels in hair altered with aging were comprehensively investigated.
Methods: Human hair was collected from 15 young (2065 years old) and 15 older (5065 years old) volunteers. Matrix-free
laser desorption/ionization IMS was used to visualize molecular distribution in the hair sections. Hair-specific ions displaying
a significant difference in the intensities between the 2 age groups were extracted as candidate markers for aging. Tissue
localization of the molecules and alterations in their levels in the cortex and medulla in the young and old groups were
determined.
Results: Among the 31 molecules detected specifically in hair sections, 2—one at m/z 153.00, tentatively assigned to be
dihydrouracil, and the other at m/z 207.04, identified to be 3,4-dihydroxymandelic acid (DHMA)—exhibited a higher signal
intensity in the young group than in the old, and 1 molecule at m/z 164.00, presumed to be O-phosphoethanolamine,
displayed a higher intensity in the old group. Among the 3, putative O-phosphoethanolamine showed a cortex-specific
distribution. The 3 molecules in cortex presented the same pattern of alteration in signal intensity with aging, whereas
those in medulla did not exhibit significant alteration.
Conclusion: Three molecules whose levels in hair altered with age were extracted. While they are all possible markers for
aging, putative dihydrouracil and DHMA, are also suspected to play a role in maintaining hair properties and could be
targets for cosmetic supplementation. Mapping of ion localization in hair by IMS is a powerful method to extract
biomolecules in specified regions and determine their tissue distribution.
Citation: Waki ML, Onoue K, Takahashi T, Goto K, Saito Y, et al. (2011) Investigation by Imaging Mass Spectrometry of Biomarker Candidates for Aging in the Hair
Cortex. PLoS ONE 6(10): e26721. doi:10.1371/journal.pone.0026721
Editor: Johanna M. Brandner, University Hospital Hamburg-Eppendorf, Germany
Received May 6, 2011; Accepted October 2, 2011; Published October 24, 2011
Copyright: ? 2011 Waki et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by grants-in-aid for the scientific research project ‘‘Machinery of bioactive lipids in homeostatis and diseases’’ from the
Ministry of Education, Culture, Sports, Science and Technology of Japan, and for the project ‘‘Development of Systems and Technology for Advanced
Measurement and Analysis’’ by Japan Science and Technology Agency. The funders had no role in study design, data collection and analysis, decision to publish,
or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: setou@hama-med.ac.jp
Introduction
Roles of human hair
Hair significantly influences the appearance and is one of the
components of the human body that determine how individuals
look for their age [1]. Hair changes chemically and physically as a
result of various environmental assaults and undergoes intrinsic
degeneration with aging, resulting in an alteration of its
appearance, e.g., color and shine; feel, e.g., wettability and softness;
and structure, e.g., formation of split ends and frizz [2]. Therefore,
hair care is a huge industry, which supplies products such as
shampoos and conditioners to clean, protect, and provide a
desirable look and feel to hair [3].
At the same time, hair is used as an index of body properties.
The advantages of hair over other commonly used samples such as
blood or urine as an indicator include ease and painlessness of
sampling, ease of storage, and the possibility of monitoring past
exposure [4,5,6]. Forensically, hair has been utilized as trace
evidence for the investigation and successful prosecution of
individuals suspected of being involved in crimes [7].
Molecules constructing human hairs
Hair keratin proteins and hair keratin-associated proteins
(KAPs), composed of large gene families, are the predominant
structure proteins in the hair [8,9]. In the human hair cortex,
keratin intermediate filaments (KIFs) are produced from hair
PLoS ONE | www.plosone.org1October 2011 | Volume 6 | Issue 10 | e26721

Page 2

keratins, cross-linked with KAPs through extensive disulfide
bonds, and rigid hair shafts are produced.
Methods for molecular research of hair
Nearly all methods reported in the literature identify analytes in
the hair by headspace solid phase microextraction-gas chroma-
tography-mass spectrometry (MS), or more recently, by gas
chromatography-tandem MS [10] and liquid chromatography-
MS [11]. However, as samples are generally prepared by the
elution of molecules with an organic solvent, these methods
provide neither visual information nor analyte localization within
the hair strand. In this regard, several techniques involving
secondary ion MS and multi-isotope imaging MS (IMS) are used
for the determination of elemental composition of a cross-section
of hair; however, detectable targets are limited to elements
[12,13,14].
Matrix-assisted laser desorption/ionization (LDI) or LDI-based
IMS enables the analysis of much larger biomolecules because of
the soft ionization principle used and is a powerful tool for
investigating biomolecules comprehensively without the use of
time-consuming extraction, purification, or separation procedures
for biological tissue sections [15,16,17]. Although IMS is used
for detecting drugs in the hair for forensic purpose [18],
comprehensive analysis of molecules in hair by using IMS is
yet to be done. Earlier, we developed an imaging mass
spectrometer with a higher spatial resolution than the original
ones [19]. This was utilized for IMS analysis of hair sections in
the present study.
Molecular markers for aging in hair
The characterization of hair aging with regard to the alteration
of molecular mass and distribution of molecules within the hair
structure is essential for the development of better cosmetic
products. It is predicted that in the future, improvements in hair-
care product development will target specific molecules [20], and
thus, the supplementation of molecules impaired during aging or
addition of their functional analogues to hair care products would
be beneficial to this section of the market.
The identification of molecules in hair that could define an
individual’s age has important forensic applications. This is
because the age, which is one of most vital pieces of information
in an investigation, cannot be determined definitively by the use of
microscopically measured indices such as hair diameter or
ellipticity [7,21].
Molecules related to aging in hair, however, have not been
comprehensively investigated, with the exception of some trace
minerals analyzed by atomic absorption spectrometry [22,23].
Therefore, as the first target for utilizing our method, we chose
investigation of biomarkers that could distinguish human age.
The objective of this article is to describe an initial assessment of
the application of IMS for comprehensive detection of aging-
related molecules in cross-sectioned hair.
Results
Visualization of molecular distribution in cross-sectioned
hair
Figure 1A provides images of hair sections obtained from
subjects aged 2065 years (hereafter termed 20 YO group) and
5065 years (50 YO group). As a typical result of IMS analysis, the
ion distribution at a mass-to-charge ratio (m/z) of 125.99, detected
with the highest intensity among hair-specific ions, is presented in
Figure 1B.
Human scalp hair is composed of a core structure of a centrally
located medulla, cortex consisting of different cell types, and
surrounding layer of cuticle cells (Figure 1C)[24]. Figure 1D
depicts the definition of regions of interest (ROI) corresponding to
cortex and medulla. As shown in Figure 1A, S1, and S2, the shape
and size of hair sections were different among samples. Thus, the
areas that were considered to correspond to the medulla were
observed only in 38 hair sections out of 90 in the light microscopic
images of the sections. Mean signal intensity of the ion at m/z
125.99 in the cortex did not show statistically significant difference
from that in the medulla (p=0.15, paired t-test; Figure 1E).
In order to confirm that cross-sectional area does not show
linear alteration with aging and thus cannot be used as an
indicator of aging, the area composed of both the cortex and
medulla was counted and compared between the 2 age groups.
There was no significant difference in cross-sectional area between
the 20 YO and 50 YO group (p=0.082, unpaired t-test; Figure 1F),
while large sectional area over 11,000 mm2was only seen in the
20-YO group (Figure 1G).
Selection of hair-specific molecules
As a preliminary step in the detection of markers for aging,
principal hair-specific molecules were extracted. The top 50
signals in the ROI in the cortex and medulla were selected from
each of signal intensity values obtained from the analysis of
subjects No. 1, 2, 3, 16, 17, and 18 (Figure S3A), and 56 ions
chosen from more than 1 of these 6 subjects were listed (Figure
S3B). Among these, 31 ions with a mean hair-specific intensity at
least 3-fold and statistically significantly (p,0.05, paired t-test)
higher than the mean background intensity (black area in
Figure 1D) were selected as hair-specific molecules (red letters in
Figure S3B).
Extraction of putative aging markers
In order to find putative aging markers, those molecules that
exhibited a significant difference in the signal intensity between the
2 age groups from among the 31 hair-specific molecules were
determined. The mean signal intensities at m/z 153.00 and 207.04
were higher in the 20 YO group than in the 50 YO group
(p=0.0064 and 0.013, respectively; unpaired t-test; Figures 2A and
C). In contrast, the ion at m/z 164.00 was observed to have a
significantly higher intensity in the 50 YO group than in the 20
YO group (p=0.0067; Figure 2B). Even when the sections whose
areas were over 11,000 mm2were excluded from the analysis to
reduce possible diameter effects, the significance of the differences
was not altered (p=0.0052, 0.0064, and 0.043 for m/z 153.00,
164.00, and 207.04, respectively; unpaired t-test; Figure 2D, E,
and F).
As already evaluated by selection of hair-specific ions, these 3
molecules were specific to hair sections (Figure 2G, H, and I).
Among the 3 putative aging markers, only the one at m/z 164.00
presented a cortex-specific distribution (p=0.70, 0.0044, and 0.97
for m/z 153.00, 164.00, and 207.04, respectively; paired t-test;
Figure 2J, K, and L). Hair-specific distribution of the molecules
and cortex-specific distribution of the molecule at m/z 164.00 in
each subject’s sample was confirmed by mapping of ion
distribution (Figure 2M, N, and O).
To confirm that the extracted molecule originated from a single
molecule observed as a peak in the histogram, ROI-specific mass
spectra were structured using values from subject No. 1 as a
typical example. As illustrated in Figures 3B to F, signal intensity at
m/z 125.99, 153.00, 164.00, and 207.04 in hair-specific ROI (red
color) were observed as individual peaks.
Aging Marker Investigation by IMS in Hair Cortex
PLoS ONE | www.plosone.org2 October 2011 | Volume 6 | Issue 10 | e26721

Page 3

Comparison of intensity of the putative markers in the
cortex and medulla between 2 age groups
To determine whether the intensity of the molecule in the
cortex and medulla differed between 20 YO and 50 YO, mean
intensity in each ROI was compared between the 2 age
groups. Figures 4A and C illustrate that the intensities at m/z
153.00 and 207.04 in the cortex were significantly higher in the
20 YO than in the 50 YO group (p=0.0081 and 0.010,
respectively; unpaired t-test). As shown in Figure 4B, the ion at
m/z 164.00 in the cortex displayed a higher intensity in the 50
YO than in the 20 YO (p=0.0097). On the other hand, no
significant difference between the 2 groups was observed in the
mean medullary intensity at any m/z value (p=0.56, 0.95, and
0.38 for m/z 153.00, 164.00, and 207.04, respectively; Figure 4D,
E, and F).
Assignment of putative aging markers
Table 1 shows the assignment of the putative aging markers
extracted above, listing the molecules which exhibited highest
probability in MS/MS analysis and the molecules assumed in
single MS analysis.
From the precursor ion at m/z 207.04, fragment ions at m/z
75.49, 133.07, 138.89, 141.84, 149.53, 151.51, 188.32, and
198.92, fragment ions at m/z 140.88 and 160.26, fragment ions at
m/z 61.06, 84.82, 136.18, 143.88, 148.54, and 191.53 were
detected at the energy level of 10, 30, and 50, respectively (Figure
S4A to E). By searching the fragments obtained at energy level of
10, 30, and 50, the precursor ion was assigned to 3,4-
dihydroxymandelic acid (DHMA), which presented the highest
relative fitness values among all precursor candidates: 81%, 98%,
and 96%, respectively.
Figure 1. Visualization of molecular distribution in hair cross-sections. (A) Images of cross-sectioned hairs are shown, numbered as per the
subject No. Each of the 3 photographs with the same number is from an independent scalp hair. Scale bar: 100 mm. (B) Ion distribution at m/z 125.99.
Scale bar: 100 mm. (C) The principal hair structures are shown. M: hair medulla. (D) ROI corresponding to the hair structures are depicted. White area:
cortex. Red area: medulla. Black area: background. Scale bar: 100 mm. (E) The signal intensity at m/z 125.99 in hair cortex and medulla is shown *:
p,0.05. (F) Cross-sectional area of hair is shown. (G) A histogram of the cross-sectional area is depicted. Black bar: 20-YO group. White bar: 50-YO
group. All values are presented as mean6standard deviation.
doi:10.1371/journal.pone.0026721.g001
Aging Marker Investigation by IMS in Hair Cortex
PLoS ONE | www.plosone.org3 October 2011 | Volume 6 | Issue 10 | e26721

Page 4

M/z 153.00 and 164.00 were speculated to be dihydrouracil
and O-phosphoethanolamine, respectively, which showed a
difference of 0.01 Da or less between the calculated m/z and
query. The ion with the highest intensity among hair-specific ions,
m/z 125.99, was assigned to 2-aminoacrylic acid.
Discussion
In the present study, we applied IMS to cross-sectioned hair
to investigate multiple molecules showing aging-related alter-
ations. Images of hair sections were obtained at high resolution,
Figure 2. Intensity and distribution of putative aging markers. (A–C) Mean signal intensity in the hair section is displayed. (D–F) Mean signal
intensity in the hair section excluding those with sectional area of 11,000 mm2or larger is displayed. (G–I) Mean signal intensity in the hair section
and background area is displayed. (J–L) Mean signal intensity in the cortex and medulla is displayed. (M–O) Each panel shows ion distribution in cross-
sectioned hair. The number indicates the subject No. Each of the 3 pictures was obtained from an independent scalp hair of a single subject. Scale
bar: 100 mm. (A, D, G, J, M) m/z 153.00. (B, E, H, K, N) m/z 164.00. (C, F, I, L, O) m/z 207.04. All values are shown as mean6standard deviation. *: p,0.05.
**: p,0.01. ***: p,10212.
doi:10.1371/journal.pone.0026721.g002
Aging Marker Investigation by IMS in Hair Cortex
PLoS ONE | www.plosone.org4 October 2011 | Volume 6 | Issue 10 | e26721

Page 5

and the distribution of ions was characterized on a micrometer
scale.
Speculated relationship between aging and tissue
localization of the extracted molecules
Dihydrouracil, presumed to be a precursor of the ion at m/z
153.00, is an intermediate metabolite of uracil [25], and DHMA,
identified as the ion at m/z 207.04, is a major metabolite of the
catecholamines [26]. Both are released into the circulation after
conversion [25,26] and can thus be transferred to matrix cells
and/or melanocytes via blood vessels in the dermal papilla [27].
Increased apoptosis of follicular melanocytes is a phenomenon
associated with aging [28]. Eumelanin synthesized in bulbar
melanocytes is transferred to matrix cells which proliferate and
differentiate into the hair shaft cortex [29]. The robust binding of
eumelanin to basic molecules by ionic interaction [30] and
aromatic carbon by hydrophobic interaction [31] renders it a
drug-binding site within the hair structure, therefore, the
hydrophobic base dihydrouracil and DHMA with dihydroxyben-
zene ring are foreseeable interaction partners. This might explain
the relationship between aging and reduction of the signal
intensities, and further investigation would be done to validate
this theory.
Dihydrouracil and DHMA were detected in a punctuated,
heterogenous manner through the hair sections. This heterogene-
ity might be explained by the diversity of cellular components of
the hair structure, since a hair fiber is composed of different cell
types including the ortho-, the meso-, and the paracortical cells
[32]. Moreover, recent study has shown that KIFs in the fiber has
several patterns in their arrangement [24]. If such structural
molecules in the hair shaft function as an absorber for the two
molecules or their precursor molecules, the heterogeneity which
was observed here will result.
O-phosphoethanolamine, tentatively assigned to the ion at m/z
164.00, is a metabolite of sphingosine-1-phosphate (S1P) [33].
Aging is associated with increased platelet activation [34], which
leads to enhanced secretion of S1P into the circulation [35]. As
observed in other tissues, increased S1P might be processed to O-
phosphoethanolamine and 2-hexadecanal by S1P-lyase following
internalization by the cells forming the hair structure [36].
The mean signal intensity at m/z 125.99, tentatively assigned to
2-aminoacrylic acid, did not significantly differ between that in the
hair cortex from in the medulla (Figure 1E). Both cortical and
medullar cells are composed mostly of bundles of KIFs and consist
of keratin proteins [24,37], while the hair medulla contains heavily
vacuolated cells. Since 2-aminoacrylic acid, also called dehydroa-
lanine, is a product of post-translational modifications observed in
keratin proteins, its localization both in the cortex and the medulla
is reasonable [38].
Cosmetic perspective
The hair shaft surface is covered with integral lipids, the only
continuous structure that plays a role in maintaining moisture,
luster, mechanical integrity, and stiffness of hair [2,39]. Oil-
containing hair cosmetics have been experimentally proven to
complement or raise the efficacy of endogenous lipids: coconut oil
Figure 3. Hair-specific mass spectra of putative aging markers.
ROI-specific mass spectra in subject No. 1 are presented. Red peaks and
blue peaks are derived from the hair section and background area,
respectively. (A) ROI selection is illustrated: ROI A as hair section and ROI
B as background area. (B) m/z 50 to 300. (C) m/z 125.99. (D) m/z 153.00.
(E) m/z 164.00. (F) m/z 207.04.
doi:10.1371/journal.pone.0026721.g003
Aging Marker Investigation by IMS in Hair Cortex
PLoS ONE | www.plosone.org5 October 2011 | Volume 6 | Issue 10 | e26721