Accepted Manuscript

Title: Assessment of Brain Derived Neurotrophic Factor in

hair to study stress responses: A pilot investigation

Authors: H. Harb, M. Gonzalez-de-la-Vara, L. Thalheimer, U.

Klein, H. Renz, M. Rose, J. Kruse, D.P. Potaczek, E.M.J.
Peters

PII: S0306-4530(17)30191-9
DOI: http://dx.doi.org/10.1016/j.psyneuen.2017.09.007
Reference: PNEC 3718

To appear in:

Received date: 24-2-2017

Revised date: 1-8-2017
Accepted date: 6-9-2017

Please cite this article as: {http://dx.doi.org/

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Assessment of Brain Derived Neurotrophic Factor in hair to study stress responses: a

pilot investigation

Runnig Title: BDNF detection in hair for the study of distress

Harb Ha,b* PhD; Gonzlez-de-la-Vara Mc,d* MD, PhD; Thalheimer Ld; Klein Ud; Renz Ha,b PhD;

Rose Me MD; Kruse Jc MD; Potaczek DPa,b MD, PhD; Peters EMJd,e MD

a
Institute of Laboratory Medicine and Pathobiochemistry, Molecular Diagnostics, Philipps
University Marburg, Research Facility at ZTI, Hans-Meerwein-Strasse 3, 35043 Marburg,
Germany, harbh@staff.uni-marburg.de, harald.renz@uk-gm.de, potaczek@staff.uni-
marburg.de
b
International Inflammation (in-FLAME) Network, Worldwide Universities Network (WUN)
c
Department of Ethology, Wild Life and Laboratory Animals, Faculty of Veterinary Medicine,
Universidad Nacional Autnoma de Mxico, Avenida Universidad 3000, 04510 Mexico City,
Mexico, marcela.gonzlez@yahoo.com.mx
d
Psychoneuroimmunology Laboratory, Department of Psychosomatics and Psychotherapy,
University of Gieen, Aulweg 123, D-35385 Gieen, Germany, Lara.Thalheimer@web.de,
undine-klein@web.de, eva.peters@eva-peters.com, johannes.kruse@psycho.med.uni-
giessen.de
e
Charit Center 12 Internal Medicine and Dermatology, Division for General Internal Medicine,
Psychosomatics and Psychotherapy: Psycho-Neuro-Immunology Skin Research Group,
Universittsmedizin Berlin, Charit Platz 1, D-19115 Berlin, Germany, eva.peters@charit.de,
matthias.rose@charite.de
Harb et al.

*equally contributing

Corresponding author: PD Dr. med. Eva Peters

Psychoneuroimmunology Laboratory, Aulweg 123; D-35385 Gieen, Germany

Phone: +49 641 99 45642 or +49 179 6968465; Fax: +49 30 80499632
Email: eva.peters@eva-peters.com

2
Harb et al.

Highlights

BDNF in hair negatively correlates with hair-cortisol and somatic complaints (SC)

H4 histone acetylation at the BDNF gene P4-promoter negatively correlates with SC

Students with high SC during the end of term exam have lower hair-BDNF

Therefore, BDNF protein levels can be determined in clipped hair

Also, cells from plucked hair can be used for epigenetic analysis of stress-effects

Abstract

To study pathogenic stress-effects in health and disease, it is paramount to define easy access

parameters for non-invasive analysis of biological change in response to stress. Hair samples

successfully provide this access for the study of hypothalamus-pituitary-adrenal axis (HPA)

changes. In this study, we assess the hair expression and corresponding epigenetic changes of

a neurotrophin essential for autonomic nervous system function and mental health: brain

derived neurotrophic factor (BDNF). In three independent studies in healthy academic

volunteers (study I: German students, N=36; study II, German academic population sample,

N=28; study III: Mexican students, N=115), BDNF protein expression or BDNF gene (BDNF)

histone acetylation was determined. Simultaneously, mental distress and distress-associated

somatic complaints were assessed by self-report. In study I, we found a negative correlation

between hair-BDNF protein level and hair-cortisol as well as between hair-BDNF and somatic

complaints, while hair-cortisol correlated positively with mental distress. In study II, we found a

negative correlation between H4 histone acetylation at the BDNF gene P4-promoter and

somatic complaints. Regression analysis confirmed confounder stability of associations in both

studies. In study III, we confirmed study I and found lower hair-BDNF protein level in volunteers

with high somatic complaints, who also reported higher mental distress during the end of term

3
Harb et al.

exams. The results indicate that BDNF protein levels can be detected in clipped hair and are

associated with somatic complaints and stress in life. In addition, we concluded that plucked

hair can provide material for the study of epigenetic changes in stress-affected tissues. These

tools can prove valuable for future studies on distress, both under experimental and field

conditions.

Keywords
Stress, hair, brain derived neurotrophic factor (BDNF), histone acetylation, epigenetic,

psychoneuroimmunology.

4
Harb et al.

1 Introduction

Sampling cortisol from clipped hair is a non-invasive, easy to perform method, which reliably

mirrors cumulative hypothalamus pituitary adrenal axis (HPA) responses to stress-exposure.

Since the first descriptions of the method it has gone through numerous validations and routine

incorporation of this biomarker into clinical settings can be expected in the future, as it has been

successfully implemented into a wide variety of laboratory and clinical studies (Kirschbaum et

al., 2009; Stalder et al., 2017; Wester and van Rossum, 2015). Cortisol levels extracted from the

most proximal cm of hair, clipped at the level of the scalp, give the most consistent results and

relate well to area under the curve changes in salivary cortisol secretion over the preceding 4

weeks, and prolonged stress exposure (Short et al., 2016). They also relate well to various

states of chronic mental distress and associated disorders (Staufenbiel et al., 2013; Wester and

van Rossum, 2015). Thereby, changes in hair-cortisol indicate clinically relevant lasting HPA

modulation, which increases with higher intensity and a longer duration of stress.

However, measurements of hair-cortisol do not always relate to ephemeral measurements of

HPA functioning, such as serum cortisol or diurnal salivary cortisol profiles, nor do they always

associate with the employed self-report assessments of mental distress (Gidlow et al., 2016).

This is a result of several factors: the transient expression of cortisol in blood and saliva as

opposed to its accumulation in targeted tissues; the often one-dimensional selection of mental

distress-measurements assessed together with the hair-cortisol, which may not include the

corresponding distress dimension; or the lack of simultaneous assessment of somatic

conditions associated with distress. To complement the outcome portfolio for stress research,

establishment of additional easy-access markers and their relationship with a spectrum of

distress measures are a worthwhile strategy. Establishment of such a marker would require

analysis of its level in the context of a variety of mental distress states, ranging from dimensions

such as present mood via prolonged perception of distress or anxiety to severe mental distress

5
Harb et al.

states such as depression. In addition, it would be instructive to know the relationship with

associated somatic complaints and standard biological outcomes indicative of distress, such as

stress mediators in serum, tissue health or immune outcomes like C-reactive protein (CRP)

serum levels or cytokine-production by peripheral blood mononucleocytes (PBMC).

It is with this in mind that we suggest that hair can provide surrogate material for the detection of

mediators representing additional stress response systems co-activated with the HPA. One

such mediator worth investigating is the neurotrophin brain derived neurotrophic factor (BDNF).

BDNF levels reportedly change in the plasma or serum, when altered HPA function in response

to continuous distress is observed (Naveen et al., 2016). It is responsible for sympathetic stress

axis function (Kasemeier-Kulesa et al., 2015) and is also hotly debated as a key player in

mental distress disorders (McEwen et al., 2015). BDNF therefore serves as a neuroendocrine

growth factor and an immune response regulator (Peters et al., 2005; Rutlin et al., 2014). The

association of hair-BDNF with certain dimensions of self-reported distress, indicators of

neuroendocrine tissue regeneration and immune response markers can therefore be

hypothesized.

Second, the chronic deregulation of HPA function may be linked to altered expression of BDNF

through epigenetic regulation of the BDNF gene (BDNF), as altered hair-cortisol measurements

link with stress-associated epigenetic changes in target tissues, and epigenetic regulation of

BDNF was shown to occur in reponse to stress (Ambrus et al., 2016; Makhathini et al., 2017;

McEwen et al., 2015; Natt et al., 2015; Simsek et al., 2015). Brain structures, such as the

hippocampus, show stress-sensitive morphological changes associate with regulation of BDNF-

expression. Corresponding stress-sensitive peripheral BDNF-expression was found in skin and

hair follicles (Panchaprateep et al., 2011; Peters et al., 2006; Peters et al., 2005). It is therefore

suggested that the living cells of the hair follicle are not only a target for BDNF but also a stress-

sensitive source and that hair can provide living cells for molecular analysis if it plucked, as

6
Harb et al.

opposed to clipped (Raab et al., 2014). In this respect, the non-invasive sampling of hair derived

cells may hold great potential as a source of surrogate material for analysis of epigenetic

response to stress. Commonly, PBMC are used to study epigenetic changes in response to

stress (Harb et al., 2016; Harb et al., 2015a; Harb et al., 2015b; Tost, 2016), as surrogate

material for analysis of histone modifications needs to contain living cells and these are easy to

obtain. However, PBMC are subjected to high turnover and variation in subtype distribution in

blood (Adalsteinsson et al., 2012), as they reflect only the time of blood collection. Their travel

through the bloodstream is often only en route to their functional destination, where their

biological characteristics most often change. As a consequence, PBMC analysis does not

always correlate with target tissue analysis in epigenetic studies (Rask-Andersen et al., 2016).

Thus, more stable and disease-relevant epigenetic changes in response to stress may be

detectable in stress-targeted tissues such as the hair follicle.

Third, if BDNF can be detected in hair, in the context of stress in a meaningful manner, it would

be instructive to test the stability of its assessment in field studies. In such studies, sample

harvesting conditions are usually less optimal than under controlled laboratory conditions and

available material for data acquisition is limited especially with respect to biological samples that

can be acquired. Thus assessment of the suitability of such a marker for field studies would be

instructive.

In summary, we hypothesized that BDNF protein can be measured in clipped hair to monitor

mental distress and associated somatic complaints and biological change. Moreover, we

suggest that by plucking hair, living cells can be obtained for corresponding epigenetic analysis.

Finally, we hypothesized that BDNF-levels in hair are consistently measurable under field study

conditions. These hypotheses were tested in three independent studies with healthy academic

volunteers. Our pilot results suggest that BDNF hair analysis provides instructive information

relating to the mental and biological consequences of stress.

7
Harb et al.

2 Materials and Methods

2.1. Ethical considerations and recruitment of participants

Three independent studies were conducted in accordance with the Code of Ethics of the World

Medical Association (Declaration of Helsinki). Studies I and II were approved by the ethics

committees of the Charit-Universittsmedizin, Berlin and the Justus-Liebig University, Gieen,

Germany; Study III by the ethics committee of Universidad Autnoma de Mxico. Mexico City,

Mexico. Recruitment of healthy volunteers followed standardized procedures as detailed in

Table 1: A. Briefly, participants were asked for permission to take one hair sample, and to self-

report distress. In addition, in study I they were asked to provide a blood sample. They were

asked not to wash or dye their hair before sampling, as this was reported to affect cortisol

measurements and potentially cell viability of hair follicle constituting cells (Hoffman et al., 2014;

Li et al., 2012). Only participants who provided informed consent were enrolled. Briefly, 36

human medicine students were included in study I, 28 academics in study II, and 115 veterinary

medicine students in study III. Basic socio-biologic data are provided in Table 1: B.

2.2. Hair sampling

Approximately 100 hairs were plucked (studies I and II) or clipped (study III) from the posterior

temporal scalp, which is directly adjacent to the posterior vertex region. This region was chosen

to enable multiple sampling, as it would allow for the collection of alternating samples from either

the left or the right side of the scalp in future studies. As the posterior vertex region is commonly

sampled in hair cortisol studies, we note the following considerations: biological features differ

little between the posterior temporal scalp and the posterior vertex: in both regions, hair growth is

hormone-independent; vasculature derives from the occipital artery and innervation derives from

the occipital nerve. Moreover, no significant differences between cortisol measures obtained in

8
Harb et al.

the temporal versus the posterior vertex region were reported in previous studies (Sauve et al.,

2007).

For studies I and II, hair was plucked, to extract the approximately 3 mm long part of the hair

shaft (tip), which is normally hidden in the scalp. Plucking hair for trichogram analysis is a well-

established routine method in dermatology. This procedure leaves no visible traces and only

induces a brief pain, which subsides within seconds after plucking, comparable to the brief pain

induced by drawing venous blood or by cosmetic epilation. When it is performed by trained

medical personal, informed subjects usually tolerate it well. In study I, the exposed tip was placed

on a microscope slide to be processed for hair morphology analysis (section 1.3) and the hair

shaft above the tip was clipped and processed accordingly (section 1.4). In study II, the living

epithelial cells, mainly of the inner root sheath and the lateral base of the hair shaft, which

commonly remain attached to the tip, were processed for epigenetic analysis, while the

remaining hair was discarded. For this, the tips with adjacent hair were immediately placed in a

drop of sterile phosphate buffered saline (PBS) and rapidly processed to obtain the living cells

(section 1.5). In study III, the most proximal cm of hair protruding from the scalp was clipped for

analysis.

2.3. Hair morphology analysis

In study I, a trichogram (Dhurat and Saraogi, 2009) was performed by two independent

investigators on digital photographs taken with a MOTIC BA 400 (Motic, Wetzlar, Germany) at

200 fold magnification. Briefly, hair tips were classified as anagen (growth phase) if a vigorous

thick bulb was attached to the plucked hair; as telogen (resting phase), if the classical club hair

appearance was present; and as catagen (transition between anagen and telogen), if they

showed neither morphology. Based on the total number of hair follicles present in the individual

sample, the percentages of anagen, telogen and catagen were calculated as described

previously (Peters et al., 2017) (Table 2).

9
Harb et al.

2.4. Extraction of cortisol and BDNF

For the extraction of cortisol and BDNF-protein, the most proximal centimeter of the hair shaft

was used. As hair in anagen gains approximately 1 cm per month, this cm was considered to

represent the past four weeks prior to sampling. This four-week period was used to measure

prolonged stress exposure. Hair was pulverized at 25 Hz for 2-3 min using a ball mill (Retsh

Tesberich M 400 Nr: 14277, Haan-Griten, Germany) and 7 mm diameter metal balls followed

by extraction procedures as detailed in Table 1: A. For extraction of each sample, 10 mg of

pulverized hair were added to a sterile eppendorf tube with high accuracy, using a precision

balance. To extract cortisol, 1 ml of methanol was added for 18.5 hrs, followed by centrifugation

and sample drying as detailed in Table 1: A. To extract BDNF, 220 l citric acid was added for

30 min, followed by centrifugation and lyophilisation as detailed in Table 1: A.

2.5. Analysis of histone acetylation

Hair tips were transferred to a precooled 15 ml tube containing PBS and centrifuged at 1000 rpm.

The supernatant was discarded and 1 ml of Trypsin/EDTA (Sigma-Aldrich, St. Louis, MO, USA)

added for 5 min at 37C and 5% CO2. The mixture was vortexed and the liquid phase transferred

into RPMI containing 10% FCS (Thermo-Fischer, Waltham, MA USA) to be centrifuged at 1000

rpm followed by re-suspending in PBS and transfer to a DNase/RNase-free tube (Starlab,

Hamburg, Germany) and another centrifugation. The cell pellet was re-suspended in 1% PFA for

8 min after which the mixture was centrifuged at 8000 rpm. The chromatin found in the pellet was

suspended in Lysis Buffer-I (For 50 ml: 0.5 ml 0.5M PIPES pH 8, 1.4 ml 3 M KCl, 0.25 ml 100%

NP40, Protease Inhibitor, 1 tablet) and stored at -20C. For chromatin immunoprecipitation

(ChIP), the chromatin was thawed and centrifuged for 5 min at 8000 rpm at RT. The chromatin

pellet was incubated for 5 min in Lysis Buffer-II (For 50 ml: 0.5 ml 1 M Tris-HCl pH 7.5, 1.5 ml 5

M NaCl, 0.5 ml 100% NP40, 0.5 g DOC, 0.25 ml 20% SDS, 0.1 ml 0.5 M EDTA pH 8, Protease

Inhibitor, 1 tablet) and placed for 3 minutes in ice, followed by sonication at below 8C for 35

10
Harb et al.

cycles of 30 seconds of sonication alternating with 30 seconds of rest. ChIP for H3 and H4

histone acetylation was performed as described previously (Harb et al., 2015a). The

immunoprecipitated DNA was analyzed by quantitative PCR with BDNF P4-promoter-specific

primers 5-GGAAAGTTGTTGGGCTGGTT-3 and 5- TGACCGACTCACTGTGACTC-3

(Metabion, Planegg, Germany). Percent enrichment in H3 or H4 histone acetylation was used for

subsequent statistics (Harb et al., 2015a).

2.6. Blood sample acquisition and processing for further analysis

Blood samples were obtained simultaneously with hair samples in study I, and immediately

processed for further analyses as detailed in Table 1: A. This was done a) to compare BDNF

and cortisol levels detected in hair with frequently performed BDNF and cortisol detection

methods b) to assess if hair-BDNF levels correlates with CRP and inflammatory responsiveness

of peripheral blood mononuclear cells (PBMC), important biologic targets of altered BDNF

expression. For serum-ELISA detection of BDNF, cortisol and CRP, serum was collected after

centrifugation in aliquots of 500 l and stored at -80C. For Cytometric Bead Array analysis

(CBA; Bender MedSystems, eBioscience, Frankfurt, Germany) PBMC were harvested by Ficoll

separation and stimulated ex vivo with an immune challenge. As a challenge,

phytohaemagglutinin (PHA, a lectin from Phaseolus vulgaris; Sigma, Sigma-Aldrich Corporation,

St. Louis, MO, USA) was chosen and added in AIM V to the PBMCs at a density of 1.25*106 cells

for 24 hrs in a 37C/5% CO2 incubator. PHA stimulation is known to activate cellular immune

responses similar to those observed under prolonged stress exposure such as activation and

cytokine shift of T cells (Cuesta et al., 2016; Newton et al., 2011). The supernatants were

collected and stored at -80C until analysis of cytokine production (section 2.8).

2.7. ELISA determination of cortisol, BDNF and CRP

For cortisol detection in hair extracts, dry hair extracts were suspended in Dulbeccos PBS

(110l/extract of 10mg pulverized hair) as recommended by the manufacturer. Measurements

11
Harb et al.

were done using the human cortisol ELISA kit for saliva (IBL International, Hamburg, Germany;

minimum detection level 0.003 g/dl; results are given in g/dl suspended hair extract which

corresponds to g/9.091 gr pulverized hair used for extraction) following the manufacturers

instructions and previous publications (Gonzalez-de-la-Vara et al., 2011; Karlen et al., 2011;

Meyer et al., 2014). For serum-cortisol, a human serum kit was used (IBL International;

minimum detection level 2.46 ng/ml; results given in ng/ml serum). For total free BDNF detection

in hair extracts, lyophilized samples were suspended in Dulbeccos PBS (Parsley, UK, ph 7.5,

110 l/extract of 10 mg pulverized hair) as recommended by the manufacturer. Measurements

were done at a dilution of 1:10 (RayBiotech, Norcross, GA, USA; minimum detection level 80

pg/ml; results are reported in ng/ml suspended hair extract which corresponds to ng/0.091 gr

pulverized hair used for extraction, study I) or 1:1000 (Emax ImmunoAssay System, Promega,

Madison, WI, USA; minimum detection level 15.6 pg/ml; results are reported in ng/ml suspended

hair extract which corresponds to ng/0.091 gr pulverized hair used for extraction, study III). To

validate accuracy of the ELISA measurements for BDNF detection in hair extracts, spiking

recovery and linearity were calculated from 9 samples. Samples were spiked with 15.6, 31.3 or

62.5 pg/ml BDNF. Recoveries for spiked test samples were calculated by comparison to the

measured recovery of spiked diluent control and yielded recovery rates of 81.44 +/- 11.70, 84.16

+/- 12.42 and 95.62 +/- 15.24 respectively. Percentage linearity was calculated by dividing the

observed concentration by the previous observed value in the dilution series divided by 2,

followed by multiplication times 100. The mean percent linearity was 80.76 +/- 12.76; 86.95 +/-

14.72 and 90,03 +/- 17.08 respectively. The parallelism test showed similar slopes for all 9

samples. For CRP determination (study I) an ELISA kit for serum was used (RayBiotech,

Norcross, GA, USA; minimum detection level 2 ng/ml; results are reported in ng/ml serum). All

ELISA measurements were done in duplicate with a Tristar LB 941 reader (Berthold

Technologies, Germany). The intra-assay coefficients of variation for all reported ELISAs were

less than 3% and all results measured were within the detection range.

12
Harb et al.

2.8. Cytometrix Bead Array determination of cytokines

After collection of supernatants from stimulated PBMCs as specified in paragraph 2.6., the

cytokine levels in study I were determined using a CBA Kit (Bender MedSystems) for

simultaneous concentrations measurements of various cytokines, representing different aspects

of stress-sensitive immune responses potentially modulated by prolonged stress exposure

through BDNF. Interleukin 1 alpha (IL1) was chosen to represent immediate innate immune

responses associated with skin barrier function and stress activation of peptidergic innervation

(Buske-Kirschbaum et al., 2007; Di Paolo and Shayakhmetov, 2016; Scheich et al., 2017 ), IL6 to

represent immune responses dominated by granulocyte activation (Ciliberti et al., 2017; Gola et

al., 2013), tumor necrosis factor alpha (TNF) to indicate activation of macrophage- and mast

cell-dominated immune responses (Bucker et al., 2015; Gola et al., 2013), interferon gamma

(IFN), IL4 and IL17 as marker cytokines for adaptive T helper cell Type 1 (TH1), 2 (TH2) and 17

(TH17) immune responses (Hansel et al., 2010; Hong et al., 2013; Kamezaki et al., 2012;

Lenarczyk et al., 2000; Oh and Ghosh, 2013; Xiang et al., 2012). A fluorescent activated cell

sorter (FACS, Becton Dickinson FACSCaliburTM, BD Biosciences, San Jose, CA, USA) (Table 2)

was used as recommended by the manufacturer (Bender MedSystems) and concentrations were

calculated using FlowCytomixPro 2.4 software (Bender MedSystems).

2.9. Self-report assessment

Test quality criteria of the questionnaires used in the studies have been described, as cited

below. Questionnaires employed in studies I and II were delivered in German. In study III,

Spanish versions of the questionnaires were employed. In studies I and II, momentary mental

well-being was assessed with the `Multidimensional Mood State questionnaire (MDMQ)

subscale `calm-nervous (Hofer et al., 2016) as an immediate indicator of the level of distressed

mood. Respondents rate statements, such as the following: right now I feel good or right now I

feel unhappy. Available rating categories range from definitely not to extremely. The

13
Harb et al.

`Screening for Somatoform Disorders (SOMS) questionnaire was used to record the presence or

absence of 53 physical complaints frequently associated with distress (e.g. nausea, diarrhea,

dizziness) during the past 7 days (Rief et al., 1998). The `Perceived Stress Questionnaire (PSQ)

was used to address prolonged stress, perceived over the past four weeks, such as tension or

worries (Fliege et al., 2005) and the subscale `state of the `State and Trait Anxiety Index (STAI)

was used to assess anxiety present over the past four weeks (Spielberger et al., 1970). In

addition, in studies I and II the subscale `depression of the Hospital Anxiety and Depression

Scale (HADS) was used to rule out clinically manifest depression (Herrmann, 1997). In study III

we used the Profile of Mood States (POMS) subscale `tension-anxiety (Andrade et al., 2010),

the Patient Health Questionaire-15 (PHQ-15) (Ros Montalban et al., 2010), the Perceived Stress

Scale (PSS) (Gonzalez-Ramirez et al., 2013), the subscale `state of the STAI short version

validated in Spanish (Perpina-Galvan et al., 2013) and the Beck Depression Inventory (BDI)

(Diez-Quevedo et al., 2001). In addition, since we found an exclusive association between hair-

BDNF and the SOMS, the POMS subscale `fatigue-inertia was evaluated in study III, as the

POMS offers analysis of this dimension and it was reported in the context of altered BDNF-levels

and somatic complaints in previous studies (Minelli et al., 2011).

2.10. Statistical analysis

For psychometric assessments, ordinal items were converted to a scale ranging from 0 to 100

points where applicable. The category "does not apply" and item non-response were coded as

missing values and no imputation techniques were used. All values were calculated as mean and

standard deviation (SD) per group and point in time. SPSS Statistics for Windows software,

version 23 (Armonk, NY, USA) and GraphPad Prism for Windows, version 6.05 (GraphPad

Software, La Jolla, CA, USA) were used to test for outliers, linearity, normal distribution and

homoscedasticity as required for the following tests: In study I and II Spearmann correlations

between variables were calculated employing the rank correlation coefficient rho. Associations

14
Harb et al.

were considered as weak if -0.3>rho<0.3, medium if 3.0>rho<4.9 or -3.0<rho>-4.9 and strong if

rho>0.5 or <-0.5 (Cohen, 1988). After logarithmic transformations to achieve approximately

normal data distribution where applicable, associations were also analyzed by multiple linear

regression models to control for potential confounders (e.g. age and sex (Wosu et al., 2013)).

Multiple linear regression models confirmed correlations. In study III, the Mann-Whitney test was

used to compare groups. All p-values given were two-tailed and p-values of less than 0.05 were

considered as significant. In addition, p-values less than 0.1 are reported as tendencies towards

significance.

15
Harb et al.

3 Results

3.1. BDNF extracted from hair inversely correlates with cortisol, hair in transition
between growth and rest and somatic complaints

In study I, Spearman correlations calculated in 36 healthy female German medical students

revealed a significant negative correlation between hair-BDNF and hair-cortisol (Table 2: A).

Since BDNF and cortisol are frequently measured in serum, hair-levels were correlated to their

simultaneously obtained levels in serum. No correlation was observed of either with morning-

serum levels of both stress mediators. Control of potentially confounding basic population

characteristics (Table 2: B) confirmed correlational analysis. Key regression analysis findings are

shown unadjusted (Figure 1).

Since BDNF was previously shown to associate with alterations in hair biology (Peters et al.,

2005), we assessed hair cycling within the study I hair sample, and found a positive correlation

between hair-BDNF and hair tips with morphology corresponding to the growth phase of the hair

cycle (anagen), and a negative correlation with hair tips corresponding to the transition between

the growth and resting phase of the hair cycle (catagen) (Table 2: B). No association was found

between hair-cortisol and hair-biology measures. Therefore, low BDNF but not low cortisol

correlated with high hair follicle transition rate between the growth stage (anagen) and the resting

stage (telogen), which is evidence for low tissue regeneration capacity (Peters et al., 2017).

In the analysis of the self-report assessment data, we found hair-BDNF did not correlate with

present mood (MDMQ), or perceived stress over the preceeding four weeks (PSQ), state anxiety

(STAI), or symptoms of depression (HADS), all of which did correlate with hair-cortisol. By

contrast, a negative correlation between hair-BDNF and somatic complaints (SOMS) present

during the past 7 days, was found. Thereby, mean SOMS `ICD10 somatization index levels and

HADS `depression reported were below cut-off for clinically manifest somatoform disorder or

depression (Herrmann, 1997; Rief et al., 1998) (Table 2: C).

16
Harb et al.

Since simultaneous capture of biomarkers to assess allostatic load is a growing issue in stress

research, we assessed correlations of hair-BDNF and hair-cortisol with key mediators of innate

and adaptive immune responses (Hostinar and Gunnar, 2013). This analysis revealed a

tendency towards significance for a negative correlation between hair-BDNF and CRP, but no

association between BDNF and any of the assayed cytokines. By contrast, hair-cortisol showed a

positive tendency towards a significant correlation with IL1 and correlated positively with

cytokines indicating both mixed TH1/17 (TH1/17) (TNF) and TH2 (IL4) adaptive immune

responses (Table 2: D).

3.2. BDNF histone acetylation in living hair cells inversely correlates with
somatic complaints

In study II, Spearman correlations calculated in 28 students and academic personal, revealed a

significant negative correlation between hair-BDNF gene P4-promoter histone H4 acetylation and

the SOMS `ICD10 somatization index (Table 3), while no correlation was observed with H3

acetylation. No correlation with either BDNF histone H3 or H4 acetylation was found with MDMQ,

PSQ, STAI or HADS. Again, association analysis was robust with respect to confounding factors

(Figure 2).

3.3. During academic stress, individuals with high somatic complaint have low
BDNF protein levels in hair

In study III we found significantly higher mental distress and lower hair-BDNF in volunteers with

high PHQ-15 levels (between 10-30) as opposed to lower mental distress and higher hair-BDNF

in volunteers with low PHQ-15 levels (between 0-9), but no differences in hair-cortisol (Fig. 3).

Inverse correlation of hair-BDNF with hair-cortisol but not `tension-anxiety, PSS `summary

score, STAI `state or BDI `summary score confirmed results of study I (Table 4). In addition, an

inverse correlation of hair-BDNF with POMS `fatigue-inertia as observed.

17
Harb et al.

4 Discussion

To define stress-parameters in hair beyond the established cortisol measures is a worthwhile

endeavor in stress research. To our knowledge, in study I we demonstrated for the first time that

BDNF can be detected in hair, and that low levels associate with high somatic complaints in

healthy individuals. This association is of medium strength, and further links with low tissue

regeneration capacity (hair cycle) and potentially with increased cell damage (high CRP).

Thereby, BDNF incorporation into the hair shaft over a period of four weeks showed a strong

negative correlation with the established hair-stress marker cortisol, which itself positively

correlated with high mental distress and inflammatory markers. This suggests that BDNF and

cortisol associate with different dimensions of self-reported distress, and supports the assertion

that continuously elevated cortisol is associated with a pro-inflammatory state (Goldsmith et al.,

2016), while BDNF is a growth factor, which protects from cellular damage. Underlying this

observation we found in a second study, that epigenetic regulation of the BDNF gene may occur,

specifically of low histone H4 acetylation which corresponds to higher DNA compaction and a

less permissive transcriptional status (Harb et al., 2016; Potaczek et al., 2017). Finally, hair-

BDNF protein detection was successful in a third independent study performed on a different

continent from studies I and II, and had a more diverse study population exposed to academic

stress. This suggests a certain stability of this measure under a variety of conditions.

At present, however, these observations constitute a pilot study and address only young healthy

individuals, mostly female students. Testing with a limited population is often done to test new

outcomes for stress research (Iglesias et al., 2015; Ullmann et al., 2016), but it is a limitation

when considering more generalizability of the results (Wright et al., 2015). Further, since both

hair-BDNF and hair-cortisol did not correlate with measurements of the same stress-mediator

transiently traceable in the blood of the healthy individuals, continuous incorporation of these

distress mediators into hair appears not to be congruent with measurement in blood. A number of

18
Harb et al.

additional validation steps can be proposed to confirm hair-BDNF measurement as an indicator

of distress, such as: additional measurement of BDNF in the second and third cm of hair,

comparison of hair-BDNF in hair from different scalp regions, or comparison of hair-levels with

area under the curve of repeated blood or salivary BDNF-measurements, covering the 4 weeks

prior to hair sampling. Moreover, the results need to be confirmed in more diverse and larger

study populations.

With respect to the association of BDNF and cortisol with different dimensions of distress, it must

be noted that in a number of studies an association between hair-cortisol and distress could not

be found in healthy individuals. In mothers living under socioeconomically difficult conditions for

example, hair-cortisol correlated with obesity - a condition found frequently among this population

- but not with perceived stress scale scores measuring uncontrollable stress over the past four

weeks (Olstad et al., 2016). Other authors found only a weak positive correlation for example in

subjects exposed to noise stress (Michaud et al., 2016) or work related stress (Gidlow et al.,

2016). In addition, low BDNF in the brain makes for a good marker of experimentally induced

depressive behavior, and its treatment in animal models of depression, and low BDNF in

peripheral blood has been reported in mental distress patients in association with high

depression scores in many studies (Macedo et al., 2015; Munkholm et al., 2016). We found a

robust association of hair-cortisol with a questionnaire addressing tension-, demands-, worries-

and joy present over the 4 weeks prior to assessment, although we found no association

between hair-BDNF and depression scores. Since we only included healthy subjects, with body

weight within generally accepted parameters, living in good socioeconomic conditions, and with

little stress in life, the results may only apply in this context.

However, while reported serum- or plasma-BDNF levels in the context of depression were not

always conclusive and sometimes even contradictory, low BDNF appears to quite consistently

associate with somatic complaints in health and disease. To discuss this phenomenon in more

19
Harb et al.

detail, the authors of a meta-analysis of 35 studies on BDNF and depression found substantial,

unexplained, between-study heterogeneity and other indications of potentially biased results

(Munkholm et al., 2016). Though respective studies are numbered, those reporting BDNF and

somatic complaints point in one direction: in healthy individuals, low serum-BDNF was not found

associated with symptoms of depression or anxiety while an association between low plasma-

BDNF and high somatization scores, measured with the symptom check list (SCL-90), was

reported (Bhang et al., 2012; Tirassa et al., 2012); the interaction between overprotection by

parents and a BDNF gene variant resulting in low BDNF-levels predicted presence of somatic

complaints measured with the brief symptom inventory in the healthy adult offspring (Ibarra et al.,

2014); BDNF was associated with measures indicating increased fatigue in healthy individuals

(Minelli et al., 2011), which is usually associated with somatic complaints; finally, alongside self-

reported symptoms of depression and anxiety, low BDNF is associated with high somatic

complaints in chronically depressed patients (Oglodek et al., 2016; Satomura et al., 2011) and

suggestive data for low BDNF in patients with somatoform disorder have been published (Kang

et al., 2016; Su et al., 2015).

With respect to potential sources of BDNF to be incorporated into hair, BDNF may be at low

levels both due to low levels in the circulation as well as low production by hair follicle epithelial

cells themselves. Various distress mouse models suggest that a low BDNF level in peripheral

blood corresponds to even lower levels in distress-affected BDNF-producing brain regions and

other tissues (Macedo et al., 2015). The hair follicle is a neuroendocrine organ that displays

striking neuroplasticity in response to stress. Both mouse experiments and human hair follicle

organ culture experiments demonstrated BDNF production and signaling is present in hair follicle

cells and altered in these cells by stress and its mediators (Botchkarev et al., 2004; Peters et al.,

2005). Such stress-induced changes of BDNF-production by hair follicle constituting cells are

associated with changes in hair follicle innervation as well as hair growth. In a mouse model for

20
Harb et al.

stress-induced premature transition of hair follicles from the growth stage (anagen) to the resting

stage (telogen), a higher density of nerve fibers was reported and the related stress-mediator

altered neurotrophin signaling and hair follicle viability in cultured human hair follicles (Peters et

al., 2006; Peters et al., 2007).

In this line of thought, it is of note that CRP opsonizes damaged cells and initiates their

elimination by complement activation and adaptive immune responses (Ablij and Meinders,

2002). CRP therefore indicates cell-damage caused, for example, by oxidative stress, a common

feature of chronic mental distress (Haapakoski et al., 2015). BDNF can balance oxidative stress

and cell damage. Treatment of cultured endothelial cells with BDNF reduces CRP levels and

inhibits oxidative stress and CRP-induced DNA damage (Noren Hooten et al., 2015) while vice

versa CRP downregulates BDNF in neuroblastoma cells (Chen et al., 2014). It is therefore not

surprising that in a number of mental distress studies, low BDNF levels in the circulation were

detected along with high CRP (Dooley et al., 2016), indicating the loss of the organisms capacity

to control damage and maintain growth.

Moreover, the negative association between hair-BDNF and hair-cortisol is striking. An inverse

relationship between cortisol and BDNF-protein or BDNF-expression altering genetic variations

detected in blood was reported to be a prerequisite for disease development by others. A variant

of the BDNF gene resulting in low BDNF production was, for example, associated with high

cortisol in response to stress in traumatized children and adults and such individuals are more

prone to development of a wide variety of non-communicable diseases later in live, including

depression (Armbruster et al., 2016; Colzato et al., 2011; Jiang et al., 2017; Simsek et al., 2015;

Tsuru et al., 2014). Animal experiments confirm this association: in mice stressed by maternal

separation, high striatal corticosterone correlated with low BDNF (Mpofana et al., 2016) and such

mice are also more vulnerable in respective distress and disease models (Bohacek and Mansuy,

2013). Hence, changes in hair-BDNF may mark the degree to which stress-damage has occurred

21
Harb et al.

in sensitive tissues in the presence of a stress-compromised HPA function (Revest et al., 2014),

paving the way for pathogenic developments.

5 Conclusion

The hair follicle is an easily accessible neurobiological unit (Munkholm et al., 2016), the BDNF

expression of which links with plasticity of peripheral innervation (Rutlin et al., 2014), stress

(Peters et al., 2012), tissue regeneration (Peters et al., 2012) and altered immune responses

(Botchkarev et al., 2004). Accordingly, in the studies reported here, we show that in healthy

individuals hair-BDNF associates primarily with somatic complaints as well as indicators of tissue

damage.Thisfinding has potential implications for the study of non-communicable diseases and

mental distress disorders. Assessment of BDNF in hair simultaneously with assessment of

mental distress and somatic complaints may provide new insights into the role of BDNF in

maladaptive stress responses and the associated neurobiological changes. Without invasive

sampling and in a field setting, alterations in HPA functioning may be measurable simultaneously

with neurotrophin expression in a relevant, stress-sensitive neurobiological tissue. This could

facilitate monitoring of treatment for example with histone acetylation inhibitors (Meylan et al.,

2016), TrkB-mimics or ligands or anti-depressants (Boulle et al., 2016). It could also facilitate the

study of the effects of life-style changes such as dietary change and physical activity change, or

the effects of psychotherapeutic intervention (Dotson et al., 2016; Tyagi et al., 2015). Moreover,

the epigenetic impact of distress on BDNF and its treatment options may be investigated in

relevant neuroendocrine tissue (Seo et al., 2016). To improve validation of the measure, future

BDNF studies may want to analyze BDNF in the most proximal cm as well as hair more distal to

the scalp to learn if prolonged stress exposures can be monitored sequentially, analog to hair-

cortisol. In addition, repeated multidimensional distress assessments in healthy individuals

compared to mental distress patients could clarify, which dimensions of self-reported distress are

consistently reflected by changes in hair-BDNF levels.

22
Harb et al.

6 Conflicts of Interest

In relation to the work described here, the authors have none.

7 Authors contributions

EP conceptualized and designed the study. MR, JK, DPP provided general support. EP, HR,

JK, MR provided funding and supervised all stages of the study. MGV, LT, UK, EP recruited

patients. MR, JK provided clinical supervision for the study. HH, MGV, LT, UK, EP collected the

data. HH, MGV, DPP performed analysis of the biological samples under the supervision of

EMJP and HR. EP performed data processing and statistical analysis with the help of Johannes

Herrmann (http://www.statistikberatung-giessen.de/). EP, MR, DPP interpreted the data. EP

drafted the manuscript and revision and coordinated the manuscript writing. All authors read,

revised critically and approved the final manuscript.

8 Acknowledgments and role of the funding sources

We gratefully acknowledge the professional language editing services of Annette Smith,

Toronto, Canada and the technical support of Sandra Laux, Susanne Tumala and Cand. med.

Yvonne Mller, Psychoneuroimmunology Laboratory, Giessen, Germany. The conduct of the

research was partially supported by funds from the Von Behring-Rntgen-Foundation (Von

Behring-Rntgen-Stiftung), the German Centre for Lung Research (DZL; 82DZL00502), and the

Universities Giessen and Marburg Lung Centre (UGMLC). Funders had no involvement in study

design; collection, analysis and interpretation of data; writing of the report; and decision to

submit the article for publication.

23
Harb et al.

References

Ablij, H., Meinders, A., 2002. C-reactive protein: history and revival. Eur J Intern Med
13, 412.

Adalsteinsson, B.T., Gudnason, H., Aspelund, T., Harris, T.B., Launer, L.J., Eiriksdottir,
G., Smith, A.V., Gudnason, V., 2012. Heterogeneity in white blood cells has potential to
confound DNA methylation measurements. PloS one 7, e46705.

Ambrus, L., Lindqvist, D., Traskman-Bendz, L., Westrin, A., 2016. Hypothalamic-
pituitary-adrenal axis hyperactivity is associated with decreased brain-derived
neurotrophic factor in female suicide attempters. Nord J Psychiatry, 1-7.

Andrade, E., Arce, C., Torrado, J., Garrido, J., De Francisco, C., Arce, I., 2010. Factor
structure and invariance of the POMS Mood State Questionnaire in Spanish. Span J
Psychol 13, 444-452.

Armbruster, D., Muller-Alcazar, A., Strobel, A., Lesch, K.P., Kirschbaum, C., Brocke, B.,
2016. BDNF val(66)met genotype shows distinct associations with the acoustic startle
reflex and the cortisol stress response in young adults and children.
Psychoneuroendocrinology 66, 39-46.

Bhang, S.Y., Kim, K., Choi, S.W., Ahn, J.H., 2012. Do levels of brain-derived
neurotrophic factor (BDNF) in plasma correlate with psychopathology in healthy
subjects? Neuroscience letters 512, 72-77.

Bohacek, J., Mansuy, I.M., 2013. Epigenetic inheritance of disease and disease risk.
Neuropsychopharmacology : official publication of the American College of
Neuropsychopharmacology 38, 220-236.

Botchkarev, V.A., Botchkareva, N.V., Peters, E.M., Paus, R., 2004. Epithelial growth
control by neurotrophins: leads and lessons from the hair follicle. Prog Brain Res 146,
493-513.

Boulle, F., Pawluski, J.L., Homberg, J.R., Machiels, B., Kroeze, Y., Kumar, N.,
Steinbusch, H.W., Kenis, G., van den Hove, D.L., 2016. Developmental fluoxetine
24
Harb et al.

exposure increases behavioral despair and alters epigenetic regulation of the

hippocampal BDNF gene in adult female offspring. Hormones and behavior 80, 47-57.

Bucker, J., Fries, G.R., Kapczinski, F., Post, R.M., Yatham, L.N., Vianna, P., Bogo
Chies, J.A., Gama, C.S., Magalhaes, P.V., Aguiar, B.W., Pfaffenseller, B., Kauer-
Sant'Anna, M., 2015. Brain-derived neurotrophic factor and inflammatory markers in
school-aged children with early trauma. Acta psychiatrica Scandinavica 131, 360-368.

Buske-Kirschbaum, A., Kern, S., Ebrecht, M., Hellhammer, D.H., 2007. Altered
distribution of leukocyte subsets and cytokine production in response to acute
psychosocial stress in patients with psoriasis vulgaris. Brain, behavior, and immunity 21,
92-99.

Chen, C.C., Hsu, L.W., Nakano, T., Goto, S., Chen, C.L., 2014. Elevation of C-reactive
protein level and its correlation with psychiatric comorbidities in recipients after liver
transplantation. Transplant Proc 46, 894-896.

Ciliberti, M.G., Albenzio, M., Inghese, C., Santillo, A., Marino, R., Sevi, A., Caroprese,
M., 2017. Peripheral blood mononuclear cell proliferation and cytokine production in
sheep as affected by cortisol level and duration of stress. Journal of dairy science 100,
750-756.

Cohen, J., 1988. Statistical power analysis for the behavioral sciences. Lawrence
Erlbaum Associates, Mahwah, New Jersey.

Colzato, L.S., Van der Does, A.J., Kouwenhoven, C., Elzinga, B.M., Hommel, B., 2011.
BDNF Val66Met polymorphism is associated with higher anticipatory cortisol stress
response, anxiety, and alcohol consumption in healthy adults.
Psychoneuroendocrinology 36, 1562-1569.

Cuesta, M., Boudreau, P., Dubeau-Laramee, G., Cermakian, N., Boivin, D.B., 2016.
Simulated Night Shift Disrupts Circadian Rhythms of Immune Functions in Humans.
Journal of immunology 196, 2466-2475.

25
Harb et al.

Dhurat, R., Saraogi, P., 2009. Hair evaluation methods: merits and demerits. Int J
Trichology 1, 108-119.

Di Paolo, N.C., Shayakhmetov, D.M., 2016. Interleukin 1alpha and the inflammatory
process. Nature immunology 17, 906-913.

Diez-Quevedo, C., Rangil, T., Sanchez-Planell, L., Kroenke, K., Spitzer, R.L., 2001.
Validation and utility of the patient health questionnaire in diagnosing mental disorders
in 1003 general hospital Spanish inpatients. Psychosomatic medicine 63, 679-686.

Dooley, L.N., Ganz, P.A., Cole, S.W., Crespi, C.M., Bower, J.E., 2016. Val66Met BDNF
polymorphism as a vulnerability factor for inflammation-associated depressive
symptoms in women with breast cancer. Journal of affective disorders 197, 43-50.

Dotson, V.M., Hsu, F.C., Langaee, T.Y., McDonough, C.W., King, A.C., Cohen, R.A.,
Newman, A.B., Kritchevsky, S.B., Myers, V., Manini, T.M., Pahor, M., Life Study, G.,
2016. Genetic Moderators of the Impact of Physical Activity on Depressive Symptoms. J
Frailty Aging 5, 6-14.

Fliege, H., Rose, M., Arck, P., Walter, O.B., Kocalevent, R.D., Weber, C., Klapp, B.F.,
2005. The Perceived Stress Questionnaire (PSQ) reconsidered: validation and
reference values from different clinical and healthy adult samples. Psychosomatic
medicine 67, 78-88.

Gidlow, C.J., Randall, J., Gillman, J., Silk, S., Jones, M.V., 2016. Hair cortisol and self-
reported stress in healthy, working adults. Psychoneuroendocrinology 63, 163-169.

Gola, H., Engler, H., Sommershof, A., Adenauer, H., Kolassa, S., Schedlowski, M.,
Groettrup, M., Elbert, T., Kolassa, I.T., 2013. Posttraumatic stress disorder is associated
with an enhanced spontaneous production of pro-inflammatory cytokines by peripheral
blood mononuclear cells. BMC psychiatry 13, 40.

Goldsmith, D.R., Rapaport, M.H., Miller, B.J., 2016. A meta-analysis of blood cytokine
network alterations in psychiatric patients: comparisons between schizophrenia, bipolar
disorder and depression. Molecular psychiatry 21, 1696-1709.

26
Harb et al.

Gonzalez-de-la-Vara, R., Valdez, R.A., Lemus-Ramirez, V., Vazquez-Chagoyan, J.C.,

Villa-Godoy, A., Romano, M.C., 2011. Effects of adrenocorticotropic hormone challenge
and age on hair cortisol concentrations in dairy cattle. Can J Vet Res 75, 216-221.

Gonzalez-Ramirez, M.T., Rodriguez-Ayan, M.N., Hernandez, R.L., 2013. The perceived

stress scale (PSS): normative data and factor structure for a large-scale sample in
Mexico. Span J Psychol 16, E47.

Haapakoski, R., Mathieu, J., Ebmeier, K.P., Alenius, H., Kivimaki, M., 2015. Cumulative
meta-analysis of interleukins 6 and 1beta, tumour necrosis factor alpha and C-reactive
protein in patients with major depressive disorder. Brain, behavior, and immunity 49,
206-215.

Hansel, A., Hong, S., Camara, R.J., von Kanel, R., 2010. Inflammation as a
psychophysiological biomarker in chronic psychosocial stress. Neuroscience and
biobehavioral reviews 35, 115-121.

Harb, H., Alashkar Alhamwe, B., Garn, H., Renz, H., Potaczek, D.P., 2016. Recent
developments in epigenetics of pediatric asthma. Current opinion in pediatrics 28, 754-
763.

Harb, H., Amarasekera, M., Ashley, S., Tulic, M.K., Pfefferle, P.I., Potaczek, D.P.,
Martino, D., Kesper, D.A., Prescott, S.L., Renz, H., 2015a. Epigenetic Regulation in
Early Childhood: A Miniaturized and Validated Method to Assess Histone Acetylation.
International archives of allergy and immunology 168, 173-181.

Harb, H., Raedler, D., Ballenberger, N., Bock, A., Kesper, D.A., Renz, H., Schaub, B.,
2015b. Childhood allergic asthma is associated with increased IL-13 and FOXP3
histone acetylation. The Journal of allergy and clinical immunology 136, 200-202.

Herrmann, C., 1997. International experiences with the Hospital Anxiety and Depression
Scale--a review of validation data and clinical results. Journal of psychosomatic
research 42, 17-41.

27
Harb et al.

Hofer, D., Thoma, M.V., Schmidlin, P.R., Attin, T., Ehlert, U., Nater, U.M., 2016. Pre-
treatment anxiety in a dental hygiene recall population: a cross-sectional pilot study.
BMC oral health 16, 43.

Hoffman, M.C., Karban, L.V., Benitez, P., Goodteacher, A., Laudenslager, M.L., 2014.
Chemical processing and shampooing impact cortisol measured in human hair. Clinical
and investigative medicine. Medecine clinique et experimentale 37, E252-257.

Hong, M., Zheng, J., Ding, Z.Y., Chen, J.H., Yu, L., Niu, Y., Hua, Y.Q., Wang, L.L.,
2013. Imbalance between Th17 and Treg cells may play an important role in the
development of chronic unpredictable mild stress-induced depression in mice.
Neuroimmunomodulation 20, 39-50.

Hostinar, C.E., Gunnar, M.R., 2013. Future directions in the study of social relationships
as regulators of the HPA axis across development. J Clin Child Adolesc Psychol 42,
564-575.

Ibarra, P., Alemany, S., Fatjo-Vilas, M., Cordova-Palomera, A., Goldberg, X., Arias, B.,
Gonzalez-Ortega, I., Gonzalez-Pinto, A., Nenadic, I., Fananas, L., 2014. The BDNF-
Val66Met polymorphism modulates parental rearing effects on adult psychiatric
symptoms: a community twin-based study. European psychiatry : the journal of the
Association of European Psychiatrists 29, 293-300.

Iglesias, S., Jacobsen, D., Gonzalez, D., Azzara, S., Repetto, E.M., Jamardo, J.,
Gomez, S.G., Mesch, V., Berg, G., Fabre, B., 2015. Hair cortisol: A new tool for
evaluating stress in programs of stress management. Life sciences 141, 188-192.

Jiang, R., Babyak, M.A., Brummett, B.H., Siegler, I.C., Kuhn, C.M., Williams, R.B.,
2017. Brain-derived neurotrophic factor (BDNF) Val66Met polymorphism interacts with
gender to influence cortisol responses to mental stress. Psychoneuroendocrinology 79,
13-19.

Kamezaki, Y., Katsuura, S., Kuwano, Y., Tanahashi, T., Rokutan, K., 2012. Circulating
cytokine signatures in healthy medical students exposed to academic examination
stress. Psychophysiology 49, 991-997.

28
Harb et al.

Kang, N.I., Park, J.I., Kim, Y.K., Yang, J.C., 2016. Decreased Plasma BDNF Levels of
Patients with Somatization Disorder. Psychiatry investigation 13, 526-530.

Karlen, J., Ludvigsson, J., Frostell, A., Theodorsson, E., Faresjo, T., 2011. Cortisol in
hair measured in young adults - a biomarker of major life stressors? BMC Clin Pathol
11, 12.

Kasemeier-Kulesa, J.C., Morrison, J.A., Lefcort, F., Kulesa, P.M., 2015. TrkB/BDNF
signalling patterns the sympathetic nervous system. Nat Commun 6, 8281.

Kirschbaum, C., Tietze, A., Skoluda, N., Dettenborn, L., 2009. Hair as a retrospective
calendar of cortisol production-Increased cortisol incorporation into hair in the third
trimester of pregnancy. Psychoneuroendocrinology 34, 32-37.

Lenarczyk, A., Helsloot, J., Farmer, K., Peters, L., Sturgess, A., Kirkham, B., 2000.
Antigen-induced IL-17 response in the peripheral blood mononuclear cells (PBMC) of
healthy controls. Clinical and experimental immunology 122, 41-48.

Li, J., Xie, Q., Gao, W., Xu, Y., Wang, S., Deng, H., Lu, Z., 2012. Time course of cortisol
loss in hair segments under immersion in hot water. Clinica chimica acta; international
journal of clinical chemistry 413, 434-440.

Macedo, I.C., Rozisky, J.R., Oliveira, C., Oliveira, C.M., Laste, G., Nonose, Y., Santos,
V.S., Marques, P.R., Ribeiro, M.F., Caumo, W., Torres, I.L., 2015. Chronic stress
associated with hypercaloric diet changes the hippocampal BDNF levels in male Wistar
rats. Neuropeptides 51, 75-81.

Makhathini, K.B., Abboussi, O., Stein, D.J., Mabandla, M.V., Daniels, W.M.U., 2017.
Repetitive stress leads to impaired cognitive function that is associated with DNA
hypomethylation, reduced BDNF and a dysregulated HPA axis. International journal of
developmental neuroscience : the official journal of the International Society for
Developmental Neuroscience 60, 63-69.

McEwen, B.S., Bowles, N.P., Gray, J.D., Hill, M.N., Hunter, R.G., Karatsoreos, I.N.,
Nasca, C., 2015. Mechanisms of stress in the brain. Nat Neurosci 18, 1353-1363.

29
Harb et al.

Meyer, J., Novak, M., Hamel, A., Rosenberg, K., 2014. Extraction and analysis of
cortisol from human and monkey hair. Journal of visualized experiments : JoVE,
e50882.

Meylan, E.M., Halfon, O., Magistretti, P.J., Cardinaux, J.R., 2016. The HDAC inhibitor
SAHA improves depressive-like behavior of CRTC1-deficient mice: Possible relevance
for treatment-resistant depression. Neuropharmacology 107, 111-121.

Michaud, D.S., Feder, K., Keith, S.E., Voicescu, S.A., Marro, L., Than, J., Guay, M.,
Denning, A., Bower, T., Villeneuve, P.J., Russell, E., Koren, G., van den Berg, F., 2016.
Self-reported and measured stress related responses associated with exposure to wind
turbine noise. The Journal of the Acoustical Society of America 139, 1467-1479.

Minelli, A., Zanardini, R., Bonvicini, C., Sartori, R., Pedrini, L., Gennarelli, M., Bocchio-
Chiavetto, L., 2011. BDNF serum levels, but not BDNF Val66Met genotype, are
correlated with personality traits in healthy subjects. European archives of psychiatry
and clinical neuroscience 261, 323-329.

Mpofana, T., Daniels, W.M., Mabandla, M.V., 2016. Exposure to Early Life Stress
Results in Epigenetic Changes in Neurotrophic Factor Gene Expression in a
Parkinsonian Rat Model. Parkinsons Dis 2016, 6438783.

Munkholm, K., Vinberg, M., Kessing, L.V., 2016. Peripheral blood brain-derived
neurotrophic factor in bipolar disorder: a comprehensive systematic review and meta-
analysis. Molecular psychiatry 21, 216-228.

Natt, D., Johansson, I., Faresjo, T., Ludvigsson, J., Thorsell, A., 2015. High cortisol in 5-
year-old children causes loss of DNA methylation in SINE retrotransposons: a possible
role for ZNF263 in stress-related diseases. Clin Epigenetics 7, 91.

Naveen, G.H., Varambally, S., Thirthalli, J., Rao, M., Christopher, R., Gangadhar, B.N.,
2016. Serum cortisol and BDNF in patients with major depression-effect of yoga. Int
Rev Psychiatry, 1-6.

30
Harb et al.

Newton, T.L., Fernandez-Botran, R., Miller, J.J., Lorenz, D.J., Burns, V.E., Fleming,
K.N., 2011. Markers of inflammation in midlife women with intimate partner violence
histories. Journal of women's health 20, 1871-1880.

Noren Hooten, N., Ejiogu, N., Zonderman, A.B., Evans, M.K., 2015. Protective Effects of
BDNF against C-Reactive Protein-Induced Inflammation in Women. Mediators of
inflammation 2015, 516783.

Oglodek, E.A., Just, M.J., Szromek, A.R., Araszkiewicz, A., 2016. Melatonin and
neurotrophins NT-3, BDNF, NGF in patients with varying levels of depression severity.
Pharmacological reports : PR 68, 945-951.

Oh, H., Ghosh, S., 2013. NF-kappaB: roles and regulation in different CD4(+) T-cell
subsets. Immunological reviews 252, 41-51.

Olstad, D.L., Ball, K., Wright, C., Abbott, G., Brown, E., Turner, A.I., 2016. Hair cortisol
levels, perceived stress and body mass index in women and children living in
socioeconomically disadvantaged neighborhoods: the READI study. Stress 19, 158-
167.

Panchaprateep, R., Korkij, W., Asawanonda, P., 2011. Brain-derived nerve factor and
neurotrophins in androgenetic alopecia. The British journal of dermatology 165, 997-
1002.

Perpina-Galvan, J., Cabanero-Martinez, M.J., Richart-Martinez, M., 2013. Reliability and

validity of shortened state trait anxiety inventory in Spanish patients receiving
mechanical ventilation. Am J Crit Care 22, 46-52.

Peters, E.M., Arck, P.C., Paus, R., 2006. Hair growth inhibition by psychoemotional
stress: a mouse model for neural mechanisms in hair growth control. Experimental
dermatology 15, 1-13.

Peters, E.M., Hansen, M.G., Overall, R.W., Nakamura, M., Pertile, P., Klapp, B.F., Arck,
P.C., Paus, R., 2005. Control of human hair growth by neurotrophins: brain-derived

31
Harb et al.

neurotrophic factor inhibits hair shaft elongation, induces catagen, and stimulates
follicular transforming growth factor beta2 expression. J Invest Dermatol 124, 675-685.

Peters, E.M., Liezmann, C., Klapp, B.F., Kruse, J., 2012. The neuroimmune connection
interferes with tissue regeneration and chronic inflammatory disease in the skin. Annals
of the New York Academy of Sciences 1262, 118-126.

Peters, E.M., Liotiri, S., Bodo, E., Hagen, E., Biro, T., Arck, P.C., Paus, R., 2007.
Probing the effects of stress mediators on the human hair follicle: substance P holds
central position. The American journal of pathology 171, 1872-1886.

Peters, E.M.J., Mller, Y., Snaga, W., Fliege, H., Reihauer, A., Schmidt-Rose, T., Max,
H., Schweiger, D., Rose, M., Kruse, J., 2017. Hair And Stress: A Pilot Study Of Hair And
Cytokine Balance Alteration In Healthy Young Women Under Major Exam Stress. PloS
one, in print.

Potaczek, D.P., Harb, H., Michel, S., Alhamwe, B.A., Renz, H., Tost, J., 2017.
Epigenetics and allergy: from basic mechanisms to clinical applications. Epigenomics 9,
539-571.

Raab, S., Klingenstein, M., Liebau, S., Linta, L., 2014. A Comparative View on Human
Somatic Cell Sources for iPSC Generation. Stem Cells Int 2014, 768391.

Rask-Andersen, M., Bringeland, N., Nilsson, E.K., Bandstein, M., Bcaro, M.O., Vogel,
H., Schrmann, A., Hogenkamp, P.S., Benedict, C., Schith, H.B., 2016. Postprandial
alterations in whole-blood DNA methylation are mediated by changes in white blood cell
composition. American Journal of Clinical Nutrition.

Revest, J.M., Le Roux, A., Roullot-Lacarriere, V., Kaouane, N., Vallee, M., Kasanetz, F.,
Rouge-Pont, F., Tronche, F., Desmedt, A., Piazza, P.V., 2014. BDNF-TrkB signaling
through Erk1/2 MAPK phosphorylation mediates the enhancement of fear memory
induced by glucocorticoids. Molecular psychiatry 19, 1001-1009.

32
Harb et al.

Rief, W., Shaw, R., Fichter, M.M., 1998. Elevated levels of psychophysiological arousal
and cortisol in patients with somatization syndrome. Psychosomatic medicine 60, 198-
203.

Ros Montalban, S., Comas Vives, A., Garcia-Garcia, M., 2010. Validation of the
Spanish version of the PHQ-15 questionnaire for the evaluation of physical symptoms in
patients with depression and/or anxiety disorders: DEPRE-SOMA study. Actas Esp
Psiquiatr 38, 345-357.

Rutlin, M., Ho, C.Y., Abraira, V.E., Cassidy, C., Bai, L., Woodbury, C.J., Ginty, D.D.,
2014. The cellular and molecular basis of direction selectivity of Adelta-LTMRs. Cell
159, 1640-1651.

Satomura, E., Baba, H., Nakano, Y., Maeshima, H., Suzuki, T., Arai, H., 2011.
Correlations between brain-derived neurotrophic factor and clinical symptoms in
medicated patients with major depression. Journal of affective disorders 135, 332-335.

Sauve, B., Koren, G., Walsh, G., Tokmakejian, S., Van Uum, S.H., 2007. Measurement
of cortisol in human hair as a biomarker of systemic exposure. Clinical and investigative
medicine. Medecine clinique et experimentale 30, E183-191.

Scheich, B., Vincze, P., Szke, ., Borbly, ., Hunyady, ., Szolcsnyi, J., Dnes, .,
Krnyei, Z., Gaszner, B., Helyes, Z., 2017 Chronic stress-induced mechanical
hyperalgesia is controlled by capsaicin-sensitive neurones in the mouse. European
journal of pain.

Seo, M.K., Ly, N.N., Lee, C.H., Cho, H.Y., Choi, C.M., Nhu le, H., Lee, J.G., Lee, B.J.,
Kim, G.M., Yoon, B.J., Park, S.W., Kim, Y.H., 2016. Early life stress increases stress
vulnerability through BDNF gene epigenetic changes in the rat hippocampus.
Neuropharmacology 105, 388-397.

Short, S.J., Stalder, T., Marceau, K., Entringer, S., Moog, N.K., Shirtcliff, E.A., Wadhwa,
P.D., Buss, C., 2016. Correspondence between hair cortisol concentrations and 30-day
integrated daily salivary and weekly urinary cortisol measures.
Psychoneuroendocrinology 71, 12-18.

33
Harb et al.

Simsek, S., Yuksel, T., Kaplan, I., Uysal, C., Alaca, R., 2015. Examining the levels of
BDNF and cortisol in children and adolescent victims of sexual abuse--a preliminary
study. Comprehensive psychiatry 61, 23-27.

Spielberger, C.D., Gorsuch, R.L., Lushene, R.E., 1970. Manual for the State-Trait
Anxiety Inventory. Consulting Psychologists Press, Palo Alto, Ca.

Stalder, T., Steudte-Schmiedgen, S., Alexander, N., Klucken, T., Vater, A., Wichmann,
S., Kirschbaum, C., Miller, R., 2017. Stress-related and basic determinants of hair
cortisol in humans: A meta-analysis. Psychoneuroendocrinology 77, 261-274.

Staufenbiel, S.M., Penninx, B.W., Spijker, A.T., Elzinga, B.M., van Rossum, E.F., 2013.
Hair cortisol, stress exposure, and mental health in humans: a systematic review.
Psychoneuroendocrinology 38, 1220-1235.

Su, S., Xiao, Z., Lin, Z., Qiu, Y., Jin, Y., Wang, Z., 2015. Plasma brain-derived
neurotrophic factor levels in patients suffering from post-traumatic stress disorder.
Psychiatry research 229, 365-369.

Tirassa, P., Iannitelli, A., Sornelli, F., Cirulli, F., Mazza, M., Calza, A., Alleva, E.,
Branchi, I., Aloe, L., Bersani, G., Pacitti, F., 2012. Daily serum and salivary BDNF levels
correlate with morning-evening personality type in women and are affected by light
therapy. Rivista di psichiatria 47, 527-534.

Tost, J., 2016. Follow the trace of death: methylation analysis of cell-free DNA for
clinical applications in non-cancerous diseases. Epigenomics 8, 1169-1172.

Tsuru, J., Tanaka, Y., Ishitobi, Y., Maruyama, Y., Inoue, A., Kawano, A., Ikeda, R.,
Ando, T., Oshita, H., Aizawa, S., Masuda, K., Higuma, H., Kanehisa, M., Ninomiya, T.,
Akiyoshi, J., 2014. Association of BDNF Val66Met polymorphism with HPA and SAM
axis reactivity to psychological and physical stress. Neuropsychiatric disease and
treatment 10, 2123-2133.

34
Harb et al.

Tyagi, E., Zhuang, Y., Agrawal, R., Ying, Z., Gomez-Pinilla, F., 2015. Interactive actions
of Bdnf methylation and cell metabolism for building neural resilience under the
influence of diet. Neurobiol Dis 73, 307-318.

Ullmann, E., Barthel, A., Petrowski, K., Stalder, T., Kirschbaum, C., Bornstein, S.R.,
2016. Pilot study of adrenal steroid hormones in hair as an indicator of chronic mental
and physical stress. Scientific reports 6, 25842.

Wester, V.L., van Rossum, E.F., 2015. Clinical applications of cortisol measurements in
hair. Eur J Endocrinol 173, M1-10.

Wosu, A.C., Valdimarsdottir, U., Shields, A.E., Williams, D.R., Williams, M.A., 2013.
Correlates of cortisol in human hair: implications for epidemiologic studies on health
effects of chronic stress. Ann Epidemiol 23, 797-811 e792.

Wright, K.D., Hickman, R., Laudenslager, M.L., 2015. Hair Cortisol Analysis: A
Promising Biomarker of HPA Activation in Older Adults. Gerontologist 55 Suppl 1, S140-
145.

Xiang, L., Del Ben, K.S., Rehm, K.E., Marshall, G.D., Jr., 2012. Effects of acute stress-
induced immunomodulation on TH1/TH2 cytokine and catecholamine receptor
expression in human peripheral blood cells. Neuropsychobiology 65, 12-19.

35
Harb et al.

9 Figure captions

Figure 1: Regression analysis of hair-brain derived neurotrophic factor (BDNF) protein

and hair-cortisol with self-report distress. F, degrees of freedom and p-values are indicated in

the figure. A: r2 = 0.147. B: r2 = 0.292.

36
Harb et al.

Figure 2: Regression analysis of H4 histone acetylation of brain derived neurotrophic

factor (BDNF) with SOMS. F, degrees of freedom and p-values are indicated in the figure. r2 =

0.274.

37
Harb et al.

A PSS B BDNF C c o r t is o l

50 *** 60 0 .8
*
40

n g /m l h a ir e x t r a c t
n g /m l h a ir e x t r a c t
0 .6
40
30
s c o re

0 .4

20
20
0 .2
10

0 0 0 .0
PHQ- PHQ+ PHQ- PHQ+ PHQ- PHQ+

Figure 3: Differences in hair- brain derived neurotrophic factor (BDNF) between Mexican

students with low and high PHQ-15 scores. Mean values and SEM are shown. * indicates

Mann Whitney U test result, P-values <0.05 one asterisk, < 0.001 - three asterisk.

38
Harb et al.

10 Tables

Table 1: A - Conditions of participant recruitment.

Study I Study II Study III
students in last 2 years students and academic mid-end education
Ocupation of education personal (semester 3, 5 or 8)
during winter semester between July and during exams at the end
Timepoint break October of the semester
in identical phases of under hormonal no control of estrus
the estrus cycle contraception cycle or contraception
baseline correlations of analysis of epigenetic analysis of hair-BDNF
Aim hair-BDNF/cortisol with: changes relevant for levels under field
serum-BDNF/cortisol, BDNF synthesis in conditions in individuals
hair biology, self-report relation to key self-report with high-versus low
measures, key markers measures of distress measures of distress
of immune function identified in study I identified in study I
Mode of advertisement at University blackboards classrooms at beginning
recruitment university internet websites of semester by personal
information mails via the university email network invitation (MG)
Inclusion written informed consent written informed
criteria successful participation in academic study/occupation consent
at least shoulder long, pigmented hair
Exclusion acute inflammatory disease acute inflammatory
criteria substance-abuse disease
presence of trauma or significant life stressors chronic illness requiring
chronic illness requiring medical attention medical attention
technical issues (e.g. blood lysis contamination, technical issues (e.g.
missing data) contamination, missing
hair washed or combed in past 2-3 days data)
hair dyed during the 10 days before sampling
Conditions 6-8 participants were invited per assessment to come any number of
of sample to the psychoneuroimmunology laboratory at 8 oclock participants were invited
collection in the morning (experimental conditions) to donate samples and
no coffee or tea, smoking or doing sports that morning file questionnaires in
blood was taken in a quiet atmosphere 30 min after their classroom during a
arrival at laboratory morning or afternoon
hair samples were taken at the end of visit teaching session
Plucking a row of hairs fastened by the distal end with a rubber a row of hairs was
and coated surgical clamp and was plucked in one swift fastened by the distal
clipping movement end with a surgical
details clamp and clipped at
the level of the scalp
Extraction cortisol: 1 ml of methanol/10 mg pulverized hair, shaker for 2 hrs/300 shakes per
details min, incubation overnight, shaker for 2 hrs (total extraction time 18.5 hrs),
evaporated in vacuum centrifuge (Savant Speed Vac. Centrifuge Evaporator,
Farmingdale, NY, USA) until dry (approximately 6-8 hrs)
BDNF: 220 l 0.1 M citric acid (Sigma, Sigma-Aldrich Corporation, St. Louis, MO,
USA) in 50% ethanol per 10 mg pulverized hair, shaker for 30 min/150 shakes per
min, lyophilised in parafilm covered tubes over weekend after freezing (Lyovac
GT2, Leybold GmbH, Cologne, Germany)
Compen- participants received 50 Euros per completed none
sation assessment

39
Harb et al.

Table 1: B - Baseline characteristics of study participants. Note that all participants did not

take oral and topical medication affecting the neuro-endocrine systems or suffered from any

disease requiring medical attention. Besides here reported results, participants served for

comparison in studies on stress experiences by patients suffering from chronic inflammatory skin

diseases or mental distress disorders to be reported elsewhere.

Study I Study II Study III

N 36 28 115
% female participants 100 96.4 78.3
mean sd mean sd mean sd
Age in years 25.91 2.27 25.85 4.87 21.93 2.23

BMI 21.32 2.19 21.64 2.54 23.72 3.45

Cigarettes per day 0.38 1.72 0.00 0.00 0.45 1.47

Cups of coffee / tea per 2.55 1.74 0.71 0.46 0.90 0.94
day

40
Harb et al.

Table 2: Spearmann correlations of hair-BDNF protein levels in 36 healthy female german

medical students during semester break. Significant correlations are bold, tendencies in

italics.

Spearmann Correlations
Hair-BDNF Hair-cortisol
in ng/ml hair in g/dl hair
extract extract
A - endocrine markers N Mean SD rho p rho p

Hair-cortisol in g/dl extract* 31 0.014 0.009 -0.511 0.004

Hair-BDNF in ng/ml extract** 36 26.47 11.59 -0.511 0.004

Serum cortisol in ng/ml 33 361.5 280.0 -0.059 0.743 -0.168 0.394

Serum BDNF in ng/ml 33 48.95 31.35 -0.287 0.105 0.300 0.121

B - hair biology markers N Mean SD rho p rho p

anagen hair bulbs in % all bulbs 32 72.58 12.49 0.433 0.012 -0.240 0.219

telogen hair bulbs in % all bulbs 32 9.251 7.128 -0.080 0.659 0.098 0.621

hair bulbs in transition in % all bulbs 32 18.16 11.15 -0.517 0.002 0.325 0.091

C - self-report measures N Mean SD rho p rho p

MDMQ `calm-nervous 33 30.28 5.43 0.262 0.141 -0.419 -0.026

SOMS `ICD10 somatization index 33 3.03 2.57 -0.480 0.005 0.292 0.131

PSQ `summary score 33 30.00 15.67 -0.176 0.328 0.595 0.000

STAI `state 33 37.06 9.36 -0.213 0.235 0.529 0.004

HADS `depression*** 33 2.64 2.41 0.026 0.887 0.487 0.008

D - inflammatory markers N Mean SD rho p rho p

Serum CRP in g/ml 33 15.58 10.59 -0.332 0.059 0.053 0.789

IL1 in pg/ml 33 156.0 147.5 0.270 0.129 0.328 0.088

IL6 in pg/ml 33 2.382 1.502 0.173 0.335 0.024 0.903

TNF in pg/ml 33 6.008 2.827 -0.232 0.193 0.398 0.036

IFN in pg/ml 33 7.739 2.228 -0.248 0.164 0.190 0.333

IL4 in pg/ml 33 90.01 59.65 -0.138 0.444 0.396 0.037

*Extract was generated by adding 1 ml of methanol to 10 mg of pulverized hair followed by lyophilization and
suspension in 110 l of PBS. ** Extract was generated by adding 220 l 0.1 M citric acid in 50% ethanol to 10 mg of
pulverized hair followed by lyophilization and suspension in 110 l of PBS.
*** cut-off for clinically manifest depression is 8.

41
Harb et al.

Table 3: Spearmann correlations of hair-BDNF acetylation levels in 28 students and

academic personel.

Spearmann Correlations
Hair bulb BDNF Hair bulb BDNF
H3 acetylation H4 acetylation
Self-report measures N Mean SD rho p rho p

MDMQ `calm-nervous 27 29.59 6.53 -0.031 0.880 0.260 0.191

SOMS `ICD10 somatization index 28 8.71 7.17 0.270 0.166 -0.420 0.026

PSQ `summary score 28 43.63 20.37 0.053 0.788 -0.270 0.164

STAI `state 28 38.43 12.72 0.117 0.553 -0.264 0.174

HADS `depression 28 6.04 3.87 0.062 0.752 -0.162 0.408

Table 4: Spearmann correlations of hair-BDNF acetylation levels in 115 healthy Mexican

students during end of semester stress.

Spearmann Correlations
Hair-BDNF
in ng/ml hair extract**
N Mean SD rho p

Hair-cortisol in g/dl extract* 115 0.224 0.149 -0.264 0.019

POMS `fatigue-inertia 115 15.443 5.524 -0.307 0.006

POMS `tension-anxiety 115 23.617 6.062 -0.132 0.248

PSS `summary score 115 25.565 8.131 -0.069 0.548

STAI `state 115 17.783 4.414 -0.092 0.428

BDI `summary score*** 115 9.881 7.372 -0.066 0.563

*Extract was generated by adding 1 ml of methanol to 10 mg of pulverized hair followed by lyophilization and
suspension in 110 l of PBS.
** Extract was generated by adding 220 l 0.1 M citric acid in 50% ethanol to 10 mg of pulverized hair followed by
lyophilization and suspension in 110 l of PBS.
*** Cut-off for clinically manifest depression is 11.

42