ACLS
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kim, S. J.
Right arrow Articles by Rockett, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kim, S. J.
Right arrow Articles by Rockett, J. C.
Annals of Clinical & Laboratory Science 36:115-126 (2006)
© 2006 Association of Clinical Scientists

Gene Expression in Head Hair Follicles Plucked from Men and Women

Sung Jae Kim1,2, David J. Dix1, Kary E. Thompson1,*, Rachel N. Murrell1,3,{dagger}, Judith E. Schmid1, Jane E. Gallagher1 and John C. Rockett1,#
1 Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC; 2 Curriculum in Toxicology, University of North Carolina, Chapel Hill, NC; and 3 Department of Environmental and Molecular Toxicology, North Carolina State University, Raleigh, North Carolina

Address correspondence to David J. Dix, Ph.D., National Center for Computational Toxicology (D343-03), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA; tel 919 541 2701; fax 919 541 1194; e-mail dix.david{at}epa.gov.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Acknowledgments and Disclaimers
 References
 
Characterizing gene expression in hair follicles can help to elucidate the hair growth cycle by delineating the genes and pathways involved in follicular growth and degeneration. The objectives of this study were to determine whether intact RNA could be extracted from a small number of plucked, unstaged hair follicles in sufficient quantity to conduct gene expression profiling, and to conduct global gene expression profiling. To this end, RNA was extracted from 1 to 3 unstaged follicles plucked from the scalp of 36 volunteers. The average quantifiable yield of RNA/follicle was 112.5 ng. Ribosomal ratios were lower than normally expected, but investigation indicated the RNA was intact. Ten of the samples were amplified and hybridized to Affymetrix genechips. On average, 2,567 of the total probe sets (8,500) were expressed in each sample; 1,422 were expressed in all 10 samples; 97 were significantly changed in one gender compared to the other, and 41 had high levels of interindividual variability. This study demonstrates that RNA of sufficient quantity and quality to use in microarray hybridizations can be obtained from as little as a single plucked human hair follicle. Genes expressed in all individuals are probably related to follicular growth and could form a starting set for developing signatures of toxicant exposure. The differentially expressed genes could be involved in producing gender and interindividual differences in hair growth.

Keywords: hair follicle, human, adult, gene expression, microarray


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Acknowledgments and Disclaimers
 References
 
Gene expression profiling (GEP) has become an increasingly important tool for: (a) understanding how cells and tissues function under normal conditions; (b) elucidating molecular mechanisms associated with aging and disease development and progression; and (c) characterizing the responses to toxicological or pharmaceutical exposures. Several groups have used microarray and quantitative real-time polymerase chain reaction (qRT-PCR) to characterize gene expression profiles in hair follicles and skin. Most of this work has been focused on identifying hair follicle cycle-associated genes [1–5] and elucidating the mechanisms of epithelial stem cell differentiation and proliferation [6,7]. The ultimate aim of such studies is to understand hair growth in order to provide potential targets for the treatment of hair loss and other skin disorders.

Hair follicle gene expression profiles might also be used as a surrogate tissue to classify and quantify individual levels of toxicant and/or pharmaceutical exposures. Many groups have already demonstrated that exposures to toxic or pharmaceutical agents cause specific gene expression changes in target and non-target tissues that are characteristic of a particular chemical or class of chemicals [8–12]. Although peripheral blood is the specimen of choice for most surrogate tissue studies, there may be occasions where it is either unavailable or less appropriate than other potential surrogate tissues. Hair follicles may in some cases offer a viable alternative, since they are easy to procure, are available from most individuals, and contain live cells that are a source of RNA required to conduct gene expression profiling.

However, if gene expression profiling of hair follicles is to be used routinely as a tool for clinical analysis, research, or forensic studies, it is important to develop simple, robust methods for the collection, storage, transportation, and analysis of samples. More information is required concerning several aspects of the collection of hair follicles in the clinical environment and their subsequent analysis by gene expression profiling.

Such information includes: (a) the best method(s) to collect, store, and transport hair follicle specimens from the clinic to the laboratory in order both to prevent ex vivo transcription and to inhibit RNA degradation; (b) the best method(s) to extract RNA from the follicles; (c) whether the extracted RNA is of sufficient quality and quantity to be used in microarray-based gene expression profiling; and (d) the range of genes expressed and their level of expression across a population of normal subjects.

To address these questions, we conducted a study in which scalp hairs were plucked from the heads of 36 adult volunteers in a clinical setting. e follicles were then transported to an analytical laboratory and RNA was extracted without regard to the size, condition, or stage of the follicles. RNA quantity and quality were determined, and gene expression profiling was conducted on a subset of the samples using Affymetrix genechips, following a preamplification step.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Acknowledgments and Disclaimers
 References
 
Hair follicle collection and storage.  Following protocol approval from the UNC Chapel Hill Biomedical Institutional Review Board, 1 to 3 hair follicles were extracted from the scalp of 36 fully informed, consenting adult subjects consisting of 19 males (average age 33 yr, range 20–40) and 17 females (average age 25 yr, range 19–40). To do this, individual hairs were grasped as near to the scalp as possible and then yanked quickly out. The bottom centimeter of each hair, containing the hair follicle, was trimmed off into a 5 ml specimen jar filled to the top with RNALater (Ambion, Austin, TX). Samples were stored at ambient temperature until they reached the analytical laboratory, a period of 1–4 hr.

RNA isolation.  Total RNA was extracted on the day of collection, or following storage at 4°C for 1–9 days in RNALater. The hair follicle samples from each individual were removed from the RNALater and placed directly into a 2 ml microcentrifuge tube containing 1 ml of TrizolTM. RNA extraction was then conducted according to manufacturer’s instructions (Invitrogen Life Technologies, CA); a handheld homogenizer (Fisher Scientific, Hampton, NH) was used to disperse the samples. At the isopropanol precipitation step, RNA was pelleted either immediately, or following storage at –80°C for up to 7 mo. After centrifugation, the pellet was washed with 70% ethanol, air dried, and resuspended in 10 µl of RNA free water (Gibco-BRL, Gaithersburg, MD). The concentration of the eluted RNA was quantified using a GeneQuant spectrophotometer (Amersham Pharmacia Biotech, Piscataway, NJ) by absorbance readings at 260 and 280 nm. Each sample was also analyzed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA) according to the manufacturer’s instructions. The 28S:18S rRNA ratio (2100 Expert software, Agilent Technologies) and RNA integrity number (RIN Beta Version Software, Agilent Technologies) were calculated. To derive a RIN number from the unusual total RNA profile obtained from hair follicles (see results), several parameters [slope threshold (0.3), height threshold (0.3), 5S anomaly region (1.0), and ribosomal ratio threshold (1.0)] had to be set at non-default values in the RIN software. The RNA to be analyzed on microarrays was cleansed with an RNeasy Micro kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. The on-column DNase I treatment option was used to eliminate genomic DNA contamination.

RNA amplification, probe labeling, and hybridization.  Because only a small amount of RNA was extracted from the samples, a double RNA amplification step was incorporated prior to microarray hybridization. Depending on sample yield, between 30 and 567 ng of total RNA was used for amplification by Ambion’s MessageAmp aRNA kit, according to the instruction manual. Briefly, first and second strand cDNA were synthesized. Unlabelled aRNA was generated by in vitro transcription with unbiotinylated NTPs. For probe preparation, aRNA was reverse transcribed with second round primers. Second-strand cDNA was synthesized with T7 Oligo(dT) primer and purified. Biotin-labeled cRNA was generated by in vitro synthesis transcription, and purified with a Qiagen RNeasy kit. Each labeled cRNA was then fragmented, added to a hybridization solution, and hybridized for 16 hr to a human genome focus (HG-Focus) genechip (Affymetrix, Santa Clara, CA) in an Affymetrix Fluidics Station 400. The chips were washed, stained with phycoerythrin-conjugated streptavidin, and amplified by biotinylated anti-streptavidin. After a final wash the arrays were scanned in a GCS3000 instrument (Affymetrix).

Microarray data analysis.  Affymetrix Microarray Suite 5.0 was used for image acquisition and analysis. A decision-matrix determines whether each transcript is reliably detected ([ie, present), marginally detected (ie, marginal), or not detected (ie, absent), and calculates signal intensities. The expression data [which were deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus database (GEO, http://www.ncbi.nlm.nih.gov/projects/geo/); GSE3058 [NCBI GEO] ] were transferred to GeneSpring 7.0 software (Agilent Technologies, Redwood City, CA) for statistical analysis. Normalization was carried out to the 50th percentile of each array, and each gene to the median, by choosing the GeneSpring normalization option. Hierarchical cluster analysis was performed using the Condition Tree Clustering option of GeneSpring. The filter function was used to identify genes expressed across all 10 samples.

To evaluate gender differences, a group comparison was made between males (n = 3) and females (n = 7). Genes were selected if they were both: (a) significantly changed in males versus females (t-test, p <0.05); and (b) ≥6-fold up- or down-regulated between any male and any female individual (as determined by pairwise comparison).

Two approaches were used to identify individual genes that had high levels if interindividual variability: (a) using pair-wise comparisons, genes with ≥6-fold change between any 2 individuals were identified, and (b) a CV (coefficient of variation), was calculated for each gene. Genes whose CV was >100% in both the males and females, or >300% in one sex, were characterized as having high interindividual variation.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Acknowledgments and Disclaimers
 References
 
RNA yield and quality.  Table 1Go shows the yield and quality parameters for the RNA extractions. The total yield of RNA from 14 of the 36 hair follicle samples was <20 ng (which in 10 µl of water was below the detection limit of spectrophotometry). e total yield from the other 22 samples ranged from 20 ng to 567 ng, with an average of 252 ng. e average yield per hair follicle from these 22 samples was 112.5 ng (range, 7 to 500 ng).


View this table:
[in this window]
[in a new window]
  		Table 1. Yield and quality control analysis of hair-follicle RNA samples.
 
Agilent Bioanalyzer electopherogram outputs enable the determination of sample quality using 2 approaches. One is to measure the 28S rRNA:18S rRNA ratios. Since the lower detection limit of the RNA Nano LabChip kits was not exceeded in 16 of the samples, these ratios were calculated for only 20 of the 36 RNA samples (Table 1Go). The mean ribosomal ratio for the 20 samples was 0.40. T-test analysis showed no difference in ratio between the samples that were processed immediately and those stored for 1 to 9 days in RNALater solution before processing (data not shown). Similarly, there was no difference in ribosomal ratios between samples processed immediately through to RNA and those stored at –80°C as an isopropanol precipitate for 4 to 7 mo following RNA extraction (see Table 1Go).

A second approach to evaluating RNA quality is to calculate the RNA integrity number (RIN) [21]. However, the bioanalyzer traces typically yielded by the hair follicle RNAs (lower 28S rRNA peak but no degradation peaks in between, see Fig. 1aGo) did not conform to the "normal" total RNA electropherogram trace typically seen in blood (Fig. 1bGo) and most other tissues. This meant that the RIN for most hair follicle RNA samples could not be calculated using the default RIN parameters, since the software did not recognize the output trace. The software default parameters were therefore adjusted (see Materials and Methods) in order to calculate the RINs for the samples selected for hybridization.


Figure 1
View larger version (21K):
[in this window]
[in a new window]
  		Fig 1. Representative Agilent bioanalyzer electropherograms showing hair-follicle total RNA from sample 26F following (a) RNA extraction and purification, and (c) RNA amplification. For comparison, panel (b) shows an electropherogram of whole blood total RNA from the same individual, and is representative of the bioanalyzer trace typically derived from intact total RNA extracted from most tissues.

 
Based on yield and 28S/18S ratios, 3 male and 7 female RNA samples were chosen for microarray hybridization. Table 2Go shows how much RNA was produced from these samples following each of the 2 rounds of preamplification. After 1 round of amplification the average increase in amount of RNA was 11,638%, and after 2 rounds 57,531%. To ensure a successful microarray experiment, RNA quality was further verified by running 1 µl of aRNA obtained from the first round of amplification on the bioanalyzer. The representative electropherogram (Fig. 1cGo) shows an expected size distribution of aRNA, thus confirming that the input RNA was of good quality.


View this table:
[in this window]
[in a new window]
  		Table 2. Effect of preamplification on RNA quantity.
 
Gene expression analysis.  Analysis of the 10 profiled samples revealed that the number of "present" calls ranged from 24–35% (data not shown). Four thousand, two hundred, and ninety-four (4,294) unique Entrez gene IDs were "present" in at least 1 of the 10 samples. On average, 30.2% (SD 4.16) of the probe sets were expressed per sample, equating to 2,720 of 8,500 unique probe sets, or 2,536 of 8,399 unique Entrez gene IDs.

Gender comparisons.  Hierarchical cluster analysis, whether using genes expressed in all subjects (Fig. 2aGo), or genes expressed in at least 1 of the 10 subjects (Fig. 2bGo), showed no clear resolution between males and females. Clustering in both genders was similar, with 2 male samples (3M, 5M) clustered together in a subgroup containing 3 female samples (26F, 30F, 34F), and the third male sample (16M) clustering with the other 4 female samples (24F, 27F, 32F, 33F). In comparing individual male and female samples, gene expression correlation coefficients ranged from 0.705 to 0.823, with an average of 0.769; for comparison within males, the correlation coefficients ranged from 0.724 to 0.944 with an average of 0.821; for the comparison within females, the correlation coefficients ranged from 0.706 to 0.929 with an average of 0.860 (data not shown).


Figure 2
View larger version (57K):
[in this window]
[in a new window]
  		Fig. 2. Hierarchical cluster analysis using (a) genes that wereexpressed in all subjects (1,436 genes), and (b) one that uses the genes (4,294) expressed in at least 1 of the 10 subjects.

 
Two thousand and seventy (2,070) genes were expressed in all male samples, and 1,482 in all female samples. Ninety-seven genes were differentially expressed between males and females (t-test, p <0.05). Eighty-nine of these were upregulated in the males compared to the females, and 8 were upregulated in the females compared to the males. Twelve were identified with levels of expression over 6-fold different between males and females (Table 3Go).


View this table:
[in this window]
[in a new window]
  		Table 3. Genes differentially expressed between genders (group comparison).
 
Interindividual analysis.  One thousand, four hundred, and thirty-six (1,436) genes were expressed in the hair follicles of all 10 subjects. Forty-five pair-wise comparisons were performed to identify genes (of the 1,436 common genes) that had >6-fold difference in expression between any 2 individuals. Twenty-eight such genes were identified (Table 4Go).


View this table:
[in this window]
[in a new window]
  		Table 4. Genes expressed in all individuals that were more than 6-fold different between at least two individuals.
 
To complement the pair-wise comparison strategy to identify genes with large differences in individual expression, CVs were calculated for all 1,436 common genes. The average CV (with normalized raw values) for males and females was 45.31% (SD 11.60) and 60.83% (SD 6.89), respectively. Variation of 26 individual genes was >3 SD from the mean. Of these, 13 (50%, see Table 4Go) overlapped with genes identified in the pair-wise comparison analysis.

The other 13 genes identified in the CV approach included: HIV-1 Tat interactive protein 2, 30kDa (Hs.90753), PTPL1-associated RhoGAP 1 (Hs.430919), propionyl Coenzyme A carboxylase alpha polypeptide (Hs.80741), poly(A) binding protein, cytoplasmic 1 (Hs.387804), metallothionein 2A (Hs.418241), cut-like 1, CCAAT displacement protein (Drosophila) (Hs.438974), cullin 3 (Hs.78946), hexokinase 2 (Hs.406266), kinesin family member 13B (Hs.15711), small acidic protein (Hs.447513), amyloid beta precursor protein binding protein 1, 59kDa (Hs.418162), immediate early response 2 (Hs.737), and defensin, beta 1 (Hs.32949 ).


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Acknowledgments and Disclaimers
 References
 
Many groups have isolated RNA from rodent and human hair follicles [1,2,5,6,22,23]. In most cases this has involved the selective microscopic dissection of anagen phase follicles from skin grafts or biopsies. e problem with such an approach in terms of human clinical or field studies involving RNA transcript analysis is that, in most cases, skin grafts or punch biopsies are not practical specimens. More appropriate for clinical studies are samples that can be taken quickly and easily by clinical staff, and cause little to no discomfort for patients and subjects. Plucked hair follicles fall into this category. However, few groups have reported using plucked hairs as a source of experimental material, and fewer still as a source of material for transcript analysis. Of those that have, Mahe et al [24], Oliveira et al [25,26], and Stark et al [27] used RT-PCR to analyze the expression of a limited number of genes in adult human hair follicles. In all cases the researchers preselected anagen phase follicles from those that were plucked.

At the time of writing there were no PubMed records of gene expression profiling of plucked hair follicles of any species using microarray analysis. is may in part be related to the amount of RNA which can be extracted from a single follicle, or perhaps the assumption that the amount is very little. In fact, Mitsui et al [23] reported extracting an average of 0.90 ± 0.34 µg total RNA from human anagen follicles microscopically dissected from skin samples. Depending on the platform, a single microarray analysis could be carried out with only 3–15 such follicles. However, the collection of skin samples to obtain hair follicles in the way described by Mitsui et al is inappropriate for most clinical studies. The ability to collect a small number of hair follicles by plucking, and then to use these directly for transcript analysis without further triaging, would facilitate the use of these specimens as a source of RNA for clinical, research, and perhaps even forensic studies.

Whereas most hair follicle studies have utilized only anagen follicles dissected from donated scalp skin, plucked hair follicles can consist of both anagen follicles (approximately 85%), and some in the telogen stage (approximately 10–15%). The catagen phase of hair growth is relatively short compared to the other 2 phases, so that only a small percentage of plucked hair follicles will be in this phase (<5%). Considering the nature of the extraction method used in this study and the potential mix of growing, resting, and transitional phase follicles, it is perhaps not surprising that the amount of total RNA extracted from the plucked human hair follicles in this study was not as high as the 0.9 µg per follicle that Matsui et al [23] isolated from dissected anagenic follicles. Brash et al [29] obtained approximately 166–300 ng of total RNA per follicle plucked from adult human volunteers, the follicles being described as "mainly anagen."

The yield obtained in the present study was somewhat variable, with 14 of 36 samples yielding less than the reliable spectrophotometic detection limit (20 ng/µl). However, the electropherogram profile of these samples indicated intact sample, and there have been reports that as little as 2–5 ng of total RNA can be preamplified and used in microarray hybridization [30,31]. Thus, although they were not tested, it may be possible that many or all of these low yield samples could still have been used in microarray hybridization

Of the 22 samples in the current study yielding >20 ng RNA in total, the average yield per hair follicle (112.5 ng) and the range of average yields per follicle (7 to 500 ng) is similar to that recovered by Brash et al [29]. These yields are insufficient to be used directly in microarray-based gene expression profiling. However, following advances in the ability to amplify RNA populations prior to probe labeling [32–34], or to amplify the probe signal following microarray hybridization [35], it is now possible to use very small amounts of RNA in microarray analysis if the RNA is of sufficient quality.

Quality, referring to both purity and integrity, is an important factor in determining whether an RNA sample should be used for microarray analysis. e presence of contaminants, such as heme, has been shown to inhibit the activity of DNA polymerase [36] and reverse transcriptase [37], and significant quantities of contaminating DNA can produce false positive hybridization data. In addition, whereas RT-PCR analysis is robust enough to work with slightly degraded total RNA, microarray analysis generally requires fully intact sample. Elimination of protein and gDNA was facilitated by the use of DNase digestion and purification columns. However, after conducting the isolation and purification process, the 28S:18S ribosomal band ratios and RINs were found in all samples to be much lower than normally expected for intact total RNA (Table 1Go); in all cases this was attributed to a smaller than expected 28S rRNA peak. This feature was particularly noticeable in samples stored at –80°C as an isopropanol precipitate for 4 to 7 mo before pelleting, and in those samples with a low total yield. One possible explanation for this is that salt in the elution buffer and/or that remaining from precipitation masked the higher molecular weight profile of the 28S rRNA peak. is salt-masking effect has been reported to be most pronounced in the lower range of RNA concentrations [38], and may lead to the erroneous conclusion that the RNA sample is degraded. Agilent thus recommends eluting or resuspending samples in deionized water before bioanalyzer analysis.

In this study, sample degradation appeared unlikely since the bioanalyzer electropherograms did not show any of the features that are typically observed in degraded RNA (multiple small peaks between the ribosomal peaks, or at the left of the trace). Furthermore, this same general trace obtained for the human hair follicle RNA samples (exemplified in Fig. 1aGo) was also observed in total RNA extracted from whisker and dorsal hair follicles of rats and mice (this laboratory, data not shown). Given the lack of degradation features, the normal appearance of the preamplified RNA (Fig. 1bGo), and the fact that the same RNA electropherogram features were observed in many samples across 3 species, it appears that total RNA extracted from a small number of hair follicles produces a unique ribosomal RNA electropherogram trace on the bioanalyzer, and that this is either a technical artifact or evidence that total RNA extracted from hair follicles is compositionally different from that typically obtained from most other tissues.

A search of PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed) shows that RNA preamplification is being increasingly accepted and used as a way to generate enough RNA to conduct microarray analysis from small samples such as hair follicles. Total RNA from hair follicles was successfully preamplified in this study prior to microarray hybridization (Fig. 1cGo). A double amplification strategy was used, which in all cases yielded sufficient RNA for 3 to 4 Affymetrix genechip hybridizations (10 µg, see Table 2Go). The data also indicate that a single round of amplification is probably sufficient if at least 180 ng of starting RNA is available. Although RNA amplification is thought sometimes to produce shortened elongation products leading to misclassification of probe sets directed to the middle and 5' region of the transcripts [34], most Affymetrix probes are designed within 600 bp of the 3' end, and therefore any truncation during RNA amplification and labeling would not be expected to introduce significant numbers of false negatives on this platform.

For the 10 samples, an average of 2,536 genes was expressed, representing 30.2% of the total genes interrogated by the genechip. A total of 1,436 genes was expressed in all 10 samples. This number is not surprising, given that scalp hair is common to both males and females and has many physiological similarities in both sexes. Indeed, hierarchical clustering analysis could not discern samples by gender. Group comparison analysis between men and women did identify a number of genes (11) that were expressed at much higher levels (>6-fold) in males than females (Table 3Go). Four of these are Y-linked genes (RPS4Y1, DDX3Y, EIF1AY, and PRKY), although none appear to have a specific documented role in hair follicle growth. Sex-linked genes, and genes expressed in both sexes but at a significantly higher level in one compared to the other, might account in part for some of the physiological differences that exist between male and female hair growth patterns.

Two complementary approaches (pair-wise comparison and CV analysis) were used to identify genes that were expressed in all individuals but had high levels of variability. Many of these genes are involved in hair growth, including the S100 proteins (found in epidermal keratinocytes [39,40]), BNIP 3 (found in the dermal papilla [41]), ID4 (dermal papilla [42]), and KRTHA8 (a keratin) and KRTAP2-4 (a keratin-associated protein), which contribute to the matrix of keratins and intermediate filaments that help form the structure of hair fibers. These variably expressed genes probably account for many of the differences in hair growth and appearance seen among individuals. At the same time, it cannot be assumed that variability of gene expression is always associated with phenotypic variation; genes with lower levels of variability may produce different phenotypes due to functional differences caused by polymorphisms, particularly in cases where the gene product is an enzyme. It has also been shown that although one of the genes differentially expressed in males (DDX3Y) is widely transcribed in a number of tissues, expression of the actual protein is limited to male germ cells by translational control [43].

Some of the variably expressed genes (eg BNIP3) have been associated with apoptosis. Such genes could be derived from catagen phase follicles, and knowledge of which genes contribute to apoptosis in hair follicles could help us understand the mechanisms involved in the hair loss, whether it is caused by age, pregnancy, fever, nutritional deficiency, disease, or exposure to toxic levels of drugs, chemicals, or radiation.

The full records of genes found to be expressed in hair follicles of each of the 10 subjects investigated in this study were deposited with the gene expression omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo). Gene expression data from hair follicles could be useful in a number of ways. Identifying genes common to all individuals can be used to enhance understanding of hair growth and regression by elucidating the gene pathways and networks involved. Comparing samples from different sexes and groups with specific hair characteristics could help us understand what genes contribute to different patterns of hair growth, texture, and color. Comparing samples from cohorts of individuals with different causes of hair loss could enhance understanding of the molecular mechanisms involved in each condition. Such an understanding could provide diagnostic or prognostic biomarkers and provide a basis for developing medical treatments. Hair follicle gene expression might be used as a means to monitor for toxicant exposure at the whole body level or individual target tissues.

As strange as it may seem, human hair follicles could be a good surrogate tissue for monitoring events in reproductive organs. Similar to reproductive organs, the growth and development of hair and hair follicles are tightly regulated by steroid hormones. Consequently, the impact of an endocrine disrupting chemical (EDC) on, for example, reproductive organs might be monitored by studying gene expression profiles in hair follicles. In this respect, it is interesting to note that at least 8 of the genes that were expressed in hair follicles (EBP, DHCR7, SQLE, FDFT1, FDPS, GGPS1, HMGCR, IDI1) have roles in sterol biosynthesis and 3 (HSD17B4, HSD17B12, SRD5A1) have roles in androgen and estrogen metabolism.

In summary, there have been many studies to characterize the global gene expression profiles of "normal" human tissues [44,45]. To the best of our knowledge, this is the first study that examines unstaged plucked hair follicles using a pre-amplification-based strategy coupled with microarray hybridization. Gene expression profiling has many potential uses for any given tissue. However, in order for it to become a routinely used tool, sample procurement, storage, transportation, and analysis must be relatively simple and robust. This study indicates that plucked, unstaged scalp hair follicles can be collected and stored for at least several days using a technically simple procedure. e RNA extracted from such follicles appears to have a unique ribosomal profile, but is still of sufficient quality for use in microarray studies. Total yields are variable, but in most cases even a single hair follicle should yield sufficient RNA to conduct microarray-based expression profiling if a preamplification step is incorporated.

Further studies are needed to verify and improve the methods described herein, and to explore fully the potential uses of hair follicle gene expression profiles in surrogate tissue analysis, diagnostic applications, and characterizing the molecular mechanisms that regulate hair growth and loss.


   Acknowledgments and Disclaimers
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Acknowledgments and Disclaimers
References
 
The authors thank Maryann Bassett, Debbie Levine, and Tracey Mantilla (US EPA) for clinical assistance, and Drs. Christopher Corton, Sid Hunter, and Sally Perreault (US EPA) for reviewing this manuscript. SJK was supported by the U.S. EPA/UNC Toxicology Research Program, Training Agreement T829472, with the Curriculum in Toxicology, University of North Carolina at Chapel Hill. RNM was supported by U.S. EPA Training Agreement CT826512010 with North Carolina State University.

Information in this document has been funded wholly by the U.S. Environmental Protection Agency. It has been reviewed by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does the mention of trade names or commercial products constitute endorsement or recommendation for use. None of the authors reported conflicts of interest.


   Footnotes
 
* Current addresses: Pharmaceutical Research Institute, Bristol-Myers Squibb, One Squibb Drive, New Brunswick, NJ 08903 Back

{dagger} Department of Pharmacology and Toxicology, Brody School of Medicine, East Carolina University, 600 Moye Boulevard, Greenville, NC 27834 Back

# Rosetta Inpharmatics LLC (a subsidiary of Merck & Co.), 401 Terry Ave. N., Seattle, WA 98109 Back


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Acknowledgments and Disclaimers
 References
 

  1. Schlake T, Boehm T. Expression domains in the skin of genes affected by the nude mutation and identified by gene expression profiling. Mech Dev 2001;109:419–422.[Medline]
  2. Lin KK, Chudova D, Hatfield GW, Smyth P, Andersen B. Identification of hair cycle-associated genes from time-course gene expression profile data by using replicate variance. PNAS 2004;101:15955–15960.[Abstract/Free Full Text]
  3. O’Shaughnessy RF, Christiano AM, Jahoda CA. The role of BMP signaling in the control of ID3 expression in the hair follicle. Exp Dermatol 2004;13:621–629.[Medline]
  4. O’Shaughnessy RF, Yeo W, Gautier J, Jahoda CA, Christiano AM. The WNT signaling modulator, Wise, is expressed in an interaction-dependent manner during hair-follicle cycling. J Invest Dermatol 2004;123:613–621.[Medline]
  5. Schlake T, Beibel M, Weger N, Boehm T. Major shifts in genomic activity accompany progression through different stages of the hair cycle. Gene Expr Patterns 2004;4:141–152.[Medline]
  6. Xu X, Lyle S, Liu Y, Solky B, Cotsarelis G. Differential expression of cyclin D1 in the human hair follicle. Am J Pathol 2003;163:969–978.[Abstract/Free Full Text]
  7. So PL, Epstein EH Jr. Adult stem cells: capturing youth from a bulge? Trends Biotechnol 2004;22:493–496.[Medline]
  8. Burczynski ME, McMillian M, Ciervo J, Li L, Parker JB, Dunn RT 2nd, Hicken S, Farr S, Johnson MD. Toxicogenomics-based discrimination of toxic mechanism in HepG2 human hepatoma cells. Toxicol Sci 2000; 58:399–415.[Abstract/Free Full Text]
  9. Bartosiewicz M, Penn S, Buckpitt A. Applications of gene arrays in environmental toxicology: fingerprints of gene regulation associated with cadmium chloride, benzo(a)pyrene, and trichloroethylene. Environ Health Perspect 2001;109:71–74.[Medline]
  10. Thomas RS, Rank DR, Penn SG, Zastrow GM, Hayes KR, Pande K, Glover E, Silander T, Craven MW, Reddy JK, Jovanovich SB, Bradfield CA. Identification of toxicologically predictive gene sets using cDNA microarrays. Mol Pharmacol 2001;60:1189–1194.[Abstract/Free Full Text]
  11. Hamadeh HK, Bushel PR, Jayadev S, DiSorbo O, Bennett L, Li L, Tennant R, Stoll R, Barrett JC, Paules RS, Blanchard K, Afshari CA. Prediction of compound signature using high density gene expression profiling. Toxicol Sci 2002;67:232–240.[Abstract/Free Full Text]
  12. Hamadeh HK, Bushel PR, Jayadev S, Martin K, DiSorbo O, Sieber S, Bennett L, Tennant R, Stoll R, Barrett JC, Blanchard K, Paules RS, Afshari CA. Gene expression analysis reveals chemical-specific profiles. Toxicol Sci 2002;67:219–231.[Abstract/Free Full Text]
  13. Kier LD, Neft R, Tang L, Suizu R, Cook T, Onsurez K, Tiegler K, Sakai Y, Ortiz M, Nolan T, Sankar U, Li AP. Applications of microarrays with toxicologically relevant genes (tox genes) for the evaluation of chemical toxicants in Sprague Dawley rats in vivo and human hepatocytes in vitro. Mutat Res 2004;549:101–113.[Medline]
  14. Fielden M, Natsoulis G, Kolaja K. A molecular basis for the prediction of renal tubular injury by drug signatures following short-term compound treatment. Toxicologist 2005;84:106 (abstract #521).
  15. Heinloth AN, Irwin RD, Boorman GA, Nettesheim P, Fannin RD, Sieber SO, Snell ML, Tucker CJ, Li L, Travlos GS, Vansant G, Blackshear PE, Tennant RW, Cunningham ML, Paules RS. Gene expression profiling of rat livers reveals indicators of potential adverse effects. Toxicol Sci 2004;80:193–202.[Abstract/Free Full Text]
  16. Amundson SA, Grace MB, McLeland CB, Epperly MW, Yeager A, Zhan Q, Greenberger JS, Fornace AJ Jr. Human in vivo radiation-induced biomarkers: gene expression changes in radiotherapy patients. Cancer Res 2004;64:6368–6371.[Abstract/Free Full Text]
  17. Rockett JC, Burczynski ME, Fornace AJ, Herrmann PC, Krawetz SA, Dix DJ. Surrogate tissue analysis: monitoring toxicant exposure and health status of inaccessible tissues through the analysis of accessible tissues and cells. Toxicol Appl Pharmacol 2004;194:189–199.[Medline]
  18. Tang Y, Lu A, Aronow BJ, Sharp FR. Blood genomic responses differ after stroke, seizures, hypoglycemia, and hypoxia: blood genomic fingerprints of disease. Ann Neurol 2001;50:699–707.[Medline]
  19. Rockett JC, Kavlock RJ, Lambright CR, Parks LG, Schmid JE, Wilson VS, Wood C, Dix DJ. DNA arrays to monitor gene expression in rat blood and uterus following 17beta-estradiol exposure: biomonitoring environmental effects using surrogate tissues. Toxicol Sci 2002;69:49–59.[Medline]
  20. Twine NC, Stover JA, Marshall B, Dukart G, Hidalgo M, Stadler W, Logan T, Dutcher J, Hudes G, Dorner AJ, Slonim DK, Trepicchio WL, Burczynski ME. Disease-associated expression profiles in peripheral blood mono-nuclear cells from patients with advanced renal cell carcinoma. Cancer Res 2003;63:6069–6075.[Abstract/Free Full Text]
  21. Mueller O, Lightfoot S, Schroeder A. RNA integrity number (RIN) standardization of RNA quality. Agilent Biotechnologies Application Notes, available at: http://geacf.cwru.edu/geacf/RNA_quality/RIN%20note.pdf [accessed 13 April 2004].
  22. Little JC, Westgate GE, Evans A, Granger SP. Cytokine gene expression in intact anagen rat hair follicles. J Invest Dermatol 1994;103:715–720.[Medline]
  23. Mitsui S, Ohuchi A, Hotta M, Tsuboi R, Ogawa H. Genes for a range of growth factors and cyclin-dependent kinase inhibitors are expressed by isolated human hair follicles. Br J Dermatol 1997;137:693–698.[Medline]
  24. Mahe YF, Buan B, Billoni N, Loussouarn G, Michelet JF, Gautier B, Bernard BA. Pro-inflammatory cytokine cascade in human plucked hair. Skin Pharmacol 1996; 9:366–375.[Medline]
  25. Oliveira IO, Lhullier C, Brum IS, Spritzer PM. The 5-alpha-reductase type 1, but not type 2, gene is expressed in anagen hairs plucked from the vertex area of the scalp of hirsute women and normal individuals. Braz J Med Biol Res 2003;36:1447–1454.[Medline]
  26. Oliveira IO, L’hullier C, Brum IS, Spritzer PM. Gene expression of type 2 17-beta-hydroxysteroid dehydrogenase in scalp hairs of hirsute women. Steroids 2003;68:641–649.[Medline]
  27. Stark K, Torma H, Cristea M, Oliw EH. Expression of CYP4F8 (prostaglandin H 19-hydroxylase) in human epithelia and prominent induction in epidermis of psoriatic lesions. Arch Biochem Biophys 2003;409:188–196.[Medline]
  28. Sperling LC. Hair anatomy for the clinician. J Am Acad Dermatol 1991;25:1–17.[Medline]
  29. Brash AR, Boeglin WE, Chang MS. Discovery of a second 15S-lipoxygenase in humans. PNAS 1997;94: 6148–6152.[Abstract/Free Full Text]
  30. Dafforn A, Chen P, Deng G, Herrler M, Iglehart D, Koritala S, Lato S, Pillarisetty S, Purohit R, Wang M, Wang S, Kurn N. Linear mRNA amplification from as little as 5 ng total RNA for global gene expression analysis. Biotechniques 2004;37:854–857.[Medline]
  31. Patel OV, Suchyta SP, Sipkovsky SS, Yao J, Ireland JJ, Coussens PM, Smith GW. Validation and application of a high fidelity mRNA linear amplification procedure for profiling gene expression. Vet Immunol Immunopathol 2005;105:331–342.[Medline]
  32. Nallur G, Luo C, Fang L, Cooley S, Dave V, Lambert J, Kukanskis K, Kingsmore S, Lasken R, Schweitzer B. Signal amplification by rolling circle amplification on DNA microarrays. Nucleic Acids Res 2001;29:E118.[Medline]
  33. Che S, Ginsberg SD. Amplification of RNA transcripts using terminal continuation. Lab Invest 2004;84:131–137.[Medline]
  34. Dumur CI, Garrett CT, Archer KJ, Nasim S, Wilkinson DS, Ferreira-Gonzalez. A. Evaluation of a linear amplification method for small samples used on high-density oligonucleotide microarray analysis. Anal Biochem 2004;331:314–321.[Medline]
  35. Karsten SL, Van Deerlin VM, Sabatti C, Gill LH, Geschwind DH. An evaluation of tyramide signal amplification and archived fixed and frozen tissue in microarray gene expression analysis. Nucleic Acids Res 2002;30:E4.[Medline]
  36. Akane A, Matsubara K, Nakamura H, Takahashi S, Kimura K. Identification of the heme compound copurified with deoxyribonucleic acid (DNA) from bloodstains, a major inhibitor of polymerase chain reaction (PCR) amplification. J Forensic Sci 1994;39:362–372.[Medline]
  37. Staudinger R, Abraham NG, Levere RD, Kappas A. Inhibition of human immunodeficiency virus-1 reverse transcriptase by heme and synthetic heme analogs. Proc Assoc Am Physicians 1996;108:47–54.[Medline]
  38. Agilent technical report: Successful analysis of low RNA concentrations with the Agilent 2100 bioanalyzer and the RNA 6000 PicoLabChip® kit. http://www.chem.agilent.com/scripts/literature.pdf.asp?iWHID=36753 [accessed 29 March 2005].
  39. Schmidt M, Gillitzer R, Toksoy A, Brocker EB, Rapp UR, Paus R, Roth J, Ludwig S, Goebeler M. Selective expression of calcium-binding proteins S100a8 and S100a9 at distinct sites of hair follicles. J Invest Dermatol 2001;17:748–750.
  40. Eckert RL, Broome AM, Ruse M, Robinson N, Ryan D, Lee K. S100 proteins in the epidermis. J Invest Dermatol 2004;123:23–33.[Medline]
  41. Farooq M, Kim Y, Im S, Chung E, Hwang S, Sohn M, Kim M, Kim J. Cloning of BNIP3h, a member of proapoptotic BNIP3 family genes. Exp Mol Med 2001; 33:169–173.[Medline]
  42. Whitehouse CJ, Huckle JW, Demarchez M, Reynolds AJ, Jahoda CA. Genes that are differentially expressed in rat vibrissa follicle germinative epithelium in vivo show altered expression patterns after extended organ culture. Exp Dermatol 2002;11:542–555.[Medline]
  43. Ditton HJ, Zimmer J, Kamp C, Rajpert-De Meyts E, Vogt PH. The AZFa gene DBY (DDX3Y) is widely transcribed but the protein is limited to the male germ cells by translation control. Hum Mol Genet 2004;13: 2333–2341.[Abstract/Free Full Text]
  44. Hsiao LL, Dangond F, Yoshida T, Hong R, Jensen RV, Misra J, Dillon W, Lee KF, Clark KE, Haverty P, Weng Z, Mutter GL, Frosch MP, Macdonald ME, Milford EL, Crum CP, Bueno R, Pratt RE, Mahadevappa M, Warrington JA, Stephanopoulos G, Gullans SR. A compendium of gene expression in normal human tissues. Physiol Genomics 2001;7:97–104.[Abstract/Free Full Text]
  45. Yanai I, Benjamin H, Shmoish M, Chalifa-Caspi V, Shklar M, Ophir R, Bar-Even A, Horn-Saban S, Safran M, Domany E, Lancet D, Shmueli O. Genome-wide midrange transcription profiles reveal expression level relationships in human tissue specification. Bioinformatics 2005;21:650–659.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kim, S. J.
Right arrow Articles by Rockett, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kim, S. J.
Right arrow Articles by Rockett, J. C.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS