Portal de Periódicos da CAPES

Use of RT-PCR and DNA Microarrays to Characterize RNA Recovered by Non-Invasive Tape Harvesting of Normal and Inflamed Skin

2004; Elsevier BV; Volume: 123; Issue: 1 Linguagem: Inglês

10.1111/j.0022-202x.2004.22729.x

1523-1747

Rita Wong, Vynga Tran, Vera B. Morhenn, She‐pin Hung, Bogi Andersen, Elaine Ito, G. Wesley Hatfield, Nicholas R. Benson,

RNA Research and Splicing

We describe a non-invasive approach for recovering RNA from the surface of skin via a simple tape stripping procedure that permits a direct quantitative and qualitative assessment of pathologic and physiologic biomarkers. Using semi-quantitative RT-PCR we show that tape-harvested RNA is comparable in quality and utility to RNA recovered by biopsy. It is likely that tape-harvested RNA is derived from epidermal cells residing close to the surface and includes adnexal structures and present data showing that tape and biopsy likely recover different cell populations. We report the successful amplification of tape-harvested RNA for hybridization to DNA microarrays. These experiments showed no significant gene expression level differences between replicate sites on a subject and minimal differences between a male and female subject. We also compared the array generated RNA profiles between normal and 24 h 1% SLS-occluded skin and observed that SLS treatment resulted in statistically significant changes in the expression levels of more than 1,700 genes. These data establish the utility of tape harvesting as a non-invasive method for capturing RNA from human skin and support the hypothesis that tape harvesting is an efficient method for sampling the epidermis and identifying select differentially regulated epidermal biomarkers. We describe a non-invasive approach for recovering RNA from the surface of skin via a simple tape stripping procedure that permits a direct quantitative and qualitative assessment of pathologic and physiologic biomarkers. Using semi-quantitative RT-PCR we show that tape-harvested RNA is comparable in quality and utility to RNA recovered by biopsy. It is likely that tape-harvested RNA is derived from epidermal cells residing close to the surface and includes adnexal structures and present data showing that tape and biopsy likely recover different cell populations. We report the successful amplification of tape-harvested RNA for hybridization to DNA microarrays. These experiments showed no significant gene expression level differences between replicate sites on a subject and minimal differences between a male and female subject. We also compared the array generated RNA profiles between normal and 24 h 1% SLS-occluded skin and observed that SLS treatment resulted in statistically significant changes in the expression levels of more than 1,700 genes. These data establish the utility of tape harvesting as a non-invasive method for capturing RNA from human skin and support the hypothesis that tape harvesting is an efficient method for sampling the epidermis and identifying select differentially regulated epidermal biomarkers. posterior probability of differential expression sodium lauryl sulfate Contact dermatitis, a common skin reaction, involves several signaling pathways. Irritant contact dermatitis (ICD) predominantly involves keratinocyte activation (Freedberg et al., 2001Freedberg I.M. Tomic-Canic M. Komine M. et al.Keratins and the keratinocyte activation cycle.J Invest Dermatol. 2001; 116: 633-640Crossref PubMed Scopus (419) Google Scholar), whereas Langerhans' cell presentation of antigen to T cells in draining lymph nodes and recognition of the offending allergen in skin by memory T cells control the initiation and expression of allergic contact dermatitis (ACD;Feghali and Wright, 1997Feghali C.A. Wright T.M. Cytokines in acute and chronic inflammation.Front Biosci. 1997; 2: 12-d26Crossref PubMed Google Scholar) Clinically, both contact dermatitides are characterized by pruritus, erythema, and edema. This commonality of the clinical signs and symptoms makes distinguishing between ICD and ACD difficult at the clinical level, particularly when symptoms are subtle. By contrast, at the molecular level, ICD and ACD are believed to be characterized by unique mRNA patterns, although the published literature is conflicting (Hoefakker et al., 1995Hoefakker S. Caubo M. van't Erve E.H. et al.In vivo cytokine profiles in allergic and irritant contact dermatitis.Contact Dermatitis. 1995; 33: 258-266Crossref PubMed Scopus (81) Google Scholar;Flier et al., 1999Flier J. Boorsma D.M. Bruynzeel D.P. et al.The CXCR3 activating chemokines IP-10, Mig, and IP-9 are expressed in allergic but not in irritant patch test reactions.J Invest Dermatol. 1999; 113: 574-578Crossref PubMed Scopus (114) Google Scholar;Morhenn et al., 1999Morhenn V.B. Chang E.Y. Rheins L.A. A noninvasive method for quantifying and distinguishing inflammatory skin reactions.J Am Acad Dermatol. 1999; 41: 687-692Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar;Ryan and Gerberick, 1999Ryan C.A. Gerberick G.F. Cytokine mRNA expression in human epidermis after patch treatment with rhus and sodium lauryl sulfate.Am J Contact Dermat. 1999; 10: 127-135PubMed Google Scholar;Ulfgren et al., 2000Ulfgren A.K. Klareskog L. Lindberg M. et al.An immunohistochemical analysis of cytokine expression in allergic and irritant contact dermatitis.Acta Derm Venereol. 2000; 80: 167-170Crossref PubMed Scopus (45) Google Scholar;Cumberbatch et al., 2002Cumberbatch M. Dearman R.J. Groves R.W. et al.Differential regulation of epidermal Langerhans cell migration by interleukins (IL)-1alpha and IL-1beta during irritant- and allergen-induced cutaneous immune responses.Toxicol Appl Pharmacol. 2002; 182: 126-135Crossref PubMed Scopus (72) Google Scholar). Documentation of simple and complex mRNA profiles is possible using RT-PCR and DNA microarray technologies. Using the technique of tape stripping, RNA can be harvested from both normal and inflamed skin (Morhenn et al., 1999Morhenn V.B. Chang E.Y. Rheins L.A. A noninvasive method for quantifying and distinguishing inflammatory skin reactions.J Am Acad Dermatol. 1999; 41: 687-692Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar) and by combining tape stripping and RNA profiling, it may be possible to non-invasively establish a diagnosis of ICD or ACD. In this study, we expand upon the work ofMorhenn et al., 1999Morhenn V.B. Chang E.Y. Rheins L.A. A noninvasive method for quantifying and distinguishing inflammatory skin reactions.J Am Acad Dermatol. 1999; 41: 687-692Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar who used adhesive tape as a substitute for punch biopsy to non-invasively sample the epidermis. Morhenn et al demonstrated that when a skin site was tape stripped 20 or more times, it was possible to recover sufficient RNA from skin cells adherent to the tapes to detect and semi-quantify specific mRNAs using the ribonuclease protection assay. We demonstrate in this work that sufficient RNA can be recovered using sequential application of as few as four small tapes. In order to document the use of tape harvesting as an accurate and reliable sampling method we performed a clinical trial in which occlusive patches containing either 1% SLS (irritant) or water (vehicle control) were applied to the mid-back of 10 subjects for 24 h. The sites were then clinically assessed and, along with normal control skin, surface cells were harvested with four applications of individual tapes and by shave biopsy. RNA was extracted from the tapes and biopsies and assayed semi-quantitatively for Il-1β, IL-8, GAPDH, and β-actin mRNA using fluorescent, quantitative RT-PCR. The results showed consistent increases in IL-1β and IL-8 mRNA in inflamed skin relative to untreated skin. We further report the successful use of tape-harvested RNA to profile normal and experimentally inflamed skin using DNA microarrays. This profile of SLS-irritated skin is the first step in the definition of RNA profiles designed to differentiate irritant from allergic skin reactions. RNA was recovered from 27 of 30 skin sites using four tapes as described in Materials and Methods. The amount of total RNA recovered was variable from site to site and subject to subject (data not shown). The average mass of RNA recovered from normal skin sites was 0.92 ng (±0.35) with a range of 0 (two samples)–3.2 ng. The average mass of RNA recovered from water-occluded skin was 0.69 ng (±0.27) with a range of 0 (one sample)–2.7 ng. SLS inflamed skin produced the greatest average yield of RNA with an average of 185 ng (±76) and a range of 0.067–747 ng. We have chosen as markers of the inflammatory process IL-1β and IL-8 mRNAs. We have accounted for differential recovery of total RNA mass in a sample by normalizing these mRNAs to an internal control, the β-actin transcript. We then calibrated this normalized RNA ratio to the analogous ratio in a control sample (normal skin) using the comparative or ΔΔCt method (described in Materials and Methods). In this study, IL-1β and IL-8 mRNA are predicted to increase relative to β-actin in response to SLS treatment, whereas the level of housekeeping mRNAs, such as β-actin and GAPDH are assumed to remain constant. We have tested the assumption that housekeeping mRNAs such as GAPDH and β-actin are unchanging among different samples by measuring the relative ratio of these two mRNAs. If the relative ratio of two housekeeping genes is unchanged in different samples, the relative ratio of GAPDH/β-actin mRNA between two samples should be equal to 1. The data in Table I reveal the fold-change of the GAPDH/β-actin mRNA ratio in SLS- and water-treated samples relative to normal skin and also water-treated skin. The tape sample data show that the average fold-change in SLS samples was 0.55, whereas the average fold-change in water-treated samples was 1.14. Biopsy samples showed similar and minor changes with SLS-treated samples having an average 0.39 fold-change, whereas water-treated samples have an average 0.53 fold-change. Individual subject data are shown in Online Table S1. Although there are some examples of statistically significant changes relative to the normal skin, these changes are not sufficient to explain the fold-changes in IL-1β and IL-8, which are reported in Table I. Thus, it is likely that our housekeeping genes do change relative levels in response to SLS and water treatment, but the magnitude of those changes is minor. Similar observations and conclusions have been reported byPaludan and Thestrup-Pedersen, 1992Paludan K. Thestrup-Pedersen K. Use of the polymerase chain reaction in quantification of interleukin 8 mRNA in minute epidermal samples.J Invest Dermatol. 1992; 99: 830-835Abstract Full Text PDF PubMed Google Scholar andGrangsjo et al., 1996Grangsjo A. Leijon-Kuligowski A. Torma H. et al.Different pathways in irritant contact eczema? Early differences in the epidermal elemental content and expression of cytokines after application of 2 different irritants.Contact Dermatitis. 1996; 35: 355-360Crossref PubMed Scopus (46) Google Scholar.Table ISummary of fold-change in GAPDH, IL-1β, and IL-8 mRNA relative to β-actin mRNA in SLS-treated and water-treated skin calibrated to control skinAverage fold-change by sampling method and treatmentaAverage fold-change in up to 10 subjects (n=8 for normal skin tape samples, n=9 for water-treated skin tape samples and n=10 for all others). The average fold-change is calculated from the subject-average ΔCt values given in Online Tables S3, S4, S7. Individual fold-change values are reported in Online Tables S1, S2, S5. Values preceded by > are lower limit estimates necessitated by the fact that the applicable mRNA was not detectable in the control (calibration) samples. The >95% confidence interval for fold-change is given in parenthesis.TapeBiopsymRNACalibratorbSLS and water samples are calibrated to normal skin samples; additionally, SLS samples are calibrated to the water sample. Calibration is described in Materials and Methods.WaterSLSWaterSLSGAPDHNormal1.14 (0.6–2.14)0.55 (0.31–0.99)0.53 (0.43–0.66)0.39 (0.29–0.53)Water10.48 (0.36–0.65)10.74 (0.55–1)IL-1βNormalND>2.66>2.49>11Water1>1.514.42 (1.42–13.76)IL-8Normal1.94 (0.63–5.99)10.32 (3.89–27.34)>1.54>52.45Water15.31 (2–14.1)134.06 (6.37–181.87)a Average fold-change in up to 10 subjects (n=8 for normal skin tape samples, n=9 for water-treated skin tape samples and n=10 for all others). The average fold-change is calculated from the subject-average ΔCt values given in Online Table S3, Table S4, Table S7. Individual fold-change values are reported in Online Table S1, Table S2, Table S5. Values preceded by > are lower limit estimates necessitated by the fact that the applicable mRNA was not detectable in the control (calibration) samples. The >95% confidence interval for fold-change is given in parenthesis.b SLS and water samples are calibrated to normal skin samples; additionally, SLS samples are calibrated to the water sample. Calibration is described in Materials and Methods. Open table in a new tab Download .jpg (.08 MB) Help with files Table S1Fold-change of GAPDH/β-actin mRNA ratio in tape and biopsy samples of SLS-treated and water-treated skin calibrated to normal skin Data in Table I reveal the average fold-change of the IL-1β/β-actin mRNA ratio in water-occluded and SLS-occluded skin relative to normal skin, in tape- and biopsy-harvested RNA samples. Individual subject data can be found in Online Table S2. Table I shows that on average the IL-1β/β-actin mRNA ratio in biopsy samples of SLS-treated skin was elevated at least 11-fold (>95% confidence interval) compared with normal skin. Furthermore, the IL-1β/actin ratio was 4.42-fold elevated in SLS- relative to water-treated samples. In six of 10 subjects water occlusion produced significant (>95% confidence) increases in the IL-1β/actin ratio but this increase was typically 2–4-fold and was always smaller than the respective effect of SLS occlusion (Online Table S3). Thus SLS-occlusion produced the most consistent elevation of the IL-1β/actin mRNA ratio but water-occlusion did effect a similar albeit smaller response. Download .jpg (.07 MB) Help with files Table S2Summary of fold-changes in IL-1β/β-actin mRNA ratio in SLS-treated and water-treated skin relative to normal skin Download .jpg (.07 MB) Help with files Table S3ΔCt values for IL-1β mRNA in normal, water-treated and SLS-treated skin samples recovered by tape and biopsy Data in Table I also reveal the fold-change of IL-1β/β-actin mRNA in tape-harvested samples of water-occluded and SLS-occluded skin relative to water-occluded and normal skin. Individual subject data are shown in Online Table S2, Table S3. In most subjects, IL-1β was undetectable in normal and water-treated, tape-harvested skin samples. The average fold-change of the IL-1β/actin ratio was estimated to be at least 2.66-fold in SLS-treated samples. Although the average fold-change for eight subjects was modest, five of these subjects showed significant (>95% confidence) IL-1β/actin ratio increases that were in qualitative agreement with the biopsy data (Online Table S3). Analysis of the remaining five samples was indeterminate because IL-1β mRNA was not detected in either the water or normal skin sample. Table I reveals the fold-change of the IL-8/β-actin mRNA ratio in water-occluded and SLS-occluded skin relative to normal skin. The data demonstrate that the IL-8/actin ratio was on average 52-fold elevated in SLS-treated skin samples relative to normal skin and 34-fold elevated relative to water-treated skin. Biopsy samples from nine of 10 untreated skin sites had undetectable IL-8 mRNA levels and the sole normal skin biopsy sample with detectable IL-8 mRNA was close to the level of detection (Online Table S4, Table S5). Thus IL-8 mRNA was generally not detectable in a biopsy of normal skin. Download .jpg (.08 MB) Help with files Table S4ΔCt values for IL-8 mRNA in normal, water-occluded and SLS-occluded skin Download .jpg (.07 MB) Help with files Table S5Summary of fold-changes in IL-8/β-actin mRNA ratio in SLS-treated and water-treated skin relative to normal skin Table I reveals that the IL-8/actin mRNA ratio was on average 10-fold increased in tape-harvested samples from SLS-occluded sites (>95% confidence interval) and 1.94-fold increased in water-treated samples (not significantly different than normal skin). We conclude that the tape data are in good qualitative agreement with the biopsy data, with a majority of inflamed sites revealing increases in IL-8 mRNA. During our analysis of IL-8 mRNA in biopsy and tape samples, we observed that although the fold-change was in qualitative agreement between the two methods, the primary ΔCt data were strikingly different (Online Table S4). As described in Materials and Methods, ΔCt is a measure of the ratio of two mRNAs in a sample. Online Table S4 shows that the ΔCt, IL-8 for tape samples is very different than that of biopsy samples—for both SLS- and water-treated skin sites—and this difference is highly significant (both p-values <0.005). The most striking example of this is revealed by a comparison of the average ΔCt, IL-8 in tape-harvested samples compared with biopsy-harvested samples of water-occluded sites. For tape-harvested samples the average (ΔCt, IL-8)water is 1.54, whereas for biopsy samples it is 9.22 (Online Table S4). The data suggest that in the cell population harvested by tape, IL-8 mRNA is much more abundant relative to β-actin mRNA than in biopsy samples. The average relative abundance can be estimated as equal to 2-(1.54–9.22) or 205. In this calculation, the biopsy sample acts as the “calibrator”, thus the tape samples have, on average, a 205-fold greater IL-8/actin mRNA ratio than biopsies. A similar calculation can be performed with the SLS sample data (Online Table S4), which also reveals a highly significant difference between tape and biopsy (ΔCt, IL-8)SLS values (p<0.005). We conclude from this data that tape harvesting recovers a distinctly different cell population than does biopsy. The success in the above analysis of tape-harvested RNA from different skin sites suggested that this RNA might be amenable to amplification and hybridization to DNA microarrays. In order to assess the reproducibility and consistency of tape-harvested RNA samples for gene expression profiling experiments, we collected three samples from the upper back of each of two healthy individuals, one male (sample C1, C2, and C3) and one female (sample A5, A6, and A9). Approximately 1 ng of total RNA was isolated and the mRNA was amplified and biotin labeled using a MessageAmp aRNA kit (Ambion Inc., Austin, Texas) as described in Materials and Methods. The resulting biotin-labeled aRNA from each sample was used for hybridization to an Affymetrix HG-U133A GeneChip. The results in Table II show the differences observed when a matrix of pairwise gene expression comparisons between two GeneChips was performed using Affymetrix Microarray Suite software. These data show an average of only 12% variance among gene measurements, regardless of whether data from different sites on the same individual or sites from different individuals are compared. Furthermore, comparing the data in quadrant three of Table II (A versus C) to the data in quadrants one (A versus A) and four (C versus C) shows that about 15% of this variance is due to either gender difference (A versus C) or inter-subject variation (A versus A or C versus C). Thus, amazingly little variance is contributed by samples obtained from different sites or from different individuals.Table IIPercentage of the measurement of gene expression reported as unchanged by the Affymetrix MAS 5.0 software for each of all possible pairwise comparisons among GeneChips (A5, A6, A9, C1, C2, and C3) hybridized with aRNA obtained from three different locations on the upper back of two subjects (A and C)GeneChip/subject ID (%)A5A6A9C1C2C3A5100A688.90100A990.8086.10100C189.8088.2087.40100C285.0085.3083.1089.60100C388.0088.0087.3088.9083.70100 Open table in a new tab To compare these data in a more quantitative manner, the three Affymetrix GeneChips each hybridized with targets from RNA samples obtained from individual A were compared with three GeneChips hybridized with targets from the three RNA samples obtained from individual C. These data were analyzed with a regularized t test (Long et al., 2001Long A.D. Mangalam H.J. Chan B.Y. et al.Improved statistical inference from DNA microarray data using analysis of variance and a Bayesian Global gene expression profiling in Escherichia coli K12. The effects of integration host factor.J Biol Chem. 2001; 276: 19937-19944Crossref PubMed Scopus (307) Google Scholar) implemented in the Cyber-T statistical program. This three-by-three comparison revealed 21,790 probe sets that exhibited gene expression levels above background for all three sites from each subject. Of these genes 1117 (5%) were differentially expressed with p-value less than 0.0035, which based on the global false positive and negative levels of this dataset corresponds to a PPDE value of 0.95. Thus, 56 of the 1117 differentially expressed genes that exceed this p-value threshold are expected to be false positives. The source of these inter-subject gene expression differences remains to be determined; however, at least one of these differences is gender based. For example, the gene with the smallest p-value and the highest PPDE value is the Y-linked ribosomal protein S4 (PRS4Y). It is likely that differences that are not gender based are a reflection of normal variation of gene expression between individuals. These data are available at http://www.igb.uci.edu. In a separate experiment, a total of nine RNA samples ranging from 1 to 10 ng were isolated by tape harvesting from three untreated, three water-occluded, and three SLS-occluded sites of each of three individuals. mRNA from each of the nine samples was amplified, biotin labeled and used for hybridization to each of nine Affymetrix HG-U133A GeneChips as shown in Figure 1. A comparison of gene expression levels between three untreated (A1, B1, C1) samples and three SLS-treated (A2, B2, C2) samples revealed 21,031 genes that exhibited expression levels above background for all samples. To assess the confidence in global changes in gene expression, methods implemented in Cyber-T were used to determine the posterior PPDE of each gene based on experiment-wide global false positive and negative gene measurement levels as described byHung et al., 2002Hung S.-P. Baldi P. Hatfield G.W. et al.Global gene expression profiling in Escherichia coli K12: The effects of leucine-responsive regulatory protein.J Biol Chem. 2002; 277: 40309-40323Crossref PubMed Scopus (115) Google Scholar andBaldi and Hatfield, 2002Baldi P. Hatfield G.W. DNA Microarrays and Gene Expression: From Experiments to Data Analysis and Modeling. Cambridge University Press, Cambridge, UK2002Crossref Google Scholar. When untreated versus SLS-occluded data are compared, the p-values for the differentially expressed genes are low and cluster toward 0. This is consistent with highly statistically significant differences among measurement levels of some genes. In fact 1771 genes are differentially expressed with a threshold of p=0.003, which corresponds to an experiment-wide probability for differential expression (PPDE value) equal to or greater than 0.99. These data are available at http://www.igb.uci.edu. A comparison of gene expression levels between three SLS-treated (A2, B2, C2) samples and three water-treated (A3, B3, C3) samples revealed 21,307 genes that exhibited expression levels above background for all samples. Based on a threshold of p=0.003, 1364 genes are differentially expressed with a PPDE value of 0.99. Of these, 1063 genes are also differentially expressed with a p-value of 0.003 and a PPDE value of 0.99 when SLS and untreated samples are compared. These data are available at http://www.igb.uci.edu. A comparison of gene expression levels between three untreated (A1, B1, C1) samples and three water-treated (A3, B3, C3) samples revealed 21,164 genes that exhibited expression levels above background for all samples. This comparison revealed no statistically significant differential expression. Nevertheless, based on our review of the genes assigned the lowest p-values, many of which are associated with inflammation, we believe that the water treatment does lead to some changes in gene expression compared with untreated control skin. These data are available at http://www.igb.uci.edu. For purposes of discussion, only the 100 genes differentially expressed with p-values less than 1.4 × 10-10 and PPDE values greater than 0.99 are discussed here, in Figure 2 and are fully described in Online Table S6. An examination of these top 100 genes most significantly altered when the SLS-treated skin samples were compared with untreated skin samples revealed that, as expected, most of these genes carry out functions related to tissue inflammation and injury (Figure 2; Online Table S6). These differentially expressed genes are proteinases, protease inhibitors, cytokines, chemokines, complement components, HLA factors, or receptors involved in immune regulation. These associations with inflammation and injury responses for many of these mostly upregulated genes are documented in the literature (Online Table S6). These results demonstrate that the tape-stripping method described here harvests RNA suitable for complete gene expression profiles of the skin that accurately reflect its pathological state. Download .jpg (.09 MB) Help with files Table S6Functional grouping of top 100 differentially expressed genes between untreated and SLS treated conditions with p-values less than 1.4 × 10-10 and PPDE values greater than 0.99 Recent advances in molecular medicine have made the possibility of molecular diagnosis a reality (Aitman, 2001Aitman T.J. DNA microarrays in medical practice.Brit Med J. 2001; 323: 611-615Crossref PubMed Google Scholar;Bertucci et al., 2001Bertucci F. Houlgatte R. Nguyen C. et al.Gene expression profiling of cancer by use of DNA arrays: How far from the clinic?.Lancet Oncol. 2001; 2: 674-682Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar;Galiegue and Casellas, 2002Galiegue S. Casellas P. Exploitation of expression profiles: examples in oncology.J Soc Biol. 2002; 196: 313-315PubMed Google Scholar;Lacroix et al., 2002Lacroix M. Zammatteo N. Remade J. et al.A low-density DNA microarray for analysis of markers in breast cancer.Int J Biol Markers. 2002; 17: 5-23PubMed Google Scholar;Whipple and Kuo, 2002Whipple M.E. Kuo W.P. DNA microarrays in otolaryngology-head and neck surgery.Otolaryngol Head Neck Surg. 2002; 127: 196-204Crossref PubMed Scopus (6) Google Scholar;Satagopan and Panageas, 2003Satagopan J.M. Panageas K.S. A statistical perspective on gene expression data analysis.Stat Med. 2003; 22: 481-499Crossref PubMed Scopus (26) Google Scholar). Through the use of microarrays and RNA profiling it is becoming increasingly clear that simple and complex cell populations can be monitored or “profiled” with the intent of understanding the physiologic state of those cells or tissues. This information is expected to lead to more accurate and possibly predictive diagnoses. We have shown here that the use of four small tape strips is an effective and basically non-invasive approach to capturing messenger RNA from the surface of skin and that this technique permits a direct quantitative and qualitative assessment of pathologic and physiologic biomarkers as a function of normal physiology. We have assayed semi-quantitatively the levels of IL-1β and IL-8 mRNA relative to β-actin in normal, water and SLS-occluded skin sites and shown that RNA from tape and biopsy samples produce qualitatively similar results. In order to account for the possibility that changes in β-actin mRNA were responsible for observed changes in the interleukin/β-actin mRNA ratios (Suzuki et al., 2000Suzuki T. Higgins P.J. Crawford D.R. et al.Control selection for RNA quantitation.Biotechniques. 2000; 29: 332-337Crossref PubMed Scopus (711) Google Scholar;Bustin, 2002Bustin S.A. Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): Trends and problems.J Mol Endocrinol. 2002; 29: 23-39Crossref PubMed Scopus (2021) Google Scholar;Tricarico et al., 2002Tricarico C. Pinzani P. Bianchi S. et al.Quantitative real-time reverse transcription polymerase chain reaction: Normalization to rRNA or single housekeeping genes is inappropriate for human tissue biopsies.Anal Biochem. 2002; 309: 293-300Crossref PubMed Scopus (490) Google Scholar)