Background
Due to the hyper-proliferation and hyper-keratinization observed in psoriatic skin, there has been substantial work examining epidermal keratinocyte and keratin dynamics
. Abnormal epidermal turnover can impair skin barrier function and tissue repair; hence keratinocyte or keratin turnover may be critical in disorders of the skin as well as in response to therapy[
1].
Dislipidemia is a further co-morbidity in this systemic disease. Skin lipid content is particularly critical for maintaining skin barrier function and regulating water loss. The stratum cornea is enriched with organized layers of intercellular lipids composed primarily of ceramides, cholesterol and fatty acids[
2]. Psoriatic skin displays abnormal expression and abnormal location/distribution of lipids, and this appears to be at least partly related to disease severity. Therefore there may be significant changes in lipid kinetics in psoriatic individuals reflecting this defect in lipid homeostasis.
We have previously developed methods to measure the kinetics of keratin[
3,
4], triglycerides, fatty acids and cholesterol[
5] using heavy water in human tissue and have also measured kinetics of complex lipids (e.g. galactocerebrosides) from brain myelin[
6].
Protein synthesis can be conveniently measured by use of a continuous label administration, rise-to-plateau approach, based on the incorporation of deuterium (
2H) from heavy water (
2H
2O) into nonessential amino acids (NEAA) in newly synthesized proteins[
6]. This
2H
2O technique has been used to measure the synthesis of proteins in muscle, bone, liver, lung and other tissues[
6‐
10]. Similarly incorporation of
2H from
2H
2O into cholesterol ester, free fatty acids or triglycerides can be used to determine the synthetic rates of lipid turnover.
When adhesive tape strips are applied to the skin surface in humans, a layer is removed composed primarily of lipids, keratin, and other epidermal constituents. The turnover rate of epidermal keratin from tape strips provides an accessible strategy for assessing psoriatic disease activity and treatment effectiveness. We have previously measured epidermal keratin synthesis by combining collection of tape strips with heavy water labeling and mass spectrometric analysis[
4]. We observed that keratin and keratinocytes have equivalent rates of fractional synthesis in a psoriatic animal model and that keratin turnover from tape strips matches keratin turnover from tissue samples[
4]. Here, we have applied our keratin synthesis method to measure skin protein turnover rates in involved and uninvolved skin of psoriatic individuals.
Discussion
There have been previous attempts to measure cell kinetics in both normal and psoriatic skin. Methods that have been used to measure the proliferative activity of the skin include bromodeoxyuridine (BrdU), DNA flow cytometry (FCM), cell cycle markers such as Ki-67 antigen and radio-isotopic tracers. Increased DNA synthesizing cells have been demonstrated in the psoriatic epidermis using methods, including tritiated thymidine and Ki-67 staining. In-vivo tritium labeling studies by Weinstein from the 1980’s predicted a cell cycle time of 311 hr in normal epidermis, 8 fold longer than the 36 hr cell cycle period in psoriatic epidermis[
20].
Using the technique applied in the present study, our previous clinical studies measuring keratin synthesis in normal subjects demonstrated a transit time of about 18 days from the start of heavy water administration until label appearance at the skin surface[
4]. A degree of variability in keratin synthesis kinetics was also observed among normal subjects. Here, we demonstrated similar appearance times for uninvolved skin from psoriatic individuals. This implies that basal turnover rates in skin are similar in psoriatic and non-psoriatic individuals, and that abnormal keratin kinetics are only observed in skin undergoing active disease. We also observed similar keratin turnover rates in uninvolved skin in different anatomic locations on the same individual, again suggesting consistent basal skin kinetics for these subjects except in areas of psoriatic plaques.
Most strikingly, labeled protein (predominantly keratin) was detected in lesional psoriatic skin within as little as three days following heavy water administration. Furthermore, as well as the rapid appearance time of labeled keratin on the skin surface, the keratin that was detected appeared to be nearly 100% newly synthesized during the labeling period. The calculated fractional synthesis rate (Figure
1) reflects the rate of protein synthesis but does not factor in the total protein present in the skin. Since keratin expression is elevated in psoriatic individuals, our observed increased fractional synthesis rate of keratin and rapid time to appearance indicate that the absolute keratin synthesis rate (mg new keratin per day) must be dramatically increased in psoriatic skin during active disease. Despite the small scale of this study, with limited subject number and sampling times, the difference in keratin turnover between involved and non-involved skin is striking and consistent.
All subjects had similar disease severity being diagnosed with severe plaque psoriasis (Table
2). Since subjects were deliberately selected to have severe plaque psoriasis we did not observe a correlation between synthesis rates and disease severity. Subject 3 had the most number of previous treatments including multiple systemic treatments and multiple biologic therapies. Despite his high severity, Subject 3 did not display the shortest or longest lag for involved or uninvolved skin (8 and 15 days respectively). We did not observe a correlation between demographics, disease severity or previous treatment with the keratin kinetics however with such a small number of subjects this study is not powered to reveal these differences.
Table 2
Subject demographics and clinical information
Age | 49 | 49 | 41 | 67 |
Gender | M | M | M | M |
BMI | 25.7 | 32.5 | 32.1 | 31.4 |
Disease severity | Severe plaque psoriasis | Severe plaque psoriasis | Severe plaque psoriasis | Severe plaque psoriasis |
Family history | No family history of psoriasis. | No family history of psoriasis. | No family history of psoriasis. | No family history of psoriasis. |
Previous therapy | UVB phototherapy and methotrexate | UVB phototherapy | UVB and psoralen UVA phototherapy, methotrexate, cyclosporine, etanercept, adalimumab, efalizumab, infliximab, and alefacept | UVB and psoralen UVA phototherapy |
Clinically the PASI score is considered the gold standard of disease activity. However the PASI score is not entirely objective and may be subject to significant inaccuracy (reviewed in[
21]). Instead we utilized trans-epidermal water loss as a quantitative measure of stratum corneum dysfunction (Figure
3). This readout could distinguish between involved and uninvolved skin, highlighting the compromised barrier function of psoriatic skin and corroborating the kinetic data, but demonstrated a high degree of noise in day to day readings across the time course which would make it unsuitable for longitudinal assessment of response to therapy in a clinical trial. We also assessed subject disease activity using an optical readout. “SquamScan” (data not shown) did not reveal any significant difference between involved and uninvolved skin. Whilst diagnosis and management of individuals by the clinician is not hampered by the noise and insensitivity of these metrics it highlights the need for quantitiative tools for use in clinical trials to assess responses to novel therapies. Future studies could apply the keratin synthesis biomarker to cross sectional studies to measure the kinetic rates of variable disease severities or longitudinal studies examining kinetics in response to treatment to develop therapeutics and further the understanding of keratin dynamics in this disorder.
The non-invasive nature of the tape strip approach described here is simple and easily applied in a clinical setting. However, by collecting total skin proteins the differential composition of skin proteins between involved and non-involved skin must be considered. This method measures the weighted average of all protein in the sample and thus is biased to the most abundant. Accordingly, we performed LC/MS/MS to identify the proteins that were being kinetically characterized.
The proteins in Table
1 are ordered by dividing the peptide count, the total number of times peptides from a given protein are observed, by the molecular weight. This yields a number which is crudely related to abundance in the sample. Although there are some problems with using this ordering method (such as variations in efficiency of peptide ionization, variation in peptide resistance to proteolysis or poor peptide chromatograph) it has been demonstrated that abundance changes between related samples can be analyzed by this method[
22].
Keratins 6A and 14 were greatly increased in apparent abundance relative to keratins 1 and 10 in psoriatic skin as compared to uninvolved skin. This data supports previous understanding of psoriatic lesions; keratins 6A and 14 are more prominently present in psoriatic than uninvolved tissue and keratins 16 and 17 are notably expressed in psoriatic tissue but are undetected in uninvolved skin[
23].
Additional observations include that greater than 98% of all peptides observed in the uninvolved skin were from keratins. Thus, protein isolated from uninvolved skin by this method is almost all keratin[
4]. Isolates from psoriatic skin are more complex; 66 proteins were detected. Keratins comprised only about 73% of all peptides observed in psoriatic skin, although the first 8 proteins were all keratins. The non-keratin proteins which appeared to be moderately abundant in this psoriatic tissue were predominantly proteins associated with epithelial organization (Plakoglobin), or epithelial innate immune responses such as inflammation (SERPINB4, S100A8) or antimicrobial (Histone H4). Many of these proteins have been previously demonstrated to be elevated in psoriatic lesions[
24‐
27]. However, because essentially 100% of protein observed in epidermal skin strips from psoriatic subjects was newly synthesized (Figure
1), the altered kinetics of label appearance in proteins from psoriatic patients cannot simply represent labeling of inflammatory proteins but primarily reflects turnover of the predominant protein, keratin.
Another point worth noting is that our approach yields an average turnover rate for the extracted proteins. This approach works well for healthy epidermis which is thought to be composed largely of layers which work their way to the surface at a fairly uniform rate. Psoriatic skin is less ordered and could conceivably vary in turnover rates of some components. In the future, use of advanced dynamic proteomic techniques to pull out turnover rates of individual proteins from the labeling of peptides may be applied with the same heavy water labeling/skin strip collection methodology that is described here.
The current technique has distinct advantages over the traditional methods of determining cell proliferation that have been described previously. Many kinetic studies involve the labeling of human subjects or ex-vivo skin biopsies with radioactive tracers. Both BrdU and 3HdT are toxic and mutagenic, providing ethical and practical considerations preventing their use. Stable isotopes such as deuterium (2H) have a long history as safe, effective methods for measuring synthesis of molecules in experimental animals and humans. Because deuterium is safe for human use, it is easily translated into clinical studies. By using stable isotopes one can label for longer periods allowing the disease to be tracked over time, for example during a treatment or intervention. Further, many historical studies use punch biopsies; this is invasive and not suitable for already damaged lesional skin, especially since injury alone may induce further exacerbations in uninvolved skin (Koebner effect). Tape stripping is a less invasive method of sampling the skin surface and also allows us to take multiple samples from the same subjects over time. No Koebner effect was observed in any subject during this study. The safety and simplicity of this tape-stripping technique enables the possibility for much larger scale clinical studies, giving a more accurate determination of kinetics based on a larger population size.
Administration of deuterium oxide and collection of skin samples is very easy making this assay highly attractive; however as with any new assay feasibility and cost must be considered. Unlike other methods of assessing protein turnover which historically have used short term i.v. infusion of radiolabelled tracer or invasive punch sampling (discussed above) in this study subjects drink pre-bottled D2O at home and are then tape stripped. In our study tape strips were collected by a single person to minimize any sampling variability between subjects in such a small cohort and because the subjects were already visiting the clinic for pre-existing visits. However skin collection could also be carried out by the subject in an outpatient setting as tape strip collection is painless and requires minimal training. Following collection tape strips are stuck onto a laminated card and can be stored in the subjects’ freezer until collection. It is conceivable to imagine a kit mailed to the subjects home of D2O, tape strips and collection cards. The abundance of keratin in the tape strip sample means protein isolation is relatively easy for any chemical laboratory. We envision this method being most applicable to interpretation of clinical trials rather than routine disease management of patients. In this scenario access to a mass spectrometry instrument and chemical lab is not limiting. Additionally, new mass spec methods such as MRM or SISCAPA, are high through put and have been used to translate biomarker approaches from discovery to clinical trials. These methods can be used to measure a targeted protein of interest from a complex biological sample and this method could be adapted to this platform.
The availability of a rapid, quantitative biomarker of disease and treatment efficacy, either protein or lipid, would have wide spread clinical applications. It would enable the basic study of disease pathogenesis and allow for the rapid assessment of novel therapeutics (decreasing the time and cost of clinical trials and reducing patient exposure to new agents with unknown side effects). We predict that our keratin synthesis biomarker would be especially suitable to detecting early response to treatment prior to detection by traditional clinical metrics. An objective marker of psoriatic disease activity, such as the keratin kinetic biomarker described here, could also be used to assess the severity of other psoriasis sub-types such as palmoplantar psoriasis, guttate psoriasis, inverse psoriasis, and erythrodermic psoriasis. These subtypes are not well measured by the current scales which were designed to evaluate plaque psoriasis. The method described here is potentially applicable to all phenotypes of psoriasis.
As well as analyzing keratin turnover as described above, we can obtain other skin constituents from the tape strip. We have previously developed methods to measure the kinetics of lipids[
5,
6] and cell proliferation[
28,
29] using heavy water in human tissue. The adaption of this tape strip technique to enable kinetic assessment of specific skin cells, lipids or proteins may yield further insights into the mechanisms behind this complex disorder. We have recently devised methods of measuring multiple protein kinetics simultaneously[
30‐
32] which could further expand the utility of this method in looking at the pathology of psoriasis or response to treatment.
Competing interests
C. Emson, G. Lindwall, K. Li and S. Turner are all employees and have stock options in Kinemed. M. Hellerstein is a consultant and stock holder at KineMed.
Authors’ contributions
CE was responsible for study design, clinical documentation, heavy water production and distribution, data analysis, data interpretation and drafting the manuscript. SF was responsible for clinical sample collection, TEWL analysis and manuscript review. GL was responsible for sample preparation for mass-spec, data analysis and interpretation and manuscript review. KL was responsible for data interpretation and manuscript review. MH was responsible for data interpretation and manuscript review. WL was responsible for subject recruitment, clinical evaluation and manuscript review. MH was responsible for study design, data interpretation and manuscript review. ST was responsible for study design, data interpretation and manuscript review. All authors read and approved the final manuscript.