Detection of IL-36γ through noninvasive tape stripping reliably discriminates psoriasis from atopic eczema.

To the Editor:

Inflammatory skin reactions, regardless of their distinct underlying pathophysiologic mechanisms, can often present with similar morphology. Although typical eczema and lesions of plaque psoriasis are easily distinguishable by experienced dermatologists, these 2 diseases often cause diagnostic difficulties when inflammation is minimal or located in certain anatomical regions, including flexures, the scalp, auricular or palmoplantar areas. In these cases diagnosis can be problematic, in particular in the primary care setting, where misdiagnosis can lead to delayed appropriate treatment, overuse of antibiotics, or both.

Histopathology, which is the current gold standard in diagnosing challenging cases, is invasive, costly, and often unavailable in the primary care setting. Furthermore, histopathologic differentiation of psoriasis from eczema in some anatomical localizations, such as the palmoplantar region, is difficult. Unfortunately, there are also no reliable blood-derived diagnostic biomarkers that cover the wide range of clinical phenotypes and disease severities, and although RNA signatures from lesional skin biopsy specimens have been described for atopic dermatitis (AD) and psoriasis, mRNA analysis can be costly and labor intensive. Therefore a simple, noninvasive, and reliable diagnostic approach would be of great clinical benefit. To address this need, this study uses a noninvasive, tape-stripping, and ELISA-based approach to investigate potential protein biomarkers that are able to discriminate eczematous from psoriatic inflammation presenting with a range of severities.

Because the epidermis is a significant source of chemokines, inflammatory lesions from patients with psoriasis and those with AD were sampled initially by means of tape stripping and analyzed for the neutrophil-recruiting chemokines CXCL1 and IL-8, as well as CCL20, which recruits IL-17/IL-22–producing cells (see the Methods section in this article's Repository at www.jacionline.org for detailed methodology). Although these chemokines were found at significantly greater levels in tape-stripping samples from lesions of patients with psoriasis compared with those with AD, receiver operating characteristic (ROC) curve analysis indicated that they would not be ideal as strong discriminators of the 2 conditions (IL-8: area under the curve [AUC], 0.83; SE, 0.0523; 95% CI, 0.726-0.931; CXCL1: AUC, 0.796; SE, 0.049; 95% CI, 0.7-0.891 [Fig 1, B]; CCL20: AUC, 0.82; SE, 0.045; 95% CI, 0.731-0.905). For all 3 chemokines, no or very low levels of protein were detected in healthy and nonlesional samples.

Fig 1

Comparison of IL-36γ (A), CXCL1 (B), and CCL17 (C) levels in tape samples from psoriatic and atopic eczema lesions. Tape stripping was performed on healthy volunteers and both lesional and nonlesional (NL) skin from patients with clinically diagnosed AD or psoriasis (Pso). Tape-derived cytokine levels were normalized to total protein content. ROC curves are depicted for all 3 parameters comparing samples from patients with eczema and those with psoriasis. ****P < .0001.

Because high levels of IL-36γ mRNA and protein have been reported in lesions of patients with psoriasis,3, 4 IL-36γ from tape samples was quantitated through use of a novel in-house sandwich ELISA (see the Methods section in this article's Online Repository). IL-36γ showed a trend for increased levels in patients with AD compared with levels seen in nonlesional or healthy skin (mean, 71/13.5/57.5 pg/μg total protein, respectively). However, IL-36γ levels within psoriatic lesions were significantly greater than those in AD lesions (mean, 719 vs 71 pg/μg total protein; Fig 1, A). No difference in IL-36γ expression was observed in healthy versus nonlesional psoriatic skin taken from ventral forearm areas.

A ROC curve (AUC, 0.987; SE, 0.0114; 95% CI, 0.965-1.01) was plotted to determine the sensitivity and specificity of IL-36γ in a diagnostic approach (Fig 1, A). At an optimal cutoff level of 214 pg/μg total protein (Youden index, 0.944), IL-36γ specificity was calculated as 100% (95% CI, 90% to 100%), and sensitivity was calculated as 94.44% (95% CI, 84.61% to 98.84%). This suggests that IL-36γ has an excellent potential as a diagnostic marker for psoriatic inflammation. When compared with CXCL1, the increased expression of IL-36γ is far more consistent, being found across all patients with psoriasis. We also measured IL-36γ expression in a range of other skin pathologies, including fungal infection, lichen planus, systemic-, subacute cutaneous-, and chronic discoid lupus erythematosus, all of which showed expression levels less than the cutoff level of 214 pg/μg total protein (data not shown).

Both CCL27 and CCL17 have been suggested as potential biomarkers for AD. CCL17 preferentially attracts IL-4/IL-13–producing lymphocytes, whereas CCL27 is involved in the homing of memory T cells to the skin. Epidermal tape stripping demonstrated CCL17, but not CCL27, to be of value in identifying AD (Fig 1, C). CCL27 levels were increased in both psoriatic and AD lesions. In those cases in which CCL17 was detectable, it pointed to an underlying AD inflammation (AUC, 0.79; SE, 0.058; 95% CI, 0.676-0.903). However, unlike IL-36γ, which was detectable in all psoriatic lesions tested, a significant number of AD samples (25%), as well as 57.4% of psoriatic samples, did not show any measurable CCL17.

The patient cohort (see Table E1 in this article's Online Repository at www.jacionline.org) included in our analysis showed prototypic plaque psoriasis or AD. However, to further illustrate the potential of tape-collected IL-36γ as a diagnostic approach for psoriasis, we investigated clinical cases of unclear diagnosis. In each case the diagnostic value of IL-36γ was correct when including dermatohistopathologic results. As an example, we show 2 clinical cases (for details, see the Methods section in this article's Online Repository).

Table E1

Demographic characteristics of patients

	Patients with AD	Patients with psoriasis	Healthy subjects
Sex
Male	28%	51%	38%
Female	72%	49%	62%
Age (y)
Mean	42	47	37
Range	21-69	18-67	25-50

The first case was given a clinical diagnosis of and treated for palmoplantar eczema in our dermatology department. However, analysis of tape stripping showed IL-36γ levels of 960 pg/μg total protein, which are indicative of psoriasis. This fact and subsequent dermatohistopathologic analysis and changes in clinical features were supportive of a corrected diagnosis to that of psoriasis (Fig 2, A).

Fig 2

Tape stripping for the diagnosis of difficult clinical cases. Shown are pictures of 2 clinical presentations in which the clinical diagnosis was unclear or challenging. Based on an IL-36γ cutoff level of 214 pg/μg total protein, tape stripping results indicated psoriatic inflammation for case A but not case B. Histopathologic examination confirmed the results. A, Palmar psoriasis was misdiagnosed initially as hand eczema. B, Chronic eczema. For further details on Fig 2, see this article's Online Repository at www.jacionline.org.

The other case was a patient with symptoms of joint pain in addition to erythrosquamous skin lesions. For the referring rheumatologist, a confirmation of these skin lesions as psoriasis would guide future diagnostic and treatment pathways in the direction of psoriatic arthritis. However, tape stripping did not confirm psoriatic inflammation, and the diagnosis for the skin lesions was confirmed to be chronic eczema (Fig 2, B).

Numerous soluble mediators, cell-surface molecules, and intracellular proteins have been described previously to be upregulated in psoriatic lesions. In our noninvasive approach we could confirm that levels of the neutrophil chemoattractants IL-8 and CXCL1, as well as CCL20, the chemoattractant for IL-17–producing cells, are increased in psoriatic compared with eczema lesions. Contrary to work showing reduced expression of CCL27 in patients with psoriasis and the value of CCL27 as a biomarker for AD, we did not find significant differences in the amount of this chemokine. The differences between our findings and those of some others is likely to be due to the method of sampling.

In conclusion, although the effects of other factors, such as systemic and topical therapeutics, age, sun exposure, and lesion chronicity, need to be investigated in future studies, the results presented here confirm IL-36γ to be a robust, specific, and reliable biomarker for psoriatic inflammation that outperforms previously reported biomarkers and is likely to withstand all challenges in real-life primary and secondary dermatologic care.

Table E2

Comparison of local severity scores of patients with AD and psoriasis

	Patients with AD	Patients with psoriasis
Mean erythema score	2.40	2.61
Mean infiltration score	2.37	2.14
Mean scaling score	1.88	2.18
Mean total lesion score	2.23	2.31

A local severity score was determined by experienced dermatologists for all patients with eczema and psoriasis included in the study.