Psoriasis is characterized by deficient negative immune regulation compared to transient delayed-type hypersensitivity reactions

Diphencyprone (DPCP) is a hapten that causes delayed-type hypersensitivity (DTH) reactions in human skin, and is used as a topical therapeutic for alopecia areata, warts, and cutaneous melanoma metastases. We examined peak DTH reactions induced by DPCP (3 days post-challenge) by comprehensive gene expression and histological analysis. To better understand how these DTH reactions naturally resolve, we compared our DPCP biopsies to those from patients with psoriasis vulgaris, a chronic inflammatory disease that does not resolve. By both microarray and qRT-PCR, we found that psoriasis lesional skin has significantly lower expression of many negative immune regulators compared to peak DPCP reactions. These regulators include: interleukin-10, cytotoxic T lymphocyte-associated 4 (CTLA4), programmed cell death 1 (PD1), programmed cell death 1 ligand 1 (PDL1), programmed cell death 1 ligand 2 (PDL2), and indoleamine 2,3-dioxygenase (IDO1). Their decreased expression was confirmed at the protein level by immunohistochemistry. To more completely determine the balance of positive vs. negative immune regulators in both DPCP reactions and psoriasis, we developed one comprehensive gene list for positive regulatory (inflammatory) genes, and another for negative regulatory (immunosuppressive) genes, through Gene Ontology terms and literature review. With this approach, we found that DPCP reactions have a higher ratio of negative to positive regulatory genes (both in terms of quantity and expression levels) than psoriasis lesional skin. These data suggest that the disease chronicity that distinguishes psoriasis from transient DTH reactions may be related to absence of negative immune regulatory pathways, and induction of these is therefore of therapeutic interest. Further study of these negative regulatory mechanisms that are present in DPCP reactions, but not in psoriasis, could reveal novel players in the pathogenesis of chronic inflammation. The DPCP system used here thus provides a tractable model for primary discovery of pathways potentially involved in immune regulation in peripheral tissues.

Introduction Diphencyprone (DPCP) is a hapten that induces delayed-type hypersensitivity (DTH) reactions in human skin, and is used therapeutically for alopecia areata 1 , warts 2 , and cutaneous melanoma metastases 3 . The mechanisms by which DPCP decreases pathogenic immunity for the promotion of hair growth in alopecia areata are incompletely understood. DPCP has been shown to alter the cytokine profile in treated alopecic scalp, in particular increasing interleukin (IL)-2 and IL-10 expression 4 . This increased IL-10 expression has been hypothesized to inhibit the lesional T cells of alopecia areata, but a comprehensive evaluation of other negative immune regulators induced by DPCP is lacking. We have previously shown that human skin responses to DPCP evolve from an inflammatory/effector peak at 3 days post-challenge to a more regulated immune response, with diminished markers of T cell activation, at 14 days. This study included comprehensive gene expression profiling, by microarray and qRT-PCR approaches, of biopsies from DPCP-challenged healthy volunteer skin at 3 days (peak reaction), 14 days (actively resolving reaction), and 120 days (4-8 months; fully resolved reaction) compared to placebo-treated skin 5 . We have also previously performed similar transcriptomic profiling of psoriasis vulgaris lesional vs. non-lesional skin. This resulted in a metaanalysis derived transcriptome (MAD3) which combined the results of 3 individual microarray experiments, in an effort to address the variability in differentially expressed genes observed between experiments 6 . In this study, we expand our previous characterizations of transient DTH reactions and chronic psoriasis biopsies by directly comparing them to each other, particularly in relation to positive and negative immune regulation.

Study subjects and skin samples/Consent
For diphencyprone (DPCP) reaction microarray, qRT-PCR, and immunohistochemistry studies, skin biopsies were obtained from 11 volunteers under a protocol approved by The Rockefeller University's Institutional Review Board (IRB Number JKR-0742). Written, informed consent was obtained from all subjects and the study adhered to the Declaration of Helsinki principles. This trial is registered at clinicaltrials.gov under NCT01452594 (https://clinicaltrials.gov/ct2/show/NCT01452594). For each volunteer, biopsies were taken of placebo-treated skin as well as DPCP reactions 3, 14, and 120 days after challenge, as previously described 5 .
For psoriatic lesional vs. non-lesional skin microarray data, we used the meta-analysis derived (MAD3) transcriptome described in 6. Psoriatic lesional tissue for qRT-PCR and immunohistochemistry studies were from deidentified residual samples of plaque-type psoriasis vulgaris from previous studies for whom no clinical characteristics are available; a psoriasis area severity index of more than 12 (moderate-to-severe psoriasis vulgaris with >10% body surface area involvement) was required for entry into these trials.

Results
Since DTH reactions naturally resolve, we sought to compare our DPCP biopsies (from 5) to those taken from patients with psoriasis vulgaris (from 6), a chronic T cell-mediated inflammatory disease that does not resolve and which, in many ways, represents amplifications of background immune circuits that exist in normal human skin 9 . To globally assess the balance of positive vs. negative immune regulators in both DPCP reactions and psoriasis using our microarray data, we developed one comprehensive gene list for positive regulatory or inflammatory genes and another gene list for negative regulatory or immunosuppressive genes (through Gene Ontology terms and literature review previously discussed in 5, Table 1 has "negative regulator" list and fold change values for DPCP day 3 and psoriasis transcriptomes, "positive regulator" list is derived from GO term 0002684 "positive regulation of immune system process" but with genes removed that are in common with GO term 0002683 "negative regulation of immune system process"). Our microarray data showed increased fold changes of many negative regulators in DPCP day 3 biopsies vs placebo-treated skin (DPCP day 3 transcriptome) compared to psoriasis lesional vs non-lesional skin (psoriasis transcriptome). For instance, CTLA4 expression was significantly increased 21.6-fold in the DPCP day 3 transcriptome, but non-significantly increased 3.7-fold in the psoriasis transcriptome. Venn diagrams show that the psoriasis transcriptome only has seven genes from the negative regulator list, while the DPCP day 3 transcriptome has 52 (Figure 1a). Although the DPCP day 3 transcriptome also has more genes from the positive regulator list than psoriasis, the odds ratio for the positive regulator list was not significantly different between these two transcriptomes. The odds ratio for the negative regulator list, however, was significantly different (Figure 1b). The altered balance between positive vs. negative regulatory transcripts in psoriasis compared to DPCP reactions can also be seen in Figure 1c which shows that DPCP transcriptomes at all time points (days 3, 14, and 120) have a higher ratio of negative to positive regulator genes than  the psoriasis transcriptome in terms of expression levels for each gene set as a whole (as opposed to number of genes as indicated in the Venn diagrams). This is despite the fact that the DPCP day 3 transcriptome has comparable expression levels of the MAD3 psoriasis transcriptome genes to actual psoriasis samples, and therefore highlights the negative regulator expression that is unique to DPCP reactions.
To confirm some of our microarray findings, we performed qRT-PCR and found that psoriasis lesional skin biopsies have significantly lower expression of many negative immune regulators compared to peak DPCP biopsies. These regulators include lymphocyte activation gene 3 (LAG3), cytotoxic T lymphocyte-associated 4 (CTLA4), indoleamine 2,3-dioxygenase (IDO1), programmed cell death 1 (PD1), programmed cell death 1 ligand 1 (PDL1), programmed cell death 1 ligand 2 (PDL2), and IL-10 ( Figure 2a). We confirmed the decreased expression of these and FAS (which by gene expression had 9.3-and 1.1-fold changes in the DPCP day 3 and psoriasis transcriptomes, respectively) at the protein level by immunohistochemistry (Figure 2b).

Discussion/conclusions
These data suggest that disease chronicity in psoriasis could be related to absence of several negative immune regulatory pathways, with the implication that strategies to obtain stable clearance/restore tolerance in skin lesions may need to focus on increasing these negative pathways. These negative immune mechanisms may be of more general importance for maintaining skin homeostasis as non-inflammatory in the presence of a large population of effector memory T cells that normally reside in skin 10 . In addition, these negative immune regulators are likely involved in the therapeutic applications of DPCP, particularly alopecia areata where IL-10 has already been implicated 4 . Further study of these regulatory mechanisms that are present in DPCP reactions, but not in psoriasis, could reveal novel factors in the pathogenesis of chronic inflammation.
The DPCP system used here provides a tractable model for primary discovery of pathways potentially involved in immune regulation in peripheral tissues.
Author contributions NG and JGK conceived the study and designed the experiments. NG carried out the research. MS-F and JCR contributed to the design of experiments and provided expertise in genomics. NG prepared the first draft of the manuscript. All authors were involved in the revision of the draft manuscript and have agreed to the final content.

Competing interests
No competing interests were disclosed. Of course, further studies are required to extricate the dynamics of immune activation in DTH vs psoriasis lesions; even data on DPCP reactions at 14 days can be found in another paper by the same authors ( ) detailed microarray and quantitative RT-PCR data on the time course of evolving Gulati , 2014 et al. lesions of psoriasis is lacking. Koebner phenomenon, guttate psoriasis and rebounds following clobetasol propionate suppression of plaque psoriasis might provide interesting models in this respect. The authors discuss the potential role of IL-10 regulation in DPCP immunotherapy of alopecia areata, and it might be interesting in this context to recall "Renbök phenomenon" or inverse Koebner phenomenon, namely the observation of normal hair growth in psoriatic plaques in patients with co-existing psoriasis and alopecia areata ( ). Mirmirani, 2015

Grant information
Pathogenic differences between DTH and psoriasis are not limited to the expression of genes involved in negative immune regulation, as illustrated by the lack of effect of blocking Th1/Th17 pathways with ustekinumab on the elicitation of ACD by patch testing ( ), even though IL-1 β Nosbaum . Negative immune regulators include cytokines such as IL-10, IL-35 and transforming growth factor β (TGF β) and membrane proteins which are greatly interdependent and subject to cross-regulation, and several cell types, including Foxp3+ regulatory T cells, monocytes and B cells can be involved in down-regulation of inflammatory and autoimmune processes. Some of these cell types are characterized by their plasticity, so their phenotype might be subject to therapeutic modulation.
In conclusion, the paper by Gulati and coworkers represents a seminal work in the study of the comparative dynamics of immune activation and downregulation in several skin disorders, with potential therapeutic implications.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. Competing Interests: In this paper, Gulati and colleagues have tested the hypothesis that psoriatic skin lesions have significantly lower expression of negative immune regulators compared to peak DTH reactions from diphencyprone (DPCP). This study was done using two previously generated data sets described in previous publications from this group: 1) a study in 11 healthy human volunteers of recall responses to DPCP at 3-day (peak) and 14-day (resolution) using immunohistochemical techniques to visualize the infiltrating immune cells and gene expression (Affymetrix HGU133 Plus 2.0 Arrays) to perform a molecular profiling of the day 3 and day 14 DPCP responses, and in 6 cases 4-8 months later (reference 5); and 2) a meta-analysis derived (MAD) transcriptome of psoriatic skin from 3 different studies using Affymetrix HGU133 Plus 2.0 arrays in which microarray raw data were available through public databases (reference 6), termed the MAD-3 transcriptome.
The results show an imbalance in the expression of negative regulatory genes in DPCP day-3 peak reactions compared to psoriatic skin lesions, favoring the expression of more negative immune regulators in the former vs. the latter group (52 genes vs 7 genes). They generated odds ratios for the positive and negative immune regulators from a list of negative regulatory and positive regulatory genes derived from a GO search. The odds ratio for negative immune regulators was significantly higher in the DPCP day-3 (peak) group than the psoriatic skin group, while the odds ratios for the positive immune regulators were not statistically significantly different between the two groups. The enrichment for negative immune regulators held true also at day-14 (resolving) and day 120 (late). These findings were confirmed by qRT-PCR of several key genes and by immunohistochemistry.
This study is technically sound and the paper clearly written, with acceptable figures and tables. The title 1.
This study is technically sound and the paper clearly written, with acceptable figures and tables. The title and the abstract are appropriate for the research described in the article. The methods are described in sufficient detail for the experiments to be replicated, although it would be useful to provide more detail about the methods underlying the calculations of the odds ratios. The interpretation of the results are reasonable based on the data and the conclusions provide a novel conceptual framework for understanding the fundamental mechanisms of skin homeostasis in psoriasis.
I had one question about the results and their interpretation. In the study from reference 5, the results suggested different patterns of immune responses after DPCP challenge, referred to as subgroups A and B, where subgroup B had higher levels of negative immune regulators at day-3 than subgroup A. Interestingly, subgroup B fit the typical kinetics of a DTH response in which T cell and DC infiltrates peaked at day-3 and were diminished at day-14; whereas, subgroup A showed an unexpected increase in the numbers of T cells and DC's at day-14 compared to day-3. It might be useful to know if the higher expression of negative immune regulators in DPCP skin compared with psoriatic skin was driven by the higher levels of these genes in subgroup B, or if it was generally true across both subgroups.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. Competing Interests: 16  In this welcome addition to the field of psoriasis transcriptomics, Gulati and colleagues sought to test whether expression of genes involved in immune regulation in transient condition, diphencyprone (DPCP) induced delayed-type hypersensitivity (DTH), was significantly different from expression of genes involved in immune regulation of a chronic T cell-mediated disease, psoriasis vulgaris. Employing microarray and qRT-PCR based approaches to profile the transcriptomes from biopsied samples, the authors found that negative immune regulatory genes were significantly overexpressed in the DPCP vs. healthy controls comparison, while such a pattern of overexpression was not observed in the psoriasis vs healthy controls comparison. This lack of overexpression of key negative regulatory genes was found in both the microarray and the qRT-PCR analyses. The authors conclude that the relative dearth of negative immune regulatory genes in psoriasis may be associated with the persistent inflammation found in psoriasis and that strategies to restore negative regulators may have therapeutic implications. This paper is technically sound, with well accepted protocol employed in all wet experiments. The statistical analysis was also performed with well accepted and sufficiently rigorous methodology. The evidence supports the authors' central claims. P-values for differential expression analysis were adjusted for multiple hypotheses by the Benjamini-Hochberg method to control for the FDR.
We believe that this paper will be of interest to researchers in the field of psoriasis transcriptomics and genetics, as well as the broader fields of autoimmunity and dermatology. The article could be strengthened by addressing the following comments and questions: The authors note that "genes with low variation and low expression in most samples were filtered The authors note that "genes with low variation and low expression in most samples were filtered out prior to the analysis". We think that the filtering threshold for variance and expression should be reported as well.
Odds ratios are reported in the results but it is not clear how they were estimated in the methods. Furthermore, interpretation of the odds ratios is unclear as well.
In Table 1, we have questions regarding the FCH and the p and FDR values. Is the FCH the absolute FC? Is it log2? We think that the labeling should be clearer. What do p and FDR values of "0E+00" indicate? Is it for p and FDR values that are high or low?
Could the authors' results be confounded by the fact that psoriatic skin has a larger proportion of keratinocytes relative to DTH skin? If the negative regulators were primarily expressed in non-keratinocyte cell types, then biopsy of psoriatic skin compared to DTH skin might show a lower proportion of negative regulators based simply on cell proportions rather than intrinsic immunologic differences.
We feel that the biology of psoriasis and DTH (or contact dermatitis) can be expanded upon in the discussion section. For instance, a number of the genes listed in Table 1  An apparent difference between the DTH model and psoriasis is that in the DTH model, antigen was transiently given, whereas in psoriasis it is possible that self-antigens are continually expressed. Might this contribute to the observed transcriptional differences? Have the authors' examined the transcriptional profile of a DTH model in which antigen is chronically given?
We have read this submission. We believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. inflammatory skin disease. Both diseases are particularly suited for this approach. DPCP-specific DTH reactions are immune mediated, resolve naturally and have well-defined self-limited kinetics. This allows for the accurate definition of time points for the analysis of inflammatory peak reactions and resolving inflammatory activity. Psoriasis instead constitutes a persistent T-cell mediated skin inflammation.
The approach to obtain insights into the mechanisms perpetuating inflammation consists of the analysis and comparison of gene expression profiles in lesional and non-lesional skin probes of both diseases. Regarding psoriasis the authors refer to meta-analysis derived transcriptome data that cover the gene expression profiles from 193 lesional and non-lesional biopsy pairs as recently published. For DPCP-treated patients 11 lesional/non-lesional sample pairs were included for different time points during inflammation. Like the psoriasis transcriptome analyses the data acquisition in DCPC samples has been peer peer-reviewed in a high ranking scientific journal.
To achieve meaningful results regarding the posed question, the authors define two groups of genes for comparison. One comprehensive gene list includes positive regulatory or inflammatory genes that enhance inflammation, the other genes particularly related to negative regulatory or immunosuppressive functions. The composition of the gene lists is based on scientifically verified Gene ontology terms. In addition to the microarray data, for select genes expression levels are verified by quantitative RT-PCR and by immunohistochemistry analyses of protein expression in skin samples.
As a result the authors report, that at the time of the peak inflammatory response DPCP-induced DTH reactions contain significantly more genes related to suppression of inflammation than psoriasis lesions. This results in an altered balance between positive . negative regulatory transcripts in psoriasis vs compared to DTH reactions, which is maintained throughout DTH healing.
The authors conclude that in psoriasis, but also beyond this disease the lack of negative immune regulatory genes may be related to inflammatory disease chronicity. They furthermore propose that DPCP reactions may serve to examine regulatory immune pathways.
The study addresses a highly relevant question. The scientific approach is novel, original and represents a truly innovative solution to obtain novel insights into immune regulation. The comparison of psoriasis and DCPC reactions is reasonable because in both of them inflammatory reactions result from T-cell driven mechanisms. The methods are sound and appropriate for the analysis. The underlying data are extensive and due to the large sample size create a reliable and credible basis for the study. Accordingly, results and conclusions are well supported. The abstract provides an adequate summary of the study. The results are clearly presented. The discussion is short, draws a clear and valid conclusion and avoids unnecessary speculations. Indeed, the insights from the study provide a novel approach to understand the chronicity of inflammation.
My personal assessment is that of an intriguing approach, which by itself is revealing and at the same time opens novel access paths into the investigation of chronic immune-mediated inflammation. I support the publication of this article in the present form without any reservation. There are no changes necessary.
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. Competing Interests: