miR-21-3p/IL-22 Axes Are Major Drivers of Psoriasis Pathogenesis by Modulating Keratinocytes Proliferation-Survival Balance and Inflammatory Response

Psoriasis is a chronic inflammatory skin disease that is mediated by complex crosstalk between immune cells and keratinocytes (KCs). Emerging studies have showed a specific psoriatic microRNAs signature, in which miR-21 is one of the most upregulated and dynamic miRNAs. In this study, we focused our investigations on the passenger miR-21-3p strand, which is poorly studied in skin and in psoriasis pathogenesis. Here, we showed the upregulation of miR-21-3p in an IMQ-induced psoriasiform mouse model. This upregulation was correlated with IL-22 expression and functionality, both in vitro and in vivo, and it occurred via STAT3 and NF-κB signaling. We identified a network of differentially expressed genes involved in abnormal proliferation control and immune regulatory genes implicated in the molecular pathogenesis of psoriasis in response to miR-21-3p overexpression in KCs. These results were confirmed by functional assays that validated the proliferative potential of miR-21-3p. All these findings highlight the importance of miR-21-3p, an underestimated miRNA, in psoriasis and provide novel molecular targets for therapeutic purposes.


Introduction
Psoriasis is a chronic inflammatory skin disease caused by local and systemic activation of innate and adaptive immunity. It affects people of all ages, with an estimated 3% prevalence worldwide [1]. Psoriatic plaques manifest as thick red irritated skin lesions caused by imbalanced keratinocytes (KCs), hyperproliferation and differentiation impairment, as a result of intricate and overlapping immune pathways [2]. This epidermal hyperplasia-a key feature of psoriatic lesions-is due to an enriched inflammatory environment, in which IL-22, an IL-10 family cytokine, is highly upregulated [3]. IL-22 is mainly produced by T H 22 and T H 17 [4]. It acts on epithelial cells that exclusively express its heterodimeric receptor (IL-10RB/IL-22RA1), resulting in STAT3 downstream-dependent activation [5,6]. IL-22 induces cytokine-specific chemokines and specific effector responses mediated by the IL-23/T H 17 axis [7]. Together with other inflammatory mediators, IL-22 stimulates KCs to produce inflammatory cytokines (IL-1β, TNF-α, IL-6), chemokines (CCL20), and antimicrobial peptides (S100A7, S100A8, S100A9, β-defensin-2), resulting in their hyperproliferation. In addition to IL-22's pleiotropic functions, IL-22-deficient Wild-type (C57BL6/JRj) mice used in this study were purchased from Janvier Labs (Saint Berthevin, France). IL-22-deficient mice (IL-22 −/− ), kindly provided by Dr. Laure Dumoutier from Ludwig Institute for Cancer Research, were bred in the animal facility of the Centre Biophysique Moleculaire UPR4301 (Orléans, France). The experiments were performed in compliance with institutional guidelines and were approved by the local ethical committee. Mice between 8 and 10 weeks of age were shaved on the back skin and received daily, for 6 consecutive days, topical applications of 60 mg imiquimod (IMQ) from a commercially available cream (5%) (Aldara Hoofddorp, Pays-Bas; 3M Pharmaceuticals). Recombinant mouse IL-22 protein (R&D systems, #582-ML-010) at 100 ng was subcutaneously injected on the shaved back 1 day before IMQ treatment and throughout the rest of the experiment.

Scoring of Skin Inflammation Severity
To evaluate the severity of back skin inflammation, a scoring system inspired by the literature was developed. Erythema and scaling were scored independently on a scale from 0 to 4-0, none; 1, slight; 2, moderate; 3, marked; 4, very marked. The sum of erythema and scaling scores were established. Then, a score for each group was averaged, expressed in arbitrary units (a.u.) and trend lines were generated to observe the changes in mouse skin lesions.

Patients and Biopsies
For this study, 4-mm punch biopsies were collected from the lesional skin of 9 patients (5 women and 3 men, with an age range 18 to 60 years) with psoriasis at the Centre Hospitalier Régional d'Orléans (CHRO). Furthermore, 6-mm punch biopsies, collected from 9 healthy donors (9 women with an age range of 41 to 60 years), were supplied from the Laboratoire BIO-EC (Longumeau, France).

Cell Stimulation
HaCaT cells were seeded at 2.5 × 10 5 cells per well of 24-well plate or 6 × 10 5 cells per well of 6-well plates for 24 h to obtain between 50% and 60% confluency. One day prior to stimulation, HaCaT cells were starved by replacing the growth medium with DMEM, with 4.5 g/L Glucose, with L-glutamine (Lonza, catalog # BE12-604F). Then, the stimulation was achieved in the same medium for 24 h with 50 ng/mL recombinant human IL-22 (R&D systems, # 782-IL010).

Luciferase Reporter Assays
One day prior to transfection, HaCaT cells were seeded onto 24-well plates at 1.5 × 10 5 cells per well. The next day, cells were at around 60% of confluence and thus transfection was performed using the PTG1 transfection reagent (Polytheragene SAS, Evry, France) with 2.5 µg pNF-CMV-luc reporter vector [25] at a DNA/Polymer weight ratio of 1/6. Four hours after transfection, the medium was replaced with DMEM-high glucose supplemented with 1% FBS overnight. Then, the cells were treated as mentioned previously. Firefly luciferase gene expression was measured using luciferin (beetle luciferin Promega (#E1601)). Luciferase activity was normalized to total protein amount using a BCA protein assay kit (Interchim (Montluçon, France) protein quantitation kit, BCAssays, UP40840A).  Total RNA and microRNA were isolated from the same piece of skin or from HaCaT  cells using a mirVana™ miRNA Isolation Kit (Ambion Life Technology, (Evry, France) #AM1560), following the manufacturer's instructions. Reverse transcription was performed on 1 µg total RNA with a RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific (strasbourg, France), #K1622) and on 2 µL microRNA with a TaqMan™ Advanced miRNA cDNA Synthesis Kit (Applied biosystems (Villebon sur Yvette, France), #A28007). Quantitative PCR (qPCR) amplifications were performed using cDNA corresponding to 100 ng total RNA using primer sets suitable for Syber green detection using the Luna ® Universal Probe qPCR Master Mix (Promega, #M3004S). On the other hand, a 1/10 dilution of the amplified cDNA of miRNA were achieved using primer sets specific for each miRNA (TaqMan Advanced miRNA Assays, Applied Biosystems, A#25576) and the TaqMan Fast Advanced Master Mix (ThermoFisher Scientific (strasbourg, France), #4444557). The relative expression of CyclinD1 (forward: CCCTCACACCAATCTCCTCA; reverse: TCTTCTTCAAGGGCTCCAGG) was measured using the comparative critical threshold (CT) method, normalized to housekeeping gene β-actin (forward: GGGCGC-CCCAGGCACCAGGT and reverse: CGTGCTCGATGGGGTACTTC) for total RNA and miR-16-5p (stable expression in cell lines and skin models used under the different conditions) for miRNA and determined using the formula 2 −∆∆CT . Primers for TaqMan MicroRNA assays were purchased from Thermo Fisher Scientific (Table 1).

Enzyme-Linked Immunosorbent Assay
To evaluate the inflammatory response, the culture medium was collected from treated keratinocytes and stored at −80 • C. The protein levels of IL-22, TNF-α, IL-1β, IL-4, and IL-6 were assayed using a DuoSet ELISA kit (R&D systems) according to the manufacturer's instructions. To evaluate the inflammatory response in the psoriasis model, the same amount of skin protein lysates from biopsies was used. The concentration of proteins was evaluated using a BCA protein assay kit (Interchim (Montluçon, France) protein quantitation kit, BCAssays, UP40840A).

Phospho-STAT3 (Y705) Assay
The HaCaT cells were seeded into 96-well dishes at a density of 1 × 10 4 cell per well, treated with 0 or 50 ng/mL IL-22 in triplicate for 24 h, and then lysed in the plate and assayed for total or pStat3 using the cell-based Stat3 ELISA kit according to the manufacturer's instructions (R&D Systems, #DYC4607B).

Immunofluorescence Staining
The HaCaT cells were fixed with 4% paraformaldehyde for 15 min at room temperature. For Ki67 staining, cells were permeabilized with 0.1% triton X-100 for 5 min at room temperature. Then, cells were incubated with 1% BSA/0.3M glycine in 0.1% PBS-Tween for 1 h to reduce non-specific protein-protein interactions. Afterwards, cells were then incubated with anti-ki67 (ab15580) at 1 µg/mL concentration overnight at 4 • C; followed by a further incubation at room temperature for 1 h with a goat anti-rabbit IgG Alexa Fluor ® 488 at (Invitrogen A11034) at 4 µg/mL. Finally, the cells were mounted with VECTASHIELD Antifade Mounting Medium with DAPI used to stain the cell nuclei in blue (vector, H-1200). Image analysis was performed using FLoid ® Cell Imaging Station microscopy.

Crystal Violet Staining
The cells were fixed with ice-cold 100% methanol for 10 min, and then stained with 0.5% crystal violet solution in 25% methanol for 15 min at room temperature. Finally, the cells were washed in water several times, until the dye stopped coming off.

Cell Metabolic Activity
A Cell Proliferation kit II (XTT-Roche) allowed us to evaluate the cell metabolic activity by assessing the activity of intracellular oxydoreductases. XTT is a colorless or slightly yellow compound that becomes brightly orange upon reduction by cellular effectors as mitochondrial oxidoreductases. HaCaT cells were seeded in 24-well plates (2.5 × 10 5 per well) and treated as indicated. An XTT assay was performed according to the manufacturer's instructions. Briefly, the XTT solution at a final concentration of 0.3 mg/mL was added into each well and the incubation was carried out for 4 h at 37 • C. During this incubation, an orange formazan solution was formed, resulting from the conversion of the yellow tetrazolium salt XTT by viable cells. The intensity of coloration was quantified by measuring the absorbance with a specific absorbance filter at 450 nm using a Victor spectrophotomer (PerkinElmer, Waltham, MA, USA).

Flow Cytometric Analysis for Ki67 and PI Staining
HaCaT cells were harvested and spun down to a pellet at 1500 rpm for 5 min. The cells were washed with 1 mL of phosphate buffered saline (PBS-1X) (Sigma-Aldrich) and centrifuged at 1500 rpm for 5 min. The cells were fixed with 4% p-formadehyde at room temperature for 30 min. Then, the cells were pelleted by centrifugation, resuspended with a permeabilization buffer (50 µL/mL PI, 100 µg/mL RNase, and 2 mM MgCl 2 ) (50 µL of buffer per 1 × 10 6 cells). Afterwards, cells were centrifuged and stained with 5 µL of Alexa Fluor 700 mouse anti-human Ki-67 antibody (BD Pharmingen, #561277) and isotype control (BD Pharmingen, #557882), followed by room temperature incubation for 30 min, protected from light. Finally, the cells were washed twice with 2 mL of permeabilization buffer, resuspended, and analyzed with a BD LSR Fortessa Flow Cytometer (BD Biosciences).

Monitoring miRNA Activity with RILES
To monitor the functionality of miR-21-5p, -21-3p, -486-5p, and -486-3p in HaCaT cells, four RILES expression plasmids-denoted as pRILES/21.5pT, pRILES/21.3pT, pRILES/486.5pT and pRILES/486.3pT-were constructed as previously described [26]. Briefly, 4 fully complementary block sequences specific to either miR-21-5p, -21-3p, -486-5p, or -486-3p were subcloned in the 3 -UTR part of the CymR transcriptional gene. A control RILES plasmid, denoted pRILES, that did not contain any targeting sequence to any miR, was included in the study. To evaluate the functionality of endogenously expressed miR in HaCaT cells, the cells were seeded onto 24-well plates at 3.5 × 10 5 cells per well. The following day, cells were transfected with 2.5 µg of RILES plasmids using PTG1 (Polytheragene SAS, Evry, France), at the pDNA/Polymer weight ratio of 1/4. After 4 h of transfection, the medium was replaced with DMEM-high glucose media supplemented with 0.5% serum overnight. Then, the cells were treated with IL-22 as described above. Firefly luciferase gene expression in cells was thereafter monitored 24 h later. Luciferase activity was normalized to total protein content using the BCA protein assay kit (Interchim, protein quantitation kit, BCAssays, UP40840A).

RNA-Seq from HaCaT Cells
Total RNA was extracted from the 3 replicates of HaCaT overexpressing 1 nM miR-21-3p mimic or negative control mimic using a mirVana™ miRNA Isolation Kit (Ambion Life Technology, #AM1560), following the manufacturer's instructions. Small RNA libraries were prepared using 0.5 ng of total RNA per sample. The RNA-seq assay was carried out on an Illumina NextSeq 500 sequencer. Sequence reads in the FASTQ format were processed and the differential expression analysis was performed. This work has benefited Cells 2021, 10, 2547 6 of 22 from the facilities and expertise of the high-throughput sequencing core facility of I2BC (Centre de Recherche de Gif-http://www.i2bc.paris-saclay.fr/, accessed on 22 July 2020).

Statistical Analysis
Unless mentioned otherwise, all the experiments in the study were reproduced in at least two independent different experiments with a minimum of triplicates performed per group. All data are presented as mean ± SEM. The two-tailed Student t-test was used to determine significance between two groups. Ordinary two-way ANOVA was used for multiple comparisons of more than two groups. The simple linear regression model was used to analyze the correlation between RNA levels of inflammatory mediators and miR-21-5p or miR-21-3p. Comparisons between multiple groups were submitted to Bonferroni's test and comparisons between 2 groups were submitted to Tukey's test. p-values less than 0.05 were considered significantly different. * p < 0.05, ** p < 0.01, *** p < 0.001.

miR-21-5p/3p Expression in IMQ-Induced Psoriasis in WT versus IL-22 −/− Mice
As a first approach, we compared the relative expression of miR-21 between WT and IL-22 −/− mice, which are known to develop a less severe form of the pathology [8] ( Figure 1). We evaluated the clinical progression of the pathology upon IMQ application in WT and IL-22 −/− mice ( Figures 1A and S1). The application of IMQ at 60 mg on the shaved back skin of WT mice for six consecutive days caused erythema, scaliness, thickening, and acute inflammation ( Figures 1A and S1). In the early onset period (day 0 to day 4), the skin developed red erythematous plaques. Then, it turned into thickened and scaly skin (day 5 to day 7). A similar application on IL-22 −/− mice showed a less severe onset of clinical psoriatic lesions ( Figures 1A and S1).
Next, we analyzed the relative expression of miR-21 between the different groups of mice. Similar relative expression levels of both miR-21 strands were obtained for WT and IL-22 −/− healthy mice. At day 7 post-IMQ application, the expression level of miR-21 was significantly increased for both miR-21 strands (5p and 3p) in WT psoriatic mice compared to WT healthy mice. An upregulation of 4.3 times and 6.3 times for miR-21-5p and miR-21-3p respectively with respect to healthy mice was recorded ( Figure 1B,C). In contrast to WT mice, IL-22 −/− psoriatic mice showed a slight increase of miR-21 expression (around 3-fold) compared to IL-22 −/− healthy mice for both miR-21 strands. Only the increased expression of miR-21-5p was statistically significant. Remarkably, WT psoriatic mice presented much higher miR-21 expression than IL-22 −/− psoriatic mice. This significant upregulation reached 1.3-fold for miR-21-5p and 2.3-fold for miR-21-3p.
Finally, we established a correlation between miR-21-5p ( Figure 1D) and miR-21-3p ( Figure 1E) expression with IL-22 mRNA and protein levels in WT mice. A significant positive correlation was noted with IL-22 mRNA and protein for both strands. Nevertheless, miR-21-3p had a higher positive correlation (p < 0.0001/p = 0.0007) than miR-21-5p (p = 0.03/p = 0.0151) ( Figure 1D,E). Thus, this result suggests a connection between miR-21-3p and the IL-22 axis in psoriasis.  Finally, we established a correlation between miR-21-5p ( Figure 1D) and miR-21-3p ( Figure 1E) expression with IL-22 mRNA and protein levels in WT mice. A significant positive correlation was noted with IL-22 mRNA and protein for both strands. Nevertheless, miR-21-3p had a higher positive correlation (p < 0.0001/p = 0.0007) than miR-21-5p (p = 0.03/p = 0.0151) ( Figure 1D,E). Thus, this result suggests a connection between miR-21-3p and the IL-22 axis in psoriasis.  These findings drove us to evaluate whether the restoration of IL-22 in the IL-22 −/− psoriatic mice would reverse the expression pattern of miR-21 described above. To do so, we treated IL-22 −/− mice with recombinant IL-22 protein and followed the miR-21 expression (Figure 3). One day before and during IMQ or vehicle application, mice received 100 ng of IL-22 or the equivalent amount of physiological serum subcutaneously on the shaved back ( Figure 3A). The miR-21 expression for both strands was assessed on day 3 (early response) and day 6 (late response) post-IMQ/vehicle application ( Figure 3B,C). The mice's body weight fluctuations are shown in Figure S2. IL-22-treated mice and control mice displayed similar levels of miR-21 expression with comparable expression profiles at days 3 and 6 post-treatments. Mice treated with IMQ showed a slight upregulation that reached around 1.26-to 2.7-fold increases for miR-21-3p and 3.2-to 4-fold increases for miR-21-5p. These upregulations rose to~5-fold in mice that received both IL-22 and IMQ treatments. These observations demonstrate a relationship between IL-22 and miR-21-3p and to a lower extent with miR-21-5p. Of note, IL-22 treatment alone was not sufficient to induce miR-21-3p expression, suggesting the implication of other inflammatory mediators. Therefore, we examined the possible correlation of key cytokines (IL-1β, TNF-α, IL-4, and IL-6) known to drive crosstalk between KCs and immune sentinels, which results in psoriasis development [27,28], with the induction of miR-21-5p or miR-21-3p expression ( Figure S3). A positive correlation trend was obtained with IL-1β, whereas IL-6 and IL-Cells 2021, 10, 2547 9 of 22 4 showed a negative correlation trend. These results suggest the interference of other inflammatory circuits downstream of the cytokine activation pathway. Nevertheless, we focused on understanding the mechanism by which IL-22 induces miR-21-3p expression, as they were strongly correlated, compared to other cytokines ( Figures 1D and S3).
fold increases for miR-21-5p. These upregulations rose to ~5-fold in mice that received both IL-22 and IMQ treatments. These observations demonstrate a relationship between IL-22 and miR-21-3p and to a lower extent with miR-21-5p. Of note, IL-22 treatment alone was not sufficient to induce miR-21-3p expression, suggesting the implication of other inflammatory mediators. Therefore, we examined the possible correlation of key cytokines (IL-1β, TNF-α, IL-4, and IL-6) known to drive crosstalk between KCs and immune sentinels, which results in psoriasis development [27,28], with the induction of miR-21-5p or miR-21-3p expression ( Figure S3). A positive correlation trend was obtained with IL-1β, whereas IL-6 and IL-4 showed a negative correlation trend. These results suggest the interference of other inflammatory circuits downstream of the cytokine activation pathway. Nevertheless, we focused on understanding the mechanism by which IL-22 induces miR-21-3p expression, as they were strongly correlated, compared to other cytokines ( Figure 1D and Figure S3).

IL-22 Induces miR-21-3p via STAT3 and NF-κB Pathways
We first looked at the miR-21 promoter region, which contains three and four potential binding elements for STAT3 and NF-κB, respectively ( Figure 4A) [29][30][31]. The latter are known to be downstream transcription factors to IL-22 signaling and have been reported to be upregulated in lesional psoriatic skin [6,[32][33][34]. Accordingly, the stimulation with 50 ng/mL IL-22 of HaCaT cells expressing the luciferase gene under a NFκB-driven promoter resulted in NF-κB nuclear translocation, as reported based on luciferase activity ( Figure 4B). The involvement of STAT3 activation was shown via increased STAT3 phos-phorylation over the total STAT3 ratio ( Figure 4C). These data were validated by functional assays by investigating IL-22-dependent miR-21 induction when STAT3, NF-κB, or both were knocked down by siRNAs ( Figure 4D). When those transcription factors were knocked down individually, the effect on miR-21 expression upon IL-22 stimulation was mitigated. However, the double knockdown of STAT3 and NF-κB resulted in clear miR-21 repression, despite IL-22 stimulation ( Figure 4D)

IL-22-Induced miR-21-3p in KCs Is Functional
To evaluate the potential functional activity of miR-21-3p in the regulation of gene expression, we employed a miR-ON functional monitoring system previously developed in our lab ( Figure 5) [26]. The RILES system relies on an engineered genetic circuit in

IL-22-Induced miR-21-3p in KCs Is Functional
To evaluate the potential functional activity of miR-21-3p in the regulation of gene expression, we employed a miR-ON functional monitoring system previously developed in our lab ( Figure 5) [26]. The RILES system relies on an engineered genetic circuit in which the expression of the luciferase reporter gene is switched ON when a miRNA of interest is processed by the RISC machinery in cells to exercise its mRNA-functional repressive function ( Figure 5A). The RILES system was engineered to bear miR-21-5p or miR-21-3p targeting sequences (pRILES/21.5pT and RILES/21.3pT, respectively). As shown in Figure 5B, the pRILES/21.5pT system was not significantly switched ON in HaCaT cells in response to several doses of IL-22 ranging from 0 to 50 ng/mL. In contrast, the RILES/21.3pT was significantly switched ON in HaCaT cells treated with the same conditions ( Figure 5B). As an internal control, with an irrelevant miR target, we demonstrated that neither the pRILES/486 5pT-nor the pRILES/486 3pT-bearing targeting sequences of miR486-5p and miR486-3p were switched ON in response to IL-22 ( Figure S4). Hence, our monitoring system showed specific IL-22/miR-21-3p induction. These observations reflect functional miR-21-3p activity in IL-22-stimulated HaCaT cells, reinforcing our data. Henceforward, we aimed at understanding the role of miR-21-3p in KC biology by defining its genetic regulatory network.

Global Changes in the Transcriptome of HaCaT Overexpressing miR-21-3p
Most studies have focused on determining one direct gene target per miRNA. In this

Global Changes in the Transcriptome of HaCaT Overexpressing miR-21-3p
Most studies have focused on determining one direct gene target per miRNA. In this study, we decided to have a global view of the gene network influenced by miR-21-3p upregulation. As a matter of fact, miRNA gene regulation is much more complex and could implicate several pathways that are interconnected, via the direct or indirect targeting of single or multiple gene sets. Therefore, we chose to use the RNA-sequencing approach to explore the transcriptomic landscape of miR-21-3p-overexpressing KCs. The overexpression of miR-21-3p was confirmed via RT-qPCR ( Figure S5). The comparison between KCs overexpressing miR-21-3p versus the negative control (NC) revealed a significant (p.adj < 0.05) set of 660 downregulated genes and a set of 884 upregulated genes among the 22,152 detected genes (Figures 6A and S6). These results indicate global changes (around 7%) in the transcriptome signature upon miR-21-3p overexpression. The impact of miR-21-3p upregulation on KCs' molecular and biological pathways was analyzed via gene-annotation enrichment analysis to identify differentially expressed genes (DEGs) using ShinyGO v0.61 and DAVID 6.7 software. Both bioinformatic tools gave similar results. The top 30 biological processes of the downregulated genes, as provided by ShinyGO ( Figure 6B), were mainly related to developmental/differentiation, locomotion/migration, cell communication, and immune system processes ( Figure S6, ShinyGO). Consistently, the DAVID analysis highlighted similar biological processes, including the interferon-gamma-mediated signaling pathway ( Figure S6, DAVID). On the other hand, almost all the upregulated genes in KCs were found in the categories related to immune responses and cellular proliferation, involving cell division, mitotic machinery, cell cycle transition, DNA replication, DNA repair, metabolic processes, RNA machinery, etc. (Figures 6C and S6, DAVID and ShinyGO). It worth pointing out the enrichment for the NIK/NF-κB signaling pathway, as well as the T cell receptor signaling and response to cytokine IL-22-dependent pathways.  The proximal region (<600 bp) of these sets of DEGs was also analyzed and significant enriched motifs in promoters were found ( Figure S8). Among the motifs that were significantly enriched in the downregulated gene set, the NR2C2 transcription factor gene, The proximal region (<600 bp) of these sets of DEGs was also analyzed and significant enriched motifs in promoters were found ( Figure S8). Among the motifs that were significantly enriched in the downregulated gene set, the NR2C2 transcription factor gene, also known as TR4, and its motif were noted in this set. This nuclear receptor and the members of this family participate in many biological processes, such as development, cellular differentiation, and homeostasis [35,36]. Most of the significant motifs include transcription factor motifs implicated in differentiation processes, such as ZNF281 for embryonic stem cell differentiation; the TCF3 motif of the E protein family, which plays a critical role in lymphopoiesis, required for B and T lymphocyte development; and MZF1 for hemopoietic development [37,38]. One motif of NFKB1 was also included in this gene set [39,40].
For upregulated genes, there was an over-representation of motifs belonging to the EF2 promotor family (2 E2F1, 1 E2F2, 1 E2F4 and 1 E2F6), of which the E2F2 gene was augmented. The E2F family plays a crucial role in cell cycle control and tumor suppressor activity [41]. In the same way, KLF7 regulates cell proliferation, differentiation, and survival [42], as well as NRF1, regulating cellular growth and nuclear genes required for respiration, heme biosynthesis, and mitochondrial DNA transcription and replication [43,44]. Another transcription factor family that was over-represented was the CxxC family, which is implicated in chromatin organization and chromatin regulation/acetylation and in methylation patterns following DNA replication. There was also the HES4 motif, a transcriptional repressor implicated in the PI3K-Akt signaling pathway, and ELK4, related to the NGF pathway and ERK signaling.
To further evaluate whether the transcriptomic change observed in our study might correlate with a similar change in situ, in KCs derived from psoriatic lesions, we took advantage of the data recently published by the team of Sonkoly on the KC transcriptome in psoriasis, which showed an IL-22 enrichment, for a comparison with our data [9]. Thereby, we established a series of comparisons between differentially expressed genes in KCs overexpressing miR-21-3p versus psoriasis lesional skin (PP) or psoriasis non-lesional skin (PN) and healthy skin (H). If most of the DEGs in KCs overexpressing miR-21-3p are distinct (1149/1544) from those found in PP and PN, all DEGs in common (395/1544-25%) concern the DEGs of PP, of which 87 were shared strictly with PP vs. H and 303 were shared with PP vs. H and PP vs. PN ( Figure 7A). The significantly enriched biological processes of these 395 genes concerned DNA replication and the cell cycle ( Figure 7B). Moreover, regardless of the number of DEGs in common, the significantly enriched biological processes of all the upregulated genes in KCs overexpressing miR-21-3p and the top 10 genes in PP were similar ( Figure 7A) [9]. These biological processes concerned DNA replication and repair, the cell cycle, and NIK/NF-kappaB signaling. Conversely, the development and differentiation processes of KCs were not shared. No significant similarly enriched biological processed (GO terms) were found with downregulated genes, even among those concerning differentiation and development (Table 2).
In the following part of this study, we intended to validate the occurrence miR-21-3pdependent cellular processes, as revealed by our RNA-seq results. For that purpose, we next assessed the effect of miR-21-3p on KC proliferation and migration. significant similarly enriched biological processed (GO terms) were found with downregulated genes, even among those concerning differentiation and development ( Table 2). In the following part of this study, we intended to validate the occurrence miR-21-3p-dependent cellular processes, as revealed by our RNA-seq results. For that purpose, we next assessed the effect of miR-21-3p on KC proliferation and migration.

miR-21-3p Promotes Keratinocyte Hyperproliferation
Since miR-21-3p overexpression upregulates genes related to cellular survival and proliferation, as suggested above, and since KC hyperproliferation is a hallmark feature of psoriatic plaque formation, we examined the impact of miR-21-3p overexpression on KC proliferation. To this end, we overexpressed or inhibited miR-21-3p in KCs, using oligomers. The efficiency of miR-21 overexpression ( Figure S5A,B) and inhibition ( Figure S5C) was confirmed via RT-qPCR analysis. Then, we assessed the proliferation via a crystal violet assay, an XTT assay, and Ki67 and PI staining for the cell cycle (Figures 8 and S10).
The crystal violet assay indicated that KC density depended on miR-21-5p and miR-21-3p expression status Figures 8A and S10A). Indeed, kinetic studies indicated that KCs overexpressing miR-21-5p or miR-21-3p presented better growth compared to negative controls and to miR-21-5p or miR-21-3p inhibition over time. We also noted a more rapid increase in proliferation for miR-21-3p-expressing KCs at an earlier time (starting from 24 h) than KCs overexpressing miR-21-5p ( Figure 8A). These observations were reinforced by the XTT results, which showed a highly significant amplified proliferation rate for miR-21-3p overexpression compared to negative controls (NC) and miR-21-5p mimics ( Figure 8B). It is noteworthy that miR-21 inhibition resulted in a limited proliferation enhancement ( Figure S10B). Accordingly, we observed an increased expression of the proliferation marker Ki67 in miR-21-overexpressing KCs, whereas Ki67 was reduced when miR-21 was inhibited ( Figure 8D). The outcome of miR-21 inhibition was modest compared to the efficiency ( Figure S10D). According to the literature, RT-qPCR data could incorrectly indicate a higher inhibition of miRNA due to masking effects [45]. Nevertheless, in our case, miR-21 inhibition gave encouraging results, in line with miRNA mimic data. In addition, cell cycle analysis revealed a significant 15% increase in S-phase-proliferating miR-21-3p-overexpressing KCs compared to NCs ( Figure 8C). Consequently, this increase was absent upon miR-21-3p inhibition ( Figure S10C). Furthermore, the percentage of Sphase-proliferating cells compared to G0-G1 cells was 13% for miR-21-5p-overexpression in KCs, whereas it reached 30% for miR-21-3p-overexpressing KCs. Given all these results, we analyzed CylcinD1, an essential regulator of G1-S phase transition that promotes cell proliferation, at the mRNA level via RTqPCR analysis, and the proteomic level via Western blot analysis ( Figure S9) [46]. Consistently, the overexpression of miR-21-3p and miR-21-5p resulted in 2.7-times and 1.7-times increases in CyclinD1 mRNA expression, respectively ( Figure S9A). At the protein level, the overexpression of both miR-21 strands caused CyclinD1 upregulation, with 33% and 39% for miR-21-5p and miR-21-3p mimics, respectively ( Figure S9B). Overall, these results support the RNAseq data, suggesting that miR-21-3p promotes the hyperproliferation of KCs.

Discussion
Despite all the findings and research advances in the biology miRNAs, skin miRNAs remain a relatively recent area of study, lacking understandings and mechanistic insights on the role and function of miRNAs in skin biology and disorders. In this study, we were interested in investigating miR-21, in particular the miR-21-3p strand, which is poorly studied in cutaneous systems. For this reason, we evaluated the expression of the two versions of miR-21 (5p and 3p) and showed their increased expression in the development of psoriasis. In the literature, only one study in addition to our results revealed the upregulation of miR-21-3p, considered as passenger strand, in psoriatic lesions, without further investigations of its role in psoriasis [23]. Interestingly, only few days ago, Li et al. showed an upregulation of miR-21-3p in NHEK co-cultured with psoriatic dermal mesenchymal stem cells, pointing to the importance of this passenger miRNA in psoriasis [47]. Although miR-21-3p expression was lower than that of miR-21-5p, our findings are

Discussion
Despite all the findings and research advances in the biology miRNAs, skin miRNAs remain a relatively recent area of study, lacking understandings and mechanistic insights on the role and function of miRNAs in skin biology and disorders. In this study, we were interested in investigating miR-21, in particular the miR-21-3p strand, which is poorly studied in cutaneous systems. For this reason, we evaluated the expression of the two versions of miR-21 (5p and 3p) and showed their increased expression in the development of psoriasis. In the literature, only one study in addition to our results revealed the upregulation of miR-21-3p, considered as passenger strand, in psoriatic lesions, without further investigations of its role in psoriasis [23]. Interestingly, only few days ago, Li et al. showed an upregulation of miR-21-3p in NHEK co-cultured with psoriatic dermal mesenchymal stem cells, pointing to the importance of this passenger miRNA in psoriasis [47]. Although miR-21-3p expression was lower than that of miR-21-5p, our findings are in agreement with previous evidence suggesting that the passenger strand is usually less abundant than the mature miRNA strand [23,48]. Similar observations were noted in ovarian cancer and in UV-induced KC inflammation models [23,49]. However, for the first time, the current study demonstrates the co-expression of leader and passenger strands of miRNAs in skin, long thought to be subject to strand selection, which is believed to be dependent on the relative thermodynamic stability of the two ends of the duplex. The leader strand is considered the effective strand, whereas, until now, no clear and solid evidence has existed concerning the fate and the mechanism of removal or degradation of the passenger strand. Some studies have reported that this passenger strand might be destined either to be degraded as a mere carrier strand or to be expressed abundantly as a potential functional guide miRNA [48,[50][51][52][53]. This study introduces new evidence on the potential functional expression of both strands, which has been underestimated.
Even though the etiology of psoriasis is still not fully understood, a complex crosstalk between KCs and the abnormal activation of the innate as well as adaptive immune responses that drive epidermal hyperplasia via a large panel of inflammatory mediators, is very well defined [27]. Therefore, we examined cytokines to search for a potential trigger of miR-21 using a correlation statistics test. We identified IL-22 as a major trigger of miR-21 in lesional skin, with a strong relationship with miR-21-3p. Furthermore, we confirmed these results in vitro using different KC cell types and human skin biopsies, as well as in vivo using an IL-22 −/− IMQ mouse model. Of note, we were expecting an upregulation of miR-21-3p expression in IL-22-treated IL-22 −/− mice, but these mice did not modulate its expression. This suggests the interference of other mediators that are not observed under steady-state conditions. Indeed, IL-1β and TNF-α were shown to stimulate miR-21-3p expression in human aortic endothelial cells [54]. Most probably, the inflammatory storm in psoriasis drives the activation of several inflammatory pathways that, in one way or another, promote miR-21-3p expression either by recruiting immune cells or by inducing the hyperproliferation of KCs. Nevertheless, we believe that the primary induction of miR-21-3p by IL-22 would initiate both KC hyperproliferation and an exacerbated inflammatory response, leading to the formation of a positive feedback loop, which may interfere with miR-21-3p transcription. However, the use of RILES technology allowed us to underline the functional activity of miR-21-3p in KCs in response to IL-22 stimulation, consequently putting this cytokine at the center of psoriasis immunopathology and miRNA modulation [55,56].
We also demonstrated that IL-22 induces miR-21-3p transcription directly or indirectly through STAT3 and NF-κB signaling, which are crucial mediators of several biological and immunological processes, including inflammation and proliferation, subsequently ensuring the connection between KCs and immunocytes in the pathogenesis of psoriasis [57,58]. These results suggest that the overexpression of miR-21 in psoriatic KCs is likely a consequence of the production of excess inflammatory cytokines, in particular IL-22 in skin lesions, and may be implicated in the occurrence of epidermal hyperplasia in psoriasis. Thus, in the forthcoming years, a global view of the psoriasis immunopathogenesis events is needed in order to understand the connections between all inflammatory actors, from cellular to molecular mediators, including miRNAs. We would like to address a curious difficulty in the detection of NF-κB signaling mediated by IL-22, as the time of detection is challenging. A possible explanation is that IL-22 upregulates miR-21-3p, which has been revealed to impair p65 nuclear translocation [54]. Hence, it is only at early stages of IL-22 stimulation that NF-κB induction detection would be possible.
In line with our hypothesis that miR-21-3p and the IL-22 axis play essential roles in psoriasis, we performed RNA-seq analysis in miR-21-3p-overexpressing KCs to gain deeper insights into the transcriptome. Our results highlight an overall upregulation in the categories related to immune responses and cellular proliferation and a downregulation of networks related to the interferon-gamma-mediated signaling pathway and differentiation processes. These observations on the proliferation-differentiation balance were confirmed by multiple functional assays that validated the capacity of miR-21-3p to boost KC proliferation and migration. A very interesting and intriguing outcome of miR-21-3p overexpression is the downregulation of IFN-γ signaling pathways. Undeniably, IFN-γ is a critical cytokine involved in several cellular processes and defenses including proliferation and innate-adaptive immunity against invading pathogens. In steady-state conditions, IFN-γ drives the growth arrest of KCs [59]. This is in line with the proliferative properties of miR-21-3p observed in our study. On the other hand, in psoriasis, IFN-γ is thought to mediate T cells' interactions with KCs. Furthermore, IFN-γ was shown to correlate with disease severity and prognosis [60]. Incontrovertibly, the downregulation of IFN-γ signaling raises a fundamental question about the nature of miR-21-3p regulation in psoriasis and points to its cell-specific function. Hence, further investigations on the function of miR-21-3p in other cellular subsets would be required in order to understand its regulatory mechanism in inflammatory diseases. Additionally, miR-21-3p overexpression upregulated genes linked to non-canonical NF-κB and T cell receptor signaling. These findings point to a probable role of miR-21-3p in exacerbating the inflammatory response. In fact, NF-κB hyperactivation is one of the main factors causing inflammation in psoriasis. The upregulation of non-canonical NF-κB signaling is dependent on NIK (MAP3K), which would mediate a long and persistent activation of the p52 complex [61]. It is plausible that NIK promotes TCR-stimulated STAT3 phosphorylation, thereby regulating the polarization of T naïve cells into T H 17 phenotypes and promoting the proliferation of KCs [62]. This could be a very interesting point to consider, as the IL-17/IL-23 axis plays an essential role in the pathogenesis of psoriasis [63,64]. IL-23 inhibitors revolutionized the management of psoriasis with their high efficacy and the fact that they were well-tolerated when used as treatments for moderate-to-severe-plaque psoriasis [63,65,66]. More recently, the strength of IL-17 has been proven with the use of the new generation of biologics against IL-17A and IL-17F. Indeed, patients with moderate-to-severe-plaque psoriasis treated with IL-17 inhibitors achieved a long-lasting PASI > 90 [67][68][69][70]. Altogether, the RNA-seq results provide accumulating evidence on the abnormal immune regulatory genes induced by miR-21-3p overexpression, which are involved in the molecular pathogenesis of psoriasis.
Finally, this study highlights the pleiotropic functions of miR-21-3p in the cutaneous system, from regulating proliferation and cell survival to potentially modulating innate and adaptive immunity. In light of these findings, the role of miR-21-3p should be further addressed in upcoming studies by covering the different cell subsets present in the skin, such as melanocytes, T cells, Langerhans cells, etc. Altogether, these results open novel therapeutic strategies in which miR-21 inhibition of both strands should be considered. Furthermore, one can imagine a combination of therapeutics, in which the naturally occurring IL-22 antagonist (IL-22BP) could serve as a complement to counterbalancing or limiting miR-21 induction, without inducing side effects as in the case of biologics.