Anti-Irritant Strategy Against Retinol Based On The Genetic Analysis Of Korean Population: A Genetically Guided Top–Down Approach

Keywords: retinol; retinoid; cosmetics; anti-irritation; genetically customized; single nucleotide polymorphisms 

Click Here To Get More Infos About Cistanche Anti-Aging Treatment

1. Introduction

Retinoids, a family of compounds derived from vitamin A that includes both synthetic and natural compounds, have been intensively studied and utilized in diverse biomedical applications such as cancer treatment (acute myeloid leukemia), cervical neoplasia, and skin disorders in the past few decades. The wide application of retinoids, especially retinoic acid, is based on the scientifific premise that retinoic acid is involved in numerous biological pathways in nature by triggering interactions between nuclear receptors that control gene expression [1–5]. One of the most striking applications of retinoids is the dermatologic approach, along with pharmaceutical drugs and cosmetic products. Since retinoids were fifirst used for the treatment of acne in the 1940s, its therapeutic effificacy against skin disorders such as actinic keratosis, psoriasis, and ichthyosis is reported [6]. As many scientifific studies and clinical studies demonstrated the therapeutic effificacy of tretinoin (all-trans retinoic acid) on photoaging, the development of retinoid derivatives for use as cosmetic ingredients was triggered. Cosmetic retinoids, which include retinyl palmitate, retinyl acetate, retinol, and retinal have been intensively investigated in the commercial market since retinoic acid was prohibited in cosmetic products by global regulations such as Annex II 375 (EC 1223/2009). Some studies have revealed that these cosmetic retinoids are also effective for the treatment of photoaging, whose pathologic characteristics involve fine and coarse wrinkles, skin roughness, and abnormal pigmentation, and even argued that they are potent as tretinoin [7]. 

However, retinoids cause irritation and some side effects characterized by erythema, scaling, burning, and itching, which are also termed retinoid dermatitis [8]. Some studies have investigated the mechanism of irritation, and it is suggested that diverse mechanisms are involved in retinoid-induced irritation, which include extensive and pervasive inflflammation characterized by the release of cytokines and infifiltration of immune cells [9], and skin barrier disruption characterized by genetic imbalance of the cornifified envelope (CE)-related factors [10,11]. In addition, the phototoxicity of retinoid and its degraded by-products due to its instability under UV and heat has been widely suggested to contribute to the retinoid-induced irritation [12]. However, these observations are insuffificient to explain the retinoid-induced irritation thoroughly, and no clear conclusion has been reached to date. For this reason, current retinol-based products usually have a “tedious” anti-inflflammatory formula, which is generally and universally used in general commercial products. Therefore, an anti-irritation strategy against retinol based on a strong scientifific background has not yet been established.

A genetic approach can be another solution for establishing an anti-irritation strategy against retinol. Genetic variations were extensively studied to discover genes associated with drug effificacy and side effects [13]. Nelson et al. showed that drugs predicted by known genetic associations account for only 2.0% of the drugs at the preclinical stage, although the proportion increases to 8.2% in approved drugs [14]. This implies that focusing on targets discovered by genetic studies would dramatically increase the chance of successful drug development, as demonstrated with drugs for cholesterol or rheumatoid arthritis [15–18]. 

Retinol treatment induces changes in the expression of transcription factors related to cellular growth [19] and structural proteins [11]. More specififically, a variety of proteins were implicated in retinol-induced irritation, which include inflflammation-related cytokines (MCP-1, TNF-a, interferons, and interleukins) [9], retinoic acid receptor (RARG) [20], pain-sensing channel transient receptor potential vanilloid 1 (TRPV1) [21], and mast cell-activating GPCR (MRGPRX2) [22]. However, most genetic studies have been restricted to a single gene or pathway, and the comprehensive mechanism of retinoid dermatitis remains to be elucidated. Therefore, a “candidate gene” analysis of retinoid-induced irritation was conducted to screen for anti-irritants against retinoid dermatitis, and the results of the same will add to the current pool of target genes for studying retinoid dermatitis.

2. Materials and Methods 

2.1. Study Design

This study was designed as shown in Figure 1. Brieflfly, phenotyping for retinolinduced irritation from the 1st clinical evaluation of 173 Koreans was performed. Through genetic analysis, genetic markers were screened, and anti-irritants and formulas were investigated. Following, a small-scale pilot study (n = 7) and large-scale clinical evaluation (2nd, n = 91) were performed to verify the anti-irritant effificacy of the newly developed formula.

Figure1.Studydesign.Upperdiagram(flowchart)depictstheoverallprocedureofthestudy.Thelowertableindicatesthe mainthreehuman-basedtests.Retinol1IUrefersto3×10−5%w/wintheformulation.(2500IU=0.075%w/w;3300IU=0.1%w/w;5000IU=0.15%w/w).

2.2. Phenotyping of Retinol-Induced Irritation

were informed of the ingredients of the cream prior to the experiment, and the participants who had experienced skin irritation to any of the ingredients in the cream, except for retinol, were excluded from the experiment. The detailed procedure is described in the Supporting Information.

2.3. Single Nucleotide Polymorphism (SNP) Microarray

Genomic DNA was extracted from saliva samples using the QIAmp Mini Prep Kit (QIAGEN, Germantown, MD, USA). Genotyping was carried out using the Global Screening assay version 2 chip according to the Illumina Infifinium High-Throughput Screening (HTS) assay protocol (Illumina, San Diego, CA, USA). Bead intensity was obtained using the iScan®instrument and subsequently used to yield genotype data using GenomeStudio®software (Illumina, San Diego, CA, USA).

Quality control of the genotype data was preferentially performed under the following conditions: Among the genotyped samples, individuals with a genotype call rate lower than 98% were excluded. For genotyped variants, SNPs with call rates lower than 98%, minor allele frequencies lower than 1%, deviation from Hardy–Weinberg equilibrium (HWE p-value < 1 × 10−6), and those having more than two alleles were excluded. Among the 173 genotyped individuals, 159 passed the QC criteria and were selected for further analysis.

2.4. Assessment of Irritation

The two types of creams, namely the control cream with 5000 IU retinol, and the anti-irritant formula (AF)-based cream with 5000 IU retinol, were administered to seven individuals (three men and four women) in the pilot study. The individuals applied thecreams on their face before bedtime in the following manner: one cream on one half of the face and another on the other half of the face. The test subjects were blinded to the cream containing the anti-irritant formula (single-blinded). After three days of application, the subjects were asked to cease the application for 4 days. Skin redness and transepidermal water loss (TEWL) were measured using a chromameter (CR-400, Konica Minolta, Osaka, Japan) and Tewameter® (TM 300 E, Courage + Khazaka electronic GmbH, Köln, Germany), respectively. To measure the pain pressure threshold (PPT), an algometer with a probe diameter of 1 mm combined with a cylinder was manufactured and used. Before the measurement, the faces of the subjects were cleaned and acclimated for 20 min in an air-conditioned room (temperature, 23 ± 2◦C; relative humidity, 50 ± 10%).

Considering that retinoid-induced irritation triggers an extremely wide range of irritation types and severity, it appears that traditional irritation-measuring guidelines mainly focusing on erythema proposed by Frosh and CTFA guidelines [23] do not effectively qualify or quantify these irritations. Therefore, we re-designed the self-evaluation guideline for skin irritation, as shown in Table S1 (see Supplementary Materials). This self-evaluation was performed in a small pilot study and a 2nd large-scale clinical evaluation. 

2.5. Cell Preparation and In Vitro Experiment

Detailed information regarding cell preparation and in vitro experimental procedures is provided in the Supplementary Information. Brieflfly, keratinocytes, fifibroblasts, mast cells, and TRPV1-overexpressing HEKs were cultured using their own culture protocols and experiments. RT-PCR, β-hexosaminidase release, and calcium inflflux were analyzed. 2.6. Statistical Analysis Statistical analysis was performed using SNP and Variation Suit (SVS) v8.9.0 (Golden Helix, Bozeman, MT, USA), and PLINK 1.90 (Cambridge, MA, USA) [24]. For candidate gene analysis, SNPs within the candidate genes (including promoter region −2 kb and downstream region 500 bp) were extracted. The statistical signifificance of the associations and odds ratios were determined by a genotype association test with an additive model. Prior to the calculation of the polygenic risk score, linkage disequilibrium (LD) was examined, and SNPs were selected using Haploview 4.2 (Cambridge, MA, USA). [25]. Alleles that showed an increased tendency for retinol-induced irritation were used for the analysis. The polygenic risk score was calculated as the weighted sum of the odds ratios of each allele [26]. Input data preprocessing, generating a random subset of participants for repeated validation, and calculation of polygenic risk score were conducted using a custom-written code in R v.4.0.3. GraphPad Prism v7.04 (GraphPad Software, San Diego, CA, USA) was used for data visualization. Error bars whose values were derived by dividing each standard deviation by the square root of the number of samples are shown in the fifigures.

In the analysis of the occurrence rate of irritation based on the dichotomous question (whether retinol cream is irritant: Y/N), the chi-squared test was performed to investigate the statistical signifificance. 

3. Results and Discussion

3.1. Proposing Target Genes for Screening of Anti-Irritants to Retinol-Induced Irritation 

3.1.1. 1st Clinical Evaluation—Topical Application of Retinol and Analysis of Its Irritant Properties

We investigated the irritant properties of retinol and its relationship with skin sensitivity in 173 Korean individuals. They applied retinol for 3 days, which was followed by a rest of 4 days a week for 3 weeks, with each concentration of retinol gradually increasing every week (2500 IU in the 1st week, 3300 IU in the 2nd week, and 5000 IU in the 3rd week). Then, we analyzed the questionnaire received after the experiment to investigate the factors related to retinol-induced irritation. 

A higher proportion of participants in the sensitive skin group reported that they experienced irritation compared to the non-sensitive skin group (Figure 2a). It was also shown that individuals who belong to the sensitive skin group were about three times more likely to have past experience of stopping cosmetic product usage due to skin irritation (Figure 2b). Among those with past irritation experiences, most individuals replied that basic cosmetics triggered irritation, which was followed by sunblock, cleanser, and cosmeceutical products (Figure 2c). A higher proportion of individuals who had skin irritation of retinol products in the past answered that they had irritation in this experiment compared to those who had not experienced irritation when using retinol products in the past (Figure 2d). This result revealed that retinol-induced irritation tends to occur repeatedly depending on individuals, which supports the hypothesis that genetic factors might affect the sensitivity to retinol. Earlier studies have shown that genetic variationssignificantly affect the bioavailability of retinol and retinoid, which support that geneticfactors could also govern retinol-induced irritation [27,28].

The type of irritation induced by retinol usage varies greatly among individuals. How.ever, stinging was most common, which comprised about three-fourths of the irritationsand was followed by burning, itching, and erythema (Figure 2e).

(c) The proportion of individuals who experience irritation when using basic cosmetic products. (d) The proportion of individuals with a past experience of irritation to retinol-containing products. (e) The types of irritation that individuals have experienced during use of retinol-containing topical products; * p-value < 0.05; *** p-value < 0.001; n.s, not signifificant.

3.1.2. Candidate Gene Analysis for Retinol-Induced Irritation

To discover genetic variants associated with skin sensitivity to retinol, we chose 14 candidate genes that are well-known for their functions in retinoid metabolism and skin sensitivity (Table S2). We referred to previous transcriptomic analysis of the 3D reconstituted skin under treatment of sensitizers [29]. 

A total of 319 SNPs located within the candidate genes were selected and used for the association analysis using the additive model. Thirty SNPs were signifificantly associated with retinol-induced irritation, none of which have been previously reported (p < 0.05; Table 1). In detail, a total of 12 SNPs were found in RARB, three in EGFR, three in CD44, two in IL18, two in IL4R, and four in BCL2. The other four genes, CD86, RXRB, MMP10, and COL6A2, included single SNPs. The discovered SNPs belonged to 10 genes, two of which were retinol-related genes, and the remaining eight were related to general skin sensitivity according to earlier studies. Among the 10 genes, we selected the most important three genes in terms of skin irritation: COL6A2, EGFR, and IL-4R. 

COL6A2, which encodes one of the three alpha chains of type VI collagen found in most connective tissues, has been shown to regulate dermal matrix assembly and fifibroblast motility [30], and it contributes to tissue remodeling and wound healing [31,32]. This defect leads to keloid formation [33] and abnormal skin phenotypes [34]. Epidermal growth factor receptor (EGFR) is an important regulator of epidermal barrier function. EGFR signaling has been shown to inhibit the competence of the cornifified envelope and disrupt tight junction barrier function in epidermal keratinocytes [35,36]. In addition to EGFR signaling, another wide range of factors such as ADAM17 (a disintegrin and metalloproteinase 17) [37] and TRP channel [38] are also believed to affect the epidermal barrier function systematically associated with EGFR. 

In the past few decades, the functions and roles of interleukin 4 (IL-4) and its receptor, IL-4R, have been extensively investigated. As a key regulator in the humoral and adaptive immune system, IL-4, which is primarily secreted from mast cells and Th2 cells, induces the differentiation of Th2 cells and stimulates B cells. Although its role is not clearly understood, many studies have pointed out that IL-4 may drive extensive pro- and inflammation processes, while its defect was shown to drive allergic disease, Alzheimer’s disease, and tumors [39]. Along the same axis for the IL-4-related immune response, IL-4R is ubiquitously expressed on various immune cells in both the innate and adaptive immune systems. IL-4R is a common receptor for both IL-4 and IL-13 [40]. Unlike IL-4, which directly affects Th2 differentiation, IL-13 is also responsible for activating mast cells, which control eosinophil function, and it has immunosuppressive and anti-inflammatory effects on macrophages by suppressing pro-inflammatory cytokines and chemokines [41]. 

3.2. Top–Down Approach: Screening Anti-Irritants In Vitro 

3.2.1. Tissue Repair-Related Genes, COL6A2, and EGFR

Based on the selected genes, we screened for anti-irritants that could modulate molecular pathogenesis, such as controlling the expression of the target genes. 

Earlier reports revealed that retinoids induce physiological and morphological changes in the skin barrier [8,42,43]. It was observed that the treatment of skin tissue with retinoic acid ex vivo and in vivo elevates transepidermal water loss (TEWL), and several studies have revealed that this disruption of skin barrier function is mediated by an imbalanced expression pattern of the cornified envelope and tight junction-related genes. The downregulation of filaggrin (FLG), loricrin, and CLDN1, and upregulation of CLDN2, CLDN4, and a significantly different expression of serpin family member genes were observed in an earlier study, which is thought to be the main cause of retinoid-induced irritation [10,11]. \We hypothesized that individuals with genetic variations in COL6A2 and EGFR tend to be susceptible to the weakening of skin barrier function by retinol, which leads to retinol-induced irritation (Figure 3a). We also considered that apart from collagen VI, collagen IV is also an important factor in the epidermal basement membrane and wound-healing process, although its exact role in the skin tissue remains unclear [44]. We hypothesized that a defificiency or abnormality of collagen VI and IV leads to susceptibility to retinol-induced irritation, and the overexpression of both types of collagen could attenuate the retinolinduced weakening of the skin barrier. Among the various substances, we observed that glucosamine increased the expression of COL6A2 and COL4A2 by 1.5-fold and 1.7-fold, respectively (Figure 3a, middle panel). The modulatory role of glucosamine on COL6A2 and COL4A2 has not been reported before, which implies that these observations are scientififically important.

In the case of EGFR, we carefully considered previous research indicating that retinoic acid induces the overexpression of aquaporin 3 (AQP3) through the EGFR/ERK pathway in human keratinocytes [45]. In the first report, we observed that not only retinoic acid but also retinol induced the AQP3 expression by approximately two-fold in fibroblasts (Figure 3a right panel). The attenuation effect of AQP3 overexpression by niacinamide (nicotinamide) is similar to that found in an earlier study. Among the various substances, trehalose and eupatilin attenuated retinol-induced AQP3 overexpression. Eupatilin (5,7-dihydroxy-3’,4’,6-trimethoxyflavone), a type of flavonoid found in Artemisa asiatica, was reported to enhance the skin barrier function under pathological conditions in vivo [46]. 

Additionally, we investigated the expression of FLG, which is a prime indicator of skin barrier function. Similar to previous in vivo studies [47], we observed that retinol signifificantly decreased the expression of FLG by approximately 45%, and this decrease was reversed by niacinamide, glucosamine, and sucralfate (Figure S1a, see Supplementary Materials).

Collectively, glucosamine, trehalose, and sucralfate could alleviate retinol-induced skin barrier disruption by modulating the expression of COL6A2, AQP3, and FLG. 

3.2.2. Inflflammatory Gene: IL-4R 

Then, we tried to identify anti-irritants related to inflflammatory genes. Based on our results of genetic analysis, it has been shown that individuals who are prone to retinol-induced irritation tend to have SNP markers for IL-4R but not IL-4. First, we investigated whether treatment with retinol modulates the expression of IL-4 or IL-4R in mast cells, which is also a primary driver of IL-4-related inflflammation. Retinol induced the overexpression of IL-4R 1.74-fold. Although retinol did not induce the overexpression of IL-4 (Figure S1b, see Supplementary Materials), it was observed that ectoine effectively decreases the expression of IL-4 of the mast cell, which is coincident with earlier studies that argued the anti-inflflammatory effects of ectoine in some disease models such as IBD (inflflammatory bowel disease) and allergic airway disease [48,49]. More specififically, earlier studies showed that ectoine could normalize the expression of IL-4 in the CNP (carbon nanoparticle)-induced lung inflflammation in an in vivo model, supporting our experimental results [50].

Crocin, glucosamine, and ectoine alleviated the retinol-provoked overexpression of IL-4R in mast cells, with a decrease of 37.03%, 41.97%, and 82.59%, respectively. Ectoine (1,4,5,6-tetrahydro-2-methyl-4-pyrimidinecarboxylic acid), a natural compound found in several halophilic bacteria, was shown to modulate many cytokines and chemokines under inflflammatory conditions, although the role of ectoine in ameliorating the expression of IL-4R has not yet been elucidated. Our observations offer novel insights into the function of ectoine. Previous study by our group showed that ectoin at an optimum concentration of 100 ppm showed the highest anti-inflflammatory effect against the effects of retinoids and negligible cytotoxicity in the cell-based in vitro experiment (data not shown). We also measured the β-hexosaminidase release rate of mast cells, which is an indicator of mast cell degranulation (Figure 3b, right panel). Ectoine did not reduce retinol-induced β-hexosaminidase release in the 400 ppm retinol, although it reduced retinol-induced β-hexosaminidase release in the 200 ppm retinol. Collectively, it can be concluded that ectoine can ameliorate IL-4 and IL-4R–related inflflammation induced by retinol. Considering the previous research that ectoine intervenes in the activation of EGFR in CNP-induced lung inflflammation [50], ectoine also appears to be benefificial for skin barrier disruption induced by retinol. 

3.2.3. Neurogenic Inflflammation, Adiponectin, and TRPV1

One distinguishing feature of retinol-induced irritation in extremely retinoid-sensitive patients is allergy-like reactions such as rapid burning and sting sensation, itchiness, rapid diffusive edema, and rash. A study revealed that retinoids, including retinol and retinoic acid, activate TRPV1 [21]. Therefore, we hypothesize that retinol-induced irritation, especially promptly occurring within a few minutes, is mediated by the activation of TRPV1 by retinol.

Figure 3. The analysis for mRNA expression to screen anti-irritants that could modulate irritation-associated molecularpathogenesis. (a) Investigation for skin barrier disruption-associated molecular pathogenesis Schematic diagram for theapproach for genetic modulation for retinol (ROL-induced skin barrier disruplion. (left) The relative mRNA expression of COL4A2 and COL6A2 when fifibroblasts were treated with glucosamine (middle) and relative mRNA expression of AQP3 (keratinocyte, right). (b) Investigation for inflflammation (mast cell driven)-associated molecular pathogenesis. Relative mRNA expression of IL-4R when RBL-2H3 was treated with 10 µM retinol and various candidates. (c) Neurogenic inflflammation mediated with TRPV1 induced by retinol and antagonistic effect by 4-t-butylcyclohexanol and omega-9; Three conditions for combination: (1) 4-t-butylcyclohexanol 25 µM and omega-9 5 µM, (2) 4-t-butylcyclohexanol 50 µM and omega-9 10 µM, and (3) 4-t-butylcyclohexanol 100 µM and omega-9 20 µM, BC, 4-t-butylcyclohexanol; OA, omega-9 oleic acid; A 50 µm scale bar (white) was shown; * p < 0.05; n.s, not signifificant; error bar was shown; ppm, parts per million. 

TRPV1, a nonselective cation activated by various physicochemical stimuli, has been widely studied, especially its role in the skin. [51–54] Although TRPV1 did not emerge in the previous genetic analysis, it should be noted that adiponectin (ADIPQ) appeared in some analysis models with low to moderate statistical signifificance. An earlier study of the Korean population showed that the “sensitive” skin phenotype appears to be related to adiponectin defificiency, which consequently contributes to the upregulation of TRPV1 [55]. TRPV1-overexpressing HEK293 cells were treated with retinol, and the activation of TRPV1 was confifirmed by imaging calcium inflflux. We observed that retinol could activate TRPV1 in a dose-dependent manner (Figure S1c, see Supplementary Materials). At concentrations above 50 µM, the agonistic effect on TRPV1 seemed to reach a plateau. Based on a previous study, we investigated whether 4-t-butylcyclohexanol and omega-9 oleic acid can antagonize TRPV1 activation induced by retinol. As expected, both substances can antagonize TRPV1, while at low dosage, omega-9 oleic acid was more potent than 4-t-butylcyclohexanol (Figure S1d, see Supplementary Materials). The combination of these two materials showed a slight synergistic effect on the inhibition of TRPV1 by retinol. Three different concentrations of the combination of 4-t-butylcyclohexanol and omega-9 showed reduced activation of TRPV1 induced by retinol, 28.25%, 43.73%, and 68.50%, respectively. Calcium inflflux into cells mediated by the activation of TRPV1 was observed under flfluorescent microscopy (Figure 3c). In conclusion, we verifified that retinol-induced irritation could also be related to neurogenic inflflammation mediated by TRPV1 activation in an in vitro model, and 4-tbutylcyclohexanol and omega-9 oleic acid could be utilized to mitigate neurogenic inflflammation by antagonizing TRPV1.

Email:wallence.suen@wecistanche.com  Whatsapp +86 15292862950