Aseptic samples of healthy skin from surgical patients were transferred to a tube with sterile saline at 4°C. All patients signed informed consent forms with full knowledge and donated their skin samples voluntarily. Prior approval of the local ethics committee was obtained (ChiCTR1800019082). This study was conducted in accordance with the declaration of Helsinki Principles. CD49f-PE, CD117-BV421, CD146-BV510, CD45-APC-Cy7, CD34-PE-Cy7 and CD29-APC were purchased from Becton Dickinson (San Jose, CA, USA). CD24-BV605 was from BioLegend (San Jose, CA, USA), and the TLR7 antibody (rabbit anti-human, sc-16245) was from Santa Cruz (Santa Cruz, CA, USA); it was used followed by donkey anti-rabbit IgG secondary antibody-Alexa Fluor 488 (Cat number: A-21206 Invitrogen, USA). Each antibody was titrated by serial dilutions. The optimized dilution of all antibodies was listed in Table1. In total, 16 healthy donors (8 males and 8 females) were included in this study with a mean age at 26 (shown in Table 2).
Isolation of human keratinocytes
The subcutaneous tissue was cut with scissors, after which the skin sample was washed twice with PBS. The sample was cut into fragments and transferred to 0.25% dispase II (neutral protease, grade II, Cat# 04942078001, Roche, Mannheim, Germany) and incubated for 1.5 hours at 37°C. The epidermis was gently separated with forceps and cut into small pieces. After digestion with 0.125% trypsin (HyClone, Cat No: SH30042.01, GE Healthcare Life Sciences, Sweden) for 8 minutes at room temperature, the trypsin was neutralized with 20% fetal bovine serum (FBS). The epidermal cells were prepared for staining after filtration with a 70 µm cell strainer, centrifugation at 350 g for 6 minutes and discarding of the supernatant.
Cell staining and flow cytometric analysis
The epidermal cells were resuspended gently, washed with PBS, and then centrifuged at 350 g for 6 minutes. After resuspension, antibody cocktail containing 2.0 μl of each antibody (CD49f, CD117, CD146, CD45, TLR7, CD24, CD34 and CD29) was added to 2 million cells with 100 µl of staining buffer. The cell suspension was incubated on ice for 30 minutes in the dark. The cell suspension was washed twice with PBS, and donkey anti-rabbit IgG secondary antibody-Alexa Fluor 488 (2.0 µl) was added, followed by incubation for 20 minutes in the dark. After 1.0 µl PI was added for another 5-minute incubation, the cell suspension was washed with PBS and centrifuged at 350 g for 6 minutes, and the cells were resuspended gently with PBS. The cell suspension was analyzed on a BD FACSAria SORP flow cytometer immediately.
Data acquisition and analysis pipeline
Acquisition and analysis were performed on a FACSAria SORP cytometer equipped with DivaV6.0 software (Becton Dickinson, San Jose, CA, USA). The instrument setup was standardized to reduce batch-to-batch shifting by daily monitoring with Rainbow beads (Spherotech). The boundary between positive and negative events was placed by fluorescence-minus-one controls. The maximum possible number of events was acquired (at least 500,000 events and preferably more). Data analysis was conducted using Cytobank (Mountain View, CA) and the FlowJo software program (TreeStar, Ashland, OR). In the analysis, a sequential gating strategy was used (Fig. 1). After excluding debris, dead cells and CD45-positive cells, data files of living epidermal cells were concatenated by group and uploaded into Cytobank. ViSNE analyzed 10,000 cells from each sample randomly. The dimensional reduction was visualized on axes identified by tSNE1 and tSNE2. Dimensional reduction and visualization of data files was performed with viSNE (viSNE setting: 10,000/sample, iterations: 7000, perplexity: 50, seed: 94138845) followed by SPADE on Cytobank or PhenoGraph (10,000/sample, iterations: 7000, perplexity: 50, seed: 42, k: 45) clustering on R. The SPADE analysis settings were as follows: target number of nodes=25 and percentage downsampling=100%. The intensity and cellular abundance of each node from each individual were exported for further analysis. Four categories: high (hi), medium (mi), low (lo), and negative (neg) was divided according to the total expression distribution of cells in each marker. The mean of the median marker expression of the cells contained in each node was then used to assign the expression of each marker to one of the four categories [45].Statistical data analysis was performed in Prism 8.2.1 (GraphPad Software Inc., La Jolla, CA, USA) and represented as the mean±SEM. Two-way ANOVA and Student’s t-tests were used to compare data among the ear, thorax and abdomen. A P value<0.05 was considered significant.