The present study received approval from the Ethics Committee of the Affiliated Hospital of Southwest Medical University (approval no. KY2022206; Luzhou, China). Prior to participation, all patients (5 males and 2 females; age range, 46-61 years) and volunteers (1 male and 3 females; age range, 37-62 years) provided written informed consent either personally or through their legal representatives. Between May 2022 and May 2023, the Affiliated Hospital of Southwest Medical University gathered seven samples of skin wound tissues from individuals with chronic skin infection, along with four samples of healthy skin tissue from volunteers, which were preserved at -80˚C for RNA sequencing. Additionally, another set of specimens, also comprising four cases in the healthy group and seven cases in the chronic infection group, were washed with pre-cooled PBS and fixed in 4% paraformaldehyde for immunohistochemical detection. Patients with chronic skin infections included post-traumatic infections and diabetic foot infections.

RNA sequencing was performed on tissue samples from chronic skin infection wounds and healthy individuals using the Illumina Novaseq 6000 sequencing platform (Illumina, Inc.). The Illumina Truseq RNA sample prep Kit (Illumina, Inc.) was used to generate the library and sequence all transcribed mRNA, involving the following six steps: Total RNA extraction; mRNA enrichment with Oligo dT; mRNA fragmentation; cDNA reverse synthesis; adapter ligation; and Illumina sequencing. The specific sequencing process entailed during the extraction was as follows: i) Total RNA from tissue samples, assessment of RNA concentration and purity using Nanodrop 2000 (NanoDrop Technologies; Thermo Fisher Scientific, Inc.), determination of RNA integrity through agarose gel electrophoresis and calculation of RNA integrity number using Agilent 2100 (Agilent Technologies, Inc.); ii) mRNAs were isolated from total RNA by A-T base pairing with polyA using oligo(dT) beads (500 times; cat. no. YH-RCZ-05; Shanghai Majorbio Pharmaceutical Technology Co., Ltd.) to analyze the transcriptomic information, based on the presence of a polyA tail structure at the 3' end of eukaryotic mRNA; iii) the mRNA was fragmented using a fragmentation buffer [DNA Purification and Fragment Screening Kit (magnetic bead method); cat. no. C03-050; Shanghai Meiji Zhuanghua Biopharmaceutical Technology Co., Ltd.], resulting in the isolation of small fragments ~300 bp in length through magnetic bead screening for sequencing; iv) A six-base random primer (random hexamers) is introduced during reverse transcription to generate single-stranded cDNA from mRNA as a template, followed by two-stranded synthesis to establish a stable double-stranded structure; v) end repair mix was added to the double-stranded cDNA structures to homogenize cohesive ends, followed by the addition of an ‘A’ base at the 3' end for joining Y-shaped joints; vi) the cDNAs obtained through PCR amplifications using Phusion DNA polymerase (NEB) for 15 cycles according to the manufacturer's instructions; recovered with 2% agarose gel, quantified using TBS380 (PicoGreen; NovaSeq Reagent Kit; Illumina, Inc.) and run in a data ratio; and vii) bridge PCR was then amplified on cBot to generate clusters, which were subsequently sequenced on the Illumina platform (Illumina, Inc.) using a PE library with a reading length of 2x150 bp.

The sequencing data, which were subjected to quality control measures [log2(CPM+4)], were analyzed using the online software idep.96 (http://bioinformatics.sdstate.edu/idep/) (10) and a network tool (http://bioinformatics.sdstate.edu/idep/) based on the R language (http://www.R-project.org/). Data quality was standardized by applying log₂ (CPM)+4(11), before the DESeq2 differential screening method [log₂ fold-change (FC)≥1 and q-value ≤0.05] (12) was used to identify DEGs between the healthy skin group and the chronically infected skin group.

GO analysis is a typical batch gene data analysis method. The GO examination of genes involved the analysis of biological process (BP), cellular component (CC) and molecular function (MF) aspects, was performed for categorizing and identifying groups of genes in subsequent stages (13). KEGG is a resource that systematically examines gene signaling pathways and connects genomic and functional data (14). The chosen DEGs were assessed using the database to study GO functions and KEGG pathways, with the results being visualized through online tools utilizing R language (https://cloud.oebiotech.com/spa#/bio/detail?number=e267d1a2-4303-44d7-a6db-d2bd2ad59e3e).

Core genes were identified through the construction of a PPI network utilizing the STRING database (https://cn.string-db.org/) (15) to understand how genes interacted at the protein level. Potential key targets were identified through this analysis. DEGs were then filtered based on their node degree and network, with a threshold comprehensive score of >0.4. The present study focused on the functional interactions among target genes and other genes associated with adapting to hypoxia, angiogenesis, cell adhesion and cell migration.

The present study adhered to the typical protocols of immunohistochemistry, including collecting materials, preparing sections, conducting antigen-antibody reactions and developing colors to identify the presence of HIF1A expression in healthy skin samples from donors and in wound tissues from patients with chronic skin infections. The specific procedures for immunohistochemistry were as follows: i) Skin tissues were fixed in 4% paraformaldehyde, embedded in paraffin wax and cut into 5-mm slices. The slices were then pressed onto anti-slip plates and placed on a 60˚C microtome for 2 h. ii) The sections were incubated twice in xylene and dehydrated in gradient ethanol solutions. iii) After microwave treatment in citrate buffer (pH 6.0), the slides were incubated in 3% hydrogen peroxide for 15 min to inhibit endogenous peroxidase activity. They were then blocked with 10% normal goat serum (cat. no. A0208; Beyotime Institute of Biotechnology) and incubated at room temperature for 10 min. iv) Sections were treated with primary antibodies (HIF-1α polyclonal antibody; 1:70; cat. no. 20960-1-AP; Proteintech Group, Inc.) as per instructions and incubated overnight at 4˚C. After rewarming for 30 min at room temperature, they were incubated with HRP-labeled secondary antibody [HRP-labeled goat anti-rabbit IgG (HL); 1:70; cat. no. A0208; Beyotime Institute of Biotechnology) for 60 min. Specimens were then washed three times in PBS for 5 min each. v) Sections were colored with DAB for 1-10 min, then washed twice in distilled water for 5 min each. Hematoxylin counterstaining was performed for 5 min, followed by dehydration with gradient alcohol and clearing with xylene twice for 5 min each. A neutral gummy film was applied for capturing images. Images were viewed using an inverted light microscope and captured at a magnification of x100 and x200. ImageJ (version 1.8.0_172) processing software (National Institutes of Health) was used for the semi-quantitative analysis of immunohistochemical staining and to determine the average optical density (AOD) value.

Statistical analysis of experimental data was conducted using SPSS 26.0 software (IBM Corp). The experimental results consist of measurements that follow a normal distribution and are represented by the mean ± standard deviation. Unpaired t-tests were used for comparing between the two means.

The gene data set obtained through RNA-seq underwent logarithmic processing and was subjected to quality control using the online analysis software idep.96 (http://bioinformatics.sdstate.edu/idep/), which primarily involved generating box plots for two datasets (Fig. 1), data density plot (Fig. 2) and principal component analysis plot (Fig. 3). The analysis results indicated that the data were comparable and the quality control of the data was satisfactory.

After the initial dataset was preprocessed using logarithmic transformation, it was analyzed using the online tool idep 0.96 (http://bioinformatics.sdstate.edu/idep/) with the criteria of FC>4 and P<0.05. In both the chronic infection group and healthy group, 900 DEGs were found, with 365 genes being upregulated and 565 genes being downregulated. The DEGs were found to be evenly distributed between the groups. The details of DEGs are shown in the bar chart (Fig. 4) and volcano chart (Fig. 5).

After screening the 900 DEGs, an analysis was conducted on their GO functions and KEGG pathways using a database, before the results were visualized with the R language-based oebiotech online tool. The examination revealed that the DEGs primarily participate in biological activities such as ‘response to hypoxia’, ‘angiogenesis’ and ‘cell adhesion’ (Fig. 6). Additionally, they served a role in various signaling pathways, including TGF-β, PI3K-Akt and IL-17 (Fig. 7).

A total of 1,104 DEGs were used to create the PPI network. From this network, 10 genes were identified as being central to the network, involved in various biological processes such as angiogenesis, protein modification, intracellular signal transduction, cell biosynthetic processes, catabolism regulation, organ morphogenesis and movement regulation (Fig. 8).

Healthy skin tissues exhibit distinct layers, such as the epidermis, containing the stratum corneum, stratum lucidum, stratum granulosum, stratum spinosum and basal cell sublayers, in addition to the dermis with its papillary and reticular layers. Immunohistochemistry was conducted on samples obtained from four individuals in the healthy skin group and seven individuals in the chronic skin infection group to assess the presence of HIF1A expression. In the chronic infection group, observations revealed thickening of the epidermal cuticle, destruction of the basal layer, proliferation of fibrous tissue and disorganized arrangement of muscle fibers. Analysis indicated the expression of HIF1A in the dermis. Semi-quantitative analysis of HIF1A was performed using ImageJ (version 1.8.0_172) software (National Institutes of Health) to determine the AOD value as a measure of HIF1A expression. The mean ± standard deviation was used to describe the experimental data. The healthy group had an AOD value of 0.207828±0.007539, whilst the chronic infection group had a value of 0.221970±0.009962. The data from the experiment followed a normal distribution. A comparison between groups was conducted using an unpaired t-test, revealing a statistical difference (P<0.05; Fig. 9). These findings suggest that HIF1A was highly expressed in the chronic infection group.