Abstract Propose To investigate the molecular mechanisms underlying the protective effects of ischemic preconditioning (IPC) in patients undergoing total knee arthroplasty. Methods [29]GSE21164 was extracted from an online database, followed by an investigation of differentially expressed genes (DEGs) between IPC treatment samples at 2 time points (T0T and T1T). Function and pathway enrichment analyses were performed on the DEGs. A protein-protein interaction network was constructed to identify hub genes according to 5 different algorithms, followed by enrichment analysis. In addition, long noncoding RNAs (lncRNAs) were identified between the T0T and T1T samples. Furthermore, a competing endogenous RNA network was predicted based on the identified lncRNA-messenger RNA (mRNA), lncRNA-microRNA (miRNA), and mRNA-miRNA relationships revealed in this study. Finally, a drug-gene network was investigated. Statistical analyses were performed using GraphPad Prism 8.0. Differences between groups were determined using an unpaired t-test. p < 0.05 was considered significant. Results A total of 343 DEGs at T0 and 10 DEGs at T1 were identified and compared with their respective control groups, followed by 100 DEGs between T0T and T1T. Based on these 100 DEGs, protein-protein interaction network analysis revealed 9 hub genes, mainly with mitochondria-related functions and the carbon metabolism pathway. Six differentially expressed lncRNAs were investigated between T0T and T1T. A competing endogenous RNA network was constructed using 259 lncRNA–miRNA–mRNA interactions, including alpha-2-macroglobulin antisense RNA 1-miR-7161-5p-iron-sulfur cluster scaffold. Finally, 13 chemical drugs associated with the hub genes were explored. Conclusion Iron-sulfur cluster scaffold may promote IPC-induced ischemic tolerance mediated by alpha-2-macroglobulin antisense RNA 1-miR-7161-5p axis. Moreover, IPC may induce a protective response after total knee arthroplasty via mitochondria-related functions and the carbon metabolism pathway, which should be further validated in the near future. Keywords: Ischemic preconditioning, Total knee arthroplasty, Differentially expressed gene, Function and pathway enrichment analysis, ceRNA network Graphical abstract [30]Image 1 [31]Open in a new tab 1. Introduction Total knee arthroplasty (TKA) is considered the most valuable and widely used method for the treatment of serious knee diseases and can effectively relieve pain and improve the patients’ quality of life.[32]^1 More than 650,000 TKAs are performed annually in the United States.[33]^2 Tourniquets are commonly used in total knee replacement (TKR) to provide better surgical vision and reduce intraoperative bleeding. However, the application of tourniquets not only leads to mechanical compression, but also induces ischemia-reperfusion (IR) injury associated with skeletal muscle and other organs, which finally causes quadriceps femoris atrophy, decreased activity function, and prolonged rehabilitation time after operation.[34]^3^,[35]^4 Ischemic preconditioning (IPC) is a process in which a tissue undergoes one or more transient injury cycles, generating a small amount of oxygen free radicals that cause an adaptive response, thereby reducing the damage to the tissue caused by subsequent long-term injury.[36]^5 IPC is commonly performed in several surgical settings including TKA. Memtsoudis et al.[37]^6^,[38]^7 suggested that IPC reduces the subsequent inflammatory response and protects tissues from IR after TKA. The protective function of IPC has been demonstrated in animal models and clinical trials.[39]^8^,[40]^9 A previous randomized controlled trial suggested that IPC could increase cerebral oxygenation after tourniquet release during TKR by improving pulmonary oxygenation.[41]^10 Recent evidence has shown that IPC exerts a protective effect on muscle strength in tourniquet-induced IR by upregulating mitofusin 2.[42]^11 Ischemia triggered by tourniquets leads to significant variations in gene expression within cells, including metabolism, genetic information processing, and other cellular processes.[43]^12 Murphy et al.[44]^13 developed an expression dataset ([45]GSE21164) from patients undergoing TKA with and without IPC treatment and identified 257 genes with differential expression (>1.5 fold change (FC)) in the IPC group at the onset of surgery (T0), and 786 genes with differential expression >1.5 fold 1 h after surgery (T1). These results suggest that IPC induces changes in gene expression in the protective response of patients undergoing TKA. However, little is known about the systemic genomic response including the changes in long noncoding RNA (lncRNA) levels induced by IPC during TKA. Therefore, in the current study, we re-analyzed the [46]GSE21164 dataset to obtain the differentially expressed genes (DEGs) induced by IPC treatment between 2 time points in TKA based on a cut-off value of p < 0.05, and |log FC| > 0.5. The results suggested that 100 genes were differentially expressed between T0 and T1 and between the IPC and control groups at a single time point. These genes were used for further protein-protein interaction (PPI) network and hub gene analyses. In addition, differential lncRNAs were identified between the 2 time points of IPC treatment, followed by hub gene-associated microRNA (miRNA) prediction. Finally, a competing endogenous RNA (ceRNA) network was constructed, followed by hub gene-drug interaction analysis. We aimed to elucidate the detailed molecular mechanism underlying the protective effect of IPC against TKA. 2. Methods 2.1. Data resource Gene expression profile dataset [47]GSE21164[48]^13 were downloaded from the Gene Expression Omnibus database ([49]http://www.ncbi.nlm.nih.gov/) based on the [50]GPL570 Affymetrix Human Genome U133 Plus 2.0 Array. In this dataset, patients in the IPC group received IPC stimulation before TKR (consisting of three 5-min tourniquet inflations of the lower extremity and a 5-min reperfusion). This dataset compared the gene expression of the quadriceps muscle biopsies of the operated knees in the IPC and control groups at the onset of surgery (T0) and 1 h after surgery (T1). Briefly, a total of 4 treatment samples (T0T) and 4 control samples at T0 (T0C), as well as 4 treatment samples (T1T) and 4 control samples at T1 (T1C), were used for subsequent analysis. 2.2. Differential expression analysis To explore the T0-DEGs (T0T vs. T0C) and T1-DEGs (T1T vs. T1C), the t-test methods in Linear Models for Microarray Data (limma, [51]http://www.bioconductor.org/packages/release/bioc/html/limma.html) package in R were used to calculate the significant p values of these genes.[52]^14 The p < 0.05 and |logFC| > 0.5 were selected as the thresholds for screening. Union-genes (un-genes) between T0-DEGs and T1-DEGs were further revealed using Jvenn software ([53]http://jvenn.toulouse.inra.fr/app/example.html).[54]^15 The results were visualized using Venn plots. In addition, with p < 0.05 and logFC >0.5, the DEGs between T1 and T0 samples obtained from un-genes were further evaluated using the limma package in R.[55]^14 Finally, the results of these common differentially expressed genes (co-DEGs) were visualized using volcano plots and heatmaps. 2.3. Enrichment analysis of co-DEGs Gene Ontology (GO; [56]http://www.geneontology.org/)[57]^16 function enrichment analysis, and Kyoto Encyclopedia of Genes and Genomes (KEGG, [58]http://www.genome.jp/kegg/pathway.html)[59]^17 pathway enrichment analysis of un-DEGs was performed using the clusterProfiler package in R.[60]^18 The GO functional categories included molecular function (MF), biological processes (BP), and cellular components (CC). The p < 0.05 and count (the gene number) ≥ 2 were considered as cut-off values of significant difference. The results were visualized using the ggplot2 package in R.[61]^19 2.4. PPI network construction According to the information in the STRING database (version: 11.0, [62]https://version-11-0.string-db.org/),[63]^20 the proteins associated with DEGs were selected using the criterion of score (medium confidence) = 0.4, and the PPI network was constructed using Cytoscape (version: 3.8.0, [64]http://www.cytoscape.org/) software.[65]^21 2.5. Hub gene investigation in PPI networks Hub genes in the PPI network were further investigated using the Cytoscape plugin CytoHubba[66]^22 to rank the nodes based on their network features. CytoHubba provides 11 topological analysis methods, including degree, edge-percolated component, maximum neighborhood centrality, density of maximum neighborhood component, maximal clique centrality, and 6 centralities (bottleneck, ecCentricity, closeness, radiality, betweenness, and stress) based on the shortest paths. The 11 topological analysis methods are summarized in [67]Table 1. Hub genes were identified using 11 topological analysis methods as previously described.[68]^23^,[69]^24 The top 20 genes were screened based on the ranking of each algorithm. Based on the UpSetR package in R,[70]^25 the hub genes were investigated by intersecting the top 20 genes of each algorithm. Table 1. Algorithms for important nodes evaluation. Centrality Options MCC The algorithm compares the paths between nodes and other nodes to evaluate the importance of paths, Nodes with low path importance are considered to be relatively important DMNC The algorithm measures the density of the maximum clique, that is, the number of connections between nodes in the maximum clique MNC It measures a subgraph with the maximum number of connected nodes. A subgraph with the maximum number of connected nodes Degree Degree means the number of neighbors of vertex. Nodes with highest degree were considered significant. EPC The algorithm represents the edge penetration component of a node in the network. The larger the component value, the more important the node is. Bottleneck The importance of the node is measured by the shortest path. If a node has the lowest degree among all the nodes in the paths, the node is considered significant. Eccentricity The eccentricity of the node means the longest path length between the node and other nodes. Nodes with small eccentricity are considered to be relatively important nodes. Closeness The algorithm measures the average distance between nodes and other nodes. Nodes with smaller average distances are considered more important in the network. Radiality The algorithm measures the distance between any two nodes. Nodes with maximum distance between any 2 nodes were considered significant. Betweenness The algorithm is based on the mediation of nodes on the shortest path. The higher the intermediation of a node, the more importance of the node in the process of information transmission in the network Stress It evaluates the number of all shortest paths on the network passing through this node. [71]Open in a new tab MCC: maximal clique centrality; DMNC: density of maximum neighborhood component; MNC: maximum neighborhood centrality; EPC: edge percolated component. 2.6. Enrichment and PPI network analyses based on hub genes The GO function and KEGG pathway enrichment analyses were performed on hub genes using clusterProfiler package in R.[72]^18 Statistical significance was set at p < 0.05. Moreover, a PPI network was constructed based on hub genes using the GeneMANIA ([73]http://genemania.org/) database, which includes different bioinformatics methods, such as co-expression and co-localization.[74]^26 2.7. ceRNA network investigation First, differentially expressed lncRNAs (DE-lncRNAs) were compared between T1T and T0T in [75]GSE21164 using the limma package in R.[76]^14 The results were visualized using volcano plots and heatmaps. Then, using miRWalk3.0 software (parameters: binding probability = 0.95; binding site position = 3UTR),[77]^27 the messenger RNA (mRNA)-miRNA interactions were revealed, followed by lncRNA-miRNA interactions determined using the lcnbaseV2 database ([78]http://carolina.imis.athena-innovation.gr/diana_tools) with score >0.85. Moreover, the lncRNA-mRNA pairs with Benjamini & Hochberg adjusted p < 0.05 and r (Pearson correlation coefficient between DE-lncRNA and hub genes) > 0.5 were enrolled for further investigation. Finally, combined with the positive co-expression relationship between mRNA and lncRNA (p > 0.7), the lncRNA-miRNA-mRNA ceRNA network relation regulated by the same miRNA was screened, and the ceRNA network was visualized using Cytoscape software.[79]^28 2.8. Drug-gene interaction prediction The drugs directly related to disease were explored using the Comparative Toxicogenomics Database (CTD) (keyword: each hub gene) with an inference score > 5.[80]^29 A drug-gene interaction network was constructed using the Cytoscape software.[81]^28 2.9. Tissue collection and quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) Three patients who underwent TKA with IPC were included in our study, which was approved by our hospital's Ethics Committee. The study protocol was conducted following the principles of the Declaration of Helsinki. The muscle tissues were extracted from each patient at the T0 and T1 time points during TKA and immediately stored at -80 °C for further use. For qRT-PCR, total RNA was extracted from tissue samples using a RNeasy RNA isolation kit (Qiagen). The real-time reaction was performed using a QuantiTect RT kit (Qiagen, Heidelberg, Germany), and the PCR assay was performed using the CFX96 Bio-Rad system. The reaction conditions were as follows: 95 °C for 2 min, followed by 36 cycles of 94 °C for 15 s, 58 °C for 20 s, and 72 °C for 15 s. Glyceraldehyde 3-phosphate dehydrogenase was set as the internal reference. The 2^−ΔΔCT method was used for quantitative analysis. The specific primers for the 9 hub genes are summarized in [82]Table 2. Table 2. Primer sequences used in RT-PCR. Gene Forward (5′-3′) Reverse (5′-3′) NDUFA9 TTGGTATTCAGGCCACACCC GCTGGCTTCACGTCTTCAAC GOT2 CGTCCGCAAGTTTGTCACTG GGCAGAAAGACATCTCGGCT MRPL43 CTCAAGCTTCGGATCCGGCCACAGCTCTCCAGTCG GGCGACCGGTGAATTCTAGCCGTGACTTCGGAAG MRPL46 GGCGCTCGAGTCGACCGGTGTTAACGGCCACAGCT CGGATCTCAAGCTTAGCGGTCCACCGTGACTTAG MRPL9 GTTACCAGAAGAGCCTATCACAC CTCTCACAGTATCAAGCCCATTT IDH3B CTGATGCACGCCGTCAAG GCCATATTCTGCACCTCACTCA HMGCL ACCACCAGCTTTGTGTCTCC GAGGCAGCTCCAAAGATGAC ISCU CCAGGTGGATGAAAAGGGGAA GCAGAGTTCCTTGGCGATGT IMP3 TCGGAAACCTCAGCGAGAAC ACTATCCAGCACCTCCCACT GAPDH CCACTAGGCGCTCACTGTTCT GCATCGCCCCACTTGATTTT [83]Open in a new tab qRT-PCR: quantitative real-time reverse transcription polymerase chain reaction; NDUFA9: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 9; GOT2: glutamic oxaloacetic transaminase 2; MRPL43: mitochondrial ribosomal protein L43; MRPL46: mitochondrial ribosomal protein L46; MRPL9: mitochondrial ribosomal protein L9; IDH3B: isocitrate dehydrogenase 3b; HMGCL: 3-hydroxy-3-methylglutaryl-CoA lyase; ISCU: iron-sulfur cluster scaffold; IMP3: insulin-like growth factor II mRNA-binding protein 3; GAPDH: glyceraldehyde 3-phosphate dehydrogenase. 2.10. Western blot analysis Total protein samples were isolated from muscle tissues collected from 3 patients at T0 and T1 by using RIPA solution. The quantified proteins were resolved by SDS-PAGE system and electroplated onto the PVDF membranes. After sealed with 5% bovine serum albumin, the membranes were incubated with the primary antibodies and secondary antibodies separately. The primary antibodies used contained NDUFA9 rabbit polyclonal antibody (Shanghai Zeye), GOT2 mouse monoclonal antibody (Invitrogen), MRPL43 rabbit polyclonal antibody (Invitrogen), MRPL46 rabbit polyclonal antibody (Invitrogen), IDH3B rabbit polyclonal antibody (Invitrogen), HMGCL rabbit polyclonal antibody (Invitrogen), iron-sulfur cluster scaffold (ISCU) mouse monoclonal antibody (Invitrogen), IMP3 rabbit monoclonal antibody (Abcam), MRPL9 rabbit polyclonal antibody (Invitrogen). Antibodies were diluted at 1:1000 for use. As secondary antibodies, we used HRP labelled goat anti-rabbit or anti-mouse IgG antibodies (1:5000, Invitrogen). Protein bands were visualized by chemiluminescent reagent and quantitatively analyzed by Image J software. The same experiment was performed in triplicate. 2.11. Statistical analysis Statistical analyses were performed using GraphPad Prism 8.0. Differences between groups were determined using an unpaired t-test. p < 0.05 was considered significant. 3. Results 3.1. DEG investigation With p < 0.05 and logFC > 0.5, 343 and 10 upregulated DEGs were revealed in T0T vs. T0C ([84]Fig. 1A) and T1T vs. T1C ([85]Fig. 1B), respectively. Heatmap analysis of DEGs in T0-DEGs and T1-DEGs showed that all samples could be separated by group at both T0 ([86]Fig. 1C) and T1 ([87]Fig. 1D). The DEGs in T0T vs. T0C and T1T vs. T1C were intersected to explore the un-genes using the Jvenn software. The results of the VENN plot analysis showed that 347 un-genes were obtained from T0T vs. T0C and T1T vs. T1C ([88]Fig. 1E). Fig. 1. [89]Fig. 1 [90]Open in a new tab The differentially expressed genes investigation. (A) The volcano plot explored the DEGs between treatment group (T0T) and control group (T0C) at the beginning of surgery (T0). (B) The volcano plot explored the DEGs between treatment group (T1T) and control group (T1C) at 1 h after surgery (T1). (C) The heatmap analysis for DEGs in T0 showed that all samples could be separated by group. (D) The heatmap analysis for DEGs in T1 showed that all samples could be separated by group in T1. (E) The Venn plot revealed 6 the common-genes (co-genes) in T0 and T1. DEGs: differentially expressed genes. 3.2. DEGs between T0T and T1T With p < 0.05 and log2FC > 0.5, 100 co-DEGs, including 1 upregulated and 99 downregulated, were revealed between T0T and T1T samples ([91]Supplementary Fig. 1A). Heatmap analysis of the DEGs between T0T and T1T showed that all samples could be separated into 2 groups ([92]Supplementary Fig. 1B). 3.3. Enrichment and PPI network analyses of co-DEGs Enrichment analysis showed that the DEGs were mainly enriched in 69 GO-BP, 15 GO-CC, 16 GO-MF, and 23 KEGG pathways. According to the p values, the top 10 GO functions and KEGG pathways are shown in [93]Supplementary Figs. 2A–2B. Moreover, with a score > 0.4, a PPI network was constructed with co-DEGs ([94]Fig. 2A). The results showed that there were 55 nodes and 68 interactions in the network. The top 20 genes were then screened using 5 methods of CytoHubba topological analysis, including maximal clique centrality, density of maximum neighborhood component, maximum neighborhood centrality, degree, and edge-percolated component based on 55 nodes in the PPI network ([95]Supplementary Fig. 3). Finally, 9 hub genes, MRPL9, IDH3B, MRPL46, MRPL43, IMP3, HMGCL, ISCU, NDUFA9, and GOT2, were identified by intersecting each set of the top 20 genes ([96]Fig. 2B). Fig. 2. [97]Fig. 2 [98]Open in a new tab The protein-protein interaction network and hub genes investigation based on common-DEGs revealed between T0T and T1T. (A) A protein-protein interaction network constructed by 55 gene nodes and 68 interactions. (B) The 9 hub genes revealed in current study: different color represented different methods of topological analysis. EPC: edge percolated component; MCC: maximal clique centrality; DMNC: density of maximum neighborhood component; MNC: maximum neighborhood centrality. 3.4. Enrichment analysis of hub genes GO and KEGG pathway analyses were performed on the 9 hub genes identified in this study. The results showed that these gene were mainly assembled in GO functions including mitochondrial gene expression (BP, GO:0140053, Genes: MRPL9, MRPL46, and MRPL43) ([99]Fig. 3A), mitochondrial matrix (CC, GO:0005759, Genes: RPL9, IDH3B, MRPL46, MRPL43, HMGCL, ISCU, NDUFA9, and GOT2) ([100]Fig. 3B), and coenzyme binding (MF, GO:0050662, Genes: IDH3B, HMGCL, NDUFA9, and GOT2) ([101]Fig. 3C). In addition, these hub genes were mainly enriched in pathways such as carbon metabolism (hsa01200, Genes: ECHS1, GOT2, IDH3G, and IDH3B) ([102]Fig. 3D). Fig. 3. [103]Fig. 3 [104]Open in a new tab The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis on hub genes. (A) The top 10 GO-BP functions assembled with hub genes. (B) Seven GO-CC functions assembled with hub genes. (C) The top 10 GO-MF functions assembled with hub genes. (D) The top 10 KEGG pathways enriched by hub genes: the X-axis represented the number of genes enriched in certain pathway, while the Y-axis represented different item of pathway; the deeper the red, the more significant the p value; the bigger the node, the more genes enriched. 3.5. PPI network investigation using hub genes A PPI network investigation was performed on 9 hub genes using the GeneMANIA database, followed by functional enrichment analysis of the hub genes ([105]Fig. 4). The results showed that these hub genes were mainly assembled into 5 functions: organellar large ribosomal subunit, mitochondrial ribosome, organellar ribosome, mitochondrial protein complex, and large ribosomal subunit. Fig. 4. [106]Fig. 4 [107]Open in a new tab The protein-protein interaction network constructed by hub genes based on GeneMANIA database. The different color in certain cycle represented different function or network and the line between 2 cycle represented interaction relation. 3.6. DE-lncRNAs analysis With p < 0.05 and log2FC > 0.5, 1 upregulated lncRNA (RP11-373D23.2) and 5 downregulated lncRNAs (ASH1L-AS1, RP11-66N24.4, GS1-358P8.4, RP11-617F23.1, and alpha-2-macroglobulin antisense RNA 1 (A2M-AS1)) were revealed between T0T and T1T ([108]Fig. 5A). Heatmap analysis of the DE-lncRNAs between T0T and T1T showed that all samples could be separated into 2 groups ([109]Fig. 5B). Fig. 5. [110]Fig. 5 [111]Open in a new tab The differentially expressed lncRNAs (DE-lncRNAs) revealed between T0T and T1T. (A) The volcano plot showed the DE-lncRNAs between 2 groups: the blue node represented down-regulated lncRNA, while the red node represented up-regulated lncRNA. (B) The heatmap analysis for DE-lncRNAs showed that all samples could be separated by different group. 3.7. ceRNA network analysis The mRNA-miRNA relation analysis based on the 9 hub genes revealed 4098 mRNA-miRNA relationships, including 1650 miRNAs. LncRNA-miRNA relationship analysis based on 6 DE-lncRNAs revealed 215 miRNAs, 5 lncRNAs, and 272 lncRNA-miRNA interactions. Then, with r > 0.7 (coordinate expression) and p < 0.05, 45 lncRNA-mRNA interactions including 9 mRNAs and 5 lncRNAs were further revealed. The scatter plots of the 4 relationships with correlation coefficient > 0.98 are shown in [112]Supplementary Fig. 4. Finally, combined with the positive co-expression relationship between mRNA and lncRNAs (p > 0.7), 259 lncRNA-miRNA-mRNA relationships, including A2M-AS1-ISCU-miR-7161-5p, were obtained to construct the ceRNA network. The results showed that there were 4 lncRNAs, 127 miRNAs, and 9 mRNAs in the ceRNA network ([113]Fig. 6). Fig. 6. [114]Fig. 6 [115]Open in a new tab The ceRNA network constructed in current study. The red node represented mRNA; the green diamond represented lncRNA; the cyan triangle represented miRNA; the pink arrow represented lncRNA-miRNA relation; the blue T-shaped line represented miRNA-mRNA relations; the grey dotted line represented lncRNA-mRNA relation. 3.8. Drug-gene interaction analysis Based on the CTD, 13 drugs associated with 9 hub genes were identified. The drug-gene interaction network is shown in [116]Fig. 7. Fig. 7. [117]Fig. 7 [118]Open in a new tab The drug-gene interaction network. The red node represented hub genes, the green diamond represented drugs predicted based on Comparative Toxicogenomics Database. 3.9. Differential expression validation of hub genes The differential expression of the 9 hub genes was validated by qRT-PCR and Western blot analysis. As depicted in [119]Fig. 8A, the mRNA expression levels of NDUFA9 (p = 0.017), GOT2 (p = 0.021), IDH3B (p = 0.031), HMGCL (p = 0.026), ISCU (p = 0.003), IMP3 (p = 0.002), and MRPL9 (p = 0.005) were significantly decreased in the muscle tissues of patients with IPC treatment at the T1 time point, compared with T0 time point, which is consistent with the bioinformatics analysis. No significant differences were observed in the expression of MRPL43 or MRPL46. Western blot ([120]Fig. 8B) elicited similar results that the protein expression levels of NDUFA9 (p = 0.030), GOT2 (p = 0.026), MRPL43 (p = 0.032), IDH3B (p = 0.023), HMGCL (p = 0.028), ISCU (p = 0.021), IMP3 (p = 0.036), and MRPL9 (p = 0.016) were significantly lower at T1 time point than that at T0 time point. MRPL46 (p = 0.298) expression showed no significant difference between T0 and T1 time point, warranting further validation. Fig. 8. [121]Fig. 8 [122]Open in a new tab Gene expression validation by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) assay and Western blot. Three paired tissues were collected at T0 and T1 time point. (A) The differential expressions of 9 hub genes were analyzed by qRT-PCR assay. ∗p < 0.05, ∗∗p < 0.01, ns: no significance. (B) Western blot analysis of the nine hub genes. ∗ indicated p < 0.05 compared with T0 time point. NDUFA9: NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 9; GOT2: glutamic oxaloacetic transaminase 2; MRPL43: mitochondrial ribosomal protein L43; MRPL46: mitochondrial ribosomal protein L46; IDH3B: isocitrate dehydrogenase 3b; HMGCL: 3-hydroxy-3-methylglutaryl-CoA lyase; ISCU: ironsulfur cluster scaffold; MRPL9: mitochondrial ribosomal protein L9; IMP3: insulin-like growth factor II mRNA-binding protein 3; 4. Discussion Accumulating evidence has revealed the protective effects of IPC against IR injury during surgery, including TKA.[123]^30^,[124]^31 However, the detailed molecular mechanisms underlying the protective effect of IPC against TKA have not been fully elucidated. In this bioinformatics analysis, 100 DEGs were identified between T0T and T1T, which were mainly associated with mitochondria-related functions and the carbon metabolism pathway. Based on these 100 DEGs, PPI network analysis revealed 9 hub genes, including ISCU. Six DE-lncRNAs were investigated between T0T and T1T. Finally, a ceRNA network was constructed using 259 lncRNA-miRNA-mRNA interactions, including A2M-AS1-miR-7161-5p-ISCU. Avoiding or reducing surgery-related injuries to important organs is an important issue in the field of surgery.[125]^32 Although IPC is a superior strategy, whose clinical application is more realistic and extensive than other surgical strategies, the complex molecular mechanism involved in protecting against IR injury is still not fully explored.[126]^33 ISCU is found to be a hub gene in IR injury protection, revealed by PPI network analysis in the present study. ISCU is an important factor in the mitochondrial electron transport chain and tricarboxylic acid cycle.[127]^42 The integration of iron-sulfur clusters is resistant to oxidative damage to the mitochondria, which further modulates energy metabolism and reduces IR injury.[128]^43 Downregulation of ISCU results in increased accumulation of iron within the mitochondria and suppression of oxidative damage in cells lacking cytosolic copper/zinc superoxide dismutase.[129]^44 Disruption of succinate oxidation by ISCU expression causes adenosine diphosphate phosphorylation and subsequent functional changes in organelles.[130]^45 To dissect the mechanism of the hub gene regulation in IR injury protection, we analyze the differential expressed lncRNAs and predicted the miRNAs. LncRNAs play an important role in IPC against IR during surgery.[131]^34 For example, lncRNA H19 is involved in myocardial IPC by increasing the stability of the nucleolin protein, which is a potential therapeutic target for myocardial injury.[132]^35 The A2M-AS1 is a lncRNA that is commonly differentially expressed in ischemic cells compared with normal cells in humans.[133]^36 Downregulation of A2M-AS1 lessens the injury caused by myocardial ischemia and reperfusion injury.[134]^37 The biological function of A2M-AS1 in human diseases involves sponging certain miRNAs.[135]^38 MiRNAs, including miR-7161, play vital roles in cell stress tolerance and survival prediction in human cancer.[136]^39 Some miRNAs are differentially expressed in human organs following extended cold ischemia but are significantly upregulated by hypoxia, indicating the important role of miRNAs during IR.[137]^40 Upregulation of A2M-AS1 promotes invasion and migration and signifies poor prognosis in cancer by sponging miR-146b to upregulate the target mRNAs.[138]^41 In this study, A2M-AS1 is downregulated in the IPC-treated group between T0 and T1. In the current study, ISCU mRNAs are overexpressed between T0T and T1T. Importantly, ceRNA network analysis shows that A2M-AS1-miR-7161-5p-ISCU is a lncRNA-miRNA-mRNA regulatory relationship. Thus, we speculated that ISCU played a key role in IR injury protection mediated by lncRNA A2M-AS1- miR-7161-5p axis. A close relationship between IPC and mitochondrial permeability transition was revealed in a previous study.[139]^46 Recent evidence suggests that mitochondria are involved in the protective effects of preconditioning.[140]^47 The anti-arrhythmic role of mitochondria-associated functions in IPC has also been demonstrated in animal models.[141]^48 Reperfusion injury involves the opening of mitochondrial permeability transition pores under the conditions of calcium overload and oxidative stress that accompany reperfusion.[142]^49 IPC inhibits mitochondrial permeability transition pore opening in mitochondria isolated from IPC organs after 30 min of global normothermic ischemia or 3 min of reperfusion.[143]^50 As mitochondria contribute to many important signaling pathways, including cardioprotection against IR, the main target of any anti-ischemic protective or post-injury therapeutic strategy should include mitochondria.[144]^51 Mitochondrial function is an important determinant of muscle quality. van Diemen et al.[145]^52 indicated that mitochondrial function, grip strength, and activity are closely associated with mobility recovery after TKA. IPC prevents skeletal muscle IR injury by increasing mitochondrial fusion after TKA.[146]^53 Thus, mitochondrial biogenesis and dynamics are involved in the protective role of IPC against IR injury after TKA. Moreover, carbon metabolism is thought to participate in TKA because of its anti-ischemic role.[147]^54 The mechanism involved in carbon release protection during IR injury was revealed in a previous study.[148]^55 Human clinical investigations have shown that carbon metabolism plays a protective role against IR injury.[149]^56 In the current study, 9 hub genes differentially expressed between T0T and T1T were identified using PPI network analysis. These genes were mainly enriched in mitochondria-related functions (mitochondrial gene expression and mitochondrial matrix) and carbon metabolism pathways. Overall, we speculate that IPC induces a protective response in TKA via mitochondria-related functions and the carbon metabolism pathway. However, there are several limitations to this study that cannot be ignored. Firstly, in the validation experiment, we included fewer cases, which may affect the accuracy of the results. Second, the key role of ISCU in lncRNA A2M-AS1-miR-7161-5p axle-mediated ischemia tolerance was not verified in cell/animal experiments and clinical trials in this study. Therefore, it is necessary to design a better experimental scheme and expand the sample size for further research. In conclusion, ISCU plays a key role in ischemic tolerance mediated by lncRNA A2M-AS1-miR-7161-5p axis. Moreover, IPC may induce a protective response after TKA via mitochondria-related functions and the carbon metabolism pathway. Large number of experiment validations are warranted. CRediT authorship contribution statement Yongli Wang: Data curation, Validation. Bencai Du: Conceptualization. Xueliang Han: Conceptualization. Lianjun Qu: Writing – original draft, Writing – review & editing. Ethical statement This study was approved by the ethics committee of Sunshine Union Hospital, and informed consent has been signed by all participants. All methods were performed in accordance with relevant guidelines and regulations. Funding Nil. Declaration of competing interests The authors declare that they have no competing interests. Footnotes Peer review under responsibility of the Chinese Medical Association. ^Appendix A Supplementary data to this article can be found online at [150]https://doi.org/10.1016/j.cjtee.2024.02.007. Appendix A. Supplementary data The following are the Supplementary data to this article. figs1. [151]figs1 [152]Open in a new tab figs2. [153]figs2 [154]Open in a new tab figs3. [155]figs3 [156]Open in a new tab figs4. [157]figs4 [158]Open in a new tab References