A microarray data analysis investigating the pathogenesis and potential biomarkers of autophagy and ferroptosis in intervertebral disc degeneration

Background: Intervertebral disc degeneration (IDD) entails complex pathological changes and causes lower back pain (LBP). However, there is still a lack of understanding of the mechanisms involved in IDD, particularly regarding the roles of autophagy and ferroptosis. The current study used microarray data to investigate the pathogenesis of IDD and potential biomarkers related to autophagy and ferroptosis in IDD. Methods: Differentially expressed genes (DEGs) were identified by analyzing the mRNA and miRNA expression profiles of IDD patients from the Gene Expression Omnibus (GEO). The protein-protein interaction network, Gene Ontology (GO) enrichment, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, and gene set enrichment analysis (GSEA) were utilized. The Human Autophagy Database (HADb) and Ferroptosis Database were used in conjunction with hub genes to identify autophagy- and ferroptosis-related genes. The Transcription Factor -hub gene-miRNA network was constructed. Lastly, the expression of DEGs in normal and degenerated nucleus pulposus cells (NPCs) was investigated via the quantitative reverse transcription polymerase chain reaction (qRT-PCR). Results: A total of 362 DEGs associated with IDD were identified. GO and KEGG analyses indicated that oxidative stress, extracellular matrix, PI3K-AKT signaling pathway, and ferroptosis were key factors in IDD occurrence. GSEA indicated that IDD was associated with changes in autophagy, iron ion homeostasis, extracellular matrix, and oxidative stress. Eighty-nine hub genes were obtained, including five that were autophagy-related and three that were ferroptosis-related. Of these, TP53 and SESN2 were the intersections of autophagy- and ferroptosis-related genes. In qRT-PCR analysis, CANX, SLC38A1, and TP53 were downregulated in degenerative NPCs, whereas GNAI3, SESN2, and VAMP3 were upregulated. Conclusion: The current study revealed aspects of autophagy- and ferroptosis-related genes involved in IDD pathogenesis, warranting further investigation.


Introduction
Lower back pain (LBP) affects people of all cultures and ages.It reportedly affects 700 million patients worldwide physically, psychologically, and economically and reduces the productivity of society by causing massive economic losses (Francisco et al., 2022).Intervertebral disc degeneration (IDD) is thought to be the most important component of LBP (Zhao et al., 2021).Intervertebral discs (IVDs) are composed of fibrous tissue and cartilage and have an avascular structure that contains three layers: the annulus fibrosus with its outer and inner parts, the central nucleus pulposus, and the endplate.The normal function of an IVD depends on the cells and stroma (Kos et al., 2019).Multiple studies have indicated that various pathological factors, such as extracellular matrix (ECM) degradation, oxidative stress, inflammatory responses, mitochondrial dysfunction, nutritional deficiency, abnormal mechanical loading, and epigenetic changes, are involved in IDD progression (Zhang G. Z. et al., 2020;Zhang et al., 2021;Liang et al., 2022).However, the regulatory mechanisms involved in IDD are complex; thus, many of them remain to be elucidated.
Autophagy is a process in which cells self-digest their own components.It nourishes cells during fasting, clearing them of excess or damaged organelles, misfolded proteins, and invading microorganisms (Levine and Kroemer 2008).In recent years, numerous studies have indicated that disordered autophagy may lead to IDD (Gong and Zhang 2021; Madhu et al., 2021).Studies investigating autophagy in IDD are still at a very early stage, and the physiological role of autophagy in normal IVDs and specific mechanisms involved in autophagy regulation in IVD cells remain unclear (Kritschil et al., 2022).More studies on relevant functions and mechanisms are needed to clarify the role of autophagy in IDD.
The term ferroptosis, coined in 2012 (Dixon et al., 2012), describes a unique form of cell death.It is driven by irondependent phospholipid peroxidation and regulated by multiple cellular metabolic events, such as redox homeostasis, mitochondrial activity, and metabolism of amino acids, lipids, and sugars, as well as many disease-related signaling pathways.Ferroptosis may be involved in many organ damage and degenerative pathologies (Jiang et al., 2021).Previous studies have suggested that ferroptosis might play a role in IDD (Zhang GZ. et al., 2020;Lu et al., 2021).
The current study used microarray data analysis to investigate IDD pathogenesis and potential biomarkers related to autophagy or ferroptosis in IDD.

Identification of differentially expressed genes and hub genes
The GEO2R network tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/) is based on the R software "LIMMA" and GEO query (Barrett et al., 2012), and it was used in the current study to identify differentially expressed mRNAs and miRNAs in healthy and IDD samples.A log2 fold change of >1 and p of <.05 were used as the criteria to identify differentially expressed genes (DEGs).The intersections of the DEGs in the two mRNA datasets and two miRNA datasets were used to obtain differentially expressed mRNAs and miRNAs.Target genes were predicted from the differentially expressed miRNAs using MIRNET (https://www.mirnet.ca/),and hub genes were identified after crossing the target genes with the differentially expressed mRNAs.VEEN diagrams were then constructed to detect and visualize the intersections of these databases.

Functional enrichment analyses of hub genes
Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses of the hub genes were performed using the R software "clusterProfiler" (Yu et al., 2012), and the results were visualized with the R software "ggplot2".

Transcription factor-hub gene-miRNA network construction
Network Analyst (https://www.networkanalyst.ca/)was used to predict transcription factors that could regulate hub genes to construct the transcription factor-hub gene-miRNA network, and this network was visualized using Cytoscape (v.3.8.0).

Autophagy-related and ferroptosisrelated genes in IDD
Autophagy-related genes were obtained from the Human Autophagy Database (HADb) (http://www.autophagy.lu/index.html) and then intersected with the hub genes to obtain the autophagy-related genes in IDD.Similarly, driver, suppressor, and marker data were obtained from the Ferroptosis Database (FerrDb) (http://www.zhounan.org/ferrdb),with removed duplication, and then intersected to obtain ferroptosis-related genes for subsequent analysis.

Isolation and culture of nucleus pulposus cells
Normal human nucleus pulposus tissue was collected, washed with phosphate-buffered saline (PBS), and then cut into pieces.After digestion with .2%collagenase type II (Thermo Fisher Scientific, United States) for 2 h, followed by .25%digestion with trypsin for .5 h, the samples were washed again with PBS.After centrifugation, they were transferred to culture flasks with Dulbecco's Modified Eagle's Medium)/ F12 medium containing 10% fetal bovine serum (Thermo Fisher Science, United States) and 1% penicillin-streptomycin (Thermo Fisher Science, United States) at 37 °C in a 5% CO 2 incubator for cell isolation and cultivation.The cell culture medium was replaced on the fifth day and every two or three days thereafter.After 80%-90% confluency was reached, nucleus pulposus cells (NPCs) were passaged.Third-passage NPCs were used in subsequent experiments.The experimental protocol was approved by the Ethics Committee of Zhujiang Hospital of Southern Medical University.

Quantitative reverse transcription-PCR
NPCs were divided into normal and degeneration groups.In the normal group, the cell culture medium was replaced every 48-72 h.In the degeneration group, the NPCs were treated with 100 ng/ml tumor necrosis factor (TNF)-α for 24 h.Total RNA was extracted and reverse-transcribed via a kit in accordance with the manufacturer's instructions (Accurate Biotechnology, China).Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed using the CFX Connection Real-Time System (Bio-Rad, United States).The 2 −ΔΔCt method was used to evaluate gene expression relative to GAPDH expression.Sequence fragments of RNAs are shown in Table 3.

Statistical analysis
Data were expressed as mean ± the standard deviation.Gene expression levels of clinical samples were compared using the independent sample t-test.Benjamini-Hochberg procedure was used to reduce the false-positive rate of multiple tests.A p of <.05 was considered statistically significant.All statistical analyses were conducted using R programming (https://www.r-project.org/,version 4.0.2).

Identification of DEGs and hub genes
A total of 1,817 upregulated and 1,689 downregulated mRNA and 470 upregulated and 538 downregulated miRNA were identified.Of these, 362 overlapping mRNA were defined as differentially expressed mRNA, and 23 overlapping miRNAs were defined as differentially expressed miRNAs (Figures 1A,B).A total of 2773 target mRNAs were predicted from these 23.Eighty-nine hub genes were identified by intersecting the 2773 target mRNAs with 362 differentially expressed mRNAs (Figure 1C; Table 1).

PPI network analysis
The PPI network consisted of 361 nodes and 1595 edges (Figure 2A).The average node degree was 8.84, and the local clustering coefficient was .397.The three proteins with the highest node degrees were ACTB (90 nodal degrees), TP53 (69 nodal degrees), and FN1 (60 nodal degrees).

GO and KEGG enrichment analyses of DEGs
GO and KEGG analyses of the DEGs obtained from the PPI network were performed to verify the functions of differentially expressed mRNAs in IDD.GO analysis results indicated changes in responses to oxidative stress, reactive oxygen species (ROS) metabolic processes, transforming growth factor beta (TGF-β) receptor signaling pathway, TGF-β, growth factor binding, collagen-containing ECM, ECM organization, mitochondrial respirasome, apoptotic signaling pathway, intrinsic apoptotic signaling pathway in response to oxidative stress, and lysosomes (Figure 2B).KEGG pathway enrichment analysis indicated that the DEGs were significantly enriched in

GSEA
Genes in GSE15227 were significantly involved in five biological functions: regulation of autophagy, autolysosomes, iron ion transport, responses to oxidative stress, and collagencontaining ECM (Figure 3A).Genes in GSE23130 were also significantly involved in five biological functions: autophagosomes, processes utilizing autophagic mechanisms, iron ion homeostasis, responses to oxidative stress, and collagen-containing ECM (Figure 3B).

GO and KEGG enrichment analyses of hub genes
In GO analysis, hub genes were mainly enriched in ROS metabolic processes and collagen-containing ECM.In KEGG pathway analysis, hub genes were mainly enriched in the PI3K- Akt signaling pathway, ECM-receptor interaction, and cellular senescence (Figure 4; Table 2).

TF-hub Gene-miRNA network
A TF-hub gene-miRNA network containing 89 mRNAs, 22 miRNAs, and 81 TFs was generated.The top targeted miRNA hub gene was TP53, which was modulated by seven miRNAs.The top targeted TF hub gene was GPBP1, which was modulated by 20 TFs.The TF that regulated the most hub genes was FOXC1, which regulated 56 hub genes.The DEM that regulated the most hub genes, 22, was hsa-mir-185-5p (Figure 5A).Six genes were obtained after the 81 predicted TFs were intersected with the 89 hub genes: TP53, DAZAP2, HLA-B, MSN, UBLCP1, and PPT1.This indicated that they are simultaneously regulated by TFs and differentially expressed miRNAs and also regulate some hub genes (Figure 5B).TP53 was evidently a key juncture of the network (Figure 5C).

Autophagy-related and ferroptosisrelated hub genes in IDD
Two hundred and twenty-two autophagy-related genes and 487 ferroptosis-related genes were obtained.Intersecting   222 autophagy-related genes with 89 hub genes resulted in the identification of five autophagy-related hub genes in IDD: CANX, GNAI3, SESN2, TP53, and VAMP3.Intersecting 487 ferroptosis-related genes with 89 hub genes resulted in the identification of three ferroptosis-related hub genes in IDD: TP53, SLC38A1, and SESN2.TP53 and SESN2 were simultaneously associated with both autophagy and ferroptosis (Figure 6A).The related TF-hub gene-miRNA network was then constructed (Figure 6B).

Discussion
IDD is an important cause of LBP and involves complex pathophysiological changes and molecular mechanisms.
Research on the occurrence and developmental mechanisms involved in IDD has been limited, particularly regarding mechanisms relating to autophagy and ferroptosis.In the current study, microarray and bioinformatics analyses were used to investigate possible molecular mechanisms involved in IDD occurrence and development, and potential biomarkers were predicted.
Both GO and KEGG enrichment analyses of DEGs and hub genes highlighted the importance of oxidative stress, ECM, and PI3K-AKT signaling pathway, and all three of these were related to autophagy and ferroptosis.Activation of the PI3K-AKT signaling pathway can increase ECM content, prevent apoptosis, promote cell proliferation, induce or prevent autophagy, mitigate oxidative damage, and contribute to adaptation to a hypoxic microenvironment, thereby preventing IDD (Ouyang et al., 2017).Exosomes of the cartilage endplate activate the PI3K/AKT/autophagy pathway, inhibit NPC apoptosis, and slow the IDD process (Luo et al., 2021).Islet amyloid polypeptide regulates gene expression responsible for ECM anabolism and catabolism and controls the interaction between apoptosis and NPC autophagy via the PI3K/AKT/mTOR and p38/JNK MAPK signaling pathways (Wu et al., 2017).XIST may act as a competitive endogenous RNA of  (Chen W et al., 2021).Bone marrow mesenchymal stem cell-derived extracellular vesicles harboring circ_0072464 inhibit NPC ferroptosis by inhibiting miR-431 and upregulating NRF2, which promotes NPC matrix synthesis and proliferation, alleviating IDD (Yu et al., 2022).Under oxidative stress, ferroprotein imbalance leads to NP intercellular iron overload, resulting in NPC ferroptosis and accelerating IDD progression in vivo (Lu et al., 2021).Homocysteine promotes IDD development by upregulating oxidative stress and ferroptosis in the nucleus pulposus by enhancing GPX4 methylation (Zhang G. Z. et al., 2020).
Previous studies have indicated that oxidative stress, ECM, and PI3K-Akt signaling pathway might be clinical biomarkers or therapeutic indicators in IDD, which is consistent with the current study.Based on the present analysis and previous studies, we suggest that the roles of autophagy and ferroptosis in IDD warrant attention.
TP53 encodes a tumor suppressor protein containing transcriptional activation, DNA binding, and oligomerization domains.The p53-mediated transcriptional suppression of SLC7A11 promotes ferroptosis in cancer cells.p53 R273H and R175H inhibit the expression of SLC7A11 by not binding to DNA and by inhibiting the activity of other transcription factors, suggesting that the expression of this core ferroptosis regulator is regulated by an integrated network of transcription factors (Chen X et al., 2021).p53 can enhance ferroptosis by enhancing the expression of SAT1 (spermidine/spermine N1-acetyltransferase 1) and GLS2 (glutaminase 2) or by inhibiting that of SLC7A11 (solute carrier family seven member 11).On the other hand, p53 suppresses ferroptosis through the direct inhibition of DPP4 (dipeptidyl peptidase 4) activity or by the promotion of CDKN1A/p21 (cyclin-dependent kinase inhibitor 1A) expression (Kang et al., 2019).The loss of TP53 prevents nuclear accumulation of DPP4, which promotes plasma membrane-associated DPP4-dependent lipid peroxidation, leading to ferroptosis (Xie et al., 2017).Previous studies have demonstrated the central role of TP53 in the ferroptosis regulatory network, controlling ferroptosis by directly regulating the expression of downstream ferroptosis-related target genes or regulating the expression of other transcription factors to control the expression of death-related genes.Our results indicated that TP53 might regulate downstream ferroptosis-related target gene SESN2 to regulate ferroptosis, while TP53 might be regulated by other transcription factors (such as DAZAP2, HLA-B, MSN, UBLCP1, and PPT1) to form a complex and integrated regulatory network of transcription factors, which was consistent with previous studies.
SESN2 encodes a member of the sestrin family of PA26related proteins.SESN2 can play an antioxidant role in sepsis, downregulating the ATF4-CHOP-CHAC1 signaling pathway and inhibiting ferroptosis in dendritic cells (Li et al., 2021).Ferroptosis-mediated SESN2 induction protects against iron overload and ferroptosis-induced liver injury, relying on NRF2 (Park et al., 2019).Based on existing studies, we believe that TP53 may be a central gene in the regulatory network of ferroptosis in IDD, and it may play an important role in IDD by regulating SESN2 or a complex transcription factor regulatory network to control ferroptosis.
CANX encodes a member of the calnexin family of molecular chaperones.Blocked mitosis in hypoxic cells may be due to the fact that after CANX knockout, the number of elongated mitochondria is increased, and the colocalization of autophagosomes and mitochondria is reduced (Wu et al.,FIGURE 6 Autophagy-related and ferroptosis-related hub genes in IDD and associated TF-hub gene-miRNA network.(A) Autophagy-and ferroptosisrelated hub genes in IDD, indicating that there are two genes simultaneously associated with both autophagy and ferroptosis.(B) Red rectangles represent TFs, green rectangles represent miRNAs, yellow and blue rectangles represent autophagy and ferroptosis genes, respectively, and yellow rectangles with a blue frame represent genes associated with both autophagy and ferroptosis.

Gene
Forward primer Reverse primer GNAI3 belongs to the G-alpha family.GNAI3 reduces autophagy flux and inhibits the autophagic degradation of hepatitis B virus antigen by enhancing the insulin-induced AKT-MTOR pathway (Wang et al., 2022).The activator of G-protein signaling 3 (AGS3), which is directly bound to light chain 3 (LC3), recruits Gαi3 to the LC3-positive membrane upon starvation and enhances autophagy by inhibiting G protein.Gαinteracting vesicle-associated protein (GIV) destroys the Gαi3-AGS three complex under the stimulation of growth factors, releases Gαi3 from the LC3-positive membrane, and inhibits autophagy by activating G protein (Garcia-Marcos et al., 2011).At present, there is no research involving GNAI3 and its encoded Gαi3 in the field of IDD.Elucidation of the role of GNAI3 in IDD may contribute to a further understanding of autophagy in IDD.
The present study had some limitations.First, the data used were downloaded from public databases, and it was not possible to evaluate whether some of those data contained errors.Second, only IDD tissue samples were analyzed, but a comprehensive analysis of venous blood samples and IDD tissue may be more helpful with respect to providing a general understanding of the potential mechanisms involved in IDD, which would be an improvement for future research.Lastly, a large amount of evidence is still needed to support our hypothesis.In subsequent studies, we will use various experiments, such as western blotting, immunohistochemistry assays, and dual luciferase reporter gene assays, to verify our conclusions and further clarify the mechanisms of IDD autophagy and ferroptosis.
In conclusion, the purpose of this study was to investigate the molecular mechanisms involved in IDD progression via comprehensive bioinformatics analysis to complement the understanding of autophagy and ferroptosis in IDD.This study might lead to the discovery of new IDD biomarkers, and more research is required to validate our results.

FIGURE 1
FIGURE 1 Differential expression of mRNAs, miRNAs, and hub genes in IDD.(A) Venn diagram of differentially expressed mRNAs in the two mRNA datasets.(B) Venn diagrams of differentially expressed miRNAs in two miRNA datasets.(C) Venn diagram of hub genes in IDD.
FIGURE 2 mRNA-related PPI networks in IDD.(A) PPI network of differentially expressed mRNAs in IDD.(B) clueGO-GO analysis of genes in PPIs that had differentially expressed mRNAs associated with IDD.(C) clueGO-KEGG analysis of genes in PPIs that had differentially expressed mRNAs associated with IDD.

FIGURE 3
FIGURE 3 GSEA of the two datasets.(A) GSEA of genes in the GSE15227 dataset.(B) GSEA of genes in the GSE23130 dataset.

FIGURE 4
FIGURE 4GO and KEGG enrichment analyses of hub genes.

FIGURE 5
FIGURE 5 TF-hub gene-miRNA network.(A) Red nodes represent TFs, green nodes represent miRNAs, and yellow nodes represent hub genes.(B) Venn diagram of hub genes acting as TFs in IDD.(C) Yellow nodes represent hub genes, and yellow nodes with red borders represent hub genes acting as transcription factors.
SLC7A11 has been confirmed to be regulated by TP53.SLC38A1 and SLC7A11 belong to different family members of solute carrier (SLC) and participate in the regulation of ferroptosis.Whether TP53 and SLC38A1 have a regulatory relationship deserves further attention and research.
IDD autophagy and regulatory axis of TP53-CANX/ TP53-VAMP3 may inspire new ideas for the research on the regulatory network of IDD autophagy.

TABLE 2
The GO and KEGG enrichment analyses of hub genes in IDD.

TABLE 2 (
Continued) The GO and KEGG enrichment analyses of hub genes in IDD.