Cuproptosis-related gene ATOX1 promotes MAPK signaling and diffuse large B-cell lymphoma proliferation via modulating copper transport

Diffuse large B-cell lymphoma (DLBCL) is a common subtype of non-Hodgkin lymphoma (NHL), highlighting the importance of studying susceptibility genes to develop personalized treatment strategies. While cuproptosis, caused by high levels of copper ions induced by ionophores, has been shown to affect cancer survival, its specific role in lymphoma is not yet clear. To investigate the involvement of upregulation-related genes in DLBCL, we employed bioinformatics techniques. Specifically, we analyzed the differentially expressed genes (DEGs) in the GSE25638 dataset using weighted gene co-expression network analysis (WGCNA) and performed functional enrichment analysis. By building a protein – protein interaction (PPI) network, candidate genes were identified. Gene set enrichment analysis (GSEA) and receiver operating characteristic (ROC) curve analysis were used to confirm the clinical diagnostic use of these genes. The effects of antioxidant 1 ( ATOX1 ) knockdown, CuCl 2 , and DCAC50 knockdown on DLBCL cells and the activation of the mitogen-activated protein kinase (MAPK) pathway were investigated by conducting in vitro experiments. Bioinformatics and in vitro experiments confirmed elevated expression of ATOX1 in DLBCL cells and tumor samples. ATOX1 knockdown led to decreased cell proliferation and G2 cell cycle arrest in vitro. Additionally, phosphorylated extracellular signal-regulated kinases 1 and 2 (P-ERK1/2) protein levels within the MAPK pathway were reduced as a result of ATOX1 knockdown, but these levels were recovered by CuCl 2 . Treatment with DCAC50 showed a dose-dependent antiproliferative effect in DLBCL cells, which was strengthened by ATOX1 knockdown. Our study demonstrated that ATOX1 may be important in DLBCL via controlling the MAPK pathway through copper transport, providing new insights into potential therapeutic strategies for DLBCL.


INTRODUCTION
Lymphomas are a class of cancerous tumors that develop from lymph nodes or other lymphoid tissues [1].These tumors are classified into two main types: non-Hodgkin lymphoma (NHL) and Hodgkin lymphoma (HL) [2,3].Diffuse large B-cell lymphoma (DLBCL), which makes up around 40% of all B-cell lymphomas, is the most common kind of NHL among them [4].The most typical clinical pathological features of lymphoma are painless lymphadenopathy, hepatosplenomegaly, cachexia, fever, anemia, and swollen lymph nodes can become painful soon after drinking alcohol [5,6].
Studies have shown that immunocompromise, genetics, smoking, and infectious factors such as Epstein-Barr (EB) virus infection, retrovirus, human herpes virus, measles virus, and Helicobacter pylori infection are all risks of lymphoma [7,8].Due to the high heterogeneity of lymphoma, the treatment effect varies widely, and the prognosis of patients is often related to different pathological types and stages.Therefore, exploring the susceptibility genes of lymphoma based on molecular biology is of much significance for clinical diagnosis and the development of individualized treatment plans.
Copper can induce a new mode of cell death called copper-dependent apoptosis (cuproptosis), which is associated with excessive copper ion concentrations [9].Copper is an essential micronutrient that is required for various physiological processes in almost all cell types.However, excessive accumulation of intracellular copper induces oxidative stress and disrupts cellular function, so copper homeostasis is tightly regulated.Recent studies have identified a new form of copper-dependent cell death, which is distinct from all other known cell death pathways.Copper-dependent apoptosis occurs through the binding of copper to the enzyme thioctanoylase in the tricarboxylic acid (TCA) cycle, which leads to subsequent protein aggregation, proteotoxic stress, and ultimately cell death [10].During this process, copper ionophore is a small molecule that can bind with copper and transport them into cells, and some copper ionophores are even used to cancer treatment [11,12].It has been demonstrated that cuprotosis affects the survival prognosis in a number of malignancies.One of the important regulators, FDX1, has been demonstrated to be substantially down-regulated in hepatocellular carcinoma (HCC), higher expression of this gene is associated with a longer survival period [13].Cuproptosis is a newly identified form of copper-driven cell death that has attracted much attention in recent years in the study of cancer pathogenesis.Due to the important role of copper and its triggered cell death in tumorigenesis, copper-based therapies show potential to inhibit tumor growth, especially in response to tumors that are insensitive to chemotherapy, and may offer new strategies for cancer treatment [14].However, it is still necessary to investigate the mechanism by which cuprotosis genes cause lymphoma.
The mitogen-activated protein kinase (MAPK) cascade is a key signaling pathway that regulates a variety of cellular processes, including proliferation, differentiation, apoptosis and stress responses.The MAPK pathway consists of three major kinases: the MAPK kinase kinase, the MAPK kinase, and the MAPK, which activate and phosphorylate downstream proteins sequentially [15].Activation of the mitogenic signaling pathway is a common oncogenic driver of many solid tumors, recent study had shown that selective inhibition of the MAPK pathway can suppress ASCL1-driven small cell lung cancer [16].In addition, recent sequencing and transcriptomics studies have demonstrated the important contribution of the RAS-MAPK pathway to the development and progression of neuroblastoma [17].The MAPK pathway plays a key role in oxidative stress, while AMPK acts as a sensor of cellular energy and participates in the regulation of energy stress response.Activation of AMPK not only induces autophagy-dependent iron death, but also activates iron death via p53, which provides a new direction for the future study of the mechanism of iron death and brings a new vision for cancer treatment strategies [18].In addition, it was found that cadmium induced apoptosis and mitochondrial damage in human bronchial epithelial cells (BEAS-2B) through the MAPK signaling pathway [19].These studies had showed that the MAPK pathway has an important regulatory role in a variety of cellular activities and disease processes, providing new potential targets for the treatment of cancer and other diseases.
In this study, we employed bioinformatics approaches to analyze the GSE25638 dataset using 129 cuproptosis-related genes obtained from the MSigDB, which is part of the GSEA platform, identifying ATOX1 as the pivotal cuproptosis-related gene.Subsequently, we explored the regulatory interactions between ATOX1 and the MAPK pathway in DLBCL cells using in vitro experiments.We also assessed the impact of CuCl2 and DCAC50 treatments on DLBCL cell growth and the MAPK signaling pathway.Through these investigations, we aim to offer novel insights into therapeutic strategies for DLBCL management.

Acquisition and analysis of the GSE25638 dataset
The Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/)provided the DLBCL-related GSE25638 dataset, which included 97 samples total.Of these, 13 normal B cell purified subpopulation samples and 26 DLBCL samples (the selection of 13 normal B cell subpopulation samples was made because these samples can represent the diversity of normal B cells.The choice of 26 DLBCL samples aims to cover the different subtypes or characteristics of DLBCL) were used for the investigation (not validate the findings using an independent cohort).Differentiallyexpressed genes (DEGs) that were upregulated and downregulated were screened using the GEO2R program (specifically, 13 normal B cell purified subpopulation samples and 26 DLBCL samples).The criteria were log2 fold change (log2FC)>1 for up-regulation and <-1 for down-regulation, both with p<0.05.

Weighted gene co-expression network analysis (WGCNA)
To find the optimal soft threshold, a gene co-expression network based on the DEGs screened in the GSE25638 dataset was constructed using the WGCNA method (signed hybrid).The genes were then separated into modules of various colors.The key module was determined by comparing the correlation between the modules and the two sets of samples in the GSE25638 dataset.Generating the cluster dendrogram between the modules and the feature gene adjacency heatmap between the modules.In order to further determine the key modules related to DLBCL, the relationship between module eigengenes (ME) and samples was calculated by Pearson correlation coefficient.

Enrichment analysis of genes in the turquoise module
To visualize the gene expression patterns of the turquoise module, a heatmap was generated depicting the expression levels of genes within this module.

Protein-Protein Interaction networks (PPI) of cuproptosis-related genes
By utilizing gene set enrichment analysis (GSEA; https://www.gsea-msigdb.org),we identified these genes associated with cuproptosis.Subsequently, using the R software package "VennDiagram," we identified genes common to the turquoise module, which exhibited the strongest correlation with the samples, and a set of 129 cuproptosis-related genes.These overlapping genes were then uploaded to the STRING database (https://string-db.org/) to obtain interaction information.The resulting data were visualized using Cytoscape software to construct and visualize the PPI network.To prioritize genes within this network, the Cytohubba plug-in in Cytoscape was employed, applying the MCC, BottleNeck, and Degree algorithms separately to the overlapping genes.Detailed steps involved accessing the STRING database online to download files compatible with Cytoscape for further analysis (separately access the STRING database online and download files to open them using Cytoscape).Only the top ten genes in each subnetwork were then chosen, and candidate genes were once more found using the "VennDiagram" package of R software.

Functional enrichment and diagnostic potential of candidate genes in DLBCL
To delineate the functional significance of the six candidate genes, we performed a Gene Set Enrichment Analysis (GSEA) single-gene Wikipathway enrichment analysis, selecting outcomes with a significance threshold set at p<0.05.After that, Receiver Operating Characteristic (ROC) curves were drawn for these genes using the R software's "timeROC" function.The Area Under the Curve (AUC) values obtained from the ROC analyses were used to critically evaluate the diagnostic predictive accuracy of these genes.Moreover, the Gene Expression Profiling Interactive Analysis2 (GEPIA2; http://gepia2.cancer-pku.cn)database was used to examine the expression profile of ATOX1 (key gene) in DLBCL.

Cell culture, treatment and transfection
From the American Type Culture Collection (ATCC), we obtained three DLBCL cell lines (OCI-LY7, DB, and U2932) as well as the B-lymphocyte cell line (HCC38BL).
The RPMI 1640 medium with 10% fetal bovine serum (FBS, Gibco) and 1% penicillinstreptomycin solution (Solarbio, China) was used to cultivate these cells.The cells were maintained at 37°C and 95% relative humidity in a humidified environment with 5%

Quantitative reverse transcription polymerase chain reaction (qRT-PCR) assay
With the use of the cDNA synthesis kit (DBI, Germany) and Trizol reagent (Invitrogen, USA), we were able to extract the RNA from transfected cells and perform reverse transcription.SYBR Green Master Mix (Applied Biosystems, USA) was used for qRT-PCR, which was carried out using an AB Fast 7500 real-time apparatus.For amplification, the primer sequences listed in the supplementary table 1 were used.After standardizing the expression levels against the internal reference GAPDH [20], the relative gene expression levels of ATOX1 were calculated using the 2 -ΔΔCt approach.

Western blotting (WB) assay
Protease and phosphatase inhibitors (Roche) were added to RIPA lysis buffer (Beyotime Imaging System from the same manufacturer was utilized to detect the bands.

Flow cytometry
Flow cytometry was used to evaluate the cell cycle distribution in DLBCL cells.
Propidium iodide (PI) staining was applied to cells by standard protocol, the cells were then analyzed using a BD FACSCanto II flow cytometer.Data interpretation and cell cycle phase quantification were executed using appropriate analytical software.

Cell Counting Kit-8 (CCK-8)
The CCK-8 kit (Qihai Biotechnology, China) was used to test cell proliferation.96-well plates were used to cultivate the transfected and DCAC50-treated cells after incubating them in CCK-8 solution for two to four hours at 37°C.The absorbance (OD) of each well was measured at 450 nm after 24, 48, and 72 hours of incubation.

Statistical analysis
The data was analyzed using GraphPad Prism (version 9.0.0,GraphPad Software, USA).All experiments were repeated at least three times and data are presented as mean ± standard error.To establish statistical significance, the Student's t-test or one-way analysis of variance (ANOVA) were performed, then the post hoc Tukey multiple comparison test was used to assess the results.The statistically significant result was defined as p < 0.05.

Screening of DEGs in DLBCL and identification of the turquoise module
We utilized the GEO2R tool from the GSE25638 dataset to detect 1608 upregulated and 324 downregulated DEGs (Figure 1A).By applying the WGCNA algorithm with a soft threshold of 14, we constructed a gene co-expression network for these DEGs (Figure 1B).Next, we classified the genes into distinct color modules based on their expression levels, depicted in Figures 1C and 1D.Evaluating the correlation between the DLBCL and normal samples, we observed a significant relationship (correlation coefficient of 0.888) between the turquoise module and the samples (p-value 4.4e-14) (Figures 1E and 1F).As a result, we identified the turquoise module as the pivotal module.

Functional enrichment analysis reveals key biological processes and pathways of turquoise module
The turquoise module had a total of 942 genes (Figure 2A), which were subjected to GO and KEGG pathway enrichment analysis.As shown in Figures 2B-2D, the enriched terms of these DEGs in BP, CC, and MF included e.g., Neutrophil degranulation, Neutrophil mediated immunity, Endoplasmic reticulum lumen, Metal ion binding.
Furthermore, KEGG pathways such as Malaria, Cytokine-cytokine receptor interaction, Focal adhesion, Complement and coagulation cascades, and Lysosome were enriched in these DEGs (Figure 2E).

Turquoise module and cuproptosis-related genes revealed six candidates through the PPI network
We identified 24 overlapping genes from a pool of 942 GSE25638-DEGs and 129 cuproptosis-related genes.These genes formed a PPI network consisting of 24 nodes and 37 edges (Figures 2F and 2G).The Cytohubba plug-in in the Cytoscape program was then utilized to generate three sub-networks using the MCC, BottleNeck, and Degree algorithms (Figures 2H-2J).By comparing the results from these three topological analysis methods using Venn diagrams, six central candidate genes were identified (Figure 2K).

ATOX1 was the hub gene related to cuproptosis in DLBCL
GSEA was performed using Wikipathway module/category on six genes, each gene was enriched with two pathways (Figures 3A-3F).GSEA showed MT1H enrichment in Vitamin B12 metabolism and Physiological and pathological hypertrophy of the heart pathways (Figure 3A), MT1X was enriched in Copper homeostasis and Urea cycle and association pathways (Figure 3B), MT2A showed enrichment in Vitamin B12 metabolism and Folate metabolism pathways (Figure 3C), ATOX1 was enriched in Pathways of nucleic acid metabolism and Sphingolipid integrated pathway (Figure 3D), CP showed enrichment in Tumor suppressor activity of smarcb1 and EDA signaling in hair follicle development pathways (Figure 3E), SLC11A2 was enriched in apoptosisrelated network due to altered notch3 in ovarian cancer and Fatty acid biosynthesis pathways (Figure 3F).Then in ROC curve analysis (all patients), ATOX1 had the highest AUC value of 0.999 (Figure 3G), indicating that it had the strongest predictive ability for DLBCL patients, followed by CP (Figure 3H, AUC=0.941),MT2A (Figure 3I, AUC=0.934),MT1X (Figure 3J, AUC=0.925),MT1H (Figure 3K, AUC=0.850).
These findings showed that ATOX1, CP, MT2A, MT1X, MT1H could be the diagnostic biomarkers for DLBCL patients.GEPIA database showed that ATOX1 gene was highly expressed in the tumor tissues of DLBCL (Figure 4A).In this study, we selected ATOX1 as the hub gene for subsequent analysis.

ATOX1 depletion induces G2 cell cycle arrest and alters cell cycle regulatory proteins in DLBCL
We used qRT-PCR and WB assays to identify ATOX1 expression in DLBCL cells.
When compared to the control cell HCC38BL, ATOX1 mRNA and protein levels were considerably higher in DB, and U2932 cells in DLBCL, where DB and U2932 exhibited significant upregulation (Figure 4B).Knockdown using si-ATOX1#1 showed greater efficiency in DB and U2932 cell lines and was selected for subsequent analysis (Figure 4C).ATOX1#1 knockdown dramatically lowered the proliferation rate of DLBCL cells, as demonstrated by CCK-8 assays (Figures 4D and 4E).The results from flow cytometry indicated that when ATOX1 was knocked down using si-ATOX1#1, there was a notable difference compared to the control group, further investigation into the cell cycle demonstrated a significant G2 phase arrest (Figures 4F and 4G).WB methods was utilized to determine the protein expression levels of cell cycle-related proteins in DLBCL cell lines following ATOX1 knockdown.The results showed that after knocking out ATOX1, cell cycle regulatory protein (Cyclin B1) was significantly downregulated, while cell cycle negative regulatory protein (p27) was significantly upregulated (Figure 4H).This supported a potential role for ATOX1 in cell cycle control in DLBCL.
WB analysis showed that protein levels were noticeably reduced for most members of the MAPK pathway in DLBCL cells treated with DCAC50, indicating potential disruption of this pathway and contributing to its anti-proliferative effects (Figures 6C   and 6D).Moreover, simultaneous knockdown of ATOX1#1 enhanced the inhibitory effects of DCAC50 on cell proliferation (Figures 6E and 6F).These findings that DCAC50 exerts a dose-dependent inhibitory effect on DLBCL cell growth, with more pronounced effects observed at higher concentrations, and knockdown of ATOX1#1 promoting its effects.

DISCUSSION
The most common form of malignant lymphoma in medical practice is DLBCL, which primarily affects middle-aged and elderly individuals [21].Studies have demonstrated that typical indications of DLBCL include high fever and a range of systemic symptoms, such as enlarged painless neck and progressive enlargement of lymph nodes in the supraclavicular region [22,23].Furthermore, personalized chemotherapy combined with immune targeted therapy is often utilized based on specific symptoms, type of pathology, and location of the tumor.However, due to its high prevalence and aggressiveness, the treatment of DLBCL is often challenging and the prognosis is usually poor [24,25].Previous research has proposed that cuproptosis may contribute to the development of malignant tumors [26,27].In this study, we employed bioinformatics and in vitro experimental investigation to examine the mechanism and clinical significance of cuproptosis-related genes in DLBCL.
cancer cell invasion and lymph node metastasis, their oxidative pathways provide new targets for the treatment of DLBCL [33].As a treatment option for DLBCL, fatty acid synthase inhibition of fatty acid synthase preferentially disrupts de novo fatty acid synthesis [34].According to ROC analysis, ATOX1, CP, MT2A, MT1X, and MT1H all had AUC values better than 0.85.The AUC value of ATOX1 was the highest, indicating that it had the strongest predictive ability for DLBCL patients.Furthermore, GEPIA database analysis revealed that the ATOX1 gene was substantially expressed in DLBCL tumor tissues.As a result, we chose ATOX1 as the hub gene.
ATOX1, a copper chaperone, has a significant impact on cellular antioxidant defense and copper homeostasis [35].It is involved as a cytoplasmic protein in the transport of intracellular copper ions to various copper-dependent enzymes, including those involved in oxidative stress protection and angiogenesis [36].ATOX1 is responsible for delivering copper to copper-dependent enzymes involved in antioxidant defense and extracellular matrix remodeling.ATOX1-mediated dysregulation of copper homeostasis is associated with increased oxidative stress and altered cellular behavior, potentially leading to cancer development [37].Research has indicated that breast cancer cell migration is influenced by the copper chaperone ATOX1 [38].Another study showed that activin A-induced migration and colony formation of colon cancer cells were enhanced by nuclear translocation of ATOX1 [39].Through a series of in vitro cell experiments, we discovered that the knockdown of ATOX1 inhibited cell growth and caused cell cycle arrest in the G1 phase of the cell cycle.Knockdown of ATOX1 resulted in downregulation of Cyclin B1 protein expression level and upregulation of p27 protein expression, implying its role in inhibiting cell cycle progression.Moreover, we found that after ATOX1 knockdown, P-ERK1/2 levels within the MAPK pathway significantly decreased.Notably, copper supplementation partially restored MAPK pathway activity that was attenuated by ATOX1 reduction.This shows that ATOX1 is an important player that may have potential oncogenic properties.
One significant signaling cascade that controls several cellular functions, such as division, growth, proliferation, and responsiveness to external stimuli, is the MAPK pathway [40].This pathway consists of a number of protein kinases that influence gene expression and cell behavior by transmitting signals from cell surface receptors to the nucleus.The MAPK pathway consists of three major kinases, namely JNK, ERK, and p38 MAP kinase [41].A sequence of phosphorylation events activate these kinases in response to a variety of stimuli, including growth factors, stress, cytokines, and hormones [42].When active, the MAPK pathway performs a variety of functions, whereas the ERK pathway is largely engaged in cell proliferation, survival, and differentiation [43].The JNK pathway is frequently activated by stress signals and plays a role in apoptosis, inflammation, and cellular damage responses [44].The p38 pathway, which responds to numerous stimuli, regulates inflammation, apoptosis, and cell cycle arrest [45].Abnormal regulation of the MAPK pathway is also linked with tumor.
Research has demonstrated that via controlling the ERK/MAPK signaling pathway and specifically targeting MEK1, miR-101 regulates cell proliferation and apoptosis in DLBCL [46].Another study found that ATOX1 is necessary for MAPK pathway activation in melanoma, which indirectly supports MEK1/2 copper binding [47].
Through the WB method, we discovered that P-ERK1/2 protein expression within the MAPK pathway was significantly reduced in DLBCL cells when ATOX1 was knocked down, which could be reversed by CuCl2.This highlights the intertwined roles of ATOX1 and copper in regulating MAPK pathway activation.DCAC50 is a small molecule inhibitor, study had confirmed that is related to the other partner protein ATOXA [48].Our observations revealed that DCAC50 administration ensures a dosedependent attenuation in DLBCL cell proliferation, with pronounced inhibition discernible at higher dosages, especially at 15 μM and 20 μM.The antiproliferative prowess of this compound seemingly hinges on its capacity to disrupt the MAPK pathway.Intriguingly, this inhibition is augmented when combined with an ATOX1 knockdown, hinting at a synergistic relationship.The interplay between DCAC50 and the MAPK pathway, particularly in the context of ATOX1, not only broadens our understanding of DLBCL pathophysiology but also furnishes potential therapeutic avenues.

Limitation
Despite the valuable insights gained from our study, it is important to acknowledge several limitations that may affect the interpretation and generalization of our findings.
First, our conclusions were largely based on database and cell line data, which provided preliminary exploratory information.However, the lack of direct tissue-level validation is a major limitation.In future studies, we plan to validate the expression of ATOX1 in different types of tissue samples by immunohistochemistry (IHC) experiments to provide stronger evidence to support our hypothesis and conclusions.Second, we have not yet validated the therapeutic effect of ATOX1 in DLBCL in in vivo experiments.
Although our study revealed the potential role of ATOX1 in regulating the MAPK pathway through copper transport in DLBCL, these conclusions were only based on preliminary experimental data.To further delve into the study and validate our hypothesis, we plan to conduct detailed in vivo experiments in animal models of DLBCL to assess the specific role of ATOX1 and its potential therapeutic effects.In conclusion, these limitations emphasized the need for further studies to validate our preliminary findings and to provide a stronger foundation for the future development of new therapeutic strategies.Furthermore, our study was a preliminary exploration of the unique role and innovative potential of the ATOX1 gene in DLBCL.Future studies will focus on integrating other related cup mutated genes and using multivariate analysis methods to establish more comprehensive and accurate prognostic prediction models for DLBCL.This will help to provide clinicians with more effective support for individualized therapeutic decision-making, thereby maximizing patient clinical management and disease prognosis.

Figure 1 .
Figure 1.Gene co-expression network of DEGs in the GSE25638 dataset.(A)

Figure 2 .
Figure 2. Turquoise module and cuproptosis-related genes revealed six candidates

Figure 3 .
Figure 3.The GSEA and ROC analysis on the candidate genes.(A-F) Enrichment

Figure 4 .
Figure 4. Regulatory effects of ATOX1 knockdown on cell proliferation, and cell

Figure 6 .
Figure 6.Evaluation of DLBCL Cell Responses to DCAC50 Treatment and ATOX1