Introduction
Multiple myeloma (MM) is a malignant hematologic tumor that originates in the bone marrow. Annually, ~588,161 individuals worldwide are diagnosed with MM ( 1 ). In 2020, among 1,278,362 cases of leukemia, lymphoma and MM, 176,404 cases of MM were reported, accounting for 14% of the total ( 2 ). Despite the unprecedented response and survival rates in MM treatment, the disease remains considered incurable due to its complex pathogenesis, recurrence and drug resistance ( 3 ). Furthermore, MM manifests with various clinical symptoms, leading to misdiagnosis and underdiagnosis, which can delay optimal treatment and pose a significant threat to patients' lives ( 4 ). Therefore, it is crucial to explore the pathogenesis of MM and identify reliable laboratory diagnostic biomarkers.
Aging is a natural and inevitable process that occurs over time in organisms ( 5 ). Pathologically, aging results from a cumulative accumulation of stress, injury, infection, immune response and metabolic disorders ( 6 ). Aging is one of the most prominent risk factors for various malignancies ( 7 ). There is substantial evidence of a bidirectional relationship between aging and malignant diseases, both of which share numerous common characteristics ( 8 ). Notably, cellular aging processes contribute to the pathogenesis of MM ( 9 ). Additionally, aging-related dysfunction of T cells and the immune system can exacerbate MM onset and progression ( 10 – 12 ). Thus, investigating aging at the molecular level may offer novel insights for the clinical translation of MM therapies.
By integrating bulk profiles with machine learning algorithms [least absolute shrinkage and selection operator (LASSO) regression and support vector machine recursive feature elimination (SVM-RFE)], this study identified TXN as an aging-associated diagnostic biomarker for MM. TXN functions as a pro-inflammatory factor by activating the NF-κB signaling pathway and generating reactive oxygen species (ROS) ( 13 , 14 ). In vitro assays further confirmed its upregulated expression in MM cell lines, highlighting its pathogenic potential.
Materials and methods
ARGs and sequencing dataset
In this study, 307 genes were selected from the Human Ageing Genomic Resources ( http://hagr.ageing-map.org/ ). The mRNA expression profile datasets GSE6477 and GSE16558 were downloaded from the Gene Expression Omnibus database ( https://www.ncbi.nlm.nih.gov/geo/ ). The GSE6477 dataset was derived from the GPL96 (HG-U133A) Affymetrix Human Genome U133A Array, while the GSE16558 dataset was derived from the GPL6244 (HuGene-1_0-st) Affymetrix Human Gene 1.0 ST Array. GSE6477 contains bone marrow CD138 + samples from 125 patients with MM and 15 normal bone marrow CD138 + controls, while GSE16558 includes bone marrow CD138 + samples from 60 patients with MM and 5 normal bone marrow CD138 + controls. The bulk data were preprocessed using R software (version 4.2.3), with batch effects removed using the ‘SVA’ package and normalization performed using the ‘normalize’ package for subsequent analysis. All datasets included the corresponding clinical information for both patients with MM and normal individuals.
Screening of differentially expressed genes (DEGs)
DEGs between MM and normal bone marrow were identified in the integrated GSE6477 and GSE16558 dataset using a corrected P-value <0.05 and |log 2 fold change (FC)|>1 as screening criteria. The ‘limma’ package was used for DEG analysis and the intersection of aging-related genes (ARGs) was further identified to obtain aging-related DEGs (ARDEGs). The Wilcoxon rank-sum test was employed to analyze the expression of ARDEGs between MM and normal samples. Cluster analysis was performed on the selected DEGs, and visualizations including heatmaps, volcano plots and box plots were generated using the ‘pheatmap’ and ‘ggplot2’ packages.
Protein-protein interactions (PPI) and correlation analysis of ARDEGs
To examine the interactions between ARDEGs, the Search Tool for the Retrieval of Interacting Genes and proteins (STRING) database ( https://string-db.org/ ) was utilized. The interaction network was constructed with a confidence score threshold of >0.7. The PPI network was visualized and further analyzed using Cytoscape software (version 3.8.1; http://cytoscape.org/ ). The Pearson correlation was calculated using the ‘correlation graph’ function in R software to identify the relationships between ARDEGs.
Gene Ontology (GO) and pathway enrichment analysis of ARDEGs
GO, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Disease Ontology (DO) enrichment analyses were performed using the ‘GOplot’, ‘KEGGplot’ and ‘DOplot’ packages in R software. GO analysis was performed for the categories biological process, cellular component and molecular function, while KEGG analysis examined the signaling pathways of ARDEGs. DO analysis was used to explore disease enrichment. A P<0.05 was considered indicative of significantly differential enrichment of DEGs.
Machine learning screening for ARGs
LASSO regression analysis of ARDEGs was conducted using the ‘glmnet’ package on the integrated GSE6477 and GSE16558 datasets. The SVM-RFE algorithm was applied using the ‘e1071’ package and the ‘caret’ package was used to intersect the results of the two algorithms to identify ARGs in MM.
Evaluating the potential value of candidate genes
To validate the potential diagnostic value of the candidate genes, receiver operating characteristic (ROC) curves were plotted and the area under the ROC curve (AUC) was calculated using the ‘pROC’ package.
Validation of ARGs in other datasets
The GSE39754 and GSE5900 datasets were used to validate the expression patterns of the candidate genes. GSE39754 was derived from the GPL5175 (HuEx-1_0-st) Affymetrix Human Exon 1.0 ST Array, while GSE5900 was derived from the GPL570 (HG-U133_Plus_2) Affymetrix Human Genome U133 Plus 2.0 Array. Both datasets included bone marrow CD138 + samples from 170 patients with MM and 28 normal bone marrow CD138 + controls, providing a basis for evaluating the diagnostic efficacy of the candidate genes. Clinical information for both patients with MM and normal individuals was included in all datasets.
Cell lines and culture
The cell lines GM12878 (human normal B lymphocyte cell line), U266 (human MM lymphocyte-like cell line) and RPMI8226 (human MM peripheral blood B lymphocyte cell line) were obtained from the Shanghai Academy of Biological Sciences. U266 and GM12878 were designated as the MM group and GM12878 was designated as the control group. All cell lines underwent short tandem repeat authentication and mycoplasma testing, conducted by the Shanghai Academy of Biological Sciences. These cell lines were cultured in complete RPMI 1640 medium supplemented with 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.) and 10% fetal bovine serum (FBS Gibco; Thermo Fisher Scientific, Inc.). Cell incubations were performed under strictly regulated conditions of 37°C and 5% CO 2 in a humidified incubator.
RNA extraction and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
Total RNA was extracted using TRIzol reagent (Takara Bio, Inc.), and its concentration, purity and integrity were evaluated with a NanoDrop™2000 spectrophotometer (Thermo Fisher Scientific, Inc.). According to the manufacturer's instructions, RT was performed using 1 µg of total RNA, processed with HiScript II Q RT SuperMix for qPCR (+gDNA wiper) (Vazyme Biotech Co., Ltd.) along with a gDNA eraser (Vazyme Biotech Co., Ltd.). The concentration, purity and integrity of the resulting cDNA were subsequently assessed using the same NanoDrop spectrophotometer. Real-time qPCR was carried out using SYBR Green MasterMix (cat. no. 11203ES50; Yeasen Biotechnology Co., Ltd.) and StepOne Software v.2.3 (Applied Biosystems; Thermo Fisher Scientific, Inc.).
The reaction conditions were as follows: Pre-denaturation at 95°C for 5 min, followed by 40 cycles of denaturation at 95°C for 30 sec, annealing at 55°C for 30 sec and extension at 72°C for 30 sec, with a final extension at 72°C for 10 min, with three biological replicates for each sample. Data analysis was conducted using the 2 −∆∆Cq method ( 15 ), normalizing against the expression levels of the reference gene GAPDH. The primer sequences used in the RT-qPCR assays are provided in Table I .
Statistical analysis
All statistical analyses in this bioinformatics study were performed using R software (version 4.2.2). Pearson correlation analysis was conducted to explore associations between different variables. A P-value or false discovery rate threshold of <0.05 was set as the criterion for statistical significance. For the experimental component, statistical analyses were conducted using GraphPad Prism (version 8.0.2; Dotmatics), with each experiment including at least three biological replicates. Values are expressed as the mean ± SD. Differences between the two datasets were evaluated using either two-way ANOVA (no post-hoc test was performed) or Student's t-test, with P<0.05 considered statistically significant.
Results
Identification of ARDEGs based on MM-retrospective analysis
After preprocessing the integrated GSE6477 and GSE16558 dataset, differential gene analysis was conducted on bone marrow CD138 + samples from 185 patients with MM and 20 normal CD138 + samples using the merged data (from GSE6477 and GSE16558 ). A correction with criteria of P<0.05 and |log 2 FC|>1 yielded 454 DEGs, which are depicted in the volcano plot and heatmap ( Fig. 1A and B ). Additionally, principal component analysis (PCA) was performed to assess the overall variance and clustering of samples, as shown in Fig. S1 . By intersecting these DEGs with the Aging-Associated Gene Library, 19 ARGs linked to MM were identified, including A2M, APOE, PIN1, FOS, KCNA3, GCLM, IRS1, PPARG, HIF1A, RB1, CETP, EGR1, IL7R, CAT, DGAT1, TXN, C1QA, MYC and JUN ( Fig. 1C ). Among these, 5 genes were upregulated, while 14 genes were downregulated in patients with MM compared to normal controls. The box plot illustrates significant expression differences of these 19 MM-related aging genes between MM and normal samples ( Fig. 1D ). PIN1, TXN and MYC were the top three upregulated genes, whereas A2M, C1QA and HIF1A were the top three downregulated genes ( Table II ).
PPI network and correlation analysis of ARDEGs
PPI analysis was performed to examine the interactions among the 19 ARDEGs ( Fig. 2A ), and the number of genes interacting with each other was quantified ( Fig. 2B ). The analysis revealed that 18 genes interacted with at least one other gene. PPARG interacted with 12 genes, JUN with 11, and both MYC and HIF1A interacted with 10 genes. Correlation analysis of the expression levels of the 18 interacting genes showed significant associations between them. Notably, the expression levels of C1QA, APOE and A2M were significantly positively correlated (r>0.6). The top three genes associated with C1QA were APOE (r=0.81), A2M (r=0.69) and CETP (r=0.56). For TXN, the top three associated genes were JUN (r=−0.41), RB1 (r=−0.41) and KCNA3 (r=−0.35) ( Fig. 2C ).
Enrichment analysis of ARDEGs
GO, KEGG and DO enrichment analyses were conducted using R software to determine the potential biological functions of the ARDEGs. The most significant GO enrichment terms included the ‘positive regulation of miRNA transcription’, ‘regulation of miRNA transcription’ (biological process), ‘RNA polymerase II transcription regulator complex’, ‘high-density lipoprotein particle’, ‘plasma lipoprotein particle’ (cellular component), ‘DNA-binding transcription factor binding’, ‘RNA polymerase II-specific DNA-binding transcription factor binding’ and ‘transcription co-regulator binding’ (molecular function) ( Fig. 3A ). KEGG enrichment analysis revealed that the ARDEGs were primarily involved in signaling pathways related to Kaposi sarcoma-associated herpesvirus infection and human T-cell leukemia virus 1 infection, suggesting their role in the dynamic regulation of the immune system ( Fig. 3B ). DO enrichment analysis showed that ARDEGs are crucial in the occurrence and development of hepatitis ( Fig. 3C ).
Screened candidate genes of MM
The LASSO regression algorithm identified 9 genes among the 19 MM ARDEGs: APOE, FOS, KCNA3, IRS1, PPARG, HIF1A, CAT, TXN and JUN ( Fig. 4A ). In comparison, the SVM-RFE algorithm identified 10 genes: TXN, RB1, A2M, FOS, JUN, GCLM, CAT, KCNA3, HIF1A and IL7R ( Fig. 4B ). A total of six candidate genes, overlapping between the two algorithms, were selected: FOS, KCNA3, HIF1A, CAT, TXN and JUN ( Fig. 4C ). Notably, all these 6 molecules are closely associated with aging.
ROC curves of six candidate genes in MM and normal bone marrow
The ‘pROC’ package was used to analyze the expression of six candidate genes in 185 MM and 20 normal bone marrow samples from the GSE6477 and GSE16558 datasets, generating ROC curves. The AUC, which combines sensitivity and specificity, was used to assess the diagnostic effectiveness of these genes. The six candidate genes demonstrated high diagnostic value for MM. Among them, TXN exhibited the highest diagnostic value in MM samples (AUC=0.941). The diagnostic values for the other genes were as follows: JUN (AUC=0.894), FOS (AUC=0.890), HIF1A (AUC=0.862), CAT (AUC=0.850) and KCNA3 (AUC=0.839) ( Fig. 5A-F ).
Validation of candidate genes in other datasets
The GSE39754 and GSE5900 datasets were used as a validation cohort to assess the accuracy of the analysis results and the expression levels of the six candidate genes. As shown in Fig. 6A-F , a significant difference in TXN and HIF1A expression was observed between MM and normal samples. TXN was overexpressed and HIF1A expression was reduced in MM samples. Notably, HIF1A expression was more variable in the MM samples. Additionally, the ROC curve analysis in the validation cohort ( Fig. 6G-L ) showed that TXN (AUC=0.707), JUN (AUC=0.527), FOS (AUC=0.503), HIF1A (AUC=0.728), CAT (AUC=0.569) and KCNA3 (AUC=0.589) exhibited diagnostic value. Among these, TXN had the highest diagnostic value and can serve as a key diagnostic marker for MM.
Verification of the key genes in the human MM cell line
To validate the bioinformatics results, RT-qPCR was performed on the human MM cell lines U266 and RPMI8226 to further analyze TXN expression. The results from the U266 and RPMI8226 cell validation tests were consistent with the bioinformatics analysis, showing significantly higher expression levels of TXN compared to normal cells. Specifically, the expression of TXN was significantly greater in U266 (P<0.01) and RPMI8226 (P<0.01) cells than in normal cells ( Fig. 7 ).
Discussion
In the present study, bioinformatics analysis and machine learning techniques were utilized to identify aging-related diagnostic and druggable targets for patients with MM. The results demonstrated that TXN could serve as a robust diagnostic biomarker for detecting MM pathogenesis. Additionally, the potential biological and molecular functions of TXN in MM were explored, offering novel insights into the disease's underlying mechanisms.
Aging is the primary risk factor for most malignancies ( 5 ). For instance, MM is an aging-related disease that poses significant health challenges for the elderly ( 16 ). Epidemiological investigation pointed out the global burden of MM is increasing in numerous countries due to an aging population ( 17 ). Studies have highlighted aging-associated epigenetic and genetic alterations that characterize both MM development and normal aging processes ( 9 ). For example, aging can destroy the immune systems, which leads to the impairment of immune surveillance and pathogenesis of MM ( 18 ). Besides, enhancer of zeste homolog 2 inhibition induces senescence via the ERK1/2 signaling pathway in MM ( 19 ). Additionally, inositol-requiring enzyme 1α inhibition enhances mitochondrial quality in CD8 + T cells, thereby improving their anti-tumor efficacy ( 20 ). Furthermore, genetic instability in patients with MM contributes to an anti-senescence effect, promoting disease progression ( 21 ). Age-related mesenchymal stromal cell senescence has also been linked to the progression from monoclonal gammopathy of undetermined significance to MM ( 22 ). In the present study, 6 molecules were identified (FOS, KCNA3, HIF1A, CAT, TXN and JUN) that are closely associated with aging. The m6A mRNA demethylase FTO in granulosa cells can slow FOS-dependent ovarian aging ( 23 ). KCNA3 is implicated in cognitive aging ( 24 ). P53/HIF-1α regulates neuronal aging and autophagy in spinal cord ischemia/reperfusion injury ( 25 ). CAT contributes to autophagy-induced vascular smooth muscle cell senescence ( 26 ). Additionally, Astragalus polysaccharide alleviates oxidative stress and senescence in osteoarthritis chondrocytes via the GCN2/ATF4/TXN axis ( 27 ). Activation of JUN has long been considered a hallmark of cellular senescence ( 28 ). The present study confirmed that TXN can serve as a promising diagnostic biomarker for MM, a malignant blood cancer characterized by uncontrolled plasma cell proliferation. TXN plays a pivotal role in the pathophysiology of MM, affecting both the cellular aging process and the maintenance of cellular redox balance. This study further explored the novel mechanisms through which TXN contributes to MM, providing valuable insights into the disease's pathophysiology. Additionally, the TXN-ALDH1L2 axis has been shown to promote the progression of colorectal cancer and radioresistance by activating the NF-κB signaling pathway ( 14 ), which is also closely associated with MM pathogenesis ( 29 ). Thus, TXN may regulate NF-κB signaling to modulate MM progression.
Despite these insights, there are still gaps in our understanding. Specifically, further studies are needed to elucidate the precise mechanisms by which TXN acts in MM, as well as to develop targeted therapeutic strategies based on TXN's regulatory pathways, particularly regarding its role in apoptosis regulation. Furthermore, additional research is required to identify therapeutic agents targeting TXN for MM treatment, which could enhance the clinical applicability of the present findings. The in vitro results obtained from the GM12878 EBV-transformed lymphoblastoid cell line should also be validated in patient samples to ensure authenticity. Besides, ARGs identified in the present study that are also involved in the pathogenesis of other disease or biological functions showed a weaker association with MM pathogenesis, indicating a need for further pre-clinical validation of these DEG mechanisms in MM. Furthermore, differences in the diagnostic performance of TXN between the training and validation sets of patients with MM suggest predictive heterogeneity. Therefore, future studies should assess TXN's diagnostic performance in multi-center cohorts to strengthen its diagnostic robustness. Lastly, the impact of TXN dysregulation on the overall survival of patients with MM should be explored in clinical studies to provide further evidence for the clinical relevance of TXN.
Supplementary Material
Supporting Data
Acknowledgements
Not applicable.
Availability of data and materials
The data generated in the present study may be requested from the corresponding author.
Authors' contributions
HL and SS conducted the research. YWu and YWa performed the data analysis. HL and SS wrote the manuscript. YuL and YiL contributed to the conception and design of the study, provided supervision and critically revised the manuscript. YuL and YiL checked and confirm the authenticity and accuracy of all the raw data. All authors have read and approved the final manuscript.
Ethics approval and consent to participate
Not applicable.
Patient consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
References
1 Kumar SK Rajkumar V Kyle RA van Duin M Sonneveld P Mateos MV Gay F Anderson KC Multiple myeloma Nat Rev Dis Primers 3 17046 2017 10.1038/nrdp.2017.46 28726797
2 Palumbo A Anderson K Multiple myeloma N Engl J Med 364 1046 1060 2011 10.1056/NEJMra1011442 21410373
3 Rajkumar SV Multiple myeloma: Every year a new standard? Hematol Oncol 37 (Suppl 1) 62 65 2019 10.1002/hon.2586 31187526 PMC6570407
4 Raab MS Podar K Breitkreutz I Richardson PG Anderson KC Multiple myeloma Lancet 374 324 339 2009 10.1016/S0140-6736(09)60221-X 19541364
5 De Winter G Aging as disease Med Health Care Philos 18 237 243 2015 10.1007/s11019-014-9600-y 25240472
6 Hou Y Dan X Babbar M Wei Y Hasselbalch SG Croteau DL Bohr VA Ageing as a risk factor for neurodegenerative disease Nat Rev Neurol 15 565 581 2019 10.1038/s41582-019-0244-7 31501588
7 Fulop T Larbi A Douziech N Levesque I Varin A Herbein G Cytokine receptor signalling and aging Mech Ageing Dev 127 526 537 2006 10.1016/j.mad.2006.01.025 16530252
8 Kanagal-Shamanna R Beck DB Calvo KR Clonal hematopoiesis, inflammation, and hematologic malignancy Annu Rev Pathol 19 479 506 2024 10.1146/annurev-pathmechdis-051222-122724 37832948
9 Urban VS Cegledi A Mikala G Multiple myeloma, a quintessential malignant disease of aging: A geroscience perspective on pathogenesis and treatment Geroscience 45 727 746 2023 10.1007/s11357-022-00698-x 36508077 PMC9742673
10 Ullrich F Bröckelmann PJ Turki AT Khan AM Chiru ED Vetter M Tresckow BV Wirth R Cordoba R Ortiz-Maldonado V Impact of immunological aging on T cell-mediated therapies in older adults with multiple myeloma and lymphoma J Immunother Cancer 12 e009462 2024 10.1136/jitc-2024-009462 39622581 PMC11624774
11 Visram A Kourelis TV Aging-associated immune system changes in multiple myeloma: The dark side of the moon Cancer Treat Res Commun 29 100494 2021 34837796 10.1016/j.ctarc.2021.100494 PMC8744501
12 Zweegman S Palumbo A Bringhen S Sonneveld P Age and aging in blood disorders: Multiple myeloma Haematologica 99 1133 1137 2014 10.3324/haematol.2014.110296 24986875 PMC4077073
13 Yang B Zhao S Polydatin regulates proliferation, apoptosis and autophagy in multiple myeloma cells through mTOR/p70s6k pathway Onco Targets Ther 10 935 944 2017 10.2147/OTT.S123398 28243129 PMC5317338
14 Yu L Guo Q Luo Z Wang Y Weng J Chen Y Liang W Li Y Zhang Y Chen K TXN inhibitor impedes radioresistance of colorectal cancer cells with decreased ALDH1L2 expression via TXN/NF-κB signaling pathway Br J Cancer 127 637 648 2022 10.1038/s41416-022-01835-1 35597868 PMC9381770
15 Livak KJ Schmittgen TD Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) method Methods 25 402 408 2001 10.1006/meth.2001.1262 11846609
16 Fotiou D Ntanasis-Stathopoulos I Gavriatopoulou M Dimopoulos MA Multiple myeloma: Current and future management in the aging population Maturitas 138 8 13 2020 10.1016/j.maturitas.2020.04.015 32631590
17 Mafra A Laversanne M Marcos-Gragera R Chaves HVS Mcshane C Bray F Znaor A The global multiple myeloma incidence and mortality burden in 2022 and predictions for 2045 J Natl Cancer Inst 117 907 914 2025 10.1093/jnci/djae321 39658225
18 Casey M Nakamura K The cancer-immunity cycle in multiple myeloma Immunotargets Ther 10 247 260 2021 10.2147/ITT.S305432 34295843 PMC8291851
19 Guo S Tang Q Gao X Hu L Hu K Zhang H Zhang Q Lai Y Liu Y Wang Z EZH2 inhibition induces senescence via ERK1/2 signaling pathway in multiple myeloma Acta Biochim Biophys Sin (Shanghai) 56 1055 1064 2024 10.3724/abbs.2024077 38804044 PMC11322866
20 Wan Y Wang J Chen M Wang J Nan F Huang H Liu Z Hou J Dual roles of IRE1α inhibition in reversing mitochondrial ROS-induced CD8(+) T-cell senescence and exerting direct antitumor effects in multiple myeloma J Immunother Cancer 13 e011044 2025 10.1136/jitc-2024-011044 40425229 PMC12107582
21 Alagpulinsa DA Szalat RE Poznansky MC Shmookler Reis RJ Genomic instability in multiple myeloma Trends Cancer 6 858 873 2020 10.1016/j.trecan.2020.05.006 32487486
22 Plakhova N Panagopoulos V Cantley MD Trainor LJ Hewett DR Clark KC Gardiner J Yong A Lee C Horvath N Age-related mesenchymal stromal cell senescence is associated with progression from MGUS to multiple myeloma Leukemia 39 1464 1475 2025 10.1038/s41375-025-02621-7 40263435 PMC12133572
23 Jiang ZX Wang YN Li ZY Dai ZH He Y Chu K Gu JY Ji YX Sun NX Yang F Li W The m6A mRNA demethylase FTO in granulosa cells retards FOS-dependent ovarian aging Cell Death Dis 12 744 2021 10.1038/s41419-021-04016-9 34315853 PMC8316443
24 Abdul Sani NF Amir Hamzah AIZ Abu Bakar ZH Mohd Yusof YA Makpol S Wan Ngah WZ Damanhuri HA Gene expression profile in different age groups and its association with cognitive function in healthy malay adults in Malaysia Cells 10 1611 2021 10.3390/cells10071611 34199148 PMC8304476
25 Liu X Wang J Shen K Jin W p53/HIF-1α regulates neuronal aging and autophagy in spinal cord ischemia/reperfusion injury Mech Ageing Dev 222 112000 2024 10.1016/j.mad.2024.112000 39515667
26 Salazar G Cullen A Huang J Zhao Y Serino A Hilenski L Patrushev N Forouzandeh F Hwang HS SQSTM1/p62 and PPARGC1A/PGC-1alpha at the interface of autophagy and vascular senescence Autophagy 16 1092 1110 2020 10.1080/15548627.2019.1659612 31441382 PMC7469683
27 Chen K Qin C Ji C Yu Y Wu Y Xu L Jiang Y Zou G Astragalus polysaccharide alleviates oxidative stress and senescence in chondrocytes in osteoarthritis via GCN2/ATF4/TXN axis Int J Biol Macromol 310 (Pt 3) 143285 2025 10.1016/j.ijbiomac.2025.143285 40253033
28 Hernandez-Segura A Nehme J Demaria M Hallmarks of Cellular Senescence Trends Cell Biol 2018 28 436 453 2018 10.1016/j.tcb.2018.02.001 29477613
29 Matthews GM de Matos Simoes R Dhimolea E Sheffer M Gandolfi S Dashevsky O Sorrell JD Mitsiades CS NF-κB dysregulation in multiple myeloma Semin Cancer Biol 39 68 76 2016 10.1016/j.semcancer.2016.08.005 27544796