Integrated transcriptomics of multiple sclerosis peripheral blood mononuclear cells identified COPG1, RPN1, and KDM3B as potential biomarkers

doi:10.21203/rs.3.rs-4685812/v1

Download PDF

Article

Integrated transcriptomics of multiple sclerosis peripheral blood mononuclear cells identified COPG1, RPN1, and KDM3B as potential biomarkers

https://doi.org/10.21203/rs.3.rs-4685812/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Among diagnostic techniques, RNA biomarkers have been poorly investigated for multiple sclerosis (MS). In this study, by the integration of GSE21942 and GSE203241 microarray profiles of peripheral blood mononuclear cells, potential biomarkers were explored. A comparison between 28 MS patients and 23 healthy controls led to the identification of 71 upregulated and 35 downregulated genes. Immune-related functional terms, particularly pathways linked to lymphocyte activation, were enriched with the differentially expressed genes (DEGs). Subsequently, key mRNAs and miRNAs were detected regarding their number of interactions in the miRNA-mRNA regulatory network. Weighted gene co-expression network analysis (WGCNA) detected a gene module highly enriched for neurodegenerative disorders. Central genes in the protein-protein interaction (PPI) network of this module were genes encoding various subunits of the respiratory chain complexes. 59 genes selected from converging results of differential expression analysis and WGCNA underwent machine learning methods and receiver operating characteristic (ROC) analysis. COPG1, RPN1, and KDM3B were subsequently identified as potential biomarkers based on their acceptable diagnostic efficacy in the integrated data, as well as in both GSE141804 and GSE146383 datasets as validation sets.

Multiple sclerosis

Biomarkers

PBMCs

COPG1

RPN1

KDM3B

Multiple sclerosis (MS) is characterized by neuroinflammation and demyelination of the central nervous system (CNS) (1). Although the exact etiology of MS remains undetermined, autoimmunity is widely acknowledged as the underlying mechanism (2). Immune cells in the periphery are implicated in the autoimmune response during progressive and relapsing-remitting courses of MS (3), particularly autoreactive T lymphocytes, which centrally mediate the destruction of the myelin sheath (4).

After the activation in the periphery, T-cells pass through the blood-brain barrier (BBB), reaching the CNS where they undergo further reactivation through interactions with local antigen-presenting cells (APCs) and autoantigen recognition (5). Among T-cells, CD4 + lineages, including Th1, Th17 and Th22 significantly contribute to the MS pathogenesis by releasing IFN-γ, IL-17 and IL-22, respectively. These cytokines eventually lead to the recruitment of other immune cells, inflammation, and increased permeability of BBB (6). CD8 + cells, comprising the largest proportion of T-cells in the CNS lesions. They express TNF upon activation, which can directly harm neurons, and disrupt the restoration of myelin sheath (7). B lymphocytes not only produce autoantibodies but also, more significantly act as APCs and express cytokines such as IL-6 and TNF-α (8). Furthermore, emerging research indicates the involvement of innate immunity cells like NK-cells (9), monocytes (10), and neutrophils (11) in the immunopathology of MS.

Magnetic resonance imaging (MRI) stands as the most common diagnostic procedure for MS (12). However, it was recently reported that overdependence on the technique can lead to an increased number of misdiagnoses (13). Exploration of peripheral blood mononuclear cells (PBMCs) transcriptome has the potential to identify potential biomarkers in neurodegenerative disorders such as Alzheimer’s and Parkinson’s disease (14, 15).

This study aimed to identify key genes and modules in MS through the integration of PBMCs gene expression data. After integrating GSE21942 and GSE203241, a total of 106 differentially expressed genes (DEGs) were identified. Weighted gene co-expression network analysis (WGCNA) showed that 59 genes had a correlation consistent with the direction of their expression levels. Moreover, the analysis singled out a module highly enriched in neurodegenerative disorders. Both random forest (RF) and least absolute shrinkage and selection operator (LASSO) methods underlined COPG1 as the most important gene. Receiver operating characteristic (ROC) analysis of each dataset revealed an acceptable diagnostic efficiency of COPG1. The result was subsequently replicated in GSE141804 and GSE146383.

Differentially expressed genes and functional enrichment

As Figs. 1a-c reveal, samples clustered regarding the disease state rather than the study following the ComBat batch-effect removal. Among 13237 analyzed genes, a total of 106 DEGs, including 71 upregulated- and 35 downregulated genes were identified (Fig. 1d). The complete list of log fold changes (LogFC) and adjusted p-values for the 71 upregulated genes and 35 downregulated genes can be found in Supplementary Table S1 and Supplementary Table S2, respectively, online. By conducting the Gene Ontology (GO) test, the top 10 terms in each GO classification, including biological processes, cellular components, and molecular functions were visualized (Fig. 1e). The DEGs were significantly enriched in the immune response-regulating cell surface receptor signaling pathway (GO:0002768), immune response-activating signaling pathway (GO:0002757), and immune response-activating cell surface receptor signaling pathway (GO:0002429) biological processes. Similarly, other biological processes reflected the involvement of the DEGs in orchestrating autoimmune processes such as differentiation of lymphocyte and mononuclear cell (GO:0030098 and GO:1903131, respectively), inflammation (GO:0050727), and reactive oxygen species metabolic process (GO:0072593). In the cellular components category, the external side of plasma membrane (GO:0009897) and blood microparticle (GO:0072562) were the most prominent terms. Tetrapyrrole binding (GO:0046906) and cytokine activity (GO:0005125) emerged as the key terms in the molecular function classification. In line with gene ontology findings, B cell receptor signaling pathway was the most enriched KEGG pathway, followed by the hematopoietic cell lineage, MAPK signaling pathway, and Th1, Th2, and Th17 cell differentiation (Fig. 1f).

miRNA-mRNA regulatory network

The inclusion criteria of having a minimum of 2 interactions for mRNAs and 3 interactions for miRNAs were applied to construct the regulatory network. CD69 and miR-493-5p had the highest number of interactions among mRNAs and miRNAs, respectively. Meanwhile, miR-124-3p, miR-181a-5p, miR-181b-5p, miR-181c-5p, miR-181d-5p, miR-450b-5p, and miR-506-3p were the other key miRNAs. Although the COPG1, RPN1, and STT3A genes have not interacted with mRNAs in the main network, they are connected to the network through miR-124-3p and miR-506-3p, which are key regulators of the network (Fig. 2).

Identification of MS-specific modules by WGCNA

After selecting 11 as the optimal soft-thresholding power (Figs. 3a-b), blockwise module construction assigned 13237 genes into 15 modules (Fig. 3c). Modules with an absolute correlation value above 0.4 and an adjusted p-value less than 0.01 were considered significant. These filtrations led to the identification of 3 positively correlated and 5 negatively correlated modules.

Among the positively correlated modules, module 9 was enriched in the FoxO signaling pathway. Meanwhile, B-cell receptor signaling pathway, hematopoietic cell lineage, and primary immunodeficiency were the most significant KEGG pathways associated with module 12. Pathway enrichment analysis failed to identify any pathway with an adjusted p-value less than 0.05 for module 14.

Within the negatively correlated modules, module 7 was highly enriched in neurodegenerative disorders, therefore it was singled out for further analysis. Module 1 was functionally associated with lysosome and phagosome pathways. The ubiquitin mediated proteolysis and phosphatidyl inositol signaling system were the most prominent biological pathways in which genes of module 2 were involved. Modules 3 and 8 were significantly enriched in the MAPK signaling pathway and protein processing in the endoplasmic reticulum, respectively (Fig. 3d).

To identify the potential biomarkers, 59 genes, including 31 upregulated and 28 downregulated genes observed in the positively and negatively correlated modules, respectively, underwent machine-based learning methods (Fig. 3e). The entire WGCNA statistics for the 59 shared genes between the analyses can be found in Supplementary Table S1 and Supplementary Table S2 online.

Identification of hub genes in the module involved in neurodegeneration

Of the 250 genes in module 7, 23 were selected as highly connected genes because they stand in the top 10% of genes in module 7 based on their degree (16). The cutoff was set because the number of interactions dropped significantly from 24 for NDUFB2 to 14 for ATP5PF (Fig. 4a).

In addition to high connectivity, hub genes also meet the criteria of having a module membership above 0.8 and a gene significance for the trait of interest exceeding 0.2 (17). By applying these cutoff thresholds, 11 genes, namely COX5B, UQCRQ, NDUFS3, UQCR10, NDUFA4, COX7B, NDUFA7, NDUFS7, NDUFA2, NDUFA1, and NDUFB2, were determined as hub genes (Fig. 4b).

The 23 central genes in the module 7 PPI network were mainly involved in the mitochondrial function and structure, revealed by gene ontology analysis (Fig. 4c). Moreover, the subcategory of neurodegenerative disorders, including Parkinson's, prion, Huntington's, Alzheimer's, and amyotrophic lateral sclerosis, contributed to the largest proportion of KEGG pathways enriched with these genes (Fig. 4d).

Identification of potential biomarkers using machine-based learning models

The 59 genes identified by both differential expression and WGCNA analyses, were subjected to RF and LASSO methods. Interestingly, COPG1 was evaluated as the most important gene by both analyses, reflecting its significance and potential as a biomarker (Fig. 5).

According to the RF results, RPN1 and KDM3B stood in second and third places, respectively, based on the MeanDecreaseGini and MeanDecreaseAccuracy values. However, in LASSO, ADAM28 was the second most prominent gene, followed by SMG8, which had a coefficient markedly lower than that of ADAM28 (Fig. 5e).

Biomarker validation

By conducting ROC analysis, the diagnostic efficacy of COPG1, RPN1, KDM3B, and ADAM28 was assessed on GSE21942 and GSE203241, as well as on GSE141804 and GSE146383 as external datasets. As Fig. 6 reveals, COPG1, RPN1, and KDM3B had an area under the receiver operating characteristic curve above 0.7 in all the examined datasets. Conversely, ADAM28 failed to pass the threshold of 0.7 in GSE141804 and GSE146383 (Fig. 6).

Despite recent advancements in the diagnosis of MS, the exploration of non-invasive biomarkers could effectively increase the accuracy of diagnosis. To this end, in an integrated microarray study of peripheral blood mononuclear cells from MS patients and control subjects, potential biomarkers were sought. A total of 106 genes were found to be differentially expressed which were functionally enriched in well-established pathways involved in the pathogenesis of MS. Gene ontology analysis reflected the contribution of DEGs to various aspects of immune system functions. Highly enriched biological processes were predominantly engaged in lymphocyte differentiation, production of immune mediators, and regulatory signaling pathways. Most significantly, immune response-regulating cell surface receptor signaling pathways, such as TLR (18), JAK/STAT (19), PD-1/PD-L1 (20), and TGFβ (21) are implicated in the immune responses in MS.

Similarly, the DEGs were associated with numerous immune-related components and complexes. The most significant cellular component term, the external side of the plasma membrane, has been reported to be enriched with common DEGs in neurodegenerative disorders (22). In line with the ranking of blood microparticles as the second most significant term in this category, elevated levels of blood microparticles originating from endothelial and platelet cells were observed in different types of MS, disrupting the barrier function of endothelial cells in vitro (23). In the molecular function category, tetrapyrrole binding and cytokine activity were the most prominent terms, followed by terms mainly associated with immune response and molecular signaling. Tetrapyrrole, heme, and haptoglobin binding-related terms were similarly enriched with genes of hemoglobin complex, including HBA1, HBD, HBG1, and HBB. Although the exact role of these genes in MS pathogenesis has yet to be elucidated, recent advancements in the field have hypothesized that these genes might be involved in functions beyond oxygen transport, such as the role of HBD in counteracting oxidative stress (24). KEGG pathways meeting the criterion of an adjusted p-value of < 0.05 were further enriched in pathways involved in autoimmune disorders such as type I diabetes mellitus, rheumatoid arthritis, inflammatory bowel disease, graft versus host disease, and allograft rejection, as well as infectious diseases, specifically Epstein-Barr virus infection, leishmaniasis, measles, and malaria. Moreover, immune-related signaling pathways highlighted the pivotal role of lymphocytes in MS, particularly Th1, Th2, Th17, and B cells. The flow cytometry analysis of peripheral blood mononuclear cells from MS patients identified MAPK signaling, the third most enriched pathway in this study, as a key pathway responsible for the multiplication and survival of immune cells, particularly B-lymphocytes (25).

Neurodegeneration in the CNS is influenced by miRNAs, which affect the development of lymphocytes and, in turn, play a key role in MS pathogenesis (26). Therefore, screening of key miRNAs was conducted by the construction of miRNA-mRNA regulatory network. In the interaction network, CD69, a conventionally acknowledged marker for T-cell activation expressed shortly after TCR and cytokine stimulation (27), exhibited the highest number of interactions. This transmembrane protein is expressed in various peripheral blood mononuclear cells, such as B- and T-lymphocytes, NK cells, and monocytes (28). Considering high connectivity as a measure of hub gene identification, CD69, PAX5, DUSP1, FASLG, and IFNG were the hub genes in the regulatory network, being in the top 10% of connectivity.

Compared to healthy controls, miR-493-5p elevation was observed in the peripheral blood leucocytes of untreated RRMS patients (29). Other key miRNAs, each interacting with four mRNAs, included miR-124-3p, miR-181a-5p, miR-181b-5p, miR-181c-5p, miR-181d-5p, miR-450b-5p, and miR-506-3p. Dysregulation of miR-124-3p has been observed in a wide spectrum of neurodegenerative disorders, including MS and Parkinson's disease. It has also been found to have therapeutic potential, as upregulation of miR-124-3p can inhibit the activation of macrophages and microglia, which are involved in the neuroinflammation in MS (30, 31).

Four members of miR-181 Family, including miR-181a-5p, miR-181b-5p, miR-181c-5p, and miR-181d-5p were of the key regulators of the DEGs. These miRNAs targeted CD69 and PAX5, the two genes with the highest number of interactions and considered hub genes in the network. The current body of literature demonstrates the altered expression of the miR-181 family in neurodegenerative disorders, including MS, Parkinson's disease, and Alzheimer's disease (32). Quantitative real-time PCR (qPCR) of PBMCs identified miR-181a-5p and miR-181b-5p as differentially downregulated and upregulated miRNAs, respectively, in MS patients (33, 34). Moreover, targeting genes involved in neuroinflammation, such as MAP2K1, CREB1, ATXN1, and ATXN3, by miR-181a-5p reflects its therapeutic potential (34). The expression of miR-181c-5p is reported to be misregulated in PBMCs, cerebrospinal fluid (CSF), and serum of relapsing-remitting multiple sclerosis (RRMS) patients (35). A study on sporadic and familial forms of MS found miR-181d-5p to be downregulated in PBMCs, as revealed by qPCR analysis (36).

Assessment of miR-450b-5p serum levels in RRMS and SPMS patients revealed its reduction and suggested it as a potential biomarker for disease progression (37). Evaluation of miR-506-3p expression levels in the peripheral blood of RRMS individuals using qPCR following fingolimod treatment, an immune-regulating therapy approved for these patients, revealed no significant alterations in its levels in MS patients and responders compared to healthy controls and non-responders, respectively (38).

Assigning 13,237 genes into 15 modules using WGCNA resulted in the identification of a set of co-expressed genes, Module 7, which was negatively correlated with MS (correlation = -0.53, p-value < 0.01) and highly enriched in neurological disorders. Functional enrichment analysis of 23 highly connected genes in the PPI network of this module demonstrated that mitochondrial functioning and pathways of neurodegenerative disorders were fully enriched with these genes.

Out of 23 highly connected genes in this module, 11 were selected as hub genes, including COX5B, UQCRQ, NDUFS3, UQCR10, NDUFA4, COX7B, NDUFA7, NDUFS7, NDUFA2, NDUFA1, and NDUFB2, based on the criteria of module membership > 0.8 and gene significance > 0.2. These hub genes encode different subunits of electron transport chain complexes, including NADH oxidoreductase (Complex I), ubiquinol-cytochrome c reductase (Complex III), and cytochrome c oxidase (Complex IV), supporting the evidence of defect in mitochondrial functioning in PBMCs of MS patients.

Mitochondrial dysfunctions have been widely established as a contributor to MS (39). However, at the peripheral level, the reduction of mitochondrial proteins and impairment of this organelle have been reported relatively recently (40). According to the current literature on MS, the downregulation of COX5B (41), NDUFS3, and NDUFA4 (42), has been documented in blood samples, while a downward trend in the expression of COX7B (43), NDUFA4, NDUFB2, and UQCRQ (44), has been reported only in brain lesions. A genome-wide association study (GWAS), reported NDUFA7 as a differentially methylated gene in both immune cells and neurons (45). Another association study found NDUFA7, NDUFS5, and NDUFS7 genes to be associated with MS (46).

Among the 59 genes selected from coordinated results of differential expression analysis and WGCNA (Fig. 3E), COPG1, RPN1, and KDM3B were selected as potential biomarkers based on the converging results of machine-based learning methods and ROC analysis (Figs. 5 and 6).

COPG1 is the gamma subunit encoding gene of the COPI complex, which is involved in vesicle transport from the Golgi to the endoplasmic reticulum (ER) and intra-Golgi traffic (47). In a bioinformatics study on MS, COPG1 was ascertained as one of the five hub mRNAs in the circRNA-miRNA-hub mRNA network that met the criterion of having a non-zero coefficient in LASSO regression analysis. Moreover, the expression level of COPG1 in MS was markedly lower than that of the control group (48). Similarly, the differential expression analysis of GSE21942 identified COPG1 as one of the most downregulated genes in PBMCs of MS patients (49). In a study on COPA syndrome, which is caused by mutations in the alpha subunit of the COPI complex, applying CRISPR/Cas9 to induce COPG1 deficiency in Th-1 cells resulted in the transcription of inflammatory genes (50). This observation suggests a potential link between COPG1 downregulation and cellular inflammation in PBMCs of MS patients.

RPN1 encodes a regulatory subunit of the 26S proteasome, which is responsible for the ATP-dependent degradation of ubiquitin-marked proteins (51). In neurons, defects in this central catalytic component of the ubiquitin-proteasome system (UPS) lead to the accumulation of misfolded proteins and subsequent neurodegeneration in the CNS (52). Meanwhile, the 26S proteasome is also involved in oxidative stress, transcription of genes, and the release of neurotransmitters. Moreover, it can indirectly affect T-cell development and migration by regulating dopamine release (53). Despite these key regulatory activities of the 26S proteasome, its consideration as a therapeutic target in MS has remained controversial. On the one hand, free proteasomes were detected in the blood specimens of MS patients, and the 26S proteasome was observed to damage the myelin sheath (54). On the other hand, 26S proteasome inhibitors result in the exacerbation of neurodegeneration as a consequence of ubiquitinated protein aggregation (53).

KDM3B encodes lysine demethylase 3B, which is responsible for demethylating histone 3 lysine 9 (H3K9) and is predominantly considered to create a transcriptionally active site (55). KDM3B downregulation found by differential expression analysis in the present study, is in accord with various reports of hypermethylation in PBMCs of MS patients (56). By conducting exome sequencing, pathogenic variants of KDM3B were reported to be associated with a range of cognitive impairments including behavior problems, attention deficit hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and epilepsy (55). Moreover, KDM3B plays a crucial role in mediation of autophagy through the regulation of autophagy-related genes (57). Autophagy, in turn, is involved in different aspects of MS pathogenesis, notably oxidative stress, myelination, and the activation of immune cells. Meanwhile, it's controversial whether inhibition or amplification of autophagy holds therapeutic potential, as its activation in different cell types is followed by contrasting consequences (58).

In conclusion, the results of the present study demonstrated COPG1, RPN1, and KDM3B to have an acceptable diagnostic efficacy across the analyzed datasets. However, further investigations are required to shed light on the mechanisms through which these genes contribute to MS pathogenesis.

Data collection and preparation

The Gene Expression Omnibus (GEO) was reviewed to find datasets incorporating both MS and healthy control PBMCs samples. The details of four retrieved datasets, including GSE21942, GSE203241, GSE141804, and GSE146383, are presented in Table 1. Using principal component analysis (PCA) and hierarchical clustering, GSM545838 from GSE21942, and GSM6165202, GSM6165212, and GSM6165227 from GSE203241 were detected as outlier samples and excluded from the analyses. After integration of GSE21942 and GSE203241, batch-effect removal was conducted using the SVA package in R. Integrated data comprised 51 samples, including 23 healthy controls and 28 MS patients.

Table 1

Microarray datasets in PBMCs of MS and healthy controls
GEO accession No.	No. Control	No. MS	Platform	Reference
GSE21942	15	13	Affymetrix Human Genome U133 Plus 2.0 Array	Unpublished
GSE203241	8	15	Affymetrix Human Genome U133A 2.0 Array	Unpublished
GSE141804	7	15	Affymetrix Human Genome U133A 2.0 Array	Unpublished
GSE146383	46	15	Affymetrix Human Genome U133A 2.0 Array	(59)

Differential expression and functional enrichment analyses

Differentially expressed genes were selected based on an adjusted p-value of less than 0.05 and a |LogFC| exceeding 0.5 using the limma (60) package. Subsequently, the ClusterProfiler (61) package was utilized to find the enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and Gene Ontology (GO) annotations. After excluding terms with an adjusted p-value above the significance threshold of 0.05, the top terms were ordered and visualized according to the proportion of involved genes.

Construction of mRNA-miRNA regulatory network

the STRING database (62) was employed to retrieve the protein-protein interactions (PPI) of the DEGs. The confidence threshold was set at 0.4 and discrete nodes were omitted from the resultant interaction network. Then, DEGs with at least 2 connections in the PPI network were used as an entry in the miRDB (63) to find miRNAs interacting with the provided list. A target prediction score of above 80 and a minimum of 3 interactions with the DEGs were used as criteria for miRNAs to be included in the final network. At last, mRNA-miRNA regulatory network was constructed using the Cytoscape v3.10.2 (64).

Weighted gene co-expression network analysis (WGCNA)

Weighted gene co-expression network analysis was carried out to identify MS-specific modules by the WGCNA (65) package. After the determination of the optimal soft thresholding power, the analysis was performed through a blockwise method. A signed WGCNA analysis was performed, with modules required to have a minimum of 30 genes. Modules were merged if the correlation of their module eigengenes was above 75 percent threshold. Lastly, modules were labeled numerically and pathway analysis was conducted for each module using the Enrichr online tool (66–68).

Moreover, for module 7, which was revealed to be highly enriched in neurodegenerative disorders, PPI network was retrieved from STRING database with a high confidence score (0.7). After excluding genes with no interaction, a network was constructed using the Cytoscape. Hub gene selection was performed by observational evaluation of the degree of connectivity in the network. KEGG pathways enriched with the hub genes were subsequently visualized by the Circlize package (69).

Biomarker identification using machine-based learning methods

The intersection of upregulated genes with modules significantly correlated positively, and downregulated genes with modules significantly correlated negatively, was subjected to machine-based learning methods. After evaluating the optimum number of trees in each split (m_try parameter), the random forest (RF) algorithm was employed to assess the classification potential of genes, using the randomForest package (70). Similarly, The Least Absolute Shrinkage and Selection Operator (LASSO) regression was applied to identify most prominent genes, using the glmnet package (71). At last, using the pROC package (72), the diagnostic efficacy of the top genes selected by both RF and LASSO methods was assessed by conducting receiver operating characteristic (ROC) analysis on each dataset. Furthermore, ROC analysis was also carried out on GSE141804 and GSE146383 to examine the robustness of the results.

Author Contributions

A.M.D.D. designed the research, acquired the data, conducted the analyses, interpreted the results, and prepared the manuscript.

Competing interests

The author declares no competing interests.

Data availability

The datasets analyzed during the current study are available in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) under the accession numbers GSE21942, GSE203241, GSE141804, and GSE146383.

Bando Y. Mechanism of demyelination and remyelination in multiple sclerosis. Clinical and Experimental Neuroimmunology. 2020;11:14-21.
Misrielal C, Mauthe M, Reggiori F, Eggen BJ. Autophagy in multiple sclerosis: two sides of the same coin. Frontiers in cellular neuroscience. 2020;14:603710.
Greiner T, Kipp M. What guides peripheral immune cells into the central nervous system? Cells. 2021;10(8):2041.
Fletcher JM, Lalor SJ, Sweeney CM, Tubridy N, Mills KH. T cells in multiple sclerosis and experimental autoimmune encephalomyelitis. Clin Exp Immunol. 2010;162(1):1-11.
Verreycken J, Baeten P, Broux B. Regulatory T cell therapy for multiple sclerosis: Breaching (blood-brain) barriers. Human Vaccines & Immunotherapeutics. 2022;18(7):2153534.
Kunkl M, Frascolla S, Amormino C, Volpe E, Tuosto L. T helper cells: the modulators of inflammation in multiple sclerosis. Cells. 2020;9(2):482.
Veroni C, Aloisi F. The CD8 T cell-Epstein-Barr virus-B cell trialogue: a central issue in multiple sclerosis pathogenesis. Frontiers in immunology. 2021;12:665718.
Van Langelaar J, Rijvers L, Smolders J, Van Luijn MM. B and T cells driving multiple sclerosis: identity, mechanisms and potential triggers. Frontiers in immunology. 2020;11:497059.
Beliën J, Goris A, Matthys P. Natural Killer Cells in Multiple Sclerosis: Entering the Stage. Frontiers in Immunology. 2022;13.
Haase S, Linker RA. Inflammation in multiple sclerosis. Therapeutic Advances in Neurological Disorders. 2021;14:17562864211007687.
De Bondt M, Hellings N, Opdenakker G, Struyf S. Neutrophils: underestimated players in the pathogenesis of multiple sclerosis (MS). International journal of molecular sciences. 2020;21(12):4558.
Rovira À, Doniselli FM, Auger C, Haider L, Hodel J, Severino M, et al. Use of gadolinium-based contrast agents in multiple sclerosis: a review by the ESMRMB-GREC and ESNR Multiple Sclerosis Working Group. European Radiology. 2024;34(3):1726-35.
Wang Y, Bou Rjeily N, Koshorek J, Grkovski R, Aulakh M, Lin D, et al. Clinical and radiologic characteristics associated with multiple sclerosis misdiagnosis at a tertiary referral center in the United States. Multiple Sclerosis Journal. 2023;29(11-12):1428-36.
Yu H, Wang F, Wu JJ, Gong J, Bi S, Mao Y, et al. Integrated transcriptomics reveals the brain and blood biomarkers in Alzheimer's disease. CNS Neurosci Ther. 2023;29(12):3943-51.
Santiago JA, Littlefield AM, Potashkin JA. Integrative transcriptomic meta-analysis of Parkinson’s disease and depression identifies NAMPT as a potential blood biomarker for de novo Parkinson’s disease. Scientific Reports. 2016;6(1):34579.
Liu Y, Gu H-Y, Zhu J, Niu Y-M, Zhang C, Guo G-L. Identification of Hub Genes and Key Pathways Associated With Bipolar Disorder Based on Weighted Gene Co-expression Network Analysis. Frontiers in Physiology. 2019;10.
He X, Xu J, Liu Y, Guo X, Wei W, Xing C, et al. Explorations on Key Module and Hub Genes Affecting IMP Content of Chicken Pectoralis Major Muscle Based on WGCNA. Animals. 2024;14(3):402.
Zheng C, Chen J, Chu F, Zhu J, Jin T. Inflammatory Role of TLR-MyD88 Signaling in Multiple Sclerosis. Frontiers in Molecular Neuroscience. 2020;12.
Kumar N, Sharma N, Mehan S. Connection between JAK/STAT and PPARγ signaling during the progression of multiple sclerosis: Insights into the modulation of T-cells and immune responses in the brain. Current molecular pharmacology. 2021;14(5):823-37.
Li H, Zheng C, Han J, Zhu J, Liu S, Jin T. PD-1/PD-L1 axis as a potential therapeutic target for multiple sclerosis: AT cell perspective. Frontiers in Cellular Neuroscience. 2021;15:716747.
Lee PW, Severin ME, Lovett‐Racke AE. TGF‐β regulation of encephalitogenic and regulatory T cells in multiple sclerosis. European journal of immunology. 2017;47(3):446-53.
Ruffini N, Klingenberg S, Schweiger S, Gerber S. Common Factors in Neurodegeneration: A Meta-Study Revealing Shared Patterns on a Multi-Omics Scale. Cells. 2020;9(12):2642.
Marcos-Ramiro B, Oliva Nacarino P, Serrano-Pertierra E, Blanco-Gelaz MA, Weksler BB, Romero IA, et al. Microparticles in multiple sclerosis and clinically isolated syndrome: effect on endothelial barrier function. BMC Neurosci. 2014;15:110.
Derakhshani A, Safarpour H, Abdoli Shadbad M, Hemmat N, Leone P, Asadzadeh Z, et al. The Role of Hemoglobin Subunit Delta in the Immunopathy of Multiple Sclerosis: Mitochondria Matters. Frontiers in Immunology. 2021;12.
Kotelnikova E, Kiani NA, Messinis D, Pertsovskaya I, Pliaka V, Bernardo-Faura M, et al. MAPK pathway and B cells overactivation in multiple sclerosis revealed by phosphoproteomics and genomic analysis. Proc Natl Acad Sci U S A. 2019;116(19):9671-6.
Cipriano GL, Schepici G, Mazzon E, Anchesi I. Multiple Sclerosis: Roles of miRNA, lcnRNA, and circRNA and Their Implications in Cellular Pathways. International Journal of Molecular Sciences. 2024;25(4):2255.
van Nierop GP, van Luijn MM, Michels SS, Melief M-J, Janssen M, Langerak AW, et al. Phenotypic and functional characterization of T cells in white matter lesions of multiple sclerosis patients. Acta Neuropathologica. 2017;134(3):383-401.
Jiménez-Fernández M, de la Fuente H, Martín P, Cibrián D, Sánchez-Madrid F. Unraveling CD69 signaling pathways, ligands and laterally associated molecules. Excli j. 2023;22:334-51.
Magner WJ, Weinstock-Guttman B, Rho M, Hojnacki D, Ghazi R, Ramanathan M, Tomasi TB. Dicer and microRNA expression in multiple sclerosis and response to interferon therapy. Journal of neuroimmunology. 2016;292:68-78.
Juźwik CA, S SD, Zhang Y, Paradis-Isler N, Sylvester A, Amar-Zifkin A, et al. microRNA dysregulation in neurodegenerative diseases: A systematic review. Prog Neurobiol. 2019;182:101664.
Zhao J, He Z, Wang J. MicroRNA-124: A Key Player in Microglia-Mediated Inflammation in Neurological Diseases. Front Cell Neurosci. 2021;15:771898.
Indrieri A, Carrella S, Carotenuto P, Banfi S, Franco B. The Pervasive Role of the miR-181 Family in Development, Neurodegeneration, and Cancer. Int J Mol Sci. 2020;21(6).
Piotrzkowska D, Miller E, Kucharska E, Niwald M, Majsterek I. Association of miRNA and mRNA Levels of the Clinical Onset of Multiple Sclerosis Patients. Biology. 2021;10(6):554.
Edgünlü TG, Yılmaz Ş G, Emre U, Taşdelen B, Kuru O, Kutlu G, Erdal ME. miR-181a-5p is a potential candidate epigenetic biomarker in multiple sclerosis. Genome. 2022;65(11):547-61.
Perdaens O, Dang HA, D'Auria L, van Pesch V. CSF microRNAs discriminate MS activity and share similarity to other neuroinflammatory disorders. Neurology Neuroimmunology & Neuroinflammation. 2020;7(2):e673.
Güllüoğlu H, Uysal H, Poyraz T, Altun Z, Kaya D, Özçelik P, İdiman E. Differences in the Differential Expression of MicroRNAs Between Patients with Familial Multiple Sclerosis and Those with Sporadic Multiple Sclerosis. Meandros Medical and Dental Journal. 2023;24(4).
Ibrahim SH, El-Mehdawy KM, Seleem M, El-Sawalhi MM, Shaheen AA. Serum ROCK2, miR-300 and miR-450b-5p levels in two different clinical phenotypes of multiple sclerosis: Relation to patient disability and disease progression. J Neuroimmunol. 2020;347:577356.
Mazdeh M, Kordestani H, Komaki A, Eftekharian MM, Arsang-Jang S, Branicki W, et al. Assessment of expression profile of microRNAs in multiple sclerosis patients treated with fingolimod. Journal of Molecular Neuroscience. 2020;70:1274-81.
Papiri G, D’Andreamatteo G, Cacchiò G, Alia S, Silvestrini M, Paci C, et al. Multiple Sclerosis: Inflammatory and Neuroglial Aspects. Current Issues in Molecular Biology. 2023;45(2):1443-70.
Gonzalo H, Nogueras L, Gil-Sánchez A, Hervás JV, Valcheva P, González-Mingot C, et al. Impairment of Mitochondrial Redox Status in Peripheral Lymphocytes of Multiple Sclerosis Patients. Frontiers in Neuroscience. 2019;13.
Safavizadeh N, Rahmani SA, Zaefizadeh M. Investigation of cytocrom c oxidase gene subunits expression on the Multiple sclerosis. Indian J Hum Genet. 2013;19(1):18-25.
Vavougios GD, Zarogiannis SG, Krogfelt KA, Gourgoulianis K, Mitsikostas DD, Hadjigeorgiou G. Novel candidate genes of the PARK7 interactome as mediators of apoptosis and acetylation in multiple sclerosis: An in silico analysis. Multiple Sclerosis and Related Disorders. 2018;19:8-14.
Dutta R, McDonough J, Yin X, Peterson J, Chang A, Torres T, et al. Mitochondrial dysfunction as a cause of axonal degeneration in multiple sclerosis patients. Annals of neurology. 2006;59(3):478-89.
Fischer MT, Sharma R, Lim JL, Haider L, Frischer JM, Drexhage J, et al. NADPH oxidase expression in active multiple sclerosis lesions in relation to oxidative tissue damage and mitochondrial injury. Brain. 2012;135(3):886-99.
Kular L, Jagodic M. Epigenetic insights into multiple sclerosis disease progression. J Intern Med. 2020;288(1):82-102.
Cree BA. Genetics of primary progressive multiple sclerosis. Handb Clin Neurol. 2014;122:211-30.
Kuliyev E, Gingras S, Guy CS, Howell S, Vogel P, Pelletier S. Overlapping Role of SCYL1 and SCYL3 in Maintaining Motor Neuron Viability. J Neurosci. 2018;38(10):2615-30.
Huang Y, Bu T, Wu L, Sun H, Gao S, Huang D, Zhang X. Identification of the circRNA/miRNA/mRNA network and hub mRNAs involved in multiple sclerosis through bioinformatics analysis. 2023.
Xu ZB, Feng X, Zhu WN, Qiu ML. Identification of key genes and microRNAs for multiple sclerosis using bioinformatics analysis. Medicine (Baltimore). 2021;100(48):e27667.
Steiner A, Schaale KH, Prigione I, De Nardo D, Dagley LF, Yu C-H, et al. Activation of STING due to COPI-deficiency. BioRxiv. 2020:2020.07. 09.194399.
Sakata E, Eisele MR, Baumeister W. Molecular and cellular dynamics of the 26S proteasome. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2021;1869(3):140583.
Ebstein F, Küry S, Papendorf JJ, Krüger E. Neurodevelopmental Disorders (NDD) Caused by Genomic Alterations of the. Protein Degradation, Aggregation, Membrane Trafficking and Exosomes in Neuronal Health and Disease. 2022.
Limanaqi F, Biagioni F, Gaglione A, Busceti CL, Fornai F. A Sentinel in the Crosstalk Between the Nervous and Immune System: The (Immuno)-Proteasome. Front Immunol. 2019;10:628.
Yadav D, Lee JY, Puranik N, Chauhan PS, Chavda V, Jin JO, Lee PCW. Modulating the Ubiquitin-Proteasome System: A Therapeutic Strategy for Autoimmune Diseases. Cells. 2022;11(7).
Diets IJ, van der Donk R, Baltrunaite K, Waanders E, Reijnders MR, Dingemans AJ, et al. De novo and inherited pathogenic variants in KDM3B cause intellectual disability, short stature, and facial dysmorphism. The American Journal of Human Genetics. 2019;104(4):758-66.
Kiselev IS, Kulakova OG, Boyko AN, Favorova OO. DNA methylation as an epigenetic mechanism in the development of multiple sclerosis. Acta Naturae. 2021;13(2):45.
Jung H, Seo SB. Histone lysine demethylase 3B (KDM3B) regulates the propagation of autophagy via transcriptional activation of autophagy-related genes. PLoS One. 2020;15(7):e0236403.
Al-kuraishy HM, Jabir MS, Al-Gareeb AI, Saad HM, Batiha GE-S, Klionsky DJ. The beneficial role of autophagy in multiple sclerosis: Yes or No? Autophagy. 2024;20(2):259-74.
Menascu S, Khavkin Y, Zilkha-Falb R, Dolev M, Magalashvili D, Achiron A, Gurevich M. Clinical and transcriptional recovery profiles in pediatric and adult multiple sclerosis patients. Ann Clin Transl Neurol. 2021;8(1):81-94.
Gentleman R, Carey V, Huber W, Irizarry R, Dudoit S. Bioinformatics and computational biology solutions using R and Bioconductor: Springer Science & Business Media; 2005.
Yu G, Wang L-G, Han Y, He Q-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics: a journal of integrative biology. 2012;16(5):284-7.
Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, et al. The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic acids research. 2023;51(D1):D638-D46.
Chen Y, Wang X. miRDB: an online database for prediction of functional microRNA targets. Nucleic acids research. 2020;48(D1):D127-D31.
Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome research. 2003;13(11):2498-504.
Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC bioinformatics. 2008;9:1-13.
Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC bioinformatics. 2013;14:1-14.
Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic acids research. 2016;44(W1):W90-W7.
Xie Z, Bailey A, Kuleshov MV, Clarke DJ, Evangelista JE, Jenkins SL, et al. Gene set knowledge discovery with Enrichr. Current protocols. 2021;1(3):e90.
Gu Z, Gu L, Eils R, Schlesner M, Brors B. circlize Implements and enhances circular visualization in R. Bioinformatics. 2014;30(19):2811-2.
Liaw A, Wiener M. Classification and regression by randomForest. R news. 2002;2(3):18-22.
Friedman J, Hastie T, Tibshirani R. Regularization paths for generalized linear models via coordinate descent. Journal of statistical software. 2010;33(1):1.
Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, Müller M. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC bioinformatics. 2011;12:1-8.

No competing interests reported.

SpplementaryTables.pdf

Download PDF

Reviewers invited by journal
16 Jul, 2024
Editor assigned by journal
16 Jul, 2024
Editor invited by journal
08 Jul, 2024
Submission checks completed at journal
05 Jul, 2024
First submitted to journal
04 Jul, 2024

You are reading this latest preprint version

Integrated transcriptomics of multiple sclerosis peripheral blood mononuclear cells identified COPG1, RPN1, and KDM3B as potential biomarkers

Status:

Version 1

Abstract

Figures

Introduction

Results

Differentially expressed genes and functional enrichment

miRNA-mRNA regulatory network

Identification of MS-specific modules by WGCNA

Identification of hub genes in the module involved in neurodegeneration

Identification of potential biomarkers using machine-based learning models

Biomarker validation

Discussion

Methods

Data collection and preparation

Differential expression and functional enrichment analyses

Construction of mRNA-miRNA regulatory network

Weighted gene co-expression network analysis (WGCNA)

Biomarker identification using machine-based learning methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1