NF-kappa B signaling pathway is associated with metformin resistance in type 2 diabetes patients

Abstract

Introduction

Metformin is an essential medicine that is most widely prescribed frontline for the treatment of Type 2 diabetes (T2D). Metformin upgraded glycemic control in T2D patients without hypoglycemic effects in patients. This assessment aims to understand molecular mechanism mechanisms in non-responder patients to metformin.

Methods

Gene expression profiles of responder and non-responder T2D patients to metformin are extracted from Gene Expression Omnibus (GEO) and are evaluated by the GEO2R program to find the significant differentially expressed genes (DEGs). The significant DEGs have been studied via action map gene ontology analyses.

Results

Results indicate that 563 significant DEGs discriminate non-responders from responder groups. “NF-kappa B signaling pathway” and 11 DEGs including BIRC3, CCL4L2, CXCL2, ICAM1, LYN, MYD88, RELA, SYK, TLR4, TNFAIP3, and TRIM25 were pointed out as core of drug resistance.

Conclusion

It can be concluded that there are differences between gene expression analysis, the response of diabetic patients to metformin. Results indicate that dysregulation of the “NF-kappa B signaling pathway” and TNFAIP3, BIRC3, RELA, MYD88, TLR4, and ICAM1 is associated with drug resistance in T2D patients.

Introduction

Metformin (1,1-dimethylbiguanide) originated from Galega officinalis planet, and has been used for centuries in folk medicine [1]. Metformin is an essential medicine introduced by the world health organization for type 2 diabetes (T2D and the most widely prescribed frontline for the treatment of this disease [2]. Studies revealed that metformin improved glycemic control in T2D patients and did not have hypoglycemic effects in those patients [3]. Metformin primarily inhibits gluconeogenesis in the liver which is still under investigation in D2 patients [4]. On the other hand, treatment of aging and the anti-cancer effects of metformin were suggested in research [5, 6]. It has been shown that metformin could reduce age-associated inflammation by increasing autophagy and mitochondrial function [7]. Metformin could also increase cardiovascular efficiency by improving vascular function and lipid profiles [8, 9]. Investigation about mechanism of metformin effect indicates that there are conflicts related to its doses variability and management [10–12].

Different proteomic research as a powerful tool can help to investigate the genes and proteins affected by metformin to show the expression of key proteins in vivo and invitro such as Ces1C and Cyp7a1 [13], vascular basement membrane proteins accumulation and vascular remodeling [14] and pancrease islets proteins modifications to prevent proinflammatory cytokines damages [15]. Bioinformatics and genomics have attracted researchers’ attention to investigate molecular mechanisms of diseases and therapeutic methods. Gene ontology and regulatory networks are useful tools for detecting molecular events in biological systems [16, 17].

Non-acceptance and response of some diabetic patients to metformin are also the categories that were discussed in this article to demonstrate key biochemical pathways and affected proteins, following the reaction of nonresponse patients to metformin administration.

Methods

Data collection

Information from 30 whole blood samples of 10 type 2 diabetic people; responders to metformin, 10 type 2 diabetic individuals; non-responders to metformin, and 10 healthy controls are recorded in GSE153315 in the GEO database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE153315). Data are characterized as Organism; homo sapiens (Human), Library strategy; RNA-Seq, and Library source transcriptomic.

Data pre-evaluation

The gene expression profiles of non-responder and responder groups were compared via the GEO2R program to find the significant DEGs involved in the drug resistance process. Evaluations were assessed via volcano plot and box plot analyses. Due to statistical mismatching, the suitable samples were selected for further study. Non-responder and responder groups were compared with controls separately.

Gene expression analysis

Based on adjusted p-value (P_adj) < 0.05, the significant DEGs for both analyses (non-responder-control and responder-control analysis) were identified. The common significant DEGs between non-responder-control and responder-control analyses which were characterized with difference of log(fold change) < 0.6 were ignored for more analysis. The remained DEGs that discriminate two analyses were considered for gene ontology enrichment and action map analysis.

Gene ontology assessment

The candidate DEGs were evaluated via gene ontology enrichment by CluePedia application of Cytoscape software. The related biochemical pathways were identified.

Action map analysis

Activation, inhibition, and expression actions were investigated for the studied significant DEGs via the CluGO application of Cytoscape software.

Statistical analysis

Adjusted p-value (Padj) < 0.05 was considered to find the significant DEGs. P-value of grouping: 0.05 and correction test Bonferroni step-down were applied to analyze the significant enriched biochemical pathways and clusters.

Results

As shown in Fig. 1, the non-responder and responder groups are separated by several significant upregulated and downregulated DEGs. Results indicate that there are 127 significant DEGs among 15,308 dysregulated genes. Box plot visualization of results (Fig. 2) revealed that the style of gene expression distribution of GSM4640264, GSM4640269, GSM4640270, GSM4640272, and GSM4640273 is different from other gene expression profiles. Analysis was repeated without the five mentioned mismatched samples as a modified analysis (see Fig. 3). As depicted in Fig. 4, significant differences between non-responder and responder groups disappeared. Analysis showed two control samples (GSM4640277-8) were not statistically comparable with the other samples and were ignored for further analysis. Results of the non-responder-control analysis are presented in Figs. 5 and 6. Based on the finding, among 17,372 dysregulated genes, 145 significant DEGs discriminate the non-responder group from healthy controls. The volcano plot and box plot of responder-control analysis are illustrated in Figs. 7 and 8. As it is shown in Fig. 7, GSM4640260 and GSM4640271 are median center as like other gene expression profiles and were considered for more analysis. This analysis is characterized by 17,159 dysregulated genes and 653 significant DEGs. After screening, 563 significant DEGs were introduced as the genes that discriminate both non-responder-control and responder-control analysis.

Fig. 1.

Volcano plot of non-responder-responder groups analysis

Fig. 2.

Box plot of non-responder-responder groups analysis

Fig. 3.

Box plot of modified non-responder-responder groups analysis

Fig. 4.

Volcano plot of modified non-responder-responder groups analysis

Fig. 5.

Box plot of non-responder-control groups analysis

Fig. 6.

Volcano plot of non-responder-control groups analysis

Fig. 7.

Box plot of responder-control groups analysis

Fig. 8.

Volcano plot of responder-control groups analysis

A number of 23 clusters of biochemical pathways including 55 terms were extracted from KEGG. A total of 55 biochemical pathways related to the 563 significant DEGs were identified. The determined pathways were grouped into 23 clusters. Considering the corrected p-value with Bonferroni step-down led to introduction of the “NF-kappa B signaling pathway” and “Herpes simplex virus 1 infection” as significant pathways. “NF-kappa B signaling pathway” and the related DEGs are presented in Fig. 9. “Herpes simplex virus 1 infection” was an isolated pathway in the visualized presentation of the map. As depicted in Fig. 9, BIRC3, CCL4L2, CXCL2, ICAM1, LYN, MYD88, RELA, SYK, TLR4, TNFAIP3, and TRIM25 are the 11 genes that linked directly to “NF-kappa B signaling pathway” pathway. Regulatory relationships between significant 563 DEGs are exposed in Figs. 10 and 11 (the isolated DEGs are not shown). For better understanding, activation, inhibition, and expression relationships between elements of the main connected component of the significant DEGs action map are provided and presented in Fig. 12.

Fig. 9.

“NF-kappa B signaling pathway” and the related DEGs. Positive co-expression of CXCL2 and SYK and TLR4 and ICAM1 is shown via the yellow connections. There are additional colors in genes as like ICAM1 relative to the central pathway because this gene was connected to the other pathways which arenot presented in this Figure

Fig. 10.

Activation relationship between significant DEGs. Direction of arrow shows activation

Fig. 11.

Inhibition and expression relationships between significant DEGs. Bar tip of red arrow shows inhibition, bar and round tips of yellow arrows refer to negative co-expression and positive co-expression respectively

Fig. 12.

Activation, inhibition, and expression relationships between significant DEGs. Direction of green arrow shows activation, Bar tip of red arrow displays inhibition, bar and round tips of yellow arrows refer to negative co-expression and positive co-expression respectively

Discussion

The box plot is the most common implementation of visualizing sample distributions [18]. Box plot analysis revealed all samples are not eligible for analysis. The selected sample based on the distribution pattern of gene expression amounts led to the analysis of the modified samples. “NF-kappa B signaling pathway” was introduced as the prominent pathway that differentiates non-responder groups from responder patients. Pathway analysis is a suitable method to explore the drug resistance process [19]. The crucial role of NF kappa B in the progress of diabetes mellitus and numerous complications that are associated with this disease is investigated and confirmed. Several pro-inflammatory cytokines activate NF kappa B in diabetes mellitus to regulate both the survival and death of β-cells [20]. NF kappa B as a family of nuclear transcription factors is associated with the pathogenesis of several continuing inflammatory progressions and regulation of immune response [21].

Number of 11 DEGs including BIRC3, CCL4L2, CXCL2, ICAM1, LYN, MYD88, RELA, SYK, TLR4, TNFAIP3, and TRIM25 were associated to “NF-kappa B signaling pathway”. Since action map is a directed PPI can be used as a screen tool to detect the critical genes. Action map analysis revealed that 82% of the associated genes including BIRC3, CXCL2, ICAM1, LYN, MYD88, RELA, SYK, TLR4, and TNFAIP3 are connected via activation, inhibition, and expression actions. More analysis indicates that activation, expression, and inhibition connect 73%, 45%, and 36% of 11 associated genes respectively. As depicted in Fig. 10 all activated or activator DEGs are included in the main connected component. Results of Fig. 11 designate that ICAM1, TLR4, and MYD88 are linked via expression connections. The significant role of these three DEGs is highlighted in Fig. 10. Due to less importance of the small subnetworks the paired DEGs are not considered for more analysis.

Investigations indicate that T2D pathogenesis is closely associated with the activation of the innate immune system and chronic low-grade inflammation. It is reported that toll-like receptor-4 (TLR4) acts as a linker between immune and metabolic systems. It is suggested that TLR4 is involved in production of proinflammatory cytokines and the diabetic deleterious effects [22]. As mentioned by Tian J et al.; TLR4/MyD88/NF-κB signaling develops the production of IL6, TNF-α, and MCP-1 which promote heart- and liver-related complications of T2DM [23]. Based on the investigation of Siddiqui K et al., intercellular adhesion molecule-1 (ICAM1) has a prominent role in microvascular complications amid with type 2 diabetic patients. (T2D) [24]. TLR4 either activates or upregulates ICAM1 (see Figs. 10 and 11).

RELA encodes a subunit of NF-κB transcription factor which is involved in the regulation of islet-specific transcriptional programs that are necessary for the preservation of glucose metabolism in healthy condition [25]. RELA has a compact relationship with elements of the main connected component especially TLR4, MYD88, BIRC3, and cxcl2 (see Fig. 10).

Due to the crucial role of human monocyte T lymphocyte-like receptor adapter protein (SCIMP)-tyrosine-protein kinase LYN- tyrosine-protein kinase SYK axis in intensification of renal injury, LYN can be associated with renal disease inflammation and T2DM [26]. Connections between LYN-SYK and the first neighbors are illustrated in Fig. 10. As shown in Fig. 12, the discussed nine associated DEGs are included in the main connected component of the action map. An important axis (TNFAIP3-BIRC3-RELA-MYD88-TLR4-ICAM1) is formed in Fig. 12. This axis can be considered as crucial element of drug resistance in patients relative to the introduced 11 key genes.

Conclusion

In conclusion, regarding the pattern of gene expression analysis, the response of diabetic patients to metformin is different. “NF-kappa B signaling pathway” and the related genes (BIRC3, CCL4L2, CXCL2, ICAM1, LYN, MYD88, RELA, SYK, TLR4, TNFAIP3, and TRIM25) play a crucial role in drug resistance patients. It seems regulation of members of the TNFAIP3-BIRC3-RELA-MYD88-TLR4-ICAM1 axis can solve the problem of drug resistance in diabetic patients. Experimental validation and a suitable sample size in future investigations are needed to approve the results of this assessment.

Authors’ contribution

Conceptualization: Vahid Mansouri, Babak Arjmand, Mostafa Rezaei Tavirani. Data curation: Farideh Razi, Mitra Rezaei. Formal analysis: Mostafa Rezaei Tavirani, Majid Rezaei Tavirani. Funding acquisition: Vahid Mansouri. Investigation: Fatemeh Bandarian. Methodology: Mitra Rezaei, Zahra Razzaghi. Project administration: Babak Arjmand, Mostafa Rezaei Tavirani. Resources: Zahra Razzaghi, Mitra Rezaei. Software: Zahra Razzaghi, Mostafa Rezaei Tavirani. Supervision: Frideh Razi, Babak Arjmand. Validation: Zahra Razzaghi. Majid Rezaei Tavirani. Visualization: Farideh Razi. Writing–original draft: Mostafa Rezaei Tavirani, Fatemeh Bandarian, Vahid Mansouri. Writing–review & editing: Babak Arjmand, Majid Rezaei Tavirani.

Data availability

Data is available upon reasonable request.

Declarations

Ethical declarations

This project is approved by ethical code IR.SBMU.RETECH.REC.1402.401 by Shahid Beheshti University of Medical Sciences.