Analysis of infiltrating immune cells and identification of related biomarkers in patients with refractory asthma

doi:10.21203/rs.3.rs-2089408/v1

Download PDF

Research Article

Analysis of infiltrating immune cells and identification of related biomarkers in patients with refractory asthma

https://doi.org/10.21203/rs.3.rs-2089408/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Patients with refractory asthma (RA) have an enhanced risk of asthma-related symptoms, mortality, and exacerbations.RA is related to innate immune deficiency. Thus, the aim of this study was to identify immune-related diagnostic genes involved in RA.

Methods

The limma R package was used to identify differentially expressed genes (DEGs) between RA and healthy control groups of induced sputum samples. The ClusterProfiler R package was used to carry out Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses of the DEGs. Single-sample gene set enrichment analysis (ssGSEA), was used to calculate the relative infiltration of 28 immune cell types. Weighted gene co-expression network analysis (WGCNA) was used to identify the modules that were correlated with the differentially infiltrated immune cells. Protein–protein interaction (PPI) network analysis was used to identify the hub genes of the network, and diagnostic genes were identified from among these hub genes by creating a diagnostic logistic regression model.The miRNA–diagnostic gene and transcription factor (TF)–diagnostic gene networks were constructed to explain the regulatory mechanisms of the diagnostic genes in RA.

Results

1844 DEGs (452 up-regulated and 1392 down-regulated) were identified between the RA and control groups. ssGSEA revealed that 17 immune cell types were significantly different between the RA and control samples. WGCNA identified two modules that were correlated with the differentially infiltrated immune cells. Next, 386 genes were identified as DEIRGs. Thereafter, the top 10 hub genes (CS, ATP5L, NDUFA4, ATP5B, COX5B, NDUFAB1, ATP5G1, NDUFA9, PARK7, and RPL8) in the PPI network, based on degree value, were identified. Next, we constructed a logistic regression model to accurately distinguish the RA and control groups; P < 0.05 for three of the genes (CS, ATP5L, and NDUFA4), which might be useful as RA diagnostic genes. Finally, the regulatory mechanisms of the diagnostic genes were explored based on the miRNA–diagnostic gene and TF–diagnostic gene networks.

Conclusion

The study identified CS, ATP5L, and NDUFA4 as RA diagnostic genes. These genes may serve as therapeutic targets for RA patients.

Refractory asthma

RNA-Seq

Differentially expressed genes

Logistic regression model

Enrichment analysis

Immunity

Refractory asthma (RA; also known as glucocorticoid-resistant asthma) refers to asthma cases that are not well controlled despite extensive glucocorticoid inhalation therapy. Up to 30% of patients with RA need regular oral corticosteroids [1]. Although RA only accounts for 5% of asthma cases, its economic cost is high and the disease is often poorly controlled, which brings great suffering to patients. The European Respiratory Society and American Thoracic Society published expert consensus opinions on RA in 1999 and 2000, respectively [2], that pointed out that although inhaled glucocorticoids and long-acting β-adrenergic receptor agonists control the symptoms of the vast majority of asthma patients, there are a small number of patients whose symptoms cannot be well controlled even with this standard treatment. Asthma cases are designated RA if they do not achieve ideal control after standard treatment for at least 6 months with the fourth-line treatment regimen, that is, two or more controller drugs [3–4].

The key reasons for the onset of RA are as follows: (1) Glucocorticoid resistance. This is mainly due to mutation of the glucocorticoid receptor gene that causes a decrease in the number of receptors and/or a decrease in affinity (due to a structural change in the receptor). (2) Abnormal levels of inflammatory factors. The high levels of eosinophils and T helper 2 (Th2)-type cytokines such as interleukin (IL-4) in the lung tissue represent an important reason for the persistent unrelieved inflammation in RA, which occurs despite treatment with large amounts of a combination of inhaled glucocorticoids. (3) Autoimmunity. More and more studies [5] have found the existence of autoimmunity in patients with RA. There are universal, nonspecific, inherited innate immune responses to certain antigenic substances, which act as the first line of defense against infection. Additionally, after these antigens stimulate the immune system, specific immune responses can be put into effect by the adaptive immune system (cellular and humoral immunity).

RA is associated with neutrophil extracellular traps [6], and the increased susceptibility to rhinovirus that occurs in asthma patients leads to increased lower airway inflammation and hyperresponsiveness [7]. Asthmatic inflammation is primarily driven by allergen-specific CD4 T cells that recruit other immune cells into the airways [8]. Moreover, the environmental microbiome influences asthma development by shaping the human microbiome, and this has received increasing attention [9]. The specific mechanism may be related to Toll-related immune regulation. Furthermore, serum neutrophil gelatinase-associated lipocalin (NGAL) is elevated in patients with asthma and airway obstruction, which may contribute to airway remodeling in asthmatics [10]. The above studies show that the pathogenesis of asthma is closely related to the body’s immune mechanism and related immune cells.

Given the important role of immunity in RA, this study aimed to identify immune-related diagnostic genes involved in RA that may act as therapeutic targets, in order to provide new ideas for the diagnosis and treatment of RA patients.

2.1 Data Source

The GSE76262 dataset was obtained from the Gene Expression Omnibus (GEO) database. There were data from 93 and 21 induced sputum cell samples from RA patients and healthy controls, respectively, including microarray data on 20,277 genes.

2.2 Identification of Differentially Expressed Genes (DEGs)

The limma R package [11] was used to process the microarray data to identify the DEGs between the RA and healthy control groups. The criteria for DEGs were adjusted p < 0.05 and |log2(fold change)|>0.5. The results were presented in a volcano plot and heatmap.

2.3 Biological Function Analyses

We performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses of the DEGs by using the ClusterProfiler R package [12].

2.4 Analysis of Immune Cell Infiltration

Single-sample gene set enrichment analysis (ssGSEA) can classify gene sets with common biological functions, chromosomal localization, and physiological regulation. The relative infiltration of 28 immune cell types in each sample was calculated using ssGSEA [13] based on the gene expression data. The Wilcoxon test (P < 0.05) was used to identify the differentially infiltrated immune cells between the RA and healthy control groups for subsequent analysis.

2.5 Weighted Gene Co-expression Network Analysis (WGCNA) and Module Functional Analysis

The WGCNA R package [14] was used to construct a co-expression network. First, using the Euclidean distance, hierarchical clustering of the gene expression in the samples was performed to detect and eliminate outliers. In the WGCNA, genes with similar expression patterns were assigned to a branch, and each branch represented a co-expression module. The pickSoftThreshold function of WGCNA was used to select the best soft threshold power (β) based on values from 1 to 20.

After calculating Pearson’s correlation coefficients between the modules and differentially infiltrated immune cell types, we selected the modules that were correlated. Furthermore, genes in the correlated modules with gene significance (GS) > 0.2 (which represents the association between a gene and a clinical trait, e.g., infiltrated immune cells) and module membership (MM) > 0.8 (which is the relationship between a module eigengene and gene expression) were identified and designated the “core module genes”.

2.6 Venn Diagram of Differentially Expressed Immune-Related Genes (DEIRGs)

The genes that were both DEGs and core module genes were designated DEIRGs, which were presumably associated with the differentially infiltrated immune cells. They were identified by using the VennDiagram R package [15] to plot a Venn diagram.

2.7 Construction of Protein–Protein Interaction (PPI) Network and Identification of Hub Genes

PPI information from the STRING database [16] is useful for predicting physical and functional interactions. All DEIRGs were uploaded to the STRING database, and the lowest minimum required interaction score was > 0.4 (medium confidence). Isolated nodes in the network were removed. CytoHubba [17], a Cytoscape plugin, was used to identify the hub genes in the network. The top 10 hub genes were selected based on degree value.

2.8 Construction of Diagnostic Prediction Model Using Logistic Regression

The expressions of the top 10 hub genes were treated as independent variables to construct a prediction model for the diagnosis of RA based on immune-related genes. The model was created by logistic regression using the glm function in R software to identify diagnostic genes (with P < 0.05 in the model) for distinguishing between RA and control samples.

2.9 Verification of Hub Genes by Receiver Operating Characteristic (ROC) Curve Analysis

ROC curve analysis was used to calculate the area under the curve (AUC), in order to compare diagnostic performance based on both specificity and sensitivity. The pROC R package [18] was used for the ROC curve analysis. The higher the AUC value, the better the ability of the diagnostic gene or model to distinguish between RA and healthy control samples.

2.10 Construction of miRNA–Diagnostic Gene and transcription factor (TF)–Diagnostic Gene Regulatory Networks

To identify the miRNAs targeting the diagnostic genes, we used the miRNet database, a network-based visual analysis tool that contains data on microRNA–target gene interactions [19]. Additionally, NetworkAnalyst [20] was used to match TFs to the diagnostic genes. The miRNA–diagnostic gene and TF–diagnostic gene networks were generated by Cytoscape [21].

3.1 Identification and Enrichment Analyses of DEGs in RA

The 114 samples (93 RA samples and 21 control samples) were normalized and fold changes were calculated by the limma R package. After setting the cutoff values (adjusted p < 0.05 and |log2(fold change)|>0.5), 1844 DEGs (452 up-regulated and 1392 down-regulated) were identified between the RA and control groups (Fig. 1). The expressions of the top 10 up-regulated genes (CLC, IL18R1, IL1RL1, P2RY14, IL1R2, SLAMF7, IL18RAP, XIST, TSIX, and PRSS33) and the top 10 down-regulated genes (CPE, LOC100129502, FAM3B, AOC3, SLC47A1, EIF1AY, CES1, STK39, RPS4Y1, and SCD) in the RA group are shown in a heatmap (Fig. 2).

GO analysis showed that 621 GO terms were enriched, comprising 445 biological process (BP), 109 cellular component (CC), and 67 molecular function (MF) terms. The most enriched terms were “neutrophil activation” and “neutrophil degranulation” in the BP category, “mitochondrial protein-containing complex” and “mitochondrial inner membrane” in the CC category, and “electron transfer activity” and “structural constituent of ribosome” in the MF category (Fig. 3). KEGG analysis showed that the DEGs were enriched in 31 pathways and the most enriched pathways were “prion disease”, “phagosome”, and “ribosome” (Fig. 4).

3.2 Analysis of Immune Cell Infiltration

ssGSEA was used to compare 28 immune cell types between the RA and control samples (Fig. 5). 17 immune cell types demonstrated a significant difference (P < 0.05, Wilcoxon test) between the RA (red) and control (blue) samples (Fig. 6). Activated CD4 T cells, type 17 T helper cells, regulatory T cells, activated B cells, memory B cells, CD56dim natural killer cells, myeloid-derived suppressor cells, plasmacytoid dendritic cells, macrophages, eosinophils, mast cells, monocytes, and neutrophils were increased in RA, while effector memory CD4 T cells, gamma delta T cells, natural killer cells, and immature dendritic cells were decreased in RA.

3.3 WGCNA

To identify genes associated with the differentially infiltrated immune cells, we performed WGCNA. First, hierarchical clustering of the samples showed that there were no outliers. Next, β = 8 (scale-free R² = 0.9) was selected as the appropriate soft-thresholding power in order to ensure a scale-free analysis (Fig. 7). Highly related genes were grouped into the same module. Six modules were identified (Fig. 8). The blue module was correlated (|Cor|>0.5, P < 0.05) with effector memory CD4 T cells, gamma delta T cells, type 17 T helper cells, activated B cells, memory B cells, natural killer cells, immature dendritic cells, macrophages, eosinophils, mast cells, monocytes, and neutrophils. The brown module was correlated (|Cor|>0.5, P < 0.05) with activated CD4 T cells, type 17 T helper cells, memory B cells, myeloid-derived suppressor cells, plasmacytoid dendritic cells, macrophages, eosinophils, mast cells, monocytes, and neutrophils (Fig. 9). Furthermore, core module genes in the blue and brown modules with GS > 0.2 and MM > 0.8 were identified. Consequently, we identified 391 (Fig. 10) and 129 (Fig. 11) core module genes in the blue and brown modules, respectively, giving a total of 520 core module genes.

3.4 Identification of Hub Genes in PPI Network

The DEGs and core module genes were analyzed, and 386 genes were identified as DEIRGs in RA (Fig. 12). To better understand the interplay among the 386 DEIRGs, we used STRING to construct a PPI network with minimum interaction score > 0.4 (Fig. 13). The top 10 genes in the network (ranked by degree value) were ATP5B, COX5B, NDUFAB1, NDUFA4, ATP5G1, NDUFA9, ATP5L, CS, PARK7 and RPL8 (Fig. 14). The degree values of the 10 hub genes are shown in Table 1.

To explore the biological functions of these 10 hub genes, we performed GO and KEGG analyses. The BP terms “cellular respiration”, “energy derivation by oxidation of organic compounds”, and “mitochondrial ATP synthesis coupled electron transport” were significantly enriched. The CC terms “mitochondrial respiratory chain complex I”, “NADH dehydrogenase complex”, and “respiratory chain complex I” were significantly enriched. The MF terms “electron transfer activity”, “NADH dehydrogenase activity”, and “NADH dehydrogenase (ubiquinone) activity” were significantly enriched (Fig. 15). Furthermore, the KEGG analysis showed that “Parkinson disease”, “pathways of neurodegeneration − multiple diseases”, and “oxidative phosphorylation” were significantly enriched (Fig. 16).

Table 1

Top 10 hub genes ranked by degree value
Gene	Degree	Rank
ATP5B	9	1
COX5B	8	2
NDUFAB1	8	2
NDUFA4	7	4
ATP5G1	7	4
NDUFA9	7	4
ATP5L	6	7
CS	5	8
PARK7	5	8
RPL8	2	10

3.5 Establishment of Diagnostic Model Based on Hub Genes

We hypothesized that the 10 hub genes may function as predictors of RA. The output of the first diagnostic logistic regression model is shown in Table 2. P > 0.05 for seven genes (RPL8, COX5B, NDUFAB1, ATP5B, NDUFA9, PARK7, and ATP5G1). After removing these genes and creating a second model with the remaining three genes (CS, ATP5L, and NDUFA4), the model fitting was satisfactory (Table 3). The model equation was:

where p^ is the probability of RA. If p^>0.5, the case is judged to be RA, otherwise the case is judged to be non-RA.

ROC curve analyses were performed to assess the diagnostic performance of CS, ATP5L, NDUFA4 independently and the second diagnostic model. The AUCs were 0.8715, 0.7438, 0.8505, and 0.9078, respectively (Fig. 17). The results demonstrated that the RA and control samples could be accurately distinguished by the diagnostic model, and CS, ATP5L, and NDUFA4 might be useful as feature parameters for the diagnosis of RA.

Table 2

Variables, coefficients, and significance in first model
	Coefficient	P value
Intercept	49.8527	0.00019
CS	-2.9673	0.02276
RPL8	2.6551	0.08075
COX5B	-1.704	0.18239
ATP5L	4.1186	0.01394
NDUFAB1	-1.7158	0.2063
ATP5B	-0.4189	0.83905
NDUFA4	-3.6126	0.03417
NDUFA9	0.3345	0.71856
PARK7	-1.5449	0.41722
ATP5G1	-0.6712	0.48852

Table 3

Variables, coefficients, and significance in second model
	Coefficient	P value
Intercept	50.2751	0.000143
CS	-2.6769	0.011357
ATP5L	2.3779	0.009108
NDUFA4	-4.632	0.004761

3.6 Prediction of miRNA–Diagnostic Gene and TF–Diagnostic Gene Networks

Subsequently, we predicted 107 miRNAs that may target the three diagnostic genes (CS, ATP5L and NDUFA4) using miRNet (Fig. 18). Next, a TF–diagnostic gene network was constructed using NetworkAnalyst to understand the possible TF–diagnostic gene transcriptional regulation mechanisms (Fig. 19). PRRX2, NKX2-5, BRCA1, GATA2, NFIC, NFYA, and SRF may regulate CS expression. ZNF354C, RELA, and EGR1 may regulate NDUFA4 expression. No potential TFs upstream of ATP5L were found.

The key reasons for the onset of RA are as follows: glucocorticoid resistance, persistent abnormal levels of inflammatory factors in the airway, airway remodeling, and genetic factors. The latest research has shown that the pathogenesis is also related to autoimmune phenomena [22]. Current conventional treatment is ineffective. This study explored immune-related genes associated with RA. We downloaded the GSE76262 dataset on RA from the GEO database. RA and healthy controls samples in the dataset were retained, and 1844 DEGs (452 up-regulated and 1392 down-regulated) were obtained. Next, three diagnostic genes (CS, ATP5L, and NDUFA4) were used to construct a diagnostic logistic regression model. ROC curves of these three diagnostic genes and the diagnostic model were constructed, and they were found to have good diagnostic performances in terms of their AUCs. Finally, miRNAs and TFs related to the three diagnostic genes were obtained using online databases, and corresponding regulatory networks were constructed.

We found significant differences in gene expression between the RA and control samples. Subsequently, GO analysis showed that the DEGs were enriched in neutrophil activation, neutrophil degranulation, mitochondrial protein-containing complexes, mitochondrial inner membrane, electron transfer activity, and structural components of ribosomes. KEGG analysis showed that the DEGs were enriched in prion diseases, phagosomes, and ribosomes. Thus, the above factors play important roles in the pathological process of RA.

We also analyzed the immune cell profile of RA. Using the GSE76262 dataset to perform ssGSEA, 17 significantly differentially infiltrated immune cell types were identified between the RA and healthy control groups. There were significant differences in the infiltration of immune cells such as myeloid-derived suppressor cells, plasmacytoid dendritic cells, immature dendritic cells, macrophages, eosinophils, mast cells, monocytes, and neutrophils. Similarly, previous studies have indicated that immune cell infiltration plays an important role in airway diseases such as asthma, including RA. Bronchial asthma is known to be a chronic inflammatory disease involving a variety of cells and cellular components. In airway diseases, innate immune cells, including dendritic cells, mast cells and innate lymphocytes, have been found to be an important part of the body’s innate immunity. In asthma in particular, innate immune cells [23], including type 2 innate lymphoid cells (ILC2) [24], play important roles. During acute exacerbations in children with neutrophilic asthma, neutrophil infiltration and serum progranulin (PGRN) levels reflect disease severity. Furthermore, autoimmunity in asthma patients can develop, with clinical manifestations such as insensitivity to glucocorticoid therapy and the development of RA [25]. In fact, autoantibodies have been found in the airways of a third of eosinophilic asthma patients, and it is believed that this autoimmunity is one of the reasons for the abovementioned treatment insensitivity.

In this study, 386 genes that were involved in RA and significantly related to immunity were found by creating a Venn diagram of the DEGs and core module genes. To further identify the key genes involved in RA, we constructed a PPI network to identify the top 10 hub genes. We then constructed a diagnostic logistic regression model. The diagnostic performances of the diagnostic model and the three individual genes in the model (CS, ATP5L, and NDUFA4) were evaluated using ROC curves. Citrate synthase (CS) exists in almost all cells capable of oxidative metabolism. ATP synthase subunit (ATP5L) is responsible for proton transport and it participates in mitochondrial function (involving ATP synthesis coupled with proton transport). NADH dehydrogenase (ubiquinone) 1 alpha subcomplex subunit 4 (NDUFA4) is encoded by the NDUFA4 gene located on human chromosome 7 p21. It is a mitochondrial respiratory chain protein that is an important subunit of mammalian mitochondrial cytochrome C oxidase (complex IV), capable of transferring electrons from NADH to the respiratory chain [26–28]. NDUFA4 and its homologues can reduce mitochondrial membrane potential and reduce reactive oxygen species (ROS) production during inflammation, leading to cytoprotection and suppression of immune responses [29]. Therefore, we speculate that RA is closely related to the immune microenvironment, and that the three key genes may affect the pathogenesis of RA patients by affecting the energy metabolism of immune cells.

In this study, bioinformatics methods were used to explore the pathogenesis of RA and its association with immunity. The three immune-related RA diagnostic genes identified (CS, ATP5L, and NDUFA4) had good diagnostic performance and may provide new therapeutic targets for use in RA. However, large-scale clinical data are still needed to further validate the findings.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

[PERSISTENT WEB LINK TO DATASETS].

Competing Interests

All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.

Funding

No funding was received to assist with the preparation of this manuscript.

Acknowledgments

Not applicable

1. Abe Y, Suga Y, Fukushima K, Ohata H, Niitsu T, Nabeshima H, Nagahama Y, Kida H, Kumanogoh A. Advances and Challenges of Antibody Therapeutics for Severe Bronchial Asthma. Int J Mol Sci. 2021 Dec 22;23(1):83. doi: 10.3390/ijms23010083. PMID: 35008504; PMCID: PMC8744863.

2. Zhang J, Dong L. Status and prospects: personalized treatment and biomarker for airway remodeling in asthma. J Thorac Dis. 2020 Oct;12(10):6090–6101. doi: 10.21037/jtd-20-1024. PMID: 33209441; PMCID: PMC7656354.

3. Schuh S, Sweeney J, Rumantir M, Coates AL, Willan AR, Stephens D, Atenafu EG, Finkelstein Y, Thompson G, Zemek R, Plint AC, Gravel J, Ducharme FM, Johnson DW, Black K, Curtis S, Beer D, Klassen TP, Nicksy D, Freedman SB; Pediatric Emergency Research Canada (PERC) Network. Effect of Nebulized Magnesium vs Placebo Added to Albuterol on Hospitalization Among Children With Refractory Acute Asthma Treated in the Emergency Department: A Randomized Clinical Trial. JAMA. 2020 Nov 24;324(20):2038–2047. doi: 10.1001/jama.2020.19839. PMID: 33231663; PMCID: PMC7686869.

4. Hekking PW, Wener RR, Amelink M, Zwinderman AH, Bouvy ML, Bel EH. The prevalence of severe refractory asthma. J Allergy Clin Immunol. 2015 Apr;135(4):896–902. doi: 10.1016/j.jaci.2014.08.042. Epub 2014 Oct 16. PMID: 25441637.

5. Pelaia C, Pelaia G, Crimi C, Maglio A, Gallelli L, Terracciano R, Vatrella A. Tezepelumab: A Potential New Biological Therapy for Severe Refractory Asthma. Int J Mol Sci. 2021 Apr 22;22(9):4369. doi: 10.3390/ijms22094369. PMID: 33922072; PMCID: PMC8122263.

6. Quoc QL, Choi Y, Thi Bich TC, Yang EM, Shin YS, Park HS. S100A9 in adult asthmatic patients: a biomarker for neutrophilic asthma. Exp Mol Med. 2021 Jul;53(7):1170–1179. doi: 10.1038/s12276-021-00652-5. Epub 2021 Jul 20. PMID: 34285336; PMCID: PMC8333352.

7. Pelaia C, Crimi C, Vatrella A, Tinello C, Terracciano R, Pelaia G. Molecular Targets for Biological Therapies of Severe Asthma. Front Immunol. 2020 Nov 30;11:603312. doi: 10.3389/fimmu.2020.603312. PMID: 33329598; PMCID: PMC7734054.

8. Chan R, RuiWen Kuo C, Lipworth B. Pragmatic Clinical Perspective on Biologics for Severe Refractory Type 2 Asthma. J Allergy Clin Immunol Pract. 2020 Nov-Dec;8(10):3363–3370. doi: 10.1016/j.jaip.2020.06.048. Epub 2020 Jul 13. PMID: 32673880.

9. Shi HL, Lan YH, Hu ZC, Yan ZN, Liu ZZ, Kadier X, Ma L, Yu JY, Liu J. Microecology research: a new target for the prevention of asthma. Chin Med J (Engl). 2020 Nov 20;133(22):2712–2720. doi: 10.1097/CM9.0000000000001127. PMID: 33031136; PMCID: PMC7647504.

10. Kawagoe J, Kono Y, Togashi Y, Ishiwari M, Toriyama K, Yajima C, Nakayama H, Kasagi S, Abe S, Setoguchi Y. Serum Neutrophil Gelatinase-associated Lipocalin (NGAL) Is Elevated in Patients with Asthma and Airway Obstruction. Curr Med Sci. 2021 Apr;41(2):323–328. doi: 10.1007/s11596-021-2350-1. Epub 2021 Apr 20. PMID: 33877549.

11. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015 Apr 20;43(7):e47. doi: 10.1093/nar/gkv007. Epub 2015 Jan 20. PMID: 25605792; PMCID: PMC4402510.

12. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012 May;16(5):284-7. doi: 10.1089/omi.2011.0118. Epub 2012 Mar 28. PMID: 22455463; PMCID: PMC3339379.

13. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005 Oct 25;102(43):15545-50. doi: 10.1073/pnas.0506580102. Epub 2005 Sep 30. PMID: 16199517; PMCID: PMC1239896.

14. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008 Dec 29;9:559. doi: 10.1186/1471-2105-9-559. PMID: 19114008; PMCID: PMC2631488.

15. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008 Dec 29;9:559. doi: 10.1186/1471-2105-9-559. PMID: 19114008; PMCID: PMC2631488. PMID: 21269502

16. von Mering C, Huynen M, Jaeggi D, Schmidt S, Bork P, Snel B. STRING: a database of predicted functional associations between proteins. Nucleic Acids Res. 2003 Jan 1;31(1):258 − 61. doi: 10.1093/nar/gkg034. PMID: 12519996; PMCID: PMC165481.

17. Chin CH, Chen SH, Wu HH, Ho CW, Ko MT, Lin CY. cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol. 2014;8 Suppl 4(Suppl 4):S11. doi: 10.1186/1752-0509-8-S4-S11. Epub 2014 Dec 8. PMID: 25521941; PMCID: PMC4290687.

18. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, Müller M. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011 Mar 17;12:77. doi: 10.1186/1471-2105-12-77. PMID: 21414208; PMCID: PMC3068975.

19. Chang L, Zhou G, Soufan O, Xia J. miRNet 2.0: network-based visual analytics for miRNA functional analysis and systems biology. Nucleic Acids Res. 2020 Jul 2;48(W1):W244-W251. doi: 10.1093/nar/gkaa467. PMID: 32484539; PMCID: PMC7319552.

20. Zhou G, Soufan O, Ewald J, Hancock REW, Basu N, Xia J. NetworkAnalyst 3.0: a visual analytics platform for comprehensive gene expression profiling and meta-analysis. Nucleic Acids Res. 2019 Jul 2;47(W1):W234-W241. doi: 10.1093/nar/gkz240. PMID: 30931480; PMCID: PMC6602507.

21. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, Amin N, Schwikowski B, Ideker T. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003 Nov;13(11):2498 − 504. doi: 10.1101/gr.1239303. PMID: 14597658; PMCID: PMC403769.

22. Hudon Thibeault AA, Laprise C. Cell-Specific DNA Methylation Signatures in Asthma. Genes (Basel). 2019 Nov 15;10(11):932. doi: 10.3390/genes10110932. PMID: 31731604; PMCID: PMC6896152.

23. Pasha MA, Patel G, Hopp R, Yang Q. Role of innate lymphoid cells in allergic diseases. Allergy Asthma Proc. 2019 May 1;40(3):138–145. doi: 10.2500/aap.2019.40.4217. PMID: 31018888; PMCID: PMC6500789.

24. Wang X, Zhao H, Ma C, Lv L, Feng J, Han S. Gallic acid attenuates allergic airway inflammation via suppressed interleukin-33 and group 2 innate lymphoid cells in ovalbumin-induced asthma in mice. Int Forum Allergy Rhinol. 2018 Nov;8(11):1284–1290. doi: 10.1002/alr.22207. Epub 2018 Sep 7. PMID: 30191679.

25. Mukherjee M, Nair P. Autoimmune Responses in Severe Asthma. Allergy Asthma Immunol Res. 2018 Sep;10(5):428–447. doi: 10.4168/aair.2018.10.5.428. PMID: 30088364; PMCID: PMC6082822.

26. Mukherjee M, Bulir DC, Radford K, Kjarsgaard M, Huang CM, Jacobsen EA, Ochkur SI, Catuneanu A, Lamothe-Kipnes H, Mahony J, Lee JJ, Lacy P, Nair PK. Sputum autoantibodies in patients with severe eosinophilic asthma. J Allergy Clin Immunol. 2018 Apr;141(4):1269–1279. doi: 10.1016/j.jaci.2017.06.033. Epub 2017 Jul 24. PMID: 28751233.

27. Pellegrino MA, D'Antona G, Bortolotto S, Boschi F, Pastoris O, Bottinelli R, Polla B, Reggiani C. Clenbuterol antagonizes glucocorticoid-induced atrophy and fibre type transformation in mice. Exp Physiol. 2004 Jan;89(1):89–100. doi: 10.1113/expphysiol.2003.002609. PMID: 15109214.

28. Dizier MH, Margaritte-Jeannin P, Pain L, Sarnowski C, Brossard M, Mohamdi H, Lavielle N, Babron MC, Just J, Lathrop M, Laprise C, Bouzigon E, Demenais F, Nadif R. Interactive effect between ATPase-related genes and early-life tobacco smoke exposure on bronchial hyper-responsiveness detected in asthma-ascertained families. Thorax. 2019 Mar;74(3):254–260. doi: 10.1136/thoraxjnl-2018-211797. Epub 2018 Oct 3. PMID: 30282721.

29. Lee CQE, Kerouanton B, Chothani S, Zhang S, Chen Y, Mantri CK, Hock DH, Lim R, Nadkarni R, Huynh VT, Lim D, Chew WL, Zhong FL, Stroud DA, Schafer S, Tergaonkar V, St John AL, Rackham OJL, Ho L. Coding and non-coding roles of MOCCI (C15ORF48) coordinate to regulate host inflammation and immunity. Nat Commun. 2021 Apr 9;12(1):2130. doi: 10.1038/s41467-021-22397-5. PMID: 33837217; PMCID: PMC8035321.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Analysis of infiltrating immune cells and identification of related biomarkers in patients with refractory asthma

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1. Introduction

2. Materials And Methods

2.1 Data Source

2.2 Identification of Differentially Expressed Genes (DEGs)

2.3 Biological Function Analyses

2.4 Analysis of Immune Cell Infiltration

2.5 Weighted Gene Co-expression Network Analysis (WGCNA) and Module Functional Analysis

2.6 Venn Diagram of Differentially Expressed Immune-Related Genes (DEIRGs)

2.7 Construction of Protein–Protein Interaction (PPI) Network and Identification of Hub Genes

2.8 Construction of Diagnostic Prediction Model Using Logistic Regression

2.9 Verification of Hub Genes by Receiver Operating Characteristic (ROC) Curve Analysis

2.10 Construction of miRNA–Diagnostic Gene and transcription factor (TF)–Diagnostic Gene Regulatory Networks

3. Results

3.1 Identification and Enrichment Analyses of DEGs in RA

3.2 Analysis of Immune Cell Infiltration

3.3 WGCNA

3.4 Identification of Hub Genes in PPI Network

3.5 Establishment of Diagnostic Model Based on Hub Genes

3.6 Prediction of miRNA–Diagnostic Gene and TF–Diagnostic Gene Networks

4. Discussion

Declarations

References

Additional Declarations

Status:

Version 1