A comprehensive bioinformatics method to screen key genes for severe burn

Abstract

Severe burn is a kind of traumatic injury, which can lead to serious financial burden, high morbidity and mortality following inflammatory response and complications. Microarray analysis has emerged as a popular tool for generating gene expression data and revealing the complex gene expression patterns. The Gene Expression Omnibus (GEO) repository at the National Center for Biotechnology Information has become the leading public repository of microarray data. This study aimed to study the mechanisms of severe burn. Microarray of GSE37069 was downloaded from GEO including blood samples from 244 severe burn patients and blood samples from 35 healthy controls. The differentially expressed genes (DEGs) between blood samples from healthy controls and patients were identified by t-test in the LIMMA package of R. Their interaction relationships were searched using STRING online software and then protein-protein interaction (PPI) network was constructed by Cytoscape. Using DAVID online tool, underlying functions of the DEGs involved in the PPI network were analyzed by functional and pathway enrichment analyses. We screened 541 DEGs in blood samples from severe burns patients compared with blood samples from healthy controls. Meanwhile, DEGs including MMP9, TIMP1, GZMK, GZMA, and GZMB showed higher degrees in the PPI networks. Moreover, they may function in severe burn through interacting with others. According to functional enrichment analysis, these DEGs were mainly involved in BP categories associated with inflammatory responses. MMP9, TIMP1, HGF, GZMA, GZMB, and GZMK might play important roles in severe burns.

Keywords

Burn injury differentially expressed genes enrichment analysis inflammatory responses protein-protein interaction network

Introduction

Severe burns can result in severe sequelae such as heterotopic ossification and significant hypermetabolism and catabolism, which are related to higher morbidity and mortality post severe burn.¹ It is reported that patients usually have hypermetabolic, debilitated, and disabled diseases for at least 2 years after severe burn injury.^2,3

After burns, inflammatory responses start immediately and are usually sustained for several months or years.⁴ The inflammatory responses include many aspects such as sepsis, which is a type of systemic inflammatory response syndrome induced by infection after severe burn.⁵ Moreover, the level of serum procalcitonin indicating severity of septic complication in patients with severe burn^6–8 and the continuous downregulation of monocyte human leukocyte antigen-DR have correlation with septic complications.^9–11 Insulin-like growth factor 1 (IGF-I)/insulin-like growth factor binding protein-3 (IGFBP-3) can reduce small bowel epithelial cell apoptosis by downregulating the Fas pathway, and then enhance gut mucosal integrity following severe burn.¹² As previously reported, tumor necrosis factor (TNF) can decrease lymphocyte apoptosis thus helping to enhance immunity and reduce risk of sepsis;¹³ the elevation of MMPs may be beneficial to tissue fibrosis and scar formation after burn injury.^14,15 Thus, the prognostic importance of MMPs and TIMP1 for burn trauma has been established.¹⁶ Some of these results have been successively testified in animal models for the drug candidates for burn trauma. However, there was poor correlation in the responses between animal models and human conditions.

In 2013, Seok et al. performed a microarray data analysis based on GSE37069 (burns), GSE36809 (tranma), endotoxemia (GSE3284), and four microarray data of sepsis (GSE13094, GSE9960, GSE13015, and GSE28750) to identify gene expression changes responding to burns, trauma, sepsis, and endotoxemia between patients/mouse models and healthy controls.¹⁷ They obtained numerous differentially expressed genes (DEGs) between burns patients and controls, and found that the DEGs and their related function terms were different between mouse models and human conditions. This demonstrated there still existed distance between clinical application and the laboratory test, and more tests should be done to identify more candidates for therapy for burn trauma.

In order to identify more genetic candidates for therapy for burn trauma, expression microarray data GSE37069 deposited in Gene Expression Omnibus (GEO) by Seok et al.¹⁷ was downloaded. DEGs in this dataset were identified and their interaction relationships were investigated by protein-protein interaction (PPI) network. Moreover, potential functions of DEGs involved in the PPI network were analyzed by functional and pathway enrichment analyses. The functions of the significant DEGs involved in significant function terms would be discussed and some of them might be considered as the candidates for severe burn therapy.

Materials and methods

Microarray data

Expression profile of GSE37069 deposited by Seok et al.¹⁷ which was based on the platform of GPL570 [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array was obtained from GEO (http://www.ncbi.nlm.nih.gov/geo/). GSE37069 included 281 blood samples from 244 severe burn patients admitted within 96 h after burn injury and with over 20% of the total body surface area of burns, and 37 blood samples from 35 healthy controls (age range, 16–55 years; seven controls were studied twice approximately 2 years apart). The informed consents were obtained from all the participants in the study of Seok et al.¹⁷

DEGs screening

Microarray data were preprocessed by Affy package.¹⁸ After background correction and quantile normalization, probe IDs were converted to gene symbols and then duplicates were removed. Finally, the gene expression matrix of the test group was obtained. The t-test method in the LIMMA (Linear Models for Microarray Data) package of R¹⁹ (http://www.bioconductor.org) was used to identify the genes that were differentially expressed between blood samples from patients and healthy controls. Next, the t-test P value was adjusted into false discovery rate (FDR) using multiple Benjamini–Hochberg tests.²⁰ Then, the genes of FDR <0.05 and |log₂fold change (FC)| >1.2 were considered as significant DEGs in this study.

PPI network construction

The interactions of proteins encoded by the DEGs selected above were searched by STRING online software,²¹ with combined score (required confidence) >0.4 was used as the cutoff criterion. Afterwards, the PPI network was visualized by using Cytoscape software (http://www.cytoscape.org).²² Then, connectivity degree analysis in the PPI network was performed to identify the potential key nodes (proteins) in the network.

Functional and pathway enrichment analysis

Gene Ontology (GO) employs a series of structured vocabularies for annotating genes and their products in biological process (BP), cellular component (CC), and molecular function (MF).²³ Kyoto Encyclopedia of Genes and Genomes (KEGG) database is composed of known genes and their functions and informs people of how molecules or genes work.²⁴ Database for Annotation, Visualization, and Integrated Discovery (DAVID) is a valuable method to classify genes and terms into several manageable modules.²⁵ Using the DAVID online tool, we performed GO functional and KEGG pathway enrichment analyses to identify the GO BP terms and pathways associated with the DEGs involved in the PPI network. The terms and pathways connected to at least three genes at the level of P value <0.05 were considered significant.

Results

DEGs analysis

After the data processing, a total of 541 DEGs were identified between blood samples from healthy controls and patients with the FDR <0.05 and |log₂FC| >1.2. Among these 541 genes which were significantly altered in burn patients, 305 genes were upregulated genes (e.g. CTSG, MPO, CAMP, PROK2, TSPO, TIMP1, LTF, HGF, STAT5B, and MMP9) and 236 were downregulated genes (e.g. CD2, CD8A, EOMES, GPR183, GZMK, GZMA, and GZMB).

PPI network analysis

We then separately searched the PPI networks of the up- and downregulated DEGs products. The PPI network of upregulated genes productions consisted of 170 nodes interacting through 384 interactions (lines) (Figure 1). Twenty-eight (16.5%, 22/170) nodes interacting with others through >8 interactions, such as translocator protein (18 kDa) (TSPO, degree = 42), myeloperoxidase (MPO, degree = 22), cathepsin G (CTSG, degree = 20), lactotransferrin (LTF, degree = 18), cathelicidin antimicrobial peptide (CAMP, degree = 16), matrix metalloproteinase 9 (MMP9, degree = 17), tissue inhibitor of metalloproteinase 1 (TIMP1, degree = 13), prokineticin 2 peptide (PROK2, degree = 8), and hepatocyte growth factor (HGF, degree = 8) which interacted with each other.

Figure 1.

The protein-protein interaction (PPI) network for upregulated genes.

The PPI network of productions of the downregulated genes consisted of 286 interactions linking to 101 nodes (Figure 2). This PPI network contained 33 nodes (32.7%, 33/101) interacting with the others through >8 interactions, such as CD2 (degree = 28), CD8A (degree = 26), granzyme K (GZMK, degree = 23), granzyme A (GZMA, degree = 22), granzyme B (GZMB, degree = 17), eomesodermin (EOMES, degree = 8), and G-protein-coupled receptor (GPR)183 (GPR183, degree = 8).

Figure 2.

The protein-protein interaction (PPI) network for downregulated genes.

Functional and pathway enrichment analysis

After the GO enrichment analysis, the DEGs in the PPI network of upregulated genes were mainly associated with the GO BP categories of defense response (including STAT5B, MPO, CTSG, LTF, CAMP, and PROK2, P value = 2.41E-15), inflammatory response (including STAT5B and PROK2,P value = 2.58E-09), and response to wounding (including STAT5B and PROK2, P value = 1.48E-08) (Table 1). Moreover, these DEGs were enriched in pathways such as hematopoietic cell lineage (P value = 1.04E-02), Fc epsilon RI signaling pathway (P value = 3.24E-02), and glutathione metabolism (P value = 4.20E-02) (Table 2).

Table 1.

The top 10 enriched Gene Ontology biological process terms for the up- and downregulated genes involved in PPI networks.

DEGs	Terms	Names	Gene number	Gene symbols	P value
Up	GO:0006952	Defense response	35	STAT5B, PGLYRP1, CXCR1, PRDX5, HP, FPR2, TLR5, AZU1, NLRC4, HIST1H2BK, LTB4R, CSF3R, LTF, CD24, PTX3, IL18R1, CR1, IL18RAP, RNASE3, KL, CAMP, HCK, S100A12, ORM1, CYBA, PROK2, TNFAIP6, LILRB2, CD55, BPI, HIST2H2BE, DEFA4, MPO, ALOX5, CTSG	2.41E-15
	GO:0006954	Inflammatory response	20	STAT5B, PGLYRP1, CXCR1, PRDX5, HP, FPR2, TLR5, AZU1, NLRC4, HIST1H2BK, LTB4R, CSF3R, LTF, CD24, PTX3, IL18R1, CR1, IL18RAP, RNASE3, KL, CAMP, HCK, S100A12, ORM1, CYBA, PROK2, TNFAIP6, LILRB2, CD55, BPI, HIST2H2BE, DEFA4, MPO, ALOX5, CTSG	2.58E-09
	GO:0009611	Response to wounding	24	CR1, IL18RAP, KL, STAT5B, CXCR1, PRDX5, FPR2, TLR5, S100A12, AZU1, PROK2, TNFAIP6, PLSCR1, CYBA, ORM1, CD55, NLRC4, F5, LTB4R, CD59, SERPINB2, CD24, ALOX5, PTX3	1.48E-08
	GO:0016052	Carbohydrate catabolic process	8	LDHA, G6PD, PYGL, HK3, PFKFB2, MGAM, PGLYRP1, CHIT1	1.91E-04
	GO:0019318	Hexose metabolic process	10	PPP1R3D, LDHA, G6PD, PPP1R3B, PFKFB3, PYGL, MAPK14, HK3, PFKFB2, GYG1	2.60E-04
	GO:0006006	Glucose metabolic process	9	PPP1R3D, LDHA, G6PD, PPP1R3B, PYGL, MAPK14, HK3, PFKFB2, GYG1	2.76E-04
	GO:0044042	Glucan metabolic process	5	PPP1R3D, PPP1R3B, PYGL, MGAM, GYG1	6.17E-04
	GO:0006073	Cellular glucan metabolic process	5	PPP1R3D, PPP1R3B, PYGL, MGAM, GYG1	6.17E-04
	GO:0005996	Monosaccharide metabolic process	10	PPP1R3D, LDHA, G6PD, PPP1R3B, PFKFB3, PYGL, MAPK14, HK3, PFKFB2, GYG1	7.49E-04
	GO:0006112	Energy reserve metabolic process	5	PPP1R3D, PPP1R3B, KL, PYGL, GYG1	1.22E-03
	GO:0005976	Polysaccharide metabolic process	7	PPP1R3D, PPP1R3B, PYGL, MGAM, PGLYRP1, GYG1, CHIT1	1.37E-03
Down	GO:0042110	T cell activation	9	CRTAM, CD3G, CD3D, CD8A, THEMIS, BCL11B, CD2, EOMES, IL7R	1.20E-06
	GO:0046649	Lymphocyte activation	10	GPR183, CRTAM, CD3G, CD3D, CD8A, THEMIS, BCL11B, CD2, EOMES, IL7R	4.22E-06
	GO:0045321	Leukocyte activation	10	GPR183, CRTAM, CD3G, CD3D, CD8A, THEMIS, BCL11B, CD2, EOMES, IL7R	2.04E-05
	GO:0030098	Lymphocyte differentiation	7	GPR183, CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R	4.52E-05
	GO:0030217	T cell differentiation	6	CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R	5.34E-05
	GO:0001775	Cell activation	10	GPR183, CRTAM, CD3G, CD3D, CD8A, THEMIS, BCL11B, CD2, EOMES, IL7R	7.71E-05
	GO:0002521	Leukocyte differentiation	7	GPR183, CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R	1.72E-04
	GO:0030097	Hemopoiesis	8	GPR183, CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R, SOD2	6.99E-04
	GO:0048534	Hemopoietic or lymphoid organ development	8	GPR183, CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R, SOD2	1.23E-03
	GO:0002520	Immune system development	8	GPR183, CD3D, CD8A, THEMIS, BCL11B, EOMES, IL7R, SOD2	1.73E-03

Table 2.

The enriched KEGG pathways for the up- and downregulated genes involved in PPI networks.

DEGs	Terms	Names	Gene number	Gene symbol	P value
Up	hsa04640	Hematopoietic cell lineage	6	IL1R2, CR1, CD55, CD59, IL4R, CSF3R	1.04E-02
	hsa04664	Fc epsilon RI signaling pathway	5	RAC2, MAPK14, FCER1G, PRKCD, MAP2K6	3.24E-02
	hsa00480	Glutathione metabolism	4	G6PD, OPLAH, RRM2, GCLM	4.20E-02
Down	hsa04660	T cell receptor signaling pathway	6	ITK, PRKCQ, CD3G, CD3D, CD8A, RASGRP1	7.70E-04
	hsa05332	Graft-versus-host disease	4	GZMB, HLA-DPA1, HLA-DMB, KLRD1	2.30E-03
	hsa04612	Antigen processing and presentation	5	KLRC3, CD8A, HLA-DPA1, HLA-DMB, KLRD1	2.35E-03
	hsa04640	Hematopoietic cell lineage	5	CD3G, CD3D, CD8A, CD2, IL7R	2.68E-03
	hsa05310	Asthma	3	FCER1A, HLA-DPA1, HLA-DMB	1.66E-02
	hsa05340	Primary immunodeficiency	3	CD3D, CD8A, IL7R	2.38E-02
	hsa05330	Allograft rejection	3	GZMB, HLA-DPA1, HLA-DMB	2.50E-02
	hsa04940	Type I diabetes mellitus	3	GZMB, HLA-DPA1, HLA-DMB	3.34E-02
	hsa04062	Chemokine signaling pathway	5	ITK, CCR3, CX3CR1, PRKACB, CCL4	3.81E-02
	hsa05320	Autoimmune thyroid disease	3	GZMB, HLA-DPA1, HLA-DMB	4.76E-02

KEGG, Kyoto Encyclopedia of Genes and Genomes.

Furthermore, the DEGs in the PPI network of downregulated genes were essentially involved in the GO BP terms of T cell activation (including EOMES, GPR183, CD2, and CD8A, P value = 1.20E-06), lymphocyte activation (including EOMES, GPR183, CD2, and CD8A, P value = 4.22E-06), and immune system development (including EOMES, GPR183, and CD8A, P value = 1.73E-03 (Table 1). These DEGs in the PPI network of downregulated genes were enriched in the KEGG pathways such as T cell receptor signaling pathway (including CD8A, P value = 7.70E-04), Graft-versus-host disease (including GZMA, P value = 2.30E-03), antigen processing and presentation (including CD8A, P value = 2.35E-03), and allograft rejection (including GZMB) (Table 2).

Discussion

Using common bioinformatics methods,²⁶ the genes of FDR <0.05 and |log₂FC| >1.2 were identified as significant DEGs, which were further screened using PPI network and enrichment analysis. In this study, a total of 541 DEGs, including 305 upregulated and 236 downregulated genes, were identified from the blood samples of severe burns patients admitted within 96 h after burn injury compared with healthy controls. These 305 upregulated genes (including STAT5B, MPO, CTSG, LTF, CAMP, and PROK2) were mainly involved in GO BP categories of defense response, inflammatory response, and response to wounding. These 236 downregulated genes (including GZMA, EOMES, GPR183, CD2, and CD8A) were essentially enriched in T cell activation, immune system development, and allograft rejection. These results indicated that those DEGs identified in this study were associated with the inflammatory response and the defense response after burn injury. And abnormal gene expressions occurred in severe burns patients admitted within 96 h after burn injury.

The dysregulation of the expressions of MMPs and their inhibitors such as TIMP1 are confirmed in diseases generating systemic inflammatory response syndrome including burn trauma.¹⁶ Ulrich et al.¹⁴ confirmed that the expression levels of MMP9 and TIMP1 showed a significant elevation only between 3 and 4 days post surgery. Moreover, the overexpression of MMP9 is demonstrated to be correlated with enhanced blood–brain barrier (BBB) permeability after traumatic brain injury,²⁷ and the increasing expression of MMPs and TIMP1 levels may be beneficial to tissue fibrosis and scar formation after burn injury.^14,15 Thus, the prognostic importance of MMPs and TIMP1 for burn trauma has been established.¹⁶ These results are inconsistent with the result in this study that the expression levels of MMP8, MMP9, and TIMP1 were upregulated in the burn patients compared with the healthy controls, demonstrating the prognostic importance of MMP8, MMP9, and TIMP1 in burn injury.

Among the DEGs after burn, genes such as STAT5B, MPO, CTSG, LTF, CAMP, and PROK2 showed a significant elevation change. All these genes are anti-inflammatory mediators. As has been shown, the constitutive activation of STAT5B is necessary for the development of Foxp3-positive Treg cell by binding directly to Foxp3 gene and enforcing it expression, which is associated with suppression of the inflammatory response.²⁸ MPO is a myeloid enzyme present in the polymorphonuclear leukemic cells,²⁹ and the synthesis of MPO is restricted to promyelocytes and selectively expressed during hemopoiesis in the granulocytic lineage.³⁰ Moreover, the expression of CAMP, a direct target of the vitamin D receptor encoding antimicrobial peptides, is strongly upregulated in myeloid cells after injury, revealing a importance of CAMP in the response to wounding.^31,32 In addition, the CTSG gene is a serine protease expressed in promyelocytes, which might influence the risk and mortality of severe sepsis in children.³³ The expression levels of all these genes have been determined by upregulation in patients after burn injury or in diseases with systemic inflammatory response via acting directly as anti-inflammatory mediators or showing anti-inflammatory activation.^34–36 Thus, the overexpression of these genes of STAT5B, MPO, CTSG, LTF, CAMP, and PROK2 revealed the anti-inflammatory responses had been greatly emerged in burn patients, and thus might be used as potential mediators to recovery from injuries.

Furthermore, as a serine protease produced by cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells, GZMA has correlation with the progression of sepsis in burn patients.³⁷ The cytolytic effector molecules GZMA and GZMB are reported to be upregulated in CTLs which are activated in severe sepsis, and their overexpression is associated with severity of sepsis.³⁸ In sepsis patients, GZMB-mediated lymphotoxicity results from megakaryocytes-platelet transcriptional alteration, thus GZMB can be served as a promising target for therapy of sepsis.³⁹ The increased expression level of GZMB has been observed during prolonged inflammation, and the upregulation of it has been implicated in several chronic inflammatory diseases. Additionally, the GZMB proteins can degrade extracellular matrix proteins including fibronectin and MMPs.⁴⁰ Moreover, the plasma GZMK levels which allows for a detailed classification of sepsis and more specific treatments can be used as a valuable marker to stage sepsis.⁴¹ In this study, however, the expression level of GZMA, GZMB, and GZMK were downregulated in burn patients, indicating that the sepsis of these patients might not be serious and that the severe burn patients had no malignant diseases with chronic and prolonged inflammatory diseases.

Other downregulated genes such as EOMES, GPR183, CD2, and CD8A were identified in this study. GPR183 also known as Epstein–Barr virus-induced gene 2 (EBI2), controls B lymphocyte migration and is expressed in macrophages and B cells.⁴² GPR183 plays an important role in the development of T-dependent early antibody responses by collaborating with the chemokine receptors such as CXCR5 to position B cells.⁴³ The CD8 molecules function as important co-receptors for class I major histocompatability complex on the surface of T lymphocytes,⁴⁴ suggesting the vital roles of them in antibody responses. The downregulation of these genes in burn patients in this study might suggest that the patients developed no antibody responses.

Conclusions

In conclusion, an integrated bioinformatics analysis has been conducted to identify genes which may be implicated in severe burn injuries in this study. We screened 541 DEGs between blood samples from severe burns patients and healthy controls. The results showed that, DEGs such as MMP9, TIMP1, PROK2, GZMA, GZMB, GZMK, and GPR183 might correlate with severe burns by being involved in the inflammatory responses. Nevertheless, further investigations are still necessary to verify the hypotheses proposed in this study and to unravel their mechanisms in severe burn injuries.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Jeschke

Chinkes

Finnerty

. (2008) The pathophysiologic response to severe burn injury. Annals of Surgery 248: 387–401.

Przkora

Barrow

Jeschke

. (2006) Body composition changes with time in pediatric burn patients. Journal of Trauma-Injury, Infection, and Critical Care 60: 968–971.

Przkora

Jeschke

Barrow

. (2005) Metabolic and hormonal changes of severely burned children receiving long-term oxandrolone treatment. Annals of Surgery 242: 384–389, discussion 390–391.

Finnerty

Herndon

Przkora

. (2006) Cytokine expression profile over time in severely burned pediatric patients. Shock 26: 13–19.

Fitzwater

Purdue

Hunt

. (2003) The risk factors and time course of sepsis and organ dysfunction after burn trauma. Journal of Trauma-Injury, Infection, and Critical Care 54: 959–966.

Lavrentieva

Kontakiotis

Lazaridis

. (2007) Inflammatory markers in patients with severe burn injury: What is the best indicator of sepsis? Burns 33: 189–194.

Barati

Alinejad

Bahar

. (2008) Comparison of WBC, ESR, CRP and PCT serum levels in septic and non-septic burn cases. Burns 34: 770–774.

Von Heimburg

Stieghorst

Khorram-Sefat

. (1998) Procalcitonin—a sepsis parameter in severe burn injuries. Burns 24: 745–750.

Venet

Tissot

Debard

A-L

. (2007) Decreased monocyte human leukocyte antigen-DR expression after severe burn injury: Correlation with severity and secondary septic shock. Critical Care Medicine 35: 1910–1917.

10.

Cheron

Floccard

Allaouchiche

. (2010) Lack of recovery in monocyte human leukocyte antigen-DR expression is independently associated with the development of sepsis after major trauma. Critical Care 14: R208.

11.

Landelle

Lepape

Voirin

. (2010) Low monocyte human leukocyte antigen-DR is independently associated with nosocomial infections after septic shock. Intensive Care Medicine 36: 1859–1866.

12.

Jeschke

Bolder

Chung

. (2007) Gut mucosal homeostasis and cellular mediators after severe thermal trauma and the effect of insulin-like growth factor-I in combination with insulin-like growth factor binding protein-3. Endocrinology 148: 354–362.

13.

Woodside

Spies

X-W

. (2003) Decreased lymphocyte apoptosis by anti-tumor necrosis factor antibody in Peyer’s patches after severe burn. Shock 20: 70–73.

14.

Ulrich

Noah

E-M

von Heimburg

. (2003) TIMP-1, MMP-2, MMP-9, and PIIINP as serum markers for skin fibrosis in patients following severe burn trauma. Plastic and Reconstructive Surgery 111: 1423–1431.

15.

Dasu

Spies

Barrow

. (2003) Matrix metalloproteinases and their tissue inhibitors in severely burned children. Wound Repair and Regeneration 11: 177–180.

16.

Nagy

Szélig

Rendeki

. (2015) Dynamic changes of matrix metalloproteinase 9 and tissue inhibitor of metalloproteinase 1 after burn injury. Journal of Critical Care 30: 162–166.

17.

Seok

Warren

Cuenca

. (2013) Genomic responses in mouse models poorly mimic human inflammatory diseases. Proceedings of the National Academy of Sciences of the United States of America 110: 3507–3512.

18.

Gautier

Cope

Bolstad

. (2004) affy—analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20: 307–315.

19.

Smyth

(2005) Limma: linear models for microarray data. In: Huber

Gentleman

(eds) Bioinformatics and Computational Biology Solutions Using R and Bioconductor. New York: Springer, pp. 397–420.

20.

Ferreira

Zwinderman

(2006) On the Benjamini–Hochberg method. Annals of Statistics 34: 1827–1849.

21.

Von Mering

Huynen

Jaeggi

. (2003) STRING: A database of predicted functional associations between proteins. Nucleic Acids Research 31: 258–261.

22.

Kohl

Wiese

Warscheid

(2011) Cytoscape: Software for visualization and analysis of biological networks. In: Hamacher

Eisenacher

Stephan

(eds) Data Mining in Proteomics. New York: Springer, pp. 291–303.

23.

Gene Ontology Consortium (2006) The gene ontology (GO) project in 2006. Nucleic Acids Research 34: D322–D326.

24.

Altermann

Klaenhammer

(2005) PathwayVoyager: Pathway mapping using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. BMC Genomics 6: 60.

25.

Alvord

Roayaei

Stephens

. (2007) The DAVID Gene Functional Classification Tool: A novel biological module-centric algorithm to functionally analyze large gene lists. Genome Biology 8: R183.

26.

Kapetanovic

Rosenfeld

Izmirlian

(2004) Overview of commonly used bioinformatics methods and their applications. Annals of the New York Academy of Sciences 1020: 10–21.

27.

Wang

H-H

Zhao

X-J

. (2008) Expression and significance of brain MMP-9 in rats at the early stages of severe scald. Journal of Yichun College 4: 035.

28.

O’Shea

Murray

(2008) Cytokine signaling modules in inflammatory responses. Immunity 28: 477–487.

29.

Kizaki

Miller

Selsted

. (1994) Myeloperoxidase (MPO) gene mutation in hereditary MPO deficiency. Blood 83: 1935–1940.

30.

Zhu

Bennett

MacGregor

. (1994) A myeloid-lineage-specific enhancer upstream of the mouse myeloperoxidase (MPO) gene. Leukemia 8: 717–723.

31.

Gombart

Borregaard

Koeffler

(2005) Human cathelicidin antimicrobial peptide (CAMP) gene is a direct target of the vitamin D receptor and is strongly up-regulated in myeloid cells by 1, 25-dihydroxyvitamin D3. FASEB Journal 19: 1067–1077.

32.

Termén

Tollin

Rodriguez

. (2008) PU. 1 and bacterial metabolites regulate the human gene CAMP encoding antimicrobial peptide LL-37 in colon epithelial cells. Molecular Immunology 45: 3947–3955.

33.

Sipahi

Pocan

Akar

(2006) Effect of various genetic polymorphisms on the incidence and outcome of severe sepsis. Clinical and Applied Thrombosis/Hemostasis 12: 47–54.

34.

Duan

Chai

Sheng

. (2009) Effect of burn injury on apoptosis and expression of apoptosis-related genes/proteins in skeletal muscles of rats. Apoptosis 14: 52–65.

35.

Gorowiec

Catalano

Norman

. (2011) Prokineticin 1 induces inflammatory response in human myometrium: A potential role in initiating term and preterm parturition. American Journal of Pathology 179: 2709–2719.

36.

Koyani

Flemmig

Malle

. (2015) Myeloperoxidase scavenges peroxynitrite: A novel anti-inflammatory action of the heme enzyme. Archives of Biochemistry and Biophysics 571: 1–9.

37.

Accardo-Palumbo

D’Amelio

Pileri

. (2010) Reduction of plasma granzyme A correlates with severity of sepsis in burn patients. Burns 36: 811–818.

38.

Napoli

Fast

Gardiner

. (2012) Increased granzyme levels in cytotoxic T lymphocytes are associated with disease severity in emergency department patients with severe sepsis. Shock 37: 257–262.

39.

Freishtat

Natale

Benton

. (2009) Sepsis alters the megakaryocyte–platelet transcriptional axis resulting in granzyme B–mediated lymphotoxicity. American Journal of Respiratory and Critical Care Medicine 179: 467–473.

40.

Parkinson

Toro

Granville

(2014) Granzyme B contributes to extracellular matrix degradation in UV-treated skin (1046.1). FASEB Journal 28: 1046.1.

41.

Rucevic

Fast

Jay

. (2007) Altered levels and molecular forms of granzyme K in plasma from septic patients. Shock 27: 488–493.

42.

Heinig

Petretto

Wallace

. (2010) A trans-acting locus regulates an anti-viral expression network and type 1 diabetes risk. Nature 467: 460–464.

43.

Hulse

Norton

Suh

. (2013) Chronic rhinosinusitis with nasal polyps is characterized by B-cell inflammation and EBV-induced protein 2 expression. Journal of Allergy and Clinical Immunology 131: 1075–1083.

44.

Banan

Rojas

Lee

W-H

. (1997) Interaction of the nuclear matrix-associated region (MAR)-binding proteins, SATB1 and CDP/Cux, with a MAR element (L2a) in an upstream regulatory region of the mouse CD8a gene. Journal of Biological Chemistry 272: 18440–18452.