Arch Gynecol Obstet DOI 10.1007/s00404-015-3649-6
GENERAL GYNECOLOGY
Identification of biological processes and genes for gestational diabetes mellitus Yile Su • Yuanzhen Zhang
Received: 30 November 2014 / Accepted: 3 February 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Aim Gestational diabetes mellitus (GDM) is one of the most usual complications of pregnancy, while the correlations between genes and their known biological processes need to be further elucidated. Methods In the current study, microarray data GSE2956 containing a list of 435 significantly modified genes (differentially expressed genes, DEGs) were used. Genes that correspond to official gene symbols were chosen and were functional annotated for Gene Ontology (GO) and pathway analyses (p B 0.05). Then, the protein–protein interaction (PPI) network and the sub network were constructed and analyzed (combined score C0.4). Results A total of 405 DEGs including 239 up-regulated and 166 down-regulated genes were screened, and they were found mainly related to adhesion and motion, stimulus–response, and wound healing, etc. Besides, a PPI network containing 217 nodes and 644 lines was obtained. Hub genes including fibronectin 1 (FN1) and insulin-like growth factor 1 (IGF1) were down-regulated, and leptin (LEP) and calmodulin 1 (CALM1) were up-regulated. Three modules in the PPI network were mined and similar functional terms enriched by DEGs of these modules were obtained. Conclusion GO terms relevant to translation and metabolic process and their related genes CREB1, ribosomal proteins and LEP, still the inflammation-related proteins (e.g., IGF1 and CALM1) and cell adhesion-related protein FN1 may work together and be essential for GDM. This study provides insight into the cooperative
Y. Su Y. Zhang (&) Department of Obstetrics and gynecology, Zhongnan Hospital of Wuhan University, No.169, Donghu Road, Wuchang District, Wuhan 430071, China e-mail: [email protected]
interactions of metabolism and immune responses and the pathogenesis of GDM. Keywords Gestational diabetes mellitus Protein– protein interaction Differentially expressed genes Module Gene ontology
Introduction Gestational diabetes mellitus (GDM) is defined as diabetes and is first recognized during pregnancy. It is one of the most usual complications of pregnancy and is associated with increased risk of developing type 2 diabetes for both mothers and offspring [1]. Besides, in undiagnosed and untreated GDM, chance for preeclampsia and cesarean sections shows increase [2, 3]. Since the high level and diversified pregnancy hormones of pregnant women suffering GDM, the pathophysiology and pathogenesis which still remain unclear would be very complicated. GDM is not the result of disproportionate secretion of proinsulin or glucagon or defective secretion of insulin [4]. Instead, it is not only related to a remarkable peripheral resistance to insulin, but also inflammatory pathways [5], vitamin D concentrations [6], oxidative stress [7] and metabolic disorder [8], etc. Besides, some genetic alternation was found to be associated with GDM. For example, b 3-adrenergic receptor [9] and transcription factor 7-like 2 [10] polymorphism may be clinically significant for GDM. In addition, gene expression level of growth hormone (GH) [11] and brain and muscle Arnt-like protein-1 (BMAL1) [12] were closely associated with GDM. However, the precise mechanisms underlying GDM, still the correlations between gene and their involved biological processes remain largely unknown.
123
Arch Gynecol Obstet
Therefore, in our current study, microarray data were used to identify possible molecular processes and genes contributing to the progress of GDM on the whole genome level. First of all, differentially expressed genes (DEGs) were analyzed for functional enrichment annotation and interaction relationships between two coding proteins. Moreover, proteins clusters (modules) in the protein–protein interaction (PPI) network were mined and functionally annotated. These identified DEGs and functional terms enriched by them may help to further understand the molecular mechanisms of GDM.
proteins. In this study, STRING was used for the PPI network construction of DEGs with combined score C0.4. Then, Cytoscape (http://wwwcytoscape.org/) [15], which is an bioinformatics software package for visualization and analysis of biological networks with high-throughput data, together with the molecular complex detection (MCODE) [16] plugin of it was applied for exploring subnetwork modules. In addition, functional enrichment analysis of modules was executed by the Biological Networks Gene Ontology (BiNGO) [17] plugin of Cytoscape with adjusted p B 0.05 (Benjamini and Hochberg multiple testing correction) as threshold.
Methods Results Microarray data and screen of differentially expressed genes
Results of differentially expressed genes
Microarray data from Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) with an access number of GSE2956 were provided by Radaelli et al. [5]. The sample for this data was placental biopsies obtained from normal pregnant women and GDM pregnancies, respectively. These data contain one sample which is the list of 435 significantly modified genes (or DEGs). The screening threshold of these DEGs was|log2 fold change (FC)| C1. In our current study, only genes with official gene symbols were selected for subsequent analysis.
As a result, total 405 DEGs including 239 up-regulated and 166 down-regulated genes with known official gene symbols were maintained, and 30 DEGs without corresponding gene symbols were removed. The most significantly upand down-regulated genes were aminopeptidase regulator of TNFR1 shedding 1 (ARTS-1) and DEAD-box RNA helicase Y (DBY), respectively.
Functional analysis of differentially expressed genes
These results of GO and KEGG pathway functional enrichment analyses (Table 1) showed that the up-regulated DEGs were related to variety of terms including vascular smooth muscle contraction, cell adhesion and motion, and stimulus– response. Meanwhile, the down-regulated DEGs were mainly related to T cell receptor signaling and wound healing.
The database for annotation, visualization, and integrated discovery (DAVID, http://david.abcc.ncifcrf.gov/) [13] program is a database integrating comprehensive set of functional annotation of a large genes list. In the present study, gene ontology (GO) enrichment analysis including biological process (BP), molecular function (MF) and cellular component (CC) was analyzed using DAVID. Besides, based on the Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/) database, which is a bioinformatics resource and is beneficial to understand the functions and genomic perspectives, DEGsrelated KEGG pathways enrichment analysis were also performed with DAVID. During these two enrichment analyses, a value of p B 0.05 was considered as statistically significant. Protein–protein interaction network construction and subnetwork mining
Functional enrichment analysis of differentially expressed genes
The protein–protein interaction network A PPI network composed of 217 nodes (genes) and 644 edges was obtained (Fig. 1). Down-regulated DEGs including fibronectin 1 (FN1), insulin-like growth factor 1 (somatomedin C) (IGF1), FBJ murine osteosarcoma viral oncogene homolog (FOS) and prolactin (PRL) with high degree centrality were hub nodes in this network, while some other hub nodes including leptin (LEP), calmodulin 1 (phosphorylase kinase, delta) (CALM1), v-myc avian myelocytomatosis viral oncogene homolog (MYC), and E1A binding protein p300 (EP300) were up-regulated. Subnetwork construction and functional analysis
Search tool for the retrieval of interacting genes/proteins (STRING, http://www.string-db.org/) [14] is an online database for predicting functional interactions between
123
Furthermore, 3 modules (13 nodes and 39 edges in module 1, 10 nodes and 24 edges in module 2, and 12 nodes and 21
Arch Gynecol Obstet Table 1 Gene ontology and pathway enrichment analysis of differentially expressed genes (DEGs) Category
Term
Count
P value
Up-regulated DEGs KEGG_PATHWAY
hsa04270:Vascular smooth muscle contraction
9
3.34E-04
KEGG_PATHWAY
hsa03010:Ribosome
7
0.002301096
KEGG_PATHWAY
hsa04510:Focal adhesion
10
0.004004092
GOTERM_BP_FAT
GO:0009719*response to endogenous stimulus
17
1.18E-05
GOTERM_BP_FAT
GO:0009725*response to hormone stimulus
16
1.51E-05
GOTERM_BP_FAT
GO:0051270*regulation of cell motion
11
6.08E-05
GOTERM_CC_FAT
GO:0042175*nuclear envelope-endoplasmic reticulum network
13
7.14E-05
GOTERM_CC_FAT
GO:0005829*cytosol
31
1.03E-04
GOTERM_CC_FAT
GO:0005789*endoplasmic reticulum membrane
12
1.92E-04
GOTERM_MF_FAT GOTERM_MF_FAT
GO:0019838*growth factor binding GO:0032403*protein complex binding
11 12
2.75E-07 1.30E-05
GOTERM_MF_FAT
GO:0005198*structural molecule activity
20
8.86E-05
Down-regulated DEGs KEGG_PATHWAY
hsa04660:T cell receptor signaling pathway
GOTERM_BP_FAT
GO:0009611*response to wounding
5
0.03983536
GOTERM_BP_FAT GOTERM_BP_FAT GOTERM_CC_FAT
GO:0005576*extracellular region
33
2.34E-06
GOTERM_CC_FAT
GO:0044421*extracellular region part
18
2.91E-04
GOTERM_CC_FAT
GO:0005615*extracellular space
14
GOTERM_MF_FAT
GO:0003707*steroid hormone receptor activity
4
0.006147058
GOTERM_MF_FAT
GO:0005179*hormone activity
5
0.009313381
GOTERM_MF_FAT
GO:0004879*ligand-dependent nuclear receptor activity
4
0.0098002
16
9.68E-06
GO:0042060*wound healing
9
9.64E-05
GO:0010893*positive regulation of steroid biosynthetic process
3
0.001153423
8.72E-04
Biological process (BP), molecular function (MF) and cellular component (CC) Count: the number of enriched DEGs
edges in module 3, respectively) were acquired (Fig. 2). Only one gene, phosphatidylinositol-4, 5-bisphosphate 3-kinase, catalytic subunit delta (PIK3CD), was down-regulated in module 1, while the other DEGs including eukaryotic translation initiation factor 4E (EIF4E), cAMP responsive element binding protein 1 (CREB1), ribosomal protein (RP) S11, S26, S10, L27, L37A, L27A involving in module 1 were up-regulated. Down-regulated DEGs including IGF1 and FOS, and up-regulated DEGs including MYC and CALM1 were present in module 2. FN1 (down-regulated) and LEP (up-regulated) pertained to the module 3. GO-BP terms of module 1 were mainly related to translation and metabolic process; those of module 2 were mainly related to regulation and stimulus–response; while those of module 3 were scattered, for example, terms of cell adhesion and organ development (Table 2).
Discussion GDM, which is widespread in pregnant women, is not clearly declared about the underlying mechanism of it. In
the current study, a whole genome scan was performed with microarray data. A total of 239 up-regulated and 166 down-regulated DEGs were screened out and were found mainly related to adhesion and motion, and stimulus–response, wound healing, etc. Besides, hub genes including FN1, IGF1 and LEP were found, and 3 modules from the PPI network were mined. Similar functional terms enriched by DEGs in these modules were obtained. Therefore, these functional terms and genes would be essential for the progress of GDM. In our current study, it was worth noting that some DEGs particularly those presented in module 1 were related to translation and metabolic process. During the period of pregnancy, carbohydrate and lipid metabolism will change to supply the growing fetus with nutrients [18]. Therefore, the activated metabolic process seems easy to demonstrate abnormalities. Furthermore, CREB1, which was the hub gene in the PPI network and module 1, encodes a transcription factor and involves in different cellular processes including the differentiation of adipose cells [19] and the synchronization of circadian rhythmicity [20]. Daniel et al. [21] found that RPS29 and RPS7 were
123
Arch Gynecol Obstet
Fig. 1 Protein–protein interaction network of differentially expressed genes (DEGs). Circle the down-regulated DEGs. Square the upregulated DEGs. Lines the correlation between genes (proteins). The
thickness of lines (edges) is proportional to the combined score. The size of nodes is proportional to the degree of them
Fig. 2 Three subnetworks of differentially expressed genes (DEGs). a subnetwork 1. b subnetwork 2. c subnetwork 3. Square the upregulated DEGs. Lines the correlation between genes (proteins). The
thickness of lines (edges) is proportional to the combined score. The size of nodes is proportional to the degree of them
123
Arch Gynecol Obstet Table 2 Gene Ontology (GO) and pathway enrichment analysis of differentially expressed genes (DEGs) in three subnetworks
CLUSTER Subnetwork1
Subnetwork2
Subnetwork3
GO BP-ID
Count
Adj. p value
6414
Translational elongation
7
3.65E-10
6412
Translation
7
4.31E-07 7.07E-06
44267
Cellular protein metabolic process
11
19538
Protein metabolic process
11
3.91E-05
44260
Cellular macromolecule metabolic process
12
3.91E-05 3.20E-05
48518
Positive regulation of biological process
9
42221
Response to chemical stimulus
8
3.20E-05
48522
Positive regulation of cellular process
8
2.54E-04
32870
Cellular response to hormone stimulus
4
2.98E-04
31589
Cell-substrate adhesion
3
1.86E-02
48513 1525 BP biological process, Count the number of enriched DEGs. Adj. p value adjust p value
Description
32501 6936
overexpressed in GDM placentas, which implied the significant roles of ribosomal proteins obtained by our study. In addition to these DEGs, some others including LEP which is a crucial regulator of lipid metabolism were highlighted in PPI network [22]. Therefore, translation and metabolic process participating in the pathogenesis of GDM may be accomplished through genes including CREB1, ribosomal proteins and LEP, etc. In addition to module 1, some other modules and genes were acquired. DEGs including IGF1, FOS, MYC and CALM1 were involved in module 2 which had a correlation with regulation and stimulus–response. Besides, similar results were obtained after our GO and KEGG pathway functional enrichment analyses. Therefore, this result implied that the stimulus–response may be important for GDM. It is well known that harmful stimuli will activate innate and adaptive immune responses and acute inflammation. Previous studies have found the possible connection between inflammatory and GDM [23, 24], but not genes that we found in the present study, at least on a certain number of them. Stimulus–response including to organic substance and endogenous, and hormone associated genes such as IGF1 [25], CALM1 [26] and FOS [27], as well as LEP [28] regulating immunity, inflammation, and hematopoiesis [29] may also participate in the progress of GDM. In addition, cell adhesion-related genes FN1 may also be associated with GDM through multiple functions, for example, regulates T cell adhesion [30]. In conclusion, two possible molecular processes including metabolism and immune responses (inflammation) were further illustrated. CREB1, ribosomal proteins and LEP may be important in the translation and metabolic process of GDM. Meanwhile, inflammation-related proteins (e.g., IGF1 and CALM1) and cell adhesion-related protein FN1 jointly participated in the immune responses.
Organ development
7
1.86E-02
Angiogenesis
3
1.86E-02
Multicellular organismal process Muscle contraction
10
1.86E-02
3
1.86E-02
Furthermore, due to the double action of LEP in these two molecular processes, metabolism and immune responses could work together in the progress of GDM, which implied the interaction and integration of various molecular processes. However, further study and experimental verification are still needed to confirm these results. Author contribution Yile Su: Data analysis, Manuscript writing; Yuanzhen Zhang: Project development, Data Collection, Manuscript writing. Conflict of interest All authors declare that they have no conflict of interests to state.
References 1. Ziegler A-G, Wallner M, Kaiser I et al (2012) Long-term protective effect of lactation on the development of type 2 diabetes in women with recent gestational diabetes mellitus. Diabetes 61:3167–3171 2. Crowther C, Hiller J, Moss J, Mcphee A, Jeffries W, Robinson J for the Australian Carbohydrate Intolerance Study in pregnant women (ACHOIS) Trila Group (2005) Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 352:2477–2486 3. Landon MB, Spong CY, Thom E et al (2009) A multicenter, randomized trial of treatment for mild gestational diabetes. N Engl J Med 361:1339–1348 4. Ku¨hl C (1991) 2 Aetiology of gestational diabetes. Baillieres Clin Obstet Gynaecol 5:279–292 5. Radaelli T, Varastehpour A, Catalano P, Hauguel-De Mouzon S (2003) Gestational diabetes induces placental genes for chronic stress and inflammatory pathways. Diabetes 52:2951–2958 6. Palomer X, Gonza´lez-Clemente J, Blanco-Vaca F, Mauricio D (2008) Role of vitamin D in the pathogenesis of type 2 diabetes mellitus. Diabetes Obes Metab 10:185–197 7. Chen X, Scholl TO (2005) Oxidative stress: changes in pregnancy and with gestational diabetes mellitus. Curr Diabetes Rep 5:282–288
123
Arch Gynecol Obstet 8. Miehle K, Stepan H, Fasshauer M (2012) Leptin, adiponectin and other adipokines in gestational diabetes mellitus and preeclampsia. Clin Endocrinol (Oxf) 76:2–11 9. Festa A, Krugluger W, Shnawa N, Hopmeier P, Haffner SM, Schernthaner G (1999) Trp64Arg Polymorphism of the b 3-Adrenergic Receptor Gene in Pregnancy: association with Mild Gestational Diabetes Mellitus. J Clin Endocr Metab 84:1695–1699 10. Shaat N, Lernmark Karlsson E, Ivarsson S, Parikh H, Berntorp K, Groop L (2007) A variant in the transcription factor 7-like 2 (TCF7L2) gene is associated with an increased risk of gestational diabetes mellitus. Diabetologia 50:972–979 11. Ma¨nnik J, Vaas P, Rull K, Teesalu P, Laan M (2012) Differential placental expression profile of human growth hormone Chorionic Somatomammotropin genes in pregnancies with pre-eclampsia and gestational diabetes mellitus. Mol Cell Endocrinol 355:180–187 12. Pappa KI, Gazouli M, Anastasiou E, Iliodromiti Z, Antsaklis A, Anagnou NP (2013) The major circadian pacemaker ARNT-like protein-1 (BMAL1) is associated with susceptibility to gestational diabetes mellitus. Diabetes Res Clin Pract 99:151–157 13. Dennis G Jr, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biol 4:P3 14. Franceschini A, Szklarczyk D, Frankild S et al (2013) STRING v9.1: protein–protein interaction networks, with increased coverage and integration. Nucleic Acids Res 41:D808–D815 15. Kohl M, Wiese S, Warscheid B (2011) Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol 696:291–303 16. Bader GD, Hogue CW (2003) An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinformatics 4:2 17. Maere S, Heymans K, Kuiper M (2005) BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448–3449 18. Butte NF (2000) Carbohydrate and lipid metabolism in pregnancy: normal compared with gestational diabetes mellitus. Am J Clin Nutr 71:1256s–1261s
123
19. Garcia Dos Santos E, Dieudonne MN, Pecquery R, Le Moal V, Giudicelli Y, Lacasa DL (2002) Rapid nongenomic E2 effects on p42/p44 MAPK, activator protein-1, and cAMP response element binding protein in rat white adipocytes. Endocrinology 143:930–940 20. Gau D, Lemberger T, Von Gall C et al (2002) Phosphorylation of CREB Ser142 regulates light-induced phase shifts of the circadian clock. Neuron 34:245–253 21. Enquobahrie DA, Williams MA, Qiu C, Meller M, Sorensen TK (2009) Global placental gene expression in gestational diabetes mellitus. Am J Obstet Gynecol 200:206. e201–e206 (e213) 22. Reidy SP, Weber J-M (2000) Leptin: an essential regulator of lipid metabolism. Comp Biochem Physiol A Mol Integr Physiol 125:285–298 23. Di Benedetto A, Russo G, Corrado F et al (2005) Inflammatory markers in women with a recent history of gestational diabetes mellitus. J Endocrinol Invest 28:34–38 24. Magee TR, Ross MG, Wedekind L, Desai M, Kjos S, Belkacemi L (2014) Gestational diabetes mellitus alters apoptotic and inflammatory gene expression of trophobasts from human term placenta. J Diabetes Complications 28(4):448–459 25. Yakar S, Liu J-L, Fernandez AM et al (2001) Liver-specific igf-1 gene deletion leads to muscle insulin insensitivity. Diabetes 50:1110–1118 26. Kahl CR, Means AR (2003) Regulation of cell cycle progression by calcium/calmodulin-dependent pathways. Endocr Rev 24:719–736 27. Wagner EF, Eferl R (2005) Fos/AP-1 proteins in bone and the immune system. Immunol Rev 208:126–140 28. Shimomura I, Hammer RE, Ikemoto S, Brown MS, Goldstein JL (1999) Leptin reverses insulin resistance and diabetes mellitus in mice with congenital lipodystrophy. Nature 401:73–76 29. Fantuzzi G, Faggioni R (2000) Leptin in the regulation of immunity, inflammation, and hematopoiesis. J Leukoc Biol 68:437–446 30. Levite M, Cahalon L, Hershkoviz R, Steinman L, Lider O (1998) Neuropeptides, via specific receptors, regulate T cell adhesion to fibronectin. J Immunol 160:993–1000
GENERAL GYNECOLOGY
Identification of biological processes and genes for gestational diabetes mellitus Yile Su • Yuanzhen Zhang
Received: 30 November 2014 / Accepted: 3 February 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Aim Gestational diabetes mellitus (GDM) is one of the most usual complications of pregnancy, while the correlations between genes and their known biological processes need to be further elucidated. Methods In the current study, microarray data GSE2956 containing a list of 435 significantly modified genes (differentially expressed genes, DEGs) were used. Genes that correspond to official gene symbols were chosen and were functional annotated for Gene Ontology (GO) and pathway analyses (p B 0.05). Then, the protein–protein interaction (PPI) network and the sub network were constructed and analyzed (combined score C0.4). Results A total of 405 DEGs including 239 up-regulated and 166 down-regulated genes were screened, and they were found mainly related to adhesion and motion, stimulus–response, and wound healing, etc. Besides, a PPI network containing 217 nodes and 644 lines was obtained. Hub genes including fibronectin 1 (FN1) and insulin-like growth factor 1 (IGF1) were down-regulated, and leptin (LEP) and calmodulin 1 (CALM1) were up-regulated. Three modules in the PPI network were mined and similar functional terms enriched by DEGs of these modules were obtained. Conclusion GO terms relevant to translation and metabolic process and their related genes CREB1, ribosomal proteins and LEP, still the inflammation-related proteins (e.g., IGF1 and CALM1) and cell adhesion-related protein FN1 may work together and be essential for GDM. This study provides insight into the cooperative
Y. Su Y. Zhang (&) Department of Obstetrics and gynecology, Zhongnan Hospital of Wuhan University, No.169, Donghu Road, Wuchang District, Wuhan 430071, China e-mail: [email protected]
interactions of metabolism and immune responses and the pathogenesis of GDM. Keywords Gestational diabetes mellitus Protein– protein interaction Differentially expressed genes Module Gene ontology
Introduction Gestational diabetes mellitus (GDM) is defined as diabetes and is first recognized during pregnancy. It is one of the most usual complications of pregnancy and is associated with increased risk of developing type 2 diabetes for both mothers and offspring [1]. Besides, in undiagnosed and untreated GDM, chance for preeclampsia and cesarean sections shows increase [2, 3]. Since the high level and diversified pregnancy hormones of pregnant women suffering GDM, the pathophysiology and pathogenesis which still remain unclear would be very complicated. GDM is not the result of disproportionate secretion of proinsulin or glucagon or defective secretion of insulin [4]. Instead, it is not only related to a remarkable peripheral resistance to insulin, but also inflammatory pathways [5], vitamin D concentrations [6], oxidative stress [7] and metabolic disorder [8], etc. Besides, some genetic alternation was found to be associated with GDM. For example, b 3-adrenergic receptor [9] and transcription factor 7-like 2 [10] polymorphism may be clinically significant for GDM. In addition, gene expression level of growth hormone (GH) [11] and brain and muscle Arnt-like protein-1 (BMAL1) [12] were closely associated with GDM. However, the precise mechanisms underlying GDM, still the correlations between gene and their involved biological processes remain largely unknown.
123
Arch Gynecol Obstet
Therefore, in our current study, microarray data were used to identify possible molecular processes and genes contributing to the progress of GDM on the whole genome level. First of all, differentially expressed genes (DEGs) were analyzed for functional enrichment annotation and interaction relationships between two coding proteins. Moreover, proteins clusters (modules) in the protein–protein interaction (PPI) network were mined and functionally annotated. These identified DEGs and functional terms enriched by them may help to further understand the molecular mechanisms of GDM.
proteins. In this study, STRING was used for the PPI network construction of DEGs with combined score C0.4. Then, Cytoscape (http://wwwcytoscape.org/) [15], which is an bioinformatics software package for visualization and analysis of biological networks with high-throughput data, together with the molecular complex detection (MCODE) [16] plugin of it was applied for exploring subnetwork modules. In addition, functional enrichment analysis of modules was executed by the Biological Networks Gene Ontology (BiNGO) [17] plugin of Cytoscape with adjusted p B 0.05 (Benjamini and Hochberg multiple testing correction) as threshold.
Methods Results Microarray data and screen of differentially expressed genes
Results of differentially expressed genes
Microarray data from Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) with an access number of GSE2956 were provided by Radaelli et al. [5]. The sample for this data was placental biopsies obtained from normal pregnant women and GDM pregnancies, respectively. These data contain one sample which is the list of 435 significantly modified genes (or DEGs). The screening threshold of these DEGs was|log2 fold change (FC)| C1. In our current study, only genes with official gene symbols were selected for subsequent analysis.
As a result, total 405 DEGs including 239 up-regulated and 166 down-regulated genes with known official gene symbols were maintained, and 30 DEGs without corresponding gene symbols were removed. The most significantly upand down-regulated genes were aminopeptidase regulator of TNFR1 shedding 1 (ARTS-1) and DEAD-box RNA helicase Y (DBY), respectively.
Functional analysis of differentially expressed genes
These results of GO and KEGG pathway functional enrichment analyses (Table 1) showed that the up-regulated DEGs were related to variety of terms including vascular smooth muscle contraction, cell adhesion and motion, and stimulus– response. Meanwhile, the down-regulated DEGs were mainly related to T cell receptor signaling and wound healing.
The database for annotation, visualization, and integrated discovery (DAVID, http://david.abcc.ncifcrf.gov/) [13] program is a database integrating comprehensive set of functional annotation of a large genes list. In the present study, gene ontology (GO) enrichment analysis including biological process (BP), molecular function (MF) and cellular component (CC) was analyzed using DAVID. Besides, based on the Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/) database, which is a bioinformatics resource and is beneficial to understand the functions and genomic perspectives, DEGsrelated KEGG pathways enrichment analysis were also performed with DAVID. During these two enrichment analyses, a value of p B 0.05 was considered as statistically significant. Protein–protein interaction network construction and subnetwork mining
Functional enrichment analysis of differentially expressed genes
The protein–protein interaction network A PPI network composed of 217 nodes (genes) and 644 edges was obtained (Fig. 1). Down-regulated DEGs including fibronectin 1 (FN1), insulin-like growth factor 1 (somatomedin C) (IGF1), FBJ murine osteosarcoma viral oncogene homolog (FOS) and prolactin (PRL) with high degree centrality were hub nodes in this network, while some other hub nodes including leptin (LEP), calmodulin 1 (phosphorylase kinase, delta) (CALM1), v-myc avian myelocytomatosis viral oncogene homolog (MYC), and E1A binding protein p300 (EP300) were up-regulated. Subnetwork construction and functional analysis
Search tool for the retrieval of interacting genes/proteins (STRING, http://www.string-db.org/) [14] is an online database for predicting functional interactions between
123
Furthermore, 3 modules (13 nodes and 39 edges in module 1, 10 nodes and 24 edges in module 2, and 12 nodes and 21
Arch Gynecol Obstet Table 1 Gene ontology and pathway enrichment analysis of differentially expressed genes (DEGs) Category
Term
Count
P value
Up-regulated DEGs KEGG_PATHWAY
hsa04270:Vascular smooth muscle contraction
9
3.34E-04
KEGG_PATHWAY
hsa03010:Ribosome
7
0.002301096
KEGG_PATHWAY
hsa04510:Focal adhesion
10
0.004004092
GOTERM_BP_FAT
GO:0009719*response to endogenous stimulus
17
1.18E-05
GOTERM_BP_FAT
GO:0009725*response to hormone stimulus
16
1.51E-05
GOTERM_BP_FAT
GO:0051270*regulation of cell motion
11
6.08E-05
GOTERM_CC_FAT
GO:0042175*nuclear envelope-endoplasmic reticulum network
13
7.14E-05
GOTERM_CC_FAT
GO:0005829*cytosol
31
1.03E-04
GOTERM_CC_FAT
GO:0005789*endoplasmic reticulum membrane
12
1.92E-04
GOTERM_MF_FAT GOTERM_MF_FAT
GO:0019838*growth factor binding GO:0032403*protein complex binding
11 12
2.75E-07 1.30E-05
GOTERM_MF_FAT
GO:0005198*structural molecule activity
20
8.86E-05
Down-regulated DEGs KEGG_PATHWAY
hsa04660:T cell receptor signaling pathway
GOTERM_BP_FAT
GO:0009611*response to wounding
5
0.03983536
GOTERM_BP_FAT GOTERM_BP_FAT GOTERM_CC_FAT
GO:0005576*extracellular region
33
2.34E-06
GOTERM_CC_FAT
GO:0044421*extracellular region part
18
2.91E-04
GOTERM_CC_FAT
GO:0005615*extracellular space
14
GOTERM_MF_FAT
GO:0003707*steroid hormone receptor activity
4
0.006147058
GOTERM_MF_FAT
GO:0005179*hormone activity
5
0.009313381
GOTERM_MF_FAT
GO:0004879*ligand-dependent nuclear receptor activity
4
0.0098002
16
9.68E-06
GO:0042060*wound healing
9
9.64E-05
GO:0010893*positive regulation of steroid biosynthetic process
3
0.001153423
8.72E-04
Biological process (BP), molecular function (MF) and cellular component (CC) Count: the number of enriched DEGs
edges in module 3, respectively) were acquired (Fig. 2). Only one gene, phosphatidylinositol-4, 5-bisphosphate 3-kinase, catalytic subunit delta (PIK3CD), was down-regulated in module 1, while the other DEGs including eukaryotic translation initiation factor 4E (EIF4E), cAMP responsive element binding protein 1 (CREB1), ribosomal protein (RP) S11, S26, S10, L27, L37A, L27A involving in module 1 were up-regulated. Down-regulated DEGs including IGF1 and FOS, and up-regulated DEGs including MYC and CALM1 were present in module 2. FN1 (down-regulated) and LEP (up-regulated) pertained to the module 3. GO-BP terms of module 1 were mainly related to translation and metabolic process; those of module 2 were mainly related to regulation and stimulus–response; while those of module 3 were scattered, for example, terms of cell adhesion and organ development (Table 2).
Discussion GDM, which is widespread in pregnant women, is not clearly declared about the underlying mechanism of it. In
the current study, a whole genome scan was performed with microarray data. A total of 239 up-regulated and 166 down-regulated DEGs were screened out and were found mainly related to adhesion and motion, and stimulus–response, wound healing, etc. Besides, hub genes including FN1, IGF1 and LEP were found, and 3 modules from the PPI network were mined. Similar functional terms enriched by DEGs in these modules were obtained. Therefore, these functional terms and genes would be essential for the progress of GDM. In our current study, it was worth noting that some DEGs particularly those presented in module 1 were related to translation and metabolic process. During the period of pregnancy, carbohydrate and lipid metabolism will change to supply the growing fetus with nutrients [18]. Therefore, the activated metabolic process seems easy to demonstrate abnormalities. Furthermore, CREB1, which was the hub gene in the PPI network and module 1, encodes a transcription factor and involves in different cellular processes including the differentiation of adipose cells [19] and the synchronization of circadian rhythmicity [20]. Daniel et al. [21] found that RPS29 and RPS7 were
123
Arch Gynecol Obstet
Fig. 1 Protein–protein interaction network of differentially expressed genes (DEGs). Circle the down-regulated DEGs. Square the upregulated DEGs. Lines the correlation between genes (proteins). The
thickness of lines (edges) is proportional to the combined score. The size of nodes is proportional to the degree of them
Fig. 2 Three subnetworks of differentially expressed genes (DEGs). a subnetwork 1. b subnetwork 2. c subnetwork 3. Square the upregulated DEGs. Lines the correlation between genes (proteins). The
thickness of lines (edges) is proportional to the combined score. The size of nodes is proportional to the degree of them
123
Arch Gynecol Obstet Table 2 Gene Ontology (GO) and pathway enrichment analysis of differentially expressed genes (DEGs) in three subnetworks
CLUSTER Subnetwork1
Subnetwork2
Subnetwork3
GO BP-ID
Count
Adj. p value
6414
Translational elongation
7
3.65E-10
6412
Translation
7
4.31E-07 7.07E-06
44267
Cellular protein metabolic process
11
19538
Protein metabolic process
11
3.91E-05
44260
Cellular macromolecule metabolic process
12
3.91E-05 3.20E-05
48518
Positive regulation of biological process
9
42221
Response to chemical stimulus
8
3.20E-05
48522
Positive regulation of cellular process
8
2.54E-04
32870
Cellular response to hormone stimulus
4
2.98E-04
31589
Cell-substrate adhesion
3
1.86E-02
48513 1525 BP biological process, Count the number of enriched DEGs. Adj. p value adjust p value
Description
32501 6936
overexpressed in GDM placentas, which implied the significant roles of ribosomal proteins obtained by our study. In addition to these DEGs, some others including LEP which is a crucial regulator of lipid metabolism were highlighted in PPI network [22]. Therefore, translation and metabolic process participating in the pathogenesis of GDM may be accomplished through genes including CREB1, ribosomal proteins and LEP, etc. In addition to module 1, some other modules and genes were acquired. DEGs including IGF1, FOS, MYC and CALM1 were involved in module 2 which had a correlation with regulation and stimulus–response. Besides, similar results were obtained after our GO and KEGG pathway functional enrichment analyses. Therefore, this result implied that the stimulus–response may be important for GDM. It is well known that harmful stimuli will activate innate and adaptive immune responses and acute inflammation. Previous studies have found the possible connection between inflammatory and GDM [23, 24], but not genes that we found in the present study, at least on a certain number of them. Stimulus–response including to organic substance and endogenous, and hormone associated genes such as IGF1 [25], CALM1 [26] and FOS [27], as well as LEP [28] regulating immunity, inflammation, and hematopoiesis [29] may also participate in the progress of GDM. In addition, cell adhesion-related genes FN1 may also be associated with GDM through multiple functions, for example, regulates T cell adhesion [30]. In conclusion, two possible molecular processes including metabolism and immune responses (inflammation) were further illustrated. CREB1, ribosomal proteins and LEP may be important in the translation and metabolic process of GDM. Meanwhile, inflammation-related proteins (e.g., IGF1 and CALM1) and cell adhesion-related protein FN1 jointly participated in the immune responses.
Organ development
7
1.86E-02
Angiogenesis
3
1.86E-02
Multicellular organismal process Muscle contraction
10
1.86E-02
3
1.86E-02
Furthermore, due to the double action of LEP in these two molecular processes, metabolism and immune responses could work together in the progress of GDM, which implied the interaction and integration of various molecular processes. However, further study and experimental verification are still needed to confirm these results. Author contribution Yile Su: Data analysis, Manuscript writing; Yuanzhen Zhang: Project development, Data Collection, Manuscript writing. Conflict of interest All authors declare that they have no conflict of interests to state.
References 1. Ziegler A-G, Wallner M, Kaiser I et al (2012) Long-term protective effect of lactation on the development of type 2 diabetes in women with recent gestational diabetes mellitus. Diabetes 61:3167–3171 2. Crowther C, Hiller J, Moss J, Mcphee A, Jeffries W, Robinson J for the Australian Carbohydrate Intolerance Study in pregnant women (ACHOIS) Trila Group (2005) Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 352:2477–2486 3. Landon MB, Spong CY, Thom E et al (2009) A multicenter, randomized trial of treatment for mild gestational diabetes. N Engl J Med 361:1339–1348 4. Ku¨hl C (1991) 2 Aetiology of gestational diabetes. Baillieres Clin Obstet Gynaecol 5:279–292 5. Radaelli T, Varastehpour A, Catalano P, Hauguel-De Mouzon S (2003) Gestational diabetes induces placental genes for chronic stress and inflammatory pathways. Diabetes 52:2951–2958 6. Palomer X, Gonza´lez-Clemente J, Blanco-Vaca F, Mauricio D (2008) Role of vitamin D in the pathogenesis of type 2 diabetes mellitus. Diabetes Obes Metab 10:185–197 7. Chen X, Scholl TO (2005) Oxidative stress: changes in pregnancy and with gestational diabetes mellitus. Curr Diabetes Rep 5:282–288
123
Arch Gynecol Obstet 8. Miehle K, Stepan H, Fasshauer M (2012) Leptin, adiponectin and other adipokines in gestational diabetes mellitus and preeclampsia. Clin Endocrinol (Oxf) 76:2–11 9. Festa A, Krugluger W, Shnawa N, Hopmeier P, Haffner SM, Schernthaner G (1999) Trp64Arg Polymorphism of the b 3-Adrenergic Receptor Gene in Pregnancy: association with Mild Gestational Diabetes Mellitus. J Clin Endocr Metab 84:1695–1699 10. Shaat N, Lernmark Karlsson E, Ivarsson S, Parikh H, Berntorp K, Groop L (2007) A variant in the transcription factor 7-like 2 (TCF7L2) gene is associated with an increased risk of gestational diabetes mellitus. Diabetologia 50:972–979 11. Ma¨nnik J, Vaas P, Rull K, Teesalu P, Laan M (2012) Differential placental expression profile of human growth hormone Chorionic Somatomammotropin genes in pregnancies with pre-eclampsia and gestational diabetes mellitus. Mol Cell Endocrinol 355:180–187 12. Pappa KI, Gazouli M, Anastasiou E, Iliodromiti Z, Antsaklis A, Anagnou NP (2013) The major circadian pacemaker ARNT-like protein-1 (BMAL1) is associated with susceptibility to gestational diabetes mellitus. Diabetes Res Clin Pract 99:151–157 13. Dennis G Jr, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biol 4:P3 14. Franceschini A, Szklarczyk D, Frankild S et al (2013) STRING v9.1: protein–protein interaction networks, with increased coverage and integration. Nucleic Acids Res 41:D808–D815 15. Kohl M, Wiese S, Warscheid B (2011) Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol 696:291–303 16. Bader GD, Hogue CW (2003) An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinformatics 4:2 17. Maere S, Heymans K, Kuiper M (2005) BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448–3449 18. Butte NF (2000) Carbohydrate and lipid metabolism in pregnancy: normal compared with gestational diabetes mellitus. Am J Clin Nutr 71:1256s–1261s
123
19. Garcia Dos Santos E, Dieudonne MN, Pecquery R, Le Moal V, Giudicelli Y, Lacasa DL (2002) Rapid nongenomic E2 effects on p42/p44 MAPK, activator protein-1, and cAMP response element binding protein in rat white adipocytes. Endocrinology 143:930–940 20. Gau D, Lemberger T, Von Gall C et al (2002) Phosphorylation of CREB Ser142 regulates light-induced phase shifts of the circadian clock. Neuron 34:245–253 21. Enquobahrie DA, Williams MA, Qiu C, Meller M, Sorensen TK (2009) Global placental gene expression in gestational diabetes mellitus. Am J Obstet Gynecol 200:206. e201–e206 (e213) 22. Reidy SP, Weber J-M (2000) Leptin: an essential regulator of lipid metabolism. Comp Biochem Physiol A Mol Integr Physiol 125:285–298 23. Di Benedetto A, Russo G, Corrado F et al (2005) Inflammatory markers in women with a recent history of gestational diabetes mellitus. J Endocrinol Invest 28:34–38 24. Magee TR, Ross MG, Wedekind L, Desai M, Kjos S, Belkacemi L (2014) Gestational diabetes mellitus alters apoptotic and inflammatory gene expression of trophobasts from human term placenta. J Diabetes Complications 28(4):448–459 25. Yakar S, Liu J-L, Fernandez AM et al (2001) Liver-specific igf-1 gene deletion leads to muscle insulin insensitivity. Diabetes 50:1110–1118 26. Kahl CR, Means AR (2003) Regulation of cell cycle progression by calcium/calmodulin-dependent pathways. Endocr Rev 24:719–736 27. Wagner EF, Eferl R (2005) Fos/AP-1 proteins in bone and the immune system. Immunol Rev 208:126–140 28. Shimomura I, Hammer RE, Ikemoto S, Brown MS, Goldstein JL (1999) Leptin reverses insulin resistance and diabetes mellitus in mice with congenital lipodystrophy. Nature 401:73–76 29. Fantuzzi G, Faggioni R (2000) Leptin in the regulation of immunity, inflammation, and hematopoiesis. J Leukoc Biol 68:437–446 30. Levite M, Cahalon L, Hershkoviz R, Steinman L, Lider O (1998) Neuropeptides, via specific receptors, regulate T cell adhesion to fibronectin. J Immunol 160:993–1000
