Identification of the functional pathways altered by placental cell exposure to high glucose: lessons from the transcript and metabolite interactome

The specific consequences of hyperglycaemia on placental metabolism and function are incompletely understood but likely contribute to poor pregnancy outcomes associated with diabetes mellitus (DM). This study aimed to identify the functional biochemical pathways perturbed by placental exposure to high glucose levels through integrative analysis of the trophoblast transcriptome and metabolome. The human trophoblast cell line, BeWo, was cultured in 5 or 25 mM glucose, as a model of the placenta in DM. Transcriptomic analysis using microarrays, demonstrated 5632 differentially expressed gene transcripts (≥± 1.3 fold change (FC)) following exposure to high glucose. These genes were used to generate interactome models of transcript response using BioGRID (non-inferred network: 2500 nodes (genes) and 10541 protein-protein interactions). Ultra performance-liquid chromatography-mass spectrometry (MS) and gas chromatography-MS analysis of intracellular extracts and culture medium were used to assess the response of metabolite profiles to high glucose concentration. The interactions of altered genes and metabolites were assessed using the MetScape interactome database, resulting in an integrated model of systemic transcriptome (2969 genes) and metabolome (41 metabolites) response within placental cells exposed to high glucose. The functional pathways which demonstrated significant change in response to high glucose included fatty acid β-oxidation, phospholipid metabolism and phosphatidylinositol phosphate signalling.


37"
Supplementary"Tables" 38" Suppl. Table 1: Gene changes identified by microarray analysis. BeWo cells were cultured in 5 39" mM for 24 hrs before being switched to culture in either 5 mM or 25 mM glucose for 48 h (n=6). RNA

40"
was pooled from BeWo cultured in 5 mM or 25 mM and analysed using Affymetrix exon arrays.

42"
analysis were performed in Bioconductor. The signal intensity of each transcript for each glucose

43"
condition is shown. The ratio and fold change in signal intensity of each transcript in BeWo cells

44"
cultured in 25 mM compared to 5 mM glucose is displayed.

45"
Suppl. Table 2. Biological functions of modules generated using the ModuLand algorithm on

46"
interactome networks of BeWo cells response to culture in 25 mM compared to 5 mM glucose.

47"
The number of significant modules generated from both inferred and non-inferred interactome

48"
networks are shown. Genes from each of the significant modules were entered into Reactome, to 49" identify possible biological functions associated with these genes. The biological function of the four 50" modules with the most significant p value from each interactome network are shown.

51"
Suppl. Table 3: Biological functions of modules generated using the ClusterOne algorithm on

52"
the interactome networks of BeWo cells response to culture in 25 mM compared to 5 mM

53"
glucose. The number of significant modules generated from both inferred and non-inferred

54"
interactome networks are shown. Genes from each of the significant modules were entered into

55"
Reactome, to identify possible biological functions associated with these genes. The biological

56"
function of the four modules with the most significant p value from each interactome network are

61"
overnight and then switched to 5 mM or 25 mM glucose conditions and culture media collected

62"
following a further 48 h. Samples were analysed using UPLC-MS and putatively named metabolite

63"
features that were significantly differently present in BeWo cell conditioned media following culture in

69"
were analysed using UPLC-MS and putatively named metabolite features that were significantly

70"
differently present in BeWo cells cultured in 25 mM compared to 5 mM D-glucose are shown (Kruskal-

71"
Wallis; p ≤0.01).  to an electrospray LTQ-Orbitrap hybrid mass spectrometer (ThermoFisher Scientific, Hemel 127# Hempstead, UK). A Hypersil GOLD column (100 x 2.1 mm, 1.9 µm; ThermoFisher Scientific, Runcorn, 128# UK) with a column temperature of 50°C was used. The mass spectrometer was used in negative 129# electrospray (ES-) and positive (ES+) ion modes separately, thus each sample was analysed twice, 130# once in each ion mode. Two solvents were applied to the samples during the run in each ion mode.

131#
The solvents (A and B) contained 0.1% formic acid in water (vol/vol) and 0.1% formic acid in methanol 132# (vol/vol), respectively. Both solvents were applied at a flow rate of 400 µL/min. Solvent A was held at 133# 100% for 0.5 min followed by a linear increase to 100%; solvent B was then held at 100% for another 134# 15.5 min. At 20.5 min a step change to 100% solvent A was performed and held for 1.5 min to 135# equilibrate. Column eluent from the first 90 seconds was transferred to waste, after which it was 136# transferred to the mass spectrometer.

137#
The Orbitrap mass analyser (mass resolution 30,000 at m/z = 400) was applied to collect a full scan