Freek G. Bouwman

Generally detected proteins in comparative proteomics--a matter of cellular stress response?

The specificity of proteins identified by proteomics as biomarkers for defined conditions or as components of biological processes and pathways is crucial. We critically reviewed differentially expressed proteins from comparative proteomic studies identified by 2‐DE followed by MS, especially with MALDI technique. Based on 66 of those studies, a list of 44 proteins is presented as generally detected proteins regardless of species, in vivo or in vitro conditions, tissues and organs, and experimental objective. Similarly, a list of 28 generally detected protein families is presented. The enriched functions linked to these generally detected proteins reveal that there are some common biological features beyond the technical limitations. Cellular stress response can be the universal reason as to why these proteins are generally expressed differentially. Using those proteins as biomarkers for cellular processes other than stress response should be done with caution. In future proteomic studies more profound approaches should be applied to look beyond these proteins to find specific biomarkers. Our results are discussed in relation to a recent viewpoint publication by Petrak et al. [Proteomics 2008, 8, 1744–1749].

Profiling of the secreted proteins during 3T3-L1 adipocyte differentiation leads to the identification of novel adipokines

AbstractAdipose tissue is an endocrine organ capable of secreting a number of adipokines with a role in the regulation of adipose tissue and whole-body metabolism. We used two-dimensional gel electrophoresis combined with mass spectrometry to profile the secreted proteins from (pre)adipocytes. The culture medium of 3T3-L1 cells during adipocyte differentiation was screened, and 41 proteins that responded to blocking of secretion by 20°C treatment and/or brefeldin A treatment were identified. Prohibitin, stress-70 protein, and adhesion-regulating molecule 1 are reported for the first time as secreted proteins. In addition, procollagen C-proteinase enhancer protein, galectin-1, cyclophilin A and C, and SF20/IL-25 are newly identified as adipocyte secreted factors. Secretion profiles indicated a dynamic environment including an actively remodeling extracellular matrix and several factors involved in growth regulation.

The Circulating Inactive Form of Matrix Gla Protein (ucMGP) as a Biomarker for Cardiovascular Calcification

Objective: Matrix γ-carboxyglutamate (Gla) protein (MGP) is a vitamin K-dependent protein and a strong inhibitor of vascular calcification. Vitamin K deficiency leads to inactive uncarboxylated MGP (ucMGP), which accumulates at sites of arterial calcification. We hypothesized that as a result of ucMGP deposition around arterial calcification, the circulating fraction of ucMGP is decreased. Here we report on the development of an ucMGP assay and the potential diagnostic utility of monitoring serum ucMGP levels. Methods and Results: An ELISA-based assay was developed with which circulating ucMGP can be determined. Serum ucMGP levels were measured in healthy subjects (n = 165) and in four patient populations; patients who underwent angioplasty (n = 30), patients with aortic stenosis (n = 25), hemodialysis patients (n = 52), and calciphylaxis patients (n = 10). All four patient populations had significantly lower ucMGP levels. In angioplasty patients and in those with aortic stenosis, some overlap was observed with the control population. However, in the hemodialysis and calciphylaxis populations, virtually all subjects had ucMGP levels below the normal adult range. Conclusion: Serum ucMGP may be used as a biomarker to identify those at risk for developing vascular calcification. This assay may become an important tool in the diagnosis of cardiovascular calcification.

Proteomics Investigations of Drug-Induced Hepatotoxicity in HepG2 Cells.

Unexpected hepatotoxicity is one of the major reasons of drugs failing in clinical trials. This emphasizes the need for new screening methods that address toxicological hazards early in the drug discovery process. Here, proteomics techniques were used to gain further insight into the mechanistic processes of the hepatotoxic compounds. Drug-induced hepatotoxicity is mainly divided in hepatic steatosis, cholestasis, or necrosis. For each class, a compound was selected, respectively amiodarone, cyclosporin A, and acetaminophen. The changes in protein expressions in HepG2, after exposure to these test compounds, were studied using quantitative two-dimensional differential gel electrophoresis. Identification of differentially expressed proteins was performed by Maldi-TOF/TOF MS and liquid chromatography-tandem mass spectrometry. In this study, 254 differentially expressed protein spots were detected in a two-dimensional proteome map from which 86 were identified, showing that the proteome of HepG2 cells is responsive to hepatotoxic compounds. cyclosporin A treatment was responsible for most differentially expressed proteins and could be discriminated in the hierarchical clustering analysis. The identified differential proteins show that cyclosporin A may induce endoplasmic reticulum (ER) stress and disturbs the ER-Golgi transport, with an altered vesicle-mediated transport and protein secretion as result. Moreover, the differential protein pattern seen after cyclosporin A treatment can be related to cholestatic mechanisms. Therefore, our findings indicate that the HepG2 in vitro cell system has distinctive characteristics enabling the assessment of cholestatic properties of novel compounds at an early stage of drug discovery.

Postprandial responses in hunger and satiety are associated with the rs9939609 single nucleotide polymorphism in FTO

BACKGROUND The common rs9939609 single nucleotide polymorphism (SNP) in the fat mass and obesity-associated (FTO) gene is associated with adiposity, possibly by affecting satiety responsiveness. OBJECTIVE The objective was to determine whether postprandial responses in hunger and satiety are associated with rs9939609, taking interactions with other relevant candidate genes into account. DESIGN Sixty-two women and 41 men [age: 31 +/- 14 y; body mass index (in kg/m(2)): 25.0 +/- 3.1] were genotyped for 5 SNPs in FTO, DNMT1, DNMT3B, LEP, and LEPR. Individuals received fixed meals provided in energy balance. Hunger and satiety were determined pre- and postprandially by using visual analog scales. RESULTS A general association test showed a significant association between postprandial responses in hunger and satiety with rs9939609 (P = 0.036 and P = 0.050, respectively). Individuals with low postprandial responses in hunger and satiety were overrepresented among TA/AA carriers in rs9939609 (FTO) compared with TT carriers (dominant and additive model: P = 0.013 and P = 0.020, respectively). Moreover, multifactor dimensionality reduction showed significant epistatic interactions for the postprandial decrease in hunger involving rs9939609 (FTO), rs992472 (DNMT3B), and rs1137101 (LEPR). Individuals with a low postprandial decrease in hunger were overrepresented among TA/AA (dominant), CC/CA (recessive), and AG/GG (dominant) carriers in rs9939609 (FTO), rs992472 (DNMT3B), and rs1137101 (LEPR), respectively (n = 39), compared with TT, AA, and/or AA carriers in these SNPs, respectively (P = 0.00001). Each SNP had an additional effect. CONCLUSIONS Our results confirm a role for FTO in responsiveness to hunger and satiety cues in adults in an experimental setting. The epistatic interaction suggests that DNA methylation, an epigenetic process, affects appetite.

Proteomic methodological recommendations for studies involving human plasma, platelets, and peripheral blood mononuclear cells

This study was designed to develop, optimize and validate protocols for blood processing prior to proteomic analysis of plasma, platelets and peripheral blood mononuclear cells (PBMC) and to determine analytical variation of a single sample of depleted plasma, platelet and PBMC proteins within and between four laboratories each using their own standard operating protocols for 2D gel electrophoresis. Plasma depleted either using the Beckman Coulter IgY-12 proteome partitioning kit or the Amersham albumin and IgG depletion columns gave good quality gels, but reproducibility appeared better with the single-use immuno-affinity column. The use of the Millipore Filter Device for protein concentration gave a 16% ( p < 0.005) higher recovery of protein in flow-through sample compared with acetone precipitation. The use of OptiPrep gave the lowest level of platelet contamination (1:0.8) during the isolation of PBMC from blood. Several proteins (among which are alpha-tropomyosin, fibrinogen and coagulation factor XIII A) were identified that may be used as biomarkers of platelet contamination in future studies. When identifying preselected spots, at least three out of the four centers found similar identities for 10 out of the 10 plasma proteins, 8 out of the 10 platelet proteins and 8 out of the 10 PBMC proteins. The discrepancy in spot identifications has been described before and may be explained by the mis-selection of spots due to laboratory-to-laboratory variation in gel formats, low scores on the peptide analysis leading to no or only tentative identifications, or incomplete resolution of different proteins in what appears as a single abundant spot. The average within-laboratory coefficient of variation (CV) for each of the matched spots after automatic matching using either PDQuest or ProteomWeaver software ranged between 18 and 69% for depleted plasma proteins, between 21 and 55% for platelet proteins, and between 22 and 38% for PBMC proteins. Subsequent manual matching improved the CV with on average between 1 and 16%. The average between laboratory CV for each of the matched spots after automatic matching ranged between 4 and 54% for depleted plasma proteins, between 5 and 60% for platelet proteins, and between 18 and 70% for PBMC proteins. This variation must be considered when designing sufficiently powered studies that use proteomics tools for biomarker discovery. The use of tricine in the running buffer for the second dimension appears to enhance the resolution of proteins especially in the high molecular weight range.

The circulating PBEF/NAMPT/visfatin level is associated with a beneficial blood lipid profile

Visfatin with the official gene name pre-B cell colony-enhancing factor 1 (PBEF) and the protein name nicotinamide phosphoribosyltransferase (NAMPT) is a recently discovered adipocyte-secreted protein that was shown by some to be associated with visceral fat and insulin resistance. To explore the link between PBEF/NAMPT/visfatin and lipid metabolism, we analyzed the relation of its plasma level with several parameters of adiposity, insulin resistance and the circulating blood lipid profile in a group of general population (n = 40) and a group of subjects who are genetically predisposed to insulin resistance and hyperlipidemia (n = 35). In both groups and pooled cohort, PBEF/NAMPT/visfatin lacked association with whole body adiposity, but correlated positively with HDL-cholesterol and negatively with triglycerides. The data suggested a negative correlation of the PBEF level with visceral fat and insulin resistance. But this negative correlation completely disappeared after adjustment for lipid profile. We concluded that circulating PBEF/NAMPT/visfatin level is an indicator of beneficial lipid profile in non-diabetic Caucasian subjects. The relation to lipid metabolism does not depend on visceral obesity and insulin resistance, but may be linked to its enzymatic function in NAD metabolism.

Comparative proteomic analysis of cell lines and scrapings of the human intestinal epithelium

BackgroundIn vitro models are indispensable study objects in the fields of cell and molecular biology, with advantages such as accessibility, homogeneity of the cell population, reproducibility, and growth rate. The Caco-2 cell line, originating from a colon carcinoma, is a widely used in vitro model for small intestinal epithelium. Cancer cells have an altered metabolism, making it difficult to infer their representativity for the tissue from which they are derived. This study was designed to compare the protein expression pattern of Caco-2 cells with the patterns of intestinal epithelial cells from human small and large intestine. HT-29 intestinal cells, Hep G2 liver cells and TE 671 muscle cells were included too, the latter two as negative controls.ResultsTwo-dimensional gel electrophoresis was performed on each tissue and cell line protein sample. Principal component and cluster analysis revealed that global expression of intestinal epithelial scrapings differed from that of intestinal epithelial cell lines. Since all cultured cell lines clustered together, this finding was ascribed to an adaptation of cells to culture conditions and their tumor origin, and responsible proteins were identified by mass spectrometry. When investigating the profiles of Caco-2 cells and small intestinal cells in detail, a considerable overlap was observed.ConclusionNumerous proteins showed a similar expression in Caco-2 cells, HT-29 cells, and both the intestinal scrapings, of which some appear to be characteristic to human intestinal epithelium in vivo. In addition, several biologically significant proteins are expressed at comparable levels in Caco-2 cells and small intestinal scrapings, indicating the usability of this in vitro model. Caco-2 cells, however, appear to over-express as well as under-express certain proteins, which needs to be considered by scientists using this cell line. Hence, care should be taken to prevent misinterpretation of in vitro obtained findings when translating them to the in vivo situation.

Genome scan for adiposity in Dutch dyslipidemic families reveals novel quantitative trait loci for leptin, body mass index and soluble tumor necrosis factor receptor superfamily 1A

OBJECTIVE: To search for novel genes contributing to adiposity in familial combined hyperlipidemia (FCH), a disorder characterized by abdominal obesity, hyperlipidemia and insulin resistance, using a 10 cM genome-wide scan.DESIGN: Plasma leptin and soluble tumor necrosis factor receptor superfamily members 1A and 1B (sTNFRSF1A and sTNFRSF1B, also known as sTNFR1 and sTNFR2) were analyzed as unadjusted and adjusted quantitative phenotypes of adiposity, in addition to body mass index (BMI), in multipoint and single-point analyses. In the second stage of analysis, an important chromosome 1 positional candidate gene, the leptin receptor (LEPR), was studied.SUBJECTS: Eighteen Dutch pedigrees with familial combined hyperlipidemia (FCH) (n=198) were analyzed to search for chromosomal regions harboring genes contributing to adiposity.RESULTS: Multipoint analysis of the genome scan data identified linkage (log of odds, LOD, 3.4) of leptin levels to a chromosomal region defined by D1S3728 and D1S1665, flanking the leptin receptor (LEPR) gene by approximately 9 and 3 cM, respectively. The LOD score decreased to 1.8 with age- and gender-adjusted leptin levels. Notably, BMI also mapped to this region with an LOD score of 1.2 (adjusted BMI: LOD 0.5). Two polymorphic DNA markers in LEPR and their haplotypes revealed linkage to unadjusted and adjusted BMI and leptin, and an association with leptin levels was found as well. In addition, the marker D8S1110 showed linkage (LOD 2.8) with unadjusted plasma concentrations of soluble TNFRSF1A. BMI gave a LOD score of 0.6. Moreover, a chromosome 10 q-ter locus, AFM198ZB, showed linkage with adjusted BMI (LOD 3.3).CONCLUSION: These data provide evidence that a human chromosome 1 locus, harboring the LEPR gene, contributes to plasma leptin concentrations, adiposity and body weight in humans affected with this insulin resistant dyslipidemic syndrome. Novel loci on chromosome 8 and 10 qter need further study.

Absence of an adipogenic effect of rosiglitazone on mature 3T3-L1 adipocytes: increase of lipid catabolism and reduction of adipokine expression

Aims/hypothesisThe thiazolidinedione (TZD) rosiglitazone is a peroxisome proliferator-activated receptor-γ agonist that induces adipocyte differentiation and, hence, lipid accumulation. This is in apparent contrast to the long-term glucose-lowering, insulin-sensitising effect of rosiglitazone. We tested whether the action of rosiglitazone involves specific effects on mature adipocytes, which are different from those on preadipocytes.Materials and methodsDifferentiated mature 3T3-L1 adipocytes were used as an in vitro model. Transcriptomics, proteomics and assays of metabolism were applied to assess the effect of rosiglitazone in different insulin and glucose conditions.ResultsRosiglitazone does not induce an increase, but rather a decrease in the lipid content of mature adipocytes. Analysis of transcriptome data, confirmed by quantitative RT-PCR and measurements of lipolysis, indicates that an altered energy metabolism may underlie this change. The pathway analysis shows a consistent picture dominated by lipid catabolism. In addition, we confirmed at both mRNA level and protein level that rosiglitazone represses adipokine expression and production, except for genes encoding adiponectin and apolipoprotein E. Moreover, transcriptome changes indicate that a general repression of genes encoding secreted proteins occurs.Conclusions/interpretationOur findings suggest that the change of adiposity as seen in vivo reflects a shift in balance between the different effects of TZDs on preadipocytes and on mature adipocytes, while the changes in circulating adipokine levels primarily result from an effect on mature adipocytes.