Altered Serum Proteins Suggest Inflammation, Fibrogenesis and Angiogenesis in Adult Patients with a Fontan Circulation

Abstract

Previous omics research in patients with complex congenital heart disease and single-ventricle circulation (irrespective of the stage of palliative repair) revealed alterations in cardiac and systemic metabolism, inter alia abnormalities in energy metabolism, and inflammation, oxidative stress or endothelial dysfunction. We employed an affinity-proteomics approach focused on cell surface markers, cytokines, and chemokines in the serum of 20 adult Fontan patients with a good functioning systemic left ventricle, and we 20 matched controls to reveal any specific processes on a cellular level. Analysis of 349 proteins revealed 4 altered protein levels related to chronic inflammation, with elevated levels of syndecan-1 and glycophorin-A, as well as decreased levels of leukemia inhibitory factor and nerve growth factor-ß in Fontan patients compared to controls. All in all, this means that Fontan circulation carries specific physiological and metabolic instabilities, including chronic inflammation, oxidative stress imbalance, and consequently, possible damage to cell structure and alterations in translational pathways. A combination of proteomics-based biomarkers and the traditional biomarkers (uric acid, γGT, and cholesterol) performed best in classification (patient vs. control). A metabolism- and signaling-based approach may be helpful for a better understanding of Fontan (patho-)physiology. Syndecan-1, glycophorin-A, leukemia inhibitory factor, and nerve growth factor-ß, especially in combination with uric acid, γGT, and cholesterol, might be interesting candidate parameters to complement traditional diagnostic imaging tools and the determination of traditional biomarkers, yielding a better understanding of the development of comorbidities in Fontan patients, and they may play a future role in the identification of targets to mitigate inflammation and comorbidities in Fontan patients.

Keywords: Fontan; angiogenesis; fibrogenesis; glycophorin-A; inflammation; leukemia inhibitory factor; nerve growth factor-ß; proteomics; signaling; syndecan-1.

Figure 1

Distinct abundance variations of proteins in patient and control serum samples (volcano plot). The plot visualizes the adjusted p-values and corresponding log-fold changes (|logFC|). p < 0.05 was considered statistically significant (horizontal red line). The |logFC| cutoffs are indicated as vertical lines. Proteins with a positive |logFC| had a higher abundance in patient samples; proteins with a negative value were in control samples. Proteins with |logFC| < 0.5 and a significant adjusted p-value are defined as differential and are displayed in blue. Proteins indicated in green either feature a |logFC| > 0.5, while not reaching the significance threshold, or feature a significant difference, while not reaching the |logFC| threshold.

Figure 2

Individual array values for the four proteins with differential abundance in patient and control samples (see Table 2). Each sample is measured by four replicate spots per array. Rhombs indicate sample group means. Whiskers indicate one standard deviation.

Figure 3

Heatmap displaying the relative expression of proteins identified as differential. Values were centered and scaled by proteins.

Figure 4

Principal component analysis for differential proteins. The scatter plot displays the first two principal components of the samples’ protein signal data based exclusively on the four differentially abundant proteins GLPA, LIF, SDC1, and NGF-β. In the plot, the location of the samples is defined by their first two principal components, i.e., linear combinations of protein features with the largest variance across the samples. Samples with a similar profile are located in close proximity. The percentages given in the axis labels describe the ratio of total variance explained by the respective principal component. Note that in the principal component analysis of the four differential proteins, the distribution of the probands suggests a clustering of patients or controls. Green, controls; red, patients.

Figure 5

ROC curves for selected parameters. Receiver operating characteristic curves and area under curves for selected traditional and proteomics serum parameters and for combinations thereof with regard to group assignment (patient vs. control). Note the exceeding performance of combinations including γGT, uric acid, or triglyceride serum concentration in combination with one or two of the proteomics variables. The dashed line represents an area under the curve of 0.5.

Figure 6

Normalized cumulated MDM (proteomics). Cumulative impact of the 526 proteomics-derived analytes examined on classification competence into diseased vs. non-diseased proband serum (dotted curve, red). The linear line (blue) represents the analytes’ rank standardized to 100%. Open circle: the 27/526 topmost ranked proteins (5% of analytes) account for 80% of classification optimum. MDM, mean decrease in the margin (classification impact score).