Basement membrane proteins in various arterial beds from individuals with and without type 2 diabetes mellitus: a proteome study.

BACKGROUND: Basement membrane (BM) accumulation is a hallmark of micro-vessel disease in diabetes mellitus (DM). We previously reported marked upregulation of BM components in internal thoracic arteries (ITAs) from type 2 DM (T2DM) patients by mass spectrometry. Here, we first sought to determine if BM accumulation is a common feature of different arteries in T2DM, and second, to identify other effects of T2DM on the arterial proteome. METHODS: Human arterial samples collected during heart and vascular surgery from well-characterized patients and stored in the Odense Artery Biobank were analysed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). We included ascending thoracic aortas (ATA) (n = 10 (type 2 DM, T2DM) and n = 10 (non-DM)); laser capture micro-dissected plaque- and media compartments from carotid plaques (n = 10 (T2DM) and n = 9 (non-DM)); and media- and adventitia compartments from ITAs (n = 9 (T2DM) and n = 7 (non-DM)). RESULTS: We first extended our previous finding of BM accumulation in arteries from T2DM patients, as 7 of 12 pre-defined BM proteins were significantly upregulated in bulk ATAs consisting of > 90% media. Although less pronounced, BM components tended to be upregulated in the media of ITAs from T2DM patients, but not in the neighbouring adventitia. Overall, we did not detect effects on BM proteins in carotid plaques or in the plaque-associated media. Instead, complement factors, an RNA-binding protein and fibrinogens appeared to be regulated in these tissues from T2DM patients. CONCLUSION: Our results suggest that accumulation of BM proteins is a general phenomenon in the medial layer of non-atherosclerotic arteries in patients with T2DM. Moreover, we identify additional T2DM-associated effects on the arterial proteome, which requires validation in future studies.

Is cardiovascular disease in patients with diabetes associated with serum levels of MMP-2, LOX, and the elastin degradation products ELM and ELM-2?

BACKGROUND: Diabetes mellitus type 2 (T2DM) is a significant risk factor for the development of cardiovascular diseases (CVDs). In a previous microarray study of internal mammary arteries from patients with and without T2DM, we observed several elastin-related genes with altered mRNA-expression in diabetic patients, namely matrix metalloproteinase 2 (MMP-2), lysyl oxidase (LOX) and elastin itself. In this study we investigate whether the serum concentrations of elastin-related proteins correlate to signs of CVD in patients with T2DM. METHODS: Blood samples from 302 type 2 diabetic patients were analysed for MMP-2, LOX, and the elastin degradation products ELM and ELM2. The results were investigated for correlations to signs of CVD in different vascular territories, as determined by myocardial perfusion scintigraphy, carotid artery thickness and ankle-brachial blood pressure index. RESULTS: T2DM patients with peripheral arterial disease (low ankle-brachial index) (PAD) display higher levels of MMP-2 and ELM compared to patients without PAD. However, none of the proteins or degradation products correlated with myocardial ischemia or a combined measure of CVD-signs, including myocardial ischemia, increased carotid thickness and decreased ankle-brachial blood pressure. CONCLUSIONS: Our results suggest that the diabetic environment affects the circulating amounts of MMP-2 and ELM in patients with PAD. However, the same connection could not be seen in diabetic patients with CVD broadly identified in three vascular territories. LOX and ELM-2 did not correlate to any type of CVD. Overall, serum levels of elastin-related molecules are only remotely related to CVD in type 2 diabetes.

A targeted multiplex mass spectrometry method for quantitation of abundant matrix and cellular proteins in formalin-fixed paraffin embedded arterial tissue.

Assessment of proteins in formalin-fixed paraffin-embedded (FFPE) tissue traditionally hinges on immunohistochemistry and immunoblotting. These methods are far from optimal for quantitative studies and not suitable for large-scale testing of multiple protein panels. In this study, we developed and optimised a novel targeted isotope dilution mass spectrometry (MS)-based method for FFPE samples, designed to quantitate 17 matrix and cytosolic proteins abundantly present in arterial tissue. Our new method was developed on FFPE human tissue samples of the internal thoracic artery obtained from coronary artery bypass graft (CABG) operations. The workflow has a limit of 60 samples per day. Assay precision improved by normalisation to both beta-actin and smooth muscle actin with inter-assay coefficients of variation (CV) ranging from 5.3% to 31.9%. To demonstrate clinical utility of the assay we analysed 40 FFPE artery specimens from two groups of patients with or without type 2 diabetes. We observed increased levels of collagen type IV α1 and α2 in patients with diabetes. The assay is scalable for larger cohorts and advantageous for pathophysiological studies in diabetes and the method is easily convertible to analysis of other proteins in FFPE artery samples. SIGNIFICANCE: This article presents a novel robust and precise targeted mass spectrometry assay for relative quantitation of a panel of abundant matrix and cellular arterial proteins in archived formalin-fixed paraffin-embedded arterial samples. We demonstrate its utility in pathophysiological studies of cardiovascular disease in diabetes.

Deep learning detects and visualizes bleeding events in electronic health records.

BACKGROUND: Bleeding is associated with a significantly increased morbidity and mortality. Bleeding events are often described in the unstructured text of electronic health records, which makes them difficult to identify by manual inspection. OBJECTIVES: To develop a deep learning model that detects and visualizes bleeding events in electronic health records. PATIENTS/METHODS: Three hundred electronic health records with International Classification of Diseases, Tenth Revision diagnosis codes for bleeding or leukemia were extracted. Each sentence in the electronic health record was annotated as positive or negative for bleeding. The annotated sentences were used to develop a deep learning model that detects bleeding at sentence and note level. RESULTS: On a balanced test set of 1178   sentences, the best-performing deep learning model achieved a sensitivity of 0.90, specificity of 0.90, and negative predictive value of 0.90. On a test set consisting of 700 notes, of which 49 were positive for bleeding, the model achieved a note-level sensitivity of 1.00, specificity of 0.52, and negative predictive value of 1.00. By using a sentence-level model on a note level, the model can explain its predictions by visualizing the exact sentence in a note that contains information regarding bleeding. Moreover, we found that the model performed consistently well across different types of bleedings. CONCLUSIONS: A deep learning model can be used to detect and visualize bleeding events in the free text of electronic health records. The deep learning model can thus facilitate systematic assessment of bleeding risk, and thereby optimize patient care and safety.