Predictive Attributes for Developing Long COVID-A Study Using Machine Learning and Real-World Data from Primary Care Physicians in Germany.

(1) In the present study, we used data comprising patient medical histories from a panel of primary care practices in Germany to predict post-COVID-19 conditions in patients after COVID-19 diagnosis and to evaluate the relevant factors associated with these conditions using machine learning methods. (2) Methods: Data retrieved from the IQVIATM Disease Analyzer database were used. Patients with at least one COVID-19 diagnosis between January 2020 and July 2022 were selected for inclusion in the study. Age, sex, and the complete history of diagnoses and prescription data before COVID-19 infection at the respective primary care practice were extracted for each patient. A gradient boosting classifier (LGBM) was deployed. The prepared design matrix was randomly divided into train (80%) and test data (20%). After optimizing the hyperparameters of the LGBM classifier by maximizing the F2 score, model performance was evaluated using several test metrics. We calculated SHAP values to evaluate the importance of the individual features, but more importantly, to evaluate the direction of influence of each feature in our dataset, i.e., whether it is positively or negatively associated with a diagnosis of long COVID. (3) Results: In both the train and test data sets, the model showed a high recall (sensitivity) of 81% and 72% and a high specificity of 80% and 80%; this was offset, however, by a moderate precision of 8% and 7% and an F2-score of 0.28 and 0.25. The most common predictive features identified using SHAP included COVID-19 variant, physician practice, age, distinct number of diagnoses and therapies, sick days ratio, sex, vaccination rate, somatoform disorders, migraine, back pain, asthma, malaise and fatigue, as well as cough preparations. (4) Conclusions: The present exploratory study describes an initial investigation of the prediction of potential features increasing the risk of developing long COVID after COVID-19 infection by using the patient history from electronic medical records before COVID-19 infection in primary care practices in Germany using machine learning. Notably, we identified several predictive features for the development of long COVID in patient demographics and their medical histories.

Radiation Response of Human Cardiac Endothelial Cells Reveals a Central Role of the cGAS-STING Pathway in the Development of Inflammation.

Radiation-induced inflammation leading to the permeability of the endothelial barrier may increase the risk of cardiovascular disease. The aim of this study was to investigate potential mechanisms in vitro at the level of the proteome in human coronary artery endothelial cells (HCECest2) that were exposed to radiation doses of 0, 0.25, 0.5, 2.0 and 10 Gy (60Co-γ). Proteomics analysis was performed using mass spectrometry in a label-free data-independent acquisition mode. The data were validated using bioinformatics and immunoblotting. The low- and moderate-dose-irradiated samples (0.25 Gy, 0.5 Gy) showed only scarce proteome changes. In contrast, an activation of DNA-damage repair, inflammation, and oxidative stress pathways was seen after the high-dose treatments (2 and 10 Gy). The level of the DNA damage response protein DDB2 was enhanced early at the 10 Gy dose. The expression of proteins belonging to the inflammatory response or cGAS-STING pathway (STING, STAT1, ICAM1, ISG15) increased in a dose-dependent manner, showing the strongest effects at 10 Gy after one week. This study suggests a connection between the radiation-induced DNA damage and the induction of inflammation which supports the inhibition of the cGAS-STING pathway in the prevention of radiation-induced cardiovascular disease.

Data independent acquisition mass spectrometry of irradiated mouse lung endothelial cells reveals a STAT-associated inflammatory response.

Purpose: Pulmonary inflammation is an adverse consequence of radiation therapy in breast cancer. The aim of this study was to elucidate biological pathways leading to this pathology.Materials and methods: Lung endothelial cells were isolated 24 h after thorax-irradiation (sham or 10 Gy X-ray) from female C57Bl/6 mice and cultivated for 6 days.Results: Quantitative proteomic analysis of lung endothelial cells was done using data independent acquisition (DIA) mass spectrometry. The data were analyzed using Ingenuity Pathway Analysis and STRINGdb. In total, 4220 proteins were identified using DIA of which 60 were dysregulated in the irradiated samples (fold change ≥2.00 or ≤0.50; q-value <0.05). Several (12/40) upregulated proteins formed a cluster of inflammatory proteins with STAT1 and IRF3 as predicted upstream regulators. The several-fold increased expression of STAT1 and STAT-associated ISG15 was confirmed by immunoblotting. The expression of antioxidant proteins SOD1 and PRXD5 was downregulated suggesting radiation-induced oxidative stress. Similarly, the phosphorylated (active) forms of STING and IRF3, both members of the cGAS/STING pathway, were downregulated.Conclusions: These data suggest the involvement of JAK/STAT and cGas/STING pathways in the genesis of radiation-induced lung inflammation. These pathways may be used as novel targets for the prevention of radiation-induced lung damage.

CREB Signaling Mediates Dose-Dependent Radiation Response in the Murine Hippocampus Two Years after Total Body Exposure.

The impact of low-dose ionizing radiation (IR) on the human brain has recently attracted attention due to the increased use of IR for diagnostic purposes. The aim of this study was to investigate low-dose radiation response in the hippocampus. Female B6C3F1 mice were exposed to total body irradiation with 0 (control), 0.063, 0.125, or 0.5 Gy. Quantitative label-free proteomic analysis of the hippocampus was performed after 24 months. CREB signaling and CREB-associated pathways were affected at all doses. The lower doses (0.063 and 0.125 Gy) induced the CREB pathway, whereas the exposure to 0.5 Gy deactivated CREB. Similarly, the lowest dose (0.063 Gy) was anti-inflammatory, reducing the number of activated microglia. In contrast, induction of activated microglia and reactive astroglia was found at 0.5 Gy, suggesting increased inflammation and astrogliosis, respectively. The apoptotic markers BAX and cleaved CASP-3 and oxidative stress markers were increased only at the highest dose. Since the activated CREB pathway plays a central role in learning and memory, these data suggest neuroprotection at the lowest dose (0.063 Gy) but neurodegeneration at 0.5 Gy. The response to 0.5 Gy resembles alterations found in healthy aging and thus may represent radiation-induced accelerated aging of the brain.

Combined Treatment with Low-Dose Ionizing Radiation and Ketamine Induces Adverse Changes in CA1 Neuronal Structure in Male Murine Hippocampi.

In children, ketamine sedation is often used during radiological procedures. Combined exposure of ketamine and radiation at doses that alone did not affect learning and memory induced permanent cognitive impairment in mice. The aim of this study was to elucidate the mechanism behind this adverse outcome. Neonatal male NMRI mice were administered ketamine (7.5 mg kg-1) and irradiated (whole-body, 100 mGy or 200 mGy, 137Cs) one hour after ketamine exposure on postnatal day 10. The control mice were injected with saline and sham-irradiated. The hippocampi were analyzed using label-free proteomics, immunoblotting, and Golgi staining of CA1 neurons six months after treatment. Mice co-exposed to ketamine and low-dose radiation showed alterations in hippocampal proteins related to neuronal shaping and synaptic plasticity. The expression of brain-derived neurotrophic factor, activity-regulated cytoskeleton-associated protein, and postsynaptic density protein 95 were significantly altered only after the combined treatment (100 mGy or 200 mGy combined with ketamine, respectively). Increased numbers of basal dendrites and branching were observed only after the co-exposure, thereby constituting a possible reason for the displayed alterations in behavior. These data suggest that the risk of radiation-induced neurotoxic effects in the pediatric population may be underestimated if based only on the radiation dose.

Comparison of methods to isolate proteins from extracellular vesicles for mass spectrometry-based proteomic analyses.

Extracellular vesicles (EVs) are cell-derived membrane-bound organelles that have generated interest as they reflect the physiological condition of their source. Mass spectrometric (MS) analyses of protein cargo of EVs may lead to the discovery of biomarkers for diseases. However, for a comprehensive MS-based proteomics analysis, an optimal lysis of the EVs is required. Six methods for the protein extraction from EVs secreted by the head and neck cell line BHY were compared. Commercial radioimmunoprecipitation assay (RIPA) buffer outperformed the other buffers investigated in this study (Tris-SDS, Tris-Triton, GuHCl, urea-thiourea, and commercial Cell-lysis buffer). Following lysis with RIPA buffer, 310 proteins and 1469 peptides were identified using LTQ OrbitrapXL mass spectrometer. Among these, 86% of proteins and 72% of peptides were identified in all three replicates. In the case of other buffers, Tris-Triton identified on average 277 proteins, Cell-lysis buffer 257 proteins, and Tris-SDS, GuHCl and urea-thiourea each 267 proteins. In total, 399 proteins including 74 of the top EV markers (Exocarta) were identified, the most of the latter (73) using RIPA. The proteins exclusively identified using RIPA represented all Gene Ontology cell compartments. This study suggests that RIPA is an optimal lysis buffer for EVs in combination with MS.

Deciphering of the Human Interferon-Regulated Proteome by Mass Spectrometry-Based Quantitative Analysis Reveals Extent and Dynamics of Protein Induction and Repression.

Interferons (IFNs) are pleotropic cytokines secreted upon encounter of pathogens and tumors. Applying their antipathogenic, antiproliferative, and immune stimulatory capacities, recombinant IFNs are frequently prescribed as drugs to treat different diseases. IFNs act by changing the gene expression profile of cells. Due to characteristics such as rapid gene induction and signaling, IFNs also represent prototypical model systems for various aspects of biomedical research (e.g., signal transduction). In regard to the signaling and activated promoters, IFNs can be subdivided into two groups. Here, alterations of the cellular proteome of human cells treated with IFNα and IFNγ were elucidated in a time-resolved manner by quantitative proteome analysis. The majority of protein regulations were strongly IFN type and time dependent. In addition to the expected upregulation of IFN-responsive proteins, an astonishing number of proteins became profoundly repressed especially by IFNγ. Thus, our comprehensive analysis revealed important insights into the human IFN-regulated proteome and its dynamics of protein induction and repression. Interestingly, the new class of IFN-repressed genes comprises known host factors for highly relevant pathogens such as HIV, dengue virus, and hepatitis C virus.

Radiation-Induced Endothelial Inflammation Is Transferred via the Secretome to Recipient Cells in a STAT-Mediated Process.

Radiation is the most common treatment of cancer. Minimizing the normal tissue injury, especially the damage to vascular endothelium, remains a challenge. This study aimed to analyze direct and indirect radiation effects on the endothelium by investigating mechanisms of signal transfer from irradiated to nonirradiated endothelial cells by means of secreted proteins. Human coronary artery endothelial cells (HCECest2) undergo radiation-induced senescence in vitro 14 days after exposure to 10 Gy X-rays. Proteomics analysis was performed on HCECest2 14 days after irradiation with X-ray doses of 0 Gy (control) or 10 Gy using label-free technology. Additionally, the proteomes of control and radiation-induced secretomes, and those of nonirradiated HCECest2 exposed for 24 h to secreted proteins of either condition were measured. Key changes identified by proteomics and bioinformatics were validated by immunoblotting, ELISA, bead-based multiplex assays, and targeted transcriptomics. The irradiated cells, their secretome, and the nonirradiated recipient cells showed similar inflammatory response, characterized by induction of interferon type I-related proteins and activation of the STAT3 pathway. These data indicate that irradiated endothelial cells may adversely affect nonirradiated surrounding cells via senescence-associated secretory phenotype. This study adds to our knowledge of the pathological background of radiation-induced cardiovascular disease.