Biomarker discovery and drug response prediction is central to personalized medicine, driving demand for predictive models that also offer biological insights. Biologically informed neural networks (BINNs), also known as visible neural networks (VNNs), have recently emerged as a solution to this goal. BINNs or VNNs are neural networks whose inter-layer connections are constrained based on prior knowledge from gene on-tologies pathway databases. These sparse models enhance interpretability by embedding prior knowledge into their architecture, ideally reducing the space of learnable functions to those that are biologically meaningful. In this systematic review—the first of its kind— we identify 86 recent papers implementing such models and highlight key trends in architectural design decisions, data sources and methods for evaluation. Growth in popularity of the approach is apparently mitigated by a lack of standardized terminology, tools and benchmarks.

Postpolymerization modifications are valuable techniques for creating functional polymers that are challenging to synthesize directly. This study presents aliphatic polycarbonates with pendant thiol-reactive groups for disulfide formation with mercaptans. The reductive responsive nature of this reaction allows for reversible postpolymerization modifications on biodegradable scaffolds. Six-membered cyclic carbonate monomers with pendant thiosulfonate groups were synthesized and polymerized using controlled organocatalytic ring-opening polymerization, yielding polymers with narrow molecular weight dispersities (Đ = 1.2) and intact reactive thiosulfonate side chains. Reversible modification with benzyl mercaptans achieved high degrees of disulfide modification. Additionally, thiol-reactive carbonate monomers were block-copolymerized onto polyethylene glycol (mPEG113) and then converted into benzyl disulfides, while the block copolymers’ hydroxyl end groups remained available for fluorescent dye labeling. The amphiphilic block copolymers self-assembled in water into micelles (∼33 nm diameter), capable of encapsulating hydrophobic molecules. These micelles successfully delivered hydrophobic dyes into macrophages, indicating the potential for intracellular drug delivery.

Trustworthiness is a major prerequisite for the safe application of opaque deep learning models in high-stakes domains like medicine. Understanding the decision-making process not only contributes to fostering trust but might also reveal previously unknown decision criteria of complex models that could advance the state of medical research. The discovery of decision-relevant concepts from black box models is a particularly challenging task. This study proposes Concept Discovery through Latent Diffusion-based Counterfactual Trajectories (CDCT), a novel three-step framework for concept discovery leveraging the superior image synthesis capabilities of diffusion models. In the first step, CDCT uses a Latent Diffusion Model (LDM) to generate a counterfactual trajectory dataset. This dataset is used to derive a disentangled representation of classification-relevant concepts using a Variational Autoencoder (VAE). Finally, a search algorithm is applied to identify relevant concepts in the disentangled latent space. The application of CDCT to a classifier trained on the largest public skin lesion dataset revealed not only the presence of several biases but also meaningful biomarkers. Moreover, the counterfactuals generated within CDCT show better FID scores than those produced by a previously established state-of-the-art method, while being 12 times more resource-efficient. Unsupervised concept discovery holds great potential for the application of trustworthy AI and the further development of human knowledge in various domains. CDCT represents a further step in this direction.

Background: Heart disease is the most prevalent cause of mortality around the world, with estimates up to 80 % being caused by modifiable risk factors. Counterfactual (CF) techniques on machine learning (ML) models based on mixed behavioral and clinical data enable generation of synthetic transitions from healthy to unhealthy (downward) or unhealthy to healthy (upward), in order to simulate pathways between the states. We used a public package to examine characteristics of the created new instances.

Methods: Our methodology employs a 3×3×2 analytical framework, integrating three ML models [naive Bayes (NB), random forest (RF), and support vector machine (SVM)] with three distinct CF generation algorithms [multi-objective CF (MOC), Nested and Interpolated Counterfactual Explanations (NICE), and What-If analysis] across two health outcome scenarios (heart disease to healthy and healthy to heart disease).

Results: Our findings reveal that the sensitivity of ML algorithms to the generation of CFs is relatively low, indicating robustness in the ML models’ performance irrespective of the CF methodology employed. Conversely, the choice of CF algorithm significantly influences the quality and characteristics of the generated CFs, particularly regarding the number of features altered. This result underscores the impact of CF methodologies on the interpretability and applicability of CFs in clinical settings, where understanding the minimal changes required to alter a health outcome is essential for targeted interventions. Our analysis of transitions from positive to negative health outcomes (Pos2Neg) and vice versa (Neg2Pos) unveils profound differences, highlighting the asymmetric nature of health state transitions.

Conclusions: This research contributes to the burgeoning field of explainable artificial intelligence (XAI) in healthcare, providing valuable insights into how different ML and CF approaches interact and impact the utility of predictive models in clinical decision-making. Our work emphasizes the careful selection of CF algorithms to enhance the actionability and relevance of CFs in health outcome predictions, paving the way for more nuanced and effective patient care strategies.

Intestinal ischemia-reperfusion (I/R) injury is an acute condition characterized by tissue damage resulting from restricted blood flow to the mesenteric vessels, leading to both local and systemic pathologies with a poor prognosis. Both ischemia and reperfusion trigger a series of cellular and molecular responses, with inflammatory cells serving as key regulators of the pathology. These interactions with the ischemic endothelium are mediated by multiple adhesion receptors. Several animal models have been established to mimic this pathology and investigate the involved molecular pathways. In this study, a microsurgical model of I/R injury is combined with intravital microscopy to visualize leukocyte rolling, adhesion, and neutrophil extracellular trap (NET) formation. This model is applied to transgenic mice deficient in endothelial PAR1 (F2r) to assess the impact of PAR1 on leukocyte rolling and NET formation 1 h after ischemia and immediately following reperfusion. In vivo, Acridine Orange leukocyte staining was employed, and NETs were visualized using a nucleic acid stain. Interestingly, reduced leukocyte adhesion and NET formation were observed in mice lacking the endothelial PAR1 receptor. This model enables the in vivo analysis of key regulators involved in I/R injury.

Background: Excess fibrotic remodeling causes cardiac dysfunction in ischemic heart disease, driven by MAP (mitogen-activated protein) kinase-dependent TGF-ß1 (transforming growth factor-ß1) activation by coagulation signaling of myeloid cells. How coagulation-inflammatory circuits can be specifically targeted to achieve beneficial macrophage reprogramming after myocardial infarction (MI) is not completely understood.

Methods: Mice with permanent ligation of the left anterior descending artery were used to model nonreperfused MI and analyzed by single-cell RNA sequencing, protein expression changes, confocal microscopy, and longitudinal monitoring of recovery. We probed the role of the tissue factor (TF)-FVIIa (activated factor VII)-integrin ß1-PAR2 (protease-activated receptor 2) signaling complex by utilizing genetic mouse models and pharmacological intervention.

Results: Cleavage-insensitive PAR2R38E and myeloid cell integrin ß1-deficient mice had improved cardiac function after MI compared with controls. Proximity ligation assays of monocytic cells demonstrated that colocalization of FVIIa with integrin ß1 was diminished in monocyte/macrophage FVII-deficient mice after MI. Compared with controls, F7fl/fl CX3CR1 (CX3C motif chemokine receptor 1)Cre mice showed reduced TGF-ß1 and MAP kinase activation, as well as cardiac dysfunction after MI, despite unaltered overall recruitment of myeloid cells. Single-cell mRNA sequencing of CD45 (cluster of differentiation 45)+ cells 3 and 7 days after MI uncovered a trajectory from recruited monocytes to inflammatory TF+/TREM (triggered receptor expressed on myeloid cells) 1+ macrophages requiring F7. As early as 7 days after MI, macrophage F7 deletion led to an expansion of reparative Olfml 3 (olfactomedin-like protein 3)+ macrophages and, conversely, to a reduction of TF+/TREM1+ macrophages, which were also reduced in PAR2R38E mice. Short-term treatment from days 1 to 5 after nonreperfused MI with a monoclonal antibody inhibiting the macrophage TF-FVIIa-PAR2 signaling complex without anticoagulant activity improved cardiac dysfunction, decreased excess fibrosis, attenuated vascular endothelial dysfunction, and increased survival 28 days after MI.

Conclusions: Extravascular TF-FVIIa-PAR2 complex signaling drives inflammatory macrophage polarization in ischemic heart disease. Targeting this signaling complex for specific therapeutic macrophage reprogramming following MI attenuates cardiac fibrosis and improves cardiovascular function.

No abstract available.

The gut microbiota has emerged as an environmental risk factor that affects thrombotic phenotypes in several cardiovascular diseases. Evidence includes the identification of marker species by sequencing studies of the gut microbiomes of patients with thrombotic disease, the influence of antithrombotic therapies on gut microbial diversity, and preclinical studies in mouse models of thrombosis that have demonstrated the functional effects of the gut microbiota on vascular inflammatory phenotypes and thrombus formation. In addition to impaired gut barrier function promoting low-grade inflammation, gut microbiota-derived metabolites have been shown to act on vascular cell types and promote thrombus formation. Therefore, these meta-organismal pathways that link the metabolic capacities of gut microorganisms with host immune functions have emerged as potential diagnostic markers and novel drug targets. In this Review, we discuss the link between the gut microbiota, its metabolites and thromboembolic diseases.

Introduction: Recent progress in cell isolation technologies and high-end omic technologies has allowed investigation of single cell sets across multiple omic domains and a thorough exploration of cellular function and various functional stages. While most multi-omic studies focused on dual RNA and protein analysis of single cell population, it is crucial to include lipid and metabolite profiling to comprehensively elucidate molecular mechanisms and pathways governing cell function, as well as phenotype at different functional stages.

Methods: To address this gap, a cellular lipidomics and transcriptomics phenotyping approach employing simultaneous extraction of lipids, metabolites, and RNA from single cell populations combined with untargeted cellular 4 dimensional (4D)-lipidomics profiling along with RNA sequencing was developed to enable comprehensive multi-omic molecular profiling from the lowest possible number of cells. Reference cell models were utilized to determine the minimum number of cells required for this multi-omics analysis. To demonstrate the feasibility of higher resolution cellular multi-omics in early-stage identification of cellular phenotype changes in pathological and physiological conditions we implemented this approach for phenotyping of macrophages in two different activation stages: MyD88-knockout macrophages as a cellular model for atherosclerosis protection, and wild type macrophages.

Results and Discussion: This multi-omic study enabled the determination of the lipid content remodeling in macrophages with anti-inflammatory and atherosclerotic protective function acquired by MyD88-KO, hence expedites the understanding of the molecular mechanisms behind immune cells effector functionality and of possible molecular targets for therapeutic intervention. An enriched functional role of phosphatidylcholine and plasmenyl/plasmalogens was shown here to accompany genetic changes underlying macrophages acquisition of anti-inflammatory function, finding that can serve as reference for macrophages reprogramming studies and for general immune and inflammation response to diseases.

Advanced age is associated with an increased susceptibility to Coronavirus Disease (COVID)-19 and more severe outcomes, although the underlying mechanisms are understudied. The lung endothelium is located next to infected epithelial cells and bystander inflammation may contribute to thromboinflammation and COVID-19-associated coagulopathy. Here, we investigated age-associated SARS-CoV-2 pathogenesis and endothelial inflammatory responses using humanized K18-hACE2 mice. Survival was reduced to 20% in aged mice (85–112 weeks) versus 50% in young mice (12–15 weeks) at 10 days post infection (dpi). Bulk RNA-sequencing of endothelial cells from mock and infected mice at 2dpi of both age groups (aged: 72–85 weeks; young: 15 weeks) showed substantially lower significant differentially regulated genes in infected aged mice than in young mice (712 versus 2294 genes). Viral recognition and anti-viral pathways such as RIG-I-like receptor signaling, NOD-like receptor signaling and interferon signaling were regulated in response to SARS-CoV-2. Young mice showed several fold higher interferon responses (Ifitm3, Ifit1, Isg15, Stat1) and interferon-induced chemokines (Cxcl10 and Cxcl11) than aged mice. Endothelial cells from infected young mice displayed elevated expression of chemokines (Cxcl9, Ccl2) and leukocyte adhesion markers (Icam1) underscoring that inflammation of lung endothelium during infection could facilitate leukocyte adhesion and thromboinflammation. TREM1 and acute phase response signaling were particularly prominent in endothelial cells from infected young mice. Immunohistochemistry was unable to detect viral protein in pulmonary endothelium. In conclusion, our data demonstrate that the early host response of the endothelium to SARS-CoV-2 infection declines with aging, which could be a potential contributor to disease severity.

The development of nano-sized carrier systems plays a fundamental role in immunodrug delivery and the treatment of cancer. Especially functional materials, coupled with a stimuli-responsive drug release, control the selective delivery of small molecular drugs to the target site and avoid systemic side effects. Based on this, we introduce a DBCO core-functionalized nanogel platform for pH-reversible conjugation of highly potent TLR7/8-activating imidazoquinolines and the selective targeting of macrophage mannose receptor (MMR/ CD206) expressed by immunosuppressive macrophages. DBCO-PEG4-amine functionalized polymethacrylates are synthesized by controlled RAFT polymerization and self-assembled into precursor micelles in polar aprotic solvents. Corresponding nanogels are generated via reactive ester chemistry, while conjugated DBCO–units are incorporated into the core, still accessible for click reaction with azide-functionalized structures. Regarding the preparation of targeted nanogels, trimannose equipped with azide moieties can be conjugated to the DBCO nanogels, revealing the efficient targeting of macrophages’ mannose receptor in vitro. Moreover, the broad applicability of the DBCO nanogel is demonstrated by the synthesis of an azide–containing 2-propionic-3-methylmaleic anhydride-based linker sensitive for the pH–reversible conjugation of secondary amine-modified immune modulators, such as IMDQ-Me. Via bioorthogonal DBCO click reaction, the immune modulator can reversibly be conjugated, affording pH-responsive drug-loaded nanogels that conserve the desired immune stimulatory effect in vitro. Overall, these findings highlight the potential of core–functionalized DBCO nanogels, a promising carrier system for pH–sensitive conjugated immunodrugs as well as an attractive platform for controlled targeting of MMR. Altogether, the versatile application of core-functionalized DBCO nanogels may pave the way for enhancing bioorthogonal multifunctionality inside nanocarrier systems that assist in addressing multiple targets in cancer immunotherapy.

The controlled pH-reversible conjugation of amine-functionalized molecules to nano-sized carrier systems is a promising achievement to enhance the efficacy of small molecular drugs at the target site. Various pH-responsive structures, such as ketals or hydrazones are accessible for drug delivery but suffer from high pH-gradients and elaborative modifications. The latter often further affects the specific activity of the released drugs. In this study, we establish the synthesis of a highly pH-sensitive bifunctional linker based on 2-propionic-3-methylmaleic anhydride. The underlying chemical structure enables the pH-reversible conjugation of different amines, although the attachment of primary amines competes with the formation of a pH-resistant imide structure. Remarkably, by analysis of the pH-reversible amidation profile in different solvents, the ring-opened amide structures are generated with primary aliphatic amines in diethyl ether. The formed conjugates rapidly phase separate from the reaction mixture and preserve the pH sensitivity of the linker system. Based on these findings, this manufacturing process is highly relevant in providing amine-conjugated 2-propionic-3-methylmaleic anhydride linkers and restoring their pH-responsiveness, particularly for primary amine-bearing drugs. This can pave their way for future applications, for instance, in nanomedicine.

Antiphospholipid antibodies (aPL) in primary or secondary antiphospholipid syndrome (APS) are a major cause for acquired thrombophilia, but specific interventions preventing autoimmune aPL development are an unmet clinical need. Although autoimmune aPL cross react with various coagulation regulatory proteins, lipid-reactive aPL, including those derived from patients with COVID-19, recognize the endolysosomal phospholipid lysobisphosphatidic acid presented by the cell surface-expressed endothelial protein C receptor. This specific recognition leads to complement-mediated activation of tissue factor (TF)-dependent proinflammatory signaling and thrombosis. Here, we show that specific inhibition of the TF coagulation initiation complex with nematode anticoagulant protein c2 (NAPc2) prevents the prothrombotic effects of aPL derived from patients with COVID-19 in mice and the aPL-induced proinflammatory and prothrombotic activation of monocytes. The induction of experimental APS is dependent on the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex, and NAPc2 suppresses monocyte endosomal reactive oxygen species production requiring the TF cytoplasmic domain and interferon-α secretion from dendritic cells. Latent infection with murine cytomegalovirus causes TF cytoplasmic domain-dependent development of persistent aPL and circulating phospholipid-reactive B1 cells, which is prevented by short-term intervention with NAPc2 during acute viral infection. In addition, treatment of lupus prone MRL-lpr mice with NAPc2, but not with heparin, suppresses dendritic-cell activation in the spleen, aPL production and circulating phospholipid-reactive B1 cells, and attenuates lupus pathology. These data demonstrate a convergent TF-dependent mechanism of aPL development in latent viral infection and autoimmune disease and provide initial evidence that specific targeting of the TF initiation complex has therapeutic benefits beyond currently used clinical anticoagulant strategies.

The commensal microbiota resides in a mutualistic relationship within the mammalian gut. It significantly influences the formation of capillary networks in the small intestine, not only during development but also in adulthood. Mucosal capillaries in small intestinal villus structures play a critical role for the uptake of dietary nutrients and immune regulation. Emerging studies have elucidated how the composition of gut microbiota can influence not only postnatal gut development regarding immune tolerance, nutrient absorption, and morphology but also the development and maintenance of blood and lymphatic capillaries within the small intestine. In particular, the analysis of gnotobiotic mouse models affirmed the importance of the gut microbiota, or even only single gut bacteria, in the remodeling of the small intestinal capillaries. Here, different epithelial-to-endothelial cross talk pathways, e.g. Paneth cell-derived signals, Toll-like receptor signaling, or tissue factor–protease activated receptor-1 signaling, have been reported to affect intestinal villus vascular remodeling in a microbiota-dependent fashion. In this review article, we will provide a comprehensive overview on the relevant microbiota–host interaction pathways, which have been revealed to influence angiogenesis and vascular remodeling in the small intestine.

In biomedical research, germ-free and gnotobiotic mouse models enable the mechanistic investigation of microbiota-host interactions and their role on (patho)physiology. Throughout any gnotobiotic experiment, standardized and periodic microbiological testing of defined gnotobiotic housing conditions is a key requirement. Here, we review basic principles of germ-free isolator technology, the suitability of various sterilization methods, and the use of sterility testing methods to monitor germ-free mouse colonies. We also discuss their effectiveness and limitations, and share the experience with protocols used in our facility. In addition, possible sources of isolator contamination are discussed and an overview of reported contaminants is provided.

The reactive ester approach provides access to various types of drug delivery systems. Either amphiphilic block copolymer micelles with hydrophobic cores can be generated for encapsulation of hydrophobic drugs, or they are (reversibly) crosslinked by polar molecules into nano(hydro)gel particles affording hydrophilic cores and coronas. Beyond short oligonucleotides complexation or covalent drug conjugation inside the core, a surface functionalization with targeting units is further possible to address a large variety of drug delivery scenarios. Interestingly, the reactive ester approach can thereby not only govern the nanocarriers' inner structure and surface property, but at the same time also provide strategies to prevent protein corona formation. These features are summarized in this article and underline the concept of reactive ester macromolecules as beneficial tool for assisting in drug delivery.

The gut microbiota is increasingly recognized as an actuating variable shaping vascular development and endothelial cell function in the intestinal mucosa but also affecting the microvasculature of remote organs. In the small intestine, colonization with gut microbiota and subsequent activation of innate immune pathways promotes the development of intricate capillary networks and lacteals, influencing the integrity of the gut-vascular barrier as well as nutrient uptake. Since the liver yields most of its blood supply via the portal circulation, the hepatic microcirculation steadily encounters microbiota-derived patterns and active signaling metabolites that induce changes in the organization of the liver sinusoidal endothelium, influencing immune zonation of sinusoids and impacting on metabolic processes. In addition, microbiota-derived signals may affect the vasculature of distant organ systems such as the brain and the eye microvasculature. In recent years, this gut-resident microbial ecosystem was revealed to contribute to the development of several vascular disease phenotypes.