Potential of Artificial Intelligence to Accelerate Drug Development for Rare Diseases.

The growth in breadth and depth of artificial intelligence (AI) applications has been fast, running hand in hand with the increasing amount of digital data available. Here, we comment on the application of AI in the field of drug development, with a strong focus on the specific achievements and challenges posed by rare diseases. Data paucity and high costs make drug development for rare diseases especially hard. AI can enable otherwise inaccessible approaches based on the large-scale integration of heterogeneous datasets and knowledge bases, guided by expert biological understanding. Obstacles still exist for the routine use of AI in the usually conservative pharmaceutical domain, which can easily become disillusioned. It is crucial to acknowledge that AI is a powerful, supportive tool that can assist but not replace human expertise in the various phases and aspects of drug discovery and development.

Transfer learning of clinical outcomes from preclinical molecular data, principles and perspectives.

Accurate transfer learning of clinical outcomes from one cellular context to another, between cell types, developmental stages, omics modalities or species, is considered tremendously useful. When transferring a prediction task from a source domain to a target domain, what counts is the high quality of the predictions in the target domain, requiring states or processes common to both the source and the target that can be learned by the predictor reflected by shared denominators. These may form a compendium of knowledge that is learned in the source to enable predictions in the target, usually with few, if any, labeled target training samples to learn from. Transductive transfer learning refers to the learning of the predictor in the source domain, transferring its outcome label calculations to the target domain, considering the same task. Inductive transfer learning considers cases where the target predictor is performing a different yet related task as compared with the source predictor. Often, there is also a need to first map the variables in the input/feature spaces and/or the variables in the output/outcome spaces. We here discuss and juxtapose various recently published transfer learning approaches, specifically designed (or at least adaptable) to predict clinical (human in vivo) outcomes based on preclinical (mostly animal-based) molecular data, towards finding the right tool for a given task, and paving the way for a comprehensive and systematic comparison of the suitability and accuracy of transfer learning of clinical outcomes.

A novel nutritional supplement prevents muscle loss and accelerates muscle mass recovery in caloric-restricted mice.

BACKGROUND: Muscle atrophy is defined as decreased muscle mass, associated with aging as well as with various chronic diseases and is a fundamental cause of frailty, functional decline and disability. Frailty represents a huge potential public health issue worldwide with high impact on healthcare costs. A major clinical issue is therefore to devise new strategies preventing muscle atrophy. In this study, we tested the efficacy of Vital01, a novel oral nutritional supplement (ONS), on body weight and muscle mass using a caloric restriction-induced mouse model for muscle atrophy. METHODS: Mice were calorically restricted for 2 weeks to induce muscle atrophy: one control group received 60% kcal of the normal chow diet and one intervention group received 30% kcal chow and 30 kcal% Vital01. The effects on body weight, lean body mass, muscle histology and transcriptome were assessed. In addition, the effects of Vital01, in mice with established muscle atrophy, were assessed and compared to a standard ONS. To this end, mice were first calorically restricted on a 60% kcal chow diet and then refed with either 100 kcal% chow, a mix of Vital01 (receiving 60% kcal chow and 40 kcal% Vital01) or with a mix of standard, widely prescribed ONS (receiving 60 kcal% chow and 40 kcal% Fortisip Compact). RESULTS: Vital01 attenuated weight loss (-15% weight loss for Vital01 vs. -25% for control group, p < 0.01) and loss of muscle mass (Vital01 with -13%, -12% and -18%, respectively, for gastrocnemius, quadriceps and tibialis vs. 25%, -23% and -28%, respectively, for control group, all p < 0.05) and also restored body weight, fat and muscle mass more efficiently when compared to Fortisip Compact. As assessed by transcriptome analysis and Western blotting of key proteins (e.g. phospoAKT, mTOR and S6K), Vital01 attenuated the catabolic and anabolic signaling pathways induced by caloric restriction and modulated inflammatory and mitochondrial pathways. In addition, Vital01 affected pathways related to matrix proteins/collagens homeostasis and tended to reduce caloric restriction-induced collagen fiber density in the quadriceps (with -27%, p = 0.051). CONCLUSIONS: We demonstrate that Vital01 preserves muscle mass in a calorically restricted mouse model for muscle atrophy. Vital01 had preventive effects when administered during development of muscle atrophy. Furthermore, when administered in a therapeutic setting to mice with established muscle atrophy, Vital01 rapidly restored body weight and accelerated the recurrence of fat and lean body mass more efficiently than Fortisip Compact. Bioinformatics analysis of gene expression data identified regulatory pathways that were specifically influenced by Vital01 in muscle.

FURIN Inhibition Reduces Vascular Remodeling and Atherosclerotic Lesion Progression in Mice.

Objective- Atherosclerotic coronary artery disease is the leading cause of death worldwide, and current treatment options are insufficient. Using systems-level network cluster analyses on a large coronary artery disease case-control cohort, we previously identified PCSK3 (proprotein convertase subtilisin/kexin family member 3; FURIN) as a member of several coronary artery disease-associated pathways. Thus, our objective is to determine the role of FURIN in atherosclerosis. Approach and Results- In vitro, FURIN inhibitor treatment resulted in reduced monocyte migration and reduced macrophage and vascular endothelial cell inflammatory and cytokine gene expression. In vivo, administration of an irreversible inhibitor of FURIN, α-1-PDX (α1-antitrypsin Portland), to hyperlipidemic Ldlr-/- mice resulted in lower atherosclerotic lesion area and a specific reduction in severe lesions. Significantly lower lesional macrophage and collagen area, as well as systemic inflammatory markers, were observed. MMP2 (matrix metallopeptidase 2), an effector of endothelial function and atherosclerotic lesion progression, and a FURIN substrate was significantly reduced in the aorta of inhibitor-treated mice. To determine FURIN's role in vascular endothelial function, we administered α-1-PDX to Apoe-/- mice harboring a wire injury in the common carotid artery. We observed significantly decreased carotid intimal thickness and lower plaque cellularity, smooth muscle cell, macrophage, and inflammatory marker content, suggesting protection against vascular remodeling. Overexpression of FURIN in this model resulted in a significant 67% increase in intimal plaque thickness, confirming that FURIN levels directly correlate with atherosclerosis. Conclusions- We show that systemic inhibition of FURIN in mice decreases vascular remodeling and atherosclerosis. FURIN-mediated modulation of MMP2 activity may contribute to the atheroprotection observed in these mice.

Adipose-derived human mesenchymal stem cells induce long-term neurogenic and anti-inflammatory effects and improve cognitive but not motor performance in a rat model of Parkinson's disease.

BACKGROUND: Mesenchymal stem cells (MSC) are easily harvested, and possess anti-inflammatory and trophic properties. Furthermore, MSC promote neuroprotection and neurogenesis, which could greatly benefit neurodegenerative disorders, such as Parkinson's disease. METHODS: MSC were transplanted one week after 6-hydroxydopamine lesioning and effects were evaluated after 6 months. RESULTS: MSC localized around the substantia nigra and the arachnoid mater, expressing pericyte and endothelial markers. MSC protected dopamine levels and upregulated peripheral anti-inflammatory cytokines. Furthermore, adipose-derived MSC increased neurogenesis in hippocampal and subventricular regions, and boosted memory functioning. CONCLUSION: Considering that hyposmia and loss of memory function are two major nonmotor symptoms in Parkinson's disease, transplants with modulatory effects on the hippocampus and subventricular zone could provide a disease-modifying therapy.

Human adipose-derived mesenchymal stromal cells increase endogenous neurogenesis in the rat subventricular zone acutely after 6-hydroxydopamine lesioning.

BACKGROUND AIMS: In Parkinson's disease (PD), neurogenesis in the subventricular zone (SVZ)-olfactory bulb (OB) axis is affected as the result of the lack of dopaminergic innervations reaching the SVZ. This aberrant network has been related to the hyposmia of PD patients, which is an early diagnostic marker of the disease. Consequently, much interest arose in finding mechanisms to modulate the SVZ-OB axis. Direct modulation of this axis could be achieved by transplantation of mesenchymal stromal cells (MSC), as it has been shown in rat and mouse PD models. However, the neurogenic effect of MSC in PD was thus far only analyzed weeks after transplantation, and little is known about effects immediately after transplantation. METHODS: We assessed the acute neuroprotective and neurogenic effects of adipose-derived MSC transplanted into the rat substantia nigra in the 6-hydroxydopamine model of PD. RESULTS: Three days after transplantation, subventricular neurogenesis was significantly increased in MSC-transplanted versus non-transplanted animals. Most MSC were found in the region of the substantia nigra and the surrounding arachnoid mater, expressing S100ß and brain-derived neurotrophic factor, whereas some MSC showed an endothelial phenotype and localized around blood vessels. CONCLUSIONS: The acute neurogenic effects and neurotrophic factor expression of MSC could help to restore the SVZ-OB axis in PD.

Human adipose-derived mesenchymal stem cells improve motor functions and are neuroprotective in the 6-hydroxydopamine-rat model for Parkinson's disease when cultured in monolayer cultures but suppress hippocampal neurogenesis and hippocampal memory function when cultured in spheroids.

Adult human adipose-derived mesenchymal stem cells (MSC) have been reported to induce neuroprotective effects in models for Parkinson's disease (PD). However, these effects strongly depend on the most optimal application of the transplant. In the present study we compared monolayer-cultured (aMSC) and spheroid (sMSC) MSC following transplantation into the substantia nigra (SN) of 6-OHDA lesioned rats regarding effects on the local microenvironment, degeneration of dopaminergic neurons, neurogenesis in the hippocampal DG as well as motor and memory function in the 6-OHDA-rat model for PD. aMSC transplantation significantly increased tyrosine hydroxylase (TH) and brain-derived neurotrophic factor (BDNF) levels in the SN, increased the levels of the glial fibrillary acidic protein (GFAP) and improved motor functions compared to untreated and sMSC treated animals. In contrast, sMSC grafting induced an increased local microgliosis, decreased TH levels in the SN and reduced numbers of newly generated cells in the dentate gyrus (DG) without yet affecting hippocampal learning and memory function. We conclude that the neuroprotective potential of adipose-derived MSC in the rat model of PD crucially depends on the applied cellular phenotype.

Metabolic alterations associated with schizophrenia: a critical evaluation of proton magnetic resonance spectroscopy studies.

By reviewing the existing (1) H-magnetic resonance spectroscopy literature in schizophrenia, the relationship of different sample characteristics and applied methodologies with metabolite alterations is explored. Furthermore, we emphasize common pitfalls and discrepancies in the methodological framework of the reviewed studies that introduce unwanted variation in findings and complicate the comparison of studies. A total of 92 studies were reviewed. Articles were retrieved by searching the Pubmed database. Care was taken to note down reliability and validity measures of each included study. Despite many methodological differences and shortcomings, progressive NAA reductions could be seen in several brain regions implicated in the pathogenesis of schizophrenia. In terms of treatment effects, cross-sectional evidence implicates a normalizing role for atypical antipsychotic medication; however, longitudinal studies remain inconclusive on this issue. Choline, creatine, and myo-inositol levels remain largely unchanged and a time-dependent role of glutamate finds confirmation in several spectroscopy studies. Other findings are less consistent and need further replication. Most studies lack power and methodological precision. Future studies should aim for standardization and for more distinguished study populations to gain more valid and reliable findings.

Effect of craving induction on inhibitory control in opiate dependence.

RATIONALE: Current neurobiological models of addiction posit that drug seeking is much more likely to occur during emotionally charged states (such as craving), as deficits in inhibitory control become more pronounced during heightened motivational states. OBJECTIVE: The purpose of this study was to examine the effect of cue-induced craving states on attention and inhibitory control within addicted individuals. METHODS: We tested the performance of 39 opiate-dependent individuals on cognitive measures of attention (Digit Span, Digit Symbol, and Telephone Search) and inhibitory control (Counting Stroop and Go-No-Go) both before and after exposure to an autobiographical craving script. A non-drug using healthy control group (n = 19) performed the same tasks before and after listening to a relaxation tape. RESULTS: Following craving induction, opiate-dependent individuals demonstrated improved performance on tests of processing speed and attentional span (consistent with the practice effect observed in controls) and increased their response errors on the Stroop task (in contrast to controls), while selective attention was unaffected. Individual differences in compulsivity mediated the association between craving and Stroop performance, such that low-compulsive (but not high-compulsive) individuals committed more response errors after craving induction. CONCLUSIONS: These findings challenge the notion of cue-induced craving as a primary trigger of disrupted cognition and drug-seeking behavior in addicted individuals, and raise the need to explore individual differences in compulsivity when addressing the links between craving and loss of control within research and clinical settings.