Multimodal workflows optimally predict response to repetitive transcranial magnetic stimulation in patients with schizophrenia: a multisite machine learning analysis

The response variability to repetitive transcranial magnetic stimulation (rTMS) challenges the effective use of this treatment option in patients with schizophrenia. This variability may be deciphered by leveraging predictive information in structural MRI, clinical, sociodemographic, and genetic data using artificial intelligence. We developed and cross-validated rTMS response prediction models in patients with schizophrenia drawn from the multisite RESIS trial. The models incorporated pre-treatment sMRI, clinical, sociodemographic, and polygenic risk score (PRS) data. Patients were randomly assigned to receive active (N = 45) or sham (N = 47) rTMS treatment. The prediction target was individual response, defined as ≥20% reduction in pre-treatment negative symptom sum scores of the Positive and Negative Syndrome Scale. Our multimodal sequential prediction workflow achieved a balanced accuracy (BAC) of 94% (non-responders: 92%, responders: 95%) in the active-treated group and 50% in the sham-treated group. The clinical, clinical + PRS, and sMRI-based classifiers yielded BACs of 65%, 76%, and 80%, respectively. Apparent sadness, inability to feel, educational attainment PRS, and unemployment were most predictive of non-response in the clinical + PRS model, while grey matter density reductions in the default mode, limbic networks, and the cerebellum were most predictive in the sMRI model. Our sequential modelling approach provided superior predictive performance while minimising the diagnostic burden in the clinical setting. Predictive patterns suggest that rTMS responders may have higher levels of brain grey matter in the default mode and salience networks which increases their likelihood of profiting from plasticity-inducing brain stimulation methods, such as rTMS. The future clinical implementation of our models requires findings to be replicated at the international scale using stratified clinical trial designs.

CD200+ fibroblasts form a pro-resolving mesenchymal network in arthritis

Fibroblasts are important regulators of inflammation, but whether fibroblasts change phenotype during resolution of inflammation is not clear. Here we use positron emission tomography to detect fibroblast activation protein (FAP) as a means to visualize fibroblast activation in vivo during inflammation in humans. While tracer accumulation is high in active arthritis, it decreases after tumor necrosis factor and interleukin-17A inhibition. Biopsy-based single-cell RNA-sequencing analyses in experimental arthritis show that FAP signal reduction reflects a phenotypic switch from pro-inflammatory MMP3+/IL6+ fibroblasts (high FAP internalization) to pro-resolving CD200+DKK3+ fibroblasts (low FAP internalization). Spatial transcriptomics of human joints indicates that pro-resolving niches of CD200+DKK3+ fibroblasts cluster with type 2 innate lymphoid cells, whereas MMP3+/IL6+ fibroblasts colocalize with inflammatory immune cells. CD200+DKK3+ fibroblasts stabilized the type 2 innate lymphoid cell phenotype and induced resolution of arthritis via CD200-CD200R1 signaling. Taken together, these data suggest a dynamic molecular regulation of the mesenchymal compartment during resolution of inflammation.

Preserving fairness and diagnostic accuracy in private large-scale AI models for medical imaging

BACKGROUND

Artificial intelligence (AI) models are increasingly used in the medical domain. However, as medical data is highly sensitive, special precautions to ensure its protection are required. The gold standard for privacy preservation is the introduction of differential privacy (DP) to model training. Prior work indicates that DP has negative implications on model accuracy and fairness, which are unacceptable in medicine and represent a main barrier to the widespread use of privacy-preserving techniques. In this work, we evaluated the effect of privacy-preserving training of AI models regarding accuracy and fairness compared to non-private training.

METHODS

We used two datasets: (1) A large dataset (N = 193,311) of high quality clinical chest radiographs, and (2) a dataset (N = 1625) of 3D abdominal computed tomography (CT) images, with the task of classifying the presence of pancreatic ductal adenocarcinoma (PDAC). Both were retrospectively collected and manually labeled by experienced radiologists. We then compared non-private deep convolutional neural networks (CNNs) and privacy-preserving (DP) models with respect to privacy-utility trade-offs measured as area under the receiver operating characteristic curve (AUROC), and privacy-fairness trade-offs, measured as Pearson's r or Statistical Parity Difference.

RESULTS

We find that, while the privacy-preserving training yields lower accuracy, it largely does not amplify discrimination against age, sex or co-morbidity. However, we find an indication that difficult diagnoses and subgroups suffer stronger performance hits in private training.

CONCLUSIONS

Our study shows that - under the challenging realistic circumstances of a real-life clinical dataset - the privacy-preserving training of diagnostic deep learning models is possible with excellent diagnostic accuracy and fairness.

Accelerating CO2 electrochemical conversion towards industrial implementation

Despite significant progress in CO2 conversion field, there remains a significant gap between fundamental research and the industrial demands. This Comment discusses key performance parameters for industrial applications and outlines current limitations in the field.

Ex vivo drug response profiling for response and outcome prediction in hematologic malignancies: the prospective non-interventional SMARTrial

Ex vivo drug response profiling is a powerful tool to study genotype-drug response associations and is being explored as a tool set for precision medicine in cancer. Here we conducted a prospective non-interventional trial to investigate feasibility of ex vivo drug response profiling for treatment guidance in hematologic malignancies (SMARTrial, NCT03488641 ). The primary endpoint to provide drug response profiling reports within 7 d was met in 91% of all study participants (N = 80). Secondary endpoint analysis revealed that ex vivo resistance to chemotherapeutic drugs predicted chemotherapy treatment failure in vivo. We confirmed the predictive value of ex vivo response to chemotherapy in a validation cohort of 95 individuals with acute myeloid leukemia treated with daunorubicin and cytarabine. Ex vivo drug response profiles improved ELN-22 risk stratification in individuals with adverse risk. We conclude that ex vivo drug response profiling is clinically feasible and has the potential to predict chemotherapy response in individuals with hematologic malignancies beyond clinically established genetic markers.

Diversity of cortical activity changes beyond depression during Spreading Depolarizations

Spreading depolarizations (SDs) are classically thought to be associated with spreading depression of cortical activity. Here, we found that SDs in patients with subarachnoid hemorrhage produce variable, ranging from depression to booming, changes in electrocorticographic activity, especially in the delta frequency band. In rats, depression of activity was characteristic of high-potassium-induced full SDs, whereas partial superficial SDs caused either little change or a boom of activity at the cortical vertex, supported by volume conduction of signals from spared delta generators in the deep cortical layers. Partial SDs also caused moderate neuronal depolarization and sustained excitation, organized in gamma oscillations in a narrow sub-SD zone. Thus, our study challenges the concept of homology between spreading depolarization and spreading depression by showing that SDs produce variable, from depression to booming, changes in activity at the cortical surface and in different cortical layers depending on the depth of SD penetration.

The integrity of dopaminergic and noradrenergic brain regions is associated with different aspects of late-life memory performance

Changes in dopaminergic neuromodulation play a key role in adult memory decline. Recent research has also implicated noradrenaline in shaping late-life memory. However, it is unclear whether these two neuromodulators have distinct roles in age-related cognitive changes. Here, combining longitudinal MRI of the dopaminergic substantia nigra-ventral tegmental area (SN-VTA) and noradrenergic locus coeruleus (LC) in younger (n = 69) and older (n = 251) adults, we found that dopaminergic and noradrenergic integrity are differentially associated with memory performance. While LC integrity was related to better episodic memory across several tasks, SN-VTA integrity was linked to working memory. Longitudinally, we found that older age was associated with more negative change in SN-VTA and LC integrity. Notably, changes in LC integrity reliably predicted future episodic memory. These differential associations of dopaminergic and noradrenergic nuclei with late-life cognitive decline have potential clinical utility, given their degeneration in several age-associated diseases.

Striated muscle-specific base editing enables correction of mutations causing dilated cardiomyopathy

Dilated cardiomyopathy is the second most common cause for heart failure with no cure except a high-risk heart transplantation. Approximately 30% of patients harbor heritable mutations which are amenable to CRISPR-based gene therapy. However, challenges related to delivery of the editing complex and off-target concerns hamper the broad applicability of CRISPR agents in the heart. We employ a combination of the viral vector AAVMYO with superior targeting specificity of heart muscle tissue and CRISPR base editors to repair patient mutations in the cardiac splice factor Rbm20, which cause aggressive dilated cardiomyopathy. Using optimized conditions, we repair >70% of cardiomyocytes in two Rbm20 knock-in mouse models that we have generated to serve as an in vivo platform of our editing strategy. Treatment of juvenile mice restores the localization defect of RBM20 in 75% of cells and splicing of RBM20 targets including TTN. Three months after injection, cardiac dilation and ejection fraction reach wild-type levels. Single-nuclei RNA sequencing uncovers restoration of the transcriptional profile across all major cardiac cell types and whole-genome sequencing reveals no evidence for aberrant off-target editing. Our study highlights the potential of base editors combined with AAVMYO to achieve gene repair for treatment of hereditary cardiac diseases.

Analysis-ready optical underwater images of Manganese-nodule covered seafloor of the Clarion-Clipperton Zone

We provide a sequence of analysis-ready optical underwater images from the Clarion-Clipperton Zone (CCZ) of the Pacific Ocean. The images were originally recorded using a towed camera sledge that photographed a seabed covered with polymetallic manganese-nodules, at an average water depth of 4,250 meters. The original degradation in visual quality and inconsistent scale among individual raw images due to different altitude implies that they are not scientifically comparable in their original form. Here, we present analysis-ready images that have already been pre-processed to account for this degradation. We also provide accompanying metadata for each image, which includes their geographic coordinates, depth of the seafloor, absolute scale (cm/pixel), and seafloor habitat class obtained from a previous study. The provided images are thus directly usable by the marine scientific community e.g., to train machine learning models for seafloor substrate classification and megafauna detection.

Electricity supply quality and use among rural and peri-urban households and small firms in Nigeria

We present a household and enterprise energy survey dataset collected within the framework of the PeopleSuN project in Nigeria in 2021. Across three Nigerian geopolitical zones, a total of 3,599 households and 1,122 small and medium-sized enterprises were surveyed. The sample is designed to be representative of rural and peri-urban grid-electrified regions of each zone. Our surveys collect data on demographic and socioeconomic characteristics, energy access and supply quality, electrical appliance ownership and usage time, cooking solutions, energy related capabilities, and supply preferences. We encourage academic use of the data presented and suggest three avenues of further research: (1) modelling appliance ownership likelihoods, electricity consumption levels and energy service needs in un-electrified regions; (2) identifying supply-side and demand-side solutions to address high usage of diesel generators; (3) exploring broader issues of multi-dimensional energy access, access to decent living standards and climate vulnerability.

HAPLN1 potentiates peritoneal metastasis in pancreatic cancer

Pancreatic ductal adenocarcinoma (PDAC) frequently metastasizes into the peritoneum, which contributes to poor prognosis. Metastatic spreading is promoted by cancer cell plasticity, yet its regulation by the microenvironment is incompletely understood. Here, we show that the presence of hyaluronan and proteoglycan link protein-1 (HAPLN1) in the extracellular matrix enhances tumor cell plasticity and PDAC metastasis. Bioinformatic analysis showed that HAPLN1 expression is enriched in the basal PDAC subtype and associated with worse overall patient survival. In a mouse model for peritoneal carcinomatosis, HAPLN1-induced immunomodulation favors a more permissive microenvironment, which accelerates the peritoneal spread of tumor cells. Mechanistically, HAPLN1, via upregulation of tumor necrosis factor receptor 2 (TNFR2), promotes TNF-mediated upregulation of Hyaluronan (HA) production, facilitating EMT, stemness, invasion and immunomodulation. Extracellular HAPLN1 modifies cancer cells and fibroblasts, rendering them more immunomodulatory. As such, we identify HAPLN1 as a prognostic marker and as a driver for peritoneal metastasis in PDAC.

Metabolic heterogeneity and cross-feeding within isogenic yeast populations captured by DILAC

Genetically identical cells are known to differ in many physiological parameters such as growth rate and drug tolerance. Metabolic specialization is believed to be a cause of such phenotypic heterogeneity, but detection of metabolically divergent subpopulations remains technically challenging. We developed a proteomics-based technology, termed differential isotope labelling by amino acids (DILAC), that can detect producer and consumer subpopulations of a particular amino acid within an isogenic cell population by monitoring peptides with multiple occurrences of the amino acid. We reveal that young, morphologically undifferentiated yeast colonies contain subpopulations of lysine producers and consumers that emerge due to nutrient gradients. Deconvoluting their proteomes using DILAC, we find evidence for in situ cross-feeding where rapidly growing cells ferment and provide the more slowly growing, respiring cells with ethanol. Finally, by combining DILAC with fluorescence-activated cell sorting, we show that the metabolic subpopulations diverge phenotypically, as exemplified by a different tolerance to the antifungal drug amphotericin B. Overall, DILAC captures previously unnoticed metabolic heterogeneity and provides experimental evidence for the role of metabolic specialization and cross-feeding interactions as a source of phenotypic heterogeneity in isogenic cell populations.

Chromatin-accessibility estimation from single-cell ATAC-seq data with scOpen

A major drawback of single-cell ATAC-seq (scATAC-seq) is its sparsity, i.e., open chromatin regions with no reads due to loss of DNA material during the scATAC-seq protocol. Here, we propose scOpen, a computational method based on regularized non-negative matrix factorization for imputing and quantifying the open chromatin status of regulatory regions from sparse scATAC-seq experiments. We show that scOpen improves crucial downstream analysis steps of scATAC-seq data as clustering, visualization, cis-regulatory DNA interactions, and delineation of regulatory features. We demonstrate the power of scOpen to dissect regulatory changes in the development of fibrosis in the kidney. This identifies a role of Runx1 and target genes by promoting fibroblast to myofibroblast differentiation driving kidney fibrosis.

Convergent synthesis of diversified reversible network leads to liquid metal-containing conductive hydrogel adhesives

Many features of extracellular matrices, e.g., self-healing, adhesiveness, viscoelasticity, and conductivity, are associated with the intricate networks composed of many different covalent and non-covalent chemical bonds. Whereas a reductionism approach would have the limitation to fully recapitulate various biological properties with simple chemical structures, mimicking such sophisticated networks by incorporating many different functional groups in a macromolecular system is synthetically challenging. Herein, we propose a strategy of convergent synthesis of complex polymer networks to produce biomimetic electroconductive liquid metal hydrogels. Four precursors could be individually synthesized in one to two reaction steps and characterized, then assembled to form hydrogel adhesives. The convergent synthesis allows us to combine materials of different natures to generate matrices with high adhesive strength, enhanced electroconductivity, good cytocompatibility in vitro and high biocompatibility in vivo. The reversible networks exhibit self-healing and shear-thinning properties, thus allowing for 3D printing and minimally invasive injection for in vivo experiments.

Next-generation hypomethylating agent SGI-110 primes acute myeloid leukemia cells to IAP antagonist by activating extrinsic and intrinsic apoptosis pathways

Therapeutic efficacy of first-generation hypomethylating agents (HMAs) is limited in elderly acute myeloid leukemia (AML) patients. Therefore, combination strategies with targeted therapies are urgently needed. Here, we discover that priming with SGI-110 (guadecitabine), a next-generation HMA, sensitizes AML cells to ASTX660, a novel antagonist of cellular inhibitor of apoptosis protein 1 and 2 (cIAP1/2) and X-linked IAP (XIAP). Importantly, SGI-110 and ASTX660 synergistically induced cell death in a panel of AML cell lines as well as in primary AML samples while largely sparing normal CD34+ human progenitor cells, underlining the translational relevance of this combination. Unbiased transcriptome analysis revealed that SGI-110 alone or in combination with ASTX660 upregulated the expression of key regulators of both extrinsic and intrinsic apoptosis signaling pathways such as TNFRSF10B (DR5), FAS, and BAX. Individual knockdown of the death receptors TNFR1, DR5, and FAS significantly reduced SGI-110/ASTX660-mediated cell death, whereas blocking antibodies for tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) or FAS ligand (FASLG) failed to provide protection. Also, TNFα-blocking antibody Enbrel had little protective effect on SGI-110/ASTX660-induced cell death. Further, SGI-110 and ASTX660 acted in concert to promote cleavage of caspase-8 and BID, thereby providing a link between extrinsic and intrinsic apoptotic pathways. Consistently, sequential treatment with SGI-110 and ASTX660-triggered loss of mitochondrial membrane potential (MMP) and BAX activation which contributes to cell death, as BAX silencing significantly protected from SGI-110/ASTX660-mediated apoptosis. Together, these events culminated in the activation of caspases-3/-7, nuclear fragmentation, and cell death. In conclusion, SGI-110 and ASTX660 cooperatively induced apoptosis in AML cells by engaging extrinsic and intrinsic apoptosis pathways, highlighting the therapeutic potential of this combination for AML.

Associations between self-reported sleep characteristics and incident mild cognitive impairment: The Heinz Nixdorf Recall Cohort Study

Associations of sleep characteristics with mild cognitive impairment (MCI) have been examined in cross-sectional, but rarely in longitudinal studies. Incident MCI and sleep characteristics were assessed in 1,890 participants of the first and second follow-up of the Heinz Nixdorf Recall study, a population-based cohort study in Germany (age at first follow-up 50-80 years, mean follow-up 5.2 years). MCI was assessed with extensive cognitive tests. Sleep questionnaires including PSQI (Pittsburgh Sleep Quality Index) were used to assess sleep quality, sleep disturbances, time asleep, and time in bed. Relative risks (RR) of developing MCI when exposed to sleep characteristics were assessed in regression models adjusted for sociodemographic and cardiovascular risk factors. Poor sleep quality (PSQI > 5) (RR = 1.43, 95% CI: 1.12-1.82, fully adjusted, reference: PSQI ≤ 5) and difficulties initiating sleep (almost nightly versus never) (RR = 1.40, 0.94-2.08) were associated with incident MCI. For time in bed, the risk of MCI was increased for ≤ 5 hours (RR = 2.86, 1.24─6.60, reference:7 to <8 hours). In this longitudinal study with older participants, MCI risk was increased in persons with poor sleep quality, difficulties initiating sleep, and short time in bed.

Probing and predicting ganglion cell responses to smooth electrical stimulation in healthy and blind mouse retina

Retinal implants are used to replace lost photoreceptors in blind patients suffering from retinopathies such as retinitis pigmentosa. Patients wearing implants regain some rudimentary visual function. However, it is severely limited compared to normal vision because non-physiological stimulation strategies fail to selectively activate different retinal pathways at sufficient spatial and temporal resolution. The development of improved stimulation strategies is rendered difficult by the large space of potential stimuli. Here we systematically explore a subspace of potential stimuli by electrically stimulating healthy and blind mouse retina in epiretinal configuration using smooth Gaussian white noise delivered by a high-density CMOS-based microelectrode array. We identify linear filters of retinal ganglion cells (RGCs) by fitting a linear-nonlinear-Poisson (LNP) model. Our stimulus evokes spatially and temporally confined spiking responses in RGC which are accurately predicted by the LNP model. Furthermore, we find diverse shapes of linear filters in the linear stage of the model, suggesting diverse preferred electrical stimuli of RGCs. The linear filter base identified by our approach could provide a starting point of a model-guided search for improved stimuli for retinal prosthetics.

Attention strengthens across-trial pre-stimulus phase coherence in visual cortex, enhancing stimulus processing

Attention selectively routes the most behaviorally relevant information from the stream of sensory inputs through the hierarchy of cortical areas. Previous studies have shown that visual attention depends on the phase of oscillatory brain activities. These studies mainly focused on the stimulus presentation period, rather than the pre-stimulus period. Here, we hypothesize that selective attention controls the phase of oscillatory neural activities to efficiently process relevant information. We document an attentional modulation of pre-stimulus inter-trial phase coherence (a measure of deviation between instantaneous phases of trials) of low frequency local field potentials (LFP) in visual area MT of macaque monkeys. Our data reveal that phase coherence increases following a spatial cue deploying attention towards the receptive field of the recorded neural population. We further show that the attentional enhancement of phase coherence is positively correlated with the modulation of the stimulus-induced firing rate, and importantly, a higher phase coherence is associated with a faster behavioral response. These results suggest a functional utilization of intrinsic neural oscillatory activities for an enhanced processing of upcoming stimuli.

Vitamin C alleviates acute enterocolitis in Campylobacter jejuni infected mice

Human foodborne infections with the zoonotic pathogen Campylobacter jejuni are on the rise and constitute a significant socioeconomic burden worldwide. The health-beneficial, particularly anti-inflammatory effects of vitamin C (ascorbate) are well known. In our preclinical intervention study, we assessed potential anti-pathogenic and immunomodulatory effects of ascorbate in C. jejuni-infected secondary abiotic IL-10-/- mice developing acute campylobacteriosis similar to humans. Starting 4 days prior peroral C. jejuni-infection, mice received synthetic ascorbate via the drinking water until the end of the experiment. At day 6 post-infection, ascorbate-treated mice harbored slightly lower colonic pathogen loads and suffered from less severe C. jejuni-induced enterocolitis as compared to placebo control animals. Ascorbate treatment did not only alleviate macroscopic sequelae of infection, but also dampened apoptotic and inflammatory immune cell responses in the intestines that were accompanied by less pronounced pro-inflammatory cytokine secretion. Remarkably, the anti-inflammatory effects of ascorbate pretreatment in C. jejuni-infected mice were not restricted to the intestinal tract but could also be observed in extra-intestinal compartments including liver, kidneys and lungs. In conclusion, due to the potent anti-inflammatory effects observed in the clinical murine C. jejuni-infection model, ascorbate constitutes a promising novel option for prophylaxis and treatment of acute campylobacteriosis.

Murine fecal microbiota transplantation lowers gastrointestinal pathogen loads and dampens pro-inflammatory immune responses in Campylobacter jejuni infected secondary abiotic mice

Conventional mice are protected from Campylobacter jejuni infection by the murine host-specific gut microbiota composition. We here addressed whether peroral fecal microbiota transplantation (FMT) might be an antibiotics-independent option to lower even high gastrointestinal C. jejuni loads in the infected vertebrate host. To address this, secondary abiotic mice were generated by broad-spectrum antibiotic treatment and perorally infected with C. jejuni by gavage. One week later, mice were stably colonized with more than 109 C. jejuni and subjected to peroral FMT from murine donors on three consecutive days. Two weeks post-intervention, gastrointestinal C. jejuni loads were up to 7.5 orders of magnitude lower following murine FMT versus mock challenge. Remarkably, FMT reversed C. jejuni induced colonic epithelial apoptosis, but enhanced proliferative and regenerative responses in the colon thereby counteracting pathogenic cell damage. Furthermore, FMT dampened both, innate and adaptive immune cell responses in the large intestines upon C. jejuni infection that were accompanied by less C. jejuni-induced colonic nitric oxide secretion. Our study provides strong evidence that novel probiotic formulations developed as alternative option to FMT in severe intestinal inflammatory morbidities including Clostridoides difficile infection might be effective to treat campylobacteriosis and lower pathogen loads in colonized vertebrates including farm animals.

Maintenance of cell type-specific connectivity and circuit function requires Tao kinase

Sensory circuits are typically established during early development, yet how circuit specificity and function are maintained during organismal growth has not been elucidated. To gain insight we quantitatively investigated synaptic growth and connectivity in the Drosophila nociceptive network during larval development. We show that connectivity between primary nociceptors and their downstream neurons scales with animal size. We further identified the conserved Ste20-like kinase Tao as a negative regulator of synaptic growth required for maintenance of circuit specificity and connectivity. Loss of Tao kinase resulted in exuberant postsynaptic specializations and aberrant connectivity during larval growth. Using functional imaging and behavioral analysis we show that loss of Tao-induced ectopic synapses with inappropriate partner neurons are functional and alter behavioral responses in a connection-specific manner. Our data show that fine-tuning of synaptic growth by Tao kinase is required for maintaining specificity and behavioral output of the neuronal network during animal growth.