Marlene Stuempflen

. Cerebral organoids enable the study of neurodevelopmental disorders in a human context. We have developed the CRISPR-human organoids-single-cell RNA sequencing (CHOOSE) system, which uses verified pairs of guide RNAs, inducible CRISPR-Cas9-based genetic disruption and single-cell transcriptomics for pooled loss-of-function screening in mosaic organoids. Here we show that perturbation of 36 high-risk autism spectrum disorder genes related to transcriptional regulation uncovers their effects on cell fate determination. We find that dorsal intermediate progenitors, ventral progenitors and upper-layer excitatory neurons are among the most vulnerable cell types. We construct a developmental gene regulatory network of cerebral organoids from single-cell transcriptomes and chromatin modalities and identify autism spectrum disorder-associated and perturbation-enriched regulatory modules. Perturbing members of the BRG1/BRM-associated factor (BAF) chromatin remodelling complex leads to enrichment of ventral telencephalon progenitors. Specifically, mutating the BAF subunit ARID1B affects the fate transition of progenitors to oligodendrocyte and interneuron precursor cells, a phenotype that we confirmed in patient-specific induced pluripotent stem cell-derived organoids. Our study paves the way for high-throughput phenotypic characterization of disease susceptibility genes in organoid models with cell state, molecular pathway and gene regulatory network readouts.

Mutations in ARID1B, a member of the mSWI/SNF complex, cause severe neurodevelopmental phenotypes with elusive mechanisms in humans. The most common structural abnormality in the brain of ARID1B patients is agenesis of the corpus callosum (ACC), characterized by the absence of an interhemispheric white matter tract that connects distant cortical regions. Here, we find that neurons expressing SATB2, a determinant of callosal projection neuron (CPN) identity, show impaired maturation in ARID1B+/− neural organoids. Molecularly, a reduction in chromatin accessibility of genomic regions targeted by TCF-like, NFI-like, and ARID-like transcription factors drives the differential expression of genes required for corpus callosum (CC) development. Through an in vitro model of the CC tract, we demonstrate that this transcriptional dysregulation impairs the formation of long-range axonal projections, causing structural underconnectivity. Our study uncovers new functions of the mSWI/SNF during human corticogenesis, identifying cell-autonomous axonogenesis defects in SATB2+ neurons as a cause of ACC in ARID1B patients.

Deep learning models for medical image segmentation can fail unexpectedly and spectacularly for pathological cases and images acquired at different centers than training images, with labeling errors that violate expert knowledge. Such errors undermine the trustworthiness of deep learning models for medical image segmentation. Mechanisms for detecting and correcting such failures are essential for safely translating this technology into clinics and are likely to be a requirement of future regulations on artificial intelligence (AI). In this work, we propose a trustworthy AI theoretical framework and a practical system that can augment any backbone AI system using a fallback method and a fail-safe mechanism based on Dempster-Shafer theory. Our approach relies on an actionable definition of trustworthy AI. Our method automatically discards the voxel-level labeling predicted by the backbone AI that violate expert knowledge and relies on a fallback for those voxels. We demonstrate the effectiveness of the proposed trustworthy AI approach on the largest reported annotated dataset of fetal MRI consisting of 540 manually annotated fetal brain 3D T2w MRIs from 13 centers. Our trustworthy AI method improves the robustness of four backbone AI models for fetal brain MRIs acquired across various centers and for fetuses with various brain abnormalities.

Colorectal cancer is the second most common cancer diagnosed in women and third most common in men. Laparoscopic resection has become the standard surgical technique worldwide given its notable benefits, mainly the shorter length of stay and less postoperative pain. The aim of this systematic review was to evaluate the current literature on postoperative pain management following laparoscopic colorectal surgery and update previous procedure-specific pain management recommendations. The primary outcomes were postoperative pain scores and opioid requirements. We also considered study quality, clinical relevance of trial design, and a comprehensive risk-benefit assessment of the analgesic intervention. We performed a literature search to identify randomised controlled studies (RCTs) published before January 2022. Seventy-two studies were included in the present analysis. Through the established PROSPECT process, we recommend basic analgesia (paracetamol for rectal surgery, and paracetamol with either a nonsteroidal anti-inflammatory drug or cyclo-oxygenase-2-specific inhibitor for colonic surgery) and wound infiltration as first-line interventions. No consensus could be achieved either for the use of intrathecal morphine or intravenous lidocaine; no recommendation can be made for these interventions. However, intravenous lidocaine may be considered when basic analgesia cannot be provided.

There are known complications for fetuses after infection with SARS-CoV-2 during pregnancy. However, previous studies of SARS-CoV-2 in pregnancy have largely been limited to histopathologic studies of placentas and prenatal studies on the effects of different SARS-CoV-2 variants are scarce to date. To examine the effects of SARS-CoV-2 variants on the placenta and fetus, we investigated fetal and extra-fetal structures using prenatal MRI. For this prospective case–control study, two obstetric centers consecutively referred pregnant women for prenatal MRI after confirmed SARS-CoV-2 infection. Thirty-eight prenatal MRI examinations were included after confirmed infection with SARS-CoV-2 and matched 1:1 with 38 control cases with respect to sex, MRI field strength, and gestational age (average deviation 1.76 ± 1.65, median 1.5 days). Where available, the pathohistological examination and vaccination status of the placenta was included in the analysis. In prenatal MRI, the shape and thickness of the placenta, possible lobulation, and vascular lesions were quantified. Fetuses were scanned for organ or brain abnormalities. Of the 38 included cases after SARS-CoV-2 infection, 20/38 (52.6%) were infected with pre-Omicron variants and 18/38 (47.4%) with Omicron. Prenatal MRIs were performed on an average of 83 days (±42.9, median 80) days after the first positive PCR test. Both pre-Omicron (P = .008) and Omicron (P = .016) groups showed abnormalities in form of a globular placenta compared to control cases. In addition, placentas in the pre-Omicron group were significantly thickened (6.35, 95% CI .02–12.65, P = .048), and showed significantly more frequent lobules (P = .046), and hemorrhages (P = .002). Fetal growth restriction (FGR) was observed in 25% (n = 5/20, P = .017) in the pre-Omicron group. SARS-CoV-2 infections in pregnancy can lead to placental lesions based on vascular events, which can be well visualized on prenatal MRI. Pre-Omicron variants cause greater damage than Omicron sub-lineages in this regard. Vienna Science and Technology Fund.