Explore the vast range of titles available.

Text-to-image models are increasingly popular and impactful, yet concerns regarding their safety and fairness remain. This study investigates the ability of ten popular Stable Diffusion models to generate harmful images, including sexual, violent, and personally sensitive material. We demonstrate that these models respond to harmful prompts by generating inappropriate content, which frequently displays troubling biases, such as the disproportionate portrayal of Black individuals in violent contexts. Our findings demonstrate a complete lack of any refusal behavior or safety measures in the models observed. We emphasize the importance of addressing this issue as image generation technologies continue to become more accessible and incorporated into everyday applications.

The origin of nonlinear responses in cells has been suggested to be crucial for various cell functions including the propagation of the nervous impulse. In physics, nonlinear behavior often originates from phase transitions. Evidence for such transitions on the single-cell level, however, has so far not been provided, leaving the field unattended by the biological community. Here, we demonstrate that single cells of a human neuronal cell line display all optical features of a sharp, highly nonlinear phase transition within their membrane. The transition is reversible and does not originate from protein denaturation. Triggered by temperature and modified by pH here, a thermodynamic approach strongly suggests that similar nonlinear state changes can be induced by other variables such as calcium or mechanical stress. At least in lipid membranes, such state changes are accompanied by significant changes in permeability, enzyme activity, elastic, and electrical properties.

An action potential is typically described as a purely electrical change that propagates along the membrane of excitable cells. However, recent experiments have demonstrated that non-linear acoustic pulses that propagate along lipid interfaces and traverse the melting transition, share many similar properties with action potentials. Despite the striking experimental similarities, a comprehensive theoretical study of acoustic pulses in lipid systems is still lacking. Here we demonstrate that an idealized description of an interface near phase transition captures many properties of acoustic pulses in lipid monolayers, as well as action potentials in living cells. The possibility that action potentials may better be described as acoustic pulses in soft interfaces near phase transition is illustrated by the following similar properties: correspondence of time and velocity scales, qualitative pulse shape, sigmoidal response to stimulation amplitude (an ‘all-or-none’ behavior), appearance in multiple observables (particularly, an adiabatic change of temperature), excitation by many types of stimulations, as well as annihilation upon collision. An implication of this work is that crucial functional information of the cell may be overlooked by focusing only on electrical measurements.

Our objective was to update the EULAR recommendations for the management of systemic lupus erythematosus (SLE), based on emerging new evidence. We performed a systematic literature review (01/2007–12/2017), followed by modified Delphi method, to form questions, elicit expert opinions and reach consensus. Treatment in SLE aims at remission or low disease activity and prevention of flares. Hydroxychloroquine is recommended in all patients with lupus, at a dose not exceeding 5 mg/kg real body weight. During chronic maintenance treatment, glucocorticoids (GC) should be minimised to less than 7.5 mg/day (prednisone equivalent) and, when possible, withdrawn. Appropriate initiation of immunomodulatory agents (methotrexate, azathioprine, mycophenolate) can expedite the tapering/discontinuation of GC. In persistently active or flaring extrarenal disease, add-on belimumab should be considered; rituximab (RTX) may be considered in organ-threatening, refractory disease. Updated specific recommendations are also provided for cutaneous, neuropsychiatric, haematological and renal disease. Patients with SLE should be assessed for their antiphospholipid antibody status, infectious and cardiovascular diseases risk profile and preventative strategies be tailored accordingly. The updated recommendations provide physicians and patients with updated consensus guidance on the management of SLE, combining evidence-base and expert-opinion.

We present DeepVesselNet, an architecture tailored to the challenges faced when extracting vessel trees and networks and corresponding features in 3-D angiographic volumes using deep learning. We discuss the problems of low execution speed and high memory requirements associated with full 3-D networks, high-class imbalance arising from the low percentage (<3%) of vessel voxels, and unavailability of accurately annotated 3-D training data—and offer solutions as the building blocks of DeepVesselNet. First, we formulate 2-D orthogonal cross-hair filters which make use of 3-D context information at a reduced computational burden. Second, we introduce a class balancing cross-entropy loss function with false-positive rate correction to handle the high-class imbalance and high false positive rate problems associated with existing loss functions. Finally, we generate a synthetic dataset using a computational angiogenesis model capable of simulating vascular tree growth under physiological constraints on local network structure and topology and use these data for transfer learning. We demonstrate the performance on a range of angiographic volumes at different spatial scales including clinical MRA data of the human brain, as well as CTA microscopy scans of the rat brain. Our results show that cross-hair filters achieve over 23% improvement in speed, lower memory footprint, lower network complexity which prevents overfitting and comparable accuracy that does not differ from full 3-D filters. Our class balancing metric is crucial for training the network, and transfer learning with synthetic data is an efficient, robust, and very generalizable approach leading to a network that excels in a variety of angiography segmentation tasks. We observe that sub-sampling and max pooling layers may lead to a drop in performance in tasks that involve voxel-sized structures. To this end, the DeepVesselNet architecture does not use any form of sub-sampling layer and works well for vessel segmentation, centerline prediction, and bifurcation detection. We make our synthetic training data publicly available, fostering future research, and serving as one of the first public datasets for brain vessel tree segmentation and analysis.

Local changes in pH are known to significantly alter the state and activity of proteins and enzymes. pH variations induced by pulses propagating along soft interfaces (e.g. membranes) would therefore constitute an important pillar towards a physical mechanism of biological signaling. Here we investigate the pH-induced physical perturbation of a lipid interface and the physicochemical nature of the subsequent acoustic propagation. Pulses are stimulated by local acidification and propagate – in analogy to sound – at velocities controlled by the interface’s compressibility. With transient local pH changes of 0.6 directly observed at the interface and velocities up to 1.4 m/s this represents hitherto the fastest protonic communication observed. Furthermore simultaneously propagating mechanical and electrical changes in the lipid interface are detected, exposing the thermodynamic nature of these pulses. Finally, these pulses are excitable only beyond a threshold for protonation, determined by the pK a of the lipid head groups. This protonation-transition plus the existence of an enzymatic pH-optimum offer a physical basis for intra- and intercellular signaling via sound waves at interfaces, where not molecular structure and mechano-enyzmatic couplings, but interface thermodynamics and thermodynamic transitions are the origin of the observations.

Molecular chaperones contribute to the maintenance of cellular protein homoeostasis through assisting de novo protein folding and preventing amyloid formation. Chaperones of the Hsp70 family can further disaggregate otherwise irreversible aggregate species such as α-synuclein fibrils, which accumulate in Parkinson’s disease. However, the mechanisms and kinetics of this key functionality are only partially understood. Here, we combine microfluidic measurements with chemical kinetics to study α-synuclein disaggregation. We show that Hsc70 together with its co-chaperones DnaJB1 and Apg2 can completely reverse α-synuclein aggregation back to its soluble monomeric state. This reaction proceeds through first-order kinetics where monomer units are removed directly from the fibril ends with little contribution from intermediate fibril fragmentation steps. These findings extend our mechanistic understanding of the role of chaperones in the suppression of amyloid proliferation and in aggregate clearance, and inform on possibilities and limitations of this strategy in the development of therapeutics against synucleinopathies. Molecular chaperones from the Hsp70 family can break up protein aggregates, including amyloids. Here, the authors utilize microfluidic diffusional sizing to assess the mechanism of α-synuclein (αS) disaggregation by the Hsc70–DnaJB1–Apg2 system, and show that single αS molecules are removed directly from the fibril ends.

Fluorescent dyes are vital for studying static and dynamic patterns and pattern formation in cell biology. Emission properties of the dyes incorporated in a biological interface are known to be sensitive to their local environment. We report that the fluorescence intensity of dye molecules embedded in lipid interfaces is indeed a thermodynamic observable of the system. Opto-mechanical coupling of lipid-dye system was measured as a function of the thermodynamic state of the interface. The corresponding state diagrams quantify the thermodynamic coupling between intensity I and lateral pressure π. We further demonstrate that the coupling is conserved upon varying the temperature T. Notably, the observed opto-mechanical coupling is not limited to equilibrium conditions, but also holds for propagating pressure pulses. The non-equilibrium data show, that fluorescence is especially sensitive to dynamic changes in state such as the LE-LC phase transition. We conclude that variations in the thermodynamic state (here π and T, in general pH, membrane potential V, etc also) of lipid membranes are capable of controlling fluorescence intensity. Therefore, interfacial thermodynamic state diagrams of I should be obtained for a proper interpretation of intensity data.

Reactive astrocytes evolve after brain injury, inflammatory and degenerative diseases, whereby they undergo transcriptomic re-programming. In malignant brain tumors, their function and crosstalk to other components of the environment is poorly understood. Here we report a distinct transcriptional phenotype of reactive astrocytes from glioblastoma linked to JAK/STAT pathway activation. Subsequently, we investigate the origin of astrocytic transformation by a microglia loss-of-function model in a human organotypic slice model with injected tumor cells. RNA-seq based gene expression analysis of astrocytes reveals a distinct astrocytic phenotype caused by the coexistence of microglia and astrocytes in the tumor environment, which leads to a large release of anti-inflammatory cytokines such as TGFβ, IL10 and G-CSF. Inhibition of the JAK/STAT pathway shifts the balance of pro- and anti-inflammatory cytokines towards a pro-inflammatory environment. The complex interaction of astrocytes and microglia cells promotes an immunosuppressive environment, suggesting that tumor-associated astrocytes contribute to anti-inflammatory responses. Astrocytes play important roles in neuroinflammatory diseases. Here the authors characterize human glioblastoma-associated astrocytes by gene expression and demonstrate their immunosuppressive role promoted by interactions with tumor and microglia cells in an organotypic model.

The complex kinetics of disease-related amyloid aggregation of proteins such as α-Synuclein (α-Syn) in Parkinson’s disease and Aβ42 in Alzheimer’s disease include primary nucleation, amyloid fibril elongation and secondary nucleation. The latter can be a key accelerator of the aggregation process. It has been demonstrated that the chaperone domain BRICHOS can interfere with the secondary nucleation process of Aβ42. Here, we explore the mechanism of secondary nucleation inhibition of the BRICHOS domain of the lung surfactant protein (proSP-C) against α-Syn aggregation and amyloid formation. We determine the 3D NMR structure of an inactive trimer of proSP-C BRICHOS and its active monomer using a designed mutant. Furthermore, the interaction between the proSP-C BRICHOS chaperone and a substrate peptide has been studied. NMR-based interaction studies of proSP-C BRICHOS with α-Syn fibrils show that proSP-C BRICHOS binds to the C-terminal flexible fuzzy coat of the fibrils, which is the secondary nucleation site on the fibrils. Super-resolution fluorescence microscopy demonstrates that proSP-C BRICHOS runs along the fibrillar axis diffusion-dependently sweeping off monomeric α-Syn from the fibrils. The observed mechanism explains how a weakly binding chaperone can inhibit the α-Syn secondary nucleation pathway via avidity where a single proSP-C BRICHOS molecule is sufficient against up to ~7-40 α-Syn molecules embedded within the fibrils. Here, the authors probe the mechanism of secondary nucleation inhibition of the BRICHOS domain of the lung surfactant protein (proSP-C) against α-Syn aggregation and amyloid formation using NMR and super-resolution fluorescence microscopy.