Leveraging Cell Painting Images to Expand the Applicability Domain and Actively Improve Deep Learning Quantitative Structure-Activity Relationship Models.

The search for chemical hit material is a lengthy and increasingly expensive drug discovery process. To improve it, ligand-based quantitative structure-activity relationship models have been broadly applied to optimize primary and secondary compound properties. Although these models can be deployed as early as the stage of molecule design, they have a limited applicability domain─if the structures of interest differ substantially from the chemical space on which the model was trained, a reliable prediction will not be possible. Image-informed ligand-based models partly solve this shortcoming by focusing on the phenotype of a cell caused by small molecules, rather than on their structure. While this enables chemical diversity expansion, it limits the application to compounds physically available and imaged. Here, we employ an active learning approach to capitalize on both of these methods' strengths and boost the model performance of a mitochondrial toxicity assay (Glu/Gal). Specifically, we used a phenotypic Cell Painting screen to build a chemistry-independent model and adopted the results as the main factor in selecting compounds for experimental testing. With the additional Glu/Gal annotation for selected compounds we were able to dramatically improve the chemistry-informed ligand-based model with respect to the increased recognition of compounds from a 10% broader chemical space.

Sensitive in vivo cell detection using size-optimized superparamagnetic nanoparticles.

Magnetic nanoparticle (MNP) enabled cell visualization with magnetic resonance imaging (MRI) is currently an intensively studied area of research. In the present study, we have synthesized polyethylene glycolated (PEG) MNPs and validated their suitability as MR cell labeling agents in in vitro and in vivo experiments. The labeling of therapeutic potent mesenchymal stem cells (MSCs) with small core and large core MNPs was evaluated. Both MNPs were, in combination with a transfection agent, stably internalized into the MSCs and didn't show an effect on cell metabolism. The labeled cells showed high contrast in MRI phantom studies. For quantification purposes, the MRI contrast generating properties of cells labeled with small core MNPs were compared with large core MNPs and with the commercial contrast agent Endorem. MSCs labeled with the large core MNPs showed the highest contrast generating properties in in vitro phantom studies and in in vivo intracranial stereotactic injection experiments, confirming the size-relaxivity relationship in biological systems. Finally, the distribution of MSCs pre-labeled with large core PEGylated MNPs was visualized non-invasively with MRI in a glioma model.

Substrate topography determines neuronal polarization and growth in vitro.

The establishment of neuronal connectivity depends on the correct initial polarization of the young neurons. In vivo, developing neurons sense a multitude of inputs and a great number of molecules are described that affect their outgrowth. In vitro, many studies have shown the possibility to influence neuronal morphology and growth by biophysical, i.e. topographic, signaling. In this work we have taken this approach one step further and investigated the impact of substrate topography in the very early differentiation stages of developing neurons, i.e. when the cell is still at the round stage and when the first neurite is forming. For this purpose we fabricated micron sized pillar structures with highly reproducible feature sizes, and analyzed neurons on the interface of flat and topographic surfaces. We found that topographic signaling was able to attract the polarization markers of mouse embryonic neurons -N-cadherin, Golgi-centrosome complex and the first bud were oriented towards topographic stimuli. Consecutively, the axon was also preferentially extending along the pillars. These events seemed to occur regardless of pillar dimensions in the range we examined. However, we found differences in neurite length that depended on pillar dimensions. This study is one of the first to describe in detail the very early response of hippocampal neurons to topographic stimuli.

Superimposed topographic and chemical cues synergistically guide neurite outgrowth.

Guidance of neuronal extensions is a complex process essential for linking neurons into complex functional networks underlying the workings of the neural system. Decades of research have suggested the ability of neuronal growth cones to integrate multiple types of cues during the extension process, but also have raised numerous still unanswered questions about synergy or antagonism between the superimposed chemical and mechanical signaling inputs. In this study, using a novel microfabricated analysis platform, we investigate the response of primary mouse embryonic hippocampal neurons to superimposed topographic and soluble chemical cues. We find that an optimal spatial frequency of topographic cues exists, maximizing the precision of the neurite extension. This optimal frequency can help the extending neurites navigate a topographically complex environment, providing pronounced directional selectivity. We also demonstrate that this cue can synergistically enhance attractive and suppress repulsive guidance by the bi-functional soluble cue Netrin-1, and eliminate the repulsive guidance by a chemorepellent Semaphorin3A (Sema3A). These results suggest that topographic cues can provide optimal periodic input into the guidance signaling processes involved in growth cone chemoattraction and can synergistically interact with chemical gradients of soluble guidance cues, shedding light on complex events accompanying the development of the functional nervous system.

Micro-sized syringes for single-cell fluidic access integrated on a micro-electrode array CMOS chip.

Very-large scale integration and micro-machining have enabled the development of novel platforms for advanced and automated examination of cells and tissues in vitro. In this paper, we present a CMOS chip designed in a commercial 0.18 µm technology with integrated micro-syringes combined with micro-nail shaped electrodes and readout electronics. The micro-syringes could be individually addressed by a through-wafer micro-fluidic channel with an inner diameter of 1 µm. We demonstrated the functionality of the micro-fluidic access by diffusion of fluorescent species through the channels. Further, hippocampal neurons were cultured on top of an array of micro-syringes, and focused ion beam-scanning electron microscopy cross-sections revealed protrusion of the cells inside the channels, creating a strong interface between the membrane and the chip surface. This principle demonstrates a first step towards a novel type of automated in vitro platforms, allowing local delivery of substances to cells or advanced planar patch clamping.