Biology subjects

Petzold, G.

Publications and source records attributed to Petzold, G..

5 recordsLinked to original sources

bioRxiv2026

Automated quantification of white matter hyperintensity confluence: A measure of spatial organisation beyond volume and visual rating scales

White matter hyperintensities (WMH) are a highly prevalent finding on FLAIR MRI scans and a prominent feature of white matter pathology across cerebrovascular and neurodegenerative diseases. Currently, WMH are assessed with visual rating scales such as the Fazekas scale or with their volume, as calculated from automatic or manual segmentations. Both methods have limitations: Visual rating scales are rater-dependent and coarse, while WMH volume does not take the confluence of lesions into account and thus disregards their spatial organisation. As an alternative, here we propose a novel automated method for quantifying the confluence of white matter hyperintensities on a continuous standardised scale between 0 and 1. The metric is based on WMH segmentations from routine MRI and quantifies the extent to which individual WMH merge into coherent lesions, independently of total lesion volume. We apply the method to QMIN-MC, a large UK memory clinic cohort, and show associations of the confluence metric with age, cognitive performance across domains, and Fazekas ratings. Participants with vascular and mixed dementia showed higher confluence than other diagnostic groups, whereas cognitively unimpaired participants showed lower confluence. However, confluence did not explain additional cognitive variance after accounting for log-transformed WMH volume. Findings were validated in DELCODE, an independent cohort of individuals with neurodegenerative disorders, replicating our original results. In this validation cohort, periventricular WMH confluence remained associated with cognition after adjustment for WMH volume. These findings introduce WMH confluence as a reproducible, automated, and fine-grained measure of lesion spatial organisation. It provides complementary information about morphological WMH severity beyond volume and is an alternative to visual rating scales. Although related to WMH volume in memory-clinic populations, confluence captures clinically interpretable information and may complement existing WMH measures for improved lesion characterisation in studies of white matter disease, ageing, and cognitive impairment.

neuroscience↗

bioRxiv2025

Molecular surface mimicry enables CRBN to target G3BP2 for degradation

Molecular glue degraders (MGDs) are small molecule compounds that repurpose the ubiquitin-proteasome system to induce degradation of challenging therapeutic targets. Clinically effective MGDs bind cereblon (CRBN), a substrate receptor of the Cullin-4/RING E3 ubiquitin ligase (CRL4CRBN), and impose a gain-of-function activity to recruit, ubiquitinate and degrade so-called neosubstrate proteins. Known neosubstrates bind the CRBN/MGD neosurface on the CRBN CULT domain through a structural G-loop recognition motif that mimics contacts of a natural CRBN degron. Here we report the binding mode of G3BP2, a CRBN neosubstrate that bypasses the G-loop requirement by engaging an unconventional binding site on the CRBN LON domain. The MGD-induced ternary complex interface does not resemble known protein-protein interactions (PPI) with CRBN. Instead, CRBN mimics an endogenous binding partner of G3BP2 and repurposes a preexisting PPI hotspot on the target protein. Our findings provide a novel generalizable concept for the rationalization of unconventional neosubstrate binding modes on CRBN, demonstrate unprecedented potential for the reprogrammability of this substrate receptor by MGDs, and offer opportunities for rational expansion of the target repertoire accessible to this modality.

molecular biology↗

bioRxiv2024

Mining the CRBN Target Space Redefines Rules for Molecular Glue-induced Neosubstrate Recognition

The CRL4CRBN ubiquitin ligase is leveraged by molecular glue degraders, small molecules that reprogram ligase specificity to induce degradation of clinically relevant neosubstrate proteins. Known CRBN neosubstrates share a generalizable {beta}-hairpin G-loop recognition motif, yet systematic exploration of the CRBN target landscape is still pending. Through computational mining of the human proteome using structure-based approaches, we predict over 1,400 CRBN-compatible {beta}-hairpin G-loop proteins across diverse target classes, identify novel mechanisms of neosubstrate recognition through structurally differentiated helical motifs and molecular surface mimicry, and validate 22 representative neosubstrates with clinical implications. This work broadens the CRBN target space, redefines rules for neosubstrate recognition and establishes a platform for the elimination of challenging drug targets by repurposing CRL4CRBN through next-generation molecular glue degraders.

biochemistry↗

bioRxiv2024

Three-photon in vivo imaging of neurons and glia in the medial prefrontal cortex with sub-cellular resolution

The medial prefrontal cortex (mPFC) is important for higher cognitive functions, including working memory, decision making, and emotional control. In vivo recordings of neuronal activity in the mPFC have been achieved via invasive electrical and optical approaches. Here we apply low invasive three-photon in vivo imaging in the mPFC of the mouse at unprecedented depth. Specifically, we measure neuronal and astrocytic Ca2+-transient parameters in awake head-fixed mice up to a depth of 1700 {micro}m. Furthermore, we longitudinally record dendritic spine density (0.41 {+/-}0.07 {micro}m-1) deeper than 1 mm for a week. Using 1650 nm wavelength to excite red fluorescent microglia, we quantify their processes motility (58.9 {+/-}2% turnover rate) at previously unreachable depths (1100 {micro}m). We establish three-photon imaging of the mPFC enabling neuronal and glial recordings with subcellular resolution that will pave the way for novel discoveries in this brain region.

neuroscience↗

bioRxiv2020

Insights into calcium signaling and gene expression in astrocytes uncovered with 129S4 Slc1a3-2A-CreERT2 knock-in mice

AO_SCPLOWBSTRACTC_SCPLOWGenetic variation is a primary determinant of phenotypic diversity within populations. In laboratory mice, genetic variation has often been regarded as a serious experimental confounder, and thus minimized through inbreeding. However, generalizations of results obtained with inbred strains need to be made with caution. Effects of genetic background on traits need to be controlled, especially when working with complex phenotypes and disease models. Here we compared behavioral parameters of C57Bl/6 - the mouse strain most widely used for biomedical research - with those of 129S4. Our data demonstrate high within-strain and intra-litter behavioral hyperactivity in C57Bl/6. In contrast, 129S4 had relatively consistent activity levels throughout life. This consistency would be advantageous for studying neurodegeneration and aging, when mice need to be analyzed for long periods. However, the majority of mouse models and transgenic tools are on a C57Bl/6 background. We recently established six popular Cre driver lines and two Cre effector lines in 129S4. To augment this collection, we genetically engineered a Cre mouse line to study astrocytes directly in 129S4, which we describe here. For functional validation, it was crossed with two Cre effector lines, each in a different genomic locus, and showed in both cases that it was functional and astrocyte-specific. Calcium currents studied with gCaMP5g-tdTomato were more heterogenous, lasted longer and had a higher amplitude in cortical compared to hippocampal astrocytes. Translatomes studied with RiboTag revealed that some genes thought to mark neurons are also expressed in astrocytes, that genes linked to a single neurodegenerative disease have highly divergent expression patterns, and that ribosome proteins are non-uniformly expressed across brain regions and cell types.

neuroscience↗