Astrocyte Subtype Marked by CD51 Drives Synapse Formation and Stroke Recovery in Mice

A single surface protein, CD51, is enough to tell apart two populations of astrocytes in the mouse brain that behave in measurably different ways, according to a study published in Nature on 21 October 2026. The new work, led by Junsung Woo and Benjamin Deneen at Baylor College of Medicine with a large team of collaborators, generated a genetically engineered mouse line that allows direct, side-by-side comparison of CD51-positive and CD51-negative astrocytes in the cortex and hippocampus, and used it to assign distinct jobs to each population across development, adulthood and brain injury.

The tool at the heart of the study is a CD51-FLP mouse line created by bacterial artificial chromosome transgenesis. The researchers inserted a FLP recombinase cassette into the Itgav locus, the gene that encodes CD51, immediately after the start codon in the first exon, using the reference transcript NM_008402.3. The BAC was introduced by pronuclear injection in collaboration with Cyagen, and all mice were maintained on a C57BL/6 background. Because FLP recombinase is expressed wherever CD51 is active, crossing these animals with FLP-responsive reporter lines allowed the team to label, sort and genetically manipulate CD51-positive astrocytes while their CD51-negative neighbours remained untouched.

With that platform in place, the authors report that CD51-positive and CD51-negative astrocytes differ in morphology, physiology and developmental behaviour in both the cortex and the hippocampus. Functional experiments then split the two populations by location and task. In the developing cortex, CD51-positive astrocytes selectively promote synapse formation. In the hippocampus, the same subpopulation supports circuit plasticity, the strengthening and reorganization of connections that underlies learning-related change. Mechanistically, the study traces these two effects to different molecules: in the cortex, CD51-positive astrocytes function through glypican 6 to promote synapse formation, whereas in the hippocampus CD51 itself, the protein encoded by Itgav, drives circuit plasticity.

The developmental origin of the subpopulation also came into view. The authors report that CD51-positive astrocytes first appear in the ventricular and subventricular zones around postnatal day one, and that their generation is tied to intrinsic transcription factor programs. In their discussion, the researchers interpret this as evidence that CD51-positive astrocytes represent a specified subtype whose functional properties are set early in development, a conclusion they note is consistent with spinal cord studies showing that developmental patterning generates distinct astrocyte subtypes. They also acknowledge a counterpoint from clonal analyses of cortical astrocytes, which suggest that morphological diversity can arise from environmental cues after cells migrate out of the germinal zones. Whether the functional identity of CD51-positive astrocytes is pre-ordained or shaped by surroundings, they write, is an important question for future work, and clonal analysis of this population could help resolve it. Neuronal activity, at least in the cortex, did not affect the generation of CD51-positive astrocytes, although the authors leave open the possibility that the functional properties of the population show plasticity in other brain regions.

The physiological characterization relied on detailed electrophysiology. In acute brain slices, long-term potentiation was induced at the Schaffer collateral pathway to CA1 using theta-burst stimulation, ten trains of four stimuli at 100 hertz separated by 200-millisecond intervals, with field excitatory postsynaptic potentials recorded in the stratum radiatum and potentiation quantified from the final five minutes of recording. Whole-cell recordings were made from cortical pyramidal neurons and hippocampal CA1 neurons at a holding potential of minus 60 millivolts, with spontaneous excitatory and inhibitory currents isolated pharmacologically using bicuculline or a combination of APV and CNQX. Calcium activity was monitored by two-photon imaging of GCaMP-labelled astrocytes in acute slices, with signals quantified using the GECIquant algorithm to extract amplitude, area under the curve, frequency and spike-width measures from somata and process microdomains.

To probe causality rather than correlation, the team deployed chemogenetics and viral perturbations. hM3Dq-based DREADD stimulation of Gq-coupled signalling was triggered by intraperitoneal CNO injections, given twice daily eight hours apart for multiday dosing or thirty minutes before behavioural testing for acute experiments. CRISPR-based knockdown targets for Gpc6 and Itgav were designed with CHOPCHOP for SaCas9, and overexpression constructs carrying the Itgav or Gpc6 open reading frames were packaged into AAV2/9 at titres between five times ten to the eleventh and seven times ten to the twelfth genome copies per millilitre. Stereotaxic injections targeted the hippocampus bilaterally and multiple cortical sites, and for experiments probing astrocyte-neuron contact, an AAV expressing the neuronal NAPA construct was injected into the thalamus to label thalamocortical projections, with astrocyte-neuron proximity assessed by confocal FRET analysis using the PixFRET plugin and normalized to astrocyte volume and projection density.

The injury experiments connected the subpopulation to disease. Focal strokes were induced by photothrombosis, in which rose bengal was administered intraperitoneally and a 3-millimetre aperture over the skull, 2 millimetres right of bregma, was illuminated for fifteen minutes to target the forelimb area of motor cortex and surrounding sensorimotor regions. Motor recovery was assessed with a grid walk assay in which mice walked spontaneously on a wire grid with 1-centimetre openings while foot faults of the impaired left forelimb were counted relative to total steps, tested before stroke and at seven and twenty-one days after. Lesion volume was quantified by tracing cortex on Nissl-stained sections and applying Cavalieri’s method. The authors report that CD51-positive astrocytes promote functional recovery after stroke through enhanced synaptic remodelling, extending the subpopulation’s role from the developing and mature brain into the injured brain.

Memory-related behaviour was examined with contextual fear conditioning, in which mice received three 1.5-milliampere foot shocks in a distinctive chamber and freezing was scored by a blinded investigator, and with a novel-place recognition task in which the discrimination index measured preference for an object moved to a new quadrant after three days of training. In the discussion, the authors argue that these results suggest a select subset of astrocytes is sufficient to support memory circuit function, paralleling recent findings that ensembles of learning-associated astrocytes operate within memory engrams. They raise the possibility that CD51-positive astrocytes are preferentially recruited to memory engrams in the hippocampus and may participate in engrams associated with other forms of learning, while framing this as a hypothesis rather than a demonstrated mechanism.

Molecular profiling complemented the functional work. Fluorescence-activated cell sorting of dissociated cortex and hippocampus, pooling tissue from three mice per replicate, separated GFP- and tdTomato-labelled astrocytes on a FACSAria II, and RNA from 100,000 sorted events per sample was sequenced on an Illumina NovaSeq X Plus. Differential expression was analysed with DESeq2 separately for each brain region, with significance thresholds of adjusted P below 0.05 and absolute log2 fold change greater than 1, and gene set enrichment was computed with fgsea and GSVA against curated astrocytic and neuronal gene sets and previously defined astrocyte populations A through E. CD51-interacting proteins were identified by anti-HA immunoprecipitation followed by mass spectrometry on a timsTOF Ultra 2, searched against the mouse UniProt proteome. The RNA-seq data have been deposited under GEO accessions GSE312439 and GSE330418.

The authors emphasize methodological rigour throughout: experiments used mice of both sexes aged two to six months, littermates were randomly allocated where possible, all analyses were blinded to group allocation, no sex-dependent effects were detected, and every experiment was repeated independently at least three times with similar results. In their discussion they position the work against earlier single-cell transcriptomic studies, arguing that mRNA profiles cannot predict physiological outputs and that prior gene-restricted astrocyte studies did not directly compare function between prospective subpopulations. Their conclusion is that astrocyte subpopulations exhibit graded functional properties, with CD51-positive astrocytes more efficient at interacting with functioning neural circuits under both normal and pathological conditions, and they speculate that additional functionally distinct astrocyte subpopulations may emerge from interactions with vascular or immune cell populations. The authors argue that their newly developed mouse line and viral tools enable such direct functional comparisons between CD51-positive and CD51-negative astrocytes.

Subject of Research: Functionally distinct CD51-positive and CD51-negative astrocyte subpopulations in the mouse cortex and hippocampus

Article Title: CD51 delineates functionally distinct astrocyte subtypes in neural circuits

Article References: Woo, J., Williamson, M. R., Choi, D., Akdemir, E., Zlatkov, T., Yu, K., Kwon, W., Maleki, E., Ko, Y., Sanchez, K., Maleki, H., Choi, J. M., Serin-Harmanci, A., & Deneen, B. (2026). CD51 delineates functionally distinct astrocyte subtypes in neural circuits. Nature. https://doi.org/10.1038/s41586-026-11186-z

Image Credits: AI Generated

DOI: 10.1038/s41586-026-11186-z

Keywords: astrocytes, CD51, Itgav, glypican 6, synapse formation, hippocampus, cortex, stroke recovery, circuit plasticity, mouse models, glial cells, neuroscience