The alpha7 integrin subunit in astrocytes promotes endothelial blood–brain barrier integrity Free

The mammalian brain contains an intricate web of blood vessels that interact with neurons and glia within multicellular complexes termed neurovascular units (Marin-Padilla, 1985). Growth factors, extracellular matrix (ECM) proteins in endothelial and parenchymal basement membranes, and adhesion and signaling receptors mediate cell–cell contact and communication to control key neurovascular functions (Paredes et al., 2018). Astrocyte end feet and pericytes closely juxtapose the abluminal surface of blood vessels and communicate with the endothelium to modulate cerebral blood flow and metabolism as well as control blood–brain barrier (BBB) integrity in health and disease (Kur et al., 2012). We recently reported that ablation of perivascular astrocytes in the postnatal brain leads to disruption of vascular endothelial barrier functions, neurological deficits and premature death (Morales et al., 2021). However, very little is understood about the genes and pathways in astrocytes that regulate endothelial cell BBB functions during brain development and physiology, including roles for cell adhesion to ECM proteins in endothelial and parenchymal basement membranes.

The gene encoding megalencephalic leukoencephalopathy with subcortical cysts 1 (MLC1) is expressed almost exclusively in astrocytes of the human and mouse brain. Loss-of-function mutations in MLC1 lead to vascular endothelial cell dysfunction and associated neurocognitive deficits (van der Knaap et al., 2012), and overexpression of MLC1 promotes brain tumor malignancy (Lattier et al., 2020). We engineered an in vivo model in which the murine Mlc1 gene drives expression of enhanced green fluorescent protein (Mlc1-EGFP) in perivascular astrocytes (Toutounchian and McCarty, 2017). This model allows selective fractionation of perivascular astrocytes from the brain and analysis of specific transcripts by quantitative RNA sequencing (Yosef et al., 2020). Several genes in perivascular astrocytes with putative involvement in BBB regulation have been identified, including Itga7, which encodes the integrin α7 subunit. Itga7 mRNA is also expressed in pericytes and vascular smooth muscle cells in the brain. Integrins are heterodimeric cell surface receptors for many ECM protein ligands (Naba et al., 2016), with integrin-mediated adhesion and signaling events playing essential functions in development and pathophysiology in multiple mammalian organs (Wong et al., 2016). Integrin α7 is a ∼130 kDa protein that dimerizes with the β1 integrin subunit (Song et al., 1993) and promotes the development and physiology of skeletal muscle (Barraza-Flores et al., 2020). Integrin α7 protein localizes to neuromuscular junctions, with global deletion of the Itga7 gene in mice leading to congenital muscular dystrophy-like phenotypes as a result of defective adhesion between muscle cells and laminin protein ligands (Mayer et al., 1997).

Laminins are heterotrimeric proteins consisting of α, β and γ subunits that are organizing components of basement membranes in multiple organs (Marshall et al., 2015). The Lama2 gene encodes the α2 subunit of laminin, which is found in laminin-211/merosin, functioning as the main ECM ligand for integrin α7β1 in skeletal muscle (Barraza-Flores et al., 2020). Genetic deletion of Lama2 in mice results in skeletal muscle degeneration that partially phenocopies defects that develop in Itga7^−/− mouse mutants (Xu et al., 1994). Loss of laminin α2 subunit expression correlates with reduced levels of α7 integrin protein in skeletal muscle cells, as reported in different knockout mouse models and in humans with LAMA2 mutations (Cohn et al., 1999). Genetic deletion of Itga7 does not augment the congenital muscular dystrophy pathologies detected in Lama2 knockout mice (Gawlik and Durbeej, 2015). These data indicate a reciprocal relationship between α7β1 integrin and laminins in the ECM, with normal expression of each component promoting stability of the ligand–receptor complex to promote neuromuscular development and homeostasis.

Here, we present in vivo and in vitro evidence that reveal essential functions for Itga7 in brain astrocytes during endothelial barrier maturation and physiology. Ablation of Itga7 in mice leads to defects in BBB integrity related to defective adhesion between astrocytes with laminins in parenchymal basement membranes. Vascular endothelial cells in Itga7^−/− mice show defects in the expression of junctional proteins important for normal BBB stability. Collectively, these data reveal essential functions for the α7 integrin subunit in the neurovascular unit in controlling endothelial barrier formation and integrity via adhesion to laminin substrates in the ECM.

In order to characterize astrocyte genes with potential links to BBB biology, we utilized a mouse model in which the endogenous Mlc1 gene drives expression of enhanced green fluorescent protein (Mlc1-EGFP). These mice express EGFP in Mlc1⁺ perivascular astrocytes (Toutounchian and McCarty, 2017). Mlc1-EGFP/+ mice were crossed with GLAST-DsRed/+ mice, which express the fluorescent reporter DsRed in astrocytes throughout the postnatal brain (Regan et al., 2007). In these double-transgenic mice, which we have previously characterized in detail (Yosef et al., 2020), EGFP and DsRed are co-expressed exclusively in perivascular astrocytes whereas non-perivascular astrocytes express DsRed but are negative for EGFP (Fig. 1A). Perivascular astrocytes fractionated from Mlc1-EGFP/+;GLAST-DsRed/+ mice showed enriched expression of Itga7 mRNA, which encodes the α7 integrin protein subunit (Fig. 1B). Lower levels of Itga7 mRNA were detected in non-perivascular cells that were DsRed⁺ but negative for EGFP. In addition, three different laminin mRNAs – Lama2, Lama3 and Lama5, which encode the α2, α3 and α5 laminin subunits, respectively – showed enriched expression in perivascular astrocyte fractions as determined by prior RNA sequencing (Fig. 1B) (Josef and McCarty, 2020). The murine integrin genes Itga3, Itga6 and Itgb1, which encode the laminin-binding integrins α3β1 and α6β1 (Halder et al., 2022; Milner et al., 2008), were not differentially expressed in Mlc1⁺ versus Mlc1⁻ astrocytes (Fig. S1).

We also queried Itga7 expression in various open-source datasets in which whole transcriptome profiling was performed using tissue and/or single cells fractionated from the brain. As shown in Fig. S2, analysis of Itga7 mRNA expression in different regions of the adult mouse brain (https://www.proteinatlas.org/) revealed the highest levels in the cerebellum with lower levels in the cortex and thalamic regions. In another mouse scRNAseq database (https://betsholtzlab.org/VascularSingleCells/database.html) (He et al., 2018), Itga7 mRNA expression was not detected in vascular endothelial cells, but was highly enriched in astrocytes and pericytes/vascular smooth muscle cells (Fig. S3). ITGA7 expression was also detected in human brain astrocytes and pericytes (Fig. S3), as determined by analysis of the human BBB single-cell RNA-sequencing database (https://twc-stanford.shinyapps.io/human_bbb/) (Yang et al., 2022).

Integrin α7 functions were next analyzed in vivo, using Itga7 wild-type (Itga7^+/+) and homozygous null (Itga7^−/−) mice, which were generated by targeted insertion of a lacZ cDNA in the endogenous Itga7 allele (Flintoff-Dye et al., 2005). Knock-in of the lacZ cDNA results in expression of β-galactosidase via the endogenous Itga7 promoter. Genotypes of Itga7^+/+, Itga7^+/− and Itga7^−/− F1 mice were confirmed using DNA from ear snips and genomic PCR (Fig. S4). β-Galactosidase expression, which serves as a surrogate marker for α7 integrin, was interrogated in fixed brain sections. Analysis of Itga7^+/+ and Itga7^−/− mice revealed α7 integrin subunit expression in astrocytes as well as pericytes and vascular smooth muscle cells in the cerebral cortex based on double immunofluorescence labeling with anti-β-galactosidase antibodies (Fig. 1C,D, Fig. S5). α7 integrin protein was detected mainly in GFAP-expressing astrocytes in the thalamus (Fig. 1E,F) and white matter regions of the cerebellum (Fig. 1G,H). Interestingly, in the Itga7^−/− brain regions analyzed we detected more GFAP-expressing astrocytes, indicative of astrogliosis. The results for integrin α7 protein expression in astrocytes are consistent with our studies of Mlc1-EGFP mice (Fig. 1) as well as with published transcriptomic analyses of Itga7 mRNA in glial and vascular cell types isolated from the neonatal mouse brain (He et al., 2018) (Figs S2, S3).

In approximately 15% of Itga7^−/− mice (n=20 of 128 postnatal mutants), we detected skeletal muscle wasting phenotypes as well as brain vascular pathologies as early as postnatal day (P) 14 and these pathologies become progressively more severe (Fig. 2A). Itga7^−/− mutants with severe phenotypes displayed reduced body weights in comparison with wild-type littermate controls. The remaining ∼85% of Itga7^−/− mutant mice, which did not develop early (<P60) neurodegenerative phenotypes, survived for more than 12 months, although these aged mutants displayed progressive BBB defects (see below).

All Itga7^−/− mice with severe phenotypes (n=20) died by P60 or required sacrifice owing to deteriorating health. Hydrocephalus as well as related neurological deficits, including paresis and seizures, were apparent in Itga7^−/− mice with early-onset neurological phenotypes (Fig. 2B,C). Next, P30 wild-type (n=3) and Itga7^−/− (n=3) mice with severe phenotypes were perfused with 10 kDa FITC-Dextran to analyze BBB integrity in vivo. FITC-Dextran can be cardiac-perfused and does not cross the intact BBB. As shown in Fig. 2D-F, pathological extravasation of FITC-Dextran was detected in the cerebral cortices of Itga7^−/− mice. We also detected abnormal FITC-Dextran extravasation in the Itga7^−/− cerebellum (Fig. S6). We also detected pathophysiological leakage of serum albumin, an abundant circulating protein, across the BBB in Itga7^−/− mice, but not in Itga7^+/+ control littermates (Fig. 2G-I).

The brain vascular pathologies in Itga7^−/− mice suggested defective α7 integrin-mediated cell adhesion and/or communication between astrocytes and vascular cell types that comprise the neurovascular unit. Therefore, the cytoarchitecture of the neurovascular unit was analyzed in brains of wild-type and Itga7^−/− mice that displayed hydrocephalus. Widespread reactive astrogliosis, as determined by changes in astrocyte morphology as well as GFAP expression, was detected in various regions of the Itga7^−/− mouse brain (Fig. 3A-G, Fig. S7). Reactive Iba1⁺ microglial cells were also detected surrounding blood vessels in the Itga7^−/− brain (Fig. 3H-N). The ultrastructural features of cerebral blood vessels and surrounding mural cells were next analyzed in P50 wild-type control and Itga7^−/− mice using transmission electron microscopy. In Itga7^+/+ control brain parenchyma, astrocyte end feet were closely juxtaposed to cerebral endothelial cells and pericytes (Fig. 4A,B). In contrast, we detected focal areas of perivascular edema associated with separation between blood vessels and surrounding astrocyte end feet in the parenchyma of the Itga7^−/− cerebral cortex (Fig. 4C,D). Reactive astrocytes migrate to regions of brain pathology to form glial scars and deposit various ECM proteins to promote tissue fibrosis, neovascularization, and immune cell-mediated neural repair (Frik et al., 2018). However, we did not detect an increase in Itga7-expressing reactive astrocytes, as monitored by β-galactosidase expression, in response to injury using a cortical wound model (Fig. S8). Similarly, β-galactosidase was not detected in Iba1-expressing reactive microglia in injured brains (Fig. S9).

Given that α7 integrin serves as a receptor for laminins, and promotes laminin protein stability in the basement membranes of skeletal muscle (Barraza-Flores et al., 2020), expression patterns of laminins were analyzed in wild-type and Itga7^−/− brain sections. We used a pan-laminin antibody, which mainly detects Lama1/laminin α1, a subunit of laminin-111 and laminin-221, which are highly expressed in the brain (Ichikawa-Tomikawa et al., 2012). As shown in Fig. 5A,C, in the wild-type brain laminin α1 protein subunit was enriched in the basement membrane surrounding blood vessels in the thalamus, cerebellum and most other brain regions. Significantly reduced levels of laminin α1 were detected in vascular basement membranes in the cerebellum of P60 Itga7^−/− mouse brains as well as in other brain regions, including the thalamus (Fig. 5B,D). Quantification of laminin protein levels revealed a significant reduction in Itga7^−/− brains (Fig. 5E). There was also a difference in the spatial expression patterns of laminins in the Itga7^−/− brains, with reduced levels in vascular basement membranes, but continued expression in surrounding neuronal cell types. These data are consistent with prior findings in Itga7^−/− mice with skeletal muscle degeneration phenotypes, showing reduced expression of the laminin α2 subunit (Mayer et al., 1997).

We detected persistent defects in neurovascular unit architecture in aged Itga7^−/− mice. Specifically, in comparison to 1-year-old (P365) Itga7^+/+ controls (Fig. 5F,H), reduced laminin α1 subunit expression in vascular basement membranes was detected in Itga7^−/− mutant brains (Fig. 5G,I). Collectively, these data reveal that α7β1 integrin is necessary for laminin α1 subunit deposition and/or stability in vascular basement membranes of the postnatal mouse brain.

To characterize adhesion and signaling functions for α7β1 integrin in neurovascular biology further, we cultured primary astrocytes from brains of Itga7^+/+ and Itga7^−/− neonatal mice. Nearly 100% of primary cells expressed the astrocyte biomarkers GFAP and/or nestin (Fig. 6A,B). In contrast, the pericyte marker NG2 (Cspg4), the endothelial marker CD31 (Pecam1) and the oligodendrocyte marker myelin basic protein were expressed at low or undetectable levels in the astrocyte cultures (Fig. S10). There was expression of α7 integrin in cultured astrocytes and in brain lysates; however, α7 integrin protein was not detected in astrocytes cultured from Itga7^−/− mouse pups (Fig. 6C,D). We next analyzed cell-surface integrin expression patterns using a membrane impermeable biotin to globally label surface proteins followed by anti-integrin immunoprecipitation. As shown in Fig. 6E, wild-type astrocytes expressed multiple cell-surface integrin heterodimers, including α5β1, α6β1, α7β1, αvβ3 and αvβ8. In contrast, the α7β1 integrin heterodimer was absent in Itga7^−/− astrocyte lysates (Fig. 6E). The expression levels and heterodimeric combinations of other integrins remained unchanged in Itga7^−/− astrocytes (Fig. 6E). Wild-type control astrocytes and Itga7^−/− astrocytes were next plated on various ECM proteins, revealing that α7 integrin is required for adhesion to laminin-111, collagens I and IV, and fibrinogen in vitro (Fig. 6F). We also detected reduced growth and survival of Itga7^−/− astrocytes in comparison with wild-type astrocytes (Fig. 6G). Cultured Itga7^−/− astrocytes also expressed reduced levels of the laminin α1 protein subunit, as determined by immunoblotting (Fig. 6H), which is consistent with the quantitative RNA-sequencing data (Fig. 1) as well as the anti-laminin immunofluorescence analysis of Itga7^−/− brain tissue (Fig. 5).

Next, we fractionated CD31⁺ vascular endothelial cells from brains of P365 adult wild-type and Itga7^−/− mice, and the expression of proteins that comprise tight junctions at the BBB was analyzed by immunoblotting (Fig. 7A). As shown in Fig. 7B,C, reduced levels of the tight junction proteins claudin 5 and ZO-1 (Tjp1) were detected in endothelial cells fractionated from Itga7^−/− mice in comparison with Itga7^+/+ controls. Adherens junctions have been reported to regulate endothelial cell barrier integrity via cross-talk with components of tight junctions (Rahimi, 2017). Along these lines, we also detected reduced expression of the adherens junction protein VE-cadherin (cadherin 5) in CD31⁺ endothelial cells isolated from Itga7^−/− mouse brains (Fig. 7B,C). Double-immunofluorescence labeling was also performed on brain sections from wild-type and Itga7^−/− mice (n=3 mice per genotype) using antibodies directed against claudin 5 and CD31 to label tight junctions in vascular endothelial cells. In comparisons with controls, reduced levels of claudin 5 protein were detected in CD31⁺ vascular endothelial cells in Itga7^−/− cerebral cortices (Fig. 7D-F) and cerebellar white matter regions (Fig. 7G-I). We also detected reduced expression of ZO-1 in endothelial cells from Itga7^−/− mice, unlike wild-type controls, which showed ZO-1 protein enrichment in CD31⁺ tight junctions (Fig. S11).

Itga7^−/− mutants that did not develop early neurological phenotypes, such as brain hemorrhage and hydrocephalus (∼85% of Itga7^−/− mice), survived for more than 18 months. However, after approximately 9-12 months, these aged Itga7^−/− mice displayed progressive wasting phenotypes, including kyphosis (Fig. S12A). Analysis of weights of P365 Itga7^+/+ and Itga7^−/− mice revealed a significant reduction in body mass (Fig. S12B). Itga7^−/− mice also developed hindlimb rigidity as well as spontaneous seizures, which were not detected in wild-type littermate controls (Fig. S13). Collectively, these data reveal that α7β1 integrin in astrocytes is required for maintenance of BBB integrity via adhesion to and stabilization of laminin protein ligands in the vascular basement membrane (Fig. 8A,B). Loss of Itga7 expression leads to diminished levels of laminins, likely owing to compromised ECM stability (Fig. 8C). We propose that these defects collectively lead to enhanced BBB permeability, linked to defective expression of junctional proteins in brain endothelial cells. In addition, the α7 integrin subunit likely interacts with other cell-surface proteins to regulate cytoplasmic signaling events in astrocytes to further promote BBB stability.

The more severe defects in BBB integrity and premature death occur in ∼15% of Itga7^−/− mice. Cerebral blood vessels in these mice, particularly in the thalamus and cortex, have abnormal morphologies, associated with edema and hemorrhage. There is some variability in the vascular pathologies within these Itga7^−/− mice. This spectrum of brain vascular pathologies may be related to functional compensation by other laminin receptors, such as α6 integrin (Krebsbach and Villa-Diaz, 2017). Indeed, a prior report has shown that deletion of Itga7 in combination with Itga6 leads to defective Schwann cell interactions with peripheral nerves, whereas single gene deletions do not cause myelination defects (Pellegatta et al., 2013). In Lama2 mutant mice, which develop muscle wasting phenotypes, αv integrin expression is increased in skeletal muscle cells (Accorsi et al., 2015). Although we do not detect changes in integrin heterodimer expression in Itga7^−/− cultured astrocytes, it is possible that there are adhesion pathways that partially compensate for loss of α7 integrin in vivo.

The remaining ∼85% of Itga7^−/− mutants live for more than a year, but display sequelae of progressive muscle wasting and neurologic defects. The variability in phenotypes and overall survival may be related to the mixed C57Bl6/129S4 background of Itga7^−/− mutants. We propose that strain-specific genetic modifiers that impact expression of laminin and/or other receptors for laminins, or additional ECM proteins, account for the phenotypic heterogeneity. Ultrastructural studies reveal that BBB defects and local edema precede the development of the progressive phenotypes that afflict these Itga7^−/− mice after 9 months, suggesting that BBB pathologies precede and contribute to the progressive phenotypes in aged Itga7^−/− mice. These data also support the suggestion that the more severe vascular phenotypes in the 15% of younger Itga7^−/− mice are the direct result of defective astrocyte-endothelial cell interactions, rather than a secondary and indirect consequence of Itga7 gene expression.

The patterns of α7 integrin expression in perivascular mural cells in different brain regions are strikingly heterogeneous. For example, in cortical regions integrin α7 protein is found in vascular smooth muscle cells and pericytes, with astrocyte expression also detected. In contrast, α7 integrin expression is detected in GFAP⁺ perivascular astrocytes in the hippocampus, thalamus and white matter regions of the cerebellum. It will be important to determine the cues that regulate differential expression of Itga7 in pericytes versus astrocytes and to decipher whether α7β1 adhesion occurs mainly via laminins or other ECM ligands. Regulatory pathways that promote Itga7 expression in skeletal smooth muscle cells may be similar to those that control expression in vascular smooth muscle cells of the brain. Control of Itga7 expression in astrocytes may involve distinct cell-autonomous transcriptional pathways because these cells originate from neuroectodermal progenitors (Zheng et al., 2022) whereas smooth muscle cells are derived mainly from the mesoderm (Tare et al., 2008). Given that we are using whole-body knockout (Itga7^−/−) mice, combined with the observation that Itga7 mRNA is detected in astrocytes as well as in brain pericytes and vascular smooth muscle cells, we cannot exclude the possibility that cells other than astrocytes are contributing to BBB maturation and integrity. It will be important to distinguish the relative roles of Itga7 in astrocytes versus pericytes/vascular smooth muscle cells using Cre/lox strategies.

Interestingly, Lama2^−/− mice, which lack laminin-211, develop variable BBB pathologies related to defects in endothelial cell tight junction protein expression (Menezes et al., 2014). We propose that α7β1 integrin in perivascular astrocytes binds to laminins (laminin-111 and laminin-211) and possibly other ECM ligands to promote communication with the vascular endothelium and provide structural stability to blood vessels. Prior studies have shown that Lama5/laminin-511 expression in cerebral vascular basement membranes promotes BBB stability via control of tight junctional proteins during brain physiology and in experimental autoimmune encephalomyelitis (Song et al., 2017; Welser et al., 2017; Wu et al., 2009). Pericyte interactions with endothelial cells via the platelet-derived growth factor pathway are crucial for BBB development (Armulik et al., 2010; Daneman et al., 2010). Given our data showing essential Itga7 functions at the BBB, it is enticing to speculate that α7β1 integrin in both astrocytes and pericytes may cooperate to induce endothelial barrier maturation and stability. The skeletal muscle defects and BBB pathologies in Itga7^−/− mice are more severe than those that develop in Lama2^−/− mutants, suggesting that there are additional ECM ligands for α7β1 integrin that have roles in stabilizing the adjacent brain vasculature. These results suggesting multiple α7β1 integrin ligands in vivo are also consistent with the broad defects in ECM detected with Itga7^−/− astrocytes in the in vitro ECM adhesion assays. In addition to laminins, our prior transcriptome profiling efforts identified genes encoding other ECM proteins, including Frem1 and Frem2, that interact with laminins and facilitate basement membrane integrity (Yosef et al., 2020). Although Frem proteins contain integrin-binding motifs (Kiyozumi et al., 2005), the cell-surface receptors that directly bind Frems in the brain have not been definitively identified.

The RGD-binding integrin αvβ8 is also enriched in astrocytes and plays roles in vascular basement membrane adhesion and endothelial cell communication (McCarty, 2020). Genetic deletion of Itgav or Itgb8 expression in astrocytes leads to profound neurovascular pathologies related to defective latent-TGFβ activation (Arnold et al., 2014; Hirota et al., 2015). We have also recently reported that truncation of the cytoplasmic tail of β8 integrin leads to extracellular adhesion defects related to latent-TGFβ activation and signaling to vascular endothelial cells during embryonic brain development (De et al., 2022). Adult Itgb8 mutant mice do not display progressive neurological deficits like we report here in Itga7^−/− mice, suggesting that the α7 integrin subunit and the β8 integrin subunit are not directly coupled in a heterodimeric complex. In human glioblastoma stem cells, both ITGB8 (Guerrero et al., 2017) and ITGA7 (Haas et al., 2017) mediate adhesion to vascular basement membranes to promote cancer growth and invasion. It will be important to further analyze possible cross-talk between α7 and β8 integrin subunits in glioblastoma cells, identify potential α7β8 integrin heterodimerization, and/or determine whether integrin affinity is impacted by laminins and/or other ECM ligands.

In addition to ECM adhesion, it is also possible that α7β1 integrin-activated signaling pathways may contribute to BBB integrity. Indeed, a prior report has shown that α7β1 integrin activates focal adhesion signaling complexes to promote fibroblast polarity and migration (Mielenz et al., 2001). Integrin α7 has two main splice variants – α7A and α7B – that impact the amino acid sequence of the cytoplasmic domain (Martin et al., 1996). It will be important to determine whether the splice variants are differentially expressed more in pericytes of the cerebral cortex versus astrocytes in other regions of the brain, such as the thalamus and cerebellum. The cytoplasmic domain splice variants may differentially regulate inside-out activation of α7β1 integrin adhesion to the ECM. It is also possible that α7 integrin intracellular signaling effectors recruited by the different cytoplasmic domain splice variants lead to the secretion of factors by astrocytes that impact BBB permeability. α7β1 integrin in perivascular astrocytes may also interact with other cell-surface receptors, such as growth factor receptor tyrosine kinases, or other transmembrane proteins, such as AGGF1 (Yu et al., 2022), to regulate their ligand affinities and/or intracellular signaling functions. We are currently investigating integrin α7 signaling pathways in astrocytes using various biochemical assays to decipher how they may impact endothelial cell barrier formation and integrity. In skeletal muscle cells, α7β1 integrin is a component of the dystrophin–glycoprotein complex (DGC), which comprises multiple transmembrane and intracellular proteins that link the cytoskeleton to the ECM and maintain skeletal muscle homeostasis and is dysfunctional in muscular dystrophies (Marshall et al., 2015). Interestingly, components of the DGC, including Aqp4, are expressed in astrocytes of the adult brain with the DGC playing roles in ECM adhesion (Waite et al., 2012); however, we understand little about their functions at the neurovascular unit, including links to α7 integrin in BBB regulation. It will be interesting to characterize interactions between α7 integrin, dystrophin, dystroglycan α/β subunits, α-dystrobrevin and other components of the DGC in astrocytes using proteomics strategies and these efforts are ongoing.

Lastly, one limitation of this study is the exclusive use of murine models to study Itga7 functions in astrocyte-endothelial cell adhesion and signaling. It will be important to link the requirements for the mouse Itga7 gene in BBB biology to similar functions at the human neurovascular unit. When we compared Itga7 and ITGA7 mRNA patterns in mouse and human endothelial cells, overlapping gene expression in astrocytes was confirmed, supporting important functions at the human BBB. Interestingly, a recent report has shown that mutations in the human ITGA7 gene contribute to adult-onset cardiovascular pathologies (Bugiardini et al., 2022). Similar to Itga7^−/− mice, humans with loss-of-function mutations in LAMA2 or ITGA7 also develop congenital muscular dystrophies (Hayashi et al., 1998; Helbling-Leclerc et al., 1995). Some patients with LAMA2-associated congenital muscular dystrophies also display associated brain vascular pathologies (Arreguin and Colognato, 2020). Hence, based on our data from mice we postulate that mutations in ITGA7 that impair normal interactions between astrocytes and blood vessels may also contribute to developmental and/or adult-onset human neurovascular disorders.

This study was reviewed and approved by the MD Anderson Cancer Center Institutional Animal Care and Use Committee (IACUC) in compliance with the National Research Council Guide for the Care and Use of Laboratory Animals. The approved protocol number is ACUF-00001108-RN02. Heterozygous and homozygous null mice in which the Itga7 locus is targeted with a lacZ cDNA insertion have been described previously (Flintoff-Dye et al., 2005). The following primers: Itga7 forward (5′-GGTGGGTGAAGGAATGAGTG-3′), Itga7 WT reverse (5′-AAAGGTAGCCCAAAGCTTGA-3′) and bgalR2 (5′-GACCTGCAGGCATGCAAGC-3′) were used to detect the targeted and wild-type alleles, respectively. Generation and characterization of the Mlc1-EGFP mouse model has been described previously (Toutounchian and McCarty, 2017).

Cerebral cortices were dissected and FACS-sorted cells (Mlc1⁺ astrocytes and Mlc1⁻ astrocytes) from the brains of adult-mice (P30) were pelleted and total RNA was extracted by following the QIAGEN RNeasy micro kit guidelines. RNA was reverse transcribed using the Invitrogen SuperScript III cDNA synthesis kit, and 100 ng cDNA was used per reaction. Quantitative RT-PCR to analyze Itga7, Lama2, Lama3 and Lama5 expression in astrocytes was performed using TaqMan primers and TaqMan Universal MasterMix II on a StepOnePlus Real-Time PCR system (Applied Biosystems). The following TaqMan primers (Thermo Fisher Scientific) were used: integrin subunit alpha-7 (Itga7, Mm00434400 m1), mouse laminin alpha 2 (Mm00550083_m1), mouse laminin alpha 3 (Lama3, Mm01254735_m1), laminin subunit alpha 5 (Lama5, Mm01222029 m1), Gapdh (Mm99999915 g1). Levels of Itga7, Lama2, Lama3, Lama5 RNA were normalized to Gapdh expression levels. Relative gene expression ratios were calculated using quantitative ΔCT methods.

Pups at P3 were genotyped, sacrificed and cerebral cortices were removed and subjected to enzymatic dissociation in a papain-containing buffer (Miltenyi Biotec, 130-095-942). Dissociated cells at high density were resuspended in low-glucose Dulbecco's Modified Eagle's Medium (DMEM) (Hyclone) containing 10% bovine calf serum with 1% penicillin/streptomycin and seeded on laminin-111 (Sigma-Aldrich)-coated dishes. Adherent astrocytes were cultured to confluence for 7-14 days. At this point, cells were shaken at 37°C for 2 h on a rotary shaker (200-250 rpm). This step removes most oligodendrocytes, microglia, and any neurons. Cells were passaged no more than three times prior to any experimentation as previously described (Hirota et al., 2015; Lee et al., 2015). Itga7-dependent cell viability was quantified by counting adherent cells grown in complete medium using the CellTiter-Glo luminescent cell viability assay kit as we have previously detailed (Chen et al., 2020). Briefly, cells were collected and 5×10³ cells were plated into 96-well plates. CellTiter-Glo reagent was added into the culture medium of each well. Plates were incubated at room temperature for 10 min and the luminescence intensity was measured with a microplate reader.

Cell adhesion assay was carried out using a ECM cell adhesion array kit (CytoSelect 48-Well Cell Adhesion Assay, Cell Biolabs, Inc.) following the manufacturer's instructions. Briefly, primary astrocytes from wild-type and Itga7^−/− mice (n=3 per genotype) cultures were passaged once they were confluent in T75 tissue culture flasks; serum-starved astrocytes were allowed to attach to an ECM-coated 48-well plate; bovine serum albumin-coated wells were used as a negative control. Standard cell culture technique involving 0.25% trypsin, 0.1% EDTA in HBSS [without calcium, magnesium and sodium bicarbonate] (MT 25053CI, Corning)] was used to detach cells followed by trypsin neutralization (ScienCell Research Laboratories, 0113) and centrifugation (250 g, 5 min, room temperature). Viable cell count was obtained (Vi-CELL Analyzer, Beckman Coulter) and cells were gently re-suspended in 4 ml of assay buffer. Subsequently, 150 µl of the cell suspension containing 1.35×10⁵ cells was added to serum-free media for 90 min. The non-adherent cells were gently removed by washing with PBS. The resulting adherent cells were fixed and stained. Finally, the stain was extracted and absorbance readings were taken at 560 nm (Synergy HTX multi-mode reader, BioTek).

Mice were anesthetized at different time points and bodies were fixed by cardiac perfusion of 4% paraformaldehyde (PFA) in PBS. Brains were removed and immersed in 4% PFA for an additional 16-24 h at 4°C. After fixation, brains were embedded in 4% agarose and sectioned at 100 µm on a vibratome and stored in PBS at 4°C. To analyze similar brain regions between Itga7^+/+ and Itga7^−/− mice, we used defined anatomic structures, such as the hippocampus, as reference points, and when necessary used online references (http://www.hms.harvard.edu/research/brain/atlas.html) to identify exact regions in sections. Sections were permeabilized and blocked with 10% donkey serum in PBS-T (1× PBS supplemented with 0.1% Triton X-100) for 1 h at room temperature, followed by an overnight 4°C incubation with primary unconjugated antibodies diluted in the blocking solution. The sections were then washed with PBS-T and incubated with secondary antibodies (1:500) in the blocking solution for 2 h. Sections were again washed three times with PBS-T, then briefly washed with PBS. The sections were mounted on pre-treated microscope slides, sealed using Vectashield with DAPI mounting media (Vector Laboratories) and kept at 4°C until imaging. Confocal Images were acquired using an Olympus FLUOVIEW FV3000 with 10×, 20× and 30× objectives. All comparative images were taken with the same laser power and gain settings in order to make qualitative comparisons between staining levels in different samples. Multiple fields of view were imaged from biological replicates.

Immunohistochemical staining was performed on formalin-fixed, paraffin-embedded Itga7^+/+ and Itga7^−/− mouse brain tissue. The tissue was deparaffinized at 65°C and rehydrated in decreasing concentrations of alcohol. Antigen retrieval was performed using a sodium citrate buffer at pH 6. The tissue was blocked in 10% serum of the same species as the secondary antibody for 1 h at room temperature. The tissue was then incubated overnight at 4°C in primary unconjugated antibodies diluted in blocking buffer. Immunohistochemical analyses were performed with the following primary antibodies: CD31 (R&D Systems, AF3628) and GFAP (Dako, Z0334). The slides were then washed in PBS and incubated with biotinylated secondary antibodies (1:500-1:1000) in the blocking buffer for 1 h at room temperature. Following another series of washes in PBS, the slides were incubated in Vector Laboratories ABC reagent (PK-4000) for 30 min at room temperature. Following more washes with PBS, ImmPACT DAB Substrate (SK-4105) was added for 5-10 min before rinsing the slides in ddH₂O and counterstaining with Hematoxylin and Eosin for 30-45 s. After rinsing in cold tap water, the slides were then dehydrated in increasing concentrations of alcohol and mounted using mounting media and stored until images were collected. Light microscope images were acquired using an Olympus BX43 with a 20× objective.

Mice were euthanized by CO₂ in accordance with IACUC regulations. Mouse brains were dissected, dissociated using the Miltenyi Biotec gentleMACS Octo Dissociator (130-096-427) and CD31⁺ cells were selected using the Miltenyi Biotec Adult Brain Dissociation Kit (130-107-677), anti-CD31 microbeads (130-097-418) and LS Columns (130-042-401), as previously described (Chen et al., 2020). Briefly, the brains were minced, enzymatically digested and homogenized by means of the gentleMACS program 37C_ABDK_01. The brain homogenate was filtered through a 70 µm cell strainer to obtain a single-cell suspension. Debris and myelin were removed by phase separation using the debris removal solution provided in the kit. Red blood cells were removed by lysis using the red blood cell removal solution. The remaining cells were selected for expression of the mouse CD31 antigen by labeling with Miltenyi Biotec anti-CD31 microbeads and then magnetically separated with Miltenyi Biotec LS columns. The final eluted CD31⁺ cell fraction was washed in PBS and either flash-frozen on dry ice or analyzed immediately.

Wild-type control and mutant animals were deeply anesthetized and then perfused with 10 ml 1 mg/ml FITC-Dextran (10 kDa, Sigma-Aldrich). Perfusion rate and volume were kept consistent between mice. Brains were removed and post-fixed in 4% PFA/PBS overnight. Sagittal brain sections were prepared with a vibratome. In addition, mouse serum albumin extravasation was quantified in brains from wild-type and Itga7^−/− mice (n=3 per genotype) using 20 Å∼ fluorescence images (n=3-5 per genotype) taken from the cerebral cortex. ImageJ software was used for quantitation of FITC extravasation (n=3 wild-type and Itga7^−/− mice) using randomly selected images (n=3-4 fields per sample) taken from the cerebral cortices.

Whole-cell lysates were collected under normal culture conditions in 1% NP-40 (50 mM Tris-HCl pH 8.0, 150 mM NaCl) lysis buffer with protease and phosphatase inhibitor tablets (Roche). Total protein was measured by BCA assay (Pierce), and then denatured at 95 C for 5 min in 5× SDS loading buffer containing 2.5% 2-mercaptoethanol (Sigma-Aldrich) and 30 μg of protein was resolved on 4-15% Tris-glycine gels. Immunoblotting was performed with nitrocellulose membranes (Bio-Rad), blocked using Odyssey TBS-based blocking buffer (LI-COR), and then incubated with specific primary antibodies diluted in blocking buffer supplemented with 0.1% Tween-20 overnight at 4°C. Target proteins were normalized to total cellular/housekeeping protein, β-actin. Secondary antibodies (IRDye 800CW goat anti-rabbit and IRDye 680RD goat anti-mouse) (LI-COR) were incubated in the dark at room temperature for 30 min. Dual-channel infrared scan and quantitation of immunoblots were conducted using the Odyssey CLx infrared imaging system with Image Studio (Ver. 5.2) (LI-COR).

For biotinylation and immunoprecipitation experiments, cell-surface proteins of confluent astrocyte cultures were biotinylated (EZ-link™Sulfo-NHS-Biotin, 21217, Thermo Fisher Scientific) followed by quenching with TBS (Tris-buffered saline, T60075-4000, RPI). Cells were lysed using the same protocol described above to obtain total soluble protein lysates and protein concentration was determined using BCA method. Then, 100-200 μg of protein lysate was incubated overnight with 1-2 μg of primary antibody at 4°C, and 10-20 μl of Protein A agarose (11134515001, Roche) (washed with cold PBS by centrifugation at 664 g for 2 min and finally resuspended in RIPA buffer) was added per tube containing protein–primary antibody conjugates and incubated at 4°C for 1 h. The primary antibody agarose conjugate was immunoprecipitated by centrifugation (664 g, 2 min, 4°C) and washed with RIPA buffer containing protease and phosphatase inhibitors. Finally, the pellet was re-suspended in nonreducing sample buffer and western blot analysis was performed using the same protocol described above. Immunoblots were scanned using Odyssey CLx infrared imaging system with Image Studio (LI-COR).

All primary and secondary antibodies used for immunoblotting, immunoprecipitation and/or immunofluorescence are summarized in Tables S1 and S2.

Cortical stab wounds were generated in both Itga7^+/+ wild-type control and Itga7^+/− heterozygous littermates to analyze astrocytic activation and recruitment to the site of injury (Pekny and Nilsson, 2005). A single incision made from the anterior pole of the skull to the posterior ridge. A 3-5 mm wound was made through the pial surface at the initial stereotactic point of entry (1.5 mm rostral, 1.5 mm anterior) using a sterile scalpel blade (no. 11). Seven days after the injury, mice in each group were anesthetized, cardiac-perfused with 4% PFA/PBS, and brain sections were analyzed. Resting brains of Itga7^+/− littermates were used as the day 0 (uninjured) control.

Immunofluorescence images were acquired using an Olympus Fluoview FV3000 confocal laser-scanning microscope. Multidimensional acquisition was conducted using z-stacks with 2.5 µm slicing intervals at a scan rate of 4 ms/pixel with a resolution of at least 1024×1024 pixels per slice and digitally compiled in FV31S-SW (version 2.4.1.198). Image acquisition parameters, including exposure time, laser power, gain and voltages, were fixed for each imaging channel. Immunohistochemistry images were captured using an Olympus BX43 light microscope. Fluorescence signal intensity of channels were measured using the standard ‘color histogram’ module in ImageJ.

Control and mutant animals were deeply anesthetized by intraperitoneal injection of Avertin (250 mg/kg). A 20-gauge needle was inserted into the left cardiac ventricle and animals were perfused with Karnovsky solution (1.25% formaldehyde, 2.5% glutaraldehyde and 0.03% picric acid in 100 mM cacodylate buffer; Polysciences, Inc.). Brains were then removed and postfixed overnight with a solution containing 3% glutaraldehyde plus 2% paraformaldehyde in 0.1 M cacodylate buffer, pH 7.3 for 1 h. After fixation, the samples were washed and treated with 0.1% Millipore-filtered cacodylate-buffered tannic acid, postfixed with 1% buffered osmium tetroxide for 30 min, and stained en bloc with 1% Millipore-filtered uranyl acetate. The samples were dehydrated in increasing concentrations of ethanol, infiltrated, and embedded in Polybed812 medium. The samples were polymerized in a 60°C oven for 2 days. Ultrathin sections were cut in a Leica Ultracut microtome, stained with uranyl acetate and lead citrate in a Leica EM Stainer, and examined in a JEM 1010 transmission electron microscope (JEOL, Inc.) at an accelerating voltage of 80 kV. Digital images were obtained using AMT Imaging System (Advanced Microscopy Techniques Corporation).

Quantification of confocal images was performed using ImageJ software (NIH, USA) (Rueden et al., 2017; Schindelin et al., 2012). GraphPad Prism 8.0 was used to plot mean values (n=3 or greater, ±s.e.m., unless otherwise indicated), compare between experimental and control groups and to determine statistical differences by unpaired Student's t-test and two-way ANOVA (Tukey post-hoc analysis) at 95% confidence intervals (α value 0.05). P<0.05 was considered statistically significant.

We thank the members of the McCarty laboratory for insightful comments on the manuscript. We also thank Dr Kechen Ban (Brain Tumor Center Animal Core) for his insightful technical advice with the cortical wound healing model and FITC-Dextran cardiac perfusions. The following NCI-funded University of Texas M.D. Anderson Cancer Center Core Facilities were instrumental in data acquisition: the shRNA and ORFeome Core for providing high-titer lentiviruses, the Research Histopathology Facility for generation of brain sections for immunohistochemical analyses, the Flow Cytometry and Cellular Imaging Facility for sorting GFP^± cells, the High-Resolution Electron Microscopy Facility for ultrastructural analyses, and the Sequencing and Microarray Facility for Sanger DNA sequencing.