Loss of pericytic integrin-β1 exacerbates blood-brain barrier damage and hemorrhagic brain injury

Abstract

BACKGROUND

Pericytes play essential roles in blood-brain barrier (BBB) regulation and stroke pathogenesis. Given that pericytes are embedded in the extracellular matrix (ECM), it is speculated that ECM-receptor interactions are involved in these functions. Integrin-β1, the most common integrin subunit that can engage multiple ECM proteins, is highly expressed in pericytes. The function of pericytic integrin β1, however, remains unknown.

METHODS

To address this question, we generated brain pericyte-specific integrin-β1 knockout mice (Itgb1^PKO) by crossing the Atp13a5-CreER with the Itgb1 floxed mice and characterized their phenotypes under homeostatic conditions and after intracerebral hemorrhage (ICH).

RESULTS

Under homeostatic conditions, Itgb1^PKO mice were grossly normal and failed to show BBB disruption or pericyte/astrocyte defects. In the collagenase-induced ICH model, however, the Itgb1^PKO mice exhibited larger hematoma volume, enhanced brain edema, aggravated BBB damage caused by both paracellular and transcellular mechanisms, reduced pericyte number/coverage and AQP4 coverage, increased neuronal death, elevated gliosis, and worsened neurological outcomes. Interestingly, hypertensive Itgb1^PKO mice demonstrated similar changes in the autologous blood model of ICH.

CONCLUSIONS

These results suggest that brain pericyte-derived integrin-β1 is dispensable under homeostatic conditions but plays a protective role in ICH likely through repairing BBB damage and regulating gliosis.

Graphical Abstract

Summary for Basic Science Manuscript:

In two mouse models of intracerebral hemorrhage, ablation of integrin-β1 in brain pericytes leads to exacerbated brain injury and worsened outcomes, highlighting its neuroprotective role in ICH.

Introduction

Stroke, a leading cause of disability and mortality worldwide, is categorized into ischemic and hemorrhagic subtypes¹. Although ischemic stroke is more common, hemorrhagic stroke has a significant higher mortality rate and poorer functional outcomes². Intracerebral hemorrhage (ICH) is a form of hemorrhagic stroke characterized by bleeding into the brain parenchyma. One key pathology of ICH is blood-brain barrier (BBB) disruption, which drives secondary injury, resulting in perihematomal edema, neuroinflammation, and neuronal injury³. The extent of BBB disruption is closely correlated with injury severity and neurological outcomes following hemorrhagic stroke^4,5.

Pericytes, mural cells that cover small blood vessels, play an indispensable role in BBB maintenance under physiological conditions⁶. Accordingly, loss of pericytes has been shown to exacerbates BBB damage and neural degeneration in Alzheimer’s disease^7,8. Similarly, a recent single-cell transcriptomic and imaging study demonstrates that early pericyte dysfunction precedes and contributes to BBB breakdown and neurovascular uncoupling following ischemic stroke⁹. These findings highlight a crucial role of pericyte defects in the pathogenesis of these diseases. Despite these insights, how pericytes regulate BBB integrity and brain injury in the context of ICH remains largely unknown.

Pericytes are embedded in the extracellular matrix (ECM)-rich basal lamina, which undergoes dynamic remodeling after ICH^3,10. It is thus believed that ECM receptors may mediate the effects of pericytes on BBB integrity and ICH outcomes. Major ECM receptors are integrins---heterodimers composed of α and β subunits. Integrin-β1, the most widely expressed β subunit, couples with various α subunits to bind multiple ECM proteins, including laminin and collagen^11,12. Pericytes highly express integrin-β1, which enables them to sense and respond to matrix cues.

In this study, we investigated the function of pericytic integrin-β1 under homeostatic conditions and after ICH using an innovative brain pericyte-specific integrin-β1 conditional knockout mouse line (Itgb1^flox/flox;Atp13a5-CreER, termed Itgb1^PKO)¹³.

Methods

Data Availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Animals

The brain pericyte-specific integrin β1 knockout (Itgb1^PKO) mice were generated by crossing the Itgb1^flox/flox mice with the Atp13a5-CreER line, which drives tamoxifen-inducible recombination specifically in capillary pericytes in the brain¹³. Cre-negative Itgb1^flox/flox littermates were used as controls. Genotyping was performed using polymerase chain reaction (Supplemental Material). All mice were maintained on a C57BL/6J background and studied at 3–5 months of age. Mice of both sexes were used in this study. Animals were housed under standard conditions with free access to food and water. All experimental procedures were designed and conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines¹⁴, adhered to the NIH Guide for the Care and Use of Laboratory Animals, and approved by the University of South Florida IACUC. Sample sizes were determined based on previous reports with similar experimental design^15–17. The genotypes of mice were blinded to the analyzers until after data acquisition.

ICH models

ICH was induced by intracerebral injection of collagenase¹⁸ and autologous blood¹⁹ as described previously. Briefly, mice were anesthetized with isoflurane and placed in a stereotaxic frame (Stoelting Co, IL, USA). For the collagenase model, collagenase (type VII-S; Sigma, St. Louis, USA; 0.225 U in 0.5 μl saline) was delivered via a 24-gauge needle through a burr hole into the striatum at the following coordinates: 0.2 mm anterior to the bregma, 2.2 mm lateral to the midline, 3.7 mm in depth below the skull. For the autologous blood model, 30μl of autologous blood was delivered via a 24-gauge needle into the same coordinates. The injection was performed slowly to minimize reflux, and the needle was left in place for 10 minutes before withdrawal. Mice with the same procedure except collagenase or blood injection were used as sham controls. At 2 and 7 days post injury (dpi), mice were perfused with PBS followed by 4% paraformaldehyde (PFA) after behavioral tests. Animals were excluded when they died before tissue collection.

Hypertension

Hypertension was induced in mice via chronic delivery of angiotensin II as described previously²⁰. Briefly, angiotensin II (1000 ng/kg/minute; Sigma chemical) dissolved in sterile saline was infused using Osmotic pumps (Alzet model 2002), which were implanted subcutaneously under anesthesia.

Systolic blood pressure measurement

Systolic blood Pressure was determined using the noninvasive tail-cuff method (Kent Scientific) as described previously²¹ (Supplemental Material). Systolic blood pressure was measured before and after (5 and 10 days) pump implantation.

Neurological deficit scoring

Functional impairment was assessed using a modified neurological deficit scale as described previously¹⁸. Specifically, body symmetry, gait, climbing, circling behavior, front limb symmetry, and compulsory circling were scored from 0 to 4, with a maximum deficit score of 24. Higher values indicate greater impairment. All assessments were performed by investigators blinded to the genotypes.

Hematoma volume

Hematoma volume was determined by cresyl violet staining on serial sections as described previously¹⁷. Brain swelling was quantified by measuring the area of the ipsilateral and contralateral hemispheres on matched coronal sections. Briefly, the percentage of hemispheric swelling was calculated as: (ipsilateral hemisphere volume - contralateral hemisphere volume) / contralateral hemisphere volume*100.

Immunohistochemistry (IHC)

Brain sections were fixed in 4% PFA for 15 minutes before staining. After extensive PBS was, sections were subjected to standard IHC procedures (Supplemental Material).

RNAscope in situ hybridization

In situ hybridization was performed using the RNAscope multiplex fluorescent reagent kit V2 (Advanced Cell Diagnostics, 323100) according to the manufacturer’s instructions and as described previously²². Sections were processed for RNAscope through the horseradish peroxidase (HRP)-blocking step using the mouse Itgb1 probe (Advanced Cell Diagnostics, 514281) to detect Itgb1 transcripts. Next, the sections were rinsed sequentially in RNAscope Wash Buffer, PBS, and 0.5% PBST, followed by incubation in blocking solution for 1.5 hours at room temperature. Then, the sections were incubated with anti-PDGFRβ (Cell Signaling, 3169S, 1:400) and anti-podocalyxin (R&D Systems, AF1556, 1:400) antibodies overnight, followed by appropriate secondary antibodies the next day. After extensive washes, the sections were mounted in Fluoromount-G with DAPI.

BBB integrity

BBB integrity was assessed using endogenous and exogenous tracers as described in our previous publications^23–25. For endogenous tracers, hemoglobin and IgG were detected by IHC. These tracers were co-stained with vascular marker podocalyxin to reveal blood vessels. The mean gray value of hemoglobin and IgG in podocalyxin-negative area was measured using ImageJ (NIH). Background signal was subtracted. For exogenous tracer, FITC-Dextran-4kD (Sigma, FD4, 250 μg in 50 μl) was injected retro-orbitally into each mouse. After 2 hours of circulation, mice were transcardially perfused with PBS followed by 4% PFA. FITC-Dextran-4kD and podocalyxin signals were revealed by IHC and quantified similarly as described above. In these analyses, 2 random fields from each section, 8–10 serial sections along the rostral-to-caudal axis of each brain, and 5–7 mice were used for quantifications. For BBB integrity under homeostatic conditions, changes in Itgb1^PKO mice were normalized to that in the controls.

Fluoro-Jade C (FJC) staining

Degenerating neurons were revealed and quantified using FJC staining as described previously²⁶. Briefly, brain sections were incubated in 1% NaOH in 80% ethanol for 5 minutes, washed in 70% ethanol and distilled water, and incubated in 0.06% KMnO₄ for 10 minutes. After extensive washes in water, the sections were transferred to 0.0001% FJC (in 0.1% acetic acid) for 10 minutes. Next, the sections were rinsed in water, dried at 56 °C, cleared in xylene, and mounted with DPX. The number of FJC⁺ cells was counted manually in the peri-hematoma region. In this analysis, 2 random images in each section, 8–10 serial sections per mouse, and 6 mice were used for quantification.

Transmission electron microscopy (TEM)

Standard TEM and HRP-TEM were performed to visualize the ultrastructure of tight junctions and quantify endothelial transcytosis, respectively, as described in our previous publications^22,27 (Supplemental Material).

Statistical analyses

Data were analyzed using GraphPad Prism. For neurological deficit scores, two-way ANOVA followed by Tukey post hoc test was used to determine statistical significance between control and Itgb1^PKO groups. For endothelial transcytotic vesicle density, one-way ANOVA followed by Tukey post hoc test was used to determine statistical significance between control and Itgb1^PKO groups. For other analyses, Mann-Whitney test was used to determine statistical significance between control and Itgb1^PKO groups due to the relatively low biological replicates. Data are presented as mean ± SD. Statistical significance was defined as p < 0.05. Sample sizes refer to biological replicates, and all analyses were conducted with investigators blinded to group identity.

Results

Characterization of Itgb1^PKO mice under homeostatic conditions

To investigate the function of pericytic integrin-β1, we generated brain pericyte-specific integrin-β1 conditional knockout (Itgb1^flox/flox;Atp13a5-CreER, termed Itgb1^PKO) mice by crossing the Itgb1^flox/flox mice with the Atp13a5-CreER line. Driven by the Atp13a5 promoter, recombination is induced in brain pericytes specifically¹³. To validate successful ablation of integrin-β1 in brain pericytes in Itgb1^PKO mice after tamoxifen administration, we performed IHC against integrin-β1 and PDGFRβ. Although colocalization of integrin-β1 and PDGFRβ was detected in both control and Itgb1^PKO brains (Figure 1A), quantitation demonstrated significantly decreased integrin-β1 fluorescent intensity in PDGFRβ⁺ areas in Itgb1^PKO mice (Figure 1B), indicating loss of integrin-β1 expression in pericytes. The residual integrin-β1 expression is likely from endothelial cells, which is immediately adjacent to pericytes²⁸. To further distinguish endothelial and pericytic integrin-β1, we performed in-situ hybridization and IHC co-staining. Robust Itgb1 mRNA was detected in PDGFRβ⁺ pericytes from control brains, but absent in those from Itgb1^PKO mice (Figure 1C). Together, these results validate successful ablation of integrin-β1 in brain pericytes in Itgb1^PKO mice.

Figure 1. Integrin-β1 is ablated in pericytes from Itgb1^PKO mice.

A, Representative images of PDGFRβ (red) and Integrin-β1 (green) staining in control and Itgb1^PKO brains. B, Quantification of Integrin-β1 fluorescent intensity in PDGFRβ-positive areas in control and Itgb1^PKO brains. *p < 0.05 (Mann-Whitney Test), n = 4 mice per group. C, Representative images of Itgb1 mRNA (red), PDGFRβ (magenta), podocalyxin (green), and DAPI (blue) staining in control and Itgb1^PKO brains. Data are shown as mean ± SD.

To determine whether pericyte-specific loss of integrin-β1 affects baseline vascular properties, we first examined BBB integrity by IHC against hemoglobin and IgG. Itgb1^PKO and control mice showed comparable hemoglobin (Figure S1A and B) and IgG (Figure S1C and D) levels in brain tissues. Similarly, we failed to detect intravenously injected 4kDa-FITC-Dextran, a small fluorescent tracer, in the brains of control or Itgb1^PKO mice (Figure S1E and F). These results indicate that Itgb1^PKO mice have intact BBB integrity under homeostatic conditions. Next, we assessed if loss of integrin-β1 in pericytes affects their structural metrics using PDGFRβ and CD31 (vessel marker) staining (Figure S2A). Quantifications revealed comparable pericyte coverage (Figure S2B) and pericyte number (Figure S2C) in Itgb1^PKO and control brains. These findings are consistent with the intact BBB integrity of Itgb1^PKO mice and indicate that pericytic integrin-β1 is dispensable for their density and association with the vasculature under homeostatic conditions. In addition, we also examined astrocytic endfoot coverage on blood vessels using AQP4 and podocalyxin staining (Figure S2D). No difference in AQP4 coverage along the vasculature was found between Itgb1^PKO and control mice under homeostatic conditions (Figure S2E), indicating intact astrocytic endfoot coverage. Collectively, these results demonstrate that pericyte-specific deletion of integrin-β1 does not alter BBB integrity, pericyte structural metrics, or astrocytic endfoot coverage in uninjured adult brains, highlighting a minimal role of pericytic integrin-β1 under homeostatic conditions.

Itgb1^PKO mice display exacerbated brain injury after ICH

We further investigated the functional significance of pericytic integrin-β1 in ICH using the most widely used collagenase model²⁹. After ICH, control mice showed a well-defined hematoma that gradually diminished over time. Although a similar trend was observed, Itgb1^PKO mice developed significantly larger hematomas at both 2 dpi and 7 dpi (Figure 2A and B), suggesting that pericytic integrin-β1 negatively regulates hematoma expansion after ICH. Brain swelling was also quantified by comparing the volume of the ipsilateral and contralateral hemispheres. Compared to the controls, Itgb1^PKO mice exhibited significantly increased brain swelling at both 2 dpi and 7 dpi (Figure 2C), indicating exacerbated edema at both acute and subacute phases after ICH. No sex-specific effects were detected in hematoma volume (Figure S3A) or brain swelling (Figure S3B). Next, the number of degenerating neurons was determined using FJC staining. In sham groups, minimal FJC⁺ were detected in control and Itgb1^PKO mice (Figure 2D and E). After ICH, however, FJC⁺ neurons were markedly increased in the perihematomal regions in both control and Itgb1^PKO mice, with the latter having more dramatic increases at both 2 dpi and 7 dpi (Figure 2D and E), indicating exacerbated neuronal death in Itgb1^PKO mice after ICH. Interestingly, male and female mice showed comparable FJC⁺ cells in each condition (Figure S3C), indicating no sex-specific changes. Furthermore, neurological function was evaluated by neurological deficit scores. Both control and Itgb1^PKO mice exhibited increased neurological deficit scores shortly after injury (Figure 2F), indicating successful induction of ICH. However, Itgb1^PKO mice demonstrated significantly higher neurological deficit scores compared to the controls up to 14 days after injury (Figure 2F), indicating worse neurological function in the acute, subacute, and chronic phases. Again, male and female mice exhibited similar changes in neurological deficit scores after ICH (Figure S3D). Together, these results suggest that loss of integrin-β1 in brain pericytes aggravates hemorrhagic brain injury equally in males and females.

Figure 2. Itgb1^PKO mice show exacerbated brain injury after ICH.

A, Representative images of Cresyl violet-stained control and Itgb1^PKO brains. Dashed lines mark injury areas. B, Quantification of hematoma volume in control and Itgb1^PKO mice. *p < 0.05 and **p < 0.01 (Mann-Whitney Test), n = 4–6 mice per group. C, Quantification of brain swelling percentage in control and Itgb1^PKO mice. **p < 0.01 (Mann-Whitney Test), n = 4–6 mice per group. D, Representative images of FJC staining in control and Itgb1^PKO brains from sham groups and after ICH. E, Quantification of FJC⁺ cells in control and Itgb1^PKO mice. *p < 0.05 and **p < 0.01 (Mann-Whitney Test), n = 6 mice per group. F, Neurological deficit scores in control and Itgb1^PKO mice from sham groups and after ICH. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, compared to controls at the same time points (Two-way ANOVA followed by Tukey post hoc test), n = 35–37 mice per group for 0–7 dpi and n = 6 mice per group for 14 dpi. Data are shown as mean ± SD.

Hypertension is a primary risk factor for ICH³⁰. To determine if similar results are observed in hypertensive conditions and increase the translational potential of our findings, we further investigated ICH outcomes in control and Itgb1^PKO mice with such risk factor. Specifically, hypertension was induced in control and Itgb1^PKO mice via chronic delivery of angiotensin II (Figure S4A) as described previously²⁰. Hypertension was successfully induced in these mice as demonstrated by the increased systolic blood pressure (Figure S4B). Next, ICH was induced in hypertensive control and Itgb1^PKO mice using the autologous blood model, the most clinically relevant ICH model³¹. Under hypertensive conditions, sham-operated control (Control^H) and Itgb1^PKO (Itgb1^PKO-H) mice failed to develop hematoma or brain swelling, although slightly higher neurological deficit scores were observed in the latter (Figure S4C–E). At 2 dpi after ICH, however, Itgb1^PKO-H mice exhibited significantly larger hematoma volume (Figure S4C) and exacerbated brain swelling (Figure S4D) compared to the controls. Elevated neurological deficit scores were observed in Itgb1^PKO-H mice after ICH, although these increases did not reach statistical significance (Figure S4E). Together, these results suggest that hypertensive Itgb1^PKO mice develop exacerbated brain injury in the autologous blood model of ICH, highlighting a beneficial role of brain pericyte-derived integrin-β1 in ICH.

Itgb1^PKO mice demonstrate aggravated BBB disruption after ICH

BBB disruption drives secondary injury and correlates with the severity of brain injury in ICH^3,32. To assess BBB integrity, we examined brain accumulation of hemoglobin and IgG, two proteins predominantly found in the blood. Negligible levels of hemoglobin were detected in sham-operated control and Itgb1^PKO mice (Figure 3A), again indicating intact BBB integrity under homeostatic conditions. After ICH, however, hemoglobin signal was detected outside podocalyxin⁺ blood vessels in both control and Itgb1^PKO mice (Figure 3A). Quantification revealed significantly increased hemoglobin levels in Itgb1^PKO mice compared to the controls at both 2 dpi and 7dpi (Figure 3B). Similarly, IgG levels were also increased in brain parenchyma of Itgb1^PKO mice compared to the controls at 2 dpi (Figure 3C and D), indicating enhanced BBB disruption after ICH. Again, male and female mice demonstrated comparable hemoglobin and IgG leakage under sham conditions and at both 2 dpi and 7 dpi (Figure S3E and F), indicating no sex-specific differences in BBB integrity. Similarly, hypertensive Itgb1^PKO mice demonstrated increased hemoglobin leakage at 2 dpi in the autologous blood model of ICH, compared to the controls (Figure S5A and B). Together, these results suggest an essential role of brain pericyte-derived integrin-β1 in BBB repair after ICH.

Figure 3. Itgb1^PKO mice show exacerbated BBB impairment after ICH.

A, Representative images of hemoglobin (red) and podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. B, Quantification of hemoglobin leakage in control and Itgb1^PKO mice. ns, non-significant, *p < 0.05, **p < 0.01 compared to controls at the same time points (Mann-Whitney Test), n = 6–7 mice per group. C, Representative images of IgG (Magenta) and podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. D, Quantification of IgG leakage in control and Itgb1^PKO mice. ns, non-significant, *p < 0.05 (Mann-Whitney Test), n = 5–6 mice per group. Data are presented as mean ± SD.

Itgb1^PKO mice show increased paracellular leakage after ICH

Brain endothelial cells express high levels of tight junction proteins, such as Claudin-5 and ZO-1, which seal intercellular gaps and prevent paracellular leakage³³. To determine whether altered tight junction integrity contributes to the enhanced BBB disruption in Itgb1^PKO mice, Claudin-5 and ZO-1 coverage of podocalyxin⁺ capillaries was quantified. Under sham conditions, high Claudin-5 and ZO-1 coverages were found in control and Itgb1^PKO mice, and no differences were detected between genotypes (Figure 4A–D). After ICH, however, reduced Claudin-5 and ZO-1 coverages were found in both control and Itgb1^PKO mice with the latter having more pronounced decreases at both 2 dpi and 7 dpi (Figure 4A–D). Consistent with these changes, TEM revealed intact, electron-dense tight junctions in control and Itgb1^PKO brains from sham-operated animals at the ultrastructural level (Figure 4E). After ICH, however, impaired tight junctions containing gaps are frequently observed in both groups with more severe defects in Itgb1^PKO brains (Figure 4E). Together, these findings suggest that pericytic integrin-β1 regulates tight junction protein expression after ICH, and that the paracellular mechanism contributes to the enhanced BBB damage in Itgb1^PKO mice after ICH.

Figure 4. Itgb1^PKO mice show aggravated paracellular leakage after ICH.

A, Representative images of claudin-5 (red) and podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. B, Quantification of claudin-5 coverage in control and Itgb1^PKO mice. ns, non-significant, **p < 0.01(Mann-Whitney Test), n = 6 mice per group. C, Representative images of ZO-1 (red) and podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. D, Quantification of ZO-1 coverage in control and Itgb1^PKO mice. ns, non-significant, *p < 0.05, **p<0.01 (Mann-Whitney Test), n = 6 mice per group. E, Representative TEM images showing the ultrastructure of TJs in control and Itgb1^PKO mice from sham groups and after ICH. Data are shown as mean ± SD. TJ, tight junction; EC, endothelial cell; AE, astrocytic endfoot; BL, basal lamina; TEM, transmission electron microscopy.

Itgb1^PKO mice exhibit enhanced transcytosis after ICH

Brain endothelial cells have extremely low rate of transcytosis, which limits transcellular leakage³⁴. To assess whether the transcellular mechanism also contributes to the exacerbated BBB disruption in Itgb1^PKO mice after ICH, we examined the expression of Mfsd2a, a suppressor of endothelial transcytosis³⁵. While Mfsd2a coverage was similar between genotypes under sham conditions, Itgb1^PKO mice showed significantly greater reduction of Mfsd2a coverage at both 2 dpi and 7 dpi compared to the controls (Figure 5A and B). Next, we further examined transcytosis in endothelial cells using HRP under TEM. No HRP-containing transcytotic vesicles (dark dots) were detected in sham-operated control or Itgb1^PKO mice (Figure 5C). After ICH. however, HRP-containing transcytotic vesicles were detected in endothelial cells from both control and Itgb1^PKO mice (Figure 5C). Quantification revealed significantly increased endothelial transcytotic vesicle density in Itgb1^PKO mice at 2 dpi compared to the controls (Figure 5D). An increasing trend was also observed in Itgb1^PKO mice at 7 dpi, although statistical significance was not achieved (Figure 5D). Subcellular distribution analysis showed substantially higher transcytotic vesicle density in the luminal membrane (Figure 5E) but not cytosol (Figure 5F) or abluminal membrane (Figure 5G) of endothelial cells in Itgb1^PKO mice at both 2 dpi and 7 dpi compared to the controls. Together, these findings suggest that pericytic integrin-β1 inhibits endothelial transcytosis after ICH, and that the transcellular mechanism also contributes to the enhanced BBB damage in Itgb1^PKO mice after ICH.

Figure 5. Itgb1^PKO mice show decreased MFSD2a coverage after ICH.

A, Representative images of Mfsd2a (red) and podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. B, Quantification of Mfsd2a coverage in control and Itgb1^PKO mice. ns, non-significant, *p < 0.05, **p < 0.01(Mann-Whitney Test), n = 6 mice per group. C, Representative TEM images showing horseradish peroxidase-containing transcytotic vesicles in endothelial cells of control and Itgb1^PKO mice from sham groups and after ICH. EC, endothelial cell; TEM, transmission electron microscopy. D, Quantification of total transcytotic vesicle density in endothelial cells from control and Itgb1^PKO mice. ns, non-significant, *p < 0.05 (One-way ANOVA followed by Tukey post hoc test), n = 19–28 capillaries per group. E, Quantifications of luminal, cytosolic, and abluminal transcytotic vesicle density in endothelial cells from control and Itgb1^PKO mice. ns, non-significant, *p < 0.05 (One-way ANOVA followed by Tukey post hoc test), n = 19–28 capillaries per group. Data are shown as mean ± SD.

Itgb1^PKO mice show pericyte defects after ICH

Since pericyte loss strongly correlates with BBB disruption^6,7,36,37, we further examined pericyte coverage and density in these mice by IHC against PDGFRβ and podocalyxin. In sham-operated animals, comparable pericyte coverage and number were observed in control and Itgb1^PKO mice (Figure 6A–C). After ICH, however, pericyte coverage and number were decreased in both groups (Figure 6A–C). Compared to the controls, Itgb1^PKO mice exhibited significantly lower pericyte coverage and number at both 2 dpi and 7 dpi (Figure 6A–C), indicating aggravated pericyte defects. Similarly, hypertensive Itgb1^PKO mice demonstrated more pronounced decrease of pericyte coverage and number at 2 dpi in the autologous blood model of ICH (Figure S5C–E). Together, these results suggest that brain pericyte-derived integrin-β1 actively regulates pericyte density and coverage after ICH.

Figure 6. Itgb1^PKO mice show pericyte defects and decreased AQP4 coverage after ICH.

A, Representative images of PDGFRβ (green) and CD31 (red) staining in control and Itgb1^PKO brains from sham groups and after ICH. B, Quantification of pericyte coverage in control and Itgb1^PKO mice. ns, non-significant, **p < 0.01 (Mann-Whitney Test), n = 6 mice per group. C, Quantification of normalized pericyte number in control and Itgb1^PKO mice. ns, non-significant, **p < 0.01 (Mann-Whitney Test), n = 6 mice per group. D, Representative images of AQP4 (red) and Podocalyxin (green) staining in control and Itgb1^PKO brains from sham groups and after ICH. E, Quantification of AQP4 coverage in control and Itgb1^PKO mice. ns, non-significant, **p<0.01 (Mann-Whitney Test), n = 6 mice per group. Data are shown as mean ± SD.

Itgb1^PKO mice demonstrate abnormal astrocyte polarity after ICH

Astrocytes cover capillaries and pericytes with their endfeet, where AQP4 is expressed exclusively. The polarized expression of AQP4 is associated with BBB integrity³⁸. To determine whether loss of pericytic integrin-β1 impacts astrocyte polarity, we examined AQP4 coverage along capillaries by IHC. Under sham conditions, AQP4 covered most capillaries in both control and Itgb1^PKO mice (Figure 6D and E). After ICH, however, significantly reduced AQP4 coverage was found in both control and Itgb1^PKO mice, with the latter demonstrating more prominent decreases at both 2 dpi and 7 dpi (Figure 6D and E). These findings indicate that pericytic integrin-β1 is involved in the regulation of astrocyte polarity after ICH.

Itgb1^PKO mice display enhanced gliosis after ICH

Neuroinflammation is a key driver of secondary brain injury and directly correlates with ICH outcome^39,40. To assess neuroinflammation, we examined microgliosis and astrogliosis by Iba-1 and GFAP staining, respectively. Ramified microglia, characterized by small cell body and long/thin processes, were found in both control and Itgb1^PKO mice under sham conditions (Figure 7A). They changed to an ameboid morphology, characterized by enlarged cell body and shortened processes, after ICH in both groups (Figure 7A). Quantification revealed comparable Iba-1⁺ area in control and Itgb1^PKO mice under sham conditions or at 2 dpi, but significantly increased Iba-1⁺ area in Itgb1^PKO mice at 7 dpi (Figure 7B), indicating exacerbated microglial activation. Similarly, comparable GFAP⁺ area was found in control and Itgb1^PKO mice under sham conditions, whereas Itgb1^PKO mice showed substantially increased GFAP⁺ area at both 2 dpi and 7 dpi (Figure 7C and D), indicating aggravated astrogliosis. These findings suggest increased gliosis in Itgb1^PKO mice after ICH. Similarly, increased microgliosis and astrogliosis were also observed in hypertensive Itgb1^PKO mice at 2 dpi in the autologous blood model of ICH (Figure S6). Together, these findings highlight an essential role of brain pericyte-derived integrin-β1 in gliosis after ICH.

Figure 7. Itgb1^PKO mice exhibit increased gliosis after ICH.

A, Representative images of Iba-1 (magenta) staining in control and Itgb1^PKO brains from sham groups and after ICH. B, Quantification of Iba-1⁺ area (%) in control and Itgb1^PKO mice. ns, non-significant, *p<0.05 (Mann-Whitney Test), n = 5–8 mice per group. C, Representative images of GFAP (green) in control and Itgb1^PKO brains from sham groups and after ICH. D, Quantification of GFAP⁺ area (%) in control and Itgb1^PKO mice. ns, non-significant, **p < 0.01 (Mann-Whitney Test), n = 6 mice per group. Data are shown as mean ± SD.

Discussion

In this study, we demonstrate that brain pericyte-specific deletion of integrin-β1 fails to affect BBB integrity under homeostatic conditions but markedly exacerbates BBB disruption and brain injury following ICH. Specifically, we find that mutant mice lacking brain pericytic integrin-β1 display enlarged hematoma volume, increased degenerating neurons, and aggravated neurological dysfunction. These abnormalities are accompanied by exacerbated vascular injury, including BBB damage due to both paracellular and transcellular leakage, pericyte defects, and impaired astrocyte polarity. Our findings identify pericytic integrin-β1 as a key regulator of vascular integrity and brain injury in ICH, which may be targeted for the treatment of ICH.

Integrin-β1-dependent signaling has emerged as a central mechanism regulating vascular stability and integrity in normal brain. For example, integrin-β1 functional blocking antibody Ha2/5 reduces tight junction protein expression in endothelial cells and increases vascular permeability in vitro and in vivo⁴¹. Similar results were observed in mutant (Itgb1^flox/flox;Mx-Cre) mice with partial loss of integrin-β1⁴². A recent study reported that the interaction between pericyte-derived vitronectin and endothelial integrin-α5β1 maintained both blood retina barrier and BBB integrity⁴³, highlighting a crucial role of endothelial integrin-β1 in vascular integrity. Unlike endothelial integrin-β1, we showed that brain pericyte-specific deletion of integrin-β1 failed to affect BBB integrity without ICH, indicating a dispensable role of pericytic integrin-β1 in BBB maintenance under homeostatic conditions. Interestingly, ablation of PDGFRβ⁺ cell-derived integrin-α8, which only partners with integrin-β1, leads to BBB disruption under homeostatic conditions⁴⁴, suggesting an important role of PDGFRβ⁺ cell-derived integrin-α8β1 in BBB maintenance. This discrepancy may be due to different Cre lines, which mark distinct populations. Specifically, the brain pericyte-specific Atp13a5-CreER line¹³ was used in our study, whereas the Pdgfrb-P2A-CreERT2 and Pdgfrb-iCre, which label pericytes, smooth muscle cells, and fibroblasts^45,46, were used to abrogate integrin-α8. Interestingly, single-cell RNAseq analysis revealed high expression levels of integrin-α8 in fibroblasts and smooth muscle cells but not pericytes⁴⁴. It is possible that deletion of integrin-α8β1 in smooth muscle cells and/or fibroblasts may indirectly compromise BBB integrity. An alternative explanation is that different integrins may have distinct functions. Brain pericytes highly express multiple β1-containing integrins, including integrin-α1β1, -α4β1, and -α7β1^47,48. Upon ablation of Itgb1, all these integrins were deleted in brain pericytes. It is possible that loss of these integrins may counteract the effect of integrin-α8β1 deletion, leading to no obvious BBB disruption in Itgb1^PKO mice.

Another key finding of this study is that deletion of integrin-β1 in brain pericytes aggravated BBB damage and brain injury after ICH, highlighting an essential role of pericytic integrin-β1 in the pathogenesis of ICH. To the best of our knowledge, this is the first brain pericyte-specific loss-of-function study for integrin-β1 in ICH. Similar neuroprotective roles of integrin-β1 have also been reported in ischemic stroke. Specifically, integrin-β1 is induced predominantly in blood vessels at 72 hours after cerebral ischemia and reperfusion injury, and its blockage reduces cerebral angiogenesis and worsens stroke outcomes, indicating a beneficial role of endothelial integrin-β1 in ischemic stroke⁴⁹. Integrin-α5β1 mediates pericyte adhesion to ECM protein perlecan and cooperates with PDGF-BB/PDGFRβ signaling to regulate pericyte recruitment during vascular repair in an in vitro model of ischemic stroke⁵⁰. A recent loss-of-function study demonstrated that ablation of integrin-α8β1 in PDGFRβ⁺ cells resulted in exacerbated BBB damage, pericyte defects, and worsened neurological recovery after ischemic stroke, establishing an crucial role of integrin-β1 in BBB repair and functional recovery after ischemic stroke⁴⁴. Together, these findings support that integrin-β1 is a molecular target with therapeutic potential in stroke.

While our findings provide direct in vivo evidence that pericytic integrin-β1 actively regulates vascular function and ICH outcomes, it has four limitations. First, we predominantly focused on the acute and subacute phases after ICH, leaving vascular remodeling and functional recovery at the chronic phase largely unexplored. Next, although hypertension was incorporated into experimental design, aging was not due to potentially high mortality. Aged mice should be used to validate these findings in the future. In addition, the α subunits that pair with integrin-β1 in brain pericytes were not identified. As an initial study to characterize the functions of integrins in brain pericytes, we focused on integrin-β1, the most widely expressed integrin subunit. Determining the α subunits that mediate these effects in brain pericytes will enable the development of highly selectively therapies. Furthermore, the molecular signaling pathways downstream of integrin-β1were not directly interrogated and warrant further investigation. Addressing these limitations in future studies will clarify the functional significance of pericytic integrin-β1 in ICH and allow precision medicine approaches with improved outcomes and reduced side effects.

Conclusions

Collectively, our results demonstrate that brain pericyte-derived integrin-β1 is a key regulator of vascular function and brain injury after ICH. Specifically, we show that loss of integrin-β1 in brain pericytes fails to affect BBB integrity under homeostatic conditions, but leads to exacerbated vascular dysfunction (BBB disruption, pericyte defects, and abnormal astrocyte polarity) and aggravated brain injury (larger hematoma volume, increased neuronal degeneration, aggravated gliosis, and worse neurological function) in two ICH models. These findings elucidate the function of brain pericyte-derived integrin-β1 in the pathogenesis of ICH and provide a novel molecular target with therapeutic potential in ICH.

Supplementary Material

Supplemental Methods

Figures S1–S6

ARRIVE Checklist

Abbreviations

BBB

blood-brain barrier

dpi

days post injury

ECM

extracellular matrix

FJC

Fluoro-Jade C

HRP

horseradish peroxidase

ICH

intracerebral hemorrhage

IHC

Immunohistochemistry

Itgb1^PKO

Itgb1^flox/flox

Atp13a5-CreER

Itgb1^PKO-H

PFA

paraformaldehyde

TEM

Transmission electron microscopy