Endothelial and mural laminin-α5 contributes to neurovascular integrity maintenance

Background Laminin-α5, a major component of the basal lamina, is predominantly synthesized by endothelial and mural cells (pericytes and vascular smooth muscle cells) in the CNS. Loss of laminin-α5 in either population fails to induce any abnormalities due to functional redundancy. Thus, the functional significance of laminin-α5 in neurovascular integrity remains unknown. Here, we hypothesize that ablation of laminin-α5 in both endothelial and mural cells increases neurovascular permeability. Methods The compound knockout mice were generated by crossing laminin-α5 floxed mice with Tie2-Cre and PDGFRβ-Cre, which target endothelial cells and mural cells, respectively. Neurovascular permeability in these mutants was determined with both exogenous and endogenous tracers. Endothelial paracellular and transcellular permeability was assessed by examining the expression of tight junction proteins and transcytosis-associated proteins. In addition, transmission electron microscopy (TEM) was used to visualize tight junction ultrastructure and endothelial caveolae vesicles. Defects in pericytes and astrocytes were investigated by examining pericyte coverage/contact and astrocyte polarity. Results Elevated neurovascular permeability was observed in the mutants. Subsequent studies found increased Caveolin-1 and decreased major facilitator superfamily domain-containing protein 2a (MFSD2A) expression, but unaltered Claudin-5 or zonula occludens-1 (ZO-1) expression. Consistent with these results, mutant mice exhibited increased endothelial caveolae vesicle number with intact tight junction structure under TEM. Additionally, pericyte coverage and contact were also decreased in the mutant mice, while astrocyte polarity was unaffected. Conclusions These results strongly indicate that endothelial and mural cell-derived laminin-α5 actively maintains neurovascular integrity via the transcellular rather than paracellular mechanism.


Background
Neurovascular unit (NVU) is a dynamic structure located at the interface of the CNS and circulation system [1,2].By tightly regulating what enters and leaves the brain, it plays an important role in maintaining brain homeostasis.Not only is NVU disfunction observed in most neurological disorders, it also actively contributes to the pathogenesis of many of them [2,3].NVU is mainly composed of endothelial cells, pericytes, astrocyte endfeet, and a non-cellular component-the basal lamina [1,2].Although how different cells regulate neurovascular integrity is well studied, the role of the basal lamina in neurovascular integrity remains largely unknown.
Considering potential functional redundancy, we hypothesize that endothelial and mural cell-derived laminin-α5 may be able to compensate for each other's loss.To test this hypothesis and investigate the functional significance of laminin-α5 in neurovascular integrity, we further generated compound mutant mice with laminin-α5 deficiency in both endothelial and mural cells and characterized their neurovascular integrity in this study.

Mice
The Laminin-α5 flox/flox mice [24] and platelet-derived growth factor receptor beta (PDGFRβ)-Cre + mice [25] were generous gifts from Drs.Jeffrey Miner and Volkhard Lindner, respectively.The Tie2-Cre + (Jax:008863) mice were obtained from the Jackson Laboratory.The α5-TPKO (Laminin-α5 flox/flox ;PDGFRβ-Cre + ;Tie2-Cre + ) mice were generated by crossing the above three mouse lines.Since PDGFRβ-Cre + ;Tie2-Cre + mice failed to show obvious defects and are indistinguishable from Laminin-α5 flox/flox mice, both were used as controls for α5-TP-KO mice.All mice were maintained on a C57Bl6 background.Mice (both sexes) at the age of 4∼6 months were used in this study.Mice were housed in the animal facility of the University of South Florida and provided with ad libitum access to food and water.The experimental protocols and procedures of this study were approved by the Institutional Animal Care and Use Committee at the University of South Florida and were conducted in accordance with the National Institutes of Health's Guide for the Care and Use of Laboratory Animals.

Genotyping
Collected mouse samples were subjected to genomic DNA isolation, followed by PCR amplification.To differentiate Tie2-Cre and PDGFRβ-Cre, promoter-specific primers were used.The primers used in this study are shown below.Tie2-Cre-Forward:

In vivo permeability assay
Neurovascular permeability was assessed in vivo as described in our previous publications [10,12,26,27].For exogenous marker, 50ul of 5 mg/ml Sulfo-NHS-Biotin (ThermoFisher, 21,217) in sterile saline was injected into control and α5-TPKO mice intravenously.After 6 h of circulation, mice were transcardially perfused with 4% PFA.In brain sections, Sulfo-NHS-Biotin was visualized with Avidin-FITC (1:200, BD, 554,057).For endogenous marker, Hemoglobin was detected by immunohistochemistry. Sulfo-NHS-Biotin and Hemoglobin were co-stained with vascular marker Podocalyxin.For quantifications, three random fields from each section, 8 serial sections along the rostral-to-caudal axis of each brain, and 5 animals were used.Due to the small number of leakage spots in each brain region, tracer leakage was determined using the whole brain to enable a more accurate quantification.

Tissue processing
Mice were anesthetized by intraperitoneal injection of inactin (100 mg/kg body weight), followed by transcardiac perfusion with PBS and/or 4% PFA.Brains were collected and subjected to serial sectioning.Briefly, 20 μm-thick serial sections were cut with Cryostat (Micro HM 550, Thermo Scientific), and eight sections evenly distributed along the rostral-to-caudal axis were collected from each brain.Brain sections were stored at − 80 °C until use.

TEM
Mice were anesthetized and transcardially perfused with PBS followed by 0.1 M sodium cacodylate buffer containing 4% PFA and 4% glutaraldehyde.Brain tissues were dissected out, fixed overnight, and post-fixed in 1% osmium tetroxide and 1% potassium ferrocyanide.Next, the collected brain tissues were en bloc stained with 2% uranyl acetate, dehydrated, and embedded in resin.An RMC MT-X microtome (Boeckeler Instruments) was used to cut ultra-thin sections, which were post-stained with 2% uranyl acetate and 1% lead citrate.Sections were examined and photographed using JEOL JEM1011 (JEOL) at 80 kV.

Angioarchitecture analysis
The open-source software "Angiotool" (National Cancer Institute, USA) was used to assess brain angioarchitecture [28].Briefly, brain sections were subjected to Podocalyxin staining.Thresholding techniques were applied to eliminate small particles, ensuring the quantification of actual vessels.Four parameters, including total vessel length, vessel density, branching index, and lacunarity, were calculated using the "Angiotool" software.Total vessel length is defined as the sum of Euclidean distances between pixels of all vessels; vessel density is expressed as the percentage of the area occupied by vessels within the explant area; branching index is defined as the number of vessel junctions per unit area; and lacunarity, an index for vascular structural nonuniformity, is reported as the average lacunarity over all box sizes.

Image analyses
In this study, z-projection images were used for analyses.Analyses were performed in three different brain regions (cortex, hippocampus, and striatum).Two random images from each section, eight serial sections evenly distributed along the rostral-to-caudal axis from each brain, and five mice were used for quantifications.Data analysis was performed by investigators blinded for the genotypes.
For Laminin, Nidogen, Collagen IV, Caveolin-1, and MFSD2a, mean fluorescent intensity was calculated using ImageJ.For tight junction protein (TJP) expression, their coverage and contact were calculated using the Angiotool software.Specifically, TJP coverage was quantified as the percentage of TJP-positive fluorescent area over Podocalyxin-positive capillary area.TJP contact was quantified as the percentage of TJP-positive fluorescent length over Podocalyxin-positive capillary length.Similarly, pericyte coverage and contact were quantified as the percentages of PDGFRβ-positive fluorescent area and length over Podocalyxin-positive capillary area and length were quantified, respectively.AQP-4 coverage and contact were quantified as the percentages of AQP-4-positive fluorescent area and length over Podocalyxin-positive capillary area and length, respectively.For endothelial vesicles, the number of endothelial pinocytotic vesicles in TEM images was manually counted and normalized to endothelial area.In this analysis, twenty-five to twenty-seven capillaries from three mice were used for quantification.

Statistical analyses
All statistical analyses were performed using the Graph-Pad Prism 10 software.The non-parametric Mann-Whitney U test was used to more accurately determine differences between two groups due to relatively low biological replicates.Significance was set at p < 0.05.Data were presented as mean ± SD.

Laminin-α5 is ablated in endothelial and mural cells in α5-TPKO mice
To generate mutants (α5-TPKO) with laminin-α5 deficiency in both endothelial and mural cells, we crossed laminin-α5 floxed mice with Tie2-Cre and PDGFRβ-Cre lines, which can be distinguished by using promoterspecific Cre primers.The α5-TPKO mice were born at the expected Mendelian ratio.Next, we performed immunohistochemical analysis against laminin-α5 to validate its deletion in α5-TPKO mice.Robust laminin-α5 signal, which co-localized well with vascular marker Podocalyxin, was observed in the cortex of control brains (Fig. 1A).Consistent with previous reports that endothelial and mural cells are the major cellular source of laminin-α5 in the CNS [4,5], markedly reduced laminin-α5 level was found in the cortex of α5-TPKO brains (Fig. 1A).Quantification revealed a significant reduction in laminin-α5 expression in the cortex of α5-TPKO mice compared to that of control mice (Fig. 1B).Similar results were observed in the hippocampus (Fig. 1C) and striatum (Fig. 1D).The residual levels of laminin-α5 in α5-TPKO brains may be from other CNS cells, including neurons and oligodendrocytes, which have been reported to synthesize laminin-α5 [29][30][31][32].These findings suggest that laminin-α5 is successfully ablated in endothelial and mural cells in α5-TPKO brains.Previous studies have shown that ablation of laminin-α2 or -α4 affects other components of the basal lamina [14,15,33].To determine if loss of laminin-α5 affects basal lamina composition, we further examined the expression of three other laminin subunits and two other components of the basal lamina.Interestingly, control and α5-TPKO mice exhibited comparable levels of laminin-α1 (Fig. 2A-D), laminin-α2 (Fig. 2E-G), laminin-α4 (Fig. 2H-J), nidogen (Fig. 2K-M), and collagen IV (Fig. 2N-P) in the cortex, hippocampus, and striatum.These results strongly suggest that loss of laminin-α5 does not affect the expression of other laminin subunits or other basal lamina components.

Neurovascular integrity is disrupted in α5-TPKO mice
To assess neurovascular integrity in α5-TPKO mice, we performed in vivo permeability assay using Sulfo-NHS-Biotin, a small exogenous tracer.Although undetectable in control brains, Sulfo-NHS-Biotin was observed in α5-TPKO brains (Fig. 3A).Quantification revealed substantially increased level of Sulfo-NHS-Biotin in α5-TPKO brains (Fig. 3B).Similarly, Hemoglobin, a large endogenous tracer, was found in α5-TPKO but not control brains (Fig. 3C and D).These results strongly indicate that neurovascular integrity is compromised in α5-TPKO mice.It should be noted that Sulfo-NHS-Biotin and Hemoglobin signals did not show a homogenous/ diffuse pattern in mutant brains.Instead, tracer leakage was only detected in a few specific spots.This sporadic pattern suggests that other factors (e.g.other laminin isoforms/ECM proteins) may compensate for the loss of endothelial and mural laminin-α5 in a region-specific manner.To determine if ablation of laminin-α5 in both endothelial and mural cells affects angiogenesis and/or vascular structure, we performed angioarchitecture analysis using Podocalyxin staining (Fig. 4A).Interestingly, control and α5-TPKO mice displayed comparable total vessel length (Fig. 4B), vessel density (Fig. 4C), branching index (Fig. 4D), and lacunarity (Fig. 4E) in the cortex, hippocampus, and striatum.These findings suggest that loss of laminin-α5 in endothelial and mural cells does not affect brain angioarchitecture, and that increased neurovascular permeability in α5-TPKO mice is not caused by abnormal vascular structure.

Endothelial paracellular mechanism is unaltered in α5-TPKO brains
One unique feature of brain endothelial cells is their expression of tight junction proteins, including ZO-1 and Claudin-5, which seal gaps between adjacent endothelial cells and form tight junctions [1,34,35].To determine if increased neurovascular permeability in α5-TPKO mice is due to this mechanism, we first examined the expression of ZO-1 and Claudin-5 in control and α5-TPKO brains.Immunohistochemical analysis showed that ZO-1 and Claudin-5 co-localized with Podocalyxin in both control and α5-TPKO brains (Fig. 5A).Quantifications revealed comparable ZO-1 coverage and contact in the cortex (Fig. 5B and C), hippocampus (Fig. 5D and E), and striatum (Fig. 5F and G) of control and α5-TPKO mice.Similar to ZO-1, no significant difference in Claudin-5 expression was found in control and α5-TPKO mice at all three brain regions (Fig. 5H-M).These results indicate a negligible role of laminin-α5 in TJP expression.Next, we further investigated the ultrastructure of tight junctions using TEM.Electron-dense tight junctions were observed in both control and α5-TPKO brains (Fig. 5N), highlighting intact tight junctions.Together, these findings suggest that endothelial paracellular mechanism is unaffected in α5-TPKO brains.

Discussion
This work provides novel and significant insights on the functional significance of laminin-α5 in neurovascular integrity.In the CNS, laminin-α5 is mainly produced by endothelial and mural cells.Previous loss-of-function studies failed to detect neurovascular leakage in mutant mice with laminin-α5 deficiency in endothelial cells [22] or mural cells [23] alone.It was speculated that loss of laminin-α5 from one cell type may be compensated by other laminin subunits (e.g.laminin-α4).However, another possibility is that loss of laminin-α5 from one cell type may be compensated by laminin-α5 from other cells.Using a compound mutant mouse line, this study clearly shows that loss of laminin-α5 from endothelial and mural cells simultaneously leads to increased neurovascular permeability, providing the first in vivo evidence that laminin-α5 actively maintains neurovascular integrity.It should be noted that we cannot exclude the possibility of potential compensation by other laminin subunits in α5-TPKO mice.However, such compensation is insufficient to maintain neurovascular integrity when laminin-α5 is ablated in both endothelial and mural cells, highlighting a possibly more important role of laminin-α5 in neurovascular integrity maintenance.These results are consistent with previous studies showing the inhibitory role of laminin-α5 in leukocyte extravasation in vivo [33,42] and barrier function in vitro [43,44].
Subsequent mechanistic study demonstrated that neurovascular leakage in α5-TPKO mice is mainly mediated by the transcellular rather than paracellular mechanism (Figs. 5 and 6).Like α5-TPKO mice, mutants lacking laminin-γ1 (both laminin-411 and − 511) in endothelial cells exhibit increased neurovascular permeability predominantly due to increased transcytosis.In sharp contrast, mutants with laminin-γ1 deficiency in astrocytes (laminin-211) [12,13] and laminin-α2 global knockout mice [14] develop neurovascular leakage with tight junction abnormalities, indicating increased paracellular leakage.Interestingly, mice missing laminin-γ1 in mural cells display age-dependent neurovascular leakage, and both paracellular and transcellular mechanisms are affected in these mutants [10].These results suggest that different laminin isoforms may regulate neurovascular integrity via distinct mechanisms.Understanding the functional significance of each laminin isoform in neurovascular integrity and their molecular mechanisms will fill the gap of knowledge in the field and open doors for new research.
Accumulating evidence supports an indispensable role of pericytes in transcellular leakage.For example, pericyte-deficient mice develop increased endothelial vesicle trafficking without obvious junction defects [38,39,45,46].It has also been reported that pericytes regulates neurovascular permeability partially via Mfsd2a, a negative regulator of caveolae-mediated transcytosis [36,37].Similarly, reduced pericyte coverage increases neurovascular permeability and endothelial transcytosis without affecting TJPs following chronic cerebral hypoperfusion [47].In addition, a recent study shows that pericytes regulate endothelial transcytosis via the vitronectin-integrin-α5 signaling pathway [48].Furthermore, loss of pericyte-derived laminin-γ1 leads to enhanced transcellular leakage and mild neurovascular leakage in an age-dependent manner [10], suggesting that pericytes may regulate neurovascular integrity via secreting laminin-γ1.Consistent with these reports, we found substantially reduced pericyte coverage in α5-TPKO mice, highlighting a critical role of laminin-α5 in pericyte coverage.It remains unclear how laminin-α5 regulates pericyte coverage and neurovascular integrity.One possibility is that laminin-α5 directly signals to pericytes and modulates their biology (e.g.vitronectin and/ or laminin-γ1 expression), which affects transcytosis in endothelial cells, leading to increased neurovascular permeability.Another possibility is that laminin-α5 binds to receptors on endothelial cells to regulate their transcytosis, which affects pericyte coverage indirectly.Alternatively, laminin-α5 may act on both pericytes and endothelial cells.These important questions will be addressed in future research.
One limitation of this study is that the receptors and signaling pathways mediating laminin-α5's effects in endothelial and mural cells remain unknown.Previous studies have shown that endothelial cells mainly express integrin-α3β1, -α6β1, -α6β4, -α5, -αv, and -β3 under physiological conditions; while mural cells predominantly synthesize integrin-α1β1, -α2β1, -α5β1, -α6β1, and -αv [4,5].Among these receptors, integrin-α3β1, -α6β1, and -α6β4 are classical laminin-binding integrin receptors [6,49].Thus, their functional significance in neurovascular integrity maintenance will be first explored in the future.Another limitation of this study lies in the Cre lines used in this study.Given that laminin is very long-lived with an extremely low turnover rate [50][51][52][53] and it takes months or years to degrade laminin synthesized before tamoxifen injection, we are unable to use tamoxifen-inducible CreER lines for laminin ablation.In this study, Tie2-Cre and PDGFRβ-Cre lines were used to ablate laminin-α5 in endothelial cells and mural cells, respectively.Although Tie2-Cre leaks into hematopoietic cells [54,55], hematopoietic cell-derived laminin-α5 is unlikely to regulate neurovascular integrity.First, hematopoietic cells, located in the periphery, are unable to deposit laminin in the neurovascular unit under homeostatic conditions.More importantly, our unpublished data show that hematopoietic cell-derived γ1-containing laminins, which include most α5-containing laminins [6], do not contribute to neurovascular integrity maintenance.In addition to mural cells, PDGFRβ-Cre also labels fibroblasts in the CNS [56][57][58].However, fibroblast-derived laminin-α5 is unlikely to maintain neurovascular integrity, since fibroblasts are found in large blood vessels rather than capillaries in the CNS [56,57].A third limitation of this study is that the compensatory mechanism upon loss of endothelial and mural laminin-α5 is largely known.Answer to this question may explain why tracer leakage in α5-TPKO mice occurs at specific sites rather than in all areas with diminished laminin-α5 expression.

Conclusions
In conclusion, we show that endothelial and mural cellderived laminin-α5 actively maintains neurovascular integrity via the transcellular rather than paracellular mechanism.These findings suggest that laminin-α5 signaling and/or transcytosis-associated molecules may be targeted to modulate neurovascular integrity.

Fig. 4
Fig. 4 Angioarchitecture analysis for cortex, hippocampus and striatum regions.A Representative images of Podocalyxin staining in the cortex of 4 ∼ 6-month-old control and α5-TPKO mice (both sexes).Scale bar: 100 μm.B-E Quantifications of total vessel length (B), vessel density (C), branching index (D), and lacunarity (E) in the cortex, hippocampus, and striatum of control and α5-TPKO mice.Non-significant, Mann-Whitney U test.n = 5 mice.Data were shown as mean ± SD

Fig. 5
Fig. 5 Endothelial paracellular mechanism is unaltered in α5-TPKO mice.A Representative images of ZO-1 (green), Claudin-5 (red), and Podocalyxin (magenta) staining in the cortex of 4 ∼ 6-month-old control and α5-TPKO mice (both sexes).Scale bar: 50 μm.B-G Quantifications of ZO-1 coverage (B, D, and F) and contact (C, E, and G) in the cortex (B and C), hippocampus (D and E), and striatum (F and G) of control and α5-TPKO mice.Non-significant, Mann-Whitney U test.n = 5 mice.Data were shown as mean ± SD.H-M Quantifications of Claudin-5 coverage (H, J, and L) and contact (I, K, and M) in the cortex (H and I), hippocampus (J and K), and striatum (L and M) of control and α5-TPKO mice.Non-significant, Mann-Whitney U test.n = 5 mice.Data were shown as mean ± SD.N Representative TEM images showing endothelial tight junctions in 4 ∼ 6-month-old control and α5-TPKO mice (both sexes).Arrows indicate tight junctions.PC, pericyte; EC, endothelial cell; Scale bar: 500 nm

Fig. 7
Fig. 7 Pericyte coverage but not astrocyte polarity is diminished in α5-TPKO mice.A Representative images of PDGFRβ (green) and Podocalyxin (red) staining in the cortex of 4 ∼ 6-month-old control and α5-TPKO mice (both sexes).Scale bar: 25 μm.B-G Quantifications of pericyte coverage (B, D, and F) and contact (C, E, and G) in the cortex (B and C), hippocampus (D and E), and striatum (F and G) of control and α5-TPKO mice.*p = 0.0159 and **p = 0.0079, Mann-Whitney U test.n = 5 mice.Data were shown as mean ± SD.H Representative images of AQP-4 (green) and Podocalyxin (red) staining in the cortex of 4 ∼ 6-month-old control and α5-TPKO mice (both sexes).Scale bar: 50 μm.I-N Quantifications of AQP-4 coverage (I, K, and M) and contact (J, L, and N) in the cortex (I and J), hippocampus (K and L), and striatum (M and N) of control and α5-TPKO mice.non-significant, Mann-Whitney U test.n = 5 mice.Data were shown as mean ± SD