Blood–brain barrier genetic disruption leads to protective barrier formation at the Glia Limitans

Pierre Mora; Pierre-Louis Hollier; Sarah Guimbal; Alice Abelanet; Aïssata Diop; Lauriane Cornuault; Thierry Couffinhal; Sam Horng; Alain-Pierre Gadeau; Marie-Ange Renault; Candice Chapouly

doi:10.1371/journal.pbio.3000946

. 2020 Nov 30;18(11):e3000946. doi: 10.1371/journal.pbio.3000946

Blood–brain barrier genetic disruption leads to protective barrier formation at the Glia Limitans

Pierre Mora ^1,^#, Pierre-Louis Hollier ^1,^#, Sarah Guimbal ^1,^#, Alice Abelanet ¹, Aïssata Diop ¹, Lauriane Cornuault ¹, Thierry Couffinhal ¹, Sam Horng ², Alain-Pierre Gadeau ¹, Marie-Ange Renault ¹, Candice Chapouly ^1,^*

Editor: Richard Daneman³

PMCID: PMC7728400 PMID: 33253145

Abstract

Inflammation of the central nervous system (CNS) induces endothelial blood–brain barrier (BBB) opening as well as the formation of a tight junction barrier between reactive astrocytes at the Glia Limitans. We hypothesized that the CNS parenchyma may acquire protection from the reactive astrocytic Glia Limitans not only during neuroinflammation but also when BBB integrity is compromised in the resting state. Previous studies found that astrocyte-derived Sonic hedgehog (SHH) stabilizes the BBB during CNS inflammatory disease, while endothelial-derived desert hedgehog (DHH) is expressed at the BBB under resting conditions. Here, we investigated the effects of endothelial Dhh on the integrity of the BBB and Glia Limitans. We first characterized DHH expression within endothelial cells at the BBB, then demonstrated that DHH is down-regulated during experimental autoimmune encephalomyelitis (EAE). Using a mouse model in which endothelial Dhh is inducibly deleted, we found that endothelial Dhh both opens the BBB via the modulation of forkhead box O1 (FoxO1) transcriptional activity and induces a tight junctional barrier at the Glia Limitans. We confirmed the relevance of this glial barrier system in human multiple sclerosis active lesions. These results provide evidence for the novel concept of “chronic neuroinflammatory tolerance” in which BBB opening in the resting state is sufficient to stimulate a protective barrier at the Glia Limitans that limits the severity of subsequent neuroinflammatory disease. In summary, genetic disruption of the BBB generates endothelial signals that drive the formation under resting conditions of a secondary barrier at the Glia Limitans with protective effects against subsequent CNS inflammation. The concept of a reciprocally regulated CNS double barrier system has implications for treatment strategies in both the acute and chronic phases of multiple sclerosis pathophysiology.

This study uncovers a critical role for the signaling molecule desert hedgehog (Dhh) in maintaining tightness of the blood-brain barrier and highlights a reciprocally regulated double barrier system in the central nervous system that relies on crosstalk between endothelial cells and astrocytes.

Introduction

In a healthy individual, the central nervous system (CNS) parenchyma is protected from the peripheral circulation by the blood–brain barrier (BBB), which tightly regulates the entry and exit of soluble factors and immune cells [1]. Importantly, during multiple sclerosis, the abnormal permeability of the BBB allows penetration into the CNS parenchyma of inflammatory cells and soluble factors such as autoantibodies, cytokines, and toxic plasma proteins, which drive lesion formation and acute disease exacerbation [2,3]. Therefore, identifying key mechanisms that promote BBB tightness is currently considered to be a main strategy for controlling leukocyte and humoral entry, preventing acute relapse and disability progression in multiple sclerosis.

Previous studies have identified the Hedgehog (HH) pathway as a regulator of BBB integrity in multiple sclerosis, HIV, and stroke [4–7]. Desert hedgehog (DHH) is expressed constitutively at the BBB in adults [8] and belongs, together with Sonic hedgehog (SHH) and Indian hedgehog (IHH), to the HH family of morphogens, identified nearly 4 decades ago in Drosophila as crucial regulators of cell fate determination during embryogenesis [9]. The interaction of HH proteins with their specific receptor Patched-1 (PTCH1) derepresses the transmembrane protein Smoothened (SMO), which activates downstream pathways including the canonical HH pathway leading to the activation of Gli family zinc finger (Gli) transcription factors, and the so-called noncanonical HH pathways, which are independent of SMO and/or Gli [10].

Interestingly, a wealth of literature published during the last decades has enabled a change in the vision of BBB structure and integrity, which has expanded to include contributions from both barrier properties of the vascular endothelial cells and the astrocytic end feet of the neurovascular unit. Within the neurovascular unit, substantial intercellular communication network involves the vascular endothelial cells and astrocytic end feet, as well as the pericytes and basement membranes within the perivascular space (PVS) [11–13]. How these signals regulate the passage of soluble factors and cells into and out of the CNS is not completely understood and is of considerable translational interest to the field of neuroimmunology. Regulatory mechanisms at the BBB include solute transporters and receptor-mediated transcytosis, and immune cells are actively prevented from crossing the BBB by low levels of immune receptors that normally permit immune trafficking. Once soluble factors and immune cells penetrate the BBB, they circulate within the PVS, a region between the basal basement membrane of the endothelial cell wall and the parenchymal basement membrane abutting the astrocyte end feet [14,15]. While it is now well established that BBB breakdown leads to soluble factor and inflammatory cell infiltration into the PVS during neuropathology, the role of the Glia Limitans is more complex. Indeed, astrocytes, described as reactive, may demonstrate opposing roles in both recruiting and restricting neuroinflammatory infiltration depending on the context [16]. Specific reactive astrocyte behaviors are likely determined by signaling events that vary with the nature and severity of CNS injury or disease. Specifically, in multiple sclerosis as well as Alzheimer’s and Parkinson diseases, it has been shown that reactive astrocytes, on one hand, produce pro-inflammatory and pro-permeability factors and on the other hand, neuroprotective factors [17–19].

Astrocyte barrier properties are not as well characterized as those of the BBB. However, several groups have highlighted barrier properties at the Glia Limitans [20–22]. Notably, endfoot–endfoot clefts, similar to those observed between endothelial cells at the BBB, have been described at the Glia Limitans and shown to be responsible for the sieving effect observed between the distribution of small and large Dextrans [23]. Moreover, under neuroinflammatory conditions, immune cell trafficking across the Glia Limitans is necessary for clinical experimental autoimmune encephalomyelitis (EAE) [21]; indeed, matrix metalloproteinase (MMP)-2 and MMP-9 proteolytically cleave dystroglycan, which anchors astrocyte end feet to the Glia Limitans basement membrane via binding to extracellular matrix molecules, allowing infiltrating leukocytes to penetrate the parenchyma [24]. Additionally, it has been shown that the scavenger receptor CXCR7 is up-regulated on the inflamed BBB endothelium, facilitating mobilization of T cells from the PVS into the CNS parenchyma [25]. Altogether, these data demonstrate that both the endothelial BBB and its basement membrane, along with the Glia Limitans and the parenchymal basement membrane, are required for immune cell trafficking across the neurovascular unit. Strikingly, our recent work has given considerable attention to a new property of reactive astrocytes: the expression of tight junction proteins (notably Claudin4 (CLDN4)) under inflammatory conditions [26]. This result provides yet another argument in favor of astrocytic barrier properties.

The first objective of our study was to decipher the role of the morphogen DHH in maintaining BBB tightness. The second objective was to demonstrate that a double barrier system comprising both the BBB and Glia Limitans is implemented in the CNS and regulated by a crosstalk going from endothelial cell to astrocytes using endothelial Dhh knockdown as a model of permeable BBB.

Here, we first demonstrate that endothelial DHH expression is down-regulated during neuroinflammation and is necessary to maintain BBB tightness. We then show that BBB opening, induced by Dhh knockdown, drives astrocyte CLDN4 expression, conferring barrier properties to the Glia Limitans, which results in the PVS entrapment of plasma proteins and inflammatory cells, both under physiological conditions and during pathology. Together, these data identify the neurovascular unit as a double barrier system whose function is controlled by the crosstalk between endothelial cells and astrocytes.

In conclusion, this work strengthens the concept of CNS double barrier system, unveiling how signals at the endothelium drive astrocyte barrier properties to protect the parenchyma during neuropathology. Consequently, taking into account both components of the neurovascular unit is of translational interest and could open the way for new therapeutic strategies notably to limit progressive multiple sclerosis pathology.

Results

DHH, but not SHH or IHH, is expressed by CNS microvascular endothelial cells and down-regulated during chronic neuroinflammation

First, we showed that DHH is expressed at the BBB in vitro using mouse CNS MECs* (Fig 1A). CNS MEC purity was assessed using platelet/endothelial cell adhesion molecule 1 (PECAM1) and zonula occludens 1 (ZO1) as positive endothelial markers and smooth muscle actin (SMA), neural/glia antigen 2 (NG2), cluster of differentiation 45 (CD45), ionized calcium binding adaptor molecule 1 (IBA1), and glial fibrillary acidic protein (GFAP) as markers of contamination by smooth muscle cells, pericytes, leukocytes, microglia, and astrocytes, respectively (Fig 1A and S1A Fig). DHH expression at the BBB was verified in vivo using human cortical sections from healthy donors (Fig 1B) and brain sections from C57BL/6 mice (S1B Fig). SHH and IHH are not expressed in the healthy BBB, and DHH is known to be stored intracellularly as well as being secreted [7]. Therefore, we here infer the CNS endothelial cells as the source of DHH within the neurovascular unit. Next, we demonstrated that Dhh is severely down-regulated at the BBB under inflammatory conditions both in vitro using human brain microvascular endothelial cells (HBMECs) treated with interleukin-1β (IL-1β) (one of the main pro-inflammatory cytokine implicated in multiple sclerosis pathophysiology) (Fig 1C and S1 Data) and in vivo (Fig 1F and S1 Data) using a preclinical model of multiple sclerosis (MOG_35-55) to induce chronic neuroinflammation in C57BL/6 mice. For this experiment, isolated spinal cord microvessels underwent a digestion step followed by a CD45⁺ T cell depletion step to discard inflammatory cell contamination induced by EAE (S1C and S1D Fig). Dhh down-regulation at the BBB is associated with the up-regulation of endothelial activation markers intercellular adhesion molecule 1 (Icam1) (Fig 1D–1G and S1 Data) and vascular cell adhesion molecule 1 (Vcam1) (Fig 1E–1H and S1 Data) and with down-regulation of mRNA markers of tight junctions (claudin5 (Cldn5) and Zo1) (Fig 1I, 1J and S1 Data).

Together, these data identify DHH as the only HH expressed in adults at the endothelial BBB. Moreover, they highlight the fact that DHH expression is down-regulated at the BBB during neuroinflammatory pathology.

Endothelial-specific Dhh inactivation induces down-regulation of adherens junction CDH5 and tight junction CLDN5 ex vivo

To test the importance of endothelial DHH expression at the BBB, we conditionally disrupted DHH expression in endothelial cells and examined the consequences on BBB integrity. To do so, we used CNS MEC cultures isolated from Cdh5-Cre^ERT2, Dhh^Flox/Flox mice (Dhh^ECKO mice), and Dhh^Flox/Flox control littermates, 2 weeks after inducing knockdown by intraperitoneal injection of tamoxifen. There is no difference between Dhh^ECKO and control mouse cell culture viability (S1E, S1F Fig and S2 Data).

We first verified the efficiency of the knockout by measuring Dhh expression in primary CNS MEC cultures obtained from Dhh^ECKO and littermate controls and showed that Dhh expression is strongly down-regulated in the knockout mice (Fig 2A and S1 Data). Moreover, CDH5, CLDN5, and ZO1 junctions are disorganized in Dhh^ECKO endothelial cells: In controls, CDH5, CLDN5, and ZO1 display a well-defined pattern of sharp contours at endothelial cell–cell contacts. In contrast, in Dhh^ECKO cultures, a broader, more irregular pattern is detected at endothelial cell–cell contacts (Fig 2D–2F). This result is consistent with the previously documented phenotype using small interfering RNA (siRNA) for DHH [8]. CDH5 and CLDN5 but not ZO1 are down-regulated in Dhh^ECKO CNS MECs compared to controls (Fig 2B, 2C and 2G–2I and S1 Data).

**(A–C)** CNS MECs were isolated from Dhh^ECKO and control mice, and **(A)** *Dhh*, **(B)** *Cldn5*, and **(C)** *Zo1* expressions were quantified by qRT-PCR. **(D–F)** Primary BMECs from Dhh^ECKO and control mice were isolated and cultured on Lab-Tek. **(D)** CDH5 (in green), **(E)** ZO1 (in green), and **(F)** CLDN5 (in red) localizations were evaluated by immunofluorescent staining of a confluent cell monolayer. Nuclei were stained with DAPI (in blue). The experiment was repeated 3×. **(G–L)** CNS MECs were isolated from Dhh^ECKO and control mice, seeded and cultured until confluency, and **(G, H)** CDH5, **(G, I)** ZO1, **(G, J)** FOXO1, **(G, K)** and p-FOXO1 expression were quantified by western blot. **(L)** BMECs were isolated from Dhh^ECKO mice and control mice, seeded and cultured until confluency and starved for 24 h. Control BMECs were then treated with DMSO and Dhh^ECKO BMECs with DMSO versus an inhibitor of FOXO1 (AS1842856). **(L)** CLDN5 expression was then quantified by qRT-PCR and **(M)** CLDN5 (in red) localization was evaluated by immunofluorescent staining of a confluent cell monolayer in Dhh^ECKO BMECs treated with DMSO versus an inhibitor of FOXO1 (AS1842856). Nuclei were stained with DAPI (in blue). The experiment was repeated 3×. **(N)** Summary outline of the up-regulation of endothelial junctions by the morphogen DHH through inhibition of FOXO1 transcriptional activity. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 Mann–Whitney U test. *P ≤ 0.05 Kruskal–Wallis test. The underlying data for Fig 2 can be found in S1 Data (https://doi.org/10.6084/m9.figshare.12625034.v6). *It is important to note that CNS endothelial cells are from a pooled source including both brain and spinal cord tissues. Therefore, the resulting cell cultures/lysates may be heterogeneous in their use of DHH. This remark applies to Figs 1–4. AKT, serine/threonine kinase 1; BMEC, brain microvascular endothelial cell; CDH5, cadherin5; Cldn5, claudin5; CNS, central nervous system; CTNNβ1, catenin beta 1; Ctrl, control; DHH desert hedgehog; FOXO1, forkhead box O1; MEC, microvascular endothelial cells; NS, non-significant; PTCH1, Patched-1; qRT-PCR, quantitative reverse transcription polymerase chain reaction; ZO1, zonula occludens 1.

We concluded that endothelial DHH expression is necessary to maintain endothelial adherens and tight junction mRNA and protein expression level at the BBB and to maintain a well-defined CDH5, ZO1, and CLDN5 pattern of sharp contours at endothelial cell–cell contacts.

DHH induces CLDN5 up regulation through the inhibition of FOXO1 transcriptional activity in vitro

We previously identified DHH as a factor facilitating the interaction between CDH5 and catenin β1 (CTNNβ1) in endothelial cells [8], and others have shown that CDH5 interacts with CTNNβ1 to inhibit transcription factor forkhead box O1 (FOXO1) [27] via PI(3)K–AKT-dependent phosphorylation, which thereby up-regulates the expression of the endothelial tight junction protein CLDN5 [27]. Therefore, we next measured the expression level of the phosphorylated form of FOXO1 (p-FOXO1) in CNS MEC cultures from Dhh^ECKO and control (Dhh^Flox/Flox) mice and demonstrated that p-FOXO1 is down-regulated in Dhh^ECKO mice compared to controls (Fig 2G, 2J–2K). We then treated CNS MEC cultures from Dhh^ECKO mice with a cell permeable inhibitor of the transcription factor FOXO1 (AS1842856), which blocks the transcription activity of FOXO1, and measured Cldn5 mRNA expression. We demonstrated that Cldn5 mRNA expression, in Dhh^ECKO CNS MEC cultures treated with the inhibitor of FOXO1, returns to the expression level of control CNS MEC cultures, unlike the Dhh^ECKO CNS MEC cultures treated with DMSO (Fig 2L and S1 Data). Additionally, in Dhh^ECKO CNS MEC cultures treated with the inhibitor of FOXO1, CLDN5 displays a well-defined pattern of sharp contours at endothelial cell–cell contacts. In contrast, in Dhh^ECKO CNS MEC cultures treated with DMSO, a broader, more irregular pattern is detected at endothelial cell–cell contacts (Fig 2M).

We concluded that endothelial autocrine DHH expression at the BBB maintains the pool of CDH5-CTNNβ1 signaling in endothelial cells, which promotes Cldn5 mRNA expression and maintains a well-defined CLDN5 pattern of sharp contours at endothelial cell–cell contacts through the inhibitory phosphorylation of the transcription factor FOXO1 (Fig 2N).

In the white matter, endothelial-specific Dhh inactivation induces BBB permeability associated with endothelial and astrocytic activation in vivo

In vivo, on spinal cord sections from control and Dhh^ECKO mice, we confirmed that the expression of adherens junction CDH5 and tight junction CLDN5, when normalized for the number and length of blood vessels, is down-regulated under resting conditions (Fig 3A–3D, S1 Data and S3A Fig) and demonstrated that it is associated with an increase accumulation of serum proteins (fibrinogen (FGB) and albumin (ALB)) [28] (Fig 3B, 3E, 3F, S1 Data, S3B and S3C Fig) around the vessels, suggesting BBB opening.

We then analyzed the activation status of both the spinal cord endothelium and Glia Limitans in Dhh^ECKO and control littermates. We chose ICAM1 as a marker of endothelial activation and GFAP as a marker of astrocyte reactivity because they are both widely published in the context of inflamed CNS tissues and represent a strong indicator of a reactive response of the CNS endothelium and Glia Limitans.

Using spinal cord sections, we revealed that ICAM1 is up-regulated at the endothelium in Dhh^ECKO mice compared to littermate controls (Fig 3G, 3I and S1 Data) and associated with a regionalized up-regulation of GFAP, a marker of astrocyte activation, in white matter (Fig 3H, 3J and S1 Data) but not gray mater (Fig 3H, 3K and S1 Data).

Here, we detected that, in the white matter, DHH expression at the endothelium controls BBB tightness and demonstrated that endothelial-specific Dhh inactivation at the BBB drives endothelial and astrocyte activation.

Dhh^ECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans

As we already demonstrated in Fig 3, Dhh^ECKO mice display BBB leakage, whereas control littermates feature a tight BBB.

Although we demonstrated BBB leakage in Dhh^ECKO mice (Fig 3B, 3E, 3F and S1 Data), we noticed that infiltrating plasmatic proteins are concentrated around the vascular area in arterioles and venules and not seamlessly distributed within the parenchyma**. To verify this observation, we quantified the distribution of immunoglobulin G (IgG) and a smaller sized dye (70 kDa fluorescein isothiocyanate (FITC) Dextran) in the 3 compartments (lumen, PVS, and parenchyma). Specifically, in control mice, there is no significant endothelial permeability, with more than 95% of IgG and 97% of 70 kDa FITC Dextran contained in the lumen of blood vessels and 5% of IgG and 3% of 70 kDa FITC Dextran segregated in the PVS area limited by the astrocytic end feet (aquaporin 4 (AQP4) or laminin (LAM) antigen) on 1 side and the vessel wall (PECAM1 or LAM antigen) on the other side, and none found in the parenchyma (S1 Data, S4A–S4E Fig and S2 Data). The quantification protocol of IgG and 70 kDa FITC Dextran distribution within the lumen, PVS, and parenchyma is described in S4 Fig. In Dhh^ECKO mice, vascular leakage is significant, but strikingly, 50% of IgG and 50% of 70 kDa FITC Dextran is contained into the PVS, while none is found in the parenchyma, indicating the presence of a secondary barrier at the Glia Limitans (S1 Data, S4A–S4E Fig and S2 Data). Interestingly, some overlap is observed between the IgG signal and the AQP4 signal, concentrated within the internal surface of the Glia Limitans (Fig 4A). Thus might reflect a potential interaction between the inner face of astrocyte end feet and IgG accumulated in the PVS, as astrocytes express Fc receptors (cell surface receptors for IgG), which play a role in both CNS health and disease [29,30]. The perivascular trapping of IgG in Dhh^ECKO mouse CNS was confirmed using co-immunostaining of IgG and LAM antigen, which mark the endothelial and astrocytic basement membranes (S5A and S5B Fig).

**(A–C)** Spinal cord sections were harvested from Dhh^ECKO mice and littermate controls and **(A)** immunostained with anti-PECAM1 (in gray), anti-AQP4 (in green), and anti-IgG (in red) antibodies (nuclei were stained with DAPI (in blue)), and **(B, C)** the distribution of IgG within the lumen, PVS, and parenchyma was quantified. We quantified the amount of IgG in the PVS and parenchyma and inferred from it the amount of IgG contained in the lumen. (Dhh^ECKO n = 6, WT n = 6). P < 0.0001, 95% CI: 0.3281 to 0.5558, chi-squared test. **(D)** Human active multiple sclerosis cortical lesions were obtained from the NeuroCEB biobank and immunostained with anti-ZO1 (in gray), anti-GFAP (in red), and anti-CLDN4 (in green) antibodies. Strong CLDN4 signal at the Glia Limitans is indicated by white arrows. Nuclei were stained with DAPI (in blue). **(E–H)** Enriched neurovascular fractions were isolated from Dhh^ECKO and control mouse CNS and **(E)** immunostained with anti-CDH5 (in gray), anti-CLDN4 (in red), and anti-GFAP (in green) antibodies (nuclei were stained with DAPI (in blue)). **(F)** Representative blots of GFAP and CLDN4 expression levels on Dhh^ECKO and control mouse enriched neurovascular fractions were shown, and **(G, H)** GFAP and CLDN4 expression levels were quantified by western blot of Dhh^ECKO and control mouse enriched neurovascular fractions. (Dhh^ECKO n = 5, WT n = 5). **P ≤ 0.01, Mann–Whitney U test. The underlying data for Fig 4 can be found in S1 Data (https://doi.org/10.6084/m9.figshare.12625034.v6). *It is important to note that CNS endothelial cells are from a pooled source including both brain and spinal cord tissues. Therefore, the resulting cell cultures/lysates may be heterogeneous in their use of DHH. This remark applies to Figs 1–4. AQP4, aquaporin 4; BBB, blood–brain barrier; CDH5, cadherin5; CI, confidence interval; CLDN4, Claudin4; CNS, central nervous system; Ctrl, control; DHH desert hedgehog; GFAP, glial fibrillary acidic protein; IgG, immunoglobulin G; NS, non-significant; PECAM1, platelet/endothelial cell adhesion molecule 1; PVS, perivascular space; WT, wild type; ZO1, zonula occludens 1.

We have previously found that reactive astrocytes express tight junctions, notably CLDN4, under inflammatory conditions in a mouse model of multiple sclerosis (EAE) [26]. Here, we found that these data are relevant to human disease since CLDN4 is also expressed by reactive astrocytes, with stronger CLDN4 labeling intensity at the Glia Limitans, in active cortical lesions from multiple sclerosis patients (Fig 4D). Based on the above results, we investigated whether the PVS entrapment of plasmatic proteins observed in Dhh^ECKO mice is linked to the expression of the tight junction CLDN4 at the Glia Limitans in response to BBB permeability (Fig 4E–4H and S1 Data). We showed that CLDN4 is expressed at the Glia Limitans in Dhh^ECKO mice but not control littermates (Fig 4E–4H and S1 Data) using isolated neurovascular enriched fractions. Small intestine samples were used as a positive control for the quantification of CLDN4 by western blot (S6A Fig).

Altogether, these results suggest that, in Dhh^ECKO mice, spontaneous BBB permeability leads to the establishment of a physical barrier at the Glia Limitans, characterized by the expression of the tight junction protein CLDN4. Therefore, in Dhh^ECKO mice, astrocytic end feet at the Glia Limitans are “preconditioned” to form a secondary barrier protecting the parenchyma. (**It is important to note that this study focuses on arterioles and venules but not capillaries. Indeed, CLDN4 is only up-regulated in Dhh^ECKO-enriched neurovascular fractions, which are 100 μm and larger in diameter (Fig 4F–4H and S1 Data). In the lysates obtained with enriched neurovascular fractions, which are 20 μm and larger, CLDN4 expression level is not different between the Dhh^ECKO and control group (S6B–S6D Fig and S2 Data). In this condition, CLDN4 expression in arterioles and venules might have been lost in the crowd of capillaries.)

Endothelial signals can drive astrocyte barrier properties at the Glia Limitans

Given the above results, we wanted to determine if astrocyte barrier formation requires signals from the endothelial BBB or from the plasmatic protein perivascular infiltrate. To do so, we first studied in vitro the response of normal human astrocytes (NHAs) to HBMEC conditioned media versus plasmatic proteins from healthy donors. HBMECs used to produce the conditioned media were treated with either the osmotic agent Mannitol or the pro-permeability factor vascular endothelial growth factor A (VEGFA) to induce BBB breakdown (S7A Fig and S2 Data) through various methods.

We demonstrated that Gfap (Fig 5A, 5E–5I and S1 Data), aldehyde dehydrogenase 1 family, member l1 (Aldh1l1) (Fig 5B and S1 Data) (markers of astrocyte reactivity), and Cldn4 mRNA expression (Fig 5D, 5A–5H, 5J and S1 Data) are up-regulated in the NHA treated with HBMEC-conditioned media but not in the NHA treated with plasma from healthy donors. vimentin (Vim) (marker of astrocyte reactivity) mRNA expression level was not modulated in any condition (Fig 5C and S1 Data).

To confirm this observation in vivo, we delivered murine VEGFA or murine plasmatic proteins into the left cerebral cortex of adult mice and evaluated the consequences on CLDN4 expression by astrocytes. PBS stereotactic administration was used as a control. Importantly, VEGFA cortical stereotactic injection has already been shown to efficiently induce BBB breakdown in mice [17,31].

In vivo, GFAP (Fig 5K–5N and S1 Data) and CLDN4 (Fig 5K–5M, 5O and S1 Data) are induced in mouse cortex after murine VEGFA and murine plasmatic protein treatments, with VEGFA having a much stronger effect than plasmatic proteins.

We concluded that permeable endothelial monolayers produce signals that can drive astrocyte reactivity and tight junction expression. Plasmatic protein involvement in controlling astrocyte barrier behavior is, however, less clear as astrocyte reactivity and tight junction expression are up-regulated in vivo but not in vitro when treated with plasma; further investigations will be necessary to identify the mechanisms involved.

Mice with endothelial Dhh inactivation display reduced disability in a model of multiple sclerosis during the onset of the disease

To examine the impact of these findings on disease severity, we investigated the phenotype of induced experimental multiple sclerosis (EAE) in Dhh^ECKO and control mice.

We observed that in control mice, neurologic deficits were observed from day 9 and increased in severity until day 18, when clinical score stabilized at a mean of 3.2, representing hind limb paralysis. In contrast, the onset of clinical signs in Dhh^ECKO mice was first seen 4 days later, and the clinical course was much milder. In Dhh^ECKO mice, disease reached a plateau at day 21 at a mean of 2.3, indicating hind limb weakness and unsteady gait, a mild phenotype (Fig 6A and S1 Data). The EAE peak score (Fig 6B and S1 Data) and average score during the time of disability (Fig 6C and S1 Data) were both decreased in Dhh^ECKO mice, but there were no significant changes in survival and mortality rates (Fig 6D and S1 Data).

**(A)** Dhh^ECKO and control mice induced with EAE were scored daily on a standard 5-point scale, nonlinear regression (Boltzmann sigmoidal). (Dhh^ECKO n = 8, WT n = 10). **(B)** Dhh^ECKO and control mice EAE peak score, **(C)** EAE average score during time of disability, and **(D)** mortality rate were quantified other the course of the disease. **(E, F)** Spinal cord EAE lesions from Dhh^ECKO mice and littermate controls were harvested at 28 days post induction or at the time of euthanasia. **(E)** Dhh^ECKO and control lesions were immunostained with an anti-MBP (in green) antibody; arrows indicate white matter loss areas. Nuclei were stained with DAPI (in blue). **(F)** MBP positive areas were quantified ((Dhh^ECKO n = 5, WT n = 5). *P ≤ 0.01, Mann–Whitney U test. The underlying data for Fig 6 can be found in S1 Data (https://doi.org/10.6084/m9.figshare.12625034.v6). AU, arbitrary units; Ctrl, control; DHH desert hedgehog; EAE, experimental autoimmune encephalomyelitis; MBP, myelin basic protein; NS, non-significant; WT, wild type.

Interestingly, the reduced clinical course of EAE in Dhh^ECKO mice was much more marked during the onset of the disease (between day 12 and day 20 post EAE induction). However, when the plateau phase was reached (after day 21 post EAE induction), the clinical score difference between the Dhh^ECKO and the control group was greatly reduced and coincided with an acceleration of the mortality rate in both groups (Fig 6A–6D and S1 Data). One explanation for this observation is that, in the Dhh^ECKO group, inflammatory cells, accumulated in the PVS, end up degrading astrocytic tight junctions by secreting proteases. This previously described phenomenon [26] explains how inflammatory cells trapped in the PVS can eventually pass through the astrocytic barrier at the Glia Limitans and thereby enter the CNS, thus inactivating Dhh at the BBB. This in turn slows disease progression until the perivascular accumulation of inflammatory cells causes the degradation of the astrocyte secondary barrier, leading to the deterioration of the clinical course of EAE in the Dhh^ECKO mice.

Importantly, the clinical course in the Dhh^ECKO mice is correlated with strikingly decreased areas of demyelination as compared to the control cohort (Fig 6E, 6F and S1 Data). Critically, these studies reveal that the clinical course and pathology of EAE are strongly reduced in Dhh^ECKO mice during the onset of the disease.

We concluded that endothelial Dhh knockdown-induced BBB opening is associated with a clinical protective effect during the onset of the disease in a model of multiple sclerosis.

Mice with endothelial Dhh inactivation display a reinforced barrier at the Glia Limitans, restraining access to the parenchyma to inflammatory infiltrate in a model of multiple sclerosis

Although Dhh^ECKO mice display equivalent FGB densities (Fig 7A and S1 Data) as well as numbers of CD45+ leukocytes (Fig 7B and S1 Data) in lesions compared to those in control mice, neuropathology in both cohorts appeared very different.

We found that while the BBB is permeable in both groups, with plasmatic protein extravasation associated with equivalent CDH5 densities, astrocyte reactivity in EAE lesions in Dhh^ECKO mice is greatly increased with GFAP immunoreactivity strongest at the Glia Limitans (Fig 7C, 7E and S1 Data). Moreover, infiltrating plasmatic proteins in Dhh^ECKO mice show less CNS parenchymal dispersion (Fig 7E, 7F and S1 Data), with 68.0% of FGB trapped into the Glia Limitans in the Dhh^ECKO cohort versus 32% in the control cohort (Fig 7F, S7B Fig and S1 Data).

In previous work from our laboratory [8], we showed that endothelial-specific deletion of Dhh in the peripheral vasculature is associated with vascular permeability and endothelial activation, notably in the lung, and that lipopolysaccharide (LPS) injection increased pulmonary neutrophil infiltration in Dhh^ECKO mice compared to control littermates. Therefore, we can hypothesize that endothelial Dhh knockdown increases the peripheral recruitment and activation of inflammatory cells that need to travel to the CNS. However, in our study, we do not observe any difference in terms of CD45+ leukocyte populations in the spinal cord of Dhh^ECKO and control mice induced with EAE MOG_35-55. What we do observe is a significant difference in terms of the repartition of these CD45+ cell populations between both groups, with 77.1% of CD45+ cells trapped into the PVS in the Dhh^ECKO cohort versus 25.1% in the control cohort (Fig 7G, 7H, S1 Data and S7C Fig).

Altogether, these data suggest less access through the Glia Limitans in the Dhh^ECKO mice compared to the littermate control mice. Finally, we demonstrated that CLDN4 expression in spinal cord EAE lesion lysates is up-regulated in Dhh^ECKO mice as compared to control mice (Fig 7I, 7J and S1 Data).

Collectively, data from Fig 4 to Fig 7 reveal that conditional loss of a key structural component of endothelial integrity at the BBB in Dhh^ECKO mice leads to increased astrocyte reactivity and implementation of barrier properties at the Glia Limitans, allowing for less diffusion of plasmatic proteins and immune cells into the CNS parenchyma than in control mice. Therefore, in Dhh^ECKO mice, astrocytic end feet at the Glia Limitans are “preconditioned” to form a barrier, explaining their ability to protect the parenchyma more efficiently during neuropathology than in controls, leading to the protective effect observed clinically. Thus, we identify BBB leakage, induced by the down-regulation of Dhh endothelial expression, as an important mechanism controlling Glia Limitans reactivity and barrier properties, and subsequently, tissue damage and clinical deficits in a model of human disease (Fig 8).

Fig 8 — Under resting conditions, in control mice, ECs express CLDN5 and CDH5, which reinforce a closed BBB. In Dhh^ECKO mice, CLDN5 and CDH5 are disrupted leading to BBB permeability; in turn, astrocytes of the Glia Limitans up-regulate CLDN4, closing the Glia Limitans and restricting incoming plasmatic proteins to the PVS. Under inflammatory condition, in both control and Dhh^ECKO mice, CLDN5 and CDH5 are disrupted leading to BBB permeability, astrocyte reactivity, and up-regulation of CLDN4. However, in Dhh^ECKO mice, astrocytic end feet of the Glia Limitans have been “preconditioned” to form a barrier, explaining their ability to trap plasmatic proteins and inflammatory cells in the PVS, and thus protecting the parenchyma more efficiently than in the controls. BBB, blood–brain barrier; CDH5, cadherin 5; CLDN4, Claudin4; CLDN5, claudin5; CTNNβ1, catenin β1; DHH desert hedgehog; EC, endothelial cells; FOXO1, forkhead box O1; PTCH1, Patched-1; PVS, perivascular space.

Discussion

While it is now well established that BBB breakdown leads to soluble factor and inflammatory cell infiltration into the PVS during neuropathology [4], the function of the Glia Limitans barrier is just starting to be unraveled [15,20,24,26]. In the present study, we have enabled a different perspective on CNS barrier organization, unveiling the existence of 2 independent, dissociable states of the astrocyte and endothelial barriers in the neurovascular unit. Indeed, we confirmed that, just like the BBB, the Glia Limitans can form a protective barrier. For the first time, to our knowledge, we have demonstrated that BBB breakdown is sufficient to induce chronic barrier properties at the Glia Limitans, and we uncovered crosstalk from endothelial cells to astrocytes that restricts access to the parenchyma of plasmatic proteins and inflammatory cells during multiple sclerosis. Moreover, we showed that in Dhh^ECKO mice, which display an open BBB, astrocytes express the tight junction protein CLDN4 under resting conditions. Therefore, under neuroinflammatory conditions, Glia Limitans in Dhh^ECKO mice is primed with stronger barrier properties protecting against the onset and severity of EAE symptoms in the Dhh^ECKO mice compared to control littermates.

However, 2 questions remain regarding barrier properties at the Glia Limitans: (1) Is this a specific aspect of the Dhh^ECKO mouse model or more broadly applicable to any model of BBB breakdown; and (2) is it observed in other CNS pathologies than inflammation?

A model of pericyte-deficient mice featuring BBB breakdown has been described; but, in this model, endothelial permeability is due to increased transcytosis and not junction degradation as in the Dhh^ECKO mouse model. Therefore, it would be interesting to study the behavior of astrocytes in this model, notably their capacity to express CLDN4 and to restrain parenchymal access to plasmatic proteins. Should astrocyte barrier properties be observed in the pericyte-deficient mouse model, thus would be suggestive for a more generalized role of BBB breakdown in driving astrocyte barrier properties, either by parallel or convergent mechanisms.

The expression of CLDN4 by astrocytes has only been identified in mouse models of CNS inflammation (acute CNS inflammation model (stereotactic injection of the pro-inflammatory cytokine IL-1β and model of multiple sclerosis: EAE MOG_35-55) [26]. However, in the present study, we demonstrated that CLDN4 is expressed by astrocytes under resting conditions in response to BBB opening in the Dhh^ECKO mouse model and after cortical stereotactic injection of the pro-permeability factor VEGFA. The next step would be to examine astrocytic CLDN4 expression in other pathologies of the CNS, notably Alzheimer’s disease, stroke, or amyotrophic lateral sclerosis in which BBB permeability has been identified as a critical pathophysiological player. This could implicate CLDN4-mediated barrier function in astrocytes as a more generalized defense against BBB opening in other chronic diseases of the CNS.

The critical role of Hh signaling in CNS neuroinflammation was first highlighted in 2011; this study revealed that during EAE, the morphogen SHH is expressed by reactive astrocytes and participates in the maintenance of BBB integrity [4]. Following this discovery, our group found that DHH is physiologically expressed at the BBB in adults [8]. Here, we demonstrated for the first time that DHH is down-regulated at the BBB during EAE and that DHH knockdown is sufficient to induce BBB permeability by inhibiting CDH5 and CLDN5 expression through the modulation of FOXO1 activity, strengthening the idea that HH signaling is essential to control BBB integrity both physiologically and under multiple sclerosis conditions. Based on our results and the literature, we hypothesize that DHH is necessary to maintain BBB tightness under physiological conditions and that DHH down-regulation under inflammatory conditions might be offset by astrocytic SHH secretion to maintain BBB homeostasis during disease progression.

Over the past years, the field has begun to acknowledge the fact that the BBB is not the sole line of defense of the CNS and that the astrocytic end feet of the Glia Limitans might play a role in restricting access to the parenchyma. Indeed, it was first described that in spinal cord injury, astrocyte scar borders corral inflammatory cells within areas of damaged tissue [32,33]. Moreover, we have found that during EAE, reactive astrocytes of the Glia Limitans form tight junctions of their own containing CLDN4 [26], a junction protein also expressed in tightly sealed epithelia [34,35]. Noteworthy is the fact that down-regulation or reorganization of CLDNs and other tight junction proteins has been implicated in permeability in various tissues, particularly the gut [19,36,37]; however, reports of dynamic tight junction protein induction resulting in functional barrier formation have been rare [38,39]. Here, we have shown for the first time that genetically induced disruption of endothelial junctions is sufficient to induce CLDN4 expression at the Glia Limitans under resting condition, identifying an inducible astrocyte barrier mediated fully by signals transmitted by the open BBB.

It has already been described, by our group and others, that astrocytes can send signals, notably VEGFA [18], thymidine phosphorylase (TYMP) [17], and SHH [4], to the BBB to modulate its state (tight versus permeable). In this study, we identify a reciprocal signaling pathway demonstrating that BBB endothelial junction disruption leads to CLDN4 expression at the Glia Limitans. Interestingly, endothelial cell capacity to send signals to neighboring cells has been previously identified, notably in the context of pericyte (mural cells associated with arterioles, capillaries, and venules) recruitment at the vascular wall [40]. Specifically, it has been shown that platelet-derived growth factor subunit b (PDGFB) is secreted from angiogenic sprout endothelium where it serves as an attractant for co-migrating pericytes, which in turn express platelet-derived growth factor receptor beta (PDGFRβ) [41]. Based on these arguments and our results, it appears highly likely that endothelial signals can be sent to astrocytes. Identifying such signals will be the aim of future studies by our group.

We then showed that in animals, which exhibit an open BBB (Dhh^ECKO mice), astrocytic end feet of the Glia Limitans form a barrier more efficiently than in the littermate controls, leading to the protective effect observed clinically in the model of multiple sclerosis. This is somehow reminiscent of what is observed in brain ischemic preconditioning where a mild nonlethal ischemic episode (preconditioning) can produce resistance to a subsequent more severe ischemic insult [42,43]. Here, inducing BBB opening and PVS plasmatic protein accumulation produces resistance to the subsequent massive inflammatory infiltration induced by multiple sclerosis development. This could account for the infrequency of recurrent multiple sclerosis relapse/lesion formation at the same location in the CNS. Interestingly, among neurons and nonneuronal cells, astrocytes are considered increasingly important in regulating cerebral ischemic tolerance [44], and a parallel can be easily drawn between these results and ours showing a major role for “preconditioned” astrocytes in the control of “chronic neuroinflammation tolerance” and protection against further relapse.

In light of the above observations, we may assume that the CNS has the ability to protect itself against isolated BBB leakage episodes through a secondary barrier at the Glia Limitans that takes over once the BBB is open. Moreover, it suggests that manipulation of the BBB and Glia Limitans in combination may have greater potential than either alone to control CNS entry of leukocytes and pro-inflammatory soluble factors in conditions such as multiple sclerosis and perhaps more widely. Indeed, taking into account both components of the neurovascular unit is of translational interest, notably to limit CNS parenchymal access to pathogenic agents by strengthening the Glia Limitans once the BBB is open in cardiovascular diseases such as brain ischemic strokes [45], neuroinfections [46], and neurodegeneration (Parkinson/Alzheimer’s diseases and vascular dementia) [47], or to facilitate parenchymal access to drugs, by opening the BBB and Glia Limitans together, in CNS tumor treatment [48]. Along similar lines, it is unknown how the barrier properties of the Glia Limitans may impact the pharmacokinetics of drugs that must enter the CNS parenchyma in conditions such as multiple sclerosis, which may account for treatment failure.

In summary, our study first demonstrates the critical role of DHH in maintaining BBB integrity. We find that DHH is down-regulated during the animal model of multiple sclerosis and that Dhh knockdown leads to BBB opening. Using Dhh knockdown as a tool to cause BBB opening, we then show that BBB permeability is sufficient to induce a secondary barrier at the Glia Limitans, mediated by CLDN4 and astrocyte reactivity. These findings not only highlight the capacity for bidirectional signaling between the endothelial BBB and the astrocytic Glia Limitans in modulating the double barriers of the CNS but also provide support for a novel concept of “chronic neuroinflammatory tolerance”, in which chronic induction of Glia Limitans barrier properties by BBB opening may lead to a protective effect against neuroinflammatory disease activity and progression.

Methods

Human tissues

Cortical sections from multiple sclerosis patients (active lesions) and healthy controls (frontal cortex) were obtained from the NeuroCEB bio bank (https://www.neuroceb.org/fr). The sections were 30 μm thick and obtained from fresh frozen samples.

Mice

Dhh Floxed (Dhh^Flox) mice were generated at the “Institut Clinique de la Souris” through the International Mouse Phenotyping Consortium (IMPC) from a vector generated by the European conditional mice mutagenesis program, EUCOMM, and described before [8].

The Cre recombinase in cadherin5 (Cdh5)-Cre^ERT2 mice was activated by intraperitoneal injection of 1-mg tamoxifen (Sigma Aldrich, St. Louis, Missouri, United States of America) for 5 consecutive days at 8 weeks of age. Mice were phenotyped 2 weeks later. Successful and specific activation of the Cre recombinase has been verified by measuring recombination efficacy in Cdh5-Cre^ERT2;Rosa26^mTmG mice (S2A Fig). Importantly, Dhh endothelial knockdown does not impact CNS angiogenesis (S2 Data) nor angioarchitecture (S2D Fig). The Cdh5-Cre^ERT2 mice and C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, Maine, USA).

Ethical statement

Human samples: The NeuroCEB bio bank and the INSERM U1034 certify that all human sections utilized for this study are ethically obtained with documented, legal permission for research use (authorization number #AC-2018-3290 obtained from the Ministry of Higher Education and Research) and in the respect of the written given consent from the source person in accordance with applicable laws and the World Medical Association (WMA) Helsinki declaration of 2013.

Animal experiments were performed in accordance with the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes and approved by the local Animal Care and Use Committee of the Bordeaux University CEEA50 (IACUC protocol #16901).

Neurovascular fraction enrichment from mouse CNS

Mouse was humanely killed by cervical dislocation, and its head was cut and rinsed with 70% ethanol. Brain and spinal cord were then harvested, and cerebellum, olfactory bulb, and white matter were removed from the brain with sterile forceps. Additionally, meninges were eliminated by rolling a sterile cotton swab at the surface of the cortex. The cortex and spinal cord were then transferred in a potter containing 2 mL of buffer A (HBSS 1X w/o phenol red (Gibco, Waltham, Massachusetts, USA), 10-mM HEPES (Gibco), and 0.1% bovine serum albumin (BSA) (Sigma Aldrich), and the CNS tissue was pounded to obtain an homogenate, which was collected in a 15-mL tube. The potter was rinsed with 1 mL of buffer A, which was added to the 2-mL homogenate. Cold 30% dextran solution was then added to the tube (V:V) to obtain a 15% dextran working solution centrifuged for 25 minutes at 3,000 g, 4°C without brakes. After centrifugation, the pellet (neurovascular components and red cells) was collected, and the supernatant (dextran solution and neural components) was centrifuged again to get the residual vessels. Neurovascular components were then pooled and resuspended in 4 mL of buffer B (HBSS 1X Ca²⁺/ Mg²⁺ free with phenol red (Gibco), 10-mM HEPES (Gibco), and 0.1% BSA (Sigma Aldrich)).

Neurovascular fraction enrichment for RT-PCR, western blots, or immunohistochemistry

After centrifugation of the cell suspension, the pellet was washed 3 times with the buffer B and filtered through a 100-μm nylon mesh (Millipore Corporation, Burlington, Massachussetts, USA). The nylon mesh was washed with 7 mL of buffer B to collect the retained enriched neurovascular fractions. The suspension was then centrifuged for 10 minutes at 1,000 g, and the pellet suspended in 300 μL of radioimmunoprecipitation assay (RIPA) lysis buffer for western blot analysis or 1,000 μL of Tri-Reagent (MRC, Cincinnati, Ohio, USA) for quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. For immunohistochemistry, the pellet was suspended in 3 mL of a solution of matrigel (Corning, Steuben, New York, USA)-Dulbecco’s Modified Eagle Medium (DMEM) 1 g/L glucose, Mg⁺, Ca²⁺ (Gibco) 1:80, distributed on a labtek (Starstedt, Nümbrecht, Germany) (1 mouse brain is needed to seed 1 labtek) and incubated for 30 minutes at 37°C. Finally, the enriched neurovascular fraction embedded in the matrigel (Corning) solution was fixed with 10% formalin for 10 minutes.

Primary culture of mouse CNS micro vascular endothelial cells (CNS MECs)

After centrifugation of the cell suspension, the pellet was washed 3 times with the buffer B and transferred in an enzyme solution (2 mg/mL Collagenase/Dispase (Roche, Bale, Switzerland), 0.147 μg/mL TLCK (Lonza, Bäle, Switzerland), and 10 μg/mL DNAse 1 (Roche)), prewarmed at 37°C, before being placed on a shaking table at maximum speed agitation at 37°C. After 30 minutes, the digestion was stopped by adding 10 mL of buffer B, the cell suspension centrifuged, and the digested neurovascular pellet washed 3 times with 3 mL of buffer B. After the 3 washing steps, the digested neurovascular pellet was resuspended in 1 mL of Mouse Brain Endothelial Cell Culture Medium (DMEM 1 g/L glucose, Mg⁺, Ca²⁺ (Gibco), fetal bovine serum (FBS) 20% (Gibco), sodium pyruvate 2% (Gibco), nonessential amino acids 2% (Lonza), FGF 1 ng/mL (PeproTech, Rocky Hill, New Jersey, USA), and gentamycin 10 mg/mL (Gibco)), and plated on a labtek (Starstedt) (1 mouse brain is needed to seed 1 labtek) previously coated with 2% matrigel (Corning) diluted in DMEM 1 g/L glucose, Mg⁺, Ca²⁺ (Gibco).

Cell culture

HBMECs (Alphabioregen–CliniSciences, Nanterre, France) were cultured in endothelial basal medium-2 (EBM-2) supplemented with EGM-2 BulletKits (Lonza). Cells from passage 3 to passage 6 were used. Before any treatment, cells were serum starved in 0.5% FBS EGM-2 medium for 24 hours. NHAs (Lonza) were cultured in astrocyte basal medium (ABM) supplemented with AGM BulletKits (Lonza). Cells from passage 2 to passage 4 were used. Before any treatment, cells were serum starved in DMEM 1 g/L glucose, Mg⁺, Ca²⁺ (Gibco) without serum for 24 hours.

Cytokines/growth factors/chemicals

Human IL-1β was purchased from PeproTech (Rocky Hills, New Jersey, USA), and Human and mouse VEGF-165 (VEGFA) were purchased from CliniSciences (Nanterre, France). Based on previous studies, Human IL-1β and Human VEGF-165 were routinely used at 10 ng/mL [19,49]. Mouse VEGF-165 was used at a concentration of 20 ng/μL. The inhibitor of FOXO1 total (AS1842856) was purchased from Merck (Kenilworth, New Jersey, USA) and was used at 100 nM [50]. D-Mannitol was purchased from Sigma Aldrich (St. Louis, Missouri, USA) and was used at 100 mM [51].

Antibodies

Anti-CLDN4 (mouse), anti-CLDN5 (mouse (tissues) and rabbit (cell culture)), anti-human ZO1 (rabbit), and anti-GFAP (rat) were from Invitrogen (Carlsbad, California, USA). Anti-CDH5 (goat) was from R&D systems (Minneapolis, Minnesota, USA). Anti-human CDH5 (mouse) and anti-human DHH (H-85) (rabbit) were from Santa Cruz Biotech (Santa Cruz, California, USA). Anti-FGB (rabbit) and anti-human PECAM1 (mouse) were from Dako (Carpinteria, California, USA). Anti-ALB (sheep) and anti-MBP (rat) were from Abcam (Cambridge, Massachusetts, USA). Anti-CD45 (rat) was from eBioscience (San Diego, California, USA). Anti-RNA binding fox-1 homolog 3 also known as neuronal nuclei antigen (NEUN) (rabbit) and anti-AQP4 (rabbit) were from Millipore (Billerica, Massachusetts, USA). Anti-LAM (rabbit) was from Sigma Aldrich (St. Louis, Missouri, USA). Anti-ZO1 (rabbit) was from Life Technologies (Grand Island, New York, USA). Anti-FOXO1 (rabbit), anti-p-FOXO1 (rabbit), and anti-β-ACTIN (rabbit) were from cell signaling (Danvers, Massachusetts, USA).

Quantitative RT-PCR

RNA was isolated using Tri Reagent (Molecular Research Center) as instructed by the manufacturer, from 3 × 10⁵ cells or from isolated mouse-enriched neurovascular fractions. For qRT-PCR analyses, total RNA was reverse transcribed with Moloney Murine Leukemia Virus (M-MLV) reverse transcriptase (Promega, Madison, Wisconsin, USA), and amplification was performed on a DNA Engine Opticon2 (MJ Research, St. Bruno, Canada) using B-R SYBER Green SuperMix (Quanta Biosciences, Beverly, Massachusetts, USA). Primer sequences are reported in Table 1.

Table 1. List of primers used for RT qPCR.

mDhh	F	5′ -CTTGGACATCACCACGTCTG- 3′
	R	5′ -ATGTAGTTCCCTCAGCCCCT- 3′
mIcam1	F	5′ -TGGCCTGGGGGATGCACACT- 3′
	R	5′-CCACCGGGCTGTAGGTGGGT-3′
mVcam1	F	5′ -CGTACACCATCCGCCAGGCA- 3′
	R	5′ -TAGAGTGCAAGGAGTTCGGGCG- 3′
mCldn5	F	5′ -GCAAGGTGTATGAATCTGTGCT- 3′
	R	5′ -GTCAAGGTAACAAAGAGTGCCA- 3′
mZo1	F	5′ -GCTAAGAGCACAGCAATGGA- 3′
	R	5′ -GCATGTTCAACGTTATCCAT- 3′
mβ-actin	F	5′ -GAAGCTGTGCTATGTTGCTCTA- 3′
	R	5′—GGAGGAAGAGGATGCGGCA- 3′
hGfap	F	5′ -TGGGTCAAAGGAAACCGGAA- 3′
	R	5′ -GAAAGTCCCAAGCCATCAGC- 3′
hDhh	F	5′ -AACAGCTTACTTCCGGCTCC- 3′
	R	5′ -CGACTCTTGTGGGCTCTGTT- 3′
hIcam1	F	5′ -ACGCCGGAGGACAGGGCATT- 3′
	R	5′ -GGGGCTATGTCTCCCCCACCA- 3′
hVcam1	F	5′ -GGCCCAGTTGAAGGATGCGGG- 3′
	R	5′ -AGAGCACGAGAAGCTCAGGAGAA- 3′
hAldh1L1	F	5′ -CCAGGGTTCTTCTTTGAGCCA- 3′
	R	5′ -CACCAGAAGCCAGGCCAAAT- 3′
hVim	F	5′ -CAGTTTTTCAGGAGCGCAAGA- 3′
	R	5′ -CAAGTTGGTTGGATACTTGCTGG- 3′
hCldn4	F	5′ -GACACTAATGAGCCTGGGAGG- 3′
	R	5′ -GTGCACAGGTCCCATTTATTGTAG- 3′
hβ-actin	F	5′ -GCTGTGCTACGTCGCCCTG- 3′
	R	5′ -GGAGGAGCTGGAAGCAGCC- 3′

Open in a new tab

F, forward; R, reverse; RT qPCR, quantitative reverse transcription polymer chain reaction.

β-actin was used as the household gene.

The relative expression of each mRNA was calculated by the comparative threshold cycle method and normalized to β-actin mRNA expression.

Western blots

Protein expression was evaluated by SDS-PAGE. Protein loading quantity was controlled using the rabbit monoclonal anti-β-actin antibody (cell signaling). Secondary antibodies were from Invitrogen. The signal was then revealed by using an Odyssey Infrared imager (LI-COR, Lincoln, Nebraska, USA). For quantification, the mean pixel density of each band was measured using Image J software (NIH, Bethesda, Maryland, USA), and data were standardized to β-actin, and fold change versus control calculated.

Stereotactic injection

Ten-week-old C57BL/6 mice (4 mice per condition) were anaesthetized using isoflurane (3% induction and 1% maintenance) (Virbac Schweiz, Glattbrugg, Germany) and placed into a stereotactic frame (Stoelting Co., Illinois, USA). To prevent eye dryness, an ophthalmic ointment was applied at the ocular surface to maintain eye hydration during the time of surgery. The skull was shaved, and the skin incised on 1 cm to expose the skullcap. Then, a hole was drilled into the skull, using a pneumatic station S001+TD783 Bien Air, until reaching the dura mater. A total of 3 μl of murine VEGFA (20 ng/μL), 3 μL of healthy mouse plasma, or 3 μL of vehicle control (PBS) were then delivered at 0.01 μl/s into the frontal cortex at coordinates of 1 μm posterior to bregma, 2 μm left of the midline, and 1.5 μm below the surface of the cortex [36].

Mice received a subcutaneous injection of buprenorphine (0.05 mg/kg) (Ceva santé animale, Libourne, France) 30 minutes before surgery and again 8 hours post-surgery to assure a constant analgesia during the procedure and postoperatively. Mice were humanely killed by pentobarbital (Richter Pharma, Wels, Austria) overdose at 24 hours post injection (dpi). For histological assessment, the brain of each animal was harvested.

Experimental autoimmune encephalomyelitis (EAE)

Ten-week-old female mice were immunized by subcutaneous injection of 300-μg myelin oligodendrocyte glycoprotein-35-55 (MOG_35–55) (Hooke Laboratories, Lawrence, Massachusetts, USA) in 200-μl Freund’s Adjuvant containing 300-μg/mL mycobacterium tuberculosis H37Ra (Hooke Laboratories) in the dorsum. Mice were administered with 500-ng pertussis toxin (PTX) intraperitoneously on day of sensitization and 1 day later (Hooke Laboratories). The emulsion provides antigen, which initiates expansion and differentiation of MOG-specific autoimmune T cells. PTX enhances EAE development by providing additional adjuvant. EAE will develop in mice 7 to 14 days after immunization (Day 0): Animals that develop EAE will become paralyzed. Disease was scored (0, no symptoms; 1, floppy tail; 2, hind limb weakness (paraparesis); 3, hind limb paralysis (paraplegia); 4, fore and hind limb paralysis; 5, death)) [31] from day 7 post immunization until day 32 post immunization. At Day 32, all the animals were euthanized by pentobarbital (Richter Pharma) overdose. For histological assessment, cervical, lumbar, and dorsal sections of each animal spinal cord, as well as the spleen, were harvested.

Immunohistochemistry

Prior to tissue collection and staining, mice were transcardially perfused with PBS (10 mL) followed by 10% Formalin (10 mL) to remove intravascular plasma proteins. Brain and spinal cord samples were either fixed in 10% formalin for 3 hours, incubated in 30% sucrose overnight, OCT embedded and cut into 9-μm thick sections, or directly OCT embedded and cut into 9 μm thick sections. Cultured cells were fixed with 10% formalin for 10 minutes. Human frozen sections were used directly without any prior treatment. Concerning the fixed sections, for CLDN4, prior to blocking, sections were soaked in Citrate (pH 7.5; 100°C). For CLDN5, prior to blocking, sections were soaked in EDTA (pH 6.0; 100°C). For CD45, sections were treated with 0.5 mg/mL protease XIV (Sigma Aldrich) at 37°C for 5 minutes. Primary antibodies were used at 1:100 except CLDN4 (1:50), FGB (1:1,000), and ALB (1:1,000). Samples were examined using a Zeiss Microsystems confocal microscope (Oberkochen, Germany), and stacks were collected with z of 1 μm.

For immunofluorescence analyses, primary antibodies were resolved with Alexa Fluor–conjugated secondary polyclonal antibodies (Invitrogen), and nuclei were counterstained with DAPI (1:5000) (Invitrogen). For all immunofluorescence analyses, negative controls using secondary antibodies only were done to check for antibody specificity.

Morphometric analysis

Morphometric analyses were carried out using NIH ImageJ software (NIH).

BBB permeability was evaluated by measuring tight junction integrity and plasmatic protein extravasation. Brain and spinal cord sections were immunostained for the expression of CLDN5/CDH5 and FGB/IgG/ALB, respectively. For each brain or spinal cord section, CLDN5+, CDH5+, FGB+, IgG+, and ALB+ areas were quantified in 20 pictures taken at the margins of the lesion area under 40× magnification. One section was quantified per spinal cord (3 different zones are displayed on the same section: 1 cervical, 1 lumbar, and 1 dorsal to get a global vision of the lesion) (per mouse).

Leukocyte densities were evaluated in sections stained for the expression of CD45 leukocyte population. For each brain or spinal cord section, CD45+ leukocytes were counted in 20 pictures randomly taken under 40× magnification. One section was quantified per spinal cord (3 different zones are displayed on the same section: 1 cervical, 1 lumbar, and 1 dorsal to get a global vision of the inflammatory lesion) (per mouse).

Demyelination was evaluated in spinal cord sections stained for the expression of MBP. For each spinal cord section, MBP+ area was quantified in 10 pictures taken in and around inflammatory lesion sites under 20× magnification. One section was quantified per spinal cord (3 different zones are displayed on the same section: 1 cervical, 1 lumbar, and 1 dorsal to get a global vision of the lesion) (per mouse).

Plasmatic protein and leukocyte infiltrate distribution at the neurovascular unit were evaluated in brain or spinal cord sections (1) triple stained for PECAM1 or CDH5 (markers of the BBB), IgG (plasmatic proteins), and AQP4 or GFAP (markers of the Glia Limitans); or (2) double stained for IgG (plasmatic proteins), FITC Dextran (exogenous tracer) or CD45 (leukocyte infiltrate), and LAM (marker of basement membranes). For each section, the distribution (between the lumen, the PVS, and the parenchyma) of IgG, 70 kDa FITC Dextran or leukocyte infiltrate was quantified for 5 to 6 neurovascular units randomly taken under 60× magnification, each 1 from a different animal. We used negative working images highlighting the endothelial BBB and astrocyte Glia Limitans and outlined them by using dotted lines. Dotted lines were then transferred to the plasmatic protein or leukocyte infiltration images so that we could quantify their distribution within the 3 compartments (lumen, PVS, and parenchyma).

Statistical analyses

Results are reported as mean ± SEM. Comparisons between groups were analyzed for significance with the nonparametric Mann–Whitney U test, the nonparametric Kruskal–Wallis test followed by the Dunn multiple comparison test when we have more than 2 groups, the chi-squared test for the distribution of plasmatic proteins and inflammatory cells in the neurovascular unit, or a nonlinear regression test (Boltzmann sigmoidal) for the EAE scoring analysis using GraphPad Prism v8.0.2 (GraphPad, San Diego, California, USA). Differences between groups were considered significant when P ≤ 0.05 (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001).

Supporting information

S1 Text

Supporting information file containing the S1 Data and S2 Data legends and DOI links (A), the Supporting Methods (B), and the associated References (C).

(DOCX)

Click here for additional data file.^{(29KB, docx)}

S1 Raw Images

Supporting information file containing the original, uncropped, and minimally adjusted images supporting all blot and gel results reported in Fig 2 panel G, Fig 4 panel F, and Fig 7 panel I as well as S6 Fig panel A and B.

(PDF)

Click here for additional data file.^{(17.6MB, pdf)}

S1 Fig. (Related to Fig 1 and Fig 2).

BMEC purity analysis shows limited contamination by other neurovascular components (A) Primary BMECs from Dhh^ECKO and control mice were isolated and cultured on Lab-Tek. ZO1 (in green) and SMA, NG2, IBA1, or GFAP (in red) localizations were evaluated by immunofluorescent staining of a confluent cell monolayer. Nuclei were stained with DAPI (in blue). The experiment was repeated 3×. Mouse brain in situ hybridization analysis highlights Dhh expression at the BBB. (B) C57BL/6 cortical cross section hybridized with the Dhh RNA probe show Dhh expression in blood vessels. A control section hybridized with the Dhh antisense RNA probe show the absence of hybridization signals. MECs isolated from the spinal cord of mice induced with EAE MOG_35-55 are highly contaminated by leukocytes (C) Spinal cord MECs were isolated from 12-week-old C57Bl/6 mice at day 13 post induction with EAE MOG_35-55 or placebo and Pecam1, Cd45, Sma, Ng2, and Gfap expressions were quantified by qRT-PCR (cycle threshold mean values) before and after the cell suspension were depleted in CD45⁺ leukocytes. β-actin is used as a reference. A MACS CD45+ cell depletion step is sufficient to eliminate leukocyte contamination in MECs isolated from the spinal cord of mice induced with EAE MOG_35-55 (D) Representative graphs of flow cytometry analysis performed on C57BL/6 mouse spinal cord MECs harvested at day 13 post EAE MOG_35-55 induction. Analysis showed no CD45+ cell population in samples depleted in leukocytes using the MACS CD45+ cell isolation kit (Miltenyi Biotec) (EAE MOG_35-55 spinal cord MECs n = 3; EAE MOG_35-55 spinal cord MECs + MACS CD45+ cell isolation n = 3). Dhh endothelial knockdown does not impact CNS MEC viability in culture (E-F). (E) Primary CNS MECs from Dhh^ECKO and control mice were isolated and cultured on Lab-Tek and immunostained with Propidium Iodide (in red) and Hoechst 33342 to label the nuclei (in blue). (F) Dhh^ECKO versus control primary CNS MEC viability was evaluated by quantifying the number of nuclei that incorporated Propidium Iodide. NS, Mann–Whitney U test. The underlying data for S1 Fig can be found in S2 Data (https://doi.org/10.6084/m9.figshare.12625085.v7).

(TIF)

Click here for additional data file.^{(7.2MB, tif)}

S2 Fig. (Related to Fig 3, Fig 4, Fig 6 and Fig 7).

Cadherin5Cre^ERT2 recombinase activation in blood vessels is successful and specific (A) Brain and spinal cord sections were harvested from Cadherin5Cre^ERT2,Rosa26^mTmG mice and littermate controls and immunostained with anti-GFP (in green) and anti-PECAM1 (in red) antibodies. Dhh endothelial knockdown does not impact CNS angiogenesis (B) Spinal cord sections were harvested from Dhh^ECKO mice and littermate controls and immunostained with an anti-IB4 (in green) antibody. IB4 positive area was quantified (Dhh^ECKO n = 7, control n = 6). (C) Cortical sections were harvested from Dhh^ECKO mice and littermate controls and immunostained with an anti-IB4 (in green) antibody. IB4 positive area was quantified (Dhh^ECKO n = 6, WT n = 6). Dhh endothelial knockdown does not impact brain angioarchitecture (D) The vascular network in the brain of Dhh^ECKO mice and control littermates was imaged by microcomputed tomography (micro-CT). NS, Mann–Whitney U test. The underlying data for S2 Fig can be found in S1 Data (individual numerical data (excel file)) and S2 Data (statistical analysis (Prism file)) (https://doi.org/10.6084/m9.figshare.12625034.v6; https://doi.org/10.6084/m9.figshare.12625085.v7).

(TIF)

Click here for additional data file.^{(8.9MB, tif)}

S3 Fig. Endothelial-specific Dhh inactivation induces BBB permeability in vivo.

(A) Spinal cord sections were harvested from Dhh^ECKO mice and littermate controls and immunostained with anti-IB4 (in green) and anti-CLDN5 (in red) antibodies. Representative IB4/CLDN5 staining was shown. (B–C) Spinal cord sections were harvested from Dhh^ECKO mice and littermate controls and immunostained with anti-CDH5 or anti-CLDN5 (in green), and anti-FGB or anti-ALB (in red) antibodies. Representative (B) CDH5/FGB and (C) CLDN5/ALB staining were shown.

(TIF)

Click here for additional data file.^{(2.7MB, tif)}

S4 Fig. (Related to Fig 4).

Dhh^ECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans. (A) Brain sections were harvested from Dhh^ECKO mice and littermate controls injected with 70 kDa FITC Dextran and (A) immunostained with an anti-LAM (in red) antibody (nuclei were stained with DAPI (in blue)). Representative LAM/FITC Dextran staining was shown. (B) Negative working images of LAM channels were used to highlight the endothelial (EBM) and astrocyte (ABM) basement membranes, using orange dotted lines. The outlines were then transferred to the FITC Dextran images to discriminate the distribution of FITC Dextran between the lumen, PVS, and parenchyma. (C–D) The distribution of FITC Dextran within the lumen, PVS, and parenchyma was quantified. (Dhh^ECKO n = 6, control n = 6) P < 0.0001, 95% CI: 0.3518 to 0.5734, chi-squared test. (E) Negative working images of AQP4/PECAM1 channels were used to highlight the endothelial (EBM) and astrocyte (ABM) basement membranes, using orange dotted lines. The outlines were then transferred to the IgG images to discriminate the distribution of IgG between the lumen, PVS, and parenchyma. The underlying data for S4 Fig can be found in S2 Data (https://doi.org/10.6084/m9.figshare.12625085.v7).

(TIF)

Click here for additional data file.^{(3.1MB, tif)}

S5 Fig. (Related to Fig 4).

Dhh^ECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans. (A) Spinal cord sections were harvested from Dhh^ECKO mice and littermate controls and immunostained with anti-LAM (in green) and anti-IgG (in red) antibodies (nuclei were stained with DAPI (in blue)). Representative LAM/IgG staining was shown. (B) Negative working images of LAM channels were used to highlight the endothelial (EBM) and astrocyte (ABM) basement membranes, using orange dotted lines. The outlines were then transferred to the IgG images to discriminate the distribution of IgG between the lumen, PVS, and parenchyma.

(TIF)

Click here for additional data file.^{(1.5MB, tif)}

S6 Fig. (Related to Fig 4).

Small intestine samples are used as a positive control for the quantification of CLDN4 expression by western blot. (A) Representative blots of CLDN4 expression level on control mouse neurovascular unit lysates and mouse small intestine lysates were shown. There is astrocyte reactivity but no astrocytic CLDN4 up-regulation at the capillary level in Dhh^ECKO mouse CNS. (B) Representative blots of GFAP and CLDN4 expression level on Dhh^ECKO and control lysates were shown. Lysates were obtained with neurovascular units, which are 20 μm and larger. (C) GFAP expression level was quantified by western blot on Dhh^ECKO and control lysates obtained with neurovascular units, which are 20 μm and larger. (D) CLDN4 expression level was quantified by western blot on Dhh^ECKO and control lysates obtained with neurovascular units, which are 20 μm and larger. (Dhh^ECKO n = 6, WT n = 6). *P ≤ 0.05, Mann–Whitney U test. The underlying data for S6 Fig can be found in S2 Data (https://doi.org/10.6084/m9.figshare.12625085.v7).

(TIF)

Click here for additional data file.^{(453.8KB, tif)}

S7 Fig. (Related to Fig 5 and Fig 7).

Both VEGFA and Mannitol induce HBMEC permeability in vitro. (A) Cultured HBMECs were treated with PBS, VEGFA, or Mannitol for 6 h, and HBMEC monolayer permeability to 70 kDa FITC Dextran was quantified. Mice with endothelial Dhh knockdown display a reinforced barrier at the Glia Limitans restraining access to the parenchyma to inflammation in a model of multiple sclerosis: (B) Negative working images of GFAP/CDH5 channels were used to highlight the endothelial (EBM) and astrocyte (ABM) basement membranes, using orange dotted lines. The outlines were then transferred to the FGB images to discriminate the distribution of FGB between the lumen, PVS, and parenchyma. (C) Negative working images of the LAM channel were used to highlight the endothelial (EBM) and astrocyte (ABM) basement membranes, using blue dotted lines. The outlines were then transferred to the CD45 images to discriminate the distribution of leukocytes between the lumen, PVS, and parenchyma. *P ≤ 0.05, ****P ≤ 0.0001 Kruskal–Wallis test. The underlying data for S7 Fig can be found in S2 Data (https://doi.org/10.6084/m9.figshare.12625085.v7).

(TIF)

Click here for additional data file.^{(5.5MB, tif)}

Acknowledgments

We thank Dr. Mary P. Heyer for her proofreading and correction of the manuscript. We thank Annabel Reynaud, Sylvain Grolleau, and Maxime David for their technical help. We thank Christelle Boullé for administrative assistance.

Abbreviations

Aldh1l1: aldehyde dehydrogenase 1 family, member l1
ALB: albumin
AQP4: aquaporin 4
ABM: astrocyte basal medium
BBB: blood–brain barrier
BSA: bovine serum albumin
CD45: cluster of differentiation 45
CDH5: cadherin5
CLDN4: Claudin4
Cldn5: claudin5
CNS: central nervous system
CTNNβ1: catenin β1
DHH: desert hedgehog
DMEM: Dulbecco’s Modified Eagle Medium
EAE: experimental autoimmune encephalomyelitis
EBM-2: endothelial basal medium-2
FBS: fetal bovine serum
FGB: fibrinogen
FITC: fluorescein isothiocyanate
FoxO1: forkhead box O1
GFAP: glial fibrillary acidic protein
HBMEC: human brain microvascular endothelial cell
HH: Hedgehog
IBA1: ionized calcium binding adaptor molecule 1
IgG: immunoglobulin G
Icam1: intercellular adhesion molecule 1
IHH: Indian hedgehog
IL-1β: interleukin-1β
IMPC: International Mouse Phenotyping Consortium
LAM: laminin
LPS: lipopolysaccharide
MEC: microvascular endothelial cell
MMP: matrix metalloproteinase
MOG_35–55: myelin oligodendrocyte glycoprotein-35-55
NG2: neural/glia antigen 2
NHA: normal human astrocytes
PDGFB: platelet-derived growth factor subunit b
PDGFRβ: platelet-derived growth factor receptor beta
PECAM1: platelet/endothelial cell adhesion molecule 1
PTCH1: Patched-1
PTX: pertussis toxin
PVS: perivascular space
qRT-PCR: quantitative reverse transcription polymerase chain reaction
RIPA: radioimmunoprecipitation assay
SHH: Sonic hedgehog
SMA: smooth muscle actin
SMO: Smoothened
Vcam1: vascular cell adhesion molecule 1
VEGFA: vascular endothelial growth factor A
Vim: vimentin
WMA: World Medical Association
ZO1: zonula occludens 1

Data Availability

All relevant data are within the paper and its Supporting Information files. All S1 Data and S2 Data files are available from the Figshare public repository database. The S1 Data DOI link is https://doi.org/10.6084/m9.figshare.12625034.v6 The S2 Data DOI link is https://doi.org/10.6084/m9.figshare.12625085.v7.

Funding Statement

This study was supported by grants from the European Council (Marie Skłodowska-Curie Actions, Individual fellowship 2019 (MSCA-IF-2019)) (grant number GA794726) and the Fondation ARSEP (Fondation pour la Recherche sur la Sclérose En Plaques (https://www.arsep.org/) (grant number ARSEP 2019/R19083GG). This study was also co-funded by the “Institut National de la Santé et de la Recherche Médicale” (project U103420G, grant number U1034SE20GA). Grant number GA794726 was received by CC, grant number ARSEP 2019/R19083GG was received by CC and grant number U1034SE20GA was received by CC. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. CC received a salary for 2 years (2019 and 2020) from the European Council as part of the grant GA794726.

References

1.McConnell HL, Kersch CN, Woltjer RL, Neuwelt EA. The Translational Significance of the Neurovascular Unit. J Biol Chem. 2017;292(3):762–70. 10.1074/jbc.R116.760215 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Bartholomäus I, Kawakami N, Odoardi F, Schläger C, Miljkovic D, Ellwart JW, et al. Effector T cell interactions with meningeal vascular structures in nascent autoimmune CNS lesions. Nature. 2009;462 (7269):94–98. 10.1038/nature08478 [DOI] [PubMed] [Google Scholar]
3.Brück W, Bitsch A, Kolenda H, Brück Y, Stiefel M, Lassmann H. Inflammatory central nervous system demyelination: correlation of magnetic resonance imaging findings with lesion pathology. Ann Neurol. 1997;42(5):783–793. 10.1002/ana.410420515 [DOI] [PubMed] [Google Scholar]
4.Alvarez JI, Dodelet-Devillers A, Kebir H, Ifergan I, Fabre PJ, Terouz S, et al. The Hedgehog pathway promotes blood-brain barrier integrity and CNS immune quiescence. Science. 2011;334 (6063):1727–1731. 10.1126/science.1206936 [DOI] [PubMed] [Google Scholar]
5.Singh VB, Singh MV, Gorantla S, Poluektova LY, Maggirwar SB. Smoothened Agonist Reduces Human Immunodeficiency Virus Type-1-Induced Blood-Brain Barrier Breakdown in Humanized Mice. Sci Rep. 2016;6:26876 10.1038/srep26876 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Xia Y, He Q, Li Y, Chen S, Huang M, Wang Y, et al. Recombinant human sonic hedgehog protein regulates the expression of ZO-1 and occludin by activating angiopoietin-1 in stroke damage. PLoS ONE. 2013;8 (7):e68891 10.1371/journal.pone.0068891 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Chapouly C, Guimbal S, Hollier P-L, Renault M-A. Role of Hedgehog Signaling in Vasculature Development, Differentiation, and Maintenance. Int J Mol Sci. 2019;20(12). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Caradu C, Couffinhal T, Chapouly C, Guimbal S, Hollier P-L, Ducasse E, et al. Restoring Endothelial Function by Targeting Desert Hedgehog Downstream of Klf2 Improves Critical Limb Ischemia in Adults. Circ Res. 2018;123 (9):1053–1065. 10.1161/CIRCRESAHA.118.313177 [DOI] [PubMed] [Google Scholar]
9.Bitgood MJ, McMahon AP. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev Biol. 1995;172 (1):126–138. 10.1006/dbio.1995.0010 [DOI] [PubMed] [Google Scholar]
10.Robbins DJ, Fei DL, Riobo NA. The Hedgehog signal transduction network. Sci Signal. 2012;5(246):re6 10.1126/scisignal.2002906 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Iadecola C. The Neurovascular Unit Coming of Age: A Journey through Neurovascular Coupling in Health and Disease. Neuron. 2017;96 (1):17–42. 10.1016/j.neuron.2017.07.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Sweeney MD, Ayyadurai S, Zlokovic BV. Pericytes of the neurovascular unit: key functions and signaling pathways. Nat Neurosci. 2016;19(6):771–83. 10.1038/nn.4288 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Cheslow L, Alvarez JI. Glial-endothelial crosstalk regulates blood-brain barrier function. Curr Opin Pharmacol. 2016;26:39–46. 10.1016/j.coph.2015.09.010 [DOI] [PubMed] [Google Scholar]
14.Engelhardt B, Coisne C. Fluids and barriers of the CNS establish immune privilege by confining immune surveillance to a two-walled castle moat surrounding the CNS castle. Fluids Barriers CNS. 2011;8(1):4 10.1186/2045-8118-8-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Engelhardt B, Ransohoff RM. Capture, crawl, cross: the T cell code to breach the blood-brain barriers. Trends Immunol. 2012;33 (12):579–589. 10.1016/j.it.2012.07.004 [DOI] [PubMed] [Google Scholar]
16.Colombo E, Farina C. Astrocytes: Key Regulators of Neuroinflammation. Trends Immunol. 2016;37 (9):608–620. 10.1016/j.it.2016.06.006 [DOI] [PubMed] [Google Scholar]
17.Chapouly C, Tadesse Argaw A, Horng S, Castro K, Zhang J, Asp L, et al. Astrocytic TYMP and VEGFA drive blood-brain barrier opening in inflammatory central nervous system lesions. Brain J Neurol. 2015;138 (Pt 6):1548–1567. 10.1093/brain/awv077 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Argaw AT, Asp L, Zhang J, Navrazhina K, Pham T, Mariani JN, et al. Astrocyte-derived VEGF-A drives blood-brain barrier disruption in CNS inflammatory disease. J Clin Invest. 2012;122 (7):2454–2468. 10.1172/JCI60842 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Argaw AT, Gurfein BT, Zhang Y, Zameer A, John GR. VEGF-mediated disruption of endothelial CLN-5 promotes blood-brain barrier breakdown. Proc Natl Acad Sci U S A. 2009;106 (6):1977–1982. 10.1073/pnas.0808698106 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Engelhardt B, Carare RO, Bechmann I, Flügel A, Laman JD, Weller RO. Vascular, glial, and lymphatic immune gateways of the central nervous system. Acta Neuropathol (Berl). 2016;132 (3):317–338. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Di Russo J, Hannocks M-J, Luik A-L, Song J, Zhang X, Yousif L, et al. Vascular laminins in physiology and pathology. Matrix Biol. 2017;57–58:140–148. 10.1016/j.matbio.2016.06.008 [DOI] [PubMed] [Google Scholar]
22.Abbott NJ, Rönnbäck L, Hansson E. Astrocyte-endothelial interactions at the blood-brain barrier. Nat Rev Neurosci. 2006;7 (1):41–53. 10.1038/nrn1824 [DOI] [PubMed] [Google Scholar]
23.Iliff JJ, Wang M, Liao Y, Plogg BA, Peng W, Gundersen GA, et al. A Paravascular Pathway Facilitates CSF Flow Through the Brain Parenchyma and the Clearance of Interstitial Solutes, Including Amyloid β. Sci Transl Med. 2012;4 (147):147ra111–147ra111. 10.1126/scitranslmed.3003748 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Song J, Wu C, Korpos E, Zhang X, Agrawal SM, Wang Y, et al. Focal MMP-2 and MMP-9 Activity at the Blood-Brain Barrier Promotes Chemokine-Induced Leukocyte Migration. Cell Rep. 2015;10 (7):1040–1054. 10.1016/j.celrep.2015.01.037 [DOI] [PubMed] [Google Scholar]
25.Cruz-Orengo L, Holman DW, Dorsey D, Zhou L, Zhang P, Wright M, et al. CXCR7 influences leukocyte entry into the CNS parenchyma by controlling abluminal CXCL12 abundance during autoimmunity. J Exp Med. 2011;208 (2):327–339. 10.1084/jem.20102010 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Horng S, Therattil A, Moyon S, Gordon A, Kim K, Argaw AT, et al. Astrocytic tight junctions control inflammatory CNS lesion pathogenesis. J Clin Invest. 2017;127 (8):3136–3151. 10.1172/JCI91301 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Taddei A, Giampietro C, Conti A, Orsenigo F, Breviario F, Pirazzoli V, et al. Endothelial adherens junctions control tight junctions by VE-cadherin-mediated upregulation of claudin-5. Nat Cell Biol. 2008;10 (8):923–934. 10.1038/ncb1752 [DOI] [PubMed] [Google Scholar]
28.Davalos D, Akassoglou K. Fibrinogen as a key regulator of inflammation in disease. Semin Immunopathol. 2012;34 (1):43–62. 10.1007/s00281-011-0290-8 [DOI] [PubMed] [Google Scholar]
29.Nitta T, Yagita H, Sato K, Okumura K. Expression of Fc gamma receptors on astroglial cell lines and their role in the central nervous system. Neurosurgery. 1992;31(1):83–7; discussion 87–88. 10.1227/00006123-199207000-00012 [DOI] [PubMed] [Google Scholar]
30.Okun E, Mattson MP, Arumugam TV. Involvement of Fc Receptors in Disorders of the Central Nervous System. Neuromolecular Med. 2010;12 (2):164–178. 10.1007/s12017-009-8099-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Gurfein BT, Zhang Y, López CB, Argaw AT, Zameer A, Moran TM, et al. IL-11 regulates autoimmune demyelination. J Immunol Baltim Md 1950. 2009;183 (7):4229–4240. 10.4049/jimmunol.0900622 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Wanner IB, Anderson MA, Song B, Levine J, Fernandez A, Gray-Thompson Z, et al. Glial scar borders are formed by newly proliferated, elongated astrocytes that interact to corral inflammatory and fibrotic cells via STAT3-dependent mechanisms after spinal cord injury. J Neurosci. 2013;33 (31):12870–12886. 10.1523/JNEUROSCI.2121-13.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Sofroniew MV. Astrocyte barriers to neurotoxic inflammation. Nat Rev Neurosci. 2015;16 (5):249–263. 10.1038/nrn3898 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Hewitt KJ, Agarwal R, Morin PJ, et al. BMC Cancer. 2006;6:186 10.1186/1471-2407-6-186 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Acharya P, Beckel J, Ruiz WG, Wang E, Rojas R, Birder L, et al. Distribution of the tight junction proteins ZO-1, occludin, and claudin-4, -8, and -12 in bladder epithelium. Am J Physiol Renal Physiol. 2004;287 (2):F305–F318. 10.1152/ajprenal.00341.2003 [DOI] [PubMed] [Google Scholar]
36.Bruewer M, Luegering A, Kucharzik T, Parkos CA, Madara JL, Hopkins AM, et al. Proinflammatory Cytokines Disrupt Epithelial Barrier Function by Apoptosis-Independent Mechanisms. J Immunol. 2003;171 (11):6164–6172. 10.4049/jimmunol.171.11.6164 [DOI] [PubMed] [Google Scholar]
37.Kebir H, Kreymborg K, Ifergan I, Dodelet-Devillers A, Cayrol R, Bernard M, et al. Human TH17 lymphocytes promote blood-brain barrier disruption and central nervous system inflammation. Nat Med. 2007;13 (10):1173–1175. 10.1038/nm1651 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Kinugasa T, Sakaguchi T, Gu X, Reinecker H. Claudins regulate the intestinal barrier in response to immune mediators. Gastroenterology. 2000;118 (6):1001–1011. 10.1016/s0016-5085(00)70351-9 [DOI] [PubMed] [Google Scholar]
39.Fujita H, Chalubinski M, Rhyner C, Indermitte P, Meyer N, Ferstl R, et al. Claudin-1 expression in airway smooth muscle exacerbates airway remodeling in asthmatic subjects. J Allergy Clin Immunol. 2011;127 (6):1612–1621.e8. 10.1016/j.jaci.2011.03.039 [DOI] [PubMed] [Google Scholar]
40.Konstantin G, Guillem G, Annika A, Christer B. Endothelial-Mural Cell Signaling in Vascular Development and Angiogenesis. Arterioscler Thromb Vasc Biol. 2009;29 (5):630–638. 10.1161/ATVBAHA.107.161521 [DOI] [PubMed] [Google Scholar]
41.Lindahl P, Johansson BR, Levéen P, Betsholtz C. Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science. 1997;277 (5323):242–245. 10.1126/science.277.5323.242 [DOI] [PubMed] [Google Scholar]
42.Yunoki M, Kanda T, Suzuki K, Uneda A, Hirashita K, Yoshino K. Ischemic Tolerance of the Brain and Spinal Cord: A Review. Neurol Med Chir (Tokyo). 2017;57 (11):590–600. 10.2176/nmc.ra.2017-0062 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kirino T, Tsujita Y, Tamura A. Induced tolerance to ischemia in gerbil hippocampal neurons. J Cereb Blood Flow Metab. 1991;11 (2):299–307. 10.1038/jcbfm.1991.62 [DOI] [PubMed] [Google Scholar]
44.Koizumi S, Hirayama Y, Morizawa YM. New roles of reactive astrocytes in the brain; an organizer of cerebral ischemia. Neurochem Int. 2018;119:107–114. 10.1016/j.neuint.2018.01.007 [DOI] [PubMed] [Google Scholar]
45.Cai W, Zhang K, Li P, Zhu L, Xu J, Yang B, et al. Dysfunction of the neurovascular unit in ischemic stroke and neurodegenerative diseases: An aging effect. Ageing Res Rev. 2017;34:77–87. 10.1016/j.arr.2016.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Hou J, Baker LA, Zhou L, Klein RS. Viral interactions with the blood-brain barrier: old dog, new tricks. Tissue Barriers. 2016;4 (1):e1142492 10.1080/21688370.2016.1142492 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Zlokovic BV. The blood-brain barrier in health and chronic neurodegenerative disorders. Neuron. 2008;57 (2):178–201. 10.1016/j.neuron.2008.01.003 [DOI] [PubMed] [Google Scholar]
48.Chacko A-M, Li C, Pryma DA, Brem S, Coukos G, Muzykantov V. Targeted delivery of antibody-based therapeutic and imaging agents to CNS tumors: crossing the blood-brain barrier divide. Expert Opin Drug Deliv. 2013;10 (7):907–926. 10.1517/17425247.2013.808184 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Liu J, Zhao ML, Brosnan CF, Lee SC. Expression of type II nitric oxide synthase in primary human astrocytes and microglia: role of IL-1beta and IL-1 receptor antagonist. J Immunol Baltim Md 1950. 1996;157 (8):3569–3576. [PubMed] [Google Scholar]
50.Karki S, Farb MG, Ngo DTM, Myers S, Puri V, Hamburg NM, et al. Forkhead box O-1 modulation improves endothelial insulin resistance in human obesity. Arterioscler Thromb Vasc Biol. 2015;35 (6):1498–1506. 10.1161/ATVBAHA.114.305139 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Hess DC, Howard E, Cheng C, Carroll J, Hill WD, Hsu CY. Hypertonic mannitol loading of NF-kappaB transcription factor decoys in human brain microvascular endothelial cells blocks upregulation of ICAM-1. Stroke. 2000;31 (5):1179–1186. 10.1161/01.str.31.5.1179 [DOI] [PubMed] [Google Scholar]

PLoS Biol. doi: 10.1371/journal.pbio.3000946.r001

Decision Letter 0

Di Jiang, PhD

1 Apr 2020

Dear Dr Chapouly,

Thank you for submitting your manuscript entitled "Blood Brain Barrier genetic disruption leads to protective barrier formation at the Glia Limitans" for consideration as a Research Article by PLOS Biology.

Your manuscript has now been evaluated by the PLOS Biology editorial staff [as well as by an academic editor with relevant expertise] and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Apr 03 2020 11:59PM.

During resubmission, you will be invited to opt-in to posting your pre-review manuscript as a bioRxiv preprint. Visit http://journals.plos.org/plosbiology/s/preprints for full details. If you consent to posting your current manuscript as a preprint, please upload a single Preprint PDF when you re-submit.

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Given the disruptions resulting from the ongoing COVID-19 pandemic, please expect delays in the editorial process. We apologise in advance for any inconvenience caused and will do our best to minimize impact as far as possible.

Feel free to email us at plosbiology@plos.org if you have any queries relating to your submission.

Kind regards,

Di Jiang,

Associate Editor

PLOS Biology

PLoS Biol. doi: 10.1371/journal.pbio.3000946.r002

Decision Letter 1

Di Jiang, PhD

30 Apr 2020

Dear Dr Chapouly,

Thank you very much for submitting your manuscript "Blood Brain Barrier genetic disruption leads to protective barrier formation at the Glia Limitans" for consideration as a Research Article at PLOS Biology. Your manuscript has been evaluated by the PLOS Biology editors, an Academic Editor with relevant expertise, and by three independent reviewers.

In light of the reviews (below), we will welcome re-submission of a much-revised version that addresses the reviewers' concerns including, particularly, the important controls requested by reviewers 1 and 3, which should be included in the revision. We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent for further evaluation by the reviewers.

We expect to receive your revised manuscript within 2 months. Please note given we are in the midst of COVID-19, we are flexible regarding turnaround time for revision.

Please email us (plosbiology@plos.org) if you have any questions or concerns, or would like to request an extension. At this stage, your manuscript remains formally under active consideration at our journal; please notify us by email if you do not intend to submit a revision so that we may end consideration of the manuscript at PLOS Biology.

**IMPORTANT - SUBMITTING YOUR REVISION**

Your revisions should address the specific points made by each reviewer. Please submit the following files along with your revised manuscript:

1. A 'Response to Reviewers' file - this should detail your responses to the editorial requests, present a point-by-point response to all of the reviewers' comments, and indicate the changes made to the manuscript.

*NOTE: In your point by point response to the reviewers, please provide the full context of each review. Do not selectively quote paragraphs or sentences to reply to. The entire set of reviewer comments should be present in full and each specific point should be responded to individually, point by point.

You should also cite any additional relevant literature that has been published since the original submission and mention any additional citations in your response.

2. In addition to a clean copy of the manuscript, please also upload a 'track-changes' version of your manuscript that specifies the edits made. This should be uploaded as a "Related" file type.

*Re-submission Checklist*

When you are ready to resubmit your revised manuscript, please refer to this re-submission checklist: https://plos.io/Biology_Checklist

To submit a revised version of your manuscript, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' where you will find your submission record.

Please make sure to read the following important policies and guidelines while preparing your revision:

*Published Peer Review*

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*PLOS Data Policy*

Please note that as a condition of publication PLOS' data policy (http://journals.plos.org/plosbiology/s/data-availability) requires that you make available all data used to draw the conclusions arrived at in your manuscript. If you have not already done so, you must include any data used in your manuscript either in appropriate repositories, within the body of the manuscript, or as supporting information (N.B. this includes any numerical values that were used to generate graphs, histograms etc.). For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5

*Blot and Gel Data Policy*

We require the original, uncropped and minimally adjusted images supporting all blot and gel results reported in an article's figures or Supporting Information files. We will require these files before a manuscript can be accepted so please prepare them now, if you have not already uploaded them. Please carefully read our guidelines for how to prepare and upload this data: https://journals.plos.org/plosbiology/s/figures#loc-blot-and-gel-reporting-requirements

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

Thank you again for your submission to our journal. We hope that our editorial process has been constructive thus far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Di Jiang, PhD

Associate Editor

PLOS Biology

*****************************************************

REVIEWS:

Reviewer #1:

Previous work from other labs have highlighted the concept of the double blood brain barrier or more accurately, the double-basement membrane structure of cerebral blood vessels (endothelial basement membrane and parenchymal basement membrane made by the glia limitans). Results suggested that penetration of the endothelial and parenchymal barriers are independent steps involving distinct molecular mechanisms.

In this study, the authors investigated this concept in a mouse model with conditional inactivation of Desert Hedghog (Dhh) in basal condition and after immunization by MOG, as a model of experimental encephalomyelitis. Although the concept is interesting, there are several major issues which limit the significance of this paper.

Figure 1:

On panel A, positive and negative controls are missing to assess the purity of the preparation and confirm that these are endothelial cells. The authors jump throughout the paper from murine to human cells without any justification. Please explain.

In situ hybridization on murine brain sections is needed to confirm production of Dhh by endothelial cells.

On panels F-H, it is unclear whether expression has been analyzed in isolated microvessels or in cultured brain endothelial cells. Authors analyzed expression of Dhh 16 days after MOG immunization, which is quite late. Please explain why this late time point has been selected.

Figure 2 and 3

The conclusion that expression of VE-cadherin and claudin 5 is decreased in Dhh KO mice is unfortunately not convincing.

Figure 2: There is a discrepancy between immunostaining (panels D-F) and qRT-PCR or Western blot data, regarding the expression of VE-cadherin and claudin 5. IF pictures show an increased expression of VE-cadherin and claudin 5 in mutant cells whereas qRT-PCR and WB quantifications show the opposite. Moreover, what is the genotype of control mice? Were these mice treated with tamoxifen? Also, same comment as above regarding brain endothelial cells vs isolated microvessels.

Figure 3: On panels A -C, what is missing is a vascular marker to quantify claudin5 and VE-cadherin area over the vessel area, to normalize for the number and length of vessels.

Methodology to analyze the BBB integrity in Dhh KO mice is not appropriate. At least mice must be transcardially perfused to remove intravascular plasma proteins which can passively diffuse in the vessel wall even after fixation (see Mestre et al, Nature Communications 2018). Moreover, because quantification of endogenous fibrinogen or albumin on sections can be challenging and is subjective, analysis of exogenous tracers such as cadaverine of Dextran is requested.

Figure 4

The conclusion that plasma leakage in Dhh KO mice is restricted by the glia limitans to the perivascular space is unfortunately not convincing. What is missing is a labelling with a pan laminin antibody to identify the endothelial and parenchymal basement membranes as described in Song et al, Cell reports 2015. Moreover, aquaporin 4 would have been a better marker of astrocytic endfeet. Indeed, as shown in panel D, GFAP labelling of astrocytic endfeet is inconstant. Also, podocalyxin, which is expressed on the luminal membrane of endothelial cells, cannot delineate the abluminal wall of the vessel.

The conclusion that "BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans » is not supported by the data in the absence of the analysis of other mouse models with BBB leakage. This might depend on the cause of BBB leakage.

Another problem is that authors did not differentiate vessel types in their analysis. The concept of double-basement membrane structure of cerebral blood vessels makes sense in arteries/ arterioles or veins/venules but not in capillaries where the parenchymal and vascular basement membranes have merged.

Figure 5

Authors used VEGF or mannitol to simulate BBB leakage in vitro in cultured cells but there is no attempt to provide evidence that VEGF or mannitol did really something on their cells.

Again, there is a discrepancy between GFAP staining shown in panels E-H and GFAP quantification.

In the in vivo experiment the "hole drilled into the cerebral cortex", as mentioned in the material and methods, is likely to have induced unspecific astrocytic changes

Figure 6

There seems to be a discrepancy between the alleged beneficial effect of inactivating Dhh on demyelination and the comparable lethality, unless the effect is marginal. Please explain.

Figure 7

Panel E shows a diffuse astrocytic reaction in Dhh KO mice whereas fibrinogen staining seems to be restricted around what seems to be an arteriole. How do the authors reconcile this finding with their conclusion?

Also, in panel H vessels from the WT and Dhh KO mice are clearly different, with no perivascular cuff in the WT and a perivascular cuff in the KO mouse. Quantification of the number of CD45+ and T cells inside or outside this cuff is valid, provided that comparable vessels (arteries or veins) are analyzed in each genotype.

Others

In panel B (Fig 4) and in panels F-G, I-J (Fig 7) authors mention % values, but the confidence interval is not indicated. Also, it is very unclear how statistics have been performed. Comparison of % is usually analyzed by a Chi-square test but this is not mentioned in the Method section.

Reviewer #2: This paper by Hollier et al describes the functional emergence of a 2nd barrier formed by the glial limitans (astrocytes) following BBB breakdown in neuroinflammation. Molecularly, show that glial limitans 'tightening' to prevent parenchymal invasion of immune cells is a consequence of signals triggered by BBB breakdown (do this using a genetic model BBB breakdown, conditional deletion of Dhh and in MS animal model, also provide evidence in human MS tissue). This is very intriguing concept and builds off their 2017 JCI paper showing upregulation of astrocytic Cldn4 in MS animal models that helps to functional blocks parenchymal immune cell invasion that promotes disease progression. In general, I think this is a well-done study and of high interest to the fields of BBB mechanisms and neuroinflammatory disease. Below are my moderate concerns about some aspect of their analysis and a list of study limitations that should be addressed in the Discussion.

Major:

Line 115: Perivascular spaces are 'seen' around large penetrating arteriole and venules but I would check the literature carefully as I am not aware of the experimental evidence for PVS around capillary vasculature, which makes up the majority of vessels in the brain. At the capillary level, the astrocytes make direct and essentially continuous contact/coverage of abluminal surface. This becomes important later as the authors are proposing to analyze the PVS - laying out the literature of the existence of a space is important to support the validity of this analysis approach

Ln 191-193, Fig. 2: These are qualitative statements that should be supported by quantitative analysis of junctional disorganization - this data is similar to what was described in their 2018 CR paper, I would be satisfied if they stated this is consistent with previously documented phenotype using siRNA for Dhh (junctional disorganization was quantified in this previous paper and looks similar)

Fig. 2: This group showed something similar in the 2018 paper Circulation Research paper, that paper also documented altered organization of Cdh5 at the junctions (with siRNA) but not a decrease in expression (but could be for experimental reasons) - the reduction of Cdh5 in these mutants is unclear, nor is the model of how Dhh is regulating this. For example they show Cdh5 down in their cultures, the mutants but why? This isn't discussed, just the decrease in Cldn5.

Fig. 3F: Please provide a low magnification image (can show current image as inset) to better demonstrate the scope of elevated ICAM expression. If the authors have this data, the ICAM with a blood vessel marker (PECAM) would help show this is vascular ICAM and how widespread it is in the spinal cord of the mutants.

Fig. 3G: Based on the image, it seems like the increase in GFAP is limited to the white matter, did quantification include all areas of spinal cord (grey and white)? This would be important to state in the methods and state if the increase was or was not regionalized (could do that by segregating/presenting the data by region).

Fig. 4G - There is insufficient detail in the methods as to how this was quantified. Recommend showing higher magnification images, also highlighting using dotted lines outlines where measurements were taken (possibly a separate diagram or image with this information). In it's current form, it's very difficult for me to see how accurate this type of analysis is for differentiated between fibrinogen in and outside the vessel. Also, the 24% in the control is a little concerning, the expectation would this is zero, again leading to some concerns about this method. A limitation of this assay is the that fibrinogen is very large, possible there is a size restriction feature of the glial limitans barrier - (70kda FITC dextran appeared to leak well away from the vessel in the 2018 paper) - two options, perform experiments with a smaller size dye to test this or discuss this as a limitation of their findings in the Discussion (possibly small size proteins could leak through the astrocytic barrier and still cause problems?)

Ln 313-315 Fig. 5K-O: Need to show that VEGFA injection into the cortex causes BBB breakdown - or cite papers that have used this method before and shown this method is effective

Fig. 7 - this data is confusing as presented w/o the non-EAE data - to understand the effects of EAE model on the two genotypes, non-EAE values for each of these parameters should be included on the histograms - in comparing data from Fig 3 and Fig 7, a lot of the same measures were done - was the analysis done the same and therefore could it be added here and stats run to make comparisons?

Ln 373: this is difficult to discern because Cdh5 expression is down, making it difficult to make out what is GFAP expression at glial limitans - remove statement or provide images/data that demonstrates this more clearly

Fig. 7F, G: Similar comment to this type of analysis in Fig. 4A/B, need much better description in methods for how this was done, a diagram or annotated imaging showing where analysis was performed - while it's very clear from the images that fibrinogen is 'trapped' it's not clear how the quantitative data was acquired

Ln 410: A model would be very helpful to bring all the observations of this paper together

Ln 429-431: I looked at Dhh and its expression in this dataset and it's actually very low vs other known endothelial genes - there could be MANY reasons for this, this was all from cortex, maybe there are regional differences, single cell can be tricky to pick up morphogens…I think the data in this manuscript is much more convincing of Dhh expression therefore I caution the authors against citing this as support for Dhh gene expression in brain endothelium

Ln 454-464 - One of the elements missing from this discussion is whether the increase in Cldn4 and astrocyte barrier induction is specific to the Dhh cKO model or is more broadly applicable to any mode of barrier opening - example in pericyte-deficient mice with BBB opening there is some problems with astrocyte-endfeet localization and the barrier opening is due to increased transcytosis no junctional degradation - bringing up these considerations is important

Discussion: Another point that should be addressed in the discussion is whether this is a VEGFA driven process or broadly applicable to other BBB disrupters, example cytokines like IFNg - VEGFA but not mannitol was able to induce secretion of factors that signaled to astrocytes to increase Cldn4, what does this mean?

Discussion: Another limitation to this study that should be discussed is it's not clear if this is a general function of Dhh to regulate the barrier or region specific, all analysis was done in the spinal cord. Further, it's also important to consider if the upregulation of astrocytic Cldn4 is limited to neuroinflammatory disease. This and their previous paper have only looked at EAE in spinal cord. What is the evidence that this could be MS specific phenomenon vs generalizable reaction to BBB opening in other context of other actute of chronic disease (stroke, TBI, AD, ALS, etc). For example, can point to their own data that VEGFA injection into the cortex induced Cldn4 upregulation in response to BBB disruptor VEGFA. Or have other studies looked at this Cldn4 upregulation in response to BBB opening?

Minor:

Ln 109: Edit sentence for clarity; also I would argue that this vision of the NVU has been building for years so suggest removing the term 'recent' and replace with language that reflects a wealth of literature that builds this idea of a multi-cellular 'BBB'.

Ln 114: I would recommend adding a little more context for the reader - maybe a sentence saying this is a highly regulated process (solute transporters and receptor mediated trancytosis) and immune cells are actively prevented from cross by low levels immune receptors that normally permit immune trafficking.

Ln 119: consider a different phrase, perhaps "is more complex"

Ln 185: Please briefly describe how Dhh was deleted from ECs - are these ECs derived from the Cdh5-CreErtl Dhh1-fl/fl mice?

Ln 236: edit "BBB" to say brain endothelium since ICAM is expressed by the endothelial cells not the BBB per se.

Ln 271: "human disease" instead of "Human"

Ln 281: typo?

Ln 420: edit "proved" to say "showed"

Ln 420-424: break up in to two sentences

Reviewer #3: In the present study Chapouly and her team provide convincing evidence for an additional role of the glia limitans in regulating CNS immunity ,when BBB integrity is impaired. They have focused on the role of Dhh and show when deleted from the endothelium in mice, BBB tightness is impaired and at the same time tightness of the glia limitans is increased possibly by the induction of tight junctions in astrocytes. DhhECKO mice develop delayed EAE with CD45+ cells found to be trapped in perivascular spaces underscoring the notion of an increased barrier property of the glia limitans in these mice.

The study combines analysis of human and mouse tissues and the finding of the cross-talk of an impaired BBB leading to improved barrier properties of the glia limitans to immune components is quite novel.

The study as it stands has however a significant number of issues that need to be addressed. Overall the authors seem to tend towards conclusions that are not yet supported by the data as presented. Thus overall downtoning of the manuscript will be required.

I am absolutely aware that in the COVID-19 pandemic a number of issues may not be solvable as they require additional experiments. If additional data cannot be provided the authors need to explain the reason and instead significantly downtone their conclusions.

The authors base their study on the statement that "Presently, it is accepted that the BBB is the sole line of defense of the CNS, restricting access to the parenchyma to inflammatory infiltration notably in the context of chronic neuroinflammation." This is certainly not the case and the authors should include mentioning and discussion of the work of e.g. the Sorokin laboratory, which has provided mechanistic insight into the role of the glia limitans in CNS inflammation. Also Joan A Abbott has written several articles highlighting the role of the glia limitans as a brain barrier. In numerous reviews including those of Weller, Owens, Ransohoff and Engelhardt data on the role of the the glia limitans has been highlighted as barrier for the immune system.

Throughout the manuscript the authors make repeatedly statements as the following that need to be corrected as suggested:

* that the adherens junction Cdh5 and tight junction Cldn5 are down regulated

o downregulation of mRNA of junctional molecules does not allow to make any statement on the integrity or overall molecular composition of a cellular junction

* Isolation of neurovascular unit

o This unit of cellular and acellular components can hardly be isolated

* Gliovascular unit

o what do the authors mean with this expression as compared to neurovascular unit?

o a network of astrocytes termed the Glia Limitans.

o The glia limitans is composed of the parenchymal basement membrane and the astrocyte end-feet rather than a network of astrocytes.

endothelial adherens and tight junction expression is maintained in DhhECKO

this could only be stated when functional and ultrastructural studies confirm this

Comments to the data in the sequence of the manuscript:

In general the authors refer to the gene names when staining or detecting proteins in IF stainings or WB -this is not correct and needs to be corrected.

Figure 1 A shows lack of expression of Shh and Ihh in the healthy brain of mice - this is in apparent contrast to the observations by the Prat lab that has identified astrocytes as source of Shh maintining BBB stability, which is mentioned by the authors. The authors should compare the primer sets used and either repeat the assay or refer to the apparent difference.

Figure 1 B does not show any overlap of VE-cadherin staining with the Dhh staining, nevertheless the authors conclude that Dhh is expressed in CNS endothelium.

Figure 1F to J: Isolation of spinal cord microvessels from mice suffering from EAE will contain other contaminants as microvessels isolated from healthy controls. Lower levels of Dhh, Claudin-5 and ZO-1 from the vascular compartment accompanied by higher levels of ICAM-1, VCAM-1 could also be obtained by coisolation of a higher number of contaminations with astrocyte end-feet, perivascular inflammatory cells or their debris or regulated expression in pericytes and smooth muscle cells. In this context the author refer to their results in Fig 1, I-J showing lower levels of mRNA for claudin-5 and ZO1, as "down regulation of tight junctions"- this is inappropriate as tight junctions are complex cell-cell-contacts including more than describing the mRNA expression of their proteins.

Figure 2A - The authors verify endothelial cell specific deletion of Dhh in vivo by determining mRNA levels for Dhh in endothelial cell cultures derived from these mice. This does of course not allow to verify the efficiency of Dhh depletion in vivo as it does now allow to determine if Dhh has been deleted with different efficiency from different parts of the CNS vasculature. It could thus be that the authors culture endothelial cells from different parts of the vascular tree as they also don't know if lack of Dhh has any effect in survival of the ECs during isolation or their growth in vitro. Furthermore, considering that Dhh is a morphogen it is mandatory to investigate how Cre-mediated deletion of Dhh affects brain angiogenesis.

Figure 3: The authors refer to ICAM-1 and GFAF as markers analyzing the "activation status of both the BBB and Glia Limitans". As outlined several times using a reference molecule to describe an entire tissue or cell is not appropriate. Although ICAM-1 is an excellent marker for EC activation, in the context of the inflamed CNS tissue the authors have to consider that ICAM-1 is also upregulated on pericytes, smooth muscle cells and astrocytes.

Figure 3A - why do the authors show two sets for VE-cadherin/Fibrinogen staining? It would be preferable to show another set e.g. for claudin-5 and albumin staining.

Figure 4B: The authors have used podocalyxin and GFAP to define localization of fibrinogen in the lumen, perivascular space and within the CNS parenchyma. The perivascular space is bordered by basement membranes, which can be detected by pan-laminin stainings which would be preferable for accurate analysis of the localization of fibrinogen. The authors thus either perform additional stainings or significantly downtone this paragraph as GFAP as an intermediate filament, which is also not detectable in all astrocytes does not allow for accurate definition of the glia limitans.

Claudin-4 is suggested to be expressed in GFAP + astrocytes- the staining for both markers looks to a large degree identical, which is confusing as GFAP is an intermediate filament whereas claudin-4 would be expected at the cell membranes and possibly focused to astrocyte end feet. Did the authors check for other claudins and could confirm specificity of this staining?

Figure 4 E: WB signal for claudin-4 looks very weak. A positive control with claudin-4 transfectants would be helpful to show.

Figure 5: The authors delivered VEGF-A or I guess what they mean plasma proteins into the left cerebral cortex of adult mice. They have used a very high volume of 3 μL which in addition was injected in only 30 seconds !!!! into the tissue - this will cause massive damage of the tissue, which needs to be considered when interpreting these data.

Figure 6: The authors show delayed onset of EAE in the DhhEC KO mice. As mentioned above without prior analysis of the impact of endothelial cell specific deletion of Dhh in the entire vasculature of these mice, conclusions from these data are difficult to make. CNS angioarchitecture is the minimum that needs to be studied, but vascular deletion of Dhh could have effects beyond the CNS not considered by the authors but still impacting on EAE which is induced by peripheral activation of T cells that need to travel to the CNS.

Last but not least the manuscript does need editing with the help of a native speaker.

PLoS Biol. 2020 Nov 30;18(11):e3000946. doi: 10.1371/journal.pbio.3000946.r003

Author response to Decision Letter 1

22 Jul 2020

Attachment

Submitted filename: response to reviewers.docx

Click here for additional data file.^{(2.3MB, docx)}

PLoS Biol. doi: 10.1371/journal.pbio.3000946.r004

Decision Letter 2

Gabriel Gasque

24 Sep 2020

Dear Dr Chapouly,

Thank you for submitting your revised Research Article entitled "Blood Brain Barrier genetic disruption leads to protective barrier formation at the Glia Limitans" for publication in PLOS Biology. I have now obtained advice from the original reviewers and have discussed their comments with the Academic Editor. You will note that reviewer 2, Julie Siegenthaler, has revealed her identity.

Based on the reviews, we will probably accept this manuscript for publication, assuming that you will modify the manuscript to address the remaining points raised by the reviewers. Having discussed their specific comments with the Academic Editor, we think that toning down some of the conclusions is sufficient and no new experiments are required.

Please also make sure to address the data and other policy-related requests noted at the end of this email.

We expect to receive your revised manuscript within two weeks. Your revisions should address the specific points made by each reviewer. In addition to the remaining revisions and before we will be able to formally accept your manuscript and consider it "in press", we also need to ensure that your article conforms to our guidelines. A member of our team will be in touch shortly with a set of requests. As we can't proceed until these requirements are met, your swift response will help prevent delays to publication.

To submit your revision, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' to find your submission record. Your revised submission must include the following:

- a cover letter that should detail your responses to any editorial requests, if applicable

- a Response to Reviewers file that provides a detailed response to the reviewers' comments (if applicable)

- a track-changes file indicating any changes that you have made to the manuscript.

*Copyediting*

Upon acceptance of your article, your final files will be copyedited and typeset into the final PDF. While you will have an opportunity to review these files as proofs, PLOS will only permit corrections to spelling or significant scientific errors. Therefore, please take this final revision time to assess and make any remaining major changes to your manuscript.

NOTE: If Supporting Information files are included with your article, note that these are not copyedited and will be published as they are submitted. Please ensure that these files are legible and of high quality (at least 300 dpi) in an easily accessible file format. For this reason, please be aware that any references listed in an SI file will not be indexed. For more information, see our Supporting Information guidelines:

https://journals.plos.org/plosbiology/s/supporting-information

*Published Peer Review History*

Please note that you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*Early Version*

Please note that an uncorrected proof of your manuscript will be published online ahead of the final version, unless you opted out when submitting your manuscript. If, for any reason, you do not want an earlier version of your manuscript published online, uncheck the box. Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us as soon as possible if you or your institution is planning to press release the article.

*Protocols deposition*

Please do not hesitate to contact me should you have any questions.

Sincerely,

Gabriel Gasque, Ph.D.,

Senior Editor,

ggasque@plos.org,

PLOS Biology

------------------------------------------------------------------------

ETHICS STATEMENT:

-- Please in your manuscript the ID number of the experimental protocol approved by your institute’s animal welfare officer and the Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES)

------------------------------------------------------------------------

DATA POLICY:

You may be aware of the PLOS Data Policy, which requires that all data be made available without restriction: http://journals.plos.org/plosbiology/s/data-availability. For more information, please also see this editorial: http://dx.doi.org/10.1371/journal.pbio.1001797

Note that we do not require all raw data. Rather, we ask for all individual quantitative observations that underlie the data summarized in the figures and results of your paper. For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5

These data can be made available in one of the following forms:

1) Supplementary files (e.g., excel). Please ensure that all data files are uploaded as 'Supporting Information' and are invariably referred to (in the manuscript, figure legends, and the Description field when uploading your files) using the following format verbatim: S1 Data, S2 Data, etc. Multiple panels of a single or even several figures can be included as multiple sheets in one excel file that is saved using exactly the following convention: S1_Data.xlsx (using an underscore).

2) Deposition in a publicly available repository. Please also provide the accession code or a reviewer link so that we may view your data before publication.

Regardless of the method selected, please ensure that you provide the individual numerical values that underlie the summary data displayed in the following figure panels: Figures 1C-J, 2A-C, 2H-L, 3C-F, 3I-K, 4CGH, 5A-D, 5IJNO, 6A-D, 6F, 7A-D, 7FHJ, S1F, S2BC, S4D, S6CD, and S7A.

NOTE: the numerical data provided should include all replicates AND the way in which the plotted mean and errors were derived (it should not present only the mean/average values).

Please also ensure that each figure legend in your manuscript include information on where the underlying data can be found and that your supplemental data file/s has/have a legend.

Please ensure that your Data Statement in the submission system accurately describes where your data can be found.

------------------------------------------------------------------------

BLOT AND GEL REPORTING REQUIREMENTS:

For manuscripts submitted on or after 1st July 2019, we require the original, uncropped and minimally adjusted images supporting all blot and gel results reported in an article's figures or Supporting Information files. We will require these files before a manuscript can be accepted so please prepare and upload them now. Please carefully read our guidelines for how to prepare and upload this data: https://journals.plos.org/plosbiology/s/figures#loc-blot-and-gel-reporting-requirements

------------------------------------------------------------------------

Reviewer remarks:

Reviewer #1: In their revised manuscript, Mora et al., have added new data and clarified some issues. However, important concerns remain to be resolved.

In figure 2, the finding that CLDN5 is downregulated in MEC from Dhh KO mice is still not convincing. Contrary to what is mentioned in the rebuttal, expression of CLDN5 has not been analyzed by immunoblot and the panel F clearly shows an upregulation of CLDN5.

The conclusion that plasma leakage in Dhh KO mice is restricted by the glia limitans to the PVS is still not convincing. 1) In figure 4, panel A, there is clearly an uptake of IgG in the astrocyte labeled by the anti-aquaporin 4 antibody. 2°) If the glia limitans restricts plasma leakage in DhhKO mice, then fibrinogen staining in Figure 3 panel shall be exclusively detected in the PVS around arteries or veins but not around capillaries, where there is a single composite basement membrane produced by endothelial cells, pericytes and astrocytic and no PVS. This does not seem to be the case in this figure 3, quality of which is quite poor.

On page 11, authors claim that Dhh inactivation at the BBB drives endothelial and astrocytic activation. Since they did not identify astrocytic activation in the gray matter (figure 3, panels H and K), this implies that there is no BBB leakage in the gray matter but this is not shown. Otherwise, this conclusion shall be toned down.

In general the quality of immunofluorescence images is rather low and should be improved.

Reviewer #2, Julie Siegenthaler: This is a revised version of Hollier et al probing the molecular basis of a 2nd barrier system formed by the glial endfeet in response to neuroinflammation. The authors have done an excellent job addressing my previous comments, including greatly improved or completely new quantitative analysis (GFAP in SC sections Fig. 3, IgG leakage in perivascular spaces Fig. 4/Supp Fig. 4, Fig. 7), also adding a helpful diagram to bring the data in non-disease vs EAE together for the reader. The authors also added new discussion points requested that are important to understand 1) is BBB breakdown a universal inducer of Cldn4 and 'strengthening' the glial limitans, 2) is the induction of Cldn4 specific to VEGFA-mediated disruption or more broadly applicable to other cytokines and 3) is this a MS/EAE phenomenon.

I did note a typo in Figure 3J should be 'White' and the text in the summary figure is rather small, enlarging would help.

I read through the other reviewer comments and the authors' response and I want to commend the authors for their efforts in addressing what were VERY extensive comments, especially considering lab shut downs and general pandemic related upheaval. Great job!

Reviewer #3: The authors have thoroughly answered all queries. They should however add the source of the VE-cadherin-CreERT2 transgenic mouse in M&M.

PLoS Biol. 2020 Nov 30;18(11):e3000946. doi: 10.1371/journal.pbio.3000946.r005

Author response to Decision Letter 2

19 Oct 2020

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file.^{(195.7KB, docx)}

PLoS Biol. doi: 10.1371/journal.pbio.3000946.r006

Decision Letter 3

Gabriel Gasque

22 Oct 2020

Dear Dr Chapouly,

On behalf of my colleagues and the Academic Editor, Richard Daneman, I am pleased to inform you that we will be delighted to publish your Research Article in PLOS Biology.

PRODUCTION PROCESS

Before publication you will see the copyedited word document (within 5 business days) and a PDF proof shortly after that. The copyeditor will be in touch shortly before sending you the copyedited Word document. We will make some revisions at copyediting stage to conform to our general style, and for clarification. When you receive this version you should check and revise it very carefully, including figures, tables, references, and supporting information, because corrections at the next stage (proofs) will be strictly limited to (1) errors in author names or affiliations, (2) errors of scientific fact that would cause misunderstandings to readers, and (3) printer's (introduced) errors. Please return the copyedited file within 2 business days in order to ensure timely delivery of the PDF proof.

If you are likely to be away when either this document or the proof is sent, please ensure we have contact information of a second person, as we will need you to respond quickly at each point. Given the disruptions resulting from the ongoing COVID-19 pandemic, there may be delays in the production process. We apologise in advance for any inconvenience caused and will do our best to minimize impact as far as possible.

EARLY VERSION

The version of your manuscript submitted at the copyedit stage will be posted online ahead of the final proof version, unless you have already opted out of the process. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with biologypress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

Thank you again for submitting your manuscript to PLOS Biology and for your support of Open Access publishing. Please do not hesitate to contact me if I can provide any assistance during the production process.

Kind regards,

Erin O'Loughlin

Publishing Editor,

PLOS Biology

on behalf of

Gabriel Gasque,

Senior Editor

PLOS Biology

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Text

Supporting information file containing the S1 Data and S2 Data legends and DOI links (A), the Supporting Methods (B), and the associated References (C).

(DOCX)

Click here for additional data file.^{(29KB, docx)}

S1 Raw Images

(PDF)

Click here for additional data file.^{(17.6MB, pdf)}

S1 Fig. (Related to Fig 1 and Fig 2).

(TIF)

Click here for additional data file.^{(7.2MB, tif)}

S2 Fig. (Related to Fig 3, Fig 4, Fig 6 and Fig 7).

(TIF)

Click here for additional data file.^{(8.9MB, tif)}

S3 Fig. Endothelial-specific Dhh inactivation induces BBB permeability in vivo.

(TIF)

Click here for additional data file.^{(2.7MB, tif)}

S4 Fig. (Related to Fig 4).

(TIF)

Click here for additional data file.^{(3.1MB, tif)}

S5 Fig. (Related to Fig 4).

(TIF)

Click here for additional data file.^{(1.5MB, tif)}

S6 Fig. (Related to Fig 4).

(TIF)

Click here for additional data file.^{(453.8KB, tif)}

S7 Fig. (Related to Fig 5 and Fig 7).

(TIF)

Click here for additional data file.^{(5.5MB, tif)}

Attachment

Submitted filename: response to reviewers.docx

Click here for additional data file.^{(2.3MB, docx)}

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file.^{(195.7KB, docx)}

Data Availability Statement

[pbio.3000946.ref001] 1.McConnell HL, Kersch CN, Woltjer RL, Neuwelt EA. The Translational Significance of the Neurovascular Unit. J Biol Chem. 2017;292(3):762–70. 10.1074/jbc.R116.760215 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref002] 2.Bartholomäus I, Kawakami N, Odoardi F, Schläger C, Miljkovic D, Ellwart JW, et al. Effector T cell interactions with meningeal vascular structures in nascent autoimmune CNS lesions. Nature. 2009;462 (7269):94–98. 10.1038/nature08478 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref003] 3.Brück W, Bitsch A, Kolenda H, Brück Y, Stiefel M, Lassmann H. Inflammatory central nervous system demyelination: correlation of magnetic resonance imaging findings with lesion pathology. Ann Neurol. 1997;42(5):783–793. 10.1002/ana.410420515 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref004] 4.Alvarez JI, Dodelet-Devillers A, Kebir H, Ifergan I, Fabre PJ, Terouz S, et al. The Hedgehog pathway promotes blood-brain barrier integrity and CNS immune quiescence. Science. 2011;334 (6063):1727–1731. 10.1126/science.1206936 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref005] 5.Singh VB, Singh MV, Gorantla S, Poluektova LY, Maggirwar SB. Smoothened Agonist Reduces Human Immunodeficiency Virus Type-1-Induced Blood-Brain Barrier Breakdown in Humanized Mice. Sci Rep. 2016;6:26876 10.1038/srep26876 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref006] 6.Xia Y, He Q, Li Y, Chen S, Huang M, Wang Y, et al. Recombinant human sonic hedgehog protein regulates the expression of ZO-1 and occludin by activating angiopoietin-1 in stroke damage. PLoS ONE. 2013;8 (7):e68891 10.1371/journal.pone.0068891 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref007] 7.Chapouly C, Guimbal S, Hollier P-L, Renault M-A. Role of Hedgehog Signaling in Vasculature Development, Differentiation, and Maintenance. Int J Mol Sci. 2019;20(12). [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref008] 8.Caradu C, Couffinhal T, Chapouly C, Guimbal S, Hollier P-L, Ducasse E, et al. Restoring Endothelial Function by Targeting Desert Hedgehog Downstream of Klf2 Improves Critical Limb Ischemia in Adults. Circ Res. 2018;123 (9):1053–1065. 10.1161/CIRCRESAHA.118.313177 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref009] 9.Bitgood MJ, McMahon AP. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev Biol. 1995;172 (1):126–138. 10.1006/dbio.1995.0010 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref010] 10.Robbins DJ, Fei DL, Riobo NA. The Hedgehog signal transduction network. Sci Signal. 2012;5(246):re6 10.1126/scisignal.2002906 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref011] 11.Iadecola C. The Neurovascular Unit Coming of Age: A Journey through Neurovascular Coupling in Health and Disease. Neuron. 2017;96 (1):17–42. 10.1016/j.neuron.2017.07.030 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref012] 12.Sweeney MD, Ayyadurai S, Zlokovic BV. Pericytes of the neurovascular unit: key functions and signaling pathways. Nat Neurosci. 2016;19(6):771–83. 10.1038/nn.4288 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref013] 13.Cheslow L, Alvarez JI. Glial-endothelial crosstalk regulates blood-brain barrier function. Curr Opin Pharmacol. 2016;26:39–46. 10.1016/j.coph.2015.09.010 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref014] 14.Engelhardt B, Coisne C. Fluids and barriers of the CNS establish immune privilege by confining immune surveillance to a two-walled castle moat surrounding the CNS castle. Fluids Barriers CNS. 2011;8(1):4 10.1186/2045-8118-8-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref015] 15.Engelhardt B, Ransohoff RM. Capture, crawl, cross: the T cell code to breach the blood-brain barriers. Trends Immunol. 2012;33 (12):579–589. 10.1016/j.it.2012.07.004 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref016] 16.Colombo E, Farina C. Astrocytes: Key Regulators of Neuroinflammation. Trends Immunol. 2016;37 (9):608–620. 10.1016/j.it.2016.06.006 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref017] 17.Chapouly C, Tadesse Argaw A, Horng S, Castro K, Zhang J, Asp L, et al. Astrocytic TYMP and VEGFA drive blood-brain barrier opening in inflammatory central nervous system lesions. Brain J Neurol. 2015;138 (Pt 6):1548–1567. 10.1093/brain/awv077 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref018] 18.Argaw AT, Asp L, Zhang J, Navrazhina K, Pham T, Mariani JN, et al. Astrocyte-derived VEGF-A drives blood-brain barrier disruption in CNS inflammatory disease. J Clin Invest. 2012;122 (7):2454–2468. 10.1172/JCI60842 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref019] 19.Argaw AT, Gurfein BT, Zhang Y, Zameer A, John GR. VEGF-mediated disruption of endothelial CLN-5 promotes blood-brain barrier breakdown. Proc Natl Acad Sci U S A. 2009;106 (6):1977–1982. 10.1073/pnas.0808698106 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref020] 20.Engelhardt B, Carare RO, Bechmann I, Flügel A, Laman JD, Weller RO. Vascular, glial, and lymphatic immune gateways of the central nervous system. Acta Neuropathol (Berl). 2016;132 (3):317–338. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref021] 21.Di Russo J, Hannocks M-J, Luik A-L, Song J, Zhang X, Yousif L, et al. Vascular laminins in physiology and pathology. Matrix Biol. 2017;57–58:140–148. 10.1016/j.matbio.2016.06.008 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref022] 22.Abbott NJ, Rönnbäck L, Hansson E. Astrocyte-endothelial interactions at the blood-brain barrier. Nat Rev Neurosci. 2006;7 (1):41–53. 10.1038/nrn1824 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref023] 23.Iliff JJ, Wang M, Liao Y, Plogg BA, Peng W, Gundersen GA, et al. A Paravascular Pathway Facilitates CSF Flow Through the Brain Parenchyma and the Clearance of Interstitial Solutes, Including Amyloid β. Sci Transl Med. 2012;4 (147):147ra111–147ra111. 10.1126/scitranslmed.3003748 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref024] 24.Song J, Wu C, Korpos E, Zhang X, Agrawal SM, Wang Y, et al. Focal MMP-2 and MMP-9 Activity at the Blood-Brain Barrier Promotes Chemokine-Induced Leukocyte Migration. Cell Rep. 2015;10 (7):1040–1054. 10.1016/j.celrep.2015.01.037 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref025] 25.Cruz-Orengo L, Holman DW, Dorsey D, Zhou L, Zhang P, Wright M, et al. CXCR7 influences leukocyte entry into the CNS parenchyma by controlling abluminal CXCL12 abundance during autoimmunity. J Exp Med. 2011;208 (2):327–339. 10.1084/jem.20102010 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref026] 26.Horng S, Therattil A, Moyon S, Gordon A, Kim K, Argaw AT, et al. Astrocytic tight junctions control inflammatory CNS lesion pathogenesis. J Clin Invest. 2017;127 (8):3136–3151. 10.1172/JCI91301 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref027] 27.Taddei A, Giampietro C, Conti A, Orsenigo F, Breviario F, Pirazzoli V, et al. Endothelial adherens junctions control tight junctions by VE-cadherin-mediated upregulation of claudin-5. Nat Cell Biol. 2008;10 (8):923–934. 10.1038/ncb1752 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref028] 28.Davalos D, Akassoglou K. Fibrinogen as a key regulator of inflammation in disease. Semin Immunopathol. 2012;34 (1):43–62. 10.1007/s00281-011-0290-8 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref029] 29.Nitta T, Yagita H, Sato K, Okumura K. Expression of Fc gamma receptors on astroglial cell lines and their role in the central nervous system. Neurosurgery. 1992;31(1):83–7; discussion 87–88. 10.1227/00006123-199207000-00012 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref030] 30.Okun E, Mattson MP, Arumugam TV. Involvement of Fc Receptors in Disorders of the Central Nervous System. Neuromolecular Med. 2010;12 (2):164–178. 10.1007/s12017-009-8099-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref031] 31.Gurfein BT, Zhang Y, López CB, Argaw AT, Zameer A, Moran TM, et al. IL-11 regulates autoimmune demyelination. J Immunol Baltim Md 1950. 2009;183 (7):4229–4240. 10.4049/jimmunol.0900622 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref032] 32.Wanner IB, Anderson MA, Song B, Levine J, Fernandez A, Gray-Thompson Z, et al. Glial scar borders are formed by newly proliferated, elongated astrocytes that interact to corral inflammatory and fibrotic cells via STAT3-dependent mechanisms after spinal cord injury. J Neurosci. 2013;33 (31):12870–12886. 10.1523/JNEUROSCI.2121-13.2013 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref033] 33.Sofroniew MV. Astrocyte barriers to neurotoxic inflammation. Nat Rev Neurosci. 2015;16 (5):249–263. 10.1038/nrn3898 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref034] 34.Hewitt KJ, Agarwal R, Morin PJ, et al. BMC Cancer. 2006;6:186 10.1186/1471-2407-6-186 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref035] 35.Acharya P, Beckel J, Ruiz WG, Wang E, Rojas R, Birder L, et al. Distribution of the tight junction proteins ZO-1, occludin, and claudin-4, -8, and -12 in bladder epithelium. Am J Physiol Renal Physiol. 2004;287 (2):F305–F318. 10.1152/ajprenal.00341.2003 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref036] 36.Bruewer M, Luegering A, Kucharzik T, Parkos CA, Madara JL, Hopkins AM, et al. Proinflammatory Cytokines Disrupt Epithelial Barrier Function by Apoptosis-Independent Mechanisms. J Immunol. 2003;171 (11):6164–6172. 10.4049/jimmunol.171.11.6164 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref037] 37.Kebir H, Kreymborg K, Ifergan I, Dodelet-Devillers A, Cayrol R, Bernard M, et al. Human TH17 lymphocytes promote blood-brain barrier disruption and central nervous system inflammation. Nat Med. 2007;13 (10):1173–1175. 10.1038/nm1651 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref038] 38.Kinugasa T, Sakaguchi T, Gu X, Reinecker H. Claudins regulate the intestinal barrier in response to immune mediators. Gastroenterology. 2000;118 (6):1001–1011. 10.1016/s0016-5085(00)70351-9 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref039] 39.Fujita H, Chalubinski M, Rhyner C, Indermitte P, Meyer N, Ferstl R, et al. Claudin-1 expression in airway smooth muscle exacerbates airway remodeling in asthmatic subjects. J Allergy Clin Immunol. 2011;127 (6):1612–1621.e8. 10.1016/j.jaci.2011.03.039 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref040] 40.Konstantin G, Guillem G, Annika A, Christer B. Endothelial-Mural Cell Signaling in Vascular Development and Angiogenesis. Arterioscler Thromb Vasc Biol. 2009;29 (5):630–638. 10.1161/ATVBAHA.107.161521 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref041] 41.Lindahl P, Johansson BR, Levéen P, Betsholtz C. Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science. 1997;277 (5323):242–245. 10.1126/science.277.5323.242 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref042] 42.Yunoki M, Kanda T, Suzuki K, Uneda A, Hirashita K, Yoshino K. Ischemic Tolerance of the Brain and Spinal Cord: A Review. Neurol Med Chir (Tokyo). 2017;57 (11):590–600. 10.2176/nmc.ra.2017-0062 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref043] 43.Kirino T, Tsujita Y, Tamura A. Induced tolerance to ischemia in gerbil hippocampal neurons. J Cereb Blood Flow Metab. 1991;11 (2):299–307. 10.1038/jcbfm.1991.62 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref044] 44.Koizumi S, Hirayama Y, Morizawa YM. New roles of reactive astrocytes in the brain; an organizer of cerebral ischemia. Neurochem Int. 2018;119:107–114. 10.1016/j.neuint.2018.01.007 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref045] 45.Cai W, Zhang K, Li P, Zhu L, Xu J, Yang B, et al. Dysfunction of the neurovascular unit in ischemic stroke and neurodegenerative diseases: An aging effect. Ageing Res Rev. 2017;34:77–87. 10.1016/j.arr.2016.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref046] 46.Hou J, Baker LA, Zhou L, Klein RS. Viral interactions with the blood-brain barrier: old dog, new tricks. Tissue Barriers. 2016;4 (1):e1142492 10.1080/21688370.2016.1142492 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref047] 47.Zlokovic BV. The blood-brain barrier in health and chronic neurodegenerative disorders. Neuron. 2008;57 (2):178–201. 10.1016/j.neuron.2008.01.003 [DOI] [PubMed] [Google Scholar]

[pbio.3000946.ref048] 48.Chacko A-M, Li C, Pryma DA, Brem S, Coukos G, Muzykantov V. Targeted delivery of antibody-based therapeutic and imaging agents to CNS tumors: crossing the blood-brain barrier divide. Expert Opin Drug Deliv. 2013;10 (7):907–926. 10.1517/17425247.2013.808184 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref049] 49.Liu J, Zhao ML, Brosnan CF, Lee SC. Expression of type II nitric oxide synthase in primary human astrocytes and microglia: role of IL-1beta and IL-1 receptor antagonist. J Immunol Baltim Md 1950. 1996;157 (8):3569–3576. [PubMed] [Google Scholar]

[pbio.3000946.ref050] 50.Karki S, Farb MG, Ngo DTM, Myers S, Puri V, Hamburg NM, et al. Forkhead box O-1 modulation improves endothelial insulin resistance in human obesity. Arterioscler Thromb Vasc Biol. 2015;35 (6):1498–1506. 10.1161/ATVBAHA.114.305139 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pbio.3000946.ref051] 51.Hess DC, Howard E, Cheng C, Carroll J, Hill WD, Hsu CY. Hypertonic mannitol loading of NF-kappaB transcription factor decoys in human brain microvascular endothelial cells blocks upregulation of ICAM-1. Stroke. 2000;31 (5):1179–1186. 10.1161/01.str.31.5.1179 [DOI] [PubMed] [Google Scholar]

PERMALINK

Blood–brain barrier genetic disruption leads to protective barrier formation at the Glia Limitans

Pierre Mora

Pierre-Louis Hollier

Sarah Guimbal

Alice Abelanet

Aïssata Diop

Lauriane Cornuault

Thierry Couffinhal

Sam Horng

Alain-Pierre Gadeau

Marie-Ange Renault

Candice Chapouly

Roles

Abstract

Introduction

Results

DHH, but not SHH or IHH, is expressed by CNS microvascular endothelial cells and down-regulated during chronic neuroinflammation

Fig 1. DHH is expressed by CNS MECs and down-regulated during chronic neuroinflammation.

Endothelial-specific Dhh inactivation induces down-regulation of adherens junction CDH5 and tight junction CLDN5 ex vivo

Fig 2. Endothelial-specific Dhh inactivation induces adherens and tight junction down-regulation through stimulation of FOXO1 transcriptional activity.

DHH induces CLDN5 up regulation through the inhibition of FOXO1 transcriptional activity in vitro

In the white matter, endothelial-specific Dhh inactivation induces BBB permeability associated with endothelial and astrocytic activation in vivo

Fig 3.

DhhECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans

Fig 4. DhhECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans.

Endothelial signals can drive astrocyte barrier properties at the Glia Limitans

Fig 5. Endothelial signals can drive astrocyte barrier properties at the Glia Limitans.

Mice with endothelial Dhh inactivation display reduced disability in a model of multiple sclerosis during the onset of the disease

Fig 6. DhhECKO mice display reduced disability in a model of multiple sclerosis during the onset of the disease.

Mice with endothelial Dhh inactivation display a reinforced barrier at the Glia Limitans, restraining access to the parenchyma to inflammatory infiltrate in a model of multiple sclerosis

Fig 7. Mice with endothelial Dhh knockdown display a reinforced barrier at the Glia Limitans restraining access to the parenchyma to inflammation in a model of multiple sclerosis.

Fig 8. Schematic of the BBB and Glia Limitans in DhhECKO versus control mice, in health and inflammatory disease.

Discussion

Methods

Human tissues

Mice

Ethical statement

Neurovascular fraction enrichment from mouse CNS

Neurovascular fraction enrichment for RT-PCR, western blots, or immunohistochemistry

Primary culture of mouse CNS micro vascular endothelial cells (CNS MECs)

Cell culture

Cytokines/growth factors/chemicals

Antibodies

Quantitative RT-PCR

Table 1. List of primers used for RT qPCR.

Western blots

Stereotactic injection

Experimental autoimmune encephalomyelitis (EAE)

Immunohistochemistry

Morphometric analysis

Statistical analyses

Supporting information

Acknowledgments

Abbreviations

Data Availability

Funding Statement

References

Decision Letter 0

Di Jiang, PhD

Roles

Decision Letter 1

Di Jiang, PhD

Roles

Author response to Decision Letter 1

Decision Letter 2

Gabriel Gasque

Roles

Author response to Decision Letter 2

Decision Letter 3

Gabriel Gasque

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Dhh^ECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans

Fig 4. Dhh^ECKO-induced BBB breakdown is sufficient to induce a secondary CNS protective barrier at the Glia Limitans.

Fig 6. Dhh^ECKO mice display reduced disability in a model of multiple sclerosis during the onset of the disease.

Fig 8. Schematic of the BBB and Glia Limitans in Dhh^ECKO versus control mice, in health and inflammatory disease.