Originally posted 1 December 2011; corrected 9 December 2011

www.sciencemag.org/cgi/content/full/science.1206936/DC1

Supporting Online Material for

The Hedgehog Pathway Promotes Blood-Brain Barrier Integrity and CNS Immune Quiescence

Jorge Ivan Alvarez, Aurore Dodelet-Devillers, Hania Kebir, Igal Ifergan, Pierre J. Fabre, Simone Terouz, Mike Sabbagh, Karolina Wosik, Lyne Bourbonnière, Monique Bernard,

Jack van Horssen, Helga E. de Vries, Frédéric Charron, Alexandre Prat*

*To whom correspondence should be addressed. E-mail: a.prat@umontreal.ca

Published 1 December 2011 on Science Express DOI: 10.1126/science.1206936

This PDF file includes:

Materials and Methods SOM Text Figs. S1 to S9 Full References

Correction 9 December 2011: The correct reference list is included.

www.sciencemag.org/cgi/content/full/science.1206936/DC1

Supporting Online Material for

The Hedgehog Pathway Promotes Blood-Brain Barrier Integrity and CNS Immune Quiescence

Jorge Ivan Alvarez, Aurore Dodelet-Devillers, Hania Kebir, Igal Ifergan, Pierre J. Fabre, Simone Terouz, Mike Sabbagh, Karolina Wosik, Lyne Bourbonnière, Monique Bernard,

Jack van Horssen, Helga E. de Vries, Frédéric Charron, Alexandre Prat*

*To whom correspondence should be addressed. E-mail: a.prat@umontreal.ca

Published 1 December 2011 on Science Express DOI: 10.1126/science.1206936

This PDF file includes

Materials and Methods SOM Text Figs. S1 to S9 All References

2

Materials and Methods Primary cell isolation & culture

Human adult CNS tissue was obtained from temporal lobe resection paths of patients undergoing surgery for intractable epilepsy, as previously described (36-39). Informed consent and ethic approval were given prior to surgery (ethic approval number HD04.046). BBB-ECs were isolated and grown in medium composed of M199 (Invitrogen) supplemented with 10% fetal bovine serum (FBS), 20% human serum (HS), EC growth supplement (5 mg/ml) and insulin-selenium transferrin premix on 0.5% gelatin-coated tissue culture flasks (reagents from Sigma). Expression of EC markers von Willebrand factor and Ulex europaeus agglutinin I binding sites until passage (P) 7-8 were confirmed. No immunoreactivity for α-myosin or glial fibrillary acidic protein (GFAP) could be detected, confirming the absence of contaminating smooth muscle cells and astrocytes. Human fetal astrocyte (HFA) isolation has previously been described (40). Briefly, fetal CNS tissue was obtained at 16-22 weeks of gestation following CIHR-approved guidelines. HFAs were cultured in complete Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% FBS, glutamax, penicillin and streptomycin (Sigma). Cultures were determined to be >90% pure by GFAP immunostaining, and used between P3-6. Once a week, ACM was harvested from confluent flasks, filtered and added to BBB-EC culture media when specified (41). Mouse CNS microvessels were isolated from minced pieces of brain and spinal cord obtained from WT and Tie2-Cre; smoc/c animals. CNS pieces were homogenized, washed with PBS and then centrifuged in dextran 30% (v/v) at 4000 g for 30 min. Supernatant was discarded and the pellet containing the BVs was washed with PBS and then processed for protein extraction. Hh agonist and antagonists hrShh (0.1 μg/ml, Sigma), Smo agonist Purmorphamine (1 μM) and ACM 40% (v/v) were used as Hh stimulating conditions. Cyclopamine (3-30 μM, Sigma), SANT-1 (0.1 μM) and the 5E1 monoclonal antibody against Shh (5 μg/mL) (Developmental Studies Hybridoma Bank, U of Iowa) were used as Hh abrogating conditions. For in vivo experiments Cyclopamine (10 mg/kg) and GDC-0449 (25 mg/kg) were used as Hh antagonists. Mice Eight to nine week-old female C57BL/6 mice (Charles River, Montreal, QC) were used for in vivo permeability assays and EAE. Smoc mice were generated by AP McMahon (42) obtained through Jax and bred on a mixed background. Tie2-Cre mice were generated by M Yanagisawa (43), obtained through Jax, and bred on a C57BL/6 background.

Immunocytofluorescence Primary cultures of BBB-ECs and HFAs were grown to confluency and fixed with 4% paraformaldehyde (PFA) or methanol. Slides were subsequently permeabilized with 0.3% triton X-100 for 5 min at room temperature (RT) and blocked by incubation in 10% goat serum (Sigma) for 1h at RT. Incubation with rat anti-Shh (R&D Systems), rabbit anti-

3

Ptch-1 (Santa Cruz Biotechnology) and rabbit anti-Smo (MBL) were done overnight at 4˚C. Following washes, the samples were incubated with biotin-conjugated anti-rat and anti-rabbit antibodies (Abs) (DakoCytomation) followed by Cy3- or FITC-labeled streptavidin (Jackson ImmunoResearch) respectively. HFAs were stained with a human anti-GFAP Cy3-conjugated Ab (Sigma). Nuclei were stained with TO-PRO3 (Molecular Probes). In all cases, control stainings were performed using isotype controls and no immunopositive cells could be detected. Stainings were visualized using a Leica SP5 confocal microscope (Leica Microsystems) and analyzed with the Leica LAS AF software. Expression of Hh receptors Ptch-1 and Smo was also assessed on CD14+ monocytes and CD4+ CD45RO+ T cultured in vitro. For that purpose, cytospined leukocytes were permeabilized, fixed with cold acetone (for 10 minutes at –20°C), immunostained and analyzed as described above.

Tissues and in situ immunostaining Brain tissue was obtained from 13 patients with clinically diagnosed and neuropathologically confirmed MS. Autopsy samples were immediately frozen in liquid nitrogen. Three cases without neurological disease were selected as controls. Control tissues were taken from the subcortical white matter, corpus callosum and cortex. White matter MS tissue samples were selected on the basis of post-mortem MRI and lesions were classified according to standard histopathological criteria, as previously published by our group (36-38). All patients and controls, or their next of kin, had given informed consent for autopsy and use of their brain tissue for research purposes. Ethic approval was given prior to autopsy. Immunostainings in Fig 5C and D, are representative of >50 lesions and NAWM sections studied from n=13 donors. The generation of Shh deficient mice has been described previously (44). Mouse embryos (either Shh−/−, Shh+/−, Tie2-Cre; smoc/c, Tie2-Cre; smo+/-, Tie2-Cre; smo+/+ and WT) were extracted and snapped frozen. Similarly, the P4, P19 and 8 wk old mice from Tie2-Cre; smoc/c and Tie2-Cre; smo+/- as well as the C57Bl/6 adult mice used for the in vivo experiments, were perfused with 5 ml of PBS and brain, spleen and liver were removed and quickly snap frozen. For immunohistochemical and fluorescent staining, 5-8 μm cryosections were cut, air dried and fixed in cold acetone for 10 min. Sections were preincubated for 30 min at RT with 10% serum from the species in which the secondary Ab was raised. Afterwards, sections were incubated overnight at 4˚C with primary Abs for MHC-II (DakoCytomation), Shh, Gli-1 (R&D Systems), Ptch-1, Smo (Santa Cruz Biotechnology), fibrinogen (Innovative Research), Apo-lipoprotein B (Abcam), mouse IgGs (Jackson Immunoresearch), laminin (DakoCytomation), PECAM-1, VE-cadherin, F4/80, CD4 (BD Biosciences), occludin, claudin-3, claudin-5, ZO-1 (Invitrogen), p120 (Santa Cruz Biotech), PDGFR-β (R&D systems) and GFAP-Cy3 (Sigma). Subsequently, for immunohistochemistry, sections were incubated with appropriate secondary biotin-labeled Abs for 1h at RT and with an avidin-biotin complex (DakoCytomation) according to the manufacturer's description. Diaminobenzidine (DAB) was used as chromogen. Between the incubation steps, sections were thoroughly washed with PBS. After a short rinse in tap water, preparations were incubated with Harris hematoxylin (Sigma) for 1 min and extensively washed in tap water for 10 min. Finally, sections were dehydrated

4

with ethanol, followed by xylene and mounted with Entellan (Canemco & Marivac). For immunohistofluorescence, sections were incubated with appropriate secondary Cy3 (DakoCytomation) and Alexa 488 (Invitrogen) labeled Abs for 1h at RT. FITC-conjugated Lycopersicon Esculentum (Tomato) lectin (Vector Laboratories) was used to provide a frame of reference for the localization of BVs and nuclei were stained with Hoechst 33258 (Invitrogen) and/or with TOPRO-3 (Molecular Probes). In all cases, Abs were diluted in PBS containing 3% blocking serum, which also served as a negative control. Stainings were visualized using either a Leica DM6000 (Leica Microsystems) microscope equipped with the OpenLab software (PerkinElmer) or with the Leica SP5 confocal microscope. Image acquisition for comparative confocal microscopy was performed using the same parameters between samples (scanning time, laser intensity, gain, contrast, pinhole size). For image analysis, the same parameters were used when acquiring images to allow direct comparison of pixel intensity values recorded using ImageJ (NIH) and the LAS AF software (Leica). Histological stainings For the luxol fast blue and hematoxylin-eosin (LHE) staining, the sections were fixed in formaldehyde 10% for 30 min at room temperature. The slides were then washed in PBS for 3 min, followed by an overnight incubation in luxol fast blue at 60ºC. Differentiation was done with a hydroquinone and sodium sulfite solution in water. Following 2-3 washes in water, sections were immersed in Harris Hematoxylin for 10 min. Acid alcohol was used for differentiation of the hematoxylin stain and the sections were then blued in NaOH 0.05%. After that, samples were dehydrated in ethanol 95% and immersed in alcoholic Eosin Y for 3 min. This was followed by dehydration in absolute ethanol 2 times for 1 min, toluene for 1 min and mounting with Entellan (Merck – Darmstadt, Germany). Peroxidase staining was performed to detect erythrocytes in frozen sections of E14, E18 and P19 murine CNS. The sections were fixed in acetone and then hydrated in PBS. Peroxidase activity was revealed with DAB and sections were subsequently treated as described above for immunohistochemistries. Murine spleen sections were used as positive controls. Detection of ferrous iron in murine tissues was performed with the Turnbull staining. For this purpose, frozen CNS sections were fixed in formaldehyde 10% for 30 min, washed in water and then treated with ferrocyanide-HCl for 10 min at 56°C. The tissues were then extensively rinse in water for 5 min and counterstained with safranin during 3 min. The sections were washed in ethanol 70% and then dehydrated and mounted with Entellan as described above. Imaging of these stainings was performed with a Leica DM6000 microscope using the Improvision OpenLab 4.0.4 software (PerkinElmer – Weltham, MA). Images were processed using Adobe Photoshop CS2 (Adobe – Mountainview, CA).

Reverse-transcription and real-time quantitative polymerase reaction Cellular RNA was isolated using the TRizol reagent (Invitrogen) and Qiagen RNeasy Mini extraction kit, according to the manufacturers’ instructions. RNA concentration and purity were determined on a spectrophotometer Ultrospec 2100 pro (GE Healthcare) at 260 nm. Samples were then treated with DNAse I (Amersham Biosciences) for 10 min at

5

37˚C followed by 10 min at 75˚C. Complementary DNA (cDNA) was obtained by mixing 3 µg of RNA with 3.3 µM random hexamer primers, 3mM dNTPs, RNase-OUT recombinant ribonuclease inhibitor, 3 nM DTT and 400 U Maloney murine leukemia virus (MMLV)-reverse transcriptase for 1h at 42˚C and 10 min at 75˚C (Invitrogen). For polymerase reactions (PCRs) 2 µl of cDNA were amplified using 5 U Taq DNA polymerase, 1.5 mM MgCl2, 0.2 mM dNTPs and 50 pmol of forward (F) and reverse (R) primers. Primers used were: Shh; F: 5’-AAAAGCTGACCCCTTTAGCC-3’; R: 5’-CTCTGAGTGGTGGCCATCTT-3’; Ptch-1; F: 5’-GTCGCACAGAACTCCACTCA-3’; R: 5’-AAGAGCGAGAAATGGCAAAA-3’; Smo; F: 5’-CAACCTGTTTGCCATGTTTG-3’; R: 5’-ACATCAGCTGAGGGCTCATT-3’; Gli-1 F: 5’-CGGGGTCTCAAACTGCCCAGCTT-3’; R: 5’-GGCTGGGTCACTGGCCCTC-3’; GAPDH; F: 5’-CAAAGTTGTCATGGATGACC-3’; R: 5’-CCATGGAGAAGGCTGGGG-3’ (all primers ordered from Invitrogen). The reaction mixtures were denatured for 10 min at 94˚C, followed by 30-35 cycles of 94˚C for 1 min, 45 sec at 60˚C and 1 min at 72˚C. After amplification, 15 µl of each sample were resolved on a 1.5% agarose gel with ethidium bromide. Real-time quantitative PCR was performed for ribosomal 18S RNA (as an internal amplification control), Gli-1, SOX-18, and claudin-5 (primers from Applied Biosystems). The amplification was performed in the presence of AmpliTaq Gold DNA Polymerase, AmpErase UNG, dNTPs with dUTP, Passive Reference (ROX), and optimized buffer components according to manufacturer’s instructions. For 18S amplification, 200 nM of internal probe conjugated to VIC dye and 50 nM of the forward and reverse primers were added. For the target genes, 12.5 nM FAM-labeled probe and 45 nM of the forward and reverse primers were added (Applied Biosystems). Each sample was run in a 96-PCR plate incubated in 7900 Fast Real Time PCR System (Applied Biosystems) for 2 min at 50°C, 10 min at 95°C and followed by 40 cycles of PCR (95°C for 15 sec and subsequent 60°C for 1 min). The relative expression of each target gene was compared to ribosomal 18S and ΔCT (CT (target) - CT (reference)) was calculated.

Western blots For western blot (WB) analysis of whole cell lysates, HFAs, BBB-ECs and isolated mouse microvessels were lysed in denaturing buffer (50 mM Tris-HCl pH 8.5 and 0.1% SDS) supplemented with protease inhibitors, sonicated and centrifuged at 10,000 g for 10 min at 4˚C. Electrophoresis of 35 µg of supernatant was performed on 6-10% SDS-polyacrylamide gels under reducing conditions. WBs of isolated mouse microvessels were loaded with 7 μg of protein. For WBs of supernatants, 100 μg of chloroform-precipitated ACM and control DMEM (supplemented ± with 0.1 μg/ml of human recombinant (hr) Shh) were used. Proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Biorad) and blocked for 1h at RT in 5% milk. Membranes were then incubated overnight with rat anti-human Shh (R&D Systems), rabbit anti-human Ptch-1 Ab (Santa Cruz Biotechnology), rabbit anti-Smo (Abcam), rabbit anti-claudin-5, rabbit anti-claudin-3, rabbit anti-occludin, rabbit anti-ZO1 (Invitrogen), mouse anti-CS-1 (Millipore, Billerica, MA), mouse anti-JAM-A, mouse anti-occludin, mouse anti-p120 (BD Biosciences), or mouse anti-β-actin (as loading control, Sigma) followed by 1h incubation with species-specific horseradish peroxidase (HRP)-conjugated Abs

6

(DakoCytomation). All signals were revealed using ECL reagents (Amersham Biosciences). WBs shown are representative of 3 distinct experiments, using 2 EC and HFA preparations and WB of mouse tight junctions are representative of 2 distinct experiments using 2 microvessel preps for each condition.

Enzyme-linked immunosorbent assay and flow cytometry Confluent BBB-ECs were treated for 24h with fresh media supplemented with Hh agonists, Hh antagonists or with IFN-γ and TNF-α (100 U/ml each, R&D Systems). Quantitative changes in cytokine production were assessed from culture supernatants by enzyme-linked immunosorbent assay (ELISA) for IL-8/CXCL8 and MCP-1/CCL2 (BD Biosciences), according to the manufacturer’s protocol. ACM treatment was not performed due to the possible secretion of IL-8/CXCL8 or MCP-1/CCL2 by astrocytes, resulting in non-conclusive analysis. Plates were read using a Bio-Tek EL800 96-well plate reader at a 450 nm wavelength and analyzed using KC Junior program (Bio-Tek). From the same experiments, extracellular stainings of confluent BBB-ECs were performed after trypsin recovery. Cells were then incubated in 10% FBS with FITC-labeled goat anti-rabbit or PE-labeled rabbit anti-goat and anti-ICAM-1 (BD Biosciences), as well as anti-VCAM-1 (BD Biosciences) for 30 min at 4°C, washed and resuspended in flow cytometry buffer [1% (v/v) FBS, 0.1% (w/v) NaN3 in PBS]. Cell staining was acquired on a LSRII and analyzed using FACSDiva software (BD Biosciences). For intracellular staining, primary cultures of HFAs were stained with anti-HLA-ABC and were then fixed and permeabilized in 4% (w/v) PFA with 0.1% (w/v) saponin in Hank’s Balanced Salt Solution (10 min at RT). Cells were then incubated with Abs against Shh or Gli-1 (R&D Systems) for 30 min at 4°C in PBS buffer containing 0.1% (w/v) saponin, 1% FBS and 0.1% (w/v) NaN3. This was followed by a wash and incubation with APC-labeled donkey anti-rat (eBioscience). Cells were resuspended in flow cytometry buffer, acquired on a BD LSR II flow cytometer and analyzed using FACSDiva software. Transendothelial resistance measurement The electrical properties of confluent monolayers of primary human BBB-ECs were measured with the Electric Cell-substrate Impedance Sensing (ECIS) methodology by using the ECIS Zθ instrument and 8W10E+ electrode arrays (Applied Biophysics, Troy, NY, USA). The TEER properties of cells grown under conditions stimulating or abrogating the Hh pathway were recorded at 4000 Hz over a period of 42 hrs. BBB-ECs were seeded at 8.0×104 cells·well−1 in 400 μL of media and in triplicates.

Permeability and migration assays Primary cultures of BBB-ECs were used to generate an in vitro model of the human BBB, as previously reported (37, 38). BBB-ECs were plated on gelatin-coated 3 µm pore size Boyden chambers (BD Biosciences) at a density of 2 x 104 cells per well in EC culture media, supplemented with 40% (v/v) ACM when applicable, and were allowed to grow for 72h to reach confluency. After 48h, stimulation and abrogation of the Hh pathway were induced. At 72h, FITC-labeled bovine serum albumin (BSA; 50 μg/ml;

7

Invitrogen), [14C]-sucrose (250 nCi/ml; MP Biomedicals) or 1 x 106 freshly purified human CD4+ lymphocytes obtained using magnetic cell sorting (MACS) isolation columns (according to the manufacturer’s instructions; Miltenyi Biotec) were added to the upper chambers. Tracer diffusion across the EC monolayer was assessed as previously shown (39, 41). The clearance and permeability of the human BBB-ECs to both tracers was calculated as previously described in a 3h time-course (45). Statistical analysis of the clearance and permeability data was performed through ANOVA. Immune cell migration was assessed 24h following lymphocyte addition by recovering the migrated cells in the lower chamber with 0.1 M EDTA, stained with trypan blue, and counted using a hemacytometer under blinded conditions. Dead cells were excluded from total counts of migrated cells (approximately 1% of cells). In addition, proliferation assays (CFSE) of CD4 T cells, pre- and post-migration were also performed.

In vivo BBB permeability To establish BBB disruption in vivo, 7-8 week old C57Bl/6 female mice (Charles Rivers Laboratories) were injected intraperitoneally (i.p.) with a single dose of Cyclopamine at a concentration of 10 mg/kg or with equal volume of its vehicle (45% Hydroxypropyl-β-cyclodextrin, HPBCD, Sigma) (n=6 animals per group). After 6h, 0.1 ml of 70 kDa FITC-conjugated dextran (1 mg/ml) was injected into the tail vein, prior to 0.1 ml pentobarbital i.p. injection. BBB disruption was also determined by intraperitoneal injection of 600 μL of Evans blue 1%, followed by animal sacrifice 1 hr post-injection. Microscopic visualization of Dextran-FITC, fibrinogen and Evans blue extravasation was done on brain cryosections using a Leica DM6000 and a confocal microscope. BVs images from distinct conditions were acquired using identical camera settings. The area (number of pixels) and fluorescent intensity (mean pixel intensity) of the extravasation markers were measured using ImageJ (NIH, Bethesda, MA). The relative extent of fluorescent extravasation and BBB disruption was then calculated by multiplying the area by the fluorescent intensity. The extent of Evans blue extravasation was measure using the LAS software (Leica Microsystems). Smo inactivation in endothelial cells We have used the Cre-loxP approach to remove Smo activity specifically in endothelial cells. The Smo conditional allele (Smoc) has two loxP sites flanking the essential first coding exon (42). In order to target specifically the endothelial cells, we used a Tie2-Cre transgenic mouse line that expresses a Cre recombinase under the control of the Tie2 promoter (43). We crossed mice heterozygous for Tie2-Cre with mice homozygous for the conditional ‘‘floxed’’ Smo. Primers for Smoc are as follows: 5′-ATGGCCGCTGGCCGCCCCGTG-3′ and 5′-GGCGCTACCGGTGGATGTGG-3′, which generate a 0.6 kb band from the targeted allele. Validation of the Tie2-Cre specificity We tested the efficiency and the specificity of the Cre recombination by crossing the Tie2-Cre line with the global double-fluorescent Cre reporter mouse line described previously (46), allowing the visualization and distinction of recombined and non-recombined cells. Briefly, a tomato construct (Tm; variant of a membrane-bound red fluorescent protein) is expressed ubiquitously under the control of a chicken beta-actin

8

core promoter with a CMV enhancer. The expression of a Cre triggers the excision of the Tm cassette and allows the expression of the downstream transgene coding for a membrane-tagged GFP.

Expansion and culture of human T lymphocytes Human T helper (Th) 17, Th1 and Th2 lymphocytes were expanded and cultured as described (36, 37, 47). In brief, peripheral blood mononuclear cells of healthy donors were isolated by density gradient centrifugation on Ficoll-Paque (GE Healthcare). Human CD14+ monocytes and memory (CD4+CD45RO+) T cells were purified using magnetic cell sorting (Miltenyi Biotec). Memory T lymphocytes (1 x 106 cells/ml) were cultured with autologous monocytes at a 5:3 ratio, and stimulated with anti-CD3 (2.5 μg/ml, eBioscience) in X-vivo 15 medium supplemented with 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (Sigma) in the absence of serum. For Th17 polarization, T lymphocytes were grown in the presence of hrIL-23 (25 ng/ml), as well as with neutralizing antibodies against IFN-γ (5 μg/ml) and against IL-4 (5 μg/ml). To support Th1 expansion, cells were cultured with hrIL-12 (10 ng/ml) and anti-human IL-4 antibody (5 μg/ml). Th2 skewing was achieved by supplementing the media with hrIL-4 (200 ng/ml), anti-human IL-12 (5 μg/ml) and anti-human IFN-γ (5 μg/ml). To activate the Hh pathway the media was supplemented with hrShh (0.1 μg/ml) or ACM (40%). The Hh pathway was blocked by addition of the 5E1 monoclonal antibody against Shh (5 μg/mL) (Developmental Studies Hybridoma Bank, U of Iowa) when required. Recombinant cytokines and antibodies were purchased from R&D Systems. Cells were collected after 5–6 days for flow cytometry analysis as well as adhesion assays with BBB-ECs.

Flow cytometry and intracellular cytokine staining of Th lymphocytes The expression of IL-17 and IFN-γ in Th lymphocytes was characterized by intracellular cytokine staining and flow cytometry, as previously described (37). In short, lymphocytes were stimulated 6h with phorbol 12-myristate 13-acetate (20 ng/ml), ionomycin (1 μg/ml) and brefeldin A (2 μg/ml). Cells were first stained for surface antigens CD6, very late antigen (VLA)-4, lymphocyte function-associated antigen (LFA)-1, CD4 and CD45RO (BD Biosciences). After that, they were fixed in 4% (w/v) paraformaldehyde and permeabilized with 0.1% saponin. Intracellular cytokine staining was performed using antibodies against human IL-17 (eBioscience) and IFN-γ (BD Biosciences). Nonspecific background staining was determined by using fluorochrome matched isotype antibodies. Samples were acquired on a BD LSR II flow cytometer and analyzed with the BD FACSDiva Software.

BBB-ECs – T lymphocytes adhesion assays

BBB-ECs were plated on gelatin-coated 8 chamber-well slides (Thermo Fisher Scientific) at a density of 2 x 104 cells per well in EC culture media supplemented with 40% (v/v) ACM when applicable, and were allowed to grow for 72h to reach confluency. After 48h, cells had one of the following treatments: hrShh, Purmorphamine, SANT-1 or

9

Cyclopamine. After 72h, 2 x 106 CFSE-labeled human CD4+ lymphocytes were added and after 1 and 2h they were fixed with 70% ethanol for 5 min. Adhesion assays for Th cells were done on BBB-ECs grown on EC culture media. At 72h, 2 x 106 lymphocytes polarized into Th17, Th1 or Th2 phenotypes in the presence or absence of hrShh (0.1 μg/ml) were added and fixed as mentioned above. Fixed BBB-ECs and T lymphocytes were stained with phalloidin-Alexa 647 (Invitrogen) following manufacturer’s instructions and the number of T lymphocytes bound to ECs were determined under blinded conditions using fluorescent and DIC images obtained on a Leica SP5 confocal microscope.

Experimental allergic encephalomyelitis induction and scoring

Eight to nine week-old female C57BL/6 mice (Charles River, Montreal, QC) were actively immunized by subcutaneously injecting 200 μg of myelin oligodendrocyte glycoprotein (MOG)35-55 peptide (Alpha Diagnostic International, San Antonio, TX) in complete Freund’s adjuvant and supplemented with 600 μg of Mycobacterium tuberculosis (Difco, Detroit, MI). 500 ng of pertussis toxin were administered i.p on days 0 and 2 post-immunization and mice were also i.p injected with GDC-0449 (25 mg/Kg; Sellek, Houston TX) or DMSO on days 0, 4, 8 and 12. Clinical signs of EAE were daily assessed according to the following scores: 0, no clinical signs; 1, limp tail; 2, hind limp weakness; 3, partial hind limb paralysis; 4, complete hind limb paralysis; 5, hind and fore limb paralysis. On day 18 post-immunization, two to three mice per group were sacrificed and their organs (brain, spinal cord, spleen and lymph nodes) were collected. Immune cells were isolated by Percoll gradient centrifugation and cell surface and intracellular staining were performed and analyzed by flow cytometry as described above. The animal protocol was approved by the Comité Institutionnel de Protection des Animaux du CHUM (N07027PAs). The EAE data presented are representative of 2 experiments, with a total of 25 animals treated with GDC-0449 and 25 animals treated with DMSO.

Statistical analysis Data were analyzed using GraphPad Prism software. Results are shown as mean and standard error of the mean (SEM) and statistical analyses included paired Student t-test and ANOVA.

10

Fig. S1.

Fig S1. Components of the Hh signaling pathway are expressed by human and mouse astrocytes and BBB-ECs. (A) RT-PCR of SHH, PTCH-1 and SMO in primary cultures of human fetal astrocytes (HFAs) (two preparations: lanes 1 and 2) and BBB-ECs (two preparations: lanes 3 and 4). (B) WB of Hh pathway components in BBB-ECs and HFAs. Shh (19 kDa cleaved form) in astrocyte-conditioned media (ACM). Human recombinant

11

Shh (hrShh) was used as a positive control. Data shown are representative of n=3 BBB-EC and astrocyte preps. (C) Expression of Hh pathway components in primary cultures of human HFAs and BBB-ECs as detected by immunofluorescent confocal microscopy (n=3 distinct preparations). (D) Specificity of the Shh antibody (green) was determined in CNS sections of E13 WT and Shh-/- mice co-stained with caveolin-1(cav1). Immunostainings for Shh, Ptch-1 and Smo (all in green) and laminin (red, vascular basement membrane) in control human brain specimen. Co-localization images demonstrate Shh expression beyond the laminin-positive basement membrane of vessels. Hh receptors Ptch-1 and Smo expression are contained within the perivascular space, as delineated by laminin-positive basement membrane. (E) Ptch-1 (red) and Smo (green) immunostaining in vessels of adult human control CNS tissue (upper panels). VE-cadherin (VE-cadh; green) a marker of ECs co-localizes with Ptch-1 expression on CNS vessels (lower panels) (n=3) (F) Shh (green) and GFAP (red, a marker for astrocytes) immunostaining in adult control mouse brain. Astrocytes (arrowheads) are shown in upper panels and astrocyte endfeet surrounding a blood vessel (asterisk) are presented in lower panels (n=3). (G) Smo (green) and PECAM-1 (red, a marker of ECs) expression in adult control mouse brain (n=5). (H) Immunostainings for PDGFR-β (green, a marker for pericytes), PECAM-1 (red), Smo (red), Shh (red) and Ptch-1 (red) in CNS sections of healthy control mouse. Neither Shh, nor Smo or Ptch-1 co-localized with PDGFR-β (n=12). Scale bars in D to H: 30 μm. In all immunostainings TOPRO-3 stained nuclei blue.

12

Fig. S2.

Fig S2. (A) Shh does not affect BBB-EC proliferation. Flow cytometry analysis of carboxyfluorescein succinimidyl ester (CFSE) fluorescence in BBB-ECs untreated (control) or treated for 24h with hrShh. Black (empty) histograms represent CFSE fluorescence in BBB-ECs at day 0. Grey histograms (filled) represent CFSE fluorescence in BBB-ECs at day 4. Data shown are representative of n=2 experiments performed in triplicate. (B) The Hh antagonist SANT-1 (100 nM) prevent astrocyte-conditioned media (ACM) induced reduction in permeability. Data shown representative of n=5 experiments,*, P<0.05. (C) Human BBB-ECs respond to ACM by upregulating Hip and Gli-1 mRNA expression. Analyzed by RT-PCR, n=2. (D) Cyclopamine injected intra-peritoneally induces changes in BBB permeability in vivo. The relative extent of BBB leakage was studied in the CNS of C57Bl/6 mice injected with Cyclopamine (10 mg/kg in 45% Hydroxypropyl-β-cyclo-dextrin, HPβCD). Animals were sacrificed 6h following Cyclopamine injection. Vessels (asterisk) of Cyclopamine-treated mice showed leakage (arrowheads) of Dextran-FITC and fibrinogen into the surrounding parenchymal tissue and perivascular accumulation of blood-derived cells (n=6 animals, >40 sections studied). TOPRO-3 stained nuclei are blue. V= ventricle. Scale bars, 30 μm. (E) ACM-mediated upregulation of JAM-1, occludin and ZO1 expression by BBB-ECs is prevented by addition of the Hh antagonist SANT-1. Data shown representative of n=2 experiments.

13

Fig. S3.

Fig S3. Junctional protein expression in Shh deficient (-/-) E13.5 animals. (A) Tight junction protein immunofluorescence of Shh-/- embryonic mouse brains showing expression of claudin-5, occludin, and ZO-1 (all in red) on lectin+ vessels (green), compared to wild type (WT) mice. Merged images including nuclear staining (TOPRO-3, in blue) are shown (bottom panels). Left corner insets demonstrate reduction of tight junction protein expression in lectin+ vessels of Shh-/- mice compared to WT mice. Scale bars, 50 μm. (B) Quantification plot of the number of lectin+ vessels expressing tight junction proteins claudin-5, occludin and ZO-1 (n=12 random sections, 3 animals per group). (C) Quantification plot of the number of lectin+ vessels in the CNS of WT and Shh-/- embryos. Error bars, mean ± SEM. ***, P<0.001.

14

Fig. S4.

Fig S4. In vivo Cre-mediated conditional deletion of the Hh signal transducer smo in ECs. (A) Schematic showing the Tomato (a red fluorescent protein)-GFP expression cassette of the reporter mouse line used in this study. Large arrowheads represent loxP target sites for Cre-mediated recombination. In absence of recombination, only Tomato is expressed. Recombination of the loxP sites removes the Tomato and first polyA sequences, leading to GFP expression. (B) CNS sections from an E18 Tie2-Cre+; Tomato-GFP+ mouse

15

embryo showing DAPI (blue), GFP and Tomato signals, demonstrating the specific expression of GFP in blood vessels (arrowheads). Scale bar: 100 μm. Tm=Tomato; GFP=Green Fluorescent Protein; pA=polyadenylation sequences. (C) Schematic of the mouse cross performed for the conditional ablation of smo in ECs. The essential first coding exon of Smo is excised after expression of the Cre recombinase. (D) Smo expression in CNS sections derived from P19 WT (left) and Tie2-Cre; Smoc/c mice (right). PECAM-1 labeling of blood vessels is included and white rectangles represent the areas shown on the right of each panel. TOPRO-3 staining (nuclei) in blue. Scale bars: 30 μm. (E) Characterization of CNS vessels in E14 Tie2-Cre; Smoc/c mice (lower panels) and WT (upper panels). Immunofluorescence for PECAM-1, claudin-3, ZO-1, p120, Tomato lectin (all BBB endothelium markers), fibrinogen (indicator of plasma protein leakage) (n=3-5 animals). White rectangles indicate areas shown on the right of each panel. Nuclei stained with TOPRO-3 in blue. Scale bars: 50 μm.

16

Fig. S5.

Fig S5. Changes in BBB permeability in Tie2-Cre; smoc/c mice. (A) CNS sections of P19 wild type (WT, upper panels) and Tie2-Cre; Smoc/c (lower panels) animals. Immunofluorescencent labeling for PECAM-1 (labeling endothelium), fibrinogen, and ApoB (labeling plasma proteins) and laminin (labeling basement membrane of blood vessels) (n=3-5 animals). White rectangles indicate areas shown on the right of each

17

panel. Arrowheads indicate laminin fragments. Nuclei stained with TOPRO-3 in blue. (B) CNS sections of 8 week old WT (upper panels) and Tie2-Cre; Smoc/c (lower panels) animals injected intra-peritoneally with Evans blue (red) and immunostained for PECAM-1 (green and labeling the blood vessel endothelium). Pixel intensity quantification of Evans blue extravasation in WT and Tie2-Cre; Smoc/c (n=4 animals per genotype) mice. Error bars, mean ± SEM. ***, P<0.001. Insets on the top right are representative images of the brains after sacrifice. (C and D) CNS sections of 8 week old WT (left panels) and Tie2-Cre; Smoc/c (right panels) mice injected intra-peritoneally with Evans blue (red) and immunostained with IgGs (C) and GFAP (D).

18

Fig. S6.

Fig S6. (A) CNS sections of 8 wk old WT (upper panels) and Tie2-Cre; Smoc/c (medium and lower panels) animals injected intra-peritoneally with Evans blue-EB (red) and immunostained for laminin-lam (green and labeling basement membrane of blood vessels) and IgGs (blue) (n=4 animals per genotype). White rectangles indicate areas

19

enhanced on the right panels. Arrowheads denote disrupted/lower laminin expression. (B) WB analysis of the junctional proteins claudin-5, occludin, ZO-1 and p120 extracted from CNS microvessels isolated from 8 wk old WT and Tie2-Cre; Smoc/c animals (n=2 mice per group). (C) Turnbull histochemistry (iron staining) of CNS sections of P19 wild type (WT, upper panel) and Tie2-Cre; Smoc/c (medium panel) animals (n=2 mice per group). CNS microvessels are indicated with asterisks. Turnbull positive control staining is included in the lower panel. (D) Peroxidase histochemistry of CNS sections of E14 and P19 wild type (WT, upper panels) and Tie2-Cre; Smoc/c (lower panels) animals (n=2 mice per group). Positive peroxidase control of spleen is included in bottom panel. Turnbull and peroxidase histochemistries revealed the absence of hemorrhages in either WT or in Tie2-Cre; Smoc/c animals. Scale bars in all panels: 30 μm.

20

Fig. S7.

Fig S7. (A) Treatment of human BBB-ECs with Shh, purmorphamine or astrocyte-conditioned media (ACM) reduces expression intensity (MFI) of ICAM-1, as assessed by flow cytometry. Cyclopamine prevents ACM downregulation of ICAM-1 (n=4, in duplicate). (B) VCAM-1 expression was unaffected by Hh pathway agonists or antagonists, (n=4, in duplicate). (C) WB analysis revealed no apparent changes in the level of expression of VCAM-1 related ligand connecting segment (CS)-1, when BBB-ECs were treated with ACM or Shh. (n=2 experiments performed in duplicate). (D) Adhesion of CD4+ T cells to human BBB-ECs. BBB-ECs were subjected to Hh

21

activation (ACM, hrShh and Purmorphamine) or inhibition (ACM + Cyclopamine) for 24h prior to adhesion assays. Human CD4+ T lymphocytes were allowed to adhere on confluent BBB-EC monolayers for 2h (n=4 donors). Error bars, mean ± SEM. *, P<0.05; **, P<0.01. (E) Representative CFSE staining of T cell proliferation pre- and post-migration across monolayers of BBB-ECs. No difference in T cell proliferation was detected following migration. (F) Expression of Hh pathway receptors (Smo and Ptch1, red) and Gli-1 (green) on CD14+ monocytes (upper panels) and CD4+ T lymphocytes (lower panels) isolated from peripheral blood of healthy donors (n=3). Scale bar 10 μm.

22

Fig. S8.

Fig S8. (A) Cytokine production by human CD4+CD45RO+ T helper lymphocytes polarized in vitro with IL-12 (Th1), IL-4 (Th2) and IL-23 (Th17). (B to D) Expression of CD6, VLA-4, IFN-γ and IL-17 by flow cytometry in human Th lymphocytes polarized in

23

the presence (blue) or absence (red) of hrShh. (B) CD6 expression on Th17 cells. (C) CD6 (left) and VLA-4 (right) expression on Th2 lymphocytes. (D) IFN-γ and IL-17 expression on Th1 lymphocytes. (E) Effect of hrShh on CD6 expression by Th2 and Th17 lymphocytes, and on IFN-γ expression by Th1 lymphocytes (n=4 donors). (F) Effect of hrShh on LFA-1, VLA-4 and IL-17 expression by Th17 lymphocytes and on CD6, LFA-1 and VLA-4 by Th1 lymphocytes (n=4 donors) (G) Flow cytometry analysis of intracellular IFN-γ and IL-17 expression by human Th1 lymphocytes in the presence of astrocyte-conditioned media (ACM, middle panel) or ACM + anti-Shh neutralizing Ab (α-Shh, 5 μg/ml, right panel). Control (untreated) Th1 lymphocytes are shown in the left panel (n=2). (H) Flow cytometry analysis of CD6 expression in Th17 lymphocytes expanded in the presence of ACM (blue) or ACM + α-Shh 5 μg/ml (green). Control (untreated) Th17 lymphocytes are shown in red and the isotype in black (n=2). In all panels, the black histogram represents isotype control. Error bars, mean ± SEM. *, P<0.05; **, P<0.01.

24

Fig. S9.

Fig S9. (A) Weight changes in EAE animals (MOG-35-55/CFA immunized C57Bl/6) receiving either GDC-0449 or DMSO at days 0, 4, 8 and 12 (n=2 experiments, 25 animals per group). Error bars, mean ± SEM. *, P<0.05; **, P<0.01. (B) Representative intracellular staining of IFN-γ and IL-17 in splenic CD4 T lymphocytes from EAE mice treated with GDC-0449 vs DMSO. (C) Quantification of IFN-γ and IL-17-expressing splenic lymphocytes from EAE mice shown in B (day 16, n=2 experiments, 4 animals per group). Error bars, mean ± SEM. *, P<0.05; **, P<0.01. (D) Representative immunofluorescent images of Smo (red) expression in CNS vessels (asterisks) from healthy controls and MS patients (NAWM and active MS lesion) (n=2-3 patients per condition). Arrowheads indicate inflammatory cells. Nuclei stained with TOPRO-3 in blue. Scale bar 30 μm

References and Notes 1. A. Dodelet-Devillers et al., J. Mol. Med. 87, 765 (2009). 2. N. J. Abbott, L. Ronnback, E. Hansson, Nat. Rev. Neurosci. 7, 41 (2006). 3. J. H. Tao-Cheng, Z. Nagy, M. W. Brightman, J. Neurosci. 7, 3293 (1987). 4. A. Prat, K. Biernacki, K. Wosik, J. P. Antel, Glia 36, 145 (2001). 5. F. E. Arthur, R. R. Shivers, P. D. Bowman, Brain Res. 433, 155 (1987). 6. A. Prat et al., Arch. Neurol. 59, 391 (2002). 7. H. Kebir et al., Ann. Neurol. 66, 390 (2009). 8. R. M. Ransohoff, P. Kivisakk, G. Kidd, Nat. Rev. Immunol. 3, 569 (2003). 9. M. Fuccillo, A. L. Joyner, G. Fishell, Nat. Rev. Neurosci. 7, 772 (2006). 10. T. Nagase, M. Nagase, M. Machida, T. Fujita, Angiogenesis. 11, 71 (2008). 11. S. Kanda et al., J. Biol. Chem. 278, 8244 (2003). 12. E. Hochman, A. Castiel, J. Jacob-Hirsch, N. Amariglio, S. Izraeli, J. Biol. Chem. 281,

33860 (2006). 13. N. Byrd, L. Grabel, Trends Cardiovasc. Med. 14, 308 (2004). 14. C. Chiang et al., Nature 383, 407 (1996). 15. J. Briscoe, J. Ericson, Curr. Opin. Neurobiol. 11, 43 (2001). 16. J. Briscoe, J. Ericson, Semin. Cell Dev. Biol. 10, 353 (1999). 17. T. Osterlund, P. Kogerman, Trends Cell Biol. 16, 176 (2006). 18. Y. Wang, J. Imitola, S. Rasmussen, K. C. O'Connor, S. J. Khoury, Ann. Neurol. 64,

417 (2008). 19. T. Seifert, J. Bauer, R. Weissert, F. Fazekas, M. K. Storch, J. Neuropathol. Exp.

Neurol. 64, 404 (2005). 20. R. D. Fontijn et al., Am. J. Physiol Heart Circ. Physiol 294, H891 (2008). 21. H. C. Bauer et al., Brain Res. Dev. Brain Res. 75, 269 (1993). 22. N. R. Saunders, C. J. Ek, M. D. Habgood, K. M. Dziegielewska, Trends Neurosci. 31,

279 (2008). 23. C. J. Ek, K. M. Dziegielewska, H. Stolp, N. R. Saunders, J. Comp Neurol. 496, 13

(2006). 24. H. Kebir et al., Nat. Med. 13, 1173 (2007). 25. A. E. Lovett-Racke, Y. Yang, M. K. Racke, Biochim. Biophys. Acta 1812, 246

(2011). 26. R. Cayrol et al., Nat. Immunol. 9, 137 (2008). 27. J. I. Alvarez, R. Cayrol, A. Prat, Biochim. Biophys. Acta 1812, 252 (2011). 28. J. Greenwood et al., Neuropathol. Appl. Neurobiol. 37, 24 (2011). 29. S. Liebner, K. H. Plate, J. Angiogenes. Res. 2, 1 (2010). 30. R. Daneman et al., Proc. Natl. Acad. Sci. U. S. A 106, 641 (2009). 31. L. L. Rubin et al., Ann. N. Y. Acad. Sci. 633, 420 (1991). 32. N. J. Abbott, J. Anat. 200, 629 (2002). 33. A. Hashizume, Y. Hieda, Biochem. Biophys. Res. Commun. 339, 996 (2006). 34. Q. G. Xu, R. Midha, J. A. Martinez, G. F. Guo, D. W. Zochodne, Neuroscience 152,

877 (2008). 35. N. M. Amankulor et al., J. Neurosci. 29, 10299 (2009). 36. I. Ifergan et al., Ann. Neurol. 60, 45 (2006). 37. K. Wosik et al., J. Neurosci. 27, 9032 (2007). 38. F. Long et al., Development 128, 5099 (2001).

39. Y. Y. Kisanuki et al., Dev. Biol. 230, 230 (2001). 40. A. Siflinger-Birnboim et al., J. Cell Physiol 132, 111 (1987). 41. M. D. Muzumdar, et al., Genesis. 45, 593 (2007). 42. A. Prat, K. Biernacki, J. P. Antel, J. Autoimmun. 24, 119 (2005).