UBC Faculty Research and Publications

SPARC expression by cerebral microvascular endothelial cells in vitro and its influence on blood-brain… Alkabie, Samir; Basivireddy, Jayasree; Zhou, Lixin; Roskams, Jane; Rieckmann, Peter; Quandt, Jacqueline A Aug 31, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12974_2016_Article_657.pdf [ 1.85MB ]
JSON: 52383-1.0313414.json
JSON-LD: 52383-1.0313414-ld.json
RDF/XML (Pretty): 52383-1.0313414-rdf.xml
RDF/JSON: 52383-1.0313414-rdf.json
Turtle: 52383-1.0313414-turtle.txt
N-Triples: 52383-1.0313414-rdf-ntriples.txt
Original Record: 52383-1.0313414-source.json
Full Text

Full Text

RESEARCH Open AccessSPARC expression by cerebral microvascularendothelial cells in vitro and its influenceon blood-brain barrier propertiesSamir Alkabie1, Jayasree Basivireddy1, Lixin Zhou1, Jane Roskams2, Peter Rieckmann3,4 and Jacqueline A. Quandt1*AbstractBackground: SPARC (secreted protein acidic and rich in cysteine) is a nonstructural, cell-matrix modulating proteininvolved in angiogenesis and endothelial barrier function, yet its potential role in cerebrovascular development,inflammation, and repair in the central nervous system (CNS) remains undetermined.Methods: This study examines SPARC expression in cultured human cerebral microvascular endothelial cells(hCMEC/D3)—an in vitro model of the blood-brain barrier (BBB)—as they transition between proliferative andbarrier phenotypes and encounter pro-inflammatory stimuli. SPARC protein levels were quantified by Westernblotting and immunocytochemistry and messenger RNA (mRNA) by RT-PCR.Results: Constitutive SPARC expression by proliferating hCMEC/D3s is reduced as cells mature and establish aconfluent monolayer. SPARC expression positively correlated with the proliferation marker Ki-67 suggesting a rolefor SPARC in cerebrovascular development. The pro-inflammatory molecules tumor necrosis factor-α (TNF-α) andendotoxin lipopolysaccharide (LPS) increased SPARC expression in cerebral endothelia. Interferon gamma (IFN-γ)abrogated SPARC induction observed with TNF-α alone. Barrier function assays show recombinant human (rh)-SPARCincreased paracellular permeability and decreased transendothelial electrical resistance (TEER). This was paralleled byreduced zonula occludens-1 (ZO-1) and occludin expression in hCMEC/D3s exposed to rh-SPARC (1–10 μg/ml)compared with cells in media containing a physiological dose of SPARC.Conclusions: Together, these findings define a role for SPARC in influencing cerebral microvascular propertiesand function during development and inflammation at the BBB such that it may mediate processes of CNSinflammation and repair.Abbreviations: BBB, Blood-brain barrier; CAM, Chorioallantoic membrane; CNS, Central nervous system;hCMEC/D3, Human cerebral microvascular endothelial cell; IFN-γ, Interferon gamma; LPS, Lipopolysaccharide; MS, Multiplesclerosis; RT-PCR, Reverse transcription polymerase chain reaction; SPARC, Secreted protein acidic and rich in cysteine;TGF-β, Transforming growth factor beta; TJ, Tight junctions; TNF-α, Tumor necrosis factor alpha; VCAM1, Vascular celladhesion molecule 1; VEGF, Vascular endothelial growth factor; ZO-1, Zonula occludensBackgroundEndothelial cells play an essential role in normal homeo-stasis of the central nervous system (CNS). In healthyindividuals, microvessels throughout most of the CNSpossess a luminal monolayer of tightly apposed endo-thelial cells situated between the blood and brainparenchyma comprising together with adjacent astro-cytes the blood-brain barrier (BBB) [1]. Cerebral endo-thelial cells are crucial for normal neurological functionas they constitute both a physical “barrier” which limitsmolecular and cellular exchange between blood andbrain compartments and a “fence” which maintains po-larity of transporters responsible for delivery of essentialnutrients and removal of potentially harmful toxins [2].These CNS endothelia derive a low permeability barrierdue to interendothelial tight junctions (TJ) occludin and* Correspondence: jquandt@pathology.ubc.ca1Department of Pathology and Laboratory Medicine, University of BritishColumbia, Vancouver, BC, CanadaFull list of author information is available at the end of the article© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 DOI 10.1186/s12974-016-0657-9claudin proteins as well as junction associated submem-branous adaptor proteins such as zonula occludens(ZO)-1 [3]. Several studies show membrane localizationof tight junction proteins are the morphological correlate ofBBB integrity and tightness [4, 5]. Barrier disruption sec-ondary to tight junction dysregulation results from reducedendovascular flow [6], hypoxia/ischemia [7], and inflamma-tory cytokines such as tumor necrosis factor-α (TNF-α) [8]and vascular endothelial-derived growth factor (VEGF) [9].Several CNS diseases including neoplasia, hereditary vascu-lar malformation, trauma, and chronic inflammatory andneurodegenerative diseases such as multiple sclerosis(MS) feature characteristics of BBB breakdown [10, 11].Characterizing factors able to influence BBB integrityand aspects of vascular remodeling during CNS inflam-mation may identify key molecules with both physio-logical and perhaps pathological roles in disease.BBB cytoarchitecture and response to stimuli are oftenexamined in a simplified treatment and effect systemcomprised of in vitro cultures of endothelial cells estab-lished from cerebral microvessels. These cells recapitulatein vivo BBB characteristics such as expression of specificendothelial markers (i.e., CD31 and VE-cadherin), BBBtransporters (i.e., GLUT-1, P-glycoprotein, transferrin),and tight junction markers (i.e., ZO-1 and occludin) andform a monolayer with low paracellular permeability andhigh transendothelial electrical resistance (TEER) con-sistent with the presence of membrane-associated tightjunctions [12–14]. In the present study, we use a well-characterized in vitro model of the BBB consisting ofimmortalized human cerebral microvascular endothelialcells (hCMEC/D3) that express and appropriatelylocalize important BBB proteins characteristic of theirin vivo counterparts [15, 16].SPARC (secreted protein acidic and rich in cysteine) isa matricellular cell-matrix modulating protein involvedin angiogenesis [17, 18] and endothelial barrier function[19]. Many cell types including endothelia, fibroblasts,and macrophages constitutively express SPARC and up-regulate its expression in tissue regions undergoing highrates of remodeling, repair, and proliferation [20]. SPARCis typically enriched where new blood vessels are beingformed, as evidenced using an in vivo chorioallantoicmembrane (CAM) model of angiogenesis [21]. In theCNS, SPARC is highly expressed in developing bloodvessels at early stages of development and down-regulated with developmental maturity (Roskams Lab,unpublished observations). This spatiotemporal patternof SPARC expression is consistent with the role forSPARC in angiogenesis and BBB establishment [22, 23].Normal physiological levels of SPARC in healthy individ-uals (0.1–0.8 μg/ml in plasma) are increased in neoplasticand inflammatory conditions (1.5–10 μg/ml in plasma)[24–26]. Increased SPARC secretion has been associatedwith carcinoma [27] and other tumors [28] such as gli-omas [29], as well as inflammatory renal disease [30, 31],and scleroderma characterized by vascular dysfunction,autoantibody production, and tissue fibrosis [32].SPARC may play a role in mediating the CNS responseto injury and repair given enhanced expression in in vivomodels of CNS damage and repair. In an in vivo corticalwound model, SPARC messenger RNA (mRNA) wasabundantly expressed in developing blood vessels prox-imal to the wound edge days 3 to 10 post-injury, sug-gesting its involvement in the vascular response to CNSinjury and cerebrovascular ischemia [22, 33]. Furthermore,SPARC may be associated with neurological recovery fol-lowing CNS injury. Proteomic screens of murine laminapropria olfactory ensheathing cell (LP-OEC)-conditionedmedia identified SPARC as the key secreted protein sup-porting neural tissue repair after damage, capable of pro-moting spinal cord repair by limiting gliotic scar andcavity formation, stimulating axonal outgrowth and direct-ing angiogenesis [34]. SPARC has been shown to promoteSchwann cell-mediated neurite outgrowth in vivo and invitro [34]. Moreover, SPARC-null OECs transplanted intocontused rat spinal cord reduced outgrowth of specificsubsets of sensory and supraspinal axons and impairedimmune response to injury, suggesting its role inneural regenerative processes and the neuroimmuneresponse to CNS injury [34].Altogether, its potential roles in modulating angiogenicand barrier parameters of vascular development andrepair [19, 23, 35, 36], its expression and influence onneural regeneration after CNS injury [33, 34], and its in-fluence on the profile and extent of immune infiltration[37, 38] make SPARC a molecule of particular interest inchronic neuroinflammatory pathologies such as MS.MethodsReagents and antibodiesRecombinant human (rh)-SPARC (R&D Systems,Minneapolis, MN), TNF-α (Invitrogen, Camarillo, CA),interferon gamma (IFN-γ), (Invitrogen, Camarillo, CA),and lipopolysaccharide (LPS) (Sigma-Aldrich, St. Louis,MO) were obtained commercially. Antibodies for immuno-blotting and immunocytochemistry included monoclonalmouse anti-human SPARC IgG1 (2.5 μg/ml; HaematologicTechnologies, Essex Junction, VT), polyclonal rabbit anti-ZO-1 IgG (1 μg/ml; Invitrogen), monoclonal mouseanti-ZO-1 IgG1 (2.0 μg/ml; Invitrogen), monoclonalmouse anti-occludin IgG1 (0.5 μg/ml; Zymed, Carlsbad,CA), monoclonal mouse anti-Ki-67 IgG1 (1:100; Millipore,Billerica, MA), and monoclonal mouse anti-humanglyceraldehyde-3-phosphate dehydrogenase IgG1 (GAPDH,1:10,000; Santa Cruz Biotechnology, Santa Cruz, CA).Secondary antibodies included Alexa Fluor 568 goatanti-mouse IgG (1:200; Invitrogen); Alexa Fluor 488Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 2 of 17goat anti-rabbit IgG (1:200; Invitrogen); and horserad-ish peroxidase (HRP)-conjugated goat anti-mouse IgG(1:5000; Jackson ImmunoResearch, Burlington, ON).Isotype-matched control abs included mouse IgG1(2.5 μg/ml; Invitrogen) and rabbit IgG (1 μg/ml;Invitrogen).Cell cultureThe hCMEC/D3 line was generously provided by Drs. B.Weksler, I. Romero, and P-O. Couraud (Cochin Institute,Paris). The cell line was established as described inWeksler et al. 2005. hCMEC/D3s were cultured incomplete media comprised of EBM-2 media (Lonza,Walkersville, MD) supplemented with 5 % fetal bovineserum (FBS) (PAA Laboratories, Etobicoke, ON), 1 %penicillin-streptomycin, 1.4 μM hydrocortisone, 5 μg/mlL-ascorbic acid, 10 mM HEPES (all from Sigma), 1:100chemically defined lipid concentrate (Invitrogen-Gibco),and 1 ng/ml basic fibroblast growth factor (bFGF) (Invitro-gen-Gibco, Grand Island, NY) at 37 °C with 5 % CO2, 95 %air, and saturated humidity. The media was replenishedevery 2 to 3 days. For immunoblotting and reverse tran-scription polymerase chain reaction (RT-PCR), the cellswere seeded at 1.2 × 10E4 cells/cm2 on 25 cm2 flasks andsix-well plates (Corning, Corning, NY) coated with type Irat tail collagen (150 μg/ml; Sigma). For immunocyto-chemistry, the cells were grown on collagen type I-coatedglass coverslips or collagenous membranes in a transwellconfiguration (MP Biomedicals, Solon, OH). hCMEC/D3cells formed confluent, contact-inhibited monolayers, andwere used consistently between passages (p)28 and 35.Growth and development of hCMEC/D3sTo test if SPARC expression changed under varied serumconditions, confluent hCMEC/D3 cultures (p30–32) weregrown in complete growth media and replenished withfresh medium containing either no FBS, 1 % FBS, 5 %FBS, or 0.1 % bovine serum albumin (BSA) (Sigma) inplace of FBS. hCMEC/D3 growth was assessed for cultureconfluency where cells were approximately 20–30 % con-fluent after 1 day in culture (DIC), 50–70 % confluentafter 4 DIC, and confluent after 6–7 DIC. Protein andmRNA was extracted from cultures at subconfluent andconfluent growth stages and subjected to immunoblottingand semi-quantitative RT-PCR. hCMEC/D3 culturesgrown on collagen membranes were fixed in 4 % para-formaldehyde (PFA)-PBS when subconfluent (50–70 %confluent) and confluent, before being probed withmonoclonal antibodies against SPARC, ZO-1, and Ki-67for immunocytochemical analysis.hCMEC/D3 exposure to SPARC or inflammatory mediatorshCMEC/D3s cultured on collagen type I-coated six-wellplates until confluent were treated 1 day later with 0.1,1, 10 μg/ml rh-SPARC (R&D Systems) and TNF-α(200 U/ml) in fresh reduced serum (1 % FBS) completemedia for 24 h. The protein levels of endothelial tightjunctions were determined by Western blotting analysis.Confluent monolayers grown on collagen membraneinserts were replenished with complete media con-taining 10 or 100 U/ml TNF-α, 100, 200, or 500 U/ml IFN-γ, and 50 ng/ml LPS, alone and in combin-ation. For cytokine treatment of cells on collagenmembranes, media containing cytokine was added tothe upper chamber.Barrier function assaysBarrier function was examined using TEER impedancemeasurements and transendothelial FITC-dextran (3 and10 kDa) diffusion assays. hCMEC/D3 (p30–32) weregrown on transwell polyester membrane inserts (0.4-μmporosity and 1.12 cm2 surface area, Corning Costar#3460) in serum-reduced (1 % FBS) complete mediauntil 2 days post-confluence, and the measured TEERhad stabilized. Media was replenished into both upper(luminal) and lower (abluminal) chambers; however,cytokine- and SPARC-containing media were only ap-plied into the upper chamber. TEER was measured dailyby EVOM-2 ohm meter using ENDOHM-12 chamberand STX2 chopstick electrodes (World-Precision Instru-ments, Sarasota, FL). Resistance of blank filters withoutcells was subtracted from those with cells, and thesevalues were multiplied by the surface area of the insertsto get the final resistance (Ω·cm2).Endothelial monolayer permeability was studied bymeasuring fluorescently labeled dextran diffusion througha confluent monolayer of hCMEC/D3 cells grown ontranswell inserts. These studies were performed by adding50 μg/ml dextran (3 or 10 kDa) tagged with Alexa Fluor488 (Molecular probes, Eugene, OR) to the luminalchamber of a transwell insert and sampling (10 μl/sam-ple) from the abluminal chamber at 30, 60, 120, and180 min into a 96-well plate, an equal volume of freshmedium replaced each time. Dextran fluorescence inthe samples was measured by a microplate fluorometer(Floroskan Ascent 374, ThermoScientific, Hudson, NH)at 485 nm (excitation) and 525 nm (emission). Rawvalues from dextran fluorescence were converted toconcentrations using a standard curve and slope of thelinear regression line. Permeability coefficient was cal-culated as described previously [4].Cell lysis and Western blottingTo lyse hCMEC/D3 cells grown on six-well plates, thecells were first washed twice with cold PBS, then incu-bated in situ on ice for 10 min in 125 μl of ice-cold NP-lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mMEDTA, and 1 % NP-40, adjusted to pH 8.0) with freshAlkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 3 of 17protease inhibitor and phosphatase inhibitor cocktailsadded before use (Roche, Laval, QC). The cells werethen scraped and collected into tubes, triturated on icethrough a 28-Gauge needle five times and centrifuged at13,500 rpm for 12 min at 4 °C. The supernatant wascollected as whole cell lysate and stored at −80 °C.hCMEC/D3 cell lysis for TJ protein analysis requiredstronger detergents, 125 μl of ice-cold radio-immunoprecipitate assay (RIPA) buffer containing20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 1 %NP-40, 0.1 % SDS, and 0.5 % sodium deoxycholate.Protein concentrations were determined by bicincho-ninic acid (BCA) protein assay kit (Sigma). Protein esti-mations based on the BSA standard absorbance curve.Lysates (20–40 μg) were diluted 1:1 in reducing 2×Laemmli sample buffer (Bio-Rad, CA) containing 0.05 %β-mercaptoethanol and heated at 100 °C for 5 min. Thesamples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE: 6 % gelfor ZO-1, 12 % gel for SPARC, 15 % gel for occludin andclaudin-5) for 1.5 h at 80–100 V and transferred tonitrocellulose membranes in wet transfer buffer (25 mMTris, 192 mM glycine, 20 % methanol) at 90 mA over-night at 4 °C. Immunoblots for SPARC, claudin-5, occlu-din, and GAPDH were blocked in 5 % skim milk in Tris-buffered saline (TBS: 1.37 M NaCl, 27 mM KCl, 0.25 MTris, adjusted to pH 7.4) for 1 h then incubated inprimary antibody diluted in 2 % skim milk-TBST(TBS + 0.01 % Tween 20 (TW20)) overnight at 4 °C(or 2 h at RT for GAPDH). After washing, HRP-conjugated secondary antibodies diluted in 2 % skimmilk-TBST (TBS + 0.01 % TW20) were added andthen again washed in TBST. Immunoblots for detec-tion of ZO-1 were blocked in 1 % BSA-TBST (TBS +0.05 % TW20) for 4 h at RT and then incubated inmonoclonal anti-ZO-1 primary antibody diluted inTBS overnight at 4 °C followed by washing in TBSTand incubation with secondary antibody. All immuno-blots were developed in enhanced chemiluminescence(ECL) substrate (ThermoFisher Scientific, Bothell,WA). Band signals were detected using a Versadocimaging system (Bio-Rad Laboratories), and banddensities were quantified by ImageJ 1.42i software(National Institutes of Health, Bethesda, MD).Semi-quantitative RT-PCRTotal RNA was extracted from cell cultures usingRNeasy Mini Kit (QIAGEN Science, Toronto, ON) ac-cording to the manufacturer’s protocol. ComplementaryDNA (cDNA) was synthesized by High Capacity RNA-to-cDNA Master Mix Kit (Applied Biosystems, FosterCity, CA). Relative quantification of PCR reactionswere performed with Platinum PCR SuperMix (Invi-trogen) using the following primer sequences for SPARCtranscript amplification 5′-AGTGCACCCTGGAGGG-CACC-3′ (Forward); 5′-TGCTTGATGCCGAAGCAGCC-3′ (Reverse) and the following primers for the GAPDHhousekeeping gene 5′-AAGGCTGGGGCTCATTTGCAG-3′ (Forward); 5′-CTGCTTCACCACCTTCTTGATG-3′(Reverse). A semi-quantitative analysis of mRNA levelswas carried out using scans with the Bio-Rad GelDoc UV system (Bio-Rad Laboratories, Hercules, CA)and differences in SPARC and GAPDH expressionwere calculated by Image Lab software (Bio-RadLaboratories).ImmunocytochemistryFollowing co-incubations, cultures were washed twicewith warm (37 °C) PBS and fixed at RT in 4 % PFA-PBSfor 10 min. Cultures were directly stored in PBS-0.01 %sodium azide at 4 °C until stained. For staining, cultureswere washed twice with PBS then incubated for 10 minin permeabilization-blocking buffer (0.1 % Triton X-100and 4 % normal goat serum (NGS) in PBS). Cultureswere incubated twice for 10 min in blocking buffer (4 %NGS-PBS) then incubated with primary antibody dilutedin blocking buffer for 1 h at RT or overnight at 4 °C.After primary antibody incubation, cultures were washedthree times for 5 min with PBS then incubated inspecific secondary antibody in blocking buffer for50 min in the dark. Cultures were washed twice inPBS, stained with DAPI nuclear stain (1:10,000) inPBS for 5 min, and again washed three times in PBS.Membranes were excised from discs using a scalpel,drained of excess PBS, and embedded in 10 μl ofProLong Gold anti-fade reagent (Invitrogen) under-neath a glass coverslip. Cultures on the glass cover-slips were mounted cell side down on 10 μl ofProLong Gold.Image acquisition and analysisDistribution analysis was performed in a blinded fashionwhere cells were distinguished and coded numericallyon a DAPI image to avoid duplication and evaluated forlevel of SPARC expression. SPARC immunoreactivitylevels were assigned to individual cells according to theclassification scheme and representative images in Table 1.Fluorescence images were captured with an Axioplan 2imaging epifluorescent microscope (Zeiss, Jena, Germany)and Axiovision 4 software (Zeiss). Confocal micrographswere captured with an Olympus Fluoview 1000 laser scan-ning confocal microscope with Nomarski optics andFV1000 Fluoview software (Olympus Corporation, TokyoJapan). Analysis was performed using Adobe Photoshopextended CS3 version 10.0 or ImageJ 1.42i software (NIH)or assessed by SPARC immunoreactivity scale. AdobePhotoshop measurements of nuclear Ki-67 intensity wereperformed by outlining individual nuclei on a DAPI/blueAlkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 4 of 17filter image using the “quick selection tool”, and measu-ring Ki-67/green filter mean pixel intensity (MPI)—theaverage intensity of all pixels within the DAPI-stained/de-lineated nuclear regions. Ki-67 positivity was relative to athreshold determined by blindly screening three images ofconfluent and subconfluent images each and denotingthose cells negative for Ki-67. A threshold limit was set byaveraging the Ki-67 MPI of those nuclei visually deemedKi-67 negative (n = 59) plus three standard deviations. Ki-67-positive nuclei were those exceeding this pre-definedKi-67 threshold. Quantification for SPARC immunofluo-rescence was performed either by regional analysis, whereregions for analysis were demarcated by thresholding thecell edge such that only the cell-covered surface area of afield/image was assessed for MPI using ImageJ or assessedbased on the SPARC immunoreactivity scale. Cell count-ing of DAPI-positive nuclei was performed using theImageJ “analyze particles” tool and automated usingImageJ macroscript.Statistical analysisGraphPad Prism version 5.01 (GraphPad Software, SanDiego, CA) was used for graph synthesis and statisticalanalysis. Parametric data expressed as mean ± standarderror of the mean (mean ± SEM) were analyzed by one-way analysis of variance (ANOVA) and Newman-Keulmultiple comparison test. Non-parametric data expressedas mean ± standard deviation (mean ± SD) were analyzedby a Kruskall-Wallis and Dunn’s multiple comparison test.Mann-Whitney comparison was performed on data com-paring two non-parametric groups/treatments. SPARCimmunoblots and functional assays were analyzed by one-way ANOVA followed by a Tukey’s honestly significantdifference (HSD) multiple comparisons analysis.ResultsThe in vitro hCMEC/D3 model of the BBBhCMEC/D3 cells grown on collagen type I-coated sur-faces had fusiform morphology and were tightly apposedforming monolayers with minimal overlap upon reach-ing confluence (Fig. 1a). The cells grown on collagenmembrane inserts (primarily type I collagen with sometype IV collagen) were comparable in morphology andgrowth rate to that of cells grown on type I collagen-coated flasks or filters and thus became our preferredculturing surface for immunocytochemistry experiments.SPARC expression during hCMEC/D3 growth andmaturationSPARC is known to be expressed in different tissues duringremodeling and repair [20], where SPARC expression cor-relates with cell propagation. SPARC protein expressionhere was first quantified by Western blot on cells at 20–30,50–70, and 100 % confluency (Fig. 1a). ImmunoblottingTable 1 SPARC immunoreactivity assessment scale for hCMEC/D3 cells in cultureSPARC levelStaining Representative immunocytochemistryDescription Approximate mean pixel intensity (MPI)0 Negligible Faint diffuse cytoplasmic staining1.9−2.31 Low Perinuclear staining2.3−2.92 Moderate Perinuclear  and diffuse cytoplasmicstaining2.9−3.53 High Perinuclear and intense cytoplasmic staining3.5−4.5Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 5 of 17Fig. 1 (See legend on next page.)Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 6 of 17and immunocytochemistry analysis showed subconfluentcultures had greater SPARC expression than confluent cul-tures. SPARC expression decreased with days in culturewith greater than 50 % down-regulation as hCMEC/D3established a confluent monolayer (Fig. 1b, c). These exper-iments provide evidence that SPARC is highly expressed inhCMEC/D3 cells during proliferative stages of growth anddown-regulated as cells establish a monolayer and becomecontact inhibited.To investigate SPARC expression during hCMEC/D3growth and as cells develop an in vitro BBB phenotype,cultures were fixed when approximately 50–70 % con-fluent (subconfluent) and confluent then subjected toimmunocytochemical staining. Confluency was visuallyassessed by phase contrast microscopy. Barrier pheno-type was confirmed by peripheral localization of ZO-1,a marker of tight junction formation. In subconfluentcultures, cells exhibited perinuclear and dense cytoplas-mic ZO-1 staining with ZO-1 rarely limited peripher-ally to cell borders (Fig. 1d). Conversely, confluent cellsexhibited tightly apposed monolayers with minimaloverlap (Fig. 1a, d) and ZO-1 bands delineating thecytoplasmic membrane in the majority of cells (Fig. 1d).Immunocytochemical analyses revealed that SPARClevels were heterogeneous among hMEC/D3 cells inculture. The intensity of SPARC expression (negligible,low, moderate, or high) in the subconfluent and conflu-ent cultures was assessed on confocal images. The sub-confluent cells predominantly exhibited low, moderate,and some high SPARC levels compared to the confluentcultures which were largely (~80 %) negligible while theremainder of cells were low in SPARC expression(Fig. 1e, f ).SPARC expression was also examined on high-resolutionconfocal micrographs by quantifying the global meanpixel intensity of the images. SPARC intensity was mea-sured in subconfluent and confluent cultures within a300 μm × 300 μm field on images pooled from replicatewells reported as relative mean pixel intensity ± stand-ard deviation (MPI ± SD). SPARC intensity was greaterin regions of subconfluent culture (3.85 ± 0.492) thanregions of confluent culture (2.83 ± 0.426). Confocalmicrographs revealed that subconfluent cultures withlow and discontinuous ZO-1 staining at cell bordershave greater SPARC immunoreactivity than confluentcultures which showed intense ZO-1 banding at cell-cell contacts (Fig. 1d) and minimal SPARC expression(Fig. 1d).SPARC expression correlates with the proliferationmarker Ki-67The Ki-67 protein (also known as MKI67) is a cellularmarker for proliferation present during all active phasesof the cell cycle (G1, S, G2, and mitosis) but typically ab-sent from resting cells (G0) [39, 40]. Consistent with arelatively quiescent versus proliferative state, cells fromconfluent cultures had a significantly reduced percentageof Ki-67-positive or bright cell nuclei than subconfluentcultures (Fig. 2a, P = 0.01). The immortalized nature ofthis cell line would be consistent with lower, oftenspeckled or punctate, but not negligible incidence of Ki-67 staining in confluent cultures. To further examine as-sociations between relative SPARC expression and Ki-67reactivity in individual cells, we focused on subconfluentcells which had SPARC expression in all ranges of ourclassification scheme. Notably, cells in subconfluent cul-tures with negligible, low, or moderate to high levels ofSPARC expression demonstrated significantly increasingfrequencies of Ki-67-positive or bright cell nuclei (Fig. 2b;P < 0.0001). Representative images demonstrate that cellsin confluent cultures were more likely to demonstratecolocalization of SPARC along with brighter Ki-67 stain-ing; conversely, confluent cultures showed faint SPARCstaining and fewer Ki-67 bright cells than observed insubconfluent hCMEC/D3 (Fig. 2c).SPARC expression is influenced by serum conditionsSerum has been defined as a potential stimulatory andmitogenic factor for cells in primary culture as well ascell lines in vitro [41]. The ability of hCMEC/D3 cul-tures to fare well in reduced serum allowed us to exam-ine the effects of serum dose on SPARC expression. Fivepercent serum-supplemented media recommended forhCMEC/D3 cell line expansion produced the highest(See figure on previous page.)Fig. 1 SPARC expression is associated with proliferation of immature hCMEC/D3 and decreased as cells mature and reach confluence. a Confluencyand monolayer intactness were visually assessed by phase contrast microscopy. b, c hCMEC/D3 cultures 20–30, 50–70, and 100 % confluent wereanalyzed for SPARC protein expression by immunoblotting. Data is pooled from two independent experiments, error bars represent standard error ofthe mean (SEM) with statistical significance by one-way ANOVA and Tukey’s comparison test **P < 0.001; ***P < 0.0001. d Representative confocalmicrograph images show greater perinuclear and cytoplasmic SPARC staining (red) in subconfluent cultures compared to confluent cultures.Subconfluent cultures exhibit discontinuous peripheral ZO-1 bands (green), whereas confluent cells were tightly apposed and organized intomonolayers with continual ZO-1 bands at interendothelial borders. Scale bar = 30 μm. e The individual cells in confluent monolayers expressless SPARC than the cells in subconfluent cultures. Levels of SPARC immunoreactivity (negligible, low, moderate, or high) were assigned according toTable 1 (Additional file 1: Figure S1). Bars represent the average of results from n = 18 and n = 10 images enumerated for low, medium, or high levelsin subconfluent (empty bar) vs. confluent (filled bar) cultures in replicate wells. f Summary of individual cell staining classification from both Fig. 1e(exp. A) and a second experiment (exp. B) with similar findings for the distribution of cells from (S) subconfluent vs. (C) confluent culturesAlkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 7 of 17observed levels of SPARC protein. One percent serumalso significantly increased SPARC compared to medialacking serum (Fig. 3a, b). Culturing cells in 0.1 % BSAresulted in SPARC expression more comparable tolevels observed in serum-free cultures (Fig. 3a, b).These data clearly demonstrate that SPARC expressionin hCMEC/D3 cells is influenced by serum concentra-tion in culture media. To enhance our sensitivity todetect increases in SPARC expression secondary toinflammatory stimuli, subsequent treatment experi-ments were carried out in the presence of 1 % serumwhere constitutive SPARC expression was demon-strated to be low.Inflammatory mediators regulate SPARC expressionin hCMEC/D3sOur in vitro approach allowed us to model the influenceof specific inflammatory molecules on cell morphologyFig. 2 SPARC expression positively correlates with a marker of proliferation Ki-67. a Representative immunocytochemistry images of subconfluenthCMEC/D3 show greater SPARC expression and higher proportions of Ki-67-positive cell nuclei than confluent cultures. Images represent SPARCstaining (red) alone (upper panel) or merged with Ki-67 staining (green) in the lower panel. Data represents one of two replicate experiments. Scalebar = 30 μm. b Percentage of cells positive for the proliferation marker Ki-67 in the subconfluent hCMEC/D3 is greater than that of the confluentcultures. Confluent hCMEC/D3 in two experiments had significantly (exp. A: P = 0.0143; exp. B: P = 0.0017) lower incidence of Ki-67-positive nucleithan the subconfluent cultures. Bars represent the average of results (20 images from triplicate subconfluent wells and 8 images from duplicateconfluent wells) grown on collagen membranes ± standard deviation (SD). Mann-Whitney comparison test, **P < 0.001. c SPARC levels positivelycorrelated with the percent of Ki-67-positive cell nuclei in subconfluent hCMEC/D3 cultures. Ki-67 staining is significantly different between groupsclassified according to increasing levels of SPARC staining from negligible to low through moderate and high combined (P < 0.0001). Barsrepresent the average of results from n = 20 images pooled from triplicate wells. Error bars indicate SEM. Data represents one of the two experimentswith similar significant results. One-way ANOVA and Tukey’s multiple comparison test, **P < 0.001, ***P < 0.0001Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 8 of 17and SPARC expression at the BBB. hCMEC/D3 culturesgrown to confluence were incubated for 12, 24, and 48 hwith TNF-α, IFN-γ, or LPS, alone or in combination.SPARC expression patterns were studied after treatmentwith pro-inflammatory cytokines TNF-α and IFN-γ, im-plicated in the pathogenesis of several CNS inflamma-tory disorders and the endotoxin LPS, a primarycomponent of the bacterial cell wall known to stimulatethe innate immune response to infection. Cell conflu-ence and morphology was assessed by phase contrastand ZO-1 immunofluorescence. SPARC expression wasquantified by immunofluorescence and immunoblotting.Untreated cultures showed tightly apposed cells withminimal overlap at the time of treatment and regular fu-siform morphology throughout the treatment period.Treatment with pro-inflammatory mediators resulted inelongated morphology and whirling into circular pat-terns along with a disruption in peripheral ZO-1 staining(data not shown). SPARC immunoreactivity was minimalin untreated confluent cultures maintained in serum-reduced media (Fig. 4a, d) and significantly increasedfollowing exposure to 10 and 100 U/ml TNF-α for 24 to48 h (Fig. 4a). 50 ng/ml LPS treatment induced SPARClevels to that observed for 100 U/ml TNF-α alone, while10 ng/ml LPS had no effect (Fig. 4b, d). Interestingly,100 U/ml TNF-α and 100 U/ml IFN-γ co-treatment ab-rogated SPARC induction otherwise observed for TNF-αalone at 24 and 48 h (Fig. 4a, d).The modulatory effect of IFN-γ on SPARC expressionin hCMEC/D3 cells was temporally delayed. Significantdecreases in SPARC protein level were not observeduntil 48 h of IFN-γ treatment (Fig. 4c) However, thedecrease in SPARC expression in response to a combi-nation of IFN-γ (100 U/ml) and TNF-α (100 U/ml) wasevident much earlier at 12 h (data not shown) and sig-nificant as early as 24 h. We further analyzed the effectsof various concentrations of IFN-γ on SPARC proteinlevel. To our surprise, higher concentrations of IFN-γ(200, 500 U/ml) did not significantly change SPARC ex-pression at 24 h and was comparable to 100 U/ml at48 h compared to untreated controls. Taken together,these results show that TNF-α alone or LPS consistentlyincreased intracellular SPARC protein expression, whileIFN-γ alone or in combination with TNF-α limits its en-hancement in hCMEC/D3 cells. SPARC expression wasalso analyzed at the individual cell level (Fig. 4d) andacross areas of confluent cells (Fig. 4e). SPARCexpression as measured by mean pixel intensity wassignificantly increased in cells treated with TNF-αalone or with LPS alone, which correlated well withWestern blot data. Treatment of hCMEC/D3 cultureswith TNF-α did up-regulate SPARC mRNA expressionin a dose-dependent manner where only 100 U/mlTNF-α reached significance in the time studied.Exogenous SPARC treatment increases cerebralendothelial barrier permeabilitySecreted SPARC binds matrix components of the base-ment membrane and modulates cell-matrix interactionsthought to affect endothelial barrier function [42]. Directevidence suggests that SPARC increases transendothelialpermeability of hCMEC/D3 monolayers and may play arole in modulating BBB permeability. To understand therole of exogenous SPARC on modulating BBB integrity,rh-SPARC was applied to hCMEC cultured on transwellinserts 1 day post-confluence. SPARC treatment wasperformed in serum-reduced media with 1 % FBS tolimit endogenous SPARC production (compared tothat observed with higher serum conditions). We as-sumed that a lower endogenous SPARC level mightclarify the effect of exogenously supplied SPARC onpermeability.TEER on hCMEC/D3 remained relatively low through-out this study. It is documented that hCMEC/D3 cultureshave low TEER under static conditions [6, 16]. BaselineTEER recordings were between 20 and 25Ω·cm2 byENDOHM-12 chambers and 55–60Ω·cm2 by STX2chopstick recordings across untreated hCMEC/D3 mono-layers. Application of rh-SPARC decreased TEER in aconcentration-dependent manner (Fig. 5a). SignificantTEER reduction was seen only at our greatest SPARC con-centration tested of 10 μg/ml at 24 h but was still onlyabout half of the TEER change observed following TNF-α(200 U/ml) treatment.Fig. 3 a, b Growth media serum concentration influences SPARCexpression in confluent hCMEC/D3 cells. SPARC protein levels werequantified by western blot analysis of lysates derived from culturespropagated in media with varied serum concentration. Relative SPARCprotein levels were normalized to GAPDH. SPARC detected in lysatesresolved at the molecular weight observed for rh-SPARC (~43 kDa).Bars represent the average of four immunoblots, two independentexperiments (exps. A and B) performed in duplicate. Error bars indicateSEM. One-way ANOVA and Tukey’s multiple comparison test, *P < 0.05Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 9 of 17Fig. 4 (See legend on next page.)Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 10 of 17Cultures used to measure TEER were also used inFITC-dextran diffusion permeability assays. Microvascularcerebral endothelia partially comprise the BBB and form afunctional barrier interface between blood and neuralparenchyma regulating vital functions such as fluid-ionbalance and essential nutrient delivery. Addition of ex-ogenous SPARC increased the permeability of hCMEC/D3 monolayers as measured by increased FITC-dextrandiffusion after 24 h. This increase was concentrationdependent where more SPARC produced a greater in-crease in permeability to both 3- and 10-kDa dextrans(Fig. 5b). A dose of SPARC greater than 0.01 μg/mlSPARC was sufficient to increase the permeability ofhCMEC/D3 monolayers. Our studies show that SPARChas direct effects on properties of the blood-brain barrier.Because TNF-α was also able to drive SPARC expression,we addressed whether or not SPARC may be contributingto TNF-α-mediated changes in permeability. Testing avariety of monoclonal as well as polyclonal antibodiesfor their potential to neutralize SPARC, we were notable to significantly influence TNF-α mediated changesin EC permeability detected by dextran diffusion (datanot shown). In this regard, the primary TNF-α inducedalterations in BBB permeability would seem to occurindependent of SPARC.SPARC alters tight junction protein levels in hCMEC/D3sTo test the ability of SPARC to alter BBB-associatedtight junction protein expression, hCMEC/D3 mono-layers were treated with rh-SPARC (0.1, 1, 10 μg/ml) orTNF-α (200 U/ml), the latter being a well-characterizedmodulator of TJ expression in BBB endothelial cells,including hCMEC/D3 cells [16]. Phase contrast mi-crographs of TNF-α and SPARC-treated hCMEC/D3monolayers showed cultures remained intact under alltreatment conditions and retained regular fusiformmorphology and apposition with minimal overcrowdingand no apparent lifting or dissociation (data not shown).Previous studies of BBB integrity and TJ expression havedescribed a barrier-promoting/protecting property forhydrocortisone [4]; hydrocortisone was included at dosesrecommended by the hCMEC/D3 providers based onpublications testing its influence on barrier phenotype.Cell morphology and growth appeared similar in thepresence or absence of hydrocortisone after 24 h (datanot shown).Addition of TNF-α decreased total expression of all TJproteins studied (Fig. 6a). TNF-α treatment of hCMEC/D3s lowered the expression of ZO-1 (0.76, P = 0.013)and occludin (0.62, P = 0.044) compared to controls. Theaddition of hydrocortisone to the media with TNF-αreturned ZO-1 expression to baseline levels, but was noteffective at restoring occludin levels to that of untreatedcells when in the presence of TNF-α (data not shown).Our experiments revealed that hCMEC/D3 cells inculture produced low levels of SPARC even under restingconditions, suggesting that levels of SPARC in cultureclosely model the physiological levels of SPARC detectedin the plasma and serum of healthy individuals (rangingfrom 0.1 to 0.5 μg/ml). In this regard, it is difficult toassess the physiological relevance of comparing cellscultured in “0” or “no SPARC” conditions which wouldexist immediately following media replacement.Exogenous SPARC was applied at physiological levelsof 0.1 μg/ml and at supraphysiological levels observedin individuals with various inflammatory disorders.Functional studies using FITC-dextran (3 and 10 kDa)diffusion and TEER showed that rh-SPARC treatment in-creases transendothelial permeability of hCMEC/D3 mono-layers in a concentration-dependent manner (Fig. 5b, c),confirming barrier disruption in cerebral monolayerswhen exposed to SPARC in physiological (0.1 μg/ml)and supraphysiological (1–10 μg/ml) concentrations.Furthermore, we were also interested in whetherSPARC altered TJ expression at the BBB. Increasingconcentrations of exogenous SPARC from 0.1 to 1 μg/ml(See figure on previous page.)Fig. 4 Inflammatory mediators differentially regulate SPARC expression in hCMEC/D3 monolayers. hCMEC/D3 were incubated for 24 and 48 h inreplenished media with TNF-α, IFN-γ, or LPS, alone or in combination. a Compared to untreated/control conditions, 10 and 100 U/ml TNF-α treatmentalone increased SPARC expression at 24 and 48 h. Bars represent the average relative value of blots from n = 3 independent experiments. Error barsindicate SEM. ANOVA and Tukey’s multiple comparison test, *P < 0.05. b The effect of varied concentrations of IFN-γ (100, 200, 500 U/ml) on SPARCexpression in hCMEC/D3 cultures was assessed by Western blot analysis. c LPS treatment enhanced SPARC levels in a dose-dependent manner at 24and 48 h by Western blotting. Pooled results from two biological replicates are shown. d The effect of inflammatory mediators on SPARC expressionin confluent hCMEC/D3 cultures was assessed in parallel by quantitative immunocytochemistry with representative images shown and summary offindings shown e. SPARC expression is shown following 100 /ml TNF-α, 100 U/ml IFN-γ, or 25 ng/ml LPS alone or in combination. Quantification waspooled from 10 images of duplicate wells for each treatment. Untreated confluent cultures showed minimal SPARC staining at 24-h TNF-α and LPStreatment alone increased SPARC levels after 24 h (P < 0.0001). Data represents one of the two experiments with similar results; error bars representSEM. One-way ANOVA and Tukey’s multiple comparison test vs. untreated controls, *P < 0.05; **P < 0.001; ***P < 0.0001. Scale bar= 30 μm. f RT-PCRwas used to quantify SPARC mRNA. 100 U/ml TNF-α treatment significantly (P < 0.05) increased SPARC mRNA expression. Bars represent theaverage densitometric value from n = 4 experiments; error bars represent SEM. One-way ANOVA and Tukey’s multiple comparison test vs.untreated controls, *P < 0.05Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 11 of 17lowered both ZO-1 (0.768, P < 0.05) and occludin (0.796,P < 0.001) protein levels (Fig. 6b). Concentrations inthe range of 1 to 10 μg/ml tended to decrease ZO-1and occludin protein levels compared to baselineSPARC serum concentrations (approximately 0.1 μg/ml)(Fig. 6a, b). Representative immunoblots and normalizedquantifications depict the trends observed (Fig. 6a).DiscussionSPARC expression during cerebral endothelial growth andbarrier establishmentThis is to our knowledge the first study to examineSPARC expression and regulation in an in vitro BBBmodel. This report describes novel in vitro evidence thatcerebral endothelia constitutively express SPARC duringproliferative stages of growth and down-regulate it as cellsform a monolayer and establish a BBB-like phenotypeas assessed by Western blot and quantitative immuno-cytochemistry. This is consistent with in vivo observationswhere SPARC expression was enriched in developingcerebrovasculature yet diminished in mature vessels(Roskams lab, unpublished observations) [22, 23, 43].hCMEC/D3 cells in our experiments were propagatedin media containing various concentrations of serumwhich reproducibly correlated with SPARC expression.Cell cultures replenished for 24 h in media with increasingserum concentrations were found to express increasedlevels of SPARC. Serum is a known mitogen in this cellline used as a positive control and stimulus of cell growthand proliferation [41]. Our experiments define serum as apositive regulator of SPARC expression in cerebral endo-thelia during proliferative stages of growth.We performed dual immunocytochemistry for SPARCand a proliferation marker Ki-67 to investigate the link-age between SPARC and endothelial proliferation—avital parameter of angiogenesis and cerebrovascular re-sponse to injury. SPARC positively correlated with Ki-67in subconfluent hCMEC/D3 cultures. Increased SPARCimmunoreactivity was associated with a higher percentageof Ki-67-positive nuclei while confluent monolayersshared low SPARC and low Ki-67 levels, providing a directcorrelation between SPARC expression and proliferation.Brain endothelial cells have been shown to co-expresshigh levels of SPARC and the marker of proliferationPCNA (proliferating cell nuclear antigen) in developingbut not mature brain capillaries [43]. SPARC produced bydividing subconfluent cerebral endothelia may promotefocal adhesion disassembly and inhibit cell spreadingmediating cell migration and proliferation during CNSangiogenesis [44].Elevated SPARC mRNA and protein expression hasbeen observed in endothelial cells involved in angiogen-esis [18, 45]. Bovine aortic and human microvascularendothelial cells that spontaneously form tube-like vesselstructures revealed that new vessel branches areenriched in SPARC and possess a high mitotic index.Confocal microscopy in an in vivo model of angiogen-esis similarly showed intense SPARC staining in small-caliber, newly formed, blood vessels and negligibleFig. 5 SPARC influences hCMEC/D3 barrier function measuredby TEER and permeability. a Confluent hCMEC/D3 on transwellinserts were replenished with media containing rh-SPARC (0.01,0.1, 1, 10 μg/ml) for 24 h then tested for TEER and transendothelialdiffusion of FITC-labeled dextran (3 and 10 kDa). Compared tothe untreated conditions, TEER measurements reflect a dose-dependent reduction in electrical impedance with increasingrh-SPARC concentration reaching significance with 1 μg/ml orgreater rh-SPARC. Permeability assays using b 3 and c 10 kDaFITC-dextran revealed exposure of hCMEC/D3 to 0.1–10 μg/mlrh-SPARC enhanced permeability to both by 24 h of exposure.One-way ANOVA and Tukey’s multiple comparison test vs.untreated controls, *P < 0.05Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 12 of 17staining in larger developmentally mature vessels [17, 36].In the heart, SPARC coincided with blood vessel neo-vascularization on aortic valves yet not normal avascularvalves [46]. Notably, SPARC expression was upregulated4.2-fold in bovine aortic endothelial cells and 10-fold inrat cerebral endothelial cell (RVEC) cultures that spontan-eously organized into capillary tubes in vitro [17, 21, 36],suggesting a role for SPARC in “de novo” morphogenesisof new microvessel angioarchitecture (vasculogenesis) andthe growth of new vessels from established ones (angio-genesis). Taken together, these data suggest SPARC ishighly expressed in developmentally new blood vesselsundergoing angiogenesis in vivo, consistent with our invitro data showing that SPARC expression in human cere-bral endothelia corresponds with proliferation and the ini-tial steps of brain microvessel formation.SPARC expression in vivo coincides with developingblood vessels during embryogenesis and postnatal devel-opment, but not in mature blood vessels of the normaladult CNS. To that effect, in situ hybridization showedSPARC mRNA levels enriched in migrating endotheliaand pia-derived blood vessels of embryonic and postna-tal tissue progressively down-regulated with maturity inadult cerebrovasculature [22, 23, 33, 47]. In the normal(unlesioned) adult CNS, SPARC protein expression ispredominantly restricted to neurogenic subventricularzones, choroid plexus, actively surveying microglia, spe-cialized radial glia (Bergmann and Müller glia), and as-trocytes [23, 34]. These data together suggest SPARC maybe expressed by cerebral endothelia in a spatiotemporalmanner consistent with a role in CNS vascularization.Regulation of SPARC expression during inflammationHuman microvascular blood vessels express elevatedlevels of SPARC at sites of injury, infection, and neoplasiarelative to healthy tissue [28, 33, 44, 48]. Expression ofSPARC in adult tissue and plasma increases in response tovarious environmental stressors, including heat shock,heavy metal toxicity, endotoxin [49], angiogenesis [21, 36],and wound repair [20]. Using an in vitro model of theBBB, this study offers the first description of inflammatoryregulation of SPARC expression in cerebral endothelia. Inthe present study, the hCMEC/D3 cells exposed to theprototypical injury response cytokine TNF-α and theendotoxin LPS increased SPARC levels compared to theuntreated cells. Notably, IFN-γ combined with TNF-αabrogated the SPARC induction otherwise observed inresponse to TNF-α alone, suggesting cross-talk betweenFig. 6 SPARC and TNF-α regulate hCMEC/D3 tight junction protein expression. a Confluent hCMEC/D3 cultures treated with SPARC (0.1, 1, 10 μg/ml)or TNF-α (200 U/ml) for 24 h were analyzed for tight junction (TJ) expression. TJ expression was quantified by immunoblot analysis of hCMEC/D3lysates with the TJ-negative human astrocytoma cell line (Astro) ccf-sttg1 (Sigma) lysate as a negative control. A reduction in endothelial TJ expressionwas observed with increasing SPARC exposure above 1 μg/ml. b Increasing exogenous SPARC from 0.1 to 1 μg/ml lowered both ZO-1 (by 23 %) andoccludin (by 20 %) expression. The bars represent the mean expression from 5 (ZO-1) and 4 (occludin) experiments, and error bars represent the SEM.**P < 0.001; ***P < 0.0001 by ANOVA and Student-Newman-Keuls multiple comparisons testAlkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 13 of 17TNF-α and IFN-γ that may negatively regulate levels ofSPARC in response to inflammation at the BBB. Althoughthe present study is the first mention of pro-inflammatoryregulation of SPARC expression in cerebral endothelia,other CNS injury models and neoplastic syndromes havehighlighted the association between inflammation andSPARC-enriched blood vessels [48, 50].In the normal adult CNS, SPARC levels are low andexpression is primarily restricted to astrocytes andmicroglia of synaptic rich regions but typically not bloodvessels. That balance appears to shift during a numberof inflammatory states and models. In a cortical woundmodel of CNS injury, SPARC mRNA was increased inblood vessels proximal to the wound [33]. During the peakphase of experimental autoimmune encephalomyelitis(EAE), an animal model of the chronic inflammatory de-myelinating disease MS, CD31-positive blood vessels be-come intensely SPARC positive by immunohistochemistry(Roskams Lab, unpublished observations). Extending be-yond the vasculature to glial cells in close proximity, an-other study demonstrated that up-regulation of SPARCsecreted by IFN-γ-treated astrocytes induced apoptosisof myelin-specific autoreactive CD4+ T cells that other-wise mediate the pathogenesis of EAE and MS, sugges-ting a role for SPARC in protecting the CNS againstautoimmune-mediated damage during the course ofneuroinflammation (Hara et al.). Furthermore, increasedbioavailability of SPARC at the neurovascular niche maypromote CNS regeneration by supporting neurite out-growth after neuronal injury [34].Although the role of SPARC at sites of injury remainsto be fully characterized, experimental evidence suggestsSPARC promotes the effector phase of inflammation byfacilitating leukocyte transmigration. A study using HUVECand primary lung and cardiac endothelial cell culturesdemonstrated that SPARC is a VCAM1 (vascular celladhesion molecule) ligand that induces cytoarchitecturalrearrangement and intercellular gap formation, two pro-cesses critical for leukocyte trafficking [38]. In SPARC-nullmice, inflammatory recruitment of neutrophils,eosinophils, and monocytes/macrophages to an in-flamed peritoneum was compromised compared tothose of the wildtype [38]. To that effect, SPARC-nullmice immunized with myelin oligodendrocyte glycopro-tein (MOG35–55) exhibited delayed onset and reducedseverity of EAE, consistent with milder demyelinationand immune infiltration, as observed by histopatho-logical examination [51]. These findings support the ar-gument that regulation of SPARC expression may playa key role in the mediation of inflammation through itsinfluence on immune trafficking into inflamed tissue.This study reports that two inflammatory molecules,TNF-α and LPS, increase SPARC expression in humancerebral endothelial cells enriching SPARC bioavailabilityat the neurovascular niche. Astrocyte-derived SPARCwas recently shown to specifically antagonize the synap-togenic effect of SPARC-like 1 and decrease the numberof axosomatic synapses in contact with lumbar spinalmotor neurons, worsening the severity of paralysis duringEAE in mice [52]. In both remitting and non-remittingEAE models, SPARC-like 1/SPARC mRNA and proteinratios inversely correlated with clinical paralysis severityscores. Decreased SPARC-like 1/SPARC ratios also corre-sponded with diminished expression of presynaptic andpostsynaptic proteins as well as the number of axosomaticsynaptic contacts. In vitro experiments using corticalastrocytes further revealed that T cell derived pro-inflammatory cytokines including TNF-α and IL-17inhibit SPARC-like 1 and promote SPARC expression,favoring synaptic retraction, while anti-inflammatorycytokines such as IL-10 did the opposite thus shiftingthe ratio towards synaptic stabilization [52]. These datataken together support a model whereby SPARC up-regulation by pro-inflammatory cytokines drives motorsynaptic disassembly, contributing to neurological defi-cits during the course of neuroinflammation. Indeed,our study provides in vitro evidence that cerebral endo-thelial cells are an abundant source of SPARC at theneurovascular niche when exposed to pro-inflammatorycytokines or bacterial endotoxin. During development,elimination of excess synaptic contacts is thought to re-fine circuit function [53]; however, whether this is anadaptive or maladaptive response in the context of CNSinflammation is unclear. It is thus reasonable to con-tend that regulation of endothelial-derived SPARCunder inflammatory conditions may contribute to tran-sient paralysis and neurological dysfunction by destabi-lizing synapses in the presence of neuroinflammation.SPARC regulation of BBB integrity and permeabilityAs BBB disruption is among the earliest events in severalCNS pathologies such as MS, we examined the influenceof SPARC on barrier function and TJ expression usingan in vitro BBB model. The involvement of SPARCin the development, maintenance, and breakdown ofblood-neural barriers is implicated by its spatiotemporallocalization at the BBB and blood-cerebral spinal fluidbarrier (BCSFB) during development and inflammation.Ependymal cells of the BCSFB lining brain ventriclesand the choroid plexus express SPARC during embryo-genesis and postnatal development but limit expressionwith maturity to barrier sites of the lateral ventricle inadult mice [23]. Analogously, brain endothelial cells ofdeveloping blood vessels arising from the pia materstrongly express SPARC during embryogenesis butdown-regulate it during early postnatal development asthey mature and establish a BBB [22]. In adult mice,SPARC is concentrated in mature astrocytic endfeet ofAlkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 14 of 17the BBB [23]. Astrocytes and the soluble mediators theyproduce can up-regulate TJ and transporter expressionon brain endothelia and are thought to be major regula-tors of BBB development [2, 54]. The localization ofSPARC to astrocytic endfeet and cerebral endotheliaplace SPARC in an ideal niche to induce, maintain, ordisrupt blood-neural barrier characteristics [11]. SPARCin astrocytic endfeet could reflect an uptake of endothelial-derived SPARC, an accumulation of astrocyte-derivedSPARC into its end feet, or both.TJ protein levels were quantified in lysates from con-fluent hCMEC/D3 cultures treated for 24 h with variousdoses of rh-SPARC or rhTNF-α. The reduction of TJprotein expression provides the first documentation ofSPARC regulation of TJ modulation in endothelial cells.An important initial consideration, here, is the physio-logical relevance of SPARC doses used. In normalhealthy individuals, cerebral endothelial cells reside inplasma concentrations of SPARC between 0.1 and 0.8μg/ml, while increased concentrations have been associ-ated with pathological conditions such as neoplasia,trauma, heart, and kidney diseases [25, 27, 30, 31]. Inour experiments, elevated SPARC concentrations, mod-eling that expected during inflammation, lowered ZO-1and occludin expression, suggesting that SPARC may in-deed be involved in mediating TJ loss and associatedBBB breakdown at supraphysiological levels. Functionalassays performed in parallel showed that increasing con-centrations of recombinant human SPARC applied ex-ogenously increased transendothelial permeability of cellmonolayers as evidenced by increased dextran diffusionand decreased TEER impedence. TNF-α applied exogen-ously as a positive control decreased total expression ofall TJ proteins studied, consistent with previous reportsin the literature [4, 8, 14]. Hydrocortisone offered someprotection against the reduction of ZO-1 but not occlu-din. This may be explained by a greater ability of hydro-cortisone to drive ZO-1 expression than occludin inhCMEC/D3 cultures [4].The prevailing understanding from these experimentswas that SPARC increases transendothelial permeabilityand disrupts endothelial barrier function in vitrothrough its effect on the relative state of endothelial dif-ferentiation. SPARC has also been shown to modulatetransendothelial permeability through protein tyrosinekinase (PTK) phosphorylation signaling. The potent PTKinhibitor, herbimycin A, diminished SPARC-inducedchanges in permeability. A marked 12-fold increase inphosphotyrosine-containing proteins was immunoloca-lized to interendothelial borders within 1 h of SPARCtreatment concomitant to barrier opening as evidencedby immunocytochemistry and BSA diffusion permeabilityassay, respectively. These findings are consistent with ourdata demonstrating decreased TJ expression and increasedbarrier permeability in SPARC-treated hCMEC/D3monolayers, suggesting a putative role for SPARC inneurological conditions such as MS, characterized bymicrovascular TJ loss and barrier disruption [10]. Clearly,SPARC has the potential to influence BBB integrity, butthe exact nature and outcome of this influence—be itreparative or pathological—is yet to be determined.In the present study, SPARC concentrations expectedduring angiogenesis, inflammation, and wound repairlowered ZO-1 and occludin expression and increasedendothelial barrier permeability. TJ depletion is onemechanism conventionally associated with BBB break-down in diseases such as MS. VEGF-induced BBB break-down is associated with occludin down-regulation andsevere disability in EAE [9]. Transient global cerebralischemia diminished SPARC mRNA expression anddepleted SPARC in the basement membranes of laser-captured microdissected (LCM) blood vessels after 24 hof reperfusion [7]. Functional assays showed that thepost-ischemic cerebral microvessels were more perme-able than the control vessels. Here, mild post-ischemichypothermia (~33 °C) attenuated loss of SPARC contentin the basement membrane and protected against BBBbreakdown, suggesting that the release of SPARC fromthe basement membrane was responsible for barrieropening. In this setting, SPARC released from the base-ment membrane after ischemic injury could act in anautocrine/paracrine manner and drive barrier disruptionand vessel leakiness through endothelial VCAM1 signal-ing. In a recent article, tumor-derived SPARC was foundto trigger endothelial paracellular permeability in primaryhuman umbilical vein endothelial cell (HUVEC) mono-layers via VCAM1 and p38 signaling, thus disruptingendothelial monolayer integrity [55]. Taken together,our study establishes SPARC as a novel regulator ofparacellular cerebrovascular permeability and TJ expres-sion, revealing a targetable interaction in CNS pathologyfor prevention of barrier disruption.ConclusionsIn summary, we describe elevated SPARC expression incerebral endothelial cells under developmental condi-tions modeling angiogenesis and differential regulationin the presence of inflammatory cytokines TNF-α, IFN-γ,and the bacterial endotoxin LPS. Exposure to SPARCat levels greater than those observed under normalphysiological conditions increases barrier permeabilityand decreases TJ expression of ZO-1 and occludin.Microvascular endothelial cell exposure to SPARC maycontribute to pathological processes in the CNS or fa-cilitate its response to injury. Our study defines SPARCas an attractive target for further scientific inquiry withregard to its role in both pathological and potentiallyreparative processes at the neurovascular niche.Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 15 of 17Additional fileAdditional file 1: Figure S1. Subconfluent hCMEC/D3 regions exhibitmore intense SPARC expression than confluent regions by regionalimmunocytochemistry analysis. SPARC intensity was measured in meanpixel intensity (MPI ± SD) for selected cell-covered regions. SPARC staining inthe subconfluent cultures (3.85 ± 0.49, n = 21 images, 7 from each triplicatewell) was greater than that in the confluent cultures (2.83 ± 0.43, n = 14images, 7 from each duplicate well). Bars represent the average of resultsfrom n = 21 and n = 14 images, respectively. Error bars represent SD. Datawere analyzed for regional SPARC MPI in one experiment. Mann-Whitneycomparison test, ***P < 0.0001. (TIF 40 kb)AcknowledgementsNot applicable.FundingThis study including the consumables and personnel support was supportedby funds from the Department of Pathology, UBC and the Brain ResearchCentre and partially supported by the MS Society of Canada. SA wassupported by a CIHR Frederick Banting and Charles Best Canada GraduateScholarship, JR by a CIHR grant, and JB by an MS Society of CanadaPost-Doctoral Fellowship. JQ is supported by an MS Society of CanadaDonald Paty Career Development Award.Availability of data and materialsAll relevant data has been shared in the manuscript; there are no relevantmaterials to be made available.Authors’ contributionsSA performed the cell culture, the immunocytochemistry, and the Westernblotting experiments to analyze tight junction and SPARC expression andperformed data collection and analyses as well as prepared the first draftof the manuscript and assisted in editing throughout. JB performed thecell culture with Western blot and permeability experiments with datacollection and analyses. LZ performed the SPARC PCR experiments.PR, JR, JB, and JQ formulated the research plan and series of experiments toaddress the aims of the study. JQ oversaw the experimental approach anddesign, reviewed the data and analyses for interpretation of the data, andoversaw the manuscript preparation including revising the manuscriptcritically for technical and scientific accuracy. All authors have reviewed themanuscript in entirety and are in agreement with the content of this work.Competing interestsThe authors declare that they have no competing interests.Consent for publicationNot applicable.Ethics approval and consent to participateNot applicable.Author details1Department of Pathology and Laboratory Medicine, University of BritishColumbia, Vancouver, BC, Canada. 2Department of Zoology, University ofBritish Columbia, Vancouver, BC, Canada. 3Department of Medicine, Divisionof Neurology, University of British Columbia, Vancouver, BC, Canada.4Sozialstiftung Bamberg, Klinikum am Bruderwald, Neurologische Klinik,Buger Str. 80, Bamberg 96049, Germany.Received: 2 February 2016 Accepted: 12 July 2016References1. Hawkins BT, Davis TP. The blood-brain barrier/neurovascular unit in healthand disease. Pharmacol Rev. 2005;57(2):173–85.2. Abbott NJ, Ronnback L, Hansson E. Astrocyte-endothelial interactions at theblood-brain barrier. Nat Rev Neurosci. 2006;7(1):41–53.3. Stamatovic SM, Keep RF, Andjelkovic AV. Brain endothelial cell-cell junctions:how to “open” the blood brain barrier. Curr Neuropharmacol. 2008;6(3):179–92.4. Forster C, et al. Differential effects of hydrocortisone and TNFalpha on tightjunction proteins in an in vitro model of the human blood-brain barrier.J Physiol. 2008;586(7):1937–49.5. Bennett J, et al. Blood-brain barrier disruption and enhanced vascularpermeability in the multiple sclerosis model EAE. J Neuroimmunol. 2010;229(1-2):180–91.6. Cucullo L, et al. Immortalized human brain endothelial cells and flow-basedvascular modeling: a marriage of convenience for rational neurovascularstudies. J Cereb Blood Flow Metab. 2008;28(2):312–28.7. Baumann E, et al. Post-ischemic hypothermia attenuates loss of the vascularbasement membrane proteins, agrin and SPARC, and the blood-brain barrierdisruption after global cerebral ischemia. Brain Res. 2009;1269:185–97.8. Aveleira CA, et al. TNF-{alpha} signals through PKC{zeta}/NF-{kappa}B to alterthe tight junction complex and increase retinal endothelial cell permeability.Diabetes. 2010;59(11):2872–82.9. Argaw AT, et al. VEGF-mediated disruption of endothelial CLN-5 promotesblood-brain barrier breakdown. Proc Natl Acad Sci U S A. 2009;106(6):1977–82.10. Minagar A, Alexander JS. Blood-brain barrier disruption in multiple sclerosis.Mult Scler. 2003;9(6):540–9.11. Kim JH, et al. Blood-neural barrier: intercellular communication at glio-vascularinterface. J Biochem Mol Biol. 2006;39(4):339–45.12. Brown RC, Morris AP, O’Neil RG. Tight junction protein expression andbarrier properties of immortalized mouse brain microvessel endothelial cells.Brain Res. 2007;1130(1):17–30.13. Poller B, et al. Regulation of BCRP (ABCG2) and P-glycoprotein (ABCB1) bycytokines in a model of the human blood-brain barrier. Cell Mol Neurobiol.2010;30(1):63–70.14. Nagyoszi P, et al. Expression and regulation of toll-like receptors in cerebralendothelial cells. Neurochem Int. 2010;57(5):556–64.15. Poller B, et al. The human brain endothelial cell line hCMEC/D3 as a humanblood-brain barrier model for drug transport studies. J Neurochem. 2008;107(5):1358–68.16. Weksler BB, et al. Blood-brain barrier-specific properties of a human adultbrain endothelial cell line. FASEB J. 2005;19(13):1872–4.17. Iruela-Arispe ML, Hasselaar P, Sage H. Differential expression of extracellularproteins is correlated with angiogenesis in vitro. Lab Invest. 1991;64(2):174–86.18. Kato Y, et al. Induction of SPARC by VEGF in human vascular endothelialcells. Biochem Biophys Res Commun. 2001;287(2):422–6.19. Goldblum SE, et al. SPARC (secreted protein acidic and rich in cysteine)regulates endothelial cell shape and barrier function. Proc Natl Acad Sci U S A.1994;91(8):3448–52.20. Bradshaw AD, Sage EH. SPARC, a matricellular protein that functions in cellulardifferentiation and tissue response to injury. J Clin Invest. 2001;107(9):1049–54.21. Iruela-Arispe ML, et al. Expression of SPARC during development of thechicken chorioallantoic membrane: evidence for regulated proteolysis invivo. Mol Biol Cell. 1995;6(3):327–43.22. Mendis DB, Brown IR. Expression of the gene encoding the extracellularmatrix glycoprotein SPARC in the developing and adult mouse brain.Brain Res Mol Brain Res. 1994;24(1-4):11–9.23. Vincent AJ, Lau PW, Roskams AJ. SPARC is expressed by macroglia and microgliain the developing and mature nervous system. Dev Dyn. 2008;237(5):1449–62.24. Malaval L, Ffrench M, Delmas PD. Circulating levels of osteonectin in normalsubjects and patients with thrombocytopenia. Bone Miner. 1990;9(2):129–35.25. Malaval L, Fournier B, Delmas PD. Radioimmunoassay for osteonectin.Concentrations in bone, nonmineralized tissues, and blood. J Bone Miner Res.1987;2(5):457–65.26. Serebruany VL, et al. Increased soluble platelet/endothelial cellular adhesionmolecule-1 and osteonectin levels in patients with severe congestive heartfailure. Independence of disease etiology, and antecedent aspirin therapy.Eur J Heart Fail. 1999;1(3):243–9.27. Porte H, et al. Overexpression of stromelysin-3, BM-40/SPARC, and METgenes in human esophageal carcinoma: implications for prognosis. ClinCancer Res. 1998;4(6):1375–82.28. Porter PL, et al. Distribution of SPARC in normal and neoplastic humantissue. J Histochem Cytochem. 1995;43(8):791–800.29. Yunker CK, et al. SPARC-induced increase in glioma matrix and decrease invascularity are associated with reduced VEGF expression and secretion.Int J Cancer. 2008;122(12):2735–43.30. Pichler RH, et al. SPARC is expressed by mesangial cells in experimental mesangialproliferative nephritis and inhibits platelet-derived-growth-factor-medicatedmesangial cell proliferation in vitro. Am J Pathol. 1996;148(4):1153–67.Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 16 of 1731. Shankland SJ, et al. Changes in cell-cycle protein expression duringexperimental mesangial proliferative glomerulonephritis. Kidney Int.1996;50(4):1230–9.32. Macko RF, et al. Increased circulating concentrations of the counteradhesiveproteins SPARC and thrombospondin-1 in systemic sclerosis (scleroderma).Relationship to platelet and endothelial cell activation. J Rheumatol. 2002;29(12):2565–70.33. Mendis DB, Ivy GO, Brown IR. SPARC/osteonectin mRNA is induced in bloodvessels following injury to the adult rat cerebral cortex. Neurochem Res.1998;23(8):1117–23.34. Au E, et al. SPARC from olfactory ensheathing cells stimulates Schwann cellsto promote neurite outgrowth and enhances spinal cord repair. J Neurosci.2007;27(27):7208–21.35. Nozaki M, et al. Loss of SPARC-mediated VEGFR-1 suppression after injuryreveals a novel antiangiogenic activity of VEGF-A. J Clin Invest. 2006;116(2):422–9.36. Iruela-Arispe ML, Diglio CA, Sage EH. Modulation of extracellular matrixproteins by endothelial cells undergoing angiogenesis in vitro. ArteriosclerThromb. 1991;11(4):805–15.37. Hara H, et al. Identification of astrocyte-derived immune suppressorfactor that induces apoptosis of autoreactive T cells. J Neuroimmunol.2011;233(1-2):135–46.38. Kelly KA, et al. SPARC is a VCAM-1 counter-ligand that mediates leukocytetransmigration. J Leukoc Biol. 2007;81(3):748–56.39. Cheutin T, et al. Three-dimensional organization of pKi-67: a comparativefluorescence and electron tomography study using FluoroNanogold.J Histochem Cytochem. 2003;51(11):1411–23.40. Verheijen R, et al. Ki-67 detects a nuclear matrix-associated proliferation-related antigen. II. Localization in mitotic cells and association withchromosomes. J Cell Sci. 1989;92(Pt 4):531–40.41. Chen PL, Easton AS. Evidence that tumor necrosis factor-related apoptosisinducing ligand (TRAIL) inhibits angiogenesis by inducing vascular endothelialcell apoptosis. Biochem Biophys Res Commun. 2010;391(1):936–41.42. Young BA, Wang P, Goldblum SE. The counteradhesive protein SPARCregulates an endothelial paracellular pathway through protein tyrosinephosphorylation. Biochem Biophys Res Commun. 1998;251(1):320–7.43. Lane TF, et al. SPARC is a source of copper-binding peptides that stimulateangiogenesis. J Cell Biol. 1994;125(4):929–43.44. Chlenski A, et al. Anti-angiogenic SPARC peptides inhibit progression ofneuroblastoma tumors. Mol Cancer. 2010;9:138.45. Reed MJ, et al. Differential expression of SPARC and thrombospondin 1 inwound repair: immunolocalization and in situ hybridization. J HistochemCytochem. 1993;41(10):1467–77.46. Charest A, et al. Distribution of SPARC during neovascularisation ofdegenerative aortic stenosis. Heart. 2006;92(12):1844–9.47. Mothe AJ, Brown IR. Differential mRNA expression of the relatedextracellular matrix glycoproteins SC1 and SPARC in the rat embryonicnervous system and skeletal structure. Brain Res. 2001;892(1):27–41.48. Chlenski A, Cohn SL. Modulation of matrix remodeling by SPARC inneoplastic progression. Semin Cell Dev Biol. 2010;21(1):55–65.49. Liu A, et al. The counteradhesive proteins, thrombospondin 1 and SPARC/osteonectin, open the tyrosine phosphorylation-responsive paracellularpathway in pulmonary vascular endothelia. Microvasc Res. 2009;77(1):13–20.50. Said NA, et al. SPARC ameliorates ovarian cancer-associated inflammation.Neoplasia. 2008;10(10):1092–104.51. Piconese S, et al. The matricellular protein SPARC supports folliculardendritic cell networking toward Th17 responses. J Autoimmun.2011;37(4):300–10.52. Blakely PK, et al. Astrocyte matricellular proteins that control excitatorysynaptogenesis are regulated by inflammatory cytokines and correlatewith paralysis severity during experimental autoimmune encephalomyelitis.Front Neurosci. 2015;9:344.53. Lopez-Murcia FJ, Terni B, Llobet A. SPARC triggers a cell-autonomous programof synapse elimination. Proc Natl Acad Sci U S A. 2015;112(43):13366–71.54. Gaillard PJ, et al. Astrocytes increase the functional expression of P-glycoprotein in an in vitro model of the blood-brain barrier. Pharm Res.2000;17(10):1198–205.55. Tichet M, et al. Tumour-derived SPARC drives vascular permeability andextravasation through endothelial VCAM1 signalling to promote metastasis.Nat Commun. 2015;6:6993.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Alkabie et al. Journal of Neuroinflammation  (2016) 13:225 Page 17 of 17


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items