THE JOURNAL OF BIOLOGICAL CHEMISTRY
`0 1994 by The American Society
`for Biochemistry and Molecular Biology Inc
`
`Vol 269 No 8 Issue of February 25 pp 60726082
`Printed in USA
`
`1994
`
`Albondinmediated Capillary Permeability to Albumin
`DIFFERENTIAL ROLE OF RECEPTORS IN ENDOTHELIAL TRANSCYTOSIS AND ENDOCYTOSIS OF NATIVE
`AND MODIFIED ALBUMINS
`
`Received
`
`for publication October 18 1993 and in revised form November 12 1993
`
`Jan E Schnitzer and Phil Oh
`From the Department of Medicine and Pathology Division of Cellular and Molecular Medicine Institute for Biomedical
`Engineering University of California San Diego School of Medicine La Jolla California 920930651
`
`Specific albumin binding to the surface of endothe
`lium initiates its transcytosis across continuous endo
`thelium via noncoated plasmalemmal vesicles Past
`work has identified several putative albumin binding
`proteins SPARC gp60 gp30 and gp18 In this study we
`examined the specific role of these proteins in the bind
`ing of bovine serum albumin BSA to endothelium The
`60kDa albumin binding protein gp60 now called al
`bondin was purified from cultured rat microvascular
`endothelial cells for antibody production Antialbondin
`antibodies agp60 specifically recognized albondin ex
`pressed by rat bovine and human endothelial cells
`EC agp60 and unlabeled native BSA inhibited 125I BSA
`binding to confluent EC monolayers and to both albon
`din and SPARC extracts immobilized on filters Modifi
`cation of BSA by maleic anhydride treatment MalBSA
`or by surface adsorption to colloidal gold particles A
`Au renders the ligand specific for gp30 and gp18 while
`eliminating its ability to inhibit 125IBSA binding to EC
`and to both albondin and SPARC extracts Mal BSA and
`AAu interacted with EC via distinct binding sites not
`recognizing native BSA EC internalization of 125I BSA is
`inhibited by unlabeled BSA and agp60 IgG but not non
`immune IgG MalBSA or AAu Conversely internaliza
`tion of modified BSA is inhibited by unlabeled modified
`BSA but not BSA or agp60 IgG Specific 125I BSA transen
`EC monolayers in vitro is inhibited 0 by unlabeled
`dothelial transport in rat lungs perfused in situ and for
`BSA and agp60 IgG but not nonimmune IgG and Mal
`BSA No specific transport of 125I Mal BSA is detected
`across bovine lung microvascular endothelial cell mono
`layers only paracellular andor fluid phase transport
`Low temperatures reduce BSA transport by 5 fold and
`Mal BSA by 2 fold Interestingly 3 fold more native BSA
`is transported than Mal BSA at 37 °C whereas at 810 °C
`they are nearly equal suggesting facilitation of BSA
`transport Cumulatively it appears that gp30 and gp18
`mediate the binding endocytosis and degradation of
`modified albumins whereas albondin mediates native
`its tran
`albumin binding which significantly enhances
`scytosis and capillary permeability
`
`Institutes of Health
`This work was supported in part by National
`Grant HL43278 and by two grantsinaid and an Established
`Investi
`gator award from the American Heart Association National and Cali
`fornia affiliates The costs of publication of this article were defrayed in
`part by the payment of page charges This article must therefore be
`in accordance with 18 USC Section
`hereby marked advertisement
`1734 solely to indicate this fact
`lb whom correspondence should be addressed and present address
`Harvard Medical School Beth Israel Hospital Dept of Pathology Re
`search North 330 Brookline Ave Boston MA 02215 Tel 617735
`4343 Fax 6177352943
`
`The endothelium lining many microvascular
`beds of many
`organs is the significant barrier to the transvascular
`transport
`of plasma proteins and metabolites from the capillary vessels to
`fluid Serum proteins circulating within the
`the interstitial
`cell surface ei
`vasculature may interact with the endothelial
`or to hinder transcapillary exchange Such
`ther to enhance
`interactions appear to be required to establish normal capillary
`permeability 17 In a phenomenon called the serum effect
`the interaction of circulating proteins such as albumin and
`cell surface of many micro
`orosomucoid with the endothelial
`beds 5 814 significantly restricts transcapillary
`vascular
`transport of a wide spectrum of molecules 17 Conversely
`specific binding mechanisms may exist to enhance
`the transen
`dothelial exchange of a selective group of ligands in order to
`provide specific nutrients to tissue cells in accordance with
`their particular needs Endothelial surface receptors or trans
`for plasma molecules including transferrin 15 16
`insulin 17 orosomucoid 8 ceruloplasmin 18 lipoproteins
`porters
`19 and even glucose 20 have been detected in select
`tissues
`and may mediate specific transport into andor across the en
`dothelium A few direct anti endothelial cell
`receptor conju
`gates have been created for targeting drugs to specific organs
`21 Vectorial
`receptor mediated transport may also occur
`in
`the opposite direction Lipoprotein lipase synthesized by un
`derlying adipocytes and myocytes is transported by a saturable
`mechanism across
`the endothelium to the lumina surface
`where it stays bound and active 22
`The serum protein albumin is a multifunctional carrier pro
`fluid homeostasis
`tein that contributes
`to cardiovascular
`through osmotic forces and permselectivity serum effects
`26 and to the transport distribution and metabolism of a
`wide variety of small bloodborne molecules including fatty
`acids sterols amino acids hormones metal ions and numer
`ous drugs 23 24 It has a dual
`role in the permeability of
`beds lined with a continuous type of en
`many microvascular
`dothelium As a major contributor to the serum effect albumin
`binds the surface of endothelium 5 914 and reduces the
`flux of water solutes and macromolecules 26
`transvascular
`Albumin binding to the endothelial glycocalyx creates a more
`via cell membrane
`restrictive endothelial barrier possibly
`transduced signals that modulate intracellular calcium levels
`25 andor the formation of an additional permselective barrier
`within transendothelial transport pathways 5 6 13 14 26
`Conversely albumin can enhance
`specific transport by acting
`as a carrier protein in the circulation for its many small ligands
`23 24 Albumin is actively involved in the presentation
`of
`important ligands such as fatty acids to a variety of tissue cells
`including adipocytes for the storage of fat 27 Leydig cells for
`hormonal production 28 and myocytes for energy production
`via oxidative metabolism 29 Because the endothelial barrier
`from the circulating
`can prevent macromolecular
`transport
`is unclear whether certain
`blood to the extravascular
`space it
`
`6072
`
`Abraxis EX2012
`Actavis LLC v Abraxis Bioscience LLC
`1PR201701101 1PR201701103 1PR201701104
`
`

`

`it
`
`via
`
`endothelia play a direct role in selective albumin transport
`is known that i many continuous endothelia have
`However
`specific albumin binding sites apparently for transcytosis
`an abundant population of noncoated plasmalemma vesicles 9
`1113 30 ii
`ligands bound to albumin such as fatty acids
`appear to increase its endothelial binding and transcytosis 24
`cultured endothelium can internalize and release al
`and iii
`bumin with little degradation 31 32 Native albumin trans
`port across continuous endothelium via noncoated plasmalem
`mal vesicles or caveolae 11 12 30 provides
`a carrier
`mechanism for the facilitated delivery of its small ligands from
`the circulating blood to the interstitium in a manner consistent
`with the nutritional
`requirements of the underlying tissue
`
`cells
`Albumin specifically binds the endothelial
`cell surface in a
`manner consistent with a receptorligand interaction 13 14
`have been sug
`albumin binding proteins
`Various putative
`gested over the past decade 3138 Two of these proteins gp30
`and gp18 were recently shown to recognize modified albumins
`much more avidly than native albumin 31 and appear
`to
`mediate the high affinity binding of modified albumins to the
`cell surface ultimately for endocytosis and lysosomal degrada
`tion 31 32 Two other albumin binding proteins gp60 14 34
`and SPARC 33 are functionally and immunologically related
`possibly via a common albumin binding domain 37
`In this study we show that direct binding of native albumin
`to the 60kDa endothelial
`cell surface protein gp60 now re
`named albondin can mediate its transendothelial
`transport
`Specific antibodies raised against purified albondin inhibit not
`only albumin binding to the endothelial
`cell surface but also
`capillary permeability to albumin by reducing its internaliza
`the endothelium In addition we
`tion and transport across
`show that modification of albumin can prevent
`its albondin
`facilitated transport and that native and modified albumins
`interact distinctly via different binding sites on the cell surface
`for differential cellular internalization and processing
`
`EXPERIMENTAL
`
`PROCEDURES
`
`and other supplies were obtained from the fol
`MaterialsReagents
`lowing sources Fetal calf serum and phosphate buffered saline PBS1
`were from Life Technologies
`Inc Gelatin was from DifcoCrystallized
`bovine serum albumin BSA was
`from ICN Biochemicals
`Triticum
`vulgaris agglutinin wheat germ agglutinin conjugated to agarose
`beads BSAagarose
`beads Nacetylglucosamine bovine transferrin
`and orosomucoid was from Sigma Dulbeccos modified Eagles medium
`DMEM was from Irvine Scientific IODOGEN 1456tetrachloro
`Triton X100 SDS and the bicinchoninic acid
`3a6adiphenylglycouril
`protein assay were from Pierce Chemical Co Na1251 was from Amer
`sham Corp All
`tissue culture plasticware was from Costar Cambridge
`MA or Corning Wilmington DE Transwell porous cell culture inserts
`24 mm 04 um pore Nucleopore polycarbonate membrane tissue cul
`turetreated were from Costar P4C1Inulin was from ICN Ferritin was
`from Polysciences Inc Warrington PA Nitrocellulose
`filter was from
`Schuell and Immobilon was from Millipore Bedford MA
`Schleicher
`Cell CultureMicrovascular
`cells derived from rat epididymal fat
`pads RFC and bovine lungs BLMVEC along with bovine aortic en
`cells HU
`dothelial cells BAEC and human umbilical vein endothelial
`VEC were obtained grown in culture and tested periodically for en
`dothelial markers as in our past work 8 13 31 37 38
`Molecular ProbesBSA was modified with maleic anhydride Mal
`to colloidal gold particles AAu as
`BSA and by surface adsorption
`described 31 32 BSA ferritin and Mal BSA were radioiodinated
`
`used are PBS phosphate buffered saline BSA
`1 The abbreviations
`bovine serum albumin DMEM Dulbeccos modified Eagles medium
`RFC rat
`fat pad capillary BAEC bovine aortic endothelial cells
`BLMVEC bovine lung microvascular
`endothelial cells HUVEC hu
`man umbilical vein endothelial cells Mal maleic anhydride AAu
`albumin adsorption to gold particles albumingold complexes PAGE
`gel electrophoresis PS permeability surface area prod
`polyacrylamide
`uct NI nonimmune serum LDL low density lipoprotein agp60 anti
`albondin serum BON antiSPARC serum PI preimmune serum
`
`Albondinmediated Endothelial Transport of Albumin
`using IODOGEN as in Ref 8 125I AAu was created by using a 1100
`ratio of radiolabeled to unlabeled BSA as in Refs 31 and 32
`BSA Cell Binding AssaysOur past ligand binding assay 8 13 was
`using BLMVEC grown
`in sixwell
`trays
`to confluence
`performed
`Briefly after washing the cell monolayers extensively at 37 and then
`inhibitor or
`4 °C they were incubated at 4 °C first with the potential
`none in DMEM for 10 min and then with 125I BSA 0103 pgml in
`the same solution After washing with DMEM at 4 °C the monolayers
`were solubilized and the lysates counted for radioactivity using a Beck
`man 5500B ycounter
`extract 100 jig was used with
`Antiserum ProductionAlbondin
`to inoculate a New Zealand White rabbit
`complete Freunds adjuvant
`Subsequent boosting injections used incomplete Freunds adjuvant with
`from the 60kDa band of a
`50 jig of antigen that was electroeluted
`this antiserum does not
`SDSpolyacrylamide gel To be certain that
`interact with albumin which can copurify with albondin before wheat
`as assessed
`by enzyme
`germ agglutinin affinity chromatography
`linked immunosorbent assay and Western blotting we routinely run
`this antiserum over a BSAagarose column The flow through is labeled
`agp60 is used for all experiments described herein and does not rec
`ognize BSA by enzyme linked immunosorbent
`assay and Western blot
`in solution of agp60 IgG 50
`ting We also tested for possible interactions
`jigm1 with 125I BSA 02 pgm1 by gel chromatography
`using a G200
`column and found that the antibody did not cause any 125I BSA to elute
`earlier at a higher molecular weight position whereas anti BSA anti
`body did indicating a lack of agp60 interaction with BSA in solution
`of 125IBSA Binding to
`Immobilized Albondin and
`Inhibition
`125I BSA binding assays were performed on extracts of
`SPARCDirect
`albondin and SPARC immobilized on nitrocellulose
`filters gp60 extract
`02505 pg or SPARC 02505 jig isolated from BAEC medium as in
`Ref 33 or purchased from Haematologic Technologies Inc Essex Junc
`tion VT human form was blotted onto nitrocellulose and air dried The
`dot blots were blocked for 1 hat 4 °C with PBS containing 5 Blotto and
`05 Triton X100 and preincubated for 2 h with the potential
`in blocking solution Then 125I BSA was added to achieve
`a final con
`of 02 ugm1 after 10 min the filters were washed three
`centration
`times for 1 min each with PBS containing 05 Triton X100 and then
`
`6073
`
`inhibitor
`
`counted for radioactivity
`Inhibition of BLMVEC Uptake of 125I labeled Native and Modified
`BSABLMVEC monolayers were washed extensively preincubated for
`10 min at 37 °C first with DMEM in the presence of the potential
`1 mgml then 1251
`inhibitors IgGs at 100 pgml and albumins at
`labeled BSA AAu or MalBSA was added to a final concentration of 02
`pgml After 20 min the medium was removed and the cells were
`washed with DMEM at 4 °C three times for 1 min each and then
`subjected as described previously 32 to Pronase digestion to determine
`internalized by the cells
`ligand specifically
`in Rat Lung in SituThe single
`Permeability Measurements
`sample injection method 39 40 was used to assess the permeability
`surface area product PS of albumin in rat
`lung in the presence of
`inhibitors of its transport The lungs of anesthetized
`various potential
`male SpragueDawley
`rats were ventilated
`and then perfused via the
`pulmonary artery as described previously 14 Briefly the right ven
`tricle was injected with 05 ml of Ringers solution at pH = 74 111 mm
`NaC1 24 mm KC1 1 mm MgSO4 55 mm glucose 5 mm HEPES 0195
`rnm NaHCO3
`11 mm CaCl2 containing
`30 um nitroprusside and 350
`of the pulmonary artery The lungs
`units of heparin before cannulation
`were flushed at 10 mm of Hg with the following solutions i oxygenated
`30 gm nitroprusside for 3 min ii oxygen
`Ringers solution containing
`ated Ringers plus blockerstinhibitors for 15 min iii
`the solution in
`step ii plus trace 125I BSA 0310 jigm1 or 125Iferritin 1 pgm1 for 2
`min and iv the solution in step i for 2 min The lungs were excised
`and cut
`groomed
`into 14 samples Each sample was weighed and
`counted for radioactivity Based on previous work 12 39 the chosen
`perfusion times not only provide albumin enough time to cross the
`endothelium which is rapidly transcytosed after only 1530 s but also
`minimize with vascular
`flushing the presence of albumin within the cell
`the cell surface In this way the detected radioactivity will gage
`or at
`tissue clearance across the endothelial barrier and will predominantly
`125I BSA that has been transcytosed and released to the in
`represent
`terstitium Also note that the flush time is too short to allow any loss of
`125I BSA to lymphatic drainage and that the perfusion pressures and
`flow rates were constant with antibody and albumin perfusions indi
`if any variation in the vascular surface area
`cating little
`The PS was calculated as experimentally and theoretically defined in
`39 using
`
`PS = Alt x C
`
`

`

`6074
`
`Albondinmediated Endothelial Transport of Albumin
`
`kDa
`
`180
`
`116
`
`58
`48
`36
`26
`
`14Z
`
`111
`
`to 0
`
`Z
`be 0
`Z Z5
`
`Xi
`
`gp6 0
`
`gp6 0
`
`kna
`180
`
`116
`84
`58
`48
`
`36
`26
`
`2
`3
`4
`5
`1
`6
`FR 1 Purification of albondin SDSPAGE analysis of fraction
`ation steps used to purify albondin silver stain of equivalent protein
`loads Lane 1 total cell lysate Lanes 26 proteins of various fractions
`of purification procedure RFC proteins were extracted using Triton
`X100 lane 2 followed by ethanol precipitation as described in Ref 8
`The precipitate was resuspended in PBS mixed for 1 h and then cen
`trifuged at 4 C The proteins in the pellet and supernatant
`are shown
`in lanes 3 and 4 respectively After the addition of detergent
`to a final
`concentration of 1 Triton X100 and 02 SDS the supernatant was
`subjected to wheat germ agglutinin affinity chromatography
`followed
`by elution with 05 m Nacetylglucosamine The eluant was dialyzed
`against 01 a PBS containing 01 Triton X100 and 002 SDS and
`then concentrated 3 fold by SpeedVac SDSPAGE analysis of this last
`fraction revealed only one band at 60 kDa by silver staining lane 5 and
`after radioiodination lane 6 as in Refs 8 and 14
`
`using
`
`lung t = perfusion time with 125I BSA =2 min
`where A = cpmg of wet
`and Cp= cpmml of 25I BSA perfusate Because recent work 25 sug
`the nor
`gested that the absence of Ca2
`from perfusates may prevent
`found with removal of
`mally large increase in hydraulic conductivities
`albumin from perforates we also performed many experiments
`Ringers solution without Ca2 We did not observe any significant dif
`ferences in BSA transport in the presence or absence of Ca2 Therefore
`all experiments were combined
`In Vitro Thansport AssayBLMVEC 12000 cellswell were seeded
`filters 033 cm2After 67 days transmonolayer
`trans
`onto rranswell
`part assays were performed similar to previous work 41 This filter
`system separates a well of a 24 well plate into a top and bottom cham
`ber by a polycarbonate filter 04pm diameter pores covered by con
`fluent BLMVEC Volumes of 100 pl for the top chamber and 600 pl for
`the bottom chamber were used to create an equal
`fluid line in each
`chamber and minimize hydrostatic
`effects The cells were washed at
`37 C with Ringers solution twice for 1 min each and then Ringers
`solution containing 05 ingen1 bovine orosomucoid once for 10 min once
`for 1 min Ringers solution containing orosomucoid was added above to
`the bottom chamber and with the desired potential
`inhibitor to the top
`chamber After 15 min the fluid in the top chamber was replaced with
`orosomucoid containing 25155A 02 pg
`Ringers solution containing
`ml and the same potential
`inhibitor To compare transport across the
`monolayers we removed 30 al of fluid from the bottom after 30 min The
`samples were counted for y radioactivity
`
`initial
`
`RESULTS
`Purification of AlbondinWe have purified the 60kDa en
`dothelial protein gp60 now called albondin for which our
`suggested a role in albumin binding 14 30
`evidence
`extracted from RFC cell monolayers us
`Albondin was initially
`ing Triton X100 followed by ethanol precipitation as described
`in Ref 14 The precipitate was resuspended in PBS containing
`1 Triton X100 and 02 SDS and was subjected to wheat
`
`germ agglutinin affinity chromatography
`followed by elution
`with 05 is Nacetylglucosamine Fig 1 shows that SDSPAGE
`analysis of this eluted fraction revealed only one band at 60
`kDa by silver staining also true for Coomassie staining data
`not shown and after radioiodination of this fraction Densitom
`etry revealed that greater than 98 of the detected signal was
`
`HUVEC
`BIMVEC
`RFC
`Fa 2 Antisera recognition of albondin in endothelial
`cell ly
`sates by inununoblotting As described previously 32 42 the pro
`teins of cultured monolayers from BLMVEC RFC and HUVEC were
`solubilized and analyzed by SDSPAGE followed by electrotransfer
`to
`Immobilon filters
`immunoblotting with the indicated sera As a
`positive control and in agreement with our past work 37 antiserum
`BON made against
`the secreted 43kna albumin binding protein
`SPARC also recognizes gp60
`
`for
`
`present
`
`in the single band Because our past experiments
`showed that antiserum BON made against
`the secreted 43
`kDa albuminbinding protein SPARC recognized albondin 37
`we used BON in slot blots and Western blots to assay each of
`our fractions for the presence of albondin We also used 125I
`BSA on slot blots for following albumin binding activity Our
`average yield in the final fraction for three different isolations
`was 47 with a 710 fold degree of purification To be certain
`that the purified 60kDa protein is indeed responsible for the
`albumin binding activity we incubated an aliquot of the final
`fraction with BSA conjugated to agarose beads and found that
`these beads extracted almost all of the 60kDa protein from the
`solution whereas equivalent
`IgG beads did not
`Specific Recognition of Albondin by AntiserumPolyclonal
`rabbit antiserum agp60 was made against albondin extract
`recognizes a 60kDa pro
`Fig 2 shows that agp60 specifically
`lysates by im
`tein in human bovine and rat endothelial cell
`munoblotting Fig 3 shows that agp60 also immtmoprecipi
`tated a radiolabeled 60kDa protein from lysates of endothelial
`cells whose surface proteins have been radioiodinated Nonim
`mune serum NI was nonreactive whereas BON recognized
`albondin as in our past work 37 Both BON and agp60 recog
`nized purified albondin by immunoblotting data not shown
`Please note that i the 43kDa protein SPARC was not de
`lysates by BON or agp60 consist
`tected in the endothelial cell
`ent with its expression only during cell migration and growth
`and cessation of expression with cell monolayer confluence 33
`37 ii SPARC was not present on the cell surface as assessed
`by radioiodination followed by inununoprecipitation light band
`near 46 kDa in Fig 3 is present
`three lanes and iii
`in all
`neither BON nor agp60 recognized the smaller putative albu
`minbinding proteins gp30 and gp18 whose presence on the
`filters was verified by blotting with A Au as in Ref 31
`agp60 Recognizes SPARCSince BON interacts with albon
`din we tested the ability of agp60 to recognize SPARC Fig 4
`shows that agp60 and BON immunoprecipitated SPARC that
`was biosynthetically radiolaheled in vitro whereas NI did not
`BON and agp60 also can immunoblot SPARC data not shown
`these two albuminbinding proteins contain
`It appears that
`common immunological epitopes
`126IBSA Binding
`Antibodies Recognizing Albondin Inhibit
`to EndotheliumNext
`the IgG fractions from various antisera
`the binding of 125I BSA to the
`were used to attempt to inhibit
`
`

`

`Albondinmediated Endothelial Transport of Albumin
`
`6076
`
`NI
`
`BON ogp60
`Etc 3 Immunoprecipitation of surface radiolabeled albondin
`from BLMVEC lysates Confluent BLMVEC monolayers were radio
`iodinated using lactoperoxidase Na251 and hydrogen peroxide and
`then processed
`for immunoprecipitation using the indicated antisera
`followed by SDSPAGE and autoradiography
`as described previously
`48 Similar results were found using RFC monolayers
`
`84
`
`48
`
`26
`
`SPARC
`
`BON agp60
`NI
`Flo 4 Recognition of SPARC by ogp60 SPARC was biosyntheti
`cally radiolabeled with 135Simethionine and imrnunoprecipitated with
`the indicated antisera from the culture medium of BAEC as in Ref 37
`
`surface of confluent monolayers of cultured BLMVEC2 Fig 5
`shows that specific 12N BSA binding can be ablated almost
`totally by agp60 IgG and unlabeled BSA in a concentration
`dependent manner The maximal
`inhibitory effect of agp60
`
`90 was comparable to the inhibition attained with a large
`molar excess of BSA Fig 6 shows that BON IgG inhibited this
`binding by greater than 80 whereas other antibodies were
`ineffective The effects of 2050 ugm1 of BON and agp60 IgG
`were comparable to the inhibition achieved with 1 mgm1 BSA
`It appears that specific antibody interaction with albondin sig
`nificantly reduces cell surface binding of albumin
`Native but Not Modified Albumins Inhibiti251BSA Binding
`to BLMVECEndothelium can bind not only native albumins
`5 1114 34 37 but also albumins modified by surface adsorp
`tion to colloidal gold particles AAu or by chemical means
`such as treatment with maleic anhydride 9 10 31 32 42 43
`
`inhibit
`
`To assess if native and modified albumins interact with the
`same binding sites we tested the ability of these albumins to
`1215I BSA binding to the BLMVEC surface As shown in
`Fig 6 unlabeled
`native BSA inhibited 1251 BSA binding
`whereas AAu and MalBSA did not even at more than 100 fold
`levels known to saturate gp30 and gp18
`molar excess and at
`
`2 Most of our past work has focused on the RFC cell monolayers 13
`143234 however we found that the RFC and HINEC monolayers do
`not form restrictive barriers on filters whereas the BLMVEC do Fur
`thermore these BLMVEC monolayers show less signs of phenotypic
`in culture and have a greater density of noncoated
`plasmalemma
`drift
`cells our unpublished ob
`vesicles than the other cultured endothelial
`servations
`
`Protein pgm1
`Flo 5 Concentration dependence of inhibitors of 25IBSAsur
`face binding to BLMVEC monolayers in vitro The indicated pro
`teins unlabeled BSA agp60 IgG or NI IgG were tested at different
`for their ability to prevent specific 251 BSA binding to
`concentrations
`confluent BLMVEC monolayers at 4 C In each experiment duplicate
`wells were assayed in the presence of a large excess of unlabeled BSA
`1070 mgm1 to assess the nonspecific background radioactivity This
`background was subtracted from all other data points as in Ref l and
`comprised less than 15 of the total The final results shown in the
`inhibitor and are ex
`graph were normalized to the controls without
`pressed as a percentage of the control The mean is given for 2
`The SD were less than 25 The asterisk indicates the detected bind
`ing in the presence of 1 mgml BSA
`
`140
`
`An
`
`Iv
`
`20
`
`inhibition of 26IBSA binding to the endo
`FIG 6 Differential
`thelial cell surface The IgG fraction of the indicated sera were used
`of 2050 pgml
`see Experimental Procedures at a concentration
`whereas BSA and Mal BSA were at 1 mgml The albumin gold com
`plexes were used at a final concentration of 2040 pgmi of BSA conju
`gated to gold which saturates binding to gp30 and gp18 31 32
`Nonspecific background was assessed and subtracted as described in
`the legend of Fig 5 The results are expressed as the percentage of the
`4 The ogp
`filters no inhibitor The mean ± SD is plotted
`control
`serum has been described previously 414 32
`
`binding 3132 Endothelial cells kinetically have two distinct
`populations of binding sites one for native albumin and an
`for modified albumins
`other
`Inhibition of Specific 251 BSA Binding to Immobilized Al
`bondin and SPARC125IBSA binding assays were also per
`formed on albondin and SPARC immobilized on nitrocellulose
`filters see Experimental Procedures in the presence or ab
`sence of various antibodies and albumins As shown in Fig 7
`the IgG fraction of BON and agp60 but not other antisera
`
`

`

`Antibodies
`
`Albumins
`
`cr
`
`0p
`
`a
`
`I
`
`LI
`
`Antibodies
`
`Albumins
`
`6076
`
`A
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`120
`
`100
`
`t
`
`CJ
`0
`
`2 0
`
`5
`
`09
`
`0 0 0
`
`2 6
`
`5
`
`rA
`<a0
`
`t
`
`0
`
`0`1
`
`71
`
`z0
`
`=
`
`2
`
`FIG 7 Inhibition of 125IBSA binding directly to albumin bind
`ing proteins on filters 125I BSA binding assays were performed on
`extracts of gp60 panel A and SPARC panel B immobilized on nitro
`cellulose filters see Experimental Procedures in the presence or ab
`are the same
`inhibitor concentrations
`sence of the indicated potential
`as in Fig 6 In each experiment nonspecific background binding was
`assessed by performing the assay in the presence of excess unlabeled
`BSA 5070 mgml and was subtracted from the counts of
`the other
`filters It was less than 25 and 35 of the total counts on the control
`filters for albondin and SPARC respectively The results are expressed
`filters no inhibitor The mean ± SD
`as the percentage of the control
`is plotted n 4 The results using the human and bovine SPARC were
`very similar and therefore were combined Our past work 31 showed
`poor binding of 125I BSA to gp30 and gp18 by blotting
`
`significantly inhibited binding to both albondin and SPARC
`This inhibitory effect was significantly diminished when these
`IgGs were preincubated with i thin filter strips containing
`immobilized albondin or SPARC processed
`for preparative
`SDSPAGE and then electrotransferred
`onto Immobilon filters
`immobilized and soluble purified SPARC or gp60 extract
`or ii
`but not i immobilized or soluble ovalbumin or ii
`filter strips
`of a nonspecific region of a preparative SDS gel of BLMVEC
`lysate transferred onto Immobilon filters data not shown Un
`labeled native BSA significantly inhibited 125I BSA binding
`whereas Mal BSA did not These results correlate well with our
`findings on native BSA binding to the cell surface and provide
`
`Albondinmediated Endothelial Transport of Albumin
`
`BLMVEC
`
`gp3018 blots
`LT
`
`V
`
`71
`
`N
`
``=
`
`t
`
`<
`
`PC
`
`cr
`
`00
`
`1
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`control
`
`of
`
`Binding
`
`ModifiedB
`
`SA
`
`FIG 8 Inhibition of modified albumin binding to the surface of
`BLMVEC and gp3018 on blots Panel A binding inhibition assays
`see Experimental Procedures were performed on BLMVEC using
`125I AAu 2 pgml or 125I Mal BSA 02 pgml with the indicated mol
`ecules 50 pgml
`for IgGs and 1 mgml for albumins Nonspecific back
`ground detected in the presence of 1 mgml unlabeled Mal BSA was less
`than 30 of the total without
`inhibitor and was subtracted from the
`other monlayers Panel B inhibition studies were done using blots of
`gp30 and gp18 as described previously 34 The results for gp30 and
`gp18 were very similarand therefore were combined The results rep
`the total detected binding to gp30 and gp18 expressed as the
`resent
`mean percentage ± SD of the control no inhibitor for n
`3
`
`that albondin binds native albumin possibly via a
`evidence
`common binding domain shared with SPARC
`Inhibition of 125I labeled Modified BSA Binding to
`Selective
`BLMVEC and gp30 1 18We also examined the ability of these
`the binding of 125I labeled modified albumins
`probes to prevent
`to BLMVEC and to gp30 and gpl 8 on blots Fig 8 shows that
`native BSA and the antibodies were quite ineffective inhibitors
`whereas MalBSA inhibited both 125I AAu and 125I Mal BSA
`binding very well Cumulatively it appears that different pro
`teins mediate the binding of native and modified albumins to
`the cell surface with gp30 and gp 18 recognizing the modified
`albumins and albondin interacting with native albumin
`Inhibition of BLMVEC Internalization of Native and Modi
`fied AlbuminsHaving dissected the cell surface binding of
`these ligands we proceeded
`to focus next on their internaliza
`tion by the cells Our recent work 31 32 showed that native
`and modified albumins are all
`internalized but are processed
`differently with only modified albumins being degraded
`Herein we attempted to inhibit
`the cellular uptake of these
`ligands by using the same group of potential
`inhibitors tested
`in the binding assays Fig 9 shows that native BSA and the IgG
`fraction of BON and agp60 inhibited specific BLMVEC uptake
`of 125I BSA but not 125I labeled modified albumins Conversely
`Mal BSA prevented both 125I Au and 125I Mal BSA internaliza
`tion but not 125I BSA uptake NI IgG did not inhibit any ligand
`uptake These results show the same inhibition profile as the
`cell surface binding studies and indicate that native and modi
`fied albumins are internalized by different molecular mecha
`nisms Albondin mediates
`native albumin internalization
`whereas gp30 and gp 18 may be responsible for modified albu
`min uptake
`Inhibition of Capillary Permeability to Albumin in
`Selective
`Rat Lungs in SituAlthough the experiments described above
`that albondin is the major endothelial plas
`provide evidence
`malemma protein mediating the cell surface binding of native
`albumin its direct
`role in the transendothelial
`transport of
`albumin is not established Therefore we also investigated the
`
`

`

`Albondinmediated Endothelial Thansport of Albumin
`
`6077
`
`10
`
`301
`
`§
`
`<
`
`119
`
`°
`
`0
`
`50
`
`40
`
`30
`
`20
`
`10
`
`Total Conc Perhised = 40 mgm1
`Ficoll = 40 BSA
`
`10
`0000
`
`001
`
`01
`
`1
`
`1
`
`10
`
`100
`
`Unlabeled BSA mgml
`of capillary permeability
`FIG 10 Concentration
`dependence
`of i25IBSA in the rat lung in situ Unlabeled BSA was used at
`the
`to inhibit 125I BSA transcapillary transport in
`indicated concentrations
`the rat lung To minimize oncotic
`effects Ficoll was added to the per
`fusate to maintain a total concentration of 40 mgml in each perfusion
`
`120
`
`100
`
`80
`
`60
`
`I1
`
`cs
`
`N R
`
`S
`
`0
`0 0
`
`F
`
`411
`
`14 o
`
`PO
`
`20
`
`125I BSA 125I AAu 125I Mal BSA
`Fic 9 Inhibition of specific BLMVEC internalization of native
`and modified albumins BLMVEC uptake of 25I labeled BSA 02
`pgm1 AAu 10 pgm1 or Mal BSA 02 pgm1 was assessed at 37 °C
`using a Pronase digestion assay see Experimental Procedures The
`indicated proteins were used to attempt to inhibit cellular uptake Non
`specific uptake was assessed in the presence of excess unlabeled ligand
`10 mgml BSA for 125I BSA or Mal BSA for both 1251 AAu and 125I
`Mal BSA and was subtracted from corresponding paired samples It
`comprised 1525 of the total radioactivity detected