HDL and apolipoprotein AI protect erythrocytes against the generation of procoagulant activity

1775

HDL and Apolipoprotein A-I ProtectErythrocytes Against the Generation

of Procoagulant Activity

Richard M. Epand, Alan Stafford, Bryan Leon, Philippa E. Lock, Ewan M. Tytler,Jere P. Segrest, G.M. Anantharamaiah

Abstract The appearance of anionic lipids on the extracel-lular surface of cells is required for the formation of theprocoagulant complex that leads to the activation of prothrom-bin. Procoagulant activity would be expected to be inhibited bysubstances that stabilize the membrane structure and henceinhibit the transbilayer diffusion of phosphatidylserine fromthe cytoplasmic to the extracellular surface of the plasmamembrane. The generation of procoagulant activity in humanerythrocytes by A23187 and Ca2+ is inhibited by apolipopro-tein A-I, its amphipathic peptide analogues, and high-densitylipoprotein (HDL). These agents do not inhibit the Ca2+

loading of erythrocytes by A23187, nor do they inhibit the

A mphipathic helixes have long been known to play/ \ an important role in the interaction of peptides

J. \ . and proteins with membranes.1 Recently, thesehelixes have been classified on the basis of the size andcharge distribution of their hydrophilic domain.2 Ingeneral, this classification also segregates peptides andproteins according to their function. Two of theseclasses are the apolipoprotein helixes (class A) and thelytic helical peptides (class L). We have shown thatthese two classes of helixes have opposite effects onseveral membrane properties, including their effects onthe bilayer-to-hexagonal phase transition temperature,leakage of liposomes, and hemolysis.3 We also demon-strated that class A helixes can inhibit leakage fromliposomes or hemolysis of human erythrocytes causedby class L helixes. This suggested that proteins contain-ing class A helixes, such as the plasma apolipoproteinA-I (apoA-I), can have a protective effect against mem-brane damage.

One of the important functions of blood is thrombo-genesis. The final step in the coagulation cascade is theconversion of prothrombin to thrombin in the presenceof factor Xa, factor Va, and Ca2+. This process occurson the surface of membranes containing anionic phos-pholipids. Such lipids are normally found on the cyto-

Received July 9, 1994; revision accepted August 23, 1994.From the Department of Biochemistry, McMaster University

Health Sciences Centre, Hamilton, Ontario, Canada (R.M.E.,A.S., B.L., P.E.L.), and the Departments of Medicine and Bio-chemistry and the Atherosclerosis Research Unit, University ofAlabama Medical Center, Birmingham, Ala (E.M.T., J.P.S.,G.M.A.).

Reprint requests to Richard M. Epand, Department of Bio-chemistry, McMaster University Health Sciences Centre, Hamil-ton, Ontario, L8N 3Z5, Canada.

O 1994 American Heart Association, Inc.

activation of prothrombin once the cells have been incubatedat 37°C with A23187 and Ca2+. Transbilayer diffusion offluorescently labeled phosphatidylserine is inhibited by apo-lipoprotein A-I. These findings indicate that class A amphi-pathic helixes as well as lipoprotein particles and liposomesinhibit the transbilayer diffusion of phospholipids and pro-coagulant activity. This activity may contribute to the pro-tective role of HDL against arteriosclerosis and thrombosis.(Arterioscler Thromb. 1994;14:1775-1783.)

Key Words • procoagulant activity • transbilayerdiffusion • apolipoprotein A-I • HDL • erythrocytes

plasmic surface of the plasma membrane, to which theblood coagulation proteins normally do not have access.However, as a result of cell lysis or the transbilayerdiffusion of anionic lipids, cell membranes can initiatethe activation of prothrombin. This process has beentermed the generation of procoagulant activity. It hasbeen demonstrated that gramicidin promotes transbi-layer diffusion of phospholipids in erythrocyte mem-branes only under conditions in which it would beexpected to promote Hn phase formation/ As a conse-quence, agents that inhibit H,, formation may inhibittransbilayer diffusion. This could have physiologicalconsequences by inhibiting the transbilayer diffusionrequired for procoagulant activity. The level of high-density lipoprotein (HDL) is known to be inverselycorrelated with the incidence of arteriosclerosis.56 HDLhas long been regarded as important for "reverse"transport of cholesterol from peripheral tissue to theliver.57 We suggest that this lipoprotein may also func-tion to stabilize membranes. We further suggest that themechanism of this stabilization may, at least in part, bethrough an interaction of the class A amphipathichelical segments of apoA-I with cell membranes.

If class A helixes stabilize membranes, then class Lhelixes would be expected to cause membrane defectsand promote procoagulant activity. However, we ob-served that several agents that promote Hn formation,including the peptides 18L and mastoparan as well asoctanol, decanol, and dodecanol, caused considerablehemolysis at concentrations required to promote proco-agulant activity. Thus, it is difficult to determine theextent of transbilayer diffusion promoted by theseagents in intact erythrocytes. It is known that Ca2+

loading of erythrocytes leads to a loss of phospholipidtransbilayer asymmetry.8 It has recently been found that

by guest on June 27, 2016http://atvb.ahajournals.org/Downloaded from

http://atvb.ahajournals.org/

1776 Arteriosclerosis and Thrombosis Vol 14, No 11 November 1994

phosphatidylinositol 4,5 -bisphosphate plays an impor-tant role in this process.9 The loss of asymmetry issustained because the higher intracellular levels of Ca2+

inhibit translocase activity.10 In addition, Ca2+ influxleads to the formation of microvesicles, but this is not amechanism for the loss of transmembrane asymmetry.11

Erythrocytes attain an enhanced permeability to Ca2+

when subjected to physiological shear stresses.12 If thisCa2+ influx caused the transbilayer diffusion of phos-pholipids, it could be of pathological significance be-cause it would promote procoagulant activity leading tothrombus formation. In this article, we demonstrate thatprotection from this nonregulated thrombogenesis isafforded by the presence of HDL.

MethodsMaterials

HDL was isolated from freshly drawn human blood bydensity gradient centrifugation. ApoA-I was then extractedand purified by reverse-phase high-performance liquid chro-matography (HPLC) as previously described.13 Synthetic pep-tides were made by solid-phase synthesis using /V-f-butoxycar-bonyl chemistry. Peptides were cleaved from the resin usinganhydrous HF and purified by reverse-phase HPLC.14 Bovineblood coagulation factor Va, factor Xa, and prothrombin werepurchased from Haematologic Technologies. /V-p-Tosyl-Gly-Pro-Arg-p-nitroanilide was purchased from the Sigma Chem-ical Co. A23187 was from Calbiochem. Bovine serum albumin(BSA) was purchased as fatty acid-free fraction V fromBoehringer Mannheim. l-Palmitoyl-2-[6-[(7-nitro-2,l,3-ben-zoxadiazol-4-yl)amino]caproyl]-in-glycero-3-phosphoserine(NBD-PS) was purchased from Avanti Polar Lipids.

Procoagulant Activity and HemolysisHuman blood was freshly drawn into tubes coated with

acid-citrate-dextrose or heparin. The erythrocytes were kept at0°C to 4°C up to the incubation step at 37°C. Blood (2 to 3 mL)was centrifuged, and the cell pellet was resuspended in 35 to40 mL of 5 mmol/L HEPES, 150 mmol/L Nad, 1 mmol/LMgSO4, 5 mmol/L K d , pH 7.4 (HEPES buffered saline[HBS]) containing 0.1 mmol/L ZnCl2. The ZnCl2 was added toprotect the erythrocytes against hemolysis.13 The cells wereagain pelleted at 4500 rpm in an SS34 rotor of a Sorvallcentrifuge at 4°C. The pellet was resuspended in the HBS withZn2+ and centrifuged twice more. A 125-/tL aliquot of thewashed, packed erythrocytes was added to 25 mL of HBScontaining 0.1 mmol/L ZnCl2 and 1.0 mmol/L Caa2 . A 950-/iLaliquot of this erythrocyte suspension was added to A23187and/or various peptides, proteins, or lipoproteins in an Eppen-dorf centrifuge tube as indicated. The final volume wasadjusted to 1.00 mL, and the solution wasvortexed. The orderof addition of the agents did not affect the final extent ofhemolysis or the generation of procoagulant activity. Themixture with the erythrocytes was then warmed to 37°C andincubated for 20 minutes with shaking. After this period, thetube was gently vortexed, and a 20-^.L aliquot of the suspen-sion was diluted with 455 fiL HBS. A 20-/iL aliquot of thismixture was used to measure procoagulant activity (see be-low). The remaining 980 /iL of erythrocytes was then spun ina microfuge for 90 seconds, and the absorbance of the super-natant at 540 nm was measured in a Perkin-Elmer spectropho-tometer. The absorbance was compared with that from anequivalent-size aliquot of erythrocytes that had been sonicatedto completely liberate all of the hemoglobin. The sonicatedsamples had an absorbance at 540 nm in the range of 1.1 to 1.3.The percentage of hemolysis was always 3±2% for the resultspresented in this article.

The 20 -/ML sample taken for the assay of procoagulantactivity was diluted with 60 yl. HBS. To the 80 /iL of this

erythrocyte suspension were added 10 /xL factor Xa solution,10 pL factor Va solution, and 10 /iL CaCl2. The tube was thenwarmed for 2 minutes at 37°C, and 10 fiL prothrombin wasadded to commence the reaction. The final concentration offactor Xa was 0.2 U/mL, which was stored in the freezer with1 mg/mL BSA in 40 mmol/L HEPES, 124 mmol/L NaCl, pH7.4, diluted 1:1 with glycerol. The final concentration of factorVa was 6 nmol/L from a stock stored as factor Xa but with theaddition of 2 mmol/L CaCl2. The final Ca2+ concentration was4 mmol/L. The prothrombin concentration was 2 jimol/L froma stock stored as above, without calcium. After addition ofprothrombin, the mixture was incubated for 5 minutes at 37°C.Any thrombin generated as a consequence of the appearanceof anionic lipids on the extracellular surface was then assayedspectrophotometrically using the synthetic substrate N-p-tosy\-Gry-Pro-Arg-p-nitroanilide. Two 40-/*L aliquots of the incu-bation mixture were removed. One aliquot was added to areference cuvette containing the phosphate buffer composedof 100 mmol/L sodium phosphate, 1 mmol/L EDTA, and 1mg/mL polyethylene glycol 6000. The polyethylene glycol isadded to avoid adsorption of thrombin to surfaces. The other40-jiL aliquot was added to a sample cuvette containing 1mmol/L Ar-p-tosyl-Gly-Pro-Arg-p-nitroanilide in the samephosphate buffer at 37°C.

The rate of hydrolysis of this peptide was monitored by therate of change of absorbance at 405 nm. The rate of reactionwas always compared with that given by sonicated erythrocytesthat had been incubated with prothrombin and factors Va andXa as for the test samples. The value for 100% procoagulantactivity was calculated by multiplying the value given by thesonicated erythrocytes by 0.75 to correct for the larger area ofthe external monolayer of small vesicles. The corrected valuewas similar to that obtained with hypotonicaUy rysed erythro-cytes but was more reproducible. The value for 100% proco-agulant activity was converted into the concentration of throm-bin generated by comparison with the rate of hydrolysis of thesynthetic substrate by thrombin standards. Blanks in theabsence of cells generated little procoagulant activity. The Km

for thrombin-catalyzed hydrolysis of N-p-tosyl-Gly-Pro-Arg-/?-nitroanilide is about 5 /xmol/L.16 Using 1 mmol/L substrate,we measured the constant rate of reaction at maximumvelocity over several minutes.

A23187 Promoted Calcium Uptake byHuman Erythrocytes

An experiment was performed to test if apoA-I affected theextent of Ca2+ accumulation in erythrocytes. This was done byincubating human erythrocytes in the presence of A23187 and43Ca2+. Cells were separated from the extracellular environ-ment by centrifugation through oil. ApoA-I, which couldhypothetically extract A23187 from the membrane, was alsoadded to test its effect on Ca2+ accumulation in erythrocytes.

For these experiments, 1 mL of blood, drawn from a healthydonor, was collected in an EDTA-treated tube. The wholeblood was mixed with 10 mL HBS and centrifuged at 3000 rpmfor 10 minutes at 4°C in a Sorvall HG-4L rotor. The serum andbuffy coat were discarded, and the erythrocytes were washedtwice more by resuspension in 10 mL HBS and recentrifuga-tion. The washed erythrocytes were resuspended in 4 mL HBSto give a hematocrit level of 10%. Calcium accumulation wasmeasured at 2% hematocrit in the presence of 100 nCi/mL of<5CaCl2 (New England Nuclear) and 1 mmol/L cold CaQ2 in atotal volume of 3.5 mL HBS, in the presence or absence of 10Atmol/L A23187, and in the presence or absence of humanapoA-Ior HDL. The higher concentration of A23187 was usedbecause the concentration of erythrocytes was higher than thatin the experiments to measure procoagulant activity. Thisresulted in the concentration of A23187 in the erythrocytemembrane being similar for the two experiments. Use of 1junoI/L A23187 gave much lower uptake of Ca2+. The mixturewas incubated at 37°C for 20 minutes in a water bath with



shaking. The reaction was terminated by centrifugation oftriplicate 1-mL samples through 100-/xL aliquots of oil (dibu-tyl phthalate:dinonyl phthalate 5:1) at 16 OOOg for 3 minutes atroom temperature.17 Aliquots (500 jiL) of each supernatantwere taken for scintillation counting. The remaining superna-tant and the oil were removed and discarded. 43Ca incorpo-rated into each erythrocyte pellet was extracted with 600 /iL of5% tricholoroacetic acid (TCA).18 The pellets were vortexedvigorously, and TCA extraction was carried out overnight at4°C. TCA-insoluble material was removed by centrifugation at16 OOOg for 3 minutes. Aliquots (500 fiL) of the TCA super-natants were taken for scintillation counting. Scintillation fluid(5 mL) was added to each vial, and 45Ca levels were deter-mined with a Beckman LS5000TD scintillation counter. Abuffer blank and triplicate standards, containing 100 nCi of45Ca, were included to allow Ca2+ accumulation to be calcu-lated in disintegrations per minute from the raw data oncounts per minute.

MicrovesiculationAlthough microvesiculation of erythrocytes is not thought to

contribute to phospholipid reorientation induced by Ca2+ ,8 wenevertheless measured the effect of apoA-I on this process.Erythrocytes treated with Ca2+ and A23187 shed microvesi-cles, which remain in the supernatant after centrifugation ofthe erythrocytes.19 Erythrocytes were washed, processed, andincubated for 20 minutes at 37°C as described in "Procoagu-lant Activity and Hemolysis," after which time the cells werecentrifuged for 5 minutes at 500g. After centrifugation, 600 fiLof the supernatant was diluted to 1.2 mL of pH 8.0 phosphatebuffer, and the solution was assayed for acetylcholinesteraseactivity30 as a measure of microvesicle formation.21 Replace-ment of NaCl with KC1 in the buffer suppresses microvesicu-lation." This was confirmed in the present study.

Transbilayer Diffusion of PhospholipidsWe used a fluorescent assay similar to that developed by

Connor and Schroit22 and that was recently applied to calcium-promoted transbilayer diffusion of phospholipids.8 Briefly,freshly prepared human erythrocytes were washed three timesin HBS by centrifugation. NBD-PS was dissolved in chloro-form and methanol (2:1), and 75 fig was deposited as a filmafter evaporation of the solvent under a stream of nitrogen anddrying in a vacuum desiccator for 2 hours. The film wassuspended by vigorous vortexing with 1.5 mL HBS, and 0.5 mLof packed erythrocytes was added to this suspension. Thiscorresponds to a ratio of approximately 1:50 of NBD-PS toerythrocyte phospholipid. The cells were then incubated at37°C for 1 hour to allow internalization of the NBD-PS. Afterthis incubation, the cells were placed on ice for 5 minutes.NBD-PS that was not internalized was removed by washing thecells nine times with ice-cold 2% fatty acid-free BSA in HBS,followed by three additional washes with ice-cold HBS alone.After the final wash, the volume of the suspension wasadjusted to 2.0 mL. The rate of exposure of NBD-PS bytransbilayer diffusion was measured by incubating 40 fih of thelabeled erythrocyte suspension in a total volume of 100 /xL inthe presence or absence of apoA-I and/or 0.5 fimolfL A23187with 1 mmol/L Cad2 for 20 minutes at 37°C. Control experi-ments demonstrated that apoA-I at the highest concentrationused (0.2 mg/mL) did not extract any NBD-PS from themembrane. At the end of the incubation period, 70 \iY. of thecell suspension was removed to prechilled Eppendorf tubescontaining 10 fiL of 120 mg/mL BSA in HBS. The tubes wereincubated on ice for 3 minutes to allow for the extraction ofany NBD-PS that had undergone transbilayer diffusion. Thetubes were then centrifuged for 2 minutes in an Eppendorfmicrofuge. Forty microliters of the supernatant was dilutedwith 2.0 mL of 2% Triton in HBS. Total cell-incorporatedNBD-PS was also determined using an aliquot from the cellsuspension in BSA before centrifugation. The relative concen-

Epand et al HDL Inhibits Transbilayer Diffusion 1777

% Procoagulant Activity

0 5 10 15 20 25 30

(ETOH + Cl1*)

A231S7

A231*7 + Of*+50 nlwA «po M.

A231S7 + Of*+25 pf/ml ipo A-I

A231S7 + Of*+12 «/ml qx> A4

SO n/ml tpo A-l

1

\?

\Fto1. Bar graph shows inhibition of Ca2+-induced procoagulantactivity by apolipoproteln A-I (apoA-I). Erythrocytes were Incu-bated for 20 minutes at 37°C with or without 1.0 /irnol/L A23187and/or 1.0 mmol/L CaClj and/or apoA-I. Procoagulant activity of100% corresponds to 7.0 nmol/L thrombin added to the finalspectrophotometrlc assay; mean of 3 determinattons±SD. ETOHIndicates ethyl alcohol.

trations of NBD-PS were determined using an SLM-AmincoSeries II spectrofluorimeter with the excitation wavelength setat 460 nm and the emission monochrometer at 534 nm.

Reconstitution of HDLApoA-I was codissolved in cholate micelles with or without

cholesterol and l-palmitoyI-2-oleoyl phosphatidylcholine(POPC) in various proportions. Equal molar amounts ofPOPC and cholate were used. In the case of dialyzed apoA-I,the molar ratio of cholate to protein was 80. The cholate wasremoved by dialysis to produce both vesicular and discoidalreconstituted HDL (rHDL) as previously described.23-24 Con-trols were made also by dialyzing cholate solutions of apoA-Ialone or of lipid alone.

Protein ConcentrationProtein concentration was determined by the method of

Lowry et al.25

Phospholipid ConcentrationPhospholipid concentration was determined by phosphate

analysis after the sample was digested using the method ofAmes.26

ResultsIncubation of erythrocytes at 37°C in the presence of

1.0 Mmol/L A23187 and 1.0 mmol/L CaCl2 causes anincrease in procoagulant activity that is linear with timeover 20 minutes. The generation of this procoagulantactivity is strongly inhibited by apoA-I (Fig 1). ApoA-Ihas no effect in the absence of Ca2+ and A23187.




(

Control(BTOH + C**)

A231S7 + CtP*

A23187 + Ci**+2C0«talJ7pA

A23U7 + OJ?*+ 100«Ai)l J7pA

AJ3U7 + C»**

A231I7 + C***

1


10 20 30 40

1

H

HH


5 10 15 20 25 30

FIG 2. Bar graph shows inhibition of CaI+-induced procoagulantactivity by the model peptide 37pA. Other conditions and abbrevi-ations are as for Fig 1. Procoagulant activity of 100% correspondsto 12.2 nmd/L thrombin; mean of 2 determinations± range.

Neither Ca2+ alone nor A23187 alone are potent pro-moters of procoagulant activity (Fig 1). The high degreeof inhibition of procoagulant activity is not shown byother peptides and proteins with class A helixes. Theapolipoprotein A-II (apoA-II) at a concentration of 60/u.g/mL (about twice the molar concentration of apoA-I)exhibits only 10% inhibition of procoagulant activity(results not presented in the figures). The peptides 18Aand Me-18A, which have some, but weaker, effects inraising the bilayer-to-hexagonal phase transition tem-perature,3 also show only 10% or less inhibition ofprocoagulant activity up to a peptide concentration of50 ixglmL (20 /xmol/L). (Also see Reference 3 for thestructure of these peptides.) The former peptides con-tain only a single a-helical segment. ApoA-I containsseveral helical segments separated by a few amino acidresidues, usually including Pro. A synthetic model am-phipathic helical peptide that more closely resemblesapoA-I in this respect is the peptide 37pA (called18A-Pro-18A in Reference 27). This peptide containstwo amphipathic a-helical segments, each correspond-ing to the 18A sequence, joined by a Pro residue, and ithas the sequence DWLKAFYDKVAEKLKEAFPD-WLKAFYDKVAEKLKEAF. This peptide has ahigher lipid affinity than 18A27 and can even activate theplasma enzyme lecithinxholesterol acyltransferase atleast as well as apoA-I.2* This peptide, 37pA, is alsocapable of inhibiting procoagulant activity (Fig 2), al-though it is much less potent than apoA-I in thisactivity.

ApoA-I is a major component of HDL. We thereforedetermined the effect of intact HDL on the appearance

CooM

A231J7 + Ct1*

A231J7 + C*1*+SV|/mlHDL

A231S7 + a 1 *+25«/mlHDL

A231S7 + Of*+ 12«/mlHDL

A231S7 + Ct1*+2.5«/mI HDL

5CW/mlHDL

1

\

FIG 3. Bar graph shows Inhibition of Ca2+-lnduced procoagu-lant activity by human plasma high-density llpoprotein (HDL).HDL concentrations are given as micrograms of protein permllllllter. Other conditions and abbreviations are as for Fig 1.Procoagulant activity 100% corresponds to 8.0 nmol/L thrombin;mean of 4 detennlnations±SD.

of procoagulant activity. This intact lipoprotein ap-peared at least as potent as the apoA-I in inhibitingprocoagulant activity (Fig 3). Concentrations of HDLand rHDL (below) are given in terms of the concentra-tion of their protein component. We compared apoA-Iand HDL in the same assay at the same proteinconcentrations and confirmed that HDL, at an equiva-lent protein concentration, was a more potent inhibitorof procoagulant activity than was apoA-I (Fig 4).

To further evaluate the cause of the increased po-tency of HDL over apoA-I, we prepared a series ofrHDL at various lipid-to-protein molar ratios. We foundthat all of the rHDLs were potent inhibitors of proco-agulant activity, as was the diaryzed lipid alone (Fig 5).This suggested that both the apolipoprotein(s) and lipidcontribute to the inhibition of procoagulant activity inHDL. We noted that solutions of apoA-I gradually losttheir potency in inhibiting procoagulant activity afterstorage for several days at 4°C. To more accuratelycompare the potency of apoA-I, lipid vesicles, andrHDL, we prepared these fresh and assayed theirpotency in inhibiting procoagulant activity (Fig 6). Inthis assay, 30 ng rHDL (1 nmol apoA-I) also contains100 nmol phospholipid. The relative molar ratio ofPOPC:cholesterol:apoA-I in this rHDL is 80:8:1. There-fore, the inhibitory effects of rHDL likely has contribu-tions from both the protein and the lipid.

We compared the potency of different normal humanplasma lipoprotein fractions with regard to their abilityto inhibit procoagulant activity. It is somewhat arbitrary




5 10 15 20 25 30

A23187 + Of*

AZ3187 + Of*+23pf/ml ipo A-I

AZJ187 + Of'+5pg/ml tpo A-I

A231J7 + Of*

A23187 + Of*+5«/mlHDL

H

1

•1

Epand er a/ HDL Inhibits Transbllayer Diffusion 1779


0 5 10 15 20 25 30

FIG 4. Bar graph shows comparison of Inhibitory potencies ofapollpoproteln A-I (apoA-l) and high-density lipoproteln (HDL).Conditions are as in Fig 1. Procoagulant activity of 100%corresponds to 9.3 nmol/L thrombln; mean of 3 determi-nations ±SD.

on what basis the lipoprotein fractions are compared.We have chosen to compare them on the basis of equalconcentrations of lipoprotein particles. However, theratio of the different lipoproteins can vary widely amongindividuals and also between postprandial and fastinglevels. We have compensated for the different proteincontents of the lipoproteins used so that, for example, ahigher protein concentration of HDL was used to give aconcentration of HDL comparable to the other lipopro-tein particles. The concentrations indicated in Fig 7refer to the protein component of the lipoproteinparticle. On a per-lipoprotein-particle basis, HDL wasthe most effective inhibitor and, at certain concentra-tions, was the only one of the three lipoprotein fractionsto inhibit procoagulant activity (Fig 7).

A possible mechanism for the inhibitory effects onprocoagulant activity is that the agents sequester theA23187 ionophore and inhibit the influx of Ca2+ intothe erythrocytes. This possibility was made less likely byour finding that the results were independent of theorder of addition. We obtained the same extent ofinhibition of procoagulant activity whether the A23187and Ca2+ were added before the apoA-I, HDL, rHDL,or dialyzed lipid. Adding the A23187 first would allowCa2+ entry into the cells before the possible extractionof A23187 by the hydrophobic agents. If the A23187and Ca2+ were added subsequently, this would makeCa2+ entry less likely if the ionophore bound theseinhibitors. Because procoagulant activity is independentof the order of addition, it is less likely that theinhibitors are acting by extracting the ionophore. We

A23187 + &1*

A23187 + Of*+30« 4fc2:l

A23187 + Of*+30« 80:8:1

A23187 + Of*+30p» 120:6:1

A231S7 + Of*- fXVtOttl

A231J7 + Of*+ 8&8.-0

_[

H

1

FIG 5. Bar graph shows inhibition of Ca2+-induced procoagu-lant activity by reconstituted high-density lipoprotein (rHDL).Concentrations given for rHDL and apolipoproteln A-I (apoA-I)are in terms of apoA-I concentration. The concentration of thedialyzed lipid (80:8:0) was adjusted to be the same as the lipidconcentration In the 80:8:1 sample. The ratios refer to the molarratios of 1-palmitoyl-2-oleoyl phosphatidylcholine (POPC):cho-lestercJ:apoA-l. Other conditions are as in Fig 1. Procoagulantactivity of 100% corresponds to 10.25 nmoi/L thrombln; mean of2 determinations±range.

further demonstrated that incubation of the erythro-cytes at 0°C in the presence of A23187 and Ca2+ did notlead to the generation of procoagulant activity over aperiod of up to 2 hours. When the cells that wereincubated with Ca2+ and A23187 for 20 minutes at 0°Cwere subsequently washed with HBS containing 1mmol/L Ca2+ and were then maintained at 37°C for 20minutes, procoagulant activity developed. This activitywas still inhibited by rHDL, lipid, or apoA-I (Fig 8).However, the washing procedure led to some loss ofprocoagulant activity (Fig 8), possibly because someCa2+ is washed out of the cells before removal of theA23187 or as a result of incomplete removal of theionophore by the buffer wash. In any case, most of theprocoagulant activity is retained despite the bufferwash. We further tested the inhibitory action of HDLand apoA-I in erythrocytes that have been loaded withCa2+ by incubation for 20 minutes at 0°C followed bywashing with Ca2+-free buffer in the absence of iono-phore. Even in these washed cells, in which extraction ofthe A23187 by the inhibitory agents could not lower theintracellular Ca2+ levels, HDL and apoA-I inhibitedprocoagulant activity (Fig 9).

We also directly measured the effect of apoA-I on theuptake of *5Ca2+ by erythrocytes in the presence ofA23187. The difficulty with this direct measurement isthat any procedure to remove extracellular 45Ca2+ can





0

Control

A231*7 + Of

FIG 6. Bar graph shows comparison of the inhibitory potenciesof reconstituted high-density lipoprotein (rHDL), apolipoproteinA-I (apoA-l), and dialyzed lipid. The rHDL and lipid were pre-pared by dialysis from a solution containing 1 -palmltoyl-2-olecylphosphatldylchollne (POPC):choJesterol:apoA-l etiolate at a mo-lar ratio of 80:8:1 and 80:8:0, respectively. Dialyzed POPC (100nmol) was added to the erythrocytes, which corresponds to theamount of lipid in rHDL containing 30 ng apoA-l. Procoagulantactivity of 100% corresponds to 14.5 nmot/L thrombin; mean of3 determlnatlons±SD.

also result in efflux of this ion from inside the cell, sincethe ionophore A23187 will still be incorporated into thecell membrane. We used a method of centrifuging thecells through oil that very rapidly stops the flux of Ca2+

across the cell membrane.17 This experiment directlydemonstrated that apoA-I was not affecting the abilityof A23187 to increase Ca2+ accumulation by the eryth-rocytes and that HDL had only a relatively smallinhibitory effect (Table 1). Higher concentrations ofapoA-I actually slightly increased Ca2+ accumulation.This may have been a result of the inhibition of calciumloss through microvesiculation by apoA-I (see below).In any case, the primary conclusion of this experiment isthat apoA-I inhibits a step in the formation of theprocoagulant complex subsequent to calcium uptake.

We also evaluated the role of microvesiculation in theobserved inhibition of procoagulant activity by apoA-Iand HDL. We found that 30 (ig/mL HDL completelysuppressed A23187- and Ca2+-promoted microvesicula-tion using Na+-containing buffers (Table 2). When K+

replaced Na+, no microvesiculation was observed eitherin the presence or in the absence of HDL or apoA-I.Despite the absence of microvesiculation in the pres-ence of K+-containing buffers, apoA-I and HDL stillmaintained a comparable inhibitory effect on the for-mation of procoagulant activity (not shown) as observedwhen microvesiculation occurred (Fig 4). This confirms

A231S7 + Of*+ l)il VLDL

A231I7 + Of*+1*1 LDL

A231J7 + Qf*


10 20 30 40 50

A231J7 + C l "+ 5 « V L D L

A23187 + Of*+ UV*LDL

A231J7 + Of*+25f<f HDL \

H

-H

FIG 7. Bar graph shows comparison of the inhibitory potencies ofdifferent plasma lipoprotein fractions. Procoagulant activity wasgenerated with A23187 and Ca2+ at 37°C. Concentrations givenare in terms of protein content and the relative amounts of thethree fractions used have been adjusted to give similar concen-trations of lipoproteins. Procoagulant activity of 100% corre-sponds to 7.8 nmol/L thrombin; mean of 2 determinations±range.VLDL Indicates very-low-density lipoprotein; LDL, tow-density li-poprotein; and HDU high-density llpoproteln.

that microvesiculation, which is not required for phos-pholipid transbilayer diffusion,8 is also not related to theinhibition of procoagulant activity by apoA-I or HDL.In addition, the ability of HDL to suppress microvesic-ulation may be another indication of its bilayer stabiliz-ing effects.

ApoA-I binds to anionic lipid more readily than tozwitterionic lipid in the liquid crystalline phase.29 It isthus possible that apoA-I and HDL inhibit procoagu-lant activity by binding to exposed anionic lipid. How-ever, when erythrocytes were incubated for 20 minutesat 37°C with Ca2+ and A23187 alone and then trans-ferred into a procoagulant assembly incubation withprothrombin, Ca2+, and factors Xa and Va, no inhibi-tion of procoagulant activity was observed with eitherapoA-I, rHDL, or dialyzed lipid at the highest concen-trations that would be transferred from the initialincubation of these agents with erythrocytes using thestandard assay procedures. However, if the concentra-tion of rHDL or dialyzed lipid required to inhibitprocoagulant activity when added to the erythrocytesincubated with A23187 and Ca2+ is instead addeddirectly to the procoagulant assay with prothrombin andfactors Xa and Va, then an approximately fourfoldincrease in procoagulant activity is observed. We haveno explanation for this phenomenon; however, it is notlikely to be related to changes in the rate of initiation of




% Procoagulant Activity TABLE 1. *»Ca2+ Uptake by Erythrocytes

0 5

A231S7 + Cm"*<+.+)

A231I7 + C»*(+,-)

A231»7 + &**(+,-)

A231J7 + &>•(+,-)+3C*f 10:1:1

A231I7 + Q?*(+,-)

A23ir? + C*"*(+,-)

^

^

10 15 20 25 30

F« 8. Bar graph shows inhibitory effects of reconstituted high-density lipoprotein, apollpoproteln A-l, and dialyzed llpid onprocoagulant activity in erythrocytes that had been preloaded withCa2+ at 0°C, followed by a buffer wash to remove the A23187 (seetext for details). Weight concentrations and molar ratios have thesame meaning as In Fig 5. The two + and - symbols inparentheses refer to the presence or absence of A23187 In thepreincubation of the erythrocytes at 0°C and in the subsequentincubation of the erythrocytes with the dialyzed samples at 37°C,respectively. Procoagulant activity of 100% corresponds to 9.0nmoVL thrombln; mean of 2 determinatlons± range.

the procoagulant complex by transbilayer diffusion ofphospholipids because this occurred before the additionof the inhibitors. These results demonstrate that theseagents do not act by directly inhibiting the conversion of


5 10 15 20

CuiUui

AD1I7 + Of(+,-)

+ 10rtHDL

A231I7

—i

FIG 9. Bar graph shows inhibitory effects of high-density lipo-protein (HDL) on procoagulant activity of erythrocytes that hadbeen preloaded with Ca2+ at 0°C, followed by a buffer wash andincubation of the cells at 37 °C In the absence of Ca2+ (incontrast to results presented in Fig 8) and In the absence ofA23187. HDL or apolipoproteln A-l (apoA-l) was added justbefore the 37°C Incubation. Procoagulant activity of 100% (forcells incubated at 37°C with Ca2+ and A23187) is 14 nmol/Lthrombin; mean of 2 determinations±range.

Conditions

No addition

10 Mmol/LA23187

10/tmol/LA23187+25

10 /imol/LA23187+50

10 ^mol/LA23187+25

10 Mmol/LA23187+50

Atg/mL

/ig/mL

MS/mL

Mg/mL

apoA-l

apoA-l

HDL

HDL

Cell-Associated*"Ca1+, dpmxio*

1.2±0.6

32±1

32±4

35.7±0.5

26±2

24±1

dpm indicates disintegrations per minute; apoA-l, apoJipopro-tein A-l; and HDL, high-density lipoprotein.

prothrombin to thrombin or its activity against syntheticsubstrates.

We also tested whether apoA-I, rHDL or dialyzedlipid could be extracting something from the erythrocytemembrane to inhibit procoagulant activity. Erythrocyteswere incubated for 20 minutes at 3°C with 30 /ig/mLrHDL, 50 fig/mh apoA-I, 100 nmol phosphate permilliliter of dialyzed phospholipid/cholesterol, or buffercontrol. The cells were then centrifuged and resus-pended in buffer with Ca2+ and A23187. Procoagulantactivity generated by Ca2+and A23187 in cells that hadbeen preincubated with any of the three inhibitoryagents was no lower than that of the buffer control. Thisresult demonstrates that the agents were not inhibitingprocoagulant activity by extracting something from theerythrocyte.

A more direct measure of transbilayer diffusion thanmeasurement of procoagulant activity is the measure-ment of the transbilayer asymmetry of labeled phopho-lipids. This is done by measuring the fraction of labeledlipid that is in the outer monolayer and is extractablewith BSA. The presence of A23187 and Ca2+ acceler-ates the rate of transbilayer diffusion (Fig 10). This issimilar to the findings reported by Devaux and hiscolleagues (Williamson et al).8 The presence of apoA-Iinhibits calcium-promoted transbilayer diffusion ofNBD-PS, but this protein has little effect on the basalrate of transbilayer diffusion.

DiscussionOur results demonstrate that intact HDL and rHDL

are potent inhibitors of A23187-mediated CaJ+-inducedprocoagulant activity. HDL may thus provide an effec-tive mechanism of inhibition of unregulated thrombosis.Both the lipid and apolipoprotein components of HDLappear to contribute to this inhibition. The principalprotein, apoA-I, is probably an effective inhibitor be-cause it contains segments of class A helixes that arebilayer stabilizers.3 This is supported by the finding that

TABLE 2. Inhibition of Mlcroveslculatlon by HDL

RelativeTreatment Mlcroveslculation, %

A23187

No addition

A23187+30 jtg/mL HDL protein

A23187, K+ replacing Na+

100 (by definition)

18±1

10±1

19±1

HDL Indicates high-density lipoprotein.




Fraction Externalized

o o o o o o o o o o o

ControlCETOH t C i * 1

A231B7 t Ca»*

A23187 + Ca**«• 100 ,Mfl./»l ipo A-l

A231B7 t Ca»»+ 300 pa/ml ipo A-I

100 >ig/al spo A-I

200 ^ ig / * ! «po A- I

1

H

Fra 10. Bar graph shows inhibition of Ca2+-induced transbi-layer diffusion of 1-palmitoyt-2-[6-[(7-nltro-2,1,3 benzoxadiazol-4-yl)amino]caproyl]-sn-glycero-3-phosphoserine. Erythrocyteswere Incubated for 20 minutes at 37 °C wtth or without 0.5 /imol/LA23187 and 1.0 mmol/L CaCI2 and/or apollpoproteln A-I (apoA-I); mean of 3 determinattans±SD. ETOH indicates ethyl alcohol.

the more potent model amphipathic helical peptide,27'28

37pA, can also inhibit procoagulant activity to someextent, although the smaller model peptide 18A haslittle effect. Neither of these model peptides is asinhibitory as apoA-I, which has several segments ofclass A amphipathic helixes and is a good bilayerstabilizer.3 However, on a weight basis or on the basis ofthe number of amphipathic helical segments, 37pA hasan activity comparable to that of apoA-I. In contrast,apoA-n has a weak effect on the generation of procoag-ulant activity. Although both the apoA-I and apoA-IIproteins of HDL have amphipathic helixes, the two haveopposite effects on the activity of lecithin xholesterolacyltransferase, and the motifs of their amphipathichelical domains can be distinguished.30 The difference ininhibitory potency of these two apolipoproteins suggeststhat there is some specificity for the inhibition of proco-agulant activity and that the inhibition is related to themarked bilayer stabilizing effects of apoA-I.

Because the agents used in this work inhibit thegeneration of procoagulant activity but do not affectprocoagulant activity after cells have been incubatedwith Ca2+ and A23187, we conclude that apoA-I isinhibiting transbilayer diffusion of phosphatidylserine.In addition, changes in the rates of transbilayer diffu-sion can be measured more directly with the use of thefluorescently labeled phospholipid, NBD-PS. This typeof assay requires the use of a phospholipid with a shortacyl chain that can be extracted from the membranewith BSA. This property, as well as the presence of thefluorescent probe, likely allows NBD-PS to undergosomewhat more rapid transbilayer diffusion than endog-enous phosphatidylserine. We used half the concentra-

tion of A23187 to monitor the redistribution of NBD-PScompared with the concentration used for the procoag-ulant assay. Higher concentrations of apoA-I wererequired to inhibit the transbilayer diffusion of NBD-PScompared with the concentration required to inhibitprocoagulant activity. This may be a consequence of thefact that the short chain NBD-PS undergoes more faciletransbilayer diffusion, which is more difficult to inhibit.In addition, one would not expect an exact correlationbetween the extent of inhibition of transbilayer diffusionand the inhibition of procoagulant activity. This is aconsequence of the formation of the procoagulant com-plex not being linearly proportional to the molar frac-tion of exposed anionic lipid.31 The fact that apoA-Iinhibits the externalization of NBD-PS provides furtherevidence that the inhibition of the rate of transbilayerdiffusion of phosphatidylserine contributes to loweringthe procoagulant activity.

We have identified an important mechanism by whichbilayer-stabilizing class A amphipathic helixes inhibitthe generation of procoagulant activity by slowing therate of phospholipid transbilayer diffusion. A majorprotein component of HDL, apoA-I, is composed ofmultiple segments of such class A helixes. Furthermore,it is likely that some apoA-I becomes incorporated intothe plasma membrane of blood cells.32 ApoA-I is knownto undergo more rapid transfer from lipoprotein parti-cles to cell membranes than do other serum apolipopro-teins.30 The presence of this apolipoprotein in cellmembranes will inhibit the spontaneous generation ofprocoagulant activity.

In addition to apoA-I, however, all of the lipoproteinfractions as well as dialyzed lipid also inhibit procoag-ulant activity, suggesting that there are other mecha-nisms by which this inhibition can occur. This couldinclude sequestering of A23187 by hydrophobic sub-stances. We have shown that a higher concentration of50 /xg/mL HDL can partially suppress 45Ca2+ uptake(Table 1). However, at 25 /xg/mL HDL this inhibition isonly marginal, and only 2.5 /xg/mL HDL is required toinhibit about 50% of the procoagulant activity (Fig 3).Therefore, sequestration of A23187 by HDL is not amajor mechanism in its inhibition of procoagulant ac-tivity, although it may contribute to the inhibitionobserved with dialyzed lipid or with the other lipopro-tein fractions. In addition, there could be direct inhibi-tion of protein components of the coagulation cascade.Lipoproteins have been shown to neutralize factor Xa,although most of this activity requires prior adsorptionof lipoprotein components by A1(OH)3.

33 However, thisindependent mechanism of inhibiting coagulation maycontribute to the increased potency of HDL over that ofapoA-I. HDL has been suggested to have a protectiveeffect against arteriosclerosis. This epidemiological cor-relation may be a consequence of the ability of thislipoprotein fraction to inhibit procoagulant activity incells with damaged or unstable membranes.

AcknowledgmentsThis study was supported by grants from the Medical

Research Council of Canada (grant MA-7654) and theNational Institutes of Health (grants HL-34343 andR55AR40734).




References1. Segrest JP, Jackson RL, Morrisen JD, Gotto AM Jr. A molecular

theory of lipid-protein interactions in the plasma lipoproteins.FEBS Lett. 1973^8:247-253.

2. Segrest JP, de Loof H, Dohlman JG, Brouillette CG, Ananthar-amaiah GM. Amphipathic helix motif: classes and properties.Proteins. 1990;8:103-117.

3. Tytler EM, Segrest JP, Epand RM, Mishra VK, Nie S-Q, EpandRF, Venlcatachalapathi YV, Anantharamaiah GM. Reciprocaleffects of apolipoprotein and rytic peptide analogs on membranes:cross-sectional molecular shapes of amphipathic a-helixes controlmembrane stability. / Bid Chan. 1993;268:22112-22118.

4. Toumois H, Henseleit U, DeGier J, DeKruijff B, Haest CWM.Relationship between gramicidin conformation dependentinduction of phospholipid transbilayer movement and hexagonalHn phase formation in erythrocyte membranes. Biochim BiophysActa. 1988;946:173-177.

5. Gordon DJ, Rifkind BM. High-density lipoprotein: the clinicalimplications of recent studies. N EnglJ Med. 1989;321:1311-1316.

6. Musliner TA, Krauss RM. Lipoprotein subspecies and risk ofcoronary disease. Clin Chem. 1988;34:B78-B83.

7. Badimon JJ, Badimon L, Fuster V. Regression of atheroscleroticlesions by high density lipoprotein plasma fraction in thecholesterol-fed rabbit. / Clin Invest 1990,85:1234-1241.

8. Williamson P, Kulick A, Zachowski A, Schlegel RA, Devaux PF.Ca2+ induces transbilayer redistribution of all major phospholipidsin human erythrocytes. Biochemistry. 1992;31:6355-6360.

9. Sulpice JC, Zachowski A, Devaux PF, Giraud F. Requirement forphosphatidylinositol 4,5-bisphosphate in the Ca!+-induced phos-pholipid redistribution in the human erythrocyte membrane. J BMChem. 1994;269:6347-6354.

10. Basse F, Gaffet P, Rendu F, Bienvenue A. Phospholipid transversemobility modifications in plasma membranes of activated platelets:an ESR study. Biochem Biophys Res Commun. 1992;189:465-471.

11. Basse F, Gaffet P, Rendu F, Bienvenue A. Transcolation of spin-labeled phospholipids through plasma membrane duringthrombin- and ionophore A23187-induced platelet activation. Bio-chemistry. 1993^32:2337-2344.

12. Larsen F, Katz S, Roufogalis B, Brooks D. Physiological sheerstresses enhance the Ca2+ permeability of human erythrocytes.Nature. 1981^94:667-668.

13. Anantharamaiah GM, Hughes TA, Iqbal M, Gawish A, Neame PJ,Medley MF, Segrest JP. Effect of oxidation on the properties ofapolipoproteins A-I and A-II. J Lipid Res. 1988;29:309-318.

14. Anantharamaiah GM. Synthetic peptide analogs of apoliproteins.Methods Enzymot. 1986;128:627-647.

15. Bashford CL, Rodrigues L, Pasternak CA. Protection of cellsagainst damage by haemolytic agents: divalent cations and protonsact at the extracellular side of the plasma membrane. BiochimBiophys Acta. 1989;983:56-64.

16. Lottenberg R, Christensen U, Jackson CM, Coleman PL. Assay ofcoagulation proteases using peptide chromogenic and fluorogenicsubstrates. Methods EnzymoL 1981;8O:341-361.

17. Blackburn WD Jr, Dohlman JG, Venkatachalapathi VY, PillionDJ, Koopman WJ, Segrest JP, Anantharamaiah GM. Apolipo-

protein A-I decreases neutrophil degranulation and superoxideproduction. / Lipid Res. 199132:1911-1918.

18. Pereira AC, Samellas D, Tiffert T, Lew VL Inhibition of thecalcium pump by high cytosolic Ca :+ in intact human red bloodcells. J PhysioL 1993;461:63-73.

19. Allan D, Thomas P. Ca2+-induced biochemical changes in humanerythrocytes and their relation to microvesiculation. Biochem J.1981;198:433-440.

20. Ellman GL, Courtney KD, Andres V Jr, Featherstone RM. A newand rapid colorimetric determination of acetylcholinesteraseactivity. Biochem Pharmacol 1961;7:88-95.

21. Allan D, Thomas P, Limbrick AR. The isolation and character-ization of 60 nm vesicles ("nanovesicles") produced during ion-ophore A23187-induced budding of human erythrocytes. BiochemJ. 1980;188:881-887.

22. Connor J, Schroit AJ. Transbilayer movement of phosphatidyl-serine in nonhuman erythrocytes: evidence that the aminophos-pholipid transporter is a ubiquitous membrane protein. Bio-chemistry. 1989;28:9680-9685.

23. Jonas A. Reconstitution of high-density lipoproteins. MethodsEnzymol. 1986;128:553-582.

24. Jonas A, Wald JH, Toohill KLH, Krul ES, Kezdy KE. Apolipo-protein A-I structure and lipid properties in homogeneous, recon-stituted spherical and discoidal high density lipoproteins. / BiolChem 1990:265:22123-22129.

25. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein mea-surement with the Folin phenol reagent. J Biol Chem. 1951;193:265-275.

26. Ames BN. Assay of inorganic phosphate, total phosphate andphosphatases. Methods EnzymoL 1966;8:115-118.

27. Anantharamaiah GM, Jones JL, Brouillette CG, Schmidt CF,Chung BH, Hughes TA, Bhown AS, Segrest JP. Studies of syn-thetic peptide analogs of the amphipathic helix; structure of com-plexes with dimyristoyl phosphatidylcholine. J Biol Chem. 1985;26O:10248-10255.

28. Chung BH, Anantharamaiah GM, Brouillette CG, Nishida T,Segrest JP. Studies of synthetic peptide analogs of the amphipathichelix; correlation of structure with function. / Biol Chem. 1985;260:10256-10262.

29. Surewicz WK, Epand RM, Pownall HJ, Hui S-W. Human apolipo-protein A-I forms thermally stable complexes with anionic but notwith zwitterionic phospholipids. J Biol Chem. 1986;261:16191-16197.

30. Anantharamaiah GM, Jones MK, Segrest JP. An atlas of theamphipathic helical domains of human exchangeable plasma apo-lipoproteins. In: Epand RM, ed. The Amphipathic Helix. BocaRaton, Fla: CRC Press; 1993:109-142.

31. Cutsforth GA, Whitaker RN, Hermans J, Lentz BR. A new modelto describe extrinsic protein binding to phospholipid membranes ofvarying composition: application to human coagulation proteins.Biochemistry. 1989;28:7453-7461.

32. Langdon RG. Serum lipoprotein apoproteins as major constituentsof the human erythrocyte membrane. Biochim Biophys Acta. 1974;342:213-228.

33. Barrowcliffe TW, Eggleton CA, Stocks J. Studies of Anti-Xaactivity in human plasma, II: the role of lipoproteins. Thromb Res.1982;27:185-195.



R M Epand, A Stafford, B Leon, P E Lock, E M Tytler, J P Segrest and G M Anantharamaiahactivity.

HDL and apolipoprotein A-I protect erythrocytes against the generation of procoagulant

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 1994 American Heart Association, Inc. All rights reserved.

Greenville Avenue, Dallas, TX 75231is published by the American Heart Association, 7272Arteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/01.ATV.14.11.17751994;14:1775-1783Arterioscler Thromb Vasc Biol.

http://atvb.ahajournals.org/content/14/11/1775World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

http://www.lww.com/reprints

Information about reprints can be found online at: Reprints:

document. Question and AnswerPermissions and Rightspage under Services. Further information about this process is available in the

which permission is being requested is located, click Request Permissions in the middle column of the WebCopyright Clearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of theArteriosclerosis, Thrombosis, and Vascular Biologyin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:


http://atvb.ahajournals.org/content/14/11/1775

http://www.ahajournals.org/site/rights/

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://atvb.ahajournals.org//subscriptions/


HDL and apolipoprotein AI protect erythrocytes against the generation of procoagulant activity

Documents

Transcript of HDL and apolipoprotein AI protect erythrocytes against the generation of procoagulant activity