promotion of lymphocyte growth by high density ...the high density lipoprotein (hdl) binding site on...

THE J O U R N A L OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry ’ and Molecular Biology, Inc.

Val. 264, No. 15, Issue of May 25, pp. 8549-8556.1989 Printed in U.S. A.

Promotion of Lymphocyte Growth by High Density Lipoproteins (HDL) PHYSIOLOGICAL SIGNIFICANCE OF THE HDL BINDING SITE*

(Received for publication, October 3, 1988)

Gunther Jurgensz, Qing-bo XU$, Lukas A. Huberg, Gunther Bock$, Helmut Howanietzll, Georg WickO(1, and Karine N. Trail15 From the $Institute for Medical Biochemistry, University of Graz, A-8010 Graz, Austria, the §Institute for General and Experimental Pathology, University of innsbruck, A-6020 fnnsbruck, Austria, and the (University Children’s Hospital, A-1090 Vienna, Austria

The characteristics and physiological relevance of the high density lipoprotein (HDL) binding site on unstimulated and mitogen activated human peripheral blood lymphocytes have been investigated. At 37 “C, specific binding/uptake of fluorescent (dioctadecylindocarbocyanine, DiI) HDL was observed by cells from healthy donors as well as by those from low density lipoprotein receptor-defective patients; mitogen activated T-blasts exhibited a markedly elevated DiI-HDL uptake compared to resting T-cells. Binding was saturable at 37 “C and of high affinity, with a K d of 5 X

M. It was blocked by anti-apoAI polyclonal antibodies (F(ab)z fraction), but not by anti-apolipoprotein (apo)E, anti-apoAII, or anti-apoB, and was inhibited competitively by HDL apoproteins and an apoAI-protein A fusion protein. T-cell associated DiI-HDL was increased by trypsin treatment (of the cells) and de- creased by activation in the presence of HDL or low density lipoprotein. Comparison of the concentration dependencies of growth promotion and specific cell association of HDL indicated that two mechanisms of lipid exchange may be in operation: one a binding- dependent mechanism of cholesterol exchange, with maximal effect in the HDL concentration range (20- 200 pg/ml) in which specific binding increases rapidly, and the other a binding-independent exchange of lipids effective at concentrations in which specific binding is saturated (300-5000 pg/ml).

High density lipoproteins (HDL)’ play an important role in whole body cholesterol homeostasis. One way in which they may do so is by mediating reverse cholesterol transport, the process by which excess cholesterol from peripheral tissues is

* This work was supported by grants from the Austrian Research Council (Project no. S41-01) and the Austrian Ministry of Science and Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

11 To whom correspondence should be addressed Inst. for General and Experimental Pathology, Fritz-Preglstr. 3, A 6020 Innsbruck, Austria. Tel.: 05222-507-2269.

’ The abbreviations used are: HDL, high density lipoprotein(s); FC, free cholesterol; CE, cholesterol esters; PBL, human peripheral blood lymphocytes; DiI, dioctadecylindocarbocyanine; LDL, low density lipoprotein; BSA, bovine serum albumin; PHA, phytohemagglu- tinin; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; apo, apolipoprotein; PA, staphylococcal protein A; FACS, fluorescence- activated cell sorter; FI, fluorescence intensity; FH, familial hyper- cholesterolemia.

transported in blood and lymph to the liver for excretion (1- 3). Another way is by providing certain tissues, for example hepatic and steroidogenic tissues, with cholesterol (4-8). The mechanism of cholesterol exchange between cells and HDL has been the subject of much investigation and controversy, in part because the outcome (delivery/removal) varies according to the cell type and the experimental procedures. Two mechanisms have been described. One involves a bidirectional equilibration of free cholesterol (FC) between cell membranes and HDL particles (9-11) and the other involves a unidirec- tional selective uptake of HDL cholesterol esters (CE) by the cells (4-6, 12-15). The result of FC exchange depends on the lipid and protein composition of both cells and HDL particles (9, 10, 16-19). It is probably biased in favor of HDL (i.e. reverse cholesterol transport) under normal circumstances in uiuo, except in the case of hepatic cells, where preferential uptake of FC may occur as a result of depletion of HDL phospholipids through the action of an extracellular hepatic lipase (20).

Although high affinity ( K d = 1-5 X lo-@ M ) binding sites for HDL have been described on many cell types, including macrophages (14,15,25), fibroblasts (21,26), and hepatocytes (25, 27-29), their nature and physiological significance are not yet known. They may (4, 5, 16, 22-24) or may not (9-11) be required to facilitate FC and/or CE exchange at the cell surface or intracellularly following receptor-mediated (retro) endocytosis (14, 15). Their importance in mediating reverse cholesterol transport has been implicated by the demonstra- tion that fibroblasts respond to cholesterol loading with increased HDL binding (16-18). The controversy extends to lymphocytes (30-32). We (31) and Schmitz et al. (30) have recently reported on the existence of a binding site for HDL on human peripheral blood lymphocytes (PBL) with Kd = 5 X M2 and 10 X M , respectively, whereas Cuthbert and Lipsky (32) have found no evidence for such a binding site. Our own binding studies were performed by flow cyto- metric analysis using dioctadecylindocarbocyanine (Di1)-labeled HDL. At 37 “C, DiI-HDL binding/uptake by PBL was specific (competed by HDL but not by LDL), saturable, and enhanced in the absence of divalent cations (31). The experiments described in this paper were designed to further characterize this interaction, and to investigate its physiological role in lipid exchange between HDL and PBL. All binding studies were performed in parallel with unstimulated and mitogen-activated peripheral blood lymphocytes.

G . Jurgens, Q. Xu, L. A. Huber, G. Bock, H. Howanietz, G. Wick, and K. N. Trail], unpublished observations.

8549

8550 HDL Interactions with Human PBL

EXPERIMENTAL PROCEDURES

Blood Donors-Healthy blood donors were males and females, aged 20-35 years. Blood with withdrawal was performed between 0700 and 1000 h, and all subjects were requested not to breakfast on that day. The LDL receptor-defective patient, from the Children’s Hospital in Vienna, was a 13-year-old male presenting with the clinically diag- nosed homozygous form of hyperlipoproteinemia, Type IIa. His lymphocytes showed very low levels (less than 15%) of LDL receptor activity compared with age-matched control^,^ similar to those of his elder brother (33).

Isolation of PBL-Anti-coagulated blood was diluted 1:2 in RPMI 1640 (Seromed, Berlin, Federal Republic of Germany) and peripheral blood mononuclear cells were isolated by density gradient centrifugation over Lympho-Paque (density 1.086 g/liter; Nyegaard & Co., Oslo, Norway) as described elsewhere (33). Monocytes were depleted by plastic adherance in tissue culture flasks (33) when lipoprotein uptake was to be measured; these lymphocyte suspensions were re- ferred to as PBL. Freshly isolated PBL were either used immediately after monocyte depletion or frozen in dimethyl sulfoxide and thawed shortly before required as described (33).

modified Dulbecco’s medium (Gibco) supplemented with 0.5% bovine Tissue Culture-Serum-free tissue culture medium was Iscove’s

serum albumin (BSA, Fraction V; Sigma, Munich, FRG), monothio- glycerol (Sigma, 10 pg/ml), and human transferrin (Sigma, 1 pglml). Mitogen responses were performed in triplicate in round-bottomed, 96-well tissue culture plates (Nunc) in Iscove’s modified Dulbecco’s medium in a final volume of 200 gl. Peripheral blood mononuclear cells (2.5 X 105/well) were stimulated with 25 pg/ml phytohemagglu- tinin (PHA-P; Difco). Stimulation was assessed by uptake of [lZ5I]5- iodo-2-deoxyuridine ([1251]dUrd; Amersham International, Bucks, Great Britain), 0.1 pCi/well added together with M fluorodeox- yuridine (Sigma) during a 3-h pulse on day 2, 3, or 4 of culture. Mevinolin suppression was performed as described previously (33, 34). Lipoproteins were dialyzed extensively against Iscove’s modified Dulbecco’s medium before testing in tissue culture. Concentrations tested are given in the figures. PHA-activated T-blasts for lipoprotein uptake studies were cultivated under similar conditions for 4 days in 50-ml Falcon tissue culture flasks (Becton Dickinson & Co., Oxnard, CA) in a final volume of 4 ml. If necessary, cell debris was removed by density gradient centrifugation over Lympho-Paque as for preparation of PBL from whole blood.

Isolation and Labeling of Lipoproteins-EDTA plasma was pooled from normolipemic, fasting (12-14 h) male and female donors, only weakly lipoprotein(a)-positive ( 4 mg/dl). Lipoproteins were pre- pared by differential centrifugation using solid KBr to adjust the density. The following fractions were obtained LDL, 1.020-1.050 g/ ml, HDL2,, 1.095-1.125 g/ml, and HDL3, 1.125-1.230 g/ml. HDLz. and HDLs were further purified two times on heparin-Sepharose affinity columns (Pharmacia Fine Chemicals AB, Uppsala, Sweden) to remove the apoE-containing fraction. Concentrations of all lipoproteins were determined gravimetrically by weighing an aliquot after

proportions: HDL3, 43% lipids of which 29% are CE and 6% FC; drying. Lipid concentrations were calculated assuming the following

LDL, 79% lipids of which 48% are CE and 10% FC. Lipoproteins were labeled with 3’,3’-dioctadecylindocarbocyanine

perchlorate (excitation 520 nm, emission 570 nm; Lambda Probes & Diagnostics, Graz, Austria) according to Pitas et al. (35) with the following slight modifications: 26 mg/ml HDL was incubated in 50 ml of lipoprotein-deficient plasma with 980 pg/ml DiI solubilized in 650 p1 of dimethyl sulfoxide for 8 h at 37 “C. After adjusting the density to 1.225 g/ml with KBr, DiI-HDL was re-isolated in the ultracentrifuge. The electrophoretic mobility of every batch of DiI- labeled HDL was compared on agarose gels (Lipidophor System, Immuno AG, Vienna, Austria) with that of unlabeled HDL. The DiI/ HDL molar ratio was approximately 30. Chloramphenicol (50 mg/ liter; Serva, Heidelberg, FRG), kallikrein inactivator (100,000 units/ liter; Trasylol, Bayer, Leverkusen, FRG) e-amino-n-caproic acid (1.312 g/liter; Sigma), and EDTA (1 g/liter; Merck, Darmstadt, FRG) were present during all steps of lipoprotein preparation and DiI labeling to prevent lipid peroxidation and apoB cleavage by contaminating bacteria or serum proteases.

HDL Apoproteins (apoHDL) ana‘ the apoAI-Staphylococcal Protein A (apoAI-PA) Hybrid Protein-HDL was dialyzed extensively against double-distilled water containing 0.01% EDTA, the pH adjusted to

3L. A. Huber, G. Jurgens, H. Howanietz, G. Bock, G. Wick, and K. N. Traill, submitted for publication.

7.4 with NaOH, lyophilized, and then delipidated by a severalfold extraction with chloroform/methanol (2:l); apoHDL were solubilized in 6 M urea and then dialyzed extensively against tissue culture medium.

The apoAI-PA hybrid protein was a gift from Drs. I. Monaco, H. M. Bond, and K. E. Howall (The European Molecular Biology Lab- oratory, Heidelberg, FRG). Its construction was described in detail by Monaco et al. (25).

Anti-apolipoprotein F(ab), IgG-Rabbit antisera against human apoAI, -AH, and -B were those described in Refs. 36 and 37. Anti- apoE was purchased from Immuno AG, Vienna, Austria. Purification of the IgG fractions was performed by ion exchange DEAE-cellulose chromatography. Absorbance of the effluents was measured at 280 nm and the first peak eluted was collected. The purified IgG was then dialyzed overnight against a 0.2 M acetate buffer, pH 4.5. Following dialysis, pepsin (Sigma) was added at a ratio of 20 mg/g IgG. Digestion was allowed to proceed for 12 h at 37 “C and then terminated by dialysis against 0.05 M Tris, pH 7.2. The fragments were then passed over a Sephadex G-100 column to remove the Fc portion. The columns were eluted with 0.05 M Tris, pH 7.2, and the first peak was collected. The protein content was determined by spectrophotometric analysis at 280 nm.

HDL Binding Studies-Lipoproteins were dialyzed against tissue culture medium prior to use. Saturation and competition studies were performed by incubating PBL for 2 h at 37 “C with the respective DiI-HDL concentrations (in BSA containing Iscove’s modified Dul- becco’s medium) with/without the indicated concentrations of unlabeled HDL or other competitors. All fluorescence measurements were performed on a FACS I11 (Becton Dickinson) equipped with an argon ion laser (model 2025, Spectra Physics, Mountain View, CA) and linked to an Apple I1 Plus computer (Apple Computer, Inc., Cuper- tino, CA). Details of FACS settings and methods of quantification of fluorescence intensity (FI) for binding studies have all been described in. detail elsewhere (31, 38). Briefly, laser excitation, 514.5 nm, 0.6- watt output power; optical filters, 520 nm interference long pass + 530 nm “cut on” color glass. Photomultiplier tube settings, 440-740 V depending on the fluorescence intensity (FI). Ten thousand cells were analyzed for each sample; unlabeled cells were always included as a zero reference to control for the signal/noise ratio a t each photomultiplier tube setting. FI was quantitated by determining the FI of the 50th (median) and 75th percentiles (FI = 75%) from the cumulative frequency (38). The dissociation constants (Kd) at 37 “C were obtained from double-reciprocal plots (Lineweaver-Burk plots) of the specific binding data.

Trypsin Treatment of Cells-PBL (1 X lo6 in 100 pl of BSA-free serum-free Iscove’s medium) were incubated with an equal volume of trypsin (Sigma; 1 mg/ml in the same medium) for 30 min at 37 “C and then washed extensively in normal BSA containing Iscove’s medium before labeling with monoclonal antibodies against the CD3 (Leu 4), CD 4 (Leu 3a) and CD 8 (Leu 2a) in direct immunofluorescence tests (all from Becton Dickinson; fluorescein isothiocyanate- labeled), or with DiI-LDL and DiI-HDL.

Statistics-Statistical analyses were performed using an unpaired Student’s t test.

RESULTS

The first indication that the lymphocyte binding site for HDL (31) may be of physiological importance was the dem- onstration that mitogen-activated T-blasts bindbake up much greater amounts of DiI-labeled HDL than nonactivated PBL-T (Fig. 1). The importance of lipids for optimal in vitro lymphocyte responsiveness has been realized since serum-free culture conditions were introduced (34, 39), but the essential lipids, which can be supplied by lipoproteins of all density fractions including HDL, have only recently been identified by Cuthbert and Lipsky as fatty acids (40). Although such cultures do not normally exhibit a requirement for an exogenous source of cholesterol, they depend on cholesterol when its endogenous biosynthesis is suppressed (40). In the experiment shown in Fig. 2, this was done by inclusion of mevinolin, a competitive inhibitor of hydroxymethylglutaryl-coenzyme A reductase, in the culture medium; the figure shows that apoE-free HDL can provide activated PBL with an exogenous source of cholesterol although they are 50-80-fold less effi-

HDL Interactions with Human PBL 8551

1

CHANNEL NUMBER

FIG. 1. FACS analysis of DiI-HDL3 uptake by resting and mitogen-stimulated PBL. FACS histograms of forward angle light scatter (an indicator of cell size) ( A ) and FI (B). PBL (2.5 X 106/ml) were incubated with DiI-HDL3 (50 rglrnl) in serum-free (BSA-containing) culture medium for 2 h at 37 "C. Unbound DiI-HDL was removed by 5 washes in a large excess of PBS. Dashed lines, resting T-cells, day 4, incubated with DiI-HDL. Solid lines, day 4 PHA- blasts, incubated with DiI-HDL. Dotted lines, zero control (FI of resting T-cells not incubated with DiI-HDL). Note the greater size and FI of T-blasts compared to unstimulated PBL, and that DiI- HDLa uptake is unaffected by incubation (without stimulation) in serum-free medium.

7 18 P E E 14

x

2 m : lipoprotein cholesterol (pg/ml)

FIG. 2. Rescue of mevinolin-suppressed PHA cultures by HDL. Peripheral blood mononuclear cells were stimulated with PHA in medium containing 0.56 KM mevinolin together with the indicated concentrations of LDL (0) or apoE-free HDL3 (0) expressed as total cholesterol (TC) concentration. Cultures were pulsed with [1251]dUrd (lZ5ZUdR) for 3 h on day 4. Data from a representative experiment. Hatched areas represent the mean k S.D. of triplicate cultures for controls: a, nonsuppressed; , mevinolin-suppressed.

cient than LDL, even when compared on the basis of lipoprotein cholesterol content. Subsequent experiments were designed to further characterize the HDL binding site on lymphocytes and demonstrate its independence of the LDL receptor, and to determine its physiological importance with respect to growth promotion under serum-free, normal, and mevinolin-suppressed conditions.

Characterization of the HDL Ligand Recognized by the Lym- phocyte ReceptorlBinding Site-We have already reported on some of the characteristics of the uptake of DiI-labeled HDL by freshly isolated, resting human PBL-T (31). In the experiments reported here the HDL binding sites on mitogen- activated PBL were investigated and compared with those on unstimulated lymphocytes. To do so, all experiments have been performed in parallel on resting and activated cells from the same donor. One apparently unique feature of HDL binding sites on resting lymphocytes which we previously described (31) was the increased uptake always observed when incubations were performed in EDTA/EGTA-containing me-

20 A

O ~ " ' ~ ' " ~ . ' ' I 0 50 100 150

i;: e 8

I

Y e

$?

0 50 100 150

apo H D L pg/ml FIG. 3. HDL apolipoproteins are important for DiI-HDL

uptake by PBL. PBL (2.5 X 106/ml) were labeled with 25 pg/ml DiI-HDL3 (0) or DiI-LDL (m) for 2 h at 37 "C together with increasing concentrations of delipidated HDL. Competition is expressed as percent of uncompeted control. A, freshly isolated resting T-cells; B, day 4 PHA-blasts.

1

B - . . 1 - . . r . . . *

0 50 100 150

apo AI-PA pglml

the binding of DiI-HDL. PBL (2.5 X 106/ml) were labeled with 25 FIG. 4. An apoAI-protein A hybrid molecule competes for

pg/ml DiI-HDLB (0) or DiI-LDL (M) for 2 h at 37 "C together with increasing concentrations of the apoAI-PA hybrid. Competition is expressed as percent of uncompeted control. A, freshly isolated resting T-cells; B, day 4 PHA-blasts.

dium or in medium lacking Ca2+/Mg2+ cations. This, however, proved to be negligable with T-blasts, so binding studies on T-blasts have been performed in normal medium, whereas those on unstimulated PBL were performed in EDTA-con-


taining medium. On both resting and activated T-cells, a specific binding of DiI-HDL was seen during a 2-h incubation at 37 "C, which was saturable and of high affinity (Kd = 2-5 X lo-' M). The importance of HDL apolipoproteins in DiI- HDL uptake by resting PBL-T and T-blasts is demonstrated in Fig. 3, a and b, where DiI-HDL uptake is effectively competed by delipidated HDL, whereas DiI-LDL uptake is unaffected. Competition studies with the apoAI-PA hybrid molecule underscored the importance of apoAI (Fig. 4). Dou- ble-reciprocal plots of DiI-HDL binding data in the concentration range of 20-1250 nM in the absence and presence of HDL apoproteins or the hybrid molecule confirmed that they are indeed competitive inhibitors for the DiI-HDL binding sites (Fig. 5). Further evidence for apoAI, rather than apoAII as ligand was obtained in studies with anti-apolipoprotein

" . 0.00 0.25 0.50 0.75 1.00 1.25 0 10 20 30 40 50 60

'p. k Y . ' " I " " I "

0.00 0.25 0.M 0.75 1.00 1.25

1 .z 1.0

0.8

0.6 i2 2 0.4

0.2

0.0

-0.2 0 10 20 30 40 M 60

DiI-HDL pM l/pM

FIG. 5. Lineweaver-Burk analysis of DiI-HDL binding with competitive inhibition by HDL apoproteins and apoAI-PA. PBL were labeled as described in the legends to Figs. 3 and 4 with the indicated concentrations of DiI-HDL in the presence or absence of the respective competitor (in a ratio of 5/1, w/w). A and B, freshly isolated PBL; C and D, day 4 PHA-blasts. A and C, binding curves; B and D, double-reciprocal plots of the binding data. m, control, DiI- HDL alone; 0, DX-HDL + HDL apoproteins; 0, DiI-HDL + apoAI- P A hybrid.

FIG. 6. Binding of DiI-HDL to PBL is blocked by anti-apoAI antibodies. DiI-HDL ( A and B) and DiI- LDL (C and D) were pre-incubated for 1 h at 37 "C with the indicated concentrations of F(ab)l rabbit IgG specific for apolipoproteins AI (m), AI1 (O), B (e), and E (0), before incubation for a further 2 h at 37 "C with resting T-cells (left) or day 4 PHA T-blasts (r ight) . Blocking by antibodies is expressed as percent of untreated control.

antibodies (Fig. 6). DiI-LDL and DiI-HDL were incubated for 1 h at 37 "C with pepsin-digested anti-apoAI, anti-apoAII, anti-apoB, or anti-apoE antibodies prior to binding studies. I t was essential to work with the F(ab)* antibody fraction for these experiments to prevent unwanted reactions which might take place via the Fc receptor. As expected, DiI-LDL uptake was totally blocked by its prior incubation with anti-apoB and was unaffected by the other three antibodies. DiI-HDL uptake, on the other hand, was blocked by its incubation with anti-apoAI but was unaffected by the others, including anti- apoAII. The lack of effect of anti-apoE further confirms that uptake of trace levels of contaminating apoE containing HDL via the LDL receptor does not substantially contribute to the measured DiI-HDL uptake.

However, further confirmation that HDL uptake is not taking place via the high affinity LDL receptor was obtained in binding experiments with cells from an FH patient showing very low lymphocyte LDL receptor activity (6-15% of normal) but almost normal HDL "receptor" activity (70-80%), as assessed by DiI-HDL uptake. The functional LDL receptor defect is evident in Fig. 7: low LDL concentrations (10 pg/ ml) which almost completely rescued the mevinolin-suppressed response of the healthy control were ineffective at rescuing that of the FH patient, whereas some rescue was seen at higher LDL concentrations when a receptor-independent component of cholesterol transfer comes into play. HDL rescue, on the other hand, did not differ between the control and FH patient (Fig. 7).

Effect of Trypsin Treatment on DiI-HDL Uptake-In contrast to the high affinity LDL receptor, the HDL receptor/ binding site on most cell types has proved to be relatively resistant to trypsin treatment. To further characterize our DiI-HDL-lymphocyte interaction we therefore investigated the effect of mild trypsin treatment of PBL on the subsequent binding/uptake of DiI-LDL and DiI-HDL. Because cell surface proteins show varying degrees of susceptibility to trypsin treatment, the procedure was first established by its effect on three lymphocyte differentiation antigens CD 3, CD 4, and CD 8 (Table I). Treatment of lymphocytes with 0.5 mg/ml trypsin for 30 min a t 37 "C reduced the percentage of cells bearing the CD 8 and CD 4 surface differentiation antigens to background levels but had no effect on percentage of CD 3-

se 40 60\ L "1 A

0 10 20 30 antibody concentration pg/ml

0 10 20 30 antibody concentration pg/ml

HDL Interactions with Human PBL a553

1Control:LDL y 2 0 -

: P l5- E X

0

aJ x (D 10- P 3

4-

a '0 5- 2

N, u)

7

0-

IControl:HDL3

.-$

I FH-R: LDL

-d

Lipoprotein concentration (pg/ml)

FIG. 7. High concentrations of LDL and HDL can rescue mevinolin-suppressed cultures from an FH patient with defective LDL receptor activity. Top, LDL; bottom, HDL. Day 4 PHA cultures were pulsed for 3 h with ['251]dUrd ('25ZUdR). Data are expressed as mean * S.D. for triplicate cultures. The control blood donor was age and sex-matched with the patient. 0, normal medium; 0, mevinolin-containing medium. Note that the LDL receptor defect of the FH patient is confirmed in these functional tests because 10 pg of LDL/ml was not sufficient to rescue mevinolin-suppressed cultures. Higher LDL concentrations did rescue this response, and the suppressive capacity of very high concentrations was indistin- guishable between the control and FH patient. The dose responsiveness of HDL rescue of control and FH cultures was also identical.

TABLE I Effect of mild trypsin treatment on the lymphocyte differentiation

antigens CD 3, CD 4, and CD 8 Trypsin CD 3 CD 4 CD 8

% Control 78 53 24 Before" 81 4 3 Afterb 84 48 4

"PBL (1 X 106/100 pl) were incubated 30 min at 37 'C with an equal volume of trypsin (1 mg/ml in BSA-free culture medium). They were then washed twice with BSA-containing medium and labeled with monoclonal antibodies in a direct immunofluorescence test.

* PBL were labeled first with monoclonal antibodies in a direct immunofluorescence test and then trypsin-treated as described above, before analysis.

positive cells; in contrast, similar trypsin treatment subsequent to labeling with monoclonal antibodies only affected the CD 8 antigen. Fig. 8 illustrates that the same trypsin treatment procedure, prior to incubation with DiI-labeled lipoproteins, totally abolishes the receptor-mediated uptake of DiI-LDL but enhances the uptake of DiI-HDL.

Physiological Significance of the HDL ReceptorlBinding Site on Lymphocytes-As a first approach to elucidating the functional significance of the lymphocyte receptorbinding site for apoAI in lipid transfer from HDL to activated PBL we compared the binding isotherms with those for enhancement of in vitro responses in the presence or absence of mevinolin. If specific binding were required to facilitate lipid exchange then the dose-dependent promotion of lymphocyte growth and/or rescue from mevinolin suppression should parallel that for DiI-HDL uptake, i.e. should occur in concentration range in which specific binding increases rapidly. On the other hand, if enhancement were first seen in HDL concentrations at which the specific DiI-HDL binding/uptake was already saturated this would indicate that receptor-independent lipid transfer, or receptor-independent holo-HDL uptake, is sufficient to provide T-blasts with the required lipids. Typical results are shown in Fig. 9. Mevinolin inhibited the ['251]dUrd uptake of day 4 control cultures by approximately 60% but in the presence of HDL concentrations of 200-300 pg/ml no inhibition was evident. In other words, rescue from mevinolin suppression took place in a concentration range in which the HDL receptorbinding site is not saturated. The enhancement of responses over and above the response of the control in normal medium is more difficult to interpret; receptor-independent exchange of lipids would seem largely responsible for this effect, although a receptor-facilitated exchange may play a role at lower concentrations.

To determine whether cellular HDL binding sites/receptor activity are finely attuned to lipid requirements, as for LDL, the effects of a 3-4-day incubation in LDL and HDL on the subsequent LDL/HDL receptor activity were compared. LDL receptors become up-regulated when resting cells are incubated in vitro in serum-free medium but not in medium containing 50-500 pg/ml LDL. T-blasts, in contrast, express measurable LDL receptor activity even in the presence of relatively high LDL concentrations (in the physiological range), perhaps reflecting their much greater requirements for lipids. In the experiment shown in Fig. 10, the effects of lipoprotein (LDL and HDL) supplementation of the culture medium on the subsequent DiI-LDL and DiI-HDL uptake by day 4 PHA-blasts was investigated. It can be seen that LDL not only down-regulates its own receptor but also down- regulates the HDL binding site, and vice uersa. In other words, both seem to be controlled in parallel according to the cell's requirements for cholesterol and/or other lipids.

DISCUSSION

Lymphocyte Binding Sites for HDL-We recently described a specific association of fluorescent labeled apoE-free HDL with human lymphocytes and drew attention to some of the features of that interaction which appeared to be unique to lymphocytes (31). In the present paper the nature and functional significance of the lymphocyte binding site for HDL have been investigated further. HDL have been shown to bind specifically to various cultured cells such as fibroblasts (16, 18, 21, 22, 26), arterial smooth muscle (16, 22), endothelial cells (24), hepatocytes (6,11,25,27-29,41), and macrophages (12, 14, 15, 25), but a specific receptor has not yet been isolated; ligand blotting studies of cell membrane proteins have identified the HDL receptor on several cell types as a protein of apparent molecular mass 110 kDa (25, 41), but radiation inactivation studies have shown that the HDL binding sites on fibroblast and hepatocyte membranes are a het- erogeneous species of low molecular mass (10-16 kDa) mole- cules (26,42). Indeed, even the nature of the ligand recognized by the receptor has not been unequivocally determined; how-


'"1 FIG. 8. Effect of mild trypsin

treatment on the DiI-HDL3 uptake by resting T-cells. PBL were trypsin- treated (as described under "Experimen- tal Procedures") prior to labeling with monoclonal antibodies against the CD 3 ( A ) , CD 4 ( B ) , and CD 8 (C) surface antigens in direct immunofluorescence assays or with DiI-LDL (D) or DiI-HDL ( E ) . FACS histograms of FI. -, before; and - - -, after trypsin treatment.

2501

I

1;o Fin

I 1

1 10 100 lo00 HDL concentration fpgtmlt

0

FIG. 9. Comparison of the binding isotherms for DiI-HDL uptake and those for enhancement of mitogen responses with1 without mevinolin. Data are expressed as mean k S.E. for six experiments (i.e. blood donors). Specific DiI-HDL3 uptake has been calculated as percent maximum specijzc FI for each subject specific uptake = total uptake (50 pg of DiI-HDL/ml) - nonspecific uptake (50 pg of DiI-HDL/ml + 1000 pg of unlabeled HDL/ml). Day 4 PHA responses with (0) and/or without (0) 0.56 PM mevinolin. All data are expressed as percent control (nonsuppressed) response. Hatched area represents the mean k S.E. of mevinolin-suppressed cultures.

ever, the importance of apoAI is becoming increasingly clear (14,25,43-45), also for lymphocytes (Ref. 30 and this paper), although a role for the other apolipoproteins (44, 45) as well as HDL lipids (17,42) cannot be ruled out.

Investigations of HDL binding are complicated by the dynamic exchange of apolipoproteins and lipids between lipoprotein particles, liposomes, and cell membranes. Choles- terol exchange, measured with 3H- or I4C-labeled FC or CE, often does not correlate with lZ5I-HDL binding (4-6, 12-15), and the relevance of apoAI receptor binding in lipid exchange between HDL and cells has been questioned (17, 42). This is the very question which we have addressed in the present work on human lymphocytes. Because a lipid label (DiI) was chosen for these studies, we are in reality tracing the move- ment of fluorescent lipids from HDL to lymphocytes and cannot determine whether this takes place by lipid transfer or by internalization of the HDL particle. Because there is still controversy regarding receptor-mediated endocytosis of holo-HDL we emphasize this point. Nevertheless, the process was dependent on the HDL apolipoproteins because it was totally competed by low concentrations of delipidated HDL; more specifically, apoAI was important because DiI-HDL uptake was also competed by the apoAI-protein A hybrid molecule constructed by Monaco et al. (25). As purified apoAI tends to aggregate in the absence of lipids, we used an apoAI .

E

CHANNEL NUMBER

PA complex. ApoAI. PA and HDL had been shown to compete effectively with each other for binding to the surface of G774 MO and SA0 hepatocytes by Monaco et al. (25). In addition, anti-apoAI blocked the subsequent recognition of DiI-HDL by the lymphocyte, whereas anti-apoAI1,-B, and -E were ineffective.

In support of apoAI1 as ligand, Schmitz et aZ. (30) recently demonstrated that tetranitromethane-modified HDL are not recognized by the lymphocyte receptor. It should be men- tioned that these investigators (30) were dissatisfied with DiI as fluorescent label, and preferred to use the dye tetra- methylrhodamine isothiocyanate. They based their criticism primarily on mixture experiments (varying the proportions of labeled and unlabeled HDL while keeping the end concentration constant) in which DiI apparently altered the affinity of HDL for its binding site on granulocytes. However, in identical mixture experiments we found no evidence for altered affinity (data not shown), confirming our previous conclusions on the suitability of DiI-labeled HDL for such studies. Our differing experiences with the DiI dye can probably be simply attributed to differences in the labeling protocols used.

A number of features of the interaction DiI-HDL and lymphocytes described by us appear to be unique to lymphocytes. First, we have shown that binding is saturable at 37 "C (31), whereas it is generally agreed that HDL binding at 37 "C is greater than at 4 "C (16, 22); saturable binding has been shown at 4 "C (22, 24) but not at 37 "C (17, 24), except for steroidogenic cells (44). Second, we have shown that Dil-HDL uptake by resting T-cells is enhanced in the presence of EDTA (and in the absence of divalent cations), indicating that cryp- tic binding sites may exist which become unmasked by these treatments (31). To our knowledge no reports exist which indicate a similar enhancement by EDTA to that described here, although binding of HDL (in contrast to LDL) is reported to be independent of divalent cations (22). Since we found that binding of DiI-HDL to cultured T-blasts was also relatively unaffected by EDTA we put forward the suggestion that the effect of EDTA on freshly isolated T-cells may reflect differences between resting cells and the cultured cell lines usually used for such binding studies. Incubation of PBL for 3-4 days without stimulation did not alter either the level of DiI-HDL uptake or this effect of EDTA.

It is generally agreed that the HDL receptor is relatively resistant to mild trypsin treatment at 4 "C (45,47) and 37 "C (12, 17), although binding has been reduced by harsher (1-5 mg/ml) treatment (26, 41, 48). Enhanced binding such as described here after mild trypsin treatment at 37 "c has not been reported, although a slight increase in binding after trypsin treatment is sometimes seen, e.g. of the apoA1-PA hybrid binding to hepatoma cells (25).4 Again the difference

H. Bond, personal communication.

HDL Interactions with Human PBL 8555

0 - 0 0 200 400 600 800 loo0 0 200 400 600 800 lo00

lipoprotein concentration pglml lipoprotein concentration @nl FIG. 10. HDL down-regulation of DiI-HDL uptake by day 4 T-blasts. PBL were activated with PHA in

serum-free medium supplemented with the indicated concentrations of apoE-free HDL (dotted lines) or LDL (solid lines). T-blasts were then washed extensively and labeled for 2 h at 37 "C with 25 pg/ml DiI-HDL (open symbols, A ) or DiI-LDL (closed symbols, B ) . DiI FI is expressed as percent control in serum-free medium. The zero references (which were identical) were unlabeled cells (i.e. background FI) or cells labeled with DiI-LDL in the presence of 12.5 mM EDTA (as a measure of nonreceptor-mediated DiI-LDL uptake).

in degree may reflect differences between cultured cell lines and freshly isolated cells.

Interestingly, both EDTA treatment (31) and mild trypsin treatment (Fig. 8) often revealed two populations of HDL- binding PBL, seen as two peaks in the FACS histograms of FI, which were not evident prior to these treatments. These were analyzed more closely in double fluorescence tests with monoclonal antibodies against various lymphocyte differentiation antigens (data not shown). Both bright and dull populations were primarily T-cells, composed of both CD 4 and CD 8 phenotypes, although there was an enrichment of CD 8- positive cells in the bright peak.

The Functional Relevance of the Lymphocyte HDL apoAI Binding Site-LDL and HDL are usually considered in terms of their respective roles in cholesterol transport to and from extrahepatic tissues. However this role-typing is not always correct (or useful) because HDL can also deliver cholesterol to certain extrahepatic tissues, e.g. steroidogenic tissues (7- 10). In addition, the other lipid components of lipoproteins, which are often ignored or considered important only in terms of their effects on cholesterol solubility, have been shown to play a more important growth-supporting role than cholesterol, at least in vitro (38). This is particularly true for cultured lymphocytes, which require an exogenous source of fatty acids or phospholipids for optimal proliferation in response to mi- togens but can synthesize sufficient cholesterol to meet their own needs (34,40). HDL, including HDL3, have been shown to be equally as efficient as LDL at promoting the growth of cultured lymphocytes (40). This has been confirmed in the present work (Fig. 2). At high concentrations in culture LDL are generally suppressive (49-55); HDL, in contrast, do not suppress mitogenesis even at very high concentrations (e.g. Fig. 2). Gospdorowicz e t al. (56) suggested that LDL suppression, and HDL lack of suppression, a t high concentrations may reflect differences in their route of internalization. We feel this is unlikely since we have recently shown that LDL suppression is seen only at concentrations where it is taken up by receptor-independent mechanisms, i.e. concentrations at which the LDL receptor is saturated.' At such concentrations, receptor-independent uptake of HDL is also evident (Fig. 9), and, although possible, it is unlikely that nonreceptor- mediated uptake of the two types of lipoprotein particle would result in their delivery to different intracellular compart- ments.

We have demonstrated here that under conditions of mevinolin suppression of cholesterol biosynthesis HDL, including

HDL3, can rescue responsiveness, albeit relatively ineffi- ciently, when compared with LDL on a cholesterol basis (total or free cholesterol). We would like to point out that our blood donors were all screened and selected as lipoprotein(a)-nega- tive to ensure that our preparations were not contaminated with lipoprotein(a), a lipoprotein which overlaps in density with LDL and HDL2.. In this respect it should be emphasized that heparin-Sepharose-purified HDL2. and HDLS were used in all of the present experiments, that HDL apolipoproteins did not compete for DiI-LDL binding, and that anti-apoE had no blocking effect on DiI-HDL uptake, all of which support our contention that we are not measuring uptake of apoE- containing HDL via the LDL receptor. In addition, the LDL receptor was excluded as possible mediator of cholesterol exchange by comparing HDL rescue of mevinolin-suppressed cultures from an LDL receptor-defective patient with those from an age-matched healthy subject. The discrepancy between our results and those of Cuthbert and Lipsky (32) may lie in the composition of the culture medium, e.g. BSA- containing medium is said to facilitate cholesterol exchange between lipoproteins and cells while reducing phospholipid exchange (10).

On the basis of our binding isotherms, we propose that the HDL binding site on lymphocytes may be of functional importance in limiting HDL concentrations to facilitate lipid transfer from HDL to cells but that at high concentrations ( i e . normal plasma concentrations) receptor-independent mechanisms would be sufficient. If this is true, one would ask what is the point of this receptor and how and why is it regulated? It has been postulated to be of importance in reverse cholesterol transport from cultured fibroblasts because receptor levels are strictly correlated with cellular cholesterol levels (16-18). We have been unable to test the relevance of this receptor in reverse cholesterol transport from lymphocytes because we have repeatedly failed to load lymphocytes with cholesterol. Using cholesterol concentrations which are very effective in loading fibroblasts, and which are effective in down-regulating the lymphocyte LDL receptor and in rescuing mitogen-stimulated PBM from mevinolin suppression, we have not been able to measure a change in DiI-HDL uptake (not shown) or a marked decrease in membrane fluidity (34). We have, however, been able to measure significantly higher levels of DiI-HDL associated with activated T-cells than with resting cells, and we have shown that the binding sites on T-blasts are also of high affinity (5 X lo-* M) and specific for apoAI. It thus appears that DiI-HDL


binding sites become up-regulated in response to a cell’s increased requirements for lipids, as is the case for LDL receptors in response to cholesterol deprivation. If this is true, then provision of an exogenous source of lipids should prevent this up-regulation, and this did indeed prove to be the case. T-blasts activated in medium containing HDL expressed less binding sites than those activated in serum-free medium. However, to our surprise LDL was also able to down-regulate the HDL binding sites, and, in the reciprocal experiments, HDL was found to down-regulate the LDL receptor. Both, therefore, seem to be under a similar type of control which is sensitive to a cell’s requirements for cholesterol and/or other lipids. The implications of this are far-reaching. For example, we recently showed that both LDL and HDL binding was significantly elevated in resting lymphocytes from healthy elderly subjects and suggested that HDL binding may be increased to compensate for increased lipid uptake via LDL, i.e. for reverse cholesterol transport (57). However, if these receptors are normally regulated in parallel on lymphocytes this might reflect a serious loss of homeostatic regulation during aging which might give rise to some of the defects in cellular immunity with age.

Acknowledgments-We thank Dr. H. A. Dresel, Boehringer Mann- heim, for his stimulating advice throughout the various stages of this work. We also thank I. Zoderer, G . Ledinski, and J. Stangl for their technical assistance, I. Atzinger for her help in preparation of the figures, and D. Schonitzer and his co-workers in the Innsbruck Blood Bank for their assistance in blood withdrawal and health screening of blood donors.

1.

2. 3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

REFERENCES Tall, A. R., and Small, D. M. (1978) N . Engl. Med. 299, 1232-

Eisenberg, S. (1984) J. Lipid Res. 25, 1017-1058 Miller, G. J. (1980) Annu. Rev. Med. 31 , 97-108 Arbeeny, C. M., Rifici, V. A., and Eder, H. A. (1987) Biochim.

Leitersdorf, E., Israeli, A., Stein, O., Eisenberg, S., and Stein, Y. (1986) Biochim. Biophys. Acta 878, 320-329

Glass, C., Pittman, R. C., Civen, M., and Steinberg, D. (1985) J. Biol. Chem. 260,744-750

Kovanen, P. T., Schneider, W. J., Hillman, G . M., Goldstein, J. L., and Brown, M. S. (1979) J. Biol. Chem. 254,5498-5505

Hwang, J., and Menon, K. M. J. (1983) J. Biol. Chem. 258,8020- 8027

Johnson, W. J., Bamberger, M. J., Latta, R. A., Rapp, P. E., Phillips, M. C., and Rothblat, G. H. (1986) J. Biol. Chem. 261,

Lundberg, B. B., and Suominen, L. A. (1985) Biochem. J. 232,

Karlin, J . B., Johnson, W. J., Benedict, C. R., Chacko, G. K., Phillios. M. C.. and Rothblat. G . H. (1987) J. Biol. Chem. 262,

1236

Biophy~. Acta. 917 , 9-17

5766-5776

219-225

12557-12564 Stein, O., Israeli, A,, Leitersdorf, E., Halperrin, G. , and Stein, Y.

(1986) Atherosclerosis 75, 151-158 Miller, N. E., Weinstein, D. B., and Steinberg, D. (1977) J. Lipid

Schmitz, G., Robenek, H., Lohman, U., and Assman, G. (1985)

Schmitz, G., Robenek, H., and Assman, G. (1986) Klin. Woch-

Oram. J. F., Brinton, E. A., and Bierman, E. L. (1983) J. Clin.

Res. 18,438-450

EMBO J. 4,613-622

enschr. 64,971-985

Invest. 72, 1611-1621 Tabas. I.. and Tall. A. R. (1984) J. Biol. Chem. 259,13897-13905 Oppenhe’imer, M.’J., Oram, J. F., and Bierman, E. L. (1987)

DeLamatre, J., Wolfiauer, G., Phillips, M. C., and Rothblatt, G.

Bamberger, M., Lund-Katz, S., Phillips, M. C., and Rothblatt, G.

Arteriosclerosis 7, 325-332

H. (1986) Biochim. Biophys. Acta 875, 419-428

H. (1985) Biochemistry 24,3693-3701

21. Brinton, E. A., Oram, J. F., Chen, C.-H., Albers, J . J., and

22. Biesbroeck, R., Oram, J. F., Albers, J. J., and Bierman, E. L.

23. Reaven, E., Chen, Y.-D. I., Spicher, M., and Azhur, S. (1984) J.

24. Tauber, J . P., Goldminz, D., Vlodowski, I., and Gospodarowicz,

25. Monaco, L., Bond, H. M., Howell, K. E.. and Cortese. R. (1987)

Bierman, E. L. (1986) J. Bwl. Chem. 261,495-503

(1983) J. Clin. Invest. 71, 525-539

Clin. Invest. 74, 1384-1397

D. (1981) Eur. J. Biochem. 119, 317-325

EMBO J.’ 6, 3253-3260. 26. Mendel. C. M.. Kunitake. S. T.. Kane. J. P.. and KemDner. E. S.

(1988) J. B&. Chem. 263,1314-1319 ’

- ,

27. Chacko, G. R. (1985) J. Lipid Res. 26,745-754 28. Chacko, G. K., Lund-Katz, S., Johnson, W. J., and Karlin, J. B.

29. Rifici, V. A., and Eder, H. A. (1984) J. Biol. Chem. 259, 13814-

30. Schmitz, G., Wulf, G., Briining, T., and Assman, G. (1987) Clin.

31. Traill, K. N., Jurgens, G., Bock, G., and Wick, G. (1987) Zmmu-

32. Cuthbert, J. A., and Lipsky, P. E. (1987) J. Biol. Chem. 262,

33. Traill, K. N., Jurgens, G., Bock, G., Huber, L. A., Schonitzer, D., Widhalm, K., Winter, U., and Wick, G. (1987) Mech. Ageing Dev. 40, 261-288

34. Traill, K. N., Offner, F., Winter, U., Paltauf, F., and Wick, G. (1988) Zmmurwbiol. 176,450-464

35. Pitas, R. E., Innerarity, T. L., Weinstein, J. N., and Mahley, R. W. (1981) Arteriosclerosis 1, 177-185

36. Jurgens, G., and Koltringer, P. (1986) Neurology 37,513-515 37. Jurgens, G., and Kostner, G. M. (1975) Immunogenetics 1,560-

569 38. Traill, K. N., Bock, G., Winter, U., Hilchenbach, M., Jurgens, G.,

and Wick, G. (1986). J. Histochem. Cytochem. 34, 1217-1222 39. Iscove, N. N. (1984) in Methods in Molecular and Cell Biology

(Barnes, D., Sirbasku, D., and Sato, G., eds) pp. 169-185, Alan

40. Cuthbert. J. A.. and Liuskv, P. E. (1986) J. Biol. Chem. 261, R. Liss, Inc., New York

(1987) J. Lipid Res. 28,332-337

13818

Chem. 33,2195-2203

nobwlogy 175,447-454

7808-7818

3620-3627 ’

- - .

41. Graham. D. L.. and Oram. J. F. (1987) J. Biol. Chem. 262,7439- 7442

42. Mendel, C. M., Kunitake, S. T., Hong, K., Erickson, S. K., Kane J. P., and Kempner, E. S. (1988) Biochim. Biophys. Acta 961,

43. Rifici, V. A., and Eder, H. A. (1984) J. Biol. Chem. 259, 13814-

44. Fidge, N. H., and Nestel, P. J. (1985) J. Biol. Chem. 260, 3570-

45. Hwang, J., andMenon, K. M. J. (1985) J . Biol. Chem. 260,5660-

46. Chen, Y.-D. I., Kraemer, F. B., and Reaven, G. M. (1980) J. Biol.

47. Chacko, G. K. (1982) Biochim. Biophys. Acta 712, 129-141 48. Mendel, C. M., Kunitake, S. T., and Kane, J. P. (1986) Biochim.

Biophys. Acta 875,59-68 49. Hui, D. Y., and Harmony, J. A. K. (1980) Proc. Natl. Acud. Sci.

U. S. A. 77,4764-4768 50. Harmony, J. A. K., and Hui, D. Y. (1981) Cancer Res. 41,3799-

3802 51. Scupham, D. W., McCarthy, B. M., and Harmony, J. A. K. (1987)

Cell. Zmmunol. 108,378-395 52. McCarthy, B. M., Okano, Y., Nakayasu, T., Macy, M., Watson,

S. R., and Harmony, J. A. K. (1987) J. Lipid Res. 28, 1067- 1077

53. Cuthbert, J. A., and Lipsky, P. E. (1984) Proc. Natl. Acud. Sci.

54. Cuthbert, J. A., and Lipsky, P. E. (1983) J. Lipid Res. 24, 1512- 1524

55. Cuthbert, J. A., and Lipsky, P. E. (1984) J. Clin. Invest. 73,992- 1003

56. Gospodarowicz, D., Lui, G.-M., and Gonzalez, R. (1982) Cancer Res. 42,3704-3713

57. Traill, K. N., Huber, L. A., Blick, G., Jurgens, J., and Wick, G. (1988) in Crossroads in Aging (Bergener, M., Ermini, M., and Stahelin, H. B., eds) pp. 125-134, Academic Press, London

188-193

13818

3575

5668

Chem. 256,9162-9167

U. S . A. 81,4539-4543

promotion of lymphocyte growth by high density ...the high density lipoprotein (hdl) binding site on...

Documents