Eur. J. Biochem. 205, 775 - 784 (1992) fj

FEBS 1992

Characterization of the chylomicron-remnant-recognition sites on parenchymal and Kupffer cells of rat liver Selective inhibition of parenchymal cell recognition by lactoferrin Marc C. M. van DIJK, Gijsbertus J. ZIERE and The0 J. C. van BERKEL

Division of Biopharmaceutics, Center for Bio-Pharmaceutical Sciences, University of Leiden, Sylvius Laboratory, The Netherlands (Received October 29, 1991/January 6, 1992) - EJB 91 1449

Upon injection of chylomicrons into rats, chylomicron remnants are predominantly taken up by parchenymal cells, with a limited contribution (8.6% of the injected dose) by Kupffer cells. In vitro storage of partially processed chylomicron remnants for only 24 h leads, after in vivo injection, to an avid recognition by Kupffer cells (uptake up to 80% of the total liver-associated radioactivity). Lactoferrin greatly reduces the liver uptake of chylomicron remnants, which appears to be the consequence of a specific inhibition of the uptake by parenchymal cells. Kupffer-cell uptake is not influenced by lactoferrin. In vitro studies with isolated parenchymal and Kupffer cells show that both contain a specific recognition site for chylomicron remnants. The Kupffer-cell recognition site differs in several ways from the recognition site on parenchymal cells as follows. (a) The maximum level of binding is 3.7-fold higher/mg cell protein than with parenchymal cells. (b) Binding of chylomicron remnants is partially dependent on the presence of calcium, while binding to parenchymal cells is not. (c) 8-Migrating very-low-density lipoprotein is a less effective competitor for chylomicron-remnant binding to Kupffer cells compared to parenchymal cells. (d) Lactoferrin leaves Kupffer-cell binding uninfluenced, while it greatly reduces binding of chylomicron remnants to parenchymal cells, The properties of chylomicron-remnant recognition by parenchymal cells are consistent with apolipoprotein E being the determinant for recognition. It can be concluded that the chylomicron-remnant recognition site on Kupffer cells possesses properties which are distinct from the recognition site on parenchymal cells. It might be suggested that partially processed chylomicron remnants are specifically sensitive to a modification, which induces an avid interaction with the Kupffer cells. The recognition site for (modified) chylomicron remnants on Kupffer cells might function as a protection system against the occurrence of these potential atherogenic chylomicron-remnant particles in the blood.

[2- 61. For the contribution of non-parenchymal cells to the total liver uptake, varying percentages were reported ranging from 6-35% [2-71. In rats and dogs [8, 93, liver uptake is the major removal route for chylomicron remnants, while in marmosets, guinea pigs and rabbits, other organs [lo] were also reported to contribute to a significant extent to the uptake of chylomicron remnants. In the rabbit, for example, about 18% of the injected dose of [3H]vitamin-A-labeled chylomicron remnants is taken up by bone marrow [l I]. Patients with type-I1 hyperlipoproteinemia [I21 and WHHL rabbits [13], which both possess genetically defective receptors for low-density lipoproteins (LDL), show a normal decay rate Correspondence to Th. J . C. van Berkel, Division of Biopharma- of chylomicron remnants, leading to the conclusion that a ceutics, Ccnter for Bio-Pharmaceutical Sciences, Sylvius Laboratory, receptor system distinct from the LDL receptor, tentatively University of Lciden, P. 0. Box 9503, NL-2300 RA Leiden, The called the remnant receptor, must be responsible for the avid Netherlands liver uptake. F ~ J x+ . 31 71 276292 In vitro studies with isolated parenchymal cells indicated Ablmviations. d, density (glml); LDL, low-density lipoprothat the recognition site for chylomicron remnants on these tein (1.025 g/ml< d < 1.055 g/ml); HDL, high-density lipoprotein (1.063g/ml< d < 1.055 g/ml); P-VLDL, P-migrating very-low-den- cell types differs from the classical LDL receptor [14] in being resistant to regulation [15, 161, independent of Ca2+ and recsity lipoprotein (d < 1.006 g/ml); GdCI3, gadoliniumchloride. Enzymes. Collagenase (EC 3.4.24.3); cholesterol esterase (EC ognizing probably only apolipoprotein E and not apoprotein B [16, 171. In recent years, much attention has been focussed 3.1.1.13).

Dietary lipids are processed by intestinal enterocytes into chylomicrons, whch are excreted in the lymphatics. Chylomicrons are processed to chylomicron remnants after the lymph drains into the bloodstream. Processing is a result of both triacylglycerol hydrolysis by lipoprotein lipase on capillary endothelial cells and exchange of surface material and apolipoproteins with high-density lipoprotein (HDL) [I]. Chylomicron remnants are rapidly cleared from the blood circulation, mainly by the liver, for which both parenchymal cells and non-parenchymal cells are reported to be responsible

776 on the isolation and characterization of the tentative remnant receptor [IS, 191, leading to the identification of apolipoprotein E-binding proteins, one of which was identical to the MgATPase [20]. More recently, Herz et al. [21] identified a 500-kDa protein which has a structural resemblance to the LDL receptor and is tentatively called the LDL-receptor-related protein. Evidence has been provided that the LDLreceptor-related protein can bind artificially apoprotein-Eenriched lipoproteins [22- 251. Recently Strickland et al. I261 showed, however, that LDL-receptor-related protein and the a2-macrog1obuh receptor are the same molecule and it was suggested that the LDL-receptor-related protein might act as a multifunctional receptor [26, 271. In order to characterize the chylomicron-remnant uptake by the liver, we felt it necessary to take recognition by parenchymal and non-parenchymal cells into account. The present investigations were aimed to investigate the criteria for chylomicron-remnant uptake in liver parenchymal and non-parenchymal cells, in order to define (a) the specific cell type within the non-parenchymal cells which is responsible for liver cell uptake, (b) the factors which determine the relative importance of the various liver cell types for uptake and (c) the characteristics of recognition of chylomicron remnants by the relevant cell types. It was recently shown that lactoferrin, a 76.5-kDa glycoprotein, with four terminal arginine residues like apolipoprotein E, inhibited the uptake of chylomicron remnants by total liver by 50% [28]. We used lactoferrin in the present study to identify the possible determining role of apoprotein E in the recognition of chylomicron remnants by liver parenchymal and non-parenchymal cells, both in vivo and in vitro. Finally, the possible physiological function of the chylomicron-remnant-uptake systems in the different liver cell types and extrahepdtic sites is discussed.

MATERIALS AND METHODS Chemicals Collagenase, pronase and cholesterol esterase were from Boehringer Mannheim. Hepes was from Merck. Bovine serum albumin (fraction Vj, heparin (sodium salt, grade I> 3,3'diaminobenzidine and polyinosinic acid (potassium salt) were from Sigma. Human lactoferrin was from Servd, Heidelberg, FRG. Metrizamide was obtained from Nycomed A/S, Oslo, Norway. Culture media were from Gibco. 4-Aminopyrazolo pyrimidine, cyclohexanedione and gadoliniumchloride (GdC1,) were from Janssen. '''1 (carrier free) in NaOH and [3H]vitamin A were from Amersham. EGTA was obtained from Fluka. All other chemicals were of analytical grade.

Animals Male Wistar rats, fed ad libitum, were used in this study. For determination of liver uptake and serum decay, rats were fasted for 16 h. For isolation of P-migrating very-low-density lipoprotein (P-VLDL), 6 - 8 rats of mass 200-220 g were maintained for 16 d on a cholesterol-rich chow (Hope Farms, Woerden, The Netherlands) that included 2% cholesterol, 5% olive oil and 0.5% cholic acid. When indicated, rats were injected via the vena penis with 20 pmol/kg GdCI, in 0.9% NaCl and 5 mM glycine, pH 3.5, 24 h before the experiment was started. GdC13 treatment of rats leads to a total liver depletion of Kupffer cells, while liver parenchymal and endothelial cells are not influenced [29].

Lipoproteins ~~

Chylomicrons were isolated from rat lymph, obtained via canulation of the thoracic duct. Briefly, rats were anaesthetized with diethyl ether and the main intestinal lymph duct was canulated with silastic tubing (external diameter 0.635 mm, internal diameter 0.3 mm). The duodenum was canulated and rats were placed in restraining cages. A lipid emulsion (Intralipid, Vitrum, Stockholm, Sweden, supplemented with 1Yocholesterol) was infused through the intestinal tubing at a rate of 1.2 - 1.5 ml/h. 4 h after surgery, lymph was collected on ice in a plastic tube containing 0.1 ml 0.1 M EDTA. Chylomicrons were isolated from rat lymph by ultracentrifugation for 1 h at 40000 rpm in a Beckman SW40 rotor at 4'C, washed 3 times with 0.9% NaCI, 8 mM sodium phosphate (NaCl/Pi) containing 1 mM EDTA, pH 7.4. For isolation of [3H]vitamin-A-labeledchylomicrons, 2 ml lipid emulsion, supplemented with 1YOcholesterol, was mixed with solvent-free 0.3 mCi [3H]vitamin A [3]. 4 h after surgery, the radioactive lipid emulsion was infused into rats and lymph was isolated similarly as described above. Chylomicron remnants were isolated from functionally hepatectomized rats [30, 311. To minimize the endogenous VLDL level in recipient animals, the rats (fasted for 16 h) were treated with 4-aminopyrazolo pyrimidine (solubilized in Tris/ HCl buffer, pH 3.5; 30 mg/kg body mass) 5.5 h before surgery and injected intraperitoneally [6]. 2 min before functionally hepatectomy, 200 U heparinlkg body mass was administered to stimulate processing of chylomicrons to chylomicron remnants. 5 - 15 mg chylomicron triacylglycerol (unlabeled, '"Ilabeled or [3H]vitamin-A-labeledchylomicrons) were injected into the bloodstream and 5 min or 30 min later (5-min remnants and 30-min remnants, respectively) blood was collected by puncture of the abdominal aorta. For '251-labelled chylomicron remnants only, the circulation time was 30 min. The radioactivity in the liver was analysed under these conditions and never exceeded 1% of the injected dose. Blood was allowed to clot for 30 min at 25°C. Serum was prepared by centrifugation for 10 min at 3000 rpm. ['251]Chylomicron remnants and unlabeled remnants were isolated according to Redgrave et al. [32]. The upper 1 ml of the KBr gradient ( d < 1.006 g/ml; d, density) was obtained by tube slicing and remnants were dialysed against NaC1/Pi 1 mM EDTA, pH 7.4, at 4'C for 20 h, with repeated changes of buffer. With [3H]vitamin-A-labeled chylomicron remnants, serum was injected into a recipient rat within 4 h of isolation or 24 h later (whole serum stored at 20°C under nitrogen). P-VLDL was obtained from rats fed as described above and blood was collected from 20-h fasted rats by puncture of the abdominal aorta. The sera were pooled and P-VLDL was isolated as described before [33]. Human LDL (1.019 g/ml < d < 1.063 g/ml) was isolated from the serum of fasted volunteers by two repetitive centrifugations, according to the procedure of Redgrave et al. [32]. LDL was removed by tube slicing and was virtually free of apolipoprotein E as described earlier [33].

Labeling of chylomicrons Chylomicrons were radioiodinated at pH 10.0 with carrierfree '''I according to a modification [34] of the ICI method [35]. Free '''I was removed by Sephadex G25 gel filtration, followed by dialysis against NaC1/Piand 1 mM EDTA, pH 7.4 for 20 h at 4"C, with repeated changes of buffer. The distribution of radioactivity in labeled chylomicron remnants (de-

777 termined after processing of chylomicrons into remnants) was 77.3 _+ 7.3% in protein, 17.9 _+ 4.5% in lipid and 6.2 k 2.3% as unbound '251. The distribution of the protein-associated radioactivity in the various apolipoproteins was determined by electrophoresis of 5 pg '251-labeled remnant protein on 5-20% SDS/polyacrylamide gels. The gels were stained with Coomassie brilliant blue and identified protein bands were cut out and counted for radioactivity. For 30-min remnants, it was found that 16.3 k 5.3% of the protein-associated radioactivity was associated with apoprotein B, 4.3 k 0.5% with apoprotein AIV, 5.0 ? 0.8% with apoprotein E, 45.8 k 5.3% with apoprotein A1 and 28.7 4.0% with apoprotein C species.

bovine serum albumin at pH 7.4. Incubations were carried out in plastic tubes (8.5 ml, Kartell) for 10 min at 37°C or 2 h at 4°C with continuously shaking (150 rpm). Every 30 min, the air was saturated with oxygen by using Carbogen (95% 0 2 / 5 % COz). After incubation, the cells were centrifuged at 50 x g for 1 min at 4°C (parenchymal cells) or 500 x g for 5 min at 4°C (Kupffer cells), and washed twice with 0.9% NaCl, 1 mM EDTA, 0.05 M Tris/HCl, 5 mM CaClz and 0.2% bovine serum albumin, pH 7.4 and twice with this solution without bovine serum albumin. Cells were lysed in 1 mlO.1 M NaOH and subsequently radioactivity and protein content were determined. Lipoprotein composition

Liver uptake and serum decay of lipoproteins Rats were anaesthetized by intraperitoneal injection of 15 - 20 mg sodium pentobarbital and the abdomen was opened. Radiolabeled lipoproteins were injected via the inferior vena cava. Where indicated, rats received an injection of lactoferrin (70 mg/kg body mass) 1 rnin prior to injection of radiolabeled lipoprotein. At the indicated times, blood samples of 0.3 ml were taken from the inferior vena cava and allowed to clot for 30 min. The samples were centrifuged for 2 min at 16000 x g and the radioactivity of 100 p1 serum samples determined. The total amount or radioactivity in the serum was calculated using the following equation: serum volume (ml) = [0.0219 x body mass (g)] 2.66 [36]. In order to determine the uptake by the liver (always measured simultaneously with serum decay) at the indicated times, liver lobules were excised and the mass determined. The amount of liver tissue tied off at the end of the experiment did not exceed 15% of the total liver mass. Radioactivity was corrected for the radioactivity in plasma assumed to be present in the tissue at the time of sampling (85 pl/g wet mass) [37]. Radioactivity was counted by combusting the liver and tissue samples in a Packard Tri Carb 306 Sample Oxidizer, always yielding a recovery of > 97% label.

+

The lipid composition of lipoproteins was determined using Boehringer enzymatic kits for triacylglycerols, phospholipids and free cholesterol. Esterified cholesterol was determined by adding 20 pg cholesterol esterase after the determination of free cholesterol. Intralipid (Boehringer) was used as an internal standard. Protein contents were determined according to Lowry et al. [41] with bovine serum albumin as a standard. The composition of chylomicrons was as follows: 91.3 & 0.9% triacylglycerol, 5.6 ? 1.1% phospholipids, 0.7 +_ 0.2% cholesterol esters, 1.3 5 0.3% free cholesterol and 2.1 5 0.3% protein. The composition of 5-min remnants was 86.2 5 1.2% triacylglycerol, 7.1 3.0% phospholipids, 3.0 0.1% cholesterol esters, 1.4 f 0.3% free cholesterol and 4.7 k 0.2% protein. The composition of 30-min remnants was 68.3 ? 1.6% triacylglycerol, 8.6 2 1 . I % phospholipids, 5.1 k 0.7% cholesterol esters, 9.1 k 1.2% free cholesterol and 9.0 1.1% protein. The thiobarbituric-acid reactivities of 5-min remnants, 30-min remnants and each after storage for 24 h before usage, were 4.3, 7.3, 4.8 and 4.9 nmol malondialdehydejmg protein, respectively and were measured as published before [42]. Cyclohexanedione modification of lactofemn

Isolation of liver cells Rats were anaesthetized and injected with the radiolabeled lipoprotein. If indicated, rats received an injection of lactoferrin (70 mg/kg body mass) into the bloodstream 1 rnin prior to injection of (1ipo)proteins. Rat liver parenchymal, endothelial and Kupffer cells were isolated by differential centrifugation and counterflow elutriation as described in detail elsewhere [38]. The contribution of the different liver cell types to the uptake of (1ipo)proteins was determined under the assumption that 92.5% of the total liver protein obtained was from parenchymal cells, 3.3% from endothelial cells and 2.5% from Kupffer cells [39,40]. Kupffer and endothelial cells were more than 95% pure as judged from peroxidase staining (0.1'YO3,3'-diaminobenzidine in 0.05 M Tris/HCl, 7% sucrose and 0.1% (by vol.) 30% H 2 0 z ,pH 7.4 for 20 rnin at 37°C). For in vim studies the different liver cells were isolated as described before [4].

Arginine residues of lactoferrin were modified according to procedures described earlier [43,44]. 100 mg lactoferrin in 1 ml 0.15 M NaCl and 0.01% EDTA was mixed with 2 ml 0.15 M 1,2 cyclo-hexanedione in 0.2 M sodium borate, pH8.1, and incubated for 17 h at 35°C. The sample was dialysed for at least 36 h against NaC1/Pi and 0.01 YOEDTA pH 7.4 at 4°C. The protein content of the sample was determined by the absorption at 280 nm. Modification of arginine residues on lactoferrin was controlled by subjecting the sample to electrophoresis in agarose gels at pH 8.8 (Tris/hippuric acid, pH 8.8). After electrophoresis, bands were visualized with Coomassie brilliant blue. For lactoferrin and cyclohexanedione-modified lactoferrin, Rf values of 0.07 and 0.14 were obtained, respectively.

RESULTS

In vitro studies with freshly isolated cells

Liver uptake and serum decay of chylomicrons, 5-min remnants and 30-rnin remnants

For in uitro studies, 1- 2 x 10' rat liver parenchymal cells viable as judged by 0.2% trypan blue staining) or Kupffcr cells were incubated with 5.0 pg/ml radiolabeled lipoproteins and indicated amounts of competitor in 0.5 ml Dulbecco's Modified Eagle's Medium supplemented with 2%

Injection of chylomicrons into anaesthetized rats leads to processing of these particles by endothelial-cell-bound lipoprotein lipase, resulting in chylomicron remnants which are rapidly taken up by the liver (Fig. 1A). As a consequence of endothelial cell binding and lipoprotein lipase processing,

( > 95%

778

C

A 125

LIVER UPTAKE

LIVER W A K E 100

J

GdCl

1)1

T D

U

a,

75

A2

-b

IW

a,

C .-

GdClJlacto

50 Y-

Lc

0

s

100

lactofenin

0

s

25

lactofenin

i-9 OY 0

0 10

30

20

0

40

20

10

30

0

40

10

time (mid

time (min)

30

40

30

40

time (mid

125

LlMR LPTAKE

SERUM DECAY

20

SERUM DECAY

-13 100

50

lactofenin

lactoferrin

0

0

10

20

time (min)

30

40

0

20

10

time (mid

30

40

0

10

20

time (mid

Fig. 1. Liver uptake and serum decay of chylomicrons, 5-min remnants and 30-min remnants. [3H]Vitamin-A-labeled chylomicrons (A, B; 7),

5-min remnants (C, D; A ) and 30-min remnants (E, F; 0)were injected into anaesthetized rats. Liver association (A, C, E) and serum decay (B. D, F) were determined with prior treatment of rats with GdCI3 (U),prior injection of lactoferrin (V)and a combined prior treatmcnt of rats with GdC13 and prior injection oflactoferrin ( 0 ) .Liver values are corrected for serum radioactivity. Values are means SE of 2- 5 rats. If not visible, error bars are covered by the symbol. lacto, lactoferrin.

the serum decay follows a characteristic biphasic pattern (Fig. 1 B). 30 rnin after injection, 92% of the injected dose is cleared from the circulation, with 84% recovered in the liver. Uptake by spleen and bone marrow account for 2.5% and 2.4%, respectively, of the injected dose. Injection of chylomicrons into the bloodstream of functionally hepatectomized rats leads to the formation of chylomicron remnants which accumulate in the blood of these animals. In this study, chylomicrons were allowed to circulate for 5 min or 30 min. Injection of 5-min remnants into the bloodstream of anaesthetized rats leads to liver uptake and serum decay values comparable with those from chylomicrons (Fig. 1C), while liver uptake of 30-min remnants proceeds much faster than that of chylomicrons (Fig. 1E). 5 min after administration, the liver uptake of chylomicrons, 5-min remnants and 30-min remnants is 26%, 29% and 67%, respectively, of the injected dose. Injection of GdC13 into rats leads to a total liver depletion of Kupffer cells 24 h after injection [29]. In our studies, GdCI, was used in vivo to identify the possible role of Kupffer cells in the liver uptake of chylomicron remnants. In GdCl,-inject-

ed rats, both liver uptake and the serum decay of chylomicrons, 5-min remnants and 30-min remnants were virtually uninfluenced (Fig. 1A - F). Injection of 70 mg/kg lactoferrin 1 min prior to administration of lipoproteins leads to a prolonged life-time of chylomicrons, 5-min remnants and 30-min remnants, as a consequence of a quantitatively important blockade of liver uptake. 30 min after injection, lactoferrin inhibited the liver uptake ofchylomicrons, 5-min remnants and 30-min remnants for 76%, 82% and 58%, respectively. Combined treatment of rats with GdC13 and injection of lactoferrin prior to the experiment results in liver uptake and serum decay values for chylomicrons and 5-min remnants which are comparable with prior lactoferrin injection alone (Fig. 1A-D). With 30-min remnants, liver uptake 2 min and 5 min after injection is significantly different (p < 0.05) from that for the single prior injection with lactoferrin. None of the above-mentioned treatments of GdC13 and/ or injection of lactoferrin resulted in a significantly different uptake from the control experiments of all three lipoproteins by spleen or bone marrow (see also Fig. 5).

779

L

PC

EC

KC

Fig. 2. In vivo association of chylomicrons and 30-min remnants with parenchymal, endothelial and Kupffer cells at different times after injection. [3H]Vitamin-A-labeledchylomicrons (A) and 30-min remnants (B) were injccted into anaesthetized rats. At 5 rnin (open bars), 15 min (hatched bars) or 30 min (crossed bars) after circulation, the liver was perfused at 8 “C. After 8 rnin of perfusion, total liver association of radioactivc lipoproteins was determined (L) and subsequently parenchymal (PC), endothelial (EC) and Kupffer (KC) cells were isolatcd by a low-temperature method. Values are means SEM of three rats.

Cellular distribution of chylomicrons and 3 0 4 1 1remnants

To analyse the liver cell types responsible for the liver uptake of chylomicrons (remnants) and 30-min remnants quantitatively, liver parenchymal, endothelial and Kupffer cells were isolated at different times after injection of radiolabeled lipoproteins by a low-temperature procedure (Fig. 2). For chylomicrons (remnants), it appears that at all times studied, parcnchyrnal cells are the major site for uptake. Within the non-parenchymal cells, mainly Kupffer cells do interact with the chylomicron remnants with a maximal contribution of 8.6% of the injected dose (1 3.5% of the total liver uptake). Initially, at 5 rnin of circulation, the ratio of parenchymal-cell uptake/Kupffer-cell uptake is somewhat lower than after I5 rnin and 30 min of circulation (4.9 compared to 6.7 and 7.4, respectively; Fig. 2A). For 30-min remnants, the relative contribution of the different cell types to the liver uptake is clearly different. Initially ( t = 5 rnin), uptake by parenchymal cells is lower than by Kupffer cells (a ratio of 0.55), while at latcr time points the relative cellular localization of the radioactivity shifts towards the parenchymal cells (ratios of 5.7 and 6.2 at 15 rnin and 30 rnin of circulation, respectively; Fig. 2 B).

L

PC

EC

KC

Fig. 3. Effect of lactoferrin on the in vivo association of chylomicrons and 30-min remnants with parenchymal, endothelial and Kupffer cells. [3H]Vitamin-A-labeledchylomicrons (A) and 30-min remnants (B) were injected into anacsthetized rats without (open bars) or with (hatched bars) prior injection of lactoferrin. 15 min after injection, the liver was perfused at 8°C. After 8 rnin of perfusion, total liver association of radioactive lipoproteins was determined (L) and subsequently parenchymal (PC), endothelial (EC) and Kupffer (KC) cells were isolated by a low-temperaturemethod. Values are means fSEM of three experiments. For both lipoproteins, values of parenchymal cells arc significantlydifferent plus or minus lactoferrin 0, < 0.05).

Lactoferrin is expected, based upon its structure, to specifically block uptake for which apoprotein E is the decissive determinant. We therefore investigated to what extent lactoferrin could discriminate between the different cell types. It appears that, both for chylomicrons and 30-min remnants, lactoferrin specifically blocks the uptake of these lipoproteins by parenchymal cells, leaving endothelial-cell uptake and Kupffer-cell uptake uninfluenced (Fig. 3 A and B). Lactoferrin inhibits uptake of chylomicrons (remnants) and 30-min remnants by parenchymal cells by 89.5% and 87.3%, respectively. Liver uptake and serum decay of chylomicron remnants which have been stored for 24 h In the previous section we utilized chylomicron remnants which were injected into recipient animals within 4 h of preparation. In order to investigate to what extent storage conditions may influence the cellular fate of chylomicron remnants, we analysed the effect of storage for 24 h at 2 0 T . Injection of 5-min remnants, which had been stored for 24 h at 20”C, into anaesthetized rats resulted in a very fast liver uptake and serum decay (Fig. 4 A and B). 5 rnin after

780 A

B

125

125

SERUM DECAY

LIVER LPTAKE

g

control

100

0 73

U

a, +J

$ .c

U lactoferrin

75

,OOt

75

I

41

.-

50

50

Lc

0

$

25

25 GdClg/lacto

0’ 0

1

10

30

20

0

40

0

20

10

time (rninl

30

40

30

40

time (mid

125 LIVER UPTAKE

loo{

SERUM DECAY GdCI3

lactoferrln

0

10

20

30

40

0

20

10

time (mid time (mid Fig.4. Liver uptake and serum decay of 5-min remnants and 30-min remnants after 24 h of storage a t 20OC. [3H]Vitamin-A-labeled 5-min remnants (A, B; A ) and 30-min remnants (C, D; 0),stored for 24 h a t 20”C, were injected into anaesthetized rats. Liver association (A, C) and serum decay (B, D) were determined with prior treatment of rats with GdCI3 (a),prior injection of lactoferrin ( V )and a combined prior treatment of rats with GdCI3 and prior injection of lactoferrin ( 0 ) .Liver values are corrected for serum radioactivity. Values are means f SE for 2- 5 rats. If not visible, error bars are covered by the symbol. lacto, lactoferrin

administration, 83% of the injected dose is recovered in the liver. compared to 29% for control experiments. Prior injection of lactoferrin had no effect on liver uptake. Prior treatment of rats with GdCI3 resulted in a greatly delayed liver uptake, while the serum decay of 5-min remnants stored for 24 h was also influenced, although to a lower extent, due to extrahepatic uptake. It was noticed that prior treatment with GdC13 resulted in a higher uptake of 5-min remnants stored for 24 h by spleen and bone marrow (19.3 1.5% and 4.0 & 0.45%, respectively of the injected dose, 30 min after administration; Fig. 5). Combined prior treatment ofrats with GdCI3 and prior injection of lactoferrin resulted in a more pronounced reduction of the liver uptake of 5-min remnants which had been stored for 24 h. Uptake by spleen and bone marrow under these conditions was similar as for GdCI3 alone (results not shown). Injection of 30-min remnants which had been stored for 24 h a t 20 C, did not influence the kinetics of liver uptake and serum decay when compared to freshly injected remnants. It appeared that 30-min remnants which had been stored for 24 h were taken up by the liver in a similar way as 30-min remnants which had not been stored (Fig. 4C and D). Prior

bone marrow

spleen

25

d

fta 5’

24H 3 0 24H 5‘

30‘

5’

24H 30’ 24H 5’

30

Fig. 5. The influence of storage at 20OC on the uptake of chylomicron remnants by spleen and bone marrow. [3H]Vitamin-A-labeled 5-min remnants ( Y ) , 30-min remnants (30’), 24-h stored 5-min remnants (24H-5’) and 24-h stored 30-min remnants (24H-30’) were injected into anaesthetized rats. 30 min after injection, uptake by spleen and bone marrow was determined with (hatched bars) or without (open bars) prior treatment of rats with GdCI,. Values are means i SE for 2- 5 rats.

781 treatment of rats with GdCl, or prior injection of lactoferrin led to an effect which was similar to that obtained with 30-min remnants which had not been stored. Uptake by spleen and bone marrow after prior treatment with GdCI3 was 6.80 1.00% and 3.50 0.83%, respectively (Fig. 5) and was similar to that of a combined prior treatment with GdC1, and prior injection with lactoferrin. Nature of the recognition sites for chylomicron remnants on parenchymal and Kupffer cells In order to determine the capacity and specificity of the recognition sites for chylomicron remnants on parenchymal and Kupffer cells, the cells were isolated and incubated in vitro with radiolabeled 30-min remnants. When the cells were incubated with increasing amounts of ['251]chylomicron remnants for 2 I1 at 4"C, with and without a 60-fold excess of unlabelcd lipoprotein, evidence for high-affinity sites on both cell types could be obtained (Fig. 6 A and B). For parenchymal cells, the maximum level of binding on a protein basis is 300 ng apolipoprotein/mg cell protein, while for Kupffer cells this value is 3.7-fold higher (1100 ng apolipoprotein/mg cell protein), with Kd values of 23 l g and 17 pg apolipoprotein/ml, respectively. It can be calculated that, assuming a size of 90 nm and a protein content of 9.0%, parenchymal cells contain approximately 9000 receptors for chylomicron remnants/cell, while this value is approximately 3200 receptors for chylomicron remnants/Kupffer cells (assuming a protein content of parenchymal cells and Kupffer cells of 1 mg protein/106 cells and 1 mg protein/1O7 cells, respectively). The calcium dependency of the binding of chylomicron remnants to parenchymal and Kupffer cells indicate that binding of chylomicron remnants to parenchymal cells is not dependent on calcium, while binding of chylomicron remnants to Kupffer cells is lower in the absence of calcium (Fig. 7A and B). In order to determine the specificity of the interaction of chylomicron remnants with Kupffer cells and parenchymal cells, competition studies were performed. Radiolabeled

chylomicron reinnants (5 pgjml) were incubated with various amounts of competitor for 10 min at 37°C (Fig. 8 A and C) or 2 h at 4°C (Fig. 8 B and D). With parenchymal cells, chylomicron remnants and B-VLDL were effective competitors, both at 4°C and at 37°C. LDL was essentially ineffective at both temperatures (Fig. 8 A and B). Binding of /?-VLDL to parenchymal cells was not dependent on calcium (results not shown). With Kupffer cells, competition of ['251]chylomicron remnants with unlabeled chylomicron remnants is very effective both at 4°C and at 37°C. In contrast to parenchymal cells, /?-VLDL is a less effective competitor. A 10-fold excess of unlabeled P-VLDL reduces the cell association of [Iz5I]chylomicron remnants with parenchymal cells by 85% (at 37 "C), while association with Kupffer cells is only reduced by 21 YO(at 37°C; Fig. 8C and D). As with parenchymal cells, LDL is ineffective as a competitor for the interaction of chylomicron remnants with Kupffer cells. The ability of lactoferrin to compete with ['2s1]chylomicron-remnant binding and/or cell-association to parenchymal and Kupffer cells was investigated in vitro in order to determine if similar structures as found in apoprotein E do interfere with the cell interaction. Increasing amounts of unlabeled lactoferrin reduce the association at 37 'C of [1251]chylomicron remnants with parenchymal cells by 67%, while association to Kupffer cells is hardly influenced (maximally 20%; Fig. 9A). Also, at 4°C binding of ['251]chylomicron remnants to parenchymal cells is reduced by 70%, while binding to Kupffer cells is, again, hardly influenced (maximally 13%; Fig.9B). To check the involvement of the arginine residues of lactoferrin in the competition, lactoferrin was modified by cyclohexanedione, which selectively reduces arginine residues of proteins. As can be seen in Fig. 9C, lactoferrin modified by cyclohexanedione was unable to compete with [lZ5I]chylomicron-remnant association to parenchymal cells and at the highest concentration of cyclohexanedione-treated lactoferrin, a maximal inhibition of 9% is observed.

B

A

3000 h

S .-

a,

2500 .-0

2000 .& = a, _oo Lc

1500

O &

b' o

I

0

5

, .

0

25

50

75

100

[Remnants] @g/ml)

125

0

25

50

75

100

125

[Remnants] (pg/rnl)

Fig.6. Binding of chylomicron remnants to rat parenchymal and Kupffer cells as a function of the chylomicron remnant concentrations. Rat parenchymal (A) and Kupffcr (B) cells were incubated for 2 h at 4°C with increasing amounts of ['251]chylomicronremnants in the absence ( A ) or presence (0)of 300 pg/ml unlabeled chylomicron remnants. The specific binding (V)is calculated by subtracting the binding with excess unlabelcd chylomicron remnants from the tolal binding. Values arc means of duplicate experiments.

782

B

A

[Mg-EGTAIM)

[Mg-EGT A h M )

[Ca2+l(mM)

[Ca2+l(mM)

Fig. 7. The effect of varying Ca2+ or EGTA concentrations on the binding of [1251]chylomicronremnants to rat parenchymal and Kupffer cells. Rat parenchymal (A) and Kupffer (B) cells were incubated for 2 h at 4°C with 5 pg/ml [1251]chylomicronremnants and with the indicated concentrations of C a z C or EGTA. The binding at 2 mM C a z + is taken as the 100% value. Values are means of two experiments. The 100% value for binding to parenchymal cells is 52.5 ng apolipoprotein/mg cell protein and for Kupffer cells 318.6 ng apolipoprotein/mg cell protein.

B parenchymal cells

120

IT

120

'c

0

0

s

s I

0'

50

120

1

K u p f f e r cells

(d

0

ml)

D

100 150 200 250 300

140 I

120

+

'c

0

0

s

s -

50

competitor

1

I

0

100 150 200 250 300

competitor C

1

parenchymal cells

T

'c

0

140 I

(wghl) 1

K u p f f e r cells

0 0

50

100 150 200 250 300

competitor (wglml) competitor (yg/ml) Fig. 8. Ability of unlabeled lipoproteins to compete with ~'251]chylomicronremnant cell association or binding to parenchymal and Kupffer cells at 37OC and 4OC. Rat parenchymal (A, €3) and Kupffer (C, D) cells were incubated with 5 pg/ml ['Z51]chylomicronremnants for 10 min at 37 C (A. C) or 2 h at 4 'C (B, D). Binding or cell-associations were determined in the absence or presence of the indicated amount of unlabeled chylomicron remnants (a), j-VLDL (A)or LDL ( 0 ) The . binding or cell association is expressed as the percentage of radioactivity obtained in the absence of competitor. Values are means -t SEM of three experiments. The 100% values for 1251-labeled-chylomicron-remnants association or binding to parenchymal cells are 93.3 f 12.6 ng apolipoprotein/mg cell protein (37'C) and 42.1 f 5.5 ng apolipoprotein/mg cell protein (4°C). For Kupffer cells, these values are 435.1 +_ 72.0 ng apolipoprotein/mg cell protein (37'C) and 332.5 f-40.6 ng apolipoprotein! mg cell protein (4'C). chylo, chylomicron.

783 125

c

.0

.$ '--i

1001

w

.-m

0

::

'

' j : 50

$\

LC.

C

a\p

25 -

0

s -

I

25-

0

DISCUSSION Chylomicrons are converted to chylomicron remnants by the action of endothelial-cell-bound lipoprotein lipase and exchange of surface material and apolipoproteins with HDL 111. The subsequent fate of chylomicron remnants is under discussion. In the rabbit, it is was reported that, in addition to the liver, 18% of the injected chylomicron remnants were taken up by bone marrow. In marmosets and guinea-pigs, bone marrow and adipose tissue, respectively, were found to be important extrahepatic sites for uptake. In rats, however, we found a nearly quantitative uptake of chylomicron remnants by the liver, provided that chylomicron remnants were utilized within 4 h of isolation. Storage of chylomicron remnants for 24 h may, however, lead to an increased spleen and bone-marrow uptake, especially when shortly processed ( 5 min) chylomicron remnants are utilized. It is unclear at the moment if the extrahepatic uptake sites, similar to these found in rabbits or marmosets, utilize similar recognition markers as the liver site, but lactoferrin might be used to investigate this interesting question. In rats, the parenchymal cells were reported to be mainly responsible for the liver uptake of chylomicron remnants [271. However, the contribution of non-parenchymal cells was considered to be significant with reported contributions which varied over 6 - 35% [2 - 71. The present investigations were aimed to characterize the structural requirements for the relative importance of recognition of chylomicron remnants by parenchymal and non-parenchymal cell types and to define the cell type within the non-parenchymal cells which is responsible for the reported uptake. The data indicate that, within the various non-parenchymal cell types, Kupffer cells specifically may interact with chylomicron remnants. It appears that the percentage of the injected dose of [3H]retinol ester-labeled chylomicrons recovered in Kupffer cells does not exceed 8.6%, while the parenchymal cells are responsible for uptake of 66% of the injected dose 15 rnin after injection. It must be realised

i\L

w-

1

o

40-

s

-

1

20 0

that only 2.5% of the total liver protein is present in Kupffer cells, so that the uptake/mg cell protein for Kupffer cells is still 4.8-fold higher than with parenchymal cells. Using preformed chylomicron remnants, the relative contribution of Kupffer cells is much higher, certainly shortly after injection. Under those conditions, 5-30 rnin after injection, the percent uptake by Kupffer cells shifts from 37.8 f 1.5% to 1 1 .O If: 2.2%. When Kupffer cells, isolated 5 min after injection, are incubated in vitro at 37 "C, it appers that retinol esters and triacylglycerols are rapidly hydrolysed and free retinol and fatty acids are secreted from the cells (unpublished results). It might therefore be suggested that the redistribution of the radiolabel in vivo, 5 - 30 min after injection of chylomicron remnants, from Kupffer cells to parenchymal cells, is the consequence of transport of free retinol from Kupffer cells to parenchymal cells. A similar pathway was reported earlier for endothelial cells after uptake of acetylated LDL [45]. In vifro,the recognition sites on parenchymal and Kupffer cells, responsible for uptake of chylomicron remnants, were characterized. High-affinity recognition sites with Kd values of 23 ng apolipoprotein/mg and 17 ng apolipoprotein/ml, respectively, were found on parenchymal and Kupffer cells. The recognition sites on Kupffer cells differ from the recognitions site on parenchymal cells in serveral respects, as follows. (a) The maximum level of binding of chylomicron remnants to Kupffer cells/mg cell protein is 3.7-fold higher than to parenchymal cells. (b) Chylomicron-remnant binding to Kupffer cells is partially dependent on calcium, while binding to parenchymal cells is not. (c) 8-VLDL is a less effective competitor for chylomicron-remnants binding to Kupffer cells than to parenchymal cells. (d) In vivo, as well as ~ P J vitro, lactoferrin reduces binding of chylomicron remnants to parenchymal cells, leaving Kupffer-cell uptake virtually unaffected. Cyclohexanedione modification of Pactoferrin destroys the ability of lactoferrin to reduce chylomicron-remnant binding, indicating that this reduction is a specific effect of the arginine residues within the protein itself.

784 The properties of chylomicron-remnant recognition by parenchymal cells are consistent with apoprotein E being the determinant for recognition. The distinct properties of the Kupffer-cell recognition site thus do point to a recognition independent of apoprotein E. The Kupffer-cell recognition site might function as an uptake site for (slightly) modified chylomicron remnants, since we show here that 5-min remnants stored for 24 h at 20°C are avidly taken up by Kupffer cells. Modification of remnants is not due to lipid (per)oxidation (as no significant increase in thiobarbituric-acid reactivities is observed). Our results indicate that during formation of chylomicronremnants, circulation (and thus processing) time is an important factor in the sensitivity of this lipoprotein to conversion into a form readily recognized by the liver macrophages. Chylomicron remnants, circulating for only a short period of time, are specifically sensitive to modification which may be related to the labile structure of the surface material of chylomicron remnants under these conditions. In patients with a defect in chylomicron-remnant processing (e. g. a defect in lipoprotein lipase activity or an apoprotein CII defect), the chylomicrons might be comparable to our experimental conditions and Kupffer celllmacrophage uptake might be triggered. Kupffer-cell uptake and processing of chylomicron remnants, coupled to the transport of the products to the parenchymal cells, may then function as a protection mechanism against extrahepatic macrophage accumulation.

REFERENCES 1. Redgrave, T. G. (1983) Int. Rev. Physiol. 28, 103 - 130. 2. Blomhoff, R., Helgerud, P., Rasmussen, M., Berg, T. & Norum, K. R. (1982) Proc. Natl Acad. Sci. USA 79, 7326-7330. 3. Blomhoff, R., Eskild, W.. Kindberg, G. M., Prydz, K. & Berg, T. (1985) J . Biol. Chem. 260, 13566-13570. 4. Van Dijk, M. C. M., Ziere, G . J., Boers, W., Linthorst, C., Bijsterbosch, M. K. & v a n Berkel, Th. J. C. (1991) Biochem. J . 279. 863 - 870. 5. Nilsson. A. & Zilversmit, D. B. (1971) Biochim. Biophys. Aria 248, 137- 142. 6. Oroot, P. H. E.. van Bcrkel. Th. J. C. & van Tol. A. 11981) Metabolism 30, 792 - 797. I. Lipiello, P. M., Dijkstra. J., Van Galen, M., Scherphof, G. & Waite, B. M. (1981) Proc. Nut1 Acad. Sci. USA 14,7454-7460. 8. Redgrave, T. G. (1970) J . Ciin. Invest. 49, 465-471. 9. Melchior, G. W.. Mahley, R. A . & Buckhold, D. K. (1981) J . Lipid Rex. 22, 598 - 609. 10. Hussain. M. M., Mahley, R. W., Boyles, J. K., Lindquist, P. A., Brecht, W. J. & Innerarity. T. L. (1989) J . Biol. Chem. 264, 17931 -17938. 11. Hussain. M. M., Mahley, R. W., Boyles, J. K., Fainaru, M., Brecht, W. J. & Lindquist, P. A. (1989) J . Biol. Chem. 264, 9571 -9582. 12. Hoeg, J. M., DeMosky, S. J. Jr, Gregg, R. E., Schaefer, E. J . & Brewer, H. B. Jr (1985) Scienre 227, 759-761. 13. Kita, T., Goldstein. J. L., Brown, M. S., Watanabe, Y., Hornick, C. A. & Havel, R. J . (1982) Proc. Nail Acad. Sci. U S A 79, 3623 - 3627. 14. Brown, M. S. & Goldstein, J. L. (1986) Science 232, 34-47. \

,

15. Arbeeny, C. M. & RiGci, V. A. (1984) J . B i d . Chern. 250, 96629666. 16. Krempler, F., Kostner, G . M., Friedl, W., Paulweber, B., Bauer, H. & Sandhofer, F. (1986) J . Clin. Invest. 80, 401 -408. 17. Cooper, A. D., Erickson, S. K., Nutik, R. & Shrewsbury, M.A. (1982) J . Lipid.Ras. 23, 42 - 52. 18. Hui, D. Y., Innerarity, T. L. & Mahley, R. W. (1981) J . Biol. Chem. 256, 5645 - 5655. 19. Hui, D. Y.. Brecht, W. J., Hall, E. A,, Friedman, G., Innerarity, T. L. & Mahley. R. W. (1986) J . B i d . Chem. 261.4256-4267. 20. Beisiegel, U., Weber, W., Havinga, J . R., Ihrkc, G., Hui, D. Y., Wernette-Hammond, M. E., Turck, C. W., Innerarity, T. L. & Mahley, R. W. (1988) Arteriosclerosis 8, 288-297. 21. H e n , J., Hamann, U., Rogne, S., Myklebost, O., Gausepohl, H . & Stanley, K. K. (1988) EMBO J . 7,4119-4127. 22. Kowal, R. C., Herx, J., Goldstein, J. L., Esser, V. & Brown, M. S. (1989) Proc. Nut1 Acad. Sci. U S A 86, 5810-5814. 23. Kowal, R. C., Herz, J., Weisgraber, K. H., Mahley, R. W., Brown, M. S. & Goldstein, J. L. (1990) J . B i d . Chern. 265, 10771 10 779. 24. Lund, H., Takahashi, K., Hamilton, R. L. & Havel, R. J. (1989) Proc. Nail Acud. Sci. USA 86,9318-9322. 25. Beisiegel, U., Weber, W., Ihrke, G., Herz, J. & Stanley, K. K. (1989) Nature 341, 162- 164. 26. Strickland, D. K., Ashcom, J. D., Williams, S., Burgess, W. H., Migliorini, M. & Argraves, W. S. (1990) f. Biol. Chem. 265, 17401 - 17404. 27. Kristensen, T., Moestrup, S., Gliemann, J., Bendtsen, L., Sand, 0. & Sottrup-Jensen, L. (1990) FEBSLett. 276, 151 -155. 28. Huettinger, M., Rctzek, H., Eder, M. & Goldenberg, H. (1988) Clin. Biochem. 21, 81-92. 29. Koudstaal, J., Dijkhuis, F. W. J. & Hardonk, M. J. (1991) in Cells ufrliehepaticsinii.soid(Wisse, E., Knook, D. L. & McCuskcy, R . S . , eds) vol. 3, pp. 87-91, The Kupffer Cell Foundation. 30. Redgrave, T. G. (1970) J . Clin. Invest. 49,465-471. 31. Mjm, 0. D., Faergeman, O., Hamilton, R. L. & Havel, R. J . (1975) J . Clin. Inves!. 56, 603-615. 32. Redgrave, T. G., Roberts, D. C. K. & West, C. E. (1975) Anal. Biochem. 65, 42 -49, 33. Harkes, L., van Duijne, A . & van Berkel, Th. J. C. (1989) Eur. J . Biochem. 180,241 -248. 34. Van Tol, A,, Van Gent, T., Van 't Hooft, F. M. & Vlaspolder, F. (1978) Atherosclerosis 29,449-457. 35. Mc Farlane, A. S. (1958) Nature 182, 53. 36. Bijsterbosch, M. K., Ziere, G. J. & van Berkcl, Th. J. C. (1989) Mol. Phnrmacol. 36,484 -489. 37. Caster, W. O., Simon, A. B. & Armstrong, W. D. (1955) Am. J . Physiol. 183, 317-321. 38. Van Bcrkel, Th. J. C., Dekker, C. J., Kruijt, J. K . & v a n Eijk, H. G. (1987) Biochem. J. 243,715-722. 39. Blouin, A,, Bolender, R. D. & Weibel, E. R. (1977) J . CeN. Bid. 72,441 -455. 40. Nagelkerkc, J. F., Barto, K . P. & Van Berkel, Th. J . C. (1983) J . Biol. Chem. 258, 12221 - 12227. 41. Lowry, 0. H., Rosebrough, R. J., Farr, A. L. & Randall, R. J. (1951) J . Biol. Chem. 193, 265-215. 42. Steinbrecher, U. P.,Witzum, J. L., Parthasarathy, S. & Steinberg, D. (1987) Arteriosclerosis 5 , 135-143. 43. Patthy, L. & Smith, E. L. (1975) J . Biol. Chem. 250, 557-564. 44. Mahley, R. W., Innerarity, T. L., Pitas, R. E., Weisgraber, K. H., Brown, J. H. & Gross, E. (1977) J . Biol. Chem. 252, 72797287. 45. Bakkeren, H. F., Kuipers, F., Vonk, R. J. & Van Berkel, Th. J. C. (1 990) Biochcm. J . 268, 685 - 691.

Characterization of the chylomicron-remnant-recognition sites on parenchymal and Kupffer cells of rat liver. Selective inhibition of parenchymal cell recognition by lactoferrin.

Upon injection of chylomicrons into rats, chylomicron remnants are predominantly taken up by parenchymal cells, with a limited contribution (8.6% of t...
991KB Sizes 0 Downloads 0 Views