The circulatory survival of FSA and α-mannosidase was determined after intravenous administration of trace amounts of ¹²⁵I-tyramine cellobiose-FSA (¹²⁵I-TC-FSA) and ¹²⁵I-α-mannosidase. Decay plots indicated efficient clearance of either probe, with 50% of injected dose being eliminated from the blood during 2–5 min (Fig. 1). The liver was the main site of uptake (Fig. 2), while a surprising finding was uptake in the lungs. Blood radioactivity after 15–20 min was 15–20% of injected dose. This equals the amount of unbound ¹²⁵I after gel filtration through a PD-10 column of a sample of the intravenously administered ligands.

Intravenuosly administered TRITC-monodisperse polymer particles (MDPP) for identification of phagocytosing KC accumulated mainly periportally in liver acini (Figs. 3A, 3B, 3C). Immunoelectron microscopy of liver sections that had been reacted with anti-TRITC-antibodies and protein A-gold revealed the presence of gold particles along the periphery of the surface of the particles, allowing a reliable identification and intracellular location of TRITC-MDPP (Figs. 4A, 4B). In contrast to these particles, FITC-FSA was taken up exclusively in LSEC-like cells lining the liver sinusoids (Fig. 3B). To distinguish LSEC from stellate cells, double immunolabeling was performed to visualize FITC-FSA and desmin in transmission electron microscopy. FITC-FSA and desmin were observed in distinct cell types along the sinusoidal lining (Fig. 5). FITC-FSA was associated with organelles judged as lysosomes of LSEC.

The number of non-parenchymal cells (NPC) obtained per liver following collagenase dispersion and isopycnic density separation in iodixanol was 280 × 10⁷ (range 50–890 × 10⁷) (weight of liver 237.1 g (43.6)) with a viability of 95.4% (2.5) as judged by trypan blue exclusion (Table 1). The corresponding figures for hepatocytes were 1880 × 10⁷ (1110 × 10⁷) and 94.1% (2.2). The cells obtained after iodixanol separation were subjected to centrifugal elutriation and collected in 4 fractions. The corresponding recoveries expressed as number of NPC and percentages of total are displayed in Table 2.

Cells, seeded on fibronectin-coated substrate, obtained from the elutriation fractions yielded LSEC cultures of varying purity (Table 3). We used in vivo (Fig. 6A) or in vitro administered FITC-FSA as a specific LSEC marker, positive reaction with anti-desmin antibodies as a specific marker of stellate cells (Fig. 7A), and a specific anti-pig macrophage antibody (Fig. 7B) or phagocytosis of TRITC-MDPP (Fig. 6B) as KC specific markers. Using these criteria, cultures resulting from elutriation fraction 1 were shown to contain 63.9% stellate cells; cultures established from fraction 2 contained 80.4% LSEC, and fractions 3 and 4 contained 66.2% and 61.0% LSEC. Cells that reacted with anti-pig-macrophage antibodies or phagocytosed TRITC-MDPP contained no FITC-FSA. Stellate cells were distinguished by immunolabeling with anti-desmin antibodies or by their content of characteristic autofluorescence from vitamin A droplets when irradiated with light of 328 nm of wavelength (12) (Fig. 6C).

The specificity of endocytosis of ¹²⁵I-FSA and ¹²⁵I-asialo-orusomucoid protein (ASOR) in cultured LSEC and hepatocytes was studied by attempting to inhibit the uptake of trace amounts of radiolabeled ligands using excess amounts of unlabeled ligands. Incubation of LSEC cultures with ¹²⁵I-FSA in the presence of excess amounts of unlabeled FSA (100 mg·mL^-1) resulted in a 90% inhibition of uptake (Fig. 8). The presence of galactose (50 mmol·L^-1) did not inhibit endocytosis of ¹²⁵I-FSA by LSEC. Incubation of hepatocytes with ¹²⁵I-ASOR in the presence of excess amounts of galactose (50 mmol·L^-1) inhibited uptake by 85%. Unlabeled FSA did not inhibit endocytosis of ¹²⁵I-ASOR by hepatocytes (Fig. 8).

First we studied the circulatory survival and anatomical distribution of FSA, a frequently used test ligand for the LSEC scavenger receptor in rat 13, and α-mannosidase, a physiological ligand for the mannose receptor of rat LSEC 14. Studies in the rat and other vertebrates have shown that ¹²⁵I-FSA is degraded very rapidly after uptake, resulting in rapid escape of radiotracer from the site of uptake. For this reason, FSA was labeled with ¹²⁵I-TC, which is trapped in the lysosomes at the cellular site of uptake, thus preventing ¹²⁵I escape from the uptake site 15. Previous studies in the rat and other vertebrates showed that α-mannosidase, after its rapid uptake by the mannose receptor, accumulates within lysosomes and is reused for several hours before being degraded 5. Therefore, α-mannosidase was labeled with ¹²⁵I in a direct, conventional manner. Both ¹²⁵I-TC-FSA and ¹²⁵I-α-mannosidase were rapidly eliminated from the circulation, with 50% of the ligands being removed during the first 2–5 min after intravenous administration. This rapid removal suggested a very efficient uptake mechanism. Monitoring of radioactivity in the organs showed that the liver contained 53% (FSA) and 62% (α-mannosidase) of injected dose, suggesting that a cell type(s) in liver was responsible for clearance via the scavenger and mannose receptors. Surprisingly, as much as 26% FSA and 18% α-mannosidase were recovered in lungs. This is clearly different than in the rat, where uptake in the lungs of these and other soluble macromolecular waste products have not been observed 3. A recent report 16 showed that ligands for studies of reticuloendothelial function were taken up in both lung and liver of pig, similarly to what we found using α-mannosidase and FSA. It was concluded from that study that ¹⁹⁸Au colloidal particles and iron oxide particles were taken up in pulmonary intravascular macrophages. The possibility that these ligands might have been taken up by scavenger endothelial cells was not mentioned in that paper.

To determine the role of different sinusoidal cells in the clearance function of pig liver, the cellular site of uptake of FITC-FSA was compared with that of TRITC-MDPP (a functional marker of phagocytosing KC), and immunoreactive desmin (a marker of stellate cells). Since light microscopy does not allow a clear distinction between particles that are truly internalized and those that are associated with the cell surface, liver tissue was prepared for electron microscopy. To enable a distinction between vitamin A-containing lipid droplets in stellate cells and internalized MDPP in KC, sections were first incubated with anti-TRITC-antibodies, then with protein A-gold. Observations of these sections revealed gold staining along the surface of the MDPP particles, corresponding to the surface localization of TRITC. Double immunolabeling showed that FITC-FSA (5 nm gold) was always associated with endothelial like lining cells that neither took up MDPP nor contained desmin (10 nm gold), indicating that the hepatic uptake in vivo of FSA was exclusively in LSEC, similar to what has been found in the rat 13.

To allow a more detailed study of the tentative scavenger function of pig LSEC, we developed a protocol for isolation of sinusoidal cells. The protocol was modified as compared to rat 17 and mouse liver. According to the literature, rat and mouse liver sinusoidal cells can be isolated in high yield and purity using isopycnic separation. We found that this method was insufficient to isolate such cells from pig due to the high number of desmin-positive cells; therefore, we included centrifugal elutriation to separate the cells according to size. Using collagenase perfusion through the portal vein, followed by differential centrifugation, isopycnic centrifugation, and centrifugal elutriation we obtained 4 fractions, of which fraction 2, containing 62 × 10⁷ (12 × 10⁷) sinusoidal cells, gave monolayer cultures on fibronectin that were 80% pure in LSEC, as judged by uptake of intravenously administered FITC-FSA. The FITC-FSA positive LSEC did not react with anti-desmin or anti-pig macrophage antibodies. Conversely, only cells that took up TRITC-MDPP reacted with the anti-pig macrophage antibodies (about 11% of the plated cells).

Specific endocytosis of FITC-FSA was used successfully for vital-staining of LSEC either in vivo by intravenous injection, or by administration to LSEC-cultures in vitro. Since FITC-FSA is a highly stable molecule that can be easily and inexpensively prepared without special equipment, and stored for years in the freezer or refrigerator, we find this probe superior to other marker molecules launched as LSEC-specific functional markers (e.g., fluorescently-labeled acetylated LDL 18) that are unstable and require expensive equipment for preparation. Although KC could be identified in vivo by their specific ability to phagocytose MDPP, this marker was disadvantageous for in vitro KC identification for the following two reasons: 1) KC that had phagocytosed intravenously administered MDPP exhibited significantly altered sedimentation properties on subseqent centrifugal elutriation; and 2) in vitro cultured KC did not phagocytose MDPP under the conditions used. These difficulties forced us to look for alternative ways of identifying the KC. Fortunately, an anti-pig-macrophage antibody kindly provided by Dr. A. Berndt, Germany, gave a specific staining of KC, but not of other types of liver cells; therefore this method was preferred as a method for identification of KC. The identification of stellate cells was at first based on demonstration of vitamin A-positive lipid droplets, that could be easily detected by autofluorescence when illuminated with light of 328 nm of wavelength. However, since vitamin A appeared to be present to a variable extent in stellate cells, we decided to employ immunostaining for desmin with a monoclonal antibody.

The observation that uptake of ¹²⁵I-FSA by cultured LSEC could be inhibited by excess amounts of FSA, but not by galactose, indicates that the cells express the same functional type of scavenger receptor as has been reported for rat LSEC 13. The finding that FITC-FSA was taken up only in LSEC in vivo and in vitro is further evidence that the porcine and rat LSEC scavenger receptors are functionally similar. The finding that LSEC isolated and cultured according to the presently developed protocol, perform a highly active scavenger-receptor-mediated endocytosis proves that pig LSEC prepared according to our protocol have preserved their in vivo scavenger function.

Although several authors have reported the presence of LSEC in cultures of porcine liver cells, methods of identification of LSEC were unable to distinguish convincingly LSEC from other types of liver cells. In a study on isolation of NPC from piglets (1–15 days of age), Caperna et al. 19 used absence of uptake of 5 μm latex beads as a characteristic of LSEC. However, our observation that KC only rarely phagocytose particles in vitro, along with the fact that other liver cell types (i.e., stellate cells and lymphocytes) are unable to perform phagocytosis, suggest that absence of phagocytosis is not a valid way of distinguishing LSEC from other liver cells. In other studies, positive immune staining of von Willebrand factor has been used to identify pig LSEC 1011. However, the fact that this marker is widely used to identify vascular endothelial cells in general 20, together with other reports claiming that LSEC are devoid, or stain only weakly for this antigen, suggest that von Willebrand factor should not be used as a specific LSEC marker. Prostacyclin PGI2 generation has been used as a LSEC-specific marker 21. However, PGI2 is a general endothelial marker, and can not be used as a LSEC-specific marker. Fluorescently labeled acetylated LDL has been used in much the same way as we used FITC-FSA as a vital stain to demonstrate the scavenger function of LSEC 22. Nevertheless, it has been shown that this probe, in contrast to FITC-FSA, is taken up in brain microvessel endothelial cells 23, as well as in peripheral macrophages 24).), suggesting that FITC-FSA is superior as a LSEC specific marker.

Monoclonal mouse anti-human desmin, clone D33, was from DAKO A/S, Denmark. Monoclonal goat anti-mouse IgG, TRITC-conjugate, was from Zymed, CA. Monoclonal rabbit anti-FITC was from DAKO A/S, Denmark. Monoclonal rabbit anti-TRITC was from Molecular Probes Inc., OR. Two mouse monoclonal antibodies (clones 2G6 and 2B10) against porcine macrophages were kindly provided by Dr. A. Berndt, Institute of Pathology, Friedrich Schiller University, Jena, Germany 25. FSA was prepared as described 26. ASOR was kindly provided by Dr. Trond Berg, University of Oslo, Norway. Bovine α-mannosidase was donated by Dr. O.K. Tollersrud, University of Tromsø, Norway.

Proteins were labeled with ¹²⁵I (carrier free Na¹²⁵I from Institute of Energiteknikk, Norway) either by a direct reaction employing Iodogen (Pierce, Rockford, IL) 27, or by conjugating the protein with ¹²⁵I-labeled TC 15. The iodinated proteins, with ¹²⁵I attached to aromatic amino acids, were separated from free ¹²⁵I on a PD-10 column (Sephadex G-25, Pharmacia, Sweden) equilibrated and eluted with PBS.

FSA (0.8 mg·mL^-1) was incubated with FITC (Sigma-Aldrich, Norway) in sodium carbonate buffer (0.5 mol·L^-1, pH 9.5) in a protein/dye weight ratio of 1:1 at 4°C overnight. Unreacted dye was removed by gel filtration through a PD-10 column equilibrated and eluted with PBS without Ca²⁺ or Mg²⁺.

Approximately 8.0 × 10¹⁰ MDPP, with a diameter of 2.0 μm (SINTEF, University of Trondheim, Norway) were incubated overnight in 1 mg·mL^-1 TRITC (ICN Biomedicals Inc., CA) in sodium carbonate buffer (0.1 mol·L^-1, pH 9.5) at 4°C. The TRITC-MDPP were washed three times in 70% ethanol, and finally in PBS to remove unbound TRITC.

Serum half-life and organ distribution of intravenously administered radiolabeled FSA and α-mannosidase were determined as described 28. Briefly, approximately 100 × 10⁶ cpm ¹²⁵I-TC-FSA or ¹²⁵I-α-mannosidase were injected into the left jugular vein. Blood sampling was started immediately by collecting 2 mL blood samples from an intravenous catheter positioned in the right jugular vein. The organs were taken out after one hour, weighed and measured in a γ-counter (Cobra II, Packard Instrument Co., Inc., CT).

The experimental protocol was approved by the local steering committee of the Norwegian Experimental Animal Board. All animals received care according to the criteria outlined in the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86–23 revised 1985). Castrated male pigs weighing 7.6 kg (0.6) (Sus scrofa domesticus, Norwegian strain) were fasted 18 hours with free access to water. Intramuscular ketamine (75 mg·kg^-1) (Parke-Davis, Sweden) and atropine sulfate 1 mg (Nycomed Pharma, Norway) were used as premedication. The pigs were tracheostomized, intubated, and ventilated on a volume controlled ventilator (Servo 900, Elema-Schönander, Sweden). Anesthesia was maintained with continuous i.v. infusions of sodium pentobarbital (4 mg·kg^-1·min^-1) (Abbott Scandinavia, Sweden) and fentanyl (0.05 mg·kg^-1·min^-1) (Alpharma, Norway). A midline laparotomy was performed. The portal vein was mobilized and the portal bloodflow was measured with a flowprobe (CardioMed, Norway). Sodium heparin (3000 IU) (Leo, Denmark) was administered intravenously. FITC-FSA (40 mg) was administered intravenously via the left jugular vein, while some pigs received TRITC-MDPP injected via the portal vein as two boluses 30 and 5 minutes prior to perfusion. The hepatoduodenal ligament was ligated, and the portal vein was cannulated. Ca²⁺-free HEPES 29 solution at 37°C was perfused in vivo with a roller pump (Gambro, AK-10, Sweden) at the same rate as the portal blood flow. The liver was weighed, thereafter perfused ex vivo with a Ca²⁺-containing HEPES-solution with 0.02% Collagenase P (Roche Diagnostics, Norway) in a basin at 37°C.

The digested liver was transferred to a different basin containing RPMI 1640 medium with 1% BSA (4°C). After removing the stroma, the suspension was subsequently filtered through 1000 and 500 μm nylon mesh. The resulting cell suspension was subjected to centrifugation at 50 g to selectively sediment hepatocytes from NPC. The pellet of hepatocytes was resuspended, and subsequently washed two more times at 50 g. Cell viability was estimated from the percentage of cells which excluded 0.4% trypan blue.

The NPC in the first and second supernatants from the low-speed centrifugations were pelleted by high-speed centrifugation (1300 g) followed by resuspension in small volume before isopycnic centrifugation in a 12% iodixanol gradient (Nycomed, Norway) to generate a band of further enriched NPC. This fraction was washed in PBS and resuspended in HBSS, supplied with 4 mmol·L^-1 EDTA, 1% FCS and 0.3% BSA, and loaded into the mixing chamber over 1 min. The NPC suspensions were fractionated in a J-21-B centrifuge fitted with a JE-6 elutriator rotor (Beckmann Instruments, Inc., CA). A pre-enrichment step was performed to further remove hepatocytes and cell aggregates with a pump flow rate of 30 mL·min^-1 and rotorspeed at 1400 rpm. Four consecutive fractions, corresponding to pump flow rates of 18.5, 32, 37 and 45 mL·min^-1, were collected from the rotor which was spun at a constant speed of 2500 rpm, at 15°C. The volume of effluent at each flow rate was 150, 250, 250 and 250 mL.

Monolayer cultures of hepatocytes and cells from each elutriation fraction were established on fibronectin-coated glass-coverslips inserted in 2 cm² culture wells and maintained in serum-free RPMI 1640 medium in a CO₂-incubator at 37°C. Cultures were fixed in 4% paraformaldehyde in PBS for 1 hour, after attachment and spreading.

To examine the distribution of FITC-FSA and TRITC-MDPP, small sections were taken at the periphery of the liver lobules during the washing perfusion, fixed in 4% paraformaldehyde, prepared and sectioned for fluorescence microscopy.

Sections and fixed cultures were embedded in antifade mounting medium (Dako Corp., CA). Examination of the mounted coverslips was performed in an Axiophot photomicroscope equipped with phase-contrast, fluorescence optics and a filter for excitation (wavelength 328 nm) of vitamin A (Zeiss Axiophot, Germany). Pictures were recorded on Kodak Ektachrome P1600X ASA (Kodak, Japan).

Whole livers were perfusion-fixed with McDowell's fixative (4% paraformaldehyde, 1% glutaraldehyde in PBS). Specimens for immunotransmission electron microscopy were prepared as described 30. Briefly, small cubes were cut out of the liver tissue, and incubated in 2.3 mol·L^-1 sucrose overnight before immersion in liquid nitrogen on aluminum pins. Ultrathin sections were cut using a diamond knife (Drukker international, The Netherlands). Before immunolabeling, sections were retrieved in sucrose and mounted on carbon coated grids. Sections were blocked for 15 minutes in 1% cold water fish skin gelatin (Sigma-Aldrich, Norway), followed by incubation with anti-TRITC (diluted in cold water fish skin gelatin), washing in PBS and incubation with protein A-gold (University of Utrecht, The Netherlands) 31 diluted in cold water fish skin gelatin for 15 minutes. Double labeling was performed in a sequential manner, using a fixative block (1% glutaraldehyde) between the first and second marker pair. The sections were then washed in PBS followed by washing in distilled water and dried in 1.8% methylcellulose and 0.3% uranyl acetate, and examined in a JEOL JEM 1010 transmission electron microscope (Tokyo, Japan) at 80 kV. Micrographs were recorded on Kodak electron microscope film, no. 4489 (Kodak, Tokyo, Japan).

LSEC and hepatocyte cultures, established in 2 cm² wells and maintained in serum-free RPMI 1640 medium, were washed and supplied with fresh medium containing 1% serum albumin and labeled proteins, ¹²⁵I-FSA or ¹²⁵I-ASOR (20 000 cpm per well). Specificity of endocytosis was studied by incubation with excess amounts of unlabeled FSA (100 μg·mL^-1) or galactose (50 mmol·L^-1). Incubations, carried out for 2 hours at 37°C to measure endocytosis, were terminated by transferring the media, along with one wash, to tubes containing 20% TCA, which precipitates only intact protein. The extent of degradation was determined by measuring the radioactivity in the pellet and supernatant that was obtained after centrifugation of the media. Cell-associated ligand was quantified by measuring the amount of radioactivity released by solubilizing the cultures in 1% (w/v) SDS. Radioactivity was measured in a γ-counter (Cobra II, Packard Instrument Co., Inc., Meridien, CT).

The values are expressed as: mean (standard deviation) unless otherwise noted.