Purification of Hepatocyte Couplets by Centrifugal Elutriation JOANNE C. WILTON,'DELYTHE. WILLIAMS,' ALASTAIRJ. STRAIN,'ROSEMARY A. PARSLOW,' J. KEVIN CHIPMAN' AND ROGERCOLEMAN' 'The School of Biochemistry, The University of Birmingham and 2The Liver Laboratories, Queen Elizabeth Medical Centre, Edgbaston, Birmingham B15 2TT,United Kingdom
An initial preparation of rat hepatocytes containing approximately 30% couplets was enriched by centrifugal elutriation. Of the couplets loaded onto the elutriator, 87%were eluted at medium flow rates of 60 to 80 ml/min at a rotor speed of 1.100 rpm; cells eluted in this range maintained a viability of more than 95%. Peak fractions were enriched in couplets to 84.5%f 2.6%. After elutriation, couplets retained the ability to secrete fluorescent cholephiles into sealed canalicular vacuoles. The preparation can now be used in hepatobiliary and hepatotoxicity studies not possible with preparations in which they are minor compo1991;14:180-183.) nents. (HEPATOLOGY
of viable couplet (and triplet) cells from a mixed population of rat hepatocytes based on a centrifugal elutriation technique.
MATERIALS AND METHODS Materials. Collagenase from Clostridium hystolyticum (0.287 U/mg) (Boehringer Mannheim, Lewes, Sussex, UK), Leibovitz-15 (L-15) tissue culture medium (Gibco, Paisley, UK) and fluorescein (sodium salt) (Sigma Chemical Co., Poole, Dorset, UK) were used. CL-FITC was synthesized as described by Fentem et al. (8). The elutriation buffer consisting of Krebs-Henseleit buffer ([pH 7.41 containing 0.1 gm/lOO ml BSA albumin, 0.1 gm/lOO ml glucose gassed with 95% oxygen/5% carbon dioxide) was Despite being of considerable use in many metabolic kept at room temperature. himals.Male Wistar rats (230 to 240 gm) bred in the studies, freshly isolated singlet hepatocytes cannot be used in hepatobiliary studies because they lack a biliary University of Birmingham had access to food and tap water ad pole (1, 2). Although redistribution of the canalicular libitum until 15 min before liver isolation. All protocols were membrane can occur in cultured hepatocytes within 24 approved by the local animal ethics committee. of Hepatocyte Couplets. The method, involving hr of preparation, leading to the formation of functional anPreparation incomplete digestion of rat liver by limited collagenase biliary poles (2, 31, these cultures are of limited use in perfusion in situ, was based on that described by Gautam et al. toxicological studies because of the loss of several drug (6), which was modified by reducing the collagenase perfusion metabolizing enzymes, particularly cytochrome P-450 to 0.03% (wt/vol) collagenase for 4.5 min. isoenzymes (4). The yield of cells was enhanced by further incubation of the Boyer et al. ( 5 ) have described a preparation of rat remaining liver tissue in a shaking water bath at 37" C for hepatocytes in which approximately 20% to 30%exist as 7 min with 30 ml collagenase solution as described previously. couplets, allowing the study of primary canalicular After filtration of the freed cells through 150 km nylon gauze, secretion in isolated hepatocytes (6,7).Fentem et al. (8) the cells were allowed to settle at 0" C for 5 min and were have shown that the selective accumulation of fluo- washed three times with L-15. The initial preparation and the obtained from secondary incubation were studied rescein and a fluorescent bile salt analog (cholyl-lysyl preparation separately. FITC [CL-FITCI) in the resealed canalicular vacuoles Centrifugal Elutriatwn. Cells were separated as a function exhibit kinetics similar to those in the whole animal (9). of their sedimentation velocity (see lo), changing only the Because cytochrome P-450activity is well maintained counterfiow of the medium; the viscosity and rotor speed in short-term hepatocyte cultures (41, the preparation of remained constant. To ensure the former, the chamber and purified viable couplets and triplets offers an improved liquid flow through it was maintained at room temperature opportunity for hepatobiliary and hepatoxicity studies (21" C) throughout the study. This temperature may have in uitro. We report here the separation and purification helped retain integrity of the cells by maintaining membrane fluidity. A standard elutriation chamber in a JE-6B elutriation rotor (Beckman RIIC Ltd, High Wycombe, UK) was mounted in a Received November 27, 1990 accepted February 21, 1991. Beckman J6-ME centrifuge. Cells were separated by mainJ. C. Wilton is sponsored by the Wellcome Trust, grant no. 030949/2/89. Address reprint requests to: Roger Coleman, D.Sc.,School of Biochemistry, taining the rotor speed at 1,100 rpm and sequentially increasing the flow of the elutriation medium through the The University of Birmingham, Edgbaston, Birmingham B15 2TT, United chamber using a peristaltic pump. Silicone tubing was used Kingdom. throughout. 31/1/29686 180
Vol. 14,No. 1, 1991
181
PURIFICATION OF HEPATOCYTE COUPLETS 100
a
+ VI
-a W
3
80
m
a
t
2
0
m
*O 15
I
R
+
60
b 0
20
40
60
1 80 90
Elutriotion flow rate rnll rnin
FIG.2. Proportions of hepatocytes present as viable couplets and triplets in each fraction collected during centrifugal elutriation. Each value is mean i- S.E.M. ( n = 8 ) .
+
._
&
k X
0
RESULTS Initial Cell Preparations. In situ liver perfusion with the collagenase conditions used previously gave a prep- % aration containing 27.4% ? 3.1% “couplets” (n = 4). =3 o The term “couplets” includes couplets and triplets. 0 20 40 60 80 90 When the residue was exposed to further collagenase E l u t r i a t i o n f l o w r a t e ml I m i n treatment in vitro, a preparation of 38.5% k 6.5% couplets was obtained (n = 4). The cells were resusFIG.1. Distribution of subpopulations of hepatocyte units eluted pended in approximately 50 ml L-15 at concentrations during a single representative experiment. Total (0)and viable ( 0 ) units are shown. Subpopulations of units are (a)singlets. (b) couplets, ranging from 1.32 to 2.79 x lo6 unitdml. Loading and E h t r i a t w n . Similar amounts of total (c) triplets and (d) multiples greater than 3. units (2.41 2 0.44 x lo7) were loaded, containing 8.0 L 0.56 x lo6 couplets (n = 8). These couplets An average of 2.41 t- 0.44 x 10’ cells was loaded into the (87%)were then recovered in the collected fractions. Figure 1describes the pattern of the cell population in separation chamber at a rate of 6 ml/min. Once loaded, the flow rate was increased. Fractions (100 ml) were collected at flow each fraction eluted during a representative single rates from 25 to 90 ml/min and kept on ice. The eluted cells elutriation run. It can be seen that the single cells elute were sedimented by centrifugation at 85 g for 5 min, and the mainly during the slower flow rates (20 to 70 ml/min) cells were resuspended in L-15 ( 4 ml). (Fig. la), followed by couplets (60 to 80 ml/min) (Fig. lb), Cell Counting and Cell Viability. Quantitation of cell pop- triplets (66 to 85 ml/min) (Fig. lc) and finally quadruulation was described in terms of units. A unit could be plets and small multiples (80 to 90 ml/min) (Fig. Id). comprised of a singlet, couplet, triplet, quadruplet or other The difference between the upper values (total units) multiple. Thus, the actual number of cells per unit was variable. Cells were counted on an improved Neubauer (Fig. 1, open circles) and the lower values (viable units) hemocytometer, and their viability was determined by trypan (Fig. 1,solid circles) represents the contribution of dead blue exclusion (11).If a single cell within a unit failed to units. It can be seen that in all cases viable units of a exclude trypan blue, the whole unit was considered dead, even given size elute at higher flow rates than dead units. though the remaining cells in the unit were viable. Although virtually 100% of cells present in the couplet Short-term Culture of Hepatocytee. Eluted cell suspensions fraction collected between 60 and 80 mumin were (2 ml at 1.0 x lo6 units/ml of L-15 medium) were plated in 30 hepatocytes, as judged by morphological criteria after mm diameter plastic tissue culture dishes and maintained at phase-contrast microscopy, we also determined the 37” C in an air atmosphere for 4 hr. proportion of contaminating nonparenchymal cells. For uptake studies (see 81, sodium fluorescein (20 ~ 1 20 ; mg/ml in isotonic NaCl) or CL-FITC (20 p1; 2 mg/ml in isotonic Cytospin preparations were stained for endogenous saline) was added to each culture dish and incubated at 37” C peroxidase as a marker of Kupffer cells and immunofor 10 min (CL-FITC) or 1hr (fluorescein) followed by washing stained for cytokeratin 19 (biliary epithelial cells), in L-15 medium. Cells were then observed at 37” C with an desmin (lipocytes) and vimentin (endothelial cells). Few inverted fluorescent microscope (Olympus IMT2-RFL). positively staining cells were observed, accounting for v,
lor
d
59
9
182
HEPATOLOGY
WILTON ET AL.
can largely be freed of contaminating singlets and dead cells until they become the predominant type of unit in the preparation. The principal contaminants are triplet cells that, functionally, are similar to the couplets. Furthermore, the virtual absence of other contaminating cell types indicates the efficient separation of hepatocyte couplets from smaller single cells. Quantitation of the enrichment, as the relative numbers of couplets progressively increase, poses a problem. The enrichment can change its apparent magnitude according to the practice of counting (e.g.,we have chosen to use the term “unit” [ l singlet = 1 unit, 1 couplet = 2 cells = 1 unit]) (see “Materials and Methods”). This gives smaller values for enrichment than if cells are used, but we believe it is more readily usable and understood, e.g.9 FIG. 3. Canalicular incorporation of fluorescein in an eluted preparation (fraction collected at 66.4 mlimin). Couplets were cultured for 5 hr, the last hour of which in the presence of fluorescein.
80 (units) couplets
+ 20 (units) singlets
=
80% couplets (units)
or less than 1%of the total cell population. In addition, more than 99% of cells stained positively for cytokeratin 18, which is expressed by hepatocytes. Reproducibility. Different values of percent couplets and percent viability in the original cell suspensions applied to the elutriator did not appear to affect the characteristics of the position of the couplets and triplets in the elutriation profile. Thus the mean proportion of couplets at different flow rates is presented for all eight preparations. It can be seen that there is a reproducible high degree of enrichment. In the range 64 to 80 ml, the viabilities of these fractions were more than 95%. In the peak fractions at 66.4 and 69.5 ml/ml, the enrichment of couplets had reached 84.5% + 2.5% (Fig. 2). Morphology and Function. In Figure 3, the appearance of a couplet peak fraction is shown after 4 hr culture followed by 1 hr in the presence of fluorescein. Five couplets are seen in fluorescence microscopy exhibiting different stages in the pumping of fluorescein into the canalicular vacuole. Evidence of dilatating vacuoles could also be seen under tungsten illumination. Contraction or collapse of these vacuoles was also seen by both types of microscopy. Couplets accumulated CL-FITC at a faster rate than fluorescein, and the bile acid analog was seen to accumulate in 10 to 15 min at 37” C. DISCUSSION
The initial hepatocyte preparation is essentially similar to that of others (6,9)having approximately 30% couplets, but many having singlets and some dead cells. The second in vitro preparation has improved couplet frequency and higher overall viability and may prove to be advantageous for some applications. Much greater enrichment in couplet frequency can be obtained by the elutriation procedure. The advantageous removal of dead before viable and singlets before couplets (and multiplets) ensures that viable couplets
160 cells (couplets) + 20 cells (singlets) = 89% couplets (cells).
Centrifugal elutriation has been used for other purposes in separations of liver cells, (e.g., separation of singlet hepatocytes [121 of sinusoidal epithelial cells [131 of Kupffer cells [141 of biliary epithelial cells [151 from liver cell suspensions and in the separation of hepatocytes from various zones of the liver [12,161). In Sumner et al.’s study (121, a population of couplets and multiples was also observed incidentally in the elutriation profile. The present elutriation study has established conditions where hepatocyte couplets have become the predominant cell type. This allows use of such preparations for studies of biliary accumulation and its malfunction by biochemical techniques that are less applicable to a couplet minority preparation, which must rely mainly on various forms of microscopy. Acknowledgments: We thank the Nuffield Foundation and the Special Trustees of the United Birmingham Hospitals for financial support. The technical assistance of D. Anderson is greatly acknowledged. REFERENCES 1. Coleman R. The biochemistry of bile secretion. Biochem J 1987;244:249-261. 2. Gebhart R. Use of isolated and cultured hepatocytes in studies on bile formation. In: Guillouzo A, Gugen-Guillouzo C , eds. Research in isolated and cultured hepatocytes. John Libbey EurotextiINSERM, 1986:353-376. 3. Petzinger E, Follman W, Acker A, Hentschel J, Zierold K, Kinne RKH. Primary liver cell cultures on gas permeable membrane as a source for the collection of primary bile. In Vitro Cell Develop Biol 1988;24:491-499. 4. Fry JR, Bridges JW.Use of primary hepatocyte cultures in biochemical toxicology. In: Hodgson E, Bend JR, Philpot R, eds. Reviews of biochemical toxicology. Amsterdam: Elsevier/North Holland, 1979:201-204. 5 . Boyer JL, Ng OC, Gautam A. Formation of canalicular spaces in isolated rat hepatocyte couplets. Trans Assoc Am Phys 1985;98: 21-29.
Vol. 14, No. 1, 1991
PURIFICATION OF HEPATOCYTE COUPLETS
6. Gautam A, Ng OC, Boyer JL. Isolated rat hepatocyte couplets in
7. 8.
9.
10. 11.
short-term culture: structural characteristics and plasma membrane reorganisation. HEPATOLO(;Y 1987:7:216-223. Boyer JL, Gautam A, Graf A. Mechanisms of bile secretion: insights from the isolated rat hepatncyte couplet. Semin Liver Dis 1988;8:308-316. Fentem JH,Foster B, Mills CO, Coleman R, Chipman JK. Biliary excretion of fluorescent cholephiles in hepatic couplets: an in vitro model for hepatobiliary and hepatotoxicity studies. Toxicol In Vitro 1990;4:452-457. Mills CO, Rahman K, Coleman R. Elias E. Cholyl-lysyl fluorescein isothiocyanate: biliary kenetics in vitro and during single-pass perfusion of isolated perfused rat liver [Abstractl. Clin Sci 1985;75(suppl. 19):9P. Saunders MJ, Sol1 AH. Cell preparation by elutriation: major and minor cell types from complex tissues. Methods Enzymol 1989; 171:482-497. Moldeus P, Hogberg J, Orrenius S. Isolation and use of liver cells. Methods Enzymol 1978;52:60-71
183
12. Sumner IG. Friedman RB, Lodala A. Characterization of hepatocyte subpopulations generated by centrifugal elutriation. Eur J Biochem 1983;134:539-545. 13. Eyrhorn S, Schayer HJ, Henninger HP, Dieter P, Herman R, Woort-Menker M, Becker H, et al. Rat hepatic sinusoidal epithelial cells in monolayer culture: biochemical and ultrastructural characteristics. J Hepatol 1988;6:23-35. 14. Rieder H, Ramadori G, Meyer zum Buschenfelde K-H. Guinea-pig Kupffer cells can be activated in uitro to an enhanced superoxide response: comparison with peritoneal macrophages. J Hepatol 1988;7:338-344. 15. Alpini G, Lenzi R, Zhai WR, Liu MH, Slott PA, Paronetto F, Tavolini N. Isolation of nonparenchymal liver cell fractions enriched in cells with biliary epithelium phenotypes. Gastroenterology 1989;97:1248-1260. 16. Siebert B, Oesch F, Steinberg P. Distribution and induction of cytochrome P-450 and two cytochrome P-450 dependent monoxygenase activities in rat liver parenchymal cell subpopulations separated by centrifugal elutriation. Arch Toxicol 1989;63:18-22.
An initial preparation of rat hepatocytes containing approximately 30% couplets was enriched by centrifugal elutriation. Of the couplets loaded onto the elutriator, 87%were eluted at medium flow rates of 60 to 80 ml/min at a rotor speed of 1.100 rpm; cells eluted in this range maintained a viability of more than 95%. Peak fractions were enriched in couplets to 84.5%f 2.6%. After elutriation, couplets retained the ability to secrete fluorescent cholephiles into sealed canalicular vacuoles. The preparation can now be used in hepatobiliary and hepatotoxicity studies not possible with preparations in which they are minor compo1991;14:180-183.) nents. (HEPATOLOGY
of viable couplet (and triplet) cells from a mixed population of rat hepatocytes based on a centrifugal elutriation technique.
MATERIALS AND METHODS Materials. Collagenase from Clostridium hystolyticum (0.287 U/mg) (Boehringer Mannheim, Lewes, Sussex, UK), Leibovitz-15 (L-15) tissue culture medium (Gibco, Paisley, UK) and fluorescein (sodium salt) (Sigma Chemical Co., Poole, Dorset, UK) were used. CL-FITC was synthesized as described by Fentem et al. (8). The elutriation buffer consisting of Krebs-Henseleit buffer ([pH 7.41 containing 0.1 gm/lOO ml BSA albumin, 0.1 gm/lOO ml glucose gassed with 95% oxygen/5% carbon dioxide) was Despite being of considerable use in many metabolic kept at room temperature. himals.Male Wistar rats (230 to 240 gm) bred in the studies, freshly isolated singlet hepatocytes cannot be used in hepatobiliary studies because they lack a biliary University of Birmingham had access to food and tap water ad pole (1, 2). Although redistribution of the canalicular libitum until 15 min before liver isolation. All protocols were membrane can occur in cultured hepatocytes within 24 approved by the local animal ethics committee. of Hepatocyte Couplets. The method, involving hr of preparation, leading to the formation of functional anPreparation incomplete digestion of rat liver by limited collagenase biliary poles (2, 31, these cultures are of limited use in perfusion in situ, was based on that described by Gautam et al. toxicological studies because of the loss of several drug (6), which was modified by reducing the collagenase perfusion metabolizing enzymes, particularly cytochrome P-450 to 0.03% (wt/vol) collagenase for 4.5 min. isoenzymes (4). The yield of cells was enhanced by further incubation of the Boyer et al. ( 5 ) have described a preparation of rat remaining liver tissue in a shaking water bath at 37" C for hepatocytes in which approximately 20% to 30%exist as 7 min with 30 ml collagenase solution as described previously. couplets, allowing the study of primary canalicular After filtration of the freed cells through 150 km nylon gauze, secretion in isolated hepatocytes (6,7).Fentem et al. (8) the cells were allowed to settle at 0" C for 5 min and were have shown that the selective accumulation of fluo- washed three times with L-15. The initial preparation and the obtained from secondary incubation were studied rescein and a fluorescent bile salt analog (cholyl-lysyl preparation separately. FITC [CL-FITCI) in the resealed canalicular vacuoles Centrifugal Elutriatwn. Cells were separated as a function exhibit kinetics similar to those in the whole animal (9). of their sedimentation velocity (see lo), changing only the Because cytochrome P-450activity is well maintained counterfiow of the medium; the viscosity and rotor speed in short-term hepatocyte cultures (41, the preparation of remained constant. To ensure the former, the chamber and purified viable couplets and triplets offers an improved liquid flow through it was maintained at room temperature opportunity for hepatobiliary and hepatoxicity studies (21" C) throughout the study. This temperature may have in uitro. We report here the separation and purification helped retain integrity of the cells by maintaining membrane fluidity. A standard elutriation chamber in a JE-6B elutriation rotor (Beckman RIIC Ltd, High Wycombe, UK) was mounted in a Received November 27, 1990 accepted February 21, 1991. Beckman J6-ME centrifuge. Cells were separated by mainJ. C. Wilton is sponsored by the Wellcome Trust, grant no. 030949/2/89. Address reprint requests to: Roger Coleman, D.Sc.,School of Biochemistry, taining the rotor speed at 1,100 rpm and sequentially increasing the flow of the elutriation medium through the The University of Birmingham, Edgbaston, Birmingham B15 2TT, United chamber using a peristaltic pump. Silicone tubing was used Kingdom. throughout. 31/1/29686 180
Vol. 14,No. 1, 1991
181
PURIFICATION OF HEPATOCYTE COUPLETS 100
a
+ VI
-a W
3
80
m
a
t
2
0
m
*O 15
I
R
+
60
b 0
20
40
60
1 80 90
Elutriotion flow rate rnll rnin
FIG.2. Proportions of hepatocytes present as viable couplets and triplets in each fraction collected during centrifugal elutriation. Each value is mean i- S.E.M. ( n = 8 ) .
+
._
&
k X
0
RESULTS Initial Cell Preparations. In situ liver perfusion with the collagenase conditions used previously gave a prep- % aration containing 27.4% ? 3.1% “couplets” (n = 4). =3 o The term “couplets” includes couplets and triplets. 0 20 40 60 80 90 When the residue was exposed to further collagenase E l u t r i a t i o n f l o w r a t e ml I m i n treatment in vitro, a preparation of 38.5% k 6.5% couplets was obtained (n = 4). The cells were resusFIG.1. Distribution of subpopulations of hepatocyte units eluted pended in approximately 50 ml L-15 at concentrations during a single representative experiment. Total (0)and viable ( 0 ) units are shown. Subpopulations of units are (a)singlets. (b) couplets, ranging from 1.32 to 2.79 x lo6 unitdml. Loading and E h t r i a t w n . Similar amounts of total (c) triplets and (d) multiples greater than 3. units (2.41 2 0.44 x lo7) were loaded, containing 8.0 L 0.56 x lo6 couplets (n = 8). These couplets An average of 2.41 t- 0.44 x 10’ cells was loaded into the (87%)were then recovered in the collected fractions. Figure 1describes the pattern of the cell population in separation chamber at a rate of 6 ml/min. Once loaded, the flow rate was increased. Fractions (100 ml) were collected at flow each fraction eluted during a representative single rates from 25 to 90 ml/min and kept on ice. The eluted cells elutriation run. It can be seen that the single cells elute were sedimented by centrifugation at 85 g for 5 min, and the mainly during the slower flow rates (20 to 70 ml/min) cells were resuspended in L-15 ( 4 ml). (Fig. la), followed by couplets (60 to 80 ml/min) (Fig. lb), Cell Counting and Cell Viability. Quantitation of cell pop- triplets (66 to 85 ml/min) (Fig. lc) and finally quadruulation was described in terms of units. A unit could be plets and small multiples (80 to 90 ml/min) (Fig. Id). comprised of a singlet, couplet, triplet, quadruplet or other The difference between the upper values (total units) multiple. Thus, the actual number of cells per unit was variable. Cells were counted on an improved Neubauer (Fig. 1, open circles) and the lower values (viable units) hemocytometer, and their viability was determined by trypan (Fig. 1,solid circles) represents the contribution of dead blue exclusion (11).If a single cell within a unit failed to units. It can be seen that in all cases viable units of a exclude trypan blue, the whole unit was considered dead, even given size elute at higher flow rates than dead units. though the remaining cells in the unit were viable. Although virtually 100% of cells present in the couplet Short-term Culture of Hepatocytee. Eluted cell suspensions fraction collected between 60 and 80 mumin were (2 ml at 1.0 x lo6 units/ml of L-15 medium) were plated in 30 hepatocytes, as judged by morphological criteria after mm diameter plastic tissue culture dishes and maintained at phase-contrast microscopy, we also determined the 37” C in an air atmosphere for 4 hr. proportion of contaminating nonparenchymal cells. For uptake studies (see 81, sodium fluorescein (20 ~ 1 20 ; mg/ml in isotonic NaCl) or CL-FITC (20 p1; 2 mg/ml in isotonic Cytospin preparations were stained for endogenous saline) was added to each culture dish and incubated at 37” C peroxidase as a marker of Kupffer cells and immunofor 10 min (CL-FITC) or 1hr (fluorescein) followed by washing stained for cytokeratin 19 (biliary epithelial cells), in L-15 medium. Cells were then observed at 37” C with an desmin (lipocytes) and vimentin (endothelial cells). Few inverted fluorescent microscope (Olympus IMT2-RFL). positively staining cells were observed, accounting for v,
lor
d
59
9
182
HEPATOLOGY
WILTON ET AL.
can largely be freed of contaminating singlets and dead cells until they become the predominant type of unit in the preparation. The principal contaminants are triplet cells that, functionally, are similar to the couplets. Furthermore, the virtual absence of other contaminating cell types indicates the efficient separation of hepatocyte couplets from smaller single cells. Quantitation of the enrichment, as the relative numbers of couplets progressively increase, poses a problem. The enrichment can change its apparent magnitude according to the practice of counting (e.g.,we have chosen to use the term “unit” [ l singlet = 1 unit, 1 couplet = 2 cells = 1 unit]) (see “Materials and Methods”). This gives smaller values for enrichment than if cells are used, but we believe it is more readily usable and understood, e.g.9 FIG. 3. Canalicular incorporation of fluorescein in an eluted preparation (fraction collected at 66.4 mlimin). Couplets were cultured for 5 hr, the last hour of which in the presence of fluorescein.
80 (units) couplets
+ 20 (units) singlets
=
80% couplets (units)
or less than 1%of the total cell population. In addition, more than 99% of cells stained positively for cytokeratin 18, which is expressed by hepatocytes. Reproducibility. Different values of percent couplets and percent viability in the original cell suspensions applied to the elutriator did not appear to affect the characteristics of the position of the couplets and triplets in the elutriation profile. Thus the mean proportion of couplets at different flow rates is presented for all eight preparations. It can be seen that there is a reproducible high degree of enrichment. In the range 64 to 80 ml, the viabilities of these fractions were more than 95%. In the peak fractions at 66.4 and 69.5 ml/ml, the enrichment of couplets had reached 84.5% + 2.5% (Fig. 2). Morphology and Function. In Figure 3, the appearance of a couplet peak fraction is shown after 4 hr culture followed by 1 hr in the presence of fluorescein. Five couplets are seen in fluorescence microscopy exhibiting different stages in the pumping of fluorescein into the canalicular vacuole. Evidence of dilatating vacuoles could also be seen under tungsten illumination. Contraction or collapse of these vacuoles was also seen by both types of microscopy. Couplets accumulated CL-FITC at a faster rate than fluorescein, and the bile acid analog was seen to accumulate in 10 to 15 min at 37” C. DISCUSSION
The initial hepatocyte preparation is essentially similar to that of others (6,9)having approximately 30% couplets, but many having singlets and some dead cells. The second in vitro preparation has improved couplet frequency and higher overall viability and may prove to be advantageous for some applications. Much greater enrichment in couplet frequency can be obtained by the elutriation procedure. The advantageous removal of dead before viable and singlets before couplets (and multiplets) ensures that viable couplets
160 cells (couplets) + 20 cells (singlets) = 89% couplets (cells).
Centrifugal elutriation has been used for other purposes in separations of liver cells, (e.g., separation of singlet hepatocytes [121 of sinusoidal epithelial cells [131 of Kupffer cells [141 of biliary epithelial cells [151 from liver cell suspensions and in the separation of hepatocytes from various zones of the liver [12,161). In Sumner et al.’s study (121, a population of couplets and multiples was also observed incidentally in the elutriation profile. The present elutriation study has established conditions where hepatocyte couplets have become the predominant cell type. This allows use of such preparations for studies of biliary accumulation and its malfunction by biochemical techniques that are less applicable to a couplet minority preparation, which must rely mainly on various forms of microscopy. Acknowledgments: We thank the Nuffield Foundation and the Special Trustees of the United Birmingham Hospitals for financial support. The technical assistance of D. Anderson is greatly acknowledged. REFERENCES 1. Coleman R. The biochemistry of bile secretion. Biochem J 1987;244:249-261. 2. Gebhart R. Use of isolated and cultured hepatocytes in studies on bile formation. In: Guillouzo A, Gugen-Guillouzo C , eds. Research in isolated and cultured hepatocytes. John Libbey EurotextiINSERM, 1986:353-376. 3. Petzinger E, Follman W, Acker A, Hentschel J, Zierold K, Kinne RKH. Primary liver cell cultures on gas permeable membrane as a source for the collection of primary bile. In Vitro Cell Develop Biol 1988;24:491-499. 4. Fry JR, Bridges JW.Use of primary hepatocyte cultures in biochemical toxicology. In: Hodgson E, Bend JR, Philpot R, eds. Reviews of biochemical toxicology. Amsterdam: Elsevier/North Holland, 1979:201-204. 5 . Boyer JL, Ng OC, Gautam A. Formation of canalicular spaces in isolated rat hepatocyte couplets. Trans Assoc Am Phys 1985;98: 21-29.
Vol. 14, No. 1, 1991
PURIFICATION OF HEPATOCYTE COUPLETS
6. Gautam A, Ng OC, Boyer JL. Isolated rat hepatocyte couplets in
7. 8.
9.
10. 11.
short-term culture: structural characteristics and plasma membrane reorganisation. HEPATOLO(;Y 1987:7:216-223. Boyer JL, Gautam A, Graf A. Mechanisms of bile secretion: insights from the isolated rat hepatncyte couplet. Semin Liver Dis 1988;8:308-316. Fentem JH,Foster B, Mills CO, Coleman R, Chipman JK. Biliary excretion of fluorescent cholephiles in hepatic couplets: an in vitro model for hepatobiliary and hepatotoxicity studies. Toxicol In Vitro 1990;4:452-457. Mills CO, Rahman K, Coleman R. Elias E. Cholyl-lysyl fluorescein isothiocyanate: biliary kenetics in vitro and during single-pass perfusion of isolated perfused rat liver [Abstractl. Clin Sci 1985;75(suppl. 19):9P. Saunders MJ, Sol1 AH. Cell preparation by elutriation: major and minor cell types from complex tissues. Methods Enzymol 1989; 171:482-497. Moldeus P, Hogberg J, Orrenius S. Isolation and use of liver cells. Methods Enzymol 1978;52:60-71
183
12. Sumner IG. Friedman RB, Lodala A. Characterization of hepatocyte subpopulations generated by centrifugal elutriation. Eur J Biochem 1983;134:539-545. 13. Eyrhorn S, Schayer HJ, Henninger HP, Dieter P, Herman R, Woort-Menker M, Becker H, et al. Rat hepatic sinusoidal epithelial cells in monolayer culture: biochemical and ultrastructural characteristics. J Hepatol 1988;6:23-35. 14. Rieder H, Ramadori G, Meyer zum Buschenfelde K-H. Guinea-pig Kupffer cells can be activated in uitro to an enhanced superoxide response: comparison with peritoneal macrophages. J Hepatol 1988;7:338-344. 15. Alpini G, Lenzi R, Zhai WR, Liu MH, Slott PA, Paronetto F, Tavolini N. Isolation of nonparenchymal liver cell fractions enriched in cells with biliary epithelium phenotypes. Gastroenterology 1989;97:1248-1260. 16. Siebert B, Oesch F, Steinberg P. Distribution and induction of cytochrome P-450 and two cytochrome P-450 dependent monoxygenase activities in rat liver parenchymal cell subpopulations separated by centrifugal elutriation. Arch Toxicol 1989;63:18-22.
