Development of a bioartificial liver: Properties and function of a hollow-fiber module inoculated with liver cells

Please download to get full document.

View again

of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Development of a bioartificial liver: Properties and function of a hollow-fiber module inoculated with liver cells
Document Share
Document Tags
Document Transcript
  Development of a Bioartificial Liver: Properties and Function of a Hollow-fiber Module Inoculated with Liver Cells JACEK OZGA,~ REDERICK WILLIAMS,^ MAN-SO0 RO, DANIEL F. NEUZIL,~ ODD D. GIORG10,4 GISELA BACKFISCH,3 ALBERT D. MOSCIONI,~ AYMOND AND ACHILLES . DEMETRIOU z epartments of ISurgery, 2Medicine and 3Pathology, Vanderbilt University School zyx f Medicine; 4Department of Chemical Engineering, Vanderbilt University; and 5Department of Veterans Affairs Medical Center, Nashville, Tennessee z 7212 z Wehave developed a bioartificial liver support system utilizing hollow-fiber bioreactor, plasma- pheresis and microcarrier cell culture technologies. Liver cells were obtained through portal vein per- fusion with zyxwvut thylenediaminetetraacetate or ethylene- diaminetetraacetate/collagenase. A mathematical model of mass transport in a hollow-fiber module, at various plasma flow velocities and system configura- tions, was developed. The bioartificial liver's ability to carry out specific differentiated metabolic liver func- tions was tested in vitro and in vivo. A reproducible large-animal model of acute ischemic liver failure was developed. Most major first-generation cyclosporine and 19-norterstosterone metabolites were isolated after substrate addition to the bioartificial liver in vitro. After bioartificial liver treatment for 6 hr (with dog or pig liver cells), dogs with acute liver failure had significantly lower serum ammonia and lactate levels and significantly higher serum glucose levels than did control animals treated with a bioartificial liver system inoculated with microcarriers alone. In addition, bio- artificial liver-treated animals had significantly higher mean systolic blood pressures than did controls. Liver cell viability at the end of the 6-hr in vivo experiment was greater than 90%. HEPATOLOGY 993; 175258-265.) The need exists to create an artificial liver support system for patients with severe liver failure who require temporary liver support. Previously described systems relied mainly on blood detoxification (1-7). More re- cently, we and others emphasized that isolated intact liver cells should be used to construct a liver support system to provide both detoxifying and synthetic func- tions (8-12). Received May 4, 1992; accepted September 15, 1992. This work was supported in part by National Institutes of Health grant DK38763-06, National Institute of Environmental Health Sciences pilot support grant 00267 and a Department of Veterans Affairs Merit Review Award. Address reprint requests to: Dr. Achilles A. Demetriou, Cedars-Sinai Medical Center, Department of Surgery, Room 8215, North Tower, 8700 Beverly Boulevard, Los Angeles, CA 90048. Copyright zyxwvutsrqp   1993 by the American Association for the Study of Liver Diseases. 0270-9139193 1.00 zyxwvutsrqpo   0.10 zyxwvutsrq  42775 In this series of experiments, we describe methods of harvesting liver cells, an experimental model of liver insufficiency, a bioartificial liver (BAL), a modeling system for prediction of BAL efficiency and the BAL's ability to carry out specific metabolic functions in vitro and provide metabolic and hemodynamic support in vivo. MATERIALS AND METHODS The animal studies reported here were performed in compliance with institutional and National Research Council guidelines for humane care of experimental animals. Preparation of Liver Cells. Liver cell isolation was carried out aseptically. Unless otherwise noted, all chemical and cell culture reagents used were purchased from Sigma Chemical Co. (St. Louis, MO). Rat liver cells were harvested from adult male Wistar rats with in situ portal vein collagenase perfusion (13). Male and female mongrel dogs 10 to 20 kg) were anesthe- tized with pentobarbital (30 mg/kg, intravenously) and given ventilator support. After total hepatectomy, each liver was placed in a sterile basin on ice-cold saline solution, and the portal vein was cannulated. The liver was perfused for 5 min with normal saline solution; this was followed by perfusion for 30 min with a 2 mmol/L EDTA solution prepared according to the method of Wang et al. (14) (solution A . The perfusate was passed through gas-permeable silicone tubing (Baxter Healthcare, McGaw Park, IL) submerged in a water bath maintained at 38 C and aerated with 100% oxygen, The oxygenated solution was perfused through the liver at 100 ml/min with a roller pump (MasterFlex; Cole-Palmer, Chicago, IL). The liver was then perfused with solution B (solution A without EDTA and NaHCO, but with 1 mmol/L CaCl,; 14) for 10 min. The pale, soft portions of liver were excised, minced and passed through 100-pm nylon mesh (Spectramesh; Spectrum Laboratory Products, Los Angeles, CA). Liver cell-enriched fractions were prepared by sedimentation 50 g for 2 mid, and viability was determined by trypan blue exclusion. Liver cells were isolated from male and female pigs weighing 10 to 12 kg. While each animal was under ketamine anesthesia (30 mag, ntravenously) the abdomen was entered through a midline incision, the hepatoduodenal ligament was dissected and all its structures were ligated and divided, except for the portal vein, which was cannulated with silicone tubing. The liver was perfused in situ with 3 L of solution A (300 ml/min) by means of a roller pump. Ten minutes later, the liver was  HEPATOLOGY zyxwv ol. 17, No. 2, 1993 TRANS zyxw   FLOW ROZGA ET z L 259 Qt ~I,FLUID ATH REDUCTOR SAMPLE PORT FCOW REDUCTOR PUMP RESERVOIR FIG. 1. Fluid recirculation loop for the study of solute transport in a hollow-fiber module. Axial flow rate zyx Q) was controlled by a pump. A three-compartment model was used to describe solute transport in the module (extrafiber space: zyxw e = volume, Ce = solute concentration; fiber lumen: Vf = volume, zyxwvutsrq f = solute concentration; fluid reservoir: Vr = volume, Cr = solute concentration). Each compartment was treated as a continuously stirred tank reactor with multiple inputs and outputs. resected and placed in a sterile basin containing a 0.1% type IV collagenase solution prepared in Ca+ +-enriched buffer (15). The solution was recirculated through the liver (300 mlimin) after passage through silicone tubing submerged in an oxygen- saturated water bath at 38” C. Thirty minutes later, the liver capsule was disrupted, and digested liver parenchyma was suspended in a large volume of ice-cold Dulbecco’s modified Eagle medium (DMEM). Suspended liver cells were then passed through a 100-pm nylon mesh, and liver cell-enriched fractions were isolated by centrifugation (50 g for 2 min). Liver cell viability was assessed with trypan blue exclusion. Liver Cell Attachment to Microcurriers. Isolated liver cells were attached to collagen-coated dextran microcarriers as previously described (16). Rat or dog cells (1.0 to 1.5 x lo8 cells) or pig cells (1 x lo9 cells) were attached to hydrated microcarriers (1.6 gm dry wt . Liver Cell Cryopreservution. After incubation, micro- carrier-attached liver cells were stored at 0” C in DMEM, 10% FCS and 5 (vol/vol) dimethylsulfoxide for 2 wk. Module Design. Hollow-fiber “bioreactor” modules (Mini Kross M22M-030-01N) were purchased from Microgon Inc. (Laguna Hills, CAI. Each module consists of a polycar- bonate cylinder (1.9-cm diameter) containing 75 cellulose ni- trate/cellulose acetate porous fibers (pore size = 0.2 pm; fiber length = 21 cm; fiber inside diameter = 660 pm, fiber wall thickness = 50 pm). The extrafiber volume is 33.5 ml and the intrafiber volume is 5 ml. The external fiber surface area is 380 cm2. Study of Mass Transfer Across Module In Vitro. The module was connected to a reservoir with silicone tubing (Cole- Palmer); a roller pump (MasterFlex) was used to control flow of the recirculating fluid, as shown in Figure 1. The reservoir initially contained 2 L PBS supplemented with glucose (400 mg/dl) or BSA 4 gm/dl). Flow through the module was maintained at a rate of 10,30 or 50 ml/min. Fluid aliquots (200 pl) were removed from the extrafiber space for determination of glucose and albumin concentrations. Glucose concentration was assayed with a Beckman Glucose Analyzer 2 (Beckman Instruments, Fullerton, CA), and albumin was assayed spec- trophotometrically with a bromocresol green albumin test kit (Sigma Chemical Co.). Experiments were carried out with two flow configurations (Fig. zy ). In the “flow-through” configu- ration, the extrafiber space of the module was hydraulically isolated from the fluid in the fibers. In the “transflow” configuration, the axial flow was reduced by 20% with a flow reducer and redirected into the transflow fluid path. A three-compartment theoretical model, shown schemati- cally in Figure 1, was used to describe solute transport in the module. The extrafiber space (Ve = volume; Ce = solute con- centration), fiber lumen (Vf = volume; Cf = solute concen- tration) and fluid reservoir (Vr = volume; Cr = solute concen- tration) were each treated as continuously stirred tank reactors with multiple inputs and outputs. The differential equations that govern the system response follow. dCe 1 dt Ve J. Cf - Cel) dCf 1 dt Vr -_ J [Ce - Cfl Q . Cr dCr 1 dt Vr (Q . Cf - Crl) Fiber flow ( ,I nd all initial conditions were preset for each experiment. Extrafiber solute concentration (Ce), measured at various time intervals, was used to estimate the unknown solvent flow rate between the extrafiber space and the fibers J). arameter estimation was carried out on a microcomputer with differential equation-solving software (Tutsim Products, Palo Alto, CA). In Vitro Bioreuctor Module Studies. The experimental setup used a system composed of 30-ml modules (Minikapi225; Microgon, Inc.). Each module was inoculated, into the ex- trafiber space, with 3 x lo7 freshly thawed microcarrier- attached rat liver cells. One hour later, cyclosporine or 19-nor testosterone was inoculated into the system (1 mg/ml). In control experiments, vehicle alone or a test drug was injected into a system containing microcarriers alone. Cyclosporine and its metabolites were extracted and ana-  260 ROZGA ET zyxwvusrq L. HEPATOLOGY ebruary 1993 Microcarrier- Attached Hepatocytes FIG. 2. BAL support system. Hollow-fiber modules were inoculated with microcarrier-attached liver cells zy insert). lyzed on HPLC (17, 18) with a Hewlett-Packard Model 1090 chromatograph with workstation (Hewlett-Packard, Avon- dale, PA). Cyclosporine metabolites were rechromatographed on a Hewlett-Packard LC-C8 analytical column (17, 18). 19-Nortestosterone and its metabolites were extracted and submitted to gas chromatographyimass spectrometry. To de- termine the glucuronides of 19-nortestosterone, we dried and hydrolyzed samples (19, 20). Analyses were performed with a Hewlett-Packard model 5988A gas chromatograph/mass spec- trometer fitted with a Hewlett-Packard HP1 methylsilicon column (16.6 mm zyxwvutsr   0.2 mm; 0.33-km film thickness). In Vivo BAL Studies: Experimental Animal Model zyxwv f Liver Failure. While under pentobarbital (30 mg/kg, intravenously) anesthesia with mechanical ventilation, male and female mongrel dogs (15 to 20 kg) underwent end-to-side portocaval shunts with placement of sutures around the common hepatic and gastroduodenal arteries. The sutures were brought out into subcutaneous pockets. Forty-eight hours later, animals were reanesthetized and supported on oxygen-enriched room air mechanical ventilation. Jugular vein and femoral artery catheters were placed. The exteriorized arterial ligatures were applied after collection of baseline blood. Additional blood was collected at 60, 120, 240 and 300 min after hepatic devascu- larization. Before creation of portocaval shunt and 2 hr after hepatic devascularization, 0.5 mgkg indocyanine green (ICG) was administered intravenously and blood samples were obtained at 5-min intervals for 20 min. Ten adult mongrel dogs were studied. All assays were carried out at the hospital clinical laboratory. ICG clearance was determined spectrophotomet- rically zyxwvuts n Vivo BAL Studies: Experimental Design. Animals were divided into three groups: group 1 consisted of control dogs (n = 6), each attached to a BAL inoculated with microcarriers alone. Group 2 dogs (n = 4) were each attached to a BAL inoculated with cryopreserved microcarrier-attached allo- geneic (canine) liver cells. Group 3 consisted of dogs (n = 6) attached to BALs inoculated with cryopreserved microcarrier- attached xenogeneic (porcine) liver cells. Plasmapheresis. After cannulation of the femoral artery and vein, heparin (100 Ukg, intravenously) was given and plasmapheresis was initiated with a Haemonetics 303 blood processor (Haemonetics Inc., Braintree, MA). After plasma separation, plasma was pumped through two parallel hollow- fiber modules at 30 mlimin. Approximately 80% of the flow was directed through the hollow-fiber lumen; the other 20% crossed the porous cellulose acetate membrane to the ex- trafiber space and exited through a side port. Plasma was reconstituted with the blood cells before it was reinfused into the animal (Fig. 2). Perfusions were carried out for 6 hours. Blood samples were obtained at 0, 60, 120 and 360 min; prothrombin time; levels of glucose, ammonia and lactate; and activity of lactate dehydrogenase (LDH) and AST were determined in the hospital clinical laboratory. At the end of the perfusion period, animals were killed with pentobarbital overdoses. Data Analysis. Data were analyzed statistically with one-way ANOVA and paired Student’s zy   tests. RESULTS Cell Isolation Cryopreservation and Viability. Cell viability after harvest of rat liver cells was greater than 85%. The initial viability of canine liver cells after EDTA perfusion was poor-less than 60%. The yield of viable liver cells harvested with this method was less than 1 x lo9. Excellent results were obtained with the EDTNcollagenase method in pigs > 95%; total viable cells = 2 x lolo . After cryopreservation, the viability of microcarrier-attached liver cells (rat, dog, pig) was 80% to 85%. In Vitro Mass Transfer Studies. The results of solute appearance in the extrafiber space are shown in Figure 3. The rate of solute appearance in the extrafiber space was a function of axial flow rate. For any axial flow rate, solute appearance was faster in the transflow than in the flow-through configuration. A significant lag period preceding the initial appearance of albumin, but not of glucose, was observed (Fig. 3a and b). The lag period decreased with increasing transfiber albumin transport.  HEPATOLOGY ol. 17, No. 2, 1993 1.0 zyxw 0 8 zyxw   zyxwv   . c zyx   zyxw 6 0 6 V C -0 0 4 .- C 0 2 0 0 ROZGA ET AL. 261 - - - - - - Solvent flux estimations correlated well with the rates of solute appearance, as shown in Figure zyxwvut . In Vitro Metabolic Studies. Cyclosporine metabolites were isolated from all modules containing microcarrier- attached rat liver cells after addition of cyclosporine. In control modules with microcarrier-attached liver cells not receiving cyclosporine, HPLC peaks corresponding to its metabolites were not detected; similarly, cyclo- sporine metabolites were not detected in modules containing microcarriers alone. Analysis of module extracts demonstrated two prominent peaks (X and Y) with zyxwvutsrq V spectral characteristics of cyclosporine and retention times corresponding to known metabolite standards (Fig. 5). Both peaks increased linearly with time after 4 hr of incubation. Peak X increased approx- imately twofold (rate = 2.8 pg/l x lo7 cellshr) and accounted for 8.5% of the srcinal dose at 8 hr of incubation; peak Y increased approximately 1 &fold (rate = 1.0 pg/l x lo7 cellshr) and accounted for 3.1% of the srcinal cyclosporine dose after 8 hr. When peaks X and Y were rechromatographed, peak X resolved into three metabolites that eluted at retention times of 14,15 and 16 min. Peak Y resolved into three metabolites that eluted at retention times of 21, 22, and 23 min. Comparison of this HPLC pattern to the pattern obtained from serum of a patient undergoing cyclo- sporine therapy demonstrated the production of iden- tical metabolites. On the basis of the comparisons of the HPLC retention times measured for these metabolites and their UV spectral assignments, we concluded that the three metabolites isolated from peak zyxwvu   represented first-generation monohydroxylation products of cyclo- sporine, two being metabolites 17 and 1, and that the metabolites under peak Y were first-generation N-demethylation products, one being metabolite 21 (18, 20-22). Metabolites of 19-nortestosterone were identified in medium after inoculation; they increased in concen- tration with time. The predominant metabolite was the glucuronide of 19-nortestosterone, which increased at an approximate rate of 134 ng/l x lo7 cellskr and accumulated, after 8 hr of culture, to about 3.5% of the level of the parent compound administered. The re- maining metabolites were isomerization products of the parent compound and amounted to less than 1 of the administered 19-nortestosterone at the end of the experiment. Identified from medium after 6 hr of cul- ture were 19-nortestosterone, 19-noreticholanolone and 19-norepiandrosterone. In Vivo B L Studies: Animal Model. Data collected from 10 dogs demonstrated that before creation of the portacaval shunt, rapid clearance of ICG took place 20 min after intravenous administration. Absorbance at 810 nm decreased from 1.90 k 0.08 to 0.52 k 0.08 (p < 0.001). After arterial ligation, near-complete re- tention of the dye was seen (2.0 0.10 to 1.80 -+ 0.09; p > 0.05). Five hours after hepatic devascularization, significant metabolic derangements were noted in all animals. Serum glucose decreased (from 94.7 zyxw   12.7 mgidl to 22.5 zyxwvut   4.31 mgidl; p < 0.021, ammonia in- creased (from 64 +- 6 +g/L to 163 k 8 kg/L; p < 0.051, lactate increased (from 1.91 +- 0.3 mmol/L to 11.3 t 1.4 I I I I 1 .o 0 8 L w .- 0 6 c V E 0 4 E . o 0.2 0 0 I I I I I I 0 10 20 30 40 50 60 70 time, minutes I I I I I I I 0 10 20 30 40 50 60 70 time, minutes FIG. 3. Glucose and albumin in the extrafiber space of the hollow-fiber module. (A) Vertical axis is glucose concentration nor- malized for the reservoir glucose concentration. (B) Vertical axis is albumin concentration normalized for the reservoir albumin concen- tration. Open symbols represent sealed-flow geometry and filled symbols represent transflow hydraulic geometry. Each symbol type represents the axial flow rate: 10 cm3/min (01, 30 cm3/min 0) and 50 cm3/min V). Each data point represents the mean from two inde- pendent experiments. mmol/L; p < 0.05) and blood pH decreased (from 7.34 .07 to 7.02 k 0.05; p < 0.05). Systolic arterial pressure had decreased from 110 2 15 mm Hg to 40 t 10 mm Hg 5 hr after devascularization (p < 0.02).  262 y OZGA ET AL. zyxwvus 25 20 zyx   zyxwv   zyxwvu   n .- 15- HEPATOLOGY ebruary 1993 I I I I z   theory, shell sealed experiment, shell sealed . xperiment, shell open - 0 theory, shell open J vs. Q for glucose transport J vs. Q for albumin transport 25 2 c .- E zyxwvutsrqp 6 15 n r’ 2 10 E a - .c u L u 5 theory, shell sealed experiment, shell sealed v theory, shell open . xperiment, shell zyxwvutsrq 0 10 20 30 40 50 60 3 Ao axial flow, cm /min I - / nL t50 10 20 30 40 50 6 3 axial flow, cm /min FIG. 4. Trans-fiber fluid flux zyxwvutsrq J s a function of axial flow CQ as determined by experiment (open symbols) and predicted from theory (dark symbols) for glucose zyxwvutsrqp A) and albumin Bi. Both sealed-flow (circles) and transflow (triangles) geometries are shown. Linear approximations are shown for each data set: linear approximations to the experimental data are extended to the vertical axis with dotted lines. ’””” 88 - X 2 HS WITHOUT 0“. I ’ . I ”. I-. ‘’ 46 48 58 52 54 Time (rnin.) FIG. 5. Cumulative increases in metabolites of cyclosporine during 8 hr of culture with microcarrier-attached rat liver cells maintained in a hollow-fiber module. Group X includes the major first-generation monohydroxylation metabolites of cyclosporine 1 and 1); roup Y includes the major N-demethylation metabolite of cyclosporine 20). S = internal standard. In addition, we saw a progressive, significant increase in xenogeneic pig liver cells produced several significant serum LDH and AST after hepatic devascularization; at changes. As shown in Table 1, after 6 hours of treatment 5 hr, LDH was 456 122 mIU/ml and AST was blood ammonia and serum lactate levels were signifi- 527 162 mU/ml. cantly lower and serum glucose level was significantly higher in both groups compared with control animals. In to BAL, systems inoculated with cryopreserved addition, the experimental groups of animals had sig- microcarrier-attached allogeneic dog liver cells and nificantly higher mean systolic blood pressures com- In Viuo BAL Animal Support. Attachment of animals
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks