Artificial Liver Support in the Third Millennium

Abstract

Analogous to the artificial kidney there is a need for an effective and safe liver support system to bridge patients with hepatic failure to liver transplantation or own liver regeneration. An overview is given of the biological and non‐‐biological systems used in clinical practice in the past and at present. The conclusion is drawn that only the biological systems might have the potential to prolong life significantly in patients with acute liver failure. The systems with this potential are summarised. Both in Europe and the USA good bioreactors are available. Most of them are based on porcine hepatocytes, which have immunological and zoonotic drawbacks. What is missing is the well differentiated human hepatocyte in sufficient amounts. Successful development of this cell will be the crown on bioartificial liver research in the third millenium.

• Bioartificial liver
• Acute liver failure
• Immortalized cells
• Hepatocyte

Background

Analogous to the artificial kidney there is a need for an effective liver support system to bridge patients with hepatic failure to liver transplantation or to own liver regeneration. In addition, artificial liver support might improve quality of life in patients with end stage liver cirrhosis. It is thinkable that chronic intermittent treatment of cirrhotic patients by such a device will reduce symptoms of tiredness, itching and hemorrhagic diathesis and will improve jaundice and severity of hepatic encephalopathy. Attempts to reach these goals by haemodialysis, haemoperfusion or plasma exchange have been disappointing, although incidentally some improvements have been obtained ((Sussman, [1996])). This was reason to develop better techniques of dialysis or plasma purification devices like diabsorption ((Biologic DT)), ((Ash et al., [1992])) or molecular adsorbance recirculation systems like MARS ((Stange et al., [1999])). Since, in contrast to the kidney, the detoxification activity of the liver, is in particular focussed at protein‐‐bound and lipid soluble compounds these newer dialysis techniques have included albumin adsorption in their system. However, it is generally assumed that a real effective artificial liver should be based on the capacity to perform the liver's multiple synthetic and metabolic functions, including detoxification and excretion ((Sussman et al., [1994])). Hybrid bioartificial systems based on the presence of active functioning hepatocytes in an extracorporeal device that can be connected to the circulation of the patient with liver failure are promising to meet these criteria. The heart of such a system consists of a bioreactor with a sufficient amount of well nourished and oxygenated viable hepatocytes immobilised on a mechanical support ((Figure 1)).

Figure 1. Schematic drawing of the extracorporeal circuit indicating the plasmaferesis device and the BAL perfusion unit.

This paper summarises the different used bioartificial systems at present and their in vivo results in patients. For results of the more “classical” blood or plasma purification systems like cross circulation, exchange blood transfusion, plasma exchange, peritoneal dialysis haemodialysis and haemoperfusion the reader is referred to literature All of them have been used in the past to treat acute or subacute liver failure. None of them have shown in a controlled trial a significant improvement of survival ((Chamuleau, [1979]; O'Grady et al., [1988]; Riordan and Williams, [2000]; Sussman, [1996]; Tygstrup et al., [1997])). The same holds for the use of isolated allogenic or xenogeneic isolated liver perfusion ((Sussman, [1996])).

Bioartificial Liver Support Systems

Extracorporeal isolated allogenic or xenogeneic liver perfusion has been used incidentally, but no controlled trials are available.

Most bio‐‐artificial systems use as a biocomponent viable hepatocytes, either freshly isolated primary hepatocytes or a hepatoma cell line.

Several variations of hepatocyte bioreactors have been developed and can be classified according to their culture technique. The hepatocytes may be:

1. encapsulated in gels

2. immobilised on collagen coated flat plates

3. cultured in‐‐or outside semipermeable hollow‐‐fibers ((directly or attached to microcarriers)).

4. cultured within a porous three dimensional matrix, allowing direct plasma perfusion conditions of the cells.

Ad 1. Encapsulation of liver cells has been successful by the use of alginate. Besides direct cell transplantation of encapsulated hepatocytes in the peritoneal cavity of experimental animals ((Sarkis et al., [1998])), bioartificial livers for extracorporeal use have been developed ((Dixit and Gitnick, [1995])). This development is still in the experimental phase and proof of principle in experimental acute liver failure has to be obtained before clinical application will be justified.

Ad 2. In 1988 Uchino ((Uchino et al., [1988])) published his BAL based on freshly isolated canine hepatocytes attached to collagen coated glassplates. Application of this BAL to anhepatic dogs showed prolonged survival in a series of 4 dogs ((Table 1)). As we know now, a too small number of animals to obtain significancy due to the large biological variation of such a model.

Table 1. Bioartificial liver support in large animals with acute liver failure

Author	Exp model, number of animals	Hybrid bioartificial system	Hepatocytes, number of cells	Survival ((hrs)) treated versus controls
Uchino et al., (([1998]))	Anhepatic dog n == 4	Collagen coated glassplates	Canine 6 billions	T: 55 ± 10.3 C: 21.3 ± 5.6
((Kelly et al., [1992]))	Anhepatic dog n == 3	Hollow fibre	C3A human 200 gram	not studied
Rozga et al., (([1993]))	Dog, liver ischemia n == 7	Hollow fibre microcarrier attached, charcoal column	porcine 6 billion	not studied
Sielaff, et al., (([1995]))	Dog, galactosamine n == 5	hollow fibre	porcine 6 billion	T: 47 hrs C: 36 hrs
Jauregui et al., (([1995]))	Rabbit, galactosamine n == 11	hollow fibre	rabbit 560 million	T: 4 out of 11 C: 1 out of 9
Gerlach, (([1996]))	Pig, anhepatic n == 9	hollow fibre		not studied
Sheil et al., (([1998]))	Pig, liver ischemia n == 5	hollow fibre		not studied
Flendrig et al., (([1999b]))	Pig, liver ischemia n == 8	non woven matrix	10–14 billions	T: 51 ± 3 C: 33 ± 3
Naruse et al., (([1998]))	Pig liver ischemia n == 4	non woven matrix	10 billions	T == 35.7 ± 2.6 C == 28.5 ± 2.6

In addition his device was not very practical and no follow‐‐up of clinical application has been published until now.

Ad 3. The principle of the hollow fibre based artificial kidney has been so far the most successful. Blood or plasma is passed through the intracapillary lumen while hepatocytes are attached to the outside of the semipermeable fibre membrane. Interaction between the cells and the plasma of the patient can occur through the semipermeable membrane ((Dixit, [1998])). The Gerlach type of hollow fibre based BAL uses a network of fibres including capillaries meant for oxygenation of the cells at site ((Gerlach, [1996])). Others ((Rozga et al., [1993])) use an extra oxygenator of the plasma before entering the BAL. Besides attachment of the liver cells to the outside of the hollow fibres, cells can also be cultured inside the lumen by using collagen gel as a contracting cell matrix. Plasma is now flowing along the outside of the hollow fibre ((Nyberg and Shatford, [1993])). However, all these hollow fibre devices have the disadvantage of indirect contact between plasma and liver cells in contrast to the in vivo plasma perfusion of the hepatocytes by the sinusoids and space of Disse.

Again significant life‐‐prolonging efficacy in experimental animal models of acute liver failure has only been shown by a minority of them as is shown in Table 1. Not all studies have looked at the effect on survival and not all studies have used a sufficient number of animals to allow valid statistical evaluation.

Different experimental animal models of sever acute liver failure have been used: the so called surgical models like complete liver ischemia or total hepatectomy or the chemical ones like galactosamine, acetaminophen or dimethylnitrosamine intoxication.

Ad 4. Others ((Naruse et al., [1998])), including ourselves ((Flendrig et al., [1999a])), have used quite a different way of cell attachment in the bioreactor. Cells are attached to a non‐‐woven poly‐‐ester fabric as a matrix. This matrix is either enclosed as different pieces in a cartridge or as a spirally wound tissue. In the AMC‐‐BAL the space between the different layers is filled with polypropylene oxygen fibres to provide at site an optimal oxygenation of the cells inside the bioreactor.

The big advance of these types of BAL's is the direct contact between plasma and the cells like the in vivo situation. Furthermore, this type of devices can easily contain a very large number of cells ((in the AMC‐‐BAL more than 20 billion of cells)), in contrast to most hollow fibre ones ((with as a positive exception the Berlin type)).

The AMC‐‐BAL was the first to show significant prolongation of survival both in small ((Flendrig et al., [1999a])) and large animals with severe hepatic failure ((Table 1)) ((Jauregui et al., [1995]; Kelly et al., [1992]; Sheil et al., [1998]; Sielaff et al., [1995])).

Several of the above mentioned BAL's have been tested in the clinic; mostly in uncontrolled studies. The only exception is the pilot controlled study of Ellis et al ((Ellis et al., [1996])), which uses a hollow fibre BAL filled with a human hepatoma cell line.

The results are shown in Table 2 ((Donini et al., [2000]; Ellis et al., [1996]; Kardassis et al., [1999]; Patzer et al., [2000]; Sussman et al., [1994]; Watanabe et al., [1997])).

Table 2. Clinical results of BAL treatments

	Number of patients
Uncontrolled studies ((phase 1))
Sussman et al, (([1994]))	11
Watanabe et al, (([1997]))	26
Kardassis et al, (([1999]))	8
Patzer et al, (([2000]))	4
Donini et al, (([2000]))	1
Chamuleau et al ((unpublished))	3
Controlled study
Ellis et al, (([1996]))	12 out of 24

Larger controlled trials are under way by the Demetriou BAL and the AMC‐‐BAL.

Future

It is well realised that the use of primary porcine hepatocytes in the BAL has major disadvantages: first, there are enormous immunological problems, even life threatening, to be anticipated by prolonged ((more than 1 week)) use, because of the risk of severe anaphylactic reactions of the recipient to porcine antigens like “serum sickness” ((Te Velde et al., [1997](( and second, the risk of transfection of porcine endogenous retroviruses ((PERV)) to the human population can not completely be excluded ((Patience et al., [1997])). However, until now no direct evidence of such a transfection has been obtained ((Paradis et al., [1999])).

The ideal solution will be the availability of sufficient amounts of viable well differentiated non‐‐tumorigenic human liver cells. In different centers research to obtain these cells is in progress. Potential approaches are genetic engineering of adult liver cells ((Kobayashi et al., [2000])) or manupilation of pluripotent stem cells in the direction of liver cells ((Kubota and Reid, [2000])).

At present, as long as these cells are not yet available, porcine hepatocytes are second best. Clinical immunological experience with xeno‐‐bioreactors in a short‐‐term ((less than 1 week)) phase 1 study in patients has, so far, been reassuring ((Baquerizo et al., [1997])).

Unsolved Problems

Characteristic for all BAL‐‐systems is the lack of a bile excretory system. We have shown ((Flendrig et al., [1997])) that porcine hepatocytes cultured in our bioreactor form bile canaliculi‐‐like structures. This implies that conjugated bile‐‐acids and bilirubin can be excreted into the circulation during in vivo application. This may result in high levels of bile‐‐salts, which might limitate the long term function of the hepatocytes in the bioreactor ((Pazzi et al., [1997])). A possible solution could be the introduction of anion exchange resins and//or silica‐‐based strong anion exchange cartridges for selective adsorption of bile acids and bilirubin ((Morimoto et al., [1989])).

As already indicated above the use of a xenogeneic hepatocyte source for the clinical application of a BAL may present cellular and humoral immunological problems and the risk of PERV transmission. The first can be reduced to an acceptable proportion by using such a BAL not longer than 1 week. The second might be made quite acceptable by using viral reducing techniques in the circuit like UV treatment of the returning plasma or virus‐‐filters ((Nyberg et al., [1999])).

The best solution of the above mentioned problems will be successful development of sufficient amounts of a well differentiated non‐‐tumorigenic human liver cell line.

Good bioreactors are available from the second millennium; the third millennium will give us the right liver cells!