Critical issues in bioartificial liver development.

End-stage liver disease accounts for over 30,000 deaths annually in the United States. Orthotopic liver transplantation is the only clinically proven treatment for patients with end-stage liver failure. A limitation of this therapy is a shortage of donor organs available. This donor organ shortage is exacerbated by the fact that the number of patients listed for transplantation has continued to increase. As a result, there has been a continuing increase in the number of patients who die waiting for a donor liver. Extracorporeal bioartificial liver devices consisting of viable hepatocytes have the potential to provide temporary support for patients with fulminant hepatic failure, thereby serving as a "bridge" to transplantation. In some patients, this temporary support would allow the native liver to regenerate function, eliminating the need for transplantation and the resulting life-long immunosuppressive therapy, all of which translates into a cost savings to the health care system. Although the bioartificial liver device is a promising technology for the treatment of liver failure, significant technical challenges remain in order to develop systems with sufficient processing capacity and of manageable size. An overview of the critical issues in the development of bioartificial liver devices is discussed.

Effect of flow on the detoxification function of rat hepatocytes in a bioartificial liver reactor.

Ethoxyresorufin-o-deethylation (EROD) can be used as a sensitive measure of hepatic detoxification function. In this study, we employed a fluorescence assay based on EROD to study the effect of varying Peclet number (or flow) on hepatic function in a microchannel flat-plate bioartificial liver (BAL) reactor containing a coculture of hepatocytes and fibroblasts. Static culture and reactor flow experiments established that: 1) a pseudo-steady-state detoxification rate could be attained at each Peclet number, 2) the steady-state detoxification rate increased nonlinearly with Peclet number (ranging from 167 to 2500), 3) the uptake rate of substrate was a linear function of cell surface substrate concentration (<1 microM), and 4) a shear stress of 10 dyne/cm2 did not adversely affect hepatic function for at least 12 h. A convection-diffusion-reaction model supports the conclusion that increased convective mass transfer of substrate to the cell surface is the primary cause of the observed increase in EROD rate with Peclet number. Our results suggest that detoxification rates can be enhanced by an order of magnitude by choosing an appropriate Peclet number. For our bioreactor configuration, this optimum corresponds to a Peclet number range of 1000-2000 at a Damkohler number of 0.55. The usefulness of the mathematical model is discussed in the context of scale-up to a clinical BAL reactor for human application.

In vitro and in vivo evaluation of albumin synthesis rate of porcine hepatocytes in a flat-plate bioreactor.

Several configurations of extracorporeal bioartificial liver devices have been developed for the potential treatment of fulminant hepatic failure or as a bridge to liver transplantation. Recently, we developed a microchannel flat-plate bioreactor with an internal membrane oxygenator in which porcine hepatocytes are cultured as a monolayer on the bottom glass surface. In the present study, we investigated synthetic function of porcine hepatocytes in the bioreactor in both in vitro and in vivo flow circuit models. In vitro, albumin synthesis was stable in the bioreactor for up to 4 days of perfusion. In vivo, with the extracorporeal connection of the bioreactor to rat vasculature, porcine albumin was detectable for 24 h in the rat plasma. We also developed a simple mathematical model to predict the in vivo porcine albumin concentration in rat plasma. These results indicate that this configuration of a microchannel flat-plate bioreactor has potential as a liver support device and warrants further investigation.

Effects of oxygenation and flow on the viability and function of rat hepatocytes cocultured in a microchannel flat-plate bioreactor.

The goal of this study was to investigate the viability and synthetic function of rat hepatocytes cocultured with 3T3-J2 fibroblasts in a small-scale microchannel flat-plate bioreactor with and without an internal membrane oxygenator under flow. Bioreactor channel heights ranged between 85 and 500 microm and medium flow rates ranged between 0.06 and 4.18 mL/min. The results showed that the bioreactor without the oxygenator resulted in significantly decreased viability and function of hepatocytes, whereas hepatocytes in the bioreactor with internal membrane oxygenator were able to maintain their viability and function. The shear stress calculations showed that, at lower wall shear stresses (0.01 to 0.33 dyn/cm(2)), hepatocyte functions, measured as albumin and urea synthesis rates, were as much as 2.6- and 1.9-fold greater, respectively, than those at higher wall shear stresses (5 to 21 dyn/cm(2)). Stable albumin and urea synthesis rates for 10 days of perfusion were also demonstrated in the bioreactor with internal membrane oxygenator. These results are relevant in the design of hepatocyte bioreactors and the eventual scaling-up to clinical devices.

Analysis of oxygen transport to hepatocytes in a flat-plate microchannel bioreactor.

Oxygen transfer to cultured hepatocytes in microchannel parallel-plate bioreactors with and without an internal membrane oxygenator was investigated with a mathematical model and the results were corroborated with fluorescence imaging experiments. The consumption of oxygen by hepatocytes was assumed to follow Michaelis-Menten kinetics. Our simulations indicate that under conditions of low Péclet (Pe) number (<80) and fixed Damlkohler number (= 0.14, corresponding to rat hepatocytes) the cells are hypoxic in the bioreactor without an internal membrane oxygenator. Under the same conditions, the bioreactor with an internal membrane oxygenator can avoid cell hypoxia by appropriate selection of membrane Sherwood number and/or the gas phase oxygen partial pressure, thus providing greater control of cell oxygenation. At high Pe, both bioreactors are well oxygenated. Experimentally determined oxygen concentrations within the bioreactors were in good qualitative agreement with model predictions. At low Pe, cell surface oxygen depletion was predicted from analytically derived criteria. Hepatocytes with oxygen dependent functional heterogeneity may exhibit optimal function in the bioreactor with the internal membrane oxygenator.

Conditions for the occurrence of large near-wall excesses of small particles during blood flow.

Prior work showed that the near-wall concentration of platelet-sized latex beads (2.38 microns diam) in flowing blood suspensions can be greater than three times the concentration in the central region of the flow. Similar methods were used to explore the dependence of the near-wall excess (NWE) of beads on the channel height and suspension composition. The bead diameter, suspending fluid viscosity, and red blood cell deformability were varied; the hematocrit was fixed at 15%. Results showed that NWEs greater than or equal to three times the central concentration were associated with shear stress, rather than with strain rate, required red cell deformability, and occurred with bead diameters of 2.2 microns or larger. The amplitude of NWEs observed in the 30- and 50-microns channels changed sharply from small to large as the wall shear rate (WSR) was increased, while those observed in 100-microns channels exhibited a more gradual dependence on WSR.

The near-wall excess of platelet-sized particles in blood flow: its dependence on hematocrit and wall shear rate.

Methods involving microscopy were used to obtain concentration profiles of platelet-sized beads during flow through glass channels. Suspensions of fluorescent latex beads (2.38 microns diam) and washed red blood cells were made from an isotonic albumin-dextrose solution. A syringe pump regulated the suspension flow through glass channels, which were either 50 or 100 microns wide; most experiments used a wall shear rate of 1630 sec-1. Via stroboscopic epifluorescence microscopy, photographs were collected on image planes parallel to the channel wall. Profiles of the bead concentration in the narrow channel direction were made by assembling counts of the focused bead images in the photographs. The results showed that a near-wall excess of the beads occurred when the suspension contained a significant fraction of red cells (over 7%). For hematocrits of 15 to 45% (the highest studied), the excess was above five times the concentration in the central region. Experiments with channels of both widths showed the region of excess beads was 5 to 8 micron thick. A series of experiments with 50-micron channels, with a suspension hematocrit of 15%, and with wall shear rates from 50 to 3180 sec-1 showed that near-wall excesses were large only for wall shear rates of 430 sec-1 and above. This work demonstrated the effects of wall shear rate (flow rate) and hematocrit on the number of platelet-sized beads near a surface and hence illustrated physical (rheological) factors that act in blood-surface interaction.

Transport of platelets in flowing blood.

Distribution and transport of platelets in flowing blood were studied experimentally using suspensions of washed red cells and fluorescent latex beads as platelet analogues. Distributions of the platelet analogues were obtained from stroboscopic epifluorescence photomicrographs of flow in 50-micron channels and from images of the cut cross sections of cryogenically frozen thin-walled 200-micron tubes. Concentration profiles of platelet analogues had a substantial near-wall excess for situations with a substantial hematocrit (greater than 10%) and a substantial wall shear rate (greater than 400 s-1). The viscosity of the suspending fluid was found to affect the size of the near-wall excess and its shear-dependent onset. Additionally, the shear-rate dependence of the near-wall excess did not occur with suspensions of hardened red cells. The excess extended a substantial distance from the wall in the 200-micron tubes and a portion of the profile could be fitted to an exponential curve. The random walk model that is used to describe enhanced platelet diffusion is envisioned as a walk (lateral platelet motion) caused by shear-induced collisions with red cells. A more comprehensive random walk model that includes biased collisions produces an effective lateral motion of convective nature in addition to a diffusional motion; it is used to explain the observed nonuniform distributions of platelet analogues.