WO1994015459A9 - Method for isolation and transplantation of specialized hepatocytes - Google Patents

Abstract

Disclosed is a method of hepatic transplantation. The method includes providing a sample of hepatocytes enriched for specialized hepatocytes, and transplanting the sample of hepatocytes into a mammal in need of hepatic transplant. The invention also includes providing a protein to a mammal by (a) providing a sample of hepatocytes enriched for specialized hepatocytes; (b) transfecting the specialized hepatocytes with DNA encoding the protein; and (c) transplanting the transfected hepatocytes into a mammal.

Description

METHOD FOR ISOLATION AND TRANSPLANTATION OF

SPECIALIZED HEPATOCYTES

Background of the Invention

The invention relates to the field of hepatocyte transplantation and human gene therapy.

One of the most important characteristics of liver tissue is its ability to regenerate (see, e.g., Leffert et al., Gastroenterology 76: 1470, 1979; Harvey et al, eds, The Principles and Practice of Medicine , Twenty- first edition, Appleton-Century-Crofts, Norwalk, Co, 1984). When liver cells are damaged by toxins, by interference with blood supply, or by obstruction to biliary flow, the remaining cells rapidly regenerate.

The exact stimulus for regeneration is unknown, but several studies suggest that humoral agents are

responsible. For example, insulin-like growth factor (IGF), glucagon, and epidermal growth factor are believed to be trophic substances for liver regeneration. More recently, hepatocyte growth factor (HGF) has been

identified as a potent mitogen for mature parenchymal hepatocytes and may be a hepatotrophic factor that acts as a trigger for liver regeneration after partial

hepatectomy and liver injury (see, e.g., Nakamura et al Nature 342: 440, 1989). In the rat, surgical removal of two-thirds of the functioning liver is followed by nearly complete restoration in a matter of days. Similar regeneration occurs in humans subjected to partial hepatectomy following trauma or in an attempt to

eliminate disease.

There is, however, a great need for methodologies to compensate for liver failure by the transplantation of cells, either to compensate for inherited deficiencies of liver function or to alleviate acquired liver

dysfunction. The present invention provides a useful method for isolating and utilizing specialized

hepatocytes not only for hepatic transplantation but also for cellular transplantation and human gene therapy.

Summary of the Invention

In general, the invention features a method of hepatic transplantation, involving providing a sample of hepatocytes enriched for specialized hepatocytes, and transplanting the sample of hepatocytes into a mammal in need of hepatic transplant. The sample of hepatocytes may be derived from self or non-self. In preferred

embodiments, the specialized hepatocytes express a plasma protein, e.g., albumin, transferrin, complement component C3, α₂-macroglobulin, fibrinogen, Factor XIII:C, Factor IX, α₁-antitrypsin or α-fetoprotein; a metabolic

protein, e.g., ornithine transcarbamylase,

arginosuccinate synthetase, glutamine synthetase,

glycogen synthetase, glucose-6-phosphatase, succinate dehydrogenase, glucokinase, pyruvate kinase, acetyl CoA carboxylase, fatty acid synthetase, alanine

aminotransferase, glutamate dehydrogenase, ferritin or alcohol dehydrogenase; a membrane protein, e.g., GLUT-1, or LDL receptor; or a structural protein, e.g.,

cytokeratin #8, or cytokeratin #18. In yet other

preferred embodiments, the specialized hepatocytes produce a carbohydrate, e.g., glycogen, glucose, or glucose-6-phosphate; a lipid, or a metabolite, e.g., urea, or glutamine. The sample of specialized

hepatocytes may be either homogeneous or heterogeneous.

In preferred embodiments, the specialized hepatocytes express a molecule, e.g., a protein, a carbohydrate, a lipid, a metabolite, capable of

preventing or treating an inherited or acquired disease.

In a second aspect, the invention features a method of providing a protein to a mammal, involving: (a) providing a sample of specialized hepatocytes; (b) transfecting said hepatocytes with DNA encoding the protein; and (c) transplanting the transfected

hepatocytes into the mammal.

In preferred embodiments, the method is used for treating an inherited, an acquired or metabolic

deficiency. For example, the hepatocyte may be

transfected with DNA encoding Factor VIII:C, Factor IX, α₁ antitrypsin, or low density lipoprotein receptor useful for treating human diseases such as hemophilia A and B, α₁ antitrypsin deficiency, and familial

hypercholesterolemia, respectively.

By the term "specialized hepatocyte" is meant a hepatocyte that produces a greater quantity of a

molecule, e.g., a protein, a carbohydrate, a lipid, or a metabolite, than the majority of hepatocytes. For

example, a specialized hepatocyte may express a

significant amount of a plasma protein (see, e.g.,

Putnam, The Plasma Proteins , Volumes I, II, and III, Academic Press, 1975), a metabolic protein (see, e.g., Jungermann et al., Trends in Biochemical Sciences

September, 1978, 198-202), a membrane protein (see, e.g, Tal et al., Endocrinology 129: 1933, 1991) or a

structural protein (see, e.g., Trevor et al., J. Biol . Chem . 260:15885, 1985). Thus, a specialized hepatocyte produces at least 10%, and preferably more than 50%, and even more preferably 90% of a particular protein, e.g., albumin, relative to a non-specialized hepatocyte. In addition, a specialized hepatocyte may also significantly express two or more proteins, e.g., albumin and

fibrinogen.

By the term "enriched" is meant that a population of hepatocytes, either homogeneous or heterogeneous, is composed of at least 75% specialized hepatocytes, preferably greater than 90%, more preferably 95% and even more preferably 99% specialized hepatocytes, e.g., hepatocytes expressing a plasma protein.

By the term "self" is meant that the hepatocytes used for transplantation, i.e., the donor cells, are derived from an autologous source, i.e, the host.

By the term "non-self" is meant that the hepatocytes used for transplantation are derived from a non-autologous source.

By the term "homogeneous" is meant that the hepatocyte sample is composed solely of a single

specialized hepatocyte, e.g., a population of albumin- expressing hepatocytes.

By the term "heterogeneous" is meant that the hepatocyte sample is composed of two or more specialized hepatocytes, e.g., a population that consists of albumin- and transferrin-expressing hepatocytes.

The invention provides methods for the treatment of liver disorders that do not require the administration of drugs or blood-derived therapeutics. The invention affords unique and critical advantages over any previous method. Specifically, the invention provides for treating any number of liver diseases, including, e.g., familial hypercholesterolemia, hemophilia, and cirrhosis, by transplantation of specialized hepatocytes that are capable of providing a molecule, e.g., a protein, a carbohydrate, a lipid, or a metabolite, necessary to prevent or treat a disease. Transplantation of

subpopulations of high-expressing hepatocytes should permit for the amplification of a desired therapeutic effect. Furthermore, the specialized hepatocytes may be useful as an in vivo recombinant protein delivery system. The specialized hepatocytes, especially plasma-protein expressing hepatocytes, may have both extraordinary secretory capabilities and an exceptional capacity for cell division (and therefore have a greater long-term survival in comparison to other hepatocytes). These characteristics are important features for the

development of a cellular transplantation system that can not only stably produce and secrete endogenous proteins necessary for the treatment of disease but also produce and secrete recombinant proteins into the circulatory system. Finally, the invention is intended to obviate problems commonly encountered in hepatic transplantation; for example, the limited supply of intact donor organs, the added immunological consequence of rejection against non-hepatocyte cellular elements, and the avoidance of the danger of surgical mortality associated with liver transplantation.

Other features and advantages of the invention will be apparent from the following description of the preferred embodiments and from the claims.

Description of the Preferred Embodiment(s) The drawings are first briefly described.

Drawings

Fig. 1 is a panel of micrographs showing the immunofluorescence of mouse liver.

Fig. 1a is an immunofluorescent micrograph showing an albumin-containing hepatocyte (green) identified by immunofluorescence with rabbit anti-mouse albumin.

Fig. 1b is an immunofluorescent micrograph showing a complement component C3-containing hepatocyte in same field (red) identified by goat anti-mouse complement component C3.

Fig. 1c is an immunofluorescent micrograph showing a double exposure of same fields as Figs. 1 a and 1b, showing simultaneously albumin (green) and C3 (red) containing cells.

Fig. 1d is an immunofluorescent micrograph showing a cluster of albumin-containing hepatocytes. Fig. 1e is an immunofluorescent micrograph from an Alb^a/Alb^c heterozygous mouse showing the reactivity of rabbit anti-mouse albumin (red).

Fig. 1f is an immunofluorescent micrograph from an Alb^a/Alb^c heterozygous mouse showing the reactivity of anti-Alb^c (green).

Fig. 2 is an illustration of the percentage of hepatocytes containing various plasma proteins from the livers of mice at various ages. Identification of Specialized Hepatocytes

Experimental Data

Material and Methods

Animals. Mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Rats were obtained from the Charles River Laboratories (Wilmington, MA).

Immunofluorescence. Mice and rats were sacrificed by CO₂ inhalation. In most experiments, the liver was perfused by passing phosphate buffered saline (PBS) into the left ventricle of the heart for approximately 3 minutes, or until the majority of blood was removed from the liver, turning it from a dark crimson to a beige color (10-15 ml for mice, 200 ml for rats). The perfused liver was then frozen on a block of aluminum on a bed of dry ice. Livers were sectioned on a Reichert-Jung 1800 Figocut E (Cambridge Instruments) at -17°C at a thickness of four microns. Two-color immunofluorescence was performed on each section following a 15 minute fixation in acetone and three washes in PBS, five minutes each.

A mixture of rabbit and goat antisera in PBS was added to each section. Sections were incubated at room temperature in a damp chamber for 45 minutes, then washed four times in PBS, five minutes each. TRITC-conjugated swine anti-rabbit antiserum (Nordic Immunological

Laboratories, Capistrano Beach, CA) and FITC-conjugated donkey anti-goat IgG antiserum (Nordic Immunological Laboratories, Capistrano Beach, CA) were mixed together in a 1/30 mixture with 1% fetal calf serum in PBS. Each section was covered with 15 microliters of this solution and left at room temperature for 45 minutes in a damp chamber. Following three washes in PBS, the slides were coverslipped (mounting solution: 1 ml PBS, 9 ml

glycerine, 10 mg ρ-phenylaminediamine).

Antisera. Rabbit anti-mouse albumin (Nordic Immunological Laboratories, Capistrano Beach, CA, Cat. Code RAM/Alb, lot 14-787); sheep anti-mouse albumin

(Cappel Research Products, Durham, NC, Cat. #0111-0344, lot 26599); rabbit anti-mouse α-fetoprotein (Miles Inc. Diagnostics Division, Kankakee, IL, Cat. Code 645611;

Allen et al., Cancer Res . 37:697, 1977); rabbit anti- mouse transferrin (ICT Inter-cell Technologies Inc., Hopwell, NJ, Cat. #A3240, lot 0138); goat anti-mouse transferrin (Cappel Research Products, Durham, NC, Cat. #0111-1441, lot 26948); rabbit anti-mouse C3 (Janssen Biochimica, Westbury, NY, Cat. #KM-1307-PO, lot 0022); goat anti-mouse C3 (Cappel Research Products, Durham, NC, Cat. #0111-0601, lot 27087); goat anti-mouse C3 FITC conj. (Nordic Immunological Laboratories, Capistrano Beach, CA, Cat. Code GAM/C3/FITC); rabbit anti-mouse fibrinogen (Research Plus Cat. #019226-D9 and Janssen Biochimica, Westbury, NY, Cat. #2394785, lot KM18-01- POl); goat anti-mouse fibrinogen (Nordic Immunological Laboratories, Capistrano Beach, CA, Cat. Code GAM/Fbg, lot 3238); goat anti-mouse α₂-macroglobulin (Nordic

Immunological Laboratories, Capistrano Beach, CA, Cat. Code GAM/A₂M, lot 14-186); rabbit anti-rat albumin

(Cappel Research Products, Durham, N.C., Cat. #0113-0342, lot 0113-0342); sheep anti-rat albumin (Cappel Research Products, Durham, NC, Cat. #0113-0344, lot 32110); rabbit anti-rat transferrin (Cappel Research Products, Durham, NC, Cat. #0113-1442, lot 32454 and Nordic Immunological Laboratories, Capistrano Beach, CA, Cat. Code RARa/Tfn, lot 353); goat anti-rat transferrin (Nordic Immunologic Laboratories, Capistrano Beach, CA, Cat. Code GARa/Tfn, lot 42-285); rabbit anti-rat C3 (Bethyl Laboratories Inc., Montgomery, TX, Cat. #A110-120, lot A110-120-3); goat anti-rat C3 (Cappel Research Products, Durham, NC, Cat. #0113-060, lot 30348); rabbit anti-rat fibrinogen (Accurate Chemical & Scientific Corp. Cat. #2397011, lot 18-01-P02); goat anti-rat fibrinogen (Cappel Research Products, Durham, NC, Cat. #0113-0821, lot 0113-0821); sheep anti-rat α-fetoprotein (Nordic Immunological

Laboratories, Capistrano Beach, CA, Cat. Code ShARa/AFP, lot 3384); swine anti-mouse rabbit IgG FITC conj. (Nordic Immunological Laboratories, Capistrano Beach, CA, Cat. Code SWAR/FITC, lot 3131); donkey anti-mouse goat IgG TRITC conj. (Nordic Immunological Laboratories,

Capistrano Beach, CA, Cat. Code DoAG/TRITC, lot 13-683).

In addition, antibodies useful for identifying any protein expressed by a specialized hepatocyte may be prepared according to standard techniques (see, e.g., Ausubel et al. Current Protocols in Molecular Biology, John Wiley & Sons, New York, 1989). For example,

monoclonal antibodies may be prepared using any liver protein, e.g., estrogen receptor, glutamine synthetase, carbamoylphosphate synthetase, produced by a specialized hepatocyte and standard hybridoma technology (see, e.g., Kohler et al., Nature 256: 495, 1975; Kohler et al. Eur. J. Immunol . 6: 511, 1976; Kohler et al., Eur. J. Immunol . 6: 292, 1976; Hammerling et al. In: Monoclonal Antibodies and T Cell Hybribomas, Elsevier, NY, 1981; Ausubel et al. supra). Furthermore, polyclonal antibodies can, if desired, also be produced (see, e.g., Garvey et al., Methods in Immunology, W. A. Benjamin, Northampton, MA, 1977). Once produced, monoclonal or polyclonal

antibodies are tested for specific recognition of their antigen by Western blot or immunoprecipitation analysis (see, e.g, methods described in Ausubel et al., supra) .

Absorptions. A mixture of rabbit anti-mouse albumin and goat anti-mouse transferrin was absorbed as follows: 25 microliters of the mix of the two antibodies were combined with a 2.5 microliter volume of protein, at concentrations of 1, .1, .01, .001 and .0001 mg/ml.

These mixtures were incubated at room temperature for 1 hour and left at 4°C overnight. Samples were spun in a microcentrifuge for 2 minutes, and the supernatants were used for immunofluorescence as described above. Residual anti-albumin antibody was assessed by the presence of green cells visualized by FITC anti-rabbit IgG, while anti-transferrin antibody was assessed by the presence of red cells visualized by TRITC labeled anti-goat IgG serum. As both anti-albumin and anti-transferrin

antibodies were mixed before absorption, each test contains an internal negative control.

Cell Quantitation. Cells were counted utilizing simultaneous two-color immunofluorescence of liver sections, as described above. Typically, 10 sections were examined, each section being placed in the clear well of a 7.5 mm diameter black teflon-templated

microscope slide. By examining 35 mm photomicrographs of both hemocytometer grating and a section of mouse liver, we determined that mouse liver contained approximately 1700 cells/mm² and using this, the immunofluorescence cell counts were converted into percentages. For many experiments, slides were photographed with an MP-4

Polaroid camera, the photograph of each well cut out with scissors and the area corresponding to the section cut out. Both parts of the photograph were weighed to estimate the percent of the well filled by the section. Such calculations on many sections from adult livers revealed that on average 70% of the well was filled, with relatively minor variations from this number, which was relied upon as a standard correction factor in many experiments. However, precise estimates of section size using Polaroid photographs were made for every analysis of cell numbers in non-adult mouse livers.

Results

Hepatocytes are Heterogeneous. Immunofluorescence tests on mouse and rat liver were carried out with more than 20 antisera to individual plasma proteins. In every case, each antiserum was found to react with only a small number of hepatocytes (Fig. 1). This narrow reactivity was identified using antisera to albumin, transferrin, complement component C3, fibrinogen, α₂-macroglobulin, and α-fetoprotein in the mouse, as well as with antisera to albumin, transferrin, fibrinogen and α-fetoprotein in the rat. Albumin-containing hepatocytes constitute somewhat less than 1% of the liver, while cells

containing each of the other proteins are present in lower amounts. As a general conclusion, it may be said that the abundance of cells of each type roughly reflects the relative rates of synthesis of the corresponding proteins (Fig. 2), with α-fetoprotein-containing cells most frequently seen in the livers of newborn animals. In contrast, immunofluorescence with other antibodies reveal widespread staining throughout the liver, e.g., carbamoylphosphate synthetase, as discussed below.

The data summarizes two color immunofluorescence experiments, which indicate that each plasma protein is produced in separate hepatocytes (Table I). Typically, the reaction of the rabbit antisera was visualized by the green fluorescence of FITC while the reaction of the goat antisera were revealed by the red fluorescence of TRITC. Two-color immunofluorescence reactions also showed a cellular identity between the hepatocytes identified by rabbit anti-mouse fibrinogen, goat anti-mouse fibrinogen, rabbit anti-rat fibrinogen and goat anti-rat fibrinogen. Other anti-mouse plasma protein antisera showed only partial cross reactivity with their rat counterparts.

See Materials and Methods section for techniques.

Each Plasma Protein Seen by Immunofluorescence is Contained in a Separate Population of Hepatocytes. More than 250 two-color immunofluorescence comparisons were carried out on mouse liver, which demonstrated that each plasma protein is located in separate population of hepatocytes, (see Fig. 1, Table I). Proteins assayed include albumin, transferrin, complement component C3, α₂-macroglobulin, fibrinogen, and α-fetoprotein in the mouse liver. For the rat liver, albumin, transferrin, fibrinogen and α-fetoprotein were examined (Table II).

The quantitative data on the separate nature of each cell type in mice is given in Table III.

A Small Number of Mouse Hepatocytes Appear to Contain More than One Protein Identifiable by Immunofluorescence.

By far, most of the plasma-protein containing cells appear to contain only one protein. However, the two-color immunofluorescence experiments with mouse liver identified a small number of cells that stain positive in both colors (Table III). These two-color cells are present in about 1/100th of the quantity of single-color cells.

Histological Appearance of Cells Containing Plasma

Proteins. Hepatocytes containing each protein are scattered throughout the liver without any obvious regional orientation. Other than containing different proteins, there are no obvious histological differences between each class of hepatocyte in adult liver. Cells containing plasma protein are scattered throughout the liver individually and in clusters (Figure 1d). Serial sections of mouse liver revealed that generally clusters can be followed on adjacent sections; usually these clusters are not large (25 or fewer cells) and have a globular shape. This size and shape is roughly in accord with the estimate of liver clonal size found by West (J. Embryol . Exp. Morpho . 36:151, 1976) from studies of aggregation chimeras.

Hepatocyte Heterogeneity is Evident in Cell

Suspensions of Hepatocytes. The appearance of cellular heterogeneity persists after hepatocytes are isolated from the liver. A single cell suspension of mouse hepatocytes was obtained by collagenase perfusion through the portal vein. Single cells were collected, tested for viability by trypan-blue exclusion, washed and pelleted. The pellet was frozen, sectioned and analyzed by

immunofluorescence; albumin- and transferrin-containing cells were present in approximately the number expected from our experience with intact liver.

Starvation-induced Suppression of Plasma-protein Synthesis Can be Seen by Immunofluorescence. Starvation is known to induce a shutdown of plasma protein synthesis (see, e.g. Morgan et al., J. Biol . Chem . 246:3500, 1971). Over a three-day period, albumin-, transferrin-, and fibrinogen-containing cells fade and disappear from the livers of mice deprived of food. Within a few hours of refeeding these mice, the plasma protein-containing cells come back to their usual appearance, with both single cells and clusters visible in the usual quantities.

Effect of Tissue Fixation on the Appearance of Hepatocyte Heterogeneity Seen by Immunofluorescence.

Most immunofluorescence experiments were carried out on cryostat sections of frozen liver, fixed after sectioning by immersion in acetone. However, the detection of liver cell heterogeneity is not dependent on the fixation step. Experiments carried out on sections of mouse liver without acetone fixation gave essentially the same results as those obtained with acetone-fixed sections, although individual protein-positive cells tended to show a granular rather than an even fluorescence and large parts of the sections detached from the glass, leaving holes and tears.

Examination of liver sections fixed and embedded by conventional methods can be problematic. The

experiments described here demonstrate that fixation by a variety of methods (see, e.g., Saint Marie, J. Hist . Cyt . 10:250, 1962 or Moorman et al., Histochemical Journal 22:653, 1990), followed by paraffin embedding, destroys the antigenicity detectable by a number of anti-albumin and anti-transferrin antisera. In addition, sections of this type showed greatly increased non-specific

background. This effect might account for some of the observations published in the literature that led to the impression that plasma protein expression is widespread.

Effect of Perfusion on the Appearance of

Hepatocyte Heterogeneity. In most experiments, livers were perfused with saline to remove the blood. Some experiments revealed, however, that with respect to the visualization of all the plasma proteins except albumin, the perfusion step could be eliminated without any apparent effect on the outcome. Albumin-containing hepatocytes are readily visualized in unperfused liver, although a markedly increased background is evident.

Increased background fluorescence is most likely due to contamination of the liver by albumin contained in the sinusoids.

Albumin-containing cells are easily visualized over the diffuse albumin background in unperfused livers. This observation can be used to approximate the

concentration of albumin in albumin-containing

hepatocytes. Albumin in blood is present in a

concentration of approximately 40 mg/ml. Its reasonable that hepatocytes contain greater than this concentration; otherwise, these cells would not be visible over the diffuse albumin background. Given that the number of albumin-containing hepatocytes reaches as high as 0.5% in adult mouse liver, and that the concentration of albumin in albumin-containing hepatocytes probably exceeds 40 mg/ml, then it follows that the total amount of

intracellular albumin contributed to the liver as a whole by the albumin-containing hepatocyte would appear to exceed 0.2 mg/gram of liver, a value that is close to the amount of pro-albumin that Morgan et al. supra measured biochemically as 0.4 mg/gm.

Serological Specificity of the Immunofluorescence Reactions. The immunochemical specificity of the

immunofluorescence reactivities was tested by a number of methods. First, immunoelectrophoresis with each anti- mouse plasma protein antiserum showed reactivity with only the expected band. Second, immunoprecipitation with each of these antisera yielded protein on SDS-PAGE with the expected molecular weight. Third, as indicated in Tables I and II, the specificity of each antibody was seen by two-color immunofluorescence experiments

comparing pairs of antisera with the same specificity (e.g., rabbit anti-albumin versus goat anti-albumin);

these studies demonstrated reactivity of both reagents in the same cells. Fourth, directly fluorescenated goat anti-mouse complement component C3 reacts with the same cells identified by rabbit anti-mouse complement

component C3. Fifth, a mouse alloantibody to an

allotypic determinant on albumin that reacts with the same hepatocytes recognized by sheep and rabbit anti- albumin, but does not react with transferrin-containing cells. Sixth, the specificity of the immunofluorescence reactions was confirmed by absorption of antisera with purified proteins (Tables IV and V). The method used was as follows: The absorption experiments were carried out with mixtures of two antisera. Such mixtures of antisera were used to provide each absorption test with an

internal control. For example, a mixture of rabbit anti- albumin and goat-anti-transferrin, when tested unabsorbed against a section of liver, will stain albumin-containing cells green and transferrin-containing cells red. When purified albumin was added to this antibody mixture, this absorption destroyed the capacity of the mixture to stain the albumin-containing cells (as detected by examination of green fluorescence) but not the capacity of the mixture to stain transferrin-containing cells (as

detected by examination of red fluorescence) (Tables IV and V). Conversely, absorption of the same antibody mixture with purified transferrin removed the capacity to stain transferrin-containing cell but not albumin- containing cells. By this approach, the specificity of the immunofluorescence reactions against mouse albumin, mouse transferrin, mouse α-fetoprotein, rat albumin, rat α-fetoprotein, rat fibrinogen and rat transferrin were confirmed. In these experiments, a mixture of rabbit anti-mouse albumin and goat anti-mouse transferrin was absorbed as described in the Materials and Methods section. As both anti-albumin and anti-transferrin antibodies were mixed before absorption, each test contained an internal negative control. The minimal concentration of transferrin required to absorb all anti- transferrin activity was found to be .01 mg/ml, with a general diminution of anti-transferrin reactivity at .001 mg/ml; the minimal concentration of albumin required to remove all anti-albumin reactivity was .1 mg/ml, with some absorption evidence at .01 and .001 mg/ml. "No" absorption indicates no diminution of reactivity even at highest protein concentrations (1 mg/ml) . Source of purified proteins: mouse transferrin #1 (Research Plus), mouse transferrin #2 (Cappel); mouse albumin (Sigma), cytochrome C (Sigma) mouse α-fetoprotein (Calbiochem). Essentially identical results were found in 3 separate albumin/transferrin absorption experiments. The α- fetoprotein absorption experiment utilized liver from a 7 day old mouse as its target tissue.

Titration of Antisera. Most antisera were checked for reactivity against hepatocytes at a range of dilutions from 1/5 to 1/250. Generally, dilution affected the brightness of the fluorescence reactions but did not influence the number of hepatocytes visualized.

Negative Controls for the Immunofluorescence

Technique. Normal rabbit and goat sera were included as controls in a majority of the experiments described here, and were never found to react with hepatocytes. In

addition, as mentioned above, after specific absorption, anti-plasma protein antisera become completely unreactive against hepatocytes.

Plasma protein-containing cells were not observed in a number of mouse tissues where synthesis is either absent or low. Tissues examined included brain, gut, testes, salivary gland, bone marrow, thyroid and lymph node. The lactating mammary gland synthesizes transferrin (but not albumin) in amounts comparable to the liver (see, e.g.,

Jordan et al., Biochim . et Biophys . Acta 174:373, 1969), and preliminary experiments have revealed that a subpopulation of these cells react with antisera to transferrin, but not with anti-albumin antisera.

Immunofluorescence Analysis of Megakaryocytes in Liver and Bone Marrow. For a short time after birth,

hematopoietic cells are present in the liver of mice and rats. Included among these hematopoietic cells are

megakaryocytes, which are known to contain fibrinogen (see, e.g., Uzan et al., Biochem . Biophys . Res . Comm . 140:543, 1986; Handagama et al. Prog. Clin . Biol . Res . 356:119, 1990; and Louache et al., Blood 15 , 77:311, 1991). Furthermore, megakaryocytes in the livers of neonatal rats were found to contain fibrinogen, as indicated by reactivity with antisera to fibrinogen, but not with antisera to albumin, α- transferrin. The same reactivities were found with respect to megakaryocytes in rat bone marrow.

Positive Controls for the Immunofluorescence Technique. In contrast to the heterogeneous distribution seen when plasma protein antisera were tested against frozen sections of liver, a number of widely expressed cellular proteins can be visualized in all hepatocytes. First, anti-estrogen receptor antisera is observed to react with the nuclei of essentially all hepatocytes in the livers of C57BL/6J strain female mice, a result in agreement with the observations of Yamashita et al., Histochem . 90:325, 1989. Second, immunofluorescence with anti-histocompatibility antigen showed reactivity with the cell membranes of virtually all mouse hepatocytes; endothelial cells in the liver are particularly reactive, but all hepatocytes are positive as well, as could be seen where adjacent hepatocytes touch. (antigens

detected: K^b on C57BL/6J, D^b on C57BL/6J,. K^k on A/J, D^d on A/J; negative controls: K^b and D^b on A/J, K^k and D^d on C57BL/6J. For antisera see, e.g., Michaelson,

Immunogenetics 17:219, 1983). Third, anti-actin antisera showed immunofluorescence reactivity throughout the liver. In accordance with previous observations of

Franke et al., Biol . Cellulaire 34:99, 1979, and

Ballardini et al., VirchoWs Archiv B Cell Pathol . 56:45, 1988, actin is present in all mouse hepatocytes, where it is concentrated on the periphery of the cell. Fourth, the enzyme carbamoylphosphate synthetase is known to be located in all hepatocytes except those located in the few rows of cells adjacent to the terminal hepatic venules (see, e.g., Moorman et al., supra) , and a similar pattern of widespread expression is found using the methods described above. Fifth, TROMA-I monoclonal antibody, which reacts with a liver-specific mouse cytokeratin, causes immunofluorescence in all mouse hepatocytes (see e.g., Howe et al.. Hist . Cyto . 34:785, 1986).

Additional liver-specific products in human liver, especially Factor VIII:C, Factor IX, LDL Receptor, and alpha-1-antitrypsin, can be determined by the methods outlined above. Other methods for identifying proteins, lipids, carbohydrates and other metabolites are known in the art (see, e.g. Jungermann et al., supra)

Transplantation

The invention involves the use of specialized subpopulations of hepatocytes in transplantation for the treatment of human diseases, e.g., disorders of hepatic metabolism, ranging from ailments associated with

carbohydrate, lipid, protein, amino acid and ammonia metabolism to illnesses connected with detoxification mechanisms. Examples of human genetic diseases (see, e.g., Wilson et al., Principles of Internal Medicine, McGraw-Hill, N.Y., 1991), described below, amenable to such treatment include, without limitation, familial hypercholesterolemia, α₁-antitrypsin deficiency, factor VIII deficiency (Hemophilia A) and factor IX deficiency (Hemophilia B).

Familial hypercholesterolemia is an autosomal dominant disorder in human patients caused by a

deficiency of the receptor that mediates the uptake of low density lipoprotein (see, e.g., Scriver et al. (eds) The Metabolic Basis of Inherited Disease, McGraw-Hill, NY, pp 1215-1250). The disease leads to elevated levels of serum cholesterol and premature development of

coronary artery disease. Patients with two abnormal alleles for low density lipoprotein receptors have severe hypercholesterolemia that is refractory to medical

therapy.

Alpha₁-antitrypsin deficiency. is a hereditary disorder characterized by reduced serum levels of α₁ - antitrypsin, a protease inhibitor that provides the major defense for the lower respiratory tract against the ravages of neutrophil elastase, a powerful destructive protease. The loss of this protective screen of the fragile alveolar walls results in emphysema (see, e.g., Wilson et al., supra) . Approximately 10 percent of children homozygous for α^antitrypsin deficiency will develop significant liver disease including neonatal hepatitis and progressive cirrhosis (Wilson et al.

supra). In adults, the most common manifestation of α₁- antitrypsin deficiency is asymptomatic cirrhosis, which may progress from a micronodular to a macronodular state and may be complicated by the development of

hepatocellular carcinoma.

Hemophilia A and hemophilia B are sex-linked inherited plasma coagulation disorders due to defects in factors VIII and factor IX, respectively (see, e.g., Wilson et al., supra). Factor VIII coagulant protein is a single-chain protein (265 kDa) which regulates the activation of factor X by proteases generated in the intrinsic coagulation pathway. It is synthesized in liver parenchymal cells and circulates complexed to the von Willebrand protein. One in 10,000 males is born with deficiency or dysfunction of the factor VIII molecule. Factor IX is a vitamin K-dependent serine protease precursor (55 kDA) involved in the intrinsic blood coagulation pathway. It is converted to an active

protease (IXa) by factor Xla or by the tissue factor-Vila complex (see, e.g., Wilson et al. supra) . Factor IXa then activates factor X in conjunction with activated factor VIII. Factor IX is synthesized in the liver.

Factor IX deficiency or dysfunction occurs in 1 in

100,000 male births.

Treatment of patients with hemophilia A involves administration of plasma products enriched in factor VIII. The current treatment of patients with hemophilia B involves administration of either frozen plasma or crude preparations of plasma, different from plasma fractions used to treat hemophilia A, enriched in the prothrombin complex proteins. Both methods are

complicated by the potential risk of exposing patients to viral contaminants, such as viral hepatitis and human immunodeficiency virus (HIV).

A description of the methods involved in isolating, enriching, culturing, preserving, transforming and transplanting specialized subpopulations hepatocytes follows.

Isolation of Hepatocytes. Those skilled in the field of hepatocellular transplantation will understand that any of a wide variety of mammalian organ donors may be used to provide specialized hepatocytes. For example, hepatocytes may be obtained, without limitation, from hepatic tissue of a chicken, a mouse, a rat, a dog, a baboon, a pig, or a human. The tissue may come from liver fragments of differing sizes provided by a living (self or non-self) or a deceased donor, e.g., from a needle biopsy, a small wedge biopsy, or a partial

hepatectomy. If the donor hepatocytes are from non-self, then donor-recipient histocompatibility is determined. Class I and class II histocompatibility antigens are determined and individuals closely matched

immunologically to the patient are selected as donors.

Furthermore, all donors are screened for the presence of transmissible viruses (e.g., human immunodeficiency virus, cytomegalovirus, hepatitis A/B). Potential donors who test negative for such infectious diseases and other metabolic disorders, e.g., hyperlipidemia, are regarded as suitable donors.

Methods for the preparation of hepatocytes from hepatic tissue suitable for culturing and transplanting are described, e.g., in Langford et al., In Vitro

Cellular & Developmental Biology 25: 174, 1989; Strom et al., Journal National Cancer Institute , 68:, 771, 1982; Kay et al., Proc . Natl. Acad . Sci . , USA 89:, 89, 1992; and Enat et al., Proc. Natl . Acad . Sci . , USA 81: 1411, 1984. In one example, surgically isolated liver from a mammalian donor, is cannulated with a catheter and

perfused according to standard methods (see, e.g., Berry et al., J. Cell Biol . 43:506, 1969; Maslansky et al., In : Weber et al . , eds, In vitro models for cancer research, vol II, Boca Raton, FL, CRC Press pp 43-60, 1985; Seglen , Exp. Cell Res . 76:25, 1973). Perfusion is performed for 10 min at 60 ml/min using calcium/magnesium-free Hank's balanced salt solution supplemented with 10 mM HEPES (pH 7.4) and 0.5 mM [ethylene

bis(oxyethylenenitrillo]-tetraacetic acid. Perfusion is continued for an additional 20 min using Williams' medium E (WME) (see, e.g., Williams, Exp Cell Res . 89: 139,

1974) supplemented with 10 mM HEPES (pH 7.4) and 100 U/ml collagenase (type I, Sigma Chemical Co., St. Louis, MO). Following collagenase perfusion, hepatocytes are

displaced by gentle shaking in a collagenase solution. The resulting hepatocytes are either plated for culturing or enriched, e.g., by cell sorting (as described below) to obtain populations of specialized hepatocytes. Cell viability is examined by trypan blue exclusion.

Enriching for Specialized Populations of

Hepatocytes. Specific subpopulations of specialized hepatocytes can be isolated and enriched from the mixed hepatocyte cell suspension, described above, by the application of a number of standard methods. Of course, the fractionation method employed to enrich for a

particular specialized hepatocyte, e.g., a plasma- protein, a metabolic protein, a membrane protein or a structural protein, involves exploiting physical and biological properties commensurate with the specialized hepatocyte sought to be isolated. Such methods include, without limitation, fluorescence-activated cell sorting (FACS), vital staining combined with FACS, manipulation of in vitro growth conditions for mitotic expansion, immunopanning, and centrifugation.

For example, one suitable cell-separation technique employs the preparation of single cell

suspensions of mixed hepatocytes, followed by sorting of these cells on a fluorescence-activated cell sorter. One sorting method involves labeling specific cells with an antibody coupled to a fluorescent dye, followed by separating the labeled cells from the unlabeled cells in an electronic fluorescence-activated cell sorter (see e.g., Horan et al., Science , 198, 148, 1977; Miller et al., J. Immunol . Methods , 47: 13, 1981; and Loken et al., J. Immunol . Methods , 50:R85, 1982). Thus, in one example, enrichment of a specialized hepatocyte expressing LDL receptor protein would involve incubating the mixed population of hepatocytes, as described above, with an antibody, e.g., a fluorescein-conjugated monoclonal antibody against LDL receptor, followed by FACS. The subpopulation of hepatocytes enriched for cells

expressing LDL receptor is then processed as described below. The FACS procedure might also be modified to enrich for other types of specialized hepatocytes, such as plasma-protein expressing hepatocytes, by executing the FACS protocol below room temperature. Furthermore, the FACS procedure may be modified to sort specialized hepatocytes, e.g., a plasma-protein expressing

hepatocytes, based on spectroscopic characteristics , i.e., the ultraviolet (UV) or fluorescence properties of a specific plasma-protein hepatocyte or more generally simply on the premise that plasma-protein expressing cells will have a higher A₂₈₀ than non-plasma-protein expressing hepatocytes that do not contain secretory granules. An alternative FACS-based protocol may be applied that utilizes methods of cytoplasmic vital staining, i.e., stains that specifically react with a specialized hepatocyte that expresses e.g, a plasma- protein, a membrane protein, a metabolic protein or a structural protein. After staining, cells are then sorted by FACS to enrich for positively stained cells.

Another method of enriching for specialized subpopulations of hepatocytes involves manipulating hepatocyte growth conditions in tissue culture media to selectively induce specific subpopulations of hepatocytes to proliferate. For example, mice undergo a renewal of α-fetoprotein synthesis after they have been injected with high levels of estradiol (Hau et al. Acta

Endocrinologica 106: 141-144, 1984). Thus, mixed

hepatocytes may be cultured in a medium containing estradiol to promote growth of α-fetoprotein plasma- protein producing hepatocytes. Similarly, the mitotic expansion of fibrinogen-expressing hepatocytes might be selectively obtained by inducing hepatocyte cultures with stimuli known to effect inflammation such as IL-1 or TNF or turpentine (see, e.g., Gauldie et al., Ann . NY Acad. Sci . 557:46, 1989). For example, in vivo exposure to turpentine (e.g., 0.1 pg-lmg/ml), a known inflammation- inducing substance, should permit for the expansion of fibrinogen-containing cells by selective mitotic

proliferation. Finally, mitotic expansion of albumin- expressing hepatocytes might simply be achieved by culturing mixed populations of hepatocytes in the absence of albumin.

Alternative methods to sorting are also available, of which immunopanning (see, e.g, Barnes et al., Cell 70: 31, 1992) is the best developed, and may be used to procure enriched populations of specialized hepatocytes with the appropriate antibody.

Finally, an additional separation technique based on centrifugation methodologies, e.g., differential or density, can be utilized to enrich for the generic class of plasma-protein expressing hepatocytes (see, e.g., Sharpe, Methods of Cell Separation , Elsevier, 1988). Since it appears that approximately 1% of total liver hepatocytes produce any of the 100 or so plasma proteins, it might reasonably be expected that these cells have significantly higher concentrations of cytoplasmic protein, due to higher amounts of intracellular secretory granules, relative to non-plasma-protein-expressing hepatocytes. Consequently, these cells may vary in their density compared with other non-secretory cells. This difference may be exploited to separate these cells by density sedimentation or centrifugation. Thus, a variant on this invention involves the isolation by sorting of this generic class of plasma protein producing cells.

Moreover, this group may be considered homogeneous with respect to the general quality of plasma protein

synthesis. Consequently, these cells may also constitute a superior population of cells for hepatocyte

transplantation.

Confirmation and assessment of the phenotypes of the isolated cells may be determined by both biochemical analysis of precursor protein and by northern blot analysis, according to standard methods (see, e.g.,

Ausubel et al., supra) , in order to determine whether the specific protein and mRNA content in each of these populations is sufficient to account for mRNA quantities of the liver as a whole.

Cell Culture. Hepatocytes are generally known in the art to be difficult cells to adapt to tissue culture, but this may not apply to specialized

hepatocytes, e.g., subtypes of plasma-protein expressing hepatocytes, which may harbor a potential for cell division not evident in ordinary unspecialized

hepatocytes.

Specialized hepatocytes are cultured and expanded according to standard methods known to those skilled in the art (see, e.g., Enat et al., Proc . Natl . Acad . Sci . 81: 1411, 1984; Langford et al., In Vitro Cellular & Developmental Biology 25: 174, 1989). For example, hepatocytes isolated as described above may be plated at a density of 3-4 X 10⁴ cells per cm² onto plates with a polycationic matrix (Primaria; Falcon, Oxnard, CA) in a hormonally defined medium (see, e.g., Enat et al., supra) supplemented with 10% fetal bovine serum; 1-6 hr later the medium is replaced with fresh hormonally defined medium, which is subsequently changed every 6-24 hr during the duration of culturing. Culture media may be further supplemented with specific growth factors and active mitogens (e.g., IL-6, IGF-1, prolactin, EGF, glucocorticoids, GM-CSF, IFNα, IFN0, IFNγ, HGF, and TGFβ) to augment growth and cell division of the hepatocytes.

Preservation of Specialized Hepatocytes.

Isolated hepatocytes may be stored and preserved by any standard method, e.g., by freezing. These methods are highly developed for the preservation of a wide variety of mammalian cells see, e.g., Rajotte et al., Cryobiology 18: 357, 1981; Daniel, ed, Methods of Mammalian

Reproduction , New York, Academic Press, 1979, p. 179; and Sandier et al., Horm . Metabol . Res . 12: 71, 1982 and should be most useful for establishing deposits of specialized hepatocytes for subsequent utilization in transplantation.

Transplantation of Specialized Hepatocytes. Once isolated the specialized hepatocytes (either transfected or untransfected) are introduced into a recipient in need of a hepatic transplant or the protein encoded by the transfected gene (see below). If the hepatocytes used for transplantation are not provided by the transplant recipient, then concomitant immunosuppression therapy is necessary, e.g., administration of the immunosuppressive agent cyclosporine (see, e.g., Physician 's Desk

Reference , Medical Economics Data, Montvale, N.J., 1992) or, e.g., by masking antigens of non-self donor

hepatocytes according to methods described by Faustman and Coe (Science 252: 1700, 1991) to prevent rejection of transplanted cells.

Several standard methods are available for the transplantation of specialized hepatocytes either

delivered directly to the liver, e.g., via the portal vein, or for deposition within other locations throughout the body; for example, some investigators have implanted hepatocytes within the spleen (see, e.g, Vroemen et al., Transplantation , 42: 130, 1986; Strom et al., supra ; Kay et al., supra ; Gupta et al., Hepatology, 14: 144, 1991; Ponder et al., supra) , fat pads (see, e.g., Jirtle et al., Am . J. Pathol . 101:115, 1980), pancreas (see, e.g., Jaffe et al., Transplantation , 45: 497, 1988), subrenal capsule (see, e.g., Ricordi et al., Transplantation 45, 1148, 1988); or on microcarrier beads in the peritoneum (see, e.g., Demetriou et al., Proc. Natl . Acad . Sci . 83: 7475, 1986).

The invention is tailored to meet the requirements of an individual patient. In one working example, a patient diagnosed with Hemophilia A may be treated by providing specialized hepatocytes that express Factor VIII:C by transplantation. Donors are chosen as

described above. The preferable donor has no family history of hemophilia or any other liver disease, and is closely matched to the recipient immunologically. Liver tissue is obtained, as described above, and Factor VIII:C expressing hepatocytes are isolated and cultured. The specialized Factor VIII:C expressing cells then may be introduced to the recipient by standard surgical

procedures. Between 10² to 10⁹ cells may be transplanted in a session. Additional transplants can be performed as required (based on assaying blood coagulation). If necessary, immunosuppressive therapy is also administered as described above.

Methods for Introduction of DNA into Hepatocyte.

The specialized populations of hepatocytes (as described in the preceding sections) may provide an excellent source of hepatocytes that can stably produce and deliver recombinant proteins into the circulatory system. Thus, the specialized hepatocytes would be isolated from a recipient, cultured and expanded, transfected or

transformed with a recombinant gene in vitro. and

subsequently transplanted into the donor organism. Once situated, the hepatocytes would manufacture and secrete the desired recombinant protein for in vivo delivery.

Thus, in one aspect of the invention, a gene encoding, for example a human low density lipoprotein receptor (see, e.g., Wilson et al., Proc. Natl . Acad . Sci . , USA 87: 8437, 1990), human α₁-antitrypsin gene (see, e.g., Kay et al., Proc . Natl . Acad . Sci . USA 89: 89, 1992), human factor VIII (see e.g., Wood et al., Nature 312:330, 1984) or human factor IX (see, e.g., Armentano et al., Proc. Natl . Acad. Sci . USA 87: 6141, 1990), is inserted into an isolated specialized

population of hepatocytes (as discussed above), either obtained from a patient suffering from a cellular defect or genetic disease such as familial hypercholesterolemia, α^antitrypsin deficiency, factor VIII or factor IX deficiency, respectively, or from an acceptable donor human (as described above) or animal.

Those skilled in the field of molecular biology will understand that any of a variety of conventional gene transfer methods may be used for introducing genes into cells. The precise method used to introduce a replacement gene, e.g., a human low density lipoprotein receptor, a human a^_^-antitrypsin gene, a human factor VIII gene or a human factor IX gene to the specialized subpopulation of hepatocytes is not critical to the invention. For example, physical methods for the

introduction of DNA into cells include microinjection (see, e.g., Capecchi et al.. Cell 22:479, 1980),

electroporation (see, e.g., Reiss et al., Biochem . Biophys . Res . Commun . 137: 244, 1986) and protoplast fusion (see, e.g., Schaffner, Proc . Natl . Acad . Sci .

U .S .A . 77: 2183, 1980) and on high-velocity tungsten microprojectiles (see, e.g., BioRad Technical Bulletin #1687, Hercules, CA; Johnston, Nature 346:776, 1990).

Chemical methods such as coprecipitation with calcium phosphate and incorporation of DNA into liposomes also have been used to introduce DNA into mammalian cells.

Finally, delivery of nucleic acids into mammalian cells can be executed through the use of viral vectors, in particular those derived from murine and avian

retroviruses (see, e.g., Gluzman et al., Viral Vectors , Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1988).

There now follows an example of a retrovirus- mediated gene transfer used for introducing a gene (e.g., LDLR, α₁-antitrypsin, factor VIII and factor IX) into an enriched population of specialized plasma-expressing hepatocytes. This example is provided for the purpose of illustrating, not limiting, the invention.

Using standard recombinant DNA methods (see, e.g., Ausubel et al., Current Protocols in Molecular Biology, John Wiley & Sons, New York, 1989), a full-length human factor IX cDNA or complete genomic sequence, with

associated regulatory sequences or sequences, e.g., promoters, enhancers, terminators, provided by a

heterologous source, is cloned into a retroviral vector. Numerous vectors useful for this purpose are generally known and have been described (see, e.g. Miller, Human Gene Therapy, 1:5, 1990; Friedman, Science , 244:1275,

1989; Eglitis and Anderson, BioTechniques , 6:608, 1988; Tolstoshev and Anderson, Current Opinion in

Biotechnology, 1:55, 1990; Sharp, The Lancet , 337:1277, 1991; Cornetta et al., Nucleic Acid Research and

Molecular Biology 36:311, 1987; Anderson, Science ,

226:401, 1984; Moen, Blood Cells , 17:407, 1991; and Miller et al., Biotechniques , 7:980, 1989). Moreover, retroviral vectors are particularly well developed and have been used in a clinical setting (Rosenberg, et al., N. Engl . J. Med, 323:370 1990). Retroviral constructs, packaging cell lines and delivery systems which may be useful for this purpose are described in the preceding references. Furthermore, vectors and transduction methods for successfully introducing a gene into

mammalian hepatocytes have been described (see, e.g., Wolff et al., Proc . Natl . Acad . Sci . USA, 84: 3344, 1987; Wilson et al., Proc . Natl . Acad . Sci . USA, 85: 3014, 1988; Wilson et al., Proc . Natl . Acad . Sci . USA, 85:

4421, 1988; Peng et al., Proc . Natl . Acad . Sci . USA, 85: 8146, 1988; Wilson et al., Proc. Natl . Acad . Sci . USA, 87:8437, 1990; Armentano et al., Proc . Natl . Acad . Sci . USA, 87: 6141, 1990; Chowdhury et al.. Science 254: 1802, 1991; Kay et al., Proc . Natl . Acad . Sci . USA, 89: 89, 1992).

For example, the human coagulation factor IX gene can be isolated from a human liver cDNA library (prepared according to standard techniques, see, e.g., Ausubel et al. supra , or from a commercially available library see, e.g, Stratagene Catalog, La Jolla, CA, #937200 human liver cDNA library from a Normal male, 49 years old) or a human genomic library (prepared according to standard techniques, see, e.g., Ausubel et al., supra , or from a commercially available library, see, e.g., Stratagene #943202 in Lambda Dash vector) using oligonucleotide probes based on the factor IX DNA sequence (Kurachi et al., Proc . Natl . Acad . Sci . , USA, 79: 6461, 1982). Once isolated, the factor IX cDNA is cloned into any suitable vector used for retroviral-mediated gene transfer as described above.

Factor IX cDNA expression is directed from any suitable promoter, e.g., the human cytomegolvirus, simian virus 40, or metallothionein. Furthermore, Factor IX cDNA expression may be driven and regulated by any natural mammalian regulatory sequences and elements. For example, enhancers/promoters known to direct preferential gene expression in the liver might be used to direct Factor IX cDNA gene expression. These

enhancers/promoters include, without limitation: α- fetoprotein (see, e.g., Hammer et al., Science , 235: 53, 1987), transthyretin (see, e.g., Yan et al., EMBO J. 9: 869, 1990), α₁-antitrypsin (see, e.g., Kelsey et al., Genes & Development 1:161, 1987), albumin (see, e.g., Pinkert et al. Genes & Development 1: 268, 1987) and metallothionein (see, e.g., Karin et al. Nature 308: 513, 1984). Alternatively, if a Factor IX genomic clone is utilized, its expression may be regulated by its cognate regulatory sequences or, if desired, by regulatory sequences derived from a heterologous source, e.g., a SV40 promoter region. Regardless of how Factor IX gene expression is regulated from the transduced specialized hepatocytes the final level of protein, after

transplantation (see above), should fall within

physiological levels of circulating factor IX found in non-affected individuals.

To assess Factor IX gene expression in transduced hepatocytes standard methods, e.g., by Southern,

Northern and Western blot analyses, may be utilized

(Ausubel et al., supra ; Armentano et al., supra) .

Furthermore, levels of recombinant Factor IX in culture supernatants can be determined by an immunoradiometric assay (IRMA) as described by Bray et al., J . Lab . Clin . Med . 107, 269, 1986. Clotting activity of purified recombinant factor IX secreted by transduced hepatocytes may be determined by a one-stage clotting assay with factor IX-deficient plasma as described by Thompson , J. Clin . Invest . , 59: 900, 1977.

Once a line of transformed or transfected specialized hepatocytes is established these cells are then transplanted according to the methods outlined above.

From the foregoing description, one skilled in the art can easily ascertain the essential characteristics of this invention, and without departing from the spirit and scope thereof, can make various changes and modifications of the invention to adapt it to various usages and conditions. Thus, other embodiments are also within the claims.

Claims

1. A method of preparing a sample of cells for hepatic transplantation, said method comprising

a) providing a sample of hepatocytes,

b) enriching said sample for specialized hepatocytes.

2. The method of claim 1, further comprising step c), admixing said enriched sample with a pharmacetically acceptable carrier.

3. A method of preparing a sample of cells for hepatic transplantation, said method comprising

a) providing a sample of hepatocytes enriched for specialized hepatocytes, and

b) admixing said sample with a pharmaceutically acceptable carrier.

4. The method of claim 1 or 3, wherein said sample of hepatocytes is obtained from self.

5. The method of claim 1 or 3, wherein said sample of hepatocytes is obtained from non-self.

6. The method of claim 1 or 3, wherein said specialized hepatocytes express a plasma protein.

7. The method of claim 1, wherein said plasma protein is albumin, transferrin, complement component C3, α₂-macroglobulin, fibrinogen, Factor XIII:C, Factor IX, α₁-antitrypsin or α-fetoprotein.

8. The method of claim 1 or 3, wherein said specialized hepatocytes express a metabolic protein.

9. The method of claim 8, wherein said metabolic protein is ornithine transcarbamoylase, arginosuccinate synthetase, glutamine synthetase, glycogen synthetase, glucose-6-phosphatase, succinate dehydrogenase,

glucokinase, pyruvate kinase, acetyl CoA carboxylase, fatty acid synthase, alanine aminotransferase, glutamate dehydrogenase, ferritin or alcohol dehydrogenase.

10. The method of claim 1 or 3, wherein said specialized hepatocytes express a membrane protein.

11. The method of claim 10, wherein said membrane protein is GLUT-1, or LDL receptor.

12. The method of claim 1 or 3, wherein said specialized hepatocytes express a structural protein.

13. The method of claim 12, wherein said structural protein is cytokeratin #8, or cytokeratin #18.

14. The method of claim 1 or 3, wherein said specialized hepatocytes express a carbohydrate.

15. The method of claim 14, wherein said carbohydrate is glycogen, glucose, or glucose-6- phosphate.

16. The method of claim 1 or 3, wherein said specialized hepatocytes express a lipid.

17. The method of claim 1 or 3, wherein said specialized hepatocytes express a metabolite.

18. The method of claim 17, wherein said metabolite is urea or glutamine.

19. The method of claim 1 or 3, wherein said sample is homogeneous.

20. The method of claim 1 or 3, wherein said sample is heterogeneous.

21. The method of claim 1 or 3, wherein said specialized hepatocytes express a molecule capable of preventing or treating an inherited or acquired disease.

22. A method of providing a sample of cells capable of expressing a protein, said method comprising: (a) providing a sample of hepatocytes enriched for specialized hepatocytes; and (b) transfecting said hepatocytes with DNA encoding said protein.