Human CYP3A4-introduced HepG2 cells: in vitro screening system of new chemicals for the evaluation of CYP3A4-inhibiting activity.

1. The aims were to attest whether HepG2-GS-3A4, a cell line into which the human CYP3A4 gene was introduced, can be used for a screening of chemicals that will inhibit CYP3A4 activity. 2. The capacity of the cells for metabolizing CYP3A4 substrates in vitro was evaluated. Also determined was the effect of CYP3A4 inhibitors and non-inhibitors on nifedipine hydroxylation. Western blot, immunohistochemostry and determination of beta-nicotinamide adenine dinucleotide phosphate (NADPH)-reductase activity were performed. 3. HepG2-GS-3A4 selectively metabolized substrates of CYP3A4 (diazepam, nordiazepam, lidocaine, atorvastatin, and nifedipine) to a greater degree than control. The metabolites were easily detected in the culture medium. Values of V(max) of HepG2-GS-3A4 were about 30- to 100-fold higher than those of the control, while values of K(m) were comparable. Pre-incubation of cimetidine and ketoconazole significantly inhibited nifedipine hydroxylation, while addition of inhibitors specific to other isoforms of CYPs had no substantial effect. The HepG2-GS-3A4 expressed a higher amount of CYP3A4 protein and mRNA than control. Most NADPH reductase activity was detected in microsomal fractions. 4 In conclusion, HepG2-GS-3A4 sufficiently and selectively metabolize substrates of CYP3A4, and inhibitors of CYP3A4 reduced the metabolism. Because the metabolites were easily detected in the culture medium, this cell might be useful for the new and easy screening of new drugs for the evaluation of CYP3A4-inhibiting activity in vitro.

In vivo estimation of bioartificial liver with recombinant HepG2 cells using pigs with ischemic liver failure.

Biological efficacy of a recombinant human hepatic cell line, glutamine synthetase transfected HepG2 (GS-HepG2), was examined with large-scale culture in a circulatory flow bioreactor and in pigs with ischemic liver failure. GS-HepG2 cells were cultured in a circulatory flow bioreactor from 5 x 10(7) to 4 x 10(9) cells for 109 days. The cells showed ammonia removal activity even under substrate (glutamic acid)-free medium, suggesting that the GS catalyzed the activity using intracellular glutamic acid that had been pooled during conventional culture. When GS-HepG2 bioartificial liver (BAL) was applied to pigs with ischemic liver failure, survival time was prolonged to 18.8 +/- 6.1 h (mean +/- SD, n = 4) from 13.8 +/- 5.4 h (n = 6) and 10.7 +/- 4.1 h (n = 6) (groups treated with cell-free BAL and treated with plasma exchange and continuous hemodiafiltration, respectively). Laboratory data indicated the tendency for improvement in increase of blood ammonia level and decline of blood coagulation indices in the GS-HepG2 BAL-treated group. The advantages and potential for the cell line as a bioreactor in BAL is also discussed, comparing to those of isolated porcine hepatocytes.

Flow cytometry: an improved method for the selection of highly productive gene-amplified CHO cells using flow cytometry.

In previous work, we clarified the relationship between the productivity and stability of gene-amplified cells and the location of the amplified gene. The location of the amplified gene enabled us to classify resistant cells into two types. One type of resistant cell group, in which the amplified genes were observed near the telomeric region, was named the "telomere type." The other type of cell group, in which the amplified genes were observed in other chromosomal regions, was named the "other type." The phenotypes of these two types of cells are very different. In this experiment, using a fluorescein isothiocyanate-labeled methotrexate (F-MTX) reagent with flow cytometry, we were easily able to distinguish between highly productive cells and the other types of cells. The level of fluorescence differed according to the difference in resistance to MTX. Based on this new finding, highly productive gene-amplified cells could be isolated from heterogeneous gene-amplified cell pools more easily than by the method of limiting-dilution assay. The limiting-dilution method requires several months to obtain highly productive gene-amplified cells, while our flow-cytometry-based method of selection requires only a few weeks.

A cDNA from human bone marrow encoding a protein exhibiting homology to the ATP1gamma1/PLM/MAT8 family of transmembrane proteins.

A cDNA clone, IWU-1, was cloned from human bone marrow. Its putative open reading frame encoded a protein of 115 amino acids with a calculated molecular mass of 12.9 kDa. The deduced amino acid sequence exhibited high homology (>68%) to members of the ATP1gamma1/PLM/MAT8 family of single transmembrane proteins, primarily in the region containing the putative transmembrane domain. The sequence at the amino-terminal side exhibited high homology (>61%) to the cytoplasmic region of the angiotensin II type 1 receptors.

Constrained optimization of L-lysine production based on metabolic flux using a mathematical programming method.

Constrained optimization for microbial fermentation was studied. For optimization, we used not the maximum principle but a nonlinear programming method because of the need to consider many metabolic reactions. In the case of L-lysine fermentation, the optimization problem in L-lysine production was formulated as a nonlinear programming problem. In general, the state equations based on material balances are represented as differential equations, but such equations which are dependent on time can not be applied to a nonlinear programming problem. Therefore, the state equations were made discrete in a time base, and a new single vector which is not dependent on time was substituted. From these formulae, the objective function and the constraints using nonlinear programming problem were defined as the amount of L-lysine produced, and as a metabolic reaction model and empirical equations, respectively. Computer program was developed to solve this constrained nonlinear programming problem. The applied algorithm of the computer programming was a sequential quadratic programming method (SQP method). When the constrained nonlinear programming problem is solved using the SQP method, the maximum amount of L-lysine produced and the optimal feeding rate of L-threonine could be calculated. From the calculated results, it was clear that introduction of the equality and inequality constraints was easy. L-Lysine at a concentration up to 75.3 g/l could be produced when the fermentation was carried out under optimal conditions.

A kinetic model for a biopanning process considering antigen desorption and effective antigen concentration on a solid phase.

A phage display library is a powerful tool for screening ligands such as antibodies and peptides that specifically recognize a target. In this study, we established a kinetic model describing the affinity selection process of phage display libraries and verified the model experimentally. Desorption of target molecules from a solid phase and orientation of the epitopes of adsorbed target molecules are taken into account in this model. The ratio of the effective antigen density to the total antigen density was estimated to be 0.0127(+/-)0.0018 when bovine pancreatic ribonuclease A (RNase A) adsorbed on polystyrene beads was recognized by an anti-RNase A single-chain Fv phage antibody. The model can faithfully describe the recovery of the phage antibody in a round of biopanning based on the effective concentration of RNase A on the beads, the desorption rate constant of RNase A from the beads, the dissociation constant and dissociation rate constant of the phage antibody from RNase A, and the time for blocking, equilibrium and washing in the biopanning process. A recommended biopanning protocol based on the model is also described.

Measurement of association rate constant of antibody-antigen interaction in solution based on enzyme-linked immunosorbent assay.

A simple and rapid method based on enzyme-linked immunosorbent assay (ELISA) was developed for measuring the association rate constant of antibody-antigen interactions. An antibody and its antigen are mixed in a solution to initiate the equilibrium reaction. At different time intervals, the amount of the free antibody in the reaction mixture is estimated by an indirect ELISA. The association rate constant was estimated by nonlinear regression against an equation introduced from the derivation of the mass balance of antigen-antibody interaction. This method can determine the association rate constant of antibodies with a dissociation rate constant up to 5 x 10(-3) s(-1). The association rate constant of a single-chain Fv (scFv) to its antigen, bovine pancreatic ribonuclease A (RNase A), determined by the present method agreed well with those determined by the fluorescence polarization method and surface plasmon resonance method. No significant difference in the association rate constant was found between the soluble anti-RNase A scFv and the same scFv displayed on a phage (5.65 +/- 0.54 x 10(4) M(-1) s(-1) and 5.96 +/- 0.56 x 10(4) M(-1) s(-1), respectively).

Control of antibody-antigen interaction using anion-induced conformational change in antigen peptide.

The binding of a monoclonal antibody to an epitope peptide was controlled by the conformational change of the epitope peptide induced by anions. We synthesized peptides in which the epitope sequence DTYRYI for the monoclonal antibody AU1 is located between amphiphilic peptides (KKLL)n and (LLKK)n. In the absence of an appropriate anion, the peptide was in a random coil state and the epitope was linear. In contrast, in the presence of an appropriate anion, the peptide exhibited an anti-parallel alpha-helical structure and the epitope was subsequently 'bent'. In the presence of 41 microM triphosphate, the association constant between the antibody and the peptide was decreased by one order of magnitude in the case of n = 3 and at least three orders of magnitude in the case of n = 4 or 5. Oligo-DNAs, as anions, dissociated the antibody-peptide complex, whereas triphosphate did not. The DNA concentrations required for 50% dissociation of the antibody-peptide complex at pH 7.5 were 4x10(-8), 1x10(-7) and 6x10(-6) M for decamer, octamer and hexamer DNA, respectively.

Amplified gene location in chromosomal DNA affected recombinant protein production and stability of amplified genes.

Previously, we established an easy and quick construction method for obtaining a stable and highly productive gene-amplified recombinant Chinese hamster ovary (CHO) cell line. With a gradual increase in methotrexate (MTX) concentration, gene-amplified cell pools had high and stable specific growth and production rates. Moreover, the phenotype of gene-amplified cells seemed to be affected by the location of the amplified gene in chromosomal DNA. We suspected that various kinds of gene-amplified cells might appear during the long-term selection to construct gene-amplified cell pools. To clarify the behavior of gene-amplified cell pools during a stepwise increase of MTX concentration, we isolated gene-amplified clones derived from gene-amplified cell pools. We compared the characteristics of isolated clones, such as the productivity of recombinant protein, stability of amplified genes, and the location of amplified genes. As a result, telomere-type clones, in which the amplified gene was located near the telomeric region, were found to be more stable and productive than other types of clones. Telomere-type clones had over 100 copies of amplified genes in the chromosomal DNA. In contrast, a large number of other types of clones had less than 10 copies of amplified genes. During long-term cultivation in the absence of MTX, in other types of clones, amplified genes rapidly decreased in the chromosomal DNA.

Efficient production of desulfurizing cells with the aid of expert system.

An expert system was used to achieve the high production of desulfurizing cells of Rhodococcus erythropolis KA 2-5-1. By adding a proper amount of sulfur containing component with the aid of the expert system, we could avoid excess feeding which resulted in the lowering of desulfurizing activity and starvation which caused serious damage to cell growth. In order to determine the addition amount by the expert system, the data of the amount of chemical elements contained in the cells were used as a reference for comparison with the medium components present. Culture experiments were carried out using a 5l jar fermentor with several kinds of media whose components were determined based on the inferred results with the aid of the expert system. We could restrict the amount of the sulfur component addition so that sulfur was a growth-limiting factor; in contrast, the amounts of other elements were sufficient for growth.As a result, the maximum specific production rate of 2-hydroxy biphenyl (2HBP) and the maximum cell concentration were 20mmolkg-drycells(-1)h(-1) and of 45g-drycellsl(-1), respectively. At 100h of cultivation, 1g/l of dibenzothiophene (DBT) was converted to 2HBP within 20h, i.e., 10mmolkg-drycells(-1)h(-1) of specific desulfurization activity, and the specific activity remained stable for a long period in the culture experiment.

Long-term culture of glutamine synthetase-transfected HepG2 cells in circulatory flow bioreactor for development of a bioartificial liver.

Glutamine synthetase (GS) is involved in an accessory pathway of ammonia removal in mammals. To develop a bioartificial liver with a human cell line, GS gene was transfected into HepG2 cells, which had no ammonia removal activity. After culturing in the presence of methionine sulfoximine (MSX), a GS inhibitor, we obtained a MSX-resistant HepG2 subline (GS-HepG2), which had amplified GS gene; ammonia removal activity was estimated to be 1/7 of that of rat primary culture hepatocytes. The cells were cultured in a circulatory flow bioreactor for 109 days, while they multiplied from 5 x 10(7) to 4 x 10(9) cells. Three days after inoculation, the ammonia level of the culture medium was lowered to a level maintained thereafter, suggesting that using recombinant cell lines for bioartificial livers enables long-term repeated treatment for hepatic failure patient. Judging from the rate of decrease in the amount of the added ammonia, the ammonia removal capability of 4 x 10(9) GS-HepG2 cells was almost equivalent to 5 x 10(8) porcine hepatocytes inoculated into the circulatory flow bioreactor. Apart from their ammonia removal activity, GS-HepG2 cells eliminated human tumor necrosis factor-alpha (TNF-alpha). Cytokine removal therefore promises to be another useful property of bioreactor cells.

Increase in desulfurization activity of Rhodococcus erythropolis KA2-5-1 using ethanol feeding.

Rhodococcus erythropolis KA2-5-1 is one of the best strains for the desulfurization of dibenzothiophene (DBT) via a sulfur-specific pathway in which DBT is converted to the end product, 2-hydroxybiphenyl, by releasing sulfite via DBT-sulfone and 2-(2'-hydroxyphenyl) benzene sulphinate. The objective of this research is to develop a culture method in order to attain a high cell density with a high level of specific desulfurization activity. Compared with glucose or glycerol, ethanol was found to be a preferable carbon source for obtaining a high specific activity (SA) of desulfurization. When the amount of DBT fed was restricted by feeding 2.9 mg-DBT/g-ethanol solution, the maximum SA and final cell concentration were 135.5 (mmol-2HBP/kg-dry cell weight-h) and 37 (g-dry cell weight/l), respectively. On the other hand, when glucose or glycerol was used as a carbon source, the SA was lower than 50 (mM-2HBP/kg-dry cell weight-h) and the final cell concentration was also lower than 27 (g-dry cell weight/l). The activities of the desulfurization enzymes in R. erythropolis KA2-5-1 grown on ethanol were remarkably higher than when the strain was grown on glucose or glycerol. It was also suggested that NADH, which is produced by the biochemical reaction of NAD with ethanol catalyzed by alcohol dehydrogenase, might contribute to the conversion of FMN to FMNH2, which is a coenzyme for the activities of desulfurization enzymes.

L-Lysine production by exponential feeding of L-threonine.

The effect of L-threonine feeding in the production phase on L-lysine production by Brevibacterium flavum, which requires L-homoserine or L-threonine for cell growth, was investigated considering the concerted inhibition by L-threonine plus L-lysine, and the metabolism related to lysine production. Exponential feeding of L-threonine increased L-lysine production to 70 g/l about three times that without feeding. From the analysis of the metabolic flux, carbon flux of L-lysine synthesis pathway in the production phase after L-threonine feeding was higher than that in the growth phase. The results show that feeding of an inhibitory substance may increase the production, especially when the substance is necessary for the continuation of cell growth and/or production.

Evaluation of stable and highly productive gene amplified CHO cell line based on the location of amplified genes.

In order to establish an easy and quick construction method for obtaining a stable and highly productive gene-amplified recombinant Chinese Hamster Ovary (CHO) cell line, variouskinds of stepwise methotrexate (MTX) selection were carriedout. The specific growth and production rates of the cell were compared with each other, and the distribution of the amplified gene location was determined using fluorescence in situ hybridization (FISH). The specific growth andproduction rates of the cell pool reached the highest levels under the selection condition in which the stepwise increase in the MTX concentration was most gradual; about 82% of amplified genes were observed near the telomeric region. During long-term cultivation without MTX, the percentage ofamplified genes near the telomeric region hardly changed, butthat of amplified genes at other regions decreased. Based on these results, stable and highly productive cell pools could be easily and quickly constructed and amplified and gradual stepwise increase of the MTX concentration. In addition, the FISH technique was powerful tool to evaluate highly productiveand stable gene-amplified cells based on the chromosomal location of the amplified gene.

An attempt to add biological functions by genetic engineering in order to produce high-performance bioreactor cells for hybrid artificial liver: transfection of glutamine synthetase into Chinese hamster ovary (CHO) cell.

In the course of immortalization, hepatocyte cell lines lose their original differentiated functions, such as ammonia removal and urea formation, drug metabolism, serum protein synthesis, etc. (Enosawa et al., Cell Transplant. 5:S39-S40; 1996). With the aim of adding lost or deficient functions and producing cell lines for the bioreactor of a hybrid artificial liver, rat glutamine synthetase (GS) gene was transfected into Chinese hamster ovary (CHO) cells, because it is able to lower the ammonia level. The GS gene-inserted pSV2 plasmid was transfected into the CHO-K1 line by electroporation. Transfected CHO (GS-CHO) cells were cultured in a glutamine-free medium containing ammonia, glutamic acid, and the GS inhibitor methionine sulfoximine (MSX). The MSX concentration was increased stepwise from 25 mumol/L to 1600 mumol/L to amplify the GS gene. In several GS-CHO sublines resistant to 300-1600 mumol/L of MSX, the specific activities of GS were increased from 0.2 x 10(4) to 1.7-2.9 x 10(4) unit/10(6) cells. When the amplified GS-CHO cells were cultured in the ammonia-containing medium, a slow but steady decrease of the ammonia level was observed when the level was high. Finally, the prospect of genetically modulated cells for bioreactors in the hybrid artificial liver is discussed.

Fluorescence polarization study of a salt bridge between a single chain Fv and its antigen ribonuclease A.

The interaction between a single chain Fv (sFv) of the monoclonal antibody 3A21 and its antigen, bovine pancreatic ribonuclease A (RNase A), was studied by site-directed mutagenesis of the hypervariable regions and fluorescence polarization analysis. The affinity constants of wild-type sFv and a mutant sFv D31A (Asp31 of heavy chain was replaced by Ala) for RNase A were found to be 2.7 x 10(7) and 4.7 x 10(6) M(-1) in PBS at pH 7.2 and 37 degrees C, respectively. Whereas the affinity constant of D31A is not affected by NaCl concentration, that of wild-type sFv is almost the same as that of D31A in the presence of more than 1 M NaCl. These results demonstrate that Asp31 of the heavy chain interacts electrostatically with a positively charged amino acid residue of RNase A.

Fluorescence polarization study of a salt bridge between a single-chain Fv and its antigen ribonuclease A.

The interaction between a single-chain Fv (sFv) of the monoclonal antibody 3A21 and its antigen, bovine pancreatic ribonuclease A (RNase A), was studied by site-directed mutagenesis of the hypervariable regions and fluorescence polarization analysis. The affinity constants of wild-type sFv and a mutant sFv D31A (Asp31 of heavy chain was replaced by Ala) for RNase A were found to be 2.7 x 10(7) and 4.7 x 10(6) M-1 in PBS at pH 7.2 and 37 degrees C, respectively. While the affinity constant of D31A is not affected by NaCl concentration, that of wild-type sFv is almost the same as that of D31A in the presence of more than 1 M NaCl. These results demonstrate that Asp31 of the heavy chain interacts electrostatically with a positively charged amino acid residue of RNase A.