Validation of fermentation processes.

The ability to prepare consistent biopharmaceutical products depends extensively on possession of banked and characterized cell substrates and on development of production processes which can be validated. While the attributes that define cell characterization have been extensively detailed by ICH and the regulatory agencies in the past decade, little has been specified regarding process validation for biological processes. The extent to which validation concepts can be applied to biological processes varies depending on the nature of the process, the nature of the product, and the level of knowledge regarding the relationship between process parameters and product quality. Expectations concerning the rigour of the validation programme should be adjusted accordingly. There is no single approach that is appropriate for all processes and products. At a minimum, there should be an attempt to define which process parameters are critical, and to focus the attention of validation efforts on these parameters.

Detection and consequences of recombinant protein isoforms: implications for biological potency.

Various types of structural variants have been observed in recombinant DNA - derived products. These isoforms include variations in post translational carbohydrate modifications where variations in site occupancy or unoccupied sites may occur. In addition, varying degrees of C-terminal processing and N-terminal substitutions have been observed. Isoforms may also be generated during processing and can include aggregated and/or chemically modified forms of the protein. Sophisticated analytical techniques exist for the identification and characterization of these structural variants. Several strategies have been used to isolate or enrich the isoform before molecular characterization. However, the effect these structural variations have on the biological activity of the product is less well understood. This may, in part, be due to the specificity and variability of the bioassay employed. This presentation describes the isolation and characterization of specific molecular isoforms for a monoclonal antibody product as well as an assessment of effects on biological activity.

Validation of a peptide mapping method for a therapeutic monoclonal antibody: what could we possibly learn about a method we have run 100 times?

Peptide mapping is a key analytical method for studying the primary structure of proteins. The sensitivity of the peptide map to even the smallest change in the covalent structure of the protein makes it a valuable 'finger-print' for identity testing and process monitoring. We recently conducted a full method validation study of an optimised reverse-phase high-performance liquid chromatography (RP-HPLC) tryptic map of a therapeutic anti-CD4 IgG1 monoclonal antibody. We have used this method routinely for over 1 year to support bioprocess development and test production lots for clinical trials. Herein we summarize the precision and ruggedness of the testing procedure and the main findings with respect to 'coverage of amino acid sequence' and limits-of-detection for various hypothetical structural variants. We also describe, in more detail, two unanticipated insights into the method gained from the validation study. The first of these is a potentially troublesome side-product arising during the reduction/alkylation step. Once the cause of this side-product was identified, it was easily prevented. We also report on subtle changes to the peptide map upon extended storage of the digest in the autosampler. These findings helped us to develop a 'robust' method for implementation in a quality control laboratory.

Quantitation of E. coli protein impurities in recombinant human interferon-gamma.

A multiple antigen ELISA for E. coli proteins (ECPs) that may be present in purified recombinant human interferon-gamma (rIFN-gamma) was developed. SDS-PAGE and Western blotting analyses showed that the assay antibodies reacted with a wide spectrum of ECPs in the standard and with ECPs in a production run. In spike recovery studies, rIFN-gamma at concentrations of 0.05 mg/mL and higher augmented the immunoreactivity of the ECPs in the standard curve (1.3-40.0 ng ECPs/mL) by approx 50%. To determine ECP content in purified rIFN-gamma, 0.2 mg/mL of rIFN-gamma was added to the standard curve diluent to compensate for enhanced immunoreactivity. The assay was precise (interassay precision of ECP controls < or = 4.1 %CV) and accurate with recoveries of 111-115% of expected for ECPs (15-40 ng/mL) spiked into purified rIFN-gamma (1 mg/mL). Linearity of dilution for ECPs spiked into rIFN-gamma was obtained (r = 0.999). Moreover, linearity of dilution was obtained for ECPs in "in-process" samples, demonstrating the required condition of antibody excess for this type of multiple antigen ELISA. ECPs were not detectable in several purified lots of rIFN-gamma. Therefore, these lots contained < 1.3 ppm ECPs.

Interaction of Cibacron Blue F3GA with glutamine synthetase: use of the dye as a conformational probe. 1. Studies using unfractionated dye samples.

Cibacron Blue F3GA dye has been used to probe subtle conformational changes in protein structure associated with the conversion of Escherichia coli glutamine synthetase (GS) between relaxed, taut, oxidized, and dissociated forms. Binding of the dye to each form of the enzyme elicits a different spectral perturbation of the dye which can be detected by difference spectroscopy. By following time-dependent changes in the difference spectrum associated with the binding of dye to the enzyme, it was demonstrated that dissociation of subunits provoked either by urea or by relaxation of the enzyme at pH 8.5 is a multiphasic process. In the presence of 3-4 M urea, dissociation of taut GS is associated with an almost instantaneous, transient increase in absorbancy of the difference spectrum at 638 nm and, after a lag, by a progressive decrease in absorbancy at 585 nm and an increase at 700 nm. The kinetics of these changes vary as a function of temperature, pH, and the concentrations of KCl, MnCl2, and urea, probably reflecting differences in the rates of GS relaxation and in the formation of aggregates of intermediate sizes. Results of direct binding measurements show that the taut and relaxed forms of GS can bind only 1-1.3 equiv of dye per subunit, whereas dissociated subunits bind up to 3.0 equiv per subunit. The Kd of the dye-taut GS complex as calculated from binding data was 0.55 microM. The binding of dye to taut GS was inhibited by its substrate, ADP, and by the allosteric effectors AMP and tryptophan. On the basis of the abilities of ADP, AMP, and tryptophan to inhibit the binding of dye to GS, dissociation constants of the respective GS-ligand complexes were 2.4, 121, and 1170 microM, respectively, in good agreement with previously determined values. From the difference spectra obtained between a given concentration of dye in a 5.0-cm cell and 10 times that concentration in a 0.5-cm cell, it was established that at concentrations greater than 5 microM a significant fraction of the dye is present as stacked aggregates. Because only the dye monomer binds to GS, the difference spectrum between dye and dye bound to GS is due in part to GS-promoted shifts in the equilibrium between stacked and unstacked dye molecules. Consequently, with increasing dye concentrations, the amplitude of the dye vs. dye + GS difference spectrum can continue to increase, even after the GS becomes saturated with dye.(ABSTRACT TRUNCATED AT 400 WORDS)

Interaction of Cibacron Blue F3GA with glutamine synthetase: use of the dye as a conformational probe. 2. Studies using isolated dye fractions.

By means of column chromatography on silicic acid, commercial preparations of Cibacron Blue F3GA have been resolved into four major subfractions (fractions I-IV). The difference spectrum between free dye and dye bound to any given form of Escherichia coli glutamine synthetase (GS) is different for each dye fraction. Moreover, uniquely different spectral perturbations are associated with the binding of any one dye fraction to the taut, relaxed, dissociated, or oxidized forms of GS. On the basis of the magnitude of the differences in the difference spectra between free dye and the dye-GS complexes, fraction II is most suitable for monitoring the interconversion of the relaxed and taut forms of GS. Fraction II can also be used to measure the fraction of oxidized (inactive) GS that is present in apparently homogeneous GS preparations. In contrast to the other three fractions, the difference spectrum obtained immediately following the binding of fraction I to GS undergoes a time-dependent change which is associated with the covalent attachment of the dye to the enzymes. Fractions II, III, and IV apparently bind to the nucleotide binding site on GS because the difference spectrum obtained with these fractions can be quenched by the subsequent addition of 1-2 mM ADP. The primary but not the secondary complex formed between GS and fraction I can also be destroyed by ADP.

Enzyme immunoassay of the glycoprotein tropic hormones--choriogonadotropin, lutropin, thyrotropin--with solid-phase monoclonal antibody for the alpha-subunit and enzyme-coupled monoclonal antibody specific for the beta-subunit.

Monoclonal antibody technology has made it possible to produce homogeneous populations of antibodies to discrete determinants on an antigen surface. We have produced monoclonal antibodies to the alpha-subunit and beta-subunits of the glycoprotein hormones choriogonadotropin, thyrotropin, and lutropin, and developed two-site simultaneous enzyme-linked immunospecific assays for these hormones. The anti-alpha-subunit monoclonal antibody was used as the solid-phase (coated tube) capture antibody for all three hormones; the anti-beta-subunit monoclonal antibodies were coupled to horseradish peroxidase (EC 1.11.1.7). Cross reactions between the closely related choriogonadotropin and lutropin were apparently greater in this method than in RIA, with use of the same antibodies. Ka of the antibodies did not appear to be as critical to sensitivity of the sandwich assay as it was for RIA. The lower limit of detection was 0.2 microgram/L after a 2-h incubation with serum sample at room temperature and a 30-min incubation with enzyme substrate at room temperature after washing away excess enzyme conjugate. Within-assay precision (CV) was very good, less than 6%.

Quantitation of aromatic residues in proteins: model compounds for second-derivative spectroscopy.

The ultraviolet absorption spectrum of proteins in 6 M guanidine is approximately that of the sum of the spectra of the constituent aromatic amino acids, phenylalanine, tyrosine, and tryptophan, plus contributions from light scattering and disulfides. A multicomponent analysis of the spectrum would theoretically permit simultaneous quantitation of each aromatic amino acid in the protein. In practice, this has not been possible, because of the similarities of the spectra of the amino acids, large differences in molar absorptivity, variable absorption by the disulfides, light scattering, and wavelength shifts which occur when the amino acids are incorporated into proteins. We describe a method for the simultaneous quantitation of the aromatic amino acids in purified proteins. We used second-derivative ultraviolet spectroscopy coupled with a statistically weighted multicomponent analysis. Use of the second derivative virtually eliminated interference from light scattering and from cystine. Empirical selection of model compounds obviated the problem of wavelength shifts. The models are N-acetylphenylalanine ethyl ester in 6 M guanidine for phenylalanine, N-acetyltyrosine ethyl ester in 55% methanol for tyrosine, and mellitin in 6 M guanidine for tryptophan. This method permits accurate, rapid quantitation of phenylalanine, tyrosine, and tryptophan in intact, denatured proteins.