Characterization of filter extractables by proton NMR spectroscopy: studies on intact filters with process buffers.

Studies were conducted to characterize potential extractables from sterilizing grade filters. The focus of this report is the 0.22 micron Durapore (hydrophilic modified PVDF) filter which is used throughout our recovery processes. The objectives of this study are (1) to identify potential filter extractables from the hydrophilic PVDF filters; (2) to show that NMR spectroscopy may be used to detect filter extractables in the presence of product and excipients; and (3) to establish levels of filter extractables obtained by extraction with a variety of buffers. The data show that the primary source of filter extractables is the hydrophilic modification of the PVDF membrane surface. Extractables from the modified hydrophilic PVDF filter include propylene glycol (PG) and soluble oligomers of the hydroxypropyl acrylate and cross-linker. Propylene glycol, arising from the hydrolysis of the hydroxypropyl acrylate, appears to be the primary extractable in buffers above pH 11. Since the 1H-NMR method can easily detect the methyl proton signals of PG, an NMR assay was developed to detect PG in the presence of buffer excipients and final product. Propylene glycol can be used as a marker for the extractables from Durapore hydrophilic PVDF filters. Although numerous buffers were used to generate extractables from the PVDF filter, significant extractables (PG and soluble oligomers) were found only in high pH extraction buffers. As a result of this finding, only a limited number of new buffers or new PVDF filters will require testing for future validation studies. Process validation studies have shown that neither PG nor soluble oligomers are at levels that impact the quality or safety of the product.

Affinity-reversed-phase liquid chromatography assay to quantitate recombinant antibodies and antibody fragments in fermentation broth.

An automated dual-column liquid chromatography assay comprised of affinity and reversed-phase separations that quantifies the majority of antibody-related protein species found in crude cell extracts of recombinant origin is described. Although potentially applicable to any antibody preparation, we here use samples of anti-CD18 (Fab'2LZ) and a full-length antibody, anti-tissue factor (anti-TF), from various stages throughout a biopharmaceutical production process to describe the assay details. The targeted proteins were captured on an affinity column containing an anti-light-chain (kappa) Fab antibody (AME5) immobilized on controlled pore glass. The affinity column was placed in-line with a reversed-phase column and the captured components were transferred by elution with dilute acid and subsequently resolved by eluting the reversed-phase column with a shallow acetonitrile gradient. Characterization of the resolved components showed that most antibody fragment preparations contained a light-chain fragment, free light chain, light-chain dimer and multiple forms of Fab'. Analysis of full-length antibody preparations also resolved these fragments as well as a completely assembled form. Co-eluting with the full-length antibody were high-molecular-mass variants that were missing one or both light chains. Resolved components were quantified by comparison with peak areas of similarly treated standards. By comparing the two-dimensional polyacrylamide gel electrophoresis patterns of an Escherichia coli blank run, a production run and the material affinity captured (AME5) from a production run, it was determined that the AME5 antibody captured isoforms of light chain, light chain covalently attached to heavy chain, and truncated light chain isoforms. These forms comprise the bulk of the soluble product-related fragments found in E. coli cell extracts of recombinantly produced antibody fragments.

Performance comparison of protein A affinity-chromatography sorbents for purifying recombinant monoclonal antibodies.

We describe the performance characteristics of five Protein A affinity-chromatography sorbents (Sepharose Fast Flow, Poros 50, Poros LP, Prosep and Streamline) for purifying a recombinant humanized monoclonal antibody from clarified Chinese hamster ovary cell culture fluid. We measured the dynamic capacity at varying flow rates, maximum capacity, pressure drop and production rate. For purified antibody, we measured yield and purity (by SDS/PAGE, the amount of DNA, the amount of host-cell proteins and the amount of Protein A). We found that, whereas all sorbents provided significant and essentially equivalent antibody purification, there were differences in capacity and pressure drop, which affected the production rate and had implications for process applications.

Purification and quantitation of tumor necrosis factor receptor immunoadhesin using a combination of immunoaffinity and reversed-phase chromatography.

The development of an automated, dual column assay to quantitate and recover the glycoprotein, tumor necrosis factor receptor immunoadhesin (TNFr-IgG) from monkey plasma, human serum, cell culture fluid and buffer samples is described. A combination of immunoaffinity and reversed-phase chromatographies are used. The targeted protein was captured using an anti-TNFr-1 monoclonal antibody immobilized on POROS resin. After non-specific adsorption had been reduced, the affinity column was placed in-line with a reversed-phase column and eluted with dilute acid. The reversed-phase column was subsequently eluted with an acetonitrile gradient and the TNFr-IgG collected and quantitated by comparison with peak areas of similarly treated standards. Detection was performed by measurement of absorbance at 214 nm. The dynamic range is from 0.5-15 microg total sample. Samples were quantitated and recovered from monkey and human pharmacokinetics samples, as well as from cell culture fluid and buffers. The lowest concentrations assayed were 100 ng ml(-1). Quantitation is reproducible, with a coefficient of variation of 2%. The procedure was used to develop a pharmacokinetic profile for the clearance of TNFr-IgG in humans and cynomolgus monkeys. Sufficient material was recovered such that the glycoforms could be identified. Additionally it has been used for process monitoring. The results compared favorably with data generated by ELISA. Optimization of the method and results are presented.

A two-dimensional protein map of Chinese hamster ovary cells.

Chinese hamster ovary (CHO) cells are used extensively for the expression of biopharmaceutical protein products. As part of our effort to better understand CHO cell physiology and protein expression changes caused by modified culture conditions, we have begun to map CHO cell polypeptides. A parental cell line reference map was established using two-dimensional gel electrophoresis with immobilized pH gradients (pH 3-10) in the first dimension and a linear acrylamide gradient (9-18%T) in the second dimension. The map is composed of over 1000 silver-stained protein spots. Protein identification is proceeding using a combination of immunostaining, NH2-terminal sequencing, and mass spectrometric analyses. Among the proteins so far identified are glucose-regulated protein 78 (GRP78), protein disulfide isomerase (PDI), galectin-1, and several heat-shock proteins. The goal is to generate a database which emphasizes those proteins most relevant to the use of CHO cells as a host for recombinant protein expression.

Perfusion immunoassay for acetylcholinesterase: analyte detection based on intrinsic activity.

A quantitative, two-column, HPLC-based assay requiring only 30 min to complete is reported. Amniotic fluid proteins are first fractionated on a size-exclusion column; the fraction containing the M(r) 280,000 neural acetylcholinesterase (AChE) is then diverted to a second column, an ImmunoDetection cartridge derivatized with an anti-AChE antibody. The immobilized antibody traps the enzyme, then substrate is flowed through the cartridge and the product is detected. For a positive result, the enzyme must have a molecular mass corresponding to the neural-AChE, be recognized by the antibody and be active in converting the substrate into product. The assay is sensitive in the clinically relevant range. The method provides rapid quantitative analysis using an automated instrument projected to be suitable for screening large numbers of samples.

Immunoaffinity purification of dietary heterocyclic amine carcinogens.

Cooking meats produces a family of heterocylic aromatic amines that are carcinogens in rodents and genotoxic in many short-term assays. Concern that these compounds may be human carcinogens has prompted us to develop immunochemical methods for quantifying these compounds in the human diet and for identifying the parent compounds and metabolites in urine and feces. Previously reported monoclonal antibodies to 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) and 6-phenyl-2-amino-1-methylimidazo[4,5-f]pyridine (PhIP) were used to purify by immunoaffinity these known mutagens and cross-reacting structural analogs from well-done cooked beef and urine samples. Materials recovered from the immunoaffinity columns were subsequently separated by HPLC to purify the known mutagens from cross-reacting chemicals that co-purify by immunoaffinity. Immunoaffinity chromatography was found to be a rapid means of quantifying individual known mutagens, with a minimum of precolumn sample clean-up required. In addition, this procedure has yielded several new mutagens present in cooked meats that are apparently structural analogs of PhIP. Immunoaffinity techniques were also used to purify metabolites from the urine of rats and humans exposed to MeIQx or PhIP. For MeIQx-exposed rats, the combination antibodies immunoconcentrated 75% of the total urinary radioactivity. Analysis of PhIP metabolites recovered from antibody columns is facilitated by the intrinsic fluorescence of PhIP and its metabolites, providing sufficient sensitivity to monitor individuals for the levels of PhIP excreted following consumption of typical western diets.

Effects of diet and folate on levels of micronucleated polychromatic erythrocytes.

We describe a study examining the effects of fried beef consumption on the frequency of micronuclei in RNA-positive polychromatic erythrocytes (PCEs) in humans. 5 splenectomized individuals participated in a 2-week experiment consisting of 3 phases. During the first and last phases the subjects refrained from eating cooked meat for 4-5 days. This was designed to clear their system of mutagens found in cooked-meat products. In the second phase each person ate a similar diet, but also consumed 4 quarter-pound well-done hamburgers per day for 4 days. Erythrocyte-folate levels were also measured for each donor. No association between diet phase and micronucleus frequencies was observed among the 2 subjects with normal levels of folate. However, among the 3 low-folate donors, the frequency of micronucleated PCEs appeared to be associated with diet. Micronucleus frequencies began to increase 1 day following onset of the beef phase, and started to decline 1 day after finishing this phase. These observations are in agreement with erythrocyte maturation kinetics following short-term exposure. A repeat study performed on one of the low-folate donors consisted of two beef phases separated by a vegetarian phase. Beef in one phase was fried (high in mutagenic amino-imidazoazaarenes [AIAs]) while the beef in the other phase was fried after first being microwave-cooked (low in AIAs). Significant increases in the micronucleus frequency were not observed in this experiment, suggesting that AIAs formed during frying were not solely responsible for the induction of micronuclei.

The effect of dose and enzyme inducers on the metabolism of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) in rats.

Male Fischer 344 rats were given a single dose of 0.03-30 mg/kg of [2-14C]2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine ([14C]PhIP), the radioactivity in urine and feces was determined over 48 h, and the major metabolites were identified and quantified. Dose had little effect on the profile of metabolites in the urine but did influence the profile in the feces. PhIP was more efficiently metabolized at higher doses. In addition, rats were pretreated with Aroclor 1254 (PCB), 3-methylcholanthrene (MC), phenobarbital (PB), PhIP and corn oil prior to a single dose of [14C]PhIP, and compared with a control group receiving [14C]PhIP only. The major metabolites in the urine and feces were quantitated for each group, as well as PhIP binding to serum proteins, hemoglobin and selected tissues. Pretreatment with MC and PCB resulted in an increase in the amount of 4'-hydroxylation of PhIP and a decrease in the amount of N-hydroxylated metabolites in the urine. Pretreatment with PB resulted in an increase in the amount of N-hydroxylated metabolites, but a decrease in 4'-hydroxylation. Pretreatment with either MC or PCB resulted in an increase in PhIP binding to the liver and kidney, while reducing the binding in other tissues. Animals pretreated with PhIP showed few significant differences from the untreated group, while pretreatment with PB in general resulted in a decrease of PhIP binding in tissues.

Azido- and nitro-PhIP, relatives of the heterocyclic arylamine and food mutagen PhIP--mechanism of their mutagenicity in Salmonella.

Azido-PhIP (2-azido-1-methyl-6-phenylimidazo[4,5-b]pyridine) and nitro-PhIP (2-nitro-1-methyl-6-phenylimidazo[4,5-b]pyridine) have been synthesized and characterized chemically. The mutagenic potencies of azido-PhIP (with photoactivation by near UV irradiation), of nitro-PhIP (without exogenous activation) and of the heterocyclic amine PhIP (with activation by rat liver S9) were evaluated by means of the reversion assay in Salmonella typhimurium. Like PhIP, azido- and nitro-PhIP were potent mutagens in the strain TA98. In addition to TA98, the strains YG1024 with increased and TA98/1,8-DNP6 deficient in acetyltransferase activity and TA98NR deficient in nitroreductase were used. Photolysis of azido-PhIP generates a very short-lived mutagen whose mutagenic potency is similar in all strains used and therefore not dependent on the acetyltransferase and nitroreductase. These findings are analogous to previous ones with azidofluorene and suggest that a short-lived DNA-binding arylnitrenium ion is formed directly by the photolysis of azido-PhIP. The nitroreductase contributes to the mutagenic potency of nitro-PhIP; in TA98NR it is 37% of that in TA98. The acetylator status of the strains influences the mutagenic potencies of PhIP and nitro-PhIP only weakly. This contrasts with findings obtained with 2-amino-3-methylimidazo[4,5-f]quinoline (IQ) and nitro-IQ and with other amino- and nitro-imidazoarenes which lose most of their mutagenic activity in the acetyltransferase-deficient strain. This atypical behavior of PhIP suggests that the presumed metabolite, N-hydroxy-PhIP, differs from the N-hydroxy metabolites of other heterocyclic amino- and nitro-imidazoarenes in that in Salmonella it is not significantly activated by O-acetylation. It must therefore either react directly with DNA or be activated in a different way. This divergent behavior of PhIP and N-hydroxy-PhIP in Salmonella may also be a key to the understanding of several 'unusual' properties of PhIP in mammalian cells and organisms.

Immunochemical detection of rodent hepatic and urinary metabolites of cooking-induced food mutagens.

Monoclonal antibodies, previously selected to bind either 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) or 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) were evaluated to determine their binding properties with several known metabolites of these compounds purified from rat hepatocyte cultures. Both 2-N- and 5-substituted MeIQx metabolites were recognized by antibodies AIA-2 and AIA-11, while antibodies AIA-1 and AIA-12 bound N-substituted metabolites only. Anti-PhIP antibodies bound N-hydroxy-PhIP, N-sulfinamide-glutathione-PhIP and N-hydroxy-glucuronide-PhIP. Immunoaffinity columns incorporating these antibodies were used to concentrate and purify both the unmetabolized parent compounds and several relatively non-polar metabolites from the urine of rats exposed either to MeIQx or PhIP. Several of these metabolites corresponded with available standards of previously identified polar conjugate metabolites, e.g. N-sulfamate-MeIQx and N(OH)-gluc-PhIP, while others are as yet unidentified. Immunoaffinity chromatography is demonstrated as a promising means of selectively concentrating metabolites of PhIP and MeIQx from urine.

Cell kinetics of rat 9L brain tumors determined by double labeling with iodo- and bromodeoxyuridine.

Rats with 9L brain tumors received intraperitoneal injections of iododeoxyuridine (IUdR) and bromodeoxyuridine (BUdR) to estimate the duration of the deoxyribonucleic acid (DNA) synthesis phase (Ts) and the potential doubling time (Tp) of individual tumors. Different sequences and intervals (2 or 3 hours) of IUdR and BUdR administration were evaluated. After denaturation, tumor sections were reacted with Br-3, a monoclonal antibody that identifies only BUdR, and then were stained immunohistochemically by the avidinbiotin complex method. An antibody that recognizes both IUdR and BUdR, IU-4, was applied to the sections and identified by the alkaline phosphatase-antialkaline phosphatase method. Nuclei labeled only with IUdR stained blue, while those labeled with BUdR or with BUdR and IUdR stained brown. The fraction of cells that either left or entered the S-phase during the time between administration of IUdR and BUdR was measured to calculate Ts and Tp, assuming that the labeled cohort completed the DNA synthesis at a constant rate. The Ts was 8.8 hours (coefficient of variation (cv) = 0.05) and the Tp was 64.2 hours (cv = 0.08). The sequence and interval of administration of IUdR and BUdR had a minimal effect on Ts and Tp. In studies of 9L cells in monolayer culture, the Ts was 9.6 hours (cv = 0.08) and the TP was 30.6 hours (cv = 0.06). Double labeling with BUdR and IUdR allows the duration of the S-phase and potential doubling time of individual brain tumors to be estimated in situ from a single biopsy specimen.

Role of metabolism on the DNA binding of MeIQx in mice and bacteria.

We report the effects of several inducers of P450 metabolizing enzymes on DNA adduct formation by 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) in C57BL/6 mice. We also examined the role of N:O-acetylation and the nitrenium ion in the genotoxicity of MeIQx, since these have been implicated in the activation of other aminoimidazoazaarenes (AIA) to DNA reactive species. Mice were given phenobarbital (PB), Aroclor 1254, beta-naphthoflavone (BNF) or corn oil, i.p., followed 3-5 days later with oral administration of MeIQx. Induction of Aroclor and BNF produced DNA with 8-fold more adducts than either the corn oil-alone or PB-treated animals. Both corn oil-alone and PB-treated animals were similar. Four major adducts were found in all cases with no differences among inducers as judged by co-chromatography. Azido-MeIQx induced calf-thymus-DNA adducts produced identical adduct profiles to those seen for the mouse DNA. Similar adduct profiles were obtained from Salmonella TA98, and the nitroreductase deficient strains (TA98NR and TA98/1,8-DNP6) exposed to MeIQx in the presence of Aroclor-induced-mouse-liver S9. Adduct frequencies in TA98/1,8-DNP6 were significantly lower than in TA98 and TA98NR. The data described in this report demonstrate that induction quantitatively increases adduct numbers but does not affect the types of DNA damage. These data also suggest that the same DNA reactive intermediates are formed in vivo as in vitro and support the hypothesis that the metabolism of MeIQx involves the P450I family of isozymes, N:O-acetyltransferases and possibly a nitrenium ion. The application of radioanalytic scanners for quantitation of 32P-postlabelling adduct maps is described.

Flow cytometric analysis of erythrocyte populations in Tn syndrome blood using monoclonal antibodies to glycophorin A and the Tn antigen.

Flow cytometric analysis employing monoclonal antibodies to the Tn antigen and glycophorin A was used to characterize the erythrocyte populations present in blood samples from individuals with Tn syndrome. Four monoclonal antibodies specific for the Tn antigen, Gal-NAc monosaccharide, on human erythrocytes were obtained from a fusion of splenocytes from a Biozzi mouse immunized with red cells from a Tn individual. These monoclonal antibodies specifically recognize GalNAc monosaccharide sites located on the erythrocyte cell surface sialoglycoproteins, glycophorin A and glycophorin B, and do not bind to fixed normal red cells presenting the Neu-NAc alpha 2-3Gal beta 1-3(NeuNAc alpha 2-6)GalNAc alpha 1-O-Ser(Thr) tetrasaccharide or to fixed neuraminidase-digested cells presenting the Gal-GalNAc disaccharide. The percentages of Tn-positive red cells in samples from six unrelated Tn donors ranged from 28 to 99%. Binding of the glycophorin A-specific monoclonal antibodies showed that the erythrocytes composing the Tn-negative fraction presented normal amounts of the M and N epitopes on glycophorin A. The presumed somatic mutational origin of Tn-positive cells was tested in blood samples from five normal donors; three possible Tn cells were observed after analysis of a total of 1.1 x 10(7) erythrocytes, suggesting that the frequency of such cells in normal individuals is less than 1 x 10(-6).

Monoclonal antibodies to 2-amino-1-methyl-6-phenylimidazo[4,5-b] pyridine (PhIP) and their use in the analysis of well-done fried beef.

Four monoclonal antibodies, PhIP-1, -2, -3 and -4, have been produced that bind to PhIP (2-amino-1-methyl-6-phenyl-imidazo[4,5-b]pyridine), a mutagenic aromatic amine produced in meat by cooking. The immunogen was an analog of PhIP conjugated to a carrier protein through the 2-amino group of PhIP, leaving the pyridine and phenyl rings unmodified. The antibodies show a range of binding specificities for PhIP and its structural analogs, with PhIP-1 being the most highly selective, requiring the entire PhIP molecule with the exception of the exocylic amino group to be present for recognition. A competition enzyme linked immunosorbent assay (ELISA) incorporating these antibodies was developed and applied to the analysis of fried meat extracts differing in mutagen content. In extracts of fried beef, more PhIP-like material was found than would be expected based on chemical estimates of the amount of PhIP, suggesting that frying produces other compounds that are structurally similar to PhIP. These PhIP-like compounds were separated from PhIP using HPLC. Addition of creatinine to the beef prior to frying increased both the mutagenicity of the extracts and the amount of PhIP-like material. Microwave pre-treatment prior to frying reduced both immunochemical reactivity and mutagenicity.

Double labeling with iodo- and bromodeoxyuridine for cell kinetics studies.

The rate of progression through the cell cycle was determined in five human glioma cell lines by a new sequential immunohistochemical staining technique. The cells were labeled first with iododeoxyuridine (IdUrd) for 1-3 hr and then with bromodeoxyuridine (BrdUrd) for 30 min. Labeled cells were identified with Br-3, a monoclonal antibody that recognizes only BrdUrd, and with IU-4, an antibody that recognizes both IdUrd and BrdUrd. Each slide was stained sequentially, first with the immunoperoxidase method for Br-3 and then with the alkaline phosphatase-anti-alkaline phosphatase method for IU-4. Cells that were positive only for IU-4 represented the fraction of S-phase cells that passed into the G2 phase during the period of incubation with IdUrd. The rates of progression measured by this method were constant in each cell line and resulted in smaller standard errors than were obtained by measurements from specimens stained singly for IdUrd and BrdUrd in different slides. The duration of the S-phase calculated from this fraction in the five cell lines ranged from 8-13 hr; the estimated potential doubling times were 25-32 hr and were very similar to the actual doubling times.