Protein modifications during antiviral heat bioprocessing and subsequent storage.

Antiviral heat treatment is routinely used in the bioprocessing of therapeutic proteins as a means of reducing viral load. However, in protein formulations containing sucrose this form of bioprocessing can lead to protein modifications. Using a model protein, hen egg white lysozyme, we investigated the effects of antiviral heat treatments in the presence of sucrose on protein integrity during subsequent long-term protein storage. Although heat treatment alone resulted in protein modification, subsequent medium- to long-term storage of both lyophilized and liquid samples at room temperature or above led to further protein modifications. The majority of these modifications were due to the formation of glycation and advanced glycation end products via the reaction of reducing sugars and their autoxidation products (derived from hydrolyzed sucrose) with function groups on the protein surface. These findings have implications for the improvement of therapeutic protein bioprocessing to ensure protein product quality.

Evaluation of protein modification during anti-viral heat bioprocessing by electrospray ionization mass spectrometry.

During the preparation of therapeutic plasma and recombinant protein biopharmaceuticals heat-treatment is routinely applied as a means of viral inactivation. However, as most proteins denature and aggregate under heat stress, it is necessary to add thermostabilizing excipients to protein formulations destined for anti-viral heat-treatment in order to prevent protein damage. Anti-viral heat-treatment bioprocessing therefore requires that a balance be found between the bioprocessing conditions, virus kill and protein integrity. In this study we have utilized a simple model protein, beta-lactoglobulin, to investigate the relationship between virucidal heat-treatment conditions (protein formulation and temperature) and the type and extent of protein modification in the liquid state. A variety of industrially relevant heat-treatments were undertaken, using formulations that included sucrose as a thermostabilizing excipient. Using liquid chromatography/electrospray ionization mass spectrometry (LC/ESI-MS) we show here that protein modifications do occur with increasingly harsh heat-treatment. The predominant modification under these conditions was protein glycation by either glucose or fructose derived from hydrolyzed sucrose. Advanced glycation end products and additional unidentified products were also present in beta-lactoglobulin protein samples subjected to extended heat-treatment. These findings have implications for the improvement of anti-viral heat-treatment bioprocesses to ensure the safety and efficacy of protein biopharmaceuticals. CopyrightCopyright 2001 John Wiley & Sons, Ltd.

Mechanisms of protein modification during model anti-viral heat-treatment bioprocessing of beta-lactoglobulin variant A in the presence of sucrose.

To ensure the safety of plasma and recombinant therapeutic proteins, heat treatment is routinely applied to these biopharmaceuticals as a means of virus inactivation. However, to maintain protein integrity during heat treatment it is necessary to use high concentrations of thermostabilizing excipients, such as sucrose, in order to prevent protein damage. In this study we describe the covalent modifications inferred to a model protein, beta-lactoglobulin A, that occur during typical and extended anti-viral heat treatments. The chemical derivation and mechanisms by which these modifications arise are addressed. Heat treatment initiated hydrolysis of sucrose to glucose and fructose, which in turn were degraded to glyoxal. Glyoxal and the free reducing sugars reacted with free amino groups in beta-lactoglobulin A to yield Maillard glycation adducts and advanced glycation end products (AGEs). The major mechanism for AGE formation was via degradation of glucose-derived Schiff-base adducts. Heat treatment and glycation of beta-lactoglobulin A resulted in thiol-disulphide interchange reactions leading to protein oligomerization. A small population of beta-lactoglobulin A non-disulphide-linked dimers were also observed with increasingly harsh heat treatments. These findings have implications for (i) improvements in the safety and efficacy of heat-treated protein biopharmaceuticals and (ii) our understanding of the mechanisms of protein glycation and AGE adduct formation.

Protein modification during antiviral heat bioprocessing.

Heat treatment is routinely used in the preparation of therapeutic protein biopharmaceuticals as a means of viral inactivation. However, in undertaking virucidal heat treatments, a balance must be found between the bioprocessing conditions, virus kill, and the maintenance of protein integrity. In this study, we utilize a simple model protein, hen egg-white lysozyme, to investigate the relationship between antiviral bioprocess conditions (protein formulation and temperature) and the extent and type of protein modification. A variety of industrially relevant wet- and dry-heat treatments were undertaken, using formulations that included sucrose as a thermostabilizing excipient. Although there was no evidence of lysozyme aggregation or crosslinking during any of the heat treatments, using liquid chromatography-electrospray ionization-mass spectroscopy (LC-ESI-MS) and peptide mapping we show that protein modifications do occur with increasingly harsh heat treatment. Modifications were predominantly found after wet-heat treatment, the major covalent modification of lysozyme under these conditions being glycation of Lys(97), by either glucose or fructose derived from hydrolyzed sucrose. The extent of sucrose hydrolysis was itself dependent on both the duration of heat treatment and formulation composition. Advanced glycation end products (AGEs) and additional unidentified products were also present in protein samples subjected to extended heat treatment. AGEs were derived primarily from initial glycation by fructose and not glucose. These findings have implications for the improvement of bioprocesses to ensure protein product quality.

Preliminary assessment of whole-blood, red-cell and platelet- leucodepleting filters for possible induction of prion release by leucocyte fragmentation during room temperature processing.

Universal leucodepletion is being introduced in the U.K. to reduce a theoretical risk of Creutzfeldt-Jakob disease (CJD) transmission. If CJD infectivity is associated with leucocytes, any cell fragmentation associated with filtration could reduce the potential benefit. Four types each of whole blood, red cell and platelet leucodepletion filters were assessed after holding of blood units for at least 4 h at 22 degrees C. In all cases the mean residual leucocyte content was <1 000 000 per unit, with only two individual filtered whole blood units having a leucocyte content exceeding this. Evidence of leucocyte fragmentation during filtration was sought but not found by assay of soluble elastase, beta-thromboglobulin and normal prion protein, as well as by isotopic labelling of leucocyte external membrane. These preliminary studies indicate that it was possible to prepare leucodepleted blood components by filtration at room temperature, and that this appeared not to be associated with overt cell fragmentation. Definitive demonstration that fragmentation does not occur requires the development of improved general (non-specific) assays for cell membrane fragments.

Virus removal from bioproducts using ultrafiltration membranes modified with latex particle pretreatment.

Ultrafiltration is an attractive process for virus removal from bioproducts owing to its high throughput as well as the fact that the operation is carried out under ambient conditions (damage to proteins is highly limited). The principal concern regarding the adoption of conventional ultrafiltration membranes for virus removal is the possibility of the virus passing through abnormally large pores or surface imperfections on the membrane surface. The chief principle behind the present work is to pretreat the membrane by blocking the abnormally large pores using latex particles. Experimental work was conducted to validate this pretreatment using the bacteriophage phi x 174 as a model virus. The results attained were highly encouraging. Different sizes of latex particles were tested by treating a 100 KD molecular weight cut-off membrane, and the transmission of phage (suspended in buffer) through this membrane assessed. In the absence of any particle pretreatment, a virus clearance of 4.78 log reduction value was observed for this membrane. The transmission of phage through the membrane could be reduced by an order of magnitude using 0.11 micron latex particles, or two orders of magnitude using a combination of 0.11 and 0.50 micron particles. The application of latex particles did not hinder the transport of protein through the 100 KD membrane. Protein sieving coefficients obtained using this membrane were 91%, 16% and 2%, for lysozyme, HSA and IgG, respectively.

No detectable alterations in immunogenicity following terminal severe dry-heat treatment of high-purity factor VIII (Liberate) and factor IX (HP9) concentrates.

We used monoclonal antibody ELISAs, antigen molecular size distribution, competition ELISA and neonatal mouse immune tolerance methods to detect potential neoantigen formation and increased immunogenicity following severe dry-heat treatment of high-purity factor VIII (Liberate) and factor IX concentrates. To provide positive controls, concentrates were heated in solution (70 degrees C for 2 h) to produce denaturation on purpose. The competition ELISA applied to factor IX proved particularly useful for quantifying differences between the positive control and the dry-heated/unheated concentrates. None of the test systems employed by us indicated any detectable neoantigen formation or any alteration in immunogenicity following terminal severe dry-heat treatment of the high-purity concentrates, and this finding is supported by clinical experience so far.

Some alternative coupling chemistries for affinity chromatography.

Three different coupling chemistries that have been tried and tested for use in affinity chromatography are described. These methods are particularly recommended for use by workers who do not have access to, or do not wish to use, complex organic chemical synthetic procedures. They have been demonstrated repeatedly to be reliable, efficient, low cost, and easily scaleable up or down in size. The periodate oxidation method works best with Sephacryl type gels and uses only low toxicity reagents and couples well to proteins with both high efficiency and high capacity. The vinyl sulfone method is more reactive and couples both carbohydrates and proteins. The bis-epoxide method, although less reactive, can be used under more extreme conditions of pH to couple otherwise unreactive molecules, such as synthetic polymers, drugs, and so forth.

Interaction of immobilised unfractionated and LMW heparins with proteins in whole human plasma.

The profile of proteins bound to immobilised heparins in hirudin-anticoagulated human plasma was analysed. In molar terms, antithrombin III was the most abundant protein bound to therapeutic doses of unfractionated heparin (M(r) = 12,000), whereas heparin cofactor II constituted < 1% of the protein bound. Histidine-rich glycoprotein was the only plasma protein likely to influence anticoagulant activity by direct competition with antithrombin III, though significant quantities of complement Factor H, fibrinogen, fibronectin, vitronectin and apolipoprotein B were also detected. Only traces of von Willebrand factor, complement factor I, inter-alpha-trypsin inhibitor, alpha 2-macroglobulin, serum amyloid P and transferrin were identified, and neither thrombospondin nor platelet factor 4 were measurable. Binding of both antithrombin III and histidine-rich glycoprotein varied with the ratio of heparin to plasma. Clexane (M(r) = 4,500) also bound antithrombin III, but both histidine-rich glycoprotein and vitronectin were quantitatively significant neutralising proteins. Neutralising proteins dominated the binding profile for Oligo H (M(r) = 2,200).

Inhibition of thrombin generation by heparin and low molecular weight (LMW) heparins in the absence and presence of platelet factor 4 (PF4).

The ability of several low molecular weight (LMW) heparins and unfractionated heparin (UFH) to inhibit thrombin generation, and their anti-Xa and anti-IIa activities, were measured in the absence and presence of platelet factor 4 (PF4). The LMW heparins studied were 2-5 times less potent, on a weight basis, than UFH as inhibitors of thrombin generation in platelet-poor plasma; the inhibition of thrombin generation by LMW heparins correlated better with their anti-IIa activity (r = 0.98) than with their anti-Xa activity (r = 0.69). At low concentrations of PF4, the activity of LMW heparins in the thrombin generation test was neutralized less than that of UFH, but at higher PF4 concentrations all their activities could be neutralized except in anti-Xa assays. These observations support the hypothesis that anti-IIa activity is important for inhibition of thrombin generation by LMW heparins in vitro. However, when all the anti-IIa activity of LMW heparins was neutralized by PF4, considerable inhibitory activity remained in thrombin generation and anti-Xa assays, indicating that a portion of the anti-Xa activity of LMW heparins also contributes towards inhibition of thrombin generation.

Development, optimization and use of an enzyme linked immunosorbent assay (ELISA) to measure factor VIII antigen utilizing monoclonal antibodies.

An enzyme linked immunosorbent assay (ELISA) has been developed to measure VIII:Ag in plasma and concentrates. The assay utilizes two commercially available monoclonal antibodies to VIII:Ag and provides an alternative to the established immunoradiometric assay (IRMA). It has the advantage of not requiring the use of radioactive material and human antibodies. The assay sensitivity is 0.006 u/ml and the interassay coefficient of variation is 6.3%. Forty-eight samples with VIII:Ag levels ranging from 0.006 to 1.5 u/ml were assayed by both ELISA and IRMA. The coefficient of correlation between the two assays was 0.89. In addition to measuring human VIII:Ag, it is also possible to detect antigen in several animal plasma and sera.

Human vascular endothelial cells catabolise exogenous glycosaminoglycans by a novel route.

Heparin and other anticoagulant glycosaminoglycans were radiolabelled with 125I and their catabolism by human vascular endothelial cells in culture was studied. Heparin, heparan sulphate and pentosan polysulphate were associated with the cellular fraction and incorporated into the subendothelial matrix, but dermatan sulphate was not found in either fraction. High molecular weight, fully desulphated carbohydrate chains were major catabolic products of all those glycosaminoglycans which were taken up by the cells. Pentosan polysulphate was not degraded further, but the catabolism of heparan sulphate, and to a lesser extent that of heparin, also yielded small oligosaccharides. Thus the first step in catabolism of exogenous glycosaminoglycans by human vascular endothelial cells appears to be complete desulphation, which destroys their biological activity, followed by depolymerisation of the carbohydrate chain. This alternative to the sequential action of lysosomal exoenzymes is dependent upon binding to the cell; thus dermatan sulphate, which is not associated with the cellular fraction, is not catabolised.

Direct dye binding--a quantitative assay for solid-phase immobilized protein.

A direct dye-binding procedure was established for the quantification of protein after its immobilization on a solid phase, using IgG and BSA as model proteins. The assay, which in the range 0-5 mg protein/ml gel correlates well with indirect protein determination by A280 as well as determination of protein hydrolyzed from the gel, is based on a modified Bradford dye-binding assay. As the protein coupled to the gel binds the dye, a decrease in A465 of the supernatant is measured. Three solid supports commonly used for protein immobilization (Sepharose, Sephadex, Sephacryl) were found to be compatible with the dye-binding assay while nonspecific dye binding was found to HEMA gels. Protein was coupled to Sephacryl S-1000 using three different activation methods (aldehyde, hydrazine, and adipic acid dihydrazide). Artifactual dye-binding was not observed using any of the three different "linkers." The assay is easily carried out and represents a useful tool, e.g., when optimizing procedures for protein immobilization.

Selection of antibodies for immunoaffinity chromatography.

A broad choice is available in immunoaffinity chromatography, of using polyclonal antibodies or monoclonal antibodies, and this chapter will concentrate mainly on the latter, although many points are common to both. Another broad distinction can be made between immobilized antigens (which may include immunoglobulins) used to prepare purified antibodies, and immobilized antibodies used to prepare pure antigens; again, this chapter will concentrate mainly on the latter. Until recently there was little rational choice possible in antibody selection for immunoaffinity chromatography-one simply had to use the individual animal species bleeds as available and hope for the best (1). The advent of monoclonal antibodies has, in principle, broadened our choice (2), but also has shown up the need for principles and techniques to produce, select, and use antibodies on a rational basis for immunoaffinity work (3, 4). Although this chapter will deal mainly with the selection process (5), it is also necessary to consider and control other aspects, such as immunization (6), fusion and cloning, preparation of solid-phase antigens and antibodies, and selection and testing of eluants (7). Unfortunately, these aspects are far from being exact sciences, and so a considerable degree of empirical testing is still needed (8). Much of the pain and tedium can be removed from such repetitive empirical testing by the sensible choice of automated or semiautomated assay techniques (e.g., enzyme-linked immunosorbent assay [ELISA], radioimmunoassay [RIA], immunoradiometric assay [IRMA], and the like) (9, 10), and assay development thus rests at the core of successful antibody selection for immunoaffinity work.

Some alternative coupling chemistries for affinity chromatography.

There are a great many methods published for use in coupling ligands to solid phases to provide affSty purification matrices for macromolecules. Frequently the authors are very enthusiastic about some feature of their work and commercial suppliers too will tell you about their strong points (but not the weak points) in their products.

Selection of monoclonal antibodies for immunoaffinity chromatography: model studies with antibodies against soy bean trypsin inhibitor.

To facilitate selection of monoclonal antibodies for immunoaffinity chromatography, an ELISA screening procedure was developed. The assay is based on the avidin-biotin system and provides a profile of the monoclonal antibody which is based on the binding characteristics of the antigen binding site when exposed to different elution reagents. The elution profiles of 5 monoclonal antibodies to soy bean trypsin inhibitor (SBTI) were determined and for 2 of the antibodies the results obtained in the ELISA were verified using column experiments. The affinity constants were determined for the same 5 monoclonal antibodies and no correlation was seen with the ease of elution. The elution profiles presented here are easily obtained and the results indicate that a general screening procedure for suitable combinations of antibodies and elution conditions can be carried out using an elution ELISA assay when modified as described herein.

Novel affinity chromatographic system for the single-step purification of glycosaminoglycans from complex systems using volatile buffers.

A new system for the isolation and purification of glycosaminoglycans (GAGs) and related molecules from complex systems such as plasma is described. Affinity chromatography which exploits the very high affinity between the polymeric base Polybrene and sulphated polysaccharides was used. A novel volatile buffer system composed of ammonium formate and formic acid, which allows the complete recovery of samples, was developed, and elution conditions were optimised for the separation and purification of GAGs of different charge densities. Using this system the losses associated with dialysis and desalting, frequently necessary preliminaries to further analysis, are avoided.

Hypotensive effects of stroma-free haemoglobin solutions attributable to adenine nucleotides.

Aqueous solutions of stroma-free human haemoglobin are being evaluated as potential oxygen-carrying resuscitation fluids. There are indications, however, that such solutions may produce toxic side-effects in vivo. Stroma-free haemoglobin solution produced a 50% fall in mean arterial pressure when infused into a small animal model despite containing very low levels of non-haem protein and phospholipid contaminants. This effect was not produced by haemoglobin solutions after extensive dialysis. Red cell-derived adenine nucleotides were found to be present in concentrations high enough to cause such a response (80-85 micrograms/ml). We have developed a chromatographic assay capable of predicting hypotension in our animal model and consider that the complete absence of adenine nucleotides must be confirmed in all studies concerning the possible toxic side-effects of stroma-free haemoglobin solutions.

Immunomodulatory properties of platelet factor 4: prevention of concanavalin A suppressor-induction in vitro and augmentation of an antigen-specific delayed-type hypersensitivity response in vivo.

The immunomodulatory properties of platelet factor 4 (PF4) have been examined in vitro and in vivo. This agent prevented the induction of concanavalin A (Con A)-induced suppressor cells in vitro in a dose-dependent manner but it did not affect the function of established Con A suppressor cells. This effect was not due to an enhanced production of interleukin-2 (IL-2) by lymphocytes exposed to PF4. The delayed-type hypersensitivity (DTH) reaction to sheep erythrocytes (SRBC) was used as a model for the generation of antigen-specific suppression in vivo. PF4 enhanced the magnitude of the swelling following SRBC challenge 10 days after sensitization by the i.p. route or following sensitization by both the s.c. and i.p. routes. These studies show that PF4 has immunomodulatory activities in well-defined models of cell-mediated immunity and suggest that this agent has a potential use in the dissection of events in antigen-specific suppression.