A methodical approach to ultra-scale-down of process sequences: application to casein removal from the milk of transgenic animals.

Scale-down models of individual operations are widely used in biopharmaceutical process development to obtain information about the performance of production-scale equipment on the basis of inexpensive and efficient laboratory-scale tests, for the purposes of validation or optimization or characterization studies. We have investigated the ability of scale-down models of whole process sequences to provide reliable information for process scale-up from laboratory- to pilot-scales of operation. Using the example of the recovery of a protein from transgenic milk, we have conducted an a priori scale-down analysis of a projected pilot-scale process sequence. A systematic approach was developed to ensure that all critical aspects of process behavior were included in the scale-down model, resulting in the creation of an accurate and reliable scale-down representation of the pilot-scale process. The data from scale-down process trials conducted at 70 and 200 mL scales of operation served to highlight crucial factors determining process performance, and proved reliable in predicting the performance of the pilot-scale process over a scaling factor of 1000.

The influence of major components on the direct chromatographic recovery of a protein from transgenic milk.

This work presents a systematic evaluation of the influence of lipids and casein on the performance of a chromatographic capture step for the recovery of a target protein from transgenic milk. Lactoperoxidase (LPO) was spiked at concentrations typical of those to be expected for transgenic proteins in commercial bovine milk and the dynamic adsorption of LPO to fixed beds of SP Sepharose FF studied in frontal analysis experiments. By removing successively selected components from whole milk, their individual influence on the dynamic adsorption behaviour of LPO could be studied. A mathematical model, fitted to the breakthrough curves of LPO, provided a quantitative measure of parameters describing mass transfer and adsorption in the column. A significant reduction in column capacity for LPO in the presence of milk or whey was recorded, which could be attributed to competing adsorption of alkaline earth metal ions to the cation exchange resin. While the high concentrations of lipids present in whole milk did strongly reduce the column permeability, no significant influence of either casein or low concentrations of lipids on the hydraulic properties of columns or on the adsorption of LPO could be detected. The results indicate that chromatography, which forms an essential part of all current large-scale processes for the recovery of proteins from transgenic milk, could potentially be moved further upstream. Alternatively, existing operations for the removal of lipid and casein could be re-designed so as to maximise product yields. This suggests that significant product losses during current pre-chromatography milk purification could be reduced or potentially even avoided.

Prediction of shear damage of plasmid DNA in pump and centrifuge operations using an ultra scale-down device.

Supercoiled circular (SC) plasmid DNA is often subjected to fluid stress in large-scale manufacturing processes. It is thus important to characterize the engineering environment within a particular unit operation as well as within the associated ancillary equipment during process design for plasmid DNA manufacture so as to avoid shear-induced degradation of the SC isoform, which would compromise product efficacy in therapeutic applications. In the past few years, ultra scale-down (USD) tools were developed within our laboratory to mimic the engineering environments experienced by biomolecules within a range of manufacturing-scale ancillary, primary recovery, and purification operations, using milliliter quantities of material. Through the use of a USD shear device, the effect of elongational strain rate on SC plasmid DNA degradation was studied in this paper, and from that, the impact of a centrifugal pump, a Mono pump, and a disk-stack centrifuge feed zone on SC plasmid DNA degradation was predicted and experimentally verified at scale. Model predictions, over the range of conditions studied, were in good agreement with experimental values, demonstrating the potential of the USD approach as a decisional tool during bioprocess design.

A capillary cytometer method to quantitate viable virus particles based on early detection of viral antigens and cellular events within single cells.

The traditional plaque forming and TCID(50) methods to determine replication competent virus titres rely on several cycles of replication and infection to generate a plaque with an incubation period of 24-72 h post-infection typically required. We developed a method to quantify infective viral particles based on early detection of cellular events by capillary cytometry. The method uses a capillary cytometer as a precise cell counter that can discriminate infected from non-infected cells. The general protocol was developed using a Guava PCA, genetically modified HSV-1 virus and polyclonal antibodies against antigens expressed on the cell membrane. Infection was detected after 1 h incubation and a plateau in the number of infected cells was observed between 7 and 9 h. A good correlation between titres obtained by the plaque forming method and the proposed method was observed for a ratio of infected to total cells between 0.5 and 0.05. The rapid and automated analysis (10 s/1000 events acquired per sample) makes the method particularly useful for high-throughput applications. The proposed method can be extended easily to determine the titre of other viruses providing a powerful tool for virology and antiviral screening.

Kinetics of whole serum and prepurified IgG digestion by pepsin for F(ab')2 manufacture.

An alternative route for the production of polyclonal F(ab')(2) fragments that might be adopted for the facile preparation of antivenoms is assessed in this work. The method involves the digestion of whole serum by free pepsin, which results in reduction of the number of processing steps commonly in use, because it avoids the initial purification of IgG's prior to their proteolytic cleavage by the enzyme. Digestion kinetics of whole serum and caprylic acid prepurified IgG using free pepsin were monitored with SDS-PAGE followed by densitometric analysis and antigen binding activity assay of the digested samples. It was observed that with equal units of pepsin activity, caprylic acid prepurified IgG was digested more rapidly than whole serum but that the overall retention of antigen binding activity was significantly greater in the latter case. The estimated first-order digestion rate parameters were 11.8 and 4.42 microM min(-)(1) for pure IgG and whole serum, respectively. The K(m) value obtained for whole serum digestion was 33 microM and that for pure IgG digestion was 43.5 microM. Calibration with undigested whole serum and pure IgG samples of known concentrations was performed using SDS-PAGE followed by image analysis. A linear relationship was observed between the protein concentration and the respective band intensity within the range of concentrations investigated (0.63-31.2 microM IgG concentration). This technique proved to be relatively rapid, reproducible, and more precise than size-exclusion chromatography as a result of its F(ab')(2)/IgG resolving power. Staining and destaining protocols were reproduced in terms of staining and destaining times, volumes added, and compositions. Furthermore, all digestion experiments were performed in duplicate sets to monitor the extent of variation of the digestion kinetic parameters measured by this method. The results obtained from this technique confirm and quantify previous observations that pepsin digestion of whole serum is slower and easier to control than digestion of pure IgG and results in higher recovery of antigenic binding activity.

Demonstration of the use of windows of operation to visualize the effects of fouling on the performance of a chromatographic step.

The high resolution afforded by packed bed chromatography makes it an indispensable operation in the downstream processing of therapeutic molecules. Packed bed performance is however inherently susceptible to changes in feed stream characteristics and fouling processes. The impact of fouling is seldom considered during the early stages of bioprocess development which is concerned with the selection of purification conditions. Instead these are performed with rigorously clarified feeds. Under such conditions, chromatography is effectively treated as an isolated step, independent from its preceding unit operations. In this study, we demonstrate how windows of operation could be used to visualize the impact of changes in the preceding clarification step on the fouling response of a subsequent cation exchange capture step. Laboratory columns (2,5 and 12 cm height) were subjected to varying fouling challenges of Escherichia coli lysate containing different amounts of solids carried over from the previous step. Changes in trans-column pressure drop and breakthrough of the target protein (Fab') were monitored. The limits of operability of the resin were determined with respect to the process material's properties. This information was used to extract the parameters for the adsorption kinetics used in the general rate (GR) model to create windows of operation for manufacturing scale operation.

Visualising fouling of a chromatographic matrix using confocal scanning laser microscopy.

Confocal scanning laser microscopy (CSLM) was used to visualise the spatial location of foulants during the fouling of Q Sepharose FF matrix in finite batch experiments and for examining the subsequent effectiveness of clean-in-place (CIP) treatments in cleaning the heavily fouled beads. Beads were severely fouled with partially clarified E. coli homogenate by contacting the beads with the foulant for contact times of 5 min, 1 or 12 h. The use of two different fluorescent dyes, PicoGreen and Cy5.5, for labelling genomic PicoGreen-labelled dsDNA and protein respectively, allowed the direct observation of the chromatographic beads. The extent of fouling was assessed by measuring the subsequent adsorption of Cy5.5-labelled BSA to the beads. Control studies established that the labelling of BSA did not affect significantly the protein properties. In the control case of contacting the unfouled matrix with Cy5.5-labelled BSA, protein was able to penetrate the entire matrix volume. After fouling, Cy5.5-labelled BSA was unable to penetrate the bead but only to bind near the bead surface where it slowly displaced PicoGreen-conjugated dsDNA, which bound only at the exterior of the beads. Labelled host cell proteins bound throughout the bead interior but considerably less at the core; suggesting that other species might have occupied that space. The gross levels of fouling achieved drastically reduced the binding capacity and maximum Cy5.5-labelled BSA uptake rate. The capacity of the resin was reduced by 2.5-fold when incubated with foulant for up to 1 h. However, when the resin was fouled for a prolonged time of 12 h a further sixfold decrease in capacity was seen. The uptake rate of Cy5.5-labelled BSA decreased with increased fouling time of the resin. Incubating the fouled beads in 1 M NaCl dissociated PicoGreen-labelled dsDNA from the bead exterior within 15 min of incubation but proved ineffective in removing all the foulant protein. Cy5.5-labelled BSA was still unable to bind beyond the outer region of the beads. A harsher CIP treatment of 1 M NaCl dissolved in 1 M NaOH was also ineffective in removing all the foulant protein but did remove PicoGreen-conjugated dsDNA within 15 min of incubation. Cy5.5-labelled BSA was able to bind throughout the bead interior after this more aggressive CIP treatment but at a lower capacity than in the case of fresh beads. The competitive adsorption of BacLight Red-labelled whole cells or cell debris and PicoGreen-conjugated dsDNA was also visualised using CSLM.