Process Changes And Product Comparability For Commercial Manufacturing

Process validation of the first commercial REMICADE manufacturing process in Leiden, The Netherlands was completed with the successful execution of five consistency batches. These were used to demonstrate reliability of the process and comparability of the product to that used in clinical trials. The results of process validation were used to support the initial licensure of REM-ICADE around the world. Not long after process validation was complete, post-approval process changes were pursued to further enhance the reliability of supply while continuing to ensure product quality. One example of these post-approval change projects, an expansion of the DPC stage, is described in detail here.

The Direct Product Capture (DPC, Stage 3) step produces a highly purified, stable process intermediate, which can be stored in a concentrated, frozen state. In 2001, two 60 cm diameter Protein A columns were employed in the Leiden manufacturing plant in a unit operation that is schematically represented in Figure 18.5. The processing of bioreactor harvests by DPC was the rate-limiting step for the production of REMICADE and the production requirements with respect to DPC could be met only at high utilization rates of both 60 cm chromatography systems.

Several scenarios identified to increase the DPC throughput were evaluated and prioritized with respect to quality and regulatory requirements, capital investments, operating cost impact, production capacity, implementation lead-time, and possible risks to all of the above.

The improvement scenario selected was an essentially linear scale-up of the DPC step that minimized the risk of significant changes to process performance or characteristics of the purified intermediate product. Scalable process parameters related to the chromatography step included the resin properties, linear flow rates, load ratio, feedstock composition, column packing, and bed

Clarified 0.45 ^m Protein A Concentration by 0.2 ,um Frozen harvests filtration chromatography ultrafiltration filtration eluate

Clarified 0.45 ^m Protein A Concentration by 0.2 ,um Frozen harvests filtration chromatography ultrafiltration filtration eluate

Clarified 0.45 ^m Protein A Concentration by 0.2 ^m Frozen harvests Nation chromatography ultrafiltration filtration eluate

FIGURE 18.5 An overview of the Leiden Direct Product Capture stage as it was operated with a 60 cm column (top) and after scale-up to an 80 cm column (bottom).

Clarified 0.45 ^m Protein A Concentration by 0.2 ^m Frozen harvests Nation chromatography ultrafiltration filtration eluate

FIGURE 18.5 An overview of the Leiden Direct Product Capture stage as it was operated with a 60 cm column (top) and after scale-up to an 80 cm column (bottom).

height. In order to scale-up the process, all of the parameters mentioned were maintained constant and column diameter was increased from 60 to 80 cm. The linear scalability of the Protein A chromatography step had been previously established in laboratory-scale experiments using columns with diameters of 2.6 and 5.0 cm and at the 60 cm manufacturing scale. Verification of comparable product quality and process performance was performed at the 80 cm scale in manufacturing.

A linear scale-up of the concentration step performed by ultrafiltration following Protein A chromatography was achieved by maintaining operating parameters including the type of membrane, its nominal pore size, flow geometry, channel height, and channel length. Important fluid dynamic parameters were maintained constant including feed and filtrate volume per membrane area, flux rates, feed, retentate, and filtrate pressures, and the concentration factor. Scale-up was achieved by a doubling of the membrane area to accommodate the 1.8-fold increase in the size of the chromatography column volume.

A schematic overview of the post-approval DPC-process is outlined in Figure 18.5. Agreements were reached with all internal stakeholders including Quality Assurance and Regulatory Affairs to perform the following validation work to support the change prior to implementation:

• Installation Qualification (IQ) of the larger chromatography columns

• IQ of the larger UF systems

• Five validation DPC runs in which all routine in-process tests and additional impurity tests were performed

• Cleaning validation of the larger chromatography column for 3 runs

• Cleaning validation of the larger UF system for 3 runs

• Additional stability testing was performed on affected batches

The following submissions in support of the described process changes were issued and approved:

• USA: Prior Approval Supplement

• Canada: Notifiable Change

These submissions contained information including:

• Description of the proposed change

• Purpose of the proposed change

• Summary of the validation work performed

• In-process test results of five consecutive DPC runs at enlarged scale

• Comparability of impurity clearance before and after the change

• Comparability of chromatographic profiles before and after the change

• Batch release data of the pre-formulated bulk (PFB) lot that contained material from the first five DPC runs performed at manufacturing scale

The implementation of the process change was governed by internal change control procedures to ensure comparability to the process defined in prior submissions. The internal change control procedure consists of initiation, review, revision, approval, and amendment (if required) of change request forms with associated implementation plans. After approval by the regulatory authorities, the process change was executed by completion of the implementation plans. Scale-up of the DPC process stage was fully implemented within 17 months and resulted in a significant throughput increase.

18.5.1 Scale-Upand Post-Approval Changes

Soon after approval of the Leiden manufacturing facility, it became apparent that demand for REMICADE would outpace production capacity. Plans to scale-up and add a second manufacturing site in Malvern, Pennsylvania were initiated in 1997. The Malvern manufacturing process consists of the same nine approved stages used to produce the drug substance in Leiden, shown in

FIGURE 18.6 1000 L perfusion bioreactor and spin filter for manufacturing of infliximab at the Malvern manufacturing site.

Figure 18.1. The Malvern facility is a twofold scale-up of the Leiden process. Changes were implemented in the manufacturing process to accommodate the capabilities of the new facility. The Malvern facility is largely hard-piped and has more automation than in Leiden. An example of this type of difference is the two different virus filtration skids employed in the facilities. A 1000 L bioreactor and a virus filtration skid employed at the Malvern manufacturing site are shown in Figure 18.6 and Figure 18.7. Additional selected changes to the process are summarized in Table 18.2.

All process changes were in place during comparability lot manufacture and were therefore included in process validation studies. Manufacturing of four consecutive comparability lots (one more than our minimum requirement of three lots) to validate the Malvern manufacturing process was initiated in October 2000. In-process controls and specifications employed in the Malvern manufacturing facility are identical to those used in the Leiden manufacturing facility. In addition to meeting all in-process specifications, clearance of impurities, host cell proteins, and host cell DNA were measured throughout the purification process. Process validation studies were conducted to demonstrate comparability of the process in the new facility with the Leiden manufacturing facility, as shown in Table 18.3.

To further demonstrate comparability, the PFB manufactured in the Mal-vern facility was subjected to additional characterization beyond routine release tests. The results of the routine release testing for four consecutive lots of PFB produced in Malvern manufacturing were compared to the results from the release tests for three Leiden PFB lots prepared in the same time frame (2001)

FIGURE 18.7 Virus filtration skid for processing of infliximab at the Malvern manufacturing site.

and three Leiden PFB process validation lots, manufactured between 1997 and 1998. It was important to demonstrate comparability to current lots being manufactured in Malvern and Leiden as well as to historical control materials that were the basis of the original process validation. All ten lots were subjected to concurrent analysis by WEHI, SDS-PAGE, IEF, and GF-HPLC assays in order to compare the lots directly (side-by-side analysis) and to minimize run-to-run variability. The PFB lots were also analyzed using additional characterization testing to demonstrate that primary structure (as determined by N-terminal sequence analysis and peptide mapping), secondary structure (as determined by circular dichroism analysis), post-translational modifications (as determined by C-terminal lysine content measurement, oligosaccharide mapping, and mass spectrometry), and hydrodynamic properties (as determined by sedimentation velocity analytical ultracentrifugation) were comparable. In addition to biochemical characterization of the product, the levels of residual impurities in the PFB prepared in the new manufacturing facility were shown to be comparable to those prepared in the previously validated facility.

0 0

Post a comment