Kirsten Strahlendorf, Kevin Harper and Ian Verhappen.
Sanofi Pasteur, (www.sanofipasteur.us), the vaccines division of Sanofi-Aventis Group, a multinational company devoted entirely to human vaccines, is in the clinical-trial stages of the development of multivalent vaccines designed to immunize against two or more micro-organisms. The equipment that was selected for small-scale batch manufacturing needed to be able to support resulting bulk formulations with two to five antigens. The system used to prepare these formulations was comprised of single-use, pre-sterilized bags, filters, connectors and tubing/transportation assemblies with associated metering and mixing equipment. An important consideration in the design of the system was that any processing downstream of the final filter is considered "closed" from the surrounding environment because doing so saves the significant expense associated with the need for ISO Class 5/Grade A clean room or isolator conditions. Doing so increases the level of sterility assurance, while also reducing the required cleaning steps and the cost and energy requirements of the system.
To meet these requirements, the equipment must be:
- Non-invasive and suited for a pre-sterile, single-use, closed process line during ingredient addition and mixing processes;
- Reliable and accurate to assure the proper ratio of each of the ingredients in the final formulation while minimizing giveaway;
- Suitable for current good manufacturing practices (cGMP) used to comply with Food & Drug Administration (FDA) and industry minimum standards.
In addition, small and mobile equipment and processing aids are preferred so they can be moved from one manufacturing area to another with minimal investment/reinvestment.
Despite the relatively minimalist nature of the equipment being used, the process itself is actually quite complex, with batch operations used to create the desired product and where the process itself uses multiple different proteins, multiple buffers, adjuvant—chemicals added to an antigen to increase the body's immunologic response—to create multiple formulations.
Here's how it works. Proteins are individually adjuvanted and diluted in intermediate bags. Once these intermediate proteins are in the proper state, they are blended into the final formulation bulk bag and further diluted.
Detailing the process further, the ingredients are primed into the processing lines followed by a flushing of the main line and filtration assembly with buffer. Ingredients are measured (volume by weight) individually into the primed and tared intermediate formulation bags. Each intermediate bag is then mixed. The final formulation bag is primed with buffer and tared, and then each of the intermediates is successively added. Once all of the intermediates have been added, the required remaining diluting buffer and adjuvant are added to the final formulation bag. The lines are sealed, and the final formulation is mixed.
Because this is an early clinical phase trial, it is effectively a small manufacturing scale operation as shown in Figure 1, and this application used bags of up to 5 L working volume. Bag sizes used are 250-mL (for bioburden sample) and 1-L, 2-L, 5-L (for product).
The commonly accepted industry estimate for a single vaccine batch is around $1M / batch in early phase trials. The consequence of a bad batch is either reprocessing or remanufacturing. Reprocessing could mean very little in terms of delays. Remanufacturing, on the other hand, can take additional weeks or months. It is therefore critical to get both the process and associated measurements right.
Obtaining the right measurements for this entire system requires an accuracy of ±2.0 g, which equates to ± 6.6% for a 30-g product amount added, and ± 0.04% for a 5000-g product amount added. Because of the nature of the system—which includes flexible tubing, flexible bags and stirring/rocking plates—the resulting measurement difficulties included a changing center of gravity. These difficulties, combined with the above-mentioned requirements for cleanliness, led to the decision that the most viable measurement option available was the use of hanging load cells.
As part of the entire system, readings from the load cells after ingredients are pumped into the hanging bags have an average percentage difference of ± 0.15% (n = 35, practical minimum and maximum weights applied) compared with target weight. The source of the error is largely due to human interventions and the dynamics of the fluid pumping into the bags rather than the accuracy of the load cell readings.
The maximum calibration range for the load cells is set for 20 kg so they would not be overloaded by a full 20-L bag. Each "production bag" has a minimum load of at least 30 g material, unless the product is being taken for a sample where the accuracy is not so critical. Bags as small as 50 mL have been tried in the system; the problem with these bags is that when they hang on the load cells, they are too light and cannot be properly stabilized. The measurement problem was compounded by the fact that the lines to which the bags are attached cause movement that affects the mass reading on the load cells, leading to inaccuracies. Therefore, it was necessary to use hanging load cells to make these measurements. With larger-sized bags, tubing holders can be used to stabilize the lines, preventing signal drift on the load cells.
As with any new process there are always a number of challenges. In this case they are associated with dispensing and weighing accuracy, mixing and scalability.
Hanging load cells were selected because the system is so dependent on accurate measurement and control. Overcoming the dispensing and weighing accuracy challenges were critical to the project's success. As a result, the selected load cells had the following features:
- Ability to withstand high lateral forces caused by the "pulling" of the tubing connecting the bags to the process and bracketing assembly and the pulsations associated with the peristaltic pumps;
- Required accuracy of 0.02% of cells;
- Readings unaffected by thermal or vibratory interference;
- Moveable load points to compensate for "shifting" of the bag on the strain gauge, not only as a result of changes in the bag profile, but also the "rocking" assembly used for mixing after weighing.
After some research, the selected weighing solution was a cantilevered weigh cell manufactured by BLH (http://cimail15.blh.com/process-weighing/). The KIS 3 Load cell is unique in that its "double" cantilevered design as shown in Figure 2 locates the load force application point directly above the strain gauges. This provides the following:
- Side-load force sensitivity is virtually eliminated, negating the effect of the tubes and rocking motion of the mixer;
- Moment stresses upon the gauges are zero, compensating for any slippage of the bag along the axis of the strain gauge, eliminating the need for stay rods and other expensive mechanical supports, while allowing for the motions to promote mixing of product down the line after weighing;
- The cantilever design balances the bending force to the center of the beam, reducing the bending stress at the mounting base by 50%;
- Shear stresses remain constant for higher repeatability;
- Calibration tolerance limit is ± 0.03%.g it up
Liquid ingredient addition (whereby the liquid has similar fluid characteristics to water) is volume added based on mass, with known specific gravity and starting and resulting concentrations. Weigh cells are a suitable measurement technology for this application as the multiple, changing masses of the ingredient additions require an in-line, multi-channel weighing instrument with panel readout.
These load cells are converted from strain gauge transducer signals to weight information using a G4 multi-channel weighing instrument. As a result, the measurement signal represents the only true applied force, thus making it possible to achieve the measurement accuracies required for this installation.
As this process migrates to larger scale production, it will be necessary move to larger bags which, due to their weight, will have to be supported with a more rigid floor-mounted structure. Fortunately, the BLH weighing system, forming a key part of the process and control system, can also be scaled to this configuration.
Accurate sensing and control is a critical part of the development of new biopharmaceuticals, not only to maximize return on investment through the elimination of reprocessing and remanufacturing, but also providing proper measurements to meet the strict documentation requirements and cGMP constraints for approval of products undergoing clinical trials.
Opportunities arise as the industry moves from stainless steel setups with fixed probes, to flexible and disposable mixing and reaction vessels, which require high-accuracy measurement irrespective of holding vessel geometry.