Future of Biopharma

In 2004, the U.S. Food and Drug Administration (FDA) published guidance regarding an initiative it calls process analytical technology, or PAT. PAT establishes guidance for biopharma manufacturers to develop next-generation manufacturing processes. The main goal of PAT is to enable manufacturers to better understand and control their bioprocesses.

In order to bring products to market faster, biopharma companies must increase the efficiency with which they conduct their experiments, tests, and manufacturing processes. In order to accomplish this, additional process parameter data must be collected to achieve a complete, accurate picture of the processes going on in their bioreactors, tanks, and other apparatus.

These data—much more than biopharma companies have ever collected before—are analyzed to gain new bioprocess insights, which are used to control and optimize those bioprocesses for maximum yield, minimum waste, and shortest cycle time. This, in turn, enables them to scale up those bioprocesses from the lab to the production floor while ensuring highest product quality.

The purpose of this article is to introduce the reader to PAT concepts and challenges, how those challenges are being met by intelligent in-line sensor technologies, and what the future holds for biopharma sensors.

Although “quality by design” (QbD) sounds like a nice marketing tagline, it’s actually so much more: it’s the central tenet of the PAT manufacturing philosophy. It’s the idea that quality checks are built into every step of the manufacturing process, not just tacked on to the end. Thus, a QbD process must be carefully designed.

This is achieved in large part by determining the appropriate parameters to measure and monitor during the production process.

PAT defines “critical process parameter” (CPP) as a process parameter whose variability has an impact on a critical quality attribute. The value of a CPP, therefore, should be monitored and controlled within prescribed limits to ensure the associated critical quality attribute remains healthy.

Similarly, “key performance indicators” (KPIs) are metrics that indicate the status of each production step, that is, the health of the process. Both CPPs and KPIs must be monitored in such a way as to show, in real time, that the process is progressing as expected to result in a quality end product.

This can be trickier than it sounds. To be effective, a CPP must have certain characteristics. For starters, the CPP must be known to impact a critical quality attribute—for example, the level of dissolved oxygen (DO) in a bioreactor is known to influence cell viability. If another parameter (such as pH) can independently impact the same attribute, it should be measured and controlled as well.

Furthermore, a CPP must be shown to be a reliable indicator under all conditions. This can be complicated, because some measurement techniques are influenced by environmental conditions such as temperature and pressure. Finally, the CPP should lend itself to real-time monitoring and control. Stopping a process to make adjustments not only reduces process efficiency, it can make batch-to-batch consistency more difficult.

For these reasons, the PAT framework favors in-line measurement of CPPs over off-line measurement. In-line measurement continuously provides process status data as close to real-time as possible. Off-line measurement provides, at best, a snapshot in time— and the results may come minutes or hours after the “snapshot” was taken, too late to make adjustments or corrections. Furthermore, off-line measurements can be skewed by the passage of time, sample preparation processes, and other environmental conditions, and therefore may suffer from accuracy issues.

Thus, a certain amount of scientific inquiry needs to go into the selection and validation of CPPs for a given bioprocess.

Fundamentally, the use of real-time CPP monitoring to automatically control bioprocessing is what QbD is all about. Expanding the number of CPPs that are monitored ensures that important process data are not missed or delayed. In this way, quality monitoring and control is built into the process.

Because bioprocesses can involve extraordinarily complex interactions among physical and chemical conditions and characteristics of the cells, it’s usually not possible to get an accurate picture of what’s going on inside that bioreactor on the basis of one or two measured parameters. In bioprocessing, the more parameters you can measure, the better.

As described earlier, off-line measurement suffers not only from delays in results, but also the fact that these “snapshots” can be taken and processed only so often. There is a real chance that an important parameter will exceed a limit and return to the acceptable range between snapshots. These “invisible” excursions can cause product quality issues that are difficult or impossible to track down. Therefore, it is critical that as many parameters as possible be monitored in real time or near real time.

Knowing what’s happening with the process at all times not only helps ensure the quality of a given batch, but helps ensure consistent quality across batches. Because you can monitor and control all critical parameters throughout the process, you can optimize the conditions in the bioreactor from start to finish—and do so the same way, every time. This increases the yield and reduces rework due to quality issues. It also means there are no surprises at the end of the process. You know what it takes to yield the “golden” batch, and you have the power to make it happen every time.

Despite its many advantages, implementing in-line measurement has its challenges. An effective in-line measurement platform must overcome these challenges.

For example, bioreactors and tanks have only so much space for sensors, so the sensors need to be small enough to fit in the available space without interfering with each other or with the bioprocess being measured. Depending on the lab, space outside the bioreactor may be at a premium as well, so the in-line sensor platform should enable deployment of additional sensors without a large increase in cables, transmitters, recording devices, calibration equipment, or other apparatus.

A related problem is that of miniaturization. Some parameters, such as levels of certain nutrients and metabolites, require assays and other offline techniques that cannot yet be performed in the confines of a sensor probe tip.

Speed is also a significant concern. As mentioned previously, the main value proposition for in-line measurement is real-time data acquisition. Measurements need to be taken, transmitted, and analyzed as close to real-time as possible, so that if adverse conditions are detected, adjustments can be made before the entire batch is lost. This means that the sensor needs to respond quickly to changes in conditions.

Another problem area is the chemical specificity of the sensors. Whereas offline measurement involves sample preparation techniques to optimize the parameter being measured under carefully controlled conditions, inline measurements need to be accurate for the parameter being measured under whatever conditions exist in the bioreactor. This means that each sensor must work under a wide range of temperature, pH, pressure, the presence of various reagents, and other factors that might otherwise degrade performance.

Finally, the sensors must be durable. In addition to the normal challenges of getting a profitable product to market, the biopharma industry must observe industry best practices, such as clean-in-place (CIP), sterilization-in-place (SIP), and various hygienic protocols. The in-line measurement platform must support these requirements and withstand multiple cleaning and sterilization cycles without affecting accuracy or precision.

Designing a platform to meet all these (and other) challenges requires careful engineering, intense attention to detail, and a close partnership with biopharma manufacturers.

The use of in-line sensor technology offers competitive advantages for the biopharma labs that adopt them. For example, labs around the world using Hamilton’s Incyte Arc technology for real-time viable cell density (VCD) measurements have seen increased productivity and reduced waste in their processes. As a result, adoption of the technology has become an industry best practice.

But what in-line parameters are currently being measured, and how reliable are they? What parameters aren’t even being measured yet using in-line sensors? As described in our white paper, Biopharma PAT: Quality Attributes, Critical Process Parameters & Key Performance Indicators at the Bioreactor, here is a brief summary of the current state:

pH: One of the most critical of all CPPs, pH typically must be a maintained in a tight range to ensure cell viability and maximize product yield. Reliable in-line sensors in both analog and digital formats are available.

Total cell density and viable cell density: How are the cells doing? A direct way to answer this question is to measure the total cell density (TCD) and viable cell density (VCD) in the environment. The main in-line sensor technologies for this key performance indicator include near-infrared sensors that detect turbidity, and capacitance probes that distinguish live cells.

Titer and product quality: For certain bioprocesses, the desired output, such as monoclonal antibodies and certain proteins, can be measured to determine the quantity and quality of the output. Numerous in-line technologies exist to conduct total protein titer and protein-specific titer for this purpose, such as spectroscopy, nuclear magnetic resonance (NMR), and capillary electrophoresis. Research is ongoing to optimize these approaches in a way that meets the requirements for sterility, durability, ease of calibration, and accuracy.

Dissolved oxygen (DO): All animal cells require oxygen to function, so it is critical to maintain a proper DO level in the bioreactor, although the acceptable range tends to be broader than that for pH. The first in-line DO sensors, based on an electrochemical cell, had high maintenance costs and increased measurement errors over time. A newer technology, the optical DO sensor, overcomes these deficiencies and has widely supplanted the older sensors.

Dissolved carbon dioxide (DCO₂): High levels of DCO₂ can influence pH in mammalian cells and degrade the production of secondary metabolites, and so must be carefully monitored. Although electrochemical-based DCO₂ sensors have been available for many years, they suffer from high maintenance costs and their reliability can be affected by many factors.

Exhaust O₂ and CO₂: In certain bioprocesses, the ratio of O₂ to CO₂ entering the bioreactor is compared to that being exhausted; the difference is used as a proxy for cell viability. Although in-line sensors for these measurements exist, their reliability is considered poor.

Nutrients and metabolites: Monitoring of these factors is important especially in fed-batch and perfusion bioprocesses, because feeding strategies can be adjusted during the process. The in-line sensors that exist are based on molecular spectrometry principles; they are complicated to set up, and their reliability is highly dependent on the off-line calibration required of them. As a result, in-line measurement of these parameters is rarely used.

Clearly, much work needs to be done in this area, and Hamilton is hard at work to develop the technologies for reliable, easily maintained in-line sensors for all bioprocessing parameters of interest.