Finesse Solutions

Many factors are currently conspiring to change the pricing models for pharmaceuticals and/or biologicals and therefore their acceptable manufacturing costs and what constitutes acceptable time to market. These changes are, in turn, driving changes in the requirements of the modern bio-processing facility.

The advent of FDA-approved bio-similars such as Omnitrope (Sandoz) and the general “offshoring” of the bio-similar industry have gathered much attention in the US and Europe.    The combined effect of recent changes in the U.S. healthcare programs and the expiration of key patents is also expected to significantly affect drug prices.  The facilities built to develop and manufacture bio-similars can benefit from high-yielding, robust new cell lines, and therefore do not require the same production volumes as previous therapeutics.   Moreover, a major goal of the emerging bio-similar production facilities is to minimize capital expenditure and diversify the risk with multi-product capability.  Additionally, the resurgence of pandemic threats and bio-terrorism has accelerated investment into vaccine production where the desired timeline from genetic analysis to first production lot is now several months and where rapid scale-up and process optimization is critical.

These drivers, namely multi-product capability, rapid scale-up and capital efficiency have accelerated the adoption of single-use systems by the cell culture community.  Over the past decade,  single-use bioreactors have evolved from simple small-volume rocking platforms with minimal measurement and control, to sophisticated systems whose capacity and capabilities rival, if not exceed, 10,000L stainless steel bioreactors.   In contrast to the stainless steel and glass of the previous decades, single-use systems allow for a higher level of automation and integration.  This is true for both raw information and for the peripheral elements that allow for control of at-line and offline sensors.   When properly executed upon, the control system can allow for a high level of efficiency while reducing the requirements on the skill level of the operator.  Interestingly, the same points also make these “smart” single-use systems attractive for reducing cost and overhead in up-stream processing in established drug production facilities.

II.    Existing control systems/platforms

The current paradigm for control of a bioreactor most frequently relies on the use of proprietary controllers (PLC’s, PAC’s, and DCS’s) where the fundamental designs have not significantly changed in the last 20 to 30 years.  The software used to program PLC and PAC type controllers often has its basis in coding techniques like “ladder logic” which is significantly older than the controller itself and has its origins in code designed primarily for use by electricians and plumbers.   The DCS systems are slower yet far more flexible than PLC’s, but also use proprietary code and lack the ability to use a scripting language which limits their ability to interface with modern computers and microcontrollers.  As the software for these proprietary controllers is at root based on antiquated and proprietary frameworks and therefore aimed at a limited market, it is almost impossible for these vendors to keep up with the pace of innovation in the multi-billion dollar modern consumer electronics and software industry.  When using these proprietary systems, the bio-processing professionals or their operators are therefore required to maintain a workforce that is knowledgeable and skilled in these limited venues.  Such workers are generally in short supply and therefore expensive to keep on staff.    It is interesting to note that while this is the norm in the bio-processing industry, even the US military and the US government are turning more towards mainline techniques except where specific environments like space, dictates otherwise.  This demonstrated recalcitrance to modernization in the bioprocessing industry is presumably due to the FDA and its published guidances related to 21CFR Part 11.  However, careful examination of these documents clearly reveals that what is dictated, almost without exception, is the “least burdensome approach” (see for example: General Principles of Software Validation; Final Guidance for Industry and FDA Staff, January 11, 2002:  http://www.fda.gov/cdrh/comp/guidance/938.pdf) and the authors of these guidances are almost without exception concerned with employing a well defined method of development, validation, tamper-proof data management (e.g.: record keeping) and documentation as opposed to  mandating a specific path of achieving the end goal. 

III.    Guiding principles for TruBio μC

With the aforementioned observations in mind, our modern bioreactor control platform, TruBio μC, based on the Wintel (Microsoft Windows operating systems/Intel microprocessors) combination has been developed.  This platform uses the latest in embedded processors (similar to those used in military systems, automated teller machines, and medical devices) and leverages off of the economics that drive the consumer electronics market.  This system, TruBio μC, has been designed to free the bio-processing professional or their operators from many of the mundane tasks that defocus them and either slows their progress towards completing their runs and minimizes situation by which the batch is not usable.

 By utilizing this type of platform, a real-time control system can be implemented and yet modern graphics and modern GUI (graphical user interface) can be constructed while easily maintaining an economically viable product.   The Wintel based GUI was designed, ironically, with the Apple iPod™  in mind.  Specifically, the goal was to make a system intuitive enough that a manual was not required to operate the system.   This approach minimizes the training burden in organizations where the workforce:

1.    Changes frequently (i.e: interns);
2.    Is not highly skilled;
3.    Has no knowledge of general principles of automation systems.

These three objectives and more have been met by using techniques now standard in computer games and in general office programs.  Objects that are “discoverable” automatically become active when the cursor is placed over them.  Almost without exception, all menu pages are “1-Click” away from the main page, allowing the user to set things up and modify them quickly and without having to navigate a complex menu tree (see the Overview page in Figure 1).  Wherever possible, standard or internationally recognized icons were used with the goal of minimizing the time to learn and navigate the system.  The final 3rd objective was achieved by careful design work up front and utilization of the processing power now available.  The details of this are described below.

Figure 1: Main page for a lab-scale glass vessel.  Note the modernized user interface and intuitive graphical menu.

IV.    TruBio μC Features
The major sets of features which are of interest to the end user are the simplicity of use and “plug-and-play” capability.  It is quite instructive to use any modern computer and operating system and note that when the peripherals are plugged in, there is no need for a service call, no need to re-wire, and most often no need to actually load a “driver” by oneself.  Why then in the 21st century should there be a need for a service call when a new sensor is plugged into the bioreactor control system?  Why should a change in pumps or mass flow controllers require the user to re-set the control system in any way? Why should the model number and serial number etc. not be available to the user the way a printer’s information is available on your desktop computer?  Why should the user not be able to perform upgrades or new plug-in loads themselves?

Another serious complication for the user is the need to continuously tune the PID (proportional, integral, differential) loops for each vessel, and often different growth phases in the run.  This need can lead to a requirement of constant attention and “tweaking” by the user and often leads to sub-optimal growth results as the incorrectly tuned loops allow the temperature or pH wander in and out of acceptable range.  The use of a modern processor/operating system again lends itself well to providing relief from these issues.
In TruBio μC, the ability to use a scripting language and utilization of communication modes common to the personal computer allow for the “smart peripherals” such as Mass Flow Controller (MFC) boxes and pump towers to be automatically recognized by the control system when their cables are connected.  The information from the peripherals that affects the control loops or sets parts of the GUI self-populates.  The current paradigm with traditional control algorithms and traditional sensors is very labor intensive – therefore usually very costly and time consuming and prone to error.  The use of manpower to do calibrations that require attention to detail, repetition, and interpretation invariably introduces both fine and gross error; data entry is notoriously tedious andloop tuning becomes a black art learned only be a few.

Figure 2: Save/load file management system in TruBio μC software.

Technology currently exists that allows for “smart” chips or tags so to be integrated into the single-use component such that the calibration and validation data can be automatically read into TruBio μC.   Additionally, inexpensive microcontrollers and flash memory are commonplace and facilitate active and stored data communication between peripherals and the main processor.  The implication here is that the sensors, transmitters, and other peripherals as well as control systems must be engineered with these features in mind.  Employing this technology, the single-use components can be accompanied with their own calibration information as well as detailed information on the wetted parts while the permanent equipment can be auto-recognized and communicate data from itself or attached single-use components back to the central processor.   All information can also be loaded directly into the control system without human intervention .

Figure 3: Data extraction and reporting in TruBio μC software. 

Using this technology, all of the material information, performance data, and calibration data can be centrally tracked and stored.  This data can be mined to create records of the materials used in each run and allows for statistical correlations between components, conditions, media etc and end performance (e.g.: titer, cell viability, feed consumptions etc) to be investigated.  This also allows for more basic tasks like automatically storing certificates of compliance and the manufacturer’s data or linking them to the run so that producing data to comply with audits or government regulation is much simpler and quicker.   The fundamental power of personal computing and data base management can be directly applied to bio-processing.  Figure 2 shows automated process configuration files that can be loaded or saved in TruBio μC software.

Once the data is read from a smart sensor or smart peripheral, a technologically matched control system like TruBio μC can make also use of the data to reduce labor costs and human error in many other ways.   For example, sensor response time, pump performance, or similar characteristic data that affects the loop performance can be combined with other knowledge to automatically modify loop settings, timing, or math functions/calculations that depend on sampling or response times.  Similarly, automated report generation can minimize the labor associated with manual data extraction, graphing, and analysis.  Figure 3 shows automated report generation using TruBio μC software; these reports can be exported as .jpg graphs or as the raw data in .csv format

Most bioprocess specialists, operators, and interns have a background in cell biology, microbiology, or chemical engineering.  In general, this background does not include deep expertise in fluid mechanics, or programming and mathematical methods.  It therefore makes sense to develop efficacious tools that minimize the need for this skill set.  The modular, plug-and-play TruBio μC system, and the paradigm shift it represents, has so far enjoyed a warm reception, as it provides a solution to a clear market need.