Downstream Process Optimization Opportunities Using On-Line and At-Line PAT Instrumentation

Traditional chemical analysis in downstream biotechnology manufacturing operations has been performed in off-line laboratories which are physically and organizationally remote, from the manufacturing site. Results of these assays are often not provided in a timeframe that is useful in controlling and monitoring the downstream process with maximum efficiency. Reports of delays (ranging from hours to days) in obtaining results of analytical monitoring of downstream processes often relegates these functions and the corresponding data to a compilation of historical databases rather than to its intended use as a tool to manage and control the real-time process under observation. Modern manufacturing processes in other industries, such as semiconductor and petrochemical processing, have long recognized and employed the advantages of performing analytical measurements at the production line and have widely used these techniques to enjoy advantages in production efficiency and product quality. The comparison and contrast to the current state of biotechnology manufacturing process efficiency and quality is particularly evident. FDA has long considered application of on-line analytical techniques as credible opportunities to improve product quality and efficiency in pharmaceutical production, including biotechnology and biological production. As a result, FDA began investigating technologies and prospects for application of automation in pharmaceutical, biotechnology, and biological manufacturing processes in ten years ago.

The purpose of the PAT initiative is to move analytical laboratory functions close to the manufacturing process and to improve manufacturing efficiencies and product quality. This would be accomplished by providing real time support and control of manufacturing processes through analysis of the process stream coupled with statistical process control and tight feedback control loops.

The PAT Initiative

The FDA PAT initiative was created formally by the Center for Drug Evaluation and Research (CDER) branch of the FDA in 2002¹ and sought to provide a framework to employ these techniques for small molecule pharmaceutical production under guidance from FDA. PAT processes² are becoming an established tool to promote and improve quality and production efficiency in the pharmaceutical manufacturing process. The successes of early implementation in small molecule manufacturing processes led early adopters in the biotech community to explore PAT techniques and processes for potential use in biotechnology applications. It has been demonstrated that the same PAT technology used in traditional pharmaceutical manufacturing can be employed in biotechnology manufacturing with the same positive effect. To that end, FDA is now focusing its PAT initiative on biotechnology and biologicals manufacturing through its Center for Biologics Evaluation and Research (CBER) and PAT offices.

Successful PAT programs are defined by a careful combined application of statistical process control (chemometrics) and use of on-line, at-line, or near-line sensors and instrumentation capable of measuring key parameters of the manufacturing system in real time. These capabilities are arguably most critical and useful in the final stages of biotechnology manufacturing, where the value of the product is at its highest. Downstream processing applications are a specific target for implementation of PAT processes because this production stage is where the desired product of the fermentation or cell culture process is separated from the complex production matrix, then concentrated, identified, quantitated, and collected for final processing and formulation. These operations are particularly suited for use of on-line, at-line, and near-line sensors and analytical instrumentation associated with corresponding chemometric and process control systems.

Statistical process control and chemometric systems are employed across a broad spectrum of industries and are not further explored here. References to general chemometrics and chemometrics specific to biotechnology and biologicals are widely available.^3,4,5 The focus of this article is on-line sensors and instrumentation specific to downstream processing.

Existing State of the Art

The primary analysis requirement in biotechnology downstream processing is molecular identification and quantitation; used to identify and quantitate the product to determine its purity after separation from its production matrix by process liquid chromatography. Therefore, the most powerful and useful analytical techniques to consider for downstream processing in the PAT context are spectroscopy and analytical scale separation methods that can characterize analytes on a molecular basis. A typical downstream process scale chromatographic technique not only purifies but concentrates the analyte and is operated at high load factors and high flow rates, therefore, the associated analytical technique must have a wide dynamic range and a fast cycle time. Typical cycle time requirements for on-line downstream assays are under 15 min, with many under 10 min. This cycle time requirement plus the need to separate, characterize, and identify the desired product in an often complex matrix composed of chemical species similar to the desired product, requires precise, high speed, purpose-designed instrumentation characterized by high accuracy, reproducibility, and resolving power. This combination of requirements limits the discussion of applicable downstream biotech PAT instrumentation to near infrared spectroscopy (NIR) and other specialty optical spectroscopy techniques, such as mass spectroscopy, liquid chromatography, and electrophoretic techniques.

NIR has enjoyed some success in pharmaceutical manufacturing operations, primarily in blending and formulations of dry bulk materials. Its use in on-line downstream processing applications for identification and quantitation of product is less established, but its possibilities should not be overlooked. Practical on-line NIR analyses for biotechnology applications have been demonstrated.⁶ Factors that will control use of NIR for downstream process applications include probe design and placement, calibration methodology, and heavy reliance of NIR on statistical analysis to calibrate and report concentration and identity of analytes. In many applications, a major advantage of NIR is non-contact. In these applications, the spectrometer is not in physical contact with the sample and, therefore enjoys long and reliable operation. In the case of a downstream process application however, the spectrometer samples the analyte by inserting an optical probe into a product transfer line at an appropriate sample point. This has the advantage that the sterility of the downstream system is maintained and that there is no physical withdrawal of sample from the process line. However, with the probe in direct contact with the fluid sample matrix, proper orientation and maintenance of this probe is critical to the technique's success. Fouling resulting from deposition of biological material on the probe quickly renders the measurement unusable. Furthermore, the low frequency characteristics of the spectra produced in NIR requires heavy reliance on calibration by multivariate solution of simultaneous equations. The sample and its matrix must be fully characterized and non-variate for the calibration to remain valid. This is not a simple task in a complex sample and matrix characteristic of biotechnology production. The speed and simplicity of instrumentation do, however, make it appropriate for consideration for downstream process applications and early applications of this technology appear promising.

Mass spectroscopy (MS) is a powerful and well appreciated tool, finding broad acceptance in a number of non-traditional applications including biotechnology. For example, mass spectrometers have been installed and used in upstream applications for measurement of headspace gas analysis in cell culture and fermentation reactors, to monitor O₂ and CO₂. Mass spectrometry can be used to determine absolute analyte identity and for quantitation—both desirable attributes. It is rapid and has high resolving capability. In downstream processing, a mass spectrometer could however, acquire a sample stream from an automated sample port or sampling system at the appropriate point in the process stream. Mass spectrometer applications in a downstream process are effectively prohibited by typical high salt content found in many biotech process streams. All known commercial fluid interfaces for mass spectrometers are characteristically intolerant of salt content in the fluid stream. Development of a successful salt resistant or salt remediating interface would open downstream processing to the power of MS. This would appear to be the primary and only limiting factor in successful, robust application of MS in biotech process applications, notwithstanding the high acquisition and operating cost of a mass spectrometer.

Liquid chromatography has been successfully employed in pharmaceutical PAT applications for decades. Of particular note are the pioneering examples of on-line high performance liquid chromatography (HPLC) applications by Eli Lilly in biotechnology applications.⁷ On-line downstream applications of HPLC for biotechnology are now widely accepted and deployed. Commercial systems for biotech process monitoring are available from Dionex and other suppliers. This technique is mature, well characterized, and well accepted in all fields of chemistry separation including biotechnology. The system also requires capture of a physical sample from the process stream while maintaining sterility of the stream. Recent advances in on-line sampling for biotechnology, including introduction of commercial products that maintain process sterility, will further increase utilization of the HPLC technique in downstream applications.⁸ The chief drawback of this technique for process characterization appears to be the complexity of fully automated HPLC systems and their attendant control processes. These include column switching, fast column equilibration, and high flow systems, all of which are strategies employed to reduce cycle times to remain consistent with downstream needs.

Electrophoretic Techniques

Electrophoretic techniques, most notably slab gel electrophoresis for protein and genetics analysis, have been a mainstay of biotechnology development analytical laboratories since its commercial inception in the 1970s. Slab gel electrophoresis is simple in concept, inexpensive to acquire, does have high operating costs, and has cycle times inconsistent with the needs of downstream processing. Capillary electrophoresis (HPCE), was introduced in the late 1980s and automates the electrophoresis process in a format similar to HPLC. Many commercial instruments are available today. Data are quantitative in contrast to slab gel techniques. It is rapid. It has recently found widespread application for protein analysis in off line laboratories operated in support of downstream applications, specifically for protein product quality, characterization, and quantitation. The technique is inherently fast, versatile in the biotechnology environment, and amenable to the demands of protein assay in the downstream process environment. The technique is employed daily in this capacity, but still has failed to gain wider appreciation and acceptance for downstream processing assay requirements because commercial instruments have been designed exclusively for standard benchtop laboratory use. Manual sample acquisition, preparation, sample introduction, and data reduction are the rule as is the fact that most HPCE systems must be operated by HPCE specialists. These instruments have remained exclusively in the province of remote laboratories operating in support of downstream operations and these instruments do not promote the advantages of PAT technology in downstream operations.

New Downstream Technology

We are introducing an on-line and at-line process analysis technology specific for protein identification and quantitation in a downstream process application. This instrument⁹ fully automates the entire process and is purposely built for on-line assays for protein identity and titer as envisioned by the FDA PAT initiative. The instrument incorporates and can be used on-line with an automatic, sanitary sample port interface, or can be used at-line with samples introduced by the instrument's autosampler. In either case, the entire sample introduction, preparation, separation, detection, and data reduction are performed automatically by the instrument without operator input. Once the sample is introduced to the instrument it is transferred directly to the sample preparation module, which delivers an appropriately prepared sample to a HPCE module that automatically performs the separation. Detection is performed by an ultraviolet absorbance detector. Data reduction and reporting are performed by a data system based on standard HPLC data and user interfaces. The concentration dynamic range of the instrument is from ?g/mL to g/mL due to the internal automatic dilution-concentration system component of the sample preparation module. Molecular weight range for the instrument is 10 to 200 kD. Calibration of the instrument is through a timed, automatic introduction of calibration standards in a manner typical of calibration for similar HPLC systems. Cycle times are consistent with the needs for most downstream processing.

Conclusion

Biotechnology manufacturing, particularly downstream processes, has many operations and opportunities suited to the premise of the PAT initiative—namely, improving production efficiency, yield, and product quality. Several technologies have been implemented in downstream processing for the purpose of on-line assays that have proven the concept. New, versatile instrumentation technologies for downstream process applications will continue to emerge as the PAT initiative becomes more fully implemented in biotechnology and biological manufacturing.

Quick Recap

George Barringer, PhD, is president and CSO of Groton Biosystems, 85 Swanson Road, Boxborough, MA 01719, 987.266.9222, fax 978.266.9223 barge@grotonbiosystems.com

References

1. www.fda.gov/cder/OPS/PAT.htm

2. PAT definition per FDA CDER. A system for designing, analyzing, and controlling manufacturing through timely measurements (during processing) of critical quality and performance attributes of raw and in-process materials and processes with the goal of ensuring final product quality. It is important to note that the term analytical in PAT is viewed broadly to include chemical, physical, microbiological, mathematical, and risk analysis conducted in an integrated manner.

3. Sharaf MA, Illman DL, Kowalski B. Chemometrics. John Wiley and Sons; 1986.

4. Brereton RG. Chemometrics: Data Analysis for the Laboratory and Chemical Plant. John Wiley and Sons; 2003.

5. Brereton RG. Chemometrics and PAT. PAT Journal. 2005 2 (3) 8-11.

6. Willis RC. PAT Pending. Modern Drug Discovery. 2004 December.

7. Cooley RE. Utilizing PAT to Monitor and Control Bulk Technology Processes. Eli Lilly and Co. University of Michigan Pharmaceutical Engineering Seminars. 2003 March 4 .

8. On-line sampling systems for biotechnology applications are supplied by Groton Biosystems, YSI Instruments, and Nova Biosystems, among others.

9. Groton Biosystems GPA1000 Process Analyzer.