September 1, 2006 (Vol. 26, No. 15)

Trend is Toward All-in-One Instrumentation

Bioprocess monitoring is rapidly becoming a tool for achieving process understanding and control and, ultimately, for realizing high yields and quality product. Understanding which analytes matter most within a cell culture or fermentation process takes considerable upfront development work. A myriad of parameters are potentially critical: temperature, agitation rate, cell density, and viability, plus concentrations of nutrients, gases, and waste products. How well control over these parameters affects the final product depends on the manufacturer’s commitment to process development and understanding.

Increasingly, bioprocessors are adopting all-in-one instruments that analyze multiple parameters and produce readouts at assigned time points. According to Bob Fox, biotechnology corporate accounts manager at Nova Biomedical (, the difference between traditional small molecule pharmaceuticals and biologics is that the former are well characterized, both for molecule and process, and therefore relatively well controlled. “In biotech, regulatory agencies are still struggling with understanding what it means to control a process with product quality as the goal.”

Bioprocessors know the product they want, but are often restricted by limitations of available analytic tools. Consequently, the FDA has felt pressure to become an agent for moving biotech into the 21st century by encouraging monitoring and testing to demonstrate that the process is in control.

Imagine a three-week manufacturing process during which only essential process conditions were analyzed, but only intermittently and offline, in a separate laboratory. That scenario would be unthinkable in the semiconductor industry, but that is how most bioprocess analytics are carried out. Manufacturing practice is trending toward more analytics through multiple offline instruments, but that level of sophistication and automation will not support current quality and process understanding initiatives. Clearly, more analytics are needed, performed online or at-line, in an automated or semi-automated fashion through a unified platform.

Nova Biomedical has consolidated all critical nutrient, metabolite, and gas tests into one instrument, the BioProfile® Analyzer, which replaces as many as six stand-alone analytical methods. Taking it a step further, Nova’s new BioProfile FLEX system measures all of these critical parameters, plus cell density/viability, and osmolality in cell culture media in one compact instrument.


According to the company, consolidating these key tests optimizes workflow, saves time and labor, minimizes sample handling, and preserves valuable laboratory space. Consolidation also streamlines data collection and reporting, and improves regulatory compliance in every phase of the bioprocess.

One option with the BioProfile Analyzer is an on-line autosampler, which can automatically sample from up to four bioreactors simultaneously as fast as every six minutes and connect to a bioreactor controller system. “An analytical group would have a difficult time taking this many samples manually,” says Harlan Polishook, marketing communications manager at Nova.

Automation improves the consistency of testing, plus adds a level of documentation, accountability, and control that was previously difficult to achieve. “One of the biggest issues today,” says Polishook, is the need for connectivity to channel data into one location and in one format and to eliminate the need for operators to make decisions.”

Like many companies in the bioprocess monitoring marketplace, YSI Life Sciences (,/a>) entered through the medical testing/monitoring route. YSI commercialized the first glucose analyzer for whole blood in 1975 and participates in this arena with its 2300 STAT Plus™ clinical blood analyzer.

The upgradeable, multiparameter 7100 MBS analyzer from YSI reportedly measures up to six process parameters simultaneously through three sensor modules containing two chemistry sensors each. Users can mix and match chemistries from among glucose, sucrose, lactate, glutamate, glutamine, ethanol, methanol, ammonium, potassium, and galactose. The last one is a recent addition at the request of one particular customer. According to Steve Grant, eastern U.S. manager at YSI, several additional biotech companies have requested the galactose sensor.

Through a partnership with Groton Biosystems (, which manufactures an online sampler that runs with the 7100 analyzer, YSI is developing products for process analytic technology (PAT). Together, the 7100-sampler combination will extract and analyze materials from up to 10 bioreactors. Grant sees opportunities for the combination in the emerging biofuels marketplace, as well as in traditional biotech.

When the FDA promulgated its PAT guidance in 2004, it envisioned process analyzers replacing more traditional analytical laboratories and services. That’s still a big part of the PAT push, but thinking has broadened to include some element of feedback and control to garner the most benefit from PAT.

Dionex ( has been active in online analysis since 1985 and in 1999 introduced a version of its analyzer that operates with Dionex analytical HPLC instruments to support biomanufacturing and other types of chemical analysis. What’s interesting here is that the Dionex DX800 process analyzer detects and quantifies amino acids and carbohydrates directly, without the need for pre- or post-column derivitization. These molecules are difficult to analyze through spectroscopy-based detectors.

“Bioreactors contain thousands of different molecules,” notes Rick Cooley, who manages the Dionex Process Analytics Center of Excellence. “Chromatography offers the ability to resolve them, and focus on single analytes to determine their impact on the product.” The DX800 acquires samples and multiplexes analytical results from up to 21 individual process streams. It also internally and automatically performs sample preparation (dilution, concentration, addition of reagents). “We’re taking operations that would normally be done offline in-lab, and performing them without human intervention,” Cooley adds.

Automation is the Key

“Life science workers have been trying to get away from doing things manually, with or on paper,” says Daren Moffat, life sciences business development manager at Invensys ( Automation in the life sciences began with human-machine interfaces (HMIs), which place a computer atop a process to allow operators to do more with less paper.

A favorite control strategy in fill-finish areas are PLCs (programmable logic controllers), computers designed for process control. PLCs tend to be point-oriented, connecting to one piece of a process, providing noncontinuous or on-off control.

A distributed control system (DCS) tends to cover a manufacturing plant more comprehensively, over wider and more continuous operations. The distinction between the two types of controller is beginning to blur, according to Moffat. “New facilities tend to specify a mix of DCS and PLCs, dictated by the process requirements.”

A Wall St. Journal article several years back blasted the life sciences industry for running its manufacturing plants with very little automation. “Potato chip manufacturers had more sophisticated systems,” Moffat observes. At the time a good deal of the blame was placed on the FDA and its burdensome regulations for validating and amending a process.

Now that the Agency has actively promoted automation, PAT, and other science-based initiatives, the industry at large has only slowly switched from manual to automated operations. But larger pharmaceutical and biotech companies have embraced automation, and manage to implement it while staying within regulatory dictates. According to Moffat, some larger firms operate at nearly “lights out” while smaller companies “still have guys walking around with clipboards.”

Moffat ties an appreciation for automation with what he calls “operational excellence,” a term he broadly defines as taking more pride in well-run manufacturing operation, and utilizing automation systems to generate data that underlie better decisions. “R&D used to get all the attention in the pharmaceutical industry, but companies now realize the value of optimizing manufacturing, of getting more products out of the same equipment, less expensively, and at higher quality.”

Automation and control is not only about the ability to have machines do what humans previously did. “The goal is to understand how a process runs so that it can be improved, how to take hundreds or thousands of data points and turn them into actionable information.”

In April, Invensys introduced its Infusion Enterprise Control System, which provides integration across multiple control and data systems in a context that managers need to see. Moffat describes the system as “closed-loop business control” as opposed to “closed-loop process control,” which takes it up a level in organization. “Infusion doesn’t just integrate systems, it brings everything already in the plant into a common information structure across one plant or multiple plants.”

Quality by Design

The philosophy of monitoring and control is changing from years past, when product quality was tested by applying quality measurements to individual samples from a lot, to designing quality in (quality by design).

“This approach makes accepting or rejecting a lot quite easy,” says Greg LeBlanc, director of product marketing at OSI Soft ( OSI, which serves oil/gas, pulp, and paper industries as well as biotechnology, offers the PI System as the data collection and distribution engine for its Real-Time Performance-Management platform for manufacturing plants and processes.

PI System serves as the time-series repository for all the events on the operations floor, delivering real-time and historical data for subsequent decision-making. PI records years of process data at high speed, in real time, and at its original fidelity, according to the company. With PI engineers can locate a specific run or series of runs, add them to a report, or apply them to a trending package to see how they fit into a time series. “You can compare any run to last year’s, or to an expected run,” says LeBlanc.

Improved communication between devices has enabled the success of products like PI. During the mid-1980s, when process monitoring and control were in their infancy, every instrument and device possessed a communication protocol. OSI Soft therefore built up a huge library of interfaces for connecting these devices. Today, most interfaces are standard, but some still employ proprietary protocols. The company still maintains about 400 off-the-shelf connectivity software packages for these devices to bring them within the PI system.

OSI Soft’s other notable control product, ProcessPoint™ 4.0, is a web-based, collaborative software tool that provides a real-time, centralized system of record for all product information from idea through scale-up to production. ProcessPoint enables customers to automate specification, recipe, and product formula management, as well as new product development. The solution eliminates errors and duplication of efforts during scaleup by automatically managing product versions as they change.

ProcessPoint also provides an audit trail and can track parameter excursions and incorporate that data into PI, so process engineers can view the history of a particular batch’s manufacturer.

Previous articleAnna Barker, Ph.D.
Next articleAffymetrix Selected to Genotype SHARE Heart Study Samples