Automation works

My personal experience of lab automation is limited to supervising a peptide synthesiser back in the late 1980s. The machine was eye-wateringly expensive — but it was soon paying its way in terms of productivity and research publications. So if I were a stake-holder in a company pondering whether to invest in a well-designed gadget that could automate a routine operation, I'd say: 'go for it'.

Pharmaceutical manufacturing, by its very nature, is well-suited to some degree of automation, with process control of batch reactions now being well-established throughout the industry. Those operations that lend themselves to standard off-the-shelf equipment such as automated liquid handling, which is good for compound storage and high-throughput screening, are becoming widespread too. Testing is also becoming increasingly automated, whether in R&D or quality control.

According to Tony Moran, sales and marketing director of Astech Projects (UK), which specializes in automation for drug delivery testing, it is the pharmaceutical companies themselves, rather than the suppliers, that have been driving the shift towards automation.

Advantages

There are many advantages to automation, and relatively few drawbacks. Higher throughput and productivity, with the ability to operate 24/7, are the most obvious benefits — and these are of particular significance to an industry under ever-increasing timeline and cost pressures. Consistency in process control is another advantage while repeatability of test methods provides more reliable data for the regulatory authorities. "A machine is very accurate and precise compared with a person," says Moran. Companies can also look forward to a reduction in overall costs through labour saving and a reduction in lost batches.

Moran provides an interesting example of labour saving and productivity improvement in a typical stability study of a metered dose inhaler (MDI) new drug application. Using 1000 samples as baseline, full automation can achieve the job using the input of less than one full-time employee. Semi-automation of MDI testing would require four to six employees for the job, while fully manual firing would take 12–18 employees.

And let us not forget the human advantages of automation, which "frees up clever chemists to use their brain, instead of their hands," according to Stephen Fleck, consulting business manager at TTP LabTech (UK), a company with 20 years experience of providing custom automation solutions in both development and production environments. There are also health and safety advantages in automation as personnel are exposed to less hazardous equipment and chemicals.

Challenges

The main drawback of automation is the cost, particularly for a GMP environment. Pharma investors are only going to vote for cost-effective systems with the shortest possible payback time, and it is up to suppliers to develop and deliver these. Moran acknowledges that customers can be conservative when it comes to bespoke automation. Therefore, Astech takes the approach of scaling-up automation by building it gradually with key modules. In this way, with time, they will have a fully automated platform. They also offer proof of concept modules and loan modules to allow for a limited risk trial of the concept.

The other challenge is to make automation more innovative to meet the needs of the biotech industry, and to make it more flexible in the context of medicinal chemistry and process development. Cell culture (the way biological medicines are produced) is both manually tedious and notoriously tricky to get right. It presents a different type of challenge to large-scale synthetic chemical synthesis, which is the basis of small molecule drug production. Fleck recalls how TTP LabTech developed a project for Amgen's (CA, USA) manufacture of its anaemia treatment, recombinant erythropoetin (EPO). A manual cell culture process had to be scaled up and its location moved. TTP LabTech's system involved setting up 40000 roller bottles, all of which had to be rotating and visible. The project required a novel approach to roller bottle storage and handling, as there was no available off-the-shelf equipment. The system, perhaps the biggest project ever in biotech manufacturing automation, took 25 man-years to complete, and contains more than 1 million components. A huge investment, but putting it into context, the cost was less than a day's EPO revenue.

There are now many products on the market to streamline biotech research, process development and manufacturing, such as automated cell culture systems from The Automation Partnership (UK) and the ClonePix FL from Genetix (UK), which uses imaging and robotics to pick out high-producing clones that manual methods might miss.

When it comes to medicinal chemistry and process development, systems must generally be more flexible than for manufacturing operations. Fleck says: "On the chemistry side, we are finding an increasing need for automation among graduate and PhD chemists who used to spend a lot of their time in the lab working on new formulations and reactions. There is now a driver towards increasing their productivity, without increasing headcount — so they are doing more in parallel using automation."

Chemists have long been able to set 10 or 20 syntheses in motion and walk away from the equipment, but such systems are inflexible, as they generally allow the variation of only one parameter at a time. They now want to explore more parameters such as solvent, temperature or process time in parallel. "We are seeing more demand for automation in organic syntheses, and in crystallization for polymorph screening and solid state development work," says Fleck. TTP LabTech has developed a 'batch-of-one' approach, where parallel synthesis allows the testing of many different reaction parameters in the one experiment. "This design of experiment approach provides a very efficient way of exploring experimental space," he says.

The future

Can any pharmaceutical operation be automated? And are we now moving towards a fully automated lab or production plant? If the demand is there, then it may be possible to push automation a very long way forward. "Engineers will often rise to the challenge and do what was previously thought impossible," says Fleck. In the future, he believes there will be an increasing role for 'intelligent automation' where an element of artificial intelligence is applied to a chemist's decision-making process. At TTP LabTech they are looking at how they might get computers to capture and mimic what a chemist really does.

Fleck also predicts a shift towards more continuous production in pharma and biotech (as has been the case in the chemical and food industries) because there is less down time with equipment laying idle. Flow chemistry also gives improved yield because of better control of reaction conditions, particularly temperature control and mixing. A further advantage of continuous over batch production is better accommodation of unstable intermediates. Automation companies will have to provide the corresponding automation technology — a challenge that lies not so much in the engineering, but more in the corresponding regulatory requirements of FDA and GMP because continuous mode requires a different set of process control data for validation.

Key points

Greater scrutiny of all aspects of pharmaceutical technology may also lead to new automation challenges, Fleck adds. For example, formulation development in pharma, healthcare and skincare tends to involve gels, pastes and viscous liquids. Until now, there has been a need to hand mix these products, but, given the demand to explore formulation more quickly, there will be increased pressure to automate, which will require technological innovation.

According to Moran, advances in technology have allowed automated platforms to become far more productive through greater computing power, allowing multifunctional parallel processing on fully integrated systems. The amount of data that can be stored and interrogated has also increased dramatically. This pleases the regulatory authorities and provides new opportunities for designing and improving processes.

The future for automation in the pharmaceutical industry looks exciting, providing the right people who can really put it into practice are found. "It is difficult to find scientists who understand engineering and vice versa," Moran admits. In other words, we need creative engineers who understand the science they are automating, such as the intricacies of growing cells, the way medicinal chemists think and the need for sterility. The success of companies has much to do with attracting people who can offer the balance of science and engineering that makes automation work.