Gently does it

Gentler methods of processing proteins using supercritical fluids promise to make them even more valuable for pharmaceuticals, say Kevin Shakesheff and Steve Howdle. 

Proteins can have marvellous pharmacological effects, but they pose new problems in the development of medicines. These problems centre on the fragile nature of protein molecules because they often ’survive’ for very short times within the body and need to be continuously manufactured by cells. So the key question is: How do you deliver these delicate molecules successfully within the patient? 

Answering this question is an urgent issue for the pharmaceutical industry. Advances in biotechnology, and post-genomics, are driving the discovery of exciting protein drugs that promise to treat major chronic diseases (Table 1). For many proteins entering clinical development there are three crucial delivery issues: patients will have to take protein drugs for long periods; oral administration is difficult because proteins are digested by enzymes in the gastrointestinal tract; and many proteins must be held at one location within the body to have a beneficial effect. 

Table 1. Some protein drugs in development

One route to address these challenges simultaneously is to wrap proteins within a biodegradable polymer delivery device. This delivery vehicle slowly releases the protein as the polymer breaks down once it reaches the desired site in the body. Not only can the medicine be taken less frequently, but delivery is restricted to the site where required. 

Despite the promise of such an approach, making polymer delivery devices poses practical problems. Organic solvents and high temperatures used in polymer processing cause many proteins to lose their function rapidly. At the University of Nottingham we (the authors) have developed new ways of processing proteins so they can be incorporated into delivery systems without using a solvent or raising the temperature above 37?C.

Our collaboration began in early 1997, when we met by chance and discussed some recent work by Steve to make porous polymer materials using supercritical carbon dioxide (scCO₂). Under supercritical conditions (T>31.1?C, P>7.38MPa) carbon dioxide is neither a liquid, nor a gas, but has properties of both (Chem. Br., August 2000, p23; February 1995, p118). Like a liquid, scCO₂ can dissolve a wide range of solutes, but like a gas it has low viscosity and high diffusivity. 

The first polymer to undergo processing in scCO₂ was poly(DL-lactic acid) (P_DLLA). The racemic stereochemistry of P_DLLA, synthesised by ring opening polymerisation of DL-lactide (Scheme 1), generates an amorphous polymer with a glass transition temperature - above which the polymer ’melts’ or liquifies - of 45?C. Conventional processing of P_DLLA typically involves either a chlorinated solvent such as chloroform or heating to a temperature that significantly lowers the polymer viscosity (>80?C). 

Scheme 1. Ring opening polymerisation of lactide to produce P_DLLA

Our breakthrough was to realise that scCO₂ acts as a molecular lubricant to the P_DLLA. The low viscosity and high diffusivity of scCO₂ ensures that the fluid penetrates very effectively into the solid polymer. Then the CO₂ molecules interact with the backbone of the polymer and act as a plasticiser to allow the polymers chains to move over each other. There is no such plasticisation effect with any other gases, for example N₂, since most gases do not enter into a molecular interaction with the polymer that causes plasticisation. 

The net result is that these molecular interactions allow the liquefaction of P_DLLA, with molecular weight in the range of 1kDa up to 150kDa, to occur at temperatures below 37?C. In fact, this plasticisation is so effective that P_DLLAcan even be liquefied at 4?C, at high scCO₂ pressures. Since this initial study we have begun to demonstrate that a wide range of polymers are plasticised by scCO₂. These include the poly(lactide-co-glycolide) copolymers. In addition, we have found that it is possible to reduce the viscosity of some semi-crystalline polymer melts by this method. For example, this is the case with poly-e-caprolactone, another important polymer for drug delivery applications. The ability to process any crystalline polymers broadens the range of potential applications of the supercritical fluid method because semi-crystalline polymers are especially useful where a high-strength material is required. 

This solvent-free and low temperature process suggested to us that as well as plasticising the polymers, scCO₂ might also provide an excellent route to adding proteins, and maintaining their activity. To see if this would work, we mixed a model protein, catalase, with P_DLLA. Under scCO₂ conditions, a helical impeller (’dough-mixer’) was employed to mix the protein and the polymer. Depressurising the supercritical fluid leads to foaming and solidification of the polymer, trapping the protein within a polymer scaffold. 

We have been able to assess the impact of this scCO₂ processing on the activity of a series of enzymes. The rate of release of catalase was controlled by the slow rate of polymer hydration and hydrolysis. Furthermore, the enzyme activity of catalase was unaffected by exposure to scCO₂. In fact, catalase function is completely unaffected by scCO₂ exposure for over one hour at pressures above 7.4 x 106Pa and at temperatures between 30 and 50?C. 

Since that first proof-of-principle experiment, we have ’processed’ many proteins in scCO₂ and have mixed these with a number of different biodegradable and non-biodegradable polymers. To date, we and other groups have looked at a number of proteins including ?-galactosidase, ribonuclease A, vascular endothelial growth factor, and bone morphogenetic protein-2; protein activity is unaffected in all cases. 

Notwithstanding, this process of ’supercritical fluid mixing’ would be of limited use but for another feature of the interaction between the polymer and the CO₂: the ability to control material architecture during the depressurisation step. As the pressure in the autoclave is lowered, the number of molecules of CO₂ plasticising the polymer chains falls. As a result, the glass transition temperature (Tg) rises and the polymer solidifies. The shape and internal architecture of the resulting solid device is controlled by the depressurisation conditions. If the CO₂ pressure is slowly released whilst the polymer is confined within a small vessel then a monolith structure with internal pores forms. 

P_DLLA can be plasticised and then ’foamed’ into a monolith. Figure 1 shows an example of the internal pore structure of such a monolith. By contrast, pressure can be released as the plasticised polymer flows through a narrow orifice into a large collecting chamber. In this second case, particles are formed as droplets of the polymer-CO₂ mixture harden, leading to a fine free-flowing powdered material. 

Fig 1. Supercritical fluid mixing can be adapted to produce porous scaffolds and microparticles

At this stage, we turned our attention to translating the benefits of supercritical fluid mixing into improved materials for specific medical applications. Tissue regeneration provided an obvious first application. Using porous biodegradable scaffolds, researchers have grown cartilage, bone, skin and a range of more complex tissues (Chem. Br., June 2000, p32). To encourage tissue formation in patients, embedded growth factors (themselves protein therapeutics) could be added to these scaffolds. 

Supercritical fluid mixing is an attractive route to achieve this because not only does it protect our growth factors during processing, but the rate of CO₂ release from the scaffold after mixing determines the size and interconnectivity of pores. At the end of the process, a porous sponge is created in which the polymer forms the major structural component and contains embedded growth factors. 

Figure 2 demonstrates that the pores of the scaffold support cell populations in the early stage of tissue formation. This scanning electron microscopy image shows human osteoblast sarcoma cells adhering to the surface of the polymer within pores and starting to produce their own natural polymer matrix. Polymer scaffolds produced by supercritical fluid mixing can also be made to deliver growth factors to promote bone and blood vessel regeneration.

Fig 2. Growing a tissue on a porous scaffold

Sizing up
Aside from polymer scaffolds, supercritical fluid mixing may prove even more useful in making drug delivery vehicles. Particles of polymer that release their drug payload after injection into the patient are the standard format for drug delivery systems. Initially, producing particles at the end of the supercritical fluid mixing process was difficult. Then, in 1998, we met an industrial team who were already tackling a similar problem but for very different reasons. 

Ferro Corporation in Ohio, US, was investigating the use of scCO₂ to create powder coatings: polymeric particles containing pigments or metals. These are the basis of paints used to coat a wide range of consumer products, from cars and buildings to fridges. Ferro has developed a relatively simple spraying process in which pigments and non-biodegradable polymers are mixed to produce particles on a batch size of hundreds of kilograms. This is ideal for powder coatings where tonnes of particles are sold at very low unit cost, but impractical for making protein formulations for the pharmaceutical industry, where only milligram quantities of protein are available. 

With this in mind, we had to learn how to scale down production, whilst maintaining particle size and yield. The challenge was to introduce the same efficient mixing and controlled pressure release on the laboratory scale, and to ensure that any losses of protein material were minimal. 

In the past two years, our attention has turned to transferring these ideas from an academic environment to industry. The UK Research Council’s Business Plan Competition provided support, advice and a mentoring coach, and it forced the team to think about our technology: how could it deliver what the pharmaceutical industry needed? 

In May 2002, our Business Plan was awarded first place and, armed with the ?25,000 prize money and additional support from the University of Nottingham, we have now set up Critical Pharmaceuticals. This new company is taking our patented ideas and doing the long, but necessary, process of validating our methods and testing new protein therapeutics. We have both maintained a curiosity-driven academic role in which we are looking for new ways to use supercritical fluids in challenging applications. Overall, supercritical fluid mixing demonstrates another example of how a clean, low temperature environment can allow us to process a delicate drug gently whilst dramatically changing the physical state of, and ultimately the morphology of, a polymer. 

Source: Chemistry in Britain

Acknowledgements

Kevin Shakesheff is professor of tissue engineering and drug delivery and Steve Howdle is professor of chemistry at the University of Nottingham.

Further Reading

• Novel therapeutic proteins: selected case studies, Klaus Dembowsky and Peter Stadler (eds). Weinheim: Wiley-VCH, 2001.
• S. M. Howdle et al, Chem. Commun., 2001, 109.
• M. S. Watson et al, Adv. Mater., 2002, 14 (24), 1802.
• X. B. Yang et al, J Bone Miner. Res., 2003, 18 (1), 47.
• E. J. Beckman, Env. Sci. Technol., 2002, 36 (17), 347A.
• T. P. Richardson et al, Nature Biotech., 2001, 19 (11), 1029.
• D. Bratton, M. Brown and S. M. Howdle, Macromolecules, 2003, in press.