Article - Issue 7, February 2001
Creating a UK Genome Valley: the role of biochemical engineering
Creating a UK Genome Valley: the role of biochemical engineering Professor Peter Dunnill OBE DSc FREng
The expression ‘Genome Valley’, coined by analogy to Silicon Valley, suggests the wealthcreating potential for the UK which could follow from new discoveries in the life sciences, particularly those in molecular genetics. Such discoveries will greatly benefit this country if we can bring them through to exploitation here. Biochemical engineering is the discipline which underpins the translation of the life-science discoveries into real outcomes such as new medicines, biopesticides and renewable chemicals. It is in the healthcare sector where the advances are most dramatic at present.
Living cells can be thought of as complex chemical factories controlled from a master computer program, the genome. Instructions from the genome are fed through molecular machines and produce catalysts, the enzymic proteins. These proteins catalyse the synthesis of a great variety of molecules from alcohol in beer or wine to antibiotics such as penicillin. There are many genes in the genome: 6,000 in a simple yeast and about 100,000 for a human. In the past the genetic instructions could be grossly manipulated by techniques such as selective breeding or by crude mutation achieved by, say, irradiation. Now we can alter the individual instructions of the genomic computer information with molecular precision. Individual genes can also be snipped out and moved between organisms or chemically synthesized and inserted into cells. A relatively easy procedure these days is to move one gene, say for a human protein, to a simpler organism.
An example shows what these techniques have already achieved in treating heart disease, one of the most common killers of the developed world. Figure 1 shows via a contrast x-ray the blockage of a main artery. The region previously served by the nutrient oxygen supply will quickly be starved and irreversible damage will occur to all tissues. Now the human system ordinarily produces an enzyme which selectively dissolves blood clots. The heart attack patient does not have enough to cope and the levels in human plasma are too low to extract this enzyme from blood donors. But the enzyme, tissue plasminogen activator, can be produced using genetic engineering techniques. The human gene for the enzyme is placed in simpler cells which can be grown in quantity in fermenters. After very careful purification the enzyme specified by the gene is given by injection in the ambulance or casualty centre and saves lives – more than a million patients have been treated over the last 10 years. Therapeutic proteins of this kind are now top-selling drugs and a new generation of vaccines is in prospect. These so-called biopharmaceuticals are expected to make up as much as half of all drugs within 10 years, representing a market of £100 billion. What is even more exciting is that these medicines often represent the first real hope of dealing with previously untreatable conditions.
The new engineering challenges
Two engineering challenges stand out for the future. The first is to cope with the increasingly complex and delicate materials that are emerging as potential medicines. The second is to speed up their development and cut the manufacturing costs. Without radical new approaches the greater complexity will push costs higher; already governments are very concerned with containing fast-growing healthcare budgets, whether private or public. Today a single injection of tissue plasminogen activator to dissolve a clot costs £1,500. A year’s treatment of a human therapeutic protein for multiple sclerosis costs £10,000 and because of this, treatment is effectively rationed in the UK. Each such medicine takes on average about 10 years to bring through from discovery to commercialisation and the cost is about £350 million.
The genetically engineered tissue plasminogen activator was synthesised by a human gene instructing a simple cell to make it. Now biologists are moving to use the gene directly for therapy. If the pure gene could be placed efficiently and safely in a patient under precise control, it would not be necessary for multiple sclerosis or diabetic patients to inject themselves many times a year with a therapeutic protein. The host of dreadful and incurable genetic diseases could be addressed and there would be a better chance of dealing with the scourge of cancer by instructing the human system to return to normality. It now seems clear that such DNA can also function as a vaccine and major efforts are being made to address diseases such as malaria and AIDS for which no vaccine yet exists.
Processing delicate materials
The first few successful examples of gene therapy have taken place but many challenges remain. One which faces engineers is that the gene which specifies a protein is a much larger entity than the protein it codes for and its less compact structure also makes it more delicate. The great size (up to one micron) causes the gene to be very sensitive to mechanical forces which occur in process equipment. At present such problems are assessed by pilot plant studies. However, this poses a particular problem with new medicines. Their development as commercial products demands extended safety testing and in the earlier phases 60% of candidates fail (on average). So until clinical promise is quite well established, serious large-scale development is too costly to risk. However, the dilemma is that once clinical promise is becoming clear the pressure to go to market is overwhelming in order to get a return on the £100 million or more typically already spent. Therefore, it is essential to compress the time needed to develop an efficient process. So far, as complexity has increased, the delays at the pilot stage have unfortunately tended to become greater.
To address this my colleagues have begun to apply methods which have some parallels with those in engineering disciplines applied to non-biological systems. Until now the complexity of biology has meant that these approaches have had very limited consideration in this field. The route that we are taking involves the techniques of scale-down and of computer-based modelling. It is widely acknowledged that refining conventional laboratory experiments often does not help to address the intrinsic constraints of fullscale process engineering. This is acutely so for biological materials. Thus it makes sense to start with the likely engineering constraints and to work backwards to small-scale mimics.
For example, taking DNA as an example, centrifuging the disrupted cells of a genetically modified type to separate the human gene in its loop of DNA from cell debris will cause no problem in a lab centrifuge because the DNA is not subjected to significant shear fields. However, when the same material passes continuously through an industrial centrifuge, in order to process several hundred litres of material, the situation is very different. There is instantaneous acceleration of every element of the fluid to the edge of the rotor with correspondingly large forces.
The shear-related damage to DNA can be assessed using a small test cell with a fast rotating disc which mimics the key features of the entry region of the industrial rotor. Such analysis demands sophisticated techniques to account for the complicated stages through which the DNA breaks down. Using fluid dynamic analysis it is possible then to link what is learned in this small device to effects in the large industrial centrifuge. As Figure 2 shows, it is close to the solid surfaces, where fluid is accelerated, that the most severe shear fields occur. Thus, by making the appropriate engineering connections it is possible to learn from a few tens of millilitres what will happen to hundreds of litres and this can be done quickly and at a low cost. Because many other biological complexes are also highly shear sensitive, such methods are broadly applicable.
Modelling complex processes
The need for modelling the underlying engineering to make this connection and the demand for computing power applies also to understanding how all the possible process options are to be assessed. The permutations of operations will be large because it can take a dozen stages to purify and formulate a biopharmaceutical. Until recently only a few of these stages had been modelled and for even fewer were these checked against large-scale experiments. We can now assemble portfolios of models which also take account of the profound interactions between the stages. For example, when biological cells must be ruptured to release products inside, the first instinct is to aim for complete breakage to achieve high product yield. However, the greater the rupturing force, the finer the cell debris and the more difficult it is to remove using industrial centrifuges, with serious consequences downstream. The sequence can now be precisely modelled. Once we can ask ‘what if’ questions at a computer loaded with process models, we can avoid having to explore all the potential routes to a genetically engineered material at a pilot scale.
The first generation of genetic engineering was concerned with changing one gene of the thousands present. However in principle there is no reason why much larger numbers should not be changed. This concept is being pursued in the new field of ‘metabolic engineering’. We can turn microorganisms into factories for synthesizing novel materials or for making more of what is needed and less of what is not. The difficulty in manipulating many genes until now has been the sheer complexity of the outcome. Successful genetic modification of microorganisms demands a model of all the cell processes. This has parallels with modelling an industrial bioprocessing plant although the number of stages is several orders of magnitude larger. Not surprisingly therefore engineers are becoming heavily involved in the necessary modelling: this new type of contribution from engineers will be common in the post-genomic era.
Changing many genes will increase the risk of unexpected outcomes. To date the use of genetic engineering for healthcare has had a good record of safety and acceptability. However, it will be vital not to be complacent and here, again, engineers have a crucial role with their formal training in assessing risk and benefit.
If macromolecules and engineered microorganisms pose processing problems, then human cells and tissue represent even bigger ones. But they are increasingly used in therapy for conditions such as Parkinson’s disease and the repair of tissues following accidents. These engineered materials are living and therefore have nutrient, and especially oxygen, demands. For cell and tissue repair, and equally for gene therapy, the material will often have to be tailored to the patient. It will need to meet the demands of the physician or surgeon (Figure 3) and optimum use will call for very close connections between scientist, engineer and (commercial or public sector) producer. This development, of ‘regenerative medicine’, could allow the UK a strong position if the data-sharing potential of the National Health Service is exploited effectively.
Where relatively small numbers of patients have to be treated individually, there is a new consequence for biochemical engineering. All the contact parts of conventional stainless steel production systems used for medicines have to be recleaned, sterilised and the outcome validated between each different material. That is simply too laborious and expensive for more patient-specific medicines produced in small quantities. Instead, plastic disposable process components will become essential, as they already are in medical laboratory-level procedures. The use of process disposables could also help start-up companies by substituting consumables expenditure, as needed, for large capital expenditure at the outset. Preliminary experience suggests that effective process sequences can be created.
The prize for industry and government
The science and biochemical engineering associated with the genome have great potential but capturing the value for the UK poses challenges to industry and government. The UK pharmaceutical industry is highly successful and spends a greater proportion of its resources on R&D than most other sectors. However, there is now a new situation. The manufacturing of simpler chemical pharmaceuticals is moving elsewhere and the reduction of British jobs in the industry is perhaps unsurprising. The general message, as in other fields of manufacture, is that the UK must address more knowledgecentred targets.
In no area of discovery does the UK have greater strength than in the life sciences. The UK was a co-founder of the field of biochemical engineering with the USA and has considerable process strengths in addressing the new materials. The country is second to the US in the number of significant biotechnology start-ups although most are not as mature as American equivalents. These UK biotechnology companies are the champions of biopharmaceuticals here. However, the start-ups face a particular development and manufacturing problem. Their venture funders are loath to pay for development facilities at an early stage because the risk of a drug failing in clinical trials is large. If UK biotechnology companies license out development it tends to be abroad to those countries, especially the USA, which embraced biopharmaceutical manufacture early. Plainly there is a danger that in this way the UK will lose perhaps 80% of the development value and fail to gain new manufacturing jobs: the US has seen a twenty-fold increase in biotechnology manufacturing jobs in the last 10 years. In the UK it may not be possible in all cases to retain manufacture but the minimum goal should be to take development to the most advanced stage possible to gain maximum value and to do this faster than others, so that the UK gains a reputation for adding value to ideas. The engineering research described earlier will help here.
The excitement of current scientific discoveries is enormous and very visible. However, genome valley will only become a UK reality if all the stages to the practical end-points are underpinned. It will be particularly important for The Royal Academy of Engineering to further extend its interface with the life sciences. The pioneering role it has taken in biomedical engineering has been immensely important in giving the area a greater national profile. While recognising current public concerns, life sciences research will nevertheless be one of the biggest influences on human progress in the new millennium. The capacity of engineers to quantify, to formally assess risk and benefit, to be accepted communicators and to lead the translation to safe products and linked services will be crucial in ensuring that we do create a UK genome valley.
I am grateful to my colleagues Parviz Ayazi-Shamlou, Mike Hoare and Nigel Titchener-Hooker for information and discussion.