Pharming, recombinant medicines and synthetic biology

Medicines that are proteins

Large or extensively modified therapeutic proteins are usually produced in living cells because correct folding and post-translational processing are difficult to reproduce by chemical synthesis. The required processing determines whether bacterial, yeast or mammalian cells are suitable. Some short peptides are synthesised chemically, and cell-free expression systems exist, so cellular production is a practical requirement for large glycoproteins rather than an absolute rule for all protein drugs.

Before recombinant DNA technology, protein medicines were extracted from an organism that already made them, and the two named examples on the specification show the consequences of that.

Factor VIII is the clotting factor missing in the commonest severe haemophilia. For decades it was concentrated from donated blood plasma, pooled from thousands of donors per batch, and in the early 1980s that pooling proved catastrophic: plasma from a single infected donor contaminated whole batches, and a large fraction of the people treated for haemophilia in that era were infected with HIV, hepatitis C or both. Recombinant factor VIII, made since the 1990s by cultured mammalian cells carrying the human gene, removed the donor pool, and with it that route of infection, entirely.

Adenosine deaminase (ADA) is the enzyme whose absence causes one form of severe combined immunodeficiency, the condition you met in the gene therapy lesson. Before gene therapy, and still alongside it, patients are treated with injections of the enzyme itself, produced recombinantly and chemically stabilised so it survives in the blood. The same condition therefore has two recombinant treatments: supplying the protein, or supplying the gene.

Mammalian cells are used in preference to a cheaper bacterial fermenter because E. coli transcribes and translates a human gene without carrying out the subsequent processing. Factor VIII requires chaperoned folding, disulfide bridges at specific positions, and sugar chains added in a Golgi apparatus; a bacterium has none of those, so it yields a polypeptide with the correct primary structure and without the required folding or glycosylation. Insulin is small and requires little post-translational modification, so it can be produced in bacteria. The required processing determines the expression host.

Recombinant medicine

A protein drug produced by cells carrying an introduced copy of the human gene, rather than extracted from blood or tissue.

Factor VIII

The clotting protein deficient in haemophilia A, produced recombinantly in mammalian cell culture since the 1990s.

Adenosine deaminase

The enzyme missing in ADA-deficient severe combined immunodeficiency, supplied to patients as a recombinant protein.

Pharming: expression in milk

Pharming uses genetically modified organisms, including plants and animals, to produce pharmaceutical substances. A tissue-specific promoter can direct expression in milk, but yield depends on the construct and species.

Mammalian cell culture is effective and expensive: the cells grow slowly, require a rich sterile medium, and typically yield milligrams per litre. A lactating transgenic animal secretes the protein into milk, which is collected by milking and is a simpler starting material to purify than cell culture medium or plasma. Reported yields vary widely between constructs and species, from milligrams to several grams per litre, so quote a figure with its source rather than a general rate.

The construct is the part to describe precisely. The human gene is joined downstream of the promoter of a milk protein gene, such as β-lactoglobulin, so the transcription factors that switch that promoter on exist only in mammary gland cells during lactation. The construct is microinjected into the nucleus of a fertilised egg, or introduced into cultured cells from which an embryo is then cloned by nuclear transfer, and the embryo is implanted into a surrogate. Microinjection can produce mosaic founders, in which only some cells carry the insert, so animals must be screened and a confirmed line established before production begins. A female of that established line carries the gene in every cell and expresses it in one tissue: her milk carries the human protein, her blood and other tissues carry very little, and purification starts from milk.

Antithrombin is the licensed example: an anticoagulant protein produced in the milk of a herd of transgenic goats and approved as a drug in Europe in 2006, the first medicine from a transgenic animal to be licensed. A herd of goats can supply quantities that would otherwise require very large numbers of blood donations.

Pharming

The use of genetically modified organisms, including plants and animals, to produce pharmaceutical substances; in livestock, classically in milk.

Mammary-specific promoter

The promoter of a milk protein gene, placed in front of an introduced gene so that the gene is transcribed only in mammary gland cells during lactation.

Welfare and regulatory considerations

The arguments about pharming are set out below on each side, with the evidence for each.

Welfare, against. Making a founder animal is inefficient: microinjection succeeds in a small percentage of eggs, so producing one transgenic goat means many surrogate pregnancies, failed implantations and discarded embryos, and where cloning is used it brings its own losses, since a large fraction of cloned pregnancies fail late or produce unhealthy offspring. Critics add a deeper objection: that redesigning an animal's genome to serve as production equipment treats a sentient creature as an instrument, whatever its day-to-day conditions.

Welfare, for. Once the founder exists, the producing herd is bred normally and kept as a dairy herd; milking does not harm the animal, and the commercial requirement for clean, consistent milk means the animals are housed and fed to a high standard. Supporters also compare the alternatives: the protein would otherwise come from pooled human plasma, with the safety record described above, or from cell culture at a cost some health systems will not meet.

Regulation. A drug from a living animal poses questions a steel fermenter does not. Milk composition varies with the animal, the season and the feed, so every batch must be purified and tested to prove it identical to the last; the herd must be certified free of infections, and of prion disease in particular, that could pass into the product; and the animals themselves must be contained, because a transgenic goat that bred into an ordinary herd would spread a pharmaceutical gene that could never be recalled. Regulators answered with closed, certified herds, animals identified and tracked individually, and approval processes that took years. The animals are also patented, which is one instance of the ownership question treated below.

Synthetic biology: parts, genomes and pathways

Recombinant DNA technology introduces one gene at a time into an organism that is otherwise unchanged. Synthetic biology assembles characterised genetic components and pathways for defined functions, treating DNA as an engineering material that is specified, synthesised to order, and assembled into circuits and pathways not found in any organism.

It has three levels. Standard parts: promoters, ribosome binding sites, genes and terminators catalogued with defined, measured behaviour, designed so that any part joins to any other, the way electronic components share standard connections. An engineer picks parts from the registry and assembles a device instead of hunting each sequence from scratch. Genome synthesis: whole genomes built chemically from their sequence data. A bacterial genome of about a million base pairs was synthesised and installed in a recipient cell in 2010, and the cell ran on it; stripped-down versions since have asked what the minimum set of genes for life actually is. Engineered pathways: several genes, from several organisms, assembled so that a cell runs a metabolic production line it never had.

The named pathway example is artemisinin, the antimalarial used in combination therapy worldwide. It is extracted from sweet wormwood, a crop whose yield and price varied substantially between years. A pathway of enzymes, some from the plant and some derived from yeast's own metabolism, was assembled in brewer's yeast, so that the yeast converts sugar into artemisinic acid, which industrial chemical steps then convert into artemisinin. Production began at scale in 2013, providing a supply that does not depend solely on the harvest.

Synthetic biology

The design and construction of new biological parts, devices and pathways, and the redesign of existing organisms, treating DNA as a specifiable engineering material.

Standard part

A DNA sequence with a defined function, catalogued and designed to be assembled with any other standard part.

Biosecurity and ownership

The same capability, DNA synthesised to order, underlies both the applications and the risks of synthetic biology. The specifications require the arguments on each side, which are paired below.

Biosecurity. Poliovirus was assembled from mail-ordered DNA fragments and published in 2002, and an extinct relative of smallpox was synthesised in 2017 for around one hundred thousand dollars, so the barrier to recreating a dangerous pathogen is money and published sequence, not rare skill. Against that: the same openness lets any public-health laboratory synthesise a new pathogen's genes within days of its sequence appearing, which is how vaccine development now starts, and the response to the risk has been screening rather than secrecy, with synthesis companies checking orders against databases of dangerous sequences and flagging the customers who place them. Screening is voluntary and coverage differs between suppliers, and whether voluntary screening is sufficient is the point at issue. Date any claim about screening practice or regulation, since both change.

Ownership. Engineered organisms and standard parts can be patented, and the case for is the usual case for patents: development costs run to hundreds of millions, and no company recovers that without a period of exclusivity, so patents are what pay for the goats, the yeast and the trials. The case against has two layers. One is access: a patented medicine is priced by its owner, and the diseases synthetic biology could treat most cheaply are concentrated where patients are poorest. The other is displacement, and artemisinin is the worked case: fermenter production stabilises the drug supply and also competes with the smallholder farmers who grow wormwood, so the same process reduces supply risk for health systems and reduces income for those growers. A boundary question is also unresolved in law: patenting one gene is established practice, and whether the design of an entire organism can be owned is not settled.