Synthetic Artemisinin on Trial
Lessons for SynBio Innovation, Communication, and Outreach
Malaria is a terrible but entirely preventable blight upon humanity. In their most recent World Malaria Report, WHO revealed that 2015 saw 212 million cases of the disease, resulting in 429,000 deaths. Over two thirds of those who lost their lives were children. Throughout the African continent, millions lack access to simple tools that could prevent and treat malaria: mosquito bed nets, insecticides, rapid diagnostic tests, and antimalarial drugs. WHO recommends artemisinin-based combination therapies (ACTs) as ‘the most effective antimalarial medicines available today’. Artemisinin, the key ingredient of ACTs, is naturally produced by Artemisia annua – the sweet wormwood plant – but historically this has proven an erratic source. In 2006, a research group led by Jay Keasling at UC Berkeley published a landmark Nature paper in which they reported ‘the engineering of Saccharomyces cerevisiae to produce high titres of artemisinic acid’, which is easily converted to artemisinin by a chemical process. This work is still held up by many today as a watershed moment for synthetic biology; Keasling hoped his remarkable strain of baker’s yeast would prove to be a cost effective, environmentally friendly, and reliable source of a key therapeutic, saving countless lives. The yeast was optimised by a spin-out company supported by a hefty Gates Foundation grant, and industrial scale production was realised in 2013. Events took an unexpected turn last year when Nature reported that no synthetic artemisinin at all was produced in 2015, and the company was selling its manufacturing capabilities. What on earth went wrong?
Five months after Keasling’s key paper, a significant but very different essay was published in American Quarterly. In Technology and Its Discontents Joel Dinerstein argues that ‘technology [has become] the American theology’: rather than looking for social, political, or economic solutions to systemic societal problems, we fall back on ‘the rhetoric of technological utopianism’. Dinerstein quotes Edward Bellamy’s 1888 science fiction novel Looking Backward to describe this phenomenon as a ‘craze for more and more and ever greater and wider inventions for economic purposes, coupled with apparent complete indifference as to whether mankind derived any ultimate benefit from them’. Crucially, Dinerstein identifies this as a ‘cultural disease’ of Euro-American society, and searches for reasons for the ‘unrealistic expectations’ we pin on technology. His central thesis is that technological advances help to maintain the myths of progress and white Western superiority; his conclusions are that we need to examine what the purpose of any ‘progress’ is, even if scientists ‘claim such questions are not within their purview’. If this all sounds a bit abstract and high-minded, note that several key players in the artemisinin story use arguments very similar to Dinerstein’s – in fact, the main disagreements in this debate arise from different interpretations of technology’s role and value, and an attitude akin to that identified by Bellamy may be the most important reason for the demise of synthetic artemisinin.
In 2000, Keasling was working on an important group of chemicals called isoprenoids, which includes ‘a huge number of valuable products’. In a Discover Magazine interview, he reflects on his initial approach: ‘we started with a focus of building up the basic pathways needed to make those chemicals, and we thought, what are we going to use this pathway for?’ A profile of Keasling in The New Yorker features some troubling quotes from its subject which paint the picture of a researcher choosing a topic he doesn’t fully understand as a vehicle for his isoprenoids project. He had ‘not a clue’ what artemisinin was, and ‘wasn’t an expert on infectious diseases’ – rather, he was ‘looking for a chemical compound that could demonstrate the utility of [his] biological tools’. It may seem a pedantic criticism, but these are not the best reasons to wade into a sociopolitical area as convoluted as malaria aid. In 2003, Keasling’s group published a significant paper detailing the engineering of a metabolic pathway in E. coli for production of amorphadiene – the first committed step in artemisinin biosynthesis. Jim Thomas of the ETC Group technology watchdog claims the first author on this paper told him that artemisinin was simply a convenient ‘molecule that could attract funding to their isoprenoid platform’ with ‘a compelling story, a funder with deep pockets, and a great PR angle’. Before you dismiss Thomas as a science-hating conspiracy theorist, the 2003 paper itself gives credence to his argument: ‘Because [these molecules] are the universal precursors to all isoprenoids, the strains developed in this study can serve as platform hosts for the production of any terpenoid compound’. Keasling’s spinout company Amyris was originally founded to refine his artemisinin-producing strain, but quickly refocused on higher value isoprenoids afterwards. The motives for synthetic artemisinin development seem especially unconvincing when we examine the controversy it caused.
At first, synthetic artemisinin received high praise from all quarters; like the ebullient ravings of a drunken reveller, it all seems a bit cringeworthy in the cold light of sober morn. Discover Magazine anointed Keasling Scientist of the Year 2006, rating the impact of his synthetic drug higher than the reclassification of Pluto as a dwarf planet, pivotal climate change research, and the first demonstration of epigenetic inheritance in mammals. The author of this piece wrote: “It is not just a technical tour de force but a humanitarian one. Keasling’s microbes will churn out the drug for a fraction of its current cost, making it accessible to much more of the world. Properly harnessed, these microbes could save millions of lives.” Berkeley News, the media relations website for UC Berkeley, called Keasling’s work ‘the first triumph of the nascent field of synthetic biology’, and promised that it ‘will be a lifesaver for hundreds of millions of people in developing countries’. Dr Robert Sebbag, a Sanofi board member, called synthetic artemisinin ‘a pivotal milestone in the fight against malaria’. In 2013, as controversy about this source of artemisinin was reaching its apex, the US Patent and Trademark Office awarded UC Berkeley its Patents for Humanity Award for ‘developing research and license agreements to provide a lower-cost, more reliable way to produce anti-malarial compounds’. Rob Carlson, Managing Director of Bioeconomy Capital and a prominent synthetic biologist, claimed that there was ‘no question that people who suffer from malaria will be better off with artemisinin produced in yeast by Sanofi’. Keasling published another Nature paper presenting synthetic artemisinin as a ‘success story’ and ‘a model for the use of synthetic biology in pharmaceutical development’. After all these long-ass verses, I’m tired, you tired: Jesus wept. The point that these many quotes have hopefully made is that there was an incredible degree of hype around Keasling’s platform from the very start; bold – even irresponsible – assertions were made about a technology in an embryonic stage of development. Synthetic artemisinin, and by extension synthetic biology, had caught the public imagination.
This is not to say that Keasling’s work is unimpressive: it is undeniably brilliant from a scientific perspective. One should not be fooled by the apparent simplicity of the three-step process the authors present in their 2006 paper – metabolic engineering is still notoriously frustrating, and each of these steps is difficult, time-consuming, and could probably be published on its own. A lot of work is packed into these four pages: differential regulation of multiple genes; genomic analysis of A. annua; evolutionary comparison of plant genes to find the enzyme that converts amorphadiene into artemisinic acid; functional characterisation of this enzyme and its partner; introduction of these genes into yeast; optimisation of the resultant strain. Essentially, this is really good science. However, at the end of the piece, the authors claim their process could produce artemisinin ‘significantly below current prices’ in a manner unaffected by ‘political climates’. The supplementary information is supposed to bolster these arguments, but contains only a short paragraph suggestive of back-of-the-envelope calculations without working shown. These hopes were to meet some pitiless economic realities.
We Are the Dollars & Cents
Soon after WHO first recommended ACTs as the gold standard of malaria treatment, it seemed synthetic artemisinin could not come soon enough. In December 2004, WHO announced that it would be able to deliver ‘only half of the 60 million [ACT] doses anticipated in 2005’ due to the shortage of artemisinin. A. annua farms in China and Vietnam had failed to supply enough of the drug due to uncertainties over the market: initially, the switch to ACTs in Africa stalled due to the high cost of a treatment course of these new drugs compared to older alternatives, and farmers saw no guarantees that they would be able to sell their crop. In response, governments encouraged these farmers to grow A. annua to meet ever-increasing demands, and tens of thousands of hectares were planted. By 2007, the price of artemisinin had fallen by over 80 percent as supply far outstripped demand, making A. annua an unprofitable crop for farmers, and putting 80 processing companies out of business. This boom-bust cycle would be seen again over the next six years, with the price of artemisinin vacillating wildly between $1100 and $200 per kilogram; a Nature analyst describes the mechanism succinctly: ‘shortages of A. annua send prices soaring, which attracts more farmers to plant it; their produce then swamps the market, depressing prices and triggering fresh shortages’.
Several efforts to stabilise the artemisinin market were initiated to combat these ‘rollercoaster’ dynamics. The Global Fund to Fight AIDS, Tuberculosis and Malaria launched the $343 million Affordable Medicines Facility for malaria (AMFm) in 2010 as a two year trial. The rationale behind AMFm was based on a study by the US Institute of Medicine that argued for a global subsidy for ACTs on the grounds that this would increase the availability and affordability of these drugs. Public health clinics were providing ACTs at just a dollar per dose, yet the majority of patients still chose to buy their antimalarials from local markets and private pharmacies at an inflated price. Instead of purchasing effective ACTs, patients were turning to cheap older drugs or artemisinin monotherapies (AMTs) – the former are useless, the latter dangerous because they can lead to the development of artemisinin resistance in the malarial parasite. The solution proffered by AMFm was to heavily subsidise ACTs, focusing in particular on the private sector so that even local businesses would sell these drugs at well under a dollar per dose. An alternative approach was championed by the UNITAID-supported Assured Artemisinin Supply System (A2S2): providing loans to farmers and artemisinin extraction companies, as well as brokering long term contracts between these parties and pharmaceutical companies. After A. annua is planted, eighteen months are required to produce artemisinin, creating a lag between demand and supply and planting and selling, which was at least partially responsible for the boom-bust cycle. Farmers would not be eager to plant A. annua when artemisinin prices were low, which could result in dangerous shortages – therefore, A2S2 aimed to incentivise producers with advance loans to ensure profitability and smooth market fluctuations. Both of these approaches may seem legitimate on paper, but they would have profoundly different levels of success.
AMFm raised concerns early on, as it was predicted to cause a spike in artemisinin demand, potentially leading to shortages. As its pilot phase was ending in 2012, Oxfam published an influential report titled Salt, Sugar and Malaria Pills which vociferously attacked the initiative as ‘a dangerous distraction from effective public health measures’. Dr Kamal-Yanni, Oxfam’s Senior Health Policy Advisor, highlighted a number of critical problems with AMFm. Evidence shows that paying for healthcare leads to patients delaying their pursuit of treatment or even going without it; even the greatly reduced cost of ACTs was too much for impoverished malaria sufferers who could not afford to pay for a full course of the drugs. Distribution of antimalarials by ‘unqualified shopkeepers and hawkers’ is far from an ideal situation – these vendors have no medical training, cannot diagnose malaria, and are motivated by profit rather than clinical needs – and not one that funding bodies should support. Proper diagnosis of malarial fevers is vital because the majority of sub-Saharan African fevers are not due to this disease: inappropriately treating these non-malarial cases with ACTs risks both the patient’s health and development of artemisinin resistance. The massively increased sales of ACTs triggered by AMFm did not correspond to increased malaria treatment, and in fact the organisation ordered five times as many treatments as the estimated number of malaria cases in its first year of operation. Finally, there was no evidence that AMFm reached the most vulnerable or reduced use of AMTs. Oxfam contended that alternative community-based approaches were delivering far more robust results and were worthier of donor investment. In Ethiopia, community health workers were trained and deployed with bed nets, rapid diagnostic tests, and antimalarials, and malaria mortality was halved in just three years. Oxfam recommended an extension of such schemes, with a focus on reaching children, pregnant women, the very poor, and inhabitants of remote areas, and called for an end to AMFm funding. In 2013, the demise of AMFm was complete.
In stark contrast, A2S2 was extremely effective: by 2016 the scheme was credited as one of the ‘main causes of [artemisinin] price stability’. By setting up contracts and communication channels between the groups involved in ACT production, A2S2 was able to encourage cooperation to ensure a stable artemisinin supply, so that by 2013 agricultural sources were meeting demands. Furthermore, WHO’s advocacy of malaria diagnostics and monitoring, and the end of AMFm, led to a drop in excessive artemisinin demand. Brilliant: problem solved! But where does synthetic artemisinin slot into these market machinations? The justifications for this source have evolved with the economic landscape to such an extent that one begins to wonder if Keasling & Co. are simply determined to shoehorn in synthetic artemisinin for any reason, even if it isn’t the best solution for malaria sufferers. When artemisinin was scarce and expensive, the yeast platform was presented as a significantly cheaper route to the drug. Then the price of botanical artemisinin fell. In 2004, the Gates Foundation set a goal of reducing the cost of each ACT treatment from $2.40 to below a dollar; a year before synthetic artemisinin reached market, the median price was already 92¢. From 2013 to 2015, the price of A. annua-derived artemisinin settled at less than $250 per kilogram. Despite an incredible $64 million of Gates Foundation funding, royalty-free licenses allowing Sanofi to use Amyris technology, and sale on a ‘no-profit-no-loss basis’, Sanofi could only drive the cost of synthetic artemisinin down to $350-400 per kilogram. ‘We have learned there is a lot more to it than cost’ admitted Jack Newman, Amyris co-founder.
The grounds shifted: now synthetic artemisinin represented a reliable and steady source that could ‘snap the drug out of its rollercoaster supply cycle’: as Newman put it, if demand were to suddenly rise, ‘you just fire up another fermenter’. Sanofi promised to initially use its production facilities only to ‘smooth out market fluctuations’, although the ETC Group pointed out the improbability of a pharmaceutical company halting production – and profits – during periods when botanical production met demand. Unfortunately (for certain interested parties) once the AMFm trial ended and other measures took effect, the boom-bust cycle stabilised, and demand fell. Malcolm Cutler of A2S2 even argued that Sanofi risked disrupting the artemisinin market further: ‘it could create huge shortages because people will stop producing the natural stuff’. Finally, Keasling claimed that if Sanofi was the only artemisinin source, sale to producers of AMTs could be stamped out, slowing the development of resistance; this ignored the efforts of governments to ban AMT imports, WHO’s restriction on AMT sale by drug companies, and A2S2’s scheme to delist extractors selling to AMT manufacturers. Simply put: there is no real need for synthetic artemisinin.
Whitey on the Moon
While the hype machine rolled on, dissenting voices emerged, growing in number and volume. Early in 2013, Sanofi launched large scale artemisinin production using an improved yeast strain that could generate 250-fold higher titres. At an A2S2-organised conference, the company announced that they had produced 39 tonnes of artemisinin, and planned to produce 60 tonnes the following year, generating alarm amongst A. annua growers as this represented over a third of the drug’s market. Later in the year, Jay Keasling popped by a Future of Nature conference in Cambridge to declare that ‘moves are afoot to replace the entire world supply’ – while ‘early on, it was not about replacing the agricultural form… now I think it’s nearly inevitable that it will shift over’. Two figures with antithetical views were in attendance: Jim Thomas of the ETC Group, an NGO generally critical of technology and synthetic biology, and Rob Carlson of Bioeconomy Capital, which funds biotech startups. Thomas was galvanised into writing a scathing Guardian piece calling Sanofi’s moves ‘a high tech assault on farmers’. He argues that A. annua farming was developed as ‘doubly beneficial’, providing a vital drug and a source of income for poor farmers. Synthetic artemisinin was not necessary to fight malaria – A. annua was already meeting needs – but simply allowed pharmaceutical companies to avoid the complexities of sourcing from many small farms, increasing their revenues. Thomas’ article has some killer lines – he compares Keasling’s suggestion that farmers switch to growing foodstuffs to Antoinette’s insouciance towards her subjects with the neat phrase ‘let them plant potatoes’ – but one of its strongest passages is a quote from Professor Hans Herren. Herren is a highly regarded expert in the field: he is President and CEO of the Millennium Institute, a World Food Prize winner, lived and worked in Africa for 27 years engaging in agriculture, health and environment research, has worked extensively with A. annua farmers, and – ironically enough – studied as a postdoc for two years in UC Berkeley. Whatever your opinion on organisations like the ETC Group, Herren’s viewpoint should carry some weight, which makes the vitriol of his quote extraordinary: ‘There is simply no rationale to have a synthetic product on the market when farmers could produce enough raw material to produce the tablets from pulverised high quality plants. Jay Keasling’s words are denigrating for farmers – he should be ashamed of his work and words. It adds nothing to solve the malaria problem at all.’ Woof.
Rob Carlson did not enjoy Thomas’ piece, dismissing it as ‘a broadside’ full of ‘glib accusations’. He joined the time-honoured scientific tradition of condescending from an unassailable position of assumed expertise: ‘Mr Thomas asserts that farmers know best, but he never himself descends to the level of looking at actual numbers’. Carlson then carries out his own analysis, relying mainly on two graphs – the first of these is one Keasling presented at the Cambridge conference. This data was taken from a Boston Consulting Group report, and appears to show that farmers in China and Vietnam would stand to make ‘substantially higher revenues planting crops other than artemisia’ – from this Carlson argues that Keasling’s ‘let them plant potatoes’ was actually sound advice. Carlson then moves on to a chart from a 2013 Nature article examining the boom-bust dynamics of artemisinin prices, and argues that ‘the farmers most at risk of contracting malaria only benefit economically when there is a shortage of artemisinin’. The data doesn’t really back up this claim, as Carlson has extrapolated from early trends to suppose that A. annua farming could only ever be profitable when demand was dangerously high – furthermore, A2S2 was already seeing success in stabilising these market swings. Carlson then goes on to make the bizarre claim that this ‘deadly paradox’ benefits ‘the artemisinin pricing cartel’, a group that controls artemisinin production and sets its price. The Nature article doesn’t mention the existence of such a group, and Google provides no answers here – Thomas later guessed that Carlson had confused the A2S2 with this nefarious cartel. He wraps up by asserting: ‘once you look at the numbers there is no argument… that we should do anything but brew artemisinin in vats’.
The blog warfare raged on: Thomas posted a decisive response to Carlson a month later in which he made several strong points which I have surreptitiously peppered throughout the above paragraphs. He highlights how the justifications for synthetic artemisinin have changed, debunks the ‘pricing cartel’, shows that synthetic artemisinin is no cheaper than botanical, argues that Sanofi’s monopoly over production wouldn’t eliminate AMTs, and raises concerns that Keasling’s statements about not wanting to drive farmers out of business ‘until we have enough installed capacity to take over the entire world supply’ are irresponsible and could lead to artemisinin shortages. Thomas exposes Keasling’s use of the Boston Consulting Group graph as misleading – the data was from a report showing how A. annua farmers need to supplement their incomes, and wasn’t meant to highlight alternatives to this crop. Thomas accuses Keasling of being ‘dismissive of farmer knowledge’: farmers understand markets and already grow foodstuffs as break crops, but need A. annua as a cash crop to keep themselves out of poverty. Thomas’ riposte seems rather definitive and – disappointingly but perhaps obviously given the state of my Twitter profile – Rob Carlson did not reply to my tweets asking if he had ever responded.
Thomas’ ETC blog post also raises some more general questions for synthetic biology. He contends that many of the problems we expect synthetic biology to address ‘emerged from the large scale and rushed application of a technology that wasn’t properly understood at the time’: global warming resulting from use of fossil fuels, biodiversity loss due to synthetic chemistry, and antibiotic resistance due to misuse of these vital drugs. There is a ‘need for real precaution and humility in the roll out of new technologies’ instead of ‘jumping to speculative claims about what a technology might do without equal enthusiasm for finding out the less desirable effects’. Thomas uses an analogy many will remember from childhood: the book There Was an Old Lady Who Swallowed a Fly imagines a woman who consumes an ever larger and more bizarre series of creatures to solve an initially small problem, until she eventually dies. Do we take a similar approach when it comes to synthetic biology? Do we take the simplistic attitude of: ‘It’s not our fault! It’s outmoded technology’s fault!’ The problem of scarce and expensive artemisinin, identified way back in 2004, was solved by simply planting more A. annua; later, the unstable market was steadied by coordinating different players in artemisinin production and providing loans; Oxfam recommends non-technological solutions to malaria involving community outreach, and has seen success so far.
Keasling clashed publicly with Thomas over their seemingly irreconcilable views on technology: ‘Keasling… is baffled by opposition to what should soon become the world’s most reliable source of cheap artemisinin. “Just for a moment, imagine that we replaced artemisinin with a cancer drug,” he said, “and let’s have the entire Western world rely on some farmers in China and Africa who may or may not plant their crop. And let’s have a lot of American children die because of that. Look at the world and tell me we shouldn’t be doing this. It’s not people in Africa who see malaria who say, “Whoa, let’s put the brakes on.”’ Cast your mind back to Dinerstein’s critique of technological advances: there is a fundamental difference here between how different groups view technology. Is it our saviour, or is it unnecessary and dangerous? Remember too Dinerstein’s claim that technology maintains ‘the myth of white Western superiority’: synthetic artemisinin was supposed to be ‘a lifesaver for hundreds of millions of people in developing countries’, but its creators dismissed the fears of farmers at risk of malaria who depend on A. annua for their livelihoods. Is there a note of condescension here? Were Keasling et al. determined that their product succeed for their own reputations (or that of synthetic biology) regardless of demonstrable benefits for malaria sufferers? A Gil Scott-Heron poem comes to mind: ‘I can’t pay no doctor bill, but Whitey’s on the moon’.
Dr. Claire Marris, a food policy researcher, has been outspoken about synthetic artemisinin, and sees it as a damning example of the poor outreach practises ingrained in the area. Synthetic biology ‘has been portrayed as an emerging field with great potential’ and high ‘sociotechnical expectations’, but those in support of it see ‘public perceptions and fears’ as a threat to its success. Therefore, the public are excluded from conversations about the social consequences and regulation of the field: boundaries are drawn ‘between the conduct of research and innovation… and downstream social implications… implying that the role of ethics, social scientists, and public engagement is relevant only to the latter’. This problem was evident in the development of synthetic artemisinin: groups such as Oxfam, with the expertise to carry out a nuanced assessment of the best ways to tackle malaria, were excluded from the conversation, while concerned NGOs like the ETC Group were dismissed as ‘baffling’ and ‘less apt to offer actual analysis’. Marris echoes Dinerstein’s conclusion, saying that ‘if [scientists] do not want to take responsibility for the real-world complexities of what happens beyond the confines of the laboratory, they should take care not to make simplistic claims’ about the outcomes of their projects. Finally, she predicts that unrealistic promises about the benefits of synthetic artemisinin will backfire and ‘tarnish the image of synthetic biology’. We live in an ‘audit society’ which demands transparency, ‘accountability and control’. There is a mounting crisis in the authority of science due to high profile failures such as the management of foot-and-mouth disease and the aggressive introduction of GM foods in the UK. The 2016 Word of the Year was ‘post-truth’ – we have ‘had enough of experts’. Now is the time for scientists to regain public trust, to learn to listen, to try to be just a little less arrogant and admit that their expertise extends only so far. While synthetic artemisinin failed to help malaria sufferers, we can take vital lessons from its story.