Rice, milk proteins and sick children in Lima: the Ventria trial revisited

The drink was simple: rice powder, salts and water, the kind of oral rehydration solution clinicians have used for decades to keep children with diarrhoea alive. What made this version different was two proteins stirred into it. Both were human proteins, normally found in breast milk. Both had been made in rice plants carrying synthetic copies of human genes, grown in fields in the United States.

In 2006 the Californian company Ventria Bioscience announced results from a trial of that drink in young children hospitalised with diarrhoea in Peru. Within weeks the study had become a byword, among critics of agricultural biotechnology, for something they had long predicted: drug-making crops grown in the open air, tested on poor children in a country with less regulatory scrutiny than the company's own.

Eighteen years on, it is worth looking again at what the trial was, which objections hold up, and what happened to the idea of turning crops into pharmaceutical factories.

The product

Ventria's platform, which it called ExpressTec, used rice grain as a protein factory. Genes for the target proteins, rewritten to suit the plant's preferred codons, were placed under the control of rice seed promoters so the proteins would accumulate in the grain. After harvest the grain was milled and the proteins extracted.

The two proteins in the Peru trial were lactoferrin and lysozyme. Lactoferrin binds iron, which starves some bacteria of a nutrient they need, and has broader antimicrobial and immune-related effects. Lysozyme is an enzyme that breaks down the cell walls of certain bacteria. Both are abundant in human milk, and the reasoning was straightforward: breastfed infants get fewer gut infections, these proteins are part of why, so adding them to a rehydration drink might shorten an episode of diarrhoea.

Rice was chosen partly because it is self-pollinating, which limits, though does not eliminate, the spread of pollen to neighbouring fields.

The trial

The study enrolled 140 young children, all under three, with acute diarrhoea and dehydration. It was run with Peruvian paediatric and nutrition research institutions, the ones named in the original complaints being in Lima. It had three arms:

• the standard glucose-based oral rehydration solution recommended by the World Health Organization;
• a rice-based solution without added proteins;
• the same rice-based solution with recombinant lactoferrin and lysozyme.

The children in the protein arm recovered faster. The headline figure was an average duration of diarrhoea of about 3.7 days, compared with about 5.2 days in the glucose solution group. The results were first presented as a conference abstract in 2006, and a full paper appeared in the Journal of Pediatric Gastroenterology and Nutrition in 2007.

The objections, one by one

Moving the trial south

The sharpest charge was regulatory arbitrage: that a US company tested an unapproved product made in a GM crop on children abroad because it would have been harder to do at home. The defence was that diarrhoea kills children in poor countries, not in California, and that a product aimed at that problem should be tested where the problem is.

Both can be true. Research ethics guidance, notably the CIOMS international guidelines, asks that research in low-income settings respond to local health needs and that successful products be reasonably available to the population that took the risk. Ventria's own business plan pointed elsewhere. The proteins were also pitched for "medical foods", supplements and nutrition products in rich markets. A trial in Lima that mainly served to build a safety and efficacy record for a supplement sold in Los Angeles would fail that test. A trial that led to a cheap rehydration product for Peruvian clinics would pass it. As far as I can tell, no such product ever reached Peruvian clinics.

The comparator

The second complaint, made by an American paediatrician and others at the time, was about design. Rice-based rehydration solutions were already known to perform at least as well as glucose solutions in some kinds of diarrhoea. Comparing the protein drink with glucose solution, critics argued, flattered the result.

There is something in this, and something missing. This was not a placebo trial in the usual sense: every child received rehydration, which is the treatment that actually saves lives. The trial also had the right internal control, the plain rice solution. The question a careful reader should ask is not "protein drink versus glucose" but "rice solution with proteins versus rice solution without", because only that comparison isolates the effect of the proteins. A headline built on the larger gap tells you more about marketing than about lactoferrin.

Consent

In Peru the public learned of the trial only after the results were out, when a human-rights association and a regional network of campaigners against GM crops denounced it. Whether parents of dehydrated children understood that the proteins came from genetically engineered rice, and that the product was unapproved anywhere, is hard to judge from outside. It is a fair question, and a familiar one. Consent obtained from frightened parents at a hospital bedside is consent under pressure, however carefully the form is worded. That is a reason for extra care in trial design, not a reason to stop all paediatric research where children are sick.

Not quite the same molecule

Critics also argued that the recombinant proteins were not identical to the natural ones. On one point the biochemistry backs them. Lactoferrin carries sugar chains, and plants attach sugars in their own style, including some sugar residues that human cells do not make and that the immune system can recognise. Whether that matters for a protein swallowed and digested, rather than injected, is a separate question, and the trial was too small to answer it.

Drugs in the food crop

The objection with the longest life was not about the trial at all. It was about where the rice grew. Ventria's plans to plant in California ran into opposition from the state's rice industry in 2004. In 2005 the company tried Missouri, until Anheuser-Busch, one of the largest buyers of rice in the country, threatened to stop buying Missouri rice if the plan went ahead. The company moved on, and in 2007 the US Department of Agriculture approved plantings of up to 3,200 acres in Kansas.

The fear was mixing. In 2002 a company called ProdiGene was fined after maize it had engineered to produce a pharmaceutical protein, grown in a Nebraska field the season before, came up again as volunteer plants among the next crop of soybeans. Some of it ended up in about half a million bushels of stored soybeans, which had to be bought up and destroyed. In 2006 an unapproved herbicide-tolerant rice line from Bayer was found in the US long-grain supply, a contamination that cost the company a settlement of $750 million with farmers in 2011. Neither case involved Ventria. Both showed that crops escape their paperwork.

What became of plant-made pharmaceuticals

The industry that the Peru trial seemed to herald did arrive, but not in the form critics feared or promoters promised.

• In 2006 the USDA approved a poultry vaccine against Newcastle disease made in plant cell culture. It was never marketed, but it was the first regulatory approval of its kind.
• In 2012 the US Food and Drug Administration approved taliglucerase alfa, an enzyme for Gaucher disease made in carrot cells grown in bioreactors.
• In 2014 ZMapp, an antibody cocktail made in a tobacco relative, was given to Ebola patients under emergency rules. In a randomised trial in 2019 it proved less effective than two competing antibody treatments.
• In 2022 Health Canada approved a plant-made COVID-19 vaccine from the company Medicago. Its owner closed the company the following year.
• In May 2023 the FDA approved pegunigalsidase alfa for Fabry disease, made in tobacco plant cells.

Note the pattern. The approvals came from plant cells in steel tanks, or plants grown indoors, not food crops in open fields. The pharmaceutical industry wanted what regulators wanted: contained, consistent production. The open-field pharma crop, the thing that worried rice farmers in Missouri, has largely faded from the agenda.

Ventria itself adapted. Its rice-made human proteins found a steady market as animal-free supplements for cell culture, sold through a division called InVitria, and in 2023 the remaining business took that name. The proteins that were once meant for sick children are now more often found in laboratory growth media.

Milk proteins without the mother

The idea of making breast-milk proteins outside a human body has not gone away. It has moved from plants to microbes. Since the mid-2010s, human milk oligosaccharides, sugars that are abundant in breast milk, have been made by fermentation with engineered bacteria and added to infant formula. Several start-ups now make lactoferrin by precision fermentation in yeast and other microbes, and at least one declared its product "generally recognised as safe" in late 2023 on the strength of its own expert panel, the US self-affirmation route that does not require FDA sign-off.

The shift is a direct result of the synthetic biology toolkit described in our piece on Synthia and synthetic cells: it is now cheaper to engineer a microbe to secrete a protein than to grow it in a field. Contained fermentation removes the gene-flow argument. It does not remove the others. Who gets tested first, what parents are told, and how regulators treat a recombinant version of a natural food component are the same questions Lima raised in 2006, and they belong in the governance debates covered in Lab life 3.0 as much as in Europe's arguments over labelling gene-edited plants.

The Peru trial is often remembered as a scandal or forgotten as a footnote. It is more useful as a case study. Some criticisms were stronger than others, and the bigger worry, pharmaceutical rice growing next to food rice, mostly resolved itself because the economics favoured steel tanks. The questions about who carries the risk of a trial, and who gets the benefit, have not resolved themselves at all.