It took ten years—from 1980 to 1990—to develop the necessary technology to introduce genes into rice. It took another nine years—from 1990 to 1999—to introduce the genes that reconstitute the pathway for provitamin A biosynthesis into the seed. And it took another five years—from 1999 to 2004—to develop Golden Rice. It is taking several more years to advance the first Golden Rice product through the regulatory approval process. Considering that Golden Rice could substantially reduce blindness (500,000 children per year) and deaths (2-3 million per year), the parsimony displayed by the responsible bodies after 20 years is hardly understandable.

During the last 20 years a vast knowledge base has been accumulated on the production and commercialisation of transgenic plants. While the Golden Rice Humanitarian Board understands that every new transgenic event must comply with regulations to guarantee the safety of the product, it has a hard time dealing with non-scientific arguments that unnecessarily delay the adoption of the technology vis-a-vis the human tragedy brought about by vitamin A deficiency. Countries where Golden Rice could provide health benefits should be provided with the opportunity and information to pursue their own independent decision-making process and should not be led by unfounded and self-serving external pressures.

The Humanitarian Board supports efforts to develop appropriate risk management strategies that include acceptable risk levels, ie where the benefits outnumber the potential danger by far, as is the case with Golden Rice. Reputed ecologists, including opponents of the technology, have so far concluded that Golden Rice poses no imaginable risk to the environment. All plants produce high amounts of carotenoids, thus their presence in the grain will not introduce any new substances into the environment nor will they provide any additional selection advantage.

An unbearable financial burden

What are the regulatory requirements standing in the way of Golden Rice deployment? First of all, the application should be for a carefully selected, regulatory clean transgenic event. Criteria are not necessarily based on scientific grounds; they include a number of requirements pertaining to the introduced genetic construct, eg the inserted DNA fragment should not have undergone multiple integrations or rearrangements, there should be no read-through across the construct borders or any residual ballast DNA. This in turn requires the production of many hundreds of transgenic events using the same DNA construct, from which the regulatory clean event is then selected. The makeup of the construct itself must have been conceived taking into account the requirements imposed by the regulatory authorities. The carefully selected event can then be used to start a series of mandatory biosafety assessment experiments expected to prove or disprove any putative biosafety hazard. The consequence of this approach is that nearly 99% of all transgenic events, and often those with the highest levels of expression, must be discarded. Already, the first step of mass production of many hundreds of similar events and the subsequent destruction of most of them is beyond reach for most public research institutions, in developing as well as in developed countries, and funding agencies are not prepared to take over such costs.

The biosafety assessment starts with event-independent studies, related to the introduced genes and their function, and are valid for all events produced with these genes. These studies are followed by exposure evaluation tests for the novel trait, its intended use and bioavailability, as would be the case for a product like beta- carotene. This study alone takes about three years, because during the pre-field trial phase the materials have to be produced in dedicated plant growth chambers and greenhouses, which is very expensive and production levels are low. Next in line are protein production and equivalence analyses for the proteins encoded by the introduced genes. For this purpose the proteins have to be isolated from the plant, characterised biochemically, and their function confirmed. Further studies include a demonstration of lack of homology to known toxins and allergens, gastric degradation studies, heat stability, acute toxicity tests in rodent feeding experiments.

This all would seem reasonable if it were not for the fact that most people have been eating these genes and their products from a number of other food sources throughout their lives. At one point, somebody even suggested to analyse whether known daffodil toxins had been introduced into Golden Rice along with the daffodil gene used to reconstitute the beta-carotene biosynthetic pathway, which totally lacks scientific basis: what has been transferred is one defined piece of DNA which is analogous to genes in other organisms, and performing the same function, which has no relation to any toxin or allergen. These studies take at least two years of intensive work in a well equipped biochemistry laboratory.

The event-dependent studies are even more cumbersome; they include:

• Molecular characterisation and genetic stability: data on single-copy effect; marker gene at same locus; simple integration; Mendelian inheritance, including phenotypic and biochemical evidence for stability over at least three generations; no potential gene disruption; no unknown open reading frames; no DNA transfer beyond borders; no antibiotic resistance gene or origin of replication; insert size limited to the minimum necessary; sequencing of insert and flanking regions.
• Expression profiling: gene expression levels at key growth stages; evidence of seed-specific expression.
• Phenotypic analysis: field performance, typical agronomic traits, yield compared to isogenic lines; pest and disease status must be same as parent (unexpected improvements are not tolerated).
• Compositional analysis: data from growing the event over two seasons at six locations in three replicates on proximates, macro and micronutrients, antinutrients, toxins, allergens; data must be generated on modified and isogenic backgrounds.
• Environmental risk assessment: this type of analysis takes 4-5 years of work by an entire research team.

It is obvious that no scientist or scientific institution in the public domain has the potential, funding or motivation to perform such lengthy, expensive biosafety experiments. It comes as no surprise then, that virtually all transgenic events that have been carried through the deregulatory process so far are—directly or indirectly—in the private sector and are restricted to high-value crops. Humanitarian projects do not fall into this category, even though they would benefit millions of people. There is a lot of goodwill in the public and in the private sectors worldwide to exploit the potential of green biotechnology for the benefit of the poor. However, without a realistic risk assessment approach, funds for public research will not be capable of doing the trick. Scientific progress would become detached from product development and the population at large would not benefit from progress.