Bt plants and non-target organisms: Jörg Romeis

“We conduct targeted tests on representative organisms.”

Jörg Romeis, Agroscope ART, Switzerland
Head of the working group GMOs in Integrated Plant Protection within the West Palaearctic Regional Section (WPRS) of the IOBC (International Organization for Biological Control). This working group initiated a project in 2005 with the aim of developing usable guidelines for assessing the risks of insect-resistant GM plants for non-target organisms, particularly insects. The working group comprises scientists from public research institutes, representatives of regulatory bodies and industry.

Lacewing: Worst-case scenario tests in the laboratory have shown that these beneficial insects are not sensitive to the Bt toxin. Jörg Romeis: “We do not regard further tests in the field to be necessary in this case.”

Monarch butterfly: Toxicological tests have shown sensitivity to the Bt toxin. Further laboratory tests are therefore necessary, with realistic intake quantities of the Bt toxin and reaction times. In addition, field tests will be conducted to clarify what quantities of toxin the caterpillars actually come into contact with.

GMO Safety: Btplants can have different effects on the ecosystem depending on the regions in which they are grown and on how they are used agriculturally. Is it possible to conduct a risk assessment that takes into account the diversity of ecosystems? What can be studied in the laboratory and what must be studied in the field?

Jörg Romeis: Ecosystems are highly complex and vary widely between regions. Nevertheless, the basic approach to the risk assessment of Bt plants is always the same. The approach has proved effective, even though it is possible that fewer species of butterfly are at risk when Bt maize is planted under Western European climate conditions than in the tropics. The procedure consists of several stages and is based on problem and risk analysis. This leads to targeted laboratory tests that are designed in such a way that they can answer a large number of the questions relating to possible risks. If the laboratory tests indicate that a particular species may be at risk, further tests are conducted in the greenhouse and, if necessary, in the field. In near-natural conditions, researchers investigate the risk in greater detail, studying for instance the probability of the particular non-target species coming into contact with the Bt protein.

GMO Safety: How exactly can your model be used in the safety assessment of Bt plants?

Jörg Romeis: Firstly, we need to describe the particular risk in detail. We call this problem analysis. The question to be answered is: Does the Bt protein have toxic effects on non-target organisms, i.e. on organisms that it is not designed to control? However, this kind of risk analysis cannot be conducted for all the non-target organisms in an ecosystem, so we have to look for suitable ‘representatives’.

The problem analysis that we carry out at the beginning helps with this. When selecting representatives, we consider various criteria: which variant of the Bt protein are we dealing with? Where is it expressed (in the leaves, pollen or only in the roots)? Is it produced in the plant throughout its life or only during particular growth phases? Which insects come into contact with the Bt protein? Is this contact direct and long-term or only occasional? And which insects ingest the Bt protein through their prey? The aim is to analyse as closely as possible which organisms are at greatest risk. The risk analysis can then focus on these groups. However, another very important criterion for selecting the representatives is that we need to be able to raise insects of this species in the laboratory and conduct standardised and informative tests on them.

The tests on representative organisms always start with worst-case scenarios. For instance, high doses of Bt protein are fed to the insects during all stages of their development, or tests are carried out with sub-lethal doses over long reaction times.

If the toxin is found to affect the insects under these worst-case conditions, further tests can be conducted in the laboratory, and also in semi-field and field conditions. At this point the purpose is to investigate the risk in realistic scenarios. We want to develop our test protocols so that we obtain meaningful information about possible risks at the laboratory test stage – in other words, we should be able to rule out a risk to non-target organisms with a high degree of certainty.

GMO Safety: Possible toxic effects on beneficial organisms are an important aspect of pesticides as well. There are recognised procedures for identifying these effects prior to market approval. Do you see parallels with the ecotoxicological testing of Bt plants?

Jörg Romeis: The procedures for the risk analysis of Bt proteins have been developed based on those for pesticides. We don’t need to reinvent the wheel here. We start with effective tests under standard conditions that fulfil the requirements of ‘good laboratory practice’. However, since a chemical is not the same as a Bt protein, the test protocols have to be adapted accordingly. We also have to tailor our selection of test organisms according to which variant of the Bt toxin is produced by the plant in question.

Let me illustrate this with an example: the Bt toxin variant used in Bt maize, the Cry1Ab protein, has a highly specific effect on caterpillars. By contrast, the Cry3Bb1 protein acts specifically on beetle larvae. We therefore believe it is important to focus particularly on butterfly species in the first instance and on beetle larvae in the second.

We are calling for test species to be selected on a case-by-case basis. For instance, risk analysis tests for future GM plants with the Cry1Ab protein should focus on butterfly species, since so far there has been no indication at all that Cry 1Ab has a direct toxic effect on organisms outside the butterfly group. In North America, the Monarch butterfly (Danaus plexippus) has become established as a test organism.

By contrast, for Cry3Bb1 events, the focus should be on testing beetle species. Suitable species would be ground beetles and ladybirds. We know that these are exposed to the Bt toxin and that they play an important role as antagonists of plant pests. Test protocols for investigating the toxic effect of transgenic protein are already available for ground beetle species (Poecilus), the predatory beetle Coleomegilla maculata found in North America, and the seven-spotted lady beetle (Coccinella septempunctata) in Europe.

In addition to these carefully selected organisms, standard organisms like the honey bee (Apis mellifera) and the brandling worm (Eisenia fetida) are usually studied as well. These species are usually selected because there are standard tests that can be conducted on them.

GMO Safety: How flexible is your model with regard to different ecosystems? The environments where Bt plants are grown can vary greatly. Can your model also be used for risk assessment in regions with great biological diversity?

Jörg Romeis: This is a very important aspect. Our model can be used in all parts of the world. The differences in ecosystems are taken into account in the problem analysis right at the beginning, before we start testing. Our model is flexible enough to allow us to add additional organisms, depending on the country and region, if they cannot be ruled out in terms of risk.

GMO Safety: How reliable and clear are the predictions which you are able to make about possible risks to beneficial and other non-target organisms? What relationships do you have to be aware of and what is simply background knowledge?

Jörg Romeis: A risk assessment is always hampered by uncertainties. For instance, if we can show that certain beneficial organisms are not susceptible to the Bt protein, we can make a relatively reliable prediction about whether or not they are at risk. It is not necessary to investigate all the possible interactions between the plant and the living environment – that would be ‘nice to know’, but not necessary for answering the question of the potential risk. The analysis should always concentrate on those species for which the greatest risk is to be expected – in the case of Bt toxins, these would be the butterfly species.

GMO Safety: How does your model take account of possible long-term effects?

Jörg Romeis: Long-term effects are generally hard to identify. But we account for them by e.g. exposing the indicator organisms to high doses of the toxin or by administering sub-lethal doses over a longer period of time. EU legislation also prescribes post-approval monitoring. This involves monitoring for unexpected effects alongside cultivation for ten years. Monitoring can also serve to dispel any remaining uncertainties from the risk analysis.