Bt Toxicity Confirmed: Flawed Studies Exposed
Researchers confirm Bt toxicity to non-target beneficial insects and show how experiments claiming to refute their results were designed not to find the effect. A fully referenced version of this article is posted on ISIS members website and is otherwise available for download here.
A new study confirms that the Cry1Ab Bacillus thuringiensis (Bt) toxin present in genetically modified (GM) crops kills the larvae of the two-spotted ladybird (Adalia bipunctata L.), a species that GM supporters claim to be unaffected by the toxin .
The study raises questions regarding the integrity of previous work published by GM proponents, whose experimental protocols were re-tested and shown to lack the scientific rigour required to pick up signs of toxicity even in target insects that the pesticide is designed to kill.
Bt toxins are present in many GM crops including cotton and maize. Monsanto’s Mon 810 Bt maize is currently approved for cultivation in Europe, although it has been banned by individual nations including Hungary, France, Austria, Germany, Greece and Luxembourg due to health and environmental concerns. Many previous studies have found effects on health and the environment (see  Bt Crops Failures and Hazards, SiS 53,  More Illnesses Linked to Bt Crops, SiS 30).
Previous Bt toxicity studies slandered by GM proponents
GM proponents claim that certain Bt toxins are effective against limited orders of insects, with Cry1Ab killing only Lepidoptera (butterflies and moths) such as the common maize pest the European cornborer. However, a peer-reviewed study published by Angela Hilbeck and colleagues at the Swiss Federal Institute of Technology in 2009 showed increased mortality in ladybirds exposed to the ‘activated’ toxin that had been coated on their food – meal moth eggs ; the team had found similar effects in green lacewings previously [5-7]. The activated version refers to the cleavage of the pro-toxin to produce the actual toxin.
In response to their original publication, a coordinated effort aimed at discrediting their findings appeared in the journal Transgenic Research, which included two highly charged critiques [8, 9] and a study led by Jörg Romeis from Agroscope, Switzerland, which failed to detect any toxicity . They concluded that the results of Hilbeck’s team were ‘false-positives’ and artefacts of a poor study design. One critique went as far as suggesting the work was ‘pseudo-science’. Agroscope, a Swiss federal governmental research organisation, is linked to the agrotech giant Syngenta, which along with Monsanto, produces Cry1Ab GM crops. These hostile attacks were triggered by the ban of Mon 810 maize in Germany based on results obtained by Hilbeck’s team among 30 other scientific publications showing harmful effects from the pesticide.
Addressing discrepancies between previous studies
The new work from Hilbeck’s team aimed to address the discrepancies between their own findings and those of their critics. First they conducted a ‘proof-of-concept’ experiment where they tested both their original protocol and Agroscope’s protocol on the target species, the European cornborer.
In the Hilbeck team’s original study, the ladybird larvae were exposed continuously for 10 days to a microbially-produced purified version of Cry1Ab or a microbially produced ‘empty’ version lacking the toxin. They were exposed through coating their food – meal moth eggs – with the toxin. The Agroscope protocol on the other hand, exposed the larvae for only 24 hours at a time through a sugar/water droplet with or without the toxin. As the larvae are carnivorous and cannot survive on a sugar diet alone, they were transferred to petri dishes with untreated moth eggs, thus giving them a period to recover from the exposure. This exposure/recovery was apparently repeated 4 times in total.
So, the aim of the new study was to understand if the differences in these protocols may have accounted for the opposing results obtained by Hilbeck’s team and by Agroscope. By testing target species that the toxin is designed to kill, any weakness in the protocol would become apparent.
Hilbeck’s team repeated the basic protocols by exposing 4 day old larvae to Bt maize as well as near isogenic non-GM maize sprayed with Bt toxins either continuously for 7 days, or for 24 hours followed by untreated non-transgenic maize for 6 days. They found high levels of mortality following continuous exposure as expected (just below 100 % with both types of exposure). The mortality rates dropped by half when animals were exposed to Bt sprayed plants for 24 hours only. Exposure to Bt maize for only 24 hours did not even cause mortality rates to rise above unexposed control groups.
An experimental protocol that cannot detect toxicity of a pesticide on a target species is clearly not fit for testing potential harm to non-target species.
In addition to insufficient exposure time, other flaws in the Agroscope experiments were noted by Hilbeck. Sugar/water droplets to which Bt toxin was added were found to dry up overnight, leaving the levels of exposure undetermined. Only one dose was tested, as opposed to three tested in the original study by Hilbeck’s team . There was no clear description of the number of animals used or the number of times the experiments were replicated, whereas performing 3 replicates is standard in laboratory studies.
Re-testing effects of Bt toxin on ladybird larvae with a new combined protocol
To counter the criticisms aimed at their previous study, Hilbeck’s team adopted a combined protocol consisting of 7 days continuous exposure to a sugar/water solution with or without the Bt toxin placed on cotton balls to prevent them drying up. After 24 hours, instead of allowing a recovery period, the cotton balls were replaced with fresh cotton balls with or without Bt toxin solutions. Additional meal moth eggs coated in the toxin were given to provide an adequate diet and ensure continuous exposure to the toxin.
After only 6 days of exposure, mean mortality rate was 40 % compared to around 25 % in unexposed larvae. The greatest difference in mortality between treated and untreated animals peaked at 4 days where there was around a 20 % increase in mortality over untreated animals, after which it began to level off.
The new work not only corroborates the team’s previous findings , but also Agroscope’s failure to detect toxicity on non-target insects .
It is important to distinguish the difference between the natural bacterial toxin and the modified version inserted into GM plants. Neither of the original studies by Hilbeck or Agroscope used the versions expressed in GM crops, which are significantly modified. Modifications are made to increase the ‘performance’ of the toxin, including changing the promoter and enhancer elements to increase production of the protein; changes in sequence to increase solubility of the toxin, as well as altering the final portions of the gene to ensure the termination of gene expression.
In reality, it is difficult for researchers to obtain the transgenes made by industry, as there are strict patent laws and resistance to giving permission to conduct independent research on their products. Previous studies have shown that the modified toxin is more toxic than their naturally produced counterparts, with green lacewings suffering from delayed development and reduced survival (see  GM Food & Feed Not Fit for "Man or Beast", ISIS Report). The new study from Hilbeck’s team was carried out with Cry1Ab toxin from another independent lab, which is not the same as that produced in Bt maize or those in earlier work that highlights these differences in toxicity. The story remains unclear as to which versions are more toxic, or if there is any non-toxic version. (The variable expression of transgenes found in different Bt crops under different environmental conditions also complicate matters (see  Scientists Confirm Failures of Bt-Crops, SiS 28)). Currently, no regulatory body requires testing of the modified transgene, which means that their effects have not been properly assessed in any version.
Attacks on researchers with ‘inconvenient’ results
The attack on scientists who publish data that happen to go against the safety of biotech products are under immense pressure from GM proponents, industry and even regulatory bodies. The work is highly scrutinised in a manner rarely seen in other non-profit-driven subject areas. As Hilbeck said in a comment piece, deliberate counter-studies and confrontational attacks have also been witnessed with other commercial products such as bisphenol A, asbestos and tobacco . The team were never given the opportunity to respond to their critiques.
In the case of Bt toxicity, this is not the first time that the researchers have faced such scrutiny; the publications on lacewing lethality [5-7] drew a similar response from some of the same authors that targeted the ladybird study.
Studies into the toxic effects of the Bt toxins are beginning to shed light on the wider effects of Bt toxins to non-target insects. This knowledge is critical to agricultural success with insects like the ladybirds playing an important biological function due to their predation on crop pests such as aphids and white flies. As it stands, the full off-target effects of the Bt toxins are not understood, with a reported 91 % of Bt toxins tested on 10 or less species, most of which are presumed target species . Independent studies have however, also linked Bt exposure to abnormal growth in snails  and caddisflies (see  Bt Crops Threaten Aquatic Ecosystems, SiS 36) and reduced fitness of water fleas . Off-target effects need to be investigated thoroughly prior to the release of such products. With Bt crops already widely commercialised, we are left with the option of withdrawing them from the market until irrefutable evidence of their safety becomes available.