Abstract:

In 2002, the Environmental Risk Management Authority of New Zealand (ERMANZ or the Authority) approved an application by the company AgResearch, Ltd. to create and dispose of genetically modified (GM) bovine. As part of its risk management strategy, the Authority imposed a requirement for monitoring soil microorganisms for uptake of transgenes by horizontal gene transfer (HGT). HGT is routinely considered in biosafety risk assessment because it may be a process that underpins eventual adverse effects to human health or the environment. 

While granting approval to the company to make GM bovine, the Authority considered that HGT-related risks were potentially non-negligible and therefore approval was contingent upon meeting regulatory controls that made the risk negligible through risk mitigation. ERMANZ’s requirements placed upon AgResearch the burden to conduct a monitoring effort capable of delivering the risk mitigating activity that the Authority sought. Using colony hybridisation and PCR, AgResearch monitored antibiotic resistance phenotypes in soil bacteria cultured from samples of soil taken from offal pits containing the carcasses of GM bovine and surrounding control sites between 2004 and 2009 in an attempt to determine if any of the antibiotic resistance in soil bacteria was caused by the uptake of transgenes originally from the GM animals.

The Centre for Integrated Research in Biosafety (INBI) at the University of Canterbury has reviewed AgResearch’s reports of these monitoring efforts released to GE Free New Zealand (in Food and Environment) under the Official Information Act.

In summary, AgResearch undertook a challenging project at the forefront of theory and practice in microbial science. However, their experiments suffered from a design that was incapable of detecting HGT with the sensitivity necessary to detect bacteria that might cause the adverse effects of concern to the Authority, including but not restricted to bacteria developing antibiotic resistance because they acquired a resistance gene used in the production of GM bovine. Notably, the sampling depth in all but one year was in the range of 2-6 m above the soil interface with the carcasses. Importantly, no study confirmed that the samples were taken from soil in contact with carcasses. 

Moreover, the suitability of control sites and the efficacy of the sampling were not demonstrated. Not just the design but the standards of follow-up on observations and determining causes of negative results (e.g. particularly from routine molecular work such as sequencing and PCR) was below what we would expect, and what we would expect to be sufficient for assurance that risk management controls were met. INBI finds that these experiments were irretrievably flawed for providing baseline data for future soil analysis, effectively monitoring HGT as a risk management strategy or influencing the assessment of the risk of HGT in future applications. We suggest ways AgResearch could have chosen to improve experimental designs and lead to more confidence-building outcomes.