Interface of biotechnology and ecology for environmental risk assessments of transgenic fish Robert H. Devlin 1 , L. Fredrik Sundstro ¨m 1 and William M. Muir 2 1 Fisheries and Oceans Canada, 4160 Marine Drive, West Vancouver, British Colombia, V7V 1N6, Canada 2 Department of Animal Science, Purdue University, West Lafayette, Indiana, 47906-1151, USA Genetically engineered fish with enhanced phenotypic traits have yet to be implemented into commercial applications. This is partly because of the difficulties in reliably predicting the ecological risk of transgenic fish should they escape into the wild. The ecological consequences of the phenotypic differences between transgenic and wild-type fish, as determined in the laboratory, can be uncertain because of genotype-by- environment effects (GXE). Additionally, we are limited in our ability to extrapolate simple phenotypes to the complex ecological interactions that occur in nature. Genetic background can also shape the phenotypic effects of transgenes, which, over time and among different wild populations, can make risk assessments a continuously evolving target. These uncertainties suggest that assessments of transgenic fish in contained facilities need to be conducted under as wide a range of conditions as possible, and that efficacious physical and biological containment strategies remain as crucial approaches to ensure the safe application of transgenic fish technology. This year marks the twentieth anniversary of the first research on transgenic fish [1]. During this time, more than 30 species of fish have been genetically engineered, including many of the major world aquaculture species (e.g. carps, tilapia, catfish and salmonids). Because demand for foods from aquatic sources continues to grow in response to human population growth and diminishing wild fish stocks, transgenic technologies are being explored for their potential to increase aquaculture production efficiency and yield. Although several commer- cially important traits are being modified (Table 1), to date, most effort has been targeted to enhancing growth and feed conversion efficiency through the transfer of growth hormone (GH) gene constructs (Figure 1). Despite initial achievements, commercial implemen- tation of transgenic fish technology for food production has remained elusive. The lack of implementation has been, in part, due to the technical challenges and regulatory requirements in addition to the significant opposition to transgenic fish within the scientific, public and aquaculture-producer sectors [2,3]. Such concerns are derived largely from legitimate (and far-fetched) specu- lation about food safety issues and the potential ecological harm that might arise if transgenic fish were to escape into nature [4]. Food safety issues from the view of the National Research Council (NRC) [4] are generally minimal and, in most cases, appear to be well managed under the existing regulatory frameworks present in many global jurisdictions. For environmental risk assessments and mitigation of transgenic fish, conceptual and theoretical bases for assessment have been elaborated [4–9]. However, for most transgenic fish, insufficient publicly accessible data are available to resolve the complex issues that are necessary both for risk assessments and to develop consumer and commercial confidence. For transgenic fish technology to move forward, empirical risk assess- ment research needs to be undertaken and presented in parallel with strain development, enabling this maturing technology to have the essential information available to support regulatory and social requirements. This article discusses the current status and uncertainties associated with empirical risk assessment research on transgenic fish, including evaluation of potential impacts, estimation of fitness and approaches for biological containment. Phenotypic effects of GH transgenesis in fish The GH gene constructs used in fish have generally followed the designs used in other vertebrate systems [10], with piscine DNA constructs now most commonly used. GH overexpression in fish can cause significant enhance- ment of growth rate, which can result in large differences in size at a particular age in several species (e.g. typically 2–10-fold, and up to 35- and 37-fold weight gain in loach and salmonids, respectively; Table 1, Figure 1) and a compression of the life history of the species [11,12]. Fast- growing GH-transgenic fish can mature to have body sizes larger than those seen in nature [13–15], but in salmonid species, which die after sexual maturity, GH-transgenic individuals reach the size of a normal wild-type adult in a shorter period of time [12,16]. The distinct effects observed among species and strains reveals the need for risk assessments to be performed on a case-by-case basis. Because GH acts on many processes in addition to growth, transgenic fish overexpressing this hormone can Corresponding author: Devlin, R.H. (devlinr@dfo-mpo.gc.ca). Available online 27 December 2005 Review TRENDS in Biotechnology Vol.24 No.2 February 2006 www.sciencedirect.com 0167-7799/$ - see front matter Crown Copyright Q 2005 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2005.12.008