Whenever a new technology is introduced there are always issues to be addressed and risks to be assessed. With Bollgard^® Bt cotton, the major issues related to two areas – potential risk to the environment, and the potential for insect resistance to develop. There is also a food/feed safety risk assessment conducted because cotton seed oil is used in food and cotton seed meal is fed to livestock.

Assessment of Environmental Risk

A. Agronomic Performance

Prior to the introduction of Bollgard^® Bt cotton in the US in 1996, detailed agronomic observations were made in extensive field trials over several years. Agronomic, pest and disease susceptibility observations confirmed that, with the exception of resistance to lepidopteran pests, Bt cotton was agronomically within the normal range of variability for commercial cotton varieties (Hamilton et al 2002). Furthermore, Bollgard^® cotton has been grown commercially in the U.S., Australia, Mexico, South Africa, China, Argentina, Colombia, Indonesia and India since initial introduction in the US. No unusual plant pest characteristics or unintended environmental effects have been observed that are attributed to the inserted DNA and expressed proteins, as confirmed by the extensive studies developed prior to, and subsequent to, regulatory approvals and market introduction. Agronomic performance of Bollgard® cotton and protection from damage by Lepidopteran insect pests have been as expected (Edge et al 2001, Benedict and Altman 2001, Gianessi and Carpenter 1999).

B. The donor organisms

The safety of the donor organisms of the Cry1Ac and nptII genes contained in Bollgard^® cotton is well established. The Cry1Ac gene encodes the insecticidal protein derived from the common soil bacterium Bacillus thuringiensis subsp kurstaki (B.t.k.). There is a history of safe use of the Cry1Ac protein in microbial Bt-based products (USEPA 1988, IPCS 2000). Microbial formulations of Bacillus thuringiensis (Bt) that contain the insecticidal protein have been registered in numerous countries worldwide, and have been safely used for control of lepidopteran insect pests for more than 40 years (Betz et al 2000). The Cry1Ac protein produced in Bollgard^® cotton event 531 is 99.4% identical in predicted amino acid sequence and comparable in biological activity to the Cry1Ac protein found in nature and in commercial Bt microbial formulations (Hamilton et al 2002). Bacillus thuringiensis and Bt microbial formulations have been shown to be very specific to the target insect pests, and do not have any deleterious effects on non-target organisms such as beneficial insects (other than closely-related lepidopterans), birds, fish, and mammals, including humans (USEPA 1988, Betz et al 2000). The NPTII protein expressed in Bollgard® cotton is chemically and functionally similar to the naturally occurring NPTII protein. The NPTII protein (donor is E. coli) is ubiquitous in the environment and found in microbes present on food and within the human digestive system (Fuchs et al 1993, USFDA 1994).

C. Effect on non-target organisms

There is extensive information about microbial preparations of Bacillus thuringiensis subsp. kurstaki (B.t.k) containing Cry proteins, including the Cry1Ac protein, which demonstrate that these proteins are non-toxic to non-target organisms (USEPA 1988, Betz et al 2000). The literature has established that the Cry1Ac protein is selective for lepidopteran insects, binds specifically to receptors on the mid-gut of lepidopteran insects and has no deleterious effect on beneficial/non-target insects. The safety of the Cry1Ac protein expressed in Bollgard^® cotton to non-target organisms was confirmed on several representative organisms (Hamilton et al 2002). With the use of in-plant Bt technology, non-target, beneficial insects are not harmed as they are with many broad spectrum insecticidal sprays (Benedict and Altman 2001). Bt protein affects a specific set of target pests, and unrelated non-target pests are not affected. However, pyrethroids have been demonstrated to affect a broad range of non-target species (Badawy and El-Arnaouty 1999). Therefore, since the use of Bollgard^® cotton has resulted in a reduction in conventional synthetic insecticide applications (Gianessi and Carpenter, 1999) increased populations of beneficial insects in Bollgard^® cotton fields are expected. Several studies have shown that predatory non-target organisms can be more active in Bollgard^® cotton as biological control agents for secondary pests (Edge et al 2001). Post commercial monitoring indicates that populations of predatory, non-target organisms are significantly higher in Bt cotton than in non-Bt cotton that was sprayed with insecticides, (Head et al In Press a, Head et al 2001, Roof and Durant 1977) and provide biological control of secondary pests. Studies have reported lower levels of secondary pests such as Spodoptera in Bt cotton related to the higher number of predator insects present (Smith 1977).

D. Potential of Bt cotton to develop as a weed

Gossypium hirsutum is well characterized and has a safe history of production under a broad range of agricultural environments. Past intensive selection to develop germplasm adapted to high productivity under agricultural conditions makes it unlikely that cotton could effectively compete and survive in the environment as a weed. Cotton is not found as a weed in the global cotton production areas. Bollgard^® cotton does not have any different weediness characteristics than other conventional cotton varieties (Hamilton et al 2002). Bollgard^® cotton does not exhibit different agronomic or morphological traits compared to controls, that would confer a competitive advantage over other species in the ecosystem in which it is grown. Also, there is little probability that any Gossypium species crossing with Bollgard^® cotton could become more weedy. Thus, there is no evidence that insertion of the Cry1Ac coding sequence into the cotton genome has had any effect on the weediness potential of the cotton plant.

E. Environmental consequences of pollen transfer

Cotton is predominantly a self-pollinating crop but can be cross-pollinated by certain insects (Niles and Feaster 1993). However, outcrossing of the Cry1Ac coding sequence from Bollgard^® cotton to other Gossypium species or to others of the malvacea family is extremely unlikely for the following reasons (Percival et al 1999):

Volunteer plants are not a significant issue and can be controlled with many registered herbicides. Thus, the environmental consequences of pollen transfer are negligible due to limited movement, natural genetic barriers that preclude outcrossing with native cotton, with no known compatibility with any wild relatives. The safety of the Cry1Ac protein is well documented and the cry1Ac gene would not confer any competitive advantage (Hamilton et al 2002).

F. Insect Resistance Management (IRM)

Several publications (Roush 1999, Benedict and Altman 2002, Fitt 2002/In Press) have discussed the potential development of resistance to Bt cotton at some length, and the reader is referred to these texts for a detailed discussion. The intent here is to provide an overview of insect resistance management strategies that have been in place for six years since Bt cotton was first commercialized in 1996.

There is no doubt that the potential development of resistance poses a significant challenge in the effective deployment of Bt cotton, but the same challenge also applies when attempting to develop effective insecticides, or other means of control. Experience with conventional breeding to enhance insect resistance in crops, and particularly experience with developing insecticides to control insect pests of cotton supports the case that an insect resistance management strategy is essential in order to preserve the durability of product effectiveness, irrespective of the source or mode of control. In the specific case of cotton pests and Bt, there is ample evidence that cotton bollworm, Helicoverpa armigera as well as other lepidopteran pests have developed resistance to a multitude of insecticides. Resistance to topical Bt spray applications has also developed in field populations of diamond back moth (Tabashink 1994). Thus, it is critically important that Bt and other genes that confer resistance to the major lepidopteran pests be managed and deployed responsibly and effectively with an IRM strategy, recognizing that different IRM strategies must be developed to meet different needs. For example, the needs of a typical small farmer growing less than a hectare of cotton in a developing country are quite different to the needs of large commercial farmers growing a large block of 100 hectares or more of Bt cotton in an industrial country. Appropriate IRM strategies have been developed in countries where Bt cotton has been commercialized, usually involving public and private sector entities working together towards the mutual objective of preserving the durability of resistance.

Whereas specific IRM strategies need to be developed to meet the needs of particular cotton production systems, the factors that impact on the development of resistance to Bt, conventional insecticides, or conventional host plant resistance are the same (Head et al in Press b; Shelton et al 2000, Roush 1997). These three factors are:

When the first Bt cotton application was submitted for consideration in the US in 1995, the inclusion of an IRM plan as part of the registration of Bollgard®, during discussions with EPA and Monsanto, was unprecedented (Roush 1997). The IRM plan was developed as a result of collaboration between USDA, universities, and Monsanto to articulate a deployment strategy over the short, medium and long term, (Table 15). The US IRM strategy features a short term program utilizing a high dose of Bt, refugia, agronomic practices that limit exposure of pests to the active protein, implemented in association with an IPM strategy and a rigorous monitoring system for the early detection of resistance. The short term strategy is fortified by a mid term strategy to develop a Bt cotton with two genes, a remediation strategy and a ‘community refuge option’ to promote grower flexibility and maximum IRM compliance. The long term strategy includes all the elements in the short and mid term plus the incorporation of host plant resistance and other novel insecticide genes, as well as defining the value of alternate hosts as contributors to the overall refuge. It is noteworthy that since its inception in 1996, the US strategy has operated effectively and that key projected products, such as Bollgard® II have already been successfully developed, approved in Australia and ready for release in the USA. Similarly, the stringent IRM in Australia, successfully implemented since 1996, has already been revised to incorporate Bollgard^® II in 2002. China has successfully implemented a different IRM strategy featuring a Bt fused gene in conjunction with CpTi and a rigorous monitoring system. Other countries growing Bt cotton including India, Indonesia, Mexico, Argentina, and South Africa, have also developed and implemented IRM strategies to meet their specific needs and have precluded the development of resistance to-date.

The use of transgenic Bt cotton, deserves continued careful attention (Gould 1998) because cotton insect pests are subject to continuous selection throughout the season. Development of resistance could jeopardize the use of Bt as a conventional spray by farmers including organic growers, which is of particular concern to many NGOs opposing biotechnology. From the time that Bt cotton was introduced in 1996, some critics have predicted that the development of resistance was imminent. Indeed, claims have been made by critics that resistance has already developed, but to-date investigation has consistently failed to confirm those claims. Whereas the risk of resistance developing is real, requiring the implementation of rigorous IRM strategies, it is equally important to acknowledge the significant benefits that have already been delivered following the planting of 13 million hectares of Bt cotton globally since 1996. Had Bt cotton not been deployed in 1996, these significant benefits would not have been realized at an enormous opportunity cost to millions of farmers who grew 13 million hectares of Bt cotton in eight countries. It is noteworthy that despite predictions to the contrary by critics, insect resistance to Bt cotton has not yet been detected in the large area of Bt cotton deployed globally. Since the introduction of Bt cotton in 1996, the Bt genes Cry1Ac, the fused gene Cry1Ab/Cry1Ac, and the cowpea trypsin CpTi gene have been successfully deployed to confer resistance against lepidopteran cotton pests. Notably, in the interim, Bollgard^® II has been developed which provides a second line of defense and more effective control. Bollgard^® II is already approved in Australia for the 2002/03 season and approval is expected for the US in 2003. Other products that are also expected to be available in the near term include a dual Bt gene cotton (Cry1Ac and Cry1F) from Dow AgroSciences in 2004 and an insect resistant cotton with a VIP gene from Syngenta in the same year. The private and public sectors in both developing (China and India) and industrial countries (USA and Australia) have active programs to develop new Bt and other novel genes as well as the incorporation of improved conventional host plant resistance.

Thus, in summary the development and implementation of IRM strategies in conjunction with the introduction of Bt cotton in 1996, and its expansion to cover approximately 4.3 million hectares in 2001, have made a significant contribution to the effective deployment of Bt genes. Society has placed a high value on Bt cotton because it can reduce by at least half the volume of broad spectrum conventional insecticides applied to cotton with significant economic, environmental and social benefits and health implications for producers, particularly small farmers in developing countries. It is reassuring to know that the initial plan to broaden and diversify the mechanisms of resistance is materializing in terms of new approved products such as Bollgard^® II and that other new Bt genes and novel resistance genes are expected in the near term. However, these expectations should not lead to complacency and any relaxing of the rigor with which Insect Resistant Management strategies are implemented by small and large farmers in both developing and industrial countries. It would be valuable now to convene an international Review of Insect Resistant Management Strategies that would consolidate the considerable knowledge and experience gained thus far, and utilize it to develop an international strategy that could guide implementation coincidentally with the further expansion of Bt cotton globally in the near term. This would be particularly important for the large number of developing countries that stand to benefit significantly from Bt cotton but require assistance to ensure that Bt cotton is deployed effectively.