Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 Current Research: What Is Known and What Are the Gaps? P articipants were invited to the workshop, by design, for their wide array of expertise. They included people with primarily research, regulatory, or land management responsibilities; those who are involved in research and development of genetically engineered organ- isms (GEOs) in different taxa (plants, trees, microbes, insects, fish) and those who focus more on the biology and ecology of wildlife and habitats; and those who work in government, academia, and nonprofits. For this reason, the planning committee began the workshop with presentations that would allow participants to get a sense of the âstate of the scienceâ in different research areas. As summarized in this chapter, the first set of presenters provided an overview of GEO research by taxa. In the following session, researchers shared lessons from other disciplines that may have applications to GEO research. Each set of presentations was followed by a short, but lively discussion period. STATUS OF RESEARCH ON EFFECTS OF GEOS ON WILDLIFE AND TERRESTRIAL AND AQUATIC HABITATS The current state of research and development and commercialization of GE plants, microorganisms, insects, and other animals is variable, rang- ing from widespread production of some GE crops to very circumscribed research, mostly through modeling and in labs and other contained set- tings, of the other taxa. Confinement so that transgenic organisms are not
10 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT released into the environment is of high concern, both from biological and legal/regulatory standpoints. Research on Effects of GE Crops â La Reesa Wolfenbarger L. LaReesa Wolfenbarger (University of Nebraska, Omaha) framed consideration of the environmental effects of GE crops by looking at research in three, interrelated categories: the impacts of GE crops on wildlife food (insects eaten by birds and other animals) in farm fields and adjacent land; impacts on wildlife in farm fields; and impacts on wildlife in land adjacent to farm fields, such as grassland, forests, riparian areas, wetlands, or streams. Wildlife Food in Farm Fields and Adjacent Habitat According to Wolfenbarger, most relevant studies have focused on the abundance of wildlife food, particularly non-target and beneficial arthro- pods, in the presence or absence of a GE crop. Although the basic research question is whether and how GE crops impact the abundance of wildlife food, she noted that an important factor to emerge was the background effects of different agricultural practices. For example, a meta-analysis by Marvier at al. (2007) showed decreases in the abundance of non-target insects of the orders Coleoptera, Hemipteran, Hymenopterans and par- ticularly Lepidoptera in fields planted with transgenic cotton expressing Bacillus thuringiensis (Bt) proteins relative to non-transgenic cotton. How- ever, the abundance of all insects were much lower in fields planted with cotton crops (transgenic or not) sprayed with insecticide. A key finding of the study was that oneâs view of what is ecologically beneficial depends on the points of comparison. About 80 percent of cotton acreage in the United States is sprayed with insecticides, said Wolfenbarger. Similarly, results of a comparison of the effects of Bt-corn on wildlife food depended on whether it was compared with insecticide-sprayed or nonsprayed corn and the types of insecticide (Wolfenbarger et al., 2008); currently, she said, about 25 percent of the U.S. corn crop (75 percent of sweet corn) is treated with insecticide. Finally, Cattaneo et al. (2006) looked at the impact of Bt-cotton and other agronomic practices on the diversity of wildlife food in farm fields relative to the diversity in adja- cent habitat, using the adjacent, uncultivated area as the baseline. They found that relative to uncultivated areas, cotton cultivation had a negative impact on ant density and a positive impact on beetle density, but those findings were irrespective of whether the crops grown were or were not transgenic. The conclusion Wolfenbarger drew from these studies was that GE
CURRENT RESEARCH 11 crops do affect wildlife food, but that variations in agricultural practices, including cultivation itself, and the use of insecticides, can have larger effects. Turning to wildlife food in adjacent habitats, Wolfenbarger presented results of a study by Rosi-Marshall et al. (2007) that looked at growth and survival of Trichopterans (an insect order that includes the caddis fly and whose larvae are aquatic) that feed on corn by-products in streams. In lab tests, the Trichopterans that fed on Bt corn by-products showed decreased growth compared to those that fed on non-Bt corn by-products. Because these small flies are a basal component of the aquatic food web, their decreased growth may be of significance. Effects on survival were only detected in the lab at exposure rates two to three times the maximum measured in field sites. Wolfenbarger was the first of several presenters at the workshop to refer to the United Kingdom Farm Scale Evaluations (FSE) as a valu- able source of information on the effects of GE herbicide-tolerant crops on wildlife and habitat (see Box 2-1). Overall, the FSE showed mixed impacts of herbicide-tolerant crops, with wildlife food populations (seeds and arthropods) increasing, decreasing, or not affected depending on the crop and type of wildlife food (Andow, 2003). One finding of the FSE was that the herbicide Atrazine was more effective in controlling weeds than glyphosate, the latter which is used in conjunction with herbicide-tolerant crops. It is known that weed diversity and abundance affect wildlife food such as arthropods, so the effectiveness of the herbicide has implications for that food supply, noted Wolfenbarger. BOX 2-1 UK Farm Scale Evaluation: GE Crops at a Landscape Level From 1999 to 2004, a large study of the environmental impact of herbicide- tolerant GE crops was conducted in the United Kingdom, known as the Farm Scale Evaluation (FSE). Sponsored by the Department for Environment Food and Rural Affairs, farmers planted transgenic and unmodified maize, rapeseed (two types), and sugar beets on 60 sites around the country to measure the effects of these crops. Biodiversity, as exhibited by weeds, seeds, and invertebrates, was measured within the fields and at their margins. The FSE not only produced valuable data for a range of different studies, but also has proven to be a good model in how to set up large-scale comparative field studies. SOURCE: U.K. Department for Environment, Food, and Rural Affairs, 2008.
12 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT Wildlife in Farm Fields Wolfenbarger turned to two studies that explored the effects of GE crops on wildlife. First, when a subset of the FSE fields in the United Kingdom were surveyed to look at bird diversity and local abundance, the differences in abundance paralleled results in the food supplyâmore weed seeds meant more seed-eating birds. Yet, overall, models based on the FSE data predicted a change in conservation status of only 1 of the 39 species studied (Butler et al., 2007). Second, researchers at the University of Tennessee did not find differences in Brazilian free-tailed bat activity in Bt and non-Bt cotton fields smaller than 200 acres, but a laboratory feed- ing trial with a small number of the bats showed active Bt toxin in their fecal samples. Follow-up studies with field-collected fecal samples are in progress (Federico et al., 2008). Wildlife in Adjacent Habitats Wolfenbarger briefly explained her own research on wildlife in habi- tats adjacent to farm fields; results were still preliminary at the time of the workshop. Her strategy is to look at high conservation priority wildlife (grassland birds and butterflies) in farmland-adjacent habitats compared to natural habitats. Farming activity is one of many variables affecting them, and the use of GE crops is one variable among many different farming practices. Research Gaps GE crops have been shown to affect the local abundance of wildlife and of wildlife food, but these can only be appreciated in the context of overall agricultural practices associated with the GE crop compared to the conventional alternative, which may have greater influence on the direc- tion and the magnitude of the change. Context, Wolfenbarger underlined, becomes key to the interpretation of the results. She added that an important knowledge gap remains in knowing how changes in wildlife food within and adjacent to farm fields will affect populations, species, and special or sensitive communities. This, she suggested, may be an area where field studies and modeling will help scientists, regulators, and the public better understand the effects of GE crops on wildlife and habitats. Research on Effects of GE TreesâChung-Jai Tsai The 20th anniversary of GE trees occurred in 2007: The first was an herbicide-tolerant poplar with the aroA gene, developed in 1987. The first
CURRENT RESEARCH 13 insect-resistant (Bt and chloramphenicol aminotransferase) transgenic spruce was developed in 1993. Field trials took place in 1989 and 1993. In her presentation, Chung-Jai Tsai (Michigan Technological University) said despite this early activity, there is far more limited field experience with transgenic trees as compared to agricultural crops. The 363 notifications of field trials for transgenic trees have accounted for only 3 percent of total notifications to the Animal and Plant Health Inspection Service (APHIS) of the U.S. Department of Agriculture (USDA) since 1989. According to Tsai, in the United States, only one company (down from seven) and five universities (down from nine) are now engaged in active trials, mostly with pine, followed by poplar, eucalyptus, and sweet gum. Growth modi- fication is the primary trait studied in active trials, followed by flowering and lignin production. Tsai reported briefly on three cases involving GE trees, all outside the United States, that provided some findings with ecological implications. The first case involved fields studies in the United Kingdom and France of birch and poplar trees engineered for lignin modification (Pilate et al., 2002). In the second case, the effects of an insect-resistant, commercially- produced poplar in China on insects were examined (Ewald et al., 2006). The third study looked at gene flow from transgenic poplars in one of these Chinese plantations (Ewald et al., 2006). The three studies did not find that transgenic trees cause significantly different effects than their non-transgenic counterparts, but Dr. Tsai stressed that all three studies raised unanswered questions and that more research is needed, par- ticularly in the long term and across multiple sites, before any definitive conclusions can be drawn. The UK-France study, which was cut short halfway through its eight- year original plan because of vandalism at the UK site, looked at the target trait (lower lignin content) but also at comparisons of the ecological effects on the herbivory (as insects, microbes, and animals fed on the trees), soil mesocosms, and decomposition between GE and non-GE trees. A similar profile of insects was observed visiting the transgenic and non-transgenic trees, and the soil microbial diversity under the trees was also similar. A slightly higher rate of root decomposition was found in the transgenic trees, a finding that is consistent with the modified lignin trait, which would make the roots more susceptible to microbial degradation. (Pilate et al., 2002). In introducing the second caseâBt poplarâTsai noted China has been particularly aggressive in the development of GE trees to overcome its dependence on imported wood. Two GE poplars (Bt poplar and a â notification is an administratively streamlined alternative to a permit, if the organism A meets certain eligibility criteria and pre-defined performance standards.
14 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT Bt/API hybrid) are now in commercial release, the first such commercial use in the world, and planted on more than 237 hectares on seven sites. She said researchers studying non-target insect communities on these sites have found the effects to be dependent on the characteristics of the site and on the scale of the plantation, but that more research is needed to reach definitive conclusions (Ewald et al., 2006). Gene flow was studied at one of these sites, since a large concern with GE trees in the field is the spread of their pollen to non-GE trees, said Tsai. The study at one plantation showed the spread of male Bt pollen was very rare at distances of 500 meters or greater, and that seed germination under field conditions was also much lower than for seeds stored at room or refrigerated temperatures. Because the site was very arid, weather condi- tions might have influenced gene flow and germination, so no general conclusions about gene flow could be drawn (Ewald et al., 2006). Tsai explained that pollen dispersal is only one aspect of gene flow that makes the logistics of field study immense. Models have attempted to address this and other aspects. One model (known as Simulation of Transgene Effects in a Variable Environment, or the âSTEVEâ model) uses spatial and landscape data, combined with field measurements, to make predictions about gene flow. However, she warned, the model, while valuable, still needs experimental data to validate and improve it, and is not a silver bullet to make reliable predictions. Tsai recapped some of the challenges related to risk assessment of GE trees. She believes the foremost challenge is the inability to perform large-scale field experiments; few studies have been permitted outside of China. In addition, Tsai said the nature of risk-benefit assessments makes measuring commercial impacts easier as compared to ecological impacts, which must be hypothetical or extrapolated. Setting the context, as the case studies showed, is an important consideration. Tsai noted that abun- dant outcrossing already exists in nature, and setting a baseline when studying the effects of GE trees is significantâespecially whether they would be planted in natural sites or, more likely, in plantations, on idle agricultural land, or even on waste sites as phytoremediation. Finally, she noted wildlife and natural habitats are in a constant state of flux, caused by climate change, human activities, and natural disasters. These changes make it more complex to predict changes that might be caused by the introduction of GE trees. Tsai reported on the beginnings of a consensus within the forest sci- â Bt/API hybrid is a complex cross of multiple poplars that result in a tree with limited seed generation and a very poor ability to germinate in natural conditions.
CURRENT RESEARCH 15 ence community, developed at two recent conferences, about studying the impact of GE trees on the environment: â¢ Greenhouse and small-scale plantings will continue, but they are insufficient to address ecological risks. â¢ Ecological risks cannot be modeled using annual systems like corn or soybeans. â¢ The efficacy of biological confinement has not been evaluated in the field over the long term. â¢ Modeling is essential, in conjunction with field data collection. â¢ Not all traits and species are equal, so prioritization must take place. â¢ Absolute versus relative risk must be taken into account. â¢ Proxy learning can come from intensively managed systems or natural hybrids. â¢ Learning by doing is required. Research Gaps What is known and especially what is not known about the effects of GE trees on the environment led Tsai to underscore what she feels is the greatest research need: the establishment of long-term research field trials, to include monitoring beyond reproductive age. She suggested the Long-Term Ecological Research (LTER) Program of the National Science Foundation and the Free Air CO2 Enrichment (FACE) facilities of the Department of Energy (see Box 2-2) as potential models for long-term ecological studies. Setting up field trials at even pre-commercial scale is beyond the resources of what an individual academic or federal labora- tory can undertake, so Tsai urged partnerships with the private sector. Tsai believes traits, species, and sites should be prioritized so that resources are allocated for the traits and species with greatest economic or ecological relevance and so that sites represent various ecological systems. She suggested that a co-facility that brings together researchers, such as FACE, may be a way to solve funding and security challenges. In terms of regulatory limitations, she shared her belief that if conditional release beyond flowering or to the point of harvest remains unallowable, research cannot move forward. â Institute for Forest Biotechnology Symposium on Genetically Engineered Forest Trees: A Workshop to Identify Priorities for Ecological Risk Assessment, May 3-4, 2007, Raleigh, NC; Institute for Forest Biotechnology Meeting on Growing Trees and Stemming Risks: Ecologi- cal Impacts Associated with the Products and Practice of Forest Biotechnology, March 19-21, 2006, Vancouver, Canada (Tree Genetics and Genomics Special Issue, April 2007).
16 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT BOX 2-2 About FACE Free Air CO2 Enrichment (or FACE) facilities are operated by the U.S. Depart- ment of Energy (DOE) Office of Biological & Environmental Research. There are four sites in the United States: a sweet gum plantation at the DOE Oak Ridge National Laboratory, a Mojave Desert location on the DOE Nevada Test Site; a loblolly pine plantation in the Duke Forest in North Carolina, and a hardwood facil- ity, including aspen, at the USDA Forest Service Harsaw Experimental Forest in Wisconsin. A poplar site (EUROFACE) is located in central Italy. The sites facilitate research that logistically would be difficult, if not impossible, for an individual researcher or group to undertake independently. For example, more than 100 scientists from 22 institutions and 9 countries are taking advantage of the FACE facility in Wisconsin, studying the effects of increasing tropospheric ozone and carbon dioxide levels on the structure and function of northern forest ecosystems. SOURCE: Aspen FACE, 2008. Research on Effects of GE FishâRobert Devlin Since the 1980s, Robert Devlin (Fisheries and Oceans Canada) reported, more than 30 species of fish have been genetically engineered by transferring a wide range of genes related to metabolism, disease resistance, reproduction, and other purposes, with growth enhancement as the principal trait studied. The consequences of different transgenes on the phenotypes of the fish are expected to differ widely, according to Devlin. For example, over- expression of growth hormone (GH) in Atlantic salmon, carp, tilapia, and other fish species produced significantly faster growth, but other charac- teristics, such as altered endocrine profiles, reduced disease resistance, and swimming ability, have also resulted. The interplay of a wide range of fish habitats, genotypes, phenotypes, and other variables could mean a myriad of consequences of GE fish. Ani- mal behaviorâin the GH case, the fact that the transgenic fish become, as Devlin called them, âfeeding machinesââcomplicates predictions of the ecological effects. Ideally, fisheries scientists would like to have data from nature, but it is not currently allowable or desirable to release fertile GE fish into aquatic ecosystems. Devlin explained that a planned introduction of transgenic fish has been proposed to assist with elimination of feral carp in Australia, but this is still in very early stages of consideration. He noted that, to his
CURRENT RESEARCH 17 knowledge, no GE fish have even been inadvertently released into nature from aquaculture production facilities, although theoretically this could occur through storms, shipping accidents, or human error. Containment is critical because recovery of fish from most aquatic environments is essentially impossible. If GE fish and aquatic organisms escaped or were released, Devlin said there may be direct effects on non-transgenic conspecifics (fish of the same species) and other organisms in the ecosystem as a result of resource competition, altered pathogen susceptibility or transfer, and indi- rect genetic effects, among others. If the transgenics survived, there could be sustained effects, if they breed with each other or with conspecific, non-transgenic fish. The framework laid out by Kapuscinski et al. (2007) can help assess the risks to the environment of GE fish. At a high level, this would involve looking at relevant components and processes in an ecosystem (using information on biotic and abiotic functions of aquatic ecosystems); the phenotype of the transgenic fish (including those traits the transgene intentionally altered, and those traits that emerged as âside effectsâ of the transgene on fish physiology or behavior) and the full range of the types of interactions likely to occur between the fish and different components of the ecosystem. When this information is gathered, the next step in the framework is to identify the likelihood of interactions, the hypothetical consequences of those interactions, and the degree of uncertainty in pre- dicting those outcomes. Instead of studies in nature, other research approaches might be used to try to understand both the fitness of a transgenic in any number of eco- systems and the consequences of survival to that ecosystem. Therefore, models to simulate these dynamics are important tools. In addition, the individual genetic, physiological, and behavioral characteristics of trans- genics can be examined in controlled lab conditions and semi-natural environments that approximate nature. Finally, Devlin noted that non- transgenic surrogates (for example, salmon with a non-transgenic growth hormone) can be released into nature to try to observe effects. From a risk assessment point of view, the question is whether these methods can generate reliable data. According to Devlin, these approaches certainly provide valuable data, but the fact that they do not fully mirror fitness in nature, especially for larger species, is problematic in terms of providing full answers to questions about the effects of GE fish on wildlife and habitats. Thus it is necessary to conduct studies under as many different exper- imental conditions as possible. Environmental conditions (for example, different levels of food availability) and rearing conditions can strongly affect phenotypes, and although these cannot be accurately replicated
18 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT outside of nature, one valuable contribution of lab experiments is their ability to reveal phenotypic differences that may be more subtly mani- fested in nature. Devlin noted that many complex phenomena that will affect conse- quences in nature need to be untangled. For example, differences in the genetic background between strains of the same fish species (for example a wild type versus domesticated strain) will also greatly influence the resulting phenotype resulting from a transgenic transformation. Antagonistic pleiotropy, in which some traits can simultaneously pro- duce beneficial and detrimental effects on fitness, is another phenomenon in need of study. He illustrated with data showing how transgenic fish in a contained experimental system outcompeted non-transgenics for food (thus growing much faster) but experienced higher mortality than the non-transgenic when a predator was introduced into the system. To contain GE fish, researchers continue to develop physical and biological containment systems, yet more needs to be done. For example, sterilization techniques have been shown to be 99.8 percent successfulâ yet even a 0.2 percent failure rate would mean the escape of large num- bers of transgenic fish into the environment so releasing fish using these techniques is not permitted. Some combination of these methods, he sug- gested, may ultimately create more complete containment. Research Needs Devlin identified six research needs to study the effects of GE fish on the environment: â¢ The development of large, variable-environment facilities to rear and assess transgenic fish in conditions that are as close to nature as possible. â¢ Assessment of whether complicating gene-by-environment (G Ã E) interactions and antagonistic pleiotropic effects are pervasive for criti- cal fitness traits. If these effects cannot be well defined, then laboratory experiments will be able to identify some of the forces at work in predict- ing fitness, but not accurately estimate magnitudes. â¢ Integration of ecosystem models with demographic and genetic models, attempting model validation with surrogate (non-GEO) models in nature. â G Ã E signifies genotype by environment interactions, in which the growth, performance, or behavior of specific genotypes (e.g., transgene, transgenic events, or other genetic entity) is affected by specific environments.
CURRENT RESEARCH 19 â¢ Development of methods for uncertainty analysis to facilitate pre- dictions and regulatory decisions. â¢ Assessment of background genetic effects on transgene phenotype. â¢ Improvement on biological containment methods to minimize exposure of transgenic fish to ecosystems, through a combination of lay- ers of containment. Research on Effects of MicrobesâMichael Allen Microbes add complexity to the discussion of the effects of GEOs on wildlife and their habitats, asserted Michael Allen (University of Califor- nia, Riverside). Although fewer than a dozen microbial traits have been approved for release, the diversity of GE microbes in development is vast and includes those developed for plant protection, improved nutrition, metal absorption, and other functions. Rather than list all the properties described in the hundreds of papers written about them, Allen suggested that he focus his presentation on the challenges and approaches to the study of microbes. First, the dispersal of microbes cannot be contained, even with the kinds of facilities described earlier in the workshop for fish. Referring to the movement of microbes, Allen observed, âif it can happen, it will.â Microorganisms can travel long distances through events such as fire, hurricanes, and human or animal movement. However, even small dis- tances can matter, especially at the interface of developed and wildland habitat. Ecologists have long observed that problems are often associated with the introduction of any exotic species into the natural environment, transgenic or not. For example, he said the presence of exotic grasses is believed to contribute to a more frequent fire cycle in California. According to Allen, issues to think about when studying GE micro- organisms include: horizontal gene transfer between microorganisms; the persistence of a GE microbe or its gene product in the environment; direct impacts on microbial populations through the creation of genetic bottlenecks; other direct effects, such as toxicity; and indirect impacts on an ecosystem through altered food webs, community structure, and nutrient cycling. Despite a large amount of literature about gene transfer between microbes, Allen said that not much is known about the outcomes of horizontal transfer or about indirect effects of the introduction of a new microbe, such as host-species shifts. Another need is to understand the effect of acute versus chronic toxicity, especially in a field environment where microbes might persist at low levels. Moreover, Allen noted it is
20 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT BOX 2-3 NEON: New Technologies to Understand Ecological Consequences The National Ecological Observatory Network has been proposed as a conti- nental-scale research program by the National Science Foundation to understand the impacts of climate change, land-use change, and invasive species on ecology. Planning for NEON is currently under way. Twenty proposed sites, one in each eco-climatic domain in the United States, are intended to operate as a national observatory, linked by advanced cyberinfra- structure to record and archive ecological data for at least 30 years. The long-term NEON data collected at these sites are expected to support improved ecological forecasting to optimize natural resource management and provide early warning of biological natural hazards. SOURCE: National Ecological Observatory Network, 2008. difficult to design experiments that look beyond direct toxicity effects to effects that move through the entire food web. A number of studies argue for the need to study ecosystem processes, but Allen believes that the technology to look at these processes is out- moded. One good opportunity, he suggested, is to link to some of the new technologies contemplated in the National Ecological Observation Net- work (NEON; see Box 2-3). For example, integrating isotopes with sensing technology, perhaps the next big development in the field of ecology, can be used to detect very subtle differences in ecosystem functions. Research Needs Allen summarized by pointing out that society has more than one alternative when it comes to making decisions about managing, for exam- ple, agricultural pests: It can give up agriculture in affected areas, it can use or not using chemical spray pesticides, it can use or not use GEOs to enhance productivity and/or reduce pest damage, it can use or not use agents that decompose toxins, to name a few. Thus, Allen believes that studies comparing the effects of GE microbes with chemical pesticides would be helpful in informing these decisions. To understand the appropriate role of microbes as it relates to these alternatives, Allen suggested a few places to focus research efforts to improve understanding of the effects of GE microbes:
CURRENT RESEARCH 21 â¢ Use new ecological monitoring technologies. â¢ Engage in long-term studies, like those in NEON, not only for natural systems but also for agricultural and urban development. â¢ Study effects based on a single versus multiple introductions over time. â¢ Study the movement, persistence, and recombination under dif- ferent event scenarios, given that âthe unusual is the usual in long-term ecological dynamics.â â¢ Study the effects of small perturbations on threatened populations under stress. â¢ Study indirect consequences and a complex systems approach of GE microbes, looking at food webs, nutrient cycling, and community composition. Research on Effects of GE InsectsâThomas Miller Thomas Miller (University of California, Riverside) focused on GE insects to conclude the round of presentations on GEO research by taxa. He noted that a new pest or disease enters California every 60 daysâand that no insect pest species has ever been eradicated by humans. None- theless, the âtoolboxâ to combat insect pests includes the development of disease-resistant plants, biological control, cultural control, chemical control, and now, although in the earliest stages, biotechnology. From Millerâs perspective, when a new pest or disease is identified, resources are allocated to develop the tools that can solve the problem, but not to study the side effects. Miller presented two examples of the development of GE insects to solve pest problems related to agricultural production. His first example was pink bollworm (Pectinophora gossypiella), which came to the United States with cottonseed. Miller and his USDA colleagues have developed a transgenic, sterile bollworm that is currently in field trials in Arizona. The transgenic bollworm would be released as a biocontrol agent. Miller noted that the effect of the bollworm (transgenic or otherwise) as either a source of wildlife food or as a predator on wildlife is not known, but as Wolfenbarger had suggested earlier, the impact of cotton cultivation itself might be more disruptive to the environment than the presence or absence of a transgenic insect. The regulatory concerns associated with the potential release of the transgenic bollworm are related to where in the bollworm genome the transgene is inserted, if the insertion is stable, the fitness effects from the insertion, and the possibility of horizontal gene movement. Because it was a partner in this research, the USDA-APHIS is paying for the field studies for environmental assessment. Millerâs second example was an approach to controlling Pierceâs dis-
22 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT ease of grapes, which is caused by a bacterial pathogen (Xylella fastidi- osa) that is vectored by an insect, the glassy-winged sharpshooter. Miller developed a transgenic version of another bacteria (Alcaligenes) found in the insects guts that he believes would displace the Pierceâs disease pathogen in the insect, leading to reduced transmission of the disease between grapevines. Regulatory approval to bring this approach into field trials has moved very slowly, in part because of the novelty of the approach. A probability model has been developed to predict effects of the introduction of the transgenic bacteria, but the model needs data that can only be derived from field trials. For field trials, regulators required that any grapevines (onto which the bacteria would be introduced) must be destroyed. Miller said that was a nonstarter for getting the cooperation of grape growers with the research. Research Needs Miller stressed the need for field trials to get data to fill in the knowl- edge gaps and added that the study of ecological impacts requires inter- disciplinary approaches. In the case of Pierceâs disease, for example, looking at the effects of the problem and potential solutions involves understanding bacteria, pathogens, insects, grapevines, immunology, wildlife, and ecology. Miller believes partnerships with industry and government can facil- itate the process. In the case of pink bollworm, USDA identified the problem and a potential solution, and provided much-needed funding. Regulatory approval by APHIS, while still complex, has also been more smoothly coordinated. He suggested perhaps USGS and FWS could play a similar pivotal role in other GEO research. Discussion on Taxa-Specific GEO Research Needs During the question-and-answer session with the five presenters, many of the comments related to the balance between field study and regulatory requirements, as well as how to choose the most relevant top- ics to investigate. Field Study A common goal across taxa is to conduct ecologically relevant studies in field settings. A few participants suggested that one potential approach for some species might be to partner with USGS and others to identify appropriate field sites where conditions and confinement are adequate. FACE, NEON, and the LTER Network seemed to possible starting points.
CURRENT RESEARCH 23 For some species, such as fish, that canât be field tested, an important step would be to develop indicators of the processes, in terms of life history, that can be most accurately mimicked in a contained facility. When the point was raised that releases can occur despite precautions and U.S. bans, for example the use of transgenic insects in other countries, Allen suggested at least using these escapes as a learning opportunity. Along those same lines, another participant suggested making better use of currently grown GE crops to study ecological effects. âWe have grown a billion acres of GE crops and have barely begun to assess the ecologi- cal effects,â he said, âso I challenge us to more effectively determine and summarize what has already happened with the global experiment that has being going on for more than a decade.â There is a need to work with regulators to resolve the impasse, as one participant termed it, between the need for high-quality ecological infor- mation and regulatory requirements to minimize environmental risk. Is it possible, he asked rhetorically, to get to the point where we do not need complete containment so that fieldwork can take place? Another partici- pant suggested that uncertainty is part of the risk assessment process and research on uncertainty should be embraced, rather than avoided. In fact, he said, it is the rare events that may be most important. Secondary Effects and Baselines Teasing out indirect effects is tough, but essential, Allen said. It is often said that tillage, soil types, climate, and other factors override the effects of GEOs, but that does not negate the possibility that a secondary effect can be critical in the long term. Allen believes more sensitive mea- surement methods will help scientists look at these more subtle effects. A participant questioned whether a comparison at the farm-field level is the appropriate baseline, since effects may go beyond the field. In those cases, he asked what should the baseline be? Wolfenbarger replied that current GE row crops do not move into habitat on the margin, but that may be an important consideration with newer technologies. She said one approach may be a gradient that encompasses a cultivated area and the natural areas around it. Another audience member wondered who decides what impacts to look for, and whether these were always anticipated to be negative, as opposed to positive impacts? Wolfenbarger suggested that this workshop, by virtue of its sponsorship by a federal agency, was an attempt to integrate societal values about what is impor- tant into how the agency will target its resources.
24 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT Selecting Research Subjects A final point relates to choosing which processes or interactions to study. How do we pick research subjects that give us meaningful data? Sometimes the less threatening systems (the release of an organism with a âneutralâ trait) are the easiest to agree on for release, but they are also less interesting in terms of what they can show. Tsai agreed the driving force to study ecological risk should be dealing with ecologically relevant traits, looking at the long term and in multiple sites. FUNDAMENTAL AND CROSSCUTTING RESEARCH FOR ASSESSING ECOLOGICAL EFFECTS OF GEOS As committee member Steven Strauss said in introducing the next plenary session on crosscutting issues, case-by-case study when looking at GEO organisms is necessary because of the diversity of traits, organ- isms, and environments. But, at the same time, this diversity can be so overwhelming that generalities to make predictions and pool resources are also needed to move forward. Presentations in this session covered three ongoing areas of research that can complement the study of GEO effects: invasion ecology, gene flow, and detection and monitoring. Research Approaches from Invasion EcologyÂâDiane Larson Diane Larson (U.S. Geological Survey) discussed the concept of inva- siveness, the attributes of invading species and their recipient environ- ments, and approaches ecologists take to study the effects of invasions. This research suggests some parallels for research into GEOs. She first reviewed two general hypotheses of invasion ecology. The enemy-escape hypothesis states that when a non-native organism finds its way into a new environment, it is subject to reduced attack (from preda- tion, parasitism, and competition) relative to its native environment. As a result, the non-native organism can increase growth and reproduction. This hypothesis, said Larson, serves as the rationale for using biological control to reduce the growth and reproduction of invaders. A related genetic-based hypothesisâthe evolution of increased competitive abil- ityâstates that escape from enemies allows an organism to shift from making defensive compounds to putting all of its effort into growth and reproduction. Both these hypotheses imply that researchers need to look at all guilds of enemies, including microorganisms, that keep an organism in checkââthe whole system,â as Larson said, ânot just one individual plant against one individual enemy.â Potential application of these hypotheses to GEO research includes anticipating whether genetic tradeoffs will make GEOs more or less competitive in an unmanaged
CURRENT RESEARCH 25 habitat; for example, if a fitness cost associated with the Bt gene would reduce its competitive ability in an unmanaged setting. Larson said that one of the most significant factors affecting the like- lihood of the invasiveness of a species, shown time and time again, is propagule pressure: the more an organism is introduced, the more likely it is to become established and subsequently invasive. She noted that this principle is also useful in determining the optimal release size of biocon- trol organisms, and may help reveal the likelihood of a GEO encountering a compatible relative in the surrounding habitat, which might allow for gene introgression. Other species attributes studied by invasion ecologists that may be relevant to GEO effects include the speciesâ use of ânovel weaponsâ that increase its competitiveness, the speciesâ life history, the existence of mutualisms that support the survival of the organism, and its tolerance of environmental amplitude (see Table 2-1). Turning from the characteristics of an invasive species to the envi- ronment in which it is introduced, the evidence suggests that the most significant attribute to consider is disturbance. Disturbance facilitates invasion, because resources like nutrients and space are freed up and the stable interactions of native species are disrupted. Larson added that the primary phase of an invasion, when the invader is gaining momentum, may present the best time to control it (Dietz and Edwards, 2006). In a GEO context, this could suggest studying the condi- tions in which a GEO will colonize beyond the disturbed area and, thus, which areas should be monitored. A second attribute of the recipient environment is biotic resistance; in other words, interactions with the native species can prevent the spread of a newcomer, and this might be related to the degree of biodiversity present in the environment. Support for this hypothesis varies, said Lar- son, and seems to be scale-dependent. A comparison of colonization in areas with varying native diversity, which might suggest way to create buffers or barriers, is one way that study of biotic resistance could apply to research in GEOs. Effects of Invasive Species Larson explained the effects of invasive species include hybridiza- tion, which can threaten native genotypes and endangered populations and can result in either in increased or decreased vigor of the new genetic combinations. According to Larson, other effects with potential applica- tion to GEO research include changes to native community structures, interactions with mutualists, relationships with other invasive species, changes in the availability and cycling of nutrients, ecosystem engineers, and predator-predator aggression (Table 2-2).
26 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT TABLE 2-1â Research On Factors That Contribute to Invasiveness, and Their Potential Application to the Study of GEOs Potential Application Correlates with Research to Research on GEO Attribute Invasiveness Approaches Impacts Propagule pressure Frequent Monitoring of Predicting introductions, introductions; likelihood of a GEO large landscape encountering a introductions analysis compatible relative in surrounding habitat Novel weapons Allelopathy Chemical ecology, Examining effects of that improve (production of ethology, use of novel root exudates competitiveness toxic or defensive realistic habitats Development of compounds), novel behaviors new predatory behaviors Life history Early sexual Matrix models Potential effect of maturity, short of life/death habitat (managed generation time, events to estimate and unmanaged) on rapid growth, survival GEOâs demographics high reproductive capacity Mutualism Assisted by Observations; Likelihood of pollinators and manipulative field dispersal of pollen seed dispersing and greenhouse or of plant to other organisms experiments areas; survival in (e.g. birds), unmanaged areas mycorrhizae Environmental Thermal Climate Potential spread of amplitude tolerance, drought envelope/ organisms modified tolerance modeling for increased environmental tolerance SOURCE: D. Larson. Larson closed with what she termed ânagging odds and endsâ from invasion ecology that might apply to GEOs: â¢ The lag times sometimes seen between the first introduction of a non-native species and their invasive effects can be long, even hundreds of years. â¢ Invasions can be cryptic.
CURRENT RESEARCH 27 TABLE 2-2â Research of Effects of Invasions and Potential Parallels to GEOs Research Potential Application Effect Impacts Approaches to GEO Research Hybridization New genetic Modeling, Spread of gene combinations experimental via hybridization, that increase or crosses in leading to decrease fitness controlled settings persistence or loss of a native of gene population Changes to native Competition, Controlled Changes in trophic community apparent field and pot/ interactions, such structure competition, mesocosm as rapid growth, predation experiments, food replacement of or interaction web predators, or analysis herbivore resistance Interactions with Pollen quality or Manipulative or Effect of GE pollen, mutualists quantity effects observational pollen dispersal on native plants, field studies, to native relatives, parasitism of comparison of potential effects fungi mycorrhizal of root exudates colonization on soil mutualists, horizontal gene transfer Relationships Invasion cascades Observational Potential interactions with other or meltdowns field and lab/ that would facilitate invasive species greenhouse invasion by the other studies species Change in Changes in Nutrient Effects of GEOs availability/ litter quality or manipulation; on litter; changes cycling of quantity; changes observational in environmental nutrients in detritivore nutrient availability community or cycling Ecosystem Creation, Observational Potential effects, but engineering modification, studies, modeling also potential utility maintenance, or in restoration destruction of habitat Predator-predator One invader Observational; Potential for a aggression against another realistic habitat rapidly growing variation GEO to influence native species SOURCE: D. Larson
28 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT â¢ Baseline information about the environment pre-invasion is impor- tant to know. â¢ Monitoring should include defining an objective threshold for action: at what point will the changes revealed through monitoring mean some action is taken? Working from the Gene and Organism and Moving UpwardâPaul Gepts As Paul Gepts (University of California, Davis) said when he intro- duced the title of his presentation, it is the âupwardâ part of the topic, beyond the organism, as the process of gene flow takes place, which is especially challenging. Variables in an organism, gene or trait, and the environment all affect the process of gene flow (see Figure 2-1). Looking at a myriad of individual cases, Gepts asserted that gene flow will take place, but that flow varies greatly by organism: a corn plant, for example, has millions of pollen grains per plant but they are wind dis- persed, as compared to a soybean plant, which has only a few thousand grains but which are insect-pollinated. Local events are most frequent, he said, but long-distance gene flow, harder to measure, also occurs. Gepts Gene or Trait -Source -Expression -Selective value -Genome location -Other Organism Environment -Plant or animal -Topography -Propagule and -Climate gamete dispersal -Wild or domesticated -Life cycle relatives -Reproduction -Pests & diseases -Other -Other FIGURE 2-1â Variables that may affect gene flow and persistence in the environment. 2-1 latest
CURRENT RESEARCH 29 described several current studies that are combining modeling with pol- len capture in the atmosphere to document long-distance gene flow. Gepts explained that variables at the gene or trait level depend on the source of the gene. In the case of a GEO, that might include another species from which the gene was taken. Gene expression, selective value, and location in the genome also affect the likelihood of the gene to move into a population and to be expressed, keeping in mind that expression is dependent on G Ã E interactions. Finally, said Gepts, environmental factors affect gene flow. Through- out the world, areas that were once centers of crop domestication have wild and domesticated relatives in proximity to each other; even in the United States and Europe, where few agricultural crops originated, crops with close wild relatives can be found. Gepts explained that a number of factors will determine gene flow and whether a gene (including a trans- gene) escapes from a domesticated to a wild plant via pollen. Whether the escape results in ultimate establishment in a different genetic background, however, is in the realm of population genetics: the level of migration between the bred and wild variety will depend on the size and diversity of the populations, inheritance characteristics, and migration. Research Needs to Study GEO Gene Flow Gepts concluded by identifying issues related to organisms, genes or traits, and the environment that would yield important information for GEOs: â¢ Organism: The dispersal ability of gametes and propagules. â¢ Gene or trait: The influence of genome location on expression. (The ability to target the location of a gene insertion in a genome would be of tremendous benefit in this regard.) â¢ Environment: The fate of transgenes and their products through a variety of methods. At present, the many factors that affect gene flow can be listed and described separately, Gepts said. But what is needed is a way to look at them in combination through development of a multifactorial, quantita- tive, integrative risk factor. Strategies for Effective Detection and MonitoringâMichelle Marvier As the third in the series of presentations on crosscutting research, Michelle Marvier (Santa Clara University, California) addressed the ques- tion of which approaches will be most useful to distinguish the ecological
30 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT impacts of GEOs against the backdrop of many other potential sources of environmental change. She explained that many confounding factors can affect wildlife and habitat, and it is difficult to tease out a specific cause why changes occur. The strategic selection of indicators, combined with spatial and temporal comparisons, can help provide answers. The key to effective monitoring, she stated, is selecting indicators that are sensitive to specific environmental changes. Meta-analysis, modeling, and literature reviews to discover other variables of ecological effects can help to identify sensitive indicators. Marvier presented a meta-analysis of more than 150 publications and unpublished reports that looked at nontarget effects of Bt crops (Marvier et al., 2007), noting meta-analyses could be useful for other topics unre- lated to transgenic effects, such as the effects of pesticides or other agri- cultural practices. She suggested that the searchable database created for the Bt meta-analysis could also be used to look at other aspects of the effect of GEOs on wildlife and habitats. A second way to select good indicators is through modeling. Marvier explained that models can identify life history traits and species that are better indicators than others; different stressors that may slow population growth, including GEOs, can then be run through the model. Simulations show a predator species is generally a more revealing indicator to monitor than a prey species. Similarly, a fast-growing species is a better indicator since differences will show up more rapidly. Thus, practical guidance from modeling suggests selecting indicator species with a higher tro- phic position, high rate of population growth, and low environmental sensitivity. A third approach is to look at the literature to consider other variables that integrate many ecological effects but do not rely on precise abun- dance estimates. Marvier provided one example measuring the mean tro- phic level of fish caught in a marine ecosystem as a proxy for measuring fishing pressure, given that a precise count of all the fish in an ecosystem is impossible. In addition to indicator species selection, monitoring requires good contrasts as focal points of study; a location where GEOs have been released contrasted with a location where they do not exist, or an environ- ment before and after the releases. She said this information is difficult to find, especially at a county level, in part because GEO releases are often kept confidential. Some researchers have developed relationships with farmers and can obtain planting information, but this is on a case-by- case basis. Marvier said maps are critical to providing spatial contrasts for GEO presence and abundance, as well as a historical perspective for â Available at: http://delphi.nceas.ucsb.edu/Btcrops/main/search
CURRENT RESEARCH 31 temporal comparisons. Local initiatives, such as the bans on GE crops in Mendocino County, California, and in the state of Vermont, can be seen as opportunities, as these areas can serve as a basis of spatial comparison. She suggested the USGSâ mapping capabilities could be used to contribute to understanding the ecosystem effects of GEOs, with precautions taken in recognition of the privacy issues associated with the data. Discussion on Crosscutting and Fundamental Research Needs During this question-and-answer session, much of the discussion centered on the feasibility of mapping because of the availability (or lack thereof) of information. Several participants pointed out that the location of large-scale releases can be determined and, thus, mapped, because of the need for an Experimental-Use Permit (EUP) for areas 10 acres and more. Marvier said smaller field trials, such as those looking at phar- maceuticals, are harder to find out and may involve traits of particular concern. Even if the information can only be made available after the trial has taken place, she urged disseminating it for monitoring purposes. Another participant suggested it is the larger releases that will yield the most important information about scale and, in the long term, the most data. A lot of information about those releases is available, but requires digging state by state. Challenges to Mapping The principal challenge to reliable mapping of GEO releases is the issue of privacy, suggested several participants. County-level data are often not available because they could be tracked to private landowners. It was suggested that a group like this workshop could stimulate the USGS or another agency to put together publicly available databases; however, privacy issues, which are of concern in Congress, make col- lection of data on private property a sensitive issue for USGS and other agencies, with no simple solution. Another participant warned about the pitfalls of relying on a map without benefit of knowing the full context, such as the confounding factors behind the decisions about whether or not to use GEOs and more detailed information about farming practices in the area. Negative Connotations Another participant felt the discussion had a pejorative tone. Land use changes over time: A farmscape may have been a forest 200 years ago and may be planted with a totally different crop in the future. Do we
32 GENETICALLY ENGINEERED ORGANISMS, WILDLIFE, AND HABITAT engage in studies to look for the negative impacts of those developments? He took issue with the language used in the session; for example, looking for indicators as warnings for negative impacts of GEOs, when, in fact, the effect may be neutral or positive. Endpoints It is hard to monitor without thinking about endpoints, suggested a participant. When designing a study to look at either pre-release risk assessment or post-release monitoring of transgenic fish, for example, it is hard to come up with indicators of impacts as the focus moves from the initial entry of an organism to its spread further into the environment. Different situations require direct measurements or indicators. The speakers responded about some endpoints to look at from their research perspectives. Gepts suggested monitoring the wild relatives of crops, such as through GIS surveys, to see how they fare in the pres- ence of transgenic crops. Larson said a matrix model may help predict potential invasive effects before a release is made, particularly in the area outside of a crop field. Marvier suggested looking at non-target effects, moving beyond the local release environment and the plant that has been manipulated. Potential Bias One participant expressed concern about potential biases in a meta- analysis if it relies on the available published literature. Marvier said she went beyond the published literature by using the Freedom of Infor- mation Act to get studies submitted by industry to the government for regulatory approvals. Predictive Power Larson was questioned about the ability to predict the invasiveness of a species that has not been previously introduced even if there is pre- invasion information about the species and the environment (but not the interaction); in fish, it is believed to be only about 70 percent. Larson said she did not think the percentage was any higher in plants. Thus, she said, invasive species studies do not provide the power of prediction that some regulators of GEOs might like, if it can be assumed that GEOs are likely to behave in the same wayâan assumption that is subject to debate.