Commercial strains of Bt used in pest and vector control, such as Btk and Bti, are all sibling species of B. cereus. Years after many products employing these strains were registered as the active ingredients of commercial insecticides, B. cereus was shown to be a relatively minor cause of food poisoning events in humans — the poisoning due to protein toxins produced during vegetative growth. The only consistent phenotypic difference between B. cereus and Bt is that the latter species produces protein parasporal bodies during sporulation.9,10 The close relationship of these two species raised concerns by some investigators in northern Europe that Bt, rather than B. cereus, may be the cause of some occasional outbreaks of food poisoning in humans. Food poisoning caused by B. cereus is due to two types of toxins: emetic toxins, which cause vomiting, and enterotoxins, which cause gastrointestinal discomfort that often leads to diarrhea.42 This raises three key questions: First, do the commercial strains of Bt encode and produce these toxins? Second, and more important, does the use of commercial Bt strains in forestry, agriculture, and vector control actually cause episodes of food poisoning? And third, if Bt products cause food poisoning, to what extent do they cause food poisoning outbreaks?
Several studies carried out over the past two decades demonstrate that the commercial strains of Bt do not encode emetic toxins but do contain genes for entero-toxins, and are capable of producing these during vegetative growth. Most available evidence, however, indicates that only low levels, if any, of these toxins are present in commercial products. This is because the supernatant that may contain these toxins is discarded or, if present, they degrade during the formulation process. Moreover, though there is reasonably good evidence that Bt has the potential to cause food poisoning via these gene products, there is no evidence that commercial or naturally occurring strains have ever caused food poisoning, even though, as the literature demonstrates, thousands of viable cells can occur in food products. For one thing, diarrhea caused by enterotoxins is caused by a combination of several gene products, and not all of these genes are present in commercial Bt strains. But even if Bt does cause mild food poisoning, these events are very rare and are more likely due to strains that originate from natural sources, such as grain and grain dust, rather than from commercial insecticides. The safety of Bt remains a somewhat controversial issue, at least in some quarters, and influences how we consider Bt crops. Therefore, to support the above overview, we review here the key literature on the B. cereus toxins that cause food poisoning, along with similar studies of Bt. We conclude this section with an assessment showing that it is unlikely Bt strains from commercial products are the cause of any outbreaks of food poisoning in humans.
22.214.171.124.1 Food Poisoning by the Emetic Toxin of B. cereus
The B. cereus emetic toxin, which induces vomiting, is a cyclic peptide known as cereulide.42-44 This peptide is denatured by digestive enzymes and onset of symptoms is normally observed soon (0.5 to 5 hours) after ingestion of contaminated food.45 This indicates that cereulide must be present at an elevated concentration (105 to 108 cells g-1) at the time of ingestion to produce both emetic and enterotoxicity.42 It has been speculated that when emetic and diarrheal symptoms occur together, it is because spores were ingested along with preformed emetic toxin.46 However, recent surveys of B. cereus group strains found no evidence that Bt strains isolated from fresh fruits and vegetables and other food sources in Danish markets (see Section 126.96.36.199.4), or commercial strains of Bt, contained the gene responsible for production of cereulide.4748 Thus, there is no evidence that Bt strains, be they from food sources or from commercial products, are the cause of food poisonings induced by cereulide and characterized primarily by vomiting.
188.8.131.52.2 Food Poisoning by Enterotoxins of B. cereus
The first case of food poisoning by B. cereus was reported in 1950, following consumption of vanilla sauce containing 3 x 107 to 108 cells of this species per milliliter.49 The symptoms occurred 10 hours after ingestion and included abdominal pain, watery diarrhea, and moderate nausea not accompanied by vomiting.49 To confirm the cause, Hauge inoculated sterile vanilla sauce with B. cereus and consumed it; diarrhea ensued after 13 hours. However, subsequent feeding experiments with B. cereus administered to human volunteers were unsuccessful in reproducing these findings.50 The mechanism of pathogenesis remained unknown for 20 years, and even now is poorly under-stood.5l Food sources determined to be most likely to cause food poisoning as a result of B. cereus contamination are those that are heated and then allowed to cool and stand prior to ingestion.42 Bacteria in the genus Bacillus, such as B. cereus, can sporulate and survive adverse conditions, such as heating or nutrient limitation, that often kill other types of bacteria. When conditions favorable to vegetative growth return, the spores germinate and the cells continue to multiply. The pasteurization of dairy products, for example, and certain other food processes produces an environment that facilitates vegetative growth of B. cereus. This has led to increased vigilance in surveillance of dairy products for contamination by B. cereus.42,52 As a result of these studies, it is now accepted that B. cereus can produce toxins capable of causing food poisoning under favorable conditions, and that this poisoning is due primarily to enterotoxins.
There are two major types of enterotoxins capable of being produced by species of the B. cereus group that cause food poisoning: hemolysin BL (HBL, a hemolytic toxin), and a nonhemolytic toxin (NHE). Each of these toxins consists of three proteins and all three are required for each toxin to produce gastrointestinal illness, which is typically characterized by mild diarrhea.45,52-54 Unlike cereulide, it is generally thought that gastrointestinal illness produced by HBL and NHE results from production of these toxins after ingestion of spores and initiation of vegetative growth. Other types of toxins also exist that act in the small intestine, such as enterotoxin T and cytotoxin K (CytK); this type of toxin is rare and has only been reported in a single case of food poisoning. Although diarrhea caused by enterotoxins can be solely due to the activity of HBL and/or NHE, each or both of these may work in concert with phos-pholipase C, sphingomyelinase, and/or proteases to produce diarrhea.46
There is a high degree of similarity between the HBL complex and the nonhe-molytic NHE complex. The HBL enterotoxin component consists of three proteins: B, Ll, and L2.42 All three are necessary to obtain full enterotoxin activity, although binary combinations can have some biological activity. The HBL complex is thought to be the primary virulence factor in B. cereus diarrhea. Some strains of B. cereus produce both HBL and NHE enterotoxin complexes, whereas other strains produce only one, and some none.42 The proteins have been characterized and the genes have been sequenced for each complex, as well as the enterotoxin T.55 With regard to enterotoxin T, the specific molecular mechanisms that produce illness remain largely unknown, but it is hypothesized that this toxin stimulates the adenylate-cyclase-cyclic-AMP system in the intestinal epithelial cells, thereby causing fluid accumulation leading to diarrhea.54 Since this enterotoxin is susceptible to low pH and proteolytic enzymes, it is unlikely to survive digestion in the stomach.45,53 It is therefore speculated that enterotoxin is produced following ingestion of a high dose of vegetative cells or spores (105 to 107), resulting in abdominal pain, watery diarrhea, and occasional nausea 8-16 hours later.42
Enterotoxin can be detected by a variety of methods, including polymerase chain reaction (PCR) assay and enzyme-linked immunosorbent assay (ELISA).
PCR primers were used to detect the different genes coding for these proteins in 22 B. cereus and 41 B. thuringiensis strains.55 The results demonstrated that all 41 B. thuringiensis strains contained at least one gene coding for either of the two protein complexes. This was also true of most of the B. cereus strains, though six of these did not have the genes to produce the HBL complex. Moreover, a significant correlation was found between the presence of a gene and the presence of other genes within the same enterotoxin complex.55 This is significant since the two commercially available immunoassay kits commonly used to detect the presence of B. cereus in food rely on detection of one protein from either the HBL or the NHE complex. The Oxoid (ELISA) test detects the L2 protein of the HBL complex that is cytotoxic, and the Tecra (BCET-RPLA) test detects one or two nontoxic proteins associated with the NHE complex.56 Therefore, a positive detection with both kits suggests that enterotoxin-producing bacilli are present and that the bacteria are likely producing all components of each enterotoxin complex. It is unclear, however, to what extent (or even if) enterotoxin T is responsible for food poisoning.42
In addition to the Tecra and Oxoid assays, several others used to determine cyto-toxicity have been proposed as a means of evaluating the activity of enterotoxin proteins produced by B. cereus and related species. The inhibition of 14 C-leucine uptake in Vero cells is characteristic of cytotoxicity and is generally observed with food-poisoning strains of B. cereus.51 The presence and activity of enterotoxin has also been measured using tetrazolium salt MTT, as it adversely affects the metabolic status of cultured CHO cells.58 In all of these tests, the strains are grown in brain-heart infusion media supplemented with 1% glucose at approximately 32°C. The cultures are grown to late exponential phase and the culture supernatant is then examined for toxicity. Since the conditions of the test are designed to maximize production of enterotoxin, these tests are an effective means of evaluating enterotoxin-producing potential but may not reflect the ability to produce illness in humans.
The phenotypic similarities of Bt and B. cereus and the significant overlap of their genomic characteristics suggest that under appropriate conditions for spore germination and vegetative growth, Bt could also produce enterotoxins similar to those of B. cereus. By screening soil isolates of Bt using commercial test kits for enterotoxin production, it was shown that 83% of new isolates tested positive for enterotoxin production.59 In another study, Bt strains were screened for enterotoxin genes using PCR, and for potential enterotoxicity by testing culture fluid for cytotoxicity.51 Six strains of Btk (H 3a3b3c) were analyzed, and five were determined to contain genes coding for enterotoxin T, the HBL complex, and the NHE complex. The superna-tants from cultures of these strains were all highly toxic to Vero cells, with the level of toxicity being similar to B. cereus strains thought to be responsible for outbreaks of food poisoning.51 In other studies, commercial strains of Btk along with B. cereus strain F4433/13 were evaluated with the Tecra NHE enterotoxin test kit.60 When grown on the media specified for this test, all commercial strains containing viable spores were determined to produce enterotoxins. Later, using commercially available test kits, these results were confirmed for commercial strains of Btk by Valent Biosciences (Libertyville, IL), a major producer of commercial Bts. However, whole beers from the fermenters used for commercial Bt production all tested negative for enterotoxin production.61
Overall, these studies indicate that under appropriate conditions, including specific media, most Bt strains can produce enterotoxins during vegetative growth. However, and importantly, despite the large quantities used in agriculture and forestry there is no known case where commercial use of Bt has been implicated in a food poisoning event. Because commercial Bt insecticides are used on food, especially vegetable and fruit crops, a slight possibility exists that enterotoxins could be produced under conditions favorable for spore germination and vegetative growth, and perhaps in quantities capable of causing food poisoning in humans. However, normal food-handling precautions make this unlikely to occur.
It is understood that expression of the requisite enterotoxin genes and production of an enterotoxic protein is a precursor to food poisoning. What remains unclear is whether, and under what circumstance, B. cereus or B. thuringiensis spores or vegetative cells necessarily lead to a host response, in this case diarrhea. Our ability to determine this is limited to evaluation using in vivo test systems. The ligated ileal loop assay in rabbits or mice has been demonstrated in multiple studies to be an effective test for determining the presence of enterotoxin-producing bacteria capable of inducing diarrheal-type food poisoning.46,51,62,63 In this assay, the sample (either culture supernatant or other material such as spores or vegetative cells) is injected into a ligated portion of the lower intestine and scored according to the quantitative degree of fluid accumulation that distends the intestine in comparison to controls. Fluid accumulation is indicative of a positive response. The diarrhetic toxin also alters the permeability of blood vessels when injected into the skin of rabbits. The vascular permeability reaction (VPR) correlates strongly with the rabbit ileal loop test.63,64 The B. cereus enterotoxin produces a positive response in both of these assay systems.63,64 Although these tests are potentially more effective than the in vitro studies, the most conclusive way to identify an enterotoxin is to study its effect when administered to humans or animals.51
The minimal dose necessary to produce diarrhea in humans has been estimated to range between 105 and 107 cells based on food poisonings where B. cereus has been isolated as a potentially causal agent.42 It has been speculated that levels of B. cereus in food of as low as 103 cfu/g would be considered "safe" for human consumption.45 Determining a maximum safe dose is further complicated because the concentration of enterotoxin produced by B. cereus strains varies by a factor of more than 100, with only high-enterotoxin-producing strains implicated as potentially causing food poisoning.53 Ingestion by test animals of much higher doses of B. cereus or Bt cells has been tolerated without incident. For example, when high doses of enterotoxin-producing Bt strains were administered to rats over a period of three weeks, there were no detectable effects, although the authors concluded that rodents may not be a sensitive test organism for investigating the potential for food poisoning with Bt.65 However, no evidence was presented that the amount of enterotoxin produced would be sufficient to cause human illness.
In a similar study, rats were challenged for four days with either irradiated spores, untreated spores, heat-activated spores and vegetative cells from either a B. cereus strain that produced high amounts of enterotoxin, or one of two strains of Bt used in production of commercial products.66 Few vegetative cells were found in fecal and intestinal samples, indicating that bacterial multiplication was minimal. High concentrations of untreated or heat-activated spores were detected up to two weeks following dosing, confirming that spore germination and subsequent vegetative stage multiplication was minimal or did not occur. None of the rats demonstrated signs of food poisoning or toxicity, which may indicate that rats may have low sensitivity to enterotoxins or simply that none were produced in rats by the isolates tested.65
Other test subjects, including humans, have only sporadically demonstrated symptoms of food poisoning following challenge with B. cereus or Bt. For example, ingestion of food artificially contaminated with B. thuringiensis var. galleriae at concentrations of 105 to 109 cells/g induced nausea, vomiting, diarrhea, and coliclike pains in the abdomen, as well as fever in three of four volunteers within eight hours. The Bt culture used in this study was not a commercial variety and the effects observed were potentially due to ingestion of exotoxin rather than production of enterotoxin.67 Feeding studies with monkeys confirmed the efficacy of the rabbit ileal loop studies in detection of diarrhea-producing strains of B. cereus, but these studies yielded mixed results, with some strains unable to induce a response in mon-keys.62,63 In an earlier study with a Bt preparation, B. thuringiensis subsp. thuringi-ensis, ingestion of 3 x 109 spores daily for 5 days produced no ill effects in 18 human volunteers.68 Five of these individuals also inhaled 100 mg of the Bt preparation for 5 days while receiving the dietary dose.68 In another study, male sheep were administered one of the following treatments for a five-month period: Dipel, Thuricide (both of which contain Btk), Thuricide carrier, or diet. The two bacterial insecticides were fed at the rate of approximately 1 x 1012 spores per day, for a cumulative load of 1.5 x 1014 spores.69 Two of the sheep receiving Dipel experienced illness during the second week, which continued through week 3. During the 16th week after administration, one sheep developed indigestion. One sheep receiving Thuricide developed indigestion on the eighth week of study and returned to normal on the ninth week. Intermittent or occasional loose stools were reported throughout the study for the Thuricide group. The researchers reported that the occasional loose stools and indigestion did not affect the health of the sheep and were most likely caused by the carrier or the observed change in the bacterial content of the rumen.69 In acute toxicity studies, rabbits were orally administered 2 x 109 spores per animal and suffered no ill effects.37 Monkeys administered Btk as either vegetative cells (1.2 x 109 cfu) or spores (1.4 x 109) suffered no diarrhea, other symptoms, or loss of appetite.70
A potential explanation for the mostly negative effects of Bt feeding is that the potential to produce enterotoxins does not mean the genes are expressed when ingested. Additionally, since the molecular basis for the toxin interaction producing a diarrheal response is unknown, it is uncertain whether the mere presence of enterotoxins is sufficient to produce food poisoning, or if other precursor proteins are necessary. A number of studies indicate that enterotoxin production in culture is promoted through availability of starch and under conditions of optimal pH and temperature but, as noted previously, these conditions might not exist in humans or animals.42,71-73 Therefore, although viable spores of Btk produced detectable entero-toxin in commercial assays and Btk was characterized as cytotoxic based on results of assays with Vero cells, feeding caused no illness. When Btk was assessed with a rabbit ileal loop test, the results were negative. Neither vegetative Bt cells, spores, nor enterotoxin extracts from culture medium elicited a response in more than two of seven animals tested, whereas in contrast, both B. cereus (4433) and cholera bacillus enterotoxin (CT) produced positive responses while physiological saline was negative. The Bt spores and cells from this culture were fed to monkeys and no effect was observed.70 The difference between observed enterotoxin-producing ability, as assessed with test kit bioassays and Vero cells, versus the lack of any response in rabbit ileal loop tests and primate feeding studies, suggests that there may be a fundamental difference between the toxin produced by B. cereus and Bt strains shown to produce a diarrhea, which mitigates the effect of possible in vivo enterotoxin production by Bt.
Most data available suggest that Bt has the capability, under appropriate conditions, of producing enterotoxins. However, the information available suggests that ingestion of foods treated with commercial Bt products does not constitute a food poisoning threat to humans. There is only one reported incident where Bt has been implicated as being potentially responsible for a case of gastroenteritis. In a reported food poisoning outbreak, Bt was isolated from the stool samples of four ill individu-als.74 One of these patients also tested positive for Norwalk virus, a known enteric pathogen. Since no other enteric pathogen was detected in three of the ill individuals testing positive for Bt, it was concluded that this bacterium could not be ruled out as a causative agent of the food poisoning. However, the symptoms of the ill patients (nausea, vomiting, and watery diarrhea) were more consistent with Norwalk virus than with Bt enterotoxin. Because Norwalk virus is substantially more virulent that Bt and was known to be present, the virus is most likely the cause of this event. Methods currently used for routine detection of Norwalk-like viruses (NLVs) in feces are based on electron microscopy. In order to achieve detection, at least 1 million virus particles per gram of feces need to be present, and only fecal samples obtained within 48 hours of the onset of symptoms are suitable for examination. A recently developed PCR test is more sensitive than electron microscopy and is able to detect NLVs in vomit and in feces up to seven days after the onset of symptoms, but this test was not available at the time. However, it is unclear how long after the outbreak stool samples were collected, and it is also unclear what test was used for identification of Norwalk virus in the outbreak described. The Bt isolates were determined to have some cytotoxicity but the link between this observed cytotoxicity in culture and the food poisoning outbreak event is insufficient to deduce causality. Furthermore, although Bt was isolated from food samples at the nursing home where the outbreak occurred, these isolates differed from the Bt recovered from the stool samples.
In virtually all of the food poisoning cases caused by bacilli, B. cereus is usually identified as the cause. It has been estimated that B. cereus may be responsible for as much as 47% of food poisoning caused by bacteria in some northern-hemisphere countries.74 The basis for this estimate is two-fold. First, it is understood that B. cereus is ubiquitous in nature, and surveys of foods have often found this and other species present.51,75,76 Second, because symptoms of B. cereus food poisoning are also relatively mild and transient, individuals potentially suffering from B. cereus food poisoning are unlikely to be hospitalized or contact a physician, resulting in a case that may not be correctly characterized or captured in public health statistics.42 Additionally, in order to confirm that food poisoning has been caused by B. cereus or other bacilli, stool samples need to be collected from afflicted individuals to show the presence of the bacteria in the absence of other pathogens capable of producing the same or similar symptoms. Without this information it is impossible to quantitatively assess the relative impact of B. cereus as a source of food poisoning, which has led several researchers to conclude that B. cereus food poisoning is underreported. Other researchers have suggested that some of these cases were, because of mis-identification, actually caused by Bt; the phenotypic characteristics of Bt and B. cereus are similar and the isolates may have not been examined for the presence of insecticidal crystals.42 In fact, even with positive identification of Bt or B. cereus, it is often difficult to rule out other causative agents. Thus, it is not even clear to what extent B. cereus is a major source of food poisoning, let alone Bt possibly misidenti-fied as B. cereus.
The lack of reports provides additional evidence that any food poisonings due to B. cereus are a relatively insignificant public health problem globally. Were they significant, greater attention would be applied to diagnosis and a consequent statistical analysis of the degree of importance. In more serious incidents or outbreaks where the etiology of the pathogen is effectively characterized, strains of B. cereus have rarely been implicated. Three diseases — norovirus infections, campylobacte-riosis, and salmonellosis — account for 70% of cases of known etiology transmitted by food.77 In England and Wales, six pathogens are responsible for 93% of cases of known etiology: nontyphoidal Salmonella, Campylobacter, Yersinia, C. perfringens, non-VTEC E. coli, and norovirus.78 Food poisonings associated with B. cereus have been most commonly reported in The Netherlands and Norway, where Salmonella and Campylobacter species are not prevalent and where foodborne illness has been a focus of research by food-control authorities.42 Despite the continued and ubiquitous prevalence of B. cereus strains in the environment and the long-term and continued global use of Bt insecticides, the number of individuals at risk of mortality or even any long-tem health effects due to exposure to these species is virtually nil.79-82
A summary of the data in literature through 2004 provides overwhelming evidence that Bt strains, especially strains from commercial bacterial insecticides, are not a cause of food poisoning in humans. Nevertheless, two recent studies from Denmark again raise the issue that Bt strains might be a cause of food poisoning, and thus these are worthy of a critical review.47,48 In essence, both provide substantial evidence that the proper use of Bt as an insecticide is not the cause of any outbreaks of food poisoning. In these studies, fresh fruits, vegetables, and various food products available in Danish markets were examined for levels of B. cereus-like bacteria, including Bt. In a study that focused on fresh fruits and vegetables,48 good evidence was provided that 23 out of 128 (17.9%) B. cereus-like strains isolated primarily from cucumbers, peppers, tomatoes, and lettuce (17 strains) probably originated from commercial application of Bt-based insecticides (8 from Btk-based insecticides, 9 from Bta-based insecticides). The other strains isolated were either non-Bt insecticide strains (27, or 21%), or non-Bt strains of B. cereus-like bacteria (78, or 60.9%). Levels of viable non-Bt bacteria on these crops were not provided, but the levels of Btk-like and Bta-like strains on some of the cucumber, tomato, and pepper samples were in the range of 104 cfu/g, a level consistent with application rates. These results are not surprising, as Bt-based insecticides are used in Europe (the source of most of these crops) to control lepidopteran pests and are exempted from residue requirements due to their long history of safety to humans. No evidence was provided, nor were there any implications in this study, that any of the strains (Bt or non-Bt B. cereus-like strains) were involved in any cases or outbreaks of food poisoning.
The second study focused on B. cereus and Bt strains in ready-to-eat foods, and provided even greater evidence that the strains originating from Bt insecticides were not the cause of any food poisoning.47 The ready-to-eat foods included everything from fresh fruits and vegetables, prepared foods such as sausage, bread, pasta, soups, and sauces, to various desserts, including a Danish dessert called ris a la mande (basically, a type of rice pudding) composed of rice boiled in milk to which almonds and whipped cream are added. In a sample of 40 B. cereus-like isolates from these foods selected for more detailed identification, 28 (70%) of these had characteristics of Bt (i.e., they either produced parasporal crystals and/or contained cryl genes, as determined by PCR). However, an even more detailed analysis indicated that only 10 (35.7%) of these strains produced crystals and were positive for cryl genes — characteristics that any isolate originating from a Bt insecticide would possess. Of these 10 isolates, 4 were from, respectively, raw sausage, pasta, bread, and honey — foods or food sources not normally treated with Bt insecticides. Thus, only 6 (15%) of the original 40 isolates selected for more detailed taxonomic analysis could have possibly had their origin from Bt bacterial insecticides. Moreover, these six isolates were from red pepper (2), cauliflower (1), leeks (1), salad (1), and figs (1) — none of which is typically associated with food poisoning caused by B. cereus group species. This study thus adds to the large body of strong evidence that Bt strains used in bacterial insecticides are highly unlikely to be the cause of any food poisoning events.
One could use the data in this study to even argue that B. cereus strains are only rarely, if ever, likely to cause food poisoning if food is treated properly. For example, the foods containing the highest levels of B. cereus-like organisms were vegetables, mainly cucumbers and tomatoes, and desserts made with milk, rice, flour, and custard (all > 104 cfu/g). Despite these amounts of viable B. cereus-like bacteria on or in these foods, this study mentions no cases or outbreaks of food poisoning associated with consumption of these foods. One would expect that if these amounts of B. cereus-like strains present a significant problem, given the popularity of these foods, cases of food poisoning would be rather common — but apparently they are not.
Was this article helpful?