Use of molecular typing to investigate bacterial translocation from the intestinal tract of chlorpyrifos-exposed rats

Although molecular typing has often been used to characterized BT in a clinical setting [24, 25], there are few literature data on the use of these techniques to detect translocation from the intestinal tract in animal models [26]. To the best of our knowledge, the present study is the first to have employed molecular typing (ERIC2-PCR and RAPD-PCR) techniques in the detection of BT in the rat. We observed BT during critical life periods in 22.5 % of the studied CPF-exposed rats. In studies of rats with ascites and carbon-tetrachloride-induced cirrhosis, the prevalence of BT ranged from 36.8 to 97.6 % [27‐30]. In a clinical study, positive MLN cultures were found for 32.1 % of patients with cirrhosis [31].

These findings indicate that the prevalence of BT depends on the type of study and the detection methods used (bacteriological culture of MLN samples). It has been reported that bacterial DNA in the biological fluids is a marker for BT in rats with cirrhosis [28, 31]; this suggests that the definition of BT has been widened to the passage of bacterial fragments (endotoxins, bacterial DNA, etc.) from the intestinal lumen to the MLNs and into the circulatory system in general. Although they came from the same species, translocated strains and non-translocated strains differed clearly in terms of the ERIC2-PCR and RAPD-PCR patterns (77.5 % of strains). As reported in the literature [26], we concluded that BT did not occur in samples when the extra-intestinal organs (liver, spleen, Peyer’s patches, MLNs, adipose tissue, and kidneys) were sterile. One could argue that the intestinal bacteria found in the extra-intestinal tissue samples resulted from bacterial contamination and not from BT. Table 3 shows that 41.5 % (22 out of 53) of the isolates in the extra-intestinal tissues were not found in intestinal tissues. Conversely, 58.5 % (31 out of 53) of the isolates in intestinal tissues were absent from extra-intestinal samples. This discrepancy might have been due to slight differences between the intestinal bacterial counts in rats with BT vs. rats without BT. Our results contradicts the literature data on BT in animal models, in which bacterial overgrowth was identified as a major BT-promoting factor [32]. In rats, anaerobic bacteria predominate in the ileum, caecum and colon. This anaerobic flora rarely translocates, and thus limits the colonization and growth of potentially translocating bacteria. The intestinal mucosa thus constitutes a barrier against BT via the transcellular route (through enterocytes) or the paracellular route (at tight junctions). Studies in animals have shown that some antibiotics (such as norfloxacin and trimethoprim–sulfamethoxazole) selectively reduce intestinal bacterial overgrowth (and thus BT) by changing the composition of the bacterial flora. In a cirrhotic rat model, the administration of propranolol or cisapride stimulated intestinal motility by speeding up intestinal transit, decreasing bacterial overgrowth, improving intestinal permeability and reducing BT. Similarly, the administration of conjugated biliary acids (cholylsarcosine and cholylglycine) suppressed intestinal bacterial overgrowth and reduced BT. Lastly, the use of antioxidants alone or in combination with probiotics (Lactobacillus) in cirrhotic rats reduced the intestinal bacterial load, BT, and oxidative stress in the intestinal wall.

Several previous studies have investigated the effect of CPF on the intestinal wall (ileum and colon) [4, 5]. Exposure to CPF decreases the mRNA expression of genes encoding the tight junction proteins (including claudin-4 and zonula-1); this increases intestinal permeability, enables BT and activates the immune system. In the present study, all the bacterial strains isolated from individual samples from CPF-exposed rats were found to be viable on agar plates, which suggests that the strains tolerated CPF doses of 1 mg/kg BW and 5 mg/kg BW in vivo. All of the isolates were then tested for their susceptibility to 20 antibiotics. Over 90 % of the detected bacterial strains were susceptible; 2 strains (5 %) of S. aureus expressed the MRSA phenotype (PEN^r OXA^r FOX^r) but were susceptible to the remaining antibiotics, 9 isolates (45 %) of E. faecalis were resistant to Er, 6 isolates (30 %) were resistant to PT, and 100 % of the isolates were resistant to L (natural resistance). For E. faecalis, 3 isolates (15 %) expressed the macrolide-lincosamide-streptogramin B (MLSB) + streptogramin A (SA) phenotype (Er^rL^rPT^r), and 12 isolates (60 %) expressed the MLSB phenotype (Er^rL^rPT^s). In strains expressing the MLSB phenotype, the genes encoding resistance are located on a 20–30 kb plasmid [33]. However, resistance is chromosome-encoded in the majority of strains isolated in France [33]. Resistance to macrolide, lincosamide, SA (PT) and SB (virginiamycin) is associated with the production of an SA-inactivating mutant acetyltransferase and an SB-inactivating mutant hydrolase [34‐36]. Expression of the MRSA phenotype and susceptibility to other antibiotics has been documented in the literature [17, 37‐40].

Methicillin resistance is associated with a specific mechanism in which a mobile genetic element [the staphylococcal cassette chromosome (SCC) mec] integrates into the S. aureus chromosome. In the S. aureus chromosome, the mec A gene encodes a specific, methicillin-resistant transpeptidase (penicillin-binding protein 2a) [38], which results in resistance to all β-lactam antibiotics (due to a low affinity for binding to β-lactam). The classic mecA gene codes for resistance to methicillin, kanamycin, gentamicin (in most cases), tobramycin, amikacin, Er, and L but not for resistance to PT and OFX. A new mecA homolog mecB strain (N315) [37, 38] and a mecC strain (LGA256) [39, 40] have been identified. The SCC elements classified as mecB or mecC share ≥70 % nucleotide sequence identity with the classic mecA gene. In the present study, the isolates were phenotypically resistant to all β-lactams but remained susceptible to other tested antibiotics and thus corresponded chromosomally to the mecC gene. Our results are in agreement with the literature data [17, 37‐40]. When considering the genotyped strains, resistant strains (B. cereus and E. faecalis) were only found in the CPF groups (both CPF1 and CPF5)—particularly in juvenile rats at D21. Naphade et al. [41] detected 9.8 and 26.2 kDa plasmids in bacterial cultures from garden soil, indicating that the ability to tolerate high concentrations of heavy metal salts, pesticides and antibiotics is a plasmid-encoded characteristic. Harishankar et al. [42] study of five model intestinal bacteria found that Lactobacillus fermentum, Lactobacillus lactis, and Escherichia coli tolerated high concentrations of CPF (>1400 µg/mL), whereas E. faecalis and Lactobacillus plantarum tolerated lower concentrations (400 and 100 μg/mL, respectively).

The three most tolerant bacteria (L. fermentum, L. lactis, and E. coli) produced an organophosphorus phosphatase that degrades CPF; the concentration of CPF was higher outside the cells (i.e. in the supernatant) than inside the cells. L. fermentum degraded 70 % of the CPF (with 3,5,6-trichloro-2-pyridinol as the end product), L. lactis degraded up to 61 % of the CPF (with CPF oxon as the end product), whereas E. coli degraded only 16 % of the CPF (with CPF oxon and diethylphosphate as the end products). This microbial variation in pesticide tolerance was dependent on the end-product degradation profiles and/or the types of end products [43, 44]. Shafiani and Malik [44] reported that a Pseudomonas spp. isolated from soil was able to tolerate up to 800 μg/mL of endosulfan, 1600 μg/mL of carbofuran and 1600 μg/mL of malathion. All of the bacterial isolates were further tested for their antibiotic susceptibility to seven different antibiotics. A bacteriological analysis of endosulfan-exposed soil by Sepperumal et al. [45] revealed the presence of Bacillus ciradans, Pseudomonas spp., Bacillus lentus, Acinetobacter spp., and B. cereus. Additional plasmid-curing experiments were performed to determine whether pesticide resistance traits in these bacteria were encoded on a plasmid or on the bacterial chromosome. Furthermore, B. ciradans and Acinetobacter spp. contained plasmids of 7 and 4 kb in size, respectively. Once BT has occurred, the isolation of viable bacteria in the liver, spleen, adipose tissue, MLNs and kidneys depends on (i) the immunological competence of the host and (ii) individual bacterial virulence factors that prevent the pathogen’s destruction [46].

We were unable to isolate anaerobic bacteria from the Enterobacteriaceae family; this constitutes a study limitation. In holoxenic rats, the stomach and bowel is mainly colonized from birth to weaning by Lactobacillus, Streptococcus and Enterococcus genera, where as Enterobacteriaceae and anaerobic bacteria are absent [47, 48]. After weaning, the pups’ bacterial flora downstream of the ileum and caecum predominantly consists of anaerobic bacteria [49]. These results are consistent with Raybaud et al. [44] observation of the absence of the Enterobacteriaceae family. However, our failure to isolate anaerobic bacteria was probably due to the detection technique used.

In conclusion, our earlier morphological and molecular analyses showed that long-lasting exposure to CPF (an organophosphate insecticide observed as food residues) altered the maturation of the rat intestine and was associated with intestinal dysbiosis and BT towards sterile organs [4, 5]. In the present study, we confirmed the occurrence of BT and validated the use of molecular typing in rats exposed to CPF in utero and during lactation.

The present study showed that molecular typing can detect intestinal BT in CPF-exposed rats; a combination of ERIC2-PCR and RAPD-PCR patterns can be used as a DNA fingerprint for intestinal microfloral strains. The most prevalent bacteria in intestinal samples from CPF-exposed rats were S. aureus, E. faecalis and B. cereus. Furthermore, chromosome- and plasmid-encoded pesticide-tolerance and antibiotic-resistance traits were observed in S. aureus and E. faecalis strains.

Lastly, we believe that our present findings provide strong evidence to show that early in utero and lactational exposure to CPF may have short- and long-lasting impacts on the digestive system. Importantly, CPF is found in cereals, fruit, vegetables and (potentially) in meat and drinking water, and is known to cross the placental barrier. Under normal circumstances, the intestinal epithelium acts as a barrier against food antigens and commensal bacteria. The latter bacteria live in symbiosis with the host and are significantly involved in physiological functions. The present data on gut homeostasis indicate that CPF disrupts the gut maturation, alters the bacterial equilibrium and enhances BT towards sterile organs. Moreover, the observed effect may be even greater during the food diversification period that occurs after weaning. Hence, we strongly believe that pesticide exposure in infancy (when the digestive system is still immature) should be avoided as much as possible.