Development of predictive models of laboratory animal growth using artificial neural networks.

Traditional regression analysis of body weight growth curves encounters problems when the data are extremely variable. While transformations are often employed to meet the criteria of the analysis, some transformations are inadequate for normalizing the data. Regression analysis also requires presuppositions regarding the model to be fit and the techniques to be used in the analysis. An alternative approach using artificial neural networks is presented which may be suitable for developing predictive models of growth. Neural networks are simulators of the processes that occur in the biological brain during the learning process. They are trained on the data, developing the necessary algorithms within their internal architecture, and produce a predictive model based on the learned facts. A dataset of Sprague-Dawley rat (Rattus norvegicus) weights is analyzed by both traditional regression analysis and neural network training. Predictions of body weight are made from both models. While both methods produce models that adequately predict the body weights, the neural network model is superior in that it combines accuracy and precision, being less influenced by longitudinal variability in the data. Thus, the neural network provides another tool for researchers to analyze growth curve data.

Objective measures of ocular irritation as a consequence of hydrogen sulphide exposure.

Hydrogen sulphide (H2S) exposure has been reported to adversely affect the eyes. Objective information regarding the extent of injury relative to varied exposure concentrations of H2S has not been published. The exfoliative eye cytology procedure has been demonstrated to provide quantitative information regarding the degree of ocular irritation and to allow for comparisons to be made between the irritants. This paper evaluates this procedure's efficiency in determining the degree of ocular irritation resulting from hydrogen sulphide exposure in rats.

Cytotoxic effects of hydrogen sulfide on pulmonary alveolar macrophages in rats.

Respiratory rates (basal and zymosan-stimulated) and cell viability were monitored in pulmonary alveolar macrophages (PAM) from rats exposed to 0, 70, 280, and 560 mg/m3 (0, 50, 200, and 400 ppm) hydrogen sulfide (H2S) gas for 4 h. Zymosan-stimulated respiratory rates were markedly reduced in PAM collected from rats exposed to 280 and 560 mg/m3 H2S; however, their basal respiratory rates were not affected. Significant decrease in cell viability was also observed in samples from 560 mg/m3 H2S-treated rats, but it remained high and unchanged in other treatments. In vitro incubation of PAM from control rats with sulfide (a precursor of H2S) and its two oxidation products, sulfite and sulfate, showed that sulfide was markedly more inhibitory to both respiratory rates than sulfite or sulfate. These treatments did not affect cell viability.

Effects of hydrogen sulfide exposure on lung mitochondrial respiratory chain enzymes in rats.

Fischer-344 rats were exposed for 4 hr to various concentrations of hydrogen sulfide (H2S) gas and killed either immediately or at 1, 24, or 48 hr after exposure. Mitochondrial fractions from lung tissues were assayed for the activities of respiratory chain enzymes. Exposure of rats to a low concentration (10 ppm) of H2S caused no significant changes in the activities of lung mitochondrial enzymes. However, exposure to sublethal concentrations of H2S (50-400 ppm) produced marked and highly significant depressions in the activities of cytochrome c oxidase and succinate oxidase complexes of the respiratory chain. The inhibition of cytochrome c oxidase activity in lungs was most severe (greater than 90%) in rats that died from acute exposure to greater than 500 ppm H2S. In rats exposed to 200 and 400 ppm H2S, a marked recovery in cytochrome c oxidase activity of lungs was observed at 24 and 48 hr postexposure. Studies in vitro with rat lung mitochondria showed that low concentrations of sulfide also caused a similar and selective inhibition of cytochrome c oxidase activity. This effect was reversed upon removal of sulfide either by washing or by oxidation with methemoglobin. The nature of sulfide inhibition of cytochrome c oxidase was noncompetitive with respect to ferrocytochrome c. Because the activities of NADH-cytochrome c reductase and succinate-cytochrome c reductase were not significantly altered by H2S exposure and in vitro treatments with low concentrations of sulfide, it is concluded that under physiological conditions H2S would block the respiratory chain primarily by inhibiting cytochrome c oxidase. Such a biochemical impairment would lead to functional (histotoxic) hypoxia in the lung tissues.

Peracute toxic effects of inhaled hydrogen sulfide and injected sodium hydrosulfide on the lungs of rats.

This study was designed to test whether intraperitoneally injected sodium hydrosulfide (NaHS) would mimic the pulmonary alterations induced by lethal peracute exposure to an atmosphere containing hydrogen sulfide. Groups of five Sprague-Dawley rats were exposed to an atmosphere of either 2317.6 +/- 547.3 mg m-3 H2S (H2S group) or no H2S (air group), or were injected intraperitoneally with a solution containing 30 mg kg-1 sodium hydrosulfide (NaHS group) or saline solution (vehicle control). Rats of the air and saline groups were killed by cervical dislocation. All rats exposed to H2S or injected with NaHS died within 3 min; however, only rats exposed to H2S showed severe respiratory distress in the agonic phase preceding death. In addition, rats in the H2S group had a notable discharge of serous fluid from the mouth and nostrils. At necropsy, all rats in the H2S group had gross and histologic evidence of pulmonary edema characterized by massive extravasation of eosinophilic fluid into the bronchoalveolar space. In contrast, the lungs of rats injected with NaHS or saline or exposed to air were unaffected. It was concluded that the edematogenic effect of H2S in the lungs cannot be reproduced by injection of NaHS. The severity of lung edema induced by a peracute exposure to H2S was extensive enough to account for death.

Use of sulfur hexafluoride and a three-dimensional sampling grid for evaluation of inhalation chamber performance.

Knowledge of the distribution of a toxic agent, for example, hydrogen sulfide, within an exposure chamber is essential for satisfactory inhalation toxicology studies. The use of a three-dimensional sampling grid is described for sampling inhalation chamber atmospheres containing various concentrations of sulfur hexafluoride gas as a substitute for the more toxic hydrogen sulfide.

Concentration-time interactions in hydrogen sulphide toxicity in rats.

Concentration-time interactions were investigated in young male and female Sprague-Dawley, Long Evans and Fischer-344 rats exposed to hydrogen sulphide for two, four or six hours. Higher concentrations caused more deaths, with no significant difference for duration of exposure. A significant sex effect was noted with 30% mortality in males and 20% in females, with no significant difference among strains. Changes in weight were significant: increasing with concentration, higher in males than in females, different among strains (Fischer-344 less than Sprague Dawley less than Long Evans), and affected by duration of exposure. Lethal concentration values (LC50 and LC10) were estimated, for the pooled data set (n = 456); the probit equation was Y = -5.74749 + 3.8259X where X is log10 dose of hydrogen sulphide in parts per million. The LC50/LC10 values were 644/298 parts per million (902/417 mg m-3) respectively. Individual probit analyses were also performed for strain, hours of exposure and sex. The LC50 and LC10 values for male, female and strain were not different. Significant differences were observed among LC50/LC10 values for hours of exposure (2 h = 587/549 parts per million, 822/769 mg m-3; 4 h = 501/422 parts per million, 701/591 mg m-3; 6 h = 335/299 parts per million, 469/491 mg m-3). There was no effect of spatial position in the exposure chamber on the distribution of mortality. All rats of all strains dying had severe pulmonary edema.

Profiles of type-II pneumocytes in rats inoculated intratracheally with bacterial lipopolysaccharide.

Ultrastructural and morphometric profiles of type-II pneumocytes (P-II) were investigated in rats killed 18 or 24 hours after a single intratracheal inoculation of bacterial (Escherichia coli) lipopolysaccharide (LPS). Inoculation with LPS induced pulmonary injury and inflammation, as measured by increased lactate dehydrogenase and alkaline phosphatase activities and increased numbers of polymorphonuclear neutrophils in fluid collected by bronchoalveolar lavage. Marked ultrastructural changes and desquamation of a few P-II developed at the time of high activity of lactate dehydrogenase and alkaline phosphatase in bronchoalveolar lavage fluid. Ultrastructural changes included swollen mitochondria and localized cisternal dilatation of the endoplasmic reticulum in which was contained membrane-bound homogenous material of medium electron density. Twenty-four hours after LPS inoculation, point-count stereologic analysis and digitizing morphometry revealed greater than 50% increase in P-II size. Changes in cell size corresponded with ultrastructural finding of swollen cells. Results obtained by point-count stereologic analysis and digitizing morphometry were highly correlated (r = 0.95). Lamellar bodies (LB) comprised 12 to 15% of P-II volume. Volume density and number of LB remained unaltered in LPS-injured P-II, and evidence of accelerated release of LB was not detected after LPS inoculation. Exudated polymorphonuclear neutrophils and pulmonary alveolar macrophages were involved actively in the phagocytosis of LB originating from necrotic and desquamated P-II. On the basis of measurement of enzyme activity (enzymes released into the bronchoalveolar space), considerable ultrastructural alterations developed in P-II when maximal LPS-induced pulmonary cell injury took place.

Histopathologic and electron microscopic studies on the acute toxicity of ochratoxin A in rats.

Ochratoxin A was given by gavage to male rats. Moribund and dead animals were necropsied, and the surviving rats, including the controls, were killed 48 hours after dosing. Many of the principal rats were moribund, or began dying, within 12 to 24 hours after dosing. Lesions suggestive of disseminated intravascular coagulation were seen by light microscopy as early as 12 hours after dosing; fibrin deposits were in the spleen, brain choroid plexus, glomerular capillaries, liver, and heart. Renal tubular nephrosis, hepatic and lymphoid necrosis, and necrotic enteritis with villous atrophy were also seen. Electron microscopy demonstrated fibrin strands mixed with degranulated platelets, necrotic leukocytes, and swollen endothelial cells in glomerular capillaries. Myocardial changes included focal supercontracted sarcomeres adjacent to intercalated disks. Swollen sarcolemma, lysed myofibrils and fragmented Z-bands with interstitial edema, vascular thrombosis, and endothelial damage were also seen. The acute pathologic changes induced by ochratoxin A in the intestine, liver, and lymphoid tissues were more obvious than the tubular nephrosis, and the development of a disseminated intravascular coagulation-like syndrome with myocardial changes was a complicating factor.

Effects of high dietary sulphur on enzyme activities, selenium concentrations and body weights of cattle.

This study was designed to assess the effects of a moderate increase in dietary sulphur (S) in cattle. Twelve animals were initially fed a basal concentrate (S = 0.2%) and then divided into two groups; one fed basal and the other high S (S = 0.75%) concentrates. Health, body weight gains, and activities of erythrocyte enzymes-glutathione peroxidase (GSH-Px), superoxide dismutase (SOD), glucose-6-phosphate dehydrogenase (G6PD), acetylcholinesterase (AChE), plasma- asparate aminotransferase (AST), and whole blood concentrations of selenium (Se) were monitored at various stages of the study. Marked increases in the activities of GSH-Px, SOD and G6PD from the pretrial values were observed upon initial feeding of basal concentrate diet. Sex related differences were not evident in enzyme activities and Se concentrations of the blood. A high linear correlation (r = 0.92) between averages of GSH-Px activity and Se concentration of blood was observed in both sexes. Increasing the amount of S in the concentrate diet (from 0.2 to 0.75%) did not produce any statistically significant change in enzyme activities and Se concentrations, body weight gains, and health of the cattle during the 85 days feeding period. The results indicate that a moderate increase in the dietary S would not impair Se and copper status or cause related disorders in cattle.

Incidence of Salmonella contamination in broiler chickens in Saskatchewan.

The incidence of Salmonella contamination in ten Saskatchewan broiler flocks varying in size from 6 200 to 14 000 was investigated from February, 1977 to April, 1979. Prior to the initial chick placement, brooding equipment, feed, water and fresh litter samples were found to be free of Salmonellae. Samples obtained from the clean and disinfected processing plant equipment before the commencement of daily operation were negative except the isolation for Salmonella anatum from the fingers of the defeathering machine in flock 4. There was no evidence of Salmonella contamination in flocks 5, 6, 8 and 10. The incidence of Salmonella was lower when cloacal swabs were taken from day old chicks fasted for 48 hours than for the same groups of chicks when carcasses were blended in nutrient broth (flocks 7 and 9). The blending of such chicks appears to be a more critical test. The serotypes isolated from eviscerated birds were the same as those isolated from used litter samples. Salmonella saintpaul was isolated from a water sample at 53 days in flock 1 and the same serotype was recovered from the intestinal contents and skin of eviscerated birds. Salmonella typhimurium was recovered from the eviscerated birds and neck samples in flock 3. In flock 4, S. saintpaul and S. anatum were isolated from 13% of the eviscerated birds sampled. Salmonella thompson, Salmonella agona and Salmonella heidelberg were recovered from 61%, 5% and 1%, respectively, of the processed carcasses sampled in flock 7.

The effects of ochratoxin A on the immune response of Swiss mice.

The acute intraperitoneal toxicity of ochratoxin A (OA) for adult female Swiss mice is presented. The seven-day LD(50) was calculated to be 48 +/- 3.2 mg/kg. Daily intraperitoneal administrations of 10 mg OA/kg resulted in 50% mortality by the tenth day of injection. Clinical symptoms included depression, huddling, roughened hair coats, humped backs and reduced weight gains. Mice injected intraperitoneally daily for 50 days with 5 mg OA/kg had a significantly (P<0.01) depressed antibody response to killed Brucella abortus. In contrast, oral administrations of OA at 4 ppm in feed for 50 days did not depress titre levels. Ochratoxin A also significantly (P<0.01 intraperitoneal; P<0.05 oral administrations) reduced body weight gain over the period of the trials. Neither oral nor intraperitoneal administration of OA for 50 days affected the response of mice to sheep red blood cells although both the number of antibody-forming cells and the number of cells per spleen were significantly lowered (P<0.01) by cyclophosphamide. Both spleen and body weights were significantly lowered (P<0.05) in the groups given OA. There was a significant depression of blast transformation (P<0.01) in mice treated intraperitoneally with either OA or cyclophosphamide and stimulated with concanavalin A; oral administration of OA did not depress blast transformation. It would appear that lower levels of exposure, e.g. 4 ppm OA in feed, do not cause depression of the immune response of mice. The depressive effect seen at much higher levels may be a result of a nonselective toxic effect.

Effects of ochratoxin A on egg production, body weight, and feed intake in white leghorn hens.

Two trials were conducted to determine whether ochratoxin A (OA) or reduced feed consumption was responsible for reduced body weight and egg production in hens. In an 8 week restricted feed intake trial two groups of White Leghorn hens received amounts of non-contaminated feed similar to that consumed in a previous trial by hens fed either noncontaminated diet or one containing 4 ppm OA. A significant (P less than .05) loss of body weight was observed during the 5th and 6th weeks in the group receiving a similar amount of feed as the OA-treated group in the previous trial, indicating that reduced feed consumption might account for the loss in body weight. Since egg production was not affected, it was possible that OA may depress egg production through a mechanism that is separate from its influence on feed intake. In an organoleptic trial White Leghorn hens were offered a choice of layer feed either with or without OA added for a three week period. There was a significant (P less than .05) reduction in the consumption of OA-contaminated feed.

Mycotoxins in animal feedstuffs and tissues in Western Canada 1975 to 1979.

Results of analyses of specimens of plant or animal origin for various mycotoxins are presented. Analyses for aflatoxins, ochratoxins and zearalenone were most frequently requested. Aflatoxin B1 was found in one of 474 specimens at a level of 60 ppb in a sample of hay. Ochratoxin A was detected in four of 148 specimens of grains and two of 19 specimens of corn at levels up to 500 ppb. Trichothecenes were qualitatively found in two of 108 specimens of forage, three of 182 specimens of feeds and one of 148 specimens of grains. Ergot was detected qualitatively in three specimens of rye and one of forage. An overall detection rate of 3.8% of potent mycotoxins suggests that acute or chronic mycotoxicoses may occasionally occur in farm livestock or poultry.

Effects of ochratoxin A on growth response and residues in broilers.

The effects of continuous feeding of graded levels (.5, 1.0, 2.0 ppm) of ochratoxin A (OA) for eight weeks to male and female broiler chickens were investigated. A depression in body weight gain was observed in all groups receiving OA. The depression was proportional to the level of exposure to OA and was more marked and prolonged in males than in females. Detectable residues of OA were observed in the liver and kidney of birds fed 2 ppm OA. Residues disappeared from liver within 24 hr and from kidney within 48 hr after withdrawal of the mycotoxin from feed. No residues of OA were found in muscle or fat.

Evaluation of brine shrimp (Artemia salina) larvae as a bioassay for mycotoxins in animal feedstuffs.

Brine shrimp larvae was tested as a possible simple biological screening system to identify specimens of animal feedstuffs that should be examined further by chemical analytical procedures for mycotoxins. All extracts of the control, nonmouldy feedstuffs increased larval mortality, this being most marked in the case of silage. Chemical and biological testing of diagnostic specimens indicated that the bioassay identified two of four chemically positive specimens and 59 of 135 chemically negative specimens and 59 identified larvicidal compounds present in normal feedstuffs gave a high percentage (56%) of false-positive bioassay results when compared to the results of chemical analyses for three mycotoxins. The use of brine shrimp larvae did not materially reduce the necessity of conducting chemical analyses for mycotoxins.

Ochratoxicosis in White Leghorn hens.

As ochratoxin-producing fungi are often isolated from poultry feeds and from cereals used in compounding these feeds, a study has been made of the effects of feeding ochratoxin A (OA) to laying birds. Four groups of White Leghorn hens were fed 0, .5, 1, and 4 ppm in the feed respectively. Egg production and feed consumption declined in the three groups given OA, while egg and body weight were depressed only by feeding higher levels of OA. Fertility and hatchability were unaffected by feeding OA. Prothrombin times were increased and total serum proteins were decreased after feeding 1 or 4 ppm for six weeks. After the withdrawal of the mycotoxin-contaminated feed, OA disappeared from the muscle after 24 hr but persisted in liver and kidney for more than 48 hr. No residues were found in fat or skin.

Acute oral ochratoxicosis in day-old White Leghorns, turkeys and Japanese Quail.

The comparative acute, oral toxicity of ochratoxin A for three day-old avian species is presented. The seven-day LD50 value for White Leghorns was calculated to be 3.4 +/-0.19 mgm./kg., for turkeys to be 5.9 +/- 0.72 mgm./kg., and for Japanese quail to be 16.5 +/- 0.56 mgm./kg., body weight. The dose-response curves are linear and parallel through one standard deviation on either side of the LD50 when log-dose is plotted against probit for survivors. It is suggested that the mechanism of action of ochratoxin A is similar in the three species, though the potency differs. The reduction in weight gain of Leghorn survivors was proportional to dose, and was observed in two separate traials over an overall dosage range from 0.2 mgm./kg. to 5 mgm./kg. The turkeys showed only a slight reduction in weight gain at doses less than 4mgm./kg., a more marked reduction being observed at higher dose levels. The quail did not show reduction of weight gain at dose levels below 10.9 mgm./kg., though the reduction was proportional to dose at higher levels. All birds dying of acute ochratoxicosis revealed a progression of symptoms from listlessness, huddling, occassionally diarrhoea, ataxia, prostration and death. Viscereal gout was observed at necropsy of the Leghorns.

Isolation of brucella abortus from two dogs in contact with bovine brucellosis.

On a farm where several cattle were serologically positive for bovine brucellosis, three dogs were found to have titres greater than 400 i.u. to Brucella abortus. The titres persisted until the dogs were killed over two months later. Two male dogs were necropsied. B. abortus was isolated from the spleen of both dogs. While farm dogs are not thought to be a major reservoir of bovine brucellosis they may be considered as possible carriers in imfected herds and should be considered during the investigation and eradication of bovine brucellosis.

Lead and mercury residues in kidney and liver of Canadian slaughter animals.

Liver and kidney samples were collected from Canadian slaughter animals during the winter of 1973-1974. A total of 256 samples were analyzed for lead. Mean lead levels of 1.02 ppm in poultry liver, 1.04 ppm in bovine liver, 1.02 ppm in bovine kidney, 0.73 ppm in pork liver and 0.85 ppm in pork kidney were found. A total of 265 samples were analyzed for mercury. Mean mercury levels of 0.003 ppm in poultry liver, 0.007 ppm in bovine liver, 0.008 ppm in bovine kidney, 0.001 ppm in pork liver and 0.013 ppm in pork kidney were found. All levels detected were below the Canadian official tolerance of 2 ppm for lead and administrative tolerance of 0.5 ppm for mercury.