Responses of Oral 2,4,6-Trinitrotoluene (TNT) Exposure to the Common Pigeon ( Columba livia ): A Phylogenic and Methodological Comparison

Abstract

Considerable concentrations of the explosive, 2,4,6-trinitrotoluene (TNT) have been found in the soil at many installations where explosives have been used, manufactured, assembled, or destroyed. To evaluate risk to avian receptors, measures of exposure are compared with a threshold level of sublethal toxicity. To date, a single feeding study has evaluated the responses of oral TNT exposure to birds with equivocal results regarding sublethal effects. The present study followed a controlled dosing regime comprising four dose groups and a control (200, 120, 70, 20, and 0 mg TNT/kg body weight [bw]-day) in the common pigeon (Columba livia) for 60 days. Overt signs of toxicity occurred with both sexes between 2 and 3 weeks of exposure. Signs included weight loss, neuromuscular effects (e.g., ataxia, tremors, etc.), and scant red feces (chromaturia). Emetic events following dosing were common and proportional to dose; however, attempts to quantify vomitus compound concentration suggests that birds were marginally successful at removing TNT following administration. Eight of 12 and 2 of 12 males and females died or were moribund in the 200 and 120 mg/kg-day groups, respectively. Changes in hematological parameters, liver, kidney, and ovary weights were related to treatment. Dose-related changes in plasma albumin and sodium concentrations were also observed. These results suggest that subchronic exposure to TNT can adversely affect the central nervous system and hematological parameters in birds. Chemical analysis of blood detected concentrations of the two primary reduction metabolites, but not parent compound, suggesting that toxicity may be due to the bioaccumulation of a toxic intermediate.

The explosive 2,4,6-trinitrotoluene (TNT) has been found in the soil at numerous U.S. Army installations where munitions are used, manufactured, or destroyed. Occurrence of detectable TNT residues in associated areas can be numerous and significant, with concentrations in soil reported as high as 67,000 μg/g (Walsh and Jenkins 1992; Hovatter et al. 1997). Often these areas (e.g., ranges) contain habitats that are used by birds; consequently, questions regarding the effects from avian exposures to TNT have developed.

Information regarding the oral toxicity of TNT in wildlife has been limited to predominantly mammalian species. Although mammalian wildlife species appear to be less sensitive than laboratory rodents, sublethal targets from repeated administration are consistent. Primary effects have been reported to include reductions in mature red blood cells (RBCs) and consequences resulting from lysis (Reddy et al. 2000; Dilley et al. 1982; Levine et al. 1984, 1990; McCain and Ferguson 1998). Although indications of adverse testicular pathology have been described in laboratory rodents (Dilley et al. 1982; Levine et al. 1984), these effects have not been reported for wildlife species.

Gogal et al. (2001) found mortality associated with northern bobwhites (Colinus virginianus) exposed to 3000 ppm TNT in feed (∼178 mg TNT/kg body weight [bw]-day); however, no other differences were found between treatments. Trends suggesting decreases in RBC parameters were found, but were not considered biologically relevant.

Because ecological risk assessments require a toxicological benchmark that represents a threshold for sublethal effects, we examined the oral toxicity of TNT using a controlled dosing design. Because these benchmarks are often intended to apply for an entire vertebrate class, we used a species from the order Columbiformes, the common pigeon (Columba livia) to investigate similarities and differences between species and methods regarding oral TNT exposure in birds.

METHODS

Test Material

TNT was synthesized at the Naval Surface Warfare Center, Indian Head Division, and determined to be >99.9% pure based on dilution of neat sample with isoamyl acetate and analyzed using gas chromatography. Following a 2-week stability study, TNT was mixed into corn oil at 2-week intervals at target concentrations of 240 mg/ml. Prior to dosing, solutions were warmed to 37°C in a water bath. Compound and vehicle were continually mixed as a slurry using an automated mixing bar throughout administration. An aliquot from each solution was randomly obtained and concentrations confirmed using the methods described previously. Slurry concentrations were 219.4 ± 5.14 mg/ml, resulting in doses 91% of target.

Animals and Treatments

Pigeons of the White Carneau strain were purchased from Palmetto Pigeon Farm (Sumter, SC, USA) and transported via climate-controlled vehicle to our American Association of Accredition for Laboratory Animal Care (AAALAC) accredited facility at U.S. Army Center for Health promotion and Prevention Medicine. Upon arrival, birds were housed individually in 24 cage units, each cage 11 inch wide × 12 inch tall × 15 inch deep constructed of 1/2 inch by 1 inch polyvinylchloride (PVC) coated wire. Each unit had two adjustable-height, automatic-nipple drinkers and individual glass jar feeders fitted with a stainless steel ring lid. Pigeons were provided with certified pelleted pigeon chow (Ralston Purina, St. Louis, MO, USA) every 3 days, which was weighed along with any spillage.

Birds were kept in constant photoperiod durations of 16-hr light: 8-hr dark with humidity maintained between 30% and 70% and the temperature between 64°F and 79°F.

Birds were quarantined/acclimated for up to 56 days before ivermectin treatment for ectoparasites (Columbicola columbae, order: mallophaga, family: Pediculidae; Pseudolynchia canariensis, order Diptera, family: Hippoboscidae; see Owiny and French 2000). Birds were then divided by sex and then randomly sorted into four treatments and one control consisting of 12 birds/sex/group (N = 120). Birds were uniquely identified by cage card and subcutaneous transponder. Treatment groups were volumetrically given oral doses equivalent to 200, 120, 70, 20 mg TNT/kg bw. Control treatments consisted of birds given the same volume of corn oil as the highest dose group.

Birds were weighed on days −1, 0, 7, 14, and weekly thereafter. Daily dose for each bird within each treatment was based on the most recent body weight for each individual for each dose group. This method ensured accurate dosing and reduced inconsistencies in dosing birds via the dietary route, where feed avoidance can occur. Dosing occurred each day from 0800 to 1100 for 60 days.

At the conclusion of the treatment period, birds were euthanized and necropsied. Immediately following euthanasia, blood was collected via cardiac puncture with a 21-gauge needle fitted onto a 3-ml syringe. Together, the needle and syringe was flushed with heparin to reduce clotting. Following blood collection, the needle was removed and the contents emptied into a 3-ml Vaccutainer containing lithium heparin. Major blood conditioning organs (e.g., liver, spleen, kidney), brain, and gonads were harvested and weighed to the nearest 0.001 g. Because individual body weights varied, organ weights were normalized and expressed as a ratio of total body weight.

Hematology

Because nitrated energetic compounds have been shown to target mature red blood cells in other laboratory animals, blood and associated clinical chemistry parameters were evaluated. Whole blood was evaluated for RBC and white blood cell (WBC) cellularity, packed cell volume (PCV), total protein solids, and five-part leukocyte differential conducted using the same standard methods provided previously (Gogal et al. 2002). Hemoglobin content (Hb) was determined from whole blood using a HemoCue Hemoglobin photometer (HemoCue AB, Angelholm, Sweden).

Clinical Chemistries

Following collection of aliquots for hematology parameters, the remaining whole blood from each heparinized tube was centrifuged to obtain plasma for clinical chemistries. A VetTest chemistry analyzer (IDEXX Laboratories, Westbrook, ME) was then used to evaluate the following plasma parameters: aspartate aminotransferase (AST; SGOT), glucose (GLU), creatinine (CREA), cholesterol (CHOL), blood urea nitrogen (BUN), uric acid (URIC), phosphorus (PHOS), albumin (ALB), total protein (TP), lactate dehyrogenase (LDH), and alkaline phosphatase (ALKP). Plasma electrolytes (e.g., sodium, chloride, potassium) were also analyzed using a VetLyte Electrolyte Analyzer (IDEXX Laboratories, Westbrook, ME).

Histopathology

Tissues were collected, trimmed, fixed in formalin, and embedded in paraffin. These tissues were then sectioned at 6 microns, stained with hematoxylin and eosin, and examined via light microscopy. Treatment was not known to the pathologist until after the histopathologic examination was completed. Because other observations were suggestive of central nervous system (CNS) damage, brain sections were evaluated with Fluor-Jade B using the methods described in Schmued and Hopkins (2000) and examined by fluorescence microscopy.

Statistical Analysis

Tests of the data for normality and equal variances for each group were conducted to satisfy the assumptions associated with parametric tests. If the data failed to fit a normal distribution or were of unequal variance distributions, they were either log transformed and reevaluated or ranked and analyzed using a two-way analysis of variance (ANOVA) on the ranks. Normally distributed data were analyzed using a two-way ANOVA using sex and treatment as variables. Body weight data were evaluated separately for each sex using repeated-measures ANOVA. If significant, a pairwise multiple comparison procedure was conducted using a Tukey test. Deviations in these procedures are noted where applicable. Statistical significance was defined at the p < .05 level. To allow for a consistent characterization of the results, means and standard errors of the mean are presented for all data. Tests were conducted using SigmaStat Ver. 3.0 (SPSS Inc., Chicago, IL, USA).

This study was conducted consistent with the standards found in Title 40 Code of Federal Regulations (CFR), Part 792, Good Laboratory Practices. The investigators and technicians adhered to the Public Health Service Policy on Humane Care and Use of Laboratory Animals (NIH 2002) and the Guide for the Care and Use of Laboratory Animals (NRC 1996).

RESULTS

Overt Observations

Mortality occurred following 14 days of exposure in both male and female high-dose groups (Table 1), with two exceptions (described further). Mortality and morbidity were often preceded with symptoms of neuromuscular origin (e.g., convulsions, tremors, ataxia), lethargy, and scant red-colored feces. Red-colored feces was more frequently observed in birds of the 200, 120, and 70 mg/kg-day groups, but not the 20 mg/kg-day or control groups (p < .001). Morbidity was often associated with a reduction in body weight, but not with feed consumption (Figures 1 and 2). Mean decrease in body weight was greater for the 200 mg/kg-day groups compared with the controls and the 20 and 70 mg/kg-day groups (p < .001). Other observations associated with the high-dose groups (120 and 200 mg/kg-day) included scant feces and hypothermia of the feet.

Although birds were sexed by the vendor before the onset of the study, two females in the 20 mg/kg-day group and one female in the 200 mg/kg-day group were found to be male. A single female in the 70 mg/kg-day and one in the 120 mg/kg-day group were removed and euthanized due to symptoms of dyspnea and congested breathing. Subsequent investigation revealed the presence of an Aspergillus infection. Given the even distribution of individuals among treatments with this condition, it was not considered to be influenced by test compound.

There was a dose-related increase in regurgitation (event/bird/day) immediately following exposure (within 15 min). This increase was greater for the 70, 120, and 200 mg/kg-day groups, but not for the controls and the 20 mg/kg-day groups (Kruskal-Wallis one-way ANOVA on ranks, Dunn’s method, df = 4, p < .05). To attempt to determine the efficiency of this response, vomitus was collected and analyzed for TNT concentration using a modification of the Environmental Protection Agency (EPA) method 8330 (N = 4). Based on these results, very little of the TNT was regurgitated (mean = 3.78% [±3.17%] of administered dose).

Organ/Body Weights

There were treatment and sex-related differences for liver and kidney/bw ratios (p < .001 for treatment, p < .005 for sex; df = 98; Figure 3). There was no interaction between sex and treatment (p > .94). No treatment-related differences or dose-related trends in brain, spleen, and testes/bw comparisons (p > .86, .43, and .19, respectively); however, there were differences in ovary/bw ratios between the 0 and 20 mg/kg-day groups and the 120 and 200 mg/kg-day groups (p = .003).

Hematology and Clinical Chemistries

PCV, Hb, RBC, and WBC counts were all affected by treatment (Table 2). No changes in leukocyte ratios were found attributable to treatment; however, there were differences between sexes. Dose-related differences were found in plasma sodium, chloride, and albumin concentrations (Table 3). Despite apparent treatment-related differences, values were within the normal ranges reported for this species.

Histopathology

Testes of four males that died/sacrificed before the end of the study from the 200 mg/kg-day group had moderate to severe hypospermatogenesis, which consisted of a pronounced reduction in the number of maturing spermatozoa in the seminiferous tubules. Mild to moderate hemosiderosis was found in the spleen of two moribund females that were sacrificed from the 200 mg/kg-day group, and mild hemosiderosis was found in the liver of three females from the 200 mg/kg-day group and one from the 70 mg/kg-day group. Because both of these findings may be attributable to various pathologic processes associated with debilitation, direct etiology could not be determined. No other changes were observed that could be attributable to treatment. Examination of brain sections with Fluoro-Jade B revealed no evidence of neuronal degeneration.

DISCUSSION

Effects from oral TNT exposure in pigeons are similar to that described for quail (Gogal et al. 2002). Although levels in quail that cause acute mortality are approximately 10-fold higher, mortality from repetitive subchronic exposures to both species appear to occur 2 to 3 weeks following initial exposures. Gogal et al. (2002) reported that 4/10 northern bobwhite (Colinus virginianus) exposed to 3000 ppm TNT in feed died or were moribund within 30 days of exposure exhibiting neuromuscular symptoms consistent with those found in this study. Levels that caused mortality from repetitive exposures in the northern bobwhite and the common pigeon were similar (178 mg/kg-day and 182 mg/kg-day for the northern bobwhite and common pigeon, respectively, following the analytical adjustment of the nominal dose estimates to the present study). These data suggest that there is little interspecific variation in this response from oral TNT exposure in these species.

This observed delay in mortality combined with the known in vivo metabolism of TNT suggests the toxic accumulation of metabolic intermediates. TNT has been found to be reduced quickly in vivo to result in either one of the monoamino derivatives, both of which tend to increase in concentration relative to exposure to the parent compound (Johnson et al. 2000; Yinon 1990). This may be due to a rate-limiting step in the secondary reduction of an additional nitro group or the conjugation of the entire molecule (Yinon 1990). Single blood samples taken from one representative for each of the TNT-exposed pigeons resulted in 16, 45, 210, and 100 μg/L of 4-amino-2,6-dinitrotoluene and 4.5, 11, 22, and 14 μg/L of 2-amino-4,6-dinitrotoluene for birds from the 20, 70, 120, and 200 mg/kg-day groups. Concentrations of parent compound were 7.2 and 2.8 for birds from the 120 and 200 mg/kg-day groups; no detectable levels of TNT were found in the other samples.

The controlled oral dosing of TNT to birds provided more descriptive data regarding effects than the previous feeding study with northern bobwhite. Differences in liver/bw, kidney/bw ratios were present between the 20 and 70 mg/kg-day and higher dose groups. Treatment-related changes in electrolytes and albumin levels coupled with these data suggest early indications of kidney and liver effects, respectively. Other treatment-related differences were found in specific hematological parameters, yet occurred at higher levels where more overt effects were observed. Body weight loss was not associated with feed consumption, suggesting that TNT exposure had no effect on appetite, yet altered individual energy budgets. This occurred either through affecting feed absorption and/or transit time or through greater energy requirements needed to maintain homeostasis. By avoiding some complications associated with feed avoidance, this exposure regime is recommended if toxicity data are intended to be used in toxicity benchmark derivation (i.e., toxicity reference values).

Packed cell volume and hemoglobin concentrations were different from controls in the 70 mg/kg-day group, red blood cell counts in the 120 mg/kg-day group, and white blood cell counts at 200 mg/kg-day group. However, comparisons of these data with reference data for C. livia (Gayathri and Hegde 1994; Ritchie, Harrison, and Harrison 1994; Cray 2000; Carpenter, Mashima, and Rupiper 2001) show that these values were within the normal ranges for this species. This comparison coupled with a lack of extramedullary hematopoetic observations and treatment-related hemosiderin incidence suggests that the hematological system is not a sensitive or biologically relevant target for TNT toxicity in birds.

Adverse effects appear to begin and be coincident with the 70 mg/kg-day group. This is also the lowest level at which emesis occurs and red-colored feces became evident. The presence of red-colored urine and/or feces is common in studies where mammals were exposed to TNT (Dilley et al. 1982; Levine et al. 1984; Gogal 2000; Reddy et al. 2000). Red-colored feces (chromaturia) were observed 0%, 9%, 56%, 60%, and 100% of the time in the control, 20, 70, 120, and 200 mg/kg-day groups, respectively. The presence of this biomarker shows promise as a noninvasive indicator of exposure and potentially of adverse effects that could have useful field applications.

In summary, C. livia were similarly sensitive to oral TNT exposures compared to northern bobwhite. Additionally, the results indicate that birds have a different manifestation of toxicity (i.e., exhibit different targets systems) from mammals that has a physiological basis. TNT-induced effects were related to alterations in CNS and apparent alterations in individual energy budgets. Despite evidence in other taxa that the hematological system may be an important target of TNT toxicity, pigeons showed little indication of these effects. Importantly, oral exposure via gavage has proven to be reliable, easy, and useful in avoiding complications associated with exposures via feed.