Materials and Methods

Animals—Twelve C3H/HeJ mice were used in the study. All procedures involving animals were conducted under the oversight of the campus attending veterinarian and were approved by the University of California, Davis, Institutional Animal Care and Use Committee.

Bacterial culture—Four strains of C pseudotubercu-losis (designated 107, 650, 115, and 312) were chosen for the experiment. These strains represent random amplification of polymorphic DNA groups 3 (strains 115 and 650), 4 (strain 312), and 7 (strain 107).¹ All were originally obtained from external abscesses on horses (strains 115 and 650 were from California, strain 312 was from Kentucky, and strain 107 was from Colorado). Isolates were confirmed to be C pseudotuberculosis on the basis of a positive result for gram staining; coryne-form morphology; appearance of white, variably hemolytic colonies on blood agar; and detection of catalase. Reduction of nitrate was determined by use of nitrate broth. Bacteria were maintained at −70°C on beads.^a After bacteria were thawed, they were plated onto 5% sheep blood agar and incubated for 48 hours at 37°C in air. Isolated colonies of bacteria were transferred to BHI broth that contained Tween-20 and incubated for an additional 48 hours at 37°C. Bacteria were harvested by centrifugation, washed twice in saline (0.9% NaCl) solution, and then adjusted to an approximate density of 2 × 10⁵ bacteria/mL in saline solution, as determined on the basis of results of a quantitative gram stain of 20 ML of bacteria in BHI broth. Simultaneously, a 20-ML loop of bacteria in saline solution was plated on sheep blood agar and incubated for 48 hours to enable us to count colony-forming units and confirm the dose administered.

The 4 isolates were evaluated for phospholipase D transcription by use of a PCR system. Total RNA was isolated by use of a commercially available kit.^b The RNA for the negative control sample from uninfected BHI broth was processed in parallel. Samples of RNA were then treated with RNase-free DNase at 37°C for 30 minutes and heat treated at 70°C for 5 minutes. First-strand cDNA was synthesized by use of random hexamers and Moloney murine leukemia virus reverse transcriptase at 37°C for 60 minutes, followed by inactivation of the reverse transcriptase for 5 minutes at 95°C. The cDNA was evaluated for phospholipase D transcription by use of a quantitative PCR assay, as reported elsewhere.⁴ Negative control samples (ie, water) were included in each assay. Final quantitation was performed by use of the comparative C_T method, where the C_T of the phospholipase D signal was standardized to the C_T for the bacterial universal 16S, which eliminated effects of variable amounts of initial cDNA and allowed comparison of phospholipase D transcription among isolates.

Experimental procedures—A preliminary experiment was conducted to determine the optimum inoculation route. Three mice were inoculated with C pseudotuberculosis. One mouse was inoculated by the ID route (in the dermis caudal to an ear), 1 mouse was inoculated by the IP route through the left inguinal region, and 1 mouse was inoculated by topical application of calcium thioglycolate^c (an irritant) on the dorsal aspect of the lumbar region, which was subsequently followed by topical application of the C pseudotuberculosis inoculum to the irritated skin. Mice were evaluated daily by a veterinarian to assess activity, appetite, body condition score, responsiveness to gentle tactile stimuli, and hydration status (which was assessed on the basis of skin turgor, tackiness of the mucous membranes, and position of the ocular globe). All 3 mice died or were euthanized by day 9 after inoculation. Necropsies were performed, and gross lesions were described.

To evaluate comparative virulence, strains chosen to infect C3H/HeJ mice were grouped as rapid-growing (strains 115 and 312) and slow-growing (strains 107 and 650) strains, as described elsewhere.¹ Prior to inoculation, quantitative bacterial culture was performed on BHI agar to evaluate growth rate. A 20-ML loop of 2 × 10⁵ bacteria/mL in BHI broth was inoculated onto the first quadrant of a BHI agar plate, and the plate was streaked for isolation of colonies. Colony count and size were assessed daily for 4 days.

Twelve C3H/HeJ mice (3 mice for each C pseudotuberculosis strain) were inoculated ID with 1 × 10⁵ bacteria in saline solution. Mice were anesthetized by SC administration of ketamine hydrochloride (20 mg/ kg) and xylazine hydrochloride (4 mg/kg). Mice were then inoculated with C pseudotuberculosis by ID injection into the dermis of the pinnae. Half of the inoculation dose was injected into the dermis of each ear. Mice were evaluated daily by a veterinarian, as described previously. On day 8, all mice were euthanized by administration of an overdose of ketamine and xylazine followed by exsanguination via cardiac puncture and cervical dislocation. Following euthanasia, necropsies were conducted.

Blood samples were analyzed at the Veterinary Medical Teaching Hospital at the University of California, Davis, by use of an automated system.^d The RBC count, WBC count, hemoglobin concentration, and mean cell volume were measured directly, whereas the Hct, mean cell hemoglobin content, and mean cell hemoglobin concentration were calculated by the instrument. Differential cell counts were performed manually on Wright-stained blood smears.

During necropsy, swab specimens of the liver, spleen, lungs, and mesenteric lymph nodes were obtained by use of a sterile bacterial loop and sequentially plated on BHI medium in 4 quadrants. Plates were incubated for 4 days. Corynebacterium pseudotuberculosis was identified as a white bacterial colony that maintained structure when moved with a sterile loop. Each plate was scored on a scale of 1 to 4 on the basis of the number of quadrants that contained colonies after incubation for 4 days.

During necropsy, lesions were recorded, and then half of the lung, spleen, skin, lymph node, and liver tissues were harvested into neutral-buffered 10% formalin, embedded in paraffin, sectioned onto glass slides, and stained with H&E. Sections were examined for evidence of bacteria and to evaluate the architecture, cellular changes, inflammatory infiltrates, necrosis, and vasculitis.

Data analysis—Data were maintained in a spreadsheet^e and analyzed by use of statistical software.^f Results of statistical tests were considered significant at values of P ≤ 0.05. Differences in phospholipase D transcription between rapidand slow-growing strains were assessed by use of t tests. Analysis of hematologic variables consisted of summary statistics followed by assessment for normality with a Kolmogorov-Smirnov test. Hematologic values in mice infected with rapid-or slow-growing bacteria were compared by use of t tests or the Wilcoxon signed rank test. Virulence of bacteria was interpreted as rapid growth on BHI agar, increases in phospholipase D transcription, or induction of se-vere clinical signs or death in a higher proportion of mice or at an earlier time point, compared with results for low-virulence strains.

Results

Bacterial culture—Plating of bacteria on BHI agar yielded pinpoint opaque white colonies for strains 115 and 312 that could be moved across the agar surface after 24 hours. By 48 hours, for these 2 strains, 10 or more white colonies that were at least 2 mm in diameter were evident in quadrant 4. In contrast, no growth was visible for strains 107 and 650 until > 48 hours, at which time pinpoint colonies were visible in quadrant 1. After 72 hours, bacteria were visible in quadrant 4, although strains 107 and 650 did not ever yield more than 5 individual colonies, and none of them were > 2 mm in diameter in quadrant 4. For further evaluation, strains 115 and 312 were grouped as rapid-growing strains, whereas strains 107 and 650 were considered slow-growing strains. The C_T values for phospholipase D transcription were low for all 4 C pseudotuberculosis strains (C_T values of 11.98, 11.14, 13.10, and 11.11 for strains 115, 312, 107, and 650, respectively). Phospholipase D transcription did not differ significantly (P = 0.67) between rapid-and slow-growing strains.

Route of inoculation—In the preliminary experiment, on day 2 after inoculation, the mouse inoculated via the ID route had lymphadenomegaly caudal to both pinnae with 2-mm nodules. The mouse inoculated via the IP route had no obvious clinical signs of disease, and the mouse in which the inoculum was applied to the skin surface had 1-mm dermal nodules in the area of the application, which were accompanied by hair loss. All mice remained active. On day 4 after inoculation, lesions on the mouse inoculated ID were approximately 3 mm in diameter, whereas the small lesions on the mouse to which the inoculum was applied to the skin surface did not become larger; the mouse inoculated IP remained clinically normal. On day 8 after inoculation, the mouse inoculated ID died, whereas the mouse inoculated IP was moribund and was euthanized. The mouse inoculated by application to the skin surface appeared healthy, except for lesions at the inoculation site. This mouse was euthanized on day 9 after inoculation.

During necropsy, the mice inoculated ID or IP had extensive lesions that involved 30% to 40% of the liver. The liver in each of these mice was swollen, pale, and yellow tinged, and multiple discrete white foci 1 to 4 mm in diameter were randomly scattered through-out the parenchyma and over the capsular surface. The spleen, heart, kidneys, and lungs were grossly normal. The mouse inoculated IP had a draining tract in the inguinal area that contained purulent exudate at approximately the site of inoculation. The tract penetrated from the dermis through the musculature of the hind limb to a distance of 0.75 cm in a cranial direction. The mouse inoculated by application to the skin surface had no obvious internal lesions but did have several small plaques in the dermis around the site of hair loss and bacterial inoculation. These plaques were not adhered to the subcutaneous tissues. On the basis of these findings, the ID route was chosen for the subsequent experiment.

Comparison among isolates—When comparing virulence among C pseudotuberculosis isolates for the 12 inoculated mice, 1 mouse inoculated with strain 312 (a rapid-growing strain) died on day 8 after inoculation. The remaining 11 mice were euthanized on day 9 after inoculation, and necropsy was performed on each mouse. All but one of the mice had obvious hepatic lesions. The liver of each mouse inoculated with a rapid-growing strain contained multiple, randomly distributed, focally extensive, sharply delineated but unencapsulated areas of coagulation necrosis (severe hepatic necrosis) that affected 30% to 40% of the parenchyma (Figure 1). Lesions were similar to those observed in mice in the preliminary study. Histologic evaluation of these areas confirmed a massive necrosis that affected all hepatocytes (regardless of zone) in multiple adjacent hepatic lobules. Numerous bacterial colonies were detected in the foci of necrosis. The lesions were unencapsulated, and there was minimal accumulation of mixed mononuclear and polymorphonuclear inflammatory cells within portal tracts. In contrast, the liver of each mouse inoculated with a slow-growing strain contained disseminated, randomly distributed foci of coagulation necrosis and neutrophilic hepatitis that affected < 10% of the parenchyma and involved only portions of each lobule without any obvious predilection for a specific zone. All mice inoculated with rapid-growing bacteria also had aural lesions, although they were often asymmetric. Aural lesions varied from a slight thickening to 1.5-mm discrete suppurative foci at the site of injection, which sometimes were accompanied by limited caseous necrosis with focally extensive edema. Edema and focally extensive caseous necrosis were evident in the dermis of all mice inoculated with slow-growing bacteria. Two mice inoculated with strain 115 had splenomegaly, and one of these mice had mesenteric lymphadenopathy. However, histologic lesions in lungs, spleens, and lymph nodes were not detected or were unremarkable.

View Full Size

Photomicrograph of a section of liver obtained from a mouse inoculated ID with rapid-growing virulent Corynebacterium pseudotuberculosis. Notice the representative hepatic lesions of massive necrosis that affected all hepatocytes (regardless of zone) in multiple adjacent hepatic lobules and the numerous bacterial colonies in the foci of necrosis. H&E stain; bar = 100 μm.

Citation: American Journal of Veterinary Research 70, 2; 10.2460/ajvr.70.2.257

• Download Figure
• Download figure as PowerPoint slide

More colonies were recovered on bacterial culture of tissues of mice inoculated with the rapid-growing strains than from mice inoculated with the slow-growing strains. Bacteria were isolated from the liver, spleen, lungs, and mesenteric lymph nodes of all mice inoculated with rapid-growing strains, whereas bacteria were isolated only from the liver and mesenteric lymph nodes of mice inoculated with slow-growing strains (Table 1). No bacteria were recovered from any tissues in 1 mouse inoculated with strain 312. There also were no lesions in the liver of that mouse, but it did have a draining tract at the site of bacterial inoculation caudal to the left pinna. In all inoculation groups, the liver was the site from which most bacteria were isolated, including for slow-growing strains.

Mice inoculated with rapid-growing C pseudotuberculosis strains had a significantly higher mean number of platelets and percentage of neutrophils but had a significantly lower mean percentage of monocytes (Table 2). There were no other significant differences in hematologic variables among the inoculation groups. Numbers of lymphocytes and RBCs, hemoglobin concentration, Hct, and mean corpuscular volume were within reference limits. On the basis of pathologic, clinical, and hematologic differences among strains, the rapidgrowing strains were interpreted to have higher virulence than the slow-growing strains.

Discussion

Infection with C pseudotuberculosis is a potentially lethal, relatively common clinical problem in horses and ungulates throughout the world, yet prevention and management of the disease, especially in horses, is challenging. Immunopathologic sequelae of infection are poorly understood, in part because of idiosyncratic responses to infection among animals, the existence of different C pseudotuberculosis strains with variances in virulence, and environmental variability in exposure and risk. A consistent, reproducible means of inducing clinically affected laboratory animals with internal abscesses associated with C pseudotuberculosis infection would be an important technique for interpreting and managing this disease in animals.

In the study reported here, we sought to develop and evaluate a method for inducing internal abscesses in mice by ID inoculation of C pseudotuberculosis. Mice in the study developed clinical signs of depression, were anorectic and unresponsiveness, and had hepatic necrosis and marked hematologic changes. The incubation period study was relatively short (approx 8 days in mice inoculated by the IP or ID routes), compared with 7 to 28 days inferred from epidemiologic data in horses.⁸ Hematologic changes were consistent with an acute course of disease, with no changes detected in erythrocyte numbers, hemoglobin concentration, or Hct. Five mice were thrombocytopenic, which was more severe in mice given the lower-virulence strains. In contrast, relative neutrophilia was detected more commonly in mice inoculated with high-virulence strains, whereas lymphocytopenia and monocytosis were detected in mice inoculated with either strain. Leukocytosis and neutrophilia have been described in affected horses, especially in conjunction with internal abscesses.⁹

In the preliminary experiment designed to choose an optimum route for inoculation, the mouse with a chemical dermal irritation followed by cutaneous application of C pseudotuberculosis failed to develop apparent systemic infection, which was in contrast to findings from another report.⁷ In that study,⁷ application of thioglycolate to BALB/c mice followed by cutaneous inoculation resulted in raised plaques on the skin that were reportedly quite distinct from the 1-mm cutaneous plaques observed in control mice with only thioglycolate application. Moreover, those mice developed infection as determined on the basis of positive culture results for subiliac lymph nodes and hepatic abscesses. Although cutaneous application of bacteria in the study reported here failed to induce infection, the anatomic distribution of lesions among mice inoculated ID and IP was similar (primarily targeting the liver), which is similar to results for mice inoculated IP and horses.^7,9

By use of the ID inoculation route, hepatic involvement was consistent. Lymphadenomegaly was also evident in the draining nodes of the ID inoculation site of many of the mice, which is consistent with lymphatic dissemination of the bacteria. Severity of the gross lesions differed remarkably in mice that received rapidgrowing bacteria, compared with the lesions in mice that received slow-growing bacteria. These results were also mirrored in the histologic lesions because mice inoculated with rapid-growing bacteria had severely affected livers, whereas mice inoculated with slow-growing bacteria had only small foci of hepatitis and necrosis. Multiple discrete accumulations of macrophages and neutrophils have been described in the livers of mice inoculated IP with C pseudotuberculosis.⁷

Findings in mice differ from findings in horses in several regards. Approximately 1% of infected horses reportedly develop ulcerative lymphangitis, 91% develop external abscesses, and 8% develop internal abscesses and bacteremia,⁹ compared with induction of liver lesions in 100% of the mice in the study reported here, even after ID inoculation at a remote site (ie, the pinna). The acute course of disease, reflected clinically and histopathologically in mice in our study, also is relatively distinct from that in horses. In horses, internal abscesses typically are encapsulated, with extensive aggregates of macrophages and lymphocytes around a necrotic center. However, it is possible that the lesions in the mice inoculated with slow-growing bacteria could have developed further over time, become encapsulated, and developed into lesions more similar to internal abscesses in horses. This suggested that use of slow-growing bacteria inoculated ID with a longer follow-up period could be an important next step in the development of this technique in mice.

To a considerable extent, results of quantitative bacterial culture also mirrored results of gross and histopathologic findings. The organ from which bacteria were cultured most commonly and in greatest numbers was the liver, followed by the lungs, spleen, and lymph nodes. In several mice, gross lesions were detected but bacterial culture yielded negative results, such as for lymph nodes in mice receiving strain 107 and the liver of a mouse receiving strain 650. Bacterial culture yielded positive results in the absence of gross lesions for the spleen in 4 mice, lungs in 3 mice, and lymph nodes in 6 mice. One mouse did not have positive results for bacterial culture, which suggested that infection may not have developed. More bacteria were recovered from more sites in mice that received rapid-growing bacteria, compared with bacterial recovery in mice inoculated with slow-growing bacteria.

The rapid-growing bacteria had consistently higher virulence than did the slow-growing strains, as determined on the basis of clinical, microbiologic, and pathologic findings; however, the cause of these differences is not known. The aroB and aroQ mutants with low virulence also have significantly lower growth rates.¹⁰ All 4 strains used in the study reported here produced comparable amounts of the extracellular toxin phospholipase D, which is an enzyme that hydrolyzes sphingomyelin in endothelium and leads to increased vascular permeability and increases the potential for bacterial dissemination.¹¹ This toxin is similar to the toxin of the brown recluse spider.¹² Mutant strains of C pseudotuberculosis that lack the ability to produce phospholipase D have a reduced ability to cause infection or chronic abscesses.¹³ The differences in virulence were reproducible despite the fact that all 4 strains originated from external abscesses of comparable severity in horses and had similar phospholipase D transcription. The virulence factor or factors responsible for these differences are not known.

In addition to the contribution of the bacteria to differential disease expression, the host and route of inoculation were important determinants of the outcome of inoculation. An ID route was used to mimic inoculation through mechanical transmission by arthropod vectors or via wound contamination. In diverse host-pathogen systems (including trypanosomes and spirochetes), the original presentation of antigens by dermal dendritic cells can have an important influence on the nature of the subsequent immune response.^14–16 Later events in infection, when clinical disease becomes evident, depend on a cascade of interdependent host-pathogen interactions starting as early as the first few hours after the pathogen is inoculated into the skin and antigen is presented by dendritic cells. An intriguing aspect of disease in mice in the study reported here was the finding that local abscesses developed but other pathologic changes were largely restricted to the liver (although bacteria were cultured in low numbers from some other organs, including the lungs). Thus, further exploration of early kinetics of infection and dissemination could help clarify the possible role of dermal immune activators in disease progression.

The use of rodents to study bacterial infectious diseases has the advantages of being efficient and relatively inexpensive, and it allows for an in vivo assessment of bacterial virulence. At least 5 studies^7,10,17–19 of C pseudotuberculosis infection in mice have been published; these studies typically were used as a means to evaluate management strategies for caseous lymphadenitis in sheep and goats. Overall, studies reveal that pretreatment with attenuated bacteria or immunomodulators (or both) could reduce but not eliminate clinical and pathologic sequelae, in part via induction of tumor necrosis factor-α and interferon-γ. We are aware of only 1 study⁷ in which investigators used a route of inoculation similar to the route tested in our study. In that study,⁷ investigators used cutaneous application of bacteria on depilated skin, which caused bacteremia, hepatic abscesses, and death in some mice. However, these findings were not reproduced in the mice of our study.

The use of ID inoculation of mice will be beneficial in the future assessment of the host immunologic response to C pseudotuberculosis–induced disease. Future studies of C pseudotuberculosis–associated internal abscesses should involve the use of additional mouse strains to evaluate specific immune defects associated with systemic disease. The C3H/HeJ mice are deficient in macrophage activation, T-helper–1 to T-helper–2 switching, and B-cell maturation, and another study¹⁷ revealed a role for tumor necrosis factor-α and interferon-γ in successful vaccination of mice against C pseudotuberculosis. The relative inability of C3H/HeJ mice to produce these cytokines could partially account for their propensity to develop internal abscesses. Thus, additional manipulation of mice could involve use of immunologically intact mice and knockout strains that lack more targeted immune functions as well as mice of various ages and that represent a diversity of other host risk factors. Induction of a more chronic course of disease than was detected in the study reported here could more accurately represent the course of the disease in horses. Ultimately, vaccination and management of C pseudotuberculosis–associated internal abscesses in horses will require better understanding of effective immune responses, including innate and adaptive immunity, cytokine responses, and specific responses to C pseudotuberculosis virulence factors.