A retrospective study of non-suppurative encephalitis in beef cattle from western Canada

Abstract

Non-suppurative encephalitis occurs sporadically in beef cattle in western Canada, leading to loss of animals. This retrospective study investigated the presence of viral, bacterial, and protozoal antigens or DNA in 37 western Canadian feedlot cattle with non-suppurative encephalitis for which a cause had not been identified. Cases were selected based on the age of the animal (> 7 months), and clinical history of recumbency and depression. The identification of rabies in 1 case stresses the importance of including this viral disease in the list of differential diagnoses. Because there was variation in the severity, distribution, and type of lesions, it is possible that there may be more than 1 cause, but failure to identify an infectious agent might also suggest that non-infectious agents could play a role.

Introduction

Disorders of the nervous system are important in cattle of all ages and are caused by infectious agents, or toxic, nutritional, neoplastic, and hereditary conditions (1,2). Idiopathic inflammatory diseases of the central nervous system (CNS), most of which correspond to non-suppurative encephalitides, occur sporadically. Non-suppurative encephalitis is a nonspecific term that embraces a series of microscopic changes such as infiltration of mononuclear leukocytes (lymphocytes, plasma cells, and histiocytes) into the Virchow-Robin space and neuropil, often accompanied by a glial reaction (1,3). This form of encephalitis can be caused by a variety of infectious agents; therefore, it is important to rely on sensitive diagnostic tests, such as immunohistochemistry (IHC) and polymerase chain reaction (PCR), in addition to other methods such as isolation of the agent to establish a more accurate diagnosis. The cause remains unknown in a great proportion of cases in cattle, pigs, cats, dogs, and humans, among other species (2–11).

This retrospective study investigated the presence of viral, bacterial, and protozoal antigens or DNA in 37 feedlot animals from western Canada with non-suppurative encephalitis for which a cause had not been identified.

Materials and methods

Thirty-seven cases of non-suppurative encephalitis from feedlot cattle, diagnosed from 1986 to 2010, and archived in the database of Prairie Diagnostic Services (PDS), Western College of Veterinary Medicine (WCVM), University of Saskatchewan were selected. Selection of animals was based on age (> 7 mo), clinical history of recumbency and depression, and a previous diagnosis of non-suppurative encephalitis for which a cause had not been determined. Hematoxylin and eosin (H&E) stained histopathological sections previously prepared from the central nervous system were reviewed and, for each case, 1 paraffin-embedded block containing 1 or more tissues, with the most severe lesions, was selected. New 5-μm thick sections from the selected blocks were stained with H&E. For each case, the grade of inflammation was categorized according to the estimated percentage of blood vessels affected and the number of cell layers that were present in the perivascular cuffs. Grades 1, 2, and 3 were assigned when the percentage of affected blood vessels was ≤ 25%, 25% to 50%, and 50% or more, respectively. A + was indicated when there were scant cells that did not form a circumferential ring or cuff around the vessel to one complete cell layer, ++ when 2 or 3 complete cell layers were present, and +++ when ≥ 3 complete cell layers were visible.

Immunohistochemistry

Sections (5-μm thick) were cut from each block and processed using standard immunohistochemical methods for avidin-biotin complex immunoperoxidase staining.

Immunohistochemistry was used to determine the presence of bovine viral diarrhea virus (BVDV; mouse 1:500) (IDEXX, One IDEXX Drive, Westbrook, Maine, USA), bovine herpesvirus type 1 (BoHV-1; mouse 1:2000) (Dr. Vikram Misra, WCVM); that would also react with bovine herpesvirus type 5 (BoHV-5), rabies (rabbit; 1:2000) (Animal Diseases Research Institute, Canadian Food and Inspection Agency, ADRI-CFIA, Nepean, Ontario), Listeria monocytogenes (rabbit 1:1000) (Difco, New Jersey, USA), Neospora sp. (rabbit 1:4000) (JP Dubey, USDA, Beltsville, Maryland, USA), Toxoplasma gondii (goat 1:10000) (VMRD Inc., Pullman, Washington, USA), and Sarcocystis spp. (rabbit 1:500) (Central Veterinary Laboratory, Weybridge, England). In 22 cases, immunohistochemistry for at least 1 of the aforementioned antigens had been requested at the time of submission. Immunohistochemistry for West Nile virus (WNV) (rabbit 1:2000) (Bioreliance Co, Rockville, Maryland, USA), was done in 5 cases in which the time of submission coincided with the usual seasonal appearance of positive cases in western Canada. Additional immunohistochemistry for Histophilus somni in 5 cases (rabbit 1:1000) (Dr. Andrew Potter, Vaccine and Infectious Disease Organization, VIDO, WCVM) and Mycoplasma bovis in 2 cases (mouse 1:800) (Veterinary Sciences Division, Sormont, North Ireland), was done based on the diagnosis of concurrent lesions, in which these 2 agents were suspected or isolated. Immunohistochemistry for parainfluenza virus type 3 (PI-3) (PDS) was done in 3 selected sections from 3 cases with severe inflammation (thalamus, midbrain, pons, and medulla oblongata). Formalin-fixed tissues were pretreated with protease for all antigens except for West Nile virus in which heat-induced epitope retrieval was applied.

Polymerase chain reaction (PCR)

Polymerase chain reaction (PCR) was done for Chlamydophila sp. in 36 cases and for ovine herpesvirus 2 (OHV-2) and caprine herpesvirus 2 (CpHV-2) in 3 cases with suspicion of vasculitis associated with malignant catarrhal fever (MCF). The DNA for PCR was obtained from paraffin-embedded tissues by standard procedures involving treatment with xylene and ethanol, sodium dodecyl sulfate, and proteinase K, extraction with phenol/chloroform, and precipitation with ethanol. For each block, if possible, a total of 5 to 10 sections (100 to 200 μm total), were obtained. Polymerase chain reaction amplification for ovine herpesvirus 2 and caprine herpesvirus 2 (3 cases) and Chlamydophila sp. (36 cases) DNA was done following published protocols (12,13).

Results

Thirty-seven histological sections of brain with the most severe lesions were selected for this study, including 15 from medulla oblongata, 10 from cerebrum, 4 from midbrain, 4 from thalamus, 2 from cerebellum, 1 from pons with cerebellar peduncles, and 1 from pons. The brainstem (midbrain, pons, and medulla oblongata) was selected in 17 cases for immunohistochemistry or PCR since it is the affected site for listeriosis and it is the best site for detection of rabies in cattle (14). The brain stem was not available or did not show inflammation in 20 cases. From these animals, the areas of the brain with the most severe inflammation were selected. The cattle ranged in age from 8 to 24 mo and included 19 males and 15 females. Gender was not recorded in the clinical history of 3 cases. Perivascular cuffing was present in all cases (Table 1), and the cuffs were chiefly composed of lymphocytes, plasma cells, and macrophages (Figure 1A). Within the perivascular cuffs of 9 sections, there were a few neutrophils which were admixed with few eosinophils in 4. Non-suppurative meningitis was identified in 14 cases with the predominant cell type being lymphocytes and plasma cells. The meningitis was categorized, subjectively, as mild in 8 cases, moderate in 2 cases, and marked in 4 cases. In 23 sections there was evidence of a glial reaction, in which microgliosis appeared to predominate (Figure 1B). Glial nodules were present in the neuropil of 8 cases and late stage of neuronophagia was clearly observed in 6 of those cases (Figure 1C). A focal area of gliosis along with edema and axonal degeneration, few associated neutrophils, and microthrombi were observed in the section from the pons with cerebellar peduncles. In 13 cases, less than 25% of the blood vessels were involved and scant lymphocytes were present in the Virchow’s space, sometimes not even forming a circumferential ring or cuff around the vessel. In 4 of these cases, along with another 5 (9 in total), concurrent lesions in other organs were described in the original diagnostic reports. Suppurative myocarditis was reported in 1 case and bronchopneumonia was present in 4 cases. In 1 of these 4 cases, polyarthritis, bronchopneumonia, and lymphocytic arteritis in the heart were also described. Immunohistochemistry for BVD was reported to be positive in the heart and Mycoplasma bovis was isolated from lesions in the affected lungs and joints. Bovine viral diarrhea virus antigen was not identified in the section of brain. In 1 of those cases, there was bronchopneumonia suspected to be caused by M. bovis; however, bacterial culture of the lung was not requested at the time of submission. In 4 other animals, fibrinous pleuropneumonia, suppurative myocarditis, ruminal tympany, and glomerulonephritis were previously diagnosed.

Table 1.

Grade of the inflammation and additional histopathologic findings in 37 cases of non-suppurative encephalitis [Modified from Bukovsky et al (6)]

		Grade of inflammation
Case number	Area of brain	Based on % of blood vesselsa	Based on number of cell layers in peri-vascular cuffb	Meningitis	Gliosis
1	Cerebrum	2	+	N	Y
4	Cerebrum	1	+	Y	Y
6, 32	Cerebrum	1	++	Y	N
12, 18, 27	Cerebrum	1	+	Y	N
16, 26	Cerebrum	1	+	N	N
20c	Cerebrum	1	++	N	Y
5c,d	Thalamus	3	+++	Y	Y
8	Thalamus	1	+++	N	Y
17	Thalamus	1	+	N	N
28c,d	Thalamus	1	+++	Y	Y
2c,d	Medulla oblongata	3	+++	Y	Y
7, 11, 23, 31, 33	Medulla oblongata	1	+++	N	Y
10e	Medulla oblongata	2	+++	N	Y
13, 30	Medulla oblongata	1	+	N	N
15c,d	Medulla oblongata	3	+++	N	Y
22, 37	Medulla oblongata	1	+++	N	N
24e	Medulla oblongata	1	+++	N	N
25, 35	Medulla oblongata	1	+	N	Y
3c,d	Midbrain	3	+++	Y	Y
9	Midbrain	1	++	N	N
19c,f	Midbrain	2	+++	Y	Y
36c,d	Midbrain	3	+++	Y	Y
14e	Pons with cerebellar peduncles	1	+++	Y	Y
34	Pons	1	+++	N	N
21	Cerebellum	1	+++	Y	Y
29	Cerebellum	1	+	Y	Y

^aPercentage of affected blood vessels: 1% = ≤25% of the blood vessels affected in the section, 2 = 25% to 50% of the blood vessels affected, 3 = >50% of the blood vessels affected.

^bNumber of cell layers in the perivascular cuffs: + scant number of cells not forming a complete concentric ring around the vessel to 1 complete cell layer, ++ 2 to 3 complete cell layers, +++ more than 3 complete cell layers.

^cGlial nodules present.

^dNeuronophagia observed.

^eVasculitis present.

^fDiagnosed with rabies.

Meningitis and gliosis: “Y” — present, “N” — absent.

Figure 1.

A — Medulla oblongata showing severe mononuclear perivascular cuffing composed of lymphocytes, plasma cells and histiocytes. Hematoxylin and eosin (H&E). Bar = 20 μm. B — Medulla oblongata showing multifocal areas of gliosis. H&E. Bar = 50 μm. C — Midbrain with glial cells surrounding and phagocytizing a neuron (late stage of neuronophagia). H&E. Bar = 50 μm. D — Midbrain showing detection of rabies antigen in the perikaryon of neurons. Avidin-biotin complex peroxidase.1:400. Bar = 50 μm. E — Pons — Marked perivascular cuffing composed of lymphocytes, histiocytes, and few neutrophils. H&E. Bar = 20 μm. F — Pons — Multifocal vacuolation in the neuropil and scattered spheroids in the white matter. Scattered neutrophils, erythrocytes and microthrombi. H&E. Bar = 50 μm.

Rabies antigen was detected in 1 case and the antigen-antibody reaction was mostly found in the perikaryon of neurons (Figure 1D). Bovine viral diarrhea virus, bovine herpesvirus (BoHV-1), L. monocytogenes, Neospora spp., Sarcocystis spp., T. gondii, West Nile virus (5 cases), and DNA for Chlamydophila spp., ovine herpes virus 2, and caprine herpesvirus 2 (3 cases with vasculitis) were not detected with immunohistochemistry or PCR, respectively. Additional immunohistochemistry for H. somni (5 cases), M. bovis (2 cases), and parainfluenza virus type 3 (3 cases) was negative.

Discussion

In this study, 36 of 37 cases of non-suppurative encephalitis from feedlot cattle tested negative for antigens or DNA of viral, bacterial, and protozoal origin that are known to cause encephalitis in cattle (4,12–23). Similar findings have been reported in previous retrospective studies in which the cause of non-suppurative encephalitis was not determined in the majority of cases (2–7,9–11). In a neuropathological and etiological retrospective study of sporadic non-suppurative meningoencephalomyelitis in Swiss cattle (11), Chlamydophila psittaci DNA was detected by PCR in 1 of 32 specimens, but immunohistochemistry did not detect chlamydial antigens. In another study of ruminants with encephalitis, the majority of which were non-suppurative, an etiology was not identified in 168/178 cases (5). Other animal species and humans are not exempt from this condition (3,6,8,10). In pigs for example, 36 of 38 natural cases of unknown etiology, non-suppurative encephalitis tested negative for 6 antigens (6).

The present study was retrospective so a methodical search for other lesions in other locations was not possible. Sections with the most severe lesions were selected and it was assumed that antigens or DNA should be present where lesions existed. Histological changes that tend to be associated with viral infections (1,3,10), such as meningitis, perivascular cuffing with predominance of lymphocytes and plasma cells, gliosis, and neuronophagia were all present in 6 cases; however, only 1 virus (rabies) was detected. It is not known whether the viruses that were investigated could have been the cause of non-suppurative encephalitis, but a number of the viral agents have been reported to cause non-suppurative encephalitis in cattle.

It is unlikely that a virus or other infectious agent which requires a vector for transmission, such as mosquitoes, could be playing a role, since most cases (21 in total) occurred in the winter. Only 5 cases, in which the time of submission coincided with the usual seasonal incidence of positive cases of WNV in western Canada, were tested for this virus. Although all of these cases were negative, reports of WNV in other ruminants exist (17), hence the importance of determining if cattle can also be affected in endemic areas. In the human and veterinary literature, authors have suggested that an immune mechanism may play a role in the pathogenesis of sporadic non-suppurative encephalitis of unknown etiology (3,5,6,8,10,24,25). It is thought that an infectious agent, most likely a virus, can induce a series of immune-mediated events in the brain that may clear the agent by the time diagnosis is attempted (26). The most frequently postulated mechanism of viral-induced immune damage mediated by T-cells is through molecular mimicry to self antigen such as glial fibrillary acidic protein, myelin-associated glycoprotein, or myelin proteolipid protein (26,27).

When the animals in the present study arrived at the feedlot, they were vaccinated against some of the viruses that were investigated. It is assumed that vaccine antigens, if responsible for the disease, might still be detected. In the cases that were studied, the time interval between vaccinations, during the fall when these animals enter the feedlot, and when clinical signs appeared were variable; however, 21 of the 37 cattle had clinical signs in the winter. In humans, post-vaccination encephalomyelitis usually affects children within 1 mo of vaccination (28). The immune-mediated damage to the nervous system is either by virus-induced molecular mimicry or contamination with CNS tissue in which the virus is propagated (29). In humans, post-infection and post-immunization encephalomyelitides are rare inflammatory diseases characterized by demyelination, which was not a feature in our cases (26,30). The disease in humans is disseminated and severe, whereas the lesions in the brain examined herein were multifocal and varied in severity.

In 13 cases, less than 25% of the blood vessels were involved and scant lymphocytes were present in the Virchow’s space, sometimes not even forming a complete cell layer. Controversy exists regarding the significance of these changes, especially when they are observed in brains from animals which die or are euthanized because of non-neurological disease. However, the animals herein were recumbent despite subtle inflammation. It is important, however, to remember that ruminants may show slight or no microscopic changes despite being affected by some neurological diseases, such as rabies (1). In 9 of the 13 cases, lesions were found in other organs, but whether these were related to non-suppurative encephalitis is not known. During systemic disease, the permeability of the blood/brain barrier can be altered through the action of cytokines and other inflammatory mediators, such as prostaglandins, allowing the influx of inflammatory cells into the brain (31). Also, with a strong immunological response in the body without involvement of the brain, elevated numbers of T-cells can be detected in the CNS (31). Histophilus somni was considered as a differential diagnosis for the non-neurological lesions in 4 of the 9 cases and there was a 5th case in which H. somni was suspected, but in which there was no concomitant disease. Suppurative myocarditis, which is a major manifestation of H. somni infection in western Canada, was described in the original report of 1 case (1). Neutrophils and microthrombi along with scattered foci of edema and axonal degeneration, as well as a prominent mononuclear perivascular cuffing with very few neutrophils were observed in the neuropil of 1 case (5th case with no other concomitant disease) (Figures 1E, F). In suspected chronic stages of H. somni infections, thought to have been influenced by antibiotic therapy, host response and virulence of the organism amongst other factors, inflammatory changes in the brain are not characterized by a suppurative response (32). Based on the absence of positive immunohistochemical results, H. somni was ruled out as a possible cause of non-suppurative encephalitis in those 5 cases. However, due to the common occurrence of H. somni infections in feedlot cattle, it would have been interesting to test for this antigen in the other 32 cases. Two animals were tested for M. bovis but these 2 animals had been treated with antibiotics. Based on the negative results with immunohistochemistry, M. bovis was ruled out as the cause of non-suppurative encephalitis. In 1 of these cases, BVDV antigen was detected in the heart, which could have been a predisposing factor for the M. bovis infection (pneumonia and arthritis) (33); however, BVDV was not identified in the section of brain examined. Bovine viral diarrhea virus has been associated mostly with CNS congenital defects due to transplacental infection; however, BVDV type 2 was detected and isolated from a 15-month-old heifer with neurological signs and a non-suppurative encephalitis in the absence of other clinical signs usually associated with BVDV. A neurovirulent strain of the virus was proposed, but it was not possible to determine whether the infection of this heifer was acquired congenitally or postnatally (16).

Apart from rabies and Listeria monocytogenes, agents that are known to affect younger animals include bovine herpesvirus types 1 and 5, BVDV, PI-3, T. gondii, N. caninum, Sarcocystis spp., and Chlamydophila spp. (15–23). Even though these infectious agents appear not to be the cause of non-suppurative encephalitis in these animals, the study of these agents was important due to the possibility of animals having persistent or latent infections. For example, BoHV-1 was isolated in 4 animals with non-suppurative encephalitis, ranging in age from 6 to 48 mo and feedlot management was considered a stress factor for the activation and dissemination of the virus (23).

This study confirms that it is still difficult to determine the etiology of non-suppurative encephalitis, even through use of extensive diagnostic techniques. Failure to identify an infectious agent might suggest that non-infectious causes could be playing a role. Further understanding of immunological mechanisms that may occur in the CNS will likely provide relevant information to the understanding of the pathogenesis of non-suppurative encephalitis (3,5,6,10). The identification of rabies stresses the importance of including this viral disease in the list of differential diagnoses of non-suppurative encephalitis in feedlot cattle from western Canada.