This study was presented in two parts in poster form at the American College of Veterinary Internal Medicine Forum, Dallas, TX, May 2002, and San Antonio, TX, June 2008. This study was performed at Colorado State University (CSU), Veterinary Teaching Hospital, Fort Collins, Colorado. Cases referred to in this study were seen at CSU, University of Wisconsin—Madison Veterinary Teaching Hospital, or by surrounding local veterinarians.
Corresponding author: Dr Natalee Holt, DVM, Tufts University Cummings School of Veterinary Medicine, 200 Westboro Road, North Grafton, MA 01760; email: email@example.com.
Background: Acute canine polyradiculoneuritis (ACP) is considered to be an animal model of the acute axonal form of Guillain-Barré syndrome (GBS) in humans. Various antecedent events have been associated with GBS, including bacterial or viral infection. The relationship between ACP and previous infection requires additional attention.
Hypothesis: We hypothesized a relationship between ACP and serological evidence of exposure to Ehrlichia canis, Borrelia burgdorferi, Toxoplasma gondii, Neospora caninum, Campylobacter jejuni, and canine distemper virus (CDV).
Animals: Eighty-eight client-owned dogs, 44 with ACP, 44 age-matched controls.
Methods: Retrospective study with stored serum samples. Serum antibodies against the target organisms were measured with commercially available assays. Sera from dogs with and without ACP that were positive for T. gondii IgG by ELISA were assayed by an IgG heavy chain-specific, Western blot immunoassay.
Results: Dogs with ACP (55.8%) were more likely to have T. gondii IgG serum antibody titers than dogs without ACP (11.4%). Serum antibodies from 8 affected dogs and 11 control dogs bound to T. gondii antigens with apparent molecular masses of 67, 61, 58, 45, 33, 24, 9, and 6 kDa. An antigen with an apparent molecular mass of 36 kDa was recognized by 2 dogs with ACP but none of the control dogs.
Conclusions: Results of this study suggest that ACP in some dogs, like GBS in some humans, may be triggered by T. gondii and a prospective study should be performed to further evaluate this potential association.
Acute canine polyradiculoneuritis (ACP) is the most commonly recognized peripheral neuropathy in dogs.1 Dogs with this disease frequently present with an acute ascending tetraparesis of varying severity, dysphonia, occasionally facial nerve weakness, and, in severe cases, respiratory compromise. A majority of dogs in the United States that develop ACP have a history of contact with a raccoon 7–14 days before developing clinical signs, thus the commonly used term “coonhound paralysis.”2 Others have identical clinical signs, electrodiagnostic evaluations, and histopathology with no history of raccoon exposure, suggesting that other predisposing factors exist.3
In previous histopathological studies, the ventral nerve roots and spinal nerves have been most severely affected in ACP, with minimal dorsal nerve root pathology. Demyelination and axonal destruction are present, with variable degrees of inflammation.4,5 Electrophysiologic studies have confirmed these findings and suggest a peripheral, predominantly motor axonopathy with the most severe axonal and myelin damage at the level of the ventral nerve root.6 Based on these electrodiagnostic and histopathological studies, ACP is considered an animal model for the acute axonal form of Guillain-Barré syndrome (GBS) in humans.7
GBS has been shown to be associated with a variety of infectious processes and events thought to trigger the immune system.8–14 Antecedent factors for which there is strong evidence of an association with GBS include Mycoplasma pneumoniae,8,15–19Campylobacter jejuni,18,20–23 HIV,24 cytomegalovirus,18,25,26 Epstein-Barr virus,18Borrelia burgdorferi,27 vaccination (rabies,28 swine influenza 29), and surgical procedures.30 In addition, Toxoplasma gondii has been associated with the syndrome in 2 people.31,32 Several theories have been proposed to explain how the immune system, stimulated in response to an inciting event, subsequently causes polyradiculoneuritis. An immune response may be mounted against the offending agent, which is antigenically similar to epitopes in peripheral nerve myelin or the nodes of Ranvier. The peripheral nerve then is targeted by the immune response via a bystander effect.32 Infection with specific infectious agents has been associated with distinct clinical variants of GBS, suggesting that the targeted antigens are not equally distributed throughout the nervous system, and that specific epitopes are associated with particular infectious agents. For example, C. jejuni gastroenteritis is associated with antibodies to the gangliosides GM1 and GD1b in peripheral nerve myelin, and the motor form of GBS. Cytomegalovirus is associated with antibodies to the GM2 ganglioside and the sensory form of the disease.18
It is unknown what event triggers the immune response in dogs with ACP and no history of exposure to raccoon saliva. As has been shown in GBS, an association may exist between exposure to certain bacteria, viruses, or protozoans and onset of ACP. The purpose of this study was to assess associations between ACP and serologic test results for a selected group of infectious agents with commercially available assays commonly utilized by practicing veterinarians in the United States.
Materials and Methods
The study was performed retrospectively with stored serum samples from dogs with suspected ACP and clinically ill dogs without neurological disease evaluated between 1999 and 2001. Each of the 44 dogs with suspected ACP received a case consultation by an ACVIM board-certified neurologist (P.C.) who concluded that the history and neurological examinations were consistent with the syndrome. In addition, electrodiagnostic evaluations (n = 19), lumbar cerebrospinal fluid (CSF) analysis (n = 15), or both electrodiagnosis and CSF analysis were performed (n = 15) for some of the dogs with suspected ACP. The initial serum samples were drawn at various time periods after the onset of clinical signs (range, 1–14 days). Additional samples collected during the course of illness were available for 6 dogs. Dogs were examined at Colorado State University Veterinary Teaching Hospital, University of Wisconsin—Madison Veterinary Teaching Hospital, or by local veterinarians in these 2 states.
To identify a control group, the records system at the Center for Companion Animal Studies and the Diagnostic Laboratory at Colorado State University was searched sequentially for dogs from Wisconsin and Colorado that had sera submitted for testing between 1999 and 2001. Samples (n = 44) were selected to match the state of origin and ages of the dogs with suspected ACP. The submission form from all control animals had to have clinical illness described, but cases with mention of neurological disease were excluded.
Sera from both groups of dogs were stored at −20 or −80°C until evaluated in this study. The majority of the samples from dogs with suspected ACP had been frozen and thawed previously for use in other studies by one of the authors (P.C.), whereas the samples from the control group had not. The Colorado State University Veterinary Diagnostic Laboratory is accredited (full service, all species) by the American Association of Veterinary Laboratory Diagnosticians meeting standards based on ISO17025/OIE (http://dlab.colostate.edu/). The laboratory is also a member of the National Animal Health Laboratory Network. The following antibody assays were performed on all samples by the trained staff of this facility by standard operating procedures maintained on file: Ehrlichia canis (indirect fluorescent antibody assay [IFA]; positive cutoff titer >1 : 80), B. burgdorferi (ELISA; positive cutoff titer >1 : 64), Neospora caninum (IFA; positive cutoff titer >1 : 100), C. jejuni (ELISA; positive cutoff titer >1 : 100), T. gondii IgG and IgM (ELISA; positive cutoff titer >1 : 64), and canine distemper virus (CDV) (serum neutralization; positive cutoff titer >1 : 20). For some sera from dogs with suspected ACP (n = 41), an ELISA for antiraccoon saliva antibodies was performed with a 1 : 2,000 dilution of pooled raccoon saliva as the antigen source.33,34 Control dogs were not tested for antiraccoon saliva antibodies.
For use in this study, a T. gondii IgG Western blot immunoassay was titrated for use with dog serum based on a previous report in cats.35 In this assay, T. gondii (RH1 strain) tachyzoites were used as the antigen source and the secondary antibody was goat anticanine IgG (heavy chain-specific). In this experiment, sera from 24 dogs with suspected ACP and 12 clinically ill control dogs that were positive in the T. gondii IgG ELISA were assayed to determine whether there was a difference in antigen recognition patterns between groups. A single investigator determined the apparent molecular masses of the recognized antigens with digital images of the blots and a commercially available software program.a
Serum antibody prevalence rates were compared between dogs with ACP and control dogs by Fisher's exact test. Comparisons between ACP and control groups for CDV titer magnitude were made by transforming reciprocal titers by natural logarithm and then analyzing by analysis of variance. Statistical significance was defined as P < .05. Results of WB are presented descriptively.
In both groups, 41 dogs were from Wisconsin and 3 dogs were from Colorado. Of the dogs with suspected ACP, 21 (47.7%) had documented raccoon exposure 7–14 days before onset of neurologic signs, whereas the remainder of the dogs either had no raccoon exposure or an unknown environmental history. Antiraccoon saliva antibodies were detected in serum from 28 dogs (68.3%) with suspected ACP. The control dogs were not analyzed in this assay and statistical comparisons between groups could not be made. The dogs in both groups ranged in age from 1 to 11 years. Of the dogs with suspected ACP, 16 dogs were male (9 castrated males), 9 were female (7 spayed), and complete signalment was not available for 20 dogs. The following breeds were represented in the group of dogs with suspected ACP: Coonhound (8), German Shepherd (6), Doberman Pinscher (4), Labrador Retriever (4), mixed breed (4), and other breeds (18).
Serological test results from both groups of dogs are summarized in Table 1. T. gondii IgG antibodies were detected by ELISA in significantly more dogs with suspected ACP than in control dogs (P= .0001). In the 6 dogs with suspected ACP and repeated samples available for assay, 2 dogs that were negative for T. gondii antibodies in the initial sample became positive on a sample collected either 24 or 29 days after the 1st sample. T. gondii IgG antibodies were detected in the serum of 15 of 28 dogs (53.6%) with suspected ACP that were positive for antiraccoon saliva antibodies and in the serum of 5 of 13 dogs (38.5%) with suspected ACP that were negative for antiraccoon saliva antibodies. These results were not significantly different.
Table 1. Seroprevalence rates for select infectious disease agents in dogs with (n = 44) and without (n = 44) clinical evidence of acute canine polyradiculoneuritis (ACP).
Fishers Exact-Test P Value
Twenty-four samples were available for this assay.
Both of these samples were also positive for T. gondii IgG.
B. burgdorferi antibodies were detected in more control dogs than dogs with suspected ACP (P= .0011). Sera from 2 dogs with suspected ACP and B. burgdorferi antibody titers >1 : 128 were assessed by Western blot immunoassay and shown to have banding patterns consistent with vaccination.b The number of dogs that were seropositive for CDV was not statistically different between groups.
In the Western blot immunoassay, T. gondii antigens were recognized by sera of 13 of 24 of the dogs with suspected ACP and T. gondii IgG detected in serum by ELISA (titer distributions = 1 : 64 [19 dogs]; 1 : 128 [5 dogs]) and by 11 of the 12 T. gondii IgG ELISA positive control dogs (titer distributions = 1 : 64 [7 dogs]; 1 : 128 [4 dogs]; 1 : 512 [1 dog]). There were 9 T. gondii immunodominant antigens recognized by the canine sera, including antigens with apparent molecular masses of 67, 61, 58, 45, 36, 33, 24, 9, and 6 kDa. The most commonly recognized antigens were the 45 and 33 kDa antigens (Table 2). The 36 kDa antigen was recognized by sera from 2 dogs with ACP and none of the dogs without ACP (Fig 1). There were no antigens recognized by sera from control dogs that were not also recognized by sera from dogs with ACP.
Table 2. Apparent molecular masses of Toxoplasma gondii antigens recognized by serum antibodies from dogs with acute canine polyradiculoneuritis (ACP) and control dogs as determined by T. gondii IgG Western blot immunoassay.
Molecular Mass (kDa)
Dogs with ACP
Dogs without ACP
After the T. gondii results were available, we attempted to contact all of the owners of dogs with ACP that also were T. gondii seropositive. Of the 8 owners that responded, all stated that their dog was fed a diet consisting of commercially processed dog foods.
In this study, we arbitrarily selected infectious agents that commonly infect dogs, have been associated with neurological disease, and have commercially available serological assays available that could be used on stored sera. A significant association between dogs with suspected ACP and the presence of T. gondii IgG in serum as determined by ELISA was made (P= .0001). However, as the study was performed retrospectively and the control groups were matched only by age and state of origin, the results should not be used to conclude a cause and effect relationship exists. In previous studies of hunting dogs that develop ACP, increased exposure to raccoons has been the most common underlying factor predicting this disease. Hunting dogs are more likely to ingest raw meat or prey, based on their lifestyles, and thus may be more likely to have previously ingested T. gondii organisms than control dogs. Although none of the owners of ACP dogs that responded to our questions concerning feeding habits intentionally fed raw meat or saw their dogs eat raw meat or prey animals, this information could have been inaccurate and similar information was not available for the control dogs. T. gondii has been shown to be endemic in the raccoon population in the United States and Canada, including the state of Wisconsin.36–38 Thus, increased exposure to raccoons could increase the risk of exposure to T. gondii as well. However, in this study, there was no difference in T. gondii IgG seroprevalence rates between dogs positive or negative for antiraccoon saliva antibodies in serum. A prospective study should be performed to further evaluate this association.
In the present study, T. gondii IgG titers, but not IgM titers, were commonly detected. This suggests that the dogs were not infected with T. gondii in the 7–10 days preceding the onset of neurologic signs. Some dogs do not develop a rise in IgM antibody during the acute phase of infection with T. gondii.39 However, it is unlikely that such a large percentage of the ACP dogs would fall into this category. It can take 4–6 weeks for an IgG titer to develop after infection with T. gondii. This suggests that the seropositive dogs in this study had been infected for at least several weeks before development of clinical signs of ACP, which would be longer than the typical time period between initial insult and resulting immune-mediated peripheral nerve pathology. The chronicity of infection is supported by the 2 dogs with follow-up samples that seroconverted between 24 and 29 days after initial sampling. Neither of the dogs had a positive IgM titer, indicating lack of an acute infection. Alternatively, it is possible that the initial titers were falsely negative titers or the follow-up samples were falsely positive. One of the dogs had 2 follow-up samples that had identical IgG titers, decreasing the likelihood that these titers were falsely positive.
In most cases of T. gondii infection, there is a chronically increased IgG titer because of continued presence of T. gondii antigen, even in the absence of clinical signs.40–42 This titer can remain increased for years. The ACP dogs in this study could have been infected with T. gondii years before developing neurologic disease, which would not fit the time period typically seen between infection and immune response resulting in clinical disease. Assuming that the immune response is mediating the peripheral nerve pathology, 1 possible explanation for the discrepancy is that the clinical signs occur when the immune response occurs, not necessarily 7–14 days after infection with the inciting organism. Another possible explanation is that the disease onset may relate more to periodic and repeated recrudescence of the disease with escalating sensitization of the immune system. This pathogenesis has not been documented with any infectious agents associated with GBS in humans. The peripheral nerve pathology in GBS and ACP is thought to be because of molecular mimicry between the inciting agent and proteins or glycolipids in peripheral nerve myelin or the axonal membrane at the nodes of Ranvier. If T. gondii does have a causal association with ACP via molecular mimicry, it is possible that it is only a particular stage in the life cycle that has antigenic similarity to structures in peripheral nerve myelin or the axolemma. This may result in a longer delay between actual infection with the organism and the component of the immune response that targets the peripheral nerves via bystander effect. Consequently, there may be a longer time period between exposure to the agent and neurologic signs. It may be worthwhile prospectively to perform multiple T. gondii titers to document any increase in IgG titer, which, if present, would support more recent exposure than was demonstrated in the present study.
In humans, infection with certain agents is associated with development of different forms of GBS. C. jejuni is strongly associated with the acute motor axonal form, whereas infection with cytomegalovirus is associated with the sensory form of GBS. We included C. jejuni in this study based on the fact that most dogs with ACP have a severe motor axonopathy, which most closely resembles the acute axonal form of GBS in humans. The hypothesized mechanism for GBS in humans is dual recognition of 1 B- or T-cell receptor with the lipopolysaccharide of C. jejuni and oligosaccharide component of the GM1 ganglioside in nerve tissue.43
Although C. jejuni is the most commonly associated organism in humans with GBS, there have been 2 case reports of humans developing GBS after exposure to T. gondii.31,32 In this study, we hypothesized that given the complexity of the Toxoplasma organism, there is a potential for molecular mimicry to occur between the surface of T. gondii and the glycoconjugates of canine nerve tissue similar to that conjectured in GBS in humans. One way to confirm the similarity in pathogenesis between ACP and GBS would be to identify a difference in antigen recognition patterns between animals both with and without a molecular mimicry disease process. The majority of the T. gondii antigens were recognized by dogs with and without ACP. The only T. gondii antigen recognized in dogs with ACP that was not also recognized by the control dogs was the antigen at 36 kDa. A 35 kDa antigen was reported as an IgG target antigen in cats experimentally infected with T. gondii.35 A 36 kDa antigen was reported not only in an extract of T. gondii tachyzoites,44 but also in pregnant women with clinical signs of toxoplasmosis.45 In sheep, cattle, and goats with toxoplasmosis, there are dominant antibody responses against proteins at 32–35 kDa.46 This antigen recognized in other species may be the same antigen recognized in the ACP dogs in our study. However, the identified molecular weight is not consistent with the molecular weight of the GM1 ganglioside, which is 1.5 kDa; however, there could be an epitope on the antigen similar to that on the ganglioside.
T. gondii antigens were detected by Western blot immunoassay in sera from 13 of the 24 dogs with suspected ACP that were T. gondii IgG antibody positive by ELISA. This most likely is related to the decreased sensitivity of the assay compared with ELISA which has been documented with other species and organisms.35 Most of the dogs with suspected ACP had the lowest possible T. gondii IgG titer of 1 : 64. In addition, the majority of the samples assayed in the ACP group had been frozen and thawed an unknown number of times, which may have affected the results of the Western blot immunoassay. Even with these limitations, 2 ACP-positive dogs had different antigen recognition patterns than control dogs. Additional work is indicated to determine whether this finding is reproducible in a larger sample set and to determine whether this antigen is immunologically similar to those associated with GBS in humans.
B. burgdorferi antibodies were detected in more control dogs than dogs with suspected ACP. However, the assay used as the screening procedure in this study cannot differentiate vaccine-induced antibodies from antibodies induced by natural exposure, and the B. burdorferi vaccine history of the 2 groups of dogs was not known. The same problem exists for the interpretation of the CDV titers in this study. In the 2 dogs with suspected ACP assessed by B. burgdorferi Western blot immunoassay, results suggested antibodies induced by vaccination. In future studies of ACP in dogs, B. burdorferi serological assays that can distinguish vaccination from natural exposure should be performed because it is possible that either natural infection or vaccination could be associated with neurological events such as ACP. Recent vaccination has been shown to be a cause of acute polyradiculoneuritis in people, dogs, and chimpanzees.47–49 Because of the retrospective nature of the study, timing of all vaccinations (including that for rabies) with respect to onset of ACP is unknown and should be investigated in a future prospective study.
The clinical, neurologic, and electrophysiologic examinations of the dogs in this study with confirmed raccoon exposure were not compared with dogs that may have developed ACP because of some other cause. The majority of dogs with ACP show clinical signs of an acute motor axonopathy. In the future, it may be worthwhile to compare these findings in raccoon-exposed versus unexposed ACP patients to determine if different subtypes of the disease also occur in dogs. The results of this study support the performance of a prospective study of ACP in dogs and further evaluation of a potential role of T. gondii in the syndrome.
aGel Doc 1000 System with Quantity One software; Bio-Rad Laboratories, Hercules, CA
bSamples submitted to New York State Veterinary Diagnostic Laboratory, Ithaca, NY
This study was supported by a grant from the College Research Council and the Center for Companion Animal Resources at Colorado State University.