Investigation of the presence of specific neural antibodies in dogs with epilepsy or dyskinesia using murine and human assays

Abstract Background Autoimmune mechanisms represent a novel category for causes of seizures and epilepsies in humans, and LGI1‐antibody associated limbic encephalitis occurs in cats. Hypothesis/Objectives To investigate the presence of neural antibodies in dogs with epilepsy or dyskinesia of unknown cause using human and murine assays modified for use in dogs. Animals Fifty‐eight dogs with epilepsy of unknown cause or suspected dyskinesia and 57 control dogs. Methods Serum and CSF samples were collected prospectively as part of the diagnostic work‐up. Clinical data including onset and seizure/episode type were retrieved from the medical records. Screening for neural antibodies was done with cell‐based assays transfected with human genes for typical autoimmune encephalitis antigens and tissue‐based immunofluorescence assays on mouse hippocampus slices in serum and CSF samples from affected dogs and controls. The commercial human und murine assays were modified with canine‐specific secondary antibody. Positive controls were from human samples. Results The commercial assays used in this study did not provide unequivocal evidence for presence of neural antibodies in dogs including one dog with histopathologically proven limbic encephalitis. Low titer IgLON5 antibodies were present in serum from one dog from the epilepsy/dyskinesia group and in one dog from the control group. Conclusion and Clinical Importance Specific neural antibodies were not detected using mouse and human target antigens in dogs with epilepsy and dyskinesia of unknown origin. These findings emphasize the need for canine‐specific assays and the importance of control groups.


| INTRODUCTION
Epilepsy is the most common chronic central nervous system (CNS) disease in dogs affecting 0.43% to 0.82% of dogs admitted to veterinary practices. [1][2][3][4] A genetic cause is frequently suspected in dogs with idiopathic epilepsy, but causal epilepsy genes are only identified for rare genetic idiopathic epilepsies with an onset in young dogs corresponding to human childhood and juvenile epilepsies. 5 The etiology of idiopathic epilepsy with onset in adult dogs remains largely unknown.
There might be a complex and multifactorial etiology with an interaction of multiple genetic risk variants and environmental factors. 6 Autoimmune mechanisms have increasing relevance in humans as a distinct etiologic category offering new therapeutic approaches. [7][8][9][10] Characteristic clinical features are sudden onset of psychiatric symptoms with rapid progression to epileptic seizures or movement disorders which could be associated with signal changes on magnetic resonance imaging (MRI), cerebrospinal fluid (CSF) pleocytosis, and the presence of neural antibodies in serum or CSF. 11 Further proof for the involvement of autoimmune mechanisms comes from positive outcomes of therapeutic trials with immunosuppressives. 9,12 Therefore, the ILAE epilepsy classification includes immune epilepsy as a distinct entity, 13 and the ILAE autoimmunity and inflammation task force recommend distinguishing between autoimmune-associated epilepsy and acute symptomatic seizures secondary to autoimmune encephalitis. 9 In animals, there is unequivocal evidence for autoimmune mechanisms in cats with limbic encephalitis (LE), orofacial seizures, and suggestive MRI and pathology findings associated with antibodies directed against leucine-rich glioma inactivated protein 1 (LGI1). The disorder in cats thus parallels anti-LGI1 LE in humans. 14-17 A further report refers to a cat with LE in association with deleted colorectal carcinoma (DCC, also known as netrin-1 receptor) autoantibodies. 18 Reports on dogs are scarce. [19][20][21][22] N-methyl-D-aspartate receptor (NMDAR) antibodies are suggested in the CSF of two dogs with meningoencephalitis of unknown etiology (MUE) and one dog with multifocal signs of neurologic disease and an abscessed maxillary tooth, and in the CSF of a polar bear (Ursus maritimus) that drowned because of epileptic seizures and NMDAR-antibody (AB) mediated encephalitis. 20,23 Furthermore, positive effects of prostaglandin G/H synthase 2 (PTGS2; also known as cyclooxygenase-2, COX-2) inhibitors in single dogs with IE and increased T helper 17 cells (Th17) support neuroinflammatory mechanisms in idiopathic epilepsy in dogs. 24,25 This study aims to screen for neural antibodies in dogs with epilepsy or suspected dyskinesia of unknown origin with commercial diagnostic murine and human assays, modified for use in dogs. We hypothesized that we would detect cross-reacting antibodies.  One dog had a history of surgical removal of an abdominal teratoma and encephalomyelitis 2 years before the onset of epilepsy (not counted here). An immune-mediated comorbidity was an exclusion criteria for control dogs.
Dogs were included in the epilepsy group if ≥2 generalized or focal epileptic seizures occurred ≥24 hours apart. 26 Paroxysmal dyskinesia was considered based on characteristic clinical features of episodes and review of video footage. 27 Episodes which were difficult to classify were also considered in this group ( Serum and CSF were collected as part of the routine diagnostic work-up. MRI was performed in all dogs with epilepsy or dyskinesia except six dogs with clinical signs and videos suggestive of idiopathic head tremor syndrome described previously (Table 1). 29 Cerebrospinal fluid was collected under general anesthesia by atlanto-occipital puncture. Routine CSF analysis (total nucleated cell counts, cytospin differential cell count, protein content, infectious disease tests as indicated) was performed. In most cases, CSF and serum (100 μL) was shipped to the immunodiagnostic laboratory, specialized for neural AB testing in human medicine, for immediate AB testing after arrival. Additional archived samples, which had been stored at À20 C or À80 C until testing, were transported frozen on dry ice.
Screening for neural antibodies was done on a tissue-based assay (TBA) on unfixed sagittal mouse brain slices (hippocampus, brain stem, and cerebellum; Euroimmun, Lübeck, Germany), and, secondly, by a

| MRI and CSF findings
Subtle MRI changes were evident in 53.8% of the dogs (28/52). Brain MRI revealed regional T2-or T2-FLAIR hyperintensity in 13 dogs  The study was designed as a screening study using available human and murine assays. The study did not provide evidence for neural antibodies including NMDAR antibodies in dogs with epilepsy of unknown cause or dyskinesia with consideration of autoimmune epilepsy with human and murine assays. Thus, with the present diagnostic methods the question remains yet unresolved whether autoimmune epilepsy is a rare condition in dogs or whether the applied test methods developed for use in humans are not reliable for investigation of dogs. Considering the selection of the dogs, 1 concern is that the results could have been negative because of methodological issues, that is, that dog antibodies did not bind to human and mouse GABAAR AB, serum titer 1:320, CSF titer 1:2 and a corresponding staining pattern on mouse brain. This dog recovered upon steroid treatment from its acutely emerging and progressive signs like tonicclonic epileptic seizures, episodic hyperexcitability alternating with episode of stupor, and intermittent circling behavior. 22 Even though the respective genes have amino acid sequences that are to ≥94% identical between humans and dogs, 20 our results highlight the need for cross-species validation of antibodies and target antigen epitopes, considering that the specificity of the AB is a critical part of IF-based methods, and the requirement to confirm positive results from IF and diagnostic assays with other or multiple detection methods. We attempted this with western blots but surprisingly the commercial NMDAR antibody failed to recognize canine NMDAR antigen despite being marketed for dogs ( Figure S3). It is a drawback that further confirmation, for example, with TBA failed in our IgLON5 AB positive dogs, and other detection methods, for example, gel electrophoresis was not pursued. 39 In the present study, all samples were tested in a two-step procedure, first by TBA, a screening test for unknown autoantibodies on unfixed sagittal mouse brain slices, and second by CBA, assuming that the TBA could serve as a screening test for yet unknown autoantibodies. 8 40 Furthermore results might also depend on the interpretation by the authors. Misinterpretation of non-specific cell staining as indicating antibody-positivity has been criticized previously. 42 We rated immunofluorescence in a stringent way. Figure 1C of the present article demonstrates the lack of background staining with a more specific anti-IgG AB in the modified cell-based assay.
Our affected group included dogs with epilepsy of unknown etiology or suspected dyskinesia, and we included dogs with unusual features, for example, new onset and rapid progression in adult dogs, behavioral and autoimmune comorbidities, reports on MRI signal changes in some of the dogs and inflammatory CSF changes in few dogs. Furthermore, nearly one-third of the dog owners reported notable behavioral changes. In three dogs the onset of behavioral changes which were described by the owners as restlessness, hyperactivity, lack of concentration, and inattention was concurrent with the onset or progression of the epilepsy suggesting a shared pathophysiology between seizures and behavioral changes. 43 In humans, sudden onset of alteration of behavior (delusions, psychosis, catatonia) and cognition combined with abnormal movements (eg, orofacial dyskinesia) are frequently seen in patients with AB against NMDAR1. 7,44,45 The AB prevalence in epilepsy score (APE score) includes aside from newonset epileptic seizures, neuropsychiatric changes, for example, agitation, aggressiveness, emotional lability, as common symptoms of patients with positive AB-status (up to 78.3%), and additionally rapidly progressive mental status changes within 1 to 6 weeks. 46 At this stage, it is premature to exclude autoimmune mechanisms in dogs with epilepsy and dyskinesia of unknown origin based on negative test results for neural antibodies with assays expressing murine and human neural antigens, considering that we included one dog with epilepsy and a history of a teratoma, which is frequently associated with NMDAR1 antibodies in humans. 47 We also included one dog with a histopathologic diagnosis of limbic encephalitis, which is frequently associated with LGI1 or CASPR2 antibodies in humans. 13,30 We were unable to identify neural AB, including LGI1 AB, in the 1 histopathologically confirmed case with LE. Low titer IgLON5 antibodies were present in 1 case and 1 control dog. Although we appreciated some parallels in clinical manifestations to IgLON5 disease in humans the finding of a low positive titer in a control dog argues against a significant association with neurologic disease in dogs. Lastly, AB status is not necessarily needed to assume immunemediated LE in humans and 7% of patients with clinical and MRI features of LE, CSF, and serum remain negative for neural antibodies. 48 Furthermore, autoimmune encephalitis with epileptic seizures can also present with normal brain MRI or CSF results in humans and cats. 7,16,49,50

| CONCLUSION
Our results highlight the need for the development of canine speciesspecific assays for neural antibodies and also the need for inclusion of control groups to assess whether there is a significant association with disease.