Thoracolumbar meningeal fibrosis in pugs

Abstract Background Thoracolumbar myelopathies associated with spinal cord and vertebral column lesions, with a similar clinical phenotype, but different underlying etiologies, occur in pugs. Objectives To further characterize the clinical and neuropathological characteristics of pugs with longstanding thoracolumbar myelopathy. Animals Thirty client‐owned pure‐bred pugs with a history of more than a month of ataxia and paresis of the pelvic limbs, suggesting a myelopathy localized to the thoracolumbar spinal cord, were included in the study. Methods Prospective clinicopathological study. Included pugs underwent a complete neurological examination and gross and histopathologic postmortem studies with focus on the spinal cord. Computed tomography (n = 18), magnetic resonance imaging (n = 17), and cerebrospinal fluid analysis (n = 27) were performed before or immediately after death. Results Twenty male and 10 female pugs had a median age at clinical onset of 84 months (interquartile range, 66‐96). Affected pugs presented with a progressive clinical course and 80% were incontinent. There was circumferential meningeal fibrosis with concomitant focal, malacic, destruction of the neuroparenchyma in the thoracolumbar spinal cord in 24/30 pugs. Vertebral lesions accompanied the focal spinal cord lesion, and there was lympho‐histiocytic inflammation associated or not to the parenchymal lesion in 43% of the pugs. Conclusions and Clinical Importance Meningeal fibrosis with associated focal spinal cord destruction and neighboring vertebral column lesions were common findings in pugs with long‐standing thoracolumbar myelopathy.


| INTRODUCTION
Myelopathy involving the thoracolumbar spinal cord is well recognized in dogs and is clinically characterized by ataxia, paresis, or paralysis of the pelvic limbs, whereas the thoracic limbs are spared. 1 The clinical presentation in dogs with thoracolumbar myelopathy is highly variable regarding clinical onset, course, and presence of spinal pain, depending partly on the underlying etiology. 2 In pugs however, thoracolumbar myelopathy with uniform clinical phenotype and clinical course, but of different etiologies, including spinal cord as well as vertebral column lesions, is described.  Spinal cord lesions, including subarachnoid diverticula (SAD) and constrictive myelopathy, have been given attention in pugs as separate diagnoses underlying myelopathy. 11,12 The histopathological features of the affected spinal cord tissues in pugs with myelopathy have been sparsely studied, mostly limited to meningeal material removed at surgery. In that respect, both SAD and constrictive myelopathy are associated with proliferation and thickening of the meninges. 3,5,11,12,17 Vertebral column lesions, such as congenital vertebral malformations (CVMs), including hemivertebra and hypo-or aplasia of the caudal articulate processes (CAPs), 4,11,21,27 and acquired intervertebral disc herniation (IVDH), 8,13 are inconsistently described in pugs presenting as thoracolumbar myelopathy.
The high prevalence of neurological gait abnormalities in pugs 6,28 and the growing evidence of thoracolumbar myelopathy with a uni-

| Inclusion criteria
Pugs were included in the study provided that they fulfilled the following criteria: a history of more than 1 month of a bilateral pelvic limb gait abnormality described by the owner as incoordination and weakness (onset of clinical signs was determined from the owner's first recognition of incoordination and weakness in the pelvic limbs), and a neurological examination confirming pelvic limb ataxia and paraparesis suggestive of a myelopathy localized to the thoracolumbar spinal cord.
For the dog to be included, the owners also agreed to donate their dog for postmortem studies after naturally occurring death or euthanasia.

| Exclusion criteria
Dogs were excluded if histopathology of the thoracolumbar spinal cord was not performed at the time the pug died or was euthanized.

| Study population
Signalment and clinical variables were included in the data set. A thorough case history, including onset and course of clinical signs, presence of incontinence (fecal/urinary), signs indicating the dog was in pain, and result of any given treatment, was obtained at the time of inclusion. All pugs underwent a complete neurological examination performed the same day they were euthanized, by a boardcertified neurologist (C. Rohdin). After euthanasia, the included pugs underwent a full postmortem examination with focus on the spinal cord and vertebral column. If practically possible, the included pugs underwent advanced diagnostic imaging, including magnetic resonance imaging (MRI) and computed tomography (CT), as well as cerebrospinal fluid (CSF) sampling and analysis, and repeat neurological examinations (by C. Rohdin). Advanced imaging was carried out before or immediately after death in order to associate any potential spinal cord lesion with the presence of vertebral column lesions, as well as to guide the histopathologic investigation of the spinal cord.
Some of the recruited dogs were also included in another study investigating the presence of CVMs in pugs with and without neurological deficits. 18

| Imaging studies
Computed tomography was performed with 2 mm slice thickness, using a dual slice CT scanner (Siemens, Somatom Spirit, Siemens Munich, Germany), or a 64 slice CT scanner (Siemens, Somatom Definition, Siemens Munich, Germany). The studies were reconstructed with a soft tissue and a bone algorithm using medium and highfrequency reconstruction algorithms. For CT studies, the examination included the entire thoracic and lumbar vertebral column and the presence of vertebral malformations, such as hemivertebra, associated kyphosis, transitional vertebrae, and hypo-or aplasia of the CAPs was evaluated.
The MRI studies were acquired using either a magnet operating at 0.2 Tesla (Esaote Vet-MR Grande, Esaote, Genoa, Italy), or at 1.5 Tesla (Siemens, Magnetom Essenza, Siemens Munich, Germany). At a minimum, the MRI study included sagittal T1-and T2-weighted images of the entire thoracolumbar spine and transverse T1-and T2-weighted images of the area of interest. Intervertebral discs were described as compressing (mildly, moderately, severely), or not compressing the spinal cord.
All CT and MRI studies were submitted to and evaluated by a single board-certified imager at a diagnostic imaging company (VetImaging of New York, New York).

| Cerebrospinal fluid analysis
The CSF samples were taken at the University Animal Teaching Hospital, Uppsala, Sweden, or at Albano Animal Hospital, Danderyd, Sweden, and analyzed within 30 minutes of collection. The analyses included total leucocyte-and erythrocyte cell counts, and differential leucocyte counts, and measurement of protein concentration (turbidimetry). Protein concentration was considered to be increased if >0.25 g/L at the cerebellomedullary cistern, or >0.40 g/L at the lumbar cistern. 29 Cerebrospinal fluid was interpreted to have pleocytosis if there were >5 nucleated cells/μL, and pleocytosis was classified by the primary cell population if 1 cell type comprised >50%. 29

| Pathology
Gross pathology was performed at the Section of Pathology, Swedish University of Agricultural Sciences (site 1) or at the National Veterinary Institute, Uppsala, Sweden (site 2).
The entire central nervous system (CNS) were carefully sampled and fixed in 10% neutral buffered formalin for transport to a boardcertified pathologist. After the spinal cord had been evacuated, the ver- After fixation, specimens of various anatomical sites of the brain and spinal cord were processed for routine histopathology, embedded in paraffin wax and sectioned at 3-4 μm. Tissues were stained using routine hematoxylin and eosin (HE), and examined by light microscopy.
Histopathology of the vertebral column was performed at site 1. Decalcified tissue was embedded using paraffin wax and processed to 4-5 μm sections. Tissues were routinely stained with HE. The overall morphology and integrity of the intervertebral discs and adjacent vertebrae, in addition to other apparent features, were assessed and recorded.

| Statistics
Descriptive statistical analyses were performed using a commercially available statistical software program (JMP Pro v. 11.2.0, Cary, North Carolina). Continuous variables are presented as means and SDs and as medians and interquartile range (IQR). Associations between clinical variables and pathological findings were investigated using the Chi-square, Fishers exact test. Level of statistical significance was set at P < .05.

| Study population characteristics
Thirty pugs fulfilled the inclusion criteria with pathology being performed by the time the pug was euthanized. Information about dog signalment, clinical variables, CSF findings, and whether or not the dog was receiving medical treatment is presented in Table 1. No pug had undergone neurosurgical intervention. All 30 pug owners described an insidious, progressive course of their pug's pelvic limb gait abnormality. Mean duration of clinical signs before veterinary consultation was 7.1 months (SD, 6.9). None of the owners perceived their dogs to be in pain, but a reluctance to go for walks was reported by the owners of 13 pugs.
At the physical examination, all 30 pugs presented with moderate to severe pelvic limb muscle atrophy. Spinal pain was not recognized in any of the pugs upon examination. In 13 pugs, a repeated neurological examination had been performed. In 3 of them, the pelvic limb flexor reflexes gradually became weaker, with loss of muscle tone in the pelvic limbs, at reexaminations performed within 9.5, 11, and 12 months after the initial examination. These 3 pugs were examined neurologically on 7, 4, and 3 occasions respectively. In 2 out of these 3 pugs, with a normal perineal reflex upon initial examination, loss of the perineal reflex and anal tone developed over time.
All 30 pugs were still ambulatory at the day of euthanasia. Twentyeight pugs were euthanized because of their gait abnormality or because of their associated incontinence. Other causes for euthanasia (n = 2) were pyometra (n = 1) and osteosarcoma of the mandibula (n = 1). The pug with pyometra had a neurological examination performed, confirming pelvic limb ataxia and paraparesis, 1 month before acute onset of clinical signs of the pyometra. The pug with osteosarcoma had a neurological examination, confirming pelvic limb ataxia and paraparesis, and a CT of the entire thoracolumbar vertebral column, with no signs of vertebral involvement, performed 2 months before euthanasia. was diagnosed by MRI in 9/16 (56%) pugs ( Figure 2). An IVDH, compressing the spinal cord at the intramedullary T2W hyperintensity, was diagnosed by MRI in 5 pugs. The disc compression was considered mild in 2 and moderate in 3 pugs.

| Imaging studies
T A B L E 1 Distribution of signalment, clinical variables, cerebrospinal fluid, and histopathological findings of lympho-histiocytic inflammation in the central nervous system of the 30 pugs of the present study. Data on signalment and incontinence is compared to pugs in the general Swedish pug population a   Table 1). The CSF of the pug with pyometra was unremarkable. The CSF of the pug with osteosarcoma showed an elevated total protein concentration with normal cell count.

| Macroscopic findings
Vertebral column pathology included hemivertebra (n = 6), with (n = 4) or without (n = 2) kyphosis, and herniating intervertebral discs (n = 15). The morphology of the CAPs could not be assessed, as they were destroyed by the processing procedure. There was occasional focal spinal cord atrophy of the thoracolumbar spinal cord.  Table 2.

| Histopathology of the CNS
Wallerian-like degeneration could be traced further cranial from the malacic spinal cord lesions, mainly through the fasciculus gracilis in the cervical spinal cord ( Figure 3E). Compared to the, in general, severe focal thoracolumbar spinal cord lesions, these neurodegenerative lesions were mild.
Nerve root entrapment and vascular encasement was found with meningeal vessels and nerve roots embedded in the circumferential meningeal fibrosis. Loss of nerve fiber density, seen involving both dorsal and ventral nerve roots ( Figure 3F), hypertrophy of the dentate ligaments, and intradural tethering was found including the thoracolumbar but also the cervical and lumbosacral spine. In addition, perivascular fibrosis with collagen deposits were found throughout the CNS ( Figure 3G), involving areas with and without meningeal proliferation.
Mild to moderate lympho-histiocytic leptomeningitis was found adjacent to the focal spinal cord destruction in 2 pugs ( Figure 3J). The pug euthanized because of pyometra had perivascular cuffs of lympho-histiocytic inflammatory cells in the brain. Evidence of lympho-histiocytic infiltration of the CNS was not confirmed in the pug with osteosarcoma.
Severe gray matter destruction was recognized in the cranial cervical spine in 3 pugs. One of these pugs had had surgery to remove a SAD in the dorsal C2-C3 spinal cord area 4 years before presenting with signs of a T3-L3 myelopathy. These lesions displayed the same features as the lesions in the thoracolumbar spinal cord, with (n = 2) or without (n = 1) ( Figure 3K)    Spinal pain was not described by any of the owners nor recognized upon examination in any of the pugs. However, reluctance to go for walks was described by more than a third of the owners. Although vascular compromise to the spinal cord in dogs, except initially, is described as non-painful, 65,66 people with ischemic spinal cord lesions describe chronic, intense pain that increase with exertion. 67 It needs to be questioned whether the decreased activity in some pugs with thoracolumbar myelopathy was because of the inability to walk normally, because of any concomitant breed related dyspnea, or, in fact, a reflection of pain. 28,68 We acknowledged that this descriptive study, having more than 1 pathologist examining the pugs, was limited by the lack of possibility for objective comparison and statistical analysis. However, the pathologists independently confirmed the major pathological findings of this study, that is, meningeal fibrosis, and concomitant spinal cord lesions, and also the presence of lympho-histiocytic CNS inflammation. Furthermore, each of them provided valuable input related to additional pathology and potential pathogenesis of pugs with longstanding thoracolumbar menin-

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
Animal Ethics Committee of Sweden (Uppsala djurförsöksetiska nämnd) C202/2014.

HUMAN ETHICS APPROVAL DECLARATION
Authors declare human ethics approval was not needed for this study.