Focal thoracolumbar spinal cord lymphosarcoma in a ferret (Mustela putorius furo)

Abstract

A 6-year-old, castrated male domestic ferret (Mustela putorius furo) was euthanized following progressive hind limb paresis and atonia of the bladder of 1-year duration. Neurological evaluation localized the lesion to the thoracolumbar spinal region, and magnetic resonance imaging showed a focal intramedullary spinal cord lesion. Histopathology revealed an extensive, unencapsulated, poorly demarcated mass within the thoracolumbar spinal cord, diagnosed as lymphosarcoma.

Primary neurological disorders are rarely reported in pet ferrets (1–3). Paresis of the pelvic limbs is a non-specific clinical sign, and is seen in a variety of conditions (2–4). Differential diagnoses for hind-end weakness include insulinomas and other metabolic disorders, distemper, cardiac disease, hypoxia, anemia, hepatic disease, caudal abdominal masses, cystic calculi, and peritonitis (2,3). Complete posterior paralysis with urinary incontinence can be caused by vertebral fractures, intervertebral disc disease (IVDD), hemivertebrae, hypoglycemia, hematomyelia associated with prolonged estrus, Aleutian disease, myelitis caused by fungal infections, Mycobacterium species, rabies virus, and neoplasia such as plasma cell myeloma, and spinal lymphosarcoma (1–6). This case report describes the clinical and pathological findings in a ferret with spinal lymphosarcoma.

Case description

A 6-year-old, castrated male domestic ferret (Mustela putorius furo) was presented to the Ontario Veterinary College Health Sciences Centre for progressive hind limb paresis of 6-month’s duration. The ferret had an initial shifting hind limb lameness which progressed to hind limb paresis and fecal incontinence. The patient had a recent 3-week history of urinary incontinence and oliguria, and the owner had been manually expressing his atonic bladder. The referring veterinarian had issued periodic prednisone [1.5 mg/kg body weight (BW), PO, q12h] and meloxicam (unknown dosage), but no clinical response was observed. Whether prednisone and meloxicam were given concurrently, and the total number of doses received, are unknown. The ferret was also previously diagnosed with adrenocortical disease and underwent a left adrenalectomy 5 y earlier. However, within a year following the surgery, related clinical signs reoccured. The ferret was medicated with leuprolide acetate (Lupron; Tap Pharmaceuticals, Abbott Park, Illinois, USA) 200 μg subcutaneously every 30 d for management of the adrenocortical disease. The ferret previously tested negative for Aleutian disease [Ferret ADV saliva antibody enzyme-linked immunosorbent assay (ELISA) test].

On physical examination, the ferret was bright, alert, responsive and very active but in poor body condition with muscle mass and fat reserves below normal limits. The base of the tail was alopecic. Hind leg paraparesis was noted and abrasions and fur loss were present on the dorsal aspect of both hind paws caused by dragging of the limbs. Abdominal palpation revealed a distended urinary bladder, and the ferret dripped urine throughout the physical examination. A complete blood (cell) count (CBC), biochemistry profile, and urinalysis were performed and the results were unremarkable.

A full neurological evaluation was performed and showed the ferret to be in normal mentation. Paraparesis and ataxia were noted when the ferret tried to ambulate. Proprioception was asymmetrically decreased in the pelvic limbs, worse on the right, and spinal reflexes and nociception were intact. Atonia of the bladder and mild caudal lumbar hyperesthesia were noted. Anal tone was within normal limits. Based on the neurologic examination, the lesion was localized to the T3 to L3 spinal cord segments, worse on the right side, suggesting an upper motor neuron lesion.

The ferret was sedated with butorphanol (Torbugesic; Fort Dodge Animal Health, Fort Dodge, Iowa, USA), 0.25 mg/kg BW, IM, and midazolam IM (Midazolam HCl Injectable; Hospira, Illinois, USA), 0.25 mg/kg BW, IM. Whole body survey radiographs (dorsoventral and lateral views) and an abdominal ultrasonography showed no abnormalities other than a distended bladder.

With a suspicion of a T3-L3 myelopathy, the ferret was presented 1 wk later for further diagnostic testing. The ferret was premedicated using the previously described sedation protocol, was masked down with 5.0% sevoflurane in oxygen, and intubated using a 2.0 mm inner diameter un-cuffed endotracheal tube. General anesthesia was maintained with 2.0% to 3.0% sevoflurane. A 24-gauge intravenous catheter was placed in the left cephalic vein and lactated Ringer’s solution with 2.5% dextrose was given at a rate of 5 mL/kg BW per hour. Thermal support was provided via a forced warm-water blanket. Anesthetic monitoring included heart rate and rhythm, capnography, and esophageal temperature using a multi-parameter patient monitor (Cardiocap/5 Anesthesia Monitor; Datex-Ohmeda, Madison, Wisconsin, USA). Magnetic resonance imaging (MRI, GE Signa 1.5 Tesla magnet) of the T3-L3 spinal region was performed and revealed a focal intramedullary spinal cord lesion from T12 to L3. Cranially, the lesion was left-sided, but caudally the lesion was dorsal and right-sided. Relative to the normal spinal cord, the lesion was hyperintense on the T2 weighted images (Figure 1), isointense on T1 weighted images, and moderately contrast enhanced. While still under anesthesia, a cervical cerebrospinal fluid (CSF) tap was performed and yielded approximately 50 μL of clear fluid, which was submitted for cytological evaluation. Total white blood cell count, differential white blood cell counts, and protein concentration were compared to published data on the CSF composition of healthy adult ferrets (7), and all values were within normal limits, except for the presence of mild hemorrhage. A polymerase chain reaction (PCR) for Aleutian disease virus on CSF was negative. The top differential diagnoses included neoplasia, inflammatory disease, and myelomalacia (secondary to trauma, ischemia, vasculitis, or other cause). The owners declined the option of fluoroscopy-guided fine-needle aspiration of the lesion. Based on the clinical, imaging, and laboratory findings and the lack of response to long-term treatment with anti-inflammatory drugs (prednisone and meloxicam), neoplasia was considered the most likely diagnosis. The owners declined radiation or chemotherapy, and the patient received no further medical therapy. Six months later, the ferret was presented for a deteriorating clinical condition. The paralysis was ascending and the ferret was now dragging his torso in addition to his hind limbs. No deep pain could be elicited in the hind limbs, although anal tone was still present. The patient was euthanized under general anesthesia with intra-cardiac pentobarbital and then submitted for complete necropsy.

Figure 1.

Sagittal MRI image revealing a hyperintense lesion at T12 to L3 (T2 weighted images).

No gross lesions were noted in the spinal cord or vertebral column, both of which were closely examined. The right adrenal gland measured 11 mm × 3 mm × 4 mm, and the pancreas grossly appeared normal. Tissues were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 4 μm, and stained with hematoxylin and eosin. Within the grey and white matter of the entire thoracolumbar spinal cord was a densely cellular, poorly demarcated, unencapsulated mass comprised of round cells arranged in sheets on a loose pre-existing stroma, interpreted as malacic neural parenchyma (Figure 2a). The constituent cells had sparse quantities of eosinophilic cytoplasm, large, centrally located, round to oval nuclei, demonstrated two-fold anisokaryosis, contained dense granular chromatin with occasional prominent individual or multiple nucleoli, and 17 mitotic figures were observed in 10 high power (40×) fields (Figure 2b). As per the World Health Organization (WHO) classification system (8), the tumor was classified as a medium-cell (1.5 to 2.0 red blood cell diameter) low grade (mean of ≤ 5 mitotic figures/40× field) lymphosarcoma of the spinal cord.

Figure 2.

Thoracolumbar spinal cord of the ferret. The spinal cord is effaced by sheets of round cells. Hematoxylin and eosin stain. (A) Bar = 500 μm. (B) Bar = 20 μm.

Within the thoracic spinal cord, 30% of myelin sheaths within the white matter of the dorsal funiculi were dilated up to 6× normal size and contained myelin-laden macrophages with large quantities of foamy cytoplasm and occasionally condensed, pyknotic nuclei (digestion chambers). A few random digestion chambers were also present within the ventral funiculi. No significant histologic lesions were observed in the cervical spinal cord or brain.

Immunohistochemistry (IHC) was pursued to further characterize the neoplasm. Sections were evaluated using IHC for T-cell (polyclonal rabbit anti-human CD3; Dako, Mississauga, Ontario) and B-cell markers (monoclonal mouse anti-human CD79a, Dako; polyclonal rabbit anti-human CD20, LabVision, Fisher Scientific, Nepean, Ontario). Non-immune rabbit serum and antibody diluent were substituted for primary antibody for negative controls for polyclonal and monoclonal antibodies, respectively. Rare small lymphocytes scattered throughout the mass were immunoreactive with anti-CD3 antibody. No neoplastic cells were immunoreactive with anti-CD79a antibody or anti-CD20 antibody. Internal control tissues (lymphoid tissue within the gastrointestinal tract) stained appropriately with all antibodies. Based on the pattern of immunoreactivity of the neoplastic round cells, the diagnosis was further refined as a null-cell or natural killer cell lymphosarcoma.

Discussion

The clinical signs and diagnostic findings in this case are consistent with the presence of a slow-growing primary spinal lymphosarcoma. Antemortem diagnosis of the specific cause of the neurologic signs and lesion noted on MRI would have required cytologic identification of neoplastic cells from either a fine-needle aspirate, which was declined by the owners, or from a CSF tap. A cervical CSF sample did not contain abnormal cells. A lumbar tap might have provided different results, but given the intraparenchymal location of the tumor, it is possible that there was simply no exfoliation into the CSF. This has been shown to occur with lymphosarcoma affecting the central nervous system (CNS) in other species (9).

Spinal lymphosarcoma in the ferret is infrequently reported. A review of the literature revealed only 1 other case of primary spinal lymphosarcoma (3,10). That ferret had acute onset of pelvic limb paresis at 22 months of age. Radiography and computed tomography (CT) revealed a soft tissue mass at the location of the 5th lumbar vertebra (L5), and associated bone lysis of L5. Lymphosarcoma was diagnosed by cytology and postmortem histopathology confirmed a malignant, invasive lymphosarcoma, which destroyed the adjacent vertebral bone. The young age of onset, acute presentation, and vertebral bone destruction are very different than what was seen in the case presented here. In the present case, the lack of vertebral damage perhaps contributed to the slow progression of clinical signs.

By comparison, extraspinal lymphosarcoma is very common in the domestic ferret. In a study evaluating neoplastic disease in 574 ferrets (11), lymphosarcoma was the third most common neoplasm diagnosed (11.9%); with the digestive (58.4%), endocrine (39.7%), and hemolymphatic (15.2%) systems most commonly affected. Another study (12) described gastrointestinal (45%), multicentric (40%), mediastinal (10%) and cutaneous lymphosarcomas (5%) in 20 ferrets. Ammersbach et al (13) evaluated the clinical and pathological changes in 29 ferrets with histologically diagnosed lymphosarcoma, classified based on the WHO (14) classification system with consideration of immunophenotype. The viscera of the abdominal cavity (38%), thoracic cavity (3.5%), both abdominal and thoracic cavities (24%), both body cavities plus other sites (27.5%), and skin (7%) were most commonly affected. Erdman et al (14) also evaluated the clinical and pathological changes in 60 ferrets with histologically diagnosed lymphosarcoma, and determined that acute onset, the presence of a mediastinal mass, lymphocytosis, and multicentric distribution were often associated with younger ferrets, whereas lymphopenia was often associated with older ferrets. The ferret in the present case showed no changes in circulating lymphocyte numbers. Erdman et al (14) also stated that survival time after diagnosis is longer in adult ferrets than in younger ferrets.

In veterinary medicine, primary lymphosarcoma affecting the CNS is rare in all species. It is most frequently reported in dogs (15,16) and cats (17), and there are few reports in other species (18). Primary CNS lymphosarcoma must be differentiated from systemic lymphosarcoma with CNS involvement. Primary CNS lymphosarcoma is most often intraparenchymal and has a peri-vascular pattern. In contrast, lymphosarcomas that involve the CNS are often extraparenchymal and affect the leptomeninges, choroid plexus, epidural space, or peripheral nerves. Diagnosis is based upon exclusion of a primary extraneural neoplasm (19,20). In the ferret described here, the spinal neoplasm was parenchymal and no other sites of neoplasia were identified.

In companion animals, there are reports of primary lymphosarcomas of both B- and T-cell origin affecting the CNS (21–24). Immunostaining was performed in the single previously reported case of a ferret spinal lymphosarcoma (10). The neoplasm stained positive for CD3 and negative for CD79, indicating that the neoplasm was of T-cell origin. Immunostaining results have been reported for lymphosarcomas of other body systems in the ferret. Of the 20 cases reported by Onuma et al (12), 88.9% were of T-cell origin and 11.1% were of B-cell origin. Ammersbach et al (13) used histomorphology and immunophenotyping to classify tumors as peripheral T-cell lymphosarcoma (58.5%), anaplastic large T-cell lymphosarcoma (17%), anaplastic large B-cell lymphosarcoma (14%), Hodgkin-like lymphosarcoma (7%), and diffuse large B-cell lymphosarcoma (3.5%).

In the current case, immunohistopathological evaluation of the neoplasm resulted in negative immunostaining for CD3, CD79a, and CD20, prohibiting further classification. Three hypotheses are proposed for the lack of immunostaining. The first is that the neoplasm is of the “null cell” (25) phenotype. Lymphosarcomas expressing this phenotype are CD3-/CD79a-, and are rare in both dogs and humans (25–27). The second is that the lymphosarcoma is of natural killer cell origin, which can also result in a CD3-/CD79a-phenotype (28). The third is that the lack of staining was due to overfixation. However, this was considered less likely given the presence of positive internal controls, which had undergone the same fixation regimen. All but 2 ferret lymphosarcomas described by Ammersbach et al (13) stained for either CD3 or CD79a. These 2 cases did not express CD3 or CD79a molecules, were classified as Hodgkin-like lymphosarcoma due to their phenotypic similarity to Hodgkin lymphosarcoma in humans, and demonstrated marked multi-organ infiltration and metastasis. Hammer et al (29) attributed the lack of IHC staining in 2 cases to overfixation in formalin.

In this case, lack of immunostaining of the neoplastic cells with antibodies recognizing T- and B-lymphocytes in formalin-fixed tissue poses not only a diagnostic challenge, but has clinical repercussions. In the current case, the predominant cell type possessed lymphocytic phenotypic characteristics, thus making the diagnosis of lymphosarcoma fairly straightforward. However, this is not always the case. In many circumstances, the diagnosis of lymphosarcoma becomes challenging, and the use of lymphocyte surface markers is extremely beneficial in the diagnostic process (30–32). For biopsy material, the inability to phenotype a lymphosarcoma may have a significant influence on clinical treatment decisions and on providing an accurate prognosis. In the case series described by Ammersbach et al (13), the mean survival times of ferrets not treated with chemotherapy following diagnosis were 5.0 and 8.4 months for T-cell and B-cell lymphosarcoma, respectively. Mean survival times of ferrets following diagnosis and treatment with chemotherapy were 4.3 and 8.8 months for T-cell and B-cell lymphosarcoma, respectively.

In summary, we describe here the prolonged and progressive clinical course, resulting in euthanasia on humane grounds, of an untreated spinal lymphosarcoma in a mature ferret. The presumptive clinical diagnosis of neoplasia was confirmed on necropsy by histopathology. Negative immunoreactivity for standard T- and B-lymphocyte markers, and the clinical importance of typing these neoplasms, highlight the importance of obtaining appropriate internal control tissue and minimizing fixation time in order to differentiate between true and artifactual lack of IHC staining. The description of additional cases of this unusual presentation of a common neoplastic condition of the ferret is necessary in order to assess the effectiveness of chemotherapeutic modalities, and to provide accurate information on prognosis.