Primary leptomeningeal gliomatosis in a domestic shorthaired cat

Abstract

A 15-y-old neutered male domestic shorthaired cat was presented with a 16-d history of hindlimb paralysis in conjunction with 1-wk duration of inappetence and lethargy. Given intractable clinical signs, development of seizures, and poor prognosis, euthanasia was elected. Gross examination revealed mild, chronic, multifocal intervertebral disk disease; however, no gross abnormalities were noted in the spinal cord. Histologic examination of the cervical, thoracic, and lumbar spinal cord and the myelencephalon revealed diffuse and variable expansion of the meninges by sheets of neoplastic round-to-polygonal cells. The cells formed sheets and clusters, supported by a variably eosinophilic, fibrillar-to-basophilic, homogeneous matrix, and contained a small amount of eosinophilic cytoplasm. The nuclei were round with finely stippled to hyperchromatic chromatin and 1–2 small nucleoli. Mild white matter degeneration was present in the dorsal and ventral funiculi multifocally throughout the spinal cord, but was most severe in the ventral lumbar sections. Immunohistochemistry revealed strong intranuclear immunoreactivity for Olig2, and intracytoplasmic immunoreactivity for glial fibrillary acidic protein, MAP2, and vimentin in the neoplastic glial cells. To our knowledge, primary leptomeningeal gliomatosis has not been reported previously in a cat.

A 15-y-old neutered male domestic shorthaired cat was presented to the Emergency and Critical Care service at the University of Florida Veterinary Hospital (UFVH, Gainesville, FL) after a 16-d history of hindlimb paralysis in conjunction with 1-wk duration of inappetence and lethargy. Prior to admission to UFVH, the patient had been treated by a referring veterinarian with a steroid (identity and dose not provided) and diagnosed with intervertebral disk disease (IVDD) on radiographs (further characterization of the radiographic findings was not provided). Upon admittance to UFVH, the cat was obtunded, hypothermic, hypotensive, and mildly dehydrated. During the physical examination, no spinal trauma or pain was noted on palpation. Bloodwork (complete blood count and serum chemistry) was unremarkable, and overnight intravenous fluids were started. The following morning, the patient had multiple seizures, and euthanasia was elected. The patient was submitted for autopsy ~6 h after euthanasia. Autopsy confirmed IVDD at T1-T2, T9-T10, and L3-L4. At these locations, there was either mild-to-moderate extrusion of degenerate disk material or, in one instance (T1-T2), a firm, smooth nodule (~3 × 5 mm) that mildly protruded into the spinal canal. The main differentials for the clinical signs and spinal cord disease in our case included IVDD, spinal cord neoplasia (lymphoma), and infectious conditions, including feline infectious peritonitis and cryptococcosis.

An autopsy was performed with sections of all major organs, as well as the entire brain and spinal cord to the level of the caudal lumbar segment. Sections were fixed in 10% neutral-buffered formalin and routinely processed for histologic analysis. In representative sections of spinal cord (from the cervical to lumbar regions) and diffusely involving the myelencephalon, the leptomeninges were diffusely and variably expanded and infiltrated by a moderately cellular neoplasm composed of small, round-to-polygonal cells forming sheets and clusters (Fig. 1A, 1B). These cells were supported by a variably eosinophilic, fibrillar-to-basophilic, homogeneous matrix, had indistinct cell borders, and contained a scant amount of eosinophilic cytoplasm. The nuclei were round and contained finely stippled to hyperchromatic chromatin and 1–2 small nucleoli (Fig. 1B). Mitoses were <1 per 400× field with a total of 2 in 10 assessed fields; anisocytosis and anisokaryosis were moderate. Rarely, cells tracked along the pial blood vessels toward the superficial spinal cord parenchyma; however, direct parenchymal invasion was absent. Within the white matter of the spinal cord, there was segmental mild-to-moderate myelin vacuolation with a few scattered digestion chambers and spheroids. These findings were most numerous and severe in the dorsal funiculi (cranial to site of compression) and the ventral funiculi (caudal to the site of compression) and were correlated with the site of disk compression. Other comorbidities included mild pulmonary emphysema and renal retention cysts.

Figure 1.

Feline leptomeningeal gliomatosis. A. The leptomeninges are diffusely expanded by a neoplastic infiltrate. H&E. B. Neoplastic cells are round-to-polygonal forming sheets and clusters, have variably discrete cell borders, and are distributed within a basophilic homogeneous-to-fibrillar matrix. H&E. C. The neoplastic cells have strong intranuclear immunoreactivity for Olig2. Leica Bond polymer refine detection DAB. D. The supporting matrix and cytoplasm of the neoplastic cells have immunoreactivity for anti-glial fibrillary acidic protein (GFAP) antibody. Inset: higher magnification highlighting the cytoplasmic processes of the neoplastic cells. GFAP immunohistochemistry. Leica Bond polymer refine detection DAB. E. Neoplastic cells have strong intracytoplasmic immunoreactivity for MAP2. Leica Bond polymer refine detection DAB. F. Approximately 5% of the neoplastic cells have intranuclear immunoreactivity for Ki67. Leica Bond polymer refine red detection.

Immunohistochemistry for CD3 (T lymphocytes), CD45B220 (lymphocytes), glial fibrillary acidic protein (GFAP; intermediate filament of astrocytes), Iba1 (macrophages), Ki67 (proliferation marker), MAP2 (microtubule-associated protein 2), Olig2 (transcription factor for glial cells), PAX5 (B lymphocytes), vimentin (intermediate filament of mesenchymal cells; Table 1), and histochemical stains (periodic acid–Schiff [PAS] with and without diastase, reticulin, Masson trichrome, and Giemsa) were performed on paraffin-embedded tissues according to standard laboratory protocols (Supplementary Table 1). All staining procedures were performed on an automated stainer (Bond Max automated IHC/ISH stainer, Leica Biosystems, Cincinnati, OH). For all antibodies, 5-µm tissue sections were deparaffinized in xylene, hydrated in graded ethanol, and subsequently blocked with 3% hydrogen peroxide. Heat epitope retrieval was performed using either: Bond epitope retrieval solution 1 (Leica Biosystems) for 10 min (vimentin) or 30 min (CD45B220, Ki67, MAP2), or Bond epitope retrieval solution 2 (Leica Biosystems) for 10 min (Iba1), 20 min (CD3), 30 min (Olig2), or 40 min (PAX5). Heat epitope retrieval was not used for GFAP and was replaced by a 5-min 3% hydrogen peroxide block. Following application of the primary antibody, secondary antibodies were either anti-mouse IgG (Leica Biosystems; CD3, Ki67, MAP2, PAX5, vimentin; 30 min), Bond polymer (Leica Biosystems; GFAP, Iba1, Olig2; 30 min), or goat anti-rat (Jackson Laboratory, Bar Harbor, Maine; CD45B220; 8 min). Tertiary antibody was Bond polymer refine red detection (Leica Biosystems; CD3, Ki67, PAX5, vimentin; 10 min) or Bond polymer refine detection DAB (Leica Biosystems; CD45B220, GFAP, Iba1, MAP2, Olig2; 10 min). Negative controls consisted of adjacent sections of tissue with irrelevant species- and isotype-matched antibodies. Positive control tissue consisted of feline lymph node (CD3, CD45B220, Iba1, PAX5), feline brain (GFAP, MAP2, Olig2), and feline small intestine (Ki67, vimentin).

Table 1.

Summary of immunohistochemical results in spinal cord sections of a cat with primary diffuse leptomeningeal gliomatosis.

Antibody	Neoplastic cell reaction	Clone	Source	Dilution
Olig2	+, nuclear	EPR2673	Abcam	1:2,000
GFAP	+, intracytoplasmic	NA	Dako/Agilent	1:5,000
MAP2	+, intracytoplasmic	HM-2	Sigma-Aldrich	1:1,000
Vimentin	+, intracytoplasmic	VIM 3B4	Dako/Agilent	1:80
CD3	–	LN10	Leica	RTU
CD45B220	–	RA3-6B2	BD Biosciences	1:50
Iba1	–	NA	Wako	1:3,000
Ki67	5%, nuclear	MIB-1	Dako/Agilent	1:50
PAX5	–	1EW	Leica	RTU

⁺= positive for antibody; – = negative for antibody; GFAP = glial fibrillary acidic protein; NA = not available; RTU = ready to use.

PAS with and without diastase and Giemsa failed to reveal positive staining in the neoplastic population. Masson trichrome and reticulin stained the normal connective tissue meshwork of the meninges, but failed to stain the matrix associated with the neoplastic cells. The nuclei of the neoplastic population had strong nuclear immunoreactivity for Olig2 (Fig. 1C). The cytoplasmic processes of the neoplastic population had strong GFAP immunoreactivity (Fig. 1D). These processes often formed stromal material in which the neoplastic cells were embedded. All neoplastic cells had strong intracytoplasmic immunoreactivity for MAP2 (Fig. 1E) and vimentin. Approximately 5% of the neoplastic cells had strong intranuclear immunoreactivity for Ki67 (Fig. 1F). Neoplastic cells lacked immunoreactivity for CD3, CD45B220, Iba1, and PAX5. There were a few scattered immunopositive CD45B220 and Iba1 cells embedded within the neoplastic population. Our findings confirm leptomeningeal gliomatosis based on GFAP, Olig2, and vimentin immunoreactivity.^3,15 Olig2 is a transcription factor that is expressed in all glial tumors; GFAP is an intermediate filament that is expressed by astrocytes. The histologic pattern, coupled with the immunohistochemical features, all strongly support the diagnosis of leptomeningeal gliomatosis.

Although, to our knowledge, primary leptomeningeal gliomatosis has not been reported previously in a cat, it has been reported in 2 dogs.² Leptomeningeal gliomatosis is characterized by widespread dissemination of neoplastic glial cells throughout the leptomeninges.^2,3,12,13,16 In humans, gliomatosis is divided into primary and secondary forms; the primary form is based on 3 criteria: no attachment of the tumor to the parenchyma, no evidence of intra-axial lesions, and leptomeningeal encapsulation of the tumor.^3,13,16 In secondary gliomatosis, neoplastic cells invade the parenchyma.^3,13,16 The neoplastic cells in our case did not infiltrate the parenchyma (white and gray matter), but did occasionally extend around superficial white matter blood vessels that are ensheathed in pia matter, therefore not representing true parenchymal invasion. Rare cases of human leptomeningeal gliomatosis have been shown to have intraparenchymal lesions late in the course of the disease; however, this involvement is minimal, does not form a distinct mass, and is still classified as primary leptomeningeal gliomatosis.¹³ Immunoreactivity for Olig2 and MAP2 supports a glial origin in our case.^1,8 Positive cytoplasmic GFAP and vimentin staining, as seen in our case, also supports a glial origin and has been reported in cases of human primary leptomeningeal gliomatosis.^3,5,15 The Ki67 labeling index in our case (5%) was fairly similar to those reported in human cases, which have been reported in up to 20% of more malignant glial neoplasms.^3,8 In a study that evaluated Ki67 expression in feline glioma, the labeling index ranged from 0% to >50%; the Ki67 results were somewhat variable when grade of the neoplasm was considered.⁶

Glial tumors including astrocytomas, oligodendrogliomas, and gliomatosis cerebri (diffuse glioma) are uncommon to rare in cats compared to their canine counterparts.^4,6,7,10 The most common primary central nervous system (CNS) tumors of cats are lymphoma (either primary or part of systemic spread) and meningiomas.^4,6 Gliomatosis cerebri is characterized by diffuse infiltration of the neuroparenchyma with preservation of the CNS architecture by neoplastic glial cells that are predominantly astrocytic or oligodendrocytic in origin.^11,12,14

In our case, it is likely that the IVDD and extrusion of disk material in conjunction with the leptomeningeal neoplastic cell population both contributed to the neurologic signs and lethargy. The neoplastic population may have partially contributed to the changes described in the white matter and was most likely present before protrusion of the disk material. However, protrusion of the disk material likely acted as the primary cause of the acute onset hindlimb paralysis. No other histologic lesions were present in the remainder of the CNS to explain the clinically described seizures. IVDD is fairly uncommon in cats, but when it does occur, it is typically seen in the lumbar region.⁹ Primary leptomeningeal gliomatosis should be considered as a differential for neurologic disease in the cat.