Immunocytochemical demonstration of feline infectious peritonitis virus within cerebrospinal fluid macrophages

Abstract

A 4-month-old female entire domestic shorthair cat presented with an acute onset of blindness, tetraparesis and subsequent generalised seizure activity. Haematology and serum biochemistry demonstrated a moderate, poorly regenerative anaemia, hypoalbuminaemia and hyperglobulinaemia with a low albumin:globulin ratio. Serology for feline coronavirus antibody was positive with an elevated alpha-1 acid glycoprotein. Analysis of cisternal cerebrospinal fluid (CSF) demonstrated markedly elevated protein and a mixed, predominately neutrophilic pleocytosis. Immunocytochemistry for feline coronavirus was performed on the CSF, with positive staining observed inside macrophages. The cat was subsequently euthanased, and both histopathology and immunohistochemistry were consistent with a diagnosis of feline infectious peritonitis. This is the first reported use of immunocytochemistry for detection of feline coronavirus within CSF macrophages. If this test proves highly specific, as for identification of feline coronavirus within tissue or effusion macrophages, it would be strongly supportive of an ante-mortem diagnosis of feline infectious peritonitis in cats with central nervous system involvement without the need for biopsy.

Case Report

A 4-month-old, female entire domestic shorthair cat presented with a history of acute onset ataxia, lethargy and apparent blindness. The cat had been rehomed from a rescue centre 3 weeks previously and had been an indoor cat in a single-cat household since that time. There was no reported history of toxin exposure, trauma or previous medical problems.

General clinical examination was unremarkable. Abnormalities on neurological examination included a depressed mentation, tetraparesis with ataxia of all limbs, delayed postural reactions for all limbs, blindness with bilaterally dilated pupils, absent menace responses and incomplete pupillary light reflexes. Ophthalmological examination was otherwise unremarkable. Soon after presentation the cat became restless and a self-limiting, generalised seizure event followed. These findings were consistent with a neurolocalisation of multifocal central nervous system (CNS) disease.

A venous blood sample was taken and complete blood cell count showed a moderate, poorly regenerative anaemia [haematocrit 13.7%, reference interval (RI) 26–45%, reticulocytes 7.1 ⋅ 10⁹/l] Serum biochemistry revealed an elevated serum total protein (96 g/l, RI 56–78 g/l), hypoalbuminaemia (23 g/l, RI 25–43 g/l) and hyperglobulinaemia (73 g/l, RI 24–47 g/l) with a low albumin:globulin ratio (A:G) (0.3, RI 0.8–2.0). Serological assays for feline leukaemia virus antigen and feline immunodeficiency virus antibody were negative (Snap Combo Plus; Idexx). Feline coronavirus (FCoV) antibody titre was high at >1280, with an elevated alpha-1 acid glycoprotein (AGP) level of 1.84 mg/ml (RI <0.5 mg/ml).

A sample of cerebrospinal fluid (CSF) was acquired under general anaesthesia from the cerebellomedullary cistern. Analysis revealed a clear, colourless appearance, with markedly elevated protein (1.83 g/l, RI <0.25 g/l) and an elevated nucleated cell count (251/μl, RI <5/μl). A mixed pleocytosis, with a predominance of non-degenerate neutrophils (60%) was observed on cytospin cytology (Figure 1). No microorganisms were observed on cytology. Cytospin preparations were also prepared on poly-L-lysine-coated slides and immunocytochemistry was performed with a monoclonal antibody against the FCoV nucleocapsid protein (mouse monoclonal, clone FIPV3-70; Novus Biologicals) using an avidin–biotin–peroxidase complex technique. ¹ Negative controls were run concurrently. Thirty percent of CSF macrophages showed positive staining consistent with the presence of intracytoplasmic FCoV antigen within these cells (Figure 2).

Figure 1

Cytocentrifuged preparation of cerebrospinal fluid demonstrating a moderate, mixed pleocytosis with a predominance of neutrophils and a smaller proportion of macrophages. Wright Giemsa, ×50 objective

Figure 2

Immunocytochemical staining of a cytocentrifuged preparation of cerebrospinal fluid. Positive staining for feline coronavirus antigen can be seen within the macrophages as intracytoplasmic brown material. ×100 objective

Pending the results of investigations and following the first seizure event, phenobarbitone treatment was started by slow intravenous injection (5 mg/kg q12h). However, shortly after recovery from general anaesthesia the cat experienced another seizure event and did not respond to two sequential doses of midazolam (0.3 mg/kg IV) and a further 5 mg/kg bolus of phenobarbitone. At this point, propofol was given to effect and a propofol constant rate infusion was started (0.1 mg/kg/min). External manifestations of the seizure activity were controlled at this rate.

In light of the clinical presentation, clinicopathological findings and results of the CSF immunocytochemistry a presumptive diagnosis of feline infectious peritonitis (FIP) was made. After discussion with the owners, it was decided to euthanase the cat given the recurrent seizure activity and grave prognosis for FIP. Consent was given for post-mortem examination. Macroscopic abnormal findings were limited to multiple, raised, white foci (2–4 mm in diameter) over the cortical surface of both kidneys. All other systems, including the CNS, were grossly normal on examination. Histopathological findings were consistent with the non-effusive form of FIP, with diffuse expansion of the meninges by moderate-to-large numbers of lymphocytes and plasma cells, and perivascular infiltration of mononuclear cells throughout the cortical grey matter (Figure 3). Perivascular infiltration of lymphocytes, macrophages and plasma cells was also found in the renal cortex, lungs, small intestine, mesenteric lymph nodes and omentum. Immunohistochemistry was performed on histological sections from formalin-fixed, paraffin-embedded tissue samples using the same monoclonal antibody against the FCoV nucleocapsid protein as for immunocytochemistry. ² The analysis was performed using automated staining equipment (DAKO Autostainer Plus) and the EnVision FLEX visualisation system (DAKO). Positive staining for FCoV antigen was observed within tissue macrophages in the meninges, kidneys and mesenteric lymph nodes, confirming a diagnosis of FIP (Figure 4).

Figure 3

Histological section of cerebral cortex showing marked, diffuse infiltration and expansion of the overlying meninges by lymphocytes and plasma cells. Haematoxylin and eosin

Figure 4

Immunohistochemical staining of the meninges for feline coronavirus antigen, demonstrating positive staining (dark brown material) within tissue macrophages, consistent with a diagnosis of feline infectious peritonitis

This is the first reported use of immunocytochemistry for the identification of FCoV antigen within CSF macrophages. Immunocytochemical analysis of the CSF has been used in the human medical field to allow differentiation of cell types,^3,4 to assist in the diagnosis of leptomeningeal metastasis ⁵ and to demonstrate the presence of various infectious agents within CSF mononuclear cells, such as human immunodeficiency virus, cytomegalovirus, Japanese encephalitis and Mycobacterium tuberculosis.^6–9 Immunocytochemistry has been reported in the detection of canine distemper virus antigen in CSF, ¹⁰ but, to our knowledge, analysis of feline CSF for infectious agents using immunocytochemistry has not been reported.

FIP is an invariably fatal disease seen in a small subset of cats infected by the ubiquitous virus FCoV. FCoV can be separated into two biotypes: a non-virulent form [feline enteric coronavirus (FECV)] and a virulent form causing FIP [feline infectious peritonitis virus (FIPV)]. ¹¹ FIPV may show increased virulence owing to an increased ability to infect mononuclear phagocytes that are normally intrinsically resistant to viral infection. ¹² FIP is a common cause of neurological disease in cats, especially in those less than 4 years old, and, in one study, was diagnosed in 48% of cats with CNS inflammatory disease.^13,14

It is widely accepted that the ante-mortem diagnosis of FIP is difficult owing to the lack of specific clinical signs, lack of pathognomonic biochemical abnormalities, and the variable sensitivity and specificity of tests routinely used in practice.^15,16 Cats with clinical signs referable only to the CNS pose a particular challenge owing to the lack of an effusion for sampling and the inherent difficulty in biopsy of the CNS ante-mortem. Findings suggestive of FIP involving the CNS include a predominately neutrophilic CSF pleocytosis and an elevated CSF total protein. In a previous study only cats diagnosed with FIP had a CSF total protein greater than 1 g/l, as seen for the cat in this report. ¹³ Owing to the presence of non-degenerate neutrophils and the absence of visible microorganisms on cytology, CSF culture was not performed in this case; however, a negative culture would add further support to a suspicion of non-septic meningeal inflammation. Anti-FCoV antibody levels in the CSF have been shown to have equivocal clinical use as they correlate with serum antibody levels and are only elevated in strongly seropositive cats. ¹⁷ Brain magnetic resonance imaging (MRI) findings reported to be consistent with FIP include ventricular dilation and ependymal or periventricular contrast enhancement.^18,19 However, these are neither pathognomonic nor observed in every case of FIP, with invasive meningeal or parenchymal biopsy still required for a definitive diagnosis. MRI was not performed in this case owing to financial constraints.

The finding of hyperglobulinaemia, as for this case, in a young cat with consistent clinical signs raises the index of suspicion for FIP. However, hyperglobulinaemia remains a non-specific finding and was reported in only 17% and 39% of cats with FIP in two studies.^13,15 While serum protein electrophoresis was not performed in this case, the presence of an electrophoretogram demonstrating increased alpha-2 and gamma globulin fractions has been reported to be suggestive of FIP. ¹⁵ The low A:G seen in this case was also supportive of a diagnosis of FIP. One study found the positive predictive value (PPV) of an A:G <0.8 to be 92%; however, a recent study found the PPV to be only 12.5% using the same cut off value.^20,21 The negative predictive value of an A:G <0.8 was found to be 100% in the latter study, making A:G better for ruling out a diagnosis of FIP for that particular population. While serum antibodies against FCoV were high in this case, antibody testing alone is of limited value as their presence neither confirms, nor their absence excludes, a diagnosis of FIP. ¹⁶ An AGP above 1.5 mg/ml, as seen in this case, was shown to be highly correlated with FIP in one study; ²² however, other reports have found elevated AGP levels to be a non-specific finding. ¹⁶ Reverse transcription-PCR for FCoV RNA was not performed in this case, but its use has been reported on a variety of different samples, including faeces, effusion, blood, CSF, tissue and saliva. It cannot currently be used to differentiate between FIPV and FECV, with both false-positive and false-negative results possible.^23,24 Therefore, while the clinical presentation and clinicopathological findings for the cat in this report were suggestive of FIP, a definitive ante-mortem diagnosis could not be made with this information alone.

Histopathological examination for distinctive perivascular and pyogranulomatous lesions is often used for confirmation/exclusion of FIP. ²³ The gold standard for a definitive diagnosis is currently considered to be the detection of intracellular FCoV antigen either by immunohistochemistry of tissue macrophages, as used in this case, or immunofluorescence antibody staining of effusion macrophages.^2,20,25 Both of these techniques have been reported to have 100% specificity for the diagnosis of FIP, as only FIPV is able to replicate within the macrophages in the sufficiently large amounts required to generate positive staining.^2,20 Positive immunocytochemical staining for FCoV antigen within CSF macrophages would be expected to have a similarly high specificity and was judged strongly supportive of FIP in this case. This increased index of suspicion ante-mortem allowed an informed decision to be made by the owner for euthanasia in a cat with an invariably fatal disease. As is the case for both immunohistochemistry and immunofluorescence antibody testing of tissue and effusion macrophages, respectively, negative staining would not rule out FIP owing to the possibility of an insufficient number of macrophages being present in the CSF sample or host anti-FCoV antibodies competing for the assay. ²³ While clinical signs referable to involvement of the CNS and/or eyes predominate in cases of non-effusive FIP, abdominal lesions are frequently found in the kidneys and mesenteric lymph nodes, as observed post mortem in this case. ²⁴ Ultrasound-guided fine needle aspiration and/or Tru-cut biopsy of affected organs could also be used to support a diagnosis of FIP following histology with or without immunochemical analysis. ²⁶

Anti-Leishmania infantum immunostaining has already been reported as useful in the diagnosis of canine leishmaniasis, and the results of this case report are encouraging for the potential future use of CSF immunocytochemistry in supporting a diagnosis of other feline infectious diseases such as toxoplasmosis, viral encephalitis or FIV.^27,28

Conclusions

This case represents the first reported use of immunocytochemistry for the identification of FCoV antigen within CSF macrophages in a cat with FIP. Further study is underway to establish a sensitivity and specificity for this technique; however, we feel that it could offer a minimally invasive, definitively diagnostic tool for the ante-mortem diagnosis of FIP in cats with CNS involvement, without the need for biopsy.