Materials and Methods

Dogs—The study protocol was approved by the Oklahoma State University Institutional Animal Care and Use Committee. Owners of all dogs with SARDS were provided with, and signed, an informed client consent form. An ophthalmic examination including biomicroscopy, indirect ophthalmoscopy, and applanation tonometry was performed in 17 dogs with SARDS. A definitive diagnosis of SARDS was made by electroretinography. In all dogs, electroretinography was performed during scotopic conditions. Eight of 17 dogs had white light testing only, 3 of 17 had white and blue light testing, and 6 of 17 had white, blue, and red light testing. There was no distinguishable waveform in all 17 dogs with SARDS. Duration from onset of blindness to examination and serum collection was 2 weeks to 6 months (1 dog) with a mean of 2 months.

Dogs with nonocular neoplasia were examined at the teaching hospital or private oncology referral practice. Neoplasia was confirmed by fine-needle aspiration and cytologic examination of the neoplasm or by histologic examination of a biopsy specimen of the neoplasm. Examinations of the anterior segment and fundus were performed by a diplomate of the American College of Veterinary Internal Medicine.

Clinically normal dogs were those with no known disease, including dogs admitted to the teaching hospital for wellness examinations, elective surgery, or orthopedic surgery, or dogs used in a teaching laboratory. These dogs had no history of ocular problems, and no ophthalmic abnormalities were detected during physical examination; however, funduscopic examinations were not performed on all dogs.

Diagnostic imaging studies—Diagnostic imaging was performed in 15 dogs with SARDS. A CBC, serum biochemical analyses, and radiography of the thorax (3 radiographic views) were performed prior to anesthetizing dogs for CT. Contrast-enhanced CT of the pituitary gland and ultrasonography of the adrenal glands were performed during general anesthesia. Ultrasonographic evaluation of the adrenal glands was performed with a 7.0-MHz electronic sector scanning transducer.^d Dogs were positioned in lateral recumbency. The left adrenal gland was examined from a left dorsal approach, and the right adrenal gland was examined from a right dorsal or right intercostal approach. Maximum length and caudal pole thickness were measured by use of the ultrasound machine software. Any subjective abnormalities of shape, echogenicity, or echotexture were recorded.^17,18 For CT, dogs were positioned in sternal recumbency. Computed tomography was performed with a fourth-generation CT scanner^e with technique settings of 120 kV (peak) and 80 mA and a small field of view. Contiguous transverse slices were obtained by use of a slice thickness of 1 mm. A repeat scan series, with the same parameters and location, was performed after IV administration of nonionic iodinated contrast medium^f (370 mg I/mL). Imaging studies were not performed in 2 dogs with SARDS because of severe cardiomegaly noted on thoracic radiographs and a perceived increased general anesthesia risk.

Western blot analysis—Blood was obtained from 17 dogs with SARDS, 53 dogs with neoplasia, and 57 clinically normal dogs for collection of serum, which was frozen at −80°C. Retinas were obtained from young adult mixed-breed dogs, all of which had a complete ophthalmic examination including biomicroscopy and indirect ophthalmoscopy performed 1 day prior to tissue collection. Immediately after euthanasia for a terminal surgery laboratory, both eyes were enucleated, each globe was incised around the equator, the vitreous was removed, and the retina was gently teased away from the periphery toward the optic disc and then cut at its attachment to the optic disc. Retinas were immediately frozen at −80°C. For western blot analysis,¹⁹ retinas were thawed and weighed, and 10 times the volume equivalent of 62.5mM Tris-HCl (pH, 6.8) was added. The mixture was sonicated 3 times for 10 seconds each to create a retinal homogenate. Electrophoresis was performed by use of a 10% SDS polyacrylamide gel. Approximately 180 μg of retinal homogenate protein was applied to each minigel with 0. 75-mm thickness. Proteins were electrophoretically transferred from the gel to a nitrocellulose membrane,^g which was cut into strips and stored in low-fat milk blocking buffer (0.1M Tris, 0.15M NaCl, 0.05% NaN₃, 0.1% Tween-20, and 5% low-fat powdered milk). Each strip contained approximately 9 μg of retinal proteins. For primary antibody incubation, individual strips were placed in serum dilutions of 1:200, 1:400, 1:1,000, and 1:3,000 for each dog. The negative control strip was incubated in low-fat milk blocking buffer alone. Positive controls included recoverin antibodies made in rabbits²⁰ diluted 1:1,000 in low-fat milk blocking buffer and arrestin antibodies made in mice^h diluted 1:1,000 in lowfat milk blocking buffer. After 16 to 18 hours' incubation at 4°C, all strips were washed individually in rinse buffer (100mM NaCl, 10mM Na₂HPO₄, 15mM NaN₃, and 1mM EDTA) twice for 15 minutes each. Secondary antibody was biotinylated goat anti-dogⁱ diluted 1:1,000 in antibody buffer (100mM NaCl, 10mM Na₂HPO₄, 15mM NaN₃, 1mM EDTA, and 0.1% bovine serum albumin) for sera and negative control, biotinylated goat anti-rabbit^j diluted 1:1,000 in antibody buffer for the antirecoverin positive control, and biotinylated horse anti-mouse^k diluted 1:1,000 in antibody buffer for the antiarrestin positive control. Secondary antibody incubation was for 1 hour. All strips were washed together in rinse buffer twice for 15 minutes each. Tertiary reagent was streptavidin horseradish peroxidase^l diluted 1:1,000 in antibody buffer for 1 hour. All strips were washed together in rinse buffer twice for 15 minutes each. Diaminobenzidine^m (0.02%) with nickel enhancement was used for substrate incubation for 3 to 6 minutes. After air drying, those bands with visibly equivalent or greater densities than 1 or both positive control bands at the 1:1,000 dilution, and present at the 1:3,000 dilution, were described as prominent bands (Figure 1).

View Full Size

Results of western blot analysis using canine retinal proteins and serial serum dilutions of 1:200, 1:400, 1:1,000, and 1:3,000 from 2 clinically normal dogs and 2 dogs with neoplasia. Positive controls (open arrows) included (a) antiarrestin 48 kd and (b) antirecoverin 23 kd. Solid arrows indicate examples of bands described as prominent; at the 1:1,000 dilution (a), the band has an equal or greater density than 1 or both positive controls, and the band can be detected at the 1:3,000 dilution (b).

Citation: American Journal of Veterinary Research 67, 5; 10.2460/ajvr.67.5.877

• Download Figure
• Download figure as PowerPoint slide

Results

Median age of dogs with SARDS (n = 17) was 8 years (range, 3 to 14 years). Ten were spayed females, 6 were castrated males, and 1 was a sexually intact male. There were 4 Dachshunds, 3 mixed-breed dogs, 2 Chinese Pugs, and 1 each of Maltese, English Springer Spaniel, Miniature Schnauzer, Yorkshire Terrier, Brittany Spaniel, Golden Retriever, American Cocker Spaniel, and Labrador Retriever.

Median age of clinically normal dogs (n = 57) was 6 years (range, 2 to 15 years); 30 (53%) were spayed females, 16 (28%) were castrated males, 7 (12%) were sexually intact females, 2 (3.5%) were sexually intact males, and the sex of 2 dogs had not been recorded. The major breeds represented (> 2 dogs) included 17 (30%) mixed-breed dogs, 6 (11%) Labrador Retrievers, and 3 (5.3%) each of Shih Tzu, Dachshund, and English Bulldog.

Median age of dogs with neoplasia (n = 53) was 10 years (range, 3 to 17 years); 26 (49%) were spayed females, 25 (47%) were castrated males, and 2 (4%) were sexually intact males. The major breeds represented (> 2 dogs) included 15 (28%) mixed-breed dogs, 7 (13%) Labrador Retrievers, 4 (8%) Boxers, and 3 (6%) each of Golden Retriever, Rottweiler, and Great Dane. Malignant lymphoma was diagnosed in 26 (49%) dogs; mast cell tumor was diagnosed in 8 (15%) dogs; osteosarcoma was diagnosed in 3 (5.6%) dogs; unspecified carcinoma, transitional cell carcinoma, and oral melanoma were diagnosed in 2 dogs each; and anal sac carcinoma, perianal carcinoma, pulmonary carcinoma, thyroid carcinoma, hepatic carcinoma, carcinomatosis, hemangiosarcoma, spindle cell sarcoma, transmissible venereal cell tumor, and peripheral nerve sheath tumor were diagnosed in 1 dog each.

Lymphopenia was detected in 3 of 17 dogs with SARDS. Abnormalities detected on serum biochemical analyses in dogs with SARDS included high activities of serum alkaline phosphatase in 9 of 17 dogs, high activities of serum alanine aminotransferase in 6 of 17 dogs, and high concentrations of serum cholesterol in 8 of 17 dogs.

Pulmonary neoplasia was not detected radiographically in any of the 15 dogs with SARDS. No pituitary gland abnormalities were detected on CT. The P:B ratio was less than 0.31 in 14 of 15 dogs and 0.72 in 1 dog. Adrenal glands in 14 dogs were ultrasonographically normal. Adrenal glands in 1 dog were bilaterally symmetrical and mildly large; the shape and echogenicity of both adrenal glands were ultrasonographically normal.

On western blot analysis, bands were identified in sera of all dogs tested. To help distinguish bands that may have indicated notable immune reactivity from lighter bands detected in all dogs, those bands classified as prominent were further evaluated. Prominent bands were detected in 16 of 57 (28%) clinically normal dogs, in 21 of 53 (40%) dogs with neoplasia, and in 7 of 17 dogs with SARDS (Table 1). Prominent band location varied from > 48 kd to < 23 kd; however, > 50% of dogs had prominent bands at or immediately adjacent to the 48-kd site. The number of dogs with prominent bands that were of greater intensity than control bands included 6 of 16 clinically normal dogs, 9 of 21 dogs with neoplasia, and 2 of 7 dogs with SARDS. Of those prominent bands that were of greater intensity than control bands, ≥ 50% were located near the 48-kd (arrestin) site. No prominent bands were identified at the 23-kd (recoverin) site. The median age of clinically normal dogs with prominent bands was 4.5 years, with 8 of 16 dogs 4 years of age or younger; 6 were mixed-breed dogs, 3 were Dachshunds, and 2 were English Bulldogs. Within the group of dogs with neoplasia (n = 21), dogs with prominent bands included 9 (43%) mixed-breed dogs, 3 (14%) Labrador Retrievers, and 2 (9.5%) Golden Retrievers. Of dogs with neoplasia with prominent bands, lymphoma was the most common neoplasia (38%), followed by mast cell tumors (29%) and carcinomas (29%). Of the SARDS dogs with prominent bands, there were more castrated males (n = 4) than spayed females (3), and mixed-breed dogs were not represented. Duration from onset of blindness to serum collection in SARDS dogs with prominent bands ranged from 2 weeks (1 dog) to 3 months (3 dogs).

Discussion

In the study reported here, the signalment of dogs with SARDS was similar to that in other studies^2,5,n in which SARDS was confirmed via electroretinography and studies^b,c in which a diagnosis of SARDS was made on the basis of clinical findings with or without electroretinography. Age-matched controls were not used in our study, and the median age for clinically normal dogs (6 years) was less than the median age of dogs with SARDS (8 years) and dogs with neoplasia (10 years); however, the age range was nearly identical in all 3 groups. In dogs with SARDS and neoplasia, signalment did not have any appreciable effect on results of western blot analysis. Interestingly, young (4 years old) dogs were more likely to have prominent bands, and Dachshunds, an overrepresented breed for SARDS, had a higher incidence of prominent bands, compared with other clinically normal dogs. Results of CBC and serum biochemical analyses were consistent with results of other studies,^1,2,5 with the most common abnormalities being lymphopenia, high serum activities of alkaline phosphatase and alanine aminotransferase, and high serum cholesterol concentrations.

In our study, results of thoracic radiography, ultrasonography, and CT were unremarkable and suggested that underlying neoplasia of the pituitary or adrenal glands, or primary lung neoplasia, as a source of ectopic retinal protein production was not present. It is possible that even with contrast enhancement, a pituitary gland microadenoma may not have been detected via CT.^21,22 Dynamic contrast-enhanced CT may have been more sensitive than contrast-enhanced CT for detection of microadenomas^23,24; however, this technology was not available for our study. Kooistra et al²⁵ found the P/B ratio useful for identifying large pituitary glands in dogs. Although the P:B ratio has not become a standard means of pituitary gland evaluation, the P:B ratio in 14 of 15 dogs with SARDS measured in our study was within the reference range (0.14 to 0.31) and was greater than the reference range in only 1 dog. Given that the results of diagnostic imaging of pituitary and adrenal glands were unremarkable, the association between retinopathy and hyperadrenocorticism-like clinical signs in dogs with SARDS is not known. Results of diagnostic imaging performed in our study did not exclude the presence of occult neoplasia elsewhere in the body.

A limitation of our study was the lack of ophthalmic examinations by a diplomate of the American College of Veterinary Ophthalmologists on all clinically normal dogs and dogs with neoplasia. Because the scope of the study encompassed 3 sites, the board-certified ophthalmologist was not always available at the time of blood collection. To minimize the effect that undetected retinopathy may have had on overall results of western blot analysis, a large number of dogs were included in the clinically normal and neoplasia groups. Undetected retinopathy would be expected to have a higher incidence in older dogs than young dogs; however, this was not detected in our study, particularly among clinically normal dogs in which half of the dogs with prominent bands were between 2 and 4 years of age.

There was no association between duration of onset of blindness and collection of blood and the presence or strength of bands on the western blot. In 1 dog, the first blood sample for serum was obtained < 2 weeks after the onset of blindness, and a second blood sample was obtained 7 weeks later. Results of the western blot were identical at both times.

Detection of multiple bands on western blots of all dogs tested was not expected. This may have been caused by the use of vascularized tissue extract, which may have caused nonspecific antibody binding on the western blot. However, this would be expected to cause reaction at the light-chain (25 kd) and heavy-chain (50 kd) sites for IgG and to cause similar bands on the negative control attributable to IgG binding with the goat anti-dog antibody. Bands identified in all dogs varied widely in location, and the negative controls had no bands. The finding of nonspecific bands is consistent with what Thirkill et al^o described as normal antibody activity with retina on western blot using sera from clinically normal humans. In that study, antibody activity was detected in serum samples in all 100 patients from < 20 kd to > 62 kd including areas corresponding to 4 known CAR antigens (62 kd, 45 kd, 40 kd, and 20 kd), but no reactivity was detected at the 23-kd site. In our study, we used the intensity of the positive control bands as a reference for a definitive retinal protein antibody-antigen reaction to assess bands that may have been relevant in each dog's serum. To help distinguish prominent bands, 4 serum dilutions for each dog were used and the strength of band reactions was compared with positive controls for that gel. The percentage of prominent bands was similar in dogs with neoplasia and SARDS and higher than that in clinically normal dogs. A limitation to comparing the percentage of prominent bands found in dogs with neoplasia and clinically normal dogs with dogs with SARDS was that the number of dogs with SARDS was relatively small, compared with the other 2 groups. The high number of prominent bands in dogs with neoplasia, compared with clinically normal dogs, may reflect an overall increase in circulating immunoglobulins in dogs with neoplasia because none of those dogs had clinical signs of retinal dysfunction. Dogs with mast cell tumors and carcinomas were overrepresented with prominent bands and dogs with lymphoma were underrepresented. However, the most prominent bands were seen in dogs with lymphoma and carcinomas.

Although 7 dogs with SARDS had prominent bands, unlike human patients with CAR, there was no consistent location of bands that would indicate a specific retinal protein-antibody interaction. On the basis of results of western blot analysis, it appears that SARDS and CAR do not have similar etiologies. The acute and total shutdown of photoreceptor function with subsequent destruction of outer segments in the absence of inflammation may indicate an immune-mediated phenomenon, although identification of the target protein and detection of antibodies remain unknown.

The 48-kd region had increased immunologic activity in clinically normal dogs, dogs with neoplasia, and dogs with SARDS. The clinical relevance of this in dogs is not known, although it also has been detected by other investigators.^p The similar finding seen in all 3 groups of dogs suggested that this immunologic activity was not associated with retinal disease. The 48-kd location may be a site of normal immunoreactivity in dogs, which may be important for future interpretation of western blot analysis in dogs with retinopathy.