Materials and Methods

Animals—Fecal samples were collected from 23 privately owned pet dogs of various ages and breeds. All dogs were ≥ 1 year of age.

Collection of fecal samples—Special tubes^a intended for collection of fecal samples from human patients were used for collection of canine fecal samples. The tubes contained a spoon attached to the lid and a plastic rod. The weight of each empty tube was measured. Fecal samples were collected in the tubes provided. The exact amount of feces added was determined by weighing the filled tube and subtracting the weight of the empty tube. Mean ± SD amount of feces was 1.0 ± 0.2 g in each tube. Fresh fecal samples were collected from each dog at the time of the first defecation of the day.

IgA extraction from feces—Fecal samples were diluted 1:5 in a low-pH glycine-extraction buffer (1M glycine; pH, 2.5) containing a protease inhibitor cocktail^b (1 tablet/25 mL). Fecal samples were dissolved in glycine-extraction buffer by use of a vortex and incubated overnight at 4°C. Incubated samples were centrifuged at 2,500 × g for 20 minutes at 4°C. Plasma filter cones^c were used to separate the supernatant from the precipitate. Supernatant was transferred into 15-mL tubes and adjusted to a pH of approximately 7.4 by the addition of 800 μL of 1M Tris-HCl buffer (pH, 8.0). This mixture was vortexed and centrifuged again under the aforementioned conditions. One milliliter of the obtained supernatant was transferred into a 1.5-mL tube. Supernatants were centrifuged again at 2,500 × g for 20 minutes at 4°C, transferred into fresh 1.5-mL tubes, and stored at −20°C until IgA concentrations were measured.

ELISA—A sandwich ELISA for the quantification of IgA in fecal extracts was established and validated. The ELISA plates that contained 96 flat-bottom wells^d were coated with affinity-purified goat anti-dog IgA (Fc region) antibodies^e that served as capture antibody. The concentration of the capture antibody solution was 100 ng/well in 0.05M sodium–carbonate monohydrate buffer (pH, 9.6). The plates were incubated for 1 hour at 37°C with constant shaking and washed 4 times with Tris-buffered saline (0.9% NaCl) solution (pH, 7.4). Nonspecific binding sites were blocked with 200 μL/well of a milk-free blocking solution.^f Plates were incubated for 1 hour at 37°C with constant shaking and washed 4 times, as described previously. All plates were treated in the same fashion. Standards were applied in duplicates from the highest to lowest concentration. Blanks and 3 control samples with known IgA concentrations and unknown samples were applied. Standards were prepared by use of a 1:2 serial dilution of a solution of IgA^g (1,000 μg/L) in Tris-buffered saline solution (pH, 7.4) containing 1% bovine serum albumin and 0.05% polyoxyethylene sorbitan monolaurate (buffer A). Standards and control samples were thawed immediately before loading. Buffer A was used as a negative control. Unknown and control samples were prepared in a 1:4,000 dilution of buffer A. Wells were each loaded with 100 μL of the designated solution. Plates were incubated for 1 hour at 37°C with constant shaking and washed 4 times, as previously described. For detection of the captured antigen, plates were incubated with the secondary antibody solution containing biotinylated anti-dog IgA antibody^h (100 ng/well) in 100 μL of buffer A. After incubation for 1 hour at 37°C with constant shaking, plates were washed 4 times as described previously and incubated with 100 μL/well of horseradishperoxidase–labeled streptavidin solution (50 ng/mL). Plates were incubated for another hour at 37°C with constant shaking, washed 4 times, and then developed for 20 minutes with 100 μL/well of 3, 3', 5, 5'-tetramethylbenzidine dihydrochloride substrate solution.ⁱ The reaction was stopped by adding a solution of 0.5M sulfuric acid (100 μL/well). Plates were read at a wavelength of 450 nm with an automated platereader.^j Standard curves were calculated by use of a 4-parameter curve fit.

Analytic validation of ELISA—The assay was analytically validated by determination of sensitivity, linearity, accuracy, precision, and reproducibility with fecal samples that were processed and then stored at −20°C until further use. Assay sensitivity was determined by calculating the mean response of 10 sets of blanks and evaluating the theoretic absorbance of the mean plus 3 SDs on the standard curve. Linearity of the assay was established by evaluation of each of the 4 samples at a 1:2,000 dilution and at dilutions of 1:4,000, 1:8,000, and 1:16,000. Spiking recovery, as a measure of accuracy, was determined by adding 0.0, 0.08, 0.16, 0.31, 0.63, 1.25, 2.50, and 5.00 mg/mL of canine IgA standard solution to the 1:4,000 diluted fecal samples. Intra-assay variability was determined by evaluating the 4 diluted fecal samples 10 times within the same assay run. Interassay variability was determined by evaluating the 4 diluted fecal samples in 10 consecutive assay runs. Intra- and interassay variabilities were calculated as CV (ie, CV% = [SD/mean] × 100).

Intraindividual variation of fecal IgA concentrations— Intraindividual variation within a fecal sample was assessed by collection of 5 fecal samples from different sections of feces from a single defecation for 5 dogs. Intraindividual variation over time was evaluated by collection of 15 fecal samples from defecations on separate days for each of the 18 healthy dogs. Fecal samples that were freshly voided from the first defecation of the day were collected in five 3-day sample collection periods as follows: days 0, 1, and 2; days 14, 15, and 16; days 28, 29, and 30; days 42, 43, and 44; and days 56, 57, and 58. Immunoglobulin A was extracted from all fecal samples and was quantified by use of the previously described ELISA method. Variability of fecal IgA concentrations was assessed for each dog by calculation of CV.

Statistical analysis—Fecal IgA concentrations of the 18 dogs from which fecal samples were repeatedly collected were analyzed by calculation of the CV for fecal IgA concentrations and comparison of mean fecal IgA concentrations.^k Various groupings of samples were evaluated to identify the most efficient sample collection strategy. It was important to determine the minimum number of samples needed that would reliably identify dogs with persistently low fecal IgA concentrations. As a first approach, a mean fecal IgA concentration for each dog was calculated from all 3 samples collected within each of the 5 sample collection periods, for a total of 15 samples/dog. Mean fecal IgA concentrations were also calculated from 3 fecal samples collected during only 1 of the 5 sample collection periods. A 4-sample mean fecal IgA concentration was calculated from samples collected on days 0, 1, 28, and 29. A 4-sample mean fecal IgA concentration was also calculated from samples collected on days 14, 15, 42, and 43. For each grouping of samples for calculation of the mean fecal IgA concentration, the intraindividual variation was calculated as the CV (ie, CV%). For each dog, mean fecal IgA concentrations obtained with the various groupings of samples were compared with the 15-sample mean fecal IgA concentration by use of a 1-way parametric ANOVA. A reference range for fecal IgA concentration in dogs was established from the central 95th percentile (2.5th to 97.5th percentile) of the 4-sample mean fecal IgA concentrations in 18 healthy dogs.

Results

IgA extraction from feces and ELISA—The described method for extraction of IgA from feces of dogs resulted in reproducible IgA concentrations when measured with the ELISA. Use of the ELISA resulted in reproducible standard curves. The sensitivity and lower limit of the working range of the assay for fecal IgA were 0.06 mg/g. Observed-to-expected ratios for serial dilutions of 4 fecal samples ranged from 111.1% to 114.9% (Table 1). Observed-to-expected ratios for spiking recovery for the same samples ranged from 67.9% to 95.9%. For intra-assay variability of 4 fecal samples, CVs were between 2.9% and 6.9%. For interassay variability of 4 fecal samples, CVs were 6.9% to 13.1%.

Intraindividual variation of fecal IgA concentrations—For IgA concentrations in 5 extracts from feces of the same defecation for 5 dogs, CVs ranged from 18.5% to 74.2% with a mean ± SD of 38.1 ± 21.9% (Table 2). For mean fecal IgA concentrations of the five 3-day sample collection periods for each of the 18 dogs, CVs ranged from 2.4% to 131.6% with a mean ± SD of 47.2 ± 29.2%. For mean fecal IgA concentrations of the two 2-day sample collection periods for each of the 18 dogs on days 0, 1, 28, and 29; days 15, 16, 43, and 44; days 14, 15, 42, and 43; days 28, 29, 56, and 57; days 1, 2, 29, and 30; and days 29, 30, 57, and 58, CVs ranged from 5.4% to 137% with a mean ± SD of 61.8 ± 30.8%. No significant (P = 0.632) difference was found between mean fecal IgA concentrations of the five 3-day sample collection periods (15 samples total) and those of the two 2-day sample collection periods (4 samples total). The reference range, determined by the central 95th percentile of the 4-sample mean fecal IgA concentration of 18 healthy dogs, was 0.22 to 3.24 mg/g.

Discussion

It has been previously reported that IgAD, the most common primary immunodeficiency in humans, also affects dogs.^6,9 However, at this time, no reliable diagnostic test is available to identify IgAD in dogs. Results of our study provide a feasible method for the quantification of fecal IgA in dogs. The described procedure allows for direct noninvasive measurement of intestinal IgA concentration. This represents a great advantage over methods in which intestinal IgA secretion is assessed from biopsy specimens.¹⁰ Immunoglobulin A concentrations in fecal extracts from dogs vary greatly. The high variability of IgA concentrations in fecal extracts is likely to be the result of the patchy distribution of intestinal IgA-producing plasma cells.^13,14 Fecal IgA concentration in dogs may also vary with diet, environment, level of stress, antigenic stimulation, gastrointestinal transit time, and vaccination status.¹⁵ This emphasizes the importance of the collection of several fecal samples from the same dog to counterbalance the physiologically occurring variation in fecal IgA concentrations. Although the more samples that can be selected the more reliable the result will be, dog owner's compliance in providing the samples, cost for shipping and analysis, and laboratory-processing capacities need to be considered. Therefore, in our study, we investigated the within-dog variability of fecal IgA concentrations over time.

A total of 15 fecal samples were collected from 18 dogs during 5 sample collection periods, with 2 weeks between periods. Our intention was to collect enough samples from 1 dog so that the IgA concentrations measured in the fecal extracts would settle around a mean value that indicates the true status of the intestinal immune response in a particular dog. On the basis of the variability of our results, analysis of a total of 4 fecal samples each collected on 2 consecutive days, with 28 days between sample collection periods, seems to satisfy this goal.

The accuracy of the ELISA may be decreased at high IgA concentrations. This can be concluded from the lowest ratio for a spiking recovery of 67.9% when the highest amount of IgA was added. Because the purpose of this assay is to detect dogs with severely low fecal IgA concentrations (< 0.06 mg/g), this should not diminish the usefulness of the ELISA to correctly identify these patients. With this in mind, the described extraction of IgA from feces and the use of the ELISA seem to result in a reliable quantification of IgA in the samples.

The molecular mechanism of IgAD is yet to be determined. It is likely that genetic and environmental aspects are involved.^16–18 Currently, no specific treatment for IgAD is available in humans or dogs. If infections occur as a result of IgAD, patients should be treated appropriately and recurrences should be anticipated.⁴

In our study, we report a practical method for the quantification of fecal IgA that takes into consideration the necessity to collect multiple fecal samples, yet balances this requirement with regard to laboratory costs and dog owner compliance in providing samples. Future studies need to investigate the incidence of IgAD in German Shepherd Dogs, compared with dogs of other breeds, by analyzing fecal IgA concentrations of a large number of German Shepherd Dogs and dogs of other breeds. If the results are suggestive of a hereditary component of this condition for German Shepherd Dogs, genetic studies may lead to the development of a genetic marker for IgAD, allowing the prevention of IgAD-positive dogs from further breeding.