Materials and Methods

Sample Population—Serum was obtained from 50 healthy pet dogs of various breeds (median age, 4.0 years; range, 0.6 to 11.6 years). Food was withheld from these dogs for at least 12 hours before collection of blood samples. The protocol for collection of blood samples was reviewed and approved by the Clinical Research Review Committee at Texas A&M University (CRRC No. 2006-30). For use of the dogs in the study, owner consent was obtained for each dog. Serum was harvested within 2 hours after collection of blood samples and stored at −80°C until further use.

Fecal samples were collected from 30 healthy pet dogs of various breeds (median age, 3.7 years; range, 0.8 to 11.0 years) by their owners and from 12 healthy, kennel-housed research Beagles (median age, 8.3 years; range, 4.0 to 11.1 years). All 42 dogs were vaccinated (at a minimum) against canine distemper virus, canine adenovirus type 2, canine parvovirus, and rabies virus except for 1 dog, within 12 months prior to collection of fecal samples. Also, all 42 dogs had been dewormed at regular intervals and had not received any medications known to affect the gastrointestinal tract.

Collection and processing of fecal samples—Fecal samples were obtained by collection after observed defecation. After collection, feces were placed in preweighed polypropylene tubes^a and immediately frozen at −20°C until further use. Feces were thawed, and aliquots of 1.3 g of feces (wet weight) were diluted 1:5 in fecal extraction buffer (20mM CH₃CO₂Na and 3mM CaCl₂ [pH, 7.6]) containing a proteinase inhibitor cocktail^b (1 tablet/25 mL). After homogenization by vigorous shaking for 30 minutes at approximately 23°C, suspensions were centrifuged for 20 minutes at 2,100 × g at 5°C. Supernatants were collected by use of serum filters^c and centrifuged for 30 minutes at 10,600 × g at 23°C. The final supernatants (fecal extracts) were stored frozen at −80°C until further use.

Production of tracer—A tracer was produced by labeling pure cCP with ¹²⁵I by use of the chloramine T method.¹⁷ Briefly, previously purified cCP^d (0.85 mg/mL) in 20mM CH₃CO₂Na and 3mM CaCl₂ (pH, 7.6) was dissolved in 0.25M sodium phosphate (pH, 7.5). Chloramine T (2 mg/mL), sodium metabisulfite (0.4 mg/mL), potassium iodide (2 mg/mL), and 10% (wt/ vol) BSA were dissolved in 0.05M sodium phosphate (pH, 7.5). The radioiodination mixture contained 8.5 μg of cCP, 8.5 μL of ¹²⁵I (0.85 mCi at the time of production), and 10 μL of chloramine T. After the mixture was stirred for 20 seconds at approximately 23°C, 100 μL of sodium metabisulfite and 860 μL of potassium iodide were added. To separate ¹²⁵I-labeled cCP and free ¹²⁵I, the mixture was loaded onto a dextran G-25 column^e equilibrated with 10% (wt/vol) BSA. The RIAB (0.05M sodium phosphate, 0.02% [wt/vol] NaN₃, and 0.5% [wt/vol] BSA [pH, 7.5]) was used as the mobile phase, and 1-mL aliquots were collected. Fractions containing protein were tested with antiserum and selected on the basis of the NSB and tracer-binding ratio for B₀ (ie, B₀/TC). The protein fraction with the lowest NSB at a B₀/TC of approximately 50% was diluted in RIAB to yield a solution with approximately 30,000 CPM, and it was then stored until used to perform an RIA.

Specific activity of the tracer was estimated by use of a method described elsewhere.¹⁸ Effects of various storage conditions on the stability of radiolabeled cCP were investigated during a period of 45 days.¹⁹ Aliquots of tracer were stored at 4° or −20°C at 2 radiochemical concentrations (approx 0.2 and 2.3 mCi/L) in polypropylene containers^f and also at 4°C and approximately 0.2 mCi/L in glass (noninert) containers.

Production of polyclonal antibodies and concentration test—Polyclonal antibodies against cCP were generated in 2 New Zealand White rabbits (Oryctolagus cuniculus). Both rabbits were repeatedly inoculated with pure cCP (in 20mM CH₃CO₂Na and 3mM CaCl₂ [pH, 7.6]) in accordance with a technique described else where.²⁰ The initial SC inoculation consisted of 200 μg of cCP emulsified in complete Freund's adjuvant and was followed by 2 booster SC inoculations with 150 μg of cCP in incomplete Freund's adjuvant at intervals of 3 weeks. Ten days after each inoculation, blood samples were collected from each rabbit to evaluate antibody concentrations in the sera. Because sera obtained from both rabbits revealed an insufficient antibody response, the protocol for antibody production was modified, and both rabbits received 200 μg of cCP in incomplete Freund's adjuvant administered SC into the prescapular nodal area at intervals of 1 month. Subsequently, one of the rabbits was injected SC in the prescapular nodal area at monthly intervals with 150, 250, 250, 150, and 150 μg of cCP in incomplete Freund's adjuvant. Blood was collected 10 days after each inoculation, and antiserum was tested.

Polyclonal antibodies were purified by affinity chromatography by use of cCP coupled to N-hydroxysuccinamide–activated beaded agarose^g and screened for IgG by use of reducing SDS-PAGE.^h Reactivity of purified anti-cCP IgG and antiserum was verified by precipitation testing against pure cCP, canine serum, and BSA by use of radial double immunodiffusion in agar gel (Ouchterlony technique).ⁱ

A simplified RIA was used to evaluate antisera in serial dilutions to obtain a B₀/TC of approximately 50%²¹ and a low NSB (< 5%). Antiserum obtained after the fifth booster injection from the rabbit that received the most inoculations was selected for the RIA procedure (final dilution, 1:11,100).

Development of RIA—The RIA procedure was conducted at approximately 23°C in polypropylene tubes. Each tube contained 100 μL of standard solutions of pure cCP (serial 2-fold dilutions from 1,000.0 to 3.9 μg/L), a serum sample (1:4 dilution) or fecal extract (1:80 dilution) in RIAB, 100 μL of antiserum diluted 1:3,700 in RIAB containing 0.05% polyoxyethylene sorbitan monolaurate, and 100 μL of tracer (approx 30,000 CPM). Standard solutions were prepared from concentrated pure cCP (0.85 mg/mL, as determined by use of the Bradford dye-binding method^j) in 20mM CH₃CO₂Na and 3mM CaCl₂ (pH, 7.6). The B₀ was determined by replacing cCP standard by RIAB. Tubes for NSB determination contained 200 μL of RIAB with a final concentration of 0.025% polyoxyethylene sorbitan monolaurate and 100 μL of radiolabeled cCP. Tubes for TC determinations contained 100 μL of tracer. Standards, serum, and fecal samples were prepared in duplicate, whereas tubes for TC, NSB, and B₀ were prepared in quadruplicate. Tubes were vortexed, and the mixtures incubated for 3 hours. To all tubes except the TC tube, 100 μL of normal rabbit serum^k diluted 1:100 in RIAB and 1 mL of commercially available precipitating solution^l containing the second antibody were added. The mixtures were vortexed and then centrifuged for 30 minutes at 3,360 × g at 10°C to allow separation of free and bound fractions. Supernatants from all tubes, except the TC tubes, were carefully decanted, and the remaining pellets were washed with 1 mL of RIAB. Radioactivity in all tubes was measured for 120 seconds in an automated gamma counter^m (82% counting efficiency). A statistical software packageⁿ was used to calculate a quartic polynomial curve fit by use of the following equation: y = f [x] = a + (b × x) + (c × x²) + (d × x³) + (e × x⁴), where y is the dependent variable, a through e are coefficients describing the shape of the curve, and × is the independent variable. On the x-axis, concentrations of cCP standard solutions were recorded on a logarithmic scale, and values on the y-axis were calculated by use of the following equation²²: y = (Bstandard/B₀) × 100, where Bstandard represents the CPM for each standard solution (after subtraction of the NSB), and B₀ is the CPM obtained for the B₀ tube (after subtraction of the NSB). Results obtained from unknown samples were plotted against the calibration curve to determine sample concentrations of cCP in serum samples or fecal extracts. Test samples containing cCP in concentrations outside the standard range of the assay were further diluted by a factor of 2 and reassayed. Aliquots of diluted serum samples and fecal extracts with low, moderate, and high cCP concentrations were stored frozen at −80°C; they were included in each use of the RIA as quality-control samples to detect shifts in the assay.

Analytic validation of the RIA—The RIA was analytically validated by determination of analytic sensitivity, working range, linearity, accuracy, precision, and reproducibility by evaluating the minimum and maximum concentration detectable by the assay, dilutional parallelism, recovery of added cCP (spiking recovery), and intra-assay and interassay variability. Analytic sensitivity of the RIA was determined²³ by analyzing 10 duplicates of the B₀ tubes in 1 assay and calculating the mean and SD of precipitate counts for the zero cCP concentration. The cCP concentration that corresponded to the mean count minus 3 SDs transposed onto the standard curve was defined as analytic sensitivity (lower limit of the working range) of the RIA. The upper limit of the working range was determined by evaluating (in 1 assay) 10 duplicates of a sample containing cCP in a concentration 7 times the highest standard and calculating the mean and SD of precipitate counts at 7,000 μg/L. The cCP concentration that corresponded to the mean count plus 3 SDs transposed onto the standard curve was defined as maximum cCP concentration detectable in canine serum and fecal extracts of the RIA (upper limit of the working range).

To determine linearity, accuracy, precision, and reproducibility of the RIA, samples with differing cCP concentrations were tested. Assay linearity was determined by use of 6 serum samples (3 samples were evaluated at dilutions of 1:2, 1:4, 1:8, and 1:16, respectively, and 3 samples with higher cCP concentrations were evaluated at dilutions of 1:8, 1:16, 1:32, and 1:64, respectively). Linearity of the assay for canine fecal extracts was assessed by evaluation of 6 samples at final dilutions of 1:100, 1:200, 1:400, 1:800, and 1:1,600. Assay accuracy was tested by spiking 6 sera and 6 fecal extracts with known concentrations of cCP (0, 31.3, 62.5, 125.0, 250.0, and 500.0 μg/L and 0, 3.1, 6.3, 12.5, 25.0, 50.0, and 100.0 μg/g for sera and fecal extracts, respectively). The percentage of standard antigen recovery was calculated as follows: (observed value/expected value) × 100. Precision of the assay was evaluated by assaying 6 serum samples and 6 fecal extracts 10 times within the same assay, followed by calculating the intra-assay CV as follows: CV = (SD/mean) × 100. Reproducibility of the assay was determined by analyzing 6 sera and 6 fecal extracts in 10 consecutive assays and calculating interassay CVs.

Existence of a position effect within the assay (end-of-run effect) was tested in sera from 5 dogs by analyzing each sample 70 times within the same assay. Mean cCP concentration calculated from the first 35 results was compared with the mean cCP concentration of the last 35 results by use of a Wilcoxon signed rank test. Species specificity of the RIA was validated by evaluating sera from 3 cats at dilutions of 1:2, 1:4, 1:8, and 1:16.

Reference interval and stability of cCP in serum or plasma—A reference interval for serum cCP concentration was established by evaluating serum samples from 50 healthy pet dogs and calculating the central 95th percentile. Serum cCP concentrations were compared between healthy male and female dogs by use of a Mann-Whitney U test. Serum cCP concentrations were compared among dogs of various age groups (< 2, 2 to 3, 3 to 5, 5 to 7, and > 7 years) by use of a Kruskal-Wallis test. Paired cCP concentration in serum and plasma (EDTA was used as an anticoagulant) obtained from 25 pet dogs, which were submitted for use in an unrelated study, was analyzed by use of a Wilcoxon signed rank test for paired datum points and calculation of a Spearman rank sum correlation coefficient.

Stability of cCP in 21 serum samples and 14 plasma samples that had been stored frozen at −20°C for 15 weeks was evaluated by repeated quantification and comparison of cCP concentrations by use of a Wilcoxon signed rank test. On the basis of the results obtained, cCP stability was determined by use of a Friedman test in another 14 serum samples that had been stored frozen at −20°C for 5 and 10 weeks.

For all statistical tests, significance was defined as values of P ≤ 0.05. All P values represented 2-sided tests.

Within-subject variation and stability of cCP in fecal samples—To evaluate distribution of cCP in feces, the within-subject variation within a single defecation was assessed by comparing cCP concentrations in 4 dogs. Five spot samples (1.3 g/sample) were collected from various portions of feces from 1 defecation of each dog and extracted as described. The remainder of the feces for the 4 dogs was diluted 1:1 in fecal extraction buffer and thoroughly blended^o for 5 minutes. Three aliquots of this suspension were obtained, and extracts were prepared as described (final dilution, 1:5). The cCP concentration was quantified in all extracts. Variability of fecal cCP concentrations was assessed for each dog by calculating the CV.

Within-subject variation over time was evaluated by collecting fecal samples on 5 consecutive days from each of the 42 healthy dogs. The cCP was extracted and quantified from all fecal samples. To identify the most efficient strategy for sample collection and determine the minimum number of samples needed for reliable identification of dogs with high fecal concentrations of cCP, various combinations of sampling days (fecal samples collected on day 1; days 1 and 2; days 1, 2, and 3; days 1, 2, 3, and 4; and days 1, 2, 3, 4, and 5) were evaluated. Variability of fecal concentrations of cCP over time was assessed for each dog by calculating the mean fecal concentration of cCP and within-subject variation (ie, CV) for various combinations of sampling days, followed by statistical comparison by use of a Friedman test.

Finally, a reference interval for fecal concentration of cCP was established from the central 95th percentile of the mean fecal concentration of cCP in 3 samples (collected on days 1, 2, and 3) from each of the 30 healthy pet dogs. Mean fecal concentration of cCP for 5 sampling days was compared between healthy pet dogs of various age groups (< 3, 3 to 6, and > 6 years) by use of a Kruskal-Wallis test. Fecal concentrations of cCP were compared between healthy pet dogs and healthy, kennel-housed research Beagles by use of a Mann-Whitney U test. Stability of cCP in feces was evaluated by quantifying cCP in extracts prepared from 5 fecal samples (thoroughly blended to ensure consistency among aliquots in regard to cCP concentration) from 4 healthy dogs. The blended samples had been stored at −80°, −20°, 4°, and 23°C for 3, 6, 9, 12, and 15 days. Fecal concentrations of cCP in aliquots stored at those 4 temperatures for the 5 durations were compared by use of a Friedman test.

Results

Immunization with pure cCP in accordance with the modified vaccination protocol yielded a moderate titer of IgG anti-cCP antibodies in only 1 rabbit; antiserum in excess (1:500 dilution) bound slightly more than 85% of radiolabeled cCP. This antiserum and purified anti-cCP IgG resulted in single strong lines of immunoprecipitate when used in agar gel double-diffusion tests against pure cCP and canine sera, but no immunocrossreactivity against BSA was detected.

Specific activity of the tracer ranged from 62 to 114 Ci/mmol (mean 91 Ci/mmol; n = 8 tracers). Repeated determination of the B₀/TC and sample analysis revealed the lowest stability of radiolabeled cCP when stored at 4°C in glass containers (Figure 1) . Decomposition was less pronounced in polypropylene containers, and improvements in stability were evident when the tracer was stored at −20°C. Optimal tracer stability was achieved for storage at −20°C in polypropylene tubes with approximate radioactivity of 0.2 mCi/L. When stored under these conditions, radiolabeled cCP yielded a fairly good sensitivity of the RIA up to 32 days with no effect on NSB, which allowed the tracer to be used for 1 month without refractionation.

View Full Size

Stability of radiolabeled cCP as measured by use of the B₀/TC when samples were stored in polypropylene tubes at 0.2 mCi/L at −20°C (black circles) or 4°C (white circles), in polypropylene tubes at 2.3 mCi/L at −20°C (black squares) or 4°C (white squares), and in glass containers at 0.2 mCi/L at 4°C (white triangles). Results represent the B₀/TC, which is the percentage of radiolabeled cCP bound to antiserum (final dilution, 1:11,100) for the B₀ tubes.

Citation: American Journal of Veterinary Research 69, 7; 10.2460/ajvr.69.7.845

• Download Figure
• Download figure as PowerPoint slide

A typical standard curve yielded a working range between 7 and 3,194 μg/L (Figure 2) . Because serum and fecal concentrations of cCP usually were > 70 μg/L and 3 μg/g, respectively, samples were appropriately diluted before quantitation (1:4 and 1:400 for serum and feces, respectively). Hence, analytic sensitivity of the RIA for serum and fecal samples was 29 μg/L and 2.9 μg/g, respectively, with a maximum detection limit of 12,774 μg/L and 1,277 μg/g, respectively. The serial dilution of serum and fecal samples with a wide range of cCP concentrations yielded dose-response curves parallel to the standard curve (Table 1). Expected and observed values for spiking recovery of the assay were highly correlated (Table 2). To test precision and reproducibility of the RIA, mean, SD, and intra-assay and interassay CV values were determined (Table 3). No end-of-run effect was detected when a sample was evaluated up to 70 times within the same assay (ie, mean cCP concentration of the first 35 aliquots of a sample did not differ significantly [P = 0.063] from the mean concentration of aliquots 36 to 70). No change in the standard curve was evident after aliquots of cCP standard dilutions had been stored at −80°C for up to 10 months. Although plasma cCP concentrations did not change significantly (P = 0.542) after 15 weeks of storage at −20°C, serum cCP concentrations decreased significantly (P < 0.001) after storage for the same number of weeks at the same storage temperature. However, serum cCP concentrations did not change significantly (P = 0.395) after storage at −20°C for 5 or 10 weeks.

View Full Size

Representative calibration plot for the estimation of cCP concentrations by use of a competitive, double-antibody RIA. A portion of the curve (gray shading) was used to quantify cCP in serum and fecal extracts. The B_standard/B₀ is determined, whereby B_standard represents the CPM for each standard solution (after subtraction of the NSB) and B₀ is the CPM obtained for the B₀ tubes (after subtraction of the NSB).

Citation: American Journal of Veterinary Research 69, 7; 10.2460/ajvr.69.7.845

• Download Figure
• Download figure as PowerPoint slide

This novel RIA was able to detect calprotectin in feline sera. The serial dilution of 3 feline serum samples yielded observed-to-expected ratios between 96.0% and 134.2% (mean ± SD, 113.9 ± 13.9%).

Serum cCP concentrations in samples from 50 healthy pet dogs ranged from 76 to 1,292 μg/L (median, 241 μg/L); the reference interval for serum cCP concentration was established as 92 to 1,121 μg/L (Figure 3) . Median serum cCP concentrations did not differ significantly (P = 0.313) between healthy male and female dogs. Because there were too few sexually intact dogs, a possible correlation between reproductive status and serum cCP concentration could not be investigated. Median serum cCP concentrations did not differ significantly (P = 0.451) among various age groups of healthy dogs. Although significantly (P < 0.001) higher cCP concentrations were detected in plasma, compared with the cCP concentrations in serum (n = 25 paired samples), there was no correlation between serum and plasma concentrations (r = 0.078; Spearman rank sum correlation; Figure 4).

View Full Size

Scatterplot of serum cCP concentrations in specimens obtained from 50 healthy pet dogs. Each symbol represents the concentration for a specific dog. The median cCP concentration (solid horizontal line) and reference interval (dashed horizontal lines) were calculated.

Citation: American Journal of Veterinary Research 69, 7; 10.2460/ajvr.69.7.845

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Scatterplot of cCP concentrations in paired serum and plasma samples obtained from 25 dogs. The median cCP concentration measured in plasma samples in which EDTA was used as an anticoagulant (2,011 μg/L) was significantly (P < 0.001) higher, compared with the median cCP concentration measured in serum samples (563 μg/L). The median cCP concentration (solid horizontal line) is indicated for serum and plasma samples, but the reference interval (dashed horizontal lines) is indicated only for the serum samples.

Citation: American Journal of Veterinary Research 69, 7; 10.2460/ajvr.69.7.845

• Download Figure
• Download figure as PowerPoint slide

For cCP concentrations in 5 spot samples collected from 1 defecation, the CV ranged from 18.4% to 59.8% (mean ± SD, 37.4 ± 14.9%; Table 4). For mean fecal concentrations of cCP in samples obtained during a 5-day sample collection period and in various numbers of samples for each of the 42 healthy dogs, the CV ranged from 0% to 165.9% (mean ± SD, 59.3 ± 42.4%) and from 0% to 178.3% (mean ± SD, 54.7 ± 41.5%), respectively. Although mean fecal concentrations of cCP were not significantly (P = 0.303) different among various combinations of sampling days, the respective CVs were significantly (P < 0.001) different between 2- and 3-day sample collections and between 2- and 4-day sample collections. Fecal concentrations of cCP in 5 samples from each of the 30 healthy pet dogs ranged from < 2.9 to 234.2 μg/g (median, 14.4 μg/g), with mean cCP concentration for 5-day sample collections ranging from < 2.9 to 122.1 μg/g (median, 15.5 μg/g). Mean cCP concentrations for 3-day sample collections in 30 healthy pet dogs ranged from < 2.9 to 142.4 μg/g (median, 20.1 μg/g); the reference interval for fecal concentrations of cCP was established as the mean for 3-day sample collections of < 2.9 to 137.5 μg/g (Figure 5) . Median values for mean fecal concentrations of cCP for 5-day sample collections did not differ significantly (P = 0.314) among healthy pet dogs of various age groups. Compared with concentrations for healthy pet dogs of various breeds, the mean fecal concentrations of cCP for 5-day sample collections were significantly (P = 0.050) higher in healthy, kennel-housed research Beagles (median, 36.5 μg/g). Median age differed significantly (P = 0.002; Mann-Whitney U test) between healthy pet dogs and kennel-housed research Beagles. Duration of storage of fecal samples at 4 temperatures did not cause significant changes in fecal concentrations of cCP (P = 0.674, 0.234, 0.674, and 0.282 for storage at −80°, −20°, 4°, and 23°C, respectively).

View Full Size

Scatterplot of fecal concentrations of cCP in samples obtained from 30 healthy pet dogs. Each symbol represents the mean cCP concentration for a 3-day sample collection for a specific dog. See Figure 3 for remainder of key.

Citation: American Journal of Veterinary Research 69, 7; 10.2460/ajvr.69.7.845

• Download Figure
• Download figure as PowerPoint slide

Discussion

An RIA for the measurement of cCP in canine serum samples and fecal extracts was successfully developed. No end-of-run effect was detected for up to 70 aliquots of the same sample. Thus, the assay may be suitable for analysis of large numbers of samples. Analysis of results for the study reported here indicated that 50 μL of serum or 1.3 g of feces is required for accurate determination of cCP with this novel RIA and that specimens can be stored frozen for up to 10 weeks (serum) and at least 2 weeks (feces) for subsequent analysis. Moreover, cCP was stable in fecal samples from healthy dogs for up to 2 weeks at ambient temperature (approx 23°C). This finding potentially eliminates the necessity to ship frozen fecal specimens for laboratory analysis of cCP concentrations.

Sensitivity of the assay was calculated as 29 μg/L of serum or 2.9 μg/g of feces, respectively, which appears to be adequate when considering the reference intervals are 92 to 1,121 μg/L for serum samples and < 2.9 to 137.5 μg/g for fecal extracts, respectively. Our intent was to create a standard curve that encompassed a wide range of cCP concentrations; thus, a relatively concentrated antiserum was chosen for the RIA, which resulted in a loss in assay sensitivity. Considering reference intervals for both types of specimens, the dilutions used for samples, and the purpose of the assay to identify dogs with increased serum or fecal concentrations of cCP, this loss in sensitivity was regarded as inconsequential. An upper limit of the assay working range of more than 3 times the upper limit of the respective reference interval, and an upper limit of the reference interval within the lower third of the respective assay working range, indicated that the RIA has the potential to be highly exact in differentiating between clinically normal dogs and dogs with inflammatory conditions.

Observed-to-expected values for dilutional parallelism and spiking recovery for serum samples and fecal extracts of dogs indicated linearity and accuracy of the RIA. Precision and reproducibility of the assay were evident by the CVs for intra-assay and interassay variability. Because of the unavailability of cCP analogues, analytic specificity of the RIA for cCP could only be determined by use of dilutional parallelism and spiking recovery of cCP in serum and fecal samples from dogs. As more canine S100 proteins or analogues become available in the future, additional research into the specificity of the novel cCP RIA will need to be conducted to clarify this issue.

In contrast to serum calprotectin concentrations in humans,²⁴ serum cCP concentrations were not significantly higher in male dogs. However, it must be emphasized that most of the male dogs tested were neutered.

Lack of correlation between paired serum and plasma concentrations of cCP indicated that serum and plasma concentrations of cCP cannot be quantified interchangeably, which is consistent with results in humans.⁹ A possible explanation could be time- or temperature-dependent activation of complement that has been reported²⁵ in stored whole blood and plasma, which is then followed by C5a-induced release of calprotectin from PMNs, as determined in vitro.²⁶ Consequently, varying the amount of time between sample collection and analysis could cause variations in plasma cCP concentrations. In contrast, serum clotting activates PMNs, which results in a release of PMN-activation markers (such as calprotectin), irrespective of time and temperature. Another possibility could be EDTA-mediated chelation of Ca²⁺, which is known to inhibit antigen-antibody reactions. In an attempt to increase specific uptake of cCP during the RIA, RIAB was supplemented with EDTA, which resulted in considerably higher cCP concentrations (data not shown). A high variability of cCP concentrations in plasma samples in the study reported here would suggest that cCP concentrations should be evaluated in serum samples.

Difficulties in generating antiserum against cCP in rabbits by vaccination in accordance with a protocol that has been used successfully for generating polyclonal antibodies against a multitude of antigens were unexpected because human calprotectin has been described as highly immunogenic in rabbits.²⁷ However, the protocol for antibody production in that study²⁷ differed slightly from the one in the study reported here with regard to intervals of injection and the adjuvant. Interestingly, modifying the injection site and increasing the amount of antigen inoculated, both of which are known to increase polyclonal antibody titers against weak immunogens, yielded a moderate anti-cCP titer in 1 rabbit. One possible explanation for an insufficient antibody response against cCP in rabbits could be extensive homology of the amino acid sequence of the antigenic sites of cCP and rabbit calprotectin (S100A8, 75.3%; S100A9, 69.6%²⁸). Conservation of the human pendants, however, was similar at 73.0% and 62.3% for dogs and rabbits, respectively.²⁸ Because the antibody response was similar in both rabbits, idiosyncrasy would appear to be unlikely. In contrast, in Ca²⁺-containing storage buffer, both cCP proteins (cS100A8 and cS100A9) assemble predominantly to a heterotetrameric complex of approximately 45 kd.^d Presumably, this increases the immunogenicity of cCP because its monomeric forms (10.3 and 14.6 kd) are close to the minimum size for an immunogen (3 to 5 kd).

Polypropylene containers were used to store cCP tracer, which decreased the loss attributable to storage to half of the loss during storage in noninert glass containers. This is consistent with a marked affinity of human calprotectin for glass surfaces in the presence of Ca²⁺ ions²⁷ and, presumably, accelerated catalytic degradation by the noninert surface.

Concentrations of cCP in fecal extracts from individual healthy dogs vary greatly; this could possibly be explained by a patchy emigration of cCP-expressing cells into the mucous membranes of the gastrointestinal tract, which has been reported in cats.²⁹ Because prolonged duration in the gastrointestinal tract may result in accumulation or concentration of fecal proteins, gastrointestinal transit time needs to also be considered as a factor in variation of fecal concentrations of cCP over time. Collection of several fecal samples from the same dog may counterbalance this physiologic variation. However, owner compliance and cost need to be considered to arrive at a reasonable compromise between reliability and economy. Consequently, within-subject variation over time was evaluated, and on the basis of the results of our study, analysis of 3 fecal samples collected on consecutive days would appear to satisfy this concern.

Significantly higher fecal concentrations of cCP in the Beagles of the research colony, compared with concentrations in pet dogs of various breeds, indicated a necessity to reevaluate the reference interval for fecal concentrations of cCP in dogs in research colonies. Because fecal concentrations of cCP did not differ significantly among pet dogs of various age groups, this finding cannot be attributed to the significantly higher age for the dogs in the research colony. A breed-specific, higher fecal concentration of cCP in Beagles needs to be considered.

Serum and fecal concentrations of cCP in healthy pet dogs were not normally distributed and contained 3 and 2 outlying datum points for serum and fecal samples, respectively. One possible explanation could be a relationship between body condition score and circulating inflammatory markers, which has been described in humans.^30,31 However, because we believed this would have been beyond the scope of the study reported here, data on body condition scores of each dog were not obtained. An association between increased serum or fecal concentrations of cCP and vaccination also needs to be considered, especially when a modified-live viral vaccine or bacterial vaccine is used.

Detection of feline calprotectin by this assay was to be expected because cross-reactivity between cCP and feline calprotectin has been reported.^d Moreover, linearity of the RIA for feline serum samples substantiates the potential application for the novel RIA in quantification of calprotectin in cats.

The RIA described in this report was sensitive, linear, accurate, precise, and reproducible, and it appears to have the potential to distinguish healthy dogs from dogs with inflammatory conditions, such as IBD. Further research into the clinical usefulness of the measurement of serum or fecal concentrations of cCP in dogs with inflammatory conditions, such as IBD, is being conducted.