Trolox (6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid), potassium chloride (KCl), sodium chloride (NaCl), potassium phosphate (KH2PO4), bathocuproinedisulfonic acid disodium salt, and copper(II) sulphate (CuSO4) were obtained from Sigma-Aldrich. Di-sodium hydrogen phosphate anhydrous (Na2HPO4) was obtained from Panreac.
The analyses were performed in the Cobas Mira Plus Biochemical Auto Analyzer (ABX Diagnostic) and in the Olympus AU400 Automatic Chemistry Analyzer (Olympus Europe GmbH).
Principle of the assay
The CUPRAC assay is based on the reduction of Cu2+ into Cu1+ by the action of the non-enzymatic antioxidants presented in the sample. The oxidant complex, consisted of Cu2+-bathocuproinedisulfonic acid (Cu2+-BCS) reacts with the antioxidants of the sample and is reduced to a Cu1+-bathocuproinedisulfonic acid (Cu1+-BCS), a stable complex which has a maximum absorbance at 490 nm . The antioxidant capacity of the sample is assumed to be equal to the extent of the complex Cu1+-BCS formation . The assay used for CUPRAC in the present study was based on the method described by Campos et al.  with some modifications.
Measurement procedure for the Cobas Mira Plus biochemical auto analyzer
In brief, a 5 μL volume of sample was pipetted. Then, 195 μL of the reagent 1 were added and a first read at 500 nm was taken. Subsequently, 50 μL of the reagent 2 were added to the reaction and incubated at 37 °C during 200 seconds. After incubation, a second read at 500 nm was taken and the difference between the first and the second read was used to calculate the antioxidant capacity of the sample. Distilled water was used for blanks.
Measurement procedure for the Olympus AU400 automatic chemistry analyzer
An amount of 195 μL of reagent 1 and 5 μL of sample were pipetted. A first read at 480 nm was taken before the addition of the second reagent. Then, 50 μL of reagent 2 were added to the mixture and incubated at 37 °C during 280 seconds. A second read at 480 nm was taken and the difference between the first and the second read was used for calculation of the antioxidant capacity of the sample. Distilled water was used instead of sample or standard for blanks.
Preparation of standards
Trolox solution at a concentration of 2.0, 1.0, 0.5, 0.25, 0.125, 0.0625 and 0.0 mmol/L were used. The results obtained for test samples were compared with a standard curve obtained with Trolox and were expressed as millimoles of Trolox equivalents per liter.
Optimization of reagents concentrations
To adjust the assay for measurements in canine serum, different concentrations of reagent 1 and reagent 2 were tested with the standards at different concentrations and also with different samples in the Cobas Mira Plus biochemical analyzer.
Reagent 1 was prepared at 0.1, 0.25, 1.0 and 1.6 mmol/L of bathocuproinedisulfonic acid disodium salt in 10 mmol/L phosphate buffer (pH 7.4) while reagent 2 was prepared at 0.1, 0.5, and 0.8 mmol/L of CuSO4 in ultrapure water.
The optimal concentrations were selected based on the production of a higher signal with lower background and a lower intra-assay imprecision calculated after analysis of one sample five times in one assay run.
Analytical validation of the assay
For the analytical validation of the CUPRAC assay, imprecision, accuracy, and sensitivity were evaluated following previously reported protocols [19–22].
Imprecision was expressed as coefficient of variation (CV) and was calculated as inter- and intra-assay variations. The CV was calculated as the standard deviation (SD) divided by the mean value (Xmean) of analyzed replicates x 100 % in the formula CV = (SD x 100 %)/ Xmean. To determine inter-assay variation, four serum samples were used. Inter-assay CV was determined by analyzing the same samples in separate runs performed on five different days. Five aliquots of each serum sample were stored in plastic vials at −20 °C until analysis. On the day of analysis, the samples were brought to room temperature prior to TACc measurement. The intra-assay CV was calculated after analysis of four samples five times in one assay run. Intra-assay CV tests were performed for all the different combination of reagents tested, although in the results only appear the values for the final concentration selected for the assay.
The accuracy was evaluated through assessment of linearity and spiking recovery. The linearity was evaluated by linearity under dilution, then duplicate determinations of TACc were made of a canine serum diluted at 1/2, 1/4, 1/8, 1/16 and 1/32 using ultrapure water. Dilution of a Trolox solution (2.0, 1.0, 0.5, 0.25, 0.125, and 0.0625 mmol/L) was also analyzed. For the spiking recovery, two canine serum samples with a known TACc concentration were mixed in different percentages (12.5, 25, 50, 75 and 87.5 %). The percentages of the measured TACc concentrations to the expected TACc concentrations were then calculated.
The detection limit was calculated on the basis of data from 20 replicate TACc determinations of ultrapure water as mean value plus 3 SDs. The lower limit of quantification (LLOQ) was calculated based on the lowest TACc concentration that could be measured within a CV less than 15 % .
Effects of hemolysis and lipemia
In order to examine the effect of hemolysis and lipemia, serum samples from three dogs were mixed with various concentrations of hemoglobin and lipids solution, respectively, and TACc was measured . To study the effects of hemolysis, fresh hemolysate was prepared by the addition of distilled water to packed, washed canine red blood cells from one dog, followed by centrifugation to remove cell debris. The hemoglobin concentration was adjusted to 80, 40, 20, 10, 5, and 0.0 g/L. Ten μL of each concentration were added to three 90 μL samples of canine serum to produce test samples with final hemoglobin concentration of 8, 4, 2, 1, 0.5, and 0.0 g/L, respectively. The 0.0 g/L concentration was reached by adding 10 μL of distilled water to 90 μL of the serum sample. Prepared samples were used to determine the TACc concentrations.
To investigate the effects of lipemia, a commercial fat emulsion (Lipofundina 20 %; Braun Medical S.S.) with triglycerides concentration of 200 g/L was serially diluted to 50, 25, 12.5, 6.25,3.125 and 0.0 g/L. Ten μL of each dilution were added to 90 μL of the serum samples and were used to determine the TACc concentration. The final concentrations of triglycerides in the samples were 5, 2.5, 1.25, 0.625, 0.3125, and 0.0 g/L (10 μL of distilled water were added to 90 μL of the serum samples).
The TACc measurements to evaluate the effect of hemolysis and lipemia in the assay were made in the Olympus AU400.
TACc levels were determined in healthy (control) dogs and dogs with inflammatory bowel disease (IBD). The control samples were from ten (seven males and three females) clinically healthy dogs of several different breeds aged between 3 and 8 years old, that were presented for routine checkups and had normal physical examination. Twelve dogs with IBD were included in this study. They were four female and eight male dogs aged between 3 and 8 years old also of different several breeds. A diagnosis of IBD was made on the basis of clinical signs (vomiting, diarrhea, and weight loss) of at least 3 weeks’ duration, and detection of lymphoplasmacytic inflammation during histologic examination of duodenal biopsy samples following the criteria of Ohta et al. . Exclusion of other causes of chronic gastrointestinal tract signs including urinalysis, abdominal ultrasound, fecal exam, complete blood count and serum biochemistry were made [25, 26].
Blood samples of the healthy and diseased dogs were collected from each via jugular or lateral saphenous venipuncture into tubes without anticoagulant. Samples were centrifuged at 3,500 x g for 5 min at 20 °C. The serum samples were stored in plastic vials at - 20 °C until analysis.
Arithmetic means, medians, intra- and inter-assay CVs were calculated by use of routine descriptive statistical procedures and computer software (Excel 2013, Microsoft; GraphPad Statistics Guide). Linearity under dilution was investigated by linear regression. To compare the TACc results from both analyzers a Spearman correlation coefficient was used. The influence of hemolysis or lipemia on TACc concentration was investigated by use of 1-way ANOVA and Dunnett posttests. Interferograms were prepared to show the differences in TACc concentrations when hemoglobin or lipids were added. Kolmogorov-Smirnov’s test was performed to assess normality of data. Comparison of the TACc concentrations between healthy dogs and dogs with IBD were made by use of Student’s t test once a parametric distribution was given. For all tests, P < 0.05 was considered as statistically significant.