Abstract
Serum cardiac troponin-I (cTnI) has been validated as a biomarker for cardiotoxicity in numerous animal models; however, owing to sensitivity issues cTnI concentrations in healthy, resting animals used in toxicology studies have not been established. Serum from healthy and isoproterenol hydrochloride (iso)-treated rats, dogs, and monkeys were assayed using the Erenna system. The Erenna cTnI assay provided sensitivity < 1 ng/L across human, rat, dog, and monkey cTnI. Linear responses (R 2= 0.99) were observed for all species. Precision studies yielded interassay CVs of curve fit quantification from 2% to 4% between 1.6 and 5000 ng/L, and 23% at 0.78 ng/L. Strong correlation (R 2= 0.99) was obtained between Erenna and Beckman Access cTnI. Concentrations of cTnI in healthy animals ranged from 1 to 9 ng/L. In longitudinal studies of iso-treated animals, the concentrations of cTnI in the control vehicle-treated groups were 10–20 ng/L for rats (N = 10) and predose values of 2–3 ng/L for dogs (N = 3). Measured with the Erenna assay system, cTnI was quantifiable at all time intervals tested in all animals treated with iso. The Erenna system provides sensitive measurement of cTnI in rats, dogs, and monkeys, makes it possible to determine small changes from normal concentrations, and provides cTnI values from small volumes of serum.
Introduction
Blood cardiac troponin-I (cTnI) concentration has been validated as a biomarker for cardiotoxicity in numerous animal models, including rat, dog, and monkey (Adin et al. 2006; Burgener et al. 2006; Feng et al. 2005; Gaze and Collinson 2005; Herman et al. 2006; O’Brien 2006; O’Brien et al. 2006). Currently a wide variety of immunoassays and platforms are available to measure human cTnI concentrations; however, little has been done to validate these systems with regard to sensitivity across species using an internationally recognized standard (Adin et al. 2006; Peetz et al. 2006; Storti et al. 2006). Furthermore, few of these systems have demonstrated sufficient sensitivity and precision to quantify the low levels of cTnI in apparently healthy, resting animals that might be used in toxicology studies. The International Life Sciences Institute/Health and Environmental Sciences Institute Cardiac Troponins Expert Working Group compared nine commercial assays used for measurement of cTnI and cTnT in animals and identified differences in species cross-reactivity, assay precision, and dynamic range (Apple et al. 2005, 2007; Pettit et al. 2007). Some assays were suboptimal for use in select species. A further challenge is that many of these assay platforms, such as the reference method used in this study, require > 100 μL of serum or plasma per determination, which makes testing small volumes of serum (e.g., individual rodents) challenging. In this report we describe the use of the Erenna System cTnI immunoassay, along with the NIST cTnI reference material, for quantifying low and high concentrations of cTnI in rat, dog, and monkey models of chemically induced cardiotoxicity, as well as its ability to measure cTnI in rat samples of low or limited volume.
Materials and Methods
The Erenna Immunoassay System (Singulex, Alameda, CA, USA), based on single-molecule counting, and the Erenna cTnI immunoassay, which uses a paramagnetic microparticle (MP) solid-phase assay format, have been described in detail previously (Todd et al. 2007; Wu et al. 2006). The current version of the assay, which was designed for use in a batch mode and has not yet been submitted for review by regulatory bodies, requires approximately two hours elapsed time and is used primarily as a research tool. Dog, monkey, and rat purified cTnIs were obtained from Hytest (Finland). Human cTnI complex #2921 was obtained from NIST (Gaithersburg, MD, USA). Briefly, the Erenna cTnI assay was performed as follows: 50 μL serum (either neat or prediluted in calibrator diluent, e.g. 2.5 μL serum + 47.5 μL calibrator diluent) or calibrator were added to 150 μL assay buffer containing MPs coated with biotinylated capture antibody. The resulting mixture was incubated for one hour in a 96-well plate. The MPs were then magnetically separated, washed, and incubated with 20 μL fluorescent dye-labeled detection antibody for thirty minutes. After five washes via magnetic separation, 20 μL of elution buffer was added, and eluted detection antibody was separated from MPs using a 384-well filter plate. The eluate was then read in the Erenna instrument system. A single lot of reagents was used in this study. Animal serum and calibrators were tested in duplicate, and the results were presented as the average of the two measurements. All animal serum were provided by Lilly Research Laboratories (Greenfield, IN, USA) and tested in a manner blinded to reference test results at Singulex (Alameda, CA, USA).
Animals
Animals were housed and fed as described previously (Higgins et al. 2003; Schultze et al. 2005; Schultze and Sullivan 2007).
Rats
Male Fischer 344 rats (F344/NHsd) were obtained from Harlan Industries, Inc. (Indianapolis, IN, USA). Five- to six-week-old male rats that weighed 80–110 g were housed individually in stainless steel cages under conditions of controlled temperature (22.2°C ± 4.4°C), humidity (20%–80%), and light cycle (light:dark, 12:12 hours). Rats were fed chow (Teklad Certified Rodent Diet, Lot 8728C, Harlan Teklad Global Diets, Madison, WI, USA) and tap water ad libitum.
Dogs
Male and female, Beagle dogs that weighed 8–14 kg were obtained from Marshall Farms (North Rose, NY, USA). Dogs were housed individually in stainless steel cages with plastic-coated grate flooring under conditions of controlled temperature (22.2°C ± 4.4°C), humidity (20–80%), and light cycle (light:dark, 12:12 hours). Dogs were fed approximately 400 g Harlan Teklad Global 2021 21% Protein Dog Diet (Madison, WI, USA) daily and allowed water ad libitum.
Monkeys
Young adult male and female cynomolgus monkeys (Macaca fasciularis) that weighed 2.5–5.0 kg were obtained from Charles River BRF, Inc. (Mauritius). Monkeys were housed individually in aluminum cages under conditions of controlled temperature (22.2°C ± 4.4° C), humidity (20%–80%), and light cycle (light:dark, 12:12). Monkeys were fed biscuits (Certified Primate Diet 5048, PMI Nutrition International, Inc., Brentwood, MO, USA) daily, supplemented with fresh fruit three times weekly, and allowed water ad libitum.
Study Design
Groups of five rats were given single subcutaneous injections of vehicle, 0.5 mg iso/kg, or 8 mg iso/kg. At 0.25, 0.5, one, two, four, eight, or twenty-four hours post-dose, groups of five rats were anesthetized, bled, humanely euthanatized, and tissues were processed for histopathologic examination. Dogs were given 4 μg iso/kg/min at an infusion rate of 0.167 mL/kg/min for approximately twenty minutes. Blood was collected from the jugular vein at one, two, four, eight, twenty-four, and forty-eight hours after iso administration. Monkeys were given 4 μg iso/kg/min at an infusion rate of 0.167 mL/kg/min for approximately twenty minutes. Blood was collected from the femoral vein at two, four, eight, twenty-four, forty-eight, seventy-two, and ninety-two hours after iso administration. Because the dogs and monkeys were not euthanatized at study conclusion, no histologic examination of heart tissue was conducted.
Institutional Compliance Statement
Rats, dogs, and monkeys were housed in facilities at Lilly Research Laboratories (Greenfield, IN, USA) that are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. All study protocols were approved by the Eli Lilly Institutional Animal Care and Use Committee.
Reference cTnI Determinations
The concentration of cTnI in frozen serum was determined using the Beckman Access cTnI assay (Fullerton, CA, USA) at Lilly Research Laboratories (Greenfield, IN, USA). A minimum of 140 μL (40 μL sample volume plus 100 μL instrument dead space) serum was required for each determination. This assay is suitable for both STAT and batch applications and has an elapsed time of approximately fifteen to twenty minutes.
Calculations of Linearity and Sensitivity
The linearity and sensitivity of the response for rat, dog, and monkey cTnI concentration were determined by preparing twofold serial dilutions of each species cTnI and testing the dilutions in the Erenna assay (using a standard curve prepared with human cTnI). The resulting data were plotted as signal versus expected as well as measured versus expected cTnI concentration (ng/L), and linear regression analysis was performed. The limits of detection (LoDs) were determined by the method of two standard deviations of signal obtained with the calibrator diluent not containing cTnI (replicates of twenty-four) divided by the slope of the regression line (signal vs. expected cTnI concentration).
Results
Erenna cTnI Assay Characterization
In this study a preliminary evaluation of analytical assay performance was performed. Limits of detection (LoDs) < 1 ng/L were obtained for each of the four species tested. Specifically, the linear responses and LoDs generated with each species were as follows: human (NIST), y = 0.89x + 2, 0.19 ng/L; rat, y = 0.66x + 0.2, 0.14 ng/L; monkey, y = 0.2x + 0, 0.72 ng/L; dog, y = 0.56x + 0, 0.58 ng/L. Linear responses (R 2= 0.99) were observed for all species. Precision of quantification of the curve fit was determined over six consecutive assay runs (over six days) using the human NIST standard. Between 1.6 and 5000 ng/L, the interassay CV of quantification ranged from 2% to 4%, and 0.78 ng/L precision was 23%. Linearity and precision of the NIST standard curve fit across the low end of the reportable range (62–0.25 ng/L) in these six assay runs were described by the linear regression line y = 1.0x − 0.18 with an r 2value of 0.999. Taken together, these preliminary data demonstrate that the Erenna immunoassay provides sensitive measurement of rat, dog, and monkey cTnI concentrations across a broad dynamic range. The human NIST standard was used in all subsequent Erenna assays to determine the concentrations of cTnI in animal serum.
cTnI Concentrations in Healthy Animals
Values obtained from healthy animals were 5.3 ng/L (male monkey, N = 3), 4.4 ng/L (female monkey, N = 3), 1.0–4.6 ng/L (dog, N = 4), and 9.5 ng/L (rat, N = 20). In longitudinal studies of iso-treated animals, the concentrations of cTnI in the control vehicle-treated groups were 9–20 ng/L for rats (N = 10) and predose values of 2–3 ng/L for dogs (N = 3). The percentage range from the mean (% CV of duplicates; standard deviation of two determinations/mean value) ranged from 1% to 7%, demonstrating good reproducibility. Identical samples tested with the reference method yielded < 20 ng/L (dogs), 20 ng/L for three rats, and < 20 ng/L for seven rats; monkey samples were unavailable for measurement.
Concordance with the Reference Method
Concordance of cTnI measurements made with the Erenna and Beckman Access were performed using serum obtained from animals in the iso study. Linear regression analysis (using samples that provided quantitative values for both assay methods) was performed, and the resulting lines and correlation coefficients comparing all measurements were as follows: rats, y = 0.68x + 145, R 2= 0.95 (seventy-seven samples); dogs, y = 3.64x − 112, R 2= 0.96 (twelve samples); monkeys, insufficient number of samples for analysis (monkey #1 all samples undetectable Beckman, and monkey #2 most samples > 5000 ng/L Singulex and > 10,000 ng/L Beckman). In addition to monkeys, the three dogs in this study showed wide variations in iso response, with four-hour peak cTnI values of 210, 590, and > 5000 ng/L measured with the Erenna assay. Results obtained from the rat comparison are depicted in Figure 1as Bland Altman bias plots. In this experiment, 2.5 μL and 140 μL of serum were used with the Singulex and Beckman assays, respectively. These data demonstrate that a small bias exists between the assays in quantification of rat cTnI (Beckman Access 1.4-fold average greater for all values considered and 1.19-fold average greater for values < 1000 ng/L). Taken together, although these results indicate differences in individual animal responses and small differences in bias between the different species and platforms, the actual numbers (i.e., trends in absolute values) are well correlated, with R 2values of 0.95–0.96.
Sample Volume
The impact of variable serum volume usage was evaluated with the Erenna assay. Preliminary experiments were performed using dilutions of rat serum covering a range of 5–50 μL sample volume. Recovery (comparison of sample diluted in calibrator diluent vs the 50 μL sample volume as the reference volume) of cTnI ranged from 85% to 107% (data not shown). In the next set of experiments, concordance in cTnI values obtained from a group of rats and using 2.5 μL and 50 μL of serum from the same animal was performed. In the case of 2.5 μL serum, the resulting values were multiplied by twenty to account for the volumetric differences before data analysis. Correlation and linear regression analysis was performed on the lowest cTnI values obtained from twenty-five rats in the iso study. Using 2.5 μL of serum, all samples provided measurable cTnI (range 3–441 ng/L) with a linear regression correlation of R 2= 0.97 (2.5 μL vs 50 μL serum). These results demonstrate that 2.5 μL of rat serum can be used to accurately measure cTnI with the Erenna assay.
Iso Study
In the iso study and measured with the Erenna assay, cTnI concentration was quantifiable at all time intervals tested (pre-dose, fifteen minutes to seventy-two hours) in all animals. As early as fifteen minutes post-administration, marked increases from baseline cTnI values, followed by decreases, were observed. Representative data, using the rat response (2.5 μL of serum) as an example, are shown in Figure 2. Cardiac TnI values obtained for identical specimens using the Beckman Access assay (140 μL of serum) are presented, as well. In the predose rats, cTnI was measurable in all ten animals with the Erenna assay and measurable in only three animals (all in the 0.5 mg/kg iso dose, Figure 2) with the Access assay. The values for dog and monkey were from individual animals, whereas the values for rat were the mean value of five rats/time point. Both assays showed parallel increases and decreases in cTnI values.
Correlation of cTnI Concentrations with Histopathologic Examination of Hearts
Minimal to marked increases in the concentration of cTnI occurred at one hour, two hours, four hours, and eight hours after administration of both 0.5 mg/kg and 8.0 mg/kg of iso, indicating cardiac myocyte injury and leakage of cardic troponin into the blood. The magnitude of the increases was higher in the rats given the 8.0 mg/kg dose. Cardiac TnI concentrations returned to baseline for both dose groups by the twenty-four-hour time point. The highest concentrations of cTnI occurred at the four- and eight-hour time points.
Histopathologic findings in the hearts of rats administered either 0.5 or 8 mg/kg of iso were similar in character and distribution but were slightly more severe in the animals given 8 mg/kg (Table 1). Myocardial lesions were most easily recognized in samples collected twenty-four hours after dosing and consisted of discrete and variably sized areas of acute myocardial necrosis scattered throughout the ventricles, with a predilection for the subendocardium and papillary muscles. The areas of acute necrosis were characterized by fragmentation of myocytes, inflammatory cell infiltration, interstitial edema, and acute hemorrhage. Histologic changes were less obvious in samples collected at the thirty-minute to eight-hour time points and consisted of occasional individual or small clusters of hyaline fibers (swollen, hypereosinophilic myocytes) and small foci of myocytes with minimal fragmentation, interpreted as peracute necrosis. Inflammatory cell infiltration at these time points was rare. In samples collected at fifteen minutes post-dose, no definitive microscopic changes were identified.
Discussion
We have demonstrated that the Erenna system is sensitive for measuring cTnI concentrations with analytical sensitivities (LoDs) < 1 ng/L, has a reporting range from 0.2 ng/L to > 5000 ng/L and can quantify rat, dog, monkey, and human cTnI concentrations, with similar reactivity across the various species. Furthermore, interassay precision of CVs < 10% at values down to 1.6 ng/L and 23% at 0.78 ng/L support its use to accurately quantify and measure changes of cTnI at low concentrations. Of note, cTnI values in rats, dogs, and monkeys were highly correlated when measured with the Singulex Erenna and Beckman Access systems. This finding is important in the context of the work performed by The International Life Sciences Institute/Health and Environmental Sciences Institute Cardiac Troponins Expert Working Group wherein performance of nine different commercial cTnI assays, including the Beckman Access, were evaluated (Apple et al. 2005, 2007; Pettit et al. 2007).
We have also shown, in small populations of serum, cTnI concentrations in the 1–20 ng/L range in healthy rats, dogs, and monkeys. These findings are preliminary, and further studies with larger populations of animals are required for the establishment of reference ranges. Nevertheless, these values are similar to those that have been reported for humans and support the models put forth by Anversa et al. on myocardium homeostasis and turnover, wherein low concentrations of cTnI should be found in blood (Anversa et al. 2006; Anversa et al. 2007; Wu et al, 2006). Our findings with various species of cTnI need to be confirmed using larger numbers of animals to establish the reference range for each species.
The measurement of cTnI concentrations in animals has been validated as a tool to detect and monitor cardiotoxicity, whether chemically or physically induced (Feng et al. 2005; Gaze and Collinson 2005; Hemalatha et al. 2006; Kurata et al. 2007; O’Brien 2006; O’Brien et al. 2006). In this study, iso-induced cardiac injury was confirmed histologically in rats. At the earliest time points, histopathological changes were not observed or were subtle, even with clear increases in cTnI concentrations. We also demonstrated individual animal variability in the response to treatment with iso in that different animals of the same species, given the same dose of iso, generated varying magnitudes of cardiotoxicity and cTnI concentrations. For example, in three dogs one hour post-iso treatment, cTnI concentrations were increased to 92, 219, and > 5000 ng/L compared to 2–3 ng/L pre-dosing. In two monkeys, values of 33 and > 5000 ng/L were noted two hours post-dosing, which gradually decreased to 1.1 and 187 ng/L by forty-eight hours, respectively.
Assay serum volume requirements are of importance in studies of small animals, especially when it is desirous to obtain and analyze samples obtained across an interval of many closely timed blood draw events. We have shown that the Erenna system provides similar results in the rat when paired 2.5 μL and 50 μL serum samples were tested. Furthermore, strong concordance in cTnI values was observed in iso-treated rat samples when 2.5 μL was used in the Erenna and 140 μL was used in the Beckman Access assays. Assay sensitivity is directly proportional to sample volume. In the case of the Erenna assay, a twenty-fold reduction in sample volume resulted in a shift of assay LoD from 0.14 ng/L to approximately 2.8 ng/L. This shifted LoD remains lower than the cTnI levels noted in healthy animals, especially rats (10–20 ng/L). Thus, it was possible to quantify cTnI concentration in healthy, nontreated rats using 2.5 μL of serum. This ability could provide significant advantages in serial monitoring of small animals. In conclusion, the Erenna cTnI assay is standardized against reference material provided by NIST, and initial data suggest accurate sensitivity at the sub-ng/L concentrations and accurate quantification from 0.78 to 5000 ng/L for human, rat, dog, and monkey serum cTnI. The sensitivity of this assay can be leveraged to provide accurate quantification of cTnI in serum samples as small as 2.5 μ L, which may provide significant advantages in serial longitudinal sampling during small animal studies. Taken together, these features provide a method for assessing potential cardiotoxicity by differentiating changes in cTnI concentrations from healthy, resting-state concentrations, which have been previously considered nonmeasurable.
Footnotes
Figures and Table
Acknowledgements
We thank Dr. Douglas Held, Ms. Laura Freitag, and Mr. Robert Freese for support in developing antibody reagents and assay format.