Qualification of Cardiac Troponin I Concentration in Mouse Serum Using Isoproterenol and Implementation in Pharmacology Studies to Accelerate Drug Development

Abstract

Cardiac troponin I is a useful biomarker of myocardial injury, but its use in mice and application to early drug discovery are not well described. The authors investigated the relationship between cTnI concentration in serum and histologic lesions in heart tissue from mice treated with isoproterenol (ISO). Cardiac TnI concentrations in serum increased in a dose-dependant manner and remained increased twenty-four to forty-eight hours after a single administration of isoproterenol. Increased cTnI concentration was of greater magnitude and longer duration than increased fatty acid binding protein 3 concentration, aspartate aminotransferase activity, and creatine kinase activity in serum. Isoproterenol-induced increases in cTnI concentrations were both greater and more sustained in BALB/c than in CD1 mice and correlated with incidence and severity of lesions observed in heart sections from both strains. In drug development studies in BALB/c mice with novel kinase inhibitors, cTnI concentration was a reliable stand-alone biomarker of cardiac injury and was used in combination with measurements of in vivo target inhibition to demonstrate an off-target contribution to cardiotoxicity. Additional attributes, including low cost and rapid turnaround time, made cTnI concentration in serum invaluable for detecting cardiotoxicity, exploring structure–activity relationships, and prioritizing development of compounds with improved safety profiles early in drug discovery.

Introduction

Cardiac toxicity is anunintended and undesirable consequence of exposure to many drugs, including medications used to treat cancer, retroviral infection, diabetes mellitus, chronic obstructive pulmonary disease, inflammation, and psychoses (O’Brien 2006). Preclinical safety assessment in animals has demonstrated high predictivity of human cardiac toxicities, with up to 80% concordance between preclinical and clinical studies (Olson et al. 2000). Histological observation of heart lesions provides indisputable evidence of cardiac injury in drug development and preclinical studies and is routinely utilized during safety assessment. However, it is not commonly employed during pharmacology studies because it is relatively time-consuming and expensive compared to measurement of a serum-based biomarker. Electron microscopy, while more sensitive than histology, is even more time-consuming and expensive. In contrast, analysis for serum-based biomarkers is cost-effective and rapid. A sufficiently sensitive and specific biomarker can be used as a stand-alone safety endpoint to prioritize compounds at early stages in drug development. Unfortunately, traditional indicators of cardiac injury, such as aspartate aminotransferase (AST) and creatine kinase (CK) activities in serum, have limited sensitivity and lack specificity for cardiac injury.

Cardiac troponin I is a specific, sensitive, and quantitative measure of cardiac injury in humans (Adams et al. 1993; Babuin and Jaffe 2005) and veterinary species (O’Brien et al. 2006). Concentrations in serum are increased as a result of direct myocardial injury, myocardial ischemia, ventricular strain caused by disease (Fromm 2007), or drug-induced toxicity (Babuin and Jaffe 2005; Adamcova et al. 2007). Cardiac TnI has thus become the biomarker of choice for monitoring cardiac injury in humans (Collinson and Gaze 2007).

Cardiac TnI is a sarcomeric protein in the troponin complex of cardiomyocytes. The complex consists of troponins I (cTnI), the inhibitory subunit; T (cTnT), the tropomyosin binding subunit; and C (cTnC), the calcium binding subunit. Upon membrane disruption, intracellular cTnI leaks into the extra-cellular environment as a monomer, part of a binary complex with cTnC, or in a ternary complex of cTnT-cTnI-cTnC (Wu et al. 1998). Cardiac troponin I is distinguished from skeletal muscle isoforms of TnI by a cardiac-specific N-terminal sequence (Larue et al. 1992), which is the basis for its diagnostic specificity for myocardial damage.

Release of cTnI from myocardium was proportional to the size and extent of tissue injury in several animal models of cardiotoxicity (O’Brien et al. 2006) and has been reported to correlate well with systolic function (Adamcova et al. 2007). Challenges remain with regard to preclinical applications, including assessment of a suitable cutoff for irreversible myocardial injury, optimal timing of sample collection (Adamcova et al. 2005), relatively limited data sets for animals, and lack of standardization between various immunoassays available to measure cTnI concentration. Important differences exist in species-specific cross-reactivity of the antibodies used in the assays available for the measurement of cTnI concentrations in laboratory rats, dogs, and monkeys, as well as the precision and accuracy of the measurements (Apple et al. 2008). There is little information available regarding the use of these same assays in mice. These challenges contribute to the limited use of cTnI in toxicology studies, even those involving compounds with previously identified cardiotoxic potential. Furthermore, strains of mice used in discovery biology studies often differ from those used in toxicology studies, contributing to uncertainty in the interpretation of toxicology measurements in “nonstandard” strains of rodent.

To extend the application of cTnI as a biomarker of cardiac injury in preclinical species, we qualified cTnI concentration in serum as a biomarker of cardiotoxicity in CD-1 and BALB/c mice treated with a nonspecific β-adrenergic agonist, isoproterenol (ISO), a well-described model of cardiac necrosis (Zhang et al. 2008). Mice, because they require relatively small amounts of compound, are a preferred species for in vivo assessment of target inhibition and efficacy. In these studies, we explored the kinetics of alterations in cTnI concentrations in dose response and time course studies, as well as the relationship of cTnI concentrations to fatty acid binding protein 3 (Fabp3) concentrations, CK and AST activities, and severity of myocardial injury detected by histopathological examination. Fabp3, also known as heart-type fatty acid binding protein, is a marker of tissue injury (Pelsers, Hermens, and Glatz 2005) with reported utility as a biomarker of cardiac necrosis (Ruzgar et al. 2006), but its use has not been described in mice. CD-1 and BALB/c mice were chosen because of their common use in toxicology and pharmacology studies, respectively. The comparison of increased cTnI concentration in serum from both strains with histology provides assurance that the results of cTnI screening in BALB/c mice will be comparable to the results of a toxicology study in CD-1 mice.

Cardiac TnI concentration in serum was then deployed as a stand-alone safety marker to detect myocardial injury caused by novel kinase inhibitors in early drug discovery studies using BALB/c mice. In vivo target inhibition was assessed concurrently in these studies. This strategy resulted in the identification of unacceptably cardiotoxic compounds very early in the discovery process and enabled prioritization of compounds that had a favorable balance between target inhibition (potency) and safety.

Methods

Animals

Animal protocols were approved by Eli Lilly and Company’s Institutional Animal Care and Use Committee. Female CD-1 and BALB/c mice were obtained at nine to eleven weeks of age from Charles River Laboratories (Wilmington, MA) or Harlan Sprague Dawley, Inc. (Indianapolis, IN). Female mice were chosen for consistency with pharmacology studies in drug development, in which female BALB/c mice are used. Mice were housed singly with access to food and water ad libitum and were allowed to acclimate for at least one week to caging, feeding, and watering conditions before any handling. Blood was collected under isoflurane anesthesia via retro-orbital sinus or abdominal vena cava. Mice were humanely euthanized either by exsanguination under anesthesia followed by organ removal or by carbon dioxide inhalation followed by cervical dislocation.

Isoproterenol Studies

ISO (Sigma-Aldrich, St. Louis, MO) was dissolved in 0.9% sodium chloride and administered in a single, sc administration at 0.1, 1, 10, or 50 mg/kg to BALB/c and CD-1 mice. Control mice were treated with 0.9% sodium chloride. Dose volume was 5 ml/kg. In a series of seven studies, blood was collected from anesthetized mice via abdominal vena cava 1, 4, 8, 24, 48, or 72 hours after injection and processed for serum. Hearts were collected for histopathologic examination 24, 48, or 72 hours after injection.

Histopathologic Examinations in Isoproterenol Studies

Hearts were fixed for histological examination in 10% neutral buffered formalin, bisected longitudinally through both ventricles, processed, embedded (both halves) in paraffin, sectioned at a thickness of 5 μm, and stained with hematoxylin and eosin. Tissue sections were examined by light microscopy and evaluated for necrosis and inflammation (cellular infiltration, primarily mononuclear). Lesions were scored as follows: 0 = normal, 1 = minimal (very few, small foci affected, approx. 2 to 5 myofibers per histological section), 2 = slight (identifiable at low magnification but estimated to involve < 8% of tissue), 3 = moderate (easily identifiable, 8% to 20% of tissue affected), 4 = marked (20% to 50% of tissue affected), and 5 = severe (>50% affected). Interpretation was based on changes in individuals within each group, but for presentation purposes, necrosis severity scores were also averaged into a group score. Following initial evaluations to establish the principal changes, tissue sections were evaluated independently by two board-certified veterinary pathologists without knowledge of strain, time, or dose. The final scores represented the consensus of both evaluators. Photomicrographs were extracted at 8X and 20X from digital images taken on a ScanScope XT (Aperio Technologies, Inc., Vista, CA) using a 20X objective.

AST, CK, Fabp3, and cTnI

AST activity and total CK activity in serum were analyzed using a Roche/Hitachi 917 automated analyzer (Indianapolis, IN) according to the manufacturer’s protocol. Fabp3 concentration in serum was measured by a mouse/rat-specific ELISA (Cat# HK403 Hycult Biotechnology, Uden, the Netherlands) as previously described (Pritt et al. 2008). Cardiac TnI concentration in serum was measured by a mouse-specific ELISA (Cat# 2010-1-HS, Life Diagnostics, Inc., West Chester, PA). Standard and sample volumes were reduced to 25 μL from the manufacturer’s recommended 100 μL per well to conserve limited sample volume. Cardiac troponin I, purified from mouse hearts by ion-exchange chromatography in conjunction with calcium-dependent affinity chromatography on troponin-C agarose (Cat# 7110, Life Diagnostics, Inc., West Chester, PA), was used to construct a standard curve after dilution in naïve mouse serum provided with the kit. Concentration of purified cTnI for standards was verified by amino acid analysis as previously described (van Wandelen and Cohen 1997). The cTnI ELISA was qualified according to published criteria (DeSilva et al. 2003). Specifically, the quantifiable range was determined in mouse serum spiked with cTnI for which total error (mean bias + interrun covariance) was less than 30%. Samples from ISO studies with cTnI concentration > 10 ng/mL were diluted 1:2 and reanalyzed.

Baseline cTnI Concentration

Sera from vehicle-treated mice were assayed for cTnI concentration to assess baseline concentrations and what effect, if any, dosing and blood collection routes had. Blood was collected from anesthetized mice via retro-orbital sinus or abdominal vena cava. Mice received a single sc administration of vehicle (0.9% saline) in studies with ISO or 1 to 4 daily oral vehicle (1% carboxymethylcellulose or 10% acacia) administrations in drug development studies.

Discrimination of On- versus Off-Target Toxicity in Pharmacology Studies

In vivo target inhibition (IVTI) was measured 2 hours after a single oral administration of a novel kinase inhibitor (dose volume = 10 mL/kg) at a range of dose levels in BALB/c mice (n = 3 to 4 mice per dose). IVTI was measured as phosphorylation status of a signal transduction protein in peripheral blood mononuclear cells (PBMCs) using phospho-specific antibodies and flow cytometry procedures similar to those previously described (Krutzik and Nolan 2003; Krutzik et al. 2004). Compound concentrations in plasma were measured using mass spectrometry. A nonlinear, sigmoidal fit of percentage IVTI versus plasma compound concentrations (Prism, GraphPad Software, Inc., La Jolla, CA) provided a quantitative relationship between IVTI and plasma compound concentration that was used to interpret serum cTnI data from toxicology studies (see below).

Cardiotoxicity was assessed using cTnI concentrations in serum (n = 3 to 12) 4 hours after the last of 4 daily, oral administrations of a range of dose levels of each compound. Limited blood volume in BALB/c mice precluded measurement of percentage IVTI in toxicology studies. However, plasma compound concentrations were measured at 1, 2, 4, 8, 16, and 24 hours (n = 2 per time) after the first administration and used to interpolate percentage IVTI from the previously established relationship between plasma compound concentration and percentage IVTI (preceding paragraph). We then calculated percentage target inhibition × hour as area under the curve and plotted these data against group mean cTnI concentration in serum to assess the relationship between target inhibition and cardiotoxicity.

Screening for Improved Margin of Safety in Pharmacology Studies

In cardiotoxicity screening studies, female BALB/c mice were treated orally with 150 mg/kg of novel kinase inhibitors sharing a common target. Test articles were administered in a 1% carboxymethylcellulose (Thermo Fisher Scientific, Waltham, MA) vehicle at 10 mL/kg. Blood was drawn from anesthetized mice 4 hours after the second daily administration via retro-orbital sinus from 5 mice per compound, processed for serum, and assayed for cTnI concentration. Comparisons were made across compounds using both IVTI and cTnI data.

A subset of compounds were advanced into more extensive toxicology studies to test the ability of cTnI measurements in screening studies to predict cardiotoxicity. In these studies, female BALB/c mice were treated daily with a range of doses for 4 days. Twenty-four hours after the fourth dose, hearts were collected for histologic examination.

Measurement of cTnI and Fabp3 Concentrations in Mouse Hearts

Left ventricles were collected at necropsy from mice and homogenized as previously described (Pritt et al. 2008). Total protein concentration was measured by Bradford assay (Pierce cat# 23236, Thermo Fisher Scientific, Rockford, IL) of each tissue supernatant sample. Cardiac TnI and Fabp3 concentrations in supernatant were determined by ELISA (described above) and expressed per microgram of total protein.

Statistical Analysis

Quantitative results for cTnI concentration were analyzed using a three-factor analysis of variance. Factors in the model included treatment, strain, time, and their interactions. Treatment effect for each strain at each time point was evaluated using a t-test at the .05 significance level. The mean difference in cTnI concentration in strain (CD-1 vs. BALB/c mice) was also tested at the .05 significance level at each time point. Shapiro-Wilk’s test (Shapiro and Wilk 1965) for normality was performed at the .01 significance level to detect outlying observations, and Levene’s test (Levene 1960) for homogeneity of variance was performed at the .01 significance level to assist in interpretation of the effects.

Results

Qualification of cTnI Assay

The lower limit of quantitation (LLQ) of the ELISA used for detection of purified mouse cTnI in mouse serum was 0.14 ng/mL; the upper limit of quantitation (ULQ) was ≥10 ng/mL. Spike recovery samples were measured in triplicate for 4 to 5 runs at 10, 5, 2.5, 1.25, 0.63, 0.31, 0.16, and 0.14 ng/mL; mean bias was 0.02, 0.27, –0.56, 0.35, 0.93, –0.23, –1.68, and 6.85%; interassay variability assessed as coefficient of variation (%CV) was 4.3, 5.5, 4.0, 4.1, 6.8, 7.6, 9.0, and 10.5, respectively.

Baseline cTnI Concentration

Cardiac TnI concentration in serum was less than the LLQ in 94.5% of vehicle-treated mice (Table 1), and no difference in baseline cTnI concentration based on blood collection route or vehicle was detectable. The few examples of quantifiable cTnI concentration in vehicle-treated mice may have been due to spontaneous myopathies or to the stress of handling. With such low incidence of detectable serological cTnI concentrations in vehicle-treated mice, most serum concentrations of cTnI above the LLQ in treatment groups were likely to be treatment-related.

Cardiac TnI Concentration in Isoproterenol Studies

In CD-1 mice, mean cTnI concentration in serum increased in a dose-dependant manner from 0.1 mg/kg to 10 mg/kg ISO during the first 24 hours after treatment (Figure 1A). The maximum mean cTnI concentration after treatment with 50 mg/kg (5.9 ng/mL) was not significantly greater than the maximum after treatment with 10 mg/kg. Thirty-five of 39 CD-1 mice treated with ISO at 1 or 10 mg/kg had cTnI concentrations in excess of LLQ in the first 24 hours after dosing. The lowest measurable increase was 0.18 ng/mL (8 hours postdose), and 21 were in excess of 1 ng/mL. After administration of ISO at 1 and 10 mg/kg, maximum cTnI concentration in serum occurred 1 to 4 hours after dosing, and in both cases, the lowest concentration (n = 5) was 1.1 ng/mL. In CD-1 mice treated with ISO at 0.1 or 50 mg/kg, cTnI concentration returned to baseline by 24 hours. Cardiac TnI concentration in CD-1 mice treated with ISO at 1 or 10 mg/kg returned to baseline by 48 hours postdose. Maximum concentration of cTnI occurred 1 hour postdose for CD-1 mice administered ISO at 0.1, 1, or 50 mg/kg and 1 to 4 hours postdose at 10 mg/kg.

In BALB/c mice, mean cTnI concentration in serum increased in a dose-dependant manner after treatment with ISO at 0.1 to 10 mg/kg ISO during the first 24 hours after treatment (Figure 1B). Within 24 hours of dosing, cTnI concentration was greater than LLQ in 42 of 43 animals treated with ISO at 1 or 10 mg/kg. The lowest concentration was 0.36 ng/ml (24 hours postdose), and 37 were greater than 1 ng/ml. In BALB/c mice treated with ISO at 0.001 and 0.01 mg/kg, cTnI was below LLQ at both times tested (4 and 24 hr). Maximum concentration of cTnI occurred 4 hours postdose in BALB/c mice administered ISO at 0.1 and 10 mg/kg and 4 to 24 hours postdose at 1 mg/kg; the lowest concentration after 1 mg/kg was 0.86 ng/ml and occurred 4 hours postdose. Cardiac TnI concentration in serum from BALB/c mice treated with ISO at 1 or 10 mg/kg returned to baseline by 72 hours postdose. BALB/c mice were not treated with ISO at 50 mg/kg.

Histological Observations in Hearts

Isoproterenol treatment-related changes in hearts included dose-, time-, and strain-dependent incidence and severity of myocardial necrosis and inflammation. Aside from one section at 48 hours (n = 5) with minimal inflammation, no important changes occurred in CD-1 mice at 0.1 mg/kg (not tabulated; hearts were not collected from BALB/c mice given 0.1 mg/kg). Changes were poorly dose-responsive at 1 and 10 mg/kg in both strains (Figure 2). Evidence for active myocardial necrosis and inflammation peaked at 48 and 72 hours, respectively, and were generally less prominent in CD-1 than BALB/c mice (Figure 3A–B). Lesion severity ranged from minimal (grade 1), represented by loss of striations (coagulative necrosis) in a few myofibers, to moderate (grade 3), which affected multiple clusters of myofibers (Figure 4; additional details of grading scheme in Methods section). Necrosis and inflammation occurred most frequently near the lumen of the left ventricle, especially near the apex, but were found in any portion of the left ventricle and rarely in other portions of the heart. Vacuolation was a minor feature and likely represented a combination of a degenerative change and evidence for edema as a component of inflammation. Inflammation was characterized by increased numbers of mononuclear leukocytes, many of which were macrophages. Increased cellularity was likely enhanced by early regenerative activities of cardiac myocytes.

Mean Fabp3 Concentration in Isoproterenol Studies

Mean Fabp3 concentration in serum was measured in mice treated with vehicle (saline) and ISO at 10 mg/kg. Fabp3 concentration was below the lower limit of quantitation (1 ng/mL) in 13 out of 15 vehicle-treated CD-1 mice. Fabp3 concentration was increased in CD-1 mice 4 hours after treatment with ISO (p = .05) and returned to control values by 8 hours postdose (n = 5; Figure 5C).

In contrast, Fabp3 concentration was above LLQ in 5 out of 6 vehicle-treated BALB/c mice. Mean Fabp3 concentration was increased in BALB/c mice after treatment with ISO; maximum concentration occurred 1 hour after dosing and was statistically significant compared to control at 1 and 4 hours after dosing (n = 5, p < .003; Figure 5C). Eight hours after dosing, mean Fabp3 concentration in serum had decreased but had not returned to baseline.

AST and CK Activities in Isoproterenol Studies

AST and total CK activities, classical markers of cardiac injury, were measured in the serum of mice treated with vehicle or ISO at 10 mg/kg. Vehicle-treated control mice from all CD-1 (n = 28) and BALB/c (n = 24–26) studies were used to establish baselines for AST and CK. Baseline AST and CK activities were comparable between strains. AST activities were minimally increased at 1 and 8 hours after treatment with ISO at 10 mg/kg (n = 4–5, p ≤ .003) and were not different from baseline values at 24 hours postdose (Figure 5E) in CD1 mice. Analyses of AST and CK activities in BALB/c mice were compromised by hemolysis, which is known to cause positive interference. No significant alterations in the activity of AST were measured at any time in BALB/c mice (n = 3–8). Total CK activity in ISO-treated CD-1 mice (n = 4–5) was increased 1 hour postdose (p ≤.001) and was not different from baseline values at 8 or 24 hours after dosing (Figure 5G). Activity of CK in BALB/c mouse serum was markedly increased at 1 hour postdose, suggestive of myocyte injury. At 8 and 24 hours post-dose, individual mice had marked increases in the activity of CK, but group mean was not different from baseline (n = 1–5).

Strain Comparison for Biomarkers

Consistent with histological observations (Figure 3A–B), concentrations of cTnI in serum were significantly greater in BALB/c than CD-1 mice at 4, 24, and 48 hours and remained increased above LLQ longer in BALB/c (48 hours) than CD-1 mice (24 hours) after treatment with ISO at 1 and 10 mg/kg (Figure 3C–D). Increases in Fabp3 concentration compared to control groups were both greater and more sustained in BALB/c mice than CD-1 mice (Figure 5D) but were not as great or sustained as increases in cTnI concentration in either strain (Figure 5B). There was no apparent strain difference in increases of AST or CK activities, and both were of lesser magnitude than changes in cTnI (Figure 5F, H).

Cardiac TnI and Fabp3 Concentrations in Left Ventricles of BALB/c and CD-1 Mice

Concentrations of cTnI and Fabp3 in supernatants from homogenized left ventricles of BALB/c mice (2.5 ± 0.3 and 13.4 ± 0.2 ng/μg total protein, n = 5) were not different from those from CD-1 mice (2.6 ± 0.5 and 14.6 ± 1.6 ng/μg total protein, n = 5). Consequently, an inherent difference in tissue concentration cannot account for either strain-based differences in the response of these biomarkers in serum from the two strains or for the differences in baseline serological Fabp3 concentration.

Discrimination of On- versus Off-Target Toxicity in Pharmacology Studies Using BALB/c Mice

To discriminate between on-target and off-target mechanisms of cardiotoxicity, we explored the quantitative relationship between IVTI and serum cTnI in BALB/c mice treated with kinase inhibitors sharing a common target. Due to limited blood volume in BALB/c mice, independent studies were conducted for IVTI and toxicological response (serum cTnI). To bridge between these studies, we measured plasma compound concentrations in both.

In the event cardiotoxicity were target-mediated, little variability should occur across compounds in plots of percentage IVTI versus serum cTnI. If the cardiotoxicity had a significant off-target component, significant variability would occur across compounds in plots of percentage IVTI versus serum cTnI. A clear difference was observed across compounds (Figure 6), suggesting a significant off-target component to the acute cardiotoxicity caused by these kinase inhibitors and justifying further exploration of structure-activity relationships to retain pharmacological activity while minimizing compound-related cardiotoxicity.

Screening for Improved Margin of Safety in Pharmacology Studies

Mice treated with kinase inhibitors in screening studies displayed a range of cTnI concentrations in serum that were not consistently related to target inhibition. Compounds were binned into high risk, moderate risk, and low risk by considering both the IVTI and the serum cTnI data (Table 2). Increased cTnI concentrations in sera from screening studies were predictive of cardiac necrosis after extended dosing (Table 3). Compounds that caused cTnI concentrations in excess of 1 ng/mL in serum from one or more mice after 2 days caused cardiac lesions observable through histopathology after 4 days of treatment.

Discussion

Biomarkers of target organ toxicity can be used as stand-alone safety endpoints in discovery biology studies during early drug development. In these studies, cTnI concentration in serum was qualified in BALB/c and CD-1 mice and then implemented as a biomarker of acute cardiac necrosis in parallel with in vivo target inhibition studies in BALB/c mice. Benefits of such a strategy included discrimination between on- and off-target mechanisms of toxicity and prioritization of the development of compounds with improved safety profiles. This strategy proved to be an efficient and quantitative method of detecting cardiotoxicity after treatment with novel kinase inhibitors in a drug discovery program.

Cardiac TnI concentration in serum was a sensitive and quantitative marker of acute cardiac injury induced by treatment with ISO in both BALB/c and CD-1 mice. Increased cTnI concentration in serum was progressive with dose of ISO and occurred concurrently with cardiac necrosis, although morphological changes detectable by histopathologic examination plateaued between 1 and 10 mg/kg. A threshold concentration of cTnI in serum was identified, above which an animal was considered to have experienced adverse cardiac injury. This was considered to be the lowest cTnI concentration in any sample at the time point during which maximum group mean concentration occurred. In both strains, this occurred from 1 to 4 hours postdose and was 0.7 ng/mL (CD-1) and 0.9 ng/mL (BALB/c).

In practice, we found that cTnI concentrations in serum from BALB/c mice after treatment with a series of novel kinase inhibitors could be binned into low-risk (BQL-0.29 ng/mL), medium-risk (0.3–1.0 ng/mL), and high-risk (>1.0 ng/mL) categories. Compounds used in 2-day screening studies from which all measured cTnI concentrations fell into the low-risk category were not observed to cause cardiac necrosis in 4-day studies. Compounds that produced cTnI concentrations in the medium-risk category demonstrated a dose margin in which compound administration did not cause cardiac necrosis. All compounds that caused cTnI concentrations in the high-risk category caused cardiac necrosis in 4-day studies. Given the lack of uniformity in the assays available to measure cTnI, the threshold concentrations are likely to vary depending on which assay is used.

Isoproterenol-induced increases in cTnI concentrations were both greater and more sustained in BALB/c than in CD-1 mice, consistent with increased incidence and severity of morphological changes (greater number of muscle fibers involved or greater total volume of injury). The increased duration was likely attributable to higher initial concentrations of cTnI in serum, resulting from more extensive injury in BALB/c mice. Differences in cTnI concentration in serum at the same dose across strains could not be explained by different concentrations of cytoplasmic cTnI in the left ventricles. Greater cTnI concentration in serum from BALB/c mice is likely attributable to a greater number of injured cardiac muscle fibers.

The cause(s) for the strain differences in response to cardiac injury induced by exposure to ISO was not determined. However, possible causes include desensitization of β-adrenergic receptors (Perrino et al. 2005), strain-dependent differences in receptor density, cardiac capacity for aerobic metabolism, or adaptive responses (Faulx et al. 2005, 2007). Desensitization may also account for the lack of dose-response in CD-1 mice treated with greater than 10 mg/kg ISO.

Cardiac TnI concentration in serum was increased in CD-1 mice treated with 0.1 mg/kg ISO. These increases were below the threshold described above and occurred without histologic evidence of myocardial necrosis. Although small areas of injury could have been missed in histological examination of a single tissue section, it seems more likely that these relatively small increases in cTnI were indicative of myocardial membrane leakage without cell death. Cells releasing cTnI likely experienced mechanical stress related to increased contractility, or a transient period of ischemia, not lasting the 10 to 20 minutes required to cause cell death (Fromm and Roberts 2001). Conceivably, this resulted in temporary membrane disruption and leakage of intracellular contents from the cytoplasm. Similar scenarios have been observed in baboons and dogs in which short episodes (15 minutes) of myocardial ischemia were induced, leading to significant increases in plasma activity of total CK as well as the more cardiac specific CK-MB, with no histological evidence of cardiac necrosis (Heyndrickx et al. 1985; Vatner, Heyndrickx, and Fallon 1986).

Approximately 2% to 4% of total cTnI exists free in the cytosol of ventricular myocytes (Mair et al. 1996; Wu and Feng 1998). The rapid, forty- to eighty-fold increases in cTnI concentration in serum in the first 1 to 8 hours after ISO treatment, followed by a gradual decline in concentration over the next 48 hours, suggested that cytosolic cTnI was released immediately upon injury. Release of sarcomere-bound cTnI may have been inhibited by the gradual process of myocyte cell death and may not have reached the circulation due to phagocytosis by infiltrating inflammatory cells. Clinical evidence of this has been observed in patients undergoing coronary artery bypass grafts in which cTnI concentration in serum increased in the first 20 minutes following reperfusion of the heart, while β-type myosin heavy chain remained within reference intervals. In excised atrial appendages from the same patients, 8% of total cTnI content was found free in the cytosol, compared to less than 0.1% of β-MHC, leading to the conclusion that the immediate increase in circulating cTnI was from the free cytosolic pool (Bleier et al. 1998).

We also compared cTnI changes with other markers of cardiac injury. Increased Fabp3 concentration in serum suggested cell membrane disruption, with considerable magnitude of change in BALB/c mice. However, the duration of increased Fabp3 concentration was relatively limited in comparison to cTnI, and it is also found in skeletal muscle (Pritt et al. 2008), confounding specific detection of myocardial injury. The differences in Fabp3 concentrations between strains, in both control and treated mice, were remarkable and unexpected. They could not be explained by differences in Fabp3 concentrations in the ventricles. The small magnitude of increases in Fabp3 in CD-1 mice were insufficient to determine whether a difference in clearance rate of Fabp3 from blood exists between mouse strains but in any case cannot explain the large differences in serum concentrations.

Increased AST and CK activities also suggested myocardial injury, but the magnitudes of change in these markers were relatively small and the time for their detection short compared to alterations in the concentrations of cTnI. These limitations, coupled with a lack of specificity for myocardium, limit their use as stand-alone cardiac biomarkers.

Recent evidence supports the use of cardiac troponin concentrations in serum, in conjunction with examination of histological sections of heart (Santucci et al. 2007) and echocardiography (Kogan et al. 2007; Adamcova et al. 2007) to explore the relationships of compound structure with pharmacology and cardiotoxicity in mice. Our experience suggests that cTnI can be used to reliably and efficiently identify xenobiotic-induced myocardial membrane disruption and necrosis without the need for more resource-intensive histological and imaging-based examinations in short screening studies.

Screening for increased cTnI concentration in serum benefited an early-stage kinase inhibitor drug discovery program. Its use as a stand-alone cardiac safety endpoint gave insight into each compound’s safety while assessing its pharmacology using the same biological model, thus shortening cycle time and reducing associated costs. Cardiac troponin I concentration provided clear evidence of cardiac toxicity as pharmacology was being assessed, allowing prioritization of compounds according to potency, bioavailability, and safety. Using cTnI as a stand-alone biomarker in mice allowed for higher throughput of compounds with less demand on resources than required to synthesize compounds for toxicology studies in larger species, such as rat. Without such screening, the lack of a relationship between target modulation and cardiotoxicity would not have been as quickly evident and compound selection for more resource intensive studies not as well informed. Subsequent studies revealed potent compounds with less cardiac liability that were advanced into more extensive toxicology studies with a greater chance of a favorable safety profile.