Advertisement

The development and application of a high-sensitivity immunoassay for cardiac myosin–binding protein C

Open AccessPublished:December 08, 2015DOI:https://doi.org/10.1016/j.trsl.2015.11.008
      Cardiac troponins (cTns) are released and cleared slowly after myocardial injury. Cardiac myosin–binding protein C (cMyC) is a similar cardiac-restricted protein that has more rapid release and clearance kinetics. Direct comparisons are hampered by the lack of an assay for cMyC that matches the sensitivity of the contemporary assays for cTnI and cTnT. Using a novel pair of monoclonal antibodies, we generated a sensitive assay for MyC on the Erenna platform (Singulex) and compared serum concentrations with those of cTnI (Abbott) and cTnT (Roche) in stable ambulatory cardiac patients without evidence of acute cardiac injury or significant coronary artery stenoses. The assay for cMyC had a lower limit of detection of 0.4 ng/L, a lower limit of quantification (LLoQ) of 1.2 ng/L (LLoQ at 20% coefficient of variation [CV]) and reasonable recovery (107.1 ± 3.7%; mean ± standard deviation), dilutional linearity (101.0 ± 7.7%), and intraseries precision (CV, 11 ± 3%) and interseries precision (CV, 13 ± 3%). In 360 stable patients, cMyC was quantifiable in 359 patients and compared with cTnT and cTnI measured using contemporary high-sensitivity assays. cMyC concentration (median, 12.2 ng/L; interquartile range [IQR], 7.9–21.2 ng/L) was linearly correlated with those for cTnT (median, <3.0 ng/L; IQR, <3.0–4.9 ng/L; R = 0.56, P < 0.01) and cTnI (median, 2.10 ng/L; IQR, 1.3–4.2 ng/L; R = 0.77, P < 0.01) and showed similar dependencies on age, renal function, and left ventricular function. We have developed a high-sensitivity assay for cMyC. Concentrations of cMyC in clinically stable patients are highly correlated with those of cTnT and cTnI. This high correlation may enable ratiometric comparisons between biomarkers to distinguish clinical instability.

      Abbreviations:

      ACS (acute coronary syndrome), AMI (acute myocardial infarction), cMyC (cardiac myosin–binding protein C), cTn (cardiac troponin), CV (coefficient of variation), DE (detected event), LoB (limit of blank), LoD (lower limit of detection), LLoQ (lower limit of quantification), MP (magnetic microparticle), NSTE-ACS (non–ST-elevation acute coronary syndrome)
      At a Glance Commentary

      Background

      Cardiac myosin–binding protein C (cMyC) is a protein with cardiac-restricted expression that we have previously shown appears in the systemic circulation after acute myocardial injury using a relatively insensitive assay. This article describes a high-sensitivity assay for cMyC, which demonstrates that it can be measured at baseline in almost all individuals, and in a stable population its concentration correlates with those for cTnI and cTnT.

      Translational Significance

      This article acts as the foundation for a study using the assay described here in patients presenting with suspected acute myocardial infarction to compare the diagnostic and prognostic performances of cMyC with cTnT and cTnI.

      Introduction

      Acute myocardial infarction (AMI) carries a poor prognosis that can be improved by timely intervention. It must therefore be rapidly identified and differentiated from other causes of chest pain.
      • Cooper A.
      • Timmis A.
      • Skinner J.
      Assessment of recent onset chest pain or discomfort of suspected cardiac origin: summary of NICE guidance.
      Cardiac necrosis biomarkers have become crucial in affirming or excluding AMI in suspected non–ST-elevation acute coronary syndromes (NSTE-ACSs) and are needed to confirm the diagnosis in an appropriate clinical context.
      • Thygesen K.
      • Alpert J.S.
      • Jaffe A.S.
      • et al.
      Third universal definition of myocardial infarction.
      Cardiac troponins (cTns) have emerged as the gold standard and are incorporated in the universal definition of AMI.
      • Thygesen K.
      • Alpert J.S.
      • Jaffe A.S.
      • et al.
      Third universal definition of myocardial infarction.
      However, the cTns have potential drawbacks and new necrosis biomarkers could prove invaluable.
      • Gerszten R.E.
      • Carr S.A.
      • Sabatine M.
      Integration of proteomic-based tools for improved biomarkers of myocardial injury.
      The concentration of cTn rises slowly after acute myocardial injury and does not peak until 16–18 hours after the onset of chest pain.
      • Morrow D.A.
      Clinical application of sensitive troponin assays.
      To triage and treat NSTE-ACS early, it is therefore necessary to heed cTn concentrations close to the 99th percentile of a healthy population.
      • de Lemos J.A.
      • Morrow D.A.
      • de Filippi C.R.
      Highly sensitive troponin assays and the cardiology community: a love/hate relationship?.
      However, triage is confounded by the assays' decreased specificity for myocardial infarction when used in this way. In addition, diagnostic sensitivity may also be poor because up to 25% of patients with an eventual diagnosis of AMI are less than this threshold at presentation.
      • Hoeller R.
      • Rubini Gimenez M.
      • Reichlin T.
      • et al.
      Normal presenting levels of high-sensitivity troponin and myocardial infarction.
      Furthermore, although initial reports suggested that these assays allow more rapid diagnosis of AMI when the event is defined by a classic cTn assay,
      • Reichlin T.
      • Hochholzer W.
      • Bassetti S.
      • et al.
      Early diagnosis of myocardial infarction with sensitive cardiac troponin assays.
      • Keller T.
      • Zeller T.
      • Peetz D.
      • et al.
      Sensitive troponin I assay in early diagnosis of acute myocardial infarction.
      this advantage is probably lost when contemporary high-sensitivity assays are also used to define the index event.
      • Hammarsten O.
      • Fu M.L.
      • Sigurjonsdottir R.
      • et al.
      Troponin T percentiles from a random population sample, emergency room patients and patients with myocardial infarction.
      These drawbacks are acknowledged in the recently updated guidelines for the management of NSTE-ACSs that adopt cutoffs substantially less than the 99th percentile to “rule-out” AMI and substantially greater than the 99th percentile to “rule-in” AMI.
      • Roffi M.
      • Patrono C.
      • Collet J.P.
      • et al.
      2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC).
      This improves sensitivity and specificity at the expense of increasing the number of patients with indeterminate troponins requiring further observation and increased testing.
      The sarcomeric protein, cardiac myosin–binding protein C (C-protein, MYBPC3, cMyBP-C, or cMyC), is abundant
      • Aye T.T.
      • Scholten A.
      • Taouatas N.
      • et al.
      Proteome-wide protein concentrations in the human heart.
      and released rapidly into the coronary effluent.
      • Jacquet S.
      • Yin X.
      • Sicard P.
      • et al.
      Identification of cardiac myosin-binding protein C as a candidate biomarker of myocardial infarction by proteomics analysis.
      Recently, we demonstrated that cMyC accumulates more rapidly in the serum than cTnT; using timed iatrogenic injury in the setting of alcohol septal ablation for hypertrophic cardiomyopathy.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      Although after coronary artery bypass surgery, cMyC disappeared more rapidly than cTnT.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      However, comparisons were hindered by an insensitive assay for cMyC (lower limit of quantification [LLoQ], 80 ng/L), which consequently could only be quantified after injury had occurred. Without a sensitive assay for cMyC it is not possible to compare its diagnostic performance for AMI in suspected NSTE-ACS with those of cTnI and cTnT. The purpose of this study was to create and validate such a high-sensitivity assay.

      Materials and Methods

      Immunoassay for cMyC

      We have previously described the creation, biophysical selection, and organ specificity of mouse monoclonal antibodies recognizing cardiac-restricted epitopes within the N-terminus of cMyC.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      Two of these antibodies, 1A4 and 3H8, were used to create a sensitive sandwich immunoassay. Subsequently, we describe the optimized assay on the Erenna platform (Merck KGaA, Darmstadt, Germany).
      Magnetic microparticles (MPs; Singulex) for capture were prepared by binding 25 μg of mouse monoclonal (1A4) per milligram of MPs. The coated MPs were diluted in assay buffer (Singulex proprietary mix with custom 450 mM NaCl and 0.5% Triton X-100) to 100 μg/mL. Serum, plasma, or analyte (recombinant C0C2 domain of cMyC
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      ) was diluted 1:1 in an equal volume of standard diluent (Singulex) and 100 μL added per well of a 96-well assay plate. Samples or standards were then exposed to 100 μL of coated MPs and agitated for 2 hours at 25°C. MPs were retained via a magnetic bed with unbound material removed in a single wash step. Fluorescently labeled mouse monoclonal (3H8) detection antibody was diluted in assay buffer (Singulex) to 100 ng/mL. To each well, 20 μL of detection antibody was added and the MPs agitated for 1 hour at 25°C, retained via a magnetic bed, and then washed 4 times to remove any unbound detection reagent. The MPs were then transferred to a new plate and all buffer was aspirated. The MPs were then exposed to 20 μL/well of elution buffer B (Singulex) for 5 minutes at 25°C before transferring to a 384-well plate containing 10 μL/well of neutralization buffer D (Singulex). Fluorescent label was then detected by single molecule counting using the Erenna system (Singulex) with a dwell time of 60 s per well. Three signal outputs were obtained from the Erenna System: detected events (DEs; low end signal), event photons (low end and higher end signal), and total photons (high end signal).

      Assessing assay performance under serum-free conditions

      Having established a refined set of assay conditions, assay performance was assessed using a 12-point standard curve. Each point consisted of three 3-fold serially diluted cMyC concentrations to S4, followed by seven 2-fold serial dilutions to S11. All dilutions were in standard diluent (Singulex). The curve ranged from 0.58 to 2000 ng/L (S1–S11) with a 0 ng/L anchor of unadulterated standard diluent (Singulex). The lower limit of detection (LoD) was defined as 2.5 × standard deviation background divided by slope, and the LLoQ was defined as the lowest point on the standard curve, which has a coefficient of variation (CV) ≤20% where the back interpolated concentration had a recovery percent bias ≤20%.
      • Findlay J.W.
      • Smith W.C.
      • Lee J.W.
      • et al.
      Validation of immunoassays for bioanalysis: a pharmaceutical industry perspective.

      Assay verification in human serum and plasma

      Interassay and intra-assay series precision was evaluated in human serum samples that were tested unadulterated and spiked with 200 ng/L of cMyC. The samples were diluted 2-fold in standard diluent (1:1 mix) before assaying 6 replicates per sample on Day 1 and 3 replicates per sample on Day 2. Spike recovery was calculated by subtracting the dilution-corrected endogenous cMyC concentration from the dilution-corrected spiked value divided by the expected value. Dilutional linearity was evaluated by serial dilution of spiked human plasma. Linearity was calculated by dividing the dilution-corrected cMyC concentration by the preceding value, expressed as a percentage.

      MyC concentrations in human serum

      Between July 2009 and January 2014, 5329 patients were referred to the Kerckhoff Heart and Thorax Center for elective coronary angiography and provided written informed consent for their participation in blood-based biomarker studies as per institutional ethics board (FF 43/2010). The research was carried out according to the Code of Ethics of the World Medical Association (Declaration of Helsinki), informed consent was obtained, and the author's institutional review board has approved the study.
      From this population we selected 360 serum samples based on the absence of obstructive stenoses (<50%) on invasive coronary angiography and normal (<14 ng/L) high-sensitivity cTnT, renal function, and liver function. Further criteria used to choose samples were prior measurement of high-sensitivity cTnI and sufficient volume of stored serum to allow duplicate measurements of cMyC (>100 μL). cTnT was measured in serum with the high-sensitivity electrochemiluminescence immunoassay (Elecsys Analyzer 2010; Roche Diagnostics, Mannheim, Germany). For the cTnT assay, the limit of blank (LoB) = 3.0 ng/L, LoD = 5.0 ng/L, and LLoQ = 13.0 ng/L. The lowest concentration measurable with a CV <10% for this assay is 13.5 ng/L. The recommended clinical decision limit (99th percentile) for rule out of AMI using this assay is 14.0 ng/L. Concentrations of cTnT less than 3 ng/L (LoB) were not returned and therefore assigned a value of 1.5 ng/L in all analyses.
      cTnI was measured in serum with the high-sensitivity chemiluminescent immunoassay (ARCHITECT STAT High Sensitive Troponin; Abbott Laboratories, Abbott Park, Illinois). For the cTnI assay, LoD = 1.2 ng/L and LLoQ = 4.7 ng/L at a CV <10%. The 99th percentile is 15.6 ng/L in women and 34.2 ng/L in men. Concentrations less than the LoD were returned and used for comparisons, because all were greater than the locally determined LoB.

      Statistical analysis

      The methods used to calculate the LoDs and LLoQs for MyC are described previously.
      The Kolmogorov-Smirnov and Shapiro-Wilk tests were used to test if cMyC, cTnT, and cTnI concentrations were distributed normally. Spearman's rank test was used to assess correlation between the serum concentrations of each biomarker and to correlate the concentration of each marker to the continuous demographic variables of the sample population. Differences in the distribution of biomarker concentration across categories of dichotomous variables were examined using the independent-samples Mann-Whitney test. Stepwise multiple logistic regression analysis was used to assess independent association between the variables and biomarker concentrations. All analyses were carried out using SPSS v22. Normally distributed data are presented as the mean ± standard deviation. Statistical significance was set at P < 0.05.

      Results

      Analytic sensitivity of the cMyC assay

      The 12-point calibration using recombinant C0C2 domain of cMyC in standard diluent is shown in Supplementary Table I. The DE counts are shown for serial dilutions >5. The linear regression relationship for S5–S12 is DE = 32.7 × [cMyC] + 46.1 (R2 = 0.9995), where [MyC] is in nanograms per liter. The LLoQ is 1.2 ng/L and the calculated LoD is 0.4 ng/L.

      Interseries and intraseries precision of the MyC assay in serum

      Sera from 5 individuals were spiked with 200 ng/L of cMyC and subjected to repeated measurement. Six repeated measurements were made on Day 1 and 3 on Day 2. Supplementary Table II shows the CVs within and between assays. The average CV within assays was 11 ± 3% and between assays was 13 ± 3%.

      Analyte recovery from human serum and plasma

      Supplementary Table III shows analyte recovery from serum and plasma samples of different individuals each spiked with 200 ng/L of recombinant cMyC. The recovery in serum was 108.0 ± 6.2% (excluding lipemic sample, 115.4 ± 15.8% with this sample included) and in plasma 107.1 ± 3.7%.
      Dilutional linearity was tested using the finalized assay in plasma (see Supplementary Table IV). The results showed more than a 16-fold range of dilutions in plasma from 8 individuals, and linearity was 101 ± 7%.

      Comparison of cMyC, cTnT, and cTnI concentrations in stable patients

      The demographics of the population cohort used to compare biomarker concentrations are shown in Table I.
      Table IDemographics of the patient population (N = 359 unless otherwise specified)
      Demographicn (%)
      Male146 (40.6%)
      Current smoker131 (36.4%)
      BMI ≥ 30124 (34.4%) [n = 358]
      Diabetes39 (10.8%)
      Hyperlipidemia181 (50.3%)
      Family history106 (29.4%)
      β-Blocker182 (50.6%)
      Statin82 (22.8%)
      ACE-I/ARB196 (54.4%)
      Aspirin159 (44.2%)
      Digitalis8 (2.2%)
      Aldosterone antagonist16 (4.4%)
      Loop diuretic56 (15.6%)
      Thiazide diuretic82 (22.8%)
      COPD24 (6.7%)
      PVD10 (2.8%)
      Pulmonary HTN7 (1.9%)
      Angina180 (50.0%)
      AF/PPM32 (8.9%) [n = 357]
      Mean (standard deviation)
      Age (y)60.0 (12.0)
      BMI29.0 (5.6) [n = 358]
      GFR (mL/min/1.73 m2)100.3 (25.7) [n = 352]
      Creatinine (mg/dL)0.8 (0.2) [n = 351]
      LVEF (%)53.7 (12.3) [n = 189]
      Systolic BP (mm Hg)134.5 (20.9) [n = 357]
      [MyC] (ng/L)17.6 (16.4)
      [TnT] (ng/L)3.4 (3.2)
      [TnI] (ng/L)3.5 (4.3)
      Abbreviations: ACE-I, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BMI, body mass index; BP, blood pressure; COPD, chronic obstructive pulmonary disease; GFR, glomerular filtration rate; HTN, hypertension; LVEF, left ventricular ejection fraction; PPM, permanent pacemaker; PVD, peripheral vascular disease.
      Angina was defined as any symptom severity using Canadian Cardiovascular Society grades 1–4. Family history defined as a first degree relative with history of coronary artery disease and/or acute myocardial infarction and/or percutaneous coronary intervention and/or coronary artery bypass graft surgery.
      Three hundred sixty serum samples with [cTnT] <14 ng/L were analyzed. In one of these samples, cMyC was less than the LLoQ. Our subsequent analysis was of the 359 patients with an evaluable cMyC. Of these 274 patients had cTnT (<5 ng/L), and 52 patients cTnI (<1.2 ng/L), concentrations less than the LoD. The resulting truncation of the leftmost portion of the concentration-frequency distribution is therefore evident for cTnT and cTnI but not for cMyC (see Fig 1). None of the concentration-frequency histograms were normally distributed. The summary statistics describing their distribution are inset in the respective panels of Fig 1. In absolute terms cMyC is approximately 5 times more abundant than either cTnI or cTnT, as previously noted.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      Our cMyC assay therefore has sensitivity at least as good as the current commercial assays for cTnT and cTnI. The question is whether the concentrations of cMyC are related to those of the cTns?
      Figure thumbnail gr1
      Fig 1Distribution of cMyC, cTnI, and cTnT concentrations among 359 patients referred for elective coronary angiography with a cTnT <14 ng/L. To validate the cMyC assay described in Supplementary Table I, Supplementary Table II, Supplementary Table III, Supplementary Table IV we examined a stable patient cohort without acute myocardial injury. Also excluding acute myocardial injury by their mode of presentation, only patients with a [cTnT] less than the 99th percentile of a healthy “normal” population (14 ng/L) were included. All patients had a [cMyC] > LLoQ. Unfortunately, 274 patients (more than half the cohort) had a [cTnT] <5.0 ng/L, the LoD of the assay. Sera with a [cTnT] <3.0 ng/L (LoB) were assigned a value of 1.5 ng/L. Similarly, 52 patients had a [TnI] <1.2 ng/L, the LoD of the assay. For TnI the sera retained the value assigned by the assay because values were greater than the locally determined LoB. The differential sensitivities of the assay are the cause for artifactual distortion of low concentration portion of the histograms. Inset in each panel are the descriptors of the biomarker concentration distribution. None of the biomarkers are normally distributed. cMyC, cardiac myosin–binding protein C; cTn, cardiac troponin; IQR, interquartile range; LLoQ, lower limit of quantification; LoD, lower limit of detection.
      Fig 2 shows the relationships between the biomarkers. Serum concentrations of cMyC, cTnT, and cTnI are all positively correlated with one another with the strongest association between [MyC] and [TnI].
      Figure thumbnail gr2
      Fig 2Relationships between cMyC, cTnI, and cTnT. All 3 biomarkers significantly correlate with one another. The correlation coefficient (Spearman's Rho) is shown on the upper right quadrant. ∗∗P < 0.01. cMyC, cardiac myosin–binding protein C; cTn, cardiac troponin.
      Because the biomarkers are co-correlated we looked at the demographic variables known to influence [cTnT] and [cTnI] to determine if they similarly influence cMyC. The continuous variables are shown in Table II as correlation coefficients and the discontinuous variables in Table III as differences in mean biomarker concentration between those with and without the demographic feature. Generally, each of the biomarkers segregates similarly and concentrations are greater in patients with comorbidities.
      Table IICorrelation coefficients (Spearman's Rho) between serum concentration of cTnI, cTnT, and cMyC and continuous variables in the sample population
      Demographic variablecTnIcTnTcMyC
      Age0.336
      P < 0.01.
      0.448
      P < 0.01.
      0.385
      P < 0.01.
      GFR−0.224
      P < 0.01.
      −0.256
      P < 0.01.
      −0.288
      P < 0.01.
      Creatinine0.197
      P < 0.01.
      0.220
      P < 0.01.
      0.284
      P < 0.01.
      LVEF−0.208
      P < 0.01.
      −0.169
      P < 0.05.
      −0.218
      P < 0.01.
      Systolic BP0.116
      P < 0.05.
      0.176
      P < 0.01.
      0.134
      P < 0.05.
      BMI0.0690.0680.011
      Abbreviations: BMI, body mass index; BP, blood pressure; cMyC, cardiac myosin–binding protein C; cTn, cardiac troponin; GFR, glomerular filtration rate; LVEF, left ventricular ejection fraction.
      P < 0.01.
      P < 0.05.
      Table IIIMean biomarker concentration in the each category of dichotomous population variable
      Demographic variableMean concentration in each group (difference in mean concentrations)
      cTnI (ng/L)cTnT (ng/L)cMyC (ng/L)
      Nonmodifiable risk factors
       Family history vs no family history3.27 vs 3.66 (0.38
      P < 0.05.
      )
      3.37 vs 3.38 (0.01)15.0 vs 18.7 (3.71
      P < 0.05.
      )
       Female vs male3.22 vs 3.69 (0.47)3.27 vs 3.49 (0.22)16.3 vs 19.4 (3.15)
      Lifestyle
       Current smoker vs nonsmoker3.12 vs 3.79 (0.67)2.89 vs 3.66 (0.78
      P < 0.05.
      )
      16.0 vs 18.5 (2.52)
       BMI ≥ 30 vs BMI < 304.07 vs 3.25 (0.82)3.48 vs 3.30 (0.18)16.2 vs 18.2 (2.04)
      Comorbidities
       Pulmonary HTN vs no pulmonary HTN8.84 vs 3.44 (5.41
      P < 0.01.
      )
      8.61 vs 3.28 (5.33
      P < 0.01.
      )
      44.4 vs 17.0 (27.32
      P < 0.01.
      )
       AF/PPM vs sinus rhythm5.88 vs 3.27 (2.61
      P < 0.01.
      )
      4.92 vs 3.20 (1.71
      P < 0.01.
      )
      27.0 vs 16.5 (10.55
      P < 0.01.
      )
       Diabetes vs no diabetes3.59 vs 3.54 (0.05)4.21 vs 3.28 (0.93)19.7 vs 17.3 (2.38)
       Hyperlipidemia vs no hyperlipidemia3.19 vs 3.90 (0.71)3.07 vs 3.69 (0.62)16.6 vs 18.6 (1.94)
       COPD vs not COPD3.98 vs 3.51 (0.47)3.88 vs 3.34 (0.53)17.3 vs 17.6 (0.38)
       PVD vs no PVD4.19 vs 3.53 (0.67)3.37 vs 3.38 (0.01)20.9 vs 17.5 (3.44)
       Angina vs no angina3.14 vs 3.95 (0.80)3.07 vs 3.69 (0.62)16.0 vs 19.2 (3.17)
      Pharmacotherapy
       β-blocker vs no β-blocker3.71 vs 3.37 (0.34)3.52 vs 3.24 (0.28)20.8 vs 14.3 (6.51
      P < 0.01.
      )
       ACE-I/ARB vs no ACE-I/ARB4.26 vs 2.68 (1.58
      P < 0.01.
      )
      3.98 vs 2.66 (1.32
      P < 0.01.
      )
      20.2 vs 14.5 (5.75
      P < 0.01.
      )
       Aspirin vs no aspirin3.13 vs 3.87 (0.75
      P < 0.01.
      )
      2.98 vs 3.70 (0.72
      P < 0.05.
      )
      15.9 vs 19.0 (3.09
      P < 0.05.
      )
       Digitalis vs no digitalis5.16 vs 3.51 (1.66)5.60 vs 3.33 (2.27
      P < 0.05.
      )
      28.1 vs 17.4 (10.77
      P < 0.01.
      )
       Aldosterone antagonist vs no aldosterone antagonist5.37 vs 3.46 (1.91
      P < 0.01.
      )
      5.86 vs 3.26 (2.59
      P < 0.01.
      )
      30.5 vs 17.0 (13.52
      P < 0.01.
      )
       Loop diuretic vs no loop diuretic5.61 vs 3.16 (2.45
      P < 0.01.
      )
      4.93 vs 3.09 (1.83
      P < 0.01.
      )
      26.9 vs 15.9 (10.98
      P < 0.01.
      )
       Thiazide diuretics vs no thiazide diuretic4.56 vs 3.24 (1.32
      P < 0.01.
      )
      4.39 vs 3.08 (1.30
      P < 0.01.
      )
      23.1 vs 16.0 (7.16
      P < 0.01.
      )
       Statin vs no statin2.82 vs 3.76 (0.94
      P < 0.05.
      )
      2.70 vs 3.58 (0.88)15.0 vs 18.3 (3.32)
      Abbreviations: ACE-I, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BMI, body mass index; cMyC, cardiac myosin–binding protein C; COPD, chronic obstructive pulmonary disease; cTn, cardiac troponin; HTN, hypertension; PPM, permanent pacemaker; PVD, peripheral vascular disease.
      In brackets is the difference in mean biomarker concentrations between the 2 categories.
      P < 0.05.
      P < 0.01.
      A stepwise multiple logistic regression analysis was performed to determine which independent variables statistically significantly predicted the serum biomarker concentrations, independently of other covariates (Table IV). In this analysis, age, gender, creatinine, pulmonary hypertension, and use of statins, loop diuretics, and β-blockers all statistically predicted cMyC (P < 0.05), R2 = 0.198, n = 346. Because left ventricular ejection fraction (LVEF) was only known for 189 samples, LVEF was omitted from the analysis to preserve sample size and statistical power. With LVEF included in the analysis, only creatinine, LVEF, and age significantly predicted cMyC (P < 0.01), R2 = 0.22, n = 183 (see Supplementary Table I, Supplementary Table II, Supplementary Table III, Supplementary Table IV, Supplementary Table V, Supplementary Table VI). The same model was applied to cTnI and cTnT, excluding LVEF as a variable. cTnI was significantly predicted by age, gender, use of angiotensin-converting enzyme inhibitor and, or angiotensin receptor blockers (ACE-I/ARBs), statins, and loop diuretics, and irregular or paced cardiac rhythm (P < 0.05), R2 = 0.153, n = 346. cTnT was significantly predicted by age, gender, family history of heart disease (see Table I for definition), pulmonary hypertension, angina, diabetes, and use of ACE-I/ARBs, aldosterone antagonists and statins (P < 0.05), R2 = 0.299, n = 346. A distinctive feature of MyC was its association with β-blocker use.
      Table IVUnstandardized coefficients (B) for those variables which independently predict biomarker concentration in stepwise linear multiple regression analysis (n = 346)
      Demographic variableUnstandardized coefficients (B)Significance
      MyC
       Creatinine10.6890.024
       Age0.2860.000
       Female−5.1280.004
       Loop diuretic6.8890.002
       Statin−5.5100.003
       B-blocker4.4360.009
       Pulmonary hypertension14.8130.024
      TnI
       Age0.0670.000
       Female−1.8790.000
       ACE-I/ARB1.1360.014
       Loop diuretic1.6390.010
       Statin−1.3770.008
       AF/PPM1.5720.047
      TnT
       Age0.1140.000
       Female−1.0120.001
       ACE-I/ARB0.8250.007
       Statin−1.4770.000
       Aldosterone antagonist2.0010.007
       Family history0.8200.013
       Pulmonary hypertension4.7710.000
       Angina−0.6030.040
       Diabetes0.9340.047
      Abbreviations: ACE-I, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; MyC, myosin-binding protein C; PPM, permanent pacemaker.

      Discussion

      We have developed a high-sensitivity assay to measure cMyC in serum or plasma. In 360 stable patients with a cTnT <14 ng/L, cMyC was quantifiable in 359 patients, a sensitivity much greater than that achieved with cTnT (4 patients with greater than LLoQ = 13.0 ng/L) or cTnI (78 patients with greater than LLoQ = 4.7 ng/L). cMyC is the first cardiac-specific marker of injury to be described since cTnT and cTnI. Generally, concentrations of cMyC were highly correlated with those of cTnI and cTnT and were influenced by the same demographic features including gender, age, renal function, left ventricular function, medication, and heart rhythm.
      The close correlation between cMyC and cTnT/cTnI is surprising because their locations within the sarcomere differ.
      • Sadayappan S.
      • de Tombe P.P.
      Cardiac myosin binding protein-C as a central target of cardiac sarcomere signaling: a special mini review series.
      cTnT and cTnI are adjacent proteins on the thin filament (actin), whereas cMyC, as its name suggests, is predominantly bound to the thick filament (myosin). The precise reason for the appearance of cardiac sarcomeric proteins in the peripheral blood of healthy individuals is not known. However, because none of these proteins are actively exported, and an intact sarcolemma is impervious to proteins >40 kDa,
      • Camilleri J.P.
      • Joseph D.
      • Amat D.
      • et al.
      Impaired sarcolemmal membrane permeability in reperfused ischemic myocardium. Ultrastructural tracer study.
      their appearance in the circulation most likely represents “stable” slow attrition and dissolution of cardiac myocytes. In such a scenario, the release rates of all cardiac-specific proteins are likely to co-correlate because they document the same fundamental process. Furthermore, the rapidity of this process is likely to be influenced by traditional cardiac risk factors explaining the correlation with gender and age, whereas the progression of this process will be documented by other measures of cardiac injury explaining the correlation with left ventricular function, pulmonary artery hypertension, and medication. The correlation with renal function is likely to have more complex explanations including the renal excretion of immunoreactive N-terminal fragments of cMyC, cTnI, and cTnT; the accumulation of waste products that increase the rate of myocyte attrition; or common factors that cause cellular injury to both the heart and the kidney.
      Although, the factors affecting cTnI, cTnT, and cMyC are broadly very similar (see Table II, Table III), serum cMyC concentration is particularly affected by β-blocker prescription with an average 6.5 ng/L higher concentration in those taking medications of this class (a relationship that continues to be significant after multiple regression analysis). A possible explanation for this exceptional dichotomy between the biomarkers may relate to protein kinase A–dependent phosphorylation of critical serine residues within the M domain of cMyC.
      • Sadayappan S.
      • de Tombe P.P.
      Cardiac myosin binding protein-C: redefining its structure and function.
      When phosphorylated, these residues more effectively guard a calpain cleavage site within cMyC.
      • Sadayappan S.
      • de Tombe P.P.
      Cardiac myosin binding protein-C as a central target of cardiac sarcomere signaling: a special mini review series.
      • Sadayappan S.
      • de Tombe P.P.
      Cardiac myosin binding protein-C: redefining its structure and function.
      Cleavage at this site releases a 40 kDa N-terminal fragment, the dominant fragment we observed in serum of patients with AMI.
      • Jacquet S.
      • Yin X.
      • Sicard P.
      • et al.
      Identification of cardiac myosin-binding protein C as a candidate biomarker of myocardial infarction by proteomics analysis.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      • Baker J.O.
      • Devaraj R.
      • Reinhold J.
      • et al.
      Cardiac myosin-binding protein C as a potential new serum biomarker of myocardial infarction.
      Interestingly this fragment may act as a “poison peptide” causing cardiac dysfunction.
      • Sadayappan S.
      • de Tombe P.P.
      Cardiac myosin binding protein-C as a central target of cardiac sarcomere signaling: a special mini review series.
      Thus, unlike cTnI and cTnT, cMyC may not just be a bystander biomarker of cardiac injury, but lie on the causal pathway leading to myocardial disease.
      Our ultimate aim is to determine if cMyC is a “better” diagnostic biomarker of acute myocardial injury than cTnT or cTnI. On the basis of our previous findings with a much less sensitive assay, after iatrogenic myocardial injury cMyC is released and cleared more rapidly from the peripheral circulation than cTnT.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
      The findings presented here are necessary stepping stones toward a large study of patients with suspected NSTE-ACS where the diagnostic utility of cMyC can be compared with cTnI and cTnT. Nonetheless, it is tempting to speculate how our present study will have impact on the diagnostic performance of cMyC in this clinical scenario. We had hoped cMyC concentrations would not be influenced by age, gender, renal function, and other cardiac risk factors. Our results clearly suggest that baseline cMyC concentrations will be higher in those at risk of an NSTE-ACS than in healthy controls. Thus, it is likely cMyC will have the same inadequacy as cTnI and cTnT in differentiating chronic increases in biomarker concentration from the minor increases associated with the start of an acute myocardial injury event. However, if our findings of faster MyC release in iatrogenic injury hold true with spontaneous myocardial injury, then the close correlation between MyC and cTnI/cTnT could become uncoupled as the biomarkers rise asymmetrically during acute injury. On the other hand, the abundance, ease of measurement, and correlation of cMyC with comorbidities may provide an advantage in the monitoring of chronic disease.
      One of the major limitations of the present study is that the study population was not healthy, and we therefore cannot estimate the 99th percentile concentration for cMyC. In choosing the population to validate our novel assay, we thought it more important to have a complete description of their demographics with the availability of other laboratory measures, including contemporary high-sensitivity cTnI and cTnT. This choice was consolidated by the lack of guidance on how rigorously to exclude covert cardiac disease in a healthy cohort and the influence this uncertainty has on the 99th percentile returned by that particular unique healthy cohort. Finally, we reasoned that the patient population we studied is more representative of those that will attend with a suspected NTSE-ACS event than a healthy younger cohort without cardiac risk factors. Nonetheless, the 99th percentiles returned in our population closely match those defined in healthy control populations for the assays we used for TnT (13 vs 14 ng/L, respectively) and TnI (19 vs 22 ng/L, respectively).
      A further limitation is that the Erenna platform on which the cMyC assay is performed is currently only available for research purposes and cannot provide the flexibility or turn-around times required for clinical use. These deficiencies could be addressed by migration to another platform or through the development of the Erenna platform.
      Another “high-sensitivity” assay for cMyC has been described recently.
      • Kuster D.W.
      • Barefield D.
      • Govindan S.
      • et al.
      A sensitive and specific quantitation method for determination of serum cardiac myosin binding protein-C by electrochemiluminescence immunoassay.
      However, this assay has a sensitivity of 2–3 orders of magnitude lower than ours and reports a mean difference in cMyC concentration between controls and patients with AMI of approximately 3-fold (∼1.5 μg/L increasing to ∼5 μg/L).
      • Kuster D.W.
      • Barefield D.
      • Govindan S.
      • et al.
      A sensitive and specific quantitation method for determination of serum cardiac myosin binding protein-C by electrochemiluminescence immunoassay.
      These concentrations are difficult to reconcile with those presented here or previously.
      • Baker J.O.
      • Tyther R.
      • Liebetrau C.
      • et al.
      Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.

      Conclusions

      We have developed and validated a sensitive assay for cMyC, which for the first time allows this cardiac-specific marker of myocardial injury to be quantified in ambulatory patients. The diagnostic performance of this assay is yet to be compared with cTnI and cTnT in the setting of NSTE-ACS.

      Acknowledgments

      Conflicts of Interest: All authors have read the journal's policy on disclosure of potential conflicts of interest. There are no relationships with industry pertaining to this work. Singulex was contracted to undertake the analyses of cMyC on a fee-for-service basis and holds no commercial interest. Marber is named as an inventor on a patent held by King's College London for the detection of cardiac myosin–binding protein C as a biomarker of myocardial injury.
      This work was supported by grants from the Medical Research Council (UK) ( G1000737 ), Guy's and St Thomas' Charity ( R060701 , R100404 ), British Heart Foundation ( TG/15/1/31518 ), and the UK Department of Health through the National Institute for Health Research Biomedical Research Centre award to Guy's & St Thomas' National Health Service Foundation Trust. High-sensitive cTnI assays were provided by Abbott Diagnostics Deutschland GmbH, Wiesbaden, Germany.
      The funding sources had no influence on the design, execution, data collection, analysis, drafting of this report, or decision to submit this manuscript.
      All authors have read the journal's authorship agreement and the manuscript has been reviewed and approved by all named authors. This manuscript has been drafted by the authors, no external editorial support was used.

      Supplementary Data

      Supplementary Table IAnalytic performance of the cMyC assay under serum-free conditions
      StandardExpected [cMyC], ng/LnMean DEObserved [cMyC], ng/L%CVRecovery
      S1200031998.553%100%
      S2666.673732.074%110%
      S3222.223227.8412%103%
      S474.07368.728%93%
      S537.043125436.863%100%
      S618.523665.3319.634%106%
      S79.263344.339.794%106%
      S84.6321894.723%102%
      S92.323106.331.855%80%
      S101.16289.51.2713%110%
      S110.583740.7520%129%
      S120351ND
      Abbreviations: cMyC, cardiac myosin–binding protein C; CV, coefficient of variation; DE, detected event; ND, not determined.
      Supplementary Table IIAnalytic precision of the cMyC assay using serum from 5 individuals with (+200) and without the addition of a 200 ng/L spike of recombinant cMyC
      IDDay 1IntraseriesDay 2Interseries
      123456MeanSD%CV123MeanSD%CV
      14.776.224.544.64.544.774.910.6513%4.976.355.090.7515%
      1 + 200249.2228.4241.6248.4286.8250.8721.749%223.9234.1214.7240.8822.19%
      220.6619.3914.7118.0315.8917.3617.672.1912%24.7920.8226.6919.813.9520%
      2 + 200322.9291.5304.6290.5361382325.4338.1512%274.1311.4246.9309.4441.814%
      371.5370.1370.666.6659.3462.4366.784.957%64.4455.3651.4563.557.0711%
      3 + 200252241.9276.5264.6271.4342.4274.7935.4713%252261.3229265.6832.312%
      413.5514.7613.6513.1311.2810.2312.771.6813%16.9717.0712.7613.712.317%
      4 + 200216.5199.2249.8273.8224.1236.6233.3226.2611%197.5203.4220.4224.5725.311%
      517.7417.9216.7615.3411.4916.1315.92.3715%17.6617.6213.4916.022.2314%
      5 + 200232.6255.3258.8228252.9254.9247.0713.195%187.8204.8216.2232.3625.411%
      Abbreviations: cMyC, cardiac myosin–binding protein C; CV, coefficient of variation; DE, detected event.
      All values are in nanograms per liter unless stated otherwise. The serum of individual number 2 was lipemic.
      Supplementary Table IIISpike recovery in human serum and plasma. Samples are sera or plasma from different individuals with (+200) and without the addition of a 200 ng/L spike of recombinant cMyC
      IDnObserved [cMyC], ng/LSD%CV% Spike recovery
      Serum
       185.090.7515
       1 + 2009240.8822.079118
       2919.813.9520
       2 + 2009309.4441.7914145
       3963.557.0711
       3 + 2009265.6832.2812101
       4913.712.3017
       4 + 2009224.5725.2711105
       5916.022.2314
       5 + 2009232.3625.4311108
      Plasma
       631.180.109
       6 + 2003206.143.182102
       736.230.132
       7 + 2003228.059.714111
       835.70.8415
       8 + 2003221.067.703108
       933.980.226
       9 + 2003212.1615.77104
       1031.280.2721
       10 + 2003227.7810.45113
       1131.20.2117
       11 + 2003212.0316.38105
       1233.730.205
       12 + 2003212.462.431104
       13310.290.414
       13 + 2003230.6811.205110
      Abbreviations: cMyC, cardiac myosin–binding protein C; CV, coefficient of variation; SD, standard deviation.
      All values are in nanograms per liter unless stated otherwise. The serum of individual number 2 was lipemic and excluded from analysis of average % spike recovery.
      Supplementary Table IVDilutional linearity of human plasma spiked with 200 ng/L of cMyC
      SamplenDilution factorMean [cMyBP-C], pg/mLSD%CVDilution corrected [cMyBP-C], ng/LDilutional linearity
      132206.143.182%412.28
      134105.427.167%421.68102%
      13849.052.094%392.4393%
      131626.831.726%429.25109%
      232228.059.714%456.09
      234123.614.544%494.43108%
      23854.844.578%438.7289%
      231628.340.151%453.43103%
      332221.067.53%442.12
      334114.142.552%456.56103%
      33854.713.096%437.6796%
      331629.551.485%472.87108%
      432212.1615.77%424.33
      434116.071.481%464.26109%
      43853.231.944%425.8192%
      431628.981.194%463.61109%
      532227.7810.45%455.56
      534108.276.326%433.0695%
      53851.021.673%408.294%
      531627.810.943%445.03109%
      632212.0216.38%424.03
      634102.623.974%410.4797%
      63848.111.393%384.8494%
      631627.040.823%432.69112%
      732212.462.431%424.91
      734101.822.192%407.2896%
      73846.981.313%375.8392%
      731626.41.335%422.35112%
      832230.6811.25%461.36
      834103.665.415%414.6590%
      83850.052.685%400.3797%
      8316281.475%447.96112%
      Abbreviations: cMyC, cardiac myosin–binding protein C; CV, coefficient of variation; SD, standard deviation.
      Supplementary Table VResults of stepwise multiple linear regression analysis for each biomarker, excluding LVEF as an independent variable (n = 346)
      Demographic variableUnstandardized coefficientsStandardized coefficientstSignificance95.0% Confidence interval for B
      BStandard errorBetaLower boundUpper bound
      MyC
       Creatinine10.6894.7140.1282.2680.0241.41819.961
       Age0.2860.0680.2184.1970.0000.1520.420
       Female−5.1281.774−0.163−2.8900.004−8.618−1.638
       Loop diuretic6.8892.2000.1603.1320.0022.56211.216
       Statin−5.5101.869−0.149−2.9480.003−9.187−1.834
       β-Blocker4.4361.6790.1422.6410.0091.1327.739
       Pulmonary HTN14.8136.5470.1132.2630.0241.93527.690
      TnI
       Age0.0670.0190.1863.5520.0000.0300.103
       Female−1.8790.444−0.218−4.2350.000−2.751−1.006
       ACE-I/ARB1.1360.4610.1322.4670.0140.2302.042
       Loop diuretic1.6390.6330.1392.5890.0100.3942.885
       Statin−1.3770.514−0.136−2.6800.008−2.388−0.367
       AF/PPM1.5720.7900.1051.9900.0470.0183.126
      TnT
       Age0.1140.0130.4318.8890.0000.0880.139
       Female−1.0120.300−0.160−3.3730.001−1.602−0.422
       Family history0.8200.3260.1192.5100.0130.1771.462
       Pulmonary HTN4.7711.2230.1803.9010.0002.3657.176
       Angina−0.6030.293−0.096−2.0590.040−1.179−0.027
       Diabetes0.9340.4680.0931.9950.0470.0131.855
       Statin−1.4770.348−0.198−4.2440.000−2.162−0.793
       ACE-I/ARB0.8250.3020.1302.7270.0070.2301.420
       Aldosterone antagonists2.0010.7320.1292.7350.0070.5623.441
      Abbreviations: ACE, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; HTN, hypertension; LVEF, left ventricular ejection fraction; MyC, myosin-binding protein C; PPM, permanent pacemaker.
      Supplementary Table VIResults of stepwise multiple linear regression analysis for each biomarker, including LVEF as an independent variable (n = 183)
      Demographic variableUnstandardized coefficientsStandardized coefficientstSignificance95.0% Confidence interval for B
      BStandard errorBetaLower boundUpper bound
      MyC
       Creatinine23.4526.0340.2613.8860.00011.54535.360
       LVEF−0.3660.094−0.263−3.9150.000−0.551−0.182
       Age0.3390.0890.2553.7910.0000.1630.516
      TnI
       Female−2.7720.687−0.280−4.0350.000−4.127−1.416
       LVEF−0.1060.028−0.257−3.7440.000−0.162−0.050
       Age0.0970.0270.2463.5280.0010.0430.151
      TnT
       Age0.1130.0170.4316.5330.0000.0790.147
       LVEF−0.0590.017−0.216−3.3980.001−0.094−0.025
       Pulmonary HTN5.2141.2620.2554.1310.0002.7237.705
       Statin−1.7970.531−0.212−3.3820.001−2.845−0.748
       Female−1.4010.417−0.213−3.3610.001−2.224−0.578
       Family history1.1300.4790.1492.3610.0190.1852.074
       ACE-I/ARB1.0360.4400.1502.3580.0200.1691.904
      Abbreviations: ACE, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; HTN, hypertension; LVEF, left ventricular ejection fraction; MyC, myosin-binding protein C.

      References

        • Cooper A.
        • Timmis A.
        • Skinner J.
        Assessment of recent onset chest pain or discomfort of suspected cardiac origin: summary of NICE guidance.
        BMJ. 2010; 340: c1118
        • Thygesen K.
        • Alpert J.S.
        • Jaffe A.S.
        • et al.
        Third universal definition of myocardial infarction.
        Circulation. 2012; 126: 2020-2035
        • Gerszten R.E.
        • Carr S.A.
        • Sabatine M.
        Integration of proteomic-based tools for improved biomarkers of myocardial injury.
        Clin Chem. 2010; 56: 194-201
        • Morrow D.A.
        Clinical application of sensitive troponin assays.
        N Engl J Med. 2009; 361: 913-915
        • de Lemos J.A.
        • Morrow D.A.
        • de Filippi C.R.
        Highly sensitive troponin assays and the cardiology community: a love/hate relationship?.
        Clin Chem. 2011; 57: 826-829
        • Hoeller R.
        • Rubini Gimenez M.
        • Reichlin T.
        • et al.
        Normal presenting levels of high-sensitivity troponin and myocardial infarction.
        Heart. 2013; 99: 1567-1572
        • Reichlin T.
        • Hochholzer W.
        • Bassetti S.
        • et al.
        Early diagnosis of myocardial infarction with sensitive cardiac troponin assays.
        N Engl J Med. 2009; 361: 858-867
        • Keller T.
        • Zeller T.
        • Peetz D.
        • et al.
        Sensitive troponin I assay in early diagnosis of acute myocardial infarction.
        N Engl J Med. 2009; 361: 868-877
        • Hammarsten O.
        • Fu M.L.
        • Sigurjonsdottir R.
        • et al.
        Troponin T percentiles from a random population sample, emergency room patients and patients with myocardial infarction.
        Clin Chem. 2012; 58: 628-637
        • Roffi M.
        • Patrono C.
        • Collet J.P.
        • et al.
        2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC).
        Eur Heart J. 2015; (Available at: http://www.ncbi.nlm.nih.gov/pubmed/26320110)
        • Aye T.T.
        • Scholten A.
        • Taouatas N.
        • et al.
        Proteome-wide protein concentrations in the human heart.
        Mol Biosyst. 2010; 6: 1917-1927
        • Jacquet S.
        • Yin X.
        • Sicard P.
        • et al.
        Identification of cardiac myosin-binding protein C as a candidate biomarker of myocardial infarction by proteomics analysis.
        Mol Cell Proteomics. 2009; 8: 2687-2699
        • Baker J.O.
        • Tyther R.
        • Liebetrau C.
        • et al.
        Cardiac myosin-binding protein C: a potential early biomarker of myocardial injury.
        Basic Res Cardiol. 2015; 110: 478
        • Findlay J.W.
        • Smith W.C.
        • Lee J.W.
        • et al.
        Validation of immunoassays for bioanalysis: a pharmaceutical industry perspective.
        J Pharm Biomed Anal. 2000; 21: 1249-1273
        • Sadayappan S.
        • de Tombe P.P.
        Cardiac myosin binding protein-C as a central target of cardiac sarcomere signaling: a special mini review series.
        Pflugers Arch. 2014; 466: 195-200
        • Camilleri J.P.
        • Joseph D.
        • Amat D.
        • et al.
        Impaired sarcolemmal membrane permeability in reperfused ischemic myocardium. Ultrastructural tracer study.
        Virchows Arch A Pathol Anat Histol. 1980; 388: 69-76
        • Sadayappan S.
        • de Tombe P.P.
        Cardiac myosin binding protein-C: redefining its structure and function.
        Biophys Rev. 2012; 4: 93-106
        • Baker J.O.
        • Devaraj R.
        • Reinhold J.
        • et al.
        Cardiac myosin-binding protein C as a potential new serum biomarker of myocardial infarction.
        Circulation. 2010; 122: A15438
        • Kuster D.W.
        • Barefield D.
        • Govindan S.
        • et al.
        A sensitive and specific quantitation method for determination of serum cardiac myosin binding protein-C by electrochemiluminescence immunoassay.
        J Vis Exp. 2013; (Available at: http://www.ncbi.nlm.nih.gov/pubmed/23963065)