Label-free detection of cardiac troponin I and complex using surface plasmon resonance for assessing acute myocardial injuries

Posted on

Introduction

Myocardial injury, a key component of cardiovascular disease, contributes significantly to global morbidity and mortality. It is identified by the presence of cardiac troponin (cTn), a protein composed of three subunits: cardiac troponin I (cTnI), T (cTnT), and C (TnC), which also serves as the established biomarker for acute myocardial infarction (AMI), commonly known as heart attack1. With the introduction of highly sensitive cTn assays, lower cTn levels are now more frequently detected1,2. However, such elevations are not exclusive to myocardial injury and infarction and can be observed in non-cardiac conditions, further complicating the diagnosis3,4.

Over the years, the urgency for rapid, reliable, and cost-effective diagnosis has stimulated the development of the so-called cTn point-of-care (PoC) devices5. The portability of these tests accelerated the diagnosis of heart attacks and streamlined the therapeutic process. The current generation of cTn PoC devices is highly sensitive, being able to detect the presence of cTn at very low concentrations, in the range of ng/L – μg/L, within minutes6. These PoCs have led to several key improvements, including rapid rule-out of heart attacks, reduced emergency room stays, and accelerated risk stratification5. However, their efficiency still falls short when compared to optimised laboratory-based protocols, hindering their widespread adoption5. This has resulted in a missed opportunity to fully leverage their benefits, unlike other PoC technologies, such as glucose and COVID-19 tests. Nevertheless, the development of PoC testing is crucial when access to laboratories is limited and should prioritise rapid turnaround, low cost, and high sensitivity and accuracy.

A key finding that may significantly impact the accuracy of cTn measurement is the continuous stepwise degradation of cTn complex (cTnTIC) following cardiac injury7. Current high-sensitivity cTn assays can detect all circulating forms of cTnI or cTnT, but they are not designed to specifically identify individual isoforms or targeted combinations, which may hold important diagnostic value8,9. Recent studies have shown that in the early hours of AMI, cTnTIC degrades from a high molecular weight complex to a low molecular weight complex, and within hours of onset, further transitions to a binary complex (cTnIC) and free cTnT7,9,10. The concentration and composition of circulating cTn forms can inform the site of injury, the ischaemic time window, and the severity of the injury10,11,12. In other conditions where cTn is also elevated, the concentration of these cTn isoforms differs broadly despite similar compositions. Patients with unstable angina predominantly show cTnIC, with low concentration of cTnTIC and free cTnI13 while patients with non-type 1 myocardial infarction (MI) display mainly low molecular weight cTnTIC and cTnIC, with little intact cTnTIC and no free, non-complexed form14. This led to a growing number of studies focusing on profiling troponin composition, which have been shown to enhance clinical specificity and sensitivity through differentiation across types of cardiac injury and infarction8,9,15. Recent evidence suggests that the ratio between free and complex cTnI may serve as a more reliable indicator of myocardial injury than the measurement of total cTnI isoforms. A study by Li et al demonstrated that the intact cTnTIC to cTnI ratio has potential for distinguishing acute from chronic myocardial injuries, with a high cTnTIC-to-cTnI ratio indicating acute injury16. However, in order to accurately detect the isoforms, the study was conducted using a combination of benchtop chemiluminescence immunoassays and commercial high-sensitivity cTn assays, an approach that may introduce variability in assay formats and complicate data interpretation. This limitation motivated us to investigate whether alternative detection methods could offer a more cost-effective approach, reduce the complexity associated with antibody labelling and enable faster turnaround times for assessing acute myocardial injury.

Photonic biosensors consist of micro and nanostructures that, when interacting with light, generate the so-called evanescent field, very sensitive to any changes in the refractive index (RI) that surrounds the biosensor structure, such as those provoked by the accumulation of the specific analytes at the sensor surface17. Therefore, they can directly detect the interactions of biomarkers with their specific bioreceptors by monitoring minute changes in the properties of the propagating light without the need for secondary labelling. This advantage can help identify multiple cTn isoforms in a given sample. Among them, Surface Plasmon Resonance (SPR) biosensors are the most prevalent photonic biosensors used in commercial settings, particularly for medical diagnosis18. They consist of a metal layer (generally gold) of about 50 nm thickness that relies on the generation of surface plasmons, special electromagnetic modes generated at the interface between the metal and a dielectric, whose propagation is very sensitive to changes in RI of the dielectric, such as those provoked by biointeraction. Numerous studies have investigated the use of biosensors to detect cardiac troponin I or T individually using innovative approaches to both enhance the effectiveness and stability of the detection and integration into point-of-care devices19,20. However, the potential of these platforms to assess the ratio between cTnI and cTn complex isoforms remains unexplored. This approach could enhance diagnostic sensitivity by capturing additional pathophysiological information.

Here, we present a label-free method using an SPR biosensor to measure the ratio of total cTnTIC to cTnI. This is achieved by the functionalisation of the sensor surface with a specific antibody that can capture and quantify both cTnTIC and cTnI isoforms. Then, we employed an unlabeled detection antibody specific to cTnTIC to differentiate from cTnI. This method allows quantitative analysis of both antibody interactions in real-time through changes in RI upon the biointeraction on the sensor surface. The signals generated by the detection antibodies to the cTn were then plotted into calibration curves. Solutions containing different ratios of these purified isomers were then used to validate the biosensor’s performance. Building on this approach, and in line with recent reports identifying the binary cTnIC isoform as a key degradation product of cTnTIC during myocardial injury9, we extended the method to test the ratio between cTnTIC, cTnIC, and cTnI. This allowed us to explore whether simultaneous monitoring of intact and degraded complexes could provide additional analytical insight into troponin composition. We evaluated the platform based on three critical parameters: selectivity for cTn isomers, signal reproducibility, and detection limit of the approach.

Results

SPR detection for the quantification of cTnTIC and cTnI isoforms

One of the key challenges with the design of conventional troponin assays is the reliance on labelled secondary antibodies, which not only limits the amount of information it offers but also the addition of multiple steps, increasing the overall cost. The label-free detection provided by an SPR biosensor offers an advantage to study specific cTnI forms, as it allows us to quantify all interactions that occur directly and in real-time on the biosensor surface. Specifically, for SPR, it is important to carefully select the capturing antibody immobilised at the sensor surface to capture cTnI and cTnTIC in one run. It is recommended that these antibodies target epitopes that are least susceptible to degradation21. Once the cTnI isomers are captured, a second antibody can be used to discern the isoform captured during the first step.

The cTn complex is released from damaged cardiac muscles into the blood and degrades into different isoforms over time (Fig. 1A). We employed two SPR channels immobilised with anti-cTnI 19c7 mAb for replicate measurements (Fig. 1B). Although one flow channel could, in principle, be used for cTnI and the other for cTnT, such a configuration would still capture complexes on both surfaces, complicating isoform resolution. We therefore adopted a sandwich assay format, consistent with previously reported methodologies12,16, to specifically differentiate cTnI and cTn complexes. This antibody is specific to the cTnI subunit and able to capture all troponin isoforms. It targets the 41–49 amino acid regions, which sit within the stable region of cTnI molecule10. A secondary detection is then carried out using anti-cTnIC 20c6 mAb, specific to a region found at the IC isomer, capable of identifying intact cTnTIC with ‘long-cTnT’ and low molecular weight cTnTIC where cTnT has degraded (see Table S1 for specificity and epitopes of the monoclonal antibodies used in this study). Using this approach, we can elucidate the concentrations and these cTn isoforms in a given sample in real-time by subtracting the concentration of cTnI isomers captured in the first interactions from the concentration of cTnI-specific isomers captured by the second interactions (Fig. 1C). To achieve this, calibration curves will be generated for each isomer at both the initial capture and secondary antibody detection stages (Fig. 1D). Finally, cTnI concentration can be obtained by subtracting the RU of cTnTIC from RU of cTnI. Figure 1 illustrates the step-by-step methodology employed in this quantification approach, highlighting how the concentration of both cTnI and complex isomers can be determined.

Fig. 1: Schematic illustration of the cTn isomers detection strategy.
figure 1

A cTn proteins release into the bloodstream following cardiac muscle damage and degrades over time. B Two-step label-free sensing approach using SPR where cTnI isomers first bind to the capture antibodies, followed by identification with detection antibodies. C Sensorgram of the two-step approach. D Flow chart quantifying cTn isomers in a solution containing cTnI and cTnTIC isomers and the assessment of cTnTIC:cTnI ratio based on their concentration.

Full size image

Characterisation of the troponin-specific sensor surface

For the covalent attachment of 19c7 mAb onto the sensing surface, we employed the generation of a well-packed self-assembled monolayer (SAM) of 11-MUA, which is commonly used in biosensing applications22. These alkanethiols have thiol groups at one end, allowing their anchoring to the gold surface via thiol-gold chemistry, and carboxylic groups at the other end, allowing their covalent coupling with the antibodies through their native amine residues23. The formation of the SAM was assessed using three conditions: gold only, 11-MUA only, and 11-MUA linked with antibodies. The results of the surface wettability (refer to Table S2) and atomic composition of these samples were compared. Bare gold surfaces exhibited a value of 61 °. With the formation of 11-MUA SAM, the WCA decreased to 37.2 °, similar to values reported elsewhere22, indicating that the surface has become more hydrophilic by the presence of these molecules. Carboxylic-terminated SAMs have been shown to have less hydrophobic surface after functionalisation24. After the primary antibody immobilisation on the SAM, the value further decreased to 28.3 °, indicating a continuous hydrophilic surface. This change was due to the addition of antibodies.

XPS analysis was performed ex situ to verify the surface composition of pre- and post-functionalisation of the gold surface (Fig. S1). To achieve this, we first studied the contribution of three elements: sulphur S2p (Fig. S1a), carbon C1s (Fig. S1b), oxygen O1s (Fig. S1c). In the S2p regions, the two peaks were observed representing the Au-S (162.5 eV) and C-SH (~164 eV) bonds. The functionalised sample displayed these peaks with similar peak intensities. In contrast, the bare gold surface showed no peak at these binding energies. The existence of both peaks in the functionalised sample confirmed the successful coating of MUA, as reported in the literature25,26. Three distinct peaks for C1s were observed at ~284 eV (C-C), ~286 eV (C-N and C-O) and ~288 eV (C=O) across control and functionalised samples. The bare gold surface showed C1s peaks, likely attributed to organic residue persisting on the surface after the cleaning process. The intensity of these peaks differed from the functionalised sample. The C-C bonds exhibited the highest intensity, which could be attributed to the presence of the long alkyl chain from 11-MUA on the surface. While the C=O bonds were present, the intensity was insufficient to confirm whether they originated solely from the MUA-only SAM. Studies have shown that cleaned gold surfaces can show minor C1s peaks27,28. This confirmed the secure anchoring of MUA molecules through covalent bonding of thiol moieties to the gold surface and a functional SAM.

Next, we verified the proper functionalisation of the antibody on the generated functional SAM at the surface of the SPR biosensor chip. The primary antibody functionalisation was carried out in situ and monitored by tracking the changes in the refractive index of the SPR biosensor channels, as shown in Fig. 2A. Successful functionalisation should result in a permanent change in the SPR baseline. The initial baseline was established with deionised water. The 11-MUA-treated surface was first activated with EDC/NHS mixture. A change in RU was observed as the mixture activated the surface carboxyl groups to create N-Hydroxysuccinimide esters. Immediately after, 10 µg/mL of cTn antibody was introduced at a lower flow rate of 5 µL/min for 3600 s, which allowed coupling of the EDC/NHS-activated carboxylic groups with the amine groups on the antibodies. An elevation in RU became apparent upon antibody introduction, signifying binding to the SAM on the surface and confirming the successful surface modification with 11-MUA. Unreacted carboxyl groups were then deactivated by flowing a solution of ethanolamine at pH 8 for 300 s. A slight decrease in RU following ethanolamine exposure indicated the removal of unbound antibodies remaining at the sensor surface. 30 µg/mL of BSA was introduced to block any uncovered areas in the functionalised SAM. Negligible change in RU was seen after 600 s, confirming successful surface coverage. The full validation of the SPR approach was presented in Fig. S2.

Fig. 2: Surface functionalisation and binding performance of the SPR immunoassay for cTnI detection.
figure 2

A Antibody immobilisation in situ where the functionalised surface was first activated with EDC/NHS solution, followed by antibody flow and deactivation with ethanolamine. The total antibody immobilised was determined by calculating the difference in signal before antibody immobilisation and after surface deactivation. B Binding of cTnTIC to capture antibody, anti-cTnI, and subsequently to detection antibody, anti-cTnIC; followed by multiple short pulses of regenerative solution, C specificity of capture antibody against cTnI isomers and myoglobin, D specificity of detection antibody anti-cTnIC, against cTnI, E specificity of capture antibody, anti-cTnI, against detection antibody, anti-cTnIC. Vertical dash lines represent injection phases, while horizontal dash lines represent the sensor baseline.

Full size image

Once the 19c7 mAb bound onto the SAM, a solution containing cTnTIC isomers was flowed through to assess the specificity of the SPR biosensor (Fig. 2B). 1 µg/mL of cTnTIC flowed through the functionalised surface, and binding was observed through a change in sensor response. Subsequently, to confirm the presence and binding of the cTnTIC complex, 2 µg/mL of 20C6 anti-cTnIC mAb was introduced, resulting in binding with the bound cTnTIC and a significant change in RU. The surface was then regenerated using short 30 s pulses of 10 mM NaOH, after which the signal returned to the initial baseline, indicating the molecules were unbound from 19c7 mAbs (Fig. 2B)29,30.

Next, the specificity of capture antibodies towards the different troponin isoforms, as well as non-specific control samples, was assessed to validate the two-step SPR approach. 1 µg/mL of cTnI, cTnTIC, and myoglobin, a common negative control in cTn biosensing, were introduced to the sensor separately. As seen in Fig. 2C, positive responses were seen with the cTnI isomers, indicating binding with 19c7 mAb, while no binding was seen with myoglobin, as the response returned to baseline after injection. This confirms that our biosensor is specific and can capture both troponin I isoforms. We then addressed cross-reactivity between the detection cTn mAb and the cTnI isomer. 2 µg/mL of 20c6 mAb was flowed over the sensor surface immobilised with cTnI (Fig. 2D). A positive response was observed when cTnI was introduced. No binding was observed with 20c6 mAb, and the response returned to the initial baseline, confirming the specificity of the antibody towards cTnTIC. The cross-reactivity of the antibody was also assessed using a sensor surface immobilised with 19c7 mAb (Fig. 2E). A change in RU was observed when 20c6 mAb was injected separately onto the sensor surface, but the response quickly returned to baseline once the sensor surface was washed with PBS. As seen in Fig. 2E, 20c6 mAb did not cross-react with 19c7 mAb. These control experiments confirmed the specificity of this approach and the opportunity of using SPR to capture troponin forms and differentiate them.

Analysis of cTnTIC and cTnI isoforms using the SPR detection approach

To quantify the different cTnI isoforms, calibration curves were generated using a series of known concentrations with PBS as the running buffer for both the capture and detection phases (Fig. 3A). These plots illustrate the ΔRU measurement of cTnI and cTnTIC, respectively, from 10 to 1000 ng/mL. Figure 3Ai, Aii represent the detection of these isomers against the capture mAb 19c7, which specifically captures all troponin forms containing cTnI. Meanwhile, Figure 3Aiii depicts the detection of cTnTIC against detection mAb 20c6. As observed, increasing concentrations of each analyte resulted in increased responses of the SPR biosensor, indicating that the SPR response correlated with the number of analytes specifically interacting with the antibodies. Calibration curves for the capture phase of each isoform individually showed an almost linear trend while the secondary detection showed a clear exponential decay. The average coefficient of variation (%CV, the ratio of the standard deviation to the mean) ranged between 11 and 13% which is within the acceptable analytical range31. The smallest amount of analyte that can be detected by the system with confidence is defined as the limit of detection or LoD32. The LoD for each plot was calculated, with the lowest being 0.213 ng/mL for anti-cTnI mAb against cTnI isomer, 0.313 ng/mL against cTnTIC and 0.337 ng/mL for anti-cTnIC mAb against cTnTIC.

Fig. 3: Simultaneous detection and quantification of cTnI and cTnTIC in a single solution.
figure 3

A Calibration plots of different cTn isomers concentrations with capture antibodies anti-cTnI 19c7 against (i) cTnI and (ii) cTnTIC and detection antibody (iii) anti-cTnIC 20c6 against cTnTIC. B Sensorgram showing the binding response of a mixture of cTnTIC and cTnI to sensor surface immobilised with anti-cTnI antibody. Detection was performed using specific anti-cTnIC antibody. Light blue region indicates the injection phase of the cTn analytes, while red represents the injection phases of anti-cTnIC detection antibody. C Percentages of different ratios of cTnTIC:cTnI in relative to total cTn concentrations. All data show mean ± standard deviation (SD) of triplicate measurements.

Full size image

To determine the biosensor’s capability to quantify the individual isomers using this approach, we examined its performance in detecting cTnTIC and I isoforms in different ratios. We prepared solutions containing these two purified isoforms to a final total concentration of 1000 ng/mL, analysed the sensor response, and quantified them based on the calibration plots in Fig. 3A. Anti-cTnIC antibody was used to detect cTnTIC. We hypothesised that each new injection of antibodies would result in a measurable change in sensor response, indicating successful binding events on the sensor surface. As seen in Fig. 3B, a change in response was seen when cTnTIC and cTnI isomers were introduced onto the sensor surface immobilised with 19c7 capture mAb. The signal was allowed to stabilise before a washing step was performed to remove unbound cTnI isomers. Anti-cTnIC mAb was then injected, and another change in response was seen as the antibodies bound to cTnTIC, followed by another washing step to remove any unbound antibodies. The final recorded signal represented the total amount of cTnTIC captured from the cTnTIC and cTnI mixture. To fully quantify the amount of cTnI isomers captured, the analysis was employed as depicted in Fig. 1D.

Since the mixed isoforms could be detected in solution, we then explored if the captured cTnI isoforms could also be quantified with values closely matching their expected proportions. As seen in Fig. 3C, five ratios were prepared to mimic the potential spectrum of acute myocardial injuries. Overall, the measured cTnTIC:cTnI ratios followed the expected trend. The measured 50:50 ratio was the closest to the expected value, 52:48, with a CV% of 9% for cTnIC and 10% for cTnI. Both the 90:10 and 10:90 ratios followed the expected ratio, with the higher concentration cTn isomer showing an underestimation of 25% cTnTIC and 14% cTnI, respectively. The signals were reproducible with CV% between 3 to 12%.

After successfully detecting two isomers, we proceeded to test with three isomers in the same sample: cTnTIC, cTnIC and cTnI. To do so, we prepared cTnIC isomer and anti-cTnT 7F4 mAbs to differentiate between cTnTIC and cTnIC. Prior to that, the specificity of capture antibody anti-cTnI towards cTnIC and cross-reactivity of anti-cTnT antibodies against anti-cTnI, cTnI and cTnIC isomers were evaluated. Anti-cTnT antibodies were introduced to a sensor surface with only immobilised anti-cTnI and no binding was observed (Fig. S3A). After the binding of cTnI to immobilised anti-cTnI, anti-cTnT mAbs were flowed through and no response was observed (Figure S3B). This was repeated with cTnIC and a positive response was observed for the specific binding of cTnIC to anti-cTnI. However, when anti-cTnT 7F4 was subsequently flowed through, no binding occurred with cTnIC (Fig. S3C). These results confirm the specificity of the capture antibodies and the lack of cross-reactivity among the tested isomers and antibodies. Following that, three additional calibration curves were prepared: (i) capture anti-cTnI mAb against cTnIC (Figure 4Ai), (ii) detection anti-cTnIC mAb against cTnIC (Fig. 4Aii) and (iii) anti-cTnT against cTnTIC (Fig. 4Aiii). As seen with Fig. 4A, similar to Fig. 3A, calibration curves for the capture phase of cTnIC showed an almost linear trend, while the secondary detection showed a clear exponential decay. The average CV% for these plots was between 12 to 15% and the LoD was 0.698 ng/mL for anti-cTnI against cTnIC, 0.241 ng/mL for anti-cTnIC against cTnIC and 0.540 ng/mL for anti-cTnT against cTnTIC.

Fig. 4: Simultaneous detection and quantification of cTnI, cTnIC and cTnTIC in a single solution.
figure 4

A Calibration plots of different cTn isomers concentrations with (i) capture antibody anti-cTnI 19c7 against cTnIC as well as detection antibodies (ii) anti-cTnIC 20c6 against cTnIC and (iii) anti-cTnT 7F4 against cTnTIC. B Flow chart quantifying three cTn isomers. C Sensorgram showing the binding response of a mixture of cTnTIC, cTnIC and cTnI to sensor surface immobilised with anti-cTnI antibody. Detection was performed using specific anti-cTnT and anti-cTnIC antibodies. Light blue region indicates the injection phase of the cTn analytes. Purple represents injection phases of anti-ctnT detection antibodies while red represents the injection phases of anti-cTnIC detection antibodies. D Percentages of 33:33:33 ratio of cTnTIC:cTnIC:cTnI relative to total cTn concentrations. All data show mean ± standard deviation (SD) of triplicate measurements.

Full size image

The calculation for three isomers was an extension of the two isomers and performed as highlighted in Fig. 4B. The RU of secondary anti-cTnT 7F4 antibody (RUB) was first used to determine the concentration of the complex form cTnTIC present in the sample. This deduced concentration of cTnTIC was then used to determine the RU of cTnIC from the calibration curve of anti-cTnIC against cTnTIC (RU20c6/TIC). Since anti-cTnIC recognises both cTnIC and cTnTIC, the RU for cTnIC alone was obtained by subtracting the total measured RU of cTnTIC and cTnIC (RUC) with RU20c6/TIC. The cTnIC concentration was then determined from the anti-cTnIC versus cTnIC plot. Next, the concentrations for cTnTIC and cTnIC were applied to their respective calibration plots for anti-cTnI binding to calculate their RU contributions (RU19c7/TIC and RU19c7/IC). Finally, the RU for free cTnI was determined by subtracting RU19c7/TIC and RU19c7/IC from the total anti-cTnI signal (RUA). This final RU for cTnI was then used to determine the true concentration of free cTnI. We hypothesised that all three isomers would bind to the immobilised anti-cTnI and a measurable change would be observed upon sequential introduction of both detection antibodies.

As seen in Fig. 4C, a change in response was observed when the mixed solution was introduced, indicating the binding of cTn isomers to the sensor surface. Once the signal stabilised, anti-cTnT was injected, resulting in a further change in response, confirming the binding of the antibody to the cTnTIC complex. This was followed by a washing step. Subsequently, anti-cTnIC was introduced, leading to an additional increase in response as both cTnIC and cTnTIC were detected. Another washing step was performed to remove any unbound antibodies. We then tested a solution containing an equimolar mixture of cTnTIC:cTnIC:cTnI at 33:33:33% (Fig. 4D). The measured ratios were close to 51:4:45% with a CV% of 1% for cTnTIC, 35% for cTnIC, and 6% for cTnI. The ratios for cTnTIC and cTnI were comparable to the 50:50% cTnTIC:cTnI mixture shown in Fig. 3C, whereas the cTnIC component showed a significant deviation in its measured concentration and exhibited a high CV% of 35%, indicating that cTnIC could not be reliably detected in the three-isomer mixture. However, cTnTIC and cTnI showed reproducible signals during this assessment, with overall CV <13%.

Discussion

Here, we presented a label-free two-step detection methodology to detect and differentiate between cTnI and cTnTIC, using the gold standard SPR photonic biosensor. The biosensor was first functionalised to ensure specificity for these isoforms, validated using control samples, and analysed with calibration curves. A percentage-based ratio was used to represent the relative proportions of cTnTIC and cTnI during acute myocardial injury. The two-step approach successfully detected and differentiated between these two isoforms. Such capability is valuable for understanding the metabolic pathways of myocardial injury, as varying levels of different cTn isomers may correspond to different stages of cardiac damage or even indicate non-cardiac conditions8,9,33.

The approach detected varying ratios of cTnTIC and cTnI, but some underestimation was observed (Fig. 3C). This may be due to competition for free antibodies on the sensor surface, as both isomers interact with the same capture layer. We hypothesise that their differing molecular weights (~80 kDa for cTnTIC and ~24 kDa for cTnI) influence this behaviour, particularly at high concentrations. In such cases, steric hindrance effects could arise, where larger complexes impede access to adjacent binding sites or restrict proper alignment for efficient binding34. Furthermore, the calibration curves for the detection antibody against cTnTIC (Fig. 3Aiii) exhibited an exponential decay at higher concentrations. This could result from partial saturation of the captured isoform on the sensor surface, leading to reduced surface accessibility. These observations suggested that surface saturation and limited antibody accessibility may contribute to reduced binding efficiency. While further improvements could be achieved by refining the sensor surface architecture, such as enhancing antibody orientation or adjusting its density35, these effects were considered acceptable for the purpose of evaluating our detection strategy.

When three cTnI-related complexes, cTnTIC, cTnIC, and cTnI, were analysed simultaneously, the biosensor was able to reliably detect cTnTIC and cTnI but not cTnIC. One possible explanation is that cTnIC may be degrading over time, contributing to the lower readings from the sensor compared to the premixed concentrations36. We found that even with purified cTnIC, the biosensor signal declined substantially after just two measurement cycles (see Fig. S4), with a CV% of 29%. This instability may arise from the lack of a stabilising agent, causing cTnIC to degrade into cTnI, oxidising to form disulfide linkages that are no longer recognised by the antibodies37. Prior work also indicates that troponin stability is temperature and buffer-dependent14,38. Greater dissociation of cTn was observed at higher temperature (37 °C versus 4 °C) and buffer formulation affects cTn stability, so our use of PBS buffer alone may have further compromised cTnIC at ambient temperature. Consistent with this, the calibration plots in Fig. 4Ai and Fig. 4Aiii showed that the LoD was substantially higher for anti-cTnI against cTnIC (0.698 ng/mL), and there was considerable signal variability in the measurements for cTnIC against anti-cTnIC. Additional testings are needed to preserve the structural integrity of cTnIC isoform to enable comprehensive characterisation of all cTn isoforms. However, this instability does not appear to affect the measurement of the key isomers, cTnTIC and cTnI, suggesting that this method remains valuable for studying the ratio between these two in a quantitative manner. Finally, while the LoD values obtained in both Fig. 3A and Fig. 4A (200–700 ng/L, reported in ng/L to align with International Federation of Clinical Chemistry standards) are higher than those of current PoC assays (typically ~1 ng/L)39, the present work using SPR demonstrated the feasibility of isoform-resolved detection.

Future work could involve categorising each degradation step through stage-specific isomers as demonstrated in a study using SPR to detect various isoforms of mRNA40. Employing a more sophisticated microfluidic system capable of precisely controlling the flow within each microchannel would greatly enhance multiplexing and controllability of the studies, enabling broader coverage of cTn isomers detection from intact to degraded fragments41. This capability may contribute to future characterisation of disease status in patients with elevated cTn levels in real-time based on the differences in their cTn composition. Additionally, translating this approach to a more integrated version or other advanced photonic biosensor platforms could be beneficial, as SPR has its limitations. For example, integrated silicon-based biosensors could be more advantageous as a point-of-care device in terms of integration, miniaturisation, and higher multiplexing capabilities, with sensitivity values around 10−7 refractive index units for bulk sensing42,43,44 and fg/mm2 for surface sensing in a direct assay45, surpassing SPR by more than one order of magnitude and offering the improved analytical sensitivity required for clinical studies. Stability of cTn isomers, especially cTnIC, may be enhanced by conducting experiments at lower temperatures (4 °C) with the addition of stabilising agents14,37.

This study identified areas for improving assay performance and sensor reliability. First, signal variability at both high and low cTn concentrations impacted quantification accuracy, likely due to antibody saturation and incomplete regeneration. Second, there was no consideration for truncated and low-molecular-weight cTnTIC complexes and cTnT in this study9,12. Incorporating antibodies that detect these isomers, as well as the degraded free cTnI and cTnT would further improve the sensitivity and accuracy of the assay. Finally, the study was conducted using purified cTn isoforms in buffer solutions. Transitioning from purified version to serum and patient samples will introduce complexities that need to be addressed, such as interference from serum components46. The next steps will involve a systematic progression from spiking experiments in serum or plasma to assess matrix effects, to testing in retrospective patient samples for comparison against established assays, and ultimately to validation in patient cohorts. This stepwise approach will ensure that assay performance is robust in complex biological environments and clinically meaningful in real-world diagnostic settings.

Methods

Reagents and buffers

All cardiac troponin isomers (cTnTIC, cTnIC, cTnI) and anti-cTn monoclonal antibodies (mAbs) (anti-cTnI 19c7, anti-cTnIC 20c6, anti-cTnT 7F4) were purchased from HyTest (Finland). The epitopes captured by these antibodies are shown in Table S1. The following reagents were purchased from Merck KGga (Germany): human myoglobin, bovine serum albumin (BSA), phosphate buffer saline (PBS) (0.01 M phosphate buffer, 0.137 M sodium chloride pH 7.4) tablets, MES hydrate, ethanolamine, 11-Mercapto-undecanoic acid (11-MUA), N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC), N-Hydroxysuccinimide (NHS), sodium hydroxide (NaOH), Tween-20, acetone and pure ethanol.

SPR biosensor

A commercial SPR-based biosensor in a Kretschmann configuration was used for this work (Pioneer FE System, Sartorius, Canada). This instrument employs angle-based interrogation. For each channel, the resonance angle is determined from periodic angular reflectivity scans known as dips. The binding responses are observed as shifts in the resonance angle, which the instrument converts into response units (RU) for real-time readout following the reference-channel subtraction. The in-built fluidic system comprises of three programmable, serial (non-parallel) channels, one of which served as the reference. Measurements were performed at ambient temperature (±0.1 −0.2 °C) with on-board temperature control, and baselines were allowed to stabilise pre- and post-injections. Data was sampled at 1 Hz.

SPR chip cleaning and functionalisation

Unmounted gold chips were purchased from Xantec (Germany) and functionalised using thiol-based alkanethiol, 11-mercapto-undecanoic acid (11-MUA), to detect specific cTn isoforms. To determine the proper functionalisation of the gold surfaces, characterisation was conducted separately on individual silicon chips coated with a thin gold layer to model the chemistry on the actual SPR sensor chips. A bare Au chip was first cleaned with acetone, ethanol, followed by deionised water on a heated plate close to the boiling point without drying out. The chip was then dried thoroughly under streams of nitrogen. The chip surface was oxidised in UV/Ozone equipment for 20 min. After oxidation, the sensor surface was cleaned with ethanol, deionised water and dried again under a nitrogen stream. The self-assembled monolayer formation on the chip was assembled by preparing an 11-MUA substrate in 95% w/v ethanol solution to a final concentration of 1 mM. The chip was immersed in the 11-MUA solution overnight in the dark at room temperature. The beaker was sealed in parafilm to prevent ethanol evaporation and wrapped in aluminium foil to protect from light. After overnight incubation, the gold chip was rinsed with ethanol and deionised water and dried with nitrogen. Antibody immobilisation was carried out by incubating the test chip in a solution of 0.2 M EDC/0.05 M NHS for 20 min23. The chip was then thoroughly washed with water and dried under nitrogen flux. Immediately after, the chip was incubated with the antibody for 1 h to allow for covalent bonding with the intermediate reactive groups generated at the sensor surface. The chips were then thoroughly washed, dried and subjected to surface characterisation analysis.

For SPR measurements, the same cleaning and functionalisation protocol was carried out. MUA-functionalised SPR chips were mounted according to Xantec’s instructions for SPR measurement. For antibody immobilisation, we carried out an in situ (in flow) approach to verify the process on the sensor chip. The chip surface was first primed with deionised water until a stable baseline was observed. A solution of 0.2 M EDC/0.05 M NHS was flowed into the flow channels at 10 µL/min for 20 min23. The primary antibody (in PBS, pH 4.3) flowed immediately after at 5 µL/min for 60 min. The reaction was deactivated with 1 M of ethanolamine at pH 8 for 5 min to avoid the covalent bonding of other molecules that may flow over the sensor in subsequent analyses. The specificity of the surface was tested with 30 µg/mL of BSA for 10 min right after47.

Surface characterisation SPR biosensor

X-ray photoelectron spectroscopy (XPS) analysis and water contact angle (WCA) analysis were performed to characterise the surface functionalisation of the gold chip. XPS was performed using a Kratos AXIS Supra XPS instrument (Kratos Analytical Limited, UK). The Au samples (1 cm × 1 cm) were mounted on the XPS stage. Samples were prepared for each step of the functionalisation process: bare gold, MUA-SAM only and MUA linked with antibody. The binding energy scale was calibrated using the Au4f7/2 peak at 83.9 eV. The survey spectra were collected from 0 to 1400 eV and high-resolution spectra were collected for each element detected (Au, C, S, N and O) with a pass energy of 40 eV. Analysis was then performed using CasaXPS software. WCA of 30 µL of distilled water droplet on each surface was measured using OCA20 goniometer (DataPhysics Instrument, Germany) to determine the surface wettability of the SPR biosensor surface. The angle of the droplet was measured using the instrument’s accompanying software.

SPR assessment of different cTn isomers

The sensor surface was primed with 1x PBS running buffer prior to any protein measurement. To measure the binding of cTn isomer to the capture and detection mAbs, cTn isomer was flowed for 10 min, followed by 2 µg/mL secondary mAb for another 10 min. The flow rates were fixed at 20 µL/min. The sensor surface was regenerated with 10 mM of NaOH in pulses of 30 s until the initial baseline was recovered. After cTn flowed through the fluidic channel and unbound cTn was washed away with PBS running buffer, the binding response was allowed to stabilise and an average of 10 s on the response curve was selected as the mean binding response. This process was also applied to the secondary antibody binding. The flow rate was 20 µL/min. cTn concentration ranging from 10 to 1000 ng/mL was measured for cTnI and cTnTIC isoforms against the change in RU (ΔRU). During the first capture, ΔRU was plotted against 10 to 1000 ng/mL concentration for cTnI and cTnTIC isomer. In the secondary capture, ΔRU obtained from the interaction of a fixed concentration of secondary mAb was plotted against the respective cTn isomer concentration captured in the first step. Commencing with the larger isomer, cTnTIC, the ΔRU resulting from the binding with detection 20c6 mAb directly indicates the concentration of cTnTIC in the solution. The ΔRU of anti-cTnI to cTnTIC was then determined. By subtracting the measure ΔRU from the ΔRU of 19c7 mAb, the concentration of cTnI is then quantified. The percentage of coefficient variation (%CV) was experimentally evaluated for each of the triplicate (or more) measurements and averaged. CV, the analytical precision of an assay, was calculated by mean/standard deviation × 100 and expressed in percentage48. The limit of detection (LoD) was calculated as three times the standard deviation of the noise signal, and the corresponding value was interpolated from the calibration curve to obtain the exact concentration.

Assessment of cTnTIC:cTnI solutions with SPR

The solutions were prepared in percentages of 100:0, 90:10, 50:50, 10:90 and 0:100 for cTnTIC and cTnI and 33:33:33 across three different cTn isomers (cTnTIC, cTnIC and cTnI) to a final total volume of 1 mL. These solutions were prepared in 1x PBS running buffer and injected at a rate of 20 µL/min for a duration of 10 min. Subsequently, 2 µg/mL of detection mAb was introduced after the baseline had stabilised, flowing at 20 µL/min for an additional 10 min. The sensor surface was regenerated with 10 mM of NaOH in pulses of 30 s until ithe nitial baseline was restored. The concentration of each isomer was determined by interpolation and deduction from three calibration curves. The deduced percentages were calculated by dividing the specific cTnI isomer concentration by the total detected cTn concentration. The data was analysed and plotted in GraphPad Prism 10 (Boston, USA) and Origin software (OriginLab, Northampton, USA).

Data availability

All data generated or analysed during this study are included in this published article and its supplementary information files or are available from the corresponding author upon reasonable request.

References

  1. Thygesen, K. et al. Fourth universal definition of myocardial infarction (2018). Eur. Heart J. 40, 237–269 (2019).

    Article  Google Scholar 

  2. Scott, I. A., Cullen, L., Tate, J. R. & Parsonage, W. Highly sensitive troponin assays — a two-edged sword?. Med. J. Aust. 197, 320–323 (2012).

    Article  Google Scholar 

  3. Airaksinen, K. E. J. et al. Novel Troponin fragmentation assay to discriminate between troponin elevations in acute myocardial infarction and end-stage renal disease. Circulation 146, 1408–1410 (2022).

    Article  Google Scholar 

  4. Vroemen, W. H. M. et al. Cardiac Troponin T: Only small molecules in recreational runners after marathon completion. J. Appl. Lab Med 3, 909–911 (2019).

    Article  Google Scholar 

  5. Cullen, L., Collinson, P. O. & Giannitsis, E. Point-of-care testing with high-sensitivity cardiac troponin assays: the challenges and opportunities. Emerg. Med. J. 39, 861–866 (2022).

    Article  Google Scholar 

  6. Apple, F. S. et al. Single high-sensitivity point-of-care whole-blood cardiac Troponin I measurement to rule out acute myocardial infarction at low risk. Circulation 146, 1918–1929 (2022).

    Article  Google Scholar 

  7. Vylegzhanina, A. V. et al. Full-size and partially truncated cardiac troponin complexes in the blood of patients with acute myocardial infarction. Clin. Chem. 65, 882–892 (2019).

    Article  Google Scholar 

  8. Li, L. et al. Design and analytical evaluation of novel Cardiac Troponin assays targeting multiple forms of the Cardiac Troponin I–Cardiac Troponin T–Troponin C Complex and fragmentation forms. Clin. Chem. 71, 387–395 (2025).

    Article  Google Scholar 

  9. Katrukha, I. A. & Katrukha, A. G. Myocardial Injury and the release of Troponins I and T in the blood of patients. Clin. Chem. 67, 124–130 (2021).

    Article  Google Scholar 

  10. Katrukha, I. A. et al. Full-size cardiac Troponin I and Its Proteolytic fragments in blood of patients with acute myocardial infarction: antibody selection for assay development. Clin. Chem. 64, 1104–1112 (2018).

    Article  Google Scholar 

  11. Katrukha, I. A. et al. Fragmentation of human cardiac troponin T after acute myocardial infarction. Clin. Chim. Acta 542, 117281 (2023).

    Article  Google Scholar 

  12. Damen, S. A. J. et al. Cardiac Troponin composition characterization after Non ST-Elevation Myocardial Infarction: Relation with Culprit Artery, Ischemic Time Window, and Severity of Injury. Clin. Chem. 67, 227–236 (2021).

    Article  Google Scholar 

  13. Giuliani, I. et al. Determination of cardiac troponin I forms in the blood of patients with acute myocardial infarction and patients receiving crystalloid or cold blood cardioplegia. Clin. Chem. 45, 213–222 (1999).

    Article  Google Scholar 

  14. van Wijk, X. M. R. et al. Cardiac troponin I is present in plasma of type 1 myocardial infarction patients and patients with troponin I elevations due to other etiologies as complex with little free I. Clin. Biochem 73, 35–43 (2019).

    Article  Google Scholar 

  15. Damen, S. A. J. et al. Cardiac troponin complexes and fragments: potential targets for improved clinical performance. Crit. Rev. Clin. Lab Sci. 62, 313–326 (2025).

    Article  Google Scholar 

  16. Li, L. et al. Characterization of Cardiac Troponin fragment composition reveals potential for differentiating etiologies of myocardial injury. Clin. Chem. 71, 396–405 (2025).

    Article  Google Scholar 

  17. Huertas, C. S., Calvo-Lozano, O., Mitchell, A. & Lechuga, L. M. Advanced evanescent-wave optical biosensors for the detection of nucleic acids: an analytic perspective. Front. Chem. 7, 724 (2019).

  18. Soler, M., Huertas, C. S. & Lechuga, L. M. Label-free plasmonic biosensors for point-of-care diagnostics: a review. Expert Rev. Mol. Diagn. 19, 71–81 (2019).

    Article  Google Scholar 

  19. Ranjbari, F., Nosrat, A., Fathi, F. & Mohammadzadeh, A. Surface plasmon resonance biosensors for early troponin detection. Clin. Chim. Acta 558, 118670 (2024).

    Article  Google Scholar 

  20. Campu, A. et al. Cardiac Troponin biosensor designs: current developments and remaining challenges. Int J. Mol. Sci. 23, 7728 (2022).

    Article  Google Scholar 

  21. Katrukha, A. G. et al. Degradation of cardiac troponin I: implication for reliable immunodetection. Clin. Chem. 44, 2433 LP–2440 (1998).

    Article  Google Scholar 

  22. Ahmad, A. & Moore, E. Electrochemical immunosensor modified with self-assembled monolayer of 11-mercaptoundecanoic acid on gold electrodes for detection of benzo[a]pyrene in water. Analyst 137, 5839–5844 (2012).

    Article  Google Scholar 

  23. Huertas, C. S., Soler, M., Estevez, M.-C. & Lechuga, L. M. One-step immobilization of antibodies and DNA on gold sensor surfaces via a poly-Adenine Oligonucleotide approach. Anal. Chem. 92, 12596–12604 (2020).

    Article  Google Scholar 

  24. Wang, M. S., Palmer, L. B., Schwartz, J. D. & Razatos, A. Evaluating protein attraction and adhesion to biomaterials with the atomic force Microscope. Langmuir 20, 7753–7759 (2004).

    Article  Google Scholar 

  25. Dadafarin, H., Konkov, E. & Omanovic, S. Electrochemical functionalization of a 316L stainless steel surface with a 11-mercaptoundecanoic Acid Monolayer: Stability Studies. Int J. Electrochem Sci. 8, 369–389 (2013).

    Article  Google Scholar 

  26. Mendoza, S. M., Arfaoui, I., Zanarini, S., Paolucci, F. & Rudolf, P. Improvements in the characterization of the crystalline structure of acid-terminated alkanethiol self-assembled monolayers on Au(111). Langmuir 23, 582–588 (2007).

    Article  Google Scholar 

  27. Mirsaleh-Kohan, N., Bass, A. D. & Sanche, L. X-ray photoelectron spectroscopy analysis of gold surfaces after removal of thiolated DNA Oligomers by ultraviolet/ozone treatment. Langmuir 26, 6508–6514 (2010).

    Article  Google Scholar 

  28. Haag, A.-L., Nagai, Y., Lennox, R. B. & Grütter, P. Characterization of a gold coated cantilever surface for biosensing applications. EPJ Tech. Instrum. 2, 1 (2015).

    Article  Google Scholar 

  29. Dillon, P. P., Daly, S. J., Manning, B. M. & O’Kennedy, R. Immunoassay for the determination of morphine-3-glucuronide using a surface plasmon resonance-based biosensor. Biosens. Bioelectron. 18, 217–227 (2003).

    Article  Google Scholar 

  30. Drake, A. W. & Klakamp, S. L. A strategic and systematic approach for the determination of biosensor regeneration conditions. J. Immunol. Methods 371, 165–169 (2011).

    Article  Google Scholar 

  31. Wood, R. How to validate analytical methods. TrAC, Trends Anal. Chem. 18, 624–632 (1999).

    Article  Google Scholar 

  32. Bhalla, N., Jolly, P., Formisano, N. & Estrela, P. Introduction to biosensors. Essays Biochem 60, 1–8 (2016).

    Article  Google Scholar 

  33. van Wijk, X. M. R. & Damen, S. A. J. Advances in cardiac Troponin composition assays: a step closer to the clinic?. Clin. Chem. 3, 387–395 (2024).

    Google Scholar 

  34. Matos, D. M. Steric hindrance: A practical (and frequently forgotten) problem in flow cytometry. Cytom. B Clin. Cytom. 100, 397–401 (2021).

    Article  Google Scholar 

  35. Duschl, C., Sévin-Landais, A. F. & Vogel, H. Surface engineering: optimization of antigen presentation in self-assembled monolayers. Biophys. J. 70, 1985–95 (1996).

    Article  Google Scholar 

  36. Liu, S. et al. Recombinant single chain cardiac troponin I-C polypeptides: superior calibration and control materials for cardiac troponin I immunoassays. Clin. Lab. 47, 19–27 (2001).

    Google Scholar 

  37. Wu, A. H. B. et al. Characterization of cardiac troponin subunit release into serum after acute myocardial infarction and comparison of assays for troponin T and I. Clin. Chem. 44, 1198–1208 (1998).

    Article  Google Scholar 

  38. Riabkova, N. S. et al. Influence of anticoagulants on the dissociation of cardiac Troponin complex in blood samples. Int J. Mol. Sci. 25, 8919 (2024).

    Article  Google Scholar 

  39. The Committee on Clinical Applications of Cardiac Biomarkers (C-CB) of the International Federation of Clinical Chemistry and Laboratory Medicine. Point of Care Cardiac Troponin I and T Assay Analytical Characteristics Designated by Manufacturer IFCC Committee on Clinical Applications of Cardiac Bio-Markers (C-CB) v062024. https://ifccfiles.com/2024/03/Point-of-Care-Cardiac-Troponin-I-and-T-Assay-Analytical-Characteristics-Designated-By-Manufacturer-v062024.pdf (2024).

  40. Huertas, C. S. et al. Site-Specific mRNA cleavage for selective and quantitative profiling of alternative splicing with label-free optical biosensors. Anal. Chem. 91, 15138–15146 (2019).

    Article  Google Scholar 

  41. Szydzik, C. et al. An automated optofluidic biosensor platform combining interferometric sensors and injection moulded microfluidics. Lab Chip 17, 2793–2804 (2017).

    Article  Google Scholar 

  42. Leuermann, J. et al. Optimizing the limit of detection of waveguide-based interferometric biosensor devices. Sensors 19, 3671 (2019).

    Article  Google Scholar 

  43. Knoerzer, M. et al. Optical frequency comb based system for photonic refractive index sensor interrogation. Opt. Express 27, 21532–21545 (2019).

    Article  Google Scholar 

  44. Diamandikos, E. et al. Label-free, sensitive, and direct detection of cardiac troponin biomarkers using frequency-locked microring resonators. Adv Sens. Res. 4 (2025).

  45. Luan, E., Shoman, H., Ratner, D. M., Cheung, K. C. & Chrostowski, L. Silicon photonic biosensors using label-free detection. Sensors 18, 3519 (2018).

    Article  Google Scholar 

  46. Katrukha, I. A. et al. Thrombin-mediated degradation of human cardiac Troponin T. Clin. Chem. 63, 1094–1100 (2017).

    Article  Google Scholar 

  47. Pawula, M., Altintas, Z. & Tothill, I. E. SPR detection of cardiac troponin T for acute myocardial infarction. Talanta 146, 823–830 (2016).

    Article  Google Scholar 

  48. Chenevier-Gobeaux, C. et al. High-sensitivity cardiac troponin assays: Answers to frequently asked questions. Arch. Cardiovasc Dis. 108, 132–149 (2015).

    Article  Google Scholar 

Download references

Acknowledgements

The authors acknowledge the facilities, resources, and scientific and technical assistance from the staff of the Micro Nano Research Facility (MNRF), RMIT Microscopy and Microanalysis Facility (RMMF), and RMIT School of Chemistry, where the study was conducted. S.J.B. acknowledges the support of the Australian Government RTP Scholarship, administered through RMIT University.

Author information

Authors and Affiliations

  1. Integrated Photonics and Applications Centre, School of Engineering, RMIT University, Melbourne, VIC, Australia

    Siew J. Beh, Arnan Mitchell & César S. Huertas

Authors

  1. Siew J. Beh
  2. Arnan Mitchell
  3. César S. Huertas

Contributions

S.J.B., C.S.H., and A.M. conceived the idea. S.J.B. and C.S.H. designed the experimental plan. S.J.B. performed all experiments. S.J.B. and C.S.H. conducted all analyses. S.J.B. and C.S.H. wrote the manuscript. S.J.B. prepared the figures. All authors read, edited, and approved the final manuscript.

Corresponding author

Correspondence to César S. Huertas.

Ethics declarations

Competing interests

All authors are co-founders of HatiSens Pty Ltd, a start-up developing point-of-care technologies for cardiac biomarker detection. HatiSens had no involvement in the conception, design, execution, or funding of this study. The work presented here was carried out using a commercial SPR technology unrelated to the company’s activities, and there was no IP creation involved.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Beh, S.J., Mitchell, A. & Huertas, C.S. Label-free detection of cardiac troponin I and complex using surface plasmon resonance for assessing acute myocardial injuries. npj Biosensing 3, 6 (2026). https://doi.org/10.1038/s44328-025-00068-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • DOI: https://doi.org/10.1038/s44328-025-00068-5

Related Posts