Dual‐Modal Iridium‐Based Self‐Immolative Chemosensors for Differential Responses against Reactive Oxygen Species and their Applications to Detect Diabetes

In diabetes, platelets are activated by several stimuli, and the activated platelets generate reactive oxygen species (ROS) to induce the aggregation of platelets followed by thrombus formation resulting in various cardiovascular diseases. Therefore, detecting ROS perturbations in platelets can provide a clue to diagnosing diabetes. In this paper, iridium‐based self‐immolative probes (1a‐1c) are reported to monitor perturbations of ROS in the blood through photoluminescence (PL) and electrochemiluminescence (ECL). The probes are designed based on an iridium complex conjugated with phenylboronic acid pinacol ester through carbamate moiety. Three probes contain distinct electron‐withdrawing groups at the ortho‐position in the benzyl linker; thus, subtle reactivity differences are expected against ROS. As expected, all three probes exhibit the most apparent PL changes against hydrogen peroxide (H2O2), but their response patterns against ROS are interestingly distinctive. Utilizing such differential ROS responsive pattern, a discrimination strategy is established using a combination of PL and ECL responses, and discrimination of platelets from diabetic and control rats is successfully demonstrated.


Dual-Modal Iridium-Based Self-Immolative Chemosensors for Differential Responses against Reactive Oxygen Species and their Applications to Detect Diabetes
Hey Young Yoon, Yecheol Bak, Seung Bin Park, Subba Rao Cheekatla, Kyung Ho Shin, Sehoon Kim,* and Jun-Seok Lee* DOI: 10.1002/admi.202202408 in the development of diabetes complications such as cardiovascular diseases. [1,5,6] In blood plasma, activated platelets are a main source of ROS and regarded as a circulating oxidative stress to induce disease progression and complications. Platelets can be activated by various stimuli including ROS and the activated platelets generate ROS inside and outside of cells. [7][8][9] In diabetes, hyperglycemia-induced oxidative stress can activate platelets in blood plasma to generate ROS and the consequently increased ROS levels of platelets would contribute to the adhesion and aggregation of platelets and thromboembolic propensity resulting in thrombus formation and atherosclerosis. These progressions will cause cardiovascular diseases, the representative diabetes complications. [8,10] Thus, ROS generated from platelets would be a potential biomarker of diabetes to differentiate diabetic platelets from normal platelets, and detection of ROS of platelets could suggest an additional criteria for diagnosis and therapies of diabetes. [3,8,11] Based on the reported mechanism of the ROS generation, superoxide generation is first step, then hydrogen peroxide (H 2 O 2 ), hydroxyl radial, hypochlorite would be generated. Although various ROS are generated in oxidative stress, most of them have a short lifetime and react with other molecules easily. [2,12] Relatively longer lifetime and stability of H 2 O 2 among other reactive oxygen species have conducted the In diabetes, platelets are activated by several stimuli, and the activated platelets generate reactive oxygen species (ROS) to induce the aggregation of platelets followed by thrombus formation resulting in various cardiovascular diseases. Therefore, detecting ROS perturbations in platelets can provide a clue to diagnosing diabetes. In this paper, iridium-based self-immolative probes (1a-1c) are reported to monitor perturbations of ROS in the blood through photoluminescence (PL) and electrochemiluminescence (ECL). The probes are designed based on an iridium complex conjugated with phenylboronic acid pinacol ester through carbamate moiety. Three probes contain distinct electron-withdrawing groups at the ortho-position in the benzyl linker; thus, subtle reactivity differences are expected against ROS. As expected, all three probes exhibit the most apparent PL changes against hydrogen peroxide (H 2 O 2 ), but their response patterns against ROS are interestingly distinctive. Utilizing such differential ROS responsive pattern, a discrimination strategy is established using a combination of PL and ECL responses, and discrimination of platelets from diabetic and control rats is successfully demonstrated.

Introduction
Oxidative stress, which arises from the overproduction of reactive oxygen species (ROS) or impaired antioxidant system, is a common physiological phenomenon in diverse diseases including diabetes, obesity, and cancer. [1][2][3][4] In diabetes mellitus whose characteristic is hyperglycemia, it is reported that high glucose level in blood plasma triggers the generation of ROS and the hyperglycemia-induced oxidative stress functions importantly www.advmatinterfaces.de development of many molecular probes, using various materials and methods such as absorption, fluorescence, electrochemistry and so on. [13][14][15][16][17][18][19][20][21][22][23][24] However, most of them were used for the studies of intracellular H 2 O 2 generated from mitochondria and reports about blood plasma H 2 O 2 were controversial due to their gradient localization and red blood cells which could scavenge the plasma ROS for the protection of body. [12,[25][26][27] Therefore, for real sample analysis, establishing new sensing strategies and development of a variety of probes are still on demand.
Herein, we developed dual-modal iridium-based self-immolative probes to patterning ROS fingerprints from platelets via photoluminescence and electrochemiluminescence. Electrochemiluminescence (ECL) is a luminescence generated from electrochemical reactions between ECL probes and coreactant such as TPrA on the electrodes. [28,29] Compared to conventional photoluminescence, ECL measurement has great advantages such as high sensitivity and low background signal which is ideal for the clinical diagnostics and environmental analysis. Thus, we combined dual-modal signals for the discrimination of blood samples from animal models between diabetes and controls. The improvement of the detection capability could be achieved by the combined data from the dual method which has its own characteristics, and it can provide diverse and accurate results based on the information-rich output. [30,31]

Design and Synthesis of Probes
Three probes (1a-1c) were designed based on the change of electron density between carbamate moiety and amine functional group on bipyridines (Scheme 1). To prepare the ancillary ligands linked the reaction substrate, phenyl boronic acid pinacol ester, by carbamate, [2,2′-bipyridine]-4,4′-dicarbonyl diazide (3) was synthesized from bipyridine carboxylic acid using diphenylphosphoryl azide (DPPA) (Scheme S1). After that, the acyl azides (3) were reacted with prepared phenyl boronic acid pinacol ester to generate carbamate moiety which is a representative electron-withdrawing group. Once the ancillary ligands were synthesized, the probes (1a-c) were prepared according to reported procedures. [32] Details of the synthesis are provided in Supporting Information.

Photophysical Properties
In virtue of the reaction of H 2 O 2 with boronic acid pinacol ester, the probes(1a-1c) enabled to form bipyridine amine which is an electron-donating group after the reaction with H 2 O 2 , inducing the changes in photo luminescence. To investigate the photophysical properties of the probes(1a-1c) in the response to H 2 O 2 , we used 5 µm probes (1a-1c) and different concentrations of H 2 O 2 in the mixture of 60% of DMSO and 40% of 10 mm phosphate buffer (pH 7.4). 1a and 1b exhibited strong emission at 520 nm, but the maximum peaks of fluorescence were shifted to 500 nm as the concentration of H 2 O 2 was increasing (Figure 1a,b). 1c con-Scheme 1. Structure and proposed ROS sensing mechanism of iridiumbased self-immolative probes.

www.advmatinterfaces.de
taining nitro group on the substrate part which can quench the emission of the probe through PeT process [33,34] showed that the fluorescence intensity at 490 nm was enhanced with increasing the concentration of H 2 O 2 (Figure 1c). Because the maximum peaks of 1a and 1b between before and after the reaction were close to each other, the fluorescence intensity ratio between 460 and 560 nm was plotted in time-and dose-dependent manner, which showed a linear increase in a range from 0 to 128 µm ( Figure S2, Supporting Information). The plot of 1c based on the emission intensity at 490 nm exhibited the same tendency to 1a and 1b. With the standard plots ( Figures S2  and S3, Supporting Information), the limit of detection (LOD) for each probe was determined to be 2.768, 2.698 and 3.084 µm, respectively (3σ/κ, Table 1). In the cases of 1a and 1b, the decreasing of emission at 560 nm could be induced by the releasing only one reaction moiety but bearing the other: the increasing at 460 nm could be attributed to the releasing all boronic acid pinacol esters. The emission of 1c might begin to increase after releasing both substrates. These results indicated that the probes (1a-1c) were able to detect H 2 O 2 through photoluminescence.
Next, we checked the selectivity of these probes (1a-1c) to various reactive species such as hypochlorite (OCl − ), singlet oxygen ( 1 O 2 ), superoxide (O 2− ), hydroxyl radical ( • OH), peroxy radical ( • OOR), oxidized glutathione (GSSG), reduced glutathione (GSH), homocysteine (Hcy) and cysteine (Cys). All three probes exhibited significant strong fluorescence intensity for H 2 O 2 and small enhancement for O 2− compared to other species (Figure 1d-f). Considering that H 2 O 2 is more stable and sustainable than superoxide in physiological milieu, the probes (1a-1c) were demonstrated to be capable for detection of H 2 O 2 selectively in biological environment.

Electrochemiluminescent Properties of Probes
To assess ECL properties of probes (1a-1c), we measured the ECL intensities of the probes (1a-1c) and (dfppy) 2 Ir-bpyNH 2 . While 1a and 1b showed similar intensity to (dfppy) 2 Ir-bpyNH 2 , the ECL intensity of 1c was quenched ( Figure S4, Supporting Information). Then, to ascertain the ECL signal changes of the probes (1a-1c) in the addition of H 2 O 2 , the ECL intensities of 1a, 1b and 1c were observed as increasing the amount of H 2 O 2 in the mixture of acetonitrile and 10 mm phosphate buffer (6:4) containing 10 mm TPrA and 100 mm TBAP. As expected, the ECL intensity of 1c was enhanced about 2.7-folds, but there were no dramatic changes with 1a and 1b (Figure 2a; Figure S2, Supporting Information). Using ECL method, limit of detection (LOD) of 1c was calculated as 1.645 µM and this value was more sensitive than the LOD value of 1c with PL method (3.084 µM). The ECL intensity of 1c was saturated with 128 µm H 2 O 2 ( Figure S3c, Supporting Information). Compared to PL response, it might be assumed that only one substrate releasing generates ECL signal and the ECL properties of the products which contained one boronate or no boronate might be similar. The selectivity of 1c was explored with 500 µm reactive species under the same condition. The ECL intensity toward H 2 O 2 was strongest among those species, however, the signals on 1 O 2 and O 2− were also notably increased. These enhanced ECL signals were presumably ascribed that the analytes affect the ECL generation mechanisms. We also considered such cross-reactivity would produce combined signal of multiple ROS in the analyte, which is potentially favorable to generate more distinctive fingerprints.
ROS patterns from platelets Platelets are reported as an important factor in diabetes and cardiovascular diseases, which can generate ROS inside and outside of cells and circulate whole body, implicating in other cell types. [8] To scrutinize the feasibility of the probes for biological samples, we applied 1c for platelets cells and lysates which were prepared from a diabetes rat and a normal rat. To avoid the interference by red blood cells, we used purified platelets from each rat model [35] 1c was treated with the platelets (2 × 10 8 cells mL −1 ) and lysates, obtaining from a diabetes rat and a normal rat, respectively. Because 1c was cell-impermeable, the response of 1c with platelets would be induced by extracellular ROS, whereas the emission from lysate solution with 1c would be generated  www.advmatinterfaces.de by intracellular ROS. After 1 h incubation at 37 °C, the fluorescent intensities were observed. As expected, the fluorescent intensities of 1c in samples from diabetes rat were increased, compared to the samples from normal rat (Figure 3a,b). Meanwhile, because catalase is responsible to degrade H 2 O 2 in biological environments, an irreversible catalase inhibitor, 3-amino-1,2,4-triazole (AT, 10 mm) [36] was pretreated in the platelet and the lysates in 1 h before the addition of 1c. The samples containing AT exhibited higher intensities than the samples without the inhibitor in both of normal and diabetes samples, which support the major contribution of fluorescence changes of 1c were originated from the H 2 O 2 from platelets and in lysates. On the other hand, there are still basal signals generated from the combined effect of all other ROS, both extracellular and intracellular samples.

Differentiation of Diabetes
In buffer condition, 1c was evaluated as the lower LOD value of using ECL method because of its intrinsic sensitivity and 1a, 1b and 1c exhibited the high selectivity using photoluminescence method. To improve the differential capability of the probes (1a-1c) for the usage in a practical sample, we applied principal component analysis (PCA) establishing a small sensor array set using photoluminescence and ECL to combine the advantages of both methods. We purified platelets from 2 mL blood of one normal rat and one diabetes rat alive. To achieve the same concentration of platelets as a practical sample, the platelet samples and lysates samples were prepared with the same volume of buffer to the blood volume we collected. The platelets and lysate solutions were treated with the probes (1a-1c) and the samples were incubated for 1 h at 37 °C. The samples containing 1c were used for the measurement of both fluorescent intensities and ECL, and the samples containing 1a and 1b were used for fluorescent intensity measurements (4 replicates). Subsequently, PCA analysis was carried out with the obtained data. As shown in Figure 3d, which was analyzed with lysate samples, the group of diabetes were located on right top side and the normal case group was located on right bottom side. The probe group was located on left side. The results from the platelets samples regarded as extracellular ROS, exhibited the similar clustering (Figure 3c). These showed that the PCA analysis using the dual-modal iridium-based selfimmolative probes array could distinguish and make a good clustering of the sample depending on the disease state.

Conclusion
In summary, we synthesized three probes (1a-1c) composing of iridium complex scaffold to generate distinctive responses against various ROS and revealed their good properties of photoluminescence and electrochemiluminescence. 1a and 1b showed major response for H 2 O 2 through photoluminescence, while 1c presented both photoluminescence and ECL methods.
To exploit the probes for practical samples, we built up an array set constituted with dual-modal probes for PL and ECL and performed PCA analysis, complementing both methods to attain high sensitivity and accuracy. Finally, the array set was showed the capability to differentiate diabetes models from a normal www.advmatinterfaces.de model. Despite further efforts are still required, this report could be of help for a future diagnosis platform and therapeutic strategies for many oxidative stress-associated diseases.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.