Detection of apoptosis by [¹⁸F]ML-10 after cardiac ischemia–reperfusion injury in mice

Ischemic heart disease is still a burden in western society [1, 2].

Ruptured atherosclerotic plaques in coronary vessels result in cardiac ischemia and loss of viable tissue. Coronary bypass surgery and percutaneous coronary intervention (PCI) aim to re-establish the perfusion of ischemic cardiac tissue [3]. Translational animal models decipher ongoing cardiac processes after injury and aim to reduce the burden of ischemic heart disease. A better understanding could transfer novel diagnostic and therapeutic applications into the daily clinical scenario, thereby reducing the number of patients suffering from ischemic heart failure.

Plenty of reviews depict the intracellular mechanism of apoptosis, a process of controlled cell death [4‐6]. In brief, the intrinsic and extrinsic pathways leading to apoptosis converge in the activation of executor enzymes, so-called caspases, and result in biochemical and morphological alteration of apoptosis (e.g., chromatin condensation, shrinkage, and breakdown of proteins and DNA [7, 8]). This process’s main feature is the externalization of phospholipid phosphatidylserine into the cell membrane [5].

Positron emission tomography (PET) could offer a feasible method for in vivo imaging myocardial damage. Imaging apoptosis induced by ischemic heart injury could implement novel diagnostics and monitor therapeutic interventions [9‐11].

The small molecule probe, 2-(5-fluoropentyl)-2-methyl-malonic acid (ML-10) (206 Da), could display a promising approach in apoptosis imaging. ML-10 is incorporated and accumulated in apoptotic cells, not viable or necrotic cells, thereby discriminating those different pathologies visualized by in vivo imaging [12].

PET probes for imaging apoptosis such as Annexin-V have been extensively studied. Annexin-V helped in detecting apoptosis in myocardial infarction, atherosclerotic plaques, and in monitoring cancer therapies [13‐16]. However, some limitations prevented the translation of Annexin-V into the clinic. Annexin-V binds to numerous phosphatidylserine head groups on cell surfaces, and is thereby limited by its specificity and its ability to discriminate apoptotic and necrotic cells, the slow clearance of non-targeted tissues, and its large protein structure (36 kDa) (reviewed in [10, 17]). Therefore, new strategies are needed for the development of imaging cell death.

[¹⁸F]ML-10 visualized apoptotic cancer cells [18] and evaluated the effect of radiation therapy on brain cancer and metastasis [19, 20].

In a murine stroke model induced by ligation of the arteria cerebri, the molecular probe [¹⁸F]ML-10 was feasible to detect ischemia, while there was no accumulation in non-ischemic brains [21]. Regarding cardiovascular disease models, [¹⁸F]ML-10 accumulation detected apoptotic cells within atherosclerosis-like lesions in rabbits [22]. Injured aortas were further evaluated by ex vivo autoradiography and histology that identified the [¹⁸F]ML-10 accumulation. Two surgical myocardial infarction models are established to mimic ischemic cardiac injury [23, 24]. The permanent ligation of the LAD artery without reperfusion results in a significant cardiac defect. On the other hand, the transient LAD ligation enables the translation of the early recanalization into an animal model, resulting in ischemia–reperfusion injury (reviewed in [25]). Recently, our group showed that [18F]ML-10 could be used to detect apoptosis after permanent ligation of the LAD in mice illustrating an untreated myocardial infarction [26]. Here, we show data of [¹⁸F]ML-10 in the ischemia–reperfusion injury, resembling the model of re-established perfusion in myocardial infarction, which represents an important model of myocardial infarction besides permanent LAD ligation.

This study aimed to assess [¹⁸F]ML-10 in mice after transient LAD ligation inducing cardiac ischemia–reperfusion by autoradiography, cardiac PET imaging, and histology.

This study is the first to assess the characteristics of [¹⁸F]ML-10 in a mouse model of cardiac ischemia–reperfusion injury. Using autoradiography, small-animal PET imaging, and histology by TUNEL stain, we assessed the feasibility of [¹⁸F]ML-10 after cardiac ischemia–reperfusion injury.

Previous studies evaluated the feasibility of [¹⁸F]ML-10 for apoptosis detection in rats after MI [33]. The accumulation of [¹⁸F]ML-10 was detected early on the first day till the third day after permanent LAD ligation. No uptake was detected at later stages, such as days 5 and 7 post-operation. Detecting apoptosis in human cardiac tissue after myocardial infarction is applicable relatively early after injury (ranging from 2 to 3 h) [34].

Previously, we could show that [18F]ML-10 accumulates in the defect area after permanent LAD ligation [26]. However, the model of transient LAD ligation inducing the ischemia–reperfusion injury better translated the clinical scenario of cardiac reperfusion of an occluded artery (e.g., by percutaneous coronary intervention or bypass surgery), and therefore preserved ejection fraction into pre-clinical research. While the permanent LAD ligation model is, on the other hand, a pivotal model for myocardial infarcts that are not treated in time, and therefore show profound scarring, left ventricular dilatation and heart failure with reduced ejection fraction [35]. Consequently, both models of myocardial infarction are distinct and of pivotal importance in basic cardiac research. The reperfusion of ischemic cardiac tissue could influence the results due to tracer availability in the ischemic area.”

Consequently, imaging early apoptosis after ischemic cardiac injury could provide a promising approach. We could detect a valid uptake of [¹⁸F]ML-10 in the autoradiography, as indicated by the increased TBR. Our data show a dynamic process since the TBR normalized 48 h after cardiac injury. In the next step, [¹⁸F]FDG PET imaging was utilized to localize the defect area in vivo after myocardial injury, as described previously [36]. We identified the infarct area by diminished [¹⁸F]FDG, and could therefore estimate the [¹⁸F]ML-10 uptake distal to the transient LAD ligation. Importantly, we could not detect a relevant uptake of [¹⁸F]ML-10 in sham-operated control mice (see Fig. 2D).

Evaluation of the %ID/g max showed localized uptake and a steady decline in [¹⁸F]ML-10 after transient LAD ligation, which is in line with the decrease in TBR after 48 h. [¹⁸F]ML-10 PET imaging estimated the %ID/g max in the infarct area determined by the diminished [¹⁸F]FDG uptake.

In our experiments, we observed a different course of [¹⁸F]ML-10 in autoradiography compared to in vivo PET imaging at the early time points after cardiac injury. This observation could be due to a technical limitation represented by comparing in vivo and ex vivo analyses. While [¹⁸F]FDG was used to localize the PET imaging analysis in vivo, the autoradiography uses the TBR, which also considers the remote signal, which could affect the ratio itself. Ma et al. [33] also observed uptake in the remote area; however, they did not provide the autoradiography data, including the myocardial target-to-background ratio. Cohen et al. described that ML-10 indicates apoptosis [12]. The TUNEL-positive cells which could be detected in the remote myocardium after infarction [37] might explain the autoradiography data.

A limitation of this study is that [¹⁸F]ML-10 uptake was not limited to cardiac injury. The previous publication on rat myocardial infarct also observed the accumulation in the chest wound after thoracotomy [33]. The murine infarct model's limitations are oropharyngeal intubation, mechanical ventilation, and thoracotomy. In our experiments, we observed a non-specific [¹⁸F]ML-10 uptake reduction over time. While the mice were intubated and mechanically ventilated during surgery, the in vivo PET imaging for 48 h was performed by isoflurane mask narcosis and spontaneously breathing mice. This could explain the apparent reduction in lung uptake. On the other hand, the cardiac [¹⁸F]FDG signal, further enhanced by isoflurane narcosis, remained stable [36]. As expected, a decent [¹⁸F]FDG background signal was observed. Since the cardiac [¹⁸F]FDG uptake was enhanced by isoflurane narcosis, we did not further evaluate the [¹⁸F]FDG activity in the infarct area. Our experimental design of using different experimental groups, instead of longitudinal PET scans, avoided the potential bias of overlapping tracer activity. Using ketamine/xylazine narcosis, which is not influencing the cardiac [¹⁸F]FDG uptake after ischemia [38], would be the accurate strategy to evaluate the overlapping activities.

If this observation can be detected and better discriminated in larger animals or patients, as shown in rats, it remains to be elucidated in future in vivo investigations.

Furthermore, we assessed the cardiac apoptosis histologically by TUNEL staining, which is the most feasible approach to assess cardiac cell apoptosis [39]. It should be considered that TUNEL staining is used to detect DNA strand breaks, while membrane potential and acidification alterations cannot be detected. Therefore, the TUNEL method is limited by detecting both apoptosis and non-specific DNA degradation [39, 40]. Before using the TUNEL staining in our experimental setup, we also evaluated the staining for active caspase-3. We did not detect a signal of active caspase-3 in these early time frames (data not shown). According to the literature, caspase-3 staining is feasible after 24–72 h [41]. Thus, active caspase-3 staining would evaluate the myocardial ischemia at later stages than in our experiments. After transient LAD ligation, TUNEL-positive cells were evident in the infarct and remote area. The reperfusion effect presumably explained by the TUNEL-positive cells decreased in the infarct and remote areas. Literature suggests that reperfusion leads to an enhancement of apoptosis (reviewed in [42]). It is common sense that reperfusion restores the energy required to complete apoptosis and can accelerate the process [40, 43, 44]. Our correlation analysis indicates that the [¹⁸F]ML-10 uptake correlates with the number of TUNEL-positive cells by proper localization. A limitation of this work and the correlation analysis is that only the hearts after PET imaging underwent further histological analysis by TUNEL stain and not the ones after autoradiography. This could potentially explain that the positive trend in the correlations of autoradiography and TUNEL does not reach statistical significance. At the same time, there is a statistically significant correlation between PET and TUNEL staining. Another explanation for the positive trends that do not reach significance is further indicative that apoptosis is not being measured exclusive of other adjacent processes that occur during injury.

ML-10 was primarily described for selective incorporation into apoptotic cells but not into viable or necrotic cells due to specific membrane potential and acidification. The incorporation of ML-10 is ceased upon membrane disruption, thereby distinguishing apoptosis from necrosis [12]. This underlines the temporal dynamics of [¹⁸F]ML-10 after an ischemic heart injury and the decline over time.

This study displays the characteristic of [¹⁸F]ML-10 after transient LAD ligation but also bears several limitations. The surgical procedure led to a severe injury to the chest, and thus the detection of cardiac [¹⁸F]ML-10 PET imaging could have interfered.

Other publications underline this limitation, which was also observed in rat myocardial infarction experiments [33]. Substrate availability by blood flow, hormonal status, blood glucose levels, and insulin sensitivity could potentially interfere with the cardiac [¹⁸F]FDG uptake [45, 46]. Ma et al. showed that using pentobarbital in rats enables accurate infarct localization. In this work, mice under isoflurane narcosis were solely used to localize the defect area by reduced [¹⁸F]FDG uptake. Of note, we cannot fully exclude differences in rats versus mice regarding the FDG uptake pattern in the heart, especially using different narcotic agents. Importantly, our experimental strategy performed the injection and imaging of [¹⁸F]ML-10 first. Thus, there is no influence induced by the latter cardiac [¹⁸F]FDG uptake. In parallel, the search for a robust nuclear apoptosis tracer, the ongoing research on cardiac hypoxia (reviewed in [47]), or hypoxia identifying [¹⁸F]fluoromisonidazole (FMISO) [48] or the evaluation of the oxidative metabolism by [¹¹C]acetate [49, 50] could display other suitable strategies.

This mouse study indicates that [¹⁸F]ML-10 is a novel tracer for cardiac imaging injury after transient LAD ligation in mice. We confirmed the uptake by a multimodal approach, including autoradiography, in vivo small-animal PET imaging, and TUNEL staining. Our results underline that [¹⁸F]ML-10 could provide a novel approach for detecting cardiac injury in an ischemia–reperfusion model. Further investigation in larger animals such as pigs using endovascular interventions, should provide more insight into the applicability of [¹⁸F]ML-10 and better clarify the translational potential.