Initially (Wild et al., 2006), Wild was the first to successfully visualize insulinomas in preclinical studies utilizing Ahx as a linker between DTPA and peptide by c-terminal extension of the Lys side chain coupling of exendin-4 (Lys40NH2 modifies Exendin-4), a landmark for GLP-1R molecular visualization. Wicki et al. (2007) later observed a 94% reduction in tumor volume within 8 days after injecting 28 MBq of radioactive peptide under the same molecular tracer. Such positive results raise expectations for driving molecular imaging in the theranostic of insulinoma. Nonetheless, the 111In is costly and has a significant radiation burden on patients. Therefore, an alternative radionuclide for SPECT/CT imaging, 99mTc, appeared in comparison research with 111In at Wild et al. (2010). The synthesized molecular compound that used exendin-4 as a substrate also demonstrated an ideal tumor background ratio and insulinoma sensitivity, making it an ideal choice for GLP-1R molecular imaging.

Regrettably, the above studies were all SPECT/CT imaging of nuclide markers until Wu’s two results (Wu et al., 2011; Wu et al., 2014) on 64Cu-exendin-4 PET/CT molecular imaging of insulinomas opened the window to new horizons of GLP-1R molecular imaging again. The ligands Cys-exendin-4 were respectively added to phosphate buffers containing BaMalSar and Mal₂Sar at pH 7.0, and the coupling products BaMalSar-exendin-4 and Mal2Sar-(exendin-4)2 were purified by HPLC with stirring at room temperature in yields greater than 90% (Figure 3). Not only did they visualize insulinomas using this molecular imaging, but also successfully visualized portal islet transplantation models. However, negative findings could also reveal that the average effective dose of 64Cu was nearly ten times higher than that of 68Ga-exendin-4, which significantly increased the radiation dose and limited its application and translation. It suggests that 68Ga may be a nuclide probe with great potential for molecular imaging of GLP-1R.

Before beginning insulinoma research with 68Ga-exendin-4, several results (Kiesewetter et al., 2012a; Kiesewetter et al., 2012b; Keliher et al., 2012; Wu et al., 2013; Yue et al., 2013; Xu et al., 2015; Li et al., 2020; Murakami et al., 2021a) with 18F indicated that 18F-exendin-4 is critical for developing molecular imaging. Surprisingly, in these 18F experiments, we saw one of the most significant drawbacks of current GLP-1R molecular imaging, namely, high renal uptake, which can lead to missed or misdiagnosed lesions located in the tail of the pancreas next to the kidney. Actually, the high renal radiation burden emerged early when Vegt et al. (2008) and Wild et al. (2010) reduced renal uptake by adding the albumin fragment FRALB, gelofusine, and polyglutamic acid infusion. However, the tedious and complicated preparation constrained subsequent development. In contrast, the 18F-exendin-4 study was performed by synthesizing new GLP-1R derivatives through diverse modifications of different amino acid structural sites of Exendin-4 to obtain a lower renal load developer. We have seen that successfully preparing different modified derivatives, including micron and nanomaterials [FBEM (Kiesewetter et al., 2012a), TCO (Keliher et al., 2012), TTCO (Wu et al., 2013), FPenM (Yue et al., 2013), and PEG (Li et al., 2020; Murakami et al., 2021a)], can reduce renal uptake to a certain extent yet the process is still complicated. It is possible that finding new analogs of GLP-1 with higher homology and a nephroprotective native structure will be a target for future molecular imaging works. Moreover, perhaps the second-generation GLP-1 analog semaglutide will be a new option.

Subsequently, the six preclinical studies (Brom et al., 2010; Sel et al., 2014; Bauman et al., 2015; Kaeppeli et al., 2019; Zhang et al., 2019) of 68Ga-exendin-4 held promise for further clinical translation and laid a solid foundation for its future unassailable position in insulinoma molecular imaging. In addition to the most common DOTA, using different chelators, including NOTA, NODAGA, DO3A, and DFO, has also compensated for the lack of singularity of other nuclide markers in the past. It was the first time in the Kaeppeli study (Kaeppeli et al., 2019) that exendin-4 was modified at position 14, where methionine (Met) was replaced by norleucine (Nle), to improve oxidative stability during labelling. And coupled to NODAGA to form [Nle14-Lys40-(NODAGA)NH2]exendin-4, referred to as Ex4NOD. A comparative study was carried out with Ex4DFO coupled to desferrioxamine DFO (Figure 4 of the two chemical structures). In particular, two comparative studies by Brom et al. (2010) and Wild et al. (2010) have shown that GLP-1R PET/CT molecular imaging has a better sensitivity and tumor background ratio for insulinoma than SPECT/CT.

We have summarized the previous preclinical studies of GLP1-R molecular imaging in insulinoma (Table 1) and discovered that from 111In in the early days of exploration to the now mature 68Ga. Even 89Zr (Bauman et al., 2015), a long half-life nucleophile, has been added to the list. It opens up more options for clinical research to be reviewed. Even so, the molecular probe for GLP-1R is still exendin-4. Such limitations and its high renal load are currently a challenge in this field, and perhaps the integration of new GLP-1R analogs into targeted therapy will become a more worthwhile prospect.

The theranostic of molecular medicine is the latest trend. Furthermore, the success of clinical translation more or less determines the future of molecular imaging of GLP-1R in insulinoma. We have likewise summarized all previous studies on insulinomas (Table 2). Nucleophiles18F, 64Cu, and 89Zr have restricted their clinical application due to some of the limitations mentioned in the previous section. Wild combines GLP-1R molecular imaging with the clinic for the first time (Wild et al., 2008; Christ et al., 2009; Wild et al., 2011), 2 years after its landmark 2006 preclinical study was published. Despite the small number of insulinoma patients included in the two clinical studies, 111In-DTPA-exendin-4 SEPCT/CT successfully detected lesions compared to the weak sensitivity of conventional imaging pairs. It opens new doors for molecular imaging of insulinomas. Subsequent SEPCT/CT molecular imaging investigations of 111In and 99mTc-labelled exendin-4 have followed (Sowa-Staszczak et al., 2013; Antwi et al., 2015; Senica et al., 2020). However, the most representative of these undoubtedly belongs to two large sample-size clinical trials by Christ in 2013 (Christ et al., 2013) and Sowa-Staszczak in 2016 (Sowa-Staszczak et al., 2016). Christ performed 111In-DTPA-exendin-4 molecular imaging in 20 of the 30 patients with EHH who were included to detect insulinomas and successfully detected 19 with a sensitivity of 95%, in contrast to 44% with conventional imaging. Such results are consistent with Sowa-Staszczak. The latter used 99mTC-HYNIC-exendin-4 in all 18 patients with insulinoma, significantly higher than the detection rate of 6/18 (33.3%) with conventional imaging. Such promising results are encouraging, and it indicates that GLP1-R molecular imaging is gradually replacing conventional imaging as the non-invasive modality of choice for the accurate preoperative localization of insulinoma.

Similarly, PET/CT studies, a representative of molecular imaging, have also revolved around 68Ga-exendin-4. Different chelator-targeted molecular probes have been successfully reported in several cases. Here, our team (Qiu et al., 2023) also successfully used 68Ga-DOTA-exendin-4 PET/CT to visualize a double lesion insulinoma in the pancreas’s tail, one of which was not detected on MRI (Figure 5). The clinical reports of Ga-exendin-4 are among the most diverse in terms of the chelating agents employed, and it has been an essential part of its success in clinical translation. Whether DOTA (Sood et al., 2019; Michalski et al., 2020; Das et al., 2021; Zhang et al., 2021), DO3A (Eriksson et al., 2014; Velikyan et al., 2017), or NOTA (Luo et al., 2015; Luo et al., 2016a; Warren et al., 2020) is reported on a case-by-case basis, the most convincing report and one that has driven the recognition of 68Ga-GLP-1R molecular imaging should be the publication of Luo’s seminal study in 2017 (Luo et al., 2016b). He prepared 68Ga-NOTA-MAL-cys40-exendin-4 by coupling NOTA-MAL with the ligand Cys40-exendin-4 and then successfully targeting 68Ga to the peptide at 100°C for 15 min. The results revealed that 68Ga-NOTA-exendin-4 PET/CT successfully detected lesions in 42 of 43 insulinoma patients, with a sensitivity of 97.7%, markedly higher than the 19.5% of 99mTc-HYNIC-TOC SPECT/CT. It is the first time that the two types of molecular imaging have been compared again in a clinical report, and PET/CT remains the obvious choice. Another study (Boss et al., 2019), also widely expected, showed a sensitivity of 85.3% for 68Ga-NODAGA-exendin-4 PET, greater than SSTR PET (54.5%) and conventional imaging (66.7%). At this point GLP-1R molecular imaging has entered the era of 68Ga-exendin-4, and their meaning here cannot be overstated.

Unfortunately, it has been greatly appreciated in molecular imaging as the most popular combination of 68Ga/177Lu for theranostics. The research is still vacant in insulinoma. As early as 2016 (Espes et al., 2016) [177Lu]-DO3A-VS-Cys40-Exendin-4 was used for the first time to successfully scan transplanted islets and quantify cell mass in an islet transplantation model. In 2019, Guleria et al. (2019) produced a molecule [177Lu]Lu-DOTA-Ahx-Lys40-Exendin-4 with high radiochemical purity. The biodistribution investigation revealed that the chemical initially accumulated in pancreatic tissue and was predominantly excreted by the kidney. Nevertheless, its substantial absorption in the kidney restricts its potential for further translation. The failure of the investigations above based on 177Lu-labeled exendin-4 to complete GLP-1R molecular imaging and targeted therapy in insulinoma models will present an emerging opportunity for our long-term research. It will also be the biggest obstacle to the current move towards GLP-1R molecular imaging into theranostic, and we need to pool our efforts to overcome this challenge.

SSTR is highly expressed in malignant tumors but rarely in GLP-1R. The results of Bongetti’s research (Bongetti et al., 2018) gave us a great warning. Although benign insulinoma accounts for 90% of all insulinomas, 10% are still malignant. Therefore, for patients with negative imaging results who remain highly suspected of insulinoma, we should be alert to whether they are malignant. At this time, we need to take the receptor imaging technology of SSTR to identify them, such as 68Ga-Dotatate-exendin-4. Do not miss or misdiagnose the patients with negative GLP-1R imaging. The study of Korner et al. (2007), other than insulinoma, illustrated that GLP-1R is also high expression in gastrinoma, MTC, 60% of pheochromocytoma, and gastrinoma. Given this, is molecular imaging targeting GLP-1 analogs feasible in these tumors? A study has since given partial answers.

Pach et al. (2013) performed [Lys (Sood et al., 2019) (Ahx-HYNIC-(99m)Tc/EDDA)NH₂]-Exendin-4 SPECT/CT imaging in four enrolled MTC patients for the first time. These results displayed that GLP-1R molecular imaging is not exclusive to insulinoma and plays a vital role in other diseases. Clinical experience has pointed to these tumors as high in receptor expression, and they all have a relatively low clinical incidence. Even with pheochromocytoma, which has a relatively high prevalence, the more established SSTR molecular imaging agent DOTATATE is now available. Hence, these factors have directly or indirectly hindered the exploration of GLP-1R molecular imaging in tumor diseases other than insulinoma. It is both a challenge and an opportunity, so searching for additional molecular probes may present promising results in these oncological diseases.

Currently, the most commonly used methods for the assessment of surviving myocardium after myocardial infarction in clinical practice include EHCO, CAG, CTCA, and MRI. In the past, these methods were used to assess myocardial survival after infarction from the level of myocardial dynamics and perfusion, but they were unable to analyze the metabolic level of the surviving myocardium from the molecular metabolism point of view in a more in-depth manner. The FDG currently used in PET/CT is taken up by myocardium with metabolic activity but not by necrotic myocardium. Thus, myocardial metabolism can be evaluated by this mechanism. Also PET dynamic myocardial perfusion imaging agents combined with 13N-NH3 provide more accurate judgment (Kim et al., 2015). In the cardiovascular, GLP-1R can improve cardiovascular function and reduce inflammation (Laviola et al., 2012). Moreover, it is involved in the protective effect of myocardial cells after myocardial ischemia. So, can infarcted myocardial cells be targeted by nuclide-labeled GLP-1R molecular imaging? The publication of two critical studies has also delivered on the promise of GLP-1R imaging in cardiovascular disease.

In the myocardial ischemia/reperfusion (MI/R) study (Gao et al., 2012), 18F-FBEM-Cys40-exendin-4 could monitor the dynamic changes of GLP-1R upregulation in the myocardial infarction/ischemia area. Meanwhile, the uptake changes observed by PET were consistent with the changes in receptor expression by Western blot analysis and histochemistry. Its protective mechanism and analogs in myocardial cells with ischemia injury are now better-understood thanks to this report, which is the first to image GLP-1R in myocardial cells that have undergone MI/R. It marks the triumphant entry of molecular imaging into the cardiovascular field. Another blockbuster study (Stahle et al., 2020) was the first application of the 68Ga-NODAGA-exendin-4 molecular probe in a myocardial infarction model. The encouraging conclusion indicated that the uptake of 68Ga-NODAGA-exendin-4 in the infarcted area was increased by 8.6 times compared with the sham operation area. The uptake of the imaging agent was correlated with the number of CD68-positive macrophages in the infarcted area and α-smooth muscle actin staining in the distal myocardium. Molecular imaging can dynamically monitor the protective effect of GLP-1R on ischemic cardiomyocytes, which brings a new auxiliary method for evaluating the therapeutic effect of myocardial ischemia and myocardial infarction.

And compared with FDG PET/CT, GLP1R molecular imaging is more friendly to patients with type 2 diabetes. It allows for a more visual assessment of GLP1R metabolic expression levels in the surviving myocardium at the molecular metabolic level. Therefore, in the future, it could complement the assessment of inflammation in the molecular metabolism of surviving post-infarction cardiomyocytes. And it can be used together with other routine tests to provide the evaluation value of surviving post-infarction cardiomyocytes from different perspectives including myocardial dynamics, perfusion and molecular metabolism. However, because only a few preclinical studies of GLP1R molecular imaging have been performed, there is a need for a larger future.

GLP-1/GLP-1R in the endocrine field has been accompanied by the introduction of semaglutide, which has become a principal receptor target. The initial release of the SUSTAIN and PIONEER phase III clinical trials established its landmark importance in treating type 2 diabetes mellitus (T2DM) and obesity (Meier, 2021). More recently, research on this new GLP-1 analog has gradually expanded into other areas, including diabetes, chronic kidney disease, cardiovascular disease, NASH, and Alzheimer’s disease (Knudsen and Lau, 2019). However, no molecular imaging studies with semaglutide have been performed, and the endocrine field is still dependent on exendin-4 and its derivatives for molecular imaging. This section reviews the progress and emerging opportunities for endocrine-related GLP-1R molecular imaging.

The expression of GLP-1R was initially detected in the nucleus tractus solitarius and hypothalamus. As the research progressed, so did other brain parts, such as the frontal cortex, hippocampus, amygdala, and substantia nigra. In addition to being primarily distributed in the pancreatic islet cell, it can appear selectively in the central nervous system (Jensen et al., 2018; Mansur et al., 2019; Gabery et al., 2020). The theranostic of central nervous system diseases targeting GLP-1R has been gradually developed. It is also a key area of focus for molecular medicine at this stage.

Molecular imaging has already become a crucial clinical tool in the diagnosis of AD and PD. For instance, the tracer 11C-PIB, as is now widely used for imaging AD age spots (Klunk et al., 2004). However, 11C has a half-life of only 20 min, limiting its use in routine examinations. 18F has a longer half-life. Therefore 18F-based PET imaging began a long quest in the disease assessment of AD. From the initial low specificity of FDG PET for pathological changes in AD to the drawbacks that amyloid PET imaging does not correlate with the severity of the disease. The unfolding of the Tau PET examination has provided significant value in the assessment of early pathological changes and disease severity in AD. And later with the 18F-florbetapir have been employed in AD as beta-amyloid (Aβ) imaging agents (Clark et al., 2012), while 18F-FDOPA has been used in PD diagnosis (Brooks, 2010). Two previous studies in PD models (Bertilsson et al., 2008; Isacson et al., 2011) have confirmed that exendin-4 can restore the nerve in the subventricular zone and hippocampus, thereby improving mice’s cognitive, motor, and memory abilities. Bu et al. (2021) still bound exendin-4 with AAV-9-A53T-α-synuclein in the PD model and observed that exendin-4 treatment attenuated the pathological accumulation of α-synuclein. In recent years, additional studies (Salameh et al., 2020; Lv et al., 2021; Zhang et al., 2022) have found that incretin receptor agonists (Exendin-4 and DA-JC4) and GLP-1/GIP dual receptor agonists DA5-CH can successfully cross the blood-brain barrier and reach the pathological area of PD, which is expected to become an ideal therapeutic drug. Exendin-4 has shown excellent results in improving cognitive function in Parkinson’s disease. The mechanism and efficacy of a novel GLP-1R analog liraglutide in PD are also evaluated (Kim et al., 2017; Mah et al., 2022; Chen et al., 2023). GLP-1R targets may usher in a new chapter in the theranostic of central nervous system diseases.

For the first time, Wang et al. (2018) illustrated the effect of age differences on GLP-1R distribution in the brain by synthesize [18F]AlF-NOTA-MAL-Cys39-exendin-4, the peptide was modified by coupling Cys39-exendin-4 to NOTA-MAL and was successfully synthesized using 18F-targeted labeling of the peptide(Figure 6). Its expression in the brains of aged mice was elevated reduced. The discovery of this molecular probe creates a better prospect for applying GLP-1R molecular imaging to neurodegenerative diseases such as PD and AD. In 2020, Liu et al. (2020) published the investigation via PET imaging in Alzheimer’s disease (AD) model, different derivatives with 18F labeling followed by microPET. [18F]FBEM-exendin-4 is more likely to cross the blood-brain barrier due to its better lipid solubility than [18F]AIF-exendin-4 based on previous findings. [18F]FBEM-Cys39-exendin-4 was compared with three other 18F-labeled derivatives of PET imaging. This tracer successfully evaluated the decrease of total GLP-1R expression in the AD model. It is the first report of GLP-1R molecular imaging in patients with neurodegenerative disease, which opens a good direction for the future in this still-unknown field. Deden et al. (2021) also used a 68Ga-labeled tracer to detect the difference in receptor intracranial distribution of obese patients compared with ordinary people. Unfortunately, 68Ga-NODAGA-exendin-4 PET was unsuitable for GLP-1R distribution imaging in the brain of obese patients.

The research on GLP-1R molecular imaging in neurodegenerative diseases and other central nervous system diseases has not achieved more extraordinary breakthroughs, limited to preclinical studies and related studies of blood-brain barrier penetration. There remains a significant gap in the comparative study of GLP-1R and neurodegenerative mechanisms and the improvement of symptoms in clinical patients. The experimental investigation of combining GLP-1R molecular imaging with other related nervous system diseases is an emerging opportunity in that uncharted territory. It could drive the neurology theranostic.