Novel CYP11B-ligand [123/131I]IMAZA as promising theranostic tool for adrenocortical tumors: comprehensive preclinical characterization and first clinical experience

Purpose Adrenal tumors represent a diagnostic and therapeutic challenge. Promising results have been obtained through targeting the cytochrome P450 enzymes CYP11B1 and CYP11B2 for molecular imaging, and [123/131I]iodometomidate ([123/131I]IMTO) has even been successfully introduced as a theranostic agent. As this radiopharmaceutical shows rapid metabolic inactivation, we aimed at developing new improved tracers. Methods Several IMTO derivatives were newly designed by replacing the unstable methyl ester by different carboxylic esters or amides. The inhibition of aldosterone and cortisol synthesis was tested in different adrenocortical cell lines. The corresponding radiolabeled compounds were assessed regarding their stability, in vitro cell uptake, in vivo biodistribution in mice, and their binding specificity to cryosections of human adrenocortical and non-adrenocortical tissue. Furthermore, a first investigation was performed in patients with known metastatic adrenal cancer using both [123I]IMTO and the most promising compound (R)-1-[1-(4-[123I]iodophenyl)ethyl]-1H-imidazole-5-carboxylic acid azetidinylamide ([123I]IMAZA) for scintigraphy. Subsequently, a first endoradiotherapy with [131I]IMAZA in one of these patients was performed. Results We identified three analogues to IMTO with high-affinity binding to the target enzymes and comparable or higher metabolic stability and very high and specific accumulation in adrenocortical cells in vitro and in vivo. Labeled IMAZA exhibited superior pharmacokinetic and imaging properties compared to IMTO in mice and 3 patients, too. An endoradiotherapy with [131I]IMAZA induced a 21-month progression-free interval in a patient with rapidly progressing ACC prior this therapy. Conclusion We developed the new radiopharmaceutical [123/131I]IMAZA with superior properties compared to the reference compound IMTO and promising first experiences in humans. Supplementary Information The online version contains supplementary material available at 10.1007/s00259-021-05477-y.


Introduction
Adrenal tumors belong to the most frequent human neoplasias with an age-dependent prevalence of 2-10% [1][2][3]. The spectrum of potential differential diagnoses largely varies; it includes non-functioning adrenocortical adenomas, representing the most frequent tumor entity, but also malignant adrenal primary tumors or metastases from other malignancies [2]. While non-functioning adrenocortical adenomas do not require further treatment, malignant or hormone-producing tumors are usually removed by surgery. Adrenocortical carcinoma (ACC) is a malignancy with an overall poor prognosis and non-satisfactory treatment in advanced disease [4][5][6][7][8][9]. The only approved drug in advanced ACC is mitotane, being associated with significant adverse effects and a response rate of maximum 24% [10,11]. Early diagnosis and surgical resection are the only curative option for these patients. However, a significant number of adrenal incidentalomas cannot be completely characterized by conventional imaging methods like computed tomography (CT) or magnetic resonance imaging (MRI) [12]. Functional imaging providing more information on the molecular background of a tumor may be of help to fill in this diagnostic gap. The radiotracers [ 11 Fig. 1) selectively bind to adrenocortical CYP11B enzymes and have been successfully used in humans for specific molecular imaging of adrenocortical tissue [13][14][15][16][17][18][19][20][21][22][23].
High uptake of [ 123 I]IMTO was reported not only for primary tumors but also for metastatic lesions in ACC. In a recent series of 11 patients with advanced ACC, we demonstrated a therapeutic potential of treatment with [ 131 I] IMTO, resulting in a lasting stabilization of the disease in the majority of patients, with only few adverse effects [24].
Although [ 123 I]IMTO represents a suitable tool for adrenal imaging and therapy, studies of the metabolic stability revealed rapid degradation in patients by esterases. IMTO derivatives resistant to rapid degradation but still maintaining high avidity to adrenal CYP11B enzymes could allow lower radiation exposure during imaging and improved imaging quality due to lower background activity and higher effective doses in targeted treatment of ACC lesions. Therefore, we aimed at establishing a new theranostic tool targeting adrenocortical tumors by development of IMTO derivatives with improved properties regarding target binding, tracer stability, and target to background ratios.

Chemistry and radiochemistry
For the preparation of the derivatives, IMTO (compound 1) was saponificated to the corresponding acid 2 and converted into esters 3 and 4 and the carboxylic amide 5 with the help of coupling agents. The trimethylstannyl precursors were synthesized from the corresponding selected iodinated arenes. Radioiododestannylation was chosen for labeling of the stannylated precursors and delivered each tracer in high yields (> 90%). Experimental details of chemistry and radiochemistry of the most important compounds IMTO, IMAZA, 3, and 4 are presented in the supplement.

Evaluation of CYP11B1 and CYP11B2 inhibition in Y1-Cyp11B1 and Y1-Cyp11B2
The human adrenocortical cell line NCI-H295 and the murine adrenocortical cancer cell line Y1 transfected with either human CYP11B1 or CYP11B2 were used for determination of IC 50 values as already described [17].

Analysis of metabolic inactivation by liver microsomes in vitro
Metabolic degradation was determined by co-incubation of the tracers with human liver microsomes (Sigma, M0317). Experiments were conducted in triplicates. Incubation mixtures consisted of liver microsomes (0.2 mg/ml), the respective 125 I-labeled compound (0.

Cell uptake studies
NCI-H295 cells (n = 250,000 per vial) were incubated with 0.1 MBq of the respective 125 I-labeled tracer at 37 °C, 5% CO 2 . For the blocking experiment, the cells were co-incubated with etomidate. Aliquots were taken and immediately placed into a 4 °C ice bath. After 5 min, the solution was centrifuged and rinsed twice with PBS/0.5% Tween 80, and the collected rinse solution was measured in a gamma counter. Assays were performed in triplicates. The intracellular distribution of the respective 125 I-labeled inhibitors and the specific binding to mitochondrial membranes were verified using a commercially available kit (Qiagen Qproteome Mitochondria isolation).

Autoradiographic binding experiments to human tissue cryosections
Human tissues were quick frozen in liquid nitrogen to block further biological processes including protein degradation and stored at -80 °C. Frozen samples from adrenocortical carcinoma (ACC; n = 3), aldosterone-producing adenoma (APA; n = 2), and cortisol-producing adenoma (CPA; n = 2), normal adrenal, kidney, and liver were sectioned into 20-µm-thick tissue slices, thaw-mounted onto Superfrost slides (PAP PEN, Zymed® Laboratories) and stored at -80 °C until processing. For the experimental set up, the tissue slices were pre-incubated for 5 min at room temperature with buffer solution (pH 7.4, 50 mM Tris). Afterwards, the slices were directly incubated with TRIS HCl buffer solution containing 2% BSA. After 1-h blocking, slides were treated with 0.1 MBq of the different compounds. As controls, slides were incubated with 0.1 MBq of the compounds in 10 µM of the non-radioactive compound to assess binding specificity. In the following, slides were washed twice with 50 mM Tris HCl buffer solution (pH 7.4). For exposure, samples were placed on Phosphor Imager plates for 30 min in dedicated lead-shielded cassettes. Autoradiographic images were analyzed with a CR35BIO Image plate scanner, and data analysis was performed with Amide software version 0.9.0.

Animal experiments
Six-to 8-week old male CD-1 mice received 0.1 MBq [ 125 I] IMTO [17] or the 125 I-labeled IMTO-analogues as intravenous injection in the tail vein. At predefined time points, mice were sacrificed (min. n = 5 per time point). Blood and urine were collected, and several organs were excised and weighed, and tracer uptake was determined by measurement in a γ-counter. The absorbed dose to the blood was estimated from assessments of the activity kinetics in total body and blood up to 48 h when the mean activity concentrations had dropped to 5% of the values at 5 min after the administration in the total body and to 1% in the blood. During therapy, the total body activity retention was measured with a survey meter in the ceiling over the patient's bed. Gamma camera whole body scans accompanied by blood sampling were acquired starting 4 days post-administration after the first and 2 days after the second therapy until dismissal. SPECT/CT imaging was performed after 4 and 5 days, respectively. Tumor evaluation was performed by [ 18 F]FDG PET/CT at baseline and 2, 5, 8, 11, 14, and 21 months after administration of the first treatment dose. Subsequently, the patient declined further follow-up imaging or additional therapeutic interventions.

Statistical analysis
IC 50 values were calculated from all values using the program GraphPad Prism5®. All results are expressed as mean ± standard deviation.

Effects of compounds on secretion of cortisol and aldosterone in vitro
The majority of the new compounds inhibited cortisol and aldosterone secretion in a dose-dependent manner both NCI-H295 cells and Y1 cells transfected with either human CYP11B1 or CYP11B2, indicating target enzyme binding. The IC 50 values for inhibition of cortisol and aldosterone in Y1 cells are displayed in Table 1.

Metabolic stability of radiolabeled compounds in vitro
The metabolic stability of [ 125 I]IMTO and three new tracers was determined by radio-HPLC; results are listed in

Effects of compounds on steroidogenesis in NCI-H295 cells
The compounds 3-tetrahydrofurylester 3 and 2,2-difluoroethylester 4 and IMAZA exhibited both strong enzyme inhibition and at least comparable metabolic stability to IMTO, so they were additionally investigated in NCI H295 cells. An accumulation of the upstream steroids 17-OH-progesterone and androstenedione was observed at lower concentrations of the compounds (0.1-10 nM) due to the block of both CYP11B enzymes. At higher concentrations, the production of the investigated hormones was completely stopped, most likely due to additional inhibition of side chain cleavage enzyme (CYP11A1), similar to the known effect of metomidate (Fig. 2).

Cellular uptake of selected radiolabeled compounds
The intracellular uptake in human adrenocortical NCI-H295 cells was analyzed for [ 125 (Fig. 3a).
To assess the specificity of all compounds, uptake experiments were furthermore performed after co-incubation with etomidate, a known inhibitor of CYP11B1 and CYP11B2. With the lowest concentration of 1 µM etomidate, the   (Fig. 3b).

Tracer binding to cellular subfractions
As both target enzymes are localized at the inner mitochondrial membrane, cellular subfractionation experiments were performed using human NCI-H295 cells. As shown in Fig. 3c (Fig. 3c). The purity of each fraction was subsequently confirmed by Western blot. was investigated in male CD-1 mice. As previously described for [ 125 I]IMTO, a fast and specific uptake in the adrenal tissue could be observed, which was followed by a rapid clearance within 240 min (Fig. 4). Also, for the other compounds, a specific uptake in the adrenal gland was observed. [ 125 I]IMAZA not only showed an outstandingly high uptake in the adrenal gland, but also demonstrated extremely low uptake values in the non-target organs ( Fig. 4 and Supplementary Table 1).

Autoradiographic experiments
Based on the results from in vitro screening, [ 125 I]IMAZA was found to be the most promising compound and was further tested in binding experiments to human tissue sections. Autoradiographic visualization of [ 125 I]IMAZA showed very strong and specific binding to ACC, APA, and CPA tissue, but no binding was detected for the control tissue (liver, n = 2; kidney, n = 1; Fig. 5). The signal was distinctively stronger for [ 125 I]IMAZA than for the reference compound

[ 123 I]IMTO and [ 123 I]IMAZA in three patients with advanced ACC
After the administration of [ 123 I]IMTO or [ 123 I]IMAZA, renal clearance was fast for both compounds but more rapid for [ 123 I]IMAZA. Apart from low activity retention in the liver and the bowel in the early phases, only blood pool activity was visible in healthy tissues in the scintigraphic images. In particular, no specific binding or persistent accumulation was apparent in the red bone marrow and the kidneys. The total body residence time was 2.2 ± 0.2 times higher for [ 123 I]IMTO than for [ 123 I]IMAZA; the residence time per liter of whole blood was 3.6 ± 1.5 times higher ( Supplementary Fig. 2). The plasma activity concentration was analyzed in patient #1 whose hematocrit was 35%. At 2 min after injection of [ 123 I]IMAZA, 3.8% of the administered activity was found per liter of plasma, which corresponds to a distribution volume of 26 L. The activity concentration was 1.16 times higher in plasma than in the whole blood. These data are consistent with the assumption that [ 123 I]IMAZA readily passes the membrane of red blood cells (plasma to whole blood activity concentration ratio expected from hematocrit, 1.13) and distributes to the rapidly exchangeable fraction of the total body water. After injection of [ 123 I]IMTO, activity concentration in plasma was 1.16 times higher than in whole blood ( Supplementary  Fig. 2).
Uptake in large tumor lesions was visible already at 10 min after tracer injection with both [ 123 I]IMTO and [ 123 I]IMAZA. Specific tumor uptake was higher, and background activity was lower for [ 123 I]IMAZA compared to [ 123 I]IMTO ( Fig. 6 and Supplementary  Fig. 3). In 5 large lesions reliably measurable for both tracers, tumor uptake in the SPECT/CT imaging at 5.7 ± 0.1 h after the administration was a factor of 2.0 (range 1.5 to 2.8) higher for [ 123 I]IMAZA, and the uptake to background ratio was higher by a factor of 4.6 (range 3.2 to 6.8). For both compounds, the effective half-life in tumor tissue was only slightly shorter than the physical half-life of I-123. The biological half-lives were too long to be derived reliably from the 123 I data.
The residence time of the activity in the whole blood was shorter for [ 131   Faint uptake in the liver and the renal pelvises was seen in the planar scintigraphic images up to 4 h. An absorbed dose of well below 0.2 Gy/GBq was estimated for both the liver and the kidney. In the later scans, the activity was only detected in the tumors except for a minor residual activity in the intestine, most likely due to additional biliary excretion.
The estimates for the mean absorbed tumor doses actually administered in the first therapy with 28.2 GBq [ 131 I]IMAZA were 133 Gy for the recurrent primary and 117-265 Gy in 3 evaluable metastatic lesions (Fig. 6e). The absorbed dose to the blood was 1.1 Gy. In a second therapy course with 30.5 GBq [ 123 I]IMAZA, the absorbed doses were 93 Gy to the primary and 145-210 Gy to the metastatic lesions. The absorbed dose to the blood was higher in the second treatment (1.7 Gy) due to a loss of renal clearance. The metastases known from the [ 123 I]IMAZA scans show comparable high uptake in the dosimetric and posttherapeutic [ 131 I]IMAZA scans; see Supplementary Fig. 4.
In our patient, we found a continuous decrease of metabolic activity and size of the tumor lesions over 8 months with a progression-free survival of 21 months and an overall survival of 30 months after initial endoradiotherapy.

Discussion
Our article describes the successful development of a novel theranostic tool for adrenal tumors. In order to improve diagnostic imaging with [ 123 I]IMTO and endoradiotherapy with [ 131 I]IMTO, we started with the hypothesis that both and 240 min after tracer injection. Results are expressed as percentage of injected dose per gram organ (%ID/g). Exact numbers on uptake in the adrenal as %ID/g for the compounds are given in the supplemental data higher avidity to the target enzymes and higher metabolic stability would result in an increased and longer lasting specific binding to target tissue and improved target to background ratios.
Starting from IMTO, a total of 50 new derivatives were synthesized and characterized. Of the most promising three new substances, we synthesized the corresponding trimethylstannyl precursors and used them to establish the radiosyntheses of the tracers. The chemistry and radiochemistry of the remaining new derivatives will be published elsewhere.
Cortisol and aldosterone synthesis were potently inhibited by the small aliphatic esters 3 and 4 and the small carboxylic acid amide IMAZA indicating high-affinity binding of these compounds to the respective enzymes. Considering the IC 50 values for inhibition of cortisol and aldosterone secretion in the low nanomolar range up to only 0.3 nM, even higher affinity to the CYP11B enzymes compared to IMTO may be assumed for two agents.
In vitro analysis of metabolic inactivation after preincubation with human liver microsomes revealed only three compounds with comparable (3,4) or higher (IMAZA) stability than IMTO. It is noteworthy that these three compounds have comparable (3,4) or lower (IMAZA) lipophilicity estimated from HPLC retention times. This fits well to the observation that the binding sites of metabolizing enzymes are generally lipophilic in nature [25]. Apart from the lower lipophilicity, the higher stability can be explained by the carboxylic amide as functional group and by the cyclic structure of IMAZA [26,27].
The target enzymes CYP11B1 and CYP11B2 are localized at the mitochondrial inner membrane, so penetrance of the cell membranes is a prerequisite for target binding. Cellular subfractionation further revealed strong uptake in the mitochondrial fraction. Both cell uptake and accumulation in the mitochondrial subfraction could be blocked by co-incubation with etomidate indicating specific CYP11B binding (Fig. 3b).
In biodistribution experiments, all radiolabeled compounds showed highly specific binding to adrenal tissue (Fig. 4B) In direct comparison, [ 123 I]IMAZA provided much higher tumor to background ratios than [ 123 I]IMTO, and a rapid clearance from the non-target tissue was observed (Fig. 6). Unbound [ 123 I]IMAZA was completely eliminated from the non-target tissue after 24 h. These specific properties might translate into a better diagnostic sensitivity. Accordingly, in our patient, several lesions were visualized by [ 123 I]IMAZA but could not be detected by [ 123 I]IMTO imaging (Fig. 6a-d).
The rapid clearance further results in a lower dose to both the whole body and the red bone marrow which is the critical dose-limiting organ.  [24]. They were furthermore reached after administration of 28.2 GBq, which is only half of the maximum possible activity. The long-term disease stabilization of 21 months and survival of 30 months after a rapid increase of the tumor masses before first treatment with [ 131 I]IMAZA demonstrate the treatment efficacy of our new radiopharmaceutical. Author contribution The project was initiated and supervised by AS and SH. Chemical and radiochemical synthesis was performed by AS and MS; preclinical evaluation including animal experiments was performed by BH, LN, and DM. PEH, CB, KH MF, AKB, and SH were responsible for patient care, imaging, and endoradiotherapy and HH for dosimetry. The manuscript was written by BH, AS, and SH and was read and approved by all authors.