Three-month old WT male and female CD-1 Swiss mice (n = 15 per sex) were obtained from Janvier (Le Genest Saint Isle, France). Animals were group-housed (six per cage) in individually ventilated cages under a 12 h light/dark cycle in a temperature- and humidity-controlled environment with food and water ad libitum. The animals were acclimatized to the facility at least one week before the start of procedures. All experiments were approved by the Ethical Committee for Animal Testing (ECD 2019-39) at the University of Antwerp (Belgium) and the European Committee Guidelines (decree 2010/63/CEE) for the care and use of animals were followed.

The mice were allocated to five cohorts, each containing three males and three females (Fig. 1). The in vivo PET/CT data was used to assess the activity concentration in the brain, lungs, heart, liver, kidneys, intestines, urinary bladder, muscle (hind leg), and whole body. In addition to the same organs delineated in in vivo scans, the ex vivo data was used to assess tracer distribution in organs that are difficult to delineate on the PET/CT images. The organs or fluids collected for ex vivo assessment were brain, eyes, lungs, heart, liver, gallbladder + bile fluid, spleen, pancreas, stomach, small intestine, large intestine, (the contents for stomach and intestines were removed during dissection) kidneys, urinary bladder, gonads, thyroid, adipose tissue, bone (tibia), muscle (hind leg) and skin (outside layer of upper thigh), blood and urine.

Overview of the experimental design for the [18F]CHDI-650 dosimetry. The study design consisted of an in vivo part where the animals were scanned via PET imaging and a subsequent ex vivo part where the animals were sacrificed, and the organs were harvested. Five cohorts with different termination time points were used, each cohort containing 3 male (M) and 3 female (F) mice, resulting in a total of 30 mice. To acquire in vivo data, the animals underwent either a 10-min static scan at 30 min and 1 h post-injection (p.i.), a 20-min static scan at 4 h and 6 h p.i. or a 2 h dynamic scan. The scans were initiated so the end of the scan corresponded to the respective time point of the animal. Additionally, the animals scanned at 6 h p.i. underwent multiple PET scans at different time points i.e., a dynamic PET scan from 0–2 h post-injection and 20 min static scans at 4 h and 6 h post-injection. At the end the PET scanning for each cohort, the mice were euthanized, and organs dissected for the ex vivo study

[18F]CHDI-650 was synthesized using an automated radiosynthesis module (Trasis AllInOne, Belgium) available at the Antwerp University Hospital, adapting the previously described process [13] to the system [14]. Radiochemical purity was greater than 98% and molar activity (Am) was 185.2 ± 27.7 GBq/μmol (range 157.3–231.9 GBq/μmol) with a decay-corrected radiochemical yield of 18.02 ± 5.82% at end of synthesis.

MicroPET/CT imaging was performed on two Siemens Inveon PET-CT scanners (Siemens Preclinical Solution, USA). Animals were placed on a single PET bed with the full body of the animals in the scanner’s field of view. The animals of the static imaging cohorts were given an intravenous (i.v.) bolus injection in the tail vein and kept separately in a cage for tracer uptake. The animals were anesthesized and placed on the scanner bed at their assigned time point. The animals of the dynamic imaging cohort were anesthesized, catheterized in the tail vein and placed on the scanner bed. For this cohort, the PET scan was initiated simultaneously with the i.v. bolus injection of the tracer. Anesthesia was induced by inhalation of isoflurane (5% for induction, and 1.5–2% for maintenance during preparation and scanning) supplemented with oxygen. Respiration and heart rate of the animal were constantly monitored using the Monitoring Acquisition Module (Minerve, France) during the entire scanning period. The core body temperature of the animals was maintained using a warm air flow. A 10 min 80 kV/500 μA CT scan was performed for attenuation and scatter correction before initiation of both the static and dynamic PET scans. PET data were acquired in list mode. An overview of the animal dosing information for both sexes is shown in Additional file 1: Table S1.

Static PET data were reconstructed into one frame of either 1 × 600 s or 1 × 1200 s depending on the timepoint while dynamic 120 min PET data were reconstructed into 45 frames of increasing length (12 × 10 s, 3 × 20 s, 3 × 30 s, 3 × 60 s, 3 × 150 s, and 21 × 300 s). All images were reconstructed in 8 iterations and 16 subsets using the 3D ordered subset expectation maximization (OSEM 3D) algorithm utilizing a list-mode iterative reconstruction with proprietary spatially variant resolution modeling for quantitative analysis [15]. Corrections for normalization, dead time, and CT-based attenuation were applied. PET image frames were reconstructed using 0.776 × 0.776 × 0.796 mm3 voxels on a 128 × 128 × 159 grid.

For in vivo data, the co-registered PET/CT image was used to delineate the organs for each timepoint with three-dimensional volumes of interest (VOIs) in PMOD 3.6 software (Pmod Technologies, Zurich, Switzerland). The organ activity inside these VOIs was determined and normalized by the injected dose. The data were pooled together (n = 3/sex/timepoint) to form time activity curves (TACs) for each organ.

For the ex vivo part of the study, at the end of the scan, the animals were sacrificed under isoflurane anesthesia, with cervical dislocation. Blood was collected via cardiac puncture; urine was collected after the voiding of the bladder during cervical dislocation. After dissection, the organs were collected, rinsed with a phosphate-buffered saline solution (PBS), weighed and radioactivity was measured using a calibrated gamma counter (Wizard2 2480, Perkin Elmer). For each organ, the counts per minute (CPM) were determined and decay corrected to the time of sacrifice. Based on the calibration curve prepared with [18F]FDG for 18F radiotracers, the CPM values were converted to kBq. The radioactivity level was normalized to the injected dose and the volume of the organ in cubic centimeters (cc) and expressed as the ratio of percentage injected activity (%IA) to cc (%IA/cc). The densities of the organs was based on the reported values in ICRP 110 [16]. The data were pooled together to obtain a TAC for each organ.

The residence time (RT) for the majority of the organs was calculated by fitting an exponential function to the TACs and integrating the function to infinity. The trapezoid rule was used in case no exponential could be fitted to the TAC. For these organs, physical decay of the activity was assumed after the last measured time point. The residence time of the large intestines was assumed to be equally distributed in the left and right colon [17]. Since no input source organs were available for the skin, muscle and fat measurements in OLINDA, these measurements were summed and put into the ‘total body/remainder’ source organ. The residence time of the full body in vivo was subtracted by the residence times of the other organs delineated, the remaining residence time was used as ‘remainder’ during the calculations. The RT for each organ was extrapolated to humans based on the following formula (18):

$$}_}}} = \left[ }}}\left( } \right)}}*\left( }\;}\;}\left( }} \right)} \right)} \right)*\left( }\;}\left( } \right)}}}\;}\;}\left( }} \right)}}} \right)} \right]$$

Human organ masses and body weights were based on models proposed in ICRP publication 89 [19]. The absorbed and total mean effective dose of [18F]CHDI-650 were estimated in humans with the dosimetry programs OLINDA/EXM 2.2 (HERMES Medical Solutions, Stockholm, Sweden) [20] and IDAC-Dose 2.1 [21]. Absorbed organ doses (Gy/MBq) were converted to equivalent organ doses (Sv/MBq) based on the weighting factors presented in ICRP 103 [22].

The in vivo and ex vivo based absorbed dose estimates measured with OLINDA and IDAC were tested for normality with the Shapiro–Wilk test. The Spearman’s rho test was used to determine the correlation between the two dosimetry programs and between the two modalities used for collection of biodistribution data. Statistical analyses were performed with GraphPad Prism (v10) statistical software. Data are represented as mean ± SD unless specified otherwise. All tests were two-tailed, and statistical significance was set at p value < 0.05.