Introduction
Positron emission tomography (PET) is well-established in oncologic radiological workflows to detect and monitor disease progression. The majority of investigations use the metabolic tracer
18F-fluorodeoxyglucose (
18F-FDG). However, there is an increasing use of peptide ligands with radiometals for specific indications to delineate molecularly specific disease types. Replacement of positron emitters with therapeutic isotopes can then be used to localize cytotoxic treatments to sites of confirmed malignancy. This paradigm of radionuclide-based theranostics has attracted a great deal of research, clinical and pharmaceutical interest [
1]. In particular
68Ga (
t1/2 = 68 min,
Emean (
β+) = 830 keV (89%)) and
177Lu (
t1/2 = 6.72 d,
Emean(
β−) = 134 keV)-labeled somatostatin receptor peptides and prostate-specific membrane antigen inhibitors have been approved for PET imaging and treatment of neuroendocrine tumor and metastatic castration-resistant prostate cancer, respectively.
Among the investigated theranostic pairs [
2],
44Sc/
47Sc is well-suited for targeted in vivo PET and beta-particle treatment, respectively [
3,
4].
44Sc has a suitable half-life of 4.04 h for centralized radiopharmaceutical production along with highly abundant positron decay (
Emean (
β+) = 632 keV (94%)) [
5,
6].
47Sc emits a low-energy
β− particle (
Emean (
β−) = 162 keV) similar to
177Lu, with the potential for treating lesions with a half-life of 80.4 h. This is well-suited to the relatively fast pharmacokinetic profiles of small peptides [
7].
44Sc can be obtained via either direction irradiation of natural or enriched calcium [
8] or the decay of
44Ti (
t1/2 = 60.6 ± 1.3 years) [
4]. The potentially long utility of the
44Ti/
44Sc generator system has many advantageous characteristics. It allows for daily elution over a long period of time, providing an attractive alternative for PET facilities that lack in-house cyclotron capabilities. The capability of
44Sc-labeled molecules have begun to be investigated in preclinical research. Multiple cancer xenografts models have been imaged, and several
44Sc-ligands have recently undergone initial clinical evaluation including [
44Sc]-PSMA617 for imaging patients with metastatic prostate cancer [
9,
10]. These in vivo studies have demonstrated that
44Sc-labeled ligands provide high contrast for disease delineation in pre-clinical xenografts and clinical patient studies.
In order to harness the potential for this isotope for theranostics, greater availability of the isotope and improved understanding of in vivo stability and pharmacodynamics is required [
11‐
13]. In this work, we sought to produce and better understand how
44Sc is excreted in vivo, and to what degree this will impact the radiation burden in research and clinical use. To address these questions, we have engineered a clinical-scale
44Ti/
44Sc generator using ZR resin [
14]. Here, we use a reversible-flow modular column design with a disposable cartridge to recover any
44Ti breakthrough and have performed PET imaging and kinetic biodistribution studies with the eluted
44Sc material. These studies provide a comprehensive evaluation of the generator and produced material, including human dosimetry estimates for more widespread clinical use.
Material and methods
All chemicals were obtained from commercial sources and were used without further purification. 44Ti/44Sc solution (111.0 MBq, 421.8 MBq/mg titanium) was obtained from Brookhaven National Laboratory, Department of Energy. Both ZR resin and 0.43-mL ZR resin cartridges were obtained from TRISKEM International. Radioactivity amounts of 44Sc were measured with a dose calibrator (CAPINTEC, CRC-15R) or a 2480 WIZARD2 automatic γ-counter (PerkinElmer). Radiochemical purity was analyzed with high-purity germanium gamma ray detector (HPGe, ORTEC, GEM-50195-S), and spectral acquisitions were acquired and analyzed by Gamma-Vision Software (version 8.0, Ametek). PEEK columns were obtained from VICI precision sampling, Inc for assembly of the ZR resin column. Deionized water (18.2 MΩcm, Rephile) and 99.999% trace-metal HCl (37 wt% in H2O) were used for preparation of 44Sc elution. Female Swiss Webster mice (6–8 weeks from Charlies River Laboratories) were purchased for in vivo pharmacokinetic studies. All radioactive material handling and animal experimentation were conducted in compliance with institutional regulations and approved by Environmental Health and Safety Radioactive Materials protocol #1169-01 and Institutional Animal Care and Use Committee protocol #22-0023.
Design and assembly of 44Ti/44Sc generator
To construct the primary column approximately 200 mg of dry ZR resin was loaded in the PEEK column (50 × 4.0 mm), and the assembled ZR resin column was pre-conditioned with 3 × 2 mL of 6.0 M HCl. 44Ti/44Sc (101.2 MBq in 1.91 mL 6.0 M HCl) mixture. The initial washout solution was reloaded into the column, twice. After loading, two equivalently sized PEEK columns (pre-conditioned with 2 mL of 0.05 M HCl were attached to each end of the primary column to assemble the 44Ti/44Sc generator. 44Sc was eluted from the 44Ti/44Sc generator with 4 mL of 0.05 M HCl (flow rate:1 mL/min via syringe pump), and the elution profile was monitored in situ with a radiodetector (γ-RAM, IN/US) and recorded with Laura software (Lablogic). The radiochemical purity of the collected 44Sc was measured with HPGe and γ-counter immediately after elution, and 3 days later. Measurements with the γ-counter used an energy window of 430–580 keV for 44Sc, and 50–230 keV for 44Ti. Since the presence of 44Sc may significantly influence the measurement accuracy of 44Ti, aliquots of the elutions were stored for several days to afford time for 44Sc decay in order to calculate the ratio of 44Sc/44Ti in the eluted solution.
Preparation of free and chelated 44Sc-NODAGA or Citrate for in vivo evaluation
After approximately 74 MBq of 44Sc was eluted into a vial with 4 mL of 0.05 M HCl solution, the 44Sc solution was adjusted to pH 6–7 with 2 M Na2CO3 (or 4 M NaOH) to prepare the 44ScCl3 solution. For preparation of 44Sc-citrate, sodium citrate (10 μL, 38.7 mM) was added to the eluted 44ScCl3 solution; and the mixture was incubated at 97 °C for 10 min. 44Sc-NODAGA was prepared with a similar procedure, except that the pH of the 44Sc solution was further adjusted with 1.0 M of ammonium acetate adjusted to pH 5 for chelation with NODAGA (10 μL, 13.6 mM). After incubation and cooling down to room temperature, 44Sc-citrate or 44Sc-NODAGA was prepared in a 30G syringe for in vivo administration.
PET imaging of animals with free and chelated 44Sc (NODAGA, Citrate)
Female Naïve Swiss Webster mice (N = 4) were injected with 3.7 MBq/400 μL of 44ScCl3 (or 44Sc-NODAGA, or 44Sc-citrate) via tail vein catheterization under 2% isoflurane anesthesia. PET imaging was performed for an initial 0.5-h on-camera dynamic image acquisition, and for 10 min static scans at 1-, 2-, and 4-h post-injection using a microPET R4 rodent scanner (Siemens). The imaged mouse was centered in the field of view and maintained under 1–2% isoflurane anesthesia during PET imaging. The calibration factor of the PET scanner was determined with a mouse-sized phantom composed of a cylinder uniformly filled with an aqueous solution of 18F with a known activity concentration. Acquisitions were recorded using an energy window of 350–700 keV and coincidence-timing window of 6 ns. PET image data were corrected for detector non-uniformity, deadtime, random coincidences and physical decay and images were reconstructed by an iterative 3D maximum a priori algorithm.
The acquired PET images were analyzed using ASIPro software (Siemens). Volume of interest (VOI) analysis of the acquired images was performed using ASIPro software, and the observed value (percent injected activity/cubic centimeter, %IA/cc) represents the mean radiotracer accumulation in the organs. The sequential radioactivity measurements (%IA/cc) were plotted over time post-administration.
Kinetic biodistribution of 44ScCl3 in naïve mice
Animals were administered 3.7 MBq/400 μL of 44ScCl3 for kinetic biodistribution studies. Four animals at each time point 5, 30, 60, 120, 240 and 1440 min post-injection were submitted for CO2 asphyxiation prior to tissue dissection. The organs of interest were collected, rinsed of excess blood, blotted, weighed, and counted with a 2480 WIZARD2 automatic γ-counter. We computed the percent of injected activity per gram of tissue (%IA/g) by normalizing the activity of each tissue to an injection standard, and the sample mass.
Estimation of human radiation dose
Biodistribution data of
44ScCl
3 in the Naïve Swiss Webster mice were extrapolated to human organs using the relative organ mass scaling method [
15‐
17]. In this method, the animal organ data reported as percent of injected activity per gram of organ,
\(\left( {\frac{{\% {\text{IA}}}}{{{\text{g}}_{{{\text{organ}}}} }}} \right)_{{{\text{mouse}}}}\), is extrapolated using the animal and human whole-body masses,
\({\mathrm{kg}}_{\mathrm{TBweight}}\), and the human organs masses,
\(\left({\mathrm{g}}_{\mathrm{organ}}\right)_{\mathrm{human}}\), employing the following equation:
$$\left( {\frac{{\% {\text{IA}}}}{{{\text{organ}}}}} \right)_{{{\text{human}}}} = \left[ {\left( {\frac{{\% {\text{IA}}}}{{{\text{g}}_{{{\text{organ}}}} }}} \right)_{{{\text{mouse}}}} \times \left( {{\text{kg}}_{{\text{TBweight}}} } \right)_{{{\text{mouse}}}} } \right] \times \left( {\frac{{{\text{g}}_{{\text{organ}}} }}{{{\text{kg}}_{{{\text{TBweight}}}} }}} \right)_{{\text{human}}}$$
The human organs masses were used as defined for adult male and female in the IDAC Dose 2.1 application [
18]. This scaling was not applied to the organs of the gastrointestinal tract. Organ integrated time-activity were determined by numerical integration of time activity data. The cumulative activity, Ã, between time 0 and the first measured time point was calculated assuming a linear increase from 0 to the first measured activity. The à between the first measured time point and the last measured time point was integrated numerically using trapezoidal approximation. The à from the last measured time point to infinity was integrated considering only the physical decay. It was assumed that the radioisotope does not relocate following the last imaging point. For walled organs (heart, large intestine, small intestine, and stomach), the residence time was assigned entirely to the organ walls; with the large intestine, the residence time was divided evenly between the right and left colons. The bone residence time was likewise evenly divided between cortical and trabecular bone [
19].
The cumulated activities for each organ were then used to compute the absorbed doses by IDAC Dose 2.1 [
18]. The mean normal-organ absorbed doses (mGy/MBq administered) and the effective dose (mSv/MBq administered) for
44ScCl
3 were calculated for standard human adults (female and male). Additionally, the biodistribution data of
44ScCl
3 were used to model the absorbed doses for
47ScCl
3. Time activity curves representing
47ScCl
3 were calculated, taking into account the different half-life of the modeled radionuclide.
Statistical analysis
Data calculated using Microsoft Excel are expressed as mean ± SD. Student’s unpaired t test (GraphPad Prism 9) was used to determine statistical significance at the 95% confidence level. Differences with p values < 0.05 were considered to be statistically significant.
Discussion
There is an increased interest in the development and implementation of theranostic nuclear medicine approaches for personalized patient management. Access to radioisotopes with desirable characteristics for quantitative PET imaging that are chemically analogous to therapeutic isotopes is an area of particular focus. Recent preclinical and clinical studies have investigated
44Sc-labeled small molecules as a promising positron-emitting diagnostic and surrogate for
47Sc-based radiotherapy. In comparison with gallium-68 (1.13 h), the half-life of
44Sc (3.97 h) affords advantages for labeling, quality control evaluation, transport logistics and the biokinetics of many tracers. The imaging characteristics for emissions from
44Sc have also been shown to be well-suited for delineation of small lesions [
20] [
21]. Cyclotron production of
44Sc through irradiation of natural calcium metal or liquid targets enables tertiary medical centers and large production facilities to produce the isotope [
22] [
6]. Alternatively, distributed generator systems that separate parent
44Ti from would enable on-site production. In this study, we built and evaluated a modular
44Ti/
44Sc generator, and further investigated the absorption, distribution, and excretion of the activity after a single intravenous injection of generator eluate in female mice.
Consistent with prior investigation [
4,
13,
14],
44Ti was efficiently loaded on the resin, and we observed that this material can re-distribute on the column following repeated
44Sc elutions. This resulted in breakthrough of
44Ti and has the potential to contaminate the radiopharmaceutical and work space for compounding. To avoid
44Ti breakthrough from the generator, a bidirectional elution approach has been employed to delay the breakthrough by others, including Filosofov et al. [
23] and Radchenko et al. [
14]. A bidirectional elution approach cannot prevent
44Ti redistribution and breakthrough completely, and the
44Ti/
44Sc generator must be re-assembled after a period of use. Therefore, we engineered the modular clinical-scale generator to allow us to: (1) recover
44Ti efficiently and conveniently; (2) replace the columns independently; and (3) load
44Ti to the ZR resin generator semi-automatically with a minimum radiation dose to the operation personnel.
It has been reported that
44Ti has a consistent absorption efficiency on ZR resin across a wide range of HCl concentrations [
14]. The
44Ti/
44Sc stock solution was provided dissolved in 6 M HCl solution. We therefore loaded to the primary ZR resin column under the condition of 6.0 M HCl with
44Ti/
44Sc. More than 99.9% of
44Ti was trapped. While
44Sc can be eluted efficiently with 4.0–6.0 M HCl, the high concentration of HCl here would complicate safe-handling and requires additional adjustment of pH conditions to reach suitable conditions for radiolabeling. Thus, a lower concentration of 0.05 M HCl was chosen to elute
44Sc. A significantly higher
44Ti breakthrough was observed in the first three elutions using this lower concentration. We put forward that
44Ti
3+ may be hydrolyzed into
44Ti(OH)
x or
44TiO
2 under the condition of less than 4 M HCl solution. During the transition from 6 M HCl to 0.05 M HCl, the absorbed
44Ti
3+ may be quickly hydrolyzed and released from the resin. After elution with 0.05 M HCl for several days,
44Sc was obtained in a consistent yield with a high
44Sc/
44Ti ratio. To further limit breakthrough of
44Ti from
44Sc elution, a disposable ZR resin cartridge (0.3 mL) was utilized at the flow outlet. Our results showed that
44Sc/
44Ti ratio was further improved by 342%. Notably these small amounts of absorbed
44Ti on this disposable cartridge can be recovered by pass through of 6 M HCl/0.65% H
2O
2 [
14], which can be combined, dried and redissolved in 0.05 M HCl solution for re-loading onto the center column of the generator.
PET imaging was performed to measure pharmacokinetic profiles of
44ScCl
3 and chelated
44Sc with either citrate or NODAGA. After intravenously injection, dynamic PET imaging showed that free
44ScCl
3 is distributed in blood (heart) and lung and that activity remained in circulation for an extended period. Ex vivo analyses recapitulated high radioactivity in the blood, heart, aorta and vena cava confirming
44Sc is mainly remained in blood with half-lives of 2.0 min (rapid phase) and 133 min (slow phase), respectively. Similar to
68GaCl
3 PET imaging results in rats [
24],
44ScCl
3 in mice was slowly excreted through the kidney and liver. Our biodistribution results also show accumulation of
44ScCl
3 in the spleen, with background levels of accumulation in all other tested organs. We hypothesize that similar to the Fe
+3 and
45Ti [
25] ion, a major portion of
44Sc
+3 is bound to transferrin after intravenous administration. This results in extended circulation times and slow kinetics of excretion without specific accumulation in the heart wall or vasculature. Neither PET imaging nor biodistribution studies identified
44ScCl
3 to be excreted via feces or the intestine. This is in contrast to the elimination of
68Ga or
64Cu, in which this gastrointestinal clearance presents a complication for interpreting preclinical imaging data. Together, these data motivate use of very high in vivo stability chelators for
44/47Sc targeted agents and for understanding of artifactual distributions.
44Sc-citrate showed a similar in vivo pharmacokinetic profile with an increased rate of clearance. The major difference identified was a higher kidney accumulation as
44Sc forms only a weak complex with citrate that may prevent rapid complexation by components in the blood. When the more stable
44Sc-NODAGA was used, activity was observed to transit into the bladder rapidly, which is consistent with the in vivo profiles of
44Sc-labeled peptides utilizing DOTA and NODAGA [
26] or other novel chelate-conjugated ligands [
27]. Further evaluation of the chemical identity of the generator output and its impact on radiotracer labeling and distribution will be conducted. This is of particular interest for future comparison of generator- and cyclotron-produced
44Sc.
Together, differences in distribution of free and chelated scandium imply that any unlabeled
44Sc/
47Sc in the solution of
44Sc/
47Sc-chelated ligand may cause a significant increase in the observed circulation time. To further clarify what amount of absorbed dose may be caused by the unlabeled
44Sc/
47Sc, the data of a kinetic biodistribution in the mice was extrapolated to male and female adults. We observed a gender difference of irradiation dose and that the free-
44Sc has a significantly higher irradiation burden than that of the conjugated
44Sc-ligands, as expected. For example, the mean effective dose of [
44Sc]-PSMA617 in male patient is 0.0389 mSv/MBq [
28], whereas the unconjugated
44Sc (current work) is 0.146 mSv/MBq. Similar results have been found in other critical organs, including spleen, liver and red marrow (0.185 vs 0.502; 0.107 vs 0.524; 0.0331 vs 0.124 mSv/MBq). These data indicate that a high purity for
44/47Sc-conjugated ligand is required for safe and effective radionuclide-based treatment in the future and provide insight into assessment of the radiation burden from decomplexed radioisotope in vivo.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.