Materials

LN2 resin (G71m resin loaded with 2-ethylhexylphosphonic acid mono-2-ethylhexyl ester (HEH[EHP])) was purchased from TrisKem International having the following physical and chemical properties: particle size = 500–100 µm, density = 0.37 g/mL, capacity = 0.16 mmol/mL resin, conversion factor DW/k’ = 1.82).

225Ac was provided by Global Morpho Pharma (GMP, La Chapelle-sur-Erdre, France) as 225Ac(NO3)3 dry film. The full activity batch was subsequently diluted in 150 µL ultra-pure and metal-free 0.05 M HCl and transferred into a 1.5 mL Eppendorf tube (Protein LoBind®).

Phosphate-buffered saline (PBS, 1X, pH 7.4) was purchased from Gibco, 0.15 M sodium chloride (NaCl, pH 5.5) and sodium acetate (NaOAc) from Sigma-Aldrich. TraceMetal Grade nitric acid (HNO3, 67–69%) was purchased from Fischer Scientific, HIPERPUR-plus hydrochloric acid (HCl, 35%) from PanReac AppliChem, and TraceSELECT™ H2O from Honeywell.

Radionuclide separation

225Ac(NO3)3 (~ 50 MBq, 150 µL 0.05 M HCl) was diluted with 450 µL 0.01 M HNO3 to a total volume of 600 µL and loaded onto 0.3 mL (Height: 19.8 mm x Diameter: 11.7 mm) LN2 column, which was pre-conditioned with at least three column volumes of 0.01 M HNO3. Afterwards, the column was washed with 800 µL of 0.01 M HNO3 (pH 2.0), 0.15 M NaCl (pH 5.5), PBS (pH 7.4), and 0.1 M NaOAc (pH 6.5). Subsequently, multiple 221Fr elutions were performed with 800 µL 0.1 M NaOAc (pH 6.5) after sufficient 221Fr in-growth time of 25–30 min. Four fractions of 150, 300, 300 and 150 µL were collected and measured immediately in a dose calibrator (ISOMED 2010, Nuvia Instruments, Dresden, Germany). Quality control was conducted retrospectively by time and energy-dependent measurements in a gamma counter (Wizard 2410, with 10 independent NaI(Tl) detectors, Perkin Elmer, Waltham, Massachusetts, USA) and a gamma spectrometer HPGe(Li) detector, Mirion Technologies, Atlanta, Georgia, USA). The activity of 221Fr and 213Bi was assessed from the photo-peaks of 218 and 440 keV, respectively.

Cell line

LNCaP cells, derived from a human lymph node metastatic lesion of prostatic adenocarcinoma, were obtained from the German Collection of Microorganisms and Cell Cultures (Leibniz Institute DSMZ, Braunschweig, Germany).

Animal model

For the biodistribution study, male SCID (CB-17/Icr-Prkdcscid/scid/Rj) mice (20–25 g, 7-weeks-old, Janvier Labs, Le Genest-Saint-Isle, France) were implanted subcutaneously (s.c.) with testosterone pellets (12.5 mg, 4 mm, prepared in-house). One day later, mice were inoculated s.c. with 5 × 106 LNCaP cells into the right flank. All animal experiments were conducted in accordance with the German Animal Welfare Law and approved by the local authorities.

Biodistribution of 221Fr in vivo

On Day 35 after inoculation, the mice (n = 3/group) were injected with a single dose of 221Fr and sacrificed at 5 min post-injection (p.i.). On Day 37 after inoculation, the mice (n = 3/group) were injected with a single dose of 221Fr and sacrificed at 15 min p.i. The procedure for each mouse was as follows: 25 min after the previous 225Ac/221Fr generator elution, a fresh elution of four fractions (150, 300, 300 and 150 µL) was taken and the second fraction was checked for pH (> 6.0 and < 7.0) and a 100 µL intravenous (i.v.) tail vein injection into one mouse followed within 2–3 min after elution. Standards of the injected volume were taken for each injected elution and measured in parallel. Throughout the study, a timer was continuously used to record exact time-points for all actions.

After 5–15 min the mouse was sacrificed and the following organs were collected: kidney, tumor, blood, liver, small intestine, salivary gland, large intestine, pancreas, spleen, muscle. The organs were halved or divided for pair organs to allow for two parallel samples for measurement 3–5 min after start of sectioning. One set was measured once by gamma spectrometry measurements HPGe(Li) detector. The other set of samples was measured several times in a gamma counter by NaI(Tl) crystal. The next mouse was injected ~ 50 min later.

Biodistribution of 213Bi in vivo

Biodistribution of 213Bi was performed analogically as for 221Fr. On Day 35 after inoculation, the mice (n = 3/group) were injected with a single dose of 213Bi and sacrificed at 5 min p.i. On Day 37 after inoculation, the mice (n = 3/group) were injected with a single dose of 213Bi and sacrificed at 15 min p.i. For the 213Bi biodistributions the elutions were approximately 2 h old to not contain any 221Fr or 217At. The three mice for each time-point were injected in 5 min intervals on one day. Standards of the injected volume were taken and measured in parallel. The collection of organs and measurements was analogous to the 221Fr biodistribution, as described in the previous paragraph.

Measurement HPGe(Li) detector

The gamma spectrometry measurements were performed with a HPGe(Li) detector purchased from Mirion Technologies, Atlanta, Georgia, USA. Measuring time was 5–6 min per sample. The activity of 221Fr and 213Bi were assessed from the gamma lines specifically attributed to these radionuclides.

Measurement NaI(Tl) detector

Wizard 2410 employing 10 independent NaI(Tl) detectors was purchased from Perkin Elmer, Waltham, Massachusetts, USA. Each sample was measured six times over a period of 40–45 min, with each measurement lasting 1 min. The acquisitions were performed in the channels of 180–240 keV for 221Fr and 380–520 keV for 213Bi.

Quantification of 213Bi and 221Fr activities in 213Bi- and 221Fr-injected mice

To determine the distribution of 213Bi and 221Fr in organs, we analyzed data from two experimental setups: mice injected with ²¹³Bi, and mice injected with 221Fr. We began by processing the data from the 213Bi-injected mice, where no 221Fr was present in the tissues. Thus, 213Bi activity was calculated by fitting an exponential decay function [Eq. (1)] to the CPM values measured in the 213Bi energy window, with constraints ensuring non-negative initial CPM and background.

For the 221Fr-injected mice, we accounted for the fact that 213Bi decay also contributes to the signal in the 221Fr window. Additionally, the calculation of 213Bi activity in these animals required fitting a more complex function derived from the Bateman equations.

Calibration was performed using 225Ac in secular equilibrium, which provided known activities for both 213Bi and 221Fr. Linear fits through the origin yielded calibration factors of fBi =1.649 +/- 0.062 Bq⁻¹ and fFr = 3.513 +/- 0.012 Bq⁻¹.

All fits were performed in OriginPro 2021 using the Levenberg–Marquardt algorithm, with instrumental weighting to account for CPM uncertainty. The calculation procedure is described in more detail in the Supplementary Material.