Production and radiochemistry of the in vivo PET generator ¹⁴⁰Nd/¹⁴⁰Pr as an imaging surrogate for F-block therapeutic radionuclides

Abstract

Theranostics, a combined approach of diagnostics and therapeutics, often employs F-block therapeutic radionuclides including ²²⁵Ac, ¹⁷⁷Lu, and ¹⁶¹Tb. While there is a lack of F-block PET imaging radionuclides, the in vivo PET generator pair ¹⁴⁰Nd/¹⁴⁰Pr can act as a theranostic imaging counterpart to the F-block therapeutic radionuclides. In this study, we explored the production and separation of high purity ¹⁴⁰Nd via the ¹⁴¹Pr(p,2n)¹⁴⁰Nd reaction route. Monoisotopic ¹⁴¹Pr targets irradiated with 20 MeV protons for 10 min with 10 µA beam current yielded 21.45 ± 0.82 MBq (580 ± 22 µCi) of ¹⁴⁰Nd. A two-step separation method was developed for the purification of ¹⁴⁰Nd from the ¹⁴¹Pr target material. Recoveries of 27.4 ± 2.1% ¹⁴⁰Nd were obtained upon separation with < 20 ppb of ¹⁴¹Pr target material in the final product. Radiolabeling of Macropa and DOTA chelators with ¹⁴⁰Nd resulted in [¹⁴⁰Nd]Nd-Macropa with a molar activity of 74.0 MBq/µmol (2.0 mCi/µmol) and [¹⁴⁰Nd]Nd-DOTA with a molar activity of 70.3 MBq/µmol (1.9 mCi/µmol). An imaging study with a phantom indicated the PET spatial resolution of ¹⁴⁰Nd/¹⁴⁰Pr was distinguishable down to 2.4 mm. This study sets the stage for the ¹⁴⁰Nd/¹⁴⁰Pr in vivo PET generator to be explored in radiopharmaceutical applications.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18929-4.

Introduction

Theranostics uses an integrated approach of diagnostic imaging by utilizing positron emission tomography (PET), or single photon emission computed tomography (SPECT), combined with targeted radionuclide therapy (TRT), which aims to improve the efficacy of treatments via selection of patients most likely to respond to the targeted radionuclide therapy using the paired imaging agent^1–3. Theranostic radiopharmaceuticals utilize the same targeting vector with different radionuclides for imaging and TRT¹. A chemically equivalent (or similar) PET diagnostic radionuclide can be used to accurately predict the chemistry, dosimetry, and pharmacokinetics of its therapeutic counterpart.

Recent advancements in radiopharmaceutical development have resulted in effective theranostic strategies, with many including F-block therapeutic radionuclides such as ¹⁶¹Tb (t_1/2: 6.9 d), ¹⁷⁷Lu (t_1/2: 6.7 d), ²²⁵Ac (t_1/2: 9.9 d), and ²²⁷Th (t_1/2: 18.7 d)^1,4. Notably, the FDA approval of [¹⁷⁷Lu]Lu-DOTATATE for treating neuroendocrine tumors (NET) and [¹⁷⁷Lu]Lu-PSMA-617 for metastatic castration-resistant prostate cancer (mCRPC) highlights the growing interest in theranostic strategies, with a number of clinical trials underway for F-block therapeutic agents, including [²²⁵Ac]Ac-PSMA-617, [²²⁵Ac]Ac-PSMA-I&T, [²²⁵Ac]Ac-J591, and [¹⁶¹Tb]Tb-PSMA I&T for mCRPC, and [²²⁵Ac]Ac-lintuzumab for myeloid leukemia^5–9. ⁶⁸Ga (t_1/2: 67.7 min, 88.9% β⁺) is often utilized to predict tumor uptake and assess treatment response for therapies such as [¹⁷⁷Lu]Lu-DOTATATE and [¹⁷⁷Lu]Lu-PSMA-617^10,11. However, its short half-life combined with different coordination chemistry can hinder the use of this radionuclide for certain applications. A critical gap remains due to the lack of PET-compatible diagnostic counterparts for many F-block therapeutic radionuclides, which drives the exploration of novel approaches^12,13.

An in vivo PET generator leverages the decay of a long-lived parent radionuclide to produce a rapidly decaying positron-emitting daughter in vivo, enabling extended PET imaging without the need for repeated administration of radiopharmaceuticals. Two notable examples of in vivo PET generators include cerium-134 (¹³⁴Ce (t_1/2: 3.2 d, EC))/lanthanum-134 (¹³⁴La (t_1/2: 6.5 min, 63.6% β⁺)) and neodymium-140 (¹⁴⁰Nd (t_1/2: 3.4 d, EC))/praseodymium-140 (¹⁴⁰Pr (t_1/2: 3.4 min, 51.0% β⁺)) from the F-block series. The decay schemes for these systems are illustrated in Fig. 1. These generator systems are anticipated to exhibit pharmacokinetics closely resembling those of other F-block elements and, owing to their similar chemical properties, may serve as effective PET imaging surrogates for therapeutic radionuclides within the same elemental series^14–16.

Fig. 1.

Decay schemes of (a) ¹³⁴Ce (t_1/2: 3.2 d). (b) ¹⁴⁰Nd (t_1/2: 3.4 d).

Production of ¹³⁴Ce is challenging due to the need for a high energy proton beam (≥ 100 MeV) to induce the nuclear reaction ¹³⁹La(p,6n)¹³⁴Ce, thereby limiting ¹³⁴Ce availability¹⁷. The in vivo generator pair ¹⁴⁰Nd/¹⁴⁰Pr is relatively unexplored, with very little chemistry published. The production of ¹⁴⁰Nd via the ¹⁴¹Pr(p,2n)¹⁴⁰Nd route exhibits a cross section > 950 millibarn (mb) at 22 MeV with a peak between 18 and 24 MeV proton energy (Fig. 2)^18–21.

Fig. 2.

Cross-sections of the ¹⁴¹Pr(p,2n)¹⁴⁰Nd reaction.

This work aimed to investigate the production of ¹⁴⁰Nd using ¹⁴¹Pr target material and the development of purification methods to obtain high purity ¹⁴⁰Nd. Further, we investigated the radiochemistry of ¹⁴⁰Nd with DOTA and Macropa chelators and conducted PET phantom imaging of the ¹⁴⁰Nd/¹⁴⁰Pr.

Methods

Materials

Praseodymium foil (99.8% purity), tantalum sheets (3N8 purity), and aluminum sheets (3N purity) were purchased from ESPI metals (Ashland, OR, USA). Normal DGA (N,N,Nʹ,Nʹ-tetra-n-octyldiglycolamide) resin, of particle size 50–100 µm, was purchased from Eichrom Technologies (Lisle, IL, USA). 1 mL polypropylene solid phase extraction (SPE) columns were purchased from MilliporeSigma (Supelco, Sigma-Aldrich, St. Louis, MO, USA). A PEEK 7.4 mm × 300 mm empty HPLC column was purchased from Analytics-Shop USA LP (Stockbridge, GA, USA). 2-Pyridinecarboxylic acid, 6,6′-[1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diylbis(methylene)]bis- (Macropa) chelator was purchased from MedchemExpress USA (Monmouth Junction, NJ, USA), and 1,4,7,10-Tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) was purchased from Macrocyclics (Plano, TX, USA). Paper backed silica SG-iTLC plates were purchased from Agilent Technologies (Santa Clara, CA, USA). All other materials were purchased from Sigma-Aldrich and Fisher Scientific as trace metal grade unless stated otherwise. Stable element detection and quantitative analysis were carried out by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7900, Agilent Technologies, Santa Clara, CA, USA). Multi-element calibration standards were purchased from Agilent Technologies (Santa Clara, CA, USA). Gamma spectra for radionuclide identification, yield determination, and radionuclidic purity were acquired using a calibrated high purity germanium detector (HPGe), Canberra GC 2018, interfaced with a DSA = 100 multichannel analyzer (Meriden, CT, USA), and Genie 2000 software was used for spectra analysis. The HPGe detector was calibrated following the methodology described in Cingoranelli et al., and dead time was maintained below 5% for all gamma ray measurements²². A dose calibrator (CRC-25R, Mirion Technologies, Florham Park, NJ, USA), cross-calibrated with the HPGe, was used for the measurement of radioactivity throughout the separation process to monitor ¹⁴⁰Nd (calibration no. 890) elution, and during radiolabeling and imaging studies. Radiochemical yields of ¹⁴⁰Nd chelates were assessed using a radio-TLC scanner (AR-2000, Eckert and Ziegler, Berlin, Germany). A variable-speed peristaltic pump (0.1-200 rpm) was acquired from Cole-Parmer (IL, USA). A micro-Derenzo phantom (outer diameter: 50 mm; hot rod diameters: 1.2-4.8 mm) was purchased from Phantech (Madison, WI, USA). PET imaging was performed using a GNEXT PET scanner (GNEXT PET/CT, Sofie Biosciences, CA, USA), and image analysis was conducted with VivoQuant software (version 4.0, VivoQuant, Perceptive Inc., Boston, MA, USA).

Target preparation and irradiation

Pr foil with 0.1 mm thickness and 99.8% purity (ESPI metals) was cut into circular 10 mm diameter pieces (50–55 mg). Targets were flushed with argon gas and stored in a desiccator to prevent excessive oxidation. A thin greenish Pr₂O₃ coating developed on targets over time and was removed with ethanol wipes before irradiation. Each ¹⁴¹Pr target was placed into a 10 mm diameter circular divot of a tantalum coin (23.6 mm diameter, 1.5 mm thick) with a 0.5 mm deep divot in the center and 1 mm backing as shown in Fig. 3a,b²³. An aluminum holder with an opening in the center was placed on the top to hold the foil in place (Fig. 3c). This target assembly was inserted into the target holder (Fig. 3d) of the cyclotron solid target station (Fig. 3e) for irradiation to produce ¹⁴⁰Nd^22,24. The entrance and exit energies of the proton beam through the target foil and tantalum backing were calculated using the Stopping and Range of Ions in Matter (SRIM-2013) simulation software²⁵. Irradiation parameters were initially optimized using a 10 µA beam current at proton energies of 18, 20, and 24 MeV for a 10-min duration. Later, irradiation currents up to 30 µA and times up to 30 min at 20 MeV were utilized for larger scale ¹⁴⁰Nd production used in purification development, radiolabeling, and imaging studies. To determine the radionuclide activity in irradiated samples, aliquots were taken from the dissolved target solutions and diluted to a final volume of 1 mL using Milli-Q water in 1.5 mL microcentrifuge tubes. These tubes were positioned facing the detector at calibrated measurement points on the high-purity germanium (HPGe) detector system, and gamma ray spectroscopy was performed for quantitative analysis.

Fig. 3.

Target design for ¹⁴¹Pr targets. The ¹⁴¹Pr foil was inserted into the divot in tantalum coin and held in place using an aluminum retainer (design adopted from Pyles et al.²⁴).

¹⁴⁰Nd purification

Irradiated ¹⁴¹Pr foils were dissolved in 15 mL of 3.5 M HCl without heating. The foil dissolved via a vigorous exothermic reaction within 60 s and the solution was allowed to cool to room temperature before proceeding. The purification was performed 24 h post irradiation. A two-step separation protocol was established for the purification of ¹⁴⁰Nd. All solvent loading steps were performed using a peristaltic pump operated at variable flow rates. For step 1, normal DGA resin (50–100 µm particle size, 4.5 g) was suspended in 50 mL of a 20:80 v/v ethanol:Milli-Q water mixture and packed into a 7.4 × 300 mm polyether ether ketone (PEEK) empty HPLC column (Analytics-Shop) (column 1) in five portions of 10 mL each, followed by a 10 mL Milli-Q water rinse. All steps using column 1 were carried out with a 1 mL/min flow rate. The prepared column was conditioned with 15 mL each of 0.5 M HCl, 2.2 M HCl and 3.5 M HCl. The dissolved irradiated foil (D (3.5 M HCl, 15 mL)) was loaded into the column. The column was washed with 440 mL of 2.5 M HCl (W1 (2.5 M HCl, 440 mL)), followed by three 13 mL fractions of 2.5 M HCl (W2–W4 (2.5 M HCl, 13 mL)) and a single 40 mL fraction of 2.2 M HCl (W5 (2.2 M HCl, 40 mL)). A final rinse with 20 mL of 0.5 M HCl was performed to remove any residual target material. The W5 fraction from column 1 was reformulated to 3.5 M HCl with the addition of concentrated HCl for use in the second separation step. For step 2, a 1 mL solid-phase extraction (SPE) column (column 2) fitted with frits was packed with 300 mg normal DGA resin and conditioned with 10 mL Milli-Q water, 10 mL 0.5 M HCl, followed by 10 mL of 3.5 M HCl. The reformulated solution (2-D (3.5 M HCl, 46 mL)) was loaded onto column 2, with the same flow rate, and the flowthrough was collected (2-FT (3.5 M HCl, 46 mL)). Column 2 was washed with 40 mL of 3.5 M HCl (2-W1 (3.5 M HCl, 40 mL)). Next, column 2 was eluted with 400 µL fractions of Milli-Q water (H₂O) and collected in 1.5 mL centrifuge tubes (2-E1, E2 (H₂O, 400 µL)) with a 0.4 mL/min flow rate. The fraction 2-E2 was used for downstream studies.

Method and product analysis

A dose calibrator (CRC-25R) was used throughout both steps to track ¹⁴⁰Nd activity (based on ¹⁴⁰Pr emissions) during loading, washing, and elution. Recovery of ¹⁴⁰Nd in each fraction at each step was evaluated by gamma ray spectroscopy using a high-purity germanium (HPGe) detector, as previously described. Samples were measured on HPGe a minimum 45 min after the fraction collection to allow for equilibrium to be established with ¹⁴⁰Nd and its daughter ¹⁴⁰Pr.

For elemental analysis, fractions were allowed to decay for one month before analysis. Aliquots from each fraction were diluted to a final volume of 5 mL in 2% nitric acid to maintain elemental concentrations below 1000 ppb. Quantitative analysis of stable impurities was performed using an Agilent 7900 ICP-MS, calibrated with multi-element ICP-MS standards according to the procedure described by Cingoranelli et al.²². A total of 14 lanthanides, including non-radioactive ¹⁴¹Pr, and 5 other common metal contaminants were quantified to assess the purity of ¹⁴⁰Nd and the extent of stable elemental contaminants in each fraction.

Radiolabeling Studies

Macropa and DOTA 30 µg/µL stock solutions were prepared in 1 M ammonium acetate (NH₄OAc) buffer and stored at − 20 °C until use. The ¹⁴⁰Nd stock was prepared in 400 µL of 0.5 M HCl and pH adjusted to pH 3 with 5 M NaOH, followed by the addition of 1 M ammonium acetate to achieve pH 6.5. A serial dilution of decreasing concentrations of Macropa (18.80 mM to 0.95 µM) and DOTA (19.50 mM to 0.99 µM) solutions were prepared from Macropa and DOTA stocks using 1 M ammonium acetate (NH₄OAc) buffer as the diluent. To each diluted chelator solution, approximately 1.85 MBq (50 µCi) of ¹⁴⁰Nd was added and allowed to react at 37 °C at 600 rpm for 30 min. Radiochemical yields were measured by instant thin layer chromatography (iTLC) using silica gel deposited paper chromatography plates. 1 µL from each reaction mixture was spotted at the bottom of the plate, which was developed using 0.1 M citric acid, pH 5.5, solution as the mobile phase. Plates were scanned on a radioTLC scanner 45 min post development with the mobile phase to allow for equilibrium to be established with ¹⁴⁰Nd/¹⁴⁰Pr. The radiochemical yields on scanned plates were analyzed using the Winscan program (Eckert & Ziegler, Berlin, Germany). The radiochemical yield (%) versus concentration of chelator (log[chelator][μmol]) was plotted using GraphPad Prism to calculate statistical mole amount of chelator for a 50% reaction. (Supplemental Fig. S5). The ¹⁴⁰Nd activity is divided by two-times the available mole of chelator at 50% reaction; this value represents apparent molar activity (AMA)²². Results are reported as MBq (mCi)/μmol.

Stability studies

Complexes of [¹⁴⁰Nd]Nd-Macropa and [¹⁴⁰Nd]Nd-DOTA were prepared by reacting ~ 1.85 MBq (50 µCi) of ¹⁴⁰Nd with 167 nmol Macropa or 173 nmol DOTA in 1 M NH₄​OAc buffer (pH 6.5) in three separate microcentrifuge tubes for each complex. Radiochemical yields exceeded 95%, as confirmed by iTLC. For stability assessment, the complexes were incubated in saline, mouse serum, and PBS (ten-fold excess volume) at 37 °C at 600 rpm for up to five days, with decomplexation monitored by iTLC at defined time points.

Phantom imaging

A micro Derenzo phantom device with hot rod sizes from 1.2, 1.6, 2.4, 3.2, 4.0, and 4.8 mm outer diameter (Fig. 9a) was filled with 3.7 MBq (100 µCi) of ¹⁴⁰Nd in 5 mL of 0.5 M ammonium acetate²². Special care was taken to avoid any bubble formation inside the rods during filling. The filled phantom was scanned for 4 h on a small animal GNEXT PET/CT scanner (GNEXT PET/CT, Sofie Biosciences, CA, USA), followed by a 5-min CT scan at a voltage of 80 kVp. PET acquisition was performed using an energy window of 350–650 keV, a tube current of 150 µA, and 720 projections. Reconstruction of the acquired PET images was carried out using a 3D-OSEM (ordered subset expectation maximization) algorithm with 24 subsets and 3 iterations, incorporating random, attenuation, and decay corrections. Image analysis was performed using VivoQuant software (VivoQuant 4.0, Perceptive Inc., Boston, MA, USA).

Fig. 9.

Derenzo phantom with rod sizes from 1.2 mm to 4.8 mm outer diameter. (a) Outline of phantom’s top view with clusters of circles representing different rod sizes. (b) PET image of phantom filled with 3.7 MBq (100 µCi) of ¹⁴⁰Nd in 5 mL of 0.5 M ammonium acetate and scanned for 4 h using a PET/CT scanner. The activity resolution of the phantom was distinguishable up to 2.4 mm hole size.

Results

Target preparation and irradiation

Among several reaction routes available for ¹⁴⁰Nd production (Supplemental Fig. S1), the reaction route ¹⁴¹Pr(p,2n) was investigated to produce ¹⁴⁰Nd. To minimize side reactions including ¹⁴¹Pr(p,n)¹⁴¹Nd (t_1/2: 2.5 h) (Supplemental Fig. S2a), ¹⁴¹Pr(p,3n)^139mNd (t_1/2: 5.5 h) (Supplemental Fig. S2c) and ¹⁴¹Pr(p,x)¹³⁹Ce (t_1/2: 137.6 d) (Supplemental Fig. S2d), a 20 MeV proton beam energy was selected. SRIM-2013 simulation results indicated the entrance beam energy of 20 MeV was reduced to 19.10 ± 0.07 MeV through a 0.1 mm ¹⁴¹Pr foil and stopped by 1 mm tantalum backing. The theoretical yield of ¹⁴⁰Nd was calculated to be 29.95 MBq (809 µCi) for a 10-min irradiation with 10 μA current using a 20 MeV to 19 MeV energy slice at the end of the bombardment (EOB). The experimental yield measured using gamma ray spectroscopy was 21.45 ± 0.82 MBq (580 ± 22 µCi) (n = 3) at EOB. While ¹⁴⁰Nd emits X-rays, it does not have any characteristic gamma ray peaks. Hence, its detection and activity calculation were based on the detection and gamma ray peak intensity of the daughter ¹⁴⁰Pr after the secular equilibrium is established. The gamma ray spectrum showed peaks at 511 keV (positron annihilation, 102.0%) and very low intensity peaks at 306.8 keV (0.15%), 751.8 keV (0.03%), and 925.2 keV (0.02%), which are characteristic of ¹⁴⁰Pr (Fig. 4). ¹⁴¹Nd activity at EOB for 10 min, 10 µA irradiation was 12.2 ± 0.8 MBq (330.1 ± 22.3 µCi) (n = 3). Irradiation at a higher proton energy (24 MeV) resulted in the observation of ^139mNd, ¹³⁹Nd, and ¹³⁹Ce addition to ¹⁴⁰Nd (see Supplemental Fig. S4). Irradiation at 18 MeV resulted in the observation of characteristic gamma peaks corresponding to the decay of ¹⁴¹Nd (t_1/2 = 2.5 h). The gamma peak energies of these radionuclides are listed in Table 1. Within 24 h post-irradiation, the activity of ¹⁴¹Nd decayed to background levels. The larger scale ¹⁴⁰Nd production at 20 MeV, 30 min irradiation time, and 30 µA proton beam current yielded 205.20 ± 17.13 MBq (5.54 ± 0.46 mCi) (n = 3) ¹⁴⁰Nd at EOB, which was used for downstream studies.

Fig. 4.

Gamma ray spectrum of aliquots taken from dissolved irradiated ¹⁴¹Pr. The 306.8 keV, 511 keV and 751.8 keV and 925.2 keV are characteristics peaks of ¹⁴⁰Pr, daughter of ¹⁴⁰Nd.

Table 1.

Half-lives, photon energies and their intensities, and particles of interest emissions and their branching ratios of the radionuclides mentioned in the manuscript. These radionuclides also emit other low energy gamma and X-ray photons and Auger and/or conversion electrons in addition to the mentioned particle emission.

Radionuclides	Half-life	Gamma photon energy (intensity)	Other particles of interest emission (branching ratio)	Remarks
161Tb	6.9 d	25.7 keV (23.2%) 48.9 keV (17.0%) 74.6 keV (10.2%) 57.2 keV (1.8%)	β− (100.0%)	It also emits other low intensity gamma photons
177Lu	6.7 d	208.4 keV (10.4%) 112.9 keV (6.2%) 321.3 keV (0.2%)	β− (100.0%)	It also emits other low intensity gamma photons
225Ac	9.9 d	99.8 keV (1.0%) 99.6 keV (0.7%) 150.1 keV (0.6%)	α (100.0%)	225Ac decay chain upto stable 209Bi emits α and β- particles. It also emits other low intensity gamma photons
227Th	18.7 d	236.0 keV (12.9%) 50.1 keV (8.4%) 256.2 keV (7.0%) 329.8 keV (2.9%) 300.0 keV (2.21%)	α (100.0%)	227Th decay chain upto stable 207Pb emits α and β- particles. It also emits other low intensity gamma photons
68Ga	67.7 min	1077.3 keV (3.2%) 1883.2 keV (0.1%) 1261.1 keV (0.1%) 805.8 keV (0.1%)	β+ (88.9%)	It also emits other low intensity gamma photons
132La	4.8 h	464.5 keV (76.0%) 567.1 keV (15.7%) 1909.9 keV (9.0%) 663.07 keV (7.8%)	β+ (42.1%)	It also emits other low intensity gamma photons
133La	3.9 h	278.8 keV (2.4%) 302.4 keV (1.6%) 290.0 keV (1.4%)	β+ (7.1%)	It also emits other low intensity gamma photons
135La	19.2 h	480.5 keV (1.5%) 874.5 keV (0.2%) 587.8 keV (0.1%)	β+ (< 0.1%)	It also emits other low intensity gamma photons
134La	6.5 min	604.7 keV (5.0%) 1554.9 keV (0.4%) 563.2 keV (0.4%)	β+ (63.6%)	It also emits other low intensity gamma photons
134Ce	3.2 d	162.3 keV (0.2%) 130.4 keV (0.2%) 300.9 keV (0.1%)		134Ce decay to short half-life radionuclide 134La. It also emits other low intensity gamma photons
139Ce	137.6 d	165.9 keV (79.9%)
139Nd	29.7 min	405.0 keV (6.9%) 1074.2 keV (2.5%) 669.0 keV (1.5%) 916.9 keV (1.5%)	β+ (24.7%)	It also emits other low intensity gamma photons
139mNd	5.5 h	113.9 keV (40%) 738.2 keV (35.1%) 982.2 keV (26.3%) 708.1 keV (26.3%)	β+ (2.3%)	It also emits other low intensity gamma photons
140Pr	3.4 min	1596.1 keV (0.49%) 306.9 keV (0.15%) 751.8 keV (0.03%) 925.3 keV (0.02%)	β+ (51.0%)	It also emits other low intensity gamma photons
140Nd	3.4 d			140Nd decays to short half-life radionuclide 140Pr
141Nd	2.5 h	1126.9 keV (0.8%) 1292.6 keV (0.5%) 1147.3 keV (0.3%)	β+ (2.6%)	It also emits other low intensity gamma photons

¹⁴⁰Nd purification

Normal DGA resin was selected for the separation of ¹⁴⁰Nd from the ¹⁴¹Pr target material. The separation was performed in 2 steps, after approximately 24 h post irradiation for short-lived side products including ¹⁴¹Nd (t_1/2: 2.5 h) and ^139mNd (t_1/2: 5.5 h), to decay. The separation profiles for ¹⁴¹Pr and ¹⁴⁰Nd are shown in Fig. 5. 94.6 ± 5.6% (n = 3) of the ¹⁴¹Pr target material was eluted with 440 mL of 2.5 M HCl in the W1 fraction in step 1. The rest of the wash steps (W2-W5) eluted ¹⁴⁰Nd along with the remainder of the ¹⁴¹Pr. The W5 fraction contained the highest amount (36.8 ± 0.4%) (n = 3) of ¹⁴⁰Nd in 40 mL of 2.2 M HCl, which was used for column 2. The HCl concentration was reformulated to 3.5 M HCl to retain ¹⁴⁰Nd during loading and washing in column 2. In the 2-E2 elution fraction of step 2, 27.4 ± 2.1% (n = 3) ¹⁴⁰Nd (decay corrected to EOB) was recovered with a ¹⁴¹Pr concentration below 20 ppb in 400 µL Milli-Q water. Due to the preceding wash step (2-W1, using 3.5 M HCl), the collected final 2-E2 fraction had an approximate acid concentration of 0.5 M HCl. This purified fraction was used for all downstream radiolabeling and imaging studies. The target material (¹⁴¹Pr 51.6 ± 1.3 mg) contained 0.050 ± 0.001% (26.0 ± 0.8 µg) cerium (Ce), 1.050 ± 0.037% (540.0 ± 19.0 µg) neodymium (Nd), and 1.070 ± 0.028% (550.0 ± 14.0 µg) gadolinium (Gd) impurities based on ICP-MS analysis. Target material impurities of 13 lanthanides and other possible contaminants analyzed by ICP-MS are reported in Supplementary data Table S1. The Ce and Gd impurities were removed with the W1 fraction and nonradioactive Nd was eluted with ¹⁴⁰Nd (Fig. 6). Trace metal impurities in the final product are listed in Table S2.

Fig. 5.

A two-step separation of ¹⁴⁰Nd from ¹⁴⁰Pr, where the first step used column 1 with 4.5 gm normal DGA resin and the second step used column 2 with 300 mg normal DGA resin. D refers to the dissolved foil solution loaded into the column. FT refers to the flow through of D from the column. W1-W5 are column washes. The W5 fraction was collected, reformulated and used for step 2(column 2). Step 2 elution fraction, 2-E2 collected  < 20 ppb ¹⁴¹Pr sample concentration and 27.4 ± 2.1% ¹⁴⁰Nd recovery in 400 µL.

Fig. 6.

Elution profile of major target material impurities Ce, Nd and Gd (All three present in dissolved fraction (D)) during the separation process of ¹⁴⁰Nd. Non-radioactive Nd travels with ¹⁴⁰Nd and was eluted in 2-E2. Ce and Gd were removed in W1 fraction.

Radiolabeling

To investigate suitable chelators for the ¹⁴⁰Nd/¹⁴⁰Pr pair, initial radiolabeling studies were performed with Macropa and DOTA chelators. A stock solution of each chelator was prepared with 30 μg/μL concentration. Macropa and DOTA radiolabeling were conducted by serially diluting both chelators with 1 M NH₄OAc, pH 6.5, and incubating with 1.85 MBq (50 μCi) of [¹⁴⁰Nd]Nd³⁺ at 37 °C, 600 rpm for 30 min. The reaction yields were evaluated by radio thin-layer chromatography (radioTLC). The retention factors (Rf) were determined to be 0.3 for [¹⁴⁰Nd]Nd-Macropa, 0.8 for [¹⁴⁰Nd]Nd-DOTA, and 1.0 for free ¹⁴⁰Nd. The molar activities obtained were 74.0 MBq (2.0 mCi)/μmol for [¹⁴⁰Nd]Nd-Macropa and 70.3 MBq (1.9 mCi)/μmol for [¹⁴⁰Nd]Nd-DOTA, each achieving radiochemical yields (RCY) > 95%. (Fig. 7).

Fig. 7.

RadioTLC scan of labeled chelators with > 95% RCY after incubating the reaction mixture at 37 °C for 30 min. (a) [¹⁴⁰Nd]Nd-Macropa(at bottom of TLC; R_f = 0.3) labeling in 1 M NH₄OAc buffer, pH 6.5 (b) [¹⁴⁰Nd]Nd-DOTA (R_f = 0.3) labeling in 1 M NH₄OAc buffer, pH 6.5. (c) Free ¹⁴⁰Nd travels with the solvent front (R_f (free ¹⁴⁰Nd): 1.0).

Stability studies

The stability of [¹⁴⁰Nd]Nd-Macropa and [¹⁴⁰Nd]Nd-DOTA complexes were evaluated in saline, mouse serum, and PBS. Both complexes demonstrated good resistance to decomplexation, remaining intact with > 95% of the radiolabeled species throughout the 5-d incubation period at 37 °C (Fig. 8).

Fig. 8.

Complex stability investigation (a) [¹⁴⁰Nd]Nd-Macropa and (b) [¹⁴⁰Nd]Nd-DOTA in saline, PBS, and mouse serum at 37 ºC incubation up to 5 days.

Phantom imaging

A micro-Derenzo phantom containing hot rods with outer diameters of 1.2, 1.6, 2.4, 3.2, 4.0, and 4.8 mm (Fig. 9a) was filled with 3.7 MBq (100 µCi) of ¹⁴⁰Nd in 5 mL of 0.5 M ammonium acetate. The filled phantom was scanned for a 4-h static scan on PET/CT scanner. The spatial resolution of ¹⁴⁰Nd/¹⁴⁰Pr was distinguishable from this scanner down to 2.4 mm diameter rod size (Fig. 9b).

Discussion

The in vivo PET generator imaging pairs, cerium-134 (¹³⁴Ce (t_1/2: 3.2 d, EC))/lanthanum-134 (¹³⁴La (t_1/2: 6.5 min, 63.6% β⁺⁾) and neodymium-140 (¹⁴⁰Nd (t_1/2: 3.4 d, EC))/praseodymium-140 (¹⁴⁰Pr (t_1/2: 3.4 min, 51.0% β⁺)) hold the potential to be used as imaging surrogates for F-block therapeutic nuclides and potentially exhibit close pharmacokinetics due to similar chemical properties^14,15. In the F- block series, ¹³²La (t_1/2: 4.8 h, 42.1% β⁺) and ¹³³La (t_1/2: 3.9 h, 7.1% β⁺) have been studied as PET imaging theranostic counterparts to the α emitter, ²²⁵Ac (t_1/2: 9.9 d) and Auger electron emitter ¹³⁵La (t_1/2: 19.5 h) from production to preclinical studies by several groups^26–32. Several studies have reported cyclotron production of ¹³²La, ¹³³La, and ¹³⁵La using ^NatBa and enriched Ba (¹³⁴Ba and ¹³⁵Ba) targets and techniques to purify these radionuclides for use in preclinical studies. A recent study carried out by Aluicio-Sarduy et al. used ¹³²La and ²²⁵Ac to radiolabel an alkylphosphocholine analogue (NM600), and this theranostic pair revealed comparative biodistribution behavior²⁶. Similarly, preclinical evaluation of [^132/135La]La-FAPI-2286 by Shirpour et al. and [¹³³La]La-PSMA-I&T by Nelson et al. revealed high cellular and tumor uptake^33,34. Although the short half-lives of ^133/132La present a limitation, these radionuclides have nonetheless proven to be suitable PET imaging surrogates for f-block therapeutic radionuclides. Recently, the ¹³⁴Ce/¹³⁴La in vivo PET generator has been explored by a few groups via ¹³⁴Ce labeled targeting vectors. For example, [¹³⁴Ce]Ce-Macropa-PEG4-YS5, [¹³⁴Ce]Ce-DOTA-trastuzumab, and [¹³⁴Ce]Ce-PSMA-617 were studied in combination with ²²⁵Ac as theranostic matches for targeted alpha therapy studies and [¹³⁴Ce]Ce-cDTPA-ATN-291 for exploratory PET imaging studies^{14–16,35,36}. Although much exploration needed to be done, these studies suggest that ¹³⁴Ce\¹³⁴La illustrate similar results to those with ²²⁵Ac. However, access to ¹³⁴Ce is limited by its production route, which necessitates high-energy proton irradiation. In the United States, production is currently restricted to Los Alamos National Laboratory and Brookhaven National Laboratory, where ^NatLa targets are bombarded with protons of ≥ 100 MeV at 100 μA using linear accelerators¹⁷. Alternatively, ¹⁴⁰Nd production can be accessed at lower proton energies that are available with medium energy cyclotrons.

There are several potential production routes for ¹⁴⁰Nd (Supplemental Fig. S1) including ¹⁴⁰Ce(³He,3n)¹⁴⁰Nd; ¹⁴⁰Ce(α,4n)¹⁴⁰Nd; ¹⁴²Ce(³He,5n)¹⁴⁰Nd; ¹⁴²Ce(α,6n)¹⁴⁰Nd; ¹⁴¹Pr(p,2n)¹⁴⁰Nd; ¹⁴¹Pr(d,3n)¹⁴⁰Nd; ¹⁴¹Pr(t,4n)¹⁴⁰Nd; ¹⁴¹Pr(³He,x)¹⁴⁰Nd; ¹⁴¹Pr(α,x)¹⁴⁰Nd^{18,19,37–39}. The ¹⁴¹Pr(p,2n)¹⁴⁰Nd reaction route has a significantly high cross section with a peak window between 18 and 24 MeV proton energies. The cross-section determinations by Aikawa et al. indicate the peak at 22.2 MeV proton energy with a cross-section value of 956 ± 85 mb (Fig. 2)¹⁸. This 18–24 MeV energy range is suitable for medium energy cyclotrons to produce yields of ¹⁴⁰Nd to sustain R&D and may potentially support clinical applications. The ¹⁴¹Pr(p,2n)¹⁴⁰Nd reaction yielded 205 ± 17 MBq (5.54 ± 0.46 mCi) ¹⁴⁰Nd in a 30 min, 30 µA irradiation. There are several side reactions possible at this proton energy window, including ¹⁴¹Pr(p,n)¹⁴¹Nd (Supplemental Fig. S2a), ¹⁴¹Pr(p,3n)^139mNd (Supplemental Fig. S2c), ¹⁴¹Pr(p,x)¹³⁹Ce (Supplemental Fig. S2d), which will lead to undesirable impurities. The gamma spectrum acquired from the dissolved irradiated ¹⁴¹Pr foil at 18 MeV shows peaks from ¹⁴¹Nd, ¹⁴⁰Pr (Supplemental Fig. S3) and at 24 MeV shows peaks of ^139mNd, ¹³⁹Nd, ¹³⁹Ce, ¹⁴⁰Pr (Supplemental Fig. S4). These findings demonstrate that a 20 MeV proton beam offers a compromise, producing high yields of ¹⁴⁰Nd while minimizing side reactions and associated contaminants, thus proving suitable for the objectives of this study.

As praseodymium (Pr) and neodymium (Nd) are adjacent lanthanides, their chemical properties are very similar, making their separation particularly challenging. To address this, a solid-phase extraction method was developed using normal DGA resin to effectively separate ¹⁴⁰Nd from ¹⁴¹Pr. The normal DGA resin has a Nd/Pr separation factor of 2.5, the highest among LN, LN2, LN3, and branched DGA resins and weight distribution constants (k´) ~ 12 for Pr and ~ 30 for Nd in 3 M HCl as reported by the manufacturer. Optimization of separation parameters, specifically the use of 2.5 M and 2.2 M HCl as wash solvents, enabled effective purification, yielding a recovery of 27.4 ± 2.1% for ¹⁴⁰Nd after a two-column separation. While no radionuclide impurities were observed, co-production of ¹⁴¹Nd is likely but difficult to observe 24 h after irradiation. Elution of Nd with Milli-Q water eliminated the need for post-separation evaporation and reconstitution steps, streamlining the workflow for radiolabeling applications. The recovery of ¹⁴⁰Nd is relatively low because the chemical properties of Nd and Pr have a high degree of similarity, which makes the Nd and Pr separation challenging. In this study, our primary goal was to obtain ¹⁴⁰Nd in high purity, which resulted in relatively low ¹⁴⁰Nd recovery yields due to the overlap in elution profiles of Nd and Pr. Nevertheless, sufficient activity for preclinical applications can be achieved through longer irradiation times and higher beam currents. In future work, optimization of the separation process using alternative resins and solvent systems, as well as liquid–liquid and liquid–solid extraction, will be explored to improve recovery. Efforts to enhance yield while maintaining radionuclidic purity will remain a focus of ongoing development. The ¹⁴¹Pr target material used in this study was commercially available at a high purity level of 99.8%. However, elemental analysis revealed trace impurities, including cerium (Ce), neodymium (Nd), gadolinium (Gd), iron (Fe), copper (Cu), and lead (Pb). The majority of these impurities except Nd were removed during the separation process. The presence of stable Nd may interfere with radiochemical applications that require high molar activity and could warrant additional pre-irradiation purification or recycling strategies in future production cycles.

The chemistry of the resulting ¹⁴⁰Nd was explored with DOTA, a widely used chelator and Macropa, an 18-membered ring chelator. Previous studies have demonstrated efficient radiolabeling of DOTA with ^132/133La and ¹³⁴Ce, as reported by Aluicio-Sarduy et al., Nelson et al., Bailey et al., Bobba et al., and others^15,16,27,34. Similarly, Thiele et al. highlighted the exceptional binding properties of Macropa with ²²⁵Ac, while Bobba et al. reported Macropa as a superior chelator to DOTA, with [¹³⁴Ce]Ce-Macropa-NH₂​ having greater binding efficiency and > 95% integrity retained in saline, PBS, human serum, and rat serum for up to seven days^13,16. Reports from Thiele et al. and Gao et al. indicate that the thermodynamic stability of lanthanide–Macropa complexes decreases with decreasing ionic radius across the series from La³⁺ to Lu³⁺, raising interest in the coordination behavior of Macropa with Nd^40,41. In this report, ¹⁴⁰Nd formed complexes with both DOTA (70.2 MBq/µmol) and Macropa (74.0 MBq/µmol). Although the presence of stable Nd and other possible metal contaminants, including Tb, Dy, Fe, and Cu, may affect the labeling and apparent molar activity, [¹⁴⁰Nd]Nd-Macropa and [¹⁴⁰Nd]Nd-DOTA show excellent stability in mouse serum, PBS and saline for 5 days with > 95% intact complexes. This suggests the suitability of DOTA and Macropa for coordinating ¹⁴⁰Nd, ensuring minimal release of free radionuclide under biological conditions, and supporting their potential use in preclinical in vivo studies.

The spatial resolution achievable with a positron-emitting radionuclide is directly related to its average positron energy. The average positron energy (< E_β+ >) from ¹⁴⁰Pr is 1.07 MeV which falls between ⁶⁸ Ga (< E_β+ >  = 0.83 MeV) and ¹³⁴La (< E_β+ >  = 1.22 MeV) average positron energies. Phantom PET imaging with the ¹⁴⁰Nd/¹⁴⁰Pr generator yielded a spatial resolution of ~ 2.4 mm, comparable to that obtained with ⁶⁸ Ga and within an acceptable range for preclinical imaging. For comparison, ¹⁸F, with a lower positron energy (< E_β+ >  = 0.25 MeV), achieves a superior resolution of ~ 1.4 mm.

¹⁴⁰Nd emits only low-energy X-rays and lacks prominent characteristic gamma-ray peaks. However, its daughter nuclide ¹⁴⁰Pr exhibits distinct gamma-ray signatures (Table 1). Accordingly, quantification of ¹⁴⁰Nd activity was performed indirectly via detection of ¹⁴⁰Pr gamma-ray emissions, which also formed the basis for dose calibrator measurements. The ¹⁴⁰Nd/¹⁴⁰Pr pair constitutes a secular equilibrium generator, and all activity measurements of separation fractions and radio-TLC plates (post-development) were acquired after a minimum of 45 min to ensure equilibrium between parent and daughter had been established.

Although the present work primarily addresses the production and purification of ¹⁴⁰Nd, the inherent challenge of an in vivo PET generator warrants discussion on the impact of the redistribution of the daughter. For ¹⁴⁰Nd-labeled radiotracers, it is the daughter radionuclide, ¹⁴⁰Pr, that emits the positron for PET imaging. However, decay recoil and/or electron ionization may cause dissociation of the daughter from the parent chelate, leading to its redistribution in vivo. Such redistribution can degrade PET image resolution and may not reflect the true biodistribution of the radiotracer. Bauer et al. (2024) reported significant early-phase redistribution of ¹³⁴La in the ¹³⁴Ce/¹³⁴La in vivo PET generator using [¹³⁴Ce]Ce-PSMA-617, although at later time points, due to internalization and membrane turnover, ¹³⁴La’s redistribution decreased considerably but was still evident³⁶. Similarly, Severin et al. demonstrated that with non-internalizing vectors such as DOTA-LM3, daughter redistribution compromises quantitative accuracy, whereas with internalizing vectors like ATN-291, at later time points post-injection, retention of the daughter within tumor cells can enhance the PET signal at the target site^14,42. Bobba et al. and Bailey et al. have likewise highlighted the potential impact of daughter radionuclide redistribution on the accuracy of PET imaging with the in vivo PET generator^15,16.

This work establishes a reliable production and purification strategy for ¹⁴⁰Nd and sets the stage for its application in preclinical theranostic studies. The favorable imaging characteristics and compatibility with established chelators support its use as a PET surrogate for F-block therapeutic radionuclides. The relatively long half-life of ¹⁴⁰Nd (3.4 days) also enables shipment and distribution to distant sites. This enhances its practicality for widespread preclinical studies and potential clinical applications, by increasing accessibility to the theranostic benefits of the ¹⁴⁰Nd/¹⁴⁰Pr generator system. Future efforts will focus on evaluating the in vivo behavior of ¹⁴⁰Nd-labeled biomolecules, developing targeted radiopharmaceuticals, and exploring translation toward first-in-human studies.

Conclusions

The ¹⁴¹Pr(p,2n)¹⁴⁰Nd reaction route was employed for the production of ¹⁴⁰Nd using a 20 MeV proton beam. Yields exceeding 200 MBq (5.5 mCi) were achieved with a 30-min irradiation, demonstrating the feasibility of this reaction for generating ¹⁴⁰Nd in quantities suitable for preclinical research and scalable for clinical applications. A Pr/Nd separation protocol was developed using normal DGA resin, resulting in > 25% recovery of ¹⁴⁰Nd and residual ¹⁴¹Pr levels below 20 ppb. Radiolabeling of Macropa and DOTA chelators with ¹⁴⁰Nd was achieved with > 95% radiochemical yield (RCY) and stable complex formation. Phantom PET imaging with a 4-h scan duration demonstrated spatial resolution down to 2.4 mm, confirming the potential of the ¹⁴⁰Nd /¹⁴⁰Pr in vivo PET generator for high-resolution molecular imaging applications.

Author contributions

M.L.G. performed all experiments, drafted the main manuscript, and prepared all figures and tables. S.J.C. led the initial radiochemical separation and targetry efforts, contributed to experimental work and data analysis, and provided editorial input on the manuscript. V.T., S.S., and H.A.H. contributed to the experimental design, assisted with data interpretation, and manuscript revisions. S.L.S. helped the phantom studies and provided input on study design, data analysis, and manuscript editing. S.E.L. conceptualized the study, provided guidance on experimental design and data analysis, supervised the project, and contributed to manuscript revisions. All authors reviewed and approved the final version of the manuscript.

Funding

This research was supported by the U.S. Department of Energy Isotope Program under grant DESC0020197 (Principal Investigator: Suzanne Lapi). Imaging studies were conducted with support from the Small Animal Imaging Core, funded by the O’Neal Comprehensive Cancer Center (P30CA013148).

Data availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.