Automated radiosynthesis of [68Ga]Ga‐PSMA‐11 and [177Lu]Lu‐PSMA‐617 on the iPHASE MultiSyn module for clinical applications

Prostate‐specific membrane antigen (PSMA)‐targeted imaging and therapy of prostate cancer using theranostic pairs is rapidly changing clinical practice. To facilitate clinical trials, fully automated procedures for the radiosyntheses of [68Ga]Ga‐PSMA‐11 and [177Lu]Lu‐PSMA‐617 were developed from commercially available precursors using the cassette based iPHASE MultiSyn module. Formulated and sterile radiopharmaceuticals were obtained in 76 ± 3% (n = 20) and 91 ± 4% (n = 15) radiochemical yields after 17 and 20 min, respectively. Radiochemical purity was always >95% and molar activities exceeded 792 ± 100 and 88 ± 6 GBq/μmol, respectively. Quality control showed conformity with all relevant release criteria and radiopharmaceuticals were used in the clinic.

GMP grade PSMA-617 was purchased from ABX (Radeberg, Germany Thin-layer chromatography (TLC) of labelled PSMA-11 samples was performed using aluminium-backed silica gel 60 F 254 strips (Merck, Darmstadt, Germany) and a 1:1 (v/v) mixture of 1 M ammonium acetate and methanol as a mobile phase. Instant TLC (iTLC) of labelled PSMA-617 samples was performed using glass microfibre iTLC-SG chromatography paper strips (Agilent, CA, USA) and 1.9% disodium hydrogen citrate as a mobile phase. Reversed-phase HPLC (RP-HPLC) was performed on a Shimadzu LC-20 equipped with a SPD-20A UV-Vis detector and a LabLogic Flow-RAM scintillation detector using a Phenomenex Kinetex C18 5 μm 150 × 4.6 mm column. Mobile phases were aqueous 0.08% TFA (A), acetonitrile (B), aq. 0.1% TFA (C) and 0.1% TFA in acetonitrile (D). Ga-PSMA-11 samples were analysed using a gradient of 5-80% D in C (v/v) over 10 min at a flow rate of 1.5 ml/min. Lu-PSMA-617 samples were analysed using a gradient of 15-35% B in A (v/v) over 15 min followed by 35-95% B in A (v/v) over 1 min and maintained for 3 min at a flow rate of 0.9 ml/min. Formulated material was analysed using a Charles River bacterial endotoxin testing unit.

| Automated production of [ 68 Ga]Ga-PSMA-11
Upon start-up and installation of the hardware cassette excluding reagents, automated pressure testing was performed on all connections. The manifolds, reactor, SPE cartridge and reagent vial spikes were flushed with argon gas. Subsequently, reagents were installed as outlined in Figure 1. Vials containing ethanol (3.5 ml), saline (10 ml) and water (10 ml) were pushed onto their respective spikes on Manifold 4, and a solution of PSMA-11 (10 μg, 9.88 nmol) in 0.25 M sodium acetate (1.6 ml) was added to the reactor via its centre port. A syringe containing 0.1 M HCl (5 ml) that was connected to the inlet of an Eckert & Ziegler IGG100 Gallium generator was placed in one of the syringe driver arms. The generator outlet was connected to Manifold 3. Finally, a 10-ml glass vial equipped with a vented 0.22-μm sterile filter was connected to manifold 2. The reagent vials were then pressurised with inert gas, and the SPE cartridge was conditioned with ethanol from the reagent vial (approximately 2 ml), washed with water (5 ml) and flushed dry. Ethanol (approximately 0.2 ml) was transferred into the reactor, and gallium-68 was eluted from the generator using 0.1 M HCl in 0.3-ml increments over 1 min of elution time. The first fraction (1.5 ml) was discarded, and the second fraction (3.5 ml) was collected into the preheated reactor. The reaction mixture was heated at 95 C for 5 min before being transferred over the SPE cartridge. Water (5 ml) was transferred first into the reactor and then loaded onto the SPE cartridge. The product was eluted from the SPE cartridge with ethanol (approximately 0.5 ml) using Syringe Driver 2 and diluted with saline (1.5 ml). The resulting solution was passed through a 0.22-μm sterile filter and collected into the product vial. The reactor, SPE cartridge and transfer tubing from Manifold 2 to the product vial were flushed with further saline (5 ml) to complete formulation of the product.

| Automated production of [ 177 Lu] Lu-PSMA-617
Automated pressure testing of all connections was performed with the manifolds, Transfer Syringe 3 (Manifold Position 6) and the reactor installed, and the reactor was flushed with inert gas. Subsequently, reagent vials and syringes were installed as described in Table 1.
[ 177 Lu]LuCl 3 in 0.04 M HCl (approximately 1 ml) was transferred from the Lu-177 vial into the reactor. One millilitre of a 0.4 M sodium acetate solution containing PSMA-617 (125 μg, 103 nmol) and 4-mg gentisic acid was transferred from Syringe 1 into the Lu-177 vial, remained there for 10 s and was transferred into the preheated reactor vial. After flushing the transfer line with inert gas, the reaction was heated to 95 C for 15 min. During the labelling reaction, half of the contents in Syringe 2, an aqueous solution containing 500-mg sodium ascorbate and 1-mg DTPA in 10-ml WFI, were drawn into  Figure 1. Gallium-68 was obtained from an Eckert & Ziegler IGG100 Gallium generator by elution with 5-ml 0.1 M hydrochloric acid. The first fraction (1.5 ml) was automatically discarded to reduce germanium-68 content and elution volume to enable easier buffering. The second fraction (3.5 ml) containing Ga-68 was collected into the reactor vial and used without prepurification or preconcentration. The reactor was charged with a small amount of ethanol prior to addition of the radioisotope to moderate radiolytic breakdown of reagents. Addition of the second generator fraction to the reactor vial containing PSMA-11 in sodium acetate resulted in a reaction pH of 4. Purification and reformulation of [ 68 Ga]Ga-PSMA-11 was performed using a polymer-based SPE cartridge (Phenomenex Strata-X) before being sterile filtered into a sterile product collection vial ( Figure S1 a-d).

| Development of automated [ 177 Lu] Lu-PSMA-617 synthesis protocol
Quantitative complexation of 177 Lu 3+ with DOTA was achieved in 0.4 M sodium acetate solution at 95 C after 15 min at pH 4.5-5. The reaction was stabilised using gentisic acid, and the final product was formulated in sodium ascorbate containing DTPA. The process was automated using the cassette based iPHASE MultiSyn module (Figure 3). Fluid transfers were performed using either gas pressure/vacuum or the two available syringe drivers. Figure 4 shows a schematic overview of the installed synthesis cassette consisting of two manifolds, one reactor vial, three syringes and a product collection vial. The reagents and transfer lines were set up as described in Table 1.
Initially, we experienced large radioactivity losses due to residual [ 177 Lu]LuCl 3 remaining in the Lu-177 vial ( Figure 5). This was improved by flushing the Lu-177 vial with the precursor solution following isotope transfer into the reactor. This led to improved Lu-177 recovery in the reactor. Residual radioactivity remaining in the Lu-177 vial decreased by 55% from 15.0% to 6.7 ± 1.8% (n = 3). Losses in the Lu-177 vial transfer tubing were reduced by 94% from 6.6% to 0.4 ± 0.1% (n = 3).  Table 1 Losses during the recovery of [ 177 Lu]Lu-PSMA-617 from the reactor were addressed by performing a saline wash of the reactor during product formulation resulting in 0.5 ± 0.0% (n = 3) of radioactivity lost in the reactor. Residual radioactivity on the sterile filter and in the manifolds remained constant at 1.2 ± 0.5% (n = 4) and 0.4 ± 0.1% (n = 4), respectively. This resulted in an overall increase in radiochemical yield of [ 177 Lu]Lu-PSMA-617 from 75% to 91.1 ± 3.8% (n = 14) at end of synthesis (EOS). The total synthesis time was 20 min including formulation and sterile filtration ( Figure S3 a-d).

| Quality control for clinical use
Quality control was performed on [ 68 Ga]Ga-PSMA-11 and [ 177 Lu]Lu-PSMA-617 according to the criteria outlined in Table 2. The reported results were obtained from 12 and 15 clinical productions, respectively. Sterile filters were tested for membrane filter integrity which was always >50 psi. Where applicable, solvent analysis was performed which showed <10% ethanol.
Production of [ 177 Lu]Lu-PSMA-617 proceeded in 91% (d.c.) radiochemical yield providing the product in a volume of 9.4 ml. Molar activity was not a mandatory F I G U R E 4 Schematic overview of iPhase MultiSyn cassette and reagent set-up for radiosynthesis of [ 177 Lu]Lu-PSMA-617 as described in Table 1 F I G U R E 5 Residual Lu-177 radioactivity before and after optimisation. Data shown as mean ± SD (n = 1-3) release criterion for [ 177 Lu]Lu-PSMA-617. Based on the starting amount of PSMA-617 (103 nmol) and the radiochemical purity, the molar activity was estimated to be ≥88 ± 6 GBq/μmol (EOS). Radiochemical purity by HPLC was always ≥95% and unreacted lutetium-177 (iTLC, R f = 0.9-1.0) was always ≤2.4% ( Figure S4 a-d).
Radionuclidic identity and purity were confirmed for both constructs. Sterility and bacterial endotoxin testing showed no positive results over the limit of detection. All productions fulfilled the quality control release criteria and were used in the clinic.

| CONCLUSIONS
In summary, the syntheses of a diagnostic and therapeutic pair of PSMA-targeting agents were automated using the same commercially available synthesiser. Production yields were comparable with other reported fully automated procedures. Both products passed quality control criteria without fail and are being used in clinical trials at the Austin Hospital.