*2.4. Radiosynthesis and Characterization of [165Er]PSMA-617*

PSMA-617 (MedChemExpress, Monmouth Junction, NJ, USA) was dissolved in 18 MΩ·cm water to a concentration of 0.1 µg/µL and distributed into 10 aliquots that were stored at –20 ◦C. A sodium acetate solution (1 M, 99.995% trace metals basis, Aldrich) was buffered with hydrochloric acid until a pH of 5.7 was obtained. Er-165 (41–52 MBq in 70–80 µL 0.01 M HCl) was added to a solution of PSMA-617 in water (1 nmol in 10 µL, 2 nmol in 10 µL, or 5 nmol in 25 µL), NaOAc aq. (1M, pH 5.7, 50 µL), and L-ascorbic acid (0.3–0.4 mg, 25 µL, ≥99.9998 trace metals basis, Honeywell-Fluka), which were reacted at 80 ◦C and 1000 rpm for 30 min. The reaction was diluted in 18 MΩ·cm water (10 mL) and loaded onto a pre-equilibrated C-18 Plus Light cartridge; after rinsing with water (10 mL), elution was performed with pure ethanol (700 µL). An aliquot (25 µL) was diluted with 18 MΩ·cm water (25 µL) and set aside for quality control analysis. Analytical HPLC was performed on [165Er]PSMA-617 using an Agilent 1260 Infinity II module coupled with an inline radioactivity detection system similar to that described in Section 2.2 but with electronic analog pulse converted to voltage (Model 106, Lawson Labs Inc., Malvern, PA, USA) and logged using an Agilent 1200 universal interface box. The separation was performed on reverse-phase C18 column (InfinityLab Poroshell 120 EC-C18, 4.6 × 100 mm, 2.7 µm, Agilent Technologies, Santa Clara, CA, USA) using a linear gradient (95 to 20% over 15 min) of 0.1% trifluoroacetic acid (Thermo Scientific in 18 MΩ·cm water) in acetonitrile (HPLC-grade, Fisher Scientific, Pittsburgh, PA, USA) at 1 mL/min.

In vitro stability of [165Er]PSMA-617 was investigated in the presence of L-ascorbic acid and freshly prepared normal human serum prepared from lyophilized powder (009- 000-001, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) reconstituted using phosphate buffered saline (2.0 mL, PBS, Lonza Bioscience). Dry [165Er]PSMA-617 (3.1 MBq) and L-ascorbic acid (0.6 mg) were dissolved in serum (300 µL) and incubated at 37 ◦C for 12 h. Aliquots of the [165Er]PSMA-617 complex were assessed by analytical HPLC at t = 1 and 12 h. The chromatograms were analyzed by integration of the [165Er]PSMA-617 peak compared to all radioactive peaks (free erbium or decomposition products) in the chromatogram.

The distribution coefficient (logD value) of [165Er]PSMA-617 was determined using a 1:1 (*v*/*v*) solution of n-octanol (Alfa Aesar, Haverhill, MA, USA) and PBS according to previously reported methods [11]. A sample of [165Er]PSMA-617 (8.2–51 MBq,

2.8–8.4 MBq/nmol) was dried and diluted with PBS and n-octanol (700 µL each). The solution was vigorously agitated for 5 min before being centrifuged at 1000 rpm for 5 min. An aliquot of PBS and n-octanol was analyzed by HPGe gamma spectrometry under identical geometries and the ratio of decay-corrected net counts per minute was used to determine the distribution coefficient. The PBS aliquot required 4–5 days of decay before acquisition of an HPGe spectrum with acceptable (<5%) dead-time. Uncertainty in the LogD value was calculated by propagation of error associated with HPGe counting statistics and in the half-life of <sup>165</sup>Er (10.36 <sup>±</sup> 0.04 h [47]). The logD experiment was repeated for five independent preparations of [165Er]PSMA-617 and logD reported as average and standard deviation of results. of PBS and n-octanol was analyzed by HPGe gamma spectrometry under identical geometries and the ratio of decay-corrected net counts per minute was used to determine the distribution coefficient. The PBS aliquot required 4–5 days of decay before acquisition of an HPGe spectrum with acceptable (<5%) dead-time. Uncertainty in the LogD value was calculated by propagation of error associated with HPGe counting statistics and in the half-life of 165Er (10.36 ± 0.04 h [47]). The logD experiment was repeated for five independent preparations of [165Er]PSMA-617 and logD reported as average and standard deviation of results. **3. Results**

617 peak compared to all radioactive peaks (free erbium or decomposition products) in

The distribution coefficient (logD value) of [165Er]PSMA-617 was determined using a 1:1 (*v*/*v*) solution of n-octanol (Alfa Aesar, Haverhill, MA, USA) and PBS according to previously reported methods [11]. A sample of [165Er]PSMA-617 (8.2–51 MBq, 2.8–8.4 MBq/nmol) was dried and diluted with PBS and n-octanol (700 µL each). The solution was vigorously agitated for 5 min before being centrifuged at 1000 rpm for 5 min. An aliquot

#### **3. Results** 3a. Ho target preparation and irradiation

1.

the chromatogram.

#### *3.1. Ho Target Preparation and Irradiation* Cold rolling, disc cutting, and spot-welding methods were well suited for the fabri-

*Molecules* **2021**, *26*, x FOR PEER REVIEW 6 of 15

Cold rolling, disc cutting, and spot-welding methods were well suited for the fabrication of cyclotron irradiation targets of tightly controlled dimensions and mass from a variety of commercial holmium metal sources. The malleability of holmium allowed for the dramatic thinning of metal foils through rolling. While thickness reduction by factors of two or three was readily achieved, when an eight-fold change in thickness was attempted, cracking around the edges was observed, as shown in supplementary Figure S1. Spot-welded holmium was well adhered to the tantalum backing and withstood proton irradiation at all investigated intensities with only minor discoloration, as shown in Figure 1. cation of cyclotron irradiation targets of tightly controlled dimensions and mass from a variety of commercial holmium metal sources. The malleability of holmium allowed for the dramatic thinning of metal foils through rolling. While thickness reduction by factors of two or three was readily achieved, when an eight-fold change in thickness was attempted, cracking around the edges was observed, as shown in supplementary Figure S1. Spot-welded holmium was well adhered to the tantalum backing and withstood proton irradiation at all investigated intensities with only minor discoloration, as shown in Figure

**Figure 1.** Representative image of a 7.9 mm ø, 270 µ m-thick, 106 mg holmium disc spot-welded to tantalum, before and after 68 min, 40 µA·h PETtrace irradiation. **Figure 1.** Representative image of a 7.9 mm ø, 270 µm-thick, 106 mg holmium disc spot-welded to tantalum, before and after 68 min, 40 µA·h PETtrace irradiation.

For the PETtrace cyclotron, a proton irradiation energy of 12.5 MeV centers the <sup>165</sup>Ho(*p,n*) <sup>165</sup>Er excitation function peak (see supplementary Figure S2 [22–24]) within the energy loss window of the protons traversing a 200–300 µm-thick holmium target. Experimental end-of-bombardment (EoB) 165Er physical yields were measured via attenuationcorrected HPGe of target discs or dose calibrator measurements of dissolved target aliquots, or CX elution fractions (Table 2). The yields show a significant dependence on the For the PETtrace cyclotron, a proton irradiation energy of 12.5 MeV centers the <sup>165</sup>Ho(*p,n*) <sup>165</sup>Er excitation function peak (see supplementary Figure S2 [22–24]) within the energy loss window of the protons traversing a 200–300 µm-thick holmium target. Experimental end-of-bombardment (EoB) <sup>165</sup>Er physical yields were measured via attenuation-corrected HPGe of target discs or dose calibrator measurements of dissolved target aliquots, or CX elution fractions (Table 2). The yields show a significant dependence on the holmium target dimensions and the irradiating cyclotron. Calculated using literature cross-sections [23,24], a 1–2 h proton irradiation of 275 µm-thick holmium results in physical <sup>165</sup>Er yields of 50 MBq·µA−<sup>1</sup> ·h <sup>−</sup><sup>1</sup> at 12.5 MeV and 39 MBq·µA−<sup>1</sup> ·h <sup>−</sup><sup>1</sup> at 11 MeV.

Experimental <sup>165</sup>Er physical yields were significantly lower than these theoretical maxima, likely because the cyclotron-integrated charge includes protons impinging on the target outside the holmium diameter. This is especially problematic for the PETtrace cyclotron, which has an oblong beam spot with full width at half maxima of 11 and 8.7 mm, as measured by autoradiography of irradiated aluminum discs. The RDS-112 cyclotron provides a significantly smaller beam spot, resulting in higher overall <sup>165</sup>Er physical yield, despite the lower irradiation energy and smaller target diameter. However, the RDS-112 cyclotron is limited to maximum irradiation current of 20 µA, a factor of 2 below that routinely used with the PETtrace cyclotron.


**Table 2.** Erbium-165 physical yields for different Ho targets and irradiation configurations.
