A kit formulation for the preparation of [89Zr]Zr(oxinate)4 for PET cell tracking: White blood cell labelling and comparison with [111In]In(oxinate)3

Background Advances in immunology and cell-based therapies are creating a need to track individual cell types, such as immune cells (neutrophils, eosinophils, chimeric antigen receptor (CAR) T cells, etc.) and stem cells. As the fate of administered cells remains largely unknown, nuclear imaging could determine the migration and survival of cells in patients. [89Zr]Zr(oxinate)4, or [89Zr]Zr-oxine, is a radiotracer for positron emission tomography (PET) that has been evaluated in preclinical models of cell tracking and could improve on [111In]In-oxine, the current gold standard radiotracer for cell tracking by scintigraphy and single-photon emission computed tomography (SPECT), because of the better sensitivity, spatial resolution and quantification of PET. However, a clinically usable formulation of [89Zr]Zr-oxine is lacking. This study demonstrates a 1-step procedure for preparing [89Zr] Zr-oxine and evaluates it against [111In]In-oxine in white blood cell (WBC) labelling. Methods Commercial [89Zr]Zr-oxalate was added to a formulation containing oxine, a buffering agent, a base and a surfactant or organic solvent. WBC isolated from 10 human volunteers were radiolabelled with [89Zr]Zr-oxine following a clinical radiolabelling protocol. Labelling efficiency, cell viability, chemotaxis and DNA damage were evaluated in vitro, in an intra-individual comparison against [111In]In-oxine. Results An optimised formulation of [89Zr]Zr-oxine containing oxine, polysorbate 80 and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) was developed. This enabled 1-step radiolabelling of oxine with commercial [89Zr]Zr-oxalate (0.1–25 MBq) in 5 min and radiotracer stability for 1 week. WBC labelling efficiency was 48.7 ± 6.3%, compared to 89.1 ± 9.5% (P < 0.0001, n = 10) for [111In]In-oxine. Intracellular retention of 89Zr and cell viability after radiolabelling were comparable to 111In. There were no significant differences in leukocyte chemotaxis or DNA damage between [89Zr]Zr-oxine or [111In]In-oxine. Conclusions, advances in knowledge and implications for patient care Our results demonstrate that [89Zr]Zr-oxine is a suitable PET alternative to [111In]In-oxine for WBC imaging. Our formulation allows rapid, stable, high-yield, single-step preparation of [89Zr]Zr-oxine from commercially available 89Zr. This will facilitate the clinical translation of cell tracking using [89Zr]Zr-oxine.


Background
Recent developments in immunology and cell-based therapies are creating a need to track the migration of individual cell types. For example, neutrophils and eosinophils in asthma and chronic obstructive pulmonary disease were shown to have different distribution patterns [1][2][3][4], and there is considerable interest in tracking T-cells [5][6][7] and dendritic cells [8] in cancer and auto-immune diseases, or stem cells in regenerative medicine [9]. There is an emerging consensus [10,11], ac-companied by recognition by drug regulators [12], that development and trials of cell-based therapies should be accompanied by methods to determine the location, survival, proliferation and differentiation of administered cells both in animal models and human subjects. Imaging the in vivo trafficking of cells radiolabelled prior to administration is a clinically acceptable, informative, non-invasive approach that can be used in human subjects, is not limited by depth and requires no biopsy. Gamma scintigraphy, and more recently single-photon emission computed tomography (SPECT), with autologous leukocytes labelled with gamma-emitting radionuclides ( 111 In, 99m Tc) has been a routine part of nuclear medicine since the 1970s [13] to detect sites of infection and/or inflammation [14,15]. Further developments in cell-based therapies [9,[16][17][18][19] will require detection of small lesions and low numbers of cells, as well as better quantification, all of which could be achieved by positron emission tomography (PET).
One of the main obstacles to the use of [ 89 Zr]Zr-oxine-labelled cells in the clinic is its cumbersome synthesis and the absence of a simple one-step formulation ("kit"), such as exists for [ 99m Tc]Tc-D,Lhexamethylene-propyleneamine oxime ([ 99m Tc]Tc-HMPAO, [ 99m Tc]Tcexametazime). This greatly restricts its clinical translation and commercial appeal for routine use in clinical trials. Furthermore, while some studies have investigated the effect of [ 89 Zr]Zr-oxine labelling on cell function [34,35], only one preclinical study has directly compared [ 89 Zr]Zr-oxine to the gold standard [ 111 In]In-oxine [30]. Here we present the first kit formulation, and a simple, good manufacturing practices (GMP)-compliant, clinically translatable protocol for using it for rapid, one-step preparation of [ 89 Zr]Zr-oxine, [ 64 Cu]Cu-oxine and [ 68 Ga]Gaoxine for radiopharmaceutical applications, which will greatly enhance access of hospitals to cell tracking by PET in clinical diagnosis and trials of cell-based therapy. We demonstrate its application in radiolabelling human white blood cells (WBC) with [ 89 Zr]Zr-oxine following a clinical protocol and provide a direct, intra-individual comparison with [ 111 In] In-oxine.
Product formation was confirmed by radio thin-layer chromatography (radioTLC) on instant TLC (ITLC)-SG paper (Macherey-Nagel) or Whatman no.1 paper (GE Healthcare) using 100% ethyl acetate (EtOAc) as the mobile phase. ITLC plates were read using a Mini-Scan™ radioTLC linear scanner (LabLogic Systems) equipped with a β + probe (LabLogic B-FC-3600). Radiochemical purity of the final product was calculated as the activity associated with the [ 89 Zr]Zr-oxine peak as a percentage of the total detected activity on the chromatogram.
To study the recovery of radiotracer from the vial, 10 μL aliquots were taken immediately after addition of [ 89 Zr]Zr-oxalate and after 15, 30, 60, 120 min, 24, 48, 72 and 168 h. The aliquots were gammacounted 7 days after addition of [ 89 Zr]Zr-oxalate and percentage recovery determined as the counts in each sample divided by the counts in the sample taken immediately after addition. For stability studies, samples were left at room temperature (RT) in the dark and analysed by radioTLC over 7 days. A diluted kit was obtained by further adding 900 μL H 2 O 5 min after addition of [ 89 Zr]Zr-oxalate to the kit formulation.
The partition coefficient (logD) of the [ 89 Zr]Zr-oxine and [ 111 In]Inoxine formulations was determined by adding 10 μL (approx. 1 MBq) of each formulation to 1 mL of a 50:50 presaturated mixture of either 1-octanol and water or 1-octanol and phosphate-buffered saline (PBS), then vortexing for 5 min. Phase separation was then accelerated by brief centrifugation. 100 μL were taken from each phase and gammacounted.

Cell isolation
White blood cell isolation was performed in accordance with guidelines for radiolabelling WBC with [ 99m Tc]Tc-exametazime and [ 111 In]Inoxine [43,44]. Briefly, peripheral venous blood (50-55 mL) was collected from healthy, male (n = 5) and female (n = 5) donors aged 22-32, in anticoagulant citrate dextrose solution A (ACD-A) blood collection tubes (BD Vacutainer #366645) using 20G needles, on two separate occasions for each donor. Cell-free plasma (CFP) was obtained by centrifuging 10-15 mL blood at 2000g for 10 min. For WBC isolation, 45 mL blood was mixed with 7 mL of HES200/0.5 (10% wt./vol. in sterile saline) and centrifuged at 8g for 45 min at room temperature. Platelets were depleted by washing the WBC layer twice with Ca 2+ /Mg 2+ -free PBS (with 10 min centrifugation at 150g). The remaining cell pellet was re-suspended in 3 mL PBS for radiolabelling. 20 min at RT with gentle swirling every 5 min. As an additional control, an aliquot of WBC was incubated with PBS only. Cells were then diluted with 50 mL PBS and centrifuged at 200g for 10 min. Supernatants and cell pellets were measured in a dose calibrator (CRC-25R, Capintec). The cells were suspended in CFP or assay medium (RPMI-1640 supplemented with 1% human serum, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin) for further experiments. Viability was assessed using the Trypan Blue dye exclusion method. Cell labelling efficiency (LE%) was calculated as: LE % ð Þ ¼ activity of cell fraction activity of cell fraction þ activity of combined supernants : For radiotracer retention studies, radiolabelled WBC were suspended in autologous CFP, in triplicate in a 24-well plate and incubated at 37°C. Cells were collected after 4 h or 24 h and viability was determined as above. Cells were diluted with PBS, centrifuged at 200g for 10 min, and supernatants and pellets were measured in a dose calibrator to determine retention using the formula above.

Chemotaxis assay
After radiolabelling, cell pellets were subjected to red blood cell (RBC) lysis by hypotonic shock. Cells were resuspended in 4.5 mL cold H 2 O for 30 s, after which isotonicity was restored by addition of 0.5 mL 10× PBS. Lysed RBC were removed by centrifugation and the remaining WBC were resuspended in assay medium at 3.5 × 10 6 cells/mL. The bottom wells of a chemotaxis plate were filled with 30 μL of assay medium, with or without 10 nM fMLP. On the top wells, 20 μL of cell suspension were then plated in triplicate and incubated for 45 min at 37°C. Remaining cells in the top wells were removed, replaced by 40 μL of 5 mM EDTA in PBS to detach cells adhering to the membrane and the plate was incubated for 30 min at 4°C. The top wells were emptied, the plate was centrifuged at 150g for 5 min and the cells in the bottom wells were counted using a haemocytometer. The chemotaxis index (CI) was calculated by dividing the number of WBC in the wells containing fMLP by the number of WBC in the wells containing medium only.

Statistical analysis
Each subject provided WBC on two separate occasions (at least 1 week apart), once for [ 89 Zr]Zr-oxine and once for [ 111 In]In-oxine labelling, enabling differences between groups to be evaluated by Student's two-tailed, paired t-test or Wilcoxon's matched pairs signed-rank test, as appropriate. When additional factors were considered, analysis was performed using a repeated-measures Mixed Model (MM) in Prism v8.2 (GraphPad Software Inc.), with Tukey's correction for multiple pairwise comparisons unless otherwise specified. Exact significance values are reported in each figure.

QC method development
To evaluate the radiochemical yield of [ 89 Zr]Zr-oxine, we optimised a simple radiopharmaceutical QC method. Initial measures by radioTLC on silica-gel impregnated glass fibre (ITLC-SG) using ethyl acetate showed no migration of unchelated 89 Zr (R f = 0, Fig. 1a), whereas [ 89 Zr]Zr-oxine frequently showed marked streaking (Fig. 1c), possibly because the interaction of silanol groups with 89 Zr leads to the dissociation of the metastable [ 89 Zr]Zr-oxine complex during migration. Therefore, ITLC-SG is not an acceptable support for the QC of this radiotracer. In contrast, TLC on Whatman no. 1 paper shows a clear separation between unchelated 89 Zr (R f = 0, Fig. 1b) and [ 89 Zr]Zr-oxine (R f = 1, Fig. 1d). A strip length of 6 cm was found to provide clear separation between the two species, with a migration time of less than 8 min (Suppl. Fig. S1).

Kit formulation optimisation
A kit formulation for [ 89 Zr]Zr-oxine requires a base to neutralise the acidic solution of 89 Zr supplied and a buffer to maintain the solution at pH 7-8. It was found that 100 μL of HEPES-buffered formulation (containing 50 μg 8-hydroxyquinoline and 52.5 μmol NaOH) was capable of buffering (pH ≥ 7.0) a maximum of 18 μL of [ 89 Zr]Zr-oxalate solution (153 mM oxalate in final product). [ 89 Zr]Zr-oxine formulated in 1 M HEPES buffer (pH 7.9) was found to rapidly adhere to glass vessels, with only 46% of the added activity recoverable from the vial 15 min after addition of [ 89 Zr]Zr-oxalate. Including 5% EtOH in the solution delayed this phenomenon but did not prevent it, with 46% recoverable activity after 1 h and less than 6% after 24 h (Fig. 2a). In contrast, addition of 1 mg/mL polysorbate 80 prevented adhesion to the glass vessel and resulted in 98.7% recovery of activity up to 1 week after addition of [ 89 Zr]Zr-oxalate (Fig. 2a, b). Reducing the concentration of polysorbate 80 resulted in minor losses of product. Replacing NaOH with NaHCO 3 led to slower formation of [ 89 Zr]Zr-oxine and reduced yields (Fig. 2c). The optimised formulation, a 100 μL solution at pH 7.9-8.0 containing 50 μg oxine, 1 M HEPES, NaOH and 1 mg/mL polysorbate 80, was stable for 7 days in concentrated format (Fig. 2d). The preparation and use of the kit formulation are schematically represented in Fig. 3 (Fig. 2e). The percentage of "free", unchelated oxine in a 20 MBq batch of [ 89 Zr]Zroxine was calculated to be >99.86% (Supplementary Material, Table S1). The partition coefficient of the formulated [ 89 Zr]Zr-oxine was found to be 0.80 ± 0.26 in water ([ 111 In]In-oxine: 1.09 ± 0.03, n = 3, P = 0.18) and 0.60 ± 0.06 in PBS ([ 111 In]In-oxine: 0.72 ± 0.11, n = 3, P = 0.08; Suppl. Fig. S2).
As well as zirconium-89, the optimised oxine formulation was shown to form lipophilic complexes of oxine with gallium-68, copper-64 and indium-111 in high yield (Suppl. Fig. S3), and therefore may be useful for labelling cells or nanomedicines with these radionuclides.

Labelling efficiency and comparison with [ 111 In]In-oxine
A detailed standard operating procedure for labelling WBCs with the kit formulation of [ 89 Zr]Zr-oxine is provided in the Supplementary Material. Using the optimised kit, we evaluated the labelling of human WBCs with [ 89 Zr]Zr-oxine, in comparison with [ 111 In]In-oxine formulated similarly to the commercial product [42]. WBCs were obtained from 10 healthy donors, following a clinical protocol for WBC labelling [44], and an intra-individual comparison of labelling with [ 89 Zr]Zroxine and [ 111 In]In-oxine was performed. In this study we have attempted to replicate real-world conditions as closely as possible, rather than standardise every single parameter (i.e. by using the same number of WBCs from each donor) or splitting WBC batches for labelling with each radiotracer. Labelling with each radiotracer was performed on separate occasions for each donor to comply with local ethical limits regarding blood donations.
The functionality of radiolabelled WBCs was tested using an in vitro chemotaxis assay, where the number of WBCs migrating in response to 10 nM fMLP was measured. The chemotaxis index of 89 Zr-labelled WBCs was 2.7 ± 1.4 (n = 9), c.f. 3.4 ± 1.9 (n = 10) for 111 In-labelled WBCs and 3.0 ± 1.0 (n = 10) for non-labelled WBCs (Fig. 5). There were no significant differences between the groups. Importantly, the chemotaxis indexes were all >1, demonstrating an active migration towards fMLP. There was no apparent trend relating the amount of activity per cell and the chemotaxis index (Suppl. Fig. S4).
To determine whether certain subtypes of WBC had preferential uptake of [ 89 Zr]Zr-oxine, radiolabelled WBC were stained with fluorescent monoclonal antibodies, automatically sorted and gamma-counted. For practical reasons, only a small fraction of the radiolabelled cells was sorted. There were large differences in average activity per cell (range 1-4 kBq/10 6 cells, Suppl. Fig. S5) between donors despite comparable labelling efficiencies (fraction of activity in cell pellet) because a fixed patient dose of [ 89 Zr]Zr-oxine (20 MBq) was used whereas the number of WBCs isolated from each donor was highly variable. Results are therefore expressed as relative uptake of 89 Zr per cell in each population (Fig. 6), taking neutrophils as reference (1.00). The relative activity of 89 Zr per cell was 1.06 ± 0.04 in lymphocytes, 0.63 ± 0.25 in eosinophils, 0.74 ± 0.10 in NK cells, 0.87 ± 0.05 in B cells, 0.85 ± 0.06 in monocytes, 0.99 ± 0.33 in platelets and 1.04 ± 0.32 in erythrocytes.
The effect of radiolabelling on DNA damage, assessed in terms of double-strand break formation, was determined by counting the   (Fig. 7). The differences between radiolabelled and non-radiolabelled samples were significant for both [ 89 Zr]Zr-oxine and [ 111 In]In-oxine, however there was no statistical difference between the two radiotracers. There was no apparent trend relating the number of foci per nucleus and the activity per cell after radiolabelling (Suppl. Fig. S6).

Discussion
[ 89 Zr]Zr-oxine was previously shown to be a useful cell labelling agent for PET imaging in animal models by our group and others [30][31][32][33][34][35][36]. In the clinic, the radiolabelling of mixed WBCs for infection imaging is one of the most common application of cell labelling, with the greatest collective experience within the nuclear medicine community. For the clinical translation of this radiotracer, we believe it is crucial to (a) simplify its synthesis to the point that it is usable with as few steps as possible by radiopharmacy operators, and (b) perform a direct comparison with the gold standard [ 111 In]In-oxine in the clinically relevant model of radiolabelled WBCs.
Our previous synthesis of [ 89 Zr]Zr-oxine from [ 89 Zr]Zr-oxalate involved the use of chloroform as a solvent and subsequent evaporation and redissolution in a small amount of dimethyl sulfoxide or ethanol [29]. While relatively simple from a research perspective, this is far from ideal for the clinic as it involves several steps, requires meticulous precision during the neutralisation step and involves organic solvents (requiring further QC tests) and solvent evaporation, with an overall radiochemical yield of 60-80%. Sato [31], requiring prior conversion of the less-reactive [ 89 Zr] Zr-oxalate [41] in which form 89 Zr is typically produced [45,46]      of the radiotracer for up to 1 week. In practice, this allows the end-user to prepare [ 89 Zr]Zr-oxine in a single manipulation simply by transferring [ 89 Zr]Zr-oxalate from its delivery vial, without modification, to an off-the-shelf vial ("kit") of oxine formulation. Alternatively, it can be used as a basis for the shipping of ready-to-use, single-or multiplepatient doses of [ 89 Zr]Zr-oxine from a central site to distant scanning centres, as has been the case previously for [ 111 In]In-oxine. The radiolabelling agent thus prepared can be added directly to a cell suspension in a procedure analogous to that used conventionally for 111 In-labelling using [ 111 In]In-oxine, as described in the Supplementary Material. In line with conventional radiolabelling protocols, the washing step after radiolabelling ensures the amount of free oxine and unchelated zirconium-89 in the final administered product is further reduced. Sterility and endotoxin testing can be performed according to local regulations.
The labelling of WBC with [ 89 Zr]Zr-oxine was reliable, consistently achieving 45-50% labelling efficiency with 160-480 million cells. This is significantly lower than that achieved with [ 111 In]In-oxine in the same conditions. As the formulations of [ 89 Zr]Zr-oxine and [ 111 In]Inoxine both contain the same total amount of oxine, in large excess (3 orders of magnitude) compared to the amount of the respective radiometals (see Supplementary Material, Table S1), it unlikely that the difference in radiolabelling efficiencies can be explained by differences in amounts of unchelated oxine. Instead, we suggest the difference is most likely due to the physicochemical properties of each radiotracer.  [30]. If imaging cells at later time points (> 24 h) is desired, then longer radiotracer retention studies are warranted. In studies with Tcells, bone marrow cells or stem cells [32,34,35,37], significant leakage (30-60%) of [ 89 Zr]Zr-oxine was observed over 2-3 days in vitro, with the rate of leakage slowing down after 3 days. However, preclinical PET images from the same studies suggest better in vivo retention of 89 Zr than in vitro, and it has been shown that PET images can be improved by administering deferoxamine intravenously to accelerate the renal clearance of 89 Zr released from radiolabelled cells [48].
Using the typical patient dose of [ 111 In]In-oxine (approx. 20 MBq) as reference, we aimed for patient doses of [ 89 Zr]Zr-oxine of about 9-10 MBq. Considering the sensitivity of current-generation clinical PET and SPECT scanners and the relatively low positron yield of 89 Zr (23%), we estimate that 9-10 MBq of 89 Zr should result in a 5-to-10fold increase in useful counts compared to 20 MBq of 111 In and therefore lead to improved image quality. Another advantage of 89 Zr over 111 In is that PET/CT facilitates signal quantification and will allow better determination of the cell numbers reaching the target organ. Sato et al. have recently performed dosimetry studies of 89 Zr-labelled NK cells in non-human primates [48], suggesting such amounts of activity to be safe for the liver and spleen. Furthermore, their study showed good quality images in a clinical PET/CT scanner with 89 Zr activities in the range of 13-44 kBq/10 6 cells, similar to the average activity of 32.9 kBq/10 6 cells achieved in the present study. Crucially, cell viability and retention of 89 Zr in cells in the presence of plasma were high and comparable to those of 111 In, suggesting that radiotracer leakage will be low after intravenous administration and that the PET signal will accurately reflect the location of intact cells, as previously observed in preclinical studies [35]. We did not perform in vivo experiments as there is no justification in this case for using unpurified human leukocytes, i.e. predominantly neutrophils, in a small animal model. Our results using flow-assisted cell sorting indicate no preferential uptake by any specific leukocyte population. In contrast, it has been shown previously that [ 99m Tc]Tc-HMPAO accumulated preferentially in eosinophils [1], with the clinical implication that [ 99m Tc]Tc-HMPAO WBC scans disproportionately represent the distribution of eosinophils. Our results suggest this phenomenon is not expected with [ 89 Zr]Zr-oxine. However,   because there is significant uptake in RBCs and platelets, better separation of these cells from WBC (e.g. using automated cell separators) will improve signal specificity and target-to-background ratio. Leukocyte chemotaxis towards pathogens and inflammatory stimuli is the biological basis for WBC imaging. We found that 89 Zr-labelled WBC retained their chemotactic properties in vitro, with no significant differences compared to [ 111 In]In-oxine and unlabelled control cells. This suggests that radiolabelling with [ 89 Zr]Zr-oxine will not affect in vivo properties of administered WBC at least within a few hours following radiolabelling and will provide clinically useful images. While both radiotracers resulted in significantly higher DNA damage compared to non-radiolabelled cells, no differences were observed between the radiotracers. Notably, our results were strengthened by the intraindividual nature of the comparison between [ 89 Zr]Zr-oxine and [ 111 In]In-oxine.
Considering the translational aspect of this technique, we also emphasise that WBC labelling was performed following the guidelines established for [ 111 In]In-oxine and [ 99m Tc]Tc-HMPAO [43,44], and thus requires no modifications from existing protocols beyond the provision of additional shielding for 89 Zr.

Conclusions
In summary, we provide a clinically applicable, kit-based method for the simple and reliable preparation of ready-to-use [ 89 Zr]Zr-oxine, in a single step, from commercially supplied [ 89 Zr]Zr-oxalate and demonstrate its use in radiolabelling WBC. Our work further demonstrates that, from a cell labelling perspective, [ 89 Zr]Zr-oxine is an appropriate PET equivalent of [ 111 In]In-oxine, and should be investigated in the clinic not only for WBC labelling in infection diagnosis, but potentially also for upcoming trials of cell-based therapies such as CAR-T and stem cells.

Declaration of competing interest
FM, PJB and RTMR have submitted a patent application in relation to this article. The authors declare that they have no other competing interests.