Automated Light-Induced Synthesis of 89Zr-Radiolabeled Antibodies for Immuno-Positron Emission Tomography

doi:10.21203/rs.3.rs-941904/v1

Download PDF

Research Article

Automated Light-Induced Synthesis of ⁸⁹Zr-Radiolabeled Antibodies for Immuno-Positron Emission Tomography

https://doi.org/10.21203/rs.3.rs-941904/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 13 Jan, 2022

Read the published version in Scientific Reports →

You are reading this latest preprint version

Clinical production of ⁸⁹Zr-radiolabeled antibodies (⁸⁹Zr-mAbs) for positron emission tomography (PET) imaging relies on the pre-conjugation of desferrioxamine B (DFO) to the purified protein, followed by isolation and characterization of the functionalized intermediate, and then manual radiosynthesis. Although highly successful, this route exposes radiochemists to a potentially large radiation dose and entails several technological and economic hurdles that limit access of ⁸⁹Zr-mAbs to just a specialist few Nuclear Medicine facilities worldwide. Here, we introduce a fully automated synthesis box that can produce individual doses of ⁸⁹Zr-mAbs formulated in sterile solution in <25 min starting from [⁸⁹Zr(C₂O₄)₄]^4– (⁸⁹Zr-oxalate), our Good Laboratory Practice-compliant photoactivatable desferrioxamine-based chelate (DFO-PEG₃-ArN₃), and clinical-grade antibodies without the need for pre-purification of protein. The automated steps include neutralization of the ⁸⁹Zr-oxalate stock, chelate radiolabeling, and light-induced protein conjugation, followed by ⁸⁹Zr-mAb purification, formulation, and sterile filtration. As proof-of-principle, ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab was synthesized directly from Herceptin in <25 min with an overall decay-corrected radiochemical yield of 20.1±2.4% (n=3), a radiochemical purity >99%, and chemical purity >99%. The synthesis unit can also produce ⁸⁹Zr-mAbs via the conventional radiolabeling routes from pre-functionalized DFO-mAbs that are currently used in the clinic. This automated method will improve access to state-of-the-art ⁸⁹Zr-mAbs at the many Nuclear Medicine and research institutions that require automated devices for radiotracer production.

Bioinorganic chemistry

Automation

89Zr

antibodies

positron emission tomography

photoradiosynthesis

Automation is common practice in the synthesis of clinical-grade radiopharmaceuticals. For example, the small-molecule, metal-based radiotracers ⁶⁸Ga-PSMA-11 and ¹⁷⁷Lu-PSMA-617 are typically produced by automated synthesis modules that perform both the radiolabeling and purification steps(1–4). In these examples, the chemical reaction is a one-step process (metal ion complexation) and efficient radiolabeling produces the isolated and fully formulated radiopharmaceutical with high decay-corrected radiochemical yields (RCYs > 80%) and radiochemical purity (RCP > 95%). It is important to note that the metalation step typically runs to full decay-corrected radiochemical conversion (RCC > 95%). Activity losses stem from inefficient transfer long fluid pathways or imperfect recovery during the purification step. More complex chemical reactions or multiple synthetic steps can also be automated but these processes are usually encountered only in the production of small-molecule ¹¹C- and ¹⁸F-labeled PET tracers(5–7). For instance, ¹¹C-radiochemistry often requires the use of non-aqueous solvents and the synthesis of reactive intermediates such as ¹¹C-methyl iodide before subsequent reactions with a precursor yield the desired radiotracer. In ¹¹C- and ¹⁸F-radiochemistry, the labeling precursors often contain protected functional groups that are unmasked by deprotection steps following the introduction of the radionuclide(8–11). This chemical flexibility facilitates the ‘Total Radiosynthesis’ of complex drug molecules where the radionuclide is incorporated in positions that are simply not accessible using the more common late-stage chemistry(9,12).

In the clinic, automation improves the reproducibility of radiotracer synthesis, and helps minimize radiation exposure to radiochemists who frequently operate with GBq amounts of radioactivity(13). Most automatic radiosynthesizers are placed inside (mini)hot cells and are operated remotely via a computer interface. Radiotracer automation also improves traceability, documentation management and compliance with regulatory requirements for quality control(14).

There are two prominent examples for the automated synthesis of ⁸⁹Zr-labeled monoclonal antibodies (⁸⁹Zr-mAbs)(3,15). In both reports, the radiolabeling reaction relies on the use of a pre-functionalized desferrioxamine-antibody conjugate (DFO-mAb). In 2016, Wright et al. performed the ⁸⁹Zr-radiolabeling of DFO-Bz-NCS-trastuzumab on a microfluidic chip followed by manual purification. The procedure was performed in 45–60 min and produced an isolated product with sufficient activity yield for multiple patient doses(15). More recently, Poot et al. reported the fully automated radiolabeling and purification of ⁸⁹Zr-mAbs in 2019(3). In this example, single patient doses of purified ⁸⁹ZrDFO-N-Succ-rituximab and ⁸⁹ZrDFO-N-Succ-cetuximab were produced in 77 minutes starting from ⁸⁹Zr-oxalate and the respective pre-functionalized DFO-mAb(3). The DFO-mAb radiolabeling precursors are produced by DFO-conjugation to lysine residues on the protein using reagents bearing the activated ester N-hydroxysuccinimide (NHS) or a benzylisothiocyanate (Bz-NCS) group which form amide and thiourea bonds, respectively(16,17). DFO-mAb conjugates are typically prepared in advance, characterized, stored, and radiolabeled on demand, and so far, this holds true for all automated and manual production of ⁸⁹Zr-mAbs. In general, this chemistry is extremely successful and has facilitated clinical translation of many ⁸⁹Zr-mAbs(18–22). However, the necessity to produce, characterize and store the DFO-mAb intermediate presents several technological and financial limitations – synthesizing sufficient material for toxicological studies is both difficult and expensive – that mean accessing ⁸⁹Zr-mAbs is not always feasible in smaller nuclear medicine facilities. In addition, the long-term storage of an intermediate can lead to questions over the stability and shelf-life of the material.

Recently, we developed a one-pot route that combines the bioconjugation and radiolabeling steps using new DFO-chelates bearing a photoactivatable aryl azide group (ArN₃)(23–26). Three key features of this alternative photoradiochemical method for making ⁸⁹Zr-mAbs are: i) rapid reaction times that depend on the rate-limiting photoactivation step; ii) high chemical tolerance of the light-induced bioconjugation process with many components of mAb formulation buffers, and iii) avoidance of the need to use pre-functionalized DFO-mAbs. The compatibility of the photolabeling process with common mAb formulation buffers, including high concentrations amino acids (e.g. histidine), surfactants (polysorbate-80), sugars (α,α-trehalose), antioxidants (ascorbate), and various salts such as phosphate buffered saline (PBS) mean that the chemistry often works without the need to pre-purify the protein from clinical-grade stocks. This critical difference between our photoradiosynthesis approach and the classic multiple step routes for making ⁸⁹Zr-mAbs led us to postulate that it could be feasible to make an automated radiosynthesizer that combines all critical steps required in the manufacture of ⁸⁹Zr-mAbs, namely, oxalic acid neutralization, bioconjugation, radiolabeling, purification and sterile formulation (Fig. 1).

Here, we report the design, manufacture, and proof-of-concept pre-clinical evaluation of ALISI – a prototype radiosynthesizer for the fully automated light-induced synthesis of ⁸⁹Zr-mAbs for immuno-positron emission tomography (PET). An overview of the design and plumbing diagram of the ALISI system is presented in Fig. 2 with full descriptions given in the Online Methods section.

Prior to developing the fully automated synthesis of ⁸⁹Zr-mAbs, we first tested different aspects of the ALISI system by using model reactions involving the bioconjugation between human serum albumin (HSA) and three photoactivatable compounds. The compounds used were RhodB-PEG₃-ArN₃, pre-radiolabeled ⁶⁸GaDFO-PEG₃-ArN₃, and the free ligand (DFO-PEG₃-ArN₃) together with ⁸⁹Zr-oxalate which forms ⁸⁹ZrDFO-PEG₃-ArN₃ in situ. Chemical structures are shown in Figs. 1 and 3A, and reaction details are presented in Supplemental Tables 1–3.

Photochemical activation kinetics were measured to assess the efficiency of the newly designed photoreactor by irradiation of a stock solution of DFO-PEG₃-ArN₃. Samples were analyzed by reverse-phase high-performance liquid chromatography (HPLC). The irreversible photoactivation of DFO-PEG₃-ArN₃ was complete within 90s, as indicated by full consumption of the starting material in HPLC (Fig. 3B).

The three model reactions were used to assess different aspects of the automated system. First, the photoactivatable fluorophore RhodB-PEG₃-ArN₃ was used to optimize the sequence and timing of the different steps in the automation protocol. The use of a highly colored (pink) RhodB-PEG₃-ArN₃ dye also helped to evaluate the completeness of all reagent transfers by visual tracking and by quantitative measurements of dilution factors using electronic absorption spectroscopy. We discovered that the lipophilic nature of the small-molecule RhodB-PEG₃-ArN₃ compound, and the related photolyzed products, means that these species stick to Sephadex columns which hare typically used for protein purification. Consequently, the fluorescently labeled RhodB-PEG₃-azepin-HSA product can be purified easily from the small-molecule byproducts on the ALISI system using standard PD-10 columns but this model fluorophore was not ideal for optimizing the purification procedure for radiolabeled proteins.

Next, test reactions between HSA and pre-radiolabeled ⁶⁸GaDFO-PEG₃-ArN₃ were used to optimize the photochemical conjugation step. Optimization was performed by maximizing the decay-corrected radiochemical conversion (RCC, measure by size-exclusion chromatography high-performance liquid chromatography, SEC-HPLC) to the labeled protein, ⁶⁸GaDFO-PEG₃-azepin-HSA. We found that pH is a critical factor in determining the yield of photo-induced bioconjugation reactions using ArN₃ species(27). Therefore, we investigated the use of different buffers with the aim of controlling pH throughout all synthetic steps. 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES) is a zwitterionic buffer commonly used in the synthesis or radiolabeled mAbs(17). In our hands, the use of HEPES at concentrations ≥0.5 M had a detrimental impact on the RCC of ⁶⁸GaDFO-PEG₃-azepin-HSA (data not shown). In contrast, reactions in sodium borate buffer (0.25M, pH8) did not quench the light-induced conjugation reaction. Experimental data showing the measured RCC versus reaction pH are presented in Fig. 3C. These data show a trend toward increased RCC in the pH window from 8.0-to-9.5, with RCC reaching 47% between pH8.0–8.7. Below pH8.0, conjugation yields decrease sharply due to progressive protonation of lysine residues. Based on these data, further reaction parameters were optimized at pH8.0.

Our goal was to minimize the manual handling of radioactive components by producing ⁸⁹Zr-mAbs directly from ⁸⁹Zr-oxalate stocks and the unfunctionalized protein. Therefore, we examined the complexation reaction to form ⁸⁹ZrDFO-PEG₃-ArN₃ on ALISI (Fig. 3D)(24,25). ⁸⁹Zr-oxalate was transferred into the reaction vial, and then neutralized by automatic transfer of the Na₂CO₃ and sodium borate solutions. Then, the DFO-PEG₃-ArN₃ solution was transferred into the magnetically stirred vial, and 10 seconds after addition, an aliquot of the reaction mixture was removed manually with a pipette and spotted immediately onto a silica gel instant-thin-layer chromatography (ITLC) strip (developed with 50mM DTPA, pH7.4). The radio-ITLC data showed quantitative radiochemical conversion to give ⁸⁹ZrDFO-PEG₃-ArN₃ (which is retained at the baseline, R_f=0.0) in < 10s. In contrast to radiolabeling of pre-functionalized DFO-mAbs which requires > 45 min, the high diffusion coefficient of the small-molecule components and gentle stirring of the reaction mixture facilitate rapid ⁸⁹Zr-complexation by DFO-PEG₃-ArN₃.

Data in Fig. 3E show the effect of changing the protein concentration on the measured photochemical conversion yields (PCY) for the synthesis of RhodB-PEG₃-azepin-HSA (purple), or RCC yields for ⁶⁸GaDFO-PEG₃-azepin-HSA (blue) and ⁸⁹ZrDFO-PEG₃-azepin-HSA (black). As with many bioconjugation reactions, photo-induced labeling using ArN₃ compounds show a steep dependence on the initial concentration of protein. Conjugation yields began to decrease when [HSA] < 10 mg mL^− 1. For each of the RhodB-PEG₃-azepin-HSA (purple trace), ⁶⁸GaDFO-PEG₃-azepin-HSA (blue), and ⁸⁹ZrDFO-PEG₃-azepin-HSA products, maximum conjugation yields were obtained at [HSA] ~ 15 mg mL^− 1 with values of 34.2±0.9%, 49.1±3.9% and 62.9±2.2%, respectively. Analysis by size-exclusion chromatography coupled to a high-performance liquid chromatography system (SEC-HPLC, Fig. 3F) confirmed the successful protein-ligation by co-elution of the HSA protein (orange trace) with that of RhodB-PEG₃-azepin-HSA (purple trace, monitoring the fluorophore at 568nm), and the radioactive traces of ⁶⁸GaDFO-PEG₃-azepin-HSA (blue trace), and ⁸⁹ZrDFO-PEG₃-azepin-HSA (black trace).

The purification of ⁸⁹Zr-mAbs typically involves using PD-10 desalting columns (Sephadex G-25, 5kDa exclusion limit) which only work well when very high RCCs are obtained, or when the small-molecule byproducts are trapped on the Sephadex (as with the RhodB-PEG₃-azepin-HSA synthesis, vide supra). However, when RCCs drop below ~ 80%, the separation efficiency of the PD-10 matrix is sub-optimal. To enhance the radiochemical purity (RCP) of the purified ⁸⁹Zr-labeled proteins, we constructed custom-made SK-10 columns that have the same geometry as PD-10 columns but are loaded with Sephadex® G-100 (150kDa exclusion limit) that increases the resolving power. SK-10 columns had a dead volume of ~ 2mL followed by a protein collection volume of 2.5mL (Fig. 4A). Small-molecules were effectively retained on SK-10 columns, where a peak-to-peak separation between the desired ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab and the byproducts was > 3mL, giving a final RCP of the purified protein fraction of > 99% starting from a crude reaction mixture that had an RCC of 42%.

Next, we investigated the fully automated production of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab using ALISI (Supplemental Table 4). Again, changing the protein concentration had an impact on both decay-corrected RCC for the protein conjugation step and on the isolated decay-corrected RCY of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab where a plateau was observed in the protein concentration range of 10–15 mg mL^− 1 (Fig. 4B). The same trend observed in reactions with HSA was seen with Herceptin. Automation successfully produced and purified ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab in < 25 min starting directly from non-purified Herceptin (14.9 mg mL^− 1 of protein) giving an isolated decay-corrected RCY of 20.1 ± 2.4% (n = 3), and an RCP > 99% (Fig. 4C). Notably, SEC analysis of the isolated ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab product revealed less than 5% of the activity was associated with an aggregated protein fraction (which elutes at a slightly shorter retention time). The activity recovered in the purified product was lower than the RCC by a factor of ~ 2. By tracking the activity along the fluid pathway, we found that ⁸⁹Zr-oxalate was loaded into the reaction vial with 82.7 ± 3.1% efficiency. The photochemical conjugation efficiency is given by the RCC (~ 45%) and loading of the crude reaction mixture onto the SK-10 column had a transfer efficiency > 99%. The purification step was, on average, 75.0 ± 4.3% efficient, while sterile filtration led to a small loss in protein-bound activity with a recovered activity efficiency of 87.7 ± 2.8%. Transfer losses account for the difference between the measured RCC and RCY values.

Qualitative testing of the ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab products for residual borate was performed by adding a few drops of an ethanolic solution of curcumin, which produces a red/orange solution due to the formation of rosocyanine dye in the presence of borate species(28). Visual detection is possible at borate concentrations down to ~ 2.5mM and tests confirmed that none of the isolated samples of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab produced by ALISI contained residual sodium borate.

In the clinic, patients undergoing PET scans with ⁸⁹Zr-mAbs typically receive a dose of 37 MBq of activity administered with between 3–100 mg of total protein(29–31). In scaled-up syntheses (Supplemental Table S4), we demonstrated the potential of ALISI to produce individual patient doses of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab. Starting from 152 MBq of ⁸⁹Zr-oxalate and 9.7 mg of trastuzumab (formulated as Herceptin), photoradiosynthesis gave an activity yield of 27.9 MBq of isolated ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab with a decay corrected RCY of 18.3%, an RCP > 99%, a chemical purity > 99%, and a molar activity A_m of 0.43 MBq nmol^− 1 of protein. If the ALISI system was qualified for use in a clinical radiopharmacy, this reaction product would be sufficient to image a patient.

Finally, to illustrate the flexibility of the ALISI radiosynthesizer, we adapted the system to automate the radiolabeling and purification of ⁸⁹Zr-mAbs via the conventional two-step approach (Fig. 5A). First, trastuzumab was recovered as a purified protein from a clinical-grade sample of formulated Herceptin by using PD-10 gel filtration. Next, the functionalized DFO-Bz-NCS-trastuzumab radiolabeling precursor was produced in accordance with the methods of Vosjan et al(16,17). Finally, ⁸⁹Zr-radiolabeling and automated PD-10 purification using ALISI gave ⁸⁹ZrDFO-Bz-NCS-trastuzumab with a decay-corrected isolated RCY of 48.3 ± 8.4% (n = 3), an RCP > 99%, and chemical purity > 99% (Supplemental Table S5) in 90 minutes. Corresponding radio-ITLC and SEC-HPLC characterization data for ⁸⁹ZrDFO-Bz-NCS-trastuzumab are shown in Figs. 5B and 5C, respectively. Importantly, in the purified sample of ⁸⁹ZrDFO-Bz-NCS-trastuzumab contained ~ 15% of ⁸⁹Zr-activity associated with an aggregated protein peak. This is common feature of the conventional two-step radiolabeling method used in the clinic, which requires removal of the mAb from the stabilizing formulation components prior to functionalization with DFO-Bz-NCS(23). The presence of ⁸⁹Zr-labeled protein aggregate is associated with accumulation of activity in the liver and spleen. In comparison, the photoradiochemical approach reduces this aggregate fraction by a factor of ~ 3 which would likely improve image contrast and reduce the radiation burden to the patient.

The ALISI radiosynthesizer was constructed by using a combination of open-source microcontrollers (Arduino) and computer-aided design (CAD), coupled with additive manufacturing(32,33). Arduino electronic components are modular and expandable. This allows for rapid and facile integration of new features during prototype development. Liquid handling components and all fluidic pathways employ commercially available, single-use, sterile tubing and three-way switching valves operated by digital servomotors. These components were chosen to facilitate future translation of ALISI to a clinical environment. The custom-made, electropolished photoreactor houses three high-powered LEDs with peak emission at 365nm (Fig. 2C). Light-induced activation of our ArN₃ reagents is rate-limiting in the photoradiosynthesis of ⁸⁹Zr-mAbs, and in comparison with manual reactions, the mirrored photoreactor of ALISI gave a 5-to-10-fold decrease in reaction time(26).

It is important to note that the bioconjugation efficiency (as measured from the PCY or RCC values) shows a strong dependence on the reaction geometry and the conditions used. For instance, the nature of the photoactive compound and protein substrate, the choice of radionuclide, the concentrations of all reagents, the buffer composition, and the pH influence conjugation efficiency. Under identical conditions, RhodB-PEG₃-ArN₃ gave consistently lower yields for HSA labeling than pre-radiolabeled ⁶⁸GaDFO-PEG₃-ArN₃, which in-turn, was lower than radiolabeling with ⁸⁹ZrDFO-PEG₃-ArN₃ generated in situ. Under our optimized conditions, the observed bioconjugation yields using an initial 1:1 stoichiometric ratio between HSA and the photoactive compound were in the range of 35–65%. Standard bioconjugation reactions used to make the DFO-mAbs, including activated ester(16) or thiourea(17) chemistry, have efficiencies in the range of ~ 20–75% (see Supplementary Methods). These data confirm that photo-induced bioconjugation is equally successful compared to current state-of-the-art methods used to make radiolabeled mAbs in clinical practice. For the photoradiosynthesis of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab from Herceptin, the RCC was ~ 45%. This is remarkable considering that the radiolabeling and bioconjugation steps are complete in < 90s, and the reaction uses non-purified mAb where the mixture contains all formulation components of clinical-grade Herceptin. We note that radiochemists performing manual syntheses of ⁸⁹Zr-mAbs are familiar with obtaining near quantitative RCYs, but in most cases, the reaction only involves the radiolabeling step (not the conjugation). Since test reactions are usually employed to predetermine the maximum molar activity of a sample, quantitative labeling is expected. In a recent study, Poot et al. automated the radiolabeling and purification of ⁸⁹Zr-mAbs from pre-functionalized DFO-mAbs with RCYs in the range 60–75%(3). With the present photochemistry, formation of ⁸⁹ZrDFO-PEG₃-ArN₃ in situ is quantitative and the decay-corrected isolated RCY of ⁸⁹ZrDFO-PEG₃-azepin-trasutzumab (20.1 ± 2.4%) encompasses all chemical reactions, transfers, and processing steps. Transfer losses are one of the main limitations of adapting manual chemistry to automated platforms. Considering the multi-step nature of the ALISI protocol, which includes all reagent transfers, acid neutralization, buffer control, radiolabeling and bioconjugation, in-line purification, sterile filtration, and product formulation, the observed RCYs are an excellent benchmark.

Purification using SK-10 columns filled with Sephadex® G-100 media provided greatly enhanced separation and improved RCP (> 99%) of the isolated ⁸⁹ZrDFO-PEG₃-azepin-trasutzumab when compared with PD-10 columns. This is expected since the Sephadex G-25 media used in PD-10 columns is primarily intended for desalting and is sub-optimal for separating large proteins from small-molecule components.

Finally, automated radiosynthesis of ⁸⁹Zr-mAbs on ALISI improved the reproducibility of the chemistry by standardizing the reaction geometry and conditions, and by decreasing the potential for user-related, irreproducible errors. Collectively, the successful synthesis and isolation of ⁸⁹ZrDFO-PEG₃-azepin-trastuzumab using our automated radiosynthesizer suggest that ALISI is potentially useful for preparing individual patient doses or small batches of ⁸⁹Zr-mAbs (and other radiolabeled proteins) on demand. With further development, we anticipate that the system can be adapted for use with other radionuclides and photoactivatable chelates to access radiopharmaceuticals for applications in radioimmunotherapy.

We developed a radiosynthesizer unit that performs the fully automated light-induced synthesis of ⁸⁹Zr-mAbs for immunoPET. The ALISI system can produce ⁸⁹Zr-mAbs in < 25 min in high radiochemical purity (RCP > 99%) starting from stock solutions of ⁸⁹Zr-oxalate, a photoactivatable DFO-PEG₃-ArN₃ chelate, and a protein of interest. Features of the automated protocol include neutralization of the oxalic acid, buffer exchange and pH correction, quantitative formation of the photoactivatable ⁸⁹ZrDFO-PEG₃-ArN₃ complex, rapid (< 90s) light-induced protein-ligation, in-line purification using novel SK-10 size-exclusion chromatography columns, and finally sterile filtration and product formulation in a biocompatible medium. After loading the reagent reservoirs, all steps are performed at the touch of a single button. The system is also high flexible, producing purified ⁸⁹Zr-mAbs via conventional labeling of a pre-functionalized DFO-mAb conjugate. With further development, we anticipate that ALISI could facilitate on-demand access to individual patient doses of ⁸⁹Zr-mAbs in Nuclear Medicine facilities that rely on automated devices for radiotracer production.

Further details are presented in the Supplemental Materials. Schematics and sample code for operating the microcontrollers are available on request.

Data and code availability statement

An electronic computer-aided design file containing the full assembly and all components of the radiosynthesizer required for 3D-printing, as well as the Arduino program (sketch) is available from the corresponding author.

Protein samples

Formulated Herceptin^TM (Roche/Genentech, South San Francisco, CA) and human serum albumin (HSA; Merck, Darmstadt, Germany) were reconstituted in water (>18.2MΩ·cm at 25^oC). Protein concentrations were determined by using a Nanodrop^TM One^C Microvolume UV-Vis Spectrophotometer.

Additive manufacturing

The ALISI radiosynthesizer was constructed by using the computer assisted design (CAD) software Solidworks2020 (Dassault Systèmes, Vélizy-Villacoublay, France). Components were prepared through additive manufacturing, laser cutting or purchased directly from commercial vendors. Additive manufacturing components were produced by selective laser sintering and made from PA2200, a fine powder based on polyamide-12. The synthesizer case was manufactured from laser cut, 5mm plates of high tensile strength, black polyoxymethylene (POM) and an aluminum profile modular assembly system (Kanya AG, Rüti, Switzerland).

Control and electronics

Electronic components are based around the single-chip ATmega328 microcontroller. Microcontrollers, breakout boards, additional circuit boards, and electronic components were either custom-manufactured or purchased from Arduino.cc (Ivrea, Italy), Adafruit Industries (New York City, NY, USA), RS Components (Frankfurt-am-Main, Deutschland), or Distrelec AG (Nänikon, Switzerland). The photoreactor consists of an electropolished stainless steel tube containing an array of three, high-powered light-emitting diodes (LEDs; Nichia, Anan, Japan) with a light output of 1.03W per LED at 365nm. Valve modulation is performed with standard digital servos (Savöx, Salt Lake City, UT, USA). All liquid transfer steps are driven pneumatically with a syringe pump controlled by a NEMA-17 bipolar stepper motor (Distrelec AG, Nänikon, Switzerland).

Liquid handling and cassettes

All components for the disposable cassette system were either custom-manufactured or purchased from commercial vendors (B. Braun Melsungen AG, Meslungen, Germany, or BD, Heidelberg, Germany). The cassettes use flexible, high chemical resistance Tygon tubing. For purification, custom-made SK-10 separation columns (Econo-Pac chromatography columns, Bio-Rad Laboratories, Hercules, CA, USA) filled with 8.3mL of pre-soaked Sephadex® G-100 (Merck, Darmstadt, Germany) were constructed. The stationary phase of the SK-10 columns was capped with filter-frits and columns were eluted with sterile PBS (pH7.4).

Synthesis and radiochemistry

The photoactivatable chelate DFO-PEG₃-ArN₃ and the fluorescent photoactivatable fluorophore (PhotoTag) RhodB-PEG₃-ArN₃ were synthesized and characterized as described previously.(24) The stock solution of ⁸⁹Zr-oxalate ([⁸⁹Zr(C₂O₄)₄]^4-(aq.) in ~1M oxalic acid) was obtained from PerkinElmer (Waltham, MA, USA; manufactured by the BV Cyclotron VU, Amsterdam, The Netherlands) and was used without further purification. Full experimental details on the model chemistry using RhodB-PEG₃-ArN₃, ⁶⁸GaDFO-PEG₃-ArN₃ and ⁸⁹Zr-oxalate in combination withDFO-PEG₃-ArN₃ are given in the Supplemental Information. All reported RCC and RCY values are decay corrected.

Setup for automated radiolabeling

Reactions with ALISI were performed on new liquid handling equipment, freshly assembled from sterile packaging. Briefly, individual reservoirs were assigned to stock solutions of Na₂CO₃(aq.), DFO-PEG₃-ArN₃, ⁸⁹Zr-oxalate, protein, reaction buffer, and sterile PBS for purification by size-exclusion chromatography (SEC) (see Supplemental Tables 1–4). A minimum volume of ~50mL was necessary to achieve adequate liquid transfer. For test reactions where reagent volumes were below this threshold, the total volume was adjusted with an appropriate volume of water.

Description of the automated ⁸⁹Zr-photoradiolabeling procedure

A schematic of the plumbing diagram and a photograph of the ALISI system set-up for the synthesis of ⁸⁹Zr-mAbs is shown in Figure 2. The radiosynthesizer unit is initialized by pushing the power button. After initialization, the system provides a digital prompt on the built-in LCD-display to indicate that the device is ready to start. Automated radiosynthesis and purification is initiated by pressing the start button. Thereafter, the radiosynthesizer transfers all reagents and components to the reaction vial, located inside the photoreactor, in the following sequence: i) ⁸⁹Zr-oxalate in ~1M oxalic acid is transferred. ii) An equal volume of 1M Na₂CO₃(aq.) is transferred. iii) DFO-PEG₃-ArN₃ in a solution of ~10% DMSO and aqueous sodium borate buffer (0.25M, pH8.0) is removed from the reservoir and first used to wash the ⁸⁹Zr-stock solution reservoir before being delivered to the reaction vial. iv) The protein solution is transferred. v) Additional sodium borate buffer (0.25M, pH8.0) is used to wash the protein reservoir and then transferred.

After the reagent transfer sequence, the reaction mixture is irradiated with 365nm light for 90s. Tests indicated that the temperature of the reaction mixture does not change during this time, but the high-powered LEDs require cooling with an aluminum heat-sink attached to a fan. After irradiation, the crude mixture is transferred to the SK-10 size-exclusion column (SEC) for purification. After automatic separation, the product fraction containing the high molecular weight protein is filtered through a standard 0.22 mm sterile filter and collected in a sterile vial. The product is formulated in sterile PBS (pH7.4), and after quality control, is ready for use in radiochemical, cellular, or in vivo assays.

Data Analysis

Data were plotted by using the GraphPad Prism 9.0 software (GraphPad Software Inc., San Diego, California USA).

ACKNOWLEDGMENTS

We thank the Swiss National Science Foundation (SNSF Professorship PP00P2_163683 and PP00P2_190093), the Swiss Cancer Research Foundation (KFS-4257-08-2017), and the University of Zurich (UZH) for financial support. This project has received funding from the European Research Council under the Grant Agreement numbers 676904, ERC-StG-2015, NanoSCAN and 101001734, ERC-CoG-2020, PhotoPHARMA. We thank Dr Amaury Guillou for synthesis of the photoactivatable compounds. We thank Sascha Weidner and the AddMan Factory at the University of Zurich.

AUTHOR CONTRIBUTIONS

S.K. and J.P.H. designed the project. S.K. performed all experiments. S.K. and J.P.H. wrote the manuscript. J.P.H supervised the project.

COMPETING INTERESTS

Nothing to declare.

Lin M, Paolillo V, Ta RT, et al. Fully automated preparation of 68Ga-PSMA-11 at curie level quantity using cyclotron-produced 68Ga for clinical applications. Appl Radiat Isot. 2020;155:108936.
Thisgaard H, Kumlin J, Langkjær N, et al. Multi-curie production of gallium-68 on a biomedical cyclotron and automated radiolabelling of PSMA-11 and DOTATATE. EJNMMI Radiopharm Chem. 2021;6:1.
Poot AJ, Adamzek KWAA, Windhorst AD, et al. Fully Automated 89Zr Labeling and Purification of Antibodies. J Nucl Med. 2019;60:691-695.
Wichmann CW, Ackermann U, Poniger S, et al. Automated radiosynthesis of [⁶⁸Ga]Ga-PSMA-11 and [¹⁷⁷Lu]Lu-PSMA-617 on the iPHASE MultiSyn module for clinical applications. J Label Compd Radiopharm. 2020:140-146.
Collet C, Vucko T, Ariztia J, Karcher G, Pellegrini-Moïse N, Lamandé-Langle S. Fully automated radiosynthesis of [¹⁸F]fluoro- C -glyco-c(RGDfC): exploiting all the abilities of the AllInOne synthesizer. React Chem Eng. 2019;4:2088-2098.
Davis RA, Drake C, Ippisch RC, Moore M, Sutcliffe JL. Fully automated peptide radiolabeling from [¹⁸F]fluoride. RSC Adv. 2019;9:8638-8649.
Collier TL, Normandin MD, Stephenson NA, et al. Synthesis and preliminary PET imaging of ¹¹C and <sup>18<sup>F isotopologues of the ROS1/ALK inhibitor lorlatinib. Nat Commun. 2017;8:15761.
Miller PW, Long NJ, Vilar R, Gee AD. Synthesis of ¹¹C, ¹⁸F, ¹⁵O, and ¹³N radiolabels for positron emission tomography. Angew Chemie - Int Ed. 2008;47:8998-9033.
Liang SH, Vasdev N. Total Radiosynthesis: Thinking Outside ‘the Box’’.’ Aust J Chem. 2015;68:1319-1328.
Rotstein BH, Liang SH, Holland JP, et al. ¹¹CO₂ fixation: A renaissance in PET radiochemistry. Chem Commun. 2013;49:5621-5629.
Jacobson O, Kiesewetter DO, Chen X. Fluorine-18 Radiochemistry, Labeling Strategies and Synthetic Routes. Bioconjug Chem. 2015;26:1-18.
Hicks JW, Wilson AA, Rubie EA, Woodgett JR, Houle S, Vasdev N. Towards the preparation of radiolabeled 1-aryl-3-benzyl ureas: Radiosynthesis of [ ¹¹C-carbonyl] AR-A014418 by [¹¹C] CO₂ fixation. Bioorganic Med Chem Lett. 2012;22:2099-2101.
Kaplan DJ, Patel JN, Liporace FA, Yoon RS. Intraoperative radiation safety in orthopaedics: A review of the ALARA (As low as reasonably achievable) principle. Patient Saf Surg. 2016;10:1-7.
Thompson S, Kilbourn MR, Scott PJH. Radiochemistry, PET Imaging, and the Internet of Chemical Things. ACS Cent Sci. 2016;2:497-505.
Wright BD, Whittenberg J, Desai A, et al. Microfluidic Preparation of a ⁸⁹Zr-Labeled Trastuzumab Single-Patient Dose. J Nucl Med. 2016;57:747-752.
Verel I, Visser GWM, Boellaard R, et al. ⁸⁹Zr Immuno-PET: Comprehensive Procedures for the Production of ⁸⁹Zr-Labeled Monoclonal Antibodies. J Nucl Med. 2003;44:1271-1281.
Vosjan MJWD, Perk LR, Visser GWM, et al. Conjugation and radiolabeling of monoclonal antibodies with zirconium-89 for PET imaging using the bifunctional chelate p -isothiocyanatobenzyl- desferrioxamine. Nat Protoc. 2010;5:739-743.
Deri MA, Zeglis BM, Francesconi LC, Lewis JS. PET imaging with ⁸⁹Zr: From radiochemistry to the clinic. Nucl Med Biol. 2013;40:3-14.
Henry KE, Ulaner GA, Lewis JS. Human Epidermal Growth Factor Receptor 2-Targeted PET/Single- Photon Emission Computed Tomography Imaging of Breast Cancer. PET Clin. 2017;12:269-288.
Lohrmann C, O’Reilly EM, O’Donoghue JA, et al. Retooling a blood-based biomarker: Phase I assessment of the high-affinity CA19-9 antibody HuMAB-5B1 for immuno-PET imaging of pancreatic cancer. Clin Cancer Res. 2019;25:7014-7023.
Berg E, Gill H, Marik J, et al. Total-Body PET and Highly Stable Chelators Together Enable Meaningful 89Zr-Antibody PET Studies up to 30 Days After Injection. J Nucl Med. 2020;61:453-460.
Nemieboka B, Sharma SK, Rao TD, et al. Radiopharmacologic screening of antibodies to the unshed ectodomain of MUC16 in ovarian cancer identifies a lead candidate for clinical translation. Nucl Med Biol. 2020;86-87:9-19.
Klingler S, Fay R, Holland JP. Light-Induced Radiosynthesis of ⁸⁹Zr-DFO-Azepin-Onartuzumab for Imaging the Hepatocyte Growth Factor Receptor. J Nucl Med. 2020;61:1072-1078.
Guillou A, Earley DF, Holland JP. Light-activated protein-conjugation and ⁸⁹Zr-radiolabelling with water-soluble desferrioxamine derivatives. Chem Eur J. 2020;26:7185-7189.
Guillou A, Earley DF, Patra M, Holland JP. Light-induced synthesis of protein conjugates and its application in photoradiosynthesis of ⁸⁹Zr-radiolabeled monoclonal antibodies. Nat Protoc. 2020;15:3579-3594.
Patra M, Klingler S, Eichenberger LS, Holland JP. Simultaneous Photoradiochemical Labeling of Antibodies for Immuno-Positron Emission Tomography. iScience. 2019;13:416-431.
Patra M, Eichenberger LS, Fischer G, Holland JP. Photochemical Conjugation and One-Pot Radiolabelling of Antibodies for Immuno-PET. Angew Chemie Int Ed. 2019;58:1928-1933.
Spicer GS, Strickland JDH. 907. Compounds of Curcumin and Boric Acid. Part II. The Structure of Rubrocurcumin. J Chem Soc. 1952:4650-4653.
Dijkers EC, Oude Munnink TH, Kosterink JG, et al. Biodistribution of 89Zr-trastuzumab and PET imaging of HER2-positive lesions in patients with metastatic breast cancer. Clin Pharmacol Ther. 2010;87:586-592.
Bensch F, Smeenk MM, van Es SC, et al. Comparative biodistribution analysis across four different 89Zr-monoclonal antibody tracers-The first step towards an imaging warehouse. Theranostics. 2018;8:4295-4304.
Bensch F, Brouwers AH, Lub-de Hooge MN, et al. ⁸⁹Zr-trastuzumab PET supports clinical decision making in breast cancer patients, when HER2 status cannot be determined by standard work up. Eur J Nucl Med Mol Imaging. 2018;45:2300-2306.
Badamasi YA. The working principle of an Arduino. In: 2014 11th International Conference on Electronics, Computer and Computation (ICECCO). IEEE; 2014:1-4.
Rossi S, Puglisi A, Benaglia M. Additive Manufacturing Technologies: 3D Printing in Organic Synthesis. ChemCatChem. 2018;10:1512-1525.

No competing interests reported.

Download PDF

Journal Publication

published 13 Jan, 2022

Read the published version in Scientific Reports →

Editorial decision: Major revision
09 Dec, 2021
Reviews received at journal
19 Nov, 2021
Reviewers agreed at journal
10 Oct, 2021
Reviewers agreed at journal
29 Sep, 2021
Reviewers invited by journal
29 Sep, 2021
Editor assigned by journal
29 Sep, 2021
Editor invited by journal
29 Sep, 2021
Submission checks completed at journal
29 Sep, 2021
First submitted to journal
27 Sep, 2021

You are reading this latest preprint version

Automated Light-Induced Synthesis of ⁸⁹Zr-Radiolabeled Antibodies for Immuno-Positron Emission Tomography

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Conclusion

ONLINE METHODS

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1