Biochemical characterization of different batches of IL2 conjugated with different chelating agents

Four different experimental batches of IL2, for labelling with Gallium-68 at room temperature, were compared.

Formula 1: hrIL2 conjugated with NODAGA-NHS (NODAGA-hrIL2);

Formula 2: hrIL2 conjugated with Mal-THP (THP-hrIL2).

Formula 3: desIL2 conjugated with NODAGA-NHS (NODAGA-desIL2);

Formula 4: desIL2 conjugated with Mal-THP (THP-desIL2);

Proleukin (Aldesleukin, Novartis, desIL2) is commercially available in lyophilized form. Human interleukin (hrIL2) was purchased by ProSpec (code CYT 209, ProSpec, Rehovot, Israel) in lyophilized form. Both desIL2 and hrIL2 were chemically modified using bifunctional chelating agents (THP or NODAGA), commercially available from CheMatech (Dijon, France).

Experiments performed with the above-mentioned formulations are summarized in Table 1.

Table 1 Summary of experiments performed with the four IL2 formulationsChemical modification of IL2 proteins with the two chelating agents

Each protein vial was reconstituted in 500 µL of water, and then divided into two aliquots (250 µL), each one used for dissolving an amount of chelating agent powders. The bifunctional chelating agents were analytically weighed in order to guarantee 40-fold excess in final reaction condition.

For the conjugation with THP, 250 µL of reconstituted protein solutions was added to 250 µL of phosphate buffer solution 0.100 mol/L at pH = 8.3 to dissolve 1.32 mg of THP. Sonication was performed using a digital ultrasonic bath (ARGOLab DU-32S) for 20 s at 40% amplitude to ensure complete dissolution and obtain a clear solution. The pH was checked after the dissolution and remained stable (7.2 < pH < 7.8).

For the conjugation with NODAGA, 250 µL of reconstituted protein solution was added to 250 µL of deionized water to dissolve 1.05 mg of NODAGA, in order to reconstitute the original conditions obtained during protein lyophilization (pH 5).

In all the four formula, limpid solutions were obtained, without modification of the final pH value after the powder dissolution. All samples were mechanically mixed using an IKA RW20 overhead stirrer at 500 rpm for 4 h at room temperature. Then, samples were dialyzed for 24 h using using Pur-A-Lyzer™ Maxi Dialysis Kit (3.5 kDa MWCO) against a dialysis buffer containing 0.271 mol/L glycerol, 0.64 mmol/L sodium dodecyl sulphate, 1.41 mmol/L NaH2PO4 and 6.27 mmol/L Na2HPO4. At the end of the dialysis, all samples were freeze-dried and lyophilized.

Protein digestion procedure

Protein samples, suspended in 50 mmol/L ammonium hydrogen carbonate, were digested for 16 h at 37 °C using trypsin. The obtained peptide blends were purified using a C18 zip-tip (Merk Life Science S.r.l., Milano, Italy), dried and re-suspended in water containing 0.2% trifluoroacetic acid.

Mass analysis

Mass analysis of pure, modified and digested samples was performed at CEINGE Laboratory (Naples, Italy). Lyophilized powders of pure and modified proteins were suspended in 50 µL of a mixture containing CH3CN/H2O (0.2% HCOOH) at 1/1 volumetric ratio. Samples were analyzed by direct injection (flow = 10 µL/min) using ESI–MS Q-TOF Premier (Waters, Milford, MA, USA). Raw data were analyzed using MassLynx 4.1 (Waters, Milford, MA, USA) software.

Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) analyses of peptide mixture

Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) analyses were carried out on a 4800 plus MALDI TOF-TOF mass spectrometer (AB Sciex) equipped with a reflectron analyzer and used in delayed extraction mode with 4000 Series Explorer v3.5 software. For the analyses, 0.50 µL of peptide mixture were mixed with an equal volume of α-cyano-4-hydroxycynnamic acid as matrix (10 mg/mL of powder in H2O 0.2% TFA in 70% CH3CN v/v), loaded onto the metallic sample plate and air-dried. MALDI-MS data were acquired over a mass range of 400–5600 m/z in the positive-ion reflector mode. Raw data were elaborated using the software provided by the manufacturer.

Identification of peptides by Liquid Chromatography Mass Spectrometry (LC–MS/MS) analysis

Peptide mixtures were re-suspended in 0.2% HCOOH and analyzed by Liquid Chromatography Mass Spectrometry (LC–MS/MS), using a 6530 QTOF LC/MS (Agilent, Santa Clara, CA, USA) system equipped with a nano-HPLC. After loading, peptide mixtures were first concentrated and desalted on the pre-column. For proteins identification, the raw data obtained from the LC–MS/MS analysis were used to search both proteins databases by an in-house version of the Mascot software. The selected parameters for peptide identification were the following: Fixed Modifications: Carbamidomethyl (C), Variable modification: Oxidation (M), Gln- > pyro-Glu (N-term Q), pyrocarbamidomethyl (N-term C) Max, missed cleavages.

Radiolabelling and quality controls (QCs)

The labelling of each batch was performed with GAIA synthesis module (Elysia-Raytest®, Belgium) using a Germanum-68/Gallium-68 (Eckert-Ziegler, Berlin, Germany) generator, under GMP condition to obtain a sterile product, with incubation at room temperature for 15 min followed by tC2 purification with EtOH to eliminate unbound Gallium-68.

Briefly, 540 MBq of a [68Ga]Ga-Cl3 solution eluted from Germanum-68/Gallium-68 generator were transferred to a sterile glass vial containing 50 µg of previously conjugated IL2. After 15 min of incubation at room temperature, a small aliquot was used to assess the labelling efficiency (LE) by instant thin layer chromatography (iTLC).

At the end of the radiolabelling process, we obtained a single dose contained 5 ml [68Ga]Ga-IL2 (20–30 μg since approximately 50% of IL2 remains in the tC2 column) in a 0.9% NaCl solution with 10% EtOH (dose 74–111 MBq; pH = 6–8).

The percentage of labelling yield (LY) was automatically calculated by the GAIA software, due to the presence of radio-detectors installed in the unit.

The percentage of radiochemical purity (RCP) was calculated by iTLC using two different mobile phases to distinguish free isotope, colloids and [68Ga]IL2:

1)

MeOH/NH4OAc (1:1), for determining the LY

2)

5% NaCl/MeOH/25% NH3 (3:1:1), for determining the presence of colloids.

Negligible amount of free or colloidal Gallium-68 was observed (< 3%) at radio-chromatogram.

For each batch the specific activity (SA), in MBq/µg was calculated by using the following formula:

$$-\text\;(\text)/\text\;(\mu\text)$$

In addition to iTLC, high-performance liquid chromatography (HPLC) was also performed to separate, identify and quantify all components.

In-vitro binding to activated T-lymphocytesIsolation of human lymphocytes subsets and phenotypic analysis with fluorescence-activated cell sorter (FACS)

To evaluate the expression of IL2R α-chain (CD25) on human peripheral blood mononuclear cells (hPBMC) we used fluorescence-activated cell sorter (FACS, Becton–Dickinson, Franklin Lakes, New Jersey, USA) analysis.

hPBMC were isolated through density gradient centrifugation from the buffy coats of 4 healthy donors, and stimulated 72 h in culture with phytohemagglutinin-M (PHA-M).

Cells were centrifuged and resuspended in PBS at optimal concentration (1 × 107/mL). The monoclonal antibodies anti-human CD25 (ThermoFisher Scientific, Roma, Italy) were added to the cells and incubated for 20 min at room temperature. Then, a small aliquot of cells was analyzed by FACS.

Kinetic cell binding assay

For kinetic binding assay, hPBMC from 3 different healthy donors were activated in culture for 3 days in presence of PHA-M (5 mg/mL) and then plated in a Petri-dish suitable for LigandTracer (Uppsala, Sweden) as described elsewhere [32, 33].

To allow the attachment of the suspended cells to the Petri-dish, the surface of glass-bottomed culture dish was treated aseptically with 4 mL of BAM (Sunbright® OE-040CS, NOF Corporation, New York, USA) (2 mg/mL in Milli-Q® water) and incubated for 30 min at room temperature [33]. After washing, 4 mL of cells (2.5 × 106/mL) were plated, left for 40 min at room temperature, then washed with PBS to remove cells not immobilized. Radiolabelled IL2 was added to the cells and counting started for 60 min. After the initial 60 min of association phase, the medium containing un-bound radioligand was gently replaced with fresh PBS, initiating the dissociation phase. The kinetic of cell binding over time, as well as the dissociation from cells, was evaluated for three of the four IL2 formulations. The drop in signal observed after 60 min corresponds to the beginning of the dissociation phase, following the replacement of the radioligand-containing solution. Raw counts per minute (cpm) of cells were corrected for background signal and for isotope decay. Fitting of the kinetic binding data was not constrained to zero at the final timepoint, allowing the model to account for residual binding. Three independent experiments were performed for each IL2 formulation.

Biodistribution studies in mice

Animal experiments were carried out in compliance with the local ethics committee and in agreement with the National rules and the EU regulation (Study 204/2018‐PR). Biodistribution studies were performed with [68Ga]Ga-NODAGA-desIL2, [68Ga]Ga-THP-hrIL2 and [68Ga]Ga-THP-desIL2 in 36 BALB/c mice (8‐week‐old females) purchased from Envigo (Indianapolis, Indiana, USA). Mice were divided in 3 groups of 12 mice each, according with the 3 different batches of IL2 to be tested.

Approximately 3.7 MBq (100 µCi of labelled IL2 in 100 μl) were injected into the tail vein of each animal.

Due to different SA of different IL2 formulations, in order to inject 3.7 MBq (100 µCi), we administered approximately 6.5 µg of [68Ga]Ga-NODAGA-desIL2, 0.6 µg of [68Ga]Ga-THP-hrIL2 and 0.5 µg of [68Ga]Ga-THP-desIL2.

After each time point (15 min, 1 h, 2 h) 4 mice from each group were anaesthetized and sacrificed; blood samples and major organs (bowel, kidneys, spleen, stomach, liver, muscle, bone and lungs) were collected and weighted. Radioactivity was counted in a single-well γ-counter (PerkinElmer, Waltham, MA, USA) and the percentage of injected dose per organ (%ID) and percentage of injected dose per gram (%ID/g) were calculated, as well as the spleen/blood, spleen/muscle and spleen/bone ratios, being the spleen the only organ on which we expect an accumulation of radiolabelled IL2 due to the presence of activated T lymphocytes in physiological conditions.

Development of a ready-to-use lyophilized kit for labelling with 68Ga

As previously mentioned, after conjugation, THP-desIL2 was freeze-dried, lyophilized and stored at -80 °C. Different vials were reconstituted after 1, 2, 3 or 6 months.

Sodium-dodecyl-sulphate poly-acrylamide gel electrophoresis (SDS-PAGE) and HPLC were performed at each time point after the reconstitution of each batch, before radiolabelling.

iTLC and HPLC were also performed after 68Ga-labelling to evaluate SA, LY and RCP.

In vitro stability over time was evaluated in human serum at 37 °C. The radiolabelled compound was incubated in 1 ml final solution and an aliquot of 100 µl was analyzed by iTLC after 10 min, 1 h, 2 h, 3 h and 4 h to calculate free Gallium-68 released from IL2.

Biodistribution and dosimetry of 68Ga-THP-desIL2 in humans

Phase-I human study [68Ga]Ga-THP-desIL2 was approved by the Ethic Committee of “Sapienza” University of Rome on the 15/05/2020 (SMART CIG: Z4B316CBDF; EudraCt number: 2020–001749-38), and all subjects signed an informed consent form.

After performing all quality controls of [68Ga]Ga-THP-desIL2, the mean and standard deviation of the administered mass and activity of [68Ga]Ga-THP-desIL2 were 25.4 ± 3.1 µg (range 22–30 µg) and 159.6 ± 36.9 MBq (range 78–206 MBq) (4.31 ± 0.9 mCi), respectively. Whole body PET/CT images were acquired at 15’, 30’, 60’, 90’, 120’, 150’ and 210’ after i.v. administration in 5 healthy male volunteers (age range: 28–45 years) to assess biodistribution and for dosimetric purposes (IDAC-Dose 2.1). THP-desIL2 was previously lyophilized and stored at -80 °C. All subjects were investigated with THP-desIL2 stored for less than 3 months and no adverse events occurred after the administration of the radiopharmaceutical.

Decay corrected whole-body PET/CT scan was acquired with a dedicated hybrid PET/CT Biograph (Siemens, Germany). After a scout CT for the definition of field of study, a low-dose CT scan (120 mA), without contrast, was acquired for attenuation correction.

Both qualitative and semi-quantitative assessment by using maximum and mean standardized uptake value (SUVmax and SUVmean, respectively) were performed, to assess the in vivo biodistribution of [68Ga]Ga-THP-desIL2 in each organ.

Semi-automatic organ segmentation was performed to calculate organ volumes and to construct time-activity curves for each organ. Data from all volunteers were pooled; average organ volumes and average time-activity curves were uploaded on IDAC-Dose2.1 software for absorbed dose and effective dose calculation.

Statistical analysis

The statistical analysis was performed by SAS v.9.4 (SAS Institute Inc., Cary, NC, USA). Continuous variables are shown as mean ± standard deviation (SD) and 95%CI (Confidence Interval). To compare values of LY, SA and RP of lyophilized kits over time, as well as the stability of [68Ga]Ga-THP-desIL2 in serum at 37 °C, we used Generalized Linear Mixed Model (GLIMMIX) for repeated measurements. The comparison of binding affinity of different IL2 formulations and of Spleen/Blood ratios in mice was performed by GLIMMIX/GLM procedure, considering Gaussian function as distribution. The normality of distribution of residuals was verified by Shapiro–Wilk test and checking Q-Q plot. Homoscedasticity was verified analyzing the studentized residuals. Post hoc analysis was performed by Tukey’s method. A p-value < 0.05 was considered statistically detectable.