Introduction
Increasing data is becoming available reporting the intense crosstalk between cancer cells and the tumor microenvironment. Tumor-associated macrophages (TAM) in the microenvironment play a key role in tumor growth and progression. Besides local immunosuppression, they influence angiogenesis and cancer cell mobility. Importantly, the TAM population consists of pro- and anti-tumoral populations residing in different tumor regions. In this respect, TAM with a high expression of MMR (MMR
hi TAM) were shown to reside in hypoxic tumor areas and were highly angiogenic and strongly immunosuppressive, characteristics that suggest a strong protumoral activity [
1‐
3].
MMR
hi TAM were reported to negatively impact therapy responsiveness and allow tumor relapse following irradiation, anti-angiogenic, or vascular disrupting therapies and chemotherapy in preclinical tumor models [
4]. A clear association between the presence of MMR
hi TAM and the (lack of) responsiveness to treatment has not been demonstrated so far in patients, likely because it is difficult to fully capture the heterogeneity of TAM presence using core needle biopsy and immunohistochemistry (IHC).
To determine MMR
hi TAM density, we propose the use of
68Ga-labeled anti-MMR single-domain antibodies fragments (sdAbs) for the non-invasive imaging of MMR
hi TAM in all cancer lesions within a patient. sdAbs, such as anti-MMR-sdAbs, are antigen-binding fragments derived from heavy-chain only antibodies of the Camelidae. Compared to conventional antibodies or antibody fragments, sdAbs are small (12–15 kDa) and they have the ability to bind antigens on hidden or unusual epitopes [
5]. sdAbs have proven their role in molecular imaging and targeted radionuclide therapy in a preclinical setting [
6‐
9], and a
68Ga-labeled compound for positron emission tomography (PET)/X-ray computed tomography (CT) imaging of HER2 was found safe and it is use straightforward in a phase I clinical trial, with low radiation burden for the patient [
10].
Previously, we have reported the generation of cross-reactive anti-mouse/human MMR-sdAbs, selection, and validation of the lead compound for imaging of MMR-expressing macrophages using radiolabeling with
99mTc and
18F [
11]. Here, we report the synthesis and validation of the clinical-grade [
68Ga]Ga-NOTA-anti-MMR-sdAb compound for PET/CT imaging, its biodistribution, dosimetry, and toxicity studies in animal models, making [
68Ga]Ga-NOTA-anti-MMR-sdAb ready for application in a phase I clinical trial.
Materials and Methods
All commercially obtained chemicals were of analytic grade. p-SCN-Bn-NOTA was purchased from Macrocyclics. 68Ga was obtained from a 68Ge/68Ga Galli Eo™ generator (IRE, Belgium). Buffers used for coupling reactions or for radiolabeling were purified from metal contamination using Chelex 100 resin (Aldrich). High purity water (Fluka) was used for radiolabeling.
Production and Purification of GMP-Grade anti-MMR-sdAb
The DNA sequence coding for the cross-reactive human-mouse-anti-MMR-sdAb, described in [
11], was cloned in the pAOXZalpha plasmid in frame with the α-factor secretory signal peptide. pAOXZalpha vector is a derivative of the pPICZalpha plasmid (Thermo Fisher), with minor changes in the multiple cloning site.
Pichia pastoris strain GS115 + His4 was transduced and a clone with stable genomic integration and high sdAb secretion in the medium was selected. The untagged sdAb was purified from the fermenter supernatant by 0.2-μm filtration, mixed-mode chromatography (MMC) capturing and polishing using anion-exchange chromatography (AEX) on NatriFlo a HD-Q Recon column (Natrix). The sdAb was buffer-exchanged to phosphate buffer saline (PBS) by tangential flow filtration and concentrated to 1.9 mg/ml. The GMP-grade sdAb was generated by Q-Biologicals (Ghent).
Chromatographic Analysis
Superdex Peptide 10/300GL (GE), using 0.1 M ammonium acetate pH 7 as eluate and flow 0.5 ml/min, was used for NOTA-anti-MMR-sdAb purification and quality control. Superdex Peptide 3.2/300 column (GE), using 0.02 M PBS/0.28 M NaCl pH 7.4 as mobile phase was used for the quality control of 69,71GaNOTA-anti-MMR-sdAb and [68Ga]Ga-NOTA-anti-MMR-sdAb. The same column was also used when performing stability studies of [68Ga]Ga-NOTA-anti-MMR-sdAb. Instant thin-layer chromatography (iTLC) was performed on silica gel (SG) (Pall Corp. Life Sciences) using 0.1 M sodium citrate pH 5.0 as eluent.
Conjugation of p-SCN-Bn-NOTA to anti-MMR-sdAb
Anti-MMR-sdAb (3 mg, 0.24 μmol) was buffer-exchanged to 0.05 M sodium carbonate buffer, pH 8.7, using PD-10 size exclusion disposable columns (GE). Protein solution (2 ml) was added to a 20-fold molar excess p-SCN-Bn-NOTA (2.6 mg, 47 μmol), pH adjusted to 8.5–8.7 with 0.2 M Na2CO3. After 2-h incubation at room temperature (RT), the pH of the reaction mixture is lowered to pH 7.4 by adding HCl 1 N. The NOTA-anti-MMR-sdAb protein solution is loaded on a size exclusion column. The collected fractions containing monomeric NOTA-anti-MMR-sdAb protein are pooled and the solution is passed through a 0.22-μm filter. The protein concentration is determined by UV absorption at 280 nm (ε = 40,660 M−1 cm−1). Three validation tests were performed in which the product was fully characterized as required for clinical application. Product was stored at − 20 °C and stability tests were performed at selected time points (T0M, T3M, T6M at − 20 °C and T3d, T7d at 4 °C).
Synthesis of 69,71GaNOTA-anti-MMR-sdAb
To a solution of NOTA-anti-MMR-sdAb (1.2 mg, 91.4 nmol) in 1 ml of 0.1 M ammonium acetate (pH 7), 250 μl of 1 M sodium acetate buffer pH 5 and 35 μl of a 60 mM solution of Ga(NO3)3·10H2O (1.83 μmol) were added, and incubated for 2 h at RT. Size exclusion purification using PD-10 column pre-equilibrated with 0.1 M ammonium acetate (pH 7) was used to remove free gallium acetate. 69,71GaNOTA-anti-MMR-sdAb was characterized by ESI-Q-ToF-MS, SDS-PAGE/Western Blot, surface plasmon resonance (SPR), and SEC. The compound was used as the reference for the identification of the radio-SEC chromatogram signals of [68Ga]Ga-NOTA-anti-MMR-sdAb (retention time [tR] = 8.4 min).
SPR
SPR measurements were performed on a Biacore T200 instrument (GE) as described previously [
11].
Preparation of [68Ga]Ga-NOTA-anti-MMR-sdAb
68Ga eluate (500–877 MBq) was added to NOTA-anti-MMR-sdAb (100–130 μg, 7.7–11 nmol) in 1 M sodium acetate buffer pH 5 (1 ml). The homogenized solution is incubated for 10 ± 1 min at RT to form [68Ga]Ga-NOTA-anti-MMR-sdAb. The solution is loaded on a PD-10 column (GE), which is pre-equilibrated with 0.9 % NaCl 5 mg/ml ascorbic acid at pH 6. The column is eluted with 0.9 % NaCl 5 mg/ml ascorbic acid at pH 6 by gravity flow and the eluate is passed through a 0.22-μm filter (Millex GV, Millipore). Radiochemical purity was evaluated using iTLC and by radio-SEC (SEC: [68Ga]Ga-NOTA-anti-MMR-sdAb tR = 10.2 min; [68Ga]Gacitrate tR = 15.9 min; iTLC-SG: [68Ga]Ga-NOTA-anti-MMR-sdAb Rf = 0, [68Ga]Gacitrate Rf = 1).
In Vitro Stability of [68Ga]Ga-NOTA-anti-MMR-sdAb
Stability of [68Ga]Ga-NOTA-anti-MMR-sdAb was tested in 0.9 % NaCl 5 mg/ml ascorbic acid at pH 6, with a 1000-fold molar excess diethylenetriaminepentaacetic acid (DTPA) and in human plasma. [68Ga]Ga-NOTA-anti-MMR-sdAb (349 ± 55 MBq, n = 4) in 0.9 % NaCl 5 mg/ml ascorbic acid pH 6 at RT was followed by iTLC up to 4 h. [68Ga]Ga-NOTA-anti-MMR-sdAb (22.7 MBq) was added to 500 μl of human plasma incubated at 37 °C up to 1 h and analyzed by SEC. To a solution of [68Ga]Ga-NOTA-anti-MMR-sdAb (22 ± 4 MBq, n = 2) was added a 1000-fold molar excess of DTPA and incubated at RT for 4 h and followed by iTLC.
Animal Models
Wild-type (WT) female C57BL/6 mice (Janvier) were used for blood curves (9 weeks old) and dosimetry studies (7 weeks old). To evaluate biodistribution and targeting specificity, female C57BL/6 WT and MMR-deficient (MMR-KO) (Janvier) (7 weeks old) mice were subcutaneously inoculated in the right flank with 3 × 106 of the 3LL-R clone of Lewis lung carcinoma cells suspended in HBSS medium while anesthetized with 2.5 % isoflurane (ABBOTT). Tumors were allowed to grow for 12 days (tumor weight of 0.662 ± 0.267 g for WT and 0.730 ± 0.268 g for MMR-deficient).
The animals were housed at 22 °C in 50–60 % humidity with a light/dark cycle of 12 h. They were kept under pathogen-free conditions and were given autoclaved food pellets and water ad libitum. For animal handling and processing of data, technicians and researchers were not blinded.
All procedures followed the guidelines of the Belgian Council for Laboratory Animal Science and were approved by the Ethical Committee for Animal Experiments of the Vrije Universiteit Brussel (license 13-272-5).
Biodistribution Studies in Tumor-Bearing Mice
3LL-R tumor-bearing mice (C57BL/6 WT (n = 5) and MMR-KO (n = 5)), anesthetized with 2.5 % isoflurane, were injected intravenously with [68Ga]Ga-NOTA-anti-MMR-sdAb (11.6 ± 0.6 MBq, 5 μg of NOTA-anti-MMR-sdAb) via the tail. Mice were euthanized at 3 h after injection, and major organs were collected, weighed, and counted against a standard of known activity in a gamma counter. Tissue/organ uptake was calculated and expressed as a percentage injected activity (%IA) per gram, corrected for decay.
Blood Curves
WT C57BL/6 mice (n = 6) anesthetized with 2.5 % isoflurane were injected intravenously with [68Ga]Ga-NOTA-anti-MMR-sdAb (4.7 ± 0.2 MBq, 5 μg of NOTA-anti-MMR-sdAb) via the tail. Blood samples were collected 2, 5, 10, 20, 30, 40, 60, 120, and 180 min after injection, via a microcapillary, and analyzed in the gamma counter to obtain a blood time–activity curve. The half-life was calculated by biexponential nonlinear regression fit (GraphPad Prism; GraphPad Software).
Dosimetry Studies
WT C57BL/6 mice (n = 6 per time point) anesthetized with 2.5 % isoflurane were injected intravenously with [68Ga]Ga-NOTA-anti-MMR-sdAb (5.1 ± 0.5 MBq, 5 μg of NOTA-anti-MMR-sdAb) via the tail. Animals were euthanized at 10, 60, 120, and 180 min post injection and major organs were collected, weighed, and counted against a standard of known activity in a gamma counter. Tissue/organ uptake was calculated and expressed as %IA and %IA/g, corrected for decay. Radiation dose estimates for the adult female and male were calculated from the biodistribution data of mice using OLINDA 1.0 software. Organ doses, effective dose, and effective dose equivalent were calculated using the appropriate weighing factors for the various organs. Full methods are described in supplemental data.
Micro-PET/CT Imaging
Mice were injected with [68Ga]Ga-NOTA-anti-MMR-sdAb (11.62 ± 0.58 MBq, 5 μg NOTA-anti-MMR-sdAb) of the tracer via the tail vein, and scanned after 60 and 150 min. Micro-PET/CT imaging was performed on a MILabs VECTor/CT. The CT scan was set to 55 keV and 615 μA, resolution of 80 μm, with a total body scan duration of 108 s. PET images were obtained using the high-energy PET collimator in spiral mode, 94 positions for whole-body imaging, with 9 s per position. Images were reconstructed with 0.6-mm voxels with 4 subsets and 2 iterations, without post reconstruction filter. Image viewing was performed with AMIDE imaging software.
Toxicity Study with NOTA-anti-MMR-sdAb
Full materials and methods of the toxicity study are described in supplemental data. A 7-day intravenous repeated dose toxicity study of the NOTA-anti-MMR-sdAb was performed (Eurofins BioPharma Product testing). The ICH M3 (R2) guideline was applied (EMA/CPMP/ICH/286/95).
NOTA-anti-MMR-sdAb or vehicle solution (PBS) was administered at 1.68-mg/kg body weight daily via intravenous route to 20 male and 20 female Balb/c mice for 7 days. Toxicity was assessed using observation, blood sampling (hematology and clinical biochemistry), necropsy, and urine analysis. Histopathological examination was performed for 40 selected organs. In addition, 10 recovery animals per group and gender were observed for 14 days following the last administration and were assessed thereafter.
Also, blood cytokine levels were determined in 5 satellite animals per group and per gender at 2 h and 6 h after single administration. Cytokine determination was performed using V-PLEX Plus Proinflammatory Panel 1 (mouse) kit.
Statistical Analysis
Quantitative data are expressed as mean ± SD and compared using the independent t test using Prism 5 (GraphPad Software, Inc.)
Discussion
In oncology research, increasing attention is going to the tumor microenvironment, with a high interest in the local immune interactions, to better understand primary and acquired resistance to different types of immunotherapy. It is hypothesized that the
in vivo imaging of key molecular markers such as programmed death-ligand 1 (PD-L1) and tumor-infiltrating lymphocytes can help to understand such resistance mechanisms [
12]. Also TAM are believed to play a role in primary resistance to immunotherapy. Macrophages found in the tumor microenvironment display a broad spectrum of molecular markers, and some of these markers are associated with a protumoral and immune-suppressive behavior. As such, macrophage mannose receptor (MMR) has been identified as a distinct marker, present on M2-polarized TAM that promote tumor growth and metastasis and inhibit local immune activation [
1‐
3]. The presence of such MMR-expressing TAM could therefore be indicative for the success rate of immune-activating therapies. To further investigate this role in cancer patients, we have developed a PET/CT imaging method using an MMR-targeting sdAb and
68Ga-labeling enabling PET/CT quantification.
The MMR-specific sdAb that recognizes both the mouse and the human MMR target was GMP-produced without any c-terminal tag. The bifunctional chelator p-SCN-Bn-NOTA was conjugated to the GMP-grade compound and radiolabeled with 68Ga. The compound proved to be stable both in the final buffer and in human plasma, and the compound was resistant to trans-chelation, confirming high stability of the [68Ga]Ga-NOTA complex. The affinity of the different compounds was around 1 nM and not affected either by conjugation of NOTA or complexation with 69,71Ga. These in vitro characterization tests confirm its suitability for further translation to patient use.
In mice, [
68Ga]Ga-NOTA-anti-MMR-sdAb accumulate specifically in MMR-expressing organs such as the liver and spleen, as well as in the 3LL-R tumors that are known to be infiltrated with high amounts of MMR
hi TAM [
1]. Specificity was confirmed by the absence of signal in MMR-deficient mice (MMR-KO), both in healthy organs and in 3LL-R tumors. The tracer was rapidly cleared from the blood
via the kidneys and urine, as observed by both PET/CT imaging and
ex vivo biodistribution analysis. The confirmation of specific targeting and rapid clearance of unbound compound provides the necessary evidence of its potential added value for subsequent use in patients.
When comparing uptake values of the [
68Ga]Ga-NOTA-anti-MMR-sdAb with the previously reported [
99mTc]Tc(CO)
3-anti-MMR-sdAb and [
18F]FB-anti-MMR-sdAbs, the tumor uptake for all three compounds is very similar in WT 3LL-R tumor-bearing mice [
11]. Uptake in extra-tumoral MMR-expressing organs, such as the liver, spleen, lymph nodes, and bone, is slightly higher for the
68Ga- compound when compared to the
18F version, but substantially lower than what was observed for the
99mTc compound. These differences may be due to different radiochemistry methods or differences in apparent molar activity (70.1, 31.5, 10.1 GBq/μmol for
99mTc,
68Ga,
18F, respectively (injection time)) [
11]. Kidney uptake was also different for the three variants, with the lowest value obtained for the [
18F]FB-anti-MMR-sdAb (7.98 ± 0.86 %IA/g at 3 h p.i.), which is about 20-fold lower than that seen with [
99mTc]Tc(CO)
3-anti-MMR-sdAb and 10-fold lower than with the here presented [
68Ga]Ga-NOTA-anti-MMR-sdAb [
11]. Similar lower kidney retention for
18F-labeled sdAb has also been reported for anti-HER2, although only for the [
18F]FSB, and not using [
18F]RL-I prosthetic group, confirming that the radiochemistry is a defining factor [
13‐
15]. Depending on the chemical structure created, different catabolites will be produced in the kidney cells, with some
18F-catabolites clearing faster from kidneys. Although the [
18F]FB-anti-MMR-sdAb compound holds promise for clinical translation and might even perform better given its lower kidney retention, our research group has chosen to initiate clinical trials with the [
68Ga]Ga-NOTA-anti-MMR-sdAb because of the ease of production and radiolabeling that could be implemented in many radiopharmacies worldwide, even in the absence of cyclotrons or synthesis modules.
For subsequent human use of [
68Ga]Ga-NOTA-anti-MMR-sdAb, the average radiation dose for a human adult was estimated based on extrapolation of normal distribution and retention in mice. The average effective dose was 0.034 and 0.027 mSv/MBq for female and male respectively, resulting in a total body radiation dose per injection of 185 MBq of
68Ga compound of 6.3 and 5.0 mSv for female and male, respectively. The organs that would receive the highest dose are the kidneys (up to 90 mGy) and the urinary bladder wall (up to 42 mGy), staying well below the kidney threshold of 7–8 Gy for potential deterministic effects [
16]. The total body irradiation dose is in the same range as a standard 2-deoxy-2-[
18F]fluoro-D-glucose PET scan (7 mSv for 370 MBq of administered activity), and could be used in daily clinical practice for diagnosis and follow-up of patients. In the toxicity study, NOTA-anti-MMR-sdAb showed no adverse effects after the 7-day repeated injected up to a dose of 1.68 mg/kg. This dose is a 1000-fold excess of what a 60-kg patient would receive when injected with 0.1 mg of the compound. Based on all preclinical data reported here, [
68Ga]Ga-NOTA-anti-MMR-sdAb can be regarded as safe for clinical translation. A phase I trial to assess safety, biodistribution and dosimetry in cancer patients will be initiated in the near future.
Other research groups have also investigated macrophage imaging using mannose receptor targeting. Many have focused on mannosylation of the targeting moiety, thereby utilizing the natural ligand of MMR [
17‐
19]. However, mannose does not only bind MMR, but also multiple other mannose-binding proteins, such as mannose-binding lectins. Such lectins are serum proteins that will bind to mannose and other types of sugars that occur on the cell surface of bacteria and yeasts, thereby facilitating opsonization by phagocytes. It can be foreseen that using mannosylated compounds for imaging will therefore not only identify MMR-expressing macrophages, but will also target other phagocytes that will engulf lectin-bound compounds.
Fluorescence imaging using more specific MMR-targeting agents has been reported using a full monoclonal anti-MMR antibody and using an MMR-binding peptide [
20‐
22]. Recently, a
125I-labeled monoclonal antibody was used for SPECT imaging [
22]. These studies confirm the promise of specifically targeting MMR for the imaging of protumorigenic macrophages, but the use of sdAb offers multiple advantages compared to approaches using mAbs given their faster blood clearance and good tissue penetration characteristics.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.