Design and development of a new vector platform for CXCR4-targeted molecular imaging and radionuclide therapy

The CXC chemokine receptor 4 (CXCR4) is a well-described G protein-coupled receptor (GCPR) that is involved in different physio-pathological processes upon binding of its endogenous ligand CXC chemokine 12 (CXCL12). CXCR4 is expressed throughout development and adulthood on a variety of cell types including lymphocytes, endothelial, epithelial and hematopoietic stem cells (HSCs). CXCR4 plays a fundamental physiological role in hematopoiesis, immune response, neurogenesis, germ cell development, cardiogenesis and vascularization, while in cancer, the CXCL12/CXCR4 axis is a crucial regulator of primary tumor growth and formation of metastases. To date, CXCR4 overexpression has been reported in more than 20 human cancer types ranging from solid tumors to hematologic malignancies. Furthermore, high levels of CXCR4 are also observed at inflammation sites, where CXCR4-expressing immune cells join to defend the host organism. Due to its multifunctional role, CXCR4 is an interesting target for drug development. The use of non-invasive molecular imaging for diagnosis has already been used for decades. In particular, 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG) is the gold standard for diagnostic positron emission tomography (PET) imaging of cancer and inflammation in the clinic. Although widely used, [18F]FDG lacks specificity as it binds to all tissues that have increased glucose metabolism. Hence, a radiopharmaceutical with improved specificity for cells or tissue involved in the above-mentioned pathologies is of high interest. A radiopharmaceutical targeting CXCR4 would increase the specificity and would be of high value for diagnosis and would allow patient selection for CXCR4-targeted radionuclide therapy (TRNT). So far, [68Ga]PentixaFor and its therapeutic companion [177Lu]PentixaTher are the best theranostic pair in the class of CXCR4-targeted radiopharmaceuticals. [68Ga]PentixaFor has been used clinically in a wide range of hematologic cancers and has been explored to detect atherosclerotic lesions and to determine the extent of infarcted tissue after acute myocardial infarction (AMI). Moreover, [68Ga]PentixaFor PET imaging has been used to guide treatment with [177Lu]PentixaTher, which has found to be responsive in patients with hematologic malignancies. Despite the clinical potential of this theranostic pair, there is still room for improvement in terms of the physical characteristics of the diagnostic ligand and the pharmacokinetic profile of the therapeutic ligand. Indeed, implementation of 68Ga-labeled tracers in clinical practice is often limited due to the low batch production yield, related to the current generation of 68Ge/68Ga generators, the high cost of generators, and relatively short half-life (67.7 min) of gallium-68. In contrast, fluorine-18 combines several advantages, such as favorable decay properties and low β+-trajectory (< 2 mm in water) and is readily produced in large quantities with a cyclotron. The half-life of 109.8 min is long enough to allow multistep synthesis, transport to remote hospitals without an on-site cyclotron, and allows a large time window between injection and scanning that is yet short enough to avoid extended irradiation of patients. Unfortunately, attempts to make a suitable fluorine-18 derivative of [68Ga]PentixaFor were unsuccessful due to the pronounced sensitivity of the PentixaFor scaffold towards even minor structural modifications, leading to strongly decreased CXCR4 binding affinity. On the other hand for [177Lu]PentixaTher, the iodination of Tyr1 led to an increase in lipophilicity compared to [68Ga]PentixaFor. As a result, hepatic activity levels remain persistently high, up to seven days post injection (p.i.) in patients, which makes accurate dosimetry studies with [68Ga]PentixaFor challenging. Therefore, this doctoral dissertation describes the evaluation of newly developed vector platforms for CXCR4-targeted molecular imaging and radionuclide therapy that can be labeled with fluorine-18 and therapeutic radionuclides without changing the pharmacokinetic profile of the radioligands. We first selected DV1-k-(DV3) as a potential vector platform for the development of CXCR4-targeting radioprobes owing its unique characteristics. It is a bivalent peptide entirely composed of D-amino acids and derived from the N-terminus of the viral macrophage inflammatory protein-2 (vMIP-II), which binds CXCR4 with high affinity (half-maximal inhibitory concentration, IC50: 3.0 nM). DV1-k-(DV3) was successfully derivatized for labeling with different radionuclides, for both diagnostic and therapeutic purposes. All DV1-k-(DV3)-based ligands displayed high in vitro binding affinity for human CXCR4 (hCXCR4) and murine CXCR4 (mCXCR4), potentially allowing translational imaging of CXCR4 upregulation in preclinical tumor and inflammation models. [18F]AlF-NOTA-DV1-k-(DV3) was produced in an automated synthesis module using a cGMP-compatible Al18F-labeling protocol, facilitating future clinical implementation. Furthermore, the diagnostic radiotracers, [18F]AlF-NOTA-DV1-k-(DV3) and [68Ga]Ga-DOTA-DV1-k-(DV3), and their therapeutic companion [177Lu]Lu-DOTA-DV1-k-(DV3), exhibit the same favorable pharmacokinetic profile in vivo, showing the potential of DV1-k-(DV3) for theranostic applications. [18F]AlF-NOTA-DV1-k-(DV3) displayed in vivo specific binding to hCXCR4 and mCXCR4 in hCXCR4-expressing tumor-bearing mice. However, the in vivo tumor uptake of [18F]AlF-NOTA-DV1-k-(DV3) did not reflect the high in vitro binding affinity for hCXCR4 of AlF-NOTA-DV1-k-(DV3). High binding affinity towards mCXCR4 of the DV1-k-(DV3) scaffold and high expression of mCXCR4 in mouse liver make evaluation of the clinical potential of this new class of CXCR4-targeting radiotracers in tumor mouse models difficult. To further improve the D-peptides as vector platform for applications in nuclear medicine, we designed a new D-peptide-based radioligand with improved CXCR4 binding affinity (IC50: 1.3 nM) compared to DV1-k-(DV3). During screening of DV1-k-(DV3), we observed that oxidation of both cysteines in the DV1 sequence resulted in complete loss of binding affinity towards CXCR4. The moderate tumor uptake kinetics of [18F]AlF-NOTA-DV1-k-(DV3) might be a consequence of in vivo oxidation. To prevent formation of internal disulfide bridges in vivo, we substituted both cysteine residues in the new vector molecule. Although hypothesized that the higher binding affinity of the new construct 2xDV1(c11sc12s) would result in improved tumor binding kinetics, [18F]AlF-NOTA-2xDV1(c11sc12s) displayed a similar tumor uptake compared to [18F]AlF-NOTA-DV1-k-(DV3). Additionally, the enhanced mCXCR4 binding affinity of [18F]AlF-NOTA-2xDV1(c11sc12s) resulted in an even higher mCXCR4-specific accumulation in the liver. Even though we initially hypothesized that the human-mouse cross reactivity of the D-peptides would have an advantage over the hCXCR4 selectivity of [68Ga]PentixaFor, the high binding affinity for mCXCR4 resulted in low concentrations of [18F]AlF-NOTA-2xDV1(c11sc12s) at the target tissue in both a hCXCR4-expressing tumor mouse model and in a mCXCR4-upregulated sterile inflammation model. The high liver expression of mCXCR4 impedes the investigation of the potential of [18F]AlF-NOTA-2xDV1(c11sc12s) as CXCR4-targeting radioligand in preclinical models and results in an unfair comparison with clinically used radiopharmaceuticals. As CXCR4 is not expressed in healthy human liver, these findings in mice are not predictive for the potential clinical performance of this novel class of CXCR4-targeting radiotracers. Hence, we evaluated [18F]AlF-NOTA-2xDV1(c11sc12s) in a non-human primate to predict its clinical biodistribution. In this species, [18F]AlF-NOTA-2xDV1(c11sc12s) displayed a similar biodistribution profile as in mice with high liver uptake. Although we expected a similar hepatic CXCR4 expression in a non-human primate as in humans, our study suggests expression of CXCR4 in liver of non-human primates, although further evaluation is needed to confirm that the observed liver uptake is indeed CXCR4-specific. From this large preclinical dataset, it can be concluded that we developed a potentially promising and versatile tool for CXCR4-targeted molecular imaging, but that only clinical evaluation allows to properly estimate its performance as a CXCR4 imaging agent. We are therefore performing the first-in-man trial with [18F]AlF-NOTA-2xDV1(c11sc12s) to evaluate and compare its clinical performance with [68Ga]PentixaFor. To overcome the challenges of the human-mouse cross reactivity in preclinical evaluation, we selected two additional vector molecules solely binding hCXCR4 for radiopharmaceutical development. Radiolabeled nanobodies (Nbs), VUN400 and VUN401, demonstrated favorable in vitro binding characteristics. However, during in vivo evaluation in wild-type mice, we observed low in vivo plasma stability of the radiolabeled Nb constructs. We first need to improve the stability of the radiolabeled constructs before we can initiate the next steps in the development process. Overall, we have evaluated a versatile platform for CXCR4-targeted molecular imaging with a wide range of future applications. Evidently, CXCR4 is an interesting target for (radio)pharmaceutical development of which its complex interactions still need to be further elucidated to better understand its role in diseases.

Publication year:2022

Accessibility:Open