Introduction
Immuno-positron emission tomography (immunoPET) is a non-invasive in vivo imaging method based on tracking and quantifying radiolabeled monoclonal antibodies (mAbs) and other related molecules, such as antibody fragments, nanobodies, or affibodies. However, the success of immunoPET in neuroimaging is limited because intact antibodies cannot penetrate the blood–brain barrier (BBB). In neuro-oncology, immunoPET has been successfully applied to brain tumors because of the compromised BBB. Different strategies, such as changes in antibody properties, use of physiological mechanisms in the BBB, or induced changes to BBB permeability, have been developed to deliver antibodies to the brain. These approaches have recently started to be applied in preclinical central nervous system PET studies. Therefore, immunoPET could be a new approach for developing more specific PET probes directed to different brain targets. The purpose of this literature review is to outline the major contributions of immunoPET to preclinical and clinical neurosciences.
Strategies to Cross the BBB
Different strategies have been used to deliver antibodies into the brain, including modifications in their physicochemical properties; physiological mechanisms, such as adsorption-mediated transcytosis (AMT) or receptor-mediated transcytosis (RMT); and induced changes in BBB permeability (Table 1).
Table 1. Summary of different strategies to cross the BBB with proteins or antibodies. Adapted from source
Strategy | Mechanism | References |
---|---|---|
Physicochemical properties modification | Poly(ethylene glycol) conjugation to increase the circulation half-life | [12,13] |
Through physiological mechanism | Cationic proteins or nanobodies that trigger adsorption-mediated transcytosis (AMT) | [14,15,16] |
Ligands or antibodies that trigger receptor-mediated transcytosis (RMT) | [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32] | |
BBB permeability changes | Open BBB with solvents such as mannitol | [19,33] |
Focused ultrasound (FUS) with microbubbles | [34,35,36,37,38] |
Antibody Engineering Strategies for ImmunoPET
Antibodies‘ capacity to cross the BBB, which is inherently limited for native antibodies, can be significantly improved by endowing the antibody with the capability to use RMT. This is achieved by generating bispecific antibodies that can bind both the RMT-associated receptor, such as TfR or insulin receptor and the target of interest in the CNS.
The modular structure of the antibody molecule makes it possible to create bispecific antibodies of many different designs, including asymmetric IgG-like constructs in which the binding specificity of one of the two inherently identical binding sites has been altered. In another common type of construct, antibody fragment(s) providing the second binding specificity are genetically fused to IgG. Typically, single chain Fv (scFv) fragments consisting of the light and heavy chain variable domains linked together with a peptide linker of approximately 15–20 amino acids are used in such constructs. Yet another way to generate bispecific constructs is to conjugate two or more antibody fragments with different binding specificities together with peptide linkers. Figure 1 shows an example of each of these three types of the bispecific antibody.
Figure 1. Examples of bispecific antibody constructs. (A) IgG-like asymmetric antibody engineered to bind an RMT-associated target with one of its binding arms and the CNS-associated target with the other binding arm. The knob-into-hole approach [41] is used to guide correct pairing of two different heavy chains, and CrossMab technology [42] is used to prevent mispairing of the light chains. (B) IgG-like antibody with scFv fragments of different binding specificity conjugated to the C-terminus of the light chains. This is a symmetric IgG-based construct. (C) di-scFv construct in which two scFvs with different binding specificities are fused together via a short peptide linker. VL and VC refer to the light chain variable and constant domains, respectively. VH, CH1, CH2 and CH3 refer to the heavy chain variable domain and the constant domains 1, 2 and 3, respectively. The roman numbers in the domain names (e.g., VL-I or VL-II) show which of the two different binding specificities the domain is associated to. Adapted from source
Radiolabeling Strategies
PET scans detect positron-emitting radionuclides, which are attached to a molecule such as a protein or an antibody. The positron is created, being short-lived, and eventually gets annihilated, converting all its mass into energy and thereby emitting two photons of 511 keV each (which is the resting energy of the electron or positron) in opposite directions. These two photons are detected in coincidence by scintillation detectors of PET scan.
Key to the labeling of antibodies, and antibodies-related molecules, is the appropriate matching between the biological half-life of the protein and the physical half-life of the radionuclide. The shorter-lived PET radionuclides 11C, 18F, 68Ga, 44Sc, and 64Cu have been used for radiolabeling antibody fragments, while the slow pharmacokinetics of intact antibodies have enabled the use of the long-lived nuclides 89Zr and 124I. PET radionuclides characteristics used to label antibodies and antibody fragments are summarized in Table 2.
Table 2. Summary of PET radionuclides characteristics used in immunoPET. Adapted from source
Radionuclide | Half-Life | Branching Ratio (Β+) (%) | Positron Energy–E Max [Mev] | Mean Positron Range (mm) |
---|---|---|---|---|
11C | 20.4 min | 99 | 0.97 | 1.2 |
18F | 109.7 min | 97 | 0.65 | 0.6 |
68Ga | 67.7 min | 89 | 1.90 | 3.5 |
44Sc | 3.97 h | 94 | 1.47 | 2.3 |
64Cu | 12.7 h | 18 | 0.65 | 0.7 |
89Zr | 78.4 h | 23 | 0.91 | 1.3 |
124I | 100.2 h | 23 | 1.54 | 4.4 |
Current Applications
Neuro-Oncology
In 2010, brain metastases were first visualized in HER2-positive breast cancer patients using [89Zr]Zr-DFO-trastuzumab. This study demonstrated brain lesions with an 18-fold higher [89Zr]Zr-DFO-trastuzumab uptake in tumors than in normal brain tissue. The brain penetration of [89Zr]Zr-DFO-trastuzumab was possible because of a disruption of the BBB at the site of the brain metastasis [69]. Another study in HER2-positive breast cancer patients using [89Zr]Zr-DFO-pertuzumab demonstrated the detection of brain metastases in these patients. In HER2-negative metastatic breast cancer patients, immunoPET following a pre-targeting approach against carcinoembryonic antigen (CEA) showed higher overall sensitivity than [18F]FDG PET imaging in disclosing metastases, including brain dissemination. In rodents, HER2-positive intracranial breast carcinoma xenografts have been shown to uptake of an 18F-labeled single-domain antibody fragment.
Figure 2. Examples of immunoPET images in preclinical and clinical neuroscience. (A) Representative fused PET/CT images of coronal and sagittal planes at 2- and 4-days post-injection containing TS543 brain tumors with [89Zr]Zr-DFO-LEM2/15. (B) PET/MR fusion images of four different pediatric patients with DIPG where four tumors showed variable uptake of [89Zr]Zr-DFO-bevacizumab (arrows). (C) Representative in vivo PET images from TgF344-AD and WT rats 3 days post-administration of [124I]I-OX265-F(ab´)2-Bapi. Adapted from source
Neurological Diseases
The first attempts to reach the brain with a preclinical immunoPET approach used poly(ethylene glycol) antibodies against amyloid-β (Aβ). 64Cu-labeled anti-Aβ mAbs 6E10, M116, and M31, showed differences in uptake between the TgCRND8 mouse model of Alzheimer’s disease and wild-type animals. In addition, radiolabeled antibodies without specific modification to cross the BBB were studied in Alzheimer‘s rodent models. However, limited penetration made them inadequate for monitoring cerebral amyloid pathogenesis.
Conclusion
The imaging diagnostic tools currently employed in neuro-oncology, computed tomography (CT) and magnetic resonance imaging (MRI), provide excellent anatomical information on the localization of brain lesions but are frequently not able to differentiate tumor tissue from other concurrent processes, such as inflammation, edema, or bleeding, resulting in under- or over-estimation of the actual extension of the tumor. In addition, following radiation and/or chemotherapy, neuro-oncologists often encounter treatment-related changes, such as pseudoprogression or necrosis. Another important issue in neuro-oncology is that the majority of GBM have gross tumor burden with an intact BBB that extends beyond the disrupted BBB. Therefore, successful treatment is only possible if an effective drug is delivered with adequate exposure to the entire population of targeted cells. The development of new antibody-based PET probes that can cross the BBB could improve the diagnosis of brain tumors and also be a great tool for managing immune and radiation therapy.
Current advances in antibody engineering have achieved successful application of immunoPET to the brain. This achievement opens the doors to improving diagnosis and treatment in neuro-oncology, neurology, and neuropsychiatry.