Antibody Engineering for Nuclear Imaging and Radioimmunotherapy

Exploring the Interplay Between Radioimmunoconjugates and Fcγ Receptors in Genetically Engineered Mouse Models of Cancer

Fcγ receptors (FcγR) are responsible for many of the interactions between immunoglobulins (IgG) and immune cells. In biomedicine, this interplay is critical to the activity of several types of immunotherapeutics; however, relatively little is known about how FcγRs affect the in vivo performance of radiolabeled antibodies. A handful of recent preclinical studies suggest that binding by FcγR-and particularly FcγRI-can affect the pharmacokinetic profiles of 89Zr-labeled radioimmunoconjugates, but there are no extant studies in immunocompetent or genetically engineered mouse models of cancer. In the investigation at hand, we synthesized and characterized 89Zr-labeled probes based on wild-type and aglycosylated variants of the CA19-9-targeting antibody 5B1 and evaluated their in vivo behavior in several murine models of cancer, including immunocompetent and FcγR-humanized mice. The aglycosylated desferrioxamine (DFO)-bearing immunoconjugate DFO-N297A5B1 displayed identical binding to CA19-9-expressing cells compared to its wild-type analogue (DFO-5B1) but exhibited dramatically attenuated affinity for several FcγR. Positron emission tomography imaging and biodistribution studies with [89Zr]Zr-DFO-5B1 and [89Zr]Zr-DFO-N297A5B1 were subsequently performed in several strains of mice bearing CA19-9-expressing BxPC3 human pancreatic ductal adenocarcinoma and B16F10-FUT3 murine melanoma xenografts. Significant differences in the pharmacokinetics of the two radioimmunoconjugates were observed in tumor-bearing immunocompromised NSG mice, but these differences failed to materialize in immunocompetent C57BL/6 and FcγR-humanized C57BL/6 mice with B16F10-FUT3 xenografts. We hypothesize that these observations are related to the presence or absence of endogenous IgG. NSG mice completely lack endogenous IgG, and thus their mFcγR are free to bind radioimmunoconjugates and alter their pharmacokinetic behavior. In contrast, C57BL/6 and FcγR-humanized C57BL/6 mice both have endogenous IgG that occupy their FcγR (murine for the former and human for the latter), precluding interactions with radioimmunoconjugates. Ultimately, these data suggest that understanding the interplay between radiolabeled antibodies and FcγR is critical during the preclinical evaluation of radioimmunoconjugates.

PET imaging of Aspergillus infection using Zirconium-89 labeled anti-β-glucan antibody fragments

PURPOSE

Invasive fungal diseases, such as pulmonary aspergillosis, are common life-threatening infections in immunocompromised patients and effective treatment is often hampered by delays in timely and specific diagnosis. Fungal-specific molecular imaging ligands can provide non-invasive readouts of deep-seated fungal pathologies. In this study, the utility of antibodies and antibody fragments (Fab) targeting β-glucans in the fungal cell wall to detect Aspergillus infections was evaluated both in vitro and in preclinical mouse models.

METHODS

The binding characteristics of two commercially available β-glucan antibody clones and their respective antigen-binding Fabs were tested using biolayer interferometry (BLI) assays and immunofluorescence staining. In vivo binding of the Zirconium-89 labeled antibodies/Fabs to fungal pathogens was then evaluated using PET/CT imaging in mouse models of fungal infection, bacterial infection and sterile inflammation.

RESULTS

One of the evaluated antibodies (HA-βG-Ab) and its Fab (HA-βG-Fab) bound to β-glucans with high affinity (KD = 0.056 & 21.5 nM respectively). Binding to the fungal cell wall was validated by immunofluorescence staining and in vitro binding assays. ImmunoPET imaging with intact antibodies however showed slow clearance and high background signal as well as nonspecific accumulation in sites of infection/inflammation. Conversely, specific binding of [89Zr]Zr-DFO-HA-βG-Fab to sites of fungal infection was observed when compared to the isotype control Fab and was significantly higher in fungal infection than in bacterial infection or sterile inflammation.

CONCLUSIONS

[89Zr]Zr-DFO-HA-βG-Fab can be used to detect fungal infections in vivo. Targeting distinct components of the fungal cell wall is a viable approach to developing fungal-specific PET tracers.

Transcatheter pseudo-vascular isolation for localization and concentration of a large molecule theranostic probe into a transgenic OncoPIG kidney tumor

INTRODUCTION

Great strides have been made identifying molecular and genetic changes expressed by various tumor types. These molecular and genetic changes are used as pharmacologic targets for precision treatment using large molecule (LM) proteins with high specificity. Theranostics exploits these LM biomolecules via radiochemistry, creating sensitive diagnostic and therapeutic agents. Intravenous (i.v.) LM drugs have an extended biopharmaceutical half-life thus resulting in an insufficient therapeutic index, permitting only palliative brachytherapy due to unacceptably high rates of systemic nontarget radiation doses to normal tissue. We employ tumor arteriole embolization isolating a tumor from the systemic circulation, and local intra-arterial (i.a.) infusion to improve uptake of a LM drug within a porcine renal tumor (RT).

METHODS

In an oncopig RT we assess the in vivo biodistribution of 99mTc-labeled macroaggregated albumin (MAA) a surrogate for a LM theranostics agent in the RT, kidney, liver, spleen, muscle, blood, and urine. Control animals underwent i.v. infusion and experimental group undergoing arteriography with pseudovascular isolation (PVI) followed by direct i.a. injection.

RESULTS

Injected dose per gram (%ID/g) of the LM at 1 min was 86.75 ± 3.76 and remained elevated up to 120 min (89.35 ± 5.77) with i.a. PVI, this increase was statistically significant (SS) compared to i.v. (13.38 ± 1.56 and 12.02 ± 1.05; p = 0.0003 p = 0.0006 at 1 and 120 min respectively). The circulating distribution of LM in the blood was less with i.a. vs i.v. infusion (2.28 ± 0.31 vs 25.17 ± 1.84 for i.v. p = 0.033 at 1 min). Other organs displayed a trend towards less exposure to radiation for i.a. with PVI compared to i.v. which was not SS.

CONCLUSION

PVI followed by i.a. infusion of a LM drug has the potential to significantly increase the first pass uptake within a tumor. This minimally invasive technique can be translated into clinical practice, potentially rendering monoclonal antibody based radioimmunotherapy a viable treatment for renal tumors.

A comparison of antibody-antigen complex sequence-to-structure prediction methods and their systematic biases

The ability to accurately predict antibody-antigen complex structures from their sequences could greatly advance our understanding of the immune system and would aid in the development of novel antibody therapeutics. There have been considerable recent advancements in predicting protein-protein interactions (PPIs) fueled by progress in machine learning (ML). To understand the current state of the field, we compare six representative methods for predicting antibody-antigen complexes from sequence, including two deep learning approaches trained to predict PPIs in general (AlphaFold-Multimer and RoseTTAFold), two composite methods that initially predict antibody and antigen structures separately and dock them (using antibody-mode ClusPro), local refinement in Rosetta (SnugDock) of globally docked poses from ClusPro, and a pipeline combining homology modeling with rigid-body docking informed by ML-based epitope and paratope prediction (AbAdapt). We find that AlphaFold-Multimer outperformed other methods, although the absolute performance leaves considerable room for improvement. AlphaFold-Multimer models of lower quality display significant structural biases at the level of tertiary motifs (TERMs) toward having fewer structural matches in non-antibody-containing structures from the Protein Data Bank (PDB). Specifically, better models exhibit more common PDB-like TERMs at the antibody-antigen interface than worse ones. Importantly, the clear relationship between performance and the commonness of interfacial TERMs suggests that the scarcity of interfacial geometry data in the structural database may currently limit the application of ML to the prediction of antibody-antigen interactions.

Site-specific bioconjugation and nuclear imaging

Monoclonal antibodies and antibody fragments have proven to be highly effective vectors for the delivery of radionuclides to target tissues for positron emission tomography (PET) and single-photon emission computed tomography (SPECT). However, the stochastic methods that have traditionally been used to attach radioisotopes to these biomolecules inevitably produce poorly defined and heterogeneous probes and can impair the ability of the immunoglobulins to bind their molecular targets. In response to this challenge, an array of innovative site-specific and site-selective bioconjugation strategies have been developed, and these approaches have repeatedly been shown to yield better-defined and more homogeneous radioimmunoconjugates with superior in vivo performance than their randomly modified progenitors. In this Current Opinion in Chemical Biology review, we will examine recent advances in this field, including the development - and, in some cases, clinical translation - of nuclear imaging agents radiolabeled using strategies that target the heavy chain glycans, peptide tags, and unnatural amino acids.

Engineered Antibodies as Cancer Radiotheranostics

Radiotheranostics is a rapidly growing approach in personalized medicine, merging diagnostic imaging and targeted radiotherapy to allow for the precise detection and treatment of diseases, notably cancer. Radiolabeled antibodies have become indispensable tools in the field of cancer theranostics due to their high specificity and affinity for cancer-associated antigens, which allows for accurate targeting with minimal impact on surrounding healthy tissues, enhancing therapeutic efficacy while reducing side effects, immune-modulating ability, and versatility and flexibility in engineering and conjugation. However, there are inherent limitations in using antibodies as a platform for radiopharmaceuticals due to their natural activities within the immune system, large size preventing effective tumor penetration, and relatively long half-life with concerns for prolonged radioactivity exposure. Antibody engineering can solve these challenges while preserving the many advantages of the immunoglobulin framework. In this review, the goal is to give a general overview of antibody engineering and design for tumor radiotheranostics. Particularly, the four ways that antibody engineering is applied to enhance radioimmunoconjugates: pharmacokinetics optimization, site-specific bioconjugation, modulation of Fc interactions, and bispecific construct creation are discussed. The radionuclide choices for designed antibody radionuclide conjugates and conjugation techniques and future directions for antibody radionuclide conjugate innovation and advancement are also discussed.

A Comparison of Antibody-Antigen Complex Sequence-to-Structure Prediction Methods and their Systematic Biases

The ability to accurately predict antibody-antigen complex structures from their sequences could greatly advance our understanding of the immune system and would aid in the development of novel antibody therapeutics. There have been considerable recent advancements in predicting protein-protein interactions (PPIs) fueled by progress in machine learning (ML). To understand the current state of the field, we compare six representative methods for predicting antibody-antigen complexes from sequence, including two deep learning approaches trained to predict PPIs in general (AlphaFold-Multimer, RoseTTAFold), two composite methods that initially predict antibody and antigen structures separately and dock them (using antibody-mode ClusPro), local refinement in Rosetta (SnugDock) of globally docked poses from ClusPro, and a pipeline combining homology modeling with rigid-body docking informed by ML-based epitope and paratope prediction (AbAdapt). We find that AlphaFold-Multimer outperformed other methods, although the absolute performance leaves considerable room for improvement. AlphaFold-Multimer models of lower-quality display significant structural biases at the level of tertiary motifs (TERMs) towards having fewer structural matches in non-antibody containing structures from the Protein Data Bank (PDB). Specifically, better models exhibit more common PDB-like TERMs at the antibody-antigen interface than worse ones. Importantly, the clear relationship between performance and the commonness of interfacial TERMs suggests that scarcity of interfacial geometry data in the structural database may currently limit application of machine learning to the prediction of antibody-antigen interactions.

Can current preclinical strategies for radiopharmaceutical development meet the needs of targeted alpha therapy?

Preclinical studies are essential for effectively evaluating TAT radiopharmaceuticals. Given the current suboptimal supply chain of these radionuclides, animal studies must be refined to produce the most translatable TAT agents with the greatest clinical potential. Vector design is pivotal, emphasizing harmonious physical and biological characteristics among the vector, target, and radionuclide. The scarcity of alpha-emitting radionuclides remains a significant consideration. Actinium-225 and lead-212 appear as the most readily available radionuclides at this stage. Available animal models for researchers encompass xenografts, allografts, and PDX (patient-derived xenograft) models. Emerging strategies for imaging alpha-emitters are also briefly explored. Ultimately, preclinical research must address two critical aspects: (1) offering valuable insights into balancing safety and efficacy, and (2) providing guidance on the optimal dosing of the TAT agent.

Good practices for 89Zr radiopharmaceutical production and quality control

BACKGROUND

During the previous two decades, PET imaging of biopharmaceuticals radiolabeled with zirconium-89 has become a consistent tool in preclinical and clinical drug development and patient selection, primarily due to its advantageous physical properties that allow straightforward radiolabeling of antibodies (89Zr-immuno-PET). The extended half-life of 78.4 h permits flexibility with respect to the logistics of tracer production, transportation, and imaging and allows imaging at later points in time. Additionally, its relatively low positron energy contributes to high-sensitivity, high-resolution PET imaging. Considering the growing interest in radiolabeling antibodies, antibody derivatives, and other compound classes with 89Zr in both clinical and pre-clinical settings, there is an urgent need to acquire valuable recommendations and guidelines towards standardization of labeling procedures.

MAIN BODY

This review provides an overview of the key aspects of 89Zr-radiochemistry and radiopharmaceuticals. Production of 89Zr, conjugation with the mostly used chelators and radiolabeling strategies, and quality control of the radiolabeled products are described in detail, together with discussions about alternative options and critical steps, as well as recommendations for troubleshooting. Moreover, some historical background on 89Zr-immuno-PET, coordination chemistry of 89Zr, and future perspectives are provided. This review aims to serve as a quick-start guide for scientists new to the field of 89Zr-immuno-PET and to suggest approaches for harmonization and standardization of current procedures.

CONCLUSION

The favorable PET imaging characteristics of 89Zr, its excellent availability due to relatively simple production and purification processes, and the development of suitable bifunctional chelators have led to the widespread use of 89Zr. The combination of antibodies and 89Zr, known as 89Zr-immuno-PET, has become a cornerstone in drug development and patient selection in recent years. Despite the advanced state of 89Zr-immuno-PET, new developments in chelator conjugation and radiolabeling procedures, application in novel compound classes, and improved PET scanner technology and quantification methods continue to reshape its landscape towards improving clinical outcomes.

Clinical Internal Dosimetry and Biodistribution of 177 Lu-DOTA-Trastuzumab in HER2-Positive Metastatic and Locally Advanced Breast Carcinoma

OBJECTIVE

The aim of this study was to assess the biodistribution and dosimetry of 177 Lu-DOTA-trastuzumab in patients with HER2-positive breast carcinoma using whole-body (WB) planar imaging at multiple time points.

PATIENTS AND METHODS

This study was a prospective evaluation of HER2-positive metastatic/locally advanced breast carcinoma patients who underwent gamma camera imaging for dosimetry and biodistribution studies by using 177 Lu-DOTA-trastuzumab. The standard diagnostic dosimetry protocol was followed, which included cold trastuzumab injection followed by in-house produced 177 Lu-DOTA-trastuzumab. Serial WB planar images (anterior and posterior) were obtained on gamma camera after the infusion of 177 Lu-DOTA-trastuzumab at multiple time points. Whole-body and organ regions of interest were drawn, and the numbers of disintegrations were obtained. The mean absorbed doses for the liver, spleen, kidneys, heart, red marrow, and tumor were obtained from OLINDA EXM v2.1.1 and ORIGIN software.

RESULTS

The study included a cohort of 21 female breast carcinoma patients. Tracer activity ( 177 Lu-DOTA-trastuzumab) was noted in the physiological organs such as the liver, spleen, kidneys, heart, as well as in the tumors. On visual analysis of 177 Lu-DOTA-trastuzumab biodistribution, the liver activity showed gradual clearance over time, and although spleen was comparatively faintly visualized than liver and similarly, kidneys were faintly visualized suggestive of the alternate route of tracer excretion. The maximum number of patients (n = 12) showed 2 components of clearance, namely, fast and slow. The average effective half-life of all the patients (including single and 2 components of clearance) was 106.25 ± 22.14 hours (84.11-128.39 hours). The mean absorbed dose for the liver, spleen, kidneys, heart, whole body, and red marrow was 1.0702 ± 0.731, 1.4114 ± 0.462, 1.4232 ± 0.364, 1.4719 ± 0.602, 0.2412 ± 0.0295, and 0.1485 ± 0.0213 mGy/MBq, respectively, by OLINDA EXM and 0.5741 ± 0.333, 0.8096 ± 0.224, 0.7943 ± 0.235, 1.8971 ± 0.713, and 0.09619 ± 0.0144 for liver, spleen, kidneys, heart and whole body respectively by ORIGIN. The absorbed radiation dose for tumor was 1.94E+2 by OLINDA EXM software and 1.78E+2 by ORIGIN software. In this study, during and after infusion of 177 Lu-DOTA-trastuzumab, no major adverse effects were noted in any patient except 1 patient who had grade 1 nausea and managed conservatively by antiemetic drug.

CONCLUSIONS

The results of our study demonstrated expected and favorable biodistribution and dosimetry with 177 Lu-DOTA-trastuzumab in HER2-positive breast carcinoma patients. We noticed the mean absorbed dose to the normal organs within the limits of maximum tolerable dose, and also tumor dose was higher than the normal liver dose. Therefore, we conclude that 177 Lu-DOTA-trastuzumab radioimmunotherapy is feasible and a safe treatment option for treating HER2-positive breast carcinoma patients.

Evaluating CD133 as a Radiotheranostic Target in Small-Cell Lung Cancer

Despite decades of work, small-cell lung cancer (SCLC) remains a frustratingly recalcitrant disease. Both diagnosis and treatment are challenges: low-dose computed tomography (the approved method used for lung cancer screening) is unable to reliably detect early SCLC, and the malignancy's 5 year survival rate stands at a paltry 7%. Clearly, the development of novel diagnostic and therapeutic tools for SCLC is an urgent, unmet need. CD133 is a transmembrane protein that is expressed at low levels in normal tissue but is overexpressed by a variety of tumors, including SCLC. We previously explored CD133 as a biomarker for a novel autoantibody-to-immunopositron emission tomography (PET) strategy for the diagnosis of SCLC, work that first suggested the promise of the antigen as a radiotheranostic target in the disease. Herein, we report the in vivo validation of a pair of CD133-targeted radioimmunoconjugates for the PET imaging and radioimmunotherapy of SCLC. To this end, [89Zr]Zr-DFO-αCD133 was first interrogated in a trio of advanced murine models of SCLC─i.e., orthotopic, metastatic, and patient-derived xenografts─with the PET probe consistently producing high activity concentrations (>%ID/g) in tumor lesions combined with low uptake in healthy tissues. Subsequently, a variant of αCD133 labeled with the β-emitting radiometal 177Lu─[177Lu]Lu-DTPA-A″-CHX-αCD133─was synthesized and evaluated in a longitudinal therapy study in a subcutaneous xenograft model of SCLC, ultimately revealing that treatment with a dose of 9.6 MBq of the radioimmunoconjugate produced a significant increase in median survival compared to a control cohort. Taken together, these data establish CD133 as a viable target for the nuclear imaging and radiopharmaceutical therapy of SCLC.

Advances and challenges in immunoPET methodology

Immuno-positron emission tomography (immunoPET) enables imaging of specific targets that play a role in targeted therapy and immunotherapy, such as antigens on cell membranes, targets in the disease microenvironment, or immune cells. The most common immunoPET applications use a monoclonal antibody labeled with a relatively long-lived positron emitter such as 89Zr (T 1/2 = 78.4 h), but smaller antibody-based constructs labeled with various other positron emitting radionuclides are also being investigated. This molecular imaging technique can thus guide the development of new drugs and may have a pivotal role in selecting patients for a particular therapy. In early phase immunoPET trials, multiple imaging time points are used to examine the time-dependent biodistribution and to determine the optimal imaging time point, which may be several days after tracer injection due to the slow kinetics of larger molecules. Once this has been established, usually only one static scan is performed and semi-quantitative values are reported. However, total PET uptake of a tracer is the sum of specific and nonspecific uptake. In addition, uptake may be affected by other factors such as perfusion, pre-/co-administration of the unlabeled molecule, and the treatment schedule. This article reviews imaging methodologies used in immunoPET studies and is divided into two parts. The first part summarizes the vast majority of clinical immunoPET studies applying semi-quantitative methodologies. The second part focuses on a handful of studies applying pharmacokinetic models and includes preclinical and simulation studies. Finally, the potential and challenges of immunoPET quantification methodologies are discussed within the context of the recent technological advancements provided by long axial field of view PET/CT scanners.

Claudin18.2-targeted cancer theranostics

Claudin 18.2 (CLDN18.2) is an emerging target for the treatment of CLDN18.2-expressing cancers such as gastric and pancreatic cancers. Cell and antibody therapies targeting CLDN18.2 are under intensive clinical trials. In this setting, how to efficiently and specifically detect CLDN18.2 expression before and after the therapies is a clinical challenge. In recent years, molecular imaging with radiolabeled antibodies or antibody fragments have shown promise in noninvasively annotating antigen expression across the body. In this Perspective, we will bring together the most recent progress on CLDN18.2-targeted imaging and therapy of solid tumors.

pH-sensitive gold nanoclusters labeling with radiometallic nuclides for diagnosis and treatment of tumor

The acidic microenvironment is one of the remarkable features of tumor and is also a reliable target for tumor theranostics. Ultrasmall gold nanoclusters (AuNCs) have good in vivo behaviors, such as non-retention in liver and spleen, renal clearance, and high tumor permeability, and held great potential for developing novel radiopharmaceuticals. Herein, we developed pH-sensitive ultrasmall gold nanoclusters by introducing quaternary ammonium group (TMA) or tertiary amine motifs (C6A) onto glutathione-coated AuNCs (TMA/GSH@AuNCs, C6A-GSH@AuNCs). Density functional theory simulation revealed that radiometal ⁸⁹Sr, ²²³Ra, ⁴⁴Sc, ⁹⁰Y, ¹⁷⁷Lu, ⁸⁹Zr, ^99mTc, ¹⁸⁸Re, ¹⁰⁶Rh, ⁶⁴Cu, ⁶⁸Ga, and ¹¹³Sn could stably dope into AuNCs. Both TMA/GSH@AuNCs and C6A-GSH@AuNCs could assemble into large clusters responding to mild acid condition, with C6A-GSH@AuNCs being more effective. To assess their performance for tumor detection and therapy, TMA/GSH@AuNCs and C6A-GSH@AuNCs were labeled with ⁶⁸Ga, ⁶⁴Cu, ⁸⁹Zr and ⁸⁹Sr, respectively. PET imaging of 4T1 tumor-bearing mice revealed TMA/GSH@AuNCs and C6A-GSH@AuNCs were mainly cleared through kidney, and C6A-GSH@AuNCs accumulated in tumors more efficiently. As a result, ⁸⁹Sr-labeled C6A-GSH@AuNCs eradicated both the primary tumors and their lung metastases. Therefore, our study suggested that GSH-coated AuNCs held great promise for developing novel radiopharmaceuticals that specifically target the tumor acidic microenvironment for tumor diagnosis and treatments.