Background
The innate and adaptive immune systems interact in a well-orchestrated manner upon recognition of an antigen. After antigen exposure, professional antigen-presenting cells (APC), such as dendritic cells (DC), lose phagocytic action simultaneously gaining antigen-presenting capability by the upregulation of major histocompatibility complex (MHC) receptors and the co-stimulatory molecules CD80 (B7.1) and CD86 (B7.2) [
1]. The interactions of CD80 and CD86 with their T cell receptor cytotoxic T lymphocyte protein 4 (CTLA-4, CD152) and CD28 belong to the best-characterized co-stimulatory pathways in immune response, determining whether antigen exposure results in immune tolerance or T cell activation [
2]. The interactions between the co-stimulatory molecules and their T cell receptors and the resulting immune response are determining factors for the course of various diseases including atherosclerosis, cancer, transplant rejection, and autoimmune diseases [
1,
3,
4].
In atherosclerosis research, Buono et al. demonstrated that LDL receptor (Ldlr) knockout (KO) mice lacking CD80 and CD86 displayed a delayed atherosclerosis progression compared to control Ldlr KO mice [
5]. The involvement of CD80 and CD86 in vascular remodeling was presented by Ewing et al. in a femoral artery cuff mouse model [
6]. We recently found significantly higher relative messenger RNA (mRNA) levels of CD80 and CD86 in human vulnerable than stable carotid plaques, supported by immunohistochemistry data [
7]. Our results were in agreement with a study on human carotid and coronary plaques by Erbel et al. [
8]. They observed a higher CD86 mRNA expression in human carotid plaques of symptomatic than asymptomatic patients. Their immunohistochemical analysis revealed that virtually all mature DCs were CD86-positive and in close proximity to activated T cells. Further evidence for an involvement of mature APCs in atherosclerosis was presented in an expression analysis of human carotid endarterectomized plaques [
9]. CD11c, CD80, CD83, and CD86 mRNA expression was significantly higher in vulnerable than stable plaques. The increase in vulnerable plaques was more pronounced for the mature subpopulation of DCs than the total number including immature DCs [
9]. Altogether, these studies indicate that advanced atheroma contain a population of fully maturated DCs and macrophages expressing co-stimulatory molecules and that these APCs could be efficient regulators of T cell activity in atherosclerosis. Antigen presentation is not restricted to lymph nodes but additionally occurs in atherosclerotic plaques as supported by co-localization data of DCs and T cells [
7,
8,
10].
Since activated APCs that express CD80 and CD86 are a primary component of atherosclerotic lesions and the amount of APCs correlates with plaque vulnerability [
9‐
11], CD80 and CD86 may be promising imaging targets in atherosclerosis. We recently evaluated CD80 as a target for non-invasive imaging by positron emission tomography (PET) [
7]. An oxodihydropyrazolocinnoline derivative labeled with carbon-11 showed high binding affinity to CD80 and bound to CD80-positive Raji xenografts and in a displaceable manner to endarterectomized human carotid plaques both in vitro. However, the low tissue distribution of the carbon-11-labeled ligand was a limiting factor in pilot PET experiments with mice [
7].
Abatacept (CTLA4-Ig, Orencia
®, Bristol Myers Squibb) and belatacept (LEA29Y, Nulojix
®, Bristol Myers Squibb) are 90-kDa fusion proteins of the extracellular domain of human CTLA-4 and a modified Fc fragment of a human IgG1, used in the therapy against rheumatoid arthritis and transplant rejection, respectively, inhibiting CD80/CD86-mediated T cell activation. Belatacept differs by two amino acids from abatacept and the CTLA-4 extracellular domain, resulting in a 10-fold increased inhibition of T cell co-stimulation in vitro [
12]. Surface plasmon resonance measurements revealed high binding affinities of belatacept to human and mouse CD80 as well as CD86 [
12,
13].
In the present proof-of-concept study, we evaluated the potential of CD80/CD86-specific belatacept for atherosclerosis imaging. We labeled belatacept with the gamma-emitting nuclide indium-111 (physical half-life 2.8 days) and characterized its binding affinity to human CD80/CD86 and its stability in plasma in vitro. The 111In-labeled probe was further evaluated by in vivo and ex vivo single-photon emission computed tomography (SPECT) with mice regarding its accumulation in CD80/CD86-positive murine tissues and human CD80/CD86-positive xenografts. Moreover, ex vivo SPECT scans were acquired in a mouse model of atherosclerosis. Finally, the in vitro binding of 111In-DOTA-belatacept to human carotid plaques was compared with histological markers of plaque inflammation.
Methods
Ethics, consent, and permissions
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committees and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual patients included in the study prior to surgery.
All applicable international, national, and institutional guidelines for the care and use of animals were followed, in particular the Swiss legislation and FELASA guidelines. All experiments with animals were approved by the veterinary office of the Kanton Zurich.
Conjugation, radiolabeling, quality control, and stability
Commercially available belatacept (Nulojix
®, Bristol Myers Squibb) was conjugated to
p-SCN-Bn-DOTA (S-2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid, B205, Macrocyclics, Dallas, TX) and labeled with indium-111 according to a published procedure [
14]. The conjugation reaction is described in Additional file
1. The conjugate was characterized by liquid chromatography-mass spectrometry (LC-MS).
For radiolabeling, each 10 μg protein conjugate was reacted with 4 MBq
111In (Mallinckrodt, Dublin, Ireland) as optimized by incubation of a fixed amount of belatacept-DOTA with increasing amounts of
111In. The reaction mixture for the radiolabeling contained 300–350 μg belatacept-DOTA (175 μL), 0.1 M ammonium acetate buffer (pH 6, 60 μL), and
111In (
111InCl
3 in 0.02 N HCl, 213–228 MBq, 215 μL, Mallinckrodt). After 1 h incubation at room temperature (rt), the reaction was quenched by the addition of 50 mM EDTA solution (10 %,
v/
v), mixed and allowed to incubate for 5 min at rt. The product was purified, and its stability in phosphate-buffered saline (PBS) and plasma was investigated as described in Additional file
1.
In vitro experiments
Cell culture, immunocytochemistry, binding studies with Raji and NCI-H69 cells and tissue staining are described in detail in Additional file
1. In brief, Raji (human Burkitt’s lymphoma cell line, ATCC CCL-86) and NCI-H69 cells (human lung small cell carcinoma cell line, ATCC HTB-119) were purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany) and were cultured according to the supplier’s protocols.
For immunocytochemistry, the following primary antibodies were used: anti-CD80 1:100 (2A2, ab86473, Abcam, Cambridge, UK), anti-CD86 1:200 (EP1158Y, ab53004, Abcam, Cambridge, UK), and anti-CTLA-4 1:100 (MO6, clone 2F1, H00001493-M06, Abnova, Taipei, Taiwan). Additionally, the following secondary antibodies were used: goat anti-rabbit IgG H&L 1:1000 (FITC, ab6717, Abcam, Cambridge, UK) and goat anti-mouse IgG H&L 1:1000 (FITC, ab6785, Abcam, Cambridge, UK). Cells were fixed in paraformaldehyde and permeabilized with methanol. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Binding affinity was determined in saturation binding and Lindmo assays with Raji and NCI-H69 cells as described in Additional file
1.
In vivo and ex vivo studies with xenograft-bearing mice
Five-week-old female CD1 nu/nu mice (Crl: CD1-
Foxn
1nu
) from Charles River Laboratories (Sulzfeld, Germany) were fed a normal chow diet ad libitum. At the age of 6 weeks, these immune-deficient mice were subcutaneously inoculated in their shoulder region with 1 × 10
7 Raji cells in 100 μL Matrigel (BD Biosciences, Oxford, UK) (4 animals both sides, 6 animals right side). NCI-H69 cells (1 × 10
7 in 100 μL Matrigel) were subcutaneously inoculated in the shoulder region of CD1 nu/nu mice 2 weeks afterwards (2 animals both sides, 6 animals with Raji xenografts, see above, opposite side). In vivo SPECT/CT scans, ex vivo autoradiography, and biodistribution experiments were performed 2 weeks after the second inoculation according to Additional file
1: Table S1. Animals were injected intravenously with 10 MBq
111In-DOTA-belatacept (25 μg, baseline) or 10 MBq
111In-DOTA-belatacept and 500 μg unlabeled belatacept (blockade). In vivo and ex vivo experiments were performed 48 h after tracer injection or as indicated.
ApoE KO atherosclerosis mouse models and C57BL/6 control mice
Eleven-week-old male C57BL/6 mice (control) and four-week-old male ApoE KO mice (B6.129P2-Apoe
tm1Unc/J) were obtained from Charles River Laboratories. Two C57BL/6 mice and two ApoE KO mice were fed a normal chow diet ad libitum. All other ApoE KO mice (
n = 4) received a modified Western-type diet containing 21 % fat, 0.25 % cholesterol, and 19.5 % casein (Kliba Nafag, Kaiseraugst, Switzerland) ad libitum. In two ApoE KO animals, fed a modified Western-type diet for 4 weeks, flow-modifying implants (cuff) were placed around the right and left common carotid arteries according to previous publications, inducing atherosclerotic plaque formation up- and downstream [
15,
16]. A detailed characterization of the mouse model under the above conditions, including a histological assessment, gene expression analysis, and oil red o staining, is shown elsewhere (Meletta et al., manuscript submitted). Mice were used for ex vivo SPECT/CT experiments 19 weeks after surgery.
111In-DOTA-belatacept (10 MBq, 25 μg, baseline) or 10 MBq
111In-DOTA-belatacept and 500 μg unlabeled belatacept (blockade) was injected into the tail vein of the animals. Mice were euthanized 48 h after tracer injection for ex vivo SPECT/CT imaging and oil red o staining of the excised aorta and carotids. The two mice with implants were in addition scanned before euthanasia.
SPECT/CT imaging
In vivo and ex vivo SPECT/CT scans were conducted with a four-head multiplexing multipinhole camera (NanoSPECT/CT, Bioscan, Washington, DC, USA). The acquisition time per view depended on the amount of radioactivity at scan start in the field of view, and the resulting scan times ranged from 20 min to 1 h (in vivo scans) and 14.5 to 15.5 h (ex vivo scans). In vivo scans were performed 48 h after tracer injection or as indicated. For ex vivo studies, animals were euthanized ∼48 h after tracer injection and SPECT/CT images were acquired after sample preparation. For in vivo SPECT/CT scans, animals were anesthetized with isoflurane (1.5–4 %) in oxygen. SPECT data were reconstructed with HiSPECT software (Scivis GmbH, Göttingen, Germany), and images were generated with the Nucline Software (Bioscan). The fused SPECT and CT data were analyzed using VivoQuant image post-processing software (inviCRO Imaging Services and Softwares, Boston, USA) and PMOD biomedical image quantification software (PMOD Technologies, Zurich, Switzerland). For the ex vivo scans, volumes of interest (VOIs, ~1 cm3) were drawn manually to include aorta and carotids. A VOI of equal size was drawn outside the tissue samples to define background radioactivity in the absence of tissue. Tissue-to-background and tissue baseline-to-blocked ratios were calculated from the averaged 5 voxels with the highest radioactivity per VOI.
Biodistribution experiments
Post-mortem biodistribution studies with inoculated CD1 nu/nu mice (Raji xenograft right and NCI-H69 xenograft left shoulder) were performed with three baseline animals and three blockade animals. Animals were euthanized with CO2 followed by cervical dislocation. Tissues, organs, and a defined volume of original injectate were weighed and counted for radioactivity in a gamma counter (Packard Cobra II Auto Gamma, PerkinElmer). Accumulated tissue radioactivity was expressed as % injected dose (Bq) per g tissue (% ID/g).
Ex vivo autoradiography
For ex vivo autoradiographic studies, CD1 nu/nu mice bearing either Raji or NCI-H69 xenografts on both shoulders were injected with ~10 MBq tracer and euthanized after 48 h as described above. Tumors were dissected, embedded in Tissue-Tek O.C.T. medium, and frozen at −80 °C, and sections of 5 μm were prepared using a cryostat. After drying, frozen sections were exposed to super-resolution imager plates (PerkinElmer, Waltham, MA, USA) for 5 min and analyzed by a Cyclone® Plus Storage Phosphor System (PerkinElmer). Radioactivity signals were quantified with the software OptiQuant (5.0, PerkinElmer).
Human atherosclerotic carotid plaques
Atherosclerotic carotid plaques were obtained from patients undergoing carotid endarterectomy (CEA) at the University Hospital of Zurich, Switzerland. Indications for CEA were stroke, transient ischemic attack, on-pump surgery to prevent perioperative stroke, ultrasound diagnosis revealing a soft and unstable plaque, or other surgical interventions (e.g. aortic valve, coronary artery surgery). Twenty-three patients were included in this study with an average age of 75.4 ± 6.8 years, 16 were of male gender, 4 females, and 3 were unknown. Plaque material was collected from all patients from the common, external, and internal carotid artery. Available details of the patients and the samples used in this study are shown in Additional file
1: Table S2. Plaque specimens were stored in RNAlater
® solution (Sigma-Aldrich, St. Louis, USA) at −80 °C. After thawing, plaques were embedded in Tissue-Tek O.C.T. medium and cryosections (20 μm) were prepared with a cryostat. Frozen sections were stored at −20 °C until further use.
In vitro autoradiography
Radiotracer binding was evaluated by in vitro autoradiography with sections of human carotid plaques (20 μm). Cryosections were thawed at rt for 30 min and preincubated for 10 min on ice in HEPES buffer (50 mM HEPES, 5 mM MgCl
2, 125 mM NaCl, 1 mM CaCl
2, pH 7.4) supplemented with 0.5 % milk powder and 40 μM gammanorm (Octapharma AG, Lachen, Switzerland). Gammanorm contains human immunoglobulins that block unspecific binding of belatacept to ubiquitous Fc receptors. Tissues were incubated with 17 nM
111In-DOTA-belatacept (specific activity 36.6 GBq/μM) diluted in the above-specified buffer (~100 μL) for 1 h on ice. For blockade conditions, additionally 10 μM unlabeled belatacept was added to the tracer solution. The sections were washed with HEPES buffer containing milk powder and gammanorm (5 min, 4 °C), 3× HEPES buffer (5 min, 4 °C), and finally 2× distilled water (10 s, 4 °C) and were dried at rt. Standards of 1 μL tracer solution dried on 0.87 cm
2 filter papers were prepared as positive controls. The sections and the standards were exposed to a super-resolution imager plate (PerkinElmer) for 30 min and 14 h. The plate was scanned by a Cyclone
® Plus Storage Phosphor System (PerkinElmer), and data was analyzed with the software OptiQuant (5.0, PerkinElmer). Specific binding was calculated from the difference between the integrated radioactivity signals per area (DLU/mm
2) under baseline and blocking conditions, divided by the value under baseline conditions. Hematoxylin and eosin (HE) staining was performed with all sections to investigate tissue morphology and score each individual plaque specimen used in autoradiography according to Additional file
1: Table S3. A direct comparison with immunohistochemistry for CD80 and/or CD86 as performed in our previous study [
7] was not performed as both
111In-DOTA-belatacept autoradiography with paraffin-embedded tissues and immunohistochemistry with cryosections were not successful. In addition, the poor resolution of the
111In autoradiography did not allow the microscopic localization of tracer accumulation in the sections. In total, 37 carotid specimens were used for autoradiography experiments. The distribution of the plaques per classification category and the individual details on patients and tissues are shown in Additional file
1: Tables S2 and S4.
Statistics
Statistical data analysis was performed with GraphPad Prism (GraphPad, La Jolla, CA, USA) or Microsoft Excel. To assess the intergroup difference of two groups, a two-tailed unpaired Student’s t test was performed. For more than two groups, data was analyzed with a one-way ANOVA with a Tukey’s multicomparison test. A p value <0.05 was considered significant.
Discussion
In this proof-of-concept study, we successfully established a radiolabeling procedure for the CD80/CD86-targeting fusion protein belatacept with indium-111 using
p-SCN-Bn-DOTA as bifunctional chelating agent. After purification, the product was obtained in high radiochemical yield and purity in a fast and robust labeling reaction. Under the applied conditions, approximately two DOTA chelators were coupled to belatacept which is the intended range to minimize undesired chelator-mediated effects on pharmacokinetics [
19]. A high stability of the radiolabeled product was determined in PBS and plasma of murine and human origin, respectively, up to 72 h allowing in vivo investigations.
The radiolabeled product bound to human Raji cells with a nanomolar
K
d value. Efforts to assess the binding affinity to mouse recombinant CD80 by differential scanning fluorimetry were not successful due to similar melting temperatures
T
m of the recombinant CD80 and belatacept. Immune-reactive affinity assays were excluded due to interferences with the Fc portion of belatacept. However, the results of this study indicate that the radiolabeling modifications did not hamper high-affinity binding to human and murine CD80/CD86 in vitro and in vivo. The contribution of Fc receptor-mediated binding to the total binding of belatacept was negligible as concluded from in vitro autoradiography with human tissue and as expected from published data [
13].
In vivo SPECT/CT scans and ex vivo autoradiography experiments indicated that
111In-DOTA-belatacept specifically accumulated in CD80/CD86-positive tissues. Besides Raji xenografts, strong and specific accumulation was observed in the lymph nodes. The lymph nodes exhibit a resident population of APCs capable of expressing CD80 and CD86 in particular during the process of antigen presentation. Moreover, biodistribution studies revealed the highest radiotracer accumulation in the spleen, a peripheral lymphoid organ, and a high specific accumulation in the salivary glands of xenograft-bearing mice. This is consistent with the fact that APCs are constituents of human and murine salivary glands being involved in immune surveillance [
20,
21]. The high non-specific hepatic radiotracer accumulation may be related to the IgG1 Fc part of the fusion protein [
22].
The accumulation of radioactivity in the Raji xenografts increased from 2 to 18 to 48 h p.i. This is in agreement with the relatively long half-life in the blood of IgG1 Fc fusion proteins. The half-life of the closely related abatacept in mice was ~90 h after i.v. injection [
23]. Indium-111 is thus a good choice for the labeling. Its long physical half-life of 67 h allows scans up to several days after tracer injection when a significant portion of
111In-DOTA-belatacept is cleared from the blood.
The high and specific accumulation of
111In-DOTA-belatacept in the lymph nodes of the xenograft-bearing animals seen in SPECT and confirmed in the biodistribution experiments looks promising towards the imaging of small CD80/CD86-rich structures. We, therefore, proceeded to an animal model of atherosclerosis where lesions are of similar or smaller dimension. However, owing to the proximity of atherosclerosis-prone areas in the aorta and the carotids to other tissues with radiotracer accumulation, e.g. the liver, the salivary glands, and the thymus, SPECT/CT scans of aorta and carotids were performed ex vivo. Furthermore, perfusion of the excised tissue allowed to exclude any blood-related tracer radioactivity, considering the long residence time of
111In-DOTA-belatacept in the blood (still 7.6 % ID/g blood 48 h after injection).
111In-DOTA-belatacept accumulated in lipid-rich atherosclerotic plaques, and tracer accumulation was higher under baseline than blockade condition. Moreover,
111In-DOTA-belatacept accumulation in the aortas of ApoE KO animals was dependent on an atherosclerosis-promoting diet. However, several lipid-rich regions were not positive in SPECT. We can only speculate on the reasons. SPECT imaging with
111In-DOTA-belatacept may not have been sensitive enough to visualize all lesions, depending on their size and tracer uptake. The signal was close to the detection limit in the ex vivo scans. Alternatively and in addition, not all lipid-rich regions may be invaded by high numbers of CD80/CD86-positive cells. As discussed under limitations, we did not further investigate these vessel walls by immunohistochemistry. However, we found that the descending aortas of ApoE KO mice fed a high-fat diet contained significantly higher mRNA levels of CD80 than the descending aortas of wild type mice under normal diet (Meletta et al., manuscript submitted). In agreement with our findings, the correlation between characteristics of plaque vulnerability and the lipid content of the lesions may be weak as derived from intravascular ultrasound measurements [
24].
In human carotid tissue samples,
111In-DOTA-belatacept binding was associated with the infiltration of immune cells. It can be assumed that the majority of inflammatory cells within a plaque are APCs [
25]. The increasing number of macrophages and mature DCs in advanced atherosclerotic lesions might promote the formation of a necrotic core due to a loss of efferocytosis activity, an intensification of the inflammatory response, and degradation of matrix components [
1,
26,
27]. This is in agreement with the observed increase in
111In-DOTA-belatacept uptake in lesions with a large lipid/necrotic core. Our results suggest that
111In-DOTA-belatacept accumulates specifically in atherosclerotic plaques, depending on their degree of inflammation and vulnerability.
Limiting factors are related to the small size of atherosclerotic plaques both in preclinical models and in humans. High specific accumulation of radioactivity and low spill-over from other tissues are required for the detection of lesions in the millimeter or submillimeter range [
28]. The long biological half-life of
111In-DOTA-belatacept is unfavorable in this respect, as it contributes to significant unspecific tissue radioactivity as observed in the in vivo scans. Imaging vulnerable plaques in coronary arteries would be even more challenging. According to the biodistribution results, radioactivity in myocardium was still relatively high 48 h p.i., although unspecific and lower than in the blood. This relatively high background level in the heart together with the respiratory and myocardium motion would hamper detection of atherosclerotic lesions in coronary arteries under these experimental conditions. A tracer with a faster clearance from the blood and CD80/CD86-negative tissues in combination with motion correction from simultaneously acquired magnetic resonance imaging (MRI) may improve the contrast between atherosclerotic lesions and myocardium in the future.
Further limitations of our study are the low number of animals used for in vivo and ex vivo imaging and the lack of control animals to study lymph node accumulation of
111In-DOTA-belatacept in the absence of xenografts or atherosclerosis. In addition, belatacept binds with similar affinity to CD80 and CD86, which may reduce selectivity for a particular type of cell activation as compared to a CD80- or CD86-selective tracer. Due to technical limitations with available antibodies, we were not able to counterstain the atherosclerotic tissues from SPECT and autoradiography for CD80 and CD86, the proteins that we found upregulated in human vulnerable atherosclerotic plaques [
7]. Finally, a fusion protein would not fulfill all requirements for clinical application as a PET imaging agent primary due to its prolonged biological half-life of 8–9 days in humans in the case of belatacept and concerns with respect to immunogenicity of proteins [
29].
Future studies have to focus on CD80- or CD86-selective small molecules and truncated versions of belatacept as well as abatacept, which has a higher affinity and selectivity for CD80 than CD86 in mouse [
13]. In this respect,
111In-DOTA-belatacept could serve as a benchmarking tracer in the preclinical development of novel CD80/CD86 imaging agents.
Acknowledgements
The authors are grateful to Susan Cohrs for the assistance with the animal experiments, acquisition of SPECT scans, and support with in vitro experiments. We thank Claudia Keller for the xenograft inoculation, performing the surgery in ApoE KO mice, and taking care of the animals and Dr. Michael Kuhlmann (European Institute for Molecular Imaging, EIMI, Münster, Germany) for the introduction to the cuff surgery. The authors thank Dr. Zoran Rancic (University Hospital Zurich) and the team of Prof. Philipp A. Kaufmann (University Hospital Zurich) for the human atherosclerotic plaque collection. We thank Sabina Wunderlin (University of Zurich) for the HE staining and Prof. Cornelia Halin (ETH Zurich) for the access to the infrastructure for immuncytochemical evaluation. The authors acknowledge the support of the Scientific Center for Optical and Electron Microscopy (ScopeM) of the ETH Zurich.
This work (RM) was financially supported by the Clinical Research Priority Program (CRPP) of the University of Zurich on Molecular Imaging (MINZ).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
RM carried out the experimental work, analyzed and interpreted the data, and wrote the manuscript. AMH supervised and was involved in the experimental work, data analysis, interpretation, and manuscript writing. PD and EF supervised and were involved in the conjugation and labeling of belatacept and the in vitro cell binding experiments. SDK supervised the project and contributed to the data interpretation and manuscript writing. All authors contributed significantly to the conception, design, and planning of the study and were involved in revising the manuscript critically for important intellectual content. All authors have given final approval of the version to be published.