Inflammation is a biological response of the immune system that is involved either directly or indirectly in many human diseases. It is a defense mechanism that is crucial for survival, but when acute inflammation becomes uncontrollable, it may also damage the adjacent healthy cells. The improper resolution of inflammatory processes leads to a variety of chronic inflammatory diseases15. Due to the wide scale of inflammatory behavior in different types of diseases, more sensitive and specific biomarkers are required to improve prevention and treatment. For example, the understanding of the autoimmune response in several diseases is mainly based on animal models, human tissue biopsy or post-mortem analysis. Repeatable, quantitative and non-invasive methods for direct monitoring of the immune response in human would therefore potentially provide an important tool in diagnosis, management and treatment monitoring. In this study we report the discovery, optimization and radiolabeling of novel CD69 targeting Affibody molecules for quantitative assessment of inflammatory processes.
CD69 is expressed on most activated immune cells, especially in the early activation process. A general marker for activated immune cells is potentially more sensitive for detecting early subclinical immune responses than a cell-specific marker. Furthermore, the background binding of CD69 to resting circulating immune cells in blood is negligible in contrast to other putative immune cell imaging targets such as CD4 or CD8. Previous studies have reported that CD69 expression is often detected on cells in samples from inflamed tissues from patients with several different diseases, including T1D, rheumatoid arthritis, psoriasis, asthma, eosinophilic pneumonia, chronic obstructive pulmonary disease (COPD), chronic bronchitis, eosinophilic chronic rhinosinusitis (ECRS), arthritis, sarcoidosis, atopic dermatitis, atherosclerosis, systemic sclerosis, multiple sclerosis, systemic lupus erythematous, granulomatosis with polyangiitis (Wegener’s granulomatosis), neuromyelitis optica and autoimmune thyroiditis, as well in transplant biopsies of rejecting grafts and T and NK cells infiltrating tumors4,6,7. These clinical findings indicate that CD69 expression is a valid activation marker for all leukocytes in tissues with ongoing inflammation. Also, this indicates the potentially broad application of CD69 targeting imaging agents in many different diseases.
Here, Affibody molecules were developed and maturated with respect to affinity towards human CD69. All variants demonstrated acceptable stability and refolding properties as assessed by CD before and after heat denaturation. The refolding is especially important in the context of radiolabeling with e.g. radiometals, as the peptide often must be heated beyond the melting temperature for a short duration during radiolabeling. Misfolding after denaturation could generate a fraction of radiolabeled peptide with decreased binding capacity.
Furthermore, the Affibody molecules were evaluated with respect to binding affinity to human CD69, as well as murine CD69. ZCD69:4 in particular demonstrated improved affinity towards both human and murine CD69, compared to the primary variant ZCD69:2. ZCD69:10 and ZCD69:12 exhibited improved affinity to human CD69. However, ZCD69:10 and ZCD69:12 had poor affinity towards murine CD69, which would make preclinical evaluation of these constructs challenging. Thus, from these in vitro data, ZCD69:4 seemed as the most promising variant.
Affinity maturation of Affibody molecules can sometimes improve the affinity on the order of magnitudes. Here, affinity towards CD69 was only improved around twofold for a few of the variants. This outcome is less than optimal, but any increase in affinity will likely lead to a corresponding increase in sensitivity for an imaging probe. The reason for the limited improvement of affinity is unclear, but not without precedent for Affibody molecule maturation. A potential reason is the binding interface between the Affibody molecule (where two alpha helices with around 10–15 variable residues is responsible for binding) and the target protein. For the identified ZCD69:# binders, only 8 residues vary, mainly in the first alpha helix (Supplemental Fig. 1). Perhaps there is not sufficient geometrical variability of the Affibody scaffold, to generate sub-nanomolar affinity binders specifically for the extracellular domain of CD69. In silico modelling is required to further understand the exact mode of action of binding of this new class of CD69 binders, but is outside of the scope of the current manuscript.
The new lead compound, ZCD69:4, exhibited an affinity towards both human and murine CD69 of approximately 30 nM, which is approaching the affinity of monoclonal affibodies, while having substantially small size and consequently faster in vivo targeting and clearance, critical for putative imaging probes.
To study the biodistribution, the variants were functionalized with a chelator, and radiolabeled with Indium-111. DOTA was, selected as it would potentially enable radiolabeling with both Indium-111 (SPECT) and Gallium-68 (Positron Emission Tomography, PET).
ZCD69:10 could not be reproducibly functionalized with DOTA for further radiolabeling, and thus discarded as a viable candidate. The radiochemical purity for all radiolabeled variants was high, in excess of 95%, and thus any signal contribution from other radiolabeled impurities can be considered negligible. As reference, a radiochemical purity in excess of 90%, with no single impurity > 5%, is often considered acceptable for clinical deliveries of PET radiopharmaceuticals.
The biodistribution of the radiolabeled Affibody molecules were evaluated by longitudinal SPECT-CT imaging, immediately from administration and for up to 72 h. The reason to use a longer half-life isotope 111In (2.8 days) in this study was to obtain longitudinal imaging biodistribution data by following the tracer kinetics for extended time period, as well as providing simpler logistics for the preclinical evaluation compared to e.g. Gallium-68.
The biodistribution and clearance of a radiolabeled construct is very important, as it determines when optimal imaging contrast can be obtained. The optimal scanning window in turn determines which radionuclides that can be used – as the radionuclide half-life must be in accordance with the optimal imaging time window. Here, we demonstrate rapid targeting and clearance of all radiolabeled Affibody molecules, i.e. low background is seen already at the 3 h time point. Importantly, the optimal imaging window is likely within 0–3 h after administration, which would enable future radiolabeling with some standard PET radionuclides including Gallium-68 (half-life 68 min) or Fluorine-18 (half-life 109 min). PET imaging has improved temporal and spatial resolution in comparison to SPECT, in addition to being quantitative, which is crucial when one attempts to measure activated immune cells over time.
As comparison, most radiolabeled larger peptides, such as antibodies, have slow clearance and an optimal imaging scanning time after a few days. The fast clearance observed with the Affibody molecules in this study highlights the importance of development of smaller peptide binders towards CD69, to enable quantitative PET imaging of this target. Furthermore, imaging with short lived (≈hours) PET radionuclides additionally yields a substantially lower radiation dose to the scanned patient, compared to radionuclides with a half-life of ≈days required for labeled antibodies. Radiation dose is especially critical when considering longitudinal imaging in young and relatively healthy populations, e.g. healthy control groups, individuals with T1D etc.
Radiolabeled peptides often exhibit renal clearance. Thus, renal uptake and retention should be minimized to reduce the predicted extrapolated absorbed radiation dose to the kidney. Therefore, we used low renal retention as one of the criteria for selecting the optimal Affibody molecule variant. Furthermore, the background binding of the radioligand should be minimized for optimal image contrast in lesions or tissues with activated CD69 expressing immune cells.
All tested Affibody molecule variants demonstrated acceptable kidney retention dose and background binding. However, also based on biodistribution, variant ZCD69:4 proved to be optimal candidate. ZCD69:8 and ZCD69:12 and exhibited similar or lower kidney dose than ZCD69:4, but on the other hand, their background binding in muscle and liver were higher. The differences between the groups in respect to kidney, liver and muscle binding were not statistically different based on a one-way ANOVA multiple-comparisons test, due to a few outliers (especially in the 111In-DOTA-ZCD69:2 group) as well as the relatively small group size (n = 3). However, the trend was robust in demonstrating improved biodistribution for e.g. 111In-DOTA-ZCD69:4 and 111In-DOTA-ZCD69:8 compared to the other variants. Further increase in group size was deemed as unnecessary for selecting the optimal variant for e.g. animal ethical considerations.
Although variant ZCD69:4 and ZCD69:8 demonstrated similarly suitable biodistribution, ZCD69:8 lacked affinity to murine CD69, which would potentially make preclinical evaluation challenging. Thus, of the variants with affinity to both murine and human CD69, ZCD69:4 exhibited the most beneficial renal and tissue background signal.
Some of the rats examined by SPECT/CT exhibited clearly detectable uptake of the CD69 binding variants in different lymph nodes across the body, especially 111In-DOTA-ZCD69:2 and 111In-DOTA-ZCD69:6. MRI scans was performed to further verify that the structures corresponded to lymph nodes (grey contrast tissue embedded in white adipose tissue, in MRI images in Fig. 5A,B). Importantly, the binding in individual lymph nodes was often sustained over several scans over up to 72 h (Supplementary Fig. 13 and 15), ruling out the possibility of an image artefact.
To explore if the lymph node binding was non-specific, we performed the same set of experiments but using a negative control affibody 111In-DOTA-ZTAQ, with an amino acid sequence not binding to CD69. No detectable uptake was observed in any lymph nodes in either of the animals when using 111In-DOTA-ZTAQ, indication that affibodies don’t accumulate non-specifically in lymph nodes. Furthermore, several different affibody molecules have previously been radiolabeled and thoroughly examined in vivo in animals and humans, and focal uptake in lymph nodes are not generally observed. Thus, the lymph node binding of the CD69 binding variants is thus potentially an active and specific process. Additionally, lymph node binding was only seen for the variants with affinity for murine CD69 – lymph node targeting was not seen in rats administered with 111In-DOTA-ZCD69:8 or 111In-DOTA-ZCD69:12 (variants lacking affinity for murine CD69, Table 1). Potentially, the restricted binding in a lymph node in a certain part of the body is indicative of an immune response against local, subclinical infection. The apparent targeting of lymph nodes will be further explored in reproducible animal models of inflammation using the optimal variant.
In summary, variant ZCD69:4 displayed the optimal properties, both regarding stability, affinity and biodistribution, and was selected as lead compound for further development of a novel class of CD69 imaging agents. Future studies will focus on functionalization of ZCD69:4 to allow radiolabeling with positron emitting nuclides (e.g. 18F and 68 Ga) to generate a construct useful for in vivo PET imaging of CD69.
