Molecular Imaging of Tumor-Infiltrating Lymphocytes in Living Animals Using a Novel mCD3 Fibronectin Scaffold

Abstract

The interaction between cancer cells and immune cells in the tumor microenvironment (TME) plays a crucial role in determining tumor growth, metastasis, and response to treatment. Tumor-infiltrating lymphocytes (TILs) in TME could be a predictive marker for treatment response in various therapeutic interventions, including chemotherapy and immunotherapy. Thus, imaging the tumor immune microenvironment is important for selecting the optimal treatment strategies in cancer therapy. The CD3 protein represents a promising target for diagnostic imaging of TILs in vivo to assess the immune state of the TME. Although many anti-CD3 antibodies have been explored for this application, the nonspecific immune activation by these antibodies limits their applications. To overcome this issue, we engineered a novel fibronectin III domain (FN3) protein binder (mCD3-FN3;11.8 kDa) against mouse CD3 antigen protein using a yeast display library to image TILs homing in vivo into the TME. We performed in vitro and in vivo assays to test the mCD3-FN3 binder purity as well as in vivo targetability in mouse models of syngeneic tumors. We used near-infrared 800 dye conjugated with mCD3-FN3 (IR800-mCD3-FN3) for in vivo tracking of TILs via optical imaging. We used three different syngeneic tumors in mice (mCD3⁺ EL4 tumor in C57BL/6 mice, mCD3^– CT26 colon tumor, and mCD3^– 4T1 breast tumor in BALB/c mice) for imaging TILs in vivo. C57BL/6 mice bearing EL4 tumors were separated into two groups (blocking [Blk] and nonblocking [Nblk]; n = 3 per group) and used for in vivo imaging. Blocking groups received 200 μg of unlabeled mCD3-FN3 2 h prior to the administration of IR800-mCD3-FN3 binder. Each mouse was administered with 25 μg of the IR800-mCD3-FN3 binder and tracked using an IVIS optical imaging system over time. C57BL/6/EL4 mice were imaged at 4 and 24 h post injection of the IR800-mCD3-FN3 binder, and mouse organs were collected at 24 h after final imaging and used for ex vivo histological imaging. In CT26 and 4T1 tumor models, TILs in TME were imaged 4, 24, and 48 h after binder injection. The NIR imaging of EL4 tumors showed that IR800-mCD3-FN3 can detect both TILs within the tumor and the tumor cells with a high signal-to-background ratio 24 h after initial binder injection with a total radiant efficiency (mean TRE ± SD) of 6.5 × 10¹⁰ ± 1.5 × 10¹⁰ [photons/s]/[μW/cm²]. The animals received preinjection of unlabeled mCD3-FN3(Blk) prior to IR800-mCD3-FN3 binder administration and showed a significant level of fluorescence signal reduction (mean TRE ± SD: 1.6 × 10¹⁰ ± 4.1 × 10⁹) in the tumor when compared to the EL4-Nblk tumors (p = 0.006). The mouse group with CT26 and 4T1 tumors where the probe can only bind to TILs within the tumor showed a specific imaging signal (mean TRE ± SD) of 1.1 × 10¹¹ ± 5.2 × 10¹⁰ and 9.5 × 10¹⁰ ± 4.6 × 10¹⁰, respectively, at 48 h p.i. For these groups, the ex vivo tumor-to-muscle ratios were 20- and 27-fold for CT26 and 4T1 tumors, respectively. These results clearly demonstrate the in vivo binding ability of the mCD3-FN3 binder to mCD3 marker expressed by T cells in the TME. The ex vivo histological analysis of tumors, and the organs of animals with EL4 tumors, and TILs imaging of CT26, and 4T1 tumors (at 48 p.i.) confirmed that the IR800-mCD3-FN3 probe was able to specifically bind to CD3 markers expressed by the T cells. In summary, both in vitro and in vivo data indicated that the engineered mCD3-FN3 binder by this study is a promising ligand for diagnostic imaging of tumors in vivo for the assessment of mCD3 expressing TILs in the TME. This can be used as a prognostic marker in evaluating tumor response to therapeutic intervention as well as a diagnostic marker in imaging tumor response to immune checkpoint blockade cancer therapies.