(F-213) Modular, targeted microparticles for visualization of heterogeneous colorectal lesions

Introduction:: Introduction: Despite increased screening efforts, colorectal cancer (CRC) remains a leading cause of cancer-related deaths in the United States. Standard colonoscopy enables identification and removal of large, adenomatous polyps, but the method’s reliance on visual identification limits its ability to detect flat lesions. Biomarkers expressed on a lesion surface present an opportunity for specific early detection of pre-cancerous lesions. However, the multivalent biological signatures of heterogeneous colorectal lesions increase the challenge of detection, requiring concurrent targeting of multiple surface biomarkers with molecular probes. High affinity recognition molecules such as antibodies or aptamers can be tethered to microparticles (MPs), which act as diagnostic contrast agents for CRC detection. Protein-L, a bacteria-derived protein that binds the kappa light chains on IgG, offers a universal route toward tailored, non-covalent attachment of antibodies to the surfaces of microparticle beacons. Here, we present the utility of this method in identifying common, but orthogonal, CRC cell surface markers, MUC1 and EpCAM, across 2D, 3D tumoroid, and in vivo rodent models, using high-contrast MPs coated with protein-L adapters for universal use across antibodies (Fig 1a).

Materials and Methods:: Materials and Methods: Mouse IgG1 antibodies against human EpCAM or MUC1 were labeled along the Fc domain with AlexaFluor fluorophores via SiteClick (ThermoFisher) azide-alkyne reaction to antibody probes on a modified sugar on the Fc region for visualization. Pre-labeled non-targeting mouse IgG1 was used as a control. Iron oxide magnetic MPs, 1 µm diameter, pre-functionalized with protein L were further decorated with the labeled antibodies (as EpCAM-MPs, MUC1-MPs, and CTRL-MPs), and confirmed through microscopy. Human CRC cell lines with high surface expression of EpCAM (HT29-luciferase) and MUC1 (HT29-MTX) were clustered to form 150-200µm 3D tumoroids, as model in vitro systems to demonstrate selective MP binding (Fig 1c). Optionally, these lines were pre-dyed, combined in ratios, and imaged to confirm even surface distribution of each cell type. In vivo orthotopic CRC models were established in nude mice by injecting high-density mixed HT29-MTX/HT29-luciferase cell suspensions within the colonic mucosa to generate heterogeneous lesions. Mature tumors were confirmed by IVIS luciferase imaging, excised, and exposed to suspensions of each functionalized MP to assess binding via confocal microscopy.

Results, Conclusions, and Discussions:: Results, Conclusions, and Discussion: Covalent attachment of AlexaFluor fluors, via click reaction, to the antibody probes was confirmed using spectroscopy. Loading of labeled antibodies onto protein-L MPs was measured via microscopy (Fig 1b), and confirmed ~110 µg of labeled IgG antibody per mg of MP as the minimum necessary concentration to ensure saturated loading onto MPs. EpCAM^hi tumoroids (HT29-luciferase) stained positive for EpCAM by conventional immunostaining, and separate staining with EpCAM-MPs yielded equivalent staining patterns, on both live and fixed tumoroids. (Fig 1d) Tumoroids of HT29-MTX cells were identified by both MUC1-MPs and EpCAM-MPs. No detection of any CRC spheroid was observed for CTRL-MPs. To prepare orthotopic rodent models, mixtures of HT29-MTX and HT29-luciferase cells were formed as spheroids and stained, and it was observed that a 70:30 ratio, respectively, permitted surface accessibility for both cell types and their epitopes (Fig 2a). To establish the in vivo model, this same ratio was used when injecting cells at an oblique angle, through a colonic catheter and via colonoscope guidance, into the rodent colonic mucosa. Mature orthotopic tumors developed within 21 days, as observed by luminescence imaging of injected luciferase, and were excised for evaluation. Tissue staining with DAPI identified nuclei and tissue morphologies, and established zones of healthy colonic villi, and dysplastic overgrowth. Dysplastic regions were confirmed as human with immunostaining for human nuclear antigen (HNA). Coincident staining with MUC1-MP and EpCAM-MP microparticles selectively identified complementary regions of human cell dysplasia, both by detection of the antibody-linked fluors and brightfield-based imaging of MPs (Fig 2b). CTRL-MP application showed extremely low baseline levels of staining (Fig 2c). Normal regions of mouse colon retained a crypt morphology, and were free of MP staining in those regions (Fig 2d).

In summary, protein L coated MPs were functionalized with antibodies against multiple targets, to recognize heterogeneous CRC surface markers across multiple in vitro and in vivo models. These molecular probes can be combined with existing imaging technologies such as confocal laser endomicroscopy or hyperpolarized MRI to achieve early, real-time, and potentially non-invasive, detection of pre-cancerous lesions.

Acknowledgements (Optional): :

References (Optional): :