Microdistribution of targeted, fluorescently labeled anti-carcinoembryonic antigen antibody in metastatic colorectal cancer: implications for radioimmunotherapy.

PURPOSE: Most radioimmunotherapy studies on radiolabeled antibody distribution are based on autoradiographic and radioluminographic data, which provide a lack of detailed information due to low resolution. We used fluorescently labeled anti-carcinoembryonic antigen (CEA) antibody (A5B7) to investigate quantitatively the kinetics and microdistribution of antibody in a clinically relevant orthotopic colorectal cancer model (LS174T) using high-resolution digital microscopy. EXPERIMENTAL DESIGN: Nude mice bearing LS174T liver orthotopic tumors received a single i.v. injection of fluorescently labeled A5B7 and were sacrificed at 10 minutes, 1 hour, or 24 hours postinjection. Before sacrifice, mice were injected with the perfusion marker Hoechst 33342. An anti-CD31 antibody was used to detect blood vessel distribution. Cryostat sections were processed with immunofluorescence procedures and analyzed with fluorescence microscopy and image analysis techniques. The fluorescence images were related to morphologic images of the same or adjacent tumor sections. RESULTS: Fluorescently labeled antibody showed rapid, selective uptake into tumor deposits, with a strong negative correlation with tumor size at 10 minutes and 1 hour (P < or = 0.01). By 24 hours, the correlation was no longer significant. The study showed movement of antibody across the tumor with time and a tendency to localize more uniformly by later time points (24 hours). The rate of antibody motility was similar in small and large tumor metastases, but small deposits showed more rapid antibody localization. Intratumoral vessels were positively related to tumor size (P < or = 0.001). CONCLUSION: The obtained data suggest that radioimmunotherapy can be highly efficient in an adjuvant or minimal residual disease setting.

Predicting response to radioimmunotherapy from the tumor microenvironment of colorectal carcinomas.

Solid tumors have a heterogeneous pathophysiology, which directly affects antibody-targeted therapies. Here, we consider the influence of selected tumor parameters on radioimmunotherapy, by comparing the gross biodistribution, microdistribution, and therapeutic efficacy of either radiolabeled or fluorescently labeled antibodies (A5B7 anti-carcinoembryonic antigen antibody and a nonspecific control) after i.v. injection in two contrasting human colorectal xenografts in MF1 nude mice. The LS174T is moderately/poorly differentiated, whereas SW1222 has a well-differentiated glandular structure. Biodistribution studies (1.8 MBq (131)I-labeled A5B7, four mice per group) showed similar gross tumor uptake at 48 h in the two models (25.1% and 24.0% injected dose per gram, respectively). However, in therapy studies (six mice per group), LS174T required a 3-fold increase in dose (18 versus 6 MBq) to equal SW1222 growth inhibition ( approximately 55 versus approximately 60 days, respectively). To investigate the basis of this discrepancy, high-resolution multifluorescence microscopy was used to study antibody localization in relation to tumor parameters (5 min, 1 and 24 h, four mice per time point). Three-dimensional microvascular corrosion casting and transmission electron microscopy showed further structural differences between xenografts. Vascular supply, overall antigen distribution, and tumor structure varied greatly between models, and were principally responsible for major differences in antibody localization and subsequent therapeutic efficacy. The study shows that multiparameter, high-resolution imaging of both therapeutic and tumor microenvironment is required to comprehend complex antibody-tumor interactions, and to determine which tumor regions are being successfully treated. This will inform the design of optimized clinical trials of single and combined agents, and aid individual patient selection for antibody-targeted therapies.