K. E. Cox1,2, S. Amirfakhri1,2, T. M. Lwin3, P. Ghosh4,5, M. Obonyo5, R. M. Hoffman6, P. J. Yazaki7, M. Bouvet1,2 1University Of California – San Diego, Department Of Surgical Oncology, San Diego, CA, USA 2VA San Diego Healthcare System, Department Of Surgery, La Jolla, CA, USA 3City Of Hope National Medical Center, Department Of Surgical Oncology, Duarte, CA, USA 4University Of California – San Diego, Department Of Cellular And Molecular Medicine, San Diego, CA, USA 5University Of California – San Diego, Department Of Medicine, San Diego, CA, USA 6AntiCancer, Inc., San Diego, CA, USA 7Beckman Research Institute of the City of Hope, Department Of Molecular Imaging And Therapy, Duarte, CA, USA
Introduction:
Gastric cancer poses a major diagnostic and therapeutic challenge with a global incidence of ~1 million cases per year. Approximately 10% of primary gastric cancers present as a diffuse thickening of the wall known as linitis plastica rather than as a solitary mass. Because of this, linitis plastica is often asymptomatic until late stages and carries a poor prognosis (median survival ranges from 5 to 17 months). In the present study, the first mouse model of linitis plastica was developed. We previously developed the first orthotopic mouse model of gastric cancer and have demonstrated the ability of an anti-carcinoembryonic antigen (CEA) monoclonal antibody (mAb) to label solitary gastric tumors and now show its ability to label infiltrative lesions.
Methods:
One million cells of the human gastric cancer cell line MKN45 (poorly differentiated gastric adenocarcinoma) were suspended in 50 µl of equal parts PBS and Matrigel. Athymic mice were anesthetized prior to making a 1-2 cm transverse incision in the left upper quadrant through which the stomach was delivered. Using a 30-gauge needle, the MKN45 cells were injected into the muscular layers of the gastric wall. The stomach was then returned to the abdomen and the incision was closed. For tumor labeling, the humanized anti-CEA M5A mAb, or non-specific IgG as a control, were conjugated with the near-infrared dye IRDye800CW. After 8 weeks, the mice received an intravenous injection of 50 µg M5A-IR800 (n = 15) or 50 µg IgG-IR800 (n = 12) and were imaged 72 hours later. Fluorescence imaging was performed using the LI-COR Pearl Imaging System. A tumor to background ratio (TBR) was calculated by dividing mean fluorescence intensity of the tumor by the adjacent stomach which was used as a background tissue.
Results:
Infiltrative tumors were identified in 21 of the 27 mice (11 treated with M5A-IR800 and 9 treated with IgG-IR800). M5A-IR800 administration resulted in bright labeling of tumors with an average TBR of 4.076 (SE±0.422). Labeling with IgG-IR800 was non-specific with an average TBR of 0.785 (SE±0.233). The average maximal fluorescence signal from a single foci within the tumors labeled with M5A-IR800 was 21 times greater than those labeled with IgG-IR800 (p-value < 0.001).
Conclusion:
The present study describes the first mouse model of linitis plastica gastric cancer. The humanized anti-CEA antibody (M5A-IR800) brightly labeled infiltrative gastric tumors that were otherwise imperceptible under bright light, with tumor to background ratios 5 times greater than the control group. This new model of linitis plastica can be used to identify additional agents for visualizing tumor and their margins during surgical resection.
