M. H. Gerber1, S. Judge2, D. Delitto1, A. Delitto2, R. Nosacka2, S. Wallet3, A. Judge2, J. Trevino1 1University Of Florida,Department Of Surgery,Gainesville, FL, USA 2University Of Florida,Department Of Physical Therapy,Gainesville, FL, USA 3University Of Florida,Department Of Oral Biology,Gainesville, FL, USA
Introduction: Cancer cachexia is a debilitating syndrome associated with pancreatic adenocarcinoma that is not well understood. Patients with pancreatic adenocarcinoma present with a wide range of cachexia that is not necessarily related to tumor burden. Current preclinical models for studying pancreatic cancer cachexia involve subcutaneously injecting highly passaged cell lines into the flanks of mice. Cellular injections lack the tumor associated stroma component and its associated local and systemic immune signals. We hypothesize that using true patient derived xenografts (PDX), we can mimic cancer cachexia and immune signaling phenotypes in a murine model.
Methods: Human pancreatic cancer PDX were created by implanting fresh surgical specimens of pancreatic cancer directly into the flanks of mice. At the time of surgery, a separate tumor sample and muscle biopsy were obtained from the patients. The tumor, spleen, and muscle tissues were collected from mice at endpoint. Expression of genes associated with cachexia were analyzed in both the human and mouse muscles. The patient tumor sample as well as the mouse tumors and spleens were lysed and analyzed using Luminex technology to detect multiple cytokines and chemokines of both human and mouse origin.
Results: Two patients with varying degrees of cachexia had PDX murine models created by implanting each graft subcutaneously into the flanks 5 mice labeled G1 and G2 with 5 mice receiving sham surgery. The two patients had varying degrees of fold changes in cachexia related genes including FoxO1, Socs3, STAT3, Atrogin-1, and MuRF1 that were significantly different compared to controls. The PDX mice mirrored the patient variability in muscle atrophy gene expression fold changes. Both G1 and G2 PDX mice had significant muscle atrophy compared to controls in the tibialis anterior muscles (G1: P < 0.001; G2: P < 0.001) and the gastrocnemius muscles (G1: P < 0.001; G2: P < 0.001). The G1 and G2 PDX tumor and spleen lysates had significant differences in expression of multiple human and mouse cytokines/chemokines including KC, IL-8, IL-6, IP-10, IL-1beta, VEGF, GRO, FGF-2, GM-CSF, TGF-alpha, MDC, and IFN-alpha. The relative cytokine and chemokine ratios in PDX G1 and G2 tumor lysates were similar in their respective patient tumor lysates.
Conclusion: Patient derived xenografts in pancreatic ductal adenocarcinoma capture cancer cachexia and preserve the cytokine/chemokine patient heterogeneity found in the local and systemic immune environments. The preservation of individual patient tumors using this PDX model will help us better understand the driving forces behind cancer cachexia and may lead towards personalized therapies.