Tumor engraftment in nude mice and enrichment in stroma- related gene pathways predict poor survival and resistance to gemcitabine in patients with pancreatic cancer

Abstract

Purpose: The goal of this study was to evaluate prospectively the engraftment rate, factors influencing engraftment, and predictability of clinical outcome of low-passage xenografts from patients with resectable pancreatic ductal adenocarcinoma (PDA) and to establish a bank of PDA xenografts.

Experimental design: Patients with resectable PDA scheduled for resection at the Johns Hopkins Hospital were eligible. Representative pieces of tumor were implanted in nude mice. The status of the SMAD4 gene and content of tumor-generating cells were determined by immunohistochemistry. Gene expression was carried out by using a U133 Plus 2.0 array. Patients were followed for progression and survival.

Results: A total of 94 patients with PDA were resected, 69 tumors implanted in nude mice, and 42 (61%) engrafted. Engrafted carcinomas were more often SMAD4 mutant, and had a metastatic gene expression signature and worse prognosis. Tumors from patients resistant to gemcitabine were enriched in stroma-related gene pathways. Tumors sensitive to gemcitabine were enriched in cell cycle and pyrimidine gene pathways. The time to progression for patients who received treatment with gemcitabine for metastatic disease (n = 7) was double in patients with xenografts sensitive to gemcitabine.

Conclusion: A successful xenograft was generated in 61% of patients attempted, generating a pool of 42 PDA xenografts with significant biological information and annotated clinical data. Patients with PDA and SMAD4 inactivation have a better engraftment rate. Engraftment is a poor prognosis factor, and engrafted tumors have a metastatic gene expression signature. Tumors from gemcitabine-resistant patients were enriched in stromal pathways.

Figure 1

Patient flow chart. 94 patients with resected PDA were included in this study. 85 patients were eligible for xenografting. Patients who received neoadjuvant therapy or had stage IV resected PDA were excluded. Out of the 69 xenografted tumors, 42 were able to engraft and grew in nude mice

Figure 2

A. Engraftment rate was higher in patients with SMAD4 deletions B. SMAD4 deletions were unrelated to the differentiation grade of the tumor

Figure 2

A. Engraftment rate was higher in patients with SMAD4 deletions B. SMAD4 deletions were unrelated to the differentiation grade of the tumor

Figure 3

Xenografts show a metastatic signature of adenocarcinoma. We collected the gene expression profiles of primary tumors (F0) and corresponding xenografts (F5 and F10). For each xenograft the first column corresponds to primary tumor (F0). Second and third column are duplicates of passage #5 (F5), while fourth and fith column are duplicates of passage #10 (F10). We tested a gene expression signature from metastatic adenocarcinoma and found that our xenografts were enriched in this metastatic gene signature.

Figure 4

Kaplan-Meier survival curves as a function of engraftment. Non engrafted patients had a decrease in the risk of death of 81%

Figure 5

Waterfall plot with gemcitabine activity in xenografts. We applied RECIST criteria to T/C to define sensitivity/resistance. We reduced the threshold of sensitivity to −15% in order to have sensitive xenografts, since none of them had a T/C <−30%. Green bars represent sensitive cases (T/C<−15%), gray bars represent stable disease (T/C −15 to 20%) and red bars represent progressive disease (T/C >20%)