Phase I Study of Epigenetic Priming with Azacitidine Prior to Standard Neoadjuvant Chemotherapy for Patients with Resectable Gastric and Esophageal Adenocarcinoma: Evidence of Tumor Hypomethylation as an Indicator of Major Histopathologic Response

Abstract

Purpose: Epigenetic silencing of tumor suppressor genes (TSG) is an acquired abnormality observed in cancer and is prototypically linked to DNA methylation. We postulated that pretreatment (priming) with 5-azacitidine would increase the efficacy of chemotherapy by reactivating TSGs. This study was conducted to identify a tolerable dose of 5-azacitidine prior to EOX (epirubicin, oxaliplatin, capecitabine) neoadjuvant chemotherapy in patients with locally advanced esophageal/gastric adenocarcinoma (EGC).Experimental Design: Eligible patients had untreated, locally advanced, resectable EGC, ECOG 0-2, and adequate organ function. 5-Azacitidine (V, 75 mg/m2) was given subcutaneously for 3 (dose level, DL 1) or 5 (DL 2) days prior to each 21-day cycle of EOX (E, 50 mg/m2; O, 130 mg/m2; X, 625 mg/m2 twice daily for 21 days). Standard 3+3 methodology guided V dose escalation. DNA methylation at control and biomarker regions was measured by digital droplet, bisulfite qPCR in tumor samples collected at baseline and at resection.Results: All subjects underwent complete resection of residual tumor (R0). Three of the 12 patients (25%) achieved a surgical complete response and 5 had partial responses. The overall response rate was 67%. The most common toxicities were gastrointestinal and hematologic. Hypomethylation of biomarker genes was observed at all dose levels and trended with therapeutic response.Conclusions: Neoadjuvant VEOX was well-tolerated with significant clinical and epigenetic responses, with preliminary evidence that priming with V prior to chemotherapy may augment chemotherapy efficacy. The recommended phase II trial schedule is 5-azacitidine 75 mg/m2 for 5 days followed by EOX chemotherapy every 21 days. Clin Cancer Res; 23(11); 2673-80. ©2016 AACR.

Conflict of interest statement

Conflicts of Interest: No author reported relevant financial interests in this manuscript with the following exceptions: Dr. Allyson Ocean received honoraria from Celgene while serving on the speaker bureau for nab-paclitaxel.

©2016 American Association for Cancer Research.

Figures

Figure 1. VEOX Decreased DNA Methylation and Induced Re-expression of TSGs in Tumor Cells

(A) Schematic of sample processing and bisulfite ddPCR analysis of DNA methylation in FFPE tumor specimens. (B) The percentage of patients with detectable DNA methylation at each locus tested is shown. (C) The percentage of tumor-associated loci tested (HPP1, TIMP3, CDKN2A, ESR1 and MGMT) with detectable DNA methylation is shown for each patient. (D) The change in methylation in the surgically resected specimen compared to the diagnostic (pre-VEOX) specimen was compared for all loci. Shown here is the average change in DNA methylation at informative tumor associated loci for each patient. The patients are ordered by the DNA methylation effect and the bars are colored by the pathologic response (CR-black, PR-grey, SD-white). The red line indicates a two-fold reduction in DNA methylation after neoadjuvant VEOX. Patients with greater than two-fold hypomethylation had better responses in this data set (p=0.03, Fisher Exact Test). (E) Patients with complete or partial response to neoadjuvant VEOX tended to have greater DNA hypomethylation of the TSGs tested (Student’s t-test, p=0.057). (F) Immunohistochemical stains for HPP1 demonstrate weak cytoplasmic staining of tumor cells (arrow) in pretreatment biopsy samples (left panel). Residual cancer in the post-treatment surgical resection specimen show more intense cytoplasmic staining for HPP1 (right panel, original magnification: 200x).

Figure 2. Survival after VEOX Therapy

(A) Overall survival (OS) is shown for all subjects after VEOX neoadjuvant chemotherapy (N=12 patients, 3 deaths). Median OS not reached. 24-month OS = 80.0% (95% CI = 40.9%, 94.6%). (B) Disease Free Survival (DFS) is shown for all subjects (N=12 patients, 6 recurrences). Median DFS = 20.5 months (95% CI = 8.7 months, upper limit not estimated). 12-month DFS = 83.3% (95% CI = 48.2%, 95.6%). 24-month DFS = 46.3% (95% CI = 17.2%, 71.4%). (C) DFS is shown by VEOX response. CR (N=3 patients, 0 recurrences): 100% DFS. PR (N=5 patients, 3 recurrences): 24-month DFS = 40.0% (95% CI = 5.2%, 75.3%). SD (N=4 patients, 3 recurrences): 24-month DFS = 25.0% (95% CI = 0.9%, 66.5%). P=0.33 by log-rank test.