Hypomethylating agents synergize with irinotecan to improve response to chemotherapy in colorectal cancer cells

Abstract

Colorectal cancer (CRC) is the second leading cause of cancer death in the United States. In the metastatic setting, the majority of patients respond to initial therapies but eventually develop resistance and progress. In this study, we test the hypothesis that priming with epigenetic therapy sensitizes CRC cell lines, which were previously resistant to subsequent chemotherapeutic agents. When multiple CRC cell lines are first exposed to 500 nM of the DNA demethylating agent, 5-aza-cytidine (AZA) in-vitro, and the cells then established as in-vivo xenografts in untreated NOD-SCID mice; there is an enhanced response to cytotoxic chemotherapy with agents commonly used in CRC treatment. For irinotecan (IRI), growth diminished by 16-62 fold as assessed, by both proliferation (IC50) and anchorage independent cell growth soft agar assays. Treatment of resistant HCT116 cell line along with in-vivo, for CRC line xenografts, AZA plus IRI again exhibits this synergistic response with significant improvement in survival and tumor regression in the mice. Genome-wide expression correlates changes in pathways for cell adhesion and DNA repair with the above responses. A Phase 1/2 clinical trial testing this concept is already underway testing the clinical efficacy of this concept in IRI resistant, metastatic CRC (NCT01896856).

Conflict of interest statement

Competing Interests: There is a competing interest: Dr. Ahuja receives research grant funding from Astex Inc. and the Van Andel Research Institute. She has licensed methylation biomarkers to Cepheid. Dr. Ahuja also has served as consultant to Johnson and Johnson. This does not alter our adherence to PLOS ONE policies on sharing data and materials.

Figures

Fig 1. Low dose HMA causes global demethylation in CRC cell lines.

(A) Colon cancer cell lines (HCT116, SW480, Caco-2, RKO, SW620, Colo205, Colo320) showed very little or no cytotoxicity. Percent viability after 72 hrs treatment with 500 nM AZA, relative to mock control. Bars represent the mean of 3 replicates ±SD; (B & C) Colon Cancer cell lines saw basal or low expression of DNMT1 protein. Upper and the lower panel is a representative blot for DNMT1 and β-Actin as a housekeeper. (D) Protein expression of DNMT1 in colon cancer cells (Caco-2-, HCT116 and SW480) changes after treatment with a various concentration of AZA. Cells were treated with AZA at the indicated concentrations for 72hr and then subjected to immunoblot analysis using anti-DNMT1 antibody. Actin served as a loading control. B, C, and D are representative blots of at least three independent experiments. (E) Quantitation of DNA methylation using bisulfite LINE-1 PCR and Pyrosequencing in Caco-2, SW480 and HCT116 (3-day treatment followed by 3 rest period). Representative LINE-1 quantitation presented for the cell lines. The pyrogram quantitates C for methylated, and T for unmethylated DNA was plotted as a line graph.

Fig 2. In vivo epigenetic therapy sensitize CRC cell line xenografts to decrease tumor burden.

Colorectal cancer cells were treated for 72 hours with AZA at 500nM and saline, or PBS inject Mock, were allowed to recover from the short-term cytotoxic effects of AZA for seven days before injection into NOD-SCID mice (n = 10 in each case). NOD-SCID mice were xenografted with various CRC cell line and monitored until mice showed 2000 mm3 tumor development. (A-G) Represents xenografted mice responded to therapy and (H & I) are the mice which did not respond to the treatment (S3 Fig). Mean tumor volume (±SEM) over time are plotted. Statistical significance determined by two-tailed paired t-test.

Fig 3. Epigenetic therapy sensitizes chemotherapy in CRC cell lines.

(A-C) Log dose response curves for CRC cell lines treated (in triplicate) with IRI for 72 hours three days’ post epigenetic therapy. Individual curves represent the percentage of viable cells (±SD) for each epigenetic pretreatment condition normalized to its untreated control cells, such that the highest values for each pretreatment state represent 100%, and 0%. Data shown from representative experiments. Visualization of drug interaction is illustrated by CI-Fa plot. The CI-Fa plot represents the combination index plotted versus Fa, the fraction of affected enzyme or biological function. CI values were calculated from each Fa (i.e. various drug concentration) for CRC cell lines Caco-2, SW480, and HCT116. Average synergism (CI<1) at Fa>0.5 for all three CRC lines. The AZA and IRI chemotherapeutic doses may be significantly reduced for combinations that are synergistic at Fa>0.5 for all three cell lines. (D-F) Caco-2, SW480, and HCT116 cells were seeded on a solidified Matrigel layer six days after epigenetic therapy. Beginning the following day, cells were treated with chemotherapy for 72 hrs. The drug was then removed, and colonies were permitted to grow 2–4 additional days. Representative Caco-2, SW480 and HCT116 colonies following treatment with 80 nM IRI. Three independent experiments (total nine replicates. (G) HCT116 colony forming assay examined for prolonged exposure to AZA as a pretreatment to IRI on HCT116 cells. HCT116 cells were pretreated with AZA for 72 hours, followed by five days’ rest and the schedule was repeated three more times before IRI treatment in soft agar assay. This modification increased colony growth inhibition in HCT116 significantly more than IRI alone, which were previously resistant.

Fig 4. Epigenetic therapy in vivo sensitize SW480 xenografts to subsequent chemotherapy.

NOD-SCID mice bearing SW480 and HCT116 xenografts were treated with 0.5 mg/kg AZA (sc, qd × 5) and 10 mg/kg IRI (i.p., day 5), or vehicle, until the end of experiment cycles. (A&D) Mean tumor volume (±SEM) over time. Statistical significance determined by two-tailed paired t-test. (B&E) Survival analysis showed that the median survival for SW480 and HCT116 tumor-bearing mice was significantly reduced in AZA and IRI treated compared to mice which were untreated tumors. (C&F) Indicates weight changes over the course of treatment. Mice were either treated with AZA alone, AZA + IRI or IRI alone or saline vehicle. Mice were treated until the end of the experiment. N = 10 mice in each.

Fig 5. Relative expression of Caco-2 and SW480 colon cancer cell lines.

Heatmap showing relative expression (Mock is shown as reference) of the Sensitization Genes in the IRI and AZA+IRI-treated samples.