Targeting the metabolic pathway of human colon cancer overcomes resistance to TRAIL-induced apoptosis

Abstract

Colon cancer is a leading cause of cancer-related mortality for which targeted therapy is needed; however, trials using apoptosis-inducing ligand monotherapy to overcome resistance to apoptosis have not shown clinical responses. Since colon cancer cells selectively uptake and rapidly metabolize glucose, a property utilized for clinical staging, we investigated mechanisms to alter glucose metabolism in order to selectively target the cancer cells and to overcome evasion of apoptosis. We demonstrate TRAIL (tumor necrosis factor-related apoptosis-inducing ligand) resistance in the majority of human colon cancers tested and utilize the glucose analog 2-deoxy-d-glucose to sensitize TRAIL-resistant gastrointestinal adenocarcinoma cells, and not normal gastrointestinal epithelial cells, to TRAIL-induced apoptosis through enhanced death receptor 5 expression, downstream modulation of MAPK signaling and subsequent miRNA expression modulation by increasing the expression of miR-494 via MEK activation. Further, established human colon cancer xenografts treated with this strategy experience anti-tumor responses. These findings in colon adenocarcinoma support further investigation of manipulation of cellular energetics to selectively overcome resistance to apoptosis and to impart tumor regressions in established colon cancer tumors.

Figure 1

2DG synergizes with TRAIL to induce cell death in TRAIL-resistant cancer cell lines but not in normal cells. (a) HeLa cells (positive control) are sensitive to TRAIL-induced apoptosis; however, the majority of human colon (HT-29, SW-620), stomach (AGS) and pancreatic cancers (PANC-1) tested were TRAIL resistant. (b–e) TRAIL sensitization upon treatment with 2DG. (f, g) Nonmalignant cell controls (InEpC, normal human intestinal epithelial cells; MMH1-1.4, hepatocytes) show no effect of 2DG treatment on TRAIL-induced apoptosis. Apoptosis is detected by caspase-3 activation. n=3, and was validated with Annexin V+PI and TMRM staining (Supplementary Figure S1). ** indicates P<0.01, ns indicates P>0.05.

Figure 2

2DG+ TRAIL enhance activation and are dependent upon the extrinsic apoptotic pathway to induce colon cancer cell death. (a) Treating TRAIL-resistant HT29 colon cancer cells with 2DG, TRAIL or 2DG+TRAIL resulted in robust caspase-8 and caspase-3 activation with combination therapy. Cleavage products p18 and p10 are indicative of caspase-8 activation while p19 and p17 represent caspase-3 activation. (b) HT-29 cells were fractionated into cytoplasmic and membrane heavy fractions. Truncated Bid (tBid) in the membrane heavy fraction was only found in the 2DG+TRAIL-treated cells. (c) HT-29 cells were treated with the general caspase inhibitor, Q-VD-OPh. Apoptosis was assessed by Annexin V/PI staining revealing inhibition of 2DG+TRAIL-induced apoptosis. (d) HT-29 cells were stably transfected with CrmA, DN-FADD plasmid constructs or an empty vector. Blockade of the extrinsic apoptotic pathway abrogated 2DG and TRAIL synergy. n=3, * indicates P<0.05, ** indicates P<0.01.

Figure 3

2DG increases DR5 expression levels with no effect on DR4, however, the increase in TRAIL-induced apoptosis requires additional mechanisms. (a) QRT-PCR of DR4 and DR5 transcripts in HT29 cells with and without 2DG treatment. (b) Immunoblotting for DR4 and DR5 indicated increased DR5 protein expression with 2DG treatment. (c) DR surface expression in various TRAIL resistant colon, gastric and pancreatic adenocarcinoma cells. (d) DR5 surface expression increases with 2DG treatment over time in HT-29 cells. (e) Apoptosis, as assessed by capsase-3 activation, in the same group of HT-29 cells increased over a 24 h time course upon treatment with 2DG+TRAIL, corresponding temporally with the increase in DR5 expression. (f) To knock down DR5 expression, HT-29 cells were transduced with the indicated lentiviral constructs and allowed to grow for 48 h. A lentivirus expressing a scrambled sequence was used as a control. (g) Cell death was significantly decreased with DR5 knockdown in 2DG+TRAIL-treated cells, however, remained significantly higher than wild-type cells treated with 2DG or TRAIL alone. n=3, ** indicates P<0.01, ns indicates P>0.05.

Figure 4

MicroRNA levels are increased with 2DG+TRAIL treatment and the effect of miR inhibitors and mimics on 2DG+TRAIL-induced apoptosis. (a) Heat map demonstrating relative increased expression levels of miR-1246, -4488, -4516 and -494 with combination treatment (shown in biological duplicates from two separate experiments). MiR-expression is upregulated and clusters in the 2DG+TRAIL-treated cells. (b) Fold change of miRs relative to untreated controls reveal increased miR-1246, -4488, -494 and -4516. Experiment was repeated in duplicate with Nanostring nCounter gene expression analysis and expression patterns were validated with RT-PCR. (c) HT-29 cells were transfected with miR inhibitors (upper panel) or mimics (lower panel). Untransfected cells and those transfected with a negative control were used as controls. Twenty-four hours post-transfection, cells were treated with 2DG, TRAIL or 2DG+TRAIL for 24 h. Inhibited expression of miRs 494 and 1246 significantly abrogated the effect of 2DG+TRAIL on apoptosis. Only overexpression of miR-494 (lower panel) sensitized cells to TRAIL-induced apoptosis even without the addition of 2DG. Apoptosis was assessed by measuring caspase-3 activation. n=3, ** indicates P<0.01, * indicates P<0.05, ns indicates P>0.05.

Figure 5

MEK signaling is necessary for 2DG+TRAIL synergy in inducing apoptosis. (a) MEK inhibition decreased apoptosis in 2DG+TRAIL-treated cells to levels comparable to TRAIL alone. (b) Cells were engineered to express constitutively activated (ca) MEK1 (1.8× compared with empty vector) or dominant-negative (dn) MEK1 (1.2× compared with empty vector). Protein expression on western blot analysis demonstrated an expected increase or decrease in phospho(p)-ERK1/2 (1.4× versus 0.8×, respectively) resulting in increased or decreased apoptosis with caMEK1 and dnMEK1, respectively. Levels of activated caspase-3 after 24 h in cells with CA-MEK were increased 2.5-fold (6.7% versus 2.65%) over control cells transfected with an empty vector, and 2.1-fold (6.7% versus 3.2%) over DN-MEK cells (c) Western blot analysis of protein expression of p-MEK and ERK was (d) adjusted for b-actin expression. (e) pMEK levels reached peak expression faster (increased slope) and to a greater degree in cells treated with 2DG+TRAIL. The slope/peak of phosphorylation over this time period for 2DG-treated, TRAIL-treated and combination-treated cells was 0.33/1.66×, 0.09/1.18× and 0.5/1.99×. Similarly, p-ERK levels increased rapidly and to a greater degree in the 2DG+TRAIL-treated cells. The slope/peak of phosphorylation for 2DG-treated, TRAIL-treated and combination-treated cells was 0.43/2.34×, 0.36/2.56× and 0.79/4.15×.

Figure 6

MEK1/2 are upstream of miR-494 and regulate its expression in colon cancer cells treated with 2DG+TRAIL. (a) miR-494 transcript levels were significantly increased in CA-MEK1-expressing cells compared with DN-MEK1 (**P=0.0002). (b) MEK inhibition with an MEK1 (PD98059) and an MEK1/2 inhibitor (U0126) caused miR-494 transcript levels in the 2DG+TRAIL-treated group to decrease 18- and 10-fold to near undetectable levels. (c, d) Caspase-3 activation was markedly enhanced with the combination treatment but was reduced to levels comparable to non-treated control cells in the presence of an MEK inhibitor. This inhibition of apoptosis was abrogated when these same cells were made to overexpress miR-494. n=3. (e) Summary diagram of the proposed pathway.

Figure 7

2DG and TRAIL cause regression of established solid tumors in vivo. (a) Representative images comparing tumor burden of an untreated mouse and one given 2DG+TRAIL before (day 0) and after (day 5) treatment following establishment of palpable tumors. (b) Photograph of representative tumors excised from mice of each experimental group after completion of treatment. (c) Tumor growth curves after initiation of treatment regimen. Data represent the percent change in tumor volume. (d) Waterfall plot where each bar represents a single mouse of the indicated treatment group. The data represent the percent change in tumor volume after 5 days of treatment relative to tumor volume on day 0. ** indicates P<0.01, * indicates P<0.05, ns indicates P>0.05.