Folinic acid, Trolox, salubrinal, succinylacetone, and β-actin antibody were purchased from Sigma-Aldrich (St. Louis, MO). Fluorouracil was from LKT Laboratories (St. Paul, MN). Oxaliplatin was from LC Laboratories (Woburn, MA). Antibodies directed toward full-length (FL)/cleaved caspase-3, FL/cleaved PARP, phospho-eIF2α (P-eIF2α), total eIF2α (T-eIF2α), P-PERK, and total-PERK were from Cell Signaling Technology (Beverly, MA). Anti-CHOP10 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-GAPDH antibody and GSK2606414 were obtained from EMD Millipore (Burlington, MA). Anti-ferritin antibody was purchased from Abcam (Cambridge, MA). Anti-rabbit 800CW and anti-mouse 680RD secondary antibody IRDyes were from LI-COR Biosciences (Lincoln, NE). The heme oxygenase inhibitor, QC-308, was purchased from AsisChem Inc. (Waltham, MA).

The human colon cancer cell lines HT29 and HCT116 were cultured in McCoy's 5A medium (Sigma Aldrich, St. Louis, MO) containing 10% heat-inactivated fetal bovine serum (FBS), penicillin (100 units/ml), and streptomycin (100 μg/ml). The non-tumorigenic colon cell line, FHC, was cultured in DMEM: F12 (1:1) medium supplemented with 10% heat inactivated FBS, 25 mM HEPES (Thermo Fisher Scientific Inc., IL), 10 ng/ml cholera toxin (Sigma Aldrich, St. Louis, MO), 0.005 mg/ml insulin (Thermo Fisher Scientific Inc., IL), 0.005 mg/ml, transferrin (Sigma Aldrich, St. Louis, MO), 100 ng/ml hydrocortisone (Sigma Aldrich, St. Louis, MO), 20 ng/mL human recombinant epidermal growth factor (Thermo Fisher Scientific Inc., IL), 100 units/ml penicillin, and 100 μg/ml streptomycin.

Cells were cultured in 96-well-plates for 48 h before treatment. Serum-free media containing different agents was added to the cells at the concentration and time period described in the figure legend. MTS reagent (Promega, Madison, WI) was then added to each well and the absorbance was measured at 495 nm as directed by the manufacturer. In the presence of MTS reagent, the absorbance reading is proportional to the number of viable cells. The half-maximal inhibitory concentration (IC₅₀) of the tested compounds is the concentration that reduces the viability of cells by 50%. IC₅₀ was calculated after 24 h of treatment (30, 31) as described previously (32–34) from dose-response curves (log drug concentration vs. percentage viability from untreated cells) generated by non-linear regression analysis with GraphPad Prism 5 software (GraphPad Software, San Diego, California).

The cells were plated in white-walled 96-well-plates and then they were cultured at 37°C for 48 h. Serum-free culture medium containing the appropriate concentration of different agents was added to the cells for the indicated period of time. Caspase-Glo 3/7 reagent (Promega, Madison, WI) was added to each well and luminescence was detected as directed by the manufacturer. The Caspase-Glo 3/7 kit measures the activity of the executioner caspases 3 and 7 using the luminogenic substrate (Z-DEVD-aminoluciferin) as substrate.

HT29, HCT116, and FHC cells were cultured in chamber slides and treated with 5 μM NSC735847, 10 μM NSC735847, or vehicle (0.1% DMSO). The cells were then washed with PBS and fixed in 4% paraformaldehyde before detection of double stranded DNA breaks that are characteristic of apoptosis. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays were performed as described by the manufacturer (in situ Cell Death Detection Kit; Roche; Indianapolis, IN). Briefly, fixed cells were incubated in permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate) for 2 min on ice. The slides were rinsed twice with PBS, 50 μL TUNEL reaction mixture was added, and then the samples were incubated for 1 h at 37°C in the dark. The slides were overlayed with antifade mounting medium containing DAPI and then visualized using a Zeiss LSM 700 confocal microscope system (Carl Zeiss Microscopy, LLC, White Plains, NY).

Western blot analysis was conducted as described previously (35). Briefly, the cells were cultured in 100 mm dishes for 48 h before treatment. The cells were then treated as described in the figure legends. For vehicle-treated samples, DMSO was added to serum-free culture medium at a maximum concentration of 0.1% (v/v). After experimentation, the cells were washed twice with phosphate buffer saline (PBS) and then 100 μl of Triton lysis buffer [25 mM HEPES, 100 mM NaCl, 1 mM EDTA, 10% (v/v) glycerol, 1% (v/v) Triton X-100], containing protease and phosphates inhibitors, was added to the cells. Protein concentrations were determined by using BCA reagents (Thermo Fisher Scientific Inc., IL). Equal concentrations of each sample were loaded onto SDS–PAGE gels and protein bands transferred to nitrocellulose membranes using semi-dry transfer cells (TRANS-BLOTSD; Bio-Rad Laboratories, Hercules, CA). The membranes were incubated for 2 h at room temperature in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE). Membranes were then incubated overnight at 4C with primary antibody followed by 1 h of incubation at room temperature with secondary antibody. All primary antibodies were used at a dilution of 1:1,000 except anti-GAPDH and anti-CHOP10 antibodies which were used at 1:10,000 and 1:500 dilutions, respectively. All secondary antibodies were used at a dilution of 1:15,000. Protein bands were visualized using Odyssey® CLx digital fluorescence imaging system (LI-COR, Lincoln, Nebraska). Band intensities were quantified by using ImageJ software (36).

Oxidative stress was measured in cultured cells by using the probe, chloromethyl-2′,7′ dichlorodihydrofluorescein diacetate (CM-H₂DCFDA or DCF) purchased from Life Technologies (Grand Island, NY). DCF is a chemically reduced form of fluorescein used as an indicator for reactive oxygen species (ROS) in cells. The cells were loaded with 5 μM DCF in phenol red-free, serum-free medium for 30 min and then the cells were treated with the appropriate drug for the indicated period of time. The cells were then trypsinized and resuspended in serum-containing, phenol red-free medium. The fluorescence of DCF was measured using an Accuri C6 flow cytometer (BD Accuri Cytometers, Ann Arbor, MI) at an excitation wavelength of 488 and emission of 533 ± 30 nm.

Total RNA was extracted from HT29 cells by utilizing TRI reagent (Sigma Aldrich, St. Louis, MO, USA) and then it was purified with the Direct-zol RNA MiniPrep Kit (Zymo Research, Irvine, CA, USA). cDNA was synthesized from purified RNA by using the iScript cDNA Synthesis Kit (Bio-Rad laboratories, Hercules, CA, USA). qRT-PCR was performed by using the iTaq Universal SYBR Green Supermix (Bio-Rad laboratories, Hercules, CA, USA) with the BioRad CFX96 thermocycler (Bio-Rad laboratories, Hercules, CA, USA). The specific primer pair sequences were 5′-CTACGTCGCCCTGGACTTCGAGC-3′ and 5′-GATGGAGCCGCCGATCCACACGG-3′ for β-actin; and 5′-CCCCCATTTGTGTGACTTCAT-3′ and 5′-GCCCGAGGCTTAGCTTTCATT-3′ for ferritin H (heavy-chain). Cycle threshold values were normalized to the internal beta-actin control. The ratio of fold change was calculated using the Pfaffl method (37).

HT29 cells were plated in 100 mm tissue culture plates for 48 h and then the cells were treated as indicated in the text. The cells were washed twice with PBS and then 100 μl triton lysis buffer (without EDTA) was added. The intracellular iron concentrations were determined using the QuantiChrom iron assay kit (BioAssay Systems; #ab83366) following the manufacturer's instructions. Briefly, iron standards of different concentrations were prepared by diluting the iron standard provided with the kit with distilled water. The diluted standards (50 μL) and sample cell lysates (50 μL) were transferred into a clear flat bottom 96-well-plate. Working reagent was prepared by mixing 20 volumes of reagent A, 1 volume reagent B and 1 volume reagent C provided with the kit. The working reagent (200 μL) was then added to the standard and sample wells. After mixing, the plate was incubated 40 min at room temperature and then the absorbance was measured at 590 nm. The iron concentration in each sample was calculated as described in the kit protocol.

All data are representative of three or more independent experiments. Data are presented as the mean ± standard error of mean (SEM). One-way analysis of variance (ANOVA) followed by Tukey's post-hoc analysis was carried out using GraphPad Prism 5 software (GraphPad Software, San Diego, California). With regard to comparisons, values were considered significant at p < 0.05.

Our previous studies showed that the DHA dimer, NSC735847, inhibited the survival of leukemia (HL-60) and prostate cancer (PC-3) cell lines (18). In the current study, we sought to determine if NSC735847 or structurally similar analogs (Figure 1A), were effective and selectively toxic toward colon cancer. Selectively toxic agents were defined as those that elicited significantly greater cytotoxicity toward human tumorigenic (HT-29 and HCT-116) than non-tumorigenic (FHC) colonocytes after 24 h of treatment (Figure 1B). In cells treated with compound (3) (NSC735847), the survival of HT-29 and HCT-116 cancer cells (IC₅₀ = 10.95 and 11.85 μM, respectively) was significantly reduced compared to non-cancerous, FHC colonocytes (IC₅₀ > 25 μM) (Figures 1B,C). Compounds (4) and (5) were more cytotoxic toward HT-29 than FHC cells, however these derivatives were inactive against HCT-116 cells. In contrast, compounds (1) and (2) were not selectively cytotoxic toward either colon cancer cell line.

To investigate whether the DHA dimers caused cell death via apoptosis, caspase 3/7 activity was measured. NSC735847 significantly increased caspase 3/7 activity in HT-29 and HCT-116 but not in FHC cells (Figure 2A). Consistent with these results, NSC735847 increased the quantity of TUNEL positive cells in both CRC cell lines however, TUNEL positive cells were absent in NSC735847-treated FHC cells (Figure 2B). Similar to the cytotoxicity study, compounds (4) and (5) preferentially induced apoptosis in HT-29 compared to FHC cells (Figure 2A) and compound (1) did not exhibit tumor selectivity. Interestingly, although compound (2) did not display selectivity in the cytotoxicity assays (Figure 1B), the data show an increase in apoptosis in HT29 compared to HCT-116 and FHC cells indicating that this compound elicits death by different mechanisms (e.g., ferroptosis, autophagy, necrosis) in the tested cell lines. Collectively, these results demonstrate that NSC735847 displays superior efficacy and selectivity compared to the structurally related analogs. As such, we investigated the mechanisms of NSC35847 cytotoxicity in our subsequent studies.

In response to ER stress, sensors including IRE1 and ATF6, and PERK are activated. PERK undergoes autophosphorylation and then it phosphorylates eIF2a, which causes a global blockade in protein synthesis to allow ER homeostasis to be reestablished (23). However, severe ER stress leads to a substantial increase in the expression of CHOP10, a protein that transactivates pro-apoptotic genes (24). Our previous studies using different cancer cell lines, showed that NSC735847 increased the expression of PERK and BiP, two ER stress regulatory proteins (18). Therefore, the current study investigated whether ER stress was required for NSC735847-induced cytotoxicity. In NSC735847 treated HT-29 cells, the phosphorylation of PERK and its substrate, eIF2α, were increased by more than 7-fold compared to vehicle-treated cells (Figure 3A). To determine if NSC735847 also activated the apoptotic ER stress pathway, we examined the expression of CHOP10 as well as the activation of caspase-3 and PARP. NSC735847 caused a marked increase in CHOP10 expression (Figure 3C) and it stimulated the cleavage of both PARP (Figure 3D) and caspase-3 (Figure 3E). Interestingly, NSC735847 did not increase the activation of PERK, eIF2a, CHOP10, PARP, or caspase 3 in non-tumorigenic, FHC colon cells (Figures 3A–E).

To evaluate whether ER stress was required for NSC735847 induced apoptosis, the ER stress pathway was blocked using the pharmacological ER stress inhibitors, GSK2606414 and salubrinal. GSK2606414 is a selective, small-molecule inhibitor of PERK that binds to its active site and inhibits ER stress induced PERK autophosphorylation (38). Pretreatment of HT29 cells with GSK2606414 decreased the phosphorylation of PERK (Figure 4A) and the expression of CHOP10 (Figure 4B). In addition, GSK2606414 suppressed NSC735847-mediated cell death (Figure 4D) and inhibited caspase 3/7 activity (Figure 4E). These findings were confirmed with salubrinal, a small-molecule inhibitor of PP1/GADD34 phosphatase activity. Salubrinal inhibits the dephosphorylation of P-eIF2α thereby promoting sustained blockade of mRNA translation which prevents ER stress from occurring (39). Consistent with the data that were obtained with GSK2626414, salubrinal blocked CHOP10 expression (Figure 4C), cell death (Figure 4D), and apoptosis (Figure 4E). Hence, the ER stress pathway is essential for the cytotoxic activity of NSC735847.

Several studies have determined that heme is required to catalyze the cleavage and subsequent activation of the endoperoxide bridge in artemisinin-type molecules (40–43). We also demonstrated that heme was needed for the cytotoxicity of NSC735847 in HL-60 and PC-3 cells (18). Therefore, the role of heme in NSC735847-treated colon cancer cells was examined by pretreating the cells with the heme oxygenase inhibitor, QC-308, or the heme biosynthesis inhibitor, succinylacetone. QC-308 is a selective inhibitor of heme oxygenase-1 (HO-1), the primary enzyme that is responsible for catabolism of heme to biliverdin, carbon monoxide, and ferrous iron (44). On the other hand, succinylacetone is an inhibitor of aminolevulinic acid dehydratase, a key enzyme in heme synthesis (45). Blockade of HO-1 activity with QC-308 increased the cytotoxicity of NSC735847 (Figure 5A). In line with this finding, the inhibition of heme synthesis with succinylacetone rescued the cells from the cytotoxic effects of NSC735847 (Figure 5B). Since our results showed that heme was essential for the anticancer activity of NSC735847, and heme has been reported to induce ER stress (46), we examined whether preventing heme synthesis with succinylacetone suppressed NSC735847-induced ER stress. As shown in Figure 5C, the inhibition of heme synthesis with succinylacetone blunted NSC735847-induced CHOP10 expression. These results indicate that heme is necessary for ER stress-induced death in NSC735847-treated cells.

The heme catabolic product, iron, is also important for the activity of artemisinin molecules (47, 48). Hence, we evaluated the impact of NSC735847 on intracellular iron levels. Cell treatment with NSC735847 caused a time-dependent increase in cellular free iron (Figure 6A). It is well-established that elevated levels of free iron cause cellular toxicity by forming ROS via Fenton reactions. To prevent this ROS-induced cytotoxicity, cells increase the expression of the iron storage protein, ferritin. Consistent with the increase in free iron, NSC735847 upregulated the expression of ferritin mRNA (Figure 6B). To determine the requirement for iron in the cytotoxicity of NSC735847, HT-29 cells treated with deferoxamine (DFO), a selective chelator that reduces intracellular iron levels (49–51). To our surprise, DFO did not prevent NSC735847-mediated cell death (Figure 6C). To confirm that the cellular iron levels were sufficiently reduced by DFO, we measured ferritin protein expression as a surrogate marker of iron levels. DFO inhibited NSC735847-mediated ferritin expression indicating that minimal quantities of free iron were present in the cells (Figure 6E). To provide additional support for this finding, iron was supplied to the cells by pretreating them with the iron-loaded transferrin molecule, holo-transferrin (HTF) (52–54). HTF did not enhance the killing effect of NSC735847, verifying that iron was dispensable for NSC735847-induced cytotoxicity (Figure 6D).

Several studies have reported that artemisinin-induced cell death is driven by the oxidative stress pathway (55). We also showed that NSC735847 increased oxidative stress in the promyelocytic leukemia cell line, HL-60 and the human prostate cancer cell line, PC3 (18). Therefore, the effect of oxidative stress on the cytotoxicity of NSC735847 in HT-29 cells was investigated by utilizing the oxidative stress probe, CM-H₂DCFDA (DCF). NSC735847 increased the intracellular oxidative stress levels as indicated by the increase in DCF fluorescence (Figure 7A). To investigate whether oxidative stress was a requirement for cell death the antioxidant, Trolox, was employed. Trolox is a water-soluble analog of vitamin E that exhibits antioxidant activity by scavenging free radicals and prohibiting lipid peroxidation (56, 57). In HT29 cells that were pretreated with Trolox, and then treated with NSC735847, a significant reduction in oxidative stress occurred (Figure 7A). However, blockade of oxidative stress with 0.25 or 0.5 mM Trolox did not rescue the cells from NSC735847-induced cytotoxicity (Figure 7B). Although 1 mM Trolox significantly increased cell survival in NSC735847-treated cells, a similar increase occurred in cells treated with 1 mM Trolox alone, indicating that the enhanced survival was caused by Trolox rather than the inhibition of NSC7358457-induced ROS (Figure 7B). Collectively, this data indicates oxidative stress is not required for the antitumor activity of NSC735847.

Most cancer chemotherapeutic regimens for advanced CRC employ multiple drugs (58). In this approach, the use of low-dose agents with different mechanisms of action is designed to increase drug activity, reduce drug resistance, and decrease overall adverse effects. One such combination, FolFOX, is composed of folinic acid (Fol; also known as leucovorin), 5-fluorouracil (F or 5-FU), and oxaliplatin (OX) and it is the standard of care treatment for advanced CRC (59). However, the OX in FolFOX often causes severe, grade 3 neurotoxicities (60) which can limit its utility. Therefore, drugs that have distinct mechanisms of action from 5-FU and Fol and exhibit selective toxicity are needed to improve the efficacy and safety of CRC therapy. Artemisinin-type drugs are well-tolerated and these agents have been reported to enhance the activity of FDA-approved antineoplastic agents including doxorubicin and gemcitabine (61). Moreover, our data show that NSC735847 is selectively toxic toward CRC cell lines (Figures 1, 2). Hence, the antitumor effect of co-administered 5-FU, Fol, and NSC735847 (FolFNSC) was examined. When administered as single agents, both NSC735847 and OX induced death in a concentration-dependent manner in HT29 and HCT116 cells although NSC735847 was lethal at substantially lower concentrations (Figures 8A,B). Next, the cytotoxic activity of FolFNSC was compared to FolFOX and at the tested concentrations, FolFNSC reduced the survival of HT29 and HCT116 cells at a lower concentration than FolFOX (Figures 8C–F). These results show that FolFNSC is an effective drug combination that exhibits greater cytotoxicity than FolFOX in colorectal cancer cells. As such, FolFNSC may be an appropriate regimen for FolFOX patients with severe OX-mediated toxicities.