Combined Parthenolide and Balsalazide Have Enhanced Antitumor Efficacy Through Blockade of NF-κB Activation

Colorectal cancer is a serious complication of ulcerative colitis and responsible for up to 15% of all deaths in patients with inflammatory bowel disease (1). Early detection and prevention strategies (such as colonoscopy, mucosal biopsies, and proctocolectomy) for colorectal cancer patients with ulcerative colitis have limitations, and therefore, there is increased interest in identifying chemopreventive agents to reduce the overall risk of colitis-associated colorectal cancer (CAC).

5-Aminosalicylate (5-ASA) is used to treat ulcerative colitis because of its ability to control and relieve inflammation. In most cases, 5-ASA is absorbed rapidly and extensively in the upper intestine and does not reach the colon (2). Sulfasalazine, the first 5-ASA–containing drug, functions by releasing an active component in the colon through the activity of azo reductase expressed by colonic bacterial. Sulfasalazine has been approved for therapeutic usage because of its ability to improve intestinal mucosal permeability (3, 4). However, the high rate of adverse effects related to sulfapyridine limits its use for patients. Balsalazide is a novel orally administered prodrug of 5-ASA in which an inert carrier molecule, 4-aminobenzoilb-alanine, is bonded to 5-ASA. Balsalazide has been shown to be more effective than sulfasalazine in the treatment of active ulcerative colitis (5, 6) and to be as effective and well tolerated as delayed release 5-ASA in the chronic treatment of ulcerative colitis (7). A recent study has shown that 5-ASA use is associated with reduced risk of colorectal cancer in patients with ulcerative colitis (8). However, the molecular mechanisms of balsalazide in the ulcerative colitis to colorectal cancer development have not been fully elucidated and the antitumor effect of balsalazide remains controversial.

NF-κB is one of the key regulators in inflammation and cancer progression (9). NF-κB activation is markedly enhanced in patients with inflammatory bowel disease and the level of activated NF-κB is significantly correlated with the severity of intestinal inflammation (10, 11). Activated NF-κB has the ability to promote the expression of various proinflammatory genes, thus strongly influencing the process of ulcerative colitis (12).

Parthenolide, a natural product, has been used for the treatment of fever and inflammatory disease. It is well known to inhibit IL-1 and TNFα-mediated NF-κB activation (13, 14). Recent studies have demonstrated that parthenolide induces apoptotic cell death through inhibition of NF-κB activation in a number of human cancers (13, 15). Several studies have also shown that parthenolide is a potent inhibitor of NF-κB activation and suppresses the expression of proinflammatory cytokines in experimental murine models (13, 16). We have also demonstrated that parthenolide inhibits phosphorylation of IκBα and NF-κB activation, resulting in initiation of apoptosis, suppression of colorectal cancer tumor growth, and CAC development (17, 18). On the basis of these observations, it can be assumed that the combination of parthenolide and balsalazide can be a promising strategy to prevent the development of ulcerative colitis to colorectal cancer, inhibiting NF-κB activation pathway.

The aim of our study was to evaluate the effects of combination therapy with parthenolide and balsalazide on the inhibition of NF-κB activation in human colorectal cancer cells and to investigate whether parthenolide and balsalazide are effective in preventing carcinogenesis from chronic colitis.

Parthenolide and z-VAD-FMK were obtained from Calbiochem. Balsalazide was provided by Chong Kun Dang Pharm. Parthenolide was dissolved in dimethylsulfoxide (DMSO; Sigma) to a concentration of 100 μmol/L and stored at −20°C in the dark. Balsalazide was dissolved in FBS-free media to a concentration of 100 mmol/L at 4°C. Annexin V-fluorescein isothiocyanate (Annexin-V FITC) and propidium iodide (PI) were purchased from Invitrogen. Hoechst 33258 was from Sigma.

HCT 116, SW480 and HT-29 (ATCC) were employed as representative human colorectal cancer cells. The cells were cultured in RPMI1640 medium supplemented with 10% FBS, 100 U penicillin, and 100 U streptomycin. The three cell lines listed above were validated by short-tandem repeat (STR) DNA fingerprinting using the Promega PowerPlex 18D System and analyzed by GeneMapper Software 5 at Cosmo Genetech Korea in January 2016.

For treatment of cells with parthenolide, the cells were subcultured in RPMI1640 medium without FBS for 24 hours. Parthenolide and balsalazide were diluted in FBS-free medium to achieve designated concentrations, and the same concentration of DMSO was applied to the cells as a control.

HCT 116, SW480, and HT-29 cells were plated at a density of 1.0 × 10⁴ cells per well in 96-well plates. Cells were treated with parthenolide and/or balsalazide for 24, 48, and 72 hours. After then, the medium was removed, 200 μL of fresh medium plus 20 μL of 3-(4, 5-dimethylthiazol-2yl)-2, 5-diphenyltetrazolium bromide (MTT, 2.5 mg dissolved in 50 μL of DMSO, Sigma) were added to each well. After incubation for 4 hours at 37°C, the culture medium containing MTT was removed and then 200 μL of DMSO was added, followed by shaking until the crystals were dissolved. Viable cells were detected by measuring the absorbance at 570 nm using a microplate reader (Molecular Devices).

Apoptotic cell death was determined by staining cells with Annexin V-FITC (Ex/Em = 488/519 nm). In brief, 1 × 10⁶ cells in a 60-mm culture dish were incubated with the designated doses of parthenolide and/or balsalazide for 24 hours. Cells were washed twice with cold PBS and then resuspended in 500 μL of a binding buffer (10 mmol/L HEPES/NaOH pH 7.4, 140 mmol/L NaCl, and 2.5 mmol/L CaCl₂) at a concentration of 1 × 10⁶ cells/mL. Annexin V-FITC (5 μL) and PI (1 μg/mL) were then added and the cells were analyzed with a FACStar flow cytometer.

Cell cycle and sub-G₁ cell analysis were determined by PI (Ex/Em = 488/617 nm) staining. In brief, 1 × 10⁶ cells were incubated in a 60-mm culture dish with the designated doses of parthenolide and/or balsalazide for 24 hours. Total cells including floating cells were then washed with PBS and fixed in 70% (v/v) ethanol. Cells were washed again with PBS, then incubated with PI (10 μg/mL) with simultaneous RNase treatment at 37°C for 30 minutes. Cellular DNA content was measured using a FACStar flow cytometer (Becton-Dickinson) and analyzed using lysis II and CellFit software (Becton-Dickinson).

Parthenolide-induced apoptosis in colon cancer cells was assessed using Hoechst 33258. The cells were treated with various concentrations of parthenolide for 24 hours, and then stained with Hoechst 33258 (1 μg/mL) at 37°C for 10 minutes. Nuclear morphology was examined under a Confocal Laser Scanning Microscope (Carl Zeiss) to identify cells undergoing apoptosis.

Cells were harvested and resolved in a lysis buffer [20 mmol/L Tris-HCl (pH 7.5), 137 mmol/L NaCl, 10% glycerol (v/v), 1% Triton X-100 (v/v), 1 mmol/L Na₃VO₄, 1 mmol/L phenylmethylsulphonylfluoride, and protease inhibitor cocktail]. After centrifugation at 16,000 × g for 15 minutes, the supernatants were used as cytoplasmic extracts. To extract the nuclear fraction, cells were resuspended in 150 μL of buffer A [10 mmol/L HEPES (pH 7.9), 1.5 mmol/L MgCl₂, 10 mmol/L KCl, 0.5 mmol/L dithiotreitol, 0.5 mmol/L phenylmethylsulphonylfluoride, 0.4% Nonidet P-40 (v/v) and protease inhibitor cocktail] for 20 minutes on ice and then centrifuged at 2,300 × g for 5 minutes. The resulting pellets were resolved in 100 μL of buffer C [20 mmol/L HEPES (pH 7.9), 420 mmol/L NaCl, 1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, 0.5 mmol/L dithiotreitol, 0.5 mmol/L phenylmethylsulphonylfluoride, and protease inhibitor cocktail] for 30 minutes on ice. After centrifugation at 16,000 × g for 15 minutes, the supernatants were used as nuclear extracts.

NF-κB activity was measured by EMSA. In brief, prior to stimulation, cells were preincubated with the indicated concentrations of parthenolide and/or balsalazide at 37°C for 24 hours. In following, cells were stimulated with TNFα (10 ng/mL), harvested by centrifugation, washed twice with ice-cold 1× PBS, and then nuclear extracts were prepared. A double-stranded oligonucleotide for NF-κB (Promega) was end-labeled with [γ-32P] ATP and purified with a G-25 spin column (Boehringer Mannheim). Nuclear extracts were incubated for 20 minutes at room temperature with a gel shift binding buffer [5% glycerol, 1 mmol/L MgCl₂, 0.5 mmol/L EDTA, 0.5 mmol/L DTT, 50 mmol/L NaCl, 10 mmol/L Tris–HCl, pH 7.5, 50 μg/mL poly(dI-dC) poly(dI-dC)] and 32P-labeled oligonucleotide. The DNA–protein complex formed was separated on 4% native polyacrylamide gel, and the gel was transferred to Whatman 3 MM paper, dried, and exposed to X-ray film, captured and analyzed by the BAS-2500 Image Analyzer (Fuji Film).

The protein concentration in cell lysates or tissue lysates was measured using a Protein Quantification kit from Bio-Rad. Fifty micrograms of protein or 30 μg of nuclear extract protein per lane was loaded onto a SDS-polyacrylamide gel. After transferring and blocking, the polyvinylidene difluoride (PVDF) membrane was probed with various antibodies (anti-p-IkB-α, anti-p65, anti-phospho-p65, anti-caspase 8, anti-caspase-9, anti-caspase-3, anti-p53, anti-PARP, anti-Bcl-xL, anti-Bcl-2, anti-VEGF, anti-MMP9, anti-cyclin D1 anti-cFLIP, anti-COX2 anti-GLUT1, anti-Hexokinase II, anti-actin and anti-Lamin B antibody). The binding of antibody to antigen was detected using enhanced ECL prime (GE Healthcare), captured and analyzed by the Las-3000 luminescent Image Analyzer (Fuji Film).

Sixty specific pathogen–free mice (Balb/C female mice, 6 week) were purchased from Orient. Mice were given ad libitum access to water and standard rodent food until they reached the desired weight (18–20 g). Mice were maintained on a 12/12 hours light/dark cycle under specific pathogen-free conditions. All procedures using the mice were reviewed and approved by Chonbuk National University Animal Care and Use Committee (approval no: CBNU 2015-0027). Twelve in each group were randomly assigned after they were weighed. Mice were intraperitoneally injected with 7.4 mg/kg body weight of azoxymethane (AOM) dissolved in physiologic saline. Seven days later, 3% dextran sulfate sodium (DSS) was given in the drinking water for 7 days, followed by 14 days of regular water. This cycle was repeated three times. Parthenolide (2 mg/kg) suspended in saline was administered by intraperitoneal injection three times a week at break time. Balsalazide (200 mg/kg) suspended in saline was administered daily by oral gavage at break time under anesthesia using isoflurane.

After finishing CAC-related procedures, all mice were fasted for 6 hours to minimize plasma glucose concentration before examination. The mice were anesthetized under 2% isoflurane and ¹⁸F-FDG (3.7 × 10⁶ Bq) was injected via the tail vein with an insulin syringe. After 30 minutes, mice were sacrificed by cervical dislocation and the entire colon was removed from the cecum to the anus. The colon was then opened longitudinally and washed twice with saline. The colons of each group were placed on the scanner bed. Micro-PET/CT images were acquired using a FLEX X-PET/X-O small-animal imaging instrument (X-PET/CT System, GE Healthcare), which combines a PET scanner and a multislice helical CT scanner. CT scans were performed with the following scanning parameters: 75 kVp and 0.25 mA. The CT images were acquired with 256 projections over 2 minutes. After acquisition of PET/CT scans, for histologic analysis, the distal colons were fixed in 10% neutral buffered formalin for 24 hours and transferred to 70% ethanol for subsequent paraffin embedding. For Western blotting, total protein was extracted using RIPA buffer (Invitrogen) and homogenizer according to the manufacturer's instructions and quantified.

Five-micron tissue sections were stained with hematoxylin and eosin, and histologic analysis was performed by a pathologist in a double-blind manner. The inflammation scores of mucosal inflammation were determined as follows (19): 0, normal morphology; 1, focal inflammatory cell infiltrate around the crypt base; 2, diffuse infiltration of inflammatory cells around the crypts or erosion/destruction of the lower one-third of the glands; 3, erosion/destruction of the lower two-thirds of the glands or loss of all the glands. Invasion depth was scored as follows (20): 0 = no invasion, 1 = invasion through mucosa, 2 = invasion through submucosa, 3 = full invasion through muscularis and into serosa. All histologic analyses were performed by a pathologist who was blinded to treatment status.

The data are presented as the mean ± SE of at least three independent experiments performed in duplicate. Representative blots are shown. All the data were entered into Microsoft Excel 5.0, and SPSS Software was used to perform two-tailed t tests or the ANOVA, where appropriate. A P value <0.05 was considered significant.

Human colorectal cancer cell lines, HCT116, SW480, and HT-29 cells were treated with various concentrations of balsalazide for 24 hours. Cell proliferation was significantly affected by treatment with 20 mmol/L balsalazide in all cell lines (Fig. 1A). Next, to examine whether parthenolide promotes balsalazide-induced death of human colorectal cancer cells, the cells were treated with parthenolide and balsalazide for 24, 48, and 72 hours. After cotreatment with 20 mmol/L balsalazide and 5 or 10 μmol/L parthenolide, the viability of cells was dramatically reduced in a dose-dependent manner of parthenolide, showing 3.4-fold, 2.9-fold, and 2.2-fold decrease for 24-hour treatment in HCT116, SW480, and HT-29 cells with 10 μmol/L parthenolide, respectively (Fig. 1B). And the cotreatment with 20 mmol/L balsalazide and 10 μmol/L parthenolide for 48 and 72 hours also showed synergistic reduction of viability (4.5-fold for 48 hours and 12-fold for 72 hours in HCT 116 cells; 2.8-fold for 48 hours and 3.3-fold for 72 hours in HT-29 cells; 4.4-fold for 48 hours and 11-fold for 72 hours in SW480 cells). For the subsequent experiments, we used the concentration of 5 μmol/L parthenolide and 20 mmol/L balsalazide for the 24-hour treatment resulted in approximately 50% inhibition of viability. HCT 116 cells showed more remarkable changes than other cells in viability with combined treatment. Moreover, population doubling time of HCT116 cells is shorter than HT-29 and SW480 cells (population doubling time: 21 hours for HCT116 cells, 23 hours for HT-29 cells, and 38 hours for SW480 cells), so it is much easier to handle for in vitro experiment. For that reason, HCT116 cells were used as a representative colorectal cancer cells for the subsequent experiments.

To confirm the above observations, Annexin-V/PI analysis was performed using FACScan (Fig. 2A). We found that treatment of HCT116 cells with a single agent induced low-level early and late apoptotic cell death [parthenolide alone (early apoptosis: 4.99%, late apoptosis: 4.98%), balsalazide alone (early apoptosis: 2.28%, late apoptosis: 3.56%)]. In agreement with cell growth inhibition, cotreatment with parthenolide and balsalazide dramatically increased the apoptotic cell death (early apoptosis: 17.11%, late apoptosis: 14.24%) by 5-fold compared with treatment with balsalazide alone, indicating that parthenolide promotes balsalazide-induced apoptosis in HCT116 cells.

We also evaluated cell-cycle modifications induced by parthenolide and balsalazide in HCT116 cells (Fig. 2B). Twenty-four hours after incubation with single or dual agents, cells were analyzed by PI staining and flow cytometric analysis. Treatment with parthenolide and/or balsalazide resulted in the presence of a sub-G₁ population, suggestive of apoptotic cell death. Peaks accounting for 9.91% and 8.12% of the overall cell population were detectable in HCT116 cells treated with parthenolide and balsalazide, respectively. A much higher sub-G₁ population (37.82%) was observed with cotreatment of parthenolide and balsalazide than with single treatment, indicating that the combination of the two agents substantially promotes apoptosis of HCT116 cells.

To understand the mechanism of cell death induced by combination treatment, apoptotic nuclear morphology was evaluated after Hoechst 33258 staining (Fig. 2C). In the control group, HCT116 cells were regular in morphology and grew fully in patches. In the parthenolide- or balsalazide-treated groups, only a few cells exhibited nuclear condensation and fragmentation. However, cotreated cells showed apoptotic characteristics, such as cell shrinkage, nuclear condensation, and fragmentation.

Inflammatory cytokines are required to activate NF-κB; therefore, TNFα was employed as a representative of NF-κB stimuli. The translocation of NF-κB to the nucleus is preceded by phosphorylation, ubiquitination, and proteolytic degradation of IκB-α (21). To determine whether inhibition of TNFα-induced NF-κB activation by parthenolide and/or balsalazide is caused by inhibition of IκB-α degradation, HCT116 cells were treated with parthenolide and/or balsalazide for 24 hours and then with TNFα (20 ng/mL) for 30 minutes. Phosphorylation and degradation of IκB-α in cytoplasmic extracts of the cells were analyzed by Western blotting. Neither phosphorylation of IκB-α nor inhibition of IκB-α phosphorylation was observed with a single-agent treatment. However, cells cotreated with parthenolide and balsalazide showed a significant inhibition in the TNFα-induced phosphorylation of IκB-α (Fig. 3A).

As the degradation of IκB-α is known to cause nuclear translocation of the p65 subunit of NF-κB, we next examined whether parthenolide and/or balsalazide modulates TNFα-induced nuclear translocation of p65. Nuclear and cytosolic extracts of HCT116 cells treated with parthenolide and/or balsalazide were analyzed by Western blotting. As shown in Fig. 3B, cotreatment of HCT116 cells with parthenolide and balsalazide resulted in a decrease in nuclear p65 level, whereas the cytosolic p65 level was increased. The results show that cotreatment with parthenolide and balsalazide also blocks TNFα-induced activation of p65.

To determine the effect of parthenolide and/or balsalazide on the DNA-binding activities of NF-κB, we performed electrophoretic mobility shift assay. TNFα-induced DNA-binding activity of NF-κB was significantly suppressed by parthenolide treatment and the DNA-binding activity of NF-κB was almost completely blocked by cotreatment with parthenolide and balsalazide (Fig. 3C). These results demonstrate that parthenolide has a greater inhibitory effect on TNFα-induced DNA-binding activity of NF-κB than balsalazide does.

Phosphorylation of p65 has been shown to regulate its transcriptional activity (22), so we examined whether parthenolide and/or balsalazide affected the post-translational modification of p65. Cells were treated with parthenolide and/or balsalazide in the presence of TNFα, and phosphorylation of p65 at serine 536 was evaluated in nuclear extracts using a 536 phosphoserine-specific p65 antibody. As shown in Fig. 3D, parthenolide inhibited phosphorylation of p65 at serine 536 in a dose-dependent manner. However, balsalazide did not prevent TNFα-mediated phosphorylation of p65 at serine 536. These results strongly suggest that parthenolide enhances the effect of balsalazide on inhibition of NF-κB activation.

Most apoptotic cell death is initiated by activation of caspases (23). To delineate the mechanisms by which treatment with parthenolide potentiates balsalazide-induced apoptosis, we next analyzed the protein expression levels of caspases by Western blotting. Treatment of HCT116 cells with parthenolide decreased the levels of full-length caspases compared with no treatment, whereas balsalazide did not show any change in the levels of caspases (Fig. 4A). However, Western blot analysis revealed that levels of full-length caspases in the cells cotreated with parthenolide and balsalazide were significantly decreased compared with the levels in the cells treated with parthenolide or balsalazide alone. Furthermore, the decrease in caspase levels in the cotreated cells was significantly blocked by pretreatment with the general caspase inhibitor Z-VAD-FMK. These data indicate that parthenolide and balsalazide activate caspases and that cotreatment with parthenolide and balsalazide promotes activation of caspases.

The activation of caspase-3 leads to the cleavage of its downstream molecular targets, PARP and p53, hallmarks of apoptosis (24). The level of p53 was significantly increased by cotreatment with parthenolide and balsalazide compared with the level in cells treated with parthenolide or balsalazide alone (Fig. 4B, first panel). The cleavage of PARP was significantly increased by treatment with parthenolide and cotreatment with balsalazide, indicating that parthenolide potentiates the apoptotic effects of balsalazide in HCT116 cells (Fig. 4B, second panel).

To investigate whether cotreatment with parthenolide and balsalazide inhibits the expression of NF-κB target gene products involved in aggressive cancers and inflammation, we analyzed expression of Bcl-xL, Bcl-2, VEGF, MMP-9, Cyclin D1, cFLIP, and COX-2 by Western blotting (Fig. 5). The expression of antiapoptotic proteins, Bcl-2, Bcl-xL, and cFLIP, was markedly reduced by cotreatment for 24 hours. Moreover, the proliferation and cell progression–related proteins, COX-2 and Cyclin D1, were suppressed. Invasion- and angiogenesis-related gene products, MMP-9 and VEGF, were also dramatically inhibited by cotreatment with parthenolide and balsalazide.

Azoxymethane (AOM) is a procarcinogen that causes formation of O⁶-methylguanine upon metabolic activation. AOM induces tumors in the distal colon of rodents and is commonly used to elicit colorectal cancer in experimental animals (25). To test the physiologic relevance of the parthenolide/balsalazide–mediated suppression of carcinogenesis from chronic inflammation in vivo, we conducted an in vivo study using a murine model of AOM/DSS–induced colon cancer. First, we compared body weights between AOM/DSS and administration of parthenolide- and/or balsalazide-treated mice (Fig. 6A). Cotreated mice significantly recovered the weight loss after the second break time. At the start of the third break time, the body weight was 16.95 ± 0.201 g for AOM/DSS mice and 20.304 ± 0.359 g for parthenolide/balsalazide cotreated mice, indicating that cotreatment with parthenolide and balsalazide prevents the loss of body weight caused by carcinogenesis.

Next, we evaluated colon length and the characteristics of colon carcinogenesis (Fig. 6B). Numerous nodular, polypoid, or caterpillar-like tumors were observed in the middle and distal colon of AOM/DSS mice. Shortening of the colon, which is another characteristic of colon carcinogenesis, was significantly improved in the cotreated mice.

We also evaluated the severity of acute murine colitis by blinded histologic injury scoring from the cecum to the distal colon (Fig. 6C). The histologic examination of colons from AOM/DSS–induced colon cancer mice showed inflammatory lesions that included total impairment of the glandular structure, mucosal ulceration, crypt damage, and the infiltration of inflammatory cells. In contrast, cotreatment with parthenolide and balsalazide markedly reduced the impairment of the glandular architecture and infiltration of inflammatory cells. Histologic grading (inflammation score and invasion depth) showed that the cotreatment significantly attenuated the overall score compared with the AOM/DSS mice.

We then assessed the metabolic activity of carcinogenesis by ¹⁸F-FDG PET/CT imaging. As shown Fig. 6D, analysis of the PET/CT images revealed clearly elevated ¹⁸F-FDG uptake in the colon of AOM/DSS–treated mice with the most intensive area of ¹⁸F-FDG uptake in the distal colon. Compared with AOM/DSS–treated mice, ¹⁸F-FDG uptake in the colon of parthenolide- or balsalazide-treated mice exhibited somewhat less uptake. Furthermore, we confirmed that ¹⁸F-FDG uptake in the colon was significantly decreased by cotreatment with parthenolide and balsalazide.

We further investigated whether protein levels induced by the NF-κB pathway in colonic tissues are affected after parthenolide and balsalazide cotreatment (Fig. 6E). Consistent with our in vitro results, rapid decline of phosphorylation of IκB-α and p65 was observed in cotreated mice colon. However, the level of p65 did not show any significant change after cotreatment with parthenolide and/or balsalazide. The expression of the representative NF-κB target gene product, VEGF, was markedly suppressed by cotreatment compared with the level observed after AOM/DSS treatment alone. Glucose transporters (GLUT) in the cell membranes and the activity of intracellular enzymes such as hexokinases (HK) mediate the intracellular accumulation of ¹⁸F-FDG (26). Therefore, levels of GLUT1 and HKII in the colonic tissues were determined by Western blotting. High expression of GLUT1 and HKІІ was observed in the colon of AOM/DSS–treated mice compared with in the colon of normal control mice. The increased levels of both proteins in the colon of mice were dramatically suppressed by treatment with parthenolide and/or balsalazide, and the expression of the proteins was well correlated with the ¹⁸F-FDG PET/CT image.

In contrast to sporadic colorectal cancers that are mainly caused by genetic factors, CAC is mostly associated with inflammatory bowel disease (IBD), and, therefore, it is understood that chronic inflammation of the colonic epithelium can induce CAC. Sporadic colorectal cancers, accounting for 65%–85% of all colorectal cancers, develop through the “adenoma-cancer” progression, while CAC develops through “inflammation-dysplasia-cancer” progression (27). Therefore, we suggest that modulation of inflammation might be an effective approach for preventing CAC cancers. In 1994, the risk of CAC among patients with ulcerative colitis was reduced through treatment with sulfasalazine, a 5-ASA derivative (28). It was also demonstrated that 5-ASA and its derivatives (i) inhibit Wnt/β-catenin signaling pathways; (ii) activate PPAR-γ; (iii) suppress NF-κB activity; and (iv) inhibit oxidative damage and DNA mutations. However, recent studies have demonstrated that there is not a statistically significant inhibitory effect of 5-ASA in colorectal cancer, raising controversy over the potency of 5-ASA and its derivatives for prevention of colonic carcinogenesis from ulcerative colitis. To our knowledge, there has been no investigation on the efficacy and underlying molecular mechanisms of balsalazide and parthenolide combination chemotherapy in vivo or in vitro.

In our previous studies, we found that parthenolide can be a potential chemopreventive and therapeutic agent for colorectal cancer and CAC (17, 18). We have also shown that parthenolide has potential to be applied in combination therapy for colorectal cancer treatment: combination with 5-fluorouracil (5-FU) can overcome 5-FU resistance in human colorectal cancer, and parthenolide sensitizes cancer cells to the TNF-related apoptosis-inducing ligand (TRAIL) in TRAIL-resistant colorectal cancer cells (29–31). These findings indicate that parthenolide may increase the efficacy of balsalazide in preventing carcinogenesis caused by chronic inflammation. The results in the current study showed that a combination treatment with parthenolide and balsalazide significantly inhibited NF-κB activation and was linked to regulation of apoptosis-related molecules and NF-κB target genes. Moreover, the combination of parthenolide and balsalazide improved AOM/DSS–induced CAC in mice clinically and histologically. These findings suggest that combination of parthenolide and balsalazide can be a useful therapeutic approach to treat CAC.

In this study, we used human colorectal cancer epithelial cells for in vitro study. However, ulcerative colitis is one of the T-cell–mediated IBD and excessive TNFα production in macrophages by T cells (32). Therefore, you may argue that human colorectal cancer cells are not reasonable to determine the effect of this therapeutic strategy in CAC. However, intestinal epithelial cells (IEC) exert their immunologic functions by processing and presenting antigens to T cells, so they can function as sensors of mucosal injury and actively participate in the mucosal response to intestinal inflammation. Chronic damage of intestinal epithelial barrier leads to development of CAC (33). By the reason, we selected commercial human colorectal cancer epithelial cells for in vitro experiments. Moreover, CAC in mice develops tumorigenesis from intestinal epithelium because of drinking water included DSS. Therefore, for this study, we consider that epithelial colorectal cancer cells for in vitro is partly correlated with in vivo observations.

A central molecule in inflammation, NF-κB, has been shown to be activated in all cells, where it regulates expression of diverse target genes that promote cell proliferation, regulate immune and inflammatory response, and contribute to pathogenesis of various diseases, including cancers (34). In chronic inflammation, NF-κB has a specific role connecting inflammation to cancer. Over 15% of malignancies are initiated by chronic inflammatory disease; for example, squamous cell carcinoma by skin inflammation, viral hepatitis by liver cancer, and IBD by colorectal cancer (35). Studies have suggested that constitutive NF-κB activation in IBDs increases the risk of colorectal cancer (10, 12, 36). A number of studies have demonstrated that parthenolide is a strong inhibitor of NF-κB activation and can effectively inhibit the expression of proinflammatory cytokines in cultured cells and experimental animal models (13, 37–42). In 1994, Zao and colleagues found that administration of parthenolide significantly reduces the severity of DSS-induced colitis through inhibition of phosphorylation of IκB-α and p65 proteins, resulting in reduction in expression of inflammatory mediators (16). Interestingly, our current study showed that treatment with parthenolide alone suppresses TNFα-induced phosphorylation of IκB-α, translocation of p65 and phosphorylation of p65 at serine 536. Balsalazide similarly downregulated phosphorylation of IκB-α and translocation of p65; however, it did not show regulation of p65 phosphorylation at serine 536, even at the maximum concentration. Moreover, the combination of parthenolide and balsalazide significantly suppressed TNFα-induced DNA-binding activity of NF-κB compared with treatment with balsalazide alone. Notably, as compared with simple translocation of p65 without phosphorylation, phosphorylation of p65 triggers strong transcriptional activity of NF-κB producing significant modulation of target gene transcription (43). To summarize the molecular mechanism of combined treatment, parthenolide potentiates balsalazide-induced inhibition of transcriptional activity of NF-κB by inhibiting of phosphorylation of p65 at serine 536.

The relation between NF-κB and apoptosis was elucidated recently by studies demonstrating that inhibition of NF-κB activation, either by the IκB super repressor or in Rel A (p65) knockout cells, results in increased apoptosis (43, 44). In particular, NF-кB is responsible for the expression of genes involved in proliferation, tumor survival, mitotic cell cycle, tumor cell invasion, and angiogenesis. Our results revealed that the combination treatment with parthenolide and balsalazide significantly downregulated NF-κB–regulated gene products such as Bcl-2, Bcl-xL, VEGF, MMP9, Cyclin D1, cFLIP, and COX2 in human colorectal cancer cells through the inhibition of transcriptional activity of NF-κB.

AOM is a procarcinogen and is metabolically activated to a potent alkylating agent forming O⁶-methyl-guanine (25). Its oncogenic potential is markedly augmented in the setting of chronic inflammation, such as that induced by repeated cycles of DSS treatment (45). The power of this model has recently been demonstrated in deciphering the epithelial versus myeloid cell contribution of IKKβ to polyp formation in the setting of inflammation (46). NF-κB activation is normally triggered in response to microbial and viral infections and proinflammatory cytokines, all of which activate the IκB kinase (IKK) complex (21). IKK phosphorylates NF-κB–bound IκBs and targets them for ubiquitin-dependent degradation, allowing liberated NF-κB dimers to enter the nucleus (47). Our results in the current study showed that phosphorylation of IκB-α and p65 in colonic tissues was dramatically reduced by cotreatment with parthenolide and/or balsalazide compared with AOM/DSS-induced CAC mice. However, p65 protein level did not change after treatment with parthenolide and/or balsalazide in colonic tissues, suggesting that NF-κB activation leads only to translocation of p65 without an increase in total protein level.

A potential limitation of our study is that we used histologic observation related to inflammation rather than carcinogenesis. Histologic scoring system such as epithelial damage and infiltration of immune cells in the mucosa/submucosa/muscularis/serosa are typical methods which evaluate severity of ulcerative colitis in animal models. However, Cooper and colleagues reported that histologic level of inflammation is significantly related to dysplasia and/or cancer showing that CAC has higher inflammation score than those of without cancer (48). Thus, our histologic analysis partially supports that combined treatment with parthenolide and balsalazide prevents carcinogenesis from chronic colitis.

Clinically, ¹⁸F-FDG PET/CT is a noninvasive imaging modality widely used for detection, staging, and follow-up of tumors, infections and inflammation (49). Spier and colleagues have also presented promising results for the use of ¹⁸F-FDG PET/CT in the assessment of inflammation in patients with IBD (50). Moreover, Hindryckx and colleagues have demonstrated that ¹⁸F-FDG PET/CT can detect DSS-induced intestinal inflammation and imaging findings are correlated strongly with histologic examination (51). In accordance with these observations, we were able to correlate ¹⁸F-FDG PET/CT images with histologic alterations in an experimental CAC model.

In conclusion, we have shown that combination of parthenolide and balsalazide exhibits synergistic suppression of NF-κB and NF-κB–regulated gene products that are associated with apoptosis, proliferation, invasion, angiogenesis, and inflammation. Moreover, our data suggest a molecular mechanism by which parthenolide enhances the effect of balsalazide on inhibition of phosphorylation of p65 at serine 536, which is critical for the transcriptional activity of NF-κB. Therefore, administration of parthenolide and balsalazide significantly inhibits the inflammation–carcinoma sequence and can be a crucial regulator of CAC.

No potential conflicts of interest were disclosed.

Conception and design: S.L. Kim, S.Y. Seo, S.T. Lee, S.W. Kim

Development of methodology: Y.R. Park, Y.N. Kim

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.H. Kim, E.-M. Kim, H.-J. Jeong, S.Y. Seo, S.T. Lee

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y.N. Kim, S.Y. Seo, S.W. Kim

Writing, review, and/or revision of the manuscript: S.L. Kim, Y.N. Kim, S.Y. Seo, S.T. Lee

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y.-C. Liu

Study supervision: I.H. Kim, S.O. Lee

Other (specify): S.H. Kim

The authors thank Professor Mie-Jae Im from Chonbuk National University Medical School for proofreading and contributions.

This work was supported by Fund of Biomedical Research Institute, Chonbuk National University Hospital.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.