Background
Colorectal cancer (CRC) has consistently ranked the top 5 most deadly malignancies in the developed countries due to its highly metastatic potential and the resistance against treatments. Despite the advance in the development of chemo- and targeted therapeutic agents, the mortality and the recurrence rates remain high in the advanced stage patients [
1]. Accumulating evidence demonstrates that these currently available agents select for and/or enrich the cancer stem-like cells [
2] leading to the eventual treatment failure. Thus, the identification of alternative or novel agents is urgently required.
Regorafenib is an approved multiple kinase inhibitor for the patients with metastatic colon cancer and failed to respond to currently available chemotherapeutic agent [
3]. Regorafenib has been shown to inhibit tumorigenesis both in vitro and in vivo via downregulating RAF/MEK/ERK signaling in different cancer types including colon, breast and renal cell carcinoma [
4]. Interestingly,
KRAS,
BRAF or
PIK3CA status was not found to be predictive for the positive regorafenib responders in the clinical settings [
5]. Thus, the precise patient population who may be benefited from regorafenib and the underlying mechanism of actions still require further investigation.
miR-34a is highly expressed in normal tissues and commonly repressed in carcinoma such as prostate cancer, breast cancer, ovarian cancer and non-small cell lung cancer. Moreover, previous studies demonstrated that ectopic expression of miR-34a suppressed cell proliferation, migration and invasion in various cancer cells, which could also contribute to drug resistance in breast cancer by targeting a variety of oncogenes [
6]. However, the role and mechanism of miR-34a in the regulation of colon cancer stem-like cells is far from being completely elucidated at present.
It was our goal to investigate the potential efficacy of regorafenib on cancer stem-like cells in this study. We first established two colon cancer cell lines resistant to fluorouracil by acclimatization. These 5-FU resistant colon cancer cells exhibited enhanced tumorigenic phenotypes including CD44+ and side-population cells, increased colony and tumor sphere formation. These observations were associated with the elevated expression of stemness markers such as Nothc1, WNT1 and β-catenin. Importantly, we showed that the treatment of regorafenib in these 5-FU resistant cancer cells suppressed the aforementioned tumorigenic phenotypes and stemness markers. The combination of regorafenib and 5-FU synergistically suppressed colon cancer viability both in vitro and in vivo. Finally, regorafenib treatment was mechanistically associated with the increased level of a tumor suppressor, miR-34a. Thus, this study provides novel and important mechanistic explanations underlying regorafenib’s ability to treat chemo-refractory colon cancer and the combination of regorafenib and 5-FU regimen may provide an improved treatment efficacy.
Methods
Chemicals and reagents
Regorafenib (BAY 73–4506, catalog No. S1178) and Fluorouracil (5-FU, catalog No. S1209) were purchased from SelleckChem. The primary and secondary antibodies used for western blotting and immunohistochemical experiments were all purchased from Cell Signaling Technology unless otherwise specified.
Generation of 5-FU resistant cell lines
Human colon cancer cell lines, HCT-116 and DLD-1 were purchased from the American Type Culture Collection (ATCC, Manassas, VA) and the cells were maintained under the conditions accordingly. To generate 5-FU resistant cells, both HCT-116 and DLD-1 cells were initially exposed to 5-FU (5 μM 72 h) and the survived cells were subsequently passaged and maintained under the same culture conditions for at least 20 more passages. The resultant 5-FU-acclimatized cells were termed HCT-116R and DLD-1R (R as 5-FU resistant line).
Side population (SP) and tumor sphere assays
We performed the side population (SP) assay to identify and quantify the cancer stem-like and/or drug resistant cancer cells. SP cells are defined as a sub-population of cells with high expression of ATP-binding cassette transporters (ABCG2) and the ability to exclude Hoechst 33,342 nuclear dye [
7]. We used FACSAria™ technology platform to determine and compare the SP cells in HCT-116, HCT-116R, DLD-1 and DLD-1R cells. Cells were first labeled with Hoechst 33342 dye (2.5 μg/mL) for 30 min at 37 °C. The control cells were treated with verapamil (50 μM, Sigma-Aldrich). Propidium iodine (PI) 1 μg/mL served to identify dead cells. After identification and cell sorting, SP cells were cultured under stem cell conditions: serum-free of HEScGRO medium, N2 supplement (Invitrogen, Carlsbad, CA), 10 ng/mL human recombinant bFGF (Invitrogen), and 10 ng/mL EGF (Invitrogen) in ultra-low attachment CoStar plates (Corning, NY). Tumor spheres were measured and those ≥ 200 μm were counted as a tumor sphere forming unit. The data calculated for the number and size of the tumor spheres is the average of three independent experiments.
Cell viability test
Sulforhodamine B (SRB) dye (Sigma-Aldrich, Chemie GmbH, Munich, Germany) was used to test the effects of selective inhibitors on cell growth and viability of SP cells. The regorafenib were dissolved in dimethyl sulfoxide (DMSO) before diluting with growth medium to a final DMSO concentration of 0.05%. The HCT-116R and DLD-1R cells were seeded into 96 well plates in growth medium at 3000 cells/well. After 24 h the medium was replaced with fresh growth medium containing the regorafenib. The cells were incubated for another 48 h. The cells were fixed with trichloroacetic acid (TCA) by gently adding 50 μL TCA (50%) to each well to a final TCA concentration of 10% with subsequent incubation for 1 h at 4 °C. The plates were then washed 5 times with tap water and air dried. The dried plates were stained with 100 μL of 0.4% (w/v) SRB prepared in 1% (v/v) acetic acid for 10 min at room temperature. The plates were rinsed quickly 4 times with 1% acetic acid to remove unbound dye, and then air dried until no moisture was visible. The bound dye was solubilized in 20 mM Tris-base (100 μL/well) for 5 min on a shaker. Optical densities were read on a microplate reader (Molecular Devices, Sunnyvale, CA) at 562 nm.
SDS-PAGE and western blotting
HCT-116R and DLD-1R cells were lysed and prepared using ReadyPrep Protein Extraction Kit (Bio-Rad, Hercules, CA) according to the vendor’s instructions. Total cell lysates (50 μg) were separated by a 10% SDS-PAGE and transferred to a PVDF (polyvinylidene fluoride) membrane via BioRad Protean III system. The blots were then blocked with 5% skim milk in PBST and incubated with primary antibodies overnight at 4 °C. All primary antibodies: β-catenin, Nanog, p65/RelA (NF-kB), Bax, p-mTOR, Wnt1, STAT3, Notch1, c-Myc, vimentin and β-actin (as loading control) were purchased from Cell Signaling unless otherwise specified. The membranes were then incubated with peroxidase-conjugated secondary antibody at room temperature for at least 1 h. Blots were washed 3 times with PBST. Signals were then detected using enhanced chemiluminescence kit and visualized using the BioSpectrum Imaging System (UVP, Upland, CA).
In vivo studies
All animal studies were performed strictly under the University’s animal experiment protocol. In the comparative tumorigenesis experiment, DLD-1 and DLD-1R cells were injected into the flank for NOD/SCID mice (1 × 106 cells/mouse; N = 5/group). In the drug treatment experiments, DLD-1R cells (1 × 106 cells/mouse; N = 5/group) were injected subcutaneously into the flank of NOD/SDCI mice. The treatments started when the tumor size reached approximately 100 mm3 determined by a caliper. Mice were randomly divided into 4 groups: Control (sham injection), 5-FU alone (30 mg/kg, 2 times/week), regorafenib only (10 mg/kg, 5 times/week), 5-FU plus regorafenib combination (5-FU, 30 mg/kg, 2 times/week and Regorafenib, 10 mg/kg, 5 times/week, respectively). Both drugs were given intraperitoneally. The change in tumor burden was expressed as a ratio (fold change in mm3) as compared to the initial tumor volume. Mice were humanely sacrificed post experiment and tumor samples were collected for further analyses.
Immunohistochemistry
Tumor tissues were fixed in 10% formalin and embedded in paraffin. Serial sections of the embedded specimens were deparaffinized and then rehydrated in a graduated fashion and stained with hematoxylin and eosin (H&E). For immunohistochemical staining, the deparaffinized slides were subjected to antigen retrieval and probed with anti-Wnt1 (1:100), anti-β-catenin (1:200), or anti-Bax (1:100) antibodies. Slides were washed and incubated with biotinylated link universal antiserum, followed by Horseradish Peroxidase Streptavidin (HRP Streptavidin). The slides were rinsed, and color was developed using 3,3-diaminobenzidine hydrochloride as a chromogen. Finally, sections were rinsed in distilled water, counterstained with Mayer’s hematoxylin, and mounted with DPX mounting medium for evaluation. Pictures were captured with a Photometrics CoolSnap CF color camera (Nikon).
Statistical analysis
Each experiment was performed in triplicates. The results were expressed as means ± SD. The significant difference between control and experimental groups was analyzed using t-test. (*, P<0.05; **, P<0.01).
Discussion
Regorafenib is a multiple kinase inhibitor which has been attributed to the survival benefits in metastatic colorectal cancer which has failed to respond to all other therapeutics [
8]. Regorafenib has been established as a specific inhibitor for receptor kinases including VEGFRs, FGFR and PDGFR [
9]. However, resistance against chemotherapeutics represents one of the key features of the presence of cancer stem-like cells. Based on this rationale, we set out to investigate the potential inhibitory effect of Regorafenib on colon cancer stem cells. We first established 5-FU resistant colon cancer cell lines by acclimatizing them under constant, low-dosage of 5-FU. We believe that these cells mimic the residual colon cancer cells post chemotherapy. In support, both HCT-116R and DLD-1R cells showed a significantly increased resistance as compared to their naïve parental counterparts. Consistently, both cells also demonstrated several cancer stem-like features including the increased percentage of CD44+ and side-population cells as well as the enhanced ability to generate tumor spheres. In support, studies have demonstrated that CD44+ cancer stem-like cells obtained from different cancer types, are found to be resistant against drug treatment and with increased expression of powerful oncogenes such as c-Myc and Wnt/β-catenin [
10‐
12] and ones identified in this study, NF-κB, Notch1 and decreased pro-apoptotic marker Bax [
13,
14]. More importantly, our 5-FU acclimatized DLD-1R cells showed enhanced tumorigenic ability as compared to their parental counterparts. This is in agreement with the notion where chemotherapy induced drug resistance observed in different cancer types including colon cancer [
2,
15,
16]. Even with the targeted therapeutic agent such as cetuximab or panitumumab, resistance has been reported in patients with metastatic colon cancer [
17] and has been implicated as a selection process for cancer stem-like cells [
18].
In this study, DLD-1R and HCT-116R cells appeared to be sensitive towards regorafenib treatment as demonstrated by the decreased number of colonies, tumor spheres and stem-like phenotypes (tumor spheres and side-populations) and markers (WNT1, mTOR/STAT3, Notch1). Our observation may explain partially why patients with metastatic and drug-resistant colon cancer responded towards the treatment of regorafenib [
19]. It is interesting to note that Notch1 expression was linked to the poor prognosis in patients with metastatic colon cancer treated with bevacizumab [
20]. According to our results, regorafenib treatment not only resulted in the suppression of Notch1 but also WNT1, CD44, and c-Myc in both DLD-1R and HCT-116R cells.
Mechanistically, regorafenib-treated colon cells (HCT-116R and DLD-1R) showed an elevated level of a various tumor suppressor microRNAs including miR-22, miR-98 and miR-34a while miR-34a as the one being elevated most prominently. The role of miR-34 has been previously linked to the inhibition of tumorigenic phenotypes in HCT-116 cells such as proliferation and metastasis [
21]. In addition, an increased level of miR-34a was shown to inhibit the metastatic potential via the suppression of MET signaling pathway in gastric cancer cells [
22]. Notably, downstream of MET signaling is STAT3 and CD44 which can be both suppressed by the treatment of regorafenib in our study. In addition, HCT-116R and DLD-1R cells transfected with exogenous miR-34a mimic molecules showed a significantly lower ability to form tumor spheres, linking the role of miR-34a to the generation and maintenance of cancer stem-like cells. Previous reports showed that the increased level of miR-34a was linked to the suppression of the generation of cancer stem cells in prostate and pancreatic cancer via directly down-regulating CD44 expression [
23‐
25], in agreement with our observation in colon cancer cells. In addition, Wnt and β-catenin pathway was shown to be one of the targets for miR-34a in breast cancer, supporting our observations in HCT-116R and DLD-1R 5-FU resistant colon cancer cells.
We also examined the potential of combining regorafenib with 5-FU to overcome drug resistance using DLD-1R and HCT-116R cells. We found that the sequential treatment of regorafenib followed by 5-FU synergistically suppressed the viability of both DLD-1R and HCT-116R cells. It is plausible that the initial treatment of Regorafenib suppressed key stemness markers such as CD44, Wnt1/β-catenin, rendering HCT-116R and DLD-1R cells more sensitive towards the subsequent 5-FU treatment. Our in vitro results were supported by our in vivo mouse experiments where the combination of regorafenib and 5-FU provided the most significant tumor inhibitory effect. The immunohistochemical analysis of the tumor samples confirmed our hypothesis that the combination treatment suppressed the WNT and β-catenin signaling, in addition of ABCG2 and increased Bax expression.
Indeed, the clinical application and/or efficacy of combination of regorafenib and 5FU is very important. In fact, there are many clinical trials being conducted or have been completed. For instance, there is one ongoing phase II clinical trial examining of Regorafenib PO plus 5-FU/LV infusion in 15 mCRC patients who progressed on prior Regorafenib monotherapy as well as 5-FU containing chemotherapy combinations. (
https://clinicaltrials.gov/ct2/show/NCT03099486?term=regorafenib%2C+5FU&cond=Colon+Cancer&rank=1). More importantly, in our study, we have provided the potential benefits of using this combination for treating drug-resistant patients through the suppression of cancer stem cells.