Gastric Cancer Cells With Stemness Properties

Published Date: 02 Nov 2017

Authors

MAO SUN 1, WEI ZHOU 2, YUN-YUN ZHANG 2, DONG-LIN WANG 3, XIAO-LING WU 1

Affiliations

1 Department of Gastroenterology, The second affiliated hospital of Chongqing medical university, Chongqing 400010; 2 Department of Radiology, 3 Oncology of Chongqing cancer hospital, Chongqing 400011. P.R. China.

Correspondence to: Professor XIAO-LING WU, Department of Gastroenterology, The Second Affiliated Hospital of Chongqing Medical University, NO. 74, Ling-Jiang Road, Chongqing 400010, P.R. China. E-mail: [email protected]

Keyword: CD44, gastric cancer, stem cell, chemoradioresistance, cancer invasion

Running title: MAO SUN et al. CD44 positive gastric cancer stem-like cells and cancer invasion

Abstract

CD44 as a cancer stem cell marker had been confirmed in a variety of human cancer cell lines and primary tumors, but whether this marker is applicable to gastric cancer still unknown. The responses of CD44+ gastric cancer stem-like cells to chemoradiation and the role it played in cancer invasion are not well understood. Human poorly differentiated gastric cancer cells was sorted to isolate CD44+ cell fractions in high purity (<1% CD44- cells). The stemness properties of CD44+ cell fraction were confirmed by two "gold standard" methods (i.e. in vivo tumorigenicity assay and in vitro spheroid colony formation assay). Afterwards, we evaluated the treatment response in both CD44+ and CD44- cell fractions which underwent chemoradiation. In general, CD44+ stem-like cells tended to respond more poorly to chemoradiation than their nonstem counterparts. Further studies revealed that CD44+ stem-like cells which scored positive in the migration and invasion assay in vitro did form invasive tumours in vivo. Therefore, we hypothesized that CD44+ stem-like cells might highly express invasion-associated gene. Consistent with our prediction, increased expression of cancer invasion-related genes matrix metalloproteinase-1, matrix metalloproteinase-2, and epidermal growth factor receptor and cyclooxygenase 2 were detected in CD44+ stem-like cells. To the best of our knowledge, it is the first report that reveals the correlation between CD44+ gastric cancer cells and cancer invasion. By selectively eliminating CD44+ stem-like cells, it would be likely to treat patients with aggressive, non-resectable gastric cancers, as well as preventing the tumour from metastasizing.

Introduction

Gastric cancer (GC) is the fourth (989,000 cases, 7.8% of the total) most common cancer and the second (738,000 deaths, 9.7% of the total) cause of cancer mortality worldwide (1). GC rates have decreased noticeably in most parts of the world, but it is still a big burden to Eastern Asia countries (mainly in China) (2). By the time symptoms occur, the cancer has often reached an advanced stage and may have also metastasized.

Clinicians regularly treated patients with appropriate doses and schedules to achieve logarithmic cancer cell kills, while unavoidably kill normal cells which undergoing rapid division. This conventional therapeutic method is just like suicide bomb attack, no one can be spared. Advances in understanding the molecular and cellular basis of cancer, current therapeutic strategies focus on inhibiting the molecular drivers of cancer. In some extent, these modern therapeutic methods can be compared to bio-guided missiles, kill the enemies spare their own. What all these strategies, whether from the past or present, treated cancer as a homogeneous, abnormal entity. Therefore, drugs targeting molecular lesions should be equally effective against all tumour cells, barring the emergence of resistant subpopulation (3).

Cancer cells are heterogeneous not only in morphology but also in functionology (i.e. marker expression, proliferation capacity and tumorigenicity). Only a minority of tumour cells, termed cancer stem cells (CSCs), have the capacity to regenerate the tumour and sustain its growth when injected into immune-compromised mice (4). Based on accumulating evidence, American Association for Cancer Research (AACR) made a consensus definition of the CSC in 2006 as "cells within a tumour that possess the capacity for self-renewal and that can cause the heterogeneous lineages of cancer cells that constitute the tumour" (5).

Studies of gastric cancer stem cells (GCSCs) began relatively late comparing to other solid tumours. A few years ago, Takaishi et al. (6) screened a series of potential stem cell markers in different human gastric cancer cell lines and demonstrated for the first time that CD44 might be an appropriate marker for stem cells. However, similar reports from clinical research were rare. To date, the theory that CSCs are subpopulation with chemoradioresistance had been verified in many solid tumours with the exception of GC (7-11). Besides, the role of gastric cancer stem-like cells played in cancer invasion remains to be elucidated.

Materials and Methods

Cells and animals. Human poorly differentiated GC cells derived from a 55 years-old woman patients with written informed consent, who never underwent chemoradiotherapy before resection. Thirty mice used were 4-week-old Balb/cA nu/nu females. They were from Shanghai Experimental Animal Centre of Chinese Academy of Science (Shanghai, China) and maintained in plastic cages (five mice /cage) in a room with constant temperature (22 ± 1°C) with a dark-light cycle (12h/12h). Animal experiments and human research were performed in accordance with the ethics code by the Ethical Committee of Chongqing medical university.

Fluorescence-activated cell sorting. For FACS cell sorting, 80% confluent cells in a 100-mm cell plate (5-10 million cells per plate) were harvested and incubated for 30 min at room temperature with 10-fold dilution of following antibodies: anti-CD44- fluorescein isothiocyanate rat monoclonal, anti-CD44- PE (both eBiosciences, San Diego, CA)). Then cells were detected by fluorescence-activated cell sorter (FACS)-LSRII flow cytometer (Becton Dickinson, USA).

Spheroid colony formation assay. The CD44 (+) and (-) fractions from FACS-sorted GC cells were inoculated into each well (20 cells per well) of ultra-low-attachment 48-well plates supplemented with 300 μl of DMEM medium plus 40 ng/mL basic fibroblast growth factor (bFGF) and 20 ng/mL epidermal growth factor (EGF, both Invitrogen, USA ). After 4 weeks, each well was examined using light microscope and total well numbers with spheroid colonies were counted. Each experiment was performed at least three times.

In vivo xenograft assay. CD44 (+) and (-) fractions were maintained in sterile DMEM supplemented with 10% FBS. Expanded to be tested for tumorigegnicity. Harvested the cells when they were sub-confluent and adjust the concentration of the cell suspension to be inoculated to 5×104/ml (CD44+) and 5×106/ml (CD44-) in PBS. Inject 0.2 ml of the cell suspension subcutaneously in the left (CD44-) and right (CD44+) hind limb of the mice, respectively. Mice were observed daily and inspected for tumour growth each week for 6 weeks.

In vitro cell migration and invasion assay. Diluted Matrigel (5 mg/ml) in serum-free DMEM. Put 100 μl of the diluted matrigel into upper chamber of 24-well millicell with 8 μm pore size insert (MILLIPORE, USA). Harvested cells and resuspended cells in media containing 0.1% FBS at a density of 1×104/ml. Put 100 μl of the cell suspension onto the matrigel. Lower chamber of the millicell was filled with 600 μl DMEM containing 10% FBS. Cells on the lower side of the insert filter were then stained with 1% Crystal Violet for 20 minutes. Take average counting on the lower side of the filter using an inverted widefield microscope. Each experiment was performed at least three times.

Cell treatment. Anti-cancer drug 5-fluorouracil (5-FU, Sigma-Aldrich) was used to assess the responses of sorted-cells. To evaluate cell viability, the sorted-cells were seeded in flat-bottomed micro culture 96-well plate (2000 cells/well) and allowed to adhere 24 h. Then, cells were treated with 5-FU (10 nM-30 μM) in phenol-free DMEM medium for 72 h. To measure reactive oxygen species (ROS) accumulation, sorted-cells were plated in 6-well plates and stimulated with 5-Fu (50 μM).

MTT assay. After the sorted-cells were treated with 5-Fu for 72 h, MTT was added to a final concentration of 0.5 mg/ml and incubated for 4 h at 37°C. The culture medium was then removed and the remaining blue precipitate was solubilized in DMSO followed by reading absorbance at 570 nm in a microplate reader. This reading was divided by the adjusted absorbance reading of untreated cells in control wells. IC50 was calculated by non-linear regression analysis using sigmoidal fitting from the sigmoidal dose-response curve. For each concentration of both drugs, five wells were analysed. Each experiment was performed at least three times.

Reactive oxygen species assay. Reactive oxygen species assay kit purchased from Beyotime Institute of Biotechnology (Haimen, Jiangsu, china) was used according to the manufacturer’s instructions. Briefly, for loading probe in situ, 2 ml 2', 7'-dichlorofluorescin-diacetate (DCFH-DA 1:1000) fluorescent probe diluted with serum-free DMEM were loaded to each well of 6-well plate. Incubated at 37°C for 20 minutes, and then washed with serum-free media three times. Stimulated the cells with 5-Fu (50 μM) for 2 hours. After stimulation, dichlorofluorescein (DCF, excitation 488/emission 525) fluorescence was assessed immediately by Flow Cytometry.

Irradiation. For colony forming assays, sorted-cells were irradiated with a Varian Clinac iX linear accelerator (VARIAN, USA) at a dose rate of 3 Gy/minute for the time required to generate dose curve of 0,2,4,6, and 8 Gy. Corresponding control were sham irradiated. Colony forming assays were performed immediately after irradiation by plating cells into 6-well culture plates. After 20 days, colonies containing more than 50 cells were counted. A radiation survival curve was generated using Albright’s method (12). For comet assays, sorted-cells were irradiated as above with 0, 2, 4Gy. The performance of the comet assay was mainly based on the method described by Olive (13). A total of 30 representative cells were investigated per slide. The "comets" were measured using image analysis software CASP. The Olive Tail Moment (OTM: median DNA migration distance × relative amount of DNA in the tail of the comet) was used to quantify DNA damage.

Real-time PCR. Quantitative real-time RT-PCR was performed using the Maxima SYBR Green/ROX qPCR Master Mix (2X) (both Fermentas, Canada). Reactions were carried out using iCycler (Bio-Rad, USA) and the results were evaluated with the iCycler real-time detection system software. Relative quantitation of target gene expression was evaluated by the comparative Ct method.

Western blot. The immunoreagents used for Western blot were rabbit monoclonal antibody against MMP-1, COX-2 (1:100) and mouse monoclonal antibody against MMP-2, EGFR (1:100). Mouse polyclonal anti-actin antibody (1:1000; All from Santa Cruz, USA) was used as loading control. The blots were developed by a standard enhanced chemiluminescence (ECL) method (Pierce, USA).

Immunofluorescence staining. CD44+ cell spheroids were fixed and blocked in PBS solution with 10% FBS. The primary antibodies (rabbit anti-MMP-1, mouse anti- MMP-2, 1:100) were incubated at 4 °C overnight. The fluorescent secondary antibodies (Goat anti-rabbit IgG antibodies conjugated with FITC, goat anti- mouse IgG antibodies conjugated with TRITC, 1:100) were added and incubated at 37 °C for 30 min. In negative controls, primary antibodies were substituted by PBS. Cells nuclei were counterstained with DAPI. Fluorescence was observed using laser scan confocal microscopy.

Histological examination. Tumour tissues were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned and stained with haematoxylin and eosin. Samples were observed to examine histological differences under microscopy at 40×magnification.

Statistical analysis. Data are presented as the mean ± standard error of the mean (SEM). Graphpad Prism 5.0 was used for statistical analysis. Statistical difference for IC50 is determined by F test. The rest of comparisons between groups were evaluated by unpaired t test. General acceptance of the level of significance was p ≤ 0.05.

Results

Confirmation of stemness properties of CD44+cells. CSC could form suspended cell spheroid under growth factor rich, low-attachment and serum-free culture conditions (14). In spheroid colony formation assay, CD44+ cells formed apparently more spheroids than CD44- cells (Table 1, p<0.01). The diameter of CD44+ spheroids reached to about 120 μm, containing approximately 1000 cells (data not shown). Tumorigenic capacity of CSCs varied from 10 to 100 fold than that of nonstem counterparts in different cancer cells (15-17). In tumorigenicity assay, individual difference variables of nude mice were excluded by injection of both cell fractions in a same mouse. In general, injection of 1×104 CD44+ cells gave rise to tumours with 80% take incidence with a relatively short latency periods (<1 week) in all mice. In contrast, injection of 1×106 CD44- cells conferred tumour formation with very low engraftment rate 27% (Table I, p<0.01). We interpret these data by the fact that CD44+ cells exhibit the intrinsic property to form tumours.

Responses of CD44+stem-like cells to chemoradiation. In MMT assay, based on the calculated IC50 values, the responses of both cell fractions to 5-FU varied significantly. Log EC50 of CD44+ cells was -5.961±0.04566 compared to that of CD44- cells -6.415±0.04231 (Figure 1A, p<0.05)). We performed ROS assay to explore the underline mechanism of chemoresistance. Consequently, fluorescence enhancement was observed in both cell fractions. However, intracellular fluorescence value in CD44+ cells was considerably lower than that of the CD44- cells (Table I, p<0.05). Cytotoxic oxidative stress result from ROS can damage DNA, RNA, proteins, and lipid components, which may lead to cell death. Most importantly, ROS-induced cell apoptosis occurred in the early stages in response to chemotherapeutic drug treatment (18). Therefore, we speculated that the reduction in oxidatively generated DNA damage in CD44+ cells due to its antioxidant ability. Meanwhile, we conducted clonogenic assay and comet assay to evaluate proliferative death of cells and quantify its DNA breaks, independently. Interestingly, the radiation survival curve of both cell fractions had a shoulder that is characterized by a comparably higher resistance at lower doses of radiation. This curve shoulder of CD44- cells disappeared at SF4Gy. In contrast, it existed in CD44+ stem-like cells till radiation dose increased to 6 Gy (Figure 1B). In comet assay, irradiated cells with damaged DNA fragments would form a tail around the DNA head after electrophoresis. Olive Tail Moment value is positively related to the damaged extent of DNA. Consistent with clonogenic assay, difference of the OTM value was not significant when irradiated with 2 Gy (p>0.05). In contrast, the difference became considerably larger when irradiation doses increased to 4 Gy (Figure 1C, p<0.01).

Enhancement of invasion capacity of CD44+stem-like cells. Invasive and migratory capacities of CD44+ stem-like cells were much higher compared with CD44- cells (Table I, p<0.01). CD44+ cells could penetrate into and pass through the matrix in vitro which means it might synthesize more MMPs. Researchers recently revealed four important cancer invasion-related genes MMP-1, MMP- 2, EGFR and COX-2 in a variety of human cancers including GC, which could manipulate the migration of tumour cells and the formation of new tumour by facilitating the release of tumour cells into the circulation (19). Therefore, we compared the expression profile of the four cancer invasion-related genes in CD44+ stem-like cells and their nonstem counterparts. This process contributed to the validation of four genes showing considerably difference of expression levels both in RNA (Figure 2A) and protein (Figure 2B). Furthermore, two representative genes, MMP-1 and MMP-2 were detected co-expressed in CD44+ cell spheroids by immunofluorescence staining. Strong immunoreactivity was detected for both proteins localized in whole cancer cells (Figure 3).

CD44+ GC stem-like cells which scored positive in the migration and invasion assay in vitro did form invasive tumours in vivo. Figure 4A show a representative mouse with progressive muscle invasion by a tumour was paralysed in its hind limb. H & E stain histology of mass demonstrated invasive tumour within muscle, consistent with poorly differentiated adenocarcinoma. Moreover, we observed histological difference between CD44+ tumours and CD44- tumours. CD44+ tumours with larger, irregular and hyperchromatic nuclei were more heterogeneous (Figure 4B).

Discussion

Many researchers paid particular attention on signalling pathways, which might mediate resistance of CSCs (20, 21). Our findings will contribute to the understanding of antioxidant ability of CSCs as another important mechanism in relation to chemoradioresistance. Expression of CD44 in GC cells enhances tumorigenicity. This finding is in line with the significance of CD44+ cancer cells in tumorigenicity of other cancers (6, 22-24). The unique and most important finding was the increased invasion capacity of CD44+ stem-like cells both in vivo and in vitro. Furthermore, we verified the expression of cancer invasion-related genes MMP-1, MMP- 2, EGFR and COX-2 were up-regulated in CD44+ gastric cancer cells, which indicated the capacity to manipulate cancer invasion, even metastasis.

Our findings may have clinical implications. Examination for CD44+ cells in the primary GC may predict the development of distant metastasis. This may facilitate patients’ selection for adjuvant chemoradiotherapy to reduce chance of recurrence after resection of primary GC. During the period of the article drafting, another research group successfully isolated GCSCs from peripheral blood of cancer patients using CD44 surface markers (25), which demonstrated that GCSCs have the possibility to be transferred to any organs of the whole body through blood circulation. Actually, nealy106 cancer cells per gram of cancer tissue shed into the bloodstream daily (26). However, it is not the reflection of the amount of distant metastasis found. So, we speculated that only a small minority of cancer cells with metastasis capacity (i.e. cancer stem cells) are the culprit for cancer invasiveness and metastasis.

At least two important features of CD44 make it to be a suitable CSC maker. CD44 is a receptor for hyaluronic acid and can also interact with other ligands such as collagens, and MMPs (27). So, CSCs with the expression of CD44 could manipulate cells invasion, migration and adhesion to matrix (28). Furthermore, CD44+ cancer cells are thought to be slow-cycling, thereby, insensitive to chemoradiotherapies (29). We also noticed that sequentially irradiated GC cells with low x-ray doses resulted in the inhibition of DNA synthesis and the accumulation of CD44+ GC cells in G0/G1 phase of the cell cycle (data not shown), which revealed the cell cycle of CD44+ GC cells could be effectively regulated to avoid DNA damage by external stimulus.

Acknowledgement

We would like to express our special gratitude to Hui-Ming Yang for her collection of xenograft tissues and Jian-Ye Xu for his technical assistance.