CD44^high/ESA^low squamous cell carcinoma cell-derived prostaglandin E₂ confers resistance to 5-fluorouracil-induced apoptosis in CD44^high/ESA^high cells

Abstract

We previously found that CD44^high/ESA^low head and neck squamous cell carcinoma (HNSCC) cells harboring high dihydropyrimidine dehydrogenase (DPD) expression exhibited potent resistance to 5-fluorouracil (5-FU)-induced apoptosis. In addition, susceptibility of HNSCC cells to 5-FU was compromised in the presence of cyclooxygenase 2 (COX2)-derived prostaglandin E₂ (PGE₂). In this study, we examined 5-FU-induced apoptosis in sorted cell populations (i.e., CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells from the HNSCC cell line A-253) to clarify the anti-apoptotic effect of PGE₂ on CD44^high cells. Notably, CD44^high/ESA^low cells upregulated PGE₂, compared with other populations. To investigate the effect of CD44^high/ESA^low cell-derived PGE₂ on CD44^high/ESA^high cells, direct and indirect co-culture assays were performed. The percentage of apoptotic cells in a culture of CD44^high/ESA^high cells was significantly reduced when they were directly and indirectly co-cultured with CD44^high/ESA^low cells. Furthermore, 5-FU-induced apoptosis of CD44^high/ESA^high cells was significantly increased in the presence of an inhibitor of the PGE₂ receptors (EP1/EP2) when CD44^high/ESA^high cells were co-cultured with CD44^high/ESA^low cells. These results suggest that CD44^high/ESA^low cell-derived PGE₂ may contribute to the inhibition of 5-FU-induced apoptosis in CD44^high/ESA^high cells. Additionally, NR4A2 knockdown enhances 5-FU-induced apoptosis in CD44^high/ESA^high cells, suggesting that PGE₂ attenuates 5-FU-induced apoptosis in an NR4A2-dependent manner in CD44^high/ESA^high cells. In conclusion, CD44^high/ESA^low cells contribute to induction of resistance to 5-FU in CD44^high/ESA^high cells through provision of PGE₂. CD44^high/ESA^low cell-targeted therapy may be effective in treatment of HNSCC.

Introduction

While the mortality rate of head and neck squamous cell carcinoma (HNSCC) has improved recently, survival rates remain lower than those of other malignant tumours, such as colorectal and breast cancers [1]. In particular, patients with advanced-stage HNSCC exhibit high rates of recurrence and metastasis [2,3]. This worsened prognosis may be associated with the behaviour of a small population of cancer stem cells that may be less sensitive to conventional cancer therapy [4]. In HNSCC, populations of CD44^high cells exhibit cancer stem cell (CSC) properties, as well as phenotypic plasticity, such as mesenchymal (i.e., CD44^high/ESA^low cells) to epithelial transition (i.e., CD44^high/ESA^high cells) [5]. In addition, CD44^high/ESA^low cells exhibited strong resistance to 5-fluorouracil (5-FU)-induced apoptosis, along with high dihydropyrimidine dehydrogenase (DPD) expression [6]. Therefore, we hypothesized that surviving CD44^high/ESA^low cells may play an important role in the poor prognosis associated with HNSCC. Additionally, we have reported that susceptibility of HNSCC cells to 5-FU might be compromised following cyclooxygenase 2 (COX2)-derived prostaglandin E₂ (PGE₂) delivery to the cancer microenvironment [7]. We suspect that PGE₂ might be involved in enhancement of the anti-apoptotic effect in HNSCC cells. However, it remains unknown whether PGE₂ can block apoptosis in CD44^high cells (i.e., HNSCC cells with CSC properties). In this study, we examined the effect of exogenous PGE₂ on 5-FU induced apoptosis in CD44^high subpopulations in HNSCC cells. In addition, we examined 5-FU-induced apoptosis of CD44^high/ESA^high cells that were co-cultured directly or indirectly with CD44^high/ESA^low cells.

Materials and methods

Cell culture and treatment

A HNSCC cell line, A-253 (American Type Culture Collection (ATCC), Manassas, VA, USA), was used in this study. Cells were cultured in a highly supplemented epithelial growth medium (Dulbecco’s Modified Eagle’s Medium: DMEM), with 10% FBS, under 5% CO₂ in air at 37°C [8]. For re-plating and for assays, cells were released into suspension using Accutase (Nakalai Tesque, Kyoto, Japan). To investigate the effect of the CD44^high/ESA^low cells on CD44^high/ESA^high cells, direct co-culture assays were performed by culturing the two cellular populations together. An indirect co-culture system, using Transwell cell culture inserts with microporous membranes (pore size: 1.0 μm) (BD Falcon, Milian, Italy) was also employed to prevent direct cell-to-cell communication. PGE₂ (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in dimethylsulfoxide (DMSO). DMSO was used as control for PGE₂. PGE₂ was added after cells had been cultured without FBS for 48 h. SC-51089 (Cayman Chemical, Ann Arbor, MI, USA) and AH-6809 (Cayman) were used as EP1 receptor and EP1/EP2 receptor antagonist, respectively. H-89 (Enzo Life Sciences, Farmingdale, NY, USA) was used as a cAMP-dependent protein kinase (PKA) inhibitor.

Fluorescence-activated cell sorting

Anti-CD44-PE-conjugated antibody (BD Pharmingen, San Diego, CA, USA) and anti-ESA-APC-conjugated antibody (BD Pharmingen) were used for fluorescence-activated cell sorting (FACS). 7-AAD (BD Pharmingen) was used to exclude dead cells during FACS analysis. According to our previously published methodology [9], samples were assayed on a Becton Dickenson FACSCalibur^TM (BD Biosciences, San Jose, CA, USA); CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells were sorted using Becton Dickenson FACSAria equipment.

Apoptosis analyses

We examined apoptosis induced by 5-FU (Wako, Osaka, Japan) in A-253 cells. 1.0 × 10⁴ cells were seeded in 6-well plates and grown for 24 h, then exposed to 25 µg/mL 5-FU for 48 h. Cells were collected and stained with 7-AAD and AnnexinV-Cy5 (BD Pharmingen), then AnnexinV-positive apoptotic fractions were analysed by FACS. Results are expressed as the mean ± SD for three independent experiments.

Introduction of green fluorescent protein (GFP) through viral infection

Introduction of GFP into the cells was performed according to our previously published method [10]. Briefly, virus host cells (HEK293FT cells) were cotransfected with a viral vector plasmid encoding EGFP (pLenti 6.3) (Invitrogen) and packaging plasmid Virapower mix (Invitrogen), using the EGFP X-treme GENE HP reagent (Roche, Basel, Switzerland). After 48 h, the viral supernatant was mixed with 8 μg/mL polybrene (Sigma-Aldrich), and then used to infect the target cells by spinoculation.

Measurement of PGE₂ concentration

Enzyme-linked immunosorbent assay (ELISA) was performed to measure PGE₂ concentrations via the Prostaglandin E₂ Monoclonal Kit (Oxford Biomedical Research, Oxford, UK), according to the manufacturer’s instructions. Briefly, after 24 h incubation of the cells in culture medium with or without FBS, culture media were harvested to determine the PGE₂ concentration. The absorbance values were measured using a microplate reader with a 450-nm filter. Results are expressed as the mean ± SD for three independent experiments.

Quantitative RT-PCR analysis

The RNAeasy Micro Kit (Qiagen, Hilden, Germany) was used for RNA extraction and the resulting total RNA was subjected to reverse transcription using the ReverTra Ace^® qPCR RT Kit (TOYOBO, Osaka, Japan). Quantification of mRNA levels was performed using the CFX Connect real-time PCR detection system (Bio-Rad, Hercules, CA, USA) and SYBR Green PCR Master Mix (TOYOBO). The reaction mixture contained 1.0 µg of cDNA, 10.0 µL of SYBR Green Mix, and 10 µmol of each pair of oligonucleotide primers. GAPDH was used as a reference control. The primer sequences were as follows: Cox2, 5’-TTGCTGGAACATGGAATTACC-3’ (sense), 5’-TGCCTGCTCTGGTCAATG-3’ (antisense); NR4A2, 5’-GTCTCAGCTGCTCGACACG-3’ (sense), 5’-TTTTGCACT-GTGCGCTTAAA-3’ (antisense); Bcl-2, 5’-CCCTGTGGATGACTGAGTAC-3’ (sense), 5’-GCATGTTGACTTCACTTGTG-3’ (antisense); Bax, 5’-GGCCCACCAGCTCTGAGCAGA-3’ (sense), 5’-GCCACGTGGGCGTCCCAAAGT-3’ (antisense); and G3PDH, 5’-GTGAACCATGAGAAGTATGACAAC-3’ (sense), 5’-ATGAGTCCTTCCACGATACC-3’ (antisense). The PCR program was as follows: initial melting at 95°C for 10 min followed by 40 cycles at 95°C for 15 sec, 57°C for 30 sec and 72°C for 40 sec. Results are expressed as the mean ± SD for three independent experiments.

Western blotting

Protein samples were solubilized in sample buffer by boiling, then separated in a 10% polyacrylamide gel and blotted onto a nitrocellulose membrane. Western blot bands were detected using an enhanced chemiluminescence Western blotting reagent (GE Healthcare, Piscataway, NJ, USA). Antibodies (all diluted at 1:1000) consisted of an anti-human EP-1 mouse monoclonal anti-body (Abcam, Cambridge, UK), an anti-human EP-2 mouse monoclonal antibody (Abcam), and an anti-human GAPDH mouse monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

siRNA knockdown

Stealth^TM siRNAs were used for NR4A2 knockdown (Life Technologies, Gaithersburg, MD, USA) and a Stealth siRNA negative control (Life Technologies) served as the control for knockdown. Cells were transfected with siRNA using HiPerFect transfection reagent (Qiagen, Hilden, Germany), according to the manufacturer’s recommendations.

Statistical methods

Statistical analysis was performed using Student’s t-test. P values < 0.05 were regarded as statistically significant.

Results

PGE₂ inhibit 5-FU-induced apoptosis of CD44^high/ESA^high cells via EP1/2 receptor

All cell fractions expressed PGE₂ receptor, EP1, and EP2 proteins, as shown by Western blotting (Figure 1A). Thus, we analysed the effect of exogenous PGE₂ on 5-FU-induced apoptosis in CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells. CD44^high/ESA^low cells exhibited the most potent resistance to 5-FU-induced apoptosis. Notably, PGE₂ significantly reduced the percentage of 5-FU-induced apoptotic cells in CD44^high/ESA^high cells at concentrations of 10 and 20 ng/mL (Figure 1B). However, PGE₂ did not significantly reduce 5-FU-induced apoptosis in CD44^high/ESA^low and CD44^low cells (Figure 1B). 5-FU-induced apoptosis was examined in the presence of PGE₂ and either EP1 or EP1/EP2 receptor inhibitor. PGE₂-inhibited apoptosis was significantly increased after the addition of either EP1 or EP1/EP2 receptor antagonist (Figure 1C).

Figure 1.

Prostaglandin E₂ (PGE₂) inhibits 5-fluorouracil (FU)-induced apoptosis of CD44^high/ESA^high cells via the EP1/2 receptor. A. EP1 and EP2 expression in CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low fractions of A-253 cells. B. 5-FU induced apoptosis in CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells in the presence of PGE₂. C. 5-FU induced apoptosis of CD44^high/ESA^high cells in the presence of PGE₂ and either SC-51089 or AH-6809. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

CD44^high/ESA^low cells produce an increased amount of PGE₂ compared than CD44^high/ESA^high cells

Next, we examined the PGE₂ production ability of sorted cells. After 24 h incubation of the cells in culture medium without FBS, culture media were collected to determine the PGE₂ concentration. CD44^high/ESA^low cells exhibited significantly higher PGE₂ concentration than CD44^high/ESA^high cells or CD44^low cells (Figure 2A). In addition, COX2 mRNA expression was significantly elevated in CD44^high/ESA^low cells, compared with CD44^high/ESA^high cells or CD44^low cells (Figure 2B).

Figure 2.

Prostaglandin E₂ (PGE₂) concentration and cyclooxygenase 2 (COX2) mRNA expression of CD44^high/ESA^low, CD44^high/ESA^high and CD44^low cells. A. PGE₂ concentration of CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells. B. COX2 mRNA expression levels of CD44^high/ESA^low, CD44^high/ESA^high, and CD44^low cells. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

Endogenous PGE₂ inhibits 5-FU-induced apoptosis of CD44^high/ESA^high cells

To investigate the effect of CD44^high/ESA^low cell-derived PGE₂ on CD44^high/ESA^high cells, direct co-culture assays were performed by culturing the two cellular populations together (Figure 3A). GFP-tagged CD44^high/ESA^low cells and CD44^high/ESA^high cells were co-cultured in the presence of 5-FU, followed by determination of the percentage of apoptotic cells within the GFP-negative CD44^high/ESA^high population by FACS. The percentage of apoptotic cells within a culture of CD44^high/ESA^high cells significantly decreased in the presence of CD44^high/ESA^low cells, relative to CD44^high/ESA^high cultures grown in the absence of CD44^high/ESA^low cells (Figure 3A). Next, indirect co-culture of CD44^high/ESA^low and CD44^high/ESA^high cells was performed using Transwell cell culture inserts, in the presence of 5-FU (Figure 3B). The percentage of apoptotic cells within a culture of CD44^high/ESA^high cells was significantly reduced when CD44^high/ESA^high cells were indirectly co-cultured with CD44^high/ESA^low cells (Figure 3B). Furthermore, the percentage of apoptotic cells within a culture of CD44^high/ESA^high cells was significantly increased by the addition of SC-51089 and AH-6809 in the presence of 5-FU, when CD44^high/ESA^high cells were indirectly co-cultured with CD44^high/ESA^low cells (Figure 3C). These results suggest that CD44^high/ESA^low cell-derived PGE₂ may contribute to the inhibition of 5-FU-induced apoptosis in CD44^high/ESA^high cells.

Figure 3.

CD44^high/ESA^low cell-derived prostaglandin E₂ (PGE₂) inhibits 5-fluorouracil (FU)-induced apoptosis of CD44^high/ESA^high cells. A. Direct co-culture assays of green fluorescent protein (GFP)-tagged CD44^high/ESA^low cells and CD44^high/ESA^high cells in the presence of 5-FU. B. Indirect co-culture assays of CD44^high/ESA^low cells and CD44^high/ESA^high cells in the presence of 5-FU. C. Indirect co-culture assays of CD44^high/ESA^low cells and CD44^high/ESA^high cells in the presence of 5-FU and either SC-51089 or AH-6809. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

PGE₂ induces NR4A2 expression via PKA dependent manner in CD44^high/ESA^high cells

We previously revealed that exogenous PGE₂ induces NR4A2 expression in a PKA-dependent manner in oral SCC cells [7]. Therefore, we aimed to examine the effect of PGE₂ on NR4A2 expression in CD44^high/ESA^low cells. CD44^high/ESA^low cells exhibited a slight increase in NR4A2 mRNA expression in the presence of PGE₂ (Figure 4A). However, NR4A2 mRNA expression was significantly upregulated in the presence of PGE₂ in CD44^high/ESA^high cells. Moreover, PGE₂-induced NR4A2 expression was inhibited by the addition of PKA inhibitor in CD44^high/ESA^high cells (Figure 4B).

Figure 4.

Prostaglandin E₂ (PGE₂)-induced NR4A2 expression was reduced in the presence of protein kinase A (PKA) inhibitor. A. NR4A2 mRNA expression in the presence of 20 ng/mL PGE₂. B. NR4A2 mRNA expression in the presence of PGE₂ and PKA inhibitor in CD44^high/ESA^high cells. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

NR4A2-knockdown enhances 5-FU-induced apoptosis in CD44^high/ESA^high cells

Finally, we examined the effect of NR4A2 siRNA knockdown on 5-FU-induced apoptosis in CD44^high/ESA^high cells. Following siRNA knockdown of NR4A2, the percentage of apoptotic cells was significantly increased (Figure 5A). In addition, we examined expression of apoptosis-related genes, such as anti-apoptotic Bcl2 and pro-apoptotic Bax, during NR4A2 siRNA knockdown in CD44^high/ESA^high cells. Bcl2 mRNA expression was significantly downregulated after siRNA knockdown of NR4A2 (Figure 5B). In contrast, Bax expression was significantly upregulated after siRNA knockdown of NR4A2 (Figure 5C).

Figure 5.

NR4A2-knockdown enhances 5-fluorouracil (FU)-induced apoptosis of CD44^high/ESA^high cells. A. The percentage of apoptotic cells after siRNA knockdown of NR4A2. B. Bcl2 mRNA expression following NR4A2 siRNA knockdown in CD44^high/ESA^high cells. C. Bax mRNA expression following NR4A2 siRNA knockdown in CD44^high/ESA^high cells. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

Discussion

Several types of cancer can produce pro-inflammatory cytokines such as PGE₂, thereby inducing localized infiltration of inflammatory cells [11-13]. PGE₂ is suspected to contribute to modulation of the cancer microenvironment (i.e., cell proliferation, anti-apoptosis, and angiogenesis, as well as migration and invasion of tumor cells) [14]. There is mounting evidence to support the oncogenic role of PGE₂ within the context of the COX2-PGE₂-EP1/2-dependent signalling pathway [15,16]. Several studies have reported a significant relationship between COX2 expression and CSC properties [17-20]. For example, there is greater sphere-forming ability within breast cancer cells that upregulate expression of COX-2 and its receptor, EP4. Notably, sphere forming ability is suspected to be a significant feature of CSCs [18]. COX-2 stimulates self-renewal of cancer stem-like cells of glioma via PGE₂ [19]. In addition, CSC repopulation has been blocked by the inhibition of the COX2-PGE₂ signalling pathway in bladder cancers [20]. These results indicate that PGE₂ may play a significant role in the maintenance of CSCs in different types of cancers. In this study, PGE₂ was involved in the inhibition of 5-FU-induced apoptosis of CD44^high/ESA^high cells, indicating that PGE₂ may contribute to resistance to chemotherapeutics in HNSCC CSCs with epithelial character.

It remains unclear which type of cells (i.e., cancer cells, cancer-associated stromal cells, blood cells, or other normal cells) are the main source of PGE₂ in HNSCC. This study revealed that CD44^high/ESA^low cells exhibit a greater capacity to produce PGE₂, compared with other populations of cells. CD44^high/ESA^low cells with both CSC and mesenchymal properties may contribute to PGE₂ production in HNSCC. In addition, we found that 5-FU-induced apoptosis was significantly blocked in CD44^high/ESA^high cells when they were co-cultured with CD44^high/ESA^low cells. Thus, PGE₂ release from CD44^high/ESA^low cells may confer anti-apoptotic capacity to CD44^high/ESA^high cells, likely via EP1/EP2 receptor activation. Previously, we found that CD44^high/ESA^low cells exhibit strong resistance to other chemotherapeutics (e.g., cisplatin and docetaxel) [9]. Therefore, CD44^high/ESA^low cells may be involved in the strong resistance to the conventional chemotherapy that is observed in HNSCC. CD44^high subpopulations with potent resistance to chemotherapeutics (e.g., CD44^high/ESA^low cells) may not only survive but also aid in the survival of other CD44^high subpopulations (e.g., CD44^high/ESA^high cells) in the presence of chemotherapeutics such as 5-FU.

NR4A2 is a member of the orphan nuclear hormone receptor subfamily and is implicated in a wide variety of biological processes [21,22]. NR4As may be capable of inhibiting tumour suppressor signalling [23]. In this study, PGE₂-induced NR4A2 expression was inhibited by a PKA inhibitor, suggesting that PGE₂ induces NR4A2 expression in a PKA dependent manner within CD44^high/ESA^high cells. The combination of Bcl2 upregulation and Bax downregulation, following NR4A2 siRNA knockdown, strongly supports the notion that NR4A2 contributes to anti-apoptosis signalling in CD44^high/ESA^high cells.

Collectively, our results indicate that CD44^high/ESA^low cells play a significant role in the induction of resistance to 5-FU within CD44^high/ESA^high cells, largely by providing PGE₂ (Figure 6). Our results suggest that cancer-associated PGE₂ may regulate chemotherapeutic resistance in HNSCC. Thus, CD44^high/ESA^low cell-targeted chemotherapies (or other potential therapies) may be effective in treatment of HNSCC.

Figure 6.

CD44^high/ESA^low cell-derived prostaglandin E₂ (PGE₂) inhibits 5-fluorouracil (FU)-induced apoptosis in CD44^high/ESA^high cells. PGE₂ release from CD44^high/ESA^low cells may contribute to enhanced resistance to 5-FU-induced apoptosis in CD44^high/ESA^high cells. (Statistical significance levels of P < 0.05, P < 0.01, and P < 0.001 are indicated by *, ** and *** respectively).

Disclosure of conflict of interest

None.