Mast cells dysregulate apoptotic and cell cycle genes in mucosal squamous cell carcinoma

Our study shows that TRAIL is up-regulated in the co-cultured SCC-25 target cells, possibly through mast cell-derived interferon-γ (IFN-γ). Up-regulation of TRAIL gene expression and protein synthesis is known to occur in Ewing sarcoma, and thyroid carcinoma cell lines following intereferon (IFN)/cytokine treatment, contributing to apoptosis of the malignant cells in an autocrine and paracrine manner [10, 11]. TRAIL serum levels in melanoma patients are found to be significantly elevated following IFN-α administration [12]. The protein encoded by this gene is a cytokine that belongs to the TNF ligand family [13]. It is an immunological apoptotic inducer that preferentially kills virus-infected, transformed and tumor cells, but spares normal cells [13, 14]. The binding of TRAIL to its receptors triggers activation of MAPK8/JNK, caspase 8 and caspase 3 [13]. Recent finding that TRAIL induces apoptosis in endothelial cells suggests that it may have an indirect anti-angiogenesis property in addition to its tumor cytotoxic effect [15]. Preclinical experiments have demonstrated the efficiency of recombinant human TRAIL and monoclonal antibody against TRAILR1 and TRAILR2 on human breast, colon, and uterine cancers [16].

BIRC4, an endogenous apoptosis inhibitor, is found to be up-regulated. The protein encoded by BIRC4 belongs to a family of highly conserved apoptosis suppressor proteins that bind to TNF receptor-associated factors, TRAF1 and TRAF2. It inhibits apoptosis induced by menadione, caspase 3 and caspase 7 [13].

HMC-1 down-regulates CDK6 expression in SCC25. CDKs are important regulators of cell cycle progression. CDK6, which first appears in mid-G1 phase, is important for G1 phase progression and G1/S transition. Together with CDK4, they regulate the activity of Retinoblastoma (Rb) tumor suppressor protein [13]. Exit from the G1 phase of the cell division cycle is regulated by phosphorylation of pRb by cyclin D/CDK4 and cyclin D/CDK6 complexes. Dysregulation of these critical kinases causes pRb inactivation resulting in deregulation of the cell cycle control. Increased expression of CDK6 has been shown as a mechanism for Rb inactivation in oral SCC [17]. Recently, the growth of melanoma cell lines has been successfully retarded in vitro by down-regulating their CDK6 gene expression with small interfering RNA [18]. Down-regulation of CDK6, coupled with reciprocal up-regulation of Rb (unpublished data, fold change 1.9, p = 0.09) will result in suppression of cell growth as observed in this study.

In addition to up-regulation of BIRC4 and CDC6, and down-regulation of CCNG2 contradict our finding of HMC-1 exerting an overall inhibitory effect on SCC25 proliferation. The DNA replication licensing protein encoded by CDC6 assembles to form one of the pre-replicative complexes required for DNA replication. CDC6 is over-expressed in dysplastic cells. CDC6 mRNA expression increases in a linear fashion in cervical squamous carcinogenesis, from normal cervix through cervical intraepithelial neoplasia to invasive cervical carcinoma [19].

CCNG2 has been suggested as a negative regulator of cell cycle progression. Its dysregulation is implicated in epithelial transformation and the early stages of human oral cancer development. Transfection of human oral SCC cell line with CCNG2 induces cell arrest in the G1 phase, resulting in a significant inhibition of cellular proliferation [20].

As some mast cells secrete a myriad of mediators, many with opposing effects, it is not surprising that genes that are known to inhibit and promote tumor proliferation are both dysregulated. The net effect on growth undoubtedly depends on an intricate interplay between these genes. Our study demonstrates an overall inhibitory effect of mast cells on the proliferation of HNSCC.

Our co-culture study represents the simplest of models to study the effects of mast cells on HNSCC, in a one-to-one relationship. In reality, mast cells are among a very heterogenous population of cells in the tumor microenvironment [4, 8]. They are exposed to a multitude of signals with different temporal patterns from the tumor cells, non-malignant resident cells (e.g., fibroblasts) and other inflammatory/immune cells (e.g., macrophages).

Our study has failed to show an unequivocal direction in the change of gene expression in the target cells, reinforcing the screening nature of microarray tests. Validation of the proposed genes with RT-PCR, and if confirmed, follow-up investigation of potential mast cell-derived mediators effecting change in the target cells e.g., TNF-α, IL-1, IL-4 and IL-6 by siRNA gene knock-down study or subjecting the co-culture supernatant to Western-blotting would be a logical approach.

Given the plasticity and versatility of mast cells, their phenotype may change, being inhibitory or stimulatory to tumor development, depending on the microenvironment [21]. It is also possible that mast cells are initially recruited to the tumors as part of the host defence system, but subsequently become enmeshed within the stroma participating in carcinogenesis [21].

The challenge to targeting mast cells in cancer treatment therefore lies in selective inhibition of tumor-promoting mediators while sparing cytotoxic ones, or identifying and blocking causative factors contributing to their unfavorable phenotypic alteration.

Human mast cell line (HMC-1) was obtained from Dr J. Butterfield (Mayo Clinic, MN, USA) and grown in Iscove's medium (with 25 mM HEPES, sodium bicarbonate, L-glutamine; without alpha thioglycerol or beta mercaptoethanol) (Gibco, NY, USA), 10% iron-supplemented bovine calf serum (Hyclone, Auckland, NZ) and 1.2 mM alpha thioglycerol (Sigma, MO, USA) at 37°C and 5% CO₂. SCC25 was obtained from American Type Culture Collection and maintained in 1:1 mixture of DMEM and Ham's F12 medium (with 1.2 g/L sodium bicarbonate, 2.5 ml L-glutamine, 15 mM HEPES and 0.5 mM sodium pyruvate supplemented with 400 ng/ml hydrocortisone) (Gibco, NY, USA) and 10% fetal bovine serum (Hyclone, Auckland, NZ) at 37°C and 5% CO₂. For the purpose of the co-culture experiment, SCC25 was subsequently harvested by trypsinization and cultured in fresh Iscove's medium as described. This did not impact negatively on the viability as determined by tryptan blue exclusion test.

SCC25 at a cell density of 3.5 × 10⁴/ml (×2.5 ml) was cultured in the lower compartment of a 35-mm co-culture well (Corning, MA, USA). 5.8 × 10⁴/ml of HMC-1 (×1.5 ml) was seeded on the polyester transwell filter (pore size 0.4 μm) (Corning, MA, USA) in the upper compartment, giving an HMC-1:SCC25 co-culture ratio of 1:1. The cells were co-cultured in Iscove's medium as specified above. HMC-1 were then degranulated with 1 μM calcium ionophore A23187 (Sigma, MO, USA). The cells were co-cultured for 72 hours. Negative controls were established in which SCC25 were cultured alone without HMC-1, and A23187 was added to the empty upper compartment of the co-culture chambers. Culture medium was renewed every 24 hours. Histamine release was adopted as a surrogate marker for measurement of mast cell degranulation [22]. 1 μM A23187 was used as secretagogue in this study as higher concentrations did not show further increase in histamine release (Figure 4), and at 1 μM, A23187 did not affect viability of cultured cells as determined by tryptan blue exclusion test. Degranulation of HMC-1 was shown to plateau at 12 hours after stimulation (Figure 4).

Co-culture experiments were conducted for 12, 24, 48 and 72 hours, at the end of which SCC25 were subjected to colorimetric quantification of proliferation and viability with MTT assay (Sigma, MO, USA). Four sets of co-culture experiments were performed for each time frame (n = 4).

Data for the co-culture experiments was subjected to analysis of variants with terms for co-culture effect, time and the interaction between co-culture effect and time, done with the log of the data, and with a random term for the culture wells.