More than 35,000 people in the United States and more than 500,000 worldwide are estimated to be diagnosed with head and neck squamous cell carcinoma (HNSCC) annually [1], [2]. The presence of metastasis in patients with head and neck cancer is common and the 5-year survival rate for patients with lymph node metastasis is approximately 25–50% [3]. Better treatments for metastatic HNSCC are urgently needed. However, our understanding of the factors that regulate metastasis in this disease is limited.

Death receptor 5 (DR5) is one of the cell surface receptors that when activated by its ligand, tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), induces the activation of the extrinsic apoptotic pathway in humans [4]. DR5 has been shown to be overexpressed in several types of cancer including colon, lung and cervical cancer [5]–[9]. Increased DR5 expression was also associated with reduced survival in non-small cell lung cancer [7], [9]. A recent mouse study has shown that deficiency of TRAIL receptor in mice (only one receptor for TRAIL in mouse) enhances lymph node metastasis (LNM) without affecting primary tumor development [10], suggesting that TRAIL receptor or TRAIL-TRAIL receptor interaction may be critical for regulation of tumor metastasis. Agonistic antibodies targeting DR5 are currently in clinical trials for treatment of various types of cancer [11]. Currently, the role of DR5 in metastasis is unknown and the expression of DR5 in primary and metastatic HNSCC has not been examined.

Caspase-8 is the first caspase activated during death receptor-initiated apoptosis [12]. There is evidence of increased expression of caspase-8 in several types of cancer including colorectal and rectal, gastric, pancreatic, and breast cancers [13], [14]–[17]. Besides, it has been also shown that caspase-8 expression is lost or inactivated in certain types of cancer such as small cell lung cancer, neuroblastoma, gastric carcinoma and hepatocellular carcinoma [18]–[25]. Loss of caspase-8 has been associated with metastasis in neuroblastoma [26]. However, it has also been recently shown that caspase-8 is associated with cell migration and can promote metastasis in apoptotic resistant cells [27], [28]. Moreover, a loss of caspase-8 was reported to be associated with unfavorable survival in childhood medulloblastoma [29]. Caspase-8 expression in HNSCC, particularly in metastatic HNSCC, has not been documented.

Thus, this study was particularly interested in comparing the expression patterns of DR5 and caspase-8 between primary HNSCC without LNM and HNSCC with LNM. To this end, we performed immunohistochemistry (IHC) to detect DR5 and caspase-8 on three groups of tumor samples from patients with either primary tumors with no evidence of LNM, primary tumors with LNM and the matching LNM.

The death receptor-mediated extrinsic apoptotic pathway plays an essential role in host immunosurveillance against tumor development, particularly metastasis [32]–[34]. In a genetic study, knockout of TRAIL receptor in mice does not affect primary tumor development, but enhances LNM [10], suggesting that TRAIL receptor is important for suppressing tumor metastasis. In human melanoma samples, a reduced DR5 expression was reported to be associated with metastatic lesions [35]. In agreement, the current study revealed a significant downregulation of DR5 expression in primary tumors with LNM (Tu^+met) and their matching LNM compared to primary tumors with no metastasis (Tu^−met). Moreover, DR5 expression was significantly reduced in PD tumors compared to MD and WD tumors. Therefore, our data on DR5 from human HNSCC samples supports an inhibitory role of DR5 in regulation of metastasis. The mechanism of DR5 in regulation of metastasis is not known. However, DR5 is involved in mediating anoikis, a form of apoptosis triggered by loss of attachment of cells from the extracellular matrix [36], [37]. Therefore, it is possible that the loss of DR5 we see in our metastatic HNSCC has contributed to the inhibition of anoikis, therefore allowing the tumor cells to escape apoptosis and migrate after detachment. It has been shown that in non-small cell lung carcinoma tissues, increased expression of DR5 correlates with PD tumors. Moreover, high DR5 expression is significantly associated with reduced overall survival [9]. However, our data clearly show that reduced DR5 expression correlates significantly with PD HNSCC. Furthermore, we did not find significant association between DR5 expression and overall survival or disease-free survival in Tu^−met group. In Tu^+met group, we found that the levels of DR5 were not associated with overall survival. However, higher DR5 was significantly associated with a worse disease-free survival. Despite the potential role of DR5 in negative regulation of metastasis as discussed above, our study did not show a survival advantage for Tu^+met HNSCC with high expression of DR5. In fact, one study has shown that TRAIL enhances the invasion of apoptosis-resistant pancreatic ductal adenocarcinoma cells in vitro and increases distant metastasis (e.g., liver) of pancreatic tumors in vivo [38], suggesting that activation of the DR5 signaling may facilitate metastasis under certain conditions. Moreover, the major mediator of DR5, caspase-8, has non-apoptotic functions that promote cell motility and migration as discussed below. Thus, whether DR5, like caspase-8, may also exert non-apoptotic functions, particularly in apoptotic resistant cells (see discussion below), should be further investigated.

We noted that DR5 staining in our study was primarily in the cytoplasm. However DR5 is known to be functional in inducing apoptosis as a membrane-bound protein. Given that DR5 expression levels do impact patient prognosis as demonstrated in this study, our data suggest that it may be interesting to study whether membrane-bound and cytoplasmic DR5 proteins exert distinct functions (e.g., apoptotic vs. non-apoptotic) under different conditions.

It has been shown that cigarette smoke impairs tumor immune surveillance and promotes invasion of cancer cells including oral carcinoma cells and tumor metastasis in experimental systems [39]–[43]. In this study, we found after examining all patient samples together that a lower DR5 expression in HNSCC was significantly associated with smokers compared to non-smokers. Thus, it would be interesting to determine if tobacco carcinogens can downregulate DR5 expression and possibly add to the effect of tumor cells escaping apoptosis or contributing to metastasis.

In addition to DR5, DR4 is another TRAIL receptor that can initiate death signaling upon TRAIL binding or overexpression [4]. Depending on tumor types, expression of DR4 has variable impact on prognosis. For example, high DR4 expression has been shown to be associated with worse disease-free survival, worse overall survival and shorter time to recurrence in colon cancer [6], whereas DR4 expression did not impact patient survival in lung, cervical and ovarian cancers [7], [44], [45]. Moreover, in breast cancer, DR4, in contrast to DR5, has been shown to be more strongly expressed in better differentiated tumors, and correlated positively with surrogate markers of a better prognosis (hormone receptor status, Bcl-2, negative nodal status), but negatively with the expression of Her2/neu and the proliferation marker Ki67 [46]. In our study, we did not stain DR4 expression in our cohort of HNSCC tissues, largely due to antibody issues. Nonetheless, it will be interesting to study DR4 expression in HNSCC and its association with LNM and prognosis in the future.

The major function of caspase-8 is to mediate apoptosis induced by death receptors including DR5. It has been recently suggested that caspase-8 can play dual roles: one as an inducer of apoptosis and one as a promoter in metastasis [47]. Caspase-8 facilitates cell death initiated from the death receptor pathway after a death ligand (e.g., TRAIL) binds a death receptor (e.g., DR5) [48]. In addition, caspase-8 can also promote cell migration [27], [49]–[51]. It has been suggested that caspase-8 can contribute to cell motility and adhesion by regulating calpain activity which controls cell migration including rac activation and lamellipodial assembly [50]. As well, the phosphorylation of procaspase-8 on tyrosine 380 and its interaction with the p85 alpha subunit of phosphatidylinositol 3-kinase was required to restore cell motility and adhesion in caspase-8 null cells [49]. However, there is evidence that a loss of caspase-8 is associated with increased metastasis. In neuroblastoma, a loss of caspase-8 prevented apoptosis by integrin-mediated cell death and therefore promoted metastasis [26]. In our study, caspase-8 expression trended towards a downregulation of expression in the metastatic group of patients, but this result was not statistically significant. Thus, it is unclear if the downregulation of caspase-8 that we observed in invasive HNSCC was playing a significant role in metastasis. It is possible that in primary tumors caspase-8 predominantly contributes to apoptosis and therefore can prevent metastasis, but in those tumor cells that escape apoptosis (i.e., are resistant to apoptosis), caspase-8 may be contributing to migration and metastasis. In our study, we found that in primary tumors with no LNM (i.e., Tu^−met) higher expression of caspase-8 correlated with better disease-free survival and overall survival, however, in tumors with LNM (i.e., Tu^+met) higher caspase-8 expression significantly correlated with worse disease-free survival and overall survival. Similar results were also generated when we analyzed the impact of caspase-8 and DR5 combination.Higher levels of both caspase-8 and DR5 in HNSCC with no LNM (Tu^−met) was significantly associated with better disease-free survival and overall survival, but was significantly correlated with poorer disease-free survival and overall survival in HNSCC with LNM (Tu^+met). Thus, it is plausible to speculate that caspase-8 (as well as DR5) in primary HNSCC without LNM (Tu^−met) may be predominantly associated with its pro-apoptotic function and thus higher caspase-8 or caspase-8 plus DR5 provides protective advantage against cancer and correlates with better survival.Whereas in HNSCC with LNM (Tu^+met) which are resistant to apoptosis, caspase-8 and even DR5 may primarily exert their non-apoptotic function, i.e., activation of PI3K and promotion of migration, and thus higher caspase-8 expression negatively impacts patient survival. A limitation of the current study is the existence of potential bias due to usage of selected patient populations and retrospective design although no tumor pretreatment was the major criteria for tumor selection. Nonetheless, our interesting findings warrant further study to demonstrate the precise role of caspase-8 as well as DR5/caspase-8 pathway in regulation of HNSCC metastasis.

In summary, our IHC analysis of caspase-8 and DR5 in HNSCC suggest that a loss or downregulation of DR5 expression and possibly caspase-8 expression may be associated with more metastatic tumors and a loss of differentiation. The overall high expression of DR5 and caspase-8 in both primary and metastatic HNSCC suggests that DR5 or caspase-8 may be a good target for therapy of HNSCC. Various agonistic DR5 antibodies such as Conatumumab (AMG655), CS-1008 and Lexatumumab have been currently tested in cancer clinical trials either as a single agent or in combination with other therapeutic agents. Early clinical trials of these agents have established the safety of the approach and showed proof-of-concept antitumor activity [52]. Thus, our current findings warrant further study on targeting DR5 for potential treatment of HNSCC.