Background
Despite tremendous efforts in molecular and clinical oncology, pancreatic ductal adenocarcinoma (PDAC) still remains one of the deadliest cancers with a mortality rate almost equal to its incidence rate [
1]. Its dismal prognosis results from the lack of early diagnostic options, its highly aggressive growth and a resistance to current radio- and chemotherapeutic treatments [
2]. Thus, identification of new prognostic markers provides a strategy to uncover still unknown players driving PDAC malignancy and potentially to identify novel therapeutic targets.
TNF-related apoptosis inducing ligand (TRAIL) and its death inducing receptors TRAIL-R1 and TRAIL-R2 are promising candidates for the development of such novel targeted strategies. The rationale behind this assumption is the original observations that i) TRAIL induces apoptosis preferentially in tumor cells leaving normal healthy cells alive; ii) tumor cells usually express high levels of either TRAIL-R1 or TRAIL-R2 or both. These facts led to the development of different TRAIL formulations as well as agonistic TRAIL-R1- or TRAIL-R2-specific antibodies for treatment of human malignancies. However, relatively soon it has been recognized that many tumor cells are resistant to TRAIL induced apoptosis, the fact explaining the disappointing results from clinical trials [
3]. In addition, it became evident that TRAIL-R1 and TRAIL-R2 may respond to TRAIL - apart from apoptosis induction - with activation of different non-apoptotic signal transduction pathways like NF-kB, ERK1/ERK2, JNK, Src and AKT [
4], which can lead to the inhibition of apoptosis as well as to cell proliferation, migration and invasion. Most importantly, TRAIL receptor signaling may enhance cancer cell invasion and metastasis in vivo [
5‐
7]. Thus, therapeutic concepts are needed which combine TRAIL receptor targeting agents with agents sensitizing tumor cells and reducing the unwanted, non-death-inducing signaling of the receptors. The important concern regarding TRAIL-receptor based anti-tumor therapy is also the observed preference for the usage of the particular TRAIL death receptor for the transmission of the TRAIL-mediated signaling in tumor cells. Generalized predictions on main death receptor responsibility for apoptosis induction in given cancer types are difficult. It is widely accepted that the preference for either TRAIL-R1 or TRAIL-R2 is a cell type specific feature. A comprehensive compilation of tumor cell lines together with their preferences for usage of the particular TRAIL death receptor is provided by Roosmalen et al. [
8]. In PDAC cells, others and we have shown that regardless of the simultaneous presence of TRAIL-R1 and TRAIL-R2 at the cell surface, these cells use predominantly TRAIL-R1 when treated with recombinant TRAIL [
9,
10]. Consequently, TRAIL-R1-targeting variants of TRAIL and agonistic TRAIL-R1 specific antibodies were expected to have higher therapeutic effects in the treatment of PDAC than they in fact do [
3]. Interestingly, a recent report revealed that some PDAC cell lines show preference for TRAIL-R2 in inducing cell death [
11] pointing to an unexpected high diversity of TRAIL receptor preference even in the same tumor entity.
Of note, under physiological conditions TRAIL has been shown to be an important effector molecule in the tumor immunosurveillance [
12,
13]. On the other hand, malignant cells themselves can produce TRAIL and this may lead to an increase of their invasive and migratory properties [
7]. Thus, the expression levels of TRAIL receptors as well as the preference for the usage of TRAIL-R1 or TRAIL-R2 in TRAIL-induced apoptotic/non-apoptotic signaling may be an essential factor determining both, the tumor initiation and progression.
The biological responses to TRAIL are attributed to the function of TRAIL receptors at the plasma membrane. Interestingly, although the intracellular presence of TRAIL-R1 and/or TRAIL-R2 has repeatedly been noticed, only recently the question of biological relevance and eventually specific functions of intracellular receptors began to be addressed. Obviously, sequestration of the receptors in the cytoplasm or in the nucleus, frequently observed in cancer, could represent one of the strategies used by these cells to escape TRAIL-induced apoptosis. Indeed, such mechanisms have been proposed for both cytoplasmic [
14‐
16] and nuclear TRAIL receptors [
16‐
18]. More recently, specific function of nuclear TRAIL-R2 has also been uncovered [
19]. In the nucleus, TRAIL-R2 interacts with the microprocessor complex and impairs the maturation of the miRNA let-7. This leads to the increased levels of the malignancy promoting factors HMGA2 and Lin28B and enhances tumor cell proliferation in vitro and in vivo [
19]. Likewise, specific functions of cytoplasmic TRAIL death receptors have also been proposed lately. Concrete, in response to endoplasmic reticulum (ER) stress, as a part of unfolded protein response (UPR), cytoplasmic TRAIL-R1 and TRAIL-R2 both are able to aggregate and induce cell death [
20,
21].
Interestingly, although numbers of immunohistological studies addressed the issue of the impact of differentially expressed TRAIL death receptors in tumor and corresponding normal healthy tissue, the clinical relevance of TRAIL receptors present in particular intracellular compartment, cytoplasm or nucleus was analyzed only sporadically. For PDAC we reported recently that high levels of nuclear TRAIL-R2 correlates with worse prognosis for PDAC patients suffering from early stage PDACs [
19]. The level of TRAIL-R2 in the cytoplasm of PDAC cells, although significantly higher than in healthy tissue, did not correlate with any clinico-pathological parameter. To the best of our knowledge, no corresponding data for TRAIL-R1 are available so far. To fill this gap, we evaluated the clinical relevance of high levels of TRAIL-R1 in tumor tissues of PDAC patients with special emphasis to the possible differential impact of its cytoplasmic and nuclear localization.
Discussion
Identification of the prognostic factors related to survival of cancer patients represents a strategy to understand the molecular mechanisms driving tumor progression and therapy resistance, and may consequently support the development of novel therapeutic strategies. The expression levels of TRAIL-R1 and TRAIL-R2 were shown to be of prognostic relevance for different tumor entities. In addition to their localization at the plasma membrane, TRAIL-R1 and TRAIL-R2 are also found in the cytoplasm and in the nucleus of several cell types. Especially in tumor cells, diminished plasma membrane but enhanced intracellular presence of these receptors was frequently observed. Importantly, despite these observations and emerging evidences for distinct compartment-specific functions of TRAIL-Rs, their cumulative expression levels - regardless of their intracellular distribution - are mainly taken into account when immunohistochemical studies are evaluated. However, recent studies on the TRAIL-R2-expression in PDAC tissues suggest the necessity of considering the intracellular distribution of TRAIL-Rs. Thus, the expression levels of TRAIL-R2 may emerge as either a positive or a negative prognostic marker, depending on the subcellular distribution (plasma membrane vs. nucleus) [
19,
24].
Immunohistochemical studies described high cytoplasmic levels of TRAIL receptors in different cancer types, e.g. colorectal cancer [
25‐
27], breast cancer cell lines [
23,
28], renal cell carcinoma [
29], NSCLC [
30,
31], melanoma [
32,
33], PDAC [
19,
24], hepatocellular carcinoma [
14], and glioblastoma multiforme [
34]. Notably, the levels of just these intracellular receptors turned out to be of prognostic relevance in different tumor types [
16]. Intriguingly, whereas high intracellular levels of TRAIL-R1 mainly correlated with positive patient’s prognosis, increased levels of TRAIL-R2 often correlated with shorter patient’s survival (for review [
16]). These observations point to the existence of different, receptor-specific activities of cytoplasmic TRAIL death receptors and, in addition, suggest anti-tumor activities of intracellular TRAIL-R1, at least in some tumor entities.
In our present study, we identified cytoplasmic TRAIL-R1 as a positive prognostic marker for patients with PDAC. Interestingly, whereas overall staining intensity showed no prognostic relevance, the number of TRAIL-R1 positive cells per tumor turned out to be important for patient’s outcome. Noteworthy, we used an antibody for immunochemistry which is able to detect membrane expressed TRAIL-R1 as it has been shown before [
23]. Specifically, patients with tumors in which more than 80% of cells showed cytoplasmic TRAIL-R1 staining had significantly prolonged survival compared to patients whose tumors presented with less than 80% cells positively stained for TRAIL-R1. In line with these findings, a significant negative correlation between number of cells with positive stained TRAIL-R1 in the cytoplasm and tumor grading was found. These results indicate that loss of cytoplasmic TRAIL-R1 may support recurrent disease with more malignant phenotype.
Little is known about the origin and sub-cytoplasmic localization of intracellular TRAIL death receptors. Cytoplasmic TRAIL death receptors have been detected in Golgi vesicles [
32], endosomes [
32] and autophagosomes [
35]. In addition, their presence in soluble cytoplasmic fractions was also reported [
19].
Likewise, the function(s) of cytoplasmic TRAIL receptors is still not fully understood. Sequestration of these receptors in autophagosomes could act as a strategy by which tumor cells escape TRAIL-induced apoptosis [
17,
18]. On the other hand, internalization of TRAIL death receptors in response to TRAIL-treatment has been demonstrated, and may represent a part of TRAIL-induced signal transduction pathway [
36]. Recently, the importance of cytoplasmic TRAIL-R1 in inducing cell death as a consequence of unresolved unfolded protein response (UPR) has been demonstrated [
20,
21]. Noteworthy, in this case TRAIL-R1 mediated cell death is independent of TRAIL. Efficient UPR activation represents a characteristic feature of many human cancers. It allows the tumor cells to survive and adapt to adverse environmental conditions, promotes dormancy and also tumor growth, progression and therapy resistance. In this context, loss of cytoplasmic TRAIL-R1 would select cancer cells which are resistant to UPR-induced apoptosis and thus cells with more aggressive phenotype.
The TRAIL-TRAIL-R system has been shown to be of crucial importance in the tumor immune surveillance [
12,
13]. Cytoplasmic TRAIL death receptors may represent a reservoir of receptors, which upon stimulation localize to the plasma membrane and boost the primary response of cells to TRAIL. Since PDAC cells preferentially utilize TRAIL-R1 to induce apoptosis in response to TRAIL, loss of TRAIL-R1 could lead to an escape of immune surveillance and accelerate the recurrent tumor growth.
Alternatively, it is also possible that cytoplasmic TRAIL-R1, via direct protein-protein interaction, sequesters TRAIL-R2 in the cytoplasm thus inhibiting its malignancy-promoting nuclear functions. According to this scenario, cells, which have lost cytoplasmic TRAIL-R1, would also present with a more aggressive phenotype.
Which, if any of these potential cytoplasmic TRAIL-R1 functions, accounts for the obviously anti-tumoral functions of cytoplasmic TRAIL-R1 remains to be elucidated. The existence of further, still unknown functions of intracellular TRAIL receptors is very likely.
Likewise, the cellular pathways leading to the observed loss of TRAIL-R1 in a subset of PDAC cells are not known yet. Recently, several mechanisms negatively regulating the cellular levels of TRAIL-R1, but not TRAIL-R2, have been registered in cancer cells. Thus, hypermethylation of TRAIL-R1 promotor leading to an epigenetic silencing of TRAIL-R1 gene was detected in ovarian cancer cells [
37]. Furthermore, at the transcriptional level, negative regulation of TRAIL-R1 promotor by GLI3 as well as miR-25-dependent decrease of TRAIL-R1 levels were reported for cholangiocarcinoma cells [
38,
39]. At the post-translational level, specific degradation of TRAIL-R1 protein was described in breast cancer and melanoma cells. Here, membrane-associated RING-CH (MARCH)-8 ubiquitin ligase targeted TRAIL-R1, but not TRAIL-R2, for lysosomal degradation. Interestingly, plasma levels of miR-25 are significantly elevated in PDAC and evaluation of the levels of miR-25 together with MIC-1 and CA19–9 was shown to be able to distinguish between PDAC, benign pancreatic disorders and other GI cancers [
40].