Re: Shi et al. Protease‐activated receptor 2 suppresses lymphangiogenesis and subsequent lymph node metastasis in a murine pancreatic cancer model. J Pathol 2014;234: 398‐409
Z. Rohan,T. Olejár,R. Matěj
DOI: https://doi.org/10.1002/path.4498
2015-05-01
Abstract:We read with great interest the article by Shi et al [1] in the November 2014 issue of The Journal of Pathology, which discussed the role of proteinase-activated receptor (PAR)-2 in the growth and metastasis of pancreatic cancer in a murine model. In addition to the well-known role of PAR-2 in the primary proliferation of various cancer cells [2,3], including pancreatic cancer [4], the results presented highlighted convincingly the importance of stromal host receptors relative to local progression in the orthotopic pancreatic cancer model. The authors recorded an increase in the average volume of primary tumours in wild-type animals compared with PAR-2−/−knockouts. However, they found that metastatic spread to abdominal organs and lymph node metastasis was more extensive in PAR-2−/− knockout animals than in wild-type animals. The authors noted the importance of the different roles of PAR-2 in metastasis, which vary depending on the model of metastasis (ie the important role of the host receptor). However, while a similar receptor (PAR-1), which is activated by thrombin, is (1) involved in cancer cell proliferation and invasion, (2) generally considered pro-carcinogenic, and (3) unambiguously considered to be a pro-metastatic factor [5], the same level of certainty regarding these events cannot be extended to PAR-2. Nevertheless, even though the data set describing this receptor in different models and/or experimental settings is small, the available results provide a good indication of it role. No differences in the size and number of lung metastases in the murine melanoma model (using injections into the tail vein) were described by Camerer et al [6]. In contrast, our team recently published data that showed larger primary B16 melanomas growing subcutaneously in PAR-2−/− knockouts, while distant metastases (mainly to the lungs) were significantly more frequently recorded in wild-type controls [7]. Additionally, survival was prolonged in PAR-2−/− knockouts. The available data make it clear that the absence of host PAR-2 leads to different outcomes depending on (1) primary tumour type (melanoma versus pancreatic cancer), (2) primary tumour location (skin versus pancreas), and (3) the model of metastatic induction (primary tumour growth versus intravenous application). Based on the three variations mentioned above, a variety of mechanisms for PAR-2 action have been discussed, including (a) antigen presentation in dendritic cells (DCs) [8]; (b) the role of PAR-2 in DC maturation [9]; and (c) the role of tissue factor/factor VIIa PAR-2 in tumour proliferation [10] and in capillary endothelia in distant locations [6,11]. Recently, Shi et al suggested an additional role for PAR-2 with regard to lymphatic vessel maturation [1]. To evaluate the influence of PAR-2 on lymphatic vessel growth, Shi et al utilized both artificial [agonist peptide (AP)] and more natural, in vivo-generated cell culture extracts to activate PAR-2. It is worth pointing out that ligand pleiotropy and receptor redundancy are characteristic features of PARs that can adversely affect virtually all PAR-based experiments and, as the authors noted in their conclusion, often lead to ambiguous results and further complicate therapeutic applications of PAR-based research. Together with PAR-2 heteroand homo-dimerization [12], and probably also transactivation [13], these features lead to the extreme variability of effects observed under different physiological, pathological, and research conditions. In vitro or ex vivo research in this field allows only limited use of well-defined proteases or APs. Nonetheless, these can help to assign a specific role to the receptor after activation by a given ligand. In an organism where complex systems of interacting proteases and anti-proteases affect virtually all receptors, PARs (namely PAR-2 in the study by Shi et al) are exposed to the influence of various proteases, including both activating and inactivating proteases, as well as several different types of protease inhibitors. Moreover, serum-activating and -inactivating proteases can also significantly affect tissue phenotype from the perspective of PAR-2 activation status. Using complex mixtures of tissues or cell culture extracts can more reliably simulate in vivo conditions; however, this makes assessing the real composition and individual protease concentrations, and their receptor affinities, problematic under physiological conditions. When combined with complex steric interactions between ligands (both natural, ie proteases, and artificial, APs) and membrane receptors, there is a potential for ‘biased’ signalling of the PAR-2 or modified downstream signalling interactions in general. Surrogate markers for PAR-2 (eg ERK1, 2 activity, used by Shi et al) and other markers of PAR activation usually reflect only one (often the dominant) signalling axis, but do not reflect other potentially co-activated cascades that may potentially modify, or even oppose, the effect of the original stimulus.