Metastatic UM still shows a high mortality rate, despite improvements in an efficient treatment of the primary tumor, in most cases. Hence, the identification and better understanding of the mechanisms involved in UM dissemination and colonization, together with the identification of new potential therapeutic targets and tools, is urgent. We have previously shown that both ADAM10 and c-Met are significantly more expressed in primary UM of patients undergoing liver metastases during follow-up, compared to those with good prognosis [14
]. Interestingly, the expression of these two pro-metastatic genes is highly correlated [14
], suggesting the possibility of a common mechanism regulating their gene expression. MiRs have been reported to contribute to cancer progression through the modulation of oncogenes or tumor suppressors. Alteration of several miRs has been described as potentially pathogenic in UM. Recently, miR181b was shown to be overexpressed in UM and to promote cell cycle progression in UM cells [48
]. On the other hand, other miRs, such as miR9 [49
], miR182 [50
], miR34b and miR34c [41
], may act as tumor suppressors in UM. In addition, miR144 [36
] has been proposed as a tumor suppressor in UM by targeting the c-Met oncogene.
The present study shows that miR122 and miR144 are expressed at low levels in three different cohorts of primary UM and also in a panel of UM cells. Low levels of miR122 and miR144 in primary tumors may contribute to the up-regulation of both pro-metastatic genes ADAM10
and may function as tumor suppressors in this disease. Indeed, we found that low expression of miR122 and miR144 was associated with higher expression of both ADAM10
in the UM TCGA dataset. Furthermore, extremely low levels of expression of miR122 and miR144 were seen in UM cell lines, which are characterized by elevated levels of ADAM10 and c-Met [14
]. Transfection of miR122 and miR144 mimics in 92.1 and UPMM3 cells leads to a reduction of ADAM10
mRNA and protein expression. Consistently with these data, we observed a reduction of proliferation, migration and shedding of c-Met in UM miR-transfected cells. In addition, an increase in G0/G1 phase of the cell cycle was present in both cell lines, while only 92.1 cells showed a concomitant increase of apoptosis. MiR122 has previously been reported to inhibit growth of hepatocellular carcinoma cells in vitro and in vivo [22
] through the inhibition of different genes, including ADAM10
. MiR122 also targets the 3’UTR of c-Met and its overexpression reduces proliferation and invasion of hepatocellular carcinoma cell lines [31
]. The present work shows that both miR122 and miR144 overexpression inhibit ADAM10 and c-Met expression and have an anti-migratory and an anti-proliferative effect on the 92.1 and UPMM3 UM cells, although in the latter the effects are detected only at later time points. These effects seem to be related to a direct targeting of miR122 and miR144 on the 3’UTR of both c-Met
. The effects of miR122 on ADAM10 and c-Met expression and on the proliferative and migratory activity of UPMM3 cells was less evident. The lower efficacy of miR122 in the UPMM3 model may relate to the low expression of this miR upon transfection, a finding that was consistently reproduced in three different transfection experiments. However, this low expression does not appear to be related to a generally low transfection efficiency of UPMM3 cells, as miR144 is well expressed upon transfection in these cells. In addition, transfection of fluorescent double strand RNA oligonucleotides produced comparable results in 92.1, OMM2.5 and UPMM3 cell lines (Figure S8
). Recent findings indicate that endogenous transcripts control specific miRNA levels in mammalian cells by target-directed miRNA degradation [51
]. One may speculate that similar mechanisms of enhanced degradation may inhibit endogenous or exogenous miR122 expression and activity in UM cells, particularly in the UPMM3 model. Differences in the functional effects of the two miRs may be also related to a different efficiency in reducing the targeted genes required to produce a functional effect, or to other possible additional targets of miR122 and miR144 not considered in this study. Chromosomal aberrations and BAP1
mutation in UPMM3 cells may influence the onco-suppressive activity of miR122 and miR144. However, the analysis of miR expression in UM TCGA [6
] found a cluster of miRs associated with monosomy 3 and BAP1
mutation, but miR122/144 are not present in this cluster, and thus may not be regulated by them.
Finally, the more efficient ability of miR144 to block cell proliferation, stimulate apoptosis and arrest cell cycle implies an important role for c-Met and ADAM10 in these biological processes in UM. If miR144 [36
] has already been proposed as a tumor suppressor by targeting the c-Met
oncogene, its role on targeting ADAM10
has not been demonstrated until now in UM. The effects of miR144 targeting both c-Met
are evident in both disomic and monosomic cells, with a prevalent effect on disomic cell line.
Our data provide further information on the influence of ADAM10 and c-Met in the biology of UM and clarify some of the molecular mechanisms related to the modulation of miRs targeting the pro-neoplastic genes ADAM10
. The possibility to counteract pro-metastatic activity of onco-miRs has been investigated by several studies using different types of delivery tools [52
]. Local delivery of miR122 and miR144 may be considered to restore tumor suppressive activity in UM and deserves to be further investigated in xenograft animal models of metastatic disease.