Information concerning miRNA expression and deregulation in MM is still lacking. Based on the hypothesis that intronic miRNAs are coordinately expressed with host transcripts [11
], we sought to identify miRNAs potentially deregulated in MM by focusing on those mapping within the intronic regions of host genes that were significantly differentially expressed in a representative panel of HMCLs profiled with U133A gene expression chips.
Following this approach, we identified six genes containing intronic miRNAs; all but one showed appreciable expression levels. For three miRNAs, miR-335, miR-342-3p, and miR-561, we demonstrated coordinated expression with their cognate protein-coding genes MEST
, and GULP1
; conversely, we did not find correlated expression of miR-628 and miR-559 and their host genes CCPG1
. Notably, miR-335, miR-342-3p, miR-561, and miR-559, but not miR-628, are sense oriented with respect to the corresponding host gene. This finding is in agreement with the evidence that intronic miRNAs are usually oriented in the same direction as the pre-mRNA, and thus could be under the control of the same regulatory motifs as their host genes and processed from the same primary mRNA transcripts regulated by PolII. On the other hand, our data may also support the previous suggestion that even miRNAs in a sense orientation to annotated genes (e.g., miR-559) may have their own regulatory motifs that can be regulated by either PolII or PolIII [12
The coregulation of these three miRNA/host gene pairs was also found in primary MM tumors. Specifically, our data showed that miR-335, miR-342-3p, and miR-561 were overexpressed in a fraction of the pathological samples with respect to normal plasma cells, without correlation to any of the known genetic alterations frequently found in MM; this finding may provide further evidence of the genetic complexity of this disease. In addition, despite the fact that miR-335 deregulation in melanoma and ovarian carcinoma was reported to be concordant with CN gain [39
], neither miR-335 nor miR-342-3p and miR-561 expression levels were significantly correlated with their corresponding locus CN in our panel of HMCLs tested by SNPs arrays. This finding indicates that DNA CN alterations may not be a critical factor affecting expression of these miRNA/host genes in MM, and suggest the occurrence of epigenetic mechanisms.
Although determining the precise contribution of each miRNA to myelomagenesis was beyond the scope of this investigation, some evidence supports the potential involvement of deregulated miRNAs/host genes in myeloma. Interestingly, miR-335 and miR-342-3p were recently reported to be involved in human cancer; depleted expression of miR-335 was found to be associated with metastatic processes in human breast cancer. Specifically, miR-335 was shown to suppress metastasis by altering cell morphology and decreasing cell motility, which would limit metastatic migration [40
]. Notably, we found that the fraction of primary myeloma samples overexpressing miR-335/MEST
also showed upregulation of genes implicated in actin polymerization and microtubule-based processes. In agreement with these data, bioinformatic tools predicted that a set of genes involved in actin cytoskeleton organization and biogenesis (DAAM1
, and RASA1
) were putative miR-335 targets (additional file 6
). Therefore, one can speculate that miR-335 may be involved in MM in the physiological mechanisms reported to be altered in breast cancer, possibly influencing different processes such as plasma cell homing into the bone marrow and/or interactions with the bone marrow microenvironment. Notably, MM patients overexpressing miR-335/MEST
also upregulated genes promoting cell proliferation, a finding apparently in contradiction with the function of miR-335 as an apoptosis permissive factor and cell cycle suppressor, as demonstrated in cortical neurosphere cultures [41
]. One possible explanation might be that the cellular context and cooperation among multiple miRNAs play a key role in the final biological effect of miRNA. Alternatively, one cannot exclude a specific effect of MEST
overexpression itself in promoting cell proliferation.
With regard to miR-342-3p, it is usually expressed in a variety of human tissues, together with its host gene EVL
. Notably, it is specifically silenced in the majority of colorectal cancers following methylation of CpG islands located upstream of EVL
, although the functional consequences of its silencing in carcinogenesis remain to be elucidated [42
]. In addition, miR-342-3p was substantially downregulated in patients with primary myelofibrosis, polycythemia vera, or essential thrombocythemia [43
], and specifically upregulated in acute promyelocytic leukemia cell lines during retinoic acid-induced differentiation [44
]. Intriguingly, EVL is an actin-associated protein that is involved in a variety of processes related to cytoskeleton remodeling and cell polarity [45
]; among miR-342-3p predicted targets, we recognized genes involved in actin cytoskeleton organization and biogenesis (FHL3
and, again, RASA1
). One can speculate that miR-342-3p and EVL
deregulation may target plasma cell homing into the bone marrow and/or interactions with the bone marrow microenvironment, much the same as for miR-335.
Finally, there is no information concerning the possible role of deregulated miR-561 and its cognate host gene GULP1
(which codes for an evolutionarily conserved adaptor protein required for efficient engulfment of apoptotic cells by phagocytes [46
]) in normal or tumor cells. Because of the high frequency of GULP1
overexpression in MM (34%) compared with normal plasma cells, both miR-561 and its cognate host gene warrant further investigation.