Commentary on microRNA Fingerprint in Human Epithelial Ovarian Cancer

miRNA signatures might seem only descriptive studies, limited to the discovery of new molecular portraits, which complete the puzzle along with gene expression information. However, all the literature up to date has demonstrated that miRNA profile analyses represented the starting point for following biological and translational studies, including the employment of miRNAs as prognostic and predictive biomarkers, as well as tools for innovative therapeutic approaches. This is actually true for all human neoplasms, but even more important for very aggressive cancers, which still represent a medical issue, as ovarian cancer is the eighth most common cancer in women and still the most lethal gynecologic malignancy. Efforts focus on identifying biomarkers that may aid in early diagnosis and reduce mortality, as well as on characterizing therapeutic targets with the aim of circumventing chemoresistance and prolonging survival at advanced stage disease.

In this context, miRNAs have provided a significant contribution. Studies concerning tumor progression in ovarian cancer have indeed shown different miRNA profiles along the course of the disease. miR-200c has been found elevated in the effusion specimens, whereas miR-145 and miR-214 were reduced. Other reports described a different miRNA profiling in primary tumors versus metastatic lesions; however, the inconsistency of the results underlines that further studies are needed to support their use as diagnostic tools.

Concerning the impact of miRNA expression on disease prognosis, data are not always consistent in different studies. miR-200 family, for example, has been associated either with poorer or better outcome, as reported in Marchini and colleagues’ studies (4). The role of this miRNA family should probably be deeper elucidated to clarify the meaning of the expression in a series of human specimens in ovarian cancer but also in other tumor types, such as breast cancer. Despite the overexpression in tumor samples, miR-200 has been clearly defined as an oncosuppressive molecule induced by p63 and p73 and able to counteract the epithelial-to-mesenchymal transition and the process of angiogenesis (5).

More recently, Bagnoli and colleagues analyzed the miRNA expression profiles in three large cohorts of samples collected at diagnosis, defining and validating a 35-miRNA–based predictor of risk of ovarian cancer relapse or progression (MiROvaR; ref. 6).

Despite the relevance of the prognostic significance of miRNA signatures, the possibility that miRNAs can play a role in the responsiveness to therapies has an even more significant impact on disease management, supporting two important applications, as predictive biomarkers and as therapeutic tools. Indeed, the high mortality of advanced stage ovarian cancer still depends on the poor responsiveness to chemotherapy. We described a panel of 23 miRNAs associated with chemoresistance (7), with an interesting role exerted by miR-484 in the regulation of angiogenic factors.

Despite these interesting reports, tissue-based markers still require an invasive procedure to obtain samples, thus determining a difficulty in the translation of these discoveries to the clinics. One of the best ways to perform early diagnosis, aid prognosis, and predict therapeutic response would be by using diagnostic or prognostic serum biomarkers. Unfortunately, there are not many reliable serum biomarkers for ovarian cancer currently used in the clinic. Circulating miRNAs, more stable and easily detectable, seem to be very promising new tools. Concerning ovarian cancer, miRNAs can be indeed detected in sera and exosomes of patients affected with cancer, and a consistency with the deregulation reported in tumor tissues has been observed. However, also in this case, the large heterogeneity of the results, based on differences in the samples analyzed and technology applied, highlights that additional studies are needed to define predictive and reliable miRNA signatures that can find a clinicopathologic application.

After our first observation that miRNA genes are located in fragile regions of the genome, in 2006, Zhang and colleagues demonstrated that miRNAs exhibit high frequency of genomic alterations in human breast and ovarian cancers and melanoma, with some shared and other peculiar features (2). Interestingly, our expression data were concordant but not completely overlapping with the above cited genomic analysis, thus suggesting the existence of different regulatory mechanisms.

Indeed, in our 2007 report on miRNA dysregulation in human ovarian cancer (1), we were one of the first groups focusing our attention on a possible epigenetic regulation of miRNA expression, in particular on the modulation of promoter methylation as a mechanism to induce the expression of protumorigenic miRNAs as miR-21, or to silence oncosuppressive miRNAs.

In 2011, The Cancer Genome Atlas Research Network analyzed mRNA expression, miRNA expression, promoter methylation, and DNA copy number in 489 high-grade serous ovarian cancers and additionally performed exon sequencing in 316 of these tumors (8).

Another mechanism that has been analyzed to explain the miRNA dysregulation in ovarian cancer is the altered levels or presence of mutations in the main enzymes of the miRNA biogenesis machinery, such as Dicer and Drosha. As discussed above, a reduced expression has been globally observed in a large percentage of tumors and associated with large tumor stage and reduced overall survival. However, this evidence still needs to be confirmed among the different ovarian cancer histotypes.

Complementary mechanisms, including transcriptional regulation, may also cooperate in miRNA deregulation of ovarian cancer. An important example of miRNA transcriptional control in ovarian cancer is represented by miR-200 family, repressed by ZEB1 and ZEB2, and miR-34, known oncosuppressive miRNA induced by P53.

As anticipated above, many miRNAs we described in our original article in 2007 (1) have been later confirmed by independent studies and functionally linked to occurrence/progression of ovarian cancer. miR-205, for instance, despite the known role as oncosuppressive molecule in several human cancers, including breast cancer, seems to act as an oncomiR in ovarian cancer, promoting cell migration. This evidence is consistent with the overexpression in tumors versus normal tissues we first reported and raises the issue that the same miRNAs might exert opposite roles in different cellular contexts, consistently with their tissue specificity and the diverse pathogenesis of human neoplasms.

Many studies have also tried to manipulate the mechanisms responsible for chemoresistance by altering the levels of miRNAs. Also in this case, miR-200 comes out again as a predominant player, being associated with treatment response through the targeting of β-tubulin III, evidence later confirmed by independent studies. However, overexpression of miR-141 in ovarian cancer cells increased resistance to the platinum-based chemotherapy, thus underlining how the role of this miRNA family in chemoresistance is still debated. It seems that the diverse function of the miR-200 depends on the cellular location of the RNA-binding protein, HuR, to the 3′UTR of the mRNA of the class III β-tubulin (9).

Let-7a expression has also been linked to the chemotherapeutic response, where the beneficial impact of the addition of paclitaxel to platinum therapy was better in patients with low let-7a levels. Bagnoli and colleagues identified a cluster of eight miRNAs located on chrXq27.3 locus that were downmodulated in early relapsing, advanced stage ovarian cancer patients and associated with a reduced sensitivity to chemotherapy.

Our own recent data underlined an interesting role for miR-302b in response to cisplatin treatment, exerted through the regulation of a master cell-cycle controller, E2F1 (10). Unpublished data confirm these effects in vivo.

We are probably far from a concrete application of miRNAs as therapeutic tools in the human setting; however, the potential translation to the clinics is fascinating and based on increasing and robust preclinical evidence. Correction of miRNA alteration can be done either by using miRNA mimics (miRNA replacement therapy) to restore loss of function or by inhibition of the upregulated oncomiRs using antisense miRNAs (miRNA inhibition therapy).

The first miRNA-based anticancer drug is an miR-34a mimic (MRX34), which entered a phase I clinical trial in 2013 for hepatocellular carcinoma. As miR-34 is frequently downregulated in ovarian cancer, this compound would potentially be useful also for the treatment of this tumor type. For MRX34, a liposome-based method of delivery is used, but several different modifications have been tested for miRNA delivery, as well as for miRNA inhibition.

However, as miRNAs may affect simultaneously the expression of a multitude of genes with potential opposite consequences, the use of miRNAs as therapeutics should be introduced with great caution. Moreover, the delivery to the correct cell type or tissue is an important aspect of efficient miRNA mimicry to prevent unwanted side effects. This said, targeting miRNAs rather than specific genes or proteins may be more effective, as they often finely tune entire pathways.

The interest in miRNAs and their involvement and possible use as markers in diagnosis, prognosis, and therapeutics in human cancer, including ovarian cancer, has expanded incredibly fast within the last decade, as we described for the first time an aberrant miRNA expression in this tumor subtype. However, the path from identifying this molecular portrait and biologically relevant miRNAs to the concrete application in clinics is still challenging. To obtain reliable and useful results, it is mandatory to perform and validate studies in larger and well-characterized cohorts.

A deeper understanding of miRNA function and their interaction with therapeutics is also still required. Nevertheless, the results obtained in this field of research are certainly exciting and encouraging, and miRNAs may have a great potential in revolutionizing cancer diagnostics and personalized medicine in the coming years.

No potential conflicts of interest were disclosed.