Introduction

Small noncoding RNAs, such as microRNAs (miRNA), are 21–25 nucleotides in length and since their discovery in 1993 have emerged as disease-associated biomarkers. MiRNAs are stable in biological fluids and thus, in theory, represent good targets for diagnostic and prognostic assays. Detection of miRNA polymorphisms and miRNA variants has been identified in patients with a predisposition to cancer development [1]. MiRNAs are well established as regulators of tumorigenesis and have different stages of expression during neoplasia. MiRNA is generated from long primary RNA polymerase II-derived transcripts through two processing reactions: nuclear Drosha cleavage and cytoplasmic Dicer cleavage, which forms mature miRNA [2]. Broadly, they control gene expression by binding to the 3′ untranslated region of target messenger RNAs (mRNAs) resulting in changes such as mRNA destabilization or translational repression. Recent studies have found the mechanism to be localized during the initiation step of mRNA translation. See [2] for an in-depth review of miRNA processing and regulation.

MiRNA expression can be altered in malignancy and plays a significant role in tumor progression, including tumorigenesis, invasion, metastasis and angiogenesis. Depending on its method of action toward a targeted protein, miRNAs may act as oncogenes when overexpressed or upregulate oncogenes by downregulating tumor suppressor genes [3]. The increasing importance of understanding the functional significance of miRNAs has made its role in gene expression a well-studied field (Tables 1, 2). Additionally, miRNAs may be tissue specific and thus have the potential to serve as unique identifiers of tumor type and origin [4]. In addition, miRNA polymorphisms have strong implications in understanding the prognosis and progression of disease. MiRNA polymorphisms can include SNPs, chromosomal variation and epigenetic defects that may interfere with miRNA-mediated regulation and can be found not only in the miRNA target gene, but also in primary, precursor and mature miRNA sequences as well as the genes involved in miRNA biogenesis and miRNA promoters [5].

Table 1 Upregulated osteosarcoma
Full size table
Table 2 Downregulated osteosarcoma
Full size table

Osteosarcoma (OS) is a leading cause of cancer-related death in adolescents, most commonly arising in the metaphyses of long bones [6]. The histologic appearance of OS can be quite diverse, including various lineages of differentiation and unusual variants (Fig. 1). Nevertheless, OS is a very aggressive tumor, which left untreated is universally fatal. With modern multimodal treatment, the 5-year survival rate of OS patients has improved to 60–70 % [7, 8]. However, for patients with metastatic disease outcomes remain poor [7, 8]. Deregulation of miRNA sequences has been recognized to have a role in osteosarcomagenesis. The implications for treatment, prognosis and the basic biology of OS will be discussed below as well as a comparison to chondrosarcoma (the second most common primary bone malignancy).

Fig. 1
figure 1

Histopathologic diversity of osteosarcoma. Although the majority of osteosarcoma variants have a similar prognosis, OS tumors have significant anatomic and histologic diversity. a Typical appearance of osteoblastic OS, with filigree type neoplastic bone. b Typical appearance of chondroblastic OS, with large areas of predominant chondroid matrix. c Typical appearance of fibroblastic OS, often demonstrating large areas of fibroblastic differentiation with inconspicuous osteoid. d Giant cell-rich OS, an unusual histologic appearance with numerous giant cells. e Parosteal OS, demonstrating typical dual fibro-osseous differentiation. f Telangiectatic OS, demonstrating areas of hemorrhage, fibrin and markedly pleomorphic tumor cells

Full size image

MiRNA and the tumor suppressor p53 in osteosarcoma

The p53 tumor suppressor lies at the nexus of cellular pathways that sense DNA damage, cellular stress and improper cell proliferation (see [9] for a review). Sporadic osteosarcoma demonstrates deletions or loss of heterozygosity in the p53 locus (17p13.1) in ~40 % of tumors [10–12]. Moreover, p53 heterozygote mice are well known to spontaneously develop OS in ~25 % of cases (see [13] for a review). He et al. [14] identified the miR-34 family as direct transcriptional targets of p53, via the use of p53-deficient cell types. Moreover, they found that ectopic expression of miR-34 induces cell cycle arrest across primary tumor cells and cell lines. Concordant with this, miR-34 downregulates a group of genes associated with the promotion of cell cycle progression. Collectively, these findings indicated that the miR-34 family acts as tumor suppressors in concert with p53 [14]. See also [15, 16] for a more detailed discussion of the miR-34 family in cancer development.

Another investigator examined the specific function of miR-34s in OS cell biology [17]. Here, He et al. examined the expression of miR-34 in 117 primary osteosarcoma samples and the in vitro roles of miR-34. Consistent with its presumed role as a tumor suppressor, miR-34 expression was decreased across tumor samples, although minimal deletions or epigenetic inactivations were observed. Consistent with the known induction of miR-34 expression by p53, those OS cell lines without p53 expression did not show miR-34 induction in the context of genotoxic stress (such as the p53 null SAOS2 cell line). Moreover, miR-34 target gene expression appeared to be partially dependent on p53 expression. Finally, miR-34 was found to induce cell cycle arrest in OS cells in vitro [17]. Thus, the miR-34 appears to have tumor suppressor properties in OS as well as other malignancies, in accordance with p53.

MiRNA and oncogenes in osteosarcoma

Another area of importance in miRNA regulation is the expression of oncogenes, which may affect tumor progression and the metastatic dissemination of sarcoma cells. Many osteosarcoma miRNAs regulate transcription factors that have been identified as inducers of pro- or anti-metastasis in steps of metastatic dissemination by the gain of migratory capacity and invasiveness or angiogenesis and resistance to apoptosis. The transcription factor TWIST promotes cell migration and invasion in synovial sarcoma and OS patients with resistance to chemotherapy [18, 19]. TWIST has been found to decrease OS cell survival against cisplatin by inhibition of the β-catenin and endothelin-1/endothelin-A receptor pathways [20, 21]. Zhou et al. [18] found that miR-33a correlated with clinical progression of OS, via downregulation of TWIST and reduced cisplatin-induced OS cell apoptosis in vitro.

MiRNA oncogenes have also been differentially expressed in OS as compared to other malignancies. Among these, using miRNA expression profiling of human OS cells, Maire et al. [22] found that miR-126 is overexpressed in human OS specimens. miR-126 has been reported to repress SPRED-1 (Sprouty-related, EVH1 domain-containing protein-1) expression, an inhibitor of MAPK (mitogen-activated protein kinase) signaling, and stimulate VEGF (vascular endothelial growth factor)-induced angiogenesis, thereby enhancing cell migration, proliferation and survival [23]. Other miRNAs known to have inhibitory effects on VEGF expression were conversely found to be downregulated among human OS samples [24, 25]. For example, miR-410 and miR-29b, known to repress VEGF expression [26], were downregulated in OS [24, 25]. In line with these observations, forced overexpression of either miR-140 or miR-29b led to reduced VEGF expression in SAOS-2 and MG-63 OS cell lines [24, 25]. Therefore, miRNAs may have a role not only in initiation of tumorigenesis but also in tumor vascularity and metastatic spread.

BCL-2 (B cell lymphoma 2) is a proto-oncogene that regulates the apoptosis pathway and increases survival by increasing resistance to apoptosis [27]. Multiple miRNAs have been found to regulate BCL-2 expression in osteosarcoma as well as other neoplasms; however, BCL-2 expression alone has not been found to correlate with osteosarcoma prognosis [28, 29]. The miR-16 family (miR-16 and miR-195) has been found to target BCL-2, and reduced expression of these miRNA correlates with resistance to chemotherapy [30, 31]. Other tumor suppressors of BCL-2 include miR-143, which was downregulated in OS cells lines and primary tumor samples, where BCL-2 was found to be the direct target of miR-143. When miR-143 expression was restored in MG-63 and U2OS OS cell lines, investigators found that the anti-tumorigenic effect of miR143 was validated in vitro through reduction in cell viability, promotion of cell apoptosis and tumor suppression [32].

MiRNA expression in osteosarcoma versus chondrosarcoma

Chondrosarcoma is the second most common primary malignant tumor of the bone. Unlike OS, chondrosarcoma is a malignancy of adulthood and older age and occurs most commonly in the pelvis (although any bone formed by endochondral ossification may be involved) [33]. Much like OS, differential expression and regulation of miRNA in chondrosarcoma is an area of recent investigation [34]. Studies have identified multiple similarities in the expression profiles of miRNA in chondrosarcoma and OS (summarized in Table 3). Probably the best example is the tumor suppressor miR-100, which is downregulated in both chondrosarcoma and OS tumors. This was determined from both primary human tissues of chondrosarcoma [35] and OS [36, 37], as well as multiple chondrosarcoma cell lines including SW1353 and OUMS-27 [38]. In addition, the forced downregulation of miR-100 was found to result in highly aggressive cell lines and cisplatin resistance, in vitro using cisplatin-resistant chondrocyte clones (CDDP CR1, CR2 and CDDP RP) [35] and in OS cell lines (HOS, MHM, OHS and OSA) [37]. In aggregate, miR-100 seems to have similar functional properties in OS and chondrosarcoma biology, although the secondary molecular mechanisms are not well understood. In fact, miR-100 may have tumor suppressor properties in various carcinomas as well, as low miR-100 expression has been observed to be a negative prognostic factor across breast, pancreatic and esophageal carcinomas [39–41]. Multiple other miRNAs are similarly dysregulated between chondrosarcoma and OS, including miR-134, miR-377 and miR-497 [38, 42–44]. However, their functional significance and/or distinct targets of gene expression have not been well defined.

Table 3 Comparison of chondrosarcoma and osteosarcoma miRNA expression
Full size table

Interestingly, several miRNAs show distinct transcriptional targets or even dichotomous changes between chondrosarcoma and OS. For example, even though miR-145 is upregulated in both chondrosarcoma and OS specimens, the target genes for the miR-145 are distinct. In chondrosarcoma cells, low miR-145 expression is associated with decreased SOX9 transcription, the major transcription factor associated with chondrogenic differentiation [45]. In OS cells, miR-145 expression was found to inversely regulate the gene ROCK1, an important activator of tumor cell motility [46–48]. This difference in regulation demonstrates that individual miRNA has multiple targets and how the impact of miRNA dysregulation may be tissue or tumor type specific [49]. Expression of miR-221 is another example of differences between OS and chondrosarcoma miRNA expression. miR-221 is a negative regulator of PTEN and the PI3K/AKT cell survival pathway and has been shown to be upregulated in OS [50] as well as carcinomas [51, 52], but downregulated in chondrosarcoma [38]. The significance of this difference in miR-221 expression profiles is not fully understood. However, it is important to note that both OS and chondrosarcoma exist as a spectrum from low-grade to high-grade sarcomas and have a large range of histologic appearances. Unfortunately in the available current studies, most sample sizes are too small to realistically take into account this diversity in tumor grade and subtype. Further research is needed to expand upon and verify these initial observations.

MiRNA as diagnostic and prognostic tool for osteosarcoma

As mentioned, miRNAs are relatively stable and may be analyzed in tissue, fluid or even paraffin-embedded samples [53, 54]. Thus, they represent attractive targets for diagnostic and prognostic assays. Differential expression of related miRNAs has been proposed as helpful in the diagnosis of certain malignancies, such as carcinoma versus lymphoma [22]. While miRNA detection may be of future utility in the diagnosis of skeletal sarcomas, this has not yet been born out in the scientific literature. In terms of prognostication of OS, various miRNA expressions have been found to regulate resistance to chemotherapeutic agents [18, 20, 21, 35, 50]. For example, overexpression of miR-140 in OS xenografts led to resistance to antifolates and fluoropyrimidine-based compounds [55]. MiR-140-induced chemoresistance was found to be caused by the regulation of HDAC4, which is involved in cell proliferation through decreased S phase and increased cell cycle arrest [55]. Additionally, overexpression of miR-503 and underexpression of miR-451 and miR-195 correlate with a poor response to chemotherapy and unfavorable prognosis in OS patients [56–58]. These initial associations between miRNA dysregulation and chemoresistance may be the introduction into more customized chemotherapeutic protocols for OS patients.

Future directions in the study of miRNA and skeletal sarcomas

MiRNA shows widespread dysregulation in skeletal sarcomas, including OS and chondrosarcoma. While not discussed here, Ewing’s sarcoma also demonstrates a unique miRNA profile [59, 60]. Although still in its relative early stages, miRNA profiling may be of utility in the diagnosis of skeletal sarcomas versus competing diagnoses, as well as to identify individuals at risk for resistance to traditional chemotherapeutic agents. While detecting and targeting miRNAs is still limited to early scientific investigations, it may well find its way into the future clinical management of skeletal sarcomas.