9.1 Introduction

MicroRNAs (miRNAs) are small non-coding RNAs that repress gene expression by regulating the stability and translation of target messenger RNAs (mRNAs) [1]. Approximately 1% of genes in different organisms encode for miRNAs. However, in mammals, more than 60% of mRNAs are predicted to be regulated by miRNAs [2]. MiRNAs can thus target multiple mRNAs modulating gene expression programs in virtually every biological process [3].

This contrasts with the observation that only few miRNA mutants are associated with obvious developmental defects [4,5,6,7,8]. Instead, many miRNAs are thought to function to fine-tune gene activity providing robustness to gene regulatory networks [9,10,11]. This serves as a mechanism to ensure proper signaling responses in the face of environmental and genetic stresses, which are often the cause of disease. Consistent with that, and despite the small number of examples associated with strong loss-of-function phenotypes, miRNAs have been shown to play important roles in human pathologies, including cancer [12]. The complexity of their regulation and the high number of potential targets for each miRNA poses the challenge of elucidating the specific targets associated with miRNA-related phenotypes and diseases.

Although bioinformatic prediction tools have been helpful in finding potential miRNA-target interactions [13], these approaches predict many false positives [14]. Thus, to establish the important miRNA-mRNA interactions—which are relevant in different cellular contexts—putative targets need to be tested and validated in vivo. The use of animal models including worms (Caenorhabditis elegans), fruit flies (Drosophila melanogaster), zebrafish (Danio rerio) and mice (Mus musculus) has been crucial for the identification of miRNA functions in development and disease [15, 16]. We focus this review on the use of Drosophila as an in vivo model to study how miRNAs influence cancer.

9.2 MiRNAs in Human Cancer

MiRNAs are frequently dysregulated in human cancers; however, the specific functions of miRNAs in tumorigenesis are often elusive [17,18,19]. Aberrant miRNA expression levels are caused by chromosomal abnormalities, changes in transcriptional control, epigenetic changes or defects in the miRNA biogenesis machinery [20]. Oncogenic miRNAs, “called oncomiRs”, are often upregulated in cancer, and facilitate tumorigenesis and disease progression. On the contrary, “tumor suppressor” miRNAs counteract tumor growth and are frequently downregulated in cancer. In fact, miRNAs have been associated with various cancer-related processes such as DNA damage response, differentiation, angiogenesis, senescence, invasion and metastasis [18,19,20,21,22,23]. MiRNA signatures can be discriminated between different types of cancer [24, 25]. Thus, miRNAs can be used as diagnostic and prognostic tools in the clinic [26]. Moreover, miRNAs are considered as tools and targets for cancer therapy. In numerous preclinical studies, miRNA expression levels are modulated via the delivery of miRNA mimics, to replenish miRNAs with tumor suppressive functions, and antimiRs, to repress oncogenic miRNAs [27, 28].

Despite the progress in understanding the role of miRNAs in cancer, there is still a gap between the observations of widespread miRNA dysregulation in cancer and functional data proving causality of aberrant miRNA expression. Thus, in vivo animal models are key to dissect the underlying mechanisms of individual miRNAs in cancer.

9.3 Drosophila Tumor Models

Cancer is a genetic disease that involves the accumulation of mutations causing, among others, increased cell proliferation, reduced apoptosis and differentiation, and the activation of invasion and metastasis [29]. Mutations affecting “driver” genes, which provide the cells with the initial potential to form tumors, have been identified. However, the identification of genes that cooperate with known cancer drivers in malignancy remains a major challenge in cancer research [30,31,32].

Drosophila is emerging as a useful model to identify genes that cooperate with driver mutations in malignancy [33,34,35,36]. Despite the obvious differences between flies and humans, using Drosophila to model cancer has distinct advantages: (1) a reduced complexity due to a lower genetic redundancy and simpler biology; (2) a short generation time that, among other benefits, allows to quickly test hypotheses and generate large scale in vivo screens; (3) a powerful genetic toolkit for targeted gene modulation in a tissue and stage-specific manner. Moreover, many of the pathways that control key cellular and physiological processes are highly conserved. In fact, nearly 75% of human disease genes have orthologs in the fly [37]. Remarkably, several signaling pathways relevant to cancer such as the Hippo [38, 39], Notch [40], and Hedgehog pathways [41] were first described in Drosophila, contributing to our understanding of the molecular mechanisms underlying tumor formation [42].

Loss of tumor suppressors such as elements of the Hippo pathway, or activation of oncogenes like Ras or Notch, leads to benign tissue overgrowth in fly imaginal tissues [39, 43, 44]. However, combining Ras or Notch activation with mutants affecting the apical-basal polarity genes scribbled (scrib) , discs large (dlg) or lethal giant larvae (lgl), drives transformation into neoplastic tumors [45, 46]. These early screens showcased the utility of Drosophila models to study oncogenic cooperation in tumorigenesis. Interestingly, loss of apical-basal polarity is a key characteristic of malignancy in human cancers, and the Scrib/Dlg/Lgl polarity module is frequently dysregulated and is associated with tumor metastasis [29, 47].

Since these seminal works, studies in Drosophila have identified numerous oncogenes and tumor suppressors involved in oncogenic cooperation. Apart from key signaling elements controlling cell growth and proliferation, other factors regulating additional cancer traits have been described in fly tumors. These include apoptosis and compensatory cell proliferation, genome stability, metabolic reprogramming, actin cytoskeletal changes, inflammation, cell competition, the tumor microenvironment, and even angiogenesis [33, 34, 36].

According to miRBase release 22 (mirbase.org), the Drosophila genome contains 258 miRNA loci, which are processed to form 469 mature miRNAs [48]. To dissect the roles of miRNAs in tumorigenesis, methods to manipulate miRNA activity in a tissue-specific fashion without affecting the animal globally are required. To that end, resources which provide a genome-wide collection of miRNA overexpression and miRNA depletion (“miRNA sponges”) transgenes are available in flies [49,50,51]. Different approaches have been used to determine the roles of miRNAs in tumorigenesis where these tools have been central. These strategies—described in detail below—include tumor miRNA transcriptome profiling followed by functional analyses (illustrated in Fig. 9.1a) and screens for modifiers of tumor-related phenotypes (illustrated in Fig. 9.1b, c).

Fig. 9.1
figure 1

Overview of the different systems used to identify miRNAs that play a role in Drosophila tumors. (a) Representation of the miRNA transcriptome profiling studies in loss-of-lgl-induced tumors. (b and c) Representation of the different studies that identify miRNAs which enhance or repress tumor phenotypes

Full size image

9.4 MiRNA Expression Changes in Drosophila Tumors

MiRNAs are aberrantly expressed in human cancers and miRNA profiles are associated with tumor development and progression. However, functional analyses to examine these correlations remain limited. Drosophila provides a tractable system to perform this kind of analysis.

Expression of oncogenic RasV12 together with loss of tumor suppressive lgl in the imaginal tissues of Drosophila leads to the formation of malignant tumors [45]. The levels of approximately 11% of all mature miRNAs (51 miRNAs) in those tumors present robust changes [52]. Clonal depletion of lgl in the wing disc results in tumorous overgrowths. In contrast, tumors are not formed when lgl is specifically depleted in the dpp domain, a band of cells adjacent to the anterior-posterior boundary of the wing disc (hereafter referred as dpp > lgl-RNAi). These backgrounds served to assess the implications of miRNAs dysregulated in RasV12-lgl tumors. Among the 28 miRNAs upregulated, 10 induce tumorigenic overgrowth when expressed in dpp > lgl-RNAi discs. Furthermore, depletion of these miRNAs in lgl clones limits tumor growth. Similarly, the miRNAs downregulated in the RasV12-lgl tumors were tested for their potential to repress tumor formation in lgl clones. In that context, 11 of the 23 miRNAs downregulated, when expressed in lgl clones, repress tumor formation and restore normal tissue organization. Interestingly, the upregulated, tumor enhancing miRNAs bantam and miR-10, and the downregulated, tumor suppressive miRNA let-7 were also identified in other Drosophila tumor models and will be discussed below. Furthermore, nearly 50% of the miRNAs identified in this study are conserved and their human homologs are involved in various cancers [52]. This analysis shows that tumor formation goes hand in hand with miRNA dysregulation and, more importantly, that many of these differentially expressed miRNAs contribute to tumorigenesis.

lgl mutant brain and imaginal discs develop neoplastic tumors [53,54,55]. Transcriptome analysis also revealed widespread changes in miRNA expression [56]. To improve the temporal resolution of the miRNA profiles, this analysis was performed at three different time-points of tumor development. 10 miRNAs were dysregulated in all tumor stages analyzed. Amongst these, let-7 , miR-210, and miR-9a were downregulated—all of which have been functionally implicated in human cancers [57,58,59]. miR-9a was amongst the top downregulated miRNAs suggesting tumor suppressive functions. Consistently, overexpression of miR-9a limited the growth of lgl mutant wing discs [56]. At the stage when tumors were fully developed, bantam levels were highly enriched [56]. This is consistent with observations from another study where bantam levels are also upregulated in lgl−, scrib−, or brat− brain tumors [60].

9.5 let-7 and bantam: Old Dogs with New Tricks—in Cancer

let-7 and bantam were amongst the first miRNAs discovered and their analysis provided important insights into miRNA mechanisms [61, 62]. More recently, let-7 and bantam have been implicated in tumorigenesis in cooperation with cancer drivers.

9.5.1 let-7

In human cancers, let-7 is the most frequently downregulated tumor suppressor miRNA and repression of let-7 is correlated with poor prognosis [63]. Furthermore, let-7 has been shown to reduce proliferation and tumor growth in cancer cell lines [58, 64]. One of the tumor suppressive mechanisms used by let-7 has been elucidated in Drosophila and involves the let-7 target chinmo [65, 66], a transcription factor involved in tumorigenesis [67, 68]. In the Drosophila eye-antennal disc, clones mutant for the epigenetic silencing regulator Polyhomeotic generate neoplastic tumors [69, 70]. These tumors show malignant traits and continue to grow when transplanted into an adult wild-type fly [69]. On the contrary, tumors generated in the larval tissue are repressed after metamorphosis and eventually eliminated in the adult fly, revealing tumor suppressive signaling during larval–adult transition [65]. The steroid hormone Ecdysone, a crucial signal coordinating metamorphosis [71], induces the expression of let-7, and chinmo downregulation by let-7 is key for tumor eviction downstream of steroid signaling during metamorphosis [65, 72].

9.5.2 bantam in Tumors of Epithelial Origin

bantam was the first miRNA discovered in flies as an element that, when overexpressed, induces tissue growth [61, 73]. bantam is a developmentally regulated miRNA and its expression is controlled by different signaling pathways such as the Hippo [74, 75], Notch [76, 77], Dpp [78], and EGFR [79] signaling pathways. bantam promotes tissue growth by inducing cell proliferation and repressing apoptosis, two processes frequently dysregulated in cancer [29, 80].

Activation of the oncogene EGFR in the wing epithelium activates the Ras/MAPK pathway and induces tissue hyperplasia [79]. However, this does not cause malignancy; cooperating factors are required for cellular transformation and neoplasia. The miRNAs miR-10, miR-375 and bantam have been found to, individually, synergize with EGFR to facilitate neoplastic transformation [81, 82]. The bantam target Suppressor of cytokine signaling at 36E (Socs36E) plays a central role in this context. EGFR induces Socs36E expression. In turn, Socs36E antagonizes EGFR signaling [83, 84], which provides a negative feedback that limits the growth-promoting role of EGFR. Socs36E also dampens the JAK/STAT pathway [84] and JAK/STAT cooperates with oncogenic Ras in malignancy [85]. Thus, bantam-mediated repression of Socs36E inactivates this homeostatic feedback and drives neoplasia (Fig. 9.2a). In analogy to this, repression of the human Socs36 ortholog SOCS5, in combination with activated RAS, promotes colony formation in a cell transformation assay [81]. In agreement with these findings, subsequent studies in human cell lines showed that the transforming activity of oncogenic RAS relies on its ability to downregulate SOCS5/6 [86].

Fig. 9.2
figure 2

bantam is involved in positive feedback loops downstream of major growth regulatory pathways to reinforce their outputs via alleviation of inhibitory elements. (a) bantam represses the EGFR and JAK/STAT-inhibitory element Socs36E. Thus, In the wing epithelium, upregulation of bantam removes this homeostatic element and facilitates the formation tumors in cooperation with EGFR. (b) In the neuroblasts, bantam represses the Notch and dMyc inhibitor Numb. Notch overactivation leads to tumor-forming neuroblasts due to bantam-mediated depletion of Numb. (c and d) Hippo pathway-mediated bantam functions are possibly conserved in mammalian miR-130a. Both miR-130a and bantam act downstream of YAP/Yki to repress the YAP/Yki inhibitory elements VGLL4 or SdBP/Tgi respectively

Full size image

9.5.3 bantam in Brain Tumors

bantam is embedded in a similar regulatory loop in neuroblasts, neural stem cells in Drosophila. In those cells, Notch plays a conserved role coordinating self-renewal and differentiation [87, 88]. Notch promotes dMyc-dependent nucleolar and cellular growth, which is key for neuroblast self-renewal [89]. bantam controls neuroblast proliferation where it targets the Notch repressor Numb [90,91,92]. Notch hyperactivation induces the formation of cancer-stem-cell (CSC)-like neuroblasts that can initiate tumors [89]. bantam is required, downstream of Notch, for the formation of CSC-like neuroblasts and tumorigenesis [93]. In this context, via repressing Numb, bantam establishes a positive feedback that reinforces Notch signaling. Furthermore, bantam-dependent repression of Numb induces Myc signaling. Thus, bantam helps to maintain neuroblast homeostasis in two ways: a) by promoting Notch signaling, and b) by facilitating Myc-dependent nucleolar and cellular growth. Interestingly, overexpression of bantam is not sufficient to drive CSC-like formation and hence bantam acts to fine tune Notch-mediated neuroblast homeostasis [93] (Fig. 9.2b).

Although bantam is not obviously conserved in mammals, its functions likely are. bantam is proposed to functionally mimic mammalian miR-130a [94]. The mammalian Yki homolog YAP controls miR-130a expression and this regulation mediates over-proliferation and tumorigenesis. miR-130a targets VGLL4, which is an inhibitor of YAP [95, 96]. Thus, by repressing a negative regulator of YAP, miR-130a provides a positive feedback loop that is critical in YAP-mediated tumorigenesis. Intriguingly, analogous to this mechanism, the Drosophila VGLL4 homolog SdBP/Tgi is regulated by bantam. Thus, bantam and miR-130a share functional characteristics: both are involved in a conserved feedback that ensures robust Hippo pathway signaling in growth control and tumorigenesis [94] (Fig. 9.2c, d).

9.5.4 bantam and Invasion

Apart from promoting cell proliferation and repressing apoptosis, bantam has been proposed to repress cell invasion [97]. Hippo signaling appears to modulate invasion and epithelial-to-mesenchymal transition in the wing epithelium through JNK. Overexpression of the Yki target bantam impairs cell invasion upon yki-depletion, while other Yki targets such as diap1 and dMyc do not alter that. In that situation, reducing bantam also phenocopies the loss of yki. Rox8 has been identified as a bantam target involved in JNK regulation downstream of the Hippo pathway [97].

9.6 MiRNAs Affect Tumorigenesis in a Context Dependent Manner

Notch signaling promotes growth in various tissues and organs, and Notch hyperactivation in Drosophila epithelial tissues leads to hyperplasia [45, 69, 98, 99]. Overexpression of the Notch ligand Delta (Dl) in the developing eye results in mild tissue overgrowth [100]. This genetic background has been used to screen for genes that cooperate with Notch in malignancy and neoplasia [101]. By using this strategy, the conserved miRNAs miR-7 and miR-8 were identified as modulators of Notch-mediated growth and tumorigenesis [102, 103]. While miR-7 was found to cooperate with Notch, miR-8 functions as a tumor suppressor inhibiting Notch-mediated tumor formation.

9.6.1 miR-7

To identify miR-7 targets contributing to the synergism between miR-7 and Notch, RNAis depleting predicted miR-7 targets were coexpressed with Dl [103]. This showed that depletion of the Hedgehog receptor interference hedgehog (ihog) reproduces the miR-7/Dl overgrowth. Direct targeting of ihog by miR-7 was validated in vivo. Moreover, repression of core Hedgehog signaling components drives tumorigenesis in the Dl-overexpression background. Reciprocally, increase in Hegdehog signaling prevents miR-7/Dl tumorigenesis. This study unraveled an unknown tumor suppressive aspect of the Hedgehog pathway in Notch-driven tumors [103].

miR-7 also controls growth of the wing epithelium, as loss of miR-7 results in small wings with defects in cell size and the cell-cycle [104]. miR-7 targets the cyclin-dependent kinase inhibitor dacapo in the germline [105]. In agreement with that, reduction in the levels of dacapo or Notch rescues the wing defects associated with loss of miR-7 [104].

In human lung and skin cancers, miR-7 is upregulated and acts as an oncogene [106]. However, miR-7 tumor suppressive functions have also been reported in numerous cancers [107]. Interestingly, these also involve miR-7-dependent regulation of the Hedgehog pathway [108]. In fact, it is frequently observed that miRNAs may act as tumor suppressors in one context and as oncogenes in another [109]. Understanding these phenomena is especially relevant in miRNA-based cancer therapy. Another miRNA showing this context dependent behavior is the member of the miR-200 family, the Drosophila miRNA miR-8 .

9.6.2 The Tumor Suppressor Side of miR-8

Overexpression of Dl in combination with the epigenetic repressors pipsqueak and lola leads to the formation of malignant tumors—this characteristic phenotype has been referred to as “eyeful” [101]. Expression of miR-8 in the eyeful background reduces tumor growth and represses metastasis formation [102]. miR-8 overexpression in the wing and eye imaginal disc induces apoptosis and growth defects, phenotypes reminiscent of a reduction in the Notch ligand Serrate [110]. These observations led to the identification of Serrate as the miR-8 target responsible for its tumor suppressive role in the eye epithelium [102]. Importantly, the human Serrate ortholog JAGGED1 is also targeted by the miR-8 orthologs miR-200c and miR-141, and, similar to the Drosophila tumor, JAGGED1-mediated prostate cancer cell proliferation is inhibited by miR-200c and miR-141 [102].

The miR-200 family is frequently dysregulated in various types of cancer and has been functionally implicated in tumorigenesis and metastasis [111, 112]. Several studies support that miR-8/200 targets and functions are conserved between flies and mammals. miR-8 in flies and miR-200 in mammals inhibit epithelial-to-mesenchymal transition (EMT) by repression of zhf1/Zeb1 and Zeb2 [113,114,115]. Furthermore, the miR-200 family inhibits cell invasion by targeting regulators of the actin cytoskeleton; similarly, miR-8 modulates the actin cytoskeleton in the neuromuscular junction and the wing epithelium [116,117,118,119,120,121]. The pesticide component trans-nonachlor was shown to inhibit miR-141 in human melanocytic cells, facilitating malignant transformation [122]. In Drosophila, trans-nonachlor also represses miR-8. Strikingly, trans-nonachlor-induced downregulation of miR-8 is epigenetically inherited over multiple generations and leads to a loss-of-weight phenotype in the offspring [122].

Despite the fact that numerous studies demonstrate tumor suppressor functions of miR-200 miRNAs, clinical data on miR-200 levels are inconsistent and suggest cancer type or even subtype dependent roles [111, 123]. For instance, high miR-200 levels are associated with improved clinical outcome in ovarian, lung, renal, basal-like breast adenocarcinomas and certain colorectal cancers [123, 124]. However, high miR-200 correlate with worse outcome in luminal breast, certain ovarian and pancreatic cancers [124,125,126]. Furthermore, functional studies suggest that miR-200 family members can act as oncogenes by repressing the tumor suppressor PTEN [127, 128]. In contrast to the tumor suppressive function of miR-8 in the context of Notch-induced growth, miR-8 was shown to cooperate with the tumor drivers EGFR [129] and Yki [130] respectively, suggesting that the dual role of miR-8/200 was maintained between flies and humans.

9.6.3 miR-8 as an Oncogenic Factor

Multiple studies show that miR-8 limits tissue growth in imaginal tissues. miR-8 represses numerous genes required for normal growth including elements involved in cytokinesis, Hippo signaling, Wingless pathway, Notch signaling, insulin signaling and cytoskeletal regulators [102, 120, 121, 129,130,131] (Fig. 9.3a). Strikingly, when miR-8 expression is combined with oncogene activation (EGFR or Yki) the observed effect is the opposite and miR-8 fuels oncogene-driven growth resulting in the development of tumors (Fig. 9.3b).

Fig. 9.3
figure 3

miR-8 generally acts as a repressor of growth, but in some contexts, it promotes tumorigenesis. (a) List of miR-8 targets relevant in tissue growth. (b) The dual role of miR-8: it inhibits Notch-induced tumors; however, miR-8 facilitates tumorigenesis together with Yki or EGFR

Full size image

As bantam , miR-8 cooperates with EGFR in tumorigenesis. Coexpression of EGFR and miR-8 causes the formation of tumors and metastasis in Drosophila larvae. These tumors are heterogeneous and are composed of a mix of normal epithelial cells and giant polyploid cells. The latter show defects in epithelial polarity, which is a common trait in neoplastic tumors [132]. During tumor progression, giant polyploid cells get selected and, in late stages of tumor development, they stem the formation of metastasis. A closer analysis revealed the presence of apoptotic corpses within giant cells suggesting that these kill and engulf surrounding cells. Consistently, genetic suppression of engulfment in those discs (EGFR + miR-8) abolishes the formation of giant cells, tumor development and metastasis. Giant tumor cells hence grow at expenses of their surrounding neighbors in a process resembling cell competition—a cell-cell interaction process first described in Drosophila by Morata and Ripoll in the early 70s [133].

The miR-8 target gene peanut (pnut) plays a central role in the formation of these tumors. Pnut encodes a Septin that is required for normal cytokinesis [134]. pnut depletion is required for miR-8 + EGFR-driven tumorigenesis, as pnut overexpression in this background rescues tumor formation. Thus, via repressing pnut, miR-8 induces cytokinesis failure and thereby, in concert with EGFR, facilitates the emergence of polyploid cells that hijack cell competition mechanisms to propagate themselves and eventually form malignant metastatic tumors [129]. Cytokinesis failure has been described to be tumorigenic in mammals and it is proposed that approximately 40% of human tumors have gone through a round of gene duplication [135]. This work [129] provides a new example whereby defective cytokinesis is associated with the formation of malignant tumors.

One of the miR-8 targets required for normal growth is the oncogene Yki. miR-8, in addition to dampen Yki levels, acts as an oncogenic partner of Yki [130]. Reminiscent of the EGFR + miR-8 tumors, a subset of yki + miR-8 cells display aberrant ploidy, possibly due to defective cleavage as a consequence of pnut downregulation. Consistently, Yki can also induce neoplasia in discs with cytokinesis failure, generated via RNAi-mediated depletion of pnut [136]. However, yki + miR-8 tumors grew bigger in size than the yki-pnut-RNAi ones suggesting that additional Yki targets are involved in the formation of those tumors. This led to the identification of the growth repressor brinker as a miR-8 target gene contributing to tumor formation downstream miR-8 [130].

Taken together, these studies demonstrate a context-dependent impact of miRNAs in tumorigenesis, which is an important consideration for the application of miRNA therapeutics.

9.7 MiRNA Biogenesis Pathway and Tumorigenesis

The canonical pathway of miRNA biogenesis is a multistep process at the end of which the mature ∼22 nucleotide long miRNA is incorporated in the RNA-induced silencing complex (RISC), directing it to the target mRNA for post-transcriptional repression. A global depletion of miRNAs by alterations in the miRNA biogenesis machinery has widespread implications in human cancer [137].

MiRNAs are transcribed into long primary transcripts (pri-miRNAs), which are further processed by an RNase III enzyme, Drosha, to form miRNA precursors (pre-miRNA) [138, 139]. In the cytoplasm, pre-miRNAs are further processed [140] by another RNase III enzyme, Dicer-1 (Dcr-1), to form a duplex, which is subsequently loaded into the Argonaute-1 protein (Ago-1) [141, 142]. The duplex is then unwound, one of the strands discarded—the ssRNA guide strand is retained—and eventually the mature silencing complex is formed [143]. The exoribonuclease Nibbler has been shown be important for 3′ end trimming of longer miRNA intermediates produced by Dcr-1 [144, 145]. Nibbler has been recently associated with tumorigenesis in flies [146]. As discussed previously, lgl mutant tumors show broad changes in the miRNA transcriptome [52, 56]. Interestingly, lgl interacts with Fragile X protein (FMRP), and with Ago-1, both of which are involved in the miRNA biogenesis machinery [56, 147, 148]. These findings insinuate that changes in miRNA expression upon loss of lgl could be a direct consequence of a dysregulated miRNA biogenesis pathway. Further studies will be required to validate this interesting hypothesis.

9.7.1 The Proto-Oncogene dMyc Senses miRNA Levels

Dcr-1 mutants show a general depletion of miRNAs and this background has been used to study how cells with reduced miRNAs behave in different developmental contexts. Even though miRNAs control nearly every biological process, Dcr-1 mutant cells in the wing primordia are viable, differentiate normally, and do not show major patterning defects [149]. The most obvious outcome of global miRNA depletion is a reduction in the levels of the proto-oncogene dMyc. As a consequence, these cells are smaller in size and show reduced proliferation rates. Mechanistically, miRNA reduction results in an accumulation of the TRIM-NHL protein Mei-P26, which triggers proteasome-dependent degradation of dMyc [149]. At the same time, dMyc induces Mei-P26 as a means to buffer its own levels, which has been shown to be a mechanism to ensure epithelial tissue homeostasis [150]. bantam is one of the miRNAs that controls Mei-P26 levels, and overexpression of bantam in cells with reduced Dcr-1 restores dMyc levels and cell size defects [149]. Thus, dMyc appears to serve as a sensor of general miRNA levels in the cell.

Cell competition is a cell-cell interaction mechanism that senses cellular fitness and mediates the elimination of suboptimal cells in a tissue [151]. Cell competition is not only relevant in normal development and homeostasis, but in some contexts it also influences tumor formation [152]. Importantly, dMyc is a central mediator of this competitive interaction. In this scenario, cells with reduced dMyc are referred to as losers and are eliminated by cells with higher dMyc, reffered to as winners [153, 154]. Consistent with the reduction in dMyc, Dcr-1 mutant cells acquire the loser status and are eliminated from the wing primordia [149]. In sum, this study suggests that cells with reduced miRNAs are identified as less fit, which causes a reduction in dMyc and their consequent elimination by cell competition.

9.7.2 Proliferation Defects in Dcr-1 Mutant Stem Cells

Multiple studies demonstrate essential roles of the miRNA machinery for self-renewal in germline stem cells (GSCs) [155,156,157]. Loss of Dcr-1 in GSCs leads to defects in cell cycle control. In that context, the cell cycle regulator dacapo is increased and a reduction of dacapo partially rescues loss of Dcr-1-dependent cell cycle defects [155]. dacapo is regulated by miR-7 and miR-278, and loss of these individually in GSCs leads to cell-cycle aberrations [105]. Loss of Dcr-1 in GSCs of adult animals leads to defective stem cell maintenance—a phenotype mimicked by loss of bantam [158]. Similarly, in neuroblasts, depletion of Dcr-1 or bantam leads to a decrease in neuroblast number due to cell proliferation defects [92]. Interestingly, similar to Dcr-1 mutant GSCs [155], these cells display elevated dacapo expression levels. Since bantam also targets dacapo in GSCs [105], the bantam-dacapo axis might contribute to the proliferation defects observed in bantam mutants.

Similar to the observations in Drosophila, the mouse ortholog of Mei-P26, TRIM32, regulates stem cell self-renewal by targeting c-Myc for proteasome-mediated degradation and by binding to Ago-1 [159]. Moreover, TRIM32 is frequently upregulated in human cancers [160] and it has been shown to target tumor suppressor p53 to promote tumorigenesis [161].

9.7.3 p53

p53 is a central tumor suppressor that mediates the response to numerous types of stress by inducing cell cycle arrest, cellular senescence, and apoptosis. Besides, p53 can also control other biological processes involved in disease progression such as metabolism, stem cell maintenance, invasion and metastasis [162]. Therefore, scrutinizing the mechanisms involved in p53 regulation is crucial towards our understanding of cancer. MiRNAs are central players suppressing tumor formation downstream of p53 [163], and downstream targets of p53 are modulated by miRNAs [163, 164]. Importantly, studies in flies showed that p53 is sensitive to miRNA levels [165]. Depletion of Dcr-1 in Drosophila leads to an increase in the expression of a transgene consisting of the dp53-3’UTR fused to GFP (dp53-sensor). The analysis of the dp53 3’UTR led to identify miR-305 as a direct regulator of dp53. [165]. dp53 is upregulated under starvation, which mediates a metabolic adjustment that increases survival in nutrient deprivation conditions. Importantly, miR-305 contributes to this adaptive response. Upon nutrient deprivation, Drosha, Dcr-1, and Ago-1 are downregulated, which leads to a reduction in miR-305 levels. This, consequently, alleviates miR-305-mediated repression of dp53 and facilitates metabolic adaptation [165]. Metabolic reprogramming is central in cancer [166]. Thus, analyzing whether miR-305 regulates dp53 and the potential implications of this axis in tumorigenesis remain to be determined.

9.8 Conclusions and Perspectives

Since the discovery of miRNAs, these regulatory molecules have been associated with virtually every cellular process. As a consequence of this, changes in miRNA expression can contribute to the initiation and development of human diseases including cancer. The main challenge in the field is to identify the relevant miRNA targets in normal development and different pathological contexts. For this, the use of animal models is key.

Studies in Drosophila tumor models establish direct implications of miRNAs as regulators of different hallmarks of cancer such as cell proliferation, apoptosis, differentiation and metabolism. However, we are likely still in the first stages towards understanding the roles that miRNAs play in disease initiation and progression. Thus, insights from Drosophila models will continue to unravel molecular mechanisms underlying miRNA-mediated tumorigenesis. Ultimately, these advances will help understanding the implications of miRNAs in human cancer.