We have previously proposed that RMS represent an arrested phase of normal development at the transition point between regulative growth and terminal differentiation, a transition regulated in part through the balance of repressive and activating bHLH protein dimers [8
]. We showed that increasing the abundance of MyoD:E-protein heterodimers tipped the balance to differentiation and proposed that this heterodimer might target an unknown central integrating function since multiple myogenic repressors were down-regulated. One prediction of this earlier model was that modulating the abundance of any of the factors that impinge on the integrating function might be sufficient to induce differentiation in RMS. Our present study supports and extends this model by demonstrating that miR-206 integrates the activity of multiple proliferative and myogenic factors and acts as a switch that transitions the RMS from growth to differentiation.
To test whether different myogenic co-factors can induce differentiation in RMS, we chose one transcriptional activator and one repressor, RUNX1 and ZNF238, respectively. RUNX1 enhances MyoD activation at a variety of MyoD targets, including ZNF238
and possibly MYOG
and the MEF
genes. In contrast, ZNF238 down-regulates multiple members of the inhibitory HES and HEY protein family and factors that drive proliferation. Our data on ZNF238
regulation, motif analysis, and gene targets in this and previous work [11
] suggest that the induction of this inhibitory factor serves two purposes: (1) to down-regulate genes that inhibit myogenesis; and (2) to interfere with MyoD binding at genes it might regulate in myoblasts. Despite differences in their direct targets, RUNX1 and ZNF238 both increase miR-206 transcription and lead to a terminally differentiated state. For RUNX1 we demonstrated a direct binding of the miR-206 promoter, whereas the mechanism for ZNF238 remains speculative, possibly through its suppression of ID
gene expression, which would increase the proportion of productive MyoD:E-protein heterodimers. While in vitro
in nature, our data on the relative ability of MyoD:E and MSC:E to shift the regulatory E-box sequences that control miR-206 expression (Figure D and Additional file 12
: Figure S8) suggest that even relatively small changes in the availability of E-protein partners could make a dramatic difference in the expression of miR-206.
, a bHLH that inhibits myogenesis [32
], suppresses the activation of miR-206 by binding an E-box required for induction by MyoD. A requirement for more than one MyoD-bound E-box to drive full target activation has been described before [33
], and it is currently unclear whether MSC is simply preventing MyoD binding or recruiting repressive factors to the locus. The fact that the miR-206 locus has acetylated H4 even when not robustly expressed suggests that MSC may have a simple obstructionist role at this locus. Future work will be necessary to determine the relative roles of MSC and MyoD at miR-206 and other myogenic targets.
Data reported in this manuscript, when combined with previous data from us and others [8
] suggests a specific model for the regulation of miR-206 that involves nested positive and negative feed-forward and feed-back loops to create a molecular switch for regulating the transition from growth to differentiation in myogenic cells (Figure ). In replicating myoblasts, a MyoD:E-protein heterodimer binds an E-box site in the regulatory regions of ID2 and ID3 [11
] creating the potential for an oscillating circuit: any increase in MyoD activity would increase ID expression, which would dampen MyoD activity by decreasing the availability of the E-protein dimer partner. However, if MyoD:E-protein heterodimers pass a threshold of activity sufficient to initiate a feed-forward circuit activating RUNX1 and ZNF238, then ZNF238 shuts off ID production by occluding the MyoD binding sites and thereby relieving the negative-feedback regulation of MyoD. The increased MyoD and RUNX1 activity can then more effectively compete with MSC on the miR-206 regulatory regions and the increased miR-206 levels feed back to inhibit MSC, and likely other growth promoting factors [8
], thereby locking the cell into a committed differentiation program. Therefore, miR-206 integrates the output of oscillating circuits and acts as a genetic switch to transition the cell from a proliferative growth phase to differentiation.
Figure 6 miR-206 integrates the output of oscillating circuits and acts as a genetic switch to transition from growth to differentiation. The experimental data support a network model composed of coupled oscillators with miR-206 functioning as a switch regulating (more ...)
Hematological malignancies are often categorized based on an arrested transition between stages of cellular differentiation. Our work suggests that the same might apply to RMS and possibly other solid tumors. bHLH factors control cell fate and differentiation in multiple cell types and a balance among bHLH dimer partners and other co-factors might establish similar ‘tipping points’ at critical genes that regulate the transition from regulative growth to differentiation. Our emerging model of multiple pathways (some functioning as oscillating circuits) integrated by switch-points for differentiation has significant implications for drug therapies to induce differentiation. Different cell types may not exhibit identical convergence of pathways. Therefore, combining multiple drugs that each has a small effect on different components might induce differentiation in the target cells while exhibiting low toxicity and few off-target effects.