Our results indicate that the lack of dyskerin can activate the translation of selected IRES-containing mRNAs. In particular, we demonstrated that dyskerin depletion stimulates the IRES-mediated translation of the mRNA encoding for the VEGF. In addition, dyskerin depletion results in the differential regulation of the translation mediated by IRES sequences present in different viral mRNA types as well as in specific cellular mRNAs.
It has been previously demonstrated that dyskerin clearly has a role in controlling the IRES-mediated translation initiation (8
). In particular, all the data published support the idea that dyskerin depletion downregulates the IRES-mediated translation (9–11
). This happens to the mRNAs encoding the anti-apoptotic factors Bcl-xL and XIAP (9
) and the tumor suppressors p27 (8
) and p53 (10
). Here we demonstrated for the first time that on different mRNAs types, dyskerin KD can upregulate the translation of selected mRNAs containing IRES elements. DKC1 KD significantly increased the ratio between the activities of two reporters of a bicistronic construct containing the VEGF-IRES in both MCF7 and MDA-MB231 breast cancer cells. The observed increase in VEGF-IRES translation is different in the two cell lines being much more evident in MDA-MB231 cells. This difference may reflect the differential capability of the two cell lines to grow in low-oxygen conditions and/or to synthesize VEGF (33
). The observed increase of the reported ratio is due mainly to an increase of VEGF-IRES–dependent reporter activity after DKC1KD; however, in some experiments also a slight downregulation of cap-dependent reporter activity could be detected. For the intrinsic characteristics of the controlled assay also this element affects the obtained ratio. However, the analysis of the individual reporter activities indicates that the ratio increase cannot be considered simply as the consequence of an overall decrease in cap-dependent translation. This is also corroborated by the data obtained from the analysis of polysomal fractions: while DKC1 KD significantly decreased the total VEGF mRNA level in the MCF7, the relative recruitment of VEGF mRNA to polysomal fraction was increased. In the MDA-MB231, the results are even more evident, as there are no significant differences in the rate of the total VEGF mRNA level, whereas an important increase in the amount of VEGF mRNA recruited to polysomes is observed. The increase in VEGF mRNA translation is associated with an increased VEGF protein secretion by both DKC1 KD MCF7 and MDA-MB231 cells. In addition, the increase in VEGF secretion is clearly caused by the VEGF mRNA IRES-mediated translation as it is reverted after treatment with the small molecule PTC299. Indeed, in contrast with the common agents that block the distal portion of the VEGF signaling pathway (17
), PTC299 is a drug that block the production of tumor VEGF at the posttranscriptional level [Dr T. Davis, personal communication and (26
)]. Additionally, by means of a stable shRNA approach, we observed that the reduction of dyskerin levels induces an increase of the clonogenic potential of breast cancer cells. This more aggressive phenotype appears to be, at least to some degree, dependent on the increase of VEGF secretion, as it is reverted on PTC299 treatment.
The analysis of 48S preinitiation complex formation further confirmed the increase in VEGF IRES translation in DKC1-depleted cells with an independent assay and provided functional insights of the mechanism that may give advantage for VEGF-IRES–mediated translation. Indeed, the formation of the complex, a critical step of mRNA translation initiation, is favored on the VEGF IRES element. As for the other assays included in this study, VEGF IRES translation in cells lacking dyskerin is different to what is reported for other IRES elements (e.g. p27 IRES; CrPV IRES) (11
). However, as the formation of the preinitiation complex constitutes only one of the steps of mRNA translation, it cannot be excluded that dyskerin KD could act also at other levels.
Given the known role of dyskerin in rRNA pseudouridylation, the differences shown in the regulation of IRES-mediated translation may be due to the qualitative changes in the ribosomes caused by DKC1 KD. Even if an intrinsic functional ribosomal defect has never been clearly demonstrated, the analysis of the mRNA associated with the polysomal fractions indicates that the changes in protein expression may be ascribed to the translational activity. To explain the observed results, a possible mechanism that may be hypothesized is that the defect in the pseudouridylation may lead to ribosome structural modifications, which in turn affect mRNA translation (27
). On the other hand, it has recently been reported that a cytoplasmic isoform of dyskerin can be expressed in human cells (35
). The function of such cytoplasmic isoform has not been fully elucidated, and its possible involvement in mRNA translation regulation cannot be excluded. On this regard, as the siRNA used in our study can target both the nuclear and the cytoplasmic dyskerin isoforms, it is not possible to exclude that the changes in VEGF mRNA translation are due to the KD of the cytoplasmic isoform.
In addition, the siRNA used in the study could have also induced some of the observed effects through an unspecific targeting of other transcripts. However, an analysis of the effect of DKC1 KD performed with different siRNA oligo separately on the expression of a list of target genes considered in this study excluded this possibility (see Supplementary Figure S10
It is also important to mention that some reports suggest the idea of dyskerin as a tumor suppressor (8–11
). This viewpoint is supported by the data presented here; also, in fact, the defect in dyskerin function can modulate the IRES-mediated translation of different subgroups of cellular mRNAs while simultaneously driving both a reduction in the expression of factors limiting cell proliferation, such as tumor suppressors, and an increase in those promoting cancer cell growth, such as growth factors.
Lastly, our results consolidated the role of dyskerin in the control IRES-mediated translation initiation. The mechanism of IRES-dependent translation activation remains little understood. On the basis of the structures and the involvement of the canonical initiation factors, the best-studied viral IRESs are divided into four types: Type 1 (e.g. poliovirus) and Type 2 (e.g. EMCV) IRESs require eIF4G and eIF4A but do not require eIF4E for their activity; Type 3 (e.g. HCV) IRESs require eIF3 but not eIF4F; Type 4 IRESs (CrPV) can support initiation without any canonical initiation factors at all (36
). As with the viral counterparts Types 1 and 2, cellular IRES elements (37
) have less necessity for initiation factors eIF4E and eIF4G. However, certain cellular IRESs, such as the VEGF, seem to have a strong necessity for eIF4G, while some others are dependent on the initiation factors eIF3 and eIF4A (28
). To date, it remains unclear whether cellular IRES-mediated translation may have distinct features with respect to viral IRES-mediated translation and whether cellular IRESs can be classified in a way similar to viral IRESes (38
). To investigate potential functional similarities between VEGF-IRES and HSP70-IRES, the two IRES elements that are both positively regulated by DKC1 KD, we investigated if HSP70-IRES translation is affected by the treatment PTC299 similarly to what occurs to VEGF-IRES. The treatment with PTC299 had, however, no significant effect on HSP70-IRES–mediated translation (Supplementary Figure S11
), indicating that effect of PTC299 on VEGF-IRES–mediated translation is specific.
Our results show that reduced dyskerin levels induce translational alterations that differentially remodulate IRES-mediated translation; thus, they indicate that the translation initiation mediated by cellular IRESs may involve different mechanisms that can be functionally classified on the basis of the effect of dyskerin downregulation on their activity.