Small nucleolar RNAs: the hidden precursors of cancer ribosomes

Abstract

Ribosomes are heterogeneous in terms of their constituent proteins, structural RNAs and ribosomal RNA (rRNA) modifications, resulting in diverse potential translatomes. rRNA modifications, guided by small nucleolar RNAs (snoRNAs), enable fine-tuning of ribosome function and translation profiles. Recent studies have begun linking dysregulation of snoRNAs, via rRNA modifications, to tumourigenesis. Deciphering the specific contributions of individual rRNA modifications to cancer hallmarks and identifying snoRNAs with oncogenic potential could lead to novel therapeutic strategies. These strategies might target snoRNAs or exploit the dependence of cancer cells on specific rRNA modification sites, potentially disrupting aberrant ribosomal translation programs and hindering tumour growth. This review discusses current evidence and challenges in linking changes in snoRNA expression to rRNA modification and cancer biology.This article is part of the discussion meeting issue 'Ribosome diversity and its impact on protein synthesis, development and disease'.

Keywords: cancer; rRNA modifications; ribosome heterogeneity; snoRNA.

Figure 1.

Ribosomal RNA modifications and ribosomal functional centres (A) Large and small subunit regions and key centres are depicted in the left panel (CP as central protuberance). They assemble into the 80S ribosome on the right where pseudouridylation (blue), 2’-O-methylation (red) and other modifications (yellow) are depicted around the key centres. mRNA tunnel and peptide exit tunnel have been added for descriptive purposes. (B) Landscape of post-transcriptional modifications and timing of protein contacts along the 28S (top), 18S (middle) and 5.8S (bottom) human rRNA. Black rectangles represent protein occupancy in the nucleolar and cytoplasmic stages of ribosome biogenesis. Modified nucleotides are binned in windows of 10, 5 and 1 respectively for the 28S, 18S and 5.8S rRNAs. Coverage by ribosomal proteins of densely modified regions is delayed in the early stages of ribosome processing, suggesting that modifications may be required for protein fixation. The statistical significance of the positive association between variable modification sites and ribosomal protein binding sites was evaluated using Fisher’s exact test (p‐value = 0.061). The table at the bottom right recapitulates the modifications (Psi as Pseudouridylation and Nm as 2’-O-methylation) and contacts in each rRNA subpart.

Figure 2.

Variability of rRNA modifications in different cellular states and tumourigenesis coefficient of variations (s.d./mean) was used to assess rRNA modification variability in high-throughput rRNA data curated from the literature. Pseudouridylation data were analysed from HydraPsiSeq datasets obtained from bone marrow cell differentiation and cell lines [78]. 2'-O-methylation data were analysed from RiboMethSeq datasets obtained from proliferating, quiescent and senescent human dermal fibroblasts [79], different grades of adult diffuse gliomas (IDH wild-type and mutant) and non-neoplastic samples [80], human cell lines [80] and primary breast tumours [81]. Pseudouridylation is set in blue, while 2’-O-methylation is in red. Opaque spheres indicate a variable site with a coefficient of variation over 0.08 in at least two studies, while translucent spheres represent stable modification sites.

Figure 3.

Diverse levels of snoRNA regulation and impact on rRNA modification patterns (A) Structures of the two classes of snoRNAs: H/ACA box snoRNPs (left, in blue) and C/D box snoRNPs (right, in red). Scaffolding proteins are shown in paler colours, while the catalytic proteins Dyskerin (DKC1) and Fibrillarin (FBL) are in darker blue and darker red, respectively. Ribosomal RNA (rRNA) is depicted in grey. (B) Pie charts showing the relationship between snoRNAs and their target in rRNA. Predictions of snoRNA–rRNA interactions were extracted from the snoDB database [93], which compiles snoRNA Atlas [94] and snoRNAbase [95] prediction data. Copies of snoRNA were considered distinct if they had an entry in snoDB. The two left pie charts display the number of snoRNAs (H/ACA on the left and C/D on the right) guiding more than one site (purple) compared with those guiding only one site (grey). The two right pie charts show the rRNA modification sites (pseudouridylation on the left and 2'-O-methylation on the right) guided by a single snoRNA (orange) or by multiple snoRNAs (green). (C) Sources of rRNA modification heterogeneity can be grouped into three main factors: (I) the snoRNA transcript abundance variability, which is affected by the expression level of the snoRNA, the availability of its core proteins and the intronic regulation of the host gene (for intronic snoRNAs); (II) the accessibility of the targeted site, which is influenced by ribosomal protein fixation, rRNA folding and proximity of rRNA sites guided by different snoRNAs; (III) heterogeneity in rRNA copies with sequence variants and heterogeneity in snoRNA copies.

Figure 4.

Dysregulated snoRNA expression impacts translation. SnoRNAs clearly associated with rRNA site modifications and dysregulated expression in cancer literature are paired with their rRNA targets in a 3D model representation of the ribosome (18S in light grey and 28S in dark grey, PDB id: 6QZP). In accordance with the C/D and H/ACA families, the nature and the direction of each modification are indicated by the colour (red for 2’-O-methylation, blue for pseudouridylation) and the shade (darker for increase, lighter for decrease in cancer), respectively. A bold border means that this rRNA modification site has been shown to impact translation when dysregulated.