SNORA56 is highly expressed and correlated with poor prognosis in CRC

To determine the differential expression of snoRNA in CRC, we performed snoRNA sequencing on five CRC and its adjacent non-tumor tissues in our previous work, then compared our data with two published snoRNA datasets from CRC tissues [25, 26]. This analysis identified the snoRNAs, SNORA1, SNORA56, SNORA27, and SNORD18B, as being differentially upregulated in all three datasets (Fig. 1A). Next, we used RT-qPCR to validate their expression in 47 paired CRC and corresponding adjacent non-tumor tissues. SNORA56, as the most upregulated, was selected for further investigation (Fig. 1D, S1A–C). SNORA56, which is derived from the tenth intron of the DKC1 gene, is 129 nucleotides long and contains the conserved H/ACA box (Fig. 1B–C). Pan-cancer analysis of The Cancer Genome Atlas (TCGA) dataset revealed that SNORA56 was most significantly enriched in CRC (Figure S2).

Next, we used RT-qPCR and FISH assay to validate SNORA56 upregulation in various CRC cell lines and CRC tissue microarrays, respectively. Our analyses revealed that SNORA56 levels in CRC cell lines (HT29, HCT8, HCT116, SW480 and Caco2, except for LoVo) were markedly higher than in the normal human intestinal epithelial cell line (HIEC) (Fig. 1E). Moreover, FISH revealed significantly higher SNORA56 levels in CRC tissues compared to the adjacent non-tumor tissues (Fig. 1I–K, S1D), suggesting that SNORA56 might influence CRC progression. Furthermore, analysis of the expression of DKC1, from which SNORA56 is derived, revealed that as with SNORA56, DKC1 mRNA levels were frequently upregulated in CRC tissues and cells (Fig. 1F–G). Moreover, in CRC tissues, there was a significant positive correlation between SNORA56 and DKC1 at the transcriptional level using Spearman rank correlation analysis (r = 0.5856, P < 0.001, Fig. 1H), implying that SNORA56 is co-transcribed with DKC1. Because DKC1 is proposed as a CRC prognostic biomarker [27], we evaluated the potential role of SNORA56 in CRC prognosis using TCGA_COAD data from SNORic [28]. This analysis found that CRC patients with higher SNORA56 levels had a poorer 5-year survival rate (Fig. 1L), indicating that SNORA56 is involved in CRC pathogenesis and highlighting SNORA56 as a potential biomarker for CRC prognosis.

Fig. 1

SNORA56 upregulation in CRC correlates with poor prognosis. (A) A Venn diagram of three published cohorts reporting upregulation of snoRNAs in CRC tissues. Cohort 1 was obtained from the published article [25]. Cohort 2 is from the Gene Expression Omnibus dataset GSE2091626. Cohort 3 is from our previously published work [29]. (B) Visualization of the genomic location of SNORA56 in its host gene, DKC1, on the UCSC Genome Browser. (C) A schematic representation of the structure of SNORA56. (D, F) The relative expression of SNORA56 and DKC1 in 47 paired CRC and adjacent non-tumor tissues was revealed using RT-qPCR. (E, G) The relative expression of SNORA56 and DKC1 in HIEC, HT29, HCT8, HCT116, SW480, Caco2, and LoVo cells was determined using RT-qPCR. (H) Analysis of the correlation between the expression of SNORA56 and DKC1. (I) SNORA56 FISH analysis and hematoxylin & eosin (H&E) staining in CRC tissue microarray. The mean density of SNORA56 signals was measured using Image Pro Plus. (J–K) SNORA56 staining intensity and the IHC score of CRC tissue microarray. (L) Kaplan Meier curve of the 5-year survival analysis in SNORA56 high and low groups using TCGA_COAD datasets from the SNORic database

SNORA56 promotes CRC cell proliferation in vitro and in vivo

To assess the biological function of SNORA56 in CRC, we transiently downregulated its expression using two independent antisense oligonucleotides (ASOs) in HCT116 and HCT8 cells. Moreover, we generated HIEC and CRC cell lines (HCT8 and HT29) in which SNORA56 was stably up- or down-regulated using lentiviral transduction of overexpressed plasmids (LV-NC and LV-SNORA56) or CRISPR/Cas9 system (sgNC and sgSNORA56), respectively. Transfection efficiency was assessed using RT-qPCR (Fig. 2A–C, S3A). Notably, stable SNORA56 overexpression or silencing did not affect DKC1 levels, suggesting that SNORA56 functions independently of DKC1 (Fig. 2D, S3B). Next, CCK-8 and colony formation analyses of the proliferative potential of CRC cells in vitro revealed that SNORA56 depletion disrupts CRC cell viability (Fig. 2E–J), whereas its upregulation markedly promotes proliferation (Figure S3C–D). Moreover, a xenograft nude mouse model of CRC using HT29 sgNC and sgSNORA56 cells was constructed through subcutaneous injection (Fig. 2K). Results revealed that the weights and volumes of tumors from the SNORA56-deficient cells (sgSNORA56) were markedly lower than in those from the control (sgNC) (Fig. 2L–M). Consistent with this observation, Ki67 levels, a canonical proliferative marker, in SNORA56 depleted xenografts were apparently decreased with the analysis of IHC, which confirmed that SNORA56 significantly promoted CRC cell proliferation in vivo (Fig. 2N).

Fig. 2

SNORA56 promotes CRC proliferation in vitro and in vivo. (A–C) The efficiency of SNORA56 silencing in CRC cell lines upon the indicated transfections was determined using RT-qPCR. (D) RT-qPCR analysis of DKC1 mRNA levels in HT29 cells stably transfected with sgNC and sgSNORA56. (E–F, H–I) CCK8 analysis of the proliferation of CRC cells upon the indicated transfections. (G, J) Colony formation analysis in CRC cells upon transient or stable SNORA56 knockdown vs. the negative control. (K) Tumors from nude mouse subcutaneous xenografts bearing HT29 sgNC cells (right) and sgSNORA56 cells (left). (L–M) Tumor weights and growth curves based on tumor volume measurements every two days. (N) IHC analyses of H&E and Ki67 in the xenograft tumors. Ki67 scores were calculated in HT29 sgNC and sgSNORA56 xenografts. Scale bar: 100 μm

SNORA56 drives tumorigenesis by promoting 28 S rRNA maturation and translation

Because snoRNAs canonically function in ribosome modification, we first examined the effect of SNORA56 on rRNA maturation. According to the snoRNA Orthological Gene Database (snOPY), SNORA56 guides 28 S rRNA U1664 pseudouridylation via a base-pairing interaction between its flanking regions and rRNA (Fig. 3A). Structure analysis revealed that although the U1664 site is located in the large ribosomal subunit, it is far from the ribosome’s functional domain (Fig. 3B), implying that SNORA56 might alter ribosomal conformation. To investigate SNORA56-mediated pseudouridylation, we mutated two interacted flanking regions of SNORA56 to their complementary matched sequences respectively to disrupt the association between SNORA56 and 28 S rRNA as previously described (Fig. 3A, MUT1 and MUT2) [6, 30]. We then examined the expression levels of SNORA56 after transfection with its mutants or non-mutated control (WT) (Fig. 3E). Also, qPCR analysis using specific primers against unprocessed or total rRNAs revealed that SNORA56 silencing markedly increased the levels of immature 28 S rRNA (Fig. 3C–D, S4B), whereas SNORA56 overexpression reduced their level (Figure S4A). However, SNORA56 levels did not have observable effects on 18 S rRNA maturation (Figure S4C–E), which is consistent with the interaction between SNORA56 and 28 S rRNA not 18 S rRNA. Furthermore, polysome profiling and OP-Puro analysis of ribosomal activity showed that SNORA56 markedly enhances the translation capacity of the ribosome in CRC cells (Fig. 3F–G, S4F–G).

Next, to determine whether the tumorigenic role of SNORA56 is mediated by its interaction with 28 S rRNA, we overexpressed SNORA56-WT, MUT1 or MUT2, in SNORA56-depleted HCT8 and HT29 cells respectively (Fig. 3E, S4H–J). This analysis revealed that overexpressing SNORA56-WT, and not the mutants, restored the levels of 28 S rRNA (Fig. 3D, S4B), implying that interaction between SNORA56 and 28 S rRNA is essential for 28 S rRNA maturation. Notably, SNORA56 mutants failed to rescue SNORA56 silencing-induced proliferation both in vitro and in vivo (Fig. 3H–O), suggesting that SNORA56 plays a tumorigenesis role relying on the pseudouridylation of 28 S rRNA at U1664 site.

Fig. 3

SNORA56 promotes 28 S rRNA maturation and translation in CRC cells. (A) Schematic representation of the interaction between SNORA56 and 28 S rRNA, and the mutated regions. (B) PyMOL analysis of the ribosomal structure and U1664, the 28 S rRNA site that is pseudouridylated by SNORA56. (C) A schematic diagram of the primers used to detect total and precursors of 18 and 28 S rRNA. Paired primers were displayed in the same color. (D-E) RT-qPCR analysis was used to reveal the relative levels of unprocessed 28 S rRNA and SNORA56 in HT29 sgNC and sgSNORA56 cells transfected with SNORA56-WT, MUT1, or MUT2. (F) Polysome profiling of HT29 cells with stable SNORA56 knockdown vs. the control. (G) OP-puro analysis of HCT29 cells with stable SNORA56 knockdown vs. the control. (H–K) CCK8 and colony formation assays of the treated HT29 and HCT8 cells. (L) Tumors from nude mice subcutaneous xenografts bearing HT29 sgNC, sgSNORA56, together with empty vector (EV), SNORA56-WT, MUT1 and MUT2, respectively. (M–N) Tumor weights and growth curves. (O) IHC analyses of Ki67 in the xenograft tumors. Scare bar: 100 μm

SNORA56 upregulates GCLC protein expression by activating its translation

To identify potential downstream targets of SNORA56 in CRC, we carried out a proteomic analysis on sgNC and sgSNORA56 cells (Fig. 2C–D). This analysis uncovered 258 differentially expressed proteins, of which 120 were upregulated and 138 were downregulated (Fig. 4A–B). Clusters of Orthologous Groups/Eukaryotic Orthologous Groups (COG/KOG) functional classification showed that 10 of the downregulated proteins are involved in translation, ribosomal structure, and biogenesis (Figure S5A). Moreover, KEGG functional enrichment analysis revealed that the downregulated proteins were enriched for the ribosome (Fig. 4C), further indicating that SNORA56 is involved in ribosome activation. Notably, the downregulated proteins were enriched for metabolic processes associated with ferroptosis, such as glutathione and cysteine and methionine metabolism (Fig. 4C), suggesting that SNORA56 might influence ferroptosis by regulating metabolism.

To test this possibility, we focused on the protein GCLC, an indispensable rate-limiting enzyme in the biosynthesis of glutathione. GCLC is downregulated upon SNORA56 silencing and it may inhibit ferroptosis by decreasing cellular peroxide levels. Consistently, transient or stable SNORA56 silencing in CRC cells significantly reduced GCLC protein levels (Fig. 4D, S5B). Furthermore, we found that SNORA56 deficiency significantly decreased the activity of glutamate cysteine ligase (Fig. 4E–F, S5C), whereas SNORA56 overexpression markedly upregulated GCLC protein and enzymatic levels (Fig. 4D, F). Moreover, IHC analysis of CRC tissue microarray revealed GCLC upregulation in paired CRC and adjacent non-tumor tissues (Fig. 4G). Further analysis indicated a significant positive correlation between GCLC protein level and SNORA56 staining intensity (Fig. 4H). Next, validation analysis of the relationship using other paired CRC tissues revealed that as with SNORA56, GCLC protein levels were markedly upregulated in tumors (Fig. 4K). However, GCLC mRNA levels were unaffected by SNORA56 (Fig. 4I–J, S5D–E), implying that SNORA56 regulates GCLC posttranscriptionally or translationally. To test this, GCLC protein stability was assessed upon treatment with cycloheximide (CHX) in HT29 and HCT8 cells to block protein synthesis and evaluate protein degradation. This analysis revealed no clear differences in GCLC protein half-life in conditions of SNORA56 deficiency or overexpression (Fig. 4L–M, S5F), suggesting that SNORA56 regulates GCLC translation and not degradation. To test this possibility, we transfected stable SNORA56-silenced HCT8 and HT29 cells with SNORA56-WT or with 28 S rRNA binding-impaired SNORA56 mutants. Importantly, the SNORA56 mutants failed to rescue GCLC protein expression (Fig. 4N), indicating that the SNORA56-induced maturation of 28 S rRNA was required for the high GCLC protein levels in CRC.

Fig. 4

SNORA56 upregulates GCLC protein by activating its translation. (A) Differential protein expression in the sgSNORA56 vs. the sgNC group. (B) Volcano plot of differentially expressed proteins. (C) KEGG pathway enrichment analysis of significantly downregulated proteins. (D) GCLC protein levels in HCT116, HCT8, HT29 and HIEC cells were determined using western blotting after indicated transfections. (E–F) GCL enzyme activity in HT29 and HIEC cells with indicated transfection. (G) IHC analysis of GCLC levels in CRC tissue microarray. Mean GCLC density was determined using Image Pro Plus. Scale bar: 100 μm. (H) Analysis of the correlation between the SNORA56 and GCLC protein levels using CRC tissue microarray. (I–J) RT-qPCR analysis of relative GCLC mRNA levels in HT29 and HIEC cells upon indicated transfections. (K) Western blot analysis of GCLC protein levels in paired CRC tissues. (L–M) Western blot analysis of GCLC protein stability in SNORA56-silenced HCT8 and HT29 cells vs. control cells after CHX treatment for indicated durations. Relative band intensities were measured on ImageJ. (N) Western blot analysis of GCLC protein levels in HT29 and HCT8 cells after indicated transfections. GCLC protein levels were normalized to GAPDH.

SNORA56 inhibits ferroptosis and promotes proliferation by upregulating GCLC protein in CRC

Because GCLC is thought to function in peroxide clearance, we investigated if SNORA56 regulates ferroptosis in CRC. Lipid reactive oxygen species (ROS) and cell death were assessed using flow cytometric analysis of HCT116, HCT8 and HT29 cells after transient or stable SNORA56 silencing. Notably, our data showed that SNORA56 deficiency markedly caused lipid ROS accumulation accompanied by increased cell death (Fig. 5A–D, S6A–D), suggesting that SNORA56 is involved in ferroptosis inhibition in CRC cells. Because labile iron contributes to ferroptosis by directly oxidizing lipids via Fenton reaction and acting as a cofactor for lipid oxidizing enzymes [31], we assessed the Fe2+/Fe3+ ratio to validate the role of SNORA56 in ferroptosis regulation. This analysis revealed a marked decrease in the Fe2+/Fe3+ ratio in the SNORA56 knockdown group (Fig. 5E, S6E), which is consistent with peroxidation in the cellular microenvironment. Similarly, the GSH/GSSG ratio was reduced by SNORA56 silencing (Fig. 5F, S6F), indicating that SNORA56 depletion triggers ferroptosis in CRC cells. However, SNORA56 overexpression in HIEC cells increased both ratios and resistance to erastin, a canonical ferroptosis inducer that blocks the system Xc- antiporter (Fig. 5E-F, L, S6G–H). Notably, only the ferroptosis inhibitor, ferrostatin-1 (Fer-1) reversed lipid ROS elevation and cell death following SNORA56 deficiency, and both the necroptosis inhibitor, necrosulfonamide, and the apoptosis inhibitor, Z-VAD-FMK, could not prevent CRC cell death (Fig. 5G–J). Moreover, analysis of the impact of SNORA56 on ferroptosis sensitivity in CRC revealed that SNORA56 downregulation markedly sensitized CRC cells to ferroptosis in a dose-dependent manner, whereas SNORA56 overexpression caused ferroptosis resistance (Fig. 5K-L, S6L). These observations highlight targeting both SNROA56 and ferroptosis inducers as a promising strategy for CRC treatment. Additionally, cell viability assays revealed that SNORA56 knockdown suppressed cell growth, which could be almost entirely restrained after treatment with Fer-1 (Figure S6I–K), implying that SNORA56 facilitates cell proliferation, at least partly, through ferroptosis resistance. These data illustrate that SNORA56 mediates ferroptosis resistance and highlight SNORA56 as a critical target for regulating ferroptosis sensitivity in CRC cells.

Since GCLC is a key rate-limiting enzyme in GSH synthesis (Fig. 5M), we next evaluated whether SNORA56-mediated ferroptosis inhibition depends on GCLC protein expression. To determine the hypothesis, we generated a GCLC overexpression system after silencing SNORA56 and verified transfection efficiency using RT-qPCR and western blotting (Figure S6M–P). As expected, GCLC overexpression rescued GCL enzyme activity in HT29 cells (Fig. 5N, S6Q), and remarkably, CRC cell proliferation was reversed upon GCLC overexpression (Fig. 5O–P, S6R–S). Moreover, lipid ROS, cell death, and ferroptosis sensitivity were partially reversed under GCLC retrieved (Fig. 5Q-R, S6T–U, X-Y), indicating that SNORA56 contributes to ferroptosis resistance in CRC mediated by GCLC. Considering SNORA56 facilitates GCLC translation via the 28 S rRNA pseudouridylation, we transfected with SNORA56-WT, MUT1 or MUT2 respectively in HCT8 and HT29 sgSNORA56 cells and examined the ferroptosis. As expected, SNORA56 mutants with depleted 28 S rRNA binding capability failed to recover the ferroptosis resistance (Fig. 5S-T, S6V-W). Next, we examined the pharmacological effect of DL-Buthionine-Sulfoximine (BSO), which inhibits glutathione synthesis and induces ferroptosis, in SNORA56-overexpressing HIEC cells. This analysis revealed that BSO markedly suppressed SNORA56-induced elevation of GCL enzyme activity (Fig. 5U), as well as SNORA56-driven proliferation and ferroptosis inhibition in HIEC cells (Fig. 5V–Y), indicating that SNORA56 promotes CRC tumorigenesis by enhancing GSH protein synthesis.

Fig. 5

SNORA56 causes GCLC-mediated CRC ferroptosis resistance and proliferation. (A, C, S, X, Q) HCT116, HCT8, HT29, and HIEC cells were transfected as indicated, stained with the BODIPY C11 probe, and subjected to flow cytometry for lipid ROS detection. (B, D, T, Y, R) HCT116, HCT8, HT29, and HIEC cells were subjected to indicated transfections, followed by propidium iodide staining and cell death analysis using flow cytometry. (E) The relative Fe2+/Fe3+ ratios in HT29 and HIEC cells after indicated transfections. (F) The relative GSH/GSSG ratios in HT29 and HIEC cells after indicated transfections. (G-J) Lipid ROS levels and cell death in HCT116 and HCT8 cells transfected with SNORA56 ASOs and control ASO, followed by treatment with DMSO, Fer-1, necrosulfonamide, and Z-VAD-FMK, were measured using flow cytometry. (K-L) The cell viability of HT29 and HIEC cells after indicated transfections and treatment with various erastin concentrations for 24 h, was determined using the CCK8 assay. (M) A schematic representation of GSH synthesis. (N, U) GCL enzyme activity in HT29 and HIEC cells after indicated treatment. (O-P, V-W) CCK8 and colony formation assays were used to assess the proliferation of treated HT29 and HIEC cells

SNORA56 is a promising CRC diagnostic biomarker and therapeutic target

To further investigate the correlation between SNORA56 expression and the clinical characteristics of CRC patients, we obtained SNORA56 expression data from the TCGA_COAD dataset (n = 394) from SNORic, along with clinical data from the UCSC Xena website and analyzed it using chi-square tests or Fisher’s exact tests to illustrate the potential relationship between SNORA56 level and clinical index. These analyses revealed a significant correlation between SNORA56 expression and histological type (P = 0.015), venous invasion (P = 0.040), and distant metastasis (P = 0.002) (Table S1–2). Importantly, Cox proportional hazards regression analysis identified high SNORA56 expression as an independent risk factor for CRC patients’ overall survival (HR = 2.223, P = 0.021, Fig. 6A). Because snoRNAs are stable in the plasma, we assessed SNORA56 levels in the plasma of CRC patients (age: 66.61 ± 11.55, n = 48) vs. healthy subjects without underlying gastrointestinal disease (age: 57.72 ± 8.67, n = 48). This analysis revealed higher SNORA56 levels in CRC plasma (Fig. 6B). We then collected 154 paired CRC tissues to evaluate the diagnosis efficacy of SNORA56. Consistent with this, SNORA56 levels were upregulated in CRC compared to the corresponding adjacent non-tumor tissues (Fig. 6C). To evaluate the diagnostic efficacy of SNORA56, analyses of the ROC curve were performed in CRC tissues and plasma, and the area under the curve (AUC) were 0.6759 and 0.7572, respectively (Fig. 6D–E), indicating that plasma SNORA56 has significant potential for CRC diagnosis. Notably, combining CEA or CA199, the canonical plasma-based CRC biomarkers, with SNORA56 markedly increased their AUC values (Fig. 6E–F), implying improved diagnostic value.

Finally, we examined the therapeutic potential of targeting SNORA56 in CRC. Because SNORA56 is involved in ferroptosis resistance, we first generated a xenograft model of CRC by subcutaneously injecting HT29 sgNC and sgSNORA56 cells into nude mice. Next, we intraperitoneally injected the mice with imidazole ketone erastin (IKE), an erastin substitute that is stable in vivo. Notably, knocking down SNORA56 markedly enhanced the sensitivity of the HT29 cells to IKE, characterized by significant suppression in tumor volumes and weights (Fig. 6H–J), highlighting the anti-CRC therapeutic potential of targeting SNORA56 while inducing ferroptosis. Moreover, GCLC and Ki67 staining was significantly weaker in the HT29 sgSNOR56 xenograft tumors than in the control, and this effect was enhanced by IKE (Fig. 6K–L), implying that SNORA56 exerts its effects on ferroptosis resistance and proliferation inhibition via GCLC. Together, these data highlight the diagnostic and therapeutic potential of SNORA56 in CRC.

Fig. 6

SNORA56 is a promising CRC diagnostic biomarker and therapeutic target. (A) Forest plot of the overall survival hazard ratios using TCGA_COAD data. (B) Relative levels of plasma SNORA56. (C) Relative SNORA56 expression in 154 paired CRC tissues was revealed by RT-qPCR. (D–E) ROC curve analysis of SNORA56 in CRC plasma and tissues. (F–G) ROC curve analysis of the CRC diagnostic potential of CEA or CA199 when combined with SNORA56. (H) Tumors from a nude mouse xenograft model that was subcutaneously injected with HT29 sgNC (right) and sgSNORA56 (left) cells and then treated with IKE or solvent respectively. (I) Plots of tumor growth measurements taken on indicated days. (J) Tumor weights. (K) IHC analysis of GCLC and Ki67 in xenograft tumors after treatment with IKE or the solvent. Scale bar: 100 μm. (L) GCLC and Ki67 IHC scores in indicated xenografts

Fig. 7

The proposed model of how SNORA56 regulates GCLC translation, thereby inhibiting ferroptosis and promoting proliferation. SNORA56, which was significantly upregulated in CRC, pseudouridylates 28 S rRNA at site U1664, thereby promoting ribosome maturation. Consequently, SNORA56 triggers GCLC translation, which drives ferroptosis resistance and proliferation in CRC.