Targeting cholesterol biosynthesis promotes anti-tumor immunity by inhibiting long noncoding RNA SNHG29-mediated YAP activation

Abstract

Anti-tumor immunity through checkpoint inhibitors, specifically anti-programmed death 1 (PD-1)/programmed death ligand 1 (PD-L1) interaction, is a promising approach for cancer therapy. However, as early clinical trials indicate that colorectal cancers (CRCs) do not respond well to immune-checkpoint therapies, new effective immunotherapy approaches to CRC warrant further study. Simvastatin is an inhibitor of 3-hydroxy-3-methylglutaryl-coenzyme A (CoA) reductase (HMGCR), the rate-limiting enzyme of the mevalonate (MVA) pathway for the cholesterol biosynthesis. However, little is known about the functions of simvastatin in the regulation of immune checkpoints or long noncoding RNA (lncRNA)-mediated immunoregulation in cancer. Here, we found that simvastatin inhibited PD-L1 expression and promoted anti-tumor immunity via suppressing the expression of lncRNA SNHG29. Interestingly, SNHG29 interacted with YAP and inhibited phosphorylation and ubiquitination-mediated protein degradation of YAP, thereby facilitating downregulation of PD-L1 transcriptionally. Patient-derived tumor xenograft (PDX) models and the clinicopathological analysis in samples from CRC patients further supported the role of the lncRNA SNHG29-mediated PD-L1 signaling axis in tumor microenvironment reprogramming. Collectively, our study uncovers simvastatin as a potential therapeutic drug for immunotherapy in CRC, which suppresses lncRNA SNHG29-mediated YAP activation and promotes anti-tumor immunity by inhibiting PD-L1 expression.

Keywords: YAP activation; colorectal cancer; immune-checkpoint PD-L1; immunotherapy; lncRNA SNHG29; simvastatin.

Graphical abstract

Figure 1

Simvastatin targets HMGCR and reduces colorectal cancer (CRC)-specific mortality (A and B) Exploring HMGCR mRNA and protein expression in most cancer categories by assembling The Human Protein Atlas (http://www.proteinatlas.org). (C) The expression of HMGCR by immunohistochemistry (IHC) staining on paraffin-embedded CRC specimens. (D) The expression levels of HMGCR in formalin-fixed paraffin-embedded (FFPE) colon cancers and normal tissues were showed in the indicated scattergram using ImageJ (n = 142). ∗∗∗p < 0.001. (E) Kaplan-Meier analysis of overall survival (OS) of CRC patients with HMGCR. (F) The mutually adjusted hazard ratios for the associations of characteristics with the risks of recurrence, CRC-specific mortality, and all-cause mortality from the DCCG database (2001–2011). (G) Kaplan-Meier analysis of OS in 415 CRC patients with and without statins therapy. (H) Cytotoxicity activity of simvastatin against CRC cell line. (I) The Gene Ontology analysis shows the enriched biological process of differentially expressed genes in the simvastatin-treated group (/fold change/ > 2; p value < 0.05). (J) Heatmap showing the difference transcripts between negative control and simvastatin-treated CRC cells (/fold change/ > 2; p value < 0.05).

Figure 2

Simvastatin represses tumor expression of PD-L1 and enhances cytotoxic T lymphocyte (CTL) infiltration in CRC patients (A) PD-L1 mRNA expression in HCT116 and RKO cells treated with simvastatin or negative control. (B) PD-L1 protein expression in HCT116 and RKO cells treated with simvastatin or negative control. (C) PD-L1 protein expression in xenograft tumors from HCT116 and RKO cells. (D) PD-L1 protein expression in xenograft tumors from RKO cells. (E) Graphic illustration of a human PDX model-based therapeutic study. PDX tumors were generated from CRC patients and treated with simvastatin or negative control. (F) In vivo analyses of tumor (left panel) and growth (right panel) in mice that were subcutaneously implanted with tumor tissues from CRC patients and treated with simvastatin or negative control (50 mg/kg) five times weekly for 3 weeks (n = 6). Results are presented as mean ± SD. ∗∗∗p < 0.001. (G) PD-L1 protein expression in the PDX model treated with simvastatin or negative control. (H) Representative immunohistochemical images were shown in randomly selected tumor. (I) Graphic illustration of C57BL/6 mice injected with 2 × 10⁶ MC38 tumor cells. At day 7 after tumor inoculation, simvastatin or negative control was intragastric administrated into tumor-bearing mice (50 mg/kg, five times weekly for 3 weeks). Anti-PD-L1 antibody was intraperitoneally injected combined with simvastatin in addition (10 μg/g, two times weekly for 3 weeks). (J) In vivo analyses of tumors (upper panel) and their growth (bottom panel) in C57BL/6 mice that were subcutaneously injected with 2 × 10⁶ MC38 tumor cells and treated with simvastatin or combined with anti-PD-L1 antibody (n = 6). Results are presented as mean ± SD. ∗∗∗p < 0.001. (K) PD-L1 protein expression in the C57BL/6 mice model treated with simvastatin or negative control. (L) CD8⁺ T cells in PBMC and TILs were analyzed by flow cytometry. Experiments were performed with at least three biological replicates and are representative of at least two independent experiments. The results are presented as mean ± SD. ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 3

Simvastatin inhibits tumor expression of lncRNA SNHG29, which promotes PD-L1 expression (A) lncRNA-sequencing analysis of simvastatin-treated and negative control cells is presented in a heatmap analysis. (B) Flow chart of screening-altered lncRNAs in indicated samples. (C) Schematic annotation of lncRNA SNHG29 genomic locus on chromosome 17 (p11.2) in humans. Green rectangles represent exons. (D) qRT-PCR analysis of differentially expressed lncRNA SNHG29 and PD-L1. (E) Western blotting tested PD-L1 level in lncRNA SNHG29-overexpressed or -knockdown CRC cells. (F) Representative ISH staining for SNHG29 and IHC staining for PD-L1, CD8, and Ki67 expression in the xenograft tumor tissues from CRC patients, treated with simvastatin or negative control (50 mg/kg) five times weekly for 3 weeks. The relative intensities of ISH and IHC staining were quantified by ImageJ software (n = 6). The density of immune cell infiltrates in the tumor was calculated as the number of positive cells per field of tissue. The results are presented as mean ± SD. ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 4

lncRNA-SNHG29 inhibition manifest enhances CTL killing targets in vitro and in vivo (A) Graphic illustration of isolation, activation, and expansion of human cytotoxic T cells and Mo-DCs to perform cytotoxicity LDH assay. (B and C) Flow cytometry analysis of cell surface markers showed in vitro-generated and -activated CD8⁺ T cells (B) and mature Mo-DCs (C). (D and E) Cytotoxic activity of CD8⁺ T cells against CRC cells treated with simvastatin or negative control (D) and mock or lncRNA SNHG29-specific siRNAs (E). The graph shows percentage of specific killing of CRC cells with various effector-to-target (E:T) ratios measured by LDH assays. (F−H) Graphic illustration of a human PDX model-based therapeutic study. PDX tumors with negative control or SNHG29-specific shRNAs were generated from CRC patients (F). At day 7 after tumor inoculation, SNHG29-specific shRNAs or negative control were intratumor injected into tumor-bearing mice (10⁹ TU, twice). Tumor volume was monitored every 7 days (G and H). (I) Representative ISH staining for SNHG29 and IHC staining for PD-L1, CD8, granzyme B, and Ki67 expression in the xenograft tumor tissues from CRC patients treated with SNHG29-specific shRNAs or negative control after 3 weeks. (J) The relative intensities of ISH and IHC staining were quantified by ImageJ software. The density of immune cell infiltrates in the tumor was calculated as the number of positive cells per field of tissue. All experiments were performed in triplicate, and results are presented as mean ± SD. ∗∗∗p < 0.001. (K) Tumor volume of PDX tumor tissues was monitored every 7 days with anti-PD-L1 antibody or immunoglobulin G (IgG) isotype intraperitoneal injection (10 μg/g, two times weekly for 3 weeks), and lncRNA SNHG29-specific shRNAs or negative control were intratumor injected (10⁹ TU, twice) (n = 5). Results are presented as mean ± SD. ∗∗p < 0.05, ∗∗∗p < 0.001.

Figure 5

lncRNA SNHG29 interacts with YAP to inhibit its phosphorylation and ubiquitination-mediated protein degradation (A) lncRNA SNHG29-associated proteins identified by RNA pulldown and mass spectrometric analyses in HCT116 cells (upper panel). Western blot analysis of pulldown product (bottom panel). (B) RIP assays in HCT116 cells. qRT-PCR analysis of RIP (upper panel). Agarose electrophoresis of PCR products (bottom panel). Experiments were performed in triplicate, and data are presented as mean ± SD. ∗∗∗p < 0.001. (C and D) lncRNA SNHG29 facilities YAP nuclear accumulation as demonstrated by immunofluorescence staining (C) and western blot (D). (E) Silencing of SNHG29 caused downregulation of CTGF but increased phosphorylation of YAP in HCT116 cells. Overexpression of SNHG29 caused upregulation of CTGF but reduced YAP phosphorylation in DLD1 cells. (F) RKO cells were transfected with SNHG29-specific siRNAs or negative control and subjected to a cycloheximide (CHX) chase assay. Immunoblot detection of YAP (left panel); IB data were quantified using ImageJ software (right panel). After 24 h, CHX (10 μg/mL) was added to the cell culture medium, and incubation was continued for 0, 4, 8, or 12 h. Error bars indicate the mean ± SD. ∗∗∗p < 0.001. (G) Ubiquitination assays of CRC cells co-transfected YAP with SNHG29-specific siRNA (left) or SNHG29 plasmid (right). The bottom panels depict the input of the cell lysates. 24 h after transfection, 10 nM MG132 was added to the 1640 culture medium, and incubation was continued for 8 h. (H) KEGG enrichment analysis of pathways, which significantly enriched in simvastatin-treated CRC cells (adjusted p ≤ 0.05). (I) Western blot showed total and phosphorylated protein of YAP, LATS1, CTGF, and PD-L1 in RKO cells treated with simvastatin or negative control. (J) YAP localization as demonstrated by immunofluorescence staining. Cells were treated with simvastatin or negative control alone or co-cultured with SNHG29 plasmid for 24 h before fixation.

Figure 6

lncRNA SNHG29 promotes the expression of PD-L1 via facilitating YAP transcriptional activity (A and B) qRT-PCR (A) and western blots analysis (B) showed a positive correlation between YAP and PD-L1 in various CRC cells. (C) IB detection using the indicated antibodies in CRC cells transfected with the indicated plasmid (left panel) or siRNAs (right panel). (D) Binding of YAP to the PD-L1 promoter was studied by the chromatin immunoprecipitation (ChIP) assay, and the coprecipitated DNA was subjected for analysis of PD-L1 by qRT-PCR (upper panel). The PCR procedures was shown by agarose gel electrophoresis (bottom panel). Experiments were performed in triplicate, and data are presented as mean ± SD. ∗∗∗p < 0.001. (E) Potential YAP binding sites in the human PD-L1 promoter between –2,000 and + 100 bp. YAP plasmid alone or with SNHG29-specific siRNA was transfected into HEK293 cells to detect the transcriptional activity of the PD-L1 promoter by a dual-luciferase reporter system. Experiments were performed in triplicate, and data are presented as mean ± SD. ∗∗p < 0.05, ∗∗∗p < 0.001. (F) IB detection using the indicated antibodies in CRC cells transfected with the indicated plasmid or siRNAs. Data shown represent three independent experiments. (G) The graphic illustration of simvastatin as a new inhibitor of PD-L1 to enhance anti-tumor immunity via inhibiting lncRNA-SNHG29-mediated YAP activation in CRC tumor progression. (H and I) The ISH staining of SNHG29 and IHC staining of YAP and PD-L1 in tumor tissues and adjacent normal tissues of CRC paraffin-embedded samples. (J) The expression levels of SNHG29, YAP, and PD-L1 in FFPE colon cancers and normal tissues were shown in an indicated scattergram using ImageJ (n = 163). ∗∗∗p < 0.001. (K) The correlation of expression of lncRNA SNHG29, YAP and PD-L1 in FFPE CRC tissues (n = 163 biologically independent samples); linear regression analysis.