The long noncoding RNA lnc-EGFR stimulates T-regulatory cells differentiation thus promoting hepatocellular carcinoma immune evasion

Abstract

Long noncoding RNAs play a pivotal role in T-helper cell development but little is known about their roles in Treg differentiation and functions during the progression of hepatocellular carcinoma (HCC). Here, we show that lnc-epidermal growth factor receptor (EGFR) upregulation in Tregs correlates positively with the tumour size and expression of EGFR/Foxp3, but negatively with IFN-γ expression in patients and xenografted mouse models. Lnc-EGFR stimulates Treg differentiation, suppresses CTL activity and promotes HCC growth in an EGFR-dependent manner. Mechanistically, lnc-EGFR specifically binds to EGFR and blocks its interaction with and ubiquitination by c-CBL, stabilizing it and augmenting activation of itself and its downstream AP-1/NF-AT1 axis, which in turn elicits EGFR expression. Lnc-EGFR links an immunosuppressive state to cancer by promoting Treg cell differentiation, thus offering a potential therapeutic target for HCC.

Figure 1. Landscape of mRNA and lncRNA expression in CD4⁺ T cells from TBLs and PBLs of human HCC patients.

(a) Hierarchical clustering analysis of 1,251 lncRNAs (left panel) and 2,012 mRNAs (right) that are differentially expressed (the threshold of significance was defined by P<0.05 (with Student t-test) and the false discovery rate) in tumour-infiltrating CD4⁺ T cells (T), paired peripheral blood CD4⁺ T cells (P), and peripheral blood CD4⁺ T cells from healthy controls (N). The clustering tree for lncRNAs and mRNAs is shown at the top. The expression values are shown in shades of red and green, indicating expression above and below the median expression value across all the samples (log scale 10, from –3 to +3), respectively. (b) Pathway analysis showing the significant pathways of the differentially expressed protein-coding genes (P<0.05 with Student t-test). (c) A portion of the co-expression network of the candidate lncRNAs and mRNAs representing the significant pathways. A green node represents one mRNA and a red node represents a lncRNA. (d) Predicted NF-AT1–lnc-EGFR–EGFR feed-forward loop.

Figure 2. Upregulated lnc-EGFR correlates with the distributions of regulatory T cells and cytotoxic lymphocytes.

(a) Relatively increased levels of lnc-EGFR and EGFR were confirmed in tumour-infiltrating T cells by comparing them with those in the peripheral blood T cells of HCC patients and healthy controls (Student’s t-test). (b) Pearson’s correlation analysis was performed to assess the correlation between lnc-EGFR and EGFR (n=67). (c) The Pearson’s correlation of lnc-EGFR expression and Treg percentage was analysed (n=67). (d) The percentage of Treg cells in the infiltrated CD4⁺ T cells was determined with flow cytometry in 67 clinical samples, which were further grouped according to the expression of lnc-EGFR. PBLs from the HCC patients and PBLs from healthy subjects were used as the controls, and Foxp3 and CD25 were used for gating CD4. (e) The subcellular location and intensity of lnc-EGFR were examined with in situ hybridization in sections from HCC patients. The expression of EGFR and Foxp3 was detected with immunohistochemistry in continuous sections from both the lnc-EGFR^high and lnc-EGFR^low groups (× 100)(n=67). (f) The expression of EGFR and marker proteins of Tregs (Foxp3) or CTLs (IFN-γ) were detected in human clinical samples. EGFR stained green, Foxp3 or IFN-γ red. Blue 4′,6-diamidino-2-phenylindole (DAPI) staining indicates nuclei (× 100). White arrow indicates positive staining. The quantitative analysis was performed in the left panel (Student’s t-test). Each experiment was performed triplicated. Data are presented as means±s.e.m. and analysed with Student t-test (*P<0.05, **P<0.01).

Figure 3. Cytoplasmic lnc-EGFR binds specifically to EGFR.

(a) Lnc-EGFR RNA pull-down assay was performed. The associated proteins were processed and subjected to Mass Spec. followed by analysis via the Proteome Discoverer program (a1) and the NCBI protein database with the Mascot search engine (a2). (b) RNA pull-down assay was performed (b1) and the associated proteins were detected with anti-EGFR antibody (b2). A schematic map of potential EGFR binding regions (R1 to 3) in lnc-EGFR (b3). Triangles indicate deletion mutations. (c) Lnc-EGFR specifically interacts with EGFR. c1: RIP assays were performed using CD4⁺ T cells transduced with either lnc-EGFR or lnc-EGFRΔR1 lentiviral particles, and anti-EGFR or anti-PDGFR antibodies. The precipitated RNAs were determined by qPCR for lnc-EGFR, lnc-EGFRΔR1, GAPDH or U6. c2: CD4⁺ T cells were transduced with different doses of lnc-EGFR lentiviral particles and the association of lnc-EGFR with EGFR was determined by RIP assay using anti-EGFR antibodies and qPCR for lnc-EGFR. c3: The amplified sequence (Lnc-EGFR range from 337 to 379 bp) was validated by Sanger sequencing. Each experiment was performed triplicated. Cytoplasmic lnc-EGFR bind specifically to EGFR. (a) Lnc-EGFR RNA pull-down assay was performed. The associated proteins were processed and subjected to Mass Spec. followed by analysis via the Proteome Discoverer program (upper) and the NCBI protein database with the Mascot search engine (lower). (b) RNA pull-down assay was performed (upper) and the associated proteins were detected with anti-EGFR antibody (lower panel). (c) A schematic map of potential EGFR binding regions (R1 to 3) in lnc-EGFR. Triangles indicate deletion mutations. (d) Lnc-EGFR specifically interacts with EGFR. RIP assays were performed using CD4⁺ T cells transduced with either lnc-EGFR or lnc-EGFRΔR1 lentiviral particles, and anti-EGFR or anti-PDGFR antibodies. The precipitated RNAs were determined by qPCR for lnc-EGFR, lnc-EGFRΔR1, GAPDH or U6. (e) CD4⁺ T cells were transduced with different doses of lnc-EGFR lentiviral particles and the association of lnc-EGFR with EGFR was determined by RIP assay using anti-EGFR antibodies and qPCR for lnc-EGFR. (f) The amplified sequence (Lnc-EGFR range from 337 to 379 bp) was validated by Sanger sequencing. Each experiment was performed triplicated.

Figure 4. Lnc-EGFR prevents the ubiquitination of EGFR by binding to Tyr1045.

(a) T cells isolated from peripheral blood of HCC patients were transduced with indicated lentiviral particles and then treated with with EGF (20 ng ml⁻¹) for indicated timepoints followed by western blotting for p-EGFR(Y1045), p-EGFR(Y1068), p-EGFR(Y1073), p-ERK1/2(T202/Y204), EGFR, ERK1/2 and β-actin. (b) Normal, healthy human T cells transduced with mock or indicated lentiviral particles were determined with real-time PCR (upper) and were further treated with EGF (100 ng ml⁻¹) for 90 min or left untreated. Whole-cell lysates were prepared and EGFR was immunoprecipitated followed by western blotting for ubiquitin. Equal loading of EGFR was determined by western blotting via anti-EGFR antibodies (lower). (c) Whole-cells lysates were prepared and c-CBL was immunoprecipitated via anti-c-CBL antibody. The presence of EGFR in the immunecomplex was determined by western blotting via anti-EGFR antibody. (d) Transduced T cells were treated with anti-CD3/anti-CD28 beads (bead-to-cell ratio of 1:1), EGF (20 ng ml⁻¹) or left untreated in the presence or absence of PD98059 (40 μM) and/or CsA (1 μM). Whole-cell lysates were prepared and subjected to Western blotting for p-ERK1/2(T202/Y204), p-MEK1/2(S217/221), p-NF-AT1(S54), ERK1/2, IL-2, MEK1/2 and β-actin. Each experiment was performed triplicated.

Figure 6. Lnc-EGFR promotes iTreg differentiation and inhibits CTL activity in vitro.

(a) Different percentages of Tregs in T cells transduced with lnc-EGFR alone or combined with EGFR shRNA lentiviral particles in the presence or absence of PD98059 (40 μM), or CsA(1 μM), or transduced with lnc-EGFRΔR1 lentivira particles in a Treg cell polarization stimulation assay (with 2 ng ml⁻¹ TGF-β). Quantitative results are shown on the right (n=6 for each group, Student’s t-test). (b) CTL suppression assay of lnc-EGFR was performed in a mixed culture of various types of CD4⁺ cells, CD8⁺ cells and OVA-induced DCs. The initial ratio of CD4⁺ cells to CD8⁺ cells was 1:1. The CD4:CD8 ratios are presented in the right panel(n=6 for each group, Student’s t-test). (c) A 3D culture system was used to simulate the tumour microenvironment, composed of a human HCC cell line (97H cells), DCs vaccinated with 97H cell lysate, various CD4⁺ cells, and CD8⁺ cells, evaluated with immunofluorescent staining for C.C3. The green stain in the right panel indicates cleaved caspase 3, whereas the green stain in the left panel indicates CD8⁺ cells. DAPI was used to stain the nuclei. The calculated data are given in the right panel(n=6 for each group, Student’s t-test). (d) The cytolytic activity of the generated CTLs based on different ratios of CD4⁺ cells was determined in a standard ⁵¹Cr-release assay with E: T ratios of 5:1, 10:1, 20:1 and 40:1. The cytolytic activity of the generated CTLs co-cultured with CD4⁺ cell treated with different groups was detected with a ⁵¹Cr-release assay and is presented in the right panel are means±s.e.m. (n=5 each group) (*P<0.05, **P<0.01). Each experiment was performed triplicated.

Figure 5. NF-AT1/AP1 complex enhances the transcription of Foxp3 lnc-EGFR and EGFR in conventional CD4⁺ T cells.

(a–d) The binding sites for AP1 and/or NF-AT1 in the Foxp3 (a), EGFR (b), and lnc-EGFR (c) promoter regions were mutated and then cloned into the pGL4 vector; the empty vector was used as the control. CD4⁺ T cells were treated with TGF-β (2 ng ml⁻¹), anti-CD3/anti-CD28 beads (bead-to-cell ratio of 1:1), EGF (20 ng ml⁻¹), PD98059 (40 μM), and /or CsA(1 μM), and their fluorescence intensity (FL) measured by comparison with the intensity of Renilla fluorescence(n=6 for each group, Student’s t-test). Expression of Foxp3 protein was confirmed with western blotting in the lower panel and its mRNA expression was confirmed with RT–PCR (d) (n=6 for each group, Student’s t-test). (e,f) Binding of specific nuclear factors was demonstrated with EMSA analysis of the promoter region of EGFR; f panel indicates lnc-EGFR. Competition experiments were performed by preincubating nuclear extracts with AP1/NF-AT1 oligonucleotides by treating with anti-CD3/anti-CD28 beads (bead-to-cell ratio of 1:1) and EGF (2 ng ml⁻¹) in the presence or absence of PD98059 (40 μM) and/or CsA (1 μM). Each experiment was performed triplicated. Data are presented as means±s.e.m. (*P<0.05, **P<0.01).

Figure 7. Lnc-EGFR enhances tumour growth in vivo by promoting Treg differentiation.

(a) Tumour growth in vivo by orthotopic transplantation with adoptive cell transfer using CD4⁺ T cells transduced with lentiviral particles overexpressing indicated vectors. Shown are representative images of tumours in the liver (left panel). Tumour volumes within different groups of the experiment (right panel). (b) Immunohistochemistry was performed in paraffin sections of the tumours for Foxp3, EGFR and IFN-γ, and the integrated optical density was calculated as mean value of five random views (× 100). (c) Foxp3 and IFN-γ were detected in frozen sections of the tumour tissues; the red staining indicates Foxp3 and the green staining indicates IFN-γ. DAPI was used to stain the nuclei (× 200). Quantitative results are listed on the right, as means±s.e.m. (n=6 for each group, Student’s t-test) (d) The distributions of Tregs and CTLs in T cells isolated from the tumours were determined by flow cytometry (n=6 for each group, Student’s t-test). Each experiment was performed triplicated. Data are presented as means±s.e.m. (*P<0.05, **P<0.01).

Figure 8. Schematic model of the forward-feedback loop.

Lnc-EGFR specifically binds to EGFR, stabilizes it through blocking its interaction with c-CBL and the subsequent ubiquitination, and sustains its activity, leading to subsequent downstream cascade activation, Treg differentiation, CTL inhibition and HCC progression.