hnRNP-A1 binds to the IRES of MELOE-1 antigen to promote MELOE-1 translation in stressed melanoma cells

Abstract

The major challenge in antigen-specific immunotherapy of cancer is to select the most relevant tumor antigens to target. To this aim, understanding their mode of expression by tumor cells is critical. We previously identified a melanoma-specific antigen, melanoma-overexpressed antigen 1 (MELOE-1)-coded for by a long noncoding RNA-whose internal ribosomal entry sequence (IRES)-dependent translation is restricted to tumor cells. This restricted expression is associated with the presence of a broad-specific T-cell repertoire that is involved in tumor immunosurveillance in melanoma patients. In the present work, we explored the translation control of MELOE-1 and provide evidence that heterogeneous nuclear ribonucleoprotein A1 (hnRNP-A1) binds to the MELOE-1 IRES and acts as an IRES trans-activating factor (ITAF) to promote the translation of MELOE-1 in melanoma cells. In addition, we showed that endoplasmic reticulum (ER) stress induced by thapsigargin, which promotes hnRNP-A1 cytoplasmic translocation, enhances MELOE-1 translation and recognition of melanoma cells by a MELOE-1-specific T-cell clone. These findings suggest that pharmacological stimulation of stress pathways may enhance the efficacy of immunotherapies targeting stress-induced tumor antigens such as MELOE-1.

Keywords: ER stress; IRES; ITAF; long noncoding RNA; melanoma; tumor antigens.

Fig. 1

RNA secondary structures upstream of MELOE‐1 ORF, redrawn from predictions by UNAFold (http://www.unafold.org), revealed typical stem‐loop elements and putative hnRNP‐A1 binding sites. (A) Schematic representation of the predicted RNA secondary structure of the 275 nt sequence upstream of MELOE‐1 ORF (AUG initiation codon in red). (B) Focus on the proximal regions of the wt, variant 1, and variant 2 internal ribosomal entry sequence (IRES), highlighting the putative hnRNP‐A1‐binding sites, circled in orange (5′CAG‐3') and blue (5′UAG‐3'). Nucleotide changes in variants 1 and 2, shown in red, are predicted not to change the rest of the IRES sequence.

Fig. 2

hnRNP‐A1 binds to MELOE‐1 internal ribosomal entry sequence (IRES). (A) Typical surface plasmon resonance (SPR) sensorgram (out of three performed) showing binding of recombinant hnRNP‐A1 (concentrations ranging from 1.95 to 250 nm) to immobilized 275 nt MELOE‐1 wt IRES (red line), variant 1 (green line), and variant 2 (blue line). (B) RT‐PCR detection of meloe RNA after immunoprecipitation with anti‐hnRNP‐A1 mAb of UV cross‐linked lysates from M117, M134, and M170 melanoma cell lines. PCR from mock‐immunoprecipitated cell lysates and from plasmid are shown as negative control and positive control, respectively.

Fig. 3

Silencing of hnRNP‐A1 reduces MELOE‐1 internal ribosomal entry sequence (IRES) activity. (A) Efficacy of siRNA‐mediated depletion of hnRNP‐A1 on M113 assessed by RT‐qPCR 48 h postlipofection. (B) FLuc/RLuc ratio (*100) was measured in M113 melanoma cell lysate 48 h post‐transfection with pRF bicistronic vectors in which Renilla luciferase (RLuc) translation is cap‐dependent and Firefly luciferase (FLuc) translation is controlled either by MELOE‐1 IRES, encephalomyocarditis virus (EMCV) IRES, or nothing (no IRES). Where indicated, cells were cotransfected with hnRNP‐A1 siRNA (siRNA#1, 10 µm, Santa‐Cruz Biotechnologies, siRNA#2, 5 µm, Qiagen) or with a universal control siRNA (5–10 µm). Data are expressed as mean ± SD (n = 7 independent experiments). P‐values were calculated using repeated‐measure one‐way ANOVA followed by Holm–Sidak’s multiple comparison test.

Fig. 4

The endoplasmic reticulum (ER) stressor thapsigargin enhances MELOE‐1 expression and hnRNP‐A1 cytosolic translocation. (A) Confocal microscopy detection of yellow fluorescent protein (YFP) expression in the M117‐YFP cell line after a 24‐h thapsigargin treatment (200 nm) or not. M117 was stably transfected with meloe cDNA in which MELOE‐1 ORF was replaced by YFP. Nuclei, stained with Hoechst 33342, are shown in red (left panel) and YFP+ cells in green (middle panel). The white bar scale represents 100 µm. (B) Percentages of YFP‐positive cells (expressed as mean ± SD) in response to increasing concentrations of thapsigargin in 6 independent experiments (over 3000 nuclei are counted in each condition). (C) Cellular sublocalization of hnRNP‐A1. M117‐YFP cells were treated with 0.5 µm thapsigargin for 24 h. Cells were harvested, and the nuclear and cytosolic fractions were analyzed by western blot for hnRNP‐A1 and tubulin levels. (D) Quantification of hnRNP‐A1 expression in cytoplasmic and nuclear fractions (n = 3 independent experiments, expressed as mean ± SEM). The level of expression of hnRNP‐A1 and tubulin in the cytosolic fraction of untreated cells was used as reference on each blot. Levels of hnRNP‐A1 are expressed relative to the levels of tubulin in each sample (internal control of even total protein loading between samples). Quantification of expression of hnRNP‐A1 was performed with Image Lab software using cytoplasmic expression in untreated cells as a reference and adjusted with tubulin expression in each sample. Considering these adjustments to show relative expression, no statistical test could be performed.

Fig. 5

Thapsigargin treatment of the M113 melanoma cell line enhances MELOE‐1 presentation to a specific CD8 T‐cell clone. (A) Interferon gamma (IFN‐γ) production by a Melan‐A/MART1_{26‐35 A27L} (top panel) and MELOE‐1_36‐44 CD8 T‐cell clone (bottom panel) was measured by intracellular staining after a 5‐h exposure to M113 (HLA‐A*0201 positive) or M6 (HLA‐A*0201 negative) melanoma cell line pretreated or not with thapsigargin (100 nm for 24 h followed by a 24‐h wash‐out period). (B) Thapsigargin treatment of M113 melanoma cell line (as in A) does not affect HLA‐A*0201 expression at the cell surface. Cells were analyzed by flow cytometry after staining with a PE coupled‐anti‐HLA‐A*0201‐specific Ab (gray: isotype control on untreated M113; blue: HLA‐A*0201 on untreated M113; green: HLA‐A*0201 on thapsigargin‐treated M113). Relative fluorescence intensity (RFI) is indicated above each histogram. (C) Compilation of five experiments performed as in (A) showing consistent increase of IFNγ production by MELOE‐1_36‐44 CD8 T‐cell clone in response to thapsigargin‐treated M113 melanoma cell line (100 nm), while the Melan‐A/MART1_{26‐35 A27L} response is not affected. Data were tested for statistical significance using a paired t‐test.