Presentation of Telomerase Reverse Transcriptase, a Self Tumor Antigen, is Down Regulated by HDAC Inhibition

Abstract

Histone deacetylases (HDACs) modify chromatin’s architecture leading to decreased gene expression, an effect that is reversed by HDAC inhibition. The balance between deacetylation and acetylation is central to many biological events including the regulation of cell proliferation and cancer, but also the differentiation of immune T cells. The effects of HDAC inhibition on the interaction between anti-tumor effector T and tumor cells are not known. Here we studied presentation of a universal self tumor antigen, telomerase reverse transcriptase (TERT), in human tumor cells during HDAC inhibition. We found that HDAC inhibition with trichostatin A was associated with a decrease presentation and diminished killing of tumor cells by cytotoxic T lymphocytes. Using gene array analysis we found that HDAC inhibition resulted in a decrease of genes coding for proteasome catalytic proteins and for tapasin, an endoplasmic reticulum resident protein involved in the Major Histocompatibility Complex Class I pathway of endogenous antigen presentation. Our findings indicate that epigenetic changes in tumor cells decrease self tumor antigen presentation and contribute to reduced recognition and killing of tumor cells by cytotoxic T lymphocytes. This mechanism could contribute to tumor escape from immune surveillance.

Introduction

Epigenetic modifications influence gene activity without altering DNA sequences from the transcriptional activity of selected genes to the individual’s susceptibility to disease (1), and include both transient modifications and stable heritable changes (2). Acetylation and deacetylation of histones impart modification on chromatin’s architecture (3) which reflect on increased or decreased gene expression through the opposing activities of histone acetyltransferases (HATs) and histone deacetylases (HDACs). Acetylation plays a critical role in the regulation of cell proliferation (4) and transcriptional silencing associated with a generalized loss of acetylation has been found in cancer (5). Levels of acetylation also play a role in the differentiation of T cells (6, 7) and in the expression of ligands for natural killer cells (8). Although mouse tumor cells treated with high dose HDAC inhibitors are immunogenic and confer partial protection from tumor growth in vivo (9), little is known on whether epigenetic changes of chromatin structure influence antigen presentation in human tumor cells.

Here we studied the effect of HDAC inhibiton on the presentation of telomerase reverse transcriptase (TRT), a bona fide universal antigen in human cancer cells (10, 11) using trichostatin A (TSA), a natural product hydroximate with high affinity for the catalytic site of class I and II histone deacetylases (HDACs) (12). We monitored the surface density of complexes formed between the major histocompatibility/Human Leukocyte Antigen (HLA)-A2 molecule and two high affinity 9mer hTRT peptides: ⁵⁴⁰ILAKFLHWL⁵⁴⁸ (p540) and ⁸⁶⁵RLVDDFLLV⁸⁷³ (p865) (11). Since telomerase is expressed in >85% of human cancers (13) and CD8 T cells specific for p540 are found in the blood of cancer patients at high frequency (14), the model system studied herein is relevant to better understanding the relation between cancer cells and the anti-tumor immune response in humans. Our findings show a reduction of HLA-A2/hTRT peptide complexes at the cell surface of TSA treated human tumor cells. More importantly, using a highly specific human CTL clone, we found that TSA treated cancer cells killing was reduced compared to untreated. This finding may be relevant to understand how immunosurvelliance of cancer cells can be affected in vivo.

Material and Methods

Cell lines and Inhibitors

The human B lymphoblastoid JY cell line and prostate cancer cells LnCap (the kind gift of Dr. A. Vitiello), melanoma cells 629.38 (the kind gift of Dr. S. Rosenberg), and TAP-deficient T2 cells (the kind gift of Dr. P. Creswell) were cultured in complete RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS). TSA, Lactacystin, Sodium butyrate, and Brefeldin A were purchased from Sigma (St Louis, MO).

Immunochemical detection of hTRT, HLA-A2/hTRT complexes and HLA-A2

Detection of HLA-A2/hTRT complexes at the cell surface.

HLA-A2/p540 and HLA-A2/p865 complexes were detected with the human Fab 4A9 and 3H2 (15), respectively. Briefly, cells were treated with TSA (1μg/ml), Brefeldin A (10 μg/ml), or lactacystin (10 μM/ml). After washing, cells were incubated with human Fab 4A9 or 3H2 (2 μg/reaction) for 1.5 hrs at +4°C. After washing cells were counterstained with FITC-conjugated goat antibody to human Fab (Jackson laboratories) for 45 min. Surface staining for HLA-A2 was performed using monoclonal antibody BB7.2 (ATTC) followed by a FITC-conjugated goat antibody to mouse IgG (eBioscience). Cells were analyzed in a FACSCalibur^™ (Becton Dickinson) using the BD CellQuest^™ data acquisition and analysis software.

Cytotoxicity assays

Buffy coats from HLA-A2.1 healthy donors were purchased from the Blood bank of Hopital Henry Mondor (Paris) in accordance with an approved IRB protocol. hTERT specific human CTL lines were generated according to published procedure (11). Briefly, peripheral blood mononuclear cells were stimulated with hTERT peptides in vitro in 24-well plates using autologous irradiated adherents cells as antigen presenting cells and IL-2 and IL-7. After three to four rounds of culture CTLs were screened in a standard ⁵¹Cr release assay. The cold target inhibition assay (11) was performed to ensure the specificity of the CTL clone. T2 cells were pulsed with either p540 or the HIV p17 gag 77–85 peptide as a control. An adaptation of this assay was used to investigate killing inhibition where ⁵¹Cr-labeled targets treated with TSA were used (Figure 6C and D). In these experiments the E:T ratio was kept constant at 20:1 whereas the cold:hot inhibitor ratio varied as indicated in the legend to the figure.

Figure 6. Reduced killing of human tumor cells treated by TSA by CTL specific for the HLA A2/p540 complex.

CTL from an individual donor immunized in vitro against p540 (11) were used as effectors (E) against ⁵¹Cr-labeled tumor targets (T) treated (18 hours) with TSA (1 μg/ml). (A) Direct lysis of TSA-treated or untreated human tumor cells (JY, LnCap and Mel 629). T2 cells pulsed with p540 served as a control for the ability of the CTL clone to lyse p540 pulsed targets. Results refer to percent (%) lysis at the E:T ratios indicated. (B) Specificity of CTL by cold target inhibition assay. 51Cr-labeled, TSA-treated JY cells (5 × 104 cells/ml) were mixed with T2 cells pulsed with p540 (10 μg/ml) (closed triangle) or T2 cells pulsed with control HIV-1 gag peptide at the indicated cold:hot target cell ratios. The E:T ratio was kept constant at 20:1. Results refer to percent lysis according to (11). Tests were run in triplicate. (C-D) TSA treatment of tumor cells decreases presentation of p540 and as a result enhances cold target inhibition. The canonical CTL reaction was set at a fixed E:T ratio of 20:1. Hot target cells were mixed at cold;hot ratio of 50:1 (C) or 250:1 (D) where black columns indicate cold T2 cells pulsed with p540 and open columns cold T2 cells pulsed with control HIV-1 gag peptide. Results are expressed as killing of tumor cells, with or without TSA treatment, in the presence of T2 cells pulsed with p540 as a proportion of killing in presence of T2 cells pulsed with control gag peptide that was used as 100% killing.

Proteasome Activity Assay

JY cells were treated with TSA (1 μg/ml), Sodium Butyrate (20 μg/ml), Lactacystin (3.8 μg/ml) or 5 μl of DMSO (untreated) in RPMI supplemented with 5% FBS for 18 hrs. Sodium Butyrate was used as a control based on the fact that to the best of our knowledge it is the only HDAC inhibitor reported to reduce proteasome activity beside TSA (16). Lactacystin served as a generic inhibitor of proteasome function. Cells were harvested, washed and resuspended in complete medium. Proteasome activity assay was performed using the cell-based luminescent assay Proteasome-Glo^™ (Promega, WI, US) in white flat-bottom 96 well-plates (Costar) with 20,000 cells/well. Luminescence was measured in a plate reader Infinite M200 (TECAN, Switzerland). Assays were done in triplicate, and repeated twice with similar results.

Deconvolution and fluorescence microscopy

Cells treated with TSA and immunostained as described above were also used for deconvolution microscopy studies. Briefly, after staining with antibodies, cells were incubated with DAPI-HCl for 15 min, and then cytospun onto glass slides. Slides were air dried and mounted in glycerol using cover slips (Fisher Scientific). The slides were examined by epifluorescence microscopy (Nikon T200 microscope) mounted on Delta Vision deconvolution microscope (Applied Precision, LLC). In some cases, where indicated, deconvolution was performed on the resultant image stacks using SoftWorx as supplied by the manufacturer (Applied Precision, LLC). Images shown in Figure 1 and 2 were volume projections of several consecutive optical sections taken from the center of the image stacks after deconvolution. In experiments where comparative quantitation of fluorescence was performed, samples to be compared were stained under identical conditions with the same reagents at the same time. Samples were imaged under identical conditions. Camera exposure time was identical and pixel intensities were kept within the linear range of the camera (CoolSnap HQ, Princeton Instruments). Quantitation was performed on volume projections composed of several consecutive optical sections. The resultant volume projections were scaled identically. Quantitation was performed using SoftWorx at Data Inspector.

Figure 1. Flow cytometry analysis of decreased presentation of HLA-A2/hTRT peptide complexes by TSA treatment.

(A) Flow cytometry analysis of the specificity of 4A9 and 3H2 antibodies for HLA-A2/hTRT complexes. (Left panel) 4A9 antibody (2 μg) in 100 μl was used to stain unpulsed TAP-deficient T2 cells or cells pulsed with 10 μg/ml of p540 or the hybrid peptides p540/p865 (ILAKFFLLV) and p540/p49 (ILAKFRRRV). (Right panel) 3H2 antibody was similarly used to stain unpulsed TAP-deficient T2 cells or cells pulsed with p865 or the hybrid peptides p865/p540 (RLLVDDLHWL) and p865/p49 (RLVDDRRRV). (B) Flow cytometry analysis of surface staining for the HLA-A2/p540 and HLA-A2/p865 complex in TSA-treated JY cells and stained by either antibody 4A9 or 3H2 (2 μg/reaction). JY cells were treated for 8 or 18 hours as indicated. Analyses were performed as detailed in Material and Methods. (C) Flow cytometry analysis of surface staining for the HLA-A2 in TSA-treated JY cells. Analyses were performed as detailed in Material and Methods. Results are representative of multiple independent experiments.

Figure 2. Deconvolution microscopy analysis of decreased presentation of HLA-A2/hTRT peptide complexes by TSA treatment.

(A) Deconvolution microscopy analysis of surface (Sfc) staining for the HLA-A2/p540 complex in untreated and TSA-treated (18 hours) JY cells. HLA-A2/p540 complexes are in green and nuclei in blue (magnification 60x). Analyses were performed as detailed in Material and Methods. Data are representative of four independent experiments. (B) Quantitation of fluorescence intensity of the optical images shown in panel A. The scale represents a gradient of pixel intensities. Analyses were performed as detailed in Material and Methods. (C) Deconvolution microscopy visualization of surface HLA-A2/p540 complexes at higher magnification (100x) comparing surface contour of untreated and TSA-treated (18 hours) JY cells. Complexes are in green and nuclei in blue.

RT-PCR

hTRT mRNA was detected by reverse transcriptase (RT)-PCR analysis. Briefly, 10⁶ cells were treated with TSA (1μg/ml) for the indicated time, and total RNA was isolated using the RNAeasy Mini Kit (Qiagen). The RNA was reverse transcribed into cDNA by RT-Omniscript (Qiagen) for 1 hour at 37°C. cDNA amplification was induced using specific primers (forward 5’-ACCAAGAAGTTCATCTCCCTGG-3’ and reverse 5’-GCTCATCTTCCACGTCAGCTC-3’). Amplification of the glyceraldehydes-3-phosphate dehydrogenase (GAPDH) mRNA served as a control for RNA extraction.

Quantitative RT-PCR (qRT-PCR)

cDNA was generated from 200ng of total RNA using Omniscript Reverse Transcriptase (Qiagen). qRT-PCR was carried out using the Taqman method on a LightCycler © 480 real-time-PCR system (Roche) and analyzed by efficiency-corrected relative quantification method. Sequences for probes and primers were as follows: tapasin (NM_003190.3)- Forward CCAGAGCCTCAGCAGGAG; Reverse GGGTGAGGACAGTCAGTACCA; Universal ProbeLibrary probe #9 (Roche Applied Science); GADPH-Forward AGCCACATCGCTGAGACA; Reverse GCCCAATACGACCAAATCC; Universal ProbeLibrary probe #60 (Roche Applied Science). Results were expressed as fold change compared to untreated cells.

DNA Microarray

Biotinylated cRNA was prepared using the Illumina RNA Amplification Kit, Catalog #1L1791 (Ambion, Inc., Austin, TX) according to the manufacturer’s directions starting with 250 ng total RNA. The labeling approach utilizes a modified Eberwine protocol (17) by which messenger RNA is converted to cDNA, followed by an amplification/labeling step mediated by T7 DNA polymerase. The cDNA and cRNA filter cartridges (Ambion) were used according to the manufacturer’s instructions for RT and IVT cleanup, respectively. For microarray analysis, the Illumina Human Expression BeadChip was used (Illumina, San Diego). Hybridization of labeled cRNA to the BeadChip, and washing and scanning were performed according to the Illumina BeadStation 500x manual. Essentially the amplified, biotin-labeled human cRNA samples were resuspended in a solution of Hyb E1 buffer (Illumina) and 25% (v/v) formamide at a final concentration of 25 ng/μL. 1.5 μg of each cRNA were hybridized. Hybridization was allowed to proceed at 55 C, for 18 hours after which, the bead array matrix was washed for 10 minutes with 1X High temperature buffer (Illumina), followed by a subsequent 10 minute wash in Wash E1BC buffer. The arrays were then washed with 100% ethanol for 10 min to strip off any remaining adhesive on the chip. A 2 min E1BC wash was performed to remove residual ethanol. The arrays were blocked for 5 minutes with 1% (w/v) casein-PBS, (Pierce). The array signal was developed via 10 min incubation with Streptavidin-Cy3 at a final concentration of 1μg/mL solution of (GE Healthcare) in 1% casein-PBS blocking solution. The Human Expression BeadChip was washed a final time in Wash E1BC buffer for five minutes and subsequently dried via centrifugation for 4 minutes at a setting of 275 rcf. The arrays were scanned on the Illumina BeadArray Reader, a confocal-type imaging system with 532 (cye3) nm laser illumination. The bead signals were computed with weighted averages of pixel intensities, and local background is subtracted. Sequence-type signal was calculated by averaging corresponding bead signals with outliers removed (using median absolute deviation). Preliminary data analysis and QC was carried out using the BeadStudio software (Illumina). Simultaneous normalization of multiple microarrays was done using the mloess method (18).

Results

HDAC inhibition downregulates hTRT presentation

Previously, it was shown that TSA treatment increases the transcription and expression of hTRT in normal (19, 20) and cancer (21) cells. Consistent with this report we found that treatment with TSA induced an increase in the transcriptional activity of the hTRT gene in the HLA-A2⁺/hTRT⁺ lymphoblastoid JY cell line. In repeated independent experiments specific mRNA amplification was consistently visible after 30 minutes, peaked at 18 hours and persisted for an additional 24 hours after TSA removal (data not shown). We reasoned that the increase in hTRT transcription could yield an increase in the number of HLA-A2/hTRT peptide complexes displayed at the cell surface (antigen presentation). Therefore, we investigated complex formation between the HLA-A2 molecule and two high affinity (~4 nM) hTRT peptides, p540 and p865. Detection of the two complexes was performed using two human recombinant Fab antibodies, 4A9 which is specific for the HLA-A2/p540 complex and 3H2 which is specific for the HLA-A2/p865 complex (15). Notably, each antibody only recognized the complex formed with the homologous hTRT peptide. Complexes with hybrid peptides comprising 5 amino acids residues of the homologous peptide and 4 amino acid residues from a different hTRT peptide did not mediate binding (Figure 1A). Surprisingly, in several independent flow cytometry analyses surface staining by either 4A9 or 3H2 was substantially reduced (Figure 1B) in TSA-treated JY cells compared with untreated cells. Diminished surface expression of the HLA-A2/hTRT peptide complex persisted for 24 hours after TSA removal (data not shown). Surface expression of HLA-A2 molecules was similar in treated and untreated cells (Figure 1C) even though in 3 out of 10 experiments a decrease (10–20%) in mean fluorescence intensity was observed after TSA treatment.

To better visualize the reduction in antigen presentation we used deconvolution microscopy focusing on the HLA-A2/p540 complex since this has a longer half life (~ 6 hours) than the HLA-A2/p865 complex (11). In untreated JY cells, HLA-A2/p540 complexes were visible with a bright surface granular staining (Figure 2A, left panel). However, fewer complexes were seen in TSA-treated vs. untreated cells (Figure 2A, right panel). A comparative, quantitative analysis of the corresponding images showed a three-fold decrease in the intensity of complex staining in TSA-treated cells (Figure 2B). The same quantitative difference was documented in three independent experiments. A higher magnification of single cells further demonstrates this phenomenon (Figure 2C).

Decreased antigen presentation was not unique to JY cells since TSA treatment (18 hrs) of the HLA-A2⁺/hTRT⁺ melanoma cell line 629.38 resulted again in transcriptional activation of hTRT Figure 3A) and, importantly, in diminished surface expression of the complex (Figure 3B). TSA treatment increased somewhat the detection of HLA-A2 molecules at the cell surface (Figure 3C). Collectively, whereas a decrease in presentation is not supported by a decrease in surface expression of HLA-A2 molecules, the data show that chromatin changes following HDAC inhibition by TSA promote hTRT transcription but decrease presentation of MHC complexes formed with high affinity hTRT peptides.

Figure 3. TSA treatment down-regulates presentation in 629.38 melanoma cells.

629.38 melanoma cells were treated with TSA (1 μg/ml) for 18 hrs. (A) RNA was extracted and tested by RT-PCR for hTRT transcription as indicated in Material and Methods. (B) Cells were cell surface stained with antibody 4A9 to visualize the HLA-A2/p540 complexes as indicated in Material and Methods. HLA-A2/p540 complexes are in green and nuclei in blue (magnification 60x). (C) HLA-A2 expression was assessed by flow cytometry.

Gene profiling analysis

MHC I/peptide complexes originate in the ER where nascent MHC I molecules are loaded with peptides that are generated in the proteasome and are translocated into the endoplasmic reticulum (ER) by the transporter associated with antigen processing (TAP). Since the MHC I/peptide cargo in the ER is a function of peptide input from the cytosol and the local synthesis of MHC I molecules, we reasoned that the three-fold lower antigen presentation concomitant with an increase in antigen expression in the cytosol could have several explanations. For instance, TSA could inhibit the proteasome and/or TAP, hence diminishing assembly of MHC I/peptide complexes in the ER. To this end we tested TAP expression levels by flow cytometry in TSA-treated cells and found no change. This lead to the further possibility that proteasome, or events in the ER-Golgi, could impair peptide generation, MHC I/peptide complex maturation and/or egress from the ER. To this end, the effect of TSA on surface expression of the HLA-A2/p540 complex was compared to that of lactacystin, a selective inhibitor of the proteasome, and Brefeldin A (BFA), a ER→Golgi transport inhibitor that also inhibits MHC I/peptide complex presentation. Lactacystin, markedly reduced presentation compared to untreated as well as TSA-treated JY cells (Figure 4A). BFA (18 hour-treatment) blocked completely surface expression of the complex (Figure 4A). Interestingly, BFA treatment for 4 hours did not affect the cell surface expression of the MHC/peptide complex (data not shown), in agreement with the half-life (~6 hours) of the complex. This suggests that the diminished presentation associated with HDAC inhibition by TSA could be due to either a partial defect of the proteasome, a defect in ER → Golgi transport, or both.

Figure 4. HDAC inhibition by TSA modulates the processing and presentation pathway in JY cells.

(A) Deconvolution microscopy visualization of surface (Sfc) expression of the HLA-A2/p540 complex by antibody 4A9 (2 μg/reaction) on JY cells treated (18 hours) with with either lactacystin (LC) (3.8 μg/ml), TSA (1 μg/ml), or Brefeldin A (BFA) (10 μg/ml). (B) Gene array profiling of JY cells treated with TSA (18 hours). Total RNA was extracted and analyzed using Illumina bead arrays. Levels of gene expression were compared to untreated JY cells. In the clustergrams, genes are grouped into proteasome genes; ER stress responsive genes; ER→Golgi resident and trafficking genes; and MHC I related genes. The fold changes are noted on the respective color scales. Two independent experiments were performed and shown.

To distinguish between these possibilities, a genome-wide expression profiling of TSA-treated (18 hours) JY cells was performed to see which gene pathway(s) might be involved in reducing antigen presentation by HDAC inhibition. Proteasome genes. Peptides for the MHC I are generated mainly by the proteasome and other cytoplasmic proteases. The 26S proteasome which is responsible for the degradation of poly-ubiquinated proteins is composed of outer structural α−subunits and inner proteolytic β-subunits which are encoded by different genes. The microarray analysis of the proteasome in TSA-treated JY cells showed an overall down-regulation of β-subunit genes (Figure 4B), with β-subunits 8 (PSMB8) and 10 (PSMB10) being down-regulated the most. Of note gene expression of the cytoplasmic peptidase TPPII was unchanged (not shown) suggesting that the down-modulation of the β-subunits of the proteasome could be responsible, in part, for a downstream diminished loading of peptides (i.e., p540 and p865) onto the MHC I molecules in the ER. ER stress genes. ER stress responsive genes, such as DNA-damage-inducible transcript 3 (Ddit3), PP1RS15A (Gadd34), Heat shock protein 5 (Grp78, Bip), Wolfram syndrome 1 (Wfs), activating transcription factor 4 (ATF4), and calreticulin, were all upregulated (Figure 4B). This indicates that a downstream effect of HDAC inhibition is the induction of ER stress. This is not surprising since TSA is known to transcriptionally modulate many (2%) genes in malignant human cells (22), and supports the notion that TSA-mediated apoptosis of tumor cells may be mediated by ER stress. Consistent with the induction of ER stress is the transcriptional activation of pro-inflammatory genes such as TNF-α (x7) and IL-23 p19 (x4) (data not shown) which are upregulated during ER stress (our unpublished data). ER and Golgi resident and trafficking related genes. Analysis of these genes showed an overall minor increase in ER→Golgi trafficking genes indicating that the diminished presentation of HLA-A2/p540 complexes by TSA treatment is unlikely to be due to a decrease in ER→Golgi trafficking. MHC I antigen presentation related genes. Transcriptional analysis of the MHC I pathway was performed to identify changes that could explain the diminished surface expression of the HLA-A2/p540 complex. Overall MHC I genes were unchanged, although, a slight negative trend was observed (Figure 4B). The increase in Bap31 (an ER resident type II transmembrane protein exerting quality control on MHC I/peptide complexes) and the decrease in Der1 (an ER retrotranslocation transmembrane protein) suggest a program to conserve MHC I molecules. In contrast, tapasin, a key regulator of TAP and MHC I peptide binding, and a gate keeper of MHC I/peptide complex stability (23), was markedly diminished. Consistently, transcription of the tapasin-like gene which shares similarity at the tri-dimensional level with tapasin was also markedly down-regulated. Finally, additional members of the peptide loading complex and MHC I maturation (TAP1, TAP2, calnexin and ERp57) showed little change, if any, with the exception of calreticulin (x2 increase).

HDAC inhibition affects the proteasome and tapasin

Gene profiling analysis proved of key value in assessing the possible downstream targets of TSA with respect self tumor antigen presentation. First, we noted that wide spectrum HDAC inhibition down-regulated transcription of most genes of the proteasome complex corresponding to a global decrease in proteasome activity of ~30% (Figure 5A). Since hTRT proteolysis requires ubiquination and is mediated by the proteasome (24) the decrease in proteasome activity by HDAC inhibition is consistent with the decrease in presentation of hTRT. Since peptide generation in the proteasome is the first step in antigen processing, the observed diminished presentation of hTRT may reflect in part an effect on the proteasome. Second, HDAC inhibition promoted ER stress, something not observed previously after TSA treatment. Several considerations suggest that ER stress may not play a crucial and direct role in diminishing presentation of hTRT since (a) hTRT transcription increases after TSA treatment whereas the unfolded protein response (ER stress) in general diminish transcription (25), and (b) treatment with thapsigargin (an inhibitor of the Ca⁺⁺ pump and inducer of ER stress) did not diminish the surface expression of the MHC I/peptide complex (data not shown). Third, a defect in ER→Golgi transport of MHC I/peptide complexes is also unlike since most genes involved in transport and trafficking underwent discrete upregulation upon TSA treatment, ruling out defects in ER→Golgi transport. Finally, selected changes in the MHC I-related genes were observed with tapasin being markedly down-regulated. Tapasin down-regulation observed in the array analysis (Figure 4B) was confirmed by quantitative RT-PCR (Figure 5B). Tapasin is involved in optimal loading of MHC I molecules with peptide and in retaining unstable MHC I/peptide complexes in the ER while favoring egress of stable, high affinity complexes (26), and in the retrieval of COPI vesicles containing MHC I molecules loaded with unstable peptides (27). Thus, its down-regulation can explain the diminished antigen presentation. Since other genes related to MHC I/peptide complex biogenesis such as calreticulin, Grp78/Bip, and COPI, were all moderately upregulated, it appears as if the down regulation of tapasin affected the quality control and sorting of MHC I/peptide complexes in and from the ER, hence playing a central role in the diminished presentation of high affinity the HLA-A2/hTRT complexes.

Figure 5. Analysis of proteasome function and tapasin’s expression after TSA treatment.

(A) Reduction of proteasome activity by TSA. JY cells were treated (18 hours) with TSA (1 μg/ml), sodium butyrate (20 μg/ml), or lactacystin (10 μM) and the activity of the proteasome quantified using Proteasome-Glo^™ assay. Cells in 5 μl of DMSO (untreated) served as a control. Tests were done in triplicate as indicated in Material and Methods. (B) Levels of tapasin gene expression as determined by qPCR. Tapasin levels in JY cells were determined at 8 and 18 hours after TSA treatment and compared to untreated cells. GAPDH is shown as a control.

Reduced killing of TSA-treated tumor cells by CD8 T cells

Recognition of MHC I/peptide complex is the structural basis for specific killing of tumor cells by cytotoxic T lymphocytes. To assess whether the decreased presentation by TSA treatment had an effect on tumor target recognition and lysis we performed killing assays using three different HLA-A2⁺/hTRT⁺ cancer cells (JY, Mel629.38 and LnCap) as targets and in vitro generated human p540-specific CTL as the effectors. The percent lysis of TSA-treated T2 cells pulsed with p540 was similar to that of untreated targets (Figure 6A), suggesting that TSA does not interfere with cell surface MHC loading with peptide, nor does it change the susceptibility to lysis of these tumor cells. Interestingly, after TSA treatment the lysis of all three tumor cells was markedly diminished compared to untreated cells (Figure 6A). The specificity of the CTL clone for the HLA-A2/p540 complex was demonstrated by a canonical cold target inhibition assay (11). As shown the lysis of JY cells was specifically and completely competed for by T2 cells pulsed with p540 but not T2 cells pulsed with the control HIV gag peptide (Figure 6B). To further prove the functional relevance of HDAC inhibition-induced reduced presentation of p540, new cold target inhibition experiments were performed using TSA-treated JY, 629 and LnCap cells as hot targets. We reasoned that if TSA treatment diminishes presentation of the HLA-A2/p540 complex rendering cells less susceptible to lysis by CTL, inhibition by cold targets pulsed with p540 would be more effective diminishing killing of hot targets. As shown, killing TSA-treated targets was markedly reduced in the presence of p540-pulsed cold inhibitors compared with control cold inhibitors (Figure 6c and D). Of note this effect was dose dependent and almost complete at the highest cold:hot ratio. Collectively, these results demonstrate that HDAC inhibition decreases presentation and this negatively affects killing by specific CTL.

Discussion

We show that HADC inhibition in human cancer cells diminishes the presentation of high affinity peptides of telomerase reverse transcriptase, a self tumor antigen. HDAC inhibition also down-regulates proteasome genes and tapasin a key ER-associated gene of the Class I pathway. Importantly, we demonstrate that subsequent to these changes tumor cells are killed less efficiently by cytotoxic T lymphocytes. Collectively, our data suggest that epigenetic modifications of tumor cells can affect self tumor antigens presentation.

Although HDAC inhibitors alter 2–5% of the genome depending on the cell type and the inhibitor (22, 28) we could identify transcriptional changes that directly relate to the MHC I pathway of antigen presentation provisionally excluding defects in transport along the secretory pathway. In light of our findings we propose that modifications associated with HDAC inhibition by TSA modulate hTRT presentation involves a) down-regulation of proteasome’s genes with diminished proteasome function (~30%); and b) down-regulation ~80%) of tapasin. Since tapasin impairs the quality control and sorting mechanisms for high affinity peptide complexes, a decrease in its expression and function would impact on the MHC I peptide repertoire, favoring the egress of MHC I molecules bound to unstable (low affinity) peptides. Several facts favor this hypothesis. Tapasin-deficient mice have defective immune responses with an altered MHC I/peptide repertoire (29). Deficiency of tapasin in mice results in the recognition of epitopes associated with impaired peptide processing (30). MHC I/peptide presentation in tapasin (−/−) human cells is subverted qualitatively and quantitatively (31). Thus, by altering the composition of the MHC I/peptide complexes in favor of low affinity peptide complexes, and by decreasing antigen presentation, HDAC inhibition reduces presentation of high affinity self antigen peptides in tumor cells. This scenario assumes that the abundance of hTRT relative to other ubiquinated proteins remains the same before and after HDAC inhibition. Furthermore, although the main effect of TSA at the transcriptional level is through histone acetylation, a direct effect on tapasin by acetylation of the double lysine-motif at the C-terminus (32) cannot be ruled out. That HDAC inhibition decreases the presentation of a self tumor antigen is new and at variance with recent findings that mouse tumor cell lines treated with TSA undergo the upregulation of genes of the MHC Class I pathway, including TAP and tapasin (33). A discrepancy between the present report and the studies in mouse cells is unclear and may reflect species differences. Of note, however, antigen presentation in mouse tumor cells was not studied with the aid of an anti-complex antibody as it was done here. Finally, the downregulation of tapasin documented here is surprising in light of the mouse data. Since HDAC inhibition opens the chromatin and transcriptionally activates many genes, downregulation of tapasin may be the indirect effect of genes not yet identified. Future studies will need to address this issue.

hTRT is a self antigen in human cancer cells (34) and it may be involved in immune surveillance since cancer patients, but not normal individuals, carry hTRT-specific CD8 T cells in their blood (14). Here, chromatin acetylation by HDAC inhibition leads unexpectedly to decreased antigen presentation of hTRT and diminished killing by human CTL (Figure 6A), suggesting that this may be yet another mechanism of immune evasion. Since previous studies in normal human cells showed that hTRT plays an important role in resetting chromatin architecture during DNA replication (35), the finding that HDAC inhibition diminishes hTRT presentation suggests that changes in the equilibrium between hTRT and chromatin that is normally in place to preserve the DNA damage response (36), is a fail-safe mechanism to oppose recognition of tumor cells by adaptive cellular T cell responses which are the main effector immune defense against cancer.

Taken together, the above considerations suggest that HDAC inhibition in tumor cells could lead to evasion of immune surveillance (37). Several reports indicate that tapasin is down-regulated in tumors (38–43) implying that its down regulation may somehow be linked with poor presentation of tumor antigens as well as poor tumor protection by T cells in vivo. Interestingly, impaired tapasin expression in tumor cells can be reversed by IFNγ (41) excluding genetic alterations as the underlying molecular mechanism of such deficiency. Since IFNγ may impart epigenetic modifications to chromatin, it appears as if epigenetic changes are at the origin of this type of immune evasion by cancer cells. The above considerations on potential evasion of immune surveillance need, however, to be seen in the broader context of the reported increase the expression of ligands for the NKG2D receptor on human natural killer (NK) cells by HDAC inhibitors (44, 45). Higher expression of these ligands would increase susceptibility to killing by NK cells hence counterbalancing the diminished killing by cytotoxic T cells.

Whether or not the epigenetic regulation of hTRT presentation reported herein applies to other self tumor antigens remains untested. Of interest, however, TSA activates the transcription of the melanoma associated MAGE A1, A2, A3, A12 genes in a variety of different cell lines (46). To conclude, the demonstration that epigenetic changes by HDAC inhibition affect the presentation of hTRT, a universal self tumor antigen, suggests a potentially important new link between epigenetic regulation of cancer cells and immune surveillance (37).