Elevation of c-MYC Disrupts HLA Class II–Mediated Immune Recognition of Human B Cell Tumors

This article is featured in In This Issue, p.1387

The c-MYC protein was first identified 30 y ago as a homolog of an avian retroviral oncogene (1). It is a transcription factor encoded by the MYC gene and boasts a target gene network encompassing ∼15% of all known genes (2–4). The c-MYC protein belongs to the family of basic region helix-loop-helix/leucine zipper transcription factors and its activity is dependent on the formation of heterodimers with MAX, upon which the heterodimers bind to regions of DNA with the CACGTG sequence motif (E-boxes) (5–7). The transcriptional effects of c-myc are thought to be exerted primarily through the recruitment of transcriptional cofactors involved in RNA polymerase II function as well as the recruitment of histone acetyl transferases, which acetylate lysine residues in histones and cause a more open structure of the chromatin allowing for increased transcription of target genes (8–10). To a lesser extent, c-myc exerts its functions on genes transcribed by RNA polymerases I and III and may repress transcription through interactions with the Miz-1 transcription factor (11). Overexpression of c-MYC also controls genes with a wide array of functions, ranging from cell cycle progression to differentiation to apoptosis (2, 12).

Transformation of cells by c-MYC protein involves numerous genes (9). Paradoxically, although c-MYC activity induces cell growth and differentiation, it also induces apoptosis. This is achieved through activation of the p53 tumor suppressor and inhibition of cyclin D1, as well as indirect suppression of antiapoptotic BCL2 and induction of proapoptotic BAX and Bim (9, 13, 14). Since its discovery, c-myc has come to be recognized as one of the most commonly activated oncogenes in human cancers and is observed in virtually all malignancies (13, 15). c-MYC protein expression is implicated in the cancer-related deaths of ∼100,000 people in the United States as well as millions worldwide every year (2, 15, 16). Among malignancies that have a known association with c-myc overexpression, Burkitt lymphoma (BL) may be the most prominent. Indeed, overexpression of c-myc is a hallmark of BL and activation of c-myc by chromosomal translocation is considered diagnostic for this lymphoid malignancy. In BL, c-myc is translocated to an Ig locus, resulting in its constitutive expression and cell transformation. The most prominent of these translocations is t(8:14) which is observed in 80% of BL, whereas t(8:22) and t(8:2) are observed with lesser frequencies (17, 18). These translocations are generated as accidents of Ab affinity maturation during the germinal center reaction through the action of activation-induced cytidine deaminase (19).

BL is a highly aggressive type of non-Hodgkin’s lymphoma that occurs most frequently in tropical climates, with a distribution closely following that of holoendemic malaria (20). A sporadic form is observed in other parts of the world and a third form is found associated with HIV. Malaria and EBV, separately or together, and HIV increase the risk of developing BL, presumably by polyclonal B cell stimulation including activation of activation-induced cytidine deaminase and by conferring antiapoptotic functions to cells hit by the c-myc translocation, but the precise contribution remains to be defined (20–23). BL is typically treated effectively with aggressive chemotherapy in young patients, but inferior responses are observed in adults (especially the elderly) and immunodeficient patients (24). In addition, older and immunodeficient patients are less tolerant of the aggressive chemotherapy required and show increased signs of treatment-associated toxicities. This gap in treatment for these patient groups highlights the need for exploration into improved treatment options that would display lower levels of toxicity. The most ideal treatments would harness the immune system of the individual to target malignant cells. In EBV-positive BL, EBV-encoded nuclear Ag (EBNA)-1 is expressed as the only viral protein and it poorly stimulates cytotoxic CD8⁺ T cells because of its low immunogenicity (25–29). As a result, CD8⁺ T cell responses to BL are weak and unsustained. Although multiple defects in class I Ag presentation and immune escape have been reported (25–29), little is known about disruption of class II presentation by malignant tumors. However, effective tumor immune responses usually involve the stimulation and maintenance of tumor specific CD8⁺ HLA class I–restricted cytotoxic T cells (CTL) and tumor-specific CD4⁺ class II–restricted helper T cells (30–32). Several groups have also shown that HLA class II–restricted CD4⁺ CTL could be generated against BL as well as non-Hodgkin’s follicular lymphoma (33–36), suggesting the feasibility of using sufficient tumor-specific CD4⁺ T cells to eliminate B cell tumors. Most B cell tumors—including BL, express class II and could be targets for helper CD4⁺ T cells. Increased expression of c-MYC imparts reduced immunogenicity to BL cells by antagonizing the NF-κB and IFN pathways (28, 29). In EBV-immortalized B-lymphoblastoid cell lines (B-LCL), c-MYC is expressed at low levels, and these cells retain a highly immunogenic phenotype with blast formation observed during growth because of expression of LMP1 that activates the canonical as well as the noncanonical NF-κB pathway (37). However, if B-LCLs are transfected to express measurable c-MYC, tumor immunogenicity is greatly diminished and suspension growth is observed, which corresponds to the phenotypes of BL cell lines (29, 38). Clearly, the expression of c-MYC plays an important role in reducing the immunogenicity of BL.

Although c-MYC expression has been described to have multiple effects on malignant cells, the role it plays in HLA class II–mediated Ag presentation and immune recognition has not been reported. Recently, it has been shown that the expression of invariant chain (Ii) regulates the repertoire of tumor peptides presented by class II-positive tumor cells (39). The absence of Ii has also been shown to facilitate CD4⁺ T cell responses (39, 40). Thus, expression levels of the components of the class II pathway such as Ii, CLIP, and HLA-DM/DO molecules in B cells as well as tumors may modulate immune recognition. In this paper, studies revealed that cells expressing high levels of c-MYC are poor stimulators of CD4⁺ T cells. c-MYC expression in these cells was tied to alterations in the expression of HLA-DM, the lysosomal thiol reductase (γ-IFN–inducible lysosomal thiol reductase [GILT]), and a 47-kDa enolase-like protein. We further show that treatment of c-MYC–overexpressing cells with a c-MYC inhibitor led to decreased c-MYC expression, restoration of HLA-DM concurrent with decreased HLA-DO molecules, and partial restoration of class II–mediated Ag presentation. Taken together, these results strongly suggest that c-MYC expression plays a key role in immune evasion via altering HLA class II Ag presentation and opens the door for novel therapeutics to reverse this deficiency in BL.

Human B-LCLs Frev, Priess, and 6.16 were cultured in IMDM (Mediatech, Manassas, VA) supplemented with 10% bovine growth serum (Hyclone, Logan, UT) and 50 U/ml penicillin and 50 μg/ml streptomycin (Mediatech) (41). The B-LCL EREB2-5 (42) was cultured in RPMI 1640 medium (Mediatech) supplemented with 10% FBS (Invitrogen, Carlsbad, CA), 50 U/ml penicillin/50 μg/ml streptomycin and 2 μM estrogen, 1% l-glutamine (Mediatech), and 50 μM 2-ME (Invitrogen) (43, 44). Human BL (Ramos) and BL-like cell lines (Nalm-6, A1, and P493-6) were maintained in complete RPMI 1640 medium supplemented with 10% FBS, 50 U/ml penicillin and 50 μg/ml streptomycin, and 1% l-glutamine (41, 45, 46). The wild-type B-LCL, Frev, constitutively expresses HLA-DR4 molecules. Priess cells are homozygous for the expression of HLA-DR4 (DRA*0101 and DRB*0401) molecules. The P493-6 cell line was cultured in the presence of 2 μM estrogen with or without 5 μg/ml tetracycline to control expression of EBNA-2 and c-myc (P493-6.DR4.est and P493-6.DR4.est.tet, respectively) (38). EREB2-5, 6.16, Nalm-6, Ramos, A1, and P493-6 cell lines were retrovirally transduced with a common allele of HLA-DR4 (DRB1*0401) with linked drug selection markers for hygromycin or histidinol resistance (47) to generate EREB2-5.DR4, 6.16.DR4, Nalm-6.DR4, Ramos.DR4, A1.DR4, and P-493-6.DR4 cell lines. 6.16.DR4 cells were further transfected with DMα and DMβ for constitutive expression of HLA-DM molecules to generate 6.16.DR4.DM (41). Frev cells were also transfected with an empty vector or c-myc to generate Frev.vec and Frev.c-myc cell lines.

T cell hybridomas 2.18a and 1.21 recognize Igκ residues 188–203 and 145–159, respectively, and were generated by immunization of DR4-transgenic mice as described previously (44, 48). The T cell hybridoma line 17.9 (provided by D. Zaller, Merck Research Laboratories, Rahway, NJ) responds to human serum albumin (HSA) residue 64–76K (49), was cultured in RPMI 1640 medium with 10% FBS, 50 U/ml penicillin/50 μg/ml streptomycin, and 50 μM 2-ME (Invitrogen). Institutional approval (HR#17159) for the use of human tissue and tumor samples was obtained. Lymph nodes and blood samples were obtained from two lymphoma patients (TB#2952 and TB#7378) through our Hollings Cancer Center Tissue Bank (Medical University of South Carolina, Charleston, SC). TB#2952 and TB#7378 tumors are EBV positive as analyzed by the Biorepository and Tissue Analysis Shared Resource of the Hollings Cancer Center at the Medical University of South Carolina. Blood samples were also obtained from two healthy individuals with written consent (approved protocol HR number 17159). Of these (C#16 and C#101) and of two tumors (TB#2952 and TB#7378), B cells were isolated using a B Cell Isolation Kit II (number 130-091-151). A portion of B cells from a healthy individual (C#16) and from a BL tumor (TB#2952) were then infected in vitro with EBV (41), and transfected with HLA-DR4 (DRB1*0401) as described previously (43). These DR4-expressing early passage cells were used in the functional class II Ag presentation and T cell recognition assays. Cells were cultured in RPMI 1640 medium with 10% FBS, 50 U/ml penicillin/50 μg/ml streptomycin, and 50 μM 2-ME (Invitrogen). An explanatory table was also included listing the type and source of the cells and peptide specificity for the T cell hybridomas (Table I).

HSA and human Igκ were purchased from Sigma-Aldrich (St. Louis, MO). HSA_64–76K (sequence: VKLVNEVTEFAKTK), human Igκ immunodominant κ_188–203 (sequence: KHKVYACEVTHQGLSS, referred to as κI), and subdominant κ_145–159 (sequence: KVQWKVDNALQSGNS, referred to as κII) peptides were produced using Fmoc technology and an Applied Biosystems Synthesizer as described previously (44, 47, 49–51). The Igκ_188–203 and Igκ_145–159 peptides were labeled as indicated at the α amino termini by the sequential addition of two molecules of Fmoc-6-aminohexanoic acid, followed by a single biotin to yield the sequence biotin-aminohexanoic acid-aminohexanoic acid-peptide. Reversed-phase HPLC purification and mass spectrometry were used to analyze the peptide and showed a peptide purity > 99%. These DR4-restricted peptides were dissolved in PBS and stored at −20°C until use.

B-LCL and BL cells were incubated with the whole Ags (Igκ or HSA), or their synthetic epitopes (Igκ_188–203, Igκ_145–159, or HSA_64–76K) for 4 h at 37°C in the appropriate cell culture media (43, 44, 51, 52). Cells were then washed and cocultured with the T cell hybridomas (12.18a and 1.21 cells recognize Igκ_188–203 and Igκ_145–159 peptides, respectively; 17.9 cells recognize HSA_64–76K peptide) for 24 h at 37°C. In separate assays, P493-6.DR4 cells were left untreated or treated for 24 h with 50 μM of the c-MYC inhibitor 10058-F4 (Calbiochem, Billerica, MA). Peptides and Ags were added for the last 4 h, and the cells were washed and cocultured with the appropriate peptide specific T cell hybridomas for 24 h. For endogenous Ag presentation assay, cells (Priess EREB2-5.DR4, A1.DR4, and P493-6.DR4), which innately express Igκ were cocultured with the appropriate epitope- or peptide-specific T cell hybridomas for 24 h (47, 51, 53). For all assays, following coculture with the T cell hybridoma, T cell production of IL-2 was quantitated by ELISA (52, 54). Assays were repeated in triplicate with SE for triplicate samples within a single experiment being reported.

Cell lysates obtained from Frev.vec, Frev.c-myc, EREB2-5.DR4, and A1.DR4 were analyzed by Western blotting as previously described for expression of HLA-DR, Ii, and HLA-DM with β-actin (Santa Cruz Biotechnology, Santa Cruz, CA) as a loading control (47, 55–57). Nuclear lysate from Frev.vec, Frev.c-myc, EREB2-5.DR4, A1.DR4, Priess, P493-6.DR4, P493-6.DR4.est.tet, 6.16.DR4.DM, Nalm-6.DR4, and Ramos.DR4 were analyzed by Western blotting for expression of c-MYC (Santa Cruz Biotechnology), EBNA-1 and EBNA-2 (a gift from Dr. E. Kremmer, Institute of Molecular Immunology, German Research Center for Environmental Health) with β-actin as the loading control. In separate assays, P493-6.DR4 cells were left untreated or treated for 24 h with 50 or 100 μM of the c-MYC inhibitor 10058-F4. Following treatment, cells were harvested, and nuclear lysate was obtained and analyzed by Western blotting for expression of c-MYC with β-actin as the loading control. Densitometry was performed using a ChemiDoc XRS station (Bio-Rad, Hercules, CA) where the protein bands were analyzed using the Quantity One 4.6.3 software (Bio-Rad). Relative protein expression levels were stated as a ratio of specific proteins expressed/β-actin for each sample. Data are representative of at least three separate experiments.

Frev, Frev.c-myc, EREB2-5.DR4, A1.DR4, P493-6.DR4, and P493-6.DR4.est/tet cells were stained with Abs against HLA-DR (L243 Ab) and HLA-DR4 (359-F10 Ab), followed by a secondary FITC conjugate (46, 50, 52). P493-6.DR4 and P493-6.DR4.est/tet cells were also stained with Abs against CLIP (cer-CLIP Ab), Ii (Pin 1.1 Ab), HLA-DM (MAP1.1-DM Ab), and HLA-DO (Mags.DO5-FITC and isotypes were a gift from L. Denzin, Memorial Sloan-Kettering Cancer Center, New York, NY). P493-6.DR4 cells treated with vehicle alone or c-MYC inhibitor 10058-F4 were also stained with Abs and matched isotype controls for the detection of CLIP and HLA-DM molecules. Primary or early passage B cells and tumors were stained with appropriate Abs as described above for the detection of HLA-DR, CLIP, HLA-DM, and HLA-DO molecules. Samples were then analyzed on FACScan using CellQuest software (BD Biosciences, Mountain View, CA). Background fluorescence was evaluated using irrelevant isotype-matched Abs (NN4 and IN-1) as described previously (46, 50, 52).

Flow cytometric peptide binding assay was used to determine peptide binding to HLA-DR4 molecules on P493-6.DR4+est/tet (c-myc^low) and P493-6.DR4 (c-myc^high) cells using a modified method (58). Briefly, P493-6.DR4 (4 × 10⁵) cells cultured under c-myc on/off conditions were washed twice with ice-cold PBS containing 1% BSA and incubated with either vehicle alone or 40 μM biotin-labeled HSA_64–76K peptide for overnight separately at 4 and 37°C. Cells were washed and stained with FITC-conjugated streptavidin (sc-2865; Santa Cruz Biotechnology) for 30 min at 4°C and analyzed by FACScan using CellQuest software (BD Biosciences).

Acid eluted surface proteins were obtained from P493-6.DR4 and P493-6.DR4.est.tet cells, run on a nonreducing gel, and stained with Coomassie blue (41). Gel plugs were excised and placed in an Eppendorf tube. Each plug was washed with 50 mM ammonium bicarbonate and destained using 25 mM ammonium bicarbonate in 50% acetonitrile. The plugs were dehydrated with 100% acetonitrile and dried in a speedvac. Each gel plug was covered with Proteomics Grade Trypsin (Sigma-Aldrich) and incubated at 37°C overnight. The supernatant was collected in a clean dry Eppendorf tube. Peptides were further extracted with one wash of 25 mM ammonium bicarbonate and three washes of 5% formic acid and 50% acetonitrile. The supernatant was collected and pooled after each wash then dried down in a speedvac to ∼1 μl. Prior to analysis, the samples were reconstituted with 10 μl of 0.1% trifluoroacetic acid. Samples were then concentrated with a C18 Ziptip (Millipore) and eluted with 0.1% trifluoroacetic acid, 50% acetonitrile, and 7.0 mg/ml α-cyano-4-hydroxycinnamic acid directly onto the MALDI target.

After protein extraction, digestion, and elution, the protein spots were dried completely, and the MALDI target plate was loaded into the Applied Biosystems 4800 Proteomics Analyzer (50). An external calibration was preformed prior to analyzing samples using the manufacturer’s standards and protocols. Samples were analyzed in batch mode using 2000 laser shots per spectrum. First, peptide mass maps were acquired over the m/z range of 800–3500 in reflectron mode with a delayed extraction time optimized for m/z 2000 by averaging 2000 scans to locate peaks of peptide origin. The next batch run performed mass spectroscopy (MS)-MS analyses to obtain sequence data on the 20 most abundant peaks from the MS analysis. Upon completion of the batch processing, the data were exported into the GPS Explorer data processing system for interpretation and identification. The MASCOT database-searching algorithm analyzed the data and summarized the results in report format. Database searches were performed using two missed cleavages and one differential modification of methionine oxidation. The top 20 matches were reviewed prior to assigning confident protein identifications.

The data are expressed as the mean (± SD) and analyzed using Student t test or one-way ANOVA, with p ≤ 0.05 considered statistically significant. A nonparametric Wilcoxon rank-sum test was also used when comparing band intensity obtained from densitometric analysis of distinct protein bands detected in Western blotting of different cells.

Although EBV-positive BL cells express EBV Ags, these cells are poorly recognized by HLA class I–restricted CD8⁺ T cells because of diminished viral epitope presentation via the HLA class I pathway. Although the Ag-specific lysis of tumors is predominantly a function of CD8⁺ T cells, HLA class II–restricted CD4⁺ T cell function is crucial in maintaining sustained immune responses to tumors (59–61). Given the high levels of c-MYC expression associated with BL, the effect of c-MYC overexpression on HLA class II Ag presentation was examined. For these studies, several approaches were used to alter c-MYC levels in human B cell lines to parallel c-MYC expression in BL (Table I). The B-LCL Frev was transfected with either an empty vector (Frev.vec) or c-myc (Frev.c-myc). The EREB2-5 cell line was generated from primary human B cells by coinfecting with an EBNA-2–deficient P3HR1 virus and recombinant EBV encoding an estrogen receptor EBNA-2 fusion protein to generate a B-LCL whose proliferation is dependent on estrogen (62, 63). These cells were then further transfected for constitutive expression of c-myc to generate the A1 cell line, capable of proliferation in the absence of EBNA-2 (64). Both EREB2-5 and A1 were transduced for expression of a common HLA-DR4 allele. Whole-cell lysates from these cells were analyzed by Western blotting for expression of c-MYC, EBNA-1 and EBNA-2 proteins (Fig. 1A). Expression of c-MYC was clearly elevated in Frev.c-myc and A1.DR4 cell lines when compared with Frev.vec and EREB2-5.DR4, respectively. For reference, the B-LCL line Priess displays low c-MYC expression. EBNA-1 is expressed at comparable levels in each cell line, whereas EBNA-2 expression is detected in Frev.vec and EREB2-5.DR4, which carries a mutant form of EBNA-2, and to a much lesser extent in the c-MYC–overexpressing cells. The c-MYC–overexpressing cells (Frev.c-myc and A1.DR4) also switched from a B-LCL (Frev.vec and EREB2-5.DR4) growth pattern (blast formation) to a BL growth pattern (suspension) (Fig. 1B). Frev.vec, Frev.c-myc, EREB2-5.DR4, and A1.DR4 were then analyzed by flow cytometry for surface expression of HLA class II DR4 (Fig. 1C). Overexpression of c-MYC did not affect expression of surface HLA class II proteins as each cell line was found to express comparable levels. The results shown in Fig. 1D indicate that c-MYC–overexpressing cells display a reduced capacity to present HSA_64–76K peptide to stimulate CD4⁺ T cells via the class II pathway.

Although c-MYC overexpression did not affect the surface expression of HLA class II proteins, it remained possible that the intracellular components of the class II pathway were affected, thus disrupting CD4⁺ T cell recognition of BL cells. Frev.vec, Frev.c-myc, EREB2-5.DR4, and A1.DR4 whole-cell lysates were analyzed by Western blotting for expression levels of the class II pathway components HLA-DR, Ii, HLA-DM, and a lysosomal thiol-reductase GILT (Fig. 2A, 2B). In parallel assays, the expression of cell surface HLA class II proteins was also analyzed (Fig. 2C). Expression levels of both intracellular and cell surface HLA-DR molecules remained unchanged or were minimally altered in the c-MYC low and c-MYC high cells. There was also no significant change observed for Ii protein expression as analyzed by the Wilcoxon rank-sum test. However, HLA-DM and GILT proteins were significantly decreased in the c-MYC-high cells as compared with c-MYC-low cells (p ≤ 0.0022), suggesting possible implications for c-MYC–induced effects on HLA class II Ag presentation.

To examine whether BL cells expressing high levels of c-MYC have a diminished capacity for HLA class II–mediated Ag presentation, whole-cell lysates were first obtained from the B-LCL 6.16.DR4.DM and two wild-type BL lines Nalm-6 and Ramos and analyzed by Western blotting for expression of c-MYC proteins (Fig. 3A). It was found that the BL lines Nalm-6.DR4 and Ramos.DR4 expressed significantly higher levels of c-MYC than the B-LCL 6.16.DR4.DM. Each cell line was also incubated with either the whole Igκ Ag or synthetic version of κ epitope Igκ_188–203 or Igκ_145–159, followed by coculture with the appropriate peptide-specific T cell hybridomas for 24 h. Following incubation, the culture supernatant was analyzed by ELISA for IL-2 as a measure of T cell stimulation. Nalm-6.DR4 and Ramos.DR4 displayed a sharply diminished capacity to process and present Igκ-derived epitopes to stimulate CD4⁺ T cells when compared with 6.16.DR4.DM (Fig. 3B, left panel). Similarly, BL cells failed to functionally present synthetic peptides to activate CD4⁺ T cells via the HLA class II pathway (Fig. 3B, right panel).

To investigate whether both exogenous and endogenous routes of class II Ag presentation are disrupted in BL, a number of B-LCL (Priess.DR4 and EREB2-5.DR4) and the c-MYC–overexpressing counterpart of EREB2-5 cells (A1.DR4 and BL-type cells) were analyzed by biochemical and functional assays. It is important to note that these cell lines innately express endogenous Igκ Ag (data not shown). Western blot analysis of whole-cell lysates from these cells showed that A1.DR4 cells expressed high levels of c-MYC, whereas B-LCL and B-LCL–type cells expressed low levels of c-MYC proteins (Fig. 3C). To further investigate whether BL-type cells expressing high levels of c-myc have a diminished capacity for HLA class II–mediated exogenous Ag presentation, cells (Priess, EREB2-5.DR4, and A1.DR4) were incubated with the whole Ag HSA for 4h (Fig. 3D, left panel). Cells were then washed and cocultured with the HSA_64–76K epitope specific T cell hybridoma (17.9) for 24 h. The production of IL-2 in the culture supernatant was quantitated by ELISA. Priess and EREB2-5.DR4 cells, which express low levels of c-MYC, efficiently presented HSA to CD4⁺ T cells, whereas high c-MYC–expressing A1.DR4 cells failed to optimally present the epitope to the same T cell hybridoma line.

To study the presentation of endogenous Ag, we used two T cell hybridoma lines, 2.18a and 1.21, which recognize κ_188–203 and κ_145–159 peptides, respectively. We then cocultured Priess, EREB2-5.DR4, and A1.DR4 cells expressing endogenous Igκ with 2.18a and 1.21 T cell hybridomas followed by analysis of IL-2 production in the culture supernatant (Fig. 3D, right panel). A1.DR4 cells, which express high levels of c-MYC failed to optimally present κ_188–203 and κ_145–159 epitopes to CD4⁺ T cells. These results strongly suggest that overexpression of c-MYC diminishes both exogenous and endogenous pathways of class II Ag presentation and CD4⁺ T cell recognition of c-myc–overexpressing B-(BL-type) cells.

To determine the role of c-MYC in diminished class II presentation in BL cells, we used the cell line P493-6.DR4, which can be reversibly shifted from an LCL-like phenotype (EBNA-2 on, exogenous c-myc off) to a BL-like phenotype (EBNA-2 off, c-myc on) by addition and withdrawal of estrogen plus tetracycline. P493-6 cells expressing DR4 were cultured in the presence (c-myc off) or absence (c-myc on) of estrogen plus tetracycline (P493-6.DR4.est.tet versus P493-6.DR4 cells) (Fig. 4A). When cultured in the presence of estrogen plus tetracycline, c-MYC expression was markedly reduced (Fig. 4A). Cell surface HLA-DR and HLA-DR4 molecules were not significantly altered in P493-6.DR4 versus P493-6.DR4.est.tet cells regardless of c-MYC expression (Fig. 4B). Cells grown under c-myc on/off conditions were used in T cell assay to determine HLA class II–mediated Ag presentation capability. When cultured in the absence of tetracycline, P493-6.DR4 cells showed a sharply diminished capacity to stimulate T cells via HLA class II as compared with P493-6.DR4.est.tet cells cultured in the presence of estrogen plus tetracycline (Fig 4C). Because P493-6 cells express endogenous κ Ag, P493-6.DR4 and P493-6.DR4.est.tet cell lines were cocultured with the κ_188–203 and κ_145–159 peptide specific T cells, followed by the quantitation of T cell production of IL-2 (Fig. 4D, right panel). Likewise, P493-6.DR4 showed a diminished capacity to present endogenous κ epitopes to stimulate CD4⁺ T cells via the class II pathway. Taken together, both exogenous and endogenous presentation of class II–restricted peptides were partially restored by downregulating c-MYC expression and shifting P493-6.DR4 cells toward an LCL-phenotype by the addition of estrogen plus tetracycline.

To investigate whether peptide binding to surface HLA-DR4 molecules was altered in c-myc–high cells, we performed binding assays using biotin-labeled HSA peptide (HSA_64-76K) and FITC-labeled avidin, followed by flow cytometric analysis. The binding of DR4-restricted HSA peptide was very low at 4°C (Fig. 4E). But there was a trend that exogenous peptide loading by class II molecules was better in low-c-myc cells as compared with that observed in high-c-myc cells. This difference was more pronounced when the binding assay was performed in fixed cells at 37°C (Fig. 4F). The HSA peptide bound much better to DR4 molecules on low-myc cells as compare with those of high-c-myc cells. These data suggest that the observed differential peptide binding to class II molecules could contribute to reduced peptide presentation and diminished T cell recognition of high-c-myc cells.

To further examine this finding, P493-6.DR4 cells were treated with the c-MYC inhibitor 10058-F4, a small molecule inhibitor that decreases c-MYC expression in tumor cells (65). Whole-cell lysates were obtained and analyzed by Western blotting for expression of c-MYC protein (Fig. 5A). A dose-dependent decrease in c-MYC expression was consistent with other reports that 10058-F4 treatment results in diminished cellular c-MYC expression. To determine whether inhibition of c-MYC can restore T cell recognition, P493-6.DR4 cells were treated with 50 μM 10058-F4 for 24 h, followed by incubation with the whole HSA for 4 h. Cells were also incubated with the synthetic version of HSA epitope (HSA_64–76K) for 4 h. After incubation, cells were washed and fixed with 1%paraformaldehyde and cocultured with the HSA epitope-specific T cell hybridoma 17.9 for 24 h. The culture supernatant was then analyzed by ELISA for IL-2 as a measure of T cell stimulation (Fig. 5B, left panel). It was found that inhibition of c-MYC expression partially restored HLA class II–mediated Ag/peptide presentation and CD4⁺ T cell recognition of BL-like P493-6 cells.

To determine whether inhibition of c-MYC can also restore the presentation of endogenous Ag, we employed A1.DR4 and P493-6.DR4 cells (both cell lines express high levels of c-MYC) and cocultured with the κ_188–203 epitope–specific T cell hybridoma line 2.181 for 24 h. After incubation, T cell production of IL-2 production in the culture supernatant was quantitated by ELISA (Fig. 5B, right panel). It was found that inhibition of c-MYC protein by 10058-F4 significantly restored HLA class II-mediated presentation of endogenous Ag Igκ. Taken together, these data suggest that overexpression of c-MYC diminishes both exogenous and endogenous pathways of class II-restricted Ag presentation and CD4⁺ T cell recognition.

Flow cytometric analysis of P493-6 cells showed that c-MYC inhibitor 10058-F4 dose-dependently reduced cell surface CLIP expression (29 versus 13 and 29 versus 7%) (Fig. 5C, upper panel), suggesting that c-MYC affects the number of peptide receptive class II complexes in the cell that may be linked to DM/DO. Further study using intracellular staining of DM/DO molecules confirmed that inhibition of c-MYC–elevated DM molecules while downregulating DO proteins (Fig. 5C, lower panel). These data suggest that the c-MYC inhibitor may partially restore CD4⁺ T cell recognition by altering DM/DO protein ratio in BL-type cells and that the disruption of c-MYC expression or function might hold potential in restoring CD4⁺ T cell recognition of BL.

Having established that treatment with the c-MYC inhibitor 10058-F4 was sufficient to partially restore class II–mediated Ag presentation in P493-6.DR4 cells, we sought to determine the mechanism by which c-MYC impairs Ag presentation via the MHC class II pathway. To this end, P493-6.DR4 cells proliferating in the absence of estrogen plus tetracycline (c-myc program) were shifted back to a B-LCL phenotype by addition of estrogen plus tetracycline (38, 63) and then analyzed by flow cytometry for expression of cell surface CLIP and intracellular Ii as well as DM/DO molecules. Although no significant change was observed in Ii expression (93 versus 94%), cell surface CLIP expression was markedly downregulated (28 versus 12%) (Fig. 6A). The shift from a BL-like (P493-6.DR4 and A1.DR4) to a B-LCL-phenotype (P493-6.est.tet) led to a sharp increase in HLA-DM molecules (mean fluorescence intensity: 194 versus 801) (Fig. 6B). This increase in HLA-DM molecules was concurrent with a slight decrease in HLA-DO molecules (mean fluorescence intensity: 274 versus 158) and a significant increase (*p < 0.001) in the DM/DO protein ratio (Fig. 6C). These data support the notion that the decrease in CD4⁺ T cell recognition of BL and BL-like P493-6 cells can be linked to alteration of the DM/DO protein ratio and increased surface CLIP expression as a result of overexpression of c-MYC protein.

We have recently shown that a 47-kDa enolase-like acid labile protein is highly expressed in B-LCL, and that this protein is nearly absent or expressed at a greatly reduced level in BL cells (41). This enolase-like protein facilitated B–T cell interactions by enhancing HLA class II peptide display to T cells. Studies were conducted to determine whether c-MYC affects the levels of this 47-kDa protein. The 47-kDa protein band was highly abundant in B-LCL type P493-6.DR4.est/tet cells as compared with BL type P493-6.DR4 cells (Fig. 6D). MALDI-TOF/TOF mass spectrometric analysis confirmed that the 47-kDa protein band was an α-enolase 1-like molecule (accession number 4503571; URL: ipi.HUMAN.V3.71). It is important to note that P493-6.DR4 cells express very high levels of c-MYC proteins relative to P493-6.DR4.est/tet cells and that elevated c-MYC disrupts CD4⁺ T cell recognition of P493-6.DR4 cells (Fig. 4). In addition, BL-type P493-6.DR4 cells express lower levels of DM proteins as compared with B-LCL–type P493-6.DR4. est/tet cells (Fig. 6A, 6B). Collectively, these data suggest that the overexpression of c-MYC downregulates HLA-DM as well as the 47-kDa enolase-like acid labile protein, leading to disruptions in immune recognition of BL.

To investigate whether high CLIP expression, a decrease in the DM/DO ratio, and impaired Ag presentation via the MHC class II pathway are also observed in primary BL, the TB#2952 lymphoma was subjected ex vivo to Western blotting and flow cytometry and was compared with an early passage of an EBV-transformed B cell line of a healthy individual (C#16). Western blot analysis showed that TB#2952 expressed higher levels of c-MYC but equivalent intracellular DRβ proteins as compared with C#16 (Fig. 7A). T cell assay demonstrated that TB#2952 lymphoma, which expressed higher levels of c-MYC, failed to stimulate CD4⁺ T cells. Likewise, early passage TB#2952 cells that had been subjected to infection with EBV (TB#2952[+EBV]) had diminished capacity to stimulate CD4⁺ T cells as compared with EBV-infected control cells (C#16[+EBV]) (Fig. 7B), indicating that in vitro EBV infection did not restore the deficiency in Ag presentation observed in BL tumor cells ex vivo.

Flow cytometric analyses of EBV-transformed early passage control C#16 and TB#2952 BL tumor cells showed comparable levels of cell surface HLA-DR molecules, whereas C#16 expressed higher levels of HLA-DM and lower levels of DO and CLIP proteins (Fig. 7C, left panels). Exactly the same was found with another BL tumor (TB#7378) and control B cell pair of a healthy individual (C#101) (Fig. 7C, right panel). These data further support our findings that the alteration of DM/DO ratio and upregulation of CLIP cause reduced CD4⁺ T cell recognition of B cell lymphoma.

In the 30 y since its discovery, the c-MYC protein has come to be recognized as one of the most commonly activated oncogenes in human cancers. As a transcription factor that controls up to 15% of all known cellular genes, the effects of c-MYC are wide ranging. Cellular functions and host defenses are frequently controlled by genes within the c-MYC target network including immune regulation (2). Despite its far-reaching effects, c-MYC protein has never been reported to have an effect on the HLA class II pathway of Ag presentation. We report in this paper for the first time, to our knowledge, that c-MYC overexpression results in decreased HLA class II–mediated immune recognition of BL.

The HLA class II pathway of Ag presentation is primarily involved in the presentation of exogenous Ags that have been internalized and degraded by APCs. Although HLA class I is expressed on every nucleated cell of the body, HLA class II expression is limited to the professional APCs: macrophages, dendritic cells, and B cells (66–69). In the class II pathway, extracellular Ags are internalized and degraded in endolysosomal compartments. Antigenic peptides are generated with the aid of cathepsins and GILT, and deficiencies in these proteins lead to the production of peptides that are not optimal for class II–mediated Ag presentation (47, 56, 70, 71). HLA class II is synthesized in the endoplasmic reticulum lumen bound to Ii, which occupies the binding groove to prevent inappropriate peptide binding and to aid in transporting class II through the Golgi and the endolysosomal compartments containing the antigenic peptides (72–74). In this study, Ii is degraded by cathepsins leaving CLIP occupying the class II binding groove (66, 68, 73, 75). HLA-DM then mediates the removal of CLIP and the binding of peptide to the class II binding groove (75–79). Removal of CLIP and class II peptide binding is partially modulated by the nonclassical HLA-DO, which inhibits the activity of HLA-DM. Shifts in the DM/DO ratio may cause an increase in destabilized cell surface class II–CLIP complexes, thus impairing the HLA class II pathway of Ag presentation (78, 80–83). Most B cell tumors including BL, express class II molecules, leading to the assumption that these cells should be ready targets for tumor-specific CD4⁺ T cells. However, we have found that the presentation of native Ag as well as peptides is influenced by elevated cellular c-MYC. Native Ag requires cellular internalization, processing, and epitope loading by class II within endocytic compartments. To better understand this defect in class II presentation in c-myc^high cells, we tested peptide binding to class II molecules at low temperature (4°C) in live cells and at 37°C in paraformaldehyde-fixed cells. We found that the peptide binding to class II was reduced in c-myc^high cells as compared with c-myc^low cells in each case. Loss of cellular DM can affect binding of exogenous peptides in multiple ways. Our previous reports demonstrated that cellular DM levels can impact peptide presentation as some peptides are endocytosed and bind intracellular class II molecules that may undergo DM editing (49, 57, 84). We similarly showed this was true of peptides that require reduction by the lysosomal-thiol reductase GILT (47, 85). Yet, peptide binding studies in this paper suggest that cellular c-MYC levels may also alter the conformation or peptide accessibility of class II molecules. Consistent with this, the conformation of class II molecules is known to be flexible, with loss of DM impacting class II structure and peptide loading (86). Inhibition of c-MYC also resulted in improved CD4⁺ T cell recognition of BL, suggesting that the alterations in peptide binding to class II molecules along with class II components DM/DO may be regulated by elevated c-MYC in tumor cells. Because the presentation of both exogenous and endogenous Ags were found to be disrupted in BL cells, c-MYC inhibitors could prove important in enhanced immune recognition of B cell lymphomas.

We have shown in this study that overexpression of c-MYC leads to decreased class II-mediated immunogenicity in BL, by regulating the expression of key components of the HLA class II pathway. Elevation of c-MYC was found to cause a shift from the immunogenic B-LCL phenotype (growth with blast formation) to the nonimmunogenic BL phenotype (growth as suspension cells) as reported previously (38). This shift in phenotype correlated with a decrease in HLA class II–mediated immunogenicity for several different peptide Ags. The cause of the c-MYC–associated decrease in immunogenicity does not seem to be linked to expression of class II proteins as both c-myc^high and c-myc^low cells expressed comparable cell surface class II levels. Rather as shown in this paper, c-MYC overexpression influences cellular expression of DM, DO, and GILT, critical components that modulate the efficiency of class II Ag presentation. GILT reduces disulfide bonds in Ags, which allows proteins to be unfolded and further processed by acidic proteases (56, 70, 71, 87). The importance of GILT for the class II pathway has been demonstrated in GILT^−/− mice, which are defective in their ability to process Ag, although they express normal levels of class II (47, 88, 89). It is plausible that the decreased immunogenicity observed in our c-myc^high cell lines is at least partially attributable to downregulation of the critical endolysosomal reductase GILT.

Second, we show that overexpression of c-MYC leads to decreased expression of the class II pathway component HLA-DM. As discussed above, HLA-DM mediates the removal of CLIP from the peptide binding groove as well as the binding of antigenic peptides. The functional importance of HLA-DM to class II–mediated Ag presentation was first recognized in B-LCL with mutations in DM, which displayed impaired class II–peptide complex formation and defects in exogenous Ag presentation (68, 76, 80, 90). In keeping with this finding, we also demonstrate that expression of cell surface CLIP is increased in c-myc^high cells. Given that one function of HLA-DM is to mediate the release of CLIP from the class II peptide binding groove, it is expected that in conditions of decreased HLA-DM expression, CLIP would remain bound to class II proteins and be expressed on the cell surface. Having demonstrated that overexpression of c-MYC was correlated with decreased class II–mediated Ag presentation and that this effect may be attributable to downregulation of GILT and HLA-DM, we next sought to determine whether inhibition of c-MYC restored Ag presentation. Cells treated with the small molecule c-MYC inhibitor 10058-F4, partially restored class II–Ag presentation. When B cells were treated with 10058-F4, cellular c-MYC levels dropped while a shift in the intracellular DM/DO ratio was observed. Treatment of B cells with 10058-F4 also resulted in decreased CLIP binding to class II molecules, a result consistent with DM editing of CLIP and the formation of increased numbers of peptide-receptive class II molecules available to present exogenous and endogenous antigenic peptides. Further analyses with primary tumors confirmed that elevated DO and CLIP molecules disrupt CD4⁺ T cell recognition of BL.

It should be noted that inhibition of c-MYC was not sufficient to fully restore class II–mediated Ag presentation. This is in keeping with our previously published (41) and current findings that a 47-kDa α-enolase 1-like acid labile protein is nearly absent or expressed at a greatly reduced level in BL, and this deficiency may attenuate class II–mediated functional Ag presentation. The expression of α-enolase molecules has been shown to downregulate c-MYC (91, 92); thus, the 47kDa α-enolase 1-like molecule may regulate c-myc as well as components of the HLA class II pathway to promote immune recognition of BL.

We thank Elisabeth Kremmer for a gift of the anti–EBNA-2 Ab. We also thank Abigail W. Lauer (Public Health Sciences at the Medical University of South Carolina) for statistical analysis.

The authors have no financial conflicts of interest.