The major histocompatibility complex I (MHC-1) presents antigenic peptides to tumor-specific CD8+ T cells. The regulation of MHC-I by kinases is largely unstudied, even though many patients with cancer are receiving therapeutic kinase inhibitors. Regulators of cell-surface HLA amounts were discovered using a pooled human kinome shRNA interference–based approach. Hits scoring highly were subsequently validated by additional RNAi and pharmacologic inhibitors. MAP2K1 (MEK), EGFR, and RET were validated as negative regulators of MHC-I expression and antigen presentation machinery in multiple cancer types, acting through an ERK output–dependent mechanism; the pathways responsible for increased MHC-I upon kinase inhibition were mapped. Activated MAPK signaling in mouse tumors in vivo suppressed components of MHC-I and the antigen presentation machinery. Pharmacologic inhibition of MAPK signaling also led to improved peptide/MHC target recognition and killing by T cells and TCR-mimic antibodies. Druggable kinases may thus serve as immediately applicable targets for modulating immunotherapy for many diseases. Cancer Immunol Res; 4(11); 936–47. ©2016 AACR.

Major histocompatibility complex class I molecules (MHC-I) generally present short peptides from either foreign or native intracellular proteins on the cell surface in an HLA-restricted manner for recognition by CD8+ T cells via their T-cell receptor (TCR; ref. 1). MHC-I is an essential protein for CD8+ cytotoxic T-cell responses, effective vaccination, adoptive T-cell therapies, hematopoietic stem cell transplantation, and organ rejection, among many important physiologic processes and therapeutic manipulations. In addition, recently developed therapeutic TCR-based constructs and TCR-mimic antibodies are directed to MHC/peptide complexes (2–5).

Although immunotherapies for cancer, infectious disease, and autoimmune disease continue to gain use as effective therapeutic strategies, the mechanisms underlying the control of presentation of foreign antigens or self-tumor antigens are only partially understood and currently not exploited clinically (6). Reduced cell-surface presentation of tumor antigens on MHC-I is an important obstacle to effective immunotherapy with adoptively transferred T cells, TCR constructs, tumor vaccines, and TCR-mimic antibodies (7–12).

We hypothesized that signaling pathways driven by kinases may also regulate surface MHC-I expression and that these could identified in loss- or gain-of-function genetic screens using specific antibodies to detect MHC-I cell-surface expression. Previously, a genome-wide screen provided evidence that regulators of MHC-II could be identified by RNAi knockdown (13). We decided to target a mesothelioma cell line for our proof of concept, due to its robust expression of HLA and the need for more effective therapies for this disease. Moreover, immunotherapies, such as the CTLA-4 blocking antibody tremelimumab, that rely on antigen presentation on MHC-I, are currently being tested in mesothelioma (14). To identify signaling pathways that regulate HLA expression in this model, we conducted an shRNA screen of currently annotated human kinases, as it affords the immediate possibility of targeting identified kinases for which inhibitors already exist. Among “hits” identified in the screen were kinases that negatively regulate HLA, including MAP2K1 (MEK1) and EGFR. In addition, we discovered that DDR2 and MINK1 increase surface MHC-I. These pathways, the effects of their inhibition, and the positive consequences of inhibition on MHC antigen presentation, TCR-based recognition of the MHC/peptide complexes and subsequent killing were explored. The use of loss- and gain-of-function screens to uncover regulators of MHC-I could have broad implications for understanding and treating multiple diseases with pathophysiology related to antigen presentation.

Cell lines and culture conditions

After informed consent on Memorial Sloan Kettering Cancer Center (MSKCC) Institutional Review Board–approved protocols, peripheral blood mononuclear cells (PBMC) from HLA-typed healthy donors and patients were obtained by Ficoll density centrifugation. The sources for obtaining the human mesothelioma cell lines JMN and Meso34 are described previously and were verified as unique cell lines by IMPACT sequencing (Supplementary Table S1; ref. 3). HEK293T, PC9, SKMEL5, UACC257, SW480, CFPAC1, H827, H1975, H1299, and A549 were obtained from ATCC between the years 2012 and 2016 and were not further validated. The TPC1 cell was obtained from the Dr. James Fagin lab, where the cell line was validated by IMPACT sequencing, and used from 2014 to 2016 (MSKCC). The B16-F10 melanoma line was originally obtained from I. Fidler (MD Anderson Cancer Center), and used from 2015 to 2016, and was not further validated. Cell lines were maintained 2 to 3 months in RPMI supplemented with 10% FBS and 2 mmol/L l-glutamine unless otherwise mentioned. HEK293T were grown in Dulbecco's modified media with 10% FBS and 2 mmol/L l-glutamine. Cells were checked regularly for mycoplasma.

ADCC

The HLA-A*02:01–positive mesothelioma cell lines JMN and Meso34, along with the melanoma cell line SK-MEL5, were used in the ADCC assay as a target. Antibodies (3 μg/mL) ESKM (15), PRAME, or its isotype control hIgG1 were incubated with target cells and fresh healthy donor PBMCs at different effector/target ratios for 6 hours, along with indicated doses of vehicle or trametinib in RPMI supplemented with 10% FBS. The supernatant was harvested, and the cytotoxicity was measured by a 51Cr release assay (Perkin Elmer).

Clonogenic killing assay

B16F10 cells were treated with either 0.1% DMSO or 1 μmol/L trametinib for 72 hours. B16F10 cells (1 × 104) were then used as targets and in vitro–activated Pmel T cells (5 × 104) as effectors isolated from the spleen of pmel (GP100) B6.Cg-Thy1a/Cy Tg(TcraTcrb)8Rest/J mice (The Jackson Laboratory).

Pooled RNAi screening

A custom shRNA library targeting the full complement of 526 human kinases was designed using miR30-adapted DSIR predictions refined with “sensor” rules (six shRNAs per gene) and constructed by PCR-cloning a pool of oligonucleotides synthesized on 12k customized arrays (Agilent Technologies and CustomArray) as previously described (16). The list of genes was obtained from KinBase Database (http://kinase.com/human/kinome/) and was manually curated. After sequence verification, 3,156 shRNAs (5–6 per gene) were combined with positive control HLA-A– and negative-control Renilla–targeting shRNAs at equal concentrations in one pool. JMN mesothelioma cells stably expressing the Tet-On rt-TA3 gene were used. This pool was subcloned into the TRMPV-Neo vector and transduced in triplicates into Tet-on JMN mesothelioma cancer cells using conditions that predominantly lead to a single retroviral integration and represent each shRNA in a calculated number of at least 1,000 cells (Fig. 1A). Transduced cells were selected for 6 days using G418 (1 mg/mL; Invitrogen); at each passage more than 3 × 107 cells were maintained to preserve library representation throughout the experiment. After induction, T0 samples were obtained (∼3 × 107 cells per replicate, n = 3) and cells were subsequently cultured in the presence of doxycycline (2 μg/mL) to induce shRNA expression. After 4 days (Tf), about 3 × 106 shRNA-expressing (dsRed+/Venus+) cells were sorted for each replicate using a FACSAriaII (BD Biosciences). DAPI-negative, dsRed+/Venus+ cells were sorted by FACS into three populations of BB7 low, BB7 middle, and BB7 high binding (Fig. 1). Genomic DNA from Tf samples was isolated by two rounds of phenol extraction using PhaseLock tubes (5′) followed by isopropanol precipitation. Deep-sequencing template libraries were generated by PCR amplification of shRNA guide strands as previously described (16). Libraries were analyzed on an Illumina Genome Analyzer at a final concentration of 8 pmol/L; 50 nucleotides of the guide strand were sequenced using a custom primer (miR30EcoRISeq, TAGCCCCTTGAATTCCGAGGCAGTAGGCA). Hits with lower than 100 reads from the Illumina HiSeq were eliminated because they were not above background.

Figure 1.

Screen for kinase regulators of surface HLA. A, a TRMPV-inducible shRNA retroviral vector was used for transducing JMN (HLA-A*02:01–positive human mesothelioma line). TRE is the Tet-responsive element, which drives expression of the fluorophore dsRed and the shRNA hairpin. The constitutive PGK promoter drives the Venus fluorophore along with Neomycin resistance cassette. B, Western blot and flow cytometry data showing knockdown of HLA-A using a TRMPV retroviral system with a positive control shRNA to HLA-A02. The shRen is a negative control shRNA designed against the Renilla gene. C, schema depicting the workflow pipeline for the screen of regulators of surface HLA-A. D, waterfall plot showing distribution of shRNA constructs against MAP2K1 and EGFR as log fold difference between the BB7-high sorted population and BB7-low sorted population. E, shRNA knockdown of MAP2K1 and EGFR in JMN cells validates them as a negative regulator of surface HLA-A. BB7.2 is an mAb specific for HLA-A02. shRNA against Renilla was used as a negative control, whereas an shRNA against HLA-A was used as a positive control. The screen was done in triplicate. Inhibition experiments were performed at least twice with similar results, and data shown are representative. The Student t test was done to compare each shRNA gene knockdown MFI to the shRen control (*, ≤0.05; **, ≤0.01; ***, ≤0.001; ****, ≤0.0001).

Figure 1.

Screen for kinase regulators of surface HLA. A, a TRMPV-inducible shRNA retroviral vector was used for transducing JMN (HLA-A*02:01–positive human mesothelioma line). TRE is the Tet-responsive element, which drives expression of the fluorophore dsRed and the shRNA hairpin. The constitutive PGK promoter drives the Venus fluorophore along with Neomycin resistance cassette. B, Western blot and flow cytometry data showing knockdown of HLA-A using a TRMPV retroviral system with a positive control shRNA to HLA-A02. The shRen is a negative control shRNA designed against the Renilla gene. C, schema depicting the workflow pipeline for the screen of regulators of surface HLA-A. D, waterfall plot showing distribution of shRNA constructs against MAP2K1 and EGFR as log fold difference between the BB7-high sorted population and BB7-low sorted population. E, shRNA knockdown of MAP2K1 and EGFR in JMN cells validates them as a negative regulator of surface HLA-A. BB7.2 is an mAb specific for HLA-A02. shRNA against Renilla was used as a negative control, whereas an shRNA against HLA-A was used as a positive control. The screen was done in triplicate. Inhibition experiments were performed at least twice with similar results, and data shown are representative. The Student t test was done to compare each shRNA gene knockdown MFI to the shRen control (*, ≤0.05; **, ≤0.01; ***, ≤0.001; ****, ≤0.0001).

Close modal

Relative representations of each individual shRNA were determined and compared in each given sorted population. We separated hits phenotypically into negative regulators (the population 1 SD below the mean fluorescence intensity) or positive regulators (the population 1 SD above the mean fluorescence intensity) of HLA-A*02:01. The ratio of the shRNA ranking between the high and low population was compared, with a high ratio indicating a putative negative regulator of surface HLA-A*02:01. The scoring criteria for a gene being a negative regulator of HLA-A*02:01 was based on having two or more shRNA constructs score in the top 5% for fold difference in relative representation between BB7 high population and BB7 low population, with other constructs scoring within 1 SD of the mean fold change. The gene products with at least two shRNA sequences in the top 5% ratio were selected for further validation by other methods. The same discovery pipeline was used for identifying positive regulators of HLA-A*02:01. For validation, the LT3GEPIR shRNA vector was used (ref. 17; Supplementary Table S2). Cells were transduced and selected with puromycin, then induced with doxycycline (2 μg/mL) for 96 hours before evaluating BB7, W6/32, ESK, or PRAME expression by flow cytometry.

Antibodies

Antibodies used for flow cytometry and Western blot analysis are described in Supplementary Table S3. Monoclonal antibodies (mAb) used for flow cytometry were specific for HLA-A02 (BB7.2), pan–HLA-ABC (W6/32), WT1 peptide RMF bound to HLA-A02 (ESK1), PRAME peptide ALY bound to HLA-A02 (Pr20), H2-Kb (AF6-88.5.5.3), and H2-Kq (KH114). Other antibodies used in this report are also listed in Supplementary Table S3.

Real-Time PCR

Total RNA was extracted using Qiagen RNA Easy Plus (Qiagen; #74134) after cells were treated for 48 hours with indicated inhibitor. RNA was converted into cDNA using qScript cDNA SuperMix (Quanta Biosciences Gaithersburg). Real-time assays were conducted using TaqMan real-time probes (Life Technologies) for human HLA-A (Hs01058806_g1), beta-2 microgobulin (β2M) (Hs00187842_m1), TAP1 (Hs00388677_m1), TAP2 (Hs00241060_m1), and TBP (Hs00427620_m1) with 50 ng cDNA. For assessment of gene expression using RT-PCR PerfeCTa. FastMix. II (Quanta), reactions were carried out in triplicates using standard thermocycling conditions (2 minutes at 50°C, 10 minutes at 95°C, 40 cycles of 15 seconds at 95°C, and 1 minute at 60°C). TBP was used as an internal control, and the ΔΔCT method was used for relative mRNA calculations.

Promoter-based studies

GLuc luciferase promoter was obtained from Genecoepia (GeneCoepia Rockville) with the β2M promoter cloned upstream of the GLuc enzyme. Normalization was done to SEAP (under the constitutively active SV40 promoter). Cells were seeded at 5E3 cells/well and treated with indicated drugs for 72 hours. Luminescence quantitation was assayed using the Secrete-Pair Dual Luminescence Assay Kit (GeneCoepia Rockville).

Flow cytometric studies

Cell lines were seeded in triplicate in a 6-well tissue culture plate at a density of 1E5 cells/well and allowed to adhere overnight. The next day, cells were treated with either vehicle control (0.1% DMSO), drugs, or inhibitors at indicated concentrations. Cells were then isolated at 72 hours after inhibitor treatment and washed with PBS. Cells were subsequently stained with BB7.2 (HLA-A02–specific mAb), W6/32 (HLA-ABC–specific mAb), or AF6-88.5.5.3 (H2-Kb–specific mAb; Ebiosciences). Cells were stained with propidium iodide for viability. Cells were analyzed on BD Accuri C6 flow cytometer.

Overexpression of β2M

Human β2M cDNA was cloned into the MSCV Puromycin vector.

Overexpression of mutant EGFR and NRAS

The pBABE retroviral vector encoding either EGFR harboring the L858R mutation was used to stably transduce the H1299 cell line using HEK293T/Amphoteric cells and were selected in puromycin (2.5 μg/mL) for 5 days. EGFR L858R was a gift from Matthew Meyerson, (Dana-Farber Cancer Institute, Boston, Massachusetts) (Addgene plasmid # 11012).

For overexpression of NRAS, the pBABE NRAS Q61K plasmid was used to transduce H827 cells similar to described above and selected in puromycin (2 μg/mL). pBabe N-Ras 61K was a gift from Channing Der (Addgene plasmid # 12543).

Small-molecule inhibitor studies

Compounds were obtained from SelleckChem. Drugs were used at subcytostatic doses by titration using the Cell Titer Glo assay (Promega). All drugs were used in vitro at indicated doses in 1% DMSO. Experiments were performed at least twice with similar results, and data shown are representative.

siRNA knockdown

The JMN cell line was treated with a control scrambled siRNA, or siRNA against STAT1, STAT3, and RelA. Cells were treated with the indicated drug 24 hours after siRNA knockdown for 72 hours before assaying for surface HLA-A by flow cytometry. shRNA construct details are given in Supplementary Table S2.

Transgenic EGFR L858R mouse model

FVB CC10-rtTA/EGFR-L858R mice were obtained as a kind gift from the Harold Varmus lab (Weill Cornell Medicine). Mice were bred in accordance with the MSKCC institutional review board under protocol 96-11-044. Mice used for the experiment were heterozygous for CC10-rtTA and EGFR-L858R as detected by quantitative PCR genotyping. At 4 to 6 weeks of age, mice were put on doxycycline via food pellets (625 mg/kg; Harlan-Teklad) for >6 weeks. Mice were imaged by anesthetizing under 2% isoflurane and lung field images were acquired on a Bruker 4.7T Biospec scanner (Bruker Biospin Inc.) magnetic resonance imager (MRI) in the small animal imaging core at the MSKCC. Images were analyzed with Osirix Imaging Software. Lung cancer in mice was confirmed to have reticulonodular appearances and consolidations by axial and coronal MR images, consistent with previous data published on the transgenic mice (18).

Mice were sacrificed once confirmed to have lung tumors (non-induced control mice were also used, which genotypically were identical but did not receive dox diet). The lungs were isolated and treated with collagenase IV in HBSS with Ca2+ and Mg2+ for 1 hour 37°C. Cells were then collected, blocked with mouse FcR block (Miltenyi), counted, and stained with mouse CD45 (30-F11; Biolegend), human EGFR (AY13 clone; Biolegend), and mouse H2-Kq (KH114, clone Abcam) antibodies. Flow cytometry analysis was performed on Fortessa (BD Biosciences).

CC10/L858R microarray data

Expression data from tissue isolated from WT and EGFR L858R transgenic mice were obtained from a previous study (GSE17373) and were selected for statistically significant data (P < 0.05) for PDCD1 (PD-1), CD274 (PD-L1), TAP1, TAP2, H2-Kd, and β2M gene expression between tumor-bearing EGFR L858R lung tissue and normal lung tissue (19, 20).

Pooled shRNA screen identified gene products regulating surface HLA-A*02:01

Loss- or gain-of-function screens serve as starting points for identifying new regulators of protein expression and function. We used an shRNA library against the 526 currently annotated human kinases to perform a custom-pooled screen. For each gene, six shRNA constructs were cloned into the TRMPV retroviral vector, a tetracycline regulated vector that couples a mir30-based shRNA to a red fluorescent protein, which allows easy tracking and sorting of cells’ productively expressing an shRNA (Fig. 1A; ref. 16). Knockdown of HLA-A*02:01 by use of an shRNA to this gene product in the same vector was tested as a positive control and caused strong knockdown by both Western blot analysis and flow cytometry (Fig. 1B).

The amount of MHC-I and antigen presentation on surface HLA-A*02:01 is an important determinant of efficacy for certain immunotherapies (21). We decided to use the human mesothelioma cell line JMN as the target for these studies, which has stable HLA-A*02:01 expression and which has been used as a target of MHC-I–directed therapies in vitro and in vivo (3). As a tool to show the impact of HLA modulation on antigen recognition and potential for TCR-based killing, we used TCR mimic antibodies that recognize peptide/MHC-I complexes. Knockdown of HLA-A substantially decreased the killing efficacy of the TCR mimic antibody ESKM against the JMN mesothelioma cell line (Supplementary Fig. S1). JMN was analyzed for presence of a predefined subset of mutations using the MSK IMPACT platform (Supplementary Table S1). No mutations or significant copy-number alterations were observed in the HLA-A*02:01 or β2M genes.

The JMN cell line was screened with an shRNA library against the human kinome, as described in Materials and Methods, for genes acting as negative or positive regulators of surface HLA-A, detected by flow cytometry with the HLA-A*02:01–specific mAb BB7.2 and fluorescence-activated cell sorting was used to sort populations based on HLA expression (illustrated as in Fig. 1C). The top 5 hits are listed (Supplementary Table S4).

Based on this analysis, MAP2K1 and EGFR were identified as important negative regulators of surface HLA-A*02:01. We chose to further investigate EGFR and MEK because of the availability of clinically approved drugs targeting these kinases both in NSCLC and metastatic melanoma, respectively (22), as well as extensive use of immunotherapy.

EGFR is a receptor tyrosine kinase that binds epidermal growth factor and is frequently found to be activated by mutation in NSCLC. Activated EGFR signals through multiple downstream pathways, including the MAPK pathway. shRNA constructs against MAP2K1 and EGFR showed a large increase in relative representation in the BB7-high sorted population versus the BB7-low population, indicative of a negative regulator of HLA-A*02:01 surface expression (Fig. 1D). We validated each of these genes with independent shRNA knockdown to the gene products and saw significant increases in HLA-A*02:01 by flow cytometry (Fig. 1E). These effects were seen not only with HLA-A*02:01 but also with total HLA-A, B, and C, suggesting coordinated control of all HLA surface expression, as measured with the W6/32 mAb (Supplementary Fig. S4). These findings were reproduced in multiple mesothelioma cell lines (Supplementary Fig. S3A and S3B). The RET protooncogene was also identified as a potential target, but was not further studied at this time because no inhibitor of adequate specificity was available (Supplementary Fig. S3C).

We identified examples of positive genetic regulators of HLA-A, including two putative positive regulators DDR2 and MINK1 (Supplementary Table S4; Supplementary Fig. S4A and S4B), and confirmed their activity as well using siRNA knockdown (Supplementary Fig. S4C). Therefore, the kinase screen discovered multiple positive and negative regulators of HLA expression, each of which in principle could be explored further for mechanism and clinical utility. The top five negative regulators evaluated were confirmed by additional study, whereas three of the five top positive regulators were validated (Supplementary Table S2).

The MAPK pathway regulates MHC-I

Multiple potent small molecule inhibitors exist for EGFR and MEK, with several already FDA approved, and others currently in clinical trials for various cancers (22). Of note, the initial screen was performed in a cell line with EGFR activation and an identified EGFR mutation (Supplementary Table S1; ref. 23). We tested, in multiple cell lines, the ability of inhibitors to phenocopy the loss of kinase expression leading to increased HLA-A expression seen with shRNA. Cell-surface HLA-A*02:01 expression increased in response to MEK inhibition for 72 hours with the selective MEK inhibitor trametinib in mesothelioma cell lines with activated MAP kinase signaling (Fig. 2A). JMN and PC9, a non–small cell lung carcinoma (NSCLC) cell line with an activating EGFR mutation (del E746–A750), responded to the EGFR inhibitor afatinib, whereas the Meso34 cell line without an EGFR mutation did not respond to afatinib at the same dose, demonstrating selectivity for activation mutations in the MAPK pathway leading to a response to HLA-A upregulation (Supplementary Table S1). We detected an effect of MAP kinase pathway inhibition on upregulation of HLA-A in the context of gain-of-function mutations or activation of other targets in the MAP kinase pathway, such as the KRAS G12V mutation in the SW480 and CFPAC-1 cell lines, the RET/PTC1 gene rearrangement in the TPC1 thyroid cell line, and the BRAF V600E mutation seen in the UACC257 and SK-MEL-5 melanoma cell lines (Fig. 2A). The MEKi trametinib did not affect surface HLA-A expression on normal PBMC cells, showing that this effect is specifically seen in cells with activated signaling.

Figure 2.

Use of selective EGFRi and MEKi increased cell-surface HLA-A expression and tumor antigen presentation, whereas activation of EGFR caused downregulation of MHC-I. A, MEK inhibition and EGFR inhibition for 72 hours with indicated inhibitors increased HLA-A (BB7 binding) by flow cytometry in JMN, Meso34, PC-9, UACC257, SK-MEL-5, SW480, and TPC1 cell lines. DMSO (1%) was used as a vehicle control. B, binding of TCR-mimic antibodies to peptide /MHC epitopes. In blue, use of ESK antibody to a peptide derived from the oncoprotein WT1 that is presented on HLA-A0201. Binding increased after inhibition of EGFR and MEK for 72 hours in JMN, Meso34, and TPC1. In red, the PRAME TCR-mimic antibody to an epitope of PRAME tumor antigen presented on HLA-A0201 on SKMEL5 cells. Experimental set-up was similar to A. C, treatment of JMN with 10 nmol/L EGF for 72 hours, causing activation of the downstream MAPK pathway, led to decreased surface HLA-A and total HLA-ABC. D, use of EGFRi erlotinib, along with MEKi trametinib, on H827 (EGFR E746del-A750 mutation), H1975 (L858R/T790M), H1299 (EGFR wt, NRAS Q61K), and A549 (EGFR wt/KRAS G12S) to alter surface HLA-ABC expression. The Student t test was done to compare each treatment to vehicle control. *P values annotated as in Fig. 1. E, Western blot analysis showing degree of inhibition of the MAP kinase pathway on a panel of NSCLC cell lines using 1% DMSO (D), 100 nmol/L erlotinib (E), 100 nmol/L afatinib (A), or 500 nmol/L trametinib (T). F, H1299 cells were transduced with retroviral vectors expressing EGFR L858R and were analyzed for surface pan–HLA-ABC using W6/32. Activation of EGFR is demonstrated by Western blot. G, EGFR inhibition upregulated surface HLA-ABC more than MEKi, despite equivalent inhibition of pERK output. H, EGFRi upregulated MHC-I despite downstream mutations causing constitutive MAPK activation. The NRAS Q61K mutation was introduced into H827 and cells were treated with EGFRi or MEKi as done in 2G. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Figure 2.

Use of selective EGFRi and MEKi increased cell-surface HLA-A expression and tumor antigen presentation, whereas activation of EGFR caused downregulation of MHC-I. A, MEK inhibition and EGFR inhibition for 72 hours with indicated inhibitors increased HLA-A (BB7 binding) by flow cytometry in JMN, Meso34, PC-9, UACC257, SK-MEL-5, SW480, and TPC1 cell lines. DMSO (1%) was used as a vehicle control. B, binding of TCR-mimic antibodies to peptide /MHC epitopes. In blue, use of ESK antibody to a peptide derived from the oncoprotein WT1 that is presented on HLA-A0201. Binding increased after inhibition of EGFR and MEK for 72 hours in JMN, Meso34, and TPC1. In red, the PRAME TCR-mimic antibody to an epitope of PRAME tumor antigen presented on HLA-A0201 on SKMEL5 cells. Experimental set-up was similar to A. C, treatment of JMN with 10 nmol/L EGF for 72 hours, causing activation of the downstream MAPK pathway, led to decreased surface HLA-A and total HLA-ABC. D, use of EGFRi erlotinib, along with MEKi trametinib, on H827 (EGFR E746del-A750 mutation), H1975 (L858R/T790M), H1299 (EGFR wt, NRAS Q61K), and A549 (EGFR wt/KRAS G12S) to alter surface HLA-ABC expression. The Student t test was done to compare each treatment to vehicle control. *P values annotated as in Fig. 1. E, Western blot analysis showing degree of inhibition of the MAP kinase pathway on a panel of NSCLC cell lines using 1% DMSO (D), 100 nmol/L erlotinib (E), 100 nmol/L afatinib (A), or 500 nmol/L trametinib (T). F, H1299 cells were transduced with retroviral vectors expressing EGFR L858R and were analyzed for surface pan–HLA-ABC using W6/32. Activation of EGFR is demonstrated by Western blot. G, EGFR inhibition upregulated surface HLA-ABC more than MEKi, despite equivalent inhibition of pERK output. H, EGFRi upregulated MHC-I despite downstream mutations causing constitutive MAPK activation. The NRAS Q61K mutation was introduced into H827 and cells were treated with EGFRi or MEKi as done in 2G. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Close modal

To confirm that the increased HLA expression on the cell-surface had important functional significance for enhanced presentation of antigens, we quantified the cell-surface MHC/peptide epitope density by use of TCR-mimic mAb selective for two well-validated tumor-associated epitopes presented by HLA-A*02:01, a WT1 peptide and a PRAME 300 peptide (24, 25). Consistent with the increased surface HLA-A*02:01 expression, we also observed increased binding of the two TCR-mimic antibodies upon inhibition of MEK and EGFR (Fig. 2B).

We confirmed the regulatory activity of the pathway in a gain-of-function experiment by further stimulating the ERK pathway with EGF. The binding of EGF to the EGFR suppressed surface HLA-A and HLA-A, -B, -C, providing additional confirmation of the importance of the MAPK pathway in regulating surface MHC (Fig. 2C).

The mechanism by which the MAP kinase pathway suppresses HLA-A was unknown. Given that many cancers have activating mutations in specific genes in the MAP kinase pathway, we investigated inhibition of the identified hits in cell lines harboring mutations in EGFR, or downstream in Ras. We used a panel of NSCLC cell lines with activating mutations in EGFR, such as the delE746–A750 in H827, or L858R/T790M mutation in H1975. The delE746–A750 confers sensitivity to erlotinib, whereas the T790M confers resistance to erlotinib and to other first-generation EGFR inhibitors, but is sensitive to afatinib (26). We also used EGFR wild-type NSCLC lines with downstream mutations, such as activating NRAS Q61K in H1299 or KRAS G12S in A549.

Use of the EGFRi erlotinib and afatinib upregulated surface MHC-I if the cell line had the sensitizing mutation, whereas all responded to trametinib MEKi (Fig. 2D and E). The sensitivity to EGFRi erlotinib and afatinib upregulating surface MHC-I was not observed with downstream activating RAS mutations. Expression of the activating EGFR mutation L858R suppressed MHC-I in H1299 NRAS Q61K mutant cell lines (Fig. 2F).

H827 responded more strongly to EGFRi by erlotinib than MEKi by trametinib, despite their similar suppression of pERK, a downstream marker of MEK activity. The combination MEKi and EGFRi was equivalent to EGFRi alone (Fig. 2G). We introduced the NRAS Q61K mutation, shown to cause resistance to EGFRi and persistent activation of the MAPK pathway in H827. Use of the EGFRi still had an effect on surface MHC-I despite no change in pERK output on the H827 NRAS Q61K cell line (Fig. 2H). This could be due to activation of parallel signaling pathways in EGFR-mutant cancers or differential stimulation of ERK. Thus, the MAPK pathway is not the only determinant of EGFR-mediated regulation of surface MHC-I. Given that both EGFR and MEK are involved in signaling via the MAP kinase pathway, these data validate the importance of this pathway in regulating surface HLA-A and MHC-I.

IFNγ is a well-known regulator of MHC-I via the JAK/STAT pathway (27, 28). We asked whether a combination of IFNγ with the kinase inhibitors would have additive effects on HLA expression (Supplementary Fig. S5A–S5C). Both IFNγ and afatinib (in EGFR-mutant H1975 lung cancer cells), and IFNγ and trametinib (in Braf-mutant SK-MEL5 and UACC257 melanoma cells), alone, each increased expression of cell-surface HLA molecules, as measured by antibodies to HLA-A*02:01 or pan–HLA-A, -B, and -C. The combination of the IFNγ and the drug had greater effects than either alone, consistent with the involvement of two different pathways. PCR analysis of TAP1 showed that this internal component of the antigen presentation machinery was also upregulated by IFNγ by 10- to 25-fold in all three cell lines. β2-Microglobulin was also upregulated 4- to 7-fold with IFNγ treatment in all three lines. There were minimal increases in these two proteins in response to the two kinase inhibitors in H1975 and UACC257. However, in SK-MEL5, trametinib increased both proteins alone and was additive with interferon gamma.

HLA-E is another component of the antigen presentation pathway that presents MHC molecule–bound peptides and may be involved in downregulating NK cell immune responses to cancers (29). IFNγ and afatinib (in EGFR-mutant lung cancer cells) did not affect HLA-E levels (Supplementary Fig. S5D). Trametinib, in Braf-mutant SK-MEL5 melanoma cells, increased cell-surface HLA-E molecules, but did not do so in UACC257 cells. IFNγ also variably upregulated HLA-E in the two melanoma lines, and the combination of drugs was more effective in increasing HLA-E in UACC257 cells (Supplementary Fig. S5E and S5F). Although an upregulation of HLA-E might be expected to partially counter the effects of upregulation of classic MHC seen in these cells, the net effect was to improve cytolytic activity.

Improving immunotherapy by inhibiting the MAPK pathway

We next tested the effects of modulating HLA-A*02:01 expression on the efficacy of immunotherapies that depend on HLA-A*02:01 upregulation and antigen presentation, by use of pmel-1 T cells expressing a TCR that reacts with gp100 and use of two different TCR-mimic antibodies whose function also depend on peptide/MHC-I expression. The TCR-mimic antibodies ESKM, which targets a peptide from WT1 in the context of HLA-A*02:01, and Pr20m, which targets a peptide from PRAME, were used as easily quantifiable surrogate tools for measuring the potential therapeutic consequences of upregulation of HLA-A*02:01–based antigen targets as a consequence of MEK inhibition. The cytotoxicity of ESKM against the JMN and Meso34 human mesothelioma cell lines were increased by MEK inhibition with trametinib (Fig. 3A and B), which was used at a noncytotoxic dose (Supplementary Fig. S6). Increased cytotoxicity of the Pr20m mAb was also observed with use of the MEKi trametinib in the SK-MEL-5 human melanoma cell line, validating this observation with multiple targets in multiple cell lines (Fig. 3C).

Figure 3.

Improving cytolysis efficacy by up-regulating cell-surface MHC-I. A, antibody dependent cellular cytotoxicity assay was performed on the JMN human mesothelioma cell line. Cells were incubated for 72 hours with either vehicle control or trametinib and subsequently exposed to either isotype antibody or ESKM in ADCC assay B, ADCC assay on Meso34 (human mesothelioma). Experimental set-up was similar to 3A. C, ADCC assay on SKMEL5 (human melanoma) using TCR-mimic mAb PRAME against the PRAME epitope, experimental setup similar to 3A. D, B16F10 cells were exposed to pmel-1 (gp100)–specific TCR T cells for 24 hours, then killing was assessed using a clonogenic assay described previously. E, B16F10 pERK protein, as measured by pERK intracellular staining, in cells treated with vehicle or 1 μmol/L trametinib. F, B16F10 MHC-I expression assessed by flow cytometry after treatment with 1 μmol/L trametinib for 72 hours. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Figure 3.

Improving cytolysis efficacy by up-regulating cell-surface MHC-I. A, antibody dependent cellular cytotoxicity assay was performed on the JMN human mesothelioma cell line. Cells were incubated for 72 hours with either vehicle control or trametinib and subsequently exposed to either isotype antibody or ESKM in ADCC assay B, ADCC assay on Meso34 (human mesothelioma). Experimental set-up was similar to 3A. C, ADCC assay on SKMEL5 (human melanoma) using TCR-mimic mAb PRAME against the PRAME epitope, experimental setup similar to 3A. D, B16F10 cells were exposed to pmel-1 (gp100)–specific TCR T cells for 24 hours, then killing was assessed using a clonogenic assay described previously. E, B16F10 pERK protein, as measured by pERK intracellular staining, in cells treated with vehicle or 1 μmol/L trametinib. F, B16F10 MHC-I expression assessed by flow cytometry after treatment with 1 μmol/L trametinib for 72 hours. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Close modal

Finally, specific killing by T cells increased after upregulating MHC-I with MEKi. The pmel-1 gp100–specific mouse T cells were more effective at killing of the gp100-positive target B16F10 melanoma cells following trametinib treatment, which correlated with pERK inhibition and MHC-I upregulation (Fig. 3D–F; ref. 30). Therefore, improved recognition as a result of the increased expression of MHC-I and its presented peptides using three different target antigens by TCR or TCR mimics had significant consequences for cytotoxic activity.

Mechanism of MAPK regulation of MHC-I

We hypothesized that the inhibition of the MAP kinase pathway might act on other components of the antigen presentation machinery in addition to MHC-I molecules, thus allowing increased epitope expression in the more abundant cell-surface HLA molecules. Indeed, EGFR and MEK inhibition produced an increase in mRNA gene expression of HLA-A along with other key components of the antigen presentation pathway and MHC-I structure, such as TAP1, TAP2, and β2M (Fig. 4A). The JMN and Meso34 cells at t = 1 hour were sensitive to trametinib at doses less than 10 nmol/L as previously reported, but required higher doses to sustain inhibition of pERK at t = 72 hours due to strong feedback (Supplementary Fig. S7A vs. S7B). Doses of trametinib were chosen over the IC50 of MEK by using pERK as a readout of MEK inhibition at 72 hours (Supplementary Fig. S7B and S7C). These data correlated with previous findings that BRAF-mutant cell lines are the most sensitive to MEK inhibition, when compared with BRAF wild-type cell lines harboring further upstream mutations (31). A time course showed maximal inhibition of MEK at 3 hours, with maximal increases of HLA-A and β2M at 72 hours (Supplementary Fig. S8). Surface HLA-A increased in a dose-dependent manner with increasing MEK inhibition in both melanoma and mesothelioma (Fig. 4B). The phenotypes observed are unlikely from off-target effects of the drug, given the dose response on pERK expression and the plateau of the dose response of surface HLA-A.

Figure 4.

MAPK signaling suppresses antigen presentation machinery and MAPK inhibition broadly upregulates antigen presentation machinery. A, MEK and EGFR inhibition for 48 hours led to increased HLA-A, along with TAP1, TAP2, and β2M in JMN, Meso34, SK-MEL-5 and UACC257, H827, and PC9. B, dose-dependent increase in surface HLA-A with increasing MEKi in JMN and SKMEL5. Cells were analyzed by flow cytometry at 72 hours. C, MEK inhibition leads to increasing amounts of HLA-A and β2M protein. Cells were treated with indicating amounts of trametinib (MEKi) for 72 hours and specific antibodies to the indicated proteins were blotted. D, EGFR inhibition led to increasing HLA-A and β2M protein. Experimental set-up similar to C. E, overexpression of β2M led to increased surface HLA-A and HLA-ABC. F, treatment of JMN with trametinib for 72 hours led to increased activity on the HLA-A and β2M promoters. The HLA-A and β2M promoter was cloned upstream of the Gaussian Luciferase gene. SEAP under the CMV promoter was used as a normalization factor. G, knockdown of STAT1, on JMN cells treated with MEKi demonstrates role in mediating surface HLA-A upregulation. JMN cells were transfected with siRNA against the genes shown and treated with either DMSO or 1 μmol/L trametinib 24 hours after siRNA transfection, then assayed by flow cytometry for surface HLA-A expression 72 hours after treatment. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Figure 4.

MAPK signaling suppresses antigen presentation machinery and MAPK inhibition broadly upregulates antigen presentation machinery. A, MEK and EGFR inhibition for 48 hours led to increased HLA-A, along with TAP1, TAP2, and β2M in JMN, Meso34, SK-MEL-5 and UACC257, H827, and PC9. B, dose-dependent increase in surface HLA-A with increasing MEKi in JMN and SKMEL5. Cells were analyzed by flow cytometry at 72 hours. C, MEK inhibition leads to increasing amounts of HLA-A and β2M protein. Cells were treated with indicating amounts of trametinib (MEKi) for 72 hours and specific antibodies to the indicated proteins were blotted. D, EGFR inhibition led to increasing HLA-A and β2M protein. Experimental set-up similar to C. E, overexpression of β2M led to increased surface HLA-A and HLA-ABC. F, treatment of JMN with trametinib for 72 hours led to increased activity on the HLA-A and β2M promoters. The HLA-A and β2M promoter was cloned upstream of the Gaussian Luciferase gene. SEAP under the CMV promoter was used as a normalization factor. G, knockdown of STAT1, on JMN cells treated with MEKi demonstrates role in mediating surface HLA-A upregulation. JMN cells were transfected with siRNA against the genes shown and treated with either DMSO or 1 μmol/L trametinib 24 hours after siRNA transfection, then assayed by flow cytometry for surface HLA-A expression 72 hours after treatment. Experiments were performed 2 to 4 times with similar results, and data shown are representative.

Close modal

Antibodies against pERK, along with total ERK1/2, were used to show dose-responsive increases in response to trametinib inversely correlated with HLA-A protein expression. The increase of β2M much greater than that of HLA complexes in multiple cell lines, consistent with the gene expression data (Fig. 4C). EGFRi with erlotinib also caused a dose-dependent increase in HLA-A and β2M (Fig. 4D). Because β2M is required for surface presentation of HLA-A, -B, and -C and stability of the MHC-I molecules on the cell surface, we investigated the potential role of β2M in controlling cell-surface HLA-A expression. Overexpression of β2M increased cell-surface HLA-A and pan–HLA-ABC, phenocopying the effect of MEK inhibition (Fig. 4E), which was regulated by multiple regulatory domains in the promoter region, including the ISRE site, E box, and NF-κB sites. Using a luciferase-based promoter assay, we demonstrated that upon addition of MEKi, a dose-dependent increase in activity on the HLA-A and β2M promoters was observed (Fig. 4F). Knockdowns of STAT1, STAT3, and RelA (a component of the NF-κB complex) were performed on JMN cells, along with treatment with MEKi. STAT1 knockdown had the largest inhibition of upregulation of surface HLA-A after MEKi, suggesting a role for STAT1 in responses to MEKi (Fig. 4G).

MAPK activation causes in vivo suppression of MHC-I and increased PD-1/L1

We confirmed that these observations on MHC regulation and antigen presentation machinery were not limited to in vitro models. Microarray profiling of the lung bearing tumors from transgenic EGFR L858R, which activates the MAPK pathway, compared with normal lungs, demonstrated suppression of mouse MHC-I and antigen presentation components H2-K/D and β2M, thereby confirming the effects of this pathway in vivo (Fig. 5A; ref. 19). Upregulation of PD-1 and PD-L1 markers in the tumors was also observed as previously published (20).

Figure 5.

Activation of the MAPK pathway via activating EGFR mutations causes in vivo suppression of MHC-I in addition to upregulation of checkpoint blockade. A, unsupervised hierarchical clustering microarray expression profiling analysis of lung tumors from CC10/L858R mice with EGFR L858R tumor–bearing lungs (right, black) or normal lungs (left, green) focusing on H2-KD, β2M, TAP1, TAP2, PD-L1(PDCD1), and PD-1 (CD274) gene expression. B, flow cytometry data of FVB CC10-rtTA/TetO EGFR L858R-expressing mice. Mice were induced with doxycycline for >6 weeks before sacrificed (mice E–G). Control mice were kept on normal diet, but genotypically identical (A–D). Lungs were isolated and stained with markers for CD45 (pan-leukocyte), hEGFR, and H2-Kq (MHC-I). C, the CD45 lung population was stained with mouse H2-kq–specific mAb. CD45hEGR population shows higher MHC-I expression than the CD45hEGFR+ population. Representative MRI images of mouse lungs are shown for two samples.

Figure 5.

Activation of the MAPK pathway via activating EGFR mutations causes in vivo suppression of MHC-I in addition to upregulation of checkpoint blockade. A, unsupervised hierarchical clustering microarray expression profiling analysis of lung tumors from CC10/L858R mice with EGFR L858R tumor–bearing lungs (right, black) or normal lungs (left, green) focusing on H2-KD, β2M, TAP1, TAP2, PD-L1(PDCD1), and PD-1 (CD274) gene expression. B, flow cytometry data of FVB CC10-rtTA/TetO EGFR L858R-expressing mice. Mice were induced with doxycycline for >6 weeks before sacrificed (mice E–G). Control mice were kept on normal diet, but genotypically identical (A–D). Lungs were isolated and stained with markers for CD45 (pan-leukocyte), hEGFR, and H2-Kq (MHC-I). C, the CD45 lung population was stained with mouse H2-kq–specific mAb. CD45hEGR population shows higher MHC-I expression than the CD45hEGFR+ population. Representative MRI images of mouse lungs are shown for two samples.

Close modal

Expression of EGFR L858R in the transgenic mice, given doxycycline for >6 weeks, was demonstrated by increased binding of a human EGFR-specific fluorescently labeled mAb (Fig. 5B). Mice were confirmed to have development of lung adenocarcinoma by MR, with development of a reticulonodular infiltrate in the lung, consistent with previous publications (18). The CD45hEGFR+ population in the lung in the EGFR L858R–expressing mice demonstrated decreased binding of a MHC-I–specific mAb by flow cytometry, when compared with a wild-type mouse that did not express the EGFR L858R mutation (Fig. 5C).

Immunotherapy of cancer is emerging as a successful and important component of treatment. MHC molecules presenting antigens are the target of multiple therapeutic strategies that involve vaccines, T cells, or TCRs, TCR-mimic antibodies, or T-cell checkpoint blockade. The last, a highly effective recent example in cancer therapy, appears to require presentation of neoantigens on MHC-I on the surface of cancer cells (32–34). Most immunotherapies have focused on enhancing intrinsic effector cell mechanisms for modulating the immune response, either by directly activating the effector T cells or by relieving their suppression. In distinct contrast, here we propose an alternative approach, whereby the antigenic targets on the cancer cells themselves are modulated to improve TCR-based killing. The ability to regulate such responses by selectively affecting target cells could have an important impact on both disease and therapy. We propose that kinases are a readily druggable pathway that might be used in conjunction with immunotherapy to enhance efficacy. The beneficial effect of the combination of immunotherapy with kinase inhibition was shown in mouse models of combined PD-1/PD-L1 blockade with MEKi (35). A second model, of adoptive T-cell therapy in combination with MEKi in BRAF-mutant murine melanoma, has demonstrated superiority to single agents alone (36). Our work has provided a new understanding of another mechanism why these combination therapies may be more effective, wherein upregulated MHC-I and antigen presentation on the target cells, essential for the adaptive immune response, improves TCR-based recognition and killing. Indeed, many of the patients treated currently with immunotherapies also receive kinase inhibitor therapies as distinct treatments.

The loss- and gain-of-function screen described here allowed unbiased interrogation of the currently annotated human kinases for their regulation of cell-surface MHC-I. We then explored mechanistically how such kinase regulators could be inhibited for altering surface expression of MHC-I, as a way of validating the screen, understanding the process, and also for extending the findings to functional modulation of a model immunotherapy proof of concept that directly depends on MHC-I presentation. In this case, we were able to specifically isolate the MHC as the sole target of the therapy by use of therapeutic TCR-mimic antibodies directed to antigens presented by MHC.

This study also provides support for the use of a flow cytometry–based loss-of-function pooled shRNA screen in the study of the regulation of other cell-surface molecules, and potentially intracellular antigens as well. This technique will allow many laboratories without robotics and high-throughput flow cytometry equipment to investigate pathways that can be easily perturbed with loss-of-function RNAi screens or other techniques, such as CRISPR loss of function.

Here, we have demonstrated the effect of the MAPK pathway on MHC-I, in vitro and in vivo. However, therapeutic applications in humans of the combination of available immunotherapies and pharmacologic kinase inhibition, to increase MHC-I surface expression and antigen presentation, will be complicated and difficult to predict, because T cells and NK cells also rely on similar kinase signaling pathways for activation. More work needs to be done to determine optimal pathways or schedules or doses to target MHC in tumor cells specifically, while sparing signaling pathways of the effector cells (37). Some studies suggest conflicting effects of MEKi on T-cell effector function, which may be dependent on the tumor model evaluated (38, 39). These effects cannot be simply modeled in mice. Empirically derived optimal dosing and schedules will be needed in in vivo models and in humans to show that use of kinase inhibition to regulate immunotherapy has therapeutic benefits, while sparing immune effector cells of the detrimental effects (39–41). These investigations will be complicated by the effects of some of the drugs on the cellular effectors themselves, the variable effects on the cancer cells depending on their specific mutations, the time frames required to upregulate the responses (about 3 days in the experiments here) and the time required for the effects to wash out of the cancer cells and the effectors.

The data provide a mechanistic explanation of how MHC-I is regulated by the MAPK pathway. MHC-I mRNA expression is regulated through upstream enhancer elements, with involvement of the NF-κB transcription factor (42, 43). MHC-I is also induced by TNF, IL1, IFNβ, and IFNγ, which upregulates HLA-A via the JAK/STAT pathway (27, 28). The CIITA transcription factor can also act on MHC-I gene expression (44). IFNγ can increase MHC-I and antigen presentation, but thus far its use has had limited applications (45). We show here that combining the kinase inhibitors with IFNγ in vitro can have additive effects on HLA expression, TAP1, and β2M. This may be of benefit in vivo as IFNγ may be elaborated locally at tumor sites from tumor-infiltrating lymphocytes at steady state or in response to other immunotherapies, such as checkpoint blockade.

MEK has been proposed by others to be a regulator of MHC-I expression. EGFR inhibition can augment MHC-I and MHC-II expression in keratinocytes (46). MEK was previously identified as a negative regulator of HLA-A*02:01 in esophageal and gastric cancer by Mimura and colleagues (47). We validated these targets in the screen as important negative regulators of MHC-I and discovered a mechanistic role of the MAP kinase pathway in regulating surface levels of MHC-I. Our data directly link to immune-oncologic applications in humans, by demonstrating potent upregulation of MHC-I in a wide variety of cancers, including melanoma and NSCLC, which are currently the subject of FDA-approved therapies which depend upon on antigen presentation, such as checkpoint blockade with mAbs ipilibumab, pembrolizumab, and nivolibumab. In addition, by demonstrating that an FDA-approved MEKi upregulated MHC-I, the results support the clinical testing of combination therapies, which could advance this concept into human therapy. We characterized these effects on MHC-I in cells with activating mutations in the MAP kinase pathway with various genotypic lesions, such as activating EGFR mutations, BRAF mutations, and RAS mutations. Finally, we also showed the mechanism was active in RET-translocated thyroid cancer. The sum of these data supports the importance of MAPK pathway in regulating MHC-I quite broadly, while providing new mechanistic insights.

Finally, our findings are immunologically significant. Upregulation induced by MEK inhibition resulted in superior cytotoxic activity of TCR-mimic antibodies (directed to specific MHC-presented antigens) and TCR-based therapy with a pmel-specific murine T-cell model. We further show, in a transgenically engineered mouse model in vivo, that activating this pathway reduces expression of the components of the antigen presentation machinery, along with MHC-I. This gain-of-function experiment is crucial to proving that activation of MAPK can cause decreased MHC-I in vivo.

Activating EGFR mutations may contribute to immune escape, due to PD-L1 expression (20). Downregulation of MHC-I, which was observed from our study, may also contribute to this finding. While demonstrating combination therapy with EGFRi and checkpoint blockade would be rational, the transgenic EGFR mice have shown dramatic tumor reduction and cures with EGFRi as monotherapy, leaving little window to show synergism in currently existing mouse models with checkpoint blockade (18). Our data also suggest that using combination therapy with MAP kinase inhibition can be powerful, not only as a direct cancer therapy to prevent growth but also indirectly to promote immunotherapy.

HLA genes are a risk factor for autoimmune diseases such as ankolysing spondylitis and multiple sclerosis (48–50). In addition to upregulation by certain kinases, we showed downregulation of MHC-I through new kinase targets. These targets are not currently addressed by immunosuppressive therapies, which inhibit the effector arm of the immune response with concomitant toxicity. These new targets warrant additional investigation into altering the course of autoimmune diseases by investigating the efficacy of specific kinase inhibitors and developing appropriate mouse models.

A requirement of many immunotherapies therapies, particularly checkpoint blockade, is the availability of recognizable antigens that are presented on MHC-I. Tumors can decrease MHC-I expression, to avoid immune system detection of the rare neoantigens created in tumors by mutations, and increase inhibitory receptor expression. By modulating expression of these limited antigens, improved clinical efficacy may be seen with certain immunotherapies in conjunction with current FDA-approved small molecules targeting EGFR and MEK. The inhibition of kinase pathways also caused a more general upregulation of the antigen presentation machinery, including Tap (responsible for transporting peptides) and β2M (responsible for stabilizing MHC-I). Many of the recently approved immunotherapies, such as blockade of CTLA-4 or PD-1, release the T-cell inhibition promoted by target tumor cells. These immunotherapies provide a promising approach to addressing multiple malignancies. By rationally combining them with targeted small-molecule inhibitors, novel synergistic treatment strategies may be developed.

D.A. Scheinberg has ownership interest in patent applications and is a consultant/advisory board member for Eureka. No potential conflicts of interest were disclosed by the other authors.

Conception and design: E.J. Brea, C.Y. Oh, E. Manchado, T. Merghoub, J.D. Wolchok, S.W. Lowe, D.A. Scheinberg

Development of methodology: E.J. Brea, C.Y. Oh, E. Manchado, R.J. Garippa, T. Merghoub, S.W. Lowe, D.A. Scheinberg

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.J. Brea, C.Y. Oh, E. Manchado, S. Budhu, G. Mo, P. Mondello, J.E. Han, C.A. Jarvis, Q. Xiang, A.Y. Chang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.J. Brea, C.Y. Oh, S. Budhu, R.S. Gejman, G. Mo, J.E. Han, D. Ulmert, R.J. Garippa, A.Y. Chang, J.D. Wolchok, N. Rosen, D.A. Scheinberg

Writing, review, and/or revision of the manuscript: E.J. Brea, C.Y. Oh, E. Manchado, J.E. Han, J.D. Wolchok, N. Rosen, S.W. Lowe, D.A. Scheinberg

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C.Y. Oh, P. Mondello, J.E. Han, D. Ulmert, D.A. Scheinberg

Study supervision: E.J. Brea, D.A. Scheinberg

Other (RNA interference studies): R.J. Garippa

We thank T. Dao, L. Dubrovsky, D. Pankov, P Lito, D. Solit, E. Pamer, R. Brentjens, M. Will, A. Lujambia, and A. Scott for their helpful discussions. We also thank Y. Li and A. Younes for use of their equipment.

The study was supported by NIH grant R01 CA 55349 (D.A. Scheinberg), P01 CA23766 (D.A. Scheinberg), Diversity Research Supplement for the P01CA023766 (E.J. Brea and D.A. Scheinberg), MARF (D.A. Scheinberg), P30 CA008748, NCI Grant NIHT32CA062948 (C.Y. Oh), NIGMS T32GM07739 (E.J. Brea and R.S. Gejman), MSKCC's Experimental Therapeutics Center,and the Lymphoma Foundation and Tudor and Glades funds. NIH F30 CA200327 (R.S. Gejman).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Agrawal
S
,
Kishore
MC
. 
MHC class I gene expression and regulation
.
J Hematother Stem Cell Res
2000
;
9
:
795
812
.
2.
Morgan
RA
,
Dudley
ME
,
Wunderlich
JR
,
Hughes
MS
,
Yang
JC
,
Sherry
RM
, et al
Cancer regression in patients after transfer of genetically engineered lymphocytes
.
Science
2006
;
314
:
126
9
.
3.
Dao
T
,
Yan
S
,
Veomett
N
,
Pankov
D
,
Zhou
L
,
Korontsvit
T
, et al
Targeting the intracellular WT1 oncogene product with a therapeutic human antibody
.
Sci Transl Med
2013
;
5
:
176ra33
.
4.
Yee
C
,
Thompson
JA
,
Byrd
D
,
Riddell
SR
,
Roche
P
,
Celis
E
, et al
Adoptive T cell therapy using antigen-specific CD8+ T cell clones for the treatment of patients with metastatic melanoma: in vivo persistence, migration, and antitumor effect of transferred T cells
.
Proc Natl Acad Sci U S A
2002
;
99
:
16168
73
.
5.
Chapuis
AG
,
Ragnarsson
GB
,
Nguyen
HN
,
Chaney
CN
,
Pufnock
JS
,
Schmitt
TM
, et al
Transferred WT1-reactive CD8+ T cells can mediate antileukemic activity and persist in post-transplant patients
.
Sci Transl Med
2013
;
5
:
174ra27
.
6.
Pardoll
DM
. 
The blockade of immune checkpoints in cancer immunotherapy
.
Nat Rev Cancer
2012
;
12
:
252
64
.
7.
Mellman
I
,
Coukos
G
,
Dranoff
G
. 
Cancer immunotherapy comes of age
.
Nature
2011
;
480
:
480
9
.
8.
Adams
JL
,
Smothers
J
,
Srinivasan
R
,
Hoos
A
. 
Big opportunities for small molecules in immuno-oncology
.
Nat Rev Drug Discov
2015
;
14
:
603
22
.
9.
Chiappinelli
KB
,
Zahnow
CA
,
Ahuja
N
,
Baylin
SB
. 
Combining epigenetic and immunotherapy to combat cancer
.
Cancer Res
2016
;
1683
90
.
10.
Campoli
M
,
Ferrone
S
. 
HLA antigen changes in malignant cells: epigenetic mechanisms and biologic significance
.
Oncogene
2008
;
27
:
5869
85
.
11.
Leone
P
,
Shin
E-C
,
Perosa
F
,
Vacca
A
,
Dammacco
F
,
Racanelli
V
. 
MHC class I antigen processing and presenting machinery: organization, function, and defects in tumor cells
.
J Natl Cancer Inst
2013
;
105
:
1172
87
.
12.
Kageshita
T
,
Hirai
S
,
Ono
T
,
Hicklin
DJ
,
Ferrone
S
. 
Down-regulation of HLA class I antigen-processing molecules in malignant melanoma
.
Am J Pathol
1999
;
154
:
745
54
.
13.
Paul
P
,
Van Den Hoorn
T
,
Jongsma
MLM
,
Bakker
MJ
,
Hengeveld
R
,
Janssen
L
, et al
A genome-wide multidimensional RNAi screen reveals pathways controlling MHC class II antigen presentation
.
Cell
2011
;
145
:
268
83
.
14.
Calabrò
L
,
Morra
A
,
Fonsatti
E
,
Cutaia
O
,
Amato
G
,
Giannarelli
D
, et al
Tremelimumab for patients with chemotherapy-resistant advanced malignant mesothelioma: an open-label, single-arm, phase 2 trial
.
Lancet Oncol
2013
;
14
:
1104
11
.
15.
Veomett
N
,
Dao
T
,
Liu
H
,
Xiang
J
,
Pankov
D
,
Dubrovsky
L
, et al
Therapeutic efficacy of an Fc-enhanced TCR-like antibody to the intracellular WT1 oncoprotein
.
Clin Cancer Res
2014
;
20
:
4036
46
.
16.
Zuber
J
,
McJunkin
K
,
Fellmann
C
,
Dow
LE
,
Taylor
MJ
,
Hannon
GJ
, et al
Toolkit for evaluating genes required for proliferation and survival using tetracycline-regulated RNAi
.
Nat Biotechnol
2011
;
29
:
79
83
.
17.
Fellmann
C
,
Hoffmann
T
,
Sridhar
V
,
Hopfgartner
B
,
Muhar
M
,
Roth
M
, et al
An optimized microRNA backbone for effective single-copy RNAi
.
Cell Rep
2013
;
5
:
1704
13
.
18.
Politi
K
,
Zakowski
MF
,
Fan
P-D
,
Schonfeld
EA
,
Pao
W
,
Varmus
HE
. 
Lung adenocarcinomas induced in mice by mutant EGF receptors found in human lung cancers respond to a tyrosine kinase inhibitor or to down-regulation of the receptors
.
Genes Dev
2006
;
20
:
1496
510
.
19.
Regales
L
,
Gong
Y
,
Shen
R
,
de Stanchina
E
,
Vivanco
I
,
Goel
A
, et al
Dual targeting of EGFR can overcome a major drug resistance mutation in mouse models of EGFR mutant lung cancer
.
J Clin Invest
2009
;
119
:
3000
10
.
20.
Akbay
EA
,
Koyama
S
,
Carretero
J
,
Altabef
A
,
Tchaicha
JH
,
Christensen
CL
, et al
Activation of the PD-1 pathway contributes to immune escape in EGFR-driven lung tumors
.
Cancer Discov
2013
;
3
:
1355
63
.
21.
Rivoltini
L
,
Barracchini
KC
,
Viggiano
V
,
Kawakami
Y
,
Smith
A
,
Mixon
A
, et al
Quantitative correlation between HLA class I allele expression and recognition of melanoma cells by antigen-specific cytotoxic T lymphocytes
.
Cancer Res
1995
;
55
:
3149
57
.
22.
Flaherty
KT
,
Robert
C
,
Hersey
P
,
Nathan
P
,
Garbe
C
,
Milhem
M
, et al
Improved survival with MEK inhibition in BRAF-mutated melanoma
.
N Engl J Med
2012
;
367
:
107
14
.
23.
Brevet
M
,
Shimizu
S
,
Bott
MJ
,
Shukla
N
,
Zhou
Q
,
Olshen
AB
, et al
Coactivation of receptor tyrosine kinases in malignant mesothelioma as a rationale for combination targeted therapy
.
J Thorac Oncol
2011
;
6
:
864
74
.
24.
Chang
A
,
Dao
T
,
Scott
A
,
Dubrovsky
L
,
Liu
C
,
Scheinberg
D
. 
A therapeutic TCR mimic monoclonal antibody for intracellular PRAME protein in leukemias
.
Am Soc Hematol
2015
;
126
:
2527
.
25.
Krug
LM
,
Dao
T
,
Brown
AB
,
Maslak
P
,
Travis
W
,
Bekele
S
, et al
WT1 peptide vaccinations induce CD4 and CD8 T cell immune responses in patients with mesothelioma and non-small cell lung cancer
.
Cancer Immunol Immunother
2010
;
59
:
1467
79
.
26.
Cross
DAE
,
Ashton
SE
,
Ghiorghiu
S
,
Eberlein
C
,
Nebhan
CA
,
Spitzler
PJ
, et al
AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated resistance to EGFR inhibitors in lung cancer
.
Cancer Discov
2014
;
4
:
1046
61
.
27.
Girdlestone
J
,
Isamat
M
,
Gewert
D
,
Milstein
C
. 
Transcriptional regulation of HLA-A and -B: differential binding of members of the Rel and IRF families of transcription factors
.
Proc Natl Acad Sci U S A
1993
;
90
:
11568
72
.
28.
Wolchok
JD
,
Vilček
J
. 
Induction of HLA class I mRNA by cytokines in human fibroblasts: comparison of TNF, IL-1 and IFN-β
;.
Cytokine
1992
;
4
:
520
7
.
29.
Kochan
G
,
Escors
D
,
Breckpot
K
,
Guerrero-Setas
D
. 
Role of non-classical MHC class I molecules in cancer immunosuppression
.
Oncoimmunology
2013
;
2
:
e26491
.
30.
Budhu
S
,
Loike
JD
,
Pandolfi
A
,
Han
S
,
Catalano
G
,
Constantinescu
A
, et al
CD8+ T cell concentration determines their efficiency in killing cognate antigen-expressing syngeneic mammalian cells in vitro and in mouse tissues
.
J Exp Med
2010
;
207
:
223
35
.
31.
Solit
DB
,
Garraway
LA
,
Pratilas
CA
,
Sawai
A
,
Getz
G
,
Basso
A
, et al
BRAF mutation predicts sensitivity to MEK inhibition
.
Nature
2006
;
439
:
358
62
.
32.
Rizvi
NA
,
Hellmann
MD
,
Snyder
A
,
Kvistborg
P
,
Makarov
V
,
Havel
JJ
, et al
Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer
.
Science
2015
;
348
:
124
8
.
33.
Snyder
A
,
Makarov
V
,
Merghoub
T
,
Yuan
J
,
Zaretsky
JM
,
Desrichard
A
, et al
Genetic basis for clinical response to CTLA-4 blockade in melanoma
.
N Engl J Med
2014
;
371
:
2189
99
.
34.
Van Allen
EM
,
Miao
D
,
Schilling
B
,
Shukla
SA
,
Blank
C
,
Zimmer
L
, et al
Genomic correlates of response to CTLA4 blockade in metastatic melanoma
.
Science
2015
;
350
:
207
11
.
35.
Liu
L
,
Mayes
PA
,
Eastman
S
,
Shi
H
,
Yadavilli
S
,
Zhang
T
, et al
The BRAF and MEK inhibitors dabrafenib and trametinib: effects on immune function and in combination with immunomodulatory
antibodies targeting PD1, PD-L1 and CTLA-4
.
Clin Cancer Res
2015
;
21
:
1639
51
.
36.
Hu-Lieskovan
S
,
Mok
S
,
Homet Moreno
B
,
Tsoi
J
,
Robert
L
,
Goedert
L
, et al
Improved antitumor activity of immunotherapy with BRAF and MEK inhibitors in BRAF(V600E) melanoma
.
Sci Transl Med
2015
;
7
:
279ra41
.
37.
Kortum
RL
,
Rouquette-Jazdanian
AK
,
Samelson
LE
. 
Ras and extracellular signal-regulated kinase signaling in thymocytes and T cells
.
Trends Immunol
2013
;
34
:
259
68
.
38.
Ebert
PJR
,
Cheung
J
,
Yang
Y
,
Kim
JM
,
Belvin
M
,
Mellman
I
, et al
MAP kinase inhibition promotes T cell and anti- tumor activity in combination with PD-L1 MAP kinase inhibition promotes T cell and anti-tumor activity in combination with PD-L1 checkpoint blockade
.
Immunity
2016
;
44
:
609
21
.
39.
Vella
LJ
,
Pasam
A
,
Dimopoulos
N
,
Andrews
M
,
Knights
A
,
Puaux
A-L
, et al
MEK inhibition, alone or in combination with BRAF inhibition, affects multiple functions of isolated normal human lymphocytes and dendritic cells
.
Cancer Immunol Res
2014
;
2
:
351
60
.
40.
Callahan
MK
,
Masters
G
,
Pratilas
CA
,
Ariyan
C
,
Katz
J
,
Kitano
S
, et al
Paradoxical activation of T cells via augmented ERK signaling mediated by a RAF inhibitor
.
Cancer Immunol Res
2014
;
2
:
70
9
.
41.
Vella
LJ
,
Andrews
MC
,
Pasam
A
,
Woods
K
,
Behren
A
,
Cebon
JS
. 
The kinase inhibitors dabrafenib and trametinib affect isolated immune cell populations
.
Oncoimmunology
2014
;
3
:
e946367
.
42.
Gobin
SJP
,
Biesta
P
,
Van den Elsen
PJ
. 
Regulation of human β2-microglobulin transactivation in hematopoietic cells
.
Blood
2003
;
101
:
3058
64
.
43.
Wolchok
D
,
Goodman
A
. 
induction cells of NF-xB may be necessary but is not sufficient for of H-2 antigens by TNF in J558L murine myeloma
.
Cytokines
1994
;
55
:
7
12
.
44.
Gobin
SJP
,
Peijnenburg
A
,
Van Eggermond
M
,
Van Zutphen
M
,
Van Den Berg
R
,
Van Den Elsen
PJ
. 
The RFX complex is crucial for the constitutive and CIITA-mediated transactivation of MHC class I and B2-microglobulin genes
.
Immunity
1998
;
9
:
531
41
.
45.
Zaidi
MR
,
Merlino
G
. 
The two faces of interferon-γ in cancer
.
Clin Cancer Res
2011
;
17
:
6118
24
.
46.
Pollack
BP
,
Sapkota
B
,
Cartee
TV
. 
Epidermal growth factor receptor inhibition augments the expression of MHC class I and II genes
.
Clin Cancer Res
2011
;
17
:
4400
13
.
47.
Mimura
K
,
Shiraishi
K
,
Mueller
A
,
Izawa
S
,
Kua
L-F
,
So
J
, et al
The MAPK pathway is a predominant regulator of HLA-A expression in esophageal and gastric cancer
.
J Immunol
2013
;
191
:
6261
72
.
48.
Fogdell-Hahn
A
,
Ligers
A
,
Gronning
M
,
Hillert
J
,
Olerup
O
. 
Multiple sclerosis: a modifying influence of HLA class I genes in an HLA class II associated autoimmune disease
.
Tissue Antigens
2000
;
55
:
140
8
.
49.
Brown
MA
,
Kenna
T
,
Wordsworth
BP
. 
Genetics of ankylosing spondylitis-insights into pathogenesis
.
Nat Rev Rheumatol
2016
;
12
:
81
91
.
50.
Robert
C
,
Kupper
TS
. 
Inflammatory skin diseases, T cells, and immune surveillance
.
N Engl J Med
1999
;
341
:
1817
28
.