Elevated immunity to cancer-expressed antigens can be detected in people with no history of cancer and may contribute to cancer prevention. We have previously reported that MHC-restricted phosphopeptides are cancer-expressed antigens and targets of immune recognition. However, the extent to which this immunity reflects prior or ongoing phosphopeptide exposures was not investigated. In this study, we found that preexisting immune memory to cancer-expressed phosphopeptides was evident in most healthy donors, but the breadth among donors was highly variable. Although three phosphopeptides were recognized by most donors, suggesting exposures to common microbial/infectious agents, most of the 205 tested phosphopeptides were not recognized by peripheral blood mononuclear cells (PBMC) from any donor and the remainder were recognized by only 1 to 3 donors. In longitudinal analyses of 2 donors, effector immune response profiles suggested active reexposures to a subset of phosphopeptides. These findings suggest that the immunogens generating most phosphopeptide-specific immune memory are rare infectious agents or incipient cancer cells with distinct phosphoproteome dysregulations, and that repetitive immunogenic exposures occur in individual donors. Phosphopeptide-specific immunity in PBMCs and tumor-infiltrating lymphocytes from ovarian cancer patients was limited, regardless of whether the phosphopeptide was expressed on the tumor. However, 4 of 10 patients responded to 1 to 2 immunodominant phosphopeptides, and 1 showed an elevated effector response to a tumor-expressed phosphopeptide. As the tumors from these patients displayed many phosphopeptides, these data are consistent with lack of prior exposure or impaired ability to respond to some phosphopeptides and suggest that enhancing phosphopeptide-specific T-cell responses could be a useful approach to improve tumor immunotherapy.

Dysregulated signaling is a hallmark of cancer cells with overactivated kinases or inhibited phosphatases driving many of the aberrant pathways that contribute to malignant transformation (1, 2). This dysregulated signaling results in overexpressed phosphorylated proteins or phosphorylation at noncanonical sites. Phosphorylated proteins undergo proteasomal degradation, and some of the resulting phosphorylated peptides can be presented by MHC molecules on cancer and Epstein–Barr virus (EBV)–transformed cells (3–13). MHC-presented phosphopeptides are recognized by T cells in a phosphate- and sequence-dependent manner (3, 4, 6, 7, 9, 12, 14–17). Therefore, MHC-presented phosphopeptides are a distinct class of tumor neoantigens.

Phosphopeptide-specific T cells can recognize cancer cell lines (6, 7, 9, 13), and many tumor types express phosphopeptides (3, 5–7, 9–11, 13); thus, phosphopeptide-specific T-cell responses may play a role in controlling tumor growth and progression. Phosphopeptide-specific cytotoxic T cells can be generated in vivo by immunization (3, 7–9, 14) or in vitro by repeated peptide stimulation (12). Adoptive transfer of phosphopeptide-specific T cells slows tumor growth in a humanized model of melanoma (7, 8), and phosphopeptide immunization delays tumor growth in a humanized model of lung cancer (9). We have shown that a vaccine designed to induce phosphopeptide-specific immune responses is safe and immunogenic in melanoma patients (8), establishing the potential therapeutic application of phosphopeptide neoantigens. Altogether, these data support further investigation into the role of phosphopeptide-specific T cells in tumor control.

There is increasing evidence that the immune system in otherwise healthy individuals performs surveillance and destruction of incipient cancer cells, developing “preexisting immune memory” to cancer antigens in the absence of clinically evident tumors. Tumor antigen–specific antibodies and T cells have been found in liver, colon, and pancreatic premalignant lesions prior to progression to clinically detectable cancer (18–20). CD4+ and CD8+ T-cell responses to myeloma-associated cancer testis antigens are evident in patients with monoclonal gammopathy of undetermined significance, a condition that can progress to multiple myeloma (21). Evidence of immunosurveillance in healthy individuals with no known premalignant lesions includes effector T cells and antibodies specific for cyclin B1, which is overexpressed in lung, colorectal, cervical, and head and neck cancers (22), and specific for hypoglycosylated MUC1, which is overexpressed in most human adenocarcinomas (23). The immunogenic exposures that generate preexisting immune memory to tumor antigens are unknown, but are thought to be incipient cancer cells or viral infections (24–27). Regardless of how it is generated, preexisting immune memory against tumor antigens may protect against future development of cancer.

We previously found that healthy donors with no evident prior cancer have robust CD8+ T-cell responses to two HLA-A2–restricted melanoma-expressed phosphopeptides and to a cohort of HLA-B7–restricted leukemia-expressed phosphopeptides (6, 7). Some of these responses arose from CD8+CD45RO+ antigen-experienced T cells, providing evidence of phosphopeptide-specific memory T cells in healthy individuals in the absence of clinically evident malignancy (6, 7). Here we investigated the pervasiveness of preexisting immune memory to more than 200 phosphopeptides in 15 healthy donors. We also identified immunodominant phosphopeptides that were recognized by the majority of donors. By characterizing memory and effector subsets, we tested the hypothesis that some immunogenic exposures to phosphopeptides were current or reoccurred over time. Finally, we evaluated memory and effector activity in ovarian cancer patients to test the hypothesis that their phosphopeptide-specific responses would be limited compared with healthy donors and that this limitation would be confined to the phosphopeptides expressed on the tumors.

Cell lines

A transfectant of the antigen-processing mutant cell line, CEMx721.174.T2 (T2 cell line; ref. 28), expressing HLA-B7 (T2-B7 cell line) was kindly received from Dr. Charles Lutz (University of Kentucky) in 1991. Low-passage T2-B7 cells were grown at 37°C with 5% CO2 in RPMI-1640 (Corning; cat#10040CV), 2 mmol/L L-glutamine (Thermo Fisher; cat #25030164), 1% penicillin–streptomycin (Thermo Fisher; cat #10378016), and 10% fetal bovine serum (FBS; Sigma; cat #F6178). 150 μg/mL Hygromycin (Thermo Fisher; cat #10687010) was used as a selective reagent for surface expression of HLA-B7. Cells were regularly assessed by flow cytometry for expression of HLA-A2 (using an APC-conjugated antibody specific for HLA-A2, clone BB7.2, BD; cat #BDB561341) and HLA-B7 (using either a PE-conjugated antibody specific for HLA-B7, clone BB7.1, Novus Biologicals; cat #NB10064159PE, or a PE-conjugated antibody specific for HLA-B7, clone REA176, Miltenyi; cat #130-118-471). Cells were confirmed as Mycoplasma negative by PCR (Universal Mycoplasma Detection Kit, ATCC; cat #30-1012K) before freeze, after thaw, and intermittently while in culture.

Peptides

MHC-presented phosphopeptides previously identified by mass spectrometry on cancer cell lines and/or primary tumor cells (Supplementary Table S1) were purchased from GenScript. Lyophilized peptides were stored at −80°C until reconstituted to 10 mg/mL in DMSO and stored in a desiccator at room temperature. Stock aliquots of individual peptides or combinations of peptides were diluted in Dulbecco's phosphate buffer solution (DPBS; Corning; cat #21-030-CVR) to 200 μg/mL and stored at −80°C.

Cytokines

Human IL2 (cat #202-IL), GM-CSF (cat #215-GM), IL4 (Cat#204-IL), IL6 (cat #206-IL), and TNFα (cat #210-TA) were from R&D. IL1β (cat #AF-200–-01B), IL7 (cat #200–07), and IL15 (cat #200-15) were from PeproTech. Prostaglandin E2 (cat #0409) was from Sigma.

Human blood and tissues

Healthy donor and patient-derived specimens were collected after informed written consent following protocols (#3467 and #10598) approved by the University of Virginia Institutional Review Board for Health Sciences Research. All human studies were performed in accordance with the Declaration of Helsinki.

Healthy donors

Blood was collected from 15 healthy donor volunteers. Characteristics of these individuals are shown in Supplementary Table S2. Inclusion criteria for the study were: ability and willingness to give informed consent, age 18 years and older, no current or past cancer diagnosis, and healthy/free from known illness at all times of blood collection. Exclusion criteria for the study were: unable or unwilling to consent to any of the study criteria and current or prior diagnosis of cancer. For continuation in the study, a donor had to be determined as HLA-A02:01+ and/or HLA-B07:02+ by genotyping and flow cytometry. Age was recorded at the time of first blood collection for this study.

Patients

Blood and tumor specimens were collected from 48 patients at the University of Virginia from August 2017 through December 2019. Tumor specimens and immune responses were analyzed in 10 patients who met all criteria for continuation in the study. Characteristics of the patients are shown in Supplementary Table S3. Inclusion criteria were: ability and willingness to give informed consent to data collection, blood collection, tumor collection at time of surgery, and HLA genotyping; age 15 years and older, suspected ovarian cancer, and no prior surgical removal of tumor. For continuation in the study, patient had to have a confirmed pathologic diagnosis of epithelial ovarian cancer following surgery, be determined as HLA-A02:01+ and/or HLA-B07:02+, and express continued consent. Exclusion criteria were as follows: unable or unwilling to consent to any of the study criteria. Age was recorded at the time of initial consent. Blood was collected at the time of phlebotomy for normal clinical care at one or more of the following disease/treatment time points: prior to surgery, after surgery, after chemotherapy, and at disease recurrence. Ascites samples were collected for research if/when removed as part of normal clinical care. Tumor specimens were collected at time of surgery. Specimens collected for immune response analyses were frozen in vapor-phase liquid nitrogen until all samples from the patient could be analyzed in the same experiment.

HLA typing

Peripheral blood mononuclear cells (PBMC) from healthy donors and patients were assessed for expression of HLA-A*02:01 and HLA-B*07:02 by genotyping and flow cytometry. Healthy donor HLA genotyping was performed in house using AllSet+ Gold SSP Low-Resolution PCR-SSP HLA-A/B/C genotyping (Invitrogen/One Lambda; cat #54340D) per the manufacturer's instructions. High-resolution HLA genotyping for a subset of healthy donors and for all ovarian cancer patients was performed by the American Red Cross using next-generation sequencing (Supplementary Table S2). For donors and patients whose genotyping indicated HLA-A2 and/or HLA-B7, cells were assessed by flow cytometry to confirm expression of HLA-A2 (clone BB7.2; BD Biosciences; RRID: AB_10646036) and HLA-B7 (clone BB7.1; Novus Biologicals, RRID: AB_964531 or clone REA176, Miltenyi; RRID: AB_2733749).

Isolation of cells from whole blood

Whole blood was collected in sodium heparin vacutainer tubes (BD; cat #366480) and separated immediately using LymphoPrep SepMate (Stem Cell; cat #07801) per the manufacturer's instructions. PBMCs were washed in DPBS and frozen in 90% human AB+ serum (Gemini; cat #100-512)/10% DMSO in gas-phase liquid nitrogen until use. In some experiments, prior to freezing, monocytes were isolated from PBMCs by adherence to plastic or using the Human CD14 Positive Selection Kit (Miltenyi; cat #130-050-201). The nonadherent or CD14neg peripheral blood lymphocyte (PBL) fraction was kept frozen in vapor-phase liquid nitrogen until use.

Isolation of cells from solid tumors

Tumors were collected in RPMI-1640 and stored at 4°C until processing. Tumor pieces were cut into 1 to 3 mm3 pieces and digested in tumor medium—RPMI-1640, 2 mmol/L L-glutamine, 1% penicillin–streptomycin, and 2.5 μg/mL Amphotericin B (Gibco; cat #15290–018)—with 35.8 U/mL dispase (Gibco; cat #17105-041) for 30 minutes at 37°C with 5% CO2, swirling every 5 to 7 minutes as described (29). Samples were washed twice in cold tumor medium/10% FBS. Samples were incubated with DNase I (Fisher; cat #NC9709009) at 37°C for 10 minutes. Cells were washed twice in cold tumor medium/10% FBS and filtered through 70-μm nylon mesh, using the end of a plunger to create a single-cell suspension. Isolated cells were frozen in 90% human AB+ serum/10% DMSO in vapor-phase liquid nitrogen until all patient samples were ready to be analyzed in parallel.

Isolation of cells from ascites

Ascites were kept on ice until processing. Cells were collected by centrifugation at 3,000 × g for 20 minutes at 4°C and washed in cold tumor medium. Samples were incubated with DNase I at 37°C for 10 minutes and washed twice with tumor medium/10% FBS and filtered through 70-μm nylon mesh, using the end of a plunger to create a single-cell suspension. Isolated cells were frozen in 90% human AB+ serum/10% DMSO in vapor-phase liquid nitrogen.

Generation of monocyte-derived dendritic cells

Monocytes were matured into dendritic cells (DC) as previously described (30). Briefly, monocytes were cultured in RPMI-1640, 10% human AB+ serum (Gemini; cat #100-512), 2 mmol/L L-glutamine, 1.5% HEPES (Thermo Fisher; cat #15630130), 1% penicillin–streptomycin with GM-CSF (1,500 U/mL) and IL4 (2,900 U/mL). Cultures were supplemented with additional GM-CSF and IL4 on day 3 or 4. On day 6 or 7, immature DCs were removed, washed, and reseeded with GM-CSF (800 U/mL), IL4 (500 U/mL), IL6 (1,000 U/mL), IL1β (10 ng/mL), TNFα (10 ng/mL), and PGE2 (1 μg/mL). Mature DCs (mDC) were harvested at day 10 and frozen until ready to use in culture.

Enrichment of CD8+ T cells from peripheral blood of healthy donors and patients

Previously frozen PBMCs or PBLs were thawed and rested overnight at 37°C with 5% CO2 in culture medium: AIM-V (Gibco; cat #12055083), 5% human AB+ serum, 1% L-glutamine, 1.5% HEPES, 0.1% β-mercaptoethanol (Sigma; cat #M3148). Cells were cultured with low-dose IL7 (1 ng/mL). The following day, CD8+ T cells were resuspended in MACS buffer (PBS, 0.1% BSA, and 2 mmol/L EDTA at pH 7.2 and filtered) and isolated by negative depletion with a Human CD8 T-cell isolation kit (Miltenyi; cat #130-096-495) following the manufacturer's instructions and filtered through 70-μm nylon mesh. Antigen-experienced CD45RO+CD8+ T cells were isolated by positive selection using anti-CD45RO microbeads (Miltenyi; cat #130-046-001).

Enrichment of CD8+ T cells from tumors

Frozen single-cell suspensions of tumor specimens were thawed and treated with DNase I for 10 minutes at 37°C, washed twice with tumor medium/10% FBS, and rested overnight in culture medium with low-dose IL2 (20 IU/mL; ref. 31). CD8+ T cells were resuspended in MACS buffer and were either isolated by positive selection with a human CD8 T-cell isolation kit (Dynabead; cat #11333D) or by fluorescence activated cell sorting (FACS).

FACS

Previously frozen PBMCs or PBLs were thawed and rested overnight in culture medium with low-dose IL7 (1 ng/mL). Previously frozen tumors were thawed, treated with DNase I as described, and rested overnight in culture medium with low-dose IL2 (20 IU/mL). Cells were washed, filtered through 70-μm nylon mesh, and resuspended in sorting medium: RPMI-1640 lacking phenol red (Gibco Thermo Fisher; cat #11835030), 1% L-glutamine, and 2% human AB+ serum. Cells were incubated with Fc block (BD Pharmingen; cat #564220) for 10 minutes at room temperature, washed, and stained for 20 minutes at 37°C followed by 10 minutes at 4°C. 7-AAD (BioLegend; cat #420403) was added immediately before sorting on the Influx Cell Sorter (BD Model #646500). CD8+ T cells were gated as live, singlet, lymphocyte, CD3+CD8+ cells and sorted into 4 subsets based on expression of CCR7 and CD45RO, using “Fluorescence Minus One” to set gating as previously described (ref. 32; Supplementary Fig. S1). To isolate naïve cells from memory stem cells, the CCR7+CD45RO subset was further gated on CD95 (Supplementary Fig. S1). In patients, the same sorting strategy was used, but the sorted subsets were combined into only 2 subsets representing either resting or recently activated cells (Supplementary Fig. S2). For FACS, we used fluorophore-conjugated antibodies specific for the following human cell-surface antigens: CD3 (UCHT-1, RRID: AB_2744392) and CCR7 (150503, RRID: AB_10561679) from BD and CD8 (RPA-T8, RRID: AB_1595443), CD45RO (UCHL1, RRID: AB_314426), and CD95 (DX2, RRID: AB_314546) from BioLegend. Sorted cells were collected in AIM-V, 20% human AB+ serum, 1% L-glutamine, 1.5% HEPES, 0.1% β-mercaptoethanol, and 1% penicillin–streptomycin. Flow cytometry files were analyzed using FlowJo version 10.7.1 (Treestar, RRID: SCR_008520).

Stimulation of CD8+ T cells

mDCs and PBMCs, PBLs, or tumor-infiltrating lymphocytes (TIL) were thawed and rested overnight prior to stimulation. On the day of stimulation, subsets of CD8+ T cells were isolated as described above. Autologous mDCs were irradiated at 3,000 rads, washed, and then peptide pulsed (10 μg/mL each of 1–9 peptides) with 3 μg/mL beta-2-microglobulin (β2m; Millipore Sigma; cat #47-582-3250) for at least 2 hours. The mDCs were washed twice before being cocultured with CD8+ T-cell subsets at a 1:10 ratio (10,000 mDCs:100,000 T cells) in culture medium with IL7 (10 ng/mL) at 37°C with 5% CO2. Due to limited numbers of patient-derived mDCs, patient CD8+ T-cell subsets were cocultured with autologous peptide-pulsed mDCs plus autologous peptide-pulsed CD4CD8 PBLs (irradiated at 4,000 rads). Three days later, IL7 (10 ng/mL) and IL15 (10 ng/mL) were added. On day 7 and then as needed, the medium was replaced with fresh medium and IL7 and IL15. Patient TIL-derived CD8+ T cells were cultured with 6,000 U/mL IL2 instead of IL7 and IL15. As positive controls, CD8+ T-cell subsets were cocultured with mDCs pulsed with one or more viral peptides derived from influenza, EBV, or cytomegalovirus (CMV), and then analyzed for responses at day 14. As a negative control, CD8+ T-cell subsets were cocultured with mDCs pulsed with Ebola peptide, NP44–52 or NP294–302, and then analyzed for responses at day 14. No responses were detected in any donor to Ebola, demonstrating that the assay assessed memory responses and did not induce de novo responses.

Assessment of phosphopeptide-specific responses by IFNγ ELISpot assay

Responses were assessed either directly ex vivo (“direct”) or after one in vitro stimulation with peptide-pulsed antigen-presenting cells and 14-day culture with cytokines (“cultured”). T2-B7 target cells were pulsed with 3 μg/mL β2 m and 10 μg/mL peptide or 0.1% DMSO alone for 2 to 4 hours at room temperature on a rocker. IFNγ Single-Color ELISpot Plates (Cellular Technology Limited; cat #hIFNγ-1M/2) were activated with 70% ethanol and incubated with primary antibody at 4°C overnight. Isolated subsets of CD8+ T cells (25,000), either direct or cultured, were plated with 75,000 target cells per well, incubated at 37°C in 5% CO2 for 18 to 20 hours, and developed using the specified reagents per the manufacturer's instructions. Spots were counted using a BioReader (BioSys Models #4000 and #7000F). Responses were determined in triplicate and quantified as the mean of peptide-pulsed over DMSO-pulsed target cells. Positive responses met 3 criteria: mean of response was at least 10 spot-forming cells (SFC)/25,000 cells over mean of background, mean of response was at least two times the mean of the background, and the standard deviations of the response and background samples were nonoverlapping. Responses meeting these 3 criteria invariably had P values of <0.05 using unpaired Student t test.

Calculation of total responses apportioned among T-cell subsets

We calculated the total number of IFNγ+ cells per 106 CD8+ T cells for each subset, and the fraction that each subset contributed to the overall cultured and direct response. First, the total number of IFNγ+ cells in each culture was calculated from the number of IFNγ+ SFCs over background per 25,000 cells in the ELISpot assay (Equation 1). For cultured samples, this number was multiplied by the total number of recovered cells in the culture (Equation 1). We normalized the number of IFNγ+ cells based on the T-cell subset's percentage out of CD8+ in the donor's blood, determined by flow cytometry (Equation 2). The percentage contribution of each subset to the total response was calculated after summing these values for all subsets.

formula
formula

Kinase recognition motifs

Peptide sequences ± 20 amino acids were compiled using BLASTp (RRID: SCR_001010; ref. 33). The kinase recognition motifs (KRM) that could generate each phosphosite were determined using: http://hprd.org/PhosphoMotif_finder, selecting for “Serine/Threonine motifs” and “Kinase Phosphatase motifs” (34). For nested KRMs, the analysis of enrichment of memory targets was performed for each set of nested motifs, from most inclusive to most restrictive.

Isolation of HLA-associated peptides

MHC class I HLA molecules were immunoaffinity purified from samples, and phosphopeptides were isolated as previously described (35). Cells from ascites were collected by centrifugation at 3,000 × g for 20 minutes at 4°C and washed in cold tumor medium. Samples were incubated with DNase I at 37°C for 10 minutes and washed twice with tumor medium/10% FBS and filtered through 70-μm nylon mesh, using the end of a plunger to create a single-cell suspension. Cells were washed twice with cold DPBS, pelleted in a 50-mL conical tube, and frozen at −80°C. Tumor pieces were cut into 1 to 3 mm3 pieces and frozen in a 50 mL conical tube at −80°C. Cells (3–4 × 109) or 1 g of tissue was lysed in a solution containing 20 mmol/L Tris-HCl, pH 8.0, 150 mmol/L NaCl, 1% CHAPS, 5 mmol/L EDTA, and 1 mmol/L PMSF. Phosphatase inhibitors (Sigma-Aldrich Cocktail II; cat #P5726 and Cocktail III; cat #P0044) were used at 1:100 dilution to prevent dephosphorylation during extraction. Lysates were centrifuged for 1 hour at 100,000 × g at 4°C, and subsequent supernatant was added to anti-HLA-A, B, C (clone W6/32, Bio X Cell, RRID: AB_1107730) bound to NHS-activated Sepharose 4 Fast-Flow (GE; cat #17-0906-01) beads. After rotation overnight at 4°C, beads were transferred to poly-prep columns (Bio-Rad; cat #731-1550) and serially washed once with lysis buffer, twice with 20 mmol/L Tris-HCL, 150 mmol/L NaCl, pH 8.0, twice with 20 mmol/L Tris-HCL, 1M NaCl, pH 8.0, and 3 times with 20 mmol/L Tris-HCL, pH 8.0. Contents were transferred to Amicon Ultra-4 centrifugal filters (Millipore; cat #UFC801024) and centrifuged at 5,000 × g at 4°C for 30 to 60 minutes until all fluid was removed from beads. Samples were then subjected to liquid chromatography/MS and phosphopeptide enrichment as previously described (35).

Statistical analyses

Statistical analyses were performed in GraphPad Prism version 9.0.0 (RRID: SCR_002798). A P < 0.05 was considered significant for unpaired Student t test. To evaluate the levels of viral responses in patients to those in healthy donors, Welch t test was performed. Correlations for donor age and percentage of phosphopeptides recognized were analyzed by calculating a Pearson R value. Ninety-five percent confidence intervals (95% CI) for proportions were calculated using the Wilson–Brown method.

CD8+ T-cell memory to cancer-expressed phosphopeptides is highly variable among healthy donors

To determine the extent to which healthy donors show preexisting immune memory to phosphopeptides, we purified CD8+CD45RO+ T cells from 15 healthy donors, including 6 HLA-A2+, 5 HLA-B7+, and 4 HLA-A2+/-B7+ donors. This T-cell population consists of central and effector memory cells (36). We evaluated IFNγ responses after in vitro stimulation with peptide-pulsed mDCs and culture with cytokines for 14 days (“cultured”). We assessed responses to 205 HLA-A2– or HLA-B7–restricted phosphopeptides previously shown to be presented on (i) one or more patient-derived tumors (leukemias, colorectal cancer, or ovarian cancer) and/or (ii) one or more cancer cell lines (melanoma, ovarian cancer, or breast cancer), and, in some cases, also (iii) EBV-transformed B-cell lines (Supplementary Table S1). There were no responses detected to Ebola peptides in any donor, consistent with assay conditions testing memory and not naïve immune responses (Fig. 1A and B). Responses were frequently detected to peptides from influenza, CMV, and EBV. All ten HLA-A2+ donors showed preexisting immune memory to at least one of 51 HLA-A2–restricted phosphopeptides (Fig. 1A), whereas 7 of 9 HLA-B7+ donors showed preexisting immune memory to at least one of 154 HLA-B7–restricted phosphopeptides (Fig. 1B). Individual donors responded to between 1 and 14 (mean = 7, median = 5.5) HLA-A2–restricted phosphopeptides and between 0 and 20 (mean = 6, median = 3) HLA-B7–restricted phosphopeptides. However, in donors analyzed later, we examined responses only to the subset of phosphopeptides recognized by at least one earlier donor (Supplementary Tables S4 and S5). No responses were detected to a subset of unphosphorylated peptides, which is consistent with the generation of phosphopeptides—but not the unphosphorylated version of the peptide—under immunogenic exposures and the absence of tolerance in that donor against the recognized phosphopeptide (Supplementary Fig. S3A and S3B). Thus, most healthy donors exhibited preexisting immune memory to some MHC class I–restricted phosphopeptides, although the breadth varied.

Figure 1.

Most—but not all—healthy donors demonstrate preexisting T-cell immune memory to phosphopeptides displayed on cancer cells. Summary response data for CD8+CD45RO+ T cells from healthy donors (HD) stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, measured in triplicate wells in an IFNγ ELISpot assay. Data represent a minimum of 3 experiments for each peptide to which a response was observed in a donor. Peptides to which responses were not observed initially may have been analyzed in fewer experiments. A, Responses of HLA-A2+ donors to the indicated HLA-A2–restricted peptides. B, Responses of HLA-B7+ donors to HLA-B7–restricted peptides. The bottom portion of B presents data for 5 HLA-B7–restricted phosphopeptides that were analyzed for HD43, HD44, HD67, and HD89 because of limited sample availability. Insufficient PBMCs were available from HD67 and HD89 to assess responses against the larger cohort of HLA-B7–restricted phosphopeptides.

Figure 1.

Most—but not all—healthy donors demonstrate preexisting T-cell immune memory to phosphopeptides displayed on cancer cells. Summary response data for CD8+CD45RO+ T cells from healthy donors (HD) stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, measured in triplicate wells in an IFNγ ELISpot assay. Data represent a minimum of 3 experiments for each peptide to which a response was observed in a donor. Peptides to which responses were not observed initially may have been analyzed in fewer experiments. A, Responses of HLA-A2+ donors to the indicated HLA-A2–restricted peptides. B, Responses of HLA-B7+ donors to HLA-B7–restricted peptides. The bottom portion of B presents data for 5 HLA-B7–restricted phosphopeptides that were analyzed for HD43, HD44, HD67, and HD89 because of limited sample availability. Insufficient PBMCs were available from HD67 and HD89 to assess responses against the larger cohort of HLA-B7–restricted phosphopeptides.

Close modal

We asked whether multiple donors recognized the same phosphopeptides, which would suggest common immunogens. Of the 26 HLA-A2–restricted phosphopeptides that were targets of preexisting immune memory (hereafter referred to as “memory targets”), 17 were recognized by 2 or more donors (0.65, 95% CI, 0.46–0.81; Fig. 1A). However, only 3 of these (pIRS21097–1105, pCDC25B38–46, and pCHEK1367–377) were “immunodominant,” in that they were memory targets in more than half of the donors. In contrast, of the 32 HLA-B7–restricted memory targets, only 12 were recognized by 2 or more individuals (0.38, 95% CI, 0.23–0.55; Fig. 1B). Although two donors (HD43 and HD44) recognized a large number of HLA-B7–restricted memory targets in common, the remaining 5 donors recognized mostly nonoverlapping memory targets, and none met the criterion for immunodominance.

We asked whether any KRMs were enriched among memory targets, which could suggest that a subset of kinases is differentially responsible for generating immunogenic phosphopeptide exposures in healthy donors. However, there was no statistically significant enrichment of any KRMs among memory targets relative to the phosphopeptides not recognized by any donor (Supplementary Fig. S4A; Supplementary Table S6). Although pathogen exposures and transformation events are expected to increase with age, the percentage of phosphopeptides recognized did not correlate with donor age (Supplementary Fig. S4B). Finally, although chronic EBV infection is prevalent in humans, there was no difference in the proportion of memory targets between those expressed (0.24, 95% CI, 0.17–0.33) and not expressed (0.33, 95% CI, 0.25–0.44) on EBV-transformed cells. These results suggest that preexisting immune memory to most phosphopeptides reflects exposures that are not common among donors.

Immunogenic reexposures to a subset of phosphopeptides occurred in two healthy donors

To test the hypothesis that immunogenic exposures to phosphopeptides occurred recently or were ongoing, phosphopeptide responses in subsets of memory and effector cells were quantitated in HD43 and HD44. These donors were chosen because of the breadth of memory targets they recognized, many of which were shared, and their expression of both HLA-A2 and HLA-B7. First, their response patterns indicative of recent or ongoing exposure to well-characterized viral antigens were established. Using CD45RO and CCR7 expression, CD8+ T cells were sorted into central memory (TCM), effector memory (TEM), T effector memory reexpressing CD45RA (TEMRA), and naïve plus memory stem cells (TN + TSCM; Supplementary Fig. S1). In some experiments, the latter population was separated using CD95. Cultured IFNγ responses to viral peptides or phosphopeptides from each subset were quantified, and the contribution of each subset to the overall CD8+ PBMC response was determined by taking into account the increase in the number of T cells during the culture period and the fractional representation of each subset in the original PBMC isolate (see Materials and Methods). A similar method was used to evaluate direct effector activity in freshly isolated PBMC after 18 to 20 hours of incubation with the same peptides.

Under resting conditions, T cells specific for antigens that arise transiently or are expressed at low level are typically found mostly as TCM (37, 38). Consistent with this, cultured influenza M158–66-specific effectors in both donors were derived predominantly (57%–83%) from TCM (Fig. 2A and B). TEM and TN + TSCM subsets contributed to a lesser extent and TEMRA contributed negligibly (Fig. 2A and B). Direct effectors to M158–66 were low-level (300 and 100–300 IFNγ+ cells/106 CD8+ T cells in HD44 and HD43, respectively) and also derived predominantly from TCM. Cultured responses to LMP2A307–315, a latent phase epitope expressed at low levels during chronic EBV infection, were also derived predominantly (71%–98%) from TCM. Direct responses for LMP2A307–315 were low level (40–200 and 15–40 IFNγ+ cells/106 CD8+ T cells in HD44 and HD43, respectively) and derived from TEM in HD44 and TCM in HD43. The contributions of the TN+TSCM subset to the cultured responses in HD44 and HD43 were largely due to TSCM, although a small contribution from the TN subset was also evident in the HD44 response to M158–66 (Fig. 2C and D). In PBMCs collected from HD44 and HD43 16 to 17 days after their annual influenza vaccinations, TCM cells were no longer the predominant contributors (less than 35%) to the cultured response; instead, TCM, TEM, and TN+TSCM contributed similarly, and there was also a low but significant contribution from TEMRA, consistent with an ongoing response (Fig. 2A and B). Direct responses from PBMCs harvested 7–8 days after vaccination shifted entirely to TEM in HD44 and modestly in HD43, again consistent with an ongoing response, although the number of direct effectors (300–500 IFNγ+ cells/106 CD8+ T cells in both donors HD43 and HD44) was only modestly increased relative to the pre-vaccine condition.

Figure 2.

T-cell responses to viral epitopes define two response patterns that distinguish recent or ongoing antigen exposure. A–D, The indicated CD8+ T-cell subsets were enriched by cell sorting, and analyzed in an IFNγ ELISpot assay either after one in vitro stimulation with the indicated viral peptide–pulsed autologous DCs and a 14-day culture (cultured), or immediately ex vivo (direct). Responses were normalized for the expansion of cultured cells over 14 days and CD8+ T-cell subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). Responses to influenza M1 were measured in PBMCs collected before the donor received the annual flu vaccine and in the absence of illness (pre-vaccine) or 16 to 17 days after receiving an influenza vaccine (post-vaccine). C, Due to low cell numbers, no data are available for TSCM responses in HD44. The color coding for each subset is based on Supplementary Fig. S1. Most plots are representative of more than 1 experiment, except HD44 cultured and direct responses to EBV BBLF2/3, HD44 cultured response to CMV pp65495–503, HD43 direct response to influenza pre-vaccine, and cultured and direct response post-vaccine. HD, healthy donor.

Figure 2.

T-cell responses to viral epitopes define two response patterns that distinguish recent or ongoing antigen exposure. A–D, The indicated CD8+ T-cell subsets were enriched by cell sorting, and analyzed in an IFNγ ELISpot assay either after one in vitro stimulation with the indicated viral peptide–pulsed autologous DCs and a 14-day culture (cultured), or immediately ex vivo (direct). Responses were normalized for the expansion of cultured cells over 14 days and CD8+ T-cell subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). Responses to influenza M1 were measured in PBMCs collected before the donor received the annual flu vaccine and in the absence of illness (pre-vaccine) or 16 to 17 days after receiving an influenza vaccine (post-vaccine). C, Due to low cell numbers, no data are available for TSCM responses in HD44. The color coding for each subset is based on Supplementary Fig. S1. Most plots are representative of more than 1 experiment, except HD44 cultured and direct responses to EBV BBLF2/3, HD44 cultured response to CMV pp65495–503, HD43 direct response to influenza pre-vaccine, and cultured and direct response post-vaccine. HD, healthy donor.

Close modal

Responses to peptides from proteins expressed at high levels during the lytic phase of chronic viral infections are typically skewed toward TEM and TEMRA (37–39). In HD44, all four cell subsets contributed substantially to cultured responses to EBV BMLF1300–308 and BBLF2/3624–632 and to CMV pp65417–426 and pp65495–503, with effectors derived from TCM contributing only 21% to 37% and effectors derived from TEMRA contributing 18% to 35% of the total response (Fig. 2A). The contribution of the TN + TSCM subset was largely due to TSCM, although a small contribution from the TN subset was also evident for BMLF1300–308 (Fig. 2C). Strong direct responses (2,500–150,000 IFNγ+ cells/106 CD8+ T cells) were also evident, predominantly from the TEM and TEMRA subsets. The cultured response of HD43 to BMLF1300–308 differed from that of HD44, in that contributions from TCM was the predominant contributor, with only minor contributions from TEM and TEMRA (Fig. 2B). The contribution of the TN+TSCM subset was similar to that of HD44 and largely due to TSCM (Fig. 2D). However, a strong direct response (14,000 IFNγ+ cells/106 CD8+ T cells) was also evident, with elevated contributions from TCM and TN + TSCM compared with HD44. The elevated representation of TSCM in HD43's direct response to BMLF1300–308 likely reflects the elevated representation of TSCM (10%) in circulating CD8+ T cells in this donor versus 2% in HD44 (Supplementary Fig. S1) and 1% to 3% in humans overall (40), but the elevated representation of TCM is surprising to us. Nonetheless, these results establish two distinct response patterns: Response Pattern 1, in which TCM are the predominant contributors (greater than 55% of the total response) to cultured responses and direct effector activity is low, associated with nonrecent or low-level antigen exposure; and Response Pattern 2, in which the contribution of TEM, TEMRA, and/or TSCM to cultured responses is enhanced at the expense of TCM (making up less than 50% of the total response) and direct effector activity is high, associated with recent or ongoing high-level antigen exposure.

Using these patterns as a guide, we evaluated responses to a cohort of phosphopeptides recognized by these two donors in blood draws taken over 13 to 16 months. However, because of sample size limitations, we were often unable to compare cultured and direct responses in the same blood draw. Consequently, we categorized response patterns based on T-cell subset distribution in cultured responses. In the initial blood draws, both response patterns were observed to different phosphopeptides. Response Pattern 1 was evident in cultured responses of HD44 to pLSP1249–258, pWWTR186–94, pSRP72466–473, pCHEK1461–471, and pPEG10248–259, (Fig. 3), and in cultured responses of HD43 to pCHEK1461–471 and pSRP72466–473 (Fig. 4). In these instances, responses in the TN+TSCM subset were exclusively from TSCM (Figs. 3B and 4B). Response Pattern 2 was evident in initial cultured responses of HD44 to pPEG10248–258 and pCDC25B38–46 (Fig. 3A), and in responses of HD43 to pPEG10248–259 and pCDC25B38–46 (Fig. 4A). Interestingly, these response patterns changed in subsequent blood draws. For HD44, the response to pLSP1249–258 shifted from Pattern 1 to 2 at +10 months. The response to pPEG10248–258 shifted from Pattern 2 to 1 at all other time points, and the response to pCDC25B38–46 remained Pattern 2 until +13 months. For HD43, Pattern 1 to 2 shifts were evident in responses to pCHEK1461–471 at +4 months and pSRP72466–473 at +3 and +13 months, and Pattern 2 to 1 shifts were evident in responses to pPEG10248–259 at +13 months and to pCDC25B38–46 at +3, +13, and +16 months. Interestingly, the time points at which Pattern 2 responses for different peptides were evident were largely distinct for both donors (Figs. 3C and 4C), indicating that the processes resulting in their immunogenic expression were independent. These results suggest that, although phosphopeptides are infrequently expressed in the population, they are repeatedly reexpressed in an individual. We evaluated direct responses to the same phosphopeptides in both donors to determine whether ex vivo effector activity was consistent with the exposure status suggested by the cultured response. Direct responses in both donors at most time points were predominantly or exclusively derived from TCM or TEM and low-level (5–472 and 0–183 SFCs/106 CD8+ T cells in HD44 and HD43, respectively; Fig. 5A and B). These responses are consistent with Pattern 1 cultured responses to the same peptides observed at many of the same time points (Figs. 3 and 4), and with direct responses to M158–66 and LMP2A307–315 (Fig. 2). The cultured responses of HD44 to pCDC25B38–46 at +3 and +4 months and HD43 cultured response to pPEG10248–259 and pCDC25B38–46 at +6 months were all Pattern 2 (Figs. 3A and 4A). The direct response of HD43 to pCDC25B38–46 was not evident, but the direct response of HD44 to pCDC25B38–46 at +3 months was predominantly TEM, consistent with viral Pattern 2 direct responses (Fig. 2). However, the direct responses of HD44 to pCDC25B38–46 at +4 months and of HD43 to pPEG10248–259 at +6 months were almost exclusively TCM (Fig. 5A and B). This is inconsistent with viral Pattern 2 direct responses in which TEM and TEMRA are predominant (Fig. 2). These direct responses mediated by TCM suggest that these cells are not resting, but may represent alternatively differentiated effectors, which may suggest distinct immunogenic exposures to phosphopeptides.

Figure 3.

Longitudinal cultured response patterns to phosphopeptides by healthy donor 44 (HD44). A and B, CD8+ T-cell subsets isolated by FACS from blood samples collected at different time points were stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, and then responses to the indicated phosphopeptides were measured in triplicate wells using an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point. A, Longitudinal responses measured in the first collected blood sample (initial) and in samples collected at the indicated times in relation to the initial sample. B, The N + SCM subset was sorted using CD95 to evaluate responses from naïve or memory stem cells. Responses shown are from 5 months (pWWTR186–94), 10 months (pPEG10248–259 and pPEG10248–258), or 13 months (pLSP1249–258, pCHEK1461–471, pSRP72466–473, and pCDC25B38–46) and are representative of at least 2 examined time points. C, Responses to the peptides in each box were considered related to Response Pattern 2 and evidence of active immunogenic exposure at the indicated time points.

Figure 3.

Longitudinal cultured response patterns to phosphopeptides by healthy donor 44 (HD44). A and B, CD8+ T-cell subsets isolated by FACS from blood samples collected at different time points were stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, and then responses to the indicated phosphopeptides were measured in triplicate wells using an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point. A, Longitudinal responses measured in the first collected blood sample (initial) and in samples collected at the indicated times in relation to the initial sample. B, The N + SCM subset was sorted using CD95 to evaluate responses from naïve or memory stem cells. Responses shown are from 5 months (pWWTR186–94), 10 months (pPEG10248–259 and pPEG10248–258), or 13 months (pLSP1249–258, pCHEK1461–471, pSRP72466–473, and pCDC25B38–46) and are representative of at least 2 examined time points. C, Responses to the peptides in each box were considered related to Response Pattern 2 and evidence of active immunogenic exposure at the indicated time points.

Close modal
Figure 4.

Longitudinal cultured response patterns to phosphopeptides by healthy donor 43 (HD43). A and B, CD8+ T-cell subsets isolated by FACS from blood samples collected at different time points were stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, and then responses to the indicated phosphopeptides were measured in triplicate wells using an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point. A, Longitudinal responses measured in the first collected blood sample (initial) and in samples collected at the indicated times in relation to the initial sample. B, The N + SCM subset was sorted using CD95 to evaluate responses from naïve or memory stem cells. Responses shown are from 9 months (pPEG10248–259), 13 months (pCDC25B38–46), or 16 months (pCHEK1461–471 and pSRP72466–473) and are representative of at least 2 examined time points. C, Responses to the peptides in each box were considered related to Response Pattern 2 and evidence of active immunogenic exposure at the indicated time points.

Figure 4.

Longitudinal cultured response patterns to phosphopeptides by healthy donor 43 (HD43). A and B, CD8+ T-cell subsets isolated by FACS from blood samples collected at different time points were stimulated once in vitro with peptide-pulsed autologous mDCs and cultured for 14 days, and then responses to the indicated phosphopeptides were measured in triplicate wells using an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point. A, Longitudinal responses measured in the first collected blood sample (initial) and in samples collected at the indicated times in relation to the initial sample. B, The N + SCM subset was sorted using CD95 to evaluate responses from naïve or memory stem cells. Responses shown are from 9 months (pPEG10248–259), 13 months (pCDC25B38–46), or 16 months (pCHEK1461–471 and pSRP72466–473) and are representative of at least 2 examined time points. C, Responses to the peptides in each box were considered related to Response Pattern 2 and evidence of active immunogenic exposure at the indicated time points.

Close modal
Figure 5.

Direct responses to phosphopeptides in healthy donors (HD) are largely consistent with cultured response patterns. Responses of CD8+ T-cell subsets isolated by FACS from HD44 (A) and HD43 (B) to the indicated phosphopeptides were measured directly ex vivo and normalized for the subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point.

Figure 5.

Direct responses to phosphopeptides in healthy donors (HD) are largely consistent with cultured response patterns. Responses of CD8+ T-cell subsets isolated by FACS from HD44 (A) and HD43 (B) to the indicated phosphopeptides were measured directly ex vivo and normalized for the subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S1. Each time point reflects one analysis of PBMCs harvested at that time point.

Close modal

Immunity to phosphopeptides in ovarian cancer patients

We next evaluated phosphopeptide-specific responses of 8 HLA-A2+ and 2 HLA-A2+/B7+ ovarian cancer patients in relation to the phosphopeptides expressed on their tumors, which were identified by mass spectrometry; characteristics of the patients are given in Supplementary Table S3. From 12 tumor specimens (1 tumor from 8 patients and 2 tumor specimens from 2 patients), we identified 242 total and 180 unique phosphopeptides (Supplementary Table S7). Individual tumors expressed between 2 and 67 (mean = 20) phosphopeptides. For analysis of patient immune responses, phosphopeptides were chosen that were expressed on at least 1 tumor and/or were common memory targets in healthy donors, because these responses demonstrated the peptides' immunogenicity and likely absence of self-tolerance.

Because cell numbers in patient PBL and tumor samples were low, CD8+ T cells were sorted and pooled into only 2 subsets, (TCM + TN) and (TSCM + TEM + TEMRA; Supplementary Fig. S2A and S2B), which still enabled discrimination between Pattern 1 and Pattern 2 responses. Fourteen-day cultured responses to M158–66 and BMLF1300–308 viral peptides were evident in 9 of 9 and 8 of 9 assessed patients, respectively (Fig. 6; Supplementary Fig. S5A). Responses to the immunodominant pIRS21097–1105 were also evident in 4 of 10 patients (0.4, 95% CI, 0.17–0.69), which was not significantly different from healthy donors (0.6, 95% CI, 0.31–0.83; Fig. 6). This included patient VTB239, whose tumor expressed pIRS21097–1105, and 2 patients who also responded to the immunodominant pCHEK1367–377. However, the fractions of patients responding to pCHEK1367–377 (0.2, 95% CI, 0.04–0.51) and the other immunodominant memory target, pCDC25B38–46 (0.0, 95% CI, 0.00–0.30), were significantly lower than the fraction of healthy donors responding to them (1.0, 95% CI, 0.72–1.0 and 0.7, 95% CI, 0.40–0.89, respectively). TILs from two of these patients also showed responses to the same phosphopeptides. There were no responses to any other phosphopeptide in these 4 patients, including 12 that were expressed on their tumors. The remaining 6 patients demonstrated no responses to any evaluated phosphopeptide, including 6 that were expressed on their tumors (Supplementary Fig. S5A). Overall, there was no difference in the fraction of phosphopeptides recognized regardless of whether they were expressed (0.06, 95% CI, 0.00–0.28) or not expressed (0.07, 95% CI, 0.03–0.15) on the patient's tumor. These data suggest a general impairment in phosphopeptide immunity in ovarian cancer patients compared with healthy donors.

Figure 6.

Immune responses to viral and phosphorylated peptides by ovarian cancer patients. CD8+ T-cell subsets (TCM + TN and TEM + TEMRA + TSCM; Supplementary Fig. S2) isolated by FACS from 4 identified ovarian cancer patients (Supplementary Table S3) were stimulated once in vitro with the indicated viral or phosphorylated peptide–pulsed autologous mDCs and irradiated CD4CD8 PBLs as antigen-presenting cells and cultured for 14 days, and then responses to the same peptides were measured in triplicate wells in an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Each datapoint reflects one analysis of PBMCs or TILs harvested at the identified time point, and is the sum of responses from both sorted subsets. T, phosphopeptide expressed on the patient's tumor as identified by mass spectrometry.

Figure 6.

Immune responses to viral and phosphorylated peptides by ovarian cancer patients. CD8+ T-cell subsets (TCM + TN and TEM + TEMRA + TSCM; Supplementary Fig. S2) isolated by FACS from 4 identified ovarian cancer patients (Supplementary Table S3) were stimulated once in vitro with the indicated viral or phosphorylated peptide–pulsed autologous mDCs and irradiated CD4CD8 PBLs as antigen-presenting cells and cultured for 14 days, and then responses to the same peptides were measured in triplicate wells in an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Each datapoint reflects one analysis of PBMCs or TILs harvested at the identified time point, and is the sum of responses from both sorted subsets. T, phosphopeptide expressed on the patient's tumor as identified by mass spectrometry.

Close modal

To determine whether expression of pIRS21097–1105 on patient VTB239′s tumor could have influenced the response in this patient, we compared the response pattern of patient VTB239 to that of patient VTB241, whose tumor did not express pIRS21097–1105 or pCHEK1367–377. Cultured PBL responses of patient VTB241 to pIRS21097–1105 and pCHEK1367–377 were predominantly derived from TCM with low cell numbers (200–600 IFNγ+ cells/106 CD8+ T cells; Fig. 7A), consistent with Response Pattern 1 and a nonrecent antigen exposure. Responses to influenza M1 and EBV BMLF1 peptides in this patient were consistent with Pattern 2, demonstrating active immunity. On the other hand, cultured responses of PBLs and TILs to pIRS21097–1105 of patient VTB239 were predominantly derived from the TEM + TEMRA + TSCM subset with high cell numbers (1,200–2,700 IFNγ+ cells/106 CD8+ T cells; Fig. 7B), consistent with Response Pattern 2. The representation of the TEM+TEMRA+TSCM subset in the pIRS21097–1105 response was substantially higher than in responses to viral peptides. This suggests that the expression of pIRS21097–1105 on patient VTB239′s tumor generated an active immune response.

Figure 7.

Phosphopeptide expression on patient tumor is associated with an active immune response. CD8+ T-cell subsets (TCM + TN and TEM + TEMRA + TSCM; Supplementary Fig. S2) isolated by FACS from ovarian cancer patients VTB241 (A) and VTB239 (B; Supplementary Table S3) were stimulated once in vitro with the indicated viral or phosphorylated peptide–pulsed autologous mDCs and irradiated CD4CD8 PBLs as antigen-presenting cells and cultured for 14 days, and then responses to the same peptides were measured in triplicate wells in an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S2. Each graph represents one analysis of PBLs or TILs.

Figure 7.

Phosphopeptide expression on patient tumor is associated with an active immune response. CD8+ T-cell subsets (TCM + TN and TEM + TEMRA + TSCM; Supplementary Fig. S2) isolated by FACS from ovarian cancer patients VTB241 (A) and VTB239 (B; Supplementary Table S3) were stimulated once in vitro with the indicated viral or phosphorylated peptide–pulsed autologous mDCs and irradiated CD4CD8 PBLs as antigen-presenting cells and cultured for 14 days, and then responses to the same peptides were measured in triplicate wells in an IFNγ ELISpot assay. Responses were normalized for the expansion of cultured cells and subset percentages in the donor's blood, as described in Materials and Methods. Responses in each subset are reported both as the number of IFNγ+ cells per 106 CD8+ T cells (left y-axis) and the subset percentage of the total measured response (right y-axis). The color coding for each subset is based on Supplementary Fig. S2. Each graph represents one analysis of PBLs or TILs.

Close modal

In this study, we characterized preexisting immune memory to more than 200 cancer-expressed phosphopeptides in a cohort of healthy donors. We also examined the possibility of ongoing effector activity over time against a subset of memory targets in 2 donors. Our analyses suggest that most phosphopeptide-specific immunity is due to less ubiquitous exposures, in that 90% of the memory targets were recognized by fewer than 30% of donors. Immunodominance was limited to 3 HLA-A2–restricted phosphopeptides and was not seen among the HLA-B7–restricted phosphopeptides. At the same time, however, there was evidence of recent or ongoing exposures to multiple phosphopeptides in 2 donors over a relatively short time frame. To the extent that we analyzed them, robust responses to phosphopeptides in healthy donors were not due to elevated precursor frequencies of naïve cells, as with MART-1 (41). Most preexisting immunity was due to TCM, and we also observed significant contribution from TSCM. The presence of these cells establishes the likely longevity of phosphopeptide-specific memory. However, ovarian cancer patients showed significantly diminished responsiveness to immunodominant peptides, and in most cases, did not show reactivity to phosphopeptides displayed on their tumors.

The presence of phosphopeptide-specific memory T cells likely reflects immunogenic exposures through cellular transformation and/or infections that alter the phosphoproteome, generating new or overexpressed MHC class I–presented antigens (1, 2, 42–44). In keeping with this, the phosphopeptides chosen for analysis were identified on patient-derived tumors, cancer cell lines, and EBV-transformed B-cell lines. In addition, although preexisting phosphopeptide-specific memory was evident in most healthy donors, the range of phosphopeptides recognized was highly variable, suggesting that the range of immunogenic exposures to different phosphopeptides varies among individuals. It is possible that additional memory targets and immunodominant phosphopeptides could have been identified if more donors, PBMC collection time points, or effector functions had been evaluated. With the exception of exposures that gave rise to immunodominant memory targets, this donor-to-donor variation is consistent with immunosurveillance of premalignant cancer cells, whose underlying kinase and phosphatase dysregulations differ among individuals. It is also consistent with the hypothesis that preexisting immunity is predominantly driven by infection with uncommon bacteria or viruses. In this regard, it is notable that some viruses dysregulate many of the same kinases and/or phosphatases that are dysregulated in cancer (42–44), potentially leading to presentation of new phosphopeptides.

It is conceivable that some donor-to-donor variation in phosphopeptide-specific memory could be due to underlying immunologic differences. Isoforms of the TAP transporter, Tapasin, TAP binding protein–related protein, and proteasomes have all been shown to restrict or enable peptide presentation (45–48). There could also be constraints in the T-cell–receptor repertoire that limit phosphopeptide recognition; however, the great diversity and high number of T-cell receptors in any individual make this unlikely (49). Although age is an influence on the number of exposures in an individual, it did not correlate with broader phosphopeptide immunity in our study. It is conceivable that some phosphopeptide-specific memory T cells have been generated by cross-reactivity to other peptides; however, we and others have previously demonstrated that several phosphopeptides are directly immunogenic in mice and humans (3, 7–9, 12, 14). Additionally, phosphopeptide-specific responses are highly dependent on the phosphate moiety and exhibit negligible cross-reactivity on the unmodified peptide (here and refs. 3–7, 10, 12, 14–17). We think it unlikely that an unphosphorylated peptide can mimic the bulkiness and charge of the phosphate group. It remains possible that the phosphopeptide-specific T cells we observed were elicited by another phosphopeptide with a distinct peptide sequence. On the other hand, donor-to-donor differences in self-tolerance based on presentation of phosphopeptides or cross-reactive self-antigens could limit the ability to generate phosphopeptide-specific T-cell responses. All of the above hypotheses may explain donor-to-donor variation in responsiveness. Regardless, donors responded to an average of only 7 of 51 HLA-A2– and 6 of 154 HLA-B7–restricted phosphopeptides. We think again that this suggests that the immunogens generating phosphopeptide-specific memory are rare infectious agents or incipient cancer cells with distinct dysregulations.

The T-cell subsets specific for most phosphopeptides at most time points in cultured responses were predominantly TCM and TSCM. This pattern is similar to Response Pattern 1 for influenza M158–66 in the absence of recent vaccination or infection, and EBV LMP2A426–434, which is expressed persistently at low level. This suggests low or negligible phosphopeptide exposure at the time of PBMC collection. However, for some phosphopeptides, we also observed time points at which TEM and TEMRA were represented at much higher levels. These patterns were similar to Response Pattern 2 for EBV BMLF1300–308, which is expressed at high level, suggesting an active phosphopeptide response. However, while direct responses to Pattern 2 viral peptides were predominantly mediated by TEM and TEMRA, direct responses to phosphopeptides were considerably more variable. Although TEM and/or TEMRA were dominant at some time points, TCM were dominant at others, and these were sometimes associated with a Pattern 2 cultured response. This suggests that immunogenic phosphopeptide exposure may induce distinct differentiation of effectors, and that this also varies over time. Nonetheless, some of these active responses occurred repeatedly within a donor, suggesting recurrent exposures to the same immunogen, while active responses to individual phosphopeptides occurred at different times. This suggests that expression of different phosphopeptides may be driven by distinct immunogens.

We also demonstrated that in a small subset of healthy donors and most ovarian cancer patients, phosphopeptide-specific IFNγ T-cell responses were lacking. In the patients this was not limited to the phosphopeptides expressed on their tumors, consistent with the possibility that phosphopeptide-specific immune tolerance was not induced by expression on their cancer cells. One possibility, in patients as well as healthy donors, is that the individuals lack prior exposure to relevant infectious agents or transformed cells expressing phosphopeptides. These individuals may also express alleles of antigen-processing pathway components that alter phosphopeptide display, leading to either lack of presentation or enhanced presentation resulting in self-tolerance. It is also possible that additional responses could have been identified by cytotoxicity, TNFα production, or multimer staining. The number of phosphopeptides evaluated in patients was limited by sample availability. Importantly, a subset of patients still responded to immunodominant phosphopeptides, and a robust effector response in PBLs and TILs to a tumor-expressed phosphopeptide was evident in one patient.

These findings create a therapeutic opportunity, in that the tumors from these patients express many phosphopeptides that they have not responded to. Along with the demonstration that phosphopeptide-specific T-cell responses correlate with delayed tumor outgrowth in humanized murine models (7–9) and recently published results demonstrating that a phosphopeptide vaccine induces immune responses in melanoma patients (8), this observation supports further investigation into augmenting or inducing tumor-specific phosphopeptide immunity in patients where it is otherwise absent and into the ability of this approach to contribute to tumor control. Furthermore, T cells specific for tumor-expressed phosphopeptides can be collected from some patients and expanded in vitro into IFNγ-producing effectors. This provides a means for isolating reactive T-cell receptors that could be utilized in recombinant T-cell receptor–adoptive cell therapy.

Further investigation is required to elucidate mechanisms by which phosphopeptide-specific T-cell memory is limited in some individuals and many cancer patients. Based on genetic and environmental factors, some individuals have a higher risk of developing cancer, which likely coincides with a higher frequency or wider breadth of spontaneously arising premalignant cells. These individuals may exhibit preexisting immune memory to a greater number of phosphopeptides. Longitudinal studies are needed to determine if the ability of patients to generate phosphopeptide-specific responses is impaired prior to the development of cancer or if preexisting responses are lost or compromised following its development. It would be particularly interesting to investigate preexisting immune memory to phosphopeptides in patients with genetic syndromes known to increase cancer susceptibility (such as Li–Fraumeni or Lynch syndrome), in patients with “premalignant syndromes” (such as monoclonal gammopathy of unknown significance), and in individuals with histories of high carcinogenic exposures (such as smoking or working in certain industries). Regardless of the reasons, the lack of responses in some donors and patients identifies a therapeutic opportunity to induce responses that are otherwise absent.

A.M. Lulu reports other support from Agenus Inc. during the conduct of the study. K.L. Cummings reports other support from Agenus Inc. during the conduct of the study; other support from Agenus Inc. outside the submitted work; and a patent for WO 2015/160928 A2 pending to Agenus Inc., a patent for US 2017/0333541 A1 pending to Agenus Inc., and a patent for AU 2019204422 A1 pending to Agenus Inc. E.D. Jeffery reports other support from Agenus Inc. outside the submitted work, as well as a patent for phosphopeptides licensed. P.T. Myers reports a patent for WO2017192969A1 issued and with royalties paid, and was an employee with Agenus Inc. at the time of this work and owned Agenus Inc. stock. C.L. Slingluff Jr reports grants from NIH during the conduct of the study; grants from Celldex, Melanoma Research Alliance, Ludwig Cancer Institute, Cancer Research Institute, GlaxoSmithKline, and Merck and other support from Immatics, CureVac, and Polynoma outside the submitted work; and patents for peptides used in vaccines pending, issued, licensed, and with royalties paid from GlaxoSmithKline and Ludwig Institute. V.H. Engelhard reports grants from USPHS and grants, personal fees, and other support from Agenus Inc. during the conduct of the study, as well as a patent related to phosphopeptides pending, licensed, and with royalties paid from Agenus Inc. No disclosures were reported by the other authors.

A.M. Lulu: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. K.L. Cummings: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. E.D. Jeffery: Data curation, investigation, methodology, writing–review and editing. P.T. Myers: Data curation, investigation, methodology, writing–review and editing. D. Underwood: Data curation, project administration, writing–review and editing. R.M. Lacy: Resources, data curation, project administration, writing–review and editing. K.A. Chianese-Bullock: Resources, data curation, project administration, writing–review and editing. C.L. Slingluff Jr: Resources, data curation, supervision, writing–original draft, project administration, writing–review and editing. S.C. Modesitt: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing, patient recruitment. V.H. Engelhard: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

The authors thank donors and patient participants, Dr. Gina Petroni for advice on data analysis, Sara Adair for phlebotomy, Kelly Smith for reading plates, and the University of Virginia Gynecologic Oncology and Surgical teams for collection of patient data and specimens. They appreciate insightful discussions and advice from the Engelhard, Slingluff, Bullock, and Rutkowski labs and from Drs. Tim Bender and Janet Cross. The authors thank the University of Virginia Flow Cytometry Core, particularly Claude Chew, for help with the cell sorting experiments and the University of Virginia Biorepository and Tissue Research Facility for help obtaining patient samples. This work was supported by USPHS Grant R01CA190665 and a Sponsored Research Agreement from Agenus Inc. (V.H. Engelhard). A.M. Lulu was supported by the USPHS Grant T32 CA009109. The University of Virginia Flow Cytometry and Biorepository and Tissue Research Cores were supported by USPHS Cancer Center Support Grant P30CA44579.

Please note that refs. 50–56 relate to Supplementary Table S1.

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.
Fleuren
EDG
,
Zhang
L
,
Wu
J
,
Daly
RJ
. 
The kinome “at large” in cancer
.
Nat Rev Cancer
2016
;
16
:
83
98
.
2.
Stebbing
J
,
Lit
LC
,
Zhang
H
,
Darrington
RS
,
Melaiu
O
,
Rudraraju
B
, et al
The regulatory roles of phosphatases in cancer
.
Oncogene
2014
;
33
:
939
53
.
3.
Zarling
AL
,
Polefrone
JM
,
Evans
AM
,
Mikesh
LM
,
Shabanowitz
J
,
Lewis
ST
, et al
Identification of class I MHC-associated phosphopeptides as targets for cancer immunotherapy
.
Proc Natl Acad Sci U S A
2006
;
103
:
14889
94
.
4.
Petersen
J
,
Wurzbacher
SJ
,
Williamson
NA
,
Ramarathinam
SH
,
Reid
HH
,
Nair
AKN
, et al
Phosphorylated self-peptides alter human leukocyte antigen class I-restricted antigen presentation and generate tumor-specific epitopes
.
Proc Natl Acad Sci U S A
2009
;
106
:
2776
81
.
5.
Meyer
VS
,
Drews
O
,
Günder
M
,
Hennenlotter
J
,
Rammensee
HG
,
Stevanovic
S
. 
Identification of natural MHC class II presented phosphopeptides and tumor-derived MHC class I phospholigands
.
J Proteome Res
2009
;
8
:
3666
74
.
6.
Cobbold
M
,
De La Peña
H
,
Norris
A
,
Polefrone
J
,
Qian
J
,
English
AM
, et al
MHC class-I associated phosphopeptides are the targets of memory-like immunity in leukemia
.
Sci Transl Med
2013
;
5
:
203ra125
.
7.
Zarling
AL
,
Obeng
RC
,
Desch
AN
,
Pinczewski
J
,
Cummings
KL
,
Deacon
DH
, et al
MHC-restricted phosphopeptides from insulin receptor substrate-2 and CDC25b offer broad-based immunotherapeutic agents for cancer
.
Cancer Res
2014
;
74
:
6784
95
.
8.
Engelhard
VH
,
Obeng
RC
,
Cummings
KL
,
Petroni
GR
,
Ambakhutwala
AL
,
Chianese-Bullock
KA
, et al
MHC-restricted phosphopeptide antigens: preclinical validation and first-in-humans clinical trial in participants with high-risk melanoma
.
J Immunother Cancer
2020
;
8
:
e000262
.
9.
Lin
MH
,
Shen
KY
,
Liu
BS
,
Chen
IH
,
Sher
YP
,
Tseng
GC
, et al
Immunological evaluation of a novel HLA-A2 restricted phosphopeptide of tumor associated antigen, TRAP1, on cancer therapy
.
Vaccine X
2019
;
1
:
100017
.
10.
Solleder
M
,
Guillaume
P
,
Racle
J
,
Michaux
J
,
Pak
HS
,
Müller
M
, et al
Mass spectrometry based immunopeptidomics leads to robust predictions of phosphorylated HLA class I ligands
.
Mol Cell Proteomics
2020
;
19
:
390
404
.
11.
Bassani-Sternberg
M
,
Bräunlein
E
,
Klar
R
,
Engleitner
T
,
Sinitcyn
P
,
Audehm
S
, et al
Direct identification of clinically relevant neoepitopes presented on native human melanoma tissue by mass spectrometry
.
Nat Commun
2016
;
7
:
13404
.
12.
Andersen
MH
,
Bonfill
JE
,
Neisig
A
,
Arsequell
G
,
Sondergaard
I
,
Valencia
G
, et al
Phosphorylated peptides can be transported by TAP molecules, presented by class I MHC molecules, and recognized by phosphopeptide-specific CTL
.
J Immunol
1999
;
163
:
3812
8
.
13.
Depontieu
FR
,
Qian
J
,
Zarling
AL
,
McMiller
TL
,
Salay
TM
,
Norris
A
, et al
Identification of tumor-associated, MHC class II-restricted phosphopeptides as targets for immunotherapy
.
Proc Natl Acad Sci U S A
2009
;
106
:
12073
8
.
14.
Zarling
AL
,
Ficarro
SB
,
White
FM
,
Shabanowitz
J
,
Hunt
DF
,
Engelhard
VH
. 
Phosphorylated peptides are naturally processed and presented by major histocompatibility complex class I molecules in vivo
.
J Exp Med
2000
;
192
:
1755
62
.
15.
Li
Y
,
Depontieu
FR
,
Sidney
J
,
Salay
TM
,
Engelhard
VH
,
Hunt
DF
, et al
Structural basis for the presentation of tumor-associated MHC class II-restricted phosphopeptides to CD4+ T cells
.
J Mol Biol
2010
;
399
:
596
603
.
16.
Mohammed
F
,
Cobbold
M
,
Zarling
AL
,
Salim
M
,
Barrett-Wilt
GA
,
Shabanowitz
J
, et al
Phosphorylation-dependent interaction between antigenic peptides and MHC class I: a molecular basis for the presentation of transformed self
.
Nat Immunol
2008
;
9
:
1236
43
.
17.
Mohammed
F
,
Stones
DH
,
Zarling
AL
,
Willcox
CR
,
Shabanowitz
J
,
Cummings
KL
, et al
The antigenic identity of human class I MHC phosphopeptides is critically dependent upon phosphorylation status
.
Oncotarget
2017
;
5
:
54160
72
.
18.
Kang
TW
,
Yevsa
T
,
Woller
N
,
Hoenicke
L
,
Wuestefeld
T
,
Dauch
D
, et al
Senescence surveillance of pre-malignant hepatocytes limits liver cancer development
.
Nature
2011
;
479
:
547
51
.
19.
Beatty
PL
,
van der Geest
R
,
Hashash
JG
,
Kimura
T
,
Gutkin
D
,
Brand
RE
, et al
Immunobiology and immunosurveillance in patients with intraductal papillary mucinous neoplasms (IPMNs), premalignant precursors of pancreatic adenocarcinomas
.
Cancer Immunol Immunother
2016
;
65
:
771
8
.
20.
Kocer
B
,
McKolanis
J
,
Soran
A
. 
Humoral immune response to MUC5AC in patients with colorectal polyps and colorectal carcinoma
.
BMC Gastroenterol
2006
;
6
:
4
.
21.
Goodyear
OC
,
Pratt
G
,
McLarnon
A
,
Cook
M
,
Piper
K
,
Moss
P
. 
Differential pattern of CD4+ and CD8+ T-cell immunity to MAGE-A1/A2/A3 in patients with monoclonal gammopathy of undetermined significance (MGUS) and multiple myeloma
.
Blood
2008
;
112
:
3362
72
.
22.
Vella
LA
,
Yu
M
,
Fuhrmann
SR
,
El-Amine
M
,
Epperson
DE
,
Finn
OJ
. 
Healthy individuals have T-cell and antibody responses to the tumor antigen cyclin B1 that when elicited in mice protect from cancer
.
Proc Natl Acad Sci U S A
2009
;
106
:
14010
5
.
23.
Richards
ER
,
Devine
PL
,
Quin
RJ
,
Fontenot
JD
,
Ward
BG
,
McGuckin
MA
. 
Antibodies reactive with the protein core of MUC1 mucin are present in ovarian cancer patients and healthy women
.
Cancer Immunol Immunother
1998
;
46
:
245
52
.
24.
Dunn
GP
,
Old
LJ
,
Schreiber
RD
. 
The immunobiology of cancer immunosurveillance and immunoediting
.
Immunity
2004
;
21
:
137
48
.
25.
Cramer
DW
,
Finn
OJ
. 
Epidemiologic perspective on immune-surveillance in cancer
.
Curr Opin Immunol
2011
;
23
:
265
71
.
26.
Su
Y
,
Tang
LY
,
Chen
LJ
,
He
JR
,
Su
FX
,
Lin
Y
, et al
Joint effects of febrile acute infection and an interferon-γ polymorphism on breast cancer risk
.
PLoS One
2012
;
7
:
e37275
.
27.
Iheagwara
UK
,
Beatty
PL
,
Van
PT
,
Ross
TM
,
Minden
JS
,
Finn
OJ
. 
Influenza virus infection elicits protective antibodies and T cells specific for host cell antigens also expressed as tumor-associated antigens: a new view of cancer immunosurveillance
.
Cancer Immunol Res
2014
;
2
:
263
73
.
28.
Anderson
KS
,
Alexander
J
,
Wei
M
,
Cresswell
P
. 
Intracellular transport of class I MHC molecules in antigen processing mutant cell lines
.
J Immunol
1993
;
151
:
3407
19
.
29.
Pribyl
LJ
,
Coughlin
KA
,
Sueblinvong
T
,
Shields
K
,
Iizuka
Y
,
Downs
LS
, et al
Method for obtaining primary ovarian cancer cells from solid specimens
.
J Vis Exp
2014
;
e51581
.
30.
Tsai
V
,
Kawashima
I
,
Keogh
E
,
Daly
K
,
Sette
A
,
Celis
E
. 
In vitro immunization and expansion of antigen-specific cytotoxic T lymphocytes for adoptive immunotherapy using peptide-pulsed dendritic cells
.
Crit Rev Immunol
1998
;
18
:
65
75
.
31.
Bobisse
S
,
Genolet
R
,
Roberti
A
,
Tanyi
JL
,
Racle
J
,
Stevenson
BJ
, et al
Sensitive and frequent identification of high avidity neo-epitope specific CD8 + T cells in immunotherapy-naive ovarian cancer
.
Nat Commun
2018
;
9
:
1092
.
32.
Abdelsamed
HA
,
Moustaki
A
,
Fan
Y
,
Dogra
P
,
Ghoneim
HE
,
Zebley
CC
, et al
Human memory CD8 T cell effector potential is epigenetically preserved during in vivo homeostasis
.
J Exp Med
2017
;
214
:
1593
606
.
33.
Altschul
SF
,
Gish
W
,
Miller
W
,
Myers
EW
,
Lipman
DJ
. 
Basic local alignment search tool
.
J Mol Biol
1990
;
215
:
403
10
.
34.
Amanchy
R
,
Periaswamy
B
,
Mathivanan
S
,
Reddy
R
,
Tattikota
SG
,
Pandey
A
. 
A curated compendium of phosphorylation motifs
.
Nat Biotech
2007
;
25
:
285
6
.
35.
Abelin
JG
,
Trantham
PD
,
Penny
SA
,
Patterson
AM
,
Ward
ST
,
Hildebrand
WH
, et al
Complementary IMAC enrichment methods for HLA-associated phosphopeptide identification by mass spectrometry
.
Nat Protoc
2015
;
10
:
1308
18
.
36.
Martin
MD
,
Badovinac
VP
. 
Defining memory CD8 T cell
.
Front Immunol
2018
;
9
:
2692
.
37.
Newell
EW
,
Sigal
N
,
Bendall
SC
,
Nolan
GP
,
Davis
MM
. 
Cytometry by time-of-flight shows combinatorial cytokine expression and virus-specific cell niches within a continuum of CD8+ T cell phenotypes
.
Immunity
2012
;
36
:
142
52
.
38.
van Aalderen
MC
,
Remmerswaal
EBM
,
Verstegen
NJM
,
Hombrink
P
,
ten Brinke
A
,
Pircher
H
, et al
Infection history determines the differentiation state of human CD8+ T cells
.
J Virol
2015
;
89
:
5110
23
.
39.
Hislop
AD
,
Gudgeon
NH
,
Callan
MFC
,
Fazou
C
,
Hasegawa
H
,
Salmon
M
, et al
EBV-specific CD8+ T cell memory: relationships between epitope specificity, cell phenotype, and immediate effector function
.
J Immunol
2001
;
167
:
2019
29
.
40.
Gattinoni
L
,
Lugli
E
,
Ji
Y
,
Pos
Z
,
Paulos
CM
,
Quigley
MF
, et al
A human memory T cell subset with stem cell-like properties
.
Nat Med
2011
;
17
:
1290
7
.
41.
Pittet
MJ
,
Valmori
D
,
Dunbar
PR
,
Speiser
DE
,
Liénard
D
,
Lejeune
F
, et al
High frequencies of naive Melan-a/Mart-1–specific Cd8+ T cells in a large proportion of human histocompatibility leukocyte antigen (Hla)-A2 individuals
.
J Exp Med
1999
;
190
:
705
16
.
42.
Guergnon
J
,
Godet
AN
,
Galioot
A
,
Falanga
PB
,
Colle
J-H
,
Cayla
X
, et al
PP2A targeting by viral proteins: a widespread biological strategy from DNA/RNA tumor viruses to HIV-1
.
Biochim Biophys Acta
2011
;
1812
:
1498
507
.
43.
Griego
SD
,
Weston
CB
,
Adams
JL
,
Tal-Singer
R
,
Dillon
SB
. 
Role of p38 mitogen-activated protein kinase in rhinovirus-induced cytokine production by bronchial epithelial cells
.
J Immunol
2000
;
165
:
5211
20
.
44.
Meineke
R
,
Rimmelzwaan
GF
,
Elbahesh
H
. 
Influenza virus infections and cellular kinases
.
Viruses
2019
;
11
:
171
.
45.
Yan
G
,
Shi
L
,
Faustman
D
. 
Novel splicing of the human MHC-encoded peptide transporter confers unique properties
.
J Immunol
1999
;
162
:
852
9
.
46.
Beutler
N
,
Hauka
S
,
Niepel
A
,
Kowalewski
DJ
,
Uhlmann
J
,
Ghanem
E
, et al
A natural tapasin isoform lacking exon 3 modifies peptide loading complex function
.
Eur J Immunol
2013
;
43
:
1459
69
.
47.
Hermann
C
,
van Hateren
A
,
Trautwein
N
,
Neerincx
A
,
Duriez
PJ
,
Stevanović
S
, et al
TAPBPR alters MHC class I peptide presentation by functioning as a peptide exchange catalyst
.
Elife
2015
;
4
:
e09617
.
48.
Groettrup
M
,
Kirk
CJ
,
Basler
M
. 
Proteasomes in immune cells: more than peptide producers?
Nat Rev Immunol
2010
;
10
:
73
8
.
49.
Robins
HS
,
Campregher
PV
,
Srivastava
SK
,
Wacher
A
,
Turtle
CJ
,
Kahsai
O
, et al
Comprehensive assessment of T-cell receptor β-chain diversity in αβ T cells
.
Blood
2009
;
114
:
4099
107
.
50.
EP3718556 target peptides for immunotherapy and diagnostics
.
Patent application
. 
2020
Dec 30.
51.
Abelin
JG
.
Identification of cancer immunotherapeutic candidates derived from dysregulated cell signaling pathways by mass spectrometry
.
Charlottesville (VA)
:
University of Virginia
; 
2013
.
52.
English
AM
.
Development and application of methodologies to facilitate peptide sequencing with mass spectrometry
.
Charlottesville (VA)
:
University of Virginia
; 
2011
.
53.
Norris
A
.
Mass spectrometry of immunologically relevant molecules
.
Charlottesville (VA)
:
University of Virginia
; 
2011
.
54.
Gerhard
DS
,
Wagner
L
,
Feingold
EA
,
Shenmen
CM
,
Grouse
LH
,
Schuler
G
, et al
The status, quality, and expansion of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC)
.
Genome Res
2004
;
14
:
2121
7
.
55.
Greenman
C
,
Stephens
P
,
Smith
R
,
Dalgliesh
GL
,
Hunter
C
,
Bignell
G
, et al
Patterns of somatic mutation in human cancer genomes
.
Nature
2007
;
446
:
153
8
.
56.
Obeng
RC
.
Tumor-associated HLA-A*0201-restricted phosphopeptides as candidates for cancer immunotherapy
.
Charlottesville (VA)
:
University of Virginia
; 
2011
.