Identification of Post-translationally Modified MHC Class I-Associated Peptides as Potential Cancer Immunotherapeutic Targets

Abstract

Over the past 3 decades, the Hunt laboratory has developed advancements in mass spectrometry-based technologies to enable the identification of peptides bound to major histocompatibility complex (MHC) molecules. The MHC class I processing pathway is responsible for presenting these peptides to circulating cytotoxic T cells, allowing them to recognize and eliminate malignant cells, many of which have aberrant signaling. Professor Hunt hypothesized that due to the dysregulation in phosphorylation in cancer that abnormal phosphopeptides could be presented by this pathway, and went on to demonstrate that this was, in fact, the case. Thereafter, the laboratory continued to sequence MHC-associated phosphopeptides and contributed several improved methods for their enrichment, detection, and sequencing. This article summarizes the most recent advancements in identification of modified MHC-associated peptides and includes the cumulative list of phosphopeptides sequenced by the Hunt lab. Further, many other post-translational modifications (PTMs) were found to modify MHC peptides, including O-GlcNAcylation, methylation, and kynurenine; in total, we present here a list of 2450 MHC-associated PTM peptides. Many of these were disease-specific and found across several patients, thus highlighting their potential as cancer immunotherapy targets. We are sharing this list with the field in hopes that it might be used in investigating this potential. Overall, the Hunt lab's contributions have significantly advanced our understanding of antigen presentation and dysregulation of PTMs, supporting modern immunotherapy and vaccine development efforts.

Keywords: MHC-associated peptides; cancer immunotherapy; immunopeptidomics; phosphorylation; post-translational modification.

Graphical abstract

Fig. 1

The MHC class I processing pathway. In all nucleated cells, cytosolic and nuclear proteins are degraded by proteasomes, transported to the ER to bind MHC molecules, and then presented on the cell surface for recognition by CD8+ T cells. Proteins are marked for degradation by ubiquitin, followed by digestion into peptides in the proteasome (1). Peptides are transported into the endoplasmic reticulum by TAP (2), where they associate with MHC molecules (3), and are then shuttled to the cell surface through the Golgi membrane (4). Finally, the peptide–MHC complex can be surveyed by CD8+ memory T cells for presence of infection or transformation (5). TAP, transporter associated with antigen processing.

Fig. 2

Optimized MHC-associated phosphopeptide enrichment procedure. Peptides are first immunopurified with a pan-class I antibody (W6/32) and eluted with acetic acid. Contaminants are removed using a HILIC-based cleanup, followed by Fischer esterification and iron-NTA enrichment of phosphopeptides. The iron is reduced with ascorbic acid which releases phosphopeptides, and then the resulting elution is desalted, separated by reversed-phase HPLC, and analyzed by mass spectrometry. All resulting spectra were manually validated. HILIC, hydrophilic interaction chromatography; NTA, nitrilotriacetic acid.

Fig. 3

Comparison of cleanup techniques for a heavily contaminated sample containing HLA peptides from cell line VMM39. HLA-associated peptides were isolated from VMM39 cells, followed by cleanup with C18 or PHEA STAGE tips. Samples were analyzed were performed on an LTQ FT Ultra. Data were searched using the Byonic node of Proteome Discoverer 3.3 against the human proteome. Identifications were considered confident if they had a Byonic score greater than 300, |logProb| greater than 3, and Proteome Discoverer considered the identification to be high confidence. Peptide quantification and chromatographic alignment were performed by Proteome Discoverer. A, PHEA cleanup (blue trace) allows for a higher peptide area and greater number of peptide identifications than no cleanup (green trace) or C18 (pink). B, Euler plot demonstrating the overlap in confident identifications between the three techniques. C, Euler plot showing overlap when the peptides were found by mass and retention time but were not identified. HLA, human leukocyte antigen; PHEA, polyhydroxyethyl aspartamide; STAGE, stop and go extraction.

Fig. 4

Ion suppression in a contaminated sample from THP-1 associated HLA peptides is reduced by HILIC cleanup. Shown here are comparisons of chromatograms and spectra after a STAGE tip cleanup (left) and subsequent HILIC cleanup (right) for THP-1 cell line. A, total ion current for the chromatographic gradient showing the removal of β2m (eluting at ∼40 min prior to HILIC cleanup). B, TIC (top) and extracted ion chromatogram for peptide SLPDFGISY (bottom) showing 20× signal intensity for the peptide after PHEA cleanup due to removal of suppressants. C, MS1 spectra at the peptide retention time shows significant noise that is removed after PHEA cleanup. With removal of the coeluting contaminants, the peptide accepted an additional charge, which allowed it to be selected for fragmentation. The peptide was then the most abundant species eluting at the time. β2m, β-2-microglobulin protein; HLA, human leukocyte antigen; HILIC, hydrophilic interaction chromatography; PHEA, polyhydroxyethyl aspartamide; STAGE, stop and go extraction; TIC, total ion chromatogram.

Fig. 5

Selective removal of beta-2-microglobulin from an HLA-associated peptide sample. Unlike reverse phase removal of β2m from samples, PHEA is able to retain peptides that elute after the protein despite its preference for hydrophilic molecules. A, total ion chromatogram (TIC) before (green) and after (blue) PHEA cleanup of a HeLa sample; after PHEA, the contaminant protein was only present at approximately 1% relative abundance. B, without removal of β2m, which elutes at ∼104 min, peptide identifications drop dramatically thereafter (green), whereas PHEA continues to have identifications at 140 min (blue), indicating selective separation of proteins from peptides. HLA, human leukocyte antigen; PHEA, polyhydroxyethyl aspartamide; TIC, total ion chromatogram.

Fig. 6

Tumor-associated phosphopeptides nominated for immunotherapeutic study.A, site-localized MHC class phosphopeptides were categorized into six sample groups. Patient tumor tissue constituted ∼50% of the samples tested and identified the second greatest number of phosphopeptides. Immortalized human cell lines identified the most phosphopeptides, however many of these were not observed in other sample groups. Xenograft cell lines, nontumor patient tissue, and T cells each identified several hundred phosphopeptides, while cadaverous brain tissue identified less than 20. B, tumor-associated phosphopeptides were found to be enriched in proteins involved in kinase signaling and transcriptional regulation when compared to MHC Class I peptides from the HLA Ligand Atlas. C, the likelihood of HLA presentation was estimated using BamQuery, which identified that 33 phosphopeptides may originate from proteins with low expression across healthy human tissues. D, of the 33 phosphopeptides with low HLA presentation likelihood, 28 peptides aligned to protein-coding genomic regions while five were predicted to come from intronic, intergenic, and noncoding regions. E, we nominate 16 tumor-associated phosphopeptides as candidate tumor antigens, including eight that have been observed in more than one patient tumor and four that were previously studied for immune reactivity as unmodified antigens. HLA, human leukocyte antigen.