Deciphering the immunopeptidome in vivo reveals new tumour antigens

Abstract

Immunosurveillance of cancer requires the presentation of peptide antigens on major histocompatibility complex class I (MHC-I) molecules^1-5. Current approaches to profiling of MHC-I-associated peptides, collectively known as the immunopeptidome, are limited to in vitro investigation or bulk tumour lysates, which limits our understanding of cancer-specific patterns of antigen presentation in vivo⁶. To overcome these limitations, we engineered an inducible affinity tag into the mouse MHC-I gene (H2-K1) and targeted this allele to the Kras^LSL-G12D/+Trp53^fl/fl mouse model (KP/K^bStrep)⁷. This approach enabled us to precisely isolate MHC-I peptides from autochthonous pancreatic ductal adenocarcinoma and from lung adenocarcinoma (LUAD) in vivo. In addition, we profiled the LUAD immunopeptidome from the alveolar type 2 cell of origin up to late-stage disease. Differential peptide presentation in LUAD was not predictable by mRNA expression or translation efficiency and is probably driven by post-translational mechanisms. Vaccination with peptides presented by LUAD in vivo induced CD8⁺ T cell responses in naive mice and tumour-bearing mice. Many peptides specific to LUAD, including immunogenic peptides, exhibited minimal expression of the cognate mRNA, which prompts the reconsideration of antigen prediction pipelines that triage peptides according to transcript abundance⁸. Beyond cancer, the K^bStrep allele is compatible with other Cre-driver lines to explore antigen presentation in vivo in the pursuit of understanding basic immunology, infectious disease and autoimmunity.

Extended Data Figure 1. (Related to Figure 1). In vitro validation of the K^bStrep allele.

a) Structural model depicting the topology of the K^bStrep protein during affinity purification. The StrepTagII engagement with Streptactin affinity resin does not interfere with peptide or B2m binding. b) Southern blot analysis of K^bStrep targeted KP* ES cells. c) Representative genotyping for WT/WT, K^bStrep/WT heterozygotes, and K^bStrep/K^bStrep homozygotes. d) Brightfield images of KP and KP/ K^bStrep pancreatic organoids pre- and post- Ad-CMV-Cre mediated transformation ex vivo. e) RT-PCR analysis of KP or KP/K^bStrep pancreatic organoids with or without Cre recombination and with or without IFN-γ treatment. Each row represents a distinct primer set showing no discernable alterations in mRNA splicing with or without StrepTagII activation. f) Representative flow cytometry plots detecting cell surface expression of PD-L1 and StrepTagII at baseline (red) and following Cre activation and IFN-γ treatment (blue). g) Quantification of the median fluorescence intensity (MFI) of StrepTagII staining in KP (orange) and KP/K^bStrep (blue) organoids in control, IFN-γ treated, Cre transformed, and Cre+IFN-γ treated samples. Data are mean ± sem (n=3). Two-sided Student’s t-test. h) Immunoblot analysis of whole cell lysate from KP or KP/K^bStrep PDAC cells after adaptation to 2D following treatment with IFN-γ. i) Immunoblot depicting affinity purification of intact MHC-I with Streptactin resin as evidenced by the co-precipitation of B2m. j) Coomassie staining of samples taken from KP or KP/K^bStrep lysates at various stages of purification. In this experiment, the elution was taken by incubating washed Streptactin resin with SDS-PAGE loading buffer.

Extended Data Figure 2. (related to Figure 1). Isolation of MHC-I complexes from PDAC in vivo.

a) Experimental illustration of samples used for immunopeptidome comparison in PDAC. b) Multiplexed immunofluorescence of a representative autochthonous PDAC tumor. White arrows indicate cancer cell nests. c) High magnification multiplexed immunofluorescence image depicting the specificity of StrepTagII staining on cancer cells. d) Schematic illustration of the traditional method of peptide extraction (top) and the method used for competitive elution of MHC-I complexes with biotin used in this study (bottom). e) Immunoblot demonstrating efficient precipitation of the MHC-I heavy and light chain with acid (1% TFA) prior to peptide clean-up with C18 ziptip. f) Representative immunoblot demonstrating purification of MHC-I specifically from KP/K^bStrep tumor bearing mice. g) Number of unique peptides identified in each sample type after filtering for length (8-11 amino acids) and NetMHCPan predicted affinity (<1000 nM). h) Amino acid length distribution of all peptides identified from 2D, Ortho, Auto, or WT samples. i) Peptide motifs of 8- and 9-mers isolated from 2D, orthotopic, and autochthonous tumors. j) Venn diagram comparison of peptides found in 2D, Ortho, or Auto samples. Peptides only found in vivo are outlined in red. k) Venn diagram comparing MHC-I peptides derived from normal pancreas in Schuster et. al. (gray) versus orthotopic transplant (light blue) or autochthonous PDAC (dark blue) in this study. l) UMAP embedding of reanalyzed pancreatic scRNAseq data from the Tabula Muris. For clarity of presentation in Extended Data Figure 2m, cells from a specific lineage were collapsed into a single cluster if they originally separated into multiple clusters (i.e. Alpha 1 and Alpha 2 → Alpha). m) Expression of gene signatures derived from genes encoding for normal pancreas peptides (gray), orthotopic PDAC peptides (light blue), and autochthonous PDAC peptides (dark blue) in all cell types of the normal pancreas as measured by scRNAseq from Tabula Muris.

Extended Data Figure 3. (related to Figure 1). In vivo validation of K^bStrep allele in KP LUAD

a) Representative H&E images demonstrating adenocarcinoma in KP and KP/K^bStrep tumors. b) Multiplexed immunofluorescence of 16-week KP/K^bStrep tumors demonstrating specific StrepTagII detection on tumor cells within the tumor microenvironment. c) Multiplexed immunofluorescence of a WT KP tumor demonstrating no detection of the StrepTagII in tumors outlined in white dotted lines. d) Gating strategy for isolating cells positive for an alveolar type 2 (AT2) phenotype from KP tumor-bearing lung tissue. e) Histograms depicting StrepTagII staining intensity across all CD45⁻ cells (left) or after gating for AT2 cells (right) in KP (pink) or KP; K^bStrep (purple) tumors. f) Quantification of StrepTagII MFI on AT2 or CD45 cells in the tumor microenvironment from KP or KP/K^bStrep tumors. Data are mean ± sem. Two-sided Student’s t-test. g) Relative abundance of CD4⁺ T cells, CD8⁺ T cells, Macrophages, and CD45⁺MHCII⁺ immune cells in the tumor microenvironment of KP and KP/K^bStrep tumors. Data are mean ± sem. Two-sided Student’s t-test. h) RT-PCR analysis of diverse tissues in KP/K^bStrep mice with and without intratracheal Adeno-SPC-Cre administration. Expression of the Strep tagged K^b allele is only present in the lung after Cre induction. i) Immunoblot depicting isolation of intact MHC-I complexes specifically from KP/K^bStrep tissue as evidenced by co-purification of B2m. j) Comparison of predicted peptide hydrophobicity (GRAVY) versus median peptide retention time in MS analysis for all peptides presented in Figure 1i. Linear regression is shown for peptides identified in Normal-Ab, Tumor-Ab, and Tumor-Strep datasets indicating no detectable differences in the biochemical features of identified peptides across methods. k) (left) Non-metric multidimensional scaling (nmds) plots depicting clusters of 8- and 9-mer peptides identified from Normal-Ab, Tumor-Ab, and Tumor-Strep samples. (right) Histogram showing the distribution of unique peptides from antibody (Ab) and Streptactin affinity purification (Strep) methods across peptide clusters identified with nmds analysis (n.s. – not significant, Fisher’s exact test). (l) Distribution of predicted peptide affinity for MHC-I peptides identified in 6 KP/K^bStrep replicates. m) Upset plot depicting the peptide identification overlap between 6 KP/K^bStrep replicates after length and affinity filtering. >77% of all identified peptides were found in at least 2/6 replicates.

Extended Data Figure 4. (related to Figure 1). Biochemical comparison of WT and Strep tagged H2-K^b.

a) Isolation strategy for KP and KP/K^bStrep cell lines. b) Histograms depicting fluorescence staining intensity of H2-K^b across KP and KP/K^bStrep cell lines at baseline (gray) or following treatment with IFN-γ (red). c) Median fluorescence intensity quantification of H2-K^b staining across KP or KP/K^bStrep cell lines at baseline (gray) or following treatment with IFN-γ (red). d) Relative H2-K^b staining intensity on KP (gray) or KP/K^bStrep (red) cell lines following incubation with brefeldin A (BFA, left) or acute stripping with acid (300 mM glycine, pH 3.0, right) for the indicated times. e) Representative immunoblot depicting relative amounts of H2-K^b immunoprecipitation with antibody (Y3-Ab), Streptactin, or antibody following Streptactin (Y3-F.T.). f) Densitometric quantification of immunoprecipitated B2m intensity following antibody (Y3), streptactin, or antibody following streptactin (Y3 after Strep) purification schemes. g) Representative immunoblot depicting Strep-tagged H2-K^b expression in KP or KP/K^bStrep cell lines following incubation with the aminopeptidase inhibitors ERAP1-in-1 or Bestatin. h) Experimental schematic for comparison of KP and KP/K^bStrep immunopeptidomes from cultured cells using a quantitative, tandem mass tag (TMT) mass spectrometry strategy. i) Immunoblot depicting the abundance of immunoprecipitated MHC-I from samples described in h). j) Quantitative abundance comparison between all peptides identified in KP and KP/K^bStrep samples. k) Illustration of the lentiviral constructs used for stable expression of SIINFEKL in KP and KP/K^bStrep cell lines. l) Flow cytometric analysis of SIINKFEKL-H2-K^b complex surface expression using 25-D1.16 antibody. m) Experimental schematic used to evaluate specific T cell killing mediated by OT-I TCR transgenic T cells. n) Representative flow cytometry histograms depicting raw data used for calculating % specific lysis. o) Quantification of OT-I T cell killing in KP and KP/K^bStrep cells.

Extended Data Figure 5. (Related to Figure 2). Analysis of the LUAD immunopeptidome throughout tumor evolution.

a) UMAP embedding of clusters used for signature expression analysis in Figure 2a. b) Gene expression profiles of cell type marker genes indicating robust clustering of known cell types in the healthy lung. c) Representative H&E stains for healthy lung (AT2), Early-, Mid- and Late-stage tumor samples. d) Peptide motifs of 8- and 9-mer peptides identified from healthy AT2 cells. e) Length and affinity characteristics of peptides identified in AT2, Early-, Mid-, and Late-stage tumors. f) Venn diagram comparison of peptides identified in bulk, healthy lung versus those identified specifically presented by normal AT2 cells. Pathways enriched by gene ontology depicted on bottom. g) Comparative analysis of gene signatures derived from peptides detected on AT2 cells versus bulk lung applied to the Tabula Muris data. Volcano plot shows the strong enrichment for an AT2 phenotype in the AT2 immunopeptidome versus bulk lung immunopeptidome. h) Comparison of AT2 and Early-, Mid-, and Late-Stage immunopeptidomes. i) Quantification of percent overlap from h). j) UMAP embedding of clusters from reanalyzing scRNAseq data from Marjonovic et. al. and used for analysis of Figure 2j. k) Loess regression analysis across all cells scored for AT2, Early, Mid, and Late peptide signatures versus the AT2, GI-Epi, and Mixed transcriptional modules. l) Signature distribution for AT2-, Early-, Mid-, and Late-Stage peptide signatures across all KP scRNAseq clusters. m) Comparison of the immunopeptidome in control tumors (gray), tumors chronically depleted of CD8 T-cells (light teal), and tumors acutely depleted of CD8 T-cells (dark teal). n) Comparison of the immunopeptidome from control tumors (gray) or tumors treated with agonistic-CD40 antibody and Flt3-L (orange). CD8 depletion experiments and CD40/Flt3L experiments were carried out independently and analyzed on different MS runs separated by ~2.5 months.

Extended Data Figure 6. (Related to Figure 3). Identification of differentially translated genes in LUAD versus AT2 cells.

a) Representative images of normal AT2 and tumor organoid cultures in 3D organotypic culture or after transient adaptation to 2D monolayer culture prior to RiboSeq and RNAseq processing. b) Denaturing PAGE gels of RNA purified from RiboLace purification. Excised bands used for RiboSeq are indicated on the right and were selected for RNA that was ~ 30 bp in length. c) Localization of RPF alignments within transcripts depicting enrichment for annotated coding sequences (CDS). d) Normalized metagene density profiles of reads from normal and tumor cells at translation initiation (left) and termination (right). Both normal and tumor metaprofiles exhibit 3 nucleotide periodicity, indicative of active translation. e) Volcano plot depicting differential mRNA expression in tumor versus normal AT2 cells. f) Volcano plot depicting differential RPF abundance in tumor versus normal AT2 cells. g) Volcano plot depicting differential translation efficiency in tumor versus normal AT2 cells.

Extended Data Figure 7. (Related to Figure 3). Transcriptomic and proteomic data associated with LUAD-unique peptides.

a) mRNA expression of genes encoding for Normal (bulk-lung), AT2, All-LUAD, or LUAD-unique peptides across normal AT2 cells, early-, mid-, and late-stage sorted tumor cells. (Adapted from Chuang et. al.). b) Subcellular compartment distribution of source proteins for peptides found in bulk Normal Lung, AT2 cells, all tumor peptides, and LUAD-unique peptides. c) Distribution of Protein length, thermal stability, or protein half-life for source proteins of peptides found in Normal Tissue, All Tumor peptides, or LUAD-unique peptides. d) StringDb analysis of source proteins for LUAD-unique peptides indicated in Figure 3a. Clusters of enriched protein families are depicted. e) Gene ontology analysis of LUAD-unique peptides from KEGG and Reactome databases. f) Expression of the LUAD-unique signature across all cells in the KP scRNAseq dataset (Marjonovic et. al.) g) Expression of the LUAD-unique signature across tumor progression in KP scRNAseq. h) Expression of LUAD-unique peptide signature across clusters in KP scRNAseq. i) Expression of individual genes encoding LUAD-unique peptides across KP timepoints. j) Correlation of the LUAD-unique peptide signature to all genes detected in scRNAseq. Genes related to antigen presentation are highlighted in red and genes related to metastasis are highlighted in blue.

Extended Data Figure 8. (Related to Figure 3). Modulation of protein folding through Hsp90 inhibition alters the immunopeptidome in vivo.

a) Experimental schematic of KP/K^bStrep tumor treatment with either vehicle control or 0.5 mg/kg/day NVP-HSP990 prior to tumor specific MHC-I isolation. b) Immunoblot analysis of purified MHC-I from Vehicle (Veh) and Hsp90 inhibitor (Hsp90i) treated tumor samples or KP control tumors. c) Length and affinity distribution of peptides found in Veh (grey) and Hsp90i (blue) treated samples. d) Venn diagram of peptides found in 12-week control tumors or Hsp90i treated samples. e) Number of Hsp90 clients giving rise to peptide in either Veh (grey) or Hsp90i (blue) samples. f) Distribution of peptides identified in Veh or Hsp90i treated samples across subcellular compartments. g) Gene ontology analysis of source proteins for peptides that were only found in Hsp90i treated samples ranked according to FDR enrichment significance. h) Comparing the rank ordered abundance of all common peptides between Hsp90i treatment and control. i) Density plots of raw peptide abundance for non-clients, synthesis clients or constitutive clients in Veh (top) or Hsp90i (bottom) treated samples. j) Rank ordered abundance of peptides derived from Non-clients (No, gray), synthesis clients (Synth, light purple) or constitutive clients (Const, dark purple) in Veh and Hsp90i treated samples. P calculated with the Kologorov-Smirnov Test. k) RNA abundance, translation rate, and melting temperatures across non-clients (No), synthesis clients (Synth) and constitutive clients (Const) that are source proteins for MHC-I presentation. P calculated with Mann-Whitney U Test. l) Comparison of peptides unique to Flt3L/aCD40 treatment and those unique to Hsp90i.

Extended Data Figure 9. (Related to Figure 4). Expression and presentation of putative tumor specific and tumor associated antigens.

a) Correlelogram and heatmap depicting transcript abundance (transcripts per million, TPM) of putative TSA genes across mouse tissues. b) Correlelogram and heatmap depicting TPM abundance of putative TAA genes across mouse tissues. c) Experimental schematic showing the derivation of samples for 2D immunopeptidomics. d) Venn diagram depicting the relationship between peptides identified by KP tumors in vivo and those identified in vitro. e) Boxplot showing the predicted affinity distributions of peptides isolated in vivo and in vitro. f) Distribution of source protein subcellular compartments for peptides identified in vivo (gray) and in vitro (green). P calculated with Fisher’s Exact test with Monte Carlo simulation. g) Volcano plot indicating differentially expressed genes between EPCAM+ cells from embryonic day 16.5 and post-natal day 28 mouse lung (Adapted from Lung Map Project). Data analyzed and P calculated with DEseq2. All genes detected are shown in grey and genes encoding for LUAD-unique peptides are indicated with black dots. h) Flow cytometry analysis of tumor-bearing lung tissue from naïve and vaccinated mice stained with control pMHC-I tetramer (SIINFEKL) or TAA tetramer (SVAHFINL). i) Peptides identified in A549 cells (Javitt et. al.) with and without treatment of IFN-γ/TNFα. Peptides derived from source proteins homologous to those using in the pooled vaccine are indicated in red. j) Heatmap depicting expression of the human homologs of putative TSAs and TAAs from this study and whether or not peptides derived from those genes were found to be presented on A549 cells from Javitt et. al. k) Heatmap depicting the RNA Expression of homologs of potential TSA and TAA genes as found in Figure 4a across all individual human tissues and 33 cancer types within TCGA.

Extended Data Figure 10. (Related to Figure 4). Mass spectrometry validation of immunogenic epitopes with synthetic peptides.

a) Mass spectrometry comparison of spectra from endogenously identified VNVYFALL peptide (Slc26a4) and a synthetic standard. b) Mass spectrometry comparison of spectra from endogenously identified SVAHFINL peptide (Prdm15) and a synthetic standard. c) Mass spectrometry comparison of spectra from endogenously identified AVLLYEKL peptide (Ift74) and a synthetic standard. In the left panel, y-, b-, and a-ions are colored in bold. In the right panel, common peaks are drawn in darker color.

Figure 1.. Design and Validation of the KP/K^bStrep Mouse Model.

a) A Cre invertible exon encoding for the StrepTagII epitope was inserted into intron 1 of the H2-K1 gene (top). Cre-recombination induces incorporation of the StrepTagII onto the amino-terminus of MHC-I (bottom). b) Schematic illustration depicting how Cre activation of K^bStrep enables tumor specific isolation of MHC-I in autochthonous tumors. c) Multiplex immunofluorescence of a representative KP/ K^bStrep lung 8 weeks post tumor initiation. White box indicates the zoomed region on the right. d) Experimental schematic depicting the different in vivo sample types to be compared in downstream analyses. e) Length distribution of peptides isolated from healthy lung (Normal-Ab), KP/K^bStrep tumor bearing lung with anti-H2-K^b antibody (Tumor-Ab), KP/K^bStrep tumor bearing lung with Streptactin affinity purification (Tumor-Strep), or “wild type” KP tumor bearing lung with Streptactin affinity purification (Tumor-WT, negative control). f) Number of unique peptides identified in each sample type after filtering for length (8-11 amino acids) and NetMHCPan predicted affinity (<1000 nM). g) Peptide motifs of 8- and 9-mers isolated from Normal-Ab, Tumor-Ab, and Tumor-Strep samples. i) Venn diagram showing the relationship between peptides identified in Healthy Lung, tumor-bearing lung with antibody (Tumor-Ab), and tumor-specific MHC-I purification with Streptactin (Tumor-Strep). Peptides unique to tumors are outlined in red.

Figure 2.. The LUAD immunopeptidome is dynamic and heterogenous throughout tumor evolution.

a) Comparison of the relative expression of gene signatures derived from Normal (gray), Tumor-Ab (pink), or Tumor-Strep (red) peptides across all cell types detected by scRNA-seq from healthy lung tissue (Adapted from Tabula Muris). b) Volcano plots for pairwise comparisons of peptide signatures for each cell type are shown on the right. P-value calculated with two-sided Student’s t-test with Bonferroni adjustment. c) Experimental schematic showing the types of samples for comparison of the immunopeptidome from the AT2 cell-of-origin through tumor progression. d) Overlap between peptides identified in Early, Mid, and Late-stage tumors versus bulk lung tissue and AT2 cells. e) Quantification of the percent overlap from Early (n=5), Mid (n=3), and Late (n=6) stage tumor peptides with Normal peptides (bulk lung + AT2). P calculated with Two-sided Student’s t-test. f) Relative expression of signatures derived from peptides identified in Early (E), Mid (M), and Late (L) stage tumors in alveolar type 2 (AT2), Club/BASC, and Basal cells in the healthy lung. P calculated with Mann-Whitney U Test. g) UMAP embedding of re-analyzed scRNAseq data from Marjonovic et. al. showing all AT2 (T 0w) and KP cells (KP 2-30w) throughout tumor progression (left) and expression of the AT2, Early-, Mid-, and Late-stage peptide signatures (right). h) Pearson correlation of AT2, Early, Mid, and Late signatures versus all gene modules described in Marjonovic et. al. i) Gene set enrichment analysis (GSEA) of genes ranked according to their correlation to the Late tumor peptide signature. Gene sets positively correlated to the Late signature are shown in red, and negatively correlated are shown in blue. j) Relative expression of the Late signature, Nxk2-1, Hmga2, or H2-K1 across clusters found in KP scRNAseq data. k) Multiplexed immunofluorescence depicting tumor-specific MHC-I expression (Strep, magenta) of a single late-stage KP/K^bStrep tumor (top) or two adjacent tumors (bottom).

Figure 3.. Transcription and translation of LUAD-unique peptides

a) Venn diagram depicting the identification of LUAD-unique MHC-I Peptides. b) Heatmap showing the relative mRNA expression (red, white, blue), mean mRNA expression (gold), and predicted affinity (green) of genes encoding for LUAD-Unique peptides throughout tumor progression compared to normal AT2 cells (adapted from Chuang et. al.). c) Workflow depicting an in silico approach to predicting tumor-specific peptides based on RNA expression and predicted affinity. Histograms show the number of LUAD-unique peptides according to whether they were predicted (grey) or not predicted (red) by RNA/affinity analysis. d) Experiment outline for ribosome sequencing by RiboLace. e) Heatmap showing the relative translation intensity (TPM) for AT2 identity genes and genes associated with KP Tumor Progression. f) Comparison of RNAseq abundance (x-axis) and RiboSeq abundance (y-axis) in tumor organoids versus AT2 organoids. Genes coordinately up or down in both RNAseq & RiboSeq are shown in blue, genes exhibiting differential translational efficiency are shown in green, and genes encoding for LUAD-unique peptides are shown in red. g) Workflow depicting an approach to predict LUAD-unique peptides based on differential translation efficiency in LUAD versus AT2 (TE) and predicted affinity. Histograms show the number of LUAD-unique peptides according to whether they were predicted (grey) or not predicted (red) by differential TE.

Figure 4.. Discovery of novel tumor antigens in LUAD

a) Workflow for identifying putative non-mutated Tumor Specific Antigens (TSA) or Tumor Associated Antigens (TAA). b) Experimental schematic for a pooled peptide vaccine strategy in naïve mice. c) Quantification of interferon-gamma (IFN-γ) ELISPOT data of splenocytes from naïve mice or mice vaccinated with pooled peptides. Each peptide in the pool was used individually to stimulate splenocytes prior to ELISPOT. n=3 mice per group. d) Comparison of in vivo bulk RNAseq, single cell RNAseq and ex vivo RNAseq and RiboSeq for genes encoding immunogenic peptides. Identification of peptides by in vitro and in vivo immunopeptidomics is also indicated. For in vivo peptide identification, the fraction of late-stage tumor samples where the peptide was identified is indicated. e) Pooled vaccine strategy for KP tumor-bearing animals. f) Flow cytometry plots depicting pMHC-I Tetramer staining for a representative TAA (SVAHFINL, Prdm15) in the lung tissue of naïve and vaccinated, tumor-bearing mice. P calculated with one-sided Mann-Whitney test. g) Transcript abundance across healthy mouse tissues for peptides included in the pooled vaccine. Detection of the peptide on AT2 cells or bulk lung tissue is also indicated. h) Model depicting the incongruence between the immunopeptidome derived from in silico prediction methods, in vitro mass spectrometry, and tumor-specific in vivo immunopeptidomics.