IFNγ Modulates the Immunopeptidome of Triple Negative Breast Cancer Cells by Enhancing and Diversifying Antigen Processing and Presentation

Abstract

Peptide vaccination remains a viable approach to induce T-cell mediated killing of tumors. To identify potential T-cell targets for Triple-Negative Breast Cancer (TNBC) vaccination, we examined the effect of the pro-inflammatory cytokine interferon-γ (IFNγ) on the transcriptome, proteome, and immunopeptidome of the TNBC cell line MDA-MB-231. Using high resolution mass spectrometry, we identified a total of 84,131 peptides from 9,647 source proteins presented by human leukocyte antigen (HLA)-I and HLA-II alleles. Treatment with IFNγ resulted in a remarkable remolding of the immunopeptidome, with only a 34% overlap between untreated and treated cells across the HLA-I immunopeptidome, and expression of HLA-II only detected on treated cells. IFNγ increased the overall number, diversity, and abundance of peptides contained within the immunopeptidome, as well increasing the coverage of individual source antigens. The suite of peptides displayed under conditions of IFNγ treatment included many known tumor associated antigens, with the HLA-II repertoire sampling 17 breast cancer associated antigens absent from those sampled by HLA-I molecules. Quantitative analysis of the transcriptome (10,248 transcripts) and proteome (6,783 proteins) of these cells revealed 229 common proteins and transcripts that were differentially expressed. Most of these represented downstream targets of IFNγ signaling including components of the antigen processing machinery such as tapasin and HLA molecules. However, these changes in protein expression did not explain the dramatic modulation of the immunopeptidome following IFNγ treatment. These results demonstrate the high degree of plasticity in the immunopeptidome of TNBC cells following cytokine stimulation and provide evidence that under pro-inflammatory conditions a greater variety of potential HLA-I and HLA-II vaccine targets are unveiled to the immune system. This has important implications for the development of personalized cancer vaccination strategies.

Keywords: cytokine stimulation; human leukocyte antigen; immunopeptidomics; mass spectrometry; proteomics; transcriptomics; triple negative breast cancer.

Figure 1

IFNγ modulates HLA expression and the immunopeptidome of MDA-MB-231 cells. (A) Flow cytometry staining of HLA class I and class II expression following IFNγ stimulation. Median fluorescence intensity (MFI) is shown across three biological replicates (mean ± SD). (B) Immunopeptidome profiling of MDA-MB-231 cells in the presence or absence of IFNγ, showing number of peptides for HLA-I and HLA-II. (C) Length analysis of unique HLA-I peptides derived from MDA-MB-231 cells with and without IFNγ treatment. (D) NetMHCpan 4.0 HLA-A, -B, and -C binding prediction of each unique 8–12 mer peptide from IFNγ treated and untreated replicates. Statistical analysis was performed by using a two-way ANOVA and significance is denoted by ****(P < 0.001). All data are from three biological replicates per condition.

Figure 2

IFNγ causes considerable remolding of the source protein immunopeptidome landscape of TNBC cells. (A) Source proteins of HLA-I peptides identified across three biological replicates of IFNγ treated (mean of 6,897 proteins) and untreated samples (mean of 5,256 proteins). (B) Comparative analysis of source protein overlap between untreated and IFNγ treated cells. (C) HLA-I source peptide log₂ intensity calculated from label free quantification from common and unique HLA-I source proteins. (D) Histogram of number of HLA-I peptides per source protein under untreated or IFNγ treated conditions. All data are from three biological replicates with mean ± SD and statistical analysis was performed by using a two-way ANOVA and significance is highlighted by ****(p < 0.001). * P < 0.1, ns, not significant.

Figure 3

IFNγ leads to increased representation of cancer-associated antigens across the HLA-I and -II immunopeptidome. (A) Pie chart showing the total number of HLA-I peptides identified from 300 cancer-associated antigens (derived from the T antigen, CTdatabase, and breast cancer specific antigens from Uniprot), either unique to untreated, unique to IFNγ treated, or common to both conditions. (B) Log₂ fold-change in the number of HLA-I peptides per protein detected following treatment with IFNγ. The 245 proteins subset are cancer-associated antigens represented across the HLA-I immunopeptidome under both conditions. (C) Overlap of cancer-associated source proteins across the HLA-I (untreated and IFNγ treated) and HLA-II (IFNγ treated) immunopeptidomes. (D) The total number of peptides from the 115 commonly identified cancer-associated antigens within the HLA-I -II immunopeptidomes, under untreated or IFNγ treated conditions. (E) A refined subset of specific or associated breast cancer antigens (41 in total), showing the overall distribution of HLA-I peptides identified across these source proteins. Proteins are ordered based on the conditions they were detected under.

Figure 4

IFNγ treatment predominantly induces components of the antigen processing and presentation machinery of MDA-MB-231 cells. (A) The number of transcripts (upper) and proteins (lower) identified from untreated and IFNγ treated cells. Data are from three biological replicates. *p = 0112 by unpaired t-test. (B) Correlation between the Log₂ fold change (IFNγ treated compared to untreated) transcripts and proteins. Significant fold changes at the protein (red), transcript (blue), or protein and transcript (green) level are indicated, with the gene symbols of those with the highest fold-changes indicated. Person correlation coefficient of 0.275, p value <0.0001. (C) The log₂ fold change of transcriptome and proteome changes within key components of the antigen processing and presentation machinery following IFNγ treatment.

Figure 5

IFNγ-mediated modulation of the immunopeptidome is only weakly reflected by changes in the transcriptome and proteome. Correlation regression plot between the log₂ fold change of the number of HLA-I peptides per protein with the associated log₂ fold change of source transcript (A) and source protein (B) following IFNγ treatment. (C) Distribution of source transcript and protein log₂ fold change, binned into categories of HLA-I peptide per protein presentation following IFNγ treatment [decreased (<0.5-fold); unchanged (0.5-fold to 1.5-fold); increased (>1.5-fold)].