Journal News

Finding neoantigens faster — advances in the study of the immunopeptidome

Laurel Oldach
February 21, 2020

The immune system may seem quiescent until an infection prods it to leap into action. In fact, it maintains active surveillance to distinguish what belongs in the body from what does not. As part of this surveillance, T cells prowl through tissues, seeking signs of something amiss. Special constitutively active mechanisms report to T cells on cell contents that otherwise would be hidden.

David Gfeller leads the computational cancer biology group at the University of Lausanne in Switzerland, where his lab works to characterize immune infiltrations at the cell level. “In order to detect cells that are infected or malignant, something intracellular has to be presented at the cell surface,” Gfeller said.

That presentation is the job of human leukocyte antigens, or HLAs, a subset of which display peptides pulled from chopped-up intracellular proteins. The repertoire of peptides they present, which varies according to HLA alleles, protein expression patterns across tissues and other factors, is called the immunopeptidome; it determines what antigens T cells can recognize and respond to.

selvanegra/iStock
An HLA protein (orange) is shown presenting a T cell receptor (violet) with a peptide antigen (green). Researchers are working to find out more about how peptides are selected for presentation.

Understanding how peptides are selected to become part of the immunopeptidome is key to harnessing adaptive immunity. For example, an effective vaccine must use a peptide that HLA proteins will take up and display. When antigen presentation goes amiss — if T cells mistakenly recognize ordinary components of the cell as dangerous invaders or overlook new mutant antigens — diseases like autoimmunity and cancer can result.

Two papers in Molecular & Cellular Proteomics highlight recent advances in the study of the immunopeptidome.

Excitement over immunopeptidomics

The first tumor-specific HLA-binding peptide was identified in 1997. But immunopeptidomics really took off in 2014, when researchers discovered that cancer treatments to boost T cell activity, overcoming immunosuppression induced by tumors, depend on T cell recognition of tumor-specific antigens. Predicting the composition of the immunopeptidome and how it differs between healthy and cancerous tissue became part of a medical frontier with enormous potential: cancer immunotherapy.

Such is the history reviewed by Juan Antonio Vizcaíno, a proteomics team leader at the European Bioinformatics Institute, and colleagues in a perspective in MCP. The authors summarize technical advances in immmunopeptidomics and findings that have linked HLA alleles with human diseases. They also look ahead to a time when immunopeptidome-wide association studies may make it possible to predict an individual’s susceptibility to autoimmunity, infection or cancer — and they discuss the possible challenges in the way.

Immunopeptidomics experiments remain technically challenging. Several fieldwide initiatives aim to pool data on the peptides presented by HLA proteins in various tissues and disease states, among them the Immune Epitope Database and the Human Immuno-Peptidome Project. Based on this hard-won knowledge about the binding preferences of HLA proteins, researchers have generated algorithms to predict which peptides HLA proteins may display, enabling computational study of the immunopeptidome. But these algorithms still fall short in predicting display of post-translationally modified peptides.

Phosphorylated immunopeptides

In addition to genetic changes, many cancers undergo major changes to signaling pathways — for example, 46% of cancer samples in a recent study carried alterations in the MAP kinase pathway. Quantifying how those signaling changes might be reflected at the cell surface is a matter of conjecture.

Michal Bassani–Sternberg, an investigator at the Ludwig Institute for Cancer Research in Lausanne, studies cancer-specific HLA ligands. “We cannot yet extract or purify phosphorylated HLA-binding peptides to a really good depth of coverage as opposed to the depth of unmodified peptides,” she said.

Until recently, this technical limitation hampered efforts to understand whether unusual phosphopeptides might be unique tumor markers in some cases of cancer.

Now, doctoral student Marthe Solleder and collaborators in Bassani–Sternberg’s and Gfeller’s labs have come up with a new bioinformatics tool for predicting phosphorylated HLA ligand peptides. They report the work in MCP.

“We collected a huge data set of peptides displayed at the cell surface that were measured by mass spectrometry,” Bassani–Sternberg said, “and we searched the spectra for modified peptides that contain a phosphate group.”

Working mostly with previously collected spectra from immunopeptidomics experiments, the team reanalyzed immunopeptidome data, uncovering phosphorylated HLA-binding peptides that previously had been overlooked. This let them make inferences about HLA proteins’ phosphopeptide binding preferences.

“For example, of the three types of HLA class I molecules, one called HLA-C is especially apt to bind to phosphorylated peptides,” Solleder said.

The algorithm may be useful for expanding the known range of targets displayed on cancer cells. Bassani–Sternberg said, “If we know where the mutations are in the genome of a patient, we can apply these prediction algorithms to predict which of the mutations is likely to be presented as a ligand on the HLAs of the patient. Similarly, we can now predict which of the known phosphorylated sites the proteins are likely to be presented as HLA ligands.”

The researchers have shared all their data and computational tools. They hope the extra information about phosphorylated peptides may help others find novel antigens. In the long term, such insights may help guide tumor-targeting treatments.

To learn more about how antigen presentation works, watch this illustrated tutorial.

 

Laurel Oldach

Laurel Oldach is a science writer for the ASBMB.

Join the ASBMB Today mailing list

Sign up to get updates on articles, interviews and events.

Latest in Science

Science highlights or most popular articles

Life's little oscillations
News

Life's little oscillations

April 05, 2020

Concentrations of molecules rise and fall, genes switch on and off, and circadian clocks keep time. These biochemical fluctuations are crucial for myriad biological needs, and cells couldn't function without them.

A small trial finds that hydroxychloroquine is not effective for treating coronavirus
News

A small trial finds that hydroxychloroquine is not effective for treating coronavirus

April 05, 2020

Despite the recent approval of this drug for use against COVID-19, questions remain as to the efficacy of this treatment.

Team Cryo ices Mass Spec’s hopes in JBC Methods Madness tourney
Journal News

Team Cryo ices Mass Spec’s hopes in JBC Methods Madness tourney

April 03, 2020

In the inaugural JBC Methods Madness Tournament, cryo­–EM/ET gave structural biologists around the world something to cheer about, beating mass spectrometry 51% to 49% —​ a margin thinner than shaved ice. 

A small army of researchers races to build a coronavirus interactome
News

A small army of researchers races to build a coronavirus interactome

April 01, 2020

Scientists at eight institutions in the U.S. and Europe have used a protein interaction map to identify 69 drugs that might work against SARS-CoV-2, the virus that causes COVID-19.

Parkinson's  Awareness Month
Health Observance

Parkinson's Awareness Month

April 01, 2020

It is the second-most common progressive neurodegenerative disease, occurring in 1% of people over the age of 60 and in 5% of people over 85.

Multiomics meets antimalarials
Journal News

Multiomics meets antimalarials

March 31, 2020

Researchers in Australia use an innovative multiomics approach to analyze a new drug against malaria parasites.