NEW YORK – Immunopeptidomics has benefited in recent years from advances in mass spectrometry and informatics technologies as well an uptick in interest from academia and pharma.
But while the field has proved useful for work in areas like antigen processing and neoantigen discovery that are key to cancer vaccine research, it remains to be seen whether it can claim a primary role in clinical efforts and drug development.
Immunopeptidomics concerns the study of peptides presented by human leukocyte antigen (HLA) molecules. These molecules play a key role in immunity, displaying peptide antigens at the cell surface that generate immune cell responses to various infections or diseases. Identification and manipulation of these antigens is key to research in areas like cancer immunotherapy, where scientists are working to trigger patients' immune systems to fight their cancers by presenting cancer- and patient-specific HLA antigens.
A key challenge to developing cancer vaccines is identifying which peptides are likely to bind to HLA molecules and thereby be presented at the cell surface. Here immunopeptidomics is playing a growing role as researchers are able to collect data on thousands of HLA-peptide pairings. This data can then be used to inform the algorithms that are commonly used to predict cancer-specific peptides, known as tumor neoantigens, that are likely to be displayed by malignant cells and which are therefore promising molecules for inclusion in vaccines.
Researchers are also using mass spec to look for tumor neoantigens directly, analyzing patient tumor samples in search of these molecules with the ultimate goal of guiding the development of personalized vaccines.
Analyzing these peptides is challenging, however. A major challenge stems from the fact that while traditional mass spec-based proteomics experiments use trypsin to digest proteins into peptides, the peptides presented by HLA molecules have been digested by a number of different enzymes. This significantly expands the search space of potential peptides researchers confront when using experimentally generated mass spectra to make peptide identifications.
Identifying tumor neoantigens is particularly difficult. Typically, researchers are working with small amounts of patient sample, making high sensitivity key. Additionally, these neoantigens may not be present in reference databases used for making peptide identifications, complicating the ID process.
Despite these challenges, the field is growing.
"I feel like we've turned the tide for immunopeptidomics," said Anthony Purcell, a professor of biochemistry and molecular biology at Monash University and longtime immunopeptidomics researcher. "Over the last four or five years, we have seen quite an ingress of researchers and a lot of interest from pharma in adopting the technology."
Immunopeptidomics "is a really central focus for all of our cancer vaccines," said Raquel Deering, senior director and head of cancer vaccine discovery and early development at Moderna. "There's no question that Moderna now sees this broadly as a critical technology."
This uptick in interest has driven mass spec vendors and informatics developers to improve their immunopeptidomic offerings, Purcell suggested, noting as well that increasing interest and developments in single-cell proteomic workflows have also pushed immunopeptidomics forward.
Among mass spec firms, Bruker has perhaps most explicitly targeted the immunopeptidomic market. When the company launched its TimsTOF SCP instrument in 2021, it specifically highlighted "immunopeptidomics for neoantigen discovery" as a key application. Torsten Müller, business development manager of proteomics at Bruker, said that the company's recently released TimsTOF Ultra instrument provides further improvements in performance for immunopeptidomic work.
"One of the reasons why I think [immunopeptidomics] is really a growing field right now is that mass spectrometers have become much more sensitive and are actually able to measure these peptides reliably, reproducibly, and [with enough depth] to actually generate the data that people can use to generate these therapies," Müller said.
"We're getting a lot more bang for our buck in terms of the data we generate," Purcell said. "We are getting a lot more depth in terms of what we can see, and we are starting to see a lot more actionable insights."
Deering said that a primary use of immunopeptidomic data within Moderna is for informing the algorithms the company uses to predict what neoantigens might be present in a patient's cancer and might warrant inclusion in a personalized vaccine.
In September, Moderna established a collaboration with German immunotherapy firm Immatics in part to gain access to data from that company's immunopeptidomic discovery platform. Moderna is also building its own internal immunopeptidomics capabilities, Deering said. The company is currently looking to hire a director to lead an immunopeptidomics group along with other employees to support that group.
Currently, most tumor neoantigen data comes from RNA and DNA, with far less data coming from immunopeptidomics, Deering said. "Immunopeptidomics is part of Moderna's process to get a sense of not only what the mutations are, but what actually is presented to the immune system, and how the immune system, at least at a T-cell level, responds," he noted.
Deering said that despite academic work showing that mass spec can detect tumor neoantigens directly from patient samples, Moderna does not view that technology as ready for this kind of work.
"I think the entire field acknowledges that it would be really hard, if not impossible, to select neoantigens for your vaccine using personalized immunopeptidomics," she said.
Largely, this is a matter of sensitivity. As is the case with proteomics generally, the sensitivity of immunopeptidomics is limited by the fact that there is no amplification technique comparable to PCR for nucleic acids.
"Immunopeptidomics is not a sensitive detection method," Deering said. "If you compare DNA or RNA sequencing data on the same sample to immunopeptidomics using the most sensitive Bruker machines, there is a small fraction of neoantigens that are detected. And the reality is that in most samples, you detect none of the neoantigens that end up going into vaccines."
Purcell likewise said that identifying neoantigens in individual patients remains a challenge.
"If you work off a genomic template, you might predict thousands, and we might find tens," he said. "The jury is still out whether that is because we still have insufficient sensitivity, whether we have issues with our searching strategies … or whether it is just that they are not on the surface of the tumor, which is, of course, the elephant in the room in terms of the conversations you have with your clinical colleagues."
Purcell said, though, that the field has made significant gains in sensitivity.
"We're able to now not just work with cancer cell lines, but now the technology is able to get at certainly resected tumors, but even with small biopsies we are able to generate informative data," he said.
Purcell said that immunopeptidomics has benefited from the explosion of interest and method development in single-cell proteomics, where researchers have devised various sample preparation and bioinformatic approaches to get the most out of extremely small samples. He noted that his facility recently purchased a CellenOne cell isolation and handling system from Scienion, which has become a common part of single-cell proteomics workflows.
"We'll steal every trick that the single-cell guys have got to improve our yields," he said.
Purcell said he believes new instrumentation will also push the field forward. His lab primarily uses the Bruker TimsTOF Pro 2 system for its immunopeptidomics work but is currently testing samples on Thermo Fisher Scientific's new Orbitrap Astral instrument. He said he has also seen immunopeptidomic data generated on the TimsTOF Ultra that "looks very impressive."
He said he expects the major advantage of the new instruments will be the ability to work with much smaller samples.
"If you can get a decent core biopsy, you have quite a bit of material," he said. "If you get a fine needle aspirate, that can be challenging, and I think these new instruments are really going to be useful for these quite limited biopsy specimens."
Deering also said she expects that over time refinements in immunopeptidomic technologies will improve performance.
In the meantime, she said that immunopeptidomics is currently sensitive enough to be useful for target discovery in Moderna's "off-the-shelf" cancer vaccine program, which aims to identify shared antigen targets the company can put into "pre-made, ready-to-go vaccines" intended for broad patient populations.
"There we have large tumors, tumors from more patients, and more time, so it is much more feasible to do target discovery with immunopeptidomics," she said.
Advances on the informatics front are also boosting the power of immunopeptidomics.
Last year, for instance, Bruker introduced a new de novo peptide sequencing software package, Paser Novor, for use on the TimsTOF platform. The company, which developed the software in collaboration with antibody sequencing firm Rapid Novor, has targeted it particularly to immunopeptidomics researchers.
De novo peptide sequencing allows researchers to identify the amino acids that compose a detected peptide simply by analyzing the experimental spectra and without comparison to a reference database, making it useful for identifying peptides like tumor neoantigens that may not be present in such databases.
The technique is still something of a niche method within proteomics, but advances like AI-based approaches to predicting fragmentation patterns and intensities for particular peptides and amino acids have improved the confidence with which algorithms are able to identify peptide sequences. The ability to train these algorithms on increasingly large datasets has also helped to refine them.
Purcell's lab often uses a hybrid approach in its immunopeptidomic work, generating reference databases specific to cancer cells or samples they are studying but also using de novo sequencing to identify potential peptides of interest not present in the reference database.
Müller said Bruker plans this year to make it possible for researchers to use the collisional cross section data generated by the TimsTOF's ion mobility system to increase the confidence of peptide IDs in immunopeptidomics experiments.
Purcell said he and his colleagues are currently involved in several clinical projects leveraging immunopeptidomics, including a proof-of-concept study with physicians at the University of West Australia looking for neoantigen targets in mesothelioma for inclusion in vaccines, as well as work with clinicians at Monash Medical Center analyzing patient material with the goal of aiding vaccine development for cancers including pancreatic.
"More and more, we're having interactions with both industry and clinicians," he noted.
That said, concrete clinical results will be key to pushing the technology forward, Purcell suggested.
"It's having that exemplar of where you worked with someone locally and were able to direct successful therapy in a patient," he said. "There's nothing like having a successful demonstration of the technology."