AUTHORS
Marijn Bakker, Luis O Correa-Medero, PhD2, Marwan Kwok, MD, PhD, MRCP, FRCPath3, Emma Duggan4, Marta Wilbrink5, Isabelle Lippincott6, Griffin M Lawrence7, Karl R. Clauser8, Kasidet Manakongtreecheep, PhD9, Claudia Ctortecka, PhD10, Cleo Forman11, William Freed-Pastor, MD, PhD12, Jenn G Abelin, PhD13, Catherine J. Wu, MD14
AFFILIATIONS
1 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States, Dana Farber Cancer Institute, Boston, Massachusetts, United States, University of Amsterdam, Amsterdam, Noord-Holland, Netherlands
2 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
3 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States, Dana Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States, Queen Elizabeth Hospital Birmingham, Birmingham, United Kingdom, University of Birmingham, Birmingham, United Kingdom
4 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
5 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
6 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
7 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
8 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
9 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
10 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States
11 Dana Farber Cancer Institute, Boston, Massachusetts, United States
12 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States, Dana Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States, Koch Institute at MIT, Cambridge, Massachusetts, United States
13 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States, Dana Farber Cancer Institute, Boston, Massachusetts, United States
14 Broad Institute of MIT and Harvard, Cambridge, Massachusetts, United States, Dana Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
ABSTRACT
Current T-cell antigen targeting immunotherapies favor mutation-derived canonical neoantigens for their high tumor specificity and immunogenicity, but their utility is limited in hematologic malignancies such as chronic lymphocytic leukemia (CLL), which characteristically have low tumor mutational burden (TMB). In these low-TMB settings, non-canonical antigen classes, HLA class I (HLA-I) peptides derived from novel unannotated open reading frames (nuORFs) or endogenous retroviruses (ERVs), offer promising shared, cancer-restricted targets capable of eliciting potent cytotoxic T-cell responses.
We hypothesized that the targetable antigen landscape of CLL could be broadened through systematic evaluation of the HLA ligandome to encompass noncanonical targets. Mass spectrometry (MS)-based immunopeptidomics is a powerful approach to identify and quantify HLA-presented peptides but has limitations in the sensitivity of detection. We recently reported Pepyrus, a high-throughput platform that substantially increases MS sensitivity by integrating reference spectral libraries within highly scalable data-independent acquisition (DIA)-MS workflows. To generate high-quality reference spectra, user-defined peptide libraries (range 200-20,000 peptides) are expressed in bacteria, purified using a SUMO tag, released by a yeast protease, and acquired by data-dependent acquisition (DDA)-MS. Here, we leverage this innovative platform to comprehensively characterize the CLL antigenic landscape of 10 samples collected from 7 patients with IGHV-unmutated CLL (8 pretreatment, 2 at relapse).
To investigate the HLA-I ligandome of CLL, Pepyrus-enabled DIA was employed. Off-the-shelf Pepyrus peptide libraries encompassing both canonical and non-canonical antigenic targets were constructed based on analysis of previously published CLL-specific and pan-cancer datasets from >40 cancer types. We built libraries comprised of: 10,390 tumor-associated antigens (TAA) derived from 49 proteins (Kowalewski, PNAS, 2015); 13,831 cancer-testis antigen (CTA) peptides derived from 92 genes with testis-specific expression (Xinpei, Cell Press, 2021); 43,451 ERV-derived peptides from 62 ERV loci enriched in CLL (La Ferlita, PNAS, 2023); and 80,368 nuORF peptides derived from 7,256 Ribo-seq confirmed ORFs previously reported in CLL and other cancer types (Ouspenskaia, Nat Biotechnol, 2022). Only peptides predicted to bind HLA-I alleles frequently expressed in CLL patients were retained in the final Pepyrus peptide libraries.
In total, we identified 2346 HLA-I peptides, including 1006 nuORFs, 619 TAAs, 412 CTAs, and 309 ERVs, in 10 CLL samples (covering 13 HLA-I alleles). Of these peptides, 59.2% (1388 peptides: 547 nuORF-, 206 TAA-, 328 CTA-, and 307 ERV-derived) were missed by conventional MS-based immunopeptidomics pipelines, demonstrating the added value of user-defined spectral libraries for antigen discovery. To confirm the CLL specificity of the identified HLA-I peptides, we cross-referenced all Pepyrus-identified peptides with healthy tissue immunopeptidomes derived from 23 diverse tissue types including lymph node, thyroid, and spleen. Their CLL-restricted nature was evident from their absence in healthy tissues, with 76.6% of the identified peptides (range 70–82% across different antigen classes) undetectable in normal cells with our high sensitivity platform. Across all 10 samples, we found a shared subset of HLA-I antigens present in ≥2 CLL patients and absent from healthy tissue, comprising 100 nuORF-, 84 TAA-, 41 CTA-, and 20 ERV-derived antigens. These antigen candidates represent potential off-the-shelf immunotherapeutic targets.
Together, these findings demonstrate that Pepyrus enables discovery of novel, cancer-restricted HLA-I antigens in CLL. These off-the-shelf spectral libraries are readily transferable to additional CLL samples and other cancer specimens with similar HLA coverage. Beyond off-the-shelf libraries, we have optimized our workflow to generate spectral libraries for personalized, non-canonical antigens arising from aberrant mRNA splicing, immunoglobulin somatic hypermutation, and structural variants. Immunogenicity studies of candidate antigens are ongoing.
DISCLOSURES
Marijn Bakker: No relevant conflicts of interest to declare.