Dysregulation of translation and protein synthesis is often observed in cancer, but whether this results in the production of targetable antigens has not been fully explored. Investigating this possibility, Weller, Bartok, and McGinnis et al. used CRISPR-Cas9 to study the effects of tRNA wybutosine (yW)-synthesizing protein 2 (TYW2) – a tRNA transferase that supports reading frame maintenance in ribosomes and shows altered expression in various tumors. Their results were recently published in Cancer Cell.

To begin, the researchers used CRISPR-Cas9 to knock out TYW2 in two melanoma cell lines, A375 and SKMEL30, and generated multiple single-cell-derived lines from each source. Mass spectrometry confirmed that yW-mediated modifications to tRNA^Phe were lost in TYW2 knockout (KO) cells, while a reporter system confirmed an increase in programmed ribosomal frameshifts (PRFs). Bulk RNAseq was then used to measure transcriptional differences between KO and wild-type (WT) cells, and this showed increased expression of tRNA modifiers, translation-related RNA binding proteins, and antigen presentation machinery, as well as decreased expression of aminoacyl-tRNA synthetases. Further, ribosome profiling showed that in KO cells, ribosomes dwelled longer on Phe codons, and the density at Phe codons (the tRNA modification driven by TYW2 affects tRNA^Phe entry to the A site of the ribosome) was relatively enriched compared to other sense codons and ribosome positions. Together these results suggested that specific pausing at Phe codons may make KO cells more susceptible to reading frame errors than WT cells.

Given the increased susceptibility to frameshifts and the increased antigen processing machinery observed in TYW2 KO cells, the researchers interrogated the immunopeptidome and found that these cells expressed numerous out-of-frame HLA-bound peptides resulting from off-frame or non-canonical translation. While some of these were also found in WT cells, most were KO cell-specific. WT cells, on the other hand, expressed very few unique off-frame peptides. Using a custom analysis pipeline (ProxyPhe), the researchers identified 99 off-frame peptides associated with frameshifting at Phe codons, and while 34 of these were specific to KO tumor cell lines, only 4 were specific to WT cells. Further, mass spectrometry analysis of proteasomes isolated from WT and KO A375 cells showed that KO and WT degradation products were distinct, with KO cells showing a higher number of detected products, and a higher number exclusive peptides (compared to WT), including peptides derived from the “ubiquitin-conjugating enzyme activity” family. These results confirmed that loss of TYW2 was associated with changes in the degradome and immunopeptidome, and with endogenous HLA presentation of aberrant peptides.

To determine whether the increased presentation of aberrant peptides could elicit immune responses, 11 previously identified peptides were synthesized and loaded onto mature DCs. When these peptide-loaded DCs were cocultured with PBMCs from healthy donors, 5 out of the 11 peptides were found to induce CD8⁺ T cell activation, as measured by 4-1BB, TNFα, and IFNγ production. In short-term coculture experiments with naive T cells, off-frame peptides were more immunogenic than known mutation-encoded neoantigens.

Next, Weller, Bartok, and McGinnis et al. evaluated whether peptides associated with aberrant translation could act as neoantigens and support antitumor immunity in vivo by knocking out Tyw2 in B2905 murine melanoma cells. Upon implantation into mouse models, WT cells grew aggressively in mice, while growth of Tyw2 KO cells was more limited. Further, overexpression of Tyw2 accelerated tumor growth and reduced ribosomal frameshifts. These differences in tumor growth were not observed in immunodeficient mouse models, nor in CD8⁺ T cell depleted mice, suggesting immune/CD8⁺ T cell-mediated control of KO tumors. Further, when T cells were isolated from naive mice and cocultured with DCs pre-loaded with tumor cell lysates, CD8⁺ T cells showed increased proliferation.

BMDCs loaded with tumor cell lysates also had a distinct MHC-I ligandome repertoire at the canonical peptide level, and a ProxyPhe-peptide search revealed 21 robustly identified aberrant peptides, 11 of which were found to KO-specific, while only 1 was WT-specific. Splenocytes isolated from mice immunized with 10 aberrant peptides showed CD8⁺ T cell reactivity against half of the target peptides upon ex vivo restimulation, and were more reactive against KO cells than those from non-immunized mice.

Looking at the TMEs of KO vs. WT tumors over time, the researchers found that CD8⁺ T increased in KO tumors between days 18 and 20, but not in WT mice. Further, while WT tumors were enriched for memory-like CD8⁺ T cells, KO tumors showed increased exhausted CD8⁺ T cells, LAG3⁺ exhausted CD8⁺ T cells, and Ki67⁺ proliferative cells at day 21, along with increased IFNγ/IFNGR1/2 signaling between CD8⁺ T cells and myeloid cells, and increased NK cell cytotoxicity, suggestive of a more inflammatory and activated TME.

Given the T cell exhaustion signature observed in CD8⁺ T cells in the Tyw2 KO TME, the researchers treated tumor-bearing mice with anti-PD-1 and found that it significantly delayed or reduced growth of KO, but not WT tumors. T cells isolated from spleens or lymph nodes of treated mice at day 27 were reactive towards multiple Tyw2 KO-specific ProxyPhe-identified peptides ex vivo.

To determine whether these results would translate to patients, the researchers evaluated data from TCGA and found that in a cohort of primary melanoma, lower TYW2 expression was associated with improved progression-free survival (PFS), while patients with higher TYW2 expression tended towards worse overall survival. In an anti-PD-1-treated cohort, lower TYW2 expression in on-treatment samples, and to a lesser extent in pre-treatment samples, was predictive of ICB response and associated with better overall survival. These observations were independent of tumor mutation burden (TMB), and predicted patient outcomes in on-treatment samples from a low-TMB cohort. Similar results were found in two other melanoma patient cohorts. Additionally, differential expression analysis of TCGA samples showed that TYW2-low tumors showed increased immune activation, CD8⁺ T cell infiltration, and cytolytic and exhausted T cell activity in tumors, with TYW2 itself emerging as the top-ranked gene for predicting patient outcomes.

Overall, these results show that aberrant translation, which is common in certain cancers, can lead to the production and presentation of antigens that can elicit antitumor immune responses. These findings could aid in predicting patient responses to checkpoint blockade and could potentially lead to the development of new immunotherapies targeting cancer-associated antigens generated through aberrant translation events.

Write-up and image by Lauren Hitchings

Meet the researcher

This week, first author Chen Weller answered our questions.

What was the most surprising finding of this study for you?
The most surprising finding for us was that we were able to detect endogenous – de novo – T cell responses against the aberrant peptides. As readers of the paper will see, we identified these peptides, demonstrated their immunogenic potential (in both humans and mice), and found strong evidence linking immune system function to tumor growth dynamics. However, we were always concerned about whether these aberrant peptides could truly elicit an immune response in vivo. Achieving this data was an incredibly exciting moment for us.

What is the outlook?
We hope to inspire further research efforts to target translation fidelity, a highly attractive source of targetable antigens in cancer. Specifically, the potential of this approach to be effective in low-TMB or MSS-S tumors is of great importance.

Who or what has been a major source of inspiration or motivation for you throughout your career?
As a PhD student in the Samuels lab, I have been fortunate to have two incredible mentors. Osnat Bartok, the lab’s staff scientist, has not only taught me invaluable skills, but also how to think critically, ask the right questions, and pursue knowledge. My PI, Yardena Samuels, has been a constant source of motivation, pushing me toward scientific excellence while giving me the freedom to grow independently. She ensured this project received the attention and resources it needed while being deeply involved in every detail. Her dedication and unwavering support have been truly inspiring, shaping my development as a scientist.