Circulating tumor‐associated and neoantigen‐specific endogenous T cells in children treated for B‐acute lymphoblastic leukemia

Like most pediatric cancers, B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is considered poorly immunogenic due to its low tumor mutational burden (TMB).1 As a result, research has primarily focused on synthetic approaches that actively induce novel immune responses targeting tumor antigens, irrespective of naturally occurring antitumor immunity.2, 3 Notable examples include bi-specific CD19-CD3 antibody and foremost chimeric antigen receptor (CAR) T cell therapy, which has revolutionized treatment and significantly improved relapse-free survival in chemorefractory pediatric BCP-ALL.4, 5 Nevertheless, severe immune-mediated toxicities, on-target off-tumor toxicity like B-cell aplasia, and disease relapse due to antigen escape associated with these treatments underscore the need for safer, more specific, and durable therapies.3 In contrast, natural immunity-based approaches, such as tumor-infiltrating lymphocyte (TIL) therapy and cancer vaccines, have shown remarkable success in adult cancers with high mutational load,6-8 yet remain largely unexplored in pediatric cancers. Interestingly, increasing evidence challenges the notion of TMB as the sole determinant of immunogenicity.9, 10 A pivotal study in the field of pediatric cancer demonstrated that patients with BCP-ALL mount, in the Bone marrow (BM), a robust endogenous CD8+ T cell response against a majority of predicted private neoantigens,11 suggesting that pediatric cancers may also exhibit substantial natural immunity. Although private neoantigens are associated with potent antitumoral immunity, tumor-associated antigens (TAAs) are also relevant targets for immune responses in this cancer type.7, 12, 13 Additionally, it is unclear whether a systemic immune response is mounted in these patients. Compared to the previously described study,11 we therefore expanded our analysis to peripheral blood (PB) and directly identified TAAs through high-throughput immunopeptidomics, while employing an experimentally validated multiparameter pipeline for refined neoantigen prediction. Building on these insights, we set out to investigate the presence of endogenous T cells recognizing tumor antigens in the PB of pediatric BCP-ALL patients after remission (Supporting Information S2: Methods and Materials). Eight pediatric patients diagnosed with BCP-ALL and treated at the Centre hospitalier universitaire vaudois (CHUV) in Lausanne were included (Supporting Information S1: Table 1). BM and PB samples were collected respectively at diagnosis (V1) and timepoints after remission subsequent to treatment (V3 or follow-up) (Supporting Information S1: Figure 1). HLA-I and HLA-II restricted TAAs and neoantigens were respectively identified and predicted. Peptide-specific T-cell reactivity was assessed using ELISpot and flow cytometry after in vitro stimulation. Of interest, circulating T cells targeting tumor-expressing antigens (TAAs and neoantigens) were identified in virtually all tested pediatric BCP-ALL patients, extending up to 2 years on remission. First, we employed a mass spectrometry-based immunopeptidomic workflow using a quadrupole Orbitrap mass spectrometer (Q Exactive HF-X) and the MaxQuant environment to identify TAAs presented across seven patients, using BM samples collected at diagnosis (V1). Overall, we manually selected 27 peptides derived from known TAAs for immunogenicity testing (Supporting Information S1: Table 2). As an initial validation step, we aimed to validate our experimental and technical approach by reproducing the observations from the seminal study of Zamora et al.11 To this end, we interrogated the BM of patient ISHOP07 against MS-identified TAAs and successfully detected T-cell responses against two out of the three tested antigens (TAA22, TAA23) (Figure 1A). We then asked whether the same T-cell reactivities would be detectable in the patient's PB at a later timepoint (V3, >3 months after initial treatment). Of interest, circulating tumor-specific T cells were identified against one of the TAAs (TAA-22) found immunogenic in BM (Figure 1B). Building on this observation, we examined PB samples from six additional patients for responses to their respective tumor antigens. Five out of six patients also demonstrated cellular responses to one or multiple TAAs (Figure 1B). Overall, amongst the six positive patients, the average response rate totaled around 57% of the tested TAAs (Figure 1C) with a magnitude of cellular responses ranging from 1.29% to 7.3% of the Staphylococcal enterotoxin B (SEB) response (Figure 1D). Interestingly, TAA-2 (GSDCTTIHY from TP53) was found immunogenic in two out of three patients (ISHOP09 and ISHOP15), both carrying HLA-A*01:01 and sharing the same genetic leukemia subtype (TCF3-PBX1), whereas the nonresponding patient (ISHOP10) had a different genetic classification (KMT2A). Furthermore, for three patients (ISHOP09, ISHOP13, and ISHOP14), we also performed whole exome sequencing of tumoral and germline DNA (V3 was used for germline DNA, demonstrating minimal disease burden) and predicted neoantigens that are likely to bind the patients' HLA allotypes as short (S) and long (L) peptides (Supporting Information S1: Table 3). For patients ISHOP09 and ISHOP13, neoantigen-specific responses in PB were evaluated using samples collected shortly after remission (V3), whereas ISHOP14 was assessed at a later follow-up timepoint (>2 years after initial treatment) to evaluate the longevity of the immune response. Patient ISHOP13 showed no antigenic response above background, but the limited sample size restricted testing to peptide pools rather than individual peptides. In contrast, ISHOP9 responded to seven out of the nine tested neoantigens, with an overall reactivity amounting to approximately 4% of the SEB reference response (Figure 2A). As mentioned above, to further assess the sustainability of endogenous T cell responses observed in PB of leukemia patients, ISHOP 14 was analyzed using a blood sample collected at a unique follow-up timepoint 2 years posttreatment. Notably this patient demonstrated a response to eight out of nine tested neoantigens, with a cumulative response reaching 32% of the SEB response (Figure 2B), that is, in a similar range relative to what was observed in BM in Zamora et al.11 The most immunogenic peptide, L-8, induced over 12% of peripheral blood mononuclear cell (PBMC) interferon (IFN)γ production. Summarized response rates and intensities observed in each patient towards private neoantigens are depicted in Figure 2C,D. Using intracellular cytokine staining, we observed that the neoantigen-specific T-cell response against L-8 was predominantly mediated by CD4+ T cells (Figure 2E), with only minimal contribution from CD8+ T cells to the overall lymphocytic response (Supporting Information S1: Figure 2). Additionally, L-8 specific CD4+ T cells were polyfunctional with over 75% of CD4+ T cells exhibiting two or three functions (IFNγ and/or tumor necrosis factor-α production or CD154 expression) (Figure 2E). Overall, our findings challenge the prevailing notion that low mutational burden restricts tumor immunogenicity, reinforcing evidence that endogenous antitumor reactivity occurs even in low-TMB cancers. Consistent with Zamora et al.'s observations, we describe a high neoantigen response rate in our cohort, surpassing the <3% recognition typically reported in adult tumors, including those with high mutational burden,10 while TAA reactivity was also significant. We also observed considerable variability in antigenic-specific responses to both TAAs and neoantigens within some patients, suggesting immunodominance, defined as the prioritization of specific antigens over broader targeting. Despite a lower cumulative immune response than reported by Zamora et al., it is important to interpret the intensity of responses in the context of chemotherapy-induced myeloablation14 and an immunosuppressive microenvironment,15 which significantly impair overall immune function. Nevertheless, the mere presence of endogenous TILs has been linked to improved treatment outcome, while their amplification further enhances clinical outcome in various malignancies,16, 17 highlighting their therapeutic potential. Another key discovery is the detection of tumor-reactive lymphocytes in PB rather than exclusively in BM. Remarkably, the presence and diversity of circulating tumor-specific cells have been correlated with treatment outcome in solid cancer.18 If peripheral responses reflect BM immunity, blood sampling could offer a practical alternative for immune profiling and therapeutic cell harvesting, reducing the need for invasive procedures. Finally, a crucial observation was the detection of antigen-specific circulating T lymphocytes beyond remission, with one patient (ISHOP14) exhibiting a robust response to private neoantigens over 2 years after initial treatment. This persistence suggests ongoing cancer-specific immune surveillance even in the absence of detectable disease, an important factor in relapse-free survival, notably following adoptive cell transfer therapies or other treatments.16, 17, 19 This finding is especially significant in pediatric BCP-ALL, where despite low relapse rates, the majority of relapses occur within the few years of diagnosis and are associated with poor outcome.20 It is important to note that for ISHOP14, analyses were conducted after a second round of in vitro stimulation. The enhanced reactivity in this patient further supports the hypothesis that aggressive immunosuppressive treatments reduce T-cell quantity and function, necessitating stronger stimuli to elicit a significant immune response. As in most pediatric studies, the limited number of participants and restricted cell availability limited data interpretation, underscoring the need for larger studies into antigen-specific T-cell responses and their clinical impact. Further research on the function and longevity of antigen-specific T cells posttreatment is also essential to evaluate their potential role in immune surveillance and relapse prevention. In conclusion, our discovery of circulating tumor antigen-reactive T cells in PB following remission provides compelling evidence against the notion that pediatric cancers lack endogenous antitumor response. Harnessing these natural responses—either through direct expansion or combination with existing therapies—offers significant potential, opening promising avenues for more research on targeted, personalized immunotherapeutic strategies and improved long-term disease control. Open access publishing facilitated by Universite de Lausanne, as part of the Wiley - Universite de Lausanne agreement via the Consortium Of Swiss Academic Libraries. Eliane Huwiler: Investigation; validation; writing—original draft; writing—review and editing; formal analysis. Anne-Christine Thierry: Methodology; data curation; formal analysis; investigation; validation; writing—review and editing. Justine Michaux: Investigation; validation; methodology; formal analysis; data curation; writing—review and editing. Huisong Pak: Investigation; methodology; validation; writing—review and editing; formal analysis; data curation. Florian Huber: Investigation; methodology; validation; writing—review and editing; formal analysis; data curation. Caroline Arber: Methodology; writing—review and editing. Michal Bassani-Sternberg: Conceptualization; methodology; investigation; data curation; supervision; validation; writing—review and editing. Alexandre Harari: Conceptualization; methodology; supervision; validation; investigation; writing—review and editing. Francesco Ceppi: Conceptualization; methodology; data curation; investigation; validation; funding acquisition; writing—original draft; writing—review and editing; supervision; formal analysis; project administration. The authors declare no conflicting interest with the submission of this article. Ethical approval for this study was obtained from the Commission cantonale d’éthique sur la recherche de l’être humain (CER-VD), under the protocol number 2016-01610 (approved November 4, 2016). This work was supported by funding from the Fondation Recherche sur le Cancer de L'Enfant (FORCE), Lausanne, Switzerland. Written informed consent was obtained from all participants, their parents, or their legal representative, after Protocol approval by the ethics review board. The mass spectrometry proteomics data are available via ProteomeXchange with identifier PXD061914. The genomic data are openly available at the European Genome-phenome Archive (EGA) with study identifier EGAS50000000925 and dataset identifier EGAD50000001355. The rest of the data that support the findings of this study are available from the corresponding author, F.C., upon reasonable request. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.