Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Negative trade-off between neoantigen repertoire breadth and the specificity of HLA-I molecules shapes antitumor immunity

Abstract

Human leukocyte antigen class I (HLA-I) genes shape our immune response against pathogens and cancer. Certain HLA-I variants can bind a wider range of peptides than others, a feature that could be favorable against a range of viral diseases. However, the implications of this phenomenon on cancer immune response are unknown. Here we quantified peptide repertoire breadth (or promiscuity) of a representative set of HLA-I alleles and found that patients with cancer who were carrying HLA-I alleles with high peptide-binding promiscuity have significantly worse prognosis after immune checkpoint inhibition. This can be explained by a reduced capacity of the immune system to discriminate tumor neopeptides from self-peptides when patients carry highly promiscuous HLA-I variants, shifting the regulation of tumor-infiltrating T cells from activation to tolerance. In summary, HLA-I peptide-binding specificity shapes neopeptide immunogenicity and the self-immunopeptidome repertoire in an antagonistic manner, and could underlie a negative trade-off between antitumor immunity and genetic susceptibility to viral infections.

This is a preview of subscription content, access via your institution

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Fig. 1: Basic properties of HLA-I allelic promiscuity.
Fig. 2: High allele promiscuity is associated with more stable peptide–HLA complexes.
Fig. 3: High allele promiscuity and survival rate in patient cohorts treated with checkpoint inhibitor therapy.
Fig. 4: HLA-I allelic promiscuity and differentiation capacity between neopeptides and corresponding nonmutated peptides.
Fig. 5: Association of high HLA genotype promiscuity with an immunosuppressive cancer microenvironment.

Data availability

Previously published clinical datasets that were reanalyzed here are available at refs. 4,23,24,25,28,54. The following databases were used in the study: the Immune Epitope Database, https://www.iedb.org/; TANTIGEN database, http://projects.met-hilab.org/tadb/; the Cancer Genome Atlas, https://portal.gdc.cancer.gov/; cBioPortal, https://www.cbioportal.org/; UniProt database, https://www.uniprot.org/. Precalculated data on T-cell dysfunction were downloaded from the TIDE server, http://tide.dfci.harvard.edu/.Source data are provided with this paper. All other data supporting the findings of this study are available from the corresponding author on reasonable request.

Code availability

The code for HLA allele promiscuity calculation is available at https://github.com/matemanc/hla_promiscuity.

References

  1. Trowsdale, J. The MHC, disease and selection. Immunol. Lett. 137, 1–8 (2011).

    CAS  PubMed  Google Scholar 

  2. Schumacher, T. N. & Schreiber, R. D. Neoantigens in cancer immunotherapy. Science 348, 69–74 (2015).

    CAS  PubMed  Google Scholar 

  3. Verdegaal, E. M. E. et al. Neoantigen landscape dynamics during human melanoma–T cell interactions. Nature 536, 91–95 (2016).

    CAS  PubMed  Google Scholar 

  4. Chowell, D. et al. Patient HLA class I genotype influences cancer response to checkpoint blockade immunotherapy. Science 359, 582–587 (2018).

    CAS  PubMed  Google Scholar 

  5. Chowell, D. et al. Evolutionary divergence of HLA class I genotype impacts efficacy of cancer immunotherapy. Nat. Med. 25, 1715–1720 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  6. Negrao, M. V. et al. PD-L1 expression, tumor mutational burden, and cancer gene mutations are stronger predictors of benefit from immune checkpoint blockade than HLA class I genotype in non-small cell lung cancer. J. Thorac. Oncol. 14, 1021–1031 (2019).

    CAS  PubMed  Google Scholar 

  7. Litchfield, K. et al. Meta-analysis of tumor and T cell intrinsic mechanisms of sensitization to checkpoint inhibition. Cell 184, 596–614 (2021).

    CAS  PubMed  PubMed Central  Google Scholar 

  8. Marty, R. et al. MHC-I genotype restricts the oncogenic mutational landscape. Cell 171, 1272–1283 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  9. McGranahan, N. et al. Allele-specific HLA loss and immune escape in lung cancer evolution. Cell 171, 1259–1271 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  10. Manczinger, M. et al. Pathogen diversity drives the evolution of generalist MHC-II alleles in human populations. PLoS Biol. 17, e3000131 (2019).

    PubMed  PubMed Central  Google Scholar 

  11. Kaufman, J. Generalists and specialists: a new view of how MHC class I molecules fight infectious pathogens. Trends Immunol. 39, 367–379 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  12. Paul, S. et al. HLA class I alleles are associated with peptide-binding repertoires of different size, affinity, and immunogenicity. J. Immunol. 191, 5831–5839 (2013).

    CAS  PubMed  Google Scholar 

  13. Košmrlj, A. et al. Effects of thymic selection of the T-cell repertoire on HLA class I-associated control of HIV infection. Nature 465, 350–354 (2010).

    PubMed  PubMed Central  Google Scholar 

  14. Havel, J. J., Chowell, D. & Chan, T. A. The evolving landscape of biomarkers for checkpoint inhibitor immunotherapy. Nat. Rev. Cancer 19, 133–150 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  15. Weiskopf, D. et al. Comprehensive analysis of dengue virus-specific responses supports an HLA-linked protective role for CD8+ T cells. Proc. Natl Acad. Sci. USA 110, E2046–E2053 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  16. Abelin, J. G. et al. Mass spectrometry profiling of HLA-associated peptidomes in mono-allelic cells enables more accurate epitope prediction. Immunity 46, 315–326 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  17. Di Marco, M. et al. Unveiling the peptide motifs of HLA-C and HLA-G from naturally presented peptides and generation of binding prediction matrices. J. Immunol. 199, 2639–2651 (2017).

    PubMed  Google Scholar 

  18. Olsen, L. R. et al. TANTIGEN: a comprehensive database of tumor T cell antigens. Cancer Immunol. Immunother. 66, 731–735 (2017).

    CAS  PubMed  Google Scholar 

  19. Robbins, P. F. et al. Mining exomic sequencing data to identify mutated antigens recognized by adoptively transferred tumor-reactive T cells. Nat. Med. 19, 747–752 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  20. Jurtz, V. et al. NetMHCpan-4.0: improved peptide–MHC class I interaction predictions integrating eluted ligand and peptide binding affinity data. J. Immunol. 199, 3360–3368 (2017).

    CAS  PubMed  Google Scholar 

  21. Harndahl, M. et al. Peptide-MHC class I stability is a better predictor than peptide affinity of CTL immunogenicity: antigen processing. Eur. J. Immunol. 42, 1405–1416 (2012).

    CAS  PubMed  Google Scholar 

  22. Fromer, M. & Shifman, J. M. Tradeoff between stability and multispecificity in the design of promiscuous proteins. PLoS Comput. Biol. 5, e1000627 (2009).

    PubMed  PubMed Central  Google Scholar 

  23. Van Allen, E. M. et al. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science 350, 207–211 (2015).

    PubMed  PubMed Central  Google Scholar 

  24. Snyder, A. et al. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N. Engl. J. Med. 371, 2189–2199 (2014).

    PubMed  PubMed Central  Google Scholar 

  25. Rizvi, N. A. et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science 348, 124–128 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  26. Mariathasan, S. et al. TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 554, 544–548 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  27. Cristescu, R. et al. Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362, eaar3593 (2018).

    PubMed  PubMed Central  Google Scholar 

  28. Riaz, N. et al. Tumor and microenvironment evolution during immunotherapy with nivolumab. Cell 171, 934–949 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Sidney, J., Peters, B., Frahm, N., Brander, C. & Sette, A. HLA class I supertypes: a revised and updated classification. BMC Immunol. 9, 1 (2008).

    PubMed  PubMed Central  Google Scholar 

  30. Samstein, R. M. et al. Tumor mutational load predicts survival after immunotherapy across multiple cancer types. Nat. Genet. 51, 202–206 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  31. Kaur, G. et al. Structural and regulatory diversity shape HLA-C protein expression levels. Nat. Commun. 8, 15924 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  32. Ghorani, E. et al. Differential binding affinity of mutated peptides for MHC class I is a predictor of survival in advanced lung cancer and melanoma. Ann. Oncol. 29, 271–279 (2018).

    CAS  PubMed  Google Scholar 

  33. Rech, A. J. et al. Tumor immunity and survival as a function of alternative neopeptides in human cancer. Cancer Immunol. Res. 6, 276–287 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  34. Duan, F. et al. Genomic and bioinformatic profiling of mutational neoepitopes reveals new rules to predict anticancer immunogenicity. J. Exp. Med. 211, 2231–2248 (2014).

    PubMed  PubMed Central  Google Scholar 

  35. Łuksza, M. et al. A neoantigen fitness model predicts tumour response to checkpoint blockade immunotherapy. Nature 551, 517–520 (2017).

    PubMed  PubMed Central  Google Scholar 

  36. Richman, L. P., Vonderheide, R. H. & Rech, A. J. Neoantigen dissimilarity to the self-proteome predicts immunogenicity and response to immune checkpoint blockade. Cell Syst. 9, 375–382 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  37. Xing, Y. & Hogquist, K. A. T-cell tolerance: central and peripheral. Cold Spring Harb. Perspect. Biol. 4, a006957 (2012).

    PubMed  PubMed Central  Google Scholar 

  38. Nurieva, R., Wang, J. & Sahoo, A. T-cell tolerance in cancer. Immunotherapy 5, 513–531 (2013).

    CAS  PubMed  Google Scholar 

  39. Binnewies, M. et al. Understanding the tumor immune microenvironment (TIME) for effective therapy. Nat. Med. 24, 541–550 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  40. Vesely, M. D., Kershaw, M. H., Schreiber, R. D. & Smyth, M. J. Natural innate and adaptive immunity to cancer. Annu. Rev. Immunol. 29, 235–271 (2011).

    CAS  PubMed  Google Scholar 

  41. Noy, R. & Pollard, J. W. Tumor-associated macrophages: from mechanisms to therapy. Immunity 41, 49–61 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  42. Walker, C., Mojares, E. & Del Río Hernández, A. Role of extracellular matrix in development and cancer progression. Int. J. Mol. Sci. 19, 3028 (2018).

    PubMed Central  Google Scholar 

  43. Henke, E., Nandigama, R. & Ergün, S. Extracellular matrix in the tumor microenvironment and its impact on cancer therapy. Front. Mol. Biosci. 6, 160 (2019).

    CAS  PubMed  Google Scholar 

  44. Sturm, G. et al. Comprehensive evaluation of transcriptome-based cell-type quantification methods for immuno-oncology. Bioinformatics 35, i436–i445 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  45. Haque, S. & Morris, J. C. Transforming growth factor-β: a therapeutic target for cancer. Hum. Vaccin. Immunother. 13, 1741–1750 (2017).

    PubMed  PubMed Central  Google Scholar 

  46. Colak, S. & ten Dijke, P. Targeting TGF-β signaling in cancer. Trends Cancer 3, 56–71 (2017).

    CAS  PubMed  Google Scholar 

  47. Das, M., Zhu, C. & Kuchroo, V. K. Tim-3 and its role in regulating anti-tumor immunity. Immunol. Rev. 276, 97–111 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  48. Manieri, N. A., Chiang, E. Y. & Grogan, J. L. TIGIT: a key inhibitor of the cancer immunity cycle. Trends Immunol. 38, 20–28 (2017).

    CAS  PubMed  Google Scholar 

  49. Gautron, A.-S., Dominguez-Villar, M., de Marcken, M. & Hafler, D. A. Enhanced suppressor function of TIM-3+ FoxP3 + regulatory T cells: immunomodulation. Eur. J. Immunol. 44, 2703–2711 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  50. Jiang, P. et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat. Med. 24, 1550–1558 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  51. Xia, A., Zhang, Y., Xu, J., Yin, T. & Lu, X.-J. T cell dysfunction in cancer immunity and immunotherapy. Front. Immunol. 10, 1719 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  52. Scott, A. C. et al. TOX is a critical regulator of tumour-specific T cell differentiation. Nature 571, 270–274 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  53. Radwan, J., Babik, W., Kaufman, J., Lenz, T. L. & Winternitz, J. Advances in the evolutionary understanding of MHC polymorphism. Trends Genet. 36, 298–311 (2020).

    CAS  PubMed  Google Scholar 

  54. Le, D. T. et al. PD-1 blockade in tumors with mismatch-repair deficiency. N. Engl. J. Med. 372, 2509–2520 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  55. Alspach, E. et al. MHC-II neoantigens shape tumour immunity and response to immunotherapy. Nature 574, 696–701 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  56. Marty Pyke, R. et al. Evolutionary pressure against MHC class II binding cancer mutations. Cell 175, 416–428 (2018).

    PubMed  Google Scholar 

  57. Rolland, M. et al. Recognition of HIV-1 peptides by host CTL is related to HIV-1 similarity to human proteins. PLoS ONE 2, e823 (2007).

    PubMed  PubMed Central  Google Scholar 

  58. McGranahan, N. & Swanton, C. Neoantigen quality, not quantity. Sci. Transl. Med. 11, eaax7918 (2019).

    PubMed  Google Scholar 

  59. Peng, M. et al. Neoantigen vaccine: an emerging tumor immunotherapy. Mol. Cancer 18, 128 (2019).

    PubMed  PubMed Central  Google Scholar 

  60. Vita, R. et al. The Immune Epitope Database (IEDB): 2018 update. Nucleic Acids Res. 47, D339–D343 (2019).

    CAS  PubMed  Google Scholar 

  61. The UniProt Consortium UniProt: a worldwide hub of protein knowledge. Nucleic Acids Res. 47, D506–D515 (2019).

    Google Scholar 

  62. Marcon, E., Hérault, B., Baraloto, C. & Lang, G. The decomposition of Shannon’s entropy and a confidence interval for beta diversity. Oikos 121, 516–522 (2012).

    Google Scholar 

  63. Bjerregaard, A.-M. et al. An analysis of natural T cell responses to predicted tumor neoepitopes. Front. Immunol. 8, 1566 (2017).

    PubMed  PubMed Central  Google Scholar 

  64. Gros, A. et al. Prospective identification of neoantigen-specific lymphocytes in the peripheral blood of melanoma patients. Nat. Med. 22, 433–438 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  65. Zacharakis, N. et al. Immune recognition of somatic mutations leading to complete durable regression in metastatic breast cancer. Nat. Med. 24, 724–730 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  66. Tran, E. et al. Immunogenicity of somatic mutations in human gastrointestinal cancers. Science 350, 1387–1390 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  67. Carreno, B. M. et al. A dendritic cell vaccine increases the breadth and diversity of melanoma neoantigen-specific T cells. Science 348, 803–808 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  68. Ogishi, M. & Yotsuyanagi, H. Quantitative prediction of the landscape of T cell epitope immunogenicity in sequence space. Front. Immunol. 10, 827 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  69. Nielsen, M. & Andreatta, M. NetMHCpan-3.0; improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length datasets. Genome Med. 8, 33 (2016).

    PubMed  PubMed Central  Google Scholar 

  70. Rasmussen, M. et al. Pan-specific prediction of peptide–MHC class I complex stability, a correlate of T cell immunogenicity. J. Immunol. 197, 1517–1524 (2016).

    CAS  PubMed  Google Scholar 

  71. Grossman, R. L. et al. Toward a shared vision for cancer genomic data. N. Engl. J. Med. 375, 1109–1112 (2016).

    PubMed  PubMed Central  Google Scholar 

  72. Cerami, E. et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data: figure 1. Cancer Discov. 2, 401–404 (2012).

    PubMed  Google Scholar 

  73. Liu, J. et al. An integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell 173, 400–416 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  74. Li, X. et al. Benchmarking HLA genotyping and clarifying HLA impact on survival in tumor immunotherapy. Mol. Oncol. https://pubmed.ncbi.nlm.nih.gov/33411982/ (2021).

  75. Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinformatics 10, 421 (2009).

    PubMed  PubMed Central  Google Scholar 

  76. Carrasco Pro, S. et al. Microbiota epitope similarity either dampens or enhances the immunogenicity of disease-associated antigenic epitopes. PLoS ONE 13, e0196551 (2018).

    PubMed  PubMed Central  Google Scholar 

  77. Sarkizova, S. et al. A large peptidome dataset improves HLA class I epitope prediction across most of the human population. Nat. Biotechnol. 38, 199–209 (2020).

    CAS  PubMed  Google Scholar 

  78. Racle, J., de Jonge, K., Baumgaertner, P., Speiser, D. E. & Gfeller, D. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. eLife 6, e26476 (2017).

    PubMed  PubMed Central  Google Scholar 

  79. Becht, E. et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 17, 218 (2016).

    PubMed  PubMed Central  Google Scholar 

  80. Finotello, F. et al. Molecular and pharmacological modulators of the tumor immune contexture revealed by deconvolution of RNA-seq data. Genome Med. 11, 34 (2019).

    PubMed  PubMed Central  Google Scholar 

  81. Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).

    CAS  PubMed  PubMed Central  Google Scholar 

  82. The Gene Ontology Consortium. The Gene Ontology Resource: 20 years and still GOing strong. Nucleic Acids Res. 47, D330–D338 (2019).

  83. De Boor, C. A Practical Guide to Splines (Springer, 1978).

  84. Wagih, O. ggseqlogo: a versatile R package for drawing sequence logos. Bioinformatics 33, 3645–3647 (2017).

    CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank T. Lenz for earlier discussions on this topic and for providing us with the raw data on HED values of HLA-I allele pairs. We also thank the anonymous reviewers for their insightful suggestions on the manuscript, and B. Győrffy. for his suggestions on survival analysis. The results here are in part based on data generated by the TCGA Research Network (https://www.cancer.gov/tcga). We thank H. Ye for providing the HLA genotype data of TCGA patients. Similarly, we thank R. Marty, J. Font-Burgada and H. Carter for sending HLA genotype data of TCGA patients for earlier analyses. The study was supported by the following research grants: The European Research Council (H2020-ERC-2014-CoG, no. 648364 – Resistance Evolution, to C.P.); ‘Célzott Lendület’ Program of the Hungarian Academy of Sciences (LP-2017–10/2017, to C.P.); ‘Frontline’ Research Excellence Program of the National Research, Development and Innovation Office, Hungary (KKP-126506, to C.P.); EVOMER (GINOP-2.3.2–15–2016–00014, to C.P. and B.P.); MolMedEx TUMORDNS (GINOP-2.3.2–15–2016–00020 and GINOP-2.3.3–15–2016–00001, to C.P., and GINOP-2.3.2–15–2016–00026, to B.P.); and ‘Frontline’ Research Excellence Program of the National Research, Development and Innovation Office, Hungary (KKP-129814, to B.P.). The project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement 739593 (to B.P., L.K. and M.M). M.M. was supported by the New National Excellence Program of the Ministry of Human Capacities (UNKP-20-5) and by the Bolyai János Research Fellowship of the Hungarian Academy of Sciences. L.A. was supported by the New National Excellence Program of the Ministry for Innovation and Technology (UNKP-20-2). G.M.B. was supported by the New National Excellence Program of the Ministry for Innovation and Technology (UNKP-19-3). B.K. was supported by the New National Excellence Program of the Ministry for Innovation and Technology (UNKP-20-4). B.T.P. was supported by the New National Excellence Program of the Ministry for Innovation and Technology (UNKP-20-3). This research work was conducted with the support of the Szeged Scientists Academy under the sponsorship of the Hungarian Ministry of Innovation and Technology (FEIF/433-4/2020-ITM_SZERZ). We thank D. Bokor for proofreading of the manuscript.

Author information

Authors and Affiliations

Authors

Contributions

M.M., L.K., B.P. and C.P. undertook conceptualization. M.M., B.K., G.M.B., B.T.P. and L.A. were responsible for methodology. Software was provided by M.M. Formal analysis was performed by M.M., B.K., G.M.B., B.T.P. and L.A. Investigation was carried out by M.M., B.K., G.M.B., B.T.P., B.P. and C.P. Resources were provided by M.M., C.P. and B.P. Data curation was undertaken by M.M., B.K., G.M.B. and B.T.P. Writing of the original draft was done by M.M., B.P. and C.P. Writing, review and editing were performed by M.M., B.P. and C.P. M.M., L.K., C.P. and B.P. supervised the project. Project administration was carried out by M.M. and C.P. Funding acquisition was the responsibility of C.P. and B.P.

Corresponding authors

Correspondence to Máté Manczinger or Csaba Pál.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Cancer thanks Pramod Srivastava and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Calculation and reliability of HLA allele promiscuity.

a, DKL values were calculated for each peptide length group separately (see Methods for details) and (b) the weighted mean of length-specific values was determined. c, d, The number of identified peptide sequences bound by a given HLA allele is variable, which may have a significant effect on the calculated Kullback-Leibler (KL) divergence values. To test whether KL divergence can be reliably determined using a minimal peptide sequence number of 400, we first selected HLA alleles with many (more than 1000) identified peptides (n = 51 alleles). After randomly selecting 400 sequences from the repertoire of each allele, we calculated KL divergence as explained in the Methods section. We iterated this process 1000 times and compared the Pr values with those calculated using the full dataset. Panel c indicates the distribution of Spearman’s ρ values and the two-sided correlation test P values. d, An example plot indicating the relationship between promiscuity values calculated with either 400 random or all peptides bound by a given allele (n = 51 alleles). e, f, The calculated promiscuity values are independent of the reference amino acid (AA) distribution used during calculation (Q on panel a). The Pr values showed strong correlation when calculated using (e) human and viral or (f) human and bacterial reference AA distributions (Spearman’s ρ: 0.99, two-sided correlation test P < 2.2 × 10−16 for both comparisons, n = 67 alleles on both panels). On plots d to f, dashed lines indicate a smooth curve fitted using cubic smoothing spline method in R (see Methods).

Source data

Extended Data Fig. 2 Sequence diversity of peptides presented by HLA class I alleles.

Amino acid preference was determined at different peptide positions bound by each of the 67 investigated HLA class I alleles. Sequence logos were visualized for 9 amino acid long peptides only84. Alleles are ordered according to their promiscuity level. The vertical axes indicate the information content of the given position in bits. The relative sizes of characters are proportional to their frequency at the given position.

Source data

Extended Data Fig. 3 The relationship between HLA allele promiscuity and (a, b) the diversity of presented peptides or (c-e) fraction of bound neoepitopes.

a, b, We found a strong correlation between the HLA allele promiscuity and sequence diversity of peptides eluted from the surface of monoallelic cell lines in studies by (a) Abelin et al. (n = 16 alleles) and (b) Di Marco et al. (n = 15 alleles). c-e, The relationship between the promiscuity of (c) HLA-A (n = 21 alleles) (d) HLA-B (n = 31 alleles) and (e) HLA-C (n = 15 alleles) alleles and the predicted fraction of bound neopeptides. The variance of allele promiscuity is shown on panels c to e. On all plots, Spearman’s ρ and the corresponding two-sided correlation test P values are indicated and dashed red lines indicate a linear regression line.

Source data

Extended Data Fig. 4 Distribution of genotype Pr in cancer immunotherapy cohorts.

The density of genotype Pr is shown for (a) all patients (n = 316), (b) melanoma patients treated with CTLA-4 inhibitors (n = 164), (c) non-small cell lung cancer (NSCLC) patients treated with PD-1 or PD-L1 inhibitors (n = 74) and (d) melanoma patients treated with PD-1 or PD-L1 inhibitors (n = 78). Dashed black vertical lines represent the 25th and 75th percentile values, while dashed red vertical lines represent the fixed genotype Pr cutoff of 2.076 used for patient classification.

Source data

Extended Data Fig. 5 The effect of HLA promiscuity on response to ICI therapy and survival of treatment naïve patients.

a, b, Response to ICI therapy in an independent melanoma cohort. Patients were stratified into low (n = 20 patients) and high (n = 11 patients) promiscuity groups using the same cutoff as in Fig. 3. On the boxplot, the P value of a two-sided Wilcoxon’s rank-sum test is indicated. Vertical lines indicate median, boxes indicate the interquartile range (IQR), horizontal lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR. On the pie chart, the P value of a two-sided Fisher’s exact test is shown. c-h, The relationship between HLA promiscuity and progression-free survival of treatment naïve melanoma (c and f) and NSCLC patients (d, e, g and h). Similarly to previous studies5, patients were classified into low (c to e) and high (f to h) mutational burden cohorts using the median as a cutoff. In each cohort, patients were classified into low and high promiscuity groups using the median as a cutoff. Two-sided log-rank test P values are shown. The vertical axes indicate the probability of progression-free survival. LUSC: lung squamous cell carcinoma; LUAD: lung adenocarcinoma. i, j, The relationship between HLA-I genotype Pr and changes in tumor size after anti-PD-1 treatment in mismatch repair deficient (i, n = 10) and proficient (j, n = 11) cancer patients. Patients were stratified into low and high genotype Pr groups using the median value as a cutoff. P values indicate the outcome of two-sided two-sample T-tests. On boxplots, vertical lines indicate median, boxes indicate the interquartile range (IQR), horizontal lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR.

Source data

Extended Data Fig. 6 HLA-I promiscuity is a major determinant of patient survival.

a, Carrying a single promiscuous HLA-I allele has no impact on survival. The patients in Fig. 1a to c were classified based on the number of promiscuous alleles in their genotypes. Alleles with promiscuity levels in the top quartile were considered to be promiscuous. Two-sided log-rank test P values are shown. In the second analysis, between-group differences were also tested for trend (see Methods). The vertical axes indicate the probability of survival. b, HLA alleles belonging to the B44 supergroup have low promiscuity. B44 alleles are highlighted in blue and have significantly lower promiscuity than alleles not belonging to this group (two-sided Wilcoxon’s rank-sum test P: 0.049). c-e, No significant association between HLA-I genotype Pr and (c) HLA homozygosity, (d) HLA-I evolutionary divergence (HED), (e) mutational burden. All patients from Fig. 3a to c were included in these analyses. c, There was no significant difference in genotype Pr between fully HLA heterozygous patients (n = 251) and the ones homozygous for at least one HLA-I locus (n = 65 patients). The P value for a two-sided Wilcoxon’s rank-sum test is indicated. Similarly, there was no significant association between (d) genotype Pr and mean HED (Spearman’s ρ: −0.02, two-sided correlation test P = 0.77, n = 316 patients) and (e) genotype Pr and tumor mutational burden of cancer immunotherapy patients (Spearman’s ρ: 0.02, two-sided correlation test P = 0.77, n = 316 patients). Dashed red lines indicate a smooth curve fitted using the cubic smoothing spline method in R (see Methods). On the boxplot, horizontal lines indicate median, boxes indicate the interquartile range (IQR), vertical lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR.

Extended Data Fig. 7 The relationship between genotype Pr and DAI.

a, The median and variance of DAI in different allele promiscuity groups. Alleles were grouped based on the 3rd quartile as a cutoff (n = 26 and 77 alleles in the high and low promiscuity group, respectively). High allele promiscuity was associated with lower median DAI, but it had no significant effect on the coefficient of variation. P values of two-sided Wilcoxon’s rank-sum tests are shown. On the boxplots, horizontal lines indicate median, boxes indicate the interquartile range (IQR), vertical lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR. b, The number and fraction of HLA-bound neopeptides in different DAI value ranges. The binding of 589 neopeptides to 103 HLA alleles was determined and DAI values were calculated for each neopeptide-human self-peptide pair. Alleles are shown in increasing order of promiscuity. For each allele, the fraction of peptide pairs belonging to different DAI categories is shown color-coded. c, The effect of genotype DAI on survival when determined as the average of allele-specific DAI values. Patients were stratified into groups using the 25th and 75th percentile of all values as cutoffs. A two-sided log-rank test P value is shown. Between-group differences were also tested for trend (see Methods). The vertical axis indicates the probability of survival. Higher genotype DAI value was associated with better survival. d, The maximum BLOSUM62 sequence similarity of neopeptides (n = 589) and immunogenic viral (n = 1038) peptides to the non-mutated human proteome. Only nine amino acid-long peptides were included in this analysis. P value of a two-sided Wilcoxon’s rank-sum test is indicated. On boxplots, horizontal lines indicate median, boxes indicate the interquartile range (IQR), vertical lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR.

Source data

Extended Data Fig. 8 The effect of HLA-associated features on TIDE dysfunction score.

a, The TIDE dysfunction score is shown as a function of HLA evolutionary divergence. There was no relationship between the two variables (Spearman’s ρ: −0.01, two-sided correlation test P = 0.96, n = 35 samples). Similarly, neither HLA homozygosity (b) nor the carrier status for B44 (c, n = 17 and 18 samples in B44 positive and B44 negative groups, respectively) or B62 (d, n = 5 and 30 samples in B62 positive and negative groups, respectively) supertype alleles affected the TIDE dysfunction score. P values indicate two-sided Wilcoxon’s rank-sum test results. On plot a, the dashed red line indicates a linear regression line. On boxplots, horizontal lines indicate median, boxes indicate the interquartile range (IQR), vertical lines indicate 1st quartile - 1.5 × IQR and 3rd quartile + 1.5 × IQR.

Source data

Supplementary information

Supplementary Information

Supplementary Tables 5–7.

Reporting Summary

Supplementary Data 1

Supplementary Tables 1–4 and 8.

Source data

Source Data Fig. 1

Statistical source data.

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 1

Statistical source data.

Source Data Extended Data Fig. 2

Statistical source data.

Source Data Extended Data Fig. 3

Statistical source data.

Source Data Extended Data Fig. 4

Statistical source data.

Source Data Extended Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 6

Statistical source data.

Source Data Extended Data Fig. 7

Statistical source data.

Source Data Extended Data Fig. 8

Statistical source data.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Manczinger, M., Koncz, B., Balogh, G.M. et al. Negative trade-off between neoantigen repertoire breadth and the specificity of HLA-I molecules shapes antitumor immunity. Nat Cancer 2, 950–961 (2021). https://doi.org/10.1038/s43018-021-00226-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s43018-021-00226-4

Search

Quick links

Nature Briefing: Cancer

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

Get what matters in cancer research, free to your inbox weekly. Sign up for Nature Briefing: Cancer