An optimized antigen–protein fusion

The fusion of an immunogenic peptide and the protein transthyretin protects the peptide antigen from proteolytic degradation, optimizes its uptake in local draining lymphatics and reduces its presentation in uninflamed distal lymphoid organs, as shown in mice.

Cancer immunotherapy has become a clinical reality by virtue of the successes of adoptive T-cell transfer and immune checkpoint blockade. Cancer vaccination may follow in their footsteps, as it should provide the ultimate means of harnessing the power of adaptive immunity for the identification and elimination of malignant cells while mounting long-lived T-cell memory responses that prevent relapse. However, efficacious cancer vaccines have proven elusive. Although subunit (or molecularly defined) vaccines can be assembled to contain the three major vaccine elements — antigen, adjuvant and delivery components — a major roadblock in their optimization is the targeting of antigens to the specific locations of the immune system where immune responses are developed (such as draining lymphatic tissue). Yet, despite clever designs and a multitude of strategies, formulating highly immunogenic antigens remains a hurdle.

In vaccine formulations, synthetic peptides can mimic tumour antigens. The peptides that optimally activate tumour-aggressive cytolytic CD8+ T-cell responses in an antigen-specific manner are generally nine to ten amino acids long1, and such short peptides are easy to formulate and manufacture. However, their in vitro antigenicity does not often translate into strong in vivo immunogenicity. In fact, the injection of synthetic peptide antigens may instead lead to the induction of immunological tolerance and to the deletion of the respective antigen-specific T cells in vivo2 through their binding to major histocompatibility complex (MHC)-I molecules on the surface of non-professional, antigen-presenting cells. Hence, to trigger productive immune responses, vaccines based on synthetic peptide antigens need to be selectively delivered to secondary lymphoid organs (lymph nodes) together with ligands for innate-immunity receptors on antigen-presenting cells. An efficient way to deliver peptide vaccines to lymph nodes is to fuse synthetic peptide antigens to poorly immunogenic carrier proteins (so as not to deviate the immune response from the tumour antigen) that, following injection, drain from the subcutaneous space into lymph nodes. Reporting in Nature Biomedical Engineering, Darrell J. Irvine, K. Dane Wittrup and colleagues now show that a biochemical fusion of a peptide antigen and a carrier protein can be optimized for efficient uptake by lymphatic vessels draining the vaccination sites, enabling the protection of antigen epitopes from tissue proteases and reducing undesired antigen presentation in distal, non-inflamed sites3. The immunization of mice with vaccine formulations containing the carrier protein transthyretin enhanced vaccine immunogenicity up to 90-fold and potentiated the immunogenicity of viral antigens, tumour-associated antigens, oncofoetal antigens and shared neoantigens.

Irvine and co-authors first examined the pharmacokinetic properties of minimally immunogenic proteins — self proteins such as serum albumin, pre-albumin (also known as transthyretin) and the Fc portion of antibodies — as carriers for a variety of tumour antigenic peptides. Analyses of the rates of absorption from the injection site into the systemic circulation (kabs) and of clearance from systemic circulation (kclear) showed that a balance must be kept to achieve minimal systemic absorption yet maximal lymphatic uptake while ensuring rapid clearance from circulation (Fig. 1a, top). The authors found that the molecular size of the protein carrier is inversely correlated to kabs, with bulkier proteins having a lower rate of systemic absorption following subcutaneous injection. In fact, molecules less than 4 kDa in molecular weight (equivalent to a peptide antigen of approximately 30 amino acids and thus larger than typical synthetic peptide antigens) were absorbed systemically, which is undesired in vaccine formulations for targeting lymphoid organs; conversely, antigen–carrier fusions were not significantly absorbed systemically. Therefore, controlling kabs via fusion with bulky proteins provides a means of increasing the availability of antigenic fusion peptides in draining lymph nodes. The authors also established that increasing kclear could further improve immunogenicity by reducing the systemic spreading of the immunogen and its exposure to distal non-inflamed lymph nodes (Fig. 1b). Notably, the incremental immunogenicity — measured in terms of frequencies of antigen-specific T cells and their functional competence — was accompanied by an increased control of tumour growth in mice receiving the fusion vaccine. Overall, the authors show that a relatively simple combination of assays (a pharmacokinetic study and a proteolytic stability test with fresh or serum-exposed antigens) can be applied to assess the kabs and kclear rates of protein carriers and thus screen for vaccine formulations with optimal pharmacokinetic properties. These parameters determine antigen biodistribution and, consequently, the relative immunogenicity of the vaccine.

Fig. 1: Optimized antigen–protein fusion for efficient vaccination.

a, Top: pharmacokinetic model for minimally immunogenic proteins such as mouse serum albumin. The protein’s absorbance rate (kabs) and clearance rate (kclear) determine its overall bioavailability after injection and its access to lymph nodes and distal lymphoid organs. Bottom: design of an antigen–protein fusion, with the peptide antigen placed at a terminus of the carrier protein (mouse serum albumin) via a polyhistidine (His) tag and flexible polypeptide protein tags. b, The subcutaneous delivery of small synthetic peptides leads to their systemic spread and to their presentation by virtually every nucleated cell expressing MHC-I molecules. In contrast, the optimized antigen–protein fusion selectively reaches the lymph nodes draining the injection sites, and can be presented by professional antigen-presenting cells. Such divergent fates determine the induction of tolerance (for unselective presentation) or of immunity (for selective presentation). Figure adapted with permission from ref. 12, Springer Nature Limited.

Different strategies for fusion proteins have been explored4,5,6; in particular, a recombinant fusion protein of the conserved influenza M2e antigen fused to human serum albumin and tested in mice induced protective humoral immunity and increased interferon responses when administered as an emulsion in incomplete Freund’s adjuvant7. In this case, the fusion served to increase the bulk of the antigen to reduce uptake into the systemic circulation and thus increase uptake in draining lymph nodes. Others have recently showed that fusing long, synthetic, antigenic peptides to the chemokine XCL1 (also known as lymphotactin) targeted the antigen to cross-presenting cells and induced protective tumour-antigen-specific T-cell responses8; here, the fusion served to target specific antigen-presenting cell populations in draining lymph nodes. In this context, Irvine and colleagues’ pharmacokinetic framework for guiding the further development and optimization of peptide–protein vaccine formulations also shows that systemically dispersed peptide antigens are not merely lost to immunity — they are, in fact, capable of hindering the immune response by inducing immunological tolerance.

The development of cancer vaccines is reaching new heights. A large variety of vaccine delivery platforms are available for testing vaccine immunogenicity and efficacy in well-controlled clinical trials. The realization of the potential of tumour neoantigens in the clinic necessitates efficient delivery vehicles in a personalized immunotherapy context. Therefore, rapidly available and nimble delivery vehicles need to be ready to use with plug-in subunits for the addition of antigens (either shared antigens or personalized neoantigens). In this context, Irvine and co-authors’ optimized antigen–protein fusion offers a way forward. However, antigen–carrier fusion proteins may create neoepitopes that interfere with the desired specificity of the elicited T-cell responses (Irvine and co-authors did not observe such responses). Moreover, the fusion constructs may be preferentially processed in the endocytic compartment, leading to a bias towards the induction of CD4 T-cell responses and, consequently, to the inefficient priming of the required antitumour CD8+ T-cell immunity. Furthermore, antigen–protein fusions crossing to the cytosolic MHC-I antigen-processing compartment may be cleaved in a way that abrogates their immunogenicity. By having the antigen placed at a terminus of the carrier protein (Fig. 1a, bottom), the authors ensured efficient proteolytic processing (Fig. 1b). From a manufacturing perspective, the use of synthetic peptide epitopes implies a trade-off between the simplicity and economy of organic peptide synthesis in polymer or particle formulations, and the challenging yet tractable recombinant expression of the peptide–carrier fusion.

Antigen–protein fusions should improve in design and translatability. Protein domains may be added to provide key characteristics such as chemokine cues that guide migration via the lymphatics, specific targeting domains for antigen-presenting cells or the adhesion to extracellular matrix components9. Also, protein domains could be linked to ligands that stimulate innate immunity10,11. Such fusion constructs may also be encoded in DNA or mRNA, as Irvine and co-authors note, and thus rely on their endogenous production at the desired sites of injection and dispense with the need for manufacturing them under standards of good manufacturing practice. The authors’ design and analytical framework may also extend to immunotherapeutic modalities for infectious diseases and autoimmunity.


  1. 1.

    Rammensee, H. G., Falk, K. & Rotzschke, O. Annu. Rev. Immunol. 11, 213–244 (1993).

    CAS  Article  Google Scholar 

  2. 2.

    Toes, R. E., Offringa, R., Blom, R. J., Melief, C. J. & Kast, W. M. Proc. Natl Acad. Sci. USA 93, 7855–7860 (1996).

    CAS  Article  Google Scholar 

  3. 3.

    Mehta, N. K. et al. Nat. Biomed. Eng. (2020).

  4. 4.

    Francis, M. J. et al. Proc. Natl Acad. Sci. USA 87, 2545–2549 (1990).

    CAS  Article  Google Scholar 

  5. 5.

    Janssen, R. & Tommassen, J. Int. Rev. Immunol. 11, 113–121 (1994).

    CAS  Article  Google Scholar 

  6. 6.

    Suzue, K. & Young, R. A. J. Immunol. 156, 873–879 (1996).

    CAS  PubMed  Google Scholar 

  7. 7.

    Mu, X. et al. J. Virol. Methods 228, 84–90 (2016).

    CAS  Article  Google Scholar 

  8. 8.

    Botelho, N. K. et al. Front. Immunol. 10, 294 (2019).

    CAS  Article  Google Scholar 

  9. 9.

    Ishihara, J. et al. Sci. Transl. Med. 9, eaan0401 (2017).

    Article  Google Scholar 

  10. 10.

    Wille-Reece, U. et al. Proc. Natl Acad. Sci. USA 102, 15190–15194 (2005).

    CAS  Article  Google Scholar 

  11. 11.

    Zom, G. G., Khan, S., Filippov, D. V. & Ossendorp, F. Adv. Immunol. 114, 177–201 (2012).

    CAS  Article  Google Scholar 

  12. 12.

    Kim, S. et al. Nat. Rev. Mater. 4, 355–378 (2019).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Pedro Romero.

Ethics declarations

Competing interests

P.R. is a member of the scientific advisory board of immatics biotechnologies GmbH. J.A.H. is co-founder of Arrow Immune, Inc., and through his university holds patents in immuno-oncology. The remaining author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Romero, P., Donda, A. & Hubbell, J.A. An optimized antigen–protein fusion. Nat Biomed Eng 4, 583–584 (2020).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing