Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Hum Vaccin. Author manuscript; available in PMC 2010 November 3.
Published in final edited form as:
Hum Vaccin. 2010 May; 6(5): 371–372.
PMCID: PMC2943984

Species neutral correlates of immunogenicity for vaccines and protein therapeutics: Fact or Science Fiction

Much like the humanoid Na’vi in the movie Avatar, “species neutral” correlates of immunogenicity have been a hot topic at recent conferences such as the ASM meeting on biodefense (Baltimore, February 18, 2010). While it might seem that developing “species neutral” measures of immune response is critically important for the development of better vaccines and safer therapeutics, like the Na’vi, the concept is closer to science fiction than many would hope.

First, some definitions. What is “species neutral”? From the perspective of vaccination and immunogenicity of protein therapeutics, “species neutral” relates to models that will mimic the progression and kinetics of an immune response regardless of the host species. Similarly a “species neutral” correlate is an indicator of an immune response that is not unique to a host species. For example, antibody titers following immunization or exposure have traditionally been considered species neutral because it has been assumed that if antibodies are generated in the animal model, they will be generated in a natural, human infection.

The hope is that such models could be used to improve the accuracy of pre-clinical immunogenicity studies, allowing drug (and vaccine) developers to identify a set of clinical correlates that will drive products’ progress more rapidly towards approval. After all, what could be better than a simple “correlate of immunogenicity” assay that could move directly from preclinical (murine, non-human primate) into the clinic?

As is often the case, fact is more complicated than fiction. There was considerable debate on the subject at the recent ASM biodefense meeting and discussion about why immune responses in animals might not reflect immune response in humans. The problem is that purported species neutral measurements of immune responses such as antibody titers, and measurement of cytokine ELISA’s may not reflect the immunological determinants that trigger or support the development of the immune response. To be more specific, measurement of antibody or even cytokine response across different species may be easy to do in a species neutral manner but the immunological determinants of titer or cytokine response are likely to be different between species. For example, whilst T cell responses can be stimulated in different host species against the same protein, the T cell epitopes (immunological determinants) that are presented to the immune system in each host are likely to be different. The immunogenic T cell epitopes are determined by factors including MHC restriction which is unmistakably species specific.

Thus “correlates” (of immunogenicity and protection) are dependent on “determinants” and determinants are not species neutral. A correlate is just that – a measure of immune response, and it is not necessarily linked to a shared determinant between species. A determinant is an aspect of the antigen or immune response that contributes directly to the immune response – such as the T cell epitopes contained within the sequence of the antigen, or the breadth of T cell response, or the B cell response, or the antibody isotype. The distinction between correlates and determinants is an important one for drug developers. Regulators often request correlate data without regard to the actual immunological determinants.

For example, the titer of IgG response is the result of a complex cascade of events, which are influenced by a variety of immunological determinants. One of the most important determinants upon initiation of this cascade is the nature of the antigen which in many instances dictates the type of immune response [1]. As the immune response develops it is ultimately the cross-talk between B and T cells (driven by B/T cell epitopes in/on the antigen, respectively) that will affect the specificity, isotype, affinity and longevity of the humoral response [2]. Characterization of factors that influence an immune response requires an understanding of the key immunological determinants as well as measuring the final outcome of an immune response i.e. IgG titer.

Since there is an absolute requirement for T cell help in order to stimulate a high affinity antibody response with B cell memory [3], it is important to assess the strength of the signal delivered by T cells. This can be achieved by measuring the quality as well as the quantity of the T cell response which will ultimately determine the type and strength of the humoral response. Indeed, since measuring B cell epitopes in proteins is notoriously difficult there has been considerable effort made to accurately measure T cell epitopes in protein sequences [4]. Unfortunately, measurement of T cell epitopes in animal models does not always translate to humans, mainly due to the differences in species’ MHC which affect the repertoire of T cell epitopes presented. This also applies with closely related species such as human and non-human primates [56], and MHC restriction can even be a factor when comparing responses among individuals of the same species.

The repertoire of MHC alleles in the human population is incredibly diverse. The highly polymorphic nature of the human MHC locus results in subtle differences in the binding pockets of MHC molecules in humans and this can be the difference between one patient’s body rejecting therapy while another’s accepts it [78]. The importance of MHC alleles in underpinning immune responses is exemplified by the fact that different MHC alleles within a single species can be correlated with either an increased or decreased risk of various autoimmune and infectious diseases [911], this fact alone limits the utility of modeling immune responses in different species (that express unrelated MHC molecules).

Since correlates depend on determinants, what methods exist for measuring the determinants of immunogenicity? T cell response can be predicted using silico immunogenicity screening, in vitro screening using naïve blood donors or exposed subjects [8], and in vivo screening using HLA transgenic mice [1214]. While there has been tremendous success in using each of these techniques, caveats remain. In silico screening of autologous proteins has to be carefully calibrated due to the presence of two types of epitopes (regulatory and effector, see reference 15). In vitro screening with “naïve” blood may give a skewed estimate of immunogenicity due to the effects of heterologous immunity, the process by which exposure to one antigen primes the immune system for a response to another, possibly unrelated, antigen [16]. In vitro culture does not accurately reflect the dynamic cross-signaling that may occur in three dimensional structures, although new culture systems are attempting to replicate that complexity in vitro [1718]. And in vivo studies, carried out using HLA transgenic mice, are hampered by the expression of only a single HLA (MHC allele) per mouse strain whereas humans have multiple HLA alleles. The new BLT Hu-SCID mice, transplanted with human immune tissue, may overcome this problem but the method promises to be tedious as each mouse is carefully crafted by hand [19].

What’s the take home message? Unfortunately, until we can find animals that are transgenic in a way that reflects the diversity and complexity of human HLA, it is likely that “species neutral” models will remain elusive. A single protein may contain multiple HLA “motifs”, or may contain none, and the number of HLA motifs may be completely different from the number of MHC motifs for the selected animal model. Having more or less motifs in the candidate antigen will lead to more or less immunogenicity and differences between species. Measures of immune response like antibody titers and gamma interferon ELISAs are determined by T cells, and while they may be “correlates”, the statistician’s mantra “correlation does not equal causation,” still applies. One must remember that measurements – or correlates - reflect determinants, which differ from species to species. The concept that a determinant of immunogencity might ever be “species neutral” is really just like the Na’vi: a science fiction fantasy.


1. Bachmann MF, Zinkernagel RM. Neutralizing antiviral B cell responses. Annu Rev Immunol. 1997;15:235–70. [PubMed]
2. Gray D, Siepmann K, Wohlleben G. CD40 ligation in B cell activation, isotype switching and memory development. Semin Immunol. 1994;6(5):303–10. [PubMed]
3. Gray D, Dullforce P, Jainandunsing S. Memory B cell development but not germinal center formation is impaired by in vivo blockade of CD40-CD40 ligand interaction. J Exp Med. 1994;180(1):141–55. [PMC free article] [PubMed]
4. De Groot AS, Cohen T, Ardito M, Moise L, Martin B, Berzofsky JA. Use of bioinformatics to predict MHC ligands and T-cell epitopes: application to epitope-driven vaccine design. In: Kaufmann SHE, Kabelitz D, editors. Methods in Microbiology: Immunology of Infection. 3. San Diego, CA: Academic Press; In Press.
5. Bontrop RE, Otting N, de Groot NG, Doxiadis GGM. Major histocompatibility complex class II polymorphisms in primates. Immunol Rev. 1999;167:339–350. [PubMed]
6. Brok HP, Uccelli A, Kerlero de Rosbo N, Bontrop RE, Roccatagliata L, de Groot NG, et al. Myelin/oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis in common marmosets: the encephalitogenic T cell epitope pMOG24-36 is presented by a monomorphic MHC class II molecule. J Immunol. 2000;165(2):1093–1101. [PubMed]
7. Barbosa MD, Vielmetter J, Chu S, Smith DD, Jacinto J. Clinical link between MHC class II haplotype and interferon-beta (IFN-beta) immunogenicity. Clin Immunol. 2006;118(1):42–50. [PubMed]
8. Koren E, De Groot AS, Jawa V, Beck KD, Boone T, Rivera D, et al. Clinical validation of the “in silico” prediction of immunogenicity of a human recombinant therapeutic protein. Clin Immunol. 2007;124(1):26–32. [PubMed]
9. Nepom GT, Erlich H. MHC class-II molecules and autoimmunity. Annu Rev Immunol. 1991;9:493–525. [PubMed]
10. Sanjeevi CB, Sedimbi SK, Landin-Olsson M, Kockum I, Lernmark A. Swedish Childhood Diabetes and the Diabetes Incidence in Sweden Study Groups. Risk conferred by HLA-DR and DQ for type 1 diabetes in 0–35-year age group in Sweden. Ann N Y Acad Sci. 2008;1150:106–11. [PubMed]
11. Mathieu A, Paladini F, Vacca A, Cauli A, Fiorillo MT, Sorrentino R. The interplay between the geographic distribution of HLA-B27 alleles and their role in infectious and autoimmune diseases: a unifying hypothesis. Autoimmun Rev. 2009;8(5):420–5. [PubMed]
12. De Groot AS, Martin WD. Reducing Risk, Improving Outcomes: Bioengineering less immunogenic protein therapeutics. In: Andy Saxon, editor. Clin Immunol. 2. Vol. 131. 2009. May, pp. 189–201. [PubMed]
13. Baker MP, Jones TD. Identification and removal of immunogenicity in therapeutic proteins. Curr Opin Drug Discov Devel. 2007;10(2):219–27. [PubMed]
14. McMurry JA, Kimball SJ, Lee D, Rivera W, Martin DB, Weiner M, Kutzler DR, Sherman H, Kornfeld De Groot AS. Epitope-Driven TB Vaccine Development: A streamlined approach using immunoinformatics, ELISpot assays, and HLA transgenic mice. Current Molecular Medicine, Curr Mol Med. 2007 Jun;7(4):351–68. [PubMed]
15. Weber CA, Mehta PJ, Ardito M, Moise L, Martin B, De Groot AS. T cell epitope: Friend or Foe? Immunogenicity of biologics in context. Series “Optimizing the Future for Biotechnology Therapies, the Key Role of Protein Engineering”, Advanced Drug Delivery Reviews. Sep 30, 2009. pp. 965–976. [PubMed]
16. Mathurin KS, Martens GW, Kornfeld H, Welsh RM. CD4 T-cell-mediated heterologous immunity between mycobacteria and poxviruses. J Virol. 2009;83(8):3528–39. [PMC free article] [PubMed]
17. Giese C, Demmler CD, Ammer R, Hartmann S, Lubitz A, Miller L, et al. A human lymph node in vitro--challenges and progress. Artif Organs. 2006;30(10):803–8. [PubMed]
18. Higbee RG, Byers AM, Dhir V, Drake D, Fahlenkamp HG, Gangur J, et al. An immunologic model for rapid vaccine assessment -- a clinical trial in a test tube. Altern Lab Anim. 2009 Sep;37(Suppl 1):19–27. [PubMed]
19. Wege AK, Melkus MW, Denton PW, Estes JD, Garcia JV. Functional and phenotypic characterization of the humanized BLT mouse model. Curr Top Microbiol Immunol. 2008;324:149–65. [PubMed]