Engineering immune evasion

One obstacle to realizing the promise of viral vectors for vaccine delivery is pre-existing immunity to such vectors. An adroit application of structure-based design points to a way around that problem.

There are still no vaccines against such devastating and widespread diseases as malaria, tuberculosis and AIDS. Because the traditional approach of live-attenuated vaccination is not feasible for most diseases, scientists have turned to molecularly engineered viruses that contain pathogen-specific gene inserts. Such viral vectors direct host cells to produce the foreign protein of interest, thus prompting a pre-emptive immune response. Among the most promising viral vectors is a form of common-cold virus known as adenovirus serotype 5. The recombinant adenovirus vectors (rAd5) cannot replicate and can be safely administered, and they elicit both of the two main types of immune response — secreted antibodies and disease-fighting T cells.

“They have taken a fresh approach to the molecular engineering of rAd5, one that has the potential to circumvent anti-vector immunity and expand the applicability of such vectors for human vaccination.”

There is a problem, however. The existing immunity to rAd5 in many adults means that the vector could be neutralized before it can have an effect. Hence the work of Roberts et al.1, described on page 239 of this issue. They have taken a fresh approach to the molecular engineering of rAd5, one that has the potential to circumvent anti-vector immunity and expand the applicability of such vectors for human vaccination.

Numerous viral vectors are being studied for use in gene-based vaccine strategies. The most commonly used vectors are derived from poxviruses, alphaviruses and adenoviruses. Among these, rAd5 is the best characterized and is perhaps the most attractive for vaccine development. As a stand-alone vaccine, rAd5 can elicit different types of T-cell immunity (those due to CD4 and CD8 cells), and more potent immune responses can be achieved with a ‘prime-boost’ approach. For example, use of vehicles known as DNA plasmids followed by boosting with rAd5 can generate durable antibody and T-cell immune responses2,3. Preclinical studies of rAd5 vaccines include vaccines against Ebola, SARS, HIV-1 and anthrax4,5,6,7, and phase II human clinical studies of rAd5 HIV-1 vectors are in progress.

But anti-vector immunity may be a serious limitation. Adenovirus serotype 5 is common — depending on the geographical region of the world, most adults are exposed to it and develop some level of immunity. This may lessen the effectiveness of rAd5 as a vaccine vector. Potential ways around this problem include the use of adenoviruses derived from other human serotypes or from non-human animal species. Indeed, there are more than 50 known human adenoviral serotypes, some of which are quite rare in the human population. The genetic manipulation required to engineer alternative serotypes is not trivial, however. The rAd5 vectors contain specific genetic deletions that render them unable to replicate. This contributes to their safety, but also means that specially engineered cells must be used to produce them. The advantages of rAd5 are that the necessary groundwork has been laid, in terms of basic molecular engineering and production of the vector, and that it has been through the regulatory approval process for use in humans. Adaptation of other serotypes will require a methodical process of research and development, and safety testing. Furthermore, preliminary data5,8 from other serotypes, such as rAd35 and rAd11, suggest that they may be less immunogenic — that is, less effective in producing immunity — than rAd5.

With this as background, Roberts and colleagues1 took advantage of our improved understanding of anti-vector immunity, coupled with structural data about viral proteins, to derive a rational approach to re-engineering the rAd5 vector. In adenoviruses, the viral DNA is surrounded by a protein shell called a capsid that contains hexon and penton subunits. Because host antibodies that neutralize rAd5 are directed against the hexon subunit, Roberts et al. studied the atomic structure of this protein to understand where antibodies would probably bind. Molecular modelling revealed that the seven hypervariable regions (HVRs) of the hexon form the outer surface of the protein, making the HVRs a likely target for antibody binding.

By exchanging all seven HVRs of rAd5 with those of the rare adenovirus serotype 48 (Ad48), the authors constructed a chimaeric adenovirus that could potentially evade the neutralizing antibody response against rAd5. The core structure of the hexon protein was not altered, so the resulting HVR-chimaeric rAd5 vectors retained their ability to grow well in culture and, importantly, the immunogenicity of the chimaeras was comparable to that of rAd5. As Roberts et al. hoped, when the HVR chimaeras were administered to mice or monkeys that had antibody immunity to rAd5, there was no decrease in the immunogenicity of the vector.

These data provide a proof-of-concept that viral vaccine vectors can be engineered to evade pre-existing immunity. The results are a tribute to the application of modern immunology and structural biology to vaccine design. The potential of this technology is considerable. One can envisage the construction of numerous HVR chimaeras that could be used to vaccinate against various pathogens. Thus, if rAd5 itself were used to vaccinate children against malaria, a chimaeric vector could still be used as an HIV vaccine. Furthermore, the use of multiple chimaeric serotypes could allow booster vaccinations to sustain the long-term immune memory response needed for durable immunity.

Yet we are still some time away from studies in humans. Vaccine developers will have to show that these new HVR-chimaeric rAd5 vectors can be manufactured, and that they have stable gene inserts, can pass regulatory review and, finally, are immunogenic in humans with pre-existing immunity. Current rAd5 vectors for HIV-1 are being evaluated in phase II human trials that will more precisely define the extent and effect of pre-existing anti-vector immunity. As we await these data, chimaeric vectors can be manufactured and tested in humans, so that we can further assess the potential effects of anti-vector immunity.


  1. 1

    Roberts, D. M. et al. Nature 441, 239–243 (2006).

    ADS  CAS  Article  PubMed  Google Scholar 

  2. 2

    Casimiro, D. R. et al. J. Virol. 77, 7663–7668 (2003).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  3. 3

    Santra, S. et al. J. Virol. 79, 6516–6522 (2005).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  4. 4

    Bangari, D. S. & Mittal, S. K. Vaccine 24, 849–862 (2006).

    CAS  Article  PubMed  Google Scholar 

  5. 5

    Shiver, J. W. & Emini, E. A. Annu. Rev. Med. 55, 355–372 (2004).

    CAS  Article  PubMed  Google Scholar 

  6. 6

    Barouch, D. H. & Nabel, G. J. Hum. Gene Ther. 16, 149–156 (2005).

    CAS  Article  PubMed  Google Scholar 

  7. 7

    Sullivan, N. J. et al. Nature 424, 681–684 (2003).

    ADS  CAS  Article  PubMed  Google Scholar 

  8. 8

    Lemckert, A. A. et al. J. Virol. 79, 9694–9701 (2005).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Mascola, J. Engineering immune evasion. Nature 441, 161 (2006).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.