Immune memory downloaded

An impressive system for retrieving large numbers of antibodies from memory B cells has been developed. It has been put into practice in an investigation of immune responses to the human immunodeficiency virus.

Infection of an individual with a virus or a bacterium triggers a vigorous response in white blood cells, some of which — B cells — are stimulated to produce antibodies that target the invading pathogen. The antibodies may be produced too late to prevent symptoms of infection, but the next contact with the same pathogen will probably be symptom-free as antibodies are rapidly deployed to clear the pathogen.

This antibody 'memory', which is crucial to vaccine efficacy, has two forms: antibodies circulating in the blood, made by a very-long-lived type of B cell in the bone marrow known as a plasma cell; and B cells in the blood that can be stimulated to make antibodies on contact with a pathogen1,2. The latter 'B-cell memory' carries a record of the antibodies an individual has made in response to a given pathogen, and is of great interest, not least in guiding the design of better vaccines. On page 636 of this issue, Scheid et al.3 describe the detailed characterization of B-cell memory responses in the context of infection with the human immunodeficiency virus. The paper contains insights that are both of a general nature and likely to be specific to HIV.

Dissection of the B-cell memory response in human blood requires individual monoclonal antibodies (specific for particular sites on pathogen molecules) to be isolated from each B cell or each set (clone) of identical B cells. Scheid et al. accomplished this tour de force by selecting single-memory B cells specific for a preparation of the surface glycoproteins of HIV, amplifying antibody genes from each cell and then producing each antibody in a cell line (Fig. 1). In principle, sufficient numbers of B cells were sampled to reflect the full response to the glycoproteins. These glycoproteins were chosen because they are the sole target of antibodies able to neutralize the virus and prevent infection. The authors studied six HIV-infected donors whose blood sera can neutralize, to varying degrees, a range of different isolates of HIV. By analysing the antibody responses of the donors in detail, it was hoped to understand the origins of this broad neutralization.

Figure 1: Retrieving a B-cell memory response: Scheid and colleagues' approach3.

The memory B-cell population is purified from the complex mixture of cells in blood by using specific cell-surface markers. Each memory B cell expresses on its surface multiple copies of a single antibody (Y-shapes). Each antibody can bind to a defined site on a given protein. The use of the protein (yellow) as a 'bait' allows the selection of all the memory B cells that make antibodies able to recognize that particular protein. In principle, any protein from any pathogen could be used for selection to interrogate an individual's memory B-cell response. Single B cells are then subjected to amplification of their antibody genes using the polymerase chain reaction (PCR), those genes then being incorporated into a cell line for producing antibodies. The end result is a large set of cell lines making monoclonal antibodies that can be individually characterized.

Scheid and colleagues did most of their work on four donors, on average isolating more than 100 monoclonal antibodies to the surface glycoprotein preparation per donor. Each antibody was exhaustively characterized at the genetic and protein levels. The antibodies from each donor could be classified into 20–50 families of antibody, with varying numbers of close relatives in each family. The sequences of the antibodies in each family are highly divergent from the sequences characteristically found in antibodies before contact with a pathogen, providing evidence that they are highly evolved to specifically recognize HIV glycoprotein. Constant exposure of the individuals' immune system to HIV over long periods is likely to be a significant factor here. Evolution is further reflected in the high affinities of the isolated antibodies for glycoprotein. Antibodies were found that bind across the whole surface of the glycoproteins, including to sites that have not been described previously.

Scheid et al. attempted to understand the neutralizing activities of the donor sera against a range of HIV isolates in terms of the activities of individual antibodies and combinations of antibodies. They were only partly successful. No single broadly neutralizing monoclonal antibodies were identified, so pools of monoclonal antibodies were tested. The pools for two donors showed neutralizing activity against representative HIV isolates, but only at high concentrations.

So it seems that further neutralizing antibodies remain unidentified in the donors. There are various possible reasons for the failure to find them — a potential disconnect between antibodies made by memory B cells and serum antibodies made by plasma cells in the bone marrow4; dysfunction of the memory B-cell compartment in HIV-infected individuals5; and Scheid and colleagues' use of a glycoprotein 'bait' that may inefficiently select memory B cells making neutralizing antibodies. The design of an optimal bait, which should ideally exactly mimic the conformation of glycoproteins on the surface of HIV, and indeed should thereby be a good vaccine candidate, is a recurring problem in this field.

One additional consideration that might help in understanding broad neutralization, using the approach of Scheid et al., is the increasing access to HIV-infected individuals with exceptional broadly neutralizing serum activity6. The identification of broadly neutralizing monoclonal antibodies that target a sizeable proportion of the huge diversity of global HIV is highly desirable, as this will favour vaccine design7. Four such antibodies are already known to exist and, thanks to novel methods like that of Scheid et al., new ones are certain to be forthcoming. The alternative possibility of a great number of antibodies, each targeting only a few HIV variants, is a less attractive basis for producing a practical vaccine.

In most of the donors studied by Scheid et al., the HIV infection is under control. In some, the virus is kept to such low levels that the individuals concerned are known as elite controllers. But we must stress that there is no convincing evidence that antibody responses are responsible for the favourable clinical course seen in some HIV-infected people8,9. In contrast, however, there is strong evidence that neutralizing antibodies can prevent infection with HIV if those antibodies are present before exposure to the virus10.

The work of Scheid and colleagues is an advance in attempts to clone human antibody responses. It will be interesting to see how the responses identified by this method compare with those obtained from other approaches and sources, such as 'gene rescue' from plasma cells of recently vaccinated individuals11, and from large repertoires or libraries of immune and naive antibodies displayed on the surface of selectable particles such as phage12. It will, of course, also be essential to take any new-won understanding of protective antibody responses at the molecular level and exploit it in designing better vaccines13.


  1. 1

    Wrammert, J. & Ahmed, R. Biol. Chem. 389, 537–539 (2008).

  2. 2

    Dörner, T. & Radbruch, A. Immunity 27, 384–392 (2007).

  3. 3

    Scheid, J. F. et al. Nature 458, 636–640 (2009).

  4. 4

    Guan, Y. et al. Proc. Natl Acad. Sci. USA 106, 3952–3957 (2009).

  5. 5

    Moir, S. et al. J. Exp. Med. 205, 1797–1805 (2008).

  6. 6

    Stamatatos, L., Morris, L., Burton, D. R. & Mascola, J. R. Nature Med. (in the press).

  7. 7

    Karlsson Hedestam, G. B. et al. Nature Rev. Microbiol. 6, 143–155 (2008).

  8. 8

    Pereyra, F. et al. J. Infect. Dis. 197, 563–571 (2008).

  9. 9

    Bailey, J. R. et al. J. Virol. 80, 4758–4770 (2006).

  10. 10

    Mascola, J. R. Curr. Mol. Med. 3, 209–216 (2003).

  11. 11

    Wrammert, J. et al. Nature 453, 667–671 (2008).

  12. 12

    Lerner, R. A. Angew. Chem. Int. Edn 45, 8106–8125 (2006).

  13. 13

    Burton, D. R. Nature Rev. Immunol. 2, 706–713 (2002).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Burton, D., Poignard, P. Immune memory downloaded. Nature 458, 584–585 (2009).

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.