Pushing the envelope: a new dawn for the role of antibodies in immune control of HIV?

Commentary by Gareth Hardy, HIV i-Base

While this years and last years Keystone Vaccine Symposia, specialising in HIV immunology, may have done little to encourage optimism that a successful vaccination strategy is on the horizon, important advances are being made in the understanding of why immunity to HIV is so hard to achieve. As a consequence of this, the sophisticated nature of HIV’s immune evasion techniques may be laid bare, such that successful stratagems can be developed.

Pessimism is being drawn from the results of VaxGen’s phase III trial of AIDSVAX and from the lack of any correlation between the magnitude of a vaccine induced antibody response and any evidence of protection from infection in this and other studies. Before we get too despondent we should take note that the perhaps falsely high expectations of these studies, especially AIDSVAX, have been based on early concepts of how immune responses in general and antibody responses in particular may confer a clinically relevant antiviral effect.

As reflected in the article above, our strategies in the past and still today often focus either on mobilising humoral immunity (antibody responses) or cellular immunity (killer T cells and related cell types). What is gradually becoming evident is that the immune system probably fails to control HIV not because both its main armaments are ineffective, but perhaps instead because it is fighting with one arm tied behind its back.

For many years immunologists have been divided into two camps; those who believe that antibodies that can neutralise the virus will hold the key to successful blockade of HIV replication and those who believe that CD8 killer (cytotoxic) T cells will do the job.

Both camps argued that with just a little tweek here and there, a bit of help and manipulation, things would come right with their respective armamentariums. This perspective is now set to be superceded by an almost Blairite third-way.

Many viral infections are quickly and efficiently eradicated by the combined force of both CD8 cytotoxic T cells and neutralising antibodies, as for example in flu virus infections. It is now becoming clear that in HIV infection this effective concert between the two arms of the immune system breaks down. This is because HIV has evolved a remarkable protective mechanism against neutralising antibodies (antibodies which bind the relevant parts of a protein to inactivate it or the organism to which it belongs). Unaided by the antiviral effect of successful humoral responses, CD8 cytotoxic T cells simply cannot keep up with viral mutation, displaying hyperactivated phenotypes and becoming exhausted.

At last year’s Keystone Symposia David Watkins of the University of Madison, Wisconsin, USA, presented data on a vaccine strategy he is using in apes. This vaccine induced powerful responses from CD8 cytotoxic T cells that protected the apes from the usual very high viral loads experienced following infection. Despite this success it was not long before the little virus that was able to replicate in the presence of this response had mutated sufficiently and the CD8 cytotoxic T cell response became considerably less effective, thus allowing viral loads to increase unchecked in these animals. This of course underscores the problem faced by relying solely on robust CD8 cytotoxic T cells to restrict HIV replication and possibly prevent infection. The story is of course more complex than this because CD4 T helper cells, which pivotally initiate both antibody and cytotoxic T cell responses, are severely impaired in function through poorly characterised means and in number as a direct result of their destruction by HIV. Thus they are unable to effectively mobilise both antibodies and cytotoxic T cells, further impairing the efficiency and capability of both arms of the immune system.

A new understanding of the role of virus-specific antibodies in HIV infection is dawning, such that it is becoming tempting to believe that neutralising antibodies may hold a big part of the key to our previous and present failures with vaccination strategies. If antibodies could be induced which block viral replication, then perhaps the loss of CD4 T helper cells and exhaustion of CD8 cytotoxic T cells could be prevented leading either to indefinite containment of disease progression or otherwise to sterilising immunity.

The new understanding of just why neutralising antibodies fail requires us to appreciate HIV’s envelope protein in minute detail. Robert Doms and James Hoxie of the University of Pennsylvania, Peter Kwong of the National Institutes of Health, USA, and Douglas Richman of the University of California, San Diego, USA have done precisely this and presented some of their findings at last years Keystone Symposia.

Gp120, the outer envelope protein of HIV, directly binds to CD4 and subsequently to other T helper cell surface molecules such as the chemokine receptors CCR5 and CXCR4. In order for gp120 to interact with its receptors CD4, CCR5 and CXCR4 it must have regions in its structure that recognise and bind them, of which the shape and chemical properties must not be subject to change by mutation, as is the case with much of the rest of the protein. These conserved regions of gp120 are the primary sites to which neutralising antibodies should bind and thus block gp120’s ability to interact with CD4 and other cell surface receptors. Antibodies that bind other parts of gp120 do not neutralise it because they do not block gp120’s essential function – to bind receptors on T helper cells – thus these are not neutralising antibodies.

Gp120’s technique for evading neutralization in this way is somewhat multifaceted, but largely relies on a combination of high variability, spatial restriction of conserved epitopes and reconfiguration following CD4 binding. The CD4 binding site on gp120 is a bridging sheet tucked into a deep crevice. Large, highly immunogenic looping regions that constantly mutate from one virus to another obscure this bridging sheet. While a given antibody may bind these looping regions and block the interaction of CD4 and the bridging sheet, easily affordable mutation of these regions in subsequent viral generations renders such a neutralising antibody clone ineffective. Thus the concept of broadly cross strain neutralising antibodies to these available sites is negated.

The CD4 binding site is however highly conserved and must remain so to maintain its integrity. While the CD4 molecule is a slender single chain molecule, antibodies are bulky, consisting of two chains. The crevice in gp120 in which the CD4 binding area resides is narrow, but allows the slender CD4 molecule to slot into position. In contrast antibodies that may be able to block the conserved CD4 binding site and neutralise gp120, regardless of quasi-species variation or mutation in other less important areas of the protein, are unable to gain access to the crevice because of their size.

Following interaction with CD4, gp120 undergoes structural changes, enabling other conserved but formerly masked areas to bind its second receptors: CCR5 or CXCR4. These changes temporarily expose other areas of the protein where antibodies can gain access and thus may be able to cross neutralise the multiple different highly variant gp120 molecules of different viral quasi-species. Such epitopes are referred to as CD4 induced or CD4i epitopes. Because these areas are exposed for a very limited time, such antibodies do not arise naturally. But if these new sites on gp120 can be introduced to the immune system in a stable form, such as in a vaccine, new neutralising antibodies should be induced which can block gp120, regardless of inter quasi-species mutation, and thus broadly inactivate free viral particles in a clinically relevant manner.

Thus the disappointment we experience when viewing VaxGens failure with the whole envelope protein preparation that was AIDSVAX, must tempered as we take on board the implications of our new understanding of the envelope. It is not the quantity of antibody a vaccine induces which is necessarily important in containing or preventing infection, but the quality and specificity of those antibodies. Unless an envelope vaccine presents something new to the immune system (such as reconfigured CD4i epitopes), we should probably expect to be disappointed.

Unfortunately, however, disappointment may not be all we experience with such results. There is some evidence that antibodies that fail to neutralise gp120 may be worse than merely ineffective. Indeed it may appear that non-neutralising gp120 antibodies actually enhance HIV infectivity and/or pathogenesis. Somewhat speculatively, the mechanism of such enhancement may lie in stabilisation of infectious HIV particles on the surface of antigen presenting cells in the lymphoid organs, or increased uptake and infection or disruption of dendritic cell populations.

At last years World AIDS Conference in Barcelona, Harriet Robinson presented work that supported just such a notion: that non-neutralising gp120-specific antibodies enhance HIV pathology (See HTB, vol 3, no. 7. Aug/Sept 2002. Immunology and Basic Science. Behind the headlines about vaccine research, pp 37). As postulated in that edition of HTB, we should be prepared to be disappointed by the results of trials that use whole envelope preparations.

The US military’s announcement at the time that they are to go ahead with the largest phase III trial of an HIV vaccine, in 16,000 volunteers in Thailand, using AIDSVAX as a component of their vaccine strategy, may now be considered with even more caution.


For an excellent recent review of the role neutralising antibodies in HIV infection see: Neutralizing antibodies against HIV – back in the major league. Flavia Ferrantelli and Ruth M Ruprecht. Current Opinion in Immunology. 2002, 14. 495-502.

For a review of the role of the HIV gp120 see: Gp120: Biological aspects of structural features. Pascal Poignard, Erica Ollmann Saphire, Paul WHI Parren and Dennis R Burton. Annual Review in Immunology. 2001, 19. 253-74.

Links to other websites are current at date of posting but not maintained.