Keystone HIV Pathogenesis and Vaccine Development Report
This year's Keystone Symposia on Molecular Mechanisms of HIV Pathogenesis (X7) and HIV Vaccine Development (X8) were held in British Columbia's Whistler Resort, Canada, from 12-18 April 2004.
This highly specialised and relatively small annual meeting is often not attended by community activists or press. The focus on basic science means that much of what is presented and discussed has little, if any, direct implications for clinical practice. However, the meeting attracts some of the world's experts on HIV immunology and pathogenesis as a forum to exchange and discuss ideas and data. From year to year, the feeling of these meetings can shift from optimism and excitement, to a mundane business-as-usual mood, to an urgent knuckle-down and crack-on intensity. This year's meeting was somewhere between the latter two.
Of significant interest was a presentation entitled "Evolution of HIV is focused in HIV-specific CD4+ T cells" by the group of Dean Hamer at the National Cancer Institute together with the group of Daniel Douek at the Vaccine Research Center, both at the NIH in Maryland.1 Douek has previously shown that HIV-specific CD4+ T cells harbour a large proportion of the pro-viral DNA that makes up the latent reservoir of virus in infected individuals. In this presentation Hamer not only showed that HIV-specific CD4+ T cells are infected, but that they are also activated by HIV itself. CD4+ T cells from patients treated with HAART early in infection (i.e., before any major loss of CD4+ count) were stimulated with HIV antigens p24 and gp120, with CMV antigen, or anti-CD3 (which stimulates all T-cells regardless of specificity), and the replication competent virus induced was sequenced. In addition proviral DNA was also sequenced from purified HIV-specific CD4+ T cells.
The sequences of HIV-1 envelope genes from viruses infecting HIV-specific CD4+ T cells was found to have an 8-12% divergence from the envelope sequences of viruses infecting CMV and anti-CD3 stimulated CD4+ T cells. Phylogenic tree analysis showed that these viral variants were diverse and distinct from viruses populating other CD4+ T cells. In general polyclonal (anti-CD3) stimulated CD4+ T cells appeared to have very homogenous sequences representing the original infecting strains, though during untreated chronic infection virus sequences were found to be highly heterogenous both in HIV-specific and polyclonal (anti-CD3 stimulated) CD4+ T cells. Hamer explained that the HIV present in HIV-specific CD4+ T cells continues to evolve even in individuals who initiated antiretroviral therapy shortly after they became infected.
Mathematical modeling based on these findings suggested that boosting HIV-specific CD4+ T cell frequency could increase viral load and decrease T cell functional help. The argument here is that while highly active antiretroviral therapy may inhibit 99.9% of viral replication, the remaining 0.1% of virus that is replicating, is doing so in HIV-specific CD4+ T cells. The reason for this is logical enough: the population of CD4+ T cells that are most likely to be continually activated in HIV infection are HIV-specific ones, even in the presence of antiretroviral therapy, due to the ongoing presence of HIV antigen which stimulates them. Such activation of these cells subsequently leads to high turnover of the virus they harbour.
Though not particularly surprising, the implications of this data are profound. This may explain why one HIV therapeutic vaccine after another cannot induce sustained HIV-specific CD4+ T cell proliferative responses. Yes, we can induce those responses, but time and time again, they emerge as a transient phenomenon only to mysteriously disappear again. Such short-term responses are the hall mark of short-lived effector T cells which have a half life of 1-2 days, and not central memory T cells which should live for many years. The lack of generation of HIV-specific central memory T cells has been a perplexing mystery for a long time. These are the kind of T cells which protect us from re-infection with measles for example, or from any of the organisms we have been vaccinated against, years after inoculation. Although the jury is still out on the precise origin of central memory T cells, one popular current theory is that a small proportion of activated effector T cells will always survive to subsequently become resting (non-activated) long lived central memory T cells. The fact that this does not seem to happen to HIV-specific CD4+ T cells could be a very important component of why our immune systems fail to control and ultimately eradicate HIV. Indeed Hamer explained that the half-life of HIV-specific CD4+ T cells, once activated, was less than 1 day, suggesting that all HIV-specific effector CD4+ T cells suffer the same fate. Hamer concluded that "The ability of HIV-specific CD4+ T cells to serve as a distinct reservoir for HIV growth and variation suggests that vaccines and treatments aimed at augmenting HIV-specific CD4+ T cell responses should be undertaken with caution." However many immunologists argue that we need to ensure preservation of these responses, perhaps by using more effective HAART regimens, which fully penetrate all anatomical and cellular compartments, thus preventing the small amount of virus replication that is taking place in HIV-specific CD4+ T cells. Indeed these responses need to be expanded in a manner in which they can be sustained, in order to help achieve long-term control of viral replication in the absence of antiretroviral treatment.
Bruce Walker of Massachusetts General Hospital, Boston, Massachusetts, presented an update on his structured treatment interruption study in primary HIV infection.2 Fourteen patients underwent up to three structured treatment interruptions. Treatment was restarted if the viral load increased to more than 5,000 copies/mL for more than 3 weeks or if the viral load increased to more than 50,000 copies/mL on any single occasion. Following interruption, 11 patients (79%) maintained control of viraemia for more than 90 days, despite lack of tissue types associated with protection. 57% achieved control of viraemia for 180 days, 43% for 369 days and 21% for 720 days. However over time there was a gradual decrease in CD4 counts and increase in viral loads. The total magnitude of CD8+ T cell responses increased 3.5, 2.1 and 1.78 fold at the first, second and third interruption and transiently detected HIV-specific CD4+ T cell proliferative responses declined with recurrence of viraemia. Walker concludes that "despite initial control of viraemia, durable immune control in persons following treated acute infection occurs infrequently."
In response to this, Dean Hamer made a passionate request to Walker that he would now denounce the practice of treatment interruptions, acknowledge the potential risks of drug resistant evolution within them, and agree that they offer limited real clinical benefit. However there was little agreement on this and Walker did not seem to share Hamer's view that structured treatment interruptions were dangerous.
One particularly interestingly element of Walker's data was his finding that CD8+ T cell responses measured by the release of the T cell cytokine interferon (IFN)-gamma in the ELISpot assay did not correlate with protection from viraemia in his patients. In contrast, measurement of HIV-1 specific CD8+ T cell proliferation revealed a very impressive correlation with protection from viraemia. Walker used a florescent dye called CFSE to stain CD8+ T cells, which binds to the cell membrane. With every round of division undertaken by proliferating cells the membrane bound concentration of CFSE halves. This assay is increasingly being used to measure cell proliferation in different laboratories. Walkers data using this assay concurs with previous data published by Migueles et al.,3 demonstrating that HIV-1 specific CD8+ T cell expression of the molecule perforin, which kills virus infected target cells, and is known to be deficient in HIV chronically infected individuals, correlates with CD8+ cellular proliferation. Thus while proliferation is coupled to effector function such as perforin production, what we are now experiencing is a gradually dawning understanding that IFN-gamma expression is not part of this picture. In fact we have known for some time that IFN-gamma expression is not linked to cellular proliferation, in the way that other cytokines, particularly interleukin-2 (IL-2) are. The implication here is that the commonly used IFN- gamma assay, now the assay of choice in many immunotherapy and vaccine trials, may not be telling us the correct information about functional T cell responses in HIV infection.
Brigitte Autran of the Hôpital Pitié-Salpétrière, Paris, France, presented the results of the first international, randomised, double blind, placebo-controlled, phase-I therapeutic vaccination trial: QUEST.4 Here 79 individuals with primary HIV-1 infection were treated with HAART >72 weeks before being randomised to one of three immunotherapy arms. Group A continued to receive ART alone, group B received the ALVAC-HIV(vCP1452) therapeutic vaccine in addition to ART and group C received both ALVAC-HIV(vCP1452) and Remune therapeutic vaccines in addition to ongoing ART. ALVAC-HIV(vCP1452) was given I/M at weeks 8, 12, 16 and 20 following randomisation in groups B and C and Remune was given I/M at weeks 0, 4, 12 and 20 following randomisation in group C. In all groups ART was discontinued 24 weeks following randomisation and patients were followed up for an additional 24 week period. The primary endpoint was a viral load <1,000 copies/mL at week 48 (24 weeks after stopping ART) without restarting ART. Secondary endpoints were maintenance of viral load <400 copies/mL throughout the 24 week ART interruption and time to reaching viral load above 1,000 copies/mL after stopping therapy. In all cases restarting HAART was considered failure in the intention to treat analysis.
Preliminary analysis of the data (vaccinated patients in groups B and C have not been unblinded) reveals that while vaccination successfully induced T cell responses measured by IFN-gamma ELISpot, the virological endpoints of this study all failed. In vaccinated patients the median p24 specific CD4 ELISpot response was 180 IFN-gamma responding lymphocytes per million peripheral blood mononuclear cells (PBMCs) (n=32) versus a median of 0 for the ART alone treated group (n=18) (p=0.006). The median CD8 IFN-gamma response to gag was similarly high for the vaccinated patients at 275 IFN-gamma responding lymphocytes per million PBMC (n=34) compared to 0 for the ART-alone treated group (n=18) (p=0.002). Of the 52 vaccinated patients, 15.4% reached the primary endpoint of a viral load <1,000 copies/mL plasma at the end of the 24-week treatment discontinuation period. Of the 27 ART-alone treated patients 22.2% reached this endpoint. There was no statistically significant difference in these values. There was also no statistical difference in the number of patients achieving viral load <400 copies/mL during the ART discontinuation period or the median number of days to a viral load more than 1,000 copies between the ART alone and vaccinated groups.
The fact that vaccination here proved immunogenic in terms of T cell IFN-gamma responses, but yet failed to translate into any discernable clinical benefit further adds credence to the notion postulated by Bruce Walker that IFN-gamma is perhaps the wrong marker of immune function to be measuring in our immunotherapy trials. It is becoming increasingly clear from the published literature that IFN-gamma production is not tied to T cell function in the manner perhaps we once thought it was. Indeed it is possible that because of this, assays measuring IFN-gamma release tend to churn out lots of positive results. These are popular as everyone likes to show positive results. Thus IFN-gamma production assays validate the immunogenicity of various strategies tested, while these responses yield very little clinical benefit because they have limited or no functional impact that could affect long-term clinical outcome. Walker advocates the CFSE dye dilution assay as an accurate measure of HIV-specific CD8+ T cell function. Functional assays for measurement of HIV-specific CD4+ T cells that offer clinically relevant alternatives to singly evaluating IFN-gamma production in the CD4 subset have previously been shown by other groups. Anna Vyakarnam's group at Kings College Hospital, London, demonstrate the superiority of IFN-gamma and IL-2 double positive intracellular staining by flow cytometry5 and Frances Gotch's group at Chelsea and Westminster Hospital also in London demonstrate the superiority of the traditional lymphocyte proliferation assay which measures incorporation of radioactive labeled thymidine into the replicating DNA of proliferating cells.6
If we are to get a handle on useful immune responses that candidate vaccines or immunotherapies should be inducing, we need to be using assays which correlate with clinical outcome. This means that immunology laboratories and investigators need to be a little more adventurous in terms of the assays with which they choose to evaluate their immunotherapy trials. Hopefully the work in this area already laid out by some groups will be verified in larger immunotherapy studies and by other groups in the not too distant future. But until then the incremental acquisition of failing immunotherapy data continues to generate a business-as-usual feel to not really understanding why our chosen immune-based interventions are not working. To this end I returned from British Columbia back home to England and to my London Immunology Lab, with a whole set of new plans for the way ahead, while the bars of the Whistler resort roared to the opening matches of a strange local game called "ice hockey" that was way beyond my comprehension.
Back to the GMHC Treatment Issues May/June 2004 contents page.