May 18, 2007
From November 27-29, 2006, the National Institutes for Allergy and Infectious Diseases (NIAID) held their second workshop on immune activation and HIV pathogenesis, attracting an impressive roster of investigators to the backwaters of Bethesda. The meeting was chaired by Cliff Lane from NIAID and Yves Levy from the ANRS in France, and organized by a committee including Zvi Grossman and Irini Sereti from NIAID, Mike Lederman from Case Western Reserve University and Guido Silvestri from Emory University.
The institution of this annual meeting (the first took place in May 2005) reflects the widespread agreement among researchers studying HIV pathogenesis that immune activation plays a key role in driving progression to AIDS. But while this agreement represents significant progress, there remain many questions as to how the virus causes immune activation, and how immune activation eventually leads to immunodeficiency. At the workshop, 37 different presenters shared data and ideas relating to these questions; this report covers some of the highlights.
Ashley Haase began the workshop by reviewing his studies of the early events following SIV infection in adult rhesus macaques. The key, dogma-challenging finding from this work is that the majority of SIV-infected cells during the first 10 days post-infection are resting CD4 T cells (it had previously been thought that activated CD4 T cells were the primary targets). Haase utilizes a technique called in situ hybridization which involves sampling sections of tissue and staining them with silver grains to identify SIV RNA and a variety of antibodies to identify CD4 T cells and activation markers such as CD69, CD25 and Ki67. Using these techniques to analyze tissue sampled from the colon, resting SIV-infected CD4 T cells can be detected by day 3 post-infection and they continue to represent the majority of infected cells at day 10. Subsequently there is an increase in the number of infected activated CD4 T cells and a concomitant decline in the number of infected resting CD4 T cells; by day 28, activated Ki67-expressing CD4 T cells make up the majority of the infected cell population.
Haase noted that in the gut lamina propria, CD4 T cell numbers are initially stable but decline precipitously (∼70%) after 6 days. Because the number of infected cells cannot fully account for this apparent loss of gut CD4 T cells, Haase and his colleagues have suggested that this represents apoptosis induced by interactions between HIV's gp120 protein and CD4 T cells. Another recent finding by Haase's group is that SIV induces a 15-30 fold increase in the apoptosis of intraepithelial lymphocytes (IEL), a population unique to the gut microenvironment. Haase suggested that this may lead to regenerative villous atrophy, a condition involving damage to the integrity of the gut mucosal barrier.
Haase has also been looking at the role of "regulatory" CD4 T cells using the marker FoxP3; his group has shown that the numbers of FoxP3-expressing cells rises in parallel with markers of immune activation, leading to the hypothesis that these responses prematurely suppress the HIV-specific immune response thereby allowing the virus to take hold (J Infect Dis. 193;5:703-12, 2006).
Offering "Fooling with Mother Nature" as an alternative title to his presentation, Rafi Ahmed discussed recent findings on PD-1, a potential therapeutic target expressed by T cells that have lost the ability to function properly (a phenomenon known as T cell exhaustion). Ahmed's group recently identified expression of the PD-1 receptor as a marker of CD8 T cell exhaustion in mice chronically infected with LCMV. Furthermore, they showed that blocking interactions between PD-1 and one of its ligands (PD-L1) could restore LCMV-specific CD8 T cell function and significantly decrease LCMV viral load. Subsequent work from the labs of Bruce Walker, Rafik Sekaly and Rick Koup has found that PD-1 expression correlates with disease progression in HIV infection and that blocking PD-1 appears to enhance HIV-specific CD8 and CD4 T cell responses in vitro.
However, while these findings may sound promising, Ahmed stressed that PD-1 likely evolved to "switch off" T cells that might otherwise attack self antigens and cause autoimmunity. He noted that early (5-10 days post-infection) blockade of PD-L1 in mice challenged with LCMV was lethal due to severe CD8 T cell mediated immunopathology. These effects were not evident in mice that had been infected for a longer period. Adding to the concern, mice bred to genetically lack PD-L1 also develop fatal immunopathology when infected with a chronic strain of LCMV. But Ahmed pointed out that PD-1 also interacts with another ligand, PD-L2, which is expressed on a far more limited array of cells (dendritic cells and macrophages) compared to PD-L1, and he is working with collaborators Gordan Freeman and Arlene Sharp to investigate whether blocking PD-1/PD-L2 interactions may be an alternative, potentially less risky strategy for reviving exhausted T cells. Encouragingly, mice lacking PD-L2 do not have adverse outcomes when chronically infected with LCMV; preliminary data suggests they have more robust LCMV-specific T cell responses and display superior control of LCMV viral load compared to normal mice.
Ahmed also briefly touched on the potential for combining anti-PD-1 strategies with therapeutic vaccination, showing results of an exploratory experiment demonstrating that anti-PD-L1 and vaccination with the gp33 epitope from LCMV led to a faster decline in viral load than anti-PD-L1 alone in chronically infected mice (vaccination alone had essentially no effect). Next steps for this work include experiments in SIV-infected macaques and also in the chimpanzee model of hepatitis C infection, since recent data has shown that PD-1 expression is also elevated on T cells from individuals with chronic HCV infection.
Mark Feinberg, formerly at Emory University in Atlanta, has conducted extensive research on SIV-infected sooty mangabeys. These animals represent an important model because, like African Green Monkeys, they rarely suffer any adverse immunological or clinical consequences of SIV infection. Yet SIV isolated from sooty mangabeys, when transferred into rhesus macaques, causes simian AIDS within a relatively short period. Feinberg has previously described one key distinction between the species in terms of their response to SIV: macaques experience chronic immune activation (as evinced by increased expression of the activation marker CD38 on T cells and increased T cell turnover) while sooty mangabeys do not.
In his talk, Feinberg described the efforts of his research group to understand this difference. The availability of microarrays allowed the investigators to analyze gene expression profiles of Mangabeys compared to humans and macaques. Feinberg reported that the key finding was that type 1 interferon response genes were strongly upregulated in HIV-infected humans, but not SIV-infected macaques. Type 1 interferons are a class of cytokines with interferon-alpha most prominent among them and the major sources of these interferons are plasmacytoid dendritic cells.
Feinberg went on to find potentially important distinctions between the species in terms of the type 1 interferon response to the virus: Mangabey peripheral blood mononuclear cells (PBMC) do not produce interferon alpha when exposed to inactivated SIV, in contrast to macaques, despite the fact that cells from both species respond similarly to influenza. In order to try and discern if the production of type I interferons contributes to disease pathogenesis, as these data imply, Feinberg suggested that strategies to block this response should be evaluated as therapies.
Chip Sheppard reviewed a theory he and colleague Michael Ascher proposed many years ago (dubbed the "panergic imnesia" hypothesis at the time). The general idea is that HIV, by binding to CD4 and CCR5 on T cells, sends signals into the cell which cause the deletion of the CD4 T cell from the repertoire. Sheppard cited extensive literature supporting this hypothesis, including data on other models of retrovirus-induced negative selection, but the reception of the idea by workshop attendees could probably best be described as frosty. Sheppard believes that trials of CCR5 antagonists may offer a means to dissect whether HIV-mediated negative selection of CD4 T cells is occurring.
Daniel Douek presented his work on bacterial products as a potential cause of immune activation in HIV infection. Douek is a leading proponent of the theory that the direct cytopathic effects of HIV cause an early depletion of gut CD4 T cells, and this depletion is central to disease pathogenesis. Seeking to explain how such CD4 depletion leads to chronic immune activation, Douek has suggested that the integrity of the gut mucosa is compromised, which in turn allows commensal "friendly" bacteria to leak into the circulation (a phenomenon known as microbial translocation). At the workshop, Douek reported that plasma levels of lipopolysaccharide (LPS), which are typically used a quantitative indicator of microbial translocation, are significantly elevated in people with chronic HIV infection and AIDS. The same held true for 10/11 SIV-infected macaques studied. Furthermore, in vitro experiments showed that chronic stimulation of monocytes/macrophages with LPS increased secretion of soluble CD14 which was also shown to be elevated in people with HIV. One exception was people with acute/early HIV infection, who did not show elevated LPS. Douek presented data suggesting that this is explained by the presence of anti-LPS antibodies (called EndoCAbs), which wane over time as people with HIV become more immunodeficient.
Douek also showed that LPS levels correlated well with CD38 expression on CD8 T cells, a marker of immune activation. In a small cohort of HIV-infected individuals treated with ART, the majority experienced a decline in LPS levels that correlated inversely with their CD4 count increase, although there was no correlation between LPS levels and CD4 counts prior to the initiation of therapy. Douek reported on a preliminary attempt to confirm that the LPS being measured was the product of microbial translocation: two SIV-infected macaques were treated with antibiotics and showed a brief (∼2 week) decline in LPS levels prior to the overgrowth of resistant bacteria.
Addressing one of the challenges to his hypothesis, Douek noted that -- as mentioned as an aside by Mark Feinberg during his talk -- SIV-infected sooty mangabeys also show depletion of CD4 T cells from the gut, despite the fact that few of these cells express CCR5 and the animals rarely develop immunodeficiency. Douek reported that Mangabeys show a transient increase in LPS levels during acute SIV infection that rapidly normalizes. He suggested that, in this species, mucosal integrity is somehow maintained or microbial translocation is contained, for as yet unknown reasons. During the question and answer period, several attendees raised the possibility that the LPS in people with HIV could be coming from other sources, such as pathogens. Plans for future studies include using PCR to look for bacterial DNA directly, and such studies may help resolve the question of whether the elevated LPS levels in people with HIV truly reflects microbial translocation from the gut.
Mike McCune discussed the potential role of an enzyme called indoleamine-2,3-dioxygenase (IDO) in HIV infection. IDO is responsible for metabolizing tryptophan and McCune cited data showing that its activity is increased in a range of chronic inflammatory diseases, both autoimmune and infectious. McCune and his colleague Jeff Mold found that IDO activity is increased approximately 5-fold in the tonsils of people with progressive HIV but not in long-term non-progressors. There was a trend toward an association with markers of immune activation (HLA DR and CD38) but it did not quite reach statistical significance. McCune noted that regulatory T cells (Tregs) have been reported to increase IDO in antigen-presenting cells and, in line with this finding, an experiment in SIV-infected macacques demonstrated that blockade of CTLA-4 (a molecule expressed on many Tregs) reduced IDO activity. In a collaboration with Chris Miller, McCune also found that IDO was significantly upregulated in SIV-infected macaques but not in SIV-infected African Green Monkeys (a species that does not develop immunodeficiency as a consequence of SIV infection). McCune closed by suggesting that inhibition of IDO should be explored as a potential therapeutic strategy.
Rene van Lier presented a review of his data demonstrating a connection between immune activation and immunodeficiency in mice genetically manipulated to express the co-stimulatory molecule CD70 on B cells. As reported in Nature Immunology in 2003, these mice experienced persistent CD4 and CD8 T cell activation leading to the excessive generation of effector/memory CD4 and CD8 T cells and the gradual depletion of the naive CD4 and CD8 T cell pools. Notably, these mice eventually developed fatal pneumocystis carinii pneumonia, suggesting their memory T cell pools were also compromised. Van Lier pointed out that CD70 expression is enhanced in HIV infection and that interactions between CD70 and its receptor CD27 are known to drive naive T cell activation and differentiation. He also showed that in the transgenic mice, monocytes were strongly activated and showed a dramatically decreased potential to migrate to tissues, which he suggested also likely contributed to the immunodeficiency observed in this model.
In order to investigate whether the effects seen in his initial experiment involved only CD70/CD27 interactions or also involved signaling through the T cell receptor (TCR), van Lier created a transgenic mouse that only expressed CD70 on T cells. In this model, only CD8 T cells were activated and differentiated into effector/memory T cells. Because murine T cells only express MHC class I (which interacts with the CD8 T cell receptor but not the CD4 T cell receptor), van Lier concluded that signaling through both the TCR and CD70/CD27 must be required for the excessive activation seen in his original study to occur.
Marc Hellerstein began with a clear statement of his main research interest, which is to "measure interesting things in living systems, using kinetics." Hellerstein has pioneered two very important techniques for evaluating T cell kinetics in humans, deuterated glucose labeling (which uses an intravenous infusion of label) and heavy water, which is simply drunk on a daily basis and therefore permits long term labeling studies to be carried out. As Hellerstein has already shown in his published work, these techniques allow detailed analysis of the turnover of labeled cells; in HIV infection, he found that there is a consistent decrease in the proportion of long-lived naive and memory T cells and an increase in the proportion of short-lived effector T cells (J. Clin. Invest. 112:956-966, 2003). In his workshop presentation, Hellerstein described his use of the heavy water labeling to explore potential sex differences in HIV pathogenesis. Based on data suggesting that women may progress at higher CD4 counts and lower viral loads, Hellerstein compared kinetic data on T cell turnover (obtained on a yearly basis) from 14 men and 17 women with CD4 counts over 500. The results showed that women had higher T cell turnover rates than men despite both groups having CD4 counts within the normal range, suggesting sex differences in immunopathogenesis that are not captured with standard surrogate markers.
Hellerstein also described two additional uses of labeling that his research group is exploring. The first is based on the hypothesis that slowing T cell recruitment into lymph nodes may reduce activation in HIV infection without necessarily leading to immunosuppression. In collaboration with Robert Busch, mouse studies have been conducted to evaluate whether particular drugs can reduce recruitment without limiting T cell division and expansion. Hellerstein stated that standard immunosuppressive agents (e.g, hydroxyurea) inhibited both recruitment and expansion, but a screen of a number of compounds uncovered one that reduced only the former. Hellerstein is pursuing further studies with this FDA-approved drug (he is not permitted to name it, however). The second novel strategy involves using heavy water to label the synthesis of collagen in order to measure fibrosis formation, which is known to occur in the lymph nodes in HIV infection. Hellerstein reported that the technique has already been successfully applied to liver and kidney fibrosis in experimental systems.
Frank Miedema delivered a lively and provocative talk describing his efforts to finally pin down the half lives of naive and memory T cells in humans using the heavy water labeling technique developed by Marc Hellerstein. He stressed that the data took 8 years to collect, and set about convincing the audience of its veracity. Miedema noted that the seminal paper on this subject was by Michie et al in 1992; this paper reported on the lifespans of CD45RA and CD45RO expressing T cells in individuals after radiotherapy. The results showed that CD45RA cells (considered naive) had a half life of 630 days whereas for CD45RO (considered memory) cells it was 182 days.
Miedema went on to describe his study, which involved a 9 week labeling period with heavy water and a stricter definition of naive and memory CD4 and CD8 T cells using additional markers. A mathematical analysis was applied to interpret the uptake and loss of label by the different T cell populations. The results were as follows:
|CD4 T Cells||CD8 T Cells|
|Naive ½ life||4.2 yrs||6.6 yrs|
|Memory ½ life||0.4yrs||0.7yrs|
Miedema also found that recently produced naive T cells had a much longer half-life than the average for naive T cells generally. Miedema stressed that this finding suggests that, while the thymus in humans only produces a relative trickle of new naive T cells in adulthood, these newly produced T cells are preferentially incorporated into the naive T cell pool at the expense of existing naive T cells. Miedema pointed out that this finding differs from that in mice, where recently produced naive T cells appear to turn over more rapidly than the pool as a whole.
The theory that depletion of gut CD4 T cells is central to HIV pathogenesis has gained significantly in popularity over the past few years. The mathematician Rob de Boer decided to court controversy by opening his presentation with a salvo at some of the assumptions that underpin this theory. De Boer noted that it is often stated that the majority of the CD4 T cells in the body reside in the gut. It turns out that this is untrue. He cited a recent review by Reinhardt Pabst and Hermann Rothkötter which states unequivocally: "It is often stated that lamina propria lymphocytes represent the largest lymphocyte pool in humans. This is not the case. Only when the number of plasma cells and in particular IgA-producing plasma cells are counted, does the lamina propria achieve such an outstanding position for plasma cells." Vitaly Ganusov, a postdoc in de Boer's lab, reviewed the primary literature and estimated that Peyer's Patches contain about 3% of the CD4 T cells in the body, while approximately 7.4% reside in the lamina propria.
De Boer also cited a paper by Sebastian Sopper from several years ago (Blood 101;4:1213-9, 2003) which sampled approximately 50% of total body CD4 T cells in a large cohort of SIV-infected macaques and found that, during the asymptomatic phase of the infection, absolute numbers of CD4 T cells are slightly increased compared to uninfected macaques, a finding incompatible with the notion -- proposed by Danny Douek and colleagues -- that SIV (and by extension, HIV) depletes more than half the memory CD4 T cell pool within weeks of infection (Sopper reported that absolute numbers only began to crash at the onset of AIDS). Sopper's paper is to this day the most comprehensive analysis of total T cell numbers in SIV infection and yet, as Mike McCune noted during the workshop, it is unfortunately rarely cited.
Moving on to the main portion of his talk, de Boer described the difficulty of interpreting data relating to T cell receptor excision circles (TREC) in HIV infection. Initially proposed as a marker of recently produced naive T cells, subsequent analyses have demonstrated that many different mechanisms can affect TREC levels: shifts in the ratio of naive and memory T cells, dilution of TREC by T cell division, increased activation and death of naive T cells, and impaired export of TREC-containing T cells from the HIV-infected thymus.
To try and get a clearer picture of what is happening to TREC in HIV infection, de Boer and colleagues collected data longitudinally from people before and after seroconversion. They found that the absolute numbers of TREC fell by about 50% between the pre-seroconversion to one year post-seroconversion timepoints, a loss which outpaced the much smaller decline in naive T cell counts. Over prolonged (2-5yr) followed up, TREC numbers stabilized as their loss paralleled the slow loss of naive and memory T cells from the peripheral blood. De Boer constructed a mathematical model to help interpret these data, which suggested a large proportion of naive T cells are driven to differentiate into memory/effector T cells during acute infection. However, at the end of his talk, a number of workshop participants pointed out that de Boer's actual actual data on loss of naive T cells appears less dramatic than the model predicted.
Derya Unutmaz described a population of CD4 T cells (dubbed effector/memory T cells) which express the CD45RA molecule typically associated with naive T cells and appear highly resistant to infection with CCR5-tropic HIV. Work in his lab suggests that the cells are able to block HIV replication at an early step after the virus entry. Unutmaz was interested to learn if having a larger proportion of effector/memory CD4 T cells expressing CD45RA might affect markers of HIV disease progression. Preliminary analyses suggest people with more effector/memory CD4 T cells expressing CD45RA versus RO have lower viral loads and higher CD4 counts, an intriguing finding that Unutmaz's group is now following up on.
Martin Meier-Schellersheim and Zvi Grossman from NIAID delivered a paired set of presentations on the subject of T cell dynamics in SIV infection, based on their collaborative work in analyzing and interpreting in-vivo DNA labeling data from Louis Picker's group (see also Picker's presentation reported below). Meier-Schellersheim described some of the challenges researchers face when interpreting labeling experiments using the radioactive label BrdU. In particular, he pointed out that while BrdU is eliminated from the system within a day after administration, 35% of peripheral blood CD4 and CD8 T cells in chronically SIV-infected rhesus macaques (RM) take up label during this short period. But extending the administration period to three weeks only increases the percentage of BrdU-labeled T cells to 60%. His explanation for this phenomenon is that a large proportion of the cells that take up label are short-lived, or have a short residency in lymph nodes and blood, and therefore quickly reach an equilibrium in which the rate of labeled cell disappearance equals the labeling rate. When BrdU administration stops, T cells that continue to divide rapidly become BrdU dim and then BrdU negative. During 1-2 days post-labeling, the ratio of bright-to-dim BrdU-labeled circulating cells drops and then stabilizes when dividing cells dilute their label beyond detection, whereas Ki-67 expression in a cohort of BrdU-labeled cells that stopped dividing is rapidly lost. Meier-Schellersheim's careful mathematical analysis of this dilution process, along with that of the loss of Ki67 expression by cells that already stopped dividing, suggests that the turnover of T cells in the lymph nodes is time-structured, where a population of rapidly dividing cells (dividing up to twice a day) gives rise to at least three discernable populations of non-dividing CD4 T cells in these macaques with half-lives of around 2 days, 10 days and several weeks, respectively.
Zvi Grossman then took over and offered a potential explanation for the BrdU data. In Grossman's view, it likely reflects overlapping local bursts of T cell activation, expansion and contraction that are initiated in the lymph nodes in response to antigen. Labeled CCR5+ and CCR5- cells disappear from blood at various rates as CCR5+ cells peak, and then decay, in bronchoalveolar lavage (BAL). Grossman noted that direct viral killing did not measurably affect effector memory CD4 T cell dynamics in BAL. Because fractional T-cell changes in blood and lymph nodes mirrored each other, and CD4 T-cell dynamics in BAL closely correlated with those documented in the lamina propria by Louis Picker, mathematical modeling was used to integrate blood and BAL data. This analysis showed that in SIV-infected rhesus macaques, the BAL (and by inference, other mucosal "effector" sites), essentially contain only short-lived CCR5+ effector-memory T cells (TEM) that depend for maintenance on a continuous TEM production from rapidly proliferating central memory cells (TCM). Indeed, the short lived T cells measured in blood mostly represent activated effector T cells expressing CCR5 that migrate to the gut and other mucosal sites. The influx of TEM to mucosal tissues is reduced several fold in infected rhesus macaques and progressively decreases, secondary to a reduced -- and declining -- production (see Picker's presentation reported below). In summary, Grossman concluded, the ability of people with HIV to generate mucosa-bound effector T cells becomes progressively compromised during the chronic phase of the disease, and this appears to be a consequence of the loss or dysfunction of the populations of long-lived resting T cells that can give rise to CCR5-expressing effector T cells upon activation (rather than the infection and killing of these latter cells in situ by HIV).
Victor Appay outlined work conducted in his lab exploring the hypothesis that HIV infection progressively exhausts cellular immunity. Appay presented a schema showing the typical stages of a T cell's life, from emerging from the thymus as a naive T cell to activation and differentiation into a central memory T cell to subsequent rounds of activation and proliferation that can eventually lead to a stage known as "terminal differentiation" and, ultimately, exhaustion. Appay reported that his group has analyzed gene expression profiles and identified clusters of genes associated with these progressive stages of differentiation. He also noted that telomeres, an indicator of cellular lifespan, progressively shorten as T cells progress through these stages.
Looking at the bigger picture of pathogenesis, Appay noted that the persistent immune activation seen in HIV infection leads to a gradual attrition of the naive T cell pool and an accumulation of terminally differentiated and exhausted T cells. He described a longitudinal analysis of a particular HIV epitope-specific CD8 T cell that dominated the response in one infected individual; at the time of the first analysis this CD8 T cell response was already showing signs of terminal differentiation as defined by the upregulation of the marker CD57. The same individual analyzed 3 years later still had this CD8 T cell response, but it was no longer dominant. A response to a different HIV epitope had emerged and now dominated the CD8 T cell response but, seemingly in turn, this CD8 T cell response was also showing signs of terminal differentiation. Appay's finding, which he and his colleagues are now investigating further, suggests that accumulation of exhausted T cells contributes to the progression of disease in people with HIV.
For several years now, Louis Picker has been focusing on delineating the key events in HIV pathogenesis using the SIV/macaque model. Picker outlined the strategies he has employed: comparing rapidly progressing animals to slower progressors, and also studying animals infected with attenuated SIVmac239deltanef versus those challenged with the virulent parental SIVmac239 virus. In work published back in 2004, Picker reported that the loss of the immune system's capacity to generate tissue-seeking, CCR5-expressing effector memory CD4 T cells (TEM cells) was a distinguishing feature of rapid disease progression. Picker concludes that the lack of these cells at mucosal surfaces is the proximal cause of opportunistic disease which, he notes, almost always first manifest at distal sites such as the lung, gut, brain, etc. At the workshop, Picker described his efforts to understand why this failure to generate TEM cells occurs.
Picker began by noting that TEM are generated by the activation of resting naive T cells or, more commonly, by the activation of resting central memory T cells (TCM). Activation is accompanied by the upregulation of CCR5, allowing HIV and SIV to exploit this aspect of CD4 T cell differentiation as a means to infect activated cells. Picker's new data suggests that the effects of SIV on TEM are (at least) twofold. There is an initial severe impact of pathogenic SIVmac239 infection on the proportion of TEM cells in extralymphoid tissues such as the gut and bronchioalveolar lavage (BAL) during the acute stage, and this decline in TEM cells continues during chronic infection (Picker calculated their half-life to be around 107 days). Simian AIDS ensues when the proportion of TEM declines to less than 0.4-0.5% of T cells in these tissues.
To try and ascertain what is driving these losses (e.g. direct killing, immune activation, or some other cause), Picker conducted BrdU labeling experiments in infected animals which showed that BAL TEM populations are highly dynamic. A short 24-hour labeling period led to a peak of BrdU uptake in proliferating CD4 T cells in blood and lymph nodes after around 2-3 days. Five days post-BrdU administration, 25% of BAL CD4 TEM cells were labeled, indicating that they represented a substantial influx of CD4 T cells that had previously acquired the BrdU label elsewhere. These BAL TEM cells were no longer proliferating as they did not express the proliferation marker Ki67. Picker explained that these data clearly demonstrate that, in SIV infection, extralymphoid CD4 TEM cells are short-lived and their numbers are maintained by a constant influx of activated CD4 T cells from other sites. Picker also showed that the pattern for CD8 TEM cells is the same in chronically infected macaques, arguing strongly that immune activation -- not direct killing -- is driving the turnover of both CD4 and CD8 TEM cells. The conclusion Picker drew from these data is that loss of TEM in the acute phase of SIV infection and subsequent immune activation causes extralymphoid TEM populations to become heavily dependent on the influx of new cells, leading to the question: why does this influx dwindle over time?
To address the question, Picker looked at the effects of SIV infection on central memory T cells; upon activation, it is these TCM cells that normally proliferate and generate tissue-seeking TEM. This differentiation process is accompanied by CCR5 upregulation, leading to the suspicion that SIV infection may directly compromise the generation of TEM. Comparing data from macaques infected with attenuated SIV to those harboring pathogenic SIV, Picker was able to confirm this suspicion: although the proliferation of memory CD4 T cells (based on Ki67 expression) in both groups was similar, 37% of the cells were differentiating into TEM in attenuated SIV infection versus only 9.8% in pathogenic infection. Provision of antiretroviral therapy rapidly normalized TEM differentiation, suggesting to Picker that direct viral killing may be responsible for this phenomenon. However, he also noted that this block in TEM differentiation did not worsen over time and thus could not fully account for declining TEM numbers. What did change over time was the absolute number of proliferating CD4 TCM cells, which showed a steady and significant diminution in progressive SIV infection but not in animals infected with attenuated SIV (these changes were not seen among CD8 TCM).
The totality of the evidence has led Picker to conclude that it is this failure in CD4 TCM homeostasis that underlies the decline in CD4 TEM production, which ultimately leads to AIDS. Although CD4 TCM express little CCR5, Picker showed that there is substantial infection and depletion of these cells from acute SIV infection onwards. Picker proposed a "two tier" model of progressive infection where there is an early and dramatic effect on CD4 TEM followed by a more covert and less efficient depletion of CD4 TCM. He also argued that persistent immune activation is critical to pathogenesis for a number of reasons: it decreases TEM lifespan thereby necessitating increased TEM production, it stimulates CD4 TCM proliferation and differentiation (which increases the susceptibility of CD4 TCM to infection and may also lead to proliferative exhaustion) and it induces tissue inflammation and fibrosis, harming the lymphoid tissue niches that are required to sustain CD4 TCM homeostasis. Picker closed by proposing that measurements of CD4 TCM numbers and functionality may provide a gauge of disease progression in HIV-infected individuals and also that therapies that may beneficially impact CD4 TCM -- such as IL-7 -- deserve careful clinical evaluation.
During an evening dinner presentation, immunologist Ron Germain from the National Institute of Allergy and Infectious Diseases (NIAID) showed some of the latest results obtained using imaging techniques to assess the behavior of immune system cells in vivo. The techniques involved are fairly recent innovations which allow the activity of mouse immune system cells to be visualized in vivo or in excised tissue sections using special microscopes and fluorescent labels to tag the cells and tissues of interest. In one study, Germain used the technique to monitor the movement of T cells in lymph nodes. He showed remarkable movies which demonstrated that, in contrast to initial suggestions that T cells wander randomly in these tissues, they in fact move along string-like pathways of fibroblastic reticular cells (FRC). These pathways form a complex traffic system with crossroads, junctions and dead ends. Dendritic cells, whose job it is to present antigens to potentially responsive T cells, gather at the crossroads like streetcorner salesmen plying their wares. The dead ends of the FRC pathways act to prevent T cells from being able to travel into the B cell zone of the lymph nodes. Germain also showed that B cells have their own similar traffic system, the follicular dendritic cell network in the B cell zone of the lymph node. This study has now been published (Immunity 25;6:989-1001, 2006) and the Quicktime movies are available free online at: www.immunity.com/cgi/content/full/25/6/989/DC1/
In a second study, the techniques were used to visualize the behavior of dendritic cells in the gut mucosa. While it is known that dendritic cells can reach through the epithelial barrier of the gut to sample antigens in the GI tract, the precise nature of this process has not been well elucidated. Germain's stunning images illustrated how the long tentacles of dendritic cells can balloon out through epithelia to grab passing antigens before seeming to deflate back behind the mucosal barrier. Additional experiments also demonstrated that the prevalence of this activity varies in different locations in the GI tract, and that it was robustly enhanced by microbial stimuli (such as Salmonella infection). Given that dendritic cells can capture HIV and act as Trojan horses that transport the virus to CD4 target cells, these findings may be very relevant to the mucosal transmission of HIV infection. These movies are also now available free online as part of the published paper (J Exp Med 203;13:2841-52, 2006): www.jem.org/cgi/content/full/jem.20061884/DC1
Germain emphasized that imaging techniques are not free of challenges; adequate resolution is needed to observe the events of interest (the reason initial lymph node imaging studies suggested random T cell movement was that FRC simply weren't visible) and great care must be taken to minimize physical perturbation of the tissues that are being imaged. Germain believes targeted in vivo imaging using micro MRI (magnetic resonance imaging) will soon allow immune system cells to be visualized in larger animals and humans.
Bob Siliciano's research group has specialized in identifying and characterizing HIV that persists in infected individuals despite long term suppressive ART. Siliciano was the first to publish data on the long-lived reservoir of infected resting memory CD4 T cells which dashed early (and over-optimistic) hopes of ART-mediated viral eradication. At the workshop, Siliciano described the surprising results of a new study that aimed to provide a more detailed description of the viruses that persist in the face of effective ART. Samples were taken every other day from individuals with well-suppressed viral loads and the genetic sequences of HIV from plasma was compared to that found in resting CD4 T cells. What Siliciano found was that most individuals displayed what he described as a "predominant plasma sequence" (PPS). These viruses were completely homogenous and in one case made up the vast majority of detectable plasma sequences in 161 samples taken over a period of two years. Furthermore, these sequences were rarely detected in infected resting CD4 T cells, suggesting a different source. Siliciano's hypothesis is that these viruses derive from a rare infection event involving a monocyte/macrophage progenitor cell; the integrated viral genome is copied without error and distributed to many progeny cells which migrate to tissues and become a permanent source of the homogenous viral transcripts detected in his study.
Steve Deeks from UCSF described his recent work with a cohort of "elite controllers." Deeks defined these individuals based on their having a viral load less than 50 copies/mL for at least two years in the absence of any therapy. So far, 50 such individuals have been identified; the average age is 47, 61% are male, 64% are of non-Caucasian ethnicity, average CD4 count is 756 (range: 528-1039) and duration of HIV infection is 14 years (range: 10-17 years). An analysis of HIV-specific T cell responses in the cohort revealed that they have significantly more HIV-specific CD4 T cells capable of making IL-2 and IL-2+interferon gamma compared to individuals with progressive disease; HIV-specific CD4 cells in the latter group typically make interferon gamma alone. These CD4 responses also seemed less differentiated in the elite controllers. HIV-specific CD8 T cells showed a similar pattern of differences. Deeks cited the classic dilemma raised by these data: is the presence of these seemingly more polyfunctional T cell responses a cause or effect of the low viral loads in elite controllers? Deeks stressed that he believes the relationship is causative, because viral suppression with ART does not lead to the generation of comparable HIV-specific CD4 and CD8 T cells. Also, protective HLA alleles -- most notably B57 -- are far more common among elite controllers, providing strong evidence that CD8 T cell responses are important in many of these individuals.
Deeks has also looked at levels of immune activation in his cohort. The activation markers CD38 and HLA-DR are far less common on their CD4 T cells compared to individuals with progressive disease, but CD38 expression on CD8 T cells is slightly elevated compared to individuals on ART with suppressed viral load. Deeks also uncovered a positive correlation between Gag-specific CD4 T cell responses and the number of CD8 T cells expressing CD38 and HLA-DR.
While T cells appear to be involved in the elite control phenomenon in a substantial proportion of cases, Deeks also noted that there are some cases where these immune responses are not easily detectable (which he dubs "non T-cell elite controllers"). Ongoing work is therefore aiming to identify additional factors that contribute to the robust restriction of HIV replication in these individuals. Deeks suggested innate immunity may play a role, and there is some evidence for increased copy numbers of the gene for CCL3L1, a potently HIV-suppressive chemokine that is a ligand for CCR5. Other potentially protective chemokine receptor genotypes do not appear to be enriched in the cohort, however. Deeks closed by citing the large collaborative study of elite controllers now being spearheaded by Bruce Walker; this effort, Deeks believes, will allow studies of sufficient power to truly tease apart the many potential factors that may be in play.
The workshop closed with a talk by Mike Lederman, who heads up a collaboration involving many different researchers interested in the role of immune activation and other under-studied phenomena in HIV pathogenesis. The group has been informally known as the Bad Boys of Cleveland (BBC) -- even though there are women involved -- but recently has adopted the more staid moniker of the Cleveland Immunopathogenesis Consortium (CIC). Lederman pointed out that data on the role of immune activation in HIV infection dates back well over a decade, particularly to the late Janis Giorgi's work correlating CD38 expression on CD8 T cells with the pace of disease progression. But in 2006, the actual mechanistic basis for this relationship still remains unknown. The CIC has been involved in a number of recent studies that may begin to shed light on this mystery, including the microbial translocation work of Danny Douek and a recent (and controversial) paper in JAMA which reported that viral load measurements are actually a poor predictor of near term CD4 T cell decline (suggesting other factors like HIV-driven immune activation are important). Because microbial translocation may also cause stimulation of the immune system via interactions between bacterial products and toll-like receptors (TLRs), Lederman has been investigating the effects of TLR stimulation on T cells in vitro. He presented preliminary data showing that TLR ligands can cause increased expression of CD38 on memory and effector CD8 T cells although HLA-DR (another activation marker upregulated in HIV infection) was unaffected.
Lederman's current thinking is that the translocation of microbial products through damaged gut mucosa likely contributes to the immune activation and cell turnover seen in HIV infection, and the CIC has a number of studies in the works that will explore this hypothesis further. The CIC is in the process of applying for a large PO1 grant from the NIH to help support these studies, and, in the discussion session at the end of the workshop, CIC member Steve Deeks explained that a key part of this effort will involve conducting small (30 people per arm), randomized 12-16 week clinical trials of interventions that might beneficially modulate pathogenesis and also shed light on whether the CIC's hypotheses are valid. Deeks listed the many interventions under consideration:
Additional workshop participants also contributed to the closing discussion. Frank Miedema and Brigitte Autran made it clear that they remain to be convinced about the importance of the gut CD4 depletion/microbial translocation theories. Derya Unutmaz had the last word, noting the importance of studying T cell activation and differentiation in HIV and stressing that researchers need to parse the different types of immune activation that seem to be in play: "the good, the bad and the ugly." With the NIAID Immune Activation & HIV Pathogenesis Workshop now an annual event, progress on all three fronts will likely be reported in 2007.