Written by David Margolis, MD, University of Texas, Southwestern Medical Center, and ACTG researcher
1. Dendritic cells and other players in the antiviral immune response
2. Viral reservoirs and viral latency
3. T Cell homeostasis and HIV infection: agreeing to disagree?
- What causes immune destruction? No firm answers.
4. Interleukin 2: the CD4 Fountain of Youth?
Dendritic cells and other players in the antiviral immune response:
During the first plenary lecture on Monday (L2), Melissa Pope discussed the role of dendritic cells (DCs) in HIV infection and pathogenesis. These cells play an important role in directing the immune response when foreign antigens are encountered. However, HIV to has adapted to subvert the function of these cells to aid in viral replication and dissemination. Dendritic cells are equipped with receptors for antigen (HIV), with cell adhesion molecules, and receptors for immunoregulatory and costimulatory signals. They can process foreign antigens encountered as they traffic throughout the body, or while waiting in mucosal or submucosal tissues. Processed antigens are then presented along with costimulatory signals to both T and B-lymphocytes, providing an optimal stimulus for an immune response.
Two types of DCs are known: myeloid DCs (MDCs) are found in epithelial, submucosal, and lymphoid tissues, as well as the blood and lymph, while plasmacytoid DCs (PDCs) traffic only in blood and lymph. DCs can be infected by HIV as they possess CD4 receptors, and can express the CCR5 receptor (if immature) or the CXCR4 receptor (when mature). HIV using the CCR5 coreceptor (ýR5 virusţ) efficiently and productively infects immature PDCs. As R5 virus usually mediates initial infection and early disease, infection of immature PDCs is likely important in establishing early infection. While MDCs are inefficiently infected, the presentation of HIV particles by MDCs to T cells provides an unfortunately optimal means of driving T cell infection and high-grade HIV production. MDCs have been said to be the ýTrojan horseÝ of HIV disease, delivering the virus to unsuspecting T cells. Understanding how to use DCs to prime an antiviral immune response, and how to prevent HIV transmission via infection of DCs are current areas of intensive investigation.
(editorial note: perhaps HIV treatment to prevent or alleviate HIV infection may develop from this research. In particular, microbicides could be developed to prevent sexual HIV transmission).
Brigitte Autran (L3) discussed immune reconstitution seen following effective HAART, the limitations of the immune response recovered, and current strategies to be tested to improve immune reconstitution. HIV antigens are presented by DCs to CD4 and CD8 cells, leading to the expansion of HIV-specific CD4 and CD8 clones. While HIV-specific CD4 clones are depleted directly and indirectly by HIV infection, sensitive assays can detect cells capable of producing interferon gamma in response to HIV antigens despite advancing immunodeficiency. The production of interferon gamma by T cells in response to a particular antigen is proof of a pre-existing mature and specific immune response. Therefore, even in advanced HIV disease, use of appropriately sensitive assays still detects a specific anti-HIV immune response. It is self-evident that this response is not always enough to prevent the development of AIDS.
Once HAART is initiated, CD4 cell counts can rise but the number of cells displaying an anti-HIV response typically wanes. This is due to the fact that the immune system sees less HIV antigen to elicit an immune response when viremia is suppressed by HAART, and also perhaps because of inappropriate antigen presentation by DCs. When treatment interruption was initially proposed, one thought was that an interruption would allow "auto-vaccination of one's immune response against the rebounding virus. Trials of interruption of therapy in most settings have found that viral RNA levels rebound to pre-therapy levels, suggesting that while the immune system can remember HIV (that is, immune responses return following a treatment interruption) this reawakened response is not more capable of controlling HIV replication.
This finding has led to a variety of current studies of therapeutic immunization during a period of HAART, followed by a treatment interruption. The goal is to teach the immune system new tricks while it is protected by HAART, in the hope of achieving better control of viral replication once therapy is halted.
In regard to such analytical treatment interruptions, Douek and colleagues from the NIH Vaccine Research Center added a cautionary note (abstr. LB7). They identified populations of T cells responding (i.e. producing interferon gamma) to either HIV antigens or (as a control) CMV antigens, and sorted these cells by flow cytometry. Then using a novel PCR technique, they compared the amount of proviral HIV DNA within the two cell populations. The proportion of either population that was infected, as evidenced by the presence of HIV DNA within them, was low. However, HIV-specific cells were significantly more likely to be infected. This could be looked at as bad news, or no news. In responding to sites of HIV replication, like a firefighter these cells place themselves at risk. The overall question that remains to be answered is whether HIV replication can be controlled by novel strategies that pairs cycles of antiviral therapy with cycles of immunotherapy, or whether slow but relentless infection of responding antiviral cells eventually erodes the immune response. Research so far has shown no evidence that therapy interruptions in persons chronically infected with HIV results in improved control of HIV when off HAART. Current studies continue to examine whether a drug holiday before initiation of salvage therapy improves response, whether early treatment within the first few months of therapy allows treatment of limited duration with better control of HIV replication, or whether immunotherapies such as HIV vaccines given during HAART allow HAART interruption with better control of viral replication. This study illustrates at the single-cell level, the cost of unrestrained viral replication.
Presentations during Monday's oral abstract session further explored the role of the macrophage in HIV infection. Brodie and colleagues (abstr. 27) found that all 24 men studied had detectable HIV DNA and RNA, primarily originating from rectal mucosal macrophages. Surprisingly, no correlation was found between plasma viral load and rectal HIV shedding. It was suggested that rectal shedding might be related to genital herpes status. This is supported by many previous studies showing that macrophages can produce prodigious amounts of virus for prolonged periods of time, and the herpes can powerfully upregulate HIV replication. Further, of significant implication for secondary HIV transmission, it appeared that suppression of plasma viremia did not guarantee lack of rectal HIV shedding. Bill Gates presciently foreshadowed this finding during his keynote address, illustrating yet again that he is not your average computer engineer, when response to a question from the audience he suggested: "Why don't you wear a condom..Ő"
In a descriptive but still impressive study, Jameson et al. (abstr. 28) illustrated the anatomic mucosal distribution of DC-SIGN and DC-SIGNR, two dendritic cell receptors capable of binding HIV and efficiently promoting HIV infection of susceptible cells. DC-SIGNR was found on 30% of the blood capillaries of cells of small intestinal mucosa, while DC-SIGN was found on the mucosal tissues of the rectum and vagina. HIV therefore has easy access to cells that can assist the virus in traversing these mucosal surfaces, and establishing infection.
In the meeting's final plenary (L9) Ashley Haase presented a visual display of initial SIV infection across the genital mucosal barrier. Using immunohistochemistry and in situ hybridization techniques to directly visualize activation markers at the cell surface, and SIV DNA within infected cells, Haase made the case that virus can immediately cross the vaginal mucosal barrier via tiny microfissures in the mucosal surface. SIV DNA within cells was seen within 2 hours of a traumatic vaginal inoculation. The majority of these infected cells over the first few days of infection are CD4+ T cells, with macrophages (ca. 10%) and dendritic cells (ca. 5%) accounting for the remainder of infected cells. These proportions appear to remain stable over time within both the blood and lymphoid tissue compartments. Paradoxically, during the first 4 weeks of infection, most of these infected cells do not display activation markers, although those cells that are activated appear to produce large amounts of virus. As infection continues, activated cells become the predominant ones that are infected, in concert with systemic immune system activation in response to viral infection.
Lessons from the monkey model
The Feinberg group at the Emory Vaccine Research Center presented the results of a unique experiment in the SIV infection model of rhesus macaques (abstr. 24). SIV in macaques results in immune activation and simian AIDS. However, the natural reservoir of SIV, sooty mangabey monkeys do not respond to SIV infection with immune activation, and do not develop simian AIDS despite high levels of viremia. It should be noted, however, that these mangabeys do not live completely at peace with SIV, as some T cell depletion may be observed.
By giving anti-CTLA4 and anti-CD40L immunoglobulin by IV infusion during primary infection with SIVmac239, the Feinberg group attempted to make the macaques tolerant of SIV, in the hope of mimicking sooty mangabey infection and blocking the development of AIDS. While this artificial costimulatory blockade blocked proliferation of CD4+ T cells and formation of lymph node germinal centers, proving that the administered immunoglobulin had done its job, treated macaques still developed simian AIDS. While this could not have been a definitive experiment, due to the many effects on the immune response induced by non-specific blockade of costimulation, it argues strongly against a simple model of AIDS in which inappropriate CD8+ response drives immunopathology
How is it that you do what you do to me?
Several groups presented findings describing if not addressing, the enduring mysteries of qualitative immunodeficiency induced by HIV replication. It seems likely that, in addition to virus-mediated CD4 destruction, other immune defects induced by HIV are critical in the emergence of AIDS.
Connors and colleagues at the NIAID (abstr. 216), measured responses to HIV peptides (gag, pol, nef, and p24) and control antigens (CMV, tetanus) in 13 subjects prior to, during, and after therapy interruption. Higher frequencies of HIV-specific CD4+ T cells were measured than expected, and HIV-specific CD4+ T cells comprised up to 1.5% of CD4+ T cells off therapy (a vigorous response). However, all HIV-specific proliferative responses were profoundly depressed during viremia, while proliferation to tetanus and CMV antigens were only somewhat diminished. PBMC collected while the subjects were aviremic time retained proliferative capacity. Connors et al. concluded that loss of proliferation is not the cause but the result of unrestricted viral replication. These findings suggest that the mechanism by which viral replication induces immune defects must be uncovered and treated, or that a durable and effective immune response (and perhaps the response to a therapeutic vaccine) must be supported or assisted by ongoing or intermittent antiviral therapy.
Studies by Lieberman and others had previously suggested that anti-HIV CD8 cells had a specific defect in the ability to produce the target-killing molecule perforin. Clerici and colleagues (abstr. 221) reported that granule-dependent (perforin and granzymes) but not granule-independent (TNF-alpha) mechanisms of target lysis were defective in CD8+ T cells of patients on HAART when compared to those not on HAART. The authors concluded that immunomodulatory therapies are needed for successful therapy of HIV infection.
However, a collaboration between the Lederman group in Cleveland and the Lieberman group in Boston suggested that loss of perforin expression in HIV-specific CD8+ cells may be a regulatory mechanism typical of many chronic viral infections. The investigators examined HAART-suppressed patients and a group of long-term non-progressors. Surface markers and cytolytic (e.g. perforin) molecules were expressed comparably on all CD8+ cells and on anti-HIV specific CD8+ cells in all patients. Further, perforin expression was not significantly different between HIV-specific, EBV-specific, and CMV-specific antiviral CD8+ cells. A defect in the ability of anti-HIV specific CD8+ cells to kill HIV-infected target cells may be a specific defect induced by HIV, and may be seen with other chronic viral infections. This may be a general mechanism by which such infections evade the host response, however overcoming this defect may still play an important role in future immunotherapies for AIDS.
T Cell homeostasis and HIV infection: agreeing to disagree?
What causes immune destruction? No firm answers.
The impact of HIV infection on turnover of T cells and their production by the thymus was the focus of a symposium and several other presentations. In comparison to the lively and contentious session covering this topic at last year's meeting, most participants in the debate (save one) seems to be nearing a consensus. The issue at hand is worthy of debate as it is central to HIV pathogenesis and therapeutics: how is the immune system damaged by HIV infection, and can the lesion be repaired. In simpler terms, it is a debate over whether AIDS develops because of the tap of CD4 cell production runs dry, the drain of T cell destruction is opened, or a little of both.
Perelson discussed results of studies by his group at Los Alamos and in collaboration with Ho, Mohri and others at the Aaron Diamond Inst. (S9 and abstr. 507). DNA in proliferating T cells was marked by 7-day infusions of labeled glucose. In HIV+ compared to uninfected subjects, increased proliferation rates for both CD4 and CD8 cells was measured, but increased death rate only for CD4 cells. This was said to support the view that CD4 depletion is primarily driven by increased cell destruction rather than decreased CD4 production. Zhang (abstr. 101) of the Aaron Diamond provided corroborating data from monkey studies, in which TRECs, DNA markers of cells recently emerged from the thymus, decreased gradually following thymectomy. An increase in the rate of TREC loss was not seen when the animals were subsequently infected with SIV. However, these experiments did not seem definitive as it was not clear that a subtle change could have been detected.
Hellerstein (abstr. 102) and McCune (S12) discussed human studies of short-term (pulse) labeling with deuterated glucose and long-term (chronic) labeling with deuterated water. By measuring labeling in cell populations sorted for memory/effector markers vs. naive markers, these investigators also found higher rates of proliferation and death in the memory/effector cell populations. They described the primary abnormality associated with HIV infection, and ameliorated by HAART to be in inability to maintain a pool of long-lived, quiescent memory T cells.
Kovacs (S10) summarized the findings several studies by the NIAID group. Using an alternate DNA label, BrdU, these investigators also detected rapidly and slowly proliferating populations of CD4 and CD8 cells, and similarly found that HIV infection served to drive more cells into the rapidly proliferating pool of cells.
While appeared to be consensus as to the role of increased proliferation in HIV pathogenesis, how this results in gradual immunodeficiency was not directly addressed. Although these groups did not discuss direct evidence of HIV-induced defects in T cell production, as they have in the past, infection of thymic T cell progenitors may contribute to immune depletion as well. In a poster presentation, the Camerini lab found that HIV-1 patient isolates that utilize the CCR5 coreceptor (formerly known as "non-syncytial forming" in cell culture) kill cells in fetal thymic organ culture (FTOC). Several R5 HIV-1 clones from rapid progressors (RP) or long-term non-progressors (LTNP) were tested. While R5 HIV-1 clones from either RP or LTNP patients did not kill cells in lymph node or tonsil organ culture, both RP and LNTP clones caused depletion of CD4+ thymocytes in FTOC. The authors concluded that LTNP status was linked to other factors, and that LTNP virus was capable of depleting new T cells in the thymus
Stoddart and coworkers at UCSF used a mouse model of HIV infection to study the pathogenic potential of R5 virus that is most often isolated in early HIV infection. Severe combined immunodeficient (SCID) mice are used, as they tolerate the implantation of human fetal thymus and liver under the kidney capsule. This organoid develops and allows the development of several lines of human blood cells within the mouse, including T cells, for periods up to 1 year.
In this model system, rapid destruction of the thymus was not observed after infection with an R5 isolate of HIV-1. However infection caused delayed destruction of the thymus. This occurred in concert with increased expression of CCR5 HIV co-receptor on T cells. This induction of CCR5 was mediated by IFN-alpha secreted by intrathymic predendritic cells in response to HIV-1 infection. These findings were yet another demonstration of a mechanism by which an HIV adaptation to a host immune response may lead to HIV pathogenesis.
Interleukin 2: the CD4 Fountain of Youth?
Kovacs (abstr. 103) and Sereti (abstr. 104) studied the turnover of T cells in subjects receiving IL-2. These findings were initiated after it was observed that some subjects, IL-2 therapy was required infrequently (i.e. once a year) to maintain CD4 counts well above 500/Ál. The turnover of T cells in HIV-infected subjects receiving suppressive HAART was measured by administration of labeled glucose or water. The labels are incorporated into newly synthesized DNA of dividing cells, allowing the measurement of new DNA in T cells and the calculation of the average life span of new T cells. The turnover of T cells from subjects receiving only HAART or HAART and intermittent IL-2 was compared. The half-life of T cells in subjects not receiving IL2 was 0.7 years, but the half -life in subjects receiving IL2 was increased to from 0.8 to 5 years. As subjects received IL2 for longer periods of time, the half-life appeared increase further.
Levy and the ANRS 079 Study Group (abstr. 514) reported encouraging clinical experience with long-term subcutaneous IL-2 therapy. Similar to the findings of the ACTG 328 study reported at ICAAC last fall, they found that IL-2 recipients experienced a faster and more substantial increase of CD4 cells, maintained over 35 months. As reported in prior studies, and consistent with the findings above, infrequent IL-2 cycles were needed to maintain response. Viral suppression was equal in the HAART alone and HAART/IL2 groups. The median CD4 increase from a baseline of ca. 350 cells/Ál was +262 (HAART) and +835 (IL-2 arm) at week 74 (p<0.0001). Again, similar to other studies, the ARNS Group did not find that the size of the latently infected CD4 cells pool was decreased by IL2 (abstr. 515).
Overall, evidence continues to build that intermittent IL2 can be tolerated and that virologic suppression can be maintained on IL2 (if not aided by it). An overarching question is whether or not one is better off living with a T cell count of 1000 or 500, and if IL2 is worth the extra 500 cells. Ongoing studies such as ESPRIT and SILCAAT seek to discover if increased CD4 counts achieved on IL2 convey clinical benefit. The findings of Kovacs and the NIAID group suggest that if clinical benefit is discovered in these large clinical endpoint trials, that the extended half-life of cells born on IL2 would make frequent administration of the cytokine unnecessary in most patients, after an initial response had been achieved.
Viral reservoirs and viral latency
Although Doug Richman's symposia discussion of viral reservoirs and dynamics was not given until late in the meeting (S22), it serves as an excellent framework for review of the new insights on the topic of latency presented at the 9th CROI. Dr. Richman reviewed the 3 phases of plasma virus decay: a) the first phase (half-life ca. 1 day) representative of infection in activated cells, b) the second phase (half-life ca. 14 days) representative of infection in cells that are chronically productive of HIV, and c) the third phase (half-life ca. 1.5 years by some estimates).
Reservoirs, be they in specific cell types or particular anatomic compartments, can contribute to viral production in the 2nd or 3rd phases of decay. Other than the well-described non-productive resting CD4 T cell, several cell types are proven or suspected reservoirs.
Indolent or low-level replication in CD4 cells lacking abundant cell surface markers of activation. Haase (L9) reported that the majority of infected cells during the first 4 weeks of infection do not display activation markers, and that over these cells persist but become a minority population among virus-producing cells.
These findings were echoed by Kinter (abstr. 269) who reported productive HIV infection of CD4+ T cells lacking markers of classic T cell activation. 3 to 19% of T cells in tonsillar histocultures lacking the expression of surface (CD25, CD69, or HLA-DR) or nuclear (Ki67, PCNA, or BrdU) activation markers associated with classic T cell activation were found to produce HIV p24. X4 HIV strains were more efficient in infecting this "resting" subpopulation of T cells than were R5 HIV strains.
Macrophages and dendritic cells may produce HIV over long periods of time. Presentations during Monday's oral abstract session further explored the role of the macrophage in HIV infection. Brodie and colleagues (27) found that all 24 men studied had detectable HIV DNA and RNA, primarily originating from rectal mucosal macrophages. Surprisingly, no correlation was found between plasma viral load and rectal HIV shedding. It was suggested that rectal shedding might be related to genital herpes status. However, of significant implication for secondary HIV transmission, it appeared that suppression of plasma viremia did not guarantee lack of rectal HIV shedding.
Other non-CD4 cell types may be productively infected with HIV. Several groups, including this author, have found that CD8 cells could be induced to express CD4 and become infected by HIV. Sullivan et al. from Rush (abstr. 407) reported the detection of activated CD4dim/CD8bright T Cells in HIV+ patients. CD4dim/CD8bright T cells were generated by exposure to SEB or anti-CD3/CD28. In contrast to prior reports, CD8+ T cells from HIV+ patients at different stages of HIV disease (early, moderate, and advanced) were still capable of up-regulating CD4 on CD8+ T cells, suggesting that the ability to generate the CD4dimCD8bright phenotype is not impacted by the stage of HIV disease.
Monocytes may harbor HIV. Fulcher et al. (abstr. 355) reported distinct patterns of HIV-1 evolution in CD14+ monocytes when compared with CD4+ T lymphocytes. CD14+ monocytes and CD4+ T lymphocytes were purified from PBMC by using both positive and negative selection from patients with and without HAART. The C2-V5 region of HIV-1 env sequences were amplified from DNA of purified cells using nested PCR, and from viral RNA of blood plasma using RT-PCR. At the first positive time point around serconversion, 8/12 (66.7%) patients studied had a homogenous virus. Of the 4 patients with heterogeneous virus around seroconversion, 3 had HIV-1 env sequences in CD14+ monocytes that were distinct from those in CD4+ T lymphocytes. Between 2 and 8 years following HIV-1 infection, 6 of 8 treated patients had distinct populations of heterogeneous virus between CD14+ monocytes and CD4+ T lymphocytes.
Finally, dendritic cells found in the germinal centers of lymphoid tissue may be a source of persistent HIV infection. This is not because they are chronically infected, but rather because extracellular virions may be trapped and preserved, as is foreign antigen, by these cells. Richman estimated that up to 1011 virions may be encased in follicular dendritic cell compartments.
However, the Pomerantz group (abstr. 494) found that peripheral blood dendritic cells were not a major reservoir in patients on HAART. Peripheral blood DCs were isolated from a cohort of HIV-1-seropositive men taking suppressive HAART. Cell sorting showed yields from 85.90 - 92.18% of relatively pure DCs isolated from patients' PBMCs. Although RNA gag and DNA RU/5 were detected in all PBMC samples isolated from these patients, these proviral and virion or viral unspliced RNA forms were not detected in the DC isolates. In addition, no replication-competent virus was demonstrated in the DC fraction, while virus was isolated from each patient's CD8+ T-lymphocyte-depleted PBMCs. Therefore, DCs may not be a major reservoir site for HIV-1 in patients on HAART with undetectable (< 50 copies/mL) plasma viral RNA.
Viral reservoirs may also be anatomical. Lymphoid tissue (as above) comprises one reservoir of persistent infection. The CNS and the mucosal tissue of the genital and gastrointestinal tract are two other important reservoirs. Discordance of drug resistance patterns and variable penetration of antiretrovirals have been reported in these anatomic compartments when compared with the blood.
Ellis et al (abstr. 49) intensively sampled blood and CSF in 12 subjects undergoing a treatment interruption. They found that CSF virus rebounded in most subjects, although viremia recurred first in plasma. Viral rebound in the CSF was often accompanied by CSF pleocytosis. The kinetics of viral rebound suggested that virus was replicating in the same cell compartment as in the blood (ie. CD4 cells). The delay in CSF rebound was thought to be due to a smaller viral population within the CNS compartment, or delayed clearance of antiretrovirals from that compartment.
(editorial note: this finding creates a concern about taking a therapy interruption that appears to result in repopulating cells and tissues that were previously drained of HIV or had major declines in the amount of HIV).
As with some other drugs, Sankatsing and colleagues reported poor penetration of lopinavir into the genital tract of HIV-1-infected men (abstr. 439). 14 HIV-1-infected patients treated with a LPV-containing regimen for a minimum of 4 weeks were studied. 4 patients had a plasma LPV concentration below 4 mg/L, the other 10 patients had a plasma concentration above 4 mg/L. The desired plasma concentration is > 4 mg/L. The LPV concentration in seminal plasma (SP) ranged between 0.046 and 3.9 mg/L (median 0.23 mg/L, IQR 0.15-0.33). There was a weak relation between the plasma concentration and the SP concentration (r = 0.51, p=0.07). The median ratio of the concentrations of LPV in SP and in plasma was only 0.034 (IQR 0.021-0.070).
Many reports dealt with the most well known reservoir of persistent infection, the resting CD4 T cell. As mentioned Dr. Richman's talk, Joe Wong (abstr. 492) found that proviral species in the lymphocyte reservoir have different rates of decay. 9 chronically HIV patients who had received 0.5-1.5 years of non-suppressive antiviral therapy including 3TC followed by suppressive therapy (plasma HIV RNA <50c/ml) were studied over a 2- to 4-year period. All subjects had developed the M184V mutation in both plasma and PBMC with non-suppressive therapy. Quantitative PBMC HIV DNA sequencing was performed to determine the proportion of wild type and 184V mutant (3TC-resistant) virus. This was used to estimate the individual decay rates of HIV DNA with the wild type and mutant genotypes. For patients without intermittent viremia clearance of HIV DNA with 184V was more rapid (T1/2:43-138 weeks, median 61 weeks) than for the 184WT (T1/2:82-infinite, median 1540 weeks). One patient with intermittent viremia and continued 3TC had a 184V half-life of 417 weeks, consistent with the effects of replenishment of the 184V population by residual replication. This may be understood to represent the fact that the drug-sensitive population in the latent reservoir was larger than the M184V population. They concluded that subsets of cells with replication competent virus persist during HAART which have intrinsically slow turnover rates that are unlikely to be affected by more potent antiviral suppression.
Fitting the same model, a phylogenetic analysis of envelope sequences re-emerging after sequential therapy interruption (abstr. 50) was reported by Martinez-Picardo and colleagues. Populations of rebounding virus changed unpredictably. This might have been due to the "firstest with the mostest," or what is known in population biology as the founder effect. If there exist several distinct viral species capable of replication, the first viral species to begin replication after each therapy interruption may become the most numerous by being the first to infect the current pool of receptive, uninfected T cells. This finding was reminiscent of that reported by the Fauci group several years ago, when subjects on prolonged and successful HAART underwent a single interruption. In some cases a novel viral species predominated upon rebound. Further, the evolution of an immune response to a rebounding population seen in a first treatment interruption might make it less likely that the population would be allowed to outgrow in a second interruption. These findings again emphasize the need for a broad antiviral immune response to contain viral replication and reseeding.
Chun and Fauci (abstr. 493) also cautioned that the detection of HIV RNA does not always mean that virus is being produced. They investigated whether the latent viral reservoir is capable of producing HIV virions in both viremic and aviremic (HAART) patients ex vivo. Resting CD4+ T cells (CD4+/CD25-/CD69-/HLA-DR-) were isolated with extreme purity (>99.9%) from 6 viremic and 7 aviremic HIV-infected patients. Cells were incubated with media +/- cyclosporin A, HAART, actinomycin D, or anti-CD3 antibody for 1-3 days. Resting CD4+ T cells from all viremic patients continued to produce readily detectable amounts of virions, as measured by Amplicor assay, in the absence of activating stimuli. The degree of viron production, while not significantly suppressed by inhibitors of cellular proliferation or by HIV-specific drugs, was abrogated by actinomycin D, a blocking agent of RNA synthesis. Low levels of virion production were also detected in the resting CD4+ T-cell compartment from a small number of aviremic patients receiving HAART. However, no virions were detected in the culture supernatants of resting CD4+ T cells from the majority of patients receiving HAART, despite the fact that cell-associated HIV RNA was readily detected. No quantifiable virions were produced by the latent viral reservoir in the majority of patients receiving HAART, despite the presence of cell-associated HIV RNA. Transcription of HIV RNA in resting CD4+ T cells may not be direct evidence for ongoing viral replication during effective therapy.
Some basic information about the fate of virus in resting cells was presented. Michael Malim's lab (abstr. 153) found that late (ie. nearly complete) HIV-1 reverse transcripts accumulate stably within resting CD4 T cells. Although it is clear that reverse transcription is very inefficient in resting cells, the Malim lab found that late reverse transcripts accumulate with time in resting T cells, despite the lack of cell activation. In contrast, within activated T cells, late reverse transcripts are degraded with time. Whether these DNA species pose a threat remains to be seen.
Two groups suggested that circular HIV DNA molecules formed by ligation of completed reverse transcripts might be stable in some cell types. The Meusing lab (abstr. 157) found that 2-LTR circles are stable for at least 1 month following infection of macrophages (not T cells) with wild type and integration-defective HIV and they appear to persist throughout the life of the cell. While the ability of 2-LTR circles to support gene expression appears to vary depending on the viral gene product, no evidence of virion production was found. However the Klotman lab (abstr, 158) found that infection of primary macrophages isolated from whole blood with either VSV-pseudotyped, integrase competent, or integrase defective and replication defective recombinant HIV-1 resulted in stable 2-LTR circles. Circles were stable in primary macrophages to 56 days post infection but in dividing cells were lost over time. The pattern of expression of multiply spliced RNA from extrachromosomal circular DNA was identical to that from integrated HIV-1 proviral DNA. Viral Tat expressed from extrachromosomal circles activated transcription from viral LTRs within the same cell but not viral LTRs in neighboring cells. The detection of circles therefore does not necessarily indicate the ongoing production of newly infected cells. No production of new infectious viral particles was reported.
Finally, in a study that has been long needed, Blankson from the Siliciano lab (abstr. 160) reported the characterization of the pre-integration state of HIV-1 latency. HIV-1 genomes in infected resting CD4+ T cells were assayed using a novel linker ligation mediated PCR (LM-PCR) assay. This method was developed to specifically detect and characterize linear forms of the HIV-1 genome. To explore the relationship between the formation of linear viral DNA in infected resting CD4+ T cells and the stability of the pre-integration state, we employed a recombinant virus expressing the enhanced green fluorescent protein (EGFP) to measure the rate with which HIV-1 decays in the pre-integration state. Resting CD4+ T cells were obtained and infected with the HIV-1 NL4-3 EGFP virus. At various times following infection, a fraction of the infected resting cells were removed and activated by mitogen. Using flow cytometry, the decay of the virus in the pre-integration state was measured over time as reduction in the fraction of infected resting T cells from which virus could be rescued (ie. express EGFP signal) following cellular activation. Analysis of the stability of the pre-integration state during this period reveals that the ability to rescue virus decays with a half-life of one day.
Therefore in summary, pre-integrated but completely reverse transcribed viral DNA appears relatively short-lived in T cells, potentially long-lived in other cells, but in general unlikely to result in production of new infectious viral particles.
Finally, several lines of investigation towards therapies to clear the resting T cell reservoir of infectious HIV were reported. Our laboratory (abstr. 161) reported further studies of the molecular mechanism of down-regulation of HIV gene expression. Our studies show that a host cellular enzyme, histone deacetylase type 1, mediates repression of HIV gene expression and inhibition of HIV Tat activation. These studies may lead to drugs that could specifically reactivate quiescent HIV.
The Pomerantz group presented a Herculean attempt to eradicate HIV (abstr. 405). 3 patients with HIV-1 infection on HAART, with < 50 copies/mL of plasma viral RNA for > 1 year, were enrolled in this trial. They continued their baseline HAART, and added DDI plus hydroxyurea (HU). After at least a month of therapy, the patients were treated in a bone marrow transplant unit with low dose OKT3 antibody, followed by a 2-week course of subcutaneous IL-2 to stimulate latent provirus. The dose of OKT3 used (400 g) stimulates T-cells rather than depleting these cells in vivo. All treated patients had only modest side effects. The HU and DDI were continued with HAART throughout this protocol, and for 5-6 months after OKT3 and IL-2 stimulation.
Replication-competent virus was undetectable after treatment and plasma viral RNA also became undetectable in each of these patients, using an ultrasensitive RT-PCR assay which quantitates to 5 copies/mL. Tonsillar biopsies were performed and no in situ hybridization for HIV-1-specific RNA was detected in lymphocytes or on follicular dendritic cells. 2 subjects stopped all antiretroviral therapy, but plasma viremia recurred after several weeks. Although this trial was unsuccessful, it should be noted that no serious toxicities were incurred.