Finding a Cure for HIV: Much Work to Do Editorial
Download the PDF here
Download the PDF here
"amount of residual infectious virus left after ART and an effective immune response are both likely to be key in achieving long-term HIV remission"
See Also (below following Editorial: [Antiretroviral-Free HIV-1 Remission and Viral Rebound After Allogeneic Stem Cell Transplantation: Report of 2 Cases]
Sharon R. Lewin, FRACP, PhD
Ann Intern Med. Published online 22 July 2014
Antiretroviral therapy (ART) for HIV infection was one of the most spectacular scientific advances in recent decades. Patients who initiate ART at the right time can now have a normal life expectancy, and ART also dramatically reduces the risk for HIV transmission. However, treatment is not perfect. Lifelong ART is required, there are associated short- and long-term toxicities, and the drugs and health systems required to keep persons in long-term care come at a considerable cost (1). Therefore, there is a significantly renewed effort to find a cure for HIV (or a way to put HIV into remission) so patients can safely stop ART but keep their virus under control.
The reasons why ART cannot currently cure HIV are complex. The major barrier is the long-term persistence of latently infected resting T-cells, with virus integrated into the host genome but not actively replicating. Latently infected cells are established early after infection; are enriched in tissue sites, such as the gastrointestinal tract; and likely have preferential expansion or homeostatic proliferation leading to an extremely long half-life (2). Other long-lived infected cells, such as naive T-cells, astrocytes, or microglia, may also play a role.
In this issue, Henrich and colleagues (3) report an in-depth study of 2 patients with HIV who had "transient" HIV remission after stopping ART after hematopoietic stem cell transplantation (HSCT) for hematologic tumors. The patients were receiving suppressive ART and had HSCT with reduced-intensity irradiation and immunosuppression with tacrolimus, sirolimus, and methotrexate.
Graft-versus-host disease developed in both patients, requiring additional treatment with prednisone. Nine months after HSCT, neither patient had detectable HIV DNA in his blood, a fairly crude measure of long-lived, latently infected resting T-cells that, if anything, overestimates the amount of residual infectious virus. HIV RNA and DNA were not detected in their blood in several subsequent tests, including using leukapharesis to collect many resting T-cells. In one patient, HIV DNA was not detected in rectal tissue. However, once ART was stopped 2 and 4 years after HSCT, the virus returned, albeit later than usual. Instead of the usual 1 to 4 weeks, it returned at 12 and 32 weeks. Viral rebound in blood was extremely rapid and was associated with classical symptoms of HIV seroconversion, and ART was reinstated, eventually resulting in effective virus control. Given that the patients had acquired a new "naive" immune system from a donor without HIV, there was no detectable HIV-specific T-cell response and waning HIV antibody levels before ART cessation.
There are some similarities but also many differences among these cases and the highly publicized "Berlin patient," Timothy Brown. The Berlin patient received an HSCT from a donor who did not express the key coreceptor for HIV (the chemokine receptor gene, CCR5) and was therefore naturally resistant to HIV, whereas the patients in Henrich and colleagues' study received stem cells from donors that expressed CCR5. The Berlin patient stopped ART soon after transplantation, whereas the patients in Henrich and colleagues' study received ART for a prolonged period to protect their new hematopoietic system from infection. There were differences in the conditioning regimens used, including the intensity of radiation and degree of immunosuppression, but importantly, the Berlin patient also had significant graft-versus-host disease. Finally and most significantly, the Berlin patient has not received ART for 6 years, has virtually no virus detected in blood or tissue, and remains the only patient truly cured of HIV (4).
There are also some similarities and differences among these cases and the highly publicized Mississippi baby (5). The Mississippi baby was infected with HIV at birth and received ART within 30 hours of delivery. Antiretroviral therapy was continued for 18 months and then stopped. Very low traces of the virus were detected in some cells, but the virus stayed at undetectable levels in plasma (for a remarkable 27 months) and then rebounded (6). As with the patients in Henrich and colleagues' study, viral rebound was significantly delayed compared with any previous reports, measures of residual virus on ART were largely negative, and there was no detectable HIV-specific immune response, at least before rebound (neither antibodies nor HIV-specific T-cells).
However, despite these differences and the fact that HSCT will never be used as a strategy to cure HIV given toxicity and cost and very early ART after infection is not a feasible intervention for persons already infected with HIV, all 4 cases have much to teach us. These lessons are likely to significantly influence future avenues of research to find an HIV cure.
What did we learn from the patients in Henrich and colleagues' study? The good news is that it is possible to significantly reduce the number of long-lived, latently infected T-cells that persist in patients receiving ART and that this reduction was associated with a significant delay in viral rebound once ART was stopped. This was also clearly demonstrated in the Mississippi baby.
The rest was not good news. We learned that the current assays we have to measure virus while patients are receiving ART are not sensitive enough to detect residual infectious virus and therefore currently cannot predict the critical clinical end point of when and whether virus rebounds after stopping ART. Treatment interruption, in the right context and with appropriate monitoring, is currently the only test of cure. A robust biomarker that can predict viral rebound off ART remains one of the most important currently unanswered questions in the field.
We also learned that long-term remission of HIV will likely need more than a very significant decrease in the number of latently infected T-cells. Other sources of virus and a boost or at least some level of effective immune control may also be important. The most sobering lesson was that these cases, along with the Mississippi baby, have raised the possibility that total elimination of every last virus or infected cell to achieve lifelong remission may not be possible.
Which component of the regimen used in Henrich and colleagues' study so dramatically reduced HIV DNA? Bone marrow irradiation and chemotherapy alone have not previously been found to reduce long-lived, latently infected cells (7). Immunomodulatory drugs, such as sirolimus, may have played a role, as recently reported in a study of patients with HIV after kidney transplantation (8). However, the process of graft-versus-host disease or "graft-versus-leukemia" (or in this case, "graft-versus-latent reservoir") may have been critical. The expanding understanding of the immunology of the graft-versus-host response after transplantation (9), including the role of the innate immune response (specifically natural killer cells) and T-cells may be key in finding new ways to eliminate latency.
More tractable and scalable approaches than HSCT and very early ART are clearly needed for the 35 million persons already infected with HIV who will all eventually require lifelong treatment. One approach is to "purge" the virus from long-lived, latently infected T-cells. Multiple latency-reversing agents (including several histone deacetylase inhibitors and the antialcoholism drug, disulfiram) have recently been examined in clinical trials of patients with HIV receiving ART. Although each of these trials have shown some modest evidence of activation of virus from latency, no study has yet shown any change in the amount of residual infectious virus or HIV DNA (1). It is likely that purging alone will not be enough and a boost to the immune system may still be needed through either vaccination or immune modulation.
The modification of CD4+ T-cells to make them resistant to HIV using gene therapy may be another promising yet invasive and expensive strategy. The elimination of CCR5 expression, a strategy inspired by the Berlin patient's success, using zinc-finger nuclease treatment of circulating T-cells has recently been shown to be feasible and safe in patients with HIV receiving ART (10). The challenge now will be to find ways to boost the numbers of modified cells. The even greater long-term challenge will be to ensure that any high-tech cure, such as gene therapy, is ultimately accessible, particularly in low-income countries.
In the early 1990s, Time Man of the Year and prominent HIV researcher David Ho paraphrased former U.S. president Bill Clinton when he said, "It's the virus, stupid," to emphasize the importance of understanding the virus's primary role in the pathogenesis of AIDS. However, in the case of finding a cure for HIV, it may not all be just about the virus. The amount of residual infectious virus left after ART and an effective immune response are both likely to be key in achieving long-term HIV remission. The patients in Henrich and colleagues' study have demonstrated that we still have much work to do.
Ann Intern Med. Published online 22 July 2014
"In summary, our results suggest that allogeneic HSCT with CCR5 wild-type donor cells may lead to loss of detectable HIV-1 from blood and rectal mucosa, but viral rebound may nevertheless occur after ART interruption despite a significant reduction in reservoir size. The definition of the nature and half-life of residual viral reservoirs is essential to achieve durable, ART-free HIV-1 remission......Allogeneic HSCT may lead to loss of detectable HIV-1 from blood and gut tissue and variable periods of antiretroviral-free HIV-1 remission, but viral rebound can occur despite a minimum 3-log10 reduction in reservoir size. Long-lived tissue reservoirs may have contributed to viral persistence. The definition of the nature and half-life of such reservoirs is essential to achieve durable antiretroviral-free HIV-1 remission......One patient developed new efavirenz resistance after reinitiation of antiretroviral therapy.....Substantial reductions in HIV-1 reservoirs were seen in 2 patients who had allogeneic HSCT with HIV-1-susceptible donor cells while receiving continuous ART. Despite readily detectable proviral DNA before HSCT (15), HIV-1 RNA, proviral HIV-1 DNA, and viral outgrowth were undetectable years after HSCT using highly sensitive assays applied to plasma, PBMCs, and gut-associated lymphoid tissue. Nevertheless, HIV-1 rebound was first detected 12 and 32 weeks after ART interruption. Rebound of viremia occurred within 2 weeks after the last negative plasma HIV-1 RNA result, and both patients developed symptoms consistent with the acute retroviral syndrome. Based on previously reported proviral DNA levels before or immediately after HSCT in both patients (96 to 144 copies/106 PBMCs) and the detection limits of sensitive quantitative assays used in this study, our findings suggest a minimum 3-log10 reduction in the number of circulating cells harboring proviral HIV-1 DNA after HSCT. Although allogeneic HSCT may lead to significant, sustained reductions in the HIV-1 reservoir, infected tissue or cell-bound virus persists. Persistence of these small numbers of residual infected cells seems to be sufficient to rekindle HIV-1 replication.
Antiretroviral-Free HIV-1 Remission and Viral Rebound After Allogeneic Stem Cell Transplantation: Report of 2 Cases
Timothy J. Henrich, MD; Emily Hanhauser, BS; Francisco M. Marty, MD; Michael N. Sirignano, BS; Sheila Keating, PhD; Tzong-Hae Lee, MD, PhD; Yvonne P. Robles, BA; Benjamin T. Davis, MD; Jonathan Z. Li, MD; Andrea Heisey, BS; Alison L. Hill, PhD; Michael P. Busch, MD, PhD; Philippe Armand, MD, PhD; Robert J. Soiffer, MD; Marcus Altfeld, MD, PhD; and Daniel R. Kuritzkes, MD
Background: It is unknown whether the reduction in HIV-1 reservoirs seen after allogeneic hematopoietic stem cell transplantation (HSCT) with susceptible donor cells is sufficient to achieve sustained HIV-1 remission.
Objective: To characterize HIV-1 reservoirs in blood and tissues and perform analytic antiretroviral treatment interruptions to determine the potential for allogeneic HSCT to lead to sustained, antiretroviral-free HIV-1 remission.
Design: Case report with characterization of HIV-1 reservoirs and immunity before and after antiretroviral interruption.
Setting: Tertiary care center.
Patients: Two men with HIV with undetectable HIV-1 after allogeneic HSCT for hematologic tumors.
Measurements: Quantification of HIV-1 in various tissues after HSCT and the duration of antiretroviral-free HIV-1 remission after treatment interruption.
Results: No HIV-1 was detected from peripheral blood or rectal mucosa before analytic treatment interruption. Plasma HIV-1 RNA and cell-associated HIV-1 DNA remained undetectable until 12 and 32 weeks after antiretroviral cessation. Both patients experienced rebound viremia within 2 weeks of the most recent negative viral load measurement and developed symptoms consistent with the acute retroviral syndrome. One patient developed new efavirenz resistance after reinitiation of antiretroviral therapy. Reinitiation of active therapy led to viral decay and resolution of symptoms in both patients.
Limitation: The study only involved 2 patients.
Conclusion: Allogeneic HSCT may lead to loss of detectable HIV-1 from blood and gut tissue and variable periods of antiretroviral-free HIV-1 remission, but viral rebound can occur despite a minimum 3-log10 reduction in reservoir size. Long-lived tissue reservoirs may have contributed to viral persistence. The definition of the nature and half-life of such reservoirs is essential to achieve durable antiretroviral-free HIV-1 remission.
Primary Funding Source: Foundation for AIDS Research and National Institute of Allergy and Infectious Diseases.
· Because HIV can be detected in patients even when they are receiving long-term suppressive antiretroviral therapy (ART) and can rebound within weeks of stopping ART, HIV therapy is currently recommended as a life-long treatment.
· Last year, 2 patients with HIV who had an allogeneic hematopoietic stem cell transplant with HIV-1-susceptible wild-type donor cells achieved sustained HIV remission many weeks after stopping ART. However, both patients have now developed detectable virus in blood, as well as the acute retroviral syndrome, and ART has been reinitiated.
· HIV can rebound even after a significant reduction in HIV reservoir size. An HIV cure remains elusive.
A major challenge in eradicating HIV-1 is the persistence of latently infected cells, which are established by integration of the viral genome into host cell chromosomes (1-2). Combination antiretroviral therapy (ART) reduces plasma HIV-1 RNA levels to below the limit of detection of clinical assays. However, low-level plasma viremia or cell-associated HIV-1 DNA are detected in most patients receiving ART, even after intensification of the antiretroviral regimen (3-5). Furthermore, the virus typically rebounds within 1 to 8 weeks after treatment interruption in patients receiving long-term suppressive ART (6-11). As a result, ART-free HIV-1 remission (that is, "functional" cure) remains elusive.
Sustained HIV-1 remission for more than 7 years has been demonstrated in a patient with a chronic infection (the "Berlin patient") who had myeloablative, allogeneic hematopoietic stem cell transplantation (HSCT) for acute myelogenous leukemia using cells from a donor with a homozygous 32-base pair deletion in the gene encoding CCR5, a coreceptor for HIV-1 (12-14). The extent of reduction in the pool of latently infected cells in the blood and other tissues required to achieve sustained HIV-1 remission is unknown.
We previously reported reduction in peripheral blood HIV-1 reservoirs after an allogeneic HSCT with reduced-intensity conditioning in 2 male patients with chronic HIV-1 infections (15). Both patients had heterozygous 32-base pair deletions in the gene encoding CCR5 but received HIV-1-susceptible, wild-type donor cells. Neither HIV-1 DNA nor HIV-1 RNA was detectable in circulating CD4+ cells or plasma, respectively, after donor cells replaced host cells under the cover of suppressive ART. Anti-HIV antibody levels and avidity decreased after HSCT, a phenomenon also seen in the Berlin patient (14-15). However, extensive sampling of tissues and large numbers of peripheral blood mononuclear cells (PBMCs) for the presence of HIV-1 is necessary to understand the full effect of allogeneic HSCT on HIV-1 persistence. Analytic treatment interruption (ATI) is necessary to establish that viral remission has been achieved. Therefore, we conducted an in-depth analysis of HIV-1 persistence in various tissues from our patients and performed closely monitored ATIs.
Tissue Collection, Apheresis, and Sample Processing
The Dana-Farber/Harvard Cancer Center Office for Human Research Studies (Boston, Massachusetts) approved this study. Both patients had previously provided consent for blood and excess tissue sampling and were aware of these results before participating in this study. After full human research committee review, we initiated a second informed consent process that offered patients the option to have PMBC collection by apheresis, cerebrospinal fluid (CSF) collection, gut-associated lymphoid tissue sampling by anoscopy, and carefully monitored ATIs after reservoir characterization with permission from their clinical care providers. Peripheral blood mononuclear cells were collected by leukapheresis and were purified by Ficoll-Hypaque density gradient centrifugation. In addition, up to 120 mL of whole blood were obtained from patients at 3-month intervals for plasma collection and isolation of PBMCs. Cells were used when fresh or cryopreserved for later testing. Rectal biopsy samples were either flash frozen or cells were disassociated and cryopreserved after Percoll centrifugation, as previously described (16-17).
Assays to Quantify and Characterize HIV-1 Reservoirs
We extracted DNA from PBMCs and gut tissue using the QIAamp DNA Blood Mini Kit or the AllPrep DNA/RNA Mini Kit. HIV-1 DNA was quantified using a sensitive real-time polymerase chain reaction (PCR) assay, as previously described (18-19). Polymerase chain reaction assays were performed in as many as 42 assay wells. A single-copy assay capable of detecting plasma HIV-1 RNA at a lower limit of detection of 0.4 copies/mL was performed (20). To determine the presence of replication-competent HIV-1 before ATI, viral outgrowth assays were performed in 30 to 32 replicate assays using aliquots of 5 million purified CD4+ T-cells, for a total of 150 million or more CD4+ cells (21-23). The Cobas TaqMan HIV-1 Test, version 2.0, was used to quantify HIV-1 RNA in CSF. Homology among proviral DNA before HSCT and primers and probes used in the quantitative PCR assays was verified for each patient, and positive controls were incorporated in each assay, as previously described (15). Cells from a patient who had autologous HSCT with detectable HIV-1 DNA were used as positive controls in the viral outgrowth experiments. A single-genome analysis of near-full-length HIV-1 envelope sequences (approximately 2.5 kb) was done on cell-associated DNA before loss of detectable HIV-1 DNA and on plasma RNA after viral rebound, as previously described (24). Maximum likelihood phylogenetic trees of single-genome sequences were constructed using PhyML.
HIV-1 Specific Immune Responses and Host Microchimerism
Longitudinal quantification of HIV-specific antibody avidity by limiting-antigen enzyme immunoassay and antibody levels using the less sensitive (1:400 diluted plasma) Vitros Anti-HIV-1 + 2 assay were performed, as previously described (25). Peripheral blood mononuclear cells were used to perform interferon-γ enzyme-linked immunospot assays incorporating overlapping peptides that represented the gag and nef consensus protein sequence of clade B HIV. Selected HIV-1 peptides (spanning gp160, vif, nef, p24, p17, reverse transcriptase, and protease) described to be presented by the patients' HLA class I molecules were used to determine HIV-1-specific cellular immune responses (26). Highly sensitive allele-specific PCR assays targeting HLA and insertion-deletion polymorphisms unique to the patient or donor were used to determine levels of host microchimerism in blood (that is, the proportion of residual host PBMCs after HSCT) (27-28). The microchimerism assay is highly specific and sensitive to a single copy of target DNA, allowing detection of host cells present as a very low proportion of the PBMC population depending on the number of cells surveyed (27).
Carefully monitored ATIs were performed incorporating weekly viral load (VL) testing by the Cobas TaqMan HIV-1 Test, version 2.0 (limit of quantification of RNA levels of 20 copies/mL and limit of detection as low as 5 to 10 copies/mL), during the first 10 weeks after ART discontinuation and every 1 to 2 weeks thereafter. Qualitative testing of HIV-1 DNA from whole blood (Quest Diagnostics) was done approximately every 2 weeks. According to our institutional research guidelines, we were only able to provide patients with live-time results from approved clinical assays during this study. We planned to restart a patient's previous ART at the first sign of plasma viral rebound (any HIV-1 RNA level >1000 copies/mL or any confirmed viral load >200 copies/mL). Large-volume blood collections for detailed immunologic and virologic reservoir characterization as described previously were planned every 12 weeks before and after cessation of ART. HIV-1 genotyping of plasma virus was done either by Quest Diagnostics or by our laboratory, as previously described (29).
Role of the Funding Source
This study was supported by the American Foundation for AIDS Research and the National Institute of Allergy and Infectious Diseases. The funding source had no role in the design or analysis of the study, the preparation of the manuscript, or the decision to submit the manuscript for publication.
Virologic and Immunologic Characteristics Before ART Discontinuation
Patient A was perinatally infected, and he had allogeneic HSCT with reduced-intensity conditioning for recurrent Hodgkin lymphoma. Patient B was a man with sexually acquired HIV who had allogeneic HSCT with reduced-intensity conditioning for the myelodysplastic syndrome after treatment of non-Hodgkin lymphoma and subsequent Hodgkin lymphoma. Both patients received sirolimus, tacrolimus, and short-course methotrexate to prevent acute graft-versus-host disease after HSCT. Patient A developed chronic graft-versus-host disease of the skin, eye, and liver and was treated with prednisone 9 months after HSCT with initial response. Patient B developed graft-versus-host disease of the skin, liver, and oropharynx 220 days after HSCT, requiring intermittent oral prednisone.
Given the lack of detectable HIV-1 from peripheral blood, as previously reported (15), patients A and B were approached again and provided written informed consent to have in-depth sampling. Leukapheresis was done 4.3 years after HSCT for patient A and 2.6 years after HSCT for patient B. Table 1 summarizes virologic tests done before ATI. No HIV-1 DNA was detected from PBMCs by sensitive quantitative PCR assay, and no replication-competent HIV-1 was recovered from coculture assays involving 150 million or more purified CD4+ T-cells from either patient. Patient B consented to have rectal biopsies, in which no HIV-1 DNA could be detected. Microchimerism testing revealed that less than 0.0010% of peripheral blood cells were of host origin 1416 and 736 days after transplantation for patients A and B, respectively.
No significant HIV-1-specific cellular immune responses were detected in PBMCs of either patient before or after HSCT and before ATI by enzyme-linked immunospot assay (<30 spots per 106 PBMCs). However, PBMCs from patient B, who had a history of cytomegalovirus infection, demonstrated activation on stimulation with pooled cytomegalovirus, Epstein-Barr virus, and influenza peptides 652 days after HSCT.
Treatment Interruption and Viral Rebound
Given the lack of detectable HIV-1 despite extensive sampling, we performed ATI with close monitoring for virologic rebound. The study team and all clinical providers, including oncologists and infectious disease specialists, were aware of results from additional reservoir testing and made joint decisions to offer ATI to each patient. After thorough discussion of the uncertain significance of virologic assays and potential risks of ATI, including viral rebound, the acute retroviral syndrome, or graft-versus-host disease exacerbation, both patients consented to interrupt ART. Figure 1 shows the clinical, virologic, and immunologic course after ART discontinuation. Table 2 shows results from frequent blood sampling for quantification of cell and plasma virus. During ATI, patient A and B had no detectable plasma RNA or cell-associated HIV-1 DNA until 12 and 32 weeks after ART cessation, respectively. Patient A continued treatment with tacrolimus for persistent graft-versus-host disease during ATI; patient B did not require treatment during this period.
In patient A, viremia was first detected in plasma by clinical VL assay 84 days after ATI (RNA level of 904 copies/mL), 14 days after a negative clinical VL test result. He was initially asymptomatic and was asked to immediately restart his previous ART regimen (tenofovir plus emtricitabine plus efavirenz), according to protocol, after follow-up testing 4 days later revealed a VL increase to 127 843 copies/mL and HIV-1 in his CSF (<20 copies/mL). His VL subsequently peaked at an RNA level of 4.2 million copies/mL 117 days after ATI, which may have been due in part to suboptimal medication adherence. HIV resistance testing done 112 days after ATI revealed a new efavirenz resistance mutation (K103N), as well as resistance to several protease inhibitors (consistent with his history of previous treatment regimens), with the exception of tipranavir and darunavir. Given the emergence of drug resistance and rapidly increasing plasma viremia of uncertain cause, the patient was prescribed tenofovir plus emtricitabine plus raltegravir plus ritonavir-boosted darunavir and reduced-dose tacrolimus 117 days after ATI. However, he developed nausea, vomiting, headaches, and fevers thought to be secondary to the acute retroviral syndrome, and he was only able to take several doses. He was evaluated at an outside hospital emergency department 120 days after ATI, where he had CSF sampling and was found to have a low-grade lymphocytic pleocytosis (11 leukocytes per high-powered field) consistent with HIV-associated meningitis, but HIV-1 RNA was not measured in his CSF at this time. His symptoms may have been exacerbated by ART and tacrolimus coadministration.
The patient was discharged but continued to have difficulty taking medication, given central nervous system symptoms and vomiting thought to be due to the acute retroviral syndrome and potential medication side effects. He was admitted to our hospital for work-up and observed reinitiation of his ART. The tacrolimus trough level peaked at 104 ng/mL (desired range, <10 ng/mL) shortly after hospital admission when receiving concomitant ritonavir. He also had an increased serum creatinine level of 150.3 μmol/L (1.7 mg/dL), but renal function and tacrolimus levels returned to normal within 4 days after both ART and tacrolimus were withheld. His symptoms resolved with decreasing viral RNA levels before ART reinitiation; ART was resumed along with very low doses of tacrolimus (0.5 mg every 3 weeks) 133 days after ATI with continued viral decay.
Patient B had frequent negative HIV-1 clinical laboratory test results during ATI and undetectable PBMC-associated HIV-1 DNA and plasma HIV-1 RNA by quantitative PCR assays (limits of detection, DNA level of 0.2 and 0.5 copies/106 PBMCs and RNA level of 0.4 copies/mL) 5 and 18 weeks after ATI (Table 2). However, he developed fevers, malaise, and fatigue 219 days after ATI, 8 days after a negative clinical VL test result. The patient presented to an outpatient urgent care center with worsening symptoms 225 days after ATI and was found to have a plasma HIV-1 RNA level of 1.9 million copies/mL.
He promptly resumed ART 228 days after ATI with tenofovir plus emtricitabine and dolutegravir, and symptoms resolved with subsequent decay of plasma viremia (Figure 1). No resistance-associated mutations were detected in plasma virus 233 days after ATI despite a history of nonnucleoside reverse transcriptase inhibitor resistance before HSCT. In addition, he had HIV-1 RNA levels of 269 copies/mL in his CSF 238 days after ATI.
Single-genome, near-full-length HIV-1 envelope sequences from rebound viremia in both patient A (44 sequences) and patient B (50 sequences) were related to peri-HSCT HIV-1 DNA sequences in phylogenetic analyses. Viral sequences from each patient after rebound were monophyletic with a high degree of intra-sample sequence homology (Appendix Figure).
HIV-1-Specific Immunity After Treatment Discontinuation
No activation of PBMCs in response to HLA-matched peptides or pooled overlapping HIV-1 peptides in enzyme-linked immunospot assays were detected in samples from patients A and B before ATI, or from samples from patient B obtained before viral rebound at 38 and 129 days after ATI (PBMCs were not collected for time points after ATI but before rebound in patient A). Activation was seen in response to an HLA-A02-restricted nef peptide in PBMCs obtained from patient B 13 days after the first detectable viral RNA (238 days after ATI). No activation to HIV-1 peptides was detected in PBMCs collected from patient A 5 days after viral rebound (89 days after ATI), but low-level activation to gag and nef peptides and higher-level activation to HLA-A02-restricted epitopes in p24 and p17 were detected in PBMCs collected 108 days after viral rebound (193 days after ATI).
Figure 2 shows HIV-1 antibody levels by less sensitive Vitros assay and avidity over time. Antibody avidity declined in both patients during virologic suppression after HSCT. Antibody levels and avidity increased after viral rebound in patient A, in whom virologic control was not immediately reestablished. Despite declining avidity after rebound in patient B, who promptly resumed ART, antibody levels increased 13 days after the first detectable VL measurement.
Substantial reductions in HIV-1 reservoirs were seen in 2 patients who had allogeneic HSCT with HIV-1-susceptible donor cells while receiving continuous ART. Despite readily detectable proviral DNA before HSCT (15), HIV-1 RNA, proviral HIV-1 DNA, and viral outgrowth were undetectable years after HSCT using highly sensitive assays applied to plasma, PBMCs, and gut-associated lymphoid tissue. Nevertheless, HIV-1 rebound was first detected 12 and 32 weeks after ART interruption. Rebound of viremia occurred within 2 weeks after the last negative plasma HIV-1 RNA result, and both patients developed symptoms consistent with the acute retroviral syndrome. Based on previously reported proviral DNA levels before or immediately after HSCT in both patients (96 to 144 copies/106 PBMCs) and the detection limits of sensitive quantitative assays used in this study, our findings suggest a minimum 3-log10 reduction in the number of circulating cells harboring proviral HIV-1 DNA after HSCT. Although allogeneic HSCT may lead to significant, sustained reductions in the HIV-1 reservoir, infected tissue or cell-bound virus persists. Persistence of these small numbers of residual infected cells seems to be sufficient to rekindle HIV-1 replication.
Our patients differed from the Berlin patient, who achieved sustained HIV-1 remission after 2 myeloablative HSCTs with resistant donor cells incapable of supporting HIV-1 replication in the event of reactivation of residual latent reservoirs (12-14). However, phylogenetic analyses indicated that only one or a few cells or latent proviruses contributed to viral rebound after ATI in our patients. The role of specific conditioning regimens and the use of various anti-inflammatory medications after HSCT and HIV-1 persistence are largely unknown and warrant further study.
HIV-1 typically rebounds 1 to 8 weeks after ART discontinuation (6-11). One person who had undetectable HIV-1 DNA in blood or tissue but very low levels of detectable infectious virus from virus outgrowth assays had viral rebound 7 weeks after ATI (22). Posttreatment control of viral replication has been seen in patients initiating ART during acute HIV-1 infection, and several of these persons had delayed rebound after stopping ART (30-31). However, the virus has been detected intermittently in patients who were treated early with subsequent control of the virus after ART discontinuation, and durable control was achieved in relatively few persons (30-31). Despite frequent sampling, neither of our patients had detectable HIV-1 in PBMCs or plasma for several months after ART discontinuation before viral rebound. In addition, ultrasensitive microchimerism testing showed that nearly all PBMCs were of donor origin years after HSCT.
A chronic, ongoing graft-versus-host reaction may have been responsible for the continued surveillance and clearance of residual recipient hematopoietic cells that survived conditioning chemotherapy for our patients and the Berlin patient, some of which happened to harbor latent HIV-1. Graft-versus-tumor and more generalized graft-versus-hematopoietic effects may exist without the development or persistence of clinical graft-versus-host disease and are mediated by innate immunity and natural killer cells in addition to T-lymphocyte activity (32-33). Furthermore, the lack of detectable HIV-1 DNA and replication-competent virus before ATI supports the hypothesis that donor cells in various tissues (such as blood and gut) were largely protected from infection by ART.
Reductions in the HIV-1 reservoir have been described in patients having myeloablative allogeneic HSCT in the setting of zidovudine monotherapy or suppressive ART (34-38). Detailed data on ART interruption after allogeneic HSCT are limited to the report of a person who had a reduction in HIV-1 DNA shortly after myeloablative HSCT and full donor chimerism (34). That patient developed grade III graft-versus-host disease of the skin and gastrointestinal tract and had rapid viral rebound within 16 days of stopping ART 4 months after transplantation (34). In contrast, our patients had been on ART for 2 or more years after HSCT (15) and achieved months of ART-free viral remission. Chronic graft-versus-host effects without clinically significant disease may have led to more profound reductions in viral reservoirs and ultimately delayed return of the virus. The longer interval between HSCT and ATI may also have contributed to a longer period of HIV-1 remission in our patients.
Long-lived tissue reservoirs, including host macrophages that are replaced more slowly than T-lymphocytes after HSCT (12), may have contributed to viral rebound. The recipient's residual pretransplant lymphoid tissue may have persisted despite a very high degree of donor blood chimerism, or donor cells that were inaccessible to peripheral blood and tissue sampling had become infected. For example, only a limited number of CD4+ T-cells were able to be surveyed from gut tissue, and more intensive sampling may have led to the detection of HIV-1. Low levels of detectable HIV-1 RNA were identified in CSF after viral rebound but were orders of magnitude lower than peripheral blood VLs. We could not obtain CSF during ATI before rebound, and further studies of tissue localization and cellular composition of this reservoir are needed.
Patients who have allogeneic HSCT experience variable HIV-specific cellular immune responses after transplantation (38). Little to no T-cell response to HIV-1 peptides was seen in our patients after HSCT or during ATI until at least 13 days after viral rebound. Thus, it seems that virus-specific adaptive immunity played little role in controlling HIV-1 replication before rebound. The extent to which nonspecific innate immunity and graft-versus-host effects influenced the duration of ART-free remission before viral rebound is not well-defined and warrants further investigation. Both patients also had a decrease in and subsequent low-level persistence of HIV-1-specific antibodies and avidity after allogeneic HSCT. Antibody levels increased shortly after viral rebound in both patients, but avidity continued to decrease in samples from patient B, who promptly reinitiated ART. Although the Berlin patient had an even further decrease in antibody levels and avidity after HSCT, his antibodies persisted after transplantation for more than 5 years (14). The sources of residual antibodies in the HSCT patients are unknown and warrant further study.
As a result of allogeneic HSCT from donors who were not exposed to HIV, the reconstituted immune systems of our patients were HIV-naive, as reflected by the absence of detectable virus-specific cellular immune responses. When HIV-1 rebound occurred, it mimicked the kinetics seen during acute HIV-1 infection (39). Given the rapid virologic rebound, the development of accompanying symptoms, and the emergence of a new nonnucleoside reverse transcriptase inhibitor mutation despite closely monitored ART reinitiation, any future studies of ATI in the setting of allogeneic HSCT should proceed with the utmost caution. Given the limited sensitivity of currently available assays for detecting viral persistence, however, analytic treatment interruption remains the only reliable means of assessing the extent of HIV-1 reservoir depletion after therapeutic interventions.
In summary, our results suggest that allogeneic HSCT with CCR5 wild-type donor cells may lead to loss of detectable HIV-1 from blood and rectal mucosa, but viral rebound may nevertheless occur after ART interruption despite a significant reduction in reservoir size. The definition of the nature and half-life of residual viral reservoirs is essential to achieve durable, ART-free HIV-1 remission.