| |
Rapid HIV RNA rebound after antiretroviral treatment interruption in persons durably suppressed in Fiebig I acute HIV infection
|
| |
| |
Download the PDF here
"Our study documented that in eight participants, ART initiated in Fiebig I did not prevent viral load rebound."
Brief Communication |
Nature Medicine (June 11 2018) - Donn J. Colby1, Lydie Trautmann2,3, Suteeraporn Pinyakorn2,3, Louise Leyre4, Amelie Pagliuzza4, Eugene Kroon1, Morgane Rolland2,3, Hiroshi Takata2,3, Supranee Buranapraditkun2,3,5,6, Jintana Intasan1, Nitiya Chomchey1, Roshell Muir7, Elias K. Haddad7, Sodsai Tovanabutra2,3, Sasiwimol Ubolyam8, Diane L. Bolton2,3, Brandie A. Fullmer9, Robert J. Gorelick9, Lawrence Fox10, Trevor A. Crowell2,3, Rapee Trichavaroj11, Robert O'Connell11, Nicolas Chomont 4, Jerome H. Kim2,13, Nelson L. Michael2, Merlin L. Robb2,3, Nittaya Phanuphak1, Jintanat Ananworanich 1,2,3,12*
and The RV411 study group
Abstract
Antiretroviral therapy during the earliest stage of acute HIV infection (Fiebig I) might minimize establishment of a latent HIV reservoir and thereby facilitate viremic control after analytical treatment interruption. We show that 8 participants, who initiated treatment during Fiebig I and were treated for a median of 2.8 years, all experienced rapid viral load rebound following analytical treatment interruption, indicating that additional strategies are required to control or eradicate HIV.
Main
HIV latency is established during acute HIV infection in long-lived CD4+ T cells and other cell types in blood and tissue sanctuaries1,2,3. These reservoirs pose a major obstacle to HIV remission (viremic control without antiretroviral therapy (ART)). Despite viral suppression by ART, most individuals experience viral load rebound within two to four weeks of stopping ART. HIV remission appears more common in a limited number of early-treated individuals analyzed4. Fiebig I stage corresponds to the first two weeks following infection when HIV nucleic acid is detected in the absence of p24 antigen (viral capsid core protein) and HIV seroconversion. The brevity of this period and non-reactivity to HIV serologic testing means that HIV is rarely diagnosed in Fiebig I. Treatment in these individuals can lead to preserved immunity, remarkably low HIV reservoir size and no seroconversion5,6, but its impact on post-treatment control is unknown. This study evaluated viremic control after analytical treatment interruption (ATI) in Fiebig I treated individuals, and had planned to enroll 15 participants in 2 stages: 8 in stage 1 and 7 in stage 2. It was stopped after stage 1 when 0 of 8 participants met the viremic control end point (Supplementary Fig. 1). This study provides proof-of-concept that viral load rebound will occur despite very early and fully suppressive ART. The sample size of eight participants did not preclude a rate of control below 30%.
At enrollment, there were 7 men and 1 woman with a median age of 29 years, an ART duration of 2.8 years, a CD4+ T count of 577 cells per mm3 and a suppressed viral load (Supplementary Table 1). All had been on efavirenz-based ART and switched from efavirenz to darunavir/ritonavir 4 weeks prior to ATI to limit resistance to that drug class. No participant achieved viral load <50 copies per ml at 24 weeks post ATI. All resumed ART following two viral load measurements >1,000 copies per ml, ranging from 1 to 2 d apart per protocol. Viral load rebound, defined as viral load >20 copies per ml, was observed at a median of 26 d (range 13-48 d) (Fig. 1a and Supplementary Fig. 2). Single copy viral load assay (SCA) on samples with viral load <20 copies per ml showed viral load <0.45 copies per ml at baseline in all participants, and two had detectable SCA during ATI prior to viral load rebound (Supplementary Fig. 3).
CD4/CD8 ratio ≤1 was associated with faster time to viral load rebound in a study of early treated people7. We also observed this in treated Fiebig I individuals (Fig. 1b). The median CD4+ T cell change at ART resumption was -9 (range -87 to 39) cells per mm3. No participant had acute retroviral syndrome, HIV-related symptoms or new drug resistance mutations (Supplementary Information and Supplementary Table 2). Viral load was <50 copies per ml in all participants by a median of 17 d (range 9-63) after ART resumption.
Phylogenetic analyses showed that sequences at time of ART initiation in acute infection and at ART resumption following viral load rebound were intermingled in the trees, with no evidence of distinct sub-clusters (Fig. 1c and Supplementary Fig. 4a-c). The convergence of different sequence analyses suggests that viral load rebound resulted from production of viral particles from latently infected cells (Supplementary analyses), supporting the survival and/or clonal expansion of latently infected CD4+ T cells during ART8,9.
The frequencies of blood CD4+ T cells harboring total HIV DNA were low pre-ART (median 66 copies per 106 CD4+ T cells, interquartile range (IQR) 22-281) and pre-ATI (median 1 copy per 106 CD4+ T cells, IQR 1-3) (Fig. 2a). Prior to ATI, our participants had an estimated ∼2 x 106 CD4+ T cells that harbored HIV DNA in the whole body. Total HIV DNA increased to a median of 17 copies per 106 CD4+ T cells (IQR 6-60) at the time of ART resumption (P = 0.03 compared to pre-ATI). However, 6 months after resuming ART, total HIV DNA returned to a median of 3 copies per 106 CD4+ T cells (IQR 2-10), similar to pre-ATI levels (P = 0.33) (Fig. 2a), and demonstrating safety of this ATI strategy with a short duration of viremia. The total and integrated HIV DNA of individual participants showed similar patterns and returned to approximate pre-ATI values (Supplementary Fig. 5). Inducible HIV RNA by the Tat/rev Induced Limiting Dilution Assay (TILDA) showed pre-ATI levels below the detection limit of 1.4 cells producing tat/rev mRNA per 106 CD4+ T cells in all participants.
In this small study, there was no association between pre-ATI total HIV DNA and time to viral load rebound10 (Fig. 2b). Similarly, pre-ATI integrated HIV DNA and 2-LTR (long terminal repeat) circles, which reflect HIV reservoir maintenance and putatively residual replication, respectively, did not predict time to viral load rebound, nor did any of the HIV reservoir markers (total and integrated HIV DNA, 2-LTR circles) assessed at pre-ART (Supplementary Fig. 6a-c). A very low frequency of HIV-infected memory CD4+ T cells at time of ART resumption was observed among the four participants assessed, with LTR or gag RNA+ cells present at 0.001-0.01% (Supplementary Fig. 7a,b). HIV DNA content in mucosal mononuclear cells from sigmoid tissue (n = 3) was also low at pre-ATI and at ART resumption (Supplementary Information). HIV RNA+ and HIV DNA+ cells in inguinal lymph node tissue were detected by in situ hybridization in participant 4,878, who consented to the biopsy both at pre-ATI and at ART resumption (vRNA+ cells 1.8 x 105 and 2.4 x 105 cells per gram and vDNA+ cells 4.2 x 105 and 5.1 x 105 cells per gram, respectively) (Supplementary Fig. 8a-d). The finding of vRNA+ cells during suppressive ART suggests ongoing virus production in the lymph nodes. HIV DNA content and Ki67+CD8+ T cells are shown in the Supplementary analyses and Supplementary Fig. 9a,b.
Reactivation of a single latently infected cell can lead to viral load rebound in the absence of ART. One mathematical model postulates that control of viremia may depend on the strength of cytotoxic T lymphocytes (CTLs) and the size of the HIV reservoir11. We posit that the rapid viral load rebound observed in this study was due to the inability to achieve a small enough pool of latently infected cells12, particularly in lymphoid tissues13, and inadequate immune control11. The reservoir in acute HIV infection may also be enriched in replication competent viruses, facilitating viral rebound14. We previously reported that CTLs generated in Fiebig I are less abundant and differentiated than in later stages, but retain greater memory and survival potential6. We therefore investigated persistence and recall of memory immune responses during ART and post ATI. In three participants expressing HLA-A11, the memory HIV-specific CTLs quantified longitudinally by tetramer staining6 persisted at low levels years after suppressive ART, and their expansion occurred faster following viral load rebound than in acute HIV infection (Supplementary Fig. 10a-d). In all participants, there was a rapid expansion of effector Ki67+CD8+ T cells at ART resumption (median 12.4%) to frequencies higher than those at baseline ATI (median 5.4%, P = 0.01) (Fig. 2c). Env-specific immunoglobulin (Ig)G level was higher at ART resumption than in acute infection in 7 of 8 participants (median 98 versus 34 µg per ml, P = 0.04) (Fig. 2d). Four of these 7 had increased levels from baseline ATI to ART resumption (median 37.8 to 187.2 µg ml-1, P = 0.07) and seroconverted after ATI (Supplementary Fig. 11), suggesting a memory B cell response. Prior to ART resumption, the highest observed plasma viral load levels were negatively associated with the frequency of effector Ki67+CD8+ T cells (r = -0.83, P = 0.02; Fig. 2e) and Env-specific IgG levels (r = -0.86, P = 0.01; Fig. 2f). It is therefore possible that pre-ATI immune responses were “too little too late” to affect time to viral load rebound, but were sufficient to affect the magnitude of viremia once expansion of HIV-specific CD8+ T cells and secretion of Env-specific IgG occurred. Our stringent ART resumption criterion based on viral load precluded assessing the effects of early ART on spontaneous viral load re-suppression after rebound. It was demonstrated previously that an initial sustained viremia (highest viral load 103-107 copies per ml) was followed by viral load <40 copies per ml in 2 of 15 (13%) early treated participants at 10 months post ATI15. The optimal design and conduct of ATI trials is unknown, but can be informed by ongoing dialogues between researchers, the community, ethicists, social scientists, and regulatory bodies.
Our study documented that in eight participants, ART initiated in Fiebig I did not prevent viral load rebound. Regardless of timing of ART, future research should aim to eliminate cells with replication competent HIV in blood and tissues through augmenting immune responses (for example, through the administration of latency reversal or immune adjuvant agents, broadly neutralizing antibodies (bNAbs), therapeutic vaccines or cell-based therapies)16. Early administration of immune interventions during acute infection should be explored as demonstrated in macaques to improve viral load control17,18. Limited viral escape in acutely treated individuals may enhance the responses to such interventions. Combination therapies (for example, multiple bNAbs and immune modulators) should be investigated as single bNAb administration in acutely and chronically treated people did not substantially delay viral load rebound19.
|
|
| |
| |
|
|
|
