Advances toward Curing HIV-1 Infection in Tissue Reservoirs

A disease of more than 39.6 million people worldwide, HIV-1 infection has no curative therapy. To date, one man has achieved a sterile cure, with millions more hoping to avoid the potential pitfalls of lifelong antiretroviral therapy and other HIV-related disorders, including neurocognitive decline. Recent developments in immunotherapies and gene therapies provide renewed hope in advancing efforts toward a sterilizing or functional cure.

ABSTRACT

A disease of more than 39.6 million people worldwide, HIV-1 infection has no curative therapy. To date, one man has achieved a sterile cure, with millions more hoping to avoid the potential pitfalls of lifelong antiretroviral therapy and other HIV-related disorders, including neurocognitive decline. Recent developments in immunotherapies and gene therapies provide renewed hope in advancing efforts toward a sterilizing or functional cure. On the horizon is research concentrated in multiple separate but potentially complementary domains: vaccine research, viral transcript editing, T-cell effector response targeting including checkpoint inhibitors, and gene editing. Here, we review the concept of targeting the HIV-1 tissue reservoirs, with an emphasis on the central nervous system, and describe relevant new work in functional cure research and strategies for HIV-1 eradication.

INTRODUCTION

Control of human immunodeficiency virus (HIV) infection can be achieved with optimum antiretroviral therapy (1), but this does not provide a disease cure. HIV persists in latently infected CD4⁺ T cells and is integrated into the host genome until cell death. The introduction of the first antiretroviral drug, zidovudine, in 1987, followed by combination antiretroviral therapy (ART) in 1996, transformed the epidemic by increasing survival and converting HIV infection from a life-threatening illness into a chronic infection. However, ART is not available to all, and compliance rates vary. Even in carefully managed patients, the consequences of lifelong medication side effects are challenging, and disease trajectory is variable. There is concern about frailty and accelerated aging with end-organ involvement. This includes HIV-associated neurocognitive disorders (HAND) and major depressive disorder (MDD), which can lead to gradual but significant functional deterioration that complicates their management (2). For these reasons, much hope has rested on finding a cure for the infection.

Despite massive efforts to eradicate HIV, challenges persist in achieving a “cure.” The history of cure strategies is one of repeated failures and false hopes. When ART was first introduced, it was thought that the size of the HIV DNA reservoir would decrease over time on prolonged therapy and the virus would eventually be cleared. Early mathematical models, such as the Ho-Shaw model, were published in high-profile journals and received wide publicity (3 – 5). Unfortunately, the ability of HIV to persist in resting CD4 T cells and tissue reservoirs almost indefinitely despite ART was underappreciated and only later fully recognized (6 – 8).

A major shift in HIV research occurred when the “Berlin patient” achieved a sterile cure following radiation treatment, chemotherapy, and a hematopoietic stem cell transplant (HSCT) from a donor with a homozygous deletion in the gene encoding the HIV coreceptor C-C chemokine receptor type 5 (CCR5Δ32) (9). Although subsequent studies have been unable to reproduce the success using homozygous CCR5Δ32 transplants (10 – 14), there have been isolated cases of suspected “cures,” such as the “Mississippi baby,” a perinatally infected child who had an undetectable viral load for 27 months after stopping ART before rebound (15, 16); this case led to the hypothesis that if children received ART at birth, treatment could be safely stopped after some time. In each case, viral rebound eventually occurred, suggesting that tissue reservoirs are established early in the disease and cannot be eradicated by ART alone. Recently, a second potential sterile-cure case, the “London patient,” was identified (17). This patient underwent a much less aggressive preconditioning for HSCT from a donor homozygous for the CCR5Δ32 allele after receiving chemotherapy for relapsed nodular sclerosing Hodgkin’s lymphoma. Unlike the Berlin patient, the London patient was homozygous for wild-type CCR5 prior to transplant and did not receive whole-body irradiation. The London patient has remained in HIV-1 remission for more than 18 months.

The term “HIV reservoir” has been variably used to describe the cells and tissues that continue to harbor HIV under optimum therapy. Some definitions include all proviruses that participate in HIV pathogenesis regardless of replication competence (18), while others include only replication-competent forms of HIV that are capable of reestablishing infection in the absence of ART (19). We prefer the former definition, since mutations and recombination events can convert a replication-incompetent viral sequence to a replication-competent virus. Further, a defective viral sequence may harbor open reading frames for individual viral proteins or produce noncoding RNA which may have biological effects. HIV reservoirs can further be divided according to scale into anatomical, cellular, and molecular reservoirs (Fig. 1), which are comprised of the organs/tissues that harbor infected cells, the HIV-containing cells themselves, and the characteristics of the HIV genomes contained within those cells, respectively. At the molecular level, the reservoir can be further subdivided into latent, persistent, and replication-defective categories. In a latent reservoir, the virus achieves complete transcriptional silencing and must be reactivated to release a replicating viral particle. HIV infection of the resting memory CD4 T cells fulfills this definition (20). However, any long-lived cell infected with the virus could serve as a persistent viral reservoir where small amounts of virus are continually or intermittently released. Macrophages likely fall into this category, though their exact contribution to chronic HIV infection is undetermined. While monocyte-derived macrophages can be persistently infected in vitro (21, 22), there is considerable debate about whether myeloid cells can support latent HIV infection (23 – 28). However, macrophages can harbor transcriptionally active unintegrated HIV DNA for more than 30 days in culture (29).

FIG 1.

Summary of types of HIV reservoirs. (A and B) There are several anatomical compartments (A) that are populated by HIV-infected cells (B). (C) The integrated provirus contained within these cells may be transcriptionally silent (latent), transcriptionally active and capable of producing infectious virions (persistent), or transcriptionally active but replication defective due to mutations or deletions in the HIV genome, leading to translation of specific viral proteins for which an open reading frame remains intact.

Data from simian immunodeficiency virus (SIV) models suggest that viral DNA (vDNA) within tissue-resident macrophages is often due to phagocytosis of infected CD4⁺ T cells rather than true infection (30, 31). The researchers observed that vDNA was contained in macrophages only in tissues that were not depleted of CD4⁺ cells (30) and that no replication-competent virus could be detected from macrophages of animals treated with ART (31). Similarly, vDNA could not be detected in alveolar macrophages isolated from HIV-positive patients on long-term ART with undetectable viral loads (31). However, others have shown that phagocytosis of infected CD4⁺ T cells can yield productive macrophage infection (32). In humanized myeloid-only mice (MoM) infected with HIV and suppressed with ART, viral rebound occurred in 3/9 (33%) mice 7 weeks after treatment was removed (33). Further, macrophages isolated from the urethras of three individuals on suppressive ART contained not only integrated vDNA but also HIV RNA, proteins, and viral particles, and they could produce replication-competent virus when stimulated with lipopolysaccharide (34). Together, these findings support the establishment of a myeloid reservoir in some HIV-infected individuals. Microglial cells and perivascular macrophages containing integrated vDNA have also been detected in postmortem central nervous system (CNS) tissue (35), which supports a myeloid reservoir in the brain. It is now important to better elucidate the characteristics of the macrophage reservoir, particularly because these cells are long-lived and resist the cytopathic effects of HIV (36).

Some cells harbor defective viral sequences. These cells, while incapable of producing infectious virus, may have open reading frames for viral proteins which may play a role in disease pathogenesis (37). There is also the possibility that either through a recombination event or via DNA repair mechanisms, viral production may occur. While these replication-defective viral sequences are poorly studied in the context of HIV infection, they have been extensively studied in the context of endogenous retroviruses, where the vast majority of the viruses are defective and may play a pathogenic role in neurodegenerative diseases and cancer (38). Hence a “sterilizing cure” should eradicate all three forms of molecular reservoirs. The terms “functional cure” and “remission” are used to describe approaches that prevent the production of infectious virus. However, it may be necessary to also control the production of all viral proteins to achieve a functional cure.

BRAIN RESERVOIR

While much is known about the lymphoid reservoirs in major end organs, the brain is difficult to study. Tissue is accessible only at autopsy, and inference during life is made by study of the cerebrospinal fluid (CSF) that bathes the brain. Substances that are unique to the CSF, such as divergence of viral strains between blood and CSF, or those found in higher concentrations in CSF than in blood are considered to be derived from the brain.

In well-controlled HIV-positive patients, immune activation can be present even when HIV is undetectable in the CSF, indicating a persistent response to the underlying infection (39). HIV proteins such as Tat have been found in the CSF, and antibody responses against Tat correlate with CSF viral load and inversely correlate with CD4 cell count (40). Interestingly, the presence and abundance of antibodies against HIV proteins in blood and CSF, especially p24, have been implicated as a representative measure of curative interventions and could potentially be used to evaluate reservoir eradication in the brain (41, 42). Studies on CSF viral escape have shown divergent CSF sequences compared to those in blood, suggesting compartmentalization of the brain reservoir in those patients (43 – 45).

While studies from the preantiretroviral era clearly show that HIV can infect cells within the brain, including microglia, astrocytes, macrophages, and even rare neurons (46, 47), it remains unknown if the reservoir persists in patients on ART. In SIV macaque models, myeloid cells in the brain are known to be a reservoir for replication-competent virus (48 – 50).

Studies using viral next-generation sequencing in the CSF and periphery have demonstrated distinct compartmentalization of viral subspecies, especially in patients with HIV-associated dementia (51). Up to 10% of patients on suppressive antiretroviral therapy experience low levels of CSF escape, characterized by detectable HIV RNA in CSF but not in blood (52). Our laboratory has found that HIV Tat protein can be detected in the CSF from >35% of HIV-positive individuals on suppressive ART (53), indicating that Tat levels may be an indirect measure of the brain viral reservoir. Collectively, these findings highlight the importance of considering multiple tissues, including the brain, in cure research strategies. Persistent Tat expression and CSF escape in otherwise well-controlled patients strongly suggest that a CNS reservoir exists in a subset of patients and may be capable of reseeding the periphery even if the memory T-cell reservoir is eradicated or silenced.

NOVEL DIRECTIONS IN HIV CURE STRATEGIES

The goal of HIV-1 remission strategies is to decrease the size of the replication-competent reservoir by (i) reactivating latent virus, termed “kick,” while improving anti-HIV immune responses to eliminate infected cells, termed “kill,” (ii) genetically engineering cells to make them resistant to HIV infection, or (iii) inducing epigenetic changes or chromatin remodeling to permanently prevent transcriptional activation of the viral genome, termed “block and lock.”

“Kick-and-kill” approaches to eliminate viral reservoirs.

Several strategies have been proposed to reactivate HIV expression from latently infected CD4⁺ T cells. Latency-reversing agents (LRAs) include the following: (i) epigenetic modifiers, including histone deacetylase inhibitors (HDACi) and histone methyltransferase inhibitors, which induce chromatin remodeling at the HIV LTR; (ii) protein kinase C (PKC) agonists and activators of the NFκB pathway; (iii) bromodomain extraterminal (BET) motif inhibitors, which enhance recruitment of P-TEFb to the HIV promoter; and (iv) cytokine/chemokine and Toll-like receptor (TLR) agonists, which activate immune signaling pathways in T cells. All of these are reviewed elsewhere (54 – 56), so our discussion is limited to agents being actively investigated in clinical trials or in early stages of development (Table 1).

TABLE 1.

Summary of active clinical trials involving latency-reversing agents^a

LRA	Secondary agent(s)	No. of patients	Status	Identifier
Nicotinamide (SIRT1 inhibitor)	Dendritic cell vaccine + auranofin + ART intensification	30	Active	NCT02961829
Vorinostat (HDACi)	ChAdV63.HIVconsv (ChAd) prime and MVA.HIVconsv boost vaccines	60	Active	NCT02336074
	Disulfiram	15	Terminated due to AE	NCT03198559
	HXTCb	12	Recruiting	NCT03212989
	Tamoxifen	30	Active	NCT03382834
	AGS-004 DC therapy	6	Terminated (AGS-004 supply unavailable)	NCT02707900 (VOR-VAX)
	VRC07-523LS	12	Recruiting	NCT03803605
Panobinostat (HDACi)	Pegylated IFN-α2a	34	Recruiting	NCT02471430
Romidepsin (HDACi)	3BNC117	30	Active	NCT02850016c
	MVA vector HIV vaccine + HIVACAR01 (personalized HIV vaccine) + 10-1074	56	Not yet recruiting	NCT03619278
	3BNC117 ab	60	Recruiting	NCT03041012
	3BNC117 ab	42	Not yet recruiting	RV 438
Valproic acid (HDACi)	Pyrimethamine	28	Recruiting	NCT03525730
Chidamide (HDACi)	None	60	Active	NCT02902185
	CAR-T or TCR-T-cell therapy	40	Recruiting	NCT03980691
Euphorbia kansui (ingenol) (PKC agonist)	None	9	Recruiting	NCT02531295
Lefitolimod (MGN1703) (TLR9 agonist)	10-1047 + 3BNC117	48	Recruiting	NCT03837756c
GS-9620 (TLR7 agonist)	None	28	Active	NCT03060447c
Pegylated IFN-α2a	None	54	Active	NCT02227277
	3BNC117 + 10-1074	21	Not yet recruiting	NCT03588715c
Recombinant human superagonist IL-15 (ALT-803/N-803)	None	10	Active	NCT02191098
	Haploidentical NK cell adoptive transfer	8	Recruiting	NCT03899480

^aBased on data from references 54 and 56. All treatments are in addition to ART, unless otherwise noted.

^bHXTC, HIV antigen expanded specific T-cell therapy.

^cThese studies involve analytical treatment interruption (ATI).

HIV transcriptional latency is mediated largely by epigenetic silencing (57, 58); hence, HIV cure research over the past decade has focused on HDACi (59, 60). There are currently 15 clinical trials using epigenetic modifiers in HIV-infected individuals, either alone or in combination with another LRA or immune-boosting therapy (Table 1). To date, none of these studies has exhibited a significant decrease in integrated DNA levels in the CD4⁺ T-cell reservoir or delay in viral rebound following treatment interruption, despite immune activation and increases in cell-associated and plasma HIV RNA (54 – 56).

TLR agonists have shown promise in nonhuman primate studies, but their use in humans is in early stages. The TLR7 agonist GS-9620, in combination with an Ad26/modified vaccinia virus Ankara (MVA) therapeutic vaccination, enhanced antiviral immune responses and delayed viral rebound in SIV-infected macaques (61). Similarly, coadministration of GS-9620 and a broadly neutralizing antibody (PGT121) prevented viral rebound following treatment interruption in simian-human immunodeficiency virus (SHIV)-infected macaques (62). This was not observed with either GS-9620 or PGT121 alone, indicating that TLR agonists require pairing with other therapeutics. A phase 1 clinical trial investigating GS-9620 in HIV-infected individuals on combination ART (NCT03060447) is under way (Table 1).

The interleukin-15 (IL-15) superagonist ALT-803, which has been shown to reactivate HIV in latently infected T cells and enhance cytotoxic T-cell responses ex vivo (63), is being evaluated in individuals on long-term ART (NCT02191098) (54). Another cytokine, polyethylene glycol-alpha interferon 2a (polyethylene glycol-IFN-α_2a), decreased HIV-1 DNA levels and delayed viral rebound after treatment interruption (64). A trial comparing the effects of a repeated low dose (1 μg/kg per week) versus a higher dose (180 mg per week) on viral reservoirs is under way (NCT02227277); it is also being investigated in combination with HIV-targeted broadly neutralizing antibodies (NCT03588715) (54).

Clinical data suggest that LRAs are insufficient to reduce the size of the viral reservoir when delivered as a monotherapy. Thus, ongoing studies combine different classes of LRAs or pair an LRA with an immune-boosting intervention such as therapeutic vaccination (Table 1) (54 – 56). Further, the antiviral immune response is impaired in chronic HIV infection due to chronic inflammation and immune exhaustion. Cytotoxic CD8⁺ T cells are unable to clear HIV-infected cells even with latency reversal, which contributes to the failure of “kick-and-kill” strategies (55, 65). Additionally, certain LRAs, such as HDACi, suppress cytotoxic T-cell responses (66). Therefore, immunomodulatory agents may be vital in successfully eliminating or reducing the viral reservoir.

Checkpoint inhibitor therapies.

Immune checkpoint molecules are cell surface receptors that regulate immune responses. Stimulatory immune checkpoint proteins provide costimulatory signals that enhance immune activation, while inhibitory immune checkpoint molecules negatively regulate immune cell function and play an important role in resolution of immune responses and maintenance of self-tolerance. During chronic viral infections, including HIV infection, upregulation of inhibitory checkpoint molecules on immune cells contributes to T-cell exhaustion characterized by loss of effector functions and failure to proliferate in response to antigen (67). Checkpoint molecules are enriched on the surface of HIV-infected CD4⁺ T cells in ART-treated individuals, and the number of cells expressing these molecules is tightly correlated with the size of the T-cell viral reservoir (68 – 70). Additionally, coinhibitory molecules such as programmed cell death 1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) function as biomarkers for low-level HIV/SIV viral replication, and in vitro blockade of these molecules restores CD4⁺ T-cell function (71, 72). HIV-infected individuals receiving checkpoint inhibitors as immunotherapy for comorbid cancer have increased immune function and stable to decreased cell-associated HIV DNA (69, 73 – 77). These reports suggest that T-cell checkpoints hold promise for the treatment of chronic infections (78, 79), including use in cure strategies for HIV reservoirs (Fig. 2).

FIG 2.

Summary of immune checkpoint receptor interactions and gene editing strategies. (1) Coinhibitory checkpoint molecules such as PD-1, CTLA-4, TIGIT, Tim-3, LAG3, BTLA-4, CD160, and 2B4 are upregulated on T cells during HIV infection and negatively regulate T-cell proliferation and effector function when bound to their cognate ligands. In contrast, the costimulatory molecules CD28, CD226, and GITR enhance T-cell activation when ligated in conjunction with TCR-mediated stimulation. Checkpoint inhibitors that block binding of the indicated molecule to its target to enhance immune function are indicated in blue boxes, while the agonist TRX518, which mimics the interaction of GITR with its ligand, is indicated in yellow. Other potential HIV cure strategies include (2a) editing of the host CCR5 gene to generate resistance to HIV infection or (2b) deletion or indel-mediated transcriptional silencing of the HIV provirus by the CRISPR/Cas9 system, (3a) posttranscriptional silencing of HIV gene expression by antisense oligonucleotides complementary to viral mRNA or novel compounds, and (3b) inhibition of Tat-mediated LTR transactivation by small-molecule inhibitors identified by high-throughput screening.

(i) PD-1 and PD-L1.

PD-1 is an immune checkpoint molecule that induces tolerance and modulates antigen-specific immune responses to infection along with its two ligands, PD-L1 and PD-L2 (80). Upregulation of PD-1 on T cells correlates with HIV disease progression, and anti-PD-1 restores T-cell effector responses against the virus (73). In an SIV model, anti-PD-1 with ART augmented antiviral T-cell responses and decreased the CD4⁺ T-cell viral reservoir (81). In vivo, exposure to anti-PD-1 decreased the number of latently infected cells, and in another HIV-positive individual, anti-PD-1 therapy reversed HIV latency (82).

Follicular helper T (Tfh) (CXCR5⁺ PD-1⁺) cells are important in providing support to maturing germinal center B cells in lymph node (LN) tissue reservoirs. These cells exhibit high PD-1 surface expression and are a major source of replication-competent latent HIV (83). Blockade of PD-1 in SIV-infected macaques enhanced anti-HIV B-cell responses, suggesting some restoration of Tfh function (84). In aviremic individuals on long-term combination ART, PD-1⁺ Tfh cells accounted for 46% of inducible replication-competent virus and 96% of infectious virus arising from the memory CD4 T-cell reservoir, suggesting a potential target for cure therapy (68).

Pembrolizumab was the first anti-PD-1 agent approved by the FDA, in 2014, as a humanized IgG4κ monoclonal antibody (80). Since then, another anti-PD-1 IgG4 human monoclonal antibody and three anti-PD-L1 monoclonal antibodies have been approved (Table 2).

TABLE 2.

Brief survey of on-market or in-development checkpoint inhibitors with HIV-targeting potential

Stage	Checkpoint	Inhibitor(s)	Trial(s) (phase)
On market	PD-1	Pembrolizumab (Keytruda), nivolumab (Opdivo)
	PD-L1	Atezolizumab (Tecentriq), avelumab (Bavencio), durvalumab (Imfinzi)
	CTLA-4	Ipilimumab, MDX-010 (Yervoy), tremelimumab (orphan designation)
In trials	TIGIT	BMS-986207	NCT02913313 (1/2)
		OMP-313M32	NCT03119428 (1)
		MTIG7192A	NCT02794571 (1)
	Tim-3	TSR-022	NCT02817633 (1)
	LAG3	Relatlimab (previously BMS-986016)	11 clinical trials (1, 1/2, 2/3)
	GITR	TRX518	NCT01239134 (1), NCT02628574 (1)
Preclinical	Anti-CD160	ELB021 (Elsalys Biotech)
	Anti-BTLA	40E4 chimericIgG1 (Merck & Co.)

The ligand to PD-1, PD-L1, decreases the proliferation of antigen-specific effector T cells. It also has affinity to B7.1 and uses this pathway to inhibit T-cell responses (85). An anti-PD-L1 monoclonal antibody, BMS-936559, was tested in a phase 1, randomized, placebo-controlled, dose-escalating trial in HIV-infected participants on antiretroviral therapy (NCT02028403). The trial was halted prematurely after enrolling one low-dose arm (0.3 mg/kg), due to the development of retinal toxicity in macaques with higher doses of BMS-936559 (86). However, HIV Gag-specific CD8⁺ T cells increased from baseline through the 28-day trial in two of the six patients who received active drug (87), suggesting enhancement of antiviral cytotoxic T-lymphocyte (CTL) responses.

Current clinical trials are focusing on PD-1 due to the presumption of a lower risk of adverse events (AE) for this target compared to PD-L1. There are three actively recruiting trials using pembrolizumab in HIV-positive individuals without cancer; we are investigators on two of these trials. The first (NCT03367754) is a double-blind, placebo-controlled phase 1 trial targeting individuals with HIV-1 infection on ART but low CD4⁺ T-cell counts (<350 cells/mm³) (Table 3). The second (NCT03239899) is an open-label, phase 1 trial targeting the HIV CNS reservoir using pembrolizumab. The third (NCT03787095) is an AIDS Clinical Trials Group (ACTG) placebo-controlled phase 1/2 dose escalation trial targeting individuals with HIV on ART who have normal CD4⁺ T-cell counts, assessing HIV-1 Gag-specific CD8⁺ T-cell responses (87).

TABLE 3.

Clinical trials using checkpoint inhibitors in HIV-infected populations

Trial	Study drug	Target(s)	Population	Phase
NCT03787095	Cemiplimab	PD1	Suppressed HIV on ART	1/2
NCT03239899	Pembrolizumab	PD-1	CNS HIV reservoir	1
NCT03367754	Pembrolizumab	PD-1	HIV with low CD4+ cell count	1
NCT02595866	Pembrolizumab	PD-1	HIV and malignant neoplasms	1
NCT03304093	Nivolumab	PD-1	HIV and non-small-cell lung cancer	2
NCT02408861	Nivolumab and ipilimumab	PD-1, CTLA-4	HIV and malignant neoplasms	1
NCT03316274	Nivolumab	PD-1	HIV and Kaposi sarcoma	1
NCT03407105	Ipilimumab	CTLA-4	HIV	1
NCT03094286	Durvalumab	PD-L1	HIV and solid tumors	2

(ii) CTLA-4.

Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is a homolog of the costimulatory molecule CD28 that negatively regulates T-cell activation when bound to its receptor B7. Since CTLA-4 and CD28 directly compete for binding to B7, the relative expression of these molecules determines whether a T cell will undergo activation or anergy. CTLA-4 expression on the cell surface is upregulated by T-cell receptor (TCR) and CD28 signaling and attenuates T-cell responses during the resolution phase of an immune response (88). CTLA-4 expression is moderately increased in HIV-specific CD4⁺ T cells from patients and is associated with higher viral loads and inversely correlated with CD4 counts (89).

In a case report of an HIV-infected patient receiving the anti-CTLA-4 therapy ipilimumab (MDX-010) for cancer, there was a decrease in plasma HIV RNA following each infusion and a slight increase in CD4 T-cell count and T-cell activation markers (90). A phase 1, open-label, dose escalation study of MDX-010 as immunotherapy in HIV-infected individuals (NCT03407105) was conducted in 2006. The results remain unpublished but have been cited as not having safety concerns (91). Combination and individual immunotherapy trials targeting CTLA-4 and PD-1 are in progress for HIV-1 related malignancies (Table 3).

In ART-treated, SIV-infected macaques CTLA-4⁺ PD1⁺ Tfh cells were more frequent than PD1⁺ CTLA4⁻ Tfh cells within LNs (70). Additionally, there was a population of CTLA-4⁺ PD1⁻ latently infected, replication-competent cells in the LN and splenic T-cell zone that persisted after ART, which was confirmed in HIV-infected individuals on ART, supporting that multiple checkpoint targets are necessary in cure strategies.

(iii) TIGIT, LAG3, and Tim-3.

Additional checkpoints such as TIGIT (T-cell immunoreceptor with Ig and ITIM domains), LAG3 (lymphocyte-activation gene 3), and Tim-3 (T-cell immunoglobulin and mucin domain containing-3) have been identified as targets of T-cell function enhancement in antiviral targeting strategies.

Tim-3 is a transmembrane protein that binds to galectin-9 and phosphatidylserine (92 – 94). It is a marker of T-cell exhaustion that is upregulated on the CD8⁺ T cells of HIV-infected individuals, correlating with HIV-1 viral load and disease progression (95, 96). CD8⁺ T cells expressing Tim-3 have impaired cytotoxic function, cytokine production, and proliferative capability (97). Blockade of Tim-3 using a recombinant Tim-3 glycoprotein or an anti-Tim-3 monoclonal antibody restores the proliferation of specific CD8⁺ T cells in response to HIV antigens (97). In HIV infection, Tim-3 expression is induced on plasmacytoid dendritic cells, resulting in reduced production of IFN-α (98). Interestingly, treating activated CD4⁺ T cells with a soluble form of the ligand for Tim-3, galectin-9, conferred resistance to HIV infection due to decreased expression of HIV coreceptors (CCR5 and CXCR4) and upregulation of the host restriction factor p21 (99). Conversely, in resting CD4⁺ T cells, which do not express high levels of Tim-3, galectin-9 enhanced HIV infection by binding to protein disulfide isomerase to alter the cell surface redox state and promote viral entry (99, 100). Thus, Tim-3 ligands must be evaluated cautiously as therapies for HIV infection.

Coexpression of PD-1, TIGIT, and LAG3 was associated with an increased frequency of CD4⁺ T cells harboring integrated HIV DNA (69), suggesting a clonal expansion of the cells. TIGIT suppresses T-cell activation by interaction with its ligand, poliovirus receptor (PVR), on dendritic cells in a manner analogous to that for CTLA-4, PD-1, and B- and T-lymphocyte attenuator (BTLA) (101).

LAG3 is a major histocompatibility complex (MHC) class II ligand, related to the CD4 superfamily, that is a coinhibitory stimulus (102). LAG3 expression on HIV-specific CD8⁺ T cells is negatively correlated with HIV RNA (102). LAG3 suppression increases T-cell proliferation and effector cytokine production. Expression of LAG3 in conjunction with Tim-3 on CD4⁺ CD25⁺ cells suppresses proinflammatory macrophage activation (103, 104). The exact mechanism of action of LAG3, like that of many of the T-cell cosignals, is likely to be influenced by the expression of other cell surface molecules. Tim-3-, TIGIT-, and LAG3-targeting immunotherapies are being evaluated in clinical trials (Table 2), though not with antiviral intent.

(iv) CD160, CD244/2B4, BTLA, and GITR.

Other targets for antibody-based therapies include the T-cell coinhibitory molecule CD160. Preclinical studies are under way in the European Union. Epstein-Barr virus (EBV)-, cytomegalovirus (CMV)- and influenza virus-specific CD8⁺ T cells expressing CD160 have reduced cytotoxic activity in response to viral antigens (105). Regulation of T-cell function by CD160 is independent of PD-1 expression, highlighting an alternate pathway to increase T-cell proliferation and CD8⁺ T-cell-mediated cell killing (Table 2) (105).

2B4 (CD244), a member of the CD2/SLAM signaling family, is an immune checkpoint molecule that is a marker of disease progression and exhaustion in HIV infection (106, 107). Abrogating 2B4 binding to its ligand CD48 using an anti-CD48 antibody increased proliferation of HIV-specific CD8⁺ T cells in response to HIV antigens, and blockade of both PD-1 and 2B4 further increased proliferation (107). However, the downstream signaling effects can be unpredictable and are dependent on cell type, ligand density, and availability of downstream signaling molecules (108).

B- and T-lymphocyte attenuator (BTLA) (CD272) is member of the immunoglobulin superfamily that negatively regulates T-cell proliferation and effector function. BTLA expression is downregulated during HIV infection, which correlates with disease progression (109). An anti-BTLA blocking antibody increased proliferation of CD8⁺ T cells from HIV-infected patients in response to HIV antigens, which was further enhanced by blockade of both PD-1 and BTLA (110). Similar effects were observed with dual blockade of PD-1 and Tim-3, and triple treatment with blocking antibodies against PD-1, BTLA, and CD160 increased production of proinflammatory cytokines compared to that with anti-PD-1 alone (110). A BTLA-targeting monoclonal IgG1 is being developed by Merck to supplement PD-1 therapies (Table 2) (111).

Costimulatory targets, such as glucocorticoid-induced tumor necrosis factor receptor (GITR), enhance CD8⁺ T-cell response and resolve chronic lymphocytic choriomeningitis virus (LCMV) infection in mice, signaling a potential role for immunotherapy in HIV disease (112). TRX518 is an agonist monoclonal antibody that mimics interaction of GITR with its ligand to activate antigen-specific T cells and suppress regulatory T-cell responses. Two phase 1 clinical studies with TRX518 in solid tumors (NCT01239134 and NCT02628574) are under way (Table 2). Other coinhibitory and costimulatory checkpoints have been identified as molecular signatures of immune exhaustion in HIV infection and are potential targets for immunological therapies (113, 114).

Gene editing.

Gene editing is a rapidly evolving field, and significant progress has been made in optimizing delivery and efficacy of these tools. However, the possibility of off-target effects due to binding at homologous sequences remains a serious concern, and there are continued challenges in increasing efficiency of mutation/excision at the targeted site and ensuring delivery to all relevant anatomical compartments. We include here a summary of current gene editing approaches that have been applied to HIV cure strategies.

(i) Genome cutting/excision.

Several genome engineering approaches have been used to (i) excise all or part of the provirus from the human genome, (ii) introduce mutations to render HIV incapable of replication, or (iii) modify the structure of HIV coreceptors on immune cells to make them resistant to infection. These approaches include zinc finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), and the clustered regularly interspaced short palindromic repeat (CRISPR) and CRISPR-associated (Cas) system. ZFNs and TALENS are artificial restriction enzymes composed of an endonuclease domain fused to a DNA binding domain. The DNA binding specificity of ZFNs is determined by a series of zinc finger repeats, each of which recognizes a specific 3-bp DNA sequence. TALENs utilize 18 repeats of a 34-amino-acid TAL effector DNA binding domain (TALE), which dictates the nucleotide it recognizes (115, 116).

The discovery of CRISPR/Cas genes has reinvigorated the field of gene editing (117, 118). This approach utilizes a short guide RNA (gRNA) containing a 20-nucleotide sequence in the 5′ end that is complementary to the target DNA and a conserved region in the 3′ end that recruits Cas9 endonuclease, which cleaves target genomic DNA. Unlike ZFNs and TALENs, the CRISPR system does not require a unique enzyme for each nucleic acid target; rather, a reusable Cas9 enzyme is paired with a variety of gRNAs that bind to different regions of the target DNA. All three nucleases silence expression primarily through insertions or deletions (indels) that are introduced during nonhomologous end joining (NHEJ)-mediated repair of double-strand DNA breaks at the target site.

It is important to consider the relative advantages and technical challenges of each of the approaches outlined above when applying these technologies to HIV therapy. For example, the small size of the genes encoding ZFNs (∼1 kb) is an advantage in terms of effective gene delivery compared to TALENs (∼3 kb) or Cas9 (∼4 kb), but ZFNs are more prone to off-target cleavage than both TALENs and CRISPR/Cas9 due to binding to genomic sites that are similar to the target sequence. TALENs offer more predictable DNA binding than ZFNs and are substantially more cost-effective to generate because they do not require extensive screening to ensure optimal binding to the target without significant off-target toxicity. However, the presence of repetitive sequences in TALENS make them unsuitable for lentiviral vectors, and constructs encoding TALENs are generally too large for incorporation into adeno-associated virus (AAV) vectors. The gRNA-guided Cas9 system, in contrast, is compatible with AAV-mediated delivery and provides significant advantages over other methods in terms of target site selection and the possibility of multiplexing gRNAs to target a single Cas9 nuclease to multiple sites. Newer versions of Cas protein are now being developed that are smaller and have multiple other advantages over Cas9.

(ii) CCR5.

The most pursued genome manipulation strategy in HIV research involves disruption of the coreceptor C-C chemokine receptor type 5 (CCR5). The CCR5Δ32 allele correlates with resistance to HIV infection, decreased viral loads, and long-term survival. The “Berlin patient,” who received cells from a donor with a homozygous CCR5Δ32 for treatment of acute myeloid leukemia, has remained HIV free for almost a decade and achieved a sterilizing cure (9, 119). Similarly, the “London patient,” who received a hematopoietic stem cell transplant (HSCT) from a CCR5Δ32 donor as part of treatment for Hodgkin’s lymphoma, has remained in remission after more than 18 months (17). Due to the low number of homozygous CCR5Δ32 donors in the population, HSCT is not a viable option for most HIV-infected patients. However, gene editing strategies have targeted CCR5. It is necessary for both alleles of the CCR5 gene to be disrupted to generate resistance to HIV infection, though patients heterozygous for the CCR5Δ32 mutation exhibit slower disease progression (120). In a phase 1/2 study in which HIV-infected patients received autologous CD4⁺ T cells transduced with a CCR5-targeting ZFN adenoviral construct, only 10 to 27% of the reinfused cells harbored the modified ccr5 gene, which was insufficient to control viral rebound during treatment interruption (121). SB-728-T and its derivatives are adenoviral CCR5-targeted ZFN constructs that are in clinical studies focused on CCR5 editing in CD4 T cells (NCT02388594, phase 1/2; NCT02225665, phase 1/2) or hematopoietic stem cells (NCT02500849, phase 1) (121). Additionally, a 5-patient study (NCT03164135) is investigating the safety and feasibility of allogeneic transplantation of CRISPR/Cas9-modified CCR5Δ32 CD34⁺ cells into HIV-positive individuals with hematological malignancies.

Effective delivery of gene editing machinery and homozygous gene disruption are the challenges of a CCR5 gene editing approach, with only a small number of cells delivered in the absence of bone marrow suppression. Even if these challenges are overcome, this strategy does not take CXCR4-tropic (X4) HIV quasispecies into account; while the vast majority of transmitted founder (TF) viruses that establish infection utilize CCR5 for entry and CCR5-tropic (R5) HIV dominates early stages of the disease (122), the virus may evolve to utilize CXCR4 (X4) in response to selective pressures such as availability of specific target cells (123, 124). This coreceptor switching (R5 to X4) occurs in approximately 50% of individuals chronically infected with HIV (124). It is important to note that in both the Berlin and London cases, CXCR4-tropic virus was not identified as a component of the patient’s reservoir, which was likely a crucial factor in achieving HIV remission. In contrast, the “Essen patient,” who received HSCT from a CCR5Δ32 donor following treatment for T-cell lymphoma, experienced rapid rebound of CXCR4-tropic virus that was present before transplantation (14). Additionally, HIV can be transmitted by cell-to-cell contact in the absence of CD4 and CCR5, as is the case for infection of astrocytes in the brain by HIV-infected lymphocytes (125). It is therefore unlikely that CCR5 editing will be an effective strategy to eliminate HIV in chronically infected individuals with a stable reservoir unless it is paired with a highly efficient ablative therapy and/or is employed only after thorough screening to ensure the absence of CXCR4-utilizing HIV isolates. Alternatively, reactivation of the viral reservoir using LRAs in combination with transplantation of HIV-resistant immune cells could eventually lead to HIV eradication; however, the presence of reactivation-resistant proviruses and inefficient killing observed in “kick-and-kill” approaches to date suggest that further optimization may be required. In terms of preventative therapy, an ethical debate surrounds the possibility of using gene editing techniques at the time of fertilization to create humans with mutations in CCR5 who might be resistant to HIV infection. CCR5-targeting strategies hold promise for a subset of patients with CCR5-restricted reservoirs but are unlikely to be broadly applicable to the majority of people living with HIV.

(iii) HIV inactivation by indel mutation or excision.

One advantage of using gene editing strategies to remove integrated HIV proviruses is that the long terminal repeat (LTR) is present at both ends of the viral genome. Therefore, a single gRNA can target both LTRs, as shown in in vitro studies using CRISPR/Cas9 (126 – 134). This excises the full integrated provirus between the target sequences, though at a much lower frequency than silencing via insertions or deletions (indels) introduced by NHEJ (126, 127, 129). Employment of two different LTR-targeting gRNAs that recognize unique sequences in the HIV promoter significantly enhances Cas9 cleavage (128, 130, 132). There is a risk of viral escape with a single gRNA due to nonlethal NHEJ-induced mutations around the Cas9 cleavage site (130, 131). Sustained inhibition of viral replication requires dual or multiplex gRNAs targeted against essential conserved regions of the HIV genome (130).

The CRISPR/Cas9 system can also cleave HIV sequences within latently infected cell lines, suggesting that Cas9 may be effective even when target sequences are partially occluded by repressive chromatin remodeling (134, 135). Delivery of AAV vectors encoding anti-HIV Cas9 constructs into rodent models of HIV gave similar results, with occasional excision of the full provirus as well as indel formation and reduced susceptibility to new infection (135). Importantly, Cas9-mediated cleavage of HIV occurred in multiple tissues, including brain, heart, kidney, lung, spleen, and liver (135). One study reported complete eradication of the latent HIV reservoir in 2/7 (29%) humanized mice following administration of long-acting, slow-effective ART (LASER ART) and HIV-targeting AAV₉-CRISPR/Cas9; no virus could be detected from lymphoid tissues, bone marrow, blood, or brain, as measured by nested PCR, digital droplet PCR, and RNAscope, and no viral rebound occurred after treatment was removed in these two animals (136). A decrease in HIV DNA was observed in the remaining 5 animals, indicating reduction in the size of blood and tissue reservoirs. In contrast, viral rebound occurred in all animals who received LASER ART or CRISPR/Cas9 treatment alone, highlighting the importance of combination therapy. There are currently no ongoing clinical trials utilizing HIV-targeted CRISPR/Cas9 constructs.

CRISPR/Cas9 technology is an incredibly valuable tool with broad applications that range from HIV research to onco-therapeutics and treatment of genetic disease. Since current methods have been insufficient to completely clear virus in animal models (135 – 137), there is a need for ways to optimize delivery of anti-HIV Cas9 constructs to all infected cells, as well as strategies to enhance the frequency of lethal indels and/or provirus excision. There is concern about the potential immunogenicity of Cas9, given its bacterial origin, when the protein is stably expressed (138). To address this, a Tat-inducible Cas9 construct has been devised that is expressed only during active HIV replication to minimize off-target effects (139). The synergy between HIV-targeting CRISPR/Cas9 constructs and slow-release ART to eradicate virus in a subset of animals (136) provides strong proof of concept for this approach. Nevertheless, the efficiency of Cas9-mediated HIV excision is likely still insufficient to remove all integrated provirus from all reservoirs in all subjects at this stage. As this technology continues to improve over time, it may be necessary to develop other therapeutic strategies to promote HIV remission in order to fill the gaps left by current ART (the “block-and-lock” strategy).

Tat antagonists.

The HIV trans-activator of transcription (Tat) is an HIV accessory protein that plays pivotal roles in HIV replication and pathogenesis. Tat is an early viral protein that binds to the transactivation response (TAR) element, an RNA structure that is found in the 5′ region of all HIV mRNA and recruits the positive transcription elongation factor (P-TEFb) to the HIV promoter (140). Tat-mediated transactivation of the HIV LTR enhances HIV transcription and is essential for productive infection (141 – 143). Intracellular Tat levels regulate the cycle of active HIV transcription and entry into latency (144). Furthermore, Tat can be released from infected cells and taken up by uninfected bystanders (145). Tat has also been shown to be directly neurotoxic (146 – 148) and to stimulate production of reactive oxygen species (ROS) (149) and proinflammatory cytokines and chemokines (150 – 155).

In vitro modeling has predicted that the disruption of Tat-TAR binding can generate stable HIV latency (156). Since current antiretroviral therapies do not target the early viral products such as Tat, experimental strategies to develop Tat antagonists are being considered. Our laboratory has developed a cell-based assay for high-throughput screening of compounds for identifying Tat antagonists that block LTR transactivation. Many putative Tat antagonists have been identified by different high-throughput screening techniques, including several candidates that have been used in phase 1/2a clinical trials. Since several of these antagonists have been reviewed extensively elsewhere (157), this review will focus on recently discovered compounds that show significant promise. A summary of known Tat antagonists and their stages of development is included in Table 4.

TABLE 4.

Putative Tat antagonists that significantly reduce HIV transcription^a

Compound(s)	Infection stage(s) tested	Cell type(s)	Screen	Development status	Reference(s)
VRX496 (Lexgenleucel-T)	Acute, chronic	CD4+ T cells, CD34+ hematopoietic stem cells	NA	Phase 2	121, 171, 174 – 178
Rev-TD-anti-TAR	Acute, chronic	CD4+ T cells	NA	Phase 1/2	172, 173
dCA	Acute, chronic	HeLa-CD4, T-cell lines, PBMCs, primary CD4+ T cells	NA	Preclinical	158 – 160, 164
Durhamycin A	Acute	HeLa-CD4, T-cell lines	LTR-reporter	Preclinical	179
WM5	Acute, chronic	HeLa, T-cell lines, PBMCs	SAR	Preclinical	180 – 183
NM13	Acute	T-cell line	SAR	Preclinical	184, 185
HM13N	Acute, chronic, latent	T-cell lines, monocytic cell lines, PBMCs	SAR	Preclinical	186
Temacrazine	Acute, chronic, latent	T-cell line, monocytic cell lines	NA	Preclinical	187
NeoR	Acute, chronic	T-cell line, promonocytic cell line, PBMCs	SAR	Preclinical	188
3-(4-Chlorophenyl)-5-methyl-N-(3-pyridinyl-methyl)pyrazolo[1,5-a]pyrimidin-7-amine (compound 791), N-{[2-(2-hydroxybenzoyl)hydrazine]carbonothiol}-4-biphenylcarboxamide (compound 833), 2-hydroxy-N′-(3-hydroxy-4-methoxybenzylidene)benzohydrazide (compound 892)	Acute, chronic, latent	T-cell lines, CEM-GXR, PBMCs, HeLa HIVrtTAΔMls (189)	SMN2 minigene reporter	Preclinical	168

^aBased on data from reference 157. Abbreviations: NA, not available; SAR, structure-activity relationship; PBMCs, peripheral blood mononuclear cells.

One of these drugs, didehydro-cortistatin A (dCA), specifically binds to Tat in the TAR binding region to inhibit transactivation of the HIV LTR (158). dCA substantially delays and reduces rebound in ex vivo cultures of CD4⁺ T cells from HIV-infected individuals even in the presence of strong stimuli such as CD3/CD28 ligation, phytohemagglutinin (PHA), and the latency-reversing agent prostratin (159, 160). dCA also induces rapid repressive chromatin remodeling of the HIV promoter that is maintained even when treatment is removed (159). Similar results have been obtained using dCA against SIV in vitro and ex vivo, which suggests that this drug is appropriate for studies in nonhuman primates (161). One of the major challenges in HIV therapy is residual low-level viremia or transient reactivation events (“blips”) that can be detected only by highly sensitive assays and persist despite ART (6, 8, 162, 163). These findings suggest that it may be possible to completely silence the HIV provirus over time by blocking Tat interaction with TAR; Tat antagonists may induce a perpetual latent state without residual viremia to achieve a functional cure via a “block-and-lock” mechanism (159). Such a strategy is especially attractive for controlling viral replication in the CNS reservoir, where delivery of gene editing therapy may be suboptimal or a more conventional “kick-and-kill” approach may lead to undesirable inflammation and neural injury. Importantly, dCA also inhibits Tat uptake and production of proinflammatory cytokines from Tat-transfected astrocytes (164). However, dCA does not bind to the cysteine-rich region of Tat, which contributes to Tat-mediated neurotoxicity through association with the N-methyl-d-aspartate (NMDA) receptor (165); therefore, it is currently unknown whether dCA can directly block Tat-mediated neurotoxicity. In vitro evidence indicates that treatment with dCA can induce mutations in the HIV promoter that increase basal transcription and reduce dependence on Tat transactivation via selective pressure; however, these mutated proviruses are less likely to enter latency and are sensitive to virus-induced cell death and immune-mediated clearance (166).

Given that Tat has a variety of functions and interacts with many cellular proteins (167), identifying compounds that selectively block Tat-TAR or Tat-PTEFb interactions may be insufficient unless the inhibitor completely shuts down proviral transcription from all reservoirs. To block other Tat functions that contribute to HIV pathogenesis, it may be necessary to identify therapeutics that directly degrade Tat protein or target Tat production at the posttranscriptional level. Recently three RNA splicing modulators, 3-(4-chlorophenyl)-5-methyl-N-(3-pyridinyl-methyl)pyrazolo[1,5-a]pyrimidin-7-amine (compound 791), N-{[2-(2-hydroxybenzoyl)hydrazine]carbonothiol}-4-biphenylcarboxamide (compound 833), and 2-hydroxy-N′-(3-hydroxy-4-methoxybenzylidene)benzohydrazide (compound 892), which reduced HIV viral output by 80 to 90% in low-micromolar ranges in vitro were identified (168). All three compounds diminished levels of HIV Gag, Env, Rev, and Tat proteins and decreased relative amounts of unspliced and singly spliced HIV transcripts. Compound 833 and compound 892 had limited off-target effects on exon inclusion of some genomic transcripts and global protein synthesis, while compound 791 significantly decreased viral replication and Tat protein production without notable off-target effects (168).

Another strategy is the use of molecules that are antisense to Tat. Several groups have used antisense oligonucleotides (ASOs) complementary to HIV sequences to either block translation or degrade viral transcripts. Recently, a 937-base antisense gene targeting HIV env under the control of an HIV promoter was incorporated into a lentiviral vector (VRX496) and transduced into CD34⁺ hematopoietic stem cells and autologous ex vivo-expanded CD4⁺ T cells. These modified T cells stably expressed the antisense gene and have been safely delivered to 65 HIV-infected subjects at six institutions to date without significant adverse events (169, 170). One study on a subset of patients who underwent ART treatment interruption reported a decrease in viral load set points in 6/8 (75%) patients who received VRX496, as well as enrichment of A→G mutations in HIV sequences at the antisense target region (171). Since tat and env have overlapping mRNA sequences, VRX496 may also inhibit Tat production. VRX496 is currently in phase 2 clinical trials (NCT00131560, NCT00295477, and NCT00622232) (Table 4).

A retroviral Rev-TD-anti-TAR construct, which expresses both a dominant-negative Rev construct and antisense TAR RNA (172), was evaluated in a phase 1/2 pilot study, but the results have not been published (NCT00001535) (Table 4) (173). Since all HIV transcripts contain the TAR sequence and Tat-TAR interaction is required for transactivation of the HIV promoter, it is likely that Rev-TD-anti-TAR would also act on Tat function and generation of tat mRNA. However, no ASOs that specifically target Tat transcripts alone have entered clinical trials.

It is worth noting that Tat antagonists, like many other potential eradication/remission strategies, are best used as combination therapy to accomplish their goal. While the fact that Tat inhibition alone in the presence of combination ART can induce additional repressive epigenetic changes to the HIV LTR and delay/prevent viral rebound in murine models (159) lends proof of concept to this strategy, further work must be done in vitro and in vivo to determine the feasibility of generating true remission via a permanently silenced (“locked”) HIV promoter.

CONCLUSION

After decades of intense research, HIV remains a chronic infection, and the number of individuals living with HIV continues to rise every year. Recent advances in gene therapy and immune therapies have provided renewed hope that a sterilizing HIV cure might be possible. Some preclinical studies for refinement of these approaches to ensure delivery to all the reservoirs and prevention of off-target effects are still necessary. Identification of pharmacological approaches that block early viral transcripts provides hope for a functional cure by providing long-term silencing of HIV. Further safety studies and means for delivery to all tissue reservoirs are still needed before clinical trials can be undertaken. A common feature in each of these approaches is the challenge posed by the CNS reservoirs due to the blood-brain barrier and the cell types infected in the brain that are different than the lymphoid reservoirs. Further, the brain cannot sustain prolonged inflammation due to bystander damage which can result in permanent neuronal injury. Nonetheless, development of immune checkpoint molecule inhibitors and other strategies that combat immune exhaustion/dysfunction, especially in combination with novel antiretroviral therapies, complementary immune therapies, vaccine work, gene editing of viral and/or host genomes, and novel anti-Tat molecules, may also provide better long-term health outcomes for HIV-infected individuals and avenues toward the goal of achieving a sterile cure.

SEARCH STRATEGY

We searched PubMed, international HIV meeting abstracts, and www.clinicaltrials.gov for HIV trial data up to and including July 2019. We also used multiple spellings, truncated nomenclatures, and abbreviations as search terms. We reviewed articles published in English resulting from these searches and the most relevant references cited in these articles.