HIV replication and latency in monocytes and macrophages

Abstract

The relevance of monocyte and macrophage reservoirs in virally suppressed people with HIV (vsPWH) has previously been debatable. Macrophages were assumed to have a moderate life span and lack self-renewing potential. However, recent studies have challenged this dogma and now suggest an important role of these cell as long-lived HIV reservoirs. Lentiviruses have a long-documented association with macrophages and abundant evidence exists that macrophages are important target cells for HIV in vivo. A critical understanding of HIV infection, replication, and latency in macrophages is needed in order to determine the appropriate method of measuring and eliminating this cellular reservoir. This review provides a brief discussion of the biology and acute and chronic infection of monocytes and macrophages, with a more substantial focus on replication, latency and measurement of the reservoir in cells of myeloid origin.

1. Introduction

Since the emergence of AIDS, it was demonstrated that both CD4 + T cells and macrophages were susceptible to HIV infection and were involved in disease progression. Early studies classified the virus as macrophage tropic, as it was able to induce syncytia in culture [1]. HIV isolates that could infect healthy donor CD4 + T cells in vitro were also able to infect tissue macrophages in several organs, including brain and lung [2,3]. Indeed, some of the direct manifestations of viral pathogenesis were directly associated with infection of tissue macrophages, leading to neurological and lung symptoms. Combined ART was broadly available to patients in 1996, and it has transformed the HIV infection from a deadly disease to a chronic but treatable illness. ART was first predicted to eliminate HIV infection, but long-term studies of virally suppressed people with HIV (vsPWH) demonstrated that, while ART dramatically reduced HIV replication to low or undetectable levels in blood, and prevented spread to uninfected cells, it does not eliminate the virus. Despite suppression of virus replication by ART, HIV DNA remains in long-lived CD4 + T cells as a latent viral reservoir with an estimated half-life of over 70 years [4,5]. ART has brought the focus of HIV research to understanding viral latency and viral reservoirs in CD4 + T cells and, more recently, in circulating monocytes and tissue macrophages [6]. During this time, there has been an increased understanding of myeloid biology. In this review, we will briefly discuss the biology of acute and chronic infection in monocytes and macrophages, with a more substantial focus on replication, latency, and measuring the reservoir in monocytes and macrophages. The primary focus will be HIV, but we will also discuss SIV models of HIV when necessary, as access to macrophages in tissues in vsPWH is often prohibitive.

2. Monocyte and macrophage ontogeny

In the last decade, there has been an increased understanding of the biology and function of myeloid cells [7]. Monocytes were previously thought to travel a linear path, from bone marrow to blood to tissue where they differentiate into macrophages. These cells were characterized as having a single phenotype described as mononuclear blood cells, with an intermediate size between lymphocytes and granulocytes, and a kidney-shaped nucleus. With the advance of cytometry, circulating monocytes in humans and other mammals have been divided in three transcriptome defined distinct subsets, identified by the expression of CD14 and CD16 on their cell surface. In humans and non-human primates (NHP), classical monocytes express CD14⁺ CD16⁻, intermediate monocytes express CD14⁺ CD16⁺, and non-classical monocytes express CD14^lo/− CD16⁺ [8,9]. Classical monocytes emerge in the blood during the post-natal stage and are derived from granulocyte-macrophage progenitors (GMP) in the bone marrow [10]. Bone marrow monocytes express increased CCR2 as the CCL2 gradient in the blood drives them into circulation [11,12]. The maturation stage from GMP proliferation to classical monocyte in the bone marrow occurs over a 36 hour period, followed by one day of circulation in blood [13]. From there, the majority of classical monocytes exit the blood or undergo apoptosis. A small percentage of monocytes transition to intermediate monocytes that circulate for an average of 4.3 days before transitioning to non-classical monocytes which circulate for an average of 7.4 days before leaving the blood or undergoing cell death [13].

Blood monocytes were considered the sole source and regular replenishers of tissue macrophages. However, recent studies using mice have demonstrated that tissue-resident macrophages originate from yolk sac progenitors during embryogenesis, and that circulating monocytes migrate to tissues and become macrophages mostly in response to inflammatory signals [14]. Embryonically derived macrophage, including Kupffer cells, microglia, Langerhans cells, lung alveolar macrophages, and splenic red pulp macrophages, seed tissues as progenitor cells during embryonic development and are long-lived and capable of self-renewal [15]. Each tissue has its own complex microenvironment that dictates the maturation and function of the macrophage. Thus, tissue resident macrophages have different phenotypes, transcriptional profiles, and use different growth factors and signaling pathways to differentiate and maintain their populations [15]. The macrophage landscape is complicated by inflammation, as blood monocytes enter tissues and differentiate into inflammatory macrophages with limited survival, with the exception that specific conditions may permit their reprogramming towards tissue-resident macrophages [16].

In vitro, macrophages can be polarized into distinct phenotypes depending on the culture media in which they are cultivated. Initially these polarized states were divided into pro-inflammatory M1 and anti-inflammatory M2. However, now it is known that macrophages present high plasticity and vary considerably according to the stimuli to which they are exposed, hindering the attempts to determine the polarized state of these cells in vivo. Overall, monocytes and macrophages are highly dynamic innate immune cells that offer significant challenges when studied ex vivo.

3. Acute and chronic infection of monocytes and macrophages

AIDS etiologic agent was named LAV and HTLV-III before being officially classified as a lentivirus and being named HIV. All lentiviruses that had previously been identified were known to infect myeloid cells and involve lymphocytes in the pathogenesis of the disease [17–20]. Early studies in the epidemic demonstrated that macrophages were infected by HIV and capable of supporting viral replication. These findings also provided evidence that macrophages are infected for long periods without succumbing to the cytotoxic effects of the virus [2,21]. Macrophages, similar to CD4 + T cells, express both the entry receptor CD4 and co-receptors CCR5 and CXCR4, which bind the envelope protein to allow entry and infection [22]. Initial infection with HIV is thought to occur by a single or select few transmitted founder viruses at the genital or rectal mucosal surfaces. Biological analysis of mathematically modeled founder sequences suggests that founder viruses are typically unable to replicate in macrophages [23,24] possibly due to the lower densities of CD4 on the macrophage surface [25]. However, this susceptibility varies according to the clades. Clade D viruses are known to replicate efficiently in macrophages [26]. Additionally, these experiments were performed in monocyte-derived macrophages (MDM), which do not represent the population of cells that are thought to be first encountered by the virus in vivo. Complementary studies in human reproductive tract and non-human primates support the idea that CD4s may be the initial cell infected by the virus [26]. This suggests that macrophages may not be an important source of viral replication at the initial site of infection.

Macrophages most likely come into play once infected CD4s have migrated to distal tissues, such as the gut and spleen. During acute infection, virus is disseminated to secondary lymphoid organs, in particular to the gut associated lymphoid tissue (GALT). There is strong evidence that most of the CD4 + T cells in the GALT, including CD4+ memory T cells are directly depleted by HIV propagation, accompanied by the loss of integrity of the intestinal barrier [27,28]. Macrophages of the gut are thought to be relatively resistant to HIV infection [29] and are more likely to orchestrate the primary antibody response against the virus which suppresses plasma viremia at the onset of the chronic phase. Circulating monocytes are also recruited to the intestinal sites of viral replication and inflammation and differentiate into inflammatory macrophages. While MDMs are permissive to HIV infection [30], their role in establishment and spread of HIV is unknown, as their half-life may be shorter than embryonically derived macrophages [31,32]. The establishment of infection in lymphoid tissues, as well as the breakdown of the intestinal barrier, leads to a dramatic increase in plasma viremia [33], likely leading to the infection of perivascular macrophages. In contrast to other tissue resident macrophages, perivascular macrophages are highly mobile and infiltrate other organs such as the lung and brain. Additionally, they have a life span of up to three months and are resistant to HIV induced cytotoxic effects.

Additional evidence points to macrophages as likely targets during acute infection, as viral DNA has been detected in tissues macrophages from multiple organs including Kupffer cells, microglia, alveolar macrophages, and intestinal macrophages. Importantly, replication-competent virus can be recovered from cultures of macrophages purified from lymphoid tissues of acutely treated macaques implying that productive infection is taking place during acute infection [34–36]. However, it remains unclear how HIV is disseminated to establish infection in these cells and tissues. Additionally, monocytes, which migrate from blood to seed multiple tissues, have been found to have HIV DNA during acute infection, and may play a key role in viral dissemination to brain and other tissues [37]. Several SIV NHP models of HIV have shown that virus can be detected in the brain as early at 3–14 days post infection [38–41]. Overall, macrophages likely play two opposing roles in the acute phase of HIV infection. On the one hand, they help to establish infection at sites of viral entry, and monocytes and perivascular macrophages disseminate the virus throughout the body, including the brain. On the other hand, macrophages are critically involved in the initiation and the orchestration of the adaptive cellular and humoral immune response, which help to diminish viral burden, leading to the reduction of viremia at the onset of chronic infection [42].

Macrophages also play a role during the untreated chronic HIV infection. Accumulating data suggest that macrophages are long-lived viral reservoirs and can store HIV particles in internal compartments. The presence of mature HIV in intracellular vesicles of macrophages was demonstrated early in the epidemic [43], and there is some controversy in the field regarding the origin of HIV within macrophages [44,45]. Regardless of the origin, they likely represent a source of virus that is inaccessible to the immune system. Previous studies have shown that infectious HIV in macrophages is protected from neutralizing antibodies, small molecules [46], and can survive in macrophage compartments for multiple weeks [47]. Similarly, the role of macrophages in HIV infection remains controversial, with some macaque studies inferring that macrophages engulf SIV infected CD4 + T cells but are not infected [48,49]. However, it is clear that macrophages contribute to untreated chronic HIV infection as HIV viral sequences within individuals become increasingly macrophage-tropic with disease progression [50]; by late stage of infection, CD4 + T cells are depleted and infected macrophages are the principle drivers of viremia [51,52]. Experimental depletion of CD4+ cells in NHP models of HIV pathogenesis have provided important insight into disease progression and the role of cells other than CD4 + T cells in disease progression. A study where CD4 + T cells were depleted prior to infection with SIVmac251 showed that similar peak viremia was observed, compared to undepleted control animals. However, CD4 depleted animals did not have a post peak decline as seen in control animals [53]. A follow up study showed that the post peak viral load was 2 logs higher compared to control animals. In addition, an expansion of pro-inflammatory monocytes was observed, as well as increased activation and infection of macrophages and microglia, which appeared to be the predominant population of productively infected cells [54]. Finally, the analysis of viral decline in plasma post ART intervention suggested that the lifespan of infected cells in the CD4 depleted animals was significantly longer than in control animals [54]. Overall, monocytes and macrophages present many barriers to productive infection, and conditions within these cells are not considered favorable for viral entry and initial replication. However, the studies presented here suggest once the virus surmounts these hurdles and integrates into the host genome it is likely to persist long-term.

4. Viral replication in macrophages

4.1. Entry

Viral entry has distinct properties in macrophages compared to CD4 + T cells. In general, HIV entry into target host cells requires the engagement of the virus surface glycoprotein, gp120, and CD4, which is present on the surface of a variety of cells at different densities, including T cells (high density) and macrophages (low density) [55–57]. The fusion of viral envelope with the cell membrane occurs by the interaction of virus glycoprotein, gp41, with either CCR5 or CXCR4, both of which are expressed by macrophages and CD4 + T cells [45,58–60]. The majority of HIV variants require a high density of surface receptor and co-receptor in order to enter a cell. This characteristic makes it possible to investigate viral tropism by sequencing the envelope alone. Studies in the Swanstrom laboratory, using envelope sequencing and an in vitro system where CD4 expression is controlled, has further defined viral tropism to include 3 primary groups: R5-T cell tropic (requiring high density CD4 and CCR5 for entry), R5-Macrophage tropic (requiring low density CD4 and CCR5 for entry) and X4-T cell tropic (requiring high density CD4 and CXCR4 for entry) [61]. One important caveat for this approach is that CD4 density in tissue macrophages is highly variable, and it is not known whether the spatial location of macrophages within tissues (compared to in cell culture or ex vivo analysis) increases the potential for viral adsorption and entry. Additionally, some studies have suggested that HIV entry can be mediate through multiple routes in macrophages, such as micropinocytosis, receptor-mediated endocytosis or via macropinosomes [62–64]. Therefore, general assumptions have been made about viral entry into macrophages based on cell culture conditions and is an area that needs further investigation in vivo.

4.2. Reverse transcription

Once the virus has entered the cell, HIV reverse transcription occurs. This process also has distinct properties in macrophages compared to CD4 + T cells, as each cell type has a unique cellular environment. In general, reverse transcription requires a basal level of nucleotides (dNTP) to be present to facilitate efficient production of proviral DNA, and without sufficient dNTPs, reverse transcription occurs suboptimally. Concentrations of cellular dNTPs vary significantly depending on the cell cycle status. Activated CD4 + T cells and macrophages differ substantially, as activated CD4s are in proliferative state and macrophages are non-proliferative. This difference in cell cycle status leads to a difference in expression of ribonucleotide reductase R2 subunit, which determines the levels of dNTPs required for viral DNA synthesis mediated by HIV reverse transcriptase (RT) [65]. In addition, the presence of highly expressed host restriction factors in macrophages can affect reverse transcription efficiency or inhibit nuclear transport of the pre-integration complex (PIC) due to a reduction in the dNTP pool available. Sterile α-motif/histidine-aspartate domain-containing protein 1 (SAMHD1), limits HIV replication in non-dividing cells, such as macrophages and resting CD4 + T cells, by depleting intracellular dNTPs [66]. In the presence of SAMHD1, cellular dNTP concentrations are suboptimal for reverse transcription of viral cDNA [67]. This low intracellular dNTP concentration causes substrate binding to RT to be the rate- limiting step in proviral DNA synthesis in macrophages, which harbor approximately 50–100 times lower dNTPs than activated CD4 + T cells [68]. Despite the low levels of dNTPs, macrophages in turn have high cellular concentrations of ribonucleoside triphosphate (rNTPs) which can be incorporated into the growing viral DNA strand in place of dNTPs [69,70]. These events, although normally infrequent, have been shown to occur in HIV infected macrophages but not T cells due to their higher frequency dNTPs in relation rNTPs. These data provide evidence that the unique ability of HIV to infect non-proliferative cells depends at least in part upon the evolutionary adaptation of its reverse transcriptase to function under conditions of limiting dNTP availability.

An additional family of viral restriction factors that are expressed in macrophages and play a role in restricting macrophage RT is Apolipoprotein B Editing Complex (APOBEC), a family of cytidine deaminases that induce G to A hypermutation and lead to the degradation of the HIV genome [71]. Macrophages express APOBEC3G, APOBEC3F and APOBEC3DE and their upregulation is IFNα-dependent. APOBEC3G restricts HIV more potently than the combined effect of APOBEC3F and APOBEC3DE [72,73], and is the dominant form found in CD4 + T cells [74]. APOBEC3A is also expressed at the high levels in monocytes, while fully differentiated macrophages and CD4 + T cells express low levels [71,75]. The difference in APOBEC3A levels between monocytes and macrophages correlates with susceptibility to HIV infection, with monocytes being more resistant to infection than macrophages [75,76]. However, it should be noted that only a small percentage (~6 %) [77] of HIV sequences from infected macrophages contain to G to A mutations compared to a larger percentage in CD4 + T cells (~20 %) [78]. These data suggest that APOBEC restriction of lentiviruses in macrophages likely occurs by other mechanisms [71]. Other viral restriction factors expressed in macrophages that may play a role in restricting macrophage RT and integration include myxovirus-resistance protein 2 (MX2), which inhibits HIV replication post entry by hindering the nuclear accumulation and integration of proviral DNA into host chromatin [79–82], and members of the RNA polymerase II-associated factor 1 (PAF1) family, which are expressed in monocytes/macrophages and repress HIV RT and proviral DNA integration [83,84].

4.3. Pre and post integration

After HIV RNA is converted into dsDNA, the newly synthesized DNA is imported to the nucleus as a pre-integration complex (PIC), comprised of HIV dsDNA, viral proteins: including RT, Vpr, integrase (IN), matrix (MA, p17), capsid protein (CA), and cellular proteins. However, in certain scenarios, viral DNA can exist in an unintegrated form. As with entry and RT, there have also been differences in the accumulation of the unintegrated viral DNA forms noted in macrophages compared to activated CD4 + T cells. In general, HIV infection causes the accumulation of the unintegrated DNA (1-LTR, 2-LTR) in both cell types, but due to a lower pool of dNTPs, specifically ATP, 2-LTR circles may accumulate to a larger degree and last longer in macrophages compared the CD4 + T cells [85]. Unintegrated viral DNA forms in macrophages can persist for up to 2 months [86,87], compared to activated CD4 + T cells where they persist for approximately one day. However they have been shown to survive longer in resting CD4 + T cells [88,89]. The reason why 2-LTR circles are lower in dividing cells vs non-dividing cells (macrophages and resting CD4 + T cells) is not definitively known. It has been possibly attributed to quick degradation during cell division [90]. The stability of 2-LTR circles in non-dividing cells may also be due to their association with host proteins, such as histones, which block the degradation process [91]. Unintegrated viral DNA in vsPWH presents a potential obstacle to effective treatment, as they may act as reservoirs and could integrate once treatment with an integrase inhibitor is discontinued [90].

Typically, integration of proviral DNA and the expression of provirus requires that a target cell is in an activated proliferative state. Distinct from other retroviruses, lentiviruses do not require cell division in order to be transported to the nucleus and integrated into host DNA [92]. HIV and other lentiviruses present a 3-strand DNA region that allows the genome to be transported to the nucleus despite cell cycle status. This structure is called a DNA flap and indicates that lentiviruses evolved to replicate in non-diving cells, such as macrophages. After proviral DNA is integrated into host DNA, HIV DNA can remain dormant within the cell. As with other life cycle steps, there are also differences reported for nuclear translocation and integration. It has been shown that RT and nuclear transport of PIC in macrophages requires an interaction between importin α and Vpr, which is not required for CD4 + T cells [93]. Reverse transcription cannot occur without this interaction, and this is thought to explain a lack of replication in monocytes since importin α is not upregulated until monocytes are differentiated into macrophages. Similar to CD4 + T cells, HIV also preferentially integrates into the transcriptionally active regions in macrophages [94–96].

4.4. Transcription

Integrated HIV DNA is flanked by the long terminal repeats (LTR). The 5′ LTR has binding sites for several transcription factors including specificity protein 1 (Sp1), activator protein 1 (AP1), nuclear factor of activated T-cells (NFAT), nuclear factor-κB (NFκB), c-myc, chicken ovalbumin upstream promoter (COUP), upstream stimulatory factor (USF), CCAAT box transcription factor/nuclear factor 1 (CTF/NF1), T cell factor 1α (TCF-1α) and the glucocorticoid receptor [97]. These transcription factors, in addition to Tat and Vpr, act together to regulate HIV transcription in infected cells. Overall, the regulation of HIV transcription in monocytes and macrophages varies considerably with the stage of cellular differentiation. Unstimulated monocytes and myeloid progenitor cells support low levels of viral replication, and activate transcription in response to cellular stimuli, similar to CD4 + T cells [98–105]. However, monocyte-derived macrophages (MDM) have increased viral replication and do not alter HIV transcription in response to cellular stimulation [100,106]. Additionally, as monocytes differentiate into macrophages, the ratio of Sp1 and Sp3 increases, resulting in increased HIV transcription [107]. This likely results in a low level of HIV transcription in circulating monocytes that evades immune detection until cells enter tissues and differentiate. NFκB regulation of HIV transcription also differs in MDMs compared to monocytes and CD4s. Not only are the components of NFκB different between cell types but expression also differs. In MDMs, NFκB is constitutively active in the nucleus, and its DNA binding activity is not increased further in response to cellular activation or differentiation [106]. This is compared to monocytes and CD4 + T cells, where cellular activation results in increased activation of NFκB and enhanced HIV replication [98,101–103].

NFAT, C/EBP, Jun and AP1 transcription factors also have distinct activities in monocytes and macrophages compared to CD4 + T cells. NFAT binds constitutively to NFκB in monocytes but requires cellular activation to bind in CD4 + T cells [108–112]. Also, NFAT5 has been shown to regulate HIV replication in monocytes and MDMs but not in T cells [111]. C/EBP proteins and their binding sites are critical for HIV replication in macrophages but not in CD4 + T cells [96]. In fact, primary macrophages cannot be infected with HIV that contains mutations in C/EBP binding sites, whereas CD4 + T cell lines can support the replication of these mutants [96]. Henderson and colleagues described the important role of C/EBPβ as regulator of the HIV LTR in macrophages and promonocytic cell lines U937 in vitro [97]. Jun levels have been shown to increase during monocyte maturation and become constitutively expressed in MDMs [113–117]. AP1 in MDMs, despite being expressed, lacks the ability to bind DNA because Ref-1, a protein that modulates the oxidation state of Fos, is not simultaneously expressed in these cells [118,119]. Other transcriptional activators such as mitogen activated protein kinases (MAPK), and Janus kinase/signal transducer and activator of transcription (JAK/STAT) have also been reported to stimulate HIV transcription in macrophages [120].

Additionally, the viral proteins Tat and Vpr have been shown to have distinct properties in monocytes and macrophages in regard to HIV transcription. Tat activity has been shown to be limited in monocytes due to the lack of sufficient levels of cyclin T1, a component of p-TEFb [121]. Differentiation into macrophages increases Cyclin T1 expression and results in strong Tat activity [54]. It appears that HIV replication requires Tat for efficient transcription and virus production in both cell types. In the absence of Tat, only low level transcription of short abortive HIV transcripts is observed [122]. Vpr has been shown to be necessary for viral replication in cells of myeloid lineage but not in T cells [123–128]. Additionally, Vpr is thought to play a specific role in viral mutation rates in monocytes and macrophages [107].

4.5. Assembly and release

HIV assembly in activated CD4 + T cells occurs the majority of the time at the surface membrane [129,130]. However, studies in macrophages demonstrated the presence of HIV virion particles in intracellular compartments, such as multivesicular bodies (MVBs) or late endosome (LE) like structures [131,132]. Proteomic analysis of host cell proteins incorporated into highly purified virions produced by macrophages, revealed the presences of many late endosomal proteins such as MHC II, CD63, and tetraspanins [133], validating results reported in several immune electron microscopy studies [131,133,134]. While the issue of endosomal acidification has been suggested as a reason why HIV cannot assemble in MVB or LE, one group reported that endosomes and lysosomes were correctly acidified but compartments containing viral components were not acidified [135]. These data suggest that HIV may have evolved a strategy for survival in macrophage intracellular compartments [44].

Among the numerous cellular factors that have been reported to be involved in HIV assembly and budding, the ESCRT cellular machinery (Endosomal Sorting Complex Required for Transport) has been shown to play a key role in the formation and release of new particles in CD4 + T cells. On the contrary, this group of proteins have not been shown to definitively play a role in macrophages. Tetherin (BST-2), however, has been shown to play a role in restricting the release of virions in macrophages. Macrophages express high levels of tetherin, whereas CD4 + T cells express none, or only low levels of this restriction factor [136–138]. Additionally, tetherin is considered to be one of the more potent HIV restriction factors identified to date, as it has been shown to inhibit both cell-free or cell-cell viral spread from primary macrophages [139]. Vpu, an accessory protein of HIV, has also been shown to play a role in virion assembly, as it downregulates tetherin from the plasma membrane. In the absence of Vpu, viral particles bud from the plasma membrane of CD4 + T cells but cannot detach. This mechanism is not fully understood in macrophages; one study showed that the presence of Vpu was unable to fully overcome tetherin-mediated restriction of particle release in this cell type [140]. Overall, HIV assembly and release has not been thoroughly studied in macrophages and needs further investigation.

5. HIV latency in monocytes and macrophages

The mechanism of HIV latency has not been completely elucidated in CD4 + T cells or macrophages, but likely results from multiple factors, such as sequestration of cellular transcription factors in the cytoplasm, epigenetic regulation, and/or the action of transcriptional repressors. However, it is likely that mechanisms that establish latency in macrophages differ from those that establish latency in CD4 + T cells. Latency can be classified into two types: pre-integration latency and post-integration latency. The pre-integration stage of latency includes unintegrated forms of HIV DNA such as 2-LTR and 1-LTR circles. As mentioned above, CD4 + T cells can harbor unintegrated HIV DNA, however, because the cells are actively replicating, it is unlikely that unintegrated DNA contributes to long-term persistence in T cells [88,141,142]. Macrophages, conversely, have been shown to contain large amounts of unintegrated HIV DNA which persists for longer periods of time (2 months) and could potentially contribute to viral persistence [86,87]. Some studies have reported detection of unintegrated HIV DNA in macrophages from patients, mostly in the brain [143,144]. This form of latency is heavily regulated by host restriction factors, SAMHD1, APOBEC3, MX2, and PAF1 as described above.

The second stage, post-integration latency, occurs after HIV proviral DNA has been integrated into the host genome, followed by silencing of HIV gene expression. Integrated provirus is considered the major source of HIV latency [145]. As previously mentioned, HIV proviral DNA preferentially integrates into euchromatin regions, where active transcription occurs, in both CD4 + T cells and macrophages [95,96,146,147]. In general, latency is thought to be induced via the recruitment of histone deacetylases (HDAC) to the 5′LTR, which then induce chromatin remodeling and suppress HIV gene expression [148,149]. HDACs have been shown to be recruited to the proviral promoter by Sp1, COUP transcription factor interacting protein 2 (CTIP2), Ying Yang 1 (YY1), C-promoter binding factor-1 (CBF-1) and Late SV40 Factor (LSF) [150–153]. Though there are few studies that have pinpointed specifically how macrophages vs CD4 + T cells induce latency, some studies have assessed the role of specific transcription factors in cells of myeloid origin. Marban and colleagues, described that CTIP2 recruits HDAC1 and HDAC2 to the 5′LTR of HIV DNA in microglial cells [154]. Further, CTIP2 was shown to interact with SUV39H1, a methyl transferase responsible for H3K9 trimethylation, which promotes the recruitment of HP1 protein to the 5′LTR leading to the local heterochromatization and induction of latency [155–157]. The role of CTIP2 was further validated in a study that analyzed the expression of CTIP2 in postmortem brain tissue of HIV infected individuals with HIV encephalitis (HIV DNA and RNA positive), patients with latent HIV (HIV DNA positive and HIV RNA negative) and no detectable HIV (HIV DNA negative) [158]. Higher levels of CTIP2 were solely observed in patients with latent HIV in the CNS, suggesting that CTIP2 may play an important role in the regulation of viral latency in microglial cells. In an additional study, lysine-specific demethylase (LSD1) was shown to regulate HIV gene expression in a synergistic way with CTIP2 in microglia [159]. LSD1 was reported to assist in the recruitment of CTIP2 and other proteins to the HIV promoter which resulted in an increase H3K4 trimethylation, and in turn repressed viral gene expression. This property of LSD1 seems to be highly specific for cells of the myeloid lineage.

In addition to the cellular transcription factors, HIV latency is also influenced by the viral protein Tat. As mentioned previously, in the absence of Tat, only a low level of HIV transcription is observed. Additionally, p-TEFb, a critical cellular cofactor for Tat, is required to produce complete transcripts of HIV. p-TEFb is composed of a catalytic subunit, cyclin-dependent kinase 9 (CDK9), and a regulatory subunit, cyclin T1 (CycT1) that are differentially regulated in monocytes and macrophages compared to CD4 + T cells [85]. Monocytes express very low levels of CycT1, which transiently increases during differentiation into macrophages. In contrast, CDK9 expression levels remain constant in monocytes and macrophages [160]. The low levels of CycT1 in monocytes result in low functional levels of p-TEFb, leading to low HIV transcription and thus viral latency in monocytes [160]. Additionally, tat-mediated LTR activation can be specifically suppressed by OTK18 in monocyte-derived macrophages. OTK18, a cellular transcription factor was shown to be highly expressed in macrophages following HIV infection and diminished virion production [161].

HIV latency can also be regulated on a post-transcriptional level in macrophages and CD4 T cells. MicroRNAs (miRNAs) can regulate host gene expression at a post-transcriptional level and have been reported to influence HIV gene expression [162–166]. miRNA-28; miRNA-150; miRNA-223; and miRNA-382 have been demonstrated to target HIV and play a role in the susceptibility of monocytes and macrophages to HIV. Inhibition of these miRNAs in monocytes results in an increase in HIV replication, while increased expression of these miRNAs in macrophages leads to a decrease of HIV replication [162,167]. These data suggest that miRNAs can play an important role in viral latency in macrophages. A better understanding of the mechanism of latency in macrophages, specifically macrophages in vivo, is essential for the eradication and or therapeutic treatment of the reservoir, as it is likely that macrophages in tissues, blood and brain play a significant role in the HIV latent reservoir.

6. Measuring the monocyte and macrophage HIV reservoir

It is estimated that 5 % of HIV proviruses within CD4 + T cells are intact and, therefore, potentially replication competent. This ratio has yet to be determined in latent monocytes or macrophages. There is substantial evidence that HIV persists in monocytes and macrophages from blood and tissues in PWH. HIV DNA has been detected in highly purified monocytes [168–171]and tissue macrophages isolated from the urethra [172], gut [154], liver [173], bronchial lavage fluid [174,175] and brain [176,177] of vsPWH. The detection of HIV DNA is the most frequently used indicator to determine if HIV persists in an individual, although it is not indicative of a functional latent reservoir and is often an overestimation. One issue with reporting total provirus as measured by qPCR is that it also detects linear unintegrated DNA and episomal forms of HIV DNA, such as 1-LTR and 2-LTR circles. As mentioned above, there is an accumulation of 2-LTR circles in macrophages due to a depleted pool of dNTPs, which can skew the results from this assay. To overcome the issue of unintegrated forms of HIV DNA, assays that specifically target integrated HIV DNA were developed. These assays take advantage of the high frequency (~11 %) of Alu short interspersed elements within the human genome as a host target to guarantee the exclusion of unintegrated forms of HIV [178]. One primer targets Alu elements while the other primer targets gag, followed by a nested PCR for the HIV LTR, enabling the detection of integrated provirus [179,180]. However, these assays are subject to variable efficiency depending on the proximity of a human Alu sequence and the HIV genome, and often requires a high number of replicates for accurate frequency estimation [180,181]. Additionally, this method has been rarely used to measure the reservoir in cells of myeloid origin with only a few publications reporting mixed results [168,182]. It will be difficult to use this method to measure the reservoir in monocytes and macrophages given the limited number of cells that are typically available and low level of HIV DNA in this cell type.

The presence of integrated HIV DNA in CD4 + T cells is not considered an accurate representation of the level of intact and/or reactivatable virus within these cells, since defective proviruses tend to outnumber intact proviruses 20 to 1 [183]. It is possible that the number of defective proviruses will differ in monocytes and macrophages and studies need to consistently assess HIV DNA in these cell populations. One major concern relative to HIV DNA quantification in monocytes or macrophages is the purity of the samples. Of the published studies, HIV DNA is detected at varying frequencies in monocytes, ranging 30–100 % of donors assessed [168–171]. However, not all studies assess contamination of monocyte preparations by CD4 + T cells. This is relevant in the context of older monocyte isolation methods, which were more prone to T cell contamination. Nevertheless, studies that define monocyte purity with stringent criteria have detected HIV DNA in highly purified monocytes from long-term suppressed individuals [6,170,184–187]. Numerous questions remain regarding whether these cells are infected or have phagocytosed virus particles, and whether the cells persist beyond the typical lifespan of a monocyte. These questions combined with the technical challenges in detecting HIV DNA in monocytes require further careful study in order to define the contributions of infected myeloid cells to the reservoir as relevant to the cure agenda.

One avenue to definitively prove whether monocyte and/or macrophages from vsPWH contain functional HIV is to complete studies that assess the replication competence of the integrated provirus within these cells. Few studies have assessed the replication competence of HIV within monocytes and/or tissue macrophages, however, and the bulk of reports have been centered on monocytes as tissues macrophages are often inaccessible. One study suggested that HIV is actively replicating in monocytes during ART suppression and detected cell associated multiply spliced and unspliced HIV RNA after 1–3 years of ART [170]. An additional study reported detecting replication competent HIV from highly purified monocytes isolated from vsPWH that had been on ART for 14–18 months [169]. There have also been limited studies completed in macrophages isolated from tissues. Some reports have described that, though macrophages contain HIV proviruses, they are not a true reservoir because the HIV DNA found in these cells is defective or restricted with regard to propagation [173,188], whereas other studies have been able to reactivate virus isolated from tissues macrophages. Replication competent HIV, for instance, was found in macrophages isolated from urethra tissues [172]. These differences may be attributed to the complex milieu of tissue macrophages both within and across various tissues in the body.

Due to the difficulty of obtaining tissue samples from vsPWH, NHP and humanized mouse models have been primarily used to understand the macrophage reservoir in tissues. DiNapoli and colleagues detected viral DNA from lymphoid tissue in resident macrophages from SIV-infected macaques that were treated with ART for at least 5 months, but they could not detect replication-competent virus from macrophages isolated from these animals [189,190], despite being able to detect replication competent virus from untreated SIV-infected animals [190]. Our group has done extensive work to assess the monocyte/macrophage reservoir in SIV-infected ART suppressed macaques. We have developed a quantitative viral outgrowth assay (QVOA), considered the gold standard in the field, to measure latency in monocytes and tissue macrophages [191]. The QVOA specifically detects replication competent viruses and quantifies the inducible reservoir, represented as infectious units per million cells (IUPM). Our group has shown in SIV-infected pigtailed macaques, receiving ART for 6–20 months, that SIV could be reactivated from MDMs, splenic macrophages, alveolar macrophages, and brain macrophages [35,36]. Additionally, we have shown that the virus produced in these macrophage cultures can infect healthy activated CD4 + T cells and propagate normally. Of particular interest, the IUPM values reported for MDM and blood CD4 + T cells as well as splenic macrophages and CD4 + T cells, were comparable, suggesting similar levels of functional latency in both cell types. Finally, we also showed that SIV DNA, but not RNA was found in all tissues indicating that the animals were fully suppressed at the time of euthanasia [35].

However, this model has received some negative feedback as it consists a highly CD4-tropic virus (SIV/17E-Fr) and a neurotropic SIV swarm (SIV/DeltaB760) that leads to an accelerated SIV disease progression, which is not representative of typical HIV progression [192]. To address these concerns our group moved to studying the more commonly used NHP model, SIVmac251 in rhesus macaques, to assess the monocyte/macrophage reservoir during ART suppression. We reported that the monocyte and macrophage reservoir could be readily detected in SIVmac251 infected rhesus macaques that had been suppressed on ART for 3 months [34]. Though the ART suppression window was shorter compared to other studies, we did report that SIV RNA was undetectable in all tissues at the time of euthanasia. In this study we had detectable IUPMs in brain macrophages, splenic macrophages, lung macrophages and MDM, similar to what was reported in the pig-tailed model. In addition, the viruses released in the macrophage QVOAs, were able to propagate in healthy activated CD4 + T cells from healthy macaques [34]. Currently, we have ongoing studies that will assess effects of long-term ART suppression on the monocyte/macrophage reservoir and assess the size of the monocyte/macrophage reservoir overtime. Together, these data suggest that macrophages isolated from different tissues harbor intact SIV genomes that are replication competent after ex vivo activation. These studies give further evidence that cells of myeloid origin likely contribute to the size of the HIV reservoir in vsPWH.

Other animal models that have proved promising in investigating the role myeloid cells play in HIV latency include the development of three humanized mouse models, bone marrow/liver/thymus (BLT) mice, T cell-only mice (ToM), and myeloid-only mice (MoM). Honeycutt and colleagues (2016) investigated the presence of replication-competent viruses in purified CD4 + T cells and monocytes from non-suppressed HIV-infected individuals using the BLT mouse model. Though the authors were unable to reactivate purified monocytes from viremic individuals to produce virus in the BLT model [193], the authors did not attempt to reactivate monocytes isolated from vsPWH. This is important as monocytes isolated during viremic infection are fragile and unlikely to survive in large numbers post rigorous selection, whereas monocytes isolated during suppressed infection are more durable. Additionally, assessing monocytes from vsPWH would have addressed if monocytes in vsPWH contain latent HIV. However, these authors did definitively show that human macrophages could support HIV replication in vivo using the MoM model. The authors found strong evidence of productive macrophage infection in tissues, specifically in brain macrophages [193]. In an additional study, MoM mice were treated with ART resulting in a significant reduction of plasma viremia [194]. ART interruption resulted in HIV rebound in 1/3 of mice and it was concluded that macrophages, in particular long-lived tissue macrophages, represent HIV reservoirs.

Other techniques have also been reported which may be useful in measuring the monocyte and macrophage reservoir. The Tat/Rev induced limiting dilution assay (TILDA) has been used to quantify the functional latent reservoir by measuring multiply spliced RNA (msRNA) that is induced in response to in vitro activation in CD4 + T cells [195–198]. However, Chateu and colleagues, determined that TILDA is significantly limited by cellular input and may make it difficult to assess HIV reactivation in macrophages as cell numbers tend to be low [199]. An additional new assay is the Intact Proviral DNA Assay (IPDA). Though this assay does not measure the reactivatable reservoirs, it does measure the number of intact proviruses vs defective, by amplifying two distinct and conserved regions in the HIV genome using digital droplet PCR (ddPCR) [200]. The measurement of intact vs total provirus is thought to be more representative of the inducible reservoir in CD4 + T cells. However, this assay has not been used to assess the monocyte/macrophage reservoir. Therefore, further studies to determine its usefulness in cells of myeloid origin are needed. Additionally, the application of in situ hybridization techniques to localize and quantify HIV RNA and DNA and the mass spectrometry-based single-cell protein quantitation technique (CyTOF) have been proven to be useful in the assessment of the HIV reservoir in various tissues with single-cell resolution. These are exciting new techniques that are being expanded to HIV latency studies. However, these assays do not directly measure replication competent virus but can shed light on the localization of latent provirus using small amounts tissue [201–204]. Additionally, the use of these assays to measure the latent reservoir in tissues can overcome the cost and cell number limitations of the QVOA and/or PCR based assays, as macrophages are difficult to obtain from tissue in sufficient numbers. These studies demonstrate that monocyte/macrophage reservoirs may not be detectable or present in all tissues but in many cases can be detected both the brain and the periphery depending on the methodology used. In addition, these studies highlight the difficulty of working with macrophages isolated from tissue and that tools specifically designed to assess the HIV reservoir in monocytes and macrophages are desperately needed.

7. Conclusions

Here we have discussed HIV infection, replication and latency in monocytes and macrophages. Monocytes and macrophages inherently present many barriers to productive infection, and conditions within these cells are not considered favorable for viral entry and initial replication. Nevertheless, once the virus surmounts these hurdles and integrates into the host genome it is likely to persist long-term. HIV infection of monocytes and macrophages is well established and relatively well understood, however, more work is needed to understand the contribution of myeloid cells to the latent viral reservoir. The majority of studies in HIV latency focus on blood as the source of latently infected cells, however, cells residing within tissues are likely to be more relevant. The short half-life of monocytes suggests that these cells cannot represent a viral reservoir. However, entry into tissues can result in maturation and their incorporation into long-lived tissue resident macrophages, which would constitute a long-lived reservoir capable of self-renewal. Studies focused on the macaque model of HIV estimate that the number of latently infected T cells and monocytes/macrophages may be comparable in blood and peripheral tissues, however, further studies are needed to determine if this is representative of vsPWH. The studies reviewed here not only show the existence of a myeloid reservoir in blood, peripheral tissues and brain, but also highlight the difficulty of working with macrophages. Specifically, designed assays need to be validated to assess the HIV reservoir in monocytes and macrophages to address these outstanding questions in the field. These questions combined with the technical challenges in detecting the HIV reservoir in monocytes and macrophages require further careful study in order to define the contributions of infected myeloid cells to the reservoir are important to the cure agenda.