Overcoming pharmacologic sanctuaries

Gut- and rectal-associated lymphoid tissue

Lymphoid tissue in the gut (gut-associated lymphoid tissue, or GALT), as well as in the rectum (rectal-associated lymphoid tissue, or RALT) is the major site of viral replication in nonvirologically suppressed patients and may also serve as a sanctuary for HIV-1. Even in the presence of long-term ART, there are reports of HIV-1 residing in this tissue [16]. Chun et al. [16] have suggested that latently infected T cells in the gut comprise a key reservoir in HIV-1-infected patients. They observed significantly higher levels of HIV-1 DNA normalized to CD4⁺ T-cell count in GALT tissue than in either resting or activated CD4⁺ T cells in the blood [16]. They found high degrees of sequence similarity between virus in the GALT and virus in peripheral blood mononuclear cells (PBMCs), suggesting the gut provides a site for further infection of circulating PBMCs. Others have shown similarity between HIV-1 sequences in the gut and plasma in HIV-1-positive individuals, suggesting that even with undetectable HIV-1 RNA in plasma, these patients still have ongoing replication in the gut, and that there is equilibrium between these two compartments [17^▪,18]. Additional studies suggest viral concentrations are higher along the entire gastrointestinal tract as compared with the blood of HIV-1+ patients treated with ART [19]. Ratios of unspliced RNA to DNA were also lower in the colon and rectum of these patients, compared with PBMCs, suggesting a low rate of HIV-1 transcription [19]. Latent virus in this reservoir can be released into the general circulation [4^▪].

The causes for these differences in viral burden are likely multifactorial. One pharmacologic explanation is the expression of drug-metabolizing enzymes and transporters is not uniform along the gastrointestinal tract; this would be expected to give rise to regional differences in antiretroviral concentrations. For example, considerable differences in expression for transporters including BCRP, multi-drug resistant protein (MRP) 2, and P-glycoprotein (P-gp) exist among patients and different sections of the gut [20,21^▪]. Small differences in P-gp, BCRP, OCTN2, and MCT1 expression occur along the length of the small intestine [20]. The colon has higher expression of MCT1, OCTN2, and MRP3 than the small intestine [20]. Human studies have shown considerable variation in cytochrome P450 (CYP) enzymes, including CYP3A expression along the length of the gastrointestinal tract, with the majority of CYP expression occurring in the proximal section of the gut [22,23]. These differences may contribute to the presence of sanctuary sites that afford the virus some degree of protection from antiretroviral suppression.

Lymph node

Mesenteric lymph nodes have been identified as a sanctuary site. Horiike et al. [10^▪▪] found effective suppression of plasma viral RNA in simian immunodeficiency virus (SIV)-infected macaques treated with ART. However, when they examined tissue from these animals, they found detectable virus levels in the spleens and lymph nodes [10^▪▪]. Detectable vDNA and vRNA in lymphoid tissue collected at necropsy has been shown in aviremic ART-treated macaques infected with a chimera of SIV containing HIV-1 reverse transcriptase [24]. These animals were treated with efavirenz, emtricitabine, and tenofovir for 26 weeks. Detectable virus was found in a variety of lymph nodes, including mesenteric, axial, inguinal, iliac, and cervical lymph nodes [24]. Bourry et al. [25] observed that in SIV-infected macaques treated with zidovudine/lamivudine/indinavir (ZDV/3TC/IDV), there were low concentrations of 3TC in lymphoid tissue. 3TC concentrations in these compartments were inversely correlated with virus persistence in these compartments [25]. Solas et al. [15] reported IDV concentrations in lymph nodes approximately two-fold greater than in plasma, whereas nelfinavir displayed a ratio of 0.58, lopinavir 0.21, and ritonavir 0.64 of drug concentration in the lymph node to drug concentration in plasma. The inability of the body to clear virus from this compartment in the presence of ART supports a hypothesis of suboptimal drug distribution and underscores one challenge to achieve complete antiretroviral suppression and eradication of the virus.

Central nervous system and genital tract

The central nervous system (CNS) and both the male and female reproductive tract are sanctuaries for HIV-1 [26,27,28^▪,29–33]. ARV distribution to these compartments has been extensively discussed and this review will not focus on these sites.

Macrophages

Macrophages, in general, and intestinal macrophages, in particular, are thought to be a reservoir for HIV-1. Zalar et al. [18] assessed the presence of virus in duodenal tissue. Similar to the findings of others, they found detectable virus despite undetectable plasma viremia. They performed flow cytometry and PCR on duodenal macrophages, and found that 0.06% of the lamina propria mononuclear cells were HIV-1-infected [18]. Whereas this percentage appeared low, given the large number of cells in this location, this actually represented a significant number of infected cells. Infected macrophages may function as a ‘Trojan horse’ for the spread of virus throughout the body, most notably to the CNS [34]. Monocytes and macrophages express efflux transporters, including P-gp and MRP, which may result in decreased intracellular concentrations of antiretrovirals medications that are substrates [35–37]. Jorajuria et al. [37] investigated the antiretroviral properties of ZDV and IDV in human monocyte-derived macrophages. They observed greater inhibition of HIV-1 replication by ZDV and IDV in cells also treated with P-gp or MRP1 inhibitors [37]. This indicates that efflux transporters expressed on macrophages maintain sub-therapeutic concentrations in these cells, and explains a potential mechanism by which macrophages may function as a sanctuary for HIV-1. It is worth noting that differences exist in the intracellular environment between macrophages and T lymphocytes, which may result in some antiretrovirals working either more or less effectively in these cells [38].

Compartment-targeted therapy

Whereas few treatments for HIV-1 have used the strategy of targeting specific compartments of ARVs for eliminating virus in these sanctuaries, this approach is being utilized for other diseases, and may hold promise for eliminating HIV-1 from restricted sanctuaries. The treatment of individuals with childhood acute lymphoblastic leukemia commonly includes intrathecal chemotherapy, with minimal long-term side effects [39]. Similarly, patients with cytomegalovirus retinitis have been treated with intravitreal ganciclovir [40]. In the treatment of HIV-1, local administration is primarily being investigated for microbicides, but may theoretically be of use for residual virus replication in sanctuary compartments. Studies assessing microbicides in primates treated with either intra-vaginal or intrarectal administration of tenofovir have displayed detectable concentrations for 24 h postadministration of the drug product [41]. Similar studies utilizing a 1% tenofovir gel vaginally in monkeys found concentrations in vaginal lymphocytes greater than the IC95 for the virus at 4 and 24 h postadministration [42].

Transporter and metabolism blockage

Strategies to increase concentrations in sanctuary sites include use of drugs to block influx and efflux transporters or the metabolism of ARVs. Ritonavir, and now cobicistat, have established roles in HIV pharmacotherapy as CYP3A enzyme inhibitors [43^▪▪,44]. In-vitro research suggests that ritonavir and cobicistat also enhance intestinal absorption. Studies with Caco-2 monolayers (a cell line derived from human epithelial colorectal adenocarcinoma cells, commonly used in in-vitro absorption studies) show increased transport of ARVs including atazanavir, darunavir, and tenofovir alafenamide fumarate (GS-7340) across the monolayers [45^▪▪]. This result is accomplished in part by inhibition of P-gp and BCRP. Similar findings have been observed for ritonavir [46^▪▪]. This inhibition of efflux transporters in the gut may lead to increased concentrations in plasma, but it is not known if this increase will have a direct effect on drug concentrations in compartments. Namanja et al. [47^▪] have recently reported an abacavir prodrug dimer that can inhibit P-gp efflux. The prodrug was designed to increase abacavir CNS penetration. They utilized in-vitro assays, as well as a brain capillary model, and found this molecule was able to inhibit P-gp in the cell culture system, and is converted back to abacavir inside cells [47^▪]. They suggest that this approach may be utilized for other antiretrovirals as a means to increase penetration past the blood–brain barrier. P-gp blockers have been studied as adjunctive treatment in cancer, but no therapeutic benefit has been established [48].

Nanomedicine

Nano-scale antiretroviral therapies (nanoART), usually defined as ARVs formulated into particles with a size of less than 100 nm, have been proposed as an approach to increase concentrations of ARVs in sanctuary sites [49^▪,50,51]. Human monocyte-derived macrophages (MDMs) exposed to nanoART consisting of atazanavir, efavirenz, and ritonavir show rapid uptake of the nanoART, and slow release over an extended time period (15–20 days) [52]. These nanoART products can be utilized to establish drug depots for a sustained response, allowing weekly, and even less frequent administration [49^▪]. Significant concentrations of nanoART products are found in a variety of tissues, including the liver, spleen, kidneys, and lungs of animals for a number of days postadministration [49^▪]. As circulating monocytes and macrophages distribute throughout the body, including to the brain and secondary lymphoid tissue, this factor has been proposed as a means to overcome barriers preventing access of antiretrovirals [52,53]. Nanoformulated ART engulfed by mononuclear phagocytes cocultured with human brain microvascular endothelial cells (HBMECs) can effectively transfer engulfed ART to these cells [54^▪▪]. When the nanoART was coated with folate they found even higher uptake in the HBMECs; folate-coated nanoART given to mice achieved three-to four-fold greater concentrations in the brain than noncoated[54^▪▪]. Kinman et al. [55] developed a lipid-IDV nanoparticle that resulted in a six-fold increase in IDV lymph node concentrations in macaques treated with the nanoparticle product as compared with a soluble formulation of IDV. Endsley and Ho [56^▪▪] have recently described peptide-coated lipid nanoparticles of IDV as a means to target the nanoparticles to CD4+ cells, and observed significantly higher antiviral activity. The development of nanoproducts may play an important role in overcoming barriers to achieving adequate concentrations in sanctuary compartments and further study is warranted [57].

Modification of drugs and prodrugs and new drug development

The process of modifying the chemical structure of existing ARVs and developing new ARVs with improved drug distribution properties offers another approach to improve pharmacokinetic/pharmacodynamic properties in sanctuary sites. Tenofovir alafenamide fumarate (GS-7340) is a prodrug of tenofovir with improved penetration into lymphoid tissue in clinical development. [58]. The drug displays variable absorption, with larger doses resulting in a greater percentage of the dose being absorbed than with lower doses, and is very stable in intestinal tissue [59^▪▪]. This situation is believed to be due to the ability of the drug to saturate intestinal transporters, including P-gp [59^▪▪]. After oral administration, GS-7340 achieves concentrations in lymph nodes 5–15-fold higher than after oral administration of tenofovir disoproxil fumarate (TDF) [60]. GS-7340 produces intracellular concentrations in PBMCs approximately 150-fold higher than in plasma. This high degree of intracellular penetration results in GS-7340 having a selectivity index (a measure of the cytotoxic concentration to the effective concentration) approximately 10-fold higher than that of TDF [60]. Another example of drug formulation technology to enhance compartment penetration is a conjugate of zidovudine and ursodeoxycholic acid [61^▪]. Ursodeoxycholic acid is capable of penetrating the blood brain barrier, and it is proposed this conjugate may allow increased concentration in the CNS. In-vitro studies utilizing human retinal pigment epithelium layers show higher permeation from the apical to basolateral compartments of the cells, as compared with the opposite, showing increased permeability of this drug product [61^▪].