Monocytes harbour replication-competent, non-latent HIV-1 in patients on highly active antiretroviral therapy

Abstract

Objective 

To determine whether HIV-1 can be recovered from blood monocytes as well as resting, memory CD4 T lymphocytes of patients on highly active antiretroviral therapy (HAART) with undetectable plasma viraemia and whether infection is active or latent.

Design 

Five patients with plasma HIV-1-RNA levels of less than 500 copies/ml for at least 3 months and less than 50 copies/ml at the time of sampling were initially selected, followed by an additional five patients with viral loads of less than 50 copies/ml for 3 months or more.

Methods 

Monocytes were isolated from blood by plastic adherence, then further purified by a second adherence step or CD3 depletion before co-culture with CD8-depleted donor peripheral blood mononuclear cells. Virus isolates were examined for mutations conferring resistance to reverse transcriptase or protease inhibitors and for genotype. The highly purified monocytes were also analysed for the presence of proviral and unintegrated viral DNA and multiply spliced (MS) viral mRNA by polymerase chain reaction.

Results 

Virus was recovered from monocytes of five patients. Sequencing of the recovered viruses did not reveal multiple drug resistance, and was consistent with a non-syncytium-inducing/CCR5 phenotype. Proviral DNA was detectable in monocytes from all subjects, and unintegrated HIV-1 DNA and MS RNA was found in four out of five populations examined.

Conclusion 

Recovery of replication-competent virus from some HAART patients indicates that monocytes can also harbour HIV-1. Detection of circular, viral DNA and spliced RNA, albeit at very low levels, in these cells suggests that their infection is recent and transcriptionally active rather than latent.

Introduction

Sustained suppression of HIV-1 replication to levels undetectable in plasma may be achieved through highly active antiretroviral therapy (HAART), using combinations of reverse transcriptase inhibitors and protease inhibitors (PI) ^[1]. Despite this, replication-competent virus can be recovered from resting, memory CD4 T cells regardless of the duration of viral suppression ^[2,3]. In addition, cell-associated HIV-1 genomic and multiply spliced (MS) RNA and transient, unintegrated forms of viral DNA have been detected within the peripheral blood mononuclear cells (PBMC) and lymphoid tissue ^[4–6], indicating continued, active viral replication rather than drug failure ^[2–4,7]. The time required for complete viral eradication by HAART, initially estimated at 2–3 years ^[8], is now thought to be considerably longer, if achievable at all ^[7,9,10]. Other potential viral reservoirs such as long-lived cells of the macrophage lineage are also likely to assume importance in HAART patients. In this report, we show that replication-competent HIV-1 can be recovered from highly purified monocytes, and provide evidence for recent infection of these cells.

Blood (30 ml) was initially obtained from individuals with less than 500 copies HIV-1 RNA/ml plasma [Quantiplex HIV RNA 2.0 bDNA assay; Bayer Diagnostics (formerly Chiron Corporation), Emeryville, CA, USA] for more than 3 months. Only cells isolated from five patients with less than 50 copies/ml by the more sensitive 3.0 assay at the time of sampling were used in this study (patients A–J, Table 1). After some success at virus isolation from these patients, another five with less than 50 copies/ml for at least 3 months were recruited (patients L–P, Table 1). Monocyte-enriched populations were isolated using plastic adherence, as previously described ^[11]. For patients A–J, a second adherence step in 24-well plates was followed by very thorough washing to remove non-adherent cells before the addition of phytohaemagglutinin-stimulated, CD8-depleted PBMC from HIV-negative donors [using Dynal (Oslo, Norway) anti-CD8 magnetic beads according to the manufacturer's protocol]. For patients L–P, the monocyte-enriched populations were further depleted of T cells using anti-CD3 magnetic beads before co-culture. Co-cultures were maintained for 3 weeks without the addition of fresh PBMC. HIV-1 RNA was quantified in culture supernatants using the Amplicor HIV-1 Monitor test (reverse transcriptase–polymerase chain reaction, RT–PCR; Roche Diagnostic Systems Inc., Branchburg, NJ, USA; patients A–J) or the 3.0 bDNA assay (patients L–P). For the detection of drug resistance mutations, RNA was extracted from culture supernatants and the HIV-1 pol gene amplified by RT–PCR and sequenced. The C4 to V4 region of the envelope was also amplified and sequenced. Phylogenetic analysis was performed with reference HIV-1 sequences ^[12] using Jukes–Cantor distance matrices and the Fitch/Margoliash tree-building algorithm, and bootstrap values were generated from 100 replicate datasets by consense (ANGIS, Canberra, Australia).

Table 1: Characteristics of patients studied and HIV-1 infection in their monocytes.

To estimate the level of T cell contamination, messenger RNA was isolated from 1 × 10⁶ purified monocytes using oligo (dT)₂₅ magnetic beads (Dynal), standardized by PCR for β-actin ^[13], then used in semi-quantitative PCR specific for T cell receptor mRNA as described previously ^[14]. All monocyte populations had a 0.1% or less T cell contamination by this method (Fig. 1a). The presence of MS HIV-1 mRNA encoding Nef, Rev, Tat, Vif and Vpr was also determined by semi-nested PCR, as previously described ^[15]. Proviral DNA was quantified according to the method of Chun et al. ^[16], whereas for the detection of circular DNA, extrachromosomal DNA was extracted then amplified with two long-term repeat (2-LTR) circle-specific primers as described previously ^[6].

Fig. 1.:   (a) Estimation of T cell contamination within highly purified patient monocyte populations by reverse transcriptase–polymerase chain reaction for T cell receptor mRNA. Poly A⁺ RNA was extracted from monocytes before co-culture and the equivalent of 1 × 10⁴, 3 × 10³ and 1 × 10³ cells used in reverse transcriptase–polymerase chain reaction (RT–PCR) with T cell receptor (TcR) primers. Three-fold serial dilutions of mRNA extracted from peripheral blood lymphocytes (PBL), equivalent to 10³–10⁰ cells, were made in purified, uninfected monocyte complementary DNA (1 × 10⁴ cell equivalents) and used as controls. The resultant 415 base pair PCR product was detected by liquid hybridization with a TcR probe ^[14], polyacrylamide gel electrophoresis (PAGE) and autoradiography. A representative sample from patient L is shown. Equivalent analyses for the other patients gave a similar or no signal from 10⁴ monocytes. (b) Semi-quantitative analysis of integrated HIV-1 DNA in monocytes from highly active antiretroviral therapy (HAART) patients. Genomic DNA was extracted from monocytes before co-culture and 10⁵–10³ cell equivalents were amplified in an inverse/nested PCR specific for proviral DNA. The 126 bp products were detected by hybridization and 8% PAGE. Representative results for patients M and N are shown (top panel). Genomic DNA extracted from 8E5 cells was used as the positive standard, and uninfected cell DNA as the negative control (C, middle panel). To control for the possible amplification of unintegrated viral DNA, dilutions of linearized, HIV-1 plasmid DNA (1000–1 copies) were added to a constant amount of uninfected monocyte DNA and amplified with both inverse and nested primers or nested primers alone (bottom panel). – Represents no DNA control. (c) Detection of two long-term repeat (2-LTR) circles in purified monocytes from HAART patients. Extrachromosomal DNA was extracted from highly purified monocytes from patients L–P and the equivalent of 3 × 10⁵ cells amplified by PCR with primers specific for 2-LTR circles. The 531 bp products were then detected by Southern hybridization. Lane L was exposed for longer than the other samples to detect the weak signal. Although no signal was evident even after long exposure for monocytes from patient 0, a strong band was readily detected in DNA from that patient's PBL. Ten-fold dilutions of extrachromosomal DNA extracted from peripheral blood mononuclear cells (PBMC) 10 days after infection with HIV-1_Ba−L were used as positive controls and DNA from 8E5 cells was used as a negative control. – Represents no DNA control. (d) Presence of multiply spliced (MS) RNA in monocytes. Poly A⁺ RNA was extracted from the monocytes of the same five patients as above, converted to cDNA and standardized by PCR with primers for β-actin (bottom panel). Three-fold dilutions were then amplified with primers specific for MS RNA by semi-nested PCR, labelled with 32P then analysed by 8% PAGE to detect the 121 bp products. Chronically infected ACH2 cells stimulated with phorbol myristate acetate were used as a positive control. C represents uninfected monocyte cDNA control.

Integrated viral DNA could be detected in all the highly purified monocyte populations (Table 1) at the level of one copy per 10³–10⁴ cells (Fig. 1b), as well as in all patient peripheral blood lymphocytes (PBL) preparations tested at similar levels (not shown). The level of proviral DNA found showed no correlation with the period of viral suppression. HIV-1 RNA was detected in co-culture supernatants from five patients (A, B, C, J and L;Table 1); by day 14 in A, B and C but not until day 21 in J and L. Virus could not be detected in any of the cultures within the first 7 days. Again, there was no correlation between virus recovery and the period with undetectable viral load (Table 1). Our results differ from those of Lambotte et al. ^[17] who also successfully recovered HIV-1 from three out of five patients after 21 days of co-culture, but only from monocytes activated by either lipopolysaccharide or Staphylococcus aureus. Unlike that group, we were able to recover virus from monocytes without their previous activation. In our hands, the activation of patient monocytes with lipopolysaccharide does not lead to an increase in active replication in these cells (unpublished data). Because HIV-1 has been detected from resting CD4 T cells within 1–7 days of co-culture ^[3], our results suggest that fewer monocytes than resting memory CD4 T cells are infected in HAART patients, or that virus is more difficult to recover from the former. Alternatively, virus recovered from monocytes may replicate more slowly. It might be argued that very low numbers of infected T cells could also produce delayed viral replication. As it has been estimated that only 0.2–16.4 per 10⁶ resting CD4 T cells harbour HIV-1 in HAART patients ^[2], this is extremely unlikely as the maximum number of purified monocytes co-cultured from any patient was 10⁷ with at most 10⁴ contaminating total T cells.

Reverse transcriptase and protease sequence was obtained for the virus isolates that yielded the highest levels of viral RNA in co-culture supernatants (A, B and C;Table 1). Sequence analysis did not reveal resistance mutations to both classes of antiretroviral drugs, although a number of polymorphisms or compensatory mutations were identified. All three isolates had the L63P substitution in protease, a major natural polymorphism seen pre-treatment ^[18,19], but also found in indinavir and ritonavir-resistant isolates as a compensatory mutation, improving replicative fitness ^[19–21]. Isolate C also had the K20R mutation found in some isolates resistant to indinavir, ritonavir or saquinavir, the A71T mutation associated with resistance to indinavir, but also found in PI-naive patients ^[18,22–24], and the I93L polymorphism found in both untreated and PI-treated patients ^[18]. Before successful HAART, this patient had been treated with zidovudine and didanosine but never with indinavir. Whereas single amino acid mutations may result in resistance to some PI, phenotypic resistance usually involves the accumulation of multiple substitutions, with three or more generally required before resistance becomes measurable (≥ 4-fold ^[25]). Therefore, although isolate C may have some resistance to indinavir, the lack of evidence for resistance to the other PI or reverse transcriptase inhibitors comprising that patient's effective treatment regimen strongly suggests that the isolate recovered was not multiply drug resistant. No additional mutations associated with resistance to any other antiretroviral agents currently or previously used in this or the other patients were found, indicating that the recovery of HIV-1 from monocytes was not caused by treatment failure.

The C4–V4 region of the envelope gene could be amplified only from isolates B and C. Sequences found within the V3 loop were consistent with those expected for non-syncytium-inducing phenotypes, with amino acids at key positions 306 and 320 being similar to those for the non-syncytium-inducing YU-2 rather than the syncytium-inducing NL4-3 reference strain. Also, their V3 sequences contained regions similar to the consensus motif that predicts CCR5 co-receptor usage, S/GXXXGPGXXXXXXXE/D ^[26]. Viruses recovered from seminal cells of HAART patients have similarly been found to be macrophage tropic ^[27], characteristic of strains capable of being sexually transmitted. In contrast, phenotypic analysis of virus produced by resting CD4 T cells from HAART patients showed a syncytium-inducing/CXCR4 phenotype, suggesting that activation of these cells can result in the production of highly cytopathic virus ^[28]. Our viral sequences were compared with all available published and unpublished sequences, and were found to be unique and to cluster well with subtype B isolates, bearing closest resemblance to North American strains (data not shown).

Circular episomal forms of HIV-1 DNA are labile by-products of replication and are indicative of recent infection ^[6]. We detected 2-LTR circles in four of the five monocyte populations examined (L, M, N and P), albeit at extremely low levels compared with those found in a control in-vitro-infected PBMC culture (Table 1, Fig. 1c). In the single negative patient (O), they were readily detected in the PBL fraction. In addition, again in four out of five monocyte populations (M, N, O, and P), we were able to detect MS RNA (Table 1, Fig. 1d), providing evidence for active viral transcription in these cells. In three of these, MS RNA was only detected at the highest number of cells tested, whereas patient N monocytes contained levels approaching those in the controls. The presence of MS RNA indicates transcriptional activity, although whether monocyte infection was also translationally active with virion production could not be determined. Although these indicators of ongoing infection have been found in the T cells of HAART patients by a number of groups ^[4–6,28], their detection in monocytes has not been published previously, only the presence of total viral DNA ^[17]. Our findings, therefore, suggest recent, low level, active infection in the monocytes as well as the resting, memory CD4 T cells of patients on successful HAART.

Cells of the macrophage lineage have long been considered potential reservoirs of HIV in vivo as a result of their longevity and the minimal cytopathic effect in these cells, with tissue macrophages generally considered to be more important than circulating monocytes ^[11,29–31]. This study provides evidence, however, that blood monocytes also constitute a continuing source of infectious virus in HAART patients regardless of the length of treatment. The presence of labile forms of HIV-1 DNA and the ability to recover virus from relatively short-lived monocytes (half-life ~3 days) suggest that these cells are newly rather than latently infected, possibly from reservoirs or sanctuary sites such as the central nervous system (CNS). Perivascular macrophages of the brain constitute a major subpopulation of macrophages that are infected with HIV ^[32]. The CNS is continuously patrolled by monocytes ^[33], but the blood–brain barrier prevents the efficient delivery of PI ^[34,35], the only clinically relevant drugs that can inhibit HIV-1 replication in chronically infected macrophages. The concentrations required, however, are considerably higher than those needed to inhibit the acute infection of T cells ^[36,37], and so macrophages of the CNS may be able to evade the effects of HIV-specific drugs that are highly effective elsewhere. Improvements in the range and efficacy of antiretroviral therapies and the consideration of additional approaches are required to achieve viral clearance. Whereas the activation-mediated diminution of latent T cell reservoirs of HIV-1 may prove useful ^[38], methods that will be effective against monocyte and macrophage reservoirs will also be necessary for eradication to be achievable.

Acknowledgements

The authors gratefully acknowledge Bayer Diagnostics for the provision of a bDNA assay kit for use in this study. They would also like to thank Christoph Koenigs for technical assistance, John Mills for his critical review of the manuscript and Sharon Lewin for useful discussions.