| |
Cellular microRNAs contribute to HIV-1 latency in resting primary CD4+ T lymphocytes
|
| |
| |
Letter
Nature Medicine 13, 1241 - 1247 (Oct 2007)
Jialing Huang1, Fengxiang Wang1, Elias Argyris1, Keyang Chen1, Zhihui Liang1,2, Heng Tian1, Wenlin Huang2, Kathleen Squires1, Gwen Verlinghieri1 & Hui Zhang1
1. Center for Human Virology, Division of Infectious Diseases, Department of Medicine, Thomas Jefferson University, Philadelphia, Pennsylvania 19107, USA.
2. Cancer Center, Sun Yatsen University, Guangzhou, Guangdong, 510060, China.
This work was supported by grants from the US National Institutes of Health (AI058798 and AI052732) to H.Z. and the National Basic Research Program of China (grant 2004CB518801) to W.H.
Abstract
The latency of human immunodeficiency virus type 1 (HIV-1) in resting primary CD4+ T cells is the major barrier for the eradication of the virus in patients on suppressive highly active antiretroviral therapy (HAART). Even with optimal HAART treatment, replication-competent HIV-1 still exists in resting primary CD4+ T cells1, 2, 3, 4. Multiple restriction factors that act upon various steps of the viral life cycle could contribute to viral latency. Here we show that cellular microRNAs (miRNAs) potently inhibit HIV-1 production in resting primary CD4+ T cells. We have found that the 3' ends of HIV-1 messenger RNAs are targeted by a cluster of cellular miRNAs including miR-28, miR-125b, miR-150, miR-223 and miR-382, which are enriched in resting CD4+ T cells as compared to activated CD4+ T cells. Specific inhibitors of these miRNAs substantially counteracted their effects on the target mRNAs, measured either as HIV-1 protein translation in resting CD4+ T cells transfected with HIV-1 infectious clones, or as HIV-1 virus production from resting CD4+ T cells isolated from HIV-1-infected individuals on suppressive HAART. Our data indicate that cellular miRNAs are pivotal in HIV-1 latency and suggest that manipulation of cellular miRNAs could be a novel approach for purging the HIV-1 reservoir.
"Latent infection is one of the most important characteristics required for all strains of HIV-1 to survive in vivo. Our work demonstrates that HIV-1 can recruit resting-cell-enriched cellular miRNAs to control the translation of viral RNA into protein, the last step in the generation of various viral antigens. Thus, a combined miRNA inhibitor panel could be used to activate latent HIV-1 for therapeutic purposes."
The miRNAs are small, noncoding RNAs that control the expression of various target genes. They can be derived from host or viral RNA. A defensive role for cellular miRNAs against viral infection has been demonstrated in plants, insects, vertebrates and mammals5, 6, 7, whereas virus-derived miRNAs can participate in the regulation of host and/or viral gene expression8, 9, 10. It has been shown that a cellular miRNA is able to enhance the replication of hepatitis C virus by an unknown mechanism11, and a recent report has indicated that cellular miR-17-5p and miR-20a have a role in regulating histone acetyltransferase Tat cofactor expression and HIV-1 replication12. However, their role in controlling HIV-1 latency has not been well characterized. It is known that virus replication does not occur in resting primary CD4+ T cells of HIV-1-infected individuals receiving suppressive HAART, even though proviral DNA and multiply spliced or unspliced viral RNA can easily be found in these cells13, 14, 15, 16, 17. The underlying molecular mechanisms responsible for this latency are still unclear. Previous studies have proposed that the mechanisms could include transcriptional inefficiency and post-transcriptional suppression4, 15, 18, 19, 20. As the major function of cellular miRNA is to inhibit protein translation, we hypothesized that one or more miRNAs could be involved in HIV-1 latency.
We wished to determine whether miRNA(s) have a direct role in the repression of HIV-1 gene expression in resting CD4+ T cells isolated from the peripheral blood mononuclear cells (PBMCs) of normal human donors (Supplementary Fig. 1 online). We created constructs for this purpose by inserting a 1.9- or 1.2-kilobase (kb) fragment of the 3' end of HIV-1 RNA, which represents the 3' untranslated region (UTR) common to almost all HIV-1 mRNAs, into the 3' UTR of the enhanced green fluorescent protein (EGFP) gene in the vector pEGFP-C1 (Fig. 1a and Supplementary Table 1 online). The fragments were derived from various HIV-1 strains such as NL4-3, 89.6, JR-CSF, LAI.2 and YU2. GFP expression from pEGFP-C1 containing these 1.2- and 1.9-kb fragments was lower than expression from the parent vector (pEGFP-C1), as measured by median fluorescent intensity (MFI) (Fig. 1b). This indicates that the element(s) responsible for the possible miRNA-mediated inhibition are mainly positioned in this 1.2-kb region and are conserved among the various HIV-1 strains. Further, six smaller fragments (fragments A-F, Fig. 1a) dissected from the 1.9-kb fragment of the 3' end of the HIV-1 NL4-3 genome (encoded on a plasmid named pNL4-3) were individually inserted into the 3' UTR of the EGFP gene in pEGFP-C1, and the derived constructs were transfected into resting CD4+ T cells. Among these six fragments, the GFP expression from fragments B, D and F was substantially decreased (Fig. 1c). Notably, GFP expression was inhibited in resting but not in activated CD4+ T cells (Supplementary Fig. 2 online). These results suggest that fragments B, D and F could harbor potential binding sites for miRNAs that are abundant in resting CD4+ T cells and that could exert an inhibitory effect on the expression of viral proteins.
To further examine the potential for miRNA inhibition of HIV-1 protein expression, we searched for putative miRNA-binding sites in the B, D and F fragments using the MicroInspector online program (Supplementary Table 2 online). On the basis of predictions from this program, we further dissected fragment B into 11 subfragments (B1 to B11), fragment D into 9 subfragments (D1 to D9) and fragment F into 4 subfragments (F1 to F4) (Supplementary Table 1). All of these subfragments, which consist of 40-76 nucleotides each, were inserted individually into the 3' UTR region of the EGFP gene in pEGFP-C1, and their effects on GFP expression were further examined. Subfragments B5, D3, D9 and F3 all harbored potential miRNA target sequences, as GFP expression from these fragments was decreased in resting primary CD4+ T cells (Fig. 2a-c). We again predicted putative cellular miRNA-binding sites in these four subfragments by employing the MicroInspector online program (Supplementary Fig. 3a-d online), and the putative binding sites were further verified by the rna22 and RNAHybrid online programs. Additionally, we compared miRNA expression in resting CD4+ T cells with that in activated CD4+ T cells by microarray analysis. Among the miRNAs that are expressed in CD4+ T cells, 31 are at least two times more abundant in resting CD4+ T cells than in activated CD4+ T cells (Supplementary Table 3 online). We determined that subfragment B5 harbors putative miR-125b- and miR-150-binding sites, subfragments D3 and D9 harbor putative miR-223- and miR-382-binding sites, respectively, and subfragment F3 harbors a putative miR-28-binding site (Fig. 2d and Supplementary Fig. 3a-d). All five of these miRNAs were enriched in resting CD4+ T cells, as determined by microarray analysis (Supplementary Table 3). It is notable that the 'seed' regions of two miRNAs (miR-28 and miR-150) do not perfectly match their target sequences (Fig. 2d). Although perfect pairing at the seed region is important for identifying miRNA-binding sites21, it has also been demonstrated that perfect seed pairing is not essential for effective translational inhibition22.
To verify that the identified putative miRNA-binding sequences (those of miR-28, miR-125b, miR-150, miR-223 and miR-382) have roles in the suppression of HIV-1 protein expression, we created mutant analogs of the binding sequences and inserted each binding sequence or its analog into the 3' UTR region of the EGFP gene in pEGFP-C1 (Fig. 2e). All of the binding sites-that is, those for miR-28, miR-125b, miR-150, miR-223 and miR-382-were sufficient to mediate strong inhibition of GFP expression in resting CD4+ T cells. The derivative mutations substantially neutralized this inhibition, further supporting our hypothesis (Fig. 2f). Notably, these binding sites are relatively conserved in the 3' ends of various HIV-1 strains (Supplementary Fig. 4 online).
To further verify the relatively abundant expression of these miRNAs in resting CD4+ T cells, we measured miRNA abundance by stem-loop RT-PCR23. The amount of miR-150 and miR-223 expressed in resting CD4+ T cells was much greater than that in activated CD4+ T cells, whereas the amount of miR-28, miR-125b, and miR-382 was also at least twice as much in resting CD4+ T cells as it was in activated CD4+ T cells. As a control, we measured the expression of three miRNAs that were not enriched in resting CD4+ T cells in our microarray analysis (miR-124a, miR-146b and miR-155). As expected, miR-124a could not be detected in the CD4+ T cells, expression of miR-146b was almost equal in both activated and resting CD4+ T cells, and the expression of miR-155 in activated cells was higher than it was in the resting cells (Fig. 2g).
To further assess the inhibitory effects of miR-28, miR-125b, miR-150, miR-223 and miR-382 on gene expression, we conducted experiments with specific, chemically synthesized antisense inhibitors of these miRNAs. Transfection of resting primary CD4+ T cells with the combination of all five inhibitors substantially counteracted the inhibitory effects of the cellular miRNAs on the GFP expression of pEGFP-C1-NL43-1.2 (Fig. 3a). Flow cytometric analysis confirmed that these miRNA inhibitors did not affect the resting status of CD4 T cells (data not shown). To examine the effects of these specific miRNA inhibitors upon HIV-1 latency in resting CD4 T cells, we mimicked in vivo HIV-1 latency by transfecting pNL4-3, an HIV-1 infectious clone, into the resting CD4+ T cells. FACS analysis indicated that these transfected cells remain in a resting state (Supplementary Fig. 5 online). Although a small number of viral particles could be produced from these cells (Supplementary Fig. 6 online), and small amounts of viral RNA and protein could also be detected in the cells (Supplementary Figs. 7 and 8 online), the amounts of cell-associated viral RNA and protein, as well as of viral particles, were substantially increased when the cells were activated with phytohemagglutinin (PHA) (Supplementary Figs. 6-8). These data indicate that this relatively simple method can effectively mimic HIV-1 latency.
Our data further indicate that neutralization of the inhibitory effects of these five miRNAs by their corresponding 2'-O-methyl-oligoribonucleotide antisense inhibitors results in increased HIV-1 production from pNL4-3-transfected cells. Although individual inhibitors only modestly affected viral production (Supplementary Fig. 9 online), the combination of the five inhibitors resulted in a substantial increase in HIV-1 production (11.3 times higher than the control) in resting CD4+ T cells, but not in activated CD4+ T cells (Fig. 3b and Supplementary Fig. 10 online). Additionally, the viral particles in the supernatants of resting CD4+ T cells treated with the five miRNA inhibitors were infectious (Fig. 3c). FACS data indicate that these miRNA inhibitors do not affect cellular proliferation status (Supplementary Fig. 5). Moreover, transfection with the combined five miRNA inhibitors could rescue the viral production of several different HIV-1 strains (Fig. 3d). Notably, transfection of these five miRNA inhibitors into resting CD4+ T cells did not substantially alter the amounts of spliced and unspliced HIV-1 mRNA (Fig. 3e), but it did increase the expression of various HIV-1 proteins (Fig. 3f), suggesting a translational regulatory pattern for these miRNAs.
Furthermore, we analyzed the possible synergistic effect of Rev on HIV-1 expression in primary resting CD4+ T cells treated with the miRNA inhibitors by transfecting the cells with a Rev-expressing vector (pcRev). Rev expression not only increased HIV-1 protein expression by itself, a result compatible with previous reports24, but also synergized with the miRNA inhibitors to promote HIV-1 production in resting CD4+ T cells (Fig. 3b,f). Our results also imply that Rev itself is regulated by miRNA (Fig. 3f).
Finally, we examined the effect of these combined miRNA inhibitors upon HIV-1 latency in resting CD4+ T cells directly isolated from HIV-1-infected individuals receiving suppressive HAART. Post-integration HIV-1 latency in these cells was confirmed by detection of integrated HIV-1 provirus in the chromosomal DNA via Alu-PCR (see Methods; Fig. 4a, top). Stimulation of the cells with PHA induced the production of a large number of viral particles (Fig. 4a, bottom), and serial passaging experiments further indicated that these viruses were replication competent (data not shown). These data suggest that the cells harboring proviral HIV-1 DNA are indeed latently infected. After transfection with the combined miRNA inhibitors, these resting CD4+ T cells generated at least ten times more HIV-1 particles than did the cells treated with a negative control inhibitor (Fig. 4b). Next, we used a coculture assay to determine whether the viruses produced after treatment with combined miRNA inhibitors were replication competent. Resting CD4+ T cells that were transfected with combined miRNA inhibitors and -irradiated to avoid activation of the cells by alloantigens or cytokines during the coculture. The cells were then cocultured with PHA-stimulated uninfected CD4+ T cells, which were able to capture the viruses budding from the resting CD4+ T cells during the first hours after coculture. These replication-competent viruses, which were captured by activated cells, were able to spread the infection, and virus production reached very high levels (Fig. 4c-e). All of these data demonstrate that cellular miRNAs contribute to post-integration latency in HIV-1-infected individuals receiving suppressive HAART.
Our results show that differentially expressed cellular miRNAs inhibit HIV-1 expression in primary resting CD4+ T cells through their interactions with the 3' end of HIV-1 RNA, thereby contributing to viral latency. It is known that every spliced or unspliced HIV-1 mRNA, with the exception of Nef-encoding mRNA, contains the 1.2-kb fragment we used (or a portion of it) in its 3' UTR25. Thus, these cellular miRNAs, which bind the 1.2-kb fragment of the HIV-1 RNA 3' end, can inhibit the translation of almost all HIV-1-encoded proteins-including Tat and Rev, which are key in the transcription and translocation of viral RNA. The resultant inefficient synthesis of Tat and Rev proteins could further enforce the viral latency. We have identified the specific binding sites of cellular miRNAs in HIV-1 RNA, and their corresponding inhibitors can effectively counteract the miRNAs' inhibitory effects on HIV-1 production. Therefore, it is likely that these cellular miRNAs, rather than virus-derived ones, play a major role in the inhibition of HIV-1 production in resting primary CD4+ T cells. However, as cellular activation was always better than combined miRNA inhibitors at stimulating virus production (Fig. 3b and Fig. 4b), other mechanisms, such as transcriptional inefficiency, could also be important in maintaining HIV-1 latency15, 18, 19, 20.
Latent infection is one of the most important characteristics required for all strains of HIV-1 to survive in vivo. Our work demonstrates that HIV-1 can recruit resting-cell-enriched cellular miRNAs to control the translation of viral RNA into protein, the last step in the generation of various viral antigens. Thus, a combined miRNA inhibitor panel could be used to activate latent HIV-1 for therapeutic purposes.
Methods
Plasmid construction.
We obtained the various infectious HIV-1 clones (pNL4-3, pYK-JRCSF, pYU2, p89.6, pLAI.2) through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, US National Institutes of Health. The sequences in the 3' end of pNL4-3 (1.9-kb fragment, 1.2-kb fragment, and fragments A-F, Fig. 1a) were amplified by PCR and directionally cloned into the 3' UTR region of the EGFP gene in the pEGFP-C1 vector (BD Biosciences). Other subfragments, predicted miRNA-targeting sites, and their mutated analogs were directly synthesized as sense and antisense oligonucleotides. These sense and antisense oligonucleotides were annealed with each other and directly inserted into the 3' UTR of the EGFP gene in pEGFP-C1. Detailed information about the oligonucleotides is provided in Supplementary Table 1.
Isolation and culture of primary CD4+ T cells.
The blood bank of Thomas Jefferson University Hospital recruited HIV-1-seronegative blood donors, whereas the clinical section of the Division of Infectious Diseases of Thomas Jefferson University Hospital recruited HIV-1-infected individuals who had received HAART for at least 3 months, had a CD4 count 400 cells/mm3 and had blood plasma HIV-1 RNA levels lower than 50 copies/ml. All of the recruited HIV-1-infected individuals gave their informed consent, and the study was approved by the Institutional Review Board of Thomas Jefferson University. The PBMCs were isolated from the whole blood by Histopaque (Sigma) sedimentation. Resting primary CD4+ T lymphocytes were then isolated from PBMCs with CD4+ T Cell Isolation Kit II (Miltenyi), and CD25+ and HLA-DR+ cells were depleted by direct immunomagnetic conjugation. The negative fraction consisted of CD4+, HLA-DR- and CD25- resting CD4+ T cells, as described previously26, 27. The primary CD4+ T cells were activated by stimulation with PHA (5 g/ml) for 48 h, then maintained with interleukin-2 (25 U/ml; Sigma).
Synthesis of short interfering RNAs and microRNA inhibitors.
We selected the miRNA gene sequences from the Sanger Center miRNA Registry at http://microrna.sanger.ac.uk/sequences/. The short interfering RNAs (siRNAs) and miRNA inhibitors were chemically synthesized by Dharmacon. Synthetic miRNA miRIDIAN antisense inhibitors (2'-O-methyl-oligoribonucleotides) for human miR-28, miR-125b, miR-150, miR-223 or miR-382 were used in our experiments individually or in combination. The miRIDIAN microRNA inhibitor negative control 2 (anti-miR neg ctl) is based on C. elegans miR-239b (mature sequence: UGUACUACACAAAAGUACUG) and has been confirmed to have minimal sequence identity with miRNAs in humans, mice and rats.
Prediction of microRNA-binding sites.
We predicted miRNA targets using the MicroInspector algorithm at http://mirna.imbb.forth.gr/microinspector/. The cutoff values for hybridization temperature and free energy were set to 37 C and -20 kcal/mol, respectively. Identified miRNA-target gene pairs were confirmed by RNAHybrid at http://bibiserv.techfak.uni-bielefeld.de/rnahybrid/submission.html and the rna22 miRNA target predictor at http://cbcsrv.watson.ibm.com/rna22.html.
MicroRNA array analysis.
A total of 20 g RNA from resting CD4+ T cells or activated CD4+ T cells was isolated with TRIzol reagent (Invitrogen). RNA processing, microarray fabrication, array hybridization and data acquisition were performed at LC Sciences.
Transfection.
We transfected 5 g HIV-1 plasmids, pEGFP-C1 plasmid or pEGFP-C1-derived plasmids with or without 100 pmol siRNAs or 100 pmol miRNA inhibitors into resting or activated primary CD4+ T cells using an Amaxa nucleofector apparatus (Amaxa Biosystems), as described previously26.
Real-time reverse transcription PCR detection.
We designed the primers for real-time RT-PCR to detect miRNA on the basis of the miRNA sequences provided by the Sanger Center miRNA Registry (http://www.sanger.ac.uk/Software/Rfam/mirna/index.shtml) (Supplementary Table 4 online). Procedures described previously were followed, with minor modifications23. The miRNAs were isolated from resting or activated CD4+ T cells with the mirVana miRNA Isolation Kit (Ambion). For detection of HIV-1 gene expression or virion-associated RNA, the total cellular RNA or virion-associated RNA was isolated with TRIzol (Invitrogen). Supplementary Table 4c lists all primers for HIV-1 gene detection. Reverse transcription reactions were performed with the iScript complementary DNA synthesis kit (Bio-Rad). Real-time RT-PCR was performed on an Applied Biosystems 7000 Sequence Detection System (Applied Biosystems). U6 RNA was used as an endogenous control for miRNA detection, and -actin mRNA was measured as an endogenous control for HIV-1 gene expression. An in vitro-synthesized HIV-1 RNA, after quantification28, was used as the external control for measuring virion-associated viral RNA. The cycle number at which the reaction crossed an arbitrarily placed threshold (CT) was determined for each target sequence, and the amount of each miRNA relative to U6 RNA (or -actin) was described using the equation 2 CT where CT is equal to CT for miRNA minus CT for U6 RNA (or -actin).
Alu-PCR.
Genomic DNA was extracted from the resting CD4+ T cells isolated from HIV-1-infected individuals. The integrated HIV-1 was detected using primer pairs specific to Alu fragments, described previously29.
Flow cytometric analysis.
Primary CD4+ T cells transfected with EGFP-encoding plasmids were subjected to flow cytometric analysis on a Beckman Coulter cytometer (BD Biosciences) 48 h after nucleofection. The cell activation status was determined by FACS analysis of CD25+ CD69+ cells as described previously26, 27.
Antibodies.
Monoclonal antibody to Rev (1G7), rabbit antiserum to Vif, goat antiserum to gp160, and monoclonal antibody to HIV-1 p24 were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, US National Institutes of Health.
|
|
| |
| |
|
|
|
