Background
HIV-1-associated immunologic and neurologic disease is dependent on the ability of the virus to infect subsets of resident immune and central nervous system (CNS) cell populations. In vitro and in vivo investigations have shown that HIV-1 infection of active CD4
+ T lymphocytes initiates a highly productive infection [
1‐
7]. In contrast, HIV-1-infected monocytic cell populations produce only limited quantities of virus due to several host-cell replication blocks including barriers that limit the reverse transcription process [
8,
9] and nuclear import [
10]. These barriers result in a more chronic infection because this cell type is more resistant to the cytopathic effects of HIV-1 gene products [
11‐
13] and has a longer lifespan in vivo. The chronic nature of HIV-1 replication in cells of the monocyte-macrophage lineage is likely a contributor to the central importance of these cells in evasion of HIV-1 detection and elimination by the immune system and the maintenance of viral reservoirs. The virus can utilize cells of this lineage as a vehicle facilitating its transport across the blood–brain barrier (BBB) and its entry into the CNS [
14‐
16], thereby promoting HIV-1-associated neuropathogenesis and the development of minor neurocognitive impairment, as well as the more severe CNS disease, HIV-1-associated dementia (HIVD). HIV-1 infection of the CNS occurs soon after infection; however, under most circumstances, prolonged productive viral replication, characterized by the formation of multinucleated giant cells with progressive loss of cognitive, behavioral, and motor deficits, is likely to occur only after severe immunosuppression and breakdown of the BBB. The pathological events that eventually lead to the development of HIVD may be initiated outside the CNS and involve the process of monocyte activation and many important events associated with passage of activated cells across the BBB. Perivascular macrophages likely play a critical role in the pathogenesis of HIVD as they are located on the parenchymal side of the BBB and the pool is continuously renewed through bone marrow-derived macrophages, particularly during CNS inflammation [
14]. Thus, the bone marrow may serve as a source of HIV-1-infected macrophages and may play a critical role in neuroinvasion and progression of CNS disease.
Genetic variation within the HIV-1 viral genome is a naturally occurring process driven by the low fidelity of reverse transcriptase, coupled with the selective pressures brought about within the host such as antiretroviral therapy, recreational drug use, immunological pressures, viral recombinatory events, host-cell phenotype, and rates of virus production [
17‐
19]. These events result in single nucleotide polymorphisms (SNPs) throughout the genome including the promoter region, designated the long terminal repeat (LTR). Genetic variation occurs within LTR binding sites where host transcription factors and viral regulatory proteins bind, altering the way the LTR drives viral transcription. The resultant viral quasispecies are likely shaped by the selective pressures operative within a variety of cellular and tissue niches that ultimately maintain specific sets of quasispecies to form viral reservoirs in susceptible cell types and end-organ tissues [
20‐
27]. The accumulation of specific LTR sequence configurations over time may also result from accumulation of poorly replicating viruses or latent proviruses in long-lived cell subsets in circulation or within viral reservoirs, such as the resting memory CD4
+ T-cells, monocytes and macrophages, and hematopoietic progenitor cells (HPCs) [
28].
Transcription factor binding sites within the LTR alternatively recruit both activating and repressing factors that can phenotypically alter the basal and inducible rates of viral gene expression [
29‐
33]. This differential regulation may ultimately impact viral replication in both a cell type- and a tissue-specific manner [
13,
21,
34,
35] that may impact the course of HIV-1 disease [
36]. Due to sequence variation within the LTR,
cis-acting transcription factor binding sites in the LTR may be functionally altered during the evolution of quasispecies, resulting in altered promoter activity [
20‐
22] and altered transcription factor binding [
23,
25‐
27]. Studies in the pre-HAART era have demonstrated that specific HIV-1 LTR CCAAT/enhancer binding protein (C/EBP) and Sp protein binding site configurations that arise as a consequence of quasispecies evolution, such as a C-to-T SNP at nucleotide position 3 in C/EBP site I within the context of a subtype B consensus sequence binding site (designated 3 T) and the C-to-T change at nucleotide position 5 in Sp binding site III (designated 5 T), were preferentially encountered in patients with more severe disease [
23,
37] and the brains of patients with HIVD [
37]. Specific nucleotide changes within the LTR, such as the 3 T and 5 T, abrogate binding of cognate transcription factors to their corresponding binding site [
23,
25‐
27,
37,
38]. Variations in the Sp GC box array of the viral promoter result in altered rates of viral gene expression and viral replication [
39‐
44]. In addition, in the HAART era the 5 T SNP in Sp site III has been found to occur almost as frequently as the consensus B (conB; derived January 2002 by the Los Alamos National Laboratory) for that site within the
DrexelMed HIV/AIDS Genetic Analysis Cohort, which currently consists of approximately 500 HIV-1-infected patients (data not shown). Furthermore, in a small subset of those patients, the 3 T and 5 T SNPs occur together. Therefore, to better understand the significance of specific SNPs found in current and past HIV/AIDS cohorts with respect to promoter function, LTR activity was examined in the TF-1 progenitor, U-937 monocytic, and the Jurkat T-cell lines within a chromatin-based environment.
Discussion
Genetic variation within the HIV-1 genome is a naturally occurring phenomenon because of the low fidelity of reverse transcriptase as well as selective pressures present within the host [
17,
18]. These viral quasispecies are potentially able to establish niches and reservoirs within HPCs, cells of the monocyte-macrophage lineage, as well as T cells that can reseed infection into the periphery and CNS of infected patients and possibly other end organs as well [
20‐
28]. Nucleotide changes within the NF-κB-proximal C/EBP and Sp transcription factor binding sites within the viral promoter have been observed in both the pre-HAART and HAART eras and have been shown to guide differential regulation of viral expression by the LTR [
20‐
22,
55‐
57]. Interestingly, the prevalence of the 3 T, 5 T, and 3T5T LTR variants in the HAART era, at least in the
DrexelMed HIV/AIDS Genetic Analysis Cohort is lower when compared to the levels observed from published studies in the pre-HARRT era [
23‐
27,
37]. However, the
DrexelMed HIV/AIDS Genetic Analysis Cohort is overall much healthier because of the prolonged use of combination antiretroviral therapy. However, we hypothesize that the prevalence of the 3 T, 5 T, and 3T5T LTR variants will become more prevalent with the development of greater disease severity across this HIV-1-infected patient population. In fact, one of the patients that has shown severe CD4 decline and neurologic impairment carries both of these variations as we have previously published [
19].
The TF-1, U-937, and Jurkat T-cell lines were used as models of HPCs, monocytes, and CD4
+ T cells, respectively, in order to study the effects of the previously described SNPs on LTR-driven gene expression within a chromatin-based microenvironment. Each cell type was stably transfected with a specific LTR variant of interest and then stored at low temperature for prolonged periods prior to experimentation. Flow cytometric analysis showed that subsequent recovery of the stored cell populations containing the HIV-1 variant LTR resulted in two different LTR-driven GFP expression phenotypes with each model cell line (TF-1, U-937, and Jurkat) and parental or variant LTR (3 T, 5 T, and 3T5T) (Figure
2). We theorized that once multiple cell samples were stored at low temperature for prolonged periods, the same expression phenotype would be maintained for each sample that was recovered from storage and passaged to approximately the same level. However, as shown in Figure
2, this was not the case; several phenotypes were associated with the specific genotypic changes within the LTRs stably transfected across all model cell lines. Alterations in gene expression from cells recovered from low-temperature storage could be caused by differences in the stably transfected cells that survive the thaw and reculturing process. Cell death is a major problem when the cell line populations are thawed and recultured, as cells with different phenotypes may be the ones surviving each time this cycle occurs. The cell clones that have very distinct expression profiles, and maintain their expression profiles during subsequent thaws and reculture from low-temperature storage, provide further evidence of this phenomenon. In another sense, a recent study showed that when rheumatoid arthritis synovial fibroblast (RASF) cells were in culture, as the culture passages increased from 2 to 8, up to 10% of the genes within this cell line are differentially expressed [
58]. In contrast to our observations, however, after long-term storage, gene expression within the RASF after passage 2 was comparable to gene expression prior to freezing [
58]. This is the first study to show that long-term storage of stably transfected cells of the HPC, monocytic, and T-cell lineages can result in alterations in gene expression patterns.
TNF-α treatment of each of the stably transfected cell line populations was performed to determine how each SNP of interest might impact the way the LTR drives GFP expression when observed from cell populations recovered from long-term low- temperature storage. Stably transfected U-937 cells containing the 3 T and 3T5T LTRs exhibited 2.34- and 2.55-fold increases in MFI, respectively, with TNF-α stimulation. The difference in LTR-driven gene transcription with LTRs containing the WT, 3 T, 5 T, and 3T5T SNPs between stably transfected TF-1 and U-937 cells could be attributed to the difference in cell types and the availability of specific transcription factors and inherent viral gene activities within those cells. Within the stably transfected TF-1 cells treated with TNF-α, the 5 T LTR resulted in a higher level of change in LTR-driven GFP expression than the WT LTR (Figure
3). This observation indicates that although this genotype (which occurs frequently in the
DrexelMed HIV/AIDS Genetic Analysis Cohort) could be involved in evading immune activation within the host under basal conditions, under stimulatory conditions it might be activated to produce a higher level of gene expression and virus production. Additionally, TNF-α stimulation of the stably transfected TF-1 and U-937 cells (Figure
3) suggested that a 3T5T LTR-containing proviral DNA could remain quiescent under certain metabolic conditions or in a basal viral gene expression mode. Only minimal intracellular stimulatory conditions would be required to drive low levels of viral gene expression and virus production within these cell populations, allowing evasion of HAART elimination of the infecting cell. The viral genome containing the 3T5T LTR could promote the evasion of effective HAART elimination, potentially contributing to the development of a niched viral reservoir for virus with specific genetic signatures that favor the quiescent or low-level gene expression mode. Subsequently, when an infected patient has a proinflammatory response to either the HIV-1 infection or an opportunistic infection, the 3T5T double-variant LTR-containing viral genotype could again result in an activation of the viral genome from a latent viral reservoir. This conclusion has been based on the increase in NF-κB binding to the 3T5T double variant and enhanced operation of these low-affinity C/EBP and Sp binding sites in the presence of elevated levels of the corresponding activation factors and a repopulation of circulating virus within the patient’s peripheral blood and lymphoid tissues and within specialized reservoirs. These changes would thereby make HAART less effective in removing and reducing virus from within the immune and nervous systems as well as other end organs.
The cell clones from each stably transfected cell line containing the WT, 3 T, 5 T, or 3T5T LTRs were developed to facilitate studies to explain the observable differences between each of the LTRs within each cell line. The low-GFP-expressing 3T5T LTR-containing TF-1 cell clone exhibited a larger difference in MFI between unstimulated and stimulated conditions compared with the intermediate-GFP-expressing cell clone. This observation may suggest that virus containing the 3T5T LTR genotype with a low enough expression profile in the body to facilitate evasion of HAART and the immune response, could, under stimulated conditions, produce a large amount of virus, reseeding infection into the body, possibly with viral strains that may be more resistant to treatment. These observations suggest that across LTRs containing each of the SNPs of interest (3 T, 5 T, and 3T5T), there is a differential expression within individual GFP-expressing clones. These results also show that within an LTR, in this case the 3T5T-containing LAI LTR, stimulation by cytokines naturally occurring during inflammation and infection can lead to differential regulation of the LTR between different cell lines. In addition, non-GFP-expressing cell clones, regardless of the LTR or cell type, could not be induced into expression with TNF-α stimulation, whereas the intermediate- and high-GFP-expressing cell clones containing each SNP of interest could be induced. This important result shows that differences in regulation of LTR-driven expression may also be attributable to the location of integration within the human genome, proviral DNA copy number, or possibly differences in epigenetic controls operative on the LTRs.
Conclusions
The experimental approach used in this study has provided a means for understanding SNP-specific LTR function within models of HIV-1-susceptible cells. However, this approach has not been without its limitations. In the model system used in this study, the function of the LTR is only observed within the context of an integrated plasmid. In addition, only the LAI backbone is used, which has been derived from an X4-tropic virus. The LAI backbone was used because (1) it has been commonly used in other scientific studies, (2) across the 12 binding sites within the LTR, it shares sequence similarity with the conB sequence except that it contains a 6G change within C/EBP site I, and (3) full-length HIV-1 LAI virus has the ability to infect all three cell line models used in this study. It would be interesting to determine the effects of the other LTR backbones, such as YU-2, an R5 virus, and 89.6, which has been defined to be an X4/R5 virus, on SNP-specific LTR driven gene expression within stably transfected TF-1, U-937, and Jurkat cells. It would also be important to determine whether the effects of these SNPs on LTR-driven gene expression observed within the stably transfected cells used in this study would remain if full-length pseudotype virus developed from molecular clones using the different HIV-1 backbones were used to infect the HPCs, monocytes, and T cells.
We have shown that the 3 T SNP within C/EBP binding site I and the 5 T SNP within Sp binding site III, both previously reported to correlate with differential binding of the C/EBP and Sp transcription factors for their cognate binding sites [
23,
25‐
27,
43,
44], cause differential regulation of gene expression by the LTR. We also showed that the combination of the two SNPs leads to low levels of LTR-driven GFP expression. The low levels of the double variant LTR-driven GFP expression could translate to a virus with a selective advantage, enhancing viral survival and facilitate the evasion of HAART therapy as well as the host immune response to viral infection. Under activating physiological conditions in the host, selected viral quasispecies of low fitness under basal conditions with appropriate
cis-acting LTR elements could retain the ability to respond to activating stimuli and replicate at higher levels. They could thereby reseed infection into numerous compartments, once again expanding the quantity of virus and spectrum of viral quasispecies in the periphery which could potentially cross the BBB reseeding virus in the CNS.
Methods
Cell culture
The TF-1 CD34+ erythromyeloid leukemia cell line (ATCC, Manassas, VA) was grown in RPMI 1640 medium with l-glutamine (Cellgro, Herndon, VA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; GemCell, West Sacramento, CA), antibiotics (penicillin and streptomycin, at a concentration of 0.04 mg/mL each; Cellgro), glucose (4.5 g/mL; Cellgro), sodium pyruvate (1 mM; Cellgro) and HEPES (10 mM; Cellgro), and recombinant human granulocyte-macrophage colony stimulating factor (2 ng/mL; eBioscience, San Diego, CA). The cells were maintained at a density between 1–5 × 105 cells/mL.
The U-937 human promonocytic cell line (ATCC, CRL-1593.2) was grown in RPMI 1640 medium with l-glutamine supplemented with 10% heat-inactivated FBS, antibiotics (penicillin and streptomycin, at a concentration of 0.04 mg/mL each), glucose (4.5 g/mL), sodium pyruvate (1 mM), and HEPES (10 mM). The cells were maintained at a density between 1–5 × 105 cells/mL.
The human T-cell line Jurkat (ATCC, TIB-152) was grown in RPMI 1640 medium with l-glutamine supplemented with 10% heat-inactivated FBS, sodium bicarbonate (0.05%), and antibiotics (penicillin, streptomycin, and kanamycin at 0.04 mg/mL each). The cells were maintained at a density between 1 × 105 and 1 × 106 cells/mL as recommended by ATCC. All cells were maintained at 37°C in 5% CO2 at 90% relative humidity.
Plasmid construction and site-directed mutagenesis
The HIV-1 LAI LTR sequence was amplified by PCR from previously published LAI-LTR-Luc reporter constructs [
59,
60]. The following primers were used for the amplification of LAI LTR: 5′-GGC-CGG-ATT-AAT-TGG-AAG -GGC-TAA-TTC-ACT-CC-3′ and 5′-GAC-CGG-TTG-CTA-GAG-ATT-TTC-CAC-AC T-G-3′ (Integrated DNA Technologies). Following amplification, the HIV-1 LAI LTR was cloned into the pEGFP-N1 vector (Clontech, Mountain View, CA), which contains the eGFP gene. The PCR product and the pEGFP-N1 vector were digested with the
Vsp I (
Ase I) and the
Bsh TI (
Age I) restriction enzymes and the digested PCR product was ligated into the digested pEGFP-N1 vector, generating the LAI LTR-GFP reporter construct.
Site-directed mutagenesis (SDM) was performed using the QuickChange mutagenesis procedure (Stratagene, La Jolla, CA) on the HIV-1 LAI LTR-GFP reporter construct to incorporate a C-to-T mutation at position 5 of the Sp site III (5 T). The following primers were used: CTT TCC AGG GAG GT G TGG CCT GGG CG and CGC CCA GGC CAC A CC TCC CTG GAA AG. The mutagenized nucleotide is underlined. This process generated the LAI LTR-5 T-GFP reporter construct. SDM was also performed on the HIV-1 LAI LTR-GFP reporter construct to incorporate a C-to-T mutation at position 3 of the C/EBP site I (3 T). The following primers were used: GCT GAC ATC GAG T TT GCT ACA AGG G and CCC TTG TAG CAA A CT CGA TGT CAG C. The mutagenized nucleotide is underlined. This process generated the LAI LTR-3 T-GFP, which was used for an additional round of SDM to incorporate a C-to-T mutation at position 5 of the Sp site III (5 T) as indicated previously. This process generated the LAI LTR-3T5T-GFP reporter constructs. For both the parental and mutated HIV-1 LTR constructs authenticity was confirmed by DNA sequencing (Genewiz, South Plainfield, NJ) and sequence analyses using Lasergene software (DNASTAR, Inc., Madison, WI).
Stable transfection of TF-1, U-937, and Jurkat cells and stably transfected cell clone development
TF-1, U-937, and Jurkat cells (ATCC) were transfected with the constructs using the AMAXA Nucleofector System and AMAXA procedure V, as described by the manufacturer (Lonza, Basel, Switzerland). Transfected cells were passaged under G418 (800 ng/mL; Mediatech, Manassas, VA) selection beginning 24 hours after transfection and passaged under G418 selection for at least 2 months to develop stably transfected populations. The TF-1 and U-937 cell populations were each serially diluted to a final concentration of 1 cell/mL media and plated into 96-well plates, where they were marked and tracked for growth. Once clonal cells became a confluent clonal population within the well, they were moved to a larger well with more media until they could be transferred to flasks and propagated normally. Basal levels of LTR-driven GFP expression within each cell population and clonal population were measured using flow cytometry.
Assessment of GFP expression utilizing flow cytometry
TF-1, U-937, and Jurkat cells were washed with flow cytometry buffer [Hanks balanced saline solution (Mediatech), FBS (3%), and NaN3 (0.02%)] and aliquots of 1 × 106 cells were fixed with paraformaldehyde (1%). Flow cytometry analysis was performed utilizing a Calibur Flow Cytometer (BD Biosciences, San Diego, CA) and analyzed the results using Flowjo version 8.8.7 software (Tree Star, Ashland, OR).
Oligonucleotide synthesis and radiolabeling
Complementary single-stranded oligonucleotides corresponding to the LAI LTR with and without the 3 T SNP in C/EBP site I and/or the 5 T SNP in Sp site III were synthesized (Integrated DNA Technologies) and annealed by brief heating at 100°C followed by slow cooling to room temperature as previously described [
37]. Blunt-ended double-stranded oligonucleotides were end-labeled using
32P-ATP and T4 polynucleotide kinase as described (Promega, Madison, WI). The sequences of the probes used in this study are as follows: LAI LTR WT: TCG AGC TTG CTA CAA GGG ACT TTC CGC TGG GGA CTT TCC AGG GAG GCG TGG CC; LAI LTR 3 T: TCG AG
T TTG CTA CAA GGG ACT TTC CGC TGG GGA CTT TCC AGG GAG GCG TGG CC; LAI LTR 5 T: TCG AGC TTG CTA CAA GGG ACT TTC CGC TGG GGA CTT TCC AGG GAG G
T G TGG CC; LAI LTR 3T5T: TCG AG
T TTG CTA CAA GGG ACT TTC CGC TGG GGA CTT TCC AGG GAG G
T G TGG CC. The SNPs are indicated as underlined nucleotides.
Electromobility shift analysis
Electrophoretic mobility shift assay binding reactions were performed as previously described [
37]. The radiolabeled LAI LTR probes were incubated with activated [TNF-α (20 ng/mL) for 24 hours] or normal nuclear extract from TF-1 cells for 30 minutes at 4°C in the presence of with poly[dI-dC)] (1 μg), bovine serum albumin (BSA), and 1× binding buffer. Reactions (20 μL, final volume) were supplemented with BSA (15 μg). The following antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) were used to determine complex formation: NF-κB p50 (H-119X), NF-κB p65 (C-20X). DNA-protein complexes were resolved at 4°C by electrophoresis for 3 hours at 200 V in a 4.5% nondenaturing polyacrylamide gel, which was then dried at 80°C for 2.5 hours and used for autoradiography.
Nuclear protein extraction was performed as previously described [
61]. Briefly, 3 × 10
7 cells were harvested for each nuclear extraction and washed once with 1× phosphate-buffered saline. The cells were incubated for 15 minutes with lysis buffer on ice and then 2 μL 10% NP-40 detergent was added and cells were inverted several times. Cells were centrifuged at 1100 RPM for 5 minutes and then resuspended, through inverting, in lysis buffer and then centrifuged again. The nuclei pellets were then resuspended in nuclear extract buffer and then incubated for at 4°C for 30 minutes on a vortexer set to a speed of 3. The nuclear lysate was centrifuged at 14,000 RPM for 10 minutes at 4°C, after which time the supernatant was removed and quantitated with respect to protein concentration as previously described [
62]. Nuclear extracts were then stored at −80°C.
Acknowledgments
These studies were funded in part by the Public Health Service, National Institutes of Health, through grants from the National Institute of Neurological Disorders and Stroke, NS32092 and NS46263, the National Institute of Drug Abuse, DA19807 (Dr. Brian Wigdahl, Principal Investigator), and under the Ruth L. Kirschstein National Research Service Award 5T32MH079785 (Jay Rappaport, PI, Brian Wigdahl PI of the Drexel subcontract). The contents of the paper are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SS contributed to the development of stably transfected cell lines, developed the stably transfected cell clones, completed all flow cytometric acquisition and analysis, contributed to study conception and design, drafted figures, and drafted the manuscript. KA constructed the LTR-plasmid constructs and contributed to the development of stably transfected populations. SD contributed to the design and completion of the electrophoretic mobility shift assays. VP helped draft the manuscript. SS, MN and BW conceived of the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.