Effect of μ-opioid agonist DAMGO on surface CXCR4 and HIV-1 replication in TF-1 human bone marrow progenitor cells

Background Approximately one-third of the AIDS cases in the United States have been attributed to the use of injected drugs, frequently involving the abuse of opioids. Consequently, it is critical to address whether opioid use directly contributes to altered susceptibility to HIV-1 beyond the increased risk of exposure. Previous in vitro and in vivo studies addressing the role of μ-opioid agonists in altering levels of the co-receptor CXCR4 and subsequent HIV-1 replication have yielded contrasting results. The bone marrow is believed to be a potential anatomical sanctuary for HIV-1. Methods The well-characterized CD34+CD38+ human bone marrow–derived hematopoietic progenitor cell line TF-1 was used as a model to investigate the effects of the μ-opioid receptor–specific peptide DAMGO (D-Ala2,N-Me-Phe4, Gly5-ol-enkephalin) on CXCR4 expression as well as infection of undifferentiated human hematopoietic progenitor cells. Results The results revealed the presence of the μ-opioid receptor-1 isoform (MOR-1) on the surface of TF-1 cells. Furthermore, immunostaining revealed that the majority of TF-1 cells co-express MOR-1 and CXCR4, and a subpopulation of these double-positive cells express the two receptors in overlapping membrane domains. Three subpopulations of TF-1 cells were categorized based on their levels of surface CXCR4 expression, defined as non-, low-, and high-expressing. Flow cytometry indicated that treatment with DAMGO resulted in a shift in the relative proportion of CXCR4+ cells to the low-expressing phenotype. This result correlated with a >3-fold reduction in replication of the X4 HIV-1 strain IIIB, indicating a role for the CXCR4 high-expression subpopulation in sustaining infection within this progenitor cell line. Conclusions These experiments provide insight into the impact of μ-opioid exposure with respect to inhibition of viral replication in this human TF-1 bone marrow progenitor cell line model.


Background
In addition to several studies linking chronic opioid use to immunomodulation [1] and increased susceptibility to bacterial infections [2], the role of opiates as potential cofactors in HIV-1 pathogenesis and disease has also been proposed. In vitro experiments that involve treatment of peripheral blood mononuclear cells with morphine prior to HIV-1 exposure resulted in increased viral replication [3]. It is now known that prolonged treatment with morphine or the selective μ-opioid receptor agonist D-Ala2, N-Me-Phe4,Gly5-ol-enkephalin (DAMGO) enhances the percentage of T cells and monocytes expressing the HIV-1 co-receptors CXCR4 and CCR5, respectively, thereby increasing the number of infected cells and the overall amount of infectious virus produced in subsequent experiments [4]. More directly, morphine treatment increases HIV-1 infection of blood monocyte-derived macrophages by upregulating CCR5 expression and inhibiting production of β-chemokines, endogenous CCR5 ligands [5].
Ongoing in vivo studies performed in the simian immunodeficiency virus (SIV)-infected rhesus macaque/ model have yielded a better understanding of the impact of prolonged morphine exposure on HIV-1 pathogenesis. Prolonged morphine exposure increased viral replication [6,7], increased the number of SIV-infected T cells [8], accelerated disease progression and neuropathogenesis [7], increased the amount of plasma virus [6,7], and increased the incidence of mortality [7]. Despite these numerous studies, a direct link between an alteration in CXCR4 or CCR5 surface expression levels and quantity of plasma virus has not been established.
The μ-opioid receptor-1 isoform (MOR-1), the best characterized isoform of the μ-opioid receptor family, has been found on cellular subsets of the immune system, as well as cells of the central nervous system, including but not limited to neurons [9][10][11]. It is possible that reported inconsistencies in the literature regarding the expression profile of CXCR4 may be attributable to a cell type-specific regulation of this chemokine co-receptor by μ-opioids. This process in turn might translate into the differential ability of μ-opioids to modulate HIV-1 replication in divergent cellular populations. To investigate the effect of μ-opioids on CXCR4 expression in human bone marrow progenitor cells, the TF-1 cell line was used; it represents a model of susceptible CD34 + /CD38 + human hematopoietic progenitor cells that are blocked at an early stage of differentiation [12]. To begin experimentation in the TF-1 cell line, experiments were performed to assess levels of MOR-1 in these cells by western immunoblot analyses, flow cytometry, and immunofluorescence microscopy. To analyze the relative surface distribution of MOR-1 and CXCR4, immunofluorescence microscopy studies were also performed. Alterations in total CXCR4 protein levels in DAMGO-treated TF-1 cells were determined using western immunoblot analyses and surface expression levels were examined using flow cytometry. We have previously demonstrated that, in addition to CXCR4, TF-1 cells express the primary HIV-1 receptor CD4 on their cell surface, thereby supporting productive infection by the HIV-1 X4-utilizing (X4) IIIB strain [13]. This observation prompted studies examining the consequence of DAMGO-mediated perturbation in CXCR4 levels on HIV-1 X4 replication in this human bone marrow-derived progenitor cell population.

Identification of MOR-1 in TF-1 cells
Western immunoblot analysis confirmed the presence of MOR-1 protein within TF-1 cells, clearly demonstrating the existence of a specific protein species at approximately 50 KDa, the expected molecular mass of human MOR-1 ( Figure 1A) [14]. As expected, the levels of MOR-1 in undifferentiated SH-SY5Y neuroblastoma cell lysates (positive control) were much higher than those observed in TF-1 lysates. In addition, the detection of MOR-1 was abrogated by preincubating the primary antibody with the MOR-1 blocking peptide. Blotting the membrane for β-actin confirmed equal loading in all lanes. Immunostaining of nonpermeabilized cells analyzed by flow cytometry ( Figure 1B) and fluorescent microscopy ( Figure 1C) clearly demonstrated that MOR-1 was expressed at the cell surface.

Relative localization of MOR-1 and CXCR4
To analyze the relative distribution of MOR-1 and CXCR4 at the cellular level, immunofluorescence studies on nonpermeabilized TF-1 cells was performed. CXCR4 and MOR-1 were shown to be co-expressed on the surface of a number of TF-1 cells. The majority of the cells quantitated (87%) expressed both MOR-1 and CXCR4 at the cell surface ( Figure 2). Far fewer cells expressed MOR-1 or CXCR4 exclusively (2% and 3% of counted cells, respectively), and 8% of cells quantitated did not express either receptor.
Interestingly, on the surface of double-positive cells, the staining for the two receptors appears to overlap, suggesting the possibility of physical proximity. A proportion of these double-positive cells co-expressed CXCR4 and MOR-1 in distinct domains on the surface that traced the circumference of the nucleus. Additionally, double-positive cells that exhibited uniform distribution of expression of the two receptors across the cell surface differed in nuclei morphology. A similar co-localization pattern of MOR-1 and CXCR4 has been previously reported in cortical neurons [10]. Given the observed changes in nuclear morphology, we proceeded to determine whether these discrete domains were specific to MOR-1 and CXCR4 or resulted from general TF-1 cellular topography. To distinguish between these two possibilities, fluorescently tagged amphipathic molecules that embed within regions of high lipid density were used to visualize the plasma membrane. This procedure clearly demonstrated that the observed discrete domains were regions of high lipid density ( Figure 3A), likely owing to membrane thinning and nuclear dispersion during M phase of the cell cycle.
Together these observations suggest a possible link between cell cycle and the observed MOR-1 and CXCR4 surface localization patterns. To better understand the influence of the phase of the cell cycle on MOR-1 and CXCR4 cell surface localization, TF-1 cells were treated with aphidicolin (10 μg/ml) for 16 hours to induce an arrest in early S phase. Following treatment, MOR-1 expression was analyzed on the surface of TF-1 cells in S phase using fluorescent microscopy and flow cytometry.
TF-1 cells in S phase expressed MOR-1 in a punctate pattern across the cell surface ( Figure 3B) and expressed MOR-1 at a higher mean fluorescent intensity (MFI) than the untreated, heterogeneous population of TF-1 cells ( Figure 3C). These observations suggested that the diverse cell surface expression patterns of MOR-1 in conjunction with CXCR4 are the result of a specific phase of the cell cycle. While the two receptors are expressed in overlapping membrane domains, suggesting the possibility of proximity, the localization in discrete membrane domains surrounding the nucleus may not be due to regulated trafficking of the two receptors.

Effects of prolonged DAMGO exposure on CXCR4 expression
To study the effects of MOR-1 signaling on CXCR4 expression, the levels of total as well as surface CXCR4 was examined in DAMGO-pretreated TF-1 cells. Western immunoblot analyses were performed on whole cell lysates prepared from TF-1 cells, in the absence or presence of two different concentrations of DAMGO (1 and 10 μM), a specific MOR-1 agonist. Total levels of CXCR4 protein (glycosylated CXCR4 monomer detected at 45 KDa) remained unchanged, suggesting that DAMGO treatment does not affect overall expression or stability of the CXCR4 protein within the cell ( Figure 4A). This observation prompted studies to determine whether the subcellular localization of CXCR4 was affected by DAMGO treatment. Immunostaining for surface CXCR4 was performed on nonpermeabilized TF-1 cells and analyzed by flow cytometry. Untreated TF-1 cells displayed a distribution among a range of MFIs, with three distinct cellular populations: a large population of CXCR4 non-expressing cells, a low-expressing cell population, and a smaller population of CXCR4 high-expressing cells ( Figure 4B). This is consistent with the heterogeneity in CXCR4 expression observed in the human hematopoietic stem cell compartment [15] and is consistent with levels of CXCR4 surface The left panel shows a subpopulation of paraformaldehyde-fixed, nonpermeabilized TF-1 cells expressing MOR-1 on the surface (detected in green with a FITC-tagged antibody directed against a region mapping within the N-terminus of MOR-1). The right panel shows a merge with DAPI (4',6-diamidino-2-phenylindole) nuclear staining of the same field. The images were taken at 60× with an Olympus IX81 deconvolution microscope. expression in primary human bone marrow cells. DAMGO (1 and 10 μM) pretreatment for 24 hours resulted in an alteration in the relative proportion of CXCR4 + cells toward the low-expressing phenotype, which was evident from an increase in the low-expressing (midrange) MFI peak and a reduction in both the non-expressing and highexpressing MFI peaks. Collectively, these studies show that total protein levels of CXCR4 remain unchanged, while surface expression levels of CXCR4 are altered following DAMGO exposure.
Effects of prolonged DAMGO exposure on HIV-1 replication in TF-1 cells Given the observed changes in the level of CXCR4 surface expression following DAMGO treatment and the understanding that co-receptor expression levels correlate with cell susceptibility to HIV-1 infection [16], the level of impact DAMGO would have on HIV-1 replication was of interest. HIV-1 core antigen (p24) assays were used to assess HIV-1 replication in TF-1 cells in the presence of DAMGO. In comparison with untreated cells, DAMGO pretreatment for 24 hours resulted in a dose-dependent decline in p24 production ( Figure 5). Specifically, p24 levels dropped from 885 ng/ml in untreated cells to 386 ng/ml in cells treated with DAMGO at 1 μM (p < 0.001) and 254 ng/ml in cells treated with DAMGO at 10 μM (p < 0.001). This decline, which ranged from 2.3-to 3.5-fold, was partially reversed upon CTAP (D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH) pretreatment to 763 ng/ml (p < 0.001). Altogether, the results suggested that exposure to μ-opioids might render bone marrow progenitor cell populations less susceptible to HIV-1 infection.

Discussion
Because HIV-1 has been reported to penetrate the bone marrow [17,18] of infected patients, it has been suggested that cellular subsets of this important compartment may serve as reservoirs for the virus. We aimed to determine the effects of the specific MOR-1 agonist DAMGO on viral replication in human bone marrow progenitor cells. The human CD34 + /CD38 + TF-1 bone marrow-derived hematopoietic progenitor cell line was used as an in vitro model because it is known to express the HIV-1 receptor CD4 as well as the co-receptor CXCR4.
MOR-1 is the most extensively characterized receptor subtype of the family and it has been identified on terminally differentiated mature immune cell populations, including human CD4 + cells, cells of the monocyte-macrophage lineage [19], and more primitive CD34 + 38 − peripheral blood and cord blood stem cells [20]. Its presence on more terminally differentiated progenitor cells and the concomitant underlying significance of its presence have, however, not been fully elucidated. In this regard, we have clearly demonstrated the presence of MOR-1 in a TF-1 human bone marrow progenitor cell population.
MOR-1 has also been shown to oligomerize with other G protein-coupled receptors (GPCRs) such as the chemokine co-receptor CCR5 [21] with an adverse impact observed on the G-protein coupling of the other oligomerization partner. Alterations in G-protein coupling, differential affinity of specific agonists, and enhanced internalization have also been reported for the μ-δ opioid receptor (MOR-DOR) hetero-oligomeric complex [22]. Therefore, the presence of MOR-1 and CXCR4 in overlapping domains on TF-1 cells leaves open the possibility of dimerization between these receptors. Such localization of MOR-1 has also been reported in neuronal membranes of the rat caudate putamen wherein μ-opioid receptors were found in cholesterol and sphingolipid-rich membrane subdomains known as lipid rafts [20], where they seem to be important in regulation of GPCR signaling, receptor phosphorylation, and membrane trafficking [23]. This observation raises interesting questions concerning the biological significance of MOR-1 localization in TF-1 progenitor cells. The results reported herein have also revealed MOR-1-mediated alteration in surface CXCR4 levels, in line with our previous studies [10,11]. It has been previously reported that CXCR4 signaling is impeded by DAMGO in rat cortical neurons [10,11]. Therefore, future studies will be focused on investigating whether molecular events triggered by MOR-1 signaling affect CXCR4 receptor recycling and/or CXCR4 signaling in bone marrow progenitor cells. These studies also provide a unique insight into the role of μ-opioids in modulating HIV-1 replication in bone marrow progenitor cells. In the setting of HIV-1 infection, downregulation of CXCR4 might translate into a reduction in HIV-1 replication under the influence of μ-opioids, thereby serving a protective function, as suggested in studies performed in macaques [24]. As shown here, DAMGO exposure resulted in a decrease in HIV-1 replication. While this decrease may be the result of a decrease in viral entry involving CXCR4, the possibility remains that this decrease is the result of inhibition of post-entry events in the viral life cycle. Paradoxically, because the SDF-1-CXCR4 axis is vital for survival of progenitor cells [25], a reduction in

Western immunoblot analysis
To identify MOR-1 at the protein level, an N-terminusspecific antibody was used that recognizes a region mapping within the extracellular domain of human MOR-1 (N-20, Santa Cruz Biotechnology, Santa Cruz, CA). To demonstrate the specificity of interaction, assays were performed under nonblocking and blocking conditions. Whole cell lysates were prepared in RIPA buffer (Pierce ECL) and used to determine protein concentrations utilizing the BCA protein assay as described by the manufacturer (Pierce ECL). An equal amount of protein was run on a 10% sodium dodecyl sulfate-polyacrylamide gel and transferred to a 0.45-μm Immobilon-P polyvinylidene fluoride membrane followed by probing with the MOR-1 polyclonal antibody and detecting with a horseradish peroxidase (HRP)-conjugated antigoat IgG antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). As a negative control, the primary antibody was preincubated with a MOR-1-specific blocking peptide (1:1) (Santa Cruz Biotechnology) before probing for the receptor protein. Because SH-SY5Y neuroblastoma cells are known to constitutively express high levels of MOR-1 [30], they were used as a positive control in the experiments. Equal loading was confirmed by stripping the membrane using the Restore Western Blot Stripping Buffer (Thermo Scientific) and reblotting for β-actin with a mouse monoclonal antibody from Sigma-Aldrich. To study the total CXCR4 protein expression resulting from prolonged DAMGO exposure, TF-1 cells were first starved for 1 hour. After washing in 1× phosphate-buffered saline (PBS), aliquots were kept untreated or treated with DAMGO (1 or 10 μM) for 24 hours followed by lysis and processing for western immunoblot analysis using a polyclonal anti-CXCR4 antibody (H-118) (Santa Cruz Biotechnology) and an HRP-conjugated antirabbit IgG antibody (Jackson ImmunoResearch Laboratories). All immunoblots were visualized using a chemiluminescent detection procedure as described by the manufacturer (Pierce ECL) on a ChemiDoc acquisition/analysis station (Bio-Rad Laboratories, Hercules, CA).

Immunofluorescence analysis
To examine cellular distribution of MOR-1 and to compare the expression pattern of MOR-1 with that of CXCR4 on the surface of TF-1 cells, immunofluorescence studies were performed. Cells were incubated on poly-D-lysine-coated coverslips. Where indicated, cells were treated for 16 hours with aphidicolin (10 μg/ml) to arrest cells in early S phase. Cells were subsequently fixed with 2% followed by 4% paraformaldehyde. So that we could visualize the surface expression of the two receptors, TF-1 cells were not permeabilized prior to staining. Sequential staining began with primary labeling with an N-terminus-specific MOR-1 antibody, N-20 (Santa Cruz Biotechnology), followed by secondary labeling with a fluorescein isothiocyanate-conjugated antigoat IgG antibody. Subsequently, CXCR4 was labeled with an N-terminus-specific phycoerythrin-conjugated antihuman CXCR4 monoclonal antibody, 1D9 (Pharmingen, BD Biosciences, San Jose, CA), followed by secondary staining with a rhodamine (TRITC)-conjugated antirat IgG antibody (Jackson ImmunoResearch Laboratories). Each step was accompanied by washing three times with PBS. Negative controls (not shown) consisted of "no primary" and "isotype primary" antibody conditions. Cells were mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA) and observed under an Olympus IX81 deconvolution microscope at 60× magnification. A total of 1000 cells were counted following three independent staining experiments, and the percentage of cells in each phenotype was calculated based on the total number of cells counted. Averages reported were calculated based on the three independent experiments, and error bars represent the standard deviation of the results.

Flow cytometry
To determine mean fluorescence intensity of surface CXCR4 on TF-1 cells, the cells were serum starved for 1 hour, then resuspended in TF-1 serum free media. Aliquots of 5 × 10 5 cells were incubated in the absence or presence of two different concentrations of DAMGO (1 and 10 μM) for 24 hours. These concentrations were selected because 10 μM represents the upper range of morphine observed in the serum of morphine-dependent animals [31]. With respect to the MOR-1 analysis, where indicated, cells were incubated in the absence or presence of aphidicolin (10 μg/ml) for 16 hours to arrest cells in early S phase. Collected cells were then washed with FACS wash buffer (Hanks balanced saline solution [Mediatech], 3% FBS, and 0.02% NaN3). So that we could visualize the surface expression of the two receptors, TF-1 cells were not permeabilized prior to staining. Cells were then reacted with a titrated amount of a monoclonal phycoerythrin-conjugated antihuman CXCR4 antibody 12G5 (R & D Systems, Minneapolis, MN) or MOR-1 (N-20, Santa Cruz Biotechnology) on ice for 30 minutes. Concentration-matched IgG2A isotype or whole IgG control antibody preparations were used as negative controls. Subsequently, cells were again washed with FACS wash buffer and fixed with 1% paraformaldehyde. We performed flow cytometry using a FACSCalibur Flow Cytometer (BD Biosciences) and analyzed the results using FlowJo version 6.1.1 software (Tree Star, Ashland, OR).

HIV-1 p24 ELISA
To assess the effects of DAMGO pretreatment on HIV-1 replication in TF-1 cells, p24 assays for measuring viral core antigen were performed. Cells were seeded into 12-well plates at a density of 0.5 × 10 6 cells/ml and pretreated for 24 hours with DAMGO (1 or 10 μM) alone or DAMGO (10 μM) and CTAP (1 μM) in TF-1 serum free media. After 24 hours, treated cells were seeded in a new 12-well plate at a concentration of 0.5 × 10 6 cells/ml/well. This step was followed by exposure to the X4-utilizing (X4) HIV-1 IIIB strain at a titer of 10 5 median tissue culture infective dose (TCID 50 ) (Advanced Biotechnologies, Inc., Columbia, MD). Two hours after infection, cells were washed with PBS, resuspended in TF-1 media containing serum, and reseeded in 12-well plates. Following incubation for 24 hours, the supernatant was collected and assayed for p24 core antigen using an Alliance p24 ELISA procedure as described by the manufacturer (Perkin Elmer, Waltham, MA). All samples were assayed in triplicate and p24 values were normalized for 1 × 10 6 cells/ml.

Statistical analysis
Significance was determined using the student t-test, considering p values of <0.05 significant.