Address correspondence and reprint requests to Dr William F. Maragos, University of Kentucky Medical Center, Department of Neurology, Kentucky Clinic, Room L-445, Lexington, KY 40536-0284,USA. E-mail: email@example.com
The human immunodeficiency virus (HIV)-1 transactivating protein Tat may be pathogenically relevant in HIV-1-induced neuronal injury. The abuse of methamphetamine (MA), which is associated with behaviors that may transmit HIV-1, may damage dopaminergic afferents to the striatum. Since Tat and MA share common mechanisms of injury, we examined whether co-exposure to these toxins would lead to enhanced dopaminergic toxicity. Animals were treated with either saline, a threshold dose of MA, a threshold concentration of Tat injected directly into the striatum, or striatal injections of Tat followed by exposure to MA. Threshold was defined as the highest concentration of toxin that would not result in a significant loss of striatal dopamine levels. One week later, MA-treated animals demonstrated a 7% decline in striatal dopamine levels while Tat-treated animals showed an 8% reduction. Exposure to both MA + Tat caused an almost 65% reduction in striatal dopamine. This same treatment caused a 56% reduction in the binding capacity to the dopamine transporter. Using human fetal neurons, enhanced toxicity was also observed when cells were exposed to both Tat and MA. Mitochondrial membrane potential was disrupted and could be prevented by treatment with antioxidants. This study demonstrates that the HIV-1 ‘virotoxin’ Tat enhances MA-induced striatal damage and suggests that HIV-1-infected individuals who abuse MA may be at increased risk of basal ganglia dysfunction.
Several lines of evidence indicate that the basal ganglia is vulnerable to damage by human immunodeficiency virus (HIV)-1 infection. During active infection, these structures contain among the brain's highest viral burden as well as HIV-1-infected macrophages and multinucleated giant cells (Kure et al. 1990), which cause loss of dopaminergic neurons in the substantia nigra (Reyes et al. 1991) and striatal atrophy (Berger and Nath 1997). Signs of parkinsonism have also been observed in the ‘acquired immune deficiency sydrome (AIDS)–dementia complex’ (Mirsattari et al. 1998). Underlying these abnormalities are decreased levels of dopamine in the CSF (Berger et al. 1994) and reduced levels of both dopamine and homovanillic acid in the caudate nucleus of patients with AIDS (Sardar et al. 1996). It has been suggested that products released from HIV-1-infected cells (e.g. ‘virotoxins’) are responsible for neuronal damage (Kolson and Pomerantz 1996; Nath and Geiger 1998). One such product is the HIV-1 gene regulatory protein Tat.
Currently, HIV-1 infection is considered an epidemic within the drug abusing community. In light of the observations that HIV-1 damages the same areas of the nervous system as those affected by chronic MA use, and that MA and Tat may share common mechanisms of toxicity, we have examined whether the HIV-1-associated protein Tat and MA modulate each other's toxic potential.
Materials and methods
Animals and supplies
Twelve-week-old (approx. 325 g) male Sprague–Dawley rats were used for animal studies. All animal use procedures were in accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the University of Kentucky Institutional Animal Care and Use Committee. Recombinant Tatl-72 HIVBRU was made and purified as described previously (Ma and Nath 1997). All chemicals used were reagent grade. MA was obtained from the National Institute on Drug Abuse (Rockville, MD, USA).
Four groups of animals were typically used in these experiments. Animals were either treated with (1) i.p. saline (control group), (2) i.p. MA, (3) striatal Tat + i.p. saline or (4) striatal Tat + i.p. MA. Animals injected with Tat were anesthetized with 50 mg/kg pentobarbital i.p. and placed in a small-animal stereotaxic apparatus. The surface of the skull was exposed and a burr hole was made at the selected site. The coordinates for the striatum were AP = + 0.7 mm, ML = 2.5 mm and DV = −4.0 mm relative to bregma. Tat was dissolved in sterile water and 20 µg Tat (in 1 µL) was injected over 1 min using a Hamilton syringe with a 30-gauge needle. To control for non-specific effects of Tat, a parallel experiment was conducted in which animals received intrastriatal injections of an equivalent amount of heat-inactivated fetal bovine serum (FBS) instead of Tat. Following injections, the needle remained in place for 1 min, after which the scalp was cleaned with sterile saline and sutured. Animals were allowed to recover from anesthesia under a heat lamp and returned to their cages once they displayed normal activity. Twenty-four hours later, Tat-injected animals were injected four times at 2-h intervals with either 5 mg/kg MA i.p. or with sterile saline using the same dosing paradigm. The concentration of Tat and MA chosen for these experiments have each been shown, in previous studies, to be ‘threshold’ and to result in no greater than 10% reduction in total striatal dopamine levels (data not shown). One week later, animals were killed by decapitation and the brains were quickly removed. Using a brain mold, a 2-mm slab of brain was cut; the striatum was dissected, frozen on dry-ice and stored at −70°C. For animals that received Tat injections, visual inspection was used to verify that the tissue slab contained the needle tract.
Effects of MA on core temperature
Because of the concern that MA-induced hyperthermia might confound our interpretation of any interactions between MA and Tat, we sought to verify that the concentration of MA used in these studies did not raise the body temperature. To accomplish this, we measured the rectal temperature 1 h before animals were treated with either saline or MA, and then 1 h after each of the four i.p. injections.
Effect of MA and Tat on core temperature
We also determined whether MA and Tat interacted to induce hyperthermia. Four groups (n = 6 animals/group) of animals were examined: (1) intraparenchymal heat-inactivated FBS + i.p. saline, (2) intraparenchymal Tat + i.p. saline, (3) intraparenchymal heat-inactivated FBS + i.p. MA and (4) intraparenchymal Tat + i.p. MA. Rectal temperatures were monitored 1 h before the first i.p. injection, 1 h following the first, second and third injections, and then 1, 3 and 5 h after the final injection.
HPLC for biogenic amines
Frozen tissue samples were sonicated in 300 µL cold 0.1 m perchloric acid containing dihydroxybenzylamine as an internal standard. Samples were centrifuged for 5 min at 12 000 g and the supernatant filtered through a 0.22-µm pore size membrane. The filtrate was diluted with HPLC mobile phase and 50 µL was injected on to a ESA Hypersil ODS C18 column (3-µm particles, 4.6 mm × 80 mm; ESA, Chelmsford, MA, USA). The flow rate was 1.4 mL/min and the mobile phase comprised 0.17 m citrate–acetate buffer, pH 4.1 (containing 50 mg/L disodium EDTA, 130–140 mg/L octanesulfonic acid and 7% methanol). Monoamines were measured with an ESA model 5200A Coulochem II electrochemical detector with a model 5011 dual-detector analytical cell (detector 1 set at + 350 mV and detector 2 at − 300 mV). Chromatograms were recorded from both detectors and peak heights were used to calculate recovery of internal standard and amounts of monoamines and metabolites.
Dopamine transporter (DAT) binding
Tat-injected animals received either MA or saline 24 h systemically after surgery. One week later, the animals were killed by decapitation, the brains frozen and 20-µm cryostat sections were obtained throughout the rostrocaudal extent of the striatum. Binding to the DAT was determined using [125I]RTI-121 (3 beta-(4-iodophenyl) tropane-2 beta-carboxylic acid isopropyl ester) (New England Nuclear Life Science Products, Boston, MA, USA). This compound binds specifically to the cocaine binding site on the DAT and, because it is iodinated, requires a shorter exposure time than tritiated radioligands such as mazindol (Boja et al. 1998). A total of four tissue sections from each brain (obtained 20 and 40 µm anterior and 20 and 40 µm posterior to the center of the needle tract) were pre-incubated in a buffer composed of 50 mm Tris–HCl, 120 mm NaCl and 5 mm KCl, pH 7.4, for 30 min at 4°C. The sections were then transferred to a binding buffer of identical composition containing 5 nm[125I]RTI-121 and incubated at 4°C for 90 min. The sections received four 15-min washes in pre-incubation buffer at 4°C, once in diluted incubation buffer for 10 s and once in deionized water for 10 s. Non-specific binding was assessed in the presence of 10 mm unlabeled mazindol. Following washing, all sections were gently dried with ambient air flow from a desktop fan. The sections were desiccated overnight under a vacuum at room temperature (23°C) and then exposed overnight (about 18 h) to Amersham Hyperfilm Bmax (Amersham, Piscataway, NJ, USA). Tissue paste standards containing various amounts of radioactivity were included in all film exposures. After the films had been developed, the sections were stained with thionin so that the autoradiographic images could be directly compared with the corresponding stained sections. Using NIH Image v.1.61 software (http://www.nih.gov), the molar quantities of bound ligand of the entire striatum in each tissue section (four sections per animal) were measured and mean values determined.
Cultures of human brain cells
Brain specimens were obtained from human fetuses of 12–14 weeks' gestational age, with consent from women undergoing elective termination of pregnancy and approval by the University of Kentucky Institutional Review Board. Neuronal cultures were prepared as described previously (Magnuson et al. 1995; Nath et al. 1996). Approximately 60% of these cells are immunoreactive for dopamine and the DAT, and are therefore considered to be dopaminergic (Turchan et al. 2001). Briefly, the cells were mechanically dissociated, suspended in Opti-MEM with 5% heat-inactivated fetal bovine serum, 0.2% N2 supplement (Gibco, Rockville, MD, USA) and 1% antibiotic solution (penicillin G 104 units/mL, streptomycin 10 mg/mL and amphotericin B 25 µg/mL), and plated in flat-bottom 96-well plates. The cells were maintained in culture for at least 1 month before conducting neurotoxicity and mitochondrial membrane potential (Ψ) experiments.
Neurotoxicity assay and measurement of Ψ
At the time of experimental treatment, the culture medium was replaced with Locke's buffer containing (in mm) 154 NaCl, 5.6 KCl, 2.3 CaCl2, 1 MgCl2, 3.6 NaHCO3, 5 glucose, 5 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, pH 7.2, and neuronal cultures were incubated with Tat (80 nm), MA (500 µm) or Tat + MA. The concentrations of Tat and MA chosen have been determined to be ‘threshold’ and result in toxicity only slightly greater or equal to that in vehicle-treated cells (Turchan et al. 2001). Cell death under each condition was determined 15 h after treatment using trypan blue exclusion as described previously. Dead neurons were counted from five fields in five predetermined locations, which were photographed and coded. A minimum of 200 cells per field was counted. Each experiment was conducted in triplicate wells and replicated at least three times.
As an index of neuronal injury, we determined the effect Tat, MA and Tat + MA on Ψ using the fluorescent dye JC-1. Once loaded into the mitochondria, JC-1 undergoes aggregate formation in the regions of high potential. The resulting spectral shift of the dye can be used to detect changes in mitochondrial activity. The green fluorescent JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcabocyanine iodine) exists as a monomer at low membrane potential. However, at higher potentials, JC-1 forms red fluorescent ‘J-aggregates’ that exhibit broad spectrum and an emission maximum at 590 nm (Smiley et al. 1991). Optical filters designed for fluorescein and tetramethylrhodamine were used to separately visualize the monomer and J-aggregate forms respectively. Ψ was monitored 6 h following treatment with the above compounds at different concentrations and time points. At this time point, treated cells are still viable and any changes in Ψ thus reflect an alteration in mitochondrial function and not number. After experimental treatment, cells were incubated for 30 min at 37°C in a 5% CO2 incubator in the presence of 10 µm JC-1 and then washed in Locke's solution. Optical measurements were acquired with excitation at 485 nm and emission at 527 nm and 590 nm. The levels of fluorescence at both emission wavelengths were quantified and the ratio of measurements was assessed. To investigate the neuroprotective properties of antioxidants, cells were incubated with either diosgenin, EUK-8 or the glutathione-mimetic Ebselen (Sigma, St Louis, MD, USA) at various concentrations followed by exposure to the viral proteins and/or MA in Locke's buffer, and cell death and Ψ were determined.
The pertinent data are given as mean ± SEM. All data were analyzed by anova with appropriate post-hoc testing.
Animals treated with a combination of MA + Tat displayed normal behavior that was indistinguishable from that of saline-treated animals or those treated with MA or Tat alone. In tissue sections stained with cresyl violet, the striatum revealed no overt pathology among the four groups. The corpus callosum appeared intact and there was no obvious evidence of cell loss or hypercellularity.
Effects of Tat ± MA on biogenic amines
As predicted, there was a synergistic interaction between MA and Tat (Fig. 1). Animals that received 20 µg Tat showed a modest (8%) reduction in striatal dopamine 7 days after surgery compared with saline-treated controls. Similarly, animals treated with 5 mg/kg MA showed an 7% reduction in striatal dopamine. Animals that were treated MA + Tat showed a dramatic decrease in striatal dopamine. In these animals, dopamine levels in the striatum were depleted by 65%.
Consistent with these observations, there was no significant change in the levels of the major dopamine metabolite dihydrophenylacetic acid (DOPAC) in animals treated individually with either Tat or MA (Fig. 2). There was, however, a nearly 30% reduction in DOPAC in animals exposed to both MA + Tat. Surprisingly, while there were no differences in striatal levels of the minor dopamine metabolite homovanillic acid (HVA) in MA-treated animals, there was an almost two-fold increase in this metabolite in Tat-treated animals. HVA returned to baseline levels in animals administered both MA + Tat (Fig. 2). No synergistic toxicity was observed in animals treated with both heat-inactivated FBS and MA (data not shown).
In contrast to the synergistic effects of MA and Tat on striatal dopamine levels, the combination of these two toxins revealed no such interaction on the serotoninergic system (Fig. 3). While MA alone did not alter striatal serotonin levels, Tat alone resulted in a small (17%) but significant ( p < 0.04) reduction in this transmitter. Although there was a slight enhancement in the decrease in serotonin when animals were exposed to Tat + MA, this was not significantly different from the loss caused by Tat alone. The serotonin metabolite 5HIAA was unchanged in animals treated with MA alone but, interestingly, increased by 30% in Tat-treated animals ( p < 0.002) but remained unchanged in animals treated with MA + Tat.
Effect of MA on core temperature
Because of our concern that any interaction between MA and Tat might be the result of hyperthermia that typically occurs with MA toxicity in rodents, we measured the core temperature of a group of rats treated with the same dosing regimen of MA used in the experiments outlined above in comparison to that of animals treated at the same time with vehicle (saline). Although there were some fluctuations in body temperature during and after MA administration, there were no significant differences between the two groups at any time point (Fig. 4).
Effect of MA and Tat on core temperature
Animals receiving intraparenchymal injections of either heat-inactivated FBS or Tat followed by systemic injections of saline did not develop a febrile response (Fig. 5), nor did animals injected with heat-inactivated FBS in the striatum followed by administration of saline. Surprisingly, animals administered Tat followed by MA became hyperthermic. In this group of animals, the temperature began to rise after the third injection of MA, peaked 1 h after the fourth injection and returned to baseline 2 h later.
In this experiment, we performed quantitative autoradiography of the presynaptic DAT using [125I]RTI-121 as an index of DA terminal integrity (Fig. 6). In control animals that received systemic injections of saline DAT binding was 24.2 ± 1.8 fmol per mg wet-weight. In animals injected with Tat and 24 h later with MA, [125I]RTI-121 binding was 10.6 ± 4.7 fmol per mg wet-weight in the ipsilateral striatum, representing a 56% reduction in binding to DAT ( p < 0.05). As an internal control, [125I]RTI-121 binding in the contralateral striatum (i.e. the side exposed to MA only) was determined to be 21.2 ± 3.3 fmol per mg wet-weight, representing a 12% reduction compared with the average binding in saline-treated animals. Thus, the reduction in [125I]RTI-121 binding to the DAT caused by either MA alone or MA + Tat paralleled the changes in DA levels measured using biochemical methods. Sections treated with mazindol did not produce images greater than film background.
Cell culture studies
Treatment with 80 nm Tat alone resulted in 2.3 + 0.44% cell death, which was not significantly different from that in controls (1.6 ± 0.13%) (Fig. 7). Cultures treated with 500 µm MA had a slightly higher degree of neuronal loss (2.5 ± 0.31%), which was significantly different from that in controls. As expected, there was a profound enhancement of cell death in cultures exposed to both MA and Tat (5.8 ± 0.62%; p < 0.001). When the amount of cell death measured in control cultures was subtracted from each treatment group, the degree of enhancement in MA + Tat-treated cultures was 300%, which is similar in magnitude to the synergistic effect observed in our in vivo studies.
In order to examine more subtle perturbations in neuronal function, we measured Ψ using the ratiometric dye, JC-1. As can be seen in Fig. 8, exposure of fetal neurons to 80 nm Tat alone had almost no effect (∼2% reduction) on Ψ. In a similar manner, exposure to 500 µm MA also failed to cause a significant change in Ψ (∼6%). However, when cells were exposed simultaneously to both Tat and MA, there was a 20% reduction in Ψ ( p < 0.01). This reduction was prevented when the antioxidant diosgenin, EUK-8 or Ebselen was included in the incubation media. Cells treated with these antioxidants alone showed no change in Ψ (data not shown) indicating that the observed protection was most probably a result of their antioxidant capacity and not a non-specific effect on mitochondrial membranes.
The present results demonstrate that the HIV-1 Tat protein and MA interact synergistically to deplete striatal dopamine levels. The interaction appears to be specific for the Tat protein since neither equivalent amounts of heat-inactivated bovine serum nor the HIV-1 coat protein gp120 (W. F. Maragos, unpublished observations) were shown to synergize with MA. The profound loss of dopamine appears to be a specific response to MA + Tat since serotonin levels were not similarly affected. Studies conducted in parallel using human fetal cortical neurons, of which approximately 60% are dopaminergic (Turchan et al. 2001), also demonstrate synergistic neurotoxicity when sublethal concentrations of Tat and MA are used. Interestingly, only a fraction of these dopamine-positive neurons are sensitive to MA + Tat (Turchan et al. 2001), suggesting that as yet unidentified factors contribute to the vulnerability of dopaminergic neurons in vitro.
Consistent with the observation that the combination of Tat + MA decreased production of dopamine was the simultaneous reduction in DOPAC. In rat, DOPAC is the major metabolite of dopamine, and a coincidental reduction in DOPAC and dopamine is suggestive of reduced dopamine synthesis. Unexpectedly, Tat alone caused a increase in levels of the minor metabolite HVA although the HVA level was normal in animals exposed to both Tat and MA. These findings suggest that Tat may have distinct effects on this alternate metabolic pathway. Alternatively, Tat may inhibit dopamine re-uptake, thereby driving dopamine through the catechol-O-methyl-transferase pathway responsible for the production of HVA. Unlike the changes in dopamine, treatment with Tat + MA had no greater effect on serotonin levels than Tat alone, which caused a modest (17%) but significant reduction in this transmitter. Thus, the dopaminergic system demonstrates preferential vulnerability to the interactions between Tat and MA.
Repeated administration of MA to experimental animals and possibly to humans results in destruction of dopamine terminals and a reduction in the levels of striatal dopamine (Wagner et al. 1980; Lorez 1981; Ricaurte et al. 1982; Trulson et al. 1985; Brunswick et al. 1992; Nakayama et al. 1993). In humans, despite prolonged periods of abstinence, alterations in the dopaminergic system may be long lasting (McCann et al. 1998; Volkow et al. 2001). The observed loss of striatal [125I]RTI-121 binding following treatment with both MA and Tat is consistent with injury to these structures and suggests that Tat may enhance the toxic potential of MA. Based on autoradiographic methods alone, however, we cannot rule out the possibility that the MA + Tat-induced reduction of striatal dopamine resulted from a disruption of storage of terminal dopamine by direct damage to the transporters associated with dopamine uptake into either synaptic terminals or vesicles. Indeed, MA, via a mechanism involving reactive oxygen species (ROS), can impair DAT function (Pogun et al. 1994; Berman et al. 1996).
Our in vitro studies using human neuronal cultures and the fluorescent probe JC-1 indicate that impaired mitochondrial function may represent a common pathway in MA + Tat-induced toxicity. Mitochondrial dysfunction has been implicated in an increasing number of neurodegenerative disorders (Beal 2000; Orth and Schapira 2001) and may be pathophysiologically relevant in HIV infection (Rustin 2001). Additionally, we showed that the combined toxicity of MA and Tat could be blocked using antioxidants with therapeutic potential. Ebselen is a selenium-containing organic compound that mimics the action of glutathione peroxidase (Wendel et al. 1997). Diosgenin is a plant-derived steroid (Accatino et al. 1998) and EUK 8 is a synthetic manganese complex that has previously been shown to protect against amyloid beta toxicity and stroke via antioxidative mechanisms (Bruce et al. 1996). Although mitochondria may be an important source of ROS under pathological conditions (Dykens 1997), the prevention of MA + Tat-induced reduction of Ψ by these antioxidants suggests that oxidative stress precipitates rather than results from mitochondrial dysfunction. That ROS can impair mitochondrial function is not unprecedented as several mitochondrial complexes can be inactivated when exposed to ROS (Zhang et al. 1990; Benzi et al. 1991; Dykens 1994). Future studies will be directed at elucidating the source of MA + Tat-generated ROS, and in evaluating the neuroprotective efficacy of these and other antioxidants and energy metabolism ‘enhancers’ in animals exposed to these two toxins.
A major concern in designing these studies was that hyperthermia, which typically occurs during exposure to MA and is necessary to generate large reductions in striatal dopamine (Bowyer et al. 1992; Ali et al. 1994; O'Callaghan and Miller 1994; Albers and Sonsalla 1995), could confound the interpretation of any observed interactions between Tat and MA. In an effort to circumvent this issue, we used the lowest concentration of MA that would generate a small reduction in striatal dopamine levels (5 mg/kg) without inducing hyperthermia. Thus, we conclude that the synergism between MA and Tat was due either to the concomitant activation of secondary pathways or physical interactions between the two compounds. The latter explanation, however, seems to be unlikely as administration of both Tat and MA on the same day failed to generate more than an additive loss of dopamine (W. F. Maragos, unpublished observations). In contrast, animals exposed to MA 24 h following striatal injections of Tat did develop hyperthermia. This effect seems to represent a specific interaction between the two toxins as animals injected with heat-inactivated FBS followed by MA did not develop hyperthermia. Although the mechanism responsible for the temperature increase is not known, the observed synergy between MA and Tat using a neuronal culture system, in which the temperature is strictly maintained, indicates that hyperthermia is not solely responsible for the occurence of neuronal injury.
Deficits in CSF levels of dopamine and its metabolites have been measured in HIV-1-infected patients (Larsson et al. 1991; Berger et al. 1994; di Rocco et al. 2000). Moreover, signs of parkinsonism including bradykinesia, postural instability and gait disturbances and sensitivity to parkinsonian-inducing agents, have been reported in patients infected with HIV-1 (For review see Berger and Nath 1997). Based on the present data, we would predict that patients infected with HIV-1 who abuse MA may be at greater risk of developing clinical signs of basal ganglia dysfunction. In support of this are the observations that HIV-infected drug abusers may develop a more severe form of encephalitis (Bell 1998) and, in some instances, an accelerated form of dementia associated with choreoathetosis (Nath et al. 2001). Although many of the patients in these studies were using MA, they were often simultaneously abusing other substances so a definitive relationship between HIV-1 infection and MA cannot be concluded. Future studies with well defined cohort populations should be able to address this important issue.
The authors thank Michael Harned and Laura Peters for their technical assistance. This work was supported in part by USPHS grants DA10115 (WAC), DA13144 (WFM) and NS39253 (AN).