Address correspondence and reprint requests to Donald M. Kuhn, Department of Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, 2125 Scott Hall, 540 E. Canfield, Detroit, MI 48201, USA. E-mail: firstname.lastname@example.org.
Methamphetamine causes persistent damage to dopamine nerve endings of the striatum. Repeated, intermittent treatment of mice with low doses of methamphetamine leads to the development of tolerance to its neurotoxic effects. The mechanisms underlying tolerance are not understood but clearly involve more than alterations in drug bioavailability or reductions in the hyperthermia caused by methamphetamine. Microglia have been implicated recently as mediators of methamphetamine-induced neurotoxicity. The purpose of the present studies was to determine if a tolerance regimen of methamphetamine would attenuate the microglial response to a neurotoxic challenge. Mice treated with a low-dose methamphetamine tolerance regimen showed minor reductions in striatal dopamine content and low levels of microglial activation. When the tolerance regimen preceded a neurotoxic challenge of methamphetamine, the depletion of dopamine normally seen was significantly attenuated. The microglial activation that occurs after a toxic methamphetamine challenge was blunted likewise. Despite the induction of tolerance against drug-induced toxicity and microglial activation, a neurotoxic challenge with methamphetamine still caused hyperthermia. These results suggest that tolerance to methamphetamine neurotoxicity is associated with attenuated microglial activation and they further dissociate its neurotoxicity from drug-induced hyperthermia.
The amphetamines are powerful stimulant drugs of abuse. Illicit use of these drugs, including methamphetamine (METH) and 3,4-methylenedioxymethamphetamine (MDMA, Ecstasy), has increased to the point that they have eclipsed cocaine and heroin abuse on a global scale (UN Office on Drugs and Crime report http://www.unodc.org/unodc/en/global_illicit_drug_trends.html). The medical, legal and societal problems associated with any rampant drug of abuse are compounded in the case of the amphetamines because many members of this pharmacological class cause persistent damage to the dopamine (DA) neuronal system. The short-term consequences of METH abuse are manifested as depletions of DA, inhibition of tyrosine hydroxylase, inactivation of the DA transporter (DAT) and reduction in function of the vesicle monoamine transporter (VMAT) (Gibb et al. 1997; Brown et al. 2000; Fleckenstein et al. 2000; Kita et al. 2003a). Additional complications emerge with longer-term use of the amphetamines including degeneration of fine, unmyelinated axons (Molliver et al. 1990) and neuronal apoptosis (Davidson et al. 2001; Cadet et al. 2003).
It has been proposed recently that microglia participate in the neurotoxicity associated with METH intoxication (Escubedo et al. 1998; Pubill et al. 2002; Guilarte et al. 2003; Pubill et al. 2003; LaVoie et al. 2004; Thomas et al. 2004a,b). Microglia are the primary antigen-presenting cells in the central nervous system. Once activated, these immune-like cells (Aloisi 2001; Streit 2002) can secrete a variety of factors such as reactive oxygen species, reactive nitrogen species, pro-inflammatory cytokines and prostaglandins, each of which can lead to neuronal damage (Kreutzberg 1996; Hanisch 2002). In the course of characterizing microglial involvement in METH neurotoxicity (Thomas et al. 2004b), we noted an attenuated microglial response in mice previously treated with drug, suggesting the possibility that tolerance to METH toxicity could be mediated at the level of microglia. The present studies confirm this possibility and show that tolerance does not develop to the hyperthermic effects of METH.
Materials and methods
(+) Methamphetamine hydrochloride, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), pentobarbital, horseradish peroxidase (HRP)-conjugated isolectin B4 (from Griffonia simplicifolia), 3,3′-diaminobenzidine (DAB), paraformaldehyde, Triton X-100, dopamine, methanol, EDTA, all buffers and HPLC reagents were purchased from Sigma-Aldrich (St Louis, MO, USA). CitriSolv and Permount were products of Fisher Scientific (Pittsburgh, PA, USA).
Female C57BL/6 mice (Harlan, Indianapolis, IN, USA) weighing 20–25 g at the time of experimentation were housed five per cage in small shoe-box cages in a light- and temperature-controlled room. Mice had free access to food and water. The Institutional Care and Use Committee of Wayne State University approved the animal care and experimental procedures. All procedures were also in compliance with the NIH Guide for the Care and Use of Laboratory Animals.
Pharmacological and physiological procedures
Mice were exposed to two different pre-treatment conditions, followed in both cases with the identical neurotoxic challenge regimen of METH. In the first case, mice were treated with a neurotoxic regimen of METH [four intraperitoneal (i.p.) injections of 5 mg/kg at 2-h intervals], which is manifested as a significant depletion of DA and robust microglial activation (Thomas et al. 2004b) and were subsequently challenged with the same METH neurotoxic regimen 7 days later, a time when the METH-induced microglial activation has returned to control levels (Thomas et al. 2004b). Controls for the drug pre-treatment and challenge conditions received i.p. injections of physiological saline on the same schedule used for METH. Injection volumes were 0.1 mL/10 g body weight. For ease of presentation, the pre-treatment/challenge conditions described above are referred to as follows: (i) control pre-treatment/control challenge (Sal/Sal); (ii) METH neurotoxic pre-treatment/control challenge (METHNT/Sal); (iii) control pre-treatment/METH challenge (Sal/METHNT); (iv) METH neurotoxic pre-treatment/METH challenge (METHNT/METHNT).
In the second design, mice were pre-treated with a tolerance regimen of METH (four i.p. injections of 2 mg/kg with a 2-h interval between each injection). This tolerance regimen was given three times (i.e. Monday, Wednesday and Friday) with a 48-h interval between each treatment. Mice receiving the tolerance regimen were subsequently challenged with the METH neurotoxic regimen 72 h later (i.e. on the following Monday). This lower dose of METH was selected for tolerance induction because it results in minimal microglial activation and has only minor effects on striatal DA content (Thomas et al. 2004b). These pre-treatment/challenge conditions are referred to as follows: (i) control pre-treatment/control challenge (Sal/Sal); (ii) METH tolerance/control challenge (METHTol/Sal); (iii) control pre-treatment/METH challenge (Sal/METHNT); (iv) METH tolerance/METH challenge (METHTol/METHNT).
Finally, in some experiments, mice were challenged with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP; four i.p. injections of 20 mg/kg with a 2-h interval between injections) 7 days after pre-treatment with the neurotoxic regimen of METH. These groups are referred to as METHNT/MPTP and controls for this group are Sal/MPTP. Body temperature was monitored by telemetry using IPTT-200 implantable temperature transponders from Bio Medic Data Systems Inc. (Seaford, DE, USA). Core body temperatures were recorded every 20 min non-invasively using the DAS-5001 console system from Bio Medic.
Lectin histochemical staining and stereology
Microglial activation was assessed by staining fixed brain sections with HRP-conjugated isolectin B4 (ILB4) as developed by Streit (Streit 1990) and as previously described in our studies with METH (Thomas et al. 2004b). All mice were killed 48 h after the last injection of the METH or MPTP neurotoxic challenge for microglial analyses. Mice were deeply anesthetized with pentobarbital (120 mg/kg) and perfused transcardially with ice-cold 4% paraformaldehyde. Brains were removed and stored overnight in fixative at 4°C. Sections of 50 μm thickness were cut through the striatum (+1.2 through −0.1 mm with respect to bregma). Sections were floated into phosphate-buffered saline (PBS) containing 0.3% H2O2 for 30 min, washed once in PBS + 0.1% Triton X-100, then incubated in fresh PBS + 0.1% Triton X-100 for an additional 30 min. Microglia were labeled with HRP-conjugated ILB4 (10 μg/mL in 0.1% Triton X-100) overnight at 4°C. Excess ILB4 was removed by three washes with PBS + 0.1% Triton X-100 (5 min each), followed by a single wash in PBS before exposure to DAB substrate (0.1 mg/mL) in PBS for 25 min. Following a single wash in PBS, sections were transferred to glass slides, air dried and dehydrated through a series of graded ethanol washes. Sections were incubated in Citrisolv for 5 min then cover-slipped under Permount. For all pharmacological studies presented below, brain sections from test drug-treated mice were processed simultaneously with controls and METH-treated mice to normalize staining among treatment groups. Microglial reactivity was viewed under the light microscope and the number of stained cells observed after various treatments was quantified by stereological analysis using NIH Image. Cell counts were sampled from a 1.2 mm2 area of the striatum by a person blinded to the treatment conditions. Cells were counted from three independent sections from all like-treated mice, bilaterally, generating an average count for each treated subject.
DA tissue content
Depletion of striatal DA after METH treatment is widely used as an index of METH-induced toxicity to DA nerve endings. DA depletion from striatum faithfully reflects other measures of DA nerve ending damage caused by METH, such as reduced tyrosine hydroxylase immunoreactivity or reduced ligand binding to the DAT. Striata were dissected from brain at the times indicated below after METH treatments and stored frozen at −80°C. Tissues were weighed and sonicated in 5 vol of 0.1M perchloric acid at 4°C. Insoluble protein was removed by centrifugation and catecholamines were adsorbed to alumina using standard protocols. After elution from alumina, the amount of DA in samples was determined by HPLC with electrochemical detection by comparison with a standard curve of authentic DA. Values for DA are reported as ng/mg tissue (wet weight) and are corrected for recovery.
Dose–response effects of METH on striatal DA content and microglial counts were tested for significance by anova. Individual treatment groups were compared with appropriate controls for DA and microglial counts with a one-way anova followed by Dunnett's multiple comparison test in GraphPad Prism 4 (GraphPad Software Inc., San Diego, CA, USA). Differences were considered significant if p < 0.05. The effects of drug treatments on core body temperature were analyzed by two-way anova.
The microglial response to a neurotoxic regimen of METH abates within 7 days (Thomas et al. 2004b), so mice were treated with a neurotoxic challenge of METH 7 days after a neurotoxic pre-treatment regimen. The results in Fig. 1 show that the neurotoxic METH challenge in naïve mice (Sal/METHNT) causes a significant reduction in striatal DA (79% reduction, p < 0.01, Dunnett's test). DA levels remain significantly depleted 9 days after the METH tolerance regimen (METHNT/Sal; 81% reduction, p < 0.01, Dunnett's test). When mice are challenged with the neurotoxic METH regimen 7 days after the same treatment (METHNT/METHNT), DA levels are not further depleted (18% of control) by comparison with mice treated with a neurotoxic regimen (Sal/METHNT) and remain substantially lower than controls (Sal/Sal). The effect of drug treatment on striatal DA was significant in all groups that received METH (p < 0.01, Dunnett's test, for all three groups relative to control).
Figure 2 presents data on the microglial response to different treatment conditions. Few activated microglia are observed in controls (Sal/Sal) in agreement with our previous results (Thomas et al. 2004b), whereas mice given a neurotoxic regimen of the drug (Sal/METHNT) show a robust microglial activation. Mice pre-treated with METH and then challenged with saline (METHNT/Sal) 7 days later show few activated microglia, confirming our previous data that microglia return to control levels within this time frame. When mice are challenged with the neurotoxic METH regimen after pre-treatment with the same regimen (METHNT/METHNT) and killed 48 h after challenge, there is little evidence of a microglial response. If the time between METH pre-treatment and challenge treatments is increased from 7 to 30 days, or if the challenge dose of METH is increased from 5 to 10 mg/kg, microglial activation is still not seen. These data are included in Fig. 2 (e and f, respectively) as well and indicate that challenge doses of METH that cause severe DA neurotoxicity and microglial activation do not provoke a subsequent microglial response after mice have been pre-treated with the same METH neurotoxic regimen. The drug effect on microglial activation was significantly different from control only in mice treated with the acute neurotoxic METH challenge (Sal/METHNT, p < 0.01, Dunnett's test). Finally, mice treated acutely with MPTP (Sal/MPTP) show a robust microglial activation in striatum as shown in Fig. 2. Treatment of mice with MPTP 7 days after a METH neurotoxic regimen (METHNT/MPTP) did not activate microglia above control levels, as shown in Fig. 2, in agreement with results observed in METHNT/METHNT-treated mice.
The METH pre-treatment regimen used above caused substantial neurotoxicity and created a floor-effect with regard to DA depletion. Therefore, we used a less severe, low-dose regimen to generate tolerance to METH neurotoxicity. The METH dose in these experiments (2 mg/kg, 4 ×, 2-h interval between injections) was selected from our previous dose–effect studies as one that caused minimal DA toxicity and microglial activation (Thomas et al. 2004b). Figure 3 shows that three treatments with 2 mg/kg (i.p., 4 ×, 2-h interval between injections) with a 48-h time period between each regimen, followed 72 h later with an acute challenge, resulted in the development of tolerance. A neurotoxic challenge with METH (Sal/METHNT) resulted in a significant (i.e. 79%) DA depletion by comparison with controls (Sal/Sal). The lower dose of 2 mg/kg METH, when given on three separate occasions (METHTol/Sal), did cause a mild depletion of DA (72% of control). However, the effects of an acute METH challenge were significantly blunted in mice previously treated with the low-dose tolerance regimen (METHTol/METHNT). METH typically reduces striatal DA levels by about 80% without pre-treatment and this response is reduced to 17% depletion by comparison with the low-dose METH tolerance regimen.
The microglial response to the lower-dose METH tolerance regimen is shown in Fig. 4. In agreement with our previous report (Thomas et al. 2004b), acute METH (Sal/METHNT) provokes a substantial activation of striatal microglia by comparison with controls (Sal/Sal). Mice that received the tolerance regimen without an acute METH challenge (METHTol/Sal) showed very little microglial activation 9 days after drug treatment. Figure 4 shows that tolerance has also developed to the effects of METH on microglial activation. The extent of microglial activation is significantly lower in mice that received the tolerance regimen before the neurotoxic challenge (METHTol/METHNT), in agreement with neurochemical findings (Fig. 3 above). The only groups that differed from control with respect to microglial activation were the Sal/METHNT and METHTol/METHNT groups (p < 0.01 for each, Dunnett's test).
The effect of the two different METH pre-treatment regimens on the body temperature response to METH challenge is shown in Fig. 5. Mice pre-treated with the neurotoxic or the tolerance regimen (METHNT/METHNT and METHTol/METHNT groups, respectively) each developed hyperthermia to a greater extent than mice given only a neurotoxic METH challenge (Sal/METHNT). Core temperatures of all groups receiving the neurotoxic challenge of METH, regardless of the pre-treatment regimen used, were significantly elevated over those of controls (Sal/Sal; p < 0.0001 by two-way anova). To quantify the differences in core temperatures among the four treatment groups throughout the 9-h time period when temperatures were monitored, the area under each curve in Fig. 5 was calculated against a baseline of 36.5°C (body temperatures in all groups remained above this level). The areas under the curves (in arbitrary units) for each treatment condition were Sal/Sal (360), Sal/METHNT (573), METHNT/METHNT (831) and METHTol/METHNT (897). Taken together, these results indicate that tolerance does not develop to the hyperthermic effects of METH. It should also be noted that the dose of METH used presently (5 mg/kg) causes a mild hyperthermia by comparion with doses of 10 mg/kg or higher (Bowyer et al. 1992, 1994; Miller and O'Callaghan 1994; O'Callaghan and Miller 1994).
Chronic amphetamine use is associated with the development of pharmacological tolerance and behavioral sensitization, two factors that are thought to sustain the addictive properties of these drugs. It is well known that tolerance develops to the neurotoxicity of METH (Gygi et al. 1996; Stephans and Yamamoto 1996; Riddle et al. 2002; Johnson-Davis et al. 2003; Johnson-Davis et al. 2004) and cross-tolerance is observed between the neurotoxic effects of METH and MPTP (Sziraki et al. 1994; Kita et al. 2003b). Tolerance also develops to MPTP-induced astrogliosis (O'Callaghan et al. 1990). Johnson-Davis et al. (2004) showed recently that the effects of a neurotoxic regimen of METH on VMAT trafficking and function developed tolerance and these investigators ruled out adaptive changes in METH pharmacokinetics and transport to the brain and reductions in drug-induced hyperthermia as mediators of this effect, at least in rats (Johnson-Davis et al. 2004). Additional studies are required to determine if altered METH pharmacokinetics would account for tolerance in mouse models. Beyond this, the mechanisms that mediate tolerance to METH neurotoxicity are not known. In view of emerging data implicating microglia as participants in this process, we decided to analyze the microglial response under conditions where tolerance to the neurotoxic effects of METH has developed.
Two different treatments were used presently to study the microglial response to METH tolerance. Previous studies from our lab showed that microglial activation was maximal 48 h after a neurotoxic regimen of METH and returned to control levels after 7 days (Thomas et al. 2004b). Therefore, mice were treated with a neurotoxic challenge of METH 7 days after pre-treatment with the same neurotoxic regimen and microglial activation was not observed. If the dose of the second METH challenge was increased from 5 mg/kg to 10 mg/kg, or if the time between METH pre-treatment and challenge treatment was lengthened to 30 days, mice were still unable to mount a microglial response. The neurotoxic pre-treatment regimen used in these experiments caused severe, persistent DA depletion, so the failure to see further reductions in DA after the second METH challenge may reflect a floor effect more so than a development of tolerance. Therefore, a lower-dose METH tolerance regimen was tested with the intent of minimizing DA depletion and microglial activation. Treatment of mice with 2 mg/kg METH for 3 days, with a 48-h interval between each regimen, followed 72 h later with an acute neurotoxic challenge, revealed that tolerance to drug neurotoxicity and microglial activation had developed. The two pre-treatment regimens used differ significantly in that the neurotoxic regimen caused extensive microglial activation, which was allowed to subside prior to the neurotoxic challenge with METH, whereas the tolerance regimen did not result in significant microglial activation. Therefore, the inability of a METH neurotoxic challenge to provoke microglial activation after a tolerance regimen is not determined by the prior activation status of the microglia. Assuming that microglial activation precedes and contributes to METH-induced nerve ending damage, it follows that microglial tolerance could result in tolerance to METH neurotoxicity.
Bowyer and colleagues were the first to make the important observation that METH intoxication results in microglial activation (Bowyer et al. 1994), and these investigators suggested that microglial activation occurred in response to neuronal damage. Therefore, it could be argued that microglial activation seen after acute METH administration does not contribute to its neurotoxicity. The same rationale could be applied to tolerance induction and would assert that microglia do not show activation because nerve endings have developed tolerance to the damaging effects of METH. However, a number of emerging results are consistent with the possibility that microglia participate in acute METH toxicity as well as in the development of tolerance to it, including the following: (i) microglia rapidly develop tolerance to various provocative stimuli (Alves-Rosa et al. 2002; Nadeau and Rivest 2002; Ajmone-Cat et al. 2003; Nakajima et al. 2003) so it is certainly possible that microglia made tolerant to METH would produce lesser amounts of those reactive species that are thought to mediate drug-induced neuronal damage (e.g. reactive oxygen species or reactive nitrogen species); (ii) METH intoxication increases the expression of numerous microglia-associated genes within hours of drug administration, well before toxicity is seen (Thomas et al. 2004b); (iii) microglial activation precedes the emergence of nerve ending toxicity (LaVoie et al. 2004; Thomas et al. 2004b) and develops along the same time course with it (Baucum et al. 2004); (iv) microglial activation caused by the amphetamines is restricted to those drugs known to cause neurotoxicity (e.g. METH, MDMA, parachloramphetamine), minimizing a role for pharmacological effects such as hyperthermia and hyperactivity as causes (Thomas et al. 2004c). It is probably not possible to classify the participation of microglia in METH neurotoxicity as either direct (i.e. causal) or indirect (i.e. reactive to damage). Clearly, considerable cross-talk is exchanged between distressed neurons and microglia to fuel the very complex process of neuronal damage (Bruce-Keller 1999; Polazzi and Contestabile 2002; Kerschensteiner et al. 2003). Therefore, we conclude that microglia participate in the acute phase of METH neurotoxicity and that an attenuation in microglial activation after repeated, intermittent METH administration mediates the adaptive tolerance to its neurotoxic effects.
The development of tolerance to METH neurotoxicity and microglial activation could be viewed as a neuroprotective response in view of the fact that tolerant animals show much less neurotoxicity after METH intoxication than mice challenged acutely with the same treatment regimen. However, the ability of prior METH exposure to arrest microglial activation (for up to 30 days in mice) could have potentially dangerous effects on the central nervous system that are not readily apparent. For example, lesions of the brain, stroke or infection are associated with microglial activation (Kreutzberg 1996; Moore and Thanos 1996; Stoll and Jander 1999). Under these conditions, microglia are thought to serve neuroprotective roles by removing damaged cell fragments or invading pathogens and by secreting various growth factors and cytokines that limit further damage or even participate in neuronal recovery (Hanisch 2002; Streit 2002). The signaling pathways within microglia that direct their responses toward neuronal protection or damage are not fully understood (Elward and Gasque 2003; Kerschensteiner et al. 2003). However, it seems clear that METH intoxication creates conditions within the brain that are unlike stroke or infection where the protective, immune-like functions of microglia are evoked. Prior, intermittent exposure to subneurotoxic doses of METH to an extent that provokes tolerance could hamper the adaptive immune response of the central nervous system (Nguyen et al. 2002) and worsen the consequences associated with subsequent neuronal damage or disease. This possibility is currently under investigation in our laboratory.
The present finding that tolerance does not develop to the hyperthermic effects of METH, regardless of the drug pre-treatment condition used (i.e. neurotoxic or tolerance regimen), agrees well with other studies (Johnson-Davis et al. 2003, 2004) and provides an additional degree of dissociation of this response from drug-induced neurotoxicity. The hyperthermic response to METH challenge observed presently was rather mild (Fig. 5) and could be explained by the use of a lower dose (i.e. 5 mg/kg per injection × 4). Higher doses (e.g. 10 mg/kg per injection × 4) and elevated ambient temperature enhance METH neurotoxicity (Miller and O'Callaghan 2003) and it is well known that the extent of DA nerve-ending damage caused by METH is highly correlated with the degree of drug-induced hyperthermia (Bowyer et al. 1992, 1994; Miller and O'Callaghan 1994; O'Callaghan and Miller 1994). The observation that METH elevates core body temperature significantly in pre-treated mice, without provoking microglial activation, further supports our finding that hyperthermia does not appear to be the signal that prompts microglial activation after METH intoxication (Thomas et al. 2004b). Hyperthermia is a clinically significant side-effect of METH ingestion (Urbina and Jones 2004). The finding that tolerance does not develop to METH-induced hyperthermia, even after it has developed to its neurotoxic effects, reinforces the dangers associated with this drug of abuse.
This research was supported by the National Institute on Drug Abuse grants DA10756 and DA14692 and by a VA Merit Award.