Contrasting dose-dependent effects of acute intravenous methamphetamine on lateral hypothalamic extracellular glucose dynamics in male and female rats

Abstract

Glucose is the brain’s primary energetic resource. The brain’s use of glucose is dynamic, balancing delivery from the neurovasculature with local metabolism. Although glucose metabolism is known to differ in humans with and without methamphetamine use disorder (MUD), it is unknown how central glucose regulation changes with acute methamphetamine experience. Here, we determined how intravenous methamphetamine regulates extracellular glucose levels in a brain region implicated in MUD-like behavior, the lateral hypothalamus (LH). We measured extracellular LH glucose in awake adult male and female drug-naive Wistar rats using enzyme-linked amperometric glucose biosensors. Changes in LH glucose were monitored during a single session after: 1) natural nondrug stimuli (novel object presentation and a tail-touch), 2) increasing cumulative doses of intravenous methamphetamine (0.025, 0.05, 0.1, and 0.2 mg/kg), and 3) an injection of 60 mg of glucose. We found second-scale fluctuations in LH glucose in response to natural stimuli that differed by both stimulus type and sex. Although rapid, second-scale changes in LH glucose during methamphetamine injections were variable, slow, minute-scale changes following most injections were robust and resulted in a reduction in LH glucose levels. Dose and sex differences at this timescale indicated that female rats may be more sensitive to the impact of methamphetamine on central glucose regulation. These findings suggest that the effects of MUD on healthy brain function may be linked to how methamphetamine alters extracellular glucose regulation in the LH and point to possible mechanisms by which methamphetamine influences central glucose metabolism more broadly.

NEW & NOTEWORTHY Enzyme-linked glucose biosensors were used to monitor lateral hypothalamic (LH) extracellular fluctuations during nondrug stimuli and intravenous methamphetamine injections in drug-naive awake male and female rats. Second-scale glucose changes occurred after nondrug stimuli, differing by modality and sex. Robust minute-scale decreases followed most methamphetamine injections. Sex differences at the minute-scale indicate female central glucose regulation is more sensitive to methamphetamine effects. We discuss likely mechanisms underlying these fluctuations, and their implications in methamphetamine use disorder.

INTRODUCTION

Methamphetamine use disorder (MUD) in people remains a considerable and persistent public health concern, with 1 million people aged 12 and older with the disorder in the United States in 2019 (1). In addition to long-term behavioral symptoms associated with MUD, methamphetamine use in both humans and animal models is also associated with significant neuroadaptations impacting catecholamine neurotransmission (2–5) as well as alterations in central glucose metabolism (6–8). For example, using positron emission tomography (PET), Volkow and coworkers (6, 8) found significantly lower glucose metabolism in recently methamphetamine-abstinent humans compared with controls in the striatum and thalamus, and that these lower levels of glucose metabolism partially recovered after protracted periods of abstinence. These differences between methamphetamine-experienced subjects and controls are important because changes in glucose metabolism may reflect far-reaching changes in regional brain function including neuronal activity, glial activation, and hemodynamics (9–16). Most of our current knowledge of methamphetamine-induced changes in glucose metabolism relies on human subjects with extensive experience with the drug. Comparatively, our understanding of acute effects of methamphetamine on central glucose dynamics is lacking, due in part to experimental and ethical constraints associated with human-subjects research. This is important because the initial, acute effects of methamphetamine may have a larger role in the etiology of MUD than currently assumed. Moreover, PET indices of glucose metabolism are primarily limited to single time points (15), and thus do not provide measures of dynamic changes in brain glucose. Preclinical studies using enzyme-linked biosensors and constant potential amperometry, which can detect extracellular concentrations of glucose, circumvent such limitations to provide measurements of dynamic changes in central glucose in response to acute methamphetamine.

Central extracellular glucose levels depend on two opposing factors: its metabolic use by brain cells (i.e., neurons and astrocytes), and its intrabrain delivery across the blood-brain barrier via gradient-dependent facilitated diffusion via the glucose transporter protein-1 (GLUT-1) (17–20). Fluctuations in extracellular glucose in the context of several drugs of abuse have been well characterized using biosensors and amperometry, albeit in a limited number of brain structures and exclusively in male rats. In the nucleus accumbens (NAc), a mesolimbic structure critical for addiction, the psychostimulants cocaine and 3,4-methylenedioxypyrovalerone (MDPV) induce distinct responses to repeated intravenous injections. Specifically, cocaine precipitates rapid second-scale and minute-scale tonic elevations in extracellular glucose. These effects indicate rapid drug-induced increases in glucose due to neural activation resulting from the interoceptive sensory effects of cocaine, and slow tonic effects resulting from increasing systemic vasoconstriction (21). Conversely, the highly dopamine-specific (22) synthetic cathinone MDPV only causes minute-scale postinjection decreases in glucose levels that are relatively stable across multiple injections (23), conspicuously lacking rapid increases characteristic of interoceptive stimulation. In the substantia nigra pars reticulata (SNpr), a part of the neighboring but functionally and anatomically distinct striatonigral circuit, cocaine causes both rapid second-scale and sustained minute-scale decreases in glucose concentrations, likely because of the differing neurophysiological properties of the neurons in the SNpr and NAc (24). Together, these findings highlight that diverse drugs impact extracellular glucose levels differently in distinct brain regions.

Although most extant research into the acute effects of psychostimulants on glucose dynamics has focused on the mesolimbic circuit and nucleus accumbens, an evolving view is that multiple circuits and central systems that interface with the mesolimbic system are also involved in regulating the physiological impact of substance use disorders, including MUD (25). One such area, the lateral hypothalamus (LH), has been traditionally studied in the diverse context of homeostatic energy regulation, nutrient sensing, responses to stress, sleep-wakefulness, substance use disorders, and sexual dimorphisms in these processes (26–32). Although the LH was one of the first brain areas identified in regulating reward-seeking behavior (33), there has been renewed interest in this area as a locus regulating addiction-related processes because of its interconnectivity with the mesolimbic system (34). As well, since sex differences mediate both the effects of methamphetamine and the neurobehavioral function of the LH, a systematic examination of sex as a factor in LH extracellular glucose regulation is needed (29, 35, 36).

Moreover, the LH is a diverse region that contains several neuronal phenotypes that are directly glucose sensitive—in other words, LH neurons that change their electrophysiological activity in accordance to extracellular glucose levels (37). This presents the possibility that, in addition to traditional neurotransmitter-based mechanisms of methamphetamine pharmacodynamics, alterations in extracellular glucose levels in response to methamphetamine may themselves influence neuronal activity of the LH at a behaviorally relevant timescale. This is supported in part for other brain regions by reports of changes in cerebral blood flow induced by acute methamphetamine. For example, acute methamphetamine has been found to induce regional hypoxia in the dorsal striatum of male rats (38), and a sustained decline in cerebral blood flow (CBF) in the cerebral cortex because of vasoconstriction in anesthetized mice (39). However, these findings remain controversial, as other studies have shown contrasting effects (40, 41) of acute methamphetamine or amphetamine possibly due to differences in experimental methodology. Consequently, the pattern of extracellular glucose fluctuations in the LH during acute methamphetamine remains unexplored and largely unknown.

The objectives of this study were to determine how extracellular glucose concentrations are changed in the LH of freely moving rats using enzyme-based glucose biosensors and high-speed amperometry in response to several types of stimuli. These stimuli included both naturalistic nondrug stimuli, cumulative doses of acute intravenous methamphetamine, and direct infusions of glucose into the blood. Although similar studies determining drug-effects have delivered drugs as boluses with long intrainjection intervals so that currents return to baseline levels (e.g., Refs. 21 and 24), cumulative dosing is a commonly used within-session approach to assess effects of drugs like methamphetamine over a range of doses (42–45). In addition, cumulative dosing is well suited for recording in cases where a drug has a long half-life [∼55 min in both male and female rats (46)]. We chose doses between 0.025 mg/kg and 0.2 mg/kg as these represent the range that is readily self-administered in both male and female rats (47–52). As part of the study design, we systematically determined if there were any sex differences in lateral hypothalamic extracellular glucose dynamics in response to these stimuli.

METHODS

Animals and Surgeries

Data from 17 male and 11 female Wistar rats (Envigo), aged matched (∼2.5–3 mo) weighing 344.71 ± 7.63 g (male) and 262.73 ± 6.89 g (female) at the time of testing were used in this study. Female rats were intact and freely cycling; as we had no a priori hypotheses on an interaction between ovarian cycle phase and the experimental variables, cycle phase was not controlled. Rats were group housed upon arrival in a climate-controlled animal colony in filtered air cages and maintained on a 12-12 light-dark cycle (lights on at 1500). Food and water were available ad libitum. Once rats underwent surgical procedures detailed in the next paragraph, they were housed singly to prevent damage to their implants but were housed within sight and sound of other rats in the colony. All procedures were approved by the University of Nebraska-Lincoln Animal Care and Use Committee and complied with the Guide for the Care and Use of Laboratory Animals (53).

No more than 24 h before the surgical procedure, rats were given carprofen (Bio-serv, 5 mg/kg, oral tablet, MD150-2) for prophylactic analgesia. Surgical procedures were performed under general anesthesia, where rats were first induced with a mixture of ketamine HCl (33 mg/kg) and xylazine (10 mg/kg) (54) injected intraperitoneally and then maintained on isoflurane (∼1%, SomnoSuite, Kent Scientific, VetOne). During the procedure, rats were prepared with a BASi cannula (Bioanalytical Systems, Inc.) for future implantations of the biosensor into the LH on the recording day. A 1.0-mm diameter hole was drilled into the skull, and the guide cannula was lowered (∼0.05 mm/s) using the arms of a stereotaxic frame (Stoelting Co.). The target coordinates for the cannula were: anteroposterior (AP) −2.9 mm, mediolateral (ML) +1.4 mm, dorsoventral (DV) −7.3 to 7.6 mm, according to the stereotaxic atlas of Paxinos and Watson (55). The skull was pretreated with acrylic primer, and the guide cannula was secured via an acrylic headmount anchored by three stainless steel bone screws. In addition, the headmount contained a pedestal that served to secure the preamplifier during the biosensor recording.

During the same surgical procedure, rats were also implanted with a chronic jugular catheter constructed out of polyurethane tubing (BTPU-040, Instech laboratories), which ran subcutaneously to the headmount and terminated in a handmade injection port secured to the same headmount. Postprocedure analgesia was sustained immediately at the end of the procedure with carprofen (5 mg/kg sc) and maintained for 2 days postprocedure (5 mg/kg, oral). Rats were allowed a minimum of 5 days postoperative recovery. During this time, jugular catheters were flushed daily with 0.15–0.2 mL of heparinized sterile saline (25 U/mL) containing gentamicin (0.1 mg/mL) to maintain patency.

Fixed-Potential Amperometry with Enzyme-Based Electrochemical Sensors

For this study, we used commercially available glucose oxidase-based biosensors (glucose sensors; Pinnacle Technology, Inc.) coupled with fixed-potential amperometry. This approach has been used and validated in multiple studies (e.g., see Refs. 21, 23, 24, and 56–58). We also employed glucose oxidase-free null sensors identically constructed to glucose oxidase biosensors to reduce the contribution of various chemical and physical interferents to better reveal dynamic fluctuations in extracellular glucose. Glucose and null sensors were prepared from Pt-IR wire of 180 µL diameter, with an active area of ∼1 mm on its distal tip. The electrode in these sensors was incorporated with an integrated Ag/AgCl reference electrode. On the active surface of the biosensor, glucose oxidase converts glucose to glucono-1,5-lactone and hydrogen peroxide (H₂O₂). The H₂O₂ was detected as an amperometric oxidation current generated by a +0.6 V applied potential (59). The contribution of ascorbic acid to the measured current was competitively reduced by colocalizing ascorbic acid oxidase enzymes on the active surface of the sensor. This enzyme converts ascorbic acid to nonelectroactive dehydroascorbate and water. In addition, a negatively charged polymer layer comprised of Nafion excluded endogenous anionic compounds (59).

In Vitro Calibration

We calibrated both types of sensors immediately before and after each in vivo experiment, similar to other published reports (e.g., see Refs. 21, 24, 58, and 60). Briefly, sensors were allowed to equilibrate to 100 mM phosphate buffered saline (PBS; pH 7.4) for ∼20 min before incrementally increasing the concentration of glucose (Sigma-Aldrich) from 0 to 0.5, 1.0, and 1.5 mM. We then added a single aliquot of ascorbic acid (250 µM; Sigma-Aldrich). To calculate the change in current after these additions, we averaged the current reported by the sensor for 30 s and subtracted this value from the preceding baseline current averaged over 15 s. Calibrations for the in vivo experiments were conducted at room temperature (∼23°C) immediately before and after each in vivo experiment. A total of two sensors were removed from the postcalibration average due to unusually high ascorbate sensitivity or glucose sensitivity, indicating damage during the explant procedure. As previously established, these sensors have increased sensitivity at body temperature (∼37°C) (61). Therefore, a separate in vitro experiment determined the difference in sensor sensitivity to the same concentration of glucose between room and body temperature. Consistent with previously established values (61), our in vitro test revealed that there was a 59.18 ± 0.71% (n = 2) decrease in sensitivity at room temperature compared with body temperature.

In Vivo Recording Protocol

In vivo electrochemical procedures occurred during the day (0900–1800) that aligned with the start of the experiment to 6 h into the rat’s dark cycle. Experiments were conducted in a round plexiglass chamber (30 cm diameter × 30 cm height).

Each rat was recorded during one acute session with either a glucose or null sensor. At the beginning of each experimental session, rats were brought from the animal colony and minimally anesthetized (∼3 min) with isoflurane only (∼3.5%, SomnoSuite, VetOne) and a precalibrated sensor (either glucose, 7011-Glucose or null, 7011-Null, Pinnacle Technology Inc.) was inserted into the brain through the implanted guide cannula. The sensor was then connected to a potentiostat (8401-HR; Pinnacle Technology Inc.) via a preamplifier (8407-BIO; Pinnacle Technology Inc.) attached to an electrical swivel (8401; Pinnacle Technology Inc.). The preamplifier was also secured to the pedestal on the headmount at this time. The injection port on the headmount was connected to polyethylene-polyvinyl chloride (BTCOEX-22; Instech Laboratories) catheter extensions that were connected to a syringe mounted in a programmable syringe pump (GenieTouch; Kent Scientific). This allowed for stress-free delivery of saline and methamphetamine from outside the cage, thus minimizing the possible detection of the intravenous injection by the rat. Testing began ∼120 min after the insertion of the sensor when the contribution of the non-Faradaic current at the start of the recording was reduced and the baseline currents were relatively stabilized (Fig. 1A).

Figure 1.

Methods. A timeline of experimental events on the day of recording (A). Locations of male (blue lines), female (gold lines) rats’ glucose biosensors (B) modified from Ref. 55 with permission. Null sensor locations in male or female rats shown as gray lines. Individual in vitro currents reported by the glucose biosensors used for male (blue circles) and female (gold circles) rats’ recordings, or those reported by null sensors (gray squares) during pre- and postcalibration validation tests (C). Mean currents reported by male and female rats’ glucose biosensors shown as blue or gold lines, respectively. Individual recordings of female glucose (gold lines) or female and male rats’ null sensors (gray lines) during novel object presentation (D, left). Means ± SE currents reported by female glucose (gold circles) and female and male null sensors (gray circles) during the same stimulus (D, middle). Means ± SE estimated change in glucose concentration from baseline (0 µM) during novel object presentation after minimizing the contribution of potential interferents to the glucose electrochemical current, individual differences in sensor sensitivity, differences in sensitivity between 23°C and 37°C (D, right).

Nondrug Stimuli

Each rat was exposed to three nondrug stimuli. These nondrug tests were important for assessing in vivo sensor performance, establishing whether LH glucose concentration ([glucose]) changes were associated with nondrug sensory stimuli of different modalities, and determining if the procedure of the intravenous injection induced any changes in LH glucose levels.

First, rats were presented with a novel object (a small glass vial). This object was manually introduced and removed from the chamber 1 min later. Approximately 15 min later, rats were exposed to a 3-min tail-touch administered by the application of a wooden clothespin on its tail approximately one-third to one-half from the distal tip of the tail. As the wooden clothespin has a hole about the same size as the tail, this stimulus served as an arousing sensory stimulus that did no harm to the rat. Normally, this stimulus evoked a behavioral response about a minute after its application, including chewing the clothespin and moderate levels of locomotion. These noncontingent stimuli were administered by an experimenter who was electrically grounded. At least 20 min later, rats were infused with one or two infusions of intravenous saline (250 µL over 14.7 s).

Cumulative Doses of Methamphetamine

We examined changes in glucose levels (Δ[glucose]) induced by four cumulative doses of methamphetamine within the same recording session. (+)-(S)-Methamphetamine HCl (a gift from the National Institutes on Drug Abuse Drug Supply Program) was prepared as a stock solution of 0.05 mg/mL in sterile saline and injected at cumulative doses of 0.025 mg/kg, 0.05 mg/kg, 0.1 mg/kg, and 0.2 mg/kg, 55 min apart from each other. The doses were delivered by maintaining the same rate of infusion per dose (17 µL/s) and changing the duration of infusion (62). The doses delivered to achieve cumulative dosing (0.025 mg/kg, 0.025 mg/kg, 0.05 mg/kg, 0.1 mg/kg) corresponded to average infusion times of 9.2 ± 0.28 s, 9.2 ± 0.28 s, 19.0 ± 0.74 s, and 36.8 ± 1.1 s, respectively. Doses reported correspond to the weight of the methamphetamine salt.

Injection of 60 mg Glucose

To further test the sensor and to assess how rapid changes in blood glucose alter its levels in the LH, we administered a 60 mg dose of glucose (60 mg in 0.6 mL of sterile saline delivered over 52.9 s). The dose of 60 mg was chosen because it has been estimated to raise blood glucose levels to approximately three times its normal level in males, increasing it to 12 mM given an average blood volume of ∼30 mL (63), and the response to this has been measured by the same approach in the nucleus accumbens (60). To eliminate any possible technical issues or disturbing the rat during this stage of the recording, this glucose injection was preceded by an injection of methamphetamine (0.015 mg) due to the dead volume of the infusion line (200 µL), injection port (∼70 µL), and jugular catheter (female 24 µL, male 30 µL). Thus, from the start, the injection period of methamphetamine and glucose corresponded to 0 s–17.6 s and 17.6 s–52.9 s, respectively.

Postrecording

At the end of each recording session, the rat was removed from the chamber, gently restrained with a hand towel, and disconnected from the preamplifier and the infusion line. Then the rat was briefly anesthetized with isoflurane, and the sensor was explanted. The sensor then immediately underwent postrecording calibration.

Rats were either immediately processed for histological procedures or underwent the procedure the next day. Briefly, rats were injected with Fatal-Plus (sodium pentobarbital, 390 mg/kg ip) and transcardially perfused with room temperature 100 mM PBS (pH 7.4) followed by 10% formalin (Sigma Aldrich). Brains were stored in formalin for at least 24 h at 4°C and then transferred to 30% sucrose in 100 mM PBS until they equilibrated. The brain was sectioned to 35 µm on a sliding microtome (American Optical 860), and the location of the sensors were verified using the stereotaxic atlas of Paxinos and Watson (Fig. 1B; 55).

Data Analysis

We sampled electrochemical data at 1 Hz as current (nA) for 1 s using the Serenia software utility (Pinnacle Technology, Inc.). While recording, events were time synchronized to the recording by an experimenter manually marking the relevant timepoint. After recording, data were initially processed using a custom Python script (64). This script parsed the recording into individual trials around user annotated events. Events analyzed at a rapid timescale had a window of up to 30 s before and 180–300 s after an experimental timepoint; slow timescale events were analyzed with a window of up to 5 min before and 50 min after the relevant event. This script also streamlined some initial analyses, including identification of individual data points falling outside of 1.5–3 times the interquartile range as possible recording artifacts, removal of these points, and replacing these with values via linear interpolation. The script then transformed the raw absolute data into relative changes by taking the baseline value before each event (5 s for rapid and 30 s for slow timescale analysis) as 0 nA. The output of this script for each trial at each step (original, interpolated, baseline transformed) was then carefully examined by three experimenters to confirm the appropriateness of the automated analyses, establish the presence of recording artifacts, and determine the overall reliability of the signal. The trial data were considered validated once all three raters reached a consensus to include these data for further analysis.

Subsequent analyses occurred in several stages. First, data for rapid timescale analysis was binned to 2-s intervals, whereas data for slow timescale analysis was binned to 30-s intervals. Then, to determine whether physical contributors or chemical interferents other than glucose (e.g., brain temperature fluctuations, catecholamines; for review see Ref. 61) were different between males and females during our recordings, we statistically assessed male and female null currents for all stimuli and events using a mixed models analysis. We found that there were no significant Fixed effects of Sex or a Sex × Time interaction for any stimuli or event at the rapid timescale, and thus for following analyses the null currents at this timescale were grouped together. We found that for some stimuli at the slow timescale and the basal level analyses detailed later, the null currents were statistically different between the males and females, and thus for the next step these were kept separate (Supplemental Fig. S1; all Supplemental Figures are available at https://doi.org/10.6084/m9.figshare.20060795). Next, to reduce the influence of these nonglucose contributors to our overall signal, we subtracted the mean changes generated by the null sensors from the changes reported by each of the glucose biosensor recordings for each stimulus, generating a current differential.

Preliminary analysis of data from the glucose sensors revealed that while the mean sensitivity of the active glucose biosensors used for the males and females were virtually identical, there was some variability within the sensitivities of individual sensors used in the experiment (Fig. 1C). We thus transformed the current differentials (nA) into glucose concentration (µM) by considering each sensor’s sensitivity toward known concentrations of glucose assessed during the prerecording in vitro calibration, while accounting for the difference in sensitivity between room temperature (∼23°C) and body temperature (∼37°C; Fig. 1D).

Differences between relative changes in mean glucose levels between males and females over time immediately after a stimulus event were analyzed using a mixed models analysis. If a significant fixed effect of sex, or significant interaction was found, a Fisher post hoc test was used to determine the time intervals wherein groups differed. If sex differences were not observed, changes in glucose levels were subsequently analyzed as a single group using a one-way mixed models analysis, followed by a Fisher post hoc comparing changes from the baseline levels preceding the stimulus.

The total effects of cumulative doses of methamphetamine over the 50-min postinjection time interval were expressed as net area under the curve (AUC) (µM × min), accounting for areas that went below baseline. Differences between the basal levels of glucose in the LH before each stimulus event during the recording were also assessed by transforming the currents reported by the glucose and null sensors during the 5 s baseline period preceding each stimulus into glucose concentration, as detailed earlier. Sex differences in these analyses were assessed with mixed models analysis, followed by a Fisher post hoc test if there was a significant fixed effect or interaction. As our age-matched male rats were larger than female rats, we also examined whether weight within each group correlated with AUC using linear regression.

Statistics were obtained using GraphPad Prism 8/9. For mixed model analyses, we used a compound symmetry covariance matrix, fitted using restricted maximum likelihood (REML). For clarity, only significant effects and interactions are statistically detailed. n values reported in figures corresponds to the number of individual glucose sensor recordings included in the analysis. The level of significance was set to α = 0.05.

RESULTS

Dynamic Fluctuations in LH [Glucose] in Response to Sensory Stimuli Differ by Modality and Sex

At the beginning of each experimental recording, we examined the responses to two nondrug stimuli. The first stimulus was the presentation and removal of a novel object to the experimental cage for 1 min. During both onset and offset of this stimulus, there were no sex differences in the [glucose] response (Fig. 2, A and B). Thus, considered together, the placement of the novel object into the cage was associated with a rapid but short-acting increase in LH [glucose] from baseline levels (Fig. 2C, one-way mixed models analysis F_30,540= 2.42, P < 0.0001). This rise became significant within the first 4.5 s of the stimulus, and reached a peak of 39.81 ± 9.25 µM, but soon began to return to baseline levels at ∼40.5 s. When the object was removed 1 min later, there was an immediate decrease in LH glucose levels that became significant ∼28.5 s after the event and reaching a nadir of −29.74 ± 12.21 µM from baseline (Fig. 2D, one-way mixed models analysis F_90,1530= 2.45). This decrease persisted significantly for ∼113 s after removal, before returning to baseline levels at ∼160 s.

Figure 2.

Rapid changes in lateral hypothalamus (LH) [glucose] induced by a brief exposure to a novel object and a tail-touch. All data are shown as means ± SE. Top graphs show changes in glucose concentrations between male (blue circles) and female (gold circles) rats. While sex differences were not found for novel object (A and B) and tail-touch start (E), compared with male rats, female rats showed decreased [glucose] after the end of tail-touch (G). Considering males and females together, both novel object presentation and the start of tail-touch were associated with rapid rises in [glucose] (C and F), within seconds of the stimulus presentation, which was significantly different from baseline. As a group, the end of the novel object presentation (D) was correlated with LH levels that transiently fell below baseline levels for 28.5–112.5 s after the event. Time intervals where male and female groups are different by Fisher post hoc test are shown as a blue bar on the X-axis (P < 0.05). Likewise, times where the males and females together are different from baseline (P < 0.05) by post hoc test are shown as red bar. Periods of stimulus presentation are marked as the shaded gray area.

The second stimulus was the application of a mild tail-touch for 3 min. Similar to the novel object, there were no sex differences during the onset of this stimulus (Fig. 2E), and together (Fig. 2F) the beginning of the tail-touch was associated with a rapid rise in LH [glucose] that became significant 2.5 s after onset and reached a peak at 47.70 ± 18.27 µM at ∼10.5 s (F_90,1530= 1.63, P = 0.0002). Unlike the onset of the tail-touch stimulus, we detected a sex difference in LH [glucose] levels at the end of this stimulus (Fig. 2G; mixed models analysis, fixed effect of Sex, F_1,15= 5.25, P = 0.037; Sex × Time interaction, F_149,2235= 2.06, P < 0.0001). Compared with the males, females showed a persistent, significantly greater decrease in [glucose] from 78.5 s after tail-touch offset, reaching a nadir of −68.46 ± 44.01 µM, which persisted for the remainder of the analysis window. At the same timepoint, males reported a Δ[glucose] of 26.01 ± 17.60 µM from baseline.

Rapid Changes in LH [Glucose] in Response to Intravenous Injections of Cumulative Doses of Methamphetamine Differ by Sex at Low Doses

Methamphetamine-induced changes in [glucose] were not related to the injection procedure itself. Following an intravenous saline injection, there was a modest variability in the dynamic patterns of Δ[glucose] between sexes, but these changes were not statistically significant. Likewise, when the groups were combined, we did not observe a significant change in [glucose] following intravenous saline (Fig. 3A).

Figure 3.

Rapid changes in lateral hypothalamus (LH) [glucose] after intravenous saline and cumulative doses of methamphetamine. All data shown as means ± SE. Graphs show changes in glucose concentration between male (blue circles) and female (gold circles) rats. Saline and doses of methamphetamine which did not result in significant differences between sexes were combined (red circles). Saline injections (A) had no significant sex differences or effects on LH [glucose] compared with baseline. The first dose of methamphetamine (B) had a rapid but modest phasic increase in LH [glucose] during the injection in males and a smaller increase in females, with a rapid decrease after the injection in females. LH [glucose] decreased from baseline for both sexes and was higher overall in males. During the second injection of methamphetamine (C) there was a slight phasic increase in LH [glucose] during the injection for both sexes, then a decrease. Males continued decreasing in [glucose] while females increased over the first 180 s. The third injection of methamphetamine (D) had a phasic increase in females in LH [glucose] like the previous injections but a biphasic increase then decrease in males. Both sexes showed [glucose] near baseline after the third injection. During the fourth injection of methamphetamine (E), there was also a biphasic increase then decrease in LH [glucose] in both sexes during the injection and then a steady decrease. Time intervals where male and females are different by Fisher post hoc are shown as a blue bar on the X-axis; time intervals where the combined group was different from pre-injection baseline by Fisher test are shown as a red bar on the X-axis. The average length of the injection is shown as the gray shaded area.

Next, cumulative doses of methamphetamine were administered and Δ[glucose] was analyzed at rapid and slow timescales (i.e., 2 s and 30 s, respectively). When analyzed at the rapid timescale, after the first 0.025 mg/kg injection of methamphetamine, both sexes decreased from baseline (Fig. 3B; fixed effect of Time, F_89,1424= 3.61, P < 0.0001), and overall Δ[glucose] after the injection varied by sex (fixed effect of Sex, F_1,16= 5.00, P = 0.040). There was no Sex × Time interaction at this dose. During this injection, qualitatively there was a rapid but modest phasic increase in [glucose] in males (31.81 ± 17.56 µM at 8.5 s) that was less apparent in females (4.82 ± 4.82 µM at 12.5 s). Due to variability, this interinjection period was not statistically different in the omnibus test. In contrast, females showed a greater decrease immediately after the injection (−45.40 ± 20.59 µM) than males (8.31 ± 15.81 µM) 50.5 s after the injection. After 180 s, male and female postinjection Δ[glucose] reached −8.67 ± 11.92 µM and −32.93 ± 14.54 µM, respectively. During the second injection where 0.05 mg/kg of total methamphetamine was delivered, there was a fixed effect of Sex (Fig. 3C; F_1,17= 6.71, P = 0.019) and an interaction of Time × Sex (F_89,1513= 2.58, P < 0.0001), but no fixed effect of Time. Qualitatively, females showed a slight phasic increase during the injection that peaked 14.5 s after the injection (15.97 ± 9.15 µM). In contrast, postinjection increases were less in males (4.99 ± 8.15 µM, at 10.5 s). After the immediate end of the injection, both sexes showed a decrease similar to the first injection. Considering the entire analysis window, male [glucose] decreased throughout, reaching −41.90 ± 13.43 µM. In contrast, female [glucose] after this injection increased again to 21.00 ± 23.97 µM by the end of the analysis window. An overall dose of 0.1 mg/kg of methamphetamine (Fig. 3D, left) induced a modest phasic rise in [glucose] during the injection in females (16.08 ± 14.66 µM), and a small decrease (−22.82 ± 19.98 µM) during the injection in males; due to high variability and the brevity of these changes, these intrainjection changes were not statistically significant. Likewise, there were no fixed effects or interactions between males and females. When both groups are considered (Fig. 3D, right), no overall changes were seen during the injection period, and both sexes had a gentle curve near baseline postinjection, ending at −16.66 ± 9.79 µM from baseline. The last total dose of methamphetamine, 0.2 mg/kg, revealed comparable results for both sexes, so that there was no fixed effect of Sex nor an interaction (Fig. 3E, left). Analyzed together, during the injection there was another small and variable biphasic rise (4.87 ± 7.97 µM) and fall where the levels continued to decrease throughout the analysis interval (Fig. 3E, right; fixed effect of Time, F_90,1260= 2.07, P < 0.0001), ending at −23.7 ± 14.58 µM from baseline.

Slow Changes in LH [Glucose] after Intravenous Methamphetamine Differ by Dose, Time, and Sex

When analyzed at the slow timescale, the first and lowest dose of methamphetamine in drug-naive rats induced a persistent decrease in LH [glucose] from baseline levels, so that by the end of the 50-min analysis window, levels were 106.21 ± 70.48 µM below baseline in males and 318.83 ± 68.93 µM in females (Fig. 4A). This decrease was significantly greater in females than males (mixed models analysis, fixed effects of Time F_99,1386= 9.33, P < 0.0001, Sex F_1,14= 6.90, P = 0.02, and a Sex × Time interaction F_99,1386= 1.53, P = 0.001). This difference appeared in females ∼11.7 min after the injection and persisted for the entire analysis window. After the second methamphetamine dose, both male and female [glucose] levels decreased after the injection (Fig. 4B; mixed models analysis, fixed effect of Time F_{99, 1485}= 12.63, P < 0.0001), so that levels in males and females were 175.58 ± 41.53 µM and 187.82 ± 63.16 µM, respectively, below baseline. There were no effects of Sex at this dose. When given an overall dose of 0.1 mg/kg, males decreased in LH [glucose] more than females around the midpoint of the analysis window, but levels reached comparable levels by the end of the period (Fig. 4C; male −166.96 ± 42.72 µM, female −148.65 ± 17.25 µM below baseline, fixed effects of Time F_99,1485= 13.11, P < 0.0001 and a Sex × Time interaction F_99,1485= 1.46, P = 0.003). During the final 0.2 mg/kg methamphetamine dose (Fig. 4D), glucose levels differed dynamically during the analysis window. [Glucose] in females fell below baseline levels in two epochs, one which reached −28.84 ± 28.74 µM at 6.5 min postinjection. Males also showed a similar initial epoch reaching −33.09 ± 19.78 at 5.5 min postinjection. After this first interval, both sexes began increasing glucose levels back toward initial levels, but males continued to increase levels so that there was a marked biphasic increase in males above baseline peaking at 49.22 ± 38.76 µM at 27.5 min after the injection. In females, this effect was much more limited, such that [glucose] never increased above the baseline level at 10.5 min postinjection. Afterward, glucose levels fell, resulting in females showing significantly less glucose −140.86 µM ± 22.73 compared with males −34.66 µM ± 32.50 at the end of the analysis window (fixed effects of Time F_99,1386= 6.23, P < 0.0001, Sex F_1,14= 5.62, P = 0.033, and a Sex × Time interaction F_99,1386= 6.85, P < 0.0001).

Figure 4.

Slow changes in lateral hypothalamus (LH) [glucose] after cumulative doses of intravenous methamphetamine. All data are shown as means ± SE. Graphs show changes in glucose concentrations between male (blue circles) and female (gold circles) rats. Although the first injection of methamphetamine (A) decreased LH [glucose], the effect was greater in females. During the 0.05 mg/kg dose, both groups decreased in [glucose] to a similar degree (B). At 0.1 mg/kg, males decreased more consistently than females, but both groups achieved similar levels at the end of the analysis period (C). During the 0.2 mg/kg dose, both males and females showed a biphasic response to methamphetamine, but in males the second effect that peaked around 27.5 min postinjection was significantly greater in magnitude (D). Time intervals where male and females are different by Fisher post hoc test are shown as a blue bar on the X-axis (P < 0.05). The time of the injection onset is shown as a dashed line.

Baseline and Integrative Measures Differ by Sex

Throughout the recording, we determined absolute baseline levels of glucose before each event. Baseline levels of LH [glucose] before the three nondrug stimuli when averaged together were 2,505 ± 24.52 µM for males, and 2,413 ± 19.10 µM for females. LH [glucose] baselines decreased throughout the recording for both sexes (Fig. 5A; fixed effect of Time F_7,119= 28.43, P < 0.0001). There was no significant effect of Sex alone. However, there was a Sex × Time interaction (F_7,119= 2.17, P = 0.0413), so that baselines decreased more in females than males after the administration of methamphetamine. In addition, the overall integrative response to methamphetamine was greater in females at a lower dose of the drug when compared with males, as indicated by sex differences in slow changes of LH [glucose] via postinjection AUC (Fig. 5B; mixed models analysis, fixed effect of Dose F_3,41= 3.37, P = 0.0276 and a Sex × Dose interaction F_3,41= 7.17, P = 0.0006). Post hoc comparisons further revealed a significant difference between males and females at 0.025 mg/kg.

Figure 5.

Changes in baseline and integrative measures of lateral hypothalamus (LH) [glucose] after natural and drug stimuli. All data are shown as means ± SE. Graphs show baseline glucose concentrations (A) and area under the curve (AUC) of changes in glucose concentrations (B) between male (blue circles) and female (gold circles) rats over the course of the procedure. Although nondrug stimuli (A) did not change baseline glucose concentrations, both males and females showed a gradual decrease after successive methamphetamine injections, with females showing a greater decrease than males. Integration of change in glucose concentrations over 50 min of each methamphetamine injection (B) revealed negative AUC for all doses except 0.2 mg/kg in males. Females showed the greatest decrease from baseline after the 0.025 mg/kg dose, whereas males showed greater decreases after the 0.05 and 0.1 mg/kg cumulative doses. Note that males show a typical U-shaped dose response with this metric. Post hoc comparisons further revealed a significant difference between males and females at 0.025 mg/kg (P = 0.0007).

Within our sample, our female rats were smaller than their age-matched male counterparts, thus sex strongly covaried with weight. However, regression analyses comparing weight with integrative effects within each sex did not find a significant relationship between these two factors (Supplemental Fig. S2).

Lateral Hypothalamic Extracellular [Glucose] Dynamics following Intravenous Glucose Administration

At the end of each experiment, we administered an intravenous injection of 60 mg glucose to test the functionality of the recording, and to assess how changes in blood glucose effects its levels in the LH. In evaluating these data, it is important to acknowledge that due to technical limitations associated with the dead volumes of the intravenous catheter assembly and to minimize disrupting the infusion lines during the experimental recording, this infusion of glucose was always preceded by an infusion of methamphetamine, which corresponded to 0.015 mg of the drug. Interestingly at this scale, no changes in glucose were observed during the methamphetamine injection (Fig. 6, A and B). There was no significant effect of Sex at either time scale; thus, the data were analyzed as a single group (Fig. 6, C and D). As the injected glucose entered the catheter, there was an immediate rise in LH [glucose] for the duration of the glucose injection of 52.5 s so that glucose levels linearly increased to 395.12 µM ± 96.65 from baseline (Fig. 6C; fixed effect of Time, F_1.429,24.29= 19.83, P < 0.0001). Once the injection completed, LH glucose levels continued to rise at a slower rate to reach a maximum of 649.19 µM ± 124.54 at 3.74 min postinjection (Fig. 6D). At the end of the ∼15-min recording interval, the [glucose] had decreased to 304.75 µM ± 65.71 from baseline. The time to t½ max was ∼14.24 min postinjection.

Figure 6.

Changes in lateral hypothalamus (LH) [glucose] after an injection of 0.015 mg methamphetamine immediately followed by 60 mg glucose. All data are shown as means ± SE. Graphs show changes in glucose concentrations between male (blue circles) and female (gold circles) rats at rapid 2-s (A) and slow 30-s (B) time resolutions. As no significant differences between sexes were found at either time resolutions, data were considered together (red circles). Both male and female rats showed a rapid increase in LH [glucose] during the injection when glucose (shaded red and orange intervals) but not methamphetamine (shaded gray intervals) was entering the body (C). After the injection completed, the level of [glucose] continued to rise at a slower rate, peaking ∼3.74 min after the injection (D). Time intervals where both male and females are different by Fisher post hoc test from baseline are shown as a red bar on the X-axis of C and D (P < 0.05). The time of the injection for (B and D) shown as a dashed line.

DISCUSSION

The major objectives of this study were to assess both rapid and slow changes in extracellular glucose concentrations in the LH of freely moving male and female rats to 1) naturalistic nondrug stimuli, 2) cumulative doses of intravenous methamphetamine that are readily self-administered, and the reactivity of the LH to 3) direct injections of intravenous glucose. We found that second-scale changes in LH extracellular glucose levels occur in response to sensory stimuli of different modalities, and that sex differences are present on some but not all nondrug stimuli. In addition, we found that while rapid changes in glucose levels in response to low doses of methamphetamine differed by sex, these differences were limited to periods after the end of the injection. Across all methamphetamine doses, the changes in glucose levels were equivocal and variable during the period of the injection. Conversely, slow changes in lateral hypothalamic glucose levels postdrug differed by dose, time, and sex in a complex pattern; integrative analysis (i.e., AUC and absolute baselines) of these results suggested that females’ lateral hypothalamic glucose dynamics were more sensitive to the effects of methamphetamine at a lower dose than males. For example, the estimated baseline levels of glucose in the LH decreased more for females than males after the administration of methamphetamine. Finally, we found that direct intravenous injection of glucose at the end of the experiment resulted in rapid and robust increases in lateral hypothalamic glucose that were similar between the sexes.

Rapid Changes in Lateral Hypothalamic Extracellular Glucose Concentrations in Response to Sensory Stimuli; Links with Neuronal Firing Activity and Changes in Local Cerebral Blood Flow

Extracellular glucose concentrations in the brain at any point in time reflects the contribution of several different processes. Most central glucose originates from the peripheral circulation, where the concentration is much higher (19, 20, 65, 66). Glucose in neurovasculature is transported across the blood-brain barrier via gradient-dependent facilitated diffusion by GLUT-1 transporters found on endothelial cells into the brain’s extracellular space (67, 68). Thus, extracellular central glucose concentrations recorded electrochemically could be the result of several related mechanisms. In line with hypotheses of neurovascular coupling and functional hyperemia, rapid increases in regional glucose concentrations are linked with local vasodilation in response to increased neural activity, possibly through astrocytic mediation (9, 17, 19, 69–72). In brain areas with phasic neuronal firing patterns like the nucleus accumbens, sensory stimulation that induces phasic neuronal activation (73, 74) is directly related to local extracellular glucose levels (24, 75). Conversely, in brain regions that are constitutively active and are inhibited during sensory stimuli, such as the substantia nigra pars reticulata, glucose levels generally show an inverted relationship with sensory activation (24). One novel finding from our current study indicates that the LH, like the nucleus accumbens (21, 24, 75), and the hippocampus (76) also experiences rapid increases in extracellular glucose levels of a similar magnitude due to sensory stimuli of different modalities. In terms of the phenotype, the LH is a heterogeneous area that contain distinct populations of gamma aminobutyric acid (GABA), melanin-concentrating hormone (MCH), and orexin expressing neurons (77–80), making the LH functionally, biochemically, and anatomically distinct from the nucleus accumbens. Yet, both orexin and MCH lateral hypothalamic neurons in male mice regulate their activity at a behaviorally relevant timescale after the presentation of a novel object (81). Orexin neuron activity rapidly increases upon novel object presentation, whereas MCH neuron activity is time locked to later periods associated with object exploration (81). Yamashita et al. (82) have shown that orexin neurons in the male mouse LH increase their activity in response to sensory stimuli that specifically have aversive qualities. Together, the rapid elevations of extracellular glucose observed in the LH of our current study strongly suggest that this change follows local increases in neuronal activity.

Examining the end of the stimuli used in this study, both the removal of the novel object and termination of the tail-touch resulted in a modest but significant decrease in extracellular glucose levels, to the extent that there was a greater significant decrease in glucose levels post-tail-touch in females than in males (Fig. 2G). The end of tail-touch stimuli in male rats appears to be associated with a decrease in impulse activity in the nucleus accumbens (73). Glucose levels after the offset of sensory stimuli used in our study suggests that lateral hypothalamic neuronal firing is briefly inhibited during this time, and that this inhibition is greater in females than in males. A parsimonious explanation is that the neurons that were active during stimuli exposure became inhibited after stimuli removal. However, it is possible that in a heterogeneous area, such as the LH, distinct population(s) of neurons could become inhibited from sensory stimulus offset, while others remain active.

Neurovascular coupling hypotheses also predict that local vasodilation increases concurrently with extracellular glucose levels to support neuronal firing (17, 24, 83). For example, local field potentials, multiunit electrophysiological activity, and CBF recordings in rat somatosensory cortex demonstrate contemporaneous increases from baseline during forepaw stimulation (84). Rapid, positively correlated changes in CBF and neuronal activity as a result of sensory stimulation have also been reported in the male and female rat barrel cortex (12, 85). These rapid changes in blood flow in response to neural activity are further linked to dynamic dilations of individual arterioles and capillaries (86). As increases in local vasodilation are positively associated with increased glucose delivery, it is likely that in our study, the immediate poststimulus increases in glucose levels seen in the LH of both male and female rats were tightly coupled with increases in neural electrophysiological activity and increased local CBF.

Interestingly, the relationship between CBF and neuronal activity is altered in anesthetized rats [(12): (urethane 1.25 g/kg ip); (87): (isoflurane 1.1%–2.1% inhalation)]. Anesthesia is an important consideration for neurophysiological experiments, as in addition to perturbations in neuronal-vascular coupling, general anesthesia depending on its type during the experiment can also disrupt CBF in diverse ways (e.g., see Ref. 88). Although these effects may depend on the detection method and its associated temporal resolution, general anesthesia also induces robust, minute-scale hyperglycemia in the nucleus accumbens of male rats when monitored by glucose biosensors [(56): (0.8–0.9 mL Equithesin iv)]. In our current study, short periods (2–3 min) of isoflurane anesthesia were used during the implant procedure at the beginning of each recording to safeguard the integrity of the biosensor and to minimize the possibility of sensor damage. Notably, this approach was consistent across all experimental groups and every rat underwent a 2-h postimplant period, reducing possible confounds from isoflurane to influence our measurements of dynamic glucose concentrations. Critically, all glucose concentration measurements during the experiment were made in awake, unanesthetized rats.

It should be noted that others studying different brain areas and tasks have observed negative correlations between measures of neuronal activity and extracellular glucose levels. For example, Li and Freeman (89) reported that dynamic glucose fluctuations in the cat primary visual cortex are phasically decreased in a time-locked manner with increased local spike activity and local field potential from the same area. This has led others to interpret that rapid phasic decreases in extracellular glucose levels in the hippocampus and dorsolateral striatum of male rats learning and performing a maze learning task is associated with increased neuronal activity (90). In addition, Béland-Millar and Messier (91) report modest decreases in extracellular glucose concentrations during wheel running in male mice at the second-by-second timescale in the primary motor cortex. Integrating these findings with ours remains challenging, as there are key experimental differences between these studies and our current work that may influence mechanistic interpretations. For example, differences in experimental data analysis, task differences, the use of anesthesia in some experimental approaches, and species differences may all contribute to an incomplete understanding of the link between fluctuations in neuronal activity and extracellular glucose concentrations. Moreover, the relationship between neuronal impulse activity and extracellular glucose levels may differ by region. Additional systematic work directly establishing how region-specific neuronal firing is associated with local glucose fluctuations is needed to clarify functional interpretations.

Acute Intravenous Methamphetamine Injections Result in Equivocal Changes in Lateral Hypothalamic [Glucose] during the Injection

In this study, cumulative doses of intravenous methamphetamine noncontingently administered to freely moving male and female rats resulted in an inconsistent change in glucose level during the injection itself. Intrainjection elevations of glucose during drug delivery result from peripheral sensory effects of some drugs of abuse (for review, see Ref. 92). For example, injections of cocaine induce robust and consistent elevations of extracellular glucose in the nucleus accumbens of male rats (21). These phasic responses do not require cocaine entering the brain, and appear largely dopamine independent (21). Given that in our current study, methamphetamine delivery resulted in inconsistent and variable intrainjection responses both between sexes and subsequent doses, such as between 0.025 mg/kg (Fig. 3B) and 0.1 mg/kg (Fig. 3D), we speculate that either the LH does not receive drug-related peripheral sensory stimuli like the nucleus accumbens, or that methamphetamine over these doses lacks peripheral-stimulating characteristics. Considering that pharmacologically similar psychostimulants like MDPV in the nucleus accumbens of male rats also fail to show robust intrainjection glucose elevations (23), the prediction that methamphetamine lacks peripheral-stimulating effects may be more likely.

Slow but Robust Decreases in Lateral Hypothalamic [Glucose] after Methamphetamine; Possible Mechanisms

At a longer timescale postinjection, methamphetamine resulted in slow but robust decreases in glucose concentrations. These effects depended on sex and the dose. Notably, females showed a greater decrease from baseline after the initial, lowest dose of methamphetamine than males. In addition, at the highest cumulative dose tested, lateral hypothalamic glucose dynamics were virtually identical for the first 20 min postinjection, and then males showed a relatively large (∼50 µM) and persistent increase in glucose that was absent in females. Overall, doses of methamphetamine generally resulted in a slow decrease in extracellular lateral hypothalamic glucose levels.

A possible influence on LH glucose levels after methamphetamine is GLUT-1 plasticity. As the main endothelial glucose transporter responsible for transport across the blood brain barrier, GLUT-1 levels are intimately related to local metabolic energy demand, cerebral blood flow, and extracellular glucose levels in a healthy brain (67). GLUT-1 expression is both experience-dependent and plastic. For example, chronic visual deprivation for 7 days in male rats results in a downregulation of GLUT-1 levels in various parts of the visual system, while water deprivation after 3 days results in an upregulation in brain osmoregulatory structures (93, 94). Moreover, GLUT-1 levels in the hippocampus and sensorimotor cortex of male mice are regulated at a behaviorally relevant timescale; GLUT-1 levels increase in these brain regions only 220 min after operant learning (95). The ability of GLUT-1 to be regulated by experience at this timescale is an important consideration for methamphetamine-related changes in glucose dynamics. Methamphetamine exposure has been shown to dose-dependently decrease GLUT-1 expression in primary human brain endothelial cell cultures, a main component of the blood-brain barrier, as well as in vivo (96). Thus, sustained decreases in extracellular glucose levels observed in our current study may reflect a dynamic reduction of GLUT-1 levels in the LH after exposure to methamphetamine at a much more rapid timescale (i.e., minutes rather than hours) than previously observed.

Another possible, although not necessarily exclusive, mechanism by which extracellular glucose levels decrease after methamphetamine is due to reductions in local CBF to the LH. The effects of amphetamines on CBF have been equivocal, with some reporting no effects in humans (97), increases in CBF in anesthetized baboons (98) and awake male humans (41), and regionally dependent, localized increases and decreases in male and female humans (40, 99). At a subminute resolution, Polesskaya et al. (39) showed that intraperitoneal methamphetamine in mice induced a biphasic initial rise in local CBF in the somatosensory cortex that peaked after ∼5 min, which then was followed by a sustained 45-min decrease in blood flow. This decrease was linked to vasoconstriction of the pial arterioles. Moreover, acute intravenous methamphetamine has been shown to decrease hypothalamic blood-oxygen-level dependent (BOLD) signal changes from baseline compared with saline or cocaine injections in anesthetized male rats for up to 20 min postinjection (100). Taken together, these findings suggest sustained decreases in local CBF due to vasoconstriction and altered hemodynamics are a possible contributor to our observed decreases in lateral hypothalamic glucose after acute methamphetamine.

Associating these changes in extracellular glucose with other neurophysiological responses remains challenging due to the paucity of preclinical data examining the intersection of methamphetamine and its effects on the LH. Using single-unit recording in awake rats, a single moderate dose of intraperitoneal methamphetamine induces temporally distinct patterns of neuronal excitation and inhibition in the prefrontal cortex of male rats (101). Elsewhere, the similar drug amphetamine changes the impulse activity of dorsal striatal neurons in a complex pattern of responses depending on their neurobehavioral function (102). The impact of methamphetamine on the LH has also been characterized by molecular approaches, such as changes in immediate early gene (IEG) expression. Changes in IEG expression such as c-fos and its protein product Fos (103) have been widely used as a biochemical marker related to increased neuronal activity (104). Increases in Fos expression have been reported in the LH of male rats after a single intraperitoneal injection of methamphetamine (105). Modest increases have also been observed after acute intravenous infusions (47). In sum, extant findings suggest that acute methamphetamine may have a complex interaction with patterns of impulse and biochemical activity in the LH at the minute-scale postinjection. However, further systematic characterization of lateral hypothalamic neuronal activity in response to acute methamphetamine would be needed to delineate mechanistic links with dynamic glucose levels at this timescale.

Decreases in Lateral Hypothalamic [Glucose] after Cumulative Doses of Methamphetamine; Implications for Glucose-Sensing Neurons

Marked decreases in extracellular glucose after acute methamphetamine may have implications for the functional neuroanatomy of the LH. Although the LH was one of the first brain regions studied in terms of reinforcement (33), it has also been extensively studied in the context of homeostatic energy regulation because neurons within the LH are nutrient sensing. That is, they change their activity based on extracellular glucose concentrations. Four different neuronal phenotypes in this brain area are glucose sensitive (27, 106–108). GABA, orexin, and neuropeptide Y neurons become electrophysiologically inhibited when extracellular glucose levels are high (37, 107). In contrast, MCH neurons present in the LH are excited when glucose levels are high (37, 107). It is notable that after some acute methamphetamine injections, lateral hypothalamic extracellular glucose levels fluctuate at the same scale as seen during direct injection of glucose in the LH or during naturalistic free drinking of glucose in the nucleus accumbens (58). Therefore, this suggests that these glucose changes induced by acute methamphetamine may also have an influence on lateral hypothalamic neuronal activity via their glucose-sensing mechanisms.

Basal Levels of Lateral Hypothalamic Extracellular Glucose and Reactivity after Direct Intravenous Injection; Comparisons with Other Structures

Baseline levels of lateral hypothalamic glucose before methamphetamine were relatively stable and similar between males and females, estimated at 2,500.16 µM, or ∼2.5 mM. The extracellular concentration of glucose in the brain has been debated because of a wide range of reported values and different experimental approaches. For example, glucose levels in the nucleus accumbens of awake male rats using electrochemical approaches similar or identical to this study have consistently provided basal concentrations in the range of 540–707 µM (21, 24, 56), and are in line with electrochemical estimates of the somatosensory cortex (109) and hippocampus (76) of awake male rats. Electrochemical estimates from the hippocampus of anesthetized male rats (2.5 mM) (110) and the anterior hypothalamus of anesthetized male and female rats (2.4 mM) (19) have been estimated to be much higher. These differences have been characterized to be the result of experimental differences related to the use of anesthesia during recordings, given that general anesthesia disrupts extracellular glucose regulation and induces central hyperglycemia (56, 76, 109, 111). However, results in our current study with awake rats are surprising, as our basal levels are similar to those of Silver and Erencinka (19), who reported measuring extracellular glucose levels in the anterior hypothalamus along with various other regions of the cortex. This opens the possibility that basal extracellular glucose concentrations vary by local brain structure and location more than previously characterized.

Another avenue for comparison is by assessing the dynamic changes in extracellular glucose concentration after a direct 60 mg injection of glucose. This has previously been reported in awake male rats for the nucleus accumbens. There, the infusion induced a peak of 184.5 ± 18.1 µM, ∼5 min after infusion (60). Comparatively, the peak change in the LH observed in this study was ∼3.5 times that of the nucleus accumbens. In addition, extracellular glucose levels in the LH appeared more rapidly, as the time to reach this peak was faster by ∼2 min, and the mean glucose concentration began to rise within seconds of the intravenous glucose infusion entering the body. Conversely, the t½ max times were similar in both regions. Given the average male and female baselines before this injection, this represented a change of ∼35%–52% above baseline. This suggests that extracellular glucose delivery in the LH after abrupt changes in the blood supply is faster and greater in scale than in the nucleus accumbens. Why baseline levels and responsiveness to direct intravenous injections of glucose are greater in the LH than the nucleus accumbens remains unclear. Recent work using quantitative gene expression indicates that the level of GLUT-1 expression in the male rat hypothalamus as a whole is ∼60% of the level in the nucleus accumbens (112). Moreover, according to Zeller et al. (68), GLUT-1 and capillary density are tightly correlated across different local brain structures in rats. Therefore, it would be expected that both GLUT-1 expression and vascularization would be lower in the LH than the nucleus accumbens, resulting in less extracellular glucose detected at any one moment. However, others have characterized the hypothalamus as having “the most luxuriant blood supply in the brain” (113). In addition, capillary vascularization across the hypothalamus is not uniform, and some (114) have reported that within subregions of the hypothalamus, the expression of GLUT-1 on capillaries is also heterogeneous. In sum, this and our results here suggest that there may be a more complex relationship between GLUT-1, blood flow, and extracellular glucose levels in the LH than previously assumed.

Sex Differences in Lateral Hypothalamic Extracellular Glucose Regulation

Our findings indicated that extracellular levels of lateral hypothalamus glucose of males and females differed during the end of the tail-touch (Fig. 2G). Namely, glucose levels rose above baseline in males but continued to decrease in females after the end of the tail-touch. One reason for this difference may be because as a mildly aversive stimulus, the tail-touch could involve a stress response. Stress responses such as increased adrenocorticotropic hormone and corticosterone release are regulated in part by the LH, whose orexin neurons make reciprocal excitatory connections to corticotropin-releasing factor (CRF) neurons in the larger hypothalamic pituitary-adrenal (HPA) axis (115). CRF neurons are considered the key regulator of the HPA axis to stress, and some studies suggest that higher hypothalamic CRF levels in females underlies sex differences in the stress response (116). Given that rapid changes after sensory stimuli like tail-touch are likely related to changes in neuronal firing and local CBF in the LH, this would suggest that both these factors are decreased in females compared with males after the end of the tail-touch, contributing to the greater sensitivity in the physiological response toward a mildly aversive stimulus.

When considering rapid changes in glucose levels after intravenous methamphetamine, significantly lower levels of glucose were detected in females than males ∼27 s after the first 0.025 mg/kg injection of methamphetamine (Fig. 3B). A parsimonious explanation for this difference might be that decreases in local hemodynamics are faster in females than in males, considering overall decreases in local cerebral blood flow in the somatosensory cortex (39) after methamphetamine. The decline in extracellular glucose observed in females at the rapid timescale may additionally reflect the beginning of the decrease also seen at the slow timescale. The second injection of methamphetamine at the rapid timescale (Fig. 3C) showed the opposite pattern as the first injection. Here, females displayed a slightly higher rapid increase than males and had higher glucose concentrations than males ∼2–3 min after the intravenous injection. The timeframe of this difference suggests that this increase in females is a result of a central effect of methamphetamine, as there was adequate time for methamphetamine to cross the blood-brain barrier after intravenous infusion. However, unlike the first injection, this increase in females was not sustained at the longer, slower timescale (e.g., Fig. 4B). That this later increase is only significant at the rapid timescale implies differences in neuronal activation in the LH after methamphetamine enters the brain. Although sex differences in the neuronal response to methamphetamine in the LH is largely unexplored, in the adjacent structures such as the paraventricular nucleus and the paraventricular thalamic nucleus, others have not found sex differences in Fos activation after acute methamphetamine in mice (117). In our study, no subsequent methamphetamine resulted in sex differences at the rapid timescale, suggesting that immediate effects of methamphetamine on lateral hypothalamic glucose levels differed across sex only at lower doses.

In contrast, intravenous methamphetamine induced dose- and sex-dependent slow changes in glucose concentration (Fig. 4). First, females had a greater decrease in glucose levels than males after the first injection. After the second dose, this difference was not significant. By the third dose, this difference was reversed so that males showed a greater decrease than females throughout most of the analysis interval. After the final injection, males and females exhibited an approximately identical response for the first 20 min postinjection, which then diverged so that males increased in lateral hypothalamic glucose levels much more than females. The potential causes for this consistently observed difference remains unclear; the long-time scale associated with this analysis suggests that significant sex differences exist in the vasoconstrictive effects of methamphetamine at this dose. Taken together, changes in lateral hypothalamic glucose at the slow timescale suggest that females are more sensitive to the vasoconstrictive effects of methamphetamine at the lowest dose than males. This was surprising, as we did not anticipate a significant sex-dependent effect at 0.025 mg/kg, since we expected physiological drug-effects at higher doses. Although our prediction was true for males, it was not for females. Interestingly, others have also reported that similar low doses are optimal for highlighting sex differences during methamphetamine self-administration (51). At more moderate doses, females appear to show tolerance to this effect more than males. At the highest dose, males consistently have a large, slow increase in glucose levels that is not present in females. Interestingly, integrative measures using area under the curve highlight these sex-dependent trends (Fig. 5B), such that males exhibit a typical U-shaped dose-response curve whereas females show a curve that is shifted to the left compared with males.

One explanation for differences in slow lateral hypothalamic glucose changes (Fig. 4 and Fig. 5B) and the greater drop in baseline levels post-methamphetamine in females (Fig. 5A) may involve differential activation of the HPA axis. Although the activation of the HPA axis involves a complex network of central circuits and pituitary activation, it ultimately raises plasma corticosterone levels, which has been implicated in the addictive liability of several drugs of abuse (e.g., Ref. 118). Importantly, acute methamphetamine in mice results in greater and more prolonged increases in corticosterone levels in females than males, with dynamic differences occurring between 30 and 120 min postinjection (117). This is relevant because higher plasma corticosterone levels have also been associated with reduced cerebral blood volume and a decrease in cerebrovascular diameters in male mice using real-time in vivo two-photon imaging (119). Another explanation for these effects may be differential vasoconstriction caused by the actions of central catecholamines including dopamine, norepinephrine, and serotonin. Amphetamines dose-dependently target the catecholamine transporters and vesicular monoamine transporters in the brain, leading to an overall increase in all three catecholamines when directly delivered to the LH (4, 120). All these catecholamines are potent central vasoconstrictors in male animal models (121–124). This may intersect with sex because endogenous levels of dopamine are regulated by ovarian but not testicular hormones (125). Moreover, the level of extracellular striatal dopamine induced by amphetamine in freely moving female rats are also regulated by ovarian hormones (126). However, additional data specific to the LH is needed to fully elucidate the mechanisms underlying the sex differences in extracellular glucose seen in our study.

Clinical Implications of Sex-Specific Lateral Hypothalamic Extracellular Glucose Regulation in Response to Methamphetamine

Our novel finding that extracellular glucose regulation in the LH in female rats is more sensitive to lower dose effects of methamphetamine contributes to an emerging understanding of sex differences in the physiological and health consequences of methamphetamine and MUD. Preclinical studies that model human MUD indicate that female rats self-administer more methamphetamine, escalate their intake, and exhibit greater locomotor activity after methamphetamine than males, suggesting greater vulnerability in females to MUD (48, 50–52). Although comparisons between animal models and human use disorders must be made carefully, it is nonetheless important to note that a clinical study has recently reported that healthy, non-MUD women are more sensitive to the acute behavioral and subjective effects of methamphetamine than non-MUD men (127). This greater sensitivity to methamphetamine may also be related to more severe psychological and MUD-related symptoms in women, as well as a faster progression of illness (e.g., Refs. 128 and 129; for review, see Ref. 35). Conversely, less sensitivity toward the subjective effects of methamphetamine in males may be related to higher poisoning and overdose death rates from methamphetamine observed in men (130, 131). Ultimately, our study indicates that these sex differences in the behavioral impact of methamphetamine, progression, and severity of MUD in people may have a neurochemical correlate and a possible physiological marker in lateral hypothalamic dynamic extracellular glucose regulation, an important component of brain health. How this factor informs both pharmaceutical and psychosocial therapeutic strategies in humans will need to be delineated further.

Limitations and Future Directions

Our study demonstrates that dynamic lateral hypothalamic extracellular glucose concentrations are changed by natural nondrug stimuli, that there is a dose effect of methamphetamine, and key differences in this regulation exist between males and females. Our results implicate a number of possible neurophysiological processes and systems that could be mediating the effects of acute methamphetamine reported here, but mechanistic conclusions remain largely speculative due to a dearth of direct data. Studies using additional or multimodal recording approaches, such as combining glucose sensing with fiber photometric monitoring of neuronal activity of subpopulations of lateral hypothalamic neurons, will be necessary to further delineate our findings. However, now that clear sex differences have been discovered, this presents novel avenues for better understanding the underlying mechanism. These may include the effects of the estrous cycle on the intersection of methamphetamine and lateral hypothalamic extracellular glucose regulation. Finally, although our study focused on the acute effects of methamphetamine, future studies will be needed to ascertain the consequences of more extensive drug experience on lateral hypothalamic extracellular glucose levels.

Conclusions

Extracellular glucose levels in the LH of freely moving awake male and female rats are dynamically and differentially modulated by both nondrug stimuli and acute doses of methamphetamine. Our novel findings add to contemporary efforts toward delineating the LH as an important brain region in drug reinforcement, energy regulation, and mediating the neurophysiological effects of MUD. Although nondrug, naturalistic stimuli induce rapid changes at the timescale of seconds, acute intravenous methamphetamine impacts LH extracellular glucose levels mostly over slow, minute timescales. Most of these changes at this timescale indicate a significant reduction in extracellular glucose levels in response to the drug, suggesting far-reaching changes in LH functionality. Significant sex differences in these methamphetamine dose effects also suggest that female rats are more sensitive to the effects of the drug at lower doses. Our data newly highlight lateral hypothalamic glucose regulation and its intersection with sex differences as a robust target for future studies examining the etiology of human MUD.

SUPPLEMENTAL DATA

Supplemental Figs. S1 and S2: https://doi.org/10.6084/m9.figshare.20060795.

DATA AVAILABILITY

All analysis code used to generate the results reported in the manuscript have been made publicly available in a GitHub repository (https://github.com/unl-nmblab/biosensing-stats).

GRANTS

This work was supported in part by the Nebraska Tobacco Settlement Biochemical Research Funds (to K.T.W.) and by the Rural Drug Addiction Research (RDAR) Center (COBRE, Grant P20GM130461 to K.T.W. and N.A.H.). N.A.H. was also partially supported by the Brain and Behavior Research Foundation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

I.R.K.K., J.A.J., and K.T.W. conceived and designed research; I.R.K.K., J.A.J., and K.T.W. performed experiments; I.R.K.K., J.A.J., C.H., and K.T.W. analyzed data; I.R.K.K., J.A.J., N.A.H., and K.T.W. interpreted results of experiments; I.R.K.K. and K.T.W. prepared figures; I.R.K.K., J.A.J., and K.T.W. drafted manuscript; I.R.K.K., J.A.J., N.A.H., and K.T.W. edited and revised manuscript; I.R.K.K., J.A.J., C.H., N.A.H., and K.T.W. approved final version of manuscript.