Acute fentanyl and xylazine co-exposure uniquely increase the excitability of dopamine type 2 receptor-expressing striatal medium spiny neurons

Abstract

Co-exposure to fentanyl, a mu-opioid receptor (MOR) agonist, and xylazine, a pan-α₂-adrenoceptor (α₂AR) agonist, results in deleterious interactions thought to ensue from the augmentation of the negative consequences of fentanyl by xylazine. Ex vivo striatal slice electrophysiology was used to investigate the effects of fentanyl (100 nM) and xylazine (10 µM) co-exposure on dopamine type 2 receptor (D2) expressing dorsolateral medium spiny neurons (MSN)s. While the acute application of either fentanyl or xylazine caused some decreases in D2 MSN firing rates at greater stimulus currents, co-exposure dysregulated D2 MSN excitability—resulting in paradoxical increases in excitability at lower stimulating currents, a more robust decrease in firing frequencies at higher stimulus currents, and overall increases in the initial momentary frequency of action potentials. The pan-opioid receptor and selective α_2A-AR antagonists, naloxone (10 µM) and BRL 44408 (1 µM), respectively, assessed the contributions of these receptor types. BRL 44408 and naloxone each partially negated the interactive effects of fentanyl and xylazine co-exposure suggesting the involvement of α_2A-ARs and MORs. Our findings underscore the unique consequences and complex nature of opioid and non-opioid receptor interactions in mediating the combined effects of fentanyl and xylazine in the striatum, which may contribute to the enhancement of fentanyl’s reinforcing properties and underlie reported neurological complications including postural abnormalities and dystonia seen with opioid use disorder.

NEW AND NOTEWORTHY

Acute fentanyl and xylazine co-exposure increased the excitability of striatal dopamine receptor- (D2) expressing medium spiny neurons (MSNs) to lower intensity currents, while strong stimuli decreased firing rates and increased the initial momentary frequency of action potentials. Acute fentanyl or xylazine exposure decreased D2 MSN excitability at higher current intensities with near additive decreases following co-exposure. Fentanyl and xylazine interact to selectively alter D2 MSN excitability potentially contributing to dystonia and enhancing fentanyl’s reinforcing properties.

Graphical Abstract

INTRODUCTION

Fentanyl is a synthetic opioid drug developed in 1959 and introduced as an intravenous anesthetic in the 1960s. Currently, it finds its clinical use in surgical procedures for local anesthesia (1, 2), in breakthrough cancer pain (3), or as an analgetic during labor or post-operative procedures (4). Like morphinan opioids, such as morphine, heroin, and oxycodone, fentanyl induces analgesia primarily via the activation of µ-opioid receptors (MOR). However, compared to morphine and other opioids, fentanyl and its analogues have much higher potency and efficacy at MORs, may take longer to metabolize and excrete (5–8), and may bind distinct pockets of MOR (9, 10), which likely contributes to a higher degree of misuse liability. Unlike morphinan or synthetic opioids such as methadone, evidence suggests that fentanyl and/or its analogues can have off-target effects on other opioid and non-opioid receptor types and/or unique biased agonist properties at MOR that may contribute to its increasing role in opioid overdose deaths (6, 11–16). The high mortality rate seen in opioid use disorder (OUD) is largely attributable to opioid-induced respiratory depression (OIRD) through actions in pontine (17, 18) and medullary (19) nuclei. Similar to other opioids, fentanyl inhibits respiration by activating MOR-expressing neurons in the parabrachial nucleus of the pontine respiratory group (17, 18), and in the nucleus tractus solitarius, the preBӧtzinger complex, and in the raphe nucleus (19, 20) in the medulla. By contrast, unlike opioids, fentanyl use increases the risk for developing wooden chest syndrome (7, 13, 21, 22), which is characterized by chest wall muscle rigidity, laryngospasm, and possibly death within minutes of fentanyl injection (21, 23). Wooden chest syndrome and certain other detrimental effects of fentanyl can be refractory to opioid antagonists (24), such as naloxone (Narcan®), suggesting that some of fentanyl’s unique actions are not mediated by opioid receptors. Some of the inability of naloxone to reverse OIRD may result from the short duration of naloxone’s action compared to fentanyl (25). Alternatively, because fentanyl is frequently adulterated with xylazine to increase its efficacy (25), unique fentanyl-xylazine interactions may additionally contribute to a decreased responsiveness to opioid antagonists (25–27) (see below).

Non-MOR-mediated deleterious effects of fentanyl may be linked to its binding with α_1A and α_1B adrenoceptors (ARs), dopamine D1, D2, and D4.4 receptors, and ability to block [³H]neurotransmitter uptake by the vesicular monoamine transporter 2 (28). We recently found that sustained fentanyl exposure can decrease the activity in striatal MSNs co-cultured with glia via a non-opioid-dependent mechanism(s) involving the activation of α₁ARs through actions in striatal neurons and/or potentially astroglia (16). Specifically, we found that acute fentanyl exposure caused modest, but significant decreases in spontaneous firing rates in striatal MSNs (16). By contrast, overnight exposure to fentanyl silenced nearly all spontaneous activity in MSNs and the suppression was unaffected by naloxone, but fully reversed by the selective α_1A adrenoceptor antagonist RS 100329 (16), suggesting the involvement of non-opioid receptors and that the duration of exposure is important for understanding fentanyl’s actions.

The recent surge in the use of xylazine, a pan-α₂ AR agonist, which is mixed with fentanyl to heighten fentanyl’s effects (29–31) to create the highly toxic drug cocktail “Tranq-dope”, has dramatically worsened the opioid crisis (32, 33). The sedative action of xylazine may exacerbate the effects of fentanyl by prolonging fentanyl’s sedative and perhaps rewarding effects (for review, see (34)). This is supported by recent data suggesting that xylazine increases the duration of fentanyl’s actions and delays withdrawal symptoms (35), an effect that is likely mediated by xylazine acting at α₂ARs. Fentanyl orthosterically binds α_2B-ARs with a K_i of 950 nM (36) and can heterologously cross-desensitize α₂ARs via a MOR-dependent mechanism following sustained exposure (37), possibly through the formation of direct MOR and α₂AR heteromeric molecular complexes (38). In combination, fentanyl and xylazine can act synergistically via α₂ARs in vivo (36, 39–43), and xylazine can exacerbate the bradycardic and respiratory depressive effects of fentanyl (40). In the striatum, α_2A- and α_2C-ARs are expressed by subpopulations of MSNs (44, 45) and α_2A-ARs are expressed by astrocytes (46, 47) suggesting that fentanyl and xylazine may act via α₂ ARs to directly affect striatal function. To explore possible interactions between fentanyl and xylazine in the striatum, we assessed the drug’s actions in D2 MSNs, which display more pronounced adaptive responses to sustained (2–5 h) fentanyl exposure (48), and more accelerated pathophysiological responses to HIV-1 Tat (49, 50), than dopamine D1 receptor-expressing (D1) MSNs.

Here, we show that acute co-exposure to fentanyl (100 nM) and xylazine (10 µM), which inhibit D2 MSN firing rates when added separately, paradoxically increases the excitability of D2 MSNs and increases adaptive changes especially at low stimulating currents. Our findings underscore the complex nature of fentanyl and xylazine interactions in the striatum.

MATERIALS AND METHODS

Animals.

All mice used in our experiments were housed in the Virginia Commonwealth University School of Medicine animal facility with unrestricted access to food and water and 12:12 h light-dark cycle (lights off at 18:00 h). Drd2-eGFP (#036931-UCD; Mutant Mouse Resource and Research Centers) mice expressing eGFP fluorescent marker exclusively in D2 MSNs were used for striatal slice electrophysiology experiments. The Virginia Commonwealth University Animal Care and Use Committee (IACUC protocol #AM10175) has approved the use of mice in this study. All experiments were conducted by following the National Institutes of Health (NIH Publication No. 85–23) ethical guidelines.

Striatal slice preparation.

Coronal striatal slices were prepared as we described earlier (16, 51). Specifically, adult male mice expressing D2-eGFP-labeled MSNs were anesthetized using 4% isoflurane for at least 3 min and maintained on isoflurane throughout the process of transcardial perfusion. Mice were perfused transcardially for 5 min with calcium-free sucrose cutting medium containing the following (in mM): 3 KCl, 4.12 MgSO₄, 1.2 NaH₂PO₄, 206 sucrose, 25 NaHCO₃, and 25 glucose. The cutting medium was aerated using a 5% CO₂ and 95 % O₂ mixture. Brains were removed, separated from the cerebellum, and glued to the cutting platform in a way most suitable to obtain coronal slices. Immediately after mounting to the cutting platform, the brains were placed into the cutting chamber of Leica VT1200 S vibratome filled with the aerated cutting medium cooled to 1–3 °C. Then the brains were cut into 350 μm-thick slices. After visual inspection, three rostral striatal slice cuts from the left and right hemispheres each were selected and kept in the aerated artificial cerebrospinal fluid (ACSF) solution containing (in mM): 125 NaCl, 3 KCl, 1.2 CaCl₂, 1.2 MgSO₄, 1.2 NaH₂PO₄, 25 NaHCO3, 25 glucose and aerated with a gas mixture of 5% CO₂ + 95% O₂ at 37 °C for 30 min for cold shock recovery and were then maintained at room temperature up to 4 hours.

Electrophysiology of neurons in striatal slices.

Patch-clamp recordings were obtained from D2 dorsolateral MSNs as described earlier (16, 51). Briefly, brain slices were placed into a recording chamber mounted on a Zeiss Axio Observer.D1 microscope. The recording chamber was constantly perfused with ACSF solution. D2-eGFP-labeled neurons were visualized using Prizmatix Ultra High-Power LED system with filters for green wavelengths, and a DAGE-MTI IR 1000 monochrome camera. Patch-clamp pipettes (6.0 – 8.5 MΏ) were pulled from borosilicate glass with a filament (catalog number BF150–86-10, Sutter Instruments, Novato, CA) on a Narishige PC-10 two-step puller (Narishige International USA, Inc, Amityville, NY). Pipettes were backfilled with intracellular solution containing (in mM): 140 K-gluconate, 10 KCl, 1 NaCl, 1 CaCl₂, 1 MgCl₂, 4 Mg-ATP, 5 EGTA, 10 HEPES, pH 7.4. Whole-cell current-clamp recordings were obtained by using MultiClamp 700B amplifier coupled with an Axon Digidata 1550A digital-analog converter and pClamp 11 software (Molecular Devices, Son Jose, CA). Action potentials recorded in the current-clamp mode were evoked by injecting −100 pA to 600 pA currents at 25 pA increments with each step lasting 500 ms and 5 s intervals between sweeps. With some exceptions, multiple neurons (n = 2 – 5) were recorded from in each mouse, and multiple mice (N = 3 – 5) were used to collect data for each experimental group; except for the fentanyl and xylazine treatment group, where we used two animals and obtained five recordings from each. Importantly, there were no experimental groups with overwhelming number of recordings obtained from the same animal. All recordings were done at 32 °C.

Offline data analyses.

Data were analyzed using IgorPro 8 (WaveMetrics, Lake Oswego, OR), Clampfit 10.7 (Molecular Devices, San Jose, CA), Prism 8 (GraphPad Software, San Diego, CA), and Microsoft Office Excel 2016 (Microsoft Corporation, Redmond, WA) software suites. To characterize the frequency of action potential firing, we counted the number of action potentials per s during 0.5 s current injection stimuli. The obtained values were plotted against injected currents. A robust action potential amplitude decline was observed in some recordings at strong depolarizing stimuli as the firing progressed. Thus, we used a threshold of a minimum 15 mV above the strong stimulus (e.g. 600 pA)-depolarized baseline to count an event as an action potential. Based on our current and earlier reported data, this threshold was sufficient to separate events with well-defined “sharp” peak and, thus, to classify them as action potentials. Importantly, the same depolarized baseline + threshold value was used to detect differences in action potentials for within the cell comparisons. The I₅₀, the injected current causing the half-maximum number of action potentials per s, rheobase, and the resting membrane potential values were used to assess drug effects during within-cell comparisons. To evaluate action potential firing adaptation, we used recordings with the highest firing frequency. We measured intervals between neighboring action potential peaks and then normalized them to the first interval value and plotted against the interval count. The first interval in the count was assigned to zero. The action potential firing adaptation plots were fitted to a linear function using the least squares method, and the slope was used for within-cell drug(s) vs. control comparisons, which we further refer to as the adaptation slope. The R² was calculated and reported to evaluate how well the linear model fits the experimental data. Additionally, we measured the initial momentary frequency, a reciprocal time interval between the first and the second action potentials of firing, and the latency, time from the current injection beginning to the peak of the first action potential of firing and plotted them against the injected current values to characterize the adaptation and firing in greater detail. To gain more mechanistic insights into fentanyl and/or xylazine effects, amplitude, half-width, threshold, and maximal velocity of the upstroke was determined for the first action potential. Specifically, the amplitude was defined as the difference between the resting membrane potential and the action potential peak. The half-width was calculated at 50% of the peak amplitude. The threshold was an absolute value of the membrane potential observed at the beginning of the action potential upstroke. The maximal velocity of the upstroke (dV_m/dt)_max) was determined by calculating the maximum value of the first derivative of the first action potential.

Statistical analyses.

Kolmogorov-Smirnov test was used to test on normal data distribution. For normally distributed groups, statistical significance (p < 0.05) was determined using a Student’s paired, two-tailed t-test for within-cell comparisons. For non-normally distributed groups, we utilized Wilcoxon matched-pairs single rank test and considered data significantly different if p < 0.05. Two-way, repeated measures ANOVA followed by Tukey’s multiple comparisons post hoc test was used to compare injected current-dependent recordings of evoked action potentials, initial momentary frequency, and the first latency before and after the acute drug applications in ex vivo striatal slices. For data that were not normally distributed, the Kruskal-Wallis, followed by Dunn’s multiple comparisons tests, were used to find statistically significant differences. All data are present as the mean ± the standard error of the mean (SEM).

Chemicals and reagents.

Fentanyl was obtained from the NIDA Drug Supply Program, Rockville, MD. BRL 44408 was purchased from Tocris (Bio-Techne, Minneapolis, MN). All other chemicals, including xylazine, were purchased from Sigma-Aldrich (St. Louis, MO).

RESULTS

Acutely and separately applied fentanyl and xylazine differentially alter excitability in D2 MSNs.

We first wanted to know how acute applications of fentanyl or xylazine would change the firing in D2 MSNs. Similar to our earlier report (16), acutely applied 100 nM fentanyl caused only a minor decrease in the number of action potentials fired during strong current injection stimulations (Fig. 1B, *p < 0.05, n = 11, N = 3; two-way repeated measures ANOVA). There were no significant changes in I₅₀, an injected current half-maximum number of action potentials firing response, (Fig. 1F), rheobase (Fig. 1E), or resting membrane potential (Fig. 1G) caused by acute fentanyl application (p > 0.05, n = 11, N = 3; Student’s paired t-test). The adaptation slope (see the definition above in the Methods) was significantly increased from 0.0279 ± 0.0050 (R² ranging from 0.69 to 0.95, 0.86 ± 0.02) in control to 0.0514 ± 0.0113 (R² ranging from 0.73 to 0.98, 0.89 ± 0.02) (Fig. 1D, **p < 0.01, n = 11, N = 3; Wilcoxon matched-pairs single rank test). The initial momentary frequency of action potential firing measured as the reciprocal of the time interval between the first and the second action potentials and the latency, time from the current injection beginning to the peak of the first action potential of firing, were not changed by acute application of fentanyl alone (Fig. 1H and I, respectively). We found that acute fentanyl application neither affected the action potential threshold (Fig. 1J) nor amplitude (Fig. 1K) across all stimuli (p > 0.05, n = 11, N = 3; two-way repeated measures ANOVA). However, acute fentanyl widened action potential half-width at injected currents ≥ 200 pA (Fig. 1L) and decreased action potential upstroke velocity at 225 pA, 250 pA, 450 pA, 475 pA, 500 pA, 550 pA, 575 pA, and 600 pA stimuli (Fig. 1M) (*p < 0.05, n = 11, N = 3, two-way repeated measures ANOVA).

Figure 1. Acutely applied fentanyl modestly reduces action potential firing frequency caused by strong saturating injected currents but promotes the action potential adaptation and widens the action potential half-width in D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 600 pA stimuli in control (A) and in the presence of 100 nM fentanyl (A’). (B) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in controls (black circles) and in the presence of 100 nM fentanyl (light blue circles). Fentanyl significantly affected D2 MSN number of action potentials when 400 pA, and 450 pA stimuli were applied (*p < 0.05, n = 11, two-way repeated measures ANOVA). (C) Action potential firing adaptation observed in (A) and (A’) is altered in the presence of 100 nM fentanyl as summarized in (D), for the linear fitting slope (**p<0.01, n = 11, paired t-test). (E - G) Fentanyl did not change rheobase (E), I₅₀ (F), or resting membrane potential (RMP) (G) (p > 0.05, n = 11, paired t-test). (H) Initial momentary frequency, a reciprocal to the duration of the first inter-action potential interval, and latency, the time between the beginning of a stimulus and a peak of the first action potential, plotted against the injected current, (I) are not altered by acute application of fentanyl in D2 MSNs (p > 0.05, n = 11, two-way repeated measures ANOVA). (J) and (K) Fentanyl does not affect the threshold and amplitude, respectively, of the first action potential (AP) (p > 0.05, n = 11, two-way repeated measures ANOVA). (L) Fentanyl significantly increased the AP width at injected currents equal or greater than 200 pA, (*p < 0.05, n = 11, two-way repeated measures ANOVA). (M) Fentanyl decreased maximal AP upstroke velocity ((dV_m/dt)_max) at 225 pA, 250 pA, 450 pA, 475 pA, 500 pA, 550 pA, 575 pA, and 600 pA stimuli (*p < 0.05, n = 11, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

Acutely applied xylazine caused a significant reduction in the number of action potentials at injected current levels equal to or greater than 325 pA (Fig. 2B, *p < 0.05, **p < 0.01, ***p < 0.001, n = 11, N = 4; two-way, repeated measures ANOVA). This effect appeared to be stronger than if fentanyl only applied acutely (compare Fig. 2B vs. Fig. 1B). Similarly to fentanyl-only application, xylazine caused no changes in I₅₀ (Fig. 2F), rheobase (Fig. 2E), and the resting membrane potential (Fig. 2G) (p > 0.05, n = 11, N = 4; Student’s paired t-test). The adaptation slope significantly increased from 0.0257 ± 0.0046 in control (R² ranging from 0.46 to 0.96, 0.79 ± 0.05) to 0.0571 ± 0.0071 in the presence of xylazine (R² ranging from 0.77 to 0.98, 0.88 ± 0.02) (Fig. 2D, ***p < 0.001, n = 11, N = 4; Student’s paired t-test). No changes caused by acute application of xylazine in momentary frequency or latency were observed in D2 MSN (Fig. 2H and Fig. 2I, respectively, p > 0.05, n = 11, N = 4; two-way, repeated measures ANOVA). The acute application of xylazine did not affect the firing threshold at most stimulating current levels with the exception of 475 pA, 500 pA, and 525 pA stimuli (*p < 0.05, n = 11, N = 4; two-way, repeated measures ANOVA) (Fig. 2 J), and had no effect on the amplitude of the action potentials (Fig. 2 K), the half-width (Fig. 2 L), or the maximal velocity of the upstroke (Fig. 2 M) (p > 0.05, n = 11, N = 4; two-way, repeated measures ANOVA). Together, these data show that fentanyl and xylazine can alter D2 MSN action potential firing in a similar way, albeit to a lesser degree for fentanyl.

Figure 2. Acutely applied xylazine reduces action potential firing frequency caused by strong saturating injected currents and promotes adaptation in D2 MSN firing rates.

(A) and (A’) Representative traces illustrate the action potential firing caused by 600 pA stimuli in control (A) and in the presence of 10 μM xylazine (A’). (B) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in controls (black circles) and in the presence of 10 μM xylazine (light blue circles). Xylazine significantly affected D2 MSN number of action potentials when 325 pA, 350 pA, 375 pA, 400 pA, 425 pA, 450 pA, 475 pA, 500 pA, 525 pA, 550 pA, 575 pA, and 600 pA stimuli were applied (*p < 0.05, **p < 0.01, and ***p < 0.001, n = 11, two-way repeated measures ANOVA). (C) Action potential firing adaptation observed in (A) and (A’) is altered in the presence of 10 μM xylazine as summarized in (D), for the linear fitting slope (***p < 0.001, n = 11, paired t-test). (E - G) Acute exposure to xylazine did not change rheobase (E), I₅₀ (F), or resting membrane potential (RMP) (G) (p > 0.05, n = 11, paired t-test). (H) Initial momentary frequency and latency (I) are not altered by acute application of xylazine in D2 MSNs (p > 0.05, two-way repeated measures ANOVA). (J) Fentanyl decreased the action potential (AP) threshold only at 475 pA, 500 pA, and 525 pA (illustrated by horizontal line and “*, p < 0.05 label”) (*p < 0.05, two-way repeated measures ANOVA, n = 11). (K - M) Fentanyl did not alter the amplitude (K), half-width (L), or the maximal velocity of the upstroke ((dV_m/dt)_max) (M) of D2 MSN action potentials (p > 0.05, two-way repeated measures ANOVA, n = 11). Data are presented as the mean ± the SEM.

Acutely co-applied fentanyl and xylazine alter robustly excitability in D2 MSNs.

We wondered if xylazine addition to fentanyl could exacerbate the effects of fentanyl observed earlier (16). We found that the co-application of 100 nM fentanyl and 10 µM xylazine caused robust changes in action potential firing in D2 MSNs (Fig. 3). The inhibitory effect of fentanyl and xylazine on the number of action potentials was profoundly and statistically significantly reduced at injected currents 450 pA or higher (Fig. 3C, *p < 0.05, **p < 0.01, n = 10, N = 2; Student’s paired t-test). Unexpectedly, the entire number of action potentials vs. injected current curve was shifted to the left resulting in higher than in control firing rates at weak and moderate current stimulations in a range from 150 pA to 250 pA (Fig. 3C, *p < 0.05, **p < 0.01, n = 10, N = 2; Student’s paired t-test). Quantitatively, the leftward shift resulted in I₅₀ from 264.2 pA ± 14.5 pA in control to 216.0 pA ± 19.5 pA (Fig. 3E, **p < 0.01, n = 10, N = 2; Student’s paired t-test) in the presence of fentanyl and xylazine. Similarly, the rheobase was reduced from 205.0 pA ± 15.3 pA in controls to 167.5 pA ± 21.4 pA (Fig. 3F, **p < 0.01, n = 10, N = 2; Student’s paired t-test) in the presence of fentanyl and xylazine. Similar to exposure to fentanyl- and xylazine alone, co-application of fentanyl and xylazine caused no changes in resting membrane potential in D2 MSNs (Fig. 3G, p > 0.05, n = 10, N = 2; Student’s paired t-test).

Figure 3. Acutely applied fentanyl and xylazine reduce action potential firing frequency caused by strong saturating injected currents but make it more sensitive to weak and mid-range stimuli, promote adaptation, increase initial momentary frequency, and decrease the first action potential amplitude and maximal velocity in D2-eGFP medium spiny neurons (MSNs).

(A) and (A’) Representative traces illustrate the action potential firing caused by 175 pA stimuli (A) and 525 pA (A’) in control and in the presence of 100 nM fentanyl and xylazine at 175 pA (B) and 525 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in vehicle-exposed controls (black circles) and in the presence of 100 nM fentanyl and 10 μM xylazine (light blue circles). Fentanyl and xylazine significantly affected MSN number of action potentials when 150 pA, 175 pA, 200 pA, 225 pA, 250 pA, 450 pA, 475pA, 500 pA, 525 pA, 550 pA, 575 pA, and 600 pA stimuli were applied (*p < 0.05, **p < 0.01, and ***p < 0.001, n = 10, two-way repeated measures ANOVA). (D) Co-applications of fentanyl and xylazine significantly increased the action potential firing adaptation linear fit slope (***p < 0.001, n = 10, paired t-test). (E) Fentanyl and xylazine decreased I₅₀ (**p > 0.01, n = 10, paired t-test). Black bar and circles and light blue bar and squares indicate control and fentanyl and xylazine group values, respectively. (F) Fentanyl and xylazine application reduced the rheobase (**p > 0.01, n = 10, paired t-test) but did not change resting membrane potential (RMP) (G) (p > 0.05, n = 10, paired t-test) compared to controls. (H) Initial momentary frequency is sensitized by fentanyl and xylazine co-application to the injected currents. The horizontal bar illustrates a range of injected currents causing significantly different (*p < 0.05, n = 10, two-way repeated measures ANOVA) momentary frequencies in control (black circles) vs. acutely applied fentanyl and xylazine (light blue circles). (I) Latency shows no significant difference between the control and fentanyl and xylazine groups (p > 0.05, n = 10, two-way repeated measures ANOVA). (J) and (L) Fentanyl and xylazine did not alter action potential (AP) threshold (J) and half-width (L), respectively (p > 0.05, two-way repeated measures ANOVA, n = 10). (K) and (M) Fentanyl and xylazine decreased the amplitude (K) and upstroke maximal velocity ((dV_m/dt)_max) (M) of the APs, respectively. Significant differences across stimuli are noted as horizontal lines and “*, p < 0.05” labels (*p < 0.05, two-way repeated measures ANOVA, n = 10). Data are presented as the mean ± the SEM.

Co-application of 100 nM fentanyl and 10 µM xylazine promoted the adaptation of action potential firing in D2 MSNs (Fig. 3D). Specifically, the adaptation slope increased from 0.0287 ± 0.0042 in control (R² ranging from 0.46 to 0.93, 0.80 ± 0.05) to 0.1113 ± 0.0161 (R² ranging from 0.88 to 0.98, 0.93 ± 0.01) in the presence of fentanyl and xylazine (Fig. 3D, ***p < 0.001, n = 10, N = 2; Student’s paired t-test). The adaptation slope fold change for acutely co-applied fentanyl and xylazine (4.63 ± 0.81, n = 10, N = 2) was significantly higher than for fentanyl-only but not xylazine-only acute applications (2.03 ± 0.61, p < 0.05, n = 11, N = 3, and 2.91 ± 0.58, p > 0.05, n = 11, N = 4, respectively; Kruskal-Wallis test followed up by Dunn’s multiple comparisons test). Interestingly, the initial momentary frequency was significantly higher at injected current values 175 pA or higher in the presence of fentanyl and xylazine (Fig. 3H, *p < 0.05, n = 10, N = 2; two-way repeated measures ANOVA). We also found that, despite the appearance of the leftward shift, the latency was not significantly changed by acute co-application of fentanyl and xylazine across all injected current values (Fig. 3I, p > 0.05, n = 10, N = 2; two-way repeated measures ANOVA). Unlike fentanyl or xylazine alone, acute co-applications of these drugs caused a significant reduction in the action potential amplitude at stimuli equal to or higher than 225 pA (Fig. 3K, *p < 0.05, n = 10, N = 2; two-way repeated measures ANOVA). Fentanyl and xylazine co-exposure did not alter the action potential threshold (Fig. 3J) and half-width (Fig.3L) (p > 0.05, n = 10, N = 2; two-way repeated measures ANOVA) but reduced the upstroke maximal velocity at 250 pA – 350 pA and 400 pA – 600 pA stimuli ranges (Fig. 3M, *p < 0.05, n = 10, N = 2; two-way repeated measures ANOVA). Together, our data suggest that acute co-application of fentanyl and xylazine causes robust alterations in D2 MSN action potential firing via mechanisms involving MOR and α2ARs.

BRL 44408, a selective α_2A-adrenoceptor antagonist, partially disrupts the interactive effects of acute co-application of fentanyl and xylazine on action potential firing rates in D2 MSNs.

We wanted to understand if inhibition of α_2AARs disrupts the interaction of acutely applied fentanyl and xylazine (Fig. 3). We found first that acute application of BRL 44408 (1 µM), the most selectiveα_2A-AR antagonist, did not alter the number of action potentials (p > 0.05, n = 9; N = 3; two-way repeated measures ANOVA) (Fig. 4C), I₅₀ (Fig. 4E), rheobase (Fig. 4F), and resting membrane potential (Fig. 4G) (p > 0.05, n = 9, N = 3; Student’s paired t-test) when compared to controls. Similarly, BRL 44408 application did not affect the adaptation slope (p > 0.05, n = 9, N = 3; Student’s paired t-test) (Fig. 4D), initial momentary frequency (Fig. 4H), and latency of action potential firing (Fig. 4I) (p > 0.05, n = 9, N = 3; two-way repeated measures ANOVA). As anticipated, BRL 44408 application did not affect the action potential threshold firing (Fig. 4J), amplitude (Fig. 4K), half-width (Fig. 4L), or upstroke maximal velocity (Fig. 4M) (p > 0.05, n = 9, N = 3; two-way repeated measures ANOVA).

Figure 4. Acutely applied BRL44408 does not change action potential firing in D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 200 pA stimuli (A) and 575 pA (A’) in control and in the presence of 1 μM BRL44408 at 200 pA (B) and 575 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in controls (black circles) and in the presence of 1 μM BRL 44408 (turquoise circles). (p > 0.05, n = 9, two-way repeated measures ANOVA). (D) Action potential firing adaptation linear fitting slope is not altered in the presence of 1 µM BRL 44408 (p > 0.05, n = 9, paired t-test). (E - G) Acute exposure to BRL44408 did not change I₅₀ (E), rheobase (F), or resting membrane potential (RMP) (G) (p > 0.05, n = 9, paired t-test). (H) Initial momentary frequency is not affected by acute application of 1 µM BRL 44408 (p > 0.05, n = 9, two-way repeated measures ANOVA). (I) Latency is not altered by acute application of 1 µM BRL 44408 (p > 0.05, n = 9, two-way repeated measures ANOVA). (J - M) BRL 44408 did not alter the action potential threshold (J), amplitude (K), half-width (L), and upstroke maximal velocity ((dV_m/dt)_max) (M), respectively (p > 0.05, n = 9, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

In the presence of BRL 44408 (1 µM), acute co-application of fentanyl (100 nM) and xylazine (10 µM) caused more modest effects if compared to fentanyl and xylazine vs. control (Fig. 5). Specifically, the number of action potentials was modestly decreased at injected currents equal to or higher than 400 pA but increased at 200 pA (Fig. 5C, *p < 0.05, **p < 0.01, ***p < 0.001, n = 11, N = 3; two-way repeated measures ANOVA). We found that I₅₀ was reduced from 228.6 pA ± 16.3 pA in the presence of only BRL 44408 to 206.3 pA ± 16.8 pA when fentanyl and xylazine were added (Fig 5E, *p < 0.05, n = 11, N = 3; Student’s paired t-test). Similarly, the rheobase was reduced from 170.5 pA ± 12.1 pA to 147.7 pA ± 14.5 (Fig. 5F, *p < 0.05, n = 11, N = 3; Student’s paired t-test). No changes were observed for the resting membrane potential (Fig. 5G, *p < 0.05, n = 11, N = 3; Student’s paired t-test). The adaptation slope was increased from 0.0334 ± 0.0046 in the presence of BRL 44408 only (R² ranging from 0.75 to 0.94, 0.87 ± 0.02) to 0.0605 ± 0.0072 in the presence of BRL 44408, fentanyl, and xylazine (R² ranging from 0.78 to 0.98, 0.90 ± 0.02) (Fig. 5D, ***p < 0.001, n = 11, N = 3; Student’s paired t-test). This effect was qualitatively identical to the fentanyl-only acute application (Fig. 1D). Similarly to fentanyl-only or xylazine-only acute applications, we observed no changes in the initial momentary frequency of action potential firing (Fig. 5H) or the latency (Fig. 5I) (p > 0.05, n = 11, N = 3; two-way repeated measures ANOVA). Interestingly, the addition of BRL 44408 to fentanyl and xylazine negated the reduction of action potential upstroke maximal velocity (Fig. 5M) observed earlier in the absence of BRL 44408 (Fig. 3M), did not affect the action potential threshold (Fig. 5J) and half-width (Fig. 5L) but preserved qualitatively small but significant reduction in action potential amplitude at 200 pA, 250 pA – 450 pA, and 550 pA – 600 pA ranges of stimuli (Fig. 5K) (*p < 0.05, n = 11, N = 3; two-way repeated measures ANOVA). Together these data indicate that α_2AARs play a critical role in the interactive effect of acutely applied fentanyl and xylazine.

Figure 5. BRL 44408 weakens the combined effect of acutely applied fentanyl and xylazine on action potential firing frequency in D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 200 pA stimuli (A) and 575 pA (A’) in the presence of 1 μM BRL 44408 (BRL) and in the presence of 1 µM BRL 44408, 100 nM fentanyl, and 10 µM xylazine at 200 pA (B) and 575 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in BRL 44408-exposed controls (turquoise circles) and in the presence of 1 µM BRL 44408, 100 nM fentanyl, and 10 μM xylazine (light blue circles). The addition of fentanyl and xylazine to BRL 44408 significantly affected MSN number of action potentials when 200 pA, 400 pA, 450 pA, 475 pA, 525 pA, 550 pA, 575 pA, and 600 pA stimuli were applied (*p < 0.05, **p < 0.01, and ***p < 0.001, n = 11, two-way repeated measures ANOVA). (D) Action potential firing adaptation linear fitting slope is altered when 100 nM fentanyl and 10 μM xylazine were added (**p < 0.01, n = 11, paired t-test). (E - G) The addition of fentanyl and xylazine to BRL 44408 decreased I₅₀ (E) and rheobase (F) (*p < 0.05, n = 11, paired t-test) but did not change resting membrane potential (G) (p > 0.05, n = 11, paired t-test). (H) Initial momentary frequency and latency (I) are not altered by the addition of fentanyl and xylazine to BRL 44408 in D2 MSNs (p > 0.05, two-way repeated measures ANOVA). (J - M) The addition of fentanyl and xylazine to BRL44408 did not affect the action potential threshold (J), half-width (L) and upstroke maximal velocity ((dV_m/dt)_max) (M) (*p < 0.05, n = 11, two-way repeated measures ANOVA) but reduced the amplitude across stimuli of various strengths (K). Horizontal lines illustrate ranges of injected currents causing significantly different (*p < 0.05, n = 11, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

Naloxone disrupts interactive effects of fentanyl and xylazine on evoked action potential firing in D2-eGFP MSNs.

We wanted to clarify the role of opioid receptors in mediating the interactive effects of fentanyl and xylazine on action potential firing in D2-eGFP MSNs. First, we found that 10 μM naloxone, a non-selective inhibitor of μ-, δ-, and κ-opioid receptors, does not affect the number of action potentials vs. injected currents, except at the strongest current stimulus of 600 pA (**p < 0.01, n = 9, N = 5; two-way, repeated measures ANOVA) (Fig. 6C). Naloxone did not affect I₅₀ (p > 0.05, n = 9, N = 5; Student’s paired t-test), rheobase (p > 0.05, n = 9, N = 5; Wilcoxon matched-pairs signed rank test), and the resting membrane potential (p > 0.05, n = 9, N = 5, Student’s paired t-test) (Fig. 6E, F, and G, respectively). Similarly, naloxone caused no significant changes in the adaptation slope from 0.0065 ± 0.0033 to 0.0179 ± 0.0103 (p > 0.05, n = 9, N = 5; Wilcoxon matched-pairs signed rank test) (Fig. 6D). Naloxone also did not affect initial momentary frequency (Fig. 6H), and the latency (Fig. 6I) (p > 0.05, n = 9, N = 5; two-way, repeated measures ANOVA). By itself, naloxone did not alter the action potential threshold (Fig. 6J), amplitude (Fig. 6K) half-width (Fig. 6L), and maximal upstroke velocity (Fig. 6M) (p > 0.05, n = 9, N = 5; two-way, repeated measures ANOVA). Together, our data indicate that naloxone alone does not affect the action potential firing in D2-eGFP MSNs.

Figure 6. Acutely applied naloxone does not change action potential firing in D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 300 pA stimuli (A) and 600 pA (A’) in control and in the presence of 10 μM naloxone at 300 pA (B) and 600 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in controls (black circles) and in the presence of 10 μM naloxone (pink circles). (**p <0.01, n = 9, two-way repeated measures ANOVA). (D) Action potential firing adaptation linear fitting slope is not altered in the presence of 10 µM naloxone (p > 0.05, n = 9, paired t-test). (E - G) Acute exposure to naloxone did not change I₅₀ (E), rheobase (F), or resting membrane potential (RMP) (G) (p > 0.05, n = 9, paired t-test). (H) Initial momentary frequency is not affected by acute application of 10 µM naloxone (p > 0.05, n = 9, two-way repeated measures ANOVA). (I) Latency is not altered by acute application of 10 µM naloxone (p > 0.05, n = 9, two-way repeated measures ANOVA). (J - M) Acutely applied naloxone did not alter the action potential threshold (J), amplitude (K), half-width (L), or upstroke maximal velocity ((dV_m/dt)_max) (M) (p > 0.05, n = 9, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

In our next experiments, we found that the addition of fentanyl (100 nM) and xylazine (10 μM) to 10 µM naloxone caused strong inhibition of the number of action potentials in a range from 300 pA to 600 pA (*p < 0.05, **p < 0.01, ***p < 0.001, n = 11, N = 3; two-way repeated measures ANOVA) but similar to acute applications of xylazine (Fig. 2B) did not exhibit the fentanyl and xylazine-induced (Fig. 3C) sensitization of the action potential firing rates to small and mid-range currents less or equal to 275 pA (p > 0.05, n = 11, N = 3; two-way, repeated measures ANOVA)(Fig. 7C). Co-applications of fentanyl and xylazine with naloxone also produced no effects on I₅₀, rheobase (p > 0.05, n = 11, N = 3; Student’s paired t-test), and resting membrane potential (p > 0.05, n = 11, N = 3; Wilcoxon matched-pairs signed rank test) compared to naloxone controls (Fig. 7E, F, and G). We found that the addition of fentanyl and xylazine to naloxone increased the adaptation slope from 0.0168 ± 0.0025 (R² ranging from 0.63 to 0.95, 0.83 ± 0.02) to 0.0569 ± 0.0153 (R² ranging from 0.81 to 0.99, 0.93 ± 0.02) (p < 0.001, n = 11, N = 3; Wilcoxon matched-pairs signed rank test) (Fig. 7D) but caused no effect on the initial momentary frequency (Fig. 7H ) or the latency (Fig. 7I) (p > 0.05, n = 11, N = 3; two-way, repeated measures ANOVA). Interestingly, the addition of naloxone to fentanyl and xylazine resulted in no significant changes in the action potential threshold (Fig. 7J), amplitude (Fig. 7K), half-width (Fig. 7L), or upstroke maximal velocity (Fig. 7M) (p > 0.05, n = 11, N = 3; two-way, repeated measures ANOVA). Together, naloxone appears to eliminate the combined effects of fentanyl and xylazine co-applications by reducing them to similar effects observed for xylazine-only applications.

Figure 7. Naloxone eliminates the combined effect of acutely applied fentanyl and xylazine on action potential firing frequency in D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 250 pA stimuli (A) and 575 pA (A’) in the presence of 10 μM naloxone and in the presence of 10 µM naloxone, 100 nM fentanyl, and 10 µM xylazine at 250 pA (B) and 575 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in naloxone-exposed controls (violet circles) and in the presence of 10 µM naloxone, 100 nM fentanyl, and 10 μM xylazine (pink circles). The addition of fentanyl and xylazine to naloxone significantly affected the MSN number of action potentials when stimuli were applied in a range from 300 pA to 600 pA (*p < 0.05, **p < 0.01, and ***p < 0.001, n = 11, two-way repeated measures ANOVA). (D) Action potential firing adaptation linear fitting slope is altered when 100 nM fentanyl and 10 μM xylazine were added to naloxone-treated neurons, (***p < 0.001, paired t-test). (E - G) The addition of fentanyl and xylazine to naloxone did not change I₅₀ (E), rheobase (F), and resting membrane potential (G) (p > 0.05, n = 9, paired t-test). (H – I) Initial momentary frequency (H) and latency (I) are not affected by the addition of fentanyl and xylazine to naloxone (p > 0.05, n = 11, two-way repeated measures ANOVA). (J - M) The addition of fentanyl and xylazine to naloxone did not alter the action potential threshold (J), amplitude (K), half-width (L), or upstroke maximal velocity ((dV_m/dt)_max) (M) (p > 0.05, n = 11, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

Synaptic neurotransmitter receptor antagonists partially attenuate the combined effects of fentanyl and xylazine on evoked action potential firing in D2-eGFP MSNs.

To isolate whether the observed effects of fentanyl and xylazine on D2-eGFP MSNs are due to intrinsic cellular effects or local synaptic changes, we next co-applied fentanyl and xylazine in the presence of antagonists for GABA-A (picrotoxin, 100 µM), NMDA (AP5, 10 µM), non-NMDA glutamate (CNQX, 20 µM), and glycine (strychnine, 1 µM) receptors, further referred to as blockers. The effect of acute fentanyl (100 nM) and xylazine (10 µM) co-exposure on evoked action potential firing (Fig. 3) was altered in the presence of the blockers (Fig. 8). Although the number of action potentials was still reduced at injected currents between 425 and 575 pA following acute application of fentanyl and xylazine, the overall leftward shift in action potential firing (Fig. 3C) was eliminated by the blockers (Fig. 8C, *p < 0.05, **p < 0.01, n = 11, N = 4; two-way repeated measures ANOVA). In addition, there was no significant change in the I₅₀ (p > 0.05, n = 11, N = 4, Student’s paired t-test), rheobase (p > 0.05, n = 11, N = 4, Wilcoxon matched-pairs signed rank test), or resting membrane potential (p > 0.05, n = 11, N = 4, Student’s paired t-test) in response to fentanyl and xylazine application when the blockers were present (Fig. 8E, F, and G).

Figure 8. An ionotropic glutamate, GABA_A, and glycine receptor antagonist cocktail (blockers) eliminated the combined effect of acutely applied fentanyl and xylazine on the firing frequency of D2-eGFP MSNs.

(A) and (A’) Representative traces illustrate the action potential firing caused by 250 pA stimuli (A) and 550 pA (A’) in the presence of GABA_A (picrotoxin, 100 µM), NMDA (AP5, 10 µM), non-NMDA glutamate (CNQX, 20 µM), and glycine (strychnine, 1 µM) receptor antagonists in the presence of 100 nM fentanyl and 10 µM xylazine at 250 pA (B) and 550 pA (B’), respectively. (C) The relationship between the amount of injected current and the number of action potentials per duration of stimulation in MSNs in antagonist-exposed controls (dark brown circles) and in the presence of blockers, 100 nM fentanyl, and 10 μM xylazine (light brown circles). The addition of fentanyl and xylazine to the above antagonists significantly affected the number of action potentials when stimuli were applied at 425 pA, 475 pA, 500 pA, 525 pA , and 575 pA (*p < 0.05, and **p < 0.01, n = 11, two-way repeated measures ANOVA). (D) Fentanyl (100 nM) and 10 μM xylazine caused a linear change in firing rate adaptation in MSNs the slope of which was altered the presence of inotropic antagonists (**p < 0.01, paired t-test). (E - G) The addition of fentanyl and xylazine to blockers did not change the I₅₀ (E), rheobase (F), or resting membrane potential (RMP) (G) (p > 0.05, n = 9, paired t-test). (H – I) Initial momentary frequency (H) and latency (I) are not affected by the addition of fentanyl and xylazine to blockers (p > 0.05, n = 11, two-way repeated measures ANOVA). (J - M) The addition of fentanyl and xylazine to blockers did not alter the action potential threshold (J), and half-width (L) (p > 0.05, n = 11, two-way repeated measures ANOVA) but reduced amplitude (K) and upstroke maximal velocity ((dV_m/dt)_max) (M) across stimuli denoted by horizontal line or *, ** labels (*p < 0.05, and **p < 0.01, n = 11, two-way repeated measures ANOVA). Data are presented as the mean ± the SEM.

We also found that acute co-application of fentanyl and xylazine in the presence of the blockers increased the adaptation slope from 0.0145 ± 0.0024 (R2 ranging from 0.79 to 0.94, 0.85 ± 0.02) to 0.0355 ± 0.0063 (R2 ranging from 0.80 to 0.97, 0.93 ± 0.02) (Fig. 8D, ***p < 0.001, n = 11, N = 3; Student’s paired t-test). The relative change in adaptation slope resulting from fentanyl and xylazine co-application with the blockers was qualitatively similar to that observed when fentanyl and xylazine were applied in the presence of BRL 44408 (Fig. 5D), but lower than when fentanyl and xylazine were co-applied alone (Fig. 3D). Finally, we did not find any significant changes in the initial momentary frequency of action potential firing (Fig. 8H)or the latency (Fig. 8I) (p > 0.05, n = 11, N = 4; two-way repeated measures ANOVA) when fentanyl and xylazine were acutely applied in the presence of the blockers. Similarly to fentanyl and xylazine effects, the addition of these drugs to the blockers resulted in a statistically significant reduction in the action potential amplitude at 250 pA – 525 pA range and 575 pA and 600 pA (Fig. 8K, *p < 0.05, **p < 0.01 n = 11, N = 4; two-way repeated measures ANOVA). Fentanyl and xylazine addition to the blockers did not alter the action potential threshold (Fig. 8J) and half-width (Fig. 8L) but reduced the upstroke maximal velocity at 250 pA – 350 pA, 400 pA – 525 pA, and 575 pA- 600 pA stimuli ranges (Fig. 8M, *p < 0.05, **p < 0.01, n = 11, N = 4; two-way repeated measures ANOVA), similar as fentanyl and xylazine vs, control effect. Overall, these findings suggest the combined effect of fentanyl and xylazine may involve modulation of synaptic input to D2-eGFP MSNs.

DISCUSSION

We utilized a striatal ex-vivo brain slice model to uncover the unique effects of fentanyl and xylazine on the firing rates of D2 MSNs. We found that the acute co-application of fentanyl and xylazine resulted in a robust reduction of the number of action potentials at strong depolarizing injected current stimulations but sensitization to weak and moderate stimulations, as well as in a decrease of I₅₀ and rheobase. Additionally, acute co-application of fentanyl and xylazine significantly increased the initial momentary frequency and promoted adaptative alterations in action potential firing rates. If applied separately, both fentanyl and xylazine demonstrated a lack of sensitization of action potential firing to the weak and mid-range injected stimulating currents, no change in I₅₀ and rheobase, no effect on the momentary frequency, and a less remarkable adaptative response. BRL 44408 appears to partially reverse the effects of acute fentanyl and xylazine co-application. BRL 44408 combined with fentanyl and xylazine qualitatively demonstrated a partial loss of sensitization to weak and mid-range current stimuli, decreased changes in I₅₀ and rheobase, less remarkable adaptation, and prevented the leftward shift in initial momentary frequency seen with increasing stimulus currents compared to combined fentanyl and xylazine alone, implying a necessary but not sufficient role of α₂ARs in fentanyl-xylazine interactions. Naloxone appears to reverse the combined effects of fentanyl and xylazine resulting in action potential firing patterns similar to those seen with xylazine alone, underscoring the critical role of opioid receptors. Similar to naloxone, synaptic inhibitors of GABA_A, NMDA and non-NMDA glutamate, and glycine receptors reversed the combined effects of fentanyl and xylazine, suggesting either a direct postsynaptic mechanism of action or indirect altered glia-neuron interactions.

Our results suggest that acutely co-applied fentanyl and xylazine interact to cause paradoxical increases in the excitability of D2 MSNs at lower current stimuli, which may contribute to xylazine’s ability to modulate the rewarding and motor effects of fentanyl. These findings are important because depending on pairing with a reward-predicting cue or a reward, increases in D2 MSN activity have been shown to indirectly increase dopamine release (52) and heighten motivational drive (53). D2 MSNs can also transmit prediction stimuli that enhance the association between cue and outcome thereby influencing learning (54). Thus, xylazine may enhance (55) or limit (56) the rewarding effects of fentanyl and alter learning by selectively modulating the response of D2 MSNs to opioids and dopamine. Furthermore, the observed fentanyl and xylazine-induced dysregulation in D2 MSN function may precede the postural changes seen with fentanyl misuse that appear to be exacerbated by co-exposure to xylazine (57). Postural changes associated with fentanyl misuse include forward head (58), hyperkyphosis, and generalized dystonia (59, 60), and resemble the camptocormia seen in Parkinson’s disease that are largely attributable to alterations in dopaminergic transmission (61–63).

Acutely co-applied fentanyl and xylazine interact to affect multiple aspects of D2 MSN function.

We have shown recently that acutely applied fentanyl could cause a modest reduction in the action potential firing frequency at very strong depolarizing stimulations and promote action potential firing adaptation (16). We also showed that sustained fentanyl exposure could inhibit striatal MSN activity via a non-opioid receptor-dependent pathway, that may be modulated via complex actions in α₁AR-expressing striatal neurons and/or glia (16). The dramatic increase in the combined use of xylazine, an α₂AR agonist, to exacerbate fentanyl’s effects prompted the question of how xylazine might further affect the already complex actions of fentanyl (29–31). Co-administration of MORs and α₂AR agonists can result in synergistic in vivo (39, 41–43) or additive in vitro (64) interactions, and depending on the outcome measured (65). α₂ARs can be expressed by astroglia (46, 47, 66) and neurons (46, 47), and α₂ARs have been shown to regulate GABA release from the astrocytes and, thus, suppress neuronal activity in hippocampal neurons (67). Most MSNs do not express MOR (68–71). Alternatively, subpopulations of striatal astrocytes, oligodendrocytes, and microglia express MORs that can affect CNS function, maturation, and pathogen-induced inflammatory responses (72–79). Thus, in addition to actions in neurons, we speculate that subpopulations of MOR-, α₁AR-, and/or α₂AR-expressing glia may contribute to the combined effects of fentanyl and xylazine.

Our findings demonstrate unique interactions between fentanyl and xylazine, not seen with either drug by itself. This was especially evident for the sensitization of action potential firing when injecting weak to moderate stimulating currents, which resulted in significant decreases in I₅₀ and rheobase, and the sensitization of initial momentary frequency to injected stimulating currents. Adaptative alterations in D2 MSN firing rates, exemplified by progressive increases in inter-action potential intervals, implies that fentanyl would have greater effects following longer duration stimuli or following bursts of spontaneous activity (16). Xylazine resulted in even more pronounced adaptations in firing frequency than with fentanyl alone that were unaffected by co-exposure to fentanyl. Thus, exposure to xylazine alone is sufficient to maximally increase the progressive escalation in inter-action potential intervals irrespective of fentanyl co-exposure. This suggests that fentanyl and xylazine act via a common final mechanism to elicit the adaptative response, and that xylazine is more potent and supersedes the effects of fentanyl in eliciting the adaptive reduction in action potential firing frequency. Lastly, the cumulative inhibitory effects of fentanyl and xylazine on D2 MSN firing rates at higher stimulating currents suggests these drugs act via a common additive pathway to selectively suppress D2 MSN excitability to more intense glutamatergic and/or dopaminergic input from cortico-thalamic and ventral tegmental afferents, respectively.

Predictably, BRL 44408 was able to disrupt the interactions between fentanyl and xylazine following acute exposure. Specifically, we observed a reduction of fentanyl and xylazine-induced sensitization of action potential firing rates when injecting modest amounts of current, and no effects of combining fentanyl and xylazine on the initial momentary frequency. BRL 44408 attenuated the inhibitory effects of combined fentanyl and xylazine on D2 MSN firing rates at higher current steps to levels seen in the present and prior studies (16) seen following the acute application of fentanyl by itself. In contrast, adding naloxone to the fentanyl and xylazine cocktail attenuated the effects of these two drugs to responses resembling those seen with xylazine alone. Thus, the use of the selective antagonist BRL 44408 confirms that α_2A adrenoceptors contribute to the interactive effects of fentanyl and xylazine on D2 MSN action potential firing rates, whereas the use of naloxone suggests that the blockade of opioid receptors alone is sufficient to mediate the effects of fentanyl. This is perhaps not unexpected since xylazine’s actions at α_2A adrenoceptors are well established (36, 39–43), although actions at α_2B and α_2C adrenoceptors (80, 81), as well as at κ opioid receptors (KORs) have been proposed (82), but unlikely to be operative in the present study. Importantly, the interactive increases on D2 MSN excitability are only revealed at lower stimulus intensities and are dependent on stimulus duration—suggesting in combination, these drugs uniquely dysregulate the flow of GABAergic, glutamatergic, and dopaminergic signals into D2 MSNs. In support of this idea, synaptic transmission was blocked by a mixture of GABA_A, NMDA and non-NMDA glutamate, and glycine receptor antagonists. The addition of fentanyl and xylazine in combination with these antagonists mimicked the effects of naloxone on firing rates, suggesting a prominent role of presynaptic afferents (extrastriatal and/or intrastriatal) in mediating the actions of fentanyl when combined with xylazine in D2 MSNs.

On a molecular level, we observed that the acute application of fentanyl was sufficient to decrease the maximal velocity of the upstroke of the first action potential, an effect that was not altered by the co-application of fentanyl and xylazine. Interestingly, both naloxone and BRL 44408 were effective in negating the decrease in maximal velocity seen with fentanyl and xylazine co-exposure. Additionally, synaptic blockade did not negate the decrease in velocity seen in the presence of fentanyl and xylazine. Together, these data suggest that fentanyl may affect the gating machinery of D2 MSN voltage-gated sodium channels via an unknown mechanism that involves opioid and potentially other receptor types, resulting in the slower action potential upstroke. We also observed a unique effect of fentanyl and xylazine co-exposure on the amplitude of the first action potential. When co-applied, fentanyl and xylazine decreased amplitude of the action potentials, and this effect was not observed when fentanyl or xylazine were applied separately or in the presence of BRL 44408 or naloxone. We speculate that fentanyl and xylazine may directly affect the ion channels that regulate neuronal firing by either downregulating sodium channels controlling the rising phase or upregulating potassium channels that repolarize D2 MSNs after firing.

Considering its potent actions at MOR, it is perhaps not surprising that fentanyl has also been reported to act as an agonist at δ opioid receptors (DORs) albeit with substantially lower affinity than at MORs (83). As noted above, xylazine reportedly can act as a full agonist at KORs with a bias towards G-protein signaling pathways (82). Thus, it is possible that some of the effects of fentanyl (and perhaps xylazine) might also be attributable to KOR or DOR activation, or possibly other receptor types (28, 83)—especially considering the high degree of lipophilicity and the likelihood that both drugs accumulate within cell membranes (6). Since the pan-opioid receptor antagonist naloxone fully negated the effects of fentanyl when combined with xylazine, we cannot rule out the possibility that fentanyl might be acting through KOR or DOR, or the outside possibility that some of xylazine’s action are mediated via KORs (82). Nevertheless, given the high efficiency and potency of fentanyl at MOR, there is little reason to assume that fentanyl’s actions are not largely or entirely mediated by MORs. Based on the collective evidence above, we propose that the acute effects of fentanyl and xylazine co-exposure are primarily mediated through actions at MOR and α_2A adrenoceptors in the present study. Lastly, unlike the opioid-dependent effects seen with acute exposure, more prolonged overnight exposure to fentanyl caused a near-complete suppression of spontaneous firing in MSNs (D1 and D2 subpopulations were not discerned) through actions mediated by α_1A adrenergic and not opioid receptors (48). Although the acute effects of fentanyl presented herein are clearly opioid-dependent, we question whether more prolonged exposure to fentanyl and xylazine might have more complex interactions involving MOR, α_2A adrenergic, and possibly α_1A adrenergic receptors and intend to explore these potential interrelationships in future studies.

Experimental model limitations.

MSNs are known for their bi-stable membrane polarization with a quiescent hyperpolarized (from −94 mV to −61 mV) “down” state mediated by inward rectifier potassium channels, and a tonically depolarized (from −71 to −40 mV ) “up” state mediated by voltage-gated A-type K⁺ channels (84, 85). The “up” state is controlled by corticothalamic inputs that are severed in ex vivo striatal slice preparations (85–88). Thus, our experimental model does not account for the potential effects of fentanyl and xylazine on the “up” state and, thus, spontaneous action potential firing in D2 MSNs. It is also noteworthy that considerable variability was evident in a small number of D2 MSNs suggesting that uncontrolled factors, e.g., lateral inhibition from intrastriatal circuitry or gliotransmission from astrocytes may be operative.

Conclusions.

We conclude that acutely applied fentanyl and xylazine i) interact to uniquely increase the excitability of D2 MSNs especially at lower stimulus thresholds, and that ii) opioid and α_2A adrenergic receptors are critical in mediating the combined effects of both drugs. The mechanisms underlying the paradoxical fentanyl and xylazine-induced interactive increases in D2 MSN excitability following acute exposure, including the potential role of the opioid- and adrenoceptor-expressing neuronal and glial types in facilitating these effects, are uncertain and their elucidation warrants further study.

DATA AVAILABILITY

Data will be made available upon reasonable request.