Cocaine induced astrocytic activation was associated with its vasoconstricting effects independent of its neuronal effects in the prefrontal cortex

Abstract

Astrocytes bridge neuronal and cerebrovascular systems, regulating neurotransmitter dynamics and modulating neurovascular coupling (NVC). Cocaine disrupts neuronal, astrocytic, and cerebrovascular system function, which in the prefrontal cortex (PFC), could interfere with executive function. Though astrocytes are known to influence neurotoxic and addictive effects of cocaine through their regulations of local neurotransmitter environments, their role in cocaine-induced vasoconstriction and dysregulated NVC remains unclear. To investigate astrocytic regulation of vascular response to cocaine, we expressed the genetically encoded calcium indicator GCaMP6f in either neurons or astrocytes alongside concurrent expression of astrocytic DREADD(Gq) in the mouse PFC. Hemodynamic changes and cellular Ca²⁺ responses to an acute cocaine challenge (1 mg/kg, i.v.) with and without DREADD(Gq) activation were recorded through a cranial window using a custom multi-wavelength imaging platform. We found that when naïve animals were exposed to acute cocaine, intracellular Ca²⁺ increased in both astrocytes (Ca²⁺_A) and neurons (Ca²⁺_N), cerebral blood vessels vasoconstricted, and oxygenated hemoglobin in tissue (HbO₂) decreased. Activation of astrocytes by DREADD(Gq) exposure also caused long-lasting increases in Ca²⁺_A and vasoconstriction. Consequently, when cocaine was administered after astrocytic activation, there was further Ca²⁺_A elevation and vasoconstriction but limited in magnitude, suggesting a ceiling effect. Meanwhile, cocaine induced increases in Ca²⁺_N were unaffected. Correlation analysis revealed a strong negative association between Ca²⁺_A and cerebrovascular diameters, while Ca²⁺_N was negatively correlated with HbO₂ and this correlation remained unchanged by astrocytic activation. These findings support astrocytes’ involvement in regulating vascular tone including cocaine’s vasoconstricting effects and its disruption of neurovascular coupling. Astrocytes present a potential therapeutic target to alleviate cerebrovascular pathology triggered by cocaine that merits further investigation.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38521-8.

Introduction

Cocaine is a highly addictive drug, and its repeated use disrupts cerebrovascular, neuronal and astrocytic function^1,2. At rest, the brain is estimated to consumes 20% of the body’s energy, with increasing demands during cognitive tasks or exposure to stimulant drugs, such as cocaine³. To meet these nutrient and metabolic demands, cerebral blood flow (CBF) dynamically adjusts to ensure adequate oxygen and nutrient delivery to active neurons through neurovascular coupling (NVC)^4,5. Astrocytes provide a cellular link between neuronal circuitry and blood vessels^6,7, with their end feet enveloping 98% of the cerebrovasculature, releasing vasoactive compounds to trigger vasodilation or vasoconstriction^8–10. Astrocytes also regulate the intrasynaptic space by buffering electrolytes^11,12, recycling neurotransmitters, and functioning as a syncytium to communicate with the surrounding neurons^13,14. Astrocytes’ bidirectional communication with both neurons and the vasculature uniquely positions them as key mediators of NVC¹⁵.

Astrocytes’ involvement in drug seeking behavior¹⁶ and in NVC¹⁷ suggests that they may play a role in cocaine’s effects on cerebral blood vessels^18,19. However, the investigation of astrocyte’s role in brain function has been challenged by the difficulties in recording their activity²⁰ as they have a more stable membrane potential than neurons and do not generate action potentials²¹, making traditional electrophysiological methods less effective for studying their dynamics²². Since astrocyte activation involves fluctuations in intracellular Ca²⁺ concentrations, the detection of astrocytic Ca²⁺ (Ca²⁺_A) signaling can be used as a correlate for their activity^23,24. Genetically encoded Ca²⁺ fluorescence indicators (GECIs), such as GCaMP6f, selectively expressed in astrocytes of transgenic mice (GFAP-Cre), have enabled researchers to advance prior constraints and image cortical astrocytic activity^25–27.

To further characterize astrocytes contribution to cocaine-induced alterations of NVC, we utilized chemogenetics receptors (DREADDs -Designer Receptors Exclusively Activated by Designer Drugs) to modulate astrocytes activation^28,29. GFAP-DREADD (Gq) was used to selectively stimulate astrocytes through G-coupled protein receptors (GPCRs)³⁰, which increase Ca²⁺ through phospholipase C activation and the release of intracellular Ca²⁺ stores^31,32. Previous studies examining astrocytic inhibition (via DREADD(Gi)) demonstrated that reducing astrocytic activity led to vasodilation (up to 9%) and attenuated cocaine-induced vasoconstriction³³. However, the role of astrocytic activation on cocaine’s induced disruption of NVC is still unclear. Here, we hypothesized that activation of astrocytes would impair cocaine induced vasoconstriction without affecting its effects on neuronal activities, thus leading to neurovascular uncoupling. To test this, we utilized our multi-modality optical image technique and genetically encoded calcium indicators (GCaMP6f) along with astrocytic activation via DREADD(Gq) to investigate the effects of astrocytic activation on cocaine induced cerebrovascular changes and astrocytic and neuronal activities in vivo. We studied the PFC as this brain region is critical for executive function and its disruption is implicated in the impairments of self-regulation and compulsive drug consumption in patients with cocaine use disorder (CocUD)³⁴. We found that cocaine induced both neuronal and astrocytic Ca²⁺ elevations together with vasoconstriction and reduced tissue hemodynamics. Astrocytic Ca²⁺_A stimulation with DREADD(Gq) resulted in long-lasting vasoconstriction, such that when cocaine was administered after their activation further vasoconstriction was limited, suggesting a ceiling effect whereas cocaine induced increases in Ca²⁺_N were unaffected. Temporal correlation analysis revealed a strong negative association between Ca²⁺_A and cerebrovascular diameters, highlighting astrocyte’s involvement in regulating vascular tone and in cocaine’s disruption of NVC.

Methods

Animals and experimental design

All experimental procedures were approved by the Institutional Animal Care and Use Committee at Stony Brook University and conducted according to the National Institutes of Health (NIH) Guidelines for Care and Use of Laboratory Animals. Transgenic GFAP-Cre (Jackson Laboratory, https://www.jax.org/strain/024098) and C57BL/6 wild type (WT) mice (Jackson Laboratory) were used for this study when they reached ~ 10 weeks of age. These experimental mice were divided into three groups: an astrocytic group, a neuronal group and a control group (summarized in Table 1). Physiological conditions of the animals (e.g., body temperature, respiration rate) were continuously monitored and recorded (Small Animal Monitoring and Gating system, model 1025 L, SA Instruments Inc.). Specifically, the body temperature was maintained as 36–37.5 °C by a customized heating pad (with voltage-controlled electric circles) and the respiration was maintained at 40–80 breaths/minute using isoflurane mixed with oxygen gas (1–2% isoflurane in 100% O₂ at ~ 1.5 L/min) to ensure similar physiological status for all mice. All experiments were performed according to the relevant guidelines, regulations, and ARRIVE guidelines.

Table 1.

Experimental design.

Group	Animals/Viral injection	Experiment	Drug challenge
Group1: Astrocytic Ca2+A, vascular diameter and hemodynamic changes with acute cocaine	GFAP-Cre mice (n = 6) AAV5.CAG.Flex. GCaMP6f.WPRE.SV40 AAV5.GFAP. hM3D(Gq). mCherry	Without DREADD (Gq) activation	Cocaine (10 min baseline + 60 min imaging)
		With DREADD (Gq) activation	a) Clozapine (10 min baseline + 30 min imaging), followed by b) Cocaine (10 min baseline + 60 min imaging)
Group2: Neuronal Ca2+N, vascular diameter and hemodynamic changes with acute cocaine	Wild type mice (n = 6) AAV5. Syn. GCaMP6f.WPRE.SV40 AAV5.GFAP. hM3D(Gq). mCherry	Without DREADD (Gq) activation	Cocaine (10 min baseline + 60 min imaging)
		With DREADD (Gq) activation	a) Clozapine (10 min baseline + 30 min imaging), followed by b) Cocaine (10 min baseline + 60 min imaging)
Group3: Control group vascular diameter changes induced by acute cocaine without DREADD (Gq) activation	Wild type mice (n = 3)	Saline injection	a) Cocaine (10 min baseline + 60 min imaging), followed by b) Saline (10 min baseline + 30 imaging), followed by c) Cocaine (10 min baseline + 60 min imaging)

Dual viral injection into mouse cortex

To detect astrocyte or neuron-specific Ca²⁺ fluorescence response to an acute cocaine challenge with or without astrocytic activation (GFAP-DREADD (Gq)) by clozapine, we used the genetically encoded calcium indicator GCaMP6f via viral delivery into the prefrontal cortex (PFC). Specifically, a mixture of two viruses (0.5 µl/each virus) AAV5.CAG.Flex.GCaMP6f.WPRE.SV40 (AV-5-PV2816,  Addgene) with AAV5.GFAP. hM3D(Gq).mCherry (50479-AAV5, Addgene) were slowly infused at 0.1 µl/min into transgenic mice (GFAP-Cre, Fig. 1c_i) to express GCaMP6f in astrocytes in the right PFC [A/P: +2; M/L: 0.75; D/V: −0.5 mm]. To express GCaMP6f into neurons, 0.5 µl AAV5.Syn.GCaMP6f.WPRE.SV40 (AV-5-PV2822, Addgene) mix with 0.5ul AAV5.GFAP. hM3D(Gq). mCherry, were infused into the same brain region of wild type mice (C57BL/6, Fig. 1c_ii). During viral injections, mice were anesthetized with isoflurane mixed with pure oxygen and their heads mounted on a stereotaxic frame and their physiology monitored. After completion of procedures, the mice were monitored daily for three days to ensure they were healthy.

Fig. 1.

A schematic illustration of experimental design and animal models. a Multi-wavelength imaging setup for GCaMP6f fluorescence, total hemoglobin (HbT), and deoxygenated hemoglobin (HbR) imaging (used to produce HbO₂ map) in the mouse prefrontal cortex (PFC). b_i-b_ii Representative images of GCaMP6f and DREADD (Gq) expression in the PFC. b_iii Representative images of overlapping GCaMP6f Ca²⁺ and DREADD (Gq) expression. b_iv Representative HbT image including arteries (red dashed line) and veins (blue dashed line). c Graphical representation of dual viral injection of GFAP-DREADD(Gq) with c_i Cre-dependent GCaMP6f into GFAP-Cre mice for astrocytes c_ii synaptophysin GCaMP6f into wild type mice for neurons. d Timeline of the experimental imaging protocol, including 1 st cocaine, astrocytic activation via DREADD (Gq) and 2nd cocaine challenge.

Cranial window implantation

To image brain function in the PFC, an optically clear cranial window (~ 2 × 2 mm²) was created on the skull over the viral injection spot. The mouse skull surrounding the viral injection spot was carefully thinned with a dental drill and then removed to expose the brain. Saline was applied to the brain tissue, and the open cortex was immediately covered with a sterile coverslip (4 × 3 mm²) and sealed with biocompatible glue. Dental cement was spread around the edges of the coverslip to further secure its attachment. After window implantation, a catheter was placed in a tail vein for drug administration during imaging. Then the animal was positioned on a custom head mount and moved to the imaging stage for in-vivo image acquisition. All surgical procedures were performed under isoflurane anesthesia.

In-vivo MIP imaging

A custom multimodality imaging platform (MIP, Fig. 1a) developed in our lab^25,35–37 was used for simultaneous imaging of astrocytic/neuronal calcium fluorescence (ΔCa²⁺_A, ΔCa²⁺_N, respectively), along with oxygenated hemoglobin (Δ[HbO₂]) from the mouse PFC in response to cocaine without or with DREADD (Gq) activation. As shown in Fig. 1, light beams at three wavelengths (488 nm, 568 nm and 630 nm) were delivered to the PFC with a time-sharing mode (10ms exposure time per channel). For GCaMP6f-Ca²⁺ fluorescence imaging, a blue LED at λ_ex = 488 nm was used for excitation and the fluorescence images are shown in (Fig. 1b_i). The DREADD(Gq) expression image (Fig. 1b_ii) was captured with λ_ex = 550 nm, where the emission filter EM2 (≥ 570 nm) was used. Images of the merged Ca²⁺ and DREADD(Gq) signals are shown in Fig. 1b_iii. The reflected images captured from λ₁ = 568 nm, and λ₂ = 630 nm light channels were used to distinguish arteries and veins and create tissue oxygenation hemoglobin map (Fig. 1b_iv). These images were detected with a sCMOS camera (synchronized with the illumination paradigm). To assess cocaine-induced dynamic changes in astrocytic and neuronal Ca²⁺ fluorescence, brain hemodynamics and astrocyte’s involvement in cocaine induced vasoconstriction, animals underwent three imaging sessions per experiment (Fig. 1d): d₁: 1 st cocaine injection (1 mg/kg, i.v. 70 min); d₂: clozapine pretreatment (0.1 mg/kg, i.p. 40 min)/saline pretreatment for control group; d₃: 2nd cocaine injection 40 min following clozapine pre-treatment (1 mg/kg, i.v. 70 min).

Image and data processing

For each experiment, three data sets were obtained from each mouse cortex: (1) [Ca²⁺] dependent fluorescence intensity changes from astrocytes (ΔCa²⁺_A) or neurons (ΔCa²⁺_N); (2) changes in vessel diameters (Δ[Φ]); (3) changes in oxygenated-hemoglobin tissue (Δ[HbO₂]). These data sets were collected without and with DREADD (Gq) activation by clozapine in response to acute cocaine challenge. Data across animals was averaged, basal and integrated reactivity response to cocaine were quantified. Besides that, to assess the impact of astrocyte activation on cocaine effects, the data was also normalized to the basal activity prior to the first cocaine injection (10 min preceding first cocaine injection).

GCaMP6f-Ca²⁺ fluorescence was selected within GCaMP6f expressing cortical areas devoid of visible blood vessels, five regions of interest (ROIs shown as yellow circle in Fig. 3a, b for astrocyte and Fig. 6a, b for neuron) selected for each animal. The astrocytic or neuronal ΔCa²⁺ changes were quantified as relative changes of fluorescence signal over their baseline, i.e., Δ[Ca²⁺]=[F(t)-F(baseline)/F(baseline)]x100%.

Fig. 3.

Cocaine-induced ΔCa²⁺ _A before and after DREADD(Gq) activation. a Representative images showing astrocytic DREADD(Gq) expression and selected ROIs. b Representative astrocytic Ca²⁺_A fluorescence ratio images showing the dynamic response to cocaine (1 mg/kg, i.v.) in PFC without (upper panel) and with DREADD(Gq) (lower panel); b_i-iii Ratio images at baseline (t=0 min), t=5 min and t = 40 min after cocaine administration. c Averaged dynamic change in ΔCa²⁺_A triggered by acute cocaine without (Coc) and with (Gq + Coc) Gq activation. d Normalized time course of change in ΔCa²⁺_A in response to first and second cocaine injection. e Quantification of Ca²⁺_A mean integration changes triggered by cocaine between Coc (shadow in orange) and Gq + Coc (shadow in blue), showed a significant attenuation when Gq was active (*p = 0.032). f Comparison of normalized ΔCa²⁺_A baseline between 1st -Coc (shadow in orange) and 2nd -Coc (shadow in blue) showing a significantly larger (*p < 0.001) increase for the 2nd -Coc. g–h Binary maps and pixel fraction analysis illustrating decreased spatial distribution of active astrocytic regions following clozapine to activate astrocyte (*p = 0.046). Values in graphs correspond to mean and standard errors.

Fig. 6.

Cocaine-induced Ca²⁺ _N dependent fluorescence changes before and after GFAP-DREADD(Gq) activation. a Representative image showing area with DREADD(Gq) expression and location of selected ROIs. b Representative neuronal Ca²⁺_N fluorescence images showing the dynamic response to cocaine (1 mg/kg, i.v.) without (upper panel) and with DREADD(Gq) (lower panel); c Averaged dynamic change in ΔCa²⁺_N triggered by acute cocaine without (Coc) and with (Gq + Coc) Gq activation. d Normalized time course of change in ΔCa²⁺_N in response to first and second cocaine injections. e Quantification of Ca²⁺_N mean integration changes triggered by cocaine between Coc (shadow in orange) and Gq + Coc (shadow in blue), did not differ.). f Comparison of normalized ΔCa²⁺_N baseline between 1 st -Coc (shadow in orange) and 2nd -Coc (shadow in blue), did not differ. g, h Binary maps and pixel fraction analysis illustrating no significant changes in the active neuronal area following clozapine. Values in graphs correspond to mean and standard errors.

The hemodynamic response Δ[HbO₂] were extracted from the reflected images. As illustrated in Figs. 6 and 7, six ROIs were selected from each animal and hemodynamic changes were calculated based on the following equation:

Fig. 7.

Comparison of cocaine-induced changes in tissue oxygenated hemoglobin Δ[HbO₂] before and after DREADD(Gq) activation. a Cortical image with ROIs location. For each animal, six ROIs were selected within the expression area. b Ratio maps of oxygenated hemoglobin (Δ[HbO₂]) before (t₁= 0 min), and at 5 and 40 min after cocaine without DREADD(Gq) (upper panel) and with DREADD(Gq) (lower panel). c Mean changes of Δ[HbO₂] triggered by cocaine without (Coc) and with (Gq + Coc) Gq activation. d Comparison of the integrated Δ[HbO₂] changes triggered by cocaine with Coc (shadow in orange) and Gq + Coc (shadow in blue), showed a significant attenuation with Gq activation (*p = 0.004). e Comparison of normalized Δ[HbO₂] baseline between 1st -Coc (shadow in orange) and 2nd -Coc (shadow in blue) showing no significant difference.

1

where ε^λ_HbO2, ε^λ_HbR are the molar extinction coefficients for HbO₂ and HbR, and R_λ1(t), R_λ2(t) are the measured diffuse reflectance at these wavelengths. R_λ1(0), R_λ2(0) are their baseline values before cocaine. L_λ1(t), L_λ2(t) are their pathlengths. The total hemoglobin concentration change can be obtained by

2

which corresponds to the cerebral blood volume change³⁶.

To assess cocaine-induced vasoconstriction and the involvement of astrocytic activation in vasoconstriction, the vessel diameters were quantified using a custom Matlab program. About 6–10 ROIs were selected from blood vessels of different sizes to monitor their diameter changes from baseline to 60 min after cocaine administration. Vessel diameter measures were performed on images acquired at λ1 = 568 nm using a Gaussian process method published by Asl et al.^33,38. Specifically, x and y coordinates of a seed point placed at the center of a vessel of the selected ROI. The centerline and direction of the vessel are then computed using the Radon transform, and trained MATLAB to start tracking from the input seed point to the next 10 points (ϕ_i, i = 1–10, within ~ 60 μm) lying in the center of the blood vessel. The diameter ϕ_i at each point is then estimated using a separate Gaussian process. The mean volume of ϕ_i (i = 1–10) was calculated as the vessel diameters at the location selected ROIs. To track the time course of vessel diameter changes, a while-loop was added to read from the first image to the last one (e.g., m = 4200 frames or based on our imaging time for each experiment above), which allowed us to continually track each frame and its change was quantified over its baseline (before cocaine) as a function of time following cocaine injection.

To identify whether the spatial distribution of cocaine-induced ΔCa²⁺_A or ΔCa²⁺_N changes was affected by astrocytic activation we developed a custom (Python-based) image analysis program to quantify the astrocytic and neuronal Ca²⁺ responses before and after DREADD(Gq) activation. The flowchart of the process is illustrated in Figure S1. Briefly, ratio maps were computed as peak divided by their baseline to normalize fluorescence intensity changes across the entire field of view. A fixed threshold (no-response area as cutoff) was then applied to binarize each map, and the fraction of pixels above the threshold was calculated to represent the responsive area.

Statistics

Results are reported as mean ± standard error (SEM). Comparisons between the two groups (e.g., before vs. after) were analyzed using Student’s T-tests and comparisons between the four groups with One-way ANOVA followed by Tukey’s HSD post hoc. Pearson correlation analyses were conducted to determine correlation coefficients (R values). Two-way repeated measure ANOVA with time and treatment factors was used to identify periods showing statistically significant differences between groups (e.g., with or without Gq activation). All statistical tests were performed using SigmaStat software (Version 12.5; Systat Software Inc.), with P < 0.05 considered significant. The sample sizes for each experiment are summarized in Table 1.

Results

To investigate whether astrocytic activation could alter the cellular and vascular response to cocaine, we recorded Ca²⁺_A/Ca²⁺_N fluorescence, vascular diameters, and hemodynamic changes in response to acute cocaine with and without DREADD(Gq) activation. The magnitude of the response was calculated by integrating the time course of parameter change until its recovery. For comparison purposes the second cocaine response was normalized to the baseline activity prior to the first cocaine injection. To assess the sensitivity of the hemodynamic and cellular signals to our intervention of astrocytic activation we compared ROIs in regions within or without DREADD (Gq) expression (Supplemental figure S2,3,4). All the results presented below are based on ROIs selected within the DREADD expression region.

Activation of DREADD(Gq) elevates astrocytic Ca²⁺_A and induces long-lasting vasoconstriction with a minimal influence on neuronal or hemodynamic changes

DREADDs is a chemogenic approach that enables subtype selective activation (Gq) or silencing (Gi) of cellular signaling (e.g., astrocytes) via clozapine activation³⁹. We used DREADD(Gq) to determine the impact of astrocytic activation on cellular and vascular activity in the mouse PFC. Figure 2 shows time-course responses of cellular and vascular activities acquired before and after clozapine injection (0.1 mg/kg, i.p.) at t = 0 min followed over 30 min. Figure 2a is a spatial ratio map post clozapine (t=25 min) normalized to its baseline (t=−5 min), revealing increases in astrocytic Ca²⁺_A (Fig. 2a_ii) localized within the Gq expression region (Fig. 2a_i) along with vasoconstriction (Fig. 2a_iii). Figure 2c summarizes the mean changes in Ca²⁺_A as a function of time post clozapine administration across animals (n = 6). DREADD(Gq) activation resulted in Ca²⁺_A elevation compared to baseline (1.325 ± 0.5689% from t = 11–30 min, P* < 0.05). Figure 2d shows mean changes in vascular diameters, revealing that DREADD(Gq) induced long-lasting vasoconstriction, with a decrease of (−3.541 ± 0.349%) over its baseline (−0.2221 ± 0.9649%). We also evaluated whether astrocytic activation would affect neuronal and hemodynamic activity. Figure 2b_ii shows a representative ratio image of ΔCa²⁺_N fluorescence at t = 25 min after DREADD(Gq) activation over the baseline (t = − 5 min), exhibiting no significant Ca²⁺_N changes (−1.232 ± 1.535%) after DREADD(Gq) activation (Fig. 2e). Meanwhile, the time course of hemodynamic ΔHbO₂ in response to astrocyte activation is shown in Fig. 2f, revealing ΔHbO₂ remained stable throughout the DREADD (Gq) activation, (−1.668 ± 0.5158%) at t = 25 post clozapine. Taken together, these results indicated that excitation of astrocytes via clozapine activation of DREADDs (Gq) did not influence baseline neuronal activity (Ca²⁺_N) nor ΔHbO₂, but induced elevation of astrocytes Ca²⁺_A and vasoconstriction.

Fig. 2.

Effects of DREADDS(Gq) activation on neuronal, astrocytic and vascular responses. a, b Ratio images of Ca²⁺_A and Ca²⁺_N fluorescence, blood vessel diameters and hemodynamic changes at t = 25 min after DREADD(Gq) activation compared to their baselines at t = − 5 min, indicating Ca²⁺_A fluorescence increased (red area in a_ii) and blood vessels vasoconstricted (blue edges in a_iii), but there were no effects on Ca²⁺_N (in b_ii) nor ΔHbO₂ (in b_iii); c–f Mean time courses of Ca²⁺_A increase in c, ∆ϕ decreases in d, but no significant Ca²⁺_N change in e, nor in Δ[HbO₂] in f after DREADD(Gq) activation.

Activation of astrocytes increases basal Ca²⁺_A limiting further Ca²⁺_A increases from cocaine

Cocaine triggered significant changes in Ca²⁺_A that differed with or without astrocytic activation. Figure 3 shows Ca²⁺ dependent fluorescence changes in response to a cocaine challenge in astrocytes without (Coc) and with (Gq + Coc) astrocytic activation. The ratio images of cocaine-induced ΔCa²⁺_A-Coc and ΔCa²⁺_A-Gq + Coc at baseline (t=0 min), 5 min, and 40 min after injection are illustrated in Fig. 3b (upper and lower panels, respectively). Figure 3a is used to confirm the ROIs were selected within the DREADD (Gq) expression area. As shown in supplemental Figure S2, activation of DREADD (Gq) increased astrocytic Ca²⁺ activity only in the expressing region. Figure 3c shows the time courses of cocaine induced Ca²⁺_A changes (n = 6) without (orange line) and with (blue line) astrocytic activation. Without astrocytic activation cocaine increased ΔCa²⁺_A (2.072 ± 0.6823%) from t = 0 to t = 4 min (P*<0.05) returning to baseline at ~ 34 min; while following DREADD(Gq) activation, there was no further increase in ΔCa²⁺_A (n = 6). Post-hoc analysis of a two-way repeated measure ANOVA showed differences between the two-time courses from baseline (t = 0) to 2 min post-cocaine (F (70,280) = 12.718, P*<0.001, Fig. 3c) indicating that after DREADD(Gq) activation further cocaine induced Ca²⁺_A increases were curtailed. The mean integrated response of ΔCa²⁺_A increase with cocaine was significantly smaller after Gq activation (Coc vs. Gq + Coc: 0.82 ± 0.07%/min vs. 0.04 ± 0.04%/min, P*=0.032, Fig. 3e) presumably reflecting a ceiling effect as ΔCa²⁺_A was already increased by the Gq activation. Meanwhile, when normalizing to the baseline prior to the first cocaine injection (1st -Coc at t = 0; Fig. 3d), the ΔCa²⁺_A of the second cocaine was higher than the 1 st -Coc from the beginning until the end of experiment. This likely reflected the fact that the 2nd -Coc was given when astrocytes were still activated resulting in a significantly higher Ca²⁺_A baseline than prior to the 1st cocaine injection (1st vs. 2nd : 0.1704 ± 0.0580% vs.1.311 ± 0.0306%, Fig. 3f, P*< 0.001). We interpret this to suggest that ΔCa²⁺_A increase associated with DREADD(Gq) activation created an astrocytic “ceiling effect” curtailing the magnitude of further ΔCa²⁺_A in response to cocaine. As shown in Fig. 3g, the spatial distribution of astrocytic Ca²⁺_A responses before and after DREADD(Gq) activation normalized to the baseline that preceded each cocaine injection indicated that the fraction of activated pixels was significantly decreased (before vs. after: 10.61 ± 3.426% vs. 3.422 ± 1.320%, P*=0.046, Fig. 3h) after DREADD(Gq) activation.

Activation of astrocytes decreases vascular diameters limiting further vasoconstriction by cocaine

Both astrocytic activation and acute cocaine decreased vascular diameters (Φ). However, when cocaine was given after astrocytic activation, the cerebral vessels were already constricted, thus further responses were limited. Figure 4a-b shows representative vascular images obtained from the multi-wavelength platform and their corresponding diameter maps without (Fig. 4a) and with (Fig. 4b) astrocytic activation at t = 0 and at t=5 min after cocaine administration. ROIs are selected within the DREADD(Gq) expression area, as shown in Supplemental S3. Without DREADD(Gq) activation, acute cocaine induced a persistent decrease in vascular diameter (ΔΦ-Coc, −5.464 ± 1.80%, t = 4–13 min P*<0.05) (orange line in Fig. 4e). Whereas, with astrocytic activation, cocaine’s effect was significantly reduced compared to its own baseline (ΔΦ-Gq + Coc, blue line in Fig. 4e). Comparison between mean time course of ΔΦ-Coc and ΔΦ-Gq+ Coc with the two-way repeated measure ANOVA showed a significant difference from t = 3 to 29 min post cocaine [F(70, 350) = 1.745; P*<0.001]. The mean integrated changes of vessel constriction were also decreased (from − 5.47 ± 1.30%/min to −1.25 ± 1.02%/min) when cocaine was given after Gq activation (Fig. 4g, P*= 0.037). We separated blood vessels into three groups based on their diameters: large vessels Φ > 60 μm, medium vessels Φ = 40–60 μm and small vessels Φ < 40 μm. Without DREADD(Gq) activation, acute cocaine induced a significant reduction in all three vessel sizes (Fig. 4c, P*< 0.05), especially in microvessels, which decreased to (−5.146 ± 1.119%). With DREADD(Gq) activation, all types of vessels had an attenuated response (Fig. 4d) compared to preceding each cocaine injection. This likely reflected a ceiling effect due to the pre-existing strong vasoconstriction from astrocytic activation, as baselines prior to 1st -Coc (−0.537 ± 0.6218%) and 2nd -Coc (−5.375 ± 0.5613%), differed significantly (Fig. 4h, P*<0.001). Indeed, after normalizing ΔΦ from cocaine to the baseline prior to the first cocaine administration (Fig. 4f), the analyses showed that after Gq activation, the vessels were already significantly constricted (−6.029 ± 1.466%) blue line in Fig. 4f prior to the second cocaine challenge and remained markedly vasoconstricted after the second cocaine injection to an even greater extent observed from cocaine without Gq activation. These vascular effects are not attributable to clozapine injection as previous studies showed it has no impact on vascular diameters^19,33.

Fig. 4.

Comparison of cocaine-induced vascular diameter changes (ΔΦ) before and after GFAP-DREADD(Gq) activation. Quantitative maps demonstrate the cerebrovascular diameter change ΔΦ in response to cocaine a before and b after DREADD(Gq) activation by clozapine. Adjacent images are skeletonization of associated vasculature with changes in vessel diameter encoded into the skeletal pixel color. c, d Averaged ΔΦ triggered by cocaine before (left panel, c) and after (right panel, d) DREADD(Gq) activation separated into large, medium and small vessels. (red: large, blue: medium and green: small), showing cocaine-induced vasoconstriction has been minimized in vascular tree after Gq activation. e Averaged dynamic ΔΦ triggered by cocaine without (Coc) and with (Gq + Coc) DREADD(Gq) activation, showing the attenuation of ΔΦ reductions with Gq activation. g Comparison between accumulated ΔΦ changes triggered by cocaine between Coc (shadow in orange) and Gq + Coc (shadow in blue) has a significant reduction (*p = 0.037). f Averaged normalized ΔΦ induced by cocaine before (1st -Coc) and after (2nd -Coc) Gq activation and h comparison of basal vascular diameter between 1 st -Coc (shadow in orange) and 2nd -Coc (shadow in blue) showing a significant baseline decrease in 2nd -Coc (*p < 0.001).

To investigate if the limited vasoconstriction induced by cocaine following DREADD(Gq) activation reflected sustained vasoconstriction from astrocytic activation, we tracked the vascular diameters in saline or clozapine pre-treated mice. Figure 5a shows the diameter changes in the saline control group in whom the 1st cocaine induced a significant decrease to (−6.869 ± 0.4029%) with gradual recovery to baseline. There were no significant changes in vascular diameter following a saline injection. When the 2nd cocaine challenge was administered (Group 3: control group), vascular diameters again decreased to (−6.563 ± 1.357%), with a gradual recovery indicating that there was no tolerance to the vasoconstricting effects of cocaine after the first injection (Fig. 5a). There were no differences between the two cocaine injections (Fig. 5c, left). In contrast, Fig. 5b shows the diameter decreases in response to the 1st cocaine (−5.464 ± 1.80%) and its gradual recovery to baseline. However, after Gq activation the diameters decreased significantly (−3.541 ± 0.349%) in contrast to saline (0.489 ± 0.165%) (P*< 0.001; Figs. 5d). Thus, normalization of diameter changes preceding the second cocaine injection showed an attenuated response (−1.212 ± 0.443%), compared to the first cocaine injection (−5.464 ± 1.80%, P*=0.006) (Fig. 5c, right) as the vessels were already constricted at the start of the 2nd cocaine injection. Taken together, this data supports astrocytic activation induced vessel constriction that may have reached the vasoconstrictive potential of the vasculature preventing further responses to the 2nd cocaine injection.

Fig. 5.

Dynamic vascular diameter changes (ΔΦ) with cocaine before, during and after GFAP-DREADD(Gq) activation. a Upper panel: Mean time course of vascular diameter changes (ΔΦ) in response to cocaine across three experimental periods: 1 st cocaine – orange; saline administration -green followed by 2nd cocaine-blue. b Lower panel: Mean time course of cocaine induced changes in vascular diameter (ΔΦ) during three experimental period: 1 st cocaine administration– orange; during DREADD (Gq) activation -purple and 2nd cocaine administration -blue, basal comparison between 1 st and 2nd cocaine administration showing further constricting during activating Gq. c Comparison of vascular diameter changes triggered by cocaine before and after saline versus clozapine administration indicates that Gq activation by clozapine significantly attenuates the ΔΦ response during the 2nd cocaine administration. d Vascular diameter changes induced by saline (green) and clozapine (purple) showed that during clozapine-induced Gq activation significantly constricts vessels compared to saline (*p < 0.001).

Activation of astrocytes does not modulate neuronal response to cocaine

Cocaine triggered significant changes in neuronal ΔCa²⁺_N that were not affected by astrocytic activation. Figure 6 shows Ca²⁺ dependent fluorescence changes in neurons before (ΔCa²⁺_N-Coc) and after (ΔCa²⁺_N-Gq + Coc) astrocytic DREADD(Gq) activation induced by acute cocaine (1 mg/kg, i.v). Figure 6a illustrates Ca_N²⁺ overlapping with DREADD (Gq) expression to ensure that the ROIs selected where within the DREADD (Gq) expression area as outside the Ca²⁺ signals were very weak (Supplemental Figure S4). Figure 6b illustrate the ratio images of cocaine-induced ΔCa²⁺_N with and without DREADD(Gq) activation at baseline (t = 0 min), t = 5 min, and t = 40 min both demonstrating a rapid influx of calcium (Fig. 6b). The mean time courses of cocaine induced ΔCa²⁺_N (n = 6, Fig. 6c) without and with clozapine show a calcium influx of (Coc vs. Gq + Coc: 4.83 ± 1.16% and 5.58 ± 1.38%, respectively). In both conditions there was a gradual recovery to baseline at ~ 50 min. A two-way repeat ANOVA showed no difference between these two-time courses [F(70, 350) = 0.281 P = 1.000]. There were also no significant differences between the mean integrated responses of ΔCa²⁺_N following cocaine injection (Coc vs. Gq + Coc: 2.704 ± 0.6604% vs. 3.062 ± 0.7618% P = 0.74, Fig. 6e). These results indicate that astrocytic activation did not affect neuronal response to cocaine. The normalized time courses of ΔCa²⁺_N changes in response to cocaine without (1st -Coc) or with Gq activation (2nd -Coc), showed similar increases that remained above baseline at t = 50 min and no differences in the preceding baselines (1st vs. 2nd : 0.00879 ± 0.00294% VS −0.0661 ± 0.143% Fig. 6f, P = 0.656). The spatial distribution analyses (Fig. 6g) of the Ca²⁺_N responses before and after DREADD(Gq) activation indicated that the activated pixels fraction was unchanged by DREADD(Gq) (before vs. after: 10.01 ± 1.685% vs. 9.382 ± 1.859%, P = 0.4, Fig. 6h).

Activation of astrocytes modulates cocaine-induced decreases in cortical hemoglobin (ΔHbO₂)

Figure 7a displays grayscale images to illustrate the location of ROIs and the ratio images at baseline (t=0 min), t=5 min and t=40 min after cocaine injection without (Fig. 7b, upper) and with DREADD(Gq) (Fig. 7b below). We compared the ROIs located inside the DREADD(Gq)expression area from those outside (Supplemental Figure S5), which showed that after Gq activation, the tissue oxygenation changes attenuated in response to cocaine inside the DREADD-expressing region only. The first cocaine injection induced a persistent decrease in Δ[HbO₂] with a maximum of (−15.27 ± 0.816%) (Fig. 7c) but following astrocytic activation, the reduction from the second cocaine was smaller (−5.904 ± 1.769%). Comparisons of the Δ[HbO₂] integrated response to cocaine with or without Gq activation differed significantly (Coc vs. Gq + Coc: −9.024 ± 0.8386%/min vs. −4.564 ± 0.6059%/min, P*=0.004 Fig. 7d). The normalized Δ[HbO₂] baselines did not differ (1st vs. 2nd : 0.0539 ± 0.0082% vs. 0.00731 ± 0.0508%, P = 0.063, Fig. 7e), indicating that DREADD(Gq) astrocytic activation did not change tissue Δ[HbO₂] (Fig. 2f). We interpret the reduction in Δ[HbO₂] despite cocaine induced neuronal activation, which should have led to an increase in blood flow to the area to indicate disruption of NVC that we ascribe to cocaine’s vasoconstricting effects^7,14.

Cocaine induced vasoconstriction ΔΦ was associated with astrocytic Ca²⁺_A

To elucidate the contribution of Ca²⁺_A and Ca²⁺_N to cocaine-induced hemodynamic changes, we studied the correlation between ΔCa²⁺, vessel diameters ΔΦ, and HbO₂ (Fig. 8) without and with astrocytic activation. Figure 8a illustrates the temporal dynamics of neuronal and astrocytic Ca²⁺, vascular diameter (Φ), and hemodynamic (Δ[HbO₂]) signals before DREADD(Gq) activation. Before DREADD activation, cocaine increased ΔCa²⁺_A and ΔCa²⁺_N, and decreased vessel diameter and Δ[HbO₂]. The temporal correlation between ΔCa²⁺_A and ΔCa²⁺_N with ΔΦ or Δ Δ[HbO₂], prior to DREADD(Gq) activation is summarized in Fig. 8c, indicating that astrocytes had a stronger correlation with vessel diameters than neurons (Ast vs. Neu: −0. 543 vs. −0.018, P*<0.001). On the other hand, the correlation of ΔCa²⁺_A with Δ[HbO₂] was significantly weaker than for ΔCa²⁺_N (Ast vs. Neu: −0.231 vs. −0.673, P*=0.012). Figure 8b shows the temporal dynamics of neuronal and astrocytic Ca²⁺, vascular diameter (Φ), and hemodynamic (Δ[HbO₂]) signals after DREADD(Gq) activation. Following astrocytic activation by clozapine, further cocaine induced increases in astrocytic Ca²⁺, in vasoconstriction and decrease in Δ[HbO₂] were attenuated. The correlations between ΔCa²⁺_A and ΔCa²⁺_N with ΔΦ were not significant after DREADD(Gq) activation as shown in Fig. 8d; though the stronger correlation between ΔCa²⁺_N and Δ[HbO₂] than for ΔCa²⁺_A (Ast vs. Neu: −0.334 vs. −0.701, P*=0.011). This suggests that while ΔHbO₂ is influenced by both astrocyte and neuron the latter has a larger influence.

Fig. 8.

Evaluation of the correlation relationship between vascular parameters and astrocytic Ca²⁺_A and neuronal Ca²⁺ _N. a, b Averaged traces of ΔCa²⁺_N (red), ΔCa²⁺_A (green), vascular diameter ΔΦ (purple), and oxygenated hemoglobin ΔHbO₂ (blue) signals before (a) and after (b) DREADD(Gq) activation. c Comparison of correlation coefficients before DREADD (Gq) of ΔCa²⁺_A (green bar) and ΔCa²⁺_N (pink bar) with ΔΦ and Δ[HbO₂] changes, respectively. d Comparison of correlation coefficients after DREADD (Gq) activation of ΔCa²⁺_A and ΔCa²⁺_N with ΔΦ and Δ[HbO₂] changes, respectively.

Discussion

In this study, astrocytes were selectively activated by chemogenetic manipulation with simultaneous recording of cerebral hemodynamics and cellular activity in response to acute cocaine. We found that without astrocytic activation, cocaine induced an increase in both neuronal and astrocytic ΔCa²⁺ along with vasoconstriction and a decrease in hemodynamics. However, when astrocytes were activated by DREADD(Gq), which by itself enhanced ΔCa²⁺_A and triggered vasoconstriction, curtailed further increases in ΔCa²⁺_A and vasoconstriction by cocaine whereas its ΔCa²⁺_N increases were unaffected. The correlation between astrocytes and vasoconstriction was disrupted by DREADD(Gq) while the correlation between astrocytes and neurons with Δ[HbO₂] was not affected.

These findings further support that cocaine disrupts NVC and that astrocyte plays an important role in regulating cocaine-induced changes in vascular tone. Specifically, that cocaine, which triggered neuronal activation alongside vasoconstriction interfered with the typical increases that would be expected from neuronal activation, resulting in neurovascular uncoupling. Thus, our findings suggest that cocaine’s activation of astrocytes may trigger vasoconstriction^40,41 and disrupt normal NVC. However, further investigation might need to distinguish these effects from cocaine’s direct effects on the cerebrovascular mural cells. Also, cocaine’s blockade of monoamine reuptake can potentiates sympathetic vasoconstriction and restricts blood flow even during heightened neuronal activity⁷.

DREADDs modulate astrocytes

Astrocytes are increasingly recognized for their multifaceted roles in brain function^16,42. However, the ability to study how neurons and glial cells communicate has been limited by the few tools available for targeted manipulation^2,43. In this respect chemogenetic tools (DREADDs) have allowed cell selective and spatiotemporal precise modulation of glial cells^32,44. Excitatory hM3Dq DREADDs stimulate astrocytes through phospholipase C (PLC) and inositol trisphosphate (IP₃) leading to intracellular Ca²⁺ release³¹. Inhibitory DREADD(Gi) suppress astrocytes by reducing cyclic adenosine monophosphate (cAMP) signaling, which regulates neurotransmitter release, synaptic plasticity, and neuroinflammation^45,46. Recently, the effect of DREADD(Gq) in astrocytic Ca²⁺ signaling was assessed for the first time in awake mice⁴⁷, documenting long-lasting increases. Here we observed a similar response in anesthetized mice which showed that DREADD(Gq), resulted in a ~ 1.4% Ca²⁺_A increase (Fig. 2c).

Manipulation of astrocytes with DREADDs has raised concern that this artificial manipulation may create conditions that do not reflect natural astrocytic activity. DREADDs themselves are modified G-coupled protein receptors and utilize the same endogenous signaling cascades activated in normal function. Importantly, the amplitude and temporal dynamics of DREADD(Gq)-evoked calcium responses have been shown to be comparable to those from endogenous Gq-GPCR activation^31,47–49.

Astrocytes activated by DREADD(Gq) do not affect neuronal response to cocaine

Astrocytes communicate with neurons modulating neuronal activity through neurotransmitter uptake and gliotransmitter release^50,51. Upon activation via DREADD(Gq), astrocytes can release glutamate that can then stimulate NMDA and AMPA receptors on neurons increasing their excitability^48,52. We previously demonstrated that the partial NMDA antagonist memantine attenuated cocaine-induced intracellular neuronal Ca²⁺ increases, significantly reducing both ΔCa²⁺ dynamics and hemodynamic changes in the rat PFC⁵³. Interestingly, in the current study, Gq activation did not alter the neuronal response to cocaine. There remains controversy as to whether activation of astrocytes through DREADD(Gq) impacts neuronal function. In the hippocampus, some studies have documented that neither the loss of nor the activation of astrocytic Gq protein receptors in the CA1 region leads to changes in excitation of pyramidal neurons^54,55. Conversely, others have reported that activation of astrocytes through Gq leads to an increase in excitatory potential or elevated neuronal activity of single cells observed by electrophysiology recording or 2-photon imaging^56,57. For our study the neuronal ΔCa²⁺ increases from cocaine were robust (~ 6%) and persisted over 40 min similar to results reported by other studies^58,59. Therefore, if a small change in neuronal activity (~ 1%) occurred when activating DREADD(Gq) (Fig. 2e), these changes may be hidden by the larger change from cocaine. Additionally, the studies to date reporting on neuronal activity focus on single cells using electrophysiology or 2-photon microscopy^60–62. These single cell observations may not generate a signal that is large enough to resolve at the tissue level with our methodology.

Cocaine induced vasoconstriction is distinct from cortical spreading depression

Given the large neuronal calcium shift induced by cocaine, one possibility is that this reflects cortical spreading depression (CSD). However, the vascular and neuronal responses we observed are distinct from those reported with CSD. Our group has examined cocaine’s effects on cortical field potentials using electroencephalography in the rat cortex⁶³. Unlike CSD, which is characterized by a propagating wave of intense depolarization followed by prolonged suppression of evoked potentials lasting up to 30 min⁶⁴, spikes in local field potentials (LFPs) in association with forepaw stimulation were observed even after cocaine administration (1 mg/kg i.v.), consistent with preserved neuronal responsiveness rather than post-SD depression. Also, optical imaging studies of CSD with GCaMP6f show a propagating wavefront of calcium activation across cortical tissue⁶⁵. In contrast, our experiments revealed a simultaneous, spatially uniform activation across the 2 × 2 mm² field followed by a slow, spatially uniform decay^66,67. This pattern is inconsistent with the spatiotemporal hallmark of CSD. Finally, CSD typically produces a biphasic vascular response, with initial transient hyperemia followed by prolonged vasoconstriction⁶⁴. While we observed sustained vasoconstriction after cocaine, we did not detect the preceding hyperemia in prior⁶⁶ nor current experiments. Together, these findings suggest a different mechanism for cocaine-induced vasoconstriction than CSD.

Activated astrocytes by DREADD(Gq) constrict blood vessels

Astrocyte activation modulates vasoconstriction and NVC through the release of vasoactive substances⁵⁹. Our group has documented the pivotal role astrocytes play in controlling vascular tone in response to different stimuli²⁵. Specifically, during electrical forepaw stimulation, we observed a gradual increase in astrocytic Ca²⁺ levels that peaked at approximately 4.7% and that was strongly correlated (R = 0.66) with the vasoconstriction that followed the hyperemic response, suggesting a role of astrocytes in resetting vascular tone post-vasodilation⁶⁸. In our current study, the activation of astrocytes via DREADD(Gq) led to an increase in astrocytic Ca²⁺ of 1.4% that was temporally associated with an average vasoconstriction of 3.5% (Fig. 2). Astrocytes regulate vascular tone through the release of substances that act on vascular smooth muscle cells, such as phospholipase A2 (PLA₂), arachidonic acid (AA), and the intermediate 20-hydroxyeicosatetraenoic acid (20-HETE)^69–71. Notably, 20-HETE is a potential vasoconstrictor⁷² that reduces blood vessel diameter and works through Ca²⁺channels such as the L-type voltage gated Ca²⁺ channels on surrounding smooth muscle-cells⁷³. Consistent with this mechanism, we previously demonstrated that the L-type Ca²⁺ channel blocker, nifedipine, attenuated vasoconstriction in rodents following cocaine administration^66,72.

Comparison of DREADD(Gq) and DREADD(Gi) effect on cocaine response

Several studies have used chemogenetic receptors (DREADDs) to manipulate cocaine-induced effects in mice^74,75. Recent publications^19,33 in our lab have focused on the inhibition of astrocytes using DREADD(Gi) and its impact on NVC and on cocaine-induced vasoconstriction. Inhibition of astrocytic activity led to a drop in astrocytic ΔCa²⁺ of ~ −3% and up to a ~ 9% increase in vascular diameters. When astrocytes were inhibited, cocaine-induced astrocytic ΔCa²⁺ rise and vasoconstriction were significantly attenuated while ΔCBFv changes were eliminated. Conversely, here we show that during astrocytic activation by DREADD(Gq), there was an increase in astrocytic ΔCa²⁺ of 1.4% and vasoconstriction of 3.5% (Fig. 2). Though, DREADD (Gq), attenuated the astrocytic ΔCa²⁺ and vasoconstriction ΔΦ in response to cocaine this likely reflected a ceiling effect since astrocytes were already activated, and vessels were constricted when second cocaine was given. The control experiment with saline injection also indicates that the prolonged vasoconstriction was not due to remaining effects from the first cocaine injection but instead the result of astrocytic activation.

Clozapine instead of CNO was used to active DREADDs, because a recent study showed that clozapine (the clozapine-N-oxide metabolite) rather than CNO itself stimulates the DREADDS receptor³⁹. The clozapine dose (0.1 mg/kg) was within the smaller range used in chemogenetic studies⁷⁶. We previously studied the effects of clozapine by itself in control experiments (no DREADD expression) and observed that clozapine did not affect neuronal activities nor CBF³³. Thus, neither neuronal activity nor baseline hemodynamics were affected by clozapine, and cocaine-induced vascular responses remained intact, suggesting minimal off-target effects of clozapine itself.

Astrocytic Ca²⁺ and neuronal Ca²⁺ drive separate components of hemodynamics and are disrupted by DREADD(Gq) activation

Astrocytes contribute to cocaine’s vasoconstrictive effects in cerebral blood vessels^14,77. We had previously reported a long lasting ΔCa²⁺_A increase with acute cocaine that was associated with cocaine-induced vasoconstriction^19,33. Our current findings are consistent with our previous work. However, this association was disrupted, following activation of astrocytes via DREADD(Gq), likely due to a sustained elevation in astrocytic calcium; suggesting that cocaine was unable to further increase ΔCa²⁺_A under these elevated conditions. The mechanism underlying cocaine-induced increases in Ca²⁺_A is unclear but it could reflect both its dopaminergic and noradrenergic effects as stimulation of dopamine D1 and D2 receptors or of alpha-1 adrenergic receptors in astrocytes increases intracellular Ca²⁺ signaling^78,79.

Meanwhile, cocaine induced increases in ΔCa²⁺_N and ΔCa²⁺_A were both inversely correlated with tissue oxygenation, the correlation was significantly stronger for ΔCa²⁺_N indicating that cocaine’s effects on neuronal activation have a greater influence than astrocytic activation in affecting tissue oxygenation and metabolism^58,80,81. Though, astrocytic activation did not affect cocaine induced changes in ΔCa²⁺_N but attenuated Δ[HbO₂], this could reflect the more localized effects of DREADD(Gq) activation than that of the area where the hemodynamic responses were measured^82,83. This study extends our previous findings by demonstrating that activation of astrocytes disrupts NVC, triggers vasoconstriction and decreases tissue oxygenation but does not impact ΔCa²⁺_N.

Limitations

While this study provides valuable insight into the role of astrocytes in cocaine-induced vasoconstriction, several limitations should be considered. First, our experiments focused only on acute cocaine exposure, which is crucial for understanding its cerebrovascular effects but future studies with chronic administration are necessary to unveil its role in the behavioral disruptions observed in animal models of cocaine addiction. Further studies are also required to investigate potential confounds in our design. In particular, two cocaine injections were administered with a two-hour interval between them to minimize lingering effects from the first cocaine injection. Though control experiments confirmed that this duration was sufficient for the vasculature to recover from the initial cocaine injection (Fig. 5), future studies should evaluate the potential cumulative effects of multiple exposures. Anesthetics may present a confounder as they can attenuate the sensitivity of neurons^67,84 and astrocytes to cocaine. In this study, the breathing rate was 40–60 breaths/min consistent with previous reports of mice under isoflurane anesthesia^85,86, which is lower than the > 100 breaths/min typical in awake animals. Though isoflurane was always administered in a gas mixture with oxygen, we acknowledge that isoflurane anesthesia can attenuate canonical NVC responses⁸⁷. Nonetheless since animals with and without DREADD(Gq) activation were imaged under the same anesthetic conditions the effect of isoflurane cannot account for the observed difference in astrocytic modulation of cocaine induced vasoconstriction.

The goal of our study was to characterize the role that astrocytes have in regulating the vascular and neuronal response to cocaine including neurovascular uncoupling. For this reason, we prioritized calcium imaging using GCaMP6f, a genetically encoded calcium indicator that we have extensively validated with our multimodal optical platform. Cocaine induces robust and reproducible increases in neuronal and astrocytic Ca²⁺, which we can record simultaneously with vascular responses across the wide cortical region^33,66. Moreover, we have previously shown that this approach is sensitive enough to resolve both single-cell activity and single-stimulus responses, making it a versatile tool for studying NVC⁶⁸. We are aware that the observed calcium changes represent the integrated influence of multiple neurotransmitters, including glutamate. Indeed, in prior studies we showed that selective blockade of NMDA receptors attenuated cocaine-induced calcium influx⁵³, confirming the contribution of glutamate to the recorded Ca²⁺ dynamics. We also recognize that cocaine triggers inflammatory cascades, including microglial activation, astrocytic reactivity, and cytokine release. However, these processes typically evolve on a slower timescale emerging over hours to days after exposure^88,89. Because our study was designed to examine rapid neuronal and vascular responses occurring within minutes of cocaine administration, we did not include inflammatory markers.

Conclusion

In summary, using our unique imaging modality with specific Ca²⁺ indicator, we were able to simultaneously assess neuronal/astrocytic Ca²⁺ dependent fluorescence, vascular diameter, and hemodynamic changes in response to acute cocaine in the cortex of mice invivo. DREADD(Gq) activation by itself increased astrocytic Ca²⁺ and induced long-lasting vasoconstriction and it attenuated the magnitude of further cocaine induced vasoconstriction under these conditions suggesting a ceiling effect both for Ca²⁺ and for degree of vasoconstriction. Combined with our previous findings using inhibitory DREADD(Gi), these results establish astrocytic Ca²⁺ signaling as a key modulator of cerebrovascular responses to cocaine. Importantly, our findings suggest astrocytes as a potential therapeutic target for cocaine induced cerebrovascular pathology.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Y.P, N.D.V, and C.D designed the research; Y.L carried out the in vivo experiments; Y.L, K.C and Y.J participated in image processing and data analysis. Y.L, K.C, C.D, N.D.V, and Y.P contributed to data interpretation, result discussions, and manuscript writing.

Data availability

All data generated or analyzed during this study are included in this published article (and its supplementary information files), which can be also available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.