Introduction

The brain is safeguarded by the blood‒brain barrier (BBB), a complex structure comprising a heterogeneous population of cells, including brain microvascular endothelial cells (BMVECs) and components of the neurovascular unit (astrocytes, pericytes, neurons, and the basement membrane). The interaction between BMVECs and the neurovascular unit is crucial for intercellular communication, ensuring the proper functioning of the central nervous system (CNS). In SUD, the demand for intercellular communication escalates due to the influx of inflammatory cells and particles into the brain parenchyma. Understanding the drivers of brain inflammation is essential for advancing our knowledge of CNS mechanics in SUD [1,2,3,4,5,6,7,8,9,10,11,12,13]. One mechanism of cell-to-cell communication involves the release and internalization of EVs, which are membrane-bound vesicles that encapsulate proteins, lipids, nucleic acids, and other cargos. EVs are generated through processes such as plasma membrane budding and endosomal system packaging, and they are released by various cell types, including those in the CNS [14, 15].

In the CNS, EVs play significant roles in neuronal activation [16] and mediate communication between neurons and astrocytes [17], oligodendrocytes [18], and microglia [19]. In individuals with SUD, EVs are thought to regulate responses to substances of abuse, including cocaine [20, 21], cannabinoids [15, 22], nicotine [23], alcohol [24], and opioids [25]. However, research on EVs in humans with SUD is limited. Most studies have utilized EVs isolated from cell lines, limiting the exploration of interactions between EVs and CNS cells in the intricate brain microenvironment—a situation that is further exacerbated by the difficulty in isolating EVs from brain tissues.

Our current study advances the field of SUD biology by using particle purification chromatography (PPLC) [22, 26] to isolate EVs from the postmortem dlPFC of individuals with SUD and without SUD (controls). We assessed the physical properties of the EVs, evaluated their cargo through lipidomic and proteomic analyses, and examined their bioactivity via functional assays of cellular uptake, as well as their effects on cellular gene expression and cell migration.

Methods

Postmortem brain tissues

The sample demographics are described in Table 1. Postmortem brain tissues were obtained from the University of Texas Health Science Center at Houston (UTHealth Houston) Brain Collection in collaboration with the Harris County Institute of Forensic Science, with approval from the Institutional Review Board and after consent from the next of kin (NOK). Medical examiner reports, including causes of death and toxicology, were obtained, and medical records were acquired when available. For SUD patients, substance-induced toxicity (cocaine and/or opioids) was confirmed as the cause of death by medical examiners and toxicology after death. All the controls died of CVD. A detailed psychological autopsy of the donor was obtained by interviewing the NOK [27], from which information on the age of onset of drug use, type of drug used, smoking and drinking history, and any comorbidities was obtained. A diagnosis of opioid use disorder (OUD) or cocaine use disorder (CUD), or designation as nonpsychiatric control (absence of any apparent psychopathology), was determined according to DSM-5 criteria after a consensus meeting where three trained clinicians reviewed the psychological autopsy and all other available records. Upon receipt of the brain, the right hemisphere was coronally sectioned, immediately frozen, and stored at −80 °C. The postmortem interval (PMI), which ranged from 13.07–39.35 h, was calculated from the estimated time of death until tissue preservation. Dissections of BA9, defined within the dlPFC between the superior frontal gyrus and the cingulate sulcus, were obtained using a 4 mm cortical punch.

Table 1 Donor demographics and sample characteristics.
Full size table

Isolation of postmortem BA9 EVs

The term EVs used in this study encompasses subgroups of exosomes, microvesicles, and other membranous vesicles that have overlapping sizes, densities, charges, and surface markers [28]. These subgroups cannot be accurately separated from each other on the basis of their physical properties. Thus, in this manuscript, we did not differentiate the different subgroups of EVs, and we characterized them as recommended by the 2018 Minimal Information for Studies of Extracellular Vesicles (MISEV) [29]. The isolation of EVs typically involves a combination of techniques such as ultracentrifugation, immunoaffinity capture, size exclusion chromatography (SEC), or PPLC. The chosen isolation method directly impacts the purity of EVs; therefore, in this study, we used PPLC to separate EVs from other bioactive extracellular components that may be copurified with EVs. We efficiently separated the EV population from analytes that may masquerade as EVs, such as liposomes and other nonlipid nonmembrane aggregates of nucleic acids and proteins, such as ECs [14, 15]. Our research group developed a protocol for isolating preparative amounts of EVs from brain tissues [22]. A schematic and workflow for the isolation of EVs are shown in Fig. 1A. Briefly, small chunks of frozen BA9 tissues, ranging from 119–161 mg (Table 1), were finely chopped and digested with collagenase III [22]. The samples were subsequently centrifuged at 500 × g, 2500 × g, and 12,000 × g to remove all the cells and cell debris prior to isolation of the EVs. The clarified samples were loaded on a 20 × 0.5 cm Sephadex G-50 size exclusion column and purified using a PPLC system as previously described [22, 26]. Fifty fractions of 200 µL were collected, and 3D UV‒Vis (230–650 nm) fractionation profiles were recorded and used to identify fractions 8–21 as EV-containing fractions [26]. A no-tissue collagenase III control was used as the background. After background subtraction and PPLC analysis, the EV-containing fractions were pooled, aliquoted, and stored in small aliquots at −80 °C until analysis.

Fig. 1: Physicochemical characterization of postmortem BA9 EVs from control individuals and individuals with SUD.
figure 1

A Infographics of the experimental reagents, tools, and protocol. B Absorbance at 280 nm of PPLC-isolated collagenase-digested BA9 eluates (left) and the PPLC R2 index detected in fractions 8–21, confirming the presence of EVs in these fractions from the control and SUD groups (n = 4/group). C Size (left, top), concentration (right, top), and zeta (ζ)-potential (left, bottom) of BA9 EVs measured by NTA (ZetaView), and total protein concentration (right, bottom) measured by the Bradford assay. N = 4 per group for the control group and n = 4 for the SUD group (CUD, n = 2; OUD, n = 2). D Representative TEM images of BA9 EVs from control and SUD individuals. E Western blot showing tetraspanin CD9, CD63, and CD81, which are EV markers present in BA9 EVs. The error bars represent the standard error of the mean. F Representative images of the western blot densitometric analysis showing the immunoreactivity of tetraspanins CD9, CD63, and CD81 in BA9 EVs.

Full size image

Assessment of EV concentration, size distribution, and surface charge

To further these analyses, EVs were diluted in 0.1X PBS (1/1000). EV size, concentration, and zeta potential (ζ-potential) were measured using nanoparticle tracking analysis (ZetaView) as described previously [22, 30]. Briefly, the EV concentration (relative abundance), size distribution, and zeta potential were determined using nanoparticle tracking analysis (NTA) (ZetaView PMX110, v8.04.02; Particle Metrix, Mebane, NC, USA) as previously described [14, 15, 22, 26, 30]. The shutter temperature was kept at 70, and the sensitivity was adjusted to 2–4 points below the noise level to capture small particles.

Transmission electron microscope (TEM)

This is an imaging method that facilitates assessment of the purity, size, and morphology of EVs. Equal volumes of BA9 EVs from each group were pooled (n = 4, n = 2, n = 2 for control, CUD, and OUD, respectively). 10 µL were spotted onto TEM grids and imaged with a Hitachi HT7700 microscope equipped with an AMT XR16M camera. Images were analyzed as described previously by us [7, 22, 26, 31].

Western blot for markers of BA9 EVs

BA9 EV markers were detected via western blotting using previously published protocols [32,33,34]. Briefly, equal amounts (50 µg) of BA9 EV protein from both the control and SUD groups were loaded into 4–20% (gradient) Mini-PROTEAN TGX precast gels and resolved by SDS‒PAGE at a constant 100 V. The separated proteins were blotted onto PVDF membranes, which were blocked with 5% milk dissolved in 1 × TBST (50 mM Tris, 150 mM NaCl, and 0.1% Tween, pH 7.6) buffer for 1 h at room temperature on a slow shaker. The membranes were incubated at 4 °C overnight with primary antibodies against CD9, CD63, and CD81. After 2 10 min washes with 1 × TBST, the membranes were incubated with the appropriate IRDye secondary antibodies at room temperature on a slow shaker for 1.5 h. The membranes were washed in 1 × TBST for 5 min prior to band detection using the Li-COR Odyssey Infrared Imaging System.

Lipid extraction

The samples were concentrated to obtain approximately 10 µL for lipid extraction. Lipids were isolated using a modified Bligh and Dyer protocol [35] with dichloromethane and methanol. Briefly, 10 µL of each sample was transferred to a glass screw-cap tube with 90 µL of water (Burdick and Jackson-Honeywell, USA) of LC‒MS grade and kept on ice for 10 min (MeOH) (Alfa Aesar-Themo Fisher, USA). Then, 2 mL of methanol MeOH and 0.9 mL of dichloromethane (DCM) (Sigma Aldrich, USA) were added. After vortexing, a DCM-MeOH sample of monophase was formed. The samples were left for 30 min at room temperature, and the aqueous and organic phases were separated by the subsequent addition of 1 mL of water and 0.9 mL of DCM. All the samples were centrifuged at 1200 rpm for 10 min, and the lower DCM organic phase was removed to a new glass tube and dried under nitrogen and a partial vacuum using a Visiprep manifold (Supelco, Bellefonte, PA). Lipid extracts were dissolved in a lipid load solution (MeOH/DCM [50:50, v/v]) supplemented with 10 mM ammonium acetate (NH4Ac) (LC/MS grade; Fisher Chemical, USA) for MS analysis.

Lipidomics

The samples were processed following our protocol for liquidization, as previously described [36], and then infused into a QTRAP 4000 mass spectrometer. Samples were analyzed using multiple precursor ion scanning (MPIS) [37] in positive and negative polarities using the following parameters: Fragments File = Fragments_pos, Mass Tolerance (Da) = 0.3, Fragments_neg, and Mass Tolerance (Da) = 0.5 spectrum peak. Direct Infusion Types in Positive: Glycerophospholipid (PC, PE) total double bonds: From 0–6; Sphingolipid (SM, Cer, GM3, GM2, GM1, GD3, GD2, GD1) total double bonds: From 0–4. Glycerolipid (DAG, TAG) total double bonds: From 0–12. Types in Negative: Glycerophospholipid (PA, PC, PE, PG, PI, PS) total double bonds: From 0–12. Lipidomic data were processed using LipidView software (version 1.3 beta) and Marker View (version 1.4) and were normalized using the most likely ratio (MLR) method, Welch t test and principal component analysis (PCA) via the Pareto scale method.

Proteomics

Protein data-independent quantitation (e.g., SWATH) and data-dependent ion spectra libraries were generated as previously described [38,39,40]. Protein identification was determined using the Paragon algorithm [41] in the ProteinPilot (SCIEX) software package, with a <5% false discovery rate (FDR). For quantitative analysis, data-independent mass spectrometric acquisition using Variable Window SWATH Acquisition [42] was conducted in triplicate and used to determine the optimal Q1 isolation windows and increase the specificity of the detected ions. Peaks were extracted with 95% peptide confidence and 1% FDR. The SWATH and MRM data were quantified using MarkerView [43] (Sciex) and normalized using MLR normalization [44] for SWATH and total area sum normalization for MRM, as described previously [45, 46]. P values for fold change were calculated using Welch’s t test in MarkerView (Sciex).

Lipidomic and proteomic data analysis

Lipidomic and proteomic data were analyzed using the multiple precursor ion scanning (MPIS) workflow [47], LipidView software (version 1.3 beta) and Marker View (version 1.4) packages, as previously described [48, 49]. Lipidomic data were analyzed following normalization to common endogenous lipid classes and total lipids using the “most likely ratio” method [36, 44]. Principal component analysis (PCA) of the data was also conducted as previously described [43, 50]. The datasets were compared using the unequal variance (Welch’s) t test unless otherwise indicated [43]. A p value less than 0.05 was considered to indicate statistical significance.

Labeling of BA9 EVs and cellular uptake assay

Pooled BA9 EVs (50 μg) were concentrated using Amicon Ultra0.5 centrifugal filter units (Millipore Sigma) by centrifuging at 14,000 × g for 5 min to remove excess PBS. EVs were resuspended in 100 μL of PBS. The EVs were labeled green or red using SYTO RNASelect Green Fluorescent Cell Stain (Cat# S32703, Invitrogen) or DiR (Cat# #60017, Biotium), respectively, following the manufacturers’ recommended protocols. Labeled EVs were purified using Exo-spin™ mini (Cat# EX03-50; Cell Guidance Systems) to eliminate excess dye. Primary human astrocytes (100,000 cells/100 µL) and HMC3 microglia seeded in each well (3 wells per 3 independent experiments) of a 96-well plate were treated with labeled BA9 EVs at 50 μg/mL and incubated for 6 h or 48 h. NucBlue Live reagent (Cat# R37605; Molecular Probes) was used to stain the nuclei of all the cells. After that, labeled EVs (50 µg/mL) were added to the cells. Images of cells with internalized fluorescently labeled EVs were obtained after 6 and 24 h using a 10× objective and a BioTek Lionheart FX automated microscope. The images were analyzed using BioTek Gen5 software to quantify fluorescence and plotted using Prism GraphPad version 10.

The HMC3 microglial cell line (CRL-3304) was purchased from ATCC and cultured according to the ATCC instructions. In accordance with the manufacturer’s instructions, the base medium used was EMEM (ATCC® 30-2003™) supplemented with 10% FBS. Human astrocytes (HAs) were purchased from Science Cell, Cat#1800. The cells were cultured in the merchant recommended media Astrocyte Medium (AM, Cat. #1801).

Isolation of RNA from glial cells treated with BA9 EVs and real-time quantitative PCR

Equivalent numbers (100,000) of human primary astrocytes and HMC3 microglia were treated with 50 µg/mL BA9 EVs for 24 h in triplicate. RNA was isolated using a Zymogen Quick RNA isolation kit per the manufacturer’s protocol. The RNA was eluted, and the eluate was measured using a NanoDrop 1000. Total RNA (1 µg) was used for cDNA synthesis. Five hundred micrograms of cDNA was used for analysis of gene expression by real-time quantitative PCR (RT‒qPCR), which was performed using a 7500 FAST machine and PowerTrack SYBR Green master mix as previously described. The primers used for RT‒qPCR are listed in Table 2.

Table 2 Sequences of primers used in this study.
Full size table

Flow cytometry

Equivalent concentrations (200 µg/mL) of control and SUD-derived EVs were resuspended in 0.1X PBS. For experimental controls, unlabeled DynabeadsTM (Cat#10608D, Invitrogen) in PBS was used as a negative control, while clarified culture supernatants of confluent monolayers of HMC3 cells and human primary astrocytes cultured with EV-depleted FBS in a 10 cm plate were used as positive controls. Two to five micrograms of biotinylated antibodies against specific cell surface markers, including GFAP, GLAST-1 (astrocytes) and IBA-1 (microglia), were coupled to streptavidin DynabeadsTM. The antibody:bead complex was added to 200 µg/mL EVs, and the mixture was incubated overnight at 4 °C with gentle mixing. The next day, the captured EVs were magnetically separated, washed three times, and then resuspended in 500 µL of MACSQuant running buffer (Miltenyi Biotec). Flow cytometry was performed using a BD FACSCelesta™ flow cytometer. Analysis was performed using FlowJo™ v10.10.

Cell migration analysis using a wound healing (scratch) assay

To perform the cell migration assay, we used our previously published wound healing assay protocol [51]. Briefly, 50,000 HC69 human microglia were seeded in silicone cell culture inserts (2-well in µ-Dish 35) mm with a defined cell-free gap within the monolayer (Ibidi) designed for precise cell migration (wound healing) assays. The cells were cultured for 5 h, after which the cells were allowed to adhere to the dish. HC69 cells were treated with 50 µg of control or SUD BA9 EVs in triplicate. Eighteen hours later, the cells were washed and incubated in serum-free media (starvation) for 2 h. Cell migration into the wound area (gap) was monitored by imaging at different time points using a BioTek Lionheart FX automated microscope. Images were analyzed to determine the wound closure rate (a typical experimental readout) by using BioTek Gen5 software.

Statistical analysis

Both datasets—Proteomics from ProteinPilot and Lipidomics from LipidView (both from Sciex)—used the inbuild t test within the MarkerView software, which does not ask for FDR info explicitly. Ingenuity Pathway Analysis (IPA) is a web-based bioinformatics inbuilt application that was used for the analysis and integration of protein-to-function data. The significance cutoff was set to a fold change (FC) > 2.0 or <−2.0 and a p value < 0.05. Ordinary one-way ANOVA or multiple comparisons with Šídák’s multiple comparisons test or two-way ANOVA with Dunnett’s correction were used to assess significant differences. When stated, an unpaired t test with Welch’s correction was also used. The detailed statistics are presented in the legend of each figure, where the p values are listed. Label-free quantitative analysis was performed using the relative abundance intensity normalized by the median, and all peptides identified were used for normalization. The expression analysis was performed considering the technical replicates available for each experimental condition following the hypothesis that each group was an independent variable. Only proteins that were present in 2 out of three technical replicates were considered, and a statistical cutoff of ANOVA > 0.05 was adopted. Protein interaction and pathway enrichment analyses were performed in Cytoscape and in silico analyses in the R environment [52]. For PCA and volcano plot analysis, MetaboAnalyst was used [53].

Results

SUD does not alter the physicochemical properties of BA9 EVs

Schematics for BA9 EV isolation and subsequent analyses are shown in Fig. 1A. The sizes, concentrations, ζ-potentials (mV), and total protein concentrations of the EVs from the controls and from the SUD patients were similar (Fig. 1B, C). TEM analysis revealed that control and SUD-derived EVs (~100–300 nm) had similar morphologies and were heterogeneous in size and electron density (Fig. 1D, E). Further analysis revealed the presence of tetraspanin markers (CD9, CD63, and CD81) in both the control and SUD patients (Fig. 1F). Given the small sample size, it is unknown whether SUD will alter EV morphology or if other EV markers are present and to what extent such markers are enriched. Nevertheless, the presence and purity of BA9 EVs prepared via rigorous PPLC devoid of most contaminants were demonstrated.

Effect of SUD on the EV lipid profile

We used an untargeted LC‒MS/MS platform that utilizes positive/negative polarity switching to perform unbiased data-dependent acquisition (DDA) through higher energy collisional dissociation (HCD) fragmentation to profile lipid ions associated with extracellular vesicles (EVs). Using the Lipid Ontology (LION) enrichment analysis web application (LION/web), a total of 98 positive (Supplementary Table 1) and 385 negative (Supplementary Table 2) polarity lipid species were identified, as displayed in the heatmaps (Supplementary Fig. 1A, B), highlighting the enrichment and clustering of various lipid species in the control and SUD groups. The identified EV-associated lipids were grouped into 3 main categories, sphingolipids, neutral lipids, and phospholipids, which were equally distributed between the control and SUD groups (Fig. 2A). Additionally, the 3 lipid categories included 16 classes of lipids (Cer, CerP, DAG, DES, GD2, GD3, GM1, GM2, GM3, GT3, Hex2Cer, Hex3Cer, HerCer, IPC, SM, and TAG) among the positive polar lipids (Fig. 2B, top) and 11 classes of lipids (CDPDAG, CL, PA, PC, PE, PG, PI, PIP, PIP2, PIP3, and PS) among the negative polar lipids (Fig. 2B, bottom).

Fig. 2: Profile of extracellular lipid classes associated with BA9 EVs from control and SUD postmortem BA9 brain tissues.
figure 2

A Main lipid categories associated with BA9 EVs. B Main lipid classes associated with BA9 EVs within the 3 lipid categories. C LION-PCA heatmap for positive (top) and negative (bottom) polar lipids associated with BA9 EVs. D, E LION enrichment analysis of BA9 EVs for D positive and E negative polar lipids associated with BA9 EVs. F Glycerophosphoinositol bisphosphate PIP2 levels in BA9 EVs.

Full size image

LION associations reveal differential enrichment of lipid species between the control and SUD groups

Next, we used a LION-PCA heatmap analysis to further dissect the lipidomic data. We identified significant LION terms (Supplementary Fig. 1C, left) or the cumulative variance explained (Supplementary Fig. 1C, right), per set number of principal components (5) for positive (Supplementary Fig. 1C, Top) and negative (Supplementary Fig. 1C, Bottom) polar lipids, and identified major differences in EV-associated lipids between SUDs and controls for positive (Fig. 2C, top) and negative (Fig. 2C, bottom) polar lipids. Inspection of the heatmap revealed a marked increase in the LION signatures of glycerophosphoinositol trisphosphate [GP09] (LION:0000018) and glycerophosphoinositol bisphosphate [GP08] (LION:0000017) in SUD BA9 EVs (Fig. 2C, bottom, orange arrows). LION enrichment analysis revealed positive and negative polar lipids that were either up- or downregulated in SUD (Fig. 2D, E), whereas glycerophosphoinositol bisphosphate PIP2 [GP08] (LION:0000017) was strongly upregulated (Fig. 2E, orange arrow; Fig. 2F). The presence of other lipid species, including DAGs and CERs, which can accumulate in SUD EVs, is acknowledged. These lipid species can potentially affect alcohol consumption and depressive‐like and anxiety‐like behaviors, as described in specific mouse brain regions and in ceramide species‐specific manners [54]. Additionally, lipidomic pathways linked to the sphingomyelinase/ceramide system may be lipotoxic or may regulate synaptic transmission, reactive oxygen species, and proinflammatory cytokine functions in the CNS [55].

Effect of SUD on the EV protein profile

We used shotgun mass spectrometry-based proteomics (LC‒MS/MS) to profile the proteome of a portion of the same BA9 EVs [56] whose lipid content was analyzed, as shown in Fig. 2. A total of 472 proteins were identified (Supplementary Table 3), 78 of which were found to be statically dysregulated after the removal of contaminants, such as keratins (p value = 0.05, FC cutoff = 20% variance). A hierarchical cluster analysis applying Pearson’s correlation based on the protein expression levels demonstrated a discernible disparity between the SUD patients and HCs, confirming the differences in the proteome cargo of the vesicles (Fig. 3A, B; Supplementary Fig. 2A). Although the heatmap shows that one SUD sample appears to be an outlier compared to the others, it is unlikely that the observed difference is related to sex since the outlier is one of the three female specimens (67,951, Table 1). For enrichment analysis, only proteins that were significantly differentially regulated (pooled p value ≤ 0.05 and adjusted p value (FDR) < 0.05) were considered, 33 of which were significantly increased and 45 of which were decreased; among them, SERPINB12, ACYP2, CAMK1D, DSC1, and FLNB were the most notable (Fig. 3C). Gene Ontology analysis revealed a pronounced abundance of proteins originating from exosomes, vesicles, and the extracellular space, confirming successful enrichment of vesicles derived from brain tissue (Fig. 3D). A substantial proportion of these proteins were implicated in inflammatory and innate immune responses (Supplementary Fig. 2B). Biological and WikiPathways analyses indicated that the proteins may be involved in cellular movement (Fig. 3E, top) and focal adhesion (Fig. 3E, bottom). Using ingenuity pathway analysis (IPA), we identified the top 5 biological processes (Table 3), with infectious diseases, organismal injury and abnormalities, metabolic disease, neurological disease, and psychological disorders as the top 5 diseases (Table 3). Signaling pathway analysis revealed that proteins associated with SUD-related EVs potentially affect two major cellular processes—cell viability and cell motility (Fig. 3F, Supplementary Fig. 2C). The upstream regulators of cellular functions related to cell death (Fig. 3G, left) and cell migration (Fig. 3G, right) were determined by IPA. Protein-to-function relationship analysis of the network molecules revealed that the human ALB, C3, and C4A/C4B proteins, which are upregulated in patients with SUD, were major hub proteins for proinflammatory cytokines (Fig. 3H, Supplementary Fig. 2D). Additionally, the C3 and C4A/C4B proteins are hub proteins for multiple neurodegenerative diseases (Fig. 3I). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [57] partner repository with the dataset identifier PXD061471.

Fig. 3: Analyses of the proteomes of BA9 EVs from control and SUD BA9 brain tissues.
figure 3

A, B Hierarchical cluster analysis assessing differences in BA9 EVs from control and SUD BA9 tissues using Pearson’s correlation reveals differences visualized as A heatmaps and B correlations. C Volcano plot analysis of the BA9 proteome. D Gene Ontology analysis showing the abundance of proteins originating from exosomes, vesicles, and the extracellular space, confirming successful enrichment of vesicles derived from brain tissue. E Gene Ontology analysis of the BA9 proteome using the WEB-based GEne SeT AnaLysis Toolkit (WebGestalt 2024). F Ingenuity pathway analysis (IPA) was used to predict ontologically related functions and molecules predicted to be increased (orange) or decreased (blue) when driven by SUD-induced changes in the proteome of BA9 EVs. The DEGs are involved in biological processes such as cell viability, cell migration, and degranulation and are not involved in disease, while the genes encoding IL1B, AGT, STAT3 and others may regulate these functions. A solid line represents a direct interaction between two gene products, and a dotted line indicates an indirect interaction. G IPA-identified upstream regulators of cellular functions related to cell death (left) and cell migration (right) mediated by BA9 EVs. A solid line represents a direct interaction between two gene products, and a dotted line indicates an indirect interaction. Orange indicates activation or increase, and blue indicates inhibition or decrease. H IPA-identified most highly rated network and canonical pathway analysis. I IPA-identified groups of ontologically related diseases predicted to be linked to the proteome of BA9 EVs.

Full size image
Table 3 Top 5 Biological Processes associated with SUD BA9 EVs.
Full size table

SUD regulates the release of EVs by microglia

The majority of EVs were positive for the microglial marker IBA 1 (Fig. 4A), suggesting that EVs were predominantly microglia rather than astrocytes. Additional analysis comparing control versus SUD patients revealed that microglia within the dlPFC may shed significantly more EVs in response to SUD than in response to control (Fig. 4B, C). Unlike the expression of IBA1, GFAP expression was significantly suppressed in SUD-derived EVs (Fig. 4C, left). The reason for this suppressed expression is unknown. However, in cocaine-administering animals, the surface area and volume of astrocytes, as well as GFAP levels, are significantly reduced [58], indicating that cocaine suppresses GFAP expression.

Fig. 4: Flow cytometry analysis of BA9 EVs isolated from postmortem brains of control and SUD individuals.
figure 4

A Pooled control and SUD BA9 EVs demonstrating the presence of glia cell-derived EVs irrespective of substance use, as shown by the plots of gates and their associated histograms. B, C Control and SUD BA9 EVs showing differences in the release of EVs by control and SUD BA9 postmortem brain tissues as visualized by B plots of gates and their histograms; C quantitative values presented in bar charts. EVs isolated from astrocyte or microglial supernatants (Sup) were used as a positive control, while beads alone in PBS served as a negative control. Statistical significance was determined by ordinary one-way ANOVA (Šídák’s multiple comparisons test) and an unpaired t test with Welch’s correction. ****p < 0.0001, ***p 0.0005, **p 0.0088–0.0051, *p 0.0167, ns nonsignificant.

Full size image

Glial cells internalize EVs

Internalization of labeled control and SUD-derived EVs by human primary astrocytes and HMC3 microglia revealed a green signal (Fig. 5A, B) that increased over time (Fig. 5C, D). At 6 h post treatment, the internalization of control EVs was significantly greater than that of SUD in astrocytes (Fig. 5C) but not in microglia (Fig. 5D); however, at 24 h post treatment, there was no significant difference in the uptake of EVs by the two cell types. The reason for the delayed uptake of EVs by astrocytes compared to HMC3 cells is unknown, but pathways regulating the uptake and trafficking of EVs in these cells may differ. A similar observation has been made in mouse primary microglia and astrocyte cultures, where the uptake of EVs was more efficient in microglia than in astrocytes [59].

Fig. 5: BA9 EVs are taken up by glial cells, and EVs alter glial cell gene expression.
figure 5

A, B Representative 10x images of A human primary astrocytes and B HMC3 microglia incubated with SytoSELECT (green fluorescence)-labeled BA9 EVs (50 μg) in the presence of NucBlue (a live cell stain, 30 µL/mL) and seeded in glass-bottom 96-well plates. Images were taken at 6 and 24 h using an automated Lionheart FX microscope (BioTek). C, D Uptake of BA9 EVs by C astrocytes and D microglia was analyzed by quantifying 3 fields of view with Gen5 software and presented as raw relative florescence units (RFUs). E, F Gene expression of E human primary astrocytes and F HMC3 microglia treated with control or SUD BA9 EVs was assessed by RT‒qPCR. Cells treated with PBS were used as a negative control. For panels C and D, statistical significance was determined by two-way ANOVA. **** p < 0.0001, *** p 0.0002 - 0.0009, ns = nonsignificant. For panels E and F, statistical significance was determined by ordinary one-way ANOVA (Šídák’s multiple comparisons test) and an unpaired t test with Welch’s correction. ****p < 0.0001, ***p 0.0002–0.0009, **p 0.0042, *p 0.0247, ns nonsignificant.

Full size image

EVs reprogram the transcriptome of glial cells

Since the protein content of SUD-related EVs, especially complement proteins C3 and C4A, was markedly different from that of control EVs, we sought to assess the effect of these EVs on the glial cell transcriptome, with a focus on C3 and C4A. In astrocytes, EVs from SUDs but not from controls significantly increased the levels of C3 and C4A mRNA after 24 h of treatment (Fig. 5E). In contrast, both control and SUD-derived EVs significantly increased the expression of the microglia C3 and C4A mRNAs (Fig. 5F).

EVs from SUD individuals promote the collective migration of microglia

Since PIP2 is enriched in SUD-related EVs (Fig. 2F) and since PIP2, along with its phosphorylated product PIP3, plays crucial roles in maintaining cell polarity and polarized trafficking of molecules in cells, we investigated the role of EVs from control and SUD individuals in directional cell migration. Using our previously described wound healing assay [51], we examined the migration of human microglia in response to a mechanical scratch wound, either in the absence (PBS) or presence of EVs isolated from control or SUD individuals. Representative images of the scratch areas at 0, 18, and 24 h are shown in Fig. 6A. Approximately 89.9 and 89.5% of the cells treated with PBS remained in the wound area at 18 and 24 h, respectively (Fig. 6A, B, left). In contrast, 93.7 and 84.3% of the cells treated with control EVs had wounds at the same time points (Fig. 6A, B, middle), indicating that EVs promote microglial migration. Notably, in the presence of EVs from SUD individuals, microglia exhibited even greater enhancement of migration, with 80.8 and 64.3% of the wound area remaining at 18 and 24 h, respectively (Fig. 6A, B, right). These findings suggest that EVs from SUD individuals significantly potentiate microglial migration beyond the levels induced by control EVs from the dlPFC.

Fig. 6: BA9 EVs regulate microglial cell migration.
figure 6

A Representative images of wound tracks and areas. The vertical white lines track the wound areas. Images were captured with a LionHeart instrument with a 4X objective. Scale = 3000 µm. B Time- and concentration-dependent plot of the wound area analyzed by quantifying 3 fields of view with Gen5 software, and the data are presented as the percent wound area. Statistical differences were assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test for panel C, ordinary one-way ANOVA with Šídák’s multiple comparisons test and an unpaired t test with Welch’s correction. ***p 0.0005, **p 0.0088, *p 0.018–0.03, ns nonsignificant.

Full size image

Discussion

In this study, we successfully isolated EVs from human postmortem brain tissue and elucidated changes occurring within the dlPFC in individuals with SUD. To our knowledge, our paired lipidomic and proteomic analyses of BA9 EVs are unprecedented in the field of SUD research. Previous studies, including ours, have focused primarily on the protein and nucleic acid composition of EVs, providing valuable insights into the roles that EVs play in human health and disease [15, 21, 22]. However, our current comprehensive analysis represents a significant advancement in understanding the impact of SUD on the brain and the role of EVs in this process.

We detected relatively high numbers of total sphingolipid and phospholipid species in BA9 EVs, with an equal distribution between the control and SUD groups. These findings align with the literature showing that EVs are enriched in sphingolipids (sphingomyelin and ceramide) and glycerophospholipids containing saturated fatty acids [60,61,62]. Prior studies have suggested that the detection of large numbers of sphingolipids in EVs may be due to contamination by lipoproteins or lipid droplets [63]. However, our use of PPLC effectively separated EVs from other extracellular particles, confirming this assumption [26, 32]. Our analysis revealed that phosphoinositides, particularly phosphatidylinositol bisphosphate (PIP2), were abundant in BA9 EVs. Although total phosphoinositides did not significantly differ between control and SUD-EV groups, PIP2 was notably elevated in SUD-EV group. PIP2 plays a critical role in maintaining cell polarity, directional cell migration, immune signaling, and inflammation [64]. The significant elevation in PIP2 in SUD-related EVs suggests functional implications that merit further investigation. Notably, the control subjects included in this study had a history of CVD or asthma, which may have affected the lipid and protein composition of brain-derived EVs.

Membrane lipids interact with proteins in various ways, potentially stabilizing the membrane environment for proteins to embed within EV membranes. We observed SUD-specific protein enrichment in BA9 EVs. The presence of tetraspanins (CD9, CD63, and CD81) correlates with the findings of previous studies [7, 32, 56, 65]. Our protein enrichment analysis indicated that SUD-related EVs may be pathological, potentially activating various diseases, including infectious, neurological, and psychological disorders, as well as organismal injury and abnormalities. Of particular interest, SUD-derived EVs exhibited elevated levels of C3 and C4A/4B complement proteins. The complement system is part of the innate immune system’s first line of defense against pathogens and plays a crucial role in neuronal health [66]. However, overactivation of the complement cascade in the brain can lead to various neurological dysfunctions [67, 68], including autism spectrum disorder (ASD) [69, 70]. While C4A has been implicated in regulating blood‒brain barrier permeability [71] and the pathophysiology of psychiatric disorders such as schizophrenia [72, 73], our findings are novel in implicating C3 and C4A complement proteins in SUD. The proteins SERPINB12, ACYP2, CAMK1D, DSC1, and FLNB are dysregulated in the EVs of SUD patients. However, studies investigating the roles of these proteins in the context of the nervous system are limited, making it difficult to fully understand their functions. However, SerpinB2 has been associated with the inhibition of macrophage migration, suggesting that decreased expression of this protein may be involved in the increase in microglial migration observed in our study [74]. Additionally, CAMK1D is also dysregulated in Alzheimer’s disease, with both mRNA and protein expression downregulated in the postmortem brain tissue of patients, which could indicate a potential link due to the neurodegenerative components shared with SUD [75, 76].

Our flow cytometric analysis identified a greater number of microglia-derived EVs in SUD BA9 patients than in controls. These findings suggest that SUD dysregulates the release of EVs in a cell type-dependent manner and that EVs derived from specific cellular origins in the brain, such as microglia or astrocytes, may have biological effects on bystander cells. The transfer of EV cargos, including biologically active nucleic acids, lipids, lipid metabolites, and proteins, has emerged as a mechanism by which cells alter biological processes within target cells and tissues, including the brain [21, 33, 34, 65]. These findings align with our findings that BA9 EVs altered the levels of C3 and C4A/4B mRNA in astrocytes and microglia. Additionally, BA9 EVs from control individuals increased microglial migration, while EVs from SUD individuals further enhanced this migration. The increased migration in the presence of BA9 EVs indicates that these EVs are functionally bioactive and that SUD enhances their activity. Although the specific mechanisms by which SUD BA9 EVs potentiate cell migration have not been determined, the significant increase in wound closure (directional cell migration toward the wound area) in cells treated with SUD EVs supports our lipidomic data, which showed PIP2 enrichment in SUD EVs. Given the increasing evidence that migration is crucial in cell‒cell communication and interactions with the microenvironment, as reviewed by Miskolci et al. [77], and therefore important in physiological and pathological processes such as wound healing, immune response, pathogen dissemination, and neuroinflammation, our findings warrant further investigation into the role of BA9-derived EVs in SUD pathophysiology.

Furthermore, the EVs used in this study were obtained from postmortem brain samples, which on their own is challenging, and the samples are limited. As a result, we evaluated microglial migration only. However, C3 colocalizes with astrocytes and promotes their survival [78]. Additionally, C3a downregulates the expression of Gfap in astrocytes for recovery after ischemic stress [79]. In our studies, we found significantly lower levels of GFAP in SUD-related EVs than in control cells (Fig. 4), and astrocytes treated with SUD-related EVs had elevated C3 mRNA levels (Fig. 4E).

Neuroinflammation is a key component of the pathogenesis of numerous CNS diseases and is thought to contribute to neural adaptations following chronic drug exposure [80]. Neuroinflammation includes reactive astrogliosis; microglial activation; and the release of cytokines, chemokines, and proinflammatory factors, which are all implicated in the brain’s response to chronic substance use [81,82,83]. Furthermore, the increased presence of microglial-derived EVs in SUD patients (Fig. 4) supports the notion that drug use heightens neuroinflammatory processes, potentially leading to neuropathology. Our finding that BA9 EVs alter the levels of C3 and C4A/4B mRNAs in astrocytes suggests that EVs could play a role in astrocyte-mediated neuroinflammation and synaptic modulation. This finding is consistent with studies showing that astrocytes play a critical role in the uptake of synaptically released glutamate and are affected by the activity levels of dopamine (DA) neurons, which are often dysregulated in people with SUD [84]. Drug-induced dysregulation of neuroimmune signaling may compromise neuronal function, exacerbate neurodegeneration, and increase neurotoxicity, all of which can contribute to drug-related behaviors through microglia and other glia-mediated synaptic remodeling [83]. The dysregulation of EV cargo in SUD, particularly the enrichment of PIP2 (Fig. 2) and the complement proteins C3 and C4A/4B (Fig. 5), points to specific pathways that may be targeted to alleviate the neuroinflammatory and neurodegenerative consequences of SUD. The role of PIP2 in cell signaling, immune responses, and maintaining cell polarity [85] underlines its potential impact on neuronal health and the progression of neuroinflammation in individuals with SUD. Similarly, the involvement of complement proteins in SUD highlights a possible mechanistic link to the broader neuroimmune response observed in these individuals. Overactivation of these pathways in SUD patients could lead to detrimental effects on synaptic integrity and cognitive function.

Overall, our study underscores the complex interplay between EVs, glial cells, and neuroinflammation in SUD. The identification of specific lipidomic and proteomic changes in EVs from SUD individuals provides valuable insights into the molecular mechanisms underlying this disease. These findings highlight potential biomarkers and therapeutic targets for addressing the neuroinflammatory and neurodegenerative effects of chronic substance use. Future research should focus on elucidating the precise roles of these EV components in mediating neuroimmune responses and their potential as targets for intervention in SUD.

Limitations

This study provides groundbreaking insights into the potential functions of BA9 EVs, leveraging human postmortem brain tissues. While the findings are promising, certain areas remain to be explored. Due to the limited sample size we were unable to perform separate analyses of specific SUD subtypes, such as CUD and OUD, or to examine their independent molecular effects. Similarly, demographic factors such as sex, age, and ancestry could not be assessed, despite their well-documented relevance in SUD pathophysiology and EV biology [86, 87]. In particular, the SUD group had a younger average age compared to the control group, and the sample was predominantly female, which may have contributed to the observed differences in EV cargo composition. Additionally, while postmortem BA9-derived EVs provide valuable molecular insights, they do not capture the dynamic nature of EV-mediated processes in vivo. These limitations reflect the preliminary nature of this study and highlight opportunities for future research. Subsequent studies with larger, more diverse cohorts are essential for examining the impact of specific substances of abuse, biological sex, and ancestry on BA9 EV composition. Notably, we are currently expanding this line of research in a larger-scale study, aiming to refine and extend these findings.

Conclusions

Our study demonstrated the effective use of novel PPLC technology to isolate BA9 EVs and elucidate their lipidomic and proteomic profiles in the brains of individuals with SUD compared to those of individuals without SUD. We identified significant alterations in EV secretion and uptake by glial cells and in the impact of these alterations on the glial transcriptome and collective migration. These findings contribute to a deeper understanding of the role of brain-derived EVs in the context of SUD, highlighting their potential involvement in pathogenic processes and their utility as biomarkers for SUD diagnosis. As the field progresses toward a more comprehensive and integrated analysis of brain-derived EVs, our study underscores the importance of examining these vesicles in understanding the molecular mechanisms of SUD. Future research should aim to determine whether the observed alterations in BA9-derived EVs are also present in peripheral EVs, such as those derived from blood, or in other brain regions. This approach will be crucial for validating EVs as reliable biomarkers and therapeutic targets for SUD.