Extracellular vesicles are increased in the serum of children with autism spectrum disorder, contain mitochondrial DNA, and stimulate human microglia to secrete IL-1β

Caucasian children (n = 20, 16 males and 4 females, 4-12 years old) diagnosed with ASD were evaluated as part of a clinical trial that was conducted at the Attikon General Hospital, Athens Medical School, Athens, Greece [33]. Children were diagnosed with ASD based on clinical assessment and corroborated by meeting the cutoff scores on both the DSM-IV-TR symptom list and the Autism Diagnostic Observation Schedule (ADOS) algorithm. Subjects were medication-free prior to a blood draw for at least 2 weeks for all psychotropic medications (4 weeks for fluoxetine or depot neuroleptics).

The exclusion criteria were (a) any medical condition likely to be etiological for ASD (e.g., Rett syndrome, fragile X syndrome, or tuberous sclerosis); (b) any neurologic disorder involving pathology above the brain stem, other than uncomplicated non-focal epilepsy; (c) contemporaneous evidence, or unequivocal retrospective evidence, of probable neonatal brain damage; (d) any genetic syndrome involving the CNS, even if the link with autism was uncertain; (e) clinically significant visual or auditory impairment, even after correction; (f) any circumstances that might possibly account for the picture of autism (e.g., severe nutritional or psychological deprivation); (g) mastocytosis (including urticaria pigmentosa); (h) history of upper airway diseases; (i) history of inflammatory diseases; and (j) history of any allergies [33]. This protocol was approved by the Attikon Hospital Human Investigation Review Committee, and all parents or legal guardians provided written informed consent.

Fasting blood was collected in serum separator vacutainer tubes (BD Biosciences, Rockville, MD). Whole blood was allowed to clot at room temperature for about 15–30 min. The samples were then centrifuged at 1000–2000×g for 10 min at 4 °C. The upper clear fraction (serum) was carefully removed and aliquoted (0.5 mL/tube) into clean plastic capped tubes. Serum was also collected from normally developing, healthy Caucasian children (n = 20, 16 males and 4 females, 4–12 years old), unrelated to the ASD subjects, who were seen for routine health visits at the Pediatric Department of the Social Security Administration (IKA) polyclinic. All ASD and control blood samples were prepared immediately and stored at − 80 °C. They were later shipped on dry ice to Tufts University for further analysis.

All blood samples were labeled only with a code number, as well as the age and sex of the respective subject. They were sent to Tufts blind without any other identifiers, such as weight, or severity of ASD in the case of patients.

Total EVs were isolated and purified from 1 mL of serum using the exoEasy Maxi Kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. Briefly, serum samples were filtered to exclude particles larger than 0.8 μm using syringe filters (EMD Millipore, Burlington, MA). Pre-filtered samples were then mixed with Buffer XBP (EMD) and were bound to an exoEasy membrane affinity spin column. The bound EVs were washed with Buffer XWP (EMD), were eluted with 400 μL Buffer XE (an aqueous buffer containing primarily inorganic salts, EMD), and were then ready for further analysis.

One drop (5 μL) of a sample containing EVs was floated on the grid storage box for 1 min. The grid was then moved to a drop of double-distilled water, and the excess liquid was removed with a filter paper and stained by floating on a small drop of uranyl formate 0.75% for 30 s. After removing the excess uranyl formate with a filter paper, the grids were examined in a TecnaiG² Spirit BioTWIN TEM (FEI Company, Hillsboro, OR, USA), and images were recorded with an AMT 2k CCD camera at a primary magnification of × 20,000–50,000 (Harvard Medical School’s Electron Microscopy Core Facility).

The concentration of EV total protein was quantified by the bicinchoninic acid (BCA) assay (Thermo Fisher Inc., Rockford, IL) using bovine serum albumin (BSA) as standard.

Extracellular vesicle-associated markers (CD9 and CD81) were determined by Western blot analysis. Protein samples of 10 μg were loaded, separated on 4–12% NuPAGE Bis-Tris gels under SDS-denaturing conditions (Invitrogen Life Technologies, Grand Island, NY) starting by 65 V for 45 min, and then increased to 90 V for another 30 min. Proteins were then electrotransferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA) followed by blocking for 1 h using 5% BSA in Tris-buffered saline containing 0.05% Tween-20. The membranes were then incubated overnight at 4 °C with the following primary antibodies at 1:1000 dilutions: CD9 and CD81 (System Biosciences, Mountain View, CA). For detection, the membranes were incubated with the appropriate secondary horseradish peroxidase (HRP)-conjugated antibody (System Biosciences) at 1:20,000 dilutions for 1 h at room temperature, and the blots were visualized by enhanced Super Signal West Pico Chemiluminescence (Fisher Scientific, Pittsburgh, PA).

The immortalized human microglia-SV40 cell line derived from primary human microglia was purchased from Applied Biological Materials Inc. (ABM Inc., Richmond, BC, Canada) and cultured in Prigrow III medium (ABM Inc., Richmond, BC, Canada) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in type-I collagen-coated T25-flasks (ABM Inc.). Microglia-SV40 maintained a specific phenotype and proliferation rate for over ten passages, during which all experiments were performed using multiple microglia thaws and sub-cultured cells. Experiments were carried out in type-I collagen-coated plates (BD PureCoat™ ECM Mimetic Cultureware Collagen I peptide plates (Becton Dickinson, Bedford, MA)). Microglia-SV40 were seeded in 6-well plates (1.0 × 10⁵ cells/well) for 24 h before the stimulation with EVs. Lipopolysaccharide (LPS) or the peptide neurotensin (NT) was used as “positive” controls. Cells were stimulated for 24–48 h, and secreted IL-1β was measured in supernatant fluids using ELISA. Cell viability was determined by Trypan blue (0.4%) exclusion.

IL-1β secretion in supernatant fluids was determined in duplicate by using commercially available ELISA kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. For all experiments, the control cells were treated with an equal volume of culture medium. The minimum detectable level for IL-1β by ELISA was 5 pg/mL.

Total DNA was extracted from EVs using Qiagen DNA Micro extraction kit (Qiagen, CA). Mitochondrial-specific DNA for 7S (mtDNA7S) was detected and quantified by real-time PCR (RT-PCR) using TaqMan gene expression assays (Mt-7S: Hs02596861_s1; GAPDH: Hu, VIC, TAMRA, Applied Biosystems, Carlsbad, CA). Samples were run at 45 cycles using Applied Biosystems 7300 Real-Time PCR System. GAPDH DNA was used to exclude any genomic “contamination.” The same amount of EV-associated DNA (50 ng/μL) from each individual ASD and control sample was used for the analysis of mtDNA.

All in vitro conditions were performed in triplicate, and all experiments were repeated at least three times (n = 3). Results are presented as mean ± SD. Data from stimulated and control samples were compared using the unpaired two-tailed Student’s t test. The significance of comparisons is denoted by p < 0.05. All analyses were performed using Graph Pad Prism 5.

The morphology of EVs was evaluated using TEM. EVs had a spherical shape with a size of approximately 100 nm surrounded by a membrane (Fig. 1a). The identity of EVs was further validated by quantitating the EV membrane-associated markers CD9 and CD81 using Western blot analysis (Fig. 1b). There was no apparent difference in the size or shape of EVs between the control and ASD samples.

The total protein concentration in EV-enriched serum samples was significantly higher in children with ASD (204.6 ± 10.6 μg/mL) as compared to the normotypic controls (161.4 ± 9.6 μg/mL), p = 0.02 (Fig. 2a).

We also measured the EV-associated mtDNA7S in ASD children and normotypic controls. The amount of mtDNA7S contained in EVs from children with ASD was found to be increased (p = 0.046) compared to the normotypic controls (Fig. 2b).

We then evaluated the effect of EVs derived from the serum of patients with ASD on human-cultured microglia secretion of IL-1β. We chose to measure this pro-inflammatory cytokine because we recently reported that microglia could secrete IL-1β in response to NT, which we had shown to be increased in the serum of children with ASD [31, 32]. Human-immortalized microglia SV40 were stimulated with EVs (1 or 5 μg/mL), quantitated as total EV-associated protein, for 24 or 48 h, and IL-1β secretion was measured by ELISA. EVs from each ASD and control sample were used separately and were not pooled. Both 1 and 5 μg/mL stimulated significant secretions of IL-1β (Fig. 3). Stimulation with 5 μg/mL for 24 h and 48 h significantly increased the secretion of IL-1β to 95.36 ± 5.31 pg/mL and 163.5 ± 13.34 pg/mL, respectively, as compared to control microglia (59.76 ± 2.03 pg/mL, p < 0.0001, and 117.7 ± 3.96 pg/mL, p = 0.008, respectively) (Fig. 3a–b). Lower amount of EVs (1 μg/mL) also stimulated IL-1β secretion at 24 h (78.47 ± 5.46 pg/mL, p = 0.02) and at 48 h (151.3 ± 13.63 pg/mL, p < 0.04), respectively (Fig. 3a–b).

EVs from normotypic children did not have a significant effect (p = 0.09) at either 1 or 5 μg/mL, as compared to the unstimulated control cells. Neurotensin (10 nM) and LPS (10 ng/mL) used as “positive” controls also increased (p < 0.0001) the secretion of IL-1β (Fig. 3).