Human induced pluripotent stem cells (hiPSCs) were generated from a healthy 35-year-old male subject using an episomal reprogramming kit (Invitrogen). Approval for the study was obtained from the University of Texas at San Antonio IRB panel and informed consent was obtained from the subject. H9 ESCs were purchased from Wicell. Both hiPSCs and ESCs showed pluripotency markers and normal karyotype throughout the study (Supplementary Figures 1A,B) and were grown on matrigel-coated plates in mTeSR medium (Stem Cell Technology) containing the Rock inhibitor, Y27632 (10 μM, Selleck). In order to fluorescently label cells for migration analysis, the Neon Transfection Kit (Invitrogen) was used to transfect cells with CRISPR-mCherry (mCh) plasmids. Cells were dissociated using accutase, then 1 × 10^6 cells were resuspended in Buffer R containing 1 μg of CRISPR plasmid (AAVS1-T2 CRISPR in pX330) and 3 μg mCh plasmid (AAVS1-Pur-CAG-mCherry). Transfections were performed at 1,100 V, 20 msec, 2 pulses and cells were immediately re-plated in mTeSR medium containing Y27632. Three days after transfection, 0.3 μg/ml puromycin was added to the media for 5 days to kill non-transfected cells.

Buprenorphine was purchased from Cayman Chemical Company. In all experiments, organoids were treated with 2 ng/ml buprenorphine to approximate concentrations observed in human placental tissue (Concheiro et al., 2010). Oxycodone was purchased from Spectrum Chemicals. To compare an abused opioid known to affect neurodevelopmental (Larson et al., 2019), 20 ng/ml oxycodone was used (Kokki et al., 2012). For chronic treatment, drugs were added at day 50 for 10 days, period of active progenitor proliferation and migration.

To grow hCS and hSS we used the method described by Pasca and colleagues with light modifications (PaŞca et al., 2015; Birey et al., 2017). Briefly, 9,000 stem cells were resuspended in 150 μl neural induction medium (DMEM/F12, 20% knock-out serum replacement, Glutamax, MEM-NEAA, 0.1 mM 2-mercaptoethanol, peni/strep) containing Y27632 (20 μM). Media was changed daily. On days 1-5, the SMAD inhibitors, dorsomorphin (5 μM, Sigma) and SB-431542 (10 μM, Tocris) were added to inhibit both the BMP and TGF-β signaling pathways. On days 6-24, media was replaced with neural medium (Neurobasal A, B-27 without Vitamin A, Glutamax, Peni/Strep) containing the growth factors, bFGF (20 ng/ml, Peprotech) and EGF (20 ng/ml, Peprotech). On day 25, spheroids were transferred to neural medium containing BDNF (20 ng/ml, Peprotech) and NT-3 (20 ng/ml, Peprotech) until day 42. hSS also received the Wnt pathway inhibitor, IWP-2 (5 μM, Selleckchem) on days 4-22 and the Smo pathway activator, SAG (100 nM, Selleckchem) on days 12-22 (Supplementary Figure 2A).

On day 60, hCS and hSS spheroids were homogenized and RNA was prepared using the Qiagen RNEasy Plus Micro Kit according to manufacturer’s instructions. Library construction, sequencing, and analysis were performed by NovoGene.

Spheroids were rinsed in DPBS, then fixed in 4% paraformaldehyde for 30 minutes - overnight at 4°C. After rinsing in DPBS, spheroids were incubated in 30% sucrose for 24-48 h at 4°C. Spheroids were embedded in OCT compound and frozen. A cryostat was used to cut 14-25 micron sections, which were mounted on gelatin-coated slides. For immunohistochemistry, slides were washed in PBS containing 0.3% Triton X-100, then blocked with 3-5% normal goat serum for 1 h. Slides were incubated in primary antibody (Table 1) overnight at 4°C in a humidified chamber. After washing, slides were incubated in secondary antibody (Jackson Immunoresearch, 1:500) for 1-2 h at room temperature. Slides were washed, DAPI (Sigma) was added to label nucleus, then coverslipped using Poly-Vinyl Alcohol (Sigma). Sections were imaged using a Leica Microscope (Spe8Info) and fluorescence intensity was analyzed using Image J software.

Human brain development, we first generated cortical spheroids and hSS were treated with vehicle, buprenorphine, or oxycodone for 10 days. On day 60, spheroids were homogenized and RNA was prepared using the Qiagen RNEasy Plus Micro Kit according to manufacturer instructions. RNA concentration and quality were determined using a NanoDrop, then total RNA was converted to cDNA using the Applied Biosystems High Capacity Reverse Transcription Kit. Real-time quantification of diluted cDNA was performed in triplicate reactions containing sample (4 ng), IDT Primetime Gene Expression Master Mix (2X), and IDT Primetime qPCR Assay (20X) on a BioRad CFX384 Real Time System. Cycling Conditions consisted of one cycle at 95°C for 3 min, followed by 40 cycles of denaturation (95°C for 15 s) and elongation (60°C for 1 min). The relative gene expression was calculated using the 2^–ΔΔCT method. The following IDT Primetime Gene Expression Assays were used: DLX1 (Hs.PT.58.4632198), DLX5 (Hs.PT.58.45646524), LHX6 (Hs.PT.58.27682011), ARX (Hs.PT.58.213.7622), NKX2.1 (Hs.PT.58.2461055), ADCY3 (Hs.PT.569.2298693), BCL11B (Hs.PT.58.27217530), MKI67 (Hs.PT.58.27920212), PAX6 (Hs.PT.58.25914558), SATB2 (Hs.PT.58.24560574), TBR1 (Hs.PT.58.24815929), TUBB3 (Hs.PT.58.20385221) and GAPDH (Hs.PT.39a.22214836).

NOP receptor-mediated inhibition of adenylyl cyclase activity was determined by measuring cAMP accumulation in the presence of the adenylyl cyclase activator, forskolin, and the phosphodiesterase inhibitor, rolipram as previously described (Berg et al., 2007). Briefly, spheroids were incubated in wash buffer (HBSS containing 20 mM HEPES, pH 7.55) for 15 min at 37°C. Rolipram (100 μM, Sigma) was added along with the following drugs: buprenorphine (5 nM), nociceptin (150 nM, Tocris), or oxycodone (60 nM) in the presence or absence of the NOP Receptor antagonist, UFP-101 (10 μM, Tocris). Forskolin (10 μM, Sigma) was then added and spheroids were incubated for an additional 15 minutes. Incubation was terminated by aspiration of the wash buffer and addition of 500 μl of ice cold ethanol. The ethanol extracts from individual spheroids were dried under a gentle airstream and reconstituted in 100 μl of sodium acetate (pH 6.2) and the cAMP content was determined by radioimmunoassay (RIA).

To record network activity, the Harvard MCS Multiwell-MEA-system, which contains a 12-electrode grid in each well of a 24 well plate, was used. Twenty-four hours before the MEA was conducted, spheroids were transferred to the MEA plate containing BrainPhys media (STEMCELL Technologies). For acute experiments, a 10-min baseline recording was followed by treatment with buprenorphine, or oxycodone in the presence of vehicle or the NOP receptor antagonist, UFP-101. Fifteen minutes later, 3 10-min recordings were performed. For the chronic experiments, hCS or fused hCS-hSS spheriods were treated with vehicle, buprenorphine, or oxycodone for 10 days. On day 60 or day 120, spheroids were transferred to MEA plates containing BrainPhys media and 3 10-min recordings were performed. Data was sampled at 20 kHz, digitized and analyzed using the Harvard Multiwell-Analyzer software with a 100 Hz high pass and 3.5 kHz low pass filter and an adaptive spike detection threshold was set at 8 times the standard deviation for each electrode. The results from each 10-min recording were averaged.

On day 30, hSS were infected with Lenti hDlx1/2b:GFP (gift from Dr. John Rubenstein). On day 40, one hSS and one hCS were transferred to an Eppendorf tube containing neural media and incubated without disruption for 3-4 days. After fusion was complete, fused spheroids were treated with buprenorphine or oxycodone from days 50-60. On day 60, fused spheroids were fixed, sliced and imaged. The total number of GFP + cells and the number of GFP + cells that migrated into the hCS were counted.

Pregnant female rats were injected intraperitoneally with 1 mg/kg buprenorphine or 10 mg/kg oxycodone on gestational days 11-21, a time point that encompasses both cortical and subcortical neurogenesis and neuronal migration (Schepanski et al., 2018). Male and female pups were weaned on postnatal day 21 and all experiments were performed in adult animals (after postnatal day 42). Animals were maintained in a temperature-controlled environment on a 14:10 h light-dark cycle and had access to food and water ad libitum. All procedures were consistent with NIH guidelines (NIH publications no. 80-23, revised 1978) and approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at San Antonio.

Adult rats were transcardially perfused with saline followed by 4% paraformaldehyde. Brains were post-fixed and cryopreserved for at least 24 hours. Coronal sections (50 μm) through the medial prefrontal cortex were blocked (5% normal goat serum), incubated with a mouse anti-Lhx6 primary antibody (Santa Cruz, 1:250), then washed and incubated with a goat anti-mouse Alexafluor488 secondary antibody (Invitrogen). Sections were mounted, coverslipped with Prolong Gold Antifade Reagent (Molecular Probes), then imaged using a Zeiss AxioObserver inverted microscope. To analyze the migration of interneurons into the cortex, 10 equidistant bins through the medial prefrontal cortex were defined to determine the distribution of cells across the cortical layers (Meechan et al., 2012). Lhx6 + cells were counted in at least 3 sections per animal by a blind experimenter.

Fetal brain bulk RNA sequencing dataset was downloaded from BrainSpan: Atlas of the Developing Human Brain¹. RPKM (reads per kilobase transcripts per million mapped reads) values for OPRL1 gene were mined and normalized for each sample using the following formula RPKM(gene)/(ΣRPKM(all genes in the sample)) X 10^6. Graphpad Prism was used to plot OPRL1 normalized counts from 8-37 weeks post conception and median values were used to draw conclusions about the temporal expression pattern.

In all figures, data are shown as mean + S.E.M and n is indicated in the figure legend. Data was analyzed by one- or two-way ANOVA and the Holm-Sidak post hoc test was used when significant interactions were present. When comparing groups with unequal variances the non-parametric Kruskal-Wallis test was used followed by Dunn’s multiple comparison test. All tests were two-tailed, and significance was determined at p < 0.05. Effect size was calculated using: https://www.psychometrica.de/effect_size.html.

To study the effect of buprenorphine during human brain development, we first generated cortical spheroids (hCS) and subpallial spheroids (hSS) from hiPSCs and ESCs using the method described by Pasca and colleagues Supplementary Figure 2A (PaŞca et al., 2015; Birey et al., 2017). Both cell lines expressed pluripotent stem cell markers (Supplementary Figure 1A) and maintained normal karyotype (Supplementary Figure 1B). RNA sequencing was used to confirm spheroid identity. Neither cortical spheroids nor subpallial spheroids expressed NANOG, a homeobox protein involved in maintaining embryonic stem cell pluripotency (Supplementary Figure 2B) (Chambers et al., 2007). Cortical spheroids expressed PAX6, a marker for dorsal forebrain progenitor cells (Georgala et al., 2011), and NEUROD6, a gene expressed ubiquitously throughout the fetal cerebral cortex (Supplementary Figure 2B) (Camp et al., 2015). These genes were not expressed in subpallial spheroids. Conversely, subpallial spheroids expressed ARX, DLX2, and NKX2-1 (Supplementary Figure 2B). DLX2 has been shown to bind to ARX and regulate the migration of GABAergic neurons (Colasante et al., 2008). Similarly, NKX2-1 is a homeobox transcription factor that regulates the fate specification and migration of GABAergic interneurons (Du et al., 2008). These subpallial markers were not observed in hCSs. Immunohistochemistry was used to demonstrate that hCS showed VZ-like structures and PAX6 expression (Supplementary Figure 2C). Cortical spheroids did not express ARX (Supplementary Figure 2C). Conversely, hSS presented both NKX2.1 + and ARX + cells (Supplementary Figure 2D). These data showed hCS and hSS resembled features of cortical and subpallial brain development in vitro.

Buprenorphine is thought to have its therapeutic effect via the μ- and κ-opioid receptors; however, buprenorphine also acts as a full agonist at the nociceptin opioid peptide (NOP) receptor (Kumar et al., 2014). Therefore, expression of these opioid receptors was first determined in hCS and hSS using RNA sequencing. In line with the pattern of opioid receptor expression in the human fetal brain (Allen Institute for Brain Science, 2008), neither the μ-, κ-, nor δ- opioid receptors were significantly expressed in either the cortical or subpallial spheroids (Figure 1A). However, the NOP receptor is expressed in both cortical and subpallial spheroids (Figure 1A). To identify the cell types expressing NOP, immunohistochemistry was used (Figures 1B-E). Robust NOP expression was observed in both cortical (Figure 1B) and subpallial (Figure 1D) spheroids at both 60 and 120 DIV. Further, NOP expression co-localized with Ki67 and Tuj1, suggesting this receptor is expressed on both proliferating cells and immature neurons (Figures 1C,E). In addition, we confirmed the expression of NOP receptor in human fetal tissue using publicly available dataset downloaded from BrainSpan, Atlas of the Developing Human Brain (See Text Footnote 1) (Figure 1F). After confirming NOP receptor expression, cAMP assays (Figures 1G-H) and multielectrode arrays (Supplementary Figures 3A-C) were performed on Day 50 cortical spheroids to determine if buprenorphine can have functional effects on the spheroids via the NOP receptor. Activation of the NOP receptor, which typically signals through Gi proteins, leads to an inhibition of adenylyl cyclase activity and decreases in cellular cAMP levels. Therefore, cAMP assays were first used to determine whether buprenorphine can activate intracellular signaling pathways via the NOP receptor. We found that buprenorphine decreased cAMP accumulation compared to vehicle-treated spheroids (Figure 1H, Two-way ANOVA: Interaction F(3,68) = 3.33, p < 0.05; Opioid F(3,68) = 11.90, p < 0.05; NOP Antagonist F(1,68) = 50.75, p < 0.05; Holm-Sidak Post Hoc Test: Vehicle-Vehicle vs. Buprenorphine-Vehicle t = 3.42, p < 0.05). These results are in line with the ability of nociceptin, the endogenous NOP receptor ligand, to decrease cAMP accumulation (Figure 1H, Holm-Sidak Post Hoc Test: Vehicle-Vehicle vs. Nociceptin-Vehicle t = 5.48, p < 0.05). Pretreatment with the NOP receptor antagonist UFP-101 completely reversed buprenorphine- and nociceptin-induced decrease in cAMP accumulation (Figure 1H, Buprenorphine-Vehicle vs. Buprenorphine-UFP t = 3.16, p < 0.05, Nociceptin-Veh vs. Nociceptin UFP t = 6.32, p < 0.05), confirming that in hCS, buprenorphine can signal through the NOP receptor.

To determine whether buprenorphine’s inhibition of cAMP signaling was accompanied by a change in neural activity, MEA recordings of cortical spheroids were then conducted to investigate changes in extracellular activity. Following acute exposure to the opioids, there was no significant effect of either oxycodone or buprenorphine on neuronal activity (Supplementary Figures 3A-C). To then determine if chronic opioid exposure affects cortical development, hCS were treated with buprenorphine or oxycodone for 10 days (day 50-60) and qPCR and MEAs were performed (Supplementary Figures 3D-G). While chronic treatment with either buprenorphine and oxycodone altered expression of some genes in hCS (Supplementary Figure 3E), neither drug produced significant changes in cortical activity (Supplementary Figures 3F-G). Similarly, when hCS were treated chronically with opioids (day 50-60), then tested at day 120, there were also no significant differences in neuronal activity (Supplementary Figures 3H-J).

After demonstrating that buprenorphine can influence the intracellular signaling in cortical spheroids via the NOP receptor, the effect of buprenorphine on subpallial spheroids was also determined. Day 60 subpallial spheroids were analyzed by qPCR (Supplementary Figures 4A) and immunohistochemistry (Figure 2A) after a 10-day treatment with oxycodone or buprenorphine. First, qPCR was used to measure mRNA expression of ARX, DLX1, LHX6, and NKX2-1, genes involved in the development and migration of interneurons (Wonders and Anderson, 2005). In contrast to oxycodone, buprenorphine slightly increased interneuron gene expression although this was not statistically significant (Supplementary Figure 4B; Two-way ANOVA: Opioid F (2,164) = 3.06, p < 0.05). Then, ARX immunolabeling was performed and a significant increase in ARX + cells in hSS after opioid exposure was observed (Figures 2B,C). Moreover, no difference was found in the number of AC3 + cells (cleaved caspase 3) after opioid treatment (data not shown). These data suggest that opioids can affect interneuron progenitors during brain development without affecting cell survival.

Migration of interneurons from the ganglionic eminences into the cortex is essential to cortical development (Wonders and Anderson, 2005). In order to determine the effect of opioids on migrating interneurons, subpallial and cortical spheroids were fused and the migrated GFP-labeled interneurons was analyzed after 20 days (Figure 3A). There was a significant effect of opioid treatment on the percent of GFP-positive cells that migrated into the cortical spheroids from the subpallial area (Figures 3B,C; One-way ANOVA F = 10.63, p < 0.05). Post hoc analysis demonstrated that buprenorphine produced a significant increase in the percentage of GFP-positive cells that migrated into the cortex compared to both control- and oxycodone-treated spheroids (Holm-Sidak: Control vs. Buprenorphine t = 4.201; Buprenorphine vs. Oxycodone t = 3.748). Next, we determined if this change in interneuron migration had an effect on network activity using MEA (Figure 3D). Cortical spheroids were treated with vehicle, buprenorphine, or oxycodone for 10 days. At day 60, MEAs were conducted to determine the effect of chronic opioid exposure on the extracellular activity of the spheroids. Although not significant, buprenorphine-treated spheroids exhibited a trend toward increased activity compared to the vehicle- and oxycodone-treated spheroids [Figure 3D, One-way ANOVA F(2,11) = 0.89, p > 0.05; Effect size: control vs. oxy = 0.049 (no effect), control vs. bup = 0.806 (large effect)].

To determine if buprenorphine can also disrupt interneuron migration in vivo, pregnant rats were injected with buprenorphine or oxycodone on gestational days 11-21 (Figure 4A). In adult offspring, the total number and distribution of Lhx6 + interneurons were analyzed throughout the medial prefrontal cortex (mPFC; Figure 4B). Lhx6, a LIM homeodomain transcription factor, is involved in the development of specific classes of GABAergic interneurons, namely somatostatin and parvalbumin subtypes, and expression is maintained in mature interneurons (Zhou et al., 2015).Prenatal opioid exposure had no effect on the total number of Lhx6 + positive cells in the mPFC (Figure 4D; One-way ANOVA F = 0.65, p > 0.05) or the thickness of the cortex (data not shown). However, the laminar distribution of Lhx6 + cells was altered by prenatal buprenorphine exposure (Figures 4C,D; Mixed effects ANOVA Layer F(2.9,20.4) = 67.45, p < 0.05). Specifically, in bins 7 and 8, there were significantly more Lhx6 + cells in the buprenorphine-treated animals (Holm-Sidak Bin 7 t = −2.4, p < 0.05; Bin 8 t = −3.9, p < 0.05). Together, these results suggest that prenatal buprenorphine exposure can alter interneuron migration into the developing cortex, and the altered distribution of interneurons can persist into adulthood.