Human cellular models to study the neurodevelopmental effects of EtOH exposure
To investigate the neurodevelopmental consequences of EtOH exposure, human cell platforms encompassing diverse neural cell types were employed. We derived functional cortical organoids and astrocytes from three hiPSC lines (Fig. 1a and Supplementary Fig. 1b–d, e) and prepared primary neuron cultures from human fetal tissue at 10-11 PCW (Fig. 1a and Supplementary Fig. 1h). Each cell type underwent moderate-high physiological concentration of EtOH exposure for one week (~4-week-old organoids, fetal neurons, and ~2–4-week-old astrocytes), followed by continued maintenance—fetal neurons and astrocytes for one week and one-to-two months for cortical organoids—in the absence of EtOH, at the end of which experimental assays were performed (Fig. 1a). Although the choice to administer 100 mM of EtOH is high, the concentration to which the cells were kept corresponds to around 20 mM or 90 mg/dL (Supplementary Fig. 1a). This treatment dosage was selected considering EtOH cytotoxicity and its concentration in amniotic fluid [36,37,38,39].
Fig. 1: Ethanol (EtOH) exposure alters proliferation and survival dynamics in human cellular models. a EtOH was applied to ~4-week-old human iPSC-derived cortical organoids, ~2–4-week-old human iPSC-derived astrocytes, and human 10–11 post-conception weeks (PCW) fetal primary neurons. b Cortical organoid diameter is decreased by EtOH exposure (Student’s t-test, t 541 = 17.05, P < 0.0001; n = 210–333 organoids/condition). c Immunohistochemistry portrays layers of Ki67+ proliferation (two months, Student’s t-test, t 64 = 4.518, P < 0.0001; three months, Student’s t-test, t 18 = 1.63, P = 0.12; n = 10–33 organoids/condition using WT83, CVB, and 4C1 cell lines). d–f EtOH exposure produces cell cycle alterations in cortical organoids (two-way analysis of variance (ANOVA), F 3,16 = 8.14, P = 0.002; n = 3 replicates/condition, from WT83, CVB, and 4C1 cell lines) (d), iPSC-derived astrocytes (two-way ANOVA, F 2,42 = 17.06, P < 0.0001; n = 8 replicates/condition, from WT83 and 4C1 cell lines) (e), and fetal neurons (two-way ANOVA, F 2,42 = 41.92, P < 0.0001; n = 8 replicates/condition) (f). g–i EtOH exposure promotes cell death, shown by CC3+ immunohistochemistry in cortical organoids (two months, Student’s t-test, t 28 = 5.785, P < 0.0001; three months, Student’s t-test, t 18 = 1.886, P = 0.076; n = 10–15 organoids/condition from WT83 cell line) (g) and Annexin+ in astrocytes (Student’s t-test, t 14 = 15.03, P < 0.0001; n = 8 replicates/condition from WT83 and 4C1 cell lines) (h) and fetal neurons (Student’s t-test, t 6 = 7.872, P = 0.0002; n = 4 replicates/condition) (i). Data are presented as mean ± standard error of the mean (s.e.m.). Scale bar = 100 µm. Full size image
Immunocytochemistry and transcriptomics indicated close fidelity between cortical organoids and human fetal cortical development. Immunocytochemistry portrayed SOX2+ progenitor regions enriched for proliferation (Ki67+) organized in ventricular zone (VZ)-like areas around a lumen, giving rise to a cortical plate-like area displaying mature neurons (NeuN+, MAP2+) that express the cortical laminar-specific markers CTIP2 and SATB2 (Supplementary Fig. 1b, c). Furthermore, transcriptomic comparison of two- and three-month-old organoids (see below) with the BrainSpan human brain gene expression reference [40] showed close correlation between organoids and developing 9–13 PCW fetal cortex (Supplementary Fig. 1d and Supplementary Table 1). Because glia cells emerge later during neurodevelopment and are thus absent or sparse in early cortical organoids [23] (Supplementary Fig. 1c), we differentiated astrocytes independently to capture astrocytic alteration attributable to EtOH exposure.
EtOH exposure alters cell cycle, proliferation, and survival
EtOH-exposed cortical organoids showed reduced diameter at two months compared to controls (P < 0.0001; Fig. 1b). Ki67+ staining was used to delineate the proliferative VZ-like region from the cortical plate-like domain, as previously described in a cortical organoid system [41]. At two months of age, a stage in organoid development with considerable cell proliferation, we observed a smaller population of Ki67+ proliferative cells in EtOH-exposed cultures compared to controls (P < 0.0001); however, at three months, this difference had narrowed (P = 0.12; Fig. 1c). The overall amount of cells (DAPI+) remained constant, independent of treatment. Next, cell cycle analysis showed that cells from EtOH-exposed organoids were less likely to be in the proliferative S and G2/M phases (P = 0.002; Fig. 1d). Similar alterations in cell cycle were also observed in EtOH-exposed astrocytes (P < 0.0001; Fig. 1e and Supplementary Fig. 1g) and fetal neurons (P < 0.0001; Fig. 1f and Supplementary Fig. 1j).
After identifying alterations in proliferation and cell cycle progression, we investigated the effect of EtOH exposure on cell survival and found a higher frequency of apoptotic cells (CC3+) in two-month EtOH-exposed organoids compared to controls (P < 0.0001; Fig. 1g). Analogous to the proliferative impairment, we observed that by three months the increased frequency of apoptosis in EtOH-exposed organoids had comparatively stabilized (P = 0.076; Fig. 1g). To evaluate cell death in astrocytes and fetal neurons, we assessed the frequency of Annexin+ cells, which was also increased in both astrocytes (P < 0.0001; Fig. 1h and Supplementary Fig. 1f) and fetal neurons (P = 0.0002; Fig. 1i and Supplementary Fig. 1i) exposed to EtOH.
Epigenetic landscape during neurodevelopment is altered by EtOH exposure
Epigenetic signaling is the principal mediator of the transcriptional response to environmental stimuli, a dynamic process in which stimulus-induced shifts in chromatin accessibility modulate differential gene expression. To define alterations in the chromatin human neurodevelopmental regulatory landscape due to EtOH exposure, we performed ATAC-seq in control and EtOH-exposed cortical organoids [42]. A high-quality library (Q30 > 90) was generated for each sample, and we detected a total of 604,998 peaks (Supplementary Tables 2 and 3); EtOH-exposed organoids exhibited 20,731 peaks that were both common between replicates and distinct from controls. Annotating the peaks by genomic region suggested preferential enrichment of intergenic regions expected to have a regulatory function (Fig. 2a, b, and Supplementary Fig. 2a–e). Analysis of chromatin motifs showed an association of EtOH exposure with enrichment of different sequences linked with proliferation and neurogenesis, including NF1, NeuroD1, and Ascl1 (Fig. 2c, d). For instance, ARHGEF10 and ARHGAP23 are linked with rho GTPases, and NRXN1 and NRXN3 produce synaptic cell adhesion molecules to promote synaptic structural integrity (Fig. 2e, f). A summary Reactome analysis confirmed the prominent neurodevelopmental consequences of EtOH, with affected functions including axon guidance, cellular adhesion, and synaptic communication (Fig. 2g).
Fig. 2: EtOH affects chromatin accessibility in regions critical for neurodevelopment. a–g ATAC-seq analysis of two-month old control and EtOH-exposed cortical organoids. a Transcription start site (TSS) enrichment plot showing TSS ± 1.0 Kb for each sample. ‘Control 1’, ‘Control 2’, ‘EtOH 1’, and ‘EtOH 2’ labels denote independent batches of organoids made from the WT83 cell line. b Venn diagram depicts peaks in EtOH-exposed and control organoids. c, d Motif enrichment in control (c) and EtOH-exposed (d) organoids. e Plotting enrichment P values identifies prominent regions of altered accessibility. f Tracks plot of select prominent genes. g Reactome analysis predicts that the effects of altered chromatin accessibility concentrate in physiological processes central to neurodevelopment. h, i Mass spectrometry analysis shows histone modifications in astrocytes, cortical organoids, and fetal neurons (hNE) attributable to EtOH-exposure; cortical organoids and astrocytes were generated from the WT83 iPS cell line. Heatmap shows EtOH exposure modifies histone methylation (ME) and acetylation (AC) patterns in diverse neural cell types; scale bar portrays relative abundance of modification across samples: red=high compared to other samples, blue=low compared to other samples (h). A recurrent pattern of modifications is observed between astrocytes, organoids, and fetal neurons; only statistically significant differences in relative modification abundance are shown (i). Full size image
In order to extend our ATAC-seq findings, we performed mass spectrometry to characterize EtOH-induced histone modifications in astrocytes, cortical organoids, and primary fetal neurons. Compared to controls, EtOH exposure was associated with specific enrichment of H3K9ac; H3K27me2,3; and H3K9me2, but we also observed a broader epigenetic shift: Histone methylation and acetylation were often decreased and increased, respectively, in EtOH-exposed cells (Fig. 2h, i and Supplementary Table 4). Histone deacetylase and methyltransferase expression were concordantly increased between all cell types in association with EtOH exposure. Moreover, EtOH appeared to induce a systematic decrease in the methylation of promoter regions (Fig. 2h, i). Even though particular histone modifications (namely, those shown in Fig. 2i) showed a statistically significant difference between EtOH-exposed cells and controls, the quantitative difference between particular modifications or patterns of modifications was often subtle, perhaps in reflection of strict epigenetic regulation [43]. Nevertheless, together these findings suggest that EtOH exposure resulted in an epigenetic shift affecting chromatin accessibility and histone post-translational modifications.
EtOH exposure is associated with transcriptomic alteration in neurodevelopmental pathways
EtOH-induced alterations in chromatin accessibility are expected to contribute to transcriptomic changes that underlie PAE neural cytopathology. We compared gene expression profiles between control and EtOH-exposed organoids and fetal neurons. Broadly, EtOH-exposed cortical organoids and fetal neurons exhibited differential global gene expression and clustered distinctly from controls (Fig. 3a; Supplementary Fig. 3a, b; and Supplementary Table 5). The affected transcriptomic pathways in cortical organoids include protein processing and cell cycle regulation (Fig. 3b and Supplementary Fig. 3c), also identified by ATAC-seq analysis. Many of the altered processes appeared to interact (Fig. 3c), underscoring the breadth of cellular impairment by EtOH. Summary gene ontology analysis in fetal neurons likewise implicated the effects of EtOH on cell cycle and cytoskeletal structure (Fig. 3d and Supplementary Fig. 3d). EtOH induces alterations in cellular signaling pathways, ligand-receptor interaction, and neurotransmission at a later stage in cortical organoids, including excitatory and GABAergic signaling and astrocytic function (Fig. 3e–h and Supplementary Fig. 3e). In addition, cAMP, PI3K-Akt, and Hedgehog signaling were impaired, leading to alterations in cell viability, calcium homeostasis, synaptic vesicle cycling, and long-term potentiation (Fig. 3i, j and Supplementary Fig. 4).
Fig. 3: EtOH alters neurodevelopmental transcriptional pathways in cortical organoids and fetal neurons. a RNAseq heatmap within cortical organoid and fetal neuron subgroups shows differential gene expression clustering by EtOH exposure. Scale indicates Pearson’s correlation coefficient. Cortical organoids are from the WT83 iPS cell line. b Visualization of prominently altered cell processes in two-month old EtOH-exposed organoids. Pathway enrichment analysis of genes altered by EtOH exposure is displayed according to total perturbation accumulation (pACC) and gene overrepresentation within pathways (pORA). Red dots represent pathways modified with a significant P value, proportional to the size of the dots. c Interactome depicting the interrelatedness of these processes. d Cross-comparative gene ontology analysis of EtOH-exposed fetal neurons shows overlapping pathway alterations with organoids. e Plot of P values for gene expression in three-month old organoids portrays range of alteration severity. f Visualization of prominently altered cell processes in three-month old EtOH-exposed organoids. g Gene ontology analysis of three-month old organoids shows gene expression alterations prominently affect neurotransmission and biochemical signaling pathways. h Interactome highlights involvement of key processes including excitatory and GABAergic signaling and astrocytic expression. i, j EtOH exposure impacts gene expression in prominent pathways with multifactorial downstream consequences. Shown are EtOH-induced changes in the cAMP signaling pathway (I) and glutamatergic synaptic function (j). Full size image
Digital spatial profiling portrays differential proteomics of EtOH-exposed cortical organoids
The combined analysis of epigenetic and transcriptomic profiles implicated synaptic connectivity and astrocytic alteration defects due to EtOH exposure, leading us to target these properties for further investigation. Western blot analysis showed EtOH-exposed cortical organoids exhibited altered content of hallmark neural proteins, including SYN1, PSD95, and GFAP (Fig. 4a, b and Supplementary Fig. 5).
Fig. 4: EtOH alters astrocytic and synaptic protein quantities and promotes differential protein abundance in cortical organoids. a, b Western blot and quantification of astrocytic and synaptic protein markers in cortical organoids during development (two-way ANOVA for each protein; *P < 0.05, **P < 0.01, ***P < 0.001; n = 3 replicates (~5–10 organoids)/condition from independent organoid batches from the WT83 iPS cell line). c Principal components analysis of protein digital spatial profiling distinguishes distinct clusters between proliferative and non-proliferative regions and between EtOH-exposed and control samples (n~10 control and EtOH-exposed organoids) (d) Heatmap showing variation in protein abundance in organoid samples between distinct spatial and EtOH-exposure clusters. e Volcano plot of proliferative (rosette) regions from control and EtOH-exposed organoids highlights the differences in expression of various proteins. Green bar represents a P value of 0.05. Full size image
In addition to altered quantity, we sought to further determine whether EtOH exposure affects spatial patterns of neural protein expression within cortical organoids using Digital Spatial Profiling [35], enhancing our analysis with commercial protein panels for neural cell subtyping and cellular neuropathology. Principal components analysis comparing spatial protein expression in cortical organoids distinguished four distinct clusters based on organoid EtOH treatment status (EtOH-exposed vs control) and regional protein localization within the organoid (rosette or non-rosette; Fig. 4c). Interestingly, upon assessing how the expression of particular proteins varies between clusters, it was observed that EtOH-exposed organoids showed a greater abundance of degenerative proteins than controls (Fig. 4d). In particular, proliferative rosette-like regions of EtOH-exposed organoids portrayed enrichment of immune response and degenerative proteins such as Neurofilament light, HLA-DR, CD-45, CD-68, TMEM119, Phospho-Tau (S199) and (S396), and TDP-43 (Fig. 4d, e). In contrast with proliferative rosettes, spatial profiling of other neural proteins, such as synaptophysin, showed the most pronounced differences in the non-rosette cluster (Fig. 4d), corresponding to the more maturely differentiated regions of the organoid. This contrasting spatial expression profile reflects these proteins’ roles in more mature neural cells, such as neurons and glia, leading to further investigation of these cell types.
EtOH exposure increases astroglia cluster content and impairs synaptogenesis and connectivity
Given convergence of the epigenetic, transcriptomic, and proteomic findings towards EtOH-induced defects in synaptic connectivity and astrocytic alterations, we sought to gain insights into a potential cellular vulnerability by quantifying cell type in cortical organoids. EtOH exposure was associated with higher percentages of cells in clusters expressing the astrocytic GFAP marker at two (P = 0.007) and three months of age (P = 0.04; Fig. 5a). Despite the astrocytic increase, the neuronal fate seemed unaffected because quantities of the cortical layer V/VI neuronal marker CTIP2 remained equivalent (Fig. 5a). Impairment of synaptogenesis was evaluated by quantifying synaptic puncta density, characterized by co-localization of the presynaptic Vglut1 and postsynaptic Homer1 proteins. Compared to controls, reduced synaptic puncta density was observed in two-month-old, EtOH-exposed organoids (P = 0.01; Fig. 5b), suggesting early EtOH exposure can impact neuronal connectivity for an extended period.
Fig. 5: EtOH exposure increases astrocytic content and impairs network connectivity. a Immunohistochemistry reveals no difference in the proportion of CTIP2+ cells (left graph; two months, P = 0.88; three months, P > 0.99; n = 3 organoids per age, per condition from WT83 and 4C1 iPS cell lines) but shows EtOH increases GFAP expression (right graph; two months, P = 0.017; three months, P = 0.013; two-way ANOVAs; n = 3 organoids per age, per condition from WT83 and 4C1 iPS cell lines). Scale bar = 50 μm. b EtOH exposure decreases co-localized synaptic puncta density (Student’s t-test, t 32 = 2.59, P = 0.01; n = 17 neurons/condition from WT83, CVB, and 4C1 iPS cell lines). Scale bar = 10 μm. c–e Multi-electrode array (MEA) analysis in WT83-based cortical organoids (c, d, n = 18 MEA wells/condition) and fetal neurons (e, f, n = 11–22 MEA wells/condition). c Representative cortical organoid MEA raster plot. d EtOH-exposed cortical organoids showed fewer spikes per minute (Student’s t-test, t 34 = 2.68, P = 0.011) and fewer bursts (Student’s t-test, P = 0.047), although not network bursts (Student’s t-test, t 34 = 0.06, P = 0.95). e Representative fetal neuron MEA raster plot. f EtOH-exposed fetal neurons showed fewer spikes per minute (Student’s t-test, t 30 = 2.68, P = 0.012) and fewer bursts per minute (Student’s t-test, t 40 = 2.60, P = 0.013), but not network bursts (Student’s t-test, t 31 = 1.78, P = 0.085). Data are presented as mean ± s.e.m. Full size image
Multi-electrode array electrophysiology recordings enable dynamic interrogation of how human neurons behave in circuits, a valuable approach given that connectivity cannot be assessed during the early stages of human corticogenesis in utero. We plated one-month-old paired control and EtOH-exposed organoids to mature in MEA plates and recorded their activity (Fig. 5c, e). Compared to controls, EtOH-exposed cortical organoids spiked fewer times per minute (P = 0.01) and had fewer bursts (P = 0.047), although synchronous network bursts remained unchanged (P = 0.93; Fig. 5d). Finally, we used primary human fetal neural tissues to validate our results. EtOH-exposed fetal neurons exhibited a similar MEA profile with fewer spikes (P = 0.01) and fewer bursts (P = 0.01) per minute, but the decrease in synchronous network bursts did not reach significance (P = 0.08; Fig. 5f). These findings showed that EtOH-induced alterations in neuronal network formation persist despite cessation of the exposure.
Pharmacological reversal of synaptic deficits associated with EtOH exposure
Having identified a series of pathophysiological changes in diverse human neural cell models exposed to EtOH, we sought to investigate whether pharmacological intervention could ameliorate the severity of the synaptic phenotypes (given their prominent representation in our findings) utilizing an abridged version of the drug screening pipeline strategy recently shown by our group to be effective [26]. Four drugs were selected for screening in EtOH-exposed organoids (Fig. 6a) based on their predicted ability to rescue synaptic impairment: Nefiracetam, PHA-543613, Donepezil, and IGF-1. Nefiracetam and PHA-543613 were chosen because our group has previously demonstrated their ability to reverse synaptic impairment associated with disease [26]. Furthermore, a compound from the same racetam class of drugs in which Nefiracetam is included, aniracetam, has previously been shown to reverse neurocognitive deficits in a mouse model of PAE by modulating synaptic transmission [44,45,46], bolstering support for trial of Nefiracetam in our human cell models of PAE. Donepezil, a compound clinically indicated as first-line treatment to slow cognitive decline and dementia, was selected because chronic alcohol usage is associated with dementia and Alzheimer’s disease, and we identified changes in neurodegenerative proteins (e.g., phospho-tau isoforms and Aβ 1-42 ) in proteomic spatial profiling in EtOH-exposed compared to control organoids.
Fig. 6: Pharmacological reversal of synaptic impairment associated with EtOH exposure. a Schematic showing the drug treatment strategy of EtOH-exposed cortical organoids. b Western blot shows a reduced amount of the presynaptic protein Synapsin in EtOH-exposed organoids that is increased by treatment with Donepezil, but not Nefiracetam (Nefi) or IGF-1. c Levels of the postsynaptic protein PSD-95 were rescued by treatment with either Donepezil or Nefiracetam, but not IGF-1. d Donepezil treatment reversed the synaptic puncta impairment observed in EtOH-exposed organoids (one-way ANOVA with Dunnett’s test for multiple comparisons, F 3,173 = 10.86, P < 0.0001; Control vs EtOH, P < 0.0001; Donepezil vs EtOH, P = 0.024; n = 48 neurites/condition from WT83, CVB, and 4C1 iPS cell lines), with lesser support for the efficacy of Nefiracetam (P = 0.066; n = 33 neurites from WT83 and CVB iPS cell lines). Dnpzl Donepezil. Nefi Nefiracetam. Data are presented as mean ± s.e.m. Full size image
Western blot of synaptic proteins and synaptic puncta co-localization were used as pilot assays to evaluate the drug candidates’ potential to reverse synaptic phenotypes observed in EtOH-exposed organoids (Fig. 6a). Each of the four drug candidates was assessed via Western blot of synaptic proteins. Only Donepezil appreciably increased the amount of presynaptic SYN1 in EtOH-exposed organoids (Fig. 6b), although PSD-95 was notably increased by treatment with either Donepezil or Nefiracetam (Fig. 6c). Donepezil and Nefiracetam were subsequently evaluated for their potential to reverse synaptic puncta impairment in EtOH-exposed organoids (P < 0.0001; Fig. 6d). Co-localized puncta density in EtOH-exposed organoids was increased by treatment with Donepezil (P = 0.024; Fig. 6d). Although the increase from treatment with Nefiracetam did not reach statistical significance (P = 0.066; Fig. 6d), the 95% confidence interval (CI) for its effect nearly completely overlaps that of Donepezil (95% CI: EtOH vs. Donepezil, [−6.75, −0.37]; EtOH vs. Nefiracetam, [−6.89, 0.17]).