Establishment and Validation of a Model for Fetal Neural Ischemia Using Necrotic Core-Free Human Spinal Cord Organoids

Abstract

Fetal spinal cord ischemia is a serious medical condition that can result in significant neurological damage and adverse outcomes for the fetus. However, the lack of an appropriate experimental model has hindered the understanding of the pathology and the development of effective treatments. In our study, we established a system for screening drugs that affect fetal spinal cord ischemia using spinal cord organoids. Importantly, we produced necrotic core-free human spinal cord organoids (nf-hSCOs) by reducing the organoid size to avoid potential complications of spontaneous necrosis in large organoids. Exposing nf-hSCOs to CoCl₂ as a hypoxia mimetic and hypoglycemic conditions resulted in significant neuronal damage, as assessed by multiple assay batteries. By utilizing this model, we tested chemicals that have been reported to exhibit beneficial effects in brain organoid-based ischemia models. Surprisingly, these chemicals did not provide sufficient benefit, and we discovered that rapamycin is a mild neuroprotective reagent for both axon degeneration and neuronal survival. We propose that nf-hSCO is suitable for large-scale screening of fetal neural ischemia due to its scalability, ease of ischemic induction, implementation of quantifiable assay batteries, and the absence of spontaneous necrosis.

Graphical Abstract

Graphical Abstract.

Significance Statement.

Fetal neural ischemia can cause permanent neurological damage, and there is a need for proper experimental models to study this condition. In this study, we successfully established and tested the validity of a model for ischemia by combining CoCl₂ treatment and hypoglycemic conditions using necrotic core-free human spinal cord organoids. This platform holds promise in identifying new drugs that could potentially mitigate the damage caused by fetal neural ischemia.

Introduction

Agenesis of blood vessels, traumatic damages, and other unexpected events can cause an ischemic condition in the developing nervous system. This condition can lead to a wide range of adverse outcomes, including spontaneous abortion, stillbirth, and infant death. In surviving infants, fetal neural ischemia can result in permanent neurological damage, including motor and sensory deficits, and intellectual disabilities.^1-3 Fetal neural ischemia can be caused by a variety of factors, including obstetrical conditions such as placenta previa and umbilical cord prolapse, as well as premature birth, fetal anomalies, and genetic mutations. Although this fetal hypoxia (HX) causes diverse medical consequences depending on the causes and locations of the insults, there are few experimental models to address the specificity and universality of the pathology, primarily the absence of experimental models.

The use of neural organoids in the study of fetal neural ischemia holds great promise.^4,5 Brain organoids have been used to investigate the effects of HX conditions in extremely premature births and to test potential treatments for this condition. Several studies have demonstrated the utility of brain organoids in analyzing the cellular mechanisms and screening drugs for fetal brain ischemia.^6-9 Particularly intriguing is the finding that treatment with small molecules like ISRIB and minocycline can prevent cell death in brain organoids. ISRIB, a small-molecule modulator of the unfolded protein response (UPR) pathway, has been shown to protect intermediate progenitors from HX-induced reduction.⁶ Similarly, minocycline, an FDA-approved small molecule, has been found to prevent the suppression of HX-related gene expression, although the exact molecular mechanism is not yet fully understood.⁷ These studies provide valuable insights into the underlying mechanisms of fetal brain ischemia and offer potential new treatments to mitigate the occurrence and severity of this condition.

Large organoids often exhibit the presence of a necrotic core due to limited oxygen and nutrient diffusion, which can negatively impact their differentiation and maturation.^10,11 To minimize necrotic core formation, various approaches have been attempted. These include vascularization of neural organoids^12-14 or slicing/cutting them into smaller sizes.¹⁵ However, these methods add complexity to the procedure, which is not favorable for large-scale drug screening. Preventing necrotic core formation is particularly crucial in modeling neural ischemia, as the presence of a necrotic core in organoids can complicate data interpretation. Recently, we developed a protocol to generate human spinal cord organoids that faithfully replicate developmental morphogenesis, cellular heterogeneity, and functional excitability.^16,17 A key feature of our protocol is the ability to control the size of organoids by adjusting the number of neural-induced cells used for reaggregation. By reducing the initial cell numbers (ICN), we successfully produced necrotic core-free human spinal cord organoids (nf-hSCOs). Importantly, our 3D culture system enables the production of a large quantity of nf-hSCOs simultaneously, making it well-suited for drug screening and quantitative assay evaluation.

With this modification made to our approach, we decided to reevaluate the effectiveness of the chemical proposed for treating fetal brain ischemia using nf-hSCOs exposed to HX and hypoglycemia (HG). This served as a proof-of-concept for our nf-hSCO-based modeling of fetal neural ischemia. By optimizing the size of the organoids and implementing sensitive assay batteries, we were able to more accurately assess the efficacy of the drugs in our model. Surprisingly, we found that these chemicals only had a minimal positive effect on our model, whereas inhibition of the mTOR pathway yielded better results. Our screening system is well-suited for large-scale drug screening and will greatly contribute to the identification of effective molecules for the treatment of fetal neural ischemia.

Materials and Methods

Stem Cell and Organoid Culture

H9 human pluripotent stem cells (hPSCs) were cultured on matrigel-coated plates (Corning, Corning, New York) using mTeSR1 medium (STMECELL Technologies, Vancouver, Canada). The hPSCs were passaged every 5 days using ReLeSR (STEMCELL Technologies). To generate hSCOs, small clumps of dissociated hPSCs were plated at high density on matrigel-coated plates in mTeSR1 medium for 2 days. After 2 days, the medium was replaced with a differentiation medium (DM) composed of DMEM/F12 (Life Technologies, San Francisco, California), 1% N2 (Life Technologies), 2% B27 (Life Technologies), 1% Penicillin/Streptomycin (Life Technologies), and 0.1% β-mercaptoethanol (Life Technologies).

For caudal neural stem cell (cNSC) induction, hPSCs were treated with 10 µM SB431542 (Tocris, Bristol, UK) and 3 µM CHIR99021 (Sigma, St. Louis, Missouri) for 3 days. After 3 days, cNSCs were dissociated into single cells using Accutase (STEMCELL Technologies). The indicated number of dissociated cells was seeded onto a 96-well low attachment plate (75 cells per well; SPL, Pocheon, Republic of Korea) in DM with 20 ng/mL basic fibroblast growth factor (bFGF; R&D, Minneapolis, Minnesota) for 4 days to allow forming 3D aggregates. In the case of nf-hSCO, 75 cells were seeded in each well of the 96-well plate. To facilitate the formation of 3D aggregates, nf-hSCOs were cultured in DM containing 10 µM Rock-inhibitor (Tocris) only on the first day. After 4 days of daily treatment with bFGF, nf-hSCOs were cultured in DM with 0.1 µM RA (Sigma) without bFGF for 6 days. The medium was changed every other day.

For organoid maturation, nf-hSCOs were grown in a 1:4 mixture of DMEM/F12 and Neurobasal medium (Life Technologies) containing 2% B27, 1% Penicillin/Streptomycin, 1% Glutamax (Life Technologies), and 0.1 µM RA. The medium was changed every 4 days.

Treatments

To induce HX condition, nf-hSCOs were cultured in a maturation medium containing 300 µM cobalt chloride (CoCl₂; Sigma). Hypoglycemic condition (HG) was induced by replacing the medium with glucose-free neurobasal medium (Thermo Fisher Scientific, Waltham, MA) mixed with 2.5 mM Glucose solution (Thermo Fisher Scientific). A combination of CoCl₂ with hypoglycemic condition (HXHG) was induced in hypoglycemic medium containing 300 µM CoCl₂. For screening of drugs, nf-hSCOs were treated with drugs 2 days prior to exposure to HXHG conditions, and subsequently exposed to HXHG condition–containing drugs (ISRIB, 10 nM; Sigma, SML0843; rapamycin, 5 μM; Sigma, R0395; metformin, 10 mM; Sigma, PHR1084; minocyclin, 2 µM; sigma, M9511). The medium was changed daily.

Live/Dead Assay

Nf-hSCOs were treated with propidium iodide/acridine orange stain (PI/AO; 1:100; Aligned Genetics, Anyang, Republic of Korea) for 1 hour at 37 ℃ incubator. In this assay, living and dead cells are labeled separately with green and red fluorescence, respectively. Therefore, the ratio of these 2 fluorescence intensities sensitively represents the viability of the cells. After 1 h, green and red fluorescence images of nf-hSCOs were captured using a Juli Stage (NanoEnteck, Seoul, Republic of Korea) with 4× objective lens. For PI/AO analysis, red/green channel signal intensities from all images were measured by mean gray value using ImageJ program.

Axon Outgrowth on 2D Using nf-hSCOs

Nf-hSCOs (day 18) were adhered onto coverslips coated with matrigel and axonal outgrowth was allowed in maturation medium for 3 days. Then the organoids were exposed to the HXHG media for 3 days. Axonal area per organoid was quantified by the measurement of the area where the tips of each axon were connected.¹⁸

Immunocytochemistry

The nf-hSCOs were fixed with 4% paraformaldehyde (PFA, Biosesang, Seongnam, Repulic of Korea) for 30 minutes at room temperature (RT), followed by washing 3 times with phosphate buffered saline (PBS) with 0.1 % Triton X-100 and immersed in 30% sucrose in PBS at 4 C overnight. Subsequently, they were embedded in Tissue-Tek optimal cutting temperature (O.C.T. compound, SAKURA, Torrance, CA) on dry ice and were sectioned to 20 µm thickness. For immunostaining, sectioned samples were washed with PBS 3 times to remove OCT compound and blocked in blocking solution (3% BSA and 0.2% Triton X-100 in PBS) for 30 minutes at RT, and incubated with primary antibodies diluted in blocking solution at 4 C overnight (antibodies are listed in Supplementary Table S1). After incubation, primary antibodies are washed 3 times with PBS and incubated in blocking solution diluted secondary antibodies for 30 minutes at RT. Secondary antibodies were washed with PBS 3 times and section samples were mounted in Crystal Mount (Biomeda, San Jose, CA).

For whole organoid staining in 3D, the incubation time for the primary and secondary antibodies was elongated to 2 days for better penetration of the antibodies. The organoids were mounted with a clearing solution (25% urea and 65% sucrose in D.W).¹⁹ All images were acquired using a Leica TCS SP8 confocal microscope.

Real Time Polymerase Chain Reaction

Total RNA samples were isolated using Trizol reagent (Invitrogen, Waltham, MA). Isolated 1 µg RNA was used to synthesize cDNA using Moloney murine leukemia virus reverse transcriptase (MMLV, Promega, Madison, WI). cDNA was amplified using gene-specific primers (primer sequences are listed in Supplementary Table S2). Quantitative real time polymerase chain reaction (RT-PCR) was performed on QuantStudio 3 machine (Thermo Fisher Scientific) using SYBR Green master mix (Enzynomics, Daejeon, Republic of Korea) combined with specific primers. To calculate fold change, all values were normalized to GAPDH expression.

Western Blot

The nf-hSCOs were lysed in RIPA buffer (Bio-solution, Suwon, Republic of Korea) containing protease and phosphatase inhibitors (Roche, Basel, Switzerland). Whole lysates were loaded onto 4%-15% gels (Bio-Rad, Erucules, California) and transferred to a 0.45 µm polyvinylidene difluoride (PVDF) membrane (Sigma). The membrane was incubated in Tris-buffered saline with Triton X-100 (TBST) containing 0.1% Tween 20 and 5% skim milk at 4 ℃ overnight with the following primary antibodies: HX-inducible factor 1-alpha (HIF1α) (1:1,000, Cell signaling, Danvers, Massachusetts, #3716) and β-actin (1:1,000, Cell signaling, #3700). The following day, membranes were washed 3 times with TBST and probed horseradish peroxidase (HRP)-conjugated secondary antibodies (Enzo Life Sciences, Farmingdale, New York) for 1 hour. To visualize chemiluminescence, a signal was developed by using enhanced chemiluminescence (ECL) detection system (Bio-Rad).

Statistics

All data were repeated at least 3 times and analyzed using an unpaired t test or one-way ANOVA with Tukey’s test. Statistical analysis was performed using GraphPad Prism 9 software.

Results

Establishment of Nf-hSCO

We developed a protocol for producing small organoids by reducing the ICN during the brief dissociation and reaggregation step of 2D-primed neuroepithelial cells (Fig. 1A). The formation of a necrotic core, characterized by condensed nuclear profiles and cleaved caspase-3 (c-Casp3) staining, was observed at approximately 350 μm (mean ± SE = 347 ± 29 μm, n = 12 positions from 3 different organoids) distance from the organoid surface. Thus, organoids with an ICN of 5000 reached the size at which necrotic core formation occurred after 14 days of culture, while small organoids with an ICN of 75 did not reach a size that exhibited a risk of necrotic-core formation until 1 month (Fig. 1B; Supplementary Fig. S1A). To examine whether different cell lines (H9, H1, and hiPSC) would affect cell death after only seeding 75 cells, we immunostained organoids from 3 different cell lines using c-Casp3 antibody. As expected, the ICN 75 organoids produced with 3 different cell lines did not exhibit c-Casp3 signals (Supplementary Fig. S1B). Consequently, the necrotic core was absent in 1-month-old organoids generated from 75 cells, in contrast to conventional organoids produced from 5000 cells (Fig. 1B). These small organoids maintained the rosette morphology of SOX2-expressing neural stem cells and exhibited neuronal differentiation, as indicated by neurofilament-M (NF-M) labeling, similar to the larger organoids (Fig. 1C). Moreover, we confirmed the presence of neural stem cells (SOX2) and neurons (DCX, NF-M) on day 19. Also, we observed the expression of Brn3a, a marker for dorsal neurons (dl1-3). On days 29 and 34 of culture, expression of dorsal marker, PAX2 (dl4, dl6), PAX3 (d0P1-dP6), PAX7 (dP3-dP6) and ventral marker, PAX2 (V0-V1), EVX1 (V0). Therefore, small organoids revealed that organoids included both mitotic progenitor cells and post-mitotic cells with approximate dorsoventral identity. After 1 month, we observed the presence of the astrocyte marker GFAP and the oligodendrocyte marker CNPase. As the organoid became mature, we confirmed an increase in the levels of the postmitotic neuronal cell body marker (NeuN), neuron marker (internexin), and astrocyte maker (GFAP). We also included HOX gene expression by using RT-PCR. Although there were minor differences in gene expression levels compared to ICN 5000, small organoids demonstrated a posterior identity, as suggested by the HOX code associated with the cervical-thoracic level. Based on our characterizations, we termed nf-hSCOs (Supplementary Fig. S2A, S2B). Previous studies have shown that necrotic core formation enhances endoplasmic reticulum (ER) stress and delays organoid maturation.²⁰ Accordingly, the basal expression of ER stress genes, such as DDIP3, P4HB, and XBP1, was significantly reduced in nf-hSCOs compared to the large organoids (Fig. 1D). Furthermore, the expression levels of mature neuronal markers, NeuN and MAP2, were also significantly higher in small organoids (Supplementary Fig. S1C), indicating improved neuronal differentiation in the nf-hSCO.

Figure 1.

Establishment of nf-hSCOs. (A) Schematic diagram illustrating the procedure for generating nf-hSCOs. (B) Measurement of the diameter changes in organoids derived from ICN 75 and ICN 5000 during growth. (C) Representative images of 1-month-old hSCOs stained with c-Casp3 (cleaved Caspase 3), SOX2, and neurofilament-M (NF-M) based on the ICN. Nuclei were counter-stained with Hoechst33342 (blue, HOE). Scale bar: 100 µm. (D) Fold-changes in the mRNA expressions of DDIT3, P4HB, and XBP1 quantified by RT-PCR. The data were normalized to GAPDH expression and presented as Mean ± SEM; **P < .01, ***P < .001 by t test, n = 3.

Establishment of CoCl₂-Mediated Ischemia Model Using Nf-hSCO

Using nf-hSCO, we investigated whether HX and/or HG could induce neurodegeneration in organoids. To induce an HX-like condition, we treated the organoids with CoCl₂ as an alternative to oxygen deprivation due to its precise controllability of concentration and ease of the treatment.²¹ Treatment with a CoCl₂ (300 μM, referred to as HX condition) was sufficient to induce cell death under both normoglycemia (NG) and HG conditions on days 2 and 3 of treatments, while NG or HG alone did not affect the viability of the organoids (Fig. 2A, 2B, example of raw dataset is shown in Supplementary Fig. S3). Thus, it appeared that HX, rather than HG, is a primary cause of ischemia-induced cell death. Although HG is not necessary for the induction of cell death, all experiments conducted hereafter combined the conditions of HG and HX in order to faithfully replicate the ischemic condition which causes oxygen and glycose deprivation.²² Western blot analysis demonstrated a marked induction of HIF1α, the main transcriptional regulator of the cellular response to the HX pathway,²³ starting from day 1 and maintaining high levels until day 3 (Fig. 2C). Consistently, well-known downstream genes of the HIF signaling pathway (PDK1 and PFKP)⁶ and HG-related genes (EDEM1 and GADD34)²⁴ were significantly upregulated under the HXHG condition (Fig. 2D). In summary, these results indicate that the HXHG condition we established faithfully mimics neural ischemia in nf-hSCO.

Figure 2.

CoCl₂-mediated ischemia model using nf-hSCOs. (A) Representative images of nf-hSCOs stained with Propidium iodide/Acridine Orange. nf-hSCOs under NG and HG conditions were treated with 300 µM CoCl₂ (HXHG) for the indicated number of days. Scale bar: 100 µm. (B) Quantification of PI/AO ratio induced by CoCl₂. The PI/AO ratio was normalized to the NG group (**P < .01, ***P < .001, One-way ANOVA with Tukey test), n = 15 from 2 independent batches. (C) Western blot analysis of HIF1α protein expression in nf-hSCOs exposed to HXHG for the indicated days. (D) Quantification of HX-related genes (PDK1 and PFKP) and HG-related genes (EDEM1 and GADD34) by RT-PCR. Organoids were subjected to NG or HXHG conditions for 24 hours (PDK1, PFKP, EDEM1, and GADD34). The data were normalized to GAPDH expression and presented as Mean ± SEM; ***P < .001 by t test, n = 3.

Neuronal Death by HXHG Condition

To further evaluate the specific characteristics of neurodegeneration caused by the HXHG condition in the nf-hSCO model, we used immunohistochemical techniques to characterize the organoids (Fig. 3). As anticipated, there was a notable increase in caspase-3-positive cells throughout the organoids following HXHG treatment, indicating a widespread occurrence of apoptotic cell death due to HXHG (Fig. 3A; Supplementary Movie). Furthermore, we observed a reduction and fragmentation of NF-M-labeled neuronal fibers in HXHG-treated organoids, suggesting that the HXHG condition also impaired the growth of neurites (Fig. 3B; Supplementary Movie). Through double-immunofluorescence staining of caspase-3 with stem cell marker (SOX2) and neuronal marker (DCX), we determined that the dying cell populations primarily consisted of neurons (Fig. 3C). Consistently, RT-PCR analyses demonstrated a significant reduction in the expression of the neuronal marker gene (Tuj1) under the HXHG condition, while the expression of the stem cell marker gene (SOX2) remained unaltered (Fig. 3D).

Figure 3.

Neuronal degeneration by HXHG condition. (A) Immunohistochemical staining of c-Casp3 in organoids under NG or HXHG conditions on day 3. Nuclei were counter-stained with Hoechst33342. Scale bar: 100 µm. (B) Immunohistochemical staining of NF-M in organoids under NG or HXHG conditions on day 3. Right panels show the large magnification views of boxed area. Scale bar: 100 µm. (C) Double immunostaining for c-Casp3 with SOX2 or DCX. Scale bar: 10 µm. Arrows indicate marker-labeled cells in the center. (D) Quantification of Tuj1 and SOX2 expression by RT-PCR on day 2 after treatments. (Mean ± SEM; ***P < .001 by t test), n = 3. (E) Representative images of the outgrowing neurites of nf-hSCOs in NG and HXHG conditions for 1 day. Upper images were obtained under brightfield, and lower images show Tuj1 immunostaining in the same field. Scale bar: 100 µm. (F) Quantification of axonal area per organoid at 1 and 3 days after treatments. (Mean ± SEM; *P < .1, ***P < .001 by t test), n = 4.

Although the presence of axonal damages by HXHG was evident in nf-hSCO, as indicated by NF-M immunostaining (Fig. 3B), quantifying the extent of axonal damage proved challenging. To address this, we carefully adhered the organoids to the dish and allowed them to outgrow their neurites for 3 days. Upon treatment with HXHG, the axons exhibited rapid degeneration as early as 1 day, characterized by a reduction in axonal area and extensive fragmentation of neurites (Fig. 3E). These damaged neurites failed to regenerate and maintained a degenerative state by day 3, while normal organoids continued to extend their axons during this time period (Fig. 3F).

Screening of Chemicals Preventing Nf-hSCO Degeneration by HXHG

In order to assess the efficacy of our nf-hSCO model in identifying potential drug candidates, we selected 4 chemicals with putative beneficial effects: ISRIB, minocycline, rapamycin, and metformin. ISRIB and minocycline have been identified as neuroprotective chemicals in organoid-based ischemia models,^6,7 and rapamycin and metformin were included due to their ability to modify cellular metabolic stress through the mTOR (mammalian target of rapamycin) and AMPK signaling pathways, respectively.^25-28

Among the 4 chemicals tested, ISRIB and rapamycin exhibited beneficial effects against HXHG insults, whereas the other 2 chemicals showed no significant effects (Fig. 4A, 4B). RT-PCR analysis targeting the previously identified HX- and HG-related genes, PDK1, PFKP, EMEM1, and GADD34, revealed a significant reduction in HXHG-induced gene expression after treatment with ISRIB and rapamycin (Fig. 4C). However, based on neurofilament (NF-M) staining, neither rapamycin nor ISRIB demonstrated a pronounced protective effect against HXHG-induced neurite reduction and fragmentation (Fig. 4D). Consistent with this observation, PCR quantification showed only a mild rescue of Tuj1 gene expression in the rapamycin-treated group and not in the ISRIB-treated group (Fig. 4E). Furthermore, rapamycin, but not ISRIB, moderately increased the axonal area in the organoids outgrowth models (Fig. 4F, 4G). These findings suggest that rapamycin can protect both neuronal cell bodies and axons from HXHG-induced neurodegeneration, whereas ISRIB-treated group can rescue neuronal cell bodies but not axons from degeneration.

Figure 4.

Screening of chemicals preventing nf-hSCO degeneration by HXHG. (A) Representative images of nf-hSCOs stained with PI/AO after drug treatments. Scale bar: 100 µm. (B) Quantification of PI/AO ratio in the drug treatment groups. Data are normalized to the NG group (**P < .01, ***P < .001, One-way ANOVA with Tukey test), n = 16-21 from 2 independent batches. (C) Fold changes of PDK1, PFKP, EDEM1, and GADD34 mRNA expression quantified by RT-PCR after 3 days of drug treatment. Data are normalized by GAPDH expression (*P < .1, ***P < .001, One-way ANOVA with Tukey test), n = 3. (D) Representative images of nf-hSCOs stained with NF-M after 2 days of drug treatment. Scale bar: 10 µm. (E) Fold changes of Tuj1 mRNA expression quantified by RT-PCR on day 2 (***P < .001, One-way ANOVA with Tukey test), n = 3. (F) Representative images of the outgrowing neurites of drug-treated nf-hSCOs in NG and HXHG for 1 day. Left images were obtained under brightfield, and right images show Tuj1 immunostaining of similar fields. Scale bar: 100 µm. (G) Quantification of axonal area per organoid on day 1 after drug treatments (**P < .01, ***P < .001, One-way ANOVA with Tukey test), n = 4.

Discussion

In this study, we developed a new organoid-based screening system to identify drug candidates for neural ischemia. Neural ischemia is caused by a combination of HX and HG, and we optimized the treatment conditions in the organoids to mimic this ischemic environment. Importantly, we used small organoids that do not develop a necrotic core. The formation of a necrotic core is mechanistically similar to ischemia, as it occurs due to poor diffusion of oxygen and nutrients into the center of the organoids.¹⁰ By using small organoids without a necrotic core, we reduced the occurrence of spontaneous ischemia-like ER stress conditions and facilitated organoid maturation. Moreover, the significant reduction in the number of cells per organoid allowed for the production of a greater quantity of organoids from the same batch of initial hPSCs. Theoretically, this modification enabled a 70-fold increase in the number of organoids generated comparing to our previous organoid production protocol with ICN 5000,¹⁶ providing a numerical advantage for large-scale quantitative screening. CoCl₂ was used as an inducer of HX based on the previous reports demonstrating its ability to mimic HX.^21,29 CoCl₂ treatment eliminated the need for specialized equipment such as a HX chamber, making it more suitable for large-scale screening. We observed that CoCl₂ treatments induced the expression of HIF1α protein and downstream signaling cascades. Notably, we observed profound neuronal death under the HXHG condition, while neural stem cells exhibited greater resistance to ischemic insults. This finding aligns with previous studies demonstrating that neurons exposed to HX experience decreased synaptic activity and subsequent neuronal death,^30,31 whereas neural stem cells are able to survive but exhibit arrested differentiation.^32,33 Overall, we successfully established an experimental model for fetal neural ischemia using organoids, enabling us to advance our understanding of this condition and screen for potential therapeutic agents.

To investigate organoid degeneration, we used 4 different assays: PI/AO live/death assays, axon outgrowth assays, RT-PCR, and immunohistochemistry. Assay provided unique insights into organoid degenerations and had its own advantages and disadvantages. The PI/AO assay, although relatively less sensitive, allowed us to detect degeneration at 3 days after treatments. This assay is simple and facilitates easy quantification, making it suitable for initial screening and high-throughput assays compared to other methods. Axon outgrowth assays, on the other hand, were more sensitive, enabling the detection of axonal degeneration as early as 1 day after treatments. However, this method does not provide information on cell body death when axon degeneration and cell death are dissociated. RT-PCR and immunohistochemistry demonstrated high sensitivities, enabling the rapid detection of degeneration. However, these methods are more complex and time-consuming, which makes them less favorable for large-scale initial screening. Therefore, a combination of these methods is desirable for a comprehensive and complementary analysis of organoid responses to degeneration.

We identified 2 potential chemicals, ISRIB and rapamycin, that showed promise in preventing HXHG-induced degeneration in our screening assays. ISRIB, known for its association with the UPR pathway, which is activated during cellular stress, was previously found effective in human cortical organoids (hCOs) for fetal ischemia.³⁴ However, in our nf-hSCO model, minocycline, which has shown rescue effects in hCOs under hypoxic conditions,⁷ did not exhibit any benefits. The reason for these discrepancies is unclear, but it is possible that the response to anti-ischemic drugs may be region-specific in the brain. Different brain regions respond differently to ischemic insult,^35,36 with the hippocampus, cerebral cortex, and striatum being particularly vulnerable.^37,38 It is also worth noting that minocycline exerts its pharmacological action through its anti-inflammatory effects on microglia in vivo.³⁹ As in vitro organoids lack functional microglia, the absence of neuroprotective effects could be associated with the absence of microglia. Rapamycin, on the other hand, demonstrated effectiveness in our study and is commonly used to prevent autophagy and metabolic insufficiency.²⁸ Its neuroprotective effects have been confirmed in rat models of cerebral ischemia, where rapamycin reduced the damaged area and improved behavioral outcomes.^40,41 These findings indicate that targeting metabolic disturbance is a promising strategy for developing drugs to treat fetal neural ischemia. Furthermore, rapamycin exhibited stronger inhibition of axon degeneration compared to ISRIB, suggesting that metabolic disturbances play a significant role in axonal degeneration.

The current studies have several limitations that will be addressed in future research. Firstly, the use of immature organoids may not fully represent the developmental stages relevant to fetal and postnatal conditions seen in clinical settings. Studies have suggested that organoids require extended culture periods of more than 2-3 months to achieve developmental stages similar to 19-24 gestational weeks in vivo.^42,43 However, prolonged culture increases the risk of necrotic core formation due to the larger size of the organoids. Future investigations will aim to achieve a more accurate representation of developmental stages while minimizing the risk of necrotic core emergence. Secondly, the use of CoCl₂ as a chemical inducer of HX, instead of oxygen deprivation, was chosen for its favorable and scalable nature in large-scale drug screening. However, it is important to acknowledge that there may be significant differences between CoCl₂-induced HX and physiological oxygen deprivation. For example, low concentrations of CoCl₂ have been shown to enhance cell proliferation and differentiation, enabling cells to adapt to critical environmental conditions.^44,45 Further studies will explore these differences and evaluate their implications. Lastly, the focus of the current research was on preventive effects using a pretreatment regimen. Early diagnosis is not yet clinically relevant, and therefore, the emphasis was placed on assessing preventive effects in individuals with genetic risk factors. However, the organoid-based chemical screening procedures can be adapted to accommodate different requirements depending on the specific research objectives.

In summary, we successfully established a model for fetal spinal cord ischemia and developed a novel drug screening platform using nf-hSCO. This platform holds promise for identifying effective chemicals for the treatment of fetal neural ischemia. Employing this improved drug screening platform may enhance the likelihood of discovering potential therapeutic compounds and accelerate the drug screening process.

Funding

This research was supported and funded by Kun-Hee Lee Seoul National University Hospital Child Cancer & Rare Disease Project, Republic of Korea (23B-001-0500), the Brain Research Program through the National Research Foundation (NRF), which is funded by the Korean Ministry of Science (NRF-2021M3E5D9021368), and by NRF (RS-2023-00225239).

Conflict of Interest

The authors indicate no potential conflicts of interest.

Author Contributions

A.S.: conception, design of the work, acquisition, analysis, interpretation of data, writing manuscript. J.R.R.: conception, design of the work, interpretation of data. B.G.K.: resources. W.S.: conception, design of the work, interpretation of data, writing manuscript.

Data Availability

The data underlying this article will be shared on reasonable requests to the corresponding author.