All cells were cultured and maintained in incubators at 37℃ and 5% CO₂. Human iPSCs were cultured on vitronectin (A31804, Thermo Fisher Scientific, USA)-coated dishes and maintained in E8^TM-based hPSC medium (A1517001, Thermo Fisher Scientific, USA) supplemented with 1X penicillin/streptomycin. Culture medium was changed every day. Human iPSCs were detached from the dish by treatment with ReLeSR^TM passaging reagent (Stemcell^TM) and were dissociated into smaller pieces by tapping. The iPSCs used throughout the study, CMC-hiPSC-003 and CMC-hiPSC-011, were provided by the National Stem Cell Bank of Korea (Korea National Institute of Health), which were originally provided by Catholic University. Human dermal fibroblasts (HDFs) were maintained in FGM2 (CC-3132, Lonza, Switzerland) medium supplemented with 10% fetal bovine serum (FBS) and 1X penicillin/streptomycin. Culture medium was changed every other day. Human umbilical cord blood-derived mesenchymal stem cells (hUCB-MSCs) were obtained and approved for use by the Boramae Hospital Institutional Review Board (IRB) as previously described. hUCB-MSCs were maintained in EGM-2 endothelial growth medium supplemented with 10% FBS and 1X penicillin/streptomycin. HUVECs (ATCC, USA) were cultured in EGM-2 medium (CC-3162, Lonza, Switzerland) supplemented with 10% FBS and 1X penicillin/streptomycin.

To generate cerebral organoids from hiPSCs, we used methods from a previous study (7). Briefly, hiPSCs were dissociated and singularized using TrypLe (12604013, Thermo Fisher Scientific, USA) and resuspended at 12,000 cells/well in EB medium (E8 media, 4 ng/ml bFGF and 50 μM ROCKi) in ultralow attached plates (CLS7007, Corning, USA). Next, EB medium was replaced with neural induction medium (DMEM/F12, 20% KSR, 1% NEAA, 0.5%, 1% penicillin/streptomycin, 0.1 mM 2-mercaptoethanol, 5 μM dorsomorphin and 10 μM SB431542) for 5 days. On day 6, the media replaced with neural differentiation media (neurobasal A medium, 2% B27 without vitamin A, 1% GlutaMAX, 20 ng/ml EGF and FGF2). After neural induction, organoids were transferred to 96-well ultralow attachment plates and assembled with vascular spheroids in neural differentiation media. On day 25, the medium was replaced with generation medium (neurobasal A medium, 2% B27 without vitamin A, 1% GlutaMAX, 20 ng/ml BDNF and NT3). On day 44, the medium was changed, and cells were maintained in maturation medium (neurobasal A medium, 2% B27 without vitamin A, 1% GlutaMAX).

To induce the formation of vascular spheroids, each cell combination (HUVEC/HDF/hUCB-MSCs) and ratio (1：4：5) were used as described in a previous study (16). To form the aggregate, premixed solutions of each cell type were suspended into 96-well low attachment dishes in EGM-2 medium supplemented with 10% FBS and 1X penicillin/streptomycin. On day 10, vascular spheroids were used for fusion with brain organoids.

Assembloids or organoids were fixed in 4% paraformaldehyde overnight at 4℃, dehydrated in 30% sucrose solution and embedded in OCT compound (Tissue Tek, Germany). Cryosections were cut into 10 μm thick sections, washed with phosphate-buffered saline (PBS), deparaffinized, rehydrated and stained with hematoxylin/eosin (HE) and picrosirius red.

For immunohistochemistry, cryosections were cut into 20 μm thick sections and mounted onto silane-coated slides (Muto, Japan). Samples were washed 3 times with PBS, blocked and permeabilized for 1 h at room temperature (RT) in 5% normal goat serum and 0.5% Triton X-100 diluted in PBS. Primary antibodies were incubated overnight at 4℃ followed by incubation with appropriate secondary antibodies and 4’,6-diamidino-2-phenylindole (DAPI) for nuclear staining. Samples were mounted using mounting media and then dried at RT overnight. All samples were imaged using an Eclipse TE200 confocal microscope (Nikon, Japan).

For quantification of cell death in organoids, cryosectioned samples were immunostained for cleaved caspase-3 (#9664, Cell Signaling, MA) and subjected to a TUNEL assay (S7165, Merck, USA). For quantification of the percentage of apoptotic cells, randomly selected images of a 400×400 μm areas containing both the progenitor and neuronal layers were analyzed. The number of cells positive for cleaved caspase-3 was counted and divided by the total number of DAPI-positive cells.

Organoid proteins were extracted using PRO-PREPTM (iNtRON, Korea). Western blotting was performed on 12∼15% acrylamide gels, and proteins were transferred onto nitrocellulose membranes. Then, the membranes were incubated with 5% bovine serum albumin (GenDEPOT, USA) and probed with antibodies against γH2AX, cleaved caspase-3, GAPDH, β-catenin, Wnt7a, GSK3β, AXIN2, and GFAP. Membranes were then washed with TRIS-buffered saline with 0.1% Tween (TBS-T) and subsequently probed with the appropriate HRP-conjugated secondary antibodies (Invitrogen, CA, USA). The blots were exposed to ECL reagent, and immunoreactive bands were analyzed using ChemiDOC (FluorChem HD2, Proteinsimple, USA). Data were quantified using ImageJ 1.38 software (NIH, Bethesda, MD).

Total RNA was isolated from the spinal cord using TRIzol (INvitron, Waltham, USA) according to the manufacturer’s protocol. The extracted RNA was quantified by a spectrophotometer (ASP-2680, ACTgene, USA). To generate cDNA, reverse transcription was performed using SuperScript III First-Strand (Invitrogen, CA, USA) according to the manufacturer’s protocol. The resulting cDNA was used for real-time PCR using SYBR green PCR master mix (Thermo Fisher Scientific, MA, USA). Expression data of the duplicated results were used for 2-ΔΔCt statistical analysis, and GAPDH expression was used for normalization.

All statistical analyses were performed using GraphPad Prisms software (version 5). One-way ANOVA followed by Newman-Keuls post hoc test and t-test were used to investigate the significance of each data. All data are presented as the mean±standard error of the mean (SEM). Significance values are indicated as *p<0.05, **p<0.01, and ns (no significance).

A previous study investigated whether the 1：9 ratio of the cell combination resulted in stable spheroids formation during the entire culture period of 10 days (16). For the formation of HUVEC spheroids, a cell combination of HUVECs, hUCB-MSCs, and HdF and a cell ratio of 1/4：5/4：5 were used. To assess the aggregation of spheroids, premixed cells were seeded onto the microwells, and cells began to aggregate after one day in culture. Cellular debris was observed after one day in culture. However, these combinations exhibited stable spheroids throughout the entire culture period of 10 days. The spheroids were round, and less cellular debris was observed 10 days after coculture of the cells. The organization of HUVECs in the spheroids was assessed using RFP-expressing HUVECs, and we observed that RFP expression was well organized and maintained in the spheroids (Fig. 1A). To assess ECM production, picrosirius red staining was performed and showed a large positive area of collagen fibers in the spheroids, suggesting increased ECM production. Hematoxylin eosin staining revealed circular organized nuclei in the center (Fig. 1B). We next examined whether cocultured HUVECs formed vascularization in the spheroids. We observed that RFP-positive HUVECs were colocalized with CD31-positive cells and VE-cadherin. These vessel structures were enveloped by a basement membrane, which was identified using immunostaining for collagen type IV (Fig. 1C). These results indicate that HUVECs can stably form vessel-like structures in coculture spheroids.

We hypothesized that the assembly of HUVEC spheroids with cerebral organoids, called assembloid human cerebral organoids (ahCO), induces the formation of vessel-like structures in cerebral organoids. Thus, we assembled HUVEC spheroids with cerebral organoids (hCO) on day 13, which is the end point of neural induction, to avoid interrupting neural induction with the assembly of HUVEC spheroids. To investigate the effect of HUVEC spheroids during the generation of organoids, cerebral organoids assembled with or without HUVEC spheroids were observed for 43 days. The growth rate of each organoid was measured during this period. We found that HUVEC spheroids effectively assembled into hCO and that both groups of organoids, hCO and ahCO, were similar in parameter and area until day 43 (Fig. 2A). To investigate vascularization in ahCO, the organization of vessel-like structures was observed by staining for several endothelial markers. On day 43, the CD31 endothelial marker was expressed in ahCO but not hCO. Furthermore, Von willebrand factor (vWF) and Cadherin 5 (CDH5) were effectively expressed in the assembled spheroids through day 60, and sprouting vessels were marked by CD31 expression, whereas hCO lacked these endothelial markers. These results demonstrate that the assembled endothelial structures were stably formed in cerebral organoids until day 60 (Fig. 2B and C). Quantitative PCR results also supported the immunostaining results, exhibiting elevated expression of calponin1, CDH5, vWF and α-SMA in vascular organoids compared to control organoids on day 43 (Supplementary Fig. S1). Furthermore, whole-mount immunostaining demonstrated that CD31-positive vessel-like structures were formed and maintained until day 60. Quantification of total vessel length demonstrated that ahCO displayed significantly increased vessel length than control organoids (Fig. 2D and Supplementary Fig. S2).

Both control and vascular organoids exhibited well-organized SOX2-positive ventricular zones. NeuN and MAP2, which mark mature neurons, were detected outside of the SOX2 expression domain. Also, the expression of neural marker (TUJ1) and cortical layer marker (CTIP2) in both organoids were identified (Fig. 3A). Consistently, high population of positive cells with TUJ1, CTIP2 and MAP2 were expressed in both organoids at day 43. Moreover, the mature neuronal marker, NeuN, was further identified in vascularized organoids relative to the control organoids (Supplementary Fig. S3). In addition to the expression of neuronal markers, we identified the tight junction marker α-ZO1 in the lumens. On day 60, we identified that the lumens of the ventricular zone (VZ) in ahCO were stained with α-ZO1 compared to those of hCO,which were only weakly stained for α-ZO1 in the lumens (Fig. 3B).

To determine the effects of vascularization on DNA integrity and cell survival in organoids, we performed western blots for rH2AX and cleaved caspase-3, markers of DNA double strand breaks and apoptosis, respectively, and performed TUNEL assays. We observed significantly decreased apoptotic cell death (TUNEL positive) in vascular organoids compared to control organoids (Fig. 4A). We also observed an increase in the protein levels of rH2AX and cleaved caspase-3 in cerebral organoids on day 60, and the protein levels of both were decreased in vascularized organoids (Fig. 4B). Furthermore, immunostaining of rH2AX and cleaved caspase-3 revealed significantly fewer positive cells in vascularized organoids than in control organoids. To determine whether DNA damage and apoptosis occurred significantly in differentiating or precursor positions, we quantified the population of rH2AX and cleaved caspase-3 cells in the ventricular zone and outer layers of both organoids after 60 days. We identified a significant increase in the population of rH2AX- and cleaved caspase-3-positive cells in control organoids compared to vascularized organoids, similar to the western blot data. Since a majority of SOX2＋ cells resided in the ventricular zone, we further quantified the number of SOX2＋ cells with rH2AX and cleaved caspase3 in the VZ, which were significantly decreased in vascular organoids compared to control organoids (Fig. 4C and D).

Wnt/β-catenin signaling is important for the formation of blood vessels, especially CNS-specific angiogenesis. Wnts may specifically regulate CNS vessel formation and function. Specifically, Wnt7a is expressed in the forebrain and ventral neural tube (15, 17). Effective activation of Wnt/β-catenin signaling activity was confirmed by western blots. We found that Wnt7a and β-catenin protein levels were elevated in vascular organoids compared to control organoids. Additionally, binding of Wnt ligands to Frizzled/LRP receptor complexes causes stabilization of β-catenin. Therefore, we assessed GSK3β and axin2 protein levels and identified activation axin2 along with activation of Wnt7a and β-catenin. Although no significant difference was confirmed in the case of GSK3β, we identified up-regulation of Axin2 (Fig. 5A). Endothelial cells signal to neural stem cells, regulating their self-renewal and differentiation into neurons. In particular, the Wnt/β-catenin pathway also regulates the proliferation of neural precursor cells during CNS development. The proliferation marker Ki67 was increased in the vascular organoids compared to the control organoids (Supplementary Fig. S4). Furthermore, Wnt7a plays critical roles in neurogenesis by activating β-catenin signaling and the specific transcription factors correlated with neural differenti-ation. To determine the difference between vascular and control organoids, we immunostained each organoid with mature neuronal markers and astrocytes. In vascular organoids, NeuN- and MAP2-positive cells, which are mature neuronal markers, were identified, and GFAP-positive cells, which are astrocytic markers, were detected more often in the vascular organoids than in the control organoids (Fig. 5B and C).