Abstract

   Human stem cells and derivatives transplantation are widely used to
   treat nervous system diseases, while the fate determination of
   transplanted cells is not well elucidated. To explore cell fate changes
   of human brain organoids before and after transplantation, human brain
   organoids are transplanted into prefrontal cortex (PFC) and hippocampus
   (HIP), respectively. Single‐cell sequencing is then performed.
   According to time‐series sample comparison, transplanted cells mainly
   undergo neural development at 2 months post‐transplantation (MPT) and
   then glial development at 4MPT, respectively. A different brain region
   sample comparison shows that organoids grafted to PFC have obtained
   cell fate close to those of host cells in PFC, other than HIP, which
   may be regulated by the abundant expression of dopamine (DA) and
   acetylcholine (Ach) in PFC. Meanwhile, morphological complexity of
   human astrocyte grafts is greater in PFC than in HIP. DA and Ach both
   activate the calcium activity and increase morphological complexity of
   astrocytes in vitro. This study demonstrates that human brain organoids
   receive host niche factor regulation after transplantation, resulting
   in the alignment of grafted cell fate with implanted brain regions,
   which may contribute to a better understanding of cell transplantation
   and regenerative medicine.

   Keywords: astrocyte, human brain organoids, neurotransmitter,
   single‐cell RNA sequencing, transplantation therapy
     __________________________________________________________________

   Grafted human brain organoids in the host brain keep forebrain cell
   fate following intrinsic expression patterns similar to fetal brain
   tissue. Grafted organoids undergo neural and astrocytic differentiation
   successively. Young human cells acquire properties similar to the brain
   regions after grafting. Neurotransmitters enhance the activity and
   morphological complexity of astrocytes in organoids.

   graphic file with name ADVS-11-2402287-g006.jpg

1. Introduction

   The adult central nervous system (CNS) has limited self‐repair capacity
   after pathogenesis or tissue impair, while young cells, which are
   equiv. to neurons in embryonic stage, or tissue replacement may be an
   effective approach to repair brain. Neural progenitor cells derived
   from human pluripotent stem cells could be considered as a young cell
   pool for cell repairing in CNS. Recently, several transplantation
   therapies using human stem cells and derivatives have been applied in
   the clinic, providing hope for neural repairment.^[ [48]^1 , [49]^2 ^]

   Recently, brain organoids have been demonstrated to be promising cell
   donors for repairing severe neural injury at tissue level.^[ [50]^3 ,
   [51]^4 ^] And transplanted human organoids have been proven to
   integrate into the recipient's neural circuit and coordinate with their
   behaviors. For example, the host vessels were observed to progressively
   infiltrate into human organoids after the transplantation of organoids
   into unilaterally injured mouse brains.^[ [52]^5 ^] Neurons in the
   implanted organoids furthermore presented more complex morphology,
   synapses, and membrane properties than that in vitro.^[ [53]^6 ^] Our
   previous work also demonstrated that grafted human brain organoids
   present long projections to the deep brain region and establish
   functional connections with the host neurons.^[ [54]^7 ^]

   Before translational cell‐based therapy, fate determination of the
   grafted organoids, as well as safety analysis, need to be
   comprehensively evaluated. In vitro, time‐series studies for the cell
   fate determination of human brain organoids have been described;^[
   [55]^8 ^] In vivo, numerous studies, including ours, have been devoted
   to study the survival, neural circuit integration, and functional
   recovery after transplantation of human pluripotent stem cells
   (hPSCs)‐derived neurons differentiation, but how the cell fate
   determination of the grafted hPSC‐derived neuron or organoid is still
   unknown. Therefore, it is essential to perform dual cell fate decoding
   of time and space for transplanted neural cells, and the interactive
   regulation of their own genes and the host micro‐environment (or niche)
   should be fully explored.

   Here, we transplanted human brain organoids into PFC or HIP of mice and
   obtained more than 30 000 high‐quality human‐derived cells.
   Transplanted cells mainly underwent neural development at 2 months
   post‐transplantation (MPT) and then glial proliferation at 4MPT,
   respectively. In different brain‐region comparisons, we found human
   cells acquired the transcriptomic properties of grafted host brain
   region. Furthermore, astrocytes transplanted in PFC were more mature
   than those in HIP, which was enhanced by receiving rich
   neurotransmitters, dopamine (DA) and acetylcholine (Ach). Our results
   may pave the road for applications of human brain organoids into
   transplantation therapy.

2. Results

2.1. Single‐Cell Transcriptome of Grafted Human Brain Organoids Recapitulates
Developing Fetal Forebrain Features

   To explore cell fate of human brain organoids after transplantation,
   the cell transcriptomic profiling of human organoids was determined by
   single‐cell RNA sequencing, with the comparison of brain organoids
   before transplantation. Human brain organoids derived from H9:
   AAVS‐ChR2‐EYFP^[ [56]^9 ^] labeled with EYFP fluorescent protein were
   implanted into the prefrontal cortex (PFC) and hippocampus (HIP) of a
   mouse (SCID‐Beige) brain, respectively. Then single‐cell sequencing was
   performed for grafted brain organoids (GBOs) in the mouse brain at 2
   months post transplantation (MPT) or 4MPT (Figure [57]1A).

Figure 1.

   Figure 1
   [58]Open in a new tab

   Single‐cell Transcriptome of Grafted Human Brain Organoids
   Recapitulates Developing Fetal Forebrain Features. A) Schematic diagram
   showing the processes of human brain organoid transplantation,
   sampling, digestion, sequencing, and analysis. B) Representative
   immunofluorescence results after human brain organoid transplantation.
   HIP_2MPT, HIP_4MPT represents the organoids at 2 and 4 months after
   transplantation to the mouse HIP, respectively; PFC_2MPT and PFC_4MPT:
   the organoids at 2 and 4 months after transplantation to the mouse PFC,
   respectively. Exogenous cell marker: EYFP; human‐derived nuclear
   marker: hN. Scale bar: 100 µm. C) UMAP visualization of the
   distribution of human cells in organoids before and after
   transplantation, with colored dots marking cell types. hBOs: human
   brain organoids; GBOs: grafted human brain organoids (n = 34388 cells
   from 20 organoids in vitro and 16 grafted mice in vivo). D) UMAP plot
   between different samples. D0: organoids acquired on the day of
   transplantation. E) Violin plots revealing differences in numbers of
   expressed genes and transcripts among types of cells. F) Dot plot
   showing the expression of characteristic genes of different cell types.
   The diameter of the dot represents the proportion of expressing cells;
   the colors show the normalized gene expression. G) The expression of
   characteristic genes (same as F) of the four major cell types in hBOs
   and GBOs. H) Heat map presenting the correlation between cells in our
   data and human brain cells in the scHCL database. Using R package
   scHCL. I) Mapping the scRNA‐seq data onto the human brain or mouse
   brain public database. J) Bar chart: the length of a bar shows the
   number of cells. Columns represent cell types; colors represent
   samples. K) The correlation between each sample and E13.5 mouse brain
   regions from Allen Brain Institute Database. L) VoxHunt showing the
   most similar brain region (E13.5 mouse brain sections) to which
   organoids were mapped. Left: coronal section; right: sagittal section.
   Reference data is from Allen Brain Institute Database. M) Heatmap
   presenting the correlation between each sample and human fetal brains
   from the BrainSpan database at different stages of development. N) The
   most similar region of a human organoid compared with human embryonic
   brain section. DF; ventral striatum (GE/STR); thalamus (DTH);
   cerebellum (URL/CB). O) Heatmap showing gene sets specifically
   expressed by each cell lineage and the corresponding enriched GO terms.

   Following a previously established protocol to differentiate human
   brain organoids in vitro,^[ [59]^7 ^] D42 human brain organoids were
   generated (Figure [60]S1A,B, Supporting Information). Then human brain
   organoids were bilaterally transplanted into PFC or HIP of mice,
   respectively. Grafted human cells were distinguished by EYFP or hN
   (Figure [61]1B; Figure [62]S1C, Supporting Information). Several blood
   vessels were found to integrate into grafted organoids within 1MPT
   (Figure [63]S1E, Supporting Information), and human synaptic labeling
   hSYN was determined on human fibers near the implants (Figure [64]S1F,
   Supporting Information), which suggested that grafted organoids
   survived and formed synaptic connections to host brain. In addition,
   nearly no pluripotent cell marker NANOG was detected in grafted
   organoids (Figures [65]S1D and [66]S2K, Supporting Information).
   Furthermore, no teratoma or malignant proliferation was found in
   transplanted animals (n > 70 mice). These results indicated that
   organoids transplantation has a low risk for tumor genesis in our
   system.

   Single‐cell suspensions were prepared from five groups, including D0
   (day 42 brain organoids that pre‐transplantation), PFC_2MPT, PFC_4MPT,
   HIP_2MPT, and HIP_4MPT. A total of 34388 human cells were obtained from
   16 grafted host brains and ≈20 human brain organoids after quality
   control (Figure [67]S2F–I, Supporting Information). On average, 2262
   transcripts and 1232 genes were generated for each cell.

   Given that the mixing of murine cells during acquisition of grafted
   organoids was inevitable, host cells were excluded to purify the
   grafted human datasheet during analysis. First, after reads were mapped
   to human‐mouse dual reference genome, ratios of human transcripts were
   calculated for each cell (Figure [68]S2A, Supporting Information). The
   results showed two aggregation peaks (low or high) which suggested
   clusters of murine or human cells, respectively. Then all the cells
   were divided into different species identities by two ratios (0.2 and
   0.8), and the cells in between were tentatively labeled as Unknown
   Cells. It is evident that three cell groups showed their own
   distribution in scatter plot, which may suggest the reliability of
   labeling (Figure [69]S2B, Supporting Information).

   The Uniform Manifold Approximation and Projection (UMAP) analysis was
   performed to investigate the diversity of human cells using Seurat. All
   the cells were unbiasedly isolated into four major cell types by global
   transcriptomes, such as, progenitor_lineage (PLC), neuron_lineage
   (NLC), astrocyte_lineage (ALC), and oligodendro_lineage (OLC)
   (Figure [70]1C,D). In total, four major cell types were identified in
   human cells according to the expression of specific genes, including
   PLC (MI67, HMGB2), NLC (STMN2, NEUROD6), ALC (GFAP, HOPX), and OLC
   (PDFGRA, MBP) (Figure [71]1F,G). Meanwhile, both GO enrichment analysis
   and comparison of our data and human brain tissue transcriptome data
   from SCHCL^[ [72]^10 ^] confirmed the validity of classification
   (Figure [73]1H,O). The proportions of the four types of cells varied
   among the five groups. Specifically, group D0 showed the highest
   proportion in progenitor cells (33.4%). The population of neurons was
   notably increased after transplantation (D0: 9.5% vs After‐graft:
   average 22.6%), which was similar to astrocytes (D0: 4.1% vs
   After‐graft: average 24.0%). Oligodendrocytes appeared only in 4MPT
   groups (Figure [74]1J), indicating that organoids matured after
   grafting. Although human‐derived oligodendrocytes were a small
   population, they were identified by co‐labeling human cytoplasmic
   marker STEM121 with oligodendrocyte marker MBP in brain slices (Figure
   [75]S1G, Supporting Information). Furthermore, human cells were
   unbiasedly subclassified and more cell subtypes were found, such as
   cycling progenitors, radial glia, and intermediate progenitors et al
   (Figure [76]S2L,M, Supporting Information). Then their identities were
   also confirmed by characteristic genes (Figure [77]S2M–Q, Supporting
   Information).

   To assess the brain regional identity of grafted cells, all samples
   were mapped onto embryonic day 13.5 (E13.5) mouse brain data from Allen
   Brain Institute.^[ [78]^11 ^] Both organoids and grafted organoids were
   found to mostly map onto dorsal telencephalon (Figure [79]1K). And
   VoxHunt algorithm supported the above conclusion in both coronal and
   sagittal section view (Figure [80]1L). In addition, human organoids
   were compared with human fetal brain data from BrainSpan, and found all
   of them showed a strong correlation with dorsal forebrain (DF)
   (Figure [81]1N). According to the developmental period, all samples
   could be separated into three stages: group D0 was close to 8–9 w human
   fetal brain, group 2MPT was similar to 12–16 w of gestation, and group
   4MPT was close to 19–26 w of gestation (Figure [82]1M). These results
   suggested that grafted human brain organoids undergo a progressive
   development and maturation.

   Taken together, the single cell transcriptome of human brain organoids
   was acquired after transplantation. All the human cells were divided
   into four major cell types and grafted organoids maintained their
   forebrain cell fate after transplantation.

2.2. Human Organoids Undergo Neural Differentiation and Astrocytic
Differentiation Successively after Transplantation

   To analyze the transcriptomic features among five samples, the single
   cell pseudo‐bulk mRNA expression of each sample was first obtained.
   Group D0 was found to be isolated from the other four groups in
   Principal Component Analysis (PCA), which indicated a substantial
   difference in transcriptome between pregraft organoids and postgraft.
   Furthermore, group 2MPT (HIP_2MPT and PFC_2MPT) was separated from
   group 4MPT (HIP_4MPT and PFC_4MPT), which showed that the diversity of
   samples in transcriptome after transplantation were mainly due to the
   experienced time of organoids. In addition, it leads to the same
   conclusion when the transcriptome‐based distance between groups and
   then clustered samples was calculated unbiasedly (Figure [83]2B).

Figure 2.

   Figure 2
   [84]Open in a new tab

   Human Organoids Undergo Neuronal and Glial Differentiation in the
   Second and Fourth Months Individually after Transplantation. A)
   Illustration showing the difference analysis of the single cell sample
   of each group. B) (left) Principal component analysis showing the
   difference in pseudo‐bulk RNA expression among the five groups of
   samples. Color: different samples, shape: different groups, and curve:
   the simulated developmental curve of groups. (right) Heat map
   describing the similarity between samples. Samples are clustered
   unbiasedly. the heatmap color represents the distance of transcriptome
   expression profiles among samples. C,E) Distribution scatterplot of
   DEGs. Dots represent the log[2]‐transformed fold change of gene
   expression in different DEGs. All red dots include genes in the first
   quadrant (in C) or the second quadrant (in E), respectively; all blue
   dots include genes in the third quadrant (in C) or the fourth quadrant
   (in E). D,F) Top terms after the enrichment analysis of GO biological
   process. Red or blue dots correspond to the gene sets in red or blue
   respectively (in C and E). The diameter of the dot represents the
   number of enriched genes. G,J) Single cell trajectories of different
   cell lineages (G: neurons, J: astrocytes). Pseudo‐time is indicated by
   the color of dot. H,K) Pseudo‐time trajectories recovered by Monocle.
   Colors represent cell subtypes. The distribution of cells along the
   trajectories reveals the order of cell development. During neuron
   development, CPCs, RGs, IPNs, and GNs appeared in the trajectory
   successively, and during astrocyte development, RGs were developed into
   Astrocytes. I,L) Pseudo‐time distribution of human cells. Colors
   represent sample groups. Black dashed lines represent the pseudo‐time
   averages of different groups.

   The differentially expressed genes (DEG) between groups were acquired
   for further analysis. There was a strong linear correlation between the
   log[2](Fold change) of DEG of group 2MPT and group D0 and those of
   group 4MPT and group 2MPT (Figure [85]2C), which proved their DEG sets
   were similar. The coupregulated gene set (red spots in 2C) was selected
   for gene ontology (GO) biological process enrichment analysis, and
   nearly all the top enriched terms were related to neural development
   and maturation, such as axonogenesis, neuron differentiation, forebrain
   development, synapse organization, and neuron development. While
   codownregulated gene set (blue spots in 2C) enriched for metabolism and
   progenitor development, for example, translational initiation,
   ribonucleoprotein complex biogenesis, mRNA metabolic process, and
   nuclear‐transcribed mRNA catabolism (Figure [86]2D). These results
   suggested that human organoids were still in the early stage of
   development before transplantation and then underwent neural
   development, differentiation, and maturation after transplantation. In
   order to explore the reasons for the enormous transcriptomic
   differences between organoids before and after grafting, gene sets
   specially belong to group 2MPT and group D0 were further investigated,
   individually. The gene set 2MPT was enriched for neuron migration,
   neural development, and dendrite development, while gene set D0 was
   associated with hypoxia, cell stress, cellular respiration, and nuclear
   division (Figure [87]S3C,D, Supporting Information). Stress‐related
   genes (HIF1A, VEGFA, ALKBH5: response to hypoxia;^[ [88]^12 , [89]^13 ,
   [90]^14 ^] ARCN1, GORASP2, ATF6B, HSPA5: endoplasmic reticulum stress
   response;^[ [91]^15 , [92]^16 ^] PGK1: glycolysis gene^[ [93]^16 ^])
   were high expressed in group D0 (Figure [94]S3E, Supporting
   Information) but decreased in group 2MPT and 4MPT, which is also
   consistent with the previous research.^[ [95]^16 ^] To eliminate
   temporal interference, a sample of human brain organoids cultured in
   vitro for ≈100 days was also compared with our grafted samples at 2MPT.
   The results reaffirmed that transplantation primarily led to the
   maturation of neurons and alleviation of stress response (Figure
   [96]S3L–O, Supporting Information).

   The log[2](Fold change) of DEG between 4MPT and 2MPT was similarly
   compared with those between 2MPT and D0 (Figure [97]2E), and the genes
   (red spots) upregulated in 4MPT and downregulated in 2MPT were focused
   on. GO enrichment analysis showed these genes were predominantly
   associated with positive regulation of gliogenesis and negative
   regulation of neurogenesis (Figure [98]2F). In contrast, all the genes
   (blue spots) upregulated in 2MPT and downregulated in 4MPT were related
   to neural development and maturation (10/top 10, data not shown).
   Therefore, human organoids mainly went through neuron development in
   2MPT and then glial genesis in 4MPT. In addition, genes upregulated
   both in 2MPT and 4MPT were enriched for neuron development and
   maturation (7/top 10) and genes downregulated were related to
   progenitor cell proliferation and development (Figure [99]S3A,B,
   Supporting Information), which indicated human organoids mainly
   underwent neural development and maturation after transplantation.

   To determine the pseudo‐time distribution of each group, single cell
   trajectories for all the human cells were reconstructed and three main
   developmental trajectory branches were found, for instance, cycling
   progenitor cells developed to radial glial cells and then to
   intermediate progenitor neurons which finally developed to
   glutamatergic neurons, radial glial cells developed to astrocytes, and
   oligodendrocyte progenitor cells developed to oligodendrocytes (Figure
   [100]S3J,K, Supporting Information). To further assess the cell lineage
   differentiation, single‐cell branching trajectory analysis was
   performed for two main cell types, neurons and astrocytes. Cycling
   progenitor cells, radial glial cells, intermediate progenitor neurons,
   and glutamatergic neurons appeared in the neural trajectory
   successively, while radial glial cells and astrocytes appeared in
   astrocyte trajectory (Figure [101]2G–H,J,K). The results showed grafted
   neurons developed over time during the first two months and showed
   limited change in 4MPT (pseudo‐time averages: 2MPT:13.40 vs
   4MPT:13.66). By contrast, astrocytes were more mature in 4MPT than in
   2MPT (pseudo‐time averages: 2MPT:6.99 vs 4MPT:8.68) (Figure [102]2I,L).
   These results suggested that the grafted organoids developed into
   neurons at the beginning and then developed into astrocytes later,
   which demonstrated the transplanted organoids matured from 2MPT to 4MPT
   (Figure [103]2E,F).

   To confirm the cell fate changes in human cells after transplantation,
   immunostaining analysis was carried out. First of all, fluorescent
   protein EYFP and human nuclear marker hN were used to identify human
   cells in the host brain. Cell proliferation marker Ki67 was detected in
   grafted organoids in 2MPT, which showed there was a progenitor cell
   pool to provide lineage expansion after grafting. Precursor cells
   (SOX2^+) took a considerable part (≈20%) in grafted organoids. In
   addition, grafted human cells mainly recapitulated dorsal telencephalon
   identity, which was confirmed by the observation of PAX6. The
   expression of doublecortin (DCX) showed the generation of a great many
   newborn neurons (Figure [104]3B,C). ≈20% of mature human neurons
   (NEUN^+ or MAP2^+) were observed in grafted organoids. Those neurons
   expressed various cortical layer markers, which suggested the cell fate
   of the cortex in grafted organoids. Furthermore, human astrocytes took
   a significant proliferation (4.5% ± 1.7% vs 9.5% ± 0.8%) from 2MPT to
   4MPT (Figure [105]S3F,G, Supporting Information), which was consistent
   with transcriptome analysis.

Figure 3.

   Figure 3
   [106]Open in a new tab

   Immunofluorescence Results Reveal that Human Brain Organoids Mainly
   Undergo Astrocyte Differentiation after Transplantation. A) The immune
   fluorescent comparison of the human brain organs after transplantation
   into mouse HIP and PFC. B) Immunostaining results of molecular
   signatures of organoids in 2MPT. HO: labeling nucleus; hN: human
   nucleus‐associated protein; SOX2: labeling neural precursor cells;
   EYFP: enhanced yellow fluorescent protein; NEUN: expressed in mature
   neuron nuclei; TBR1 and CTIP2: cortical neuron layer markers; MAP2:
   mature neuron marker; TUJ‐1: stained at mature neuronal axon; DCX:
   newborn neuron marker; PAX6: stained at DF precursor cells; GFAP:
   labeling astrocytes. C) Staining results of mouse brain sections 4
   months after organoid transplantation. Scale bar: 20 µm.

   In summary, through the transcriptome analysis of each sample, human
   organoids were found to undergo neural lineage development from DO to
   2MPT and experience glial lineage proliferation in 4MPT, which was
   consistent with physiological development process of human fetal brain.

2.3. Grafted Human Cells Acquire Properties Similar to the Implanted Host
Brain Regions

   In order to further investigate the transcriptomic diversity of human
   cells transplanted in PFC and HIP, these four DEGs were compared with
   each other (Figure [107]4B). As the gene number of DEG was generally
   used to assess the diversity of samples, the number of DEG in different
   brain region samples (HIP2MPT vs PFC2MPT or HIP4MPT vs PFC4MPT) was
   almost equiv. to that in time‐series samples (4MPT vs 2MPT), which
   suggested that the diversity of human cells between brain regions was
   also necessary to be further analyzed. To characterize DEGs in
   different brain‐region samples, multiple single‐cell transcriptome
   datasets of HIP and PFC grafts from adult mice were collected for
   pseudo‐bulk PCA analysis.^[ [108]^17 , [109]^18 , [110]^19 , [111]^20
   ^] HIP or PFC samples from adult mice were aggregated respectively,
   which demonstrated the different cell characteristics at transcriptomic
   levels (Figure [112]4C). To simplify the subsequent analysis, two
   samples (HIP_p28 and PFC_p29 in Figure [113]4C) from the Mousebrain
   Atlas were used and their representativeness of the total samples was
   demonstrated (Figure [114]S4A, Supporting Information). In addition,
   the identities of these two groups were confirmed by observing the
   expression of brain region specific genes, such as Satb2, Tbr1, Prox1,
   and Zbtb20 (Figure [115]4D).^[ [116]^21 , [117]^22 , [118]^23 ,
   [119]^24 , [120]^25 ^]

Figure 4.

   Figure 4
   [121]Open in a new tab

   Differences in Human Organoids Grafted into Distinct Brain Regions
   Resemble Differences in Mouse Brains. A) Schematic diagram showing the
   difference analysis of the single‐cell samples between PFC and HIP in
   mice. B) Numbers of DEGs among four groups (P.adjust < 0.05; Percent
   (PCT)> 0.1; Log[2]FoldChange> 0.25). C) PCA diagram showing the
   aggregation of pseudo‐bulk transcriptomes between different groups (2
   conditions: mouse HIP and mouse PFC) of samples. Left panel: total
   cells; middle panel: neurons; right panel: astrocytes. p23 represents
   postnatal day 23. D) Bar chart revealing the expression of HIP or PFC
   related genes (from public date) in the two representative samples from
   the Mousebrain Altas database. Log2FC: Log[2](fold change); PCT:
   percent. E) Single cell pseudo‐time trajectories of three groups of
   cells (PFC cells, HIP cells, and Progenitor cells). F) Distribution of
   pseudo‐time showing that the overall mouse hippocampal cells are less
   mature than the prefrontal cells. RGs: Radial glial cells; SVZ:
   Subventricular zone cells. G) Pseudotime analyses for normalized
   expression of marker genes (RGs, Neurons, and Astrocytes) along the
   pseudo‐time. H,J) Volcano plot showing the DEGs (HIP vs PFC) in mouse
   neurons or astrocytes, respectively. Red dots show genes upregulated in
   HIP; blue dots show genes upregulated in PFC; p‐adj: adjusted P‐values.
   I,K) Representative enriched GO terms (red bar: HIP; blue bar: PFC) of
   DEGs of mouse neurons and astrocytes, respectively. L) GOCircle plot
   representing the expression of enriched genes (enriched for GO terms in
   I or K) of human neurons or astrocytes that grafted into HIP and PFC.
   The enriched gene sets expressed in humans were consistent with those
   in mice. M) Density plot of DEGs of neurons showing that DEGs in our
   and public dataset were predominantly distributed in the first
   quadrant; x: grafted human organoids in 2MPT (PFC vs HIP); y: mouse
   brain samples(Mousebrain_cortex_p28 vs Mousebrain_hip_p28, in C).

   To determine the distinction between mouse HIP and PFC, neural
   precursor cells from mouse embryonic telencephalon were analyzed,^[
   [122]^20 ^] and the PFC matured more during development than the mouse
   HIP with the comparison of pseudo‐time trajectories (Figure [123]4E,F),
   including neurons and astrocytes (Figure [124]S4D, Supporting
   Information). Additionally, HIP_p28 was also compared with PFC_p23, and
   PFC_p23 was found to be more mature than HIP_p28 (Figure [125]S4B,C,
   Supporting Information). Then the DEGs of adult mouse neurons and
   astrocytes were identified individually (Figure [126]4H,J), and GO
   biological process enrichment analysis was performed. GO terms of
   astrocytes were pointed to three categories of genes: those were
   crucial for astrocytic development (e.g., HIP: translational
   initiation; PFC: regulation of fiber organization), astrocyte function
   (e.g., HIP: carboxylic acid biosynthetic process and membrane
   biogenesis; PFC: cellular respiration and iron ion transport),
   interaction of different types of cells (e.g., HIP: neural precursor
   cell proliferation; PFC: modulation of chemical synaptic transmission)
   (Figure [127]4K). In addition, a similar analysis was performed in
   neurons, and GO terms were also enriched for three clear categories of
   genes: “neural development”, “neural function”, and “interaction”
   (Figure [128]4I). Furthermore, the expression of genes, which presented
   in the enriched GO terms (all the three categories of GO terms in
   Figure [129]4I,K) of the transplanted brain regions, were observed in
   grafted human cells (neuron or astrocyte), and the transplanted human
   cells were found to obtain similar transcriptome properties to host
   brain regions, where the GBOs located (sum of log[2](fold change) >0 or
   <0) (Figure [130]4L), as the genes differentially expressed between
   grafted human organoids were also largely differentially expressed
   between PFC and HIP of mouse brain(Figure [131]4M).

   Taken together, transcriptomic divergence in mouse brain regions was
   identified, and transplanted human‐derived cells were found to acquire
   properties similar to those of adjacent host brain tissue.

2.4. Neurotransmitters Enhance Human Astrocyte Maturation in PFC

   Human astrocytes have been implicated various brain disease
   pathologies.^[ [132]^26 , [133]^27 ^] To determine whether the
   characteristics of grafted human astrocytes might be affected in
   different host brain regions, further enrichment analysis was
   performed. Since the main developmental stage of grafted human
   astrocytes was 4MPT in our analysis, the DEG among astrocytes between
   HIP_4MPT and PFC_4MPT (Figure [134]5B) was acquired. The gene set which
   was highly expressed in HIP_4MPT was enriched for glial cell fate
   commitment and neuron maturation (Figure [135]5C), while the gene set
   which was highly expressed in PFC_4MPT was enriched for terms related
   to astrocyte function (Figure [136]5D), such as ATP metabolic process,
   response to cadmium ion, cellar response to cadmium ion and regulation
   of growth. These results suggested that grafted human astrocytes in PFC
   might be more functionally mature, which was further confirmed by the
   comparison of pseudo‐time distribution between PFC_4MPT and HIP_4MPT
   (Figure [137]5E). Gene set enrichment analysis (GSEA) was also
   performed, and astrocytes of PFC_4MPT was enriched for Gap
   junction‐related signaling pathways, which was closely associated with
   astrocyte function (Figure [138]5F–H). In addition, neurons in PFC_2MPT
   were also more mature than those in HIP_2MPT (Figure [139]5E).

Figure 5.

   Figure 5
   [140]Open in a new tab

   Human Astrocytes are More Mature in PFC_4MPT than in HIP_4MPT due to
   Neurotransmitter Signaling. A) Differences analysis of grafted human
   astrocytes between PFC_4MPT and HIP_4MPT. B) Volcano plot showing the
   DEGs in astrocytes between PFC_4MPT and HIP_4MPT. Red dots represent
   genes upregulated in HIP_4MPT; blue dots show genes upregulated in
   PFC_4MPT; the gray dots represent genes that were not significantly
   differentially expressed between the groups. C,D) The top enriched
   terms in GO enrichment analysis of upregulated genes in HIP_4MPT (red
   dots in B) or upregulated genes in PFC_4MPT (blue dots in B)
   respectively. Color of dots: adjusted P‐values; diameter of dots:
   number of enriched genes. E) Pseudo‐time distribution of human neurons
   in 2MPT and human astrocytes in 4MPT. Colors show samples; dashed lines
   show the averages of pseudo‐time distribution. F) GSEA results showing
   the enriched terms (P.adjust < 0.05, comparison of PFC_4MPT and
   HIP_4MPT). Colors represent adjusted P‐values. G) Differences in the
   gap junction pathway between PFC_4MPT and HIP_4MPT, indicating a
   decrease in the HIP_4MPT group. H) Violin plot showing the expression
   levels of representative genes enriched in the gap junction pathway. I)
   Human astrocyte enriched top KEGG pathways in PFC_4MPT. Colors
   represent the ‐log[10](adjusted P). J,K) Expression levels of genes
   related to dopaminergic (DA) or acetylcholinergic (ACh) synapse pathway
   between PFC_4MPT and HIP_4MPT, respectively.

   To reveal why transplanted human astrocytes in PFC showed more mature
   characteristics in transcriptomic level, DEG between PFC_4MPT and
   HIP_4MPT was obtained and then kyoto encyclopedia of genes and genomes
   (KEGG) pathway enrichment analysis was performed. The results showed
   two neurotransmitter synapses (dopaminergic and cholinergic) and gap
   junctions were in the top enriched KEGG pathways (Figure [141]5I).
   Moreover, the high expression of genes in PFC_4MPT was related to
   dopaminergic and cholinergic synapses (Figure [142]5J,K), which
   suggested two neurotransmitters may play important roles in regulating
   the maturation of grafted human astrocytes.

   Together, human astrocytes in PFC_4MPT were found to be more
   functionally mature at transcriptomic level than HIP_4MPT, which was
   mainly due to the effects of dopaminergic and cholinergic
   neurotransmitters.

2.5. Morphological Complexity of Both Human and Mouse Astrocytes is Increased
in Mouse PFC

   To analyze the effects of neurotransmitters, DA and Ach, comparison of
   histology between PFC and HIP of SCID mice was carried out.
   Dopaminergic neurites (TH^+) were abundant in the PFC but rarely in the
   HIP (9.8 ± 0.81 µm/100µm^2 vs 1.7 ± 0.24 µm/100µm^2; P < 0.0001). There
   were more cholinergic positive neurites (VACHT^+) in PFC than HIP (2.3
   ± 0.19/100µm^2 vs 1.0 ± 0.17/100µm^2; P < 0.0001) (Figure [143]6A–C).
   These results suggested that the higher density of dopaminergic and
   cholinergic neurites in mouse PFC may contribute to higher expression
   of these neurotransmitters in the PFC, and hinted that grafted human
   astrocytes may receive more dopaminergic and cholinergic signals. Adult
   mouse astrocytes were used to assess the interaction between
   dopaminergic and cholinergic neurons and astrocytes. Therefore,
   colocalization of dopaminergic neurites and mouse astrocytes (GFAP^+)
   was analyzed. The colocalization was found to be higher in PFC than in
   HIP on total, proximal (<40 µm from soma), and distal astrocyte fiber
   (Figure [144]6D,E), which was consistent with cholinergic neurites
   (Figure [145]6F,G; Figure [146]S5C–F, Supporting Information).
   Meanwhile, human astrocytes (STEM123^+) were significantly colocalized
   with both dopaminergic and cholinergic neurons in PFC than in HIP in
   4MPT (Figure [147]6H,I). Therefore, both endogenous host and grafted
   human astrocytes received more dopaminergic and cholinergic signals in
   PFC.

Figure 6.

   Figure 6
   [148]Open in a new tab

   Histological Results Suggest Two Types of Neurons Enhance the
   Morphological Complexity of Astrocytes after Transplantation. A,B)
   Immunostaining presenting the density of dopaminergic and cholinergic
   neurons in PFC and HIP, respectively, in the adult mouse brain. TH: a
   marker for dopaminergic neurons, VACHT: a marker for cholinergic
   neurons. C) Comparison of the density of dopaminergic and
   acetylcholinergic neurons in A or B (n = 3 mice). D) Top images:
   Fluorescence image showing the colocalization of dopaminergic neuron
   processes and mouse astrocyte fibers in PFC and HIP, respectively.
   Bottom images: 3D reconstruction images reconstructed by Imaris (not
   the same astrocyte). GFAP: astrocyte marker. E) Quantification of
   colocalization of neuron processes and astrocyte fibers (proximal
   fiber, the distal fiber, and the overall fiber) in Figure D. The
   boundary (proximal and distal) was 40 µm from the center of the cell
   body (n = 3 mice). F,G) Immunostaining and quantification were
   performed on acetylcholinergic neurons in PFC and HIP (n = 3 mice). H)
   Representative colocalization images of human astrocyte fibers and
   mouse neuron processes (acetylcholinergic or dopaminergic neurons) in
   4MPT. STEM123: a human astrocyte marker. I) Quantification of
   colocalization of human astrocyte fibers and mouse neuron processes in
   H (n = 6 mice). J–L) Representative immunostaining images and
   quantification of the morphological complexity of human astrocyte
   processes in the HIP or PFC in 4MPT. Right panel in J shows a higher
   magnification of the boxed astrocytes in left panel in J;
   Sholl‐analysis, total branch length, number of primary branches, number
   of terminal branches, and effective area of cell bodies were compared
   (n = 6 mice). M–O) Comparison of morphology of mouse astrocytes in the
   HIP or PFC was performed, similar to J‐L. mPFC: mouse PFC; mHIP: mouse
   HIP (n = 3 mice). Data are represented as mean ± SEM; scale bar for
   stained images: 20 µm; scale bars for reconstructed images: 10 µm;
   two‐way ANOVA in K and N, two‐sided Student's t‐tests in C, E, G, I, L,
   and O, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

   Morphological divergence in human astrocytes between PFC and HIP was
   further analyzed. Morphology of human astrocytes was significantly more
   complex in PFC than in HIP in 4MPT (p = 0.0396). Furthermore, the
   branch length and effective areas of astrocytes were higher in PFC than
   in HIP, which also contributed to a more complex morphology
   (Figure [149]6J–L). Notably, morphological complexity of human
   astrocytes did not increase in PFC compared with HIP at 2MPT (p =
   0.6871) (Figure [150]S5H–J, Supporting Information). Additionally, the
   mouse astrocyte morphology in PFC and HIP was assessed and the
   morphological complexity was found to be significantly increased in PFC
   compared with that in HIP, too (p = 0.0057). Moreover, the total branch
   length, terminal branch number, and effective areas of astrocytes were
   significantly lower in the HIP than in PFC (Figure [151]6M–O; Figure
   [152]S5K, Supporting Information). The above results suggested that
   grafted human astrocytes in 4MPT showed more complex morphology in PFC
   than in HIP, similar to the host astrocytes.

   In summary, histology results showed the morphology of both human and
   mouse astrocytes were more complex in PFC than in HIP, which might be
   due to the higher density of dopaminergic and cholinergic neurons.

2.6. Neurotransmitters Enhance Functional Activity and Increase the
Morphological Complexity of Astrocytes

   To further determine how neurotransmitters (DA and ACh) affect
   astrocytes astrocyte morphology and functionality, purified human
   astrocytes were generated from human pluripotent stem cells in vitro.^[
   [153]^28 ^] Moreover, purified mouse primary astrocytes isolated from
   neonatal mouse brains were also acquired (Figure [154]7A; Figure
   [155]S6F, Supporting Information). Intracellular calcium and glutamate
   uptake are considered as the major functional indicators of
   astrocytes.^[ [156]^8 , [157]^29 ^] Therefore, the effect of
   neurotransmitters on the functional activity of astrocytes was tested.
   Intracellular calcium was increased dramatically (∆F[DA]/∆F[CONT] =
   3.5; ∆F[ACh]/∆F[CONT] = 2.7) in human astrocytes (GFAP^+TUJ‐1^−) with
   the treatment of DA or ACh, which indicated neurotransmitters might
   induce calcium responses and activate human astrocytes
   (Figure [158]7B,C,E). Additionally, there was no significant difference
   in the response time between DA treatment group and Ach treatment group
   (Figure [159]7D). Notably, correlation analysis revealed that
   astrocytes in group ACh and group DA clustered individually, which may
   indicate that each neurotransmitter has a different effect on
   astrocytes (Figure [160]7F). In addition, mouse astrocytes were
   examined and similar results were observed: both DA and ACh activated
   mouse astrocytes, no difference in response time, and clustering in
   correlation analysis (Figure [161]S6A–E, Supporting Information).
   Furthermore, low concentrations of transmitters (0.1 mM DA or 0.1 mM
   ACh) did not induce calcium responses in human astrocytes (Figure
   [162]S6G,H, Supporting Information). So, neurotransmitters may enhance
   the astrocyte function of intracellular calcium.

Figure 7.

   Figure 7
   [163]Open in a new tab

   Neurotransmitters Activate Cultured Astrocytes and Increase Their
   Morphological Complexity. A) Illustration showing the functional tests
   of human and mouse astrocytes in vitro. B) Representative snapshots of
   calcium imaging (by Rhod‐4) in human astrocytes at several time points.
   Left panels: snapshots at 0 s, 20 s, and 50 s in the test are shown
   respectively. CONT: control group, DA: dopamine group, ACh:
   acetylcholine group; DA+ACh: both transmitters group. Administration at
   10 s. Right panels: Cells in snapshots were verified astrocyte identity
   by immunostaining. C) Averaged calcium imaging traces. Line represents
   the mean of fluorescence intensity; colors represent groups (n = 8
   astrocytes for each group). D,E) Quantification of the response time of
   calcium signals (confirmed by (F/F0‐1)>10%) or the maximum amplitude of
   calcium signals in each group (n = 8 astrocytes for each group). F)
   Heatmap showing the correlation of calcium activity among astrocytes in
   four groups. Samples are clustered unbiasedly; red box indicates high
   correlation. G–I) Neurotransmitter ACh did not affect glutamate
   clearance capacity of human astrocytes in vitro (n = 3 replicates). J)
   Representative staining images of human astrocytes in vitro.
   Neurotransmitters were added for 12 hours. The white arrows indicate
   representative astrocytes from each group. K) Quantification of
   sholl‐analysis of four groups in J (n = 15 cells for each group). L,M)
   Comparison of morphology of primary astrocytes from neonatal mice among
   four groups, similar to J‐K. Insets showing the astrocytes processes
   tracing of fluorescent images (Cont, n = 13; DA, n = 14; ACh, n = 12;
   DA+ACh, n = 10). Data are represented as mean ± SEM; Scale bars in B:
   20 µm; scale bars in J, L: 100 µm; two‐way ANOVA in I, K, and M,
   two‐sided Student's t‐test in D, *p < 0.05, **p < 0.01, ***p < 0.001,
   ****p < 0.0001.

   Previous studies have shown that mature astrocytes take up
   neurotransmitters to maintain a homeostasis in synapse, which plays a
   vital role in preventing excitement toxicity.^[ [164]^30 ^] Therefore,
   a glutamate assay kit was used to assess whether neurotransmitters
   affected the ability of astrocytes to uptake glutamate in vitro. ACh
   did not affect the human astrocyte function of glutamate uptake
   (Figure [165]7G–I; Figure [166]S7D, Supporting Information).

   Furthermore, we examined the morphology of cultured human astrocytes
   with the treatment of DA or Ach neurotransmitters. Compared with the
   control group, the astrocytic morphological complexity significantly
   increased in neurotransmitter groups (Figure [167]7J,K). The total
   branch length of human astrocytes also significantly increased(CONT:
   112.2 ± 22.4 µm; DA: 175.0±20.8 µm; ACh: 226.5 ± 51.9 µm) (Figure
   [168]S7A, Supporting Information). Similarly, the morphology of mouse
   primary astrocytes also significantly increased with the treatment of
   neurotransmitters (Figure [169]7L,M). The total branch length and the
   number of terminal branches increased in the presence of
   neurotransmitters (Figure [170]S7C, Supporting Information). DA and ACh
   did not increase the astrocyte stress response, as evidenced by the
   absence of change in the GFAP^+ area (Figure [171]S7B, Supporting
   Information). Furthermore, after the treatment of 0.1 mM ACh, there was
   no significant change in morphological complexity, the total branch
   length, or the number of terminal branches of mouse astrocytes (Figure
   [172]S7E–J, Supporting Information), which may be associated with the
   results of calcium transient test.

   In conclusion, both DA and ACh increased the astrocytic activity, and
   increased morphological complexity of astrocytes in vitro, which was
   consistent with the in vivo results after organoid transplantation. The
   host brain micro‐environment has an effect on the cell fate of similar
   human organoids, which is mainly due to neurotransmitter signaling.

3. Discussion

   Recent studies showed that human brain organoids integrated into animal
   brains after grafting, which provides a new strategy for neural repair.
   Previously, we reported the survival and neural projection of human
   brain organoids after transplantation into the mouse cortex, as well as
   functional integration with host cells and behavioral improvement.^[
   [173]^7 ^] In this study, we transplanted human brain organoids and
   obtained the single‐cell atlas of human brain organoids after
   transplantation, by removing the transcriptomic data from host cells.
   We did not detect cell overgrowth or pluripotent genes. Above all, we
   have uncovered a developmental landscape of gene expression and lineage
   specification, and performed dual decoding of space and time after
   organoid grafting.

   Our research explores the developmental process of human brain
   organoids after transplantation. Early studies revealed that progenitor
   cells of the fetal brain first differentiated into neurons and then
   glial cells.^[ [174]^31 , [175]^32 ^] In this study, we found
   consistent results in grafted organoids, which indicates that grafted
   organoids still follow the intrinsic pattern of gene regulation and
   expression in the host brain. Alternatively, we performed a comparison
   of correlated brain region samples. The number of DEGs of grafted
   organoids in HIP and PFC was close to that of grafted organoids at 2MPT
   and 4MPT. In addition to intrinsic developmental patterns, the host
   brain environment is important for the cell fate decision of grafted
   organoids. Previous research showed that the main cell types of HIP and
   PFC of adult mice are similar;^[ [176]^17 ^] thus, we speculate that
   the brain region variations are related to cell function. Our data
   verified via enrichment analysis of biological processes as well as
   pseudo‐time analysis of total cells. Furthermore, we find that
   transplanted human cells are similar to the host brain region in
   transcriptional profile.

   These results also suggest that there is an unreported repairing
   mechanism in transplantation therapy. Early studies mainly focused on
   the exogenous effects of grafted cells to repair CNS, such as
   nutritional effects, cell renewal, and neuron projection.^[ [177]^9 ,
   [178]^33 , [179]^34 , [180]^35 ^] Interestingly, we found transplanted
   human cells obtained similar transcriptional features to the host brain
   region, which may be attributed to the effect of the host environment.
   Given that the fundamental principle of cell transplantation for CNS
   repair is to regenerate the damaged brain region, the findings suggest
   that the host environment also plays a crucial role in the endogenous
   effects of transplantation repair. Furthermore, due to the impact of
   the environment on transplanted cells, our research results have
   affirmed the efficacy of the orthotopic transplantation strategy.
   However, the diseased internal environment may negatively affect cell
   transplantation.

   In this study, we also described the paradigm of fate changes in human
   brain organoids before and after transplantation. Previous research has
   indicated there is a substantial alleviation of cell stress response
   after organoid grafting;^[ [181]^16 ^] another report on
   transcriptional comparison of cortical organoids before and after
   transplantation showed that the expression of mature neuron genes is
   upregulated after organoid transplantation.^[ [182]^6 ^] In our study,
   first we made a comparison of samples before and after grafting,
   time‐series samples, and distinct region samples, and found the
   transcriptome diversity of the organoids was greatest before and after
   grafting. Furthermore, we found that this effect may be due to neuron
   development and maturation, cell projection, and neuron migration as
   well as an alleviation of hypoxia stress reactions after organoid
   grafting. Therefore, our research suggests better development and
   maturation of grafted human cells compared with brain organoids in
   vitro, which enhances the importance of in vivo animal models for stem
   cell research.

   Our research demonstrated the key role of neurotransmitters DA and ACh
   in brain region variations. Recently, the brain region variations have
   been extensively reported, and astrocytes play an important role in
   it.^[ [183]^36 ^] Specifically, astrocytes display diverse morphologies
   across brain regions.^[ [184]^37 ^] In this study, we found the
   diversity of human astrocytes in PFC and HIP might be due to the
   endogenous transmitters (DA and Ach), since neither neuron subtype
   (dopaminergic or cholinergic) was found in our grafted organoids. Then
   we found distinctions of density of endogenous mouse neurons between
   brain regions. Moreover, we find the conservativeness of the effect of
   DA and ACh on both human and mouse astrocytes. Therefore, our results
   suggest that neurotransmitters are crucial for the brain regions
   variations in HIP and PFC.

   In this study, we also described the vascularization of organoids after
   transplantation. Successful vascularization was found to be important
   for organoid survival after grafting.^[ [185]^5 ^] We found the blood
   vessels in grafted human organoids grew from the host, and organoids in
   4MPT were more extensively vascularized than 2MPT (Figure [186]S3P,Q,
   Supporting Information). There was no significant difference in the
   vascularization of grafted organoids between HIP and PFC groups (Figure
   [187]S4E–G, Supporting Information).

   Our scRNA sequencing data revealed that there was a proportion of
   neural progenitor cells in organoids after transplantation. However,
   the majority of progenitor cells in our study were neural precursor
   cells(high PAX6 and NESTIN expression, low KI67 expression), which are
   restricted to the fate of neuron lineage, and could serve as a cell
   pool for regeneration. We found the proportion of cycling cells was ≈2%
   according to our sequencing data, which may decrease safety concerns.
   Additionally, there were almost no Ki67‐positive cells among grafted
   cells, as shown by immunostaining, indicating a low proportion of the
   devising cells. In addition, we did not observe any malignant
   proliferation in any of the mice after transplantation.

   Our results underscore the effect of neurons on the development and
   function of astrocytes. Currently, the role of astrocytes in providing
   nutritional support to neurons has been reported, but the effect of
   neurons on astrocytes is not yet clear; the effect of neurotransmitters
   on neurons has been widely studied,^[ [188]^39 , [189]^40 ^] whereas
   the effects of neurotransmitters on glial cells have been relatively
   less explored. In our study, two neurotransmitters stimulated the
   calcium activity of astrocytes and enhanced their morphological
   complexity. Therefore, we revealed the important role of
   neurotransmitters in the development of astrocytes, reflecting the
   complexity of interaction networks between neurons and astrocytes.

   In summary, our results report the cell fate changes of brain organoids
   after transplantation, and revealed that the host micro‐environment
   might regulate the cell function and fate of astrocytes in grafted
   organoids. These results might contribute to subsequent in vivo
   adjustments of cell fate of organoids and regulation of newborn neuron
   projections, and our study provides a theoretical foundation for the
   application of human brain organoids in neural repair.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   Conceptualization was carried out by Y.L. and M.L.; Methodology was
   prepared by M.L. and S.‐B.X.; Software was dealt with by S.‐B.X., X.L.,
   and Y.H.; Investigation was conducted by S.‐B.X., X.‐R.L., and P.F.;
   The acquisition of resources was done by Y.C. and X.M.; S.‐B.X., X.L.,
   H.X., S.W., C.C., and L.X.R. helped by Writing – Original Draft;
   Writing – Review & Editing were done by Y.L., X.G., and M.L.;
   Visualization was by S.‐B.X. and L.X.R.; Y.L., M.L., and X.G.
   supervised the project; Y.L. and X.G. performed funding acquisition.

Supporting information

   Supporting Information
   [190]ADVS-11-2402287-s001.pdf^ (2.2MB, pdf)

   Supporting Information
   [191]ADVS-11-2402287-s002.docx^ (26.6KB, docx)

Acknowledgements