Integrative Single-Cell Transcriptomics and Epigenomics Mapping of the Fetal Retina Developmental Dynamics
The underlying mechanisms that determine gene expression and chromatin accessibility in retinogenesis are poorly understood. Herein, single-cell RNA sequencing and single-cell assay for transposase-accessible chromatin sequencing are performed on human embryonic eye samples obtained 9–26 weeks after conception to explore the heterogeneity of retinal progenitor cells (RPCs) and neurogenic RPCs. The differentiation trajectory from RPCs to 7 major types of retinal cells are verified. Subsequently, diverse lineage-determining transcription factors are identified and their gene regulatory networks are refined at the transcriptomic and epigenomic levels. Treatment of retinospheres, with the inhibitor of RE1 silencing transcription factor, X5050, induces more neurogenesis with the regular arrangement, and a decrease in Müller glial cells. The signatures of major retinal cells and their correlation with pathogenic genes associated with multiple ocular diseases, including uveitis and age-related macular degeneration are also described. A framework for the integrated exploration of single-cell developmental dynamics of the human primary retina is provided.
The eye is a vital and highly specialized visual organ, and the retina is the most important component of vision production in the eye. The retina primarily comprises six types of neurons and several types of glial cells.
Therein, photoreceptor cells, including cones and rods in the outer nuclear layer (ONL), receive and process light signals from the external environment.
The interneurons, including amacrine cells (ACs), bipolar cells (BCs), and horizontal cells (HCs) in the inner plexiform layer (IPL), inner nuclear layer (INL), and outer plexiform layer (OPL), deliver signals from the photoreceptor cells to retinal ganglion cells (RGCs) in the ganglion cell layer (GCL).
Inside the RGCs, these light signals are converted to electrical signals and transmitted to the brain.[1]
The primary types of retinal glial cells include Müller glial cells (MGCs) and microglia.
MGCs mainly act as mediators to assist photoreceptor cells in light absorption, provide nutrients to neurons, and remove metabolic waste,[2, 3]
whereas microglia are resident immune cells in the retina and central nervous system that have a critical role in the maintenance of normal homeostasis and immune surveillance of these systems.[4]
The retina primarily develops from retinal progenitor cells (RPCs), which differentiate into six classes of neurons and one class of glial cells at chronologically separate, yet frequently overlapping, intervals during development. Before becoming terminal cells, RPCs typically transition through a precursor cell stage. Distinct evolutionary fates are selected as the RPCs reach saddle points, where they segregate into different and progressively restricted precursor cell states and finally develop and mature into terminal cells.[5-9] Although terminal-cell specification has received considerable attention, it remains unclear how heterogeneity develops within RPCs.
The general process of retinal development is meticulously regulated by cell-type-specific transcription factors (TFs) that recruit chromatin effectors to repurpose the chromatin and promote new retinal cellular characteristics.[10]
Dysregulation of any step in this process can cause varying degrees of visual dysfunction and congenital diseases, including retinoblastoma, Leber congenital amaurosis, and autosomal recessive retinitis pigmentosa.[10-13]
Meanwhile, the development of retinal cell lineages remains to be investigated, and prospective critical factors have not yet been thoroughly characterized.
Recently, the single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) and single-cell RNA sequencing (scRNA-seq) have proven effective for the analysis of human embryonic retinal development. More specifically, scRNA-seq has been employed to investigate the differentiation trajectory of RPCs and unique subpopulations of embryonic retinal cells. The associated studies have reported that myriad TFs have crucial roles in directing the development of specific lineages, such as nuclear factor I (NFI) in the regulation of cell-cycle exit and generation of late-born retinal cell types, and atonal bHLH transcription factor 7 (ATOH7) in the specification of cone photoreceptors.[14-17]
Meanwhile, gene expression patterns of important lineage-defined TFs are primarily regulated by epigenetic programs that are influenced by changes in chromatin accessibility and can be detected by scATAC-seq. For instance, the sequences of various TF cascades responsible for determining cell fate have been verified at the epigenomic level from a developing human retina database using scATAC-seq.[18-20] However, integrated scRNA-seq and scATAC-seq datasets from the same human embryonic eye sample are lacking; hence, transcriptomic and epigenetic results have the potential to be better matched.
Here, we performed an integrative analysis of scATAC-seq and scRNA-seq on individual cells obtained from human embryonic eyes to capture the dynamic transcriptomic and epigenetic landscapes of the developing human eye at single-cell resolution.
To this end, we probed the precursors of MGCs and the intrinsic connection among neurogenic RPCs (NPCs).
We then constructed developmental trajectories and gene regulatory networks (GRNs) for RPC-derived cells. (Figure1A).
In this way, we defined the continuous evolution of TF motif activity related to neuronal specification and identified the co-dependence of TF motif accessibility along these trajectories.
By investigating whether RE1 silencing transcription factor (REST) influences RPC and MGC fate, we found that treatment of retinospheres with a REST inhibitor induced neurogenesis with a regular spatial arrangement and decrease in MGCs.
Finally, we identified differences between embryonic macrophages and microglia and combined disease-related genes from retina-related diseases to characterize relationships between risk factors and specific retinal cell types.
Collectively, this study highlights neurogenic cell fate determination
processes and elucidates a portion of the gene regulation involved in
retinal development.
(Figure1A).
Using scRNA-seq, most cells were divided into 17 groups according to the normalized expression of known markers (Figure 1D,E; Table S2, Supporting Information). Due to the limited number of MGCs (38 cells, 0.0505% of total cell count) and microglia (11 cells, 0.0146% of total cell count) directly mapped to the chromatin landscape in scRNA-seq, subsequent analysis was not performed (Tables S3 and S4, Supporting Information). We then integrated the gene expression levels with corresponding gene activity scores and found that major cluster annotations in the retina of matched cells were consistent. A group of RPCs expressing CCND1, HES1, and SOX2 was identified, and several NPC markers, including ATOH7, HES6, and DLL3, were expressed by cells in one scRNA-seq cluster. Additionally, a group of RGCs expressing NEFM, SNCG, and GAP43, as well as genes related to HC identification (ONECUT1 and ONECUT2), was detected. Moreover, a group of cells expressing AC-related genes (GAD2 and TFAP2A) was observed. Among the BC clusters, a group of cells expressing VSX1 and TMEM215, and distinct clusters of photoreceptor cells, including cones (expressing PRDM1 and THRB) and rods (expressing NRL, RHO) were detected. We further annotated a cluster of cells highly expressing RGR, GPX3, and RLBP1 as MGCs and determined that RGR might be more specific to MGCs than RLBP1 during the embryonic period. Furthermore, we observed gene expression in clusters of retinal pigment epithelia (RPE) (RPE65 and TTR), fibroblasts (DCN and MGP), melanocytes (PMEL and TYRP1), endothelial cells (PECAM1 and ENG), lens cells (CRYGC and LIM2), and corneal epithelium (KRT5). Numerous markers exhibited dynamic gene activity scores in related scATAC-seq clusters (Figure 1F,G). The emergence of these major retinal cell types reflected normal retinal development (Figure S1A, Supporting Information).
Considering that RPCs have a two-branch developmental tendency, including glial cell orientation and neuronal orientation,[6, 14, 16] we extracted RPCs, NPCs, and MGCs, as well as six major neurons for pseudo-time analysis via two different approaches (Palantir and Monocle3). The results verified that RPCs could differentiate into glia or NPCs, and transitional NPCs could develop into neurons of three lineages (Figure 1H,I; Figure S1B, Supporting Information).
Collectively, we generated a collaborative database that describes retinal development, laying the groundwork for subsequent analyses.
Construction of Gene Regulatory Networks in Retinal Progenitor Cell-Producing Cells at the Transcriptomic and Epigenetic Levels
To identify which TFs are crucial for the differentiation of RPCs into major retinal cells and for the maintenance of their specific cell end-states, we performed Python implementation of a single-cell regulatory network of inference and clustering (SCENIC) analysis. Moreover, we assessed the relative accessibility to chromatin and enrichment of motifs for these TFs (Figure 2).
The results predicted a series of TFs, with highly similar transcriptomic and epigenomic levels, as vital for the fate of specific cell types (Figure 2A–C; Figure S2A, Supporting Information). UMAP embedding of SCENIC TF activity revealed four developmental branch endpoints derived from RPCs, which were demarcated and represented by respective TFs reported to affect the development of related lineages (Figure 2D), such as SOX8 for the MGC lineage, POU4F3 for RGC lineage, TFAP2B for HC/AC lineage, and LHX4 and NEUROD1 for BC/PH lineage.[21, 23-25]
Next, we screened critical cell-type-specific TFs at the epigenetic level from the top predicted TFs; the 10 genes regulated by at least two key TFs, according to the weight value, were selected to depict the top GRNs of related TFs for each cell linage (Figure 2E; Figure S2B,C, Supporting Information). For instance, the Iroquois homeobox protein family TFs (IRX1 and IRX2) exhibited a marked influence on genes required for RGC development, such as EBF3, NRN1, and NEFL,[26, 22] whereas ONECUT family members had regulatory effects on HC/AC lineage-specific factors, including PROX1 and BARHL2.[27-29]
In summary, we identified fate-determining TFs for RPCs and their products, constructed.... more in link
Inga kommentarer:
Skicka en kommentar