Synergistic inhibition of Notch signaling and forced cell cycle re-entry drive Müller glia reprogramming in uninjured mouse retina

eLife Assessment

This study shows that combining forced cell cycle re-entry with Rbpj deletion enhances Müller glia dedifferentiation and promotes their conversion into retinal neuron-like cells in the uninjured mouse retina. It provides a valuable strategy for improving Müller glia-mediated neurogenesis and advancing regenerative potential in the mammalian retina. Overall, the data are convincing. The authors have also addressed concerns regarding Müller glia function, cell survival, and the limitations of neuronal maturation and integration, further strengthening the conclusions of the study.

https://doi.org/10.7554/eLife.111251.3.sa0

Significance of the findings:

Valuable: Findings that have theoretical or practical implications for a subfield

Landmark
Fundamental
Important
Valuable
Useful

Strength of evidence:

Convincing: Appropriate and validated methodology in line with current state-of-the-art

Exceptional
Compelling
Convincing
Solid
Incomplete
Inadequate

During the peer-review process the editor and reviewers write an eLife Assessment that summarises the significance of the findings reported in the article (on a scale ranging from landmark to useful) and the strength of the evidence (on a scale ranging from exceptional to inadequate). Learn more about eLife Assessments

Abstract
Introduction
Results
Discussion
Materials and methods
Data availability
References
Article and author information
Metrics

Abstract

In regenerative species, such as teleost fish, Müller glia (MG) autonomously re-enter the cell cycle after injury and give rise to functional retinal neurons. In contrast, the loss of retinal neurons in mammals is irreversible due to the limited proliferative and regenerative ability of MG. Various strategies have been developed to induce proliferation of mature mouse MG with or without injury, yet most MG daughter cells retain glial cell fate. Here, we found that MG progenies maintain high Notch signaling, which may constrain their neurogenic potential. Conditional deletion of Rbpj, the central transcriptional effector of Notch, induced limited MG-to-neuron conversion in mature MG without proliferation. However, Rbpj deletion, combined with forced MG proliferation by overexpressing Ccnd1 and suppressing Cdkn1b, significantly promoted MG dedifferentiation and ectopic expression of the neuronal marker Otx2 in MG daughter cells in uninjured mouse retina. Combining Notch inhibition with MG cell cycle re-activation not only increased the numbers of bipolar- and amacrine-like cells generated from MG but also promoted the further differentiation toward ON-cone, OFF-cone, and rod-bipolar subtypes. Single-nucleus RNA and ATAC sequencing data revealed that Notch inhibition facilitated the formation of MG-derived progenitor-like cells while MG proliferation increased chromatin accessibility of neurogenic genes. Notably, most MG-derived cells survived long term despite incomplete maturation. Together, our findings delineate how Notch inhibition and MG proliferation, alone or in combination, influence the regenerative potential of MG in the mammalian retina.

Introduction

Müller glia share a common lineage with retinal neurons, serving as the only glial cell type differentiated from retinal progenitor cells (RPCs) while retaining a transcriptomic profile similar to their progenitors (Roesch et al., 2012; Jadhav et al., 2009; Cepko et al., 1996). Beyond their role in maintaining retinal homeostasis, MG function as a latent stem cell population in lower vertebrates. In zebrafish, retinal injury triggers a robust regenerative response wherein quiescent MG dedifferentiate, re-enter the cell cycle, and undergo asymmetric division to self-renew and generate multipotent retinal progenitors. These progenitor cells subsequently differentiate to regenerate all major retinal neuron types, replacing damaged cells (Todd and Reh, 2022; Lahne et al., 2020; Wan and Goldman, 2016). However, this regenerative capacity is progressively restricted across vertebrate evolution. While MG in young chicks can initiate a single mitotic cycle and transition to MG-derived progenitor cells (MGPCs), MGPCs only differentiate into amacrine cells and bipolar cells (Fischer, 2005; Fischer and Reh, 2001). In mice and humans, MG lack the intrinsic capacity to autonomously re-enter the cell cycle or regenerate lost neurons, representing a major barrier to treating retinal degenerative diseases (Jadhav et al., 2009; Wan and Goldman, 2016; Bringmann et al., 2009).

Overcoming MG quiescence is the first step toward regeneration. We previously demonstrated that the high level of Cdkn1b, a cell cycle inhibitor, and the low level of Ccnd1 in mature mouse MG prevent them from re-entering the cell cycle. Concurrent downregulation of Cdkn1b and upregulation of Ccnd1, which was delivered via a single AAV vector termed the Cell Cycle Activator (CCA), synergistically promote MG proliferation. CCA-driven MG proliferation further promoted the dedifferentiation of MG following mitosis, but the vast majority of MG eventually reverted to their glial identity (Wu et al., 2025). Similarly, direct activation of the Wnt/β-catenin signaling pathway or bypassing the Hippo pathway via overexpressing Hippo non-responsive form of YAP (YAP5SA) also led to spontaneous re-entry into the cell cycle and transiently reprogramming into a progenitor cell-like state, but no regeneration of mature neurons was reported (Yao et al., 2016; Rueda et al., 2019; Hamon et al., 2019). These findings indicate that while cell proliferation is necessary to expand the MG pool and initiate dedifferentiation, it is insufficient to drive functional neurogenesis, implying that the existence of additional molecular barriers that enforce glial identity.

Notch signaling represents a primary candidate for this barrier. During development, Notch acts as a binary switch, maintaining the RPC pool and specifying glial fate. Downregulation of Notch is required for neuronal differentiation, whereas sustained activity promotes MG formation (Rueda et al., 2019; Mills and Goldman, 2017; Shimojo et al., 2008). After development, this pathway remains active in mature MG, enforcing quiescence in the uninjured retina (Mills and Goldman, 2017; Conner et al., 2014). In zebrafish, injury induces the downregulation of Notch receptors and downstream target genes Hes/Hey family, thereby derepressing the proneural genes Ascl1 to promote neurogenesis (Wan and Goldman, 2016; Mills and Goldman, 2017; Goldman, 2014; Campbell et al., 2021; Elsaeidi et al., 2018). In the mammalian central nervous system, Notch signaling actively suppresses neurogenesis in contexts ranging from the cortex to the cochlea (Li et al., 2025; Shu et al., 2019; Zamboni et al., 2020; Magnusson et al., 2014). Recent studies in the mouse retina have shown that disrupting Notch, particularly in combination with knockout of Nuclear factor I a/b/x (Nfi a/b/x) or overexpressing the Yamanaka factor, octamer-binding transcription factor 4 (Pou5f1), enhance MG reprogramming efficiency (Le et al., 2025; Le et al., 2024).

In this study, we investigated why postmitotic MG fail to yield neurons. We found that Notch signaling remains active in the postmitotic MG-derived cells, which may prevent their diversion toward a neuronal fate. We hypothesized that inhibition of Notch signaling would unlock the neurogenic potential of proliferating MG. By combining CCA treatment with MG-specific deletion of the Notch transcriptional effector, Recombination signal binding protein for immunoglobulin kappa J region (Rbpj), we achieved robust reprogramming of MG into bipolar- and amacrine-like neurons in uninjured adult mouse retina. Single-nucleus RNA sequencing (snRNA-seq) and RNA in situ hybridization revealed that these newborn neurons exhibit a retinal neuron transcriptional signature, expressing markers characteristic of amacrine cells (ACs) and subtypes of bipolar cells (BCs). Single nucleus ATAC-sequencing (snATAC-seq) analysis demonstrated that CCA treatment induced chromatin opening at key neurogenic genes, a priming event that is functionally capitalized upon by Notch inhibition. Notably, the reprogrammed MG displayed long-term survival up to 9 months post-treatment. Together, our study demonstrated the synergistic effects of cell proliferation and Notch inhibition on MG reprogramming in the absence of retinal injury, providing molecular and temporal insights into this process.

Results

CCA-induced MG progeny maintains high Notch signaling

Previously, we developed an AAV_7m8-GFAP-Ccnd1-Cdkn1b shRNA vector, designated as CCA, which enables simultaneous Ccnd1 overexpression and Cdkn1b knockdown specifically in MG (Figure 1a; Wu et al., 2025). When injected intravitreally, this vector achieves near-complete MG transduction near the injection site and drives more than half of the MG population to enter the cell cycle once (Wu et al., 2025). Following mitosis, these MG transiently and partially dedifferentiated, but the vast majority of postmitotic MG revert to their original glial identity by four months post-treatment (Wu et al., 2025). Fewer than 1% MG completely lost MG identity, as shown by negative Sox9 staining and decreased GFP signal (Figure 1—figure supplement 1a and b), while expressing high levels of Otx2, a pro-neuronal marker (Figure 1—figure supplement 1c).

Figure 1 with 7 supplements see all

Download asset Open asset

MG and MG-derived cells maintained high levels of Notch signaling.

(a) Schematic representations of the intravitreal injection of AAV_7m8-GFAP-*Ccnd1-Cdkn1b* shRNA-WPRE (CCA). (b) The overview process of MG proliferation induced by CCA. (c) *Hes1* mRNA in situ hybridization in the control and the *Glast^CreERT;Rosa26^Sun1-GFP* mouse retinas harvested at four months post CCA injection. (d) Magnified views of the highlighted regions in (c). n=3 mice.

To understand this blockade in neurogenesis, we analyzed the single-cell RNA sequencing (scRNA-seq) data from our previous study (Wu et al., 2025) to characterize the activity of Notch signaling in postmitotic MG. We found that key Notch signaling components, including the receptor Notch1, the central regulator Rbpj, and the downstream effector Hes1, remained highly expressed in postmitotic MG following cell cycle re-activation (Figure 1—figure supplement 1d–g). In contrast, these genes were expressed at negligible levels in native retinal neurons, such as rod photoreceptors (Figure 1—figure supplement 1f and g).

We confirmed this finding using RNA in situ hybridization to assess Hes1 mRNA levels in the Glast^CreERT;Rosa26^Sun1-GFP mice, in which MG nuclei were specifically labeled by nuclear membrane-localized Sun1-tagged GFP. In control retinas, Hes1 mRNA was abundant in MG within the inner nuclear layer (INL) but minimal in the photoreceptors residing in the outer nuclear layer (ONL) (Figure 1c and d). In CCA-treated retinas, even in the MG-derived cells that had migrated to the ONL, high Hes1 mRNA levels were maintained (Figure 1c and d). Given the established role of Notch as a suppressor of neurogenesis, we hypothesized that this persistent Notch activity constitutes the primary barrier preventing CCA-treated MG from differentiating into neurons. Consequently, we reasoned that inhibiting Notch signaling in MG would remove this brake and promote their successful reprogramming into neurons.

Rbpj deletion in late RPC promotes rod genesis at the expense of MG and bipolar cells

Rbpj is the essential nuclear mediator of Notch receptors (Campbell et al., 2022; Borggrefe and Oswald, 2009; Yoon and Gaiano, 2005). Rbpj knockout in the retina led to efficient inhibition of Notch signaling, as validated by lowered Hes1 mRNA expression in Rbpj^-/- cells (Figure 1—figure supplement 2a–e). Glast is expressed in RPCs in neonatal mice and later restricted to the specified MG (Trimarchi et al., 2008; Nelson et al., 2011). Prior to assessing MG-specific Rbpj knockout, we first induced Rbpj deletion in the late RPCs by administering Tamoxifen (TAM) to Glast^CreERT;Rbpj^flox/flox; Rosa26^{tdTomato(tdT)} mice at postnatal day 1 (P1), then analyzed retinas at P12 (Figure 1—figure supplement 3a and b). Rbpj^-/- retinas had a significantly higher proportion of late-born rod photoreceptors, labeled by tdT in the ONL, compared to the control and Rbpj^+/- retinas, whereas the early-born cone photoreceptors, identified by the mCAR marker, were unaffected (Figure 1—figure supplements 3c, d and 4a–c). Conversely, Rbpj deletion significantly impaired the generation of MG (Sox9+) and BCs (Otx2+ in the INL) from late RPCs (Figure 1—figure supplements 3c, e and 5a–c). The generation of ACs, labeled by HuC/D and Pax6 in the INL, and retinal ganglion cells (RGCs), labeled by Rbpms in the RGC layer, remained unaffected (Figure 1—figure supplements 6a–c and 7a–c). These developmental data confirm that Rbpj removal biases progenitors toward a photoreceptor at the expense of MG and BCs, consistent with the previous report using Notch1 knockout in the late PRCs (Jadhav et al., 2006b).

Rbpj loss in mature MG triggers inefficient direct glia-to-neuron conversion

We next tested whether Notch inhibition alone is sufficient to reprogram mature MG. We induced Rbpj deletion in fully developed retinas at P28 and analyzed them at 3 weeks and 4 months post-treatment (Figure 2a). At 3 weeks, all Rbpj KO MG expressed the glial marker Sox9 (Figure 2b–d). However, by 4 months, approximately 7.6% of GFP-positive MG-derived cells had lost Sox9 and 7.5% had begun expressing the neuronal marker Otx2 (Figure 2e, Figure 2—figure supplement 1a–c). These findings indicate that Notch inhibition alone elicits a slow, progressive dedifferentiation in a subset of MG over time. Unlike in the developmental context, no photoreceptor-like cells were formed. Of note, these Rbpj-deficient MG did not incorporate EdU (Figure 2—figure supplement 2a and b), indicating that Notch inhibition unlocks a slow, inefficient transdifferentiation process without inducing cell cycle activation.

Figure 2 with 2 supplements see all

Download asset Open asset

*Rbpj* deletion in adult MG induces limited dedifferentiation.

(a) Schematic illustration of MG dedifferentiation and reprogramming experiment. (b) Representative immunostaining of Sox9 on retinal sections from *Glast^CreERT;Rosa26^Sun1-GFP* and *Glast^CreERT;Rbpj^flox/flox;Rosa26^Sun1-GFP* mice in different timepoints post TAM injection. The white arrows refer to GFP+ Sox9- cells. (c) Magnified views of the highlighted regions in (b). (d) Percentage of GFP+ Sox9- cells in overall GFP+ cells. n=3 mice, data are presented as mean ± SEM. ns = not significant, ****p<0.0001, by one-way ANOVA with Tukey’s post hoc test. (e) Percentage of GFP+ Otx2+ cells in overall GFP+ cells. n=3 mice, data are presented as mean ± SEM. ns = not significant, **p<0.01, by one-way ANOVA with Tukey’s post hoc test.

Notch inhibition reduces but does not abolish CCA-induced MG proliferation

Since direct conversion depletes MG, we aimed to combine CCA-induced proliferation to expand the MG pool with Rbpj KO-driven reprogramming. Because Notch inactivation drives premature cell-cycle exit during development (Jadhav et al., 2006b; Yaron et al., 2006; Jadhav et al., 2006a), we first determined if Rbpj deletion would antagonize the mitogenic activity of CCA. We administrated TAM from P28 to P35 to induce MG labeling and Rbpj knockout, followed by CCA injection at P36 (Figure 3—figure supplement 2a). Subsequently, EdU was administrated intraperitoneally daily for 21 days, spanning the major time window of MG proliferation (Figure 3—figure supplement 1a–c), and the mice were harvested to assess MG proliferation rates (Figure 3—figure supplement 2a).

Remarkably, even in the absence of Notch signaling, CCA maintained a robust capacity to drive MG proliferation. Although the total number of EdU+ MG was moderately decreased in Rbpj^-/- mice compared to wild type and Rbpj^+/- controls, a substantial population of MG still successfully proliferated (Figure 3—figure supplement 2b and c). Reversing the orders of TAM and CCA treatments yielded similar robust results (Figure 3—figure supplement 2d–g). While this moderate reduction could reflect a minor anti-proliferative effect of Notch inhibition, it is also possible that Rbpj deletion down-regulates GFAP promoter activity in dedifferentiated MG and thereby lowers vector expression. Crucially, the persistence of numerous proliferated cells demonstrates that Ccnd1 overexpression and Cdkn1b suppression overrode these factors. This confirms that these downstream cell-cycle regulators act as potent drivers of proliferation independent of upstream Notch signaling (Rowan et al., 2008).

Combining Rbpj deletion and CCA treatment induces robust MG dedifferentiation and ectopic expression of Otx2

We next evaluated the reprogramming efficiency of the combined treatment of CCA and Rbpj KO in Glast^CreERT;Rbpj^flox/flox;Rosa26^{tdT or Sun1-GFP} mice. Ideally, Rbpj deletion should be induced immediately after MG proliferation. However, CCA-driven MG proliferation in adult mice is not a synchronized process, with most MG proliferations occurring from week 1 to week 6 after CCA injection (Figure 3—figure supplement 1a–c). To ensure precise lineage tracing and minimize confounding effects of CCA on gene expression, we adopted the protocol of administering TAM immediately prior to CCA injection (Figure 3—figure supplement 3a). At 3 weeks post-treatment, MG in all groups retained Sox9 expression (Figure 3—figure supplement 3b–d). By 4 months, approximately 27.8% of GFP-positive MG-derived cells in the combination group had lost Sox9 expression, a significantly higher fraction than in the CCA alone (1.5%) or Rbpj KO alone (7.6%) groups (Figure 3—figure supplement 3b–e). Importantly, the total number of Sox9+ MG remained higher in the combination group than that in the control, indicating that the combined treatment did not deplete the glial pool (Figure 3—figure supplement 3f).

The enhanced dedifferentiation was accompanied by progressively more robust neurogenesis over time. Lineage tracing revealed no Otx2⁺ MG-derived cells in the combination group at 3 weeks post-treatment (Figure 3a–d). By 2 months, approximately 9.2% of MG-derived cells had acquired Otx2 expression, and this proportion continued to rise to 27.4% by 4 months (Figure 3b–e). This progressive increase indicates that MG-to-neuron conversion in the combination group unfolds gradually over months rather than as an abrupt switch. Compared with the CCA-alone and Rbpj deletion-alone groups, the percentage of Otx2⁺ MG-derived cells in the combined treatment group increased by approximately 20-fold and 3-fold, respectively (Figure 3e, Figure 3—figure supplement 4a). Injection of a control AAV_7m8-GFAP-GFP vector did not increase the percentage of Otx2+ MG (Figure 3—figure supplement 5a–c), indicating that the enhanced MG reprogramming was driven by the transgenes rather than by the AAV vector itself or the injection procedure.

Figure 3 with 5 supplements see all

Download asset Open asset

*Rbpj* KO and CCA synergistically increase Otx2+ cell formation from MG.

(a) Schematic illustration of the neurogenesis assessment experiment. (b) Representative immunostaining of EdU and Otx2 on retinal sections. The white arrows refer to tdT+ Otx2+ cells. (c) Magnified views of the highlighted regions in (b). The white arrows refer to tdT+ Otx2+ cells. (d) Percentage of tdT+ Otx2+ cells in overall tdT+ cells. 3W: 3 weeks, 2M: 2 months, 4M: 4 months, n=3 mice, data are presented as mean ± SEM. ns = not significant, ***p<0.001, by one-way ANOVA with Tukey’s post hoc test. (e) Percentage of tdT+ Otx2+ cells in overall tdT+ cells. n=3 mice, data are presented as mean ± SEM. ns = not significant, *p < 0.05, ***p<0.001, ****p<0.0001, by one-way ANOVA with Tukey’s post hoc test. (f) Representative immunostaining of EdU and Otx2 or Crx on retinal sections. (g) Percentage of tdT+ EdU+ Otx2+ cells in ONL tdT+ EdU+ cells. n=3 mice, data are presented as mean ± SEM. ns = not significant, ****p<0.0001, by one-way ANOVA with Tukey’s post hoc test. (h) Percentage of tdT+ EdU+ Crx+ cells in ONL tdT+ EdU+ cells. n=3 mice, data are presented as mean ± SEM. ns = not significant, ****p<0.0001, by one-way ANOVA with Tukey’s post hoc test.

Given that photoreceptor loss is a leading cause of incurable blindness worldwide (Molday and Moritz, 2015; Hartong et al., 2006), reprogramming MG into photoreceptors represents a promising therapeutic strategy. We examined whether MG-derived cells adopted photoreceptor identities. In the combination group, a subset of postmitotic MG in the ONL upregulated photoreceptor markers, including Otx2 (12.8% of ONL tdT+ EdU+ MG) and Crx (8.2% of ONL tdT+ EdU+ MG), consistent with potential progression toward photoreceptor fate (Figure 3f–h). The soma of some MG-derived cells in the ONL also adopted a photoreceptor cell-like circular morphology but without any structure of outer segment (Figure 3f). However, only about 2% of ONL MG-derived cells expressed Nrl, an important factor regulating the rod photoreceptor specification (Mears et al., 2001; Figure 3—figure supplement 4b and c), which may limit maturation into fully developed rod photoreceptors.

In summary, these findings collectively demonstrate that the combined treatment of CCA and Rbpj deletion significantly enhances both the dedifferentiation of MG and the subsequent generation of retinal neurons compared to either treatment alone.

Rbpj deletion promotes the formation of neuronal progenitor cells from MG

To dissect the transcriptional trajectory of reprogramming, we performed snRNA-seq on purified MG from Glast^CreERT;Rbpj^flox/flox;Rosa26^Sun1-GFP mice (Figure 4a). We analyzed four experimental groups, the untreated control (Ctrl), CCA treatment (CCA), Rbpj deletion (Rbpj KO), and combined Rbpj deletion and CCA treatment (Rbpj KO+CCA), at 1 week, 3 weeks, and 4 months post-treatment (Figure 4a).

Figure 4 with 3 supplements see all

Download asset Open asset

snRNA-seq analysis shows that *Rbpj* KO and CCA promote MG reprogramming.

(a) Schematic illustration of the snRNA-seq experiment. To induce *Rbpj* deletion and GFP expression in MG, we administered TAM for seven consecutive days in *Glast^CreERT;Rosa26^Sun1-GFP* and *Glast^CreERT;Rbpj^flox/flox;Rosa26^Sun1-GFP* mice from P28. At P36, CCA was injected into the CCA and *Rbpj* KO+CCA groups, while control and *Rbpj* KO mice were uninjected by any AAV vector. 3–4 retinas from mice at indicated ages were pooled together for nuclei extraction, and MG nuclei were then isolated by GFP signal via FACS for snRNA-seq. (b) UMAP plot of snRNA-seq data from four treatment groups (Ctrl, CCA, *Rbpj* KO, *Rbpj* KO+CCA) at different timepoints, and separated UMAP based on different timepoints. Clusters were identified based on known marker gene expression. (c) Proportions of cell clusters within Ctrl, CCA, *Rbpj* KO and *Rbpj* KO+CCA groups at different timepoints. (d) Feature plots highlighting the cluster of quiescent MG (*Kcnj10*), proliferating MG (*Mki67*), reactivated MG (*Gfap*), MGPC (*Dll1*, *Ascl1*), transitional MG (*Mt-Nd4*), AC-like MG (*Elavl3*), and BC-like MG (*Otx2*). (e) Heatmap showing the expression of top 50 differentially expressed genes (DEGs) of 3-week reactivated MG between CCA and *Rbpj* KO+CCA (p<0.05). (f) Heatmap showing the expression of top 50 DEGs of 4-month MGPC between *Rbpj* KO and *Rbpj* KO+CCA (p<0.05).

No significant batch effects were observed across the treatment groups (Figure 4—figure supplement 1a). Following quality control filtering and the exclusion of native mature neurons, which were uniformly distributed across all three timepoints and treatment groups, approximately 3500–12,000 cells remained for downstream snRNA sequencing analysis (Figure 4—figure supplement 1a and b). The uniform manifold approximation and projection (UMAP) visualization of snRNA-seq data revealed dynamic cell cluster transitions over time (Figure 4b and c). Clustering analysis identified eight MG-related populations: quiescent MG, reactivated MG, proliferating MG, MGPCs, transitional MG, AC-like MG, and BC-like MG, annotated using known retinal markers (Figure 4b–d, Figure 4—figure supplements 1c, d and 2a, b).

At 1 week, prior to peak transgene expression, MG in the control and Rbpj KO groups remained quiescent, expressing high levels of MG genes (Aqp4, Rlbp1, Kcnj10) (Figure 4b–d, Figure 4—figure supplements 2a–c and 3a). In contrast, CCA treatment drove cells into a ‘reactivated’ state marked by upregulation of the gliosis genes (Gfap, Vim), likely a response to AAV infection (Figure 4b–d, Figure 4—figure supplements 2a–c and 3a). At this time, no proliferating cells were observed in the CCA-injected eyes, as the AAV-mediated transgene expression had not reached the level required to drive the cell cycle.

At 3 weeks, the divergence between treatments became evident. In the CCA-treated groups, MG (6.4% in CCA and 2.2% in Rbpj KO+CCA) proliferated (Figure 4—figure supplement 2c). A subset of MG in Rbpj KO (3.4%) and Rbpj KO+CCA (4.0%) upregulated neurogenic factors (Neurog2, Ascl1, Dll1), marking them as MGPCs (Figure 4b–d, Figure 4—figure supplements 2a–c and 3a). A small fraction (<1%) of nascent AC-like cells (expressing Elavl3, Rbfox3, Caln1) and BC-like cells (expressing Gsg1, Pcdh17, Lgr5) emerged in Rbpj KO MG at this stage (Figure 4b–d, Figure 4—figure supplements 2a–c and 3a). By 4 months, while CCA-only MG had largely reverted to quiescence with minimal neurogenesis, the Rbpj KO and Rbpj KO+CCA groups sustained high percentages of MGPCs (21.5% and 23.5%, respectively) (Figure 4b–d, Figure 4—figure supplements 2b, c and 3a), confirming that Notch inhibition drives MGPC formation. Neurogenesis progressed significantly by this stage, with the Rbpj KO alone and Rbpj KO+CCA treatments yielding a higher percentage of AC-like cells (Rbpj KO, 3.2%; Rbpj KO+CCA, 2.2%) and BC-like cells (Rbpj KO, 9.9%; Rbpj KO+CCA, 18.3%) compared to the CCA-only group (Figure 4b–d, Figure 4—figure supplements 2b, c and 3a). Additionally, a unique ‘transitional’ cluster enriched for mitochondrial genes (mt-Cytb, mt-Nd4) appeared in the Rbpj KO+CCA group (Figure 4b–d, Figure 4—figure supplements 2b, c and 3a), suggesting a metabolically active state conducive to reprogramming.

To elucidate the mechanism by which Rbpj deficiency promotes neurogenic potential, we performed differential gene expression analysis on 3-week reactivated MG, a critical inflection point between quiescence and commitment to MGPC. At this time point, the CCA-treated reactivated MG cluster exhibited heightened expression of genes associated with quiescent MG (Hes5, Lgals9) (Figure 4e). The elevated expression levels of genes linked to quiescent MG may potentially guide reactivated MG back to a quiescent state (Figure 4e). In contrast, the Rbpj KO+CCA-treated reactivated MG cluster demonstrated extensive expression of neurogenic genes, including Neurog2, Zbtb16, Myrip, and Prkar2b. (Figure 4e). Neurog2 and Zbtb16 are well-established downstream genes regulated indirectly by the Notch–Rbpj pathway through the pro-neural differentiation programs (Leino et al., 2023; Oproescu et al., 2021; Henke et al., 2009). Their upregulation therefore reflects secondary effects of Rbpj loss and supports enhanced neurogenic reprogramming in the Rbpj KO+CCA group. Interestingly, in Rbpj KO+CCA-treated reactivated MG clusters, there was a significant suppression of genes associated with the interferon (IFN) pathway, including Ifi211, Ifit202b, and Mx2 (Figure 4e). As IFN signaling is known to play a crucial role in regulating the plasticity of MG (Nelson et al., 2012; Todd et al., 2016; Jorstad et al., 2020), its downregulation, coupled with the activation of neurogenic factors, likely catalyzed the formation of MGPCs.

DEG analysis of 4-month MGPC populations revealed distinct transcriptional programs between the Rbpj KO and Rbpj KO+CCA groups (Figure 4f). The Rbpj KO+CCA MGPCs upregulated genes associated with early retinal progenitor maintenance (Nebl, Eya2) and retinal development (Lrmda, Ptpre, Grik3) (Figure 4f). Conversely, the Rbpj KO MGPCs preferentially expressed gliosis-related genes, such as Prss56, Optn, Col8a1, Col22a1, and Col11a1 (Figure 4f; Minegishi et al., 2013; Hoang et al., 2020; Prieto-López et al., 2024). Given that retinal gliosis impedes MG reprogramming in mammals (Mills and Goldman, 2017), these data support the idea that combining Rbpj deletion with CCA enhances regenerative potential by boosting early progenitor programs while mitigating the gliotic responses.

CCA promotes BC subtype differentiation of the Rbpj-deficient MG

To investigate the heterogeneity of the newly generated neuron, we performed sub-clustering analysis focused specifically on the BC-like cells, which constituted a major reprogrammed neuron-like cluster. Using established BC subtype markers (Shekhar et al., 2016), we identified three distinct BC subtypes: OFF-cone BCs (identified by Pcdh17, Zfhx4, Lrrtm3, Chrm2), ON-cone BCs (identified by Isl1, Pde8a, Esrrg, Ryr3, Cdh9, Grm6), and rod BCs (identified by Isl1, Prkca, Cep112, Lrrtm4) (Figure 5a–c). Comparative analysis revealed that Rbpj deletion and CCA generated a higher abundance of these BC subtypes than either CCA or Rbpj deletion treatment alone (Figure 5d).

Figure 5 with 3 supplements see all

Download asset Open asset

CCA and *Rbpj* deletion drive BC subtype differentiation.

(a) Subclusters of the BC-like population. The BC-like MG (outlined in red) was used for subclustering analysis. (b) Feature plot of BC-like subtypes showing the cluster of OFF-cone BC (*Pcdh17*, *Zfhx4*), ON-cone BC (*Isl1*, *Grm6*), and rod BC (*Isl1*, *Prkca*, *Cep112*). (c) Dot plot showing gene expression and cell percentages for OFF-cone BC, ON-cone BC, and rod BC. (d) Percentage of BC subtypes in different treatment groups. (**e, f**) *Pcdh17* and *Grm6* mRNA in situ hybridization in the Ctrl and *Rbpj* KO+CCA groups. The white dashed boxes indicate the position of the enlarged images. (**g, h**) Representative immunostaining of EdU and PKCα on retinal sections of the Ctrl and *Rbpj* KO+CCA samples. The white dashed boxes indicate the position of the enlarged images. (i) Percentage showing GFP+ *Grm6*+ cells in overall GFP+ cells, GFP+ *Pcdh17*+ cells in overall GFP+ cells and GFP+ PKCα+ cells in overall GFP+ cells in the Ctrl and *Rbpj* KO +CCA groups. n=3 mice, *p<0.05, **p<0.01, ****p<0.0001, by unpaired two-tailed Student’s t-test. (j) Representative immunostaining of MG-derived cells with bipolar cell morphology.

To validate the snRNA-seq findings, we performed RNA in situ hybridization on Pcdh17, Grm6, and immunostaining on PKCα to identify OFF-cone BCs, ON-cone BCs, and rod BCs, respectively (Figure 5e–h). No MG in the control group co-localized with any of these BC markers. The proportions of newborn BC subtypes among MG-derived cells, as quantified by staining, were similar to the proportions observed in the snRNA-seq data (Figure 5i). Furthermore, some MG-derived neuron-like cells, which are EdU+Otx2+tdT+, exhibited morphological changes towards BC, characterized by the retraction of their apical glial processes and the adoption of BC nucleus shapes (Figure 5j). Collectively, these results indicate that Rbpj KO+CCA promoted the generation and maturation of BC subtypes formation.

To better understand the trajectory of neuronal regeneration, we performed pseudotime analysis (Figure 5—figure supplement 1a). This revealed a clear and sequential neurogenic order, progressing from resting MG to neurogenic MGPCs, and finally to BC-like and AC-like cells. This progression was closely associated with the downregulation of MG-specific genes (e.g., Hes5, Kcnj10, Slc1a2) and the upregulation of neurogenic factors (e.g., Neurog2, Dll1, Eya2) (Figure 5—figure supplement 1b–d). The Rbpj KO+CCA treatment group exhibited a more rapid shift in the gene-to-cell ratio trend line towards the BC formation branch compared to the CCA-alone and Rbpj KO-alone groups (Figure 5—figure supplement 1e), which may explain the increased diversity and abundance of BC-like cells in the Rbpj KO+CCA group.

We also investigated whether AC subtypes formed from the MG. Although this cluster could be separated into two distinct sub-clusters, they did not clearly correspond to specific AC subtypes using known markers (Yan et al., 2020), such as Chat for Starburst ACs, Dab1 for A17 ACs, and Slc6a9 for nGnG ACs (Figure 5—figure supplement 2a). Immunostaining for HuC/D confirmed the formation of AC-like cells, consistent with the snRNA-seq detection (Figure 5—figure supplement 2b and c). Quantification further confirmed that the loss of Rbpj promotes AC formation compared to the CCA-alone treatment group (Figure 5—figure supplement 2d and e).

While the neuron-like clusters were best classified as BC-like and AC-like based on their distinct marker gene expression, they also exhibited mixed expression of genes associated with other retinal neuronal types, including Tubb3, Myt1l, and Grin1, which are the signature genes of RGC, and Crx, Prom1, Epha10, Gucy2e, Scg3, which are signature genes associated with photoreceptor lineages (Figure 5—figure supplement 3a), suggesting that the regenerated cells exist in a hybrid state. Notably, we did not detect Nrl expression in the newborn neuron clusters, which is essential for rod photoreceptor specification (Mears et al., 2001; Figure 5—figure supplement 3a). The discrepancy in Nrl detection between snRNA-seq and prior immunofluorescence staining suggests that Nrl+ cell generation is stochastic and occurs at low frequency (Figure 3—figure supplement 4b and c, Figure 5—figure supplement 3a). Moreover, relative to the Rbpj KO group, neuron-like clusters in the Rbpj KO+CCA group showed upregulated expression of genes typically enriched in both RGCs and photoreceptors, such as Tmem132d, Cntn5, Ryr2, Mef2c, Meis2, and Cngb1 (Figure 5—figure supplement 3b).

Together, these mixed lineage signatures indicate that the newly generated neurons remain in a plastic, incompletely committed state and likely require additional cues, such as Nrl, to achieve full maturation and terminal specification. Moreover, combining Rbpj deficiency with CCA treatment enhances the acquisition of abundant and distinct neuronal characteristics in the newborn neurons.

Increased chromatin accessibility of neurogenic factors underlies CCA-induced neuronal formation from MGPCs

We next sought to elucidate how CCA promotes neurogenesis. We hypothesized that CCA-induced MG proliferation remodels the chromatin landscape to facilitate reprogramming, a mechanism analogous to those observed in somatic cell reprogramming and cardiac regeneration (Monroe et al., 2019; Papp and Plath, 2013). To test this hypothesis, we profiled chromatin accessibility using snATAC-seq on MG nuclei purified from four groups at 4 months post-treatment, including the untreated control, CCA, Rbpj KO, and the combined Rbpj KO+CCA group (Figure 6a).

Figure 6 with 1 supplement see all

Download asset Open asset

snATAC-seq analysis suggests increased chromatin accessibility of the neurogenic genes by CCA.

(a) Schematic illustration of the snATAC-seq experiment. (b) Proportions of cell clusters within Ctrl, CCA, *Rbpj* KO and *Rbpj* KO+CCA groups. (c) UMAP plot of snATAC-seq data from four treatment groups (Ctrl, CCA, *Rbpj* KO, *Rbpj* KO+CCA). (d) Feature plots highlighting the cluster of quiescent MG (*Kcnj10*), KO MG (*Hes5*), active MG (*Bal3*), MGPC (*Dll1*), AC-like MG (*Caln1*), BC-like MG (*Otx2*). (e) Dot plot showing gene activity and cell percentages for quiescent MG, KO MG, active MG, MGPC, AC-like MG, BC-like MG. (f) Volcano plot showing the differential gene activity of MG and active MG within CCA treatment group (p<0.05). Red dots indicate the genes upregulated in active MG and blue dots indicate genes downregulated in active MG. (g) WikiPathways enrichment analysis of upregulated genes in active MG (p<0.05). (h) Increased chromatin accessibility at the Neurod2 locus was observed in the active MG. The black arrow shows the transcription direction.

No batch effects were detected across treatments, and clusters were annotated using canonical cell type-specific markers (Figure 6—figure supplement 1a–c). After quality control and removal of native mature neurons, which uniformly represented across groups, approximately 8000–11,000 cells remained for downstream analysis (Figure 6b and c). Clustering resolved six cell types based on marker accessibility: resting MG, active MG, Rbpj KO MG (KO MG), MGPC, AC-like, and BC-like, with distinct compositional shifts across treatments (Figure 6b and c). As expected, MG in the control group exhibited a quiescent chromatin state, with high accessibility at classic MG marker genes (e.g., Kcnj10, Rlbp1, Aqp4) (Figure 6d and e). A unique active MG cluster, showing markedly increased accessibility at loci involved in cell cycle regulation (e.g., Firre, Cdc45, Bcl3), was characterized in the CCA-treated group (Figure 6d and e). In Rbpj-deficient MG, accessibility at Notch pathway effectors (Hes1, Hes5, Heyl) was markedly reduced relative to resting MG (Figure 6d and e). Notably, the Rbpj KO+CCA combination produced more pronounced chromatin changes associated with MGPC, AC-like, and BC-like states compared to Rbpj KO alone, underscoring the contribution of CCA to chromatin remodeling (Figure 6d and e).

A direct comparison of resting MG versus active MG within the CCA treatment group revealed broad increases in chromatin accessibility (Figure 6f), with upregulated loci linked to retinal development and neurogenesis, including Neurod2, Dll1, and Otx2 (Figure 6f). WikiPathways enrichment indicated strong associations with dopaminergic neurogenesis (Figure 6g). Genome browser tracks further showed elevated accessibility across the Neurod2 promoter and gene body in active MG relative to resting MG, consistent with the activation of this key neurogenic regulator (Figure 6h).

In summary, our snATAC-seq data support a model wherein CCA-induced proliferation remodels the chromatin landscape of MG, priming them for reprogramming by opening key neurogenic loci.

MG-derived daughter cells exhibit long-term survival

To evaluate long-term outcomes of MG reprogramming, a cohort of mice treated with Rbpj KO and CCA were aged to 9 months post-treatment (Figure 7a). At this late time point, MG-derived cells, marked by the tdT reporter, displayed a more regular, circular nuclear morphology reminiscent of mature retinal neurons, though fully mature rod photoreceptors were not detected (Figure 7b). Quantification showed an upward trend in the proportion of Otx2+ tdT+ cells among total tdT+ cells, consistent with a gradual, ongoing MG-to-neuron reprogramming process (Figure 7b and c). However, both the percentage of Otx2+ EdU+ tdT+ cells within the EdU +tdT+ population and the absolute number of EdU + cells declined slightly over time (Figure 7d and e). This data suggests that while most MG-derived daughter cells persist long-term, a small subset of proliferated MG-derived neurons may ultimately undergo apoptosis.

Figure 7 with 2 supplements see all

Download asset Open asset

Neurons derived from MG exhibited long-term survival capabilities.

(a) Schematic illustration of the experiment assessing long-term cell survival. (b) Representative immunostaining of EdU and Otx2 on retinal sections. The white arrows refer to tdT+Otx2+ cells. (c) Percentage of tdT+ Otx2+ cells in overall tdT+ cells, n=3 mice, data are presented as mean ± SEM. ns = not significant, **p < 0.01, ***p<0.001, by one-way ANOVA with Tukey’s post hoc test. (d) Percentage of tdT+EdU+ Otx2+ cells in overall tdT+EdU+ cells, ns = not significant, **p < 0.01, by one-way ANOVA with Tukey’s post hoc test. (e) Number of EdU+ cells per 500 µm, ns = not significant, by one-way ANOVA with Tukey’s post hoc test.

To evaluate retinal structure following treatment, we performed ZO1 staining to visualize the outer limiting membrane (OLM) (Figure 7—figure supplement 1a). The OLM integrity in the Rbpj KO+CCA group at 9 months was comparable to that of untreated controls, suggesting that sufficient MG persist to support junctional complexes (Figure 7—figure supplement 1b). Next, to assess whether the combined Rbpj KO and CCA treatment compromises retinal function, we performed structural and functional analyses at 4 months post-treatment (Figure 7—figure supplement 2a). Optical coherence tomography (OCT) revealed that retinal layer organization and ONL thickness were comparable among the uninjected eyes, GFP AAV-injected control eyes, and CCA-treated eyes, demonstrating that overall retinal architecture was well preserved (Figure 7—figure supplement 2b and c). Consistently, optomotor response testing revealed no significant differences in visual acuity across groups, suggesting that visual function remained intact (Figure 7—figure supplement 2d). Furthermore, electroretinography (ERG) demonstrated that scotopic and photopic a- and b-wave amplitudes were unaffected by the treatment, confirming that light responses from photoreceptor and inner retinal neurons were preserved (Figure 7—figure supplement 2e–h). Taken together, these findings demonstrate that combined Rbpj KO and CCA treatment achieves neurogenesis without compromising retinal structure and functional visual circuitry.

In summary, our findings demonstrate that the majority of MG-derived cells possess inherent long-term stability, and that maintaining an adequate MG population is a prerequisite for achieving stable and functional regeneration.

Discussion

Retinal degeneration is characterized by the progressive loss of retinal neurons, leading to irreversible vision impairment. In teleost fish, retinal injury triggers functional regeneration, with MG serving as the source of regenerated neurons. In contrast, adult mammalian retinas lack regenerative capacity because MG fail to re-enter the cell cycle, dedifferentiate, and subsequently differentiate into neurons. In our previous study, we showed that simultaneous knockdown of Cdkn1b and overexpression of Ccnd1 effectively drives mouse MG to re-enter the cell cycle in the absence of injury. However, most postmitotic MG underwent only transient dedifferentiation, and only a small fraction (~1%) differentiated into neuron-like cells (Wu et al., 2025). In the present study, scRNA-seq and RNA in situ hybridization revealed persistent activation of Notch signaling in postmitotic MG, which restricts their reprogramming potential. Conditional deletion of Rbpj, the core transcriptional effector of Notch signaling, in late-stage RPCs promoted ectopic rod photoreceptor formation at the expense of MG differentiation, whereas Rbpj deletion in mature MG alone produced only limited neurogenic effects, suggesting that epigenetic landscape changes may contribute to more limited reprogramming potential from late RPC-to-MG transition. Notably, combining Rbpj knockout with the proliferative stimulus CCA treatment markedly enhanced neurogenesis from postmitotic MG. Immunofluorescence showed that ~26% of MG daughter cells expressed neuronal marker Otx2 and HuC/D. snRNA-seq confirmed these results and revealed a clear trajectory of MG-derived neurogenesis, in which cells transitioned from a glial state toward BC-like and AC-like identities. Subtype analysis identified three BC subtypes: OFF-cone BCs, ON-cone BCs, and rod BCs, among MG-derived BC-like cells, whereas AC-like cells could not be further classified. MG-derived neuron-like cells also expressed genes characteristic of RGCs and photoreceptors, indicating enhanced lineage promiscuity. This hybrid transcriptomic profile, together with the immature morphology of newborn neurons, suggests incomplete or inefficient reprogramming. Mechanistically, snATAC-seq revealed that CCA treatment increases chromatin accessibility at pro-neurogenic loci following MG proliferation, likely facilitating the improved reprogramming efficiency of Rbpj^-/- MG. Notably, the majority of reprogrammed neurons survived for up to 9 months, while the MG pool remained intact, underscoring their preserved function to support retinal homeostasis.

MG are essential for maintaining retinal structural and functional integrity. They contribute to the blood-retinal barrier, regulate extracellular ion and neurotransmitter homeostasis, provide metabolic support to neurons, and mediate inflammatory responses (Reichenbach and Bringmann, 2013; Bringmann et al., 2006). Thus, preserving MG number and function during reprogramming interventions is critical, as extensive MG loss or dysfunction of MG could disrupt retinal homeostasis and exacerbate degeneration (Goldman, 2014). In zebrafish, MG regenerate neurons through asymmetric division, generating one self-renewing MG and one progenitor cell, thereby maintaining the glial pool (Lenkowski and Raymond, 2014; Gorsuch and Hyde, 2014). In our study, combined CCA treatment and Rbpj deletion effectively stimulated MG proliferation and reprogramming in the adult mouse retina without overt MG depletion. Because both daughter cells lacked Rbpj, a classic asymmetric division with one daughter having high Notch signaling and the other low is unlikely. Nonetheless, our strategy successfully regenerated neuron-like cells while maintaining the MG pool, as more MG were generated to replace the reprogrammed MG. The long-term survival of newborn neurons and the stability of retinal architecture and function support the integrity of the MG pool. Although our strategy differs from zebrafish MG self-renewal, it still offers a viable regenerative approach that preserves glial homeostasis, a critical prerequisite for maintaining retinal homeostasis. Achieving a balance between neurogenesis and MG preservation represents a central challenge in glial-based regeneration. Excessive or widespread reprogramming risks depleting MG below a critical threshold or impairing their supportive functions. Future strategies should emphasize transient, partial, or spatially confined reprogramming to maintain a functional MG pool. For instance, replication-incompetent retroviral vectors (Noctor et al., 2001; McMahon and McDermott, 2006) to knockout Rbpj in one daughter cell of a divided MG could ensure that the sibling cell remains a supportive glial cell, thereby maintaining retinal homeostasis while generating new neurons.

Multiple studies have explored the proliferative potential of MG in adult mice, including Wnt pathway activation, Hippo pathway inhibition, and forced manipulations of cell cycle regulators (Wu et al., 2025; Yao et al., 2016; Rueda et al., 2019). However, unlike zebrafish, proliferating mouse MG rarely give rise to mature retinal neurons, regardless of the proliferation-inducing method. Our time-course snRNA-seq analysis of CCA-treated retinas confirmed this limitation: MG re-entered and exited the cell cycle but largely reverted to a quiescent glial state. What is the functional role of MG proliferation in retinal regeneration? Our findings imply that proliferation actively facilitates reprogramming rather than merely expanding MG numbers. Specifically, we demonstrate that CCA-induced proliferation triggers widespread chromatin opening at key neurogenic gene loci, including Neurod2, Dll1, and Otx2, in active MG compared to resting MG (Figure 6f–h). This principle parallels findings in somatic cell reprogramming, where DNA replication and cell division facilitate the erasure of lineage-restrictive epigenetic marks to enhance reprogramming efficiency (Papp and Plath, 2013; Ruiz et al., 2011; Hanna et al., 2009; Wolffe, 1991). Similarly, proliferative neural stem cells (NSCs) exhibit higher chromatin accessibility at pro-neural genes Otx2 and Ascl1, compared with aged NSCs that show reduced proliferative capacity (Yeo et al., 2023; Li et al., 2024). Enhancing NSCs proliferation with betacellulin upregulated Ascl1 and Nestin, thereby boosting neurogenesis in the olfactory bulb and dentate gyrus (Gómez-Gaviro et al., 2012). Crucially, while this replication-coupled chromatin remodeling is insufficient to drive neurogenesis, it creates a permissive, epigenetically reset window. When combined with Rbpj deletion, which relieves Notch-mediated repression of downstream neurogenic factors such as Ascl1, Neurog2, Otx2, and Neurod2, to enable their effective occupancy and activation of the newly accessible chromatin. Taken together, our findings provide strong evidence that cell-division-mediated epigenetic resetting is conserved in MG and is essential for overcoming the barriers to retinal neurogenesis.

The Notch signaling pathway plays a pivotal role in regulating both cell proliferation and cell fate specification throughout retinal development (Jadhav et al., 2006b; Yaron et al., 2006). During development, Notch-mediated lateral inhibition ensures that diverse retinal cell types are generated in a spatially and temporally coordinated manner. Deletion of Notch1 or downstream effectors such as Hes1 and Rbpj disrupts this balance, causing premature cell-cycle exit of RPCs and increased production of early-born neurons, particularly photoreceptors, at the expense of later-born cell types, including MG (Jadhav et al., 2006b; Riesenberg et al., 2009). Conversely, forced expression of Hes1 in progenitor cells drives gliogenesis and promotes MG differentiation (Furukawa et al., 2000). Aligning with this paradigm, Rbpj deletion during retinal development redirected the fate of late-stage progenitor cells, producing more mature rod photoreceptors at the expense of MG and BCs. In the mature retina, sustained Notch signaling maintains MG quiescence (Mills and Goldman, 2017; Campbell et al., 2022). Its conserved role across species has been well documented: inhibition of the Delta-Notch3-Hey1/Id2b pathway triggers zebrafish MG proliferation and regeneration (Campbell et al., 2021; Fogerty et al., 2022; Sahu et al., 2021). In the post-hatch chick retina, Notch inhibition using a γ-secretase inhibitor likewise enhances retinal regeneration (Hayes et al., 2007). In adult mice, simultaneous suppression of Notch signaling and deletion of Nfia/b/x synergistically promotes MG conversion into BCs and ACs (Le et al., 2024). Similarly, virus-mediated Pou5f1 overexpression combined with Notch inhibition boosts the neurogenic competence of mammalian MG (Le et al., 2025). Collectively, these findings position Notch inhibition as a central mechanism for unlocking MG plasticity. Together, these studies highlight Notch inhibition as a critical gatekeeper of MG plasticity. However, both our work and others indicate that Notch inhibition primarily generates BC- and AC-like neurons, with limited evidence for mature photoreceptor formation. This underscores an enduring barrier to producing the full repertoire of retinal neurons and suggests that additional lineage determinants, such as Nrl, may be required to guide MG toward photoreceptor fates.

Furthermore, we acknowledge that the precise morphological features, synaptic connectivity, and electrophysiological properties of MG-derived cells remain largely uncharacterized in the current study. As demonstrated by both our immunostaining and snRNA-seq data, the MG-derived neuron-like cells exhibit an incompletely mature transcriptomic profile with hybrid lineage signatures. While we observed initial morphological changes toward a BC identity, including retraction of apical glial processes and adoption of BC-like nuclear morphology (Figure 5j), these cells lacked the fully elaborated dendritic and axonal arbors characteristic of mature BC. Moreover, we did not observe any MG-derived cells with photoreceptor outer segment or typical RGC morphology. Consequently, it is highly probable that functional synaptic connectivity and mature electrophysiological properties have not yet been established at the current stage of reprogramming. Future studies incorporating synaptic marker analysis, morphological reconstruction, and electrophysiological recordings will be essential to determine whether further optimization of our reprogramming strategy can drive these MG-derived neurons toward full functional maturation and successful circuit integration.

Regenerating retinal neurons from endogenous MG represents a promising frontier in regenerative medicine, yet the long-term survival of newly generated neurons is essential for clinical translation. Normally, mature retinal neurons depend on intrinsic survival mechanisms and extrinsic trophic support from the microenvironment, including trophic factors, such as brain-derived neurotrophic factor (BDNF), ciliary neurotrophic factor (CNTF), glial cell-line derived neurotrophic factor (GDNF), and nerve growth factor (NGF), synaptic integration for functional activity-dependent survival, and metabolic support from the RPE and vasculature (Kimura et al., 2016; Kolomeyer and Zarbin, 2014). In contrast, transplanted neurons often suffer low survival rates due to poor migration, inadequate support, limited synaptic integration, and immune rejection (Ahmed et al., 2022; Oswald et al., 2021; Soucy et al., 2023). For example, xenotransplantation of mouse induced pluripotent stem cell/mouse embryonic stem cell-derived RGCs exhibits only minimal medium-term survival in the host retina owing to immune barriers (Ahmed et al., 2022; Oswald et al., 2021; Soucy et al., 2023). While the survival of endogenous MG-derived neurons has been underexplored, our study demonstrates a remarkably high long-term survival rate, with ~80% of postmitotic MG-derived newborn neurons persisting for at least 9 months following combined Rbpj deletion and CCA treatment. We speculate that it may involve successful apical migration to their appropriate retinal layers (e.g., ONL and INL), where they can access layer-specific trophic support and integrate into local microenvironments. Their laminar positioning likely facilitates contact with presynaptic and postsynaptic partners, as well as access to survival signals from MG and photoreceptor cells, which are critical for preventing apoptosis. To further enhance the survival of newborn neurons from MG, future strategies should focus on recapitulating the endogenous survival mechanisms of the retina. This could involve co-expression of anti-apoptotic genes such as Bcl-2 or Xiap (Zadro-Lamoureux et al., 2009; Donovan et al., 2006) and supplemental trophic support (e.g., BDNF, CNTF), alongside synaptic integration through molecular guidance cues such as Semaphorins or Netrins (Dent et al., 2004; Alto and Terman, 2017). Regenerating neurons within a degenerating retinal environment, where vacated synaptic positions may facilitate new connections, may further enhance neuronal integration. The efforts to enhance the survival of regenerated neurons will be essential for advancing regenerative therapies.

Beyond survival, the appropriate laminar positioning of MG-derived neurons remains another challenge for retinal in vivo reprogramming. During retinal development, precise laminar positioning of neurons is guided by a coordinated interplay of cell-intrinsic transcriptional programs and extrinsic cues, including cell adhesion molecules, guidance factors, and interactions with neighboring cells. In adult retina, many of these developmental cues are no longer present or active, which likely contributes to the failure of MG-derived neurons to migrate to their appropriate laminar positions. Interestingly, the vast majority of our divided MG cells remained localized within the ONL. Because the ONL is the physiological niche of photoreceptors, this preferential position could serve as an advantageous baseline layout for driving targeted photoreceptor differentiation in future work. In summary, our study establishes that combined Rbpj deletion and CCA-induced proliferation efficiently reprograms adult mouse MG into diverse retinal neurons that exhibit remarkable long-term survival, providing a robust platform for dissecting the molecular logic of mammalian retinal regeneration. Building upon this framework, future efforts to refine subtype specification, guide laminar positioning, and enhance synaptic integration will be critical to translating endogenous MG reprogramming into a viable therapeutic strategy for restoring vision in retinal degenerative diseases.

Materials and methods

Animals

Rbpj^flox/flox mice (strain: Rbpj^em2Lutzy) (JAX number: #034200), tdTomato reporter mice (strain B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J) (JAX number: #007914) (Madisen et al., 2010), Sun1-GFP reporter mice (strain B6.129-Gt(ROSA)26Sor^{tm5.1(CAG-Sun1/sfGFP)Nat}/MmbeJ) (JAX number: #030952) (Mo et al., 2015), and Glast^CreERT reporter mice (strain Tg(Slc1a3-cre/ERT)1Nat/J) (JAX number: #012586) (Wang et al., 2012) mice were purchased from the Jackson Laboratory. All mice were kept on a 12/12 hour light/dark cycle in the Laboratory Animal Research Unit, City University of Hong Kong. All animal procedures performed were approved by the Hong Kong Department of Health under Animals Ordinance Chapter 340 (Ref: 20-130 in DH/HT&A/8/2/5 Pt.2) and by the City University of Hong Kong Animal Ethics Committee (Ref: A-0264). Mice were randomly assigned to experimental and control groups without sex restriction. All mice that received successful intravitreal injections were included in experiments and subsequent statistical analyses.

AAV production

Request a detailed protocol

AAV production was carried out following the previously described procedure (Hoang et al., 2023). In brief, to produce the recombinant AAV_7M8-GFAP-Ccnd1-Cdkn1b shRNA-WPRE and AAV_7m8-GFAP-GFP-WPRE, the HEK293T cells were transfected with a mixture of pAAV vector transgene plasmid, rep/cap packaging plasmid, and adenoviral helper plasmid. At 96 hours post-transfection, both the culture medium and the transfected cells were harvested for AAV collection. The collected AAV underwent further purification through ultra-centrifugation in the iodixanol (OptiPrep) gradient at 147,000 × g at 4°C for 90 minutes. The iodixanol in the AAV solution was washed three times with PBS using Amicon 100K columns (EMD Millipore), and about 30 µL of the final volume AAV was collected for downstream applications. The virus titration was determined through the protein SDS-PAGE method.

Tamoxifen injection

Request a detailed protocol

To induce Cre recombinase expression in the majority of MG, tamoxifen (Sigma, dissolved in corn oil) was administered via intraperitoneal injection at a dosage of 50 mg/kg.

Intravitreal AAV injection

Request a detailed protocol

To stimulate MG proliferation, mice were intravitreally injected with AAV_7m8-GFAP-Ccnd1-Cdkn1b shRNA-WPRE and AAV_7m8-GFAP-GFP-WPRE at a final concentration of 4E13 vg/mL. Briefly, the eyelid was gently manipulated with tweezers to expose the eyeball. Subsequently, 1 μL of the virus was precisely introduced into the intravitreal space using a custom angled glass pipette controlled by a FemtoJet (Eppendorf). The treatment was administered to the right eye of the animal, while the left eye remained untreated as a control.

Retinal cryosection and immunohistochemistry

Request a detailed protocol

The mice were humanely euthanized using CO₂ and cervical dislocation. Before enucleation, the eyeballs were marked ventrally to indicate the AAV injection site, and then the retinas were carefully dissected in PBS and fixed in 4% paraformaldehyde (PFA) at room temperature for 30 minutes. The fixed retinas underwent three-times washes with PBS (Product #10010023, Thermo Fisher Scientific) and were subsequently dehydrated in sequential sucrose solutions of 5%, 15%, and 30% for 15, 30, and 60 minutes, respectively. The retina was then immersed in a solution of optimal cutting temperature (OCT) and 30% sucrose in a 1:1 ratio at 4°C overnight before being embedded in cryomolds in a specific orientation for sectioning.

After freezing the tissue below –20°C, a series of 20 μm sections were cut and mounted on glass slides using a cryostat machine (Thermo HM525NX Cryostat). For immunostaining, retinal sections were first blocked in 3% bovine serum albumin (BSA, #A9647, Sigma-Aldrich) and 0.1% Triton X-100 in PBS (PBST) for 30 minutes at room temperature, followed by overnight incubation with primary antibodies at the recommended dilution at 4°C. Primary antibodies used in this study included goat anti-Otx2 antibody (1:200, AF 1979; R&D Systems), mouse anti-HuC/D (1:200, A21271; Thermo Fisher Scientific), rabbit anti-Sox9 antibody (1:500, AB5535; Millipore), rabbit anti-mCAR (1:200, AB15282, Millipore), rabbit anti-Nrl (1:150, AF2945, R&D Systems), mouse anti-PKCα (1:200, sc-8393, Santa Cruz), rabbit anti-Crx (1:500, PA5-32182, Thermo Fisher Scientific), goat anti-Sox2 (1:500, AF2018, R&D Systems), rabbit anti-Pax6 (1:500, AB2237, Millipore), rabbit anti-ZO1 (1:500, 61-7300, Thermo Fisher Scientific), and rabbit anti-Rbpms (1:500, ab152101, Abcam). Following primary antibody incubation, the samples were washed thrice with PBST before incubation with a mixture of DAPI (0.5 μg/mL) and secondary antibodies in the dark for 2 hours at room temperature.

Subsequently, the retinal sections were washed and mounted with an anti-fade solution before microscopy or storage. Slide images were captured using Nikon A1HD25 High Speed and Large Field of View Confocal Microscope. Histological measurements and image processing were conducted using ImageJ software. The percentage of MG expressing various marker genes was quantified within retinal regions exhibiting near-complete AAV transduction.

EdU incorporation and detection

Request a detailed protocol

5’-ethynyl-2’-deoxyuridine (EdU, 50 mg/kg, Abcam ab146186) was intraperitoneally injected for 21 days after CCA administration to label the cells in the S phase. EdU staining was performed following the instructions of the Click-iT EdU Alexa Fluor 488 or 647 Imaging Kit (C10337, Thermo Fisher Scientific).

In situ RNA hybridization

Request a detailed protocol

In the study, in situ RNA hybridization was conducted utilizing the RNAscope Multiplex Fluorescent Detection Reagents V2 kit (Advanced Cell Diagnostics) following standard commercial procedures. Initially, retinas were carefully dissected, fixed in 4% PFA, dehydrated using a sucrose solution, and embedded in OCT medium. Subsequently, the retinas were sectioned into 20 μm slices and placed on SuperFrost Plus glass slides (Epredia).

After removing the OCT with PBS and dehydrating further with 50%, 70%, and 100% ethanol, the retinal sections were incubated with a GFP antibody (AB_2307313; Aves Labs) overnight at 4°C. Post a triple wash with PBST (PBS with 0.1% Tween-20), the sections were hybridized with RNA probes: Hes1 probe (Cat# 417701-C2 RNAscope Probe-Mm-Hes1-C2), Pcdh17 probe (Cat# 489901-C2 RNAscope Probe-Mm-Pcdh17-C2), and Grm6 probe (Cat# 511611 RNAscope Probe-Mm-Grm6) for 2 hours at 40°C. Following the RNA hybridization process, the slides were stained with secondary antibodies (Jackson ImmunoResearch) and DAPI for 2 hours at room temperature.

The resulting fluorescent signals were observed and captured using a Nikon A1HD25 High-Speed and Large Field of View Confocal Microscope. The mRNA levels within the GFP-labeled nuclei membrane were then quantified by measuring the signal intensity level using ImageJ.

MG sorting and snRNA library preparation and sequencing

Request a detailed protocol

Three or four fresh retinas were dissected from adult mice (strain: Glast^CreERT;Rbpj^flox/flox;Rosa26^Sun1-GFP and Glast^CreERT;Rosa26^Sun1-GFP, aged as specified in ‘Results’) under RNase-free conditions in cold PBS on ice to minimize RNA degradation. Dissection tools (forceps, scissors) were sterilized with 70% ethanol and RNase Zap (Product #AM9780, Thermo Fisher Scientific) to eliminate RNase contamination. Retinas were frozen in liquid nitrogen until use. Retinal tissues were homogenized in 500 µL of ice-cold NP-40 lysis buffer (Product #74385, Sigma-Aldrich) supplemented with 1×protease inhibitor cocktail (Product #P8340, Sigma-Aldrich) and 1 U/µL RNase inhibitor (Product #EO0381, Thermo Fisher Scientific) using a pellet pestle (AST-YMB-15, Axyste). Homogenization was performed on ice with 15–20 gentle strokes to avoid excessive shearing of nuclei, followed by incubation on ice for 5 minutes to ensure complete lysis of cytoplasmic membranes while preserving nuclear integrity. The homogenate was filtered through a 70 µm MACS SmartStrainer (Cat # 130-110-915, Miltenyi Biotec) pre-rinsed with cold PBS to remove cell debris and intact cells. The filtrate (containing nuclei) was centrifuged at 500×g for 5 minutes at 4°C to pellet nuclei, and the supernatant was discarded. The nuclear pellet was resuspended in 1 mL of 1% BSA in PBS (1% BSA-PBS), supplemented with 1 U/µL RNase inhibitor, to neutralize residual lysis buffer and stabilize nuclei. To enrich for MG nuclei (a rare population comprising 2–3% of total retinal cells), fluorescence-activated cell sorting (FACS) was performed using a Sony SH800Z cell sorter equipped with a 488 nm laser for GFP excitation and a 525/50 nm emission filter. Nuclei were gated based on forward scatter (FSC) and side scatter (SSC) to exclude debris and aggregates, and GFP-positive events were sorted into RNase-free 1.5 mL tubes containing 100 µL of 1% BSA-PBS with RNase inhibitor. Sorting was performed at 4°C with a flow rate of 1000–2000 events/second. Approximately 200,000 GFP+ nuclei were collected per sample and centrifuged at 500×g for 5 minutes at 4°C. The pellet was resuspended in 1% BSA-PBS (with RNase inhibitor) and adjusted to a concentration of 600–800 nuclei/µL using a Countess II FL Automated Cell Counter (Thermo Fisher Scientific), ensuring compatibility with 10x Genomics Chromium systems for optimal Gel Bead-in-Emulsion (GEM) formation.

snRNA sequencing libraries were prepared using the Chromium Next GEM Single Cell 3' GEM, Library & Gel Bead Kit v3.1 (Cat# 1000269, 10x Genomics), Chromium Next GEM Chip G Single Cell Kit (Cat# 1000127, 10x Genomics), and Chromium Controller iX. Briefly, 20 µL of the nuclei suspension (containing ~18,000 nuclei) was loaded into a well of the Chromium Next GEM Chip G, along with reverse transcription (RT) reagents, gel beads, and partitioning oil. GEMs were generated using the Chromium Controller iX, with each GEM containing a single nucleus, a gel bead (with barcoded oligonucleotides), and RT reagents. GEMs were transferred to 0.2 mL 8-tube strips (951010022, Eppendorf) and subjected to RT in a ProFlex PCR System (4484073, Thermo Fisher Scientific) with the following program: 53°C for 45 minutes, 85°C for 5 minutes, and hold at 4°C. Resulting cDNA was amplified using 11 cycles of PCR (98°C for 3 minutes; 12 cycles of 98°C for 15 seconds, 63°C for 20 seconds, 72°C for 1 minute; final extension at 72°C for 1 minutes) to generate sufficient material for library construction. Amplified cDNA was purified using SPRIselect beads (Beckman Coulter) at a 1:1.8 ratio (cDNA:beads) to remove primers and impurities. Purified cDNA was fragmented enzymatically (32°C for 5 minutes) using the Fragmentation Enzyme Mix (included in the 10x kit), followed by end repair, A-tailing, and ligation of Illumina-compatible adapters. Final libraries were purified with SPRIselect beads (1:1 ratio) and quantified using a Qubit 4 Fluorometer (Thermo Fisher Scientific).

Libraries were sequenced by Novogene (Beijing) on an Illumina NovaSeq 6000 platform using paired-end 150 bp reads. Each sample was allocated 100 Gb of sequencing data, ensuring an average depth of ~35,000 reads per nucleus to capture robust transcriptomic profiles, including low-abundance transcripts relevant to MG function and differentiation. This expanded method provides detailed, reproducible parameters for each step, including reagent supplements, centrifugation conditions, FACS gating, and sequencing metrics, to support rigorous replication of the experiment.

Preprocessing, filtering, and clustering of snRNA data

Request a detailed protocol

The raw sequencing reads (paired-end 150 bp) were processed using Cell Ranger software (v6.1.2, 10x Genomics) following the manufacturer’s recommended pipeline (here). For each sample, count was used to align reads to the mouse reference genome (mm10/GRCm38) and quantify gene expression. Cell Ranger’s built-in quality control metrics were used to generate filtered matrices, excluding potential cell-free RNA or debris-associated barcodes.

Subsequently, quality control, filtering, dimensional reduction, and clustering of the data were performed utilizing the Seurat package in R (Stuart et al., 2019). Doublets and cells with less than 200 or more than 6000 expressed features and a percentage of mitochondrial transcripts exceeding 20% were excluded from further analysis. For the remaining cells, UMAP dimension reduction based on eight principal components (PCs) was implemented, and cells were clustered using the graphical clustering method within Seurat. Cell types were identified by leveraging known marker genes.

snATAC library preparation and sequencing

Request a detailed protocol

Retinal MG nuclei were isolated as described in the snRNA-seq library preparation section. To permeabilize nuclear membranes for chromatin accessibility, isolated nuclei were treated with lysis buffer containing 5% digitonin (Thermo Fisher Scientific) for 1 minute on ice, then immediately washed and resuspended in diluted Nuclei Buffer (prepared by diluting 20X Nuclei Buffer, PN-2000207, 1:20 in nuclease-free water) to a final concentration of approximately 4000 nuclei/µL.

snATAC-seq libraries were generated using the Chromium Next GEM Single Cell ATAC Kit v2 (PN-1000406, 10x Genomics) following the manufacturer’s protocol (CG000496 Rev C). In brief, permeabilized nuclei were incubated with Transposition Mix (ATAC Buffer B and ATAC Enzyme B) at 37°C for 30 minutes. Transposed nuclei were mixed with Master Mix (Barcoding Reagent B, Reducing Agent B, Barcoding Enzyme) and loaded onto a Chromium Chip H with ATAC Gel Beads v2 and Partitioning Oil to form GEMs using the Chromium Controller iX. GEMs were thermocycled (72°C 5 minutes; 98°C 30 seconds; 12 cycles of 98°C 10 seconds, 59°C 30 seconds, 72°C 1 minute; hold 15°C). Post-GEM cleanup used Dynabeads MyOne SILANE followed by double-sided SPRIselect purification. Indexed libraries were PCR-amplified for seven cycles and purified again with SPRIselect. Library quality was assessed using an Agilent 2100 Bioanalyzer and sequenced by Novogene on an Illumina NovaSeq 6000 (PE150), with 120  Gb generated per sample.

Preprocessing, filtering, and clustering of snATAC-seq data

Request a detailed protocol

The raw sequencing reads (paired-end 150 bp) were processed using Cell Ranger ATAC software (10x Genomics) following the manufacturer’s recommended pipeline. The workflow included: Illumina BCL files were converted to FASTQ format using cellranger-atac mkfastq, which assigned reads to individual samples based on index barcodes. For each sample, cellranger-atac count was used to align reads to the mouse reference genome (mm10/GRCm38) and generate fragment files quantifying chromatin accessibility.

Subsequently, quality control, filtering, dimensional reduction, and clustering of the data were performed utilizing the SnapATAC2 package in Python (Fang et al., 2021; Zhang et al., 2024). Fragment files were imported and processed with sample-specific quality control parameters. Cells were filtered based on the following criteria: minimum fragment counts of 9000, minimum transcription start site enrichment (TSS enrichment) score of 6, and maximum fragment counts of 90,000 to exclude low-quality nuclei and potential doublets. A tile matrix with 500 bp bins was generated, and the top 250,000 most variable features were selected for downstream analysis. Doublet detection was performed using the Scrublet algorithm implemented in SnapATAC2, and predicted doublets were removed from further analysis.

For dimensional reduction, spectral embedding (similar to Latent Semantic Indexing) was performed on the filtered cells, followed by UMAP for visualization. Cell clustering was performed using the Leiden algorithm after constructing a k-nearest neighbor graph. Cell types were identified by integrating gene activity scores derived from chromatin accessibility profiles with known marker genes and comparison to snRNA-seq data from matched samples.

Electroretinography (ERG)

Request a detailed protocol

Mouse retinal function was assessed by ERG using the Espion E3 System (Diagnosys LLC), with a protocol adapted from previous studies on wild-type mice to characterize rod and cone responses (Wu et al., 2025). Mice were dark-adapted overnight, anesthetized with a ketamine/xylazine mixture (100/10 mg/kg), and their pupils were dilated with 5% phenylephrine (Mydrin-P, Santen) and 0.5% tropicamide for 5 minutes. Corneas were kept hydrated with gel. Gold-wire electrodes were placed on each cornea, with reference and ground electrodes in the mouth and tail, respectively. All procedures were performed under dim red light. Scotopic responses were elicited by 530 nm flashes of increasing intensity (0.01–30 cd·s/m²). For photopic recordings, mice were light-adapted for 5 minutes at 10 cd·s/m² to suppress rod activity, followed by flashes of 30 cd·s/m² on the same background. The amplitudes and implicit times of a- and b-waves were recorded for analysis.

Optical coherence tomography (OCT)

Request a detailed protocol

In vivo OCT was performed using a Bioptigen SD-OCT system (Envisu R4310, Germany). Mice were anesthetized by intraperitoneal injection of ketamine/xylazine (100/10 mg/kg) and kept warm on a heating pad. Corneas and pupils were treated with 0.5% proxymetacaine (Provain-POS) and a mixture of 0.5% tropicamide and 0.5% phenylephrine (Mydrin-P, Santen), and hydrated with lubricating drops (Systane Ultra, Alcon) during imaging. Radial volume scans (1.7 mm diameter, 1000 A-scans/B-scan, 8 B-scans/volume, 24 frames/B-scan) were centered on the optic nerve. ONL thickness was measured 0.6 mm from the optic nerve head using ImageJ.

Optomotor response test

Request a detailed protocol

Mouse visual acuity was assessed using an Optometry System (Cerebral Mechanics Inc). Testing was performed with a grating drifting at 12°/second and 100% contrast. The right eyes and left eyes were tested independently using counterclockwise and clockwise grating rotations, respectively. A staircase procedure was applied, with the observer progressing from low to high spatial frequencies to determine acuity thresholds. Each animal was tested for ~10–15 min per session.

Statistics

The data were expressed as mean ± standard error of the mean (s.e.m.). Sample sizes for each experiment were specified in the figure legend. Statistical analysis involved conducting one-way or two-way ANOVA followed by the Tukey’s test for comparing multiple groups, while the unpaired two-tailed Student’s t-test was employed for comparing two groups. Blinded quantification was performed by an independent investigator during assessment of regenerated neurons. All in vivo experiments were performed with at least three independent laboratory replicates. Data shown represent biological replicates corresponding to individual mice; no technical replicates were used for quantitative analysis.

Data availability

snRNA and snATAC sequence data were deposited in GEO (accession number GSE331267).

The following data sets were generated

1. Liao B
2. Lyu C
3. Jiang Y
4. Liu S
5. Wong W
6. Zhang J
7. Tsang H
8. Xie J
9. Chen L
10. Zhang Q
11. Xiong W
(2026) NCBI GEO
ID GSE331267. Synergistic Inhibition of Notch Signaling and Forced Cell Cycle Re-entry Drive Müller Glia Reprogramming in Uninjured Mouse Retina.

https://ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE331267

References

1. Ahmed M
2. Kojima Y
3. Masai I
(2022) Strip1 regulates retinal ganglion cell survival by suppressing Jun-mediated apoptosis to promote retinal neural circuit formation
eLife 11:e74650.

https://doi.org/10.7554/eLife.74650
- PubMed
- Google Scholar
1. Alto LT
2. Terman JR
(2017) Semaphorins and their signaling mechanisms
Methods in Molecular Biology 1493:1–25.

https://doi.org/10.1007/978-1-4939-6448-2_1
- PubMed
- Google Scholar
1. Borggrefe T
2. Oswald F
(2009) The Notch signaling pathway: transcriptional regulation at Notch target genes
Cellular and Molecular Life Sciences 66:1631–1646.

https://doi.org/10.1007/s00018-009-8668-7
- PubMed
- Google Scholar
(2006) Müller cells in the healthy and diseased retina
Progress in Retinal and Eye Research 25:397–424.

https://doi.org/10.1016/j.preteyeres.2006.05.003
- PubMed
- Google Scholar
(2009) Cellular signaling and factors involved in Müller cell gliosis: neuroprotective and detrimental effects
Progress in Retinal and Eye Research 28:423–451.

https://doi.org/10.1016/j.preteyeres.2009.07.001
- PubMed
- Google Scholar
1. Campbell LJ
2. Hobgood JS
3. Jia M
4. Boyd P
5. Hipp RI
6. Hyde DR
(2021) Notch3 and DeltaB maintain Müller glia quiescence and act as negative regulators of regeneration in the light-damaged zebrafish retina
Glia 69:546–566.

https://doi.org/10.1002/glia.23912
- PubMed
- Google Scholar
(2022) Retinal regeneration requires dynamic Notch signaling
Neural Regeneration Research 17:1199–1209.

https://doi.org/10.4103/1673-5374.327326
- PubMed
- Google Scholar
1. Cepko CL
2. Austin CP
3. Yang X
4. Alexiades M
5. Ezzeddine D
(1996) Cell fate determination in the vertebrate retina
PNAS 93:589–595.

https://doi.org/10.1073/pnas.93.2.589
- PubMed
- Google Scholar
1. Conner C
2. Ackerman KM
3. Lahne M
4. Hobgood JS
5. Hyde DR
(2014) Repressing notch signaling and expressing TNFα are sufficient to mimic retinal regeneration by inducing Müller glial proliferation to generate committed progenitor cells
The Journal of Neuroscience 34:14403–14419.

https://doi.org/10.1523/JNEUROSCI.0498-14.2014
- PubMed
- Google Scholar
1. Dent EW
2. Barnes AM
3. Tang F
4. Kalil K
(2004) Netrin-1 and semaphorin 3A promote or inhibit cortical axon branching, respectively, by reorganization of the cytoskeleton
The Journal of Neuroscience 24:3002–3012.

https://doi.org/10.1523/JNEUROSCI.4963-03.2004
- PubMed
- Google Scholar
(2006) Decreased expression of pro-apoptotic Bcl-2 family members during retinal development and differential sensitivity to cell death
Developmental Biology 291:154–169.

https://doi.org/10.1016/j.ydbio.2005.12.026
- PubMed
- Google Scholar
1. Elsaeidi F
2. Macpherson P
3. Mills EA
4. Jui J
5. Flannery JG
6. Goldman D
(2018) Notch suppression collaborates with Ascl1 and Lin28 to unleash a regenerative response in fish retina, but not in mice
The Journal of Neuroscience 38:2246–2261.

https://doi.org/10.1523/JNEUROSCI.2126-17.2018
- PubMed
- Google Scholar
1. Fang R
2. Preissl S
3. Li Y
4. Hou X
5. Lucero J
6. Wang X
7. Motamedi A
8. Shiau AK
9. Zhou X
10. Xie F
11. Mukamel EA
12. Zhang K
13. Zhang Y
14. Behrens MM
15. Ecker JR
16. Ren B
(2021) Comprehensive analysis of single cell ATAC-seq data with SnapATAC
Nature Communications 12:1337.

https://doi.org/10.1038/s41467-021-21583-9
- PubMed
- Google Scholar
1. Fischer AJ
2. Reh TA
(2001) Müller glia are a potential source of neural regeneration in the postnatal chicken retina
Nature Neuroscience 4:247–252.

https://doi.org/10.1038/85090
- PubMed
- Google Scholar
1. Fischer AJ
(2005) Neural regeneration in the chick retina
Progress in Retinal and Eye Research 24:161–182.

https://doi.org/10.1016/j.preteyeres.2004.07.003
- PubMed
- Google Scholar
1. Fogerty J
2. Song P
3. Boyd P
4. Grabinski SE
5. Hoang T
6. Reich A
7. Cianciolo LT
8. Blackshaw S
9. Mumm JS
10. Hyde DR
11. Perkins BD
(2022) Notch inhibition promotes regeneration and immunosuppression supports cone survival in a zebrafish model of inherited retinal dystrophy
The Journal of Neuroscience 42:5144–5158.

https://doi.org/10.1523/JNEUROSCI.0244-22.2022
- PubMed
- Google Scholar
1. Furukawa T
2. Mukherjee S
3. Bao Z-Z
4. Morrow EM
5. Cepko CL
(2000) rax, hes1, and notch1 promote the formation of müller glia by postnatal retinal progenitor cells
Neuron 26:383–394.

https://doi.org/10.1016/S0896-6273(00)81171-X
- Google Scholar
1. Goldman D
(2014) Müller glial cell reprogramming and retina regeneration
Nature Reviews. Neuroscience 15:431–442.

https://doi.org/10.1038/nrn3723
- PubMed
- Google Scholar
(2012) Betacellulin promotes cell proliferation in the neural stem cell niche and stimulates neurogenesis
PNAS 109:1317–1322.

https://doi.org/10.1073/pnas.1016199109
- PubMed
- Google Scholar
1. Gorsuch RA
2. Hyde DR
(2014) Regulation of Müller glial dependent neuronal regeneration in the damaged adult zebrafish retina
Experimental Eye Research 123:131–140.

https://doi.org/10.1016/j.exer.2013.07.012
- PubMed
- Google Scholar
1. Hamon A
2. García-García D
3. Ail D
4. Bitard J
5. Chesneau A
6. Dalkara D
7. Locker M
8. Roger JE
9. Perron M
(2019) Linking YAP to müller glia quiescence exit in the degenerative retina
Cell Reports 27:1712–1725.

https://doi.org/10.1016/j.celrep.2019.04.045
- PubMed
- Google Scholar
1. Hanna J
2. Saha K
3. Pando B
4. van Zon J
5. Lengner CJ
6. Creyghton MP
7. van Oudenaarden A
8. Jaenisch R
(2009) Direct cell reprogramming is a stochastic process amenable to acceleration
Nature 462:595–601.

https://doi.org/10.1038/nature08592
- PubMed
- Google Scholar
(2006) Retinitis pigmentosa Prevalence and inheritance patterns
Lancet 368:1795–1809.

https://doi.org/10.2174/138920211795860107
- PubMed
- Google Scholar
1. Hayes S
2. Nelson BR
3. Buckingham B
4. Reh TA
(2007) Notch signaling regulates regeneration in the avian retina
Developmental Biology 312:300–311.

https://doi.org/10.1016/j.ydbio.2007.09.046
- PubMed
- Google Scholar
1. Henke RM
2. Savage TK
3. Meredith DM
4. Glasgow SM
5. Hori K
6. Dumas J
7. MacDonald RJ
8. Johnson JE
(2009) Neurog2 is a direct downstream target of the Ptf1a-Rbpj transcription complex in dorsal spinal cord
Development 136:2945–2954.

https://doi.org/10.1242/dev.035352
- PubMed
- Google Scholar
1. Hoang T
2. Wang J
3. Boyd P
4. Wang F
5. Santiago C
6. Jiang L
7. Yoo S
8. Lahne M
9. Todd LJ
10. Jia M
11. Saez C
12. Keuthan C
13. Palazzo I
14. Squires N
15. Campbell WA
16. Rajaii F
17. Parayil T
18. Trinh V
19. Kim DW
20. Wang G
21. Campbell LJ
22. Ash J
23. Fischer AJ
24. Hyde DR
25. Qian J
26. Blackshaw S
(2020) Gene regulatory networks controlling vertebrate retinal regeneration
Science 370:eabb8598.

https://doi.org/10.1126/science.abb8598
- Google Scholar
1. Hoang DA
2. Liao B
3. Zheng Z
4. Xiong W
(2023) Mutation-independent gene knock-in therapy targeting 5’UTR for autosomal dominant retinitis pigmentosa
Signal Transduction and Targeted Therapy 8:100.

https://doi.org/10.1038/s41392-022-01308-0
- PubMed
- Google Scholar
(2006a) Notch activity permits retinal cells to progress through multiple progenitor states and acquire a stem cell property
PNAS 103:18998–19003.

https://doi.org/10.1073/pnas.0608155103
- PubMed
- Google Scholar
(2006b) Notch 1 inhibits photoreceptor production in the developing mammalian retina
Development 133:913–923.

https://doi.org/10.1242/dev.02245
- PubMed
- Google Scholar
(2009) Development and neurogenic potential of Müller glial cells in the vertebrate retina
Progress in Retinal and Eye Research 28:249–262.

https://doi.org/10.1016/j.preteyeres.2009.05.002
- PubMed
- Google Scholar
1. Jorstad NL
2. Wilken MS
3. Todd L
4. Finkbeiner C
5. Nakamura P
6. Radulovich N
7. Hooper MJ
8. Chitsazan A
9. Wilkerson BA
10. Rieke F
11. Reh TA
(2020) STAT signaling modifies Ascl1 chromatin binding and limits neural regeneration from muller glia in adult mouse retina
Cell Reports 30:2195–2208.

https://doi.org/10.1016/j.celrep.2020.01.075
- PubMed
- Google Scholar
1. Kimura A
2. Namekata K
3. Guo X
4. Harada C
5. Harada T
(2016) Neuroprotection, growth factors and BDNF-TrkB signalling in retinal degeneration
International Journal of Molecular Sciences 17:1584.

https://doi.org/10.3390/ijms17091584
- PubMed
- Google Scholar
1. Kolomeyer AM
2. Zarbin MA
(2014) Trophic factors in the pathogenesis and therapy for retinal degenerative diseases
Survey of Ophthalmology 59:134–165.

https://doi.org/10.1016/j.survophthal.2013.09.004
- PubMed
- Google Scholar
(2020) Reprogramming müller glia to regenerate retinal neurons
Annual Review of Vision Science 6:171–193.

https://doi.org/10.1146/annurev-vision-121219-081808
- PubMed
- Google Scholar
1. Le N
2. Vu TD
3. Palazzo I
4. Pulya R
5. Kim Y
6. Blackshaw S
7. Hoang T
(2024) Robust reprogramming of glia into neurons by inhibition of Notch signaling and nuclear factor I (NFI) factors in adult mammalian retina
Science Advances 10:eadn2091.

https://doi.org/10.1126/sciadv.adn2091
- PubMed
- Google Scholar
1. Le N
2. Awad S
3. Palazzo I
4. Hoang T
5. Blackshaw S
(2025) Viral-mediated Pou5f1 (Oct4) overexpression and inhibition of Notch signaling synergistically induce neurogenic competence in mammalian Müller glia
eLife 14:e106450.

https://doi.org/10.7554/eLife.106450
- PubMed
- Google Scholar
(2023) Zbtb16 mediates a switch between Fgf signalling regimes in the developing hindbrain
Development 150:dev201319.

https://doi.org/10.1242/dev.201319
- PubMed
- Google Scholar
1. Lenkowski JR
2. Raymond PA
(2014) Müller glia: Stem cells for generation and regeneration of retinal neurons in teleost fish
Progress in Retinal and Eye Research 40:94–123.

https://doi.org/10.1016/j.preteyeres.2013.12.007
- PubMed
- Google Scholar
1. Li M
2. Guo H
3. Carey M
4. Huang C
(2024) Transcriptional and epigenetic dysregulation impairs generation of proliferative neural stem and progenitor cells during brain aging
Nature Aging 4:62–79.

https://doi.org/10.1038/s43587-023-00549-0
- PubMed
- Google Scholar
1. Li XJ
2. Morgan C
3. Li L
4. Zhang WY
5. Chrysostomou E
6. Doetzlhofer A
(2025) The Notch ligand Jagged1 plays a dual role in cochlear hair cell regeneration
Nature Communications 16:8169.

https://doi.org/10.1038/s41467-025-63053-6
- PubMed
- Google Scholar
1. Madisen L
2. Zwingman TA
3. Sunkin SM
4. Oh SW
5. Zariwala HA
6. Gu H
7. Ng LL
8. Palmiter RD
9. Hawrylycz MJ
10. Jones AR
11. Lein ES
12. Zeng H
(2010) A robust and high-throughput Cre reporting and characterization system for the whole mouse brain
Nature Neuroscience 13:133–140.

https://doi.org/10.1038/nn.2467
- PubMed
- Google Scholar
1. Magnusson JP
2. Göritz C
3. Tatarishvili J
4. Dias DO
5. Smith EMK
6. Lindvall O
7. Kokaia Z
8. Frisén J
(2014) A latent neurogenic program in astrocytes regulated by Notch signaling in the mouse
Science 346:237–241.

https://doi.org/10.1126/science.346.6206.237
- PubMed
- Google Scholar
1. McMahon SS
2. McDermott KW
(2006) A comparison of cell transplantation and retroviral gene transfection as tools to study lineage and differentiation in the rat spinal cord
Journal of Neuroscience Methods 152:243–249.

https://doi.org/10.1016/j.jneumeth.2005.09.008
- PubMed
- Google Scholar
1. Mears AJ
2. Kondo M
3. Swain PK
4. Takada Y
5. Bush RA
6. Saunders TL
7. Sieving PA
8. Swaroop A
(2001) Nrl is required for rod photoreceptor development
Nature Genetics 29:447–452.

https://doi.org/10.1038/ng774
- PubMed
- Google Scholar
1. Mills EA
2. Goldman D
(2017) The regulation of notch signaling in retinal development and regeneration
Current Pathobiology Reports 5:323–331.

https://doi.org/10.1007/s40139-017-0153-7
- PubMed
- Google Scholar
1. Minegishi Y
2. Iejima D
3. Kobayashi H
4. Chi ZL
5. Kawase K
6. Yamamoto T
7. Seki T
8. Yuasa S
9. Fukuda K
10. Iwata T
(2013) Enhanced optineurin E50K-TBK1 interaction evokes protein insolubility and initiates familial primary open-angle glaucoma
Human Molecular Genetics 22:3559–3567.

https://doi.org/10.1093/hmg/ddt210
- PubMed
- Google Scholar
1. Mo A
2. Mukamel EA
3. Davis FP
4. Luo C
5. Henry GL
6. Picard S
7. Urich MA
8. Nery JR
9. Sejnowski TJ
10. Lister R
11. Eddy SR
12. Ecker JR
13. Nathans J
(2015) Epigenomic signatures of neuronal diversity in the mammalian brain
Neuron 86:1369–1384.

https://doi.org/10.1016/j.neuron.2015.05.018
- PubMed
- Google Scholar
1. Molday RS
2. Moritz OL
(2015) Photoreceptors at a glance
Journal of Cell Science 128:4039–4045.

https://doi.org/10.1242/jcs.175687
- PubMed
- Google Scholar
1. Monroe TO
2. Hill MC
3. Morikawa Y
4. Leach JP
5. Heallen T
6. Cao S
7. Krijger PHL
8. de Laat W
9. Wehrens XHT
10. Rodney GG
11. Martin JF
(2019) YAP partially reprograms chromatin accessibility to directly induce adult cardiogenesis in vivo
Developmental Cell 48:765–779.

https://doi.org/10.1016/j.devcel.2019.01.017
- PubMed
- Google Scholar
1. Nelson BR
2. Ueki Y
3. Reardon S
4. Karl MO
5. Georgi S
6. Hartman BH
7. Lamba DA
8. Reh TA
(2011) Genome-wide analysis of Müller glial differentiation reveals a requirement for Notch signaling in postmitotic cells to maintain the glial fate
PLOS ONE 6:e22817.

https://doi.org/10.1371/journal.pone.0022817
- PubMed
- Google Scholar
1. Nelson CM
2. Gorsuch RA
3. Bailey TJ
4. Ackerman KM
5. Kassen SC
6. Hyde DR
(2012) Stat3 defines three populations of Müller glia and is required for initiating maximal müller glia proliferation in the regenerating zebrafish retina
The Journal of Comparative Neurology 520:4294–4311.

https://doi.org/10.1002/cne.23213
- PubMed
- Google Scholar
(2001) Neurons derived from radial glial cells establish radial units in neocortex
Nature 409:714–720.

https://doi.org/10.1038/35055553
- Google Scholar
(2021) New insights into the intricacies of proneural gene regulation in the embryonic and adult cerebral cortex
Frontiers in Molecular Neuroscience 14:642016.

https://doi.org/10.3389/fnmol.2021.642016
- PubMed
- Google Scholar
1. Oswald J
2. Kegeles E
3. Minelli T
4. Volchkov P
5. Baranov P
(2021) Transplantation of miPSC/mESC-derived retinal ganglion cells into healthy and glaucomatous retinas
Molecular Therapy. Methods & Clinical Development 21:180–198.

https://doi.org/10.1016/j.omtm.2021.03.004
- PubMed
- Google Scholar
1. Papp B
2. Plath K
(2013) Epigenetics of reprogramming to induced pluripotency
Cell 152:1324–1343.

https://doi.org/10.1016/j.cell.2013.02.043
- PubMed
- Google Scholar
(2024) The mechanics of the retina: Müller glia role on retinal extracellular matrix and modelling
Frontiers in Medicine 11:1393057.

https://doi.org/10.3389/fmed.2024.1393057
- PubMed
- Google Scholar
1. Reichenbach A
2. Bringmann A
(2013) New functions of Müller cells
Glia 61:651–678.

https://doi.org/10.1002/glia.22477
- PubMed
- Google Scholar
1. Riesenberg AN
2. Liu Z
3. Kopan R
4. Brown NL
(2009) Rbpj cell autonomous regulation of retinal ganglion cell and cone photoreceptor fates in the mouse retina
The Journal of Neuroscience 29:12865–12877.

https://doi.org/10.1523/JNEUROSCI.3382-09.2009
- PubMed
- Google Scholar
(2012)
Gene expression changes within Müller glial cells in retinitis pigmentosa

Molecular Vision 18:1197–1214.
- PubMed
- Google Scholar
1. Rowan S
2. Conley KW
3. Le TT
4. Donner AL
5. Maas RL
6. Brown NL
(2008) Notch signaling regulates growth and differentiation in the mammalian lens
Developmental Biology 321:111–122.

https://doi.org/10.1016/j.ydbio.2008.06.002
- PubMed
- Google Scholar
1. Rueda EM
2. Hall BM
3. Hill MC
4. Swinton PG
5. Tong X
6. Martin JF
7. Poché RA
(2019) The hippo pathway blocks mammalian retinal müller glial cell reprogramming
Cell Reports 27:1637–1649.

https://doi.org/10.1016/j.celrep.2019.04.047
- PubMed
- Google Scholar
(2011) A high proliferation rate is required for cell reprogramming and maintenance of human embryonic stem cell identity
Current Biology 21:45–52.

https://doi.org/10.1016/j.cub.2010.11.049
- PubMed
- Google Scholar
1. Sahu A
2. Devi S
3. Jui J
4. Goldman D
(2021) Notch signaling via Hey1 and Id2b regulates Müller glia’s regenerative response to retinal injury
Glia 69:2882–2898.

https://doi.org/10.1002/glia.24075
- PubMed
- Google Scholar
1. Shekhar K
2. Lapan SW
3. Whitney IE
4. Tran NM
5. Macosko EZ
6. Kowalczyk M
7. Adiconis X
8. Levin JZ
9. Nemesh J
10. Goldman M
11. McCarroll SA
12. Cepko CL
13. Regev A
14. Sanes JR
(2016) Comprehensive classification of retinal bipolar neurons by single-cell transcriptomics
Cell 166:1308–1323.

https://doi.org/10.1016/j.cell.2016.07.054
- PubMed
- Google Scholar
(2008) Oscillations in notch signaling regulate maintenance of neural progenitors
Neuron 58:52–64.

https://doi.org/10.1016/j.neuron.2008.02.014
- PubMed
- Google Scholar
1. Shu Y
2. Li W
3. Huang M
4. Quan YZ
5. Scheffer D
6. Tian C
7. Tao Y
8. Liu X
9. Hochedlinger K
10. Indzhykulian AA
11. Wang Z
12. Li H
13. Chen ZY
(2019) Renewed proliferation in adult mouse cochlea and regeneration of hair cells
Nature Communications 10:5530.

https://doi.org/10.1038/s41467-019-13157-7
- PubMed
- Google Scholar
1. Soucy JR
2. Todd L
3. Kriukov E
4. Phay M
5. Malechka VV
6. Rivera JD
7. Reh TA
8. Baranov P
(2023) Controlling donor and newborn neuron migration and maturation in the eye through microenvironment engineering
PNAS 120:e2302089120.

https://doi.org/10.1073/pnas.2302089120
- PubMed
- Google Scholar
1. Stuart T
2. Butler A
3. Hoffman P
4. Hafemeister C
5. Papalexi E
6. Mauck WM
7. Hao Y
8. Stoeckius M
9. Smibert P
10. Satija R
(2019) Comprehensive integration of single-cell data
Cell 177:1888–1902.

https://doi.org/10.1016/j.cell.2019.05.031
- PubMed
- Google Scholar
1. Todd L
2. Squires N
3. Suarez L
4. Fischer AJ
(2016) Jak/Stat signaling regulates the proliferation and neurogenic potential of Müller glia-derived progenitor cells in the avian retina
Scientific Reports 6:35703.

https://doi.org/10.1038/srep35703
- PubMed
- Google Scholar
1. Todd L
2. Reh TA
(2022) Comparative biology of vertebrate retinal regeneration: restoration of vision through cellular reprogramming
Cold Spring Harbor Perspectives in Biology 14:a040816.

https://doi.org/10.1101/cshperspect.a040816
- PubMed
- Google Scholar
(2008) Individual retinal progenitor cells display extensive heterogeneity of gene expression
PLOS ONE 3:e1588.

https://doi.org/10.1371/journal.pone.0001588
- PubMed
- Google Scholar
1. Wan J
2. Goldman D
(2016) Retina regeneration in zebrafish
Current Opinion in Genetics & Development 40:41–47.

https://doi.org/10.1016/j.gde.2016.05.009
- PubMed
- Google Scholar
1. Wang Y
2. Rattner A
3. Zhou Y
4. Williams J
5. Smallwood PM
6. Nathans J
(2012) Norrin/Frizzled4 signaling in retinal vascular development and blood brain barrier plasticity
Cell 151:1332–1344.

https://doi.org/10.1016/j.cell.2012.10.042
- PubMed
- Google Scholar
1. Wolffe AP
(1991) Implications of DNA replication for eukaryotic gene expression
Journal of Cell Science 99 (Pt 2):201–206.

https://doi.org/10.1242/jcs.99.2.201
- PubMed
- Google Scholar
1. Wu Z
2. Liao B
3. Ying J
4. Keung J
5. Zheng Z
6. Ahola V
7. Xiong W
(2025) Simultaneous Cyclin D1 Overexpression and P27kip1 knockdown enable robust müller glia cell cycle reactivation in uninjured mouse retina
eLife 13:e100904.

https://doi.org/10.7554/eLife.100904.2
- Google Scholar
1. Yan W
2. Laboulaye MA
3. Tran NM
4. Whitney IE
5. Benhar I
6. Sanes JR
(2020) Mouse retinal cell atlas: molecular identification of over sixty amacrine cell types
The Journal of Neuroscience 40:5177–5195.

https://doi.org/10.1523/JNEUROSCI.0471-20.2020
- PubMed
- Google Scholar
1. Yao K
2. Qiu S
3. Tian L
4. Snider WD
5. Flannery JG
6. Schaffer DV
7. Chen B
(2016) Wnt regulates proliferation and neurogenic potential of müller glial cells via a Lin28/let-7 miRNA-dependent pathway in adult mammalian retinas
Cell Reports 17:165–178.

https://doi.org/10.1016/j.celrep.2016.08.078
- PubMed
- Google Scholar
(2006) Notch1 functions to suppress cone-photoreceptor fate specification in the developing mouse retina
Development 133:1367–1378.

https://doi.org/10.1242/dev.02311
- PubMed
- Google Scholar
1. Yeo RW
2. Zhou OY
3. Zhong BL
4. Sun ED
5. Navarro Negredo P
6. Nair S
7. Sharmin M
8. Ruetz TJ
9. Wilson M
10. Kundaje A
11. Dunn AR
12. Brunet A
(2023) Chromatin accessibility dynamics of neurogenic niche cells reveal defects in neural stem cell adhesion and migration during aging
Nature Aging 3:866–893.

https://doi.org/10.1038/s43587-023-00449-3
- PubMed
- Google Scholar
1. Yoon K
2. Gaiano N
(2005) Notch signaling in the mammalian central nervous system: insights from mouse mutants
Nature Neuroscience 8:709–715.

https://doi.org/10.1038/nn1475
- PubMed
- Google Scholar
(2009) Effects on XIAP retinal detachment–induced photoreceptor apoptosis
Investigative Opthalmology & Visual Science 50:1448.

https://doi.org/10.1167/iovs.08-2855
- PubMed
- Google Scholar
(2020) A widespread neurogenic potential of neocortical astrocytes is induced by injury
Cell Stem Cell 27:605–617.

https://doi.org/10.1016/j.stem.2020.07.006
- PubMed
- Google Scholar
1. Zhang K
2. Zemke NR
3. Armand EJ
4. Ren B
(2024) A fast, scalable and versatile tool for analysis of single-cell omics data
Nature Methods 21:217–227.

https://doi.org/10.1038/s41592-023-02139-9
- PubMed
- Google Scholar

Article and author information

Author details

Baoshan Liao

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

Competing interests
inventor on a pending patent application (Priority No. 18/820,216) related to the CCA vector used in this study

"This ORCID iD identifies the author of this article:" 0009-0007-6599-624X
Chengshang Lyu

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Methodology

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-0971-2851
Yuqing Jiang

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Data curation, Methodology

Competing interests
No competing interests declared
Shanggong Liu

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Data curation, Methodology

Competing interests
No competing interests declared
Waiho Wong

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Data curation, Methodology

Competing interests
No competing interests declared
Jiadong Zhang
1. Department of Biomedical Engineering, College of Biomedicine, City University of Hong Kong, Hong Kong, China
2. Hong Kong Centre for Cerebro-Cardiovascular Health Engineering (COCHE), Tai Po, China
Contribution
Methodology

Competing interests
No competing interests declared
Hoyin Tsang

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Data curation, Methodology

Competing interests
No competing interests declared
Junxi Xie

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Data curation, Methodology

Competing interests
No competing interests declared
Lingxi Chen

Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Methodology

Competing interests
No competing interests declared
Qinrong Zhang

Department of Biomedical Engineering, College of Biomedicine, City University of Hong Kong, Hong Kong, China

Contribution
Methodology

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0001-8893-514X
Wenjun Xiong
1. Department of Biomedical Sciences, College of Biomedicine, City University of Hong Kong, Hong Kong, China
2. Key Laboratory of Biochip Technology, Biotech and Health Centre, Shenzhen Research Institute of City University of Hong Kong, Shenzhen, China
Contribution
Conceptualization, Supervision, Funding acquisition, Investigation, Project administration, Writing – review and editing

For correspondence
wenjun.xiong@cityu.edu.hk

Competing interests
inventor on a pending patent application (Priority No. 18/820,216) related to the CCA vector used in this study

"This ORCID iD identifies the author of this article:" 0000-0001-6836-2807

Funding

Research Grants Council, University Grants Committee (11103819)

Wenjun Xiong

Research Grants Council, University Grants Committee (11102922)

Wenjun Xiong

Research Grants Council, University Grants Committee (11100723)

Wenjun Xiong

Health and Medical Research Fund (05160276)

Wenjun Xiong

Health and Medical Research Fund (06172466)

Wenjun Xiong

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This research was funded by Research Grants Council Hong Kong Project (11103819, 11102922, and 11100723), Hong Kong Health and Medical Research Fund Project (05160276 and 06172466), TUNG Biomedical Sciences Foundation, and Ming Wai Lau Center for Reparative Medicine Research Associate Program.

Ethics

All animal experiments were carried out in strict compliance with the national guidelines for the care and use of laboratory animals and were approved by the Hong Kong Department of Health under Animals Ordinance Chapter 340 (Ref: (20-130) in DH/HT&A/8/2/5 Pt.2) and by the City University of Hong Kong Animal ethics committee (Ref: A-0264). All animal handling, housing, behavioral tests, and surgical procedures were performed in accordance with standardized laboratory animal welfare protocols to minimize pain, distress, and suffering throughout the experimental period. Animals were maintained under controlled environmental conditions with free access to standard food and water. Appropriate anesthesia and euthanasia procedures were applied whenever necessary to reduce animal discomfort. All experimental protocols were reviewed and officially authorized by the ethics committee prior to the commencement of the study.

Version history

Sent for peer review: March 10, 2026
Preprint posted: March 12, 2026
Reviewed Preprint version 1: April 28, 2026
Reviewed Preprint version 2: August 18, 2026
Version of Record published: September 17, 2026

Cite all versions

You can cite all versions using the DOI https://doi.org/10.7554/eLife.111251. This DOI represents all versions, and will always resolve to the latest one.

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.