{"id":74034,"date":"2023-09-28T20:00:00","date_gmt":"2023-09-29T00:00:00","guid":{"rendered":"https:\/\/platohealth.ai\/human-retinal-ganglion-cell-neurons-generated-by-synchronous-bmp-inhibition-and-transcription-factor-mediated-reprogramming-npj-regenerative-medicine\/"},"modified":"2023-09-29T23:56:07","modified_gmt":"2023-09-30T03:56:07","slug":"human-retinal-ganglion-cell-neurons-generated-by-synchronous-bmp-inhibition-and-transcription-factor-mediated-reprogramming-npj-regenerative-medicine","status":"publish","type":"post","link":"https:\/\/platohealth.ai\/human-retinal-ganglion-cell-neurons-generated-by-synchronous-bmp-inhibition-and-transcription-factor-mediated-reprogramming-npj-regenerative-medicine\/","title":{"rendered":"Human retinal ganglion cell neurons generated by synchronous BMP inhibition and transcription factor mediated reprogramming – npj Regenerative Medicine","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
NEUROG2, ATOH7, ISL1, and POU4F2 participate in RGC development and maturation. To explore the idea that their forced expression could convert stem cells directly into RGCs, we used a doxycycline (dox)-inducible third-generation Tet-ON (TetO) promoter coupled to a polycistronic N<\/i>EUROG2<\/i>, A<\/i>TOH7<\/i>, I<\/i>SL1<\/i>, and P<\/i>OU4F<\/i>2<\/i> (NAIP2) gene cassette (Fig. 1a<\/a>). After transient transfection into IMR90.4 POU4F2-p2A-tdTomato cells (Fig. 1b<\/a>) we observed that untreated controls (CTLs) had a PSC-like morphology (Fig. 1c<\/a>) and dox treated TetO-NAIP2<\/i> transfected cells were PSC-like but also included sparsely populated neurons (Fig. 1d<\/a>). These were not POU4F2-tdTomato+ and the neuronal looking cells were quickly overtaken by proliferating PSCs. To confirm that the reporters were functioning properly, the IMR90.4 POU4F2-p2A-tdTomato and WA09 POU4F2-p2A-h2b-mNeonGreen PSC lines integrated with TetO-NAIP2<\/i> were differentiated as 3D retinal organoids and confirmed to express fluorescent proteins only in RGCs (Fig. 1e, f<\/a>). Encouraged by the partial morphological conversion of stem cells into RGCs, we next stably integrated the TetO-NAIP2<\/i> cassette into the CLYBL<\/i> SHS, a site that supports sustained transgene expression22<\/a><\/sup>. Integration was accomplished with high-fidelity EnCas12a and an appropriate cRNA for gene targeting (Fig. 1g<\/a>). Despite stable integration, as determined by the acquisition of zeocin resistance, dividing non-neuronal flat cells remained problematic for the long-term growth of the induced neurons.<\/p>\n a<\/b> A conceptual diagram for a 3G Tet-ON system where doxycycline binds to constitutively expressed rtTA to induce polycistronic expression of NEUROG2<\/i>, ATOH7<\/i>, ISLET1<\/i>, and POU4F2<\/i> (NAIP2<\/i>) leading to conversion of RGC-like neurons. b<\/b> Electroporation approach for the transient transfection of the NAIP2 transgene package. c<\/b> PSCs that were transiently transfected with control (empty) or d<\/b> NAIP2<\/i> plasmids and treated with dox for 3 days. Black arrows = PSCs, white arrows = converted neurons. Brightfield and fluorescent images of e<\/b> the POU4F2-tdTomato+ reporter in a 66-day-old retinal organoid in the IMR90.4 genetic background and f<\/b> a 65-day-old POU4F2-h2b-mNeonGreen+ retinal organoid in the WA09 genetic background. g<\/b> U6 promoter-driven expression of an AsCas12a cRNA targeting the CLYBL safe harbor site for insertion of a zeocin selectable Tet-inducible transgene cassette. Scale c<\/b>, d<\/b>\u2009=\u2009100\u2009\u00b5m, e<\/b>, f<\/b>\u2009=\u2009400\u2009\u00b5m.<\/p>\n<\/div>\n<\/div>\n Simultaneous expression of NEUROG2<\/i> and dual SMAD inhibition (to block BMP and TGF-\u03b2) can generate human patterned forebrain-like induced neurons (hpiNs)10<\/a><\/sup>. To explore whether this could be adapted for RGCs, we simultaneously expressed NAIP2<\/i> and inhibited BMP signaling. IMR90.4 POU4F2-tdTomato reporter PSCs harboring empty (CTL) or NAIP2<\/i> cassettes were treated with 1.0\u2009\u03bcg\/mL dox (condition 1), 100\u2009nM LDN (condition 2), or LDN plus dox (condition 3) and assessed for changes in morphology (Fig. 2a<\/a>). Empty TetO cassette integrated CTL cells treated with LDN plus dox failed to generate neurons (Fig. 2b<\/a>), whereas TetO-NAIP2<\/i> integrated cells (condition 3) developed long branched neurites by day 6 (Fig. 2c<\/a>). Interestingly, while dual SMAD inhibition with both LDN and SB-431542 (SB) was reported to produce patterned forebrain neurons, its combined use did not improve RGC-iN differentiation more than LDN alone (Supplementary Fig. 1<\/a>). Although the NAIP2<\/i> combination worked well with BMP inhibition, we wanted to evaluate the contribution of each TF so we integrated each gene (NEUROG2<\/i>, ATOH7<\/i>, ISLET1<\/i>, or POU4F2<\/i>) individually as Tet-inducible cassettes and compared their morphologies after LDN\/dox induction (Fig. 2d\u2013g<\/a>). NEUROG2<\/i> expression led to efficient neural conversion over the first 2 days with long neurites by day 6 (Fig. 2d<\/a>), whereas neurite formation in NAIP2-nc (NAIP2 integrated and zeocin selected but not clonally isolated) RGC-iNs typically occurred in as little as 24\u2009h, indicating a faster rate of differentiation than NEUROG2<\/i> alone. POU4F2<\/i> overexpressed cells (Fig. 2e<\/a>) formed stout-looking neurons with stubby neurites that appeared only partially converted, ATOH7<\/i> overexpressed cells (Fig. 2f<\/a>) became a mixture of flat cells and sparsely populated neurons and ISL1<\/i> cells (Fig. 2g<\/a>) became flat with no recognizable neuronal features. Overall, only NAIP2<\/i> and NEUROG2<\/i> expressing cells produced robust levels of neurons.<\/p>\n a<\/b> Timeline of LDN193189 and\/or dox treatments of POU4F2-tdTomato PSCs with an integrated empty cassette (control) or NAIP2 cassette (conditions 1\u20133). Brightfield images of dox treated b<\/b> empty cassette control cells, c<\/b> NAIP2-nc, d<\/b> NEUROG2, e<\/b> POU4F2, f<\/b> ATOH7, and g<\/b> ISLET1 cells after 6 days. NAIP2-nc = not clonally selected. h<\/b> Quantification of %POU4F2+ cells at day 6 relative to DAPI in the empty cassette control and conditions 1\u20133 illustrated in (a<\/b>). *P<\/i>\u2009<\u20090.05, **P<\/i>\u2009<\u20090.01, ***P<\/i>\u2009<\u20090.001, ns P<\/i>\u2009>\u20090.05, n<\/i>\u2009=\u20093. Error bars are reported as standard deviation (SD). i<\/b>\u2013n<\/b> NAIP2<\/i> RGC-iNs from different genetic backgrounds differentiated and imaged in brightfield and POU4F2+ tdTomato fluorescence for clonally selected i<\/b>, j<\/b> IMR90.4 and k<\/b>, l<\/b> GM23720 PSCs, and m<\/b>, n<\/b> brightfield and POU4F2\u2009+\u2009h2b-mNeonGreen fluorescence for WA09 PSCs. Click-iT EdU staining in o<\/b> undifferentiated PSCs and 2-day LDN\/dox treated p<\/b> control, q<\/b> NEUROG2, and r<\/b> NAIP2 cells co-stained with DAPI. Arrows in O-R indicate EdU+ cells. Scale b<\/b>\u2013g<\/b>, i<\/b>\u2013n<\/b>\u2009=\u2009100\u2009\u00b5m, o<\/b>\u2013r<\/b>\u2009=\u2009200\u2009\u00b5m.<\/p>\n<\/div>\n<\/div>\n To corroborate the identity of RGCs, we utilized the POU4F2–<\/i>tdTomato reporter to visualize real-time changes in endogenous POU4F2. Control and NAIP2 PSCs were treated with dox (condition 1), LDN (condition 2), or both (condition 3) and changes in tdTomato expression were evaluated by microscopy. After 1 week, control cells failed to express tdTomato, NAIP2 conditions 1 and 2 had low levels of tdTomato+ and NAIP2 condition 3, which received both LDN and dox, had the highest levels (~40%) of tdTomato+ cells with respect to total DAPI + cells (Fig. 2h<\/a>).<\/p>\n During RGC-iN formation many flat cells persisted after an initial burst of neurogenesis. We reasoned that cell heterogeneity was limiting conversion so we clonally selected PSCs to obtain PSC lines with greater differentiation potential. Each PSC clone was independently differentiated and the resulting morphologies were observed to range from enriched neurons (Supplementary Fig. 2a, d, e<\/a>) to unwanted flat cells (Supplementary Fig. 2b, c, f<\/a>). For all subsequent experiments, a PSC clone with a high conversion efficiency (NAIP2; Fig. 2a<\/a>) was used. To ensure that conversion was not cell-line dependent, we introduced TetO-NAIP2<\/i> into two additional PSC lines for a total of 3 genetic backgrounds. GM23720 iPSCs and WA09 ESCs were made with POU4F2-p2A-tdTomato and POU4F2-p2A-h2b-mNeonGreen reporters, respectively. In all NAIP2 lines, LDN\/dox induced efficient neural conversion (Fig. 2i\u2013n<\/a>). Further supporting the idea that NAIP2 was driving differentiation, Click-iT EdU (5-ethynyl-2\u2019-deoxyuridine) staining showed a reduced level of cell cycling beyond what was seen in NEUROG2 controls (Fig. 2o\u2013r<\/a>). Overall, this confirmed that the RGC-iN conversion triggered by NAIP2<\/i> led to early cell-cycle exit and differentiation and was possible across genetic backgrounds.<\/p>\n We also sought to explore how TFs alone or in combination influenced POU4F2-tdTomato reporter expression (Fig. 3a\u2013l<\/a>). Compared with controls that showed no fluorescence (Fig. 3a<\/a>), NAIP2 cells showed high levels of POU4F2-tdTomato+ (Fig. 3b<\/a>). NEUROG2<\/i> overexpressing neurons exhibited no POU4F2-tdTomato fluorescence at four days, however, a small population of POU4F2-tdTomato+ cells emerged by 6 days (Fig. 3c<\/a>; white arrows). This was not surprising since POU4F2 is also present in the developing forebrain and NEUROG2<\/i> induces forebrain-like conversion10<\/a><\/sup>. Though ATOH7<\/i> led to far fewer neurons, those that formed were POU4F2-tdTomato+ (Fig. 3d<\/a>). Flat ISL1<\/i> overexpressing cells were largely POU4F2-tdTomato negative (Fig. 3e<\/a>) and POU4F2<\/i> overexpressing cells were only weakly fluorescent (Fig. 3f<\/a>). This data suggested that while NEUROG2<\/i>, ATOH7,<\/i> and POU4F2<\/i> each contributed, their combined activity was far more effective in triggering neurite outgrowth and endogenous POU4F2-tdTomato gene expression.<\/p>\n a<\/b>\u2013l<\/b> Representative images of PSCs induced for 6 days with LDN\/dox and different TFs alone (N, A, I, P2) or in combination (NAIP2, NA, NAI, AI, NP, NAP2, IP2) with imaging for DAPI (left) and POU4F2-tdTomato+ (right). TdTomato panels are intentionally uniformly overexposed so that samples with weaker fluorescence (d<\/b>\u2013l<\/b>) could be detected. m<\/b> Percent of POU4F2-tdTomato+ cells on day 6 quantified as POU4F2+ cells relative to DAPI. For the WA09 background, this was determined by measuring h2b-mNeonGreen+ relative to DAPI. n<\/b> NAI iNs and o<\/b> NAIP2 RGC-iNs showing POU4F2-tdTomato expression (left) and TUJ1 staining (right). NAI cells were additionally labeled with DAPI. p<\/b> %TUJ1+ neurons and q<\/b> %TUJ1+\/POU4F2+ co-labeled RGC-iNs. r<\/b> A diagram summarizing the key events driving differentiation from immature PSCs to mature RGCs. White arrows=overlapping tdTomato+ signal; yellow arrows=absence of a signal. *P<\/i>\u2009<\u20090.05, **P<\/i>\u2009<\u20090.01, ***P<\/i>\u2009<\u20090.001, ****P<\/i>\u2009<\u20090.0001, ns P<\/i>\u2009>\u20090.05, n<\/i>\u2009=\u20093. Error bars are reported as standard deviation (SD). Scale a<\/b>\u2013l<\/b>\u2009=\u2009200\u2009\u00b5m; n<\/b>, o<\/b>\u2009=100\u2009\u00b5m.<\/p>\n<\/div>\n<\/div>\n Next, we evaluated how neural conversion was influenced by different binary and ternary TF combinations including NEUROG2<\/i>–ATOH7<\/i> (NA<\/i>), NEUROG2<\/i>–ATOH7<\/i>–ISL1<\/i> (NAI<\/i>), ATOH7<\/i>–ISL1<\/i> (AI<\/i>), NEUROG2<\/i>–POU4F2<\/i> (NP2<\/i>), NEUROG2<\/i>–ATOH7<\/i>–POU4F2<\/i> (NAP2<\/i>), and ISL1<\/i>–POU4F2<\/i> (IP2<\/i>) (Fig. 3g\u2013l<\/a>). In terms of morphology and POU4F2-tdTomato expression, NAIP2 was superior to all binary and ternary TF combinations. The next best was NA (Fig. 3g<\/a>) which had more POU4F2-tdTomato+ neurons than NEUROG2 but less than NAIP2. NAI (Fig. 3h<\/a>) produced fewer converted cells and lower levels of tdTomato than NA. Experiments performed with AI<\/i>, NP2 NAP2<\/i> and IP2<\/i> overexpression (Fig. 3i, j, k, l<\/a>), resembled ISL1 in that most cells were flat and POU4F2-tdTomato negative. Comparisons between different TF combinations (Fig. 3m<\/a>) also demonstrated that NAIP2 was the most efficient, with conversion efficiencies of 88%, 94%, and 90% in IMR90.4, GM23720, and WA09 PSCs, respectively. While both NAI and NAIP2 generated TUJ1+ neurons (Fig. 3n, o<\/a>) at efficiencies of 98% and 79%, respectively (Fig. 3p<\/a>), the number of POU4F2-tdTomato+ neurons was significantly higher for NAIP2 (88%) than for NAI (21%) (Fig. 3q<\/a>). This demonstrated that while NAI (and other TF combinations) are efficient in making neurons, NAIP2 is the most efficient in making POU4F2\u2009+\u2009RGC-like cells. A working model of RGC conversion highlights the cumulative importance of NAIP2 TFs for RGC conversion (Fig. 3r<\/a>).<\/p>\n To interrogate the transcriptional identities of induced neurons we carried out RNA-seq on PSCs (n<\/i>\u2009=\u20093), dox-treated CTL (n<\/i>\u2009=\u20095), and NEUROG2<\/i>, NA,<\/i> and NAIP2<\/i> expressing cells (n<\/i>\u2009=\u20094 each) at 1 week (Supplementary Table 1<\/a>). Principal component analysis (PCA) from DESeq2 demonstrated that biological replicates were consistent across experiments and clustered separately for PSCs, controls and RGC-iNs (Fig. 4a<\/a>-upper panel). NEUROG2, NA, NAIP2, and NAIP2-nc neurons were also clearly distinct from one another confirming that they had different gene expression profiles (Fig. 4a<\/a>-lower panel). Global population level differences were illustrated with a Venn diagram depicting the number of genes expressed in each group (normalized count >100) (Fig. 4b<\/a>). While there was much overlap, treatment groups were distinct from one another. Visualization of the top 500 genes in NAIP2 RGC-iNs (Fig. 4c<\/a>) revealed that NEUROG2, NA, and NAIP2 groups were more related than controls or PSCs. Zeocin-treated non-clonally selected NAIP2-nc cells consisted of flat cells and neurons while zeocin-treated and clonally selected NAIP2 cells were highly enriched for neurons. The gene expression differences observed in Fig. 4c<\/a> were consistent with these morphological characteristics. Recognizing the importance of clonal selection, all further analysis was conducted with clonal-selected NAIP2 cells. The top 50 highly differentially expressed genes (DEGs) in NAIP2 cells, including the 4 individual NAIP2<\/i> TFs, were further compared with publicly available single-cell sequencing data which showed a strong similarity to human RGCs as opposed to other retinal neurons (Supplementary Fig. 3a\u2013f<\/a>)23<\/a><\/sup>. We also used DAVID pathway analysis (Fig. 4d<\/a>) to identify biologically relevant signaling pathways in NAIP2 cells relative to control cells. UP_KEYWORDS and UP_SEQ_FEATUREs identified ligand-gated ion channels, voltage-gated ion channels, neurogenesis and potassium\/chloride channels, which are all important for neural cell identity and function. Likewise, KEGG_PATHWAYS identified calcium signaling, neuroactive ligand interactions and cAMP\/MAPK signaling. In all neurons (NAIP2, NA and NEUROG2) pan-neuronal markers (e.g., RBFOX3, TUBB3, SYT1, DLG4<\/i>, and SYP<\/i>) were detected (Fig. 4e<\/a>). By contrast, NAIP2 samples expressed higher levels of RGC enriched genes (e.g., POU4F1<\/i>, POU6F2, NRN1, SYT13<\/i>, and NEFL<\/i>), thus supporting their RGC identity. Likewise, we performed RNA-seq on WA09 and GM23720 genetic background NAIP2 cells which similarly expressed RGC genes (Supplementary Fig. 4<\/a>).<\/p>\n a<\/b> Principal Component Analysis (PCA) plot of day 6 CTL, NAIP2 and PSC (top) and NEUROG2, NA, NAIP2-nc and NAIP2 cells (bottom). b<\/b> Venn diagram of the number of genes expressed in each sample (normalized count >10). c<\/b> Heatmap of the top 500 genes expressed in the NAIP2 samples. d<\/b> Bar graph depicting \u2212log10(FDR) of significantly differentially expressed pathways between NAIP2 and CTL samples identified by DAVID in Uniprot (top) and KEGG Gene Ontology (bottom) databases. e<\/b> Scatterplots showing normalized counts of pan-neuronal and RGC marker genes in each treatment group. f<\/b> Heatmaps of genes expressed within the neurogenesis (left), axon guidance (center), and growth factor (right) pathways. g<\/b> Volcano plot highlighting log2(FC) gene expression between the CTL and NAIP2 samples plotted with respect to \u2212log10(FDR). Immunostaining for VIM and GLAST in control cells (left) and PAX6, ISL1, POU4F1, and MAP2 in NAIP2 samples (right). FDR = false discovery rate. FC = fold change. Scale g<\/b>\u2009=\u200985\u2009\u00b5m. NAIP2-nc = RGC-iNs from non-clonally selected PSCs.<\/p>\n<\/div>\n<\/div>\n During early RGC growth, there is an active process of axon outgrowth. Both ISLR2<\/i> and ELAVL3<\/i> (Fig. 4f<\/a>), which are involved in RGC axonogenesis, were expressed in NAIP2 cells but not in PSCs or controls24<\/a><\/sup>. Functional guidance cues also steer axons towards\/away from cellular targets. While some ephrin ligands involved in cell-to-cell communication were highly expressed in non-neuronal controls (e.g., EFNB1<\/i>, EFNB2<\/i>, EPHA2<\/i>) many were upregulated in neurons (e.g., EFNA3<\/i>, EFNB3<\/i>, EPHA5<\/i>, and EPHB2)<\/i>, and in RGC-iNs their expression increased from 1 to 4 weeks (Supplementary Fig. 5a, b<\/a>). ROBO2\/3 and DCC receptors, which mediate RGC axonogenesis, were present in neurons at 1 week (Supplementary Fig. 5c, d<\/a>)25<\/a><\/sup>. While <\/a><\/div>\n
BMP inhibition and TF overexpression synergistically enhance the morphological conversion of iNs<\/h3>\n
<\/a><\/div>\n
Endogenous POU4F2 reporter expression accompanies RGC conversion<\/h3>\n
Individual transcription factors induce partial reprogramming<\/h3>\n
<\/a><\/div>\n
NAIP2 induced RGC-iNs exhibit RGC-like transcriptional profiles<\/h3>\n
<\/a><\/div>\n