ABSTRACT
The proneural transcription factor Achaete-scute complex-like 1 (Ascl1) is a major regulator of neural progenitor fate, implicated both in neurogenesis and oligodendrogliogenesis. Ascl1 has been widely used to reprogram non-neuronal cells into induced neurons. In vitro, Ascl1 induces efficient reprogramming of proliferative astroglia from the early postnatal cerebral cortex into interneuron-like cells. Here, we examined whether Ascl1 can similarly induce neuronal reprogramming of glia undergoing proliferation in the postnatal mouse cerebral cortex in vivo. Toward this, we targeted cortical glia at the peak of proliferative expansion (i.e., postnatal day 5) by injecting a retrovirus encoding for Ascl1 into the mouse cerebral cortex. In sharp contrast to the very efficient reprogramming in vitro, Ascl1-transduced glial cells were converted into doublecortin-immunoreactive neurons only with low efficiency in vivo. Interfering with the phosphorylation of Ascl1 by mutation of six conserved proline-directed serine/threonine phosphorylation sites (Ascl1SA6) has been previously shown to increase its neurogenic activity in the early embryonic cerebral cortex. We therefore tested whether transduction of proliferative glia with a retrovirus encoding Ascl1SA6 improved their conversion into neurons. While in vitro glia-to-neuron conversion was markedly enhanced, in vivo reprogramming efficiency remained low. However, both wild-type and mutant Ascl1 reduced the relative number of cells expressing the astrocytic marker glial fibrillary acidic protein (GFAP) and increased the relative number of cells expressing the oligodendroglial marker Sox10 in vivo. Together, our results indicate that the enhanced neurogenic response of proliferative postnatal glia to Ascl1SA6 versus Ascl1 observed in vitro is not recapitulated in vivo.
INTRODUCTION
The postnatal mammalian brain is largely devoid of persistent neurogenesis, except from specialized niches such as the subependymal zone of the lateral ventricle and the subgranular zone of the dentate gyrus (Denoth-Lippuner and Jessberger, 2021). In all other brain regions, neurons lost due to disease or injury cannot be replaced, resulting in irreversible circuit dysfunction and functional impairments. Harnessing the neurogenic potential of glia to produce new neurons by direct lineage reprogramming has emerged as an approach for potential repair of diseased circuits in non-neurogenic brain areas such as the cerebral cortex (Peron and Berninger, 2015).
The basic helix-loop-helix (bHLH) transcription factor Ascl1 directly transactivates target genes and thereby orchestrates multiple aspects of cortical development including cellular proliferation, cell cycle exit, and neural differentiation (Castro et al., 2011; Guillemot and Hassan, 2017). Notably, Ascl1 controls GABAergic neurogenesis by regulating expression of homeobox genes of the distal-less gene family (Dlx genes) in progenitors of the ventral telencephalon (Casarosa et al., 1999; Poitras et al., 2007). We previously demonstrated that expression of Ascl1 in mouse postnatal cortical astrocytes in vitro was sufficient to reprogram them into functional neurons endowed with GABAergic identity (Berninger et al., 2007; Heinrich et al., 2010). Remarkably, Ascl1 can also reprogram cultured cells of human origin, including fibroblasts and pericytes, into neurons in vitro (Karow et al., 2012; Chanda et al., 2014). This raises the question whether it can also induce a neurogenic fate in vivo. For instance, it remains unknown whether glia of the cortical parenchyma can be reprogrammed into neurons in vivo with similar efficiency as in vitro when forced to express Ascl1 during their proliferative expansion, which peaks around postnatal day 5 (Ge et al., 2012).
Ascl1 function is tightly regulated by post-translational modifications, including phosphorylation (Dennis et al., 2019), which ultimately affects cell fate decisions. Notably, increased RAS/ERK signaling diverts Ascl1 from its neurogenic role and promotes a proliferative glial program (Li et al., 2014). bHLH transcription factors share an evolutionarily conserved serine/threonine phosphorylation residue in the L-H2 junction of the bHLH domain (Quan et al., 2016), but also harbor unique phosphorylation sites outside of this domain (Guillemot and Hassan, 2017). Most notably, similar to other bHLH transcription factors (Oproescu et al., 2021), Ascl1 is regulated by proline-directed serine threonine kinases, such as ERK (Li et al., 2014). Remarkably, preventing phosphorylation-dependent regulation of Ascl1 activity by mutating all six serines of the conserved serine-proline (SP) phospho-sites to alanine, a mutation here referred to as Ascl1SA6, has been found to increase its neurogenic activity in the embryonic day (E) 12.5 cerebral cortex (Li et al., 2014). This finding led us to hypothesize that using the Ascl1SA6 mutant variant could promote glia-to-neuron conversion both in vitro and in vivo.
Consistent with this hypothesis, our results show that Ascl1SA6 is more efficient than Ascl1 in converting postnatal cortical glia into neurons in vitro. However, Ascl1 and Ascl1SA6 had only limited reprogramming efficiency in vivo. Instead, we observed a reduction in the relative number of transduced cells expressing the astrocytic marker GFAP and a concomitant increase in the relative number of cells expressing the oligodendroglial lineage marker Sox10. This data suggests that, irrespective of its phosphorylation state, Ascl1 may preferentially promote an oligodendrogliogenic fate in proliferative postnatal cortical glia in vivo.
MATERIALS AND METHODS
Cell Culture
Postnatal cortical astrocytes were isolated from cortices of C57Bl6/J mice between postnatal day 5-7 days (P5-7), which were obtained from the Translational Animal Research Center of the University Medical Center Mainz.P5-P7 astrocytes were cultured as previously described (Heinrich et al., 2011; Sharif et al., 2021). Briefly, after isolation, cells were expanded for 7-10 days in Astromedium: Dulbecco’s Modified Eagles Medium, Nutrient Mixture F12 (DMEM/F12, Gibco, 21331-020); 10% Fetal Bovine Serum (FBS, Invitrogen, 10270-106); 5% Horse Serum (Invitrogen, 16050-130); 1x Penicillin/Streptomycin (Invitrogen, 15140122); 1x L-GlutaMAX Supplement (Invitrogen, 35050-0380); 1x B27 Supplement (Invitrogen, 17504001); and supplemented with 10ng/µl Epidermal Growth Factor (EGF; Peprotech, AF-100-15) and 10 ng/µl basic-Fibroblast Growth Factor (FGF-2; Peprotech, 100-18B). Cells were incubated at 37°C in 5% CO2. When cells reached 70-80% confluency, cells were detached with 0.05% Trypsin EDTA (Life Technologies, 15400054) for 5 min at 37°C. Cells were subsequently seeded onto poly-D-lysine hydrobromide-coated (PDL; Sigma, P0899) glass coverslips (12mm, Menzel-Gläser, 631-0713) in 24-well plates at a density of 50000-80000 cells/well in 500 µl Astromedium supplemented with 10 ng/µl EGF and 10 ng/µl FGF-2.
Plasmids and retroviruses
Moloney Murine Leukaemia Virus (MMLV)-based retroviral vectors (Heinrich et al., 2011) were used to express Ascl1 and Ascl1SA6 under control of the chicken β-actin promoter with a cytomegalovirus enhancer (pCAG). A GFP or DsRed reporter was cloned in behind an Internal Ribosome Entry Site (IRES). To generate the pCAG-Ascl1-IRES-DsRed/GFP and pCAG-Ascl1SA6-IRES-DsRed/GFP retroviral constructs, a cassette containing the coding sequences flanked by attL recombination sites was generated through the excision of the coding sequences for Ascl1 and Ascl1SA6 from the pCIG2 parental vectors (Li et al., 2014) via XhoI/SalI double restriction. Isolated fragments were inserted into the pENTRY1A Dual Selection (Invitrogen) intermediate vector linearized via SalI. The final retroviral constructs were subsequently obtained via recombination catalyzed by the LR Clonase II (Invitrogen, 11791020), which substituted the ccdB cassette in the destination vector pCAG-ccdB-IRES-DsRed or pCAG-ccdB-IRES-GFP with Ascl1 or Ascl1SA6 coding sequences. Transduction with MMLV-based retroviral vectors encoding only the fluorescent protein GFP or DsRed behind an IRES under control of pCAG promoter (pCAG-IRES-DsRed/pCAG-IRES-GFP) (Heinrich et al., 2011) was used for control experiments. Viral particles were produced using gpg helper free packaging cells to generate Vesicular Stomatitis Virus Glycoprotein (VSV-G)-pseudotyped retroviral particles (Ory et al., 1996). Viral stocks were titrated by transduction of HEK293 cultures. Viral titers used were in the range of 107 TU/ml.
Retroviral transduction
After seeding the cells and letting them attach for 4h in the incubator, cells were transduced with 1 µl retrovirus/well and incubated at 37°C in 8% CO2. One day later, treated medium was removed and substituted with 500 µl of B27 Differentiation Medium: DMEM/F12; 1x Penicillin/Streptomycin; 1x L-GlutaMAX Supplement; 1x B27 Supplement. Cells were treated again with 1 µl/well of retrovirus. One day later, the culture volume was brought to 1 ml/well with fresh B27 Differentiation Medium. Cells were kept in culture for a total of 7 days in vitro before fixation for immunocytochemical analyses.
Immunocytochemistry
Cells were fixed with 4% paraformaldehyde (PFA, Sigma, P6148) for 10-15 min and washed 3 times with 1xPBS (Gibco, 70013-016) before storage at 4°C. Washed cells were first incubated for 1 h at room temperature (RT) with blocking solution (3% bovine serum albumin [BSA, Sigma, A7906] and 0.5% Triton X-100 [Sigma, X100] in 1xPBS) and then with primary antibodies for 2-3 h at RT. After 3 washes with 1xPBS, cells were incubated with secondary antibodies for 1h at RT. Cells were then counterstained with DAPI (Sigma, D8417) diluted 1:1000 in blocking solution, then washed 3 time in 1xPBS before being mounted with Aqua Polymount (Polysciences, 18606-20). The following primary antibodies were used: β-Tubulin III (Mouse IgG2b, 1:1000, Sigma, T8660); Green Fluorescent Protein (GFP, Chicken, 1:300, AvesLab, GFP-1020); GFAP (rabbit, 1:1000, Dako, Z0334); Red Fluorescent Protein (RFP, rat, 1:400, Chromoteck, 5F8). Secondary antibodies were diluted 1:1000 and were conjugated to: A488 anti-chicken (donkey, Jackson Immunoresearch, 703-545-155); Cy3 anti-mouse (goat, Dianova, 115-165-166); Cy3 anti-rat (goat, Dianova, 112-165-167); Cy5 anti-rabbit (goat, Dianova, 111-175-144).
Animals and Animal Procedures
The study was performed in accordance with the guidelines of the German Animal Welfare Act and the European Directive 2010/63/EU for the protection of animals used for scientific purposes and was approved by the Rhineland-Palatinate State Authority (permit number 23 177 07-G15-1-031). For retroviral injections, male and female C57Bl6/J pups kept with their mother were purchased from Janvier Labs. Mice were kept in a 12:12 h light-dark cycle in Polycarbonate Type II cages (350 cm2). Animals were provided with food and water ad libitum and all efforts were made to reduce the number of animals and their suffering. Before the surgery, animals received a subcutaneous injection of Carprofen (Rimadyl®, Zoetis, 4 mg/kg of body weight, in 0.9% NaCl [Amresco]). Anaesthesia was induced by intraperitoneal (i.p.) injection of a solution of 0.5 mg/kg Medetomidin (Pfizer), 5 mg/kg Midazolam (Hameln) and 0.025 mg/kg Fentanyl (Albrecht) in 0.9% NaCl. Viruses were injected in the cerebral cortex using glass capillaries (Hirschmann, 9600105) pulled to obtain a 20 µm tip diameter. Briefly, a small incision was made on the skin with a surgical blade and the skull was carefully opened with a needle. Each pup received a volume of 0.5-1 μl of retroviral suspension targeted to the somatosensory and visual cortical areas. After injection, the wound was closed with surgical glue (3M Vetbond, NC0304169) and anesthesia was terminated by i.p. injection of a solution of 2.5 mg/kg Atipamezol (Pfizer), 0.5 mg/kg Flumazenil (Hameln) and 0.1 mg/Kg Buprenorphin (RB Pharmaceutials) in 0.9% NaCl. Pups were left to recover on a warm plate (37°C) before returning them to their mother. Recovery state was checked daily for a week after the surgery.
Tissue preparation and immunohistochemistry
Animals were lethally anesthetized with a solution of 120 mg/kg Ketamine (Zoetis) and 16 mg/kg Xylazine (Bayer) (in 0.9% NaCl, i.p.) and transcardiacally perfused with pre-warmed 0.9% NaCl followed by ice-cold 4% paraformaldehyde (PFA, Sigma, P6148). The brains were harvested and post-fixed for 2 h to overnight in 4% PFA at 4°C. Then, 40 μm thick coronal sections were prepared using a vibratome (Microm HM650V, Thermo Scientific) and stored at -20°C in a cryoprotective solution (20% glucose [Sigma, G8270], 40% ethylene glycol [Sigma, 324558], 0.025% sodium azide [Sigma, S2202], in 0.5 X phosphate buffer [15mM Na2HPO4·12H2O [Merck, 10039-32-4]; 16mM NaH2PO4 ·2H2O [Merck, 13472-35-0]; pH 7.4]).
For immunohistochemistry, brain sections were washed three times for 15 min with 1X TBS (50mM Tris [Invitrogen, 15504-020]; 150 mM NaCl [Amresco, 0241]; pH7.6) and then incubated for 1.5 h in blocking solution: 5% Donkey Serum (Sigma, S30); 0.3% Triton X-100; 1X TBS. Slices were then incubated with primary antibodies diluted in blocking solution for 2-3 h at RT, followed by an overnight incubation at 4°C. After three washing steps with 1X TBS, slices were incubated with secondary antibodies diluted blocking solution for 1 h at RT. Slices were washed twice with 1X TBS, incubated with DAPI dissolved in 1X TBS for 5 min at RT and washed three times with 1X TBS. For mounting, slices were washed two times with 1X Phosphate Buffer (30 mM Na2HPO4·12H2O [Merck, 10039-32-4]; 33 mM NaH2PO4 ·2H2O [Merck, 13472-35-0]; pH 7.4) and were dried on Superfrost (Thermo Fisher Scientific) microscope slides. Sections were further dehydrated with toluene and covered with cover-glasses mounted with DPX mountant for histology (Sigma, 06522) or directly mounted with Prolong™Gold (Invitrogen, P36930). The following primary antibodies were used: Achaete scute-like1 (Ascl1, mouse IgG1, 1:400, BD Pharmingen, 556604); Doublecortin (DCX, goat, 1:250, Santa Cruz Biotechnology, sc-8066); Green Fluorescent Protein (GFP, chicken, 1:1000, AvesLab, GFP-1020); Glial Fibrillary Acidic Protein (GFAP, rabbit, 1:300, Dako, Z0334); Ionized calcium-binding adapter molecule 1 (Iba1, rabbit, 1:800, Wako, 16A11); mCherry (chicken, 1:300, EnCor Biotechnology, CPCA-mCherry); Red Fluorescent Protein (RFP, rabbit, 1:500, Biomol, 600401379S); SRY-Box 10 (Sox10, goat, 1:100, Santa Cruz Biotechnology, sc-17342). Secondary antibodies were made in donkey and conjugated with: A488 (anti-chicken, 1:200, Jackson Immunoresearch, 703-545-155); A488 (anti-rabbit, 1:200, Invitrogen, A21206); A647 (anti-rabbit, 1:500, Invitrogen, A31573); A488 (anti-mouse, 1:200, Invitrogen, A21202); A647 (anti-mouse, 1:500, Invitrogen, A31571); Cy3 (anti-chicken, 1:500, Dianova, 703-165-155); Cy3 (anti-goat, 1:500, Dianova, 705-165-147); Cy3 (anti-mouse, 1:500, Invitrogen, A10037); Cy3 (anti-rabbit, 1:500, Dianova, 711-165-152); Cy5 (anti-goat, 1:500, Dianova, 705-175-147).
Imaging and data analysis
Images were acquired using a TCS SP5 (Leica) confocal microscope (Institute of Molecular Biology, Mainz, Germany) equipped with four PMTs, four lasers (405 Diode, Argon, HeNe 543, HeNe 633) and a fast-resonant scanner. Images were taken with a 20x dry objective (NA 0.7) or a 40x oil objective (NA 1.3). For imaging of brain sections, serial Z-stacks spaced at 0.3 μm-1.25 μm distance were acquired to image the whole thickness of the section. Alternatively, imaging was performed using an Axio Imager.M2 fluorescent microscope equipped with an ApoTome (Zeiss) at a 20x dry objective (NA 0.7) or a 63x oil objective (NA 1.25).
For in vitro experiments, biological replicates (n) were obtained from independent cultures prepared from different animals. For each n, the value corresponds to the mean value of two technical replicates (i.e., two coverslips). Cell quantifications were performed on 4×4 tile scans (individual tile size: 624,70×501,22 μm). For in vivo experiments, n corresponds to the number of animals. Quantifications were performed on equally spaced sections (240 or 480 µm) covering the whole area with transduced cells. Tile scans were acquired with a serial Z-stack spaced at 1.25 μm distance.
For images used for illustration, the color balance of each channel was uniformly adjusted in Photoshop (Adobe). If necessary, Lookup Tables were changed to maintain uniformity of color coding within figures. When appropriate, a median filter (despeckle) was applied in Fiji to pictures presenting salt-and-pepper noise, and noise was filtered via removal of outlier pixels.
Multiple sequence alignment of the Ascl1 and Ascl1SA6 protein sequences was performed in Clustal Omega (RRID:SCR_001591).
Statistical analysis
The number of independent experiments (n) and number of cells analyzed are reported in the main text or figure legends. Data are represented as means ± SD. Statistical analysis was performed in SPSS Statistics 23 V5 (IBM). Normality of distribution was assessed using Shapiro-Wilk test and the significance of the differences between groups was analyzed by One-Way ANOVA followed by Bonferroni post-hoc test. P-values are indicated in the figures or figure legends. Graphs were prepared in GraphPad Prism 5.
RESULTS
Ascl1SA6 improves neuronal reprogramming efficiency from cultured postnatal cortical astroglia compared to wild-type Ascl1
Our earlier work showed that Ascl1 can reprogram cultured postnatal astroglia into neurons (Berninger et al., 2007; Heinrich et al., 2010). More recently, overexpression of Ascl1SA6 in embryonic cortical progenitors was found to enhance neuronal differentiation compared with wild-type Ascl1 (Li et al., 2014). Here, we examined whether Ascl1SA6 can increase the glia-to-neuron reprogramming capacity of Ascl1 in postnatal astroglial cultures. For this purpose, we first cloned murine Ascl1 and Ascl1SA6 sequences into retroviral vectors for transduction of proliferative glia. Figure 1A depicts the six serine-to-alanine (SA6) substitutions resulting from the targeted mutation of the Ascl1 coding sequence. Astroglial cultures prepared from P5 mice were transduced with retroviruses encoding for Ascl1 or Ascl1SA6 together with a reporter gene (GFP or DsRed). A retrovirus encoding only a reporter gene was used as control (Figure 1B). Neuronal reprogramming efficiency was evaluated seven days after transduction by immunocytochemistry directed against the neuronal marker β-Tubulin III. After transduction with control virus, virtually no β-Tubulin III-immunoreactive cells were found (0.1±0.2%, 1398 transduced cells analysed, n=3 biological replicates; Figure 1C,D). In contrast, consistent with our previous findings (Berninger et al., 2007; Heinrich et al., 2010), astrocytes transduced with Ascl1 acquired a neuronal-like elongated morphology and expressed β-Tubulin III (27.3±3.8%, 3061 transduced cells analysed, n=4 biological replicates; Figure 1C,D). Strikingly, the proportion of converted cells doubled upon overexpression of Ascl1SA6 (51.1±7.0%, 3462 transduced cells analysed, n=3 biological replicates; Figure 1C,D).
These results indicate an increased neurogenic potential of Ascl1SA6 in glia-to-neuron conversion when expressed in postnatal astrocytes in vitro.
Ascl1 or Ascl1SA6 converts postnatal cortical glia into neurons with low efficiency in vivo
We next tested whether, like in vitro, proliferative cortical glia can be efficiently reprogrammed towards a neuronal fate in vivo. Cortical glia greatly expands during the first postnatal week by local proliferation (Ge et al., 2012; Clavreul et al., 2019), To target proliferative glia, retroviruses were injected into the cerebral cortex at postnatal day five (P5). We then analyzed the identity of the transduced cells by immunohistochemical analysis at three days post injection (3 dpi), first with the control virus only (Figure 2A). We found that virtually all transduced cells were immunopositive for glial markers (Figure 2B). The majority of transduced cells were immunoreactive for the astroglial marker GFAP (62.8±8.1%, 753 transduced cells analysed, n=3 mice; Figure 2B,C), and the remaining were oligodendroglial cells immunoreactive for Sox10 (32.3±6.1%, 753 transduced cells analysed, n=3 mice; Figure 2B,C). Rarely, we found transduced cells immunoreactive for the microglial marker Iba1 (1.0±0.9%, 578 transduced cells analysed, n=3 mice; Figure 2B,D). Importantly, none of the control-transduced cell expressed the immature neuronal marker DCX (0.0±0.0%, 578 transduced cells analysed, n=3 mice; Figure 2B,D). These results indicate that retroviruses injected in the P5 mouse cerebral cortex in vivo specifically transduce astroglial and oligodendroglial lineage cells.
We next injected control, Ascl1- or Ascl1SA6-encoding retroviruses and investigated whether these bHLH genes could reprogram P5 proliferative glia into neurons by immunohistochemical analysis at 12 dpi. Ascl1 was effectively expressed in cells transduced with Ascl1 and Ascl1SA6, while absent from control-transduced cells (Figure 3A). Control-transduced cells lacked DCX expression (0.0±0.0%, 2157 transduced cells analysed, n=3 mice; Figure 3B,C), confirming that the control vector did not induce a cell fate switch. In contrast to our findings in vitro, Ascl1- and Ascl1SA6-transduced cells largely remained immunonegative for DCX (Figure 3B), with only a small number of cells exhibiting an immature neuron-like morphology and expressing DCX (Ascl1: 4.6±1.6%, 720 transduced cells analysed, n=3 mice, and Ascl1SA6: 6.9±0.2%, 409 transduced cells analysed, n=3 mice) (Figure 3B,C). Together, our results indicate that despite the strong neurogenic potential of Ascl1 and Ascl1SA6 in vitro (Figure 2), these bHLH genes can only reprogram postnatal cortical glia into neurons with low efficiency in vivo.
Ascl1 expression in postnatal cortical glia increases the relative number of cells expressing oligodendroglial markers
Given that only a few Ascl1 or Ascl1SA6 transduced cells were converted into neurons, we examined whether the remaining cells nevertheless had responded to the reprogramming factors and downregulated glial markers. We therefore analyzed the expression of the pan-oligodendroglial marker Sox10 and the astroglial marker GFAP in Ascl1- and Ascl1SA6-transduced cells (Figure 4A-C). Consistent with our analysis at 3 dpi (Figure 2), control-transduced cells at 12 dpi were glial cells, predominantly astrocytes, with two thirds of the cells expressing GFAP (63.3 ± 11.3%, 1885 transduced cells analysed, n=3 mice) and another third expressing Sox10 (35.6±12.1%, 1885 transduced cells analysed, n=3 mice; Figure 4A,C). As expected, the expression of GFAP and Sox10 was mutually exclusive in control-transduced cells (0.3±0.6% of GFAP/Sox10-positive cells, 1885 transduced cells analysed, n=3 mice; Figure 4B,C and Supplementary Movie 1). Following transduction with both Ascl1 variants, we observed a marked alteration in the expression of glial markers. Strikingly, only one fifth of transduced cells expressed exclusively GFAP (Ascl1, 18.7±3.1%, 848 transduced cells analysed, n=4 mice; Ascl1SA6, 20.4±6.1%, 573 transduced cells analysed, n=3 mice). Interestingly, in Ascl1-transduced cells, the reduction in GFAP expression was concomitant with a large increase in the relative number of Sox10-only expressing cells (70.0±7.7%, 848 transduced cells analysed, n=4 mice; Figure 4C). The same trend was observed following Ascl1SA6 overexpression (50.7±3.1%, 573 transduced cells analysed, n=3 mice; Figure 4C), albeit without reaching statistical significance. Instead, a significant increase in the relative number of cells co-expressing Sox10 and GFAP was observed in Ascl1SA6-transduced cells (Figure 4B,C and Supplementary Movie 3). The detection of GFAP/Sox10-immunopositive cells following transduction with both Ascl1 variants (Ascl1, 4.5±2.6%, 848 transduced cells analysed, n=4 mice; Ascl1SA6, 17.1±6.0%, 573 transduced cells analysed, n=3 mice, Figure 4B,C and Supplementary Movies 2 and 3) may capture cells in a “mixed” glial state. These results indicate that although largely failing to redirect towards neurogenesis, proliferative glial cells appear to be responsive to Ascl1 or Ascl1SA6 by turning on Sox10 expression.
DISCUSSION
In the present study, we assessed potential reprogramming of glia during their proliferative expansion in the early postnatal cerebral cortex by overexpression of either wild-type Ascl1 or a mutant variant, in which the six conserved serine-proline motifs located outside of the bHLH domain had been mutated (Ascl1SA6), thereby rendering Ascl1 unresponsive to regulation by phosphorylation (Li et al., 2014). We provide evidence that Ascl1SA6 is more efficient than wild-type Ascl1 in converting postnatal astroglia into neurons in vitro. Furthermore, we show that the reprogramming efficiency of both Ascl1 and Ascl1SA6 in vivo is surprisingly low in the early postnatal cortex compared to the results obtained in vitro. Interestingly, while only few Ascl1-transduced cells converted into neurons, we observed a relative shift from GFAP-positive cells to Sox10-positive cells, suggesting an increase in the number of cells of the oligodendroglial lineage at the expense of astroglia.
Our results indicate that Ascl1 reprograms proliferative postnatal cortical glia into neurons with low efficiency in vivo. This is in agreement with previous studies reporting inefficient neuronal reprogramming following retrovirus- or lentivirus-mediated expression of Ascl1 in reactive glia in the adult lesioned cortex (Heinrich et al., 2014), adult striatum (Niu et al., 2015) and adult lesioned spinal cord (Su et al., 2014). In contrast to these findings, another study reported very efficient reprogramming of glia into mature neurons following adeno-associated virus (AAV)-mediated expression of Ascl1 in the dorsal midbrain, striatum and somatosensory cortex (Liu et al., 2015). However, misidentification of endogenous neurons as glia-derived neurons was recently reported following AAV-mediated expression of Neurod1, possibly due to transgene sequence-specific effects in cis (Wang et al., 2021). Thus, one possible explanation for the apparent discrepancy with regard to the efficiency of Ascl1 to induce reprogramming in vivo is that AAV-mediated expression of Ascl1, similarly to Neurod1, resulted in labelling of endogenous neurons. Future studies combining AAV-mediated expression of reprogramming factors such as Ascl1 with genetic lineage tracing are required to clarify the origin of seemingly induced neurons (Wang et al., 2021; Leaman et al., 2022).
The apparent difference in reprogramming potency of Ascl1 in vitro and in vivo could be attributed to various factors. 1) Enhanced intrinsic cell plasticity of cultured astrocytes as compared to astrocytes in vivo despite both being in a similar proliferative state. The protocol employed here to culture and reprogram astrocytes may enhance their competence to undergo cell fate conversion. Indeed, a previous study showed that allowing these astrocytes to mature in vitro even only for few days prior to proneural factor activation resulted in a drastic decrease in reprogramming rate, an effect that could be attributed to activation of the REST/coREST repressor complex and accompanying epigenetic maturation (Masserdotti et al., 2015). In vivo, REST/coREST complex activity may be already higher, thereby safeguarding glial identity against Ascl1-induced neurogenic reprogramming. 2) Another important difference obviously consists in considerably more complex local environment in which these glial cells find themselves in vivo. Nearly nothing is known about the influence that other cell types exert on cells that successfully undergo reprogramming or fail to do so in vivo. However, in vitro studies have shown that human pericytes undergoing reprogramming by Ascl1 and Sox2 pass through a neural stem cell-like stage during which they become responsive to several intercellular signaling pathways including Notch signaling (Karow et al., 2018). Thus, it is conceivable that signaling molecules as well as extracellular matrix components secreted by cells within the local environment could impinge on early and perhaps more vulnerable reprogramming stages, thereby curtailing progression towards neurogenesis.
Our retroviral vectors were found to target proliferative cells of both the astroglial and oligodendroglial lineage. The overall very low conversion efficiency suggests that not only astroglia, but also cells of the oligodendroglial lineage possess effective safeguarding mechanisms that protect against acquiring a neurogenic fate. In fact, these safeguarding mechanisms are effective even when confronted with a powerful transcription factor with pioneer factor activity, such as Ascl1 (Wapinski et al., 2013; Raposo et al., 2015; Park et al., 2017) or a mutant variant with even increased neurogenic capacity (Woods et al., 2022). Thus, despite being in a proliferative state, astroglia and oligodendrocyte progenitor cells could be potentially less plastic than their adult reactive counterparts in the injured adult brain (Sirko et al., 2013; Magnusson et al., 2014; Faiz et al., 2015; Nato et al., 2015).
While Ascl1 did not induce neurogenic conversion in cells of the astroglial and oligodendroglial lineages, we observed a significant shift in the ratio of virus-transduced astroglial to oligodendroglial cells. Intriguingly, the same shift was observed when using the more neurogenic mutant Ascl1SA6. Several mechanisms could account for that: enhanced expansion of the oligodendroglial cells by activating Ascl1-mediated proliferative programs (Castro et al., 2011); conversely, enhanced cell cycle exit of astrocytes expressing Ascl1; enhanced or reduced survival of oligodendroglial or astroglial lineage cells, respectively; finally, conversion of astroglial cells towards an oligodendroglial fate. The latter would be consistent with the fact that Ascl1 is known to contribute to oligodendrogliogenesis in the developing and adult brain (Parras et al., 2004; Parras et al., 2007). Moreover, studies in the adult hippocampus have previously shown that similar retroviral expression of Ascl1 in neural stem cells, contrary to expectation, promoted oligodendrogliogenesis instead of GABAergic neurogenesis (Jessberger et al., 2008; Braun et al., 2015). Intriguingly, if this latter scenario is indeed the case, the fact that Ascl1SA6 caused a similar shift as wildtype Ascl1 seems to preclude an interpretation according to which acquisition of an oligodendroglial fate would require Ascl1 to be phosphorylated. One line of evidence arguing for at least some oligodendrogliogenic reprogramming by Ascl1 consists in the fact that both variants caused the emergence of GFAP and Sox10 double-positive cells, which were not observed in control virus transduced brains, potentially hinting at an intermediate state between the two glial lineages. Be that as it may, future studies will be needed to distinguish between the mutually non-exclusive mechanisms of action delineated here.
In sum, our study reveals that proliferative glia in the healthy postnatal cerebral cortex are safeguarded against potential neurogenic fate conversion induced by pioneer transcription factors such as Ascl1. Further work will be needed to assess whether additional factors synergizing with Ascl1, such as Sox2 (Heinrich et al., 2014), Dxl2 (Lentini et al., 2021) or Bcl2 (Gascon et al., 2016) could help overcoming these potent safeguarding mechanisms in vivo.
Author Contributions
Methodology, investigation and formal analysis C.G., N.M., S.P.; Writing – Original Draft, C.G., N.M., S.P.; Funding Acquisition, Conceptualization, Visualization, Writing – Review and Editing, C.S., B.B., S.P. All authors contributed to the article and approved the submitted version.
Funding
This study was supported by grants of the German Research Foundation (BE 4182/7-1; CRC1080, project number 221828878) and Wellcome Trust (206410/Z/17/Z) to B.B. and by the Inneruniversitäre Forschungsförderung Stufe I of the Johannes Gutenberg University Mainz to S.P.; by core funding to the Francis Crick Institute from Cancer Research UK, The Medical Research Council, and the Wellcome Trust (FC001002); N.M. was supported by a fellowship from the Human Frontiers Science Program (HFSP Long-Term Fellowship, LT000646/2015).
Data availability statement
The original contributions presented in the study are included in the article/Material, further inquiries can be directed to the corresponding author.
Acknowledgments
We are grateful to the members of the Berninger laboratory for their helpful comments and critical feedback over the course of this study. We acknowledge support by the Microscopy Core Facility of the Institute of Molecular Biology (IMB) Mainz.