Induction of Neuronal Cells in the Injured Adult Cerebral Cortex upon Forced Expression of Sox2 and Ascl1
(A) Schematic diagram of experimental procedures. The red bar shows the stab wound lesion restricted to the upper layers of the cortex. The red dashed line shows the area where cells transduced by retroviral vectors (RVs) are distributed. dpi, days postinjection; IHC, immunohistochemistry; mo, months.
(B and C) Absence of spontaneous neurogenesis in the injured adult cortex in control conditions. (B) The micrograph depicts DSRED⁺ cells following injection of a control retrovirus encoding Dsred only (pCAG-IRES-Dsred) at 10 dpi. Most transduced cells exhibit a glial-like morphology. (C) Micrograph of the same field of view as shown in (B), revealing that neither transduced nor untransduced cells express DCX (white). The white dashed line marks the cortical surface.
(D–F) Virtual absence of induced neurogenesis in the injured adult cortex following forced expression of Ascl1. (D) The micrograph depicts DSRED⁺ cells transduced by the retrovirus encoding Ascl1 at 11 dpi (Ascl1-IRES-Dsred). Note the glial-like morphology of most transduced cells. (E) Micrograph of the same field of view as shown in (D), revealing that DSRED⁺ transduced cells do not express DCX (white). (F) High-magnification view of the area boxed in (D) and (E), showing double immunostaining for DSRED and DCX (white).
(G–L) Forced expression of Sox2 and Ascl1 induces neurogenesis in the injured adult cortex. (G) Triple immunostaining for DSRED, GFP, and DCX reveals appearance of numerous induced neuronal cells expressing DCX (white) in the injured cortex following coexpression of Sox2 (Sox2-IRES-Gfp) and Ascl1 (Ascl1-IRES-Dsred), as shown at 10 dpi. (H) Micrograph of the same field of view as depicted in (G), showing the neuronal morphology of DCX⁺ cells (white). (I) Numbers of induced reporter⁺/DCX⁺ neurons expressed as mean percentages ± SEM of the total number of transduced reporter⁺ cells following injection of a retrovirus encoding Dsred only (control; n = 3 mice), Ascl1 (n = 3 mice), Sox2 (n = 4 mice), or Sox2/Ascl1 (n = 3 mice). Statistical analysis was performed with Mann-Whitney U-test (^∗p ≤ 0.05). (J and K) High-magnification views of the area boxed in (G) and (H), respectively, showing the density and neuronal morphology of DCX⁺ cells (white). The arrowhead points to a DCX⁺ cell extending a long and ramified process. (L) Example of a DCX⁺ neuronal cell (white) induced upon expression of Sox2 and Ascl1.
(M–O) Example of a DCX⁺ neuron (white, arrowhead; O) boxed in (G) and (H), which appears to be induced by forced expression of Sox2 only (green, arrowhead; N) in absence of Ascl1 expression (red, arrowhead; M), as revealed by the white dashed line in (M) that mirrors the position of the depicted GFP⁺ cell in (N). Yellow arrowheads indicate the neuronal process of the cell in (N) and (O).
The scale bars represent 60 μm (B–E), 25 μm (F), 55 μm (G and H), 17 μm (J and K), and 10 μm (L–O). See also and .

Induction of Neuronal Cells in the Injured Adult Cerebral Cortex upon Forced Expression of Sox2 Alone
(A–D) Forced expression of Sox2 alone induces DCX⁺ neurons in the injured adult cerebral cortex. (A) Double immunostaining for GFP and DCX reveals appearance of induced neurons expressing DCX (white) in the injured cortex following Sox2 expression (Sox2-IRES-Gfp). Inset: high magnification of the boxed area showing GFP reporter expression in DCX⁺ cells. (B) Micrograph of the same field of view as depicted in (A), showing the neuronal morphology of the DCX⁺ cells (white, arrowheads). Inset: high magnification of the boxed area. (C and D) The micrographs depict a large cluster of densely packed DCX⁺ cells (white) induced by forced expression of Sox2 (green), located at the site of injury on top of the area depicted in (A) and (B). Inset in (D): high magnification of the boxed area (a single confocal plane is shown).
(E–G) The micrographs depict DCX⁺ neurons (red, arrowheads; E and F show the same cell) induced by forced expression of Sox2 alone (green, F; inset in G) at 10 dpi. Note the high level of complexity and ramification of their neuronal processes.
(H–N) The micrographs depict two GFP⁺/NeuN⁺ neurons (arrowheads 1 and 2) at 21 dpi induced by forced expression of Sox2 alone. (H) GFP immunostaining illustrating the high level of complexity and ramification of neuronal processes. (I) High-magnification view of the area boxed in (H), showing that these induced neurons express the mature neuronal marker NeuN (white, arrowheads 1 and 2). (J–M) High-magnification views showing GFP (J and L) and NeuN (K and M) immunoreactivity of the induced neurons (1 and 2). Single confocal optical planes are shown. (N) The micrograph shows the same field of view as shown in (H) and reveals loss of DCX expression in NeuN⁺ mature neurons.
(O and P) Morphological maturation of neuronal cells induced by Sox2 alone or Sox2/Ascl1. (O) Schematic diagram illustrating the progressive morphological maturation of DCX⁺ induced neuronal cells that were categorized at ∼12 dpi into five subsequent classes (A–E) according to their morphology. Class A: the cells are already DCX⁺ although still exhibiting a glial-like morphology with several processes; class B: the cells round up loosing their processes and exhibit a cytoplasmic ring of DCX; class C: the cells extend a single major DCX⁺ process, which is up to twice the size of their soma in length; class D: the cells extend a long unbranched DCX⁺ process, which is more than three times the size of their soma in length; and class E: the cells display a complex morphology and extend long DCX⁺ processes that show several ramifications. Reconstructions of reporter⁺/DCX⁺ cells based on the reporter signal are shown. (P) Numbers of Sox2- and Sox2/Ascl1-induced DCX⁺ cells in each class (O), expressed as mean percentages ± SEM of the total number of induced DCX⁺ cells at ∼12 dpi.
(Q) Location of Sox2- and Sox2/Ascl1-induced DCX⁺ cells throughout the injured cortical layers at ∼12 dpi. Numbers of DCX⁺ cells in each location (i.e., superficial, upper, and deeper layers) expressed as mean percentages ± SEM of the total number of induced DCX⁺ cells. Sox2: 258 DCX⁺ cells analyzed; n = 4 mice. Sox2/Ascl1: 184 DCX⁺ cells analyzed; n = 3 mice.
The scale bars represent 26 μm (A and B), 30 μm (C and D), 10 μm (E; I; and insets in A, B, and D), 6 μm (F), 15 μm (G and inset in G), and 20 μm (H and N). See also and .

Genetic Fate Mapping Demonstrates Reprogramming of NG2 Glia into Induced Neurons
Following tamoxifen-induced recombination, Sox10-iCreER^T2/GFP mice received a stab wound injury followed 3 days later by injection of Sox2-encoding retrovirus (Sox2-IRES-Dsred; A–C) or two retroviruses encoding Sox2 and Ascl1 (Ascl1-IRES-Dsred; D–F).
(A–C) Neuronal conversion of fate-mapped NG2 glia upon Sox2 expression. The micrographs depict a DSRED⁺ transduced cell (A; 12 dpi) that is converted into a DCX⁺ neuron (B, white) upon Sox2 expression and is immunoreactive for GFP (C), demonstrating its oligodendroglial origin.
(D–F) Neuronal conversion of fate-mapped NG2 glia following coinjection of Sox2- and Ascl1-encoding retroviruses. The micrographs show two DSRED⁺ transduced cells (D; 10 dpi; arrowheads) converted into DCX⁺ neurons (E, white, arrowheads) that also express GFP (F, arrowheads). Insets in (E): high magnification of the boxed areas in (D–F) highlighting the GFP⁺ soma of induced neurons (1 and 2).
The scale bars represent 12 μm (A–F) and 4 μm (insets in E).

Induced Neuronal Cells Exhibit Voltage- and Time-Dependent Conductances and Receive Synaptic Inputs
(A) Schematic diagram of experimental design.
(B) The micrograph depicts an example of DSRED⁺ cell 24 days after coinjection of Sox2- (Sox2-IRES-Dsred) and Ascl1 (Ascl1-IRES-Dsred)-encoding retroviruses, exhibiting a neuronal morphology selected for patch-clamp recording. Arrowheads point to processes emerging from the cell body. Insets: DSRED fluorescence and bright field picture of the cell body with the patch pipette.
(C) Ability of a DSRED⁺ cell to generate action-potential-like voltage changes in response to depolarizing current injection. Changes in membrane potential in response to de- and hyperpolarizing currents are shown.
(D) The action-potential-like signals induced by suprathreshold current pulses are blocked by TTX (red trace).
(E and F) Voltage-clamp recording of the same DSRED⁺ cell as shown in (D), demonstrating the presence of a Na⁺ current blocked by TTX. (E) Two current traces recorded in absence of TTX and at the indicated command potentials are shown. Holding potential (HP): −70 mV; step potential duration: 300 ms. Traces are depicted superimposed. (F) Similar recordings in the same cell following application of TTX. Currents were corrected for leakage currents.
(G and H) Examples of spontaneously occurring synaptic inputs recorded in a DSRED⁺ transduced cell. (G) A voltage-clamp recording for 4 min is shown. Nine synaptic-event-like responses are identified. (H) Events detected in (G) were superimposed with respect to their onset and averaged (red trace). Blue line: monoexponential fit to the average decay of these events.
The scale bars represent 10 μm (B, insets). See also and .

Induced Neurons Receive Synaptic Connections from Endogenous GABAergic Neurons
(A) The micrograph shows DSRED⁺-induced neurons following coinjection of Sox2- and Ascl1-encoding retroviruses (the same cells are also shown in D). Lower inset: high magnification of the boxed area showing some spine-like protrusions on the process of the DCX⁺ neuron (arrowheads). Upper inset: zoomed view of the area boxed in the lower inset showing the neck of one spine-like protrusion (arrowhead; a single confocal plane is shown).
(B–B′′) Close apposition of GAD65/GAD67⁺ puncta on the soma and processes of DCX⁺ induced neurons. (B) The micrograph depicts a GFP⁺ neuron induced by forced expression of Sox2 at 11 dpi. Inset: High-magnification view of the area no. 1 boxed in (B), showing several GAG65/GAG67⁺ puncta (red) in close apposition with the GFP⁺ process of the induced neuron. (B′) Same field of view as shown in (B), revealing DCX immunoreactivity of the GFP⁺ transduced cell. (B′′) High-magnification views of the area no. 2 boxed in (B), showing several GAG65/GAG67⁺ puncta (red) in close apposition with the GFP⁺ soma of the induced neuron. Upper panel: single confocal plane of the top of the cell revealing several puncta distributed on the cell surface. Lower panel: single optical confocal plane sectioning the cell in its middle, revealing the presence of puncta located around the cell body.
(C–H′) GFP⁺ axonal fibers and varicosities originating from GABAergic neurons wrap around the soma and processes of DCX⁺ induced neurons at 23 dpi. (C) GFP-expressing inhibitory neurons (GIN) transgenic mice received a coinjection of Sox2- and Ascl1-encoding retrovirus 3 days after stab-wound injury. The micrograph shows an overview of the stab-wounded area with the majority of DSRED⁺ transduced cells being located within the injured area, whereas several GFP⁺ GABAergic neurons are located on both sides of the lesion. Note the presence of several GFP⁺-intermingled dendritic and axonal processes. (D) High magnification of the area boxed in (C), depicting a DSRED⁺-induced neuron surrounded by GFP⁺ endogenous GABAergic interneurons and their processes. Inset: zoomed view of the area boxed in (D), illustrating the DCX immunoreactivity (white) of this induced neuron. (E and E′) 3D reconstruction of the same induced neuron shown in (D) (boxed area), depicting GFP⁺ axonal varicosities impinging on its soma and processes, arguing for innervation by local interneurons (arrowheads). 3D front (E) and back (E′) views are shown. Insets in (E and E′): high-magnification views of the area boxed in (E) and (E′). (F–H′) Example of another close apposition of GFP⁺ axons and axonal varicosities on the soma and processes of a DSRED⁺-induced neuron (F; arrowheads in H and H′). (G) High magnification of the same neuron depicted in (F), illustrating its DCX immunoreactivity (white). Inset: magnified view of a GFP⁺ axonal fiber and its varicosities wrapping around the induced neuron (arrowheads). (H and H′) 3D reconstruction of the same neuron shown in (F) (boxed area) and (G). 3D front (H) and back (H′) views are shown.
The scale bars represent 10 μm (A), 6 μm (B and B′), 70 μm (C), 14 μm (D), 7 μm (inset in D), 5 μm (E and E′), 9 μm (F), 7 μm (G), and 3.5 μm (H and H′).

Forced Expression of Sox2 Does Not Induce Neuronal Conversion of Cortical Glial Cells in Absence of Stab Wound Injury
(A) Schematic diagram of experimental procedures. The red dashed line shows the area where cells transduced by the lentiviral vector are distributed.
(B–D) Absence of neurogenesis induced by forced expression of Sox2 in the noninjured adult cortex. (B) The micrograph depicts GFP⁺ cells following injection of a Mokola-pseudotyped lentivirus encoding Sox2 expression (Sox2-IRES-Gfp) at 10 dpi. Most transduced cells exhibit a glial morphology. (C) Micrograph of the same field of view as shown in (B), revealing high levels of SOX2 expression (red) in all GFP⁺ transduced cells. (D) Micrograph of the same field of view as shown in (B) and (C), depicting that none of the transduced cells is DCX⁺ (white).
(E–H) GFP⁺ transduced cells are astrocytes and NG2 glia. (E and F) Double immunostaining for GFP and S100β (red), showing astrocytes transduced by the lentivirus (arrowheads). (G and H) Double immunostaining for GFP and OLIG2 (red), showing one transduced cell of the oligodendroglial lineage (arrowhead).
(I) Numbers of GFP⁺ cells in each category (i.e., astro, oligo, neuron, and DCX) are expressed as mean percentages ± SEM of the total number of transduced GFP⁺ cells following injection of the Mokola- or LCMV-pseudotyped lentiviruses encoding Sox2 (pooled data; n = 3 mice).
The scale bars represent 70 μm (B–D), 12 μm (E and F), and 6 μm (G and H). See also .