Transient Cnp expression by early progenitors causes Cre-Lox-based reporter lines to map profoundly different fates

Abstract

NG2 expressing oligodendroglial precursor cells are ubiquitous in the central nervous system and the only cell type cycling throughout life. Previous fate mapping studies have remained inconsistent regarding the question whether NG2 cells are capable of generating certain types of neurons. Here, we use CNP-Cre mice to map the fate of a sub-population of NG2 cells assumed to be close to differentiation. When crossing these mice with the ROSA26/YFP Cre-reporter line we discovered large numbers of reporter-expressing pyramidal neurons in the piriform and dorsal cortex. In contrast, when using Z/EG reporter mice to track the fate of Cnp-expressing NG2 cells only oligodendroglial cells were found reporter positive. Using BrdU-based birth dating protocols and inducible NG2CreER:ROSA26/YFP mice we show that YFP positive neurons are generated from radial glial cells and that these radial glial cells display temporary and low level activity of certain oligodendroglial genes sufficient to recombine the Cre-inducible reporter gene in ROSA26/YFP but not in Z/EG mice. Taken together, we did not obtain evidence for generation of neurons from NG2 cells. Our results suggest that with an appropriate reporter system Cnp activity can be used to define a proliferative subpopulation of NG2 cells committed to generate oligodendrocytes. However, the strikingly different results obtained from ROSA26/YFP versus Z/EG mice demonstrate that the choice of Cre-reporter line can be of crucial importance for fate mapping studies and other applications of the Cre-lox technology. GLIA 2017;65:342-359.

Keywords: Cnp; NG2 cells; fate mapping; radial glial cells; transient Cre expression.

FIGURE 1:

Reporter expression by NG2 cells in the CNP-Cre:Z/EG and CNP-Cre:ROSA26/YFP lines. A: Illustration of postnatal day 6 murine brain partial frontal section with anatomical structures analyzed in this study. D. Ctx, Dorsal Cortex; V. Ctx, Ventral Cortex; Piri. Ctx, Piriform Cortex; CA1, Hippocampal Cornus Ammonis Region 1; CC, Corpus Callosum; Fim, Fimbrae. B: Fraction of NG2 cells being GFP+ or YFP+ at P6 in the given gray and white matter regions. CNP-Cre driven GFP and YFP expression by a subpopulation of NG2 cells. Note that in the CNP-Cre:ROSA26/YFP line, a significantly greater proportion of reporter positive cells among NG21 OPCs exists as compared to the CNP-Cre:Z/EG line (Two-way Anova, Bonferroni post-hoc test, alpha = 0.05). C–D: Single optical section confocal scan of the Dorsal Cortex of CNP-Cre:Z/EG (C) and CNP-Cre:ROSA26/YFP (D) mice at P6 co-immunolabeled for GFP (green), NG2 (red) and Nissl deep red (blue). All arrowheads point to GFP−/NG2+ cells. All arrows point to GFP+/NG2+ cells. E–F: Coronal sections through the Corpus Callosum of CNP-Cre:Z/EG (E) and CNP-Cre:ROSA26/YFP (F) mice at P6 double immunolabeled for GFP (green), NG2 (red) and Nissl deep red (blue). All arrowheads point to GFP-/NG2+ cells. All arrows point to GFP1/NG2+ cells. All scale bars: 10 mm. G: A larger fraction GFP+ NG2+ cells co-labeled with anti-GPR17 antibodies when compared to NG2 cells being GFP-. White arrows indicate NG2 cells, red arrow marks GPR17 positive cells and green asterisks denote GFP+ cells. NG2 cells (cells with consistent labeling of their membrane at least around the nucleus) only showed GPR17 labeling within the soma. Cells showing GRP17 positive processes were found to be NG2−. Scale bar: 10 mm. H: As above but instead of the GFP channel the nuclear counter stain is shown in green for clearer cell identification.

FIGURE 2:

CNP-Cre driven GFP and YFP expression in mature oligodendrocytes. A: Fraction of GFP1 cells that are NG2+ (black bars) in Dorsal Cortex and Corpus Callosum of CNP-Cre:Z/EG mice at P10 and P30. Red bars show the fraction of GFP+ cells representing CC11 oligodendrocytes at P30. B-C: Single optical section confocal scan of the Dorsal Cortex (B) and Corpus Callosum (C) of P10 and P30 (upper and lower row, respectively) CNP-Cre:Z/EG mice immunolabelled for GFP/NG2 or GFP/CC1. D: Fraction of YFP+ cells that are NG2+ (black bars) and CC1± (red bars) in the P10 and P30 Dorsal Cortex and Corpus Callosum of CNP-Cre:ROSA26/YFP mice. E–F: Single optical section confocal scan of the Dorsal Cortex (E) and Corpus Callosum (F) of P10 and P30 CNP-Cre:ROSA26/YFP mice immunolabelled for GFP/NG2 or GFP/CC1. Asterisks indicate large YFP+ cell bodies being NG2− and CC1− seen only in CNP-Cre:ROSA26/YFP mice. All arrowheads point to GFP1/NG2− or GFP1/CC1-cells. All arrows point to GFP1/NG2+ or GFP+/CC11 cells. All scale bars: 10 µm.

FIGURE 3:

CNP-Cre driven YFP, but not GFP expression in astrocytes. A–D: Single optical section confocal scan of the Dorsal Cortex of P30 CNP-Cre:Z/EG (A) and CNP-Cre:ROSA26/YFP (B) mice double immunolabelled for GFP (green) and GFAP (red). (C) and (D) as before but for the Ventral Cortex of P30 mice. In the CNP-Cre:Z/EG line no astrocytes were GFP positive. On the other hand, YFP positive cells in the ventral cortex of the CNP-Cre:ROSA26/YFP line very frequently expressed GFAP (arrow) and represented 0.12 ± 0.02 of all YFP+ cells in the ventral cortex (total of 179 YFP+ cells, n = 4 mice). Note the overlap of GFP and GFAP expression in the primary and secondary processes of the astrocyte (D). All arrowheads point to GFP+/GFAP-cells. scale bars: 20 µm. E: Fraction of GFP+/YFP+ cells that are GFAP+ in the given gray and white matter regions of P30 CNP-Cre:Z/EG (n = 4 mice) and CNP-Cre:ROSA/YFP mice (n = 4 mice). Student’s unpaired t test was used to indicate significant difference, * indicates P < 0.05. D.Ctx, Dorsal Cortex; CC, Corpus Callosum; V.Ctx, Ventral Cortex.

FIGURE 4:

Widespread CNP-Cre driven YFP expression in neurons. A: Overview scan of YFP/NeuN double labelling showing the typical distribution of YFP positive cells in CNP-Cre:ROSA26/YFP mice. Note the dense group of YFP+ and NeuN+ large cells in layer II of the medial part of the dorsal cortex (dashed line). B–D: Single optical section confocal scan of the Dorsal Cortex (B) Ventral Cortex (C) and Piriform Cortex (D) of P30 CNP-Cre:ROSA26/YFP mice double immunolabelled for YFP (green) and NeuN (red). Arrows point to YFP1/NeuN+ cells. Arrowheads point at YFP+/NeuN-cells, likely representing OPCs. Note the numerous NeuN positive large somata. 0.55 ± 0.07 (total of 607 YFP+ cells, n = 4 mice), 0.37 ± 0.08 (total of 188 YFP1 cells, n = 4 mice) and 0.7 ± 0.05 (total of 353 YFP+ cells, n = 4 mice) of YFP+ cells that were NeuN+ mature neurons in the dorsal, ventral and piriform cortices, respectively. E: As above but dorsal cortex of P30 CNP-Cre:Z/EG mice. All scale bars: 20 mm. F: Fraction of GFP+ or YFP+ positive cells that are NeuN+ in the Dorsal, Ventral and Piriform Cortices of P30 CNP-Cre:Z/EG and CNP-Cre:ROSA/YFP mice. Two-way Anova with Bonferroni post-hoc test was used to test significant difference. D.Ctx, Dorsal Cortex; V.Ctx, Ventral Cortex; Piri.Ctx, Piriform Cortex.

FIGURE 5:

Embryonic but not postnatal BrdU pulse-chase labels YFP+ neurons. A: BrdU administration strategy to E12.5 CNP-Cre:R-OSA26/YFP mice via oral gavage, followed by analysis at P30. B–C: Single optical section confocal scan of the Ventral Cortex (B) and Piriform Cortex (C) of embryonic BrdU administered animals at P30 double immunolabelled for YFP (green) and BrdU (red). Arrows point to YFP+/BrdU+ neurons. D: BrdU administration strategy to postnatal CNP-Cre:ROSA26/YFP mice via i.p injections from P10-P13, followed by analysis at P32. E–F: Single optical section confocal scan of the ventral cortex (E) and piriform cortex (F) of postnatal BrdU injected animals at P32 double immunolabelled for YFP (green) and BrdU (red). Arrowheads point to YFP+/BrdU-neurons. All scale bars: 20 mm. G: Fraction of YFP+ neurons that are BrdU1 at both postnatal and embryonic BrdU administration time points in the ventral and piriform cortices. D.Ctx, Dorsal Cortex; V.Ctx, Ventral Cortex; Piri.Ctx, Piriform Cortex.

FIGURE 6:

Embryonic CNP-Cre driven reporter expression in radial glial cells, neurons and NG2 cells exclusively in ROSA26/YFP Cre reporter mice. A: Overview confocal scan of the dorsal cortex of E16 CNP-Cre:ROSA26/YFP mice double immunolabelled with GFP (green) and NG2 (red). Arrowheads point at NG2 negative YFP+ cells in the cortical plate, intermediate cortical layer and ventricular neuroepithelium, likely representing neurons, general progenitors and radial glia cells, respectively. Arrows point at NG21 YFP+ cells which we only found in the intermediate cortical layer. Only the small fraction of 0.14 ± 0.03 of YFP cells in this region expressed NG2 (total of 762 YFP+ cells, n = 3 mice). B: Single optical section confocal scan of the ventricular zone of E16 CNP-Cre:ROSA26/YFP mice double immunolabeled with GFP (green) and BLBP (red). BLBP identifies a fraction of 0.28 ± 0.07 (total of 435 YFP+ cells, n = 3 mice) of YFP+ cell located in the ventricular neuroepithelium as radial glia cells (arrows). Arrowheads point at BLBP negative YFP1 cells. C: GFP-NG2 double labelling of the E16 cortical plate of CNP-Cre:Z/EG mice. In this and other regions of the embryonic brain only vascular cells were positive for GFP and NG2. All scale bars: 20 µm.

FIGURE 7:

Radial glia-like cells display Ng2 promoter activity and generate neurons in the embryonic and adult dorsal cortex. A: Single optical section confocal scan of the ventricular zone of E19.5 NG2creER™:ROSA26/YFP administered tamoxifen at E16.5 double immunolabelled for YFP (green) and NG2 (red). Arrowheads point at YFP+/NG2-cells. Clear YFP positivity of radial glia-like cell indicates activity of Ng2 promoter and lack of NG2 immunoreactivity suggests transient nature of this activity. B: Single optical section confocal scan of the ventricular zone of E19.5 NG2creER™:ROSA26/YFP administered tamoxifen at E16.5 double immunolabelled for YFP (green) and BLBP (red). Arrows point at YFP+/BLBP+ cells. BLBP labelling demonstrates a radial glia-like phenotype of most YFP1 cells. Arrowhead points at YFP+/BLBP-cells. C: In the early dorsal cortex of the same mice large YFP+ and NG2-cells are observed which likely represent pyramidal neurons of the dorsal cortex (arrowheads). Note the smaller and brighter YFP+ cells (arrows) representing NG2+ cells seen at significant number at this age. D–F: Single optical section confocal scan of the dorsal cortex (D) ventral cortex (E) and piriform cortex (F) of P30 NG2creER™:ROSA26/YFP administered tamoxifen at E16.5 and double immunolabelled for YFP (green) and NeuN (red). Arrows point to YFP+/NeuN+ cells. Upon embryonic induction of Cre-recombinase, widespread reporter expression is observed in NeuN+ neurons. All scale bars: 20 µm.