Heterogeneity in oligodendrocyte precursor cell proliferation is dynamic and driven by passive bioelectrical properties

Abstract

Oligodendrocyte precursor cells (OPCs) generate myelinating oligodendrocytes and are the main proliferative cells in the adult central nervous system. OPCs are a heterogeneous population, with proliferation and differentiation capacity varying with brain region and age. We demonstrate that during early postnatal maturation, cortical, but not callosal, OPCs begin to show altered passive bioelectrical properties, particularly increased inward potassium (K⁺) conductance, which correlates with G1 cell cycle stage and affects their proliferation potential. Neuronal activity-evoked transient K⁺ currents in OPCs with high inward K⁺ conductance potentially release OPCs from cell cycle arrest. Eventually, OPCs in all regions acquire high inward K⁺ conductance, the magnitude of which may underlie differences in OPC proliferation between regions, with cells being pushed into a dormant state as they acquire high inward K⁺ conductance and released from dormancy by synchronous neuronal activity. Age-related accumulation of OPCs with high inward K⁺ conductance might contribute to differentiation failure.

Keywords: CP: Neuroscience; Kir channels; aging; bioelectricity; cell cycle; cell states; glia; oligodendrocyte precursor cell; potassium; proliferation; stem cell.

Graphical abstract

Figure 1

Cortical, but not callosal, OPCs develop inward conductance during early CNS maturation (A and C) Cells were selected by their EYFP expression (NG2-EYFP-knockin mice) in the cortex (CTX; A) and corpus callosum (CC; C). During whole-cell patch-clamp recording, OPCs were dye-filled with Lucifer yellow (LY; green) and post hoc labeled for NG2 (red). (B and D) Current/voltage (I/V) curves of steady-state currents in response to 20 mV steps from −134 to +26 mV in OPCs between P6 and P210 in CTX (B) and CC (D). (E–I) Comparison of I/V curves between CTX and CC OPCs at different ages. (J–L) In cortical OPCs, the resting membrane potential, membrane resistance, and inward conductance (see STAR Methods) change significantly at P10–P22 compared to P6–P9. These parameters then stay constant until P180–P210 except for the inward conductance, which further increases at P25–P34. (M–O) In callosal OPCs, the resting membrane potential, membrane resistance, and inward conductance change substantially only at P100–P120. Dots on bar graphs represent individual recorded cells. The p values are from two-way ANOVA with Šídák’s multiple comparisons test (E–I), with ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001, or from one-way ANOVA or Welch’s ANOVA (J–O; top) with post hoc Holm-Bonferroni tests (J–O; bottom). Data are shown as the mean ± SEM. See also Figure S1.

Figure 2

Cortical OPCs have different properties in areas of high neuronal density (Α) Gaussian fit of the log₁₀-transformed inward conductance (nS) in cortical OPCs (n = 264), where the sum of two Gaussians, R² value of 0.96, fits the data better than a single Gaussian. The red dashed line indicates the 90th percentile of the two Gaussian curves (equal to 2.35 nS). (B) Delineation (white circle) of an area occupied by a recorded NG2⁺ OPC (red) and dye filled with LY (green). Neuronal somata were labeled with anti-NeuN antibody (white). (C) The number of NeuN⁺ cells, or neuronal density, around OPCs with high inward conductance (high S; blue dots) was significantly higher compared to OPCs with a low inward conductance (low S; orange dots). (D and E) OPCs surrounded by a high number of NeuN⁺ cells were significantly more hyperpolarized (D) and had lower membrane resistance (E) compared to OPCs with a low number of NeuN⁺ cells in their surroundings. (F) The inward conductance was higher in OPCs with a high number of NeuN⁺ cells around compared to OPCs with a low number of NeuN⁺ cells in their surroundings. (G) The inward conductance was higher in OPCs co-cultured with a high density of neurons compared to OPCs cultured alone. Orange, blue, and gray dots represent individual recorded cells. The p values for (C)–(G) are from unpaired two-tailed t tests or t tests with Welch’s correction. Data are shown as the mean ± SEM. See also Figure S3.

Figure 3

Neuronal activity induces spontaneous long inward currents in cortical OPCs (A) Increasing neuronal firing by omitting Mg²⁺ in the aCSF evokes spontaneous long inward currents (SLICs) in OPCs, detected both in voltage-clamp and in current-clamp mode. (B) SLICs are reversibly blocked by tetrodotoxin (TTX), suggesting that they are dependent on neuronal activity. (C) OPCs and neurons were simultaneously patch clamped and dye-filled with Lucifer yellow (LY). OPCs were labeled for NG2 (red). (D) Dual patch-clamp recordings from a neuron in current-clamp and an OPC in voltage-clamp mode. Neurons in disinhibited slices (0 Mg²⁺ aCSF) fire spontaneous action potentials, which slightly precede SLICs in OPCs, highlighted by a gray box. (E and F) Dual voltage-clamp recordings from a neuron and an OPC show that adding 2 mM Mg²⁺ to the aCSF reduces the amplitude of synaptic inputs in neurons and the detection of SLICs in OPCs. Synaptic inputs to neurons slightly precede SLICs in OPCs, highlighted by a gray box and zoomed in (E′). (G) OPCs showing SLICs have a higher number of NeuN⁺ cells in their surroundings. (H) OPCs with SLICs show higher inward conductance compared to OPCs without SLICs. (I–K) OPCs with SLICs are more hyperpolarized and have lower membrane resistance and higher inward conductance than OPCs without SLICs. Dots represent individual recorded cells. The p values are from two-way ANOVA with Šídák’s multiple comparisons test (H), with ^∗p < 0.05 and ^∗∗p < 0.01, or unpaired two-tailed t tests or t tests with Welch’s correction (G and I–K). Results are given as the mean ± SEM. See also Figures S4 and S5.

Figure 4

SLICs are mediated by inward rectifying K⁺ channels (A) Differential interference contrast image of the mouse cortex depicting the simultaneous triple recording from an OPC using patch clamp, extracellular field potentials (FPs) recording to assess neuronal activity, and K⁺-selective microelectrodes (K⁺-ISM) to record changes in extracellular K⁺ concentration ([K⁺]_e). (B) Bath perfusion with 0 Mg²⁺ aCSF leads to spontaneous neuronal activity in mouse cortical slices as shown by FP recording accompanied by a release of K⁺ ions from neurons into the extracellular space recorded by K⁺-ISM and SLICs in OPCs. Representative traces are shown. (C) The amplitude of SLICs does not correlate with [K⁺]_e. (D) SLICs occur mostly in OPCs with low membrane resistance. (E and F) Adding Ba²⁺ blocks SLICs in OPCs but does not affect neuronal activity nor the release of K⁺ ions from active neurons. (G) Dual patch-clamp recordings from a neuron and an OPC in voltage-clamp mode showing that addition of nortriptyline blocks SLICs in OPCs but neuronal depolarizations are preserved. (H) Quantification of the percentage of SLICs blocked by TTX, Ba²⁺ (non-specific K⁺ channel blocker), nortriptyline (NT; inward rectifying K⁺ channel blocker), and a combination of 30 μM VU0134992 (VU) + 30 μM ML133 (ML) hydrochloride (specific Kir4.1 and Kir2.1 channel blockers, respectively). (I) Both nortriptyline and the combination of VU + ML block the inward currents in OPCs almost entirely (Welch’s ANOVA, p = 3.4 × 10⁻⁶). The numbers of recorded cells are indicated in the bars (H), while dots represent individual recorded cells (C, D, and I). The p values are from linear regression (C and D) or from Welch’s ANOVA followed by Holm-Bonferroni post hoc tests (I). Data are presented as the mean ± SEM. See also Figure S6.

Figure 5

Cortical OPCs with high inward K⁺ conductance reside in the G1/G0 cell cycle phase (A) Schematic showing the fluorescent labeling seen at different points in the cell cycle for fluorescent ubiquitination-based cell cycle indicator 2a (Fucci2a) mice. (B and C) Membrane currents evoked by stepping the membrane potential of recorded cells from −134 to +26 mV in 20 mV voltage steps in the NG2-EYFP:Fucci2a mouse cortex either in green (cycling) or in red (non-cycling) OPCs. (D) The resting membrane potential did not differ between the cycling (green) and the non-cycling (red) OPCs in the mouse cortex. (E and F) The membrane resistance of non-cycling (red) OPCs was significantly lower (E) and the inward conductance (F) was significantly higher compared to cycling (green) OPCs. Green or red dots indicate individual recorded cells. The p values were calculated by unpaired two-tailed t test. Data are shown as the mean ± SEM.

Figure 6

Increasing inward K⁺ conductance in OPCs diminishes their proliferation (A and B) Membrane currents evoked by stepping the membrane potential from −134 to +26 mV in 20 mV voltage steps in OPCs recorded in the corpus callosum of PDGFRɑ-CreER^T2:Kir2.1-mCherry mice or PDGFRɑ-CreER^T2:tdTomato reporter mice. (C) I-V curve of steady-state currents in response to 20 mV steps between −134 and +36 mV in Kir2.1 and tdTom mice. Kir2.1 channel overexpression in OPCs increases predominantly inward K⁺ currents. (D) I-V curve of steady-state currents in response to 20 mV steps between −134 and +36 mV in Kir4.1 channel-overexpressing and control cultured OPCs. Kir4.1 channel overexpression in OPCs produces passive (ohmic) K⁺ currents. (E–G) mCherry⁺ OPCs in Kir2.1 mice have significantly lower membrane potential, lower membrane resistance, and higher inward conductance compared to tdTom⁺ control OPCs. (H) Cultured OPCs overexpressing Kir4.1 channels have higher inward conductance compared to control cultured OPCs. (I) Quantification of OPC proliferation in Kir2.1 mice and their WT littermates in brain slices stained with DAPI (blue), EdU (green), mCherry (red), and NG2 (white). mCherry⁺/EdU⁻ OPCs are indicated with yellow arrowheads and mCherry⁻/EdU⁺ OPCs are indicated with white arrowheads. (J) Proliferation is reduced in Kir2.1-mCherry⁺ OPCs compared to controls. (K) Proliferation is reduced in Kir4.1 channel-overexpressing cultured OPCs compared to controls. (L) Proliferating OPCs in NG2-EYFP mice were immunolabeled with Ki67 at different ages. At P2, before the onset of high inward conductance, OPCs in the cortex and corpus callosum proliferate at similar rates. After the onset of high inward conductance in cortical OPCs, callosal OPCs proliferate more compared to cortical OPCs. At P80, the proliferation rates in both areas become similarly low. (M) OPCs co-cultured with neurons proliferate significantly less compared to OPCs cultured alone. (N) The rate of proliferation is comparable between young cortical and adult callosal OPCs, as well as between P10–P22 cortical and callosal OPCs overexpressing Kir2.1 channels. Individual dots indicate individual recorded cells (E–H), average in an animal (J, L, and N), or average on a coverslip (K and M). The number of animals was 5–6 per group. The p values are from two-way ANOVA with Šídák’s multiple comparisons test (C and D), with ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001, unpaired two-tailed t tests (E–H, K, M, and N) or mixed-effects analysis (J and L). Results are given as the mean ± SEM. See also Figures S6 and S7.

Figure 7

Increased neuronal activity leads to a decrease in inward K⁺ conductance and an increase in proliferation in OPCs (A) Adult NG2-EYFP mice were stereotaxically injected into the somatosensory cortices with either pAAV8-hSyn-hM3D(Gq)-mCherry (Gq) or pAAV8-hSyn-mCherry (control) viruses. (B) Representative images of cFos (green) and NeuN (white) staining in control and Gq mice. Yellow arrowheads point to NeuN⁺/cFos⁺ cells. (C) The number of cFos⁺ cells was higher in Gq animals compared to controls. (D) Representative traces from cortical extracellular field potential recording in acute brain slices from Gq mice superfused either with aCSF or with aCSF containing 10 μM CNO. (E) The frequency of neuronal firing was increased by 10 μM CNO in brain slices from Gq mice. (F) Timeline of the experiment. (G) EYFP⁺ OPCs (left, green; right, white) surrounding mCherry⁺ transduced neurons (red) were whole-cell patch clamped and dye-filled with Lucifer yellow (LY; green). (H–J) The resting membrane potential in OPCs was not different between Gq and control mice. However, increasing neuronal activity increased membrane resistance and decreased the inward K⁺ conductance in OPCs. (K) Timeline of the experiment in Kir2.1 mice. (L–N) The resting membrane potential and inward conductance in OPCs overexpressing Kir2.1 channels were not affected by increased neuronal activity. However, the membrane resistance in OPCs overexpressing Kir2.1 channels slightly increased in Gq mice compared to controls. (O and P) Timeline of EdU experiments in NG2-EYFP mice and Kir2.1-overexpressing mice with chemogenetically increased neuronal activity. (Q) To examine OPC proliferation following neuronal activity, slices were stained with DAPI (blue), EdU (white), and EYFP (green). Areas with mCherry⁺ neurons (red) were imaged. Arrowheads indicate EdU⁺ proliferating OPCs. (R) To examine OPC proliferation following neuronal activity in Kir2.1 mice, slices were stained with EdU (green), mCherry to label OPCs overexpressing Kir2.1 channels, and NG2 (white). Arrowheads indicate EdU⁺ proliferating mCherry⁻ and mCherry⁺ OPCs. (S) Proliferation was increased in OPCs of NG2-EYFP-Gq mice, but not in Kir2.1-Gq mice. Gray dots represent individual recorded cells (H–J and L–N) or animals (C and S) or brain slices (E). The p values are from unpaired two-tailed t tests (C, H–J, and L–N), paired t test (E), or one-way ANOVA (S; top) with Holm-Bonferroni post hoc tests (S; bottom). Data are shown as the mean ± SEM.