Genesis of a functional astrocyte syncytium in the developing mouse hippocampus

Abstract

Astrocytes are increasingly shown to operate as an isopotential syncytium in brain function. Protoplasmic astrocytes accomplish this function by evolving into a spongiform morphology, cytoplasmically connecting into a syncytium, and expressing a high density of K⁺ conductance. However, these critical structural and functional features are absent in neonatal astrocytes, which imposes a basic question of when a functional syncytium evolves in the developing brain. We show that the cellular structure and the spatial organization of an astrocyte syncytium reach stationary levels by postnatal day (P) 15 in the hippocampal CA1 region. Functionally, astrocytes begin to uniformly express a mature level of passive K⁺ conductance by P11. We then used syncytial isopotentiality measurement to monitor the maturation of the astrocyte syncytium. In uncoupled P1 astrocytes, the substitution of endogenous K⁺ by a Na⁺-electrode solution ([Na⁺]_p) resulted in the total elimination of the physiological membrane potential (V_M), and outward K⁺ conductance as predicted by the Goldman-Hodgkin-Katz (GHK) equation. As more astrocytes are coupled to each other through gap junctions during development, the [Na⁺]_p-induced loss of physiological V_M and the outward K⁺ conductance is progressively compensated by the neighboring astrocytes. By P15, a stably established syncytial isopotentiality (−73 mV), and a fully compensated outward K⁺ conductance appeared in all [Na⁺]_p- recorded astrocytes. Thus, in view of the developmental timeframe wherein a singular syncytium is anatomically established and functionally capable of equilibrating K⁺ ions, an astrocyte syncytium becomes fully operational at P15 in the mouse hippocampus.

INTRODUCTION

Astrocytes establish a syncytial network in the brain (Gutnick et al. 1981). However, this distinctive feature is absent in neonatal astrocytes (Binmoller and Muller 1992; Schools et al. 2006; Zhong et al. 2016), which is consistent with the time course of astrocyte genesis that starts within the later embryonic and early postnatal timeframe (Ge et al. 2012; Levison and Goldman 1993; Magavi et al. 2012; Tsai et al. 2012). What remains unknown is the time that newborn astrocytes fully converge into a syncytial network.

Astrocytes mainly use connexin 43 and 30 to connect into a continuous cytoplasmic syncytium (Dermietzel et al. 1989; Kunzelmann et al. 1999; Nagy et al. 1999), which is essential for the diffusion and equilibration of ions, metabolites, and signaling molecules within a syncytium (Khakh and McCarthy 2015; Kuga et al. 2011; Langer et al. 2012; Newman 2001; Rouach et al. 2008; Verkhratsky 2006; Verkhratsky and Nedergaard 2018). Additionally, a strong electrical coupling constantly equalizes the membrane potentials of astrocytes into an isopotential network, termed syncytial isopotentiality (Huang et al. 2018; Kiyoshi et al. 2018; Ma et al. 2016). Recent evidence further shows the essential role of syncytial isopotentiality for the functional microglia-astrocyte-neuron coupling in normal brain function (Du et al. 2022). The disruption of this network feature has also been shown to contribute to the pathology in animal models of epilepsy and depression (Aten et al. 2022a; Wang et al. 2020a). Thus, a full understanding of the structural and functional bases underlying syncytial isopotentiality is fundamentally important for further exploring the physiological and pathological roles of astrocytes in the brain.

Another important question is how to define a syncytium that is functionally operational. Homeostatic regulation of K⁺ ions is a frequently discussed astrocyte function. To execute this function, a fully connected syncytium through gap junctions and a mature level expression of membrane passive K⁺ conductance are crucial determinants for the intra-syncytium buffering of K⁺ ions (Kofuji and Newman 2004; Orkand et al. 1966; Ransom 1996). Interestingly, newborn astrocytes lack gap junctional coupling and are deficient in functional leak K⁺ conductance (Binmoller and Muller 1992; Schools et al. 2006; Zhong et al. 2016). Therefore, it is fundamentally important to know the time at which a syncytium becomes operational for K⁺ homeostasis. Further, a functional readout that can discern the efficiency of K⁺ buffering within a syncytium should have a profound application for exploring the pathological role of dysfunctional astrocyte networks in neurological disorders.

To address these fundamental questions, we followed the postnatal development of astrocytes in the hippocampal CA1 striatum radiatum and examined the maturation in terms of astrocyte cellular morphology, syncytial organization, K⁺ channel expression, syncytial isopotentiality, and the buffering capacity within a syncytium for K⁺ ions. We show that a coordinated maturation of these features ultimately gives rise to a functionally mature astrocyte syncytium at P15.

MATERIALS AND METHODS

Animals

All the experimental procedures were performed following a protocol approved by the Animal Care and Use Committees of The Ohio State University. C57BL/6N and BAC-Aldh1l1-eGFP transgenic mice (Yang et al. 2011) aged postnatal days 1–21 were used in the present study. Hippocampal astrocytes from P1–21 mice of both sexes were used.

Preparation of acute hippocampal slices

Hippocampal slices were prepared as described previously (Ma et al. 2014; Ma et al. 2012). Briefly, mouse brains were rapidly removed from skulls after anesthesia with 8% chloral hydrate in 0.9% NaCl and placed into ice-cold oxygenated (95%O₂/5%CO₂) slice-cutting artificial cerebrospinal fluid (aCSF) with reduced Ca²⁺ and increased Mg²⁺ (in mM): 125 NaCl, 3.5 KCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 0.1 CaCl₂, 3 MgCl_2, and 10 Glucose. Coronal hippocampal slices (250 μm) were cut at 4 °C with a Vibratome (Pelco 1500) and transferred to the oxygenated normal aCSF (in mM): 125 NaCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 3.5 KCl, 2 CaCl₂, 1 MgCl₂ and 10 Glucose (Osmolality, 295 ± 5 mOsm/L; pH 7.3–7.4). Then the slices were recovered at room temperature for at least 1 hour before electrophysiological recordings.

Sulforhodamine 101 staining

In some experiments, the slices after the recovery were transferred to a beaker with a slice-holding basket containing 0.6 μM Sulforhodamine 101 (SR101) in aCSF at 34 °C for 30 min. Then, the slices were transferred back to normal aCSF at room temperature before the experiment. Slices from BAC-Aldh1l1-eGFP transgenic mice were mounted and images were immediately taken after SR101 staining to analyze the co-localization of SR101 with Aldh1l1-eGFP in CA1 stratum radiatum region using a confocal microscope (LSM510, Carl Zeiss).

Optical tissue clearing and immunohistochemistry

Tissue clearing was performed according to the protocols described in these reports (Susaki et al. 2014; Susaki et al. 2015; Tainaka et al. 2014). Briefly, 4% paraformaldehyde-fixed brain slices were immersed in ScaleCUBIC reagent-1 for 3 days at room temperature. Brain slices were mounted with fresh ScaleCUBIC reagent-1 for 1 hour before imaging. Images were acquired by single photon confocal microscopy (SP8, Leica).

Fresh dissociation of single hippocampal astrocytes

The protocol for fresh dissociation of single hippocampal astrocytes was described previously (Du et al. 2015). Briefly, coronal hippocampal slices at 250 μm thickness were sectioned from P21–25 mice and then incubated in the oxygenated aCSF. Thereafter, three slices were incubated in SR101 for 30 min. The CA1 regions were dissected out from slices after incubation, cut into small pieces (1 mm²), and transferred into a 1.5 ml Eppendorf tube containing oxygenated aCSF supplemented with 24U/ml papain and 0.8 mg/ml L-cysteine for 7 min incubation at 25^°C. The loosened tissues after papain digestion were gently triturated 5–7 times into a cell suspension. The cell suspension was transferred to the recording chamber and was precipitated for 5 min. Only single dissociated astrocytes were used in this study.

Imaging acquisition

A fluorescent imaging system, Polychrome V system (Till Photonics, Germany) was used for the identification of astrocytes from Aldh1l1-eGFP or SR101 staining astrocytes in slices. This system was also used for high-resolution visualization of small glial soma for whole-cell astrocyte recording.

Electrophysiology

For brain slice recording, individual hippocampal slices were transferred to the recording chamber mounted on an Olympus BX51WI microscope, with constant perfusion of oxygenated aCSF (2.0 ml/min). Astrocytes in the CA1 stratum radiatum region were identified based on SR101 staining or Aldh1l1-eGFP signal visualized through an infrared differential interference contrast (IR-DIC) video camera. Whole-cell patch-clamp recordings were performed using a MultiClamp 700B amplifier and pClamp 9.2 software (Molecular Devices, Sunnyvale, CA). Borosilicate glass pipettes (Warner Instrument, Hamden, CT) were pulled from a Micropipette Puller (Model P-87, Sutter Instrument). The recording electrodes had a resistance of 4.0–4.5 MΩ when filled with the electrode solution. The standard pipette solution, or [K⁺]_P, contained (in mM) 140 KCl, 0.5 CaCl₂, 1.0 MgCl₂, 5 EGTA, 10 HEPES, 3 Mg-ATP, and 0.3 Na-GTP (280 ± 5 mOsm/L, pH 7.25–7.35). In a subset experiment, the 140 mM KCl was substituted by 140 mM of K⁺-gluconate to reduce the intracellular Cl⁻ content. Consistent with our previous report, altering intracellular Cl⁻ content did not have a major impact on astrocyte V_M (Xie et al. 2007), and thus the data were merged for statistical analysis. In Na⁺-based pipette solution, or [Na⁺]_P, the 140 mM KCl was fully substituted by equimolar NaCl in the referred experiments. The whole-cell membrane current (I) in voltage-clamp mode and V_M in the current-clamp mode were amplified by a MultiClamp 700B amplifier and the data acquisition was controlled by PClamp 9.2 program. The liquid junction potential was compensated before forming the cell-attached mode for all recordings. A minimum of 2 GΩ seal resistance was required before rupturing the membrane into the whole-cell recording mode. In the current clamp recording, the input resistance (R_in) was measured by the “Resistance test” protocol in PClamp 9.2 software (pulse: 63 pA/600 ms) before and after recording with the standard [K⁺]_P. When R_in varied greater than 10% during recording, the recordings were discarded. In the experiments of studying isopotentiality, we used a 140 mM Na⁺ pipette solution. The R_in was periodically monitored by injection of negative current pulses (60 pA/0.5 s). Note that in all V_M recordings in this report, the R_in was continuously monitored to ensure consistent recording quality, e.g., Figure 3a. The V_M was read either in “I = 0” mode or measured directly in current clamp mode with no holding currents. Astrocytes with a resting membrane potential more positive than −75 mV in both brain slices and freshly isolated single astrocytes were discarded. All the experiments were conducted at room temperature.

Figure 3. Dependence of quasi-physiological V_M,SS on the number of coupled astrocytes in a syncytium.

a. [Na⁺]_p V_M recording from a freshly isolated astrocyte (left) and a syncytial coupling astrocyte at P21 (right). In either case, the V_M,_I is a readout of the cell’s resting V_M. The V_{M, SS} in single astrocyte, and astrocyte in situ respectively followed the GHK predicted V_M for [Na⁺]_p and a quasi-physiological V_M established by syncytial coupling. A negative current pulse (63 pA/600 ms) was periodically injected during recording to ensure an unchanged R_in throughout the recording. Two representatives R_in tests (t,₁ and t,₂) are shown in expanded scale on the right panel. b. The V_{M, SS} recorded with [Na⁺]_p containing 0.1 mM LY from a P1, P9, and P15 astrocyte as indicated, where the establishment of a quasi-physiological V_{M, SS} is developmentally dictated by the increasing number of LY dye coupled astrocytes. c. Using V_M,SS as a readout, the syncytial isopotentiality requires a minimum of 7–9 astrocytes that couple to the [Na⁺]_p recorded astrocyte in situ. **: p < 0.005, NS: p > 0.05.

Lucifer Yellow dye loading

0.1 mM Lucifer Yellow (LY) was included in the standard [K⁺]_P or the [Na⁺]_P (Stewart 1978), which did not alter the osmolarity (280 ± 5 mOsm/L) and the pH (7.25–7.35) of these solutions. Note that astrocytic expression of connexin (Cx) 30 begins after P16 (Kunzelmann et al. 1999; Nagy et al. 1999). Because Cx30 lacks LY permeability (Manthey et al. 2001), the LY dye-coupling analysis was restricted to animals younger than P15 in the present studies. After the whole-cell LY dialysis for 20 min, the hippocampal slices were fixed in 4% paraformaldehyde for 1 hour at room temperature and then washed in PBS three times. The number of coupled LY⁺ astrocytes in the syncytia was counted from images acquired from a confocal microscope (LSM510, Carl Zeiss).

Chemical reagents

SR101 was purchased from Invitrogen (New York, NY). All other chemicals and salts used in intracellular and extracellular solutions were purchased from Sigma-Aldrich. The 100 μM meclofenamic acid (MFA) or 100 μM BaCl₂ was dissolved directly in aCSF.

Data analysis

The patch clamp recording data were analyzed by Clampfit 9.0 (Molecular Devices, Sunnyvale, CA) and Origin 8.0 (OriginLab, North Hampton, MA). Data are presented as mean ± SEM. Statistical analysis was performed using One-way ANOVA. Two-Way repeated measures ANOVA was used to compare differences in the currents recorded at each of the command voltage steps in separate groups (Figure 7). The significance level was set at p < 0.05. Boltzmann function here is used to fit the postnatal development of syncytial coupling based on the data of the median steady-state V_M (V_{M, SS}) recorded with [Na⁺]_p at each postnatal day in the following form (Figure 4b) (Dubois et al. 2009).

$$y=A2+\frac{(A1-A2)}{1+{exp}^{\left(\frac{x-x0}{dx}\right)}}$$

Y is the V_{M, SS} recorded with [Na⁺]_p at a given postnatal day x, A1 is the V_{M, SS} at the P1 where the syncytial isopotentiality is absent, A2 is the V_{M, SS} where the syncytial isopotentiality is fully established in astrocyte syncytium, i.e., P15 onward, x0 is the midpoint reflecting the inflection of the curve, and dx is the time constant associating with the rate of change.

Figure 7. Syncytial coupling enables a high efficient redistribution of K⁺ ions within a syncytium.

a. Illustrations of a functionally mature syncytium for efficient spatial transfer of K⁺ conductance (I_{K, syncytium}) between a [Na⁺]_p recoded astrocyte and coupled neighbors (a1), and inhibition of syncytial coupling by MFA (a2). b1. The intact passive K⁺ conductance in a [Na⁺]_p recorded astrocyte in situ (b1), and subsequent elimination of outward-going K⁺ conductance by MFA (b2). b3. Shows a significant contribution of I_K,syncytium to both inward-going and outward-going K⁺ conductance. c. In the presence of MFA, the removal of 3.5 mM K⁺ from bath aCSF further eliminated the inward-going passive conductance in a reversible manner. d. In the presence of MFA, the inward-going passive K⁺ conductance was significantly inhibited by a Kir_4.1 inhibitor, 100 μM BaCl₂. # in b3: p values range between 0.00074–1.23E-12. £ in e1: 0 mM [K⁺]_o comparing to 3.5 mM [K⁺]_o, p values range between 0.01404 to 6.52E-12; & in e2: 100 μM Ba²⁺ comparing to aCSF, p values range between 0.4938 to 2.82E-34. Further statistical details are available in Supporting information – Table 4.

Figure 4. Development of syncytial isopotentiality in the hippocampal CA1 region.

a. An illustration of developmental integration of astrocytes into a syncytium through gap junctional coupling. b. Development of syncytial isopotentiality revealed from 397 V_{M, SS} recordings from P1 -P21, where a stable quasi-physiological V_M,SS appears around P13–15. The developmental transition of the V_M, _SS from the solitary/weakling coupled astrocytes to a fully coupled syncytium followed nicely to a Boltzmann growth function (Adj. R-Square (R²) value, 0.98165; reduced Chi-square (X²) value, 13.72091). c. Using V_M,SS = −60.8 mV as a cut-off value to differentiate the astrocytes into those coupled to either less or greater than 7–9 astrocytes at each postnatal day, where a stably established syncytium appeared around P14. *: p < 0.05, **: p < 0.005, NS: p > 0.05.

Coefficient of variation (CV) is used to access the extent of variability of data in relation to the mean in the following formula:

$$CV=\frac{\sigma }{u}$$

where σ is the standard deviation and μ the mean of the dataset.

The rectification index (RI) was calculated by dividing the current amplitudes induced by the command voltages of +20 mV (y1) over −180 mV (y2) (Du et al. 2016a; Wang et al. 2020b). RI was used in analyzing the developmental changes in the ratio of inward- over outward-going passive conductance (Figure 6), and the ability of astrocytes for intro-syncytial equilibration of K⁺ conductance (Figure 8).

Figure 6. Developmental maturation of leak K⁺ conductance.

a-b. Three whole-cell current profiles (a) and their corresponding I-V plots (b) show the developmental transition of astrocytes from a predominant expression of voltage-gated IK_a and IK_d with a weak expression IK_in to the solely leak type K⁺ conductance (passive conductance). c. The development of inward K⁺ conductance (activated at −180 mV) and outward K⁺ conductance (activated at +20 mV) with the animal ages. The VRAs disappear at P10–12, meanwhile, both inward and outward currents do not show further change at this developmental time point. d. Rectification index (RI) analysis of developmental changes of the inward-going over outward-going K⁺ conductance within and between VRA and PA phenotypes. e. illustration of development maturation of leak K⁺ conductance with animal ages. *: p < 0.05, **: p < 0.005, ***: p < 0.001, NS: p > 0.05.

Figure 8. An operational astrocyte syncytium emerges at P15 in the developing hippocampus.

a-c. To evaluate the efficacy of a syncytium for spatial transfer of K⁺ ions within a syncytium, astrocytes at different ages were recorded with a [Na⁺]_P in voltage-clamp mode. The cells were held at their actual resting V_M, and the voltage commands (V_COM) were in the range of Δ ±100 mV (20 mV increment). a. A solitary astrocyte showed negligible hyperpolarization-induced endogenous inward currents (I_{K, endogenous}) due to low-density expression of leak Kir_4.1 channels, nor the depolarization-induced outward-going K⁺ currents deriving from the coupled syncytium (I_{K, syncytium}) due to the absence of gap junction coupling with other astrocytes. b. A recorded P9 astrocyte has a significantly increased I_{K, endogenous}, resulting from up-regulation of intrinsic Kir_4.1 and recruitment of additional inward K⁺ conductance from neighboring astrocytes (I_{K, syncytium}). Note that the I_{K, syncytium} is the K⁺ conductance of neighboring astrocytes in response to the V_COM applied to the recorded astrocyte. c. Shows an operational syncytium where the I_{K, syncytium} made a full compensation for the entire missing outward-going component of passive K⁺ conductance, as well as a significant amount of inward-going passive K⁺ conductance. Because the percentage of compensation for the latter cannot be quantified, only outward-going K⁺ conductance is labeled as I_{K, syncytium}. d. Rectification index (RI) analysis as a readout of the developmental increase in the I_{K, syncytium}. Two comparations are analyzed; the astrocytes recorded with [Na⁺]_p at different age groups and comparisons of these recordings with [K⁺]_p. *: p < 0.05, **: p < 0.005, ***: p < 0.001, NS: p > 0.05.

$$RI=\frac{y1}{y2}={/}_{+20\text{mV}}/{/}_{-180\text{mV}}$$

RESULTS

Developmental maturation in astrocyte gap junctional coupling

Consistent with the lack of dye coupling of astrocytes in the early developing hippocampus and visual cortex (Binmoller and Muller 1992; Schools et al. 2006), we have previously shown the absence of electrical coupling in a large population of newborn astrocytes in mouse hippocampal CA1 region (Zhong et al. 2016). To further examine this question, here, we used Aldh1l1-eGFP reporter mice to identify astrocytes. To begin, we first confirmed the astrocytic identity of eGFP-expressing cells in this reporter mouse by analyzing the colocalization of the eGFP signal with a commonly used astrocytic marker SR101 at different postnatal developmental times. In line with our previous findings (Zhong et al. 2016), SR101 and Aldh1l1-eGFP showed robust colocalization throughout development (Figure 1a–c, Supporting information -Table 1). Therefore, in the following studies, Aldh1l1-eGFP and SR101 were interchangeably used in different experiments for the identification of astrocytes.

Figure 1. Developmental transition of astrocytes from solitary individuals to syncytial building blocks.

Astrocytes at P3 (a), P12 (b), and P21(c) that could be identified from BAC-ALDH1L1-eGFP mice or SR101 staining in the hippocampal CA1 stratum radiatum: the yellow merges show a robust colocalization of eGFP⁺ astrocytes with another astrocytic marker SR101 at all animal ages (a-c). d-f. Patch electrode LY loading revealed a solitary astrocyte at P1 (d), multiple LY dye coupled astrocytes at P8 (e), and a fully developed syncytium at P21 (f). g. An age-dependent increase in LY dye coupled astrocytes with the animal ages. h. Dual patch recording from a pair of SR101 identified neighboring astrocytes in the CA1 region. i-k: The current steps that were alternately injected to a pair of recording astrocytes (i) for identification of uncoupled solitary astrocytes (j) and coupled astrocytes (k). ***: p < 0.001, NS: p > 0.05.

First, we used LY dye coupling analysis to observe the developmental formation of astrocyte syncytium. At P1, LY was always restricted in the recorded SR101⁺ astrocytes. This was consistent with our previous finding that newborn astrocytes lack gap junctional coupling, and therefore, are mostly solitary individual cells (n = 7) (Figure 1d, g) (Zhong et al. 2016). The number of dye-coupled astrocytes increased with age: 1.6 ± 0.2 cells at P2–3 (n = 17, p = 0.992), 3.2 ± 1.4 cells at P4–6 (n = 6, p = 0.861), 5.1 ± 0.6 cells at P7–9 (n = 17, p = 0.714), and 10.3 ± 1.6 cells at P10–12 (n = 13, p = 0.0006) (Figure 1e, g). Starting from P12, LY spreads into hundreds of astrocytes (Figure 1f, g). This observation was consistent with previous reports from others and us (Binmoller and Muller 1992; Schools et al. 2006; Xu et al. 2010), thus, offering us an initial clue that the astrocyte syncytium should be established around P12.

Electrical coupling measurement offers a higher sensitivity to detect gap junctional coupling (Stephan et al. 2021). Thus, we next used dual patch recording to further characterize the developmental maturation of gap junctional coupling of astrocytes (Figure 1h, i). Experimentally, command currents were injected into one cell in a pair and induced voltage responses were recorded from both cells (Figure 1i–k). The presence of voltage responses in the second cell indicates the electrical coupling between the recorded astrocytes (Figure 1k), otherwise, the recorded cells were uncoupled astrocytes (Figure 1j). The absence of electrical coupling does not rule out the coupling of the recorded astrocytes to other neighbors, thus, the percentage of uncoupled astrocytes provides an estimate of astrocytes that are not fully integrated into a singular syncytium at a given developmental time point. In 20 pairs of immediate adjacent astrocytes from P1–3, electrical coupling was absent in 6 recorded pairs, or 30% (12/40 cells) of astrocytes were not fully integrated into a singular syncytium. At P17–19, electrical coupling was absent in only 1 out of 17 recorded pairs, or only 5.9% (2/34 cells) of astrocytes were not coupled into a syncytium. Electrical coupling could be detected from all 21 pairs of astrocytes at P21–28. The cellular center-to-center distance between the two recording electrodes was 21.5 ± 1.52 μm (n = 20 pairs), 32.5 ± 3.43 μm (n = 17 pairs), and 38.5 ± 3.13 μm (n = 21 pairs) for P1–3, P17–19, and P21–28 groups, respectively.

Together, the dye and electrical coupling analyses convincingly show a progressive transition of newborn astrocytes from solitary individuals to the “building bricks” of a singular astrocyte syncytium (Figure 1g) (Zhong et al. 2016). Note that, compared to a total absence of dye coupling (Figure 1d, g), 60% of astrocytes show electrical coupling at P1–3, indicating that weak but non-uniform electrical coupling already occurs in a large population of newborn astrocytes. This observation illustrates again the need for electrical coupling measurement to assess whether they are coupled (Stephan et al. 2021; Xu et al. 2010).

Developmental maturation in the spatial organization of an astrocyte syncytium

The number of glial cells, but not neurons, rapidly increases from 6% at birth to 50% of the total brain cells at the adult level (Bandeira et al. 2009). However, little is known about the relevance of a total increase in glial population to the developmental change in astrocyte density change in astrocytes, nor its regulatory role in the spatial organization of astrocyte syncytium. To answer this basic question, we used CUBIC tissue clearing to make Aldh1l1-eGFP brain slices optically transparent. In doing so, the GFP⁺ astrocytes in the hippocampal CA1 stratum radiatum region could be readily visualized and analyzed (Figure 2a). Within the dormant time (P1–3) with no change in the number of neurons and only a slight increase in non-neuronal cells (Bandeira et al. 2009), astrocyte density significantly increased (Figure 2b): P1: 26.1 ± 0.5×10³ cells/mm³ (n = 9); P3: 29.8 ± 1.6×10³ cells/mm³ (n = 6, p = 0.013). Astrocyte density then fell out at P9 (19.6 ± 0.4×10³ cells/mm³, n = 6, p = 0.00009) and P15 (14.7 ± 0.7×10³ cells/mm³, n = 8, p = 0.0008). Compared with P15, no further change occurred in P21 (13.3 ± 0.5×10³ cells/mm³, n = 9, p = 0.646) (Figure 2b). Thus, following a transient increase from P1 to P3, astrocyte density declines over the first two postnatal weeks to arrive at a stationary level by P15.

Figure 2. Developmental maturation of astrocyte syncytial organization in CA1 stratum radiatum region.

a.Optically transparent (CUBIC) hippocampal stratum radiatum tissues prepared from P3 (a1) and P15 (a2) ALDH1L1-eGFP transgenic mice, respectively. Both images were comparably rendered with the same volume. More astrocytes in proximity with each other, indicated by eGFP⁺ cells, appeared in P3 CA1 syncytium (a1) than that of P15 syncytium (a2). b. The density of astrocytes progressively decreases over syncytial development. c. Interastrocytic distance increases during the maturation of the syncytium. d. Illustrations of polarity analysis of astrocytes at the indicated animal ages. e-g. Summaries of polarity analysis: polarity index (e), process length (f), and coefficient of variation (CV) (g). *: p < 0.05, **: p < 0.005, ***: p < 0.001, NS: p > 0.05.

We then examined the interastrocytic distance that is critical for the formation of gap junctional coupling at the interfaces of astrocytic domains (Aten et al. 2022b; Bushong et al. 2004). There was no significant change from P1 (18.2 ± 0.5 μm, n = 45) to P3 (17.9 ± 0.6 μm, n = 45, p = 1.000), but that was followed by a significant increase at P9 (32.4 ± 0.6 μm, n = 45, p = 7.29E-08), and P15 (40.3 ± 1.1 μm, n = 45, p = 8.08E-08, Figure 2c). No further change occurred at P21 (42.0 ± 1.0 μm, n = 45, p = 0.585, Figure 2c), therefore, the interastrocytic distance also reaches a stable level at P15.

In summary, although the total glial population increases during brain development (Bandeira et al. 2009; Ge et al. 2012), this event only corresponds with a transient increase in astrocyte density during the dormant period. Astrocyte density then continuously falls out in the following two postnatal weeks until it reaches a steady level by P15. A developmental expansion of hippocampal volume and a large-scale apoptotic astrocytic death may collectively contribute to our observations (Krueger et al. 1995; West 1990).

Maturation in astrocyte process arborization and domain formation

A mature form domain structure of astrocytes is another determinant to achieve an adequate strength of syncytial coupling (Aten et al. 2022a; Bushong et al. 2004). To examine this question further in the developing brain, we loaded LY to the individual astrocytes after the syncytium was uncoupled by 100 μM MFA in bath aCSF for over one hour (Schools et al. 2006; Zhou et al. 2006), which effectively inhibits gap junctional coupling by 99.3% (Ma et al. 2016). The uncoupling enhanced LY accumulation in the recorded cell, which, therefore, facilitated the visualization of the delicate processes for morphological analysis (Du et al. 2022; Xu et al. 2010; Xu et al. 2014).

The developmental change in the domain shape of astrocytes was analyzed by the polarity index. Specifically, two maximal lengths of an astrocyte, one running in parallel with, and another perpendicular to the pyramidal neuron layer were respectively defined as X and Y. The polarity index is defined as the quotient (Y/X) (Figure 2d1–d3). A polarity index value greater than 1 indicates the domain polarity that is preferentially orientated towards the pyramidal layer.

The polarity index at P3 (2.16 ± 0.06, n = 4) was the highest compared to those at P15 (1.45 ± 0.11, n = 4, p = 0.001) and P21 (1.55 ± 0.11, n = 4, p = 0.004). The polarity index becomes comparable between P15 and P21 (p = 0.739) (Figure 3e); thus, a mature ellipsoid domain shape (with respect to astrocyte morphology) appears to be established at P15 (Aten et al. 2022b; Cali et al. 2019).

We next examined the changes in the length of astrocytic processes during development. The process length did not differ across the age groups: P3 (32.9 ± 2.9 μm, n = 24), P15 (32.7 ± 1.7 μm, n = 24, p = 0.997), and P21 (28.7 ± 1.2 μm, n = 24, p = 0.334) (Figure 2f). However, the process length varied from cell to cell in P3 astrocytes. Therefore, we introduced the coefficient of variation (CV) analysis to further analyze this question. Indeed, the CV in P3 (45.09 ± 5.03%, n = 4) was significantly higher than P15 (22.02 ± 2.91%, n = 4, p = 0.004) and P21 (21.28 ± 2.76%, n = 4, p = 0.004). However, no further change was observed in the CV between P15 and P21 (p = 0.989) (Figure 2g).

These morphometric analyses revealed that the cellular structure of hippocampal astrocytes reaches an initial maturity by P15, an observation that is consistent with a previous study (Bushong et al. 2004).

Maturation in astrocyte syncytial isopotentiality

To identify the time hippocampal astrocytes fully connect into a syncytial network, we combined LY dye coupling and syncytial isopotentiality analyses to trace the functional maturation of the astrocyte syncytium in the mouse hippocampal CA1 region (Du et al. 2022; Kiyoshi and Zhou 2019; Stephan et al. 2021).

To begin, it is necessary to reiterate the rationale for the syncytial isopotentiality measurement. In [Na⁺]_p whole-cell recording, the rupture of uncoupled astrocytes results in an immediate deflection of V_M to the cells’ resting V_M at ~ −76 mV, defined as V_M,_I. The substitution of endogenous K⁺ content by the Na⁺ ions from the [Na⁺]_p progressively shifts the V_M towards ~ 0 mV at the steady-state level (V_M,_SS) which is anticipated by the GHK equation (Figure 3a, left) (Ma et al. 2016). In a syncytial coupled astrocyte, the V_M,SS is retained (or voltage-clamped) at a quasi-physiological level (−73 mV) by the neighboring astrocytes through gap junctional coupling (Figure 3a, right). It should be noted that a V_{M, SS} recording remains at a constant level for hours in a functionally mature syncytium (e.g., >P15), resulting from a minimized ionic movement within an isopotential network as the charge carriers (mainly Na⁺ and K⁺). The biophysical rationale and experimental validation of this notion are available in our previous publications (Ma et al. 2016; Stephan et al. 2021). Additionally, in the events of strengthening or weakening of the syncytial coupling, the V_M,SS shifts further toward the physiological V_M of the neighboring astrocytes (i.e., −76 mV) or the GHK predicted V_M at 0 mV, respectively (Kiyoshi et al. 2018; Stephan et al. 2021).

To evaluate the relationship between the number of coupled astrocytes with the levels of the quasi-physiological V_M,SS, we added 0.1 mM LY to [Na⁺]_p and recorded astrocytes from P1 - P15 (Figure 3b–c). In solitary astrocytes, the V_M,_SS was at the level of −16.8 ± 3.3 mV (n = 6). The V_M,_SS in 2–3, and 4–6 dye-coupled astrocytes was −19.2 ± 5.0 mV (n = 10, p = 0.998), and −34.0 ± 5.5 mV (n = 9, p = 0.190), respectively (Figure 3c). The V_M,_SS of a cluster of 7–9 coupled astrocytes started to be significantly closer to the physiological V_M of neighboring astrocytes: −60.8 ± 5.4 mV (n = 10, p = 0.003). Importantly, the V_M,_SS of the syncytia containing 7–9 astrocytes did not differ from the syncytium containing more than 10 astrocytes (−72.5 ± 3.4 mV, n = 8, p = 0.456) (Figure 3c), which confirms our previous finding that, for a given astrocyte, a minimum of 7–9 connected neighbors is a necessity to ensure the recorded cell has a V_M close to its neighbors at a quasi-physiological level (Ma et al. 2016).

Finally, we used V_M,SS as a readout to evaluate the development of syncytial isopotentiality in the hippocampal CA1 region (Figure 4b, Supporting information -Table 2, n = 397). The V_M,SS in P1–3 astrocytes (P1: V_M,SS, −15.8 ± 5.4 mV, n = 16; P2: V_M,SS, −22.6 ± 4.4 mV, n = 21, p = 1.000; P3: V_M,SS, −19.3 ± 3.8 mV, n = 24, p = 1.000) leaned to the GHK predicted V_M for [Na⁺]_p, indicating the lack of syncytial coupling of neonatal astrocytes. The syncytial coupling significantly increased at P9 compared to the P3 (V_M,SS, −43.8 ± 4.0 mV, n = 35, p = 0.004), and further strengthened at P12 (−63.1 ± 3.5 mV, n = 39, p = 0.020). From P12 onward, syncytial isopotentiality becomes comparable: P13: −72.1 ± 2.8 mV, n = 20, p = 0.984; P14: −73.3 ± 1.1 mV, n = 11, p = 0.994; P15: −73.8 ± 1.1 mV, n = 11, p = 0.990. Compared to P21 (−74.8 ± 0.8 mV, n = 8), a stable level of V_M,SS appeared at P12 (p = 0.993).

The developmental transition of the V_M,_SS followed nicely to a Boltzmann growth function (Adj. R-Square (R²) value, 0.98165; reduced Chi-square (X²) value, 13.72091), where the syncytial isopotentiality (V_M,_SS) reaches a stationary level at the postnatal day 13–15 (Figure 4b).

We used two additional analyses to validate that the syncytial isopotentiality is indeed established around P13-P15. First, we used CV to assess the variability of V_M,SS dataset from P1 to P15 (Supporting information-Table 2). This analysis revealed a continuous decline of CV from P1 (97.6 ± 26.7%, n = 3) to P12 (32.65 ± 22.86%, n = 5, p = 0.120). However, the changes in CV reached the first significant point at P13 compared to P1 (14.22% ± 16.31%, n = 3, p = 0.037), and continued to P15 (4.5% ± 1.7%, n = 2, p = 0.038) compared to that of P1, favoring a view of maturation of syncytial isopotentiality around P13 - P15 (Supporting information-Table 2). In the second analysis, the V_M,SS = −60. 8 mV is identified as a cut-off value to differentiate astrocytes that were coupled to either less or greater than 7–9 astrocytes (Figure 3c). Accordingly, 397 [Na⁺]_p -recorded astrocytes were sorted out into two classes: coupled with <7–9 and >7–9 astrocytes, at each postnatal day (Figure 4c, Supporting information-Table 3). This analysis showed that starting from P14 onward, every recorded cell was coupled to a syncytium with >7–9 astrocytes (Figure 4c).

In summary, based on our observations and analyses, a conservative estimate of the time by which the syncytial isopotentiality reaches maturity should be P15 in the hippocampal CA1 stratum radiatum region (Figure 4a, Figure 9c).

Figure 9. Anatomical and functional maturation of an operational astrocyte syncytium.

Over the course of animal development, the cellular structure and syncytium organization of astrocytes reach initial maturity by P15 (a). b. The expression of K⁺ conductance reaches an initial maturation level by P11. c. The maturation of syncytial coupling, measured by V_{M, SS}, reaches an initial maturation around P15. d. Following the structural and functional maturation of the features noted in a, b, and c, an operational syncytium for efficient spatial transfer of K⁺ ions within a syncytium appears around P15 in the mouse hippocampal CA1 region.

Developmental maturation in astrocyte leak K⁺ conductance

The syncytial isopotentiality results from a spatiotemporal summation of the K⁺ conductance of individual astrocytes in a syncytium; thus, a mature level expression of leak K⁺ channels is critical to maintaining this network feature (Du et al. 2020; Zhou et al. 2021). To determine the time by which the expression of astrocyte leak K⁺ conductance reaches maturity, we analyzed the expression of K⁺ conductance from P1–21 astrocytes.

At P1, every astrocyte exhibited a variably rectifying whole-cell current-to-voltage (I-V) relationship, termed variably rectifying astrocyte (VRA) (Figure 5a), resulting from a combined expression of voltage-gated outward transient K⁺ (K_a), delayed rectifying K⁺ (K_d), and inward rectifying K⁺ (K_in) channels (insets below the Figure 5a) (Zhong et al. 2016; Zhou and Kimelberg 2000). The passive astrocyte (PA), characterized by a linear I-V conductance, first appeared in P2 (Figure 5b), then increased in number with age, until it becomes the only phenotype by P11 (Figure 5c). VRAs and PAs can be differentiated based on two key features. First, a more hyperpolarizing resting V_M, −85.1 ± 0.2 mV (n = 124) in VRAs compared to PAs, −83.5 ± 0.2 mV (n = 109, p = 2.0E-7) (Figure 5d1). Second, a significantly higher R_in in VRAs (R_in, 47.2 ± 5.6 MΩ, n = 150) than that of the PAs (12.3 ± 0.6 MΩ (n = 116, p = 1.0E-7) (Figure 5e1).

Figure 5. Developmental maturation of astrocyte leak K⁺ conductance.

a-b. The whole-cell current profile of one variably rectifying astrocyte (VRA) and one passive astrocyte (PA). The command voltages (V_com) from a holding of −80 mV stepping in the range from −180 mV to +20 mV with 20 mV increments. The combined expression of three major K⁺ channel conductances in VRA: the voltage-gated transient (IK_a), delayed rectifying (IK_d) and inwardly rectifying (IK_in) K⁺ channels that can be isolated based on the biophysical properties of these channels (insets, details see (Seifert et al. 1999; Zhou and Kimelberg 2000)). c. The developmental transition of functional immature VRA and mature PA phenotypes where PA becomes the only electrophysiological phenotype after P11. d. VRAs exhibit a significantly more hyperpolarizing V_M than PAs independent of developmental times (d1). Also, the VRAs remained a constant V_M. The V_M in PAs was initially comparable with VRAs before P6, then shifted towards a more depolarizing value at P7–9 onward. e. The R_in in VRAs continually falls from P1 to P7–9 and becomes stabilized before VRA disappears after P11. The R_in in PAs also falls from P2–3 to P10–12 and then stabilizes thereafter. *: p < 0.05, **: p < 0.005, NS: p > 0.05.

Variably rectifying astrocytes are functionally immature astrocytes:

This conclusion is drawn from the following observations. First, VRA is the only electrophysiological phenotype at P1 that predominantly expresses voltage-gated IK_a and IK_d, a feature that has been associated with proliferating cells (Pardo 2004). Second, the percentage of VRAs declines over time and disappears by P11 (Figure 5c). Third, there is an age-dependent decline in the R_in from 83.4 ± 19.2 MΩ at P1 (n = 39) to 45.2 ± 3.8 MΩ at P2–3 (n = 61, p = 0.037), 26.8 ± 3.0 MΩ at P4–6 (n = 27, p = 0.003), and 16.0 ± 2.0 MΩ at P7–9 (n = 18, p = 0.010). The R_intends to further decline at P10–12 (10.4 ± 1.0 MΩ, n = 5) (Figure 5e2), indicating an age-dependent upregulation in leak K⁺ conductance. Fourth, in line with the same notion, the outward K⁺ conductance (activated at V_com = +20 mV) increased significantly from P1 (3.0 ± 0.2 nA, n = 39) to P10–12 (11.9 ± 0.5 nA, n = 5, p = 2.0E-8). Within the same time, the inward current (activated at V_com = −180 mV) increases from −1.9 ± 0.2 nA (n = 39) to −12.4 ± 0.6 nA (n = 5, p = 2.0E-8) (Figure 6a–c). Fifth, this age-dependent disappearance of voltage-gated IK_a and IK_d, and an increase in leak K⁺ conductance in VRA could be quantitatively revealed by the rectification index (RI) analysis: the RI declined significantly from P1 (2.65 ± 0.37, n = 39) to P4–6 (1.28 ± 0.12, n = 27, p = 0.003) until a plateau that appears at P10–12 (0.96 ± 0.04, n = 5, p = 0.993) (Figure 6d).

Passive astrocytes are functional mature astrocytes:

The electrical passivity is a hallmark of the functional maturation of astrocyte K⁺ conductance (Kafitz et al. 2008; Zhou et al. 2021; Zhou et al. 2006), and this feature has been further quantitatively characterized by a RI value around 0.9 (Du et al. 2016a; Wang et al. 2013). Here we show that this characteristic feature remained throughout development (Figure 6d). In the present study, we have revealed further details through the lens of development. For example, there is a clear developmental transition from VRA to PA phenotypes (Figure 5c). Second, the resting V_M of PAs undergo a depolarizing shift from P1 to P21: the V_M at P2–3 is more hyperpolarized (−85.4 ± 0.8 mV, n = 9) than those at P10–12 and onward (−82.1 ± 0.4 mV, n = 42, p = 0.011) (Figure 5d2). A more depolarized V_M around −78 mV was previously reported from animals older than P21 (Du et al. 2016b; Ma et al. 2016; Wang et al. 2013). Third, an age-dependent increase in leak K⁺ conductance is indicated by a decline in R_in from P2–3 (18.4 ± 3.3 MΩ, n = 9) to P10–12 (9.7 ± 1.0 MΩ, n = 34, p = 0.003), and this low R_in remained from this point onward (Figure 5e2). Fourth, the outward (I_outward) and inward currents (I_inward) of PAs increased significantly during development. The I_outward increased from P7–9 (9.1 ± 0.7 nA, n = 40) to P10–12 (12.0 ± 0.5 nA, n = 18, p = 0.043). The I_inward increased from P7–9 (−10.0 ± 0.8 nA, n = 40) to P10–12 (−14.0 ± 0.8 nA, n = 18, p = 0.010) (Figure 6a–c). Overall, these features reach a stationary level around P11–12. Furthermore, regardless of age, the overall I_outward and I_inward in PAs, (I_outward: 9.7 ± 0.4 nA and I_inward: −10.7 ± 0.5 nA, n = 93) were larger than VRAs (I_outward: 4.8 ± 0.3 nA, p = 2.0E-8 and I_inward: −4.1 ± 0.3 nA, p = 2.0E-8, n = 151). However, it must be noted that the low R_in in PAs imposes a severe voltage deficiency in voltage-clamp recording, therefore, the passive conductance cannot be accurately analyzed in PAs (Zhou et al. 2021).

In summary, at P11, the PA begins to be the only phenotype at P11 with a stabilized resting V_M and low R_in, indicating a mature level of leak K⁺ conductance is established at this developmental point (Figure 6e).

Maturation in the capacity of an astrocyte syncytium for buffering of K⁺ ions

The ability of astrocytes to buffer K⁺ is a well-recognized astrocyte function (Kimelberg 2010; Orkand et al. 1966; Verkhratsky and Nedergaard 2018), which requires a high efficiency of spatial equilibration of K⁺ ions within a syncytium (Du et al. 2018; Ma et al. 2016). Here we evaluate the time by which an astrocyte syncytium reaches an initial maturation for this critical function with our newly designed electrophysiological approach.

Specifically, this evaluation uses [Na⁺]_P and voltage-clamp recording mode (Figure 7a1– b1) where the inward-going K⁺ conductance remains because of the influx of K⁺ ions from the aCSF (3.5 mM) through leak K⁺ channels (Figure 7b1). However, the outward-going K⁺ conductance is eliminated due to the substitution of the endogenous K⁺ by the Na⁺ ions from the [Na⁺]_p. In actuality, the outward-going K⁺ conductance remains intact (Figure 7b1), resulting from the transfer of junctional K⁺ conductance from neighboring astrocytes (Du et al. 2018; Ma et al. 2016; Stephan et al. 2021; Wang et al. 2020b) (Figure 7b1).

Indeed, decoupling of syncytium (by 100 μM MFA) (Figure 71b) significantly eliminated the outward-going K⁺ conductance in [Na⁺]_p-recorded astrocytes (aCSF, n = 11, MFA, n = 10; p = 7.4E-4 at step V_com Δ+60 mV, p = 1.E-7 at step V_com Δ+80 mV, p = 1.0E-12 at step V_com Δ+100 mV) (Figure 7a2, b1–3, Supporting information - Table 4). In this set of experiments, we next removed 3.5 mM K⁺ from aCSF, which resulted in an anticipated abolishment of the inward-going K⁺ conductance (n = 5, p = 0.01404 at step V_com Δ−40 mV, p = 8.5E-5 at step V_com Δ−60 mV, p = 5.8E-8 at step V_com Δ−80 mV, p = 6.5E-12 at step V_com Δ−100 mV) (Figure 7c1–3, e1, Supporting information - Table 4). Note that decoupling also resulted in a significant reduction in the inward-going K⁺ conductance (Figure 7b3, Supporting Information-Table 4), indicating that a significant amount of the inward-going K⁺ conductance is also derived from junctional K⁺ conductance (I_{K, syncytium}). Additionally, we showed that the inward-going K⁺ conductance can be significantly inhibited by 100 μM BaCl₂ (n = 8), confirming that inward rectifiers, including Kir_4.1, are the major channels mediating inward-going leak K⁺ conductance (Figure 7d1–3, e2, Supporting information - Table 4) (Zhou et al. 2021).

These experiments demonstrated that the outward-going K⁺ conductance in a [Na⁺]_p recorded astrocyte is “outsourced” from the coupled neighbors and the amplitude of this current component is a readout of the capacity of a syncytium for spatial transfer of K⁺ ions.

Now, turning back to the original question of the time in which an astrocyte syncytium becomes operational in the brain, in Figure 8a we show a [Na⁺]_P recording from a P3 solitary astrocyte. Only a tiny inward-going endogenous K⁺ current (I_{K, endogenous}) could be induced, which is consistent with a low-density expression of inward rectifier K⁺ conductance early in life (Figure 6a) (Seifert et al. 2009; Zhong et al. 2016). Meanwhile, there was a total absence of outward-going K⁺ conductance from syncytium (I_{K, syncytium}), attributed to the lack of gap junction coupling (Figure 8a). Shown in Figure 8b is a recorded P9 astrocyte (Figure 3b) where there was a significant and moderate increase in I_{K, endogenous} and I_{K, syncytium}, respectively. An increase in intrinsic K⁺ conductance expression and a partial connection to neighboring astrocytes altogether contribute to a significant increase in I_{K, endogenous}, whereas the outward-going K⁺ conductance was mostly made up by the I_{K, syncytium}. Shown in Figure 8c was a P15 (Figure 3b) where a functionally mature syncytium made a full compensation for the passive K⁺ conductance; most noticeably the missing outward-going K⁺ component.

It should be noted that the exact contribution of I_K,syncytium to the inward-going K⁺ conductance could not be quantified, therefore, we denoted only outward-going K⁺ conductance as the I_K,syncytium in Figure 8. However, we emphasized this important conclusion, i.e., a substantial contribution of I_K,syncytium to the inward-going K⁺ conductance (Figure 7b1–3), in the summary illustration in Figure 9d.

Quantitatively, both I_K,endogenous and I_{K, syncytium} increase with age and reaches a stationary level at P15. The outward current (I_K,syncytium) increased from P1–3 (0.3 ± 0.1 nA at ΔV_com=100 mV, n = 9) to P10–12 (1.9 ± 0.7 nA, n = 11, p = 0.045) and further increased to P13–15 (5.8 ± 1.0 nA, n = 5, p = 7.0E-7). The inward current (I_K,endogenous) in [Na⁺]_p -recorded astrocytes also gradually increased during the development. The I_K,endogenous (at ΔV_com = −100 mV) for P1–2, P7–9, P10–2 and P13–15 were −0.6 ± 0.1 nA (n = 9), −3.2 ± 0.2 nA (n = 46, p = 2.0E-4), −4.3 ± 0.7 nA (n = 11, p = 2.0E-5) and −7.2± 0.5 nA (n = 5, p = 4.0E-9), respectively. Both I_K,endogenous and I_{K, syncytium} from P13–15 astrocytes were comparable with P21 (I_{K, endogenous}: 7.1 ± 0.7 nA, n = 4, p =0.999; I_{K, syncytium}: 6.9 ± 0.7 nA, n = 4, p = 0.660).

To further evaluate the maturation state of syncytium for intra-syncytial buffering of K⁺ ions, we introduced the RI analysis. The first RI analysis was performed from recordings made with [Na⁺]_p for the age-dependent increase in I_{K, syncytium}. The RI values showed no significant change during early development from P1–3 (0.58 ± 0.10, n = 9), P4–6 (0.38 ± 0.05, n = 19, p = 0.243) to P10–12 (0.39 ± 0.06, n = 11). A significant increase occurs at P13–15 (0.81 ± 0.13, n = 5, p = 0.003), and that remained constant until P21 (0.99 ± 0.02, n=4, p = 0.825) (Figure 9C). We next used RI of mature astrocytes recorded with [K⁺]_p (0.98 ± 0.01, n = 4) as control and compared it with [Na⁺]_p recorded astrocytes at different age groups. We found a significant deficiency in RI (as a readout for I_{K, syncytium}) in early age groups: P1–3 (0.58 ± 0.10, n = 9, p =0.040), P4–6 (0.38 ± 0.05, n = 19, p = 5.0E-5), P7–9 (0.38 ± 0.03, n = 46, p = 1.0E-5), and P10–12 (0.39 ± 0.06, n = 11, p = 2.0E-4). The RI of [Na⁺]_p recorded astrocytes became comparable with those recorded with [K⁺]_p at P13–15 (0.81 ± 0.13, n = 5, p = 0.883) and P21 (0.99 ± 0.02, n = 4, p =1.000). These analyses corroborated the notion that an astrocyte syncytium becomes operational for intra-syncytium buffering of K⁺ ions around P15 in the mouse hippocampal CA1 region (Figure 9d).

DISCUSSION

Astrocytes establish the largest syncytial network across the brain. Although this was unequivocally shown by dye coupling over 4 decades ago (Gutnick et al. 1981), the question of the time in which an astrocyte syncytium becomes functionally operational in the developing brain remains unknown. Here we show, for the first time, that an astrocyte syncytium reaches an initial maturity in terms of the syncytial organization and efficacy for intra-syncytium equilibration of K⁺ ions at P15 in the developing mouse hippocampus (Figure 9).

Newborn astrocytes are solitary cells

The first wave of astrocyte production peaks around P1–2 and mainly derives from the asymmetric division of glial progenitors and the direct transformation of subventricular zone (SVZ) radial glia (Levison and Goldman 1993; Magavi et al. 2012; Tsai et al. 2012). The newborn astrocytes lack gap junctional coupling, and therefore, appear to be mostly solitary individuals in contrast to mature astrocytes that are aggregated into a singular syncytium (Zhong et al. 2016). In the present study, we have provided additional evidence in support of this notion.

First, none of the P1 astrocytes showed LY dye coupling to neighboring astrocytes (Figures 1, 3), and the electrical coupling was also absent in 30% of newborn astrocytes (Figure 1h–k). The latter method has the highest sensitivity at detecting gap junctional coupling (Stephan et al. 2021). Second, in the syncytial isopotentiality measurement, most of the P1–3 astrocytes were either weakly coupled or completely uncoupled to a syncytium (Figure 4). Interestingly, this feature differs strikingly from those astrocytes generated later in the P6–13 postnatal brain, where astrocytes are produced mainly through symmetrical division from differentiated astrocytes so that they are already integrated as part of syncytium at birth (Ge et al. 2012).

The cellular structure and syncytial organization of astrocytes reaches maturity at P15

An ellipsoid shape of astrocytes is an essential anatomical necessity to fit in a functional syncytium as a building block (Aten et al. 2022b). We show that the domain morphology and the pattern of syncytial organization of astrocytes undergo substantial remodeling during development. First, in mouse hippocampal CA1 region, neonatal astrocytes are highly polarized along the apical dendrites of the pyramidal neuron layer with extensive filamentous processes. As astrocytes mature, their cellular morphology becomes more rounded, and the processes ramify into a spongiform morphology (Figure 2). Such dramatic alteration in cellular structure is most pronounced within the first 10 postnatal days and reaches a stationary level around P15 (Figure 2). These observations are consistent with the reports by Bushong and colleagues (Bushong et al. 2004; Bushong et al. 2002).

Along with the change in astrocyte cellular morphology, there is an equally notable alteration in the syncytial organization; specifically, a decrease in astrocyte cell density and an increase in the interastrocytic distance with age (Figure 2). The refinement of the process arborization, domain polarization, and syncytial anatomy is likely a critical step for the establishment of mutually exclusive domains and interastrocytic gap junction coupling.

Astrocytes converge into a syncytium by P15

Now the genetic/molecular programs directing the wiring of neuron-astrocyte networks in the developing and adult brain is an active research area (Akdemir et al. 2020; Farmer et al. 2016; Holt et al. 2019; Martin et al. 2015; Stogsdill et al. 2017; Stork et al. 2014). In the present study, we have focused on the structural-functional aspect of the time astrocytes integrate into an operational system.

Although the intrinsic and extrinsic molecular programs that guide the integration of these astrocytes into a syncytium remain elusive, the solitary individuals seemingly do need to communicate with one another, modify their process arborizations and dock the hemichannels from each side to form functional interastrocytic gap junctions. How exactly these sophisticated processes are regulated is an interesting question to be determined in the future. In contrast, a more pronounced expansion of astrocyte population towards the end of the second postnatal week is mainly produced by the symmetrical division of differentiated astrocytes (Ge et al. 2012). These astrocytes do not physically separate from the existing syncytium and have already adopted a passive membrane K⁺ conductance. It is conceivable that a significant advantage of taking this cellular division process is to avoid going through the same developmental process as those astrocytes produced in the early stage of asymmetrical cell division. Evidently, as more astrocytes are produced from P6 - P13, the immature electrophysiological phenotype (VRA) never appears after P11(Ge et al. 2012). The developmental continuity in terms of the upregulation of K⁺ conductance, the emergence of syncytial isopotentiality, and the increased efficacy of K⁺ ion redistribution in a syncytium shows no sign of interruption over the following postnatal development (Figures 2, 4, 5, 6, and 8).

Astrocytic leak K⁺ conductance is a shared feature of a syncytium

We have previously shown that the newborn astrocytes start to share their membrane leak K⁺ conductance as early as postnatal day 2. Specifically, a small fraction of astrocytes begins to “outsource” K⁺ conductance from neighboring astrocytes through their better-established gap junctional coupling. This converts some low leak K⁺-expressing VRAs to the high leak K⁺-expressing PAs, and that the PA phenotype can be totally reverted back to VRA after decoupling (by 100 μM MFA) (Zhong et al. 2016).

In the present study, the following evidence further supports the notion that the copiously expressed K⁺ conductance in the individual astrocytes is a shared feature of an astrocytic network. First, there is a clear transition from VRA to PA over the course of development (Figure 5). VRAs exhibit a continuous decline in R_in and growing inward and outward K⁺ conductances until the VRA phenotype disappears at P12. Second, in both VRAs and PAs, a progressive sharing of K⁺ conductance is shown by a decrease in R_in from P2 to P9 (Figure 5e), and an increase in inward and outward passive conductance from P2 to P15 (Figure 6c). Third, in a syncytial coupled astrocyte, depletion of endogenous K⁺ content by [Na⁺]_p is anticipated to abolish the outward-going K⁺ conductance. We have speculated in the past that the unexpected outward-going K⁺ passive conductance observed from [Na⁺]_p recording is junctional conductance deriving from the neighboring astrocytes (Ma et al. 2016). Now, our new finding, i.e., the outward-going K⁺ conductance can be removed by inhibition of gap junctional coupling, directly confirms our early speculation (Figure 7).

An operational syncytium for buffering of K⁺ ions emerge at P15

A [Na⁺]_P recording method and an associated computational model have been used to quantitatively monitor the dynamic change in astrocyte syncytial isopotentiality (Kiyoshi et al. 2018; Ma et al. 2016). In the present study, this method is further evolved to evaluate the capacity of K⁺ spatial buffering with a syncytium. As noted in the above section, the “concentration clamping” of an astrocyte by the [Na⁺]_p results in near total elimination of outward K⁺ conductance due to the depletion of intracellular K⁺ ions, and this has been demonstrated from single freshly dissociated astrocytes (Du et al. 2018; Kiyoshi et al. 2018; Ma et al. 2016). This experimental condition also selectively increases the membrane resistance (R_M) in the outward direction, which allows the depolarizing voltage commands to spatially expand into the neighboring astrocytes that drive the efflux of K⁺ ions through the leak K⁺ channels. This is the ionic mechanism underlying the outward-going K⁺ conduction in [Na⁺]_p-recorded astrocytes (Figures 7, 8). Therefore, the outward-going K⁺ conductance from a [Na⁺]_P-recorded astrocyte is mainly derived from the neighboring astrocytes. Accordingly, we have revealed an age-dependent increase in the compensatory outward-going K⁺ currents (I_{K, syncytium}) in [Na⁺]_p-recorded astrocytes (Figures 8, 9d). This observation has allowed us to conclude that an operational syncytium for K⁺ buffering initially arises at P15.

Although our experiments focused on the age- and syncytial coupling-dependent capacity of a syncytium to compensate for the outward-going K⁺, this conclusion should be equally applicable to the inward-going K⁺ conductance, which has been demonstrated by the suppression of this current component after uncoupling of syncytium (Figure 7b). By extension, the [Na⁺]_p recording method should have broad usage in the future to examine the syncytial function under both physiological and pathological conditions.

A functional readout for the maturity of an astrocyte syncytium

The frequently discussed K⁺ spatial buffering hypothesis is still subject to full experimental validation (Ransom 1996; Stephan et al. 2021). As noted above, in the present study, we have extended [Na⁺]_P recording mode to examine the efficacy of a syncytium in redistributing K⁺ ions. Together with syncytial isopotentiality measurement, the same recording method provides additional insights into the operation of an astrocyte network; in this case, is the age-dependent maturation of astrocyte syncytium in mouse hippocampus.

We have previously shown that by experimentally creating a “K⁺-deficient” astrocyte inside a mature astrocyte network with a [Na⁺]_P, a nearly identical passive K⁺ conductance retains (Ma et al. 2016). Now we demonstrate that the outward-going K⁺ is primarily stemming from the coupled syncytium, and the levels of compensation for the outward-going K⁺ depend on the strength of syncytial coupling that increases along the course of syncytial development. In the present study, we also show that the compensatory capacity of the syncytial K⁺ conductance to a [Na⁺]_p-recorded astrocyte critically depends on two interrelated factors: 1) the levels of expression of leak K⁺ conductance in individual astrocytes and 2) the levels of the maturity of syncytial gap junctional coupling (Figures 4, 5, 6, 9).

Although an isopotential network minimizes the diffusion of electrode Na⁺ into the syncytium, there should still be a large Na⁺ flow into the syncytium. The fact that a V_{M, SS} can remain constant for hours indicates a powerful Na⁺ extrusion capacity of an astrocyte syncytium. This is consistent with the observation of large amounts of mitochondrial networks in astrocytes to meet the energy demand for this intracellular Na⁺ clearance function (Aten et al. 2022b).

The syncytial isopotentiality measurement has already been used to study the mechanism among microglia-astrocyte-neurons crosstalk (Du et al. 2022), and changes in the functional state of an astrocyte syncytium in diseased conditions (Aten et al. 2022a; Wang et al. 2020a). Now a variation of this method can be used to examine the efficacy of the K⁺ buffering within a syncytium under both physiological and pathological conditions in the future.

Concluding remarks

In summary, within the first two postnatal weeks, the newborn astrocytes undergo a rapid refinement in their morphology, spatial organization, K⁺ channel expression, and intercellular gap junction coupling that transforms them from solitary individuals to a functionally mature syncytial network. In the hippocampus, an operational syncytium initially arises at postnatal day 15.

MAIN POINTS.

• Astrocyte syncytium is anatomically established at P15.

• Astrocyte leak K⁺ conductance reaches a mature level of expression at P11.

• Syncytial isopotentiality emerges in the hippocampus by P15.

• Astrocyte syncytium is functionally operational at P15 in the hippocampus.