Cortical layer 1 and layer 2/3 astrocytes exhibit distinct calcium dynamics in vivo

Abstract

Cumulative evidence supports bidirectional interactions between astrocytes and neurons, suggesting glial involvement of neuronal information processing in the brain. Cytosolic calcium (Ca(2+)) concentration is important for astrocytes as Ca(2+) surges co-occur with gliotransmission and neurotransmitter reception. Cerebral cortex is organized in layers which are characterized by distinct cytoarchitecture. We asked if astrocyte-dominant layer 1 (L1) of the somatosensory cortex was different from layer 2/3 (L2/3) in spontaneous astrocytic Ca(2+) activity and if it was influenced by background neural activity. Using a two-photon laser scanning microscope, we compared spontaneous Ca(2+) activity of astrocytic somata and processes in L1 and L2/3 of anesthetized mature rat somatosensory cortex. We also assessed the contribution of background neural activity to the spontaneous astrocytic Ca(2+) dynamics by investigating two distinct EEG states ("synchronized" vs. "de-synchronized" states). We found that astrocytes in L1 had nearly twice higher Ca(2+) activity than L2/3. Furthermore, Ca(2+) fluctuations of processes within an astrocyte were independent in L1 while those in L2/3 were synchronous. Pharmacological blockades of metabotropic receptors for glutamate, ATP, and acetylcholine, as well as suppression of action potentials did not have a significant effect on the spontaneous somatic Ca(2+) activity. These results suggest that spontaneous astrocytic Ca(2+) surges occurred in large part intrinsically, rather than neural activity-driven. Our findings propose a new functional segregation of layer 1 and 2/3 that is defined by autonomous astrocytic activity.

Figure 1. Distinct cytoarchitecture of cortical layer 1 and layer 2/3.

A, Representative in vivo images of layer 1 (L1) and layer 2/3 (L2/3). The Ca²⁺ indicator Oregon Green 488 BAPTA-1 (OGB-1, green) stained both neurons and astrocytes while Sulforhodamine 101 (SR101, red) stained astrocytes. In L1, most of the cells were astrocytes, indicated by loading with OGB-1 and SR101 (red–yellow). In L2/3, neurons (SR101-negative) outnumber astrocytes (SR101-positive). Time course of Ca²⁺ transients of the numbered astrocytes (1–6) in L1 image is presented in Figure 2A (for L2/3, see Figure S1). Movies of Ca²⁺ transients for L1 and L2/3 are shown in Video S1 and S2. B, The packing densities of astrocytes and neurons are plotted against the depth from cortical pia. L1 (blue circle) and L2/3 (red triangle) were determined by the physical depth and appearance of astrocyte-neuron ratio. Counts of astrocyte and neuron were 19.8±5.6 and 8.9±4.5 in L1 (n = 38 imaging sites), 12.6±4.8 and 63.0±17.1 in L2/3 (n = 34) per imaged area (∼320×320 µm²) (mean±SD in all figures, otherwise noted). While L1 has more astrocytes than layer 2/3, neurons dominated in L2/3 (***p<0.001, Student's t test). C, Immunohistochemical staining for S100β (astrocyte marker, red) and NeuN (neuron marker, green) of a coronal cortical section. Glial dominance in L1 and neuronal dominance in L2/3 is evident. Scale bar: A, 100 µm; C, 200 µm.

Figure 2. More astrocytes elicit spontaneous Ca²⁺ surges in layer 1 than in layer 2/3.

A, Representative time course of spontaneous Ca²⁺ surges of astrocytes in L1. Numbers (1–6) correspond to astrocytes in Figure 1A. Each trace shows normalized fluorescence intensity of Ca²⁺ indicator OGB-1 from an individual astrocyte. Small red dots indicate period of Ca²⁺ surge. Vertical position of each trace was adjusted arbitrary to improve visibility. Scale bar: 50%. B, Number of astrocytes with at least one spontaneous Ca²⁺ surge (“active astrocyte”) per imaged area is plotted against the depth from the cortical pia. L1 had significantly more active astrocytes than L2/3. ***p<0.001, Student's t test. C, Proportion of active astrocyte for each experiment is plotted against the imaged depth. L1 had significantly higher percentage of active astrocytes than L2/3. ***p<0.001, Student's t test. D, Histograms (solid bars) and cumulative distribution (solid lines) of Ca²⁺ surge duration of astrocytes in L1 (blue) and L2/3 (red) are plotted. Vertical dotted lines indicate the mean duration of spontaneous Ca²⁺ surge. There was no significant difference in Ca²⁺ duration between L1 and L2/3. E, Histograms (solid bars) and cumulative distributions (solid lines) of Ca²⁺ surge peak of astrocytes in L1 (blue) and L2/3 (red) are plotted. L1 astrocytes had significantly higher spontaneous Ca²⁺ surge peak values. Vertical dotted lines indicate the mean Ca²⁺ surge peak (***p<0.001, Student's t test). F, Ca²⁺ surge peak values are plotted against Ca²⁺ surge duration in L1 (blue circle) and L2/3 (red triangle). There was a slight but significant correlation between Ca²⁺ peak value and duration in L1 (r≈0.24), but little correlation was observed in L2/3 (r≈0.07).

Figure 3. Ca²⁺ surges occurrence in astrocytes are non-uniform.

The distribution of astrocytic spontaneous Ca²⁺ surge frequency in L1 (blue bar, left graph) or L2/3 (red hashed bar, right graph) was compared with the Poisson distribution (grey bar) that represents unbiased occurrence probability of Ca²⁺ surges among astrocytes. Distribution of experimental data deviated significantly from the Poisson distribution (L1: χ² = 4586, d.f. = 5, p<0.001; L2/3: χ² = 1316, d.f. = 2, p<0.001). “7+” and “4+” indicate summed counts of Ca²⁺ surge-number equal to or larger than 7 and 4, respectively. Note the large deviation between Poisson distribution and experimental results, especially at Ca²⁺ surge counts 0∼1 and 7+ (L1) or 4+ (L2/3).

Figure 4. Condition of craniotomy assessed with spontaneous Ca²⁺ surge occurrence and immunohistochemistry.

A, Number of spontaneous Ca²⁺ surges of astrocytes of all experimental data was plotted against imaged time. Red line indicates the cumulative proportion. Apparently uniform occurrence probability of spontaneous Ca²⁺ surge was observed. B, Comparison of number of Ca²⁺ surges during the first and the last 10 min of the imaging period. No significant difference was observed (p = 0.32). C, Wide field view of immunostained cortex with astrocyte specific antibody, GFAP (left). Square indicates the imaged area with craniotomy, which is displayed in an expanded view (right, ‘craniotomy’) and compared with the contralateral side of the corresponding cortical area (middle, ‘contra’). Comparable shape of astrocyte was acknowledged. D, Wide field view of immunostained cortex with microglia specific antibody, Iba-1 (left). Square indicates the imaged area with craniotomy, which is displayed in an expanded view (right, ‘craniotomy’) and compared with the contralateral side of the corresponding cortical area (middle, ‘contra’). Similar morphology of microglia was observed. Scale bar: C, D, 1 mm (left); 100 µm (middle and right).

Figure 5. Spatio-temporal correlation of Ca²⁺ activity of active astrocytes was low.

A, Mean cross-correlation of Ca²⁺ sensitive fluorescence intensity (ΔF/F₀) of pairs of “active astrocytes” (astrocyte with as least one Ca²⁺ surge) in L1 (2300 pairs of active astrocytes) and L2/3 (181 pairs). Correlation coefficients (r-value) for both L1 and L2/3 suggest that temporal correlation among active astrocytes is weak (r<0.15). B, Relationship between distance of active astrocytes and the magnitude of correlation at zero time lag. The correlation coefficients were −0.07 and −0.21 for L1 and L2/3, respectively, showing little association between cell to cell distance and synchrony of Ca²⁺ surges when computed over the whole population.

Figure 6. Pairs of synchronously active astrocytes were located closer.

A, Histograms of cell to cell distance of astrocyte pairs. “All pairs” contains pairs of astrocytes whether they have Ca²⁺ surge or not during the 30 min imaging session (L1: 7625 pairs, L2/3: 2843 pairs). “Active all” includes pairs of an astrocyte which has at least one Ca²⁺ surge (L1: 2481 pairs, L2/3: 181 pairs). Pairs of “active astrocyte” were separated into two categories: “active p>0.05” and “active p<0.01” according to their synchrony of Ca²⁺ surges. Synchrony of Ca²⁺ surge between a pair of astrocytes was estimated with Fisher's exact test (see method). Highly synchronized “active p<0.01” pairs are in closer proximity than other pairs of astrocytes (p<0.001 in L1, p<0.02 in L2/3. ANOVA followed by Tukey's HSD test). Vertical dotted lines indicate the mean distance of astrocyte pairs. Cumulative plot for the distributions are plotted in B. C, Histograms of distance of all the imaged astrocytes separated in the order of physical proximity. “1st closest pair” indicates the distribution of distances of a pair of closest astrocytes. The histogram shows that pairs of astrocyte with synchronized Ca²⁺ surge (“active p<0.01”) are not necessarily the closest pairs.

Figure 7. Low corelation between cortical EEG states and spontaneous astrocytic Ca²⁺ surges.

A, Representative EEG spectrogram of cortex shows alternating epochs of synchronized and desynchronized neuronal activity under urethane anesthesia (upper graph). Synchronized and desynchronized activities are characterized by high EEG power of slow wave (0.5–2 Hz) and theta range frequency (3–4 Hz), respectively. To quantify states of EEG activity, power ratio of EEG frequency of 0.5–2 Hz over 3–4 Hz was calculated (lower graph). B, Representative EEG traces for synchronized and desynchronized states. Synchronized activity is characterized by high amplitude slow wave. Desynchronized state of EEG is characterized as smaller and rapid activity. Scale: 1 sec (horizontal), 100 µV (vertical). C, Histogram of EEG power ratio of [0.5–2 Hz] and [3–4 Hz] during whole imaging session (blue bar) and during periods with astrocytic spontaneous Ca²⁺ activity (red bar). Vertical dotted lines indicate the mean ratio. Mean ratio during whole imaging session and periods with astrocytic Ca²⁺ surges did not differ significantly (Wilcoxon rank sum test, p = 0.35 for L1, p = 76 for L2/3).

Figure 8. Spontaneous Ca²⁺ surges of astrocytes were not affected by pharmacological inhibitors.

Percentage of active astrocytes in inhibitor experiments was compared with control experiments. The mGluR5 blocker MPEP, or the P2Y receptor blocker PPADS, the voltage gated sodium channel blocker TTX, or the muscarinic acetylcholine receptor blocker atropine had no effects on the occurrences of spontaneous astrocytic Ca²⁺ surges.

Figure 9. Dynamics of Ca²⁺ activities of astrocytic processes differ between layer 1 and layer 2/3.

Astrocyte's soma and processes were identified with astrocyte specific SR101 image (A). Three primary processes (proc) of an astrocyte were arbitrarily selected and two rectangular ROIs were placed at each primary process. One ROI was placed at proximal (p) part of the process and the other was at distal (d) part. Ca²⁺ activity was monitored throughout the imaged session for each of the chosen region (B). As a control, one ROI was placed at neuropil at least 100 µm away from the astrocyte. The control ROI had the same size as the ROI of astrocytic soma. Representative image (A) and Ca²⁺ traces (B) are shown for an astrocyte in L1. For an astrocyte in L2/3, see Figure S3. C, Pearson's linear correlations (r-value) between Ca²⁺ traces of ROIs were calculated for L1 (30 processes of 10 astrocytes from 3 animals, upper panel) and L2/3 (30 processes of 10 astrocytes from 3 animals, lower panel). The correlation coefficient of ‘within-process’ (e.g. proc1p vs. proc1d) ROIs was high in both L1 & L2/3. Note that the between-processes correlation was high only in L2/3 (e.g. proc1p vs. proc2p in layer 2/3). D, Comparisons of the average correlation coefficient of within-process and between-process for L1 and L2/3. Note that while the correlation coefficient of ‘within process’ is high (r>0.3) in both L1 and L2/3, the correlation of ‘between process’ is high only in L2/3. *p<0.05, ***p<0.001; one-way ANOVA followed by Tukey's HSD test. Scale bar: A, 20 µm; B, 100%.