Differential subcellular Ca²⁺ signaling in a highly specialized subpopulation of astrocytes

Abstract

Recent evidence suggests that astrocytes do not serve a mere buffering function, but exhibit complex signaling pathways, disturbance of which contributes significantly to the pathophysiology of CNS diseases. Little is known regarding the intracellular signaling pathways in specialized optic nerve head astrocytes (ONHAs), the major glia cell type in non-myelinated optic nerve head. Here we show the differential subcellular expression of intracellular Ca²⁺ channels in ONHAs. Expression of type 1 and type 3 inositol-1-4-5,-trisphosphate receptors (IP₃Rs) in the endoplasmic reticulum and type 2 IP₃Rs the nuclear envelope causes differential Ca²⁺ release from intracellular stores in nuclear vs. cytosolic compartments. Our study identifies differential distribution and activity of Ca²⁺ channels as molecular substrate and mechanism by which astrocytes independently regulate Ca²⁺ transients in both cytoplasm and nucleoplasm, thereby controlling genomic and non-genomic cellular signaling, respectively. This provides excellent targets for therapeutics restoring pathological disturbances of intracellular Ca²⁺ signaling present in glaucoma and other neurodegenerative disorders with astrocyte involvement.

Introduction

Over the past two decades, the appreciation of astrocytes has changed from considering them merely as metabolic buffering and support system to attributing them a crucial function in synaptic transmission and nervous system physiology (Dallerac, et al., 2013), This is largely the result of several studies in brain astrocytes showing that this cell type employs complex intracellular signaling pathways including calcium signaling (Cornell-Bell, et al., 1990, Muller, et al., 2013, Tong, et al., 2013, Yan, et al., 2013) thought to be unique for neurons (Berridge, 1998). However, detailed mechanistic analyses of such signaling pathways to identify both their molecular components and biophysical and cell biological properties, at the level of necessary detail of work done with neurons, are still needed for astrocytes in most CNS diseases and disease models. Optic nerve head astrocytes (ONHAs) are the major glia cell type in the non-myelinated optic nerve head, where they are the major contributor to extracellular matrix synthesis during development and throughout life (Hernandez, 2000). In glaucoma, a disease affecting more than 60 million people worldwide (Quigley, 2011), pathological changes include altered astrocyte gene and protein expression resulting in activation and extracellular matrix remodeling with little known regarding underlying signaling pathways. Investigation of the intracellular signaling pathways in ONHAs may thus provide novel avenues for target identification and drug discovery.

Cellular Ca²⁺ homeostasis is tightly regulated and changes in the intracellular Ca²⁺ concentration control important physiological processes, especially in cell types such as neurons where they include neurotransmitter release and gene expression (Duncan, et al., 2010). In neurons and brain astrocytes alike, these changes are mediated by either Ca²⁺ influx through (voltage-gated) plasma membrane Ca²⁺ channels (VGCCs) or release from intracellular Ca²⁺ stores (Beck, et al., 2004, Gleichmann and Mattson, 2011, Nag, 2011, Venance, et al., 1997, Worley, et al., 1987). Intracellular Ca²⁺ release is driven mainly by inositol-1,4,5,-trisphosphate receptors (IP₃Rs) and ryanodine receptors (RyR) (Berridge, 1998, Koulen and Thrower, 2001), with polycystin-2 ion channels largely serving to amplify their activity upon prolonged depolarization (Koulen, et al., 2002, Koulen and Thrower, 2001). In addition to these ion channels found in membranes of intracellular Ca²⁺ stores, mechanisms controlling Ca²⁺ uptake into intracellular stores or Ca²⁺ extrusion into the extracellular space contribute to cellular Ca²⁺ homeostasis (Koulen and Thrower, 2001).

Genetic, age-related and pathological changes in Ca²⁺ signaling underlie a significant number of neurological and neurodegenerative disorders, including glaucoma (Duncan, et al., 2010, Gleichmann and Mattson, 2011, Payne, et al., 2013). Targeting aberrant Ca²⁺ signaling in neurodegeneration is widely considered a promising strategy and the target of significant drug development efforts (Payne, et al., 2013), and identifying the molecular determinants of intracellular Ca²⁺ signaling in ONHAs is critically important for drug discovery for glaucoma and related disorders affecting the ON and ON head.

Herein we identified differentially distributed inositol-1,4,5-trisphosphate receptors (IP₃Rs) and ryanodine receptors (RyRs) in ex vivo rat ONHAs based on immunocytochemistry and optical imaging of pharmacologically-elicited changes in the intracellular Ca²⁺ concentration. Our data are the first report of a detailed structure-function relationship of Ca²⁺ signaling in this cell type of high clinical relevance and provide a novel basis for future glioprotection studies of ONHAs.

Materials and Methods

Primary culture of rat optic nerve head astrocytes

The protocol for the present studies was approved by the Institutional Animal Care and Use Committee at the University of Missouri – Kansas City, and was executed in accordance with the ARVO Statement on the Use of Animals in Ophthalmic and Visual Research.

The isolation of ONHAs was modified from the report by Murphy et al.(Murphy, et al., 2010) Three months old male Brown Norway rats were euthanized and optic nerve heads with approximately 2 mm optic nerve stump attached were dissected and cultured retina side down in a well of a 24-well plate (TTP, Midwest Scientific, St. Louis, MO), in Dulbecco’s Modified Eagle’s Medium (Lonza, Walkersville, MD) supplemented with 20% fetal bovine serum (PAA Laboratories, GE Healthcare Bio-Sciences Corp., Piscataway, NJ) 100 U/mL penicillin, and 100 µg/mL streptomycin. At 5 days in vitro (DIV), astrocytes had migrated onto the plate as confirmed under a tissue culture microscope (DM IL, Leica Microsystems, Buffalo Grove, IL). Remaining tissue and media were aspirated by vacuum suction and fresh media added. Media was subsequently refreshed every 72 hr. At DIV 10, cells had reached approximately 80% confluency in the well and were passaged into a T25 tissue culture flask (TPP, Midwest Scientific, St. Louis, MO) by trypsinization (0.25% Trypsin, 2.21 mM EDTA in Hank’s Balanced Salt Solution; MediaTech Inc., Manassas, VA). ONHA culture was successfully maintained for more than 10 passages at a sub-culturing ratio of 1:4 every 72 hr. Passages 3 – 6 were used for the experiments described herein.

Immunocytochemistry and image acquisition

Immunocytochemistry was performed essentially as described previously (Kaja, et al., 2011, Kaja, et al., 2011, Kaja, et al., 2012). Briefly, ONHAs were seeded onto poly-L-lysine coated glass coverslips (BD Biosciences, San Jose, CA) at a density of 25,000 cells. Cells were fixed in 4% paraformaldehyde in 0.1 M phosphate buffered saline for 15 min. The primary antibodies against glial fibrillary acidic protein (GFAP), IP₃Rs and RyRs have previously been validated extensively for use in immunocytochemistry and are listed in Table 1. Secondary antibody was AlexaFluor^® 488 labeled goat anti-rabbit IgG (1:2,000 dilution; Life Technologies, Carlsbad, CA). Cells were co-labeled with AlexaFluor^® 647 Phalloidin (1:200 dilution; Life Technologies, Carlsbad, CA) to visualize actin filaments, and Hoechst 33258 (120 ng/µL; Enzo Life Sciences Inc., Farmingdale, NY) to label cell nuclei. Cells were mounted using AquaPolymount (Polysciences Inc., Warrington, PA). Images were acquired using a Leica SP5X WLL microscope (Leica Microsystems Inc., Buffalo Grove, IL). Single optical sections are shown.

Table 1.

Antibodies used for immunocytochemistry

Epitope	Host	Cat#	Supplier	Dilution	Reference
anti-GFAP	Rabbit	Ab4674	AbCam, Cambridge, MA	1:2,000
Anti-IP3R1	Goat	sc-6093	Santa Cruz Biotechnology, Santa Cruz, CA	1:50	(Payne et al., 2013a)
Anti-IP3R2	Goat	sc-7278	Santa Cruz Biotechnology, Santa Cruz, CA	1:50	(Payne et al., 2013a)
Anti-IP3R3	Goat	sc-7277	Santa Cruz Biotechnology, Santa Cruz, CA	1:50	(Payne et al., 2013a)
anti-RyR1	Rabbit	AB9078	Chemicon; EMD Millipore, Bellerica, MA	1:15,000	(Medina-Ortiz et al., 2007; Payne et al., 2013a)
anti-RyR2	Rabbit	AB9080	Chemicon; EMD Millipore, Bellerica, MA	1:15,000	(Medina-Ortiz et al., 2007; Payne et al., 2013a)
anti-RyR3	Rabbit	AB9082	Chemicon; EMD Millipore, Bellerica, MA	1:15,000	(Medina-Ortiz et al., 2007; Payne et al., 2013a)

We used the above subtype-specific antibodies to determine the localization and distribution of IP₃Rs and RyRs in primary cultured rat ONHAs. All antibodies are well-characterized and their specificity for their respective epitopes has been shown previously.

Optical imaging of the intracellular calcium concentrations

Calcium imaging experiments and analysis were performed essentially as described previously.(Payne, et al., 2013) Experiments were conducted in Extracellular Solution buffer (ECS; 137 mM NaCl, 5 mM KCl, 1 mM Na₂HPO₄, 1 mM MgSO₄, 10 mM HEPES, 22 mM D-(+)-glucose, and 1.8 mM CaCl₂. Final concentrations of the excipient, dimethylsulfoxide (Sigma, St Louis, MO), were maintained at 0.1%. Cells were seeded as detailed above and incubated for 20 min with 1 µM Fura-2 acetoxymethyl ester (AM) (Life Technologies, Carlsbad, CA) in growth media supplemented with 3 mM sodium pyruvate (Lonza, Walkersville, MD) to counter potential effects of byproducts of the AM hydrolysis reaction.(Tiffert, et al., 1984) Coverslips were rinsed in ECS supplemented with 55 mM KCl to fill intracellular stores prior to being mounted for imaging on a Leica DMI 6000 fluorescent microscope (Leica Microsystems Inc., Buffalo Grove, IL) with an Orca-R2 C10600 digital camera (Hamamatsu, Bridgewater, NJ) with DG5 fast wavelength switcher (Sutter, Novato, CA) controlled by MetaFluor (Molecular Devices, Sunnyvale, CA) and Fura2 filter set 71000av2 (D340xv2, D380xv2, 400DCLP, D510/40m; Chroma, Bellows Falls, VT) and allowed to equilibrate in ECS for at least 5 minutes prior to imaging. Cells were imaged with a HCX PL APO 40x/1.25–0.75 oil objective (Leica Microsystems, Buffalo Grove, IL). Mounted coverslips were constantly perfused with ECS buffer warmed to 37°C by a temperature controller (TC-344B; Warner Instruments, Hamden, CT) and an inline heater/cooler (SC-20; Warner Instruments, Hamden, CT). Drugs were delivered through a software-controlled perfusion system (VC-8T; Warner Instruments, Hamden, CT) driven by high flow peristaltic pumps (P720/66; Instech, Plymouth Meeting, PA) at speed of 2 ml/min. For RyR-mediated responses, caffeine-treated coverslips were imaged at 1 Hz and were perfused with 50 mM caffeine in ECS solution for 300 s, followed by 120 s in ECS buffer. Cells did not respond to depolarization with 55 mM KCl (data not shown). Maximum release of intracellular Ca²⁺ was obtained from the same cells by perfusion with 2.5 µM ionomycin in ECS for 120 s, 5 min after return of the signal to baseline. For IP₃R-mediated responses, cells were incubated and recorded with 10 µM Bt₃Ins(1,4,5)P₃/AM (AG Scientific, San Diego, CA) for 20 minutes in a total of 300 µl with a heated cover. Images were recorded at 2 Hz, to reduce photobleaching.

For pharmacological experiments, dantrolene (DAN; 20 µM; Enzo, Farmingdale, NY) and Xestospongin D (Xes D; 10 µM; EMD Millipore, Gibbstown, NJ), inhibitors of RyR and IP₃R, respectively, were applied 30 min prior to application of the respective agonists, and the concentration maintained during the experiment.

Regions of Interest (ROIs) were selected based on fluorescent and brightfield images, identifying cell perimeter, cytosol and nucleus. Data was background corrected utilizing the mean fluorescence signal intensity from 5 randomly placed background ROIs in the field of view, normalized for surface area and section volume and exported to Microsoft Excel (Microsoft Corporation, Redmond, WA) for analysis, as described previously (Payne, et al., 2013). Specifically, a single ROI was used for the total cellular response by outlining the cell based on the DIC image. For subcellular analysis, a single ROI was used for the nuclear Ca²⁺ response, whereas 3–5 ROIs were placed in the cytosol. Signal intensity of cytosolic ROIs was averaged to obtain the mean fluorescence intensity value for each cell.

Data Analysis and Statistics

Statistical analysis was conducted using Prism version 5 (Graphpad Software Inc., La Jolla, CA). Experimental groups were compared by Student’s t-test or one-way analysis of variance (ANOVA) with Newman-Keuls post hoc test, as appropriate. Ca²⁺ transients were analyzed with respect to amplitude, the maximal intracellular calcium release from ER stores, and presented as mean F/F₀ max, and as mean area under the curve (AUC), a correlate measure of total calcium release. This experimentally-defined AUC was calculated from the baseline (F₀) to the amplitude (F/F₀ max) plus the next 100 time points (equaling 100s for caffeine experiments) or 250 time points (equaling 500s for IP₃-AM experiments), and is shown as integrated relative fluorescence units (iRFU). Transients were defined by thresholding at 5% above baseline (F/F₀ > 1.05). The 5% threshold was applied after normalization and was used to distinguish artifacts or noise from specific signals in a standardized way. The 5% threshold served as inclusion criterion for cells responding at different intensities, thus permitting differentiation of lower cytosolic responses compared with stronger nuclear responses. Data was compiled from the number of cells (n) derived from a minimum of three to five separate experiments for each experimental condition. When referring to the number of cells responding to pharmacological stimuli, n represents a single experiment, consisting of 3–5 coverslips. For histograms, binned data (in F/F₀ max = 0.05 bins, based upon the applied 5% threshold) was fitted with a non-linear regression for Gaussian distribution (1,000 iterations) and data contingency tables were compared with the Pearson’s chi squared (χ²) test, as described by us previously.(Payne, et al., 2013)

Results

ONHAs exhibit a differential subcellular expression of intracellular Ca²⁺ channels

Ex vivo ONHAs showed strong immunoreactivity for glial fibrillary acidic protein (GFAP), identifying them positively as astrocytes (Fig. 1). Furthermore, ONHAs showed distinct plasma membrane labeling for the glutamate aspartate transporter (GLAST; data not shown). Cells were of a single morphology typical for CNS astrocytes, and no other cell types were observed. ONHAs were difficult to detect by DIC using confocal microscopy due to their thinness (Fig. 1). Backscatter images, obtained using 647 nm wavelength laser line, highlight adhesion and morphology of ONHAs.

Figure 1. Cultured ONHAs are positive for GFAP.

Representative images for primary cultured rat ONHAs (passage 3) stained for GFAP, colabeled with AlexaFluor^® 647 Phalloidin and Hoechst 33258. Backscatter image generated with 647 nm laser illumination shows morphology and adhesion of ONHA. A differential interference contrast (DIC) image is shown for comparison. Scale bar: 25 µm.

We determined the expression and distribution of all three known subtypes of IP₃Rs, type 1 IP₃R (IP₃R1), type 2 IP₃Rs (IP₃R2) and type 3 IP₃Rs (IP₃R3). We identified significant immunoreactivity for IP₃R1 and IP₃R2 in primary cultured rat ONHAs using immunocytochemistry (Fig. 2). Immunoreactivity was mostly cytosolic for IP₃R1s (Fig. 2A), characteristic of expression on the ER (Koulen, et al., 2000), whereas IP₃R2 showed strong nuclear localization, characteristic of expression on the nuclear envelope (Fig. 2B). We did not detect IP₃R3 immunoreactivity in primary cultured rat ONHAs.

Figure 2. Immunocytochemical localization of IP₃Rs in primary cultured rat ONHAs.

Representative images are shown for staining against IP₃R1 (A), IP₃Rs (B), and IP₃R3 (C), using AlexaFluor^® 488-labeled secondary antibody. Cells were co-stained using AlexaFluor^® 647 Phalloidin, labeling actin filaments, and with the nuclear label Hoechst 33258. Note the punctate, cytosolic immunoreactivity indicative of ER localization for IP₃R1, compared to the strong nuclear immunoreactivity of IP₃R2. We did not detect significant IP₃R3 immunoreactivity in primary cultured rat ONHAs. Scale bars are 25 µm.

Similarly, we identified strong immunoreactivity for all subtypes of RyR in primary cultured rat ONHAs (Fig. 3). All RyR subtypes showed punctate immunoreactivity reminiscent of cytosolic and ER localization. Co-labeling with the nuclear dye Hoechst 33258 did not reveal any differentially distributed elevated localization to the nuclear envelope as seen for IP₃R2s (Fig. 2B).

Figure 3. Immunocytochemical localization of RyRs in primary cultured rat ONHAs.

Representative images are shown for staining against RyR1 (A), RyR2 (B), and RyR3 (C), using AlexaFluor^® 488-labeled secondary antibody. As for IP₃R staining in Fig.1, cells were co-stained using AlexaFluor^® 647 Phalloidin and Hoechst 33258. Note the punctate, cytosolic immunoreactivity indicative of ER localization typical for RyRs. We identified all RyR subtypes in primary cultured rat ONHAs, with RyR2 showing the strongest Scale bars are 25 µm.

IP₃Rs and RyRs contribute functionally to intracellular Ca²⁺ release from stores

In order to identify the potential functional relevance of such a differential distribution of intracellular Ca²⁺ channels we next used receptor-specific pharmacology to elicit Ca²⁺ release from intracellular stores and performed quantitative imaging of intracellular Ca²⁺ concentration transients.

In order to determine the contribution of IP₃Rs, we measured the rise in intracellular Ca²⁺ concentration upon stimulation with the IP₃R-selective agonist, IP₃ (10 µM IP₃-AM). Representative raw cellular responses (excitation: 380 nm, not corrected for background) presented as heat maps (Fig. 4A) are shown alongside the matching background corrected and normalized single cell traces (F/F₀ max; Fig. 4B). Stimulation resulted in a 19.0 ± 1.3% increase of normalized fluorescence intensity over baseline (F/F₀ max; n=47; Fig. 4C) and resulted in an experimentally determined AUC for IP₃ stimulation of 168 ± 18 iRFU (n=45; Fig. 4D). Using a threshold of 5%, we calculated a responder rate of 81 ± 13% (Fig. 4E). To test the specificity of the response, we pre-incubated cells with the IP₃R blocker, Xes D, for 30 min. Xes D reduced both AUC (n=18; P<0.001; Fig. 4A) and F/F₀ max (n=18; P<0.001; Fig. 4B). The number responding cells was reduced significantly by 68% (P<0.05; Fig. 4C) by the presence of Xes D.

Figure 4. Functional contributions of IP₃Rs to intracellular Ca²⁺ release in primary cultured rat ONHAs.

In order to determine the contribution of IP₃Rs to Ca²⁺ release from intracellular stores, we quantified the rise in intracellular Ca²⁺ concentration upon stimulation with the IP₃R-selective agonist, IP₃, using optical imaging and the Ca²⁺ fluorophore, Fura-2. For delivery, we used the membrane-permeable molecules IP₃-AM (10 µM) and Fura-2-AM (1 µM). Representative cellular responses (A) and traces (B) are shown. Warmer colors indicate higher fluorescence intensity. IP₃ stimulation resulted in a mean 19.0 ± 1.3% increase of normalized fluorescence intensity over baseline (F/F₀ max; n=47; C). This corresponded to an experimentally determined AUC of 168 ± 18 iRFU (n=45; D). In our experiments, 81 ± 13% of cells responded to IP₃ stimulation (n=3 separate experiments, with at least 15 cells per experiment; E). In order to determine the specificity of the response, we pre-incubated cell with the IP₃R blocker, Xestospongin D (Xes D), for 30 min, and experiments performed with Xes D co-application. Xes D resulted in a 15.4 ± 2.5% reduction of F/F₀ max (n=18, P<0.001; C), and reduced AUC significantly by 145 ± 30 iRFU (n=18, P<0.001; D). The number of cells responding to IP₃ in the presence of Xes D was reduced significantly by 68% (P<0.05; E). * P<0.05; *** P<0.001. Scale bar: 10 µm.

In a separate set of experiments, we measured RyR-mediated Ca²⁺ release using pharmacological stimulation of the receptor (50 mM caffeine). Representative cellular responses, obtained from uncorrected raw images (excitation: 380 nm) presented as heat maps are shown in Fig. 5A. The matching traces are plotted in Fig. 5B. Mean fluorescence intensity increased by 13.0 ± 0.8% over baseline (F/F₀ max; n=38; Fig. 5C) following caffeine stimulation. The experimentally determined AUC was 19.5 ± 1.5 iRFU (n=38; Fig. 5D). The number of responding cells was again calculated applying a 5% threshold and was 56 ± 12% (Fig. 5E). Following a 30 min pre-incubation with the RyR inhibitor, DAN (20 µM), for 30 min, peak fluorescence intensity (F/F₀ max; n=49, P<0.001; Fig. 5C) and AUC (n=49, P<0.001; Fig. 5D) were statistically significantly reduced. Similarly, the number of responding cells was significantly lower during DAN application (P<0.05; Fig. 5C). Representative traces are shown (Fig. 5D). The matching cellular responses are obtained from uncorrected raw images (excitation: 380 nm) presented as heat maps (Fig. 5E).

Figure 5. Functional contributions of RyRs to intracellular Ca²⁺ release in primary cultured rat ONHAs.

Caffeine (50 mM) was used as RyR agonist to quantify the contributions of RyR to stimulus-induced intracellular Ca²⁺ release. Representative cellular responses (A) and traces (B) are shown. Warmer colors indicate higher fluorescence intensity. We observed an average 13.0 ± 0.8% increase of normalized fluorescence intensity over baseline (F/F₀ max; n=38; C). The experimentally determined AUC for RyR stimulation was 19.5 ± 1.5 iRFU (n=38; D). In our experiments, 56 ± 12% of cells responded to caffeine stimulation (n=3 separate experiments, with at least 15 cells per experiment; E). In order to determine the specificity of the response, we pre-incubated cells with the RyR inhibitor, dantrolene sodium (DAN; 20 µM), for 30 min, and experiments performed with DAN co-application. DAN reduced F/F₀ max (3.0 ± 0.7%, n=49, P<0.001; C) and AUC significantly by 14.6 ± 2.0 iRFU (n=49, P<0.001; D). The number of cells responding to caffeine in the presence of DAN differed significantly by 35 ± 15% (P<0.05; E). * P<0.05; *** P<0.001. Scale bar (in A): 10 µm.

Evidence for independent regulation of nuclear and cytosolic Ca²⁺ signaling in ONHAs

In order to account for a preferential distribution of ICCs on the membranes of the ER or the membranes of the nuclear envelope, an ER sub-compartment, we performed a subcellular analysis of our Ca²⁺ imaging data. Caffeine-induced, i.e. RyR-mediated Ca²⁺ release did not differ significantly between nuclear and cytosolic areas, as quantified by AUC (P=0.13; Fig. 6A) and the maximum fluorescence intensity F/F₀ max (P=0.67; Fig. 6B). DAN blocked caffeine-mediated Ca²⁺ release similarly in the nucleus and the cytosol (Fig. 6A, B). Nuclear and cytosolic amplitudes of RyR-mediated Ca²⁺ transients binned at 5% intervals (Fig. 6C) were found to be Gaussian normally distributed (R²=0.45 and R²=0.71, respectively) and not to be statistically significantly different (Pearson’s χ² test, P=0.41). Time to threshold was lower for nuclear responses, compared with cytosolic responses of caffeine-stimulated intracellular Ca²⁺ release (n=32, P<0.001; Fig. 7A). In contrast, time to peak (F/F0 max) was similar between nuclear and cytosolic responses (n=32, P=0.81; Fig. 7B).

Figure 6. Subcellular localization of stimulus-induced intracellular Ca²⁺ release.

Given the differential subcellular localization of ICCs on the membranes of ER or the membranes of the nuclear envelope, we performed a subcellular analysis of our Ca²⁺ imaging data. No statistically significant difference was identified between nuclear and cytosolic Ca²⁺ release upon caffeine stimulation, neither at the level of AUC (A) nor F/F₀ max (B). DAN blocked caffeine-mediated Ca²⁺ release similarly in both regions (A, B). Amplitudes of the RyR-mediated Ca²⁺ transients binned at 5% intervals are Gaussian normally distributed, and the distributions for not statistically significantly different between cytosol and nucleus (P=0.41; C). In contrast, AUC of IP₃R-mediated Ca²⁺ release was 39% higher in the nucleus compared with the cytosol (n=42, P<0.05; D). Similarly, F/F₀ max was 28% higher in the nucleus vs. the cytosol (n=42, P<0.05; E). Both, binned nuclear and cytosolic F/F₀ max amplitudes could be fitted with a Gaussian distribution and were highly significantly different (P<0.001; F).

Figure 7. Kinetic analysis of IP₃- and caffeine-stimulated intracellular Ca²⁺ responses.

We measured the kinetic responses of the increase in intracellular Ca²⁺ concentration, following exposure to the agonists, IP₃-AM and caffeine. (A) Time to reach the 5% threshold was significantly faster for nuclear responses, compared with cytosolic responses (228.1 ± 6.4 sec vs. 262.5 ± 10.0 sec, n=32, P<0.01). (B) Similarly, time to peak (F/F₀ max) was faster for nuclear vs. cytosolic responses (343.3 ± 13.7 sec vs. 361.6 ± 14.7 sec, n=32, P<0.001), indicative of significant nuclear localization of RyR2 receptors. (C) IP₃-AM stimulation resulted in a faster nuclear response to threshold, compared with cytosolic kinetics (357.1 ± 41.0 sec vs. 398.4 ± 43.5, n=46, P<0.05), supporting evidence for significant expression of nuclear IP₃R2 receptors in the membranes of the nuclear envelope. (D) In contrast, time to peak (F/F₀ max) was similar between nuclear and cytosolic responses (1063 ± 72 sec. vs. 1023 ± 68 sec, n=46, P=0.69).

In contrast, the AUC of IP₃R-mediated Ca²⁺ release was 39% higher in the nucleus vs. the cytosol (n=42, P<0.05; Fig. 6D). Similarly, F/F₀ max was 28% higher in the nucleus compared with the cytosol (n=42, P<0.05; Fig. 6E). Both, nuclear and cytosolic F/F₀ max amplitudes, binned at 5% intervals, could be fitted with a Gaussian normal distribution (R²=0.71 and R²=0.82, respectively) and were highly significantly different when comparing responses using the Pearson’s χ² test (P<0.001; Fig. 6F). Time to threshold was significantly faster for nuclear responses (n=42, P<0.05, Fig. 7C), while time to peak was similar between nucleus and cytosol (n=42, P=0.678, Fig. 7D).

Discussion

We herein describe intracellular Ca²⁺ signaling pathways in primary cultured adult rat ONHAs. To our knowledge, this is the first report of a quantitative analysis of Ca²⁺ signaling with a focus on intracellular Ca²⁺ channels in this highly clinically relevant cell type.

All antibodies used for immunocytochemistry have been previously validated and their specificity has been shown in a number of different cells and tissues. Furthermore, we performed appropriate controls to test for specificity, including omission of the primary antibody and pre-incubating the primary antibody with 100-fold excess of the peptide used to generate the antibodies (data not shown). We used the ratiometric fluorophore, Fura-2, to measure the stimulus-induced rise in the intracellular Ca²⁺ concentration. Fura-2 has been well established for live-cell imaging and the estimated K_D for Ca²⁺ is approximately 224 nM in the cytosol of living cells (Grynkiewicz, et al., 1985, Haworth and Redon, 1998). Furthermore, due to ratiometric sampling, Fura-2 provides for experimental control over dye loading concentration, loading efficiency, illumination intensity, and photobleaching, in addition to having a very low affinity for divalent cations other than Ca²⁺ (Grynkiewicz, et al., 1985).

In order to measure IP₃R- and RyR-mediated intracellular Ca²⁺ release, we used the specific agonists IP₃-AM (10 µM) and caffeine (50 mM), respectively. These concentrations are within the physiological or pharmacologically relevant range of concentrations to stimulate IP₃R and RyR activity, respectively (Berridge, 1998, Leite, et al., 2003). Furthermore, we have shown specificity of the responses by using the IP₃R and RyR blockers, Xes D and DAN, respectively (Stutzmann, 2005, Zalk, et al., 2007). Currently, there is no selective polycystin-2 inhibitor available, making it thus impossible to test any non-specific contribution of polycystin-2 to IP₃-AM- or caffeine-induced Ca²⁺ responses pharmacologically. However, polycystin-2 channels require prolonged depolarization and Ca²⁺ induced activation to mediate Ca²⁺ release (Koulen, et al., 2002), making their involvement unlikely.

Using immunocytochemistry, we detected IP₃R1-specific immunoreactivity in primary cultured rat ONHAs. The diffuse punctate staining pattern is characteristic of cytosolic and ER localization, as seen for IP₃R1 in other cell types (Kaja, et al., 2011, Medina-Ortiz, et al., 2007). In contrast, IP₃R2 immunoreactivity was very strong in the nuclear envelope, consistent with reports for other cell types (Kaja, et al., 2011, Koulen, et al., 2005, Leite, et al., 2003, Medina-Ortiz, et al., 2007). This is the first description of such a molecular substrate of differential subcellular Ca²⁺ signaling in astrocytes. Imaging of the intracellular Ca²⁺ concentration allowed us to quantify the functional contribution of IP₃Rs to Ca²⁺ release. The maximum Ca²⁺ release for representative Ca²⁺ transients was calculated to be 654 nM, using the method originally described by Grynkiewicz and colleagues (Grynkiewicz, et al., 1985). Besides IP₃Rs, we identified specific immunoreactivity for all subtypes of RyRs in cultured ONHAs. RyR2 immunoreactivity was reminiscent of nuclear envelope and Golgi staining, while RyR1 and RyR3 immunoreactivity was largely localized to the ER, and to a lesser extent to the nuclear envelope. Overall, the maximum Ca²⁺ release following caffeine stimulation was 566 nM, calculated using the same method (Grynkiewicz, et al., 1985).

We conclude that the major driving force of intracellular Ca²⁺ release are IP₃Rs, specifically IP₃R2 and to a lesser extent IP₃R1, while RyR contribute to Ca²⁺ signaling through Ca²⁺-induced Ca²⁺ release, leading to the amplification of intracellular Ca²⁺ signals generated by IP₃Rs or plasma membrane driven influx of Ca²⁺ (Berridge, 1998).

While the membranes of the ER and the nucleus are contiguous, the nuclear envelope comprises IP₃-sensitive stores separate from those of the ER (Stehno-Bittel, et al., 1995). In neuronal cells, IP₃Rs show a distinct expression profile with type 1 and type 3 expressed in the ER, while type 2 IP₃Rs preferentially localize in the membranes of the nuclear envelope (Kaja, et al., 2011, Koulen, et al., 2005, Leite, et al., 2003). Our detailed subcellular quantification of Ca²⁺ transients, which showed significantly larger nuclear vs. cytosolic/ER transients, consistent with stronger IP₃R2 expression in the nucleus. The higher levels of Ca²⁺ release were evident both at the level of the maximum amplitude and the integrated density of the Ca²⁺ transient. Furthermore, time to threshold was faster for nuclear vs. cytosolic transients. In this context it is important to note that IP₃R2 possesses the highest ligand affinity of all IP₃R subtypes, activating at concentrations well below those needed for IP₃R1 and IP₃R3 (Koulen and Thrower, 2001, Koulen, et al., 2005, Leite, et al., 2003) and while IP₃R1 shows a pronounced inhibition by higher Ca²⁺ concentrations, IP₃R2 do not inactivate as readily under conditions of prolonged high cytosolic Ca²⁺ concentrations (Koulen and Thrower, 2001). We have previously reported the selective up-regulation of IP₃R2 under conditions of oxidative stress in neuronal HT-22 cells (Kaja, et al., 2011) indicating that the high physiological relevance of IP₃R2 for normal ONHA Ca²⁺ signaling could become even more pronounced under disease conditions associated with oxidative stress. Of particular relevance, aberrant Ca²⁺ signaling has been described both in preclinical glaucoma models and glaucoma patients (reviewed in Crish and Calkins, 2011) and is considered a viable target for neuroprotection (Duncan, et al., 2010, Payne, et al., 2014, Payne, et al., 2013).

The subcellular quantification of RyR-mediated Ca²⁺ release did not reveal statistically significant differences between cytosolic regions and the nuclear envelope, i.e. there is a significant number of RyRs in the nuclear envelope that mediate Ca²⁺ release from dedicated nuclear stores. Nuclear RyR-elicited transients reached threshold and peak amplitude faster than cytosolic ones. To our knowledge this is the first report of differential nuclear vs. cytosolic RyR-mediated Ca²⁺ transients in primary cells. Based on the specific distribution of intracellular Ca²⁺ channels, and the subcellular quantification of Ca²⁺ transients, we conclude that IP₃R2 signaling is linked to RyR2 mediated Ca²⁺ release from the nuclear envelope, while all RyR subtypes amplify IP₃R1-specific Ca²⁺ signals from the ER (Fig. 8).

Figure 8. Schematic representation of intracellular Ca²⁺ signaling in primary cultured rat ONHAs – differential distribution of intracellular Ca²⁺ channels provides the molecular substrate for distinct and temporally and spatially independent subcellular Ca²⁺ signaling patterns.

IP₃R2 is the major subtype of IP₃Rs in ONHAs, expressed differentially in the membranes of the nuclear envelope, which are contiguous with the membranes of the ER. Ca²⁺ release from IP₃R2 in the nuclear envelope into the nucleoplasm is amplified through RyR2 by Ca²⁺-induced Ca²⁺ release (CICR; black dotted arrows), leading to Ca²⁺-mediated regulation of gene expression. All types of RyR contribute to Ca²⁺ release into the ER via amplification of IP₃R1-mediated Ca²⁺ release potentially mediating non-genomic functions of Ca²⁺signaling in ONHAs.

Taken together, IP₃R2 is the major subtype of IP₃Rs in ONHAs; Ca²⁺ release from the nuclear envelope is likely amplified through RyR2. This differential pattern of Ca²⁺ signaling in ONHAs therefore potentially regulates gene expression (Hwang, et al., 2009, Koulen, et al., 2008) (Fig. 8) to adapt to changes in the astrocytic milieu and synaptic signaling, in accordance with the generally accepted view of astrocytes contributing critically to synaptic signaling and homeostasis (Dallerac, et al., 2013).

Our current knowledge regarding signaling pathways in ONHAs upstream and downstream of intracellular Ca²⁺ release is very limited. However, intracellular Ca²⁺ release through endothelin A receptor activation (Murphy, et al., 2011) and G_q-mediated IP₃ generation leading to activation of IP₃Rs (Biber, et al., 1997) has been documented in primary cultures of rat ONHAs. Our results indicate that the molecular and functional framework for these phenomena is composed of a complex differentially distributed set of intracellular Ca²⁺ release channels, but more importantly have also direct mechanistic relevance to help explain reported changes in Ca²⁺ signaling in response to stress and/or injury. For instance, hydrostatic pressure-induced intracellular Ca²⁺ release has been found to occur via RyRs, resulting in extracellular signal-regulated kinase (Erk) phosphorylation (Mandal, et al., 2010), and to involve signaling pathways that control intracellular Ca²⁺ signaling effector proteins in assays measuring the migratory ability of ONHAs and potentially underlying ONH remodeling in glaucoma (Tezel, et al., 2001). In similar studies of human ONHAs in vitro, a molecular substrate responsible for differential Ca²⁺ signaling was inferred based on phenotypic studies and pharmacological activation of purinergic receptors (Chen, et al., 2009). Similarly, partial transection of the ON, a widely used model for ocular drug discovery, resulted in the dysregulation of Ca²⁺ homeostasis, increased levels of oxidative stress and increased immunoreactivity of both the GluR1 subunit of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor and aquaporin 4 (Wells, et al., 2012).

Conclusion

The data presented herein is the first report of quantification of intracellular Ca²⁺ signaling and its molecular correlates in ONHAs. ONHAs utilize a set of differentially distributed intracellular Ca²⁺ channels to control intracellular homeostasis. Our data provides a foundation for future studies investigating potential changes in Ca²⁺ signaling in ONHAs in glaucomatous retinopathy and other disorders affecting the optic nerve and optic nerve head.

Highlights.

• First description of differential IP₃R-mediated Ca²⁺ signaling in astrocytes.

• IP₃Rs are the major driving force of intracellular Ca²⁺ release in ONHAs.

• Differentially-distributed RyRs in ONHAs contribute to Ca²⁺ signaling via CICR.

• ONHAs have larger nuclear vs. cytosolic/ER Ca²⁺ transients.

• Ca²⁺ signaling in ONHAs is a druggable target for glaucoma and related disorders.

Abbreviations

AM

acetoxymethyl ester

DAN

dantrolene

ECS

extracellular solution

IOP

intraocular pressure

IP₃R

inositol-1,4,5,-trisphosphate receptor

ON

optic nerve

ONHA

optic nerve head astrocyte

Ph

Phalloidin

RGC

retinal ganglion cell

RyR

ryanodine receptor

Xes D

Xestospongin D