The IG-8 line of heterozygous transgenic mice (C57BL/6N), which express green fluorescent protein (GFP) signals in SACs in the retina under the control of the metabotropic glutamate receptor 2 promoter, was previously reported for this transgenic line (Watanabe et al., 1998; Yoshida et al., 2001). Transgenic mice were backcrossed to C57BL/6J mice. The mice of both sexes were used at postnatal day 3 (P3), P9, P15, and P28. Mice were housed with a 12:12-h light:dark cycle in a temperature-controlled room. Animals were euthanized by neck dislocation 2 h after lights on. Fresh water and a rodent diet were supplied ad libitum. When the dark rearing of pups was necessary, pregnant mice were reared in the dark before the birth of pups. Dark reared mice were euthanized with the same time schedule for mice housed with a 12:12-h light:dark cycle. The eyes were enucleated and hemisected, and the retinas were isolated from the sclera in Ringer’s solution (in mM: 115 NaCl, 5 KCl, 26 NaHCO₃, 2 CaCl₂, 1 MgCl₂, 1.1 NaH₂PO₄, and 20 D-glucose, pH 7.4, bubbled with 95% O₂ and 5% CO₂). Isolated retinas were used for whole-mount, slice, or dissociated preparations.

Whole-mount retinas were incubated in Ringer’s solution at room temperature for > 30 min before patch-clamp recordings or dye injections. Details on the methods used for retinal slice preparation were previously described (Kaneda et al., 2008; Ishii and Kaneda, 2014; Ishii et al., 2017). In brief, isolated retinas were placed on a membrane filter (pore size, 0.45 μm; Advantec Toyo, Tokyo, Japan) with the retinal ganglion cell (RGC) side down and sliced at a thickness of 150 μm in Ringer’s solution. All experimental procedures were conducted at room temperature.

Whole-cell recordings were made from ON- or OFF-type SACs in all retinal quadrants identified by a GFP fluorescent signal when viewed under a fluorescent microscope (BX51WI; Olympus, Tokyo, Japan). Retinas were superfused at a rate of 2 ml/min with Ringer’s solution. Data recordings were conducted in Ringer’s solution containing 1 μM SR95531 (Sigma-Aldrich, Taufkirchen, Germany) to block IPSCs at room temperature. Other drugs were also applied to the bath holding the retina. Meclofenamic acid (MFA) was purchased from Sigma-Aldrich and SKF38393 and SCH23390 from Nacalai Tesque, Inc (Kyoto, Japan). Patch pipettes with a resistance of 7–9 MΩ (when filled with an intracellular solution) were fabricated from borosilicate glass. The composition of the intracellular solution was as follows (in mM): 120 CsCl, 5 EGTA, 0.5 CaCl₂, 10 HEPES, 5 ATP-2Na, and 1 GTP-3Na (pH adjusted to 7.2 with CsOH). Recordings were made with a patch clamp amplifier (Axopatch-200B; Molecular Devices, San Jose, CA, USA) connected to a Digidata 1322A interface and pCLAMP 10.3 software (Molecular Devices). Data were sampled at 10 kHz after passing a low-pass filter at 5 kHz. All recordings were started at least 5 min after achieving the whole-cell recording configuration.

The Cm of SACs was calculated according to a previous study (Lindau and Neher, 1988). The whole-cell current evoked by a step pulse (steps of −2 mV with a duration of 10 msec) at a holding potential of −70 mV was averaged (5–10 sweeps). Averaged capacitive currents were used to calculate Cm (Figure 1A, blue area). The effects of drugs (MFA, SKF38393, and SCH23390) on Cm were evaluated by comparing Cm recorded during and before (control) drug application. Cm for the control was calculated by averaging Cm during the last 5 min within the control, and Cm during drug application was calculated from the peak value for Cm during drug application.

Whole-mount retinas were used for the dye injection. Since the cell bodies of OFF-type SACs were located in the inner nuclear layer in the whole-mount preparation, it was difficult to inject Neurobiotin into OFF-type SACs without damaging other cells. Therefore, experiments were only performed on ON-type SACs. Under the whole-cell clamp mode, ON-type SACs were held at −60 mV for 20 min. Electrodes were filled with Neurobiotin™ (Vector Laboratories, Burlingame, CA, USA, SP-1120) at ∼2% wt/vol and ∼290 mOsm in cesium chloride internal solution. Electrode resistance was 10∼12 MΩ. After tracer filling, the retina was fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB) for 20 min. The retina was then incubated in 0.1 M PB containing 0.3% Triton-X 100 for 15 min. After the incubation in 0.1 M PB containing 1% Block Ace (Dainippon Pharmaceutical Co., Ltd., Osaka, Japan) for 1 h, samples were reacted with the primary antibodies, rabbit anti-RNA-binding protein with multiple splicing (RBPMS), a marker of RGCs (Kwong et al., 2010; Rodriguez et al., 2014) (working dilution, 1:500) (Abcam, Cambridge, MA, USA, ab194213), and rat anti-GFP (working dilution, 1:500) (Nacalai Tesque, 04404-84) in 0.1 M PB containing 0.4% Block Ace at 4°C for 48 h. Samples were then allowed to react with streptavidin-conjugated Alexa Fluor 594 (working dilution, 1:500) (Invitrogen, Carlsbad, CA, USA, S11227) and the secondary antibodies, Alexa Fluor 488-conjugated donkey anti-rat IgG (working dilution, 1:1,000) (Life Technologies, Carlsbad, CA, USA, A11006) and either Alexa Fluor 647-conjugated donkey anti-rabbit IgG (working dilution, 1:500) (Life Technologies, A31573) or Alexa Fluor Plus 405-conjugated goat anti-rabbit IgG (working dilution, 1:1,000) (Thermo Fisher Scientific, Waltham, MA, USA, A48254) in 0.1 M PB at room temperature for 2 h. Nuclear staining was performed with 4′,6-diamidino-2-phenylindole (DAPI) (Nacalai Tesque, 11034-56). The retina was mounted in Fluoro-KEEPER Antifade Reagent (Nacalai Tesque, 12593-64) for observations. Fluorescent images were captured using confocal microscopes (FV1200; Olympus, Tokyo, Japan, and LSM-980; Carl Zeiss, Jena, Germany).

Dissociated cells were prepared as previously described (Ishii et al., 2017). Briefly, the isolated retina obtained from P9 or P28 mice was incubated for 30–40 min in an external solution (which comprised, in mM, 135 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 glucose, and 5 HEPES; pH adjusted to 7.4) containing 2.5 U/ml papain (Worthington Biochemical, Freehold, NJ, USA) and its activator, L-cysteine (0.1 mg/ml) bubbled with 100% O₂ at 37°C. The enzyme treatment was stopped by washing the retina with the external solution containing 0.1 mg/ml of bovine serum albumin. The retina was then triturated with a Pasteur pipette. Retinal cells were re-suspended at a final concentration of 2–3 × 10⁶ cells/ml in the external solution containing 0.5 μg/ml propidium iodide. GFP-positive cells (1 × 10⁴ cells per sample) were collected by fluorescence-activated cell sorting (FACS) (Aria II, BD Biosciences, Franklin Lakes, NJ, USA) as SACs. Propidium iodide-positive cells were excluded as dead cells. The accuracy of the collection of GFP-positive cells was visually inspected by fluorescent microscopy. Scatter data were displayed by FlowJo_v10.7.2 (BD Biosciences). Collected GFP-positive cells were used in the gene expression analysis.

Gene expression analyses were performed as previously described (Ishii et al., 2015), except for the methods of extracting RNA and reverse transcription. Total RNA was extracted using ISOSPIN Cell & Tissue RNA (Nippon Gene, Japan) from GFP-positive cells. Reverse transcription to obtain complementary DNA (cDNA) was conducted using ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Japan). The resulting cDNA was amplified with gene-specific primers and SYBR Premix Ex Taq II (TaKaRa, Japan). A real-time polymerase chain reaction analysis was performed using Thermal Cycler Dice Real-time System Single (TaKaRa). Reactions were performed at 95°C for 30 s, followed by 40 cycles at 95°C for 5 s and 60°C for 30 s. All processes were conducted according to the manufacturer’s instructions. Primer sequences, product sizes, and accession numbers are listed in Table 1.

Statistical analyses were performed using Prism 7.0 (GraphPad Software, La Jolla, CA, USA) and R (version 4.1.1) (Ihaka and Gentleman, 1996). Grouped data in Figures 1, 2, 6 were presented in box plots, with the central line showing the median and the lower and upper edges of the box indicating 25 and 75% of data, respectively. We used a parametric or non-parametric test depending on the results of normality and homoscedasticity tests. If any dataset did not exhibit a normal distribution or homoscedasticity, non-parametric tests were performed. Kruskal–Wallis tests were used for multiple comparisons with the Steel-Dwass test to examine the significance of differences (see Figure 1). The Mann–Whitney test was used when two unpaired groups were compared (see Figures 2, 4, 5, 6). Comparisons between recordings from the same cell before and after the application of pharmacological agents were performed using paired t-tests (see Figures 3, 5). Details on the results of individual statistical analyses are described in the figure legends.

In ON- and OFF-type SACs, a change was observed in Cm during development. In ON-type SACs, the average Cm recorded in the slice preparation was 16.6 pF at P3 (Figure 1B). Average Cm peaked at P9 and then decreased to 3.4 pF by P28. Changes in Cm in OFF-type SACs were similar to those in ON-type SACs (Figure 1C). In biological membranes, the Cm of the unit membrane area was 1 μF/cm² and was independent of the cell type (Hille, 1992). Under this condition, when we assume that SAC is a spherical cell with a soma diameter of 10 μm (Taylor and Smith, 2012), the estimated Cm of single SAC is ∼3 pF. The decrease observed in Cm after eye-opening was in contrast to previous findings showing that the soma size of SACs did not significantly differ during postnatal development (at P5, P10, P15, P20, P25, and P30) (Zhang et al., 2005) and also that an elongation of the dendrite length of SACs, indicating an increase in biological membranes, occurred throughout development (Wong and Collin, 1989). Therefore, large Cm at P3 and P9 implied that SACs were electrically coupled with other cells via gap junctions in the early postnatal stage.

To examine whether gap junctional coupling formed between SACs and other cells in the early postnatal stage, the Neurobiotin tracer was injected into ON-type SACs (white arrow, Figures 2A, E, F, H–J) to visualize cells that coupled with ON-type SACs. Many dye-coupled cells were detected at P9 (Figures 2A, H). Some dye-coupled cells showed strong fluorescence (white arrowheads), while others showed weak fluorescence (gray arrowheads). In the present study, dye-coupled cells with weak fluorescence were excluded from analyses because they may have included second-order connections. Dye-coupled cells were radially distributed around dye-injected SACs (16.5 ± 4.5 cells, mean ± SD). On the other hand, only a few dye-coupled cells with small soma were observed at P28 (1.2 ± 1.4 cells, mean ± SD) (Figures 2A, B). The colocalization of DAPI, a nuclear marker, was observed in dye-coupled cells, indicating that Neurobiotin signals did not reflect the varicosities of SAC dendrites. The number of dye-coupled cells was significantly lower at P28 than at P9 (Figure 2D). At P9, dye-coupled cells did not overlap with GFP, a marker of SACs (Figures 2E, I, K), whereas the majority of dye-coupled cells showed immunoreactivity for RBPMS, a marker of RGCs (Figures 2J, K, white arrowheads). At P28, dye-coupled cells did not overlap with GFP. However, further identification was not performed because of the limited number of dye-coupled cells. These results suggest that large Cm in SACs was due to the formation of electrical coupling via gap junctions and that dye-coupled cells in ON-type SACs were heterologous, not homologous cells.

We examined the effects of MFA, a gap junction inhibitor, on Cm. In ON-type SACs, the application of 300 μM MFA to a whole-mount preparation significantly reduced Cm at P9 (Figures 3A, B). The inhibitory effects of MFA on Cm peaked within 2 min and persisted during the application of MFA. After the washout of MFA, Cm recovered to the control level. The inhibitory effects of MFA on Cm were also observed at P3, P15, and P28 (Figure 3B). In OFF-type SACs, we investigated the effects of MFA on Cm at P3 only because OFF-type SACs in the slice preparation at P9 deteriorated and did not recover after the application of MFA. In OFF-type SACs at P3, we observed a significant decrease in Cm after the application of MFA (Figure 3B).

The retina uses multiple types of connexins to form various types of gap junctions (Guldenagel et al., 2000; Volgyi et al., 2013). The expression of multiple types of connexin genes has also been reported in SACs (Yan et al., 2020). Therefore, we investigated whether a postnatal change occurred in the expression levels of connexin genes [connexin 23 (Cx23), 36 (Cx36), 43 (Cx43), 45 (Cx45), and 50 (Cx50)] (Table 1) in SACs. A fraction of SACs was separated as GFP-positive cells by FACS (Figure 4A, dotted green rectangle). The Ct value of EGFP mRNA, the internal standard in the present study, did not significantly differ between P9 and P28 (Figure 4B). All connexins examined were detected at P9. However, Cx50 was detected in 2 out of 8 samples. The relative expression of Cx23 and Cx43 was significantly lower at P28 than at P9, whereas that of Cx36 and Cx45 did not significantly differ between P9 and P28 (Figure 4C).

In the vertebrate retina, dopamine is synthesized and released from dopaminergic amacrine cells (Roy and Field, 2019). Dopamine release is detected before eye-opening and increases after eye-opening (Melamed et al., 1986; Casini et al., 2004). The dopamine D1 receptor (D1R) regulates gap junctional coupling between retinal neurons (Bloomfield and Volgyi, 2009). Moreover, SACs have been reported to express Drd1 encoding D1R at P19 (Yan et al., 2020). In our experiments, GFP-positive cells also expressed Drd1 at P9 and P28 (Figure 5A). Therefore, we investigated whether dopamine regulated gap junctional coupling in ON-type SACs using a whole-mount preparation. The application of 10 μM SKF38393, a D1R agonist, slightly reduced Cm at P9, but did not exert any effects on Cm at P28 (Figures 5B, C). On the other hand, the application of 10 μM SCH23390, a D1R antagonist, did not increase Cm at P9 or P28 (Figures 5D, E). The present results suggest that gap junctional coupling by SACs in the early postnatal stage was regulated by D1R, similar to other gap junctional coupling types in the retina (Hampson et al., 1992; Yadav et al., 2019; Banerjee et al., 2020).

Visual experience has been reported to play an important role in functional and morphological maturation in the retina (Sernagor and Grzywacz, 1996; Zhang et al., 2005; Tian, 2008). Since mice open their eyes by P12–14 (Ford and Feller, 2012; Hoy and Niell, 2015), we investigated whether the deprivation of visual experience retarded the maturation process of gap junctional coupling by SACs. We compared the Cm of ON-type SACs between normal light/dark cycle-reared and dark-reared mice at P9 (before eye-opening) and P15 (after eye-opening) using a whole-mount preparation. No significant differences were observed in Cm between normal light/dark cycle- and dark-reared mice at both ages (Figure 6), whereas variance at P15 was larger in dark-reared mice (112.8) than in normal light/dark cycle-reared mice (13.7) (p < 0.01, the F-test). This result suggests that the deprivation of visual experience did not retard the maturation process of gap junctional coupling by SACs.