Skip to main content
NCBI home page
Journal of Neurophysiology logoLink to Journal of Neurophysiology
. 2018 Aug 8;120(4):2121–2129. doi: 10.1152/jn.00322.2018

Bistratified starburst amacrine cells in Sox2 conditional knockout mouse retina display ON and OFF responses

Todd L Stincic 1, Patrick W Keeley 2, Benjamin E Reese 2,3, W Rowland Taylor 1,4,✉
PMCID: PMC6230791  PMID: 30089022

Abstract

Cell-intrinsic factors, in conjunction with environmental signals, guide migration, differentiation, and connectivity during early development of neuronal circuits. Within the retina, inhibitory starburst amacrine cells (SBACs) comprise ON types with somas in the ganglion cell layer (GCL) and dendrites stratifying narrowly in the inner half of the inner plexiform layer (IPL) and OFF types with somas in the inner nuclear layer (INL) and dendrites stratifying narrowly in the outer half of the IPL. The transcription factor Sox2 is crucial to this subtype specification. Without Sox2, many ON-type SBACs destined for the GCL settle in the INL while many that reach the GCL develop bistratified dendritic arbors. This study asked whether ON-type SBACs in Sox2-conditional knockout retinas exhibit selective connectivity only with ON-type bipolar cells or their bistratified morphology allows them to connect to both ON and OFF bipolar cells. Physiological data demonstrate that these cells receive ON and OFF excitatory inputs, indicating that the ectopically stratified dendrites make functional synapses with bipolar cells. The excitatory inputs were smaller and more transient in Sox2-conditional knockout compared with wild type; however, inhibitory inputs appeared largely unchanged. Thus dendritic stratification, rather than cellular identification, may be the major factor that determines ON vs. OFF connectivity.

NEW & NOTEWORTHY Conditional knockout of the transcription factor Sox2 during early embryogenesis converts a monostratifying starburst amacrine cell into a bistratifying starburst cell. Here we show that these bistratifying starburst amacrine cells form functional synaptic connections with both ON and OFF bipolar cells. This suggests that normal ON vs. OFF starburst connectivity may not require distinct molecular specification. Proximity alone may be sufficient to allow formation of functional synapses.

Keywords: development, electrophysiology, retina, synaptic transmission, transcription factor

INTRODUCTION

The retina provides a convenient model to analyze the role of transcription factors in orchestrating neural determination, which ultimately guides neural migration, differentiation, and connectivity (for review see Reese 2011). Here we focus on starburst amacrine cells (SBACs), a well-defined neural population that is easily recognized as the only cholinergic neurons in the retina (Masland and Mills 1979). Mirror-symmetric ON and OFF populations of SBACs are distinguished by the localization of their somata in the ganglion cell layer (GCL) and the inner nuclear layer (INL), respectively (Hayden et al. 1980; Vaney et al. 1981), and by their sharp dendritic stratification in the ON and OFF strata in the inner plexiform layer (IPL) (Famiglietti 1983b; Vaney 1984). Within those narrow strata, SBAC dendrites exhibit strong self-avoidance. The dendrites of an individual cell rarely cross over one another, giving rise to their characteristic “starburst” morphology (Lefebvre et al. 2012; Tauchi and Masland 1984; Vaney 1984); however, adjacent cells display a high degree of dendritic overlap (Keeley et al. 2007; Tauchi and Masland 1984; Vaney 1984) and form reciprocal inhibitory synapses with one another (Lee et al. 2010). The spatially asymmetric inhibitory connections that SBACs make with cofasciculating direction-selective ganglion cells (DSGCs) form the basis of the directional responses of these cells (Briggman et al. 2011; Fried et al. 2002; Lee et al. 2010; Taylor and Vaney 2002; Wei et al. 2011).

The ON and OFF SBAC dendritic strata are established before the arrival of the axon terminals of retinal bipolar cells (Morgan et al. 2006; Stacy and Wong 2003). Their development is guided at least in part by semaphorin6A signaling via plexinA2 receptors and Megf10 signaling between homotypic SBACs (Ray et al. 2018; Sun et al. 2013). Recent evidence demonstrates that the transcription factor Sox2 is also critical for the control of SBAC development, including soma positioning and dendritic stratification. When Sox2 is conditionally eliminated, migration of the ON and OFF SBACs into the GCL and INL is altered; somata are added to the INL, and a corresponding number are lost from the GCL (Whitney et al. 2014). This disruption likely reflects a deficit in subtype specification rather than defective migration, because although many choline acetyltransferase (ChAT)-positive cells exhibit bistratified dendritic arbors, the dendrites stratify within the ON and OFF sublaminae of the IPL at levels appropriate for SBACs. Moreover, the SBACs with somas in the INL, which are normally OFF-type cells, lack expression of the purine receptor P2X2, which is normally expressed heavily in OFF SBACs (Whitney et al. 2014). Together these observations indicate that differentiation of both ON and OFF SBACs is abnormal when Sox2 expression is eliminated.

ON and OFF bipolar cells normally make synaptic connections with ON and OFF SBACs, respectively (Famiglietti 1983a). Here we asked whether neuronal connectivity between both ON and OFF bipolar cells is established with bistratified SBACs. Bipolar cells express different cadherins, which are believed to confer laminar specificity and postsynaptic target connectivity (Duan et al. 2014). Might bistratified SBACs in the GCL lacking Sox2 still retain their subtype (ON) specificity through selective wiring mediated by molecular recognition, or do these cells now receive input from both ON and OFF bipolar cells? To address this question, we compared the excitatory synaptic inputs to SBACs in wild-type (Sox2-WT) and Sox2-conditional knockout (Sox2-CKO) mice. The data show that, unlike WT SBACs, the nominally ON-type SBACs in Sox2-CKO retinas exhibit ON-OFF excitatory responses, suggesting that the laminar positioning of an SBAC dendritic arbor is sufficient to confer bipolar cell connectivity. Curiously, the magnitude and time course of the ectopic OFF inputs and the “normal” ON inputs are smaller and more transient than those seen in WT retinas, suggesting that Sox2 also modulates other properties of synaptic physiology in SBACs.

MATERIALS AND METHODS

All procedures involving animals complied with the National Institutes of Health guidelines for animal use and a protocol approved by the Institutional Animal Care and Use Committees at Oregon Health & Science University. Three Sox2-WT and ten Sox2-CKO animals were transferred from the Animal Resource Center at the University of California, Santa Barbara to Oregon Health & Science University. Sox2-CKO mice were generated by crossing breeders homozygous for both Chat-Cre and Rosa-YFP reporter and heterozygous for the floxed Sox2 allele, thereby yielding Sox2-WT and Sox2-CKO mice from the same litters (mice heterozygous for the floxed Sox2 allele were not used) (Whitney et al. 2014). As described previously, all SBACs in the Sox2-CKO retina retain their cholinergic status, even if their somata are slightly smaller than those in Sox2-WT retinas. In whole mount preparations, these cells in the GCL (Fig. 1, A and B), like those in the INL (not shown), have lost their Sox2-immunopositive status in the Sox2-CKO retina, amidst a collection of Sox2-immunopositive Müller end feet, arising from the population of Sox2-immunopositive Müller cells situated in the INL (Fig. 1, C and D).

Fig. 1.

Fig. 1.

Open in a new tab

Sox2 is selectively eliminated from starburst amacrine cells (SBACs). A and B: SBACs retain their cholinergic status (red) in Sox2 conditional knockout (Sox2-CKO) retina, whereas only the SBACs in wild-type (Sox2-WT) retina retain their Sox2-immunopositive status (cyan). GCL, ganglion cell layer; ChAT, choline acetyltransferase. C and D: Müller glial somata remain Sox2-immunopositive, giving rise to the population of immunopositive Müller glial end feet in A and B. INL, inner nuclear layer. Calibration bar, 20 µm.

All mice were between 2 and 4 mo of age at the time of each experiment. Animals had ad libitum access to food and water and were kept on a 12:12-h light-dark cycle. Experiments were performed during the circadian day. For the physiological studies, mice were dark-adapted for at least 1 h before the retinas were isolated. The animals were deeply anesthetized by intraperitoneal injection of pentobarbital sodium (0.25 ml, 50 mg/ml) and killed by cervical dislocation immediately after enucleation. Retinas were isolated and maintained under dim red or infrared illumination for all subsequent procedures.

Immunofluorescence and single-cell injections.

For immunolabeling studies, mice were given a lethal injection of pentobarbital sodium and then intracardially perfused with 2–3 ml of physiological saline followed by 50 ml of 4% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.2 at 20°C). Eyes were removed and fixed by immersion for an additional 15 min. Whole retinas were dissected from the eyes and then immunostained with antibodies to ChAT (1:50; Millipore, no. AB144P) and to Sox2 (1:200; Abcam, no. AB97959), all as previously described (Whitney et al. 2014). For single-cell injections, whole retinas were dissected from deeply anesthetized mice and then fixed by immersion in the same fixative for 20 min. Single SBACs were impaled with a micropipette filled with Alexa 546 dye to label the dendritic arbor, and then retinas were subsequently immunostained for the presence of ChAT to reveal the ON and OFF starburst plexuses, all as previously described (Whitney et al. 2014).

Electrophysiology.

Whole cell patch recordings from SBAC somata, targeted because of their small spherical somata and lack of action potentials, were performed with borosilicate glass electrodes with a resistance of 4–8 MΩ. Electrodes were filled with an intracellular solution containing (in mM) 128 methanesulfonate, 6 CsCl, 10 Na-HEPES, 1 EGTA, 2 Mg-ATP, 1 Na-GTP, 5 phosphocreatine, 3 QX-314, and 0.1 spermine. The pH was adjusted to 7.4 with CsOH. Reagents were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated. To improve voltage clamp at positive potentials, potassium was replaced by cesium to suppress current through voltage-gated potassium channels. Voltages were adjusted by –10 mV to correct for the liquid junction potential.

Radial cuts were made on retinal halves to facilitate flat mounting on Whatman Anodiscs, which were then placed in a recording chamber and held in place with a harp. The preparation was constantly perfused with Ames medium (United States Biological, Salem, MA) at a flow rate of 3–4 ml/min and maintained at 32–34°C. Alexa 488, Alexa 594, or a mix was added to the pipette to visually confirm the identity of SBACs at the end of recordings.

Light stimulation and recording.

Stimuli generated on a CRT computer monitor at a refresh rate of 60 Hz were projected through the ×20 water-immersion microscope objective (numerical aperture = 0.95) onto the photoreceptor outer segments. Percentage contrast was defined as C=100(LStim−LBack)(LBack), where LStim and LBack are the stimulus and background intensities, respectively. The contrast of all stimuli was 80%. The retina was adapted to photopic levels by setting the screen intensity to ∼105 photons·μm−2·s−1 throughout the experiment. Center-surround sizes in Fig. 3 were measured by recording responses to stimulus spots with a range of diameters. The amplitudes of the current responses as a function of spot diameter were fitted with a difference-of-Gaussians function, to estimate the center and surround extents (Rodieck 1965). Light-evoked excitatory and inhibitory synaptic conductances were estimated from the current-voltage relations of the net light-evoked postsynaptic currents, as described previously (Borg-Graham 2001; Taylor and Vaney 2002; van Wyk et al. 2006; Venkataramani and Taylor 2010).

Fig. 3.

Fig. 3.

Open in a new tab

Average responses to spot stimuli. Membrane current recorded in starburst amacrine cells (SBACs) at a holding potential of −70 mV in wild-type (Sox2-WT, black) and Sox2 conditional knockout (Sox2-CKO, red) cells. A: average current traces for 6 cells. Bars above traces show light stimulus timing and contrast. The size of the stimulus spot, centered on the receptive field, is indicated above the timing bars. Shaded regions in traces show SEs. B: excitatory postsynaptic current (EPSC) amplitudes from A measured as a function of stimulus spot diameter at time points indicated by symbols in A. Solid lines show fits to a difference-of-Gaussians function. Rescaling the Sox2-CKO data (red dashed line) to match the peak magnitude of Sox2-WT data demonstrates the largely unchanged spatial properties. C and D: average EPSCs for 5 cells during negative contrast (dark spot) stimulation. Format and layout identical to A and B.

Excitatory postsynaptic current analysis.

Data analysis and figure preparation were performed with custom procedures in IGOR Pro (WaveMetrics, Tigard, OR). Excitatory inputs were recorded by holding at the chloride reversal potential (−70 mV). The power spectral density (PSD; see Fig. 5B) functions were calculated from 96 s of data in six Sox2-WT cells and 126 s of data in six Sox2-CKO cells. The PSD was calculated from the Fourier transform of the current records with routines implemented in the IGOR Pro software. Fourier transforms were performed on multiple 1.5-s segments of continuous baseline current. PSDs were calculated from the Fourier transforms for each cell and then averaged across cells to produce the data in Fig. 5.

Fig. 5.

Fig. 5.

Open in a new tab

Spontaneous excitatory postsynaptic currents (sEPSCs) recorded at a holding potential of −70 mV in wild-type (Sox2-WT, black) and Sox2 conditional knockout (Sox2-CKO, red) starburst amacrine cells (SBACs). A, left: representative segments of baseline currents. Gray trace shows the control (black) after high-pass filtering to remove the slow fluctuations. This allowed for efficient threshold detection of the fast sEPSCs as illustrated at right. Asterisks mark 4 detected sEPSCs that crossed the threshold. B: average power spectra were calculated from 96 s of data in 6 Sox2-WT and 126 s of data in 6 Sox2-CKO cells. Shading shows SD. C: average sEPSCs calculated from average sEPSCs derived from the same 6 cells (see materials and methods). Left: arrows show the estimation of the amplitude of the slow sEPSC components. Right: the fast component on an expanded timescale. Solid blue lines show exponential fits to the decay phase, with time constants of 0.45 ms (Sox2-WT) and 0.34 ms (Sox2-CKO). The corresponding amplitudes obtained from these exponential fits were −29 and −24 pA. Shading shows SDs calculated for the average sEPSCs across the 6 cells, after subtraction of the mean offset current.

The same data used for the PSD analysis were analyzed to detect spontaneous excitatory postsynaptic currents (sEPSCs). sEPSCs were isolated in a two-step process. First, the large, slow fluctuations in the membrane current, particularly prominent in the WT cells, were removed by high-pass filtering (e.g., see Fig. 5A). Second, the fast sEPSCs were identified as events that crossed a negative threshold of −15 pA in the filtered trace (see Fig. 5A, right). The same threshold was used for both WT and CKO cells. Average EPSCs were generated from 200-ms segments excised from the original, unfiltered records. Average sEPSCs in each cell were calculated from between 228 and 706 events excised from the current recordings in each of the six Sox2-WT cells. Similarly, the average sEPSC in each of the Sox2-CKO cells was calculated from between 165 and 854 events in each cell. A total of 2,695 and 2,530 events contributed to the overall averages for the Sox2-WT and Sox2-CKO data in Fig. 5.

RESULTS

ON-type SBACs in Sox2-CKO mice receive input from ON and OFF bipolar cells.

The ON-type SBACs, with somata in the GCL, were targeted on the basis of relatively small soma size and an absence of spike responses during extracellular recordings. SBACs were more difficult to find in Sox2-CKO retinas as their somata were smaller than in Sox2-WT retinas (Whitney et al. 2014), making them harder to distinguish from other nonspiking, amacrine cells. Cells were confirmed to be SBACs on the basis of dendritic morphology revealed by filling the cells with fluorescent dye (Alexa) during the recordings. As previously described (Whitney et al. 2014), such filled cells frequently show a less radially symmetric dendritic field (Fig. 2, A and B), with a variable proportion of the dendritic arbor mispositioned to the outer portion of the IPL. Figure 2, C and D, show the boxed region indicated in Fig. 2, A and B, with the ON and OFF strata shown separately (Fig. 2, C and D, top), and include the immunopositive cholinergic plexus along with three other cells for both Sox2-WT and Sox2-CKO retinas. Together, they show these two prominent alterations, namely, a reduction in the extent of the dendritic arbor in the ON stratum and the addition of ectopic dendrites ramifying in the OFF stratum.

Fig. 2.

Fig. 2.

Open in a new tab

Dendritic arbors of starburst amacrine cells (SBACs) in the Sox2 conditional knockout (Sox2-CKO) retina are bistratified. A and B: examples of single injected SBACs from wild-type (Sox2-WT) and Sox2-CKO retinas, showing the altered dendritic morphology of SBACs deprived of Sox2. C and D: portions of the dendritic arbor that lie within the 2 cholinergic immunolabeled plexes for these same 2 cells (top and highlighted box in A and B) and for 3 additional cells (bottom) in each condition. In Sox2-WT cells all of the dendritic arbor is contained within the ON stratum (shown in red, C), whereas Sox2-CKO cells show a reduced density of its dendritic arbor in the ON stratum (shown in red, D) and ectopic dendrites present in the OFF stratum (shown in green, D). The entire cholinergic plexus was immunolabeled as well (shown in blue), confirming that the bistratified dendritic arbors localize to the ON and OFF cholinergic strata. Calibration bar, 20 µm.

Our first goal was to determine whether these cells received excitatory inputs from bipolar cells appropriate for the cell type (ON, defined by somal positioning) or received excitatory inputs from both ON and OFF bipolar cells. To this end, we isolated EPSCs by holding the cells close to the chloride reversal potential. We stimulated the cells with light or dark spots (relative to a constant gray background; contrast was 80%) and measured the amplitudes of the EPSCs. We used a range of stimulus diameters to obtain estimates of the spatial extent of the center excitation and any inhibitory surround effects. WT SBACs displayed sustained EPSCs, either during a bright spot presentation or upon termination of a dark spot (Fig. 3, A and C). The magnitudes of the EPSCs in Sox2-CKO cells (Fig. 3, A and C) were smaller and more transient than those seen in Sox2-WT cells. For bright spot stimuli, EPSCs in the Sox2-WT and Sox2-CKO cells were evoked only at the onset (ON response) of the flashes (Fig. 3, A and B). An inward current was seen in the Sox2-WT cells at the fixed measurement time point 100 ms after termination of the bright flash (Fig. 3B). This current is probably not a separate OFF response but appears to result from slow turnoff of the ON EPSC activated during the light step. It was not seen in the Sox2-CKO cells (Fig. 3B), presumably because of faster shutoff of the light-evoked response. For dark spots, EPSCs were evoked at both the onset (OFF response) and termination (ON response) of the flashes (Fig. 3, C and D). Note the suppression of the tonic inward current during the dark stimulus for Sox2-WT cells (Fig. 3, C and D), as has been reported previously (Stincic et al. 2016; Taylor and Wässle 1995).

Previous anatomical analysis has shown that the stratification of ON SBACs in Sox2-CKO mice is disrupted, with asymmetric ON and OFF arbors; however, the total area enclosing both dendritic arbors is similar to that seen for the ON arbors in normal mice (Whitney et al. 2014). The area-response curves illustrated in Fig. 3, B and D, allowed us to compare the physiological receptive field sizes calculated from the EPSCs. In WT cells, the ON response elicited by the onset of a bright spot (Fig. 3B, left) or the termination of a dark spot (Fig. 3D, left) could be fit by a difference-of-Gaussians function, with center sizes (2 × σ) of 190 and 240 µm, respectively. Center sizes of 230 and 280 µm were measured in Sox2-CKO cells (Fig. 3, B and C). The data set is not sufficient to establish whether the larger estimates for center size are significant; however, the previous morphological analysis indicates that the anatomical spread of the dendrites is not larger in Sox2-CKO cells (Whitney et al. 2014). The corresponding surround sizes measured in Sox2-WT cells of 610 and 500 µm were smaller than the surrounds of 890 and 1,060 µm measured in Sox2-CKO cells.

In addition to excitatory inputs from bipolar cells, SBACs receive inhibitory inputs from other SBACs and also potentially from other amacrine cells (Chen et al. 2016; Ding et al. 2016; Lee and Zhou 2006; Taylor and Wässle 1995). To examine the magnitude and kinetics of both excitatory and inhibitory synaptic inputs to the SBACs, we recorded light responses over a range of holding potentials (Fig. 4). Average whole cell leakage currents in SBACs were approximately twofold smaller in Sox2-CKO cells (Fig. 4A). Since the total dendritic length of SBACs in Sox2-WT and Sox2-CKO cells are approximately the same (Whitney et al. 2014), the smaller leak currents suggest an increase in the specific membrane resistivity in the latter. Along with the increased input resistance, the excitatory conductance was smaller in Sox2-CKO cells relative to Sox2-WT cells (Fig. 4B), in line with the data shown in Fig. 3. On the other hand, the amplitudes of the inhibitory conductance in Sox2-WT and Sox2-CKO cells were similar (Fig. 4C).

Fig. 4.

Fig. 4.

Open in a new tab

Calculated excitatory and inhibitory conductances activated by 300-µm-diameter centered spots of positive and negative contrast. A: average membrane currents recorded at a range of membrane potentials between −90 and +30 mV. Stimulus contrast and timing indicated beneath the traces. B and C: excitatory (GE) and inhibitory (GI) conductances calculated for the data shown in A (see materials and methods).

Spontaneous release is altered in Sox2-CKO SBACs.

SBACs display spontaneous EPSCs under steady background illumination (Peters and Masland 1996; Petit-Jacques et al. 2005; Taylor and Wässle 1995; Vlasits et al. 2014). In the recordings presented here the spontaneous input appeared to comprise slow fluctuations with rapid events superimposed. Visual inspection suggested that the sEPSCs were qualitatively different in the Sox2-WT and Sox2-CKO cells. The Sox2-WT cells displayed slow fluctuations in the membrane current that were not evident in the Sox2-CKO cells (Fig. 5A). We quantified this difference by measuring the PSD in six Sox2-WT and six Sox2-CKO cells (Fig. 5B). The power at low frequencies was lower in the Sox2 CKO recordings, consistent with the loss of the slow fluctuations observed in the raw data. In addition to the slow fluctuations, rapid sEPSCs were evident in both Sox2-WT and Sox2-CKO cells. To examine these events in more detail, we isolated them with a threshold-crossing criterion (see materials and methods). The Sox2-WT data were high-pass filtered to remove the slow fluctuations and allow isolation of the rapid events. For each threshold crossing time, we excised a 200-ms segment from the original current record and added it to the average. The resulting average sEPSCs are shown in Fig. 5C for Sox2-WT and Sox2-CKO cells. The time courses of the fast events detected by the threshold crossing were measured as the time constants for exponential fits to the decay phases of the sEPSCs in Fig. 5C, right. The time constants were 0.45 and 0.34 ms for Sox2-WT and Sox2-CKO cells, respectively, similar to a time constant of 0.33 ms measured for similar events in rabbit SBACs (Taylor and Wässle 1995). Similarly, the amplitudes of the fast sEPSCs measured from the exponential fits were essentially the same at −29 and −24 pA for Sox2-WT and Sox2-CKO cells. The slow component in Sox2-WT cells had a 10–90% rise time of ~19 ms and an initial decay time constant of ~26 ms. The analysis revealed the presence of a slower component of the sEPSCs in the Sox2-WT cells that was much reduced in the Sox2-CKO cells. The amplitude of this slow component, measured from the sEPSCs 5 ms after the peak of the fast events (Fig. 5C, left), was −16.8 pA for the Sox2-WT cells but only −2.8 pA in the Sox2-CKO cells. In summary, analysis of the spontaneous excitatory inputs to the SBACs revealed the presence of fast and slow EPSC components. The fast component appeared to be similar in amplitude and time course in the Sox2-WT and Sox2-CKO cells, whereas the slow component appeared to be substantially reduced in the Sox2-CKO retinas.

DISCUSSION

Sox2 is an essential transcription factor during mammalian embryogenesis. It serves to maintain early cells in a pluripotent state and, in concert with a few other transcription factors, is sufficient for reestablishing pluripotency from otherwise differentiated cells (Feng and Wen 2015). It is also critical for organogenesis, including normal eye development, as mutations in Sox2 yield anophthalmia and other ocular dystrophies (Bardakjian and Schneider 2011). Curiously, Sox2 is downregulated as retinal development proceeds yet is retained into maturity in retinal Muller glia (Surzenko et al. 2013), astrocytes (Kautzman et al. 2018), and the two populations of SBACs. We have previously shown that Sox2 is critical for the subtype specification of the two populations of SBACs that differ in their somal positioning, dendritic stratification, and distinct P2X2 receptor expression patterns (Whitney et al. 2014). SBACs in the Sox2-CKO retina retain many WT properties (including their cholinergic status, their total number, their positioning within the INL and GCL, and the targeting of their dendrites to their normal depths within the IPL) yet lack features that fully discriminate them. In the absence of Sox2, although these cells stratify only within the ON and OFF cholinergic strata in the IPL, they lose the ON vs. OFF specificity for dendritic targeting, and many cells become bistratified. Nevertheless, they might retain other intrinsic instructions critical for producing two distinct populations in the two cellular layers. For example, they may retain an intrinsic instruction to connect selectively with ON vs. OFF bipolar cells.

The primary goal of this study was to determine whether ectopically bistratifying SBACs receive functional synaptic inputs from both ON- and OFF-type bipolar cells. The physiological analysis supports four key conclusions. First, SBACs located in the GCL, which are normally ON-type cells, receive both ON and OFF light-evoked excitatory inputs in Sox2-CKO retinas, consistent with input from ON and OFF bipolar cells. Second, compared with Sox2-WT cells, the excitatory inputs are more transient in Sox2-CKO cells, suggesting that the synaptic connections that form in either ectopically or entopically localized dendrites are not normal. Analysis of sEPSCs confirmed the loss of a slow component that was tightly correlated with fast sEPSCs. Third, the magnitude of the excitatory synaptic inputs to Sox2-CKO cells is reduced compared with Sox2-WT cells. And finally, unlike the excitatory synapses, the inhibitory synaptic inputs to SBACs are largely similar in Sox2-CKO and Sox2-WT cells.

The smaller excitatory inputs activated by light (ON responses) in the Sox2-CKO cells compared with WT cells (Fig. 3A) may be due in part to the reduction in the dendritic arbor positioned within the ON stratum of these bistratifying SBACs (Whitney et al. 2014). The reduced ON arbors would present fewer opportunities to make connections with ON-type bipolar cells, yielding a reduction in the average peak ON response. Similarly, the size of the ectopic OFF arbor in these cells may limit the size of the OFF response. Homeostatic mechanisms could also contribute. The input resistance of the Sox2-CKO cells was higher than that of Sox2-WT cells, as evident from the smaller leak currents observed in Fig. 4A. A higher input resistance will mean a larger voltage change per unit synaptic current. Perhaps the smaller synaptic currents in the Sox2-CKO cells result in part from homeostatic mechanisms that downregulate synaptic gain to regulate overall excitability (Wefelmeyer et al. 2016). Moreover, if the total dendritic length and membrane area of SBACs in Sox2-CKO and Sox2-WT mice were similar (Whitney et al. 2014), then the change in input resistance might indicate changes in the specific membrane resistivity.

Analysis of the light-evoked synaptic inputs demonstrated that the sustained ON excitation seen in Sox2-WT cells was replaced with more transient ON and OFF excitation in the Sox2-CKO cells (Fig. 3). The shift from sustained to more transient excitatory inputs was echoed in the analysis of the spontaneous excitatory currents observed in the absence of light stimulation. Background spontaneous activity has been documented in SBACs from rabbit (Peters and Masland 1996; Taylor and Wässle 1995) and mouse (Petit-Jacques et al. 2005; Vlasits et al. 2014). Both species display fast sEPSCs along with slower fluctuations in the inward current, in line with the present results (Fig. 5). Our analysis of fast sEPSCs reveals for the first time that these fast sEPSCs are temporally correlated with slower spontaneous fluctuations in the membrane current (Fig. 5C). After filtering out the slow current fluctuations (Fig. 5A, left) and thresholding to detect fast sEPSCs (Fig. 5A, right), we generated the average fast sEPSC by averaging epochs from the original unfiltered records. If the fast and slow currents were uncorrelated, the average fast sEPSC should have appeared on a flat baseline with an inward current offset equal to the mean membrane current over the analyzed data. Instead, the average fast sEPSC revealed the presence of temporally correlated slower EPSC-like current fluctuations (Fig. 5C). It remains unclear whether there are tightly correlated fast and slow glutamate release processes or a single process with complex kinetics. In any case, with identical analysis procedures, the slow component was not seen in the Sox2-CKO cells (Fig. 5), suggesting a fundamental difference in the properties in excitatory synapses in these retinas. A comparison of other studies in mouse suggests that the balance between sustained and transient excitation to SBACs can vary considerably, from mostly transient EPSCs (Hoggarth et al. 2015; Petit-Jacques and Bloomfield 2008) to a more balanced mix of transient and sustained (Stincic et al. 2016; Vlasits et al. 2014), similar to the present data. The reasons for these differences in time course remain unclear. It is possible that the same underlying mechanisms are activated but to different extents because of differences in adaptation state or stimulus conditions. It is also interesting to note that the light-evoked OFF EPSCs that appear in the Sox2-CKO cells are activated at the onset of a dark spot but not at the termination of a bright spot (Fig. 3), whereas the ON EPSCs are activated at both the onset of a bright spot and the termination of a dark spot. This asymmetry between the ON and OFF inputs seems consistent with the difference in rectification for synaptic transmission in the ON and OFF pathways (Chichilnisky and Kalmar 2002; Liang and Freed 2010). Under this scenario, ON bipolar cells tend to rest above release threshold during background illumination (and modulate output above and below a tonic baseline level; see Fig. 3, A and C, and Fig. 4B), while OFF bipolar cells tend to rest below release threshold. Thus, upon return to background at the termination of a nonpreferred contrast stimulus, ON bipolar cells exceed release threshold and increase glutamate release whereas OFF bipolar cells do not. In this context it is interesting to note that OFF responses have been observed in ON SBACs under some conditions (Vlasits et al. 2014), and the authors proposed that such OFF responses resulted from differential adaptation effects produced by localized rod saturation combined with horizontal cell-mediated surround signals arising from cones. We failed to observe similar adaptation effects in our experiments, possibly because of our lower stimulus contrast.

Recent work has established that ON SBACs in the mouse receive excitatory inputs from sustained type 7 and more transient type 5 ON bipolar cells that form synaptic contacts at different distances from the soma (Ding et al. 2016; Kim et al. 2014). It is tempting to speculate that the present results reflect differences in the ability of the sustained bipolar cells to make synaptic connections with SBACs in the Sox2-CKO retinas. Under this scenario, both type 5 and type 7 cells would contribute fast sEPSCs, whereas type 5 cells would contribute additional slow events, which would be highly correlated with some or all of the fast events. The loss of the slow events in Sox2-CKO cells might be explained by an impaired ability of SBACs to form synapses with type 5 ON bipolar cells. Furthermore, the results indicate that the OFF inputs on ectopically located dendrites avoid connections with sustained OFF bipolar cell subtypes (e.g., type 2) (Greene et al. 2016). Since we could not directly detect the slow sEPSCs, it remains possible that they can appear in the absence of the fast sEPSC. Further work is needed to test this hypothesis.

In contrast to the excitatory input, the amplitude of the inhibition appeared to be similar in Sox2-WT and Sox2-CKO cells (Fig. 4). Moreover, area-response analysis revealed the presence of surround suppression of excitatory inputs to Sox2-CKO cells (Taylor and Wässle 1995). The spatial extent and strength of the surround also appeared to be relatively normal in the Sox2-CKO cells (Fig. 3B). These results suggest that both the feedforward inhibition onto the SBACs and the inhibitory connections onto the bipolar cells presynaptic to the SBACs are intact and functioning fairly normally. Altogether, the results indicate pronounced changes in the direct excitatory inputs to the SBACs but milder effects on inhibitory circuits in the absence of Sox2.

The processes of SBACs, like other retinal neurons, are thought to form selective synaptic contacts by virtue of their targeted distribution within the IPL, where they engage in homophilic or heterophilic molecular interactions with their synaptic partners (Duan et al. 2014; Krishnaswamy et al. 2015; Yamagata and Sanes 2008). SBACs in the GCL have recently been shown to express Cntn5, and its expression in these ON SBACs is critical for the formation of the inner stratum of dendrites in bistratifying DSGCs (Peng et al. 2017). Cntn5 is also expressed by those same ON-OFF DSGCs and is regulated by Satb1. Loss of this transcriptional regulator yields a loss of Cntn5 from those DSGCs, yielding a comparable reduction of the inner processes of their bistratifying dendritic arbors, rendering them monostratified, producing only OFF responses (Peng et al. 2017). The present study, by contrast, has shown that the loss of Sox2 has turned a monostratifying ON amacrine cell into a bistratifying ON-OFF one. It remains to be seen whether these SBACs in the GCL also have lost Cntn5 expression and whether there is any downstream consequence for the DSGC.

A recent study suggests that ON and OFF bipolar cells are attracted to the strata occupied by their respective ON and OFF SBACs, even when those SBACs generate ectopic microstratifications in association with their primary ON or OFF strata (Ray et al. 2018). The present results suggest that specification of laminar position, orchestrated by Sox2 well before bipolar cell differentiation, is enough to generate ON vs. OFF bipolar cell connectivity. Other factors direct bipolar cells to the ON vs. OFF divisions in the IPL (Duan et al. 2014), and once in position they make connections with any SBAC dendrites, since the SBACs (in the Sox2-CKO retina) lack specificity for a bipolar cell’s ON vs. OFF status. In this scenario, normal Sox2-dependent ON vs. OFF specification may be due to gene expression that prevents dendrites from entering the OFF vs. ON divisions of the IPL, respectively, perhaps mediated by repulsive factors (see, e.g., Matsuoka et al. 2011). Once a bipolar terminal has reached one of the cholinergic strata, other Sox2-dependent factors intrinsic to the SBACs may be important for forming functionally normal synapses.

GRANTS

This research was supported by National Eye Institute Grants EY-014888, EY-022070 (W. R. Taylor), and EY-019968 (B. E. Reese).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.L.S., B.E.R., and W.R.T. conceived and designed research; T.L.S. and P.W.K. performed experiments; T.L.S., P.W.K., B.E.R., and W.R.T. analyzed data; T.L.S., P.W.K., B.E.R., and W.R.T. interpreted results of experiments; T.L.S., P.W.K., B.E.R., and W.R.T. prepared figures; T.L.S., B.E.R., and W.R.T. drafted manuscript; T.L.S., P.W.K., B.E.R., and W.R.T. edited and revised manuscript; T.L.S., P.W.K., B.E.R., and W.R.T. approved final version of manuscript.

REFERENCES

  1. Bardakjian TM, Schneider A. The genetics of anophthalmia and microphthalmia. Curr Opin Ophthalmol 22: 309–313, 2011. doi: 10.1097/ICU.0b013e328349b004. [DOI] [PubMed] [Google Scholar]
  2. Borg-Graham LJ. The computation of directional selectivity in the retina occurs presynaptic to the ganglion cell. Nat Neurosci 4: 176–183, 2001. doi: 10.1038/84007. [DOI] [PubMed] [Google Scholar]
  3. Briggman KL, Helmstaedter M, Denk W. Wiring specificity in the direction-selectivity circuit of the retina. Nature 471: 183–188, 2011. doi: 10.1038/nature09818. [DOI] [PubMed] [Google Scholar]
  4. Chen Q, Pei Z, Koren D, Wei W. Stimulus-dependent recruitment of lateral inhibition underlies retinal direction selectivity. eLife 5: e21053, 2016. doi: 10.7554/eLife.21053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chichilnisky EJ, Kalmar RS. Functional asymmetries in ON and OFF ganglion cells of primate retina. J Neurosci 22: 2737–2747, 2002. doi: 10.1523/JNEUROSCI.22-07-02737.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Ding H, Smith RG, Poleg-Polsky A, Diamond JS, Briggman KL. Species-specific wiring for direction selectivity in the mammalian retina. Nature 535: 105–110, 2016. doi: 10.1038/nature18609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Duan X, Krishnaswamy A, De la Huerta I, Sanes JR. Type II cadherins guide assembly of a direction-selective retinal circuit. Cell 158: 793–807, 2014. doi: 10.1016/j.cell.2014.06.047. [DOI] [PubMed] [Google Scholar]
  8. Famiglietti EV., Jr On and off pathways through amacrine cells in mammalian retina: the synaptic connections of “starburst” amacrine cells. Vision Res 23: 1265–1279, 1983a. doi: 10.1016/0042-6989(83)90102-5. [DOI] [PubMed] [Google Scholar]
  9. Famiglietti EV., Jr “Starburst” amacrine cells and cholinergic neurons: mirror-symmetric on and off amacrine cells of rabbit retina. Brain Res 261: 138–144, 1983b. doi: 10.1016/0006-8993(83)91293-3. [DOI] [PubMed] [Google Scholar]
  10. Feng R, Wen J. Overview of the roles of Sox2 in stem cell and development. Biol Chem 396: 883–891, 2015. doi: 10.1515/hsz-2014-0317. [DOI] [PubMed] [Google Scholar]
  11. Fried SI, Münch TA, Werblin FS. Mechanisms and circuitry underlying directional selectivity in the retina. Nature 420: 411–414, 2002. doi: 10.1038/nature01179. [DOI] [PubMed] [Google Scholar]
  12. Greene MJ, Kim JS, Seung HS; EyeWirers . Analogous convergence of sustained and transient inputs in parallel on and off pathways for retinal motion computation. Cell Rep 14: 1892–1900, 2016. doi: 10.1016/j.celrep.2016.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hayden SA, Mills JW, Masland RM. Acetylcholine synthesis by displaced amacrine cells. Science 210: 435–437, 1980. doi: 10.1126/science.7433984. [DOI] [PubMed] [Google Scholar]
  14. Hoggarth A, McLaughlin AJ, Ronellenfitch K, Trenholm S, Vasandani R, Sethuramanujam S, Schwab D, Briggman KL, Awatramani GB. Specific wiring of distinct amacrine cells in the directionally selective retinal circuit permits independent coding of direction and size. Neuron 86: 276–291, 2015. doi: 10.1016/j.neuron.2015.02.035. [DOI] [PubMed] [Google Scholar]
  15. Kautzman AG, Keeley PW, Nahmou MM, Luna G, Fisher SK, Reese BE. Sox2 regulates astrocytic and vascular development in the retina. Glia 66: 623–636, 2018. doi: 10.1002/glia.23269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Keeley PW, Whitney IE, Raven MA, Reese BE. Dendritic spread and functional coverage of starburst amacrine cells. J Comp Neurol 505: 539–546, 2007. doi: 10.1002/cne.21518. [DOI] [PubMed] [Google Scholar]
  17. Kim JS, Greene MJ, Zlateski A, Lee K, Richardson M, Turaga SC, Purcaro M, Balkam M, Robinson A, Behabadi BF, Campos M, Denk W, Seung HS; EyeWirers . Space-time wiring specificity supports direction selectivity in the retina. Nature 509: 331–336, 2014. doi: 10.1038/nature13240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Krishnaswamy A, Yamagata M, Duan X, Hong YK, Sanes JR. Sidekick 2 directs formation of a retinal circuit that detects differential motion. Nature 524: 466–470, 2015. doi: 10.1038/nature14682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lee S, Kim K, Zhou ZJ. Role of ACh-GABA cotransmission in detecting image motion and motion direction. Neuron 68: 1159–1172, 2010. doi: 10.1016/j.neuron.2010.11.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lee S, Zhou ZJ. The synaptic mechanism of direction selectivity in distal processes of starburst amacrine cells. Neuron 51: 787–799, 2006. doi: 10.1016/j.neuron.2006.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lefebvre JL, Kostadinov D, Chen WV, Maniatis T, Sanes JR. Protocadherins mediate dendritic self-avoidance in the mammalian nervous system. Nature 488: 517–521, 2012. doi: 10.1038/nature11305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Liang Z, Freed MA. The ON pathway rectifies the OFF pathway of the mammalian retina. J Neurosci 30: 5533–5543, 2010. doi: 10.1523/JNEUROSCI.4733-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Masland RH, Mills JW. Autoradiographic identification of acetylcholine in the rabbit retina. J Cell Biol 83: 159–178, 1979. doi: 10.1083/jcb.83.1.159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Matsuoka RL, Nguyen-Ba-Charvet KT, Parray A, Badea TC, Chédotal A, Kolodkin AL. Transmembrane semaphorin signalling controls laminar stratification in the mammalian retina. Nature 470: 259–263, 2011. doi: 10.1038/nature09675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Morgan JL, Dhingra A, Vardi N, Wong RO. Axons and dendrites originate from neuroepithelial-like processes of retinal bipolar cells. Nat Neurosci 9: 85–92, 2006. doi: 10.1038/nn1615. [DOI] [PubMed] [Google Scholar]
  26. Peng YR, Tran NM, Krishnaswamy A, Kostadinov D, Martersteck EM, Sanes JR. Satb1 regulates contactin 5 to pattern dendrites of a mammalian retinal ganglion cell. Neuron 95: 869–883.e6, 2017. doi: 10.1016/j.neuron.2017.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Peters BN, Masland RH. Responses to light of starburst amacrine cells. J Neurophysiol 75: 469–480, 1996. doi: 10.1152/jn.1996.75.1.469. [DOI] [PubMed] [Google Scholar]
  28. Petit-Jacques J, Bloomfield SA. Synaptic regulation of the light-dependent oscillatory currents in starburst amacrine cells of the mouse retina. J Neurophysiol 100: 993–1006, 2008. doi: 10.1152/jn.01399.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Petit-Jacques J, Völgyi B, Rudy B, Bloomfield S. Spontaneous oscillatory activity of starburst amacrine cells in the mouse retina. J Neurophysiol 94: 1770–1780, 2005. doi: 10.1152/jn.00279.2005. [DOI] [PubMed] [Google Scholar]
  30. Ray TA, Roy S, Kozlowski C, Wang J, Cafaro J, Hulbert SW, Wright CV, Field GD, Kay JN. Formation of retinal direction-selective circuitry initiated by starburst amacrine cell homotypic contact. eLife 7: e34241, 2018. doi: 10.7554/eLife.34241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Reese BE. Development of the retina and optic pathway. Vision Res 51: 613–632, 2011. doi: 10.1016/j.visres.2010.07.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rodieck RW. Quantitative analysis of cat retinal ganglion cell response to visual stimuli. Vision Res 5: 583–601, 1965. doi: 10.1016/0042-6989(65)90033-7. [DOI] [PubMed] [Google Scholar]
  33. Stacy RC, Wong RO. Developmental relationship between cholinergic amacrine cell processes and ganglion cell dendrites of the mouse retina. J Comp Neurol 456: 154–166, 2003. doi: 10.1002/cne.10509. [DOI] [PubMed] [Google Scholar]
  34. Stincic T, Smith RG, Taylor WR. Time course of EPSCs in ON-type starburst amacrine cells is independent of dendritic location. J Physiol 594: 5685–5694, 2016. doi: 10.1113/JP272384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sun LO, Jiang Z, Rivlin-Etzion M, Hand R, Brady CM, Matsuoka RL, Yau KW, Feller MB, Kolodkin AL. On and off retinal circuit assembly by divergent molecular mechanisms. Science 342: 1241974, 2013. doi: 10.1126/science.1241974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Surzenko N, Crowl T, Bachleda A, Langer L, Pevny L. SOX2 maintains the quiescent progenitor cell state of postnatal retinal Muller glia. Development 140: 1445–1456, 2013. doi: 10.1242/dev.071878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Tauchi M, Masland RH. The shape and arrangement of the cholinergic neurons in the rabbit retina. Proc R Soc Lond B Biol Sci 223: 101–119, 1984. doi: 10.1098/rspb.1984.0085. [DOI] [PubMed] [Google Scholar]
  38. Taylor WR, Vaney DI. Diverse synaptic mechanisms generate direction selectivity in the rabbit retina. J Neurosci 22: 7712–7720, 2002. doi: 10.1523/JNEUROSCI.22-17-07712.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Taylor WR, Wässle H. Receptive field properties of starburst cholinergic amacrine cells in the rabbit retina. Eur J Neurosci 7: 2308–2321, 1995. doi: 10.1111/j.1460-9568.1995.tb00652.x. [DOI] [PubMed] [Google Scholar]
  40. van Wyk M, Taylor WR, Vaney DI. Local edge detectors: a substrate for fine spatial vision at low temporal frequencies in rabbit retina. J Neurosci 26: 13250–13263, 2006. doi: 10.1523/JNEUROSCI.1991-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Vaney DI. “Coronate” amacrine cells in the rabbit retina have the “starburst” dendritic morphology. Proc R Soc Lond B Biol Sci 220: 501–508, 1984. doi: 10.1098/rspb.1984.0016. [DOI] [PubMed] [Google Scholar]
  42. Vaney DI, Peichl L, Boycott BB. Matching populations of amacrine cells in the inner nuclear and ganglion cell layers of the rabbit retina. J Comp Neurol 199: 373–391, 1981. doi: 10.1002/cne.901990305. [DOI] [PubMed] [Google Scholar]
  43. Venkataramani S, Taylor WR. Orientation selectivity in rabbit retinal ganglion cells is mediated by presynaptic inhibition. J Neurosci 30: 15664–15676, 2010. doi: 10.1523/JNEUROSCI.2081-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Vlasits AL, Bos R, Morrie RD, Fortuny C, Flannery JG, Feller MB, Rivlin-Etzion M. Visual stimulation switches the polarity of excitatory input to starburst amacrine cells. Neuron 83: 1172–1184, 2014. doi: 10.1016/j.neuron.2014.07.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wefelmeyer W, Puhl CJ, Burrone J. Homeostatic plasticity of subcellular neuronal structures: from inputs to outputs. Trends Neurosci 39: 656–667, 2016. doi: 10.1016/j.tins.2016.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wei W, Hamby AM, Zhou K, Feller MB. Development of asymmetric inhibition underlying direction selectivity in the retina. Nature 469: 402–406, 2011. doi: 10.1038/nature09600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Whitney IE, Keeley PW, St. John AJ, Kautzman AG, Kay JN, Reese BE. Sox2 regulates cholinergic amacrine cell positioning and dendritic stratification in the retina. J Neurosci 34: 10109–10121, 2014. doi: 10.1523/JNEUROSCI.0415-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yamagata M, Sanes JR. Dscam and Sidekick proteins direct lamina-specific synaptic connections in vertebrate retina. Nature 451: 465–469, 2008. doi: 10.1038/nature06469. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Neurophysiology are provided here courtesy of American Physiological Society

RESOURCES