| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Departments of Physiology and Biophysics and Ophthalmology, University of Arkansas for Medical Sciences, Little Rock, Arkansas 72205; 2 Institute of Neuroscience, Shanghai Institutes of Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China
Submitted 24 November 2003; accepted in final form 12 December 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
20 s) but shared a similar frequency and propagation speed with the IR waves. Simultaneous Ca2+ imaging in VZ and multi-electrode array recording in the ganglion cell layer (GCL) revealed close spatiotemporal correlation between spontaneous VZ and IR waves, suggesting a common source of initiation and/or regulation of the two waves. Pharmacological studies further showed that all drugs that blocked IR waves also blocked VZ waves. However, the muscarinic antagonist atropine selectively blocked VZ but not IR waves at this developmental stage, indicating that IR waves were not dependent on VZ waves, but VZ waves likely relied on the initiation of IR waves. Eliciting IR waves with puffs of nicotinic or non-N-methyl-D-aspartate agonists in GCL produced atropine-sensitive waves in the VZ, demonstrating a unique, retrograde signaling pathway from IR to VZ. Thus differentiated neurons in the IR use spontaneous, rhythmic waves to send both forward signals to the central visual targets and retrograde messages to the developing cells in the VZ. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
). Once a cell has undergone its final division, it migrates away from the VZ toward its final destination and begins to differentiate. The first cells leaving the VZ are ganglion and amacrine cells, which form the inner layers of the retina. The newly differentiated ganglion and amacrine cells soon undergo rhythmic, spontaneous bursts of excitation, which propagates laterally in the form of waves (Catsicas and Mobbs 1995; Copenhagen 1991; Feller 1999; Katz and Shatz 1996; O'Donovan 1999; Wong 1999; Zhou 2001b) and has been shown in mammals to influence the development of precise retinogenicular connectivity (Penn et al. 1998).
Although the role of retinal waves in central visual development has been investigated extensively, relatively little attention has been paid to the possible involvement of retinal waves in the development of the retina itself. One of the important issues has been whether spontaneous retinal waves are present only among differentiated neurons in the inner retina (IR) or they also exist among retinal progenitor cells in the VZ. Because the developmental states of VZ and IR cells are drastically different, a presence of retinal waves in both VZ and IR would suggest new and diverse developmental roles played by the waves. Indeed, ventricular Ca2+ transients have been reported in embryonic chick retinal slices (Catsicas et al. 1998), suggesting wave-like activities in the retinal VZ. However, it could not be determined in retinal slices whether the transients actually propagated laterally as waves nor was it clear what drove the VZ transients. Moreover, because similar Ca2+ transients have never been found in the VZ of the mammalian retina despite considerable efforts to identify spontaneous activity in various retinal layers (Wong et al. 1995), it has been assumed that retinal waves in mammals are restricted to differentiated neurons in the IR (Catsicas et al. 1998).
In this study, we investigated whether propagating Ca2+ waves exist in the VZ of the developing rabbit retina and, in particular, where spontaneous retinal waves initiate and propagate in the mammalian retina. Our results show that Ca2+ waves existed in the VZ of the developing rabbit retina; these waves had distinctive propagation patterns, unique pharmacological properties, and prolonged Ca2+ dynamics; and retinal waves appeared to be initiated in the IR and spread to the VZ. Preliminary results of this study were published in abstract form (Zhou et al. 2002).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Retinal flatmount and slice preparations were prepared from New Zealand white rabbits aged between embryonic day E24 and E31 (day of birth) as previously described (Zhou 2001a; Zhou and Fain 1995; Zhou and Zhao 2000). All procedures involving the use of animals were in accordance with the National Institutes of Health guidelines as implemented by the university animal care and use committee. In brief, eyes were isolated from rabbits immediately following death by an overdose of pentobarbital (ip). Retinas were isolated from the eyes at 10°C in oxygenated, HEPES-buffered Ames medium, which was modified from Ames medium (Ames and Nesbett 1981) by replacing NaHCO3 with 20 mM HEPES (pH adjusted to 7.4 with NaOH). Each retina was cut into two to four pieces and flat-mounted on filter paper (Type HABP; Millipore, Bedford, MA) with either the ganglion cell layer (GCL) or the VZ facing paper. Retinal flatmounts were loaded with the calcium indicator dye Fura 2-AM (10 µM) or Fluo 4-AM (5 µM) (Molecular Probes, Eugene, OR) in the presence of 0.001% pluronic acid in oxygenated, HEPES-buffered Ames medium at 30° C for 1–2 h. Under this condition, only cells near the retinal surface facing the solution were adequately loaded with the dye. Depending on the specific purpose of the experiment, Ca2+ imaging was performed in one of the following configurations. Conventional Ca2+ imaging was made with an intensified cooled CCD camera (I-Pantamax, Roper Scientific Instruments, Princeton, NJ) under a fixed-stage, upright microscope (Olympus BX50WI, Olympus USA, New York, NY) using either a x10 (UMPlanFl, NA = 0.3, Olympus USA) or a x100 (UMPlanFl, NA = 1.00, Olympus USA) water-immersion objective lens. Two-photon Ca2+ imaging was made under an upright microscope (Axioskop2 configured with LSM510, Carl Zeiss, Thornwood, NY) with a x40 water-immersion objective lens (Achroplan, NA = 0.76, Carl Zeiss). To measure ventricular cell responses to puffs of neurotransmitter agonists at the GCL, Ca2+ imaging was made from Fluo 4-AM-loaded VZ cells under an inverted microscope (Olympus IX70, Olympus USA) with a x10 objective lens (NA = 0.3, UPlanFl, Olympus USA). Fura-2-loaded cells were imaged at an excitation wavelength of 380 nm (712.7 nm during 2-photon imaging) and displayed a decreased emission intensity (at 500 nm) when free [Ca2+]i was elevated (Grynkiewicz et al. 1985). Fluo-4-loaded cells were excited at 450 nm, and the emission intensity (at 500 nm) increased when free [Ca2+]i was elevated.1
To measure spontaneous spikes in the GCL and Ca2+ waves in the VZ simultaneously, a piece of Fura 2-AM-loaded retina was placed, GCL facing down, on a multielectrode array (MEA) (Multi Channel Systems, Reutlingen, Germany). The retina was then covered by a piece of transparent membrane filter (Millicell-CM, Millipore) that was held in place by a platinum ring. The MEA was made of 60 planar electrodes spaced 200 µm apart in an 8 x 8 array without electrodes at the four corners. The electrodes in the MEA were built on a glass substrate and connected to a 60-channel amplifier system (MEA 60, Multi Channel Systems) for multi-channel recording. Simultaneous Ca2+ imaging from the ventricular surface of the same retina was made with a cooled CCD camera (I-Pentamax) through a x10 water-immersion objective lens (NA = 0.3, UplanFl, Olympus USA) under an upright microscope (BX 50WI, Olympus USA). The acquisition of multielectrode and Ca2+-imaging data were triggered simultaneously. The 8 x 8 electrode array covered a 1.4 x 1.4 mm2 area, which was slightly larger than the 1.15 x 1.15 mm2 area imaged by the CCD camera under the x10 objective lens (Fig. 8B).
|
2 ml) or pressure injection with a Picospitzer-II (General Valve, Fairfield, NJ). Data analysis
Fluorescence images were analyzed with the assistance of Axon Imaging Workbench (AIW) (Axon Instruments) and MetaMorph (Universal Imaging, Downingtown, PA) software. To monitor the intracellular Ca2+ activity, oval zones were drawn around individual dye-loaded cells or local areas of the retina. The average fluorescence intensity in each zone was plotted as a function of time F(t). The relative fluorescence change (
F/Fo) was defined as [F(t) – Fo(t)]/Fo(t), where the baseline intensity Fo(t) was defined as the intensity immediately before a wave occurred. In most cases, Fo(t) was determined by drawing a straight line or a curve connecting the intensity values measured in the absence of waves. To calculate relative changes in fluorescence intensity produced by bath application of thapsigargin (Fig. 3), the baseline intensity during drug application was determined by extrapolating the baseline intensity curve measured immediately before drug application. Difference images (F) of a wave were made by subtracting a control image (averaged from 4 frames recorded immediately before a wave started) from images recorded during the wave. The speed of the wave was measured as the rate of wavefront displacement in the direction of wave propagation using AIW (Zhou and Zhao 2000) or MetaMorph. The duration of a Ca2+ transient was defined as the 10%-amplitude width on the F/Ft curve. Measurements of wave dynamics were expressed as means ± SD.
|
F(t)/F(t) was calculated based on the Auto-Regressive Integrated Moving Average (ARMA) model (Box et al. 1994) using SAS software (version 8.2, SAS Institute, Cary, NC) over a recording period of 2,400 s (2,400 pairs of sample points) with the lag time between –60 and +60 s in 1-s intervals. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
To determine whether correlated activity occurs spontaneously in the VZ of the developing mammalian retina, isolated retinas from E24-P0 rabbits were mounted on filter paper (ganglion cells facing paper) and loaded with the Ca2+-sensitive dye Fura 2-AM or Fluo 4-AM. Ca2+ imaging from the ventricular surface of the retina under a x10 objective lens revealed spontaneous intracellular Ca2+ increases that propagated laterally in the form of waves (Fig. 1A). These waves occurred rhythmically once every 1–5 min and lasted 10–20 s at each occurrence. To find out if the fluorescence signals were produced by changes in [Ca2+]i in ventricular cells, we imaged Fura-2 fluorescence from cells near the ventricular surface under an x100 objective lens. Figure 1B shows a fluorescence image of such cells in an E29 retina. At this magnification, individual cells at the ventricular surface were clearly visible, and changes in their [Ca2+]i were measured as F/F from small circular zones drawn around the cell bodies. As exemplified by the signals measured from seven randomly selected cells shown in Fig. 1C, most cells near the ventricular surface displayed rhythmic, spontaneous Ca2+ transients that were highly correlated among neighboring cells, indicating that ventricular cells directly mediated the propagating Ca2+ waves.
|
F/F reaching as high as 50% in Fura 2-AM-loaded cells (Fig. 2A). In contrast, because of the lack of sufficient Fura 2-AM diffusion into deeper layers of the retina, the measured fluorescence intensity decreased dramatically as two-photon optical sections were made at 20 and 40 µm below the VZ surface (Fig. 2B), indicating that IR cells had negligible contribution to the F/F measured at the VZ surface. The lack of sufficient dye loading in deeper layers of the retina under our experimental condition (with GCL attached to filter paper) was also evident from fluorescence images taken from vertical slices of the dye-loaded retina (Fig. 2C). Thus the fluorescence signals measured from the VZ surface represented Ca2+ activities of cells near the VZ surface. These results demonstrated for the first time the existence of rhythmic, propagating VZ waves in the developing mammalian retina.
|
To determine whether neurotransmitter interactions played a role in the propagation of VZ waves, the effects of antagonists of common neurotransmitter receptors were tested. During the period between E24 and E31, IR waves in rabbits are mediated mainly by nicotinic neurotransmission (Zhou and Zhao 2000). Application of the nicotinic receptor antagonist hexamethonium (50–100 µM) completely blocked the VZ wave in a partially reversible manner (Fig. 3A), indicating a critical role of nicotinic activation in the formation of VZ waves. Interestingly, the muscarinic antagonist atropine, which did not affect the IR waves at these ages (Zhou and Zhao 2000), completely blocked VZ wave (Fig. 3B), suggesting the muscarinic system is pivotal in mediating ventricular, but not IR, waves. The muscarinic effect on the VZ wave was likely mediated by the M1 receptors because pirenzepine, an antagonist considerably selective for M1 receptor at the concentration used (2 µM), completely blocked the VZ wave, whereas the M2 receptor antagonist gallamine had no effect on the wave (Fig. 3C). It is not clear if the VZ wave also required other subtypes (M3–M5) of muscarinic receptor activation.
To investigate whether ACh could directly influence the excitability of ventricular cells, nicotine and muscarine were puffed locally to the ventricular surface of Fura 2-AM-loaded E29 retinas. As shown in Fig. 4 (left), the muscarine puff evoked a robust Ca2+ response that was inhibited by atropine, consistent with the previous finding that ventricular cells in the rabbit retina respond to exogenous muscarinic agonists (Wong 1995). Puffing nicotine evoked a much smaller, yet clearly detectable Ca2+ increase, which could be blocked by hexamethonium (Fig. 4, right). The response persisted in the presence of a cocktail of antagonists containing 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), D(–)-2-amino-7-phosphonoheptanoic acid (AP-7), picrotoxin, strychnine, atropine, and 18-
glycyrrhetinic acid (18-GA), which should block most indirect responses. Thus VZ cells expressed functional muscarinic and, to a much less degree, nicotinic receptors, suggesting that ACh might act directly on VZ cells to mediate the VZ wave.
|
), suggesting the hydrolysis of acetylcholine had a profound impact on VZ, but not IR waves. Thus in addition to ACh, the endogenous AChE activity was also a key factor in the regulation of Ca2+ dynamics in VZ cells.
In contrast to cholinergic antagonists, blockers of ionotropic glutamate and GABA receptors (CNQX, AP-7, and picrotoxin), applied either singly or in combination, did not block VZ waves in E24-P0 rabbits (Fig. 5). Thus as with early IR waves (before P1) in rabbits (Zhou and Zhao 2000), glutamate and GABA receptor-mediated fast neurotransmissions were not essential in generating VZ waves at this stage. However, blocking adenosine receptors with aminophylline (500 µM) blocked both VZ and IR waves (Fig. 5). Adenosine receptor activation has been previously shown to regulate the dynamics of IR waves in ferret and mouse retinas (Singer et al. 2001; Stellwagen et al. 1999).
|
30 min did not block either VZ or IR waves. Application of thapsigargin evoked a large, slowly decaying increase in intracellular Ca2+ in ventricular cells (Fig. 3F), presumably due to blockade of uptake of Ca2+ into intracellular Ca2+ stores, but spontaneous VZ waves remained atop even though the decaying phase of the wave appeared prolonged to some degree. Thus Ca2+ release from thapsigargin-sensitive intracellular stores was not required for the generation of VZ waves in contrast to the Ca2+ transients found in ventricular cells of the developing neocortex (Owens et al. 2000). Furthermore, application of the gap junction blocker, 18-GA (75 µM) or octanol (150 µM), also effectively inhibited both VZ (Fig. 3G) and IR waves (Zhou et al. 2002). This result suggests a critical involvement of gap junctions in the formation of spontaneous IR and VZ waves, although it is also not clear whether the effect of gap junction blockers occurred mainly in the IR or VZ or both. Gap junction blockers have also been shown to block IR waves induced by calcium channel agonists in the mouse retina (Singer et al. 2001). A quantitative summary of the pharmacological effects on VZ waves is shown in Fig. 5. Calcium dynamics and spatiotemporal patterns of VZ waves
The spatial and temporal patterns of the VZ wave were characterized by Ca2+ imaging under a x10 objective lens. In most cases, the waves propagated from one region of the VZ surface to another with both the wavefront and the center of the wave propagating laterally. In some cases, however, the wave mainly expanded in size, showing clear propagation of the wavefront but little lateral displacement of the wave center. Although the shapes and trajectories of VZ waves varied from one wave to another, the overall wave patterns remained similar between E24 and P0. Figure 6 shows two examples of spontaneous ventricular waves. A common feature of these waves was the spatial boundary formed between successive waves. As shown in Fig. 6A, a clear boundary was formed between two consecutive waves: the first one (marked 1) moved into the frame from the left, and the second one (marked 2) propagated downward from the top of the frame, observing the boundary formed by wave 1. The boundaries between neighboring VZ waves closely resembled those observed in the ganglion cell layer of the developing rabbit retina (Zhou and Zhao 2000), which are thought to be a result of the refractory process following a spontaneous IR wave (Feller et al. 1997). While the propagation trajectories of most VZ waves were irregular, some waves displayed trajectories that were spatially symmetric. Figure 6B shows a VZ wave that propagated in a circular trajectory and stopped abruptly when the wavefront reached the point of wave initiation. This type of pattern was also observed in IR waves in rabbits (Zhou 2001a).
|
F/F) was also larger in VZ than that in IR. These results suggested different cellular mechanisms responsible for Ca2+ signaling during these two waves. Given that the temporal pattern of spontaneous Ca2+ transients may encode specific developmental cues (Gu and Spitzer 1995), the distinctively different Ca2+ dynamics between VZ and IR waves may be associated with functional differences that are yet to be understood.
|
Correlation between VZ and IR waves
To test whether VZ and IR waves were correlated in space and time, Ca2+ imaging of VZ and multielectrode array (MEA) recording of IR were made simultaneously on the same retina. MEA recording and Ca2+ imaging have been used previously in GCL of the ferret and mousse retina (Bansal et al. 2000; Demas et al. 2003; Feller et al. 1996; Meister et al. 1991; Wong et al. 1995), and both methods gave comparable measurements of the wave speed, frequency, and trajectory. To use the two methods simultaneously, isolated E29 and E30 retinas were loaded with Fura 2-AM and placed, GCL facing down, on a 60-channel MEA for multi-channel recording of spontaneous spikes in the GCL, while fluorescence signals from the ventricular surface of the same retinas were measured with Ca2+ imaging under an upright microscope. The position of each electrode in the MEA was mapped to a corresponding position in the VZ image by taking a bright-field image of the MEA under trans-illumination prior to fluorescence imaging (Fig. 8B).
Figure 8A shows an example of simultaneous recording of spikes propagating in the GCL and a Ca2+ wave in the VZ. The MEA is represented in the figure by a square, with small open circles indicating the positions of the 60 electrodes. Electrodes that detected spontaneous spikes are represented by large filled circles, which are coded with different shades of gray according to the spike rate detected (Fig. 8A). A time series of the MEA recording (Fig. 8A, top 2 rows) shows the spike activity as a function of time in 1-s intervals. Likewise, a time series of the difference fluorescence (F) images from the VZ shows the propagation of a VZ wave during the same time period (Fig. 8A, bottom 2 rows). In this example, both the GCL and the VZ wave initiated in the upper part of the frame and propagated downward before spreading to both the left and the right. Figure 8B shows the propagation trajectories of this pair of VZ and GCL waves, demonstrating a close spatial correlation between the two. Similar analysis was performed on all of the 32 VZ waves detected in this experiment. Each of the 32 VZ waves could be paired with a corresponding IR wave, resulting in 32 pairs of corresponding VZ and GCL waves from this experiment. Out of these 32 pairs, 24 pairs showed clearly identifiable propagation trajectories/directions. Remarkably, in each of the 24 individual pairs, the VZ wave always had a propagation direction/trajectory that was within 30° of the direction/trajectory of the simultaneous GCL wave, suggesting that not only was the initiation of the VZ and IR waves correlated, but the two waves continued to be linked during propagation. The remaining eight pairs of VZ and IR waves in this experiment were either localized waves or waves that occurred too close to the edge of the frame for the propagation direction to be determined accurately. Nonetheless, in each of these eight pairs the corresponding VZ and GCL waves always appeared within the same quadrant of the image frame. These results clearly demonstrated that VZ and GCL waves were closely correlated in space.
To compare the relative timing of VZ and GCL waves, the raw spike data and the spike rate recorded by an arbitrarily selected electrode (electrode 53 in Fig. 8B) were plotted together with the fluorescence signals (F/F) measured simultaneously at the corresponding VZ area (position 53 in Fig. 8B). Although the peak spike rate and the peak F/F were offset by a few seconds, the overall temporal correlation between the VZ and GCL activities was apparent (Fig. 8C). This correlation became more evident when the spike rate r(t) measured at electrode 53 and the fluorescence change F(t)/F(t) measured at position 53 were compared for the entire 2,400-s recording period (Fig. 9A, notice only 21 of the 32 VZ waves moved past position 53). Indeed, cross-correlation analysis of the two traces in Fig. 9A found a high degree of correlation between r(t) and F(t)/F(t), with the signal-to-noise ratio at the peak of the correlation equal to 20, which was highly significant (P < 0.0001). Similar results were obtained from an E30 retina (Fig. 9, B and D) with the signal-to-noise ratio at the peak of the correlation equal to 16, again extremely significant (P < 0.0001). Such a close correlation was only found between corresponding (matching) VZ and IR positions. When the spike rate at an electrode was compared with the VZ fluorescence at a distant location, many VZ and IR waves were no longer correlated between these positions (Fig. 9E) although correlation still remained (with a longer lag time) between large VZ and IR waves that propagated past both locations.
|
F(t)/F(t) seen in Fig. 8C. However, because of the intrinsic kinetic difference between spikes and Ca2+ responses (in general, spikes are expected to lead Ca2+ responses even in the same cell), it could not be concluded with certainty whether this lag time in fact represented a delay in the onset of VZ waves with respect to IR waves. However, it was clear from our data that VZ and IR waves were highly correlated in the temporal domain. This close temporal correlation, together with the demonstrated spatial correlation between VZ and IR waves, suggested that the two waves were initiated or regulated by a common source. It should be noted that a small number of GCL spikes in Fig. 9(A and B), predominantly with very lower spike rates, did not correspond to any VZ wave. Thus even though every VZ wave was correlated with a GCL burst, some small GCL bursts did not match any VZ waves, possibly because these small GCL bursts were too weak to generate a wave or to propagate to the VZ. We next used pharmacological tools to determine if VZ waves were initiated or regulated by the IR waves or vice versa. Application of atropine during simultaneous imaging and MEA recording completely blocked VZ waves but did not affect GCL waves in the same retina (Fig. 10). In contrast, every drug that was found to block IR waves also blocked VZ waves (Fig. 5). These results clearly demonstrate that the initiation of IR waves was independent of the presence of VZ waves. Taken together, our data suggest that retinal waves were initiated in IR and propagated to the VZ.
|
To demonstrate directly that waves of excitation were able to spread from IR to VZ, we elicited VZ waves by selectively exciting the IR. Because the VZ cells were insensitive to direct puffs of kainate (KA; Fig. 11B, n = 3), we used KA puffs at GCL to selectively excite IR cells (Fig. 11C). Although spontaneous IR waves were not mediated by glutamate at this age, KA could evoke robust excitation from cholinergic amacrine cells and ganglion cells (Lee and Zhou, unpublished observation) and induce wave-like IR responses, which were insensitive to atropine as expected (Fig. 11C). These responses propagated only a short distance, usually as a two-dimensional plane wave with a circular wavefront. As shown in Fig. 11A, KA puffs at the GCL surface evoked Ca2+ waves at the VZ surface. Like spontaneous VZ waves, which formed spatial boundaries with their neighbors (e.g., the 2 spontaneous VZ waves shown in Fig. 11A, a and b), puff-induced VZ waves also formed spatial boundaries with subsequent spontaneous waves (Fig. 11A, c and d), and they also caused a refractory period, during which a second puff could not induce a wave in the same area (Fig. 11A, e and f). Similarly, during the refractory period produced by a spontaneous VZ wave, a puff at GCL could not induce a VZ wave (Fig. 11A, g and h). Similar to the spontaneous VZ waves, KA-evoked VZ waves could also be completely blocked by atropine, indicating the involvement of muscarinic receptor activation in the generation of VZ waves. This result also confirmed that the fluorescence signals detected from the VZ were not out-of-focus signals from the IR. Because VZ waves were evoked by KA puffs at the IR, but not by KA puffs at the VZ, this experiment clearly demonstrated a retrograde signaling pathway by which excitation in the IR was propagated to the VZ. Similar results were obtained with puffs of the nicotinic agonist 1,1-dimethyl-4-phenylpiperzinium iodide (DMPP) at the GCL, which also produced atropine-sensitive, wave-like responses in VZ (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Our results demonstrated for the first time in the mammalian retina that spontaneous rhythmic Ca2+ transients existed in the VZ and propagated laterally as waves. We showed with simultaneous MEA and optical recordings that the VZ and IR waves were highly correlated in both space and time, indicating a common source of initiation or regulation of the two waves. Pharmacological comparisons between VZ and IR waves suggest that waves were initiated in IR and propagated to the VZ. Puffing exogenous agonists to GCL evoked VZ waves, demonstrating the capability of the developing mammalian retina to spread rhythmic, wave-like activity from IR to VZ.
The spontaneous VZ wave found in this study seemed similar to the rhythmic Ca2+ transients previously reported in E11 chick retina (Catsicas et al. 1998), although the dynamics and pharmacology of VZ transients in the chick retina are not available for comparison. Thus the VZ waves may represent a developmental phenomenon conserved in both lower vertebrates and mammals. Ventricular Ca2+ transients have also been reported in early (E5) developing chick retina and shown to play a role in regulating the cell cycle (Pearson et al. 2002). However, because these early transients in chick VZ appear mainly as small bursts in single cells and only occasionally spread to neighboring cells, it seems unlikely that they belong to the category of propagating retinal waves, which emerge in chick GCL only after E11 (Wong et al. 1998). Because the spatial and temporal dynamics of spontaneous Ca2+ transients, and not merely the transients per se, is believed to be developmentally important (Gu and Spitzer 1995), an understanding of the developmental function of retinal VZ transients would require the knowledge of the spatiotemporal pattern, mechanism of generation, and developmental sequence of these VZ events. We showed that the spontaneous VZ activity in E24-P0 rabbits was distinctive in several important aspects. First, the activity was correlated among a large population of VZ cells and propagated laterally as a wave with a well-defined wavefront, spatial boundary, and refractory period. Second, the VZ wave was closely correlated with the IR wave, yet it was distinct in dynamics, showing a much more prolonged Ca2+ kinetics. Third, the VZ wave was mediated by nicotinic and muscarinic receptor activation and gap junction communication; and it seemed to depend on Ca2+ entry but not intracellular Ca2+ release. These findings are important for future understanding of the role of VZ waves in retinal development. Because VZ waves were most prominent in embryonic rabbit retina, it is presently difficult to manipulate the waves in vivo for functional studies. Studies in an animal species, in which the VZ waves occur postnatally, may allow direct assessment of the developmental role of VZ waves in the mammalian retina. It also remains to be determined how VZ waves in rabbit progress beyond the period (E24-P0) examined in this study.
Retrograde signaling via retinal waves
Our simultaneous MEA recording and Ca2+-imaging results demonstrated that VZ and IR waves were closely correlated in both time and space domains, suggesting a common source of the initiation and regulation of the two waves. Previous results from chick retinal slices also show that the ventricular transients were nearly synchronized with the activities in the GCL. However, due to limited temporal resolution, it could not be distinguished in chick retinal slices whether the Ca2+ transients are propagated from the inner to the outer retina or vice versa (Catsicas et al. 1998). We now determined that spontaneous waves in rabbit are initiated in the IR and propagated to the VZ. This conclusion was based on three lines of evidence: VZ waves were always correlated with IR waves, but a small number of IR waves could exist without correlated VZ waves; atropine blocked VZ but not IR waves, whereas blocking IR waves always resulted in VZ waves being blocked as well; puffing KA or DMPP on IR resulted in a retrograde spread of wave-like activities to the VZ.
It remains to be determined how spontaneous activities spread from IR to the VZ. Our pharmacological results indicate a number of factors that might be involved, including muscarinic receptor activation, acetylcholinesterase activity, and gap junction communication. VZ cells directly responded to muscarine puffs and seemed to receive endogenous cholinergic input during VZ waves. One source of this cholinergic input may be the diffusion of ACh released by cholinergic amacrine cells in the IPL because both cholinergic immunoreactivity (Ahmad and Zhou 2003) and starburst cells (data not shown) have been found in the rabbit retina at this age. It is also possible that ACh is released near the VZ, by immature horizontal cells (Kim et al. 2000) because cholinergic markers (ChAT, VAChT) have been found transiently in the outer mammalian retina during a developmental period comparable to the appearance of VZ waves in rabbit (Ahmad and Zhou 2003; Kim et al. 1999, 2000). A common downstream effect of muscarinic receptor activation is the release of Ca2+ from intracellular stores. However, the apparent insensitivity of both IR and VZ waves to thapsigargin seems to contradict a critical involvement of such a pathway in VZ wave generation, suggesting different muscarinic effects. The exact signal transduction mechanism responsible for the muscarinic action on retinal waves remains to be elucidated.
The muscarinic interactions during VZ waves seemed to be regulated tightly by the AChE activity because neostigamine, which did not have a significant effect on IR waves, greatly enhanced or induced VZ waves (Fig. 3). The effects of AChE inhibitor on VZ waves are also consistent with an age-dependent expression of AChE activity previously reported in the VZ of the mammalian retina (Hutchins et al. 1995). Both muscarinic activation and AChE activity have been implicated in early developmental events, such as cell proliferation, differentiation, and neurite outgrowth (Cheon et al. 2001; Layer 1991; Ma et al. 2000; Pearson et al. 2002; Robitzki et al. 1997). Rhythmic Ca2+ transients have also been reported in the VZ of the developing neocortex (Owens and Kriegstein 1998; Owens et al. 2000), and muscarinic activation has been shown to induce slowly traveling waves of neural activity in developing neocortex (Peinado 2000).
In addition to cholinergic interactions, gap junctions also contributed significantly to the formation of VZ and IR waves (Fig. 3). Gap junction coupling has been observed not only among IR cells in the developing mammalian retina (Penn et al. 1994), but also between ganglion cells and cells that extend processes spanning the depth of the retina in E11 chick retina (Catsicas et al. 1998). It is possible that gap junctions play a role in the back propagation of retinal waves, although our study could not pinpoint the location of these critical gap junctions. Interestingly, it has been shown in developing ferret (Johnson et al. 1999, 2001) and rat (Reye et al. 2002) that immature photoreceptors send processes directly into IPL and GCL during a transient developmental period that matches the period of spontaneous VZ waves in rabbit. Thus it is possible that these immature photoreceptor processes and other radial processes, such as those formed by cells undergoing interkinetic migration, might form a part of the structural basis for the back propagation of IR waves to the VZ. It will be interesting to find out if any of these transient radial processes express cholinergic receptors or gap junctions. It also remains to be determined whether/how differentiating neurons in the outer retina participate in the VZ waves.
Although the exact functional role of VZ waves remains to be understood, our finding of the existence of VZ waves and the retrograde propagation of spontaneous retinal waves from the IR to the VZ suggests a novel pathway in the developing mammalian retina. This transient pathway runs opposite to the principal direction of information flow in the adult vertebrate retina and may allow newly differentiated neurons in the IR to regulate the activity of ventricular cells in the outer retina. Thus spontaneous retinal waves in the developing IR may send both forward messages to the central visual targets and retrograde signals to the VZ in the outer retina.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
GRANTS
This study was supported by grants to Z. J. Zhou from the National Eye Institute (RO1 EY-01894), Research to Prevent Blindness, and National Natural Science Foundation of China (30028005).
|
FOOTNOTES |
|---|
1 Supplementary material for this article (3 movie clips) is available online at http://jn.physiology.org/cgi/content/full/01129.2003/DC1. 
Address for reprint requests and other correspondence: Z. J. Zhou, Dept. of Physiology and Biophysics, Mail Slot 505, University of Arkansas for Medical Sciences, Little Rock, AR 72205 (E-mail: zhoujimmy{at}uams.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Ames A and Nesbett FB. In vitro retina as an experimental model of the central nervous system. J Neurochem 37: 867–877, 1981.[CrossRef][ISI][Medline]
Bansal A, Singer JH, Hwang BJ, Xu W, Beaudet A, and Feller MB. Mice lacking specific nicotinic acetylcholine receptor subunits exhibit dramatically altered spontaneous activity patterns and reveal a limited role for retinal waves in forming ON and OFF circuits in the inner retina [In process citation]. J Neurosci 20: 7672–7681, 2000.
Box G, Jenkins G, and Reinsel G. In: Time Series Analysis: Forecasting and Control (3rd ed.). Upper Saddle River, NJ: Prentice-Hall, 1994, p. 407–461.
Catsicas M, Bonness V, Becker D, and Mobbs P. Spontaneous Ca2+ transients and their transmission in the developing chick retina. Curr Biol 8: 283–286, 1998.[CrossRef][ISI][Medline]
Catsicas M and Mobbs P. Retinal development. Waves are swell. Curr Biol 5: 977–979, 1995.[CrossRef][ISI][Medline]
Cheon EW, Kuwata O, and Saito T. Muscarinic acetylcholine receptors in the normal, developing and regenerating newt retinas. Brain Res Dev Brain Res 127: 9–21, 2001.[CrossRef][Medline]
Copenhagen DR. Synaptic transmission in the retina. Curr Opin Neurobiol 1: 258–262, 1991.[CrossRef][Medline]
Demas J, Eglen SJ, and Wong RO. Developmental loss of synchronous spontaneous activity in the mouse retina is independent of visual experience. J Neurosci 23: 2851–2860, 2003.
Feller MB. Spontaneous correlated activity in developing neural circuits. Neuron 22: 653–656, 1999.[CrossRef][ISI][Medline]
Feller MB, Butts DA, Aaron HL, Rokhsar DS, and Shatz CJ. Dynamic processes shape spatiotemporal properties of retinal waves. Neuron 19: 293–306, 1997.[CrossRef][ISI][Medline]
Feller MB, Wellis DP, Stellwagen D, Werblin FS, and Shatz CJ. Requirement for cholinergic synaptic transmission in the propagation of spontaneous retinal waves. Science 272: 1182–1187, 1996.[Abstract]
Grynkiewicz G, Poenie M, and Tsien RY. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J Biol Chem 260: 3440–3450, 1985.
Gu X and Spitzer NC. Distinct aspects of neuronal differentiation encoded by frequency of spontaneous Ca2+ transients. Nature 375: 784–787, 1995.[CrossRef][Medline]
Hutchins JB, Bernanke JM, and Jefferson VE. Acetylcholinesterase in the developing ferret retina. Exp Eye Res 60: 113–125, 1995.[CrossRef][ISI][Medline]
Johnson PT, Raven MA, and Reese BE. Disruption of transient photoreceptor targeting within the inner plexiform layer following early ablation of cholinergic amacrine cells in the ferret. Vis Neurosci 18: 741–751, 2001.[CrossRef][ISI][Medline]
Johnson PT, Williams RR, Cusato K, and Reese BE. Rods and cones project to the inner plexiform layer during development. J Comp Neurol 414: 1–12, 1999.[CrossRef][ISI][Medline]
Katz LC and Shatz CJ. Synaptic activity and the construction of cortical circuits. Science 274: 1133–1138, 1996.
Kim IB, Lee EJ, Kim KY, Ju WK, Oh SJ, Joo CK, and Chun MH. Immunocytochemical localization of nitric oxide synthase in the mammalian retina. Neurosci Lett 267: 193–196, 1999.[CrossRef][ISI][Medline]
Kim IB, Lee EJ, Kim MK, Park DK, and Chun MH. Choline acetyltransferase-immunoreactive neurons in the developing rat retina. J Comp Neurol 427: 604–616, 2000.[CrossRef][ISI][Medline]
Layer PG. Cholinesterases during development of the avian nervous system. Cell Mol Neurobiol 11: 7–33, 1991.[CrossRef][ISI][Medline]
Ma W, Maric D, Li BS, Hu Q, Andreadis JD, Grant GM, Liu QY, Shaffer KM, Chang YH, Zhang L, Pancrazio JJ, Pant HC, Stenger DA, and Barker JL. Acetylcholine stimulates cortical precursor cell proliferation in vitro via muscarinic receptor activation and MAP kinase phosphorylation. Eur J Neurosci 12: 1227–1240, 2000.[CrossRef][ISI][Medline]
Meister M, Wong RO, Baylor DA, and Shatz CJ. Synchronous bursts of action potentials in ganglion cells of the developing mammalian retina. Science 252: 939–943, 1991.
O'Donovan MJ. The origin of spontaneous activity in developing networks of the vertebrate nervous system. Curr Opin Neurobiol 9: 94–104, 1999.[CrossRef][ISI][Medline]
Owens DF, Flint AC, Dammerman RS, and Kriegstein AR. Calcium dynamics of neocortical ventricular zone cells. Dev Neurosci 22: 25–33, 2000.[CrossRef][ISI][Medline]
Owens DF and Kriegstein AR. Patterns of intracellular calcium fluctuation in precursor cells of the neocortical ventricular zone. J Neurosci 18: 5374–5388, 1998.
Pearson R, Catsicas M, Becker D, and Mobbs P. Purinergic and muscarinic modulation of the cell cycle and calcium signaling in the chick retinal ventricular zone. J Neurosci 22: 7569–7579, 2002.
Peinado A. Traveling slow waves of neural activity: a novel form of network activity in developing neocortex. J Neurosci 20: RC54, 2000.
Penn AA, Riquelme PA, Feller MB, and Shatz CJ. Competition in retinogeniculate patterning driven by spontaneous activity. Science 279: 2108–2112, 1998.
Penn AA, Wong RO, and Shatz CJ. Neuronal coupling in the developing mammalian retina. J Neurosci 14: 3805–3815, 1994.[Abstract]
Reye P, Sullivan R, and Pow DV. Distribution of two splice variants of the glutamate transporter GLT-1 in the developing rat retina. J Comp Neurol 447: 323–330, 2002.[CrossRef][ISI][Medline]
Robinson SR. Development of the mammalian retina. In: Vision and visual dysfunction, edited by Cronly-Dillon JR. London, UK: Macmillan, 1991, p. 69–138.
Robitzki A, Mack A, Hoppe U, Chatonnet A, and Layer PG. Regulation of cholinesterase gene expression affects neuronal differentiation as revealed by transfection studies on reaggregating embryonic chicken retinal cells. Eur J Neurosci 9: 2394–2405, 1997.[CrossRef][ISI][Medline]
Singer JH, Mirotznik RR, and Feller MB. Potentiation of L-type calcium channels reveals nonsynaptic mechanisms that correlate spontaneous activity in the developing mammalian retina. J Neurosci 21: 8514–8522, 2001.
Stellwagen D, Shatz CJ, and Feller MB. Dynamics of retinal waves are controlled by cyclic AMP. Neuron 24: 673–685, 1999.[CrossRef][ISI][Medline]
Wong RO. Cholinergic regulation of [Ca2+]i during cell division and differentiation in the mammalian retina. J Neurosci 15: 2696–2706, 1995.[Abstract]
Wong RO. Retinal waves and visual system development. Annu Rev Neurosci 22: 29–47, 1999.[CrossRef][ISI][Medline]
Wong RO, Chernjavsky A, Smith SJ, and Shatz CJ. Early functional neural networks in the developing retina. Nature 374: 716–718, 1995.[CrossRef][Medline]
Wong WT, Sanes JR, and Wong ROL. Developmentally regulated spontaneous activity in the embryonic chick retina. J Neurosci 18: 8839–8852, 1998.
Zhou ZJ. A critical role of the strychnine-sensitive glycinergic system in spontaneous retinal waves of the developing rabbit. J Neurosci 21: 5158–5168, 2001a.
