|Home | About | Journals | Submit | Contact Us | Français|
The neurotransmitter serotonin is synthesized in the retina by one type of amacrine neuron but accumulates in bipolar neurons in many vertebrates. The mechanisms, functions and purpose underlying of serotonin in bipolar cells remain unknown. Here, we demonstrate that exogenous serotonin transiently accumulates in a distinct type of bipolar neuron. KCl-mediated depolarization causes the depletion of serotonin from amacrine neurons and, subsequently, serotonin is taken-up by bipolar neurons. The accumulation of endogenous or exogenous serotonin by bipolar neurons is blocked by selective reuptake inhibitors. Exogenous serotonin is specifically taken-up by bipolar neurons even when serotonin-synthesizing amacrine neurons are destroyed; excluding the possibility that serotonin diffuses through gap junctions from amacrine into bipolar neurons. Further, inhibition of monoamine oxidase (A) prevents the degradation of serotonin in bipolar neurons, suggesting that MAO(A) is present in these neurons. However, the vesicular monoamine transporter (VMAT2) is present only in amacrine cells suggesting that serotonin is not transported into synaptic vesicles and re-used as a transmitter in the bipolar neurons. We conclude that the serotonin-accumulating bipolar neurons perform glial functions in the retina by actively transporting and degrading serotonin that is synthesized in neighboring amacrine cells.
Our current understanding of the use of transmitters by neurons is based on the axiom that a neuron synthesizes the transmitters that are released at presynaptic terminals. Once released from a terminal, the fate of the transmitter is limited to four possibilities; (1) the transmitter binds and activates pre- and postsynaptic receptors, (2) the transmitter is degraded by enzymes within or surrounding the synapse, (3) the transmitter is taken-up by the neuron that released it (re-uptake), or (4) the transmitter is taken-up, and subsequently degraded, by glia with processes that flank the synapse. Here we describe a fifth possibility which involves synthesis of the transmitter in one cell type that subsequently is released and actively transported into a neighboring neuron where it is degraded.
Serotonin is an amino acid-derived transmitter that is synthesized by discrete types of neurons in the central nervous system. However, a few studies have suggested that serotonin can accumulate in neurons that do not synthesize it. Lebrand and colleagues have shown that during mouse embryonic development, thalamic neurons do not synthesize serotonin but transiently take-up exogenous serotonin through high affinity transporters located on thalamocortical axons and terminals, and that glutamatergic neurons may co-release serotonin as a ‘borrowed’ transmitter (Lebrand et al. 1996). Similarly, Upton and colleagues have shown that retinal ganglion cells take-up serotonin during embryonic and early postnatal development, even though these cells do not synthesize serotonin (Upton et al. 1999). In addition, Whitworth and colleagues have shown that serotonin transporters (SERTs) are transiently expressed in thalamocortical neurons during development (Whitworth et al. 2002). Taken together, these findings indicate that during neuronal development, serotonin is transiently transported into distinct types of neuronal cells that do not synthesize this transmitter.
Serotonin-immunoreactive bipolar neurons have been described in the retinas of several vertebrate species. In the skate retina, Schuette and Chappel have demonstrated that a subset of bipolar cells show increased serotonin-immunoreactivity in the presence of high potassium-Ringer in in vitro eye-cup preparations. Further, a SERT inhibitor blocks the uptake of exogenous serotonin by bipolar cells, but not by amacrine cells (Schuette and Chappell 1998). The same authors provide evidence that OFF bipolar neurons acquire serotonin from large amacrine neurons in Xenopus retina (Schutte 1994). Similarly, in the chicken retina, a population of bipolar neurons is weakly immunoreactive for serotonin during late-stages embryonic development (Rios et al. 1997). However, the roles and mechanisms of serotonin accumulation in retinal bipolar neurons remain unknown.
In the current study, we demonstrate that a distinct type of bipolar cell in the mature chicken retina actively transports serotonin that is injected into the eye or is synthesized and released by amacrine cells. We determine the morphological characteristics and immunohistochemical profile of the serotonin-accumulating bipolar cells. We also provide evidence that serotonin is not synthesized by the bipolar neurons, but is specifically taken-up and degraded in these cells. A distinct type of amacrine neuron is the solitary source of serotonin in the retina, whereas the accumulation of serotonin in bipolar neurons relies upon active transport.
The use of animals was in accordance with the guidelines established by the National Institutes of Health and the Ohio State University. Newly hatched leghorn chickens (Gallus gallus domesticus) were obtained from the Department of Animal Sciences at the Ohio State University and kept on a cycle of 12 hours light and 12 hours dark (lights on at 7:00 am). Chicks were housed in a stainless steel brooder at about 28°C and received water and Purinatm chick starter ad libitum.
Chickens were anesthetized via inhalation of 2.5% isoflurane in oxygen at a flow rate of 1.5 l/min. Injections were made using a 25 μl Hamilton syringe and a 26-gauge needle with a beveled, curved tip. Penetration of the needle was consistently made through the upper eyelid into the dorsal quadrant of the vitreous chamber. In all the experiments, 20 μl of the vehicle containing the test compound was injected into the experimental (right) eye, and 20 μl of the vehicle alone was injected into the control (left) eye. The vehicle was sterile saline containing bovine serum albumin, 50 μg/ml, as the carrier. Test compounds included serotonin (5-Hydroxytryptamine; 1 μg per dose; Sigma-Aldrich), NMDA (N-methyl-D-aspartate; 1000 nmol per dose; Sigma-Aldrich), Quis (200 nmol per dose; Sigma-Aldrich), KCl (potassium chloride; 3.5M per dose; Sigma-Aldrich), Zimelidine dihydrochloride (200 ng per dose; Sigma-Aldrich), 6-Nitroquipazine maleate (200 ng per dose; Sigma-Aldrich), Sertraline hydrochloride (90ng per dose; Sigma-Aldrich), Clorgyline (200 ng per dose; Sigma-Aldrich) and Pargyline (200 ng per dose; Sigma-Aldrich). Assuming a total volume of about 1ml for a P14 eye, the initial maximum concentration of the injected compounds are 4.7μM for serotonin, 1mM for NMDA, 0.2mM for Quis, 70mM for KCl, 0.512μM for Zimelidine dihydrochloride, 0.534μM for 6-Nitroquipazine maleate, 0.262μM for Sertraline hydrochloride, 0.648μM for Clorgyline and 1.022μM for Pargyline.
We used 4 different injection paradigms: (1) Paradigm A – on post-hatch day 14 (P14), the right eye received a single injection of serotonin and the left eye received vehicle. Retinas were harvested at various time points 10 mins-24 hrs later. (2) Paradigm B – on P14, the right eye received an injection of the test compound (NMDA or Quis) and the left eye received vehicle. On P14, both eyes received a single injection of serotonin. Retinas were harvested 2.5 hrs later. (3) Paradigm C – on P14, the right eye received a single injection of the test compound (KCl and/or SSRI, Clorgyline or Pargyline) and the left eye received vehicle. Retinas were harvested 0.5 hrs or 2 hrs later. (4) Paradigm D – on P14, the right eye received a single injection of the test compound + serotonin and the left eye received serotonin alone. Retinas were harvested 2 hrs later.
Tissues were fixed, sectioned, and sequential double-immunolabeled with primary antibodies raised in the same species was performed as described elsewhere (Fischer et al. 2007; Fischer et al. 2008) Working dilutions and sources of antibodies used in this study included; rabbit anti-serotonin polyclonal antibody (Courtesy Dr. G. Bishop originally developed by Dr. R. Ho, The Ohio State University), mouse anti-Lim3 used at 1:50 (67.4E12; Developmental Studies Hybridoma Bank (DSHB), University of Iowa), mouse anti-Islet1 used at 1:50 (402.D6; DSHB), rabbit anti-Prox1 used at 1:400 (Novus), mouse anti-PKC used at 1:100 (BD Pharmingen), rat anti-glycine used at 1:1000 (Dr. D. Pow; Univeristy of Newcastle), mouse anti-calbindin used at 1:150 (Swant), rabbit anti-calretinin used at 1:1000 (Swant), and rabbit anti-VMAT2 used at 1:1000 (Affinity Bio-Reagents). None of the observed labeling appeared to be due to secondary antibody or fluorophore because sections labeled with secondary antibodies alone were devoid of fluorescence. Secondary antibodies included goat-anti-rabbit-Alexa488, goat-anti-mouse-Alexa488/568, goat-anti-rat-IgG-Alexa488, goat-anti-mouse-IgG-Alexa568 (Invitrogen) diluted to 1:1000 in PBS plus 0.2% Triton X-100.
Retinas from 3 chicks at P7 were pooled and placed in 3.0 ml of TRI Reagent® (Sigma-Aldrich). Total RNA was isolated as described elsewhere (Fischer et al. 2008). The web-based program Primer3 from the Whitehead Institute for Biomedical Research (http://frodo.wi.mit.edu/) was used to design primers for PCR. Primer sequences were as follows: VMAT2 Forward 5′ ACG ATG AAG AGA GAG GCA AC 3′, VMAT2 Reverse 5′ CAC CTA TGG GAT AGG ACT GG 3′. Predicted product size was 881 base pairs. PCR reactions were performed by using standard protocols and an Eppendorf thermal cycler. 20 –mer T7 and T3 RNA polymerase initiation sites were added to the 5′ ends of the forward and reverse primers respectively in order to generate probes off the PCR product directly for in situ hybridization. PCR products were run on an agarose gel to verify the predicted product sizes and purified using the ChargeSwitch-Pro PCR clean-up kit (Invitrogen).
Standard procedures were used for in situ hybridization, as described elsewhere (Fischer and Reh 2002; Fischer et al. 2004). Digoxigenin-labeled riboprobes were generated from the purified PCR product synthesized by using a kit provided by Roche (Alameda, CA) and stored at -80°C until use. Postnatal (P14) eyes were dissected in RNase-free Hanks' balanced salt solution (HBSS), fixed overnight at 4°C in 4% PFA buffered in 0.1 M dibasic sodium phosphate (pH 7.4), and embedded in OCT compound. Cryosections were processed for in situ hybridization as described previously (Fischer and Reh 2002; Fischer et al. 2004).
Photomicrohistograms were obtained using a Leica DM5000B microscope equipped with epifluorescence and Leica DC500 digital camera. Confocal microscopy was done with a Zeiss LSM 510 at the Hunt-Curtis Imaging Facility at the Department of Neuroscience at the Ohio State University. Confocal stacks of images were obtained for 1 μm-thick optical sections by using a 20× objective (0.75 NA) and multi-track, narrow-pass emission filter settings to exclude the possibility of fluorescence bleeding across channels. Images were optimized for color, brightness, and contrast, and double-labeled ones were overlaid by using Adobe PhotoshopTM6.0. Cell counts were made from at least five different animals, and means and standard deviations calculated on data sets. To avoid the possibility of region-specific differences within the retina, cell counts were consistently made from the central region of retina for each data set.
Immunofluorescence was quantified by using ImagePro 6.2. Identical illumination, microscope and camera settings were used to obtain images for quantification. Areas (1000 × 150 pixels or 290 × 43.5 μm) were sampled from 5.4 MP digital images. These areas were randomly sampled over the INL where the nuclei of the bipolar and amacrine neurons were observed. Measurements were made for regions containing pixels with intensity values of 72 or greater (0 = black, 255 = saturated green); a threshold that included labeling in the bipolar or amacrine neurons. The total area was calculated for regions with pixel intensities >72. The average pixel intensity was calculated for all pixels within threshold regions. The density sum was calculated as the total of pixel values for all pixels within threshold regions. These calculations were determined for INL regions sampled from six different retinas for each experimental condition.
Serotonin is known to be synthesized and released by a subset of amacrine neurons in most vertebrate species (Pourcho 1996; Vaney 1986; Rios et al. 1997; Millar et al. 1988). Additionally, there are a few reports of accumulation of serotonin in bipolar neurons of the chicken retina during development, with the levels declining after hatching (Rios et al. 1997). Studies in lower vertebrate species such as frog (Schuette and Chappell 1998) and skate (Schutte 1994) have suggested that serotonin is transported into bipolar neurons. However, this notion has not been unambiguously tested. Using the post-hatch chicken eye as a model system, we tested whether serotonin is merely being scavenged by bipolar neurons and degraded, or is actively taken-up by transporters and subsequently used as a transmitter by the bipolar neurons.
Serotonin-immunoreactivity was observed at high levels in a small population of amacrine cells with large somata found near the middle of the INL and large dendritic fields in the IPL (Fig. 1a), consistent with previous reports (Zhu et al. 1992; (Schutte 1994). In addition, we observed weak serotonin-immunoreactivity in the somata of presumptive bipolar neurons located near the middle of the INL (Fig. 1a). The levels of serotonin-immunoreactivity were not different in the amacrine or bipolar cells in retinas obtained under day-light or night-time illumination (data not shown). For all of the following data, procedures were performed during the middle of the day. We tested whether intraocular delivery of serotonin would increase serotonin-immunoreactivity in bipolar cells. Accordingly, 1μg of serotonin was injected into the vitreous chamber of the eye and retinas were harvested at various time-points after injection (Fig. 1a-g). Serotonin-immunoreactivity was found to be higher within 10 mins of injection (Fig. 1a), peaked within 2.5 hrs (Fig. 1e), declined rapidly, and returned to baseline levels by 4.5 hrs (Fig. 1f) and beyond until 24hrs (Fig. 1g). It is expected that an injection of serotonin would result in transient accumulation and degradation as is evident when we plot the area and intensity of the relative serotonin-immunofluorescence within the distal INL (bipolar cell bodies and dendrites) across the different timepoints (Figs. 1h, e). The quantitative immunofluorescence supports this notion and verifies the utility of this method to measure relative levels of immunolabeling. These findings indicate that exogenous serotonin is transiently taken-up by a population of bipolar neurons and is cleared rapidly thereafter. There was an increase in serotonin-immunoreactivity in the small population of amacrine cells too, within 2.5 hrs of serotonin delivery (Figs. 1b-e). In addition, neurites in the inner plexiform layer (IPL), presumably from both the amacrine and the bipolar neurons, appeared to transiently accumulate serotonin following an intraocular injection (Figs. 1a-g).
Immunofluorescence was not measured in the terminals of bipolar neurons in the IPL because it was impossible to distinguish the axon terminals from the processes of amacrine neurons. Although it is likely that the uptake of serotonin is prevalent at the terminals of bipolar neurons, elevated levels of serotonin-immunofluorescence in the somata and dendrites of bipolar neurons occurred, presumably, following uptake at the axon terminals.
In order to characterize the type of bipolar cell that accumulates serotonin, we injected eyes with serotonin, harvested the retinas 2 hours later and labeled the retinal sections with antibodies to well-known markers of bipolar neurons - Lim3, Prox1, Islet1, PKC, calretinin, calbindin, and glycine, along with serotonin. Members of the Lim-domain family of transcriptional factors are known to be constitutively expressed by bipolar neurons in the retina (Edqvist et al. 2006; Fischer et al. 2008). We observed that a subset of Lim3-positive bipolar neurons accumulated serotonin; these cells had somata located near the middle of the INL, vitread to most other types of bipolar cells (Fig. 2a-d). All SAB cells (n=121) counted were positive for Lim3. The SAB cells were weakly positive for Prox1 (Fig. 2e-h). None of the SAB cells were positive for glycine, Islet1, calretinin, calbindin or PKC (Figs. 2i-n). We compared the morphology of SAB cells to the bipolar cell-types described by Cajal by using Golgi-impregnation (Fig 2o). The SAB cells have somata that were located in the center of the INL and bistratified axon terminals in both the distal and proximal IPL (Fig. 2o). Their dendritic terminals branched out and established endings adjacent to calbindin-positive photoreceptor pedicles in the distal OPL (Fig. 2l and n). The SAB cells do not form a Landolt's club (the enlarged apical dendrite of a bipolar cell) unlike most other types of bipolar neurons. However, we cannot exclude the possibility that serotonin fails to accumulate within the Landolt's club. The SAB cells are distinctly different from the types of bipolar cells described by Cajal in the avian retina (Fig. 2o).
It is possible that exogenous serotonin accumulates in bipolar neurons because of active transport and diffusion through gap junctions from the serotonergic amacrine cells to the bipolar cells. NMDA and Quis are known to destroy the serotonergic amacrine cells, in addition to many other types of amacrine neurons, whereas bipolar cells escape excitotoxic cell death. Consistent with a previous report (Fischer et al. 1998), we found that NMDA and Quis destroyed the vast majority of serotonin-producing amacrine neurons within 7 days of the treatment (Figs. 3b, c and f). Occasionally, we observed a residual serotonin-positive amacrine neuron with abnormal dendrites remaining in the damaged IPL (not shown). Bipolar neurons in damaged retinas had little or no serotonin-immunoreactivity (not shown). Thus, either the serotonin-immunoreactive bipolar neurons were destroyed by the excitotoxins, or these bipolar neurons survived but were no longer detected because their source of serotonin, the amacrine neurons, was ablated. To test these possibilities, we injected serotonin into NMDA and Quis-treated eyes. Intense serotonin-immunoreactivity was seen in bipolar neurons in NMDA and Quis-treated retinas (Figs. 3b and c), whereas few amacrine cells were detected (Fig. 3f). In the absence of serotonin-producing amacrine neurons, exogenous serotonin was taken-up by the bipolar neurons, suggesting that these cells express serotonin transporters and survive NMDA and Quis treatments. With the majority of serotonergic amacrine cells destroyed, we were able to distinguish the axon terminals of the bipolar neurons that stratify in both distal and proximal strata of the residual IPL.
KCl is known to cause membrane depolarization and neurotransmitter release from neurons. We hypothesized that neurons would release serotonin with KCl-mediated depolarization, hence decreasing serotonin-immunoreactivity. To test this hypothesis, we injected saline or KCl into eyes and harvested the retinas 2 hours later. Although we observed a significant decrease in serotonin-immunofluorescence in amacrine neurons, their levels were significantly enhanced in bipolar neurons (Figs. 4a-b). This finding suggests that KCl-mediated depolarization stimulates amacrine neurons to release serotonin, thereby depleting stores and reducing immunofluorescence. It is also possible that KCl-mediated depolarization reverses the direction of SERT in amacrine cells to deplete stores of serotonin. Subsequent to KCl-stimulated release from amacrine cells, serotonin is taken-up by nearby bipolar neurons. Alternatively, KCl causes an increase in the rate of transport of serotonin, or depolarization of inhibitory amacrine cells may feed back to inhibit the depolarization of bipolar neurons, thereby preventing the release of serotonin from bipolar neurons. The percentage change in the intensity of serotonin-immunofluorescence is shown in Fig 4c.
Accumulation of serotonin in bipolar neurons may occur through serotonin transporters. Accordingly, a specific SERT reuptake inhibitor, zimelidine, was injected with serotonin into the eye. Zimelidine significantly reduced the serotonin-immunofluorescence within bipolar neurons (Figs. 4d-g, j, k), while causing no change in serotonin-immunofluorescence in amacrine neurons (Figs. 4d-g, j, k). Similar results were seen with two additional SERT inhibitors, sertraline (Figs. 4h, i, l, m) and nitroquipazine maleate (data not shown). Since KCl induced an accumulation of serotonin in bipolar neurons, we sought to determine whether this accumulation was due to SERT-mediated uptake of the serotonin. KCl was injected with or without zimelidine and retinas were harvested 30 minutes later. Serotonin-immunofluorescence was significantly diminished in both amacrine and bipolar neurons treated with KCl and zimelidine (Figs. 4n-r). Our results are consistent with the hypothesis that the SAB cells normally take-up serotonin that is released from amacrine neurons.
Monoamine oxidases (MAO) catalyze the oxidative deamination of monoamines to inactivate and clear transmitters. MAO(A) specifically oxidizes serotonin, norepinephrine and epinephrine while MAO(B) oxidizes phenethylamine. We tested whether MAO inhibitors influence the endogenous levels of serotonin-immunofluorescence in retinal cells. Eyes were injected with MAO(A) inhibitor (clorgyline) or MAO(B) inhibitor (pargyline) and retinas were harvested 2 hours later. We observed that clorgyline (Fig 5a-b), but not pargyline (Fig 5c-d), increased the serotonin-immunofluorescence in bipolar neurons and amacrine neurons, suggesting that degradation via MAO(A) was inhibited. Further, neurites in the IPL were more intensely immunoreactive for serotonin in the presence of clorgyline (Fig 5b). The total area and intensity of serotonin-immunofluorescence in the somata of SAB cells in the distal INL was significantly increased by the MAO(A) inhibitor, but not with the MAO(B) inhibitor (Fig 5e-h). This finding suggests that the enzymes for serotonin degradation are present in the SAB cells.
To determine whether SAB cells utilize serotonin as a neurotransmitter, we labeled retinas for VMAT2, a protein that is required to load synaptic vesicles with serotonin. VMAT2 has been shown to be present in the IPL processes of the rat retina (Ostergaard et al. 2007). We found that VMAT2-immunoreactivity is present in the cell bodies and proximal neurites of serotonergic amacrine cells. None of the somata of the SAB cells were immunoreactive for VMAT2, indicating that these cells do not utilize serotonin as a neurotransmitter. In addition, VMAT2-immunolabeling was observed in the distal and proximal strata of the IPL (Figs. 6b, c, e, f, h). Most of the VMAT2-positive neurites in the IPL were concentrated in a distal stratum of the IPL (Figs. 6b and e). Additionally, VMAT2 is present in neurites in the middle and proximal strata of the IPL. Serotonin-positive processes co-localized with VMAT2 in both distal and proximal strata of the IPL (Figs. 6a, c, d and f). Further, VMAT2-positive / serotonin-negative processes were present in a distinct layer that sits directly above the distal serotonergic processes in the IPL and in the centre of the IPL (Fig. 6f). The serotonin-negative, VMAT2-positive neurites in the IPL were those of TH-positive dopaminergic amacrine cells (Figs. 6g-i). Consistent with the findings of immunolabeling studies, VMAT2 mRNA was shown to be expressed in the cell bodies of amacrine cells but not bipolar cells in the INL (Figs. 6j and k).
We report here that bipolar cells in the retina are capable of accumulating serotonin within neurites and cell bodies, even though they do not synthesize this neurotransmitter. Serotonin is likely to be synthesized by amacrine cells in the inner nuclear layer. Studies have shown that tryptophan hydroxylase 1 (TPH1), the serotonin biosynthetic enzyme, is expressed in the chicken retina (Chong et al. 1998; Iuvone et al. 1999). TPH1 is expressed predominantly in the photoreceptors of the outer nuclear layer and at low levels in presumptive amacrine cells of the inner nuclear and ganglion cell layers (Chong et al. 1998; Liang et al. 2004). Further, a population of amacrine cells, but not bipolar cells, has been reported to be immunoreactive for anti- phenylalanine hydroxylase using an antibody that also recognizes tryptophan hydroxylase (Haan et al. 1987; Cotton et al. 1988). Thus, serotonin has been proposed to be a neurotransmitter in amacrine cells.
Our findings are consistent with the notion that some types of bipolar cells accumulate serotonin in different vertebrate retinas (Witkovsky et al. 1984; Zhu et al. 1992; Gabriel et al. 1993; Schutte 1994; Zhu et al. 1995; Schlemermeyer and Chappell 1996). However, the role of serotonin accumulation in the SAB cells remains unclear. It is possible that one function of the SAB neurons is to degrade transmitter that is released from another type of neuron - a function that is normally ascribed to glial cells. Our data indicate that the SAB cells utilize SERT to take-up serotonin that is normally released from a distinct type of amacrine cell. We found that an MAO(A) inhibitor increased levels of serotonin in both amacrine and SAB cells during daylight conditions. This finding indicates that normally some of the serotonin that is transported into the SAB cells is degraded via MAO(A) and that the serotonin originates from amacrine cells. However, we cannot exclude the possibility that the MAO(A) inhibitor blocks serotonin degradation in amacrine cells or Müller glial cells thereby enhancing serotonin levels in bipolar cells. Although we can detect the mRNA from whole retinal extracts, it is unclear which cell types in the retina express MAO(A).
Subsequent to uptake, the SAB cells degrade some of the serotonin, performing a function that is expected of glia (such as astrocytes or Müller glia) but not of neurons. It is widely believed that monoamine transporters are expressed selectively on presynaptic cells that produce the specific monoamine (Gainetdinov and Caron 2003). The purpose of SERT expression on the pre-synaptic terminal is to remove the transmitter that escapes the cleft and allow for re-cycling of the transmitter for subsequent synaptic transmission. However, our data suggest that serotonin transporters can be expressed by neurons that do not re-use the transmitter for synaptic release. To the best of our knowledge, there are currently no reports of neurons in the mature nervous system that actively transport and degrade serotonin that is released from a nearby orthologous neuronal cell type.
A small population of amacrine cells is the only type of neuron in the retina that uses serotonin as a synaptic transmitter. Although, serotonin can be detected in both amacrine and bipolar neurons in the retina (current findings and references), VMAT2 is only detected in the serotonergic and dopaminergic amacrine cells. Although the bipolar cells accumulate serotonin, these cells do not express the transporter to load synaptic vesicles. These findings exclude the possibility that SAB cells utilize serotonin as a ‘borrowed’ neurotransmitter.
It is unlikely that the bipolar cells use VMAT1 to load synaptic vesicles with serotonin. VMAT1 has not been identified in Gallus gallus and has only been shown to be present in non-neuronal cells in other species including rat pinealocytes (Hayashi et al. 1999), adrenal medulla and intestinal endocrine cells (Hansson et al. 1998) (Weihe et al. 1994), human neuroendocrine tissue (Erickson et al. 1996), whereas VMAT2 is known to be present in neurons in the brain(Schutz et al. 1998) including the retina (Upton et al. 1999; Ostergaard et al. 2007). Differential studies on the expression of VMAT1 and VMAT2 have shown that VMAT2 is predominantly expressed in the central and peripheral nervous system whereas VMAT1 is only expressed outside the nervous system (Peter et al. 1995). Given the evidence in the literature, it can be assumed that (a) there may be no orthologue for VMAT1 in chicken and / or (b) VMAT1 is not present in the neurons of the chicken retina.
It is unlikely that the SAB cells convert serotonin to melatonin using the enzyme N-acetyltransferase (NAT). NAT is known to be expressed in photoreceptors (Tosini et al. 2006) (Niki et al. 1998). However its presence has not been documented in bipolar neurons. If the SAB cells do not express NAT, then it may be possible that some of the serotonin that the SAB cells scavenge from the amacrine cells is transported to the outer retina and is, subsequently, somehow transferred to photoreceptors that convert the serotonin into melatonin. The sources and the mechanisms by which serotonin reaches photoreceptors remain uncertain. It is well established that retinal melatonin is synthesized in a circadian manner by photoreceptors (Iuvone et al. 2005; Lundmark et al. 2007; Tosini et al. 2008). Further, reports have suggested that the expression and activity of tryptophan hydroxylase in the retina are regulated by circadian rhythms (Thomas and Iuvone 1991; Green and Besharse 1994; Chong et al. 1998; Liang et al. 2004) indicating that diurnal changes in metabolism influence serotonin synthesis. We assessed whether the accumulation of serotonin in the SAB cells was also dependent upon photoperiod but found no differences, indicating that the accumulation of serotonin in bipolar neurons may not be modulated by ambient light levels (data not shown). It remains possible that there are subtle diurnal changes in serotonin accumulation that were not detected by our methods.
The SAB cells have a unique morphology among previously described bipolar cells. Ramon y Cajal described seven morphologically distinct types of bipolar cells in the chicken retina (Cajal 1972). The SAB cells that we currently describe appear to be different from those described by Cajal with regard to the pattern and distribution of terminal arbors in the OPL and IPL. Further, the SAB cells do not form a Landolt's club, unlike 6 of the 7 types of bipolar cells described by Cajal. It is possible there may be bias against labeling the SAB cells via silver chromate impregnation. For example, rare cell types like bullwhip cells have gone unnoticed by this technique and were described only recently (Fischer et al. 2005; Fischer et al. 2006). Likewise, here we provide the first detailed description of the morphology and immunohistochemical profile of the SAB cells; loading with exogenous serotonin has allowed this. However, we cannot exclude the possibility that serotonin is excluded from the Landolt's club, which might otherwise indicate that the SAB could, indeed, be included among the bipolar types that were described by Cajal.
The serotonin-accumulating bipolar (SAB) cells represent a distinct type neuron that has not been well characterized previously. Under normal conditions, the SAB cells utilize SERT to accumulate serotonin that is synthesized and released from a minor population of amacrine cells in the retina. Some of the serotonin that is transported into the SAB cells is degraded via MAO(A). We conclude that the SAB cells perform functions normally ascribed to glial cells.
We thank Drs. Christophe Ribelayga, Howard Gu and Georgia Bishop for comments that contributed to the final form of this paper. Confocal microscopy was performed at the Hunt-Curtis Imaging Facility at the Department of Neuroscience of The Ohio State University. The Lim3 and Islet1 antibodies developed by Dr T. Jessell and were obtained from the Developmental Studies Hybridoma Bank developed under auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. This work was supported by a grant (EY016043-03) from the National Institutes of Health.