A novel method for inducing nerve growth via modulation of host resting potential: gap junction-mediated and serotonergic signaling mechanisms

A major goal of regenerative medicine is to restore the function of damaged or missing organs through the implantation of bioengineered or donor-derived components. It is necessary to understand the signals and cues necessary for implanted structures to innervate the host, as organs devoid of neural connections provide little benefit to the patient. While developmental studies have identified neuronal pathfinding molecules required for proper patterning during embryogenesis, strategies to initiate innervation in structures transplanted at later times or alternate locations remain limited. Recent work has identified membrane resting potential of nerves as a key regulator of growth cone extension or arrest. Here, we identify a novel role of bioelectricity in the generation of axon guidance cues, showing that neurons read the electric topography of surrounding cells, and demonstrate these cues can be leveraged to initiate sensory organ transplant innervation. Grafts of fluorescently labeled embryological eye primordia were used to produce ectopic eyes in Xenopus laevis tadpoles. Depolarization of host tissues through anion channel activation or other means led to a striking hyperinnervation of the body by these ectopic eyes. A screen of possible transduction mechanisms identified serotonergic signaling to be essential for hyperinnervation to occur, and our molecular data suggest a possible model of bioelectrical control of the distribution of neurotransmitters that guides nerve growth. Together, these results identify the molecular components of bioelectrical signaling among cells that regulates axon guidance, and suggest novel biomedical and bioengineering strategies for triggering neuronal outgrowth using ion channel drugs already approved for human use.

The online version of this article (doi:10.1007/s13311-014-0317-7) contains supplementary material, which is available to authorized users.

The field of regenerative medicine has made remarkable progress in the wake of the molecular revolution, owing largely to cross-disciplinary approaches and growing convergence between developmental biology, neurobiology, bioengineering, and genetics. Already visible on the horizon are a number of exciting new treatments for a variety of human nervous system impairments, including paralysis, deafness, and blindness. Clinical trials are underway for a range of sensory augmentation devices, including retinal prosthetics and a new generation of cochlear implants [1–6], and recent findings include the generation of full vertebrate eyes in culture or in vivo [7–9].

The regeneration of damaged sensory structures such as eyes, or the implantation of bioengineered sensory (or even effector) components, requires afferent innervation of the host for information to be transmitted. In natural development, these connections are orchestrated by a suite of overlapping attractive and repulsive secreted signals that direct axon pathfinding to particular targets. In the case of the eye these signals include Sonic Hedgehog, Ephrin, Netrin, Semaphorin, and Slit family members, with various combinations being required for the exit of retinal ganglion cells from the base of the optic cup, the journey of the neurons from the eye to the brain, the crossing of the 2 optic nerves at the optic chiasma, and, finally, penetration of the visual center within the brain [10–14]. While the pathways involved in retinal ganglion pathfinding have been well documented in the context of early development, how eyes respond to these signals when they are spatially or temporally perturbed is not understood. Answering this question is essential for regenerative medicine approaches to augment, exchange, or replace sensory systems in an already developed organism, or to understand and repair birth defects affecting the visual system.

Insight may be gained from a number of studies in which researchers create morphologically complete ectopic eyes in vertebrates through the transplantation of developing eye primordia [15–19], misexpression of a master eye control genes, including Otx2, Pax6, Rx1, and Six3 [20–25], or establishment of eye-specific bioelectrical states in embryonic tissues [9]. Cellular tracking revealed that in some cases the retinal ganglion cells of ectopic eyes penetrated the host, growing toward the brain or spinal cord [16, 26], though the signals guiding the transplanted neurons have yet to be identified. Further, in at least one species, Xenopus laevis, it has been shown that animals can use these ectopic eyes in color learning assays, even when they are located far from the head of the individual [27–29]. The mechanism underlying this observation currently remains unclear, however, and a major goal is to determine what signals instruct ectopic eyes to innervate specific targets so they could be leveraged in molecular and regenerative medicine.

Bioelectricity is known to be a powerful instructive signal that cells and tissues use in a surprisingly diverse array of developmental and regenerative contexts [30–33], and large literature exists on the ability of extracellular electric fields, both endogenous and exogenous, to direct neuronal growth and orientation [34–48]. The ability of another aspect of cell biophysics—resting potential gradients—to dictate tissue identity and large-scale organization has begun to be exploited but has not yet been applied to the control of innervation in vivo [49, 50]. While the mechanistic studies of electric field effects on neurons have focused on the neuronal cells themselves, there have been no investigations of how V mem of the neurons environment may affect their outgrowth decisions.

Here, we asked whether and how the bioelectric topography of the host body could be leveraged to regulate the innervation emerging from ectopically transplanted organs. Our goals were to: 1) test the specific hypothesis that innervation from ectopic implants was sensitive to the resting potential states of cells in the microenvironment; 2) determine the molecular mechanisms by which this signaling occurs; and 3) establish a proof-of-principle of using ion channel drugs to modulate nerve growth after microsurgery. Using eye primordium grafts in Xenopus embryos, a model system that greatly facilitates insight into mechanisms of neurodevelopmental pathways [51], ectopic eyes were produced at caudal locations. Lineage tracers marking the donor tissue revealed that when host tissues’ V mem was depolarized, these ectopic eyes drastically hyperinnervated the host, extending axons through the majority of the fin and trunk of the developing tadpole. Functional experiments probing the downstream mechanism underlying the phenotype revealed serotonin (5-hydroxytryptamine, 5-HT) signaling (a charged neurotransmitter molecule known to be transported between cells along voltage gradients) to play an essential role in mediating the hyperinnervation effect. These results have a number of implications for both sensory and regenerative biology, revealing the mechanistic details of a novel growth control pathway for ectopic nerves that may underlie behavioral brain–body plasticity, and identifying a new endogenous role for embryonic 5-HT. The ability to control axon outgrowth and attraction through manipulation of membrane voltage suggests new approaches for guiding neurons across damaged areas, directing neurons to specific targets within the host, and promoting afferent innervation by bioengineered sensory implants.

Such approaches have the potential to increase the efficacy of treatments for a wide range of nervous system impairments, and to extend the applicability of existing electroceuticals [52, 53].

Xenopus laevis embryos were obtained through in vitro fertilization according to standard protocols [54]. Animals were reared in 0.1× Marcs Modified Ringer solution (MMR), pH 7.8, at 22 °C for the first 24 h (prior to surgery) and 16 °C thereafter. All animals were raised in a 12 h : 12 h light : dark cycle and were staged according to Nieuwkoop and Faber [55]. Animal density was a maximum of 30 individuals per 100 ×25 mm Petri dish, with media exchanges occurring every 3 days. For experiments involving high chloride media, 0.1× MMR was supplemented with additional NaCl to raise the chloride level from the 10.5 mM base level to a final concentration of 65 mM. This treatment has been previously been shown to raise the intracellular concentration of chloride following ivermectin exposure [56]. All experimental procedures involving Xenopus embryos were approved by the institutional animal care and use committees and Tufts University Department of Laboratory Animal Medicine under protocol M2011-70

Fluorescent donor animals were created through microinjection of capped synthetic tdTomato mRNA [57]. mRNA was synthesized using standard message machine kits (Life Technologies, Carlsbad, CA, USA) and injected into animals in 3 % Ficoll solution using a pulled capillary, targeting all 4 cells of a stage 3 embryo to maximize expression. Three hours after microinjection the embryos were washed in 0.1× MMR and returned to the incubator for 24 h, at which time they were sorted for expression of the fluorescent construct using an Olympus BX-61 microscope (Olympus, Tokyo, Japan) equipped with a Hamamatsu ORCA AG CCD camera (Hamamatsu, Shizuoka-ken, Japan). The hyperpolarizing channel Kir4.1 has been described previously [58], as has the dominant negative gap junction construct 3243H7 (H7) [59, 60].

Prior to microsurgery, the vitelline membranes of stage 23 tdTomato+ donors and wild-type (WT) recipients were removed and the embryos anesthetized in a 0.02 % tricaine solution, pH 7.5, in 0.1× MMR. Using surgical forceps, the presumptive eye primordium was excised from tdTomato+ donors with care taken to avoid any underlying or surrounding tissue. Using the same forceps, a small slit was created in the posterior tissue of the WT recipient, directly ventral to the neural tube. The donor tissue was then gently inserted into this wound and allowed to heal for 10 min. To avoid any confounding effects of laterality on transplants, only the left eye was harvested from the donor, and all recipients received grafts to their left sides. In addition, the transplant was positioned in its natural proximal/distal orientation, that is with the eye facing “outwards” from the body of the animal. Transplanted tissue generally fused with the host within minutes and animals were washed 2× in fresh 0.1× MMR before being returned to an incubator at 16 °C. To generate eyeless tadpoles, stage 34 animals were anesthetized for 10 min in 0.02 % tricaine solution. Surgical forceps were used to create a small opening in the epidermis overlying the eye, after which the entire structure was removed. Operated animals were kept anesthetized for 30 min before being washed twice with 0.1× MMR and returned to an incubator at 16 °C.

Visualization of host innervation was performed by whole mount immunohistochemistry [61]. Animals were fixed overnight in MEMFA [100 mM 3-(N-morpholino)propanesulfonic acid (pH 7.4), 2 mM ethylene glycol tetraacetic acid, 1 mM MgSO4, 3.7 % (v/v) formaldehyde], washed in phosphate buffered saline Tween-20 (PBST), and blocked for 1 h at room temperature with 10 % normal goat serum in PBST. Samples were then rocked overnight at 4 °C in monoclonal antiacetylated alpha-tubulin antibody (Sigma T7451; Sigma, St. Louis, MO, USA), diluted 1 : 500 in PBST +10 % goat serum. Following primary exposure, samples were washed 3 times for 15 min in PBST before a 60-min secondary incubation with AlexaFlour-555 conjugated secondary at 1 : 1000 diluted in PBST. Following secondary incubation, samples were washed 3 times for 15 min in PBST and imaged on an Olympus BX-61 microscope.

Animals were exposed to pharmaceutical compounds immediately following surgery, stage 24, through imaging at stage 46. The only exception was para-chlorophenylalanine, in which case the embryos were exposed directly after fertilization, throughout their entire development, and during surgical procedures. All pharmaceuticals were diluted from stock solutions in dimethyl sulfoxide and were refreshed every 3 days. For the indicated stages, animals were exposed in 0.1× MMR to altanserin, 10 μM; cyanopindolol, 50 μM; fluoxetine, 10 μM; ivermectin, 1 μM; lindane, 280 μM; metergoline, 3.3 μM; para-chlorophenylalanine, 12.5 mM; SB258719, 50 μM; 5-HT, 5 mM; tricaine mesylate, 318 μM; and tropisetron, 10 μM. Drug concentrations used in the present study were determined through toxicity screens and were applied at levels that did not result in lethality or observable developmental defects.

To image transplanted eye innervation of the host, animals were first anesthetized for 10 min in 0.02 % tricaine solution, pH 7.5, after which they were transferred individually to depression slides and photographed on an Olympus BX-61 microscope. Donor fluorescence was detected with a tetramethylrhodamine filter set (Olympus U-MRFPHQ), which was then overlaid with standard bright field images. Animals were imaged at both 40× and 100× magnifications, and raw widefield stacks were used to resolve the regions of interest.

Tissue concentrations of 5-HT were determined by high-performance liquid chromatography with fluorometric detection using a modification of a published methods for the determination of 5-HT in blood and brain [62, 63]. Briefly, animals (n =1–10 per tube) were sonicated (two or three 1-s pulses, medium power) in 100–500 μl of an ice-cold solution containing 1 % ascorbic acid and 10 ng/ml of the internal standard N-methylserotonin. Following perchloric acid deproteinization (10 % volume of 3.4 M HClO4), samples were centrifuged at 10,000 × g and the supernatants directly injected on the high-performance liquid chromatography system.

Statistical analyses were performed using Prism v. 5 (GraphPad Software, La Jolla, CA, USA). All data were collected as binomial outputs (presence or absence of a phenotype) and analyzed using nonparametric statistics. For comparisons between control and treated animals, Fisher’s exact tests were performed. In the cases where multiple comparisons were made within an experiment (including the analysis of 5-HT and gap junction blocker results) Bonferroni corrections were employed to correct for experiment-wide α level. To quantify the innervation in control and ivermectin-exposed animals, Sholl analysis was performed on raw widefield image stacks from each treatment using ImageJ v. 1.49 (NIH, Bethesda, MD, USA). After processing with a threshold algorithm, concentric circles with and increasing radius of 50 μm were drawn starting at the region of image closest to the ectopic eye and the axons crossing each radius counted using an automated function. Analysis began 200 μm from the center of the circle to reduce noise due to over exposure of fluorescence near the ectopic eye itself. Comparisons between control and treated animals were performed by two-way analysis of variance.

To begin to investigate the bioelectrical cues that guide innervation of ectopic organs, we established an assay in X. laevis: eye primordium transplants [27], which result in complete eyes developing at the site of the graft. Eyes could be transplanted rostrally or caudally, with success in >95 % of performed microsurgeries, and were shown through behavioral analysis to be functional [27, 29]. To visualize any innervation arising from donor tissue in the host, eye primordium was taken from animals ubiquitously expressing the fluorescent protein tdTomato. Fluorescence revealed that ectopic eyes created in the tails of Xenopus tadpoles often resulted in no innervation of the host, or in some cases small amounts of innervation in the tail or trunk of the animal (Fig. 1a–c).Fig. 1Ectopic eyes hyperinnervate hosts in response to chloride channel activation. (a) Diagram of eye primordium graft location (red eye) and location imaged for innervation analysis (yellow shaded area =40×; gray shaded area =100×). (b) Eye primordium removed from a donor tdTomato-expressing Xenopus and transplanted to caudal locations of a wild-type (WT) recipient produces ectopic eyes at the site of the graft (scale bar =500 μm). (c) Fluorescent labeling of donor tissue reveals that ectopic eyes occasionally innervate the host, with small numbers of axons observed in the fin and trunk of the tadpole (scale bar =100 μm). (d), In the presence of the glutamate-gated chloride channel activator ivermectin (IVM), ectopic eyes hyperinnervate the host, with labeled axons present throughout large portions of the fin and trunk of the recipient. (e) High magnification reveals extensive branching of donor axons in response to IVM.

(f) Sholl analysis comparing WT (n =15) and IVM-exposed (n =14) embryos reveals significantly more axons (P <0.01) in treated animals at both proximal and distal locations to the transplanted eye. (g) Time lapse imaging of ectopic eye innervation shows extensive axon remodeling in response to IVM, including extension, retraction, and degradation over 4 days. Loopback structures were observed in which axons appear to self-cross, after which the axon shows blebbing and degradation (white arrows; scale bar =100 μm). Error bars indicate ±1 SEM. Asterisks represent values which significantly differ from controls (2-way analysis of variance, followed by Bonferroni post hoc analysis)

To test the functional roles of polarization state of the host tissue in determining the quality of innervation, we took advantage of the chloride channel activator ivermectin (targeting glycine gated chloride channel, glutamate gated chloride channel, and gamma-aminobutyric acid (GABA) A receptor family members), which depolarizes cells expressing these channels through chloride efflux upon binding [56, 64]. Application of ivermectin after surgery had a dramatic effect on ectopic eye innervation, 37 % of treated animals displayed a drastic hyperinnervation phenotype (n =59; P <0.01) (Fig. 1d,e), with labeled donor axons present throughout the tailfin and trunk of affected tadpoles. The phenotype appeared to be biphasic, tadpoles showed either little/no innervation, or large amounts of axons were observed covering the entire ventral tailfin. Hyperinnervation was quantified using Sholl analysis: affected animals showed a significant increase in axons at both proximal and distal locations relative to untreated transplanted eyes (2-way analysis of variance; P <0.01) (Fig. 1f).

Time-lapse imaging revealed that neurons began to emerge from the graft site 2 days after transplantation at stage 38, and continued to grow through the remainder of the experiment. In addition to growth, neurons demonstrated extensive remodeling throughout the study. A number of individual “loopback” patterns were observed, where axons appeared to reverse direction and cross themselves (Fig. 1gi,ii). In these cases, neuronal blebbing was detected within 24 h of loopback, and within 2 days the entire neuron was degraded (Fig. 1giii,iv). In other cases axons were seen to retract before extending in a different direction. As noted in previous studies of glycine receptor chloride channel-expressing cell depolarization in Xenopus [56], the overall development of the host animals was completely normal, though late-stage animals were paralyzed owing to the silencing effect of depolarization on the host musculature.

We next asked whether the hyperinnervation was induced by depolarization of the host tissue or the donor tissue. To address this question, fluorescent donors were cultured in medium containing ivermectin from fertilization through the time the primordium was excised, after which the transplant tissue was washed in standard media before being grafted into hosts. Ivermectin is known to be an irreversible activator of chloride channels [65], and ivermectin-exposed tadpoles are paralyzed for days, confirming that treated tissue retains its bioelectrical state long after the drug is removed. Analysis of these transplants revealed that when only donor tissue was exposed to ivermectin, the resultant tadpoles did not display any significant differences in innervation compared with untreated embryos (n =32; P =0.55 (Fig. 2a). In some cases a small number of axons were observed exiting the ectopic eyes (Fig. 2ai) in both treated and control tadpoles, but animals largely showed no innervation following transplantation, indicating ivermectin exerts its effect not by altering the properties of the ectopic nerves themselves, but by changing the properties of their host environment. Taken together, these data reveal that manipulation of chloride flux can strongly and specifically increase innervation from ectopic organs.Fig. 2Ectopic eye hyperinnervation is location-specific and requires depolarization of the host. (a) Donor eye primordium treated with ivermectin (IVM) prior to implantation does not result in hyperinnervation of the host (eye is off frame to the left of the image). (ai) High magnification of donor eyes treated with IVM reveals a small number of eyes extend few projections into the host, but not in significant levels on hyperinnervation.

(b) When native tadpole eyes are extirpated and replaced with labeled donor eye tissue, optic nerves (white arrow) emerge from replacement structures and target the optic tecta of the host. (c) When native eyes are replaced with labeled donor eyes in the presence of IVM, hyperinnervation is not observed. Transplanted eyes send single optic nerves to the optic tecta of the host (white arrow). (d) Xenopus eyes can be surgically removed at tadpole stages, resulting in eyeless animals. (e) Ectopic eyes induced in eyeless animals do not hyperinnervate the host. WT = wild-type

Given that ectopic nerve growth was radically altered by ivermectin exposure, we next checked whether the host’s own normal innervation was similarly disrupted. To determine whether the observed phenotype was specific to ectopic nerve, or whether depolarization caused the same degree of hyperinnervation of the native nervous system, immunohistochemistry was performed against the neuronal marker acetylated tubulin. Culturing animals in ivermectin, regardless of whether they possessed ectopic eyes or not, did not result in any detectable large-scale neural patterning defects in the host animal (Supplementary Fig. 1). Remarkably, the native optic nerves, somite patterns, and lateral line all appeared normal and were certainly not affected to the extent observed in transplanted nerves, indicating the effect of ivermectin targets transplanted nerve but not the host’s normal neural anatomy.

Why are ectopic nerves hyperinnervating while the native innervation is unaffected by ivermectin? We first asked if this distinction was due to the surgery (the transplantation history of the ectopic eye tissue) or simply to its aberrant location within the host. To determine if ectopic eye hyperinnervation was location specific, the native eye of a developing tadpole was removed and replaced with labeled donor tissue. In this case, the surgery transplants an eye into its normal native location, and in the absence of ivermectin, donor eyes extend optic nerves to the optic tectum of the host in apparently the same manner as native eyes (Fig. 2b). These optic nerves appeared identical to those of WT animals, and obeyed the same lateralization rules, penetrating the brain hemisphere opposite to the eye location. When the same surgeries were performed in the presence of ivermectin the results were unchanged, transplanted eyes extended optic nerves to the midbrain of the host, and no hyperinnervation was observed (n =20; P =1.00 (Fig. 2c). These results confirmed that the effect of ivermectin on transplanted eyes was location specific: exposure of eyes in the posterior resulted in hyperinnervation of the host, while those in the anterior, normal eye field, did not.

An additional possibility was that the presence of native eyes suppressed innervation by ectopic structures, and that this suppression was simply alleviated by ivermectin exposure. To test this hypothesis, microsurgeries were performed on stage 30 tadpoles to remove their native eyes. This stage was chosen because time lapse revealed it was prior to the stage when ectopic eyes normally innervate the host. Earlier time periods were ruled out as Xenopus tadpoles regenerate eyes lost earlier in development. All microsurgeries performed were successful, producing eyeless tadpoles that did not contain any other detectable developmental defects compared with WT siblings (Fig. 2d). When ectopic eyes were added to eyeless tadpoles at posterior locations, in no cases was hyperinnervation observed (n =32; Fig. 2e). In the majority of cases no axons were seen exiting the ectopic eyes, and when labeled neurons were observed they were few in number and length. These results indicated that the presence of native eyes in a tadpole does not inhibit innervation by ectopic eyes.

Exposure to ivermectin depolarizes cells expressing glutamate/glycine-gated chloride channels by locking the targets in an open position, leading to Cl– loss from chloride-rich Xenopus cells [56]. To determine if the phenotype caused by ivermectin was a specific effect of chloride signaling (or even an off-target effect of the drug), or was a response to membrane depolarization per se (regardless of which ion’s conductance caused it), experiments were performed to examine ectopic eye innervation in response to voltage changes driven by chloride, potassium, and sodium channels.

The intracellular concentration of chloride ions in developing tadpoles can be as high as 60 mM, while that of standard media is 10 mM, leading to a cellular loss of chloride (and membrane depolarization) following ivermectin exposure. To reverse this effect and hyperpolarize GlyCl-expressing cells, a supplemented medium was created with a 65 mM chloride concentration; the same strategy was previously used to characterize the function of the glycine receptor-expressing cells during embryogenesis [56]. Animals were raised in this medium following transplantation (during the same time periods as drug exposure) and host innervation was visualized as in the previous experiments. Imaging revealed that chloride supplemented media significantly inhibited ivermectin induced hyperinnervation of the host [n =37 (P =0.05); 19 % hyperinnervation in high chloride treatment compared with 37 % with ivermectin only). Most animals showed no innervation, and those in which axons were observed showed a greatly reduced number and simpler branching pattern (Fig. 3a,ai) compared with ivermectin only-treated animals (e.g., Fig. 1d,e). No developmental defects were noted in animals raised in high chloride solutions compared with siblings in standard media, indicating the effect on ectopic eyes was not the result of any large-scale developmental defects. Given the ability of high extracellular chloride to inhibit the hyperinnervating effect of ivermectin precisely as predicted by the Goldman equation, we conclude that the effect is not due to an ion-independent function of ivermectin.Fig. 3Ectopic eye hyperinnervation is a result of membrane voltage depolarization, not restricted to chloride flux.

(a) While ectopic eyes hyperinnervate hosts in response to ivermectin (IVM), supplementing the media with high chloride levels (hyperpolarizing cells affected by the drug) inhibits the extent of innervation and branching pattern. (b) Injecting mRNA coding for the hyperpolarizing potassium channel Kir4.1 inhibits hyperinnervation of ectopic eyes in response to IVM exposure. (c) Treating animals with the sodium transport inhibitor tricaine mesylate, which hyperpolarizes affected cells, inhibits hyperinnervation of ectopic eyes in response to IVM exposure

We next asked whether the effect was chloride-specific by attempting to inhibit the ivermectin effect with a hyperpolarizing potassium channel. mRNA encoding the hyperpolarizing Kir4.1 potassium channel [56, 58, 66] was injected into both cells of host embryos at the 2-cell stage for ubiquitous expression. At stage 24, labeled eye primordia were transplanted into the Kir4.1-expressing donors, after which the animals were cultured in ivermectin. Visualization of donor tissue indicated a suppression of hyperinnervation, with most ectopic eyes completely failing to innervate the host (only 3 of 24 hyperinnervated; P <0.01) (Fig. 3b,bi). This result reveals that it is the voltage state of the host (not the donor tissue) that is crucial for the effect because introduction of the potassium channel mRNA into the host alone was sufficient. To further bolster the K+ and Cl– data, we next used a sodium-based method. The paralytic tricaine mesylate inhibits sodium currents in affected cells, thus causing hyperpolarization, and has been shown to reduce regeneration of central nervous system (CNS) structures in Xenopus tadpoles [67]. Culturing tadpoles in this compound in conjunction with ivermectin following transplantation revealed that, as with Kir4.1, innervation was reduced to the levels found in untreated tadpoles (1 of 34; P <0.01) (Fig. 3c, ci). Taken together, these 3 results support bioelectrical signaling as a key regulator of the hyperinnervation phenotype.

Thus, depolarization through ivermectin is convenient and sufficient to drive the hyperinnervation effect, while hyperpolarization through chloride level alteration, potassium channel misexpression, or sodium channel blockade all inhibit the effect of ivermectin. Together, these results provide evidence that hyperinnervation by ectopic eyes is not strictly dependent on chloride channel activation specifically, but rather is a consequence of membrane potential depolarization, which can be induced by a variety of ion movements.