EAAT5 in the Rod Bipolar Cell and Oscillatory Potentials in Diabetes

Washington University in St. Louis Washington University Open Scholarship

Arts & Sciences Electronic Theses and Dissertations Arts & Sciences

Spring 5-15-2020

Electrophysiology of the Inner Retina in Health and Disease: EAAT5 in the Rod Bipolar Cell and Oscillatory Potentials in Diabetes

Gregory William Bligard Washington University in St. Louis

Follow this and additional works at: https://openscholarship.wustl.edu/art_sci_etds

Part of the Neuroscience and Neurobiology Commons, and the Ophthalmology Commons

Recommended Citation Bligard, Gregory William, "Electrophysiology of the Inner Retina in Health and Disease: EAAT5 in the Rod Bipolar Cell and Oscillatory Potentials in Diabetes" (2020). Arts & Sciences Electronic Theses and Dissertations. 2168. https://openscholarship.wustl.edu/art_sci_etds/2168

This Dissertation is brought to you for free and open access by the Arts & Sciences at Washington University Open Scholarship. It has been accepted for inclusion in Arts & Sciences Electronic Theses and Dissertations by an authorized administrator of Washington University Open Scholarship. For more information, please contact [email protected]. WASHINGTON UNIVERSITY IN ST. LOUIS

Division of Biology and Biomedical Sciences Neurosciences

Dissertation Examination Committee: Peter D. Lukasiewicz, Chair Erik D. Herzog Timothy E. Holy Daniel Kerschensteiner Steven J. Mennerick

Electrophysiology of the Inner Retina in Health and Disease: EAAT5 in the Rod Bipolar Cell and Oscillatory Potentials in Diabetes by Gregory W. Bligard

A dissertation presented to The Graduate School of Washington University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

May 2020 St. Louis, Missouri

Table of Contents List of Figures ...... iv Acknowledgments ...... vi Abstract ...... ix Chapter 1: Background and Introduction ...... 1 1.1 The Neural Circuitry of the Inner Retina ...... 1 1.2 Excitatory Amino Acid Transporters ...... 4 1.3 Electroretinography in Diabetic Retinopathy ...... 6 1.4 Objectives of the Dissertation ...... 7 Chapter 2: The Glutamate Transporter EAAT5 Mediates Rod Bipolar Cell Autoinhibition ...... 9 2.1 Background ...... 9 2.2 Materials and Methods ...... 12 2.2.1 Animal Protocols ...... 12 2.2.2 Retinal Dissection ...... 13 2.2.3 Whole-Cell Recordings ...... 13 2.2.4 Morphological Identification ...... 14 2.2.5 Solutions and Drugs ...... 14 2.2.6 Light Stimulation ...... 16 2.2.7 Data Analysis ...... 16 2.2.8 Modeling ...... 16 2.3 Results ...... 17 2.3.1 The Mode of Retinal Presynaptic Inhibition Does Not Switch with Light Intensity...... 17 2.3.2 GABA and Glycine Mediate Alloinhibition onto Rod Bipolar Cells...... 20 2.3.3 EAAT5 Mediates Autoinhibition onto Rod Bipolar Cells...... 24 2.3.4 EAAT5 Inhibition Modulates Signaling to AII Amacrine Cells...... 31 2.4 Discussion ...... 35 Chapter 3: Functional Deficits Precede Structural Lesions in Diabetic Retinopathy ...... 40 3.1 Background ...... 40 3.2 Materials and Methods ...... 45 3.2.1 Animals ...... 45 3.2.2 Electroretinography ...... 46 ii

3.2.2 Statistical Analysis ...... 46 3.3 Results ...... 47 3.4 Discussion ...... 56 Chapter 4: Summary and Future Directions ...... 59 4.1 EAAT5 ...... 59 4.2 Diabetic Retinopathy ...... 61 References ...... 64

iii

List of Figures Figure 1.1: The retina is a well-organized, laminar structure composed of five distinct classes of neurons...... 1

Figure 1.2: Rod pathways in the mouse retina utilize cone circuitry...... 3

Figure 1.3: EAATs are both glutamate transporters and chloride channels...... 5

Figure 2.1: Conventional inhibition is mediated by a variety of sources...... 10

Figure 2.2: EAAT5-mediated presynaptic inhibition is gated by glutamate and has two potential sources ...... 12

Figure 2.3: Differential interference contrast photomicrograph of a pipette and rod bipolar cell (RBC) ...... 14

Figure 2.4: Example RBC IPSCs...... 18

Figure 2.5: The mode of light-evoked inhibition to rod bipolar cells does not depend on light intensity...... 19

Figure 2.6: Inhibition onto cone bipolar cells is exclusively conventional...... 20

Figure 2.7: Cartoon depicting intracellular and puff application of receptor antagonists...... 21

Figure 2.8: Puff application of GABA/glycine blockers successfully blocks the light-evoked IPSC...... 22

Figure 2.9: Results of intracellular and puff application of receptor antagonists...... 23

Figure 2.10: Intracellular application of TBOA blocks the response to glutamate...... 24

Figure 2.11: Voltage pulse input...... 25

Figure 2.12: Voltage pulse example traces with and without GABA/gly blockers...... 26

Figure 2.13: Voltage pulse example traces with and without TBOA...... 27

Figure 2.14: EAAT5 is the primary mediator of autoinhibition in RBCs...... 28

Figure 2.15: EAAT5-mediated autoinhibition is dependent on synaptic activation...... 29

Figure 2.16: EAAT5-mediated autoinhibition is dependent on glutamate release...... 30

Figure 2.17: EAAT5-mediated autoinhibition is dependent on synaptic activation and glutamate release: Summary ...... 31

Figure 2.18: Outputs of in-silico model of the RBC – AII amacrine synapse...... 33

Figure 2.19: The synaptic sensitivity of the RBC synapse with and without EAAT5...... 35

Figure 3.1: High-fat diet fed mice develop obesity and glucose intolerance...... 43

Figure 3.2: At an advanced age, high-fat diet fed mice develop typical lesions of diabetic retinopathy...... 44

Figure 3.3: High-fat diet feeding promotes retinal vascular permeability...... 45

Figure 3.4: ERG analysis parameters...... 47

Figure 3.5: High-fat diet feeding does not alter major retinal responses to light stimuli...... 49

Figure 3.6: Three-month high-fat diet fed mice do not have decreased OP amplitudes...... 50

Figure 3.7: Six-month high-fat diet fed mice do not have decreased OP amplitudes...... 51

Figure 3.8: Twelve-month high-fat diet fed mice have decreased OP amplitudes...... 52

Figure 3.9: Twelve-month high-fat diet mice have decreased OP amplitudes: Example traces. .. 53

Figure 3.10: Delays in OP implicit times are seen in high-fat diet animals before development of vascular lesions...... 55

Acknowledgments This has been a long journey, and I would never have gotten here without the help of a remarkable number of people along the way. There are far too many to name them all, but I would like to acknowledge at least a few of them here.

First and foremost, I obviously need to thank Peter for making all of this possible. Peter is a caring, supportive mentor and a pleasure to work for. He has a wealth of knowledge not only of the intricacies of his field, but also of the intricacies of navigating a career in the sciences, and he somehow makes the time to be a connoisseur of both gourmet food and professional theater as well. I will always be in awe of his effortless ability to network at scientific conferences. I’m grateful for the personal and professional training that I’ve received from Peter, and I will do my best to live up to his example.

Of course, nothing in the lab would ever be possible without the tireless efforts of Jim DeBrecht.

From mice to equipment to reagents and solutions to software, there isn’t a single thing in the lab that would ever be in the right place or running properly without his hard work. Jim taught me the lab’s dissection technique, taught me how to keep track of my tools when working with night-vision goggles, and taught me how to code in three different programs. Whenever I needed something built or repaired, Jim always knew the best way to do it, and without his advice I’d probably still be fumbling around trying to fix something on my rig right now. I can’t properly repay Jim for all of his help, but hopefully explaining to him that tournament teams with small numbers next to their names are the better ones helps at least a little.

I’ve also had the pleasure of several other wonderful lab-mates who have come and gone. Rob

Purgert, Lateece Griffin, and Christine Emnett have all been wonderful people to work with. I

vi am particularly indebted to Rob; while he graduated and moved on not long after I arrived, he was extremely generous with his help and advice, going so far as to compile a massive manual of step-by-step guides he gifted to the rest of us after his defense. Everything I know about the mystical secrets to getting a patch on a healthy cell I learned from talking to Rob, and his diligence in thorough and accurate note-keeping and documentation set a standard that I have completely failed to live up to. (Although I’m not sure how any mere mortal could.)

I am extremely grateful to Dr. Rithwick Rajagopal for reaching out to me to do ERGs on the diabetic retinopathy project. Rithwick, along with Sheng Zhang, Li Yin, and Clay Semenkovich, did all of the work characterizing the metabolic course of the high-fat diet mouse model.

Rithwick’s enthusiasm for translational science is infectious, and he is a pleasure to collaborate with. I am also very thankful for the assistance of Dr. Robert Smith at the University of

Pennsylvania for sharing the use of his NeuronC modeling program and for sharing his expertise in modeling to answer my questions about how to adapt his simulations to my EAAT5 project. I would also like to thank my thesis committee, Steve Mennerick, Tim Holy, Daniel

Kerschensteiner, and Erik Herzog, for all of their help and advice and for always providing helpful perspective on my work. Outside of my project, Dr. Rajendra Apte has been an additional mentor, and I would like to thank him for taking the time to provide career advice.

I also greatly appreciate the help of the entire MSTP administration, especially Wayne

Yokoyama, Brian Sullivan, Liz Bayer, Christy Durbin, and Linda Perniciaro, who are all so adept at helping students navigate the difficulties of bouncing between the medical and graduate programs. Without them none of this would have happened.

vii

I am also grateful for the financial support that I have received for this work, which was funded by the NIH National Eye Institute training grant 5T32EY013360.

Finally, I wouldn’t be the person that I am without my family. My parents have always encouraged me and been positive role models, and without their support and upbringing I would never have made it this far. And, of course, I need to thank my beautiful and brilliant wife,

Kathy. She is a constant inspiration and supports me in more ways than I can count. She makes me a better person, both by her example and by always encouraging me to be the best version of myself I possibly can. I’m glad I came to St. Louis for many reasons, but the most important is that I was lucky enough to meet Kathy.

Greg Bligard

Washington University in St. Louis

May 2020

viii

ABSTRACT OF THE DISSERTATION

Electrophysiology of the Inner Retina in Health and Disease:

EAAT5 in the Rod Bipolar Cell and Oscillatory Potentials in Diabetes

Gregory W. Bligard

Doctor of Philosophy in Biology and Biomedical Sciences

Neurosciences

Washington University in St. Louis, 2020

Professor Peter D. Lukasiewicz, Chair

The inner retina is a highly tractable location for the study of neural networks in the central nervous system and improving our understanding of the inner retina is important for the treatment of a number of ophthalmic diseases. This dissertation aims to improve understanding of the inner retina in two parts: first, a basic neuroscientific project concerning normal neuronal processing in the inner retina at a cellular level, and second, a translational ophthalmologic project regarding the development in the inner retina of a common disease.

In mammals, modulation of the amplitude of dim visual signals primarily occurs at axon terminals of rod bipolar cells. While such effects are largely mediated by conventional inhibitory neurotransmitter receptors, excitatory amino acid transporter 5 (EAAT5) also contributes a significant unconventional inhibitory activity. EAATs clear glutamate from the synapse, but they also have an intrinsic glutamate-gated chloride conductance. EAAT5 in particular has very little glutamate reuptake activity, and primarily acts as an inhibitory glutamate-gated chloride channel.

In my primary project I demonstrate the relative contribution of EAAT5-mediated inhibition onto rod bipolar cells in response to physiologically relevant stimuli in retinal slice explants. In

ix this setting, I also compare the roles of autoinhibition, in which excitation of the cell leads to inhibition onto itself, and alloinhibition, which is driven by other bipolar cells. Contrary to prior reports, I found that EAAT-mediated inhibition can be elicited by dim light stimuli. I also found that while conventional inhibitors mediate nearly all alloinhibition onto rod bipolar cells, EAAT5 is the primary mediator of rod bipolar autoinhibition. Finally, in silico modeling demonstrates that EAAT5-mediated autoinhibition may have a substantial impact on postsynaptic excitation of the downstream AII amacrine cell. My combined results suggest that EAAT5 autoinhibition in rod bipolar cells modulates the rod signaling pathway, possibly tuned to prevent long-term changes in baseline depolarization while still allowing single-photon signals to pass in the dark and transient signals to pass in the twilight.

Obesity predisposes humans to type 2 diabetes, which may result in diabetic retinopathy, commonly understood to be fundamentally a vascular pathology. In my secondary project I seek to characterize the course of electroretinographic dysfunction in a high-fat diet–induced model of diabetes in mice. Mice weaned to chow or to a high-fat diet were tested for electroretinographic dysfunction at 3, 6 and 12 months of age. There is no evidence of microvascular disease in these mice until 12 months. However, at as early as 6 months, high fat–fed mice demonstrated increased latencies and reduced amplitudes of oscillatory potentials compared with controls.

These electroretinographic abnormalities were correlated with glucose intolerance. The results suggest that retinal disease in the diabetic milieu may progress through neuro-retinal stages long before the development of vascular lesions representing the classic hallmark of diabetic retinopathy, establishing a model for assessing novel interventions to treat eye disease.

Chapter 1: Background and Introduction 1.1 The Neural Circuitry of the Inner Retina The retina has long served as an experimentally tractable location for the study of the central nervous system, due in part to its physical accessibility, its relatively simple circuitry compared to the brain, and its precise, layered organization (Figure 1.1) [1]–[3]. Nevertheless, the vertebrate retina is still a complex neural network, and significant neuronal processing and filtering of visual information takes place in the retina itself before any visual signal even reaches the optic nerve [4]–[12].

Figure 1.1: The retina is a well-organized, laminar structure composed of five distinct classes of neurons. (Figure from [1].) Photoreceptors (PR) are hyperpolarized by light and synapse onto bipolar cells (BC) and horizontal cells (HC) in the outer plexiform layer (OPL). Bipolar cells then synapse onto ganglion cells (GC) and amacrine cells (AC) in the inner plexiform layer (IPL).

Visual signaling begins with rod and cone photoreceptors, which continuously depolarize in darkness and release glutamate [13], [14]. Following the absorption of a photon, photoreceptor

1 opsins, acting via a G-protein-coupled receptor (GPCR) signal cascade, can close membrane channels, hyperpolarize the cell, and decrease glutamate release [15]–[17]. Photoreceptors transmit this information in parallel onto bipolar cells at synapses in the outer plexiform layer

(OPL) [13]. Cone bipolar cells come in a number of varieties [18], but two main classes: in response to glutamate, OFF cone bipolar cells depolarize via ionotropic AMPA or kainate receptors [19], whereas ON cone bipolar cells hyperpolarize via sign-inverting metabotropic mGluR6 receptors [20]. Visual information passes from cone bipolar cells to ganglion cells via glutamatergic neurotransmission in the inner plexiform layer (IPL) [21]–[23], and from there, ganglion cell axons form the optic nerve, which passes visual information to the brain [24], [25].

Meanwhile, rods only synapse onto a single type of rod bipolar cell, which also receives its input via the mGluR6 metabotropic receptor [26]. Rod bipolar cells, in turn, synapse in the IPL onto

AII amacrine cells [27]–[30]. Rod and cone pathways converge where AII amacrines make gap junctions with ON cone bipolar cells and glycine-mediated inhibitory synapses onto OFF cone bipolar cells (Figure 1.2) [31]–[33]. Additional lateral communication and signal modulation between these pathways takes place through horizontal cells at the OPL and many forms of amacrine cells in the IPL [34]–[40].

Figure 1.2: Rod pathways in the mouse retina utilize cone circuitry. (Figure from [33].) Rods synapse onto a single class of rod bipolar cell, which synapses onto the AII amacrine cell. The AII generates parallel streams of ON and OFF by forming excitatory electrical synapses with ON cone bipolar cells and inhibitory glycine synapses with OFF cone bipolar cells, respectively.

Astoundingly, this neural network is capable of interpreting visual stimuli presented over more than 12 orders of magnitude of ambient luminance [41], [42], a feat that is only possible with sophisticated neuronal gain control systems. Such flexibility is achieved in part by a transition from rod-mediated vision in darkness to cone-mediated vision in bright light [43], [44], but rod vision alone still operates over a luminous range that far exceeds the dynamic range that ganglion cells would be capable of encoding without mechanisms for signal gain control. Gain modulation is attributable both to adaptation within photoreceptors themselves as well as to adaptation of the neural network downstream of the photoreceptors [45]–[50]. Our understanding

3 of primary photoreceptor adaptation has become quite sophisticated, but the mechanisms behind retinal network adaptation are less well understood [51].

1.2 Excitatory Amino Acid Transporters Ion channel gating and neurotransmitter transport are mechanistically distinct processes that are typically performed by separate proteins. Ion channels are pores that allow the diffusion of ions along preexisting electrochemical gradients, while neurotransmitter transporters actively pump a substrate against its electrochemical gradient [52]. There are, however, a few novel neurotransmitter transporters that can also act as ion channels, including known transporters for dopamine, GABA, glutamate, norepinephrine, proline, and serotonin [53]–[61].

The excitatory amino acid transporters (EAATs) are one such family of neurotransmitter transport proteins. EAATs are members of the solute carrier family 1 (SLC1), which also includes two neutral amino-acid transporters, ASCT1 and ASCT2 [62], and are distinct from vesicular glutamate transporters. The EAAT family is best known for those members that act to terminate signaling in the CNS by removing released glutamate from synapses following neurotransmission. However, in certain members of the EAAT family, glutamate binding can instead open a chloride conductance through the transporter that suppresses synaptic activity, although its role in CNS signaling is poorly understood [63], [64].

Five mammalian EAATs have been discovered and are designated EAAT1 through EAAT5

[64]–[66]. EAAT1, also known as SLC1A3 and GLAST1 (glutamate aspartate transporter 1)

[67], is present in glial cells throughout the central nervous system and at high levels in

Bergmann glia of the cerebellum. EAAT2, also known as SLC1A2 and GLT1 (glutamate transporter 1) [68], is almost exclusively glial, and is widespread and highly abundant throughout

4 the central nervous system. EAAT3, also known as SLC1A1 and EAAC1 (excitatory amino acid carrier 1) [69], is present in neurons throughout the CNS. EAAT4 (SLC1A6) is found in cerebellar Purkinje cells and in the forebrain [62], [70]–[72]. EAAT5 (SLC1A7) is localized to rod photoreceptors and rod bipolar cells in the retina [65], [73].

Figure 1.3: EAATs are both glutamate transporters and chloride channels. (Figure from [74].) A schematic diagram of an EAAT is shown. Glutamate is co-transported with three Na+ and one H+, along with the counter-transport of one K+. Glutamate and Na+ also activate a stoichiometrically uncoupled chloride conductance.

The two functions of these proteins are not thermodynamically coupled. The rate of glutamate uptake is independent of chloride concentration, instead dependent on the cotransport of three sodium ions and one proton and the counter-transport of one potassium ion (Figure 1.3) [75].

Meanwhile the transporter’s chloride conductance is dependent on glutamate and sodium binding to the transporter, as well as membrane potential, but is thermodynamically independent of the rate or direction of glutamate transport (27-31) [76]–[81]. In addition, the balance between these two functions varies enormously between individual members of the EAAT family [82], [83].

EAAT1, EAAT2, and EAAT3 have little chloride conductance and play an important role in glutamate reuptake in the cells that express them [84], [85]. EAAT4 and EAAT5 have little

5 glutamate uptake activity, but instead play a novel role as a presynaptic inhibitory signaling mechanism in the cerebellum and retina, respectively [64], [86].

1.3 Electroretinography in Diabetic Retinopathy Diabetic retinopathy is an end-organ complication of diabetes mellitus and is a leading cause of blindness worldwide [87], [88]. It can take many years for a diabetic patient to develop symptoms of retinopathy, but patients can progress quite rapidly to blindness once symptoms appear [89]. In many ways current treatments are inadequate, and there is a critical need for a better understanding of the basic pathogenic mechanisms of this disease.

The traditional understanding of diabetic retinopathy is that the metabolic defects of diabetes cause microvascular damage, inflammation, and loss of endothelial integrity in the retina. These, in turn, lead to grossly detectable changes such as microaneurysms, hemorrhages, macular edema, and vascular proliferation which all finally lead to neuronal damage in the inner retina and loss of vision. Thus, diabetic retinopathy is treated clinically with drugs and procedures aimed at reducing vascular leakage, proliferation, and edema. However, deficits in the oscillatory potential (OP) component of the electroretinogram (ERG) have been observed by a number of researchers in the retinas of patients prior to visible microvascular defects [90]–[93]. In fact, patients with diabetes but without funduscopic retinopathy were found with decreased OP amplitudes as early as 1962 [94]. Collectively, these studies suggest that functional changes may actually occur before vascular pathology.

In the clinic, however, diabetic patients are still typically first seen and treated when they present with signs and symptoms attributable to retinal vascular pathology; electrophysiological visual testing is not regularly used to assess diabetic patients [95], [96]. Thus, a better understanding of

6 the neuronal deficits in the early-stage diabetic retina is ripe for clinical benefit; the identification of early functional changes might actually allow the prevention of later vascular pathologies before they even occur.

1.4 Objectives of the Dissertation This thesis is intended to improve our understanding of the neural circuitry of the inner retina as a project in two independent parts. In the first and larger part, presented in Chapter 2, I have focused on a basic scientific question relating to retinal network signal modulation. Recently, evidence has arisen that an amino acid transporter, EAAT5, plays a substantial role in this inhibitory process. In Chapter 2, I present the results of experiments designed to expand our understanding of the novel role of this protein in inner retinal visual processing. Specifically, I begin by correcting a previously-published theory regarding the circumstances under which

EAAT5 inhibition is actually active. I then examine whether EAAT5 primarily mediates autoinhibition or alloinhibition in rod bipolar cells, a question for which evidence supporting both possibilities had previously existed. Finally, I demonstrate the functional significance of

EAAT5-mediated inhibition on visual processing.

Revealing the function and mechanisms behind EAAT5-driven inhibition in retinal circuitry will not only fill the gaps in our understanding of visual processing but will also undoubtedly improve our understanding of neural circuit modulation in general, including a better understanding of the many roles that EAAT-mediated inhibition may play in other parts of the

CNS. In addition, a better understanding of retinal circuitry is needed to elucidate the mechanisms underlying common eye diseases like macular degeneration, retinitis pigmentosa, congenital blindness, and diabetic retinopathy. In the second part of this dissertation, presented in

Chapter 3, I shift focus from basic science to such a disease, concentrating in particular on the 7 neuronal effects of the early stages of diabetic retinopathy on the inner retina. Specifically, I present an analysis of electroretinograms (ERGs) recorded from high-fat diet (HFD) and control mice over 12 months. In so doing, I characterize the ERG changes that take place as diabetic retinopathy develops and demonstrate that despite the traditional understanding of diabetic retinopathy, electrophysiological deficits actually precede structural lesions in these mice.

Chapter 2: The Glutamate Transporter EAAT5 Mediates Rod Bipolar Cell Autoinhibition 2.1 Background The retina is capable of responding to visual stimuli over more than 12 orders of brightness magnitude [41], [42], necessitating substantial gain modulation of neural signals to avoid response saturation and maintain sensitivity. In the rod pathway, gain modulation occurs primarily at the synapse between the rod bipolar and the AII amacrine cell [97]. It has long been understood that, in response to presentation of light stimuli, a variety of amacrine cell populations release GABA and glycine onto rod bipolar cell axon terminals, which activate

GABAA, GABAC, and glycine receptors, evoking an inhibitory chloride conductance [98]–[106].

These conventional inhibitors mediate both autoinhibition, in which excitation of the cell leads to feedback inhibition onto itself, and alloinhibition, in which inhibition is driven by other bipolar cells and occurs independently of the activity of the cell of interest. Specifically, conventional autoinhibition of the rod bipolar cell is mediated via the A17 amacrine cell reciprocal synapse while conventional alloinhibition is mediated via various lateral amacrine cell pathways (Figure

2.1) [107], [108].

Figure 2.1: Conventional inhibition is mediated by a variety of sources. Rod bipolar cells receive presynaptic GABA-mediated autoinhibition from A17 amacrine cells and additional presynaptic alloinhibition from other GABAergic and glycinergic amacrine cells.

However, in addition to conventional inhibitors, rod bipolar cells are also subject to a presynaptic inhibition mediated by excitatory amino acid transporter 5 (EAAT5) [65]. Transporters in the

EAAT family are thusly named for their role in removing glutamate from synapses following neurotransmission. Glutamate binding to EAATs can also open a chloride conductance through the transporter that suppresses synaptic activity [63], [64]. These two functions are not stoichiometrically coupled; the rate of glutamate uptake is independent of chloride concentration, while the transporter’s chloride conductance is dependent on glutamate, sodium, and membrane potential [76], [77]. In addition, the relative activity of each of these two functions varies greatly between individual members of the EAAT family [82], [83].[63], [64][62], [63][53], [54] For example, while EAAT1, EAAT2, and EAAT3 have little chloride conductance and play an important role in glutamate reuptake in the cells that express them [84], [85], EAAT4 and

EAAT5 play a significant role in presynaptic inhibition but little role in glutamate uptake [86],

[109]. Compared to other forms of presynaptic inhibition, EAAT5-mediated inhibition has slow

10 kinetics, with a decay time constant of around 130 ms, as opposed to 34 ms for GABACR- or 2 ms for GABAAR-mediated inhibitory currents [110], [111]. In the retina, EAAT5 is expressed in the axon terminals of photoreceptors and rod bipolar cells [112].

Wersinger et al. demonstrated that rod bipolar EAAT5 can mediate autoinhibition, since glutamate release from the cell of interest can evoke EAAT5-mediated currents at its own synaptic terminal [112]. It has also been shown that EAAT5-mediated alloinhibition is possible, because glutamate spillover from neighboring rod bipolars can evoke EAAT5 currents [113].

However, both of these studies either used nonphysiological electrical stimulation of cells or bath-applied antagonists of conventional inhibition. Previous work suggested that the mode of presynaptic inhibition onto rod bipolar cells switches from conventional inhibitors to EAAT5- mediated inhibition as a function of the intensity of the light stimulus, and there has been speculation that EAAT5 inhibition mediates light adaptation in the rod pathway [111]. In response to light, the depolarizing response of rod bipolar cells is limited by inhibitory neurotransmitters [114], and exogenous blockade of GABA and glycine is known to upregulate glutamate release from bipolar cells [99], [115], so the degree to which auto- or alloinhibition is important under physiologically-relevant conditions remains unknown (Figure 2.2). Currently, neither the source of glutamate that activates EAAT5 in the rod bipolar cell, nor the function of this transporter in visual processing is clear.

Figure 2.2: EAAT5-mediated presynaptic inhibition is gated by glutamate and has two potential sources indicated by the numbered red arrows: 1) autoinhibitory glutamate released from the rod bipolar of interest onto itself and 2) alloinhibitory glutamate released from neighboring rod bipolar cells.

In this chapter, I investigate these questions. I first show that both forms of inhibition occur at all light intensities, suggesting that the mode of inhibition does not switch across light intensities in the manner previously described. I then show that although EAAT5 is not a significant contributor to alloinhibition, it is the main source of autoinhibition to rod bipolar cells. Finally, in-silico modeling of the rod bipolar – AII amacrine cell synapse shows that EAAT5-mediated autoinhibition has a substantial effect on synaptic transmission and may be important in maintaining rod bipolar cell signaling within the sensitive region of the synaptic activation curve.

2.2 Materials and Methods

2.2.1 Animal Protocols Animal protocols were approved by the Washington University School of Medicine Animal

Studies Committee.

2.2.2 Retinal Dissection Mice of either sex (28 – 60 days of age; C57BL/6J strain; The Jackson Laboratory) were dark- adapted overnight and sacrificed using carbon dioxide. The retina was isolated, and 400 µm slice preparations were made from the dorsal retina. All dissection procedures were performed in complete darkness under infrared illumination. Dissection medium (see section 2.2.5 Solutions and Drugs) was continuously oxygenated. Retinal preparations were stored in an oxygenated dark box at room temperature until experimental use. Dissections were performed from 10 to 11 am, and recordings were always performed in the afternoon.

2.2.3 Whole-Cell Recordings Whole-cell patch recordings were made from bipolar cell somata in retinal slices by viewing them with an upright microscope under infrared illumination. The sample chamber was curtained to prevent light leakage from infrared eyepieces reaching the samples. Extracellular media was continuously applied to the samples via a preheated perfusion system positioned near the slice preparation. Electrodes were made from borosilicate glass with a P97 Flaming/Brown puller

(Sutter Instruments) and had resistances of approximately 5 MΩ. All recordings were made at

30°C. Signal software (Cambridge Electronic Design) was used to generate electrical command outputs and acquire and store data. Recordings were sampled at 5 kHz and filtered at 2 kHz with the four-pole Bessel filter on the Axopatch 200B (Molecular Devices). Liquid junction potentials were corrected at the beginning of each recording. Depending on the experiment, cells were either voltage clamped at 0 mV, the reversal potential for nonselective cation channels, isolating

- Cl currents, or were held at -60 mV with ECl at 0 mV. In the latter case, the mGluR6-mediated cation currents ran down quickly and did not contaminate the Cl- currents.

2.2.4 Morphological Identification Sulforhodamine B (0.005%) was included in the intracellular solution for all recordings to identify the cell type and to confirm correct placement of puff pipettes at the rod bipolar axon terminal. After electrophysiological recordings were finished, cells were morphologically identified by their axon terminal ramification morphology and location in the inner plexiform layer (IPL) (Figure 2.3) [116], [117].

Figure 2.3: Differential interference contrast photomicrograph of a pipette and rod bipolar cell (RBC) filled with fluorescent dye (Sulforhodamine B 0.005%) following whole-cell recording in a retinal slice.

2.2.5 Solutions and Drugs

Retinal dissections were performed in HEPES-buffered extracellular Ringer’s solution containing (in mM): 137 NaCl, 2.5 KCl, 2.5 CaCl2, 1 MgCl, 28 glucose, 10 HEPES, 4 Na-L- lactate, 2 Na-pyruvate, 0.5 L-glutamine, adjusted to pH 7.4 by NaOH. Physiological recordings 14 were performed in bicarbonate-buffered extracellular solution containing (in mM): 125 NaCl, 2.5

KCl, 2 CaCl2, 1 MgCl2, 26 NaHCO3, 20 glucose, 1.25 NaH2PO4,4 Na-L-lactate, 2 Na-pyruvate,

0.5 L-glutamine. Extracellular solutions were continuously aerated with 95% O2 and 5% CO2 and the pH was 7.4 at 30°C. For experiments in which cells were held at 0 mV the intracellular solution was (in mM): 120 CsOH, 120 D-gluconic acid, 1 CaCl2, 1 MgCl2, 11 EGTA, 10 Na-

HEPES, 10 TEA-Cl, 5 ATP-Mg, 1 GTP-Na, adjusted to pH 7.2 by CsOH. For experiments in which cells were held at -60 mV, the intracellular solution was (in mM): 125 CsCl, 25 HEPES,

10 TEA-Cl, 0.5 EGTA, 2.5 ATP-Mg, 2.1 GDPβS, adjusted to pH 7.2 by CsOH.

When bath-applied, conventional inhibitory receptor antagonists were the GABAA receptor antagonist (–)-bicuculline methobrimide (50 µM), GABAC receptor antagonist (1,2,5,6-

Tetrahydropyridin-4-yl)methylphosphinic acid hydrate (TPMPA) (50 µM), and glycine receptor antagonist strychnine (1 µM). When puff-applied, the concentrations used were 250 µM bicuculline, 250 µM TPMPA, and 5 µM strychnine. Puff-applied glutamate was at concentration

2.5 mM. EAATs were blocked with DL-threo-β-benzyloxyaspartic acid (TBOA) (50 µM extracellularly or 3 mM intracellularly). Excitatory glutamate receptors were generally not targeted pharmacologically, except in Figure 2.10, when AMPA and NMDA receptors were blocked by bath application of 6-cyano-7-nitroquinoxaline-2,3-dione disodium salt hydrate

(CNQX, 5 µM), and D-(–)-2-amino-5-phosphonopentanoic acid (D-AP5, 50 µM), respectively.

Puff-applied drugs were dissolved in a HEPES-buffered extracellular solution identical to the one used for retinal dissections, loaded into pipettes similar to those used for patch recordings, and were puffed onto rod bipolar cell terminals using a Picospritzer II (Parker Hannifin). Bath- applied drugs were dissolved into the recording solution described above and aerated using the same source of gas and perfused into the recording chamber using the same delivery system as 15 the control solution. Drugs were given at least 4 minutes to perfuse into and be washed out of the chamber before the next round of recordings, and therefore experiments with bath-applied drugs were performed in batches; otherwise, the order of experiments was randomized.

2.2.6 Light Stimulation Full-field light stimuli were generated using a 505 nm LED projected through the optical path of the microscope, passing through a diffusion filter to ensure even lighting over the entire field and a wide enough lens to illuminate the entire retinal slice. Coarse control of the light intensity was achieved by inserting neutral-density filters of varying strength into the optical path. Fine control of light intensity was achieved by varying the current through the LED.

2.2.7 Data Analysis Charge transfer was measured using the area function in Signal, and difference in charge transfer was calculated by simple subtraction. Standard errors of the mean (SEMs) for difference in charge transfer were calculated by addition of root mean squares of SEMs of the underlying charge transfers. P-values were calculated using two-tailed t-tests. In experiments comparing the same conditions across multiple light intensities, corrections for multiple comparisons were made using the Holm–Bonferroni method.

2.2.8 Modeling The rod-bipolar to AII-amacrine cell synapse was simulated in NeuronC [118]. Each cell was modeled as a sphere (diameter 4 µm), and each cell’s membrane was modeled as a first-order resistor-capacitor circuit. The rod bipolar cell explicitly included voltage-dependent calcium

2 2 channels (L-type, non-inactivating, 8 mS/cm ) and calcium pumps (Vmax = 20 nA/cm , Km = 20

µM, non-electrogenic). Calcium diffused through radial shells (n = 10) to and from an internal core. All other gated and non-gated channels were represented by the global membrane

16 resistance of the rod bipolar cell, which was calibrated to match our experimental data (Rm = 400

Ω-cm2). Rod bipolar cell membrane capacitance was set to calibrate the membrane time constant

2 (40 ms, Cm = 100 µF/cm ). The synapse itself was simulated by a series of filters that defined the transfer function for vesicle release and a temporal impulse function (release rate proportional to

[Ca], tau = 2 ms), as well as the size of the readily-releasable pool (20) and its replenishment rate

(50 Hz). For simplicity, release was set to be noiseless, i.e. only the average rate was simulated.

The postsynaptic currents were generated by binding of the simulated glutamate release to a sequential-state Markov model of a non-desensitizing AMPA receptor/channel. A current clamp in the rod bipolar was driven by pre-recorded data from the experiments on real rod bipolar cells,

2 2 and a voltage-clamp in the AII (Rm = 100 MΩ-cm , Cm = 1 µF/cm ) was set to the membrane resting potential (-70 mV) to record post-synaptic currents. The presynaptic [Ca2+] was plotted from the Ca2+ diffusion shell closest to the simulated cell membrane.

2.3 Results

2.3.1 The Mode of Retinal Presynaptic Inhibition Does Not Switch with Light Intensity. Light stimuli evoke inhibitory Cl- currents on rod bipolar axon terminals that are mediated by a combination of amacrine cell activation of GABA and glycine receptors and glutamate-driven activation of EAAT5 [98], [111]. I investigated the relative contribution of each of these pathways in response to light stimuli. Unexpectedly, however, GABAA, GABAC, and glycine receptor antagonists reduced inhibition only minimally in response to dim stimuli, while further addition of EAAT5 antagonist TBOA [119] blocked such inhibition to a greater degree (Figure

2.4).

Figure 2.4: Example RBC IPSCs. Inhibitory post-synaptic currents (IPSCs) evoked by a brief (200 ms) flash of light from a rod bipolar cell (RBC) voltage clamped at 0 mV and bath-applied inhibitory receptor antagonists are shown. RBC IPSCs evoked by a dim flash of light (400 photons/µm2/s) from a black background were blocked very little by a mixture of GABA and glycine (gly) blockers (50 µM bicuculline, 50 µM TPMPA, 1 µM strychnine). Subsequent addition of the EAAT antagonist TBOA (50 µM) blocked the current.

In response to bright light stimuli, bath application of GABA and glycine receptor blockers significantly reduced inhibition onto rod bipolar cells, but additional bath application of TBOA blocked the inhibitory response further (Figure 2.5). These results differ from earlier work suggesting that in the dark-adapted retina, low-intensity light stimuli elicit presynaptic inhibition onto rod bipolar cells mediated by GABA and glycine, while high-intensity light stimuli fail to elicit GABA- or glycine-mediated inhibition, and instead elicit EAAT5-mediated inhibition

[111]. My results are not consistent with this hypothesis of an inhibitory switch.

Figure 2.5: The mode of light-evoked inhibition to rod bipolar cells does not depend on light intensity. Inhibitory post-synaptic currents (IPSCs) were evoked by a brief (200 ms) flash of light from bipolar cells voltage clamped at 0 mV and bath-applied inhibitory receptor antagonists. Normalized inhibitory response by light intensity for RBCs is shown. All stimuli were presented in the context of a zero-light background. Conventional inhibitory receptor blockers (open circles) had minimal effect at any light intensity compared to controls (filled circles), while TBOA (filled triangles) caused a significant reduction in inhibition at all light intensities above 10 ph./µm2/s. (n = 43 cells) Error bars indicate SEM. * indicates p-value < 0.05.

The EAAT5-mediated current was unique to rod bipolar cells, as conventional inhibitory blockers eliminated all inhibition to cone bipolar cells (Figure 2.6). This validates previous work and is consistent with the finding that EAAT5 is found in rod but not cone bipolar cells [111],

[112]. All of these experiments were performed in dark-adapted retinas, and the dimmest stimuli that elicited an inhibitory response in either rod or cone bipolar cells were below cone sensitivity,

19 suggesting that rod input can contribute inhibition onto cone pathways in response to dim stimuli

[120], [121].

Figure 2.6: Inhibition onto cone bipolar cells is exclusively conventional. Inhibitory post-synaptic currents (IPSCs) are evoked by a brief (200 ms) flash of light from bipolar cells voltage clamped at 0 mV and bath-applied inhibitory receptor antagonists. Normalized inhibitory response by light intensity for cone bipolar cells is shown. All stimuli were presented in the context of a zero-light background. Conventional inhibitory receptor blockers alone (open circles) completely ablated the inhibitory response compared to controls (filled circles). (n = 4 cells) Error bars indicate SEM. * indicates p-value < 0.05.

2.3.2 GABA and Glycine Mediate Alloinhibition onto Rod Bipolar Cells. I was surprised by the capacity of bath-applied TBOA to block inhibitory responses to such a significant extent in rod bipolar cells. Others have shown that when presynaptic inhibition is blocked at bipolar cell terminals, downstream excitatory post-synaptic currents are enhanced

[99], [115]. I considered that the simultaneous bath-application of GABA and glycine receptor 20 blockers may have disinhibited cells throughout the network, causing abnormally high synaptic activity in neighboring bipolar cells, potentially leading to artificially enhanced EAAT5- mediated inhibition via glutamate spillover. To test this possibility, I measured the light-evoked inhibitory response in the presence and absence of intracellular TBOA applied through the recording pipette [111], and in the presence or absence of a puff application of GABA and glycine receptor blockers on the axon terminal (Figure 2.7).

Figure 2.7: Cartoon depicting intracellular and puff application of receptor antagonists. Diagram illustrates experimental paradigm to limit network effects of drug application. RBCs were recorded in the presence and absence of intracellular TBOA (3 mM) in the recording pipette, while a second puff pipette with conventional inhibitor antagonists (250 µM bicuculline, 250 µM TPMPA, 5 µM strychnine) was targeted at the axon terminal.

In the presence of puff-applied GABA and glycine antagonists, rod bipolar cells demonstrated substantial reduction of inhibition in response to light (Figure 2.8). In fact, conventional

21 inhibitory blockers significantly reduced inhibitory input when puff-applied, regardless of the presence or absence of intracellular TBOA.

Figure 2.8: Puff application of GABA/glycine blockers successfully blocks the light-evoked IPSC. A rod bipolar cell was voltage clamped at 0 mV and a 30 ms flash of light (106 ph./µm2/s) was given in the presence (gray) and absence (black) of a puff starting 100 ms before the onset and ending 100 ms after the offset of light.

In contrast, internal application of TBOA had no significant effect on the amount of inhibition onto rod bipolar cells (Figure 2.9).

Figure 2.9: Results of intracellular and puff application of receptor antagonists. Puff application of GABA/glycine antagonists (dark bars) blocks light-evoked IPSCs compared to controls (white bars) while intracellular TBOA (left) has no significant effect compared to controls (right). RBCs were voltage clamped at 0 mV (n = 12 cells). Error bars indicate SEM. * indicates p < 0.05.

To confirm that intracellular TBOA in my preparation was capable of blocking glutamate- mediated inhibition, I tested the responses of control and TBOA-filled rod bipolars to glutamate puff application at the axon terminal and demonstrated that internal TBOA reduced the responses to puffs of glutamate (Figure 2.10). Overall, these results indicate that conventional inhibitors, but not EAAT5, mediate alloinhibition onto rod bipolar cells, in agreement with previous studies

[98]–[100].

Figure 2.10: Intracellular application of TBOA blocks the response to glutamate. RBCs with (grey) and without (black) intracellular TBOA were voltage clamped at -60 mV with ECl = 0 and 5 µM CNQX and 50 µM D-AP5 bath-applied to block AMPA and NMDA receptors, then subjected to a brief 2.5 mM glutamate puff. Left: two example traces. Right: charge transfer for each category. (n = 8 cells) Error bars indicate SEM. * indicates p < 0.05.

2.3.3 EAAT5 Mediates Autoinhibition onto Rod Bipolar Cells. Usage of a static voltage clamp paradigm in which the cell was clamped at a fixed voltage had prevented the activation of voltage-gated calcium channels in the cell of interest, blocking excitatory output and therefore blocking any type of autoinhibition in my recordings. Thus, I developed a moving voltage clamp experiment in order to test the extent of EAAT5-mediated autoinhibition under physiologically relevant conditions. I applied to rod bipolar cells a voltage pulse input modeled to duplicate the excitation pattern of a rod bipolar cell in response to a light pulse of matching intensity and duration [114] (Figure 2.11). The initial current responses from such traces would necessarily include leakage currents changing in response to the shift in voltage, meaning that raw amplitudes would not be informative. However, by comparing 24 responses from the same cell in the presence and absence of a given extracellularly-applied inhibitory receptor antagonist, the difference between the two responses could measure the relevant level of inhibition. I also measured the alloinhibitory responses of these cells via a fixed voltage and stimulation with light, to better facilitate direct comparison between auto- and alloinhibition under comparable conditions of resting membrane potential and ECl.

Figure 2.11: Voltage pulse input. Input was modeled to mimic the excitatory profile of an RBC in response to the light stimuli used in this chapter based on data in [114].

In this manner, I compared the autoinhibitory response of rod bipolar cells in the presence and absence of conventional inhibitory blockers and observed small but measurable levels of GABA- and glycine-mediated inhibition (Figure 2.12).

Figure 2.12: Voltage pulse example traces with and without GABA/gly blockers. RBCs stimulated from a baseline of -60 mV with the voltage pulse pictured in Figure 2.11 without (grey) and with (black) 2 s puff application of GABA/gly blockers. GABA/gly blockers elicited a small but measurable difference that could be attributed to autoinhibition from conventional inhibitory sources. ECl = 0 mV.

Meanwhile, the difference between TBOA and TBOA-free responses demonstrated that EAAT5 plays a substantial role in autoinhibition (Figure 2.13).

Figure 2.13: Voltage pulse example traces with and without TBOA. RBCs stimulated from a baseline of -60 mV with the voltage pulse pictured in Figure 2.11 without (grey) and with (black) extracellular application of TBOA. TBOA elicited a larger difference than conventional blockers that could be attributed to autoinhibition from EAAT5. ECl = 0 mV.

Overall, these results allow us to make direct comparisons, painting a more complete picture of the extent to which each source of inhibition impacts rod bipolar cells in response to physiologically-relevant stimuli. They demonstrate that autoinhibition onto rod bipolar cells is mediated by both conventional channels and by EAAT5, but that EAAT5 is the most significant mediator of autoinhibition (Figure 2.14, left). At the same time, they also confirm that EAAT5 does not play any significant role in alloinhibition, which is largely mediated by conventional

GABA and glycine receptors (Figure 2.14, right).

Figure 2.14: EAAT5 is the primary mediator of autoinhibition in RBCs. ECl = 0 mV. Left: Summary data for autoinhibition experiments depicted in Figures 2.12 and 2.13. Data represents the difference in charge transfer between two comparator conditions in individual cells. GABA/glycine-mediated autoinhibition and EAAT5-mediated autoinhibition were significantly different from each other and were also each significantly different from zero. (n = 24 cells) Right: Summary data for alloinhibition experiments. RBCs were held at -60 mV and flashed for 100 ms at 105 ph./µm2/s. TBOA was present or absent in the intracellular media. GABA and glycine blockers were puff-applied. Data represent the difference in charge transfer between populations of cells. GABA/glycine mediated alloinhibition was significantly different from both EAAT5-mediated alloinhibition and from zero. EAAT5-mediated alloinhibition was not significantly different from zero. (n = 10 cells) Error bars indicate SEM. * indicates p < 0.05.

To demonstrate that the difference between the TBOA and TBOA-free traces in my autoinhibition experiments was strictly attributable to glutamate release from the cell of interest,

I performed two control experiments. In the first, I held the cell at a reduced baseline of -100 mV, while applying a stimulus of the same amplitude, duration, and shape as before, in order to keep the cell below the threshold for synaptic activation (Figure 2.15). In the second, I

28 depolarized the cell to zero potential for an extended period of time, and then briefly returned the cell to -60 mV and applied the standard stimulus, aiming to deplete the readily-releasable vesicle pool prior to stimulation (Figure 2.16).

Figure 2.15: EAAT5-mediated autoinhibition is dependent on synaptic activation. A rod bipolar cell was stimulated from a baseline of -100 mV with the voltage pulse in Figure 2.11, without (grey) and with (black) extracellular application of TBOA. A lack of TBOA-sensitivity confirms that the autoinhibition attributed to EAAT5 is dependent on synaptic activation.

Figure 2.16: EAAT5-mediated autoinhibition is dependent on glutamate release. A rod bipolar cell was stimulated from a baseline of -60 mV with the voltage pulse in Figure 2.11 immediately following depolarization to -30 mV for 60 s. Lack of TBOA-sensitivity confirms that EAAT5-mediated autoinhibition is dependent on vesicle release.

In both cases, the TBOA-blockable effect was completely abolished (Figure 2.17), confirming that the observed role of EAAT5 in autoinhibition is dependent on synaptic vesicle release.

Figure 2.17: EAAT5-mediated autoinhibition is dependent on synaptic activation and glutamate release: Summary data representing the difference in charge transfer between two comparator conditions in individual cells is shown. No stim indicates a cell held at -60 mV without applying a voltage stimulus. “Stim,” “-100 mV baseline stim,” and “depolarization before stim” indicate experiments depicted in 2.13, 2.15 and 2.16, respectively. (n = 24 cells) Error bars indicate SEM. * indicates p < 0.05.

2.3.4 EAAT5 Inhibition Modulates Signaling to AII Amacrine Cells. Previous work has established that rod bipolar cell synapses are biphasic, with a transient component that encodes increases in contrast, and a sustained component that encodes luminance, mediated by depletion of the readily-releasable vesicle pool [122]. To investigate the role of EAAT5-mediated inhibition in modulating signaling at this synapse, I compared the outputs of an in-silico model of the rod bipolar – AII amacrine cell synapse when given inputs modulated by autoinhibition as measured by the above experiments. When stimulated with current inputs matching the responses of real cells in the presence and absence of TBOA, a model rod bipolar without EAAT5-mediated inhibition exhibited a two-fold increase in the peak

31 activation of a downstream AII amacrine cell (Figure 2.18, bottom), demonstrating that EAAT5- mediated inhibition is substantial enough to act as an important modulator of signaling at this synapse. More dramatically, I also found that local intracellular calcium ion concentration at the rod bipolar cell pre-synapse was increased 34-fold (Figure 2.18, top).

Figure 2.18: Outputs of in-silico model of the RBC – AII amacrine synapse. The virtual RBC was stimulated with an excitatory input taken from the autoinhibition results in either the absence or presence of TBOA (see Figure 2.13). Top: The local intracellular calcium ion concentration at the RBC pre- synapse. In our stimulus paradigm, under normal conditions, [Ca2+] peaks at only 0.46 µM. In the absence of EAAT5-mediated inhibition, however, [Ca2+] peaks at 15.8 µM. Bottom: the excitatory post-synaptic current in the AII amacrine cell. The control trial elicited a peak excitation of 10.5 pA, while the trial without EAAT5-mediated inhibition peaked at 21.9 pA. 33

Since this change in [Ca2+] was accompanied by a comparatively modest increase in peak AII activation, I hypothesized that in the absence of EAAT5-mediated inhibition the model cell was operating in the saturation plateau region of the synaptic activation curve, beyond the sensitive range of synaptic response. To test this hypothesis, I measured the sensitivity of both models to small changes in input amplitude and found that the output of the control model was indeed 32- fold more sensitive to variation in inputs than the model lacking EAAT5 mediated inhibition

(Figure 2.19). This meant that the control cell was indeed operating along a much steeper region of synaptic activation relative to the TBOA cell, supporting my hypothesis. Overall, these results suggest that EAAT5-mediated autoinhibition onto rod bipolar cells may be important for gain modulation that acts to keep the rod bipolar synapse within the more sensitive part of its activation range.

Figure 2.19: The synaptic sensitivity of the RBC synapse with and without EAAT5. The inputs to each RBC were increased and decreased by 1 pA, and the resulting change in AII peak excitation was quantified as a ratio of input current to output current, representing the sensitivity of the synapse to encode small stimulus changes. The control cell was 32 times more sensitive than the TBOA cell. This suggests that EAAT5- mediated autoinhibition keeps the RBC within the sensitive range of synaptic activation.

2.4 Discussion This study characterizes the relative contribution of EAAT5-mediated inhibition onto rod bipolar cells in response to physiologically relevant stimuli. It also compares the roles of EAAT5- mediated autoinhibition and alloinhibition. While conventional inhibitors mediate nearly all alloinhibition onto rod bipolar cells, EAAT5 is the primary mediator of rod bipolar autoinhibition and may have a substantial impact on postsynaptic excitation of the downstream

AII amacrine cell. Overall, these findings greatly increase understanding of the function of

EAAT5-associated chloride channels in rod bipolar cells.

While activation of EAAT5 by light was observed, my results do not agree with the idea of an inhibitory switch in which EAAT5 specifically enables inhibition in response only to bright stimuli and conventional inhibitory pathways mediate inhibition in response only to dim stimuli

[111]. Unlike the previously published work, I found EAAT-mediated inhibition across a much broader range of light intensities. I attempted to attribute the discrepancy between the differing results to a number of possibilities, including the wait time between cellular break-in and beginning the experiment, presence or absence of glutamine in the extracellular media, delay between application of flashes, thickness of the slice preparations, duration of the light stimuli, temperature of the prep, use of dorsal vs. ventral retina, and color of the presented light, but attempts to replicate the previously published finding by changing any of these variables were all unsuccessful (data not shown). In the end, in order to completely assure myself of the validity of my findings, I assayed more than ten times as many cells as the matching experiment in the cited paper.

I have also shown that glutamate spillover likely activates EAAT5 inhibition less than previously suggested. In the past, extensive EAAT-mediated spillover transmission in rod bipolar cells has been reported [111], [113], but I believe these findings should be viewed with caution because, to my knowledge, they have always depended on either excessive direct stimulation of the neighboring cell or network-wide blockade by conventional inhibitors, which I have here demonstrated leads to non-physiologic enhancement of EAAT5-mediated inhibition through disinhibition of bipolar cells throughout the network. Ichinose & Lukasiewicz [111] demonstrated that EAAT5-mediated inhibition onto rod bipolar cells is spatially limited, consistent with the understanding that for glutamate to be released near the rod bipolar cell synapse necessitates synaptic activity from either the cell in question or a neighboring one close

36 enough to lead to spillover, which would preclude any form of long-distance lateral inhibition.

My results here are consistent with this understanding and also bring it a step further by demonstrating that, under physiological stimulus conditions and in the absence of any exogenous network blockade of GABA or glycine, EAAT5 does not play a major role in alloinhibition.

Nonetheless, I have demonstrated an important role for EAAT5 in rod bipolar cells: autoinhibition. It is well-characterized that conventional autoinhibition onto rod bipolar cells is mediated by A17 amacrine cells, which form a reciprocal GABAergic synapse with them [123],

[124]. Up to this point, however, study of total autoinhibition onto rod bipolar cells has been neglected. This is likely because inhibitory inputs are typically studied by clamping the cell of interest at a fixed voltage, preventing the cell’s voltage from changing and precluding glutamate release, which necessarily eliminates EAAT5-mediated autoinhibitory effects from study.

However, my findings demonstrate that the overlooked EAAT5 inhibitory feedback pathway is actually the most prominent source of autoinhibition in a dark-adapted rod bipolar cell.

Furthermore, my modeling data show that this inhibition is sufficient to have a substantial effect on signal transmission through the rod bipolar – AII synapse and may be acting to keep the rod bipolar cell within the sensitive part of its activation curve.

Because autoinhibition does not include input from parallel pathways, it cannot carry spatial information, but rather must mediate temporal processing. In particular, I speculate that autoinhibition in the rod bipolar cell specifically acts as a gain modulator, allowing the cell to transmit sudden changes in signaling intensity, but acting as negative feedback to push the cell back towards an equilibrium in response to longer stimuli. In addition, complex control over this filtering effect could be effectively accomplished by separate mechanisms with differing kinetics. EAAT-mediated inhibition onto rod bipolar cells is known to have slower kinetics than 37 conventional inhibition [110], [111]. This suggests that EAAT5-mediated inhibition may be acting in concert with conventional A17-mediated autoinhibition to define a sophisticated temporal gain modulator.

In particular, at very low light intensities, the delayed autoinhibition from EAAT5 may act to enhance the relatively slow single-photon signal in the rod pathway. At low light intensities

(around 10-3 – 10-1 photons/µm2/s), the rod bipolar cell is tuned to transmit single photon signals from rods. Others have shown that the transient timing of the rod bipolar vesicle release function is the correct duration to pass a single photon signal [27], [122]. EAAT5 autoinhibition may be just delayed enough to allow the single-photon transient to pass, and then inhibit the rod bipolar cell to counter long-term baseline depolarization. Thus, by regulating the baseline slowly,

EAAT5 autoinhibition may prevent depletion of the readily-releasable pool and maintain the right level to depolarization to allow single-photon signals to be transmitted. At medium backgrounds (around 10-1 – 102 photons/µm2/s), EAAT5 autoinhibition may instead be acting to allow the contrast encoding described by Oesch and Diamond [122]. At brighter backgrounds than that, rod bipolar cells are highly depolarized (20 – 30 mV above their dark resting level), and their Ca2+ channels are likely to be inactivated when depolarized that far.

Overall, this study increases understanding of the roles of EAATs in the retina. The removal of glutamate from retinal synapses is primarily performed by EAAT1, also known as GLAST, on

Müller glial cells [125]–[127]. EAAT2 and EAAT3, which can be found on retinal neurons, are similarly biased towards glutamate transport and exhibit very little chloride conductance but appear to have a much smaller role in bulk removal of glutamate from synapses [127], [128].

EAAT5, on the other hand, plays almost no role in glutamate transport [86], [109], but instead mediates autoinhibition in rod bipolar cells, and likely plays the same role in photoreceptors as 38 well. EAAT4 has a high-affinity glutamate-gated chloride activity similar to EAAT5, but is primarily expressed in the cerebellum [64], where it may be acting similarly to EAAT5 in the retina.

Gain modulation in sensory-system neurons is important for transmitting information on stimulus changes despite a wide range of baseline stimulus intensities, and the glutamate- activated inhibition of EAAT5 makes it a natural candidate for gain modulation. My data supports the idea that EAAT5 is playing this role at the rod bipolar synapse.

Chapter 3: Functional Deficits Precede Structural Lesions in Diabetic Retinopathy 3.1 Background Diabetic retinopathy is an end-organ complication of either type 1 or type 2 diabetes mellitus

[87]. It is commonly understood to be the result of damage to the microvasculature and neurons of the retina, and is a leading cause of blindness worldwide [88]. Its prevalence is increasing at an alarming rate, particularly in developed countries where the incidence of type II diabetes has been rapidly increasing, likely due to the current epidemic of obesity and overnutrition. The

National Eye Institute projects that more than 14.6 million Americans alone will have diabetic retinopathy by the year 2050 [129]. It can take many years for a diabetic patient to develop symptoms of retinopathy, but patients can progress quite rapidly to blindness once symptoms appear [89]. Disease progression involves vascular endothelial growth factor (VEGF), and anti-

VEGF therapy is commonly used in diabetes care [130], [131]. However, anti-VEGF agents require prolonged frequent administration, are not always effective, and may have adverse effects [132]. Long-term VEGF antagonism in mice may induce vision loss [133]. On the other hand, other treatments for diabetic retinopathy are also associated with serious adverse effects

[134]. There is a critical need for a better understanding of the basic pathogenic mechanisms of this disease.

The traditional understanding of diabetic retinopathy is that diabetes causes vascular leakage and edema in the retina which in turn leads to neuronal damage and loss of vision. Thus, diabetic retinopathy is treated clinically with drugs and procedures aimed at reducing vascular leakage, proliferation, and edema. However, deficits in the oscillatory potential (OP) component of the

40 electroretinogram (ERG) have been observed in retinas of both patients and diabetic animal models [135]. Oscillatory potentials represent processing in the inner retina and have been attributed to amacrine cells and ganglion cells [136]. Conversely, the ERG components that represent photoreceptor and bipolar cell function are unchanged when inner retinal defects are first detected. In morphological studies, diabetic mice lose several classes of amacrine and ganglion cells via apoptosis [137], [138]. Apoptotic losses, however, occur months after diabetes onset, whereas functional defects in these neurons can be observed within weeks [135], suggesting that early deficits of neuronal signaling or integration may possibly precede more severe changes.

In the clinic, diabetic patients are typically first seen and treated when they present with signs and symptoms attributable to retinal vascular pathology; electrophysiological visual testing is not regularly used to assess diabetic patients [95], [96]. Thus, a better understanding of the neuronal deficits in the early-stage diabetic retina may lead to earlier intervention and prevention of later vascularization, edema, and blindness.

A number of rodent models have been developed and used for the study of diabetic retinopathy

[139]. Genetic models will spontaneously develop diabetes; commonly used genetic models include the Ins2Akita mouse, a model of type 1 diabetes in which insulin deficiency results from a point mutation in the Insulin-2 gene, and the db/db mouse, a model of type 2 diabetes that lacks the leptin receptor, which leads to overfeeding, obesity, and insulin resistance [140], [141]. The neurotoxin streptozotocin as also widely used to induce beta cell loss and subsequent insulin deficiency to model type 1 diabetes in wild-type rodents [142]. Each of these approaches has important limitations, however. In general, single-locus genetic models such as the db/db and

Ins2Akita mice fail to represent the true multifactorial nature of human diabetes. Rather, some 41 elements of their pathology may be unrelated to diabetes but instead to the consequences of their underlying genetic alteration. More specifically, the Ins2Akita mouse is generally reflective of the overall physiology of type 1 diabetes but Ins2Akita mice do not develop retinopathy [143], and leptin signaling affects the retina [144], suggesting that db/db mice are not an ideal model for obesity-related diabetic retinopathy. Meanwhile, streptozotocin is used as a model for type 1 diabetes because it directly kills pancreatic β-cells and causes rapid diabetic disease, but it also has off-target effects including immunosuppression [145], nephrotoxicity [146], hepatotoxicity

[147], and neurotoxicity [148]–[150]. The undesired toxic effect on neuronal tissue obviously confounds the interpretation of retinal changes. In addition, streptozotocin-treated mice do not require insulin for survival, unlike humans with type 1 diabetes.

Another animal model is the high-fat diet obese mouse. Rodents fed a diet of 42% fat develop weight gain and subsequent insulin resistance and diabetes [151], which better reflects human obesity-related type 2 diabetes. Animals fed the high-fat diet developed hyperinsulinemia that peaked at 6 months of age and was less marked by 12 months (Figure 3.1). This pattern likely reflects initial beta-cell hypersecretion, followed by beta-cell dysfunction and loss mirroring the progression of human type 2 diabetes. Until recently, these mice had not been studied systematically for diabetic retinopathy, but these animals do in fact develop eye disease [152].

Figure 3.1: High-fat diet fed mice develop obesity and glucose intolerance. (Figure taken from [152].) Male C57BL/6J mice were fed standard chow or a high-fat diet providing 42% of energy from fat. Insulin levels from fasting serum samples were measured using an ELISA-based method at 3, 6, and 12 months of age. Values represent mean ± SEM from at least five independent experiments in each group. ****P < 0.0001.

In human non-proliferative diabetic retinopathy, vascular signs such as intraretinal microvascular anomalies, venous beading, dot-blot hemorrhages, and microaneurysms are detectable by ophthalmoscopy. In diabetic rodents, the equivalent identifiable pathologic features are capillary basement membrane thickening, atrophy of small capillaries, and a loss of pericytes known histologically as “pericyte ghosts” [139]. When retinas were flat mounted and trypsin-digested to stain for microvessel skeletons, there were no differences in the mean number of atrophic capillaries or pericyte ghosts between high-fat diet and chow-fed mice at 6 months of age (Figure

3.2). However, at 12 months of age high-fat diet animals showed more atrophic capillaries and pericyte ghosts compared with controls [152].

Figure 3.2: At an advanced age, high-fat diet fed mice develop typical lesions of diabetic retinopathy. (Figure taken from [152].) Retinal vasculature networks were analyzed in 6- and 12-month-old animals by trypsin digest of isolated and fixed retinas, followed by periodic acid Schiff/hematoxylin staining. High-fat diet fed animals develop significantly more atrophic capillaries (left) and pericyte ghosts (right) at 12 months, but not 6 months, of age, compared with chow-fed controls. Shown are quantifications of vascular lesions, as assessed by a masked grader. Values represent mean ± SEM from 10 independent experiments in each group. **P < 0.01, ***P < 0.001 by ANOVA.

Evans Blue dye is commonly used as a vascular permeability assay. In normal controls, Evans

Blue dye injected into the bloodstream stays within the plasma compartment for hours, but in experimental diabetes, Evans Blue dye leakage is detectable in the vitreous and neural retina

[153]. High-fat diet fed animals have more Evans Blue leakage into the aqueous compartment than controls at 12 months but not 6 months (Figure 3.3) [152].

Figure 3.3: High-fat diet feeding promotes retinal vascular permeability. (Figure taken from [152].) Evans Blue dye (25 mg/kg) was injected intravenously into anesthetized animals and allowed to circulate for 180 minutes. Quantification of dye fluorescence from aqueous paracentesis samples demonstrates significantly greater leakage in high-fat diet fed animals compared with controls at 12 months of age but not at 6 months of age. Values represent mean ± SEM. *P < 0.05 by ANOVA.

Here I perform electroretinography on high-fat diet mice over time to determine the sequence of vascular and neurologic dysfunction in diabetes induced by feeding without administration of a potential neurotoxin. My findings show that high-fat diet feeding produces retinopathy in which neurologic dysfunction precedes vascular dysfunction.

3.2 Materials and Methods

3.2.1 Animals Protocols followed the Association for Research in Vision and Ophthalmology Statement for the

Use of Animals and were approved by Washington University in St. Louis. C57BL/6J mice were free of RD1 or RD8 mutations. Animals were weaned onto Purina 4043 chow (13% kcal from fat, 62% kcal from carbohydrate, 25% kcal from protein) or Harlan Teklad TD 88137 HFD (42% kcal from fat, 43% kcal from carbohydrate, and 15% kcal from protein).

3.2.2 Electroretinography A UTAS BigShot System (LKC Technologies, Inc., Gaithersburg, MD) was used. Mice (≥5 for each group) were dark adapted overnight. Under red light illumination, animals were anesthetized with ketamine (80 mg/kg total body mass) and xylazine (15 mg/kg lean body mass).

Pupils were dilated with 1% atropine sulfate. Body temperature was maintained at 37°C with a heating pad. Contact lens electrodes were placed bilaterally with appropriate reference and ground electrodes.

The stimulus consisted of a full-field white light flash (10 µs) in darkness or in the presence of dim (30.0 candela [cd]/m2) background illumination after 10 min adaptation time. The response was recorded over 231 ms plus 25 ms of pretrial baseline. Between 5 and 10 repeated trials were averaged for each luminance, with 10 repeats used for the dimmest flashes, and 5 for the brightest. Raw data were processed using MATLAB software (MathWorks, Natick, MA). The amplitude of the a-wave was measured from the average pretrial baseline to the most negative point of the average trace, and the b-wave amplitude was measured from that point to the highest positive point, after subtracting oscillatory potentials (OPs). The eye with the larger recorded b- wave amplitude was used for each mouse. OPs were isolated using a digital Butterworth 25 Hz high-pass filter and quantified using root-mean-square analysis of filtered waves, normalized to the maximal b-wave amplitude. The log luminance of the stimulus (log [cd · s/m2]) was calculated based on the manufacturer's calibrations.

3.2.2 Statistical Analysis Distributions of electroretinography (ERG) responses across diet groups at different light intensities are described by means ± SEM. Differences were compared using two-way ANOVA, followed by a Bonferroni post hoc test for multiple comparisons.

3.3 Results Although the primary clinical hallmark of diabetic retinopathy is vascular disease, antecedent or coincident electrophysiological changes also occur in humans [94], [154]–[156]. It is less clear whether neural effects are involved in the etiology of diabetic retinopathy in rodents. To determine whether neurologic dysfunction occurs early in high-fat diet-induced diabetic retinopathy, I performed full-field, stimulus-evoked scotopic and photopic electroretinograms

(ERGs) in high-fat diet fed mice and chow-fed controls.

After dark adaption overnight, scotopic ERG was performed with full-field white stimulus given across increasing intensities. Thereafter, a 10-minute period of light adaptation was performed against a dim white background of 30-cd/m2 luminance. Photopic ERG was recorded at increasing stimulus intensities. Raw ERG data were then analyzed for a-wave (scotopic reading only) and b-wave (in both scotopic and photopic conditions) amplitudes and implicit times. The raw waves were processed through a digital high-pass filter to isolate early oscillatory potentials

(OPs), as shown in Figure 3.4.

Figure 3.4: ERG analysis parameters. ERG responses were recorded from animals in scotopic conditions after overnight dark adaptation. A high-pass filter was applied to the raw ERG to isolate the OPs for separate analysis from the a-wave and b-wave measurements.

As expected, maximal ERG responses to white light stimuli for both a-waves and b-waves declined with age (Figure 3.5). At no age did I observe a difference in amplitudes or implicit times in scotopic a-waves or b-waves between HFD and control groups. There were no differences in photopic b-wave characteristics between different groups at any age (data not shown).

Figure 3.5: High-fat diet feeding does not alter major retinal responses to light stimuli. ERG responses were recorded from animals in scotopic conditions after overnight dark adaptation. ERG a-wave amplitudes recorded at 3, 6, and 12 months of age are plotted against stimulus luminance (left), with corresponding b-wave amplitudes (right). Values represent the mean ± SEM from at least five animals at each age. Decline in amplitudes of responses are seen with increasing age. At all tested ages, no significant differences were observed between animals on different diets with regard to either major ERG component (ANOVA).

Amplitudes of summed OPs did not differ between dietary groups at 3 or 6 months of age

(Figures 3.6, top and 3.7, top). However, at 12 months of age, average OP amplitudes were 49 significantly reduced in HFD-fed animals compared with chow-fed controls at all stimulus intensities (P < 0.05) except for the dimmest (−2.4 log [cd · s/m2]) and brightest (0.89 log [cd · s/m2]) (Figure 3.8, top). Examples of the preferential decline in OP amplitudes in HFD-fed animals compared with controls, with relative preservation of a-wave and b-wave amplitudes, are shown in Figure 3.9.

Figure 3.6: Three-month high-fat diet fed mice do not have decreased OP amplitudes. Visual function in high fat diet or chow-fed animals was assessed at 3 months of age by ERG under scotopic, dark- adapted conditions. OPs were extracted from the raw ERG trace using a high-pass filter. The root mean square (RMS) of all OP peaks and troughs were normalized to the amplitude of the maximal b wave. RMS of OP amplitudes are plotted against stimulus luminance at 3 months of age (top). Values represent the mean ± SEM from at least five animals. Bottom: Correlation plot of normalized OP amplitude with the area under the curve (AUC) for the intraperitoneal glucose tolerance test (GTT) are shown for animals at 3 months of age. Each data point represents a unique animal. Pearson correlation analyses were performed for animals fed each diet at each tested age, with R2 and P values shown.

Figure 3.7: Six-month high-fat diet fed mice do not have decreased OP amplitudes. Visual function in high-fat diet or chow-fed animals was assessed at 6 months of age by ERG under scotopic, dark- adapted conditions. OPs were extracted from the raw ERG trace using a high-pass filter. The root mean square (RMS) of all OP peaks and troughs were normalized to the amplitude of the maximal b wave. RMS of OP amplitudes are plotted against stimulus luminance at 6 months of age (top). Values represent the mean ± SEM from at least five animals. Bottom: Reductions in OP amplitudes correlate with glucose intolerance beginning at 6 months of age. Correlation plots of normalized OP amplitude with the area under the curve (AUC) for the intraperitoneal glucose tolerance test (GTT) are shown for animals at 6 months of age. Each data point represents a unique animal. Pearson correlation analyses were performed for animals fed each diet, with R2 and P values shown.

Figure 3.8: Twelve-month high-fat diet fed mice have decreased OP amplitudes. Visual function in high-fat diet or chow-fed animals was assessed at 12 months of age by ERG under scotopic, dark- adapted conditions. OPs were extracted from the raw ERG trace using a high-pass filter. The root mean square (RMS) of all OP peaks and troughs were normalized to the amplitude of the maximal b wave. RMS of OP amplitudes are plotted against stimulus luminance at 12 months of age (top). At 12 months of age, mice fed the HFD demonstrate significant declines in OP amplitudes relative to controls. Values represent the mean ± SEM from at least five animals. *P < 0.05 by ANOVA. Bottom: Reductions in OP amplitudes correlate with glucose intolerance beginning at 6 months of age and continuing at 12 months. Correlation plots of normalized OP amplitude with the area under the curve (AUC) for the intraperitoneal glucose tolerance test (GTT) are shown for animals at 12 months of age. Each data point represents a unique animal. Pearson correlation analyses were performed for animals fed each diet at each tested age, with R2 and P values shown in the lower right corner of each graph.

Figure 3.9: Twelve-month high-fat diet mice have decreased OP amplitudes: Example traces. Visual function in high-fat diet or chow-fed animals was assessed by ERG under scotopic, dark-adapted conditions. OPs were extracted from the raw ERG trace using a high-pass filter. Representative ERG tracings from mice fed each of the two indicated diets are shown, with the HFD-fed mice displaying reductions in OP amplitudes compared with chow-fed controls.

HFD feeding promotes insulin resistance, but individual mice show variable glucose tolerance responses. To determine whether mice with greater glucose intolerance (i.e., higher area under the curve values during a glucose tolerance test) also have a greater reduction in OP amplitudes, we correlated OP amplitudes with area under the curve of the glucose tolerance test for individual mice at different ages. At 3 months (Figure 3.6, bottom), no correlation was found between glucose tolerance and OP amplitudes in either dietary group. However, at 6 months

(Figure 3.7, bottom), the area under the curve of the glucose tolerance test was correlated with lower OP amplitudes in the high-fat diet group (Pearson R2 = 0.60, P = 0.04). In contrast, no such correlation was found for chow-fed animals of the same age (R2 = 0.02, P = 0.81). At 12 months (Figure 3.8, bottom), the correlation between glucose tolerance and lower OP amplitudes

53 was also present in high-fat diet animals (Pearson R2 = 0.51, P = 0.009) but not in chow-fed controls (R2 = 0.12, P = 0.5).

Prompted by this correlation, I determined whether OP kinetics may be affected in the high-fat diet-induced diabetes model by measuring the latency time for each of the first four OPs recorded on scotopic ERG at 3, 6, and 12 months (Figure 3.10). Beginning at 6 months of age, a significant delay in implicit time for the first two OPs (OP1 and OP2) elicited by a −0.60 log (cd

· s/m2) white light flash was observed in HFD-fed animals (27.1 ± 2.2 ms) compared with controls (24.3 ± 1.8) at OP1 (P < 0.05). This delay persisted through 12 months, with an average

OP1 implicit time of 25.6 ± 2.0 ms in high-fat diet fed animals compared with 24.2 ± 1.4 ms in chow-fed controls (P < 0.05) and an average OP2 implicit time of 32.0 ± 2.6 ms in high-fat diet fed animals compared with 30.5 ± 1.9 ms in chow-fed controls (P < 0.05). These data are consistent with the notion that diet-induced insulin resistance promotes early neurologic deficits in the retina. By ERG analysis, these deficits are manifested as reductions in amplitude and delays in OP responses to white light stimuli under scotopic conditions. Thus, electrophysiologic abnormalities detected at 6 months precede the development of vascular abnormalities at 12 months (Figures 3.2 and 3.3) in a diabetes model induced by a clinically relevant diet without genetic manipulations or administration of toxins.

Figure 3.10: Delays in OP implicit times are seen in high-fat diet animals before development of vascular lesions. Implicit times of the first four OPs (OP1–4) were measured from ERGs generated by a 0.25 cd · s/m2 white light stimulus. Recordings were performed at each indicated age. At 6 months of age, HFD-fed mice show significant delays in the implicit times of the first two OP peaks compared with chow-fed mice. This difference persists through 12 months of age. Shown are mean values ± SEM for each of the four measured peaks, with individual values for each animal tested superimposed (at least five animals were used in each diet group at each age). *P < 0.05 by ANOVA. 55

3.4 Discussion Neurologic activity in the retina is metabolically demanding [157], and it is not surprising that retinal neural function might be altered in diabetes. Nevertheless, diabetic retinopathy is considered first and foremost a disease of blood vessels. In humans and mice, the earliest visible signs of the disease are generally believed to be vascular. Because of this, it has largely been assumed that early neural insults are resultant from vascular damage in the retina. However, in this chapter, I demonstrate in an animal model that in diet-induced diabetes retinal function is disturbed before the development of detectable defects in the retinal vasculature.

Using a clinically relevant diet, it has been shown that the development of obesity, glucose intolerance, and insulin resistance in mice is associated with pericyte damage and microvascular atrophy by 1 year of age [152]. Using ERG, I show that a functional neurologic deficit occurs by

6 months, before vascular injury. This functional abnormality is correlated with glucose intolerance and persists to 12 months.

This work is novel because the relationship between early ERG dysfunction and retinopathy development is confounded in other mouse models. In the streptozotocin diabetes model, ERG and structural vascular abnormalities have both been identified [158]–[160], but streptozotocin causes off-target neurologic effects [149], [150]. Streptozotocin also causes rapid onset of severe disease that obscures any potential early neural contribution. In the db/db mouse, which develops a disease more similar to type 2 diabetes due to leptin receptor deficiency, massive reductions in b-wave amplitude have been recorded at around 6 months of age [161], but these findings coincide with vascular defects, including morphological abnormalities and abnormal leakage.

Leptin signaling also has trophic and angiogenic effects in the retina that could have accounted for the ERG deficits seen in these mice [144]. Diet-induced metabolic disease in rodents has been shown to produce retinal inflammation and degeneration [162], [163], and a rat model of type 2 diabetes manifests features of retinal inflammation and vascular disease coincident with severe metabolic derangements [164], but these studies did not report ERG responses.

The high-fat diet model is a slow-onset disease, accurately representing the pathophysiology of human type 2 diabetes complications such as retinopathy. Even after 1 year, high-fat diet mice do not lose inner retinal thickness [152], confirming that disease in these animals is mild and similar to the early stages of human retinopathy [165], [166]. My findings that OP disturbances precede the appearance of vascular disease are consistent with data from humans, where electrophysiology is altered prior to visible microvascular disease, frequently involving the same specific waveforms [155], [167].

The ERG changes shown here are small but extremely specific to OPs. Unlike the large ERG effects caused by mutations in visual cycle proteins, ERG changes in diabetes are expected to be modest. Full-field ERG represents a summed response of the neural retina to light. However, just like in human diabetic retinopathy, I would not expect experimentally induced rodent retinopathy to be uniform. In humans, multifocal ERG identifies preferentially affected regions that later develop classic vascular lesions of retinopathy [93], [168], but unfortunately multifocal recording was not possible in rodents as of the time this work was performed. Compared to visual function assays, structural and functional assays of blood vessel integrity offer more specificity due to greater spatial resolution. Given the relatively lower resolution for full-field

ERG in mice, it is therefore meaningful that functional deficits correlated with glucose intolerance precede the detection of vascular defects, since I would surmise that a detection bias 57 would likely work in favor of histology, not ERG. It is also known that OP characteristics can lead to artifacts in b-wave recordings [169], [170], but I did not detect b-wave changes between groups, so this potential confounder is not an issue here.

Distorted OPs are believed to represent perturbations of the inner retina, specifically dysfunctional amacrine, bipolar, and retinal ganglion cells [171], [172]. Thus, my findings suggest that inner retinal networks may be important early targets of diabetic disease.

It is of course possible that early ERG dysfunction in humans and mice could be caused by microvascular disease that cannot be detected with current technologies. Alternatively, diet- induced diabetes may disturb neuroglial-vascular coupling. Neurovascular coupling is mediated by VEGF, semaphorin 3a, nitric oxide, and arachidonic acid metabolites, all of which are known to be affected by diabetes [173]–[176]. Nevertheless, my findings in this chapter are only correlative. Additional studies are required to demonstrate how high-fat diet-induced diabetes affects interactions between neurons and vasculature. I believe the high-fat diet mouse model will continue to be useful for studying the interactions between early neuronal, glial, and neurovascular changes in diabetic retinopathy. Doing so will be important, because finding new targets or biomarkers in the high-fat diet model may accelerate the improvement of clinical interventions for this debilitating disease.

Chapter 4: Summary and Future Directions 4.1 EAAT5 In Chapter 2, I demonstrated the relative contribution of EAAT5-mediated inhibition onto rod bipolar cells in response to physiologically relevant stimuli in retinal slice explants. In this setting, I also compared the roles of autoinhibition, in which excitation of the cell leads to inhibition onto itself, and alloinhibition, which is driven by other bipolar cells. Contrary to prior reports, I found that EAAT5-mediated inhibition can be elicited by dim light stimuli. I also found that while conventional inhibitors mediate nearly all alloinhibition onto rod bipolar cells, EAAT5 is the primary mediator of rod bipolar autoinhibition. Finally, in silico modeling demonstrated that EAAT5-mediated autoinhibition may have a substantial impact on postsynaptic excitation of the downstream AII amacrine cell.

I speculated that autoinhibition in the rod bipolar cell may specifically act as a gain modulator, allowing the cell to transmit sudden changes in signaling intensity, but acting as negative feedback to push the cell back towards an equilibrium in response to longer stimuli. In particular,

I hypothesized that EAAT-mediated inhibition onto rod bipolar cells may have the correct kinetics to enhance the relatively slow single-photon signal in the rod pathway at very low light intensities. While my data fit this hypothesis, it still requires experimental validation, and further work to investigate this question is warranted. Future experiments should present retinas with extremely dim stimuli in the presence and absence of EAAT5-mediated inhibition. If the hypothesis is correct, single-photon stimuli should cause a longer period of depolarization in the absence of EAAT5, such that the threshold for detection of an initial single-photon stimulus is

59 unchanged, but the downstream AII amacrine cell is better able to resolve sequential single- photon excitations with EAAT5-mediated inhibition at the rod bipolar cell intact.

Of course, this also raises the question of how the EAAT5 system responds to light adaptation more generally. While my work here did investigate responses to varying intensities of light, the stimuli presented here were in the context of a pure black background, on completely dark- adapted preparations. I have done some experiments (not shown) with light-adapted preparations, but these were grossly light-adapted into the low photopic range, where rod bipolar cells will be highly depolarized even at baseline and their Ca2+ channels will be largely inactivated, eliminating autoinhibition. I have not done any experiments in which rod bipolar cells were more subtly light-adapted, to around 0.1 photons/µm2/s, for example. Repeating under such conditions the kind of experiments I have performed here would be quite informative in investigating the role that EAAT5 may play in light adaptation.

All of these studies, and more, would also be greatly facilitated by the development of a bipolar- cell specific genetic knockout of EAAT5. During the time I have done the work presented here, the only available EAAT5 mutants were constitutive knockouts, and since EAAT5 is also present in photoreceptors, such mice were not useful for light-stimulus experiments. However, genetic technologies have progressed to the point where it should now be fairly straightforward to generate an EAAT5 selective knockout driven by a bipolar cell specific promoter. Such a model would provide a valuable check on my conclusions here, and also a valuable tool for further experiments. A genetic paradigm would also allow for more global assays of the effects of

EAAT5, including experiments to assess the role of EAAT5-mediated inhibition on visual parameters like visual sensitivity and contrast sensitivity, which can be assessed using optokinetic reflex or water maze behavioral testing or ganglion cell recording experiments. 60

Finally, while the findings of this work have elucidated the function of EAAT5 in the inner retina specifically, it is likely that it may have general implications for other neurotransmitter transporters that can act as ion channels elsewhere. In particular, the closely-related EAAT4 is already known to have a glutamate-gated chloride conductance similar to that of EAAT5 and is also known to be found in neuronal axon terminals – in this case in cerebellar Purkinje cells.

Future studies should investigate whether EAAT4 is playing the same functional role in the cerebellum that EAAT5 is in the inner retina.

4.2 Diabetic Retinopathy In Chapter 3, I characterized the course of electroretinographic dysfunction in a high-fat diet– induced model of diabetes in mice. Mice weaned to chow or to a high-fat diet were tested for electroretinographic dysfunction at 3, 6 and 12 months of age. There was no evidence of microvascular disease in these mice until 12 months. However, at as early as 6 months, high fat fed mice demonstrated increased latencies and reduced amplitudes of oscillatory potentials (a marker reflective of inner retinal neurons) when compared with controls. These electroretinographic abnormalities were correlated with glucose intolerance.

The results suggest that disease in the diabetic milieu may progress through neuronal stages in the inner retina long before the development of vascular lesions representing the classic hallmark of diabetic retinopathy. However, the neuronal types and the cellular and synaptic processes involved in the functional deficits are unknown. Identifying these components could possibly lead to novel interventions to treat eye disease.

When action potentials in the retina are blocked with tetrodotoxin, oscillatory potentials are reduced [177], [178]. This suggests that oscillatory potentials represent the contributions of

61 spiking neurons, meaning amacrine cells, ganglion cells, or both. Thus, action potentials may be compromised in amacrine or ganglion cells in the early stages of diabetic retinopathy. To determine whether this is the case, and, if so, which cells, future work should compare spiking activity in ganglion cells and spiking amacrine cells in high-fat diet and control mice. This should be feasible to do via single-cell patch-clamp recordings like those used here in Chapter 2.

If changes in innate spiking ability are observed, it should then be possible to further isolate the specific molecular changes involved by comparing the activation and inactivation properties of pharmacologically-isolated sodium and potassium channels in these cells.

Alternatively, it may be that altered inputs to amacrine or ganglion cells are responsible for the changes in ERGs I have observed here. For example, reduced excitatory input or increased inhibitory input to ganglion cells may lead to the reduced oscillatory potentials seen here. Future work should also investigate this possibility, by comparing the light-evoked excitatory or inhibitory post-synaptic currents onto ganglion cells from high-fat diet and control mice. From there, changes can be further narrowed down to changes in neurotransmitter release versus changes in receptor responsiveness by comparing the effects of paired pulse protocols versus the effects of spontaneous postsynaptic current recordings.

Of course, the question of whether or not the early electrical changes observed in these mice actually precede or coincide with changes in visual function also remains. In humans, data has demonstrated that changes in the ERG oscillatory potential precede measurable visual loss [90]–

[94]. If this is the case, the high-fat diet model would be most clinically useful if it also recapitulates this aspect of disease. Future work should assess the visual function of high-fat diet mice versus controls using visual parameters like contrast detection, spatial resolution, temporal

62 resolution, and light adaptation, which can all potentially be assessed using optokinetic reflex or water maze behavioral testing.

Finally, future work should investigate the possibility that electroretinographic testing could become more routinely used to screen for retinopathy in diabetic patients. First in patients, and now in mice, it has been shown that the oscillatory potential is a very early marker of disease progression that precedes the development of vascular changes in the retina. Clinically, however, routine screening is almost exclusively funduscopic, meaning that patients are not detected as having retinopathy until vascular changes have already begun. Perhaps because oscillatory potentials are a relatively obscure parameter compared to a-waves and b-waves, and because current treatments are geared toward halting vascular progression anyway, electroretinographic screening is not typically used. Indeed, it may not yet be practical to screen patients with ERGs in the general diabetic population, but researchers studying retinopathy therapeutics may benefit from ERG testing to identify the development of retinopathy in clinical study populations, which could, in itself, lead to the development of therapeutics that can be used in the earliest stages to prevent retinopathy progression. Thus, there is the potential for great clinical benefit if these findings can be leveraged to detect retinopathy in patients at an earlier stage.

References [1] E. D. Eggers, T. Ichinose, B. T. Sagdullaev, and P. D. Lukasiewicz, “Retinal GABA receptors and visual processing: a model system for presynaptic inhibition,” Cellscience Reviews, vol. 2, no. 3, Jan. 2006.

[2] R. H. Masland, “Neuronal diversity in the retina,” Current Opinion in Neurobiology, vol. 11, no. 4, pp. 431–436, Aug. 2001.

[3] E. D. Eggers and P. D. Lukasiewicz, “Interneuron Circuits Tune Inhibition in Retinal Bipolar Cells,” J Neurophysiol, vol. 103, no. 1, pp. 25–37, Jan. 2010.

[4] R. H. Masland, “The fundamental plan of the retina,” Nat Neurosci, vol. 4, no. 9, pp. 877– 886, Sep. 2001.

[5] S. Nirenberg and M. Meister, “The Light Response of Retinal Ganglion Cells Is Truncated by a Displaced Amacrine Circuit,” Neuron, vol. 18, no. 4, pp. 637–650, Apr. 1997.

[6] F. S. Werblin and J. E. Dowling, “Organization of the retina of the mudpuppy, Necturus maculosus. II. Intracellular recording.,” J Neurophysiol, vol. 32, no. 3, pp. 339–355, May 1969.

[7] B. B. Boycott, J. E. Dowling, and H. Kolb, “Organization of the Primate Retina: Light Microscopy,” Philos. Trans. R. Soc. Lond., B, Biol. Sci., vol. 255, no. 799, pp. 109–184, Mar. 1969.

[8] J. E. Dowling and B. B. Boycott, “Organization of the primate retina: electron microscopy,” Proc. R. Soc. Lond., B, Biol. Sci., vol. 166, no. 2, pp. 80–111, Nov. 1966.

[9] C. Enroth-Cugell and J. G. Robson, “The contrast sensitivity of retinal ganglion cells of the cat,” J Physiol, vol. 187, no. 3, pp. 517–552, Dec. 1966.

[10] H. B. Barlow, R. M. Hill, and W. R. Levick, “Retinal ganglion cells responding selectively to direction and speed of image motion in the rabbit,” J Physiol, vol. 173, no. 3, pp. 377– 407, Sep. 1964.

[11] S. W. Kuffler, “Discharge Patterns and Functional Organization of Mammalian Retina,” J Neurophysiol, vol. 16, no. 1, pp. 37–68, Jan. 1953.

[12] S. W. Kuffler, “Neurons in the Retina: Organization, Inhibition and Excitation Problems,” Cold Spring Harb Symp Quant Biol, vol. 17, pp. 281–292, Jan. 1952.

[13] H. Wässle, “Parallel processing in the mammalian retina,” Nat Rev Neurosci, vol. 5, no. 10, pp. 747–757, Oct. 2004.

[14] D. A. Baylor, M. G. F. Fuortes, and P. M. O’Bryan, “Receptive fields of cones in the retina of the turtle,” J Physiol, vol. 214, no. 2, pp. 265–294, Apr. 1971. 64

[15] J.-Z. Chuang, Y. Zhao, and C.-H. Sung, “SARA-Regulated Vesicular Targeting Underlies Formation of the Light-Sensing Organelle in Mammalian Rods,” Cell, vol. 130, no. 3, pp. 535–547, Aug. 2007.

[16] M. E. Burns and V. Y. Arshavsky, “Beyond Counting Photons: Trials and Trends in Vertebrate Visual Transduction,” Neuron, vol. 48, no. 3, pp. 387–401, Nov. 2005.

[17] M. M. Slaughter and R. F. Miller, “2-amino-4-phosphonobutyric Acid: A New Pharmacological Tool for Retina Research,” Science, vol. 211, no. 4478, pp. 182–185, Jan. 1981.

[18] K. K. Ghosh, S. Bujan, S. Haverkamp, A. Feigenspan, and H. Wässle, “Types of bipolar cells in the mouse retina,” The Journal of Comparative Neurology, vol. 469, no. 1, pp. 70– 82, 2004.

[19] S. H. DeVries, “Bipolar Cells Use Kainate and AMPA Receptors to Filter Visual Information into Separate Channels,” Neuron, vol. 28, no. 3, pp. 847–856, Dec. 2000.

[20] N. Vardi, R. Duvoisin, G. Wu, and P. Sterling, “Localization of mGluR6 to dendrites of ON bipolar cells in primate retina,” The Journal of Comparative Neurology, vol. 423, no. 3, pp. 402–412, 2000.

[21] P. D. Lukasiewicz, J. E. Lawrence, and T. L. Valentino, “Desensitizing glutamate receptors shape excitatory synaptic inputs to tiger salamander retinal ganglion cells,” J. Neurosci., vol. 15, no. 9, pp. 6189–6199, Sep. 1995.

[22] R. Nelson, E. V. Famiglietti, and H. Kolb, “Intracellular staining reveals different levels of stratification for on- and off-center ganglion cells in cat retina,” J Neurophysiol, vol. 41, no. 2, pp. 472–483, Mar. 1978.

[23] E. V. Famiglietti, A. Kaneko, and M. Tachibana, “Neuronal Architecture of On and Off Pathways to Ganglion Cells in Carp Retina,” Science, vol. 198, no. 4323, pp. 1267–1269, Dec. 1977.

[24] R. W. Rodieck, “Visual Pathways,” Annual Review of Neuroscience, vol. 2, no. 1, pp. 193– 225, 1979.

[25] B. B. Boycott and H. Wässle, “The morphological types of ganglion cells of the domestic cat’s retina,” J Physiol, vol. 240, no. 2, pp. 397–419, Jul. 1974.

[26] A. Nomura, R. Shigemoto, Y. Nakamura, N. Okamoto, N. Mizuno, and S. Nakanishi, “Developmentally regulated postsynaptic localization of a metabotropic glutamate receptor in rat rod bipolar cells,” Cell, vol. 77, no. 3, pp. 361–369, May 1994.

[27] J. H. Singer and J. S. Diamond, “Sustained Ca2+ Entry Elicits Transient Postsynaptic Currents at a Retinal Ribbon Synapse,” J. Neurosci., vol. 23, no. 34, pp. 10923–10933, Nov. 2003.

[28] F. Muller, H. Wassle, and T. Voigt, “Pharmacological modulation of the rod pathway in the cat retina,” J Neurophysiol, vol. 59, no. 6, pp. 1657–1672, Jun. 1988.

[29] E. Raviola and R. F. Dacheux, “Excitatory dyad synapse in rabbit retina,” PNAS, vol. 84, no. 20, pp. 7324–7328, Oct. 1987.

[30] R. F. Dacheux and E. Raviola, “The rod pathway in the rabbit retina: a depolarizing bipolar and amacrine cell,” J. Neurosci., vol. 6, no. 2, pp. 331–345, Feb. 1986.

[31] P. Sterling, M. A. Freed, and R. G. Smith, “Architecture of rod and cone circuits to the on- beta ganglion cell,” J. Neurosci., vol. 8, no. 2, pp. 623–642, Feb. 1988.

[32] E. V. Famiglietti Jr. and H. Kolb, “A bistratified amacrine cell and synaptic circuitry in the inner plexiform layer of the retina,” Brain Research, vol. 84, no. 2, pp. 293–300, Feb. 1975.

[33] M. R. Deans, B. Volgyi, D. A. Goodenough, S. A. Bloomfield, and D. L. Paul, “Connexin36 Is Essential for Transmission of Rod-Mediated Visual Signals in the Mammalian Retina,” Neuron, vol. 36, no. 4, pp. 703–712, Nov. 2002.

[34] U. Grünert and H. Wässle, “Immunocytochemical localization of glycine receptors in the mammalian retina,” The Journal of Comparative Neurology, vol. 335, no. 4, pp. 523–537, 1993.

[35] J. Zhang, C. S. Jung, and M. M. Slaughter, “Serial inhibitory synapses in retina,” Vis. Neurosci., vol. 14, no. 3, pp. 553–563, Jun. 1997.

[36] S. A. Bloomfield, D. Xin, and T. Osborne, “Light-induced modulation of coupling between AII amacrine cells in the rabbit retina,” Vis. Neurosci., vol. 14, no. 3, pp. 565–576, Jun. 1997.

[37] D. Xin and S. A. Bloomfield, “Tracer coupling pattern of amacrine and ganglion cells in the rabbit retina,” The Journal of Comparative Neurology, vol. 383, no. 4, pp. 512–528, 1997.

[38] B. Roska, E. Nemeth, and F. S. Werblin, “Response to Change Is Facilitated by a Three- Neuron Disinhibitory Pathway in the Tiger Salamander Retina,” J. Neurosci., vol. 18, no. 9, pp. 3451–3459, May 1998.

[39] P. B. Cook and J. S. McReynolds, “Lateral inhibition in the inner retina is important for spatial tuning of ganglion cells,” Nat. Neurosci., vol. 1, no. 8, pp. 714–719, Dec. 1998.

[40] C. Zhang and M. A. McCall, “Receptor targets of amacrine cells,” Visual Neuroscience, vol. 29, no. 01, pp. 11–29, 2012.

[41] S. L. Mills and S. C. Massey, “Differential properties of two gap junctional pathways made by AII amacrine cells,” Nature, vol. 377, no. 6551, pp. 734–737, Oct. 1995.

[42] R. Shapley and C. Enroth-Cugell, “Chapter 9 Visual adaptation and retinal gain controls,” Progress in Retinal Research, vol. 3, pp. 263–346, 1984. 66

[43] R. W. Rodieck and W. A. Rushton, “Cancellation of rod signals by cones, and cone signals by rods in the cat retina.,” J Physiol, vol. 254, no. 3, pp. 775–785, Jan. 1976.

[44] P. Gouras and K. Link, “Rod and cone interaction in dark-adapted monkey ganglion cells,” J Physiol, vol. 184, no. 2, pp. 499–510, May 1966.

[45] A. Stockman and L. T. Sharpe, “Into the twilight zone: the complexities of mesopic vision and luminous efficiency,” Ophthalmic and Physiological Optics, vol. 26, no. 3, pp. 225– 239, 2006.

[46] F. Naarendorp, Y. Sato, A. Cajdric, and N. P. Hubbard, “Absolute and relative sensitivity of the scotopic system of rat: Electroretinography and behavior,” Visual Neuroscience, vol. 18, no. 04, pp. 641–656, Jul. 2001.

[47] L. J. Frishman and P. A. Sieving, “Evidence for two sites of adaptation affecting the dark- adapted ERG of cats and primates,” Vision Research, vol. 35, no. 3, pp. 435–442, Feb. 1995.

[48] C. Enroth-Cugell and R. M. Shapley, “Adaptation and dynamics of cat retinal ganglion cells,” J Physiol, vol. 233, no. 2, pp. 271–309, Sep. 1973.

[49] H. B. Barlow and W. R. Levick, “Three factors limiting the reliable detection of light by retinal ganglion cells of the cat,” J Physiol, vol. 200, no. 1, pp. 1–24, Jan. 1969.

[50] J. E. Dowling, “The Site of Visual Adaptation,” Science, vol. 155, no. 3760, pp. 273–279, Jan. 1967.

[51] S. A. Bloomfield and R. F. Dacheux, “Rod Vision: Pathways and Processing in the Mammalian Retina,” Progress in Retinal and Eye Research, vol. 20, no. 3, pp. 351–384, May 2001.

[52] N. C. Danbolt, “Glutamate uptake,” Progress in Neurobiology, vol. 65, no. 1, pp. 1–105, Sep. 2001.

[53] S. Mager, J. Naeve, M. Quick, C. Labarca, N. Davidson, and H. A. Lester, “Steady states, charge movements, and rates for a cloned GABA transporter expressed in Xenopus oocytes,” Neuron, vol. 10, no. 2, pp. 177–188, Feb. 1993.

[54] S. Mager et al., “Conducting states of a mammalian serotonin transporter,” Neuron, vol. 12, no. 4, pp. 845–859, Apr. 1994.

[55] A. Galli, R. D. Blakely, and L. J. DeFelice, “Norepinephrine transporters have channel modes of conduction,” PNAS, vol. 93, no. 16, pp. 8671–8676, Aug. 1996.

[56] A. Galli, C. I. Petersen, M. deBlaquiere, R. D. Blakely, and L. J. DeFelice, “Drosophila Serotonin Transporters Have Voltage-Dependent Uptake Coupled to a Serotonin-Gated Ion Channel,” J. Neurosci., vol. 17, no. 10, pp. 3401–3411, May 1997.

[57] A. Galli, L. D. Jayanthi, I. S. Ramsey, J. W. Miller, R. T. Fremeau, and L. J. DeFelice, “l- Proline and l-Pipecolate Induce Enkephalin-Sensitive Currents in Human Embryonic Kidney 293 Cells Transfected with the High-Affinity Mammalian Brainl-Proline Transporter,” J. Neurosci., vol. 19, no. 15, pp. 6290–6297, Aug. 1999.

[58] S. L. Ingram and S. G. Amara, “Arachidonic Acid Stimulates a Novel Cocaine-Sensitive Cation Conductance Associated with the Human Dopamine Transporter,” J. Neurosci., vol. 20, no. 2, pp. 550–557, Jan. 2000.

[59] S. L. Ingram, B. M. Prasad, and S. G. Amara, “Dopamine transporter–mediated conductances increase excitability of midbrain dopamine neurons,” Nature Neuroscience, vol. 5, no. 10, pp. 971–978, Oct. 2002.

[60] M. W. Quick, “Regulating the Conducting States of a Mammalian Serotonin Transporter,” Neuron, vol. 40, no. 3, pp. 537–549, Oct. 2003.

[61] S. Risso, L. J. DeFelice, and R. D. Blakely, “Sodium-dependent GABA-induced currents in GAT1-transfected HeLa cells.,” J Physiol, vol. 490, no. Pt 3, pp. 691–702, Feb. 1996.

[62] Y. Kanai and M. A. Hediger, “The glutamate/neutral amino acid transporter family SLC1: molecular, physiological and pharmacological aspects,” Pflugers Arch - Eur J Physiol, vol. 447, no. 5, pp. 469–479, Feb. 2004.

[63] S. Eliasof and F. Werblin, “Characterization of the glutamate transporter in retinal cones of the tiger salamander,” J. Neurosci., vol. 13, no. 1, pp. 402–411, Jan. 1993.

[64] W. A. Fairman, R. J. Vandenberg, J. L. Arriza, M. P. Kavanaught, and S. G. Amara, “An excitatory amino-acid transporter with properties of a ligand-gated chloride channel,” Nature, vol. 375, no. 6532, pp. 599–603, Jun. 1995.

[65] J. L. Arriza, S. Eliasof, M. P. Kavanaugh, and S. G. Amara, “Excitatory amino acid transporter 5, a retinal glutamate transporter coupled to a chloride conductance,” PNAS, vol. 94, no. 8, pp. 4155–4160, Apr. 1997.

[66] J. L. Arriza, W. A. Fairman, J. I. Wadiche, G. H. Murdoch, M. P. Kavanaugh, and S. G. Amara, “Functional comparisons of three glutamate transporter subtypes cloned from human motor cortex,” J. Neurosci., vol. 14, no. 9, pp. 5559–5569, Sep. 1994.

[67] T. Storck, S. Schulte, K. Hofmann, and W. Stoffel, “Structure, expression, and functional analysis of a Na(+)-dependent glutamate/aspartate transporter from rat brain,” PNAS, vol. 89, no. 22, pp. 10955–10959, Nov. 1992.

[68] G. Pines et al., “Cloning and expression of a rat brain L-glutamate transporter,” Nature, vol. 360, no. 6403, pp. 464–467, Dec. 1992.

[69] Y. Kanai and M. A. Hediger, “Primary structure and functional characterization of a high- affinity glutamate transporter,” Nature, vol. 360, no. 6403, pp. 467–471, Dec. 1992.

[70] R. D. O’Shea, “Roles and regulation of glutamate transporters in the central nervous system,” Clinical and Experimental Pharmacology and Physiology, vol. 29, no. 11, pp. 1018–1023.

[71] Y. Dehnes, F. A. Chaudhry, K. Ullensvang, K. P. Lehre, J. Storm-Mathisen, and N. C. Danbolt, “The Glutamate Transporter EAAT4 in Rat Cerebellar Purkinje Cells: A Glutamate-Gated Chloride Channel Concentrated near the Synapse in Parts of the Dendritic Membrane Facing Astroglia,” J. Neurosci., vol. 18, no. 10, pp. 3606–3619, May 1998.

[72] A. Massie, F. Vandesande, and L. Arckens, “Expression of the high-affinity glutamate transporter Eaat4 in mammalian cerebral cortex,” Neuroreport, vol. 12, no. 2, pp. 393–397, Feb. 2001.

[73] D. V. Pow and N. L. Barnett, “Developmental expression of excitatory amino acid transporter 5: a photoreceptor and bipolar cell glutamate transporter in rat retina,” Neuroscience Letters, vol. 280, no. 1, pp. 21–24, Feb. 2000.

[74] R. M. Ryan and R. J. Vandenberg, “A channel in a transporter,” Clinical and Experimental Pharmacology and Physiology, vol. 32, no. 1–2, pp. 1–6, 2005.

[75] N. Zerangue and M. P. Kavanaugh, “Flux coupling in a neuronal glutamate transporter,” Nature, vol. 383, no. 6601, pp. 634–637, Oct. 1996.

[76] M. S. Sonders and S. G. Amara, “Channels in transporters,” Current Opinion in Neurobiology, vol. 6, no. 3, pp. 294–302, Jun. 1996.

[77] J. I. Wadiche, S. G. Amara, and M. P. Kavanaugh, “Ion fluxes associated with excitatory amino acid transport,” Neuron, vol. 15, no. 3, pp. 721–728, Sep. 1995.

[78] J. I. Wadiche and M. P. Kavanaugh, “Macroscopic and microscopic properties of a cloned glutamate transporter/chloride channel,” J. Neurosci., vol. 18, no. 19, pp. 7650–7661, Oct. 1998.

[79] B. Billups, D. Rossi, and D. Attwell, “Anion Conductance Behavior of the Glutamate Uptake Carrier in Salamander Retinal Glial Cells,” J. Neurosci., vol. 16, no. 21, pp. 6722– 6731, Nov. 1996.

[80] J. I. Wadiche, J. L. Arriza, S. G. Amara, and M. P. Kavanaugh, “Kinetics of a human glutamate transporter,” Neuron, vol. 14, no. 5, pp. 1019–1027, May 1995.

[81] R. J. Vandenberg, J. L. Arriza, S. G. Amara, and M. P. Kavanaugh, “Constitutive Ion Fluxes and Substrate Binding Domains of Human Glutamate Transporters,” J. Biol. Chem., vol. 270, no. 30, pp. 17668–17671, Jul. 1995.

[82] M. J. Palmer, H. Taschenberger, C. Hull, L. Tremere, and H. von Gersdorff, “Synaptic Activation of Presynaptic Glutamate Transporter Currents in Nerve Terminals,” J. Neurosci., vol. 23, no. 12, pp. 4831–4841, Jun. 2003.

[83] T. S. Otis, Y. C. Wu, and L. O. Trussell, “Delayed clearance of transmitter and the role of glutamate transporters at synapses with multiple release sites,” J. Neurosci., vol. 16, no. 5, pp. 1634–1644, Mar. 1996.

[84] D. Y. Tse, I. Chung, and S. M. Wu, “Pharmacological inhibitions of glutamate transporters EAAT1 and EAAT2 compromise glutamate transport in photoreceptor to ON-bipolar cell synapses,” Vision Research, vol. 103, pp. 49–62, Oct. 2014.

[85] T. S. Otis, M. P. Kavanaugh, and C. E. Jahr, “Postsynaptic Glutamate Transport at the Climbing Fiber-Purkinje Cell Synapse,” Science, vol. 277, no. 5331, pp. 1515–1518, Sep. 1997.

[86] D. V. Pow, N. L. Barnett, and P. Penfold, “Are neuronal transporters relevant in retinal glutamate homeostasis?,” Neurochemistry International, vol. 37, no. 2–3, pp. 191–198, Aug. 2000.

[87] R. N. Frank, “Diabetic Retinopathy,” New England Journal of Medicine, vol. 350, no. 1, pp. 48–58, 2004.

[88] Prevention of Blindness and Visual Impairment, “Global Data on Visual Impairments 2010,” World Health Organization. [Online]. Available: http://www.who.int/blindness/publications/globaldata/en/. [Accessed: 08-May-2018].

[89] R. Klein, B. E. K. Klein, S. E. Moss, M. D. Davis, and D. L. DeMets, “The Wisconsin Epidemiologic Study of Diabetic Retinopathy: III. Prevalence and Risk of Diabetic Retinopathy When Age at Diagnosis Is 30 or More Years,” Arch Ophthalmol, vol. 102, no. 4, pp. 527–532, Apr. 1984.

[90] D. Papakostopoulos, J. C. D. Hart, R. J. M. Corrall, and B. Harney, “The scotopic electroretinogram to blue flashes and pattern reversal visual evoked potentials in insulin dependent diabetes,” International Journal of Psychophysiology, vol. 21, no. 1, pp. 33–43, Jan. 1996.

[91] K. Holopigian, W. Seiple, M. Lorenzo, and R. Carr, “A comparison of photopic and scotopic electroretinographic changes in early diabetic retinopathy.,” Invest. Ophthalmol. Vis. Sci., vol. 33, no. 10, pp. 2773–2780, Sep. 1992.

[92] Y. Shirao and K. Kawasaki, “Electrical responses from diabetic retina,” Progress in Retinal and Eye Research, vol. 17, no. 1, pp. 59–76, Jan. 1998.

[93] A. M. Palmowski, E. E. Sutter, M. A. Bearse, and W. Fung, “Mapping of retinal function in diabetic retinopathy using the multifocal electroretinogram.,” Invest. Ophthalmol. Vis. Sci., vol. 38, no. 12, pp. 2586–2596, Nov. 1997.

[94] D. Yonemura, T. Aoki, and K. Tsuzuki, “Electroretinogram in Diabetic Retinopathy,” Arch Ophthalmol, vol. 68, no. 1, pp. 19–24, Jul. 1962.

[95] D. S. Fong et al., “Retinopathy in Diabetes,” Diabetes Care, vol. 27, no. suppl 1, pp. s84– s87, Jan. 2004.

[96] D. E. Singer, D. M. Nathan, H. A. Fogel, and A. P. Schachat, “Screening for diabetic retinopathy,” Ann. Intern. Med., vol. 116, no. 8, pp. 660–671, Apr. 1992.

[97] F. A. Dunn, T. Doan, A. P. Sampath, and F. Rieke, “Controlling the Gain of Rod-Mediated Signals in the Mammalian Retina,” J. Neurosci., vol. 26, no. 15, pp. 3959–3970, Apr. 2006.

[98] E. D. Eggers and P. D. Lukasiewicz, “GABAA, GABAC and glycine receptor-mediated inhibition differentially affects light-evoked signalling from mouse retinal rod bipolar cells,” J Physiol, vol. 572, no. 1, pp. 215–225, Apr. 2006.

[99] E. D. Eggers and P. D. Lukasiewicz, “Multiple pathways of inhibition shape bipolar cell responses in the retina,” Visual Neuroscience, vol. 28, no. 01, pp. 95–108, 2011.

[100] E. D. Eggers, M. A. McCall, and P. D. Lukasiewicz, “Presynaptic inhibition differentially shapes transmission in distinct circuits in the mouse retina,” J Physiol, vol. 582, no. 2, pp. 569–582, Jul. 2007.

[101] P. D. Lukasiewicz and F. S. Werblin, “A novel GABA receptor modulates synaptic transmission from bipolar to ganglion and amacrine cells in the tiger salamander retina,” J. Neurosci., vol. 14, no. 3, pp. 1213–1223, Mar. 1994.

[102] P. D. Lukasiewicz, B. R. Maple, and F. S. Werblin, “A novel GABA receptor on bipolar cell terminals in the tiger salamander retina,” J. Neurosci., vol. 14, no. 3, pp. 1202–1212, Mar. 1994.

[103] C.-J. Dong and F. S. Werblin, “Temporal Contrast Enhancement via GABAC Feedback at Bipolar Terminals in the Tiger Salamander Retina,” J Neurophysiol, vol. 79, no. 4, pp. 2171–2180, Apr. 1998.

[104] T. Euler and H. Wässle, “Different Contributions of GABAA and GABAC Receptors to Rod and Cone Bipolar Cells in a Rat Retinal Slice Preparation,” J Neurophysiol, vol. 79, no. 3, pp. 1384–1395, Mar. 1998.

[105] N. Menger, D. V. Pow, and H. Wässle, “Glycinergic amacrine cells of the rat retina,” The Journal of Comparative Neurology, vol. 401, no. 1, pp. 34–46, 1998.

[106] P. D. Lukasiewicz, E. D. Eggers, B. T. Sagdullaev, and M. A. McCall, “GABAC receptor-mediated inhibition in the retina,” Vision Research, vol. 44, no. 28, pp. 3289– 3296, Dec. 2004.

[107] A. E. Chávez, W. N. Grimes, and J. S. Diamond, “Mechanisms Underlying Lateral GABAergic Feedback onto Rod Bipolar Cells in Rat Retina,” J. Neurosci., vol. 30, no. 6, pp. 2330–2339, Feb. 2010.

[108] W. N. Grimes, J. Zhang, C. W. Graydon, B. Kachar, and J. S. Diamond, “Retinal Parallel Processors: More than 100 Independent Microcircuits Operate within a Single Interneuron,” Neuron, vol. 65, no. 6, pp. 873–885, Mar. 2010.

[109] N. Schneider, S. Cordeiro, J.-P. Machtens, S. Braams, T. Rauen, and C. Fahlke, “Functional Properties of the Retinal Glutamate Transporters GLT-1c and EAAT5,” J. Biol. Chem., vol. 289, no. 3, pp. 1815–1824, Jan. 2014.

[110] E. D. Eggers and P. D. Lukasiewicz, “Receptor and Transmitter Release Properties Set the Time Course of Retinal Inhibition,” J. Neurosci., vol. 26, no. 37, pp. 9413–9425, Sep. 2006.

[111] T. Ichinose and P. D. Lukasiewicz, “The Mode of Retinal Presynaptic Inhibition Switches with Light Intensity,” J. Neurosci., vol. 32, no. 13, pp. 4360–4371, Mar. 2012.

[112] E. Wersinger et al., “The glutamate transporter EAAT5 works as a presynaptic receptor in mouse rod bipolar cells,” J Physiol, vol. 577, no. 1, pp. 221–234, Nov. 2006.

[113] M. L. Veruki, S. H. Mørkve, and E. Hartveit, “Activation of a presynaptic glutamate transporter regulates synaptic transmission through electrical signaling,” Nat Neurosci, vol. 9, no. 11, pp. 1388–1396, Nov. 2006.

[114] T. Euler and R. H. Masland, “Light-Evoked Responses of Bipolar Cells in a Mammalian Retina,” J Neurophysiol, vol. 83, no. 4, pp. 1817–1829, Apr. 2000.

[115] B. T. Sagdullaev, E. D. Eggers, R. Purgert, and P. D. Lukasiewicz, “Nonlinear Interactions between Excitatory and Inhibitory Retinal Synapses Control Visual Output,” J. Neurosci., vol. 31, no. 42, pp. 15102–15112, Oct. 2011.

[116] K. K. Ghosh, S. Bujan, S. Haverkamp, A. Feigenspan, and H. Wässle, “Erratum: Types of bipolar cells in the mouse retina.,” The Journal of Comparative Neurology, vol. 476, no. 2, pp. 202–203, 2004.

[117] V. Pignatelli and E. Strettoi, “Bipolar cells of the mouse retina: A gene gun, morphological study,” J. Comp. Neurol., vol. 476, no. 3, pp. 254–266, Aug. 2004.

[118] R. G. Smith, “NeuronC: a computational language for investigating functional architecture of neural circuits,” Journal of Neuroscience Methods, vol. 43, no. 2, pp. 83– 108, Jul. 1992.

[119] Y. Shigeri et al., “Effects of threo-β-hydroxyaspartate derivatives on excitatory amino acid transporters (EAAT4 and EAAT5),” Journal of Neurochemistry, vol. 79, no. 2, pp. 297–302, 2001.

[120] L. T. Sharpe and A. Stockman, “Rod pathways: the importance of seeing nothing,” Trends in Neurosciences, vol. 22, no. 11, pp. 497–504, Nov. 1999.

[121] E. B. Trexler, W. Li, and S. C. Massey, “Simultaneous Contribution of Two Rod Pathways to AII Amacrine and Cone Bipolar Cell Light Responses,” Journal of Neurophysiology, vol. 93, no. 3, pp. 1476–1485, Mar. 2005.

[122] N. W. Oesch and J. S. Diamond, “Ribbon synapses compute temporal contrast and encode luminance in retinal rod bipolar cells,” Nature Neuroscience, vol. 14, no. 12, p. 1555, Dec. 2011.

[123] C.-J. Dong and W. A. Hare, “Temporal Modulation of Scotopic Visual Signals by A17 Amacrine Cells in Mammalian Retina In Vivo,” Journal of Neurophysiology, vol. 89, no. 4, pp. 2159–2166, Apr. 2003.

[124] W. N. Grimes et al., “Complex inhibitory microcircuitry regulates retinal signaling near visual threshold,” Journal of Neurophysiology, vol. 114, no. 1, pp. 341–353, May 2015.

[125] P. Kugler and A. Beyer, “Expression of glutamate transporters in human and rat retina and rat optic nerve,” Histochem Cell Biol, vol. 120, no. 3, pp. 199–212, Sep. 2003.

[126] T. Rauen, W. R. Taylor, K. Kuhlbrodt, and M. Wiessner, “High-affinity glutamate transporters in the rat retina: a major role of the glial glutamate transporter GLAST-1 in transmitter clearance,” Cell Tissue Res., vol. 291, no. 1, pp. 19–31, Jan. 1998.

[127] T. Harada et al., “Functions of the two glutamate transporters GLAST and GLT-1 in the retina,” PNAS, vol. 95, no. 8, pp. 4663–4666, Apr. 1998.

[128] T. Rauen, J. D. Rothstein, and H. Wässle, “Differential expression of three glutamate transporter subtypes in the rat retina,” Cell Tissue Res., vol. 286, no. 3, pp. 325–336, Dec. 1996.

[129] “Diabetic Retinopathy | National Eye Institute.” [Online]. Available: https://www.nei.nih.gov/eyedata/diabetic. [Accessed: 14-May-2018].

[130] L. P. Aiello et al., “Vascular Endothelial Growth Factor in Ocular Fluid of Patients with Diabetic Retinopathy and Other Retinal Disorders,” http://dx.doi.org/10.1056/NEJM199412013312203, Jan. 2010.

[131] N. Ferrara, L. Damico, N. Shams, H. Lowman, and R. Kim, “Development Of Ranibizumab, An Anti–vascular Endothelial Growth Factor Antigen Binding Fragment, As Therapy For Neovascular Age-related Macular Degeneration,” Retina, vol. 26, no. 8, pp. 859–870, Oct. 2006.

[132] D. V. Do et al., “Ranibizumab for Edema of the Macula in Diabetes Study: 3-Year Outcomes and the Need for Prolonged Frequent Treatment,” JAMA Ophthalmol, vol. 131, no. 2, pp. 139–145, Feb. 2013.

[133] T. Kurihara, P. D. Westenskow, S. Bravo, E. Aguilar, and M. Friedlander, “Targeted deletion of Vegfa in adult mice induces vision loss,” J Clin Invest, vol. 122, no. 11, pp. 4213–4217, Nov. 2012. 73

[134] M. C. Gillies et al., “Intravitreal Triamcinolone Prior to Laser Treatment of Diabetic Macular Edema: 24-Month Results of a Randomized Controlled Trial,” Ophthalmology, vol. 118, no. 5, pp. 866–872, May 2011.

[135] E. L. Fletcher, J. A. Phipps, M. M. Ward, T. Puthussery, and J. L. Wilkinson-Berka, “Neuronal and glial cell abnormality as predictors of progression of diabetic retinopathy,” Curr. Pharm. Des., vol. 13, no. 26, pp. 2699–2712, 2007.

[136] L. Wachtmeister, “Oscillatory potentials in the retina: what do they reveal,” Progress in Retinal and Eye Research, vol. 17, no. 4, pp. 485–521, Oct. 1998.

[137] M. J. Gastinger, R. S. J. Singh, and A. J. Barber, “Loss of Cholinergic and Dopaminergic Amacrine Cells in Streptozotocin-Diabetic Rat and Ins2Akita-Diabetic Mouse Retinas,” IOVS, vol. 47, no. 7, pp. 3143–3150, Jul. 2006.

[138] M. J. Gastinger, A. R. Kunselman, E. E. Conboy, S. K. Bronson, and A. J. Barber, “Dendrite Remodeling and Other Abnormalities in the Retinal Ganglion Cells of Ins2Akita Diabetic Mice,” Invest. Ophthalmol. Vis. Sci., vol. 49, no. 6, pp. 2635–2642, Jun. 2008.

[139] R. Robinson, V. A. Barathi, S. S. Chaurasia, T. Y. Wong, and T. S. Kern, “Update on animal models of diabetic retinopathy: from molecular approaches to mice and higher mammals,” Disease Models & Mechanisms, vol. 5, no. 4, pp. 444–456, Jul. 2012.

[140] A. J. Barber et al., “The Ins2Akita Mouse as a Model of Early Retinal Complications in Diabetes,” Invest. Ophthalmol. Vis. Sci., vol. 46, no. 6, pp. 2210–2218, Jun. 2005.

[141] E. Midena et al., “Studies on the Retina of the Diabetic db/db Mouse,” ORE, vol. 21, no. 2, pp. 106–111, 1989.

[142] J. Babel and P. Leuenberger, “A long term study on the ocular lesions in streptozotocin diabetic rats,” Albrecht Von Graefes Arch Klin Exp Ophthalmol, vol. 189, no. 3, pp. 191– 209, Jan. 1974.

[143] S. McLenachan, X. Chen, P. G. McMenamin, and E. P. Rakoczy, “Absence of clinical correlates of diabetic retinopathy in the Ins2Akita retina,” Clinical & Experimental Ophthalmology, vol. 41, no. 6, pp. 582–592, Aug. 2013.

[144] X.-F. Zhao, J. Wan, C. Powell, R. Ramachandran, M. G. Myers, and D. Goldman, “Leptin and IL-6 Family Cytokines Synergize to Stimulate Müller Glia Reprogramming and Retina Regeneration,” Cell Reports, vol. 9, no. 1, pp. 272–284, Oct. 2014.

[145] Y. D. Muller et al., “Immunosuppressive Effects of Streptozotocin-Induced Diabetes Result in Absolute Lymphopenia and a Relative Increase of T Regulatory Cells,” Diabetes, vol. 60, no. 9, pp. 2331–2340, Sep. 2011.

[146] D. J. Perry, “Nephrotoxicity of Streptozocin,” Annals of Internal Medicine, vol. 96, no. 1, p. 122, Jan. 1982.

[147] E. Kume et al., “Hepatic changes in the acute phase of streptozotocin (SZ)-induced diabetes in mice,” Experimental and Toxicologic Pathology, vol. 55, no. 6, pp. 467–480, Jan. 2004.

[148] A. Kraska et al., “In Vivo Cross-sectional Characterization of Cerebral Alterations Induced by Intracerebroventricular Administration of Streptozotocin,” PLOS ONE, vol. 7, no. 9, p. e46196, Sep. 2012.

[149] S. Rai, P. K. Kamat, C. Nath, and R. Shukla, “Glial activation and post-synaptic neurotoxicity: The key events in Streptozotocin (ICV) induced memory impairment in rats,” Pharmacology Biochemistry and Behavior, vol. 117, pp. 104–117, Feb. 2014.

[150] S. Shoham, C. Bejar, E. Kovalev, and M. Weinstock, “Intracerebroventricular injection of streptozotocin causes neurotoxicity to myelin that contributes to spatial memory deficits in rats,” Experimental Neurology, vol. 184, no. 2, pp. 1043–1052, Dec. 2003.

[151] D. A. Towler, M. Bidder, T. Latifi, T. Coleman, and C. F. Semenkovich, “Diet-induced Diabetes Activates an Osteogenic Gene Regulatory Program in the Aortas of Low Density Lipoprotein Receptor-deficient Mice,” J. Biol. Chem., vol. 273, no. 46, pp. 30427–30434, Nov. 1998.

[152] R. Rajagopal, G. W. Bligard, S. Zhang, L. Yin, P. Lukasiewicz, and C. F. Semenkovich, “Functional Deficits Precede Structural Lesions in Mice With High-Fat Diet–Induced Diabetic Retinopathy,” Diabetes, vol. 65, no. 4, pp. 1072–1084, Apr. 2016.

[153] Q. Xu, T. Qaum, and A. P. Adamis, “Sensitive Blood–Retinal Barrier Breakdown Quantitation Using Evans Blue,” Invest. Ophthalmol. Vis. Sci., vol. 42, no. 3, pp. 789–794, Mar. 2001.

[154] M. A. Bearse et al., “A multifocal electroretinogram model predicting the development of diabetic retinopathy,” Progress in Retinal and Eye Research, vol. 25, no. 5, pp. 425–448, Sep. 2006.

[155] Bresnick GH, Korth K, Groo A, and Palta M, “Electroretinographic oscillatory potentials predict progression of diabetic retinopathy: Preliminary report,” Arch Ophthalmol, vol. 102, no. 9, pp. 1307–1311, Sep. 1984.

[156] S. E. Simonsen, “The value of the oscillatory potential in selecting juvenile diabetics at risk of developing proliferative retinopathy,” Acta Ophthalmol (Copenh), vol. 58, no. 6, pp. 865–878, Dec. 1980.

[157] M. Wong-Riley, “Energy metabolism of the visual system,” Eye Brain, vol. 2, pp. 99– 116, Jul. 2010.

[158] H. A. Hancock and T. W. Kraft, “Oscillatory Potential Analysis and ERGs of Normal and Diabetic Rats,” Invest. Ophthalmol. Vis. Sci., vol. 45, no. 3, pp. 1002–1008, Mar. 2004.

[159] K. Kohzaki, A. J. Vingrys, and B. V. Bui, “Early Inner Retinal Dysfunction in Streptozotocin-Induced Diabetic Rats,” Invest. Ophthalmol. Vis. Sci., vol. 49, no. 8, pp. 3595–3604, Aug. 2008.

[160] Q. Li, E. Zemel, B. Miller, and I. Perlman, “Early Retinal Damage in Experimental Diabetes: Electroretinographical and Morphological Observations,” Experimental Eye Research, vol. 74, no. 5, pp. 615–625, May 2002.

[161] P. Sapieha et al., “Omega-3 polyunsaturated fatty acids preserve retinal function in type 2 diabetic mice,” Nutrition & Diabetes, vol. 2, no. 7, p. e36, Jul. 2012.

[162] J. E. Mancini, G. Ortiz, J. O. Croxatto, and J. E. Gallo, “Retinal upregulation of inflammatory and proangiogenic markers in a model of neonatal diabetic rats fed on a high- fat-diet,” BMC Ophthalmology, vol. 13, p. 14, Apr. 2013.

[163] A. C. Marçal et al., “Diet-induced obesity impairs AKT signalling in the retina and causes retinal degeneration,” Cell Biochem. Funct., vol. 31, no. 1, pp. 65–74, Jan. 2013.

[164] K. Noda, S. Nakao, S. Zandi, D. Sun, K. C. Hayes, and A. Hafezi-Moghadam, “Retinopathy in a novel model of metabolic syndrome and type 2 diabetes: new insight on the inflammatory paradigm,” The FASEB Journal, vol. 28, no. 5, pp. 2038–2046, Feb. 2014.

[165] S. A. Elhabashy, N. S. Elbarbary, K. M. Nageb, and M. M. Mohammed, “Can optical coherence tomography predict early retinal microvascular pathology in type 1 diabetic adolescents without minimal diabetic retinopathy? A single-centre study,” Journal of Pediatric Endocrinology and Metabolism, vol. 28, no. 1–2, pp. 139–146, 2014.

[166] S. Vujosevic and E. Midena, “Retinal Layers Changes in Human Preclinical and Early Clinical Diabetic Retinopathy Support Early Retinal Neuronal and Müller Cells Alterations,” Journal of Diabetes Research, 2013. [Online]. Available: https://www.hindawi.com/journals/jdr/2013/905058/. [Accessed: 20-May-2018].

[167] H. Asi and I. Perlman, “Relationships between the electroretinogram a-wave, b-wave and oscillatory potentials and their application to clinical diagnosis,” Doc Ophthalmol, vol. 79, no. 2, pp. 125–139, 1992.

[168] Y. Han, M. A. Bearse, M. E. Schneck, S. Barez, C. H. Jacobsen, and A. J. Adams, “Multifocal Electroretinogram Delays Predict Sites of Subsequent Diabetic Retinopathy,” Invest. Ophthalmol. Vis. Sci., vol. 45, no. 3, pp. 948–954, Mar. 2004.

[169] C. J. Layton, R. Safa, and N. N. Osborne, “Oscillatory potentials and the b-Wave: partial masking and interdependence in dark adaptation and diabetes in the rat,” Graefes Arch. Clin. Exp. Ophthalmol., vol. 245, no. 9, pp. 1335–1345, Sep. 2007.

[170] M. H. Sims, “Partial masking of the canine electroretinogram by oscillatory potentials. The problem of frequency bandwidth,” J. Vet. Intern. Med., vol. 4, no. 1, pp. 40–42, Feb. 1990.

[171] H. Heynen, L. Wachtmeister, and D. van Norren, “Origin of the oscillatory potentials in the primate retina,” Vision Research, vol. 25, no. 10, pp. 1365–1373, Jan. 1985.

[172] L. Wachtmeister and J. E. Dowling, “The oscillatory potentials of the mudpuppy retina.,” IOVS, vol. 17, no. 12, pp. 1176–1188, Dec. 1978.

[173] D. Attwell, A. M. Buchan, S. Charpak, M. Lauritzen, B. A. MacVicar, and E. A. Newman, “Glial and neuronal control of brain blood flow,” Nature, vol. 468, no. 7321, pp. 232–243, Nov. 2010.

[174] A. Cerani et al., “Neuron-Derived Semaphorin 3A Is an Early Inducer of Vascular Permeability in Diabetic Retinopathy via Neuropilin-1,” Cell Metabolism, vol. 18, no. 4, pp. 505–518, Oct. 2013.

[175] A. Mishra and E. A. Newman, “Inhibition of inducible nitric oxide synthase reverses the loss of functional hyperemia in diabetic retinopathy,” Glia, vol. 58, no. 16, pp. 1996–2004, Dec. 2010.

[176] K. Okabe et al., “Neurons Limit Angiogenesis by Titrating VEGF in Retina,” Cell, vol. 159, no. 3, pp. 584–596, Oct. 2014.

[177] S. Viswanathan, L. J. Frishman, and J. G. Robson, “The Uniform Field and Pattern ERG in Macaques with Experimental Glaucoma: Removal of Spiking Activity,” Invest. Ophthalmol. Vis. Sci., vol. 41, no. 9, pp. 2797–2810, Aug. 2000.

[178] C.-J. Dong, P. Agey, and W. A. Hare, “Origins of the electroretinogram oscillatory potentials in the rabbit retina,” Visual Neuroscience, vol. 21, no. 04, pp. 533–543, 2004.