ROLE of ARGINASE 2 by Esraa Farid Aly Shosha October

NEUROVASCULAR DEGENERATION FOLLOWING RETINAL ISCHEMIA REPERFUSION INJURY: ROLE OF ARGINASE 2

Esraa Farid Aly Shosha

Submitted to the Faculty of the Graduate School

of Augusta University in partial fulfillment

of the Requirements of the Degree of

(Doctor of Philosophy)

October

2017

ACKNOWLEDGEMENTS

I would like to thank my mentor, Dr. Ruth Caldwell, for giving me the opportunity to join her lab and pursue my studies under her guidance. It was a great pleasure to work under her mentorship. Dr. Caldwell not only teaches you how to do research, but sets a role model of a great scientist and human being. I am blessed to be mentored by her.

I would like to extend my gratitude to the current lab members: Ms. Tahira Lemtalsi, Dr.

Abdelrahman Fouda, Ms. Zhimin Xu, Dr. Modesto Rojas and a previous lab member, Dr.

Chintan Patel. I also would like to thank Dr. Priya Narayanan for her continuous guidance and help during my period of studies. Dr. Narayanan and Ms. Lemtalsi, thank you for being my family in Augusta.

I would like to acknowledge my committee members for their time and scientific contribution: Dr. David Fulton, Dr. William Caldwell, Dr. Mohamed Al-Shabrawey and

Dr. Priya Narayanan. I also would like to thank the thesis reader, Dr. Anilkumar Pillai.

This work wouldn’t have been accomplished without the help of our collaborators: Dr.

William Caldwell’s lab, Dr. Mohamed Al-Shabrawey’s lab and Dr. David Fulton’s lab.

I also would like to thank the vascular biology center (VBC) graduate program for providing an intense training environment. I specially thank Dr. Rudolf Lucas for his great support and help. Thanks to the staff of the graduate school for the great help.

I dedicate this thesis to my mother (Mrs. Mervat Ahmed) and my father (Dr. Farid Shosha) for having faith in me, raising me the way I am and encouraging me in times of difficulties throughout my life. I also would like to extend my sincere gratitude to my brother (Dr.

Mohamed Shosha) and my sister (Dr. Alaa Shosha) for your invaluable support and humor.

Special thanks and dedication go to my small family for their sacrifice along the way to make this dream come true. To my dear husband (Dr. Abdelrahman Fouda), for your love, patience, friendship and support. To my precious daughters (Alia and Farida), you are the inspiration for me to keep going on.

ABSTRACT

ESRAA SHOSHA Neurovascular Degeneration Following Retinal Ischemia Reperfusion Injury: Role of Arginase 2 (Under the direction of RUTH B. CALDWELL)

Ischemic retinopathies such as retinopathy of prematurity, central retinal artery occlusion and diabetic retinopathy are leading causes of visual impairment and blindness.

These pathologies share common features of oxidative stress, activation of inflammatory pathways and neurovascular damage. There is no clinically effective treatment for these conditions because the underlying mechanisms are still not fully understood. In the current study, we used a mouse model of retinal ischemia reperfusion (I/R) insult to explore the underlying mechanisms of neurovascular degeneration in ischemic retinopathies. The arginase enzyme utilizes the L-arginine amino acid for the production of L-ornithine and urea. Here, we investigated the role of the mitochondrial arginase isoform, arginase 2 (A2) in retinal I/R induced neurovascular injury. We found that retinal I/R induced neurovascular degeneration, superoxide and nitrotyrosine formation, glial activation, cell death by necroptosis and impairment of inner retinal function in wild type (WT) mice. A2 homozygous deletion (A2-/-) significantly protected against the neurovascular degeneration after retinal I/R. That was attributed to decreased oxidative stress and glial activation. A2 deletion protected against I/R induced retinal function impairment. Using

Optical coherence tomography (OCT), we evaluated the retinal structure in live animals and found that A2-/- retinas showed a more preserved structure and less retinal detachment.

To investigate the underlying mechanisms of A2 induced vascular damage after I/R, we used an in vitro model of oxygen glucose deprivation/ reperfusion (OGD/R) in bovine retinal endothelial cells (BRECs). Analysis of oxidative metabolism showed impaired mitochondrial function. We also found an increase in dynamin elated protein 1 (Drp1), a mitochondrial fission marker. Mitochondria labeling studies showed fragmented mitochondria after OGD/R. Arginase inhibition reduced mitochondrial fragmentation in

OGD/R insult.

This dissertation presents A2 as a new therapeutic target in reducing neurovascular damage in ischemic retinopathies.

KEY WORDS: retinal ischemia, arginase, neurovascular degeneration, mitochondria

Table of Contents

Chapter Page

1 Chapter I: Introduction ...... 3

A) Statement of the problem and specific aims of the overall project ...... 3

B) Literature review ...... 6

C) Rationale ...... 25

2 Chapter II: Arginase 2 Promotes Neurovascular Degeneration During

Ischemia/Reperfusion Injury (published manuscript) ...... 27

3 Chapter III: Mechanisms of Retinal Ischemia/Reperfusion Injury: Arginase 2 and the Mitochondria (unpublished data) ...... 62

4 Chapter IV: Discussion ...... 85

5 Chapter V: Summary ...... 93

References ...... 96

List of Tables

Table 1: List of abbreviations used in the text ...... 1

Table 2: Time points of mice sacrifice after I/R and rationale for each ...... 54

List of Figures

Figure 1: Summary of the proposed study ...... 5

Figure 2: Scheme for L-arginine catabolism ...... 6

Figure 3: Scheme for polyamines production ...... 7

Figure 4: Oxygen induced retinopathy (OIR) mouse model ...... 14

Figure 5: Schematic diagram for retinal structure ...... 17

Figure 6: Retinal blood supply ...... 18

Figure 7: Retinal ischemia reperfusion (I/R) mouse model ...... 21

Figure 8: Increased expression of arginase 2 following I/R injury ...... 33

Figure 9: A2 deletion prevented I/R injury-induced retinal thinning ...... 35

Figure 10: A2 deletion reduced I/R injury-induced necroptic cell death ...... 37

Figure 11: A2 deletion prevented I/R injury-induced loss of ganglion cell layer neurons ...... 38

Figure 12: A2 deletion prevented I/R injury-induced microvascular degeneration ...... 40

Figure 13: A2 deletion limited I/R injury-induced glial activation ...... 41

Figure 14: A2 deletion prevented I/R injury-induced oxidative stress ...... 44

Figure 15: A2 deletion prevented I/R injury-induced impairment of inner retina function ...... 46

Figure 16: Possible mechanisms of neurovascular degeneration after retinal I/R ...... 48

Figure 17: A2 deletion reduces I/R injury-induced increases in Drp1 ...... 68

Figure 18: A2 expression and cell stress response increase in endothelial cells after

OGD/R ...... 69

Figure 19: Arginase inhibition protects against OGD/R induced mitochondrial fission ...71

Figure 20: OGD/R induces mitochondrial dysfunction ...... 73

Figure 21: A2 deletion preserves retinal structure and prevents retinal detachment ...... 75

Figure 22: Summary of the role of A2 in inducing neurovascular degeneration following retinal I/R injury ...... 95

Table 1

Abbreviation Term A1 Arginase 1 A2 Arginase 2 A2-/- A2 Homozygous Knockout ABH 2-(S)-Amino-6-Boronohexanoic Acid AD Alzheimer’s Disease ADMA Asymmetric Dimethylarginine APAO N(1)-acetylpolyamine oxidase BDNF Brain-Derived Neurotrophic Factor BH4 Tetrahydrobiopterin BRECs Bovine Retinal Endothelial Cells BSA Bovine Serum Albumin CNS Central Nervous System CRAO Central Retinal Artery Occlusion DR Diabetic Retinopathy Drp1 Dynamin Related Protein 1 eNOS or NOS III Endothelial Nitric Oxide Synthase ERK Extracellular Signal-Regulated Kinase FBS Fetal Bovine Serum GCL Ganglion Cell Layer INL Inner Nuclear Layer iNOS or NOS II Inducible Nitric Oxide Synthase IOP Intraocular Pressure IPL Inner Plexiform Layer LPS Lipopolysaccharide MCP-1 Monocyte Chemotactic Protein 1 Mff Mitochondrial Fission Factor Mfn1 Mitofusion 1 Mfn2 Mitofusion 2

Table 1 (Continuation)

Abbreviation Term MDL [N,N’-bis(2,3-butadienyl)-1,4 butanediamine]; MiD 49 Mitochondrial Dynamics Proteins Of 49 MiD51 Mitochondrial Dynamics Proteins Of 51 MPK Mitogen Activated Protein Kinase MPT, MPTP Mitochondrial Permeability Transition Pore NFL Nerve Fiber Layer nNOS or NOS I Neuronal Nitric Oxide Synthase NO Nitric Oxide NOS Nitric Oxide Synthase OAT Ornithine Amino Transferase OCR Oxygen Consumption Rate ODC Ornithine Decarboxylase OGD/R Oxygen Glucose Deprivation/Reperfusion OIR Oxygen Induced Retinopathy ONL Outer Nuclear Layer OPA1 Optic Atrophy1 OPL Outer Plexiform Layer ROP Retinopathy Of Prematurity ROS Reactive Oxygen Species SD-OCT Spectral Domain Optical Coherence Tomography SMO Spermine Oxidase SSAT Spermine Spermidine Acetyl Transferase TBI Traumatic Brain Injury TNFα Tumor necrosis factor alpha WT Wild Type

Chapter I:

INTRODUCTION

A. Statement of the Problem and specific aims of the overall project

Vision loss or impairment is among the top causes of disabilities in adults aged 18 or older. Ischemic retinopathies are a major cause of visual impairment and blindness. They include diabetic retinopathy (DR), acute glaucoma, retinopathy of prematurity (ROP) and central retinal artery occlusion (CRAO). According to National Eye Institute, the number of people affected by diabetic retinopathy and glaucoma will double from 2010 to 2050.

Visual impairment and blindness cause a considerable economic burden for affected persons, their caregivers and society at large, which increases with the degree of visual impairment for all assessed cost categories as well as intangible effects (Koberlein et al.,

2013) . A study estimated a total financial burden of major visual disorders of $35.4 billion comprised of $16.2 billion in direct medical costs, $11.1 billion in other direct costs, and

$8 billion in productivity losses (Rein et al., 2006). Until now, there is no clinically effective treatment for ischemic retinopathies because the mechanisms by which they happen are far from clear.

Arginase, the ureahydrolase enzyme, is implicated in vascular dysfunction in cardiovascular diseases. Arginase hydrolyzes the semi essential amino acid L-arginine for the production of ornithine and urea. Nitric oxide synthase (NOS) utilizes the same substrate for the production of L-citrulline and nitric oxide (NO). Arginase competes with

NOS for its substrate which limits nitric oxide availability and can cause NOS uncoupling.

Arginase has two isoforms; arginase 1 (A1) is the cytosolic isoform and arginase 2 (A2) is the mitochondrial isoform. Both isoforms are expressed in the retina, in different retinal layers and cells. Our group has shown that arginase activity and expression are increased in diabetic models and in the oxygen induced retinopathy (OIR) model of retinopathy of prematurity. Arginase inhibition or deletion protected against diabetes induced retinal vasorelaxation impairment and increased oxidative stress. Deletion of A2 significantly protected against neuronal and vascular degeneration in the OIR model. In the current study we utilized a mouse model of retinal ischemia reperfusion (I/R) insult to study the involvement of A2 in neurovascular degeneration injury. Previous studies have also reported the involvement of mitochondrial dysfunction in I/R injury. Hence, we hypothesized that A2 promotes neurovascular degeneration in retinal ischemia reperfusion injury via a mechanism involving increased oxidative stress, glial activation and mitochondrial dysfunction (Fig.1). We studied this through the following two aims:

Aim 1: Test the hypothesis that I/R causes neurovascular degeneration through upregulation of A2

Aim 2: Test the hypothesis that targeting arginase 2 mitigates neurovascular degeneration through suppressing glial activation, oxidative stress and mitochondrial dysfunction

Figure 1: Summary of the proposed study In the current study, we propose to investigate the effect of retinal I/R on arginase 2 (A2) expression and the cellular sources involved. We will also investigate the role of A2 and effect of A2 deletion in inducing cell death and neurovascular injury following retinal I/R. Question marks point to the aims proposed to be investigated in this study

B. Literature review

B.1 Arginase and L-arginine metabolism

B.1.1. Arginase enzyme

Arginase is a Binuclear Manganese Metalloenzyme. It hydrolyses the semi essential amino acid L-arginine for the production of L-ornithine and urea. It plays an important role for the detoxification of nitrogenous compounds in the liver in the urea cycle. L-ornithine is utilized for the production of polyamines and L-proline (Fig.2). Polyamines including putrescine, spermidine and spermine are produced through sequential enzymatic reactions on L-ornithine (Fig.3) (Caldwell et al., 2015). Polyamines are crucial for cell proliferation, ion channel regulation and neuroprotection under physiological conditions. L-proline can also be produced from ornithine through the chemical reaction catalyzed by ornithine aminotransferase (OAT). L-proline is used for collagen formation (Caldwell et al., 2015).

Figure 2: Scheme for L-arginine catabolism (from Caldwell et al, 2015) L-arginine amino acid is hydrolyzed by arginase to L-ornithine and urea. L- ornithine is used for the production of L-proline and polyamines. L-arginine is also used for the production of nitric oxide by nitric oxide synthase

Figure 3: Scheme of polyamines synthesis (from Caldwell et al, 2015) L-ornithine is used for the production of putrescine, spermidine and spermine through different enzymatic reactions.

Arginase enzyme exists in two isoforms, arginase 1 and arginase 2 (Ash et al., 2000).

Arginase 1 (A1), the cytosolic isoform, is strongly expressed in the liver (Morris, 2002) and arginase 2 (A2), the mitochondrial isoform, is expressed primarily in extra-hepatic tissues, especially the kidney (Miyanaka et al., 1998). The physiological role of A2 is poorly understood but it has been speculated to be important for polyamines and L-proline synthesis (Wu et al., 1998). Both isoforms are also found in other tissues including brain and retina (Caldwell et al., 2015; Patel et al., 2013). A1 plays a significant role in the urea cycle for detoxification of ammonia. Mice lacking both copies of A1 suffer from severe hyperammonemia and die postnatally at 10-12 days (Iyer et al., 2002). Mice lacking both copies of A2 are healthy and do not have a phenotype except for mild hypertension beginning after 8-10 weeks of age (Huynh et al., 2009). This highlights the crucial role of arginase 1 under physiological conditions for detoxification of ammonia and production of

polyamines and L-proline. The specific role of arginase 2 in normal physiological processes are not well understood.

B.1.2. Nitric Oxide Production

Nitric oxide (NO) can be produced by the nitric oxide synthase (NOS) enzymes. There are three isoforms of NOS enzymes; neuronal (nNOS or NOS I), inducible (iNOS or NOS II) and endothelial (eNOS or NOS III) NOS. NOS catalyzes the conversion of L-arginine to

L-citrulline and NO. Tetrahydrobiopterin (BH4) is required for this enzymatic reaction.

NO is a major autocrine and paracrine signaling molecule that regulates many functions including vascular tone, neurotransmission, immune response, and oxidative stress. NO also induces vasodilation, inhibits platelet aggregation and prevents leucocyte adhesion to the vessel wall, which prevents vascular inflammation. In addition, NO controls vascular smooth muscle proliferation and can stimulate post-natal angiogenesis. eNOS derived NO can activate endothelial progenitor cells (Forstermann et al., 2012).

Arginase can regulate NO production through competing with NOS enzymes for the common substrate L-arginine. When arginase activity or expression is increased, less L- arginine will be available for NO production. This will decrease NO bioavailability and increase superoxide and peroxynitrite formation instead, a phenomenon known as NOS uncoupling. NOS uncoupling is involved in cardiovascular pathologies including dilated cardiomyopathy, ischemia reperfusion injury, endothelial dysfunction, atherosclerosis, hypertension and diabetes mellitus (Roe et al., 2012). Other factors can also induce NOS uncoupling such us depletion of BH4, increased Asymmetric dimethyl-l-arginine (ADMA) the endogenous inhibitor of eNOS and S-Glutathionylation of eNOS (Forstermann et al.,

2012).

B.1.3. L-Arginine Metabolism

L-arginine is a semi-essential amino acid. It can be obtained from the diet, protein turnover or de novo synthesis (Morris, 2016). Arginine metabolism is a complex process, yet very well studied. However, the exact physiological and pathological roles of its metabolites are still under investigation. It is important to highlight that L-arginine is catabolized by two major reactions catalyzed by arginase for production of urea and L-ornithine or catalyzed by NOS for the production of L-citrulline and NO. L-citrulline is recycled back to L- arginine by Aminosuccinate synthase (ASS) and Aminosuccinate lyase (ASL). In a few tissues, including the liver and the lining of the gastro-intestinal tract, L-ornithine can be recycled back to L-arginine through the sequential enzymatic reaction of ornithine transcarbamylase (OTC), ASS and ASL (Fig.2) (Caldwell et al., 2015).

It is documented that L-arginine concentration in the blood and within the endothelial cells is 50 µM and 800 µM, respectively (Bode-Boger et al., 1994; Gold et al., 1989). It is important to note that that the affinity of different NOS isoforms for L-arginine is much higher (Km= 2-20 µM) than the affinity of arginase for L-arginine (Km=2-20 mM) (Wu et al., 1998). However, the Vmax of arginase is 1000 times higher than Vmax of NOS, suggesting similar rates of substrate usage at low L-arginine concentrations (Wu et al.,

1998).

B.1.4. Arginine Paradox

The arginine paradox refers to the phenomenon that exogenous administration of L- arginine increases NO production despite the fact that NOS is theoretically saturated with

L-arginine and the concentration of L-arginine is highly exceeding the amount needed for maximum NO production (Elms et al., 2013a). One explanation of this phenomenon could

be that eNOS is colocalized with the L-arginine depleting enzyme, arginase, which can deplete L-arginine locally and decrease NO production (Elms et al., 2013a). Another explanation could be that there are poorly interchangeable pools of L-arginine inside the cells which need replenishment from the extracellular L-arginine environment or the L- arginine recycling pathway. This means that local depletion of the L-arginine could account for this paradox (Dioguardi, 2011). However, it is important to note that chronic administration of L-arginine did not increase NO production or improve endothelial function. Instead, it increased the byproducts of the arginase reaction: urea and L-ornithine not L-citrulline, the byproduct of NOS reaction (Wilson et al., 2007).

B.1.5. Arginase Expression Regulation

How arginase expression is regulated in vascular diseases is still not completely understood. However, arginase expression can be regulated by different factors including inflammatory mediators, oxidative stress and cell stress response. In bovine endothelial cells, inflammatory cytokines induced gene expression of A1 and A2 which was mediated through transactivation of epidermal growth factor receptors by the cytokines (Nelin et al.,

2005). It has been reported that oxidative species like peroxynitrite and hydrogen peroxide increase arginase activity/expression in bovine aortic endothelial cells through protein kinase C (PKC) mediated activation of RhoA/Rho kinase pathway (Chandra et al., 2012).

Using the same in vitro model, angiotensin II treatment has been shown to induce arginase expression through activation of RhoA/ROCK/p38 MAPK pathways leading to endothelial dysfunction (Shatanawi et al., 2011). Oxidized LDL has been reported to upregulate A2 in vascular endothelial cells by LOX-1 receptor activation and downstream Rho/ROCK signaling (Ryoo et al., 2011). At the transcription level, histone deacetylase 2 was found to

suppress A2 expression in human aortic endothelial cells (Pandey et al., 2014b). A2 mRNA was induced by NO in immortalized human retinal pigment epithelial cells (Ohnaka et al.,

2011). Hypoxia has been shown to induce A2 mRNA and protein levels through (hypoxia- inducible factors-2) HIF-2 (Krotova et al., 2010).

B.2. Arginase as a Therapeutic Target

B.2.1. Cardiovascular Diseases

Arginase has been found to be involved in many disease conditions such as; diabetes, hypertension, aging, atherosclerosis and myocardial ischemia reperfusion injury and obesity (Bagi et al., 2013; Berkowitz et al., 2003; Caldwell et al., 2015; Chung et al., 2014;

Johnson et al., 2005; Romero et al., 2012; Romero et al., 2008; Toque et al., 2013;

Tratsiakovich et al., 2013b; Yang et al., 2006, 2013). For the last two decades, the arginase enzyme has been highlighted as a key player in many disease conditions. Arginase is a critical regulator of nitric oxide synthesis and vascular function as mentioned above

(Durante et al., 2007). Excessive arginase induces NOS uncoupling through competing with NOS for the common substrate (Durante et al., 2007). In a rat model of salt induced hypertension, arginase inhibition restored endothelial dependent vasodilation (Johnson et al., 2005). It has been shown that A1 mediates increased blood pressure and contributes to vascular endothelial dysfunction in deoxycorticosterone acetate-salt hypertension (Toque et al., 2013). Arginase activity and expression are increased in diabetic aorta (Romero et al., 2008). Diabetic mice lacking one copy of A1 and both copies of A2 exhibit better EC- dependent vasodilation and less vascular stiffness and coronary fibrosis compared with diabetic wild type (WT) and AI(+/+) AII(-/-) mice (Romero et al., 2012). This suggests a major involvement of A1 in diabetes induced vascular dysfunction. In a clinical trial,

arginase inhibition protects against endothelial dysfunction caused by ischemia reperfusion in patients with coronary artery disease (Kovamees et al., 2014). Myocardial infract size and coronary microvascular dysfunction decrease with arginase inhibition following ischemia reperfusion in a rat model (Gronros et al., 2013; Jung et al., 2010). These studies clearly highlight the role of arginase as a therapeutic target for vascular dysfunction.

Arginase is not only implicated in vascular dysfunction, but it has been reported to affect the smooth muscle cells proliferation and remodeling as well. It can also induce aberrant vessel wall remodeling and neointima formation through induction of polyamines production (Durante et al., 2007). Arginase inhibition decreased polyamines production and smooth muscle cells growth (Wei et al., 2001). Enhanced arginase expression in both endothelial cells and smooth muscle cells has been shown to be associated with intimal hyperplasia in premenopausal human uterine arteries (Loyaga-Rendon et al., 2005). After arterial injury, arginase expression was increased which contributed to the remodeling response (Peyton et al., 2009).

B.2.2. Central Nervous System Diseases

Several studies have reported that arginase not only plays a role in cardiovascular health but it is also involved in central nervous system (CNS) diseases like stroke, Alzheimer’s disease (AD) and traumatic brain injury (TBI) (Ahmad et al., 2016; Caldwell et al., 2015;

Hunt et al., 2015; Petrone et al., 2016; Villalba et al., 2017). In a rat model of permanent focal ischemia, A1 expression and activity were significantly increased in the cortical lesion where the inflammatory cells are found (Quirie et al., 2013). A1 increase was correlated with an early and long-lasting A1 upregulation in macrophage and astrocytes.

A1 has a delayed downregulation in neurons near the lesion. The expression pattern of A1

was similar to that of brain-derived neurotrophic factor (BDNF) which suggests a role of

A1 in neuroplasticity. A recent study in patients with acute ischemic stroke found a significant increase in A1 in both gene and serum protein levels with increased acute ischemic stroke severity (Petrone et al., 2016). In a permanent cerebral artery occlusion mouse model, A2 deletion aggravated brain infarct size and neurological deficits (Ahmad et al., 2016). In Alzheimer’s disease models, data about A1 and A2 was contradictory.

Overexpression of A1 in the CNS reduced phosphorylation of Tau, and limited inflammation and hippocampal atrophy (Hunt et al., 2015). In another study of AD in mice,

A2 mRNA levels in brain were found to be increased with no change in A1 mRNA (Colton et al., 2006). A recent study has shown that endothelial dependent vasorelaxation was restored by L-arginine supplementation or arginase inhibition after TBI insult through a mechanism involving A1 induced eNOS uncoupling (Villalba et al., 2017). Another study has shown an improved cerebral blood flow in A2 knock out mice suffering from TBI

(Bitner et al., 2010). Further studies are needed to elucidate the distinct roles of A1 and A2 in CNS injuries.

B.2.3. Retinopathies

Our group has extensively studied the role of arginase in different models of retinopathies.

Ischemic retinopathy occurs in a variety of visual disorders like diabetic retinopathy, retinal vein occlusion, glaucoma and retinopathy of prematurity. These pathologies share common features including oxidative stress, neurodegeneration, inflammation, activation of glial cells and vascular damage (Cuenca et al., 2014; Downie et al., 2007; Hartnett, 2015;

Kowluru et al., 2015; Osborne et al., 2004; Zhang et al., 2011). Excessive arginase activity is well documented to increase oxidative stress, inflammation and neurovascular damage

as discussed above. We and others have reported that arginase is expressed in the retina in different retinal layers and cells. A1 is expressed in the ganglion cell layer, inner nuclear layer (INL) and in Müller glia cells (Patel et al., 2013; Zhang et al., 2009). A2 is expressed in the INL in horizontal cells and in the inner plexiform layer (IPL) and nerve fiber layer

(Narayanan et al., 2011; Patel et al., 2013).

We first showed in a mouse model of ocular inflammation that glial and microglial A1 expression was significantly increased with lipopolysaccharide (LPS) stimulation (Zhang et al., 2009). Inflammation, as indicated by increases in tumor necrosis factor α (TNFα) and monocyte chemoattractant protein-1 (MCP-1) expression and leukostasis, was significantly reduced with arginase deletion or inhibition (Zhang et al., 2009). Arginase is also implicated in ROP models (Narayanan et al., 2011; Stevenson et al., 2010). Oxygen induced retinopathy (OIR) is a widely used model to study ROP in rodents (Smith et al.,

1994).

Figure 4: Oxygen induced retinopathy (OIR) mouse model from (Connor et al, 2009) Neonatal mice with their mothers are kept at room air from birth till postnatal day 7 (P7), then exposed to 75% oxygen till P12. This induces loss of immature vessels and slows the development of the normal retinal vasculature leading to central vaso-obliteration. At p12, mice are returned to room air which triggers both normal and pathologic vessel regrowth and leads to neovascularization (NV) formation. NV reaches its maximum severity at P17. Afterwards, NV starts to regress with almost no vessel loss or NV visible at P25.

In this model, one week old mouse pups are maintained in 75% oxygen (hyperoxia phase) for 5 days then transferred to room air conditions (ischemia phase) (Fig.4) (Smith et al.,

1994).

In the OIR model, A2 was significantly induced at the start of the ischemic phase

(Narayanan et al., 2011). A2 deletion protected against apoptosis, glial activation and impaired retinal function (Narayanan et al., 2011). In the hyperoxia phase, A2 was also increased at post-natal day 9 (P9) (Narayanan et al., 2014). Deletion of A2 protected against retinal neurodegeneration by a mechanism involving polyamines metabolism. (Narayanan et al., 2014) Increased spermine oxidase (SMO) expression was involved in hyperoxia- mediated neurodegeneration (Narayanan et al., 2014). In the same mouse model, A2 was shown to have a crucial role in inducing vaso-obliteration as well. A2 knock out mice showed less vaso-obliteration with decreased peroxynitrite and superoxide formation as well as normalization of NOS activity (Suwanpradid et al., 2014). Polyamine oxidation has been also implicated in the hyperoxia-induced vascular injury. A recent study in our lab showed that inhibition of polyamine oxidation with MDL 72527 significantly reduced the hyperoxia-induced vascular injury, which was associated with decreased microglial activation and inflammatory cytokines and chemokines (Patel et al., 2016).

In diabetic models, our group has shown that arginase is involved in diabetes induced neurovascular injury (Elms et al., 2013b; Patel et al., 2013; Rojas et al., 2017). Ex-vivo and in vivo studies showed an endothelial dependent vasodilation impairment in diabetic mice

(Elms et al., 2013b). Homozygous deletion of A1 significantly protected against diabetes induced endothelial dysfunction (Elms et al., 2013b). Arginase expression and activity

were markedly increased in diabetic retina and endothelial cells exposed to high glucose

(Patel et al., 2013). Beneficial effects of arginase inhibition or deletion were attributed to decreased oxidative stress and leukocyte adhesion, and enhanced NO formation in diabetic models (Patel et al., 2013).

These studies emphasize the crucial role of the two arginase isoforms in different models of ischemic retinopathy. Arginase is strongly involved in neurovascular dysfunction and injury in several visual disorders, suggesting arginase as a potential therapeutic target for alleviating ischemic retinopathies.

B.3. Retinal Ischemia

B.3.1. Retina

Retina is the light sensitive tissue at the back of the eye. It is responsible for the vision. The retina is a laminar structure with both neuronal and vascular component. The different layers of the retina are as follows: nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL) and outer nuclear layer (ONL), photoreceptors layer and retinal pigment epithelial (RPE) layer

(Fig.5)

The photoreceptors are responsible for processing the light into biochemical signals which will be transmitted as electrical signals through the different retinal layers and the optic nerve and into the brain. The retina consists of different neuronal cells: ganglion cells, horizontal cells, bipolar cells and amacrine cells and the neurons are supported by glial cells, microglial cells, vascular cells and the pigment epithelium. The retina is a part of the central nervous system with the advantage of being more accessible that the brain. It is easily functionally tested and has a well-defined laminar structure (London et al., 2013).

Figure 5: Schematic diagram of retinal structure ( this figure was originally published in BERNE & LEVY PHYSIOLOGY, Chapter 8, authors: Bruce Koeppen and Bruce Stanton, Copyright Elsevier, 2010

Thus, the retina can be used as a noninvasive approach to study CNS diseases like stroke, multiple sclerosis, Parkinson disease and Alzheimer’s disease (London et al., 2013).

B.3.2. Retinal Blood Supply

In mammals, the retina is supplied by a dual circulation. The inner retina receives its blood supply from branches from the central retinal artery which arises directly from the ophthalmic artery. The artery enters with the optic nerve in close proximity with the retinal vein. Then, the retinal artery divides irregularly and dichotomously over the retina. The main retinal vessels form 2 capillary plexuses; the superficial capillary network within the nerve fiber layer, and the deep capillary network lying between the inner nuclear layer and

the outer plexiform layer (Fig.6). The outer retina gets its supply from the choroidal circulation. When the blood supply is interrupted, the ischemia outcome and the cells affected depend on which circulation is injured. Inner retina ischemia happens with occlusion of the central retinal artery, while the outer retina is affected by choriocapillaris failure (Osborne et al., 2004). Complete retinal ischemia requires the occlusion of both the ophthalmic artery and the posterior ciliary artery.

Figure 6: Retinal blood supply (This figure was originally published in Ophthalmology 4th edition, Chapter 9, Authors: Myron Yanoff and Jay S. Duker, 2014, Copyright Elsevier)

B.3.3. Pathophysiology of Retinal Ischemia

Retinal ischemia happens when the retinal blood circulation can not meet the metabolic demands of the retina. This can happen due to different causes, general and local. General circulatory failure conditions include severe left ventricular failure, and hypovolemic shock. It happens most commonly in local circulatory failure like central retinal artery occlusion (Osborne et al., 2004).

Generally, the deprivation of cellular nutrients and oxygen in the ischemic phase activates waves of depolarization and molecular events known as “ischemic cascade”. This cascade consists of vicious circle of cell depolarization, calcium influx, oxidative stress, energy depletion, and glutamatergic stimulation (D'Onofrio et al., 2013; Osborne et al., 2004). The energy depletion resulting from the loss of the normal blood flow causes membrane depolarization with the excessive release of neurotransmitters including the excitatory amino acid glutamate (Osborne et al., 2004). This leads to calcium accumulation inside the cell through the activation of specific receptors including the ionotropic glutamate receptors such as NMDA (N-methyl-D-aspartate) and AMPA (α-amino-3-hydroxyl-5- methyl-4-isoxazole-propionate) receptors. The intracellular calcium interacts with intracellular proteins leading to degradation of structural cellular elements and impairment of normal homeostasis. Anaerobic glycolysis and cytotoxic edema will lead to acidosis which further increases the edema formation and causes mitochondrial dysfunction. Free radical species formation will be increased because of degradation of membrane phospholipids and accumulation of NO. This will activate inflammatory mediators which will cause a secondary neuronal injury through the release of cytokines, chemokines and phospholipases. Eventually, cell death will be induced by these myriad events and lead to neurovascular degeneration (Osborne et al., 2004).

Ischemia reperfusion (I/R) injury is a pathological condition characterized by a period of restriction of the blood flow to certain areas followed by restoration of the blood flow and reoxygenation. The severity of cell dysfunction and I/R injury depends on the magnitude and duration of the ischemia. In general, reperfusion (restoration of blood supply, nutrients and oxygen) remains the mainstay approach for all ischemia therapies. However,

reperfusion per se can elicit a pathological response that exacerbate the injury.

Surprisingly, the restoration of the blood flow is usually associated with exacerbation of the ischemic injury and profound inflammatory response (Eltzschig et al., 2011). The mechanisms underlying the I/R injury are complex and multifactorial. They involve the increases reactive oxygen species (ROS) production, endothelial dysfunction, calcium overload, opening of the mitochondrial permeability transition pore, and a sterile inflammatory response (Eltzschig et al., 2011). The cellular responses after retinal I/R are multi-faceted and include the Toll-like receptor signaling activation which leads to inflammasome formation, activation of microglia/macrophages, blood retinal barrier breakdown which causes leukocyte infiltration, increase in regulatory T cells, release of inflammatory cytokines and activation of cell stress response (Minhas et al., 2016).

B.3.4. Retinal ischemia reperfusion (I/R) model

Several in vivo and ex vivo animal models have been described in the literature to induce retinal ischemia (Osborne et al., 2004). In the current study, we adopted a model of high intraocular pressure (IOP) induced retinal ischemia. This model was first described by

Smith and Baird in 1952 in rats and then was improved by Buchi et al in 1991 (Buchi et al., 1991; Smith et al., 1952). In this model, the I/R insult is induced by cannulation of anterior chamber with a needle connected to a sterile saline reservoir. This reservoir is elevated up to 160 cm which will increase the IOP up to 110 mmHg (Fig.7). This induces global retinal ischemia, then after certain period of ischemia, the needle will be withdrawn.

This will allow for reperfusion to happen (Osborne et al., 2004; Yokota et al., 2011).

Figure 7: Retinal ischemia reperfusion (I/R) mouse model. A needle connected to a saline reservoir is inserted in the anterior chamber of the eye and then the saline reservoir is elevated which increases the IOP. Then needle is removed to allow for reperfusion

This model produces pathological changes that can be extrapolated to central retinal artery occlusion, acute glaucoma and certain aspects of diabetic retinopathy. This model has been extensively used to study acute neurovascular injury mechanisms. Both neurodegeneration and vascular degeneration have been reported in this model (Zheng et al., 2007). In a rat model, the ganglion cell layer (GCL) cells start to die from day 2 after I/R insult (Zheng et al., 2007). TUNEL-positive cells were reported in the inner retina starting from 8 hours after I/R and this continues through 2, 3 and 7 days (Lam et al., 1999; Yokota et al., 2011;

Zheng et al., 2007). In this model, cells also die by mechanisms other than apoptosis.

Several studies have reported that cells in inner retina die by programmed cell necrosis

(necroptosis) as well (Dvoriantchikova et al., 2014; Rosenbaum et al., 2010). Retinal thickness has been reported to decrease after I/R injury (Rosenbaum et al., 1998; Yokota et al., 2011).

At the vascular level, TUNEL-positive capillary cells were reported at 2 days after the insult (Zheng et al., 2007). At 7 days, there were more TUNEL-positive capillary cells

(Zheng et al., 2007). The increased death of capillary cells leads to degeneration of some capillaries and the formation of what is known as acellular capillaries. Acellular capillaries are empty basement membrane sleeves which do not have endothelial cells. After I/R insult, the acellular capillary formation increased dramatically at 7 days with a maximum increase at 14 days (Zheng et al., 2007). Vascular permeability has been reported to increase in this model (Abcouwer et al., 2013; Abcouwer et al., 2010; Muthusamy et al.,

2014; Tong et al., 2013). In a rat model of I/R using Evans Blue leakage method, it has been reported that retinal vessels were leaky after the injury starting from 15 minutes

(Muthusamy et al., 2014). The retinal vessels remained leaky at 4 hours and up to 48 hours after the injury (Abcouwer et al., 2013; Abcouwer et al., 2010). Disruption of the blood retinal barrier proteins occludin, Zonula occludens-1 and claudin-5 was reported as well

(Muthusamy et al., 2014)

B.3.5. Arginase and I/R injury

Arginase has been identified as a key player in I/R injury in different organs including liver, heart and kidney (Erbas et al., 2004; Jung et al., 2010; Langle et al., 1995). Arginase inhibition by different agents has been shown to ameliorate the myocardial and hepatic I/R injury in several animal models (Gronros et al., 2013; Jung et al., 2010; Kovamees et al.,

2014; Reid et al., 2007; Tratsiakovich et al., 2013a) Increases in TNFα could be the reason why A1 is increased after myocardial I/R (Gao et al., 2007; Zhang et al., 2010). In diabetic rats subjected to myocardial I/R, A2 expression was significantly increased in the heart

(Tratsiakovich et al., 2017). Arginase inhibition protected against myocardial I/R in type 1

diabetic rats through NOS dependent pathways (Tratsiakovich et al., 2017). Less is known about the role of arginase in I/R injury in the brain, however studies in a model of

Alzheimer’s disease have shown that arginase inhibition protected mice from AD-like pathology (Kan et al., 2015). In the current study, we are focusing on studying the role of

A2 in CNS injury using a retina model.

Arginase 2 is the mitochondrial isoform of the arginase enzyme. Both A1 and A2 are encoded by distinct genes mapped to separate chromosomes. They share around 60% of their amino acid residues, with 100% homology in areas crucial for enzyme activity

(Vockley et al., 1996; Wu et al., 1998). A2 is the major isoform in the prostate and kidney

(Vockley et al., 1996). However, it is also expressed in the brain and retina (Patel et al.,

2013; Vockley et al., 1996; Yu et al., 2001). A2 function is thought to be primarily for the production of ornithine which is the precursor for proline, glutamate and polyamines synthesis. A2 has been reported to be implicated in cardiovascular diseases and central nervous system diseases.

B.4. Mitochondrial dysfunction

Mitochondria are highly dynamic organelles. They undergo a simultaneous fission (the division of single organelle into two or more independent organelles) and fusion (the combination of two organelles to form one organelle) and in a balanced manner. Fusion is physiologically needed during cell stress and starvation. It is important for the mitochondria at these conditions to share the matrix metabolites (Pagliuso et al., 2017). On the other hand, mitochondrial fission is important during mitosis for organelle distribution.

It is also important for normal neuronal function; it is needed for proper distribution of mitochondria to neuronal synapses and to support synaptic functions (Pagliuso et al., 2017).

Large GTPase proteins play an important role in the mitochondrial fusion and fission machinery. Mitochondrial fission depends on the dynamin related protein-1 (Drp1) and its recruitment receptors on the outer mitochondrial membrane: mitochondrial fission factor

(Mff), the mitochondrial dynamics proteins 49 and 51 (MiD49 and MiD51) and mitochondrial fission protein 1 (Fis1) (Pagliuso et al., 2017). Mitochondrial fusion depends on the mitofusion 1 (Mfn1) and mitofusion 2 (Mfn2) proteins for the mitochondrial outer membrane and optic atrophy 1 (Opa1) for the mitochondrial inner membrane.

Mitochondrial function is not only the energy production through the electron transport chain and oxidative phosphorylation; it plays a major role in generation of ROS. ROS are important participants in cell signaling mechanisms and energy homoeostasis. The mitochondrial membrane potential is important to maintain energy production.

Mitochondrial failure leads to loss of membrane potential, decreased energy production and excessive production of ROS, which activate pro-apoptotic and apoptotic signals.

Thus, healthy mitochondria are crucial to maintain normal cell function, cell survival and homeostasis. Mitochondrial dysfunction has been reported to occur in cardiovascular and neuronal diseases (Bakthavachalam et al., 2017; Chan et al., 2017).

Given that the retina is part of the CNS, retinal I/R injury highly resembles cerebral I/R injury. Mitochondrial dysfunction in cerebral ischemia happens due to the reperfusion injury and calcium overload (Bakthavachalam et al., 2017). Opening of the mitochondrial permeability transition pore (MPT) is considered to be the leading cause of mitochondrial death (Javadov et al., 2013). MPT opening causes mitochondria to become further depolarized. This causes uncoupling of the electron transport chain, disrupts oxidative phosphorylation and induces osmotic swelling and outer membrane rupture. This leads to

the release of the mitochondrial proteins, which activates the caspase cascade. This subsequently causes irreversible DNA damage and neuronal death. The uncoupled oxidative phosphorylation elicits cytochrome c release, which activates apoptotic signals.

Further, MPT opening induces mitochondrial fission/fusion and mitophagy

(Bakthavachalam et al., 2017). The severity of the mitochondrial dysfunction depends on the cell type and duration of ischemia.

C. Rationale

Literature review has extensively supported the role of arginase in the pathogenesis of several vascular and neuronal diseases like hypertension, atherosclerosis, obesity, aging, diabetes complications, retinopathies, stroke, Alzheimer’s disease, Parkinson’s disease and ischemia reperfusion injury (Ahmad et al., 2016; Bitner et al., 2010; Caldwell et al., 2015;

Hunt et al., 2015; Johnson et al., 2005; Jung et al., 2010; Langle et al., 1995; Patel et al.,

2013; Raup-Konsavage et al., 2017a; Reid et al., 2007; Romero et al., 2008; Yang et al.,

2013). The arginase effects have been attributed to several mechanisms including decreased NO bioavailability, which induces endothelial dysfunction and enhanced inflammation along with elevated oxidative stress including the production of superoxide and peroxynitrite. We have previously shown that retinal I/R injury was associated with an increase in NADPH oxidase NOX2 expression (Yokota et al., 2011). NOX2 deletion significantly protected against the neurodegeneration through a mechanism involving decreased peroxynitrite and superoxide formation, and less glial activation (Yokota et al.,

2011). It also dampened ERK phosphorylation and NF-kB activation (Yokota et al., 2011).

Our preliminary studies showed that A2 was decreased in NOX2 knockout mice

(unpublished data) which suggested that A2 could be downstream of NOX2 signaling. A2

has been shown to be upregulated and involved in the vascular dysfunction associated with aging, obesity and retinopathy (Caldwell et al., 2015; El Assar et al., 2016; Patel et al.,

2013; Shin et al., 2012; Yu et al., 2014). After traumatic brain injury, A2 expression is increased and TBI-induced impairment of cerebral blood flow is prevented by A2 deletion

(Bitner et al., 2010). We have previously shown that A2 deletion protected against neurovascular degeneration in the OIR mouse model through decreasing glial activation, and increasing bioavailable NO (Narayanan et al., 2011; Suwanpradid et al., 2014). Studies from our group and others suggest that A2 plays a crucial role in CNS injury. Being the mitochondrial isoform, A2 may be highly associated with mitochondrial function.

However, this has never been studied. Healthy mitochondria are essential to maintain the neurovascular health. Therefore, we looked at the mitochondrial fission as a potential mechanism of A2 induced neurovascular dysfunction after retinal I/R along with other mechanism including oxidative stress, cell stress response and glial activation. Based on the extensive literature review and our published and preliminary studies, we hypothesized that A2 induces neurovascular degeneration, and retinal function impairment after retinal

I/R insult through a mechanism involving glial activation, increased oxidative stress and mitochondrial fission (Fig.1). We tested this hypothesis by conducting retinal I/R insult in wild type (WT) and A2 homozygous knockout (A2-/-) mice. We evaluated the neurovascular injury and retinal function in relation to A2 expression. Our current study highlights the role of A2 in promoting CNS I/R injury which could be used as a therapeutic target for stroke and ischemic retinopathies.

Chapter II-

PUBLISHED MANUSCRIPT

Arginase 2 Promotes Neurovascular Degeneration During

Ischemia/Reperfusion Injury

Esraa Shosha*1,2,3, Zhimin Xu*1,2,3, Harumasa Yokota4, Alan Saul2,5, Modesto Rojas1,2,3, R.

William Caldwell6, Ruth B. Caldwell#1,2,3 and S.Priya Narayanan#2,7

* These authors contributed equally to this work

# Co-corresponding authors

1- Vascular Biology Center, Augusta University, Augusta, GA, USA

2- Vision Discovery Institute, Augusta University, Augusta, GA, USA

3- Charlie Norwood VA Medical Center, Augusta, GA, USA

4- Department of Ophthalmology, Asahikawa Medical University, Asahikawa, Japan

5- Department of Ophthalmology, Augusta University, Augusta, GA, USA

6- Department of Pharmacology and Toxicology, Augusta University, Augusta, GA,

USA

7- Department of Occupational Therapy, College of Allied Health Sciences, Augusta

University, Augusta, GA, USA

Corresponding Authors:

S. Priya Narayanan, PhD Ruth B. Caldwell, PhD Assistant Professor Professor Department of Occupational Therapy Vascular Biology Center College of Allied Health Sciences Medical College of Georgia Augusta University Augusta University 987 St. Sebastian Way, CL 3008 1459 Laney Walker Blvd, CB 3209A Augusta, GA 30912 Augusta, GA 30912 E-mail: [email protected] E-mail: [email protected] Office Phone Number: 706-721-6060 Office Phone Number: 706-721-6145

This study was originally published at Cell Death and Disease 2016 Nov 24;7(11):e2483. doi: 10.1038/cddis.2016.295.

Abstract:

Introduction: Retinal ischemia is a major cause of visual impairment and blindness and is involved in various disorders including diabetic retinopathy, glaucoma, optic neuropathies and retinopathy of prematurity. Neurovascular degeneration is a common feature of these pathologies. Our lab has previously reported that the ureahydrolase Arginase 2 (A2) is involved in ischemic retinopathies. Here, we are introducing A2 as a therapeutic target to prevent neurovascular injury after retinal ischemia reperfusion (I/R) insult.

Methods: Studies were performed with mice lacking both copies of A2 (A2-/-) and wild type (WT) controls (C57BL6J). I/R insult was conducted on the right eye and the left eye underwent sham surgery. Retinas were collected for analysis at different times (3h-4wk post injury). Neuronal and microvascular degeneration were evaluated using NeuN staining and vascular digests, respectively. Glial activation was evaluated by GFAP expression. Necrotic cell death was studied by propidium iodide (PI) labelling and western blot for RIP-3. Arginase expression was determined by western blot and quantitative RT-

PCR. Retinal function was determined by electroretinography (ERG).

Results: A2 mRNA and protein levels were increased in WT I/R. A2 deletion significantly reduced ganglion cell loss and microvascular degeneration and preserved retinal morphology after I/R. Glial activation, ROS formation and cell death by necroptosis were significantly reduced by A2 deletion. ERG showed improved positive scotopic threshold response (pSTR) with A2 deletion.

Conclusions: This study shows for the first time that neurovascular injury after retinal I/R is mediated through increased expression of A2. Deletion of A2 showed to be beneficial in reducing neurovascular degeneration after I/R.

Introduction:

Retinal ischemia is a major cause of visual impairment and blindness and is involved in various disorders including diabetic retinopathy, acute glaucoma and retinopathy of prematurity. These pathologies share common features including oxidative stress, neurodegeneration, inflammation, activation of glial cells and vascular damage

(Cuenca et al., 2014; Downie et al., 2007; Hartnett, 2015; Kowluru et al., 2015; Osborne et al., 2004; Zhang et al., 2011). Retinal ischemia reperfusion (I/R) injury models have been widely used to study the mechanisms of neuronal and vascular damage in ischemic retinopathy (Osborne et al., 2004; Wang et al., 2011; Zheng et al., 2007). I/R injury occurs upon restoration of tissue blood supply after a period of ischemia. Although there has been much emphasis on studying protective measures to reverse or reduce I/R insult mediated tissue damage, so far, there is no clinically effective treatment. This is mainly because the molecular mechanisms by which neurovascular dysfunction and injury happen are far from clear.

Degeneration of neurons and increased vascular permeability have been reported in a rat model of I/R injury (Abcouwer et al., 2010). Our group has shown that activity of the superoxide generating enzyme NOX2 NADPH oxidase is crucially involved in I/R mediated neuronal damage in retina (Yokota et al., 2011). Studies from our laboratory also have shown a significant role of the urea cycle enzyme, arginase, in mediating neuronal and vascular injuries in other models of retinal disease (Narayanan et al., 2011; Narayanan et al., 2014; Patel et al., 2013; Suwanpradid et al., 2014).

Arginase exists in two isoforms, arginase 1 and arginase 2 (Ash et al., 2000). Arginase 1

(A1), the cytosolic isoform, is strongly expressed in the liver (Morris, 2002) and arginase

2 (A2), the mitochondrial isoform, is expressed primarily in extra-hepatic tissues, especially the kidney (Miyanaka et al., 1998). Both isoforms are also found in other tissues including brain and retina (Caldwell et al., 2015; Patel et al., 2013) and have been linked to diseases, such as hypertension, aging, I/R injury in heart and kidney, and diabetes complications (Durante et al., 2007; Romero et al., 2012; Romero et al., 2008). Arginase has been extensively studied as a key player in I/R injury in different organs including liver, heart and kidney (Erbas et al., 2004; Jung et al., 2010; Langle et al., 1995). Studies in models of myocardial and hepatic ischemia/reperfusion injury have shown a beneficial role of arginase inhibition by different agents (Gronros et al., 2013; Jung et al., 2010;

Kovamees et al., 2014; Reid et al., 2007; Tratsiakovich et al., 2013a). While less is known about the role of arginase in I/R injury in brain, studies in a model of Alzheimer’s disease

(AD) have shown that treatment with an inhibitor of arginase protected mice from AD-like pathology (Kan et al., 2015). Upregulation of A2 has also been implicated in vascular dysfunction associated with aging, obesity and retinopathy (Caldwell et al., 2015; El Assar et al., 2016; Patel et al., 2013; Shin et al., 2012; Yu et al., 2014). Previous reports have shown that A2 expression increases after traumatic brain injury (TBI) and that TBI-induced impairment of cerebral blood flow is prevented by A2 deletion (Bitner et al., 2010). These studies along with our previous studies underscore a crucial role for A2 in CNS injury. In the current study we investigated the role of arginase 2 in retinal neurovascular injury following I/R insult.

Results:

Increased arginase 2 expression during retinal I/R injury

Our previous studies in a model of oxygen-induced retinopathy (OIR) have shown that retinal neurovascular injury is associated with increases in A2 expression and that deletion of A2 limits both neuronal and vascular injury (Narayanan et al., 2011; Narayanan et al.,

2014; Suwanpradid et al., 2014). Based on these findings, we hypothesized that A2 is involved in I/R-induced retinal injury. In order to examine this hypothesis, we used quantitative RT-PCR and western blot analyses to examine arginase expression after I/R or sham injury. These studies showed A2 mRNA and protein levels were significantly increased within 3 hours after I/R as compared to the sham control (Fig. 8A, B, C). In contrast, levels of A1 mRNA (Fig. 8A) and protein (data not shown) were slightly reduced at this time.

Figure 8: Increased expression of arginase 2 following I/R injury A) qRT-PCR showing increased A2 mRNA levels and decreased A1 mRNA levels 3 hours after I/R in WT retinas. Results are presented as a fold change of WT sham. ***p<0.001 vs. WT sham, *p<0.05 vs. WT Sham. For A2 mRNA studies, WT sham: n=8 and WT I/R: n=10. For A1 mRNA studies, WT sham: n=8 and WT I/R: n=9. Data is presented as mean± SEM B) Western blot analysis with quantification (C) showing increased A2 protein levels in WT retinas 3 hours after I/R injury. The red arrow points to the band of interest. Any other band are nonspecific. Molecular size marker is showing on the left. Results are presented as a fold change of WT sham. No A2 protein was detected in A2-/- retinas. **p<0.01 vs. WT sham. WT sham: n=6 and WT I/R: n=7. Data is presented as mean± SEM

We and others have shown that I/R injury results in a marked distortion and thinning of the neural retina (Rosenbaum et al., 1998; Schmid et al., 2014; Yokota et al., 2011). To assess the role of A2 in this I/R induced tissue damage, we determined the impact of A2 deletion on retinal morphology at 7 days after I/R injury. Quantitative analysis of H&E stained retinal sections using Image J confirmed a significant thinning of the WT I/R retinas as compared to sham controls (Fig. 9). However, A2 deletion preserved retinal morphology and prevented retinal thinning (Fig. 9A, B). Since the inner nuclear layer (INL) is highly sensitive to I/R injury, we also measured INL thickness in response I/R injury. This analysis showed a 19% decrease in INL thickness in WT I/R retinas compared to sham controls, whereas INL thickness in A2-/- I/R retinas was comparable to both sham control groups (Fig.9C). These data suggest that the I/R-induced increase in A2 expression is critically involved in mediating I/R-induced retinal injury.

Figure 9: A2 deletion prevented I/R injury-induced retinal thinning A) H&E images of frozen retinal sections shows loss of cells in the GCL and INL and 7 days after I/R injury. WT sham: n=4, n=5/ group for others. Scale bar 50µM. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer. B) Quantification of total retinal thickness. Results are presented as a percent of WT sham. *p<0.05 vs. WT sham. WT sham: n=4, n=5/ group for others. Data is presented as mean± SEM C) Quantification of INL thickness. Results are presented as a percent of WT sham. *p<0.05 vs. WT sham, #p<0.05 vs. WT I/R. WT sham and A2-/- Sham: n=5, WT I/R and A2-/- I/R: n=4. Data is presented as mean± SEM

Reduction in retinal cell death and improved neuronal survival by arginase 2 deletion

The improvement in retinal morphology in the A2-/- I/R mice suggested that A2 plays a critical role in retinal injury during I/R. Previous studies in a rat model of I/R injury have shown that retinal neurons die by necroptosis, a caspase independent mechanism of programmed cell death (Dvoriantchikova et al., 2014; Rosenbaum et al., 2010). In the current study, we examined the impact of A2 deletion on I/R-induced necroptosis by evaluating uptake of propidium iodide (PI) after I/R injury. Plasma membrane permeability is an early feature of necroptosis. PI is membrane impermeable in living cells, but when the plasma membrane is permeable it can enter the cell to bind DNA. We observed a significant increase in PI positive cells in WT retinas within 6 hours following I/R injury.

PI-labeled cells were mainly localized in GCL and INL of I/R retinas (Fig. 10A). The number of PI positive cells in the inner retinal layers was significantly reduced with A2 deletion (Fig. 10B). To further examine the role of necroptosis in I/R-induced injury we assessed the expression of receptor interacting protein-3 (RIP-3) by western blotting. RIP-

3 is a critical regulator of programmed necrosis/necroptosis (Moriwaki et al., 2013). A significant increase in RIP-3 protein levels was evident at 6 hours post-injury in WT I/R retinal lysates compared to sham controls (Fig. 10C, D). The I/R-induced increase in RIP-

3 was largely reduced in A2-/- I/R retinas. These results suggest that A2 mediates I/R induced cell death through a mechanism involving necroptosis.

Figure 10: A2 deletion reduced I/R injury-induced necroptic cell death A) PI images and quantification (B) of PI-positive cells showing increases in necroptic cells in the WT retina at 6 hours after I/R injury. Results are presented as a fold change of WT sham. ***p<0.001 vs. WT Sham, *p<0.05 vs. WT I/R. WT sham and A2-/- Sham: n=3, WT I/R: n=5, A2-/- I/R: n=4. Scale bar 50µM. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer. Data is presented as mean± SEM C) Western blot and quantitative analysis (D) showing increases in RIP-3 expression in WT retina at 6 hours after I/R injury. The red arrow points to the band of interest. Any other band are nonspecific. Molecular size marker is showing on the left. Results are presented as a fold change of WT sham. ***p<0.001 vs. WT sham, #p<0.05 vs. WT I/R. WT sham: n=9, WT I/R: n=12, A2-/- I/R: n=7, A2-/- sham: n=6. Data is presented as mean± SEM

Degeneration of GCL neurons is another major hallmark of retinal damage following I/R insult (Lam et al., 1999; Takahashi et al., 1992; Zheng et al., 2007). In order to assess the impact of A2 deletion on I/R-induced loss of GCL neurons, we prepared anti-NeuN immunolabeled retinal flat mounts from I/R treated WT and A2-/- retinas and their respective sham controls. Consistent with our previous observations, there was a marked decrease in the number of NeuN positive cells in the GCL layer of WT I/R retina compared to WT sham control (Fig. 11). Quantification using confocal imaging and Image J analysis showed a 40% decrease in GCL neurons in WT I/R retina compared to sham control.

Deletion of A2 significantly blunted this effect (Fig. 11B). The relative density of surviving

GCL neurons in the A2-/- I/R retina was not significantly different from that in the sham control retinas.

Figure 11: A2 deletion prevented I/R injury-induced loss of ganglion cell layer neurons A) NeuN labeling and quantification (B) showing decreases in ganglion cell layer (GCL) neurons in WT retina at 7 days after I/R. Results are presented as a percent of WT sham. ***p<0.001 vs. WT sham, #p<0.05 vs. WT I/R. WT sham: n=6, WT I/R: n=8, A2-/- I/R: n=3, A2-/- sham: n=4. Scale bar 100µM. Data is presented as mean± SEM

Protection of the retinal microvasculature from I/R injury by arginase 2 deletion

In addition to the neuronal injury, I/R also induces degeneration of the retinal microvasculature. It has been reported that vascular cells start to die within 2 days after retinal I/R injury with maximum losses at 7 days (Zheng et al., 2007). The vascular cell death leads to the formation of empty basement sleeves, which are termed acellular capillaries (Zheng et al., 2007). Our previous study has shown that A2 deletion significantly reduced retinal vascular injury in the OIR mouse model (Suwanpradid et al.,

2014). Based on these findings, we investigated the impact of A2 deletion on the formation of acellular capillaries after retinal I/R injury. We prepared vascular digests and quantified the number of acellular capillaries in bright field images using Image J software. This analysis showed a profound increase in the number of acellular capillaries (Fig. 12A) in

WT I/R retinas 14 days after I/R insult compared to their respective sham controls. The mean number of acellular capillaries in WT I/R vasculature is 148/mm2 compared to

13/mm2 in sham controls (Fig. 12B). Deletion of A2 significantly reduced the formation of acellular capillaries in I/R retina (35/mm2).

Figure 12: A2 deletion prevented I/R injury-induced microvascular degeneration A) Vascular digests and quantification (B) showing increases in acellular capillaries (arrows) in WT retina at 14 days after I/R injury. ***p<0.001 vs. WT sham, #p<0.001 vs. WT I/R. A2-/- sham: n=5, n=8/group for others. Scale bar 50µM. Data is presented as mean± SEM

Prevention of glial activation and oxidative stress by arginase 2 deletion

To further assess the effect of A2 deletion on I/R induced retinal injury we examined glial cell activation by analyzing expression of glial fibrillary acidic protein (GFAP). GFAP is normally expressed in astrocytes and its upregulation is linked to retinal injury (Lewis et al., 2003). Müller cells show increased GFAP expression under conditions of retinal injury and stress. We assessed GFAP expression by immunofluorescence and western blotting studies 5 days post-injury. Immunolocalization analysis showed increased GFAP expression in astrocytes and Müller cells in WT I/R retinas compared to sham controls

(Fig. 13A). GFAP was markedly increased in Müller cells indicating their activation in response to I/R injury. GFAP expression in the A2-/- retinas was largely unaffected by I/R injury. Western blot analysis further confirmed the immunolabeling results. GFAP levels were significantly increased in WT I/R retinas compared to sham controls. GFAP

expression in the A2-/- I/R retinas was markedly reduced as compared to WT I/R retinas and was similar to that in the sham controls (Fig. 13B & C). These data suggest that A2 deletion protected against glial cell activation and limited the cell stress response following retinal I/R injury.

Figure 13: A2 deletion limited I/R injury-induced glial activation A) Immunofluorescence images showing increases in GFAP immunoreactivity in WT retina at 5 days after I/R injury. WT I/R: n=4, n=3/group for others. Scale bar 50µM. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer. B) Western blot analysis and quantification (C) showing increases in GFAP expression in WT retina at 5 days after I/R injury. The red arrow points to the band of interest. Any other band are nonspecific. Molecular size marker is showing on the left. Results are presented as a fold change of WT sham. **p<0.01 vs. WT sham, #p<0.01 vs. WT I/R. WT Sham and A2-/- sham: n=3, WT I/R and A2-/- I/R: n=4. Data is presented as mean± SEM

Oxidative stress is another key mediator of I/R-induced retinal neurovascular injury. In order to assess the involvement of A2 in I/R induced oxidative stress we determined the effects of A2 deletion on the formation of superoxide and peroxynitrite. We used dihydroethidium (DHE) imaging of fresh frozen retinal sections to assess superoxide formation. When superoxide is produced it binds to DHE and oxidizes it to ethidium which binds to DNA and fluoresces red. We have previously shown that 6 hours after I/R injury there is a significant increase in superoxide formation in WT retinas (Yokota et al., 2011).

This was confirmed in our present study. Quantification of DHE assay images using fluorescence microscopy and Metamorph imaging system analysis showed that the I/R injury induced a significant increase in superoxide formation in the WT retinas and that this increase was largely prevented in the A2-/- retinas (Fig. 14A, B). Because DHE can be oxidized by other reactive oxygen species (ROS) in addition to superoxide, (Zhao et al.,

2005) we performed control studies using superoxide dismutase (SOD) to demonstrate specificity of the DHE reaction for superoxide.

Another marker of oxidative stress is peroxynitrite. Peroxynitrite (ONOO-) is a short-lived molecule at physiological pH, but it has been shown to nitrate protein tyrosine residues and thus can serve as a peroxynitrite biomarker. We have previously shown that nitrotyrosine formation is significantly increased 6 hours after I/R injury (Yokota et al., 2011). Therefore we examined ONOO- formation indirectly by western blotting analysis with an anti- nitrotyrosine antibody. A significant reduction in the level of nitrated proteins was observed in A2-/- I/R retina compared to WT I/R (Fig. 14C & D). These findings further support the involvement of A2 in I/R injury-induced oxidative stress.

The p38 mitogen-activated protein kinase is one of the mitogen-activated protein kinases that is activated in response to cell stress. Phosphorylation of p38 is an indicator of its activation. We used western blotting to assess p38 phosphorylation in samples collected 3 hours after injury and observed a significant increase in phospho-p38 in WT I/R retinal lysates compared to sham controls (Fig. 14E & F). Phosphorylation of p38 was significantly reduced in A2-/- I/R retinas compared to WT I/R retinas.

Figure 14: A2 deletion prevented I/R injury-induced oxidative stress A) DHE images of superoxide formation and quantification of the fluorescence intensity (B) showing increases in superoxide formation in WT retina at 6 hours after I/R injury. Results are presented as a fold change of WT sham. *p<0.05 vs. WT sham, #p<0.01 vs. WT I/R. n=3/ group for the 4 groups. Scale bar 50µM, RFI= relative fluorescence intensity. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer. Data is presented as mean± SEM C) Western blot analysis with quantification (E) showing decreases in nitrotyrosine formation in A2-/- retina at 6h after I/R injury. Molecular size marker is showing on the left. Results are presented as fold change of WT I/R. *P<0.05 vs. WT I/R. WT I/R: n=6, A2-/- I/R: n=5. Data is presented as mean± SEM D) Western blot analysis with quantification (F) showing decreases in p38 phosphorylation in A2-/- retina at 3 hours after I/R injury. Molecular size marker is showing on the left. *p<0.05 vs WT I/R. n=4/group for the 2 groups, OD= optical density. Data is presented as mean± SEM

Prevention from I/R-induced impairment of retinal function by arginase 2 deletion

To assess the effect of A2 deletion on function of the inner retina, we used dark-adopted

(scotopic) electroretinography (ERG) to record the positive scotopic threshold response

(pSTR). The pSTR is thought to reflect activity of the proximal retinal portion, i.e. amacrine and ganglion cells, of the sensitive rod circuit (Saszik et al., 2002). This analysis showed that the amplitude of the pSTR was significantly reduced in WT I/R mice compared to WT sham 4 weeks after the injury (Fig. 15). This I/R-induced impairment of the pSTR was significantly lessened in the A2-/- I/R retinas at higher intensity stimulation.

Figure 15: A2 deletion prevented I/R injury-induced impairment of inner retina function A) Amplitudes of the positive scotopic threshold responses (pSTRs) are plotted against stimulus intensity showing impaired function of the inner retinal neurons at 4 weeks after I/R injury. **p<0.01 vs. WT sham, $p<0.0001 vs. WT sham, #p<0.0001 vs. WT I/R. WT sham and WT I/R: n=7, A2-/- sham and A2-/- I/R: n=8. Data is presented as mean± SEM B) Averaged responses to 5 ms flashes are shown for the two highest intensities in panel A showing improved response of A2-/- I/R at 5x10-5 candela.second/meter2 (cd.s/m2) (top panel) and significant protection at 10-4 cd.s/m2 (bottom panel).The pSTR amplitude was measured at 110 ms after the flash.

Discussion:

In the present study, we demonstrate for the first time that arginase 2 enzyme is crucially involved in I/R-induced neurovascular degeneration. We show that deletion of A2 gene preserved retinal morphology, improved neuronal survival and decreased the formation of acellular capillaries while reducing oxidative stress, gliosis, phosphorylation of p38 MAPK and necroptosis-mediated cell death following I/R injury. These protective effects were associated with an improvement in function of the inner retinal neurons as shown by ERG analysis.

Our previous studies in the mouse model of I/R injury have shown that expression and activity of NOX2 NADPH oxidase are involved in the neuronal cell death. We found that

NOX2 deletion prevented I/R-induced increases in oxidative stress, gliosis and activation of cell stress pathways and protected against neuronal loss (Yokota et al., 2011).

Interestingly, western blot analyses of retinal protein samples from the same animals showed that levels of the A2 protein were increased in the WT retina after I/R injury. This increase was completely prevented by deletion of NOX2, suggesting that A2 is a downstream target of NOX2-induced increases in oxidative stress. (Unpublished data, H.

Yokota and R.B. Caldwell). The mechanism by which NOX2 activation increases expression of A2 is unclear. However, we believe that superoxide produced from activated

NOX2 along with other ROS which are markedly increased after I/R plays an important role in A2 upregulation which leads eventually to cell death and I/R induced neurovascular degeneration (Fig. 16). Relatively little is known about regulation of A2 expression.

However, previous studies have shown that oxidative species can increase A1 expression in vascular endothelial cells through protein kinase C-mediated activation of the RhoA/Rho

kinase pathway (Chandra et al., 2012). Studies using mesenteric arteries from streptozotocin-induced diabetic rats have shown a link between increased formation of phospho-p38 MAPK, increased expression of A1 and A2 and vascular dysfunction

(Pernow et al., 2015). These effects were reversed by p38 MAPK inhibition. A mechanism involving inflammation is also possible given that A2 expression in macrophages has been shown to be increased by Liver X Receptor (LXR) (Marathe et al., 2006).

Figure 16: Possible mechanisms of neurovascular degeneration after retinal I/R Cartoon showing possible sources of superoxide and other reactive oxygen species (ROS) which can increase A2 expression. A2 upregulation depletes L-arginine needed for NO production by NOS. This leads to NOS uncoupling with production of more superoxide which reacts with NO to form peroxynitrite, further increasing ROS levels. A2 upregulation can also increase SMO expression and activity, further increasing ROS production. These mechanisms can activate cell death pathways leading to neurovascular degeneration after .- retinal I/R insult. I/R: ischemia/reperfusion, O2 : superoxide, NOX2: NADPH oxidase 2, ROS: reactive oxygen species, A2: arginase 2, SMO: spermine oxidase, NOS: nitric oxide synthase, ONOO-: peroxynitrite, NO: nitric oxide. Question marks point to potential mechanisms that will be investigated in future studies.

Arginase has been identified as a potential therapeutic target for treatment of coronary artery disease (CAD) in a rat model (Jung et al., 2010). An arginase inhibitor has been shown to protect against endothelial dysfunction caused by ischemia-reperfusion in patients with CAD (Kovamees et al., 2014). Recent reports have implicated A2 in endothelial dysfunction through eNOS uncoupling in obese and aged mice, (Shin et al.,

2012; Yu et al., 2014) as well as in impairment of cerebral blood flow after traumatic brain injury (Bitner et al., 2010). Additionally, recent studies from our laboratory have demonstrated the involvement of A2 in neurovascular damage in a neonatal mouse model of ischemic retinopathy (Narayanan et al., 2011; Narayanan et al., 2014; Suwanpradid et al., 2014).

In the present study, we demonstrate a significant increase in A2 mRNA and protein levels as early as 3 hours after I/R insult. This early elevation of A2 expression implies a primary role in the pathological process. While the role of arginase in retinal I/R injury has not been studied until now, our previous studies have shown that both arginase isoforms are expressed in the retina (Patel et al., 2013). We also have found that streptozotocin-induced diabetes causes an increase in retinal arginase activity (Elms et al., 2013b; Patel et al.,

2013) and that inhibition of arginase activity or deletion of one copy of A1 and both copies of A2 reduces diabetes-induced retinal inflammation (Patel et al., 2013). Furthermore, our studies in a mouse model of oxygen-induced retinopathy (OIR) have shown that homozygous deletion of A2 prevents hyperoxia-induced neuronal-glial injury and improves retinal function (Narayanan et al., 2011). Studies in the same model showed that

A2 deletion also prevents oxidative stress and limits retinal vascular degeneration and pathological neovascularization (Suwanpradid et al., 2014). We have also found that A2 is

highly expressed in the horizontal cells and in the ONL after OIR (Narayanan et al., 2011;

Narayanan et al., 2014). A2 induced cell death in other retinal layers could be mediated through signals initiated by cell-cell interactions or release of soluble cytotoxic factors from activated cell types such as micro- or macro-glia. This will be investigated in future studies.

The role of oxidative stress and inflammation in neuronal cell loss and formation of acellular capillaries after I/R insult in retina is well established (Chen et al., 2009; Wei et al., 2011; Yokota et al., 2011). Furthermore, studies have shown that that the transcription factor NF-E2-related factor 2 (Nrf2) plays a significant role in protecting cells from I/R- induced neurodegeneration by a mechanism involving decreases in oxidative stress (Xu et al., 2015). We have previously shown that A2 deletion prevented oxidative stress and reduced hyperoxia induced retinal vascular degeneration in a model of OIR (Suwanpradid et al., 2014). Our group has also shown that retinal I/R insult significantly increases superoxide and nitrotyrosine formation by mechanisms involving activation of NOX2

(Yokota et al., 2011). In the present study, we are introducing A2 as a new player in mediating oxidative stress and neurovascular degeneration following retinal I/R. We show that deletion of A2 prevented the I/R-induced increases in superoxide and nitrotyrosine formation. Taken together with our finding that deletion of NOX2 prevents I/R injury induced increases in A2 expression, the latter result suggests that NOX2-induced increases in oxidative stress are due in part to the upregulation of A2. It should be noted that there is a difference in relative intensity of the DHE reaction in the INL and GCL which is consistent with the patterns of PI labeling in these layers. In contrast, the lack of PI labeling in the ONL in combination with the intense DHE reaction suggests that the photoreceptors

are dying at a different time and/or are less sensitive to oxidative injury than the INL and

GCL neurons.

Further studies are needed to elucidate the mechanisms by which A2 increases oxidative stress. However, our investigations in models of OIR and diabetic retinopathy suggest that increases in superoxide and peroxynitrite formation subsequent to arginase-induced depletion of the L-arginine supply to NOS may be involved (Fig.16) (Elms et al., 2013b;

Patel et al., 2013; Suwanpradid et al., 2014). Our studies in the OIR model have also identified a downstream element in the arginase pathway, polyamine metabolism by spermine oxidase (SMO), as a source of oxidative stress and damage to retinal neurons

(Narayanan et al., 2014). Oxidation of spermine by SMO has been shown to lead to production of hydrogen peroxide and the reactive aldehyde 3-amino propanal (Casero et al., 2009). Both can damage DNA, RNA, proteins and membranes. Our studies in the OIR model have shown that hyperoxia-induced neuronal cell death is associated with increases in SMO expression along with elevated levels of hydrogen peroxide formation and that treatment with a SMO inhibitor reduces oxidative stress and prevents the neuronal injury

(Narayanan et al., 2014). It is possible that A2 deletion mediated neuro vascular protection in I/R injury is mediated via polyamine signaling pathway, and will be investigated in our future studies (Fig. 16).

Activation of Müller glial cells, characterized by increased GFAP expression, is a well- established injury marker in retinal disease conditions such as I/R, diabetic retinopathy and

OIR (Cho et al., 2011; Narayanan et al., 2011; Zeng et al., 2000). This was confirmed in our study. Further, the I/R-induced activation of Muller cells was prevented by A2 deletion, demonstrating a protective effect of deletion on glia. This is consistent with our previous

observation that A2 deletion abrogates glial activation in the OIR model (Narayanan et al.,

2011).

Death of neurons in the inner retina following I/R can occur through different mechanisms including apoptosis, necrosis and programmed necrosis, i.e. necroptosis (Dvoriantchikova et al., 2014; Rosenbaum et al., 2010). In the present study, PI labelling showed a significant increase in PI positive cells in inner retina at 6 hours after I/R. The number of PI positive cells was significantly reduced in the retinas of A2 deficient mice. PI staining showed relatively reduced signal in the GCL compared to the INL in WT retinas. However, NeuN labelling showed a severe loss of GCL neurons. This could mean that GCL neurons are dying by other mechanism and/or at different time points. PI labelling was done 6 hours after I/R, but neurodegeneration was assessed 7 days after I/R. It should be noted that while

PI labels cells dying by necroptosis, it also stains late apoptotic and early necrotic cells.

Thus, we examined expression of RIP-3, a marker for necroptotic cell death. Previous studies have reported that RIP-3 expression is rapidly increased in cells of the GCL and

INL in the rat retina following acute increases in intraocular pressure (Huang et al., 2013).

It was also reported that RIP-3 immunolabeling was co-localized with the PI positive cells in the GCL that indicates the involvement of RIP-3 in necroptotic cell death. Our western blotting analysis showed a significant increase in RIP3 protein expression following I/R, which was significantly reduced in A2 deficient retinas. These results indicate that A2 is involved in the I/R-induced necroptosis.

In conclusion, our study demonstrates for the first time that A2 is crucially involved in I/R- induced retinal neurovascular injury, through a mechanism involving increased oxidative

stress, glial activation and necroptotic cell death. Our study suggests that A2 can be considered as a therapeutic target to decrease neurovascular degeneration after I/R injury.

Materials and Methods:

Animals and ischemia/reperfusion (I/R) insult

All procedures with animals were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the institutional animal care and use committee (Animal Welfare Assurance no. A3307–01).

All surgeries were performed under anesthesia, and all efforts were made to minimize suffering. We used male Wild-type C57BL6 (WT) and Arginase-2 deficient (A2-/-) mice on C57BL6 background maintained in our animal facility. These mice were subjected to ischemia reperfusion (I/R) injury in the right eye. I/R was induced as previously described

(Yokota et al., 2011). Briefly, mice (10-12 weeks old) were anesthetized with (73 mg/kg

Ketamine Hydrochloride and 7.3 mg/kg Xylaxine Hydrochloride, ip), 1% tropicamide

(Akorn, Lake Forest, IL) was used to dilate the pupil, and topical anesthesia (0.5% proparacaine hydrochloride (Akorn) was applied to cornea. A 30-gauge needle cannulated in the anterior chamber of the right eye was used to infuse sterile saline. The intraocular pressure (IOP) was raised to 110 mm Hg by elevating the saline reservoir and held for 40 minutes. To confirm ischemia we looked at the whitening of the anterior segment of the globe and blanching of episcleral veins (Da et al., 2004). The left eye was used as control.

The mice were sacrificed at various times after I/R depending on preliminary studies and existing literature as follow in (table 2):

Justification & Outcome measures Time after I/R References

qRT-PCR and western blot (Gronros et al., 2013; Hein 3 hours for A2 et al., 2003)

(Rosenbaum et al., 1998; Thickness measurements 7 days Yokota et al., 2011)

(Huang et al., 2013;

PI labeling and necroptosis 6 hours Rosenbaum et al., 2010;

Shibuki et al., 2000)

(Yokota et al., 2011; Zheng Neurodegeneration 7 days et al., 2007)

Microvascular degeneration 14 days (Zheng et al., 2007)

(Cho et al., 2011; Yokota Glial activation 5 days et al., 2011)

Oxidative stress and (Shibuki et al., 2000; 6 hours Nitrotyrosine formation Yokota et al., 2011)

ERG 4 weeks (Nashine et al., 2015)

Quantitative RT-PCR

Total RNA was extracted from frozen retinal tissue using a RNA isolation kit (Invitrogen,

Carlsbad, CA) as instructed by the manufacturer. RNA was converted to cDNA using M-

MLV reverse transcriptase (Invitrogen, Carlsbad, CA). Quantitative PCR was performed using an ABI StepOne Plus Thermocycler (Applied Biosystems) with TaqMan gene

expression assays (Invitrogen, Carlsbad, CA). The probes of TaqMan assay used to detect

A1, A2 and hypoxanthine phosphoribosyltransferase (HPRT) as internal control, were

Mm00475988_m1, Mm00477592_m1 and Mm00446968_m1. Data was normalized to

HPRT and the fold change between levels of different transcripts was calculated by the CT method.

Neurodegeneration evaluation

Neuronal degeneration was assessed as previously described with minor modification

(Yokota et al., 2011). The neuronal cell marker NeuN was used to label surviving neurons in retinal whole flat mounts. Eye balls were collected 7 days after I/R or sham surgery and fixed overnight in 4% paraformaldehyde (PFA) at 4°C. Retinas were dissected, permeabilized, blocked and then labeled with anti-NeuN (Millipore Cat. # MAB377,

Billerica, MA) in 37°C for 2 hours. Then retinas were incubated with Alexa488 anti-mouse

IgG overnight. After flatmounting, retinas were imaged using a confocal microscope (LSM

510; Carl Zeiss, Thornwood, NY). Four images were taken in the midperiphery of each retina and the NeuN positive cells were counted using ImageJ software. The result was presented as a percent of NeuN-positive cell numbers in the ganglion cell layer (GCL) of the I/R eyes compared to the sham eyes.

Retinal vasculature isolation and measurement of acellular capillaries

Trypsin digestion method was used for isolating the retinal vasculature as previously described with minor modification (Chou et al., 2013). Eyeballs were removed and fixed with 4% paraformaldehyde overnight. Retinas were carefully dissected and incubated in distilled water at room temperature with gentle shaking at least 24 hours. Then the retinas were digested with 3% trypsin (Difco Trypsin 250, Becton Dickinson and Company,

Sparks, MD) in 0.1 M Tris buffer (pH 7.8) for 1.5 hours at 37°C on an orbital shaker (50 rpm). Retinas were washed carefully in several changes of fresh distilled water until no more non-vascular tissues were observed under the microscope. The isolated retinal vasculature was air-dried on the silane coated slide and stained with periodic acid-Schiff

(PAS) and hemotoxylin. Acellular capillaries were counted in ten random areas of the mid- retina. The field area was calculated using Image J software. The number of acellular capillaries were divided by the field area to get number of acellular capillaries per 1 millimeter square (mm2) of retina.

Retinal thickness measurement

Retinal thickness was studied on retinal frozen sections from WT and A2-/- animals 7 days after the I/R injury. Cross sections with optic nerve attachment were prepared (10 μm) followed by hematoxylin and eosin (H&E) staining for morphological observation. Images were taken 162 μm away from the optic nerve head and four sections with 20 μm apart from each other were used. Whole retina or individual retinal layer (inner nuclear layer,

INL) thicknesses at 3 different distances from optic nerve head was determined using

ImageJ software and averaged. Averaged retinal thickness was presented as percentage compared to the contralateral sham eyes.

Reactive oxygen species production

Dihydroethedium (DHE) method was used to evaluate superoxide formation as described previously (Suwanpradid et al., 2014). Briefly, fresh frozen sections were preincubated with or without superoxide dismutase-polyethylene glycol (400 U/mL, Sigma-Aldrich, St.

Louis, MO) for 30 minutes, followed by reaction with DHE (10 µM) for 15 min at 37°C.

Superoxide oxidizes DHE to ethidium bromide which binds to DNA and fluoresces red

(Miller et al., 1998). The fluorescence microscope (AxioVision; Carl Zeiss, Thornwood,

NY) was used to obtain the DHE images immediately after incubation. DHE was excited at 488 nm with an emission spectrum of 610 nm. Six images/slide were taken and

Computer assisted morphometry (Metamorph Image System; Molecular Devices, PA) was used to analyze images for fluorescence intensity.

Immunofluorescence

Eyes were enucleated, fixed in 4% PFA (overnight, 4°C). The next day, eye balls were washed in PBS and then incubated with 30% sucrose overnight at 4oC. Then eye balls were snap frozen in optimal cutting temperature (OCT) solution. Cryostat sections (10 µm) were obtained, mounted on glass slides, permeabilized with 1% Triton for 10 minutes, and blocked in 10% normal goat serum for 1 hour. Sections were then incubated in different primary antibodies at 4oC overnight. The primary antibody used is GFAP (Dako Cat. #

Z0334, Carpinteria, CA, 1:200). On the next day, sections were incubated for 1 hour at room temperature in fluorescein-conjugated secondary antibody (Molecular Probes, Grand

Island, NY, 1:500), washed in PBS, and covered with mounting medium and DAPI

(Vectashield; Vector Laboratories, Burlingame, CA).

Western blot analysis

Retinal protein extracts were prepared using RIPA buffer (Millipore, Billerica, MA) containing 1x protease and phosphatase inhibitors (Complete Mini and phosSTOP, respectively; Roche Applied Science, Indianapolis, IN). We separated the proteins on SDS-

PAGE and then transferred them to nitrocellulose membrane (Millipore, Billerica, MA), blocked in 5% milk (Bio-Rad, Hercules, CA). The membranes were then incubated with different primary antibodies overnight at 4oC. The primary antibodies we used are:

Arginase-2 (Santa Cruz Biotechnology Cat. # Sc-20151, Dallas, Texas, 1:500), phospho p38 (Cell Signaling Technology Cat. # 4511, 1:500), total p38 (Cell Signaling Technology

Cat. # 9212, 1:500) GFAP (Sigma-Aldrich Cat. # G6171, St. Louis, MO, 1:500), RIP-3

(Santa Cruz Biotechnology Cat. # SC-135170, Dallas, Texas, 1:500), Tubulin (Sigma-

Aldrich Cat. # T-9026, St. Louis, MO, 1:5000) and β-actin (Sigma-Aldrich Cat. # A1978,

St. Louis, MO, 1:5000). Primary antibodies were diluted in either 5% milk or 5% bovine serum albumin (BSA). Peroxynitrite formation was determined indirectly by western blot analysis of nitrotyrosine using polyclonal anti-nitrotyrosine antibody (Millipore Cat. # 05-

233, Billerica, MA, 1:500 in 1% BSA). The next day, membranes were washed in TBST

(Tris-buffered saline with 0.5% Tween-20) and horseradish peroxidase-conjugated secondary antibodies (GE Healthcare, Piscataway, NJ) were added (1:5000 for tubulin and actin, 1:4000 for nitrotyrosine and 1:1000 for others). Enhanced chemiluminescence system (GE Healthcare Bio-Science Corp., Piscataway, NJ) was used to detect immunoreactive proteins. Data were quantified by densitometry using image J and normalized to loading control.

Propidium iodide labeling and detection of necrotic cells

Propidium iodide (PI) was used to label necrotic cells as previously described (Rosenbaum et al., 2010) Briefly, mice were subjected to I/R injury. At 5 hours after the I/R injury PI

(P-3566, Sigma, 5 mg/kg) was injected intraperitoneally and mice were sacrificed one hour later with carbon dioxide (CO2). Eyeballs were harvested and snap-frozen in OCT compound. For detecting PI-positive cells, retinal sections (10 µm) were fixed in 100% ethanol for 10 min at room temperature and quantification of necrotic cells was performed on 3 sections from each sample using a Carl Zeiss Anxioplan2 fluorescence microscope.

Electroretinogram (ERG)

Mice were dark-adapted overnight, anesthetized with ketamine/xylazine, and prepared under dim red lighting. The eyes were treated with drops of proparacaine, tropicamide, and phenylephrine. A rectal probe controlled a heating pad to maintain temperature at 37oC. A ground electrode was placed in the tail, and reference electrodes in each cheek. Silver thread electrodes were placed on each eye, and a drop of hypromellose was added to improve electrical contact and protect the cornea from drying. An optic fiber was then positioned just in front of each pupil. Visual stimuli were generated by an LED device, with the light from the LED defocused and filtered before arriving at the optic fiber launcher in order to provide extremely dim luminances, ranging from about 10-6 to 10-4 candela.second/meter2 (cd.s/m2). Testing consisted of a set of 5 milliseconds (ms) flashes over a range of intensities, randomly interleaved with a probability distribution emphasizing intensities just above threshold (which is around 4x10-6 cd.s/m2.). Responses were averaged over 10-100 trials at each intensity, and positive (pSTR) and negative

(nSTR) scotopic threshold responses were measured at 110 ms and 200 ms, respectively

(Saszik et al., 2002) after the flash that occurred 500 ms into each 2 s trial. The pSTR and nSTR amplitudes had floors at 0 µV. The results were averaged across the mice in each group (WT and A2 -/-), and the differences between the sham and I/R eyes were used to estimate the effects of the knockout on the damage caused by the I/R.

Statistical analysis

Results were presented as mean ± SEM. Statistical analysis was performed by GraphPad

Prism 7 (GraphPad Software Inc., La Jolla, CA, USA). Values were tested if they follow normal distribution by the same software. Two-way ANOVA followed by Tukey test for

multiple comparisons or, the student’s t-test (two-tailed) in case of single comparisons, were used. P≤0.05 was considered statistically significant. Outliers were checked by online outlier calculator. Experiments were repeated at least twice. For ERG studies, two-way

ANOVAs were computed to gauge the effects of the genotype across stimulus intensities, and the effects at individual intensities were assessed by t-tests after Holm-Bonferroni correction for the multiple comparisons.

Acknowledgments:

Authors thank Ms. Tahira Lemtalsi and Ms. Ji Xing for technical assistance.

This research was completed in partial fulfillment of requirements for E. Shosha’s PhD degree. The work was supported in part by grants from The National Institute of Health

(NIH grant R01-EY11766 (RBC)), the Department of Veterans Affairs, Veterans Health

Administration (RBC), Office of Research and Development, Biomedical Laboratory

Research and Development (BX001233), American Heart Association (AHA

15PRE25560007 (ES), 11SDG7440088 (SPN)) and the Culver Vision Discovery Institute at Augusta University. The contents do not represent the views of the Department of

Veterans Affairs or the United States Government. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of Interest:

The authors declare no conflict of interest.

List of abbreviation:

I/R: ischemia/reperfusion, A1: arginase 1, A2: arginase 2, AD: Alzheimer’s disease, TBI: traumatic brain injury, CNS: central nervous system, WT: wild type, A2-/-: arginase 2 deficient mice, IOP: Intraocular pressure, HPRT: hypoxanthine phosphoribosyltransferase,

PFA: paraformaldehyde, GCL: ganglion cell layer, PAS: periodic acid-Schiff, H&E: hematoxylin and eosin, INL: inner nuclear layer, DHE: dihydroethedium, OCT: optimal cutting temperature, BSA: bovine serum albumin, PI: Propidium iodide, CO2: carbon dioxide, ERG: Electroretinogram, pSTR: positive scotopic threshold response, nSTR: negative scotopic threshold response, OIR: oxygen-induced retinopathy, RIP-3: receptor interacting protein kinase-3, GFAP: glial fibrillary acidic protein, ROS: reactive oxygen species, SOD: superoxide dismutase, ONOO-: peroxynitrite, LXR: Liver X receptor, CAD: coronary artery disease, Nrf2: NF-E2-related factor 2, SMO: spermine oxidase

Chapter III

Mechanisms of Retinal Ischemia/Reperfusion Injury: Arginase 2 and

the Mitochondria (Unpublished Data)

Esraa Shosha1,2,3, Abdelrahman Fouda1,2,3, Tahira Lemtalsi1,2,3, Zhimin Xu1,2,3,

Ahmed Ibrahim2,4,7, Mohamed Al-Shabrawey2,4, R. William Caldwell5, S.Priya

Narayanan2,6, Ruth B. Caldwell1,2,3

1- Vascular Biology Center, Augusta University, Augusta, GA, USA

2- Vision Discovery Institute, Augusta University, Augusta, GA, USA

3- Charlie Norwood VA Medical Center, Augusta, GA, USA

4- Department of Oral Biology, Dental College of Georgia, Augusta, GA

5- Department of Pharmacology and Toxicology, Augusta University,

Augusta, GA, USA

6- Department of Occupational Therapy, College of Allied Health Sciences,

Augusta University, Augusta, GA, USA

7- Department of Biochemistry, Faculty of Pharmacy, Mansoura University,

Mansoura, Egypt.

Abstract:

Introduction: Retinal ischemia is implicated in several blinding diseases such as acute glaucoma, central vein occlusion and retinopathy of prematurity. Until now, there is no clinically effective treatment because the underlying mechanisms are still not fully understood. We have previously shown that the mitochondrial arginase isoform, Arginase

2 (A2) is implicated in retinal ischemia reperfusion (I/R) injury. We found that A2 promotes neurovascular degeneration and impairs retinal function after retinal I/R through increasing oxidative stress, cell death by necroptosis and glial activation. However, the exact mechanisms of A2-induced neurovascular damage are not yet known. In the current study, we hypothesized that A2 induces neurovascular damage by a mechanism involving mitochondrial dysfunction.

Methods: We used wild type (WT) and A2 knockout (A2-/-) mice subjected to I/R insult on the right eye. The left eye was used as sham control. For in vitro studies, bovine retinal endothelial cells (BRECs) were subjected to oxygen glucose deprivation/reperfusion insult

(OGD/R). Western blot analysis was used to evaluate different protein levels. Rhodamine

123 was used to study mitochondrial structure. Seahorse XFe96 analyzer was used to evaluate mitochondrial function. Optical coherence tomography was used to evaluate retinal structure in mice.

Results: We found that dynamin related protein 1 (Drp1) levels were significantly increased 3 and 6 hours after I/R insult. A2 deletion protected against the I/R induced Drp1 expression. In BRECs, A2 expression was increased after OGD/R along with increased p38 phosphorylation and PARP cleavage. Drp1 expression levels were increased after

OGD/R which was significantly reduced with Arginase inhibition. Mitochondria labeling

showed a fragmented morphology after OGD/R which was partially rescued by arginase inhibition. Seahorse analysis showed an impaired mitochondrial function after OGD/R.

OCT evaluation showed a preserved retinal structure and less retinal detachment in A2-/-

I/R retinas compared to WT.

Conclusions: This study is the first to show that A2 could affect mitochondrial dynamics through inducing mitochondrial fragmentation which could cause mitochondrial dysfunction and distorted retinal structure

Introduction:

Retinal ischemia is a prominent feature that happens in many blinding eye diseases like central vein occlusion, acute glaucoma and retinopathy of prematurity. There is a critical need for developing new therapeutic targets for retinal ischemic diseases. Until now, there is no clinically effective treatment because the molecular mechanisms are unclear. Retinal ischemia is characterized by sequential events of calcium overload, oxidative stress, inflammation and neurovascular degeneration (Minhas et al., 2016;

Osborne et al., 2004). Retinal ischemia reperfusion (I/R) models have been widely used to study the molecular mechanism of retinal ischemic disease (Shosha et al., 2016; Yokota et al., 2011; Zheng et al., 2007). We have previously shown that the reactive oxygen species

(ROS) generating enzyme NOX2 NADPH oxidase plays a role in promoting neurodegeneration after retinal I/R insult (Yokota et al., 2011). NOX 2 deletion significantly protected against I/R induced neuronal cell death through a mechanism involving decreases in glial activation, oxidative stress and activation of extracellular signal-regulated kinase (ERK), mitogen activated protein kinase (MPK) and nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB) (Yokota et al., 2011). In a recent study from our group, we found that the ureahydrolase arginase enzyme is involved in I/R induced neurovascular injury (Shosha et al., 2016).

Arginase hydrolyzes the amino acid L-arginine to form ornithine and urea. It plays an important role in ammonia detoxification through the urea cycle and in cell proliferation through formation of L-ornithine which is a precursor for polyamines and L-proline

(Caldwell et al., 2015). It exists in 2 isoforms, arginase 1 (A1) the cytosolic isoform and arginase 2 (A2) the mitochondrial isoform (Ash et al., 2000). They are expressed in

different organs including liver, kidney, brain and prostate (Morris, 2002). Both isoforms are expressed in the retina (Narayanan et al., 2011; Narayanan et al., 2014; Patel et al.,

2013). Arginase activity has been implicated in I/R insult in different organs including the heart, kidney, and liver (Erbas et al., 2004; Raup-Konsavage et al., 2017a; Tratsiakovich et al., 2017). In our recent study, we found that A2 expression is significantly increased after retinal I/R insult (Shosha et al., 2016). A2 deletion significantly protected against I/R induced neurodegeneration and preserved inner retinal function. The data indicated that A2 promotes neurovascular degeneration through a mechanism involving increased formation of superoxide and peroxynitrite, glial activation and p38 phosphorylation (Shosha et al.,

2016). We have also shown that A2 is involved in neurodegeneration and vascular injury in a mouse model of oxygen induced retinopathy (Narayanan et al., 2011; Narayanan et al.,

2014; Suwanpradid et al., 2014). The damage was attributed to increased oxidative stress,

NOS uncoupling, glial activation and activation of polyamine pathways. The A2 enzyme is implicated in other central nervous system injuries as well, like Alzheimer’s disease and traumatic brain injury (Bitner et al., 2010; Hansmannel et al., 2010).

Since A2 is the mitochondrial isoform, in the current study, we investigated its potential role in inducing neurovascular degeneration in retinal I/R model and OGD/R in vitro by altering mitochondrial structure and function.

Results:

A2 deletion reduces I/R injury-induced increases in Drp1:

Mitochondria are highly dynamic organelles. They undergo mitochondrial fission and fusion in a well-balanced manner depending on the cell type and context. Mitochondrial dynamics have been implicated in retinal I/R injury. After I/R injury, the mitochondrial morphology in ganglion cells and their axons converts to fission type accompanied by greatly increased level of the fission protein, dynamin related protein-1 (Drp1) (Park et al.,

2011). A selective inhibitor of Drp1, mdivi-1, inhibits such I/R-induced morphological changes (Park et al., 2011). Consistently, overexpression of the fusion protein optic atrophy

(Opa1) in the retinas of DAB/2J mice, a spontaneous murine glaucoma model, represses the ganglion cell loss (Ju et al., 2010). Opa1 undergoes cleavage during ischemia/reperfusion injury and introduction of a cleavage-resistant isoform of Opa1 was found to inhibit retinal injury-induced neurodegeneration (Sun et al., 2016). In the present study, we evaluated the expression levels of Drp1 at 3 and 6 hours after I/R. At 3 hours after the insult, the Drp1 expression was significantly increased in both WT and A2-/- retinas compared to both sham controls (Fig. 17A & B). At 6 hours after the insult, Drp1 expression in the WT retinas was still significantly elevated compared to WT sham and

A2-/- sham (Fig. 17C & D). Drp1 protein levels in the A2-/- I/R retinas were slightly reduced and not significantly different from levels of the both sham controls. This data suggests that A2 deletion could be protective against the neurovascular degeneration through a mechanism involving decrease in I/R-induced mitochondrial fission.

3 hours after I/R

6 hours after I/R

Figure 17: A2 Deletion Reduces I/R injury-Induced Increases in Drp1: A) Western blot analysis with quantification (B) showing increased Drp1 levels in WT I/R retinas 3 hours after the insult compared to sham controls. **p<0.01 vs. WT Sham and A2- /- Sham, #p<0.05 vs WT Sham and A2-/- Sham, n=3-4. Data is presented as mean±SEM C) Western blot analysis with quantification (D) showing increased Drp1 levels in WT I/R retinas compared to sham controls 6 hours after I/R. A2-/- I/R had less Drp1 expression which was not significantly altered from sham controls. ***p<0.001 vs. WT Sham and A2- /- Sham. n=3-4. Data is presented as mean±SEM

A2 expression and cell stress response increase in endothelial cells after OGD/R

To better asses the role of A2 in promoting vascular endothelial cell degeneration, we used bovine retinal endothelial cells (BRECs) as an in vitro model. We stressed these cells with oxygen glucose deprivation (OGD) to simulate the ischemic phase and then returned the cultures to complete media to simulate the reperfusion (R). After 6 hours of OGD followed by 6 hours of R, we found a significant increase in A2 expression (Fig. 18A & C). At the same time points, phosphorylation of the cell stress marker p38 mitogen-activated protein kinase was significantly increased (Fig. 18B & D). Arginase inhibition with 2(S)-amino-

6-boronohexanoic Acid (ABH) significantly reduced p38 activation (Fig. 18B & D). To assess the effect of A2 on cell death, we measured the cleavage of Poly (ADP-ribose)

polymerase (PARP) as a marker for cell death. The cleaved PARP was significantly increased after OGD/R and this effect was slightly reduced with arginase inhibition but this decrease did not reach statistical significance (Fig. 18B & E).

Figure 18: A2 expression and cell stress increase in endothelial cells after OGD/R A) Western blot analysis with quantification (C) showing increased A2 protein levels after 6h OGD followed by 6h R. *p<0.05 vs. Normoxia, n=3, Data is presented as mean±SEM B) Western blot analysis with quantification (D) & (E) showing increased phosphorylation levels of p38 and PARP cleavage at 6h OGD followed by 6h R. ABH treatment significantly reduced the phosphorylation of p38 and slightly decreased PARP cleavage. **p<0.01 vs. Normoxia, #p<0.01 vs. OGD/R, ****p<0.0001 vs. Normoxia, n=3, Data is presented as mean±SEM

Arginase inhibition protects against OGD/R induced mitochondrial fission

Several studies have reported that mitochondrial fission is increased after OGD in both endothelial cells and neurons (Wang et al., 2016; Yuan et al., 2017; Zhao et al., 2014). We examined the expression levels of Drp1 after OGD/R. We found an increase in Drp1 protein levels after 6h OGD/6h R (Fig. 19A & B). This increase was significantly prevented with arginase inhibition (Fig. 19A & B). The suggested increase in mitochondrial fission was confirmed by use of a mitochondrial labelling assay. We labeled the mitochondria in live endothelial cells with the cationic dye Rhodamine 123. In normoxic conditions, the mitochondrial morphology was normal with tubular highly interconnected mitochondria

(Fig. 19C). After OGD/R, the mitochondria were punctate and fewer in number with reduced fluorescence intensity (Fig. 19C). With arginase inhibition, the mitochondria were restored and resumed a more normal morphology (Fig. 19C).

Figure 19: Arginase inhibition protects against OGD/R induced mitochondrial fission A) Western blot analysis with quantification (B) showing a slight increased Drp1 expression after 6h OGD followed by 6h R. ABH treatment significantly reduced its expression. **p<0.01 vs. OGD/R, n=3. Data is presented as mean±SEM C) Mitochondria labeling using Rhodamine 123 showing elongated and interconnected (arrowheads) mitochondria under normoxic conditions versus punctate (arrows) and fragmented mitochondria after 6h OGD followed by 1h R. ABH treatment reduced mitochondrial fragmentation. Scale bar= 10 µM

OGD/R induces mitochondrial dysfunction

The effect of OGD/R on mitochondrial function of endothelial cells has not been well studied. In the current study, we used the Seahorse XFe96 instrument to study the effect of

OGD/R on endothelial cells mitochondrial function. BRECs were subjected to 5 hours

OGD followed by 1 hour “reperfusion”. At the end of the experiment time course, mitochondrial function was measured by tracking the oxygen consumption rate (OCR) following injections of selected compounds. The injection compounds were: Oligomycin which inhibits ATP synthase (complex V), FCCP which uncouples oxygen consumption from ATP production, and rotenone and antimycin A which inhibit complexes I and III, respectively. Basal respiration rate was measured before injecting any compounds. This measurement showed that basal respiration was decreased after OG/R compared to normoxic conditions (Fig. 20A). The OCR was further decreased with Oligomycin injection and was increased with FCCP injection in normoxic conditions but not in OGD/R conditions (Fig. 20A). FCCP, an uncoupling agent, simulates physiological energy demand conditions. The endothelial cells under normoxic/ normal glucose conditions responded to increased energy demand as shown by increased OCR. The endothelial cells subjected to

OGD/R failed to do this. Maximal respiration rate measured after FCCP injection was significantly reduced with OGD/R (Fig. 20A & B). Spare respiratory capacity calculated by subtracting the basal respiration from the maximal respiration was significantly reduced after OGD/R compared to normoxic conditions (Fig. 20A & B). Cells treated with the non- selective arginase inhibitor showed no difference between the OGD/R and OGD/R+ABH conditions. This could be due to the fact that ABH inhibits both A1 and A2 which have opposing effects in this model. Future studies using selective inhibition of A2 vs A1 or

genetic strategies to knockdown A2 vs A1 expression are needed to elucidate the specific role of A2 in mitochondrial function.

Figure 20: OGD/R induces mitochondrial dysfunction A) Oxygen consumption rate (OCR, pmol/min) over time in minutes was measured using Seahorse XFe96 extracellular flux analyzer. The graph shows a decreased OCR with 5h OGD followed by 1h R. ABH treatment under normoxic and OGD/R conditions did not affect OCR levels. n=9-44. Data is presented as mean±SD B) Maximal respiration calculated after adding FCCP showing a significant decrease after OGD/R compared to normoxic conditions. ABH treatment did not affect this parameter. #p<0.0001 vs. normoxia. C) Spare respiratory capacity showing a significant decrease after OGD/R compared to normoxic conditions. ABH treatment did not affect this parameter. #p<0.001 vs. normoxia

A2 deletion preserves retinal structure and prevents retinal detachment

Previous studies from our group and others have shown that retinal layers are highly disrupted after I/R insult (Schmid et al., 2014; Shosha et al., 2016; Yokota et al., 2011).

We have previously evaluated the retinal thickness using histological approaches. In the current study, we used a more relevant clinical approach; the optical coherence tomography

(OCT). OCT is used in humans and animals as a non-invasive approach to identify and assess retinal abnormalities. Using this approach in live mice, we found that WT retinal layers were disrupted with an obvious retinal detachment and edema (arrows, Fig. 21A) at

7 days after the insult. When compared to WT mice, A2-/- mice showed a more preserved retinal structure and no edema or retinal detachment. We evaluated the same mice again at

5 weeks after the injury and found that the edema in WT retinas were resolved (data not shown). Examination of the eyes showed a distorted fundus in the WT mice, which was prevented in A2-/- mice (Fig.21B). This data demonstrates the clinical relevance of A2 as a therapeutic target to alleviate retinal ischemic retinopathies.

Figure 21: A2 Deletion Preserves Retinal Structure and Prevents Retinal Detachment A) Optical coherence tomography (OCT) images showing marked distortion of retinal layers, retinal detachment (yellow arrows) and edema in WT I/R retinas 1 week after injury (n=5). A2-/- I/R retinas (n=4) showed more preserved retinal structure and less retinal detachment. B) Fundus images showing distorted fundus in WT I/R mice 1 week after injury (n=5). That was markedly prevented in A2-/- I/R (n=4).

Discussion:

In the present study, we investigated the role of A2 in promoting retinal disruption and inducing mitochondrial fission and dysfunction in I/R insult. We found that A2 deletion protected against the disruption of retinal layers and limited retinal edema after I/R insult.

Drp1 expression was significantly increased 3 and 6 hours after the injury, deletion of A2 reduced its expression. In endothelial culture studies, A2 expression was increased after

OGD/R along with cell death and cell stress markers. Drp1 expression and mitochondrial morphology studies showed that mitochondrial fission was implicated in this in-vitro model. Inhibition of arginase by ABH protected against the structural alterations but did not improve mitochondrial function.

Drp1 expression was shown to be increased significantly in the early neurodegenerative events in a mouse model of ischemic retina (Park et al., 2011). The same study also showed that Drp1 is expressed in ganglion cell layer (GCL), inner nuclear layer (INL), inner plexiform layer (IPL) and outer plexiform layer (OPL) of the retinas of normal mice. After

I/R, Drp1 immunoreactivity was significantly increased in the GCL (Park et al., 2011).

Treatment with a Drp1 inhibitor significantly protected against retinal I/R induced cell death (Park et al., 2011). The relation between A2 and mitochondrial fission has never been studied. Being the mitochondrial isoform, A2 could be playing an essential role in affecting mitochondrial dynamics under stress conditions. However, it is important to note that A2 can be translocated from mitochondria to cytosol by action of the mitochondrial processing peptidase enzyme (Pandey et al., 2014a). In the current study, we show a novel potential role of A2 in inducing mitochondrial fission in retinal I/R insult. We found that Drp1 expression was significantly increased in WT retinas early after I/R. This is was partially

blocked by A2 deletion, suggesting a potential involvement of A2 in inducing Drp1 expression in retinal I/R injury. However, more studies are needed to explore the exact mechanism of A2 induced Drp1 expression. Drp1 activation and translocation to mitochondria can be regulated by post-translational modifications including phosphorylation/dephosphorylation, nitrosylation and SUMOylation (Anzell et al., 2017).

Serine 616 and serine 637 are important sites for Drp1 phosphorylation. The outcome of phosphorylation of those sites depends on upstream regulation which makes the Drp1 induced mitochondrial fission tightly regulated in a complex fashion (Anzell et al., 2017).

Future studies looking at the different post-translational modifications of Drp1 in relation to A2 expression levels are needed to further explore the underlying mechanisms.

Oxygen glucose deprivation/reperfusion has been widely used as an in-vitro model to study the molecular mechanisms of I/R injury. Using endothelial cells subjected to OGD/R, we showed that A2 expression along with cell stress and cell death markers are increased.

Mitochondrial fission has been described using the same model as well (Peyton et al., 2009;

Yuan et al., 2017). The current study is the first to show that in bovine retinal endothelial cells, Drp1 expression and mitochondrial fission are increased after OGD/R. Furthermore, the increase in Drp1 expression was significantly reduced with arginase inhibition which further supports the involvement of A2 in promoting mitochondrial fission in retinal I/R.

This study is the first to show that OGD/R insult in retinal endothelial cells not only affects the mitochondrial dynamics, but it affects the mitochondrial function as well. We utilized

Seahorse technology to demonstrate that endothelial cell mitochondria are impaired after

OGD/R. Basal respiration was decreased with the OGD/R insult, along with reduced maximal respiration response and spare capacity. This demonstrates that mitochondrial

function is impaired after I/R insult in vitro which could be playing a role in inducing cell death. Indeed, it has been reported that mitochondria are involved in retinal ganglion apoptotic cell death after I/R through cyclophilin D mediated mitochondrial permeability transition pore (MPTP) opening (Kim et al., 2014). Further studies in this model are needed to explore the mechanisms of mitochondrial dysfunction-induced cell death.

One limitation of the current study is unavailability of a specific arginase 2 inhibitor. ABH inhibits both A1 and A2. Further studies using genetic manipulations in vitro are needed to delineate the distinct roles of A1 and A2 in inducing mitochondrial dysfunction during retinal I/R injury. Another limitation is that the studies need to be repeated with larger sample sizes to evaluate statistical significance and demonstrate reproducibility of the trends shown by these preliminary data.

Not only does mitochondrial fission play a role in retinal I/R injury, other mechanisms like mitophagy, mitochondrial fusion and mitochondrial permeability transition pore opening could be involved in the pathology as well. In brain ischemic models, mitochondria have been widely studied. Mitophagy is showed to be critical for the protection against cerebral ischemia (Shen et al., 2017; Yuan et al., 2017). Parkin, a mediator of mitophagy, has been shown to be a protective against OGD insult through promoting degradation of Drp1 (Tang et al., 2016). As the retina is part of the CNS, the findings found in cerebral ischemia could be extrapolated to retinal ischemic injury as well. Future studies exploring those potential mechanisms are needed to further understand the relation between A2 and mitochondrial injury.

Spectral domain OCT (SD-OCT) is widely used to assess retinal structural alterations in both humans and in animal models. A previous study using a mouse model of I/R utilized

the SD-OCT to show that there is a progressive retinal thinning after I/R especially in the inner plexiform layer and inner nuclear layer (Kim et al., 2016). In our current study, we used the same technique to show the disruption in retinal layers along with edema in the outer retina. When A2 was deleted, the retina showed a better preserved structure after I/R.

This finding confirms our earlier finding using histological approaches that A2 induced retinal thinning after I/R insult (Shosha et al., 2016). In a central artery occlusion mouse model of retinal ischemia, fundus examination showed edema and whitening of the retina

(Goldenberg-Cohen et al., 2008). Here, we show an abnormal fundus in WT mice compared with A2-/- mice after the insult. These findings highlight the clinical role of A2 as a potential therapeutic target.

To summarize, this study is the first to show that A2 can promote neurodegeneration and endothelial cell stress through inducing mitochondrial dysfunction and promoting mitochondrial fission. Further studies are still needed to explore the exact mechanisms.

Materials and Methods:

Animals and ischemia/reperfusion (I/R) insult

All surgeries were performed under anesthesia, and all efforts were made to minimize suffering. Wild-type C57BL6 (WT) and Arginase-2 deficient (A2-/-) mice on C57BL6 background maintained in our animal facility were used for our in vivo experiments. The ischemia reperfusion (I/R) injury in the right eye was done as previously described (Shosha et al., 2016). Mice (10-12 weeks old) were anesthetized with (73 mg/kg Ketamine

Hydrochloride and 7.3 mg/kg Xylaxine Hydrochloride, ip), 1% tropicamide (Akorn, Lake

Forest, IL) was used to dilate the pupil, and topical anesthesia (0.5% proparacaine hydrochloride (Akorn) was applied to cornea. The anterior chamber of the right eye was cannulated with a 30-gauge needle to infuse sterile saline. The saline reservoir was elevated to raise the intraocular pressure and that was held for 40 minutes. To confirm ischemia we looked at the whitening of the anterior segment of the globe and blanching of episcleral veins (Da et al., 2004). The left eye was used as control. We sacrificed the mice at different time points depending on the outcome measures and existing literature.

Cell culture and oxygen glucose deprivation/reperfusion model:

Bovine retinal endothelial cells (BRECs) from passages 4-8 were used for the cell culture studies. BRECs were isolated by our group as previously described (Behzadian et al.,

1995). Cells were grown in M199 media containing 10% Fetal bovine serum (FBS), 1%

Penicillin/Streptomycin (Gimini, West Sacramento, CA) and 10% cell systems complete media until reached 70-80% confluency. For OGD/R experiments, cells were shifted to serum starvation M199 media containing 0.2% FBS and 0.1% bovine serum albumin

(BSA) overnight. The next day, media was changed to glucose free DMEM media (Thermo

Fisher, Waltham, MA) supplemented with 1% penicillin/streptomycin and the cells were kept in <1% oxygen in hypoxia chamber for the certain periods of time as mentioned in the results part. To achieve reperfusion, after the specified OGD time, media were changed to normal complete media and cells were kept in normoxic conditions for another period of time. Cells were then processed for western blotting or mitochondria labeling as explained later. Cells were treated with or without the arginase inhibitor 2(S)-amino-6- boronohexanoic Acid (ABH) (100µM).

Mitochondrial labeling:

BRECs were cultured in glass bottom dishes and OGD/R was done as explained above. At the end of the experiment time, VectacellTM Rhodamine 123 (Vector Laboratories,

Burlingame, CA) was used to label mitochondria in live cells according to manufacturer’s instructions. Briefly, cells were washed three times with modified phosphate buffered

+ saline containing 1mM CaCl2 and 0.5 mM MgCl2 (PBS ) (Thermo Fisher, Waltham, MA).

The cells were incubated with the Rhodamine 123 staining solution in 37oC for 15 minutes.

Then, the cells were rinsed three times with PBS+. Zeiss LSM 780 Inverted Confocal microscopy (Carl Zeiss AG , Oberkochen, Germany) was used to image the live cells using

63x lens. Several Images were taken randomly to cover the whole field.

Seahorse XFe96 Mito stress test:

Mito Stress test (Agilent, Catalogue # 103015-100, Santa Clara, CA) and Seahorse XFe96

(Agilent, Santa Clara, CA) were used to evaluate the mitochondrial dysfunction. Briefly,

Seahorse cell culture plates were used for growing the cells. BRECs (p4-6) were seeded in the Seahorse cell culture plate coated with 0.2% gelatin in all the wells except A1, A12,

H1 and H12 wells which were used as background wells. Plate was left under the hood for

1 hour to ensure even distribution of cells, then cells were checked under microscope and put in the incubator. Cells were kept to grow in normal complete media for 24 hours. Next day, cells were switched to serum free media overnight and then subjected to OGD 5h and

R 1 hour as explained above. The day before the assay, the Seahorse media was prepared according to manufacturer’s instructions and supplemented with 2.5 mM glutamine

(Gimini, West Sacramento, CA) and 5.5 mM glucose (Sigma, St. Louis, MO). At the day of the assay, we measured the PH of the media and kept it at 7.4± 0.1. Then, we ran the

Mito stress test according to manufacturer’s instructions. The concentrations of the injections compounds used were as follow: Oligomycin (2µM), FCCP (2µM) and

Rotenone/antimycin A (0.5µM). The data were collected and analyzed using the Wave software (Agilent).

Spectral Domain Optical Coherence Tomography (SD-OCT):

SD-OCT was done as previously described (Ibrahim et al., 2016; Ibrahim et al., 2015).

Briefly, 2% isoflurane was used to anesthetize the mice and their pupils were dilated using

1% tropicamide eye drops. Mice were placed on the imaging platform of the Phoenix

Micron III retinal imaging microscope supplemented with OCT imaging device (Phoenix

Research Laboratories, Pleasanton, CA). To keep they eye moist, Genteal gel was applied during imaging. Several images were taken for each mouse eye.

Western blot analysis

Western blotting was done as previously described (Shosha et al., 2016). Briefly, retinal or endothelial cells lysates were extracted using RIPA buffer (Millipore, Billerica, MA) containing 1x protease and phosphatase inhibitors (Complete Mini and phosSTOP, respectively; Roche Applied Science, Indianapolis, IN). Proteins were separated on SDS-

PAGE and transferred to nitrocellulose membrane (Millipore, Billerica, MA) and then blocked in 5% milk (Bio-Rad, Hercules, CA). The primary antibodies we incubated overnight were: Arginase-2 (Santa Cruz Biotechnology Cat. # Sc-20151, Dallas, Texas,

1:500), phospho p38 (Cell Signaling Technology Cat. # 4511, Danvers, MA, 1:500), total p38 (Cell Signaling Technology Cat. # 9212, Danvers, MA, 1:500), Drp1 (Santa Cruz

Biotechnology Cat. # Sc-271583, Dallas, Texas, 1:500) , PARP (Cell Signaling

Technology Cat. # 9542, Danvers, MA, 1:500), Tubulin (Sigma-Aldrich Cat. # T-9026, St.

Louis, MO, 1:5000) and β-actin (Sigma-Aldrich Cat. # A1978, St. Louis, MO, 1:5000).

Primary antibodies were diluted in either 5% milk or 5% bovine serum albumin (BSA).

The next day, membranes were washed in TBST (Tris-buffered saline with 0.5% Tween-

20) and horseradish peroxidase-conjugated secondary antibodies (GE Healthcare,

Piscataway, NJ) were added (1:5000 for tubulin and actin and 1:1000 for others). Enhanced chemiluminescence system (GE Healthcare Bio-Science Corp., Piscataway, NJ) was used to detect immunoreactive proteins. Data were quantified by densitometry using image J and normalized to loading control.

Statistical analysis

Data is presented as mean± SEM. GraphPad Prism 7 (GraphPad Softwar Inc., La Jolla,

CA, USA) was used for statistical analysis. Two-way ANOVA or One-way ANOVA followed by Tukey test for multiple comparisons or, the student’s t-test (two-tailed) in case of single comparisons, were used. P≤0.05 was considered statistically significant. For cell culture studies, experiments were repeated at least twice in different independent batches of cells. Only one set of experiments is presented in the cell culture figures.

Acknowledgments:

This research was completed in partial fulfillment of requirements for E. Shosha’s PhD degree. The work was supported in part by grants from The National Institute of Health

(NIH grant R01-EY11766 (RBC)), the Department of Veterans Affairs, Veterans Health

Administration (RBC), Office of Research and Development, Biomedical Laboratory

Research and Development (BX001233), American Heart Association (AHA

15PRE25560007 (ES), 11SDG7440088 (SPN)) and the Culver Vision Discovery Institute at Augusta University. The contents do not represent the views of the Department of

Veterans Affairs or the United States Government. The funders had no role in study design, data collection and analysis, or preparation of the manuscript.

Chapter IV

DISCUSSION

The aim of this dissertation was to investigate the potential deleterious role of the isoenzyme arginase 2 (A2) in retinal ischemia injury. Retinal ischemia is implicated in several eye diseases such as retinopathy of prematurity, acute glaucoma, central retinal artery occlusion and diabetic retinopathy. Our previous studies have shown that

A2 is implicated in the neurovascular injury in the mouse model of oxygen induced retinopathy (Narayanan et al., 2011; Narayanan et al., 2014; Suwanpradid et al., 2014).

Deletion of A2 protected against the neurovascular degeneration through decreasing oxidative stress and preventing neuroglial activation (Narayanan et al., 2011;

Narayanan et al., 2014; Suwanpradid et al., 2014). Thus, we hypothesized that induced

A2 promotes neurovascular degeneration in I/R injury through glial activation, increased oxidative stress and induction of mitochondrial dysfunction. To investigate this, we adopted a mouse model of retinal I/R insult. We also employed in vivo and in vitro experiments with genetic as well as pharmacological blockade of arginase.

Our results have provided novel findings. Firstly, we have shown for the first time that

A2 is central in I/R-induced neurovascular degeneration. Deletion of A2 gene (using A2

KO mice) preserved retinal morphology and function, improved neuronal survival and decreased the formation of acellular capillaries after I/R. Secondly, we have investigated the possible underlying mechanisms involved in A2 mediated neurovascular degeneration. A2 KO mice showed reduction in I/R-induced oxidative stress and less glial activation and necroptosis-mediated cell death. Furthermore, we have shown that A2 can mediate its deleterious effects through a mechanism involving mitochondrial fission and fragmentation.

Arginase is the enzyme that hydrolyzes the semi-essential amino acid L-arginine for the production of L-ornithine and urea. Arginase has been extensively studied in relation to

NO signaling. It is very well established that arginase causes vascular dysfunction through competing with NOS for their common substrate L-arginine which limits its availability and decreases NO production. NO is a potent vasodilator, it also decreases platelet aggregation, leukocyte adhesion and inflammation (Caldwell et al., 2015). This study introduces a new concept and demonstrates the role of arginase in promoting neurovascular dysfunction after I/R insult. In the current study, we demonstrated that arginase not only decreases NO bioavailability, but also induces glial activation, programmed cell death and mitochondrial fragmentation.

Published literature has delineated the involvement of arginase in I/R injury in different organs. It has been shown to mediate renal, (Raup-Konsavage et al., 2017b) myocardial

(Tratsiakovich et al., 2017; Tratsiakovich et al., 2013b) and hepatic (Jeyabalan et al., 2008;

Reid et al., 2007) ischemia reperfusion injury. A recent paper has shown a neuroprotective

role of A2 in permanent ischemic stroke as well as NMDA induced excitotoxic brain injury in mice (Ahmad et al., 2016). The discrepancy between this study and our results could be due to the different injury models employed (permanent ischemia and chemical excitotoxicity) and difference in pathophysiology between brain and retinal ischemia.

Our results in the retinal I/R injury model are consistent with previous reports from the lab examining A2 in a model of retinopathy of prematurity (ROP). Our lab has previously shown that A2 deletion reduces neuro-glial injury, oxidative stress and retinal vascular degeneration, and improves retinal function in ROP (Narayanan et al., 2011; Suwanpradid et al., 2014). Furthermore, A2 mediates neurodegeneration in ROP through increased polyamine oxidation (Narayanan et al., 2014).

We have previously shown that A2 is expressed in cells of the inner nuclear layer, inner plexiform layer and nerve fiber layer of the retina (Patel et al., 2013). Staining of retinal flat mounts and sections has shown that A2 is colocalized with calbindin which is a marker for horizontal cells (Narayanan et al., 2011). Our preliminary studies have also shown that

A2 is expressed in retinal endothelial cells and mixed retinal neuronal cultures. As the retina has both a vascular and a neuronal component and A2 is expressed in both types of cells, A2 could be an important player in different retinal injuries.

The mechanisms that regulate A2 expression are still being explored. In this study, we have shown that I/R insult induced an increase in A2 expression. Oxidative stress and inflammation are important components of retinal I/R insult (Chen et al., 2009; Fang et al.,

2015; Yokota et al., 2011). It has been reported that oxidative stress induces arginase expression and activity (Chandra et al., 2012). Proinflammatory mediators including lipopolysaccharide, tumor necrosis factor and interferon gamma induce arginase

expression (Pernow et al., 2013). So, in this model, oxidative stress and inflammation are likely to be the key players in increasing A2 expression after retinal I/R.

In addition to being a target of oxidative stress, A2 can also be a source of oxidative stress.

We and others have previously shown that A2 induces oxidative stress in different injury models (Cho et al., 2015; Huang et al., 2016; Suwanpradid et al., 2014). In the current study, we further confirmed that A2 induces superoxide and peroxynitrite formation. The exact mechanisms of how A2 induces oxidative stress are still not fully understood. This could be due to downstream signaling involved in polyamine metabolism. It has been reported that A2 induces neurodegeneration in the hyperoxia phase of the OIR mouse model due to increased polyamine oxidation and induced expression of spermine oxidase

(Narayanan et al., 2014). Oxidation of polyamines by spermine oxidase has been shown to induce formation of hydrogen peroxide and the reactive aldehyde 3-amino propanal

(Casero et al., 2009). Indeed, the inhibition of polyamine oxidation protected against the vascular injury in OIR model (Patel et al., 2016). Another possible mechanism of A2- induced oxidative stress is NOS uncoupling. It is well established that increased arginase activity and expression induces NOS uncoupling (Ming et al., 2004; Shatanawi et al., 2015;

Suwanpradid et al., 2014). Uncoupled NOS produces superoxide which will react with NO to form peroxynitrite. A2-induced eNOS uncoupling has been shown to be mediated through p38 MAPK pathway which was also activated in our model (Yu et al., 2014) . This supports the concept that A2-induced oxidative stress after retinal I/R could be due to uncoupled NOS.

This dissertation proposed mitochondrial fission as a potential downstream target of A2 in

I/R induced neurovascular injury. Several studies have revealed excessive mitochondrial

fragmentation or fission during ischemia or I/R in different organs (Disatnik et al., 2013;

Kim et al., 2011; Ong et al., 2010). In studies using OGD/R model in mouse N2a neuroblastoma cells, a profile of fragmented mitochondrial was found (Tang et al., 2016).

In immortalized neuronal cell lines exposed to OGD/R, mitochondrial fission was observed which was accompanied by release of cytochrome c (Sanderson et al., 2015). These findings were further confirmed in vivo in rats exposed to global brain I/R (Kumar et al.,

2016). These reports highlight the association of excessive mitochondrial fragmentation in

I/R and further support our novel findings in the retinal I/R model.

Surprisingly, in spite of A2 being localized in the mitochondria, no studies have ever looked at the relation of A2 to mitochondrial dynamics or biogenesis. This study is the first to show a potential involvement of A2 upregulation in promoting mitochondrial fission.

Our preliminary studies using genetic deletion of A2 in vivo and pharmacological inhibition of arginase in vitro showed a trend towards decreased Drp1 expression after I/R insult. However, further study is needed to confirm this result. The exact mechanism of how A2 might reduce Drp1 expression is yet to be determined. One possible mechanism could be through the receptor interacting protein 3 (RIP3)/Drp1 axis. A previous study has shown that RIP3 can activate Drp1 and induce its translocation in mitochondria which causes mitochondrial damage (Wang et al., 2014). Here, we first showed that A2 caused neuronal cell death by necroptosis through the induction of RIP3 expression. Then we showed that A2 deletion or inhibition partially inhibited the elevation of Drp1 expression after the insult. Based on the existing literature and our preliminary data, we speculated that A2 induces mitochondrial fragmentation through the activation of RIP3 which activates Drp1 and causes its translocation to mitochondria, this will lead eventually to

neurovascular degeneration after I/R. Here, we have also shown that OGD/R injury is associated with mitochondrial dysfunction in cultured BRECs. However, further studies are needed to determine the specific role of A2 in this dysfunction.

Several studies have reported the interplay between mitochondrial dynamics and mitophagy (Anzell et al., 2017). Mitophagy is the clearance of the dysfunctional mitochondria by autophagy. Mitochondrial dynamics and mitophagy are interconnected and tightly regulated to maintain healthy mitochondria. It has been reported that mitochondrial fission is apparently essential for mitophagy (Anzell et al., 2017). The role of mitophagy in I/R injury has been controversial. In a rat model of permanent cerebral artery occlusion, cerebral ischemia preconditioning was found to be neuroprotective via increasing autophagosome formation (Sheng et al., 2010). Recently, this increased autophagy has been found to include mitophagy. (Zuo et al., 2014) On the other hand, inhibition of autophagy reduced the infarct size in the same mouse model along with increased cell viability in neuronal cultures subjected to OGD. (Zhang et al., 2013) The role of mitophagy has never been studied in retinal I/R injury. However, autophagy has been implicated in different retinal injuries. (Lin et al., 2014) It has been shown in a cobalt chloride hypoxia model that autophagic cell death is observed in retinal ganglion cell cultures along with increased levels of hypoxia-inducible factor 1-alpha (HIF1A) and Bcl-

2/E1B-19kDa interacting protein 3 (BNIP3, a mitophagy player) (Yang et al., 2012).

Hence, mitophagy could be studied as a potential downstream event consequent to the A2 induced mitochondrial fragmentation after retinal I/R insult.

In addition to the potential adverse effects of I/R-induced increases in A2 expression on mitochondrial function, the upregulation of A2 may also promote retinal injury by

enhancing the infiltration of inflammatory leukocytes into the retina. Inflammatory leukocytes express high levels of inducible NOS (iNOS) and produce large amounts of NO which is damaging to retinal cells. Indeed, ongoing studies in our lab suggest that I/R induced increases in A2 expression are associated with an increase in iNOS expression and a decrease in A1 expression. Conversely, A2 deletion increases A1 expression. This increase in A1 expression may limit I/R-induced retinal injury by a mechanism involving decreases in NO production by iNOS in myeloid-derived cells. Our preliminary studies support this concept. Our data show that A1 deletion worsens I/R-induced neurovascular injury, while administration of the recombinant A1 is protective. Moreover, mice lacking

A1 in myeloid-derived cells show worsened pathology after IR (unpublished data).

The opposite roles of the arginase enzyme isoforms may be due to their different cell specific expression and sub-cellular localization. While A2 is mitochondrial, A1 is cytosolic. A1 is expressed in retinal glia and A2 is expressed in neurons (Caldwell et al.,

2015). Both isoforms are also expressed in macrophage and endothelial cells, and their expression pattern differs in conditions of injury and disease (Yang et al., 2014). After experimental stroke, A1 has been reported to be strongest in microglia/macrophages with less expression in astrocytes. A1 upregulation in myeloid cells limits NO production by iNOS and promotes a reparative macrophage phenotype (Munder et al., 1999). While A1 is a well-accepted marker of “M2-like” anti-inflammatory myeloid cells, A2 expression is associated with pro-inflammatory macrophages. Therefore, the two isoforms appear to play opposite roles in macrophages (Yang et al., 2014). M1 macrophage polarization with LPS leads to A2 upregulation (Jin et al., 2015). On the other hand, IL-4- induced M2 polarization (alternative activation) upregulates A1 (Sheldon et al., 2013). Further study is

needed to investigate the interaction and reciprocal regulation of the two arginase isoforms.

One limitation of the in vitro studies in this dissertation is the lack of specificity of the arginase inhibitor used. To date, there is no specific inhibitor for either A1 or A2. ABH blocks the activity of both A2 and A1. In the OGD/R in vitro studies, the effect of arginase inhibition was not very dramatic. This could be due to the fact that ABH blocks both the beneficial effects of A1 and the detrimental effects of A2 which may partially mask the beneficial effects of A2 inhibition. Another limitation of the current study is the use of global knockout of A2. To address this point, genetic tools like A2 overexpression or A2 siRNA or CRISPR/Cas9 and specific cell deletion of A2 need to be used for future studies.

This dissertation provides a better understanding of the mechanisms of retinal I/R pathophysiology through delineating the deleterious role of A2. Current ongoing studies in the lab will shed further light on the role of A1. Collectively, these studies will pave the way for therapeutic targeting of the arginase enzymes to benefit patients with blinding diseases involving ischemic injury such as diabetic retinopathy, retinopathy of prematurity, glaucoma and retinal vein occlusion.

In conclusion, this study is the first to show that retinal I/R insult induces A2 expression.

We found that A2 causes neurovascular damage through increased oxidative stress, glial activation and cell death by necroptosis. Also, we have found that A2 induces mitochondrial fragmentation through increasing the expression of Drp1. This study shows that targeting A2 could be a beneficial intervention to alleviate neurovascular injury after retinal and CNS I/R insults.

Chapter V

SUMMARY

This dissertation introduces arginase 2 (A2) as a novel player in retinal ischemia reperfusion (I/R) injury. In the current study, we found that A2 expression increased after retinal I/R insult in mice. Our studies have shown that A2 is expressed in both neurons and endothelial cells of the retina. The increase in A2 expression is associated with increases in oxidative stress as measured by increased superoxide and peroxynitrite formation which were significantly reduced by A2 deletion. The oxidative stress further increases A2 expression. We have also shown that A2-induced neurovascular degeneration involves glial activation and cell death by necroptosis (Fig. 22). The involvement of A2 in this process was confirmed by A2 deletion, which significantly reduced both glial activation and cell death. Given the mitochondrial localization of A2, we examined its potential role in mitochondrial dysfunction during I/R injury. This analysis showed that A2 could be playing a detrimental role in retinal I/R insult through promoting mitochondrial fragmentation and altering mitochondrial oxidative metabolism.

Taken together, our results showed that A2 induced neurovascular degeneration, impairment of retinal function and disruption of retinal structure in retinal I/R injury are mediated by a mechanism involving increases in oxidative stress, glial cell activation and mitochondrial dysfunction. Future studies using A2 specific inhibitors and genetic manipulations are needed to explore the exact mechanisms.

Figure 22: Summary of the role of A2 in inducing neurovascular degeneration following retinal I/R injury Retinal I/R induces A2 expression in neurons and endothelial cells. The elevated A2 levels promote glial activation, increase oxidative stress, and mitochondrial fragmentation by increasing Drp1 expression. Collectively, these events will lead to cell death (manifested by increases in RIP3 expression) and neurovascular injury.

REFERENCES

Abcouwer, S. F., Lin, C. M., Shanmugam, S., Muthusamy, A., Barber, A. J., & Antonetti, D. A. (2013). Minocycline prevents retinal inflammation and vascular permeability following ischemia-reperfusion injury. Journal of Neuroinflammation, 10, 149. doi:10.1186/1742-2094-10-149 Abcouwer, S. F., Lin, C. M., Wolpert, E. B., Shanmugam, S., Schaefer, E. W., Freeman, W. M., . . . Antonetti, D. A. (2010). Effects of ischemic preconditioning and bevacizumab on apoptosis and vascular permeability following retinal ischemia- reperfusion injury. Investigative Ophthalmology and Visual Science, 51(11), 5920- 5933. doi:10.1167/iovs.10-5264 Ahmad, A. S., Shah, Z. A., & Dore, S. (2016). Protective Role of Arginase II in Cerebral Ischemia and Excitotoxicity. J Neurol Neurosci, 7(2). Anzell, A. R., Maizy, R., Przyklenk, K., & Sanderson, T. H. (2017). Mitochondrial Quality Control and Disease: Insights into Ischemia-Reperfusion Injury. Molecular Neurobiology. doi:10.1007/s12035-017-0503-9 Ash, D. E., Cox, J. D., & Christianson, D. W. (2000). Arginase: a binuclear manganese metalloenzyme. Metal Ions in Biological Systems, 37, 407-428. Bagi, Z., Feher, A., Dou, H., & Broskova, Z. (2013). Selective up-regulation of arginase-1 in coronary arteries of diabetic patients. Frontiers in Immunology, 4, 293. doi:10.3389/fimmu.2013.00293 Bakthavachalam, P., & Shanmugam, P. S. (2017). Mitochondrial dysfunction - Silent killer in cerebral ischemia. Journal of the Neurological Sciences, 375, 417-423. doi:10.1016/j.jns.2017.02.043 Behzadian, M. A., Wang, X. L., Jiang, B., & Caldwell, R. B. (1995). Angiostatic role of astrocytes: suppression of vascular endothelial cell growth by TGF-beta and other inhibitory factor(s). Glia, 15(4), 480-490. doi:10.1002/glia.440150411 Berkowitz, D. E., White, R., Li, D., Minhas, K. M., Cernetich, A., Kim, S., . . . Hare, J. M. (2003). Arginase reciprocally regulates nitric oxide synthase activity and contributes to endothelial dysfunction in aging blood vessels. Circulation, 108(16), 2000-2006. doi:10.1161/01.cir.0000092948.04444.c7 Bitner, B. R., Brink, D. C., Mathew, L. C., Pautler, R. G., & Robertson, C. S. (2010). Impact of arginase II on CBF in experimental cortical impact injury in mice using MRI. Journal of Cerebral Blood Flow and Metabolism, 30(6), 1105-1109. doi:10.1038/jcbfm.2010.47 Bode-Boger, S. M., Boger, R. H., Creutzig, A., Tsikas, D., Gutzki, F. M., Alexander, K., & Frolich, J. C. (1994). L-arginine infusion decreases peripheral arterial resistance and inhibits platelet aggregation in healthy subjects. Clinical Science (London, England: 1979), 87(3), 303-310. Buchi, E. R., Suivaizdis, I., & Fu, J. (1991). Pressure-induced retinal ischemia in rats: an experimental model for quantitative study. Ophthalmologica, 203(3), 138-147. Caldwell, R. B., Toque, H. A., Narayanan, S. P., & Caldwell, R. W. (2015). Arginase: an old enzyme with new tricks. Trends in Pharmacological Sciences, 36(6), 395-405. doi:10.1016/j.tips.2015.03.006

Casero, R. A., & Pegg, A. E. (2009). Polyamine catabolism and disease. Biochemical Journal, 421(3), 323-338. doi:10.1042/bj20090598 Chan, S. H. H., & Chan, J. Y. H. (2017). Mitochondria and Reactive Oxygen Species Contribute to Neurogenic Hypertension. Physiology (Bethesda, Md.), 32(4), 308- 321. doi:10.1152/physiol.00006.2017 Chandra, S., Romero, M. J., Shatanawi, A., Alkilany, A. M., Caldwell, R. B., & Caldwell, R. W. (2012). Oxidative species increase arginase activity in endothelial cells through the RhoA/Rho kinase pathway. British Journal of Pharmacology, 165(2), 506-519. doi:10.1111/j.1476-5381.2011.01584.x Chen, B., Caballero, S., Seo, S., Grant, M. B., & Lewin, A. S. (2009). Delivery of antioxidant enzyme genes to protect against ischemia/reperfusion-induced injury to retinal microvasculature. Investigative Ophthalmology and Visual Science, 50(12), 5587-5595. doi:10.1167/iovs.09-3633 Cho, J. S., Oh, Y. J., Kim, O. S., & Na, S. (2015). The effects of arginase inhibitor on lung oxidative stress and inflammation caused by pneumoperitoneum in rats. BMC Anesthesiology, 15, 129. doi:10.1186/s12871-015-0112-y Cho, K. J., Kim, J. H., Park, H. Y., & Park, C. K. (2011). Glial cell response and iNOS expression in the optic nerve head and retina of the rat following acute high IOP ischemia-reperfusion. Brain Research, 1403, 67-77. doi:10.1016/j.brainres.2011.06.005 Chou, J. C., Rollins, S. D., & Fawzi, A. A. (2013). Trypsin digest protocol to analyze the retinal vasculature of a mouse model. J Vis Exp(76), e50489. doi:10.3791/50489 Chung, J. H., Moon, J., Lee, Y. S., Chung, H. K., Lee, S. M., & Shin, M. J. (2014). Arginase inhibition restores endothelial function in diet-induced obesity. Biochemical and Biophysical Research Communications, 451(2), 179-183. doi:10.1016/j.bbrc.2014.07.083 Colton, C. A., Mott, R. T., Sharpe, H., Xu, Q., Van Nostrand, W. E., & Vitek, M. P. (2006). Expression profiles for macrophage alternative activation genes in AD and in mouse models of AD. Journal of Neuroinflammation, 3, 27. doi:10.1186/1742- 2094-3-27 Cuenca, N., Fernandez-Sanchez, L., Campello, L., Maneu, V., De la Villa, P., Lax, P., & Pinilla, I. (2014). Cellular responses following retinal injuries and therapeutic approaches for neurodegenerative diseases. Progress in Retinal and Eye Research, 43, 17-75. doi:10.1016/j.preteyeres.2014.07.001 D'Onofrio, P. M., & Koeberle, P. D. (2013). What can we learn about stroke from retinal ischemia models? Acta Pharmacologica Sinica, 34(1), 91-103. doi:10.1038/aps.2012.165 Da, T., & Verkman, A. S. (2004). Aquaporin-4 gene disruption in mice protects against impaired retinal function and cell death after ischemia. Investigative Ophthalmology and Visual Science, 45(12), 4477-4483. doi:10.1167/iovs.04-0940 Dioguardi, F. S. (2011). To give or not to give? Lessons from the arginine paradox. J Nutrigenet Nutrigenomics, 4(2), 90-98. doi:10.1159/000327777 Disatnik, M. H., Ferreira, J. C., Campos, J. C., Gomes, K. S., Dourado, P. M., Qi, X., & Mochly-Rosen, D. (2013). Acute inhibition of excessive mitochondrial fission after myocardial infarction prevents long-term cardiac dysfunction. J Am Heart Assoc, 2(5), e000461. doi:10.1161/jaha.113.000461

Downie, L. E., Pianta, M. J., Vingrys, A. J., Wilkinson-Berka, J. L., & Fletcher, E. L. (2007). Neuronal and glial cell changes are determined by retinal vascularization in retinopathy of prematurity. Journal of Comparative Neurology, 504(4), 404-417. doi:10.1002/cne.21449 Durante, W., Johnson, F. K., & Johnson, R. A. (2007). Arginase: a critical regulator of nitric oxide synthesis and vascular function. Clinical and Experimental Pharmacology and Physiology, 34(9), 906-911. doi:10.1111/j.1440- 1681.2007.04638.x Dvoriantchikova, G., Degterev, A., & Ivanov, D. (2014). Retinal ganglion cell (RGC) programmed necrosis contributes to ischemia-reperfusion-induced retinal damage. Experimental Eye Research, 123, 1-7. doi:10.1016/j.exer.2014.04.009 El Assar, M., Angulo, J., Santos-Ruiz, M., Ruiz de Adana, J. C., Pindado, M. L., Sanchez- Ferrer, A., . . . Rodriguez-Manas, L. (2016). Asymmetric dimethylarginine (ADMA) elevation and arginase up-regulation contribute to endothelial dysfunction related to insulin resistance in rats and morbidly obese humans. Journal of Physiology, 594(11), 3045-3060. doi:10.1113/jp271836 Elms, S., Chen, F., Wang, Y., Qian, J., Askari, B., Yu, Y., . . . Fulton, D. J. (2013a). Insights into the arginine paradox: evidence against the importance of subcellular location of arginase and eNOS. American Journal of Physiology: Heart and Circulatory Physiology, 305(5), H651-666. doi:10.1152/ajpheart.00755.2012 Elms, S. C., Toque, H. A., Rojas, M., Xu, Z., Caldwell, R. W., & Caldwell, R. B. (2013b). The role of arginase I in diabetes-induced retinal vascular dysfunction in mouse and rat models of diabetes. Diabetologia, 56(3), 654-662. doi:10.1007/s00125-012- 2789-5 Eltzschig, H. K., & Eckle, T. (2011). Ischemia and reperfusion--from mechanism to translation. Nature Medicine, 17(11), 1391-1401. doi:10.1038/nm.2507 Erbas, H., Aydogdu, N., & Kaymak, K. (2004). Effects of N-acetylcysteine on arginase, ornithine and nitric oxide in renal ischemia-reperfusion injury. Pharmacological Research, 50(5), 523-527. doi:10.1016/j.phrs.2004.04.005 Fang, I. M., Yang, C. M., & Yang, C. H. (2015). Chitosan oligosaccharides prevented retinal ischemia and reperfusion injury via reduced oxidative stress and inflammation in rats. Experimental Eye Research, 130, 38-50. doi:10.1016/j.exer.2014.12.001 Forstermann, U., & Sessa, W. C. (2012). Nitric oxide synthases: regulation and function. European Heart Journal, 33(7), 829-837, 837a-837d. doi:10.1093/eurheartj/ehr304 Gao, X., Xu, X., Belmadani, S., Park, Y., Tang, Z., Feldman, A. M., . . . Zhang, C. (2007). TNF-alpha contributes to endothelial dysfunction by upregulating arginase in ischemia/reperfusion injury. Arteriosclerosis, Thrombosis, and Vascular Biology, 27(6), 1269-1275. doi:10.1161/atvbaha.107.142521 Gold, M. E., Bush, P. A., & Ignarro, L. J. (1989). Depletion of arterial L-arginine causes reversible tolerance to endothelium-dependent relaxation. Biochemical and Biophysical Research Communications, 164(2), 714-721. Goldenberg-Cohen, N., Dadon, S., Avraham, B. C., Kramer, M., Hasanreisoglu, M., Eldar, I., . . . Bahar, I. (2008). Molecular and histological changes following central retinal artery occlusion in a mouse model. Experimental Eye Research, 87(4), 327-333. doi:10.1016/j.exer.2008.06.014

Gronros, J., Kiss, A., Palmer, M., Jung, C., Berkowitz, D., & Pernow, J. (2013). Arginase inhibition improves coronary microvascular function and reduces infarct size following ischaemia-reperfusion in a rat model. Acta Physiologica (Oxford, England), 208(2), 172-179. doi:10.1111/apha.12097 Hansmannel, F., Sillaire, A., Kamboh, M. I., Lendon, C., Pasquier, F., Hannequin, D., . . . Lambert, J. C. (2010). Is the urea cycle involved in Alzheimer's disease? Journal of Alzheimer's Disease, 21(3), 1013-1021. doi:10.3233/jad-2010-100630 Hartnett, M. E. (2015). Pathophysiology and mechanisms of severe retinopathy of prematurity. Ophthalmology, 122(1), 200-210. doi:10.1016/j.ophtha.2014.07.050 Hein, T. W., Zhang, C., Wang, W., Chang, C. I., Thengchaisri, N., & Kuo, L. (2003). Ischemia-reperfusion selectively impairs nitric oxide-mediated dilation in coronary arterioles: counteracting role of arginase. FASEB Journal, 17(15), 2328-2330. doi:10.1096/fj.03-0115fje Huang, J., Rajapakse, A., Xiong, Y., Montani, J. P., Verrey, F., Ming, X. F., & Yang, Z. (2016). Genetic Targeting of Arginase-II in Mouse Prevents Renal Oxidative Stress and Inflammation in Diet-Induced Obesity. Frontiers in Physiology, 7, 560. doi:10.3389/fphys.2016.00560 Huang, J. F., Shang, L., Zhang, M. Q., Wang, H., Chen, D., Tong, J. B., . . . Xiong, K. (2013). Differential neuronal expression of receptor interacting protein 3 in rat retina: involvement in ischemic stress response. BMC Neuroscience, 14, 16. doi:10.1186/1471-2202-14-16 Hunt, J. B., Jr., Nash, K. R., Placides, D., Moran, P., Selenica, M. L., Abuqalbeen, F., . . . Lee, D. C. (2015). Sustained Arginase 1 Expression Modulates Pathological Tau Deposits in a Mouse Model of Tauopathy. Journal of Neuroscience, 35(44), 14842- 14860. doi:10.1523/jneurosci.3959-14.2015 Huynh, N. N., Andrews, K. L., Head, G. A., Khong, S. M., Mayorov, D. N., Murphy, A. J., . . . Chin-Dusting, J. P. (2009). Arginase II knockout mouse displays a hypertensive phenotype despite a decreased vasoconstrictory profile. Hypertension, 54(2), 294-301. doi:10.1161/hypertensionaha.108.121731 Ibrahim, A. S., Mander, S., Hussein, K. A., Elsherbiny, N. M., Smith, S. B., Al-Shabrawey, M., & Tawfik, A. (2016). Hyperhomocysteinemia disrupts retinal pigment epithelial structure and function with features of age-related macular degeneration. Oncotarget, 7(8), 8532-8545. doi:10.18632/oncotarget.7384 Ibrahim, A. S., Tawfik, A. M., Hussein, K. A., Elshafey, S., Markand, S., Rizk, N., . . . Al- Shabrawey, M. (2015). Pigment epithelium-derived factor inhibits retinal microvascular dysfunction induced by 12/15-lipoxygenase-derived eicosanoids. Biochimica et Biophysica Acta, 1851(3), 290-298. doi:10.1016/j.bbalip.2014.12.017 Iyer, R. K., Yoo, P. K., Kern, R. M., Rozengurt, N., Tsoa, R., O'Brien, W. E., . . . Cederbaum, S. D. (2002). Mouse model for human arginase deficiency. Molecular and Cellular Biology, 22(13), 4491-4498. Javadov, S., & Kuznetsov, A. (2013). Mitochondrial permeability transition and cell death: the role of cyclophilin d. Frontiers in Physiology, 4, 76. doi:10.3389/fphys.2013.00076

Jeyabalan, G., Klune, J. R., Nakao, A., Martik, N., Wu, G., Tsung, A., & Geller, D. A. (2008). Arginase blockade protects against hepatic damage in warm ischemia- reperfusion. Nitric Oxide, 19(1), 29-35. doi:10.1016/j.niox.2008.04.002 Jin, Y., Liu, Y., & Nelin, L. D. (2015). Extracellular signal-regulated kinase mediates expression of arginase II but not inducible nitric-oxide synthase in lipopolysaccharide-stimulated macrophages. Journal of Biological Chemistry, 290(4), 2099-2111. doi:10.1074/jbc.M114.599985 Johnson, F. K., Johnson, R. A., Peyton, K. J., & Durante, W. (2005). Arginase inhibition restores arteriolar endothelial function in Dahl rats with salt-induced hypertension. American Journal of Physiology: Regulatory, Integrative and Comparative Physiology, 288(4), R1057-1062. doi:10.1152/ajpregu.00758.2004 Ju, W. K., Kim, K. Y., Duong-Polk, K. X., Lindsey, J. D., Ellisman, M. H., & Weinreb, R. N. (2010). Increased optic atrophy type 1 expression protects retinal ganglion cells in a mouse model of glaucoma. Molecular Vision, 16, 1331-1342. Jung, C., Gonon, A. T., Sjoquist, P. O., Lundberg, J. O., & Pernow, J. (2010). Arginase inhibition mediates cardioprotection during ischaemia-reperfusion. Cardiovascular Research, 85(1), 147-154. doi:10.1093/cvr/cvp303 Kan, M. J., Lee, J. E., Wilson, J. G., Everhart, A. L., Brown, C. M., Hoofnagle, A. N., . . . Colton, C. A. (2015). Arginine deprivation and immune suppression in a mouse model of Alzheimer's disease. Journal of Neuroscience, 35(15), 5969-5982. doi:10.1523/JNEUROSCI.4668-14.2015 Kim, B. J., Silverman, S. M., Liu, Y., Wordinger, R. J., Pang, I. H., & Clark, A. F. (2016). In vitro and in vivo neuroprotective effects of cJun N-terminal kinase inhibitors on retinal ganglion cells. Molecular Neurodegeneration, 11, 30. doi:10.1186/s13024- 016-0093-4 Kim, H., Scimia, M. C., Wilkinson, D., Trelles, R. D., Wood, M. R., Bowtell, D., . . . Ronai, Z. A. (2011). Fine-tuning of Drp1/Fis1 availability by AKAP121/Siah2 regulates mitochondrial adaptation to hypoxia. Molecular Cell, 44(4), 532-544. doi:10.1016/j.molcel.2011.08.045 Kim, S. Y., Shim, M. S., Kim, K. Y., Weinreb, R. N., Wheeler, L. A., & Ju, W. K. (2014). Inhibition of cyclophilin D by cyclosporin A promotes retinal ganglion cell survival by preventing mitochondrial alteration in ischemic injury. Cell Death & Disease, 5, e1105. doi:10.1038/cddis.2014.80 Koberlein, J., Beifus, K., Schaffert, C., & Finger, R. P. (2013). The economic burden of visual impairment and blindness: a systematic review. BMJ Open, 3(11), e003471. doi:10.1136/bmjopen-2013-003471 Kovamees, O., Shemyakin, A., & Pernow, J. (2014). Effect of arginase inhibition on ischemia-reperfusion injury in patients with coronary artery disease with and without diabetes mellitus. PloS One, 9(7), e103260. doi:10.1371/journal.pone.0103260 Kowluru, R. A., Kowluru, A., Mishra, M., & Kumar, B. (2015). Oxidative stress and epigenetic modifications in the pathogenesis of diabetic retinopathy. Progress in Retinal and Eye Research, 48, 40-61. doi:10.1016/j.preteyeres.2015.05.001 Krotova, K., Patel, J. M., Block, E. R., & Zharikov, S. (2010). Hypoxic upregulation of arginase II in human lung endothelial cells. American Journal of Physiology: Cell Physiology, 299(6), C1541-1548. doi:10.1152/ajpcell.00068.2010

100

Kumar, R., Bukowski, M. J., Wider, J. M., Reynolds, C. A., Calo, L., Lepore, B., . . . Sanderson, T. H. (2016). Mitochondrial dynamics following global cerebral ischemia. Molecular and Cellular Neurosciences, 76, 68-75. doi:10.1016/j.mcn.2016.08.010 Lam, T. T., Abler, A. S., & Tso, M. O. (1999). Apoptosis and caspases after ischemia- reperfusion injury in rat retina. Investigative Ophthalmology and Visual Science, 40(5), 967-975. Langle, F., Roth, E., Steininger, R., Winkler, S., & Muhlbacher, F. (1995). Arginase release following liver reperfusion. Evidence of hemodynamic action of arginase infusions. Transplantation, 59(11), 1542-1549. Lewis, G. P., & Fisher, S. K. (2003). Up-regulation of glial fibrillary acidic protein in response to retinal injury: its potential role in glial remodeling and a comparison to vimentin expression. International Review of Cytology, 230, 263-290. Lin, W. J., & Kuang, H. Y. (2014). Oxidative stress induces autophagy in response to multiple noxious stimuli in retinal ganglion cells. Autophagy, 10(10), 1692-1701. London, A., Benhar, I., & Schwartz, M. (2013). The retina as a window to the brain-from eye research to CNS disorders. Nature Reviews: Neurology, 9(1), 44-53. doi:10.1038/nrneurol.2012.227 Loyaga-Rendon, R. Y., Sakamoto, S., Beppu, M., Aso, T., Ishizaka, M., Takahashi, R., & Azuma, H. (2005). Accumulated endogenous nitric oxide synthase inhibitors, enhanced arginase activity, attenuated dimethylarginine dimethylaminohydrolase activity and intimal hyperplasia in premenopausal human uterine arteries. Atherosclerosis, 178(2), 231-239. doi:10.1016/j.atherosclerosis.2004.09.006 Marathe, C., Bradley, M. N., Hong, C., Lopez, F., Ruiz de Galarreta, C. M., Tontonoz, P., & Castrillo, A. (2006). The arginase II gene is an anti-inflammatory target of liver X receptor in macrophages. Journal of Biological Chemistry, 281(43), 32197- 32206. doi:10.1074/jbc.M605237200 Miller, F. J., Jr., Gutterman, D. D., Rios, C. D., Heistad, D. D., & Davidson, B. L. (1998). Superoxide production in vascular smooth muscle contributes to oxidative stress and impaired relaxation in atherosclerosis. Circulation Research, 82(12), 1298- 1305. Ming, X. F., Barandier, C., Viswambharan, H., Kwak, B. R., Mach, F., Mazzolai, L., . . . Yang, Z. (2004). Thrombin stimulates human endothelial arginase enzymatic activity via RhoA/ROCK pathway: implications for atherosclerotic endothelial dysfunction. Circulation, 110(24), 3708-3714. doi:10.1161/01.cir.0000142867.26182.32 Minhas, G., Sharma, J., & Khan, N. (2016). Cellular Stress Response and Immune Signaling in Retinal Ischemia-Reperfusion Injury. Frontiers in Immunology, 7, 444. doi:10.3389/fimmu.2016.00444 Miyanaka, K., Gotoh, T., Nagasaki, A., Takeya, M., Ozaki, M., Iwase, K., . . . Mori, M. (1998). Immunohistochemical localization of arginase II and other enzymes of arginine metabolism in rat kidney and liver. Histochemical Journal, 30(10), 741- 751. Moriwaki, K., & Chan, F. K. (2013). RIP3: a molecular switch for necrosis and inflammation. Genes and Development, 27(15), 1640-1649. doi:10.1101/gad.223321.113

101

Morris, S. M., Jr. (2002). Regulation of enzymes of the urea cycle and arginine metabolism. Annual Review of Nutrition, 22, 87-105. Morris, S. M., Jr. (2016). Arginine Metabolism Revisited. Journal of Nutrition, 146(12), 2579s-2586s. doi:10.3945/jn.115.226621 Munder, M., Eichmann, K., Moran, J. M., Centeno, F., Soler, G., & Modolell, M. (1999). Th1/Th2-regulated expression of arginase isoforms in murine macrophages and dendritic cells. Journal of Immunology, 163(7), 3771-3777. Muthusamy, A., Lin, C. M., Shanmugam, S., Lindner, H. M., Abcouwer, S. F., & Antonetti, D. A. (2014). Ischemia-reperfusion injury induces occludin phosphorylation/ubiquitination and retinal vascular permeability in a VEGFR-2- dependent manner. Journal of Cerebral Blood Flow and Metabolism, 34(3), 522- 531. doi:10.1038/jcbfm.2013.230 Narayanan, S. P., Suwanpradid, J., Saul, A., Xu, Z., Still, A., Caldwell, R. W., & Caldwell, R. B. (2011). Arginase 2 deletion reduces neuro-glial injury and improves retinal function in a model of retinopathy of prematurity. PloS One, 6(7), e22460. doi:10.1371/journal.pone.0022460 Narayanan, S. P., Xu, Z., Putluri, N., Sreekumar, A., Lemtalsi, T., Caldwell, R. W., & Caldwell, R. B. (2014). Arginase 2 deficiency reduces hyperoxia-mediated retinal neurodegeneration through the regulation of polyamine metabolism. Cell Death & Disease, 5, e1075. doi:10.1038/cddis.2014.23 Nashine, S., Liu, Y., Kim, B. J., Clark, A. F., & Pang, I. H. (2015). Role of C/EBP homologous protein in retinal ganglion cell death after ischemia/reperfusion injury. Investigative Ophthalmology and Visual Science, 56(1), 221-231. doi:10.1167/iovs.14-15447 Nelin, L. D., Chicoine, L. G., Reber, K. M., English, B. K., Young, T. L., & Liu, Y. (2005). Cytokine-induced endothelial arginase expression is dependent on epidermal growth factor receptor. American Journal of Respiratory Cell and Molecular Biology, 33(4), 394-401. doi:10.1165/rcmb.2005-0039OC Ohnaka, M., Okuda-Ashitaka, E., Kaneko, S., Ando, A., Maeda, M., Furuta, K., . . . Ito, S. (2011). Induction of arginase II mRNA by nitric oxide using an in vitro model of gyrate atrophy of choroid and retina. Investigative Ophthalmology and Visual Science, 52(3), 1493-1500. doi:10.1167/iovs.10-5516 Ong, S. B., Subrayan, S., Lim, S. Y., Yellon, D. M., Davidson, S. M., & Hausenloy, D. J. (2010). Inhibiting mitochondrial fission protects the heart against ischemia/reperfusion injury. Circulation, 121(18), 2012-2022. doi:10.1161/circulationaha.109.906610 Osborne, N. N., Casson, R. J., Wood, J. P., Chidlow, G., Graham, M., & Melena, J. (2004). Retinal ischemia: mechanisms of damage and potential therapeutic strategies. Progress in Retinal and Eye Research, 23(1), 91-147. doi:10.1016/j.preteyeres.2003.12.001 Pagliuso, A., Cossart, P., & Stavru, F. (2017). The ever-growing complexity of the mitochondrial fission machinery. Cellular and Molecular Life Sciences. doi:10.1007/s00018-017-2603-0 Pandey, D., Bhunia, A., Oh, Y. J., Chang, F., Bergman, Y., Kim, J. H., . . . Romer, L. H. (2014a). OxLDL triggers retrograde translocation of arginase2 in aortic endothelial

102

cells via ROCK and mitochondrial processing peptidase. Circulation Research, 115(4), 450-459. doi:10.1161/circresaha.115.304262 Pandey, D., Sikka, G., Bergman, Y., Kim, J. H., Ryoo, S., Romer, L., & Berkowitz, D. (2014b). Transcriptional regulation of endothelial arginase 2 by histone deacetylase 2. Arteriosclerosis, Thrombosis, and Vascular Biology, 34(7), 1556-1566. doi:10.1161/atvbaha.114.303685 Park, S. W., Kim, K. Y., Lindsey, J. D., Dai, Y., Heo, H., Nguyen, D. H., . . . Ju, W. K. (2011). A selective inhibitor of drp1, mdivi-1, increases retinal ganglion cell survival in acute ischemic mouse retina. Investigative Ophthalmology and Visual Science, 52(5), 2837-2843. doi:10.1167/iovs.09-5010 Patel, C., Rojas, M., Narayanan, S. P., Zhang, W., Xu, Z., Lemtalsi, T., . . . Caldwell, R. B. (2013). Arginase as a mediator of diabetic retinopathy. Frontiers in Immunology, 4, 173. doi:10.3389/fimmu.2013.00173 Patel, C., Xu, Z., Shosha, E., Xing, J., Lucas, R., Caldwell, R. W., . . . Narayanan, S. P. (2016). Treatment with polyamine oxidase inhibitor reduces microglial activation and limits vascular injury in ischemic retinopathy. Biochimica et Biophysica Acta, 1862(9), 1628-1639. doi:10.1016/j.bbadis.2016.05.020 Pernow, J., & Jung, C. (2013). Arginase as a potential target in the treatment of cardiovascular disease: reversal of arginine steal? Cardiovascular Research, 98(3), 334-343. doi:10.1093/cvr/cvt036 Pernow, J., Kiss, A., Tratsiakovich, Y., & Climent, B. (2015). Tissue-specific upregulation of arginase I and II induced by p38 MAPK mediates endothelial dysfunction in type 1 diabetes mellitus. British Journal of Pharmacology, 172(19), 4684-4698. doi:10.1111/bph.13242 Petrone, A. B., O'Connell, G. C., Regier, M. D., Chantler, P. D., Simpkins, J. W., & Barr, T. L. (2016). The Role of Arginase 1 in Post-Stroke Immunosuppression and Ischemic Stroke Severity. Transl Stroke Res, 7(2), 103-110. doi:10.1007/s12975- 015-0431-9 Peyton, K. J., Ensenat, D., Azam, M. A., Keswani, A. N., Kannan, S., Liu, X. M., . . . Durante, W. (2009). Arginase promotes neointima formation in rat injured carotid arteries. Arteriosclerosis, Thrombosis, and Vascular Biology, 29(4), 488-494. doi:10.1161/atvbaha.108.183392 Quirie, A., Demougeot, C., Bertrand, N., Mossiat, C., Garnier, P., Marie, C., & Prigent- Tessier, A. (2013). Effect of stroke on arginase expression and localization in the rat brain. European Journal of Neuroscience, 37(7), 1193-1202. doi:10.1111/ejn.12111 Raup-Konsavage, W. M., Gao, T., Cooper, T. K., Morris, S. M., Jr., Reeves, W. B., & Awad, A. S. (2017a). Arginase-2 mediates renal ischemia-reperfusion injury. American Journal of Physiology: Renal Physiology, 313(2), F522-f534. doi:10.1152/ajprenal.00620.2016 Raup-Konsavage, W. M., Gao, T., Cooper, T. K., Morris, S. M., Jr., Reeves, W. B., & Awad, A. S. (2017b). Arginase-2 Mediates Renal Ischemia/Reperfusion Injury. American Journal of Physiology: Renal Physiology, ajprenal.00620.02016. doi:10.1152/ajprenal.00620.2016 Reid, K. M., Tsung, A., Kaizu, T., Jeyabalan, G., Ikeda, A., Shao, L., . . . Geller, D. A. (2007). Liver I/R injury is improved by the arginase inhibitor, N(omega)-hydroxy-

103

nor-L-arginine (nor-NOHA). American Journal of Physiology: Gastrointestinal and Liver Physiology, 292(2), G512-517. doi:10.1152/ajpgi.00227.2006 Rein, D. B., Zhang, P., Wirth, K. E., Lee, P. P., Hoerger, T. J., McCall, N., . . . Saaddine, J. (2006). The economic burden of major adult visual disorders in the United States. Archives of Ophthalmology, 124(12), 1754-1760. doi:10.1001/archopht.124.12.1754 Roe, N. D., & Ren, J. (2012). Nitric oxide synthase uncoupling: a therapeutic target in cardiovascular diseases. Vascular Pharmacology, 57(5-6), 168-172. doi:10.1016/j.vph.2012.02.004 Rojas, M., Lemtalsi, T., Toque, H. A., Xu, Z., Fulton, D., Caldwell, R. W., & Caldwell, R. B. (2017). NOX2-Induced Activation of Arginase and Diabetes-Induced Retinal Endothelial Cell Senescence. Antioxidants (Basel), 6(2). doi:10.3390/antiox6020043 Romero, M. J., Iddings, J. A., Platt, D. H., Ali, M. I., Cederbaum, S. D., Stepp, D. W., . . . Caldwell, R. W. (2012). Diabetes-induced vascular dysfunction involves arginase I. American Journal of Physiology: Heart and Circulatory Physiology, 302(1), H159-166. doi:10.1152/ajpheart.00774.2011 Romero, M. J., Platt, D. H., Tawfik, H. E., Labazi, M., El-Remessy, A. B., Bartoli, M., . . . Caldwell, R. W. (2008). Diabetes-induced coronary vascular dysfunction involves increased arginase activity. Circulation Research, 102(1), 95-102. doi:10.1161/circresaha.107.155028 Rosenbaum, D. M., Degterev, A., David, J., Rosenbaum, P. S., Roth, S., Grotta, J. C., . . . Savitz, S. I. (2010). Necroptosis, a novel form of caspase-independent cell death, contributes to neuronal damage in a retinal ischemia-reperfusion injury model. Journal of Neuroscience Research, 88(7), 1569-1576. doi:10.1002/jnr.22314 Rosenbaum, D. M., Rosenbaum, P. S., Gupta, H., Singh, M., Aggarwal, A., Hall, D. H., . . . Kessler, J. A. (1998). The role of the p53 protein in the selective vulnerability of the inner retina to transient ischemia. Investigative Ophthalmology and Visual Science, 39(11), 2132-2139. Ryoo, S., Bhunia, A., Chang, F., Shoukas, A., Berkowitz, D. E., & Romer, L. H. (2011). OxLDL-dependent activation of arginase II is dependent on the LOX-1 receptor and downstream RhoA signaling. Atherosclerosis, 214(2), 279-287. doi:10.1016/j.atherosclerosis.2010.10.044 Sanderson, T. H., Raghunayakula, S., & Kumar, R. (2015). Neuronal hypoxia disrupts mitochondrial fusion. Neuroscience, 301, 71-78. doi:10.1016/j.neuroscience.2015.05.078 Saszik, S. M., Robson, J. G., & Frishman, L. J. (2002). The scotopic threshold response of the dark-adapted electroretinogram of the mouse. Journal of Physiology, 543(Pt 3), 899-916. Schmid, H., Renner, M., Dick, H. B., & Joachim, S. C. (2014). Loss of inner retinal neurons after retinal ischemia in rats. Investigative Ophthalmology and Visual Science, 55(4), 2777-2787. doi:10.1167/iovs.13-13372 Shatanawi, A., Lemtalsi, T., Yao, L., Patel, C., Caldwell, R. B., & Caldwell, R. W. (2015). Angiotensin II limits NO production by upregulating arginase through a p38 MAPK-ATF-2 pathway. European Journal of Pharmacology, 746, 106-114. doi:10.1016/j.ejphar.2014.10.042

104

Shatanawi, A., Romero, M. J., Iddings, J. A., Chandra, S., Umapathy, N. S., Verin, A. D., . . . Caldwell, R. W. (2011). Angiotensin II-induced vascular endothelial dysfunction through RhoA/Rho kinase/p38 mitogen-activated protein kinase/arginase pathway. American Journal of Physiology: Cell Physiology, 300(5), C1181-1192. doi:10.1152/ajpcell.00328.2010 Sheldon, K. E., Shandilya, H., Kepka-Lenhart, D., Poljakovic, M., Ghosh, A., & Morris, S. M., Jr. (2013). Shaping the murine macrophage phenotype: IL-4 and cyclic AMP synergistically activate the arginase I promoter. Journal of Immunology, 191(5), 2290-2298. doi:10.4049/jimmunol.1202102 Shen, Z., Zheng, Y., Wu, J., Chen, Y., Wu, X., Zhou, Y., . . . Zhang, X. (2017). PARK2- dependent mitophagy induced by acidic postconditioning protects against focal cerebral ischemia and extends the reperfusion window. Autophagy, 13(3), 473-485. doi:10.1080/15548627.2016.1274596 Sheng, R., Zhang, L. S., Han, R., Liu, X. Q., Gao, B., & Qin, Z. H. (2010). Autophagy activation is associated with neuroprotection in a rat model of focal cerebral ischemic preconditioning. Autophagy, 6(4), 482-494. doi:10.4161/auto.6.4.11737 Shibuki, H., Katai, N., Yodoi, J., Uchida, K., & Yoshimura, N. (2000). Lipid peroxidation and peroxynitrite in retinal ischemia-reperfusion injury. Investigative Ophthalmology and Visual Science, 41(11), 3607-3614. Shin, W. S., Berkowitz, D. E., & Ryoo, S. W. (2012). Increased arginase II activity contributes to endothelial dysfunction through endothelial nitric oxide synthase uncoupling in aged mice. Experimental and Molecular Medicine, 44(10), 594-602. doi:10.3858/emm.2012.44.10.068 Shosha, E., Xu, Z., Yokota, H., Saul, A., Rojas, M., Caldwell, R. W., . . . Narayanan, S. P. (2016). Arginase 2 promotes neurovascular degeneration during ischemia/reperfusion injury. Cell Death & Disease, 7(11), e2483. doi:10.1038/cddis.2016.295 Smith, G. G., & Baird, C. D. (1952). Survival time of retinal cells when deprived of their blood supply by increased intraocular pressure. American Journal of Ophthalmology, 35(5 2), 133-136. Smith, L. E., Wesolowski, E., McLellan, A., Kostyk, S. K., D'Amato, R., Sullivan, R., & D'Amore, P. A. (1994). Oxygen-induced retinopathy in the mouse. Investigative Ophthalmology and Visual Science, 35(1), 101-111. Stevenson, L., Matesanz, N., Colhoun, L., Edgar, K., Devine, A., Gardiner, T. A., & McDonald, D. M. (2010). Reduced nitro-oxidative stress and neural cell death suggests a protective role for microglial cells in TNFalpha-/- mice in ischemic retinopathy. Investigative Ophthalmology and Visual Science, 51(6), 3291-3299. doi:10.1167/iovs.09-4344 Sun, Y., Xue, W., Song, Z., Huang, K., & Zheng, L. (2016). Restoration of Opa1-long isoform inhibits retinal injury-induced neurodegeneration. Journal of Molecular Medicine (Berlin, Germany), 94(3), 335-346. doi:10.1007/s00109-015-1359-y Suwanpradid, J., Rojas, M., Behzadian, M. A., Caldwell, R. W., & Caldwell, R. B. (2014). Arginase 2 deficiency prevents oxidative stress and limits hyperoxia-induced retinal vascular degeneration. PloS One, 9(11), e110604. doi:10.1371/journal.pone.0110604

105

Takahashi, K., Lam, T. T., Edward, D. P., Buchi, E. R., & Tso, M. O. (1992). Protective effects of flunarizine on ischemic injury in the rat retina. Archives of Ophthalmology, 110(6), 862-870. Tang, J., Hu, Z., Tan, J., Yang, S., & Zeng, L. (2016). Parkin Protects against Oxygen- Glucose Deprivation/Reperfusion Insult by Promoting Drp1 Degradation. Oxidative Medicine and Cellular Longevity, 2016, 8474303. doi:10.1155/2016/8474303 Tong, N., Zhang, Z., Zhang, W., Qiu, Y., Gong, Y., Yin, L., . . . Wu, X. (2013). Diosmin alleviates retinal edema by protecting the blood-retinal barrier and reducing retinal vascular permeability during ischemia/reperfusion injury. PloS One, 8(4), e61794. doi:10.1371/journal.pone.0061794 Toque, H. A., Nunes, K. P., Rojas, M., Bhatta, A., Yao, L., Xu, Z., . . . Caldwell, R. W. (2013). Arginase 1 mediates increased blood pressure and contributes to vascular endothelial dysfunction in deoxycorticosterone acetate-salt hypertension. Frontiers in Immunology, 4, 219. doi:10.3389/fimmu.2013.00219 Tratsiakovich, Y., Gonon, A. T., Krook, A., Yang, J., Shemyakin, A., Sjoquist, P. O., & Pernow, J. (2013a). Arginase inhibition reduces infarct size via nitric oxide, protein kinase C epsilon and mitochondrial ATP-dependent K+ channels. European Journal of Pharmacology, 712(1-3), 16-21. doi:10.1016/j.ejphar.2013.04.044 Tratsiakovich, Y., Kiss, A., Gonon, A. T., Yang, J., Sjoquist, P. O., & Pernow, J. (2017). Inhibition of Rho kinase protects from ischaemia-reperfusion injury via regulation of arginase activity and nitric oxide synthase in type 1 diabetes. Diabetes & Vascular Disease Research, 14(3), 236-245. doi:10.1177/1479164116687935 Tratsiakovich, Y., Yang, J., Gonon, A. T., Sjoquist, P. O., & Pernow, J. (2013b). Arginase as a target for treatment of myocardial ischemia-reperfusion injury. European Journal of Pharmacology, 720(1-3), 121-123. doi:10.1016/j.ejphar.2013.10.040 Villalba, N., Sackheim, A. M., Nunez, I. A., Hill-Eubanks, D. C., Nelson, M. T., Wellman, G. C., & Freeman, K. (2017). Traumatic Brain Injury Causes Endothelial Dysfunction in the Systemic Microcirculation through Arginase-1-Dependent Uncoupling of Endothelial Nitric Oxide Synthase. Journal of Neurotrauma, 34(1), 192-203. doi:10.1089/neu.2015.4340 Vockley, J. G., Jenkinson, C. P., Shukla, H., Kern, R. M., Grody, W. W., & Cederbaum, S. D. (1996). Cloning and characterization of the human type II arginase gene. Genomics, 38(2), 118-123. doi:10.1006/geno.1996.0606 Wang, H., Zheng, S., Liu, M., Jia, C., Wang, S., Wang, X., . . . Guo, Y. (2016). The Effect of Propofol on Mitochondrial Fission during Oxygen-Glucose Deprivation and Reperfusion Injury in Rat Hippocampal Neurons. PloS One, 11(10), e0165052. doi:10.1371/journal.pone.0165052 Wang, L., Li, C., Guo, H., Kern, T. S., Huang, K., & Zheng, L. (2011). Curcumin inhibits neuronal and vascular degeneration in retina after ischemia and reperfusion injury. PloS One, 6(8), e23194. doi:10.1371/journal.pone.0023194 Wang, X., Jiang, W., Yan, Y., Gong, T., Han, J., Tian, Z., & Zhou, R. (2014). RNA viruses promote activation of the NLRP3 inflammasome through a RIP1-RIP3-DRP1 signaling pathway. Nature Immunology, 15(12), 1126-1133. doi:10.1038/ni.3015 Wei, L. H., Wu, G., Morris, S. M., Jr., & Ignarro, L. J. (2001). Elevated arginase I expression in rat aortic smooth muscle cells increases cell proliferation.

106

Proceedings of the National Academy of Sciences of the United States of America, 98(16), 9260-9264. doi:10.1073/pnas.161294898 Wei, Y., Gong, J., Yoshida, T., Eberhart, C. G., Xu, Z., Kombairaju, P., . . . Duh, E. J. (2011). Nrf2 has a protective role against neuronal and capillary degeneration in retinal ischemia-reperfusion injury. Free Radical Biology and Medicine, 51(1), 216-224. doi:10.1016/j.freeradbiomed.2011.04.026 Wilson, A. M., Harada, R., Nair, N., Balasubramanian, N., & Cooke, J. P. (2007). L- arginine supplementation in peripheral arterial disease: no benefit and possible harm. Circulation, 116(2), 188-195. doi:10.1161/circulationaha.106.683656 Wu, G., & Morris, S. M., Jr. (1998). Arginine metabolism: nitric oxide and beyond. Biochemical Journal, 336 ( Pt 1), 1-17. Xu, Z., Cho, H., Hartsock, M. J., Mitchell, K. L., Gong, J., Wu, L., . . . Duh, E. J. (2015). Neuroprotective role of Nrf2 for retinal ganglion cells in ischemia-reperfusion. Journal of Neurochemistry, 133(2), 233-241. doi:10.1111/jnc.13064 Yang, L., Tan, P., Zhou, W., Zhu, X., Cui, Y., Zhu, L., . . . Chen, H. (2012). N- acetylcysteine protects against hypoxia mimetic-induced autophagy by targeting the HIF-1alpha pathway in retinal ganglion cells. Cellular and Molecular Neurobiology, 32(8), 1275-1285. doi:10.1007/s10571-012-9852-0 Yang, Z., & Ming, X. F. (2006). Endothelial arginase: a new target in atherosclerosis. Current Hypertension Reports, 8(1), 54-59. Yang, Z., & Ming, X. F. (2013). Arginase: the emerging therapeutic target for vascular oxidative stress and inflammation. Frontiers in Immunology, 4, 149. doi:10.3389/fimmu.2013.00149 Yang, Z., & Ming, X. F. (2014). Functions of arginase isoforms in macrophage inflammatory responses: impact on cardiovascular diseases and metabolic disorders. Frontiers in Immunology, 5, 533. doi:10.3389/fimmu.2014.00533 Yokota, H., Narayanan, S. P., Zhang, W., Liu, H., Rojas, M., Xu, Z., . . . Caldwell, R. B. (2011). Neuroprotection from retinal ischemia/reperfusion injury by NOX2 NADPH oxidase deletion. Investigative Ophthalmology and Visual Science, 52(11), 8123-8131. doi:10.1167/iovs.11-8318 Yu, H., Iyer, R. K., Kern, R. M., Rodriguez, W. I., Grody, W. W., & Cederbaum, S. D. (2001). Expression of arginase isozymes in mouse brain. Journal of Neuroscience Research, 66(3), 406-422. doi:10.1002/jnr.1233 Yu, Y., Rajapakse, A. G., Montani, J. P., Yang, Z., & Ming, X. F. (2014). p38 mitogen- activated protein kinase is involved in arginase-II-mediated eNOS-Uncoupling in Obesity. Cardiovascular Diabetology, 13(1), 113. doi:10.1186/s12933-014-0113- z Yuan, Y., Zheng, Y., Zhang, X., Chen, Y., Wu, X., Wu, J., . . . Chen, Z. (2017). BNIP3L/NIX-mediated mitophagy protects against ischemic brain injury independent of PARK2. Autophagy, 1-13. doi:10.1080/15548627.2017.1357792 Zeng, X. X., Ng, Y. K., & Ling, E. A. (2000). Neuronal and microglial response in the retina of streptozotocin-induced diabetic rats. Visual Neuroscience, 17(3), 463-471. Zhang, C., Wu, J., Xu, X., Potter, B. J., & Gao, X. (2010). Direct relationship between levels of TNF-alpha expression and endothelial dysfunction in reperfusion injury. Basic Research in Cardiology, 105(4), 453-464. doi:10.1007/s00395-010-0083-6

107

Zhang, W., Baban, B., Rojas, M., Tofigh, S., Virmani, S. K., Patel, C., . . . Caldwell, R. B. (2009). Arginase activity mediates retinal inflammation in endotoxin-induced uveitis. American Journal of Pathology, 175(2), 891-902. doi:10.2353/ajpath.2009.081115 Zhang, W., Liu, H., Rojas, M., Caldwell, R. W., & Caldwell, R. B. (2011). Anti- inflammatory therapy for diabetic retinopathy. Immunotherapy, 3(5), 609-628. doi:10.2217/imt.11.24 Zhang, X., Yan, H., Yuan, Y., Gao, J., Shen, Z., Cheng, Y., . . . Chen, Z. (2013). Cerebral ischemia-reperfusion-induced autophagy protects against neuronal injury by mitochondrial clearance. Autophagy, 9(9), 1321-1333. doi:10.4161/auto.25132 Zhao, H., Joseph, J., Fales, H. M., Sokoloski, E. A., Levine, R. L., Vasquez-Vivar, J., & Kalyanaraman, B. (2005). Detection and characterization of the product of hydroethidine and intracellular superoxide by HPLC and limitations of fluorescence. Proceedings of the National Academy of Sciences of the United States of America, 102(16), 5727-5732. doi:10.1073/pnas.0501719102 Zhao, Y. X., Cui, M., Chen, S. F., Dong, Q., & Liu, X. Y. (2014). Amelioration of ischemic mitochondrial injury and Bax-dependent outer membrane permeabilization by Mdivi-1. CNS Neuroscience & Therapeutics, 20(6), 528-538. doi:10.1111/cns.12266 Zheng, L., Gong, B., Hatala, D. A., & Kern, T. S. (2007). Retinal ischemia and reperfusion causes capillary degeneration: similarities to diabetes. Investigative Ophthalmology and Visual Science, 48(1), 361-367. doi:10.1167/iovs.06-0510 Zuo, W., Zhang, S., Xia, C. Y., Guo, X. F., He, W. B., & Chen, N. H. (2014). Mitochondria autophagy is induced after hypoxic/ischemic stress in a Drp1 dependent manner: the role of inhibition of Drp1 in ischemic brain damage. Neuropharmacology, 86, 103-115. doi:10.1016/j.neuropharm.2014.07.002

108