survey of 61 (2016) 164e186

Available online at www.sciencedirect.com ScienceDirect

journal homepage: www.elsevier.com/locate/survophthal

Major review Autoregulation and neurovascular coupling in the head

Daniele Prada, PhDa,*, Alon Harris, MS, PhD, FARVOb, Giovanna Guidoboni, PhDa,b,c, Brent Siesky, PhDb, Amelia M. Huang, BSb, Julia Arciero, PhDa a Department of Mathematical Sciences, Indiana University-Purdue University Indianapolis (IUPUI), Indianapolis, Indiana, USA b Department of Ophthalmology, Indiana University School of , Indianapolis, Indiana, USA c LabEx IRMIA, University of Strasbourg, Strasbourg, Alsace, France article info abstract

Article history: Impairments of autoregulation and neurovascular coupling in the optic nerve head play a Received 8 May 2015 critical role in ocular , especially glaucomatous . We critically Received in revised form review the literature in the field, integrating results obtained in clinical, experimental, and 2 October 2015 theoretical studies. We address the mechanisms of autoregulation and neurovascular Accepted 2 October 2015 coupling in the optic nerve head, the current methods used to assess autor- Available online 20 October 2015 egulationdincluding measurements of optic nerve head blood flow (or volume and velocity)dblood flow data collected in the optic nerve head as pressure or metabolic Keywords: demand is varied in healthy and pathologic conditions, and the current status and autoregulation in the optic nerve potential of mathematical modeling work to further the understanding of the relationship head between ocular blood flow mechanisms and diseases such as . neurovascular coupling ª 2016 Elsevier Inc. All rights reserved. ocular biomechanics and mechanical influences on autoregulation metabolic autoregulation effects of impaired blood flow regulation optic neuropathies glaucoma blood flow regulation modeling

* Corresponding author: Mr. Daniele Prada, PhD, Department of Mathematical Sciences, Indiana University-Purdue University Indi- anapolis (IUPUI), 402 North Blackford Street, LD 270, Indianapolis, IN 46202, USA. Tel.: þ1 317 274 3460. E-mail address: [email protected] (D. Prada). 0039-6257/$ e see front matter ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.survophthal.2015.10.004 survey of ophthalmology 61 (2016) 164e186 165

1. Introduction Table 1 e Table with abbreviations Abbreviation Full name Glaucoma is an optic neuropathy characterized by progressive 20-HETE 20-hydroxy-eicosatetraenoic acid death of retinal ganglion cells (RGCs) and irreversible visual loss. BP Blood pressure Glaucoma is the second leading cause of blindness world- CCD Charge coupled device 177 wide, and yet its etiology and treatment remain unclear. The CDI Color Doppler imaging main modifiable risk factor in glaucoma patients is elevated CO2 Carbon dioxide (IOP)1,45,110,118,121; however, a high percent- CRA Central retinal age of individuals with elevated IOP (a condition called ocular CRV Central retinal vein CSFp Cerebrospinal fluid pressure hypertension) never develop glaucoma,103 and many glaucoma dAR Dynamic autoregulation patients continue to experience disease progression despite EC Endothelial cell lowering IOP to target levels or have no history of elevated IOPea EDV End diastolic velocity condition called normal tension glaucoma (NTG).203 EET Epoxyeicosatrienoic acid Several studies suggest correlations between impaired ET Endothelin ocular blood flow and glaucoma.56,60,67,84,85,90,241 In healthy IOP Intraocular pressure LDF Laser Doppler flowmetry conditions, vascular beds exhibit an intrinsic ability to main- LSFG Laser speckle flowgraphy tain relatively constant blood flow over a large range of arterial MAP Mean arterial pressure pressures. This autoregulatory behavior is recognized in most NB Normalized blur 3,169 162 21 vascular bedsdincluding the , brain, heart, kid- NO Nitric oxide ney,178 skeletal muscle,61 and gut129dbut the effectiveness of NOS Nitric oxide synthase autoregulation differs among these vascular beds according to NTG Normal tension glaucoma importance of function. For example, the brain and kidney OCT Optical coherence tomography ONH Optic nerve head receive stable flow over a range of arterial pressure,32,162 OPP Ocular perfusion pressure whereas autoregulation in other beds such as the gut is less PCA Posterior ciliary artery effective. In the eye, the retinal and optic nerve head (ONH) PGE2 Prostaglandin E2 vascular beds are known to exhibit autoregulation, though to PI Pulsatility index differing extents. Details and experimental measures of PO2 Oxygen partial pressure autoregulation are better established in the than in the POAG Primary open-angle glaucoma ONH. In experiments assessing hemodynamic responses to PSV Peak systolic velocity RGC Retinal ganglion cell light stimulation,68,69,184,186 blood flow in the retina and ONH RI Resistive index seems to be highly correlated to increased neural activity. This RLTp Retro laminar tissue pressure 130 phenomenon is called neurovascular coupling. ROS Reactive oxygen species In glaucoma the location of damage to nerve cells is sAR Static autoregulation hypothesized to be predominantly in the ONH,176 and thus a SBR Square blur ratio clearer understanding of the factors affecting the blood supply SNFL Superficial nerve fiber layer SSADA Split spectrum amplitude to the ONH is necessary to determine how this may be decorrelation angiography compromised and potentially contribute to the pathophysi- ology of glaucoma. The aim of this review is to 1) summarize the mechanisms of autoregulation and neurovascular coupling that function in astrocytes, a particular type of glial cell.28 For the purpose of the ONH; 2) describe the current ability to assess autor- description, the anatomy and vascular supply of the ONH is egulation in the ONH using methodologies capable of deter- best divided into 4 regions, from anterior to posterior seg- mining ONH blood flow (or volume and velocity); 3) compare ments (see Fig. 1). data on blood flow for varying pressure or metabolic needs in The most anterior part of the ONH is the superficial nerve the ONH to assess autoregulation in healthy and pathologic fiber layer (SNFL). Some vascular details of this layer can be conditions; and 4) describe the current status of ophthalmic resolved on ophthalmoscopy examination or angiography. A research and support the potential of mathematical modeling part of the appearance of the SNFL comes from light back- to further the understanding of the relationship between scattered from deeper tissue.51 Immediately behind the SNFL ocular blood flow mechanisms and ocular diseases such as is the “prelaminar region,” which lies adjacent to the peri- glaucoma. In order to help the reader, a list of the acronyms papillary . Posterior to the prelaminar region, the used in this paper is provided in Table 1. “laminar region” is composed of the lamina cribrosa, a structure consisting of fenestrated connective tissue beams through which the RGC axons pass on their path from the 2. Anatomy and vascular supply of the ONH retina to the optic nerve. Finally, the “retrolaminar region” lies posterior to the lamina cribrosa. It is marked by the 2.1. Anatomy beginning of axonal myelination and is surrounded by meninges. The ONH is where RGC axons leave the eye through the scleral The lamina cribrosa bears the translaminar pressure portion of the neural canal, forming bundles separated by difference: the difference between the IOP, which is the 166 survey of ophthalmology 61 (2016) 164e186

Fig. 1 e Anatomy and vascular supply of the optic nerve head (ONH). The ONH includes the superficial nerve fiber layer, the prelaminar region, the laminar region, and the retrolaminar region.

pressure in the intraocular space, and the retrolaminar astrocytes. From there, they may go into the astrocytes or tissue pressure, which is the pressure in the retrolaminar percolate in the extracellular space between them, ultimately region. The retrolaminar tissue pressure is usually lower reaching the adjacent axons. than the IOP and is strongly correlated to the cerebrospinal fluid pressure and the pressure in the subarachnoid space 2.2. Vascular supply of the optic nerve when cerebrospinal fluid pressure > 2mmHg(1mmHgz 133.3224 Pa).134,142 A hoop stress is The vascular system nourishing the ONH is quite complex89,92 also transferred to the lamina by the .28 There is and shows high interindividual and intraindividual vari- evidence that an annulus of collagen fibrils exists around ability.94,95,211 An important anatomic distinction between the the scleral canal in the peripapillary sclera. These fibrils different portions of the ONH is that blood flow to the ONH is appear to be oriented mostly radially in the periphery of primarily supplied by the posterior ciliary (PCAs), the lamina.79,188,240 The peripapillary annulus significantly whereas the SNFL receives oxygenated blood primarily from reduces the IOP-related expansion of the scleral canal and retinal arterioles.155 These small vessels, called “epipapillary shields the lamina from high-tensile stress. The radially vessels,” originate in the peripapillary SNFL and run toward oriented fibrils in the lamina periphery reinforce the lam- the center of the ONH (see Fig. 2). ina against transversal shear stresses and reduce laminar In approximately 30% of all people, a cilioretinal artery bending deformations.79 The lamina cribrosa remodels may be present and supply the temporal SNFL. This artery, if into a thicker, more posterior structure, which in- present, may be a direct branch of the ciliary or choroidal corporates more connective tissue after chronic IOP arteries, emerging from the temporal SNFL of the ONH and elevation.80,188 extending laterally along the papillomacular bundle. The In the prelaminar, laminar, and retrolaminar regions, RGC retinal arteries and the cilioretinal arteries lack anastomotic axons are surrounded by astrocytes, which are believed blood exchange in the case of an artery occlusion, leading to to maintain the homeostasis of the extracellular environ- an ischemic infarct in the area supplied by the artery or its ment. In particular, astrocytes remove potassium and gluta- branches.161 mate from the extracellular space, provide cellular support to The prelaminar region is mainly supplied by direct the axons, and synthesize extracellular matrix macromole- branches of the short PCAs and by branches from the circle of cules.28,101 In the prelaminar and retrolaminar region, it is Zinn-Haller (see Fig. 3). presumed that nutrient delivery to the axons occurs both via The circle of Zinn-Haller, if present, is a complete or diffusion and advection.28 In the laminar region, the extra- incomplete ring of arterioles within the perineural sclera cellular matrix of laminar beams lies in between capillaries formed by the confluence of branches of the short PCAs. The and astrocytes. Consequently, nutrients likely diffuse from arterial circle branches into the prelaminar region, lamina laminar capillaries, across the endothelial and pericyte cribrosa, and retrolaminar pial system and supplies the peri- basement membranes, through the extracellular matrix of papillary choroid. This vascular ring can be recognized in vivo the laminar beams, across the basement membranes of using indocynanine green videoangiography in highly myopic survey of ophthalmology 61 (2016) 164e186 167

Fig. 2 e Superficial nerve fiber layer (SNFL). The SNFL receives oxygenated blood primarily from retinal arterioles. These small vessels, called epipapillary vessels, originate in the peripapillary SNFL and run toward the center of the optic nerve head.

.153 These vessels exhibit an anastomotic blood the lamina is viewed as a set of stacked perforated sheets exchange,17 but it is unclear whether this exchange can containing vessels, with pores in each sheet aligned to create counterbalance an insufficiency of a single PCA. There is also tunnel for bundles of nerve fibers to exit from the eye.8,9 evidence of direct arterial supply to the prelaminar layer Unlike in vivo imaging, histology imaging suffers from arising from the choroidal vasculature,92 even though the distortions because of the loss of pressure (IOP, intracranial extent to which it contributes to the perfusion of the region is pressure, and blood pressure), distortions during tissue prep- still a matter of debate.89 Blood flow to the laminar region is aration, and tissue degradation after death. Nevertheless, care provided by centripetal branches of the short PCAs (see Fig. 4). is needed when comparing optical coherence tomography Such a 3-D architecture differs from, without effectively (OCT) results to histology because of differences in optical denying, what is proposed in some histology studies, where resolution and sampling density. OCT has significantly worse

Fig. 3 e Prelaminar region. The prelaminar region is mainly supplied by direct branches of the short posterior ciliary arteries (PCAs) and by branches of the circle of Zinn-Haller. The circle of Zinn-Haller, if present, is a complete or incomplete ring of arterioles within the perineural sclera formed by the confluence of branches of the short PCAs. 168 survey of ophthalmology 61 (2016) 164e186

Fig. 4 e Laminar region. Blood flow to the laminar region is provided by centripetal branches of the short PCAs. The centripetal branches arise either directly from the short PCAs or from the circle of Zinn-Haller. The lamina cribrosa is shown as a 3-D network, as suggested in recent in vivo imaging studies based on optical coherence tomography (OCT)106,146,215,235,236 and in finite element modeling studies of the lamina cribrosa microarchitecture.29,55,188

lateral resolution when compared with electron microscopy intraorbital optic nerve is innervated, but innervation stops (at or other forms of microscopy used to study the lamina cri- least) anterior to the lamina cribrosa, and it does not follow brosa, and it likely overemphasizes beams, compared to his- the branches of the CRA inside the eye.243 Neurotransmitter tology.236 Hence, many questions still need to be answered to receptors, however, are present on the surface of retinal ves- characterize the 3-D geometry of the lamina cribrosa sels.58,104 In addition, normal retinal vessels lack fenestra- accurately.209 tions.161 Hence, vasoactive hormones cannot leak from The centripetal branches arise either directly from the capillaries and reach the muscular coat of nearby arterioles short PCAs or from the circle of Zinn-Haller. These precapil- where they can influence blood flow. The branches of the PCA lary branches perforate the outer boundary of the lamina and that feed the intrascleral portion of the optic nerve may or then branch into an intraseptal capillary network, which runs may not be innervated and/or fenestrated. Such knowledge is inside the laminar beams. It is still unclear whether there are crucial to understand how blood flow is regulated in the ONH. anastomoses between the capillary or precapillary bed of the Venous drainage of the ONH occurs primarily through the laminar region, the prelaminar region, and the SNFL region. If central retinal vein (CRV). In the SNFL, blood is drained directly these anastomoses exist, it is unclear whether they play a role into the retinal veins, which then join to form the CRV. In the when a sudden (or slowly progressive) vascular occlusion on prelaminar, laminar, and retrolaminar regions, venous the precapillary or intracapillary level happens.161 The retro- drainage occurs via the CRV or axial tributaries to the CRV. laminar region is supplied by the (CRA) and the pial system (see Fig. 5). The pial system is an anas- tomosing network of capillaries located immediately within the pia mater. The pial system originates from the circle of 3. Techniques for in vivo studies of ONH Zinn-Haller and may also be fed directly by the short PCAs. hemodynamics The branches of the pial system extend centripetally to nourish the axons of the optic nerve. The CRA may supply As described in section 2, the complex vasculature of the several small intraneural branches in the retrolaminar region. ONH is comprised of small diameter vessels arranged in Some of these branches may also anastomose with the pial an intricate 3-dimensional geometry. At present, no system. technology allows a noninvasive measurement of volu- In the ONH the capillaries form a continuous network metric blood flow in absolute units; however, some he- throughout its entire length, being continuous posteriorly modynamic measurement techniques provide surrogates with those in the rest of the optic nerve and anteriorly with for ONH blood flow in arbitrary units. Four of these the adjacent retinal capillaries.7,122 It is unclear whether this measurement techniques for in vivo studies of ONH he- implies that blood flow regulation is similar122 or not92 in modynamics are discussed and compared in the following both vascular regions, independent of the arterial source. sections. Table 2 summarizes their main features, ad- Critical questions remain unanswered. The CRA within the vantages, and limitations. survey of ophthalmology 61 (2016) 164e186 169

Fig. 5 e Retrolaminar region. The retrolaminar region is supplied by the central retinal artery (CRA) and the pial system. The pial system is an anastomosing network of capillaries located immediately within the pia mater.

3.1. Laser Doppler flowmetry through a specific part of a capillary at a given time. The main advantage of LDF is its ability to measure 3 different Laser Doppler flowmetry (LDF) is a noninvasive method of hemodynamic parameters; however, LDF only provides assessing blood flow and perfusion in the ONH. LDF is based measurements of blood perfusion in arbitrary (and not ab- on the Doppler effect. It measures the shift in frequency that solute) units, which limits its usefulness in a clinical occurs when light is scattered by the red blood cells moving setting.183 Moreover, LDF measurements depend signifi- through capillaries. LDF uses a fundus camera and a com- cantly on the depth of the sampled tissue. This depth de- puter system to detect these changes in frequency. This termines the relative contribution to the Doppler signal of information is used to calculate 3 hemodynamic parame- the superficial layers, the layers supplied by the CRA, and ters: velocity, blood volume, and blood flow within the ONH. the deeper layers supplied by the short PCAs. Blood flow Velocity is defined as the average speed of red blood cells autoregulation may or may not differ within these vascular traveling through capillaries and is proportional to the beds. In a study on monkey eyes, LDF appeared to be more mean change in Doppler shift frequency. Blood volume is heavily influenced by blood flow changes in the more su- defined as the number of red blood cells in the given sample. perficial layers of the ONH than in deeper ones, but to what Blood flow or flux is defined as the flux of red blood cells extent remains uncertain.187

Table 2 e Techniques for in vivo studies of ONH hemodynamics Technique Measurements Location of Advantages Limitations measurement

Laser Doppler Velocity, volume, and flow ONH Multiple hemodynamic No absolute measurements; flowmetry in arbitrary units parameters no interindividual comparisons OCT angiography Flow in arbitrary units ONH High quality images, fast No absolute measurements; difficult to localize measurements Color Doppler Velocity , Vessel selective, no need for No absolute measurements; imaging CRA, and short PCAs dilation, clear media, or velocity measurements only; fixation expert operator required Laser speckle Velocity, flow in arbitrary ONH Time evolution of velocity at No absolute measurements; no flowgraphy units the same site of the same eye interindividual comparisons

CRA, central retinal artery; OCT, optical coherence tomography; ONH, optic nerve head; PCA, posterior ciliary artery. 170 survey of ophthalmology 61 (2016) 164e186

3.2. OCT angiography The major advantages of CDI are that it is noninvasive, vessel selective, reproducible, and does not require pupil OCT angiography, a combination of high speed OCT and a new dilation, clear media, or fixation. CDI is limited in its ability to 3D angiography system called split-spectrum amplitude- measure only velocity (not flow) and calculate vascular decorrelation angiography, is a noninvasive method used to resistance and requires for an experienced operator to obtain estimate blood flow in the ONH, especially within the micro- accurate results.87 CDI has particular difficulty imaging and circulation.108 It computes the flow index, which is a surrogate interpreting small vessels, and the PCAs are close to the size for blood flow in arbitrary units. limit that can be studied. OCT is a technique that takes cross-sectional images of a biologic tissue using a low-coherence interferometer. These 3.4. Laser speckle flowgraphy cross-section images are captured using a low-coherence beam directed at the target tissue. The light signals reflect Laser speckle flowgraphy (LSFG) is a noninvasive method of off of the tissue back to the interferometer, which stacks a measuring blood flow and velocity in the ONH. LSFG measures series of longitudinal tomographic b-scans to derive a blood flow by using the laser speckle phenomenon, which is an 3-dimensional image. Doppler OCT, a commonly used subtype interference event that occurs when laser light scatters off of a of OCT, can detect the Doppler frequency shift of the reflected diffusing tissue. This creates a speckle pattern that varies in light, providing additional information on blood flow. The proportion to the velocity of red blood cells and thus represents split-spectrum amplitude-decorrelation angiography algo- capillary blood flow. The faster the velocity of the red blood rithm allows 3-dimensional angiography to be done 4 times cells, the greater the rate of pattern variation. Although the faster than previous algorithms and also improves blood flow velocity cannot be measured directly, the normalized blur and detection, creates better visualization of the microvascula- square blur ratio values can be calculated as quantitative in- ture, and removes motion errors automatically.105 dicators of blood velocity. Normalized blur values are well OCT angiography has many advantages over Doppler correlated with blood flow measurements simultaneously OCT. Doppler OCT can only quantify blood flow in large taken with the hydrogen gas clearance method, colored mi- superficial vessels of the ONH and cannot visualize the crospheres technique, and other methods in the ONH, , e microvasculature.105 OCT angiography minimizes the pul- choroid, and retina.218,220 226 The distribution of blood flow can satory bulk motion noise along the axial direction and be displayed in a 2 dimensional color-coded map, which re- optimizes flow detection along the transverse direction.108 flects the time variation of the speckles at each pixel point.217 As with all measuring techniques, OCT angiography has This allows for visualization of blood flow in real time. limitations. One of the main disadvantages is that blood LSFG uses a diode laser, image sensor, infrared charge- flow from superficial layers can be projected to deeper coupled device camera, and digital charge-coupled device layers, thereby incorrectly indicating that the imaged blood camera. The diode laser and image sensor are used for laser flow is a few layers deeper than its location in vivo.108 Also, speckle measurements. The laser is focused on the image split-spectrum amplitude-decorrelation angiography cannot sensor and creates a speckle pattern, which is scanned at 512 distinguish between perfusion defects caused by damaged scans/second.217 The digital charge-coupled device camera tissue or ischemia,109 and ONH blood flow estimates are measures vessel diameter and takes pictures of the fundus. provided in arbitrary units. Despite these limitations, OCT The advantages of LSFG are that its results are adequately angiography is a very useful tool to measure blood flow in reproducible and that the change in velocity at the same site the ONH. of the same eye can be followed over time. A major disad- vantage of LSFG is that the meaning of its measurement is not 3.3. Color Doppler imaging clearly understood and does not allow for intereye or inter- individual comparisons.87 Color Doppler imaging (CDI), also known as color Doppler ultrasound, is a noninvasive procedure that allows the user to visualize a color-coded image of blood velocity against a gray- 4. Evidence of blood flow autoregulation in scale image of the surrounding structures. This technique the ONH uses the principle of Doppler frequency shift to measure blood flow velocity in absolute units. Various transducers are used Autoregulation is the intrinsic ability of vascular beds to main- to measure the Doppler frequency shift, which produces color tain relatively constant blood flow over a large range of pressure, pixels.239 The color red represents blood flowing toward the while meeting the metabolic demand of the tissue. Autor- ultrasound probe, whereas blue represents blood flowing egulation is evaluated most often on a flow versus pressure away from the probe.214 graph, where pressure may be expressed as mean arterial pres- CDI is most commonly used to measure the peak systolic sure (MAP), IOP, or ocular perfusion pressure (OPP) (see Fig. 6). velocity (PSV) and end diastolic velocity (EDV), which are then OPP refers to the arterovenous pressure difference driving used to measure the resistive index [RI ¼ (PSVEDV)/PSV] and blood flow through the intraocular vasculature. The intraoc- pulsatility index [PI ¼ (PSVEDV)/Tmax, where Tmax is the ular venous pressure is very close to IOP,15,46,73,81 and thus, time averaged peak velocity]. These values estimate resis- OPP is usually estimated as the difference between the arterial tance to blood flow caused by the microvasculature distal to BP and IOP in the upright position. OPP may be defined as the site of measurement. The RI is particularly suitable for mean, systolic, or diastolic OPP. Mean OPP is typically calcu- investigating the low resistance retrobulbar vasculature.239 lated as survey of ophthalmology 61 (2016) 164e186 171

4.1. Evidence of ONH autoregulatory capacity during OPP changes in healthy subjects

Autoregulation In several animal70,186,212 and human studies,166,181 the ONH plateau vascular bed was shown to maintain autoregulatory capacity over a wide range of perfusion pressures. Autoregulatory ca- pacity is conventionally assessed by a “two-point” blood flow measurement: blood flow or other hemodynamic parameters are measured before and after the OPP is artificially modified by a step challenge in either the IOP or the systemic arterial BP.

Steady-state blood flow If blood flow changes significantly from normal after the pressure challenge, then autoregulation is said to be impaired; OPP if blood flow remains nearly constant over the pressure change, autoregulation is said to be intact. Fig. 6 e A schematic representation of a static When the pressure step challenge is applied rapidly, it autoregulation curve describing the relationship between induces 2 phases of hemodynamic response: 1) an initial steady-state blood flow responses in the ONH and OPP. transient, or dynamic, phase lasting a few seconds during ONH, optic nerve head; OPP, ocular perfusion pressure. which the vasculature tries to return blood flow to its original level by adjusting vascular resistance and 2) a steady-state phase when transient blood flow changes have equilibrated to a steady state level. High temporal resolution blood flow 2 mean OPP ¼ MAP IOP measures made during the initial phase following pressure 3 changes are referred to as dynamic autoregulation (dAR), where whereas those made during the steady-state phase are 1 120 MAP ¼ diastolic BP þ ðsystolic BP diastolic BPÞ: referred to as static autoregulation (sAR). To date, most 3 studies assessing autoregulatory capacity in the ONH have The factor 2/3 accounts for the drop in BP between the been limited to sAR. brachial and ophthalmic artery when the subject is seated182 In this section, some important clinical studies that and the fact that the orbital arteries are further downstream. address dAR and sAR in the ONH in healthy conditions are In clinical studies, the brachial arterial pressure has been reviewed. often considered as representative of systemic BP and is used as the basis for calculating the ophthalmic arterial pressure 4.1.1. Dynamic autoregulation studies in the calculation of OPP. The pressure in the brachial artery, The dAR refers to the transient vascular changes preceding however, is not a precise predictor of the pressure in the the equilibrated steady state. Interestingly, dAR was pointed ophthalmic artery. It is not clear how accurately the above out as a more sensitive indicator of cerebral blood flow formula approximates the difference between the ocular autoregulation than sAR since dAR is better correlated to arterial BP and the brachial arterial pressure because of the neuronal activities than sAR.160,191 Moreover, a dynamic blood hydrostatic column effect when an individual is sitting.30 We flow response contains both time and frequency information also cannot assume that the difference between the ocular that can effectively reveal potential autoregulation dysfunc- and brachial arterial pressures is the same in normal and tion. In fact, dAR measurements have become the standard diseased vascular beds. Moreover, blood flow is determined method for assessing cerebral blood flow autoregulation in not only by OPP but also by vascular tone. Regulation of blood cerebral diseases.230 flow may occur through changes in vascular resistance A blood flow time course similar to that described in the (vasoconstriction or vasodilation, see section 5) indepen- brain was observed in the ONH of rabbits220 and nonhuman dently of changes in OPP. Physical exercise may or may not primates119 after an acute increase in IOP. The time course of cause a change both in cardiac output and in the net resis- relative ONH blood flow changes from baseline (IOP ¼ 10 mm tance of the many microvascular pathways in parallel. Hg) to elevated IOP (IOP ¼ 30 mm Hg), and then back to Equating venous pressure to IOP is also sometimes erro- baseline, was tracked for 3 different ranges of BP in mon- neous. For example, if the cerebrospinal fluid pressure is keys.119 In the high-BP group, there was no significant change higher than IOP, the venous pressure must exceed cerebro- in ONH blood flow during the IOP alterations; however, the spinal fluid pressure in the subarachnoid space for the CRV to same IOP alterations caused a significant ONH blood flow remain patent, and thus venous pressure could not be change in the 2 lower BP groups, suggesting that autor- approximated by IOP. In this way, the estimate for OPP in- egulation of the ONH is deficient in the lower BP groups. volves systematic errors; however, even if more reliable for- To characterize dAR in the ONH, time-domain parameters mulas for computing the OPP have been proposed,46 the were extracted from high temporal resolution blood flow previously mentioned relations are consistently used in measurement obtained using an LSFG device in a group of clinical studies. A reliable, direct measure of OPP would of monkeys.120 A rapid OPP decrease induced by a sudden IOP course be desirable, but without this, care is needed when step increase evoked a transient and reproducible dAR in the interpreting blood flow regulation studies. ONH. The duration of blood flow decrease was much shorter 172 survey of ophthalmology 61 (2016) 164e186

than that of the pressure change. In other words, while the IOP disk locations of some individuals appeared unable to regulate was still increasing, the blood flow had already started blood flow under even a minimal IOP challenge. recovering. This observation suggests that autoregulation is Effects of elevated OPP on the ONH blood flow of healthy activated immediately after IOP elevation. volunteers were investigated by increasing MAP through iso- Responses of OPP, ONH blood flow, and vascular resistance metric exercise (squatting).144 An LDF probing laser beam was to increases in BP were investigated by hand gripping.39 BP directed at temporal disk areas of the ONH, at least 200 mm and ONH blood flow parameters were simultaneously and from the disk margin and outside of the physiologic cup. In the continuously measured by LDF. Healthy subjects could be range of OPP between 56 4 and 80 5 mm Hg (30% 8%), subdivided into 2 groups because of marked differences in the there was no significant variation of mean velocity, volume, efficiency and OPP range of autoregulation: in one group, and flux of red blood cells, but vascular resistance increased autoregulation was found to be highly efficient once OPP was by about 50%. These results suggest that the maintenance of increased by approximately 15% above baseline; in another constant blood flow is achieved by an increase in vascular group, autoregulation was found to be less efficient. These resistance. Similar results were obtained in another study.171 findings are in accordance with other studies.24 Regulation of ONH blood flow was investigated during The dAR studies mentioned above are summarized in combined changes in IOP and systemic BP. ONH blood flow Table 3. was measured in monkeys using LSFG during artificial changes in IOP and BP.119 When IOP was increased to 30 mm 4.1.2. Static autoregulation studies Hg and MAP was normal (102 mm Hg), no significant change The sAR refers to steady-state responses of ONH blood flow to was observed in ONH blood flow, indicating that autor- a wide range of OPP values. These responses constitute a egulation was functioning. However, when IOP was increased classic autoregulation curve or pressure-flow relationship (see to 30 mm Hg and MAP was reduced to 56 mm Hg, significant Fig. 6). This plateau of the autoregulation curve indicates the reductions in ONH blood flow were observed, suggesting that range of OPP where autoregulation is functioning. When the autoregulation was unable to maintain blood flow at low OPP. OPP values are outside the range defined by this plateau, These blood flow results, as well as those obtained in other autoregulation fails and blood flow will gradually decrease or studies,234,237 must be interpreted in the context of where and increase passively as OPP changes. how they were measured. Microspheres were used to measure blood flow in the ONH blood flow was investigated in 40 healthy subjects various compartments of the monkey optic nerve following using continuous LDF during a separate increase in IOP and manometric IOP elevation.70 Small changes in IOP which MAP as well as during their combined elevation.24 The laser reduced OPP as low as 29 mm Hg had small effects on prel- beam was directed toward the neurovascular rim. MAP was aminar blood flow. At OPP less than 29 mm Hg, prelaminar increased by isometric exercise consisting of squatting, and flow was proportional to OPP. Laminar blood flow was nearly IOP was raised via suction cup. During both experiments, the unchanged even at high IOP (low OPP). change in ONH blood flow was less pronounced than the OPP was decreased by increasing IOP with a scleral suction change in OPP, indicating autoregulation. cup in normal human subjects.68,166,181 An LDF probing laser The sAR studies mentioned in the previous paragraphs are beam was directed at the ONH tissue at temporal and nasal disk summarized in Table 4. areas, at least 200 mm from the disk margin and outside the In Figure 7, OPP-ONH blood flow relationships reported in a physiologic cup. Autoregulation was active for OPP values as few studies24,70,144,166,181 have been gathered. Each data set low as 15e20 mm Hg (IOP ¼ 40e45 mm Hg),181 or until IOP has been normalized to its corresponding baseline OPP and reached 45 mm Hg.166 Apparently, the ONH vasculature was ONH blood flow estimate. Care is needed to interpret this not fully dilated at this point, because diffuse luminance flicker figure correctly. First, different species and different tech- increased ONH blood flow even more.68 Regional variability in niques for estimating ONH blood flow were involved in these 6 the degree of autoregulation has been reported.166 Some optic studies. Second, in one study,70 monkeys were supine,

Table 3 e Summary of dynamic autoregulation studies in healthy subjects Study Methods Technique Main findings

Liang et al119 Combined changes in IOP and MAP in monkeys LSFG The duration of the changes in ONH blood flow from baseline to a peak and to a steady state was significantly delayed in subjects with medium or low, but not high, BP Liang et al120 Sudden IOP step increase in monkeys LSFG While the IOP was still increasing, the blood flow had already started recovering Chiquet et al39 MAP elevation in healthy humans LDF Marked differences in the efficiency and OPP range of autoregulation were found among subjects, in accordance with another study24

BP, blood pressure; IOP, intraocular pressure; LDF, Laser Doppler flowmetry; LSFG, laser speckle flowgraphy; MAP, mean arterial pressure; ONH, optic nerve head; OPP, ocular perfusion pressure. survey of ophthalmology 61 (2016) 164e186 173

Table 4 e Summary of static autoregulation studies in healthy subjects Study Methods Technique Main findings

Geijer et al70 IOP elevation in healthy monkeys Microspheres The laminar and retrolaminar portions of the ONH showed higher autoregulation capabilities than the prelaminar portion Riva et al181 IOP elevation in healthy humans LDF ONH blood flow was relatively constant to an OPP as low as 15e20 mm Hg Pillunat et al166 IOP elevation in healthy humans LDF ONH blood flow was regulated to an IOP up to 45 mm Hg Movaffaghy et al144 MAP was elevated in healthy humans LDF In the range of OPP between 56 4 and 80 5mm Hg, there was no significant variation of mean velocity, volume, and flux of red blood cells, but vascular resistance increased by about 50% Liang et al119 Combined changes in IOP and MAP in monkeys LSFG When IOP was increased to 30 mm Hg and MAP was normal (102 mm Hg), ONH blood flow did not change significantly. Instead, blood flow decreased when IOP was increased to 30 mm Hg and MAP was reduced to 56 mm Hg Boltz et al24 Combined and separate IOP and MAP elevation in LSFG ONH blood flow increased with OPP for OPP healthy humans values of 66% above baseline. ONH blood flow decreased for OPP values of 40% below baseline Wang et al237 IOP elevation in monkeys LSFG ONH blood flow is effectively regulated for OPPs of approximately 41 mm Hg and above

IOP, intraocular pressure; LDF, Laser Doppler flowmetry; LSFG, laser speckle flowgraphy; MAP, mean arterial pressure; ONH, optic nerve head; OPP, ocular perfusion pressure.

whereas, in the others, human volunteers were standing. autoregulation components. Dynamic and static aspects are Third, some data were obtained by increasing IOP, whereas 2 related, but different, components of the autoregulation other data were obtained others by increasing MAP. Last, process. Importantly, measuring dAR could succeed in displaying all the data in a single graph presumes the use of a revealing potential autoregulation dysfunction in situations common baseline. Despite these limitations, Figure 7 shows where the conventional “two-point” method measured dur- that the static OPP-blood flow curve in the ONH is consistent ing the sAR phase would fail. with the curve schematized in Figure 6. On reviewing a number of clinical studies addressing 4.2. Clinical studies of static blood flow autoregulation dynamic and static blood flow autoregulation in the ONH in the ONH in pathologic conditions in healthy conditions, note the different sensitivity of Under pathologic conditions, ONH blood flow regulation may be 50 disrupted.5 Impaired autoregulation in the ONH associated with altered blood flow has been observed in experimental dia- betes204 and hypercholesterolemia.205 In glaucoma, and espe- cially in NTG, impaired autoregulation in the ONH has been 0 speculated to be an important risk factor for the progression of the disease.43,60,75,91,156 Glaucoma patients have been suggested to exhibit abnormal autoregulation especially in response to

Geijer (1979), prelaminar region acute changes in OPP, as reviewed in the following paragraphs. -50 Geijer (1979), laminar region The autoregulatory control of retrobulbar blood flow in ONH blood flow Pillunat (1997), ONH Riva (1997), ONH response to postural challenge was investigated in NTG Movaffaghy (1998), ONH patients in comparison with primary open-angle glaucoma Boltz (2013), ONH 66

(mean % change from baseline) patients and healthy volunteers. PSV, EDV, and RI in the short -100 PCAs were recorded after a change from sitting upright to a -100 -50 0 50 100 OPP (mean % change from baseline) supine body position using CDI. Ten minutes after postural change, blood flow velocities in the short PCAs remained un- Fig. 7 e Pressure-flow relationships for ONH blood flow in 6 changed in controls, whereas a significant increase of PSV and experimental studies in healthy monkeys (Geijer et al.,70 EDV occurred in both glaucoma groups. The RI in the short circles and asterisks), and in healthy humans (Pillunat PCAs was significantly lower in the NTG group when compared et al.,166 squares), (Riva et al.,181 downward-pointing to healthy and primary open-angle glaucoma individuals. The triangles), (Movaffaghy et al.,144 upward-pointing authors suggested that the unaltered flow velocities in the triangles), (Boltz et al.,24 crosses). ONH, optic nerve head; short PCAs of healthy controls might indicate tight autor- OPP, ocular perfusion pressure. egulatory control. In contrast, the accelerated flow in the short 174 survey of ophthalmology 61 (2016) 164e186

PCAs exhibited by NTG and primary open-angle glaucoma range of major eye diseases, it is incredibly important to patients might indicate an insufficient compensatory response quantify its effects on ONH blood flow regulation. to postural change. However, it is not clear to what extent these In a study of the ONH of control and glaucomatous monkey interpretations are consistent with the experimental data. It is eyes,237 static blood flow autoregulation was characterized, unknown how the supine position affects intraocular venous and impaired autoregulation was tested as a potential mech- pressure and, in turn, OPP.46 It is accepted that, in the supine anism involved in the reduction of ONH blood flow observed position, there is a different interaction between orbital venous in previous studies.47,234 Autoregulation curves were created pressure and intraocular venous pressure than in the upright based on a series of relative ONH blood flow changes, each position. This interaction is affected by the tissue pressure, measured with an LSFG device in response to an acute OPP which experiences huge variations due to differences in body decrease induced by instantaneous IOP elevation monitored mass and the hydrostatic water column effect on venous across stages of glaucoma. Within the ONH of the control pressure.30 Consequently, it is hard to establish how the supine eyes, blood flow was effectively regulated within the OPP position impacts OPP and, hence, blood flow. range from 41 mm Hg to 115 mm Hg. When OPP was below Chronic unilateral elevation of IOP was induced in mon- 41 mm Hg, blood flow declined linearly with OPP. Autor- keys by laser treatment to the trabecular meshwork, leading egulation curves of control and glaucomatous eyes were not to glaucomatous damage.234 The authors found compromised significantly different from each other. Thus, the authors blood flow in the anterior ONH (including SNFL, prelaminar argued that sAR is not a predominant factor accounting for the tissue, and lamina cribrosa) of glaucomatous eyes measured reduced ONH blood flow observed in the monkey model.47,234 by both LSFG and the microsphere method. ONH blood flow All of these clinical studies addressing autoregulation in was measured for IOP ¼ 40 mm Hg in both control and glau- pathologic conditions are summarized in Table 5. comatous eyes. In a parallel study, longitudinal changes in basal blood flow of the ONH were investigated in the same monkey model.47 There are a couple of possible explanations 5. Mechanisms of blood flow regulation for the reduced flow observed in these 2 studies during the later stages of glaucoma. If there is a reduced amount of Several important response mechanisms combine to cause neural tissue requiring nutrition, the flow would be regulated changes in vascular tone that lead to blood flow regulation to a to avoid supplying more blood than what is necessary for the particular tissue. A complete overview of the biochemistry of remaining tissue. ONH blood flow could also be reduced if all the mediators and modulators involved is beyond the autoregulation is impaired. scope of this review. We will focus on studies relevant for the There was no evidence of altered autoregulation in the control of blood flow in the ONH. ONH in glaucoma in some studies. Regulation of ONH blood flow in response to an increase of OPP was investigated 5.1. Mechanical influences through a challenge in systemic BP induced by isometric exercise in NTG patients and in age-matched healthy volun- An important class of autoregulation is seemingly more teers.171 ONH blood flow parameters, as determined by LDF, dependent on mechanical influences. In the myogenic did not indicate abnormal blood flow regulation in either response, vascular smooth muscle cells constrict as intra- healthy subjects or NTG patients; however, NTG patients vascular pressure, and consequently the circumferential wall showed a greater percent increase in vascular resistance tension, is elevated. Indeed, circumferential wall tension compared to the normal subjects for a similar percent depends on both the vessel radius R and transmural pressure increase in OPP in both groups during squatting. DP, defined as the difference between the intravascular and To our knowledge, neither dynamic nor static mechanisms extravascular pressures. According to Laplace’s law, as DP of ONH blood flow regulation have been quantified in patients rises, the vessel wall stretches, leading to increased wall with arterial hypertension. Arterial hypertension can poten- tension T: tially interfere with ONH blood flow in many ways.94 For T ¼ DP R: example, atherosclerosis resulting from prolonged arterial hypertension can cause inadequate changes in vessel size in In the myogenic response, arterioles respond to the response to fluctuations in OPP.88 Chronic arterial hyperten- increased wall tension by constricting to reduce the vessel sion can also cause the range of autoregulation to shift to radius R and restore the wall tension to a normal level. The higher levels to adapt to high BP.97 Such an adjustment makes myogenic response has been observed in the ONH. More the patient more tolerant to high BP but less tolerant to low BP. effective blood flow autoregulation was observed in the ONH In this situation, a sudden reduction of BP in hypertensive of healthy humans during an increase in MAP than IOP.24 Such subjects can cause anterior ischemic optic neuropathy.96,98 behavior could be compatible with a myogenic mechanism of Systemic hypertension is associated with pathologic autoregulation. changes in retinal vasculature.123 Endothelial function of the The contractility of smooth muscle cells involved in the retinal vasculature is impaired in early essential hyperten- myogenic response is regulated by ECs. In addition to chemical sion,50 and hypertensive is associated with agents,ECsrespondtomechanicalstimuli,suchasfluidshear endothelial dysfunction111,112 (see section 5 for a discussion stress and stretch.38 Externally applied forces with a clear di- about the role of endothelial cells (ECs) in regulating blood rection, such as shear stress from pulsatile flow or uniaxial flow). Hence, since systemic hypertension is linked to a wide stretch, induce the release of vasoactive substances which then survey of ophthalmology 61 (2016) 164e186 175

Table 5 e Autoregulation studies in pathologic conditions Study Methods Technique Main findings

Pournaras et al171 MAP elevation in NTG patients LDF NTG patients did not show abnormal ONH blood flow regulation but only a greater percent increase in vascular resistance compared to healthy subjects Galambos et al66 Investigation of autoregulatory control of CDI NTG and POAG patients exhibited an retrobulbar blood flow in response to postural insufficient compensatory response to changes in NTG and POAG patients postural change, leading to accelerated flow in the short PCAs Shibata et al205 IOP elevation in hypercholesterolemic rabbits LSFG Hypercholesterolemia induced impairment in the autoregulation of ONH blood flow and deterioration in visual function and histology Shibata et al204 IOP elevation in rabbits with induced diabetes LSFG Autoregulation was disrupted both in the and induced gap junction uncoupling animals who were induced diabetes and in those who received gap junction uncouplers Wang et al234 Chronic unilateral IOP elevation induced in Microspheres and LSFG Blood flow was compromised in the SNFL, monkeys by laser treatment to the trabecular prelaminar, and laminar tissues, possibly due meshwork to impaired autoregulation Cull et al47 Chronic unilateral IOP elevation induced in LSFG A 2-phase pattern of ONH blood flow monkeys by laser treatment to the trabecular alteration was observed in treated animals. meshwork ONH blood flow increased during the earliest stage (while retinal nerve fiber layer thickness was within 10% of baseline) followed by a linear decline strongly correlated with loss of RGCs Wang et al237 Chronic unilateral IOP elevation induced in LSFG Chronic IOP elevation caused no remarkable monkeys by laser treatment to the trabecular change to the static autoregulation of meshwork glaucomatous eyes

CDI, Color Doppler imaging; IOP, intraocular pressure; LDF, Laser Doppler flowmetry; LSFG, laser speckle flowgraphy; MAP, mean arterial pressure; NTG, normal tension glaucoma; ONH, optic nerve head; PCA, posterior ciliary artery; POAG, primary open-angle glaucoma; RGC, retinal ganglion cell; SNFL, superficial nerve fiber layer.

determine ECs remodeling. This remodeling phenomenon in- the presence of oxygen and superoxide and longer in the cludes changes in the orientation of ECs cytoskeletal fibers,38 the absence or lower concentrations of oxygen.20,140 Because NO morphology of cell surface49,152 and/or cell stiffness99,137,194,195 to cannot be stored in vivo, regulation of NO production is minimize the alterations in intracellular stress. Whereas ECs controlled at the level of the biosynthesis.198 Hence, NO syn- remodeling results from molecular signaling due to mechano- thesis largely depends on the amount of nitric oxide synthase transduction, this adaptive behavior, in turn, modulates the (NOS). Inhibiting NOS in cats resulted in a decrease in ONH molecular signaling. Thus, a close coupling exists between blood flow, both at baseline and with light flicker stimulation mechanics and vascular endothelium. In addition, experiments of neuronal activity.27 Administration of a nonspecific NOS have shown that myogenic tone of the PCAs is regulated by inhibitor, NG-monomethyl-L-arginine, appeared to reduce prostaglandin formation and the release of endothelium- baseline ONH blood flow in humans.128,199 NO also exerts derived relaxing factor by arteries.150 To our knowledge, regu- antioxidative and neuroprotective effects by interacting with lation of myogenic tone and sheer stress responses in the ONH reactive free radicals such as superoxide anions, hydroxyl, vasculature has not been studied. Further research is needed to and peroxyl lipid radicals.40 investigate the contribution of mechanical influences on ONH ET-1, a potent vasoconstrictor that is released from 8,156,179,180,200,232 blood flow regulation. ECs, acts on 3 types of ET receptors: ETA,ETB1,

and ETB2. More precisely, ETA receptors are present in vascular 5.1.1. Nitric oxide and endothelin-1 smooth muscle and mediate vasoconstriction.156,179,200 The

Two important vasoactive substances released by the ECs stimulation of ETB1 receptors, which are located on ECs, results after local stimulation by, for example, shear stress from in vasodilation via clearance of ET-1 and the release of NO and 200 167 pulsatile flow or uniaxial stretch, are nitric oxide (NO) and prostacyclin. ETB2 receptors mediate vasoconstriction. endothelin-1 (ET-1). A constant balance between the opposing Despite the opposing roles of the B1 and B2 receptors, a net functions of NO and ET-1 is necessary for proper regulation of vasodilation is shown to result from the stimulation of the ETB 156,157,232 238 the vascular system. NO is a potent vasodilator subtype receptors. Pharmacologically blocking ETA and ETB secreted by smooth muscle cells that causes the dilation of receptors resulted in a significant increase in blood flow to the arterioles via activation of smooth muscle cells and the dila- retina, choroid, and ONH in patients with glaucoma as well as tion of capillaries via pericytes.93 Owing to the radical nature healthy subjects, suggesting a potential role for endothelial of NO, it has a short half-life, which is made even shorter in blockers to be used in the treatment of glaucoma.180 176 survey of ophthalmology 61 (2016) 164e186

5.2. Metabolic response in the blood velocity of the CRA with hypercapnia, but a decrease in blood velocity in the OA.210 Although the prepon- Metabolic regulation results in the adjustment of blood flow derance of data suggests that blood vessels dilate or constrict during inadequate or excessive nutrient supply, because of in response to CO2 levels, it is possible that hypercapnia- either a change in the metabolic activity of the tissue or a induced vasodilation is mainly because NO is brought into change in flow by virtue of cardiovascular factors such as play by lowered levels of oxygen, which would increase NO atherosclerosis or systemic vasospasm. No matter the cause, half-life. Future research further defining the mechanisms of when tissues are not receiving the appropriate amount of this is needed. blood flow, an adjustment is made by control mechanisms. Overall, the vascular response to tissue demand is influenced 5.2.3. Nitric oxide by the metabolic conditions in the tissue such as levels of NO was shown to mediate vasodilation in response to marked oxygen, carbon dioxide levels (mediated by pH), NO, ET-1, and hypoxia in the forearm resistance vessels in healthy other chemical signals.157 For example, when the eye is humans22 and in response to increased myocardial oxygen exposed to a flickering light, the number of action potentials demand.4 As mentioned in the previous sections, the role of and the consequent need for (re)polarization of the RGC axon NO in hypercapnia-related vasodilation is not clearly membranes in the ONH is increased with each “on” and “off” understood. light switch. In response to this increased metabolic demand, e blood flow in the ONH appears to increase.184 186 Importantly, 5.3. Neurovascular coupling metabolism in the various parts of RGC axons may be controlled independently. In the prelaminar and lamina cri- Experiments measuring vascular responses to light stimula- brosa regions, RGC axons are unmyelinated and acquire a tion have suggested that ONH blood flow and neuronal myelin sheath only after passing the posterior boundary of the activity of RGCs are coupled.68,185,186 Traditionally, in the ONH lamina cribrosa.11 The unmyelinated portions of the axons are and other regions like the retina and the brain, vascular in need of energy to repolarize the axons along the entire responses to changes in neural activity were assumed to be surface, whereas, in the myelinated portion, only the axons at controlled exclusively by a local metabolic feedback system in the nodes of Ranvier need to be repolarized. Thus, mito- which neural activity leads to energy demand and thus chondria are numerous and metabolic activity is much higher vasodilation. This idea has been challenged, however, in the unmyelinated nerves, but less in the myelinated following the discovery that feedforward mechanisms portion.19 According to the metabolic hypothesis of blood flow mediate the vascular energy supply by neuronal activity.16 regulation, tissue perfusion and tissue metabolism are tightly According to these mechanisms, active neurons either send coupled in such a way that any reduction in arterial inflow a signal directly to blood vessels or activate astrocytes to causes an increase of vasodilator metabolites in the tissue.157 release vasoactive agents onto the vessels to increase local blood flow, a process known as “functional hyperemia.” Both 5.2.1. Oxygen cases involve neurotransmitter signaling, particularly via All living tissues require a steady supply of oxygen to meet, glutamate. In the brain, blocking glutamate receptors reduces but not exceed, their nutritional needs. Oxygen tension in a functional hyperemia but does not affect the energy use tissue provides an important indication of its current meta- associated with neuronal activity, providing evidence that bolic status. In hypoxic conditions, ONH blood flow is regu- metabolic feedback and neurovascular feedforward mecha- lated to establish a healthy partial pressure of oxygen in the nisms can be distinguished and both contribute to blood flow tissue. Under moderate hypoxic conditions, the intravascular regulation in response to neuronal activity.151,216 The inter- tissue’s partial pressure of oxygen at 200 mm depth within the play between metabolic feedback and neuronal feedforward ONH was observed to be constant, reflecting, probably, mechanisms should then be taken into account whenever autoregulation of blood flow.25 Decreased partial pressure of considering clinical studies addressing ONH vascular oxygen was found in the ONH when IOP was increased (OPP responses to light stimulation. was decreased) above 45 mm Hg.23 Both oxygen metabolism in Astrocytes exhibit extremely complex and dynamic the mitochondria of optic nerve cells and tissue oxygen ten- behaviors. One of their functions is to control the extracellular sion were observed to decrease when hypoxia was induced.149 environment for neurons; for example, by clearing the space of In contrast, under hyperoxic conditions, capillary blood flow potassium after action potentials pass by and taking up excess in the ONH decreases via vasoconstriction.115 glutamate that leaks from synapses.100,101 Astrocytes also play a fundamental role in regulating blood flow to the brain.16 As- 5.2.2. Carbon dioxide trocytes are the predominant glial cell type in the nonmyelin- 19 CO2 is a vasodilator in all vascular beds at the posterior pole of ated ONH in most mammalian species. Structural similarities the eye.200 The mechanisms driving this vasodilation response suggest that retinal and ONH astrocytes may behave similarly are not clearly understood. In human197 and animal studies,196 to cerebral astrocytes.10 Retinal and ONH astrocytes surround hypercapnia-induced vasodilation was shown to depend on blood vessels,8,233 and this close relationship further supports a NO, whereas, in other animal studies, the vasodilation was potential contribution of astrocytes to blood flow regulation in shown to be independent of NO.52,72 Hypercapnia triggers an the retina and ONH. increase in flow to the short PCAs.189 Patients with untreated There is evidence of neurovascular coupling in the retina primary open-angle glaucoma have shown a normal increase and ONH occurring via glial signaling mediated by vasoactive survey of ophthalmology 61 (2016) 164e186 177

þ agents released onto the vessels (not through nerve fibers and endfeet to open, releasing K onto vessels.59 This potential þ synapsis with glutamate receptors on contractile cells). The role of K in neurovascular coupling has not been tested either retinal vascular branches are devoid of such innervation once in the retina or in the ONH yet. they emerge onto the retinal surface, although the CRA within the intraorbital optic nerve is innervated.243 Yet, retinal ves- 5.3.2. Nitric oxide sels do retain receptors for various neurotransmitters on their NO is believed to mediate neurovascular coupling in the cer- þ surface.58,104 To our knowledge, the innervation of the arte- ebellum.242 Changes in neuronal activity induce Ca2 riolar branches of the ONH has not been studied yet. Future increases within neurons and activate neuronal NOS, leading research further elucidating the mechanisms of neuro- to the production of NO. NO is then able to permeate mem- þ vascular coupling in the retina and the ONH is essential. branes and diffuse to blood vessels, where it opens K chan- Axon-glial signaling pathways have been observed in the nels and relaxes vascular smooth muscle cells, leading to optic nerve.41,64,107 Following IOP elevation, autoregulation of vasodilation and increased blood flow.34 NO has also been ONH blood flow was not maintained in rabbit eyes where glial suggested to mediate neurovascular coupling in the ONH. cells were selectively impaired by a gliotoxic compound, L-2- Increased NO levels were found in the ONH of humans during aminoadipic acid, indicating a possible involvement of astro- changes in neuronal activity due to light flicker stimuli.26 In cytes in ONH blood flow autoregulation.206 The cell processes addition, in the retina of humans54 and the retina and the of astrocytes are connected to each other via gap junctions190 ONH of cats,114 systemic administration of nonselective NOS forming a functional syncytium that allows astrocytes to inhibitors reduces flicker-evoked increases in blood flow. communicate and maintain control of the ionic and metabolic These results would appear to support NO as an important homeostasis in the ONH. Decreased gap junction communi- mediator of neurovascular coupling in the retina and the ONH. cation between ONH astrocytes was reported under condi- However, evidence has been found that NO could be pre- tions of elevated IOP.133 Closure of gap junctions interrupts dominantly a modulator of neurovascular coupling, but not a the continuity of astrocyte intercellular communication, mediator. In the whole-mount rat retina,138 at NO levels less causing loss of cell-cell contact and homeostatic regulation. than 100 nmol/L, flickering lighteinduced vasodilation, but Uncoupling gap junctions resulted in impaired ONH blood not vasoconstriction. At increased NO levels, smaller vasodi- flow autoregulation in healthy rabbits.204 Such conditions lation was observed, while vasoconstriction became more may lead to RGC axonal loss and optic disk remodeling char- common. At 10 mmol/L NO, a flickering light evoked large acteristic of glaucomatous optic neuropathy.100 vasoconstriction, mediated by 20-HETE. It is still not clear how In studies on brain slice and isolated retina, evidence is NO inhibits flicked-induced vasodilation. The modulatory found that astrocytes contribute to blood flow regulation by mechanism of NO has been suggested to be due to its producing and releasing metabolites of arachidonic acid. After inhibitions of P450 epoxygenase, which synthesizes the glutamate is released at synapses following neuronal activity, vasodilator EETs. With less EETs being produced, vasodilation some of the released neurotransmitter escapes the synaptic will be smaller and vasoconstriction mediated by 20-HETE will cleft and activates metabotropic glutamate receptors on prevail.148 þ astrocytes, increasing Ca2 in astrocytes.170 The increased 2þ Ca activates phospholipase A2 and results in arachidonic 5.3.3. Adenosine acid production. The build-up of arachidonic acid leads, in Adenosine, which is produced when adenosine triphosphate turn, to the production of its metabolites, including prosta- is hydrolyzed, is a vasodilator that contributes to functional glandin E2 and epoxyeicosatrienoic acids (EETs), which dilate hyperemia in the . Blocking adenosine receptors nearby arterioles.74,138,163,164,219,245 In brain slices and in the has been shown to reduce the increase in blood flow evoked þ isolated retina, increased Ca2 can also cause vessels to by neuronal activity.113 Given its function in the brain, aden- constrict.44,138,145 This is made possible by the conversion of osine is also thought to be involved in the control of ocular arachidonic acid to 20-hydroxy-eicosatetraenoic acid (20- blood flow. Intravenous administration of adenosine was HETE).138 When the synthetic enzymes for EETs and prosta- shown to cause increased choroidal and ONH blood flow in 168 glandin E2 are inhibited, glial-evoked vasodilation is reduced humans and was established as an important participant in by 88%.138,139 Under physiologic conditions, the effect of the mediating retinal blood flow autoregulation during hypoxia 71 vasodilators prostaglandin E2 and EETs surpasses the effect of and hypotension in piglets. 20-HETE, and vasodilation occurs. However, under non- physiologic conditions such as hyperoxia or elevated NO 5.3.4. Lactate levels, 20-HETEemediated vasoconstriction dominates.148 Lactate, which is produced by anaerobic glycolysis and released by both glial cells and neurons, is thought to 5.3.1. Potassium contribute to neurovascular coupling as a metabolic negative- Glial cells might also cause vasodilation (and hence increased feedback mediator. Blood flow in the human retina is sensitive blood flow) by releasing potassium ions. Modest increases in to changes in blood lactate levels,69 but lactate responses in þ extracellular K concentration cause the hyperpolarization of the ONH have not been established yet. þ smooth muscle cells, a reduced influx of Ca2 through voltage- gated channels, and vasodilation. When glutamate released 5.3.5. Oxygen from neurons activates astrocyte metabotropic glutamate Oxygen modulates neurovascular coupling in brain tissue by þ receptors, the resultant rise in astrocyte Ca2 causes large- altering the synthesis of glial and neuronal messengers and by 2þ þ conductance Ca -activated K channels in astrocyte altering the levels of lactate and adenosine. O2 is needed for 178 survey of ophthalmology 61 (2016) 164e186

the synthesis of NO and the vasoactive messengers derived feedforward coupling mechanisms, but this still needs to be from arachidonic acid. At in vivo levels of O2, the synthesis of further investigated.

NO and 20-HETE is expected to be limited by the amount of O2 In contrast to retinal capillaries, choroidal capillaries are available.16 More specifically, 20-HETE formation is sup- fenestrated and have sparse pericytes.161 The retinal tissue, pressed in the presence of low O2 concentrations, leading to a part of the central nervous system, has a preserved blood- reduced vasoconstriction when arachidonic acid is generated brain barrier due to the tight junctions between the retinal 161,179 in astrocytes. A lower O2 also results in a reduced amount of pigment epithelial cells. Circulating vasoconstrictive NO being present to inhibit the formation of vasodilatory EETs hormones, such as angiotensin, epinephrine, and so forth, 2þ 157,161,200 in tissues. In addition, when O2 concentrations decrease, the raise intracellular Ca when they occupy receptors. lack of energy for adenosine triphosphate synthesis causes an In vascular smooth muscle, as well as in pericytes, an þ increase in the level of extracellular adenosine, which binds to increase of intracellular Ca2 induces contraction of the cell. adenosine A2A receptors on vascular smooth muscle cells to Consequently, ECs produce more NO, which relax vascular depress vessel constriction.16 Moreover, low oxygen concen- smooth muscle and pericytes. In this way, the circulating trations induce a decrease in the rate of oxidative phosphor- hormones leak through fenestrated vessels, reach the ylation relative to the rate of glycolysis, resulting in lactate muscular coat of vessels in the region, causing vasoconstric- production. Monocarboxylate transporters release the lactate tion in the skin and many visceral organs. In contrast, central into the extracellular space, where it reduces the reuptake of nervous vessels, where hormones can only affect ECs and are prostaglandin E2 by the prostaglandin transporter, promoting prevented from reaching the contractile coat of the vessels to vasodilation.74 Importantly, it is possible that oxygen modu- the blood-tissue barrier, exhibit vasodilation.157,161,200 It is lates neurovascular coupling by affecting NO half-life. Future unclear whether the branches of the PCAs that feed the research to further elucidate the mechanisms involved is intrascleral portion of the optic nerve are innervated and/or needed. fenestrated. Such knowledge is fundamental to understand how the intrascleral papillary tissue reacts to various insults, 5.4. The anatomic agents of blood flow regulation in the including abnormally high IOP. ONH

Understanding which vessels in the ONH exhibit blood flow regulation is still an open problem. In general, the structures 6. Pathologic effects of impaired blood flow that control blood flow to a particular tissue are the resistance regulation in the ONH arterioles (arterial branches smaller than 40 mm), which change diameter actively to achieve autoregulation.172 Capil- Impaired blood flow regulation has been suggested to render laries have smaller diameters than arterioles. Because resis- the ONH more susceptible to damage by compromising blood tance to flow is inversely proportional to the fourth power of supply (leading to ischemia) and, consequently, oxygen the vessel radius (Poiseuille’s law),65 capillaries exhibit high delivery (leading to hypoxia) in potentially dangerous situa- resistance individually; However, they are numerous and in tions of reduced BP, increased IOP, and/or increased local parallel array outweigh the reduction in vessel size, having a metabolic demands.12,117,141 However, the mechanisms large cross-sectional area that does not contribute much to through which the damage occurs are still not completely the net resistance between arteries and veins.246 In addition to understood.175,231 Evidence suggests that ischemia and the arterial role in autoregulation, flow to a local capillary related hypoxia might influence the astrocytes in the ONH101 network is controlled by a precapillary sphincter (where it and/or the mitochondria of RGC axons,158 and that the ulti- exists), a band of smooth muscle located where capillaries mate death of RGCs can occur via apoptosis158,175 and/or originate from small arterioles.6 Finally, blood flow could also autophagy.124 be controlled by changes in the resistance of the capillaries Astrocytes in the ONH are considered to be quiescent in themselves because of the contractile properties of pericytes.6 normal conditions,18 but they appear to become reactive in The retinal and optic nerve capillaries lack precapillary response to pathologic conditions and contribute to changes sphincters but have abundant pericytes.62,63,161 It has been in the ONH environment, which do not favor the survival or shown that pericytes respond to carbon dioxide concentra- growth of the RGC axons.125 During ischemia, hydrogen tions (mediated by pH), oxygen levels (affected by changes in peroxide was observed to be abundantly produced and to NO concentrations), and adenosine levels.161 Vascular smooth promote pathologic excitatory amino acid release and muscle cells respond in a very similar manner.36,82,136 Peri- swelling of reactive astrocytes.102 In glaucoma, ONH astro- cytes on isolated retinal capillaries were found to constrict or cytes were observed to have lower basal levels of antioxidant þ dilate in response to neurotransmitters, following Ca2 alter- glutathione, which plays an important role in protecting the ations.173 In situ in brain slices, pericytes constricted in mitochondrial electron transport chain from damage by response to noradrenaline and dilated in response to gluta- oxidative stress in both astrocytes and neurons.132 Evidence mate,165 and in the isolated retina, blocking ionotropic suggests that reactive astrocytes also produce NO in supra- receptors of the neurotransmitter g-aminobutyric acid was normal quantities, and this excessive production seems to shown to constrict capillaries, suggesting that endogenous have neurotoxic effects on the RGC axons.126,143,147,244 transmitter release could regulate capillary diameter.83,165 Ischemia, hypoxia, and perfusion instability can also dam- Thus, capillaries are equipped to participate in the control of age the astrocyte-astrocyte gap junctions, and, as a conse- ONH blood flow through metabolic feedback and neuronal quence, compromise the homeostasis of the RGC axons101 and survey of ophthalmology 61 (2016) 164e186 179

reduce the autoregulation capabilities of the ONH vasculature, from the superficial layers. In addition to the difficulties thereby rendering the ONH even more susceptible to damage. related to structural and functional imaging of the ONH, Impaired autoregulation might also cause perfusion there are scientific challenges in designing experimental and instability in the ONH,141 which, in turn, might compromise clinical studies capable of disentangling the complex inter- the normal function of mitochondria in the RGCs and their play between chemical, mechanical, and hemodynamic axons.158 Mitochondria play a critical role in the maintenance factors that contribute to the pathogenesis of optic neu- of cellular homeostasis and they are abundantly present in ropathies. Given these challenges, mathematical modeling the RGCs, both inside and outside the eye .158 Mito- provides a unique tool that can play a significant role in chondria are involved in numerous metabolic functions, advancing the understanding of ONH in health including the metabolism of oxidative energy, the regulation and disease. A mathematical model can serve as a virtual of intracellular calcium levels and pH, and the production of laboratory where the influence of each factor acting on the reactive oxygen species, which promote and regulate system can be singled out and investigated, from a theo- apoptosis.35,201 retical viewpoint, in isolation or in relation with other fac- Apoptosis (“self-killing”) and autophagy (“self-eating”) tors. In the following, we review the main contributions to are 2 different mechanisms that regulate cell life. The rela- the modeling of the biomechanics, perfusion, and autor- tionship between apoptosis and autophagy is complex in the egulation in the ONH and indicate some interesting future sense that, under certain circumstances, autophagy consti- directions of research. tutes a stress adaptation that avoids cell death (and sup- presses apoptosis), whereas in other cellular settings, it 7.1. Biomechanics of the ONH constitutes an alternative cell-death pathway.131 Ischemia and hypoxia increase the reactive oxygen species produc- Alterations in the ONH biomechanical response to changes in tion rate from mitochondria35,42,159 and favor the IOP have been identified as a major factor in the pathogenesis mitochondria-mediated cellular apoptosis; however, of glaucoma.28,174 Particular attention has been devoted to reactive oxygen species accumulation stimulates cellular the mathematical description of the mechanical stresses and authophagy under conditions of nutrient deficiency,192 strains arising within the lamina cribrosa, which is thought hypoxia, ischemia, and perfusion instability86 and other to be a primary site of axonal injury in glaucoma.176 A linear pathologic conditions.57 model of elastic mechanics theory on the bending of thin circular plate was developed for the lamina cribrosa.53 Such an idealization allowed quantitative estimates to be obtained 7. Mathematical modeling of ONH blood flow of the extent to which the degree of fixity offered by the regulation connection with the sclera, the pretension caused by scleral expansion, and the ratio between flexural and in-plane In the previous sections, we have reviewed experimental and stiffness influence the mechanical response of the lamina clinical evidence of blood flow autoregulation in the ONH, the cribrosa to IOP. An idealized, analytical microstructural mechanisms contributing to autoregulation and the patho- model of the ONH load bearing tissues based on an octagonal logic consequences of vascular dysregulation. Despite sig- cellular solid description of the porosity within the lamina nificant recent advances in the understanding of ONH blood identified the material and geometrical properties of the flow autoregulation, important questions remain unan- sclera as major determinants in the strain distributions swered. What are the relative contributions of various within the lamina.193 The analysis also showed that much mechanisms, including responses to mechanical, metabolic, larger strains are developed perpendicular to the major axis and neurovascular stimuli, in achieving ONH blood flow of an elliptical canal rather than in a circular canal. Eye- autoregulation? Do these relative contributions change in specific finite element models based on experimentally health and disease? Is the vascular energy supply by neural reconstructed geometries have been used recently.188,207,208 activity in the ONH mediated by feedforward mechanisms, as These models are used to study in depth influences of ge- it happens in the brain? Are capillaries involved in ONH blood ometry and material properties of the ONH to changes in IOP flow autoregulation? To which extent do impairments in dAR and to investigate growth and remodeling mechanisms in and sAR mechanisms contribute to glaucomatous damage? glaucoma, including adaptation of tissue anisotropy, tissue Can dynamic and static mechanisms be independently thickening/thinning, tissue elongation/shortening, and tissue impaired? migration. Macro- and micro-scale strains are proposed as The quest for answers to these questions is hindered by potential control mechanisms governing mechanical home- e limitations in both the technological and scientific tools ostasis.76 80 Further development of these sophisticated currently available to the community. Major limitations in finite element models may benefit from the recent advances the current technologies for ONH measurements include the in OCT imaging aimed at providing a more accurate charac- fact that LDF, LSFG, and OCT angiography only provide ONH terization of the architectural microstructure within the blood flow estimates in arbitrary units and intereye lamina cribrosa.209 comparison is problematic. LDF appears to be mainly influ- enced by blood flow changes in the more superficial layers of 7.2. Perfusion of the ONH the ONH than deeper ones, but to what extent remains un- certain. OCT angiography provides measurements for the Perfusion of the ONH results from the complex interplay ONH deeper layers which may contain spurious projections between BP, which provides the driving force for the blood 180 survey of ophthalmology 61 (2016) 164e186 through the vasculature, “mechanical stress,” which acts as 7.4. Conclusions and future directions external forces on the vessels, and “vascular regulation,” which mediates vessel dilation/constriction to compensate The mathematical modeling of the interplay between for changes in the system. To the best of our knowledge, only biomechanics, perfusion and autoregulation in the ONH is one model combines biomechanics and hemodynamics in still at its early stages, but is quickly attracting attention. the lamina cribrosa.33 In this model, the lamina cribrosa is Elucidating the complex interactions of ONH perfusion and treated as a 2 dimensional poroelastic material, where blood tissue structure in health and disease using current imaging vessels are viewed as pores in a solid matrix. The presence of methodologies is difficult, and mathematical modeling pro- blood vessels in the tissue defines a vascular porosity N vides an approach to solving these limitations. One of the which changes with the local state of stress and strain. The main difficulties lies in the fact that the biophysical phe- tissue permeability, which defines the ability of the porous nomena governing the ONH physiology occur at different material to allow fluid passing through it, is assumed pro- scales in time and space. For example, BP, IOP and retro- portional to the square of N; the solid matrix is assumed to laminar tissue pressure oscillate within a cardiac cycle (1 behave as nonlinear elastic material. Blood flow is driven by second), but also during the course of a day (24 hours). The the difference between the arterial pressure in the short PCAs characteristic time for the onset of autoregulation is and the venous pressure in the CRV. The lamina cribrosa approximately 20 seconds,119 and the biomechanically deforms under the combined action of IOP, retrolaminar induced remodeling of the collagen network in the ONH takes tissue pressure and scleral tension. This exploratory 2- several months or years. The accurate modeling and simu- dimensional analysis suggested that the degree of fixity at lation of multiscale problems is still an open area of research, the conjunction with the sclera has a strong influence on the and therefore the modeling of the ONH offers stimulating distributions of stresses and strains, as suggested in other opportunities for groundbreaking activities from both the studies,53,193 but also on the blood flow within the lamina. In clinical and theoretical viewpoints. The application of dy- particular, the inner surface of the lamina was found to be namic modeling to reveal the mechanistic interplay of pre- more susceptible to experiencing reduced blood supply viously unseen physiologic relationships holds the potential following IOP elevation. Despite the many simplifying to advance medical care in ophthalmic disease and provide assumptions adopted in the model, most importantly the patients and clinicians new hope for future diagnosis and choice of a 2-dimensional geometry, the satisfactory quali- . tative agreement between experimental data and numerical simulations encourages further investigation of poroelastic models to describe the complex mechanisms governing ONH 8. Method of literature search perfusion.

A literature search using the PubMed and Web of Science 7.3. Autoregulation of ONH blood flow search engines and available library databases was used with reference cross-matching to obtain relevant peer reviewed Mathematical models have been used to describe autor- articles published on blood flow autoregulation and ocular egulation of blood flow in various organs, including biomechanics and hemodynamics. The article search 2,48,213,229 37,116,127,202 13,31,227 brain, kidney, and skeletal muscle. included available published studies from 1900 to April 2015. To date, though, the theoretical modeling of autoregulation Separate searches were performed by the primary author in the ONH has not yet been addressed. Our group has (D.P.), and 3 independent researchers (G.G., A.M.H., and J.A.) developed a vascular wall mechanics model to predict the until relevant articles were identified. A few selected articles relative importance of regulatory mechanisms in achieving published before 1990 are included for their clinical relevance, 14 blood flow regulation in the retina. Resistance vessels are but the review is mainly based on articles published in the last assumed to respond to changes in pressure, shear stress, 2 decades. The following were the major search terminologies carbon dioxide (CO2), and the downstream metabolic state used: “autoregulation in the optic nerve head,” “neurovascular communicated via conducted responses. The model quan- coupling,” “ocular biomechanics and hemodynamics,” “me- tifies the relative contributions of different mechanisms to chanical influences on autoregulation,” “metabolic autor- retinal autoregulation and shows that the combined effects of egulation,” “effects of impaired blood flow regulation,” “optic the metabolic and CO2 responses are critical for achieving neuropathies,” “glaucoma,” “blood flow regulation modeling.” retinal autoregulation. It would be extremely interesting to Articles published in languages other than English were also apply a similar modeling approach to the autoregulation in included during the literature search. the ONH, combine it with the poroelastic model mentioned above,33 and explore the relative contributions of the various mechanisms described in section 5. Most of the models for blood flow autoregulation, including the one developed by our Acknowledgments group for the retinal circulation,14 are not time dependent. Thus, they are suitable to study static, but not dynamic, The authors acknowledge the valuable contribution of Ales- autoregulation. The dAR has been addressed by models for sandra Cantagallo, who realized the anatomic drawings (Figs. brain154,228 and kidneys.135 1e5) published in this article. survey of ophthalmology 61 (2016) 164e186 181

references endothelial injury from nitric oxide and superoxide. Proc Natl Acad Sci U S A. 1990;87(4):1620e4 21. Berne RM, Knabb RM, Ely SW, Rubio R. Adenosine in the local regulation of blood flow: a brief overview. Fed Proc. 1. The Advanced Glaucoma Intervention Study (AGIS): 7. The 1983;42(15):3136e42 relationship between control of intraocular pressure and 22. Blitzer ML, Lee SD, Creager MA. Endothelium-derived nitric visual field deterioration. The AGIS Investigators. Am J oxide mediates hypoxic vasodilation of resistance vessels in Ophthalmol. 2000;130(4):429e40 humans. Am J Physiol. 1996;271(3 Pt 2):H1182e5 2. Alastruey J, Moore SM, Parker KH, et al. Reduced modelling 23. Blumenroder S, Augustin AJ, Koch FH. The influence of of blood flow in the cerebral circulation: coupling 1-D, 0-D intraocular pressure and systemic oxygen tension on the and cerebral auto-regulation models. Int J Numer Methods intravascular pO2 of the pig retina as measured with Fluids. 2008;56(8):1061e7 phosphorescence imaging. Surv Ophthalmol. 3. Alm A, Bill A. Ocular and optic nerve blood flow at normal 1997;42(Suppl 1):S118e26 and increased intraocular pressures in monkeys (Macaca 24. Boltz A, Schmidl D, Werkmeister R, et al. Regulation of optic irus): a study with radioactively labelled microspheres nerve head blood flow during combined changes in including flow determinations in brain and some other intraocular pressure and arterial blood pressure. J Cereb tissues. Exp Eye Res. 1973;15(1):15e29 Blood Flow Metab. 2013;33(12):1850e6 4. Altman JD, Kinn J, Duncker DJ, Bache RJ. Effect of inhibition 25. Bouzas EA, Donati G, Pournaras CJ. Distribution and of nitric oxide formation on coronary blood flow during regulation of the optic nerve head tissue PO2. Surv exercise in the dog. Cardiovasc Res. 1994;28(1):119e24 Ophthalmol. 1997;42(Suppl 1):S27e34 5. Anderson DR. Introductory comments on blood flow 26. Buerk D, Riva C. Adenosine enhances functional activation autoregulation in the optic nerve head and vascular risk of blood flow in cat optic nerve head during photic factors in glaucoma. Surv Ophthalmol. 1999;43(Suppl 1):S5e9 stimulation independently from nitric oxide. Miscovasc Res. 6. Anderson DR. Glaucoma, capillaries and pericytes. 1. Blood 2002;64(2):254e64 flow regulation. Ophthalmologica. 1996;210(5):257e62 27. Buerk DG, Riva CE, Cranstoun SD. Nitric oxide has a 7. Anderson DR. Ultrastructure of the optic nerve head. Arch vasodilatory role in cat optic nerve head during flicker Ophthalmol. 1970;83(1):63e73 stimuli. Microvasc Res. 1996;52(1):13e26 8. Anderson DR. Vascular supply to the optic nerve of primates. 28. Burgoyne CF. A biomechanical paradigm for axonal insult Am J Ophthalmol. 1970;70(3):341e51 within the optic nerve head in aging and glaucoma. Exp Eye 9. Anderson DR. Ultrastructure of human and monkey lamina Res. 2011;93(2):120e32 cribrosa and optic nerve head. Arch Ophthalmol. 29. Burgoyne CF, Downs JC, Bellezza AJ, et al. The optic nerve 1969;82(6):800e14 head as a biomechanical structure: a new paradigm for 10. Anderson D, Hoyt W, Hogan M. The fine structure of the understanding the role of IOP-related stress and strain in the astroglia in the human optic nerve and optic nerve head. pathophysiology of glaucomatous optic nerve head damage. Trans Am Ophthalmol Soc. 1967;65:275e305 Prog Retin Eye Res. 2005;24(1):39e73 11. Andrews RM, Griffiths PG, Johnson MA, Turnbull DM. 30. Caprioli J, Coleman AL. Blood Flow in Glaucoma D. Blood Histochemical localisation of mitochondrial enzyme activity pressure, perfusion pressure, and glaucoma. Am J in human optic nerve and retina. Br J Ophthalmol. Ophthalmol. 2010;149(5):704e12 1999;83(2):231e5 31. Carlson BE, Arciero JC, Secomb TW. Theoretical model of 12. Araie M, Crowston J, Iwase A, et al. Clinical relevance of blood flow autoregulation: roles of myogenic, shear- ocular blood flow (OBF) measurements including effects of dependent, and metabolic responses. Am J Physiol Heart general medications or specific glaucoma treatment, in Circ Physiol. 2008;295(4):H1572e9 Weinreb RN, Harris A (eds) Ocular Blood Flow in Glaucoma. 32. Carlstrom M, Wilcox CS, Arendshorst WJ. Renal Amsterdam, Kugler Publications; 2009, pp 59e128 autoregulation in health and disease. Physiol Rev. 13. Arciero JC, Carlson BE, Secomb TW. Theoretical model of 2015;95(2):405e511 metabolic blood flow regulation: roles of ATP release by red 33. Causin P, Guidoboni G, Harris A, et al. A poroelastic model blood cells and conducted responses. Am J Physiol Heart Circ for the perfusion of the lamina cribrosa in the optic nerve Physiol. 2008;295(4):H1562e71 head. Math Biosci. 2014;257:33e41 14. Arciero J, Harris A, Siesky B, et al. Theoretical analysis of 34. Cavet M, Vittitow J, Impagnatiello F, et al. Nitric oxide (NO): vascular regulatory mechanisms contributing to retinal an emerging target for the treatment of glaucoma. Invest blood flow autoregulation. Invest Ophthalmol Vis Sci. Ophthalmol Vis Sci. 2014;55(8):5005e15 2013;54(8):5584e93 35. Chan DC. Mitochondria: dynamic organelles in disease, 15. Attariwala R, Giebs CP, Glucksberg MR. The influence of aging, and development. Cell. 2006;125(7):1241e52 elevated intraocular pressure on vascular pressures in the cat retina. Invest Ophthalmol Vis Sci. 36. Chen Q, Anderson DR. Effect of CO2 on intracellular pH and 1994;35(3):1019e25 contraction of retinal capillary pericytes. Invest Ophthalmol e 16. Attwell D, Buchan AM, Charpak S, et al. Glial and neuronal Vis Sci. 1997;38(3):643 51 control of brain blood flow. Nature. 2010;468(7321):232e43 37. Chen J, Sgouralis I, Moore LC, et al. A mathematical model of 17. Awai T. Angioarchitecture of intraorbital part of human the myogenic response to systolic pressure in the afferent e optic nerve. Jpn J Ophthalmol. 1985;29(1):79e98 arteriole. Am J Physiol Renal Physiol. 2011;300(3):F669 81 18. Bachoo RM, Kim RS, Ligon KL, et al. Molecular diversity of 38. Chien S. Mechanotransduction and endothelial cell astrocytes with implications for neurological disorders. Proc homeostasis: the wisdom of the cell. Am J Physiol Heart Circ e Natl Acad Sci U S A. 2004;101(22):8384e9 Physiol. 2007;292(3):H1209 24 19. Balaratnasingam C, Kang M, Yu P, et al. Comparative 39. ChiquetC,LacharmeT,RivaC,etal.Continuousresponseof quantitative study of astrocytes and capillary distribution in optic nerve head blood flow to increase of arterial blood pressure optic nerve laminar region. Exp Eye Res. 2014;121:11e22 in humans. Invest Ophthalmol Vis Sci. 2014;55(1):485e91 20. Beckman JS, Beckman TW, Chen J, et al. Apparent hydroxyl 40. Chiueh CC. Neuroprotective properties of nitric oxide. Ann N radical production by peroxynitrite: implications for Y Acad Sci. 1999;890:301e11 182 survey of ophthalmology 61 (2016) 164e186

41. Choi I, Chiu S. Expression of high-affinity neuronal and glial 64. Fu C, Sretavan D. Ectopic vesicular glutamate release at the glutamate transporter in the rat optic nerve. Glia. optic nerve head and axon loss in mouse experimental 1997;20(3):184e92 glaucoma. J Neurosci. 2012;32(45):15859e76 42. Christophe M, Nicolas S. Mitochondria: a target for 65. Fung YC. Biomechanics: circulation. New York, Springer; 1997 neuroprotective interventions in cerebral ischemia- 66. Galambos P, Vafiadis J, Vilchez S, et al. Compromised reperfusion. Curr Pharm Des. 2006;12(6):739e57 autoregulatory control of ocular hemodynamics in glaucoma 43. Chung H, Harris A, Evans D, et al. Vascular aspect in the patients after postural change. Ophthalmology. pathophysiology of glaucomatous optic neuropathy. Surv 2006;113(10):1832e6 Ophthalmol. 1999;43(Supplement I):S43e50 67. Galassi F, Giambene B, Varriale R. Systemic vascular 44. Chuquet J, Hollender L, Nimchinsky EA. High-resolution dysregulation and retrobulbar hemodynamics in normal- in vivo imaging of the neurovascular unit during spreading tension glaucoma. Invest Ophthalmol Vis Sci. depression. J Neurosci. 2007;27(15):4036e44 2011;52(7):4467e71 45. Coleman AL, Caprioli J. The logic behind target intraocular 68. Garho¨ fer G, Resch H, Weigert G, et al. Short-term increase of pressure. Am J Ophthalmol. 2009;147(3):379e80 intraocular pressure does not alter the response of retinal 46. CostaV,HarrisA,AndersonD,etal.Ocularperfusion and optic nerve head blood flow to flicker stimulation. Invest pressure in glaucoma. Acta Ophthalmol. Ophthalmol Vis Sci. 2005;46(5):1721e5 2014;92(4):e252e66 69. Garho¨ fer G, Zawinka C, Huemer K, et al. Flicker light- 47. Cull G, Burgoyne C, Fortune B, Wang L. Longitudinal induced vasodilation in the human retina: effect of lactate hemodynamic changes within the optic nerve head in and changes in mean arterial pressure. Invest Ophthalmol experimental glaucoma. Invest Ophthalmol Vis Sci. Vis Sci. 2003;44(12):2216e27 2013;54(6):4271e7 70. Geijer C, Bill A. Effects of raised intraocular pressure on 48. David T, Alzaidi S, Chatelin R, Farr H. Coupled retinal, prelaminar, laminar, and retrolaminar optic nerve autoregulation models. 13th International Conference on blood flow in monkeys. Invest Ophthalmol Vis Sci. Biomedical Engineering, Vols 1-3. 2009;23(1-3):1896e1899. 1979;18(10):1030e42 49. Davies PF. Overview: temporal and spatial relationships in 71. Gidday J, Park T. Adenosine-mediated autoregulation of shear stress-mediated endothelial signalling. J Vasc Res. retinal arteriolar tone in the piglet. Invest Ophthalmol Vis 1997;34(3):208e11 Sci. 1993;34(9):2713e9 50. Delles C, Michelson G, Harazny J, et al. Impaired endothelial 72. Gidday JM, Zhu Y. Nitric oxide does not mediate function of the retinal vasculature in hypertensive patients. autoregulation of retinal blood flow in newborn pig. Am J Stroke. 2004;35(6):1289e93 Physiol. 1995;269(3 Pt 2):H1065e72 51. Delori FC, Pflibsen KP. Spectral reflectance of the human 73. Glucksberg MR, Dunn R. Direct measurement of retinal ocular fundus. Appl Opt. 1989;28(6):1061e77 microvascular pressures in the live, anesthetized cat. 52. Donati G, Pournaras CJ, Munoz JL, Tsacopoulos M. [The role Microvasc Res. 1993;45(2):158e65 of nitric oxide in retinal vasomotor regulation]. Klin Monbl 74. Gordon GR, Choi HB, Rungta RL, et al. Brain metabolism Augenheilkd. 1994;204(5):424e6 dictates the polarity of astrocyte control over arterioles. 53. Dongqi H, Zeqin R. A biomathematical model for pressure- Nature. 2008;456(7223):745e9 dependent lamina cribrosa behavior. J Biomech. 75. Grieshaber M, Mozaffarieh M, Flammer J. What is the link 1999;32(6):579e84 between vascular dysregulation and glaucoma? Surv 54. Dorner GT, Garhofer G, Kiss B, et al. Nitric oxide regulates Ophthalmol. 2007;52(Suppl 2):S144e54 retinal vascular tone in humans. Am J Physiol Heart Circ 76. Grytz R, Girkin CA, Libertiaux V, Downs JC. Perspectives on Physiol. 2003;285(2):H631e6 biomechanical growth and remodeling mechanisms in 55. Downs JC, Roberts MD, Burgoyne CF, Hart RT. Multiscale glaucoma. Mech Res Commun. 2012;42:92e106 finite element modeling of the lamina cribrosa 77. Grytz R, Meschke G. A computational remodeling approach microarchitecture in the eye. Conf Proc IEEE Eng Med Biol to predict the physiological architecture of the collagen fibril Soc. 2009;2009:4277e80 network in corneo-scleral shells. Biomech Model 56. Drance SM, Douglas GR, Wijsman K, et al. Response of blood Mechanobiol. 2010;9(2):225e35 flow to warm and cold in normal and low-tension glaucoma 78. Grytz R, Meschke G. Constitutive modeling of crimped patients. Am J Ophthalmol. 1988;105(1):35e9 collagen fibrils in soft tissues. J Mech Behav Biomed Mater. 57. Essick EE, Sam F. Oxidative stress and autophagy in cardiac 2009;2(5):522e33 disease, neurological disorders, aging and cancer. Oxid Med 79. Grytz R, Meschke G, Jonas JB. The collagen fibril architecture Cell Longev. 2010;3(3):168e77 in the lamina cribrosa and peripapillary sclera predicted by a 58. Ferrari-Dileo G. Beta 1 and beta 2 adrenergic binding sites in computational remodeling approach. Biomech Model bovine retina and retinal blood vessels. Invest Ophthalmol Mechanobiol. 2011;10(3):371e82 Vis Sci. 1988;29(5):695e9 80. Grytz R, Sigal IA, Ruberti JW, et al. Lamina cribrosa 59. Filosa J, Bonev A, Straub S, et al. Local potassium signaling thickening in early glaucoma predicted by a microstructure couples neuronal activity to vasodilation in the brain. Nat motivated growth and remodeling approach. Mech Mater. Neurosci. 2006;9:1397e403 2012;44:99e109 60. Flammer J, Orgu¨ l S, Costa V, et al. The impact of ocular blood 81. Guidoboni G, Harris A, Cassani S, et al. Intraocular flow in glaucoma. Prog Retin Eye Res. 2002;21(4):359e93 pressure, blood pressure, and retinal blood flow 61. Folkow B, Sonnenschein RR, Wright DL. Loci of neurogenic autoregulation: a mathematical model to clarify their and metabolic effects on precapillary vessels of skeletal relationship and clinical relevance. Invest Ophthalmol Vis muscle. Acta Physiol Scand. 1971;81(4):459e71 Sci. 2014;55(7):4105e18 62. Frank RN, Dutta S, Mancini MA. Pericyte coverage is greater 82. Haefliger IO, Chen Q, Anderson DR. Effect of oxygen on in the retinal than in the cerebral capillaries of the rat. Invest relaxation of retinal pericytes by sodium nitroprusside. Ophthalmol Vis Sci. 1987;28(7):1086e91 Graefes Arch Clin Exp Ophthalmol. 1997;235(6):388e92 63. Frank RN, Turczyn TJ, Das A. Pericyte coverage of retinal and 83. Hall CN, Reynell C, Gesslein B, et al. Capillary pericytes cerebral capillaries. Invest Ophthalmol Vis Sci. regulate cerebral blood flow in health and disease. Nature. 1990;31(6):999e1007 2014;508(7494):55e60 survey of ophthalmology 61 (2016) 164e186 183

84. Hamard P, Hamard H, Dufaux J. Blood flow rate in the 105. Huang D, Swanson EA, Lin CP, et al. Optical coherence microvasculature of the optic nerve head in primary open tomography. Science. 1991;254(5035):1178e81 angle glaucoma. A new approach. Surv Ophthalmol. 106. Inoue R, Hangai M, Kotera Y, et al. Three-dimensional high- 1994;38(Suppl):S87e93, discussion S4. speed optical coherence tomography imaging of lamina 85. Hamard P, Hamard H, Dufaux J, Quesnot S. Optic nerve head cribrosa in glaucoma. Ophthalmology. 2009;116(2):214e22 blood flow using a laser Doppler velocimeter and 107. Jeffery G, Sharp C, Malitschek B, et al. Cellular localisation of haemorheology in primary open angle glaucoma and normal metabotropic glutamate receptors in the mammalian optic pressure glaucoma. Br J Ophthalmol. 1994;78(6):449e53 nerve: a mechanism for axon-glia communication. Brain 86. Hariharan N, Zhai P, Sadoshima J. Oxidative stress Res. 1996;741(1-2):75e81 stimulates autophagic flux during ischemia/reperfusion. 108. Jia Y, Morrison JC, Tokayer J, et al. Quantitative OCT Antioxid Redox Signal. 2011;14(11):2179e90 angiography of optic nerve head blood flow. Biomed Opt 87. Harris A, Januleviciene I, Siesky B, et al. Clinical Express. 2012;3(12):3127e37 measurements of ocular blood flow, in Weinreb RN, Harris A 109. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude- (eds) Ocular Blood Flow in Glaucoma. Amsterdam, Kugler decorrelation angiography with optical coherence Publications; 2009, pp 19e56 tomography. Opt Express. 2012;20(4):4710e25 88. Harris A, Jonescu-Cuypers C, Martin B, et al. Simultaneous 110. Kass MA, Heuer DK, Higginbotham EJ, et al. The Ocular management of blood flow and IOP in glaucoma. Acta Hypertension Treatment Study: a randomized trial Ophthalmol Scand. 2001;79(4):336e41 determines that topical ocular hypotensive medication 89. Harris A, Jonescu-Cuypers CP, Kagemann L, et al. Atlas of delays or prevents the onset of primary open-angle ocular blood flow. Oxford, UK, Butterworth-Heinemann; 2 glaucoma. Arch Ophthalmol. 2002;120(6):701e13, discussion ed, 2010 829e30. 90. Harris A, Rechtman E, Siesky B, et al. The role of optic nerve 111. Katsi V, Vlachopoulos C, Souretis G, et al. Association blood flow in the pathogenesis of glaucoma. Ophthalmol between retinal microcirculation and aortic stiffness in Clin North Am. 2005;18(3):345e53, v. hypertensive patients. Int J Cardiol. 2012;157(3):370e3 91. Harris A, Sergott R, Spaeth G, et al. Color Doppler analysis of 112. Klein R, Sharrett AR, Klein BE, et al. Are retinal arteriolar ocular vessel blood velocity in normal-tension glaucoma. abnormalities related to atherosclerosis?: The Am J Ophthalmol. 1994;118(5):642e9 Atherosclerosis Risk in Communities Study. Arterioscler 92. Hayreh SS. The blood supply of the optic nerve head and the Thromb Vasc Biol. 2000;20(6):1644e50 evaluation of itdmyth and reality. Prog Retin Eye Res. 113. Ko K, Ngai A, Winn H. Role of adenosine in regulation of 2001;20(5):563e93 regional cerebral blood flow in sensory cortex. Am J Physiol. 93. Hayreh SS. Factors influencing blood flow in the optic nerve 1990;259(6):H1703e8 head. J Glaucoma. 1997;6(6):412e25 114. Kondo M, Wang L, Bill A. The role of nitric oxide in 94. Hayreh SS. The 1994 Von Sallman Lecture. The optic nerve hyperaemic response to flicker in the retina and optic nerve head circulation in health and disease. Exp Eye Res. in cats. Acta Ophthalmol Scand. 1997;75(3):232e5 1995;61(3):259e72 115. Langhans M, Michelson G, Groh M. Effect of breathing 100% 95. Hayreh SS. Inter-individual variation in blood supply of the oxygen on retinal and optic nerve head capillary blood flow optic nerve head. Its importance in various ischemic in smokers and non-smokers. Br J Ophthalmol. disorders of the optic nerve head, and glaucoma, low- 1997;81(5):365e9 tension glaucoma and allied disorders. Doc Ophthalmol. 116. Layton AT. Mathematical modeling of kidney transport. 1985;59(3):217e46 Wiley Interdiscip Reviews-Systems Biol Med. 96. Hayreh SS, Joos KM, Podhajsky PA, Long CR. Systemic 2013;5(5):557e73 diseases associated with nonarteritic anterior ischemic optic 117. Leske MC. Ocular perfusion pressure and glaucoma: clinical neuropathy. Am J Ophthalmol. 1994;118(6):766e80 trial and epidemiologic findings. Curr Opin Ophthalmol. 97. Hayreh SS, Servais GE, Virdi PS. Fundus lesions in malignant 2009;20(2):73e8 hypertension. V. Hypertensive optic neuropathy. 118. Leske MC, Heijl A, Hussein M, et al. Factors for glaucoma Ophthalmology. 1986;93(1):74e87 progression and the effect of treatment: the early 98. Hayreh SS, Zimmerman MB, Podhajsky P, Alward WL. manifest glaucoma trial. Arch Ophthalmol. Nocturnal arterial hypotension and its role in optic nerve 2003;121(1):48e56 head and ocular ischemic disorders. Am J Ophthalmol. 119. Liang Y, Downs J, Fortune B, et al. Impact of systemic 1994;117(5):603e24 blood pressure on the relationship between intraocular 99. Helmke BP. Molecular control of cytoskeletal mechanics by pressure and blood flow in the optic nerve head of hemodynamic forces. Physiology (Bethesda). 2005;20:43e53 nonhuman primates. Invest Ophthalmol Vis Sci. 100. Hernandez M. The optic nerve head in glaucoma: role of 2009;50(5):2154e60 astrocytes in tissue remodeling. Prog Retin Eye Res. 120. Liang Y, Fortune B, Cull G, et al. Quantification of dynamic 2000;19(3):297e321 blood flow autoregulation in optic nerve head of rhesus 101. Hernandez MR, Miao H, Lukas T. Astrocytes in glaucomatous monkeys. Exp Eye Res. 2010;90(2):203e9 optic neuropathy. Prog Brain Res. 2008;173:353e73 121. Lichter PR, Musch DC, Gillespie BW, et al. Interim clinical 102. Hirrlinger J, Gutterer JM, Kussmaul L, et al. Microglial cells in outcomes in the Collaborative Initial Glaucoma Treatment culture express a prominent glutathione system for the Study comparing initial treatment randomized to defense against reactive oxygen species. Dev Neurosci. medications or . Ophthalmology. 2000;22(5-6):384e92 2001;108(11):1943e53 103. Hollows FC, Graham PA. Intra-ocular pressure, glaucoma, 122. Lieberman MF, Maumenee AE, Green WR. Histologic studies and glaucoma suspects in a defined population. Br J of the vasculature of the anterior optic nerve. Am J Ophthalmol. 1966;50(10):570e86 Ophthalmol. 1976;82(3):405e23 104. Hoste AM, Boels PJ, Brutsaert DL, De Laey JJ. Effect of alpha- 123. Liew G, Campbell S, Klein R, et al. Ten-year longitudinal 1 and beta agonists on contraction of bovine retinal changes in retinal microvascular lesions: the atherosclerosis resistance arteries in vitro. Invest Ophthalmol Vis Sci. risk in communities study. Ophthalmology. 1989;30(1):44e50 2011;118(8):1612e8 184 survey of ophthalmology 61 (2016) 164e186

124. Lin WJ, Kuang HY. Oxidative stress induces autophagy in 145. Mulligan SJ, MacVicar BA. Calcium transients in astrocyte response to multiple noxious stimuli in retinal ganglion endfeet cause cerebrovascular constrictions. Nature. cells. Autophagy. 2014;10(10):1692e701 2004;431(7005):195e9 125. Liu B, Chen H, Johns TG, Neufeld AH. Epidermal growth 146. Nadler Z, Wang B, Wollstein G, et al. Repeatability of in vivo factor receptor activation: an upstream signal for transition 3D lamina cribrosa microarchitecture using adaptive optics of quiescent astrocytes into reactive astrocytes after neural spectral domain optical coherence tomography. Biomed Opt injury. J Neurosci. 2006;26(28):7532e40 Express. 2014;5(4):1114e23 126. Liu B, Neufeld AH. Expression of nitric oxide synthase-2 147. Neufeld AH, Hernandez MR, Gonzalez M. Nitric oxide (NOS-2) in reactive astrocytes of the human glaucomatous synthase in the human glaucomatous optic nerve head. optic nerve head. Glia. 2000;30(2):178e86 Arch Ophthalmol. 1997;115(4):497e503 127. Loutzenhiser R, Bidani A, Chilton L. Renal myogenic 148. Newman E. Functional hyperemia and mechanisms of responsedKinetic attributes and physiological role. Circ Res. neurovascular coupling in the retinal vasculature. J Cereb 2002;90(12):1316e24 Blood Flow Metab. 2013;33(11):1685e95 128. Luksch A, Polak K, Beier C, et al. Effects of systemic NO 149. Novack RL, Stefansson E, Hatchell DL. Intraocular pressure synthase inhibition on choroidal and optic nerve head blood effects on optic nerve-head oxidative metabolism measured flow in healthy subjects. Invest Ophthalmol Vis Sci. in vivo. Graefes Arch Clin Exp Ophthalmol. 2000;41(10):3080e4 1990;228(2):128e33 129. Lundgren O. Autoregulation of intestinal blood flow: 150. Nyborg NC, Nielsen PJ. The level of spontaneous myogenic physiology and pathophysiology. J Hypertens Suppl. tone in isolated human posterior ciliary arteries decreases 1989;7(4):S79e84 with age. Exp Eye Res. 1990;51(6):711e5 130. Mackenzie PJ, Cioffi GA. Vascular anatomy of the optic nerve 151. Offenhauser N, Thomsen K, Caesar K, Lauritzen M. Activity- head. Can J Ophthalmol. 2008;43(3):308e12 induced tissue oxygenation changes in rat cerebellar cortex: 131. Maiuri MC, Zalckvar E, Kimchi A, Kroemer G. Self-eating and interplay of postsynaptic activation and blood flow. self-killing: crosstalk between autophagy and apoptosis. Nat J Physiol. 2005;565(Pt 1):279e94 Rev Mol Cell Biol. 2007;8(9):741e52 152. Ohashi T, Ishii Y, Ishikawa Y, et al. Experimental and 132. Malone PE, Hernandez MR. 4-Hydroxynonenal, a product of numerical analyses of local mechanical properties oxidative stress, leads to an antioxidant response in optic measured by atomic force microscopy for sheared nerve head astrocytes. Exp Eye Res. 2007;84(3):444e54 endothelial cells. Biomed Mater Eng. 2002;12(3):319e27 133. Malone P, Miao H, Parker A, et al. Pressure induces loss of 153. Ohno-Matsui K, Futagami S, Yamashita S, Tokoro T. Zinn- gap junction communication and redistribution of Haller arterial ring observed by ICG angiography in high connexin-43 in astrocytes. Glia. 2007;55(10):1085e98 . Br J Ophthalmol. 1998;82(12):1357e62 134. Marek B, Harris A, Kanakamedala P, et al. Cerebrospinal fluid 154. Olufsen MS, Nadim A, Lipsitz LA. Dynamics of cerebral pressure and glaucoma: regulation of trans-lamina cribrosa blood flow regulation explained using a lumped parameter pressure. Br J Ophthalmol. 2014;98(6):721e5 model. Am J Physiol Regul Integr Comp Physiol. 135. Marsh DJ, Sosnovtseva OV, Chon KH, Holstein-Rathlou NH. 2002;282(2):R611e22 Nonlinear interactions in renal blood flow regulation. Am J 155. Onda E, Cioffi GA, Bacon DR, Van Buskirk EM. Physiol Regul Integr Comp Physiol. 2005;288(5):R1143e59 Microvasculature of the human optic nerve. Am J 136. Matsugi T, Chen Q, Anderson DR. Adenosine-induced Ophthalmol. 1995;120(1):92e102 relaxation of cultured bovine retinal pericytes. Invest 156. Orgu¨ l S, Gugleta K, Flammer J. Physiology of perfusion as it Ophthalmol Vis Sci. 1997;38(13):2695e701 relates to the optic nerve head. Surv Ophthalmol. 137. Matthews BD, Overby DR, Mannix R, Ingber DE. Cellular 1999;43:S17e26 adaptation to mechanical stress: role of integrins, Rho, 157. Orgu¨ l S, Meyer P, Cioffi G. Physiology of blood flow regulation cytoskeletal tension and mechanosensitive ion channels. and mechanisms involved in optic nerve perfusion. J Cell Sci. 2006;119(Pt 3):508e18 J Glaucoma. 1995;4(6):427e43 138. Metea M, Newman E. Glial cells dilate and constrict blood 158. Osborne NN. Mitochondria: their role in ganglion cell death vessels: a mechanism of neurovascular coupling. J Neurosci. and survival in primary open angle glaucoma. Exp Eye Res. 2006;26(11):2862e70 2010;90(6):750e7 139. Mishra A, Hamid A, Newman E. Oxygen modulation of 159. Osborne NN, Wood JP, Chidlow G, et al. Ganglion cell death neurovascular coupling in the retina. Proc Natl Acad Sci U S in glaucoma: what do we really know? Br J Ophthalmol. A. 2011;108(43):17827e31 1999;83(8):980e6 140. Moncada S, Palmer RM, Higgs EA. Nitric oxide: physiology, 160. Panerai R. Assessment of cerebral pressure autoregulation in pathophysiology, and pharmacology. Pharmacol Rev. humansda review of measurement methods. Physiol Meas. 1991;43(2):109e42 1998;19(3):305e38 141. Moore D, Harris A, Wudunn D, et al. Dysfunctional 161. Pasquale L, Jonas JB, Anderson DR. Anatomy and physiology, regulation of ocular blood flow: a risk factor for glaucoma? in Weinreb RN, Harris A (eds) Ocular Blood Flow in Clin Ophthalmol. 2008;2(4):849e61 Glaucoma. Amsterdam, Kugler Publications; 2009, pp 3e13 142. Morgan WH, Yu DY, Alder VA, et al. The correlation 162. Paulson OB, Strandgaard S, Edvinsson L. Cerebral between cerebrospinal fluid pressure and retrolaminar autoregulation. Cerebrovasc Brain Metab Rev. tissue pressure. Invest Ophthalmol Vis Sci. 1990;2(2):161e92 1998;39(8):1419e28 163. Peng X, Carhuapoma JR, Bhardwaj A, et al. Suppression of 143. Motallebipour M, Rada-Iglesias A, Jansson M, Wadelius C. cortical functional hyperemia to vibrissal stimulation in the The promoter of inducible nitric oxide synthase implicated rat by epoxygenase inhibitors. Am J Physiol Heart Circ in glaucoma based on genetic analysis and nuclear factor Physiol. 2002;283(5):H2029e37 binding. Mol Vis. 2005;11:950e7 164. Peng X, Zhang C, Alkayed NJ, et al. Dependency of cortical 144. Movaffaghy A, Chamot S, Petrig B, Riva C. Blood flow in the functional hyperemia to forepaw stimulation on human optic nerve head during isometric exercise. Exp Eye epoxygenase and nitric oxide synthase activities in rats. Res. 1998;67(5):561e8 J Cereb Blood Flow Metab. 2004;24(5):509e17 survey of ophthalmology 61 (2016) 164e186 185

165. Peppiatt CM, Howarth C, Mobbs P, Attwell D. Bidirectional optic nerve and retina. Prog Retin Eye Res. control of CNS capillary diameter by pericytes. Nature. 2005;24(2):183e215 2006;443(7112):700e4 188. Roberts MD, Grau V, Grimm J, et al. Remodeling of the 166. Pillunat L, Anderson D, Knighton R, et al. Autoregulation of connective tissue microarchitecture of the lamina cribrosa human optic nerve head circulation in response to increased in early experimental glaucoma. Invest Ophthalmol Vis Sci. intraocular pressure. Exp Eye Res. 1997;64(5):737e44 2009;50(2):681e90 167. Pollock DM, Keith TL, Highsmith RF. Endothelin receptors 189. Roff EJ, Harris A, Chung HS, et al. Comprehensive and calcium signaling. FASEB J. 1995;9(12):1196e204 assessment of retinal, choroidal and retrobulbar 168. Polska E, Ehrlich P, Luksch A, Fuchsja¨ger-Mayrl G. Effects of haemodynamics during blood gas perturbation. Graefes adenosine on intraocular pressure, optic nerve head blood, Arch Clin Exp Ophthalmol. 1999;237(12):984e90 and choroidal blood flow in healthy humans. Invest 190. Rose C, Ransom B. Gap junctions equalize intracellular Naþ Ophthalmol Vis Sci. 2003;44(7):3110e4 concentration in astrocytes. Glia. 1997;20(4):299e307 169. Porsaa K. Experimental studies on the vasomotor 191. Rosengarten B, Hecht M, Kaps M. Brain activity affects innervation of the retinal arteries. Acta Ophthalmol. dynamic but not static autoregulation. Exp Neurol. 1941;19(S18):9e201 2007;205(1):201e6 170. Porter JT, McCarthy KD. Hippocampal astrocytes in situ 192. Russell FD, Hamilton KD. Nutrient deprivation increases respond to glutamate released from synaptic terminals. vulnerability of endothelial cells to proinflammatory insults. J Neurosci. 1996;16(16):5073e81 Free Radic Biol Med. 2014;67:408e15 171. Pournaras C, Riva C, Bresson-Dumont H, et al. Regulation of 193. Sander EA, Downs JC, Hart RT, et al. A cellular solid model of optic nerve head blood flow in normal tension glaucoma the lamina cribrosa: mechanical dependence on patients. Eur J Ophthalmol. 2004;14(3):226e35 morphology. J Biomech Eng. 2006;128(6):879e89 172. Pries AR, Secomb TW, Gaehtgens P. Biophysical aspects of 194. Satcher R, Dewey CF Jr, Hartwig JH. Mechanical remodeling blood flow in the microvasculature. Cardiovasc Res. of the endothelial surface and actin cytoskeleton induced by 1996;32(4):654e67 fluid flow. Microcirculation. 1997;4(4):439e53 173. Puro DG. Physiology and pathobiology of the pericyte- 195. Sato M, Ohashi T. Biorheological views of endothelial cell containing retinal microvasculature: new developments. responses to mechanical stimuli. Biorheology. Microcirculation. 2007;14(1):1e10 2005;42(6):421e41 174. Quigley HA. Glaucoma. Lancet. 2011;377(9774):1367e77 196. Sato E, Sakamoto T, Nagaoka T, et al. Role of nitric oxide in 175. Quigley HA. Neuronal death in glaucoma. Prog Retin Eye Res. regulation of retinal blood flow during hypercapnia in cats. 1999;18(1):39e57 Invest Ophthalmol Vis Sci. 2003;44(11):4947e53 176. Quigley HA. Pathophysiology of the optic nerve head in 197. Schmetterer L, Findl O, Strenn K, et al. Role of NO in the O2 glaucoma, in McAllister JA, Wilson RP (eds) Glaucoma. and CO2 responsiveness of cerebral and ocular circulation in London; Boston, Butterworths; 1986 humans. Am J Physiol. 1997;273(6 Pt 2):R2005e12 177. Quigley HA, Broman AT. The number of people with 198. Schmetterer L, Polak K. Role of nitric oxide in the control of glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. ocular blood flow. Prog Retin Eye Res. 2001;20(6):823e47 2006;90(3):262e7 199. Schmidl D, Boltz A, Kaya S, et al. Role of nitric oxide in optic 178. Rein H. Vasomotorische regulationen. Ergebnisse der nerve head blood flow regulation during an experimental Physiologie. 1931;32(1):28e72 increase in intraocular pressure in healthy humans. Exp Eye 179. Resch H, Garhofer G, Fuchsjager-Mayrl G, et al. Endothelial Res. 2013;116:247e53 dysfunction in glaucoma. Acta Ophthalmol. 200. Schmidl D, GarhoferG, Schmetterer L. Thecomplex interaction 2009;87(1):4e12 between ocular perfusion pressure and ocular blood 180. Resch H, Karl K, Weigert G, et al. Effect of dual endothelin flowdrelevance for glaucoma. Exp Eye Res. 2011;93(2):141e55 receptor blockade on ocular blood flow in patients with 201. Schon EA, Manfredi G. Neuronal degeneration and glaucoma and healthy subjects. Invest Ophthalmol Vis Sci. mitochondrial dysfunction. J Clin Invest. 2003;111(3):303e12 2009;50(1):358e63 202. Sgouralis I, Layton AT. Theoretical assessment of renal 181. Riva CE, Cranstoun SD, Petrig BL. Effect of decreased ocular autoregulatory mechanisms. Am J Physiol Renal Physiol. perfusion pressure on blood flow and the flicker-induced 2014;306(11):F1357e71 flow response in the cat optic nerve head. Microvasc Res. 203. Shah R, Wormald RP. Glaucoma. BMJ Clin Evid. 2011;2011 1996;52(3):258e69 204. Shibata M, Oku H, Sugiyama T, et al. Disruption of gap 182. Riva C, Falsini B, Logean E. Flicker-evoked responses of junctions may be involved in impairment of autoregulation human optic nerve head blood flow: luminance versus in optic nerve head blood flow of diabetic rabbits. Invest chromatic modulation. Invest Ophthalmol Vis Sci. Ophthalmol Vis Sci. 2011;52(5):2153e9 2001;42(3):756e62 205. Shibata M, Sugiyama T, Hoshiga M, et al. Changes in optic 183. Riva CE, Geiser M, Petrig BL, Beijing PRCOBFRA. Ocular blood nerve head blood flow, visual function, and retinal histology flow assessment using continuous laser Doppler flowmetry. in hypercholesterolemic rabbits. Exp Eye Res. Acta Ophthalmol. 2010;88(6):622e9 2011;93(6):818e24 184. Riva CE, Grunwald JE, Petrig BL. Autoregulation of human 206. Shibata M, Sugiyama T, Kurimoto T, et al. Involvement of retinal blood flow. An investigation with laser Doppler glial cells in the autoregulation of optic nerve head blood velocimetry. Invest Ophthalmol Vis Sci. 1986;27(12):1706e12 flow in rabbits. Invest Ophthalmol Vis Sci. 185. Riva CE, Harino S, Petrig BL, Shonat RD. Laser Doppler 2012;53(7):3726e32 flowmetry in the optic nerve. Exp Eye Res. 207. Sigal IA, Flanagan JG, Tertinegg I, Ethier CR. Modeling 1992;55(3):499e506 individual-specific human optic nerve head biomechanics. 186. Riva C, Hero M, Titze P, Petrig B. Autoregulation of human Part I: IOP-induced deformations and influence of geometry. optic nerve head blood flow in response to acute changes in Biomech Model Mechanobiol. 2009;8(2):85e98 ocular perfusion pressure. Graefe’s Arch Clin Exp 208. Sigal IA, Flanagan JG, Tertinegg I, Ethier CR. Modeling Ophthalmol. 1997;235(10):618e26 individual-specific human optic nerve head biomechanics. 187. Riva C, Logean E, Falsini B. Visually evoked hemodynamical Part II: influence of material properties. Biomech Model response and assessment of neurovascular coupling in the Mechanobiol. 2009;8(2):99e109 186 survey of ophthalmology 61 (2016) 164e186

209. Sigal IA, Wang B, Strouthidis NG, et al. Recent advances in a theoretical and experimental study. Microvasc Res. OCT imaging of the lamina cribrosa. Br J Ophthalmol. 1998;56(3):233e52 2014;98(Suppl 2):ii34e9 228. Ursino M, Giannessi M. A model of cerebrovascular 210. Sines D, Harris A, Siesky B, et al. The response of retrobulbar reactivity including the circle of willis and cortical vasculature to hypercapnia in primary open-angle glaucoma anastomoses. Ann Biomed Eng. 2010;38(3):955e74 and . Ophthalmic Res. 2007;39(2):76e80 229. Ursino M, Lodi CA. A simple mathematical model of the 211. Singh S, Dass R. The central artery of the retina. II. A study of its interaction between intracranial pressure and cerebral distribution and anastomoses. Br J Ophthalmol. 1960;44:280e99 hemodynamics. J Appl Physiol (1985). 1997;82(4):1256e69 212. Sossi N, Anderson D. Effect of elevated intraocular pressure 230. van Beek A, Claassen J, Rikkert M. Cerebral autoregulation: on blood flow. Occurrence in cat optic nerve head studied an overview of current concepts and methodology with with iodoantipyrine I 125. Arch Ophthalmol. special focus on the elderly. J Cereb Blood Flow Metab. 1983;101(1):98e101 2008;28(6):1071e85 213. Spronck B, Martens EG, Gommer ED, van de Vosse FN. 231. Vasudevan SK, Gupta V, Crowston JG. Neuroprotection A lumped parameter model of cerebral blood flow control in glaucoma. Indian J Ophthalmol. 2011;59(Suppl): combining cerebral autoregulation and neurovascular S102e13 coupling. Am J Physiol Heart Circ Physiol. 232. Venkataraman S, Flanagan J, Hudson C. Vascular reactivity 2012;303(9):H1143e53 of optic nerve head and retinal blood vessels in glaucomada 214. Srikanth K, Kumar MA, Selvasundari S, Prakash ML. Colour review. Microcirculation. 2010;17(7):568e81 Doppler imaging of ophthalmic artery and central retinal 233. Wang L, Burgoyne C, Cull G, et al. Static blood flow artery in glaucoma patients with and without diabetes autoregulation in the optic nerve head in normal and mellitus. J Clin Diagn Res. 2014;8(4):VC01e2 experimental glaucoma. Invest Ophthalmol Vis Sci. 215. Srinivasan VJ, Adler DC, Chen Y, et al. Ultrahigh-speed 2014;55(2):873e80 optical coherence tomography for three-dimensional and en 234. Wang L, Cioffi GA, Cull G, et al. Immunohistologic evidence face imaging of the retina and optic nerve head. Invest for retinal glial cell changes in human glaucoma. Invest Ophthalmol Vis Sci. 2008;49(11):5103e10 Ophthalmol Vis Sci. 2002;43(4):1088e94 216. St Lawrence KS, Ye FQ, Lewis BK, et al. Measuring the effects 235. Wang L, Cull G, Piper C, et al. Anterior and posterior optic of indomethacin on changes in cerebral oxidative nerve head blood flow in nonhuman primate experimental metabolism and cerebral blood flow during sensorimotor glaucoma model measured by laser speckle imaging activation. Magn Reson Med. 2003;50(1):99e106 technique and microsphere method. Invest Ophthalmol Vis 217. Sugiyama T, Araie M, Riva CE, et al. Use of laser speckle Sci. 2012;53(13):8303e9 flowgraphy in ocular blood flow research. Acta Ophthalmol. 236. Wang B, Nevins JE, Nadler Z, et al. Reproducibility of in-vivo 2010;88(7):723e9 OCT measured three-dimensional human lamina cribrosa 218. Sugiyama T, Utsumi T, Azuma I, Fujii H. Measurement of microarchitecture. PLoS One. 2014;9(4):e95526 optic nerve head circulation: comparison of laser speckle 237. Wang B, Nevins JE, Nadler Z, et al. In vivo lamina cribrosa and hydrogen clearance methods. Jpn J Ophthalmol. micro-architecture in healthy and glaucomatous eyes as 1996;40(3):339e43 assessed by optical coherence tomography. Invest 219. Takano T, Tian GF, Peng W, et al. Astrocyte-mediated control Ophthalmol Vis Sci. 2013;54(13):8270e4 of cerebral blood flow. Nat Neurosci. 2006;9(2):260e7 238. White LR, Bakken IJ, Sjaavaag I, et al. Vasoactivity 220. Takayama J, Tomidokoro A, Ishii K, et al. Time course of the mediated by endothelin ETA and ETB receptors in isolated change in optic nerve head circulation after an acute porcine ophthalmic artery. Acta Physiol Scand. increase in intraocular pressure. Invest Ophthalmol Vis Sci. 1996;157(2):245e52 2003;44(9):3977e85 239. Williamson TH, Harris A. Color Doppler ultrasound 221. Tamaki Y, Araie M, Fukaya Y, et al. Effects of lomerizine, a imaging of the eye and . Surv Ophthalmol. calcium channel antagonist, on retinal and optic nerve head 1996;40(4):255e67 circulation in rabbits and humans. Invest Ophthalmol Vis 240. Winkler M, Jester B, Nien-Shy C, et al. High resolution Sci. 2003;44(11):4864e71 three-dimensional reconstruction of the collagenous 222. Tamaki Y, Araie M, Kawamoto E, et al. Non-contact, two- matrix of the human optic nerve head. Brain Res Bull. dimensional measurement of tissue circulation in choroid 2010;81(2-3):339e48 and optic nerve head using laser speckle phenomenon. Exp 241. Wong TY, Mitchell P. The eye in hypertension. Lancet. Eye Res. 1995;60(4):373e83 2007;369(9559):425e35 223. Tomita K, Araie M, Tamaki Y, et al. Effects of nilvadipine, a 242. Yang G, Zhang Y, Ross M. Attenuation of activity-induced calcium antagonist, on rabbit ocular circulation and optic increases in cerebellar blood flow in mice lacking neuronal nerve head circulation in NTG subjects. Invest Ophthalmol nitric oxide synthase. Am J Physiol Heart Circ Physiol. Vis Sci. 1999;40(6):1144e51 2003;285(1):H298e304 224. Tamaki Y, Araie M, Tomita K, Tomidokoro A. Time-course of 243. Ye XD, Laties AM, Stone RA. Peptidergic innervation of the changes in nicardipine effects on microcirculation in retina retinal vasculature and optic nerve head. Invest Ophthalmol and optic nerve head in living rabbit eyes. Jpn J Ophthalmol. Vis Sci. 1990;31(9):1731e7 1996;40(2):202e11 244. Yuan L, Neufeld AH. Activated microglia in the human 225. TamakiY,AraieM,TomitaK,UrashimaH.Effectsofpranidipine, glaucomatous optic nerve head. J Neurosci Res. a new calcium antagonist, on circulation in the choroid, retina 2001;64(5):523e32 and optic nerve head. Curr Eye Res. 1999;19(3):241e7 245. Zonta M, Angulo MC, Gobbo S, et al. Neuron-to-astrocyte 226. Tomita K, Tomidokoro A, Tamaki Y, et al. Effects of signaling is central to the dynamic control of brain semotiadil, a novel calcium antagonist, on the retina and microcirculation. Nat Neurosci. 2003;6(1):43e50 optic nerve head circulation. J Ocul Pharmacol Ther. 246. Zweifach B, Lipowski H. Pressure-flow relations in blood and 2000;16(3):231e9 lymph microcirculation, in Renkin EM, Michel CC, Geiger SR 227. Ursino M, Colantuoni A, Bertuglia S. Vasomotion and blood (eds) Microcirculation. 4. Bethesda, Md, American flow regulation in hamster skeletal muscle microcirculation: Physiological Society; 1984, pp 251e307