Aqueous Humor Outflow Assessment

Abstract

In the past decade, many new pharmacological and surgical treatments have become available to lower intraocular pressure (IOP) for glaucoma. The majority of these options have targeted improving aqueous humor outflow (AHO). Concurrent to new treatments, research advances in AHO assessment have led to the development of new tools to structurally assess AHO pathways and to visualize where aqueous is flowing in the eye. These new imaging modalities have uncovered novel AHO observations that challenge traditional AHO concepts. New behaviors include segmental, pulsatile, and dynamic AHO which may have relevance to disease and the level of therapeutic response for IOP-lowering treatments. By better understanding the purpose and regulation of segmental, pulsatile, and dynamic AHO, it may be possible to innovate new glaucoma treatments aimed at these new AHO behaviors.

Glaucoma is a cause of progressive and irreversible optic neuropathy that is a leading cause of blindness world-wide¹. Glaucoma treatment equates to risk factor management with the biggest (and only reversible) risk factor being elevated intraocular pressure (IOP)². IOP is determined by the balance of aqueous humor production and aqueous humor outflow (AHO), and thus glaucoma treatment strategies have revolved around manipulating these factors.

Over the past 15 years, there has been an unprecedented explosion of outflow-targeted treatments to lower IOP for glaucoma. In the last year, two new drugs targeting conventional aqueous humor outflow have come to market, a rho-kinase inhibitor³ and a nitric-oxide (NO) donor⁴. Additionally, there has been an explosion of minimally invasive glaucoma surgeries (MIGS), most of which ablate⁵ or bypass⁶ the trabecular meshwork (TM)⁷. The challenge has been that these treatments do not always lower IOP or have side-effects, raising the question as to whether we fully understand AHO biology. Research spurred by this variable clinical success has now demonstrated that AHO is actually much more complicated than originally anticipated.

AHO naturally occurs through two pathways, the conventional (trabecular) and unconventional (uveoscleral) pathways. Both are modeled by the Goldman equation: IOP = (F_in - F_out)(R) + EVP (IOP = intraocular pressure [mm Hg]; F_in = aqueous production [μl/min]; F_out = unconventional outflow [μl/min]; R = conventional outflow resistance [mm Hg*min/μl]; and EVP = episcleral venous pressure [mm Hg])⁸. Another AHO pathway exists in the subconjunctival space that is only accessed during fistulizing glaucoma surgeries. The movement of aqueous humor can be considered for each of the pathways, but for the purpose of this review, the focus will be on the conventional/trabecular pathway given the greater abundance of treatment options and because current tools are better designed for it.

Tools to Study AHO

Today, there are many new tools to study AHO, including structural methods and functional fluid-flow assessment.

1. Structural Methods from the Laboratory to the Clinic.

Traditional static detection methods mainly relied on gross histological anatomy and electron microscopy combined with corrosion casting techniques (using different materials, like gelatin and neoprene) to reveal Schlemm’s canal and its outflow connections^9–11. However, these methods were only a static depiction of AHO pathways and did not provide a real-time AHO representation in live human eyes.

Anterior-segment optical coherence tomography (AS-OCT [swept-source and spectral domain]) is a modern method that non-invasively images AHO structural pathways in live humans at high-speed and resolution (~ 5 μm). The advantages of AS-OCT are that it provides the opportunity for longitudinal study^12–14. AS-OCT imaging of the TM, Schlemm’s canal (SC), collector channels (CC), and the intrascleral venous plexus correlates well with corrosion casting results¹⁵. AS-OCT also makes it possible for circumferential imaging around the limbus with three dimensional reconstruction of AHO pathways in live human eyes. However, there are some limitations to AS-OCT imaging. It cannot image unconventional or uveoscleral AHO. The exact identity of some of the reflective features such as an arc-like hypodensity in the angle have not been firmly established histologically. Not every lumen in the anterior segment necessarily contributes to AHO. For example, there are arteries, lymphatics, and veins that do not pass aqueous in the anterior segment. Also, circumferential and comprehensive mapping of AHO pathways around the limbus in live human eyes is very time-consuming because the distance mapped is large (2πR around the limbus) and relevant structures like CCs are small. Lastly, the relationship between structural pathways to fluid flow characteristics are not fully clear. It is uncertain for example if a larger lumen always equates to more flow with lower pressure.

2. Functional (Flow-Based) Methods, Including Static and Real-Time Tools.

Complementing structural imaging, flow-based AHO assessments visualize where aqueous fluid flows regardless of anatomic details. In static methods using labeled-beads in the laboratory in animal and human eyes, tracers are trapped in the TM allowing for quantification of TM utilization in histological section^16–24. However, a key limitation of static methods is that safety has not been established for humans and visualization of tracers require histological techniques that are not suitable for live individuals.

Real-time functional imaging methods include the episcleral venous fluid waves (EVFW)^25–27, canalography/channelography, and aqueous angiography in operating rooms. In EVFWs, irrigation from phacoemulsification units are used to clear blood from episcleral veins to identify them as being associated with AHO. Canalography/channelography started with canaloplasty surgery where SC was accessed either using a cannula or fiberoptic probe to deliver tracers into SC^28–30. This allowed visualization of AHO pathways around the limbus; however, the contribution of TM to AHO were not included^28–31. Aqueous angiography is a novel physiologic AHO assessment modality that delivers tracers, such as fluorescein and indocyanine green, into the anterior chamber followed by imaging using an angiographic camera³². The advantage is that aqueous angiography can be performed in live human subjects.

Combined, all of the above tools have been useful to better understand the physiology of AHO and have led to the discovery of several new AHO behaviors.

New AHO Behaviors

Using the above tools, a few new features of AHO have been more greatly appreciated or newly discovered: segmental, pulsatile, and dynamic AHO^{33, 34}.

Segmental AHO

Possibly the most significant novel AHO behavior that has been confirmed by modern tools has been the observation of segmental AHO. Segmental AHO directly challenges current AHO medical education. AHO is near-uniformly taught using single cross-sectional images of the anterior segment from the eye starting with aqueous humor production at ciliary processes. Next aqueous humor flows between the lens and iris into the anterior chamber. Then aqueous humor exits outflow pathways (conventional and unconventional) through the angle. The problem is that because a single two-dimensional image is used to capture all of AHO, students of the eye must generalize this flat description 360-degrees around the globe, circumferential and around the limbus. The implied outcome is that AHO is circumferential and uniform around the eye.

Instead, there were always clues that outflow was segmental. Structurally, histological sections of AHO pathways showed lumens of varying and inconsistent sizes in both normal and glaucomatous eyes. Today, structural OCT has shown that in living subjects that AHO pathways continue to be non-uniform and that the number and placement of CCs distal to SC are varied¹⁴.

However, it was tracer-based studies that truly drove the idea of segmental AHO. In the laboratory, fluorescent beads or spheres of various sizes had been introduced into eyes of multiple species^16–24. Invariantly, these tracers became trapped in the TM. This trapping was segmental with certain regions showing greater or lesser trace density. This observation has been replicated in rodents, cows, and humans. Percent-effective filtration length became a quantitative way to express the segmentalization by dividing the total length of TM showing trapped beads by the total length of TM assessed¹⁶. While clearly demonstrating segmental AHO, the disadvantage of bead-based studies was that the safety of these materials in live human eyes had not been established such that live-human experiments were not possible. Further, most of the studies did not employ real-time imaging given the need to enucleate eyes and process them for histological section and/or laboratory microscopy to see the tracer.

Thus, aqueous angiography was created to overcome some of these challenges. Aqueous angiography was developed in a bedside-to-bench-to-bedside spirit. Given the intention of translating AHO imaging to live humans, aqueous angiography started with instruments and tools from the clinic and operating room. Side-port blades, stock-cocks, tubing, syringes, and anterior chamber maintainers used in routine human surgery were acquired. Fluorescent tracers such as fluorescein and indocyanine green (ICG) were chosen because they were FDA-approved for human-use in retinal vascular angiography. Intraocular tracers, while not FDA-approved, they are routinely used off-label as intraocular stains for membrane peels during pars plana vitrectomy and capsulorhexis for phacoemulsification. For the angiographic camera, the Spectralis OCT+HRA (Heidelberg Engineering) was chosen because it imaged fluorescent tracers and could be used for anterior- and posterior-segment OCT, all FDA-approved functionalities.

After tracer delivery into the anterior chamber in the laboratory, post-limbal segmental AHO was seen in multiple species (pig, cow, and human) using multiple tracers (fluorescein and ICG)^{32, 35, 36}. Similar results were seen in live non-human primates and humans^37–40 (Fig. 1). In support of prior MIGS literature, most of the angiographic AHO was seen nasal⁴¹. A limbal-based line scan method applied to a group of live humans undergoing fluorescein aqueous angiography showed that ~ 45% of angiographic outflow was nasal with ~ 23/23/9% as superior/inferior/temporal, respectively³⁸.

Figure 1. Segmental Aqueous Angiographic Aqueous Humor Outflow in Live Humans.

Row A) Left eye of a 73 year-old male and Row B) Right eye of a 69 year-old female undergo aqueous angiography during cataract surgery. Each patient looked inferior, nasal, superior, and temporal to perform imaging in the 4 cardinal positions. Green and red arrows denote regions with and without local/segmental aqueous angiographic outflow signal, respectively.

Angiographic AHO patterns were reminiscent of outflow pathways including the intrascleral venous plexus (web-like interconnections) leading to aqueous and episcleral veins that took typical Y-shaped configurations (Fig 2). Clear and direct imaging of SC and CC were likely not occurring given their greater depth that would scatter fluorescent emission. However, it was notable that with careful observation, regions of prominent distal AHO were located nearby to peri-limbal “hot-spots” while regions with less distal AHO were not (Fig. 2). These “hot-spots” may be fuzzy reflections of AHO immediately distal to SC and CC.

Figure 2. Aqueous Angiographic Aqueous Humor Outflow Features in Live Humans.

A) Right eye of a 74 year-old male and B) Left eye of a 60 year-old female undergo aqueous angiography during cataract surgery. Green and red asterisks denote regions with and without local angiographic outflow. White arrows point out web-like outflow patterns likely representing the intrascleral venous plexus. Yellow arrows denote aqueous and episcleral veins given their Y-shaped appearance. Black arrows denote pre-limbal regions of increased local signal intensity more often associated with areas of distal outflow

With any new imaging modality, validation was needed. Here there were two basic tools. First, the Spectralis was concurrently an angiographic camera and an OCT. Thus, anterior-segment OCT was performed on- and off-angiographic structures^{32, 35–38, 40}. Excellent correlation was noted between angiographic signal and OCT intra-scleral lumens that appeared capable of carrying AHO. OCT off-angiographic structures did not show lumens or showed random lumens which could represent other luminal structures known to exist in the scleral wall (arteries, lymphatics, or veins not associated with AHO).

Validation was also performed using fluorescent and fixable dextrans. These tracers were large molecular weight sugars with two modifications. First, they had fluorescent attachments viewable by the fluorescein settings on the Spectralis. Second, they had lysine groups that allowed for fixation trapping of the tracer to nearby tissue. The first modification made this tracer viewable by aqueous angiography, showing segmental AHO in cow, pig, and human eyes^{32, 35, 36}. The second modification allowed for paraformaldehyde to “trap” the tracers in place after aqueous angiography was performed^{32, 35, 36}. Histological sectioning in areas with (but not without) distal aqueous angiographic signal showed TM tracer trapping. This observation validated that aqueous angiographic AHO imaging was based on tracer flow through the adjacent TM. It further reminded investigators of the importance of the TM. Today, while greater appreciation is growing for the contribution of distal AHO pathways to overall AHO homeostasis, it’s important to not forget the pivotal roles we know that the TM plays.

What regulates and controls segmental AHO is currently unclear. Similar to the discussion for dynamic AHO (below), there are a few possibilities. Segmental AHO could be controlled locally at the TM because trapping of tracer in the TM is segmental^16–24. Another option is that segmental AHO is set by the distal AHO pathways. A distal region could show no signal because pathways simply did not exist or because of vessel contraction⁴². The argument against absence is that in limited live human imaging, low-flow aqueous angiographic regions can still show a patent SC. What remains to be determined is whether there are structural AHO pathways differences between regions of low- and high-flow. Imaging in live patients is difficult because they are blurred [dilated], sedated, and do not follow instructions well during surgery. Additional evidence that distal AHO pathways do exist in low-flow regions comes from the result of increased angiographic AHO after trabecular bypass in post-mortem eyes³⁵ and eyes of living human glaucoma patients³⁹ (see below section on Clinical Relevance). Thus, the TM has to be at least partially involved in regulation of segmental AHO.

Pulsatile AHO

AHO pulsatility was first observed when the aqueous and episcleral venous junction was appreciated. Decades ago, Asher identified aqueous veins and their connection to episcleral veins via clear aqueous from aqueous veins mixing with blood from episcleral veins to create laminar flow columns of aqueous alternating with blood in distal episcleral veins^{43, 44}. In this arrangement, a glass rod could be placed on the distal episcleral vein and used to assess if the proximal episcleral venous blood would reflux into the aqueous vein (negative result) or if aqueous from the aqueous vein would reflux into the proximal episcleral vein (positive result). In any given eye, veins could be “positive” or “negative.” Glaucomatous eyes were shown to have more “negative” veins, consistent with diminished aqueous leaving the eye^{43, 44}.

Careful observations at the aqueous and episcleral vein junctions showed pulsatility⁴¹. Since episcleral veins carry blood flow driven by a beating heart, it was more interesting that aqueous joining blood in episcleral veins was pulsatile as well⁴¹.

The most robust work regarding pulsatile AHO came from structural-OCT evaluation, specifically phase-contrast OCT (ph-OCT). ph-OCT used software to process data normally acquired but not analyzed in traditional structural OCT assessment⁴⁵. Ph-OCT compared differences between consecutive B-scans to identify motion from static areas in the scan⁴⁵. Ph-OCT showed pulse-dependent TM motion in enucleated non-human primate and live human eyes^{12, 13, 46, 47}. In live human imaging, cardiac parameters were simultaneously recorded, and there was significant correlation between pulsatile TM motion and the cardiac pulse measured at the finger¹².

Aqueous angiography has now further confirmed pulsatile AHO by directly visualizing it. During real-time imaging in live non-human primates and live humans, the signal intensity of AHO was observed to be pulsatile^{37, 40}. While aqueous angiography has never been performed with synchronized cardiac measurements, aqueous angiography pulsatility rates in non-human primates (120 +/− 52 beats per minute [mean +/− SD])⁴⁰ correlated well with published non-human primate heart rates (70–180 beats per minute [resting] and 150–270 beats per minutes [exercise]). Thus, aqueous angiography pulsatility was consistent with a cardiac origin and ph-OCT results^{48, 49}.

While the overall data suggested that pulsatile AHO was cardiac in origin, the source of cardiac influence on the eye is less clear. Aqueous humor production may play a role. Aqueous humor production arises from the ciliary processes in the posterior segment and comes from passive diffusion, active secretion, and ultrafiltration⁵⁰. Among these three mechanisms, ultrafiltration driven by arteries in ciliary processes may be the source of pulsatile AHO as the aqueous is pushed into the eye.

Alternatively, a more likely scenario involves the choroid. The choroid is a large vascular bed in the posterior-pole of the eye that nourishes the outer retina⁵¹. Implicated in glaucoma and AHO, choroidal expansion is a leading hypothesis behind malignant glaucoma^{52, 53}. Because the eye has a rigid scleral shell with a fixed internal volume, choroidal expansion would consume space in the eye that could compress the vitreous. This could lead to decreased vitreous hydraulic conductivity, vitreous hydration, and forward movement of the lens/iris diaphragm that is often seen in malignant glaucoma^{52, 53}. In this case, the choroid could act as a cardiovascular “piston” in the back of the eye. A pulsatile piston driven by the heart rate could then serve to cause pulsatile AHO.

Besides its observation, the importance of pulsatile AHO is unclear and strongly needing of further study. One interesting parameter that is less explored and maybe interesting for the eye and AHO is the idea of a pressure set-points⁴¹. All pulsatile systems in the body display set-point behavior. Set point behavior is where pressure considerably below and above certain levels preclude the observation of pulsatility. Systemic blood pressure (BP) is the best example. As any medical student knows, when checking blood pressure using a stethoscope and sphygmomanometer, the first step is to increase the pressure in the cuff around the arm to a very high level. Auscultation of the brachial artery only yields a pulsatile sound when pressure in the cuff is lowered to a certain “set-point.” This point and when the pulsatility is first heard is the systolic BP. Then as the cuff pressure lowers even more, the pulsatile sound disappears at another “set-point.” This is a diastolic BP. Dysregulation in this system with altered set-points (systolic or diastolic BP) are well-known to be associated with significant morbidity and mortality. The set-points for observing pulsatile AHO in the eye and its relationship to disease is currently not known. However, as pulsatility can now be finally appreciated structurally (using OCT) and visually (using aqueous angiography), set-point studies in the eye can be performed.³⁷

Dynamic AHO

Dynamic AHO was a completely novel discovery that came from aqueous angiography imaging. Dynamic AHO was the most unexpected real-time AHO imaging finding because it refuted the classical depiction that outflow pathways were fixed and unchanging.

Dynamic AHO is different from pulsatile AHO because in pulsatile AHO, the AHO patterns are set. There is only a repetitive back and forth unidirectional increase and decrease in angiographic signal intensity. Dynamic AHO is instead defined by shifting AHO patterns. Some regions of the eye can open up to allow for additional AHO in previously low-flow regions, and other regions of the eye can shut down previously high-flow regions.

Dynamic AHO was only discovered when tracer-based studies for living subjects became available^{37, 40}. Thus, the development of aqueous angiography was critical. Large regions of AHO had to be viewable in living subjects in a real-time fashion. Structural studies, including histology and OCT, provided a too microscopic and cross-sectional view. Laboratory bead-based methods were not available for live humans because of safety concerns with intraocular injection of beads. Thus, the ability to detect dynamic AHO only became possible after aqueous angiography was developed and applied to living non-human primates and human subjects^37–40.

A few features regarding dynamic AHO are salient. First, these were infrequent events. During non-human primate study, during ~9 minutes of combined video recording, a total of 16 dynamic events were observed⁴⁰. Second, positive (increasing AHO) and negative (decreasing AHO) dynamic events could be seen simultaneously in a single eye³⁷. Lastly, dynamic events were sometimes associated with ocular motion³⁷. With respect to living subjects, this is relevant because extraocular muscle contraction is known to cause increased IOP⁵⁴. Thus, how eye motions impact angiographic AHO may be important for overall AHO and IOP homeostasis.

The reason why eyes have dynamic AHO is not currently understood. Why does AHO need to increase or decrease in different regions of the eye? To speculate, stable vision requires stable optics which would be best achieved by stable IOP. Under real-world conditions where IOP may vary (such as rubbing the eye that can initially increase IOP followed by ocular decompression that transiently decreases IOP), dynamic AHO can be a way to stabilize fluctuations by adjusting regional AHO.

In this case, it becomes very important to understand the mechanisms behind dynamic AHO. Currently, this is unclear, but there are several potential mechanisms. First, aqueous angiography performed using fixable fluorescent dextrans showed that local dextran trapping in the TM aligned with distal angiographic AHO location in pigs, cows, and human eyes^{32, 35, 36}. Thus, the pattern and location of TM/proximal segmental outflow mirrors that of distal segmental AHO. This potentially positions the TM as a possible regulatory check-point for dynamic AHO. Second, distal AHO pathways (aqueous and episcleral veins) are known to have muscular walls. Thus, local contraction/dilation of these pathways could also regulate dynamic AHO⁴². Third, dynamic AHO could occur passively due to influences such as ocular motion.

Dynamic AHO has also been seen in terms of the direction of flow. Veins in and around the eye and face are known to be valveless such that there is not much to prevent AHO in opposite directions⁵⁵. Real-time aqueous angiography has shown this (Clip 1). In this case, it is hard to imagine TM or local aqueous vein contractility controlling the direction of flow such that direction may be controlled by other factors like shifting pressure-gradients.

The Clinical Relevance of Better Understanding Aqueous Humor Outflow

Improved knowledge of AHO using structural and flow-based techniques not only improves the basic understanding of fluid flow in the eye but there is potential for clinical relevance. If the core reason for why IOP is elevated in glaucoma is better clarified, this then provides the opportunity to develop novel IOP-lowering strategies targeted at new pathways and mechanisms. Further, current therapies can be better understood and potentially augmented, both pharmacological and surgical.

Disease Modeling

The reason why IOP is elevated in glaucoma is because of increased AHO resistance⁵⁶. However, what is not fully understood is why resistance is pathologically increased. In the normal eye, the main source of resistance is located at the TM/SC border⁵⁷. Mechanistically, cellular hypotheses have been put forth to explain retarded AHO disease in glaucoma including impacted pore formation to transit aqueous across cells⁵⁸. Molecularly, other hypotheses have held that there is increased “plaque” material in the TM and/or deposition of extracellular matrix that could also diminish AHO⁵⁹.

The reason why new AHO behaviors are so important is because they alter how we interpret previous findings. Clearly, finding the exact cause of increased AHO resistance has been difficult, and research has accumulated slowly over decades. Possible explanations for this include glaucoma being a heterogeneous group of disease. Segmental AHO may also be a contributing factor. For example, if AHO is not circumferential and uniform around the limbus, prior research regarding TM biology could have been heavily influenced by where the TM biology was sampled. Were histological/imaging methods or TM cells grown in culture from high-flow or low-flow regions? Segmental AHO adds additional layers of biological variance that could have made prior findings difficult to appreciate or replicate

That is why new investigation taking segmental AHO into account is so important. Currently, a few reports have described the differences between high- and low-flow regions using bead-based methods. Versican is a large proteoglycan which has been shown to be elevated in low-flow regions¹⁷. Gene-expression and proteomic approaches have been taken also showing segmental difference including increased collagen VI and matrix metalloproteinase 3 in high-flow regions^{18, 19}. Our group has replicated some of these findings using aqueous angiography determined segmental AHO in post-mortem human eyes. Additionally, we studied pro-fibrotic pathways showing increased TGF-β, TGF-β receptor, thrombospondin-1 (a TGF-β pathway activator) and downstream effectors such as α-smooth muscle actin and the fibronectin-EDA isoform (but not total fibronectin) in low-flow regions (Saraswathy et al., 2018 ARVO B0133). Biomechanical studies have now also shown an increased elastic modulus (tissue stiffness) in low-compared to high-flow regions¹⁸.

Segmental AHO in the normal eye may hold further clues as to why AHO resistance is elevated in disease, particularly in its relationship to dynamic AHO. Increased AHO resistance in disease could be due to an expansion of normal low-flow regions. In this case, it becomes very important to not just understand the difference between low-flow and high-flow but potentially what the intermediate-flow regions are like. It becomes important to better understand how dynamic AHO occurs. The biology of those AHO regions that can be modified (for the better or worse) may be the best target for IOP lowering therapy that aims to return diseased tissue (low-flow regions) to a more native (high-flow) state.

To accomplish all of this, new models are needed. Certainly, studying human tissue or live humans are best. Imaging has already been done here. However, experimental designs are limited since human tissue is difficult to acquire for molecular analyses. Aqueous angiography is also invasive and done during routine clinical-care which demands prompt attention to surgical completion for best visual outcome. Thus, animal models are still an important place to start. Currently, OCT and aqueous angiographic study of a primary congenital glaucoma cat model with a mutation in a latent TGF-β binding protein has shown altered aqueous angiographic AHO and OCT AHO luminal appearance (McLellan et al., 2018 ARVO C0195). Early research is also being performed on dog models of primary open angle glaucoma also showing aqueous angiographic derangements (personal communication: Gillian McLellan).

Pharmacological Treatment of Glaucoma

Better understanding the biology of AHO also includes understanding how current drugs influence new AHO behaviors. By understanding this, it opens the door to develop new pharmacological treatments as well.

The oldest IOP-lowering therapy came from the Nigerian calabar bean which was a source of physostigmine and acted as a parasympathomimetic agent^60–62. Today, parasympathomimetic agents fall into two general classes, direct and indirect. Direct agents such as pilocarpine act by direct stimulation of muscarinic receptors on ocular muscles to cause contraction. Indirect agents, like physostigmine, instead inhibit acetylcholinesterase to elevate endogenous acetylcholine levels at the motor endplate to cause the same effects. With either direct or indirect activation, the result is contraction of the longitudinal ciliary muscle in the eye to pull down the scleral spur, open the TM, and decrease TM outflow resistance⁶².

Pharmacologically, a Rho-kinase (ROCK) inhibitor³ and NO-derivative⁴ became commercially available as IOP-lowering eye drops in 2018 and represent a new class of drug called cytoskeletal relaxing agents. Laboratory evaluation demonstrated that they dropped AHO resistance^{63, 64}. Focused on the TM, these pathways mechanistically impacted TM cytoskeleton, contractility, and extracellular matrix to promote easier AHO^65–68. However, long-studied in other systems, NO-derivatives are also well known to vasodilate and are first-line treatments for acute coronary syndrome and vasospastic angina⁶⁹. Decades-old animal studies have shown that NO-derivatives lower EVP in some species as an TM-independent mechanism to lower IOP⁷⁰. In the porcine eye, NO was shown to dilate regions of focal constriction in the distal AHO pathways⁷¹. During ROCK inhibitor clinical trials, up to 50% of patients demonstrated ocular surface vessel dilation³. Aqueous humor dynamic measurements under ROCK inhibitor treatment also showed statistically significant EVP reduction in humans⁷². Therefore, while undoubtedly impacting the TM, the question has also been raised as to whether these drugs can lower IOP through mechanisms outside or past the TM as well.

How muscarinic drugs and cytoskeletal relaxing agents impact segmental, pulsatile, and dynamic AHO is currently unknown. An area of active investigation, the impact of these drugs on the TM or distal outflow pathways may vary depending on whether low- or high-flow regions are being studied in the first place. It will be important to learn whether these drugs actually work by enhancing high-flow regions or by improving low-flow regions. This emphasis on potentially improvable regions again stresses the idea that AHO is not fixed, consistent with dynamic and rescuable (see Surgical Treatment below) AHO. By exactly isolating regions of AHO that can be improved and studying it, new molecular targets may be able to be pharmacologically targeted.

Surgical Treatment of Glaucoma

Given that variable results of MIGS spurred much of the above new structural and tracer-based AHO research, it was natural that considerable attention had been directed at better understanding MIGS.

The bulk of available MIGS target the TM either via ablation or TM bypass. Given that the majority of AHO resistance resides at the TM/SC interface⁵⁷, ablating this interface and creating a direct communication between the anterior chamber and SC was considered to be an good way to lower IOP. While safe, fast, and easy, trabecular MIGS were variably successful which caused the need to better understand AHO as a way to improve MIGS results^{5, 6}. Segmental AHO may play a central role here.

In segmental AHO, not only is AHO variable across an eye but different eyes show different segmental patterns. Thus, since trabecular MIGS are near universally conducted from a temporal approach aimed at bypassing or ablating the nasal TM, it was considered whether segmental AHO contributed to MIGS results variability. Was it better to target MIGS to high-flow regions since these areas were stable? However, an argument could be made that improving high-flow regions would be insufficient since high-flow regions already had good AHO. Instead, should surgery be targeted to low-flow regions to rescue those areas? However, maybe low-flow regions showed limited AHO in the first place due to inadequate anatomy. With new outflow imaging tools, including aqueous angiography, testing these questions was possible

First, an experimental design was needed. Fortunately, since the Spectralis imaged fluorescein and ICG, a two-dye system was possible. Starting in bovine eyes, it was shown that ICG aqueous angiography, followed by fluorescein aqueous angiography provided similar segmental patterns³⁶. Thus, initial ICG aqueous angiography could establish the baseline AHO pattern of an eye, an experiment could be performed, and fluorescein aqueous angiography could be used to query the impact.

This was first tested using trabecular bypass in post-mortem human eyes³⁵. The biological question that was asked was whether low-flow regions were capable of being rescued. ICG aqueous angiography established baseline patterns. Control touch of the TM near low-flow regions did not lead to improved fluorescein aqueous angiographic AHO. However, placement of a trabecular bypass stent in a low-flow region did lead to a ~ 17-fold improvement (p = 0.043) in fluorescein angiographic outflow.

Recently, the exact same paradigm was tested in living glaucoma patients³⁹. After establishing a baseline ICG aqueous angiographic pattern, trabecular bypass stents were placed in high- or low-flow regions with angiographic outcomes as a primary end-point. Multiple patterns were seen. In some cases, trabecular bypass led to no improvement. In some cases, new recruitment was seen where low flow-regions were rescued just as in the laboratory. This could be transient or long-lasting. In other cases, trabecular bypass in baseline high-flow regions led to faster and more dominant AHO near trabecular bypass stents as opposed to creating new AHO patterns. In the future, the next steps are to understand which patterns and whether trabecular bypass in different locations can yield improved outflow facility and IOP-lowering.

Future of AHO Imaging

In the future, additional AHO outflow pathways will be studied. Currently, real-time imaging modalities for uveoscleral outflow in living subjects are not available. Due to its deeper position (which makes fluorescence emission more difficult to observe because of tissue-induced scatter) and parallel nature to conventional AHO (which is faster such that any tracer in the anterior chamber would preferentially show conventional over unconventional AHO) imaging uveoscleral outflow will be harder. However, with ongoing development of uveoscleral MIGS^{73, 74}, it is possible that research in this space will be spurred the same way that trabecular MIGS inspired modern tools.

Also, another AHO pathway exists. The subconjunctival space has been used for decades in trabeculectomies and glaucoma drainage devices. Today, MIGS are also either available or under development to create blebs and access the subconjunctival space using minimal surgeries as well^{75, 76}. This has re-invigorated the old question of where aqueous flows in blebs. Options include connections from blebs to the systemic vasculature or subconjunctival lymphatics^{77, 78}. Moderns imaging tools such as OCT are just now being applied to studying these pathways⁷⁹. Modifications of aqueous angiography in our lab is currently underway to study subconjunctival AHO.

In conclusion, new IOP-lowering treatment modalities (MIGS and new drugs) have driven the development of new tools to better understand AHO. As new AHO behaviors (segmental, pulsatile, and dynamic) are now just beginning to be appreciated and explored in normal biology, glaucoma, and IOP-reducing treatments; the future of AHO imaging is wide. Now, the next steps are to improve the basic understanding of AHO in order to leverage it for better IOP-lowering treatments and perseveration of vision in glaucoma for the future.

Supplementary Material

Clip 1

Clip1: Bi-directional Aqueous Angiographic Aqueous Humor Outflow: Left eye of a 72 year-old undergoes aqueous angiography during cataract surgery. Here, the focus is on the nasal aspect of the eye where there is homogeneous post-limbal angioraphic signal. Inferior aqueous angiographic signal arises and moves more superior-nasal initially (~4 seconds) but then reverses and shifts inferior (12 seconds), demonstrating bi-directional aqueous humor outflow.