Towards a Chromatic Pupillometry Protocol for Assessing Melanopsin-Driven Post-Illumination Pupil Response in Basic Science and Clinical Investigations

TOWARDS A CHROMATIC PUPILLOMETRY PROTOCOL FOR ASSESSING MELANOPSIN-DRIVEN POST-ILLUMINATION PUPIL RESPONSE IN BASIC SCIENCE AND CLINICAL INVESTIGATIONS

Shaobo Lei

A thesis submitted in conformity with the requirements for the degree of Master of Science Institute of Medical Science University of Toronto

Towards a Chromatic Pupillometry Protocol for Assessing Melanopsin-Driven Post-Illumination Pupil Response in Basic Science and Clinical Investigations

Shaobo Lei

Master of Science

Institute of Medical Science University of Toronto

2016 Abstract

The pupillary light reflex (PLR) is mediated by intrinsically photosensitive retinal ganglions cells

(ipRGCs), a sub-group of retinal ganglion cells that contain photopigment melanopsin.

Melanopsin activation drives a sustained pupil constriction after the offset of light stimulus, this so-called post-illumination pupil response (PIPR) is an in vivo index of melanopsin-driven ipRGC photoactivity. PIPR can be assessed by chromatic pupillometry, but consensus on a standardized PIPR testing protocol has not been reached yet. The purpose of this thesis is to develop an optimized PIPR testing methodology, and to use it to investigate clinical and basic science questions related to melanopsin and ipRGCs. Based on previous pilot work on full-field chromatic pupillometry, a new and repeatable method was developed to measure PIPR induced by hemifield, central-field and full-field light stimulation. This chromatic pupillometry system was then used to investigate a series of basic science and clinical questions related to melanopsin and ipRGCs.

Acknowledgments

I would like to take this opportunity to express my gratitude to a number of people who have helped me to see through this thesis project. Without their contributions, support, advice and encouragement, this thesis would not be possible.

First of all, my deepest gratitude is extended to my supervisor, Dr. Agnes Wong, for offering me this amazing opportunity to study in Canada. This research project has greatly expanded my academic horizon, and opened up many opportunities for my future career.

My sincere appreciation is also extended to my co-supervisor, Dr. Herbert Goltz. Thank you for your guidance throughout the process. If it was not for your contributions, support and advice, this thesis would not be what it is today.

Next, I would like to express my gratitude to members of my program advisory committee, Dr. Graham Trope and Dr. John Flanagan, for their invaluable input and support in this project.

Special thanks to Manokaraananthan Chandrakumar, Jaime Sklar, Alan Blakeman, Luke Gane and Arham Raashid for their technical support.

Last, but certainly not least, I would like to thank all of my family for supporting me over the years. Of course, a special mention has to go to my lovely wife Jingwen, and our daughter Sophia. Thank you for being with me and giving me unconditional love and unwavering support.

This is for all of you who have made this thesis a reality. Thank you again.

iii

Contributions

Author and degree candidate, Dr. Shaobo Lei, made primary contributions to the conception and design of the study, acquisition, analysis and interpretation of data, drafting the thesis and revising it.

Program supervisors, Dr. Agnes Wong and Dr. Herbert Goltz made contributions in this thesis by providing experimental apparatus, participating study designing and thesis editing.

Ms. Jaime Sklar and Mr. Manokaraananthan Chandrakumar participated in data acquisition and analysis.

Mr. Alan Blakeman and Mr. Luke Gane provided technical support in the setup of the experimental apparatus; Mr. Gane also wrote the computer script that was used to analyze data.

Mr. Arham Raashid provided support for statistical analysis

Table of Contents

Acknowledgments ...... iii

Table of Contents ...... v

List of Tables ...... ix

List of Figures ...... x

Chapter 1 ...... 1

1 Literature Review ...... 1

1.1 General introduction ...... 1

1.2 Discovery and history of melanopsin-containing intrinsically photosensitive retinal ganglion cells ...... 3

1.2.1 Early behavioral observations in non-rod, non-cone animals ...... 3

1.2.2 The discovery of a non-visual photopigment: melanopsin ...... 4

1.2.3 The discovery of intrinsically photosensitive retinal ganglion cells ...... 5

1.2.4 Summary ...... 6

1.3 Anatomical and morphological features of ipRGCs ...... 7

1.3.1 ipRGCs population and distribution ...... 7

1.3.2 Morphological features of typical ipRGCs: M1 cells ...... 7

1.3.3 Morphological features of atypical ipRGCs: M2-M5 ...... 8

1.3.4 Central projections of ipRGCs ...... 9

1.3.5 Summary ...... 10

1.4 Physiological properties of ipRGCs ...... 11

1.4.1 Light absorption of melanopsin ...... 11

1.4.2 Menalopsin-driven phototransduction ...... 12

1.4.3 Electrophysiology of ipRGCs ...... 15

1.5 Functions of melanopsin and the ipRGC system ...... 21

1.5.1 Melanopsin and ipRGC function in non-image-forming photosensation ...... 21 v

1.5.2 Melanopsin and ipRGC functions in image-forming vision ...... 26

1.6 Chromatic Pupillometry: in vivo assessment of melanopsin-driven ipRGC photoactivity ...... 28

1.7 Preparation technical development: full-field chromatic pupillometry assessment of the melanopsin-driven post-illumination pupil response ...... 32

1.7.1 Rationale for developing full-field chromatic pupillometry ...... 32

1.7.2 Apparatus and experiment protocols ...... 33

1.7.3 Results: full-field vs central field PIPR in intensity and duration trials ...... 34

1.7.4 Discussions ...... 40

Chapter 2 ...... 45

2 Aims and Hypothesis ...... 45

2.1 Introduction ...... 45

2.2 Hemifield, Central-Field and Full-Field Chromatic Pupillometry for Assessing the Melanopsin-driven Post-illumination Pupil Response: A Methodological Study ...... 46

2.3 The effect of red light exposure on pre-existing PIPR: Implementing PIPR as an in- vivo index of melanopsin photoactivity in basic science research ...... 47

Chapter 3 ...... 49

3 Hemifield, Central-Field and Full-Field Chromatic Pupillometry for Assessing the Melanopsin-driven Post-illumination Pupil Response ...... 49

3.1 Introduction ...... 49

3.2 Methods ...... 51

3.2.1 Participants ...... 51

3.2.2 Apparatus ...... 51

3.2.3 Experimental Conditions and Procedure ...... 55

3.2.4 Data Analysis ...... 56

3.3 Results ...... 58

3.3.1 Post-Illumination Pupil Response (PIPR) ...... 58

3.3.2 Maximal Pupil Constriction (MPC)...... 63

3.4 Discussion ...... 67

Chapter 4 ...... 70

4 The Effect of Red light Exposure on Pre-existing Melanopsin-Driven Post-illumination Pupil Response ...... 70

4.1 Introduction ...... 70

4.2 Methods ...... 74

4.2.1 Participants ...... 74

4.2.2 Apparatus ...... 74

4.2.3 Testing conditions and protocols ...... 74

4.2.4 Data Analysis ...... 76

4.3 Results ...... 78

4.3.1 Experiment 1 ...... 78

4.3.2 Experiment 2 ...... 80

4.4 Discussion ...... 87

Chapter 5 ...... 91

5 General discussion ...... 91

5.1 Refinements of chromatic pupillometry PIPR testing ...... 91

5.2 Characteristics of melanopsin-driven post-illumination pupil response ...... 94

5.3 Applications of hemifield, central-field and full-field chromatic pupillometry induced PIPR ...... 96

Chapter 6 ...... 100

6 Conclusions ...... 100

Chapter 7 ...... 102

7 Future directions ...... 102

7.1 Using hemifield, central field and full-field chromatic pupillometry to investigate melanopsin-driven post-illumination pupil response in glaucoma patients ...... 102

7.1.1 Introduction ...... 102

vii

7.1.2 Methods ...... 104

7.1.3 Preliminary testing data ...... 105

7.2 Using chromatic pupillometry to investigate retinal dystrophies ...... 107

References ...... 110

viii

List of Tables

Table 1. Properties of Three Types of Photoreceptors: Cones, Rods and IPRGCs ...... 18

Table 2. Summary of Post-illumination Pupil Response (PIPR) Testing Conditions Published in Literature ...... 30

Table 3. PIPR values of each test trial from the 5 stimulation conditions...... 60

Table 4. Mean PIPR and p values for pairwise comparisons during the 5 testing conditions...... 61

Table 5. Test-retest reliability of PIPR measured during hemifield, central-field and full-field stimulation...... 62

Table 6. MPC values for each test trial from the 5 stimulation conditions...... 64

Table 7. Mean MPC and p values for pairwise comparisons during the 5 testing conditions...... 65

Table 8. Test-retest reliability of MPC measured during hemifield, central-field and full-field stimulation ...... 66

List of Figures

Figure 1. Mean normalized PIPR tracings to 1 s stimuli of varying intensity, Data from 10 visually-normal participants...... 35

Figure 2. Comparison of PIPR induced using 400 cd/m2 central-field stimuli vs 100-400 cd/m2 full-field stimuli ...... 36

Figure 3. Mean normalized pupil size from 10 to 30 s post-stimulation offset (PIPR(10-30 s)) and pupil size at 6 s post-stimulation offset (PIPR(6 s))...... 38

Figure 4. Mean PIPR to 100 cd/m2 and 400 cd/m2 full-field stimulation of varying duration from 10 visually-normal observers...... 39

Figure 5. PIPR(10-30 s) as a function of stimulus duration...... 40

Figure 6. Full-field chromatic pupillometry system ...... 53

Figure 7. The shutter panel apparatus...... 54

Figure 8. Mean hemifields, central-field and full-field PIPR from 10 visually-normal subjects. 59

Figure 9. Mean pupil responses from 10 visually-normal subjects in Experiment 1 ...... 79

Figure 10. Mean pupil responses to “red only” reference stimulations from 10 visually-normal subjects...... 81

Figure 11. Mean pupil responses from 10 visually-normal subjects in Experiment 2...... 82

Figure 12. Comparisons of PIPR tracings of “blue+red” conditions vs. “blue only” control from 10 visually normal participants ...... 83

Figure 13. Mean pupil responses from 7 visually-normal subjects in Experiment 2 ...... 84

Figure 14. Comparisons of PIPR tracings from “blue+red” conditions vs. “blue only” controls from 7 visually normal participants ...... 85

Figure 15. Comparisons of maximum pupil constriction (MPC) induced by “red only” stimuli vs. red stimuli presented on top of blue-light-induced PIPR (“Red on PIPR”) ...... 86

Figure 16. Individual PIPR testing results from visually normal participants and participants with glaucoma...... 106

Figure 17. Waveforms of comprehensive chromatic pupillometry testing...... 109

xi 1

Chapter 1

1 Literature Review

1.1 General introduction

Vision, a perception of the environment by interpreting information that is contained in light, is the primary source of sensory information for humans. As the only light-sensing neuro- tissue of our body (Nelson and Zucker 1981; Foster, Provencio et al. 1991), the retina contains two types of photoreceptor: rods and cones. For the last 150 years, these two types of photoreceptors were assumed to code all visual information. Rods are responsible for scotopic and mesopic vision, owing to their high sensitivity to dim light. Cones have three sub-types that are sensitive to long (L-cone), medium (M-cone) and short wavelength (S-cone) visible light. Distributed in high density in the posterior pole of retina, cones are responsible for photopic central vision and color vision. The information coded by the rods and cones is further processed in the inner retina that is then transmitted to the midbrain and then visual cortex, creating a real time image representation of our surroundings, integrating light brightness, contrast, color, object shape, size and movement (Demb 2008; Gollisch and Meister 2010). The amount and complexity of information our eyes extract from light is so enormous and our awareness of the world is so dominated by our sense of vision that it is difficult to comprehend the emerging evidence that our retina carries light sensory information that is separate from conscious vision.

About a decade ago, scientists made a major discovery that there is a parallel, subconscious, non-imaging-forming retinal photo sensing pathway mediated by a non-rod, non-cone photoreceptor (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002; Berson 2003): a small subset of retinal ganglion cells that response to light stimulation on their own with depolarizing action potential. These so-called intrinsically photosensitive retinal ganglion cells (ipRGCs) utilize a novel photopigment called melanopsin (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002). The

melanopsin-driven intrinsic photo-activity of ipRGCs represent as tonic signal of ambient light irradiance (Dacey, Liao et al. 2005), which is the primary afferent sensory input for circadian rhythm photoentrainment and some subconscious light reflexes, such as the pupillary light reflex (PLR). The decade following this major discovery has seen rapid expansion of literature on melanopsin and the ipRGC-mediated non-visual photoperception pathway, yet little is known about the involvement of melanopsin and ipRGCs in retinal diseases (Feigl and Zele 2014). Based on updated understanding of the ipRGC-mediated PLR pathway, a technique using monochromatic light called “chromatic pupillometry” to selectively induce pupil response driven by rods, cones and ipRGCs has emerged (Gamlin, McDougal et al. 2007). The melanopsin- driven intrinsic ipRGCs photoactivity can be assessed in vivo by using chromatic pupillometry to measure the sustained pupil constriction after the offset of high intensity blue light stimulation. This so-called post-illumination pupil response (PIPR) is a promising objective tool to evaluate this novel aspect of retinal function. However, consensus has not yet been reached on a standardized chromatic pupillometry protocol to test PIPR. The overarching purpose of this thesis is to refine the current PIPR testing methodologies and use them to investigate melanopsin/ipRGC-related basic science and clinical questions.

In this chapter, the history of the discovery of melanopsin and ipRGC, and the literature on cellular morphology, anatomy and physiology of this inner retina light-sensing pathway are reviewed.The functional properties of melanopsin/ipRGC system in mammals will be reviewed with emphasis on using PIPR as an in vivo index of the melanopsin-driven intrinsic photoactivity of ipRGCs in mammalian animals and humans in healthy and diseased states.

1.2 Discovery and history of melanopsin-containing intrinsically photosensitive retinal ganglion cells

1.2.1 Early behavioral observations in non-rod, non-cone animals

The first literature that indicates the presence of a photoreceptor in the inner retina can be traced back to 1927. Clyde Keeler, who was then a graduate student in Harvard University, observed seemingly paradoxical preservation of pupillary light reflex in mice lacking rods and cones due to outer retinal degeneration (Keeler 1927). Based on this observation, Keeler speculated on the presence of an unknown type of light-sensing cells in the inner retina. In a paper entitled ‘Iris movements in blind mice’ (Keeler 1927), Keeler wrote “…we may suppose that a rodless mouse will not see in the ordinary sense. Nevertheless, we can imagine the possibility of other forms of stimulation by light, such as through absorption by pigment cells, the contraction of the iris, or direct stimulation of the internal nuclear or ganglionic cells in the case of absence or faulty development of the external nuclear layer or of the rods.’’

Over some 70 years following Keeler’s remarkable paper, it was noted by many other investigators that mice lacking functional rods and cones maintained a normal 24-hour day/night activity cycle and their circadian rhythm could still be phase-shifted by artificially adjusting the ambient lighting cycle, a phenomenon called circadian photoentrainment (Ebihara and Tsuji 1980; Foster, Provencio et al. 1991; Freedman, Lucas et al. 1999; Lucas, Freedman et al. 2001). Yet when mice’s eye balls were surgically removed, the animal’s circadian rhythms could no longer be photoentrained (Yamazaki, Goto et al. 1999). By the end of 20th century, it became clear that neither rods nor cones are required for circadian photoentrainment, and that the mammalian circadian clock is regulated by additional photoreceptors in retina (Freedman, Lucas et al. 1999). Furthermore, by carefully analyzing the pupil light response in mice lacking rods and cones (rodless+coneless), Lucas et al. revealed that the response kinetics and spectral sensitivity of the residual retinal photo activity differed from those of rods and cones(Lucas, Douglas et al. 2001). The evidence of non-rod, non-cone photoreceptors in experimental rodent animals reconciled well with the clinical observation that some blind patients with photoreceptor

diseases still have well synchronized circadian rhythms, relatively preserved pupillary light response and photophobia (Klerman, Shanahan et al. 2002; Zaidi, Hull et al. 2007).

1.2.2 The discovery of a non-visual photopigment: melanopsin

The hunt for the putative photoreceptor responsible for preserved circadian clock and pupillary light reflexes in mammals lacking rods and cones was assisted by the parallel photobiology and genetic studies of extra-ocular photopigments in non-mammalian vertebrates. Unlike mammals, where the eye ball is the only site of photoperception (Nelson and Zucker 1981; Foster, Provencio et al. 1991), non-mammalian vertebrates like birds, fishes and amphibians possess a large variety of extra-ocular photoreceptive sites, such as the pineal gland, deep-brain photoreceptors and dermal photoreceptors (Shand and Foster 1999). All novel photopigments isolated from these non-visual photosensing sites/tissues consist of an opsin protein bound to a vitamin-A chromophore, which is structurally similar to the well-studied visual opsin of rods and cones (Peirson, Halford et al. 2009). However, up until 1998, none of the identified non-visual photopigment genes were found to be expressed in the mammalian genome. In 1998, Provencio at al. isolated a non-visual photopigment from the photosensitive dermal melanophores of Xenopus laevis (African clawed frog) (Provencio, Jiang et al. 1998). It was then given the name melanopsin. The coding messenger RNA of melanopsin was subsequently found in both mouse and human genomes by the same research team (Provencio, Rodriguez et al. 2000). The study also found that in humans melanopsin is expressed only in the eye. In situ hybridization histochemistry showed that melanopsin expression is exclusive to cells within the inner retina (Provencio, Rodriguez et al. 2000). The presence of melanopsin in human inner retina indicated that some retinal ganglion cells may contain melanopsin and carry the task of non-image-forming photosensing. Shortly afterward, it was collectively demonstrated by multiple teams that the retinohypothalamic tract, the fiber bundle that connects retina to the suprachiasmatic nuclei (SCN, the site of mammalian circadian rhythm centre), was exclusively formed by axons of melanopsin-containing retinal ganglion cells (Gooley, Lu et al. 2001; Hannibal, Hindersson et al. 2002). These findings strongly suggested that these ganglion cells were the mysterious third photoreceptor predicted by Keeler almost 80 years ago. The last piece of the puzzle was whether these retinal ganglion cells were indeed intrinsically photosensitive.

1.2.3 The discovery of intrinsically photosensitive retinal ganglion cells

The final piece of the puzzle was put in place in 2002 by two papers published in the journal Science (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002). Berson and colleagues employed retrograde tracer to label the small group of giant retinal ganglion cells that innervate SCN (Berson, Dunn et al. 2002). Patch clamp recording revealed that all these SCN-innervating retinal ganglion cells were intrinsically photosensitive: even when the rod and cone photoreceptors were pharmaceutically blocked or mechanically removed, they still responded to light stimulation with a depolarizing action potential, confirming that these retinal ganglion cells were indeed a third class of photoreceptors (Berson, Dunn et al. 2002). Therefore they were termed intrinsically photosensitive retinal ganglion cells (ipRGCs). Berson et al. (Berson, Dunn et al. 2002) also reported the sensitivity, spectral tuning, and kinetics of the light response of ipRGCs. When isolated from rods and cones’ influence, the ipRGCs were selectively sensitive to short wavelength light, with a peak absorption rate at around 480nm (blue light). They had a high threshold intensity and long integration time; it took long exposure to bright light to stimulate them. Most remarkably, ipRGCs responded to steady illumination with sustained tonic depolarization over a long period of time, a unique sensory property that is suitable for their hypothesized purpose of detecting ambient light irradiance for circadian rhythm photoentrainment and pupil size regulation. These photobiology characteristics were also consistent with the behavioral responses previously described in rodless+coneless mice (Lucas, Douglas et al. 2001). Berson and colleagues’ elegant work was complemented by an associated paper published in the same issue of Science. Hattar and colleagues (Hattar, Liao et al. 2002) identified ipRGCs using the same combination of retrograde tracing and single cell recording technique as described by Berson et al. (Berson, Dunn et al. 2002). By staining the flat mount preparation of retina for melanopsin immunoreactivity, they confirmed that all ipRGCs were invariably melanopsin-positive, whereas conventional ganglion cells lacking intrinsic light responses were all melanopsin-negative. They also employed a transgenic mouse model to demonstrate the projections of melanopsin-containing retinal ganglion cells. In this animal model, β-galactosidase, a marker enzyme, was targeted to the melanopsin gene locus so that all melanopsin-containing ipRGCs would express β-galactosidase. Histochemical staining showed

that β-galactosidase–positive RGC axons projected primarily to the SCN and the olivary pretectal nucleus (OPN, the pupillary light reflex centre), (Sun and May 2014) Stained fibers also “sparsely innervated the ventral lateral geniculate (vLGN) but did not innervate the dorsal lateral geniculate (dLGN)” (Hattar, Liao et al. 2002), as the LGN is the primary relay for the image- forming visual pathway, it seemed that ipRGCs did not have the functional properties for direct image formation. The authors concluded therefore that melanopsin-containing ipRGCs are “generally involved in non–image-forming visual function” (Hattar, Liao et al. 2002).

1.2.4 Summary

The discovery of melanopsin and melanopsin-containing intrinsically photosensitive retinal ganglion cells provides clear evidence that rods and cones are not the only types of photoreceptors in our retina, light information is also coded by a class of non-rod and non-cone inner retina photoreceptors, and transmitted to the brain in a parallel pathway from that of the image-forming visual system. This marked the beginning of a new line of research in visual science.

1.3 Anatomical and morphological features of ipRGCs

1.3.1 IpRGCs population and distribution

Intrinsically photosensitive retinal ganglion cells represent a small subset of highly specialized retinal ganglion cells. Hattar et al.’s study revealed an ipRGC count of only 2300- 2600 in rat retina (Hattar, Liao et al. 2002). With a total population of 100,000 retinal ganglion cells in the rat retina, (Crespo, O'Leary et al. 1985) these number accounted for 2-3% of the total retinal ganglion cell population in rats. In human retina, the proportion of ipRGCs in the total population of retinal ganglion cells is much lower: Dacey et al. reported about 3000 melanopsin- positive retinal ganglion cells within a human retina flat mount preparation. Assuming 1.5 million ganglion cells in the human retina (Wässle, Grünert et al. 1990). the melanopsin-positive cells represent only 0.2% of the total. (Dacey, Liao et al. 2005) The distribution of ipRGCs in rodent retina is asymmetric, with slightly higher density in superior and temporal quadrants than in the rest of retina (Hattar, Liao et al. 2002; Berson, Castrucci et al. 2010). In human retina, however, ipRGCs are more evenly distributed within the four quadrants, with higher density around the macular area except the foveola, where ipRGCs are almost absent. This is presumably to avoid interfering the acuity of central vision (Dacey, Liao et al. 2005).

1.3.2 Morphological features of typical ipRGCs: M1 cells

IpRGCs were initially described as a uniform, morphologically distinct subtype of retinal ganglion cells. The morphological features revealed in the early years of ipRGCs research include “a big cell body and long, sparse dendritic processes narrowly monostratified at the outermost lamina (OFF sublamina) of the inner plexiform layer (IPL) of the retina”(Hattar, Liao et al. 2002; Provencio, Rollag et al. 2002; Berson 2003). Most of the cell bodies (95%) are located in the ganglion cell layer (GCL) of retina, with remainder being in the inner nucleus layer (INL) (Dacey, Liao et al. 2005). Their dendritic fields are large, spanning about 500 µm (by far the largest dentritic field diameter of any retinal ganglion cells identified so far) (Dacey, Peterson et al. 2003; Dacey, Liao et al. 2005), forming an extensively overlapping meshwork within IPL. These cells invariably show intense melanopsin-immunoreactivity and strong intrinsic photosensitivity (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002; Dacey, Liao et al.

2004; Dacey, Liao et al. 2005; Schmidt, Taniguchi et al. 2008; Do, Kang et al. 2009). Their axons project to the SCN as well as other circadian rhythm related areas, and the pupil motion centre OPN as well(Berson, Dunn et al. 2002; Hattar, Liao et al. 2002; Morin, Blanchard et al. 2003; Sollars, Smeraski et al. 2003; Hattar, Kumar et al. 2006).

Since the original description of ipRGCs, their anatomical features and diversity have been under constant refinement. Now it is clear that they are not a homogeneous population of cells, but instead consist of numerous morphologically and functionally distinct cell types. The most “classical” ipRGCs as described above have now been classified as M1 cells (Hattar, Kumar et al. 2006).

1.3.3 Morphological features of atypical ipRGCs: M2-M5

Soon after the description of the M1-type, two new cell types, M2 and M3, were characterized. M2 ipRGCs are distinct from M1 by the fact that their processes stratify in the inner sublamina (ON sublamina) instead of outer sublamina (OFF sublamina) of the IPL. Besides the differences in dendritic stratification, M2 ipRGCs have even “larger cell body sizes and larger, more branched dendritic arbors than M1 cells” (Warren, Allen et al. 2003; Viney, Balint et al. 2007; Baver, Pickard et al. 2008; Schmidt, Taniguchi et al. 2008; Schmidt and Kofuji 2009; Berson, Castrucci et al. 2010; Schmidt, Chen et al. 2011; Schmidt, Do et al. 2011). M3 cells are bistratified and have dendrites located in both the OFF and ON sublamina of the IP;otherwise, M3 cells are similar to M2 cells in terms of cell body size and dendritic tree complexity (Schmidt and Kofuji 2011). Most recently, using a transgenic melanopsin reporter mouse line, investigators identified two more new ipRGCs types: M4 and M5. The melanopsin level of M4 and M5 is so low that it was not detectable by even by the most sensitive melanopsin immunohistochemistry procedure (Lin, Wang et al. 2004; Ecker, Dumitrescu et al. 2010; Estevez, Fogerson et al. 2012). Yet they did consistently show weak melanopsin-dependent intrinsic photosensitivity (Estevez, Fogerson et al. 2012). M4 cells have the largest cell bodies of all ipRGC subtypes. They stratify in the ON sublamina of IPL, and have slightly larger and more highly branched dendritic arbors than M2. In contrast, M5 ipRGCs also stratify in the ON

sublamina of IPL and they have small, bushy dendritic arbors branched uniformly around the cell body (Ecker, Dumitrescu et al. 2010).

1.3.4 Central projections of ipRGCs

Identifying the photoreceptors that set the circadian clock was the initial motivation for ipRGC research, so many early studies chose the circadian rhythm centre, the suprachiasmatic nucleus SCN of hypothalamus, as their starting point. Gooley and colleagues were the first to demonstrate that the retinohypothalamic tract was exclusively formed by axons of melanopsin- positive retinal ganglion cells (Gooley, Lu et al. 2001). Soon afterwards, ipRGCs were shown to send “dense projections to the SCN and other brain areas” that are related to circadian rhythm regulation, such as the intergeniculate leaflet (IGL) and the ventral lateral geniculate nucleus (vLGN) (Hattar, Liao et al. 2002). IpRGCs also provide primary innervation to the pupil motion centre, the olivary pretectal nucleus (OPN), then complete the loop of the pupillary light reflex through the Edinger-Westfal nucleus-ciliary ganglion-iris sphincter pathway (Berson 2003; Hattar, Kumar et al. 2006). Scattered fibers reach lateral and ventrolateral preoptic areas, influencing the secretion of reproductive hormones from the pituitary (Hattar, Kumar et al. 2006). A number of fibers reach medial amygdala, an area that plays a key role in innate emotional behaviors (Keshavarzi, Sullivan et al. 2014). Other regions receiving ipRGC input include the supraoptic nucleus, the ventral subparaventricular zone, and the lateral habenula (Gooley, Lu et al. 2003; Hattar, Kumar et al. 2006). Noseda et al. demonstrated that there is a connection between ipRGCs projections and the trigeminal system through the posterior hypothalamus. (Noseda, Kainz et al. 2009) The authors suggested that this could explain pain and photophobia in migraine patients who lack pattern vision (Noseda, Kainz et al. 2009).

All the above-mentioned central projections were contributed by typical ipRGCs (M1 cells), while M2-M5 cells provide substantial synaptic input to regions involved in image- forming vision, such as the dorsal lateral geniculate nucleus (dLGN) and the superior colliculus (SC) (Ecker, Dumitrescu et al. 2010), suggesting melanopsin’s influence on image-forming vision. In contrast, M1 cells only provide sparse fibers to the dLGN and the SC.

1.3.5 Summary

IpRGCs are spread across the retina. The morphological diversity of melanopsin- containing ipRGC implies diverse functional properties. Although the properties of responses elicited by each subtype remain to be further elucidated, it is generally accepted that typical ipRGCs (M1) form a retinal irradiance detecting “network”, and provide a primary afferent signal for most “classical” non-image-forming photo responses, such as circadian rhythm photoentrainment and pupil size regulation, while atypical ipRGCs (M2-M5) have significant influences on the conventional visual pathway, and may play regulating roles in image-forming vision (Dacey, Peterson et al. 2003; Dacey, Liao et al. 2004; Dacey, Liao et al. 2005; Ecker, Dumitrescu et al. 2010; Schmidt, Chen et al. 2011; Schmidt, Do et al. 2011; Estevez, Fogerson et al. 2012).

1.4 Physiological properties of ipRGCs

1.4.1 Light absorption of melanopsin

Much like the opsins in rods and cones, melanopsin is a transmembrane G-protein- coupled receptor that binds 11-cis retinal as its chromophore at resting state (Walker, Brown et al. 2008). After absorbing a photon, 11-cis retinal transforms into all-trans retinal, which subsequently drives conformational changes and eventually an activated state of melanopsin. The spectral sensitivity of mammalian melanopsin has an absorption peak at around 480 nm, within the range of short wavelength blue light, which has been demonstrated consistently in studies of purified rat melanopsin, the spectral tuning of ipRGCs, the behavioral response of animal models, and also in physiological measurements from human subjects (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Gamlin, McDougal et al. 2007; Do, Kang et al. 2009; Mure, Cornut et al. 2009; Do and Yau 2010).

Once being activated, the opsin’s chromophore conformation needs to be converted from all-trans back to 11-cis again to regain its photosensitivity. In many vertebrate photoreceptors (e.g. rods and cones) located in close proximity to underlying retinal epithelial cells (RPE), the thermally unstable all-trans retinal disassociates from the photoreceptor and is transported to the RPE. Subsequently, the all-trans retinal is converted to 11-cis in RPE through a multi-enzymatic biochemical cascade called the “visual cycle”. Taking advantage of ample 11-cis retinal supply from RPE, the bare opsin simply binds another 11-cis retinal and becomes photosensitive once again. In contrast, the melanopsin-containing retinal ganglion cells are located on the other side of the retina away from the RPE supply of 11-cis retinal, which raises an intriguing question about the regenerating mechanism of melanopsin. There is growing evidence that melanopsin is capable of functioning as a “bistable” opsin (Lucas 2006; Mure, Rieux et al. 2007; Rollag 2008; Mure, Cornut et al. 2009; Matsuyama, Yamashita et al. 2012; Sexton, Golczak et al. 2012), meaning that the photoisomerized chromophore does not dissociate from the opsin. Instead, the stable active (“meta”) state of the chromophore is reversed to a stable resting state through subsequent absorption of light. This property of the opsin is referred as “bistability”, a common opsin regenerating mechanism employed by many invertebrate photoreceptors (Koyanagi,

Kubokawa et al. 2005). The reversing absorption wavelength of bistable opsin is usually longer than the activating wavelength. Mure and colleagues (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009) demonstrated in vivo melanopsin bistability by measuring long-wavelength light potentiation of the pupil response to blue light. They also demonstrated that the spectrum of the reversing reaction of melanopsin was distinct from that of the forward reaction, and has its absorption peak at 587 nm, within the range of orange-red light (Mure, Cornut et al. 2009). However, the long-wavelength light potentiation of ipRGC firing was not observed in vitro (Mawad and Van Gelder 2008). More recently, a photochemistry study on purified rat melanopsin provided conflicting evidence, namely that the backward and forward spectra may overlap, with both peaking within the blue light range (467 nm vs. 476 nm) (Matsuyama, Yamashita et al. 2012). Therefore, the bistability property of melanopsin and its physiological consequences remain to be further elucidated.

1.4.2 Menalopsin-driven phototransduction

Phototransduction is the process by which light stimulation is converted to an electrical signal by the photoreceptor. Based on differences in the phototransduction cascade, photoreceptors can be divided into two broad categories: ciliary photoreceptors (usually found in vertebrate retina, with rods and cones being the most extensively studied) and rhabdomeric photoreceptors (found in invertebrate animals) (Hardie 2001; Arendt 2003; Arendt, Tessmar- Raible et al. 2004; Hardie and Postma 2008). The photoreceptor cascade in ciliary photoreceptor rods and cones has been well characterized: absorption of light causes isomerization (11-cis to all-trans) of the vitamin A chromophore bound to opsin protein, which leads to activation of the

G-protein transducin (a member of the G i/o family of G-proteins), resulting in phosphodiesterase activation and hydrolysis of cGMP to GMP (Arendt, Tessmar-Raible et al. 2004). Decreasing cGMP concentration results in closure of cyclic nucleotide-gated cation channels, leading to a hyperpolarizing receptor potential (Arendt 2003; Fu and Yau 2007). In comparison, the phototransduction cascade of rhabdomeric photoreceptor classically starts with G q/11-type G- protein activation, leading to the activation of phospholipase C (PLC) and the generation of 1,2- diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3) following the breakdown of phosphatidylinositol 4,5-bisphosphate (PIP2), subsequently resulting in the influx of Ca2+

through transient receptor potential (TRP) ion channels in the cell membrane, eventually the occurrence of depolarizing action potentials (Hardie 2001).

Melanopsin protein is structurally similar to invertebrate opsins, and it mediates a depolarizing photo response (Yau and Hardie 2009; Fain, Hardie et al. 2010). Therefore, initial studies on the phototransduction cascade of ipRGCs were based on the hypothesis that ipRGCs use a G q/11-type G-protein signaling pathway as described in rhabdomeric photoreceptors (Provencio, Jiang et al. 1998). Using pharmacological and transgenic approaches, investigators have since identified some key components in the melanopsin phototransduction cascade. (see discussion below)

1.4.2.1 Activation of G-proteins

G-proteins, also known as guanine nucleotide-binding proteins, are a family of proteins that function as intracellular molecular switches. All opsins are coupled to a specific member of the G-protein family to convert external light stimulation into intracellular signals. There have been over 20 subtypes of G-proteins identified so far, they are categorized into 4 groups: G s, G i/o, G q/11 ad G 12/13 based on the biochemical cascade they trigger (Davignon, Barnard et al. 1996). As discussed above, due to high homology to invertebrate opsin and the fact that melanopsin- containing cells respond to light with a depolarizing potential, it was assumed that melanopsin is coupled with G q/11-type protein (Provencio, Jiang et al. 1998). The most compelling evidence came from Graham and colleagues’ study (Graham, Wong et al. 2008), where ipRGCs’ intrinsic photo response was completely abolished by application of GPant-2a, a specific inhibitor of the

G q/11 class of G-proteins, but not by peptide GPant-2, a specific inhibitor of the G i/o class, confirming that phototransduction of ipRGCs is G q/11 dependent (Graham, Wong et al. 2008).

The G q/11 class itself consists of 4 members: Gαq, Gα11, Gα14, and Gα15 (Davignon, Barnard et al.

1996). mRNA of the first three sub-units were expressed in ipRGC (with Gα14 being the most commonly detected), but G α15 was not (Peirson, Oster et al. 2007; Graham, Wong et al. 2008). The exact G-protein sub-type that mediates melanopsin phototransduction in ipRGC remains to be determined.

1.4.2.2 Activation of phospholipase C

The next step in the classic G q/11 signaling pathway is the activation of PLCβ isoforms, which subsequently generate 1,2-diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3) by breaking down phosphatidylinositol 4,5-bisphosphate (PIP2) (Hubbard and Hepler 2006). By administrating the PLC inhibitor U73122, Graham and colleagues (Graham, Wong et al. 2008) completely abolished the melanopsin-driven ipRGC photoactivity, indicating that PLC is a key effector enzyme in the melanopsin phototransduction cascade. Among four identified PLCβ subtypes (PLCβ1, 2, 3, 4), PLCβ4 is most retina-specific. In PLCβ4 -/- gene knockout mouse line, the intrinsic photosensitivity of ipRGCs was almost completely diminished (Xue, Do et al. 2011). This finding further expands our knowledge of PLCβ subtype involved in melanopsin signaling.

There is also a growing body of literature about the downstream effectors following the activation of PLCβ4. Strong evidence indicates that DAG and IP3, two products of PIP2 hydrolysis, are not essential in the melanopsin phototransduction cascade (Graham, Wong et al. 2008). Investigators hypothesize that it is the breakdown of PIP2 itself that controls the opening of several ion channels, and subsequently leads to an action potential. When PIP2 is replenished, the post-stimulus potential will return to its resting state. To test this hypothesis, Graham and co- workers (Graham, Wong et al. 2008) use a drug called wortmannin to block the synthesis of PIP2. According to their hypothesis, when the rate of PIP2 restoration is reduced, the termination of post-stimulus potential would be delayed. Indeed, when wortmannin was added to the pipette solution, ipRGCs showed a prolonged photo response. However, more work is needed to determine the exact mechanism of how PIP2 controls the gating of ion channels.

1.4.2.3 Activation of transient receptor potential channel

The transient receptor potential (TRP) channel is a group of cation channels that exist widely on the plasma membrane of numerous animal cell types. TRPC (“C” for canonical) is a TRP subfamily first described in Drosophila fruit fly photoreceptors (Montell 2005). TRPC is permeable for all cations, with selectivity for calcium over sodium. Because melanopsin’s structure and phototransduction cascade highly resemble those of drosophila opsin, it was speculated that TRPC is responsible for mediating the initial influx of depolarizing current.

Findings of subsequent studies indeed supported this hypothesis: electrophysiological studies of ipRGC showed that light-induced current reverses at around 0 mV (Do, Kang et al. 2009; Schmidt and Kofuji 2009), suggesting a nonspecific cationic channel involvement. There is also evidence that calcium is the main carrier of light-induced current in ipRGC (Warren, Allen et al. 2006). Pharmaceutical studies showed that many drugs known to suppress TRPC channels can block the ipRGC photo response (Warren, Allen et al. 2006; Hartwick, Bramley et al. 2007). The most compelling evidence once again came from a genetic knockout animal line: Xue and colleagues (Xue, Do et al. 2011) showed that in double knockout mice lacking both TRPC6 and TRPC7 (two members of the TRPC subfamily expressed in retina), the photo response of the M1-type ipRGC was abolished. Single-knockout of TRPC6, TRPC7 or other members of the TRPC subfamily did not appear to alter the intrinsic photosensitivity of M1-ipRGCs. The identification of TRPC subfamily members that mediate the melanopsin-driven photoactivity is a major breakthrough in our understanding of melanopsin phototransduction. However, many key components in his cascade, such as the mechanism by which PLCβ activation triggers that opening of TRPC channels, and downstream action of other ion channels following the opening of TRPC channels, remain to be further elucidated.

1.4.3 Electrophysiology of ipRGCs

1.4.3.1 Intrinsic photo response of ipRGCs

Unlike rod and cone photoreceptors that respond to light with hyperpolarizing potentials in an analog manner, ipRGCs code digitally via action potential spikes. It has been demonstrated that the melanopsin-driven ipRGC phototransduction has a remarkably high gain: absorption of a single photon is sufficient to result in a recordable increase in spike firing (Do, Kang et al. 2009). This feature is brought about by some unique electrophysiological properties of ipRGCs. First, the photo current of the single photo response is relatively high, around 1-2.3 pA at body temperature, which is larger than that of that of rods and 100 times that of cones (Do, Kang et al. 2009; Schmidt and Kofuji 2009). Second, the single-photo-response is slow and long lasting, with an integration time of about 8 s, which is much longer than that of rods/cones and most other invertebrate photoreceptors (Do, Kang et al. 2009; Do and Yau 2010). The long integration

time of the single-photo-response may prevent it from being filtered by the membrane time constant, and also prolongs the after-stimulus effect (Do, Kang et al. 2009). Last but not least, ipRGCs exhibit a low spontaneous spiking rate in darkness (Wong, Dunn et al. 2007; Do, Kang et al. 2009), indicating that ipRGCs are operating at a membrane potential near the spiking threshold, so that a small depolarization caused by single photon absorption is sufficient to make the cells spike more frequently. It has been suggested that this high amplification is a compensating mechanism for the very low density of melanopsin in the retina (Do, Kang et al. 2009; Do and Yau 2010). In addition to the sparse distribution of ipRGCs, the density of melanopsin within ipRGCs is also low, with only a few molecules per square micron, compared to ~25,000 per square micron for rods and cones, which employ a specialized folded outer segment membrane to store a large amount of photopigment (Belenky, Smeraski et al. 2003; Do, Kang et al. 2009; Do and Yau 2010). The low density of melanopsin is thought to avoid interference with the photon absorption of image-forming photoreceptor rods and cones, which lie deeper in the retina (Lucas 2013). Therefore, even under bright illumination, ipRGCs capture very few photons, thus a high photo response gain is required in order to signal ambient light irradiance.

Despite the high gain in melanopsin phototransduction, the extremely low density of melanopsin molecules causes low sensitivity of the ipRGC intrinsic photo response. It was shown that half-saturation of the ipRGC response to flashes at their maximum absorption wavelength requires an intensity of 107 photons/μm2, ~106-fold higher than that for rods and ~104-fold higher than for cones at their corresponding maximum absorption wavelengths (Do, Kang et al. 2009; Do and Yau 2010), which makes ipRGCs much less sensitive photoreceptors than rods and cones.

Another remarkable feature of ipRGC electrophysiology is the ability to faithfully encode light energy over a prolonged period of time. When stimulated with long duration stimuli, the spiking rate builds relatively slowly (3s to 200 ms depending on the stimulation intensity) (Do, Kang et al. 2009) and reaches a maximal sustained firing rate that is “linearly proportional to light intensity” (Berson 2003; Dacey, Liao et al. 2005; Fu, Liao et al. 2005; Wong, Dunn et al.

2005; Do, Kang et al. 2009). When the bright light is left on, the firing rate of ipRGCs is “remarkably steady and sustained, without fatigue or adaptation to the continuous light stimulation” (Fu, Liao et al. 2005). When the light is turned off, ipRGCs do not immediately stop firing, but show a prolonged post-stimulus potential that lasts for tens of seconds to minutes (Berson 2003; Fu, Liao et al. 2005). A linear relation can also be found between the total number of ipRGCs spiking and the total amount of light energy that the ipRGCs are exposed to (Dacey, Liao et al. 2005). This unique “photon-counting” ability is ideal for the functional role of ipRGCs as an ambient light irradiance detectors for circadian rhythm and tonic pupil size regulation. The electrophysiological properties of ipRGCs, in comparison to rods and cones, are summarized in Table 1.

CONE ROD ipRGC

Functions photopic, central scotopic, peripheral pupillary response,

vision vision circadian rhythm, and

other non-visual

functions

Number 5 million 120 million 4000-12000

Location inner segment: outer inner segment: outer ganglion cell/inner

nuclear layer nuclear layer plexiform layers

outer segment: outer segment:

photoreceptor layer photoreceptor layer

Photopigment lodopsins Rhodopsin Melanopsin

Photosensitivity low extremely high Low

Receptive field Very small Small Very large

Peak absorption 420 (S), 534(M), 564 498 482

wavelength (nm) (L)

Patterns of light Rapid onset, early Rapid onset, Delayed onset and

pupillary response adaptation sustained on prolonged firing

continuous exposure

Table 1. Properties of Three Types of Photoreceptors: Cones, Rods and IPRGCs

1.4.3.2 Extrinsic synaptic input to ipRGCs

Beside their melanopsin-driven intrinsic photosensitivity, all ipRGCs also receive extensive extrinsic synaptic input that originates from rods and cones, and which is mediated via bipolar cells and amacrine cells (Wong, Dunn et al. 2007). As discussed above, rods and cones respond to light increments with hyperpolarizing potentials, resulting in a reduced release of the neurotransmitter glutamate. The second-order neurons, the bipolar cells, can be divided into two sub-groups depending on the glutamate receptor they employ. The bipolar cells that employ the metabotropic receptor harness glutamate as an inhibitory signal, so they reverse the signal of rods and cones, and show a depolarizing response on light incrementation. Therefore they are called ON bipolar cells. The remaining bipolar cells employ an ionotropic receptor that utilizes glutamate as an excitatory neurotransmitter. They hyperpolarize at light increment and depolarize at decrement, so they are called OFF bipolar cells (Kolb and Nelson 1995). Conventionally, the axons of ON bipolar cells terminate and make synaptic contact with the dendrites of ON retinal ganglion cells in the innermost sublamina (ON-sublamina) of IPL. Similarly, OFF bipolar cells stratify and contact OFF-RGCs in the outermost sublamina (OFF- sublamina) of IPL.

The classic ipRGCs (M1-cells) arborize in the OFF-sublamina, but paradoxically, they receive mostly ON bipolar cells input (Wong, Dunn et al. 2007; Schmidt and Kofuji 2010).IpRGCs also receive weak input from OFF cone bipolar cells (Wong, Dunn et al. 2007). Under physiological conditions, this minor extrinsic OFF response (spiking at the offset of light) of ipRGCs is buried by the withdrawal of the ON response, and is only observable after pharmacological blockage of amacrine cells and ON bipolar cells (Wong, Dunn et al. 2007). The physiological importance of the extrinsic OFF input to ipRGCs is still unclear. Dacey and colleagues reported a primate ipRGC with a prominent L+M-cone ON/S-cone OFF pattern of synaptic input. (Dacey, Liao et al. 2005). Although this feature may not be universal for all ipRGCs, it indicates that some ipRGCs may play a role in coding color opponency and modulating color perception (Dacey, Liao et al. 2005).

Putting all the evidence together, ipRGCs receive net excitatory synaptic input predominantly through ON bipolar cells. Interestingly, the synaptic input to ipRGCs is quite sustained compared to the synaptic input to the conventional RGCs (although not as sustained and long-lasting as compared to the melanopsin-driven response) (Wong, Dunn et al. 2007; Do and Yau 2010). IpRGCs can thus use both their extrinsic and intrinsic photo activity to signal light irradiance for long periods of time. In addition, rod/cone-driven extrinsic ipRGC photoactivity are faster and more sensitive than the melanopsin-driven intrinsic ipRGC photoactivity(Wong, Dunn et al. 2007), It has been suggested that combining the rod/cone-driven extrinsic photoactivity and melanopsin-driven intrinsic photoactivity expand the dynamic range of ipRGCs to steadily encode dim light and rapid change in light intensity (Lucas 2013).

1.5 Functions of melanopsin and the ipRGC system

1.5.1 Melanopsin and ipRGC function in non-image-forming photosensation

IpRGCs integrate the photo activity of all three known retinal photoreceptors (rod, cones and melanopsin-driven intrinsic ipRGC photo responses) to provide an afferent signal for a wide range of reflex and sub-conscious photo bio-responses responsible for many aspects of physiological and behavioral statuses, collectively referred as “non-image-forming visual functions”. Using genetically modified animal lines, investigators have gradually detailed the functional roles of melanopsin and melanopsin-containing ipRGCs.

1.5.1.1 Circadian rhythm photoentrainment

Most organisms living on earth (except some cave animals) are exposed to a 24-hour day/night cycle. The endogenous circadian clock has evolved to regulate their physiology and behaviours according to the varying demands throughout the day. Needless to say, these internal clocks are useless if they are not synchronized with the 24-hour day/night cycle of the real world. (A compelling example is jetlag, where our internal circadian clock is desynchronised and acting against our behavioural needs). Not surprisingly, the primary environmental synchronizing cue of circadian rhythm is the natural cycle of ambient light. Since the internal circadian cycle is not exactly 24 hours, this photo-induced synchronizing process known as photoentrainment is occurring constantly. Unlike non-mammalian vertebrates such as fish, birds, and reptiles that have extra-ocular circadian photoreceptors (Shand and Foster 1999; Helfrich-Förster, Winter et al. 2001; Vigh, Manzano et al. 2002), photoentrainment of mammalian animals originates solely from the retina of the eyes, this is confirmed by the fact that eye removal completely abolishes circadian photoentrainment (Freedman, Lucas et al. 1999; Yamazaki, Goto et al. 1999). It is now clear that the retinohypothalamic tract that connects retina to the superior chiasmatic nucleus, the endogenous circadian centre, is predominantly formed by axons of ipRGCs (Hattar, Liao et al. 2002; Hannibal, Hindersson et al. 2004). In order to demonstrate the functional roles of ipRGC cells, selective in vivo ipRGC ablation was induced by administration of diphtheria toxin to a transgenic mouse model where the diphtheria toxin receptors are expressed on ipRGCs, or by

intraocular injection of melanopsin antibody linked to saporin toxin (Göz, Studholme et al. 2008; Güler, Ecker et al. 2008; Hatori, Le et al. 2008). Photoentrainment is lost after these induced selective lesions in ipRGCs, meaning that ipRGCs are important generators and primary conduits for the photoentrainment afferent signal to the circadian clock. IpRGCs do not perform the task alone, but rather they receive synaptic input from conventional rod/cone circuits within the retina. Therefore, the light response of ipRGCs in the real world is a summation of rod/cone-mediated activities and intrinsic melanopsin-driven activity. How rods, cones and melanopsin work together to provide photoentrainment has been the subject of extensive study.

The fact that even extremely dim light near the threshold intensity of vision has significant influence on circadian rhythm indicates an important contribution from rods (Lall, Revell et al. 2010). There is also evidence that photoentrainment of rods is not restricted to very dim light. Genetically engineered “rod only” mice (lacking cones and melanopsin, but with an intact relay function of ipRGCs) entrain well to artificial dark/light cycles with the light portion set at 500 photopic lux (equivalent to normal indoor lighting levels), indicating that rods are a reliable source of irradiance coding over a moderate range of light intensities (Altimus, Güler et al. 2010). But what would happen if the lighting level was above the saturation point of rods? Under such conditions, cones are responsible for our vision; can they replace rods as irradiance detectors under high intensity lighting? It appears that cones are not up to this task: “cones only” mice (lacking both rods and melanopsin, preserved ipRGCs) failed to entrain reliably to dark/light cycles in the laboratory (Lall, Revell et al. 2010). When presented with a bright light pulse, both ipRGCs (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005) and neurons in SCN (Brown, Wynne et al. 2011) show a rapid increase in cell firing that subsides quickly to a lower level of sustained excitation. Pharmaceutical and genetic manipulation (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Wong, Dunn et al. 2007) suggested that the rapid phase of the response is driven by rods/cones, and the sustained cell response under constant bright light exposure and after the offset of light is driven by melanopsin. Light wavelengths that can activate cones but not melanopsin induced strong firing at lights on, but drove very little sustained firing under continuous light exposure (Brown, Wynne et al. 2011). These data indicate that cones adapt to bright light too quickly to provide sustained irradiance coding for photoentrainment. On the other

hand, melanopsin-driven ipRGC activity has a high threshold and sustainable excitability under bright light, providing a good compensation for the functional disadvantage of rods and cones. As discussed earlier, it has been long observed that a retina lacking rods and cones can still photoentrain (Ebihara and Tsuji 1980; Foster, Provencio et al. 1991; Freedman, Lucas et al. 1999; Lucas, Freedman et al. 2001), while melanopsin knock-out mice can entrain but the magnitude of behaviour response is substantially reduced (Panda, Sato et al. 2002; Ruby, Brennan et al. 2002; Hattar, Lucas et al. 2003). Putting all this evidence together, it is safe to conclude that cones, rods and ipRGCs all participate in non-image-forming irradiance detection for photoentrainment, with ipRGCs being the primary afferent signal relay. Each known photoreceptor contributes a different quality of information about the dynamic environmental lighting. Cones encode sudden increases in light intensity, leaving rods and ipRGC to act at low and high intensity levels respectively, allowing the circadian clock to respond to gradual changes in environmental lighting over a wide range of irradiance levels.

1.5.1.2 Pupillary light reflex

Intrinsically photosensitive retinal ganglion cells also project to the olivary pretectal nucleus (OPN), forming the afferent limb of the pupillary light reflex (Hattar, Liao et al. 2002). This direct connection to the midbrain pupil movement centre by ipRGCs explains why genetically manipulated rodless+coneless mice (Lucas, Freedman et al. 2001) and blind human patients due to severe photoreceptor disease can still have a preserved pupil light reflex (PLR) to high irradiance stimuli. Subsequent work by Hatori (Hatori, Le et al. 2008) and colleagues showed that ablation of ipRGCs almost completely eliminates photoentrainment and PLR in mice, revealing the central role of ipRGCs in PLR pathway.

It has been known for many decades that PLR has transient as well as sustained steady- state components (Young and Kennish 1993). For example, a 10 s long bright light stimulus will induce a rapid pupil constriction until it reaches a minimum pupil size (usually taking just a few seconds). This rapid phase response is followed by a gradual pupil re-dilation (a phenomenon called “pupil escape”) to a steady state of tonic pupil constriction that continues until the end of the light illumination (Kawasaki and Kardon 2007). Under certain circumstances, the steady-

state pupil constriction can even persist beyond the offset of light stimulus, which is called the post-illumination pupil response (PIPR) (Young and Kennish 1993). Historically, due to lack of knowledge of melanopsin and ipRGCs, the origin of differential PLR components has been the subject of debate. Ever since the discovery of ipRGCs and their role in the PLR pathway, numerous investigators have made attempts to elucidate the relative contributions of rods, cones and melanopsin to PLR. A growing body of evidence (Gooley, Lu et al. 2001; Dacey, Liao et al. 2005; Gamlin, McDougal et al. 2007; McDougal and Gamlin 2010) suggests that under photopic conditions, melanopsin is primarily responsible for maintaining the tonic pupil constriction; rods and cones contribute mostly to the transient phase of PLR for scotopic and photopic conditions respectively, the cone-driven pupil constriction adapts to light considerably and contributes little after prolonged exposure, while the rod-driven response adapts less and contributes significantly to the maintenance of tonic pupil size under steady illumination when the irradiance level is below the threshold of melanopsin (McDougal and Gamlin 2010). Study involving pharmaceutical blockage of rods/cones and pupillometry recording demonstrated that the PIPR after the offset of light stimulus is driven by melanopsin-mediated phototransduction of ipRGCs in both macaques and humans(Gamlin, McDougal et al. 2007). Interestingly, the model of the photoreceptors’ influence on PLR is very similar to that of circadian photoentrainment described earlier: cones encode the dynamic increase in irradiance and cause strong rapid pupil constriction, then they adapt to light causing a pupil escape, while rods and melanopsin provide long term steady regulation of pupil size at low and high irradiance level respectively.

Besides sustained pupil constriction under constant illumination and PIPR, melanopsin also makes a significant contribution to the maximum constriction of the rapid phase pupil response. In a wild-type mouse stimulated with flashing light of increasing intensity, the pupil reaches maximum constriction over about 6 log units of intensity. In melanopsin knock-out animals, however, pupil reacts normally until it reaches 80% of maximum constriction, further increases in intensity do not make the pupil constrict more (Lucas, Hattar et al. 2003; Panda, Provencio et al. 2003; Semo, Peirson et al. 2003; Barnard, Appleford et al. 2004). On the other hand, in mice lacking rods and cones, the threshold intensity of PLR is much higher than that of wild-type, but the pupil can constrict fully to bright light (Lucas, Hattar et al. 2003). These

findings indicate that melanopsin’s contribution to pupil constriction only starts at relatively high intensity levels, but it is required for the pupil to fully constrict.

1.5.1.3 Light suppression of activity

Mice and many other nocturnal rodents show reduced levels of activity and locomotion under bright light illumination, a phenomenon called “negative masking” (Mrosovsky 1999). Mice lacking melanopsin show reduced negative masking compared to wild-type, although the masking effect starts at normal intensity thresholds, but it fades away after prolonged illumination, as animals gradually adapt to the masking and resume normal activity, indicating that negative masking is not sustainable without the contribution from melanopsin (Mrosovsky and Hattar 2003). In contrast, in mice lacking rods/cones, negative masking shows an action spectrum similar to that of ipRGCs, appearing at a higher intensity threshold, and reaching the same end-point as wild-type in response to bright light (Thompson, Blaner et al. 2001).

In summary, rods, cones and ipRGCs all participate in mediating negative masking in nocturnal animals. The melanopsin-driven photo response is mandatory for maximal and sustained negative masking.

1.5.1.4 Light suppression of pineal melatonin

Melatonin is a neuro-hormone regulating a wide range of physiological responses. In mammals, circulating melatonin is predominantly released by the pineal gland of the brain. The release of melatonin is regulated by ambient light. In both nocturnal and diurnal animals, circulating melatonin begins to rise after sunset, reaching its peak during the night. Bright light exposure as short as several minutes can suppress pineal melatonin production. In humans, the spectral sensitivity of this response is distinct from those of rod/cone photoreceptors, and is rather consistent with that of the melanopsin photo response, which has peak absorption at 480 nm. Under constant illumination with melanopsin-activating blue light, the suppression of pineal melatonin last for hours, while melanopsin-silent red light only induced transient suppression (Thapan, Arendt et al. 2001; Lockley, Brainard et al. 2003). Mice with outer retinal degeneration are visually blind but light suppression of pineal melatonin appears to be normal (Lucas,

Freedman et al. 2001). These findings collectively suggest a major contribution from melanopsin in mediating light suppression of pineal melatonin synthesis.

1.5.1.5 Light regulation of sleep

Light has a profound effect on sleeping in both nocturnal and diurnal animals. In wild- type mice under a 12 hr:12 hr light:dark cycle, a pulse of light illumination in the dark period induces sleeping, while a period of dark in the light phase induces awakening. These effects are diminished in melanopsin knock-out mice. The melanopsin knock-out mice also show perturbations in sleep, and sleep approximately 1 hour less than wild-type (Tsai, Hannibal et al. 2009) , suggesting a sleep-modulating role for melanopsin. These patterns of sleep photo- regulation may apply to diurnal animals like humans in an opposite manner. It has been suggested that a “sign inversion” mechanism exists in hypothalamus such that light enhances wakefulness while dark promotes sleep (Brown and Piggins 2007).

1.5.2 Melanopsin and ipRGC functions in image-forming vision

It was conventionally believed that melanopsin and ipRGCs play little if any role in image-forming vision. Following the discovery of this novel photoreceptor, the concept of “non- image-forming vision” was introduced to summarize the major functions of this cell class. However, the boundaries between “non-image forming vision” and “image forming vision” have never been clear-cut.

As discussed earlier in this thesis, ipRGCs are morphologically and functionally diverse. Several subtypes of ipRGCs project directly and heavily to LGN and SC, brain areas that provide important relays for conventional image-forming vision (Ecker, Dumitrescu et al. 2010). This is further supported by electrophysiological evidence that melanopsin photoperception drives the responses in a large proportion of neurons in the mouse dLGN. A direct LGN projection of ipRGC encoding color opponency and light irradiance had also been observed in non-human primates (Dacey, Liao et al. 2004; Dacey, Liao et al. 2005). It was initially suggested that the melanopsin/ipRGC system may play a modulating role in pattern vision (Kawasaki and Kardon 2007), however recent evidence indicates that the contribution of melanopsin and ipRGC in

pattern vision may be more direct and more important than initially thought. Ecker and colleagues showed that mice lacking rod/cone phototransduction, and thus entirely dependent on melanopsin for light detection, have profoundly reduced but still measurable visual acuity, and can perform tasks that require pattern vision (Ecker, Dumitrescu et al. 2010). Schmidt and co- workers (Schmidt, Alam et al. 2014) employed an ultra-sensitive melanopsin assay to show that the well-known alpha ON ganglion cells capable of contrast detection contain low levels of melanopsin, and have intrinsic photosensitivity. These alpha ON ganglion cells were further identified as M4 ipRGCs that had been described previously (Schmidt, Alam et al. 2014). Their experiment further showed that melanopsin knock-out mice have reduced contrast sensitivity compared to wild-type, indicating melanopsin’s influence in contrast detection. Interestingly, intrinsically photosensitive alpha ON ganglion cells are also found in non-mammalian vertebrates, this finding can be viewed as evidence of evolutionary conservation (Rajaraman 2012).

In summary, exciting new evidence has gradually revealed a concept of “melanopsin- based vision”, and raised many intriguing questions: What exactly does melanopsin contribute to vision? Can patients with outer retinal degeneration see with the melanopsin pathways alone? How are ipRGCs affected in disease? Investigations are underway to address these questions.

1.6 Chromatic Pupillometry: in vivo assessment of melanopsin- driven ipRGC photoactivity

Among many photo responses and bio-effects mediated by melanopsin and ipRGCs, the pupillary light reflex is the most fast-acting and most observable one. PLR thus provide an easily accessible assessment of the ipRGC-mediated non-visual photoperception pathway. However, the ipRGC-mediated afferent signal for PLR is a mixture of photo responses from rods, cones and melanopsin (see discussion in 1.5.1.2), and so the real challenge is to isolate melanopsin- driven photoactivity from the influence of rods and cones. The spectral sensitivity of intrinsic melanopsin-mediated phototransduction is relatively narrow with a peak at around 480 nm (short wavelength light) (Brainard, Hanifin et al. 2001; Dacey, Liao et al. 2005). The intrinsic melanopsin-driven photo transduction has lower sensitivity compared with rod-mediated and cone-mediated photo transduction. However, once activated, melanopsin drives a sustained ipRGC firing that lasts well beyond the offset of light stimulus. Given its central role in mediating the pupillary reflex, the ipRGCs’ unique electrophysiological property reconciles very well with the long-observed phenomenon that pupil is able to tonically maintain constriction under constant light exposure and stay constricted after the cessation of a short wavelength high intensity light stimulus (Young and Kennish 1993). Gamlin and colleagues were the first team to provide in vivo evidence that the sustained component of the pupillary light response is indeed primarily driven by melanopsin (Gamlin, McDougal et al. 2007). They reported that after pharmacological blockade of rods and cones’ signaling in monkey eyes, continuous light induced sustained pupil constriction that slowly decayed after the offset of light. The pupil response’s spectral sensitivity, slow kinetics and irradiance coding properties matched the known electrophysiological features of ipRGCs. They further reported pupil responses to chromatic light stimuli in normal human eyes. When stimulated with luminance-matched red and blue light for 10 seconds, rapid pupil constriction was induced by both red and blue stimuli, but the maximum amplitude of pupil constriction was greater and more sustained under continuous blue light exposure. The authors suggested that the pupil response to red light was presumably driven by cones (mediated by ipRGCs), while the greater and more sustained pupil constriction to blue light compared to red light represented the melanopsin-driven intrinsic photoactivity of ipRGCs

superimposed on cone-driven extrinsic activity. Interestingly, the persistent pupil constriction after the offset of light, namely the post-illumination pupil response (PIPR), was observed after blue light but not subsequent to red light. The PIPR was largely unchanged after blockage of rods and cones. The authors therefore suggested using PIPR as an index of the melanopsin- driven ipRGC contribution to PLR.

Gamlin and colleague’s experiment (Gamlin, McDougal et al. 2007) laid the foundation of a novel technique called “chromatic pupillometry”, which utilizes light stimuli of different wavelengths, intensities and durations to selectively assess the contributions of the rod, cone and melanopsin responses to PLR, based on their differential spectral sensitivities and kinetics of photo responses. This technique’s ability to assess the newly discovered melanopsin-driven intrinsic photoactivity of ipRGC was particularly attractive to both vision science researchers and ophthalmologists.

In the early years of chromatic pupillometry development, Kardon and coworkers (Kardon, Anderson et al. 2009) described using a Ganzfeld bowl positioned 75 mm away from the subject to present 100 cd/m2 red and blue central field chromatic stimulation of 60° × 90° viewing angle. They took the difference in pupil size at the end of 13 s of red and blue light exposure as their index of the melanopsin response. This protocol, however, did not induce substantial PIPR. It is questionable whether this protocol stimulates melanopsin adequately. Park and coworkers (Park, Moura et al. 2011) used the same chromatic pupillometry system and refined the protocol. They provided evidence in favor of using dim blue stimuli (0.01 cd/m2) to induce the rod-driven pupil response, bright red stimuli (398 cd/m2) to induce the cone-driven pupil response, and bright blue stimuli (398 cd/m2) of 1 s duration to induce a sustained PIPR as an index of the melanopsin-driven intrinsic activity of the ipRGCs. This apparatus and protocol have since been adopted by a few other groups (Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012; Münch, Léon et al. 2012). Several other research teams have devised their own chromatic pupillometry apparatus and protocols to measure the melanopsin photo response in both healthy and diseased populations (Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Herbst, Sander et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Herbst,

Sander et al. 2012; Herbst, Sander et al. 2013; Roecklein, Wong et al. 2013; Nissen, Sander et al. 2014). The parameters of chromatic pupillometry methodologies in the literature are summarized in Table 2.

Stimulation

Author(year) Color (wavelength) Size Intensity Duration Main Index of Melanopsin Activity Gamlin (2007) Multiple steps, from Blue Central 36° 9-15 log 10 s PIPR during 15-30 s (492 nm) to Red (613 nm) quanta/cm2/s after offset of stimuli

Kardon (2009,2011) Blue (467±17 nm) Central 60° × 90 ° (V 1, 10, 100 cd/m2 13 s per intensity Sustained pupil Red (640±10 nm) × H) (continuous steps of step, 39 s in total constriction at 13 s increment) after constant illumination Kankipati (2010) Blue (470 nm) Central 60° 13 log quanta/cm2/s 10 s PIPR during 10-40 s Red (623 nm) after offset of stimuli Fiegl (2011) Blue (488 nm) Central 7.15° 14.2 log 10 s exponential Red (610 nm) quanta/cm2/s function of PIPR Herbst (2011, 2012, Blue (470 nm) Not clearly 300 cd/m2 20 s PIPR during 10-30 s 2013), Nissen (2012, Red (660 nm) described, after offset of stimuli 2014) presumably full- field or near full- field Park (2011), Kawasaki Same as Kardon (2009,2011) 398 cd/m2 1 s PIPR at 6 s after offset (2012) of stimuli Munch (2012) Same as Kardon (2009,2011) 14 log quanta/cm2/s 1 s and 30 s PIPR at 6 s after offset of stimuli Roeckleina (2013) Blue (467.7 nm) Central 29° 13.7 log 30 s PIPR during 10-40 s Red (632.9 nm) quanta/cm2/s after offset of stimuli

Table 2. Summary of Post-illumination Pupil Response (PIPR) Testing Conditions Published in Literature

Among disease populations tested with chromatic pupillometry, patients with outer retinal degeneration such as retinitis pigmentosa, Leber’s congenital amaurosis (LCA) and achromatopsia show a reduced rod or cone-driven pupil response, but their melanopsin-driven PIPR can be well preserved (Kardon, Anderson et al. 2009; Kardon, Anderson et al. 2011; Park, Moura et al. 2011; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012), while patients with inner retinal disease, such as glaucoma and optic neuropathy, may show impaired PIPR (Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Herbst, Sander et al. 2013; Nissen, Sander et al. 2014). It is noteworthy that impaired PIPR was only observed in advanced glaucoma but not in early glaucoma, indicating that melanopsin-containing ipRGCs are initially

relatively resilient to glaucomatous damage (Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011). Roecklein and coworkers (Roecklein, Wong et al. 2013) reported that PIPR is reduced in patients with seasonal affective disorder (SAD), a disease closely related to sun light exposure, indicating an important role for melanopsin phototransduction in the genesis of SAD.

Chromatic pupillometry measurement of PIPR has also been used in basic science research as an index of in vivo melanopsin activity. Roecklein and colleagues showed that certain single nucleotide polymorphism (SNP) genotypes of melanopsin are associated with reduced PIPR. This suggests that the genetic variations of melanopsin may have a functional impact on melanopsin phototransduction (Roecklein, Wong et al. 2013). Mure and coworkers reported that long duration pre-exposure to red light enhances the sustained pupil response and PIPR to blue light in visually healthy human subjects. They claimed that the enhancing effect of red light on PIPR was in vivo evidence of melanopsin bistability (Mure, Cornut et al. 2009).

Collectively, these data strongly suggest that chromatic pupillometry measurement of PIPR has important applications for clinical and basic science research as an in vivo index of melanopsin-driven intrinsic ipRGC photoactivity, which represents a novel aspect of inner retinal function. However, the PIPR testing conditions reported in literature are highly variable: most of the reported protocols employ a bright blue light stimulus of 10 s duration or longer presented to dark-adapted eyes. This is obviously not patient-friendly for clinical use. Efforts to standardize and optimize the methodology used for PIPR testing may further expand the application of the chromatic pupillometry technique in basic science and clinical investigation.

1.7 Technical development: full-field chromatic pupillometry assessment of the melanopsin-driven post-illumination pupil response

1.7.1 Rationale for developing full-field chromatic pupillometry

Using chromatic pupillometry to measure PIPR holds promise as a new diagnostic and therapeutic outcome measurement tool to assess the melanopsin-driven intrinsic photoactivity of ipRGCs, which represents a whole new aspect of inner retinal function; however, the testing conditions for PIPR have not been fully optimized. All existing studies (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Kankipati, Girkin et al. 2010; Kankipati, Girkin et al. 2011; Kardon, Anderson et al. 2011; Nissen, Sander et al. 2011; Kawasaki, Crippa et al. 2012; Münch, Léon et al. 2012) have used central-field blue-light stimulation of high intensity and long duration to induce PIPR. The visual angle of the central-field stimuli ranged from 7.5° to 60°×90°, which is not ideal because it does not stimulate all ipRGCs, which are distributed across the entire retina (Hattar, Liao et al. 2002). Given the photon-counting property of ipRGCs to measure the retinal irradiance (Dacey, Liao et al. 2005), inadequate stimulation of the peripheral retina may be a reason why all existing studies using central-field stimuli required long duration stimulation (10-30 s) to induce a recordable PIPR. Although one protocol (Park, Moura et al. 2011) reduced the duration of stimulation to 1 s, in our experience, most participants still found it difficult to tolerate the stimulus especially at higher intensities, and they had difficulties keeping their eyes open. While stimulus intensity and duration can be adjusted easily in most of the previously described chromatic pupillometry systems, adjusting the stimulus area/location has not been investigated. Such a methodology may potentially have important clinical implications in diseases with localized inner retina damage such as glaucoma, branch retina vein occlusion, maculopathy, and anterior ischemic optic neuropathy, among others.

Before I began my graduate study, I led a project as a research fellow in my current supervisor’s lab to refine the existing PIPR testing methodologies by devising a reliable and subject-friendly chromatic pupillometry protocol, where the intensity, duration and stimulus area/location can be adjusted for investigations of clinical and basic science questions related to

melanopsin and ipRGCs. We sought to develop a full-field chromatic pupillometry apparatus, and test this apparatus with a series of experiments on visually normal human subjects to understand the characteristics of PIPR under full-field stimulation, and to address the hypothesis that full-field stimulation can induce PIPR more efficiently than a previously described (Kardon, Anderson et al. 2009) 60° × 90° central field stimulation. We also compared the single time point post-illumination pupil measurement technique (Park, Moura et al. 2011) with prolonged measurement as an index of PIPR. (Lei, Goltz et al. 2014)

1.7.2 Apparatus and experiment protocols

We modified a Ganzfeld screen originally designed for full-field ERG testing to present monochromatic light stimuli (Blue: 467±17 nm, Red: 640±10 nm), and an infrared video eye tracker (ViewPoint EyeTracker ® system, Arrington Research, Scottsdale, AZ) to monitor the pupil response. This chromatic pupillometry apparatus was identical to that described in Kardon and co-workers (Kardon, Anderson et al. 2009, Kardon, Anderson et al. 2011) and Park and colleagues (Park, Moura et al. 2011) except for some modifications to our eye tracker. The scene camera was removed to allow the participants to place their foreheads against the edge of the Ganzfeld screen for full-field stimulation. To increase the quality of the eye tracking, an extra miniature near-infrared illumination diode was added in front of each camera (peak wavelength 940 nm, radiant intensity 40 mW, 18° angle of emitted light; Fairchild Semiconductor, San Jose, CA) to provide additional infrared illumination to each eye. The study consisted of two experiments. Experiment 1 (intensity trials) compared the PIPR in response to full-field stimulation of increasing intensities from 0.1 to 400 cd/m2 at a fixed 1 s duration. For comparison, PIPR was also induced using a 60°×90° central-field blue stimulus of 400 cd/m2 as used by Park and co-workers. (Park, Moura et al. 2011). Experiment 2 (duration trials) was conducted on two separate days testing two specific flash intensities (100 cd/m2 red and blue stimuli on one day and 400 cd/m2 red and blue stimuli on the second day) with ten different durations from 4-1000 ms. The PIPR at various durations was compared for the two specific intensities. All experiments were conducted during the day between 8 am and 2 pm.

1.7.3 Full-field vs. central field PIPR in intensity and duration trials

Ten visually-normal subjects were tested in each experiment. The results of experiment 1 (Figure 1) show that at all tested intensity steps, a melanopsin-silent red stimulus of 1 s duration induced a rapid constriction of the pupil, which rapidly re-dilated to reach baseline about 10 s after stimulus offset. In contrast, the blue stimulus of photopically-matched luminance induced greater and more sustained pupil constriction overall. When the intensity of the stimulus was low (0.1-1 cd/m2), however, the difference in the PIPR tracings between red and blue stimuli was minimal, indicating that melanopsin is not adequately stimulated in this intensity range. As intensity increased from 3.16 to 400 cd/m2, the PIPR were increasingly more sustained. When central-field stimulation was compared with full-field stimulation, the PIPR induced by the 400 cd/m2 blue central-field stimuli (Figure 1l) was less sustained than the responses induced by 200 and 400 cd/m2 blue full-field stimuli (Figure 1j and 1k). The response induced by the 400 cd/m2 blue central-field stimulation was comparable to those induced by the 100 and150 cd/m2 blue full-field stimuli (Figure 2).

Figure 1. Mean normalized PIPR tracings from 1 s stimuli of varying intensity, Data from 10 visually-normal participants. Panels a-k show PIPRs to full-field red and blue stimuli at 11 intensity levels from 0.1 to 400 cd/m2. Panel l shows PIPRs to 400 cd/m2 central-field stimulation.

Figure 2. Comparison of PIPR induced using 400 cd/m2 central-field stimuli (dashed line) vs 100-400 cd/m2 full-field stimuli. The response induced by the 400 cd/m2 blue central-field stimulation (dashed line) was comparable to those induced by the 100 and 150 cd/m2 blue full-field stimuli (solid lines).

Using averaged pupil size from a 20 s interval from 10-30 s after the offset of stimulus as 2 PIPR index (PIPR(10-30s)), the greatest PIPR was induced with 400 cd/m blue full-field stimulus, which was significantly greater than the PIPR induced by the 400 cd/m2 blue central-field stimulus as shown in Figure 3. In contrast, during red stimulation (full and central-field), the induced PIPR was small and did not differ significantly across all the stimulus intensities.

After the intensity trials, we conducted a duration trial at 100 cd/m2 and 400 cd/m2 intensity levels. With 400 cd/m2 blue stimuli, pupil constriction was increasingly sustained when duration increased from 4 ms to 400 ms; however, further increases in duration (greater than 400 ms) did not result in a more sustained pupil response, as shown in Figure 4. This is verified quantitatively by the mean PIPR(10-30 s), which increased gradually until the duration reached the 200 ms step, with last 5 duration steps (200 ms, 400 ms, 600 ms, 800 ms and 1000 ms) not differing significantly among one another. The PIPR (10-30 s) induced by the last 4 duration steps (400 ms, 600 ms, 800 ms and 1000 ms) of 400 cd/m2 full-field blue stimuli were significantly greater than the PIPR (10-30 s) induced by 1000 ms central field blue stimulus. In contrast, with 2 100 cd/m blue stimulation, PIPR(10-30 s) increased monotonically with increased duration, indicating that 100 cd/m2 blue stimuli were not strong enough to induce saturated PIPR for durations between 4 to 1000 ms. The 100 and 400 cd/m2 red stimuli induced no or very little

PIPR(10-30 s) that did not change significantly with increased duration (Figure 5).

Figure 3. Mean normalized pupil size from 10 to 30 s post-stimulation offset (PIPR(10-30 s)) and pupil size at 6 s post-stimulation offset (PIPR(6 s)). The highest PIPR(10-30 s) was induced by 1 s, 400 cd/m2 blue full-field stimuli, which was significantly greater than the mean 2 PIPR(10-30 s) induced by the 1 s, 400 cd/m blue central-field stimulus (far right column indicated by the arrow). Bars represent ± SE.

Figure 4. Mean PIPR to 100 cd/m2 and 400 cd/m2 full-field stimulation of varying duration from 10 visually-normal observers.

Figure 5. PIPR(10-30 s) as a function of stimulus duration. Error bars represent ± 1 SE. Full- 2 field blue stimulation at 400 cd/m for 400-1000 ms induced significantly greater PIPR(10-30s) than the central-field 400 cd/m2 stimulation for 1000 ms.

1.7.4 Implications: Relation between PIPR and stimulus intensity, duration and area

One of the major findings of this body of work is that a strong PIPR can be induced using full-field stimulation of lower intensity and shorter duration than existing central-field protocols (Park, Moura et al. 2011), confirming our prediction that full-field stimulation is more effective in inducing PIPR. We used a modified eye tracker configuration to allow participants to be positioned optimally for full-field stimulation. Although subjects were closer to the ColorDome

Ganzfeld screen when the eyes were simulated with full-field stimuli vs. central-field stimuli, the difference in viewing distance between these conditions was only 75 mm. The physical attenuation of light intensity over this distance is negligible, thus the stimulus light projected onto the central part of the retina can be considered equally intense under both conditions. The difference in PIPR induced by full-field vs. central-field conditions is most likely attributable to additional retinal recruitment.

Our preliminary work is the first to demonstrate that a significant PIPR recordable 10-30 seconds after the offset of the stimulus can be induced in vivo with a strong blue flash of only a few hundred milliseconds. Specifically, we found that PIPR from 10-30 seconds after stimulus cessation became saturated at around 200-400 ms with 400 cd/m2 blue full-field stimuli. Because 400 cd/m2 is the maximum intensity our apparatus can generate, it remains unknown whether similarly strong PIPR can be induced with even shorter exposure if higher light intensities were used, although higher light intensity will be more uncomfortable for the subjects, particularly for photophobic patients.

2 In our experiment, detectable PIPR(10-30 s) emerged at around 3.16 cd/m (approximately equivalent to 12 log quanta/cm2/s [Pianta and Kalloniatis 2000; Park, Moura et al. 2011] ) given a mean resting pupil diameter of 6.4 mm in our subjects), and increased steadily with increasing stimulus intensity up to 400 cd/m2 ( 14.3 log quanta/cm2/s). The action spectrum of these responses reconcile well with the observed dynamic range of melanopsin for 470 nm light in vitro (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005), and matches the response range of PIPR induced with 10 s blue light stimulation in human subjects (approximately 12-15 log quanta/cm2/s) (Gamlin, McDougal et al. 2007). In addition, the PIPR recorded in this pilot work is selectively sensitive to short wavelength light; so both the sensitivity and kinetics are consistent with the known features of the melanopsin-driven photoresponse in vitro (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005). These findings indicate that the PIPR induced with our short duration testing conditions is most likely mediated by melanopsin-driven intrinsic ipRGC activity.

Previous studies (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Kardon, Anderson et al. 2011; Münch, Léon et al. 2012) using central-field bright blue light stimuli of 10- 20 s duration induced a response profile characterized by a rapid pupil constriction upon stimulus onset, followed by a sustained component of pupil constriction under constant illumination, then a rapid dilation at stimulus offset, followed by a sustained post-illumination pupil constriction of lesser magnitude. It is noteworthy that the characteristics of the pupillary responses to a short bright blue flash used in the current study are fundamentally different: the response is characterized by a rapid constriction that is sustained well after the offset, and lacking a rapid dilation upon stimulus offset. The rapid dilation upon stimulus offset is likely due to the photoreceptors’ (rods and cones) OFF pathway response being superimposed on the intrinsic melanopsin-driven activity. This is supported by Gamlin’s study (Gamlin, McDougal et al. 2007) where the OFF effect was recorded by both pupillometry and intracellular recording from ipRGCs. When the synaptic input from the image-forming photoreceptors was blocked pharmacologically, the OFF effect was absent (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Gamlin, McDougal et al. 2007). We hypothesize that under our testing conditions, the spikes of intrinsic melanopsin-driven ipRGC firing occur after the offset of the blue flash, so that the intrinsic firing can carry on without being affected by the activity of the OFF pathway. If this hypothesis is true, PIPR induced by a short flash of blue light would represent a “pure” intrinsic ipRGC activity that is not being “contaminated” by the image-forming photoreceptor OFF effect.

The data in these experiments also show that PIPR is a function of stimulus intensity, duration and retinal area: higher intensity, longer duration and a larger stimulated retinal area all contribute to a larger and more sustained PIPR. This phenomenon was expected considering the precise photon-counting ability of ipRGCs that was first described in Dacey and colleagues’ study (Dacey, Liao et al. 2005), where a highly linear relation between the total number of cell firing spikes and intensity (in log quanta/cm2/s) was demonstrated by intracellular recording. It has been suggested that ipRGCs’ long, sparsely branching dendrites form a largely overlapping network that cover the entire retina as an irradiance detector (Dacey, Liao et al. 2005). The relation between PIPR and stimulus intensity, duration and retinal area delineated in our study

supports the idea that the ipRGC network codes intensity-dependent spatial and temporal summation of retinal irradiance.

The ability of our protocol to induce a large PIPR with full-field stimulation of only 200- 400 ms is of considerable clinical significance. First, a short flash greatly decreases participants’ discomfort compared to long duration bright light stimulation in darkness. Second, a short flash enhances the consistency of the amount light exposure, since it is practically impossible for participants to keep their eyes open throughout a 10-20 s duration of bright light exposure. Eye blinking and squinting causes inconsistent light exposure, and both can be largely avoided with our protocol using a 200 to 400 ms short “flash”. The pupil constriction during light stimulation is also minimized with a short flash, so the retinal irradiance during exposure is more consistent. The shorter the exposure, the less likely it is that the pupillary response will be affected by other factors such as attention, accommodation and fatigue (Kahneman and Beatty 1966; Kahneman 1973). Third, compared to central-field stimulation, full-field stimulation provides stimulation to the entire retina, and may be useful in assessing ipRGC activity as an input signal to other biological functions such as circadian rhythm regulation, or as an index of remaining inner retinal function in end-stage diseases. The full-field saturating response may not be suitable for all clinical or research applications, however, with an updated understanding of the relation between PIPR and stimulus intensity, duration and area, investigators may also be able to tailor their PIPR testing paradigm to target a particular basic research or clinical question.

In this pilot study, besides PIPR(10-30 s), we also evaluated another measurement interval as an indicator of the post illumination pupil response: PIPR(6 s) which is defined as pupil size at 6 s after the offset of light stimulus as previously described and recommended by Park and coworkers (Park, Moura et al. 2011). The results showed that in both experiments, PIPR(6 s) responses to blue stimuli of varying intensity and duration are less linear and plateaued much earlier than those of PIPR(10-30 s). In contrast, red light stimulation, which primarily induces a cone response, generated a slow increase in PIPR(6 s) with increasing intensity and duration. We observed that the pupil response to red light stimulation subsided to a plateau near baseline levels after 10 s post illumination. In addition, it has been well established that the response latency of

the melanopsin-driven ipRGC activity is typically several hundred milliseconds to several seconds (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Johnson, Wu et al. 2010) (negatively associated with stimulus intensity), so at 6 seconds after the offset of light, the melanopsin may not be fully activated yet, especially when the stimulus intensity is low. These results suggest that at 6 s post illumination, there is still a significant proportion of cone-driven responses contributing to the pupillary light response. In contrast, the PIPR(10-30 s) to red stimulation was minimal at all stimulation levels. Furthermore, as discussed earlier, PIPR(10-30 s) induced in our experiment closely matches the action range of melanopsin. Therefore we conclude that PIPR(10-

30 s) is the more appropriate index of the ipRGC-driven post illumination pupil responses tested.

In summary, compared to the existing protocols that used central-field stimulation, full- field stimulation induces a large PIPR with lower stimulus intensities and dramatically shorter durations, indicating that PIPR represents the information of an intensity-dependent spatial and temporal summation of retinal irradiance coded by melanopsin-driven ipRGC activity. This updated information about the relation between PIPR and stimulus intensity, duration and area will allow us to tailor our PIPR testing paradigm to target a particular investigation question, and greatly facilitate the development of a convenient and comfortable technique to assess ipRGC function for emerging basic science and clinical investigations.

Chapter 2

2 Aims and Hypothesis 2.1 Introduction

This thesis consists of two separate studies. The first study (Chapter 3) is a methodological study on the development of a novel chromatic pupillometry protocol to induce, record and compare melanopsin-driven PIPR by stimulating different areas/locations of retina. The second study (Chapter 4) implements the new PIPR testing protocol to investigate the in vivo mechanism of melanopsin bistability.

2.2 Hemifield, Central-Field and Full-Field Chromatic Pupillometry for Assessing the Melanopsin-driven Post- illumination Pupil Response: A Methodological Study

Chromatic pupillometry testing of PIPR is an objective tool for assessing melanopsin- containing retinal ganglion cell function (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Kardon, Anderson et al. 2011; Park, Moura et al. 2011), however, an optimal PIPR testing protocol has not been developed yet. While stimulus intensity and duration can be adjusted easily in most of the previously described chromatic pupillometry systems (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Mure, Cornut et al. 2009; Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Roecklein, Wong et al. 2013; Nissen, Sander et al. 2014), adjusting the stimulus area/location has not been investigated. Such a methodology may potentially have important clinical implications. Based on a full-field, high intensity, short duration PIPR testing protocol determined previously, (Lei, Goltz et al. 2014) this study aims to further develop a novel chromatic pupillometry device, where a monochromatic light stimulus can be directed to specific areas/regions of the retina to induce melanopsin-driven PIPR. This device will be evaluated by testing upper and lower hemifield, central-field, and full-field PIPR in a group of visually normal subjects. Given that melanopsin-driven ipRGC intrinsic photoactivity is responsible for detecting light irradiance, (Berson 2003; Dacey, Liao et al. 2005; Fu, Liao et al. 2005; Wong, Dunn et al. 2005; Do, Kang et al. 2009) We hypothesize that hemifields, central field and full-field stimulation will induce increasingly larger PIPR, and that upper and lower hemifield PIPR will be symmetric in normal subjects. The test-retest reliability of hemifields, central-field and full-fields PIPR will also be evaluated.

2.3 The effect of red light exposure on pre-existing PIPR: Implementing PIPR as an in-vivo index of melanopsin photoactivity in basic science research

Melanopsin is a vitamin A-based opsin employing 11-cis-retinal as its chromophore (Provencio, Jiang et al. 1998). After absorbing a photon, 11-cis-retinal is photoisomerized into all-trans-retinal, causing conformational changes in the melanopsin which lead to the formation of “meta-melanopsin” (the activated iso-form of melanopsin) and subsequently trigger phototransduction cascade in the melanopsin-containing ipRGCs (Walker, Brown et al. 2008). Subsequently, the all-trans-retinal chromophore of melanopsin has to be converted back to 11- cis-retinal in order to maintain ipRGC photosensitivity. Unlike rod and cone photoreceptors that rely on the visual cycle in RPE cells to regenerate 11-cis-retinal, melanopsin containing ipRGCs are located in the inner retina; it is still unclear how ipRGCs regenerate their melanopsin chromophore in vivo without easy access to the supply of 11-cis-retinal from RPE. A growing body of literature (Lucas 2006; Mure, Rieux et al. 2007; Rollag 2008; Mure, Cornut et al. 2009; Matsuyama, Yamashita et al. 2012; Sexton, Golczak et al. 2012) suggests that the activated all- trans-retinal could be photoisomerized back to 11-cis-retinal by absorbing photon energy at a different wavelength. This photo-driven switchable state of chromophore is call “opsin bistability”.

While there is consensus that melanopsin activation (forward reaction) is selectively sensitive to short wavelength light, with maximum absorption being at around 480 nm (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Gamlin, McDougal et al. 2007; Do, Kang et al. 2009; Mure, Cornut et al. 2009; Do and Yau 2010), conflicting evidence exists regarding the spectral sensitivity of melanopsin photo-regeneration (backward reaction) (Mure, Rieux et al. 2007; Mawad and Van Gelder 2008; Rollag 2008; Mure, Cornut et al. 2009; Matsuyama, Yamashita et al. 2012). Mure and co-workers (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009) provided in vivo evidence that pre-exposure to long-wavelength red light potentiates the photo response to short-wavelength light (as measured by spiking of SCN neurons, the pupillary light reflex and the negative masking effect), a feature of bistable opsin. By computing the spectral tuning of red light potentiation of the sustained pupil response to blue light, they reported that the putative

spectral sensitivity of meta-melanopsin is red-shifted, with maximal absorption at 587 nm (Mure, Cornut et al. 2009). However, a similar paradigm failed to induce potentiation in ipRGCs cells firing in an in vitro experiment (Mawad and Van Gelder 2008), which argues against this “blue forward, red backward” hypothesis. Furthermore, a recent photochemical study on purified rat melanopsin protein showed that melanopsin and meta-melanopsin have essentially overlapping spectral sensitivity, with both peaking in the range of blue light (467 vs. 476 nm), (Matsuyama, Yamashita et al. 2012) but so far there is no in vivo evidence to support this “blue-forward, blue- backward” theory.

The activation of melanopsin drives a persistent firing of ipRGCs resulting in a prolonged PIPR (Gamlin, McDougal et al. 2007), which is an in vivo index of melanopsin photoactivity (Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Herbst, Sander et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Park, Moura et al. 2011; Herbst, Sander et al. 2012; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012; Münch, Léon et al. 2012; Herbst, Sander et al. 2013; Roecklein, Wong et al. 2013; Lei, Goltz et al. 2014; Nissen, Sander et al. 2014; Lei, Goltz et al. 2015). In order to provide in vivo evidence for the mechanism of melanopsin chromophore photo-regeneration, this study employs a blue-red “double flash” experimental chromatic pupillometry paradigm to investigate the effect of red light exposure on pre-existing PIPR induced by blue light. Based on the assumption that subsequent absorption of light energy converts the activated melanopsin back to its resting state resulting in attenuated PIPR, it is predicted that if the “red-light-reversing” theory is true, red light exposure presented after a melanopsin-activating blue light stimulus will attenuate the blue- light-induced PIPR.

Chapter 3

3 Hemifield, Central-Field and Full-Field Chromatic Pupillometry for Assessing the Melanopsin-driven Post-illumination Pupil Response 3.1 Introduction

While measuring PIPR using chromatic pupillometry holds promise as an objective means of assessing melanopsin-containing retinal ganglion cell function independent of photoreceptors, (Kawasaki and Kardon 2007; Kardon, Anderson et al. 2009; Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Kardon, Anderson et al. 2011; Park, Moura et al. 2011; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012; Roecklein, Wong et al. 2013; Lei, Goltz et al. 2014; Nissen, Sander et al. 2014), consensus has not yet been reached on a standardized clinical protocol. Originally, PIPR was induced with prolonged exposure to high intensity short wavelength light (duration ranging from a few seconds to a few minutes) (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Mure, Cornut et al. 2009; Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Roecklein, Wong et al. 2013; Nissen, Sander et al. 2014), which is not practical as a clinical test. Recently, we described a chromatic pupillometry apparatus and testing conditions that can induce a large PIPR with a full-field blue light flash of only a few hundred milliseconds, and we found that PIPR is a function of stimulus intensity, duration and retinal area stimulated. (Lei, Goltz et al. 2014) While stimulus intensity and duration can be adjusted easily in most of the previously described chromatic pupillometry systems, adjusting the stimulus area/location has not been investigated. Such a methodology may potentially have important clinical implication in diseases with localized inner retina damage such as glaucoma, branch retina vein occlusion, maculopathy, and anterior ischemic optic neuropathy, among others. The first goal of this study was to develop a PIPR testing method in which stimulation area/location can be adjusted to present hemifield, central-field and full-field stimulation. It is predicted that hemifields, central field and full-field stimulation will induce increasingly larger

PIPR, and that upper and lower hemifield PIPR will be symmetric in normal subjects. The second goal was to validate this PIPR testing method by evaluating the test-retest reliability of hemifield, central-field and full-field PIPR in visually normal subjects.

3.2 Methods

3.2.1 Participants

Ten visually-normal subjects who had normal or corrected-to-normal vision (20/20 Snellen visual acuity or better) participated in the study (6 females, mean age 30 years, age range 19-56 years). All participants underwent a screening ophthalmology exam, which include visual acuity, contrast sensitivity, eye position and movement tests, colour vision assessment (HRR test, Richmond Products, NM, USA), intraocular pressure assessment using slit-lamp tonometry, slit- lamp assessment of iris structures, anterior chamber angle and non-dilated fundus exam. Inclusion criteria include: 1). Age from 17 to 70 years; 2). Corrected visual acuity of 20/20 or better; 3). Refractive error between +5.00D to -5.00D spherical equivalent; 4). No known visual abnormalities. Exclusion criteria include: 1. Intraocular surgery 2. Eye trauma 3. Optic neuropathy 4. Glaucoma 5. Shallow anterior chamber 6. Narrow angle 7. Uveitis 8. Diabetic neuropathy 9. Retinal detachment 10. Myopia greater than -5.00D 11. Hyperopia greater than +5.00D 12. Any other conditions that may compromise the integrity of the pupillary light reflex pathway upstream of the retina

The experiments were conducted monocularly, with the right eye being stimulated and recorded. The left eye was patched. The study was approved by the Research Ethics Board at The Hospital for Sick Children. All the procedures adhered to the guidelines of the Declaration of Helsinki. Informed consent was obtained from each participant.

3.2.2 Apparatus

The chromatic pupillometry system consisted of two components, a Ganzfeld screen and an infrared video-based spectacle frame-mounted eye tracker. The Ganzfeld screen (Espion V5 system with the ColorDome LED full-field stimulator; Diagnosys LLC, Lowell, MA) that was

originally designed for full-field ERG testing was used to present full-field light stimulation. This Ganzfeld screen can generate a wide range of flash intensities from 0.0001 to 400 cd/m2 (-4 to 2.6 log) for blue (467±17 nm) light and 0.0001 to >400 cd/m2 (-4 to >2.6 log) for red (640±10 nm) light. The binocular eye-tracking camera system used near-infrared (940 nm) illuminating diodes (Arrington Research, Scottsdale, AZ) to record the changes in pupil diameter at a sampling rate of 60 Hz. The chromatic pupillometry hardware was identical to that described in Kardon and co-workers (Kardon, Anderson et al. 2009; Kardon, Anderson et al. 2011) and Park and colleagues (Park, Moura et al. 2011) except for some modifications to our eye tracker: the scene camera was removed to allow the participants to place their forehead against the edge of the Ganzfeld screen for full-field stimulation. (Figure 6)

An additional shutter panel was custom-built to present central-field and hemifield stimuli. The shutter panel has a back plate with a round opening of 50 mm in diameter (the size of the opening can be changed by changing the back plate), which resulted in 30° central field stimulation when positioned 95 mm away from the subject’s eye. A removable rectangular

Figure 6. Full-field chromatic pupillometry system. (A) Author wearing the Arrington eye tracker with miniature infrared LEDs and cameras. (B) Pupil size is monitored with a video-based eye tracker at a sample rate of 60 Hz. (C, D) Full-field blue and red light stimulation presented with the Colordome Ganzfeld screen.

hemifield occluder can be inserted in the front of the shutter panel to block the upper or lower half of the 30° central field (Figure 7). In order to ensure the hemifield stimulation was aligned

Figure 7. The shutter panel apparatus. The shutter panel comprises 3 components: (1) main frame; (2) back plate with 50 mm diameter opening; and (3) hemifield occluder with 2 horizontally-aligned red LEDs on both the upper and lower edges. When hemifield stimulation is presented, the shutter panel is attached to the opening of Ganzfeld dome. The participant is instructed to align the center fixation LED in the back of the Ganzfeld dome in the middle between the 2 horizontally-aligned LEDs on the edge of the hemifield occluder. The illumination of the occluder LEDs is controlled by a toggle switch. For full- field stimulation, the shutter device is removed and the forehead of the subject rests against the edge of the Ganzfeld dome. accurately, two red mini LED lights (3 mm in diameter) were installed on the upper and lower edge of the hemifield occluder. For example, to stimulate the upper hemifield, the hemifield

occluder was positioned at the bottom of the panel to block the lower half of the central opening. Two red LED lights on the upper edge of the hemifield occluder flanked the central fixation red LED light in the Ganzfeld screen. The subject was instructed to align all three lights while maintaining fixation on the central one, thereby ensuring that only one half of the 30° central field was stimulated.

3.2.3 Experimental Conditions and Procedure

Prior to the experiment, subjects were exposed to an indoor laboratory environment with ambient lighting levels ranging from 80-400 lux for at least 2 hours. During the pupillometry recordings, participants were seated in a darkened room (0 lux) with their head rested on a chinrest. The Ganzfeld was positioned 95 mm away from the participant’s eyes (measured from the shutter panel) during the presentation of blue light (467±17 nm) for lower hemifield, upper hemifield and 30° central-field stimulation. The order of the three conditions was randomized. The shutter panel was then removed and the subject was repositioned closer to the Ganzfeld screen, with the forehead touching the upper edge of the Ganzfeld screen opening, then full-field red (640±10 nm) and blue stimulation was presented. Experimental trials were always initiated with 10 s of dim amber light (590 ± 7 nm, 0.3 cd/m2) pre-exposure (to ensure precisely 90 s of dark adaptation before each light stimulation), followed by 90 s of dark adaptation, which was then followed by red or blue light stimulation (400 cd/m2, 400 ms) in darkness. The pupillary response was recorded in real-time at 60 Hz, starting from 5 s prior to the onset of light stimulation until 35 s after its offset. The overhead room lighting (200 lux) was turned on afterwards and the participants were allowed to take a short break (30 s to 2 minutes) before starting another trial to prevent carry-over effects and fatigue. Each condition was repeated 3 times in experiment session 1. Within one month of the first session, a second session was carried out during which all subjects were tested with lower and upper hemifield stimulation only, with each condition repeated 3 times in randomized order. All experiments were conducted during the day between 8 am and 2 pm.

3.2.4 Data Analysis

Data from the eye tracker were analyzed offline using a custom-written script (MatLab; MathWorks Inc., Natick, MA). A median (window length of 0.5 s) and low-pass filter (fourth- order, zero-phase Butterworth) with a cut-off frequency of 5 Hz were applied to remove eye blink artifacts. The filtered data were inspected visually in a graphical user interface (GUI) to ensure data quality and detect artifacts. The filtered data were then normalized to the baseline pupil size calculated from the mean pupil size during a 5 s period before the onset of each stimulus (i.e., Normalized pupil size = Absolute pupil size / Baseline pupil size). Two parameters were measured: (1) PIPR—mean pupil size over a 20 s interval from 10-30 s after the offset of light stimulation. Our previous work found that the cone-driven pupil responses subsided within 10 s after the offset of light stimuli, so this measurement is expected to represent “pure” melanopsin activity. (2) Maximal Pupil Constriction (MPC)—the smallest pupil size following light stimulation. This parameter primarily represents the rapid phase pupil constriction driven by rods and cones, but may receive melanopsin influence under certain conditions (Lucas, Hattar et al. 2003; Panda, Provencio et al. 2003; McDougal and Gamlin 2010). For both parameters, a smaller value represents greater pupil constriction.

Statistical analyses were performed using SPSS 19.0 (IBM Corporation. Armonk, NY). Differences in mean PIPR and MPC were compared using separate one-way repeated measures ANOVAs across 5 different testing conditions: (1) lower hemifield (lower half of the central 30° field) stimulation using blue light (400ms, 400cd/m2); (2) upper hemifield (upper half of the central 30° field) stimulation using blue light (400ms, 400cd/m2); (3) central-field (both halves of the central 30° field) stimulation using blue light (400ms, 400cd/m2); (4) full-field stimulation using blue light (400ms, 400cd/m2); and (5) full-field stimulation using red light (400ms, 400cd/m2). Within each condition, the means of each repeated test were also compared using one-way repeated measures ANOVA. All post hoc pairwise comparisons were adjusted for multiple comparisons using the Bonferroni method. A p value of <0.05 was considered statistically significant.

The hemifield PIPR and MPC measurements of the two recording sessions (session 1: tests 1-3 vs. session 2: tests 4-6) were compared using paired sample t tests and Pearson correlation.

The intra-subject coefficient of variation (CV) for PIPR and MPC were calculated for each subject, then the median and range of CV from all 10 subjects were reported for each condition. The CV is analogous to the signal-to-noise ratio, and is defined in this context as the ratio of the standard deviation (SD) of the repeated measures of PIPR (or MPC) to the mean changes of pupil diameter: CV (PIPR) = SD/ (1- mean PIPR) and CV (MPC)= SD/ (1- mean MPC). The correlation between repeated tests was reported as the intra-class correlation coefficient (ICC) along with its 95% confidence interval. The ICC assesses measurement reliability by comparing the variability of different measures on the same subject to the total variation across all measures and all subjects. (Vincent and Weir 1994; Shand and Foster 1999; Barrett 2001) The formula for the ICC is: Var (B)-Var (W)/Var (B) + Var (W), where Var (W) is the pooled variance within subjects, and Var (B) is the variance of the measurements between subjects. Three different models of ICC were calculated for PIPR and MPC from all conditions: (MacLennan 1993) (1) ICC(1,1)—one-way random single measure i.e., measures were randomly repeated on each subject with reliability calculated from a single measurement; (2) ICC(1,3)—as above, but with reliability calculated by taking the average of 3 random measurements; and (3) ICC(1,6)—for the hemifield conditions, the extra 3 repeated measurements from the second session were combined with the first 3 repeated measurement to calculate ICC(1,6) with reliability calculated by taking the average of 6 random measurements.

3.3 Results

3.3.1 Post-Illumination Pupil Response (PIPR)

PIPR values for each test trial and mean pupil responses from the 5 stimulation conditions are shown in Table 3 and Figure 8 respectively. There were significant differences in mean PIPR between the 5 conditions (F(4, 36) =62.68, p<0.001). Post-hoc analysis revealed no statistically significant difference in PIPR between the mean of upper and lower hemifield stimulation (p=0.996), the waveforms of which were largely overlapping. The hemifield PIPR measures from the two sessions (session 1: tests 1-3 vs. session 2: tests 4-6) were highly correlated—there was no statistically significant difference between the hemifield PIPR from session 1 and session 2: the lower hemifield PIPR from session 1 was 0.91±0.05 (mean±SD) and 0.90±0.04 from session 2 (p=0.55, Pearson’s r = 0.81); the upper hemifield PIPR from session 1 was 0.90±0.04 and 0.90±0.05 from session 2 (p=0.63, Pearson’s r = 0.79). Full-field blue PIPR was significantly greater than that of central-field and hemifield responses (p<0.001); however, the differences between central-field PIPR and upper/lower hemifield PIPRs did not reach statistical significance (p=0.069 and 0.078 respectively). Full-field red stimulation induced significantly smaller PIPR than the blue stimulation across all conditions. All p values of pairwise comparisons of PIPR are summarized in Table 4.

The coefficient of variation and intra-class correlation coefficient data for each condition are shown in Table 5. The first three repetitions of PIPR for the upper and lower hemifield stimulations had ICC(1,3) means of 0.63 and 0.77 respectively. To test if more repetitions would increase the mean ICC, we repeated upper and lower hemifield stimulation 3 additional times in a second experimental session. After combining 6 repeated hemifield tests, the ICC(1,6) increased to 0.87 for the lower hemifield and 0.88 for the upper hemifield.

Figure 8. Mean hemifield, central-field and full-field PIPR from 10 visually-normal subjects. Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of blue or red light stimulation (400ms, 400cd/m2). Figure 2a, mean pupil response tracings for the 5 stimulation conditions; Figure 2b and 2c, mean PIPR pupil size changes (1-PIPR) and mean MPC pupil size changes (1-MPC). The 2 sessions of hemifield recording are plotted separately, designated as “#1” and “#2”. Error bars represent standard error (SE). LHF, Lower Hemifield; UHF, Upper Hemifield; CF, Central-field; and FF, Full-field.

Lower Upper Central-Field Full-Field Full-Field Red

Hemifield Hemifield Blue

Test 1 0.92±0.05 0.90±0.05 0.84±0.08 0.64±0.12 0.98±0.03

Test 2 0.91±0.06 0.88±0.08 0.83±0.10 0.67±0.10 1.00±0.03

Test 3 0.90±0.06 0.92±0.03 0.82±0.10 0.68±0.08 0.98±0.02

Test 4 0.90±0.04 0.90±0.06 N/A N/A N/A

Test 5 0.89±0.06 0.90±0.06 N/A N/A N/A

Test 6 0.92±0.04 0.90±0.05 N/A N/A N/A

ANOVA F (5, 45)=1.15 F (5, 45)=1.07 F (2, 18)=1.06 F (2, 18)=2.26 F (2, 18)=0.96

p=0.339 p=0.375 p=0.355 p=0.146 p=0.402

Table 3. PIPR values of each test trial from the 5 stimulation conditions.

Upper Hemifield Central-field Full-field Blue Lower Hemifield

PIPR = 0.91±0.04 PIPR = 0.90±0.04 PIPR = 0.83±0.09 PIPR = 0.66±0.09

Upper Hemifield p=0.996 PIPR = 0.90±0.04

Central-field p=0.078 p=0.069 PIPR = 0.83±0.09

Full-field Blue p<0.001 p<0.001 p<0.001 PIPR = 0.66±0.09

Full-field Red p=0.003 p=0.004 p=0.004 p<0.001 PIPR = 0.99±0.01

Table 4. Mean PIPR and p values for pairwise comparisons during the 5 testing conditions. Mean PIPR ± standard deviation from 10 individual participants’ data obtained by averaging 6 repeated tests from each hemifield condition and 3 repetitions for central-field and full-field conditions. p values (in italics) were obtained by post-hoc analysis of one-way repeated measures ANOVA, adjusted for multiple comparisons by Bonferroni correction. Significant p values are denoted in bold.

Lower Upper Central-Field Full-Field Full-Field Red

Hemifield Hemifield Blue

Median CV 0.38 0.30 0.16 0.10 0.87

(range) (0.06, 0.62) (0.04, 0.57) (0.02, 0.27) (0.04,0.23) (-809.30, 3.36)

ICC(1,1) 0.53 0.36 0.87 0.84 -0.26

(CI 95%) (0.14, 0.83) (0.01, 0.74) (0.69, 0.96) (0.60,0.95) (-0.42, 0.15)

ICC(1,3) 0.77 0.63 0.95 0.94 -0.44

(CI 95%) (0.32, 0.94) (0.02, 0.90) (0.87, 0.99) (0.83,0.98) (-1.68, 0.48)

ICC(1,6) 0.87 0.88 N/A N/A

(CI 95%) (0.69, 0.96) (0.71, 0.97) N/A

Table 5. Test-retest reliability of PIPR measured during hemifield, central-field and full- field stimulation. Intra-subject coefficient of variation (CV) = SD/(1- mean PIPR). This measure was calculated using 6 trial repetitions for hemifield conditions and 3 repetitions for central-field and full-field conditions. The median and range for 10 individual subjects are reported. ICC(1,1) and ICC(1,3) were calculated from 3 repetitions. ICC(1,6) for the hemifield conditions was calculated by combining 6 repetitions across 2 sessions. (CI, 95% confidence interval)

3.3.2 Maximal Pupil Constriction (MPC).

Similar to hemifield PIPR, the hemifield MPC measures from the two sessions were also highly correlated—there was no statistically significant difference between the hemifield MPCs from session 1 and session 2: the lower hemifield MPC of session 1 was 0.60±0.04 and 0.60±0.05 for session 2 (p=0.88, Pearson’s r = 0.85); the upper hemifield PIPRs from session 1 were 0.59±0.05 and 0.60±0.04 for session 2 (p=0.59, Pearson’s r = 0.82).

MPC values for each test trial are shown in Table 6. There was a significant difference between the five conditions for mean MPC (F(4, 36) =68.24, p<0.001). The differences in mean MPC in all pairwise comparisons were statistically significant except for upper hemifield vs. lower hemifield (p=0.998) and lower hemifield vs. full-field red stimulation (p=0.104). All p values for the pairwise comparisons of MPC are summarized in Table 7. The coefficient of variation and intra-class correlation coefficient data for each condition are shown in Table 8. The first three repetitions of MPC for the lower and upper hemifield stimulations had mean ICC(1,3) of 0.92 and 0.93 respectively. After combining 6 repeated hemifield tests, the ICC(1,6) increased to 0.95 for the lower hemifield and 0.94 for the upper hemifield.

Lower Upper Central-Field Full-Field Blue Full-Field Red

Hemifield Hemifield

Test 1 0.61±0.04 0.59±0.05 0.53±0.05 0.42±0.06 0.67±0.06

Test 2 0.61±0.05 0.59±0.06 0.54±0.05 0.44±0.06 0.65±0.06

Test 3 0.60±0.05 0.60±0.05 0.52±0.05 0.44±0.07 0.66±0.05

Test 4 0.60±0.06 0.60±0.04 N/A N/A N/A

Test 5 0.60±0.05 0.60±0.05 N/A N/A N/A

Test 6 0.61±0.06 0.60±0.05 N/A N/A N/A

ANOVA F(5,45)=0.92 F(5,45)=0.37 F(2,18)=2.25 F(2,18)=0.77 F(2,18)=2.26

p=0.436 p=0.725 p=0.144 p=0.427 p=0.152

Table 6. MPC values for each test trial from the 5 stimulation conditions.

Lower Hemifield Upper Hemifield Central-field Full-field Blue

MPC = 0.60±0.05 MPC = 0.60±0.04 MPC = 0.33±0.04 MPC = 0.43±0.06

Upper Hemifield p=0.998

MPC = 0.60±0.04

Central-field p=0.001 p=0.001

MPC = 0.33±0.04

Full-field Blue p<0.001 p=0.001 p<0.001

MPC = 0.43±0.06

Full-field Red p=0.104 p=0.042 p=0.001 p<0.001

MPC = 0.66±0.05

Table 7. Mean MPC and p values for pairwise comparisons during the 5 testing conditions. Mean MPC ± standard deviation from 10 individual participants’ data obtained by averaging 6 repeated tests from the individual hemifield conditions and 3 repetitions for central-field and full-field conditions. p values (in italics) were obtained by post-hoc analysis of one-way repeated measures ANOVA, adjusted for multiple comparisons by Bonferroni correction. Significant p values are denoted in bold.

Lower Upper Central-Field Full-Field Full-Field Red

Hemifield Hemifield Blue

Median CV 0.06 0.05 0.05 0.02 0.04

(range) (0.10, 0.04) (0.11, 0.02) (0.11, 0.01) (0.04, 0.01) (0.11, 0.01)

ICC(1.1) 0.79 0.86 0.70 0.76 0.88

(CI95%) (0.51, 0.94) (0.66, 0.96) (0.38, 0.90) (0.46, 0.92) (0.70, 0.97)

ICC(1.3) 0.92 0.93 0.88 0.90 0.90

(CI95%) (0.76, 0.98) (0.85, 0.97) (0.64, 0.97) (0.72, 0.97) (0.72, 0.97)

ICC(1.6) 0.95 0.94 N/A N/A N/A

(CI95%) (0.89, 0.99) (0.86, 0.98)

Table 8. Test-retest reliability of MPC measured during hemifield, central-field and full- field stimulation. Intra-subject coefficient of variation (CV) = SD/(1- mean MPC). This measure was calculated using 6 trial repetitions for the hemifield conditions and 3 repetitions for the central-field and full-field conditions. The median and range for 10 individual subjects are reported. ICC(1.1) and ICC(1.3) were calculated from 3 repetitions. ICC(1.6) for the hemifield conditions was calculated by combining 6 repetitions across 2 sessions. (CI, 95% confidence interval)

3.4 Discussion

This study described a simple device designed and constructed to work with a commercially available Ganzfeld stimulator to induce hemifield, central-field and full-field PIPR. This device is easy to install and remove without alteration to the structure or function of the Ganzfeld stimulator itself. The test is well tolerated by the participants. With 400 cd/m2, 400 ms blue light stimuli, we recorded differentiable PIPR and MPC from 10 visually normal subjects and found that the hemifield, 30° central-field and full-field stimuli induced increasingly larger PIPR and MPC. These findings are consistent with our previous observation that PIPR is a function of stimulus intensity, duration, and in particular, retinal area stimulated (Lei, Goltz et al. 2014). As we expected, mean responses of upper and lower hemifield stimulation are highly symmetric, indicating that there is no systematic bias in our hemifield stimulation apparatus nor in our normal subjects.

In this study, we used intra-class correlation coefficient (ICC) as our assessment measure of test-retest reliability. ICC describes how closely a set of repeated measurements resemble each other. It quantifies the direction (+/-) and the strength of the relation between test-retest scores by estimating their linear relation, yielding a value between +1 and -1.24. It has been suggested that, as a general rule, a value of over 0.90 is considered high, between 0.80 and 0.90 as moderate, and under 0.80 as low reliability for using an instrument for individual decision-making (Shrout and Fleiss 1979; Vincent and Weir 1994; Lei, Goltz et al. 2014). Herbst and co-workers (Herbst, Sander et al. 2011) described a custom-built chromatic pupillometry system using 20 s of continuous bright blue light (300 cd/m2) to induce PIPR, but the area of stimulation was not specified. After analyzing 2 repeated measurements of PIPR, they reported an ICC of 0.80. In this study, for full-field stimulation, we achieved single measure PIPR reliability of 0.84 and excellent reliability of 0.94 after averaging 3 measures. Similarly, for 30° central-field stimulation, we obtained single measure PIPR reliability of 0.87 and excellent reliability of 0.95 after averaging 3 measures. The better reliability we observed may be attributed to the properties of our testing method. First, the duration of the light stimuli we used was substantially shorter (400 ms vs. 20 s), which minimizes the effect of eye blinking and squinting, resulting in more

consistent light exposure between trials. Second, the intensity of the light stimuli we used was higher (400 cd/m2 vs. 300 cd/m2), which induced a larger and more reliable PIPR.

For hemifield stimulation, however, not only is the PIPR amplitude smaller, the test- retest reliability is also substantially lower than that during central-field and full-field stimulation. Given the fixed intensity and duration of stimuli we used, as well as the smaller area of stimulation, the PIPR amplitude from either hemifield is expected to be smaller than that from central-field or full-field stimulation. The smaller amplitude of pupil constriction during the 10 to 30 s post-illumination interval following hemifield stimulation makes it more susceptible to changes induced by loss of mental engagement and other natural fluctuations of pupil contractility (Kahneman and Beatty 1966; Kahneman 1973; McLaren, Erie et al. 1992), which may lead to a reduced signal-to-noise ratio. To test whether using more repeated measures would achieve acceptable test-retest reliability, we added 3 more trials for upper and lower hemifield stimulation. The ICC for 6 averaged measures increased to 0.87 (from 0.63 for 3 measures) for lower hemifield and to 0.88 (from 0.77 for 3 measures) for upper hemifield, which is considered moderate reliability and is generally acceptable for clinical use (Shrout and Fleiss 1979; Vincent and Weir 1994). These findings emphasize the need to use the mean of multiple trials as the index when the PIPR values are expected to be small.

Measuring PIPR as an index of melanopsin-driven ipRGC activity is a promising new tool to assess inner retinal function independent of conventional photoreceptors (rods and cones). Using predominately central field focal chromatic stimulation, Kardon and coworkers provided evidence that a clinical chromatic pupillometry protocol could assess differentially the rod and cone-driven rapid phase responses and melanopsin-driven steady state and post-illumination response (Kardon, Anderson et al. 2009; Kardon, Anderson et al. 2011; Park, Moura et al. 2011; Kawasaki, Crippa et al. 2012). The central-field PIPR testing method has also been used by other groups (Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012; Münch, Léon et al. 2012; Feigl and Zele 2014; Nissen, Sander et al. 2014), including investigations into diseases such as retinitis pigmentosa (Kardon, Anderson et al. 2011; Kawasaki, Crippa et al.

2012), Leber's hereditary optic neuropathy (Park, Moura et al. 2011), glaucoma (Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011) as well as the circadian rhythm of ipRGC activity (Münch, Léon et al. 2012). Our previous work further refined the testing protocol for PIPR using full-field stimulation (Lei, Goltz et al. 2014). We induced consistent PIPR with a full-field stimulus of only a few hundred milliseconds. To the best of knowledge, all previous studies regarding melanopsin-driven PIPR testing used either central or full-field stimulation, but within subject comparison of PIPR from different sub-regions of retina has not been investigated. Although a multi focal “perimetry-like” pupillography technique has also been developed (Maddess, Ho et al. 2011; Chang, Arora et al. 2013), the induced pupil responses are usually small and transient, and it is still unclear whether multi focal pupillography can adequately induce the melanopsin-driven response in a consistently detectable manner. The ability of our method to induce and compare full-field, central-field and hemifield PIPR is of considerable clinical significance. While full-field stimulation can be used to assess PIPR as an index of generalized melanopsin-driven ipRGC function across the whole retina, central-field stimulation is more appropriate in diseases confined to the posterior pole such as maculopathy (Augood, Vingerling et al. 2006). Differential hemifield responses are particularly useful in conditions where retinal ganglion cell damage is topographically asymmetric, e.g., early glaucoma and anterior ischemic optic neuropathy (Hart and Becker 1982; DeLeón-Ortega, Carroll et al. 2007). Analyzing differential PIPR responses from full-field, central-field and hemifield stimulation may also facilitate the localization of retinal damage and further expand the clinical utility of PIPR testing.

In summary, we have described a practical method to induce full-field, central-field and hemifield PIPR as indices of melanopsin-containing retinal ganglion cell function. Full-field and central-field PIPR have good test-retest reliability with either a single measure or an average of multiple measures. For hemifield PIPR, however, the reliability of single measure estimates are low; accordingly, we recommended using the average of multiple measures to attain acceptable reliability. Investigators may use this information when interpreting their PIPR test results.

Chapter 4 4 The Effect of Red light Exposure on Pre-existing Melanopsin-Driven Post-illumination Pupil Response 4.1 Introduction

Intrinsically photosensitive retinal ganglion cells (ipRGCs) are a third photoreceptor class in the retina. (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002) They contain a novel photopigment, melanopsin, which was first described as a non-visual opsin in the dermal melanophores of African clawed frogs (xenopus laevis) (Provencio, Jiang et al. 1998; Provencio, Rodriguez et al. 2000). Melanopsin mediates a depolarizing phototransduction cascade in ipRGCs, allowing these cells to fire action potentials on their own upon light stimulation (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002). IpRGCs also receive synaptic input originating from rods and cones, mainly via the ON bipolar cell pathway (Wong, Dunn et al. 2007). The primary physiological function of the ipRGC pathway is to signal ambient light irradiance (Lucas 2013). It has been demonstrated that melanopsin-driven intrinsic ipRGC photoactivity has a unique “photo-counting” ability that can provide sustained tonic coding under constant illumination. This feature was thought to compensate for the fast-adapting rod/cone photoresponse (Dacey, Liao et al. 2004; Dacey, Liao et al. 2005). IpRGCs integrate melanopsin-driven intrinsic photoactivity and rod/cone-driven extrinsic photoactivity to provide the afferent signal for non- image-forming visual functions, such as circadian rhythm photoentrainment and the pupillary light response (Do and Yau 2010; Lucas 2013). There is growing evidence that melanopsin phototransduction may have a significant influence on conventional image-forming visual perception as well (Ecker, Dumitrescu et al. 2010; Schmidt, Chen et al. 2011; Schmidt, Do et al. 2011; Schmidt, Alam et al. 2014).

Much like rhodopsin in rods and photopsins in cones, melanopsin is also a vitamin A- based photopigment that employs 11-cis-retinal as its chromophore (Provencio, Jiang et al. 1998). After absorbing a photon, 11-cis-retinal is photoisomerized into all-trans-retinal, causing conformational changes in the opsin and subsequently triggering the phototransduction cascade

in the photoreceptor (Ebrey and Koutalos 2001). The activated opsin has to re-bind 11-cis- retinal for the photoreceptor to regain its photosensitivity. Benefiting from close proximity to underlying retinal pigmented epithelium (RPE), rods and cones rely on RPE cells to recycle their all-trans-retinal and supply them with 11-cis-retinal: the all-trans-retinal is dissociated from the photoreceptor and transferred to RPE cells, in which the all-trans-retinal is converted back to 11- cis conformation and made available for rods and cones through a multi-enzymatic chemical reaction cascade called the “visual cycle” (Ebrey and Koutalos 2001). IpRGCs, however, are located in the inner retina, far away from the RPE layer, which raises an intriguing question: how does melanopsin maintain its photosensitivity without ample supply of 11-cis-retinal?

There is growing evidence (Lucas 2006; Mure, Rieux et al. 2007; Rollag 2008; Mure, Cornut et al. 2009; Matsuyama, Yamashita et al. 2012; Sexton, Golczak et al. 2012) suggesting that melanopsin’s chromophore regeneration is at least partially light-dependent: the isomerized chromophore (all-trans-retinal) does not leave the ipRGC; instead, it remains stably attached to activated melanopsin (meta-melanopsin), and is converted back to 11-cis conformation by absorbing light energy at a different wavelength. This inter-switchable state of chromophore is call “opsin bistability”. This opsin regeneration mechanism is commonly found in photoreceptors in invertebrates animals, such as Limulus and Drosophila (Koyanagi, Kubokawa et al. 2005).

While there is consensus that melanopsin activation (the forward reaction) is selectively sensitive to short wavelength light, with maximum absorption being at about 480 nm (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Gamlin, McDougal et al. 2007; Do, Kang et al. 2009; Mure, Cornut et al. 2009; Do and Yau 2010), conflicting evidence exists regarding the spectral sensitivity of melanopsin photo-regeneration (the backward reaction) (Mure, Rieux et al. 2007; Mawad and Van Gelder 2008; Rollag 2008; Mure, Cornut et al. 2009; Matsuyama, Yamashita et al. 2012). Mure and co-workers (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009) showed that pre-exposure to long-wavelength red light potentiates the photoresponse to short-wavelength light (as measured by spiking of SCN neurons, the pupillary light reflex and negative masking effect), a feature of bistable opsin. They reported that the putative spectral sensitivity of meta- melanopsin is red-shifted, with maximal absorption at 587 nm (Mure, Cornut et al. 2009).

However, a similar paradigm failed to induce potentiation in ipRGCs cell firing in an in vitro experiment, which argues against this “blue forward, red backward” hypothesis (Mawad and Van Gelder 2008). Furthermore, a recent photochemical study on purified rat melanopsin protein showed that melanopsin and meta-melanopsin have essentially overlapping spectral sensitivity, with both peaking in the range of blue light (467 vs. 476 nm) (Matsuyama, Yamashita et al. 2012), but so far there is no in vivo evidence to support this “blue-forward, blue-backward” theory yet.

Since the recovery of bistable opsin is light-dependent, after being activated with an intense light exposure near its maximum absorption, a photoreceptor that employs a bistable opsin often shows pro-longed response in the dark, which reflects an insufficient down-stream deactivating mechanism and a long half-life of the activated bistable opsin in the absence of subsequent stimulation (Hillman, Hochstein et al. 1983; Wang and Montell 2007; Hardie and Postma 2008). This electrophysiological property gives rise to another classic experimental paradigm to investigate opsin bistability: a prolonged post-stimulation potential is induced first by stimulating the photoreceptor with a pulse of intense light near the optimal excitation wavelength, then a second intense light exposure at a longer wavelength can be used to see if the prolonged response can be quenched (Hillman, Hochstein et al. 1983; Qiu and Berson 2007). This paradigm has been used extensively in investigating opsin bistability in invertebrate photoreceptors (Hillman, Hochstein et al. 1983). A number of studies have demonstrated that melanopsin-containing ipRGCs indeed show a prolonged depolarizing post-stimulation potential after intense exposure to melanopsin-activating blue light, which is consistent with the features of a bistable-opsin-driven photoresponse (Berson, Dunn et al. 2002; Hattar, Liao et al. 2002; Dacey, Liao et al. 2004; Dacey, Liao et al. 2005). This prolonged post-stimulation potential leads to a sustained pupil constriction beyond the offset of the light stimulus (Gamlin, McDougal et al. 2007). This so-called post-illumination pupil response (PIPR) has been widely accepted as an in vivo index of the melanopsin-driven intrinsic ipRGC photo response (Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Herbst, Sander et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Park, Moura et al. 2011; Herbst, Sander et al. 2012; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012; Münch, Léon et al. 2012; Herbst, Sander et al. 2013;

Roecklein, Wong et al. 2013; Lei, Goltz et al. 2014; Nissen, Sander et al. 2014; Lei, Goltz et al. 2015).

In order to provide evidence regarding the in vivo recovery mechanism of melanopsin, we conducted two experiments to investigate the effect of long wavelength red light exposure on pre-existing PIPR by presenting red light stimuli of variable intensity and duration at different time points of blue light-induced PIPR. We predicted that if the spectral sensitivity of meta- melanopsin is red-shifted, red light exposure will attenuate the pre-existing PIPR, and if the spectral sensitivity of meta-melanopsin is essentially overlapping with that of melanopsin, red light exposure will have no effect on PIPR.

4.2 Methods

4.2.1 Participants

Ten visually-normal subjects who had normal or corrected-to-normal vision (20/20 Snellen visual acuity or better) participated in the experiment 1 and 2 respectively (not every subject participated in both experiments). All participants underwent a screening ophthalmology exam, which include visual acuity, contrast sensitivity, eye alignment and movement tests, colour vision assessment (HRR test, Richmond Products, NM, USA), intraocular pressure assessment using slit-lamp tonometry, slit-lamp assessment of iris structures, anterior chamber angle and non-dilated fundus exam. Inclusion and exclusion criteria were previously described in Chapter 3 (section 3.2). The study was approved by the Research Ethics Board at The Hospital for Sick Children. All the procedures adhered to the guidelines of the Declaration of Helsinki. Written informed consent was obtained from each participant.

4.2.2 Apparatus

The technical specifications of our chromatic pupillometry apparatus have been. described earlier in this thesis (3.2.2). Only full-field stimulation was used in this study.

4.2.3 Testing conditions and protocols

All subjects were exposed to an indoor laboratory environment with ambient lighting levels ranging from 80-400 lux for at least 2 hours prior to the experiment. During the pupillometry recordings, participants were seated in a quiet darkened room (0 lux) with their head resting on a chinrest, with the forehead touching the upper edge of the Ganzfeld screen opening to receive full-field chromatic stimulation

4.2.3.1 Experiment 1

Experiment 1 was conducted monocularly, with one eye patched, and the fellow eye stimulated and recorded. The pupillary light response was induced with light stimuli under 4 conditions:

Condition 1: 400 ms red light (red 400 ms)

Condition 2: 2600 ms red light (red 2600 ms)

Condition 3: 400 ms blue light (blue 400ms)

Condition 4: 400 ms blue light immediately followed by 2600 ms red light (blue + red)

The intensity of red (640±10 nm) and blue (467±17 nm) light stimuli was set at 400 cd/m2 for this experiment. Conditions 1 and 2 were tested in a single trial, which began with 5 s of dim amber light (590 ± 7 nm, 0.3 cd/m2) pre-exposure (in order to ensure precisely 90 s of dark adaptation before light stimulation), followed by 90 s of dark adaptation allowing pupil to reach a resting state, which was then followed by 3 flashes of 400 ms red stimuli and 3 flashes of 2600 ms red stimuli in 15 s interval. Pupil response was recorded in real-time at 60 Hz, starting from 5 s prior to the onset of first light stimulation until 15 s after the offset of the last flash. Then the room light was turned on, and subject was allowed to remove their chin from the chinrest and to take a short break (20 s to 1 minute) before starting another trial.

Conditions 3-4 were each repeated 3 times in 3 separate trials. Similarly, each trial began with 5 s of dim amber light pre-exposure and 90 s dark adaptation, followed by a flash of the experimental light stimulus. The PIPR was recorded from 5 s prior to the onset until 60 s after the offset of stimulus. Subjects were allowed to take break after each trial to prevent fatigue and carry-over of PIPR.

A 500 ms sound beep was presented 5 s prior to and at the onset of each experimental flashes as an auditory cue to remind subjects to keep their eye open for stimulation. The order of testing conditions was randomized.

4.2.3.2 Experiment 2

Experiment 1 was conducted monocularly, the effectiveness of red light exposure might be limited by the fact that it was presented to an already-constricted pupil. To maximize the retinal irradiance of red light, we conducted experiment 2, where both eyes were stimulated with

one eye being pharmaceutically dilated, with the pupil response recorded from the non-dilated eye. The pupil of non-dominant eye was dilated with topical 1% tropicamide and 2.5% phenylephrine ophthalmic solution. Both eyes were stimulated simultaneously, while pupil response was recorded from the non-dilated dominant eye.

Experiment 2 started with a red-light-only reference trial, where 9 flashes of red light stimuli of increasing intensity and duration (1, 3.16, 10, 31.6, 100, 316, 1000 cd/m2 for 1 s, 1000 cd/m2 for 5 s and 1000 cd/m2 for 10 s) were presented in 15 s intervals. The red light-only control trial was followed by 9 double-flash (blue+red) trials and a blue-only single-flash trial tested in randomized order. For the double-flash conditions, a blue stimulus of 400 cd/m2 intensity was presented for 200 ms to induce PIPR, then red stimuli of increasing intensity and duration (1, 3.16, 10, 31.6, 100, 316, 1000 cd/m2 for 1 s, 1000 cd/m2 for 5 s and 1000 cd/m2 for 10 s, 9 steps) were presented at 9 s after the offset of each PIPR-inducing blue stimulus. For the single-flash trial, only a 400 cd/m2, 200 ms blue flash was presented. Similar to experiment 1, all trials started with 5 s of dim amber light pre-exposure followed by precisely 90 s dark adaptation to make sure that the pupils were in the same resting state when the first flash was presented. Auditory cues were also given in the same fashion as experiment 1.

4.2.4 Data Analysis

Data from the eye tracker were filtered and inspected using a custom-written processing program as described in Chapter 3 (session 3.2.4). The filtered data were then normalized to the baseline pupil size calculated from the mean pupil size during a 5 s period before the onset of each stimulus (i.e., Normalized pupil size = Absolute pupil size / Baseline pupil size). It was anticipated that the putative red light attenuating effect would alter the long-term recovery tendency of PIPR, therefore the PIPR measurement was defined as the average normalized pupil size over a 20 second period from 10 to 30 seconds after the offset of blue light stimuli in experiment 1, and over a 35 second period from 25 to 60 second after the offset of the first blue light stimuli in experiment 2 (smaller value means greater PIPR). We did not anticipate “red only” control conditions to produce significant PIPR, therefore, PIPR was not measured from these

trials. Maximum pupil constriction (MPC) after light stimulation was also measured and compared.

Statistical analyses were performed using SPSS 19.0 (IBM Corporation. Armonk, NY). Differences in mean PIPR between “blue-red” and “blue only” conditions were compared using one-way repeated measures ANOVAs with post hoc pairwise comparisons adjusted for multiple comparisons using the Bonferroni method. A value of p<0.05 was considered statistically significant.

4.3 Results

4.3.1 Experiment 1

Mean pupil responses from all tested conditions from 10 visually normal participants (6 females, mean age 31.5 years, age range 20-57 years) are summarized in Figure 9: as expected, while the reference “red only” stimulations did not induce sustained PIPR, a 400 cd/m2, 400 ms blue flash induced rapid pupil constriction followed by slow-decaying PIPR, which is consistent with the pattern of the melanopsin-driven pupil response. (Park, Moura et al. 2011; Lei, Goltz et al. 2014; Lei, Goltz et al. 2015) Interestingly, the PIPR induced by combined “blue+red” stimulation (400 ms blue flash followed by 2,600 ms red light exposure, M=0.698, SD=0.079) was not statistically different from “blue only” stimulation (M=0.707, SD=0.071, t(9)=0.619, p=0.551). The pupil tracings of these two conditions were substantially overlapping, except that the rapid constriction in response to the “blue+red” stimulus reached an MPC end-point that was greater than for the 400 ms blue light and 2,600 ms red light alone, there was a significant effect on MPC as confirmed by one-way repeated measures ANOVA: Wilks’ Lambda=0.022,

F(3,6)=87.339, p<0.001, with post-hoc pairwise comparison on MPC of “blue only” (M=0.431, SD=0.053) and “blue+red” condition (M=0.350, SD=0.037, p=0.006), and that the rapid redilation phase after the offset of “blue+red” stimulus briefly fell below the blue-light induced PIPR. In a further examination of the trough of the re-dilation phase, we sampled pupil size data over a 1 s interval centred at 3.5 s after the offset of “blue+red” stimulus, and compared with the data of “blue only” PIPR. (Figure 10) The results showed that in this time-window, “blue+red” PIPR (M=0.555, SD=0.053) was significantly lower than “blue only” PIPR (M=0.502,

SD=0.043, t(9)=-4.743, p=0.001), confirming a transient reduction in PIPR after the offset of red light.

Figure 9. Mean pupil responses from 10 visually-normal subjects in Experiment 1, Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of stimuli. Note the PIPR tracing of the “blue only” condition (400 ms blue flash) and “blue+red” condition (400 ms blue flash followed by 2,600 ms red light exposure) are highly overlapping; magnified box: comparing pupil size data over a 1 s interval centred at 3.5 s after the offset of “blue+red” stimuli to “blue only” condition, values represent mean±SD, p=0.001. Figure 9b, comparison of “blue only” and “blue+red” PIPR measurements, there is no significant difference. Figure 9c, comparison of MPC among all tested conditions, “b-lue+red” stimulus induced significant greater MPC than “red only”

and “blue only” stimuli (one-way repeated measure ANOVA, p<0.006). Error bars represent standard deviation.

4.3.2 Experiment 2

Experiment 1 was conducted monocularly, but the effectiveness of the red light exposure might have been limited by the fact that it was presented to an already-constricted pupil. To maximize the retinal irradiance of red light, we conducted experiment 2, where both eyes were stimulated with one eye being pharmaceutically dilated, with the pupil response recorded from the non-dilated eye. We presented red light of variable intensity and duration at 9 s after the offset of a 400 cd/m2 , 200 ms melanopsin-activating blue flash.

Ten visually normal subjects (7 females, mean age 32.1 years, age range 20-57 years) participated in experiment 2. All participants successfully completed the “red-only” and “blue- only” reference trials. Mean pupil tracings for the “red only” reference conditions are summarized in Figure 10: no significant PIPR was induced even with the most intense red light (1000 cd/m2) presented for 10 s, indicating that the red light exposure in this experiment did not activate melanopsin. When red light exposure was presented after the blue flash, 3 participants had significant eye squinting, although none of these 3 photophobic participants closed their eyes during red light exposure, the narrowing palpebral fissures caused artifact in the pupil tracings when the red light was turned on, but the PIPR measurements were not affected because of the time delay between stimulation and PIPR measurement. Therefore, the data from these 3 subjects were included for PIPR comparison. As shown in Figure 11 and Figure 12, mean PIPR tracings of “blue-red” trials are essentially overlapping with “blue only” control. There was no statistically significant difference in PIPR measurements among tested “blue-red” and “blue only” conditions (F(9,81)= 0.976, p=0.466, one-way repeated measure ANOVA).

The pupil response tracings were inspected further by excluding the 3 individuals who produced artifacts (Figure 13 and Figure 14). The PIPR measurements still did not differ from each other (F(6, 54)= 1.075, p=0.378, one-way repeated measure ANOVA). Similar to the results of experiment 1, compared to the “red only” reference, red light exposures presented after the

melanopsin-activating blue light induced greater MPC, all pair-wise comparisons of MPC (red light reference vs. red light presented on blue light PIPR) reached statistical significance (Figure 15). A transient reduction in PIPR following the offset of red light was also observed in most of the “blue-red” double flash conditions, but the PIPR difference at 3.5 s after the offset of red light (“blue-red” vs “blue” only) did not reach statistical significance (p>0.1).

Figure 10. Mean pupil responses to “red only” reference stimulations from 10 visually- normal subjects.

Figure 11. Mean pupil responses from 10 visually-normal subjects in Experiment 2. Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of stimuli. PIPR was induced with a 400 cd/m2, 200 ms full-field blue light stimulus presented to both eyes simultaneously; red light exposure of variable intensity and duration was presented at 9 s after the offset of the blue flash. Artifacts were mainly from 3 individuals who had difficulty keeping their eyes open during the red light exposure. Grey area: PIPR measurement was defined as mean normalized pupil size over a 35 second period from 25 to 60 seconds after the offset of the blue flash. Mean±SD of PIPR for all 10 conditions is represented in the inset bar graph.

Figure 12. Comparisons of PIPR tracings of “blue+red” conditions vs. “blue only” control from 10 visually normal participants. Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of stimuli. PIPR was induced with a 400 cd/m2, 200 ms full-field blue light stimulus presented to both eyes simultaneously; red light exposure of variable intensity and duration was presented at 9 s after the offset of the blue flash. Text underneath the curves indicates intensity and duration the red light exposure in “Blue+Red” conditions.

Figure 13. Mean pupil responses from 7 visually-normal subjects in Experiment 2. Data from 3 photophobic individuals who produced pupil tracing artifacts were deleted in order to reveal the waveform of rapid phase pupil response during and immediately after the red light exposure. Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of stimuli. PIPR was induced with a 400 cd/m2, 200 ms full- field blue light stimulus presented to both eyes simultaneously; red light exposure of variable intensity and duration was presented at 9 s after the offset of the blue flash. Grey area: PIPR measurement was defined as mean normalized pupil size of a 35 second period from 25 to 60 seconds after the offset of the blue flash. Mean±SD of PIPR was represented in the inset bar graph.

Figure 14. Comparisons of PIPR tracings from “blue+red” conditions vs. “blue only” controls from 7 visually normal participants. Data from 3 photophobic individuals who produced artifacts were deleted in order to reveal the waveform of the rapid phase pupil response during and immediately after the red light exposure. Pupil diameter data were normalized to the mean of 5 s of baseline recording prior to the onset of stimulation. PIPR was induced with a 400 cd/m2, 200 ms full-field blue light stimulus presented to both eyes simultaneously; red light exposure of variable intensity and duration was presented at 9 s after the offset of the blue flash. Legends underneath the curves indicate the intensity and duration the red light exposure in “Blue+Red” conditions.

Figure 15. Comparisons of maximum pupil constriction (MPC) induced by “red only” stimuli vs. red stimuli presented on top of blue-light-induced PIPR (“Red on PIPR”). Smaller values represent greater pupil constriction. Error bars represent standard deviation. (n=7, 3 individuals who produced artifacts during the red light exposure were excluded from this analysis). Compared to the “red only” reference, red light exposure presented after the melanopsin-activating blue light induced greater MPC, all pair-wise comparisons reached statistical significance (p<0.001, paired sample t-test).

4.4 Discussion

The primary goal of this study was to investigate whether blue light-induced pre-existing PIPR can be cut short by subsequent exposure to long wavelength red light. By presenting red light stimuli of a wide range of intensities (1-1000 cd/m2) and duration (1-10 s) at two different time points in the blue light-induced PIPR (immediately after the blue light exposure in experiment 1 vs. 9 s after blue light in experiment 2), we found no significant difference in PIPR between “blue-red” double-flash conditions and the blue-only control. Our findings argue against the previous assertion that long wavelength red light converts activated melanopsin back to its resting state (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009). It is a general feature demonstrated in many invertebrate bistable opsins that the maximum absorption wavelength of the backward reaction is usually longer than that of the forward reaction (Ritter, Zimmermann et al. 2004; Wang and Montell 2007; Hardie and Postma 2008). Therefore, early investigations into melanopsin bistability were based on the assumption that the maximum absorption wavelength for meta-melanopsin is red-shifted from that of melanopsin. Mure and co-workers (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009) previously conducted a series of experiments to show that pre-exposure using 620 nm long-wavelength red light potentiates many in vivo photo responses to melanopsin-activating blue light, including the spiking of SCN neurons, sustained pupil constriction and the negative masking effect. The authors suggested that these findings are in vivo evidence of melanopsin bistability based on the principle that pre-exposure to red light drives more meta-melanopsin back to its resting state, thus enhancing the sensitivity of the melanopsin-driven response. By computing the spectral tuning of red light potentiation of the sustained pupil response to blue light, they reported a putative spectral sensitivity of meta- melanopsin that peaks at 587 nm (Mure, Cornut et al. 2009). If this “blue forward, red backward” hypothesis for melanopsin bistability is true, the red light stimuli from the current study (maximum 1000 cd/m2 for 10 s) should convert a significant proportion of meta-melanopsin induced by the 400 cd/m2 blue flash of 200-400 ms, and thus quench the PIPR mediated by the previous melanopsin-driven post-stimulus ipRGC potential. However, such an attenuating effect was not observed in any of the double-flash conditions in our experiment. Even after the most intense red light exposure (1000 cd/m2 , approximately equivalent to 15 log quanta/cm2/s, for

10s), pupil tracings quickly returned to the PIPR trajectory pre-determined by previous blue light stimulation, and remaining very similar for up to 60 s. The absence of a red light attenuating effect on PIPR echoes the findings of Mawad and co-workers’ study, namely that red light pre-exposure did not induce potentiating on ipRGCs cell firing in an in vitro experiment (Mawad and Van Gelder 2008), which suggested that the in vivo potentiating effect that was observed by Mure et al. (Mure, Rieux et al. 2007; Mure, Cornut et al. 2009) may be a downstream effect rather than a consequence of melanopsin bistability (Mawad and Van Gelder 2008; Rollag 2008).

The absence of a red light attenuating effect on blue light-induced PIPR rather supports the idea emerging from a previous study, namely that the spectral sensitivity of melanopsin and meta-melanopsin being close to one other. In a study by Matsuyama and colleagues (Matsuyama, Yamashita et al. 2012), spectroscopic measurements on a large amount of purified rat melanopsin revealed that the maximum absorption wavelength of meta-melanopsin was only minimally longer than that of melanopsin (476 vs. 467 nm); it could be argued that the spectral sensitivities of melanopsin activation and regeneration are essentially overlapping, with both peaking in the range of blue light, which effectively explains our observation that long- wavelength red light neither induces nor suppresses the melanopsin-driven PIPR. With constant blue light exposure, Matsuyama and colleagues (Matsuyama, Yamashita et al. 2012) were able to generate a steady mixture of melanopsin (11-cis) and meta-melanopsin (all-trans), indicating that melanopsin activation and regeneration reactions are likely happening concurrently, the resultant 11-cis:all-transretinal ratio following constant illumination will be dependent on the relative conversion rates between forward and backward reactions (Rollag 2008; Matsuyama, Yamashita et al. 2012). If these photochemical findings can be translated directly into functional physiological properties, the melanopsin-containing ipRGCs would become “self-regenerating” blue light sensors, no matter how long the ipRGCs are stimulated, there will always be a certain amount of 11-cis-retinal-binding melanopsin available for subsequent photo stimulation. This putative property of melanopsin thus provides a convenient explanation for the fact that melanopsin is highly resilient to photo bleaching in vivo (Sexton, Golczak et al. 2012), and that

ipRGCs can provide sustained irradiance coding even under prolonged illumination near the maximum absorption wavelength (Dacey, Liao et al. 2004; Dacey, Liao et al. 2005).

By presenting red light stimuli during the PIPR activity, our experimental paradigm inevitably induced a rapid phase cone-driven extrinsic ipRGC response superimposed on the pre- existing melanopsin-driven intrinsic ipRGC activity. Therefore, our data should also be interpreted from the perspective of integration of ipRGC extrinsic/intrinsic photo activity. In both experiments, red stimuli superimposed on the melanopsin-driven PIPR induced further pupil constriction to a point that would not be reached by red light stimulation alone; this phenomenon is most clearly demonstrated by the comparison between 1000cd/m2, 10 s red light on PIPR vs. 1000cd/m2, 10 s red light alone. It was previously demonstrated in a growing body of literature that melanopsin activity is required for the pupil to reach maximal constriction; without the contribution from melanopsin, cone activity alone would not drive the pupil constriction beyond 80% of its maximum (Lucas, Hattar et al. 2003; Panda, Provencio et al. 2003; Semo, Peirson et al. 2003; Barnard, Appleford et al. 2004). Our findings clearly demonstrate the ability of the ipRGCs to temporally summate the cone-driven extrinsic photoactivity and melanopsin-driven intrinsic photoactivity to provide pupil size regulation. It is also noteworthy that immediately after the offset of red stimuli, the pupil briefly re-dilates to an extent below the PIPR before returning to the PIPR trajectory. This phenomenon was consistently observed in the averaged responses to most of the blue-red double flash conditions, it lasts only a few seconds, with the greatest disparity being at around 3-4 seconds after the offset of red light. By sampling this part of the data and comparing it to the blue-only PIPR, the difference was statistically significant in some of the double flash conditions. Considering the transient nature of this effect, it is very unlikely that this can be attributed to melanopsin bistability. We hypothesize that the transient suppression of pre-existing PIPR after the offset of red light stimulation is due to overshoot of iris dilator muscle or due to disaffiliation of the ipRGC pathway caused by the removal of cone input via ON bipolar cells. The physiological significance of this effect remains to be elucidated.

In summary, using a novel blue-red double flash chromatic pupillometry paradigm, we found a transient suppression of PIPR immediately after the withdrawal of red light stimuli, but no sustained attenuating effect of red light exposure on the pre-existing PIPR. This in vivo evidence argues against the previous hypothesis that melanopsin employs a long wavelength light-dependent mechanism to regenerate its 11-cis retinal chromophore. Rather it supports the idea that melanopsin activation and regeneration have overlapping spectral sensitivity, with both peaking in the range of blue light. The experimental paradigm described in this study may be useful for investigating the interaction between rod/cone-driven intrinsic ipRGC activity and melanopsin-driven intrinsic ipRGC activity.

Chapter 5

5 General discussion

The primary goal of this thesis was to develop a chromatic pupillometry system measuring the post-illumination pupil response as a convenient, reliable and objective tool to assess the newly discovered melanopsin-driven intrinsic ipRGC photoactivity in vivo, and to use this tool to answer clinical and basic science questions related to the melanopsin/ipRGC system. The experimental work consists of two parts: first, a clinically-oriented methodological study on hemifield, central-field and full-field PIPR testing. The results showed that upper and lower hemifield PIPRs in normal subjects are symmetric, and that hemifield, central-field, full-field stimulation induced increasingly greater PIPR, findings that are expected based on the current understanding of melanopsin photoactivity. Good test-retest reliability was demonstrated in hemifield, central-field and full-field PIPR. In effect, these findings validate this chromatic pupillometry system as a tool to test melanopsin-driven PIPR. The usefulness of this tool was subsequently demonstrated in a basic-science-oriented experiment implementing PIPR testing as an in vivo index of melanopsin photoactivity to investigate the effect of red light exposure on pre-existing PIPR. The results showed that red light exposure has no long term PIPR-attenuating effect; the withdrawal of red light exposure, however, causes transient suppression of PIPR, which provided insights into the chromophore regeneration mechanism of melanopsin and the interaction between extrinsic and intrinsic ipRGC activity.

5.1 Refinements of chromatic pupillometry PIPR testing

We successfully combined two commercially available devices into a chromatic pupillometry system: chromatic light stimuli were presented with a Ganzfeld screen stimulator originally designed as part of a electrophysiology system for full-field ERG and EOG testing (ColorDome LED full-field stimulator with the Espion V5 operating system; Diagnosys LLC, Lowell, MA); pupil reaction was monitored with a binocular head-mounted eye-tracking camera system (Arrington Research, Scottsdale, AZ). The Ganzfeld stimulator integrates a xenon light

bulb and monochromic LEDs: red (640±10 nm), amber (590±7 nm), green (535 ±15 nm ), blue (467±17 nm) as light sources to generate full-spectrum (white) and monochromatic light stimuli over a wide range of intensities and durations (technical documents from Diagnosys LLC, Lowell, MA). The blue light of the Colordome (467±17 nm) matches the peak absorption of melanopsin (～480 nm), the red light spectrum is narrow and well-distanced from the absorption spectrum of melanopsin. Therefore, blue light was chosen to stimulate the melanopsin-driven pupil response and red light was used as a melanopsin-silent reference condition. The eye tracker employs miniature near-infrared (940 nm) illuminating diodes and infrared cameras mounted on a light-weight spectacle frame to record changes in pupil diameter binocularly at a sampling rate of 60 Hz. The Ganzfeld stimulator and eye tracker were connected through a synchronizing circuit (Arrington Research, Scottsdale, AZ).

This combination of hardware as a chromatic pupillometry system was first described by Kardon and his colleagues (Kardon, Anderson et al. 2009). In their initial study, the Ganzfeld screen was positioned 75 mm away from the subject to present light stimuli of approximately 60° × 90° viewing angle (Kardon, Anderson et al. 2009). The reason for this unusual positioning of Ganzfeld screen was not clearly described in the paper; we presumed that it was due to the bulky forehead scene camera on the eye tracker frame which prevented the subjects from positioning their faces closer to the opening of the bowl-shaped Ganzfeld screen. In an effort to enhance this chromatic pupillometry setup, we removed the scene camera that was of no use in pupil recording -- a simple modification allowing the subject place their forehead against the edge of the Ganzfeld screen for full-field stimulation – the way the screen was designed to be used. In our earlier work (Lei, Goltz et al. 2014) on full-field chromatic pupillometry measuring PIPR, we found that full-field stimulation induced significantly greater PIPR than 60° × 90° central- field stimulation of the same intensity and duration. We showed that a sustained PIPR of more than 30 seconds can be induced with an intense full-field blue light stimulation of only a few hundred milliseconds duration. This “brief flash” PIPR stimulation protocol minimizes participant’s discomfort compared to previously described 10-30 seconds long duration bright light stimulation in darkness (Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Herbst, Sander et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Herbst, Sander et al.

2012; Herbst, Sander et al. 2013; Roecklein, Wong et al. 2013; Nissen, Sander et al. 2014). It also avoids the inconsistency in light exposure caused by eye blinking and squinting (See discussion in 1.7.4). More importantly, the offset of the long duration stimulation causes a rapid pupil redilation before the steady state PIPR, which is presumably caused by the withdrawal of the cone-driven extrinsic ipRGC activity. As shown in our experiment in Chapter 4, this rapid pupil redilation can cause transient suppression of melanopsin-driven PIPR, and may be problematic depending on how the PIPR measurement is sampled. This potential confounding effect can be eliminated by using a brief blue flash with duration less than 1 s that does not cause rapid pupil redilation at the offset of stimulation. The “brief flash” PIPR stimulating protocol was subsequently used and was well-tolerated by participants in the experiments described in chapters 3 and 4 of this thesis.

As a test of retinal function, it is clinically useful to induce and compare PIPRs from different retinal regions. However, previously published chromatic pupillometry papers either used central-field or full-field stimuli; a “regional PIPR” testing protocol has not been developed (Gamlin, McDougal et al. 2007; Kardon, Anderson et al. 2009; Mure, Cornut et al. 2009; Kankipati, Girkin et al. 2010; Feigl, Mattes et al. 2011; Kankipati, Girkin et al. 2011; Nissen, Sander et al. 2011; Roecklein, Wong et al. 2013; Nissen, Sander et al. 2014). To further refine the PIPR testing protocol, we designed and constructed a novel occluder to regulate the exposed area of the Ganzfeld stimulator, this add-on device is easy to install and remove without structural or functional changes to the Ganzfeld stimulator itself, so that intra-subject hemifield, central-field and full-field PIPR can be induced in a convenient manner. With 400 cd/m2, 400 ms brief blue flash stimuli, we recorded differentiable PIPR from 10 visually normal subjects with good test-retest reliability. We found that the hemifields, 30° central-field and full-field stimuli induced increasingly larger PIPR. These findings are consistent with our previous observation that PIPR is a function of stimulus intensity, duration, and in particular, retinal area stimulated (Lei, Goltz et al. 2014). As we expected, the mean responses to upper and lower hemifield stimulation are highly symmetric, indicating that there is no systematic bias in our hemifield stimulation apparatus nor in our normal subjects. The clinical applications of the hemifield, central-field and full-field PIPR testing will be further discussed in sections 5.3 and 7.2.

In chapter 4, we described a novel “blue-red” double flash protocol to study the effect of red light exposure on blue light–induced PIPR. In addition to providing in vivo evidence to support the hypothesis that melanopsin chromophore regeneration is not dependent on long wavelength light, this paradigm effectively produced cone-driven extrinsic ipRGC activity superimposed on pre-existing melanopsin-driven intrinsic ipRGC activity. Interestingly, we noted the summation effect of extrinsic/intrinsic ipRGC activity when the red light was ON, and a transient inhibition effect immediately after the red light OFF response (see discussion in section 4.4). The results indicate that this protocol may have important application in investigating the interaction between extrinsic and intrinsic ipRGC activity.

Besides the above-mentioned major refinements in the PIPR testing protocols, we also made multiple minor modifications and improvements to the current PIPR testing methodology, such as adding extra infrared LEDs to enhance the pupil tracking quality, a custom-built circuit producing auditory cues to help participants maintaining mental engagement and attention during pupillometry recording, and a period of pre-exposure to ensure the same duration of dark adaptation before each trials. Although the benefit of these minor changes were not directly evaluated in our experiments, after extensive testing in both normal subjects and in pilot studies of participants with ocular disease, we are confident that they are useful refinements to PIPR testing, and should be included in future development of chromatic pupillometry.

In summary, many novel modifications and refinements to multiple aspects of current PIPR testing methodology have been devised in the making of this thesis. From a knowledge translation perspective, these refinements are practical and easily implemented; therefore, they may have wide-spread positive impact on melanopsin/ipRGC related research.

5.2 Characteristics of melanopsin-driven post-illumination pupil response

It has been generally accepted that melanopsin-driven intrinsic ipRGC photoactivity provides steady irradiance coding at high intensity level with selective sensitivity to short wavelength light (Berson, Dunn et al. 2002; Dacey, Liao et al. 2005; Gamlin, McDougal et al.

2007; Do, Kang et al. 2009; Mure, Cornut et al. 2009; Do and Yau 2010). A remarkable “photon- counting” ability of melanopsin has been demonstrated by a number of studies, where the spikes in individual melanopsin-driven ipRGC activity showed a precise linear relation to the amount of light energy that they had been exposed to (Berson 2003; Dacey, Liao et al. 2005; Fu, Liao et al. 2005; Wong, Dunn et al. 2005; Do, Kang et al. 2009). These electrophysiological properties of melanopsin are well reflected in the melanopsin-driven PIPR. Data from our initial experimental work on full-field PIPR showed that when the stimulus intensity was low, PIPR was minimal, and the pupil responses to red and blue light were similar to each other. Sustained blue light- induced PIPR appeared when the intensity reached 10 cd/m2, a threshold much higher than that for rod or cone photoactivity. Further increase in stimulation intensity caused a linear increase in PIPR up to 400 cd/m2. It is noteworthy that 400 cd/m2 was the highest intensity our apparatus could produce using blue light, and to our knowledge, blue light with intensity higher than 400 cd/m2 has not been tested in human subjects. Linearity also exists in the duration-response relation over a large span of duration steps. This dual-linearity relationship (intensity-response and duration-response) is consistent with the “photo-counting” properties of melanopsin-driven ipRGC activity, and indicates that light intensity and duration can be interchanged proportionally to produce PIPR of similar amplitude or kinetics, facilitating the prediction regarding and the quantitative analysis of PIPR responses.

Considering the irradiance detecting property of melanopsin phototransduction in ipRGCs, and the fact that full-field stimulation induced greater PIPR than 60° × 90° central-field stimulation, it was not surprising to see in the experiment described in Chapter 3 that increasingly larger PIPR was induced with upper/lower hemifields, central-field and full-field stimulation, with the two hemifield conditions inducing essentially overlapping PIPR. More interestingly, the waveforms of the pupil tracing from hemifield, central-field and full-field stimulation are essentially parallel to each other, suggesting that these responses have a similar proportion of melanopsin’s influence; the ipRGC-mediated pupillary light reflex pathway may employ a simple linear spatial summation of melanopsin photoactivity. When the stimulated retinal area increases, the amplitude of the melanopsin-driven pupil response increases, but the waveform remains the same. Apparently this relationship would be limited by the endpoint of

maximum pupil constriction: once the full pupil constriction is reached, further increase of spatial summation of melanopsin activity will appear as a more sustained PIPR.

The experiments in Chapter 4 further reveal some characteristics of the interaction between cone-driven extrinsic ipRGC activity and melanopsin-driven PIPR: when presented on a pedestal of blue-light induced PIPR, red light exposure induced further pupil constriction to a level that would not be reached by red light stimulation alone, which reflects the cone-driven extrinsic ipRGC activity superimposed on the melanopsin-driven intrinsic ipRGC activity. Under these circumstances, the cessation of this cone-driven extrinsic influence led to a transient minor suppression of PIPR that lasted only a few seconds; the mechanism and physiological significance of this phenomenon remain to be elucidated. Most importantly, our data show that red light exposure did not cause a long-term attenuating effect on PIPR. It was quite remarkable to see that even after 10 s of 1000 cd/m2 red light exposure, the pupil quickly returned to the prior PIPR trajectory that was determined by the melanopsin-activating blue light presented over 9 seconds prior to the onset of the red light. This PIPR behavioural pattern supports the idea that melanopsin-driven ipRGC activity may confer a form of “photic memory” that faithfully encodes short wavelength light irradiance that the eye has been exposed to, which leads to a sustained component of pupil response that is largely unmodifiable by subsequent extrinsic ipRGC photoactivity.

In summary, the experimental work of this thesis has provided valuable information about the kinetic characteristics of melanopsin-driven PIPR, allowing investigators to tailor the PIPR testing paradigm to target a particular investigation question. It may also be used as a reference when predicting or interpreting PIPR test results.

5.3 Applications of hemifield, central-field and full-field chromatic pupillometry induced PIPR

Ever since the discovery of melanopsin and melanopsin-containing ipRGCs, there has been a fast growing interest in testing the melanopsin-driven PIPR in health and disease as an objective in vivo assessment of inner retinal function. This idea is attractive to visual

physiologists and ophthalmologists for many reasons: firstly and most obviously, the melanopsin-mediated retinal light-sensing pathway represents a novel aspect of inner retinal function, and we are still in the early stages of understanding the basic mechanisms of this system.How important this pathway is to the performance of our visual system, and how it is affected in retinal diseases remains largely unknown. PIPR as an objective and relatively convenient assessment of melanopsin photoactivity is particularly useful in providing in vivo evidence to address melanopsin/ipRGC-related clinical questions. Secondly, clinical observations indicate that ipRGCs may be relatively spared in the advanced stages of many retinal diseases. For example, early chromatic pupillometry studies also showed that impaired PIPR was observed in advanced glaucoma, but not in early glaucoma, which is consistent with the evidence from experimental animal glaucoma models that melanopsin-containing ipRGCs are relatively resilient to glaucomatous damage compared to conventional retinal ganglions cells. It has also been observed that in patients with severe retinitis pigmentosa, despite non-recordable ERG, the pupillary light reflex is still well preserved. In fact, a preserved pupil reaction to light in blind patients was one of the earliest clues that eventually led to the discovery of melanopsin- containing ipRGCs (Klerman, Shanahan et al. 2002; Zaidi, Hull et al. 2007). This evidence hinted at a potentially important clinical application of chromatic pupillometry--to monitor disease progression and the efficacy of therapeutic intervention in the end-stage of retinal diseases, where the other conventional functional indices, such as ERG and visual fields, are diminished. Last but not least, while conventional functional assessments including visual acuity, contrast sensitivity, visual field and ERG are all dependent on phototransduction of rods and cones, melanopsin-driven PIPR is purely an index of inner retinal function independent of rods and cones, as evidenced by the fact that PIPR remained largely unchanged even when the synaptic input from outer retinal was pharmaceutically blocked or mechanically removed (Gamlin, McDougal et al. 2007). This unique property may be very useful in certain clinical circumstances. For example, a 65 year-old completely blind man with end-stage outer retinal degeneration wanted to know if he would benefit from the newly approved retinal implants that restore vision by using a microelectrode array to stimulate retinal ganglion cells. In this clinical scenario, ERG and visual fields are not useful because of the lack of rod and cone function. However, melanopsin-driven PIPR may provide critical information to objectively and

quantitatively determine how much retinal ganglion cell function is preserved, provided that the functional status of ipRGCs can serve as a proxy of the overall health of retinal ganglion cells.

The knowledge generated in this thesis facilitates the application of PIPR testing as a functional assessment of melanopsin/ipRGC in both clinical and basic science research setting. In Chapter 3, we described a methodology to induce and compare differential PIPR using upper and lower hemifield, central-field and full-field stimulation. The results showed that hemifield, central-field and full-field stimuli induce increasingly greater maximum pupil constriction and PIPR, indicating that melanopsin-driven pupil response may have a linear spatial summation property. As expected, upper and lower hemifield blue light stimuli induced similar maximum pupil constriction and PIPR, with the pupil tracings being essentially overlapping. Being able to adjust the stimulation area and induce PIPR from different regions of retina using one pupillometry setup may facilitate the localization of retinal damage and expand the clinical utility of PIPR testing. Because full-field PIPR assesses the melanopsin-driven ipRGC function across the whole retina, it can be used when the involvement of ipRGC is expected to be generalized, such as end-stage glaucoma or retinal dystrophies, or in conditions where melanopsin-driven photo-sensing function is considered as a whole, such as seasonal affective disorder or circadian rhythm-related investigations. Central-field stimulation is more appropriate when retina disease is confined to the posterior pole such as maculopathy (Augood, Vingerling et al. 2006). Comparing upper and lower hemifield responses are particularly useful in conditions where retinal ganglion cell damage is topographically asymmetric, e.g. early glaucoma and anterior ischemic optic neuropathy (Hart and Becker 1982; DeLeón-Ortega, Carroll et al. 2007).

Besides clinical applications, PIPR testing obviously has important applications in basic science research as well. In Chapter 4, we investigated the effect of red light exposure on blue- light induced PIPR in an effort to provide in vivo evidence of the mechanism of melanopsin in vivo chromophore regeneration and the interaction between extrinsic/intrinsic ipRGC activities. The ability of our full-field chromatic pupillometry protocol to induce large PIPR that lasts tens of seconds with a brief blue flash makes it a convenient platform to conduct a novel experimental “double flash” paradigm: A brief melanopsin-activating blue light flash was employed to induce

PIPR, followed by a second impulse of red light to see if the kinetics of PIPR can be altered. The results showed that red light exposure induced further pupil constriction superimposed on the underlyinge PIPR, demonstrating integration of cone-driven and melanopsin-driven ipRGC activity, followed by a rapid redilation at the offset of red light that leads to minor transient suppression of PIPR, which is most likely due to a disfaciliating effect caused by the withdrawal of synaptic input from cones. The pupil tracing then returns to and follows the trajectory of PIPR pre-determined by the melanopsin-activating blue light. These data suggest that the process of melanopsin chromophore recovery may be independent of red light, which is consistent with latest photochemical evidence that the melanopsin at resting state and the activated meta- melanopsin have overlapping spectral sensitivity-both are selectively sensitive to blue light (see discussion in section 4.4). In addition to providing insight into the mechanism of melanopsin bistability, from a knowledge translation prospective, this study clearly demonstrated the usefulness of the PIPR testing methodology that we developed. The “double flash” protocol we described in this study may also be used in clinical investigations, for example, some inherited retinal dystrophies, such as congenital stationary night blindness (CSNB) and achromatopsia, are known to have abnormal patterns of pupil response at light-ON and OFF, but with unknown mechanisms. The “double flash” paradigm may reveal the abnormalities in the “rod/cone-ipRGC” pathway in this patient population.

100

Chapter 6 6 Conclusions

In conclusion, the experimental work of this thesis developed a novel chromatic pupillometry system where the stimulation area can be adjusted to induce melanopsin-driven PIPR from upper and lower hemifield, central-field and full-field stimulation. The results showed that upper/lower hemifields, central-field and full-field stimulation induced increasingly larger PIPR, with the upper and lower hemifield PIPR being symmetric and highly overlapping. Full- field and central-field PIPR have good test-retest reliability with either a single measure or an average of multiple measures. Acceptable test-retest reliability of hemifield PIPR can be obtained by using the average of multiple measures.

The full-field PIPR testing protocol was then used to investigate the effect of red light exposure on pre-existing PIPR. It was demonstrated that impulses of red light exposure presented over a large range of intensities, durations and at two different time points in the PIPR time course induced further pupil constriction superimposed on the pre-existing PIPR, demonstrating integration of cone-driven and melanopsin-driven ipRGC activity, followed by a rapid re-dilation at the offset of red light that leads to minor transient suppression of PIPR, which is most likely due to iris dilator muscle overshoot or brief suppression of ipRGC caused by the withdrawal of cone-driven input to ipRGC. The pupil tracing then returns to and follows the trajectory of PIPR pre-determined by the melanopsin-activating blue light. The absence of a red light attenuating effect on PIPR argues against the idea that the melanopsin chromophore regeneration may be dependent on red light, rather it is consistent with latest photochemical evidence that the melanopsin at resting state and the activated meta-melanopsin are both selectively sensitive to blue light.

The experimental apparatus and paradigm described in this thesis represent refinements to the current methodology of using chromatic pupillometry to investigate melanopsin-driven intrinsic ipRGC photoactivity. The results enhance the understanding of the characteristics of

101

melanopsin-driven PIPR, and may have important applications in both clinical and basic visual science research.

102

Chapter 7

7 Future directions

Knowledge and experience of using chromatic pupillometry to measure PIPR as an in vivo index of melanopsin-driven intrinsic ipRGC photoactivity has been accumulating, and intriguing questions have been emerging during the course of the above-described thesis projects, paving the road for further investigation. As part of my graduate studies, I have directly contributed to the initiation of several projects using chromatic pupillometry to investigate ocular diseases. Although these on-going projects have not produced enough data for them to be included as separate chapters in this thesis due to the pace of patient recruitment, I would like to briefly describe them here as future directions.

7.1 Using hemifield, central field and full-field chromatic pupillometry to investigate melanopsin-driven post- illumination pupil response in glaucoma patients

7.1.1 Introduction

Glaucoma is a grouping of eye diseases that result in loss of retinal ganglion cells in a characteristic pattern, with increased intraocular pressure (IOP) being the most important and only modifiable risk factor. The functional assessment of glaucoma has long relied on automated perimetry for detection of the glaucomatous visual field damage, which is a time-consuming and subjective test. The clinical application of visual field testing in glaucoma is limited by the fact that the relation between the proportional losses of ganglion cells and visual field function is nonlinear: 30%-50% percent of retinal ganglion cells may have died before the earliest visual field changes can be detected (Harwerth, Carter-Dawson et al. 1999). In late stage disease, it may be difficult to monitor disease progression in advanced glaucoma when patients’ visual function becomes too poor to perform perimetry testing. There is a need to improve the functional assessment of retinal ganglion cell loss in glaucoma.

103

PIPR is an in vivo index of melanopsin-driven intrinsic photoactivity of ipRGCs. Chromatic pupillometry measurement of PIPR may offer a new objective assessment of a novel aspect of retinal ganglion cell function in glaucoma. However, very little is known about the ipRGC involvement in glaucoma. Evidence from flat-mounted rat retinas indicated that surgically induced ocular hypertension for 12 weeks does not cause significant ipRGC morphologic changes or cell loss, whereas non–melanopsin-labeled superior colliculus- projecting retinal ganglion cells do exhibit significant loss (Li, Chen et al. 2006). It was inferred that ipRGCs are more resistant to hypertension injury (Li, Chen et al. 2006). In another similar study, reduced ipRGC density was observed, but there was no change in soma size and dendritic morphology in the remaining ipRGCs (Wang, Lu et al. 2008). There is, however, anecdotal evidence showing no immunoreactivity of the ipRGC photopigment melanopsin in one patient with longstanding glaucoma (Hannibal, Hindersson et al. 2004). Moreover, patients with advanced glaucoma show reduced melatonin suppression, indicating that ipRGC function may be affected (Pérez-Rico, de la Villa et al. 2010). Recently Kankipati et al. (Kankipati, Girkin et al. 2011) demonstrated reduced ipRGC-mediated PIPR in advanced glaucoma using chromatic pupillometry. In a separate study, Feigl et al (Feigl, Mattes et al. 2011) also showed that PIPR was reduced in advanced glaucoma, but not in early glaucoma. These studies indicated that PIPR has potential to become a clinical indicator of progressive changes in glaucoma.

Glaucoma does not result in uniform loss of ganglion cells across the retina, rather the damage is often asymmetric, with the lower half of the retina more severely affected than the upper half (Jakobs, Libby et al. 2005). Glaucoma hemifield analysis that tests for asymmetry between the superior and inferior visual fields is a commonly used perimetry test that enhances early detection of the disease (Asman and Heijl 1992). However, hemifield comparison of ipRGC photosensitivity has not been investigated in glaucoma patients. In this ongoing study, we are investigating the melanopsin-driven PIPR in normal controls and patients with glaucoma of variable severity using the optimized hemifield, central-field and full-field PIPR testing protocols developed in Chapter 3 of this thesis. We hypothesize that early glaucoma will show asymmetric hemifield PIPR, thus allowing early detection of ipRGC impairment in glaucoma. We predict reduced central-field and full-field PIPR will be seen in intermediate and advanced glaucoma.

104

We will also investigate the correlation between PIPR and conventional structural and functional assessments in glaucoma, such as the thickness of the retinal ganglion cell layer and visual field function.

7.1.2 Methods

Clinical records and perimetry databases from a glaucoma clinic are being reviewed to identify patients ranging in age from 17-70 with an established diagnosis of primary open angle glaucoma (POAG), low tension glaucoma (LTG), normal tension glaucoma (NTG) or juvenile open angle glaucoma (JOAG). Normal adult control participants are being recruited through advertising flyers. Information including age, gender, refractive error and ocular history are recorded for each participant. For patients, their visual fields are also being obtained from the perimetry database. All participants undergo an ocular examination, where visual acuity, intraocular pressure, iris structures, status of lens and retina are assessed. Any individual who has any of the following conditions in one or both eyes is excluded: 1. Intraocular surgery except uncomplicated cataract extraction and IOL implant, uncomplicated trabeculectomy or glaucoma valve implant > 1 year 2. Glaucoma suspect or ocular hypertension 3. Currently use of tamsulosin 4. Currently use of mydriatic or miotic eye drops 5. Refractive error > 5 diopters 6. Eye trauma 7. Optic neuropathy unrelated to glaucoma 8. Secondary glaucoma (pseudoexfoliation, pigment dispersion glaucoma, neovascular glaucoma) 9. Shallow anterior chamber 10. Angle closure glaucoma 11. Uveitis 12. Diabetic retinopathy 13. Retinal detachment 14. Lens opacity score>2 (for nuclear, subcapsular, and cortical cataracts)

105

15. History of CNVM/injections, or geographic atrophy causing <20/200 vision 16. Any other conditions that may compromise the integrity of the pupillary light reflex pathway upstream of the retina

Glaucoma patients are sub-stratified into 4 sub-groups according to the severity of their glaucomatous visual field damage as measured by mean deviation (MD) on Humphrey perimetry.

 Early Glaucoma: MD < -6.00 dB

 Moderate Glaucoma: MD of -6.00 to -12.00 dB

 Advanced Glaucoma: MD of -12.01 to -20.00 dB

 End-Stage Glaucoma: Unable to perform Humphrey visual fields attributable to central scotoma or visual acuity of 20/200 or worse, attributable to primary open-angle glaucoma.

We aim to recruit 20 patients from each of the 4 sub-groups and 20 age-matched control participants. PIPR from upper and lower hemifield, central-field and full-field is being measured using the apparatus and protocols described in Chapter 3.

7.1.3 Preliminary testing data

Despite that the number of tested subjects in this on-going study is still too small to draw any scientific conclusions, preliminary individual data have demonstrated some interesting patterns of PIPR changes in glaucoma. Sample data from glaucoma patients are demonstrated in comparison to normal controls in Figure 16. The data show that PIPR waveforms in a patient with early glaucoma are similar to that of the normal control, suggesting normal ipRGC photoactivity in early glaucoma; Hemifield and central-field PIPR is severely reduced in a patient with advanced glaucoma, but full-field PIPR is much less impacted. Interestingly, the maximum pupil constriction to red stimulation is normal, indicating non-linear relation between intrinsic and extrinsic ipRGC photoactivity. Diminished PIPR was recorded in a patient with

106

end-stage glaucoma, suggesting that both intrinsic and extrinsic ipRGC photoactivity are severely impaired.

These data were sampled from a few individual patients, so they should be interpreted with caution. More subjects are needed to properly evaluate the utility of PIPR testing as an objective functional assessment for glaucoma.

Figure 16. Individual PIPR testing results from visually normal participants and participants with glaucoma. (a) Hemifield, central-field and full-field PIPR in a normal

107

participant. (b) PIPR waveforms in a patient with early glaucoma are similar to those of the normal control, suggesting normal ipRGC photoactivity in early glaucoma. (c) Hemifield and central-field PIPR is severely reduced in a patient with advanced glaucoma, but full-field PIPR is much less impacted. Interestingly, the maximum pupil constriction to red stimulation is normal, indicating non-linear relation between intrinsic and extrinsic ipRGC photoactivity. (d) Diminished PIPR in a patient with end-stage glaucoma, suggesting that both intrinsic and extrinsic ipRGC photoactivity are severely impaired.

7.2 Using chromatic pupillometry to investigate retinal dystrophies

Besides inducing and measuring melanopsin-driven PIPR, chromatic pupillometry can also evaluate rod and cone input to the ipRGC pathway (Park, Moura et al. 2011). It can potentially be an adjunct to ERG testing in the diagnosis and monitoring of inherited retinal dystrophies (Park, Moura et al. 2011; Kawasaki, Crippa et al. 2012). As an objective assessment of retinal function in these disease populations, chromatic pupillometry may be particularly useful in the two following situations. Firstly, in the advanced stage of progressive degenerative outer retinal dystrophies, such as retinitis pigmentosa, ERG is non-recordable despite that patients may still have a moderate number of functioning rods/cones and preserved ipRGCs (Kardon, Anderson et al. 2011; Kawasaki, Crippa et al. 2012; Kawasaki, Munier et al. 2012). In this situation, chromatic pupillometry evaluating rod, cone and melanopsin contributions to the pupillary light reflex could serve as a substitute for ERG to monitor disease progression into the end stage. Secondly, there are many clinical phenomena suggesting that certain subtypes of congenital retinal dystrophies may have an abnormal pattern of circuitry in the ipRGC pathway. For example, a paradoxical pupil constriction at the offset of light stimulation is observed clinically in patients with congenital stationary night blindness (CSNB) and achromatopsia (Barricks, Flynn et al. 1977; Flynn, Kazarian et al. 1981; Price, Thompson et al. 1985; Simon, Abraham et al. 2004). The mechanism for this strange pupillary light reflex is still unclear. Based on the updated knowledge that the afferent signal for the pupillary light reflex is mediated by

108

ipRGCs integrating rod/cone-driven extrinsic photoactivity and their own melanopsin-driven intrinsic photoactivity, this paradoxical pupil response may suggest abnormally enhanced rod or cone input via the OFF-bipolar cells to ipRGCs. Furthermore, photophobia is a prominent clinical feature in many retinal dystrophies (Digre and Brennan 2012) such as achromatopsia, (Simon, Abraham et al. 2004) bradyopsia (Kuburas, Thompson et al. 2014) and Leber’s congenital amaurosis (Weleber, Francis et al. 2013). There is growing evidence suggesting that the ipRGC-mediated irradiance measuring pathway consists of the afferent limb of photophobic sensation (Noseda, Kainz et al. 2009; Noseda and Burstein 2011; Digre and Brennan 2012). Meanwhile, abnormally enhanced and prolonged rod or cone-driven pupil responses have also been reported in patients with achromatopsia (Lisowska, Lisowski et al. 2015) and in an animal model of bradyopsia (Kuburas, Thompson et al. 2014). Collectively, these findings suggest that abnormally enhanced rod/cone input to the ipRGC pathway may cause photophobia in congenital retinal dystrophies. Chromatic pupillometry may be a useful tool to reveal the putative abnormal retinal circuitry in the ipRGC irradiance measuring pathway in these disease populations.

In preparation for investigating retinal dystrophies, we have created a comprehensive chromatic pupillometry protocol to investigate the contributions of rod/cone-driven extrinsic ipRGC activity and melanopsin-driven intrinsic ipRGC activity to the pupillary light reflex: After 30 minutes of dark adaptation, alternating dim blue and red flashes (0.1 cd/m2, 4 ms) will be presented in 15 seconds intervals to induce dark-adapted (scotopic) rod and cone-driven pupil responses, followed by 15 s of exposure to alternating red and blue light of increasing intensity (0.1, 1, 10, 100 cd/m2) to assess the light-ON and light-OFF pupil responses, then bright red and blue flashes (400 cd/m2, 400 ms) will be used to induce the light-adapted (photopic) cone response and melanopsin-driven PIPR. (Figure 17) This study is currently in the pilot-testing stage to characterize the chromatic pupil response in patients with retinal dystrophies.

109

Figure 17. Waveforms of comprehensive chromatic pupillometry testing. (a) pupil responses induced by 0.1 cd/m2, 400 ms dim blue and red light stimuli presented to dark- adapted eyes, assessing the scotopic rod and cone-driven responses. (b) Pupil responses stimulated by 400 cd/m2, 400 ms bright red and blue flashes to light-adapted eyes, assessing the photopic cone-driven response and the melanopsin-driven PIPR. (c) Step-wise long exposures to alternating red and blue light inducing the light-ON and light-OFF pupil responses.

110

References

Altimus, C. M., A. D. Güler, et al. (2010). "Rod photoreceptors drive circadian photoentrainment across a wide range of light intensities." Nature neuroscience 13(9): 1107-1112.

Arendt, D. (2003). "Evolution of eyes and photoreceptor cell types." International Journal of Developmental Biology 47(7/8): 563-572.

Arendt, D., K. Tessmar-Raible, et al. (2004). "Ciliary photoreceptors with a vertebrate-type opsin in an invertebrate brain." Science 306(5697): 869-871.

Asman, P. and A. Heijl (1992). "Glaucoma Hemifield Test. Automated visual field evaluation." Archives of ophthalmology 110(6): 812-819.

Augood, C. A., J. R. Vingerling, et al. (2006). "Prevalence of age-related maculopathy in older Europeans: the European Eye Study (EUREYE)." Archives of ophthalmology 124(4): 529-535.

Barnard, A. R., J. M. Appleford, et al. (2004). "Residual photosensitivity in mice lacking both rod opsin and cone photoreceptor cyclic nucleotide gated channel 3 [alpha] subunit." Visual neuroscience 21(05): 675-683.

Barrett, P. (2001). "Assessing the reliability of rating data." Retrieved March 25: 2002.

Barricks, M. E., J. T. Flynn, et al. (1977). "Paradoxical pupillary responses in congenital stationary night blindness." Archives of ophthalmology 95(10): 1800-1804.

111

Baver, S. B., G. E. Pickard, et al. (2008). "Two types of melanopsin retinal ganglion cell differentially innervate the hypothalamic suprachiasmatic nucleus and the olivary pretectal nucleus." European Journal of Neuroscience 27(7): 1763-1770.

Belenky, M. A., C. A. Smeraski, et al. (2003). "Melanopsin retinal ganglion cells receive bipolar and amacrine cell synapses." Journal of Comparative Neurology 460(3): 380-393.

Berson, D. M. (2003). "Strange vision: ganglion cells as circadian photoreceptors." Trends in neurosciences 26(6): 314-320.

Berson, D. M., A. M. Castrucci, et al. (2010). "Morphology and mosaics of melanopsin‐ expressing retinal ganglion cell types in mice." Journal of Comparative Neurology 518(13): 2405-2422.

Berson, D. M., F. A. Dunn, et al. (2002). "Phototransduction by retinal ganglion cells that set the circadian clock." Science 295(5557): 1070-1073.

Brainard, G. C., J. P. Hanifin, et al. (2001). "Action spectrum for melatonin regulation in humans: evidence for a novel circadian photoreceptor." The journal of Neuroscience 21(16): 6405-6412.

Brown, T. M. and H. D. Piggins (2007). "Electrophysiology of the suprachiasmatic circadian clock." Progress in neurobiology 82(5): 229-255.

Brown, T. M., J. Wynne, et al. (2011). "Multiple hypothalamic cell populations encoding distinct visual information." The Journal of physiology 589(5): 1173-1194.

Chang, D. S., K. S. Arora, et al. (2013). "Development and Validation of an Associative Model for the Detection of Glaucoma Using Pupillography." American journal of ophthalmology 156(6): 1285-1296. e1282.

112

Crespo, D., D. D. O'Leary, et al. (1985). "Changes in the numbers of optic nerve fibers during late prenatal and postnatal development in the albino rat." Developmental Brain Research 19(1): 129-134.

Dacey, D. M., H.-W. Liao, et al. (2004). "Melanopsin-expressing ganglion cells in primate retina project to the LGN and signal both color and irradiance." Journal of Vision 4(11): 19-19.

Dacey, D. M., H.-W. Liao, et al. (2005). "Melanopsin-expressing ganglion cells in primate retina signal colour and irradiance and project to the LGN." Nature 433(7027): 749-754.

Dacey, D. M., B. B. Peterson, et al. (2003). "Fireworks in the primate retina: in vitro photodynamics reveals diverse LGN-projecting ganglion cell types." Neuron 37(1): 15-27.

Davignon, I., M. Barnard, et al. (1996). "Gene Structure of MurineGna11andGna15: Tandemly Duplicated Gq Class G Protein α Subunit Genes." Genomics 31(3): 359-366.

DeLeón-Ortega, J., K. E. Carroll, et al. (2007). "Correlations between retinal nerve fiber layer and visual field in eyes with nonarteritic anterior ischemic optic neuropathy." American journal of ophthalmology 143(2): 288-294. e282.

Demb, J. B. (2008). "Functional circuitry of visual adaptation in the retina." The Journal of physiology 586(18): 4377-4384.

Digre, K. B. and K. Brennan (2012). "Shedding light on photophobia." Journal of neuro- ophthalmology: the official journal of the North American Neuro-Ophthalmology Society 32(1): 68.

113

Do, M. T. H., S. H. Kang, et al. (2009). "Photon capture and signalling by melanopsin retinal ganglion cells." Nature 457(7227): 281-287.

Do, M. T. H. and K.-W. Yau (2010). "Intrinsically photosensitive retinal ganglion cells." Physiological reviews 90(4): 1547-1581.

Dumitrescu, O. N., F. G. Pucci, et al. (2009). "Ectopic retinal ON bipolar cell synapses in the OFF inner plexiform layer: contacts with dopaminergic amacrine cells and melanopsin ganglion cells." Journal of Comparative Neurology 517(2): 226-244.

Ebihara, S. and K. Tsuji (1980). "Entrainment of the circadian activity rhythm to the light cycle: effective light intensity for a Zeitgeber in the retinal degenerate C3H mouse and the normal C57BL mouse." Physiology & behavior 24(3): 523-527.

Ebrey, T. and Y. Koutalos (2001). "Vertebrate photoreceptors." Progress in retinal and eye research 20(1): 49-94.

Ecker, J. L., O. N. Dumitrescu, et al. (2010). "Melanopsin-expressing retinal ganglion-cell photoreceptors: cellular diversity and role in pattern vision." Neuron 67(1): 49-60.

Estevez, M. E., P. M. Fogerson, et al. (2012). "Form and function of the M4 cell, an intrinsically photosensitive retinal ganglion cell type contributing to geniculocortical vision." The Journal of Neuroscience 32(39): 13608-13620.

Fain, G. L., R. Hardie, et al. (2010). "Phototransduction and the evolution of photoreceptors." Current Biology 20(3): R114-R124.

Feigl, B., D. Mattes, et al. (2011). "Intrinsically photosensitive (melanopsin) retinal ganglion cell function in glaucoma." Investigative ophthalmology & visual science 52(7): 4362-4367.

114

Feigl, B. and A. J. Zele (2014). "Melanopsin-expressing intrinsically photosensitive retinal ganglion cells in retinal disease." Optometry & Vision Science 91(8): 894-903.

Flynn, J. T., E. Kazarian, et al. (1981). "Paradoxical pupil in congenital achromatopsia." International ophthalmology 3(2): 91-96.

Foster, R., I. Provencio, et al. (1991). "Circadian photoreception in the retinally degenerate mouse (rd/rd)." Journal of Comparative Physiology A 169(1): 39-50.

Freedman, M. S., R. J. Lucas, et al. (1999). "Regulation of mammalian circadian behavior by non-rod, non-cone, ocular photoreceptors." Science 284(5413): 502-504.

Fu, Y., H.-W. Liao, et al. (2005). "Non-image-forming ocular photoreception in vertebrates." Current opinion in neurobiology 15(4): 415-422.

Fu, Y. and K.-W. Yau (2007). "Phototransduction in mouse rods and cones." Pflügers Archiv- European Journal of Physiology 454(5): 805-819.

Gamlin, P. D., D. H. McDougal, et al. (2007). "Human and macaque pupil responses driven by melanopsin-containing retinal ganglion cells." Vision research 47(7): 946-954.

Gollisch, T. and M. Meister (2010). "Eye smarter than scientists believed: neural computations in circuits of the retina." Neuron 65(2): 150-164.

Gooley, J. J., J. Lu, et al. (2001). "Melanopsin in cells of origin of the retinohypothalamic tract." Nature neuroscience 4(12): 1165-1165.

115

Gooley, J. J., J. Lu, et al. (2003). "A broad role for melanopsin in nonvisual photoreception." The Journal of Neuroscience 23(18): 7093-7106.

Göz, D., K. Studholme, et al. (2008). "Targeted destruction of photosensitive retinal ganglion cells with a saporin conjugate alters the effects of light on mouse circadian rhythms." PloS one 3(9): e3153.

Graham, D. M., K. Y. Wong, et al. (2008). "Melanopsin ganglion cells use a membrane- associated rhabdomeric phototransduction cascade." Journal of neurophysiology 99(5): 2522- 2532.

Güler, A. D., J. L. Ecker, et al. (2008). "Melanopsin cells are the principal conduits for rod–cone input to non-image-forming vision." Nature 453(7191): 102-105.

Hannibal, J., P. Hindersson, et al. (2002). "The photopigment melanopsin is exclusively present in pituitary adenylate cyclase-activating polypeptide-containing retinal ganglion cells of the retinohypothalamic tract." The Journal of neuroscience: the official journal of the Society for Neuroscience 22(1): RC191-RC191.

Hannibal, J., P. Hindersson, et al. (2004). "Melanopsin is expressed in PACAP-containing retinal ganglion cells of the human retinohypothalamic tract." Investigative ophthalmology & visual science 45(11): 4202-4209.

Hardie, R. and M. Postma (2008). "1.05 Phototransduction in Microvillar Photoreceptors of Drosophila and Other Invertebrates."

Hardie, R. C. (2001). "Phototransduction in Drosophila melanogaster." Journal of Experimental Biology 204(20): 3403-3409.

116

Hart, W. M. and B. Becker (1982). "The onset and evolution of glaucomatous visual field defects." Ophthalmology 89(3): 268-279.

Hartwick, A. T., J. R. Bramley, et al. (2007). "Light-evoked calcium responses of isolated melanopsin-expressing retinal ganglion cells." The Journal of Neuroscience 27(49): 13468- 13480.

Harwerth, R. S., L. Carter-Dawson, et al. (1999). "Ganglion cell losses underlying visual field defects from experimental glaucoma." Investigative ophthalmology & visual science 40(10): 2242-2250.

Hatori, M., H. Le, et al. (2008). "Inducible ablation of melanopsin-expressing retinal ganglion cells reveals their central role in non-image forming visual responses." PloS one 3(6): e2451.

Hattar, S., M. Kumar, et al. (2006). "Central projections of melanopsin‐expressing retinal ganglion cells in the mouse." Journal of Comparative Neurology 497(3): 326-349.

Hattar, S., H.-W. Liao, et al. (2002). "Melanopsin-containing retinal ganglion cells: architecture, projections, and intrinsic photosensitivity." Science 295(5557): 1065-1070.

Hattar, S., R. J. Lucas, et al. (2003). "Melanopsin and rod–cone photoreceptive systems account for all major accessory visual functions in mice." Nature 424(6944): 75-81.

Helfrich-Förster, C., C. Winter, et al. (2001). "The circadian clock of fruit flies is blind after elimination of all known photoreceptors." Neuron 30(1): 249-261.

Herbst, K., B. Sander, et al. (2012). "Intrinsically photosensitive retinal ganglion cell function in relation to age: a pupillometric study in humans with special reference to the age-related optic properties of the lens." BMC ophthalmology 12(1): 4.

117

Herbst, K., B. Sander, et al. (2013). "Unilateral anterior ischemic optic neuropathy: chromatic pupillometry in affected, fellow non-affected and healthy control eyes." Frontiers in neurology 4.

Herbst, K., B. Sander, et al. (2011). "Test–retest repeatability of the pupil light response to blue and red light stimuli in normal human eyes using a novel pupillometer." Frontiers in neurology 2.

Hillman, P., S. Hochstein, et al. (1983). "Transduction in invertebrate photoreceptors: role of pigment bistability." Physiol Rev 63(2): 668-772.

Hoshi, H., W.-L. Liu, et al. (2009). "ON inputs to the OFF layer: bipolar cells that break the stratification rules of the retina." The Journal of Neuroscience 29(28): 8875-8883.

Hubbard, K. B. and J. R. Hepler (2006). "Cell signalling diversity of the Gqα family of heterotrimeric G proteins." Cellular signalling 18(2): 135-150.

Jakobs, T. C., R. T. Libby, et al. (2005). "Retinal ganglion cell degeneration is topological but not cell type specific in DBA/2J mice." The Journal of cell biology 171(2): 313-325.

Johnson, J., V. Wu, et al. (2010). "Melanopsin-dependent light avoidance in neonatal mice." Proceedings of the National Academy of Sciences 107(40): 17374-17378.

Kahneman, D. (1973). Attention and effort, Citeseer.

Kahneman, D. and J. Beatty (1966). "Pupil diameter and load on memory." Science 154(3756): 1583-1585.

118

Kankipati, L., C. A. Girkin, et al. (2010). "Post-illumination pupil response in subjects without ocular disease." Investigative ophthalmology & visual science 51(5): 2764.

Kankipati, L., C. A. Girkin, et al. (2011). "The post-illumination pupil response is reduced in glaucoma patients." Investigative ophthalmology & visual science 52(5): 2287.

Kardon, R., S. C. Anderson, et al. (2009). "Chromatic pupil responses: preferential activation of the melanopsin-mediated versus outer photoreceptor-mediated pupil light reflex." Ophthalmology 116(8): 1564-1573.

Kardon, R., S. C. Anderson, et al. (2011). "Chromatic pupillometry in patients with retinitis pigmentosa." Ophthalmology 118(2): 376-381.

Kawasaki, A., S. V. Crippa, et al. (2012). "Characterization of pupil responses to blue and red light stimuli in autosomal dominant retinitis pigmentosa due to NR2E3 mutation." Investigative ophthalmology & visual science 53(9): 5562-5569.

Kawasaki, A. and R. H. Kardon (2007). "Intrinsically photosensitive retinal ganglion cells." Journal of Neuro-ophthalmology 27(3): 195-204.

Kawasaki, A., F. L. Munier, et al. (2012). "Pupillometric quantification of residual rod and cone activity in leber congenital amaurosis." Archives of ophthalmology 130(6): 798-800.

Keeler, C. E. (1927). "Iris movements in blind mice." American Journal of Physiology--Legacy Content 81(1): 107-112.

119

Keshavarzi, S., R. K. Sullivan, et al. (2014). "Functional properties and projections of neurons in the medial amygdala." The Journal of Neuroscience 34(26): 8699-8715.

Klerman, E., T. Shanahan, et al. (2002). "Photic resetting of the human circadian pacemaker in the absence of conscious vision." Journal of biological rhythms 17(6): 548-555.

Kolb, H. and R. Nelson (1995). The organization of photoreceptor to bipolar synapses in the outer plexiform layer. Neurobiology and clinical aspects of the outer retina, Springer: 273-296.

Koyanagi, M., K. Kubokawa, et al. (2005). "Cephalochordate melanopsin: evolutionary linkage between invertebrate visual cells and vertebrate photosensitive retinal ganglion cells." Current Biology 15(11): 1065-1069.

Kuburas, A., S. Thompson, et al. (2014). "Photophobia and Abnormally Sustained Pupil Responses in a Mouse Model of Bradyopsia." Investigative ophthalmology & visual science 55(10): 6878.

Lall, G. S., V. L. Revell, et al. (2010). "Distinct contributions of rod, cone, and melanopsin photoreceptors to encoding irradiance." Neuron 66(3): 417-428.

Lei, S., H. C. Goltz, et al. (2014). "Full-field Chromatic Pupillometry for the Assessment of the Postillumination Pupil Response Driven by Melanopsin-Containing Retinal Ganglion Cells." Investigative ophthalmology & visual science 55(7): 4496-4503.

Lei, S., H. C. Goltz, et al. (2015). "Test–retest reliability of hemifield, central-field, and full-field chromatic pupillometry for assessing the function of melanopsin-containing retinal ganglion cells." Invest Ophthalmol Vis Sci 56: 1267-1273.

120

Li, R. S., B.-Y. Chen, et al. (2006). "Melanopsin-expressing retinal ganglion cells are more injury-resistant in a chronic ocular hypertension model." Investigative ophthalmology & visual science 47(7): 2951-2958.

Lin, B., S. W. Wang, et al. (2004). "Retinal ganglion cell type, size, and spacing can be specified independent of homotypic dendritic contacts." Neuron 43(4): 475-485.

Lisowska, J., L. Lisowski, et al. (2015). "Chromatic pupillography in patients with Achromatopsia." Investigative ophthalmology & visual science 56(7): 575-575.

Lockley, S. W., G. C. Brainard, et al. (2003). "High sensitivity of the human circadian melatonin rhythm to resetting by short wavelength light." J Clin Endocrinol Metab 88(9): 4502-4505.

Lucas, R., S. Hattar, et al. (2003). "Diminished pupillary light reflex at high irradiances in melanopsin-knockout mice." Science 299(5604): 245-247.

Lucas, R. J. (2006). "Chromophore regeneration: melanopsin does its own thing." Proceedings of the National Academy of Sciences 103(27): 10153-10154.

Lucas, R. J. (2013). "Mammalian inner retinal photoreception." Current Biology 23(3): R125- R133.

Lucas, R. J., R. H. Douglas, et al. (2001). "Characterization of an ocular photopigment capable of driving pupillary constriction in mice." Nature neuroscience 4(6): 621-626.

Lucas, R. J., M. S. Freedman, et al. (2001). "Identifying the photoreceptive inputs to the mammalian circadian system using transgenic and retinally degenerate mice." Behavioural brain research 125(1): 97-102.

121

MacLennan, R. N. (1993). "Interrater reliability with SPSS for Windows 5.0." The American Statistician 47(4): 292-296.

Maddess, T., Y.-L. Ho, et al. (2011). "Multifocal pupillographic perimetry with white and colored stimuli." Journal of glaucoma 20(6): 336-343.

Matsuyama, T., T. Yamashita, et al. (2012). "Photochemical properties of mammalian melanopsin." Biochemistry 51(27): 5454-5462.

Mawad, K. and R. N. Van Gelder (2008). "Absence of long-wavelength photic potentiation of murine intrinsically photosensitive retinal ganglion cell firing in vitro." Journal of biological rhythms 23(5): 387-391.

McDougal, D. H. and P. D. Gamlin (2010). "The influence of intrinsically-photosensitive retinal ganglion cells on the spectral sensitivity and response dynamics of the human pupillary light reflex." Vision research 50(1): 72-87.

McLaren, J. W., J. C. Erie, et al. (1992). "Computerized analysis of pupillograms in studies of alertness." Invest. Ophthalmol. Vis. Sci 33(3): 671-676.

Montell, C. (2005). "The TRP superfamily of cation channels." Science Signaling 2005(272): re3-re3.

Morin, L. P., J. H. Blanchard, et al. (2003). "Retinal ganglion cell projections to the hamster suprachiasmatic nucleus, intergeniculate leaflet, and visual midbrain: bifurcation and melanopsin immunoreactivity." Journal of Comparative Neurology 465(3): 401-416.

Mrosovsky, N. (1999). "Masking: history, definitions, and measurement." Chronobiology international 16(4): 415-429.

122

Mrosovsky, N. and S. Hattar (2003). "Impaired masking responses to light in melanopsin- knockout mice." Chronobiology international 20(6): 989-999.

Münch, M., L. Léon, et al. (2012). "Circadian and wake-dependent effects on the pupil light reflex in response to narrow-bandwidth light pulses." Investigative ophthalmology & visual science 53(8): 4546-4555.

Münch, M., L. Léon, et al. (2012). "Circadian and wake-dependent effects on the pupil light reflex in response to narrow-bandwidth light pulses." Investigative Ophthalmology and Visual Science 53(8): 4546-4555.

Mure, L. S., P.-L. Cornut, et al. (2009). "Melanopsin bistability: a fly's eye technology in the human retina." PLoS One 4(6): e5991.

Mure, L. S., C. Rieux, et al. (2007). "Melanopsin-dependent nonvisual responses: evidence for photopigment bistability in vivo." Journal of biological rhythms 22(5): 411-424.

Nelson, R. J. and I. Zucker (1981). "Absence of extraocular photoreception in diurnal and nocturnal rodents exposed to direct sunlight." Comparative Biochemistry and Physiology Part A: Physiology 69(1): 145-148.

Nissen, C., B. Sander, et al. (2011). "The effect of pupil size on stimulation of the melanopsin containing retinal ganglion cells, as evaluated by monochromatic pupillometry." Frontiers in neurology 2.

Nissen, C., B. Sander, et al. (2014). "Monochromatic pupillometry in unilateral glaucoma discloses no adaptive changes subserved by the ipRGCs." Frontiers in neurology 5.

123

Noseda, R. and R. Burstein (2011). "Advances in understanding the mechanisms of migraine- type photophobia." Current opinion in neurology 24(3): 197-202.

Noseda, R., V. Kainz, et al. (2009). A retinal-thalamic pathway for photophobia during migraine. Cephalalgia, WILEY-BLACKWELL PUBLISHING, INC COMMERCE PLACE, 350 MAIN ST, MALDEN 02148, MA USA.

Østergaard, J., J. Hannibal, et al. (2007). "Synaptic contact between melanopsin-containing retinal ganglion cells and rod bipolar cells." Investigative ophthalmology & visual science 48(8): 3812-3820.

Panda, S., I. Provencio, et al. (2003). "Melanopsin is required for non-image-forming photic responses in blind mice." Science 301(5632): 525-527.

Panda, S., T. K. Sato, et al. (2002). "Melanopsin (Opn4) requirement for normal light-induced circadian phase shifting." Science 298(5601): 2213-2216.

Park, J. C., A. L. Moura, et al. (2011). "Toward a clinical protocol for assessing rod, cone, and melanopsin contributions to the human pupil response." Investigative ophthalmology & visual science 52(9): 6624.

Peirson, S. N., S. Halford, et al. (2009). "The evolution of irradiance detection: melanopsin and the non-visual opsins." Philosophical Transactions of the Royal Society B: Biological Sciences 364(1531): 2849-2865.

Peirson, S. N., H. Oster, et al. (2007). "Microarray analysis and functional genomics identify novel components of melanopsin signaling." Current Biology 17(16): 1363-1372.

124

Pérez-Rico, C., P. de la Villa, et al. (2010). "Evaluation of functional integrity of the retinohypothalamic tract in advanced glaucoma using multifocal electroretinography and light- induced melatonin suppression." Experimental eye research 91(5): 578-583.

Pianta, M. J. and M. Kalloniatis (2000). "Characterisation of dark adaptation in human cone pathways: an application of the equivalent background hypothesis." The Journal of physiology 528(3): 591-608.

Price, M. J., H. S. Thompson, et al. (1985). "Pupillary constriction to darkness." British journal of ophthalmology 69(3): 205-211.

Provencio, I., G. Jiang, et al. (1998). "Melanopsin: An opsin in melanophores, brain, and eye." Proceedings of the National Academy of Sciences 95(1): 340-345.

Provencio, I., I. R. Rodriguez, et al. (2000). "A novel human opsin in the inner retina." The journal of Neuroscience 20(2): 600-605.

Provencio, I., M. D. Rollag, et al. (2002). "Anatomy: Photoreceptive net in the mammalian retina." Nature 415(6871): 493-493.

Qiu, X. and D. Berson (2007). "Melanopsin bistability in ganglion-cell photoreceptors." Invest Ophthalmol Vis Sci 48: 612.

Rajaraman, K. (2012). "ON ganglion cells are intrinsically photosensitive in the tiger salamander retina." Journal of Comparative Neurology 520(1): 200-210.

Ritter, E., K. Zimmermann, et al. (2004). "Transition of rhodopsin into the active metarhodopsin II state opens a new light-induced pathway linked to Schiff base isomerization." Journal of Biological Chemistry 279(46): 48102-48111.

125

Roecklein, K., P. Wong, et al. (2013). "The post illumination pupil response is reduced in seasonal affective disorder." Psychiatry research 210(1): 150-158.

Rollag, M. D. (2008). "Does melanopsin bistability have physiological consequences?" Journal of biological rhythms 23(5): 396-399.

Ruby, N. F., T. J. Brennan, et al. (2002). "Role of melanopsin in circadian responses to light." Science 298(5601): 2211-2213.

Schmidt, T. M., N. M. Alam, et al. (2014). "A role for melanopsin in alpha retinal ganglion cells and contrast detection." Neuron 82(4): 781-788.

Schmidt, T. M., S.-K. Chen, et al. (2011). "Intrinsically photosensitive retinal ganglion cells: many subtypes, diverse functions." Trends in neurosciences 34(11): 572-580.

Schmidt, T. M., M. T. H. Do, et al. (2011). "Melanopsin-positive intrinsically photosensitive retinal ganglion cells: from form to function." The Journal of Neuroscience 31(45): 16094- 16101.

Schmidt, T. M. and P. Kofuji (2009). "Functional and morphological differences among intrinsically photosensitive retinal ganglion cells." The Journal of neuroscience 29(2): 476-482.

Schmidt, T. M. and P. Kofuji (2010). "Differential cone pathway influence on intrinsically photosensitive retinal ganglion cell subtypes." The Journal of neuroscience 30(48): 16262- 16271.

126

Schmidt, T. M. and P. Kofuji (2011). "Structure and function of bistratified intrinsically photosensitive retinal ganglion cells in the mouse." Journal of Comparative Neurology 519(8): 1492-1504.

Schmidt, T. M., K. Taniguchi, et al. (2008). "Intrinsic and extrinsic light responses in melanopsin-expressing ganglion cells during mouse development." Journal of neurophysiology 100(1): 371-384.

Semo, M. a., S. Peirson, et al. (2003). "Melanopsin retinal ganglion cells and the maintenance of circadian and pupillary responses to light in aged rodless/coneless (rd/rd cl) mice." European Journal of Neuroscience 17(9): 1793-1801.

Sexton, T. J., M. Golczak, et al. (2012). "Melanopsin is highly resistant to light and chemical bleaching in vivo." Journal of Biological Chemistry 287(25): 20888-20897.

Shand, J. and R. Foster (1999). The extraretinal photoreceptors of non-mammalian vertebrates. Adaptive mechanisms in the ecology of vision, Springer: 197-222.

Shrout, P. E. and J. L. Fleiss (1979). "Intraclass correlations: uses in assessing rater reliability." Psychological bulletin 86(2): 420.

Simon, G. B., F. Abraham, et al. (2004). "Pingelapese achromatopsia: correlation between paradoxical pupillary response and clinical features." British journal of ophthalmology 88(2): 223-225.

Sollars, P. J., C. A. Smeraski, et al. (2003). "Melanopsin and non-melanopsin expressing retinal ganglion cells innervate the hypothalamic suprachiasmatic nucleus." Visual neuroscience 20(06): 601-610.

127

Sun, W. and P. J. May (2014). "Central pupillary light reflex circuits in the cat: I. The olivary pretectal nucleus." Journal of Comparative Neurology 522(18): 3960-3977.

Thapan, K., J. Arendt, et al. (2001). "An action spectrum for melatonin suppression: evidence for a novel non‐rod, non‐cone photoreceptor system in humans." The Journal of physiology 535(1): 261-267.

Thompson, C. L., W. S. Blaner, et al. (2001). "Preservation of light signaling to the suprachiasmatic nucleus in vitamin A-deficient mice." Proceedings of the National Academy of Sciences 98(20): 11708-11713.

Tsai, J. W., J. Hannibal, et al. (2009). "Melanopsin as a sleep modulator: circadian gating of the direct effects of light on sleep and altered sleep homeostasis in Opn4−/− mice." PLoS biology 7(6): e1000125.

Vigh, B., M. Manzano, et al. (2002). "Nonvisual photoreceptors of the deep brain, pineal organs and retina." Histology and histopathology 17(2): 555-590.

Vincent, W. and J. Weir (1994). Statistics in kinesiology, Human Kinetics.

Viney, T. J., K. Balint, et al. (2007). "Local retinal circuits of melanopsin-containing ganglion cells identified by transsynaptic viral tracing." Current Biology 17(11): 981-988.

Walker, M. T., R. L. Brown, et al. (2008). "Photochemistry of retinal chromophore in mouse melanopsin." Proceedings of the National Academy of Sciences 105(26): 8861-8865.

Wang, H.-z., Q.-j. Lu, et al. (2008). "Loss of melanopsin-containing retinal ganglion cells in a rat glaucoma model." Chinese Medical Journal (English Edition) 121(11): 1015.

128

Wang, T. and C. Montell (2007). "Phototransduction and retinal degeneration in Drosophila." Pflügers Archiv-European Journal of Physiology 454(5): 821-847.

Warren, E. J., C. N. Allen, et al. (2003). "Intrinsic light responses of retinal ganglion cells projecting to the circadian system." European Journal of Neuroscience 17(9): 1727-1735.

Warren, E. J., C. N. Allen, et al. (2006). "The light‐activated signaling pathway in SCN‐ projecting rat retinal ganglion cells." European Journal of Neuroscience 23(9): 2477-2487.

Wässle, H., U. Grünert, et al. (1990). "Retinal ganglion cell density and cortical magnification factor in the primate." Vision research 30(11): 1897-1911.

Weleber, R. G., P. J. Francis, et al. (2013). "Leber congenital amaurosis."

Wong, K. Y., F. A. Dunn, et al. (2005). "Photoreceptor adaptation in intrinsically photosensitive retinal ganglion cells." Neuron 48(6): 1001-1010.

Wong, K. Y., F. A. Dunn, et al. (2007). "Synaptic influences on rat ganglion‐cell photoreceptors." The Journal of physiology 582(1): 279-296.

Xue, T., M. Do, et al. (2011). "Melanopsin signalling in mammalian iris and retina." Nature 479(7371): 67-73.

Yamazaki, S., M. Goto, et al. (1999). "No evidence for extraocular photoreceptors in the circadian system of the Syrian hamster." Journal of biological rhythms 14(3): 197-201.

Yau, K.-W. and R. C. Hardie (2009). "Phototransduction motifs and variations." Cell 139(2): 246-264.

129

Yoshimura, T. and S. Ebihara (1996). "Spectral sensitivity of photoreceptors mediating phase- shifts of circadian rhythms in retinally degenerate CBA/J (rd/rd) and normal CBA/N (+/+) mice." Journal of Comparative Physiology A 178(6): 797-802.

Young, R. S. and J. Kennish (1993). "Transient and sustained components of the pupil response evoked by achromatic spatial patterns." Vision research 33(16): 2239-2252.

Zaidi, F. H., J. T. Hull, et al. (2007). "Short-wavelength light sensitivity of circadian, pupillary, and visual awareness in humans lacking an outer retina." Current Biology 17(24): 2122-2128.