Pupillary Light Reflex Reveals Melanopsin System Alteration in the Background of Myopia-26, the Female Limited Form of Early-Onset High Myopia

Abstract

Purpose: The purpose of this study was to evaluate pupillary light reflex (PLR) to chromatic flashes in patients with early-onset high-myopia (eoHM) without (myopic controls = M-CTRL) and with (female-limited myopia-26 = MYP-26) genetic mutations in the ARR3 gene encoding the cone arrestin.

Methods: Participants were 26 female subjects divided into 3 groups: emmetropic controls (E-CTRL, N = 12, mean age = 28.6 ± 7.8 years) and 2 myopic (M-CTRL, N = 7, mean age = 25.7 ± 11.5 years and MYP-26, N = 7, mean age = 28.3 ± 15.4 years) groups. In addition, one hemizygous carrier and one control male subject were examined. Direct PLRs were recorded after 10-minute dark adaptation. Stimuli were 1-second red (peak wavelength = 621 nm) and blue (peak wavelength = 470 nm) flashes at photopic luminance of 250 cd/m². A 2-minute interval between the flashes was introduced. Baseline pupil diameter (BPD), peak pupil constriction (PPC), and postillumination pupillary response (PIPR) were extracted from the PLR. Group comparisons were performed with ANOVAs.

Results: Dark-adapted BPD was comparable among the groups, whereas PPC to the red light was slightly reduced in patients with myopia (P = 0.02). PIPR at 6 seconds elicited by the blue flash was significantly weaker (P < 0.01) in female patients with MYP-26, whereas it was normal in the M-CTRL group and the asymptomatic male carrier.

Conclusions: L/M-cone abnormalities due to ARR3 gene mutation is currently claimed to underlie the pathological eye growth in MYP-26. Our results suggest that malfunction of the melanopsin system of intrinsically photosensitive retinal ganglion cells (ipRGCs) is specific to patients with symptomatic MYP-26, and may therefore play an additional role in the pathological eye growth of MYP-26.

Early onset high myopia (eoHM) is often the result of a single gene mutation which, in turn, displays a Mendelian form of inheritance. Among the genes most often involved in Mendelian eoHM, is the X-linked ARR3, encoding the cone arrestin. The resulting disease, sometimes referred to as Myopia-26 (MYP-26; MIM: 301010), was first reported in an Asian cohort¹ and was more recently documented in Caucasian families.^2,3 Counterintuitively, MYP-26 is a female-limited disease: heterozygous female patients display eoHM that progresses to a degenerative stage with aging, whereas hemizygous male patients for the mutant allele retain good vision. This disease is associated with myopic fundus alterations caused by the pathological eye elongation, but no abnormalities typical of cone dystrophy are present. Low visual acuity is found in female patients but not in male patients,^3–6 as in the cone arrestin null mouse model,⁷ whereas color vision deficiency is present in both hemizygous male and heterozygous female patients.^3,4 Importantly, the cone dysfunction indicated by the above symptoms is not a consequence of the eye elongation but seem to be directly caused by the ARR3 gene mutations.⁴ It may play a role in the pathological mechanism leading to high myopia (HM), but the gender-specific nature of this connection is still under investigation.⁴ 

In cones, ARR3 or cone arrestin (MIM: 301770), the product of the ARR3 gene, binds the light-activated phosphorylated receptors to shutoff opsin signaling during the phototransduction cascade. Therefore, its alteration affects the photoreceptor deactivation.^8,9 Gu et al. recently reported that patients with MYP-26 may show normal S-cone and abnormal L/M cone function, probably due to compensatory mechanisms of the rod arrestin expressed in S-cones.⁴ The authors raised a hypothesis to explain the female-limited profile of HM in MYP-26. They stated that the random X-chromosome inactivation causing a heterogeneous pattern of L/M-cone dysfunction may induce pathological eye growth due to the resulting constitutive contrast signal in the retina, whereas L/M-cones homogeneously affected in hemizygous male patients would not.⁴ 

In addition to its role in the phototransduction cascade, cone arrestin interacts with D2-like dopamine receptors, mainly the D4 receptor (DRD4), highly expressed in photoreceptors.¹⁰ DRD4 is involved in the circadian control of rod/cone gap junction coupling^11,12 and in the light-dependent intrinsic dopamine-melatonin retinal circadian clocks which, in turn, influence and are influenced by the function of melanopsin-expressing intrinsically photosensitive retinal ganglion cells (ipRGCs) in the vertebrate retina.^11,13 The ipRGC-melanopsin system disruption by either genetic^14,15 or environmental factors (i.e. light)^16–18 has been consistently reported to alter refractive development toward a myopic shift. The current assumption is that the melanopsin-driven endogenous circadian clock in the retina plays a crucial role in emmetropization.¹⁹ The potential influence of the melanopsin system on myopic development is also supported by the association between the prevalence of myopia and the time spent outdoors under natural sun light.^20,21 However, to establish a more complete picture on the influence of the time spent outdoors on myopic development, the circadian regulation of gene expression needs to be considered. Recently, a meta-analysis of a large population from European ancestry revealed that genetic factors controlling circadian rhythm are involved in the development of myopia and refractive errors.²² Overall, ARR3 gene mutations could prove to be an extreme example of genetic factors influencing ipRGC-melanopsin function, and thereby emmetropization. 

The assessment of ipRGC-melanopsin system integrity is feasible through the ipRGCs controlling the pupillary light reflex (PLR), by investigating the postillumination pupillary response (PIPR) specifically originating in the ipRGCs.^23–27 Whereas photoreceptors (cones and rods) sending inputs to ganglion cells are responsible for the pupillary constriction, melanopsin expressed in ipRGCs is the main generator of PIPR²⁴ under a spectral environment (short-wavelength) providing the optimal condition for melanopsin system activation in primates.²⁸ 

In this study, we recorded PLRs from female patients with MYP-26 with ARR3 gene mutations, from an asymptomatic male carrier, from patients with other forms of eoHM, and from emmetropic controls. 

Methods

The procedures were performed in accordance with the Declaration of Helsinki (1964) and its later amendments or comparable ethical standards. Written informed consent was obtained from all participants included in the study, or their legal representatives. The protocols and procedures have been approved by the National Scientific and Research Ethics Committee of the Medical Research Council of Hungary (registration number 58542-1/2017/EKU). 

Participants

The clinical and demographic characteristics of the patients with eoHM enrolled in this study are summarized in Table 1. Early onset was defined as myopia initiation between 2 and 3 years of age, based on patients’ anamnesis. HM was declared as spherical equivalent (SE) ≤−6.0 Diopters (D). Participants were 14 female patients with eoHM (see below), 12 female healthy volunteers, 1 male carrier, and 1 male control. All eoHM subjects underwent ophthalmological and genetic examinations partially reported in Széll et al. (2021).³ Refractive error was determined by non-cycloplegic autorefraction using an automated kerato-refractometer (except for patients under 18 years of age measured under cycloplegia). Best-corrected visual acuity (BCVA) was measured using trial lenses with the correction providing the best visual acuity. 

Table 1.

 

Characteristics of Patients Included in This Study

 

Seven female patients (mean age = 28.3 ± 15.6 years, mean SE = −10.25 ± 3.22 D), all related to each other (see the pedigree in Fig. 1) displayed ARR3 heterozygous mutant genotype (the MYP-26 group), with one of the alleles harboring an early nonsense mutation that completely inactivates the encoded protein (c.214C>T, p.Arg72*).³ The other seven female patients (mean age = 25.7 ± 11.5 years, mean SE = −9.55 ± 2.46 D) showed ARR3 homozygous wild type genotype (the myopic control [M-CTRL] group). Emmetropic controls (E-CTRLs) were age-matched (mean age = 28.6 ± 7.8 years) healthy female volunteers with negligible refractive errors (SE = −0.50 to +0.50 D), normal BCVA in both eyes, and absence of ocular or systemic diseases that could affect the central nervous system. In addition, results from 2 male subjects were included: one asymptomatic male subject member of the family of the inter-related ARR3 heterozygous women (37 years old, SE = −1.00/−2.00 D, subject III/3 in the pedigree) with confirmed hemizygous ARR3 gene mutation, and one age-matched control subject. 

Figure 1.

 

Pedigree of the family with X-linked female-limited eoHM (Myopia-26).

Two-Color Pupillary Light Reflex

In this study, we used the handheld RETeval portable system to record the two-color PLR, as recently reported by Asakawa et al.³⁰ This was a single session/single trial measurement, except for the male carrier who underwent PLR recordings twice in the same session. Figure 2A shows that the PLR protocol started with 10-minute dark-adaptation to allow the largest possible and stable baseline pupil diameter (BPD) to be recorded prior to light stimulation. After dark-adaptation, PLR to the red short-flash was recorded on the right eye, whereas the left eye was kept completely occluded. After 2 minutes of the dark-adaptation interval, the same measurement was performed on the left eye, while the right eye was covered. After another 2-minute dark-adapted interval, the same procedure was repeated for the blue short-flash stimulation. A dark-adaptation interval of 2 minutes was again applied between the measurements of the 2 eyes. If a large artifact or a measurement error was detected by the examiner, that condition was retested after a 2-minute dark-adaptation period. 

Figure 2.

 

Pupillary light reflex (PLR) protocol, origins, responses , and metrics. (A) The PLR protocol applied alternating dark-adaptation (black) and recording (grey) periods. The latter lasted 1 minute each, and comprised a 1 second red or blue light flash, delivered 10 seconds after the start of recording. (B) Left: The five types of human retinal photoreceptors are differentially activated depending on the flash stimulus: the red flash mainly activates the L/M cones (darker drawing), whereas the blue flash activates ipRGCs directly and the outer retinal photoreceptors with the L-cones being less activated (lighter drawing). Right: From a representative emmetropic control subject of the study, the PLR to the red flash (upper plot) elicited fast pupil constriction and redilation with weak postillumination pupil response (PIPR) at 6 and 30 seconds after stimulus offset. On the other hand, the PLR to the blue flash generated a sustained pupil constriction, that is, a strong PIPR after the peak pupil constriction (PPC) and before recovery to the baseline. (C) PLR metrics and units.

The integrated infrared camera provided 0.08 mm/pixel resolution at the iris and 28.3 hertz (Hz) recording rate. Visual stimuli were 1-second flashes of light. The photopic luminance was 250 cd/m² in both red (relative spectral energy between 590 and 650 nm; peak wavelength = 621 nm) and blue (relative spectral energy between 430 and 510 nm; peak wavelength = 470 nm) light stimuli. It has been well established that long-wavelengths fall away from the peak of the melanopsin absorption spectrum. Therefore, PLR to the red light contain insignificant direct activation of the ipRGC-melanopsin system. On the other hand, the short wavelengths are able to activate melanopsin directly,^24,31 besides activating the outer retinal photoreceptors. Figure 2B shows that, accordingly, PLR to red (long-wavelength) flash shows a fast pupil constriction and redilation, which is driven by cone inputs to ipRGCs, whereas PLR to the blue (short-wavelength) flash shows rod/cone controlled fast pupil constriction, followed by the slow PIPR driven by the ipRGC-melanopsin system.²³ 

Data Analysis

PLRs to the red and blue flashes of light were analyzed with MatLab (MathWorks Inc., Natick, MA, USA) custom-written program. As shown in Figures 2B and 2C, the BPD was calculated for both responses elicited by the red and blue flashes. The BPD was the average pupil diameter during the 1-second recording preceding each stimulus onset. The peak pupil constriction (PPC) was calculated as the difference in normalized pupil diameters between BPD and maximal constriction after flash offset. The PIPR was calculated twice for each condition (red and blue): at 6 seconds (PIPR6) and at 30 seconds (PIPR30) after flash offset. PIPR6 and PIPR30 indicated the differences between BPD and pupil diameters measured 6 and 30 seconds after the flash, respectively, normalized to the BPD. PPC as well as PIPR were normalized to correct for individual differences in BPD, as previously recommended.^25,32 Statistical comparisons were applied to pooled right and left eyes for the female groups. Genotype and group comparisons were performed using 1 or 2-way ANOVAs depending on the presence of a within-subject repeated measure. The P values < 0.05 were considered statistically significant. Special attention was paid to the potentially lower variation of the data deriving from the MYP-26 group (comprising related individuals only), as it might lead to easier rejection of the null-hypothesis of statistical tests, and consequently an increased rate of false positive statistical comparisons. For no parameter, however, did the MYP-26 group display the lowest SD among the 3 groups analyzed (see Tables 1, 2), indicating that this potential source of error was unlikely to mislead the study. 

Table 2.

 

The Results of PLR Measurements

 

Results

Ophthalmological examinations revealed normal anterior and posterior segments of the eye, except for the myopic fundus alterations, high myopic refractive error of SE −5.50 to −16.00 D, and reduced BCVA for the patients with MYP-26. The M-CTRLs also showed normal anterior and posterior segments and HM between SE −6.00 and −14.00 D, but all these patients achieved full BCVA = 1.0 (decimal). Statistical comparisons showed similar mean age among the three groups (F_(2,25) = 0.157; P = 0.856) as well as comparable SE between groups MYP-26 and M-CTRL (P = 0.999). However, BCVA was significantly reduced in patients with MYP-26 as compared to the M-CTRL and E-CTRL groups (F_(2,51) = 115.990; P < 0.001). All subjects in the E-CTRL group showed no relevant refractive error (SE = −0.50 to +0.50 D) and had normal ophthalmological findings. The male subjects (carrier and healthy control) had normal anterior and posterior segments, low myopic refractive errors, and full BCVA for the age-matched male control and the dominant eye of the male carrier. The non-dominant eye of the male carrier showed reduced BCVA due to exotropic amblyopia. No other ocular diseases were observed in any of the patients. 

The fundus appearance and the optical coherence tomography (OCT) images of patients with MYP-26 showed no characteristics of cone dystrophy, such as bull's eye, macular geographic atrophic lesions, or outer retinal changes (i.e. disruption and loss of the photoreceptor layer), as opposed to what was expected based on the ARR3 knockout animal model.⁷ Retinal morphology was rather characteristic of HM and comparable in both the M-CTRL and MYP-26 groups. The Meta-Analyses for Pathologic Myopia (META-PM) classification system was used to categorize myopic retinal changes.²⁹ Category 0 (C0 = normal retinal structure) was observed in the E-CTRL and in the male carrier. Category 1 (C1 = tessellated retina) was observed in all M-CTRL subjects, whereas category 1 or 2 (C2 = diffuse chorioretinal and peripapillary atrophy) was observed in patients with MYP-26. OCT images revealed normal retinal and choroidal structure for the M-CTRL and male carrier patients, whereas moderate to extreme attenuation of the choroidal thickness without disruption or loss of the photoreceptor layer corresponding to C2 fundus appearance for the patients with MYP-26. Advanced stages of pathological fundus alterations (META-PM C3 and C4), signs of posterior staphyloma or visual media opacity (cataract, corneal, or vitreous opacities), that could affect PLR were not observed. 

Baseline Pupil Diameter

Prior to the flash onset, pupil size was recorded for 9 seconds. The BPD was the average of the pupil diameter recorded during the last second before the flash onset for the red condition BPD1 and for the blue condition BPD2. First, we found that BPD was highly replicable based on the coefficient of variance (CoV) which was less than 5% in the 3 groups (mean CoV: E-CTRL = 4.5%, M-CTRL = 2.7%, and MYP-26 = 3.4%) considering the two BPD measurements. The 2-way ANOVA within-subject repeated measure (BPD1/BPD2) was applied to compare the mean BPDs among the 3 groups. The mean BPDs are shown in Table 2. BPDs were statistically comparable among the groups, irrespective of considering one outlier subject from the MYP-26 group with relatively small BPDs (outlier included: group effect: F_(2,49) = 0.313 and P = 0.733; BPD*condition interaction: F_(2,49) = 2.016 and P = 0.144), or not (outlier excluded: group effect: F_(2,47) = 0.296 and P = 0.745 and BPD*condition interaction: F_(2,47) = 2.060 and P = 0.139). When searching for potential associations between BPD and age, we found significant negative correlations (BPD1 × age: r = −0.608, P = 0.021 and BPD2 × age: r = −0.611, P = 0.020; Supplementary Fig. S1). It is worth mentioning that BPDs recorded after dark-adaptation may not be comparable to pupil sizes of patients with myopia reported by previous studies due to differences in light conditions. 

Peak Pupil Constriction

Figure 3 shows PPC amplitude and time to peak constriction comparisons among the three groups for the red (upper plots) and the blue (bottom plots) conditions. The PPC to the red flash was reduced in patients with myopia. The mean PPCs are shown in Figure 3A and Table 2. One-way ANOVA showed that PPC was significantly lower in myopes (group effect: F_(2,49) = 8.780; P = 0.001). Bonferroni post hoc (between groups) comparisons revealed a myopic effect on PPC to the red flash because both myopic groups were similarly reduced (M-CTRL versus MYP-26: P = 0.999) compared to E-CTRL (P = 0.02). Time to peak constriction after the red flash offset (Fig. 3B) was comparable between the myopic groups (mean M-CTRL = 2.04 ± 0.19 seconds and MYP-26 = 2.07 ± 0.20 seconds) and they were both similar to the control values (mean E-CTRL = 2.03 ± 0.16 seconds). Significant differences among the groups were not observed for the time to peak constriction to the red flash (group effect: F_(2,49) = 0.244; p = 0.784). Overall, myopia seems to influence PPC to the red flash in our cohort, as shown in detail in Figure 4A. 

Figure 3.

 

Peak pupil constriction (PPC) normalized to baseline for E-CTRL (empty symbols), M-CTRL (light color), MYP-26 (dark color). (A) Mean PPCs to the red flashes were statically different among the groups with patients with myopia showing decreased values (less constriction, marked with asterisks) compared to the emmetropic control group. (B) Comparable time to peak (approximately 2 seconds for the 3 groups) was observed for the PPC to the red flash. (C) The PPC to the blue flash was statistically similar among the groups. (D) Time to peak constriction to the blue flash was highly variable, as expected. No statistical differences were observed among the groups for the time to peak constriction to the blue flash. * Significant (P < 0.05) values with post hoc Bonferroni test when compared to control group (E-CTRL).

Figure 4.

 

Averaged pupillary light reflex. The PLR traces to the red (A) and blue (B) flashes of averaged responses from each group: E-CTRL (N = 24 eyes, thin traces), M-CTRL (N = 14 eyes, traces with the intermediate thickness) and MYP-26 (N = 14 eyes, thicker traces). The PLR was characterized by a pupillary constriction briefly after light stimulation (dotted lines at time 0) followed by the pupil dilation after light offset. In the blue condition, the pupil redilation is slower originating from a sustained component. The smaller plots show the magnification of the peak pupil constriction (PPC) and the postillumination pupil response (PIPR) for the red and the blue conditions, respectively.

PPC to the blue flash was comparable (group effect: F_(2,49) = 2.206 and P = 0.121) among the groups with very similar means (see Fig. 3C, Table 2). Moreover, the mean time to peak constriction (Fig. 3D) was similar among the groups (E-CTRL = 2.63 ± 0.65, M-CTRL = 2.61 ± 0.46, and MYP-26 = 2.54 ± 0.39), and more variable compared to the time to peak of the PPC to the red flash, as expected. The statistical comparisons revealed no significant differences (group effect: F_(2,49) = 0.129 and P = 0.879) among the control and myopic groups for the time to peak. 

Postillumination Pupillary Response

Figure 5 shows PIPR to the red (upper plots) and the blue (bottom plots) conditions at 6 seconds (PIPR6) and at 30 seconds (PIPR30) after flash offset. PIPR6 values are higher in the blue condition compared to the red condition, as expected (see Table 2). Group comparisons showed that PIPRs to the red flash were similar for both PIPR6 (Fig. 5A) and PIPR30 (Fig. 5B) among the groups (PRPR6 group effect: F_(2,49) = 0.390 and P = 0.679 and PIPR30 group effect: F_(2,49) = 0.265 and P = 0.768). 

Figure 5.

 

Postillumination pupillary response (PIPR) results for E-CTRL (empty symbols), M-CTRL (light color), and MYP-26 (dark color). (A) Mean PIPR6 to the red flash were similar among the groups. (B) Comparable PIPR30 to red flash was also observed. (C) The PIPR to the blue flash was significantly decreased for the MYP-26 group (marked with an asterisk) compared to both control groups E-CTRL and M-CTRL. (D) PIPR30 to the blue flashes were similar among the groups. * Significant (P < 0.05) values with post hoc Bonferroni test when compared to either control group (E-CTRL and M-CTRL).

In contrast, significantly reduced PIPR6 was found in the blue flash condition for the MYP-26 group (Fig. 4B shows the average PLR traces and PIPR in detail). The mean E-CTRL PIPR6 to the blue flash (Fig. 5C) was similar to the mean of the M-CTRL group. However, the MYP-26 group showed significantly lower PIPR6 compared to both control groups E-CTRL and M-CTRL. The 2-way ANOVA reflected these differences (group effect: F_(2,49) = 10.392 and P < 0.001), whereas the post hoc Bonferroni test showed that the differences were limited to MYP-26 compared to E-CTRL (P < 0.001) and compared to M-CTRL (P = 0.003), whereas PIPR6 was comparable between the control groups (E-CTRL × M-CTRL P = 0.999). Interestingly, the PIPR6 of the male carrier was comparable to PIPR6 from an age-matched male control (mean of the eyes = 0.51). The pupil constriction at PIPR6 was 0.55 (mean of the eyes) in both the test and retest. These PIPR6 values were similar to female E-CTRL PIPR6 = 0.51 ± 0.06 and female M-CTRL = 0.50 ± 0.05, but higher than MYP-26 PIPR6 (mean = 0.42 ± 0.07) values (see Table 2). 

The reduced PIPR6 in patients with MYP-26 means that after the peak constriction, the pupil was more rapidly redilated. However, affected PIPR at 6 seconds after the blue flash offset was later recovered, considering that PIPR30 was comparable among the three groups (group effect: F_(2,49) = 0.216 and P = 0.806). The means are shown in Figure 5D and Table 2. 

Figure 6 shows that there was no correlation between PIPR6 and age (M-CTRL: r = −0.148, P = 0.613 and MYP-26: r = −0.069, P = 0.815), SE (M-CTRL: r = −0.005, P = 0.987 and MYP-26: r = 0.144, P = 0.623), and BCVA (MYP-26: r = −0.229, P = 0.431). 

Figure 6.

 

Testing for correlation of PIPR6 (after the blue flash) and various parameters of M-CTRL and patients with MYP-26. PIPR6 to the blue flash as a function of age (A) M-CTRL and (B) MYP-26, spherical equivalent (SE) (C) M-CTRL and (D) MYP-26 and best-corrected visual acuity (BCVA) (E) both myopic groups, pooled. There were no significant correlations in either test. Larger symbols (A and C) represent two overlapping points and the extra-large symbol (A) represents three overlapping points.

Discussion

The PLR has its primary origin in retinal photoreceptors: rods expressing rhodopsin (RHO gene; sensitivity spectral peak between 490 and 500 nm), cones expressing opsins sensitive to short (OPN1SW gene; approximately 420 nm), middle (OPN1MW gene; approximately 530 nm), and long wavelengths (OPN1LW; approximately 560 nm)³³ and ipRGCs expressing melanopsin (OPN4 gene; 482 nm).²⁸ The proportion of activation provided by each photoreceptor type depends on the light intensity³⁴ and the spectral distribution (“color”) of the light.³¹ The PIPR following the PPC is strongly affected by the spectral characteristics of the light. It has been well established that short-wavelength (blue) light may directly activate melanopsin expressing ipRGCs,^28,35 which is the neural substrate of the sustained PIPR component elicited by blue light at intensities of around 100 cd/m² and brighter.²⁴ 

In the present study, PLRs recorded to long- and short-wavelength brief flashes of light revealed an association between weaker PIPR6 under blue-light condition and MYP-26. The possible pathomechanisms involved in the ipRGC-melanopsin system dysfunction and speculations regarding genotype to phenotype relationships are discussed below. 

The first observation is that HM itself has little effect on PLRs elicited by red and blue brief flashes of light. This conclusion was achieved by comparing PLR results from patients with eoHM harboring intact ARR3 alleles (the M-CTRL group) and the emmetropic controls (E-CTRL group). Here, we show (see Fig. 4A) that the PPC to the red flash may be slightly affected by HM. It is worth mentioning, that (i) the genetic background of the M-CTRL group is not known, and (ii) the present study used dark-adapted single flashes to record PLRs which differs from the method reported by Mutti et al.³⁶ to study refractive error effects on PLRs. In the latter work, the authors applied the repeated light exposure method, finding that the PLR changes under light-adaptation, presumably due to the modulation of melanopsin inputs by the repeated stimulation.³⁶ Nevertheless, completely preserved PIPR in patients with eoHM without ARR3 gene mutations showed here corroborates previous studies consistently failing to find associations between mild to moderate myopia and PIPR.^32,37–40 To date, only one such study dealt with HM, showing that red and blue PLRs of a single patient with HM (SE = −9.25) were similar to that of emmetropes and patients with mild to moderate myopia.³² 

The second observation – with an exception of one patient with MYP-26 showing bilateral miosis, was that scotopic BPD recorded prior to flash onset was similar among the three groups: E-CTRL, M-CTRL, and MYP-26. This observation is congruent with previous findings of comparable pupil size in emmetropes and myopes, whereas it was significantly higher in myopes as compared to hyperopes.⁴¹ Conversely, refractive status has been reported to impact light-adapted pupil size with the larger pupils associated with the myopic status.^40,42–44 In the present study, no associations were found between BPD and myopia, under the dark-adapted condition. 

The main finding of the present study was the decrease of the sustained component of the PIPR to the blue flash, indicating the weaker signal originating from the ipRGC-melanopsin system.²³ This was exclusively observed in ARR3 heterozygous (MYP-26) female patients of the affected family (see Fig. 4B), evidencing that cone-arrestin is required for the proper functioning of the ipRGC-melanopsin system. Whether the disruption of PIPR is caused by dysfunctional outer retinal photoreceptors or it is due to circadian rhythms dysregulated by the altered cone-arrestin in MYP-26 retinas remain to be investigated. Normal PIPR6 to blue light in male carriers with hemizygous ARR3 gene mutation remains to be further confirmed. However, the description of a single case in the present study suggests that ipRGC-melanopsin dysfunction might be restricted to female patients and, therefore, involved in the pathological eye growth of patients with MYP-26. 

Photoreceptor dysfunction influencing the activity of the retinal ganglion cells (RGCs) in patients with MYP-26 was previously demonstrated by our group based on electrophysiological results showing reduced and delayed pattern electroretinogram (PERG).³ Therefore, it could be argued that the altered PIPR6 to the blue flash reported here was simply the effect of a general RGC dysfunction. However, it is known that the contribution of ipRGCs to the PERG is not significant, as normal PERG was found in an experimental mouse model (OPN4^-/-) lacking melanopsin expression.⁴⁵ It is worth mentioning that reduced and delayed PERG amplitudes are biomarkers of moderate⁴⁶ and high⁴⁷ myopia, regardless of the genetic background, possibly caused by the pathological eye elongation disrupting ganglion cells’ function in general. Still, this study found that patients with eoHM without the genetic alteration of the ARR3 gene showed PIPRs comparable to emmetropes, contrary to what one would expect if the general RGC dysfunction would lead to a weakened ipRGC-melanopsin signal. Our previous study found PERG alterations (indicating a general RGC dysfunction) in both, female patients and male carriers.³ In this study, a male carrier did not display an altered PIPR6, therefore, the generalized RGC dysfunction linked to this ARR3 gene mutation does not seem to necessarily cause ipRGC dysfunction. 

Nevertheless, the outer retinal photoreceptors certainly contribute to the PLR elicited by the red and blue flashes,³¹ as shown in Figure 2B. However, rod as well as L/M-cone inputs alone contribute minimally to ipRGCs signaling neurons in the olivary pretectal nucleus (OPN), the subcortical substrate of PLR.⁴⁸ S-cones, on the other hand, are possible contributors of OPN/PLR activation⁴⁸ with a peak spectral sensitivity (420 nm) close to that of the melanopsin.²⁸ It has been recently shown that S-cone function is apparently preserved in patients with MYP-26,⁴ which is expected considering that ARR1, the rod arrestin, is highly concentrated, even more expressed than ARR3, in S-cones.⁴⁹ In L/M-cones, on the other hand, ARR1 is much less expressed, whereas ARR3, which is highly concentrated in cone synaptic pedicles, is the main cone-arrestin in these cells⁴⁹ showing unique modulatory roles.⁵⁰ This could mean that S-cone dysfunction is less likely to lead to PIPR defects and, therefore, may indicate the direct effect of ARR3 gene mutations on ipRGC-melanopsin system dysfunction. 

A further question would be consequently whether the ipRGC-melanopsin system disruption suggested by the PLR changes in patients with MYP-26 stimulate pathological eye growth in female patients? It has been demonstrated that the ipRGC-melanopsin system is a key player in emmetropization.^14,19,51 Pathological eye growth and myopic shifts were observed in mouse models with disrupted melanopsin system with,¹⁴ but also without¹⁵ a functional rod and cone system. Moreover, under form-deprivation conditions, the proper activation of the ipRGC-melanopsin system is required to slow myopic progression. The melanopsin system dysfunction was associated to three times higher myopic shifts in mice.¹⁴ Interestingly, ipRGCs provide excitatory inputs to dopaminergic amacrine cells,⁵² probably the only cell type synthetizing dopamine in the retina.⁵³ This is most likely the mechanism underlying melanopsin signaling-related form-deprivation myopic development.¹⁴ It became evident that the normal levels of retinal dopamine is essential for the postnatal regulation of eye growth.^54,55 The lower the dopamine level the higher the susceptibility to develop form-deprivation^56,57 and spontaneous⁵⁸ myopia. In addition, it has been demonstrated in animal models that increasing dopamine levels in the retina by genetic or pharmacological manipulation inhibits myopia development.⁵⁴ In fact, dopamine is involved in both experimentally induced and spontaneous myopia (for a review, see Feldkaemper and Schaeffel 2013⁵⁹) and the administration of levodopa inhibits the development of form-deprivation^60,61 and lens-induced⁶² myopia by modulating dopamine D2-like receptor, mainly DRD4 dopamine receptor. Through retinal volume transmission, dopamine interacts with DRD4 dopaminergic receptors highly expressed in photoreceptors and from which desensitization/internalization is controlled by cone-arrestin, in addition to beta-arrestin.¹⁰ Photoreceptor DRD4 dopamine receptors have been proposed to influence photoreceptor function,⁶³ retinal light-adaptation,⁶⁴ electrical coupling between rods and cones, especially under mesopic light condition,⁶⁵ and retinal melatonin synthesis.^66,67 These are all DRD4/dopamine-dependent mechanisms related to visual conditions known to impact emmetropization in humans.^68–72 Although outer retinal photoreceptors are the main modulators of the dopamine pathway, ipRGC-melanopsin system certainly influences the retinal levels of dopamine.⁵⁷ 

An overall speculation accordingly could be that the mechanisms of (i) cone dysfunction causing form-deprivation/retinal defocus, as recently proposed elsewhere,⁴ and (ii) the alteration in the ipRGC-melanopsin system shown here, which is known to affect the dopaminergic pathways and, consequently, eye growth,⁷³ may both impact the pathogenesis of MYP-26. Further investigations are required to evaluate whether the ipRGC-melanopsin system dysfunction introduces a complex or multifactorial phenomena to the pathological eye growth in MYP-26 and possibly to myopization in other inherited retinal diseases. Our study is limited to seven affected women and one carrier male, all related. PLR measurements should therefore be extended to patients with MYP-26 independent of the family investigated here to verify the general nature of our findings. Nevertheless, the recently proposed cone dysfunction stimulating eoHM in female patients⁴ could be further re-interpreted in the light of retinal circadian rhythms influencing myopia development.^51,74 

As reported earlier by our team⁷⁵ and other groups,⁷⁶ the ipRGC-melanopsin system dysfunction is associated with sleep disturbances. Therefore, PIPR alterations in patients with MYP-26 could be associated with sleep symptoms which was, nevertheless, not the case in our patients. However, an important remark regarding the present data is that ipRGC-melanopsin system driven PIPR is disrupted but not completely suppressed in MYP-26. It is rather able to reach full pupil redilation at 30 seconds after flash offset, which is comparable to emmetropes and patients with eoHM without ARR3 gene mutations. 

The present results emphasize that PLR is a powerful noninvasive technique allowing the objective measurement of retinal cells’ integrity.^23,27,77 

Conclusions

We detected weaker sustained postillumination pupil constriction in female patients with MYP-26, but not in the male carrier of the ARR3 gene mutation. This finding indicates the dysfunction of the ipRGC-melanopsin system exclusively in patients with myopia, which may act as a further pathomechanism of pathological eye elongation, in addition to the previously reported L/M cone dysfunction.⁴ 

Acknowledgments

Supported by the National Research, Development, and Innovation Fund of Hungary to MTSB (NKFI OTKA-PD grant number 134799); Tempus Public Foundation (Hungary) to MTSB (TKA grant number 167949), and the Biotechnology National Laboratory of Hungary to T.F. 

Disclosure: M.T.S. Barboni, None; N. Széll, None; Z. Sohajda, None; T. Fehér, None