In this study, we report a family displaying a heritable form of eoHM, where the disease is manifested only in females. Compilation of the pedigree permitted the identification of carrier males, and revealed that their female offspring areexclusively affected, which suggested an X-linked dominant, female-limited inheritance. Whole exome sequencing of two individuals indeed revealed a nonsense-mutation within the coding region of a gene on the X-chromosome, namely ARR3. Sanger sequencing of the respective locus in a total of 16 female family members unveiled a perfect correlation between the presence of the mutant allele and the high myopia phenotype. This is the first report of a mutation in ARR3 causing hereditary eoHM, called Myopia-26 in a Caucasian family. Three Chinese families have been reported earlier to display a similar, X-linked dominant, female-limited transmission of eoHM(6). In those cases the ARR3 was found to carry c.893C>A (p.Ala298Asp), c.298C>T (p.Arg100*) and c.239T>C (p.Leu80Pro) mutations, respectively. The mutant allele identified in our study (c.214C>T, p.Arg72*) is therefore novel. The earlier publication on Myopia-26 lacked a detailed phenotypic description of the patients, and did not attempt to explain the pathomechanism of the disease. Our main goals from this point onwards were therefore to carry out a thorough ophthalmologic investigation of the family and use the acquired information, along with literature data to build reasonable hypotheses on the molecular mechanism of pathogenesis.
ARR3 encodes a 388 amino acid-long visual arrestin with multiple names (Arrestin 3, Arrestin 4, Cone-arrestin, Retinal cone arrestin-3, X-arrestin), we refer to it as X-arrestin. Besides its key role in the phototransduction process in retinal cones, it is also expressed in pinealocytes of the pineal gland (10). Arrestins make up an important family of proteins, with the primary function of desensitizing phosphorylated G-protein coupled receptors (GPCRs). Arrestin 1 and X-arrestin bind to opsins (hence called visual arrestins), while -arrestin 1 and 2 bind to numerous other types of GPCRs. Arrestin 1 has very high preference for opsins found in retinal rods and cones, whereas X-arrestin has a fairly high binding capacity to non-opsin binding partners as well, and therefore has more diverse synaptic roles (11).
Our knowledge about the function and cell type-specific expression of X-arrestin is, at this time based mostly on experimental data derived from animal models. X-arrestin is expressed in all cone types of the human retina (12), however it displays a weaker expression in the S-cones of mice (13). Arrestin 1, on the other hand is detectable in rods and S-cones of baboons, but not in LM cones (14). In the cones of knockout mice, Arrestin-1 seems to provide a functional replacement for X-arrestin(15). This experimental dataset allows us to formulate two reasonable, albeit incomplete hypotheses on the pathogenesis of myopia in ARR3-mutant patients. We refer to these as the cone- and the ganglion cell-hypothesis, respectively. The cone-hypothesis assumes that Arrestin-1 expression in humans is present in S-cones, but not in LM cones, as seen in baboons (14), so an X-arrestin defect would lead to limited arrestin function in LM, but not in S cones. Since arrestins are responsible for the desensitization of opsins, decreased arrestin function in LM-cones would mean their increased activity, and the “sensitization” to red/green visual stimuli. Such selective cone dysfunction could explain the onset of myopia the following way. The physical phenomenon of chromatic aberration leads to shorter wavelengths forming an image in a more anterior, and longer wavelengths forming an image in a more posterior plane (Figure S1A). Normally, the measure of luminance contrast is maximized during accommodation, and long-wavelengths form an image behind the photoreceptors. In patients with a relatively increased sensitivity of L-cones, the posterior image will produce a stronger stimulus (Figure S1B). As a result, a higher luminance contrast will be attained upon increased accommodation and by ocular elongation, two hallmarks of myopia pathogenesis (16). Although accommodation excess in itself may not be sufficient to cause myopia (17), the phenomenon of image-forming behind the retina, called hyperopic defocus has been shown to provoke ocular elongation in numerous animal studies (18, 19). Briefly, since blue light is claimed to have a protective effect against myopia, the relative weakening of the blue light stimulus upon the loss of X-arrestin can explain the eventual development of myopia in these patients (20).
The selectively altered function of various cone types, however, cannot be tested with standard photopic 3.0 ERGs. Due to the quite extensively overlapping spectral sensitivities of different photopigments(21), these tests reflect the summed activity of all three retinal cone types. Photopic 3.0 ERGs indeed, were normal and showed no alteration in our patients (Figure 3,). L, M and S-cones responses can be isolated electrophysiologically by recording the light adapted ON/OFF-ERG and the S-cone ERG. Similar to the PhNR, these recordings are an extension of the full-field ERG (22)which enable characterisation of the different cone types, including bipolar cell interactions.
Our ganglion cell-hypothesis attributes the development of refractive error to the dysfunction of retinal ganglion cells (RGC). To better understand this connection, one must acknowledge that apart from their primary role of transmitting visual information from photoreceptors to higher cerebral visual centres, a subset of RGCs called intrinsically photosensitive retinal ganglion cells (ipRGCs) have an additional role (23). As their name suggests, they can detect light directly through their photosensitive protein called melanopsin. At the same time, they also transduce the signal originating from rod and cone photoreceptor cells, analogously to classical RGCs (24). Classical and ipRGCs are interconnected horizontally by amacrine cells, which allow them to influence the activity of one another (25). IpRGCs and their light sensitive protein, melanopsin are primarily responsible for non-image forming visual functions such as circadian rhythms or pupil reactions (26-28). They have recently been discovered to play a role in conscious, image-forming visual perception as well (27). Eye development is connected to both image-forming and non-image forming light detection pathways and accordingly refractive error may be a consequence of the derailment of either.
There is an increasing body of evidence supporting that in the image-forming pathway, light plays a key role in emmetropization and refractive error development, and besides the intensity, the spectral composition of the light stimulus is just as crucial (29, 30). As opposed to opsins, melanopsin is most sensitive to shorter wavelengths of the spectrum, i.e. blue light (31). Besides the anti-myopic effect of blue light attributed to the myopic defocus it causes on the retina (discussed above) (20), it has a further protective effect mediated in part by dopamine through pre- and postsynaptic connections of ipRGCs to dopaminergamacrin cells (32). Dopamine has been long acknowledged as a retinal neurotransmitter acting against myopia development, and it has also been evidenced that blue light stimulates a larger amount of dopamine release than other wavelengths do (32). Accordingly, a disruption of ipRGC function may result in the alteration of the wavelength composition of the perceived light with a chromatic aberration shifted towards longer wavelengths of the spectrum, along with decreased dopaminergic activity. Both issues reduce the protective effect of blue light against myopia, potentially leading to the development of a progressive refractive error.
The non-image forming visual functions of ipRGCs, such as circadian rhythm photoentrainment also play an important role in eye development (33). IpRGCs and melanopsin mediate circadian cycles both endogenously in the retina (again, through dopamine release) and via a systemic route comprising the hypothalamic suprachiasmatic nucleus (SCN) and the pineal gland through the inhibition of melatonin release in pinealocytes(33). The circadian clock influences ocular development, and disruption of the circadian cycle has been found to elongate eye components and yield myopia in various myopia models (34). Therefore, either the primary defect of ipRGCs or the primary dysfunction of pinealocytes (or both) could cause the refractive error seen in our patients. Although the prior is difficult to explain (discussed below), the latter (pineal malfunction) is highly probable due to the fact that pinealocytes normally express the X-arrestin. Melatonin, the product of pinealocytes has been shown to inhibit retinal dopamine synthesis (35), modulate D2 dopamine-receptor expression in the retina of chicks (36) and abolish diurnal cycling of dopamine levels in goldfish retina (37). These observations could strongly support the possibility that pinealocyte malfunction caused by ARR3 mutations lead to altered (probably increased) melatonin levels, which in turn cause myopia by impairing the diurnal rhythms of the eye.
Currently, the most obviously missing piece of both the cone- and the ganglion cell-hypothesis is the cause of RGC dysfunction displayed on the PERG recordings. Direct linkage to the ARR3 mutation would require ARR3 expression in RGCs, which was not detectable in mice (15). However, the promoter of the human ARR3 and its murine orthologue are markedly different, which may result in disparate cell type specific expression as well (11). Another possibility would be the secondary malfunction of RGCs, resulting from the altered activity of pinealocytes. This could be mediated by the humoral control of retinal dopaminerg transmission by the pineal gland (described above), or the direct effect of melatonin on RGCs via their MT1 and MT2 melatonin receptors (38). The details of this control are currently missing, it is nevertheless noteworthy that myopes have higher melatonin levels than non-myopes(39). Finally, altered cone function, resulting from reduced X-arrestin levels may also negatively influence RGC activity. We nevertheless have no reason to believe that the cone- and the ganglion cell hypotheses are mutually exclusive, or exclude other pathomechanisms.
Another major shortcoming of both the cone- and the ganglion cell hypothesis is the lack of explanation for the female-limited heredity pattern of myopia. It is especially curious that the central macular dysfunction seems to be present also in males, without leading to eoHM. We assume the presence of a “rescue mechanism” in males, or in other words, the lack of a pathological process that would lead to an axial length elongation in response to the central retinal dysfunction. Sex-dependent differences in retina function have been described in mice, and the risk of certain retinal diseases have been shown to be sex hormone-dependent in humans(40). Further physiology and molecular biology studies are required however to unveil the exact mechanisms responsible for the observed female-limited phenotype. Such research may also shed light on why the mutant allele is dominant in females. In the course of molecular studies however, the limitations of animal models must always be kept in mind, despite their great value. For example, an age related cone dystrophy was suggested in Arr4-/- mice (Arr4 being the murine orthologue of ARR3) based on immune-histochemical findings and the pronounced diminishment in photopic flash and flicker ERGs (9). In contrast, no generalized cone dysfunction could be evidenced in our patients carrying ARR3 mutation, either male or female, according to the electrophysiological and ophthalmological phenotypic characterization.
From the clinical point of view, our next investigative steps seem well defined: i) cone-specific ERGs (S-cone ERGs and ON/OFF ERGs) to isolate individual (L, M, or S) cone responses (41) and thus support or exclude our selective cone dysfunction hypothesis; ii) post-illumination pupil response (PIPR) to test melanopsin expressing ipRGC function (21) and thus shed light on the extent of ipRGC damage. iii) long-term follow-up of the progression of a potential POAG monitoring IOPs,visual field defects, optic nerve head appearances and RNFL OCTs.