Early Macular Involvement in Non-syndromic Retinitis Pigmentosa

Abstract

Retinitis pigmentosa (RP) is the most common inherited retinal disease (IRD), with a worldwide prevalence of about 1:4000. Functionally, RP is a rod-cone dystrophy (RCD), where rod photoreceptors are affected earlier and more severely than cone photoreceptors. As a consequence, RP typically manifests with night blindness beginning in adolescence followed by concentric constriction of visual field, while central visual loss usually occurs later in life. The molecular bases of this disorder, however, are highly heterogeneous, with over 100 genes implicated, and in some RP subtypes macular manifestations occur relatively early in the disease course. The subsequent rapid decline in visual acuity in such patients blurs the line between RP and other IRDs, namely cone-rod dystrophies (CORDs), and increases the difficulty in narrowing down the differential diagnosis. This chapter aims to review the features of non-syndromic RP caused by mutations in genes that have been commonly associated with early macular involvement and to provide an updated overview on recent preclinical or clinical studies addressing these rare diseases.

1. Introduction

Retinitis pigmentosa (RP) is an inherited retinal disease (IRD) that causes progressive vision impairment commonly leading to legal blindness, with an estimated worldwide prevalence of about 1:4000 [1], but highly variable based on geographic area. The highest frequency of occurrence has been reported among the Navajo Indians (1:1878), from Northern China (1:1000), and from southern and central India (1:600–750), while countries, such as Switzerland (1:7000) and England (1:4800), have a lower prevalence.

Non-syndromic RP, which accounts for approximately 70–80% of all RP cases [2], can be classified according to its inheritance pattern in: autosomal dominant (AD) RP (20–25%), autosomal recessive (AR) RP (15–20%), and X-linked (XL) RP (5–15%). The remaining 40–50% of patients are designated as simplex RP, representing isolated cases with only one affected family member [3, 4]. In addition, 20–30% of patients present with syndromic forms of the disease, where the retinal disorder is accompanied by extraocular abnormalities [2].

RP is a rod-cone dystrophy (RCD) as there is a primary degeneration of rod photoreceptors, while cone photoreceptors are involved in more advanced diseases. Traditionally, the distinction between RCD and cone/cone-rod dystrophy (COD/CORD) has been established based on the pattern of generalized retinal dysfunction displayed on full-field electroretinograms (ERGs). In RCDs, scotopic dim flash (0.01 cd/m²) ERGs show a decrease in amplitude and a delay in implicit time that exceed those measured on bright single flash (3.0 cd/m²) or 30 Hz flicker ERGs, while the opposite occurs in CORDs. While this definition often allows to make an unequivocal diagnosis and to provide patients with accurate information on likely progression of symptoms, some cases are not as clear-cut. For example, since in RCDs and CORDs functional damage usually precedes structural change, patients can have no recordable scotopic and photopic ERGs at relatively early disease stages. In these circumstances, when ophthalmoscopy does not reveal obvious signs of RP, retinal imaging, such as optical coherence tomography (OCT) or fundus autofluorescence (FAF), can help identify biomarkers (e.g., subfoveal sparing of outer retinal layers on OCT or altered autofluorescence at the posterior pole) that are consistent with a certain pattern of dysfunction (rod versus cone-dominated). Similarly, psychophysical exams, such as microperimetry and chromatic full-field perimetry, can isolate rod and cone contributions and guide physicians in counseling their patients.

In classic RP, nyctalopia is often the initial symptom, followed by a progressive constriction of the visual field. Macular function and visual acuity (VA) are usually relatively preserved until later stages. However, in certain RP subtypes, the degeneration of central photoreceptors and retinal pigment epithelium (RPE) occurs earlier than one would expect based on the degree of visual field constriction. As discussed above, these subtypes are often characterized by extinguished scotopic and photopic ERGs which, in the presence of macular atrophy, can make differentiating the diagnosis with CORDs challenging.

RP with early macular involvement is a particularly debilitating as patients experience a rapid decline in best-corrected visual acuity (BCVA) in addition to the progressive visual field constriction that is typical in RP. These patients often meet both criteria for legal blindness as defined in the United States (i.e., BCVA of 20/200 or less in the better eye and visual field less than 20 degrees in the better eye) at a young age. Some genes have been more often associated with early macular atrophy and a more aggressive phenotype. In order to provide patients with accurate prognostic information, it is important for the IRD specialist to be aware of these genotype-phenotype correlations and up-to-date with current research.

In this chapter, we will review the molecular biology and clinical features of RP-associated genes that most often result in early macular involvement, and we will provide an overview of relevant research efforts.

2. Common causative genes of non-syndromic RP with early macular involvement

In this section, we will provide an overview of RP-associated genes that cause non-syndromic RP with early macular involvement, focusing on those that are being investigated in preclinical studies and clinical trials.

2.1 Non-syndromic retinal ciliopathies

Ciliopathies are defined as a group of disorders resulting from impaired function of the cilium, a highly specialized and evolutionarily conserved organelle in eukaryotes [5]. Cilia are classified into two subtypes: motile, mainly localized in epithelial cells of the lung, ventricles of the brain and spermatozoa, and nonmotile (or primary), which are ubiquitous [6]. While motile cilia have the main function of generating movement, primary cilia are involved in cell signaling and trafficking [7]. In photoreceptors, the primary cilium is responsible for connecting the outer segments (OS) to the inner segments (IS), where it plays a critical role in transport, and it is estimated that every minute approximately 2000 opsin molecules are delivered to the OS through the cilium [8, 9, 10].

Primary cilia dysfunction may present as a single organ disorder, including an isolated retinal degeneration (e.g., XLRP), or it can affect multiple organs resulting in a syndromic ciliopathy, such as Bardet-Biedl syndrome (BBS) [5]. Although some ciliary gene-associated forms of RP, such as those caused by biallelic variants in MAK, are reported to spare central vision until well into adulthood [11], patients with mutations in connecting cilium genes are often reported to have a more severe RCD with early macular degeneration [12, 13].

Herein, we will review the molecular and clinical features of genes expressed in the primary cilium of photoreceptors that often cause non-syndromic RP with early macular involvement.

2.1.1 RPGR

RPGR is located on the X chromosome and is comprised of 19 exons encoding the retinitis pigmentosa GTPase regulator, a protein involved in photoreceptor ciliary function. Alternative splicing results in more than 10 different isoforms [14, 15]. The two major isoforms are RPGR^Ex1–19, the full-length variant and RPGR^ORF15, which encompasses the first 14 exons of the gene and the open reading frame (ORF) 15 region, derived from an alternatively spliced exon 15 with part of intron 15 (Figure 1). While RPGR^Ex1–19 is expressed throughout the body, RPGR^ORF15 is primarily expressed in the connecting cilia of retinal photoreceptors. Approximately 80% of pathogenic variants in RPGR are located in the ORF15 region [16], and the remaining 20% are found in exons 1–14. No disease-causing mutations have been reported in exons 16–19 [14].

Figure 1.

Structural diagram of the retinitis pigmentosa GTPase regulator (RPGR) gene and two common splicing isoforms.

RPGR pathogenic variants can manifest as a rod-cone dystrophy (RCD), which is the most common phenotype, a cone/cone-rod dystrophy (COD/CORD), and a macular dystrophy (MD) [17, 18]. Genotype-phenotype correlations are controversial. Some authors report that variants in exon 1–14 are characterized by a more severe disease than ORF15 variants [17, 19, 20], while others report the opposite [18, 21, 22]. There seems to be a broader agreement on the fact that patients with mutations in the ORF15 region exhibit greater phenotypic variability [17, 19, 23], and that variants at the 3′-end are more often responsible for CORD, while 5′ variants frequently cause RCD [24].

RPGR is responsible for approximately 80% of XLRP which, compared to other forms of RP, generally displays an earlier onset and a more rapidly progressive course (Figure 2). Macular abnormalities, including cystoid macular edema (CME), vitreo-macular traction (VMT), epiretinal membrane (ERM), full-thickness macular hole (FTMH), and lamellar macular hole (LMH), are common. Early macular involvement in XLRP patients is also associated with a reduced sensitivity on microperimetry, which does not seem to correlate with older age [21]. High myopia is frequent with a prevalence of up to 80% and is associated with a more rapid decline in BCVA [18, 21].

Figure 2.

Multimodal imaging of a 35-year-old male hemizygous for a pathogenic variant in RPGR. Pseudocolor fundus photos (A, B) show diffuse bone spicule pigmentation and macular RPE mottling, although the view of the posterior pole is partially obscured by the patient’s posterior subcapsular cataract. FAF (C, D) better highlights the atrophic and dystrophic changes in the macular region. OCT (E, F) shows a severe attenuation of outer retinal structures, with a near complete loss of the EZ band, accounting for the patient’s poor BCVA (20/400 OU) despite his young age. FAF = fundus autofluorescence; OCT = optical coherence tomography; EZ = ellipsoid zone; BCVA = best-corrected visual acuity.

Female carriers show high phenotypic variability and asymmetry between affected eyes [25]. The presentation can range from asymptomatic to severe disease indistinguishable from that seen in males [26]. Even in asymptomatic subjects, carrier status can often be detected with clinical examination and fundus autofluorescence (FAF), which sometimes show a typical radial tapetal reflex and radial pattern of hyperautofluorescence, respectively [27, 28].

There are a number or reasons why RPGR stands out as an optimal candidate for gene replacement therapy, including its small size fitting the cargo capacity of the common viral vectors, the loss-of-function (LOF) mechanism underlying its pathogenic variants, along with the relatively high prevalence and more severe disease burden in the IRD population. However, the repetitive purine-rich sequence of ORF15 is prone to spontaneous mutations, making it a challenge to produce a safe and effective therapeutic vector. Two main strategies have been adopted to address this problem. The first consists of using a truncated form of RPGR^ORF15 since moderate shortening of its length improves stability while preserving function [29]. The second strategy is through codon optimization, which aims to modify the codon composition to make it more stable without altering the amino acid sequence [29]. There are five ongoing clinical trials, each of which has adopted either of these two strategies, with promising results.

2.1.2 RP1

RP1 is a four-exon gene encoding oxygen-regulated protein 1, a photoreceptor-specific microtubule-associated protein, which is involved in photoreceptor development, organization of the OS, and regulation of ciliary trafficking [30]. The RP1 locus was first identified in 1982 in a large family with AD RP [31], and the gene linked to the disease was named retinitis pigmentosa 1 (RP1, OMIM 603937) [32], which is now known to be one of the most frequent causative genes in AD RP, with a prevalence as high as 10% [33]. Later publications revealed that variants in RP1 can also cause AR RP [34, 35, 36], and a recent study depicted that mutations in this gene are a relatively common cause of AR RCD [37].

Generally, AD RP tends to have a milder presentation, and RP1-associated AD RCD is no exception, with patients typically being diagnosed in their 20s and exhibiting classical signs of RP with relatively preserved central vision and visual field [33], as shown in Figure 3. Incomplete penetrance and variable expressivity have also been reported in families with dominantly inherited RP1 mutations [33, 38]. Biallelic pathogenic variants lead to more severe disease (Figure 4), with early onset and frequent central involvement, which has been variably described in previous reports as macular RPE degeneration [36], macular stippling [34], and macular atrophy [39]. BCVA often declines in the second decade of life, sometimes progressing to hand motion (HM) or light perception (LP) by the third or fourth decade [39].

Figure 3.

Multimodal imaging of a 65-year-old female heterozygous for a pathogenic variant in RP1. Compared to AR forms of RP1-retinopathy, AD forms are less severe. Pseudocolor fundus photos (A, B) show the classic clinical triad of optic disc pallor, vessel attenuation, and bone spicule pigmentation. FAF (C, D) shows a perifoveal hyperautofluorescent ring, speckled midperipheral hypoautofluorescent changes, and some peripheral patches of deep hypoautofluorescence, especially nasally. OCT (E, F) shows relatively preserved outer retinal layers, including ONL, EZ, and IZ, subfoveally and parafoveally. BCVA is 20/25 OU. FAF = fundus autofluorescence; OCT = optical coherence tomography; ONL = outer nuclear layer; EZ = ellipsoid zone; IZ = interdigitation zone; BCVA = best-corrected visual acuity.

Figure 4.

Multimodal imaging and psychophysical assessment of a 27-year-old male, harboring two frameshift variants in the RP1 gene. Pseudocolor fundus photos (A, B) show extensive bone spicule pigmentation, as well as atrophic changes in the macula, which appear as hypoautofluorescent lesions on FAF (C, D). OCT B-scan through the fovea (E, F) shows outer retinal atrophy OU and some parafoveal hyporeflective cysts in the INL OS. Microperimetry (G, H) reveals an unstable eccentric fixation OD, which is fairly stable OS. Macular sensitivity is severely depressed in both eyes, with no measurable threshold at most loci. On KVF (I, J) only target size V4e could be detected in a small (<5°) central island, as well as in a temporal crescent in the far periphery. FAF = fundus autofluorescence; OCT = optical coherence tomography; INL = inner nuclear layer; KVF = kinetic visual field.

RP1 mutations responsible for both AD and AR forms are mostly truncating (i.e., nonsense and frameshift) and result in premature termination codons, while missense mutations are rare [33]. Variants near the 5′ end tend to generate transcripts that are sensitive to nonsense-mediated decay (NMD), an mRNA surveillance mechanism that leads to degradation of the transcripts with introns in the 3′ untranslated region [40]. These mutations create a null allele causing LOF and have been associated with recessive RP, suggesting that haploinsufficiency of RP1 protein does not cause RP in humans [39]. AD RP1-related RCD is typically caused by pathogenic variants located within a hot spot (spanning approximately p.500 to p.1053) in exon 4 [33], which tend to escape NMD due to the lack of intronic structure downstream [39]. Mutations in this region are the vast majority and mutant proteins are expected to cause a deleterious effect on photoreceptors [39], either through gain-of-function (GOF) or a dominant negative effect (DNE). In mouse models, it has been shown that when nonsense variants are located within the hot spot, then photoreceptor degeneration can be delayed or prevented by adding wild-type Rp1 protein, suggesting a DNE rather than a GOF for at least some variants [41, 42].

These observations are of translational interest as shedding light on the molecular mechanisms underlying RP1-related IRDs is crucial in selecting the most appropriate therapeutic strategies. GOF mutations are not amenable to gene replacement therapy alone, while LOF and, potentially, DNE variants are. However, the pathogenic mechanism of many dominant mutations is still unclear; thus, suppression of the mutant protein may be required in addition to delivery of a healthy gene. A recent study showed that a CRISPR-Cas9 approach successfully reduced RP1 expression in Hap1-EF1a-RP1 cells, providing proof-of-concept of efficacy of gene editing in the treatment of patients with RP1-associated dominant IRDs [43].

2.1.3 PCARE

PCARE (previously known as C2orf71) encodes the photoreceptor cilium actin regulator protein (PCARE), an intracellular protein exclusively expressed in the retina, specifically at the initiation site of OS discs [44]. Although its function is not fully understood, it has been proposed that PCARE acts as an actin-associated protein and interacts with other molecular complexes, such as WASF3, to regulate the actin-driven expansion of the ciliary membrane during the development of new OS in photoreceptors [44].

Biallelic pathogenic variants in PCARE cause RP54 (OMIM #613428) [45, 46, 47] and, less frequently, CORD [48, 49]. Moreover, a digenic inheritance has been proposed, although evidence for this is confined to case reports suggesting a possible pathogenic effect in combination with other genes, such as CEP250 [50] and RP1L1 [51].

RP54 is more prevalent in Swiss [52] and Chinese [53] subjects, in whom PCARE is responsible for 15 and 8% of RP cases, respectively. This form of RCD is characterized by early age of onset and cone involvement with frequent macular degeneration [45, 46, 47], as shown in Figure 5. Recently, two animal models of PCARE-retinopathy have been developed, including the zebrafish PCARE [54] and a knockout of the mouse PCARE homolog, BC027072 (BC−/−) [55]. Zebrafish, such as humans, are diurnal animals with a cone-rich retina, but they have duplicated genes (pcare1 and pcare2), so decreasing or abolishing the function of pcare1 does not entirely mimic the retinal phenotype of PCARE-related retinopathy. Mice share 79% of the amino acid sequence identity of proteins coded by IRD-causing genes, but they are nocturnal animals without a cone-rich region comparable to the fovea. Despite these shortcomings, loss of function of both the zebrafish and mouse orthologs has been shown to lead to outer nuclear layer (ONL) thinning and OS disorganization, resulting in early-onset alterations in visual behavior, suggesting that PCARE is likely implicated in outer retinal development. Moreover, there are six dog breeds whose AR rod-cone dysplasia type 4 has been linked to the c.3149_3150insC variation in C17H2orf71 (PCARE human homolog), which could represent a feasible natural-occurring model to help develop new treatments [44]. PCARE is a two-exon gene spanning a 12.5 kb region (with one of the exons containing 95% of the coding sequence) [55] and therefore exceeds the cargo capacity of adeno-associated viral (AAV) vectors. However, limitations related to size are likely to be overcome in the near future as dual vector strategies are currently being developed to allow the delivery of large genes through a standard AAV gene therapy approach.

Figure 5.

Multimodal imaging and psychophysical assessment of a 46-year-old female with two pathogenic variants in PCARE. Pseudocolor fundus photos (A, B) show extensive regions of atrophy and dots of hypopigmentation with large peripheral areas of nummular punched-out atrophy 360 degrees OU. FAF (C, D) reveals a large central zone of hypoautofluorescence sparing a very small island, as well as large mid and far peripheral patches of hypoautofluorescence OU. OCT (E, F) displays extensive loss of outer retinal structures with minimal foveal sparing, hyperreflective granular deposits in the outer retina, along Bruch’s membrane and inner choroid. Static perimetry (G) shows a severely decreased sensitivity. BCVA of this patient is 20/50 OD and light perception OS. FAF = fundus autofluorescence; OCT = optical coherence tomography; BCVA = best-corrected visual acuity.

2.2 MERTK

MERTK encodes the widely expressed MER receptor tyrosine kinase, a transmembrane protein involved in a signal transduction pathway that regulates numerous cellular processes. In the retina, MERTK is expressed in the apical membrane of RPE cells and plays a role in the initiation steps of OS phagocytosis [56]. Failure of the RPE-mediated phagocytic pathway leads to accumulation of toxic debris in the subretinal space and to subsequent photoreceptor degeneration. To date, more than 90 MERTK variants have been identified in patients with AR RCD, Leber congenital amaurosis (LCA), and, more rarely, CORD [57]. MERTK-related RP or RP38 (OMIM #613862) accounts for approximately 2% of all RP cases and is generally characterized by childhood-onset nyctalopia and early macular involvement with decrease in BCVA often occurring after 2 years of age [58]. Clinically, RP38 can present, most notably in young patients, with discrete whitish dots on fundoscopy that are hyperautofluorescent on FAF consistently with impaired RPE phagocytosis [58]. Optical coherence tomography (OCT) displays a severe retinal thinning with loss of outer retinal layers and debris-like material in the subretinal space [59]. Foveal hyperautofluorescence in the absence of peripheral bone-spicule pigmentation, even when ERGs show a clear rod-cone pattern of dysfunction, is a common finding (Figure 6) and highlights the importance of a combined structural and functional assessment [57].

Figure 6.

Multimodal imaging of a 13-year-old female with two variants in MERTK. Although the macular abnormalities are not evident on pseudocolor fundus photos (A, B), FAF (C, D) shows foveal hyperautofluorescence, which is a common finding in MERTK-retinopathy. OCT (E-H) shows rapid progression over time: Compared to baseline scans (E and G), acquired when patient was 11 years old, the 2-year follow-up imaging (F and H) reveals loss of the small subfoveal island of preserved EZ, corresponding to a significant decrease in BCVA (from 20/30 OD and 20/40 OS to 20/300 OD and 20/100 OS). FAF = fundus autofluorescence; OCT = optical coherence tomography; EZ = ellipsoid zone.

MERTK has several features that make it an appealing target for gene therapy, including expression in the RPE, which is efficiently targeted by AAV vectors, small size of the cDNA sequence, and availability of preclinical models for vector validation. AAV-mediated retinal gene therapy was first tested in the Royal College of Surgeons (RCS) rat, which has recessive mutations in the Mertk gene, with promising results [60, 61, 62].

A gene augmentation therapy phase one trial in six patients with biallelic MERTK mutations was conducted in Saudi Arabia, and the results were published in 2016 [63]. The primary end point was safety at the 2-year follow-up, while secondary outcome measures included BCVA, full-field stimulus test (FST), and central macular thickness (CMT) on OCT. Subretinal injection of the vector was not associated with major side effects, and 50% of patients demonstrated increased BCVA though only one of them maintained these results [64]. In one patient, BCVA improved from <20/6400 at baseline to 20/40 90 days posttreatment but was 20/800 2 years after surgery. Although there was limited functional improvement, it should be noted that four out of six patients had a baseline BCVA <20/6400 and foveal atrophy, a scenario in which an increase in BCVA is not to be expected. A long-term follow-up to assess safety will be performed up to 15 years after treatment.

2.3 CERKL

CERKL encodes ceramide kinase-like protein, which has many isoforms due to extensive alternative splicing and multiple translational start sites [65]. Although the pathophysiology of this protein has yet to be fully elucidated, it has been shown that overexpression of CERKL protects cells from oxidative stress-induced apoptosis [66]. Biallelic pathogenic variants in CERKL have been associated with isolated MD, CORD, and RCD or RP26 (OMIM #608380), the latter representing a major cause of AR RP in the Spanish [67], Finnish [68], and Yemenite Jewish [69] populations.

However, unlike the majority of other IRDs, in CERKL-retinopathy the boundaries between these categories are not clearly cut. Many patients have a similar degree of rod and cone dysfunction on ERG, and even those with an electrophysiological pattern of RCD or CORD present overlapping features, an early drop in BCVA with central scotomas in patients functionally classified as RCD, as well as nyctalopia and progressive constriction of visual field in those with a diagnosis of CORD [70]. Although the rate of BCVA decline is extremely variable, this was found to be higher than other genotypes and especially faster in the 25–40 age group, suggesting preserved BCVA until the third decade of life, followed by a more rapid decline [70]. Indeed, a common finding in early stages is the presence of foveal atrophy with sparing of a small central island and a relatively retained BCVA, which severely diminishes once the atrophy encompasses the foveola in more advanced disease (Figure 7) [68, 70, 71].

Figure 7.

Multimodal imaging of a male harboring two variants in CERKL with a 6-year follow-up. Only OS is shown. OD had similar findings. At baseline (A, C, E), patient was 32 years old and displayed mild RPE mottling on pseudocolor fundus photos (A), while FAF (C) showed diffuse hyperautofluorescence around the fovea and speckled hypoautofluorescent changes in the midperiphery. OCT (E) revealed a relatively large area of EZ sparing. At this time, BCVA in OS was 20/20. Six years later (B, D, F), imaging shows a dramatic progression of the disease. Foveal atrophy is now evident on fundus photo (B), corresponding to a dense hypoautofluorescent lesion on FAF (D), which also shows tiny midperipheral hyperautofluorescent dots. OCT displays severe subfoveal outer retinal atrophy, with no detectable EZ (F). Functionally, these anatomic changes corresponded to a drop in BCVA OS from 20/20 to 20/100. FAF = fundus autofluorescence; OCT = optical coherence tomography; EZ = ellipsoid zone; BCVA = best-corrected visual acuity.

Other typical retinal changes include peripheral punched-out areas of chorioretinal atrophy with little to no pigment deposits and fine macular dots, which are hyperautofluorescent on blue autofluorescence (BAF) [70]. The FAF pattern suggests that these lesions are at least partially composed by lipofuscin-like material and that they may correspond to photoreceptor debris resultant from impaired RPE phagocytosis, similar to what is observed in MERTK-retinopathy [70]. These findings can help distinguish between CERKL-retinopathy and ABCA4-retinopathy, the latter being one of the most common initial diagnoses for macula-involving IRDs before genetic testing is performed [68, 70]. Given the link between CERKL and oxidative stress, antioxidants—which are generally recommended to all IRD patients—might be particularly beneficial in these cases. On the other hand, although presenting with a relatively favorable window of opportunity, gene therapy is hampered by the size of the cDNA, which would not fit into a regular AAV vector and will require more complex editing strategies. Recently, a new Cerkl mouse model was generated to provide insights into the molecular pathways downstream of this gene and to help design effective treatments [72].

2.4 PROM1

PROM1 encodes prominin-1, a 5-transmembrane domain glycoprotein [73], which is expressed ubiquitously in plasma membrane protrusions. Seven alternative splice variants have been reported [74], two of which are highly expressed in the retina [75]. In photoreceptors, prominin-1 localizes at the base of rod and cone OS, where it is involved in disc membrane morphogenesis [74], and subsequent photopigment sorting [76]. More recently, PROM1 has also been associated with the regulation of photoreceptor autophagy in RPE cells [77]. It has also been shown that PROM1 is required for the maintenance of physiological levels of ABCA4 and RDH12 [78], which is consistent with the assumption that this gene is also involved in the regulation of the visual cycle, and that it may play a role on lipofuscin accumulation indirectly through ABCA4 dysfunction [79].

PROM1-retinopathy encompasses a broad spectrum of phenotypes, including MD, cone dystrophy (COD), CORD, and, more rarely, RCD and LCA [80, 81, 82, 83, 84, 85, 86, 87]. This preferential cone involvement can be explained by the fact that cone opsins are more prone to mislocalization than rhodopsin [88] together with the wider distribution of prominin-1 in the cone OS [73, 89].

Regardless of the pattern of dysfunction on electrophysiological testing, PROM1-retinopathy is generally characterized by macular degeneration, which tends to be mild and isolated (i.e., normal ERG) in dominant forms, or severe with associated with panretinal involvement (manifesting as COD, CORD or RCD) in recessive forms [80, 81]. Other genotype-phenotype correlations, for example, with the type of variant, have proven less meaningful since similar phenotypes have been observed in patients with homozygous missense and splice variants compared with age-matched patients with truncating variants [80].

The vast majority of RP cases caused by pathogenic variants in PROM1 is AR [81, 84, 90], although AD inheritance has also been reported [81, 87]. AR PROM1-related RP is often difficult to distinguish from CORD because of the severely reduced photopic responses on ERG, the virtually constant macular involvement, and the possible absence of the typical ophthalmoscopic features (optic disc pallor, attenuated retinal vessels, peripheral bone spicules) [84]. Data on AD forms is scant in the literature (Figure 8). The largest series of AD PROM1-related RP describes the phenotype of eight members of an Italian family harboring the R373C mutation, a variant that generally results in isolated maculopathy [83, 87]. Reduced BCVA was the most common presenting symptom, with night blindness being noticed increasingly over time, reversed from the typical sequence for RP. Moreover, all patients displayed bull’s eye maculopathy (BEM) that gradually progressed into macular atrophy. A variable combination of the hallmark fundus abnormalities seen in RP was also observed [87].

Figure 8.

Multimodal imaging and psychophysical assessment of a 60-year-old female heterozygous for a pathogenic variant in PROM1. Pseudocolor fundus photos (A, B) show a large area of central atrophy, with sparing of a tiny foveal island, as better highlighted on FAF (C, D). OCT scans show an ERM (more prominent OD) and a severe loss of outer retinal layers, with a more preserved subfoveal region. On KVF (G), all isopters are constricted, and patient was not able to detect targets smaller than III4e (red line) centrally OD, while I4e (green) and I3e (yellow) could still be seen OS in a small central area of 2.1 deg2 and 0.7 deg2, respectively, likely due to the larger preserved foveal island. Patient had a BCVA of 20/200 OD and 20/50 OS. FAF = fundus autofluorescence; OCT = optical coherence tomography; ERM = epiretinal membrane; KVF = kinetic visual field; BCVA = best-corrected visual acuity.

Although there are no ongoing clinical trials, PROM1 is a gene that holds potential for future development of novel treatments. PROM1 is a small gene (2.5 kb) that mostly causes disease through recessive truncating variants, resulting in LOF or missense variants acting through a DNE [80], both mechanisms amenable to gene replacement therapy. Moreover, there is a natural occurring mouse model [91] that can further investigate PROM1’s pathophysiology, in addition to transgenic [74] and Prom1 knock-out models [78].

2.5 IMPG2

IMPG2 encodes the interphotoreceptor matrix proteoglycan-2 (IMPG2), formerly known as IPM 200 or SPACRCAN [92], which is localized in the interphotoreceptor matrix (IPM). The IPM, composed of glycoproteins and proteoglycans, fills the space between individual photoreceptors, as well as between photoreceptors and the RPE [93]. The IPM regulates multiple functions, including intercellular communication, neovascularization, cell survival, membrane turnover, photoreceptor differentiation and maintenance, retinoid transport, matrix turnover, and the precise alignment of the photoreceptors to the optical light path [92, 94]. Rods and cones synthesize IMPG2 and secrete it in the IPM [95], where it binds to polysaccharides, such as hyaluronan, and it anchors in the plasma membrane of photoreceptors, thereby fixating these cells to the retinal extracellular matrix [96]. Additionally, IMPG2 has calcium-binding potential, which suggests it plays a role in sequestering extracellular calcium released by photoreceptors in response to light [96].

Pathogenic variants in IMPG2 are responsible for RP [97, 98, 99] and MD, the latter often displaying a vitelliform phenotype [97, 100, 101, 102]. IMPG2-related maculopathy is generally caused by heterozygous missense variants, with only one previous report of an adult patient who harbored a homozygous missense IMPG2 variant with a mild MD [97]. IMPG2-associated RP presents as a recessive retinal dystrophy with early macular involvement (Figure 9). The two largest published series on IMPG2-related RP report some form of macular damage in the majority of patients [97, 98]. Macular abnormalities ranged from subtle RPE mottling, to a BEM, to macular chorioretinal atrophy, often with subsequent severely decreased BCVA. Longitudinal data is available for one patient, who showed progression of BEM to atrophy covering the entire fovea [98]. In most patients, the onset of the disease is either during childhood or during the early teens [97, 98, 99]. Of note, a recent publication showed that all [3] tested heterozygotes from families with AR IMPG2-related RP showed carrier signs in the form of macular focal thickening detected by OCT [99].

Figure 9.

Multimodal imaging of a 49-year-old male with two pathogenic variants in the IMPG2 gene. FAF (A, B) shows dense central hypoautofluorescence surrounded by mottled hyperautofluorescence, as well as large nummular areas of hypoautofluorescence in the mid and far periphery. OCT (C, D) displays in both eyes and ERM and severe thinning of outer retinal structures, including foveal loss of EZ line. BCVA is 20/1000 OD and 20/150 OS. FAF = fundus autofluorescence; OCT = optical coherence tomography; EZ = ellipsoid zone; BCVA = best-corrected visual acuity.

Despite the current lack of trials, preclinical research is being conducted on retinal organoids (ROs) generated from patient-derived induced pluripotent stem cells (iPSCs) and gene-edited embryonic stem cells to model human IMPG2-RP in vitro [103]. All ROs harboring IMPG2 mutations exhibited lacked an OS layers, in contrast to isogenic control, providing a model for advanced RP and a robust platform to study IMPG2-retinopathy [103].

3. Conclusions

In conclusion, we provided a thorough description of RP genotypes associated with severe phenotypes characterized by childhood-onset and early macular involvement. These subtypes of RP are rare, difficult to manage, and challenging to treat; therefore, biotechnology and pharmaceutical companies are often unwilling to invest in research on them. In this chapter, we summarized the features, expected course, and recent preclinical or clinical studies addressing these orphan diseases and discussed relevant prognostic and therapeutic implications.