Bardet-Biedl syndrome-7 (BBS7) shows treatment potential and a cone-rod dystrophy phenotype that recapitulates the non-human primate model

Abstract

Purpose:

To provide a detailed ophthalmic phenotype of two male patients with Bardet-Biedl Syndrome (BBS) due to mutations in the BBS7 gene

Methods:

Two brothers ages 26 (Patient 1, P1) and 23 (P2) underwent comprehensive ophthalmic evaluations over three years. Visual function was assessed with full-field electroretinograms (ffERGs), kinetic and chromatic perimetry, multimodal imaging with spectral domain optical coherence tomography (SD-OCT), fundus autofluorescence (FAF) with short- (SW) and near-infrared (NIR) excitation lights and adaptive optics scanning light ophthalmoscopy (AOSLO).

Results:

Both siblings had a history of obesity and postaxial polydactyly; P2 had diagnoses of type 1 Diabetes Mellitus, Addison’s disease, high-functioning autism-spectrum disorder and −12D myopia. Visual acuities were better than 20/30. Kinetic fields were moderately constricted. Cone-mediated ffERGs were undetectable, rod ERGs were ~80% of normal mean. Static perimetry showed severe central cone and rod dysfunction. Foveal to parafoveal hypoautofluorescence, most obvious on NIR-FAF, co-localized with outer segment shortening/loss and outer nuclear layer thinning by SD-OCT, and with reduced photoreceptors densities by AOSLO. A structural-functional dissociation was confirmed for cone- and rod-mediated parameters. Worsening of the above abnormalities was documented by SD-OCT and FAF in P2 at 3 years. Gene screening identified compound heterozygous mutations in BBS7 (p.Val266Glu: c.797 T > A of maternal origin; c.1781_1783delCAT, paternal) in both patients.

Conclusions:

BBS7-associated retinal degeneration may present as a progressive cone-rod dystrophy pattern, reminiscent of both the murine and non-human primate models of the disease. Predominantly central retinal abnormalities in both cone and rod photoreceptors showed a structural-functional dissociation, an ideal scenario for gene augmentation treatments.

Introduction

Bardet-Biedl syndrome (BBS) describes an autosomal recessive, genetically heterogeneous condition, characterized by the presence of a photoreceptor degeneration variably associated with truncal obesity, postaxial polydactyly, hypogonadism, renal and cardiac abnormalities, and variable cognitive impairment (1–9). Prevalence estimates are variable and range from 1:18,000 to 1:160,000(5). Retinal degeneration is the most consistent feature(10). Indeed, BBS is one of the most common forms of syndromic retinitis pigmentosa (RP) (5,11). Although there is variability in disease severity, patients typically present with nyctalopia and visual field defects in the first decade of life, progressing to severe central vision loss and legal blindness at a relatively young age (5,12–14). At some point in the disease patients develop a pigmentary retinopathy indistinguishable from the classic appearance of RP, in the majority of the patients associated with a rod-greater-than-cone dystrophy (RCD) pattern of dysfunction, less frequently a cone-rod (CRD) dystrophy, rarely a cone dystrophy (COD) pattern (12–16).

Mutations in up to 22 genes have been identified to date, all encoding proteins expressed in primary cilia, including retinal photoreceptors (6,17,18). BBS is thus considered a ciliopathy, with genetic, mechanistic and phenotypic overlaps with other syndromic ciliopathies such as McKusick-Kaufmann, Alström, Meckel Gruber, Joubert, Senior-Løken, and Mainzer-Zaldino syndromes (6,11,17). The genetic heterogeneity and the role of modifying genes appears to explain the variable severity of the retinal phenotype as well as of the associated systemic abnormalities, even among family members who harbor identical mutations (4,5,14,19). Of the currently known disease-causing genes, eight (BBS1, BBS2, BBS4, BBS5, BBS7, BBS8, BBS9, and BBS18) encode proteins that play a role in assembly of a protein complex known as the BBSome, which plays a crucial role in protein trafficking (11,20–26). Other genes play a role in the assembly (BBS6, BBS10, and BBS12), or protein trafficking (BBS3, BBS14, and BBS17) within cilia (5,27,28). The precise function of other genes remains unknown but are thought to also be closely related to the BBSome. Specifically within the retina, abnormalities are thought to be related to structural and/or functional damage to the cilium connecting photoreceptor inner and outer segments, likely from accumulation of mislocalized proteins (26,29). By far, mutations in BBS1 and BBS10 explain most of the BBS cases.

BBS7 was identified as a disease-causing gene in 2003(4). Although BBS7 mutations account for only 3–7% of the BBS cases, the encoded protein is part of the core of the BBSome complex (11,18,19,24,26,30). The few descriptions of the BBS7 human retinal phenotype are limited and suggest a higher disease burden with a greater expression of cardinal features than patients with other gene mutations such as BBS1 or BBS8 (31–34). The paucity of descriptions of the BBS7 human phenotype contrasts with the availability of animals models of this specific form of BBS, including the recent description of a non-human primate (NHP) model (35). In this report we present a detailed phenotypic description of two patients with BBS7 and draw direct comparisons with the animal models in hopes the exercise will help us gain a better understanding not only of BBS7, but of the much larger group of syndromic retinal ciliopathies, including their treatment potential by gene therapy.

Methods

Two patients underwent a complete ophthalmic examination and testing as part of their standard of care, complemented by detailed retinal phenotyping over the course of 3 years as research. Informed consent was obtained; procedures adhered to the Declaration of Helsinki and were approved by the institutional review board (protocol #815348). Automated static perimetry was performed using a modified Humphrey Field Analyzer (HFA II–i, Carl Zeiss Meditec, Dublin, CA) using a 200-ms long, 1.7° diameter stimuli, presented at 2° intervals along a horizontal profile that extended to 30° of eccentricity. Light-adapted perimetry was performed with an achromatic stimulus on a white (10 cd.m⁻²) background; two color dark-adapted perimetry was performed with 500 nm and 650 nm stimuli (36,37). Spectral sensitivity differences in the dark-adapted state were used to define photoreceptor mediation of the stimuli. The sensitivity profile corresponds to the retinal region scanned with spectral domain optical coherence tomography (SD-OCT; see below). A standard full-field electroretinogram (ffERG) was recorded using a computer-based system (Espion e3, Diagnosys LLC, Littleton, MA), following the recommendations of the standards of the International Society for Clinical Electrophysiology of Vision (ISCEV) (38,39). Rod-mediated responses were elicited with a dim white flash (0.01 phot-cd.s.m⁻²). A white flash (3 phot-cd.s. m⁻²) in the dark adapted state was used to evoke a combined rod-cone response and in light-adapted conditions in response to a 1 Hz or 30 Hz flicker to elicit cone-mediated ERGs.

SD-OCT, en-face near infrared (NIR) reflectance (REF) and fundus autofluorescence (FAF) imaging using NIR and short-wavelength (SW) excitation wavelengths was performed using a Spectralis-HRA system (Heidelberg Engineering GmbH, Heidelberg, Germany). Segmentation of the SD-OCT images was performed with the built-in segmentation software of the Spectralis system, supervised to ensure correct identification of the different laminar boundaries by examining the longitudinal reflectivity profiles (LRPs) with a publicly available software (http://imagej.nih.gov/ij/links.html)(37) (40).

A custom adaptive optics scanning light ophthalmoscope (AOSLO) was also used in this study (41,42). Patients were aligned to the AOSLO using a dental impression. Wavefront sensing was done using an 848 nm superluminescent diode with a full-width at half-maximum bandwidth (FWHM) of 26 nm (Superlum, Ireland) and aberration correction was provided by a 97 actuator deformable mirror (Alpao SAS, France). Confocal and split-detection images were acquired at 17.85 frames per second over a 1° by 1° field of view using a superluminescent diode centered at 795 nm with FWHM of 15.3 nm (Superlum, Ireland) and a photomultiplier tube (PMT, Hamamatsu Corporation, Japan). The patient was instructed to fixate (using the imaged eye) at a target while the AOSLO image sequences were acquired along the temporal meridian. A custom strip-registration algorithm was used for intra-frame strip-based registration and dewarping of the AOSLO images (41,43,44). Reference frames for registration were chosen automatically from the image sequences and 50 frames were registered and averaged (45,46). Averaged images were then automatically montaged using a previously described algorithm (47). Regions of interest (ROI) were at the fovea and along the temporal meridian. Cones, rods, and RPE cells were identified within the ROIs in a semi-automated fashion, using a previously described algorithm, and densities were extracted from the cell locations (48,49).

Results

The proband in this study is a 26-year-old male patient (Patient 1, P1) referred for ophthalmic evaluation accompanied by his asymptomatic brother and mother. He had an ophthalmic history of moderate myopic astigmatism and two strabismus surgeries at ages 10 and 14. His medical history was significant for postaxial polydactyly of the right foot. He was obese (BMI = 50.2) but had no history of diabetes mellitus, kidney disease, or hypogonadism. There was no known parental consanguinity. Parents ancestry was German-Scottish on the maternal side, Irish-Welsh on the paternal side. There was no family history of retinal degeneration or blindness; an uncle, two aunts and grandfather on the maternal side are reportedly myopes. The patient reported blurred vision for the preceding 6 months associated with photophobia necessitating use of sunglasses and a hat. He reported no difficulty with driving at night or negotiating obstacles. He noted difficulty distinguishing dark blue colors. On initial evaluation, best corrected visual acuity was 20/30 in the right eye and 20/20 in the right eye (Spherical equivalent, SE, −0.50D) in each eye. He demonstrated a right intermittent exotropia. Color vision testing (Farnsworth, D15), intraocular pressures, and anterior segment examinations were normal. Fundus examination demonstrated a tilted optic nerves and parafoveal depigmentation as a subtle bull’s eye maculopathy (Figure 1a, P1). The fundus exam appearance did not change over the 3-year follow up period. SW-FAF imaging showed an incomplete parafoveal annulus of tenuous hypo-autofluorescence, more obvious in supero-nasal parafovea that surrounds a small ring of juxtafoveal hyper-autofluorescence (Figure 1a). By contrast, there was a complete parafoveal annulus of dense hypo-autofluorescence on NIR-FAF imaging that surrounds tightly a normal appearing foveal center (Figure 1a).

Figure 1.

En-face multimodal imaging in BBS7 patients. (a) Wide angle color fundus photography and fundus autofluorescence elicited with short-wavelength (SW-FAF) and near-infrared (NIR-FAF) excitation lights in both patients. The images correspond to an intermediate visit on their follow up (P1, age 29; P2, age 22) with the highest quality, wide-angle color fundus photography available. Normal appearance of SW-FAF and NIR-FAF imaging are shown as insets of P2. Vertical yellow arrows point to depigmented halo around the foveal center in P1 and to a darker than the surrounding center in P2. Diagonal red arrow on the SW-FAF image of P2 points to juxtafoveal region of increased SW-FAF, which is close to the normal ring of increased autofluorescence caused in the normal subject by a local reduction of the macular pigment optical density in the juxtafovea (shoulder of the macular pigment) in a large proportion of otherwise normal subjects. Representative normal FAF images are shown as insets forcomparison. (b) SD-OCT horizontal, 8 mm cross sections through the fovea of the patients. Nuclear layers are labeled in P1: outer nuclear layer = ONL, inner nuclear layer = INL, ganglion cell layer = GCL. Outer retinal sublaminae are labelled (diagonal arrows) according to conventional nomenclature: 1. Outer limiting membrane (OLM), 2. Inner segment ellipsoid region (ISe or EZ), 3. The contact cylinder between the apical RPE microvilli and the photoreceptor outer segments tips, or interdigitation zone (IZ), 4. Basal RPE and Bruch’s membrane (RPE/BrM). The outer plexiform layer (OPL) is also labelled in P2. Vertical yellow arrows in the patients denote juxtafoveal segment with abrupt outer retinal (OLM, EZ and IZ) changes where the retina transitions from a near normal appearance on SW- and SW-FAF imaging to deep hypoautofluorescence in (a). Scale bars to the left. T, temporal, N, nasal retina.

The patient’s 22-year-old only brother (P2) was asymptomatic. He had an ophthalmic history notable for high myopia in both eyes (SE −11.50). On questioning medical history was notable for similar postaxial polydactyly, Type 1 Diabetes Mellitus, Addison’s disease, obesity (BMI = 32.3), and high-functioning autism spectrum disorder. His best corrected visual acuity was 20/30 in each eye. Color vision testing revealed multiple errors with a tritan axis of confusion. Intraocular pressures were normal and anterior segment examination was unremarkable. Fundus examination revealed a myopic conus and a depigmented peripapillary and nasal fundus appearance with visualization of the choroidal vasculature (Figure 1a, P2). The normal SW-FAF signal is centered by a dark fovea resulting from pre-RPE screening by the macular pigment (MP) of the SW excitation lights used in this work; the normal NIR-FAF is characterized by a local maximum near the foveal center caused by greater content of melanin within taller central RPE cells (Figure 1a, insets) (50–52). SW-FAF in P2 demonstrated a darker than normal foveal center surrounded by a ring of juxtafoveal hyper-autofluorescence (Figure 1a, P2, red arrow). The local hyperautofluorescence exceeded the normal increase in SW-FAF signal in the juxtafovea caused by a local reduction in the MP optical density (MPOD) or ‘shoulder’ of the MPOD, illustrated in a representative normal subject shown (Figure 1a, inset, red arrow) (51,53–56). The central round hypoautofluorescent region had a similar appearance on NIR-FAF also surrounded by a juxtafoveal ring of mild NIR hyperautofluorescence (Figure 1a).

SD-OCT imaging was used to explore the underlying structural abnormalities in cross-section (Figure 1b). P1 showed obvious foveal and juxtafoveal thinning of the photoreceptor outer nuclear layer (ONL), including the photoreceptor axons. At the foveal center there is approximation of the inner segment ellipsoid band (EZ) to the apical retinal pigmented epithelium and Bruch’s membrane (RPE/BrM) signals, likely from shortening and/or loss of photoreceptor outer segments (Figure 1b) (40,57–63). While the EZ and outer limiting membrane (OLM) are visible at the foveal center, the EZ becomes undiscernible in the juxtafoveal region (<0.75 mm of eccentricity), where the OLM abuts the apical RPE (Figure 1b, between yellow arrows) (64,65). This segment corresponds to the area of hypoautofluorescence on SW-FAF and NIR-FAF that is slightly eccentric toward the nasal side of the fovea (Figure 1a). With increasing eccentricity into the parafovea and pericentral retina, the retinal lamination regains a normal appearance with the exception for the existence of a darker than usual band separating the EZ signal from the signal originating from the tip or distal photoreceptor outer segments as they interdigitate with the apical RPE, conventionally known as the interdigitation zone (IZ) band (60,63). The foveal center in P2 appears to be normal in thickness and the only visible abnormality is the loss of the IZ signal with approximation of the EZ to the apical RPE with an intervening hyporreflective signal. In the juxtafovea there is a brief interruption of the EZ band (Figure 2b, between yellow arrows). The association of polydactyly, obesity, endocrine abnormalities and an outer retinal degeneration was suggestive of Bardet-Biedl syndrome. Genetic testing by Next Generation Sequencing identified two novel variants in the BBS7 gene segregating in a compound heterozygous state with the phenotype in both patients: p.Val266Glu:c.797 T > A, maternally inherited, predicted to be damaging by both Polyphen-2 and REVEL, and c.1781_1783delCAT, an in-frame deletion of p.Ser594, inherited from their father. Both patients were negative for mutations in BBS1, BBS2, BBS4, BBS5, BBS9, BBS10, BBS12, and ALSM1 genes (Molecular Vision Laboratories, Panel v1). The involvement of other BBS variants was not pursued.

Figure 2.

Longitudinal changes in retinal structure over a three-year interval in BBS7. (a) NIR-FAF images in both patients at two visits. (b) Magnified 2 mm-long SD-OCT cross-sections from the two visits. Horizontal dashed bar at the bottom of NIR-FAF panels in (a) delimit the horizontal extent sampled by the OCT scans. Vertical dashed lines define the peripheral boundary of the juxtafoveal area hypoautofluorescence on NIR-FAF at the first visit that corresponds with a juxtafoveal segment where the EZ band is interrupted. Vertical short solid line parallel to the dashed lines in temporal retina denotes the location of the re-emergence of the EZ band at 3 years which corresponds with the centrifugal movement of this transitional zone. Asterisks = juxtafoveal increased posterior signal scattering due to RPE depigmentation. Arrows point to linear intraretinal hyperreflectivities that may reflect both intraretinal pigment migration and/or Muller cell hypertrophy. Arrowhead points to an hyporreflective area apical to the basal RPE/BrM that appears on follow up in P1. Vertical bars in P2 compared the length of the distance between the EZ and the RPE/BrM at baseline compared to 3 years of follow up.

Longitudinal changes and structural–functional relationships in BBS7

Structural changes over a 3-year period were clearly detectable on NIR-FAF in P1 with further loss of the signal and both a centripetal and centrifugal expansion of the juxtafoveal annulus of hypoautofluorescence, whereas P2, the younger brother, showed no discernible changes (Figure 2a). The hypoautofluorescent lesion is delineated in both patients by a thin hyperautofluorescent contour. On SD-OCT there was further foveal thinning in P1, the EZ band no longer detectable within the foveal center (Figure 2b). The juxtafoveal regions of hypoautofluorescence on NIR-FAF co-localized with segments on the SD-OCT cross-sections with increased signal scattering posterior to the RPE (Figure 2b, asterisks), and intraretinal hyperreflectivities (Figure 2b, diagonal arrows), that may reflect intraretinal pigment migration, Muller cell hypertrophy, or changes in reflectivity of degenerating photoreceptors. The transitional zones where the EZ is lost correspond with the hyperautofluorescent boundary line that surrounds the hypoautofluorescent lesions in each patient (Figure 2b, vertical dash lines). On follow-up of P1, signals from the apical RPE in the segment with total loss of the inner and outer segment in juxtafoveal nasal retina are substituted by a hyporreflective void, suggesting RPE degeneration and/or loss following the earlier photoreceptor abnormalities (Figure 2b, arrowhead). Changes in the younger brother were limited to the foveal center with a narrower EZ-to-RPE/BrM distance 3 years after his initial baseline visit (Figure 2b, overlaid vertical bars). In both brothers there were small centrifugal movements of the interruption of the EZ band towards the temporal parafovea (Figure 2b).

The peripheral visual field extent measured with kinetic perimetry and a V-4e target was moderately constricted in both patients, severely constricted to the smallest size I-4e target. Full-field electroretinograms (ffERGs) showed non-detectable cone-mediated signals whereas rod ERGs were ~80% of the normal amplitudes (Figure 3a). Central retinal function assessed with light- and dark-adapted automatic static perimetry revealed abnormally reduced cone-mediated light-adapted sensitivities by at least one log unit across the central retina, except near fixation where sensitivities approached the lower limit of normal (Figure 3b). Contrasting with the normal retina-wide rod functioning by ffERG, there was also definite rod dysfunction across the central in both patients, locally severe in the nasal retina of P1 where sensitivity losses exceeded 2 log units (Figure 3b). In contrast to this central dysfunction, particularly the cone dysfunction by electroretinography, quantitation of the central retinal structure showed a different picture (Figure 3c). The thicknesses of the central ONL, which includes the photoreceptor nuclei and their laterally displaced axonal projections, and of the ‘outer retinal sublaminae’, a term that corresponds to the distance from the photoreceptor inner segment ellipsoid region (conventionally named inner segment ellipsoid band EZ) to the RPE/BrM were comparatively preserved in both patients (Figure 3c) (57,66,67). At the foveal center the ONL thickness was near the lower limit of normal in P2, thinner than normal in P1. The ONL remained at the lower limit of normal or slightly thinner than normal in both patients, with a local dip between 0.5 and 1 mm of the foveal center (Figure 3c).

Figure 3.

(a) Standard full-field ERGs in the patients compared with a representative normal subject (gray traces). (b) Light-adapted achromatic and dark-adapted chromatic (500 nm) horizontal sensitivity profiles in the patients compared with the normal range (gray bands, normal mean ± 2SD). Sensitivities to the 500-nm stimulus are confirmed mediated by rods through the use of spectral sensitivity (500 nm – 650 nm) differences. Hatched bar denotes the location of the blind spot. N, nasal. T, Temporal. (c) Thickness of the outer nuclear layer (ONL) along the horizontal meridian at eccentricities that co-localize with the sensitivities measured with static perimetry in (B). Gray bands: normal limits (mean ± 2SD).

The modest changes in the overall thickness of the photoreceptor ONL layer in our patients contrasted with the depth of the retinal dysfunction, in particular the severe and retina-wide loss of cone function, prompted exploration of changes of the distal photoreceptors and apical RPE with the use of LRP analyses (Figure 4) (40,62,63,68). LRPs can be used to unambiguously visualize the signal transitions of the OCT cross-sections. In the normal retina major signal troughs correspond to the nuclear layers, while broad hyperscattering signal peaks correspond to the plexiform and retinal nerve fiber layers (Figure 4a). The LRP waveform for these major components have a normal appearance in the BBS7 patients (Figure 4a). A magnified view to the LRP segment distal to the OLM in the normal retina reveals closely spaced hyperscattering peaks that vary in number and position with increasing distance from the foveal center (Figure 4b). (40,58,59,62,63,68–70) The peaks correspond to the ellipsoid region of the photoreceptor inner segments or ellipsoid band (ISe or EZ), the contact cylinder between the apical RPE and the photoreceptor outer segments tips, conventionally termed the interdigitation zone (IZ), and the basal RPE and Bruch’s membrane signals (RPE/BrM) (40,58–63,68–73). Within the rod-free foveola the distance between the ISe and the EZ peak, which relates to the length of the cone photoreceptor outer segments, is slightly shorter in P2 with milder changes, but severely shortened in P1 (Figure 4b, red segment on LRPs). A short distance from the foveal center in nasal retina of a normal subject illustrates the complexity of the signals. Here the peak that corresponds to the COST/IZ peak at the foveola moves proximally (or superficially) (Figure 4b, red segment in normal tracing), the remaining segment of the LRP to the RPE/BrM peak represents the distal end of the rod photoreceptor outer segment (ROS) (Figure 4b, i segment in normal trace). Both patients do not show the peak associated with the cone outer segment signal and the EZ-to-RPE/BrM distance is shorter compared to normal (Figures 4d and 4c). The combination of a moderately thin ONL with obvious outer segment abnormalities for both cone and rod photoreceptors suggests reduced cone photoreceptor densities and the presence of rods with abnormally shorten outer segments, providing an explanation for the sensitivity losses for both photoreceptor mechanisms (Figure 4c).

Figure 4.

(a) Magnified 2.5 mm horizontal SD-OCT cross-sections from the fovea into the nasal retina in the two patients on their first visit are compared with a representative normal subject. Overlaid white traces are LRPs from a location ~2 mm nasal from the foveola. (b) LRPs segments distal to the outer limiting membrane (OLM) are magnified to explore changes in the outer retinal sublaminae. LRPs from the foveola and 2 mm nasal [white boxes overlaid on OCT scans in (A)] are compared with normal LRPs (gray traces). The various peaks that correspond with the different outer retinal sublaminae are labelled as in Figure 1. At the foveola, LRP segments colored red denote the signal between the ellipsoid region of photoreceptor inner segments (ISe or EZ) and the contact cylinder between the apical RPE microvilli and the cone outer segment tips (COST), which relates to the length of the foveolar cone outer segment length. LRP segments in blue at the 2 mm location denote the signal that bridges the distance between the ISe and the apical RPE/BRM, which relates to the distance spanned by rod, and intermingled cone (overlapping red segment) outer segments. (c) Thickness of the ONL and the length of the ISe-to-RPE/BrM in patients compared to the lower limit of normal (mean – 2SD) for both parameters (gray horizontal bars). F, foveola. N and T are measures from 2 mm nasal and 3.6 mm temporal to the fovea, respectively.

To further explore the photoreceptor abnormalities AOSLO imaging was performed in both patients (Figure 5 P1, Figure 6 P2). P1 showed a tightly packed inner segment cone mosaic at the foveal center surrounded by a region of cone loss corresponding to the bull’s eye lesion (Figure 5). Peak cone density within the foveal region was measured as 31,736 cones/mm² which is 27% of normal density (normal foveal density: 119,000 ± 23,300 cones/mm²)(48). The foveal cones and a pocket of cones at 0.6 mm temporal to the fovea displayed waveguiding of outer segments on confocal AO images, however the retained waveguiding signal appeared mottled and dim in comparison to brightly reflective and Gaussian-shaped profiles of normally waveguiding cones (Figure 5, confocal 0.6 mm Temporal). Cone density in this location was 13,991 cones/mm² or approximately 44% of normal density (normal 31,500 cones/mm2). Beyond this location, cone density was further reduced; at 1.3 mm temporal cone density was 21% of normal (3,331 cones/mm² compared to 15,800 cones/mm²) and did not exhibit waveguiding outer segments (Figure 5). Though rods at this location did retain waveguiding outer segments, rod density was reduced to 59,000 rods/mm² (normal rod density: 80,000 rods/mm²) (74–77). RPE density was on the high side of normal at this location and was measured as 6,330 RPE cells/mm. P2 displayed a similar adaptive optics ophthalmoscopy phenotype (Figure 6) as P1. Peak cone density in the fovea was 97,814 cones/mm² or 82% of normal. Foveal cones retained normal waveguiding reflectance properties, although there was a small region in the superior-nasal parafovea that displayed a patch of abnormally dim waveguiding cones (Figure 6). The cone inner segment mosaic was intact at this location. Cones at 0.5 mm temporal (Figure 6) were reduced in density (20,186 cones/mm2 or 52% of normal) and displayed the abnormal mottled waveguide appearance described at the same location in P1. Beyond this location, cone density was further reduced and cones did not waveguide (Figure 6). Cone density at 1.9 mm temporal was 2,577 cones/mm2 (22% of normal) and rod density was reduced to 56,262 rods/mm (49% of normal). Overall, adaptive optics imaging showed wide-spread loss of the photoreceptors, with a more advanced phenotype in the cone outer segments than the inner segments.

Figure 5.

Adaptive optics images in P1. Top. Split-detection adaptive optics montage of the inner segment mosaic. Widespread loss of cone inner segments is observed. Bottom. Magnified regions of interest. Split-detection adaptive optics imaging at the fovea (large image) reveals an intact mosaic, cross-hair located at the reduced peak cone density of 31,736 cones/mm². Asteriskscorrespond to the bull’s eye lesion where the cone mosaic is no longer visible. Split-detection, confocal and dark-field adaptive optics images at 0.6 mm Temporal () show a pocket of retained cone inner and waveguiding outer segments and retinal pigment epithelium, respectively. At 1.3 mm Temporal (eccentricity of) cone and rod densities are reduced, cones do not normally waveguide while rods do and retinal pigment epithelial cell density is normal.

Figure 6.

Adaptive optics images in P2. Top. Confocal adaptive optics montage of the photoreceptor mosaic. Widespread loss of cone waveguiding is observed albeit with abnormally retained waveguiding in the temporal parafovea and normal-appearing waveguiding in the foveal center. Bottom. Magnified regions of interest. Confocal adaptive optics imaging at the fovea (large) reveals an intact and waveguiding mosaic, cross-hair located at the reduced peak cone density of 97,814 cones/mm². Arrowheads identify a region of dimly waveguiding cones, perhaps a precursor to a bull’s eye lesion. Split-detection and confocal adaptive optics images at 0.5 mm Temporal again show the cone inner segment mosaic with mottled but retained waveguiding. At 1.9 mm Temporal cone and rod densities are reduced and cones do not normally waveguide while the rods do. Dark-field images show the RPE mosaic.

Discussion

To date, there are only limited descriptions of the retinal phenotype of BBS7. The reported cases have shown a severe juvenile-onset pigmentary retinopathy with non-detectable ERGs or small residual cone ERG signals, invariably interpreted as being within the spectrum of retinitis pigmentosa (32,78–82). In the present study, we describe two brothers from a non-consanguineous family with novel compound heterozygous mutations in BBS7 who presented in the third decade of life with minimal photophobia and blurred vision in the oldest patient, and with virtually no visual symptoms in his highly myopic brother. Visual acuities were within normal limits and there were only subtle color vision abnormalities. Both brothers had associated systemic abnormalities, more numerous in the youngest brother with subclinical retinal changes, fitting a clinical diagnosis of BBS. The retinal exam was dominated by severe juxtafoveal loss of the photoreceptor outer segments and ONL thinning that worsened within a 3-year interval. Foveolar thinning, while obvious in the oldest subject, was nearly undetectable in P2, the youngest brother. Most interestingly, full-field cone ERGs were non-detectable, which contrasted with modest reductions of rod photoreceptor function, somewhat reminiscent to the pattern observed in achromatopsia(83). Static perimetry showed central cone and rod dysfunction in a pattern that may correspond with the earliest stages of a cone-rod dystrophy(84). This pattern of retina-wide cone>rod ERG abnormalities, while previously unreported for BBS7, is well modelled in the recently described non-human primate model of the disease(35).

Variability of expression of the retinal and systemic abnormalities has long been recognized in BBS, dating back to the pre-molecular era, making BBS an example of pleiotropism (10,14,85). The cases reported herein are encouraging as they suggest a molecular diagnosis of BBS7 should not automatically equate with early and severe vision loss (10–13). They also provide a view into the mildest abnormalities in this specific BBS subtype and thus a possible insight into the pathophysiology of the disease. Of interest, macular changes with severely reduced visual acuities have been documented and nystagmus reported in BBS7, which indicate severe, early central involvement in this specific form of BBS, and suggests a spectrum of severity, with our cases representing the mildest end of the spectrum (33,86,87). In the previously reported cases, a fast and early phase of cone (or central) photoreceptor degeneration may have preceded, with the disease later converging into a severe retinal degeneration indistinguishable from RP, as described in other forms of cone-rod dystrophy (88). Such scenario may explain the different functional classification and apparent discrepancy of the phenotype between our patients and earlier BBS7 reports. Follow up of our cases into the future will help further elucidate this possibility.

The reasons for the initially severe cone dysfunction, milder rod involvement, and relative preservation of the retinal structure in our patients and reported in BBS are unclear and deserves further exploration. The similarities with the recently described NHP model of the disease, however, are extraordinary and deserve mention as they were interpreted as a model of classical RP(35). Both the BBS7-NHP model and the BBS7 patients reported herein (and some features reported in the BBS7 literature) show a predominantly central degeneration with significant cone dysfunction (33,35,86,87). Central involvement with bull’s eye maculopathies and central atrophic lesions with predominant cone dysfunction have been repeatedly reported in BBS and seem to be regaining attention (11,14–16,89–91). Underrepresentation of the pattern in the BBS literature may result from the extension, as noted above, of a predominant central disease into a retina-wide degeneration, with convergence of an earlier cone>rod phenotype into a severe, retina-wide photoreceptor degeneration involving rods and cones similarly, a sequence that may be better recognized with earlier molecular diagnoses(88). Experimental work suggests there may be greater susceptibility for cone degeneration in certain forms of BBS (BBS3, BBS5, BBS6, BBS8, BBS21) (25,27,92–94). And a recent study reported a cone-rod dystrophy pattern with a very similar phenotype to that described in our patients in several forms of the disease (BBS1, BBS5, BBS6, BBS10, BBS12) with predominance (5/7 patients) of mutations in genes that encode proteins within the chaperonin complex (15). Juxtafoveal changes and relative preservation of the retina with increasing eccentricity was also documented in that series (15). Among the syndromes overlapping manifestations with BBS, Alström syndrome classically shows a severe early cone dysfunction, as well as in INPP5E-associated retinopathies, isolated or as part of Joubert syndrome (5,95–100). The severe and early cone dysfunction described in achromatopsia (ACHM) and blue cone monochromatism (BCM), a molecularly heterogenous group of conditions, resembles the severe cone dysfunction of BBS and Alström syndrome, albeit without the severe rod degeneration that accompanies the latter two conditions (73,101,102). Points of contact between the mechanism of these conditions through the function of the BBSome and/or with the unique physiology of the cone photoreceptor may explain the similarities and eventually reveal a common mechanism (71,103).

In BBS there is greater juxtafoveal/parafoveal involvement, at least initially, than there is foveal involvement, which contrasts with the primarily foveolar and minimally progressive disease in ACHM. The reason for the initial regional predilection for the juxtafovea with relative sparing of the foveola observed in BBS, even in cases with a rod-cone pattern of degeneration in BBS, is unknown. One possibility is the additional involvement of rod photoreceptors, which become more numerous immediately outside of the foveal center (74,104). The cases described here had also locally severe RPE depigmentation and loss, including at the foveal center of P2 when local changes were limited to mild shortening of the cone outer segment, a pattern that has been described in BBS, as well as in ACHM. It is possible that the BBS ciliopathy interferes directly or indirectly with processes critical for the interdependent photoreceptors and RPE cells (71,105,106). Using AOSLO we documented severely reduced cone photoreceptor densities all across the central retina and to a similar extent in both patients, independent of the degree of degeneration. The finding may be consistent with an early cone loss phase and/or a developmental failure as proposed for blue cone monochromatism(73). The presence of a relatively intact ONL in the same regions where the loss of the photoreceptor distal organelles was documented by AOSLO suggests relative sparing of the cells, at least for some time, after the loss of the inner and outer segments. Symptomatic presentation may relate to a phase when the disease crosses into a degenerative stage starting with the cells immediately surrounding the foveal center. Determining the factors that lead to this outcome warrants further study.

Restoration of the physiology of a renewable organelle of terminally differentiated neurons, such as the photoreceptor outer segment, makes retinal ciliopathies, including BBS and related conditions, attractive targets for intervention (107,108). There are now several examples of improvements of the retinal and systemic phenotype of several BBS animal models through various approaches, including forms of gene therapy (17,107,109–115). Retinal gene therapy in the bbs1, bbs4, bbs10 and bbs17 mouse models have shown promising structural and functional results, suggesting that proper BBSome complex formation can be restored along with restored ciliary protein trafficking with improved retinal function (108,109,112,113,115). Similarly encouraging results have been demonstrated in the treatment of NPHP5- and CEP290-associated ciliopathies, which involve proteins intrinsically associated with the BBSome (116,117). The phenotype of the BBS7 siblings described herein include a disproportionate loss of cone and rod vision with relatively preserved photoreceptors. This clinical picture likely represents not a unique phenotype, but rather, a view through a time-window of an otherwise progressive retinal degeneration where interventions may have the best odds at restoring vision, as has been already demonstrated in proof-of-concept experimental studies in BBS and in patients with early and severe inherited retinal degenerations (108,118–123). Although proof-of-concept treatments for BBS7 disease do not exist today, the possibility of variations of the BBS7 phenotype described herein exists. This, together with the availability of the BBS7 NHP animal model of the disease, provide the needed motivations for developing future interventions for BBS7(35). Thus, despite the low prevalence of BBS7 as a cause of BBS, the important role of the protein in the pathophysiology of the much larger group of retinal ciliopathies as well as the existence of an ideal animal model, position BBS7 as an ideal target for experimental treatments that will both, increase our understanding of these conditions and ultimately move treatments into the clinic for patients suffering from these complex syndromes (17,18,86).