|Home | About | Journals | Submit | Contact Us | Français|
Albinism, an inherited disorder of melanin biosynthesis, disrupts normal retinal development, with foveal hypoplasia as one of the more commonly associated ocular phenotypes. However the cellular integrity of the fovea in albinism is not well understood – there likely exist important anatomical differences that underlie phenotypic variability within the disease and that also may affect responsiveness to therapeutic intervention. Here, using spectral domain optical coherence tomography (SD-OCT) and adaptive optics (AO) retinal imaging, we obtained high-resolution images of the foveal region in six individuals with albinism. We provide a quantitative analysis of cone density and outer segment elongation demonstrating that foveal cone specialization is variable in albinism. In addition, our data reveal a continuum of foveal pit morphology, roughly aligning with schematics of normal foveal development based on post-mortem analyses. Different albinism subtypes, genetic mutations, and constitutional pigment background likely play a role in determining the degree of foveal maturation.
Albinism is an inherited disorder of melanin biosynthesis, associated with absent or reduced melanin pigment in the eye, and often in the skin and hair. Oculocutaneous albinism type 1 (OCA1;MIM 203100) is a recessive disorder in which individuals have a mutation in the tyrosinase gene on chromosome 11q14.3. Those who never develop any melanin pigment within the eye, hair, and skin have OCA1A, while those with development of some melanin pigment due to a “leaky” mutation that allows residual enzyme activity have OCA1B (King, Pietsch, Fryer, Savage, Brott, Russell-Eggitt, Summers & Oetting, 2003). Ocular albinism (OA1; MIM 300500), is characterized by X-linked inheritance and has been mapped to Xp22.3-Xp22.2. The affected males typically have normal skin and hair pigment, but will usually have all the ocular-visual manifestations of albinism, including iris transillumination, macular translucency, photosensitivity, refractive errors, astigmatism, nystagmus, impaired stereopsis, altered retinostriate decussation, and reduced visual acuity (Oetting, Summers & King, 1994, Summers, 2009).
The normal human fovea underlies the majority of our visual function, including color vision and high spatial acuity vision, and is characterized by an avascular zone, an increase in cone photoreceptor density, and an excavation of inner retinal neurons (Hendrickson, 2005). Foveal manifestations of albinism can include absence of a foveal avascular zone (Abadi & Pascal, 1989, Gregor, 1978), foveal hypoplasia (absence of a foveal pit) (Kinnear, Jay & Witkop, 1985, Oetting et al., 1994), and loss of an annular reflex (Lee, King & Summers, 2001). Reductions in visual acuity has been attributed to absent melanin synthesis and the absence of a foveal pit (Seo, Yu, Kim, Choung, Heo & Kim, 2007), however certain patients with some melanin pigment in their maculae due to residual enzyme activity still show decreased visual acuity and others with poorly defined foveal pits have relatively preserved visual acuity (Harvey, King & Summers, 2006, Summers, 1996). As such, significant work remains in elucidating the relationship between foveal maldevelopment and visual function in albinism (Jeffery, 1997). This lack of clarity stems in part from ambiguity surrounding foveal anatomy in albinism, specifically with regard to foveal cone specialization.
There are but a few cases in the literature that deal directly with the issue of cone specialization in human albinism, and they are far from unified in their findings. In a postmortem analysis, Fulton et al. (1978) found that central cone density in a patient with OCA was significantly decreased compared to values normally found in the parafoveal region (Fulton, Albert & Craft, 1978). In another histological study on an OCA retina, increased cone density was observed in the fovea compared to the periphery, though it was not reported whether the foveal density was consistent with normal values (Akeo, Shirai, Okisaka, Shimizu, Miyata, Kikuchi, Nishikawa, Suzumori, Fujiwara & Majima, 1996). Wilson, Mets, Nagy & Kressel (1988) measured grating acuity in two OCA patients and found a defect in spatial processing in the central retina that was best explained by an increased spacing of the central cone photoreceptors. Based on multifocal electroretinogram findings, it has been inferred that the density of cone photoreceptors across the central retina is homogeneous in OCA (Kelly & Weiss, 2006) and in ocular albinism (OA) (Nusinowitz & Sarraf, 2008), consistent with an underdeveloped macular region.
Recent advances in high-resolution imaging make it possible to quantitatively assess foveal morphology in the living retina. Optical coherence tomography (OCT) offers high axial resolution and enables visualization of retinal lamination, while adaptive optics (AO) provides high lateral resolution and allows direct visualization of the cone photoreceptor mosaic. Here we applied these imaging tools to characterize the foveal cone specialization in four individuals with OCA1B and two with OA1. We observed varying degrees of foveal maturity and cone specialization in the albinotic retina, which roughly align with schematics of normal foveal development based on post-mortem analyses (Isenberg, 1986, Mann, 1950, Provis, Diaz & Dreher, 1998, Springer, 1999). This supports the idea that, in general, normal foveal development is arrested in individuals with albinism (Wilson et al., 1988). As such, examining foveal morphology and cone specialization in albinism may offer valuable insight into the process of normal foveal development. In addition, as previous work suggests that it may be possible to recover some aspect of foveal maturity in albinism through L-DOPA supplementation (Ilia & Jeffery, 1999, Ilia & Jeffery, 2000, Lopez, Decatur, Stamer, Lynch & McKay, 2008, Reis, Ventura, Kubrusly, de Mello & de Mello, 2007), better, more quantitative measures of foveal architecture in albinism are needed. The tools described here should prove useful for dissecting retinal versus cortical roles in vision loss in albinism, evaluating therapeutic strategies for albinism, and for more accurately characterizing post-natal foveal development in the normal retina.
This study adhered to the tenets of the Declaration of Helsinki and was approved by the Children's Hospital of Wisconsin Institutional Review Board. Participants provided written informed consent after explanation of the nature and possible consequences of the study. Two subjects with X-linked OA1 (both male), four with OCA1B (three male and one female), and 167 control subjects without albinism (72 males, 95 females with a mean age of 32.6) were recruited for retinal imaging. The control subjects excluded individuals with color vision deficiency or history of ocular surgery or other diagnosed retinal abnormalities. Fundus photographs were obtained from the subjects with albinism and were evaluated for fundus pigmentation, foveal reflex, foveal vasculature, and general anatomical features. These photos, in addition to clinical examination, were used to classify each of the subjects with albinism by phenotype. A summary of clinical characteristics can be found in Table 1.
Volumetric images of the macula were acquired using the Zeiss Cirrus HD-OCT (Carl Zeiss, Meditec, Dublin CA, USA). Volumes were 6mm × 6mm and consisted of 128 B-scans (512 A-scans/B-scan). At least three replicate scans were acquired to assess repeatability of fixation and allow an estimation of the preferred retinal locus of fixation. For each subject with albinism, images containing a Cirrus LSO (laser scanning ophthalmoscope) map, retinal thickness map, and volume scan boundaries were extracted and registered to a color fundus image using i2k Align Retina software (DualAlign, LLC, Clifton Park, NY USA). The LSO images with OCT boundaries contain a cross hair depicting the center of the scan, which by design, corresponds to fixation for that particular volume scan. After all replicate Cirrus LSO maps were registered to the color fundus image, the center of fixation was identified for each image, and their absolute positions within the fundus recorded. The center of mass of these locations was calculated to derive a single estimate of the preferred retinal locus of fixation. In addition, the position of the center of the retinal doming was approximated, and its position within the fundus recorded for each subject. The expected location of the fovea was also calculated based on data from (Rohrschneider, 2004). Some subjects had cyclotropia, and their photos were corrected prior to calculation of the expected foveal location (Bixenman & Von Noorden, 1982). In all cases, the locus of fixation was close to the expected foveal location, however errors in this calculation could affect our absolute topographical analysis of cone density.
High-resolution SD-OCT (Bioptigen, Inc., Durham, NC) imaging was performed in all six subjects with albinism and the 167 controls. Line scan sets were acquired (1000 A-scans/B-scan; 100 repeated B-scans) through the center of the foveal center (either the actual fovea or expected foveal location when there was no visible pit). Scans were registered and averaged as previously described to reduce speckle noise in the image (Tanna, Dubis, Ayub, Tait, Rha, Stepien & Carroll, 2009).
The inner limiting membrane (ILM), external limiting membrane (ELM), inner segment-outer segment junction (IS/OS), and two layers representing the retinal pigment epithelium (RPE1, RPE2) were manually segmented. Demarcated retinal layers were then interpolated and layer thicknesses were extracted using programs developed with Matlab™ (Mathworks, Natick, MA). The distance between the ILM and RPE2 provided total retinal thickness, the distance between the center of the ELM and center of the IS/OS junction provided inner segment (IS) thickness, while the distance between the center of the IS/OS and center of RPE1 provided the outer segment layer (OS) thickness. The scan length of each OCT scan set was corrected for inter-individual differences in axial length based on Bennett, Rudnicka, & Edgar (1994). Interpolated thickness values for each subjects' IS and OS layer were binned and averaged at 0.1 mm increments between 2 mm nasal and 2 mm temporal retina. The relative foveal lengthening of the IS (or OS) layer was computed by dividing the foveal IS (or OS) thickness by the average IS (or OS) thickness at 1.75 mm nasal and temporal to the fovea.
An AO fundus camera was used to acquire images of the cone mosaic. The subjects' eye was dilated and accommodation suspended using one drop each of phenylephrine hydrochloride (2.5%) and tropicamide (1 %). The AO system housed at the Medical College of Wisconsin has been described in detail elsewhere (Rha, Schroeder, Godara & Carroll, 2009), a brief summary of the system is given here. The eye's monochromatic aberrations are compensated for using a 52-channel deformable mirror (Imagine Eyes, Orsay, FRANCE). Retinal images are acquired by illuminating the retina with a 1.8-degree diameter, 500 ms flash from a fiber-coupled near infrared source (λ = 837.8 nm; Δλ= 14.1 nm). To reduce speckle noise in the image the spatial coherence of the laser was reduced by using 110 meters of multimode step index fiber (Fiberguide Industries, Stirling, NJ USA). A back-illuminated scientific-grade 12-bit charge-coupled device, the Cam1M100-SFT (Sarnoff Corporation, Princeton, NJ), captured images of the retina. During each 500 ms flash, continuous images of the retina are collected at a frame rate of 167 fps with 6 msec exposure. Individual frames are registered then averaged using a MatLab-based algorithm (MathWorks, Natick, MA) for subsequent analysis.
Individual cones in each image were identified using a semi-automated cone identification program based on the work of Li & Roorda (2007). As the results of this algorithm depend on the spatial filter applied to the image, we added an objective procedure for setting this filter. The user selected a group of cones manually, and the mean nearest neighbor distance of these cones was used to set the spatial filter. The filtered image was then fed through the automated algorithm to identify individual cones. From this data cone density (cones/mm2) was calculated as previously described (Carroll, Neitz, Hofer, Neitz & Williams, 2004). Regularity of the cone mosaic was assessed using a previously described Voronoi analysis (Baraas, Carroll, Gunther, Chung, Williams, Foster & Neitz, 2007). In all cases, the approximate retinal location of the cone mosaic images relative to the incipient foveal center was determined based on the fundus images using common blood vessel patterns and that subjects' suspected locus of fixation.
To qualitatively assess gross foveal topography, we acquired both cross-sectional and volumetric images of the central macula using SD-OCT (Fig. 1). The high-resolution foveal cross sections from the Bioptigen SD-OCT system (Fig. 1, panels A-F) and the macular thickness maps obtained from Cirrus HD-OCT volumes (Fig. 1, panels I-N) reveal a continuum of foveal maturity across the albinotic subjects. Some individuals have nearly planar thickness across the macula (Fig. 1, panels A, B, I, J), one has marked doming in his retinal thickness (Fig. 1, panels C & K), and the others have a shallow foveal depression and some excavation of inner retinal neurons in a region of increased retinal thickness (Fig. 1, panels D-F & L-N). The “doming” of the macular region is thought to precede the formation of a fovea centralis (Provis et al., 1998, Springer, 1999). The two normals (Fig. 1, panels G & H and O & P) represent the minimum and maximum foveal pit depth observed in the 167 normal controls in this study and illustrate the degree of variation in normal foveal pit morphology (Dubis, McAllister & Carroll, 2009). Among those subjects with albinism without a pit, there was no discernable lengthening of the photoreceptor OS at the fovea whereas among those with a rudimentary pit and the normal subjects, there was visible lengthening of the photoreceptor OS in the central fovea.
As the fovea develops, the IS and OS of the foveal cones become longer, and this process is thought to continue as long as 4 years after birth (Yuodelis & Hendrickson, 1986). As a metric of foveal cone specialization, we developed a quantitative approach to assess IS and OS maturation based on the high-resolution SD-OCT images. Shown in Figure 2 are images from one of the normal controls, specifying the identity of the different layers in the parafoveal (Fig. 2A) and foveal regions (Fig. 2B) on an SD-OCT scan. The IS can be clearly demarcated on the OCT scans as the distance between the ELM & IS/OS junction, while the OS length is given by the distance between the IS/OS junction and the OS/RPE interface (RPE1). We mapped the IS and OS photoreceptor lengths along the horizontal meridian for the 167 normal subjects and found a pronounced lengthening of the OS layer in the fovea relative to that in the parafovea (Fig. 2C). There was a slight increase in IS length in the fovea compared to that in the parafovea (Fig. 2D). At the fovea, the IS are 1.34 times longer than in the periphery while the OS are 1.87 times longer in our normal subjects (Fig. 2E). This lengthening is a well-described feature of the adult retina (Yuodelis & Hendrickson, 1986) and the magnitude of the lengthening observed here in vivo is consistent with that reported from histology (Hoang, Linsenmeier, Chung & Curcio, 2002). The subjects with albinism without a visible pit on SD-OCT had minimal to no foveal lengthening of the IS and OS whereas the individuals with a rudimentary pit had IS and OS lengthening that was within the normal range.
During foveal development there is also significant central migration of the cone photoreceptors, resulting in a decrease in their diameter and an increase in packing density (Diaz-Araya & Provis, 1992, Yuodelis & Hendrickson, 1986). We examined this by imaging the cone photoreceptor mosaic using an adaptive optics fundus camera. Images of the cone mosaic from 1-degree superior retina from a normal control (Fig. 3A) and two of the subjects are shown (Fig. 3B & C). An image from a subject with OCA1B (Fig. 3B) reveals a gradual decrease in cone packing density moving from left (inferior retina) to right (superior retina). A corresponding image from a subject with OA1 (Fig. 3C) reveals more uniform cone packing density. Colored strips below each image are Voronoi diagrams from the center of the corresponding image, which allows easier visualization of cone packing. The gradual decrease in cone packing for the normal control (Fig. 3A) and OCA1B subject (Fig. 3B) can be seen as an increase in the area of the Voronoi polygons, whereas the image from the OA1 subject (Fig. 3C) has more uniformly-sized polygons. We measured the percentage of cones with 6-sided (green) Voronoi domains as a metric of regularity (Baraas et al., 2007, Li & Roorda, 2007, Pum, Ahnelt & Grasl, 1990) and found that the normal subject had 57% (Fig. 3A), the OCA1B subject had 53% (Fig. 3B), and the OA1 subject had 53% (Fig. 3C). These are consistent with previously published values for the normal retina at this eccentricity (Baraas et al., 2007, Li & Roorda, 2007). Cone density values for all subjects are shown in Figure 3E, and there is variation in the degree of cone packing moving towards the fovea. Importantly, in two of the patients shown not to have significant OS lengthening (JC0125, AD0063), there was a progressive increase in cone packing toward the fovea. This appears to be consistent with earlier data that cone packing and OS lengthening are decoupled events, specifically that cone IS reach adult diameters before cone OS reach adult lengths (Yuodelis & Hendrickson, 1986). These measurements provide yet another level of resolution for quantifying foveal maturity with in vivo retinal imaging.
In one of the subjects with albinism (JC0140), we were able to visualize melanin pigment clumping throughout the central fovea (Fig. 4A-E). Pigment clumping was observed in the fundus photograph (see Supplementary Fig. 1 online). While in the normal retina there are intraretinal variations in cone reflectivity (Pallikaris, Williams & Hofer, 2003) and low frequency undulations in image intensity across adaptive optics images of the cone mosaic (Fig. 4F), this individual had discrete patches of hypo- and hyper-reflective retina. We posit that the hyper-reflective areas contain little or no melanin resulting in increased light scatter, while the hypo-reflective areas represent the melanin clumping. Interestingly, the cones within the brighter patches are variable in reflectivity, just as those within the darker areas are, suggesting that melanin distribution may not be the primary determinant to individual variation in cone reflectivity in AO images. It would be interesting to monitor how these pigmentary clumps change over time in this individual, as well as to examine other individuals with disruptions in melanin biosynthesis and distribution (such as Menkes disease, Wilson disease, Usher Syndrome type 1B, Hermansky-Pudlak syndrome; see King, Oetting, Summers, Creel & Hearing (2007) for a review) for possible retinal manifestations that may not have been previously recognized.
Our anatomic data obtained with AO & SD-OCT reflect the lack of consensus regarding the albinotic fovea in the literature. We have shown that foveal specialization can be measured in vivo on multiple levels – from morphology of the foveal pit to the length of the cone outer segment. We observed varying degrees of foveal maturity and cone specialization in the albinotic retina, which roughly align with schematics of normal foveal development based on post-mortem analyses (Isenberg, 1986, Mann, 1950, Provis et al., 1998, Springer, 1999). Thus we conclude that there is not a single picture that captures foveal morphology in albinism; rather there is a continuum of foveal maturity associated with the condition. This is likely based on a number of factors, including albinism subtype (OA versus OCA), the specific genetic mutation, and the natural pigment background of the patient. At first glance, the retinas from subjects with OCA1B tended to be more mature than those from the OA1 subjects. However no genetic analysis was done for this study, so more work needs to be done to determine the exact nature of any genotype-phenotype correlation within or across classifications of albinism.
Contrary to a recent report (Marmor, Choi, Zawadzki & Werner, 2008), none of our subjects with albinism had complete foveal cone specialization. This discrepancy may have to do with examining different albinism subtypes, though Marmor et al. did not provide any quantitative analysis so it is difficult to compare their results with our data. However, our data do support the idea that significant foveal cone specialization can occur in the absence of a foveal pit. In addition, Hendrickson, Djajadi, Erickson & Possin (2006) found that in anencephaly, foveal cone specialization (cone packing and elongation) occurred in the absence of a fully formed foveal pit. These data conflict with the “passive” model of foveal development, which predicts that in the absence of a foveal pit centripetal cone packing would not occur (Hendrickson, Troilo, Possin & Springer, 2006, Springer, 1999, Springer & Hendrickson, 2004a, Springer & Hendrickson, 2004b, Springer & Hendrickson, 2005).
A qualitative nomenclature is often used to describe the fovea in albinism and related conditions. Foveal hypoplasia literally refers to the absence of a pit, however it has been suggested that the negative functional connotations that accompany this term are reason for a different classification – fovea plana (Marmor et al., 2008). In light of our data, it is clear that neither of these qualitative terms universally applies to albinism and are inadequate descriptors to capture even the purely anatomical phenotype associated with albinism or other disorders where foveal pit morphology is compromised. Thus, we propose that a more quantitative description of the central retina be adopted, given the emerging availability of non-invasive high-resolution imaging tools. This is especially relevant given the significant variability observed among the albinotic fovea (Chong, Farsiu, Freedman, Sarin, Koreishi, Izatt & Toth, 2009, Harvey et al., 2006, Seo et al., 2007); foveal pit maturity and cone specialization are not all-or-none phenomena.
What limits vision in the albinotic visual system? For one, the abnormal retino-cortical projections in albinism (Guillery, Okoro & Witkop, 1975) presumably compromise stereo vision, though both the degree of chiasmal misrouting (Hoffmann, Tolhurst, Moore & Morland, 2003, von dem Hagen, Houston, Hoffmann & Morland, 2007) and reduction in stereo acuity (Lee et al., 2001) are variable. With regard to spatial vision, Wilson et al. (1988) proposed that “the deficits in central monocular spatial vision of albinos result from the increased spacing and (to a lesser extent) the decreased outer segment length of their central cones related to those of normal young adults.” This is supported by the finding that reduced cortical volume in a region of the brain that corresponds to the cortical representation of the central retina has been observed in albinism (von dem Hagen, Houston, Hoffman, Jeffery & Morland, 2005), presumably due to a reduced number of foveally-driven retinal ganglion cells. As we only recorded clinical acuity, it is difficult to draw any conclusions in our patients, however the ability to characterize cone density and OS length in vivo should enable further examination of this issue when paired with improved measures of acuity and contrast sensitivity.
In animal studies, it has been shown that the degree of retinal pigmentation is related to the extent of chiasmal misrouting (LaVail, Nixon & Sidman, 1978, Sanderson, Guillery & Shackleford, 1974), with lower pigmentation levels being associated with a greater shift in the line of decussation into temporal retina. This was recently also shown to be true in humans with albinism (von dem Hagen et al., 2007), and the authors proposed that if pigmentation were to determine both the extent of foveal hypoplasia and misrouting, then a correlation between acuity and misrouting would be predicted (von dem Hagen et al., 2007). In addition, the level of ocular pigmentation has been shown to affect the clinical features of albinism (Lee, King & Summers, 2001, Dorey, Neveu, Burton, Sloper & Holder, 2003). Such a correlations might help explain differences in vision across albinism subtypes, though further investigation is required. There is a strong relationship between foveal abnormalities and cortical disruption, given recent findings in albinism from MRI, fMRI, and VEP studies (Neveu, Von Dem Hagen, Morland & Jeffery, 2008, von dem Hagen et al., 2007, von dem Hagen et al., 2005, Dorey et al., 2003). However a deficiency of these studies is that no accompanying measures of foveal maturity or cone specialization were made. With the large degree of variation in cone specialization observed in our patients with albinism, it would be interesting to revisit these cortical studies in patients for whom the foveal architecture has been quantified in order to develop a more comprehensive model of reduced visual function in albinism.
Given that foveal development continues after birth, it should soon be possible to visualize with high temporal resolution the progression of the latter stages of foveal development in humans with some of the imaging modalities used here. In particular, handheld SD-OCT technology is now commercially available and has been used to image the infant retina under anesthesia (Chong et al., 2009, Scott, Farsiu, Enyedi, Wallace & Toth, 2009) and also in the awake state (Chavala, Farsiu, Maldonado, Wallace, Freedman & Toth, 2009, Maldonado, Izatt, Sarin, Wallace, Freedman, Cotten & Toth, 2010). As future treatments aimed at increasing melanin pigment production are likely to be more successful when applied earlier in development, such technology will be required to identify the most favorable candidates and to assess treatment response in these young retinas.
JC is the recipient of a Career Development Award from Research to Prevent Blindness. This study was supported by NIH Grants P30EY001931, T32EY014537, Fight for Sight, The E. Matilda Ziegler Foundation for the Blind, RD and Linda Peters Foundation, The Gene and Ruth Posner Foundation, Hope for Vision, unrestricted departmental grants to the Medical College of Wisconsin and University of Minnesota from Research to Prevent Blindness, and the Albinism & Related Disorders Research Fund at the Minnesota Medical Foundation. The authors thank Dr. Alfredo Dubra for sharing image registration software, Dr. Pooja Godara, Brett Schroeder, & Phyllis Summerfelt for assistance with retinal imaging, and Dr. Janice Burke, Dr. Thomas Connor, Jr., & Melissa Wagner-Schuman for helpful discussions. The authors thank the patients and their families for their time and assistance with the study.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.