«R91W mutation in Rpe65 leads to milder early-onset retinal dystrophy due to the generation of low levels of 11-cis-retinal Marijana Samardzija1,*, ...»
HMG Advance Access published October 12, 2007
R91W mutation in Rpe65 leads to milder early-onset retinal dystrophy due to the generation of low
levels of 11-cis-retinal
Marijana Samardzija1,*, Johannes von Lintig2, Naoyuki Tanimoto3, Vitus Oberhauser2, Markus
Thiersch1, Charlotte E. Remé1, Mathias Seeliger3, Christian Grimm1 and Andreas Wenzel1,§
Laboratory for Retinal Cell Biology, Dept. Ophthalmology, University of Zurich, Frauenklinikstr. 24,
8091 Zurich, Switzerland; Inst. of Biology I, Animal Physiology and Neurobiology, University of Downloaded from http://hmg.oxfordjournals.org/ by guest on November 24, 2016 Freiburg, Hauptstr. 1, D-79104 Freiburg, Germany; 3 Ocular Neurodegeneration Research Group, Inst. for Ophthalmic Research, University of Tuebingen, Schleichstr. 4/3, D-72076 Tuebingen, Germany * Address correspondence to: Laboratory for Retinal Cell Biology, Dept. Ophthalmology, University
Zurich, Frauenklinikstr. 24, 8091 Zurich, Switzerland. Tel.: 41-44-2553905; Fax: 41-44-2554385; E-mail:
© The Author 2007. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail: email@example.com 1
We generated R91W knock-in mice to understand the mechanism of retinal degeneration caused by this aberrant Rpe65 variant. We found that in contrast to Rpe65 null mice, low but substantial levels of both Downloaded from http://hmg.oxfordjournals.org/ by guest on November 24, 2016 RPE65 and 11-cis-retinal were present. While rod function was impaired already in young animals, cone function was less affected. Rhodopsin metabolism and photoreceptor morphology were disturbed, leading to a progressive loss of photoreceptor cells and retinal function. Thus, the consequences of the R91W mutation are clearly distinguishable from an Rpe65 null mutation as evidenced by the production of 11- cis-retinal and rhodopsin as well as by less severe morphological and functional disturbances at early age.
Taken together, the pathology in R91W knock-in mice mimics many aspects of the corresponding human blinding disease. Therefore, this mouse mutant provides a valuable animal model to test therapeutic concepts for patients affected by RPE65 missense mutations.
The vitamin A derivative 11-cis-retinal is the chromophore of rod and cone visual pigments. Absorption of light leads to an 11-cis to all-trans isomerization followed by dissociation of all-trans-retinal from the protein moiety (opsin) of the visual pigment holo-complex. The regeneration of the visual chromophore is a complex protein-mediated process termed the visual cycle (1). The crucial all-trans to 11-cis-retinoid isomerization reaction step takes place in the retinal pigment epithelium (RPE) and is catalyzed by RPE65 (2-4). Rpe65-deficient mice lack 11-cis-retinal and consequently no rhodopsin is detectable in their eyes
intermediates of the visual cycle (5).
More than 60 disease-associated mutations have been identified in the human RPE65 gene (summarized in http://www.retina-international.com/sci-news/rpe65mut.htm; see also (7)). The broad spectrum of mutations includes point mutations, splice-site defects, deletions, and insertions. Mutations in RPE65 are estimated to account for approximately 11% of all autosomal recessive childhood-onset retinal dystrophy cases (8). Patients suffering from mutations in the RPE65 gene are alternatively diagnosed as autosomal recessive Retinitis Pigmentosa (arRP), autosomal recessive retinal dystrophy, early-onset severe retinal dystrophy (EOSRD) or Leber’s congenital amaurosis type II (LCAII) (8-10). These rather diverse diagnoses reflect the phenotypic heterogeneity of the underlying disease. Most patients suffering from mutations in the RPE65 gene are diagnosed early in childhood as severely visually impaired, with night blindness, distinctly restricted visual fields and nystagmus, but no photophobia (summarized in (11)).
Retinal function (ERG) is undetectable after dark adaptation and is severely impaired following light adaptation. However, many of these patients have sufficient vision to attend elementary school. Vision is then gradually lost, resulting in blindness almost invariably in the third decade of life (11).
Two naturally occurring models (rd12 mouse and Briard dog) and the genetically engineered Rpe65 knock-out mouse (Rpe65-/-) have been useful for the delineation of RPE65 protein function in the visual cycle (5, 12, 13). These models have meanwhile been used in pre-clinical gene replacement therapy
All of the above models represent a “null situation” for RPE65, in which the visual cycle has never been functional. However, current literature indicates that more than 50% of RPE65 mutations in patients are missense mutations. Some of these missense mutations presumably produce mutant versions of RPE65 with some residual enzymatic activity. Recently, we characterized three consanguineous families carrying the R91W mutation in RPE65 (23). All affected family members had useful cone-mediated vision in the first decade of life, suggesting that the mutant RPE65 protein is expressed and possesses residual function. The consequences of a partial loss of RPE65 activity caused by a missense mutation have not
under such conditions.
To assess this specific pathology we generated the Rpe65-R91W knock-in mouse as a model for the human disease and analyzed the effect of the mutation on retinal function, visual cycle and morphology.
RESULTS Generation of R91W knock-in mice. Gene targeting in mouse ES cells was used to modify exon 4 of the Rpe65 gene such that codon 91 changed from arginine to tryptophan (R91W) (Fig. 1A). In humans, the R91W mutation is caused by a single point mutation (TGATGG). However, in mice arginine 91 is encoded by a CGA codon, therefore two point mutations were introduced into codon 91 (CGATGG).
The introduction of these two mutations resulted in the loss of a TaqI restriction site, facilitating the genotype analysis (Fig. 1A-C).
The gene targeting strategy is shown in Fig. 1B. In addition to mouse Rpe65 genomic DNA carrying the R91W mutation, the targeting vector contained a floxed neo resistance and a DT cassette as selection markers. Sequencing of the full genomic DNA insert of the targeting vector confirmed the presence of the R91W mutation and the selection markers in an otherwise wild-type sequence (Fig. 1A). The linearized construct was electroporated into coisogenic TC1 ES cells, and correctly targeted ES cell-clones were
(Fig. 1C and data not shown).
These clones were used to generate germ-line competent chimeric mice. A chimeric male was mated to coisogenic 129S6 females to propagate the line. The resulting heterozygous (Rpe65R91Wneo) F1 mice were bred with a germ-line Cre-deleter mouse line (129S6-Tg(Prnp-GFP/Cre)1Blw/J)) (24) to excise the neo cassette (Fig. 1B). The resulting offspring was heterozygous for the R91W mutation (R91W/wt). The only foreign sequence in addition to the R91W substitution was a single intronic loxP site. Finally, we interbred the heterozygous R91W/wt mice to obtain pure homozygous 129S6/SvEvTac-Rpe65tm1Lrcb
and the above mentioned loxP site.
Rpe65 expression in mutant mice. Quantitative RT-PCR, immunoblotting and immunohistochemistry showed that the mutant protein is expressed in R91W knock-in but not in Rpe65-/-mice (Fig. 2A-C).
While the expression of Rpe65 mRNA was only slightly reduced in R91W mice (Fig. 2A), immunoblotting revealed that the steady state protein levels of the RPE65R91W mutant variant were reduced by 95% as compared to wild-type mice (Fig. 2B and data not shown). The mutant protein was detected in the RPE and thus had correctly localized (Fig. 2C).
Lecithin retinol acyltransferase (LRAT) and cellular retinaldehyde binding protein (CRALBP) are functionally connected to RPE65 and the visual cycle (25-28). Expression of both proteins was comparable in eyecups of 8 week-old wild-type, R91W knock-in and Rpe65-/- animals with a tendency of increased LRAT levels in knock-out mice (Fig. 2B). Levels of rod opsin were slightly reduced in retinas of R91W mice and – as previously reported (5) – in retinas of Rpe65-/- animals.
Retinoid analysis. The absence of RPE65 results in the arrest of the visual cycle, causing an accumulation of retinyl esters in the RPE and visual chromophore deficiency (5). Based on this observation, we expected that the dramatically reduced protein levels of the RPE65R91W variant (Fig. 2B)
dark-adapted wild-type, R91W/wt heterozygous, R91W homozygous and Rpe65-/- mice at different ages.
HPLC analysis of whole eye preparations of 4, 8, 12, 24 and 40 week-old animals demonstrated the presence of 11-cis-retinal in R91W homozygous mice, though at low levels between 2.5% and 6.3% as compared to age-matched wild-type animals (Table 2). As expected, no 11-cis-retinal was detectable in Rpe65-/- mice at any age (Table 2 and (5, 29)). All-trans-retinal, all-trans-retinol and 9-cis-retinal levels were comparable between R91W and Rpe65-/- mice. The retinoid content and composition in R91W/wt heterozygous animals did not differ from wild-type mice at all tested ages (Table 2).
Nevertheless, we detected a strong and almost linear (13 pmol/day; R2=0.93) accumulation of retinyl esters during the first 24 weeks of life. Between 24 and 40 weeks no further increase was detected (2186 and 2108 pmol/eye, respectively; Table 2) suggesting that a plateau was reached in R91W mice. In Rpe65-/- animals retinyl ester levels increased linearly (9 pmol/day; R2=0.96) throughout 40 weeks (Table 2). Electron microscopy revealed that this accumulation of retinyl esters was accompanied by the formation of lipid droplets in the RPE of R91W and, as previously reported, in Rpe65-/- mice (5) (Fig. 3).
Rhodopsin content and regeneration. We next determined dark-adapted rhodopsin levels, which reached 30 pmol/retina (Fig. 4A) in R91W mice representing 6% of wild-type rhodopsin. This percentage was comparable to the amount of 11-cis-retinal observed in 4 week-old R91W mice, which corresponded likewise to 6% of wild-type levels (Table 2).
The expression levels of RPE65 protein govern the kinetics of rhodopsin regeneration (30, 31). Given the reduced amount of RPE65 in R91W mice (see Fig. 2), we expected that rhodopsin would regenerate with a much slower rate. To allow a maximal rhodopsin production, we kept R91W mice in darkness for 4, 10 and 22 days. By the time of analysis the mice were 6-7 weeks old. Maximal rhodopsin levels detected after 22 days in darkness were 42 pmol, which was still less than 10% of wild-type animals (Fig.
4A, see Fig. 4C for the wild-type levels).
exposed for 10 min to 5’000 lux, which bleaches 90% of rhodopsin in wild-type mice (27). Animals were returned to darkness to allow regeneration of the visual pigment. Non-exposed R91W mice contained 27 pmol of rhodopsin (Fig. 4B), which was bleached by light exposure, reducing it to 1 pmol. Even after prolonged dark-adaptation (5 days) only 44% (12 pmol) of the pre-bleach value (27 pmol) was detectable.
Thus, following a strong bleach, regeneration of rhodopsin was inefficient and reached a plateau at about half the maximal level of non-bleached, dark-adapted animals. Maximal dark-adapted rhodopsin levels were similar in R91W/wt heterozygous (538 pmol/retina) and wild-type (560 pmol/retina) (Fig. 4C). The
in R91W/wt mice, indicating slightly slower rhodopsin regeneration kinetics in the heterozygous situation.
Photoreceptor function in R91W mutant mice. Next, we studied the consequences of the reduced 11cis-retinal levels on retinal function. Scotopic ERG recordings revealed that higher flash intensities were needed to induce an electrical response in R91W as compared to wild-type mice (Fig. 5A top panel). By assessing the luminance required to generate a half maximal b-wave amplitude (32), the reduction in light sensitivity in 8 weeks-old R91W mice was determined to be approximately 2.5 log units as compared to wild-type. Therefore, R91W mice are about one log unit more sensitive to light than Rpe65-/- animals (6).
In age series testing of 8, 12, 24 and 40 week-old animals the sensitivity threshold remained unaltered, while the b-wave amplitude was reduced with increasing age in R91W, this being especially prominent between 12 and 24 weeks (Fig. 5A top and bottom panels). From 24 to 40 weeks of age no further reduction in b-wave amplitude was detected. Thus, while the sensitivity of the retina was not prone to age-related reduction, the maximal response size was.
In order to test cone function, ERG responses were recorded under photopic conditions (Fig. 5B).
Notably, there was no difference in threshold sensitivity or in amplitude of the b-wave between 8 weekold R91W and wild-type control animals (Fig. 5B top and bottom panels). Given the low amount of
waveforms were altered in a way similar to that observed in Rpe65 null mice (6), indicating that rods are active under those conditions. Therefore, it appears that the photopic signals are mixed responses containing both rod and cone system components.
A comparison of wild-type and R91W/wt heterozygous animals resulted, as expected, in no difference between the genotypes under scotopic or photopic conditions (not shown).
Assessment of retinal morphology. The results of the tests for retinal function, both scotopic and