Mouse Models of RP59: Comparison
Please note this is a comparison between Version 1 by Steven J. Fliesler and Version 3 by Conner Chen.

Retinitis pigmentosa-59 (RP59) is a rare, recessive form of RP, caused by mutations in the gene encoding DHDDS (dehydrodolichyl diphosphate synthase). DHDDS forms a heterotetrameric complex with Nogo-B receptor (NgBR; gene NUS1) to form a cis-prenyltransferase (CPT) enzyme complex, which is required for the synthesis of dolichol, which in turn is required for protein N-glycosylation as well as other glycosylation reactions in eukaryotic cells. The mouse is the most commonly used vertebrate animal model in experimental biology, there have been efforts to generate murine models of RP59.

  • cis-prenyltransferase
  • DHDDS
  • dolichol
  • nogo-B receptor
  • RP59

1. Mouse Models of RP59

The mouse is the most commonly used vertebrate animal model in experimental biology, including in studies relevant to eye and vision research [1][2][48,49]. The mouse retina is a “duplex” retina (i.e., possesses both rod and cone photoreceptors), but the rods heavily dominate the photoreceptor population (ca. 97%), while the cones are far less numerous. Additionally, unlike the human or nonhuman primate retina, the mouse retina lacks a cone-rich macula or fovea. Nonetheless, the fundamental cell biology and physiological processes that are extant in the mouse and human retina are quite comparable. Further, DHDDS (dehydrodolichyl diphosphate synthase) is s largely conserved in mammals; specifically, murine and human DHDDS exhibit 92.4% sequence identity with a 100% query coverage (E = 0.0) (using BLASTP algorithm) [3][4][5][13,50,51]. Hence, there have been efforts to generate murine models of retinitis pigmentosa-59 (RP59), using both knock-in and knock-out genetic manipulation of the Dhdds gene.

2. K42E Dhdds Knock-In Mouse

Using CRISPR-Cas9 gene-editing technology, a viable mouse model of RP59 has been generated (on a C57BL/6J background), and the ocular phenotype of mice either heterozygous (DhddsK42E/+) or homozygous (DhddsK42E/K42E) for this mutation has been reported [6][52]. The morphological organization of the retina of Dhdds mutants and age-matched wildtype (WT) controls was evaluated in vivo, using spectral domain-optical coherence tomography (SD-OCT), a non-invasive and quantitative analytical method. Surprisingly, no obvious structural abnormalities were observed as a function of Dhdds mutation, at least up to postnatal (PN) 12 months. Quantification of the total thickness of the neural retina as well as the thickness of the outer nuclear layer (ONL), the latter being specifically relevant to the health and persistence of photoreceptors, was performed to assess retinal structure. No overt differences in these quantifiable parameters were observed, comparing mutant and age-matched wild type (WT) mice. Hence, unlike human RP59, this mouse model did not appear to exhibit obvious retinal degeneration, even up to one year of age.
Retinal frozen sections were probed with antibodies against glial fibrillary acidic protein (GFAP; a biomarker of astrocytes and glia) and opsin (the visual pigment apoprotein; a biomarker for rod photoreceptors, especially the outer segments (OS)).
Despite the seemingly normal appearance of the retina, anti-GFAP immunoreactivity was markedly elevated in DhddsK42E/K42E mice, relative to controls, and the radial labeling pattern reaching from the internal limiting membrane (ILM, the vitreoretinal interface) to the ONL was indicative of gliotic reactivity. However, opsin immunolocalization was comparable in mutant and control retinas, and there was no evidence of mislocalization of opsin (i.e., opsin was almost exclusively localized to the rod OS, rather than being partially distributed along the plasma membrane of the rod inner segment or down the axonal process or to the synaptic ending, as is commonly observed when rods undergo degeneration). Additionally, concanavalin-A (Con-A) lectin cytochemistry, with and without PNGase-F treatment, was performed to assess protein glycosylation status of DhddsK42E/K42E and WT mice. [PNGase-F treatment was done to simulate the scenario of a glycosylation defect, and to demonstrate that the lectin cytochemistry approach could detect such a defect, if present, in the mutant retinas.] Con-A labeling and PNGase-F sensitivity were comparable in mutant and control retinas. Hence, there was no evidence of globally compromised protein N-glycosylation in homozygous K42E Dhdds mutant retinas.
That initial report did not include any electrophysiological (electroretinography, ERG) analysis. However, subsequently, it has been discovered that this knock-in mouse model of RP59 exhibits initially subtle and then progressively more marked ERG defects. Starting at about PN 1 month of age, there are progressive reductions in the dark-adapted (scotopic; rod-driven) and light-adapted (photopic; cone-driven) ERG b-wave amplitudes, while the a-wave amplitudes exhibit no significant reductions [7][53], i.e., a “negative b-wave”. Those findings suggest there is defective transmission of visual information from rod and cone photoreceptors to their respective bipolar cell populations. Additionally, a more recent study has further evaluated the presence of structural abnormalities in the retinal pigment epithelium (RPE) and in the inner retina of the homozygous K42E Dhdds mutant mouse at PN 18 months, compared to age-matched C57BL/6J (WT) mice [8][54]. Grossly, retinal histology was comparable in both the mutant and WT mice. However, ectopic rod photoreceptor nuclei were found intermittently in the OPL of mutant retinas, and second-order neuronal processes were reduced, especially in the periphery. Pyknotic nuclei also were observed in the outer and inner nuclear layers (ONL, INL), and TUNEL-positive cells (consistent with apoptotic cell death) were far more numerous in retinas of the mutant than in WT mice. Ultrastructural features of the OPL and INL were mostly comparable in both mutant and WT retinas, however, some photoreceptor cell bodies and their synaptic terminals displayed darkened cellular material (consistent with impending cell death). In addition, the RPE basal infoldings adjacent to Bruch’s membrane often were disorganized and, on occasion, moderately to severely degenerated RPE cells were observed in K42E mutant mice. Hence, despite the grossly normal appearance of the mutant retina, these results suggest that defective signal transmission between photoreceptors and inner retinal neurons, as well as RPE dysfunction and compromised viability may be significant contributors to the etiology of RP59.
In a preliminary study (L. Surmacz and E. Swiezewska, unpublished results), it was found that this global K42E Dhdds knock-in mouse is still competent to synthesize dolichols; however, the isoprenylog isoforms in retina, liver, and brain have shorter than normal chain lengths: Dol-17 predominates (Dol-17 >> Dol-18 >> Dol-19), whereas in controls the dominant species are Dol-18 and Dol-19. In fact, the total amount of dolichol in tissues from homozygous K42E mutant mice is greater than (rather than less than) normal. This finding of a shift to shorter chain length dolichol species in all tested tissues is consistent with what has been reported for human RP59 patient plasma and urinary dolichol profile, where the Dol-18/Dol-19 ratio from DHDDSK42E/K42E and DHDDST206A/K42E patients is ca. 3, while for unaffected controls the ratio is ca. 1, and for heterozygotes it is ca. 1.5 [9][7]. In fact, the dolichol isoform profile now is recognized as a useful companion diagnostic tool for a range of CDGs (congenital disorder of glycosylations) [10][27]. However, analysis of dolichol chain length and total dolichol content has not been performed on tissue biopsies of RP59 patients, nor has it been demonstrated that RP59 patient tissues or bodily fluids exhibit decreased levels or loss of total dolichol content compared to unaffected human subjects. The role of shortened dolichol chain length and/or dolichol content in RP59-associated retinal degeneration remains to be investigated.

3. Rod-Specific Dhdds Knockout Mouse

Using Cre-lox technology, mice that express Cre recombinase under the control of the rod opsin promoter (Rho-iCre75; [11][55]) were mated with Dhddsflx/flx mice, harboring loxP sites flanking Dhdds exon 3, (both on a C57BL/6J background) to generate mice that had Dhdds ablated selectively in rod photoreceptor cells, starting at PN day 7 [12][56]. This approach more closely models the degenerative effects of RP59-associated severe mutations with expected null DHDDS activity (e.g., W64X) [3][12][13][14][6,13,14,56]. At PN 4 weeks of age (allowing sufficient time for complete maturation of the retina), the structure of retinas of Dhdds knockout mice was comparable to that of age-matched control mice (the latter being Dhddsflx/flx iCre mice, rather than WT), as determined by SD-OCT as well as correlative histological analysis; yet, there were subtle, but measurable, ERG deficits in the mutant retinas (predominantly in scotopic a-wave amplitudes), compared to age-matched controls.
By PN 5 weeks of age, however, about 50% of the photoreceptors had died and dropped out, rod outer segment (OS) lengths were dramatically reduced, and the ERG deficits (scotopic and photopic) were comparably profound. By PN 6 weeks of age, there were few if any remnant photoreceptor cells and ERG responses were nearly extinguished. Importantly, at PN 4 weeks (i.e., prior to any obvious histological defects in the retina were manifest), the total dolichol content of the neural retina was decreased by ca. 50% in mutant mice, compared to age-matched controls. Additionally, at this time point, the ONL in mutant mouse retinas was devoid of Dhdds mRNA, as detected by in situ hybridization, whereas mRNA content of the INL in both mutant and control retinas was comparable (serving as an internal control), thus validating the cell type-specificity and efficiency of the Dhdds ablation. In addition, at PN 5 weeks, the rod-specific Dhdds knockout retina exhibited marked gliotic reactivity, as evidenced by dramatically elevated GFAP levels, measured by Western blot analysis as well as immunofluorescence confocal microscopy. There also were signs of a localized neuroinflammatory process and phagoptosis, as Iba-positive cells (consistent with activated microglia) were observed to invade the outer retina, some of which were seen to engulf TUNEL-positive (dying/dead) photoreceptor nuclei, and the levels of ICAM (an inflammatory cytokine) increased by >5-fold in mutant retinas, relative to age-matched controls. In addition, as observed in the DhddsK42E/K42E mouse model, lectin cytochemical analysis revealed the lack of any obvious protein N-glycosylation defect in the retina of this rod-specific Dhdds knockout model. Additionally, Western blot analysis of mutant vs. control retinas showed no glycosylation defect either in rod opsin (the most prominent glycoprotein in the vertebrate retina [15][57]) or in LAMP2 (a glycoprotein biomarker for lysosomal membranes [16][58]).
In summary, selective ablation of Dhdds in rod photoreceptors results in a rapid, severe, and irreversible retinal degeneration, primarily involving the outer retina (photoreceptor layer). This would be expected to result in prevention of dolichol synthesis in rod photoreceptors and, in turn, a marked loss of dolichol content of the neural retina. Notably, this scenario is quite different from what occurs in the homozygous DhddsK42E/K42E mutant retina (see above).

4. RPE-Specific Dhdds Knockout Mouse

Using the same general strategy as employed to generate the rod-specific Dhdds knockout model, homozygous floxed Dhdds mice were mated with a Cre recombinase mouse line under the control of the VMD2 (vitelliform macular degeneration 2) promoter (both lines on a C57BL/6J background), to ablate Dhdds selectively in RPE cells [17][59]. Although the primary defect was localized to the RPE, there were concomitant morphological abnormalities evident in the neural retina as well, including a progressive retinal degeneration, cell loss, and thinning, apparent initially at about PN 1 month. By PN 3 months, pathological features were evident in the RPE and photoreceptor cells, although non-uniformly, across the retina. RPE cells were observed ectopically in the photoreceptor layer, there was patchy loss of photoreceptor cells, and the external limiting membrane (ELM) descended toward and abutted the RPE. Consistent with this marked retinal degeneration, progressively worsening scotopic and photopic ERG deficits were observed at PN 1, 2, and 3 months. Unexpectedly, however, electrophysiological defects also were observed (although substantially less severe) in heterozygous Dhdds mutants. [Note: typically, recessive diseases, by definition, do not manifest in heterozygotes.] This observation suggests the possibility of a functional change in the cis-prenyltransferase (CPT) enzyme complex that occurs prior to any obvious retinal structural changes and suggests that 50% CPT activity may be insufficient to fulfill its biological function in the RPE. Additionally, this finding predicts that carriers of Dhdds mutations also may develop visual dysfunction, depending on the nature of the mutation and other factors, such as genetic background and environment.
Given the multiple essential functions that the RPE plays in supporting the physiological health of the neural retina [18][60], it is not surprising that a primary molecular defect such as Dhdds ablation in the RPE would eventually result in compromising the health of the underlying neural retina. These results suggest that RPE dysfunction likely contributes significantly to the observed DHDDS mutation-initiated pathology in RP59, and that the underlying disease mechanism may transcend simple disruption of glycosylation.

5. Nogo-B Receptor Mutants as RP59 Models

Homozygous NgBR (Nus1) knockout in mice results in early embryonic (E6.5 or earlier) lethality, while heterozygotes are viable and do not exhibit a pathological phenotype [19][61]. Humans harboring a R290H Nus1 mutation exhibit profound musculoskeletal and neurological pathology, as well as macular lesions and other visual system defects [19][61]; however, this disease is distinct from RP59. Additionally, fibroblasts from patients harboring the R290H NUS1 mutation have been shown to have defective protein glycosylation [19][61]. To date, there have been no reports of Nus1 mouse mutants as RP59 models.

6. Emerging New Mouse Models of RP59

Another DHDDS mutation identified in RP59 patients is T206A [20][21][4,5]. This has only been reported thus far as a compound heterozygous mutation with K42E in patients. A recent preliminary report [22][62] has described the generation and initial characterization of new RP59 mouse models, consisting of targeted knock-in of T206A/T206A (homozygous), T206A/+ (heterozygous), and compound heterozygous T206A/K42E Dhdds mutations. ONL thickness and total neural retinal thickness measurements were comparable to WT control values for all of these Dhdds mouse mutant lines (by SD-OCT). INL thickness in all homozygous mutants, however, were markedly reduced. ERG a-wave amplitudes (scotopic and photopic) were comparable to WT values, but the ERG b-/a- wave amplitude ratios at saturating flash intensities (scotopic and photopic) were significantly lower than WT values for T206A/T206A and T206A/K42E Dhdds mutants (i.e., a “negative b-wave” phenotype). Additionally, the photopic b-/a- wave ratio difference was greater for T206A/K42E than for T206A/T206A mutants. T206A/+ mutant amplitude ratio values were comparable to WT control values. These results are consistent with (and extend) those obtained with homozygous K42E/K42E Dhdds mutant mice (K42E INL measurements have not been reported) and implicate defective photoreceptor-to-bipolar cell synaptic transmission in these RP59 mouse models. They also suggest that the K42E mutation is more strongly pathological than is the T206A mutation. To date, a knock-in R98W Dhdds mutant mouse RP59 model has not been reported.

7. Retinal Degeneration in a Drosophila Dhdds Knockdown Model

Although not a vertebrate, it should be noted that a Drosophila model of RP59 has been generated and characterized [23][63]. Drosophila offers a tractable model system readily amenable to genetic manipulation, and it contains a genetic ortholog (CG10778) to the DHDDS gene. Targeted RNAi-mediated knockdown of this gene was embryonic lethal. However, targeted expression of CG10778-RNAi using the glass multiple reporter (GMR)-Gal4 driver (GMR-DHDDS-RNAi) in the eye disc and pupal retina at the larval stage caused an unusual retinal degeneration phenotype. Photoreceptors R2 and R5 exhibited a nearly normal rhabdomere structure (the invertebrate counterpart to the “outer segment” of vertebrate photoreceptors), yet exhibited cytopathological features in the nuclear region, whereas other photoreceptors exhibited retinal degeneration in all regions. Additionally, rhodopsin levels were dramatically reduced in mutant vs. wildtype flies, while there was massive amplification (accumulation) of endoplasmic reticulum (ER) membranes in the photoreceptors. These results indicate that the CG10778 gene product is essential for the normal development of the Drosophila retina. By extension, despite the known, considerable differences in retinal architecture between flies and humans, the results suggest that DHDDS may be essential for the normal development of the vertebrate retina.
Video Production Service